ABSTRACT Title of Thesis: IDENTIFYING HIGHLY CONSERVED PATHOGENICITY GENES IN CHESTNUT BLIGHT AND POWDERY MILDEW FUNGI AS TARGETS FOR NOVEL FORMS OF HOST RESISTANCE Bruce J. Levine, Master of Science, 2019 Thesis Directed By: Dr. Shunyuan Xiao, Professor, Department of Plant Science and Landscape Architecture A bioinformatic search of the genomes of chestnut blight fungus, Cryphonectria parasitica (Cp), and the Arabidopsis powdery mildew fungus, Golovinomyces cichoracearum (Gc), yielded six suspected pathogenicity genes with homologues in both species. Deletion of these genes by homologous gene replacement was attempted in Cp, with one success, TG4. The TG4-knockout strain showed changes in phenotype and reduced fungal virulence against chestnut. TG4 appears to be a promising target for host-induced gene silencing (HIGS) in transgenic American chestnut. The use of homologues from genetically tractable species like Cp can help overcome the obstacles to performing reverse genetics on intractable, biotrophic fungi such as Gc. Experiments underway involving the silencing and ectopic overexpression of the Gc homologues of the target genes provide a rapid method to study Cp genes, including to screen additional candidate genes as future targets for HIGS. IDENTIFYING HIGHLY CONSERVED PATHOGENICITY GENES IN CHESTNUT BLIGHT AND POWDERY MILDEW FUNGI AS TARGETS FOR NOVEL FORMS OF HOST RESISTANCE by Bruce J. Levine Thesis submitted to the Faculty of the Graduate School of the University of Maryland, College Park, in partial fulfillment of the requirements for the degree of Master of Science 2019 Advisory Committee: Professor Shunyuan Xiao, Chair Dr. Jianhua Zhu Dr. John H. Payne Dr. Dongxiu Zhang © Copyright by Bruce J. Levine 2019 Dedication To my wife, Laura Goertzel, who, when I said “I think I want to quit my job at age, 52, even though our children are still young, and go get a degree in a subject in which I have no formal background, “ replied, “I think you should.” Thanks for your support, then, now and forever. ii Acknowledgements This thesis, and the path of discovery that it describes would not have been possible without the help of many others. Dr. Shunyuan Xiao took me on and accommodated my very particular interest in chestnut blight. He helped me craft and carry out a research plan that would potentially contribute to the restoration of the American chestnut, while also building on the Xiao lab’s work mapping and characterizing the molecular interactions between plants and their fungal pathogens. I am very grateful to my fellow lab members, Qiong Zhang, Harley King, Xianfeng Ma, Yuheng Yang, Lili Wang, Xiaoting Chen, and Frank Coker, and especially to Ying Wu for her work on selecting the target genes, and other bioinformatics aspects of the project. My advisory committee members, Dr. Jianhua Zhu, Dr. John Payne, and Dr. Dongxiu Zhang, provided support, encouragement and critically important advice over the course of my research. My fellow graduate students in the Department of Plant Science and Landscape Architecture were extremely helpful in critiquing my research along the way, and providing useful advice and suggestions, from how to get help on statistics to techniques in fungal culture. Throughout my degree program, I feel I learned as much from my fellow students as from my instructors. iii The American Chestnut Foundation, especially the Maryland chapter, provided many of the materials and a good deal of labor, necessary to make this project work. Chestnut scientists Dr. Donald Nuss, Dr. Mark Double, Dr. John Carlson, Dr. Jared Westbrook, Dr. Kim Steiner, Dr. Hill Craddock, Tom Saielli, Sara Fitzsimmons, Dr. Angus Dawe provided me with critically important technical advice drawn from their vast experience. Among the many scientists who advised me, my most heartfelt thanks go to Dr. Fred Hebard, Chief Scientist Emeritus of the American Chestnut Foundation, who has indulged me in answering random questions about chestnut for over 20 years, since long before either of us knew I would be playing scientist for real. His generosity in sharing his time and knowledge, and the joy he so clearly derives from his life’s work, are a great part of what inspired me to take this path myself. iv Table of Contents Dedication …………………………………………………………………………... ii Acknowledgements ……………………………………………………………...…. iii Table of Contents …………………………………………………………………….v List of Tables ……………………………………………………………………….vii List of Figures ………………………………………………………………………viii List of Abbreviations ………………………………………………………………...ix Chapter 1: Research Narrative ………………………………………………………..1 Introduction …………………………………………………………………...1 Background and Literature Review …………………………………………..2 Interactions between plants and pathogenic fungi ……...............2 The demise of American chestnut as a case study ……………………5 Rationale and Significance of this Research ………………………………12 The growing importance of molecular genetics………………………12 The Arabidopsis-powdery mildew model pathosystem ……………. 13 The merits of the chestnut-Cp pathosystem as model …..………….15 Defining pathogenicity broadly for target gene selection ………...…16 Research Objectives …………………………..…………………………….18 Chapter 2: Genetic Study of six Highly-Conserved Fungal Genes in Cp………….20 Introduction ………………………………………………………………….20 Materials and Methods ………………………………………………………20 Identification of target genes ………………………………………20 Creation of gene disruption constructions …………………………21 Spheroplast preparation ……………………………………………26 Spheroplast transformation ………………………………………...29 Regeneration of putative transformants …………………………….29 Isolation of monokaryon knockout strain mycelia ………………….31 Examination of knockout strain phenotype………………………33 Examination of knockout strain virulence in planta ………………...34 Detached stem assay ……………………………………… 35 Small stem assay ………………………………………….. 37 Results and Discussion …………………………………………………….. 39 Profiles of target genes ……………………………………………...39 TG1: a putative autophagy-related protein………………... 41 TG2: a putative 60s ribosomal subunit P2 acidic protein……43 TG3: a putative cell wall mannoprotein ……………………..44 TG4: a putative pre-protein translocase ……………………47 TG5: a putative mitochondrial carrier protein ………………49 TG6: a possible effector protein with an ML-domain ………50 The Generation and Characterization of Knockout Strains …………..53 Multiple attempts at transformation …………………………53 Use of a novel medium to induce spore germination ……….55 Observations of the TG4 knockout strain in vitro …………58 Detached stem assay of the TG4 knockout strain ......... 64 Small stem assay of the TG4 knockout strain ………………. 65 v Chapter 3: Conclusions, Reflections and Future Directions ………………………67 General conclusions ………………………………………………………..67 Specific conclusions concerning the role of TG4 in Cp ……………………69 Further research and future directions ………………………………………71 Reflections on methodology ………………………………………………75 Appendix I: Fungal Growth Media Used in this Study …………………….…83 Appendix II: Spheroplast Preparation, Transformation and Regeneration ……….86 Bibliography ……………………………………………………………………….91 vi List of Tables 1. Cryphonectria parasitica genes known to be involved in fungal virulence 2. Primers sequences and melting points used in the creation of gene disruption constructs 3. PCR and other amplification methods used in the generation of gene disruption constructs 4. Cryphonectria parasitica genes targeted in this study 5. Partial or full knockout cultures created in this study 6. Result of in vitro assays of DK80 and dTG4A-8 vii List of Figures 1. Diagram of homologous gene replacement by overlapping PCR 2. Cryphonectria parasitica spheroplasts as seen in a hemocytometer 3. Cryphonectria parasitica fruiting bodies oozing spores 4. Formula for normalized canker length 5. Decision tree for assignment of qualitative scores for chestnut blight cankers 6. Gel showing fragment length of DK80 and TG4 knockout strains 7. Mycelial diameter of DK80 and dTG4A-8 at 7 dpi, in vitro 8. Dry weight of DK80 and dTG4A-8 mycelia at 7 dpi, in vitro 9. DK80 and dTG4A-8 in vitro assay on three types of media 10. Difference in colony morphology between DK80 and dTG4A-8 11. DK80 and dTG4A-8 cultures grown in darkness 12. Comparison of DK80, dTG4A-8 and control detached stem inoculations at 24 dpi 13. Mean normalized canker lengths in detached stem assay, 24 dpi 14. Mean days of survival in small stem assays, 98 dpi 15. Cryphonectria parasitica strains show varying degrees of hygromycin tolerance 16. Parasitic Penicillium colonies growing on Cryphonectria parasitica hyphae 17. Comparison of repeatedly subcultured and fresh DK80 cultures viii List of Abbreviations APHIS Animal and Plant Health Inspection Service CBA Chestnut bark agar CHV Chestnut hypovirus CIM Chestnut induction medium Cp Cryphonectria parasitica DLP Dicer-like protein dsRNA Double-stranded RNA EHM Extra-haustorial membrane EPC Endothia parasitica complete ER Endoplasmic reticulum ET Ethylene ETI Effector-triggered immunity Gc Golovinomyces cichoracearum GFP Green fluorescent protein HIGS Host-induced gene silencing hph Hygromycin phosphotransferase HR Hypersensitive response IBBR Institute for Bioscience and Biotechnology Research JA Jasmonic acid mRNA Messenger RNA NCL Normalized canker length PDA Potato dextrose agar ix PM Powdery mildew PTI PAMP-triggered immunity RNAi RNA interference SA Salicylic acid siRNA Small interfering RNA RISC RNA-induced silencing complex WA Water-agar x Chapter 1: Research Narrative Introduction The long-term goal of this thesis research project is to expand our understanding of the molecular interactions between plants and plant-pathogenic fungi for developing host resistance. We sought to take advantage of two distinct pathosystems to identify key fungal genes, characterize their potential roles in pathogenesis, and explore host- induced gene silencing (HIGS) to develop antifungal resistance in plants. The first pathosystem is the interaction between American chestnut (Castanea dentata) and a necrotrphic pathogen Cryphonectria parasitica (Cp) which results in the devastating chestnut blight disease. The second pathosystem is the interaction between Arabidopsis thaliana and a biotrophic pathogen, Golovinomyces cichoracearum (Gc), one species which leads to a common disease called powdery mildew. Specifically, homologous recombination-based gene replacement will be used to delete conserved fungal genes in Cp to investigate their potential role in fungal survival and pathogenesis. Time permitting, homologous Gc genes encoding candidate secreted effector proteins (CSEPs) may be expressed in Arabidopsis to test if they can suppress host immunity. Candidate fungal genes identified will be targeted in Gc via HIGS using Arabidopsis. Finally, HIGS would be deployed to target essential Cp genes identified to engineer Cp resistance in American chestnut 1 Background and Literature Review Interactions between plants and pathogenic fungi The study of plant disease is as old as agriculture itself. For thousands of years, humanity has tried to prevent or mitigate the effects of disease on crops through cultural practices and the artificial selection of resistant cultivars. Only recently however, have we begun to understand the biological, biomolecular and genetic basis of plant disease resistance and susceptibility. Key discoveries since the 1950s have made it possible to discover the details about how plants defend themselves from pathogens, and how pathogens can overcome plant defenses in order to parasitize their hosts (Bent 2018). More recent advances in genetics and molecular biology have given us new tools to explore the nature of plant disease more deeply, and continually help us refine our understanding of plant disease. A certain degree of basal resistance to fungi, bacteria, oomycetes, nematodes, (as well as to herbivores and abiotic stress) is found in all plants. In nature, most plants are resistant to most potential microbial pathogens thanks to the existence of robust preformed and inducible defense barriers (Bent 2018). Preformed (or constitutive) defense barriers of plants include the cellulosic plant cell wall, surface waxes, and other protective compounds and structures that make living plants an inhospitable environment for all but the most adapted microbes (Bigeard 2014). “Plant pathogens” is a term that refers to the subset of microbes that have evolved to produce enzymes, toxins and/or physical structures that enable them to overcome these constitutive defenses, penetrate host plant tissues and derive nutrients from them. 2 The induced defenses that plants use against broad categories of microbes, serve as a second tier of defense against the specialized, plant-adapted pathogens that overcome plants’ constitutive defenses. Pattern recognition receptors (PRRs), proteins embedded in plant cell membranes, can detect compounds which are commonly associated with broad categories of potential pathogens (Boutrot and Zipfel 2017). These pathogen- or microbe-associated molecular patterns (PAMPs or MAMPs) include signature compounds of fungi and bacteria, such as chitin and flagellin. Compounds released by damage to the plant cell wall (damage associated molecular patterns, or DAMPS) can also trigger PRRs. When microbes can get past a plant’s constitutive defenses and establish themselves on, or damage the plant surface, they trigger a defensive response, generally known as PAMP-triggered immunity (PTI) or MAMP-triggered immunity. In PTI, microbes or damage trigger the PRRs, which in turn initiate chemical signaling cascades between the plant cell plasma membrane and the nucleus (Bigeard 2014). This results in the mobilization of various energy- intensive but highly effective defense responses on the part of the plant, including the production or increased production of anti-microbial compounds, and secretion of enzymes involved in reinforcement and repair of the plant cell wall (Bent 2018, Boutrot and Zipfel 2017). Plant pathogens have also evolved ways to overcome PTI. They secrete proteins and other compounds, collectively known as effectors, which can disrupt the PTI response. Effectors that pathogens use to disrupt PTI also vary widely in architecture 3 and function, and tend to be host-specific, allowing a given effector to overcome the defenses of just one or several closely-related plant host species (Kim et al, 2018, Jones and Dangl 2016, Lo Presti 2015). Through the process of attack and counter-attack, plants have evolved a third tier of defense known as effector triggered immunity (ETI). In ETI, plant immune receptors (often called resistance or R proteins) can recognize the presence or virulence activity of certain specific effectors and subsequently trigger even more robust defensive responses in plants (Jones and Dangl 2016). Unlike in PTI, plant immune receptors that detect pathogen effectors are generally intracellular and belong to a nucleotide- binding and leucine-rich-repeat (NB-LRR) superfamily highly conserved across all plant species. During ETI, an R protein gets activated upon recognition of a specific pathogen effector, which leads to elevated biosynthesis of salicylic acid (SA) and production of anti-microbial compounds. ETI is often, although not always, associated with the hypersensitive response (HR). HR is a form of programmed cell death at the site of infection, which can stop some pathogens, especially biotrophs such as powdery mildew, by depriving them of the resources they need to survive (Bent 2018, Jones and Dangl 2006). In addition to the tight regulation of PTI and ETI by various molecular components downstream of the respective immune receptors, plant defenses are also modulated by phytohormones, such as SA, jasmonic acid (JA) and ethylene (ET). These hormones are also involved in systemic acquired resistance (SAR) and induced systemic 4 resistance (ISR). SAR and ISR are broad spectrum forms of resistance induced by local infection by microbes, including avirulent fungal endophytes and beneficial root-colonizing bacteria. Kuhn et al reported that basal defenses against fungal pathogens, at least in Arabidopsis, rely primarily on PRR signaling to block penetration, but that phytohormone-mediated signaling comes into play once the pathogen has penetrated the plant surface (Kuhn et al 2016). Every pathogen-host relationship (pathosystem) is unique and involves multiple, sometimes hundreds, of molecular interactions at the cellular level, governed by genes accumulated over millions of years of co-evolution (Kim et al 2018). The co- evolution can be compared to an arms race that endows both sides with arsenals of molecular weapons and defenses, developed in response to each other over time, through a series of incremental genetic changes. This gradual change allows successful pathogens and hosts to remain in ecological equilibrium with each other. Our knowledge of the nature of these molecular interactions remains incomplete, however, even for highly-conserved interactions common to all plants. A better understanding of these molecular interactions will enable us to understand the disequilibrium that leads to specific diseases and develop strategies to control them. The demise of American chestnut as a case study in host – pathogen disequilibrium Background: The near complete eradication of the once-prevalent American chestnut from the eastern forests of North America, which began in the early 1900s, is a perfect case 5 study in host-pathogen disequilibrium. The causal agent of chestnut blight, the ascomycete fungus Cp, originated in Asia (Rigling and Prospero 2018, Gruenwald 2012), and researchers have described its accidental introduction to North America as a classic example of disasters that can arise when pathogens are introduced to new environments or host species (Anagnostakis 1987). The arsenal of offensive genes Cp had developed through eons of co-evolution with Asian chestnut species allowed it to decimate the nearly defenseless American chestnut population, that had never been exposed to it, within a few decades of introduction. Cp has been endemic in the American chestnut’s entire natural range since the mid-1900s (Rigling and Prospero 2018, Steiner 2017). Asian Castanea species, including Japanese chestnut (C. crenata), Chinese chestnut (C. mollissima), and two other Chinese species, C. henryi, and C. seguinii, show variable but generally high levels of resistance to Cp (Steiner 2017, Zhang 1998). Cp is a minor, superficial disease in Asian chestnut forests, and only causes significant damage when trees are stressed by anthropogenic or environmental factors (Zhang 1998). These Asian chestnut species can also survive in good health in North America due to their blight resistance, but have not replaced American chestnut in the wild during the century since blight arrived, perhaps because they are not ideally adapted to American forests. Asian chestnut species are generally smaller than, and lack the timber form of American chestnut (Schlarbaum et al 1992). American and European chestnut species, including C. dentata, Alleghany chinkapin (C. pumilla), Ozark chinkapin (C. ozarkensis) and European chestnut (C. sativa), none of which 6 co-evolved with Cp, show variable but low levels of resistance to the fungus (Prospero and Rigling 2017). The molecular basis for Cp virulence and host resistance is poorly understood. Despite over 100 years of research, the genes that are responsible for resistance in Asian chestnut species have not yet been identified (Steiner 2017). Efforts under way to restore American chestnut to North American forests take various approaches that do not require specific knowledge of naturally-occurring, resistance genes. These include: • selective intercrossing between the most resistant surviving American chestnuts (Griffin 2006), • the use of the chestnut hypovirus (CHV), a mycovirus, as a biocontrol against Cp (Milgroom and Cortesi 2004), • backcross breeding of Chinese or Japanese chestnut to the American chestnut background (Steiner 2017, Hebard 2005 and 2014), • and the use of biotechnology to introduce novel forms of resistance into the American chestnut background (Newhouse et al 2014). Except for the biotechnology approach, in which an oxalate oxidase gene from wheat was inserted in the chestnut genome (Newhouse et al 2014), these approaches rely on the introgression of naturally occurring resistance genes from Asian populations into a blight-susceptible American population. Though the loci, sequences, and functions of these naturally occurring resistance genes remain elusive (Steiner 2017), 7 researchers have uncovered some clues about the mechanisms behind the resistance of Asian chestnut species. Studies of the inheritance pattern (Hebard 2005 and 2014) of blight resistance in Chinese-American hybrid trees, for example, support the hypothesis that Chinese chestnut’s resistance to Cp involves multiple genes which each contribute partially to resistance. Several researchers have also isolated compounds or chemical fractions present in Chinese chestnut bark that inhibit fungal growth (Gao and Shain 1995, McCarrol and Thor 1979, Samman et al 1979), suggesting a difference in pre-formed or constitutive defenses between Chinese and American chestnut. Studies on the histology of Chestnut blight infections in susceptible and resistant trees, however, also point to induced responses that differ in amplitude between resistant and susceptible trees. Chinese chestnut and resistant hybrids show an ability to contain the spread of Cp infection through the rapid lignification of wound periderm around the infection site, while more susceptible trees are slower to lignify, allowing the mycelial fan of the fungus to grow through the defensive wound periderm (Hebard et al 1994). Studies of gene expression in Cp-infected and non-infected American and Chinese chestnut tissues (Barakat et al 2009 and 2012) are consistent with the findings of Hebard et al (1994). Upon infection with Cp, Chinese chestnut shows high levels of expression of defense-related genes, followed later by a much smaller spike in metabolic genes associated with tissue repair. In American chestnut, on the contrary, the amplitude of increased expression of defense-related genes is relatively low, but 8 the subsequent spike in metabolic/repair-associated genes is much higher, likely because the damage done by the pathogen is greater. This suggests that even though Chinese and American chestnut can detect fungal infection, but that the PTI response of Chinese chestnut is faster or more effective, and/or that ETI plays a stronger role in the defenses of Asian chestnut species than it does in C. dentata. Cp can grow as a saprophyte on numerous woody plant species, and it can be a significant pathogen on certain oak species which, like chestnut, are members of the Fagaceae family (Roane et al 1986). However, the fungus is only severely pathogenic on chestnut species (the Castanea genus). This suggests that the highly conserved, broad-spectrum, constitutive and PTI defenses found in non-host plants are sufficient to protect them from Cp, and that the pathogen has adapted to produce effectors that can disable or bypass the particular forms of these defenses that are specific to the Castanea genus and its closest relatives. The fact that Castanea species that co-evolved with Cp are resistant, but isolated populations such as C. dentata are not, also suggests that ETI developed in Asian chestnut species under the selective pressure of Cp. Most of our current knowledge about the molecular interactions between Cp and chestnut comes from investigation of the naturally-occurring mycovirus CHV, which has effectively prevented the eradication of the C. sativa in Europe (Rigling and Prospero 2018). CHV infects Cp and can significantly reduce its virulence. The capsidless virus is widespread in Europe and Asia, and is transmitted horizontally by 9 anastomosis between mycelia, and vertically through asexual spores (but not sexual spores). Research into viral hypovirulence has revealed several Cp genes suppressed by the virus which are essential for virulence, some of which appear to have regulatory functions and others of which (summarized in table 1 below) appear to encode proteins directly involved in pathogenesis. These genes are involved in the biosynthesis of secreted enzymes or phytotoxins, and many are co-regulated through Cp’s G-protein signaling pathway (Dawe et al 2004). Table 1. Cp genes identified through research into CHV as being involved in fungal pathogenicity or virulence in chestnut. Gene Putative function Reference CHB1 Cellobiohydrolase involved in Wang and Nuss 1995 breakdown of cellulose CRP Cryparin, an abundant Cp Zhang et al 1994 hydrophobin associated with fruiting body eruption KEX2 Protease necessary for virulence but Jacob-Wilk et al 2009 not vegetative growth LAC3 Extracellular laccase (phenoloxidase) Chung et al 2008 necessary for virulence Kim et al 1995 PRB1 Subtilisin-like protease involved in Shi 2014 both vegetative growth and virulence OAH Oxaloacetylhydrolase enzyme Havir and Anagnostakis 1985 necessary for the production of the phytotoxin oxalic acid While virulence factors revealed by CHV could serve as targets for engineered forms of resistance to Cp, they do not appear to be effector proteins, and are not necessarily responsible for the Cp’s unique pathogenicity on Chestnut. They appear to support fungal processes which are necessary, but not sufficient by themselves, for Cp to colonize and severely parasitize live chestnut tissue. For example, the Cp OAH gene 10 is essential for production of the secreted phytotoxin oxalic acid, and the suppression of this gene renders Cp avirulent against chestnut (Havir and Anagnostakis 1985). There are, however, numerous fungi that produce oxalic acid (Dutton 1996), among which only Cp is a significant chestnut pathogen. Most likely, Cp’s ability to invade chestnut tissue and kill cells with oxalic acid is only possible because the fungus also produces as-yet-unidentified effectors that overcome chestnut’s constitutive and PTI defenses. Improved sequencing technology, and the recent completion of an American chestnut genome (Schmutz et al, not yet published) will help researchers identify and characterize some of the natural resistance genes in highly-resistant Chinese- American hybrid chestnuts (Steiner 2017, Westbrook 2018). This information will greatly improve the efficiency of programs to breed for resistance. However, even if such programs are successful in capturing and fixing major Chinese chestnut resistance genes in a mostly American hybrid population, there is reason to believe that this may not bring about equilibrium between host and pathogen. Zhang et al (1998), in their analysis of the dynamics between Cp and the blight-resistant C. molissima in Chinese chestnut forests, describe the relationship as a “hybrid system” in which the disease is kept under control by both host resistance and CHV infection of the fungus. Though CHV has been released, and also found to occur naturally in North America, it has not become a widespread or durable biocontrol as it has naturally in Europe or Asia. The reasons for this remain poorly understood (Milgroom and Cortesi 2004). It is therefore likely that novel forms of resistance 11 will be a necessary to supplement the introgression of naturally occurring Asian resistance genes in the overall effort to restore the American chestnut. Rationale and significance of this research The growing importance of molecular genetics for crop protection and environmental integrity The population of the world is projected to reach 9.7 billion in 2050 (United Nations 2017). This will put great pressure on a global agricultural system that has limited new land to put under cultivation, and which already relies heavily on non-renewable fossil fuels and diminishing fresh water sources to meet current demand. To meet the challenge of increasing agricultural yields, an obvious priority is to reduce the amount of food lost to plant disease. Currently, an estimated 30% of crops planted are lost to pre- or post-harvest disease, with plant-pathogenic fungi accounting for most of this loss (Bent 2018). Climate change and human behavior are exacerbating this problem by bringing crops and wild plants into contact with pathogens which have not previously threatened them (Cline 2007). Despite advances in biotechnology and genetic engineering, most improvements in resistance to pathogens of economically important plants still rely on traditional breeding to capture naturally occurring genetic resistance (Chrispeels 2018). There is no guarantee that this approach will remain sufficient to stay ahead of accelerating 12 changes in plant-pathogen dynamics, a problem which has implications for food security and for environmental integrity. New developments in biotechnology, however, offer the possibility of developing novel forms of disease resistance for both crops and endangered wild plants, and to protect and improve agricultural productivity and ecosystem integrity. Realizing the potential of biotechnology to combat fungal disease threats will require a better understanding of the molecular interactions between pathogenic fungi and their plant hosts. The Arabidopsis-powdery mildew model pathosystem Many important discoveries about plant defenses against fungal and other microbial pathogens have been achieved through studies of model species, such as Arabidopsis, the first plant to have its genome fully sequenced. In addition to the availability of a reference genome, Arabidopsis also offers the advantages of a short life cycle (as little as 6 weeks from seed to seed), being easy to self- or cross-pollinate, and amenability to genetic modification. Arabidopsis is also susceptible to certain species of powdery mildew (PM) fungi, and there are mutant Arabidopsis lines that show either resistance to PM species that are well-adapted to Arabidopsis, or susceptibility to PM species that are poorly adapted to Arabidopsis wild-type plants. Forward genetic studies of Arabidopsis have thus enabled researchers to identify Arabidopsis genes associated with resistance or susceptibility to PM. This has revealed important details about the molecular interactions that either allow or prevent PM infection in Arabidopsis, and homologues of such host components are found in other plant species. 13 Two important examples of discoveries that have emerged from examination of the PM-Arabidopsis pathosystem are RPW8-mediated broad-spectrum resistance, and loss-of-MLO (mlo)-mediated complete resistance against PM fungi. RPW8.2, a member of a small family of broad-spectrum resistance genes identified and characterized by Xiao et al (1997 and 2001), encodes a resistance protein that is targeted to, and functions at the extrahaustorial membrane (EHM), a host-derived membrane that surrounds the PM feeding structure known as the haustorium. The EHM is the principal interface between host and pathogen, and the primary place at which effector proteins enter host cell and nutrients are taken up by the fungus. RPW8-mediated resistance is SA-dependent, involves the accumulation of hydrogen peroxide, and can lead to HR response in the infected cell (Xiao et al 2001). mlo- mediated broad-spectrum and durable resistance to PM fungi in barley has been employed in agriculture for close to a century (Buschges et al 1997, Piffanelli et al 2004). The Arabidopsis MLO2 gene, along with its close homologs MLO6 and MLO12 plays a similar role in PM penetration of host cells (Consonni et al 2006). Mutant plants with non-functional MLO2 or multiple non-functional MLO family genes show strong resistance to penetration by PM fungi (Kuhn et al 2016). However, the molecular basis of both RPW8- and mlo-mediated resistance to PM fungi remains to be elucidated (Kuhn et al 2016). One of the main reasons that little is known about RPW8 and mlo-mediated resistance from the pathogen’s perspective is because PM fungi are genetically intractable. To 14 date, no one has perfected any method to make stable, targeted mutations in the genomes of biotrophic fungi such as PM. This means that reverse genetic methods, such as gene deletion and gene over-expression can only be applied to the host in the PM-Arabidopsis pathosystem. Genetic exploration of the pathogen is very limited in the case of Arabidopsis-PM interaction, and the genetic basis of fungal virulence in PM fungi consequently remains poorly understood. The merits of the chestnut-Cp model system Unlike biotrophic PM fungi, Cp as a nectrotrophic pathogen is genetically tractable (Nuss 2011). Reverse genetic studies with Cp have allowed researchers to conduct knockout studies for candidate genes and identify essential factors in fungal virulence in chestnut (table 1). There are several advantages of using Cp as a model species for studying fungal virulence or pathogenicity. These include: (1) Cp can be easily cultured in vitro and stored at -20°C for many years; (2) the virulence of Cp fungal strains in chestnut can be quickly assessed through various types of controlled inoculations; (3) vegetative Cp spores and mycelia are haploid, resulting in phenotypic expression of relevant mutations; (4) Cp quickly produces uni-nucleate, haploid asexual spores in culture, making it possible to isolate monokaryon knockout strain mycelia through a variety of screening methods; and finally (5) the availability of DK80, a mutant strain of Cp, which is highly virulent but has an impaired non- homologous end joining capability, makes genetic transformation of the fungus by homologous recombination highly efficient (Nuss 2011). In addition, a well- annotated reference genome is available for the standard virulent Cp strain EP-155 15 (U.S. Department of Energy, Joint Genome Institute, https://genome.jgi.doe.gov/Crypa2). By using this ascomycete fungus as a surrogate for related biotrophic fungi, such as PM species, we can overcome some of the limitations on performing reverse genetic studies with biotrophs, at least for the rapid screening of candidate genes with homologues present in both species. Defining pathogenicity broadly for target gene selection This study seeks to identify genes that are highly conserved in ascomycete plant pathogenic fungi and play a role in fungal pathogenesis in plant hosts. Our focus goes beyond fungal gene products that act on the plant host, such as digestive enzymes, toxins and effector proteins, and covers a broader concept of pathogenicity that includes internally-acting fungal genes that enable fungi to survive in the well- defended, hostile and often nutrient-poor environment of a live host. We also emphasize candidate genes that are highly conserved across species, and we do so for two reasons: (1) that we will be able to use reverse genetic methods (particularly gene knockout) to explore the function of genes that exist in a genetically tractable fungus (Cp) to identify important virulence factors for further study in intractable biotrophic fungus (Gc); and (2) that it may lead to the development of novel forms of resistance that may be applicable to multiple plant-pathogenic fungi. More immediately, the information we obtain from the Cp – chestnut pathosystem through genetic modification of the Cp fungus may be useful for us to functionally characterize candidate homologous PM genes through host-induced gene silencing (HIGS). 16 HIGS is a relatively recently developed method for engineering plants with resistance to specific pathogens. HIGS relies on RNA interference (RNAi), a process that occurs in eukaryotic cells (Weiberg et al 2014.) In RNAi, a host cell produces enzymes known as dicer-like proteins (DLPs), which cleave double stranded RNA (dsRNA) molecules into 21-22 base pair short fragments, known as small interfering RNAs (siRNAs). These siRNA fragments are then incorporated into RNA-induced silencing complexes (RISCs), which are also produced by the host cell. The siRNA fragments in the RISCs serve as template to bind complementary messenger RNA (mRNA) molecules and guide the RISCs to cleave the target mRNA and /or disrupt its translation, thus silencing expression of the gene from which the mRNA was transcribed (Weiberg et al 2014). Though RNAi was originally understood as a defense against dsRNA viruses and/or a process for regulating host gene expression, it is now known that there is also two- way trafficking of dsRNA between pathogens and plants (Weiberg et al 2014, Baulcombe 2015, Han and Luan 2015). There are fungal pathogens that export dsRNA to plants where they silence host genes, and plants that export dsRNA to silence pathogen genes (Cai et al 2018). HIGS technology involves the artificial insertion of genes into a host that encode dsRNA matching target genes from a pathogen’s genome. These dsRNAs can be trafficked into the pathogen where they help silence the target gene. Though the silencing is not always complete, HIGS can at least down-regulate target genes. It can be used experimentally, as a substitute for gene knockout, or as a novel form of defense in genetically modified plants. HIGS- based defenses have been successfully demonstrated in several plant species, 17 including in papaya against ring spot virus (Gonsalves 1998), in barley against Fusarium head blight fungus (Fusarium graminearum) (Nowara et al 2010, Koch et al 2013), and in banana against Panama disease (Fusarium oxysporum f.sp. cubensis) (Ghag et al 2014). Research objectives This research examines the function of six highly conserved genes found in plant- pathogenic fungi, focusing on the roles they play, with a view to discovering new targets for fungal gene disruption that could be employed in novel forms of defense. The six genes in this study have homologues in Cp and six PM species, including the Arabidopsis pathogen Gc, the genomes of which have been sequenced and annotated (Wu et al 2018). We screened homologous genes found in Cp and Gc, emphasizing candidates whose PM homologues are upregulated in the haustorium and likely to be involved in pathogenicity. (Note: gene selection methodology and results are discussed in greater detail in chapter 2). We attempted to delete these genes, by homologous gene replacement in Cp, and observe the effect on the fungus in vitro and in planta. After studying the selected genes in the Cp-chestnut pathosystem, we intend to silence their homologues in Arabidopsis by means of HIGS, and compare the effect on fungal virulence to that observed in Cp. This research is exploratory in nature and is intended to generate additional testable hypotheses about specific genes or types of genes that may serve as targets for developing novel forms of defense against fungal pathogens in genetically modified plants. 18 The research presented in this thesis has 4 main objectives: 1. To test the concept of using a genetically tractable necrotrophic pathogen as a surrogate for reverse genetic studies of homologous genes in a genetically intractable, biotrophic pathogen, 2. To functionally characterize previously unstudied fungal genes for their potential roles in pathogenesis, 3. To improve our understanding of molecular interactions between hosts and pathogens in the Arabidopsis-PM pathosystem, and 4. To identify Cp genes as targets for HIGS-mediated resistance to Cp in American chestnut 19 Chapter 2: Genetic Study of six Highly-Conserved Genes in Cp Introduction The Chestnut-Cp interaction and Arabidopsis-Gc interaction, each have distinct features. While Cp is amenable to genetic manipulation (Nuss 2011), genetic modification of chestnut trees is difficult and time consuming (Newhouse et al 2014). Conversely, while Arabidopsis is genetically amenable, Gc is not. In this work, we sought to use both pathosystems in a complementary. To bypass technical barriers that prevent reverse genetic research on biotrophic fungi such as Gc, we targeted homologues of Gc genes found in the genetically transformable necrotrophic plant pathogen Cp. Any of these homologous genes whose deletion in Cp results in reduced virulence in its host (chestnut) would not only be promising targets for novel forms of resistance to chestnut blight, but also candidates as possible pathogenicity genes in Gc. While targeted mutagenesis/gene knockout is not possible in Gc, HIGS could be used to suppress gene expression. The first step was to identify promising target genes. Materials and methods Identification of target genes Our criteria for selecting homologous gene pairs across the two ascomycete fungal species (Cp and Gc) included high-level protein sequence homology (E<10-6), the presence of a predicted N-terminal signal peptide, the absence of predicted transmembrane domains, and increased expression of the Gc homologues in haustoria 20 (Wu et al 2018). We rejected genes whose homologues had been shown to be essential for survival (i.e. lethal when deleted) in the model ascomycete yeast Saccharomyces cerevisiae (Sc). These criteria were designed to capture a broad range of secreted proteins involved in colonization of, and adaptation to, the host, rather than just those that resemble proteins involved in known pathogenesis pathways. We did not include in our selection criteria information about putative gene function, or other characteristics generally considered typical of effector proteins or pathogenesis-related genes, such as short protein sequence, cysteine richness or lack of homologues outside of Gc and Cp (Kim et al 2016). Target gene selection began with browsing the genome of Gc strain UCSC1 (Genbank accession number MCR00000000.1), using SignalP3.0 (www.cbs.dtu.dk/services/SignalP-3.0/) to predict potential N-terminal secretion signal peptides and TMHMM 2.0 (www.cbs.dtu.dk/services/TMHMM) to predict transmembrane domains in the mature peptides. We then did BlastP searches of the resulting list of Gc genes against the genome of Cp strain EP155 (U.S. Department of Energy Joint Genome Institute, genome.jgi.doe.gov/Crypa2). The Cp EP155 strain is the parent strain of the Cp DK80 mutant strain used for transformation in this project. Simultaneously, we also did a BlastP search for homologues of the same genes in the genome of Sc strain S288C (GenBank accession number PRJNA128, ncbi.nlm.nih.gov/bioproject/128), to screen out genes found to be essential in this extensively studied genome. 21 Using the criteria above, we identified six potential pathogenicity genes with homologues in Cp and Gc. The six genes were designated as Target Gene one through six (hereafter TG1-TG6), in priority order based on the degree to which homologues have been reported in plant-pathogenic fungi, and the levels of amino acid sequence identity between the Cp and Gc homologues. None of these six genes appear to have been previously studied or characterized in either species. However, BlastP searches of the predicted proteins in Cp produced numerous homologues for each, and several of the genes have been characterized or studied in Sc or other species of fungi. Creation of gene disruption constructs We attempted to develop knockout strains of Cp for each target gene using homologous gene replacement techniques described by Churchill et al (1990). The Cp strain DK80, obtained from Dr. Dongxiu Zhang at the U.S. Department of Agriculture’s Agricultural Research Service in Beltsville, MD, is a mutant strain derived from the standard virulent research isolate of Cp, EP-155, with a gene essential for non-homologous end joining deleted (Lan et al 2008). DK80 can be transformed by homologous recombination with up to 85 percent efficiency (Nuss 2011) by incubating DK80 spheroplasts in an osmotic solution at room temperature with chimeric fragments of DNA whose flanking sequences match targeted portions of the DK80 genome (Churchill et al 1990). Chimeric DNA fragments, to be used as gene disruption/knockout constructs, were designed using Benchling molecular cloning web tools (Benchling.com). One type of 22 chimeric fragment was developed for each target gene. Each included a common marker cassette from the pKAES173 plasmid that contains a hygromycin resistance enzyme (hygromycin phosphotransferase – hph) controlled by a constitutive fungal promoter from Aspergillus nidulans. We obtained the pKAES173 plasmid from the Nuss lab at the University of Maryland’s Institute for Bioscience and Biotechnology Research (IBBR). Flanking sequences matching those of the relevant target gene, were fused to the 5’ and 3’ ends of the marker cassette using overlapping PCR. Flanking sequences were cut from DK80 genomic DNA using outer forward (FSFPs) and reverse (FSRPs) primers matching 200-600 bp of the 5’ and 3’ flanking sequences of the target genes and overlapping primers that included one part matching the other flanking sequences fused to 20-22 bp segments complementing the relevant end of the marker cassette. These PCRs resulted in the amplification of target gene flanking sequences with 20-22 bp overhangs matching the marker cassette. The marker cassette itself was amplified from the pKAES173 plasmid by PCR with simple forward and reverse primers (Marker cassette forward and reverse primers – MCFP and MSRP). All primers were designed with the web-based molecular biology platform Benchling (Benchling.com) and manufactured by Eurofins. All PCR reactions were performed with Takara Extaq high fidelity polymerase and Extaq 10x PCR buffer. Primer sequences and melting temperatures are provided in table 2. 23 Table 2. Primers used in the amplification of chimeric gene disruption constructs for TG1-TG6, and their melting points. FSFP and FSRP refer to flanking sequence forward and reverse primers. MCFP and MCRP: marker cassette forward and reverse primers. OFP and ORP: overlapping forward and reverse primers. The common sequences between the OFPs/ORPs and the marker cassette¸ and melting points, are indicated in bold. Primer 5’ to 3’ sequence Tm MCFP TGCAGCCCGGGGGATCCATAA 62°C MCRP CGACGTTGTAAAACGACGGCCA 60°C TG1-FSFP AGACTTGCCATTTCTTCTCTCCT 61°C TG1-FSRP GCATATGAGTCTTTGAGCAAAACGA 61°C TG1-OFP TGGCCGTCGTTTTACAACGTCGGGGTACGGCATCAGCGA 60°C /54°C TG1-ORP TTATGGATCCCCCGGGCTGCAGAGTGATGTCGACGTGAAAAGA 62°C /55°C TG2-FSFP AAGTCAGAGAAGGGGAAAGTGA 64°C TG2-FSRP CGGTGGACTTCGGTTGACT 64°C TG2-OFP TGGCCGTCGTTTTACAACGTCGCTCAACTTCGTCCCTCCGT 60°C/55°C TG2-ORP TTATGGATCCCCCGGGCTGCAGAGTCGAACGGTGTGTCGT 62°C/57°C TG3-FSFP CCCTCGGTTGCTCAGTATATCA 65°C TG3-FSRP GGAAGAAGAGGTGGCGGTA 64°C TG3-OFP TGGCCGTCGTTTTACAACGTCGTCGCAGATCTCTGATGGTAAGTT 60°C /55°C TG3-ORP TTATGGATCCCCCGGGCTGCACGTTGTTGTTTTGTGGCGTTTA 62°C /55°C TG4-FSFP ACAAGATGTCGTGGTATTACTAGGA 62°C TG4-FSRP GCTTGGAATTTGGTGGTGGA 65°C TG4-OFP TGGCCGTCGTTTTACAACGTCGTGAAAGGAAAAAGCCAGGTTGA 60°C /55°C TG4-ORP TTATGGATCCCCCGGGCTGCAGCTCCTCCCAGATTGCAGAT 62°C /56°C TG5-FSFP GCAGGGTGAACCTGATTTCTCTACCACATCAAAAT 63°C TG5-FSRP CAGGCAAACAATGCCTGCCAGCTTAT 62°C TG5-OFP TGGCCGTCGTTTTACAACGTCGTTGGGGGAGGTGGGATCTCAAGTCA 60°C /63°C TG5-ORP TTATGGATCCCCCGGGCTGCACACCTGATGTTATACAAGACCAAGTGGTTGTCAA 62°C /62°C TG6-FSFP TAATGTGAGCAGGAGCATCTTGACGAAGTGTTT 63°C TG6-FSRP GTGGTTTTAACACTTTACTAGAGGCGCATATTTACCATCATATATTA 62°C TG6-OFP TGGCCGTCGTTTTACAACGTCGCTATGACTGACAAGTGGACGCCGCT 60°C /63°C TG6-ORP TTATGGATCCCCCGGGCTGCATTTCTATTGACTTTGAGCAAGTACTCGTGCA 62°C /60°C For each target gene, full chimeric fragments were generated by overlapping PCR using the forward primer of the 5’ flanking sequence as a forward primer and the reverse primer of the 3’ flanking sequence as a reverse primer. During the PCR, denatured marker cassette segments annealed to the overlapping tails of the flanking sequences resulting in fusion and amplification of all three fragments. A diagram of the chimeric fragment assembly is provided in figure 1. The resulting PCR products were purified through gel electrophoresis in 1% agarose gel and measured with a standard 1 kb DNA ladder (Fermentas Generuler). The bands corresponding to the 24 full length of the expected chimeric fragment were cut from the gel and purified using a ThermoScientific GeneJET Gel Extraction Kit. Figure 1. Schematic of the overlapping PCR process to create chimeric gene disruption constructs from separate fragments of DNA. Concentrations of chimeric gene disruption constructs amplified by overlapping PCR for TG1-TG4 and TG6 ranged from 14.6 to 35.2 ng/μl, too low to efficiently produce the 5-10 mg required for fungal spheroplast transformation (Churchill et al 1990). To increase concentrations, chimeric fragments for TG1, TG2, TG3, TG4 and TG6 were inserted into and amplified in E. coli bacteria, and plasmid DNA was extracted (ThermoScientific GeneJet miniprep kit). Multiple attempts produced successful results for different target genes using different vectors, bacteria and extraction methods (table 3). Because the DNA concentration of TG5 overlapping PCR product was relatively high, bacterial amplification was not necessary. TG5 fragments were produced with overlapping PCR, using 5 cycles without primers to encourage the flanking sequences to anneal to the marker cassette fragments, followed by 30 cycles with the 5’ forward and 3’ reverse primers added. 25 Table 3. Methods used for amplification of the TG1-TG6 gene disruption constructs. The pGEM-T EasyVector and T-4 ligase were purchased from Promega. The pGXT plasmid was obtained from Dr. Guoliang Wang, Ohio State University. Target Initial Plasmid used for Bacteria type Amplification/Extraction Final DNA Gene DNA transformation method conc. conc. (ng/μl) TG1 18 ng/μl pGEM-T EasyVector/T4 Invitrogen OneShot Miniprep 441 ng/μl ligase Top10 chemically competent cells TG2 34 ng/μl pGEM-T EasyVector/T4 Invitrogen OneShot Miniprep 287 ng/μl ligase Top10 chemically competent cells TG3 15 ng/μl pGEM-T EasyVector/T4 Invitrogen OneShot Miniprep 278 ng/μl ligase Top10 chemically competent cells TG4 18 ng/μl pGXT plasmid/T4 ligase Invitrogen OneShot Miniprep, followed by 35 39 ng/μl Top10 chemically cycles of PCR amplification competent cells using ExTaq polymerase TG5 107 ng/μl n/a n/a Overlapping PCR 107 ng/μl TG6 64 ng/μl pGXT plasmid/T4 ligase Invitrogen OnShot Miniprep 98 ng/μl Top10 electro- competent cells Spheroplast preparation Spheroplasts of Cp strain DK80 were prepared following the protocol described in Appendix II. DK80 was grown on PDA medium, and aerial hyphae were harvested by swirling through the aerial layer with a pipette tip, accumulating a small sphere (approximately 5 mm in diameter) of hyphal tissue. The sphere was transferred to a 1.5 mLcentrifuge tube, washed with distilled water, and ground with 50 twists of a small plastic micro-pestle. The contents were transferred to 100 mL of potato dextrose broth in a sterile 250 mL Erlenmeyer flask and incubated on a bench top over three days until about half the volume of the medium was occupied by white, cloudlike mycelial masses. 26 The mycelium was strained and transferred to sterile 50 mL falcon tube, and spun down at 3700 rcf for 5 minutes at room temperature. The mycelial pellet was washed with distilled water, and spun down again at 3700 rcf for 3 minutes, after which the washing step was repeated one more time. The mycelium was removed from the tube and gently blotted on sterile filter paper to remove excess water and then placed and re-suspended in a new sterile 50 mL falcon tube containing 25 mL of digestion buffer, freshly prepared in the same tube. (The protocol for the preparation of digestion buffer, and other buffers and media used in the spheroplast preparation, transformation and regeneration steps is provided in Appendix II). The mycelium and digestion buffer was incubated horizontally overnight (approximately 16 hours) in a 30°C shaker at 50 rpm, resulting in a homogenous cloudy suspension. After removal from the incubator, cold, sterile trapping buffer was added to overlay the spheroplast suspension until the tube was filled to 50 mL, with care taken not to disrupt the spheroplast layer. The suspension and trapping buffer were spun down at 3,700 x g for 5 minutes at 4°C. Spheroplasts were collected at the interface with a 1,000 μl pipette and transferred to a sterile 50 mL falcon tube, and diluted with two volumes of 1M sorbitol solution, and mixed gently but thoroughly. The suspension was then evenly distributed to 1.5 mL centrifuge tubes, and spun down again at 6,000 rcf at 4°C for 5 minutes. Supernatant was removed by pipette, and the spheroplasts were suspended in a single 200 μl volume of STC, which was transferred from tube to tube to resuspend all of the pellets in all of the tubes in a single volume. The final volume of spheroplasts suspended in STC was 27 spun down again at 6,000 rcf for 5 minutes. The supernatant was removed and replaced with 1 mL of STC, which formed a slightly cloudy homogeneous solution. Approximately 10 μl of spheroplast suspension was observed on a hemocytometer under a dissecting microscope at 250x magnification. Spheroplasts were counted and concentration was estimated at 2 x 106 spheroplasts per mL (figure 2). The solution was then diluted with one part PTC and 0.05 parts DMSO per 4 parts STC, then aliquoted into 1.5 mL tubes and placed in a -80°C freezer for storage in 50 μl units. (Appendix II provides formulations for all buffers related to spheroplast preparation and transformation.) Figure 2. Cp spheroplasts under a hemocytometer. 28 Spheroplast transformation Genetic transformation to create Cp knockout strains was accomplished by incubation of freshly prepared or thawed spheroplasts in a polyethylene glycol (PEG)-based osmotic buffer containing high concentrations (~10 μg of gene disruption constructs per transformation) at room temperature, using methods described by Churchill et al (1990). (The protocol and recipes for all buffers and solutions are provided in Appendix II.) For each transformation, a DNA solution containing the relevant gene disruption construct was placed in 1.5 mL centrifuge tubes with enough distilled water to reach a volume of 10 μl. Fifty µl of spheroplast suspension was added to each tube, gently mixed and allowed to chill on ice for 30 minutes. Then, 500 μl of PTC was added to each tube, mixed gently and incubated at room temperature for 25 minutes. One mL of STC was then added to each tube to stabilize the osmotic pressure, stop the transformation process, and create a ready-to-use suspension of spheroplasts. Regeneration of putative transformants Following the transformation step, the spheroplast suspension was transferred to labelled 85 mm petri dishes in serial dilutions of 2 μl, 20 μl and 200 μl (for first transformation of TG1-TG4) or 5 ul, 20 μl and 80 μl (for second and third transformations of TG1-TG3, TG5-TG6). One control dish was inoculated with the untransformed DK80 spheroplast suspension. Ten mL of regeneration medium (see Appendix I) was added to each dish and mixed with the spheroplast suspension by 29 swirling. Dishes were allowed to solidify under the laminar flow hood and covered. After 16 hours of incubation on the benchtop another layer of regeneration medium with hygromycin was added, allowed to solidify and cool, and then covered and moved to incubate on the benchtop. (The hygromyicin concentration in the upper layer of regeneration medium was 50 mg/mL in the first set of transformations of TG1-TG4, and then attempted at 30, 40 and 50 mg/mL in subsequent attempts with TG1-TG3, TG5 and TG6 . The plates were observed for 3-5 days for the emergence of hyphae on the surface of the regeneration medium. Up to twelve small colonies per knockout strain were named with capital letters and then transferred to individual new petri dishes containing potato dextrose agar (Difco Bacto PD broth plus agar - PDA) medium amended with the selective dose of hygromycin (PDA+hyg) and allowed to grow out. The putative transformant mycelia were labelled by knockout strain followed by the letter identifying the specific colony transferred (e.g. TG1A, TG2C). Plates containing putative transformed mycelia were allowed to grow out on the bench top until the mycelium was large enough to exhibit different morphology the center of the colony than at the growing margin. Two 5 mm punches of each putative transformant and of the untransformed DK80 control were taken from the margin of the mycelium. One of each was transferred to the center of a fresh 60 mm PDA+hyg plate, and another to the center of a 60 mm plate of hygromycin-free PDA. This resulted in two cultures (+hyg and -hyg) of each putative transformant paired with two cultures of the DK80 control (+hyg and -hyg). Each four-plate set was allowed 30 to grow out on the bench top until the most rapidly growing mycelium of the four approached the edge of the petri dish (generally 3 days). At this point the transformants were photographed and phenotypic observations noted. Isolation of monokaryon knockout strain mycelia Tissue from the +hyg culture for each putative transformant and from one DK80 -hyg culture was harvested when aerial hyphae were relatively abundant, and genomic DNA was separately extracted from these samples using a tissue grinder and SDS- based DNA extraction method. Separate PCR analyses were carried out using each sample of genomic DNA, using the forward and reverse primers for the hygromycin resistance marker cassette, and using the flanking sequence primers for each target gene as forward and reverse primers for the whole gene disruption construct. Electrophoresis of the PCR products was carried out in a 1% agarose gel, with a 1 kb ladder (Fermentas GeneRuler) used to gauge DNA fragment size. DK80 samples were used as negative (non-transformed) controls. The presence of a band for PCR reactions using marker cassette primers identified putative transformants. Differences in length between the original target genes and the gene disruption construct that replaced them revealed which samples were transformants, which were wild-type DK80 and which were heterokaryons, containing both the intact and knockout genomes. Because of Cp’s propensity for anastomosis, wild-type spores that germinated in the regeneration medium fused with knockout spores. No mononkaryon mutant colonies were expected or observed at this stage. 31 PDA and PDA+hyg plates containing heterokaryon colonies were incubated on the benchtop for up to 10 days to produce conidia. The heterokaryon colonies generally did not produce fruiting bodies on PDA+hyg medium, so spores in most cases had to be harvested from colonies grown on PDA. Spores were collected by pipetting 10 μl drops of sterile distilled water onto ripe fruiting bodies to allow spores to disperse into the water droplet, then drawing the droplet back into the pipette and transferring it to to a 1.5 μl tubes containing 1 mL of distilled water for dilution (figure 3). Serial dilutions of suspension were transferred to fresh +hyg plates and spread over the surface with sterilized a glass spreader. Any mycelia that grew on these selective medium plates were subsequently transferred to fresh +hyg plates, allowed to grow out until DNA could again be extracted, and analyzed by PCR to identify true monokaryon knockout cultures. Figure 3. Ripe Cp fruiting bodies at 340x magnification. Black arrows indicate where pycnidia exude a water-soluble spore suspension that can be diluted and collected by pipette. 32 Examination of knockout strain phenotype To observe phenotypic changes resulting from the deletion of target genes, nine 5 mm plugs of monokaryon mycelia for each knockout strain were transferred from the +hyg media, where they were grown and maintained, to three types of petri dishes containing different –hyg media: Endothia parastica complete (EPC) medium to observe growth under nutrient-rich conditions, water-agar (WA) medium to observe growth under nutrient-poor conditions, and chestnut bark agar (CBA) medium to simulate the environment present in American chestnut bark. (Formulas for these media are provided in appendix I.) There were three replications for each type of medium. Mycelial diameters were measured daily with a ruler to chart the growth of the fungal strains, and observations made concerning the appearance of pigment and fruiting bodies. Photographs containing one plate each, representing the medium mycelial diameter for each medium type, were taken at regular intervals. Because diameter increases faster for mycelia in nutrient-restrained conditions (Jennings and Lysek, pg. 13), diameter can be a poor indication of biomass accumulation. To measure relative biomass accumulation, nine additional plugs of knockout strain and nine of untransformed DK80 mycelia were made on plates with cellophane layers placed over the three types of medium. The cellophane was pre- cut to have a uniform dry weight prior to inoculation, and was peeled off with the 33 mycelium at seven dpi, blotted with sterile filter paper to remove liquid that occasionally accumulates between the mycelium and the cellophane (this step was only necessary for CBA medium), and weighed in wet conditions (weight when peeled off the medium) and dry conditions (after the mycelium-laden cellophane was allowed to dry at least 24 hours in a 37°C incubator. The average weight of the non- inoculated cellophane controls was subtracted from the average weight for each mycelium type, and the results used to compare the relative accumulation of biomass by the knockout strain versus wild type DK80, using a Student’s t-test (α=0.05). Examination of knockout strain virulence in planta While in vitro observations can reveal some differences between the wild type and knockout strains, there may be other relevant environmental factors in the live bark of chestnut trees that can influence fungal growth and development. In addition, inoculations into live American chestnut tissue are necessary to observe any effect of the gene knockouts on the virulence of Cp. Inoculations into live mature American chestnut trees was not feasible due to the lack of available non-blighted trees, and due to restrictions on the release of genetically modified pathogens into the environment. Therefore two other types of in planta assays were used to observe fungal growth and development in American chestnut tissue: a detached stem assay widely used in Cp research (Elliston 1978, Jacob-wilk et al 2009, Levine-Double personal communication), and a small stem assay developed in cooperation between the author and the science staff of the American Chestnut Foundation (Saielli and Levine 2019). 34 The detached stem assay For the detached stem assay, non-blighted branches 4-8 cm in diameter and 0.5 – 1.0 meters in length were harvested from dormant American chestnut trees in orchards maintained by the Maryland Chapter of the American Chestnut Foundation in January 2019. The ends of the branch segments were sealed with melted food grade wax to preserve moisture and put in a -20 freezer for storage until use. When removed for inoculation, stems were left to thaw to room temperature, bathed in a 10% dilution of household bleach for 15 minutes, then allowed to dry on newspaper. Inoculations sites were chosen at points along the stems that were away from branches or damaged bark, and at least 10 cm apart from each other. Sites were marked with a randomly assigned site number with typing correction fluid. Five mm plugs of inoculum grown on PDA were placed, mycelium side in, into holes drilled through the bark of the stems with an ethanol-sterilized 3/8 inch steel punch in a cordless drill. Inoculum was sealed in place with segments of masking tape. Inoculated stems were placed in sealed, translucent plastic boxes in a greenhouse at 25°C, with approximately 70% humidity and a natural daylight cycle. The length and width of resulting Cp cankers were measured weekly, and estimates of a normalized canker length (NCL) were made. NCL is the square root of the area of an ellipse calculated using the actual length and width of the canker (figure 4). The NCL is a linear measurement that allows us to compensate for differences in the ratio of canker length to canker width that results from variance in diameter between stems. 35 𝜋𝜋𝜋𝜋𝜋𝜋 𝑁𝑁𝑁𝑁𝑁𝑁 = � 4 Figure 4. Formula for normalized canker length, where l = observed canker length and w = observed canker width. Observations were noted concerning the emergence and phenotype of fruiting bodies. Fifteen replicates for each knockout strain and for the inoculations with DK80 were randomly distributed among the available inoculation sites, along with 15 inoculations with sterile agar medium as negative controls, 15 of the virulent Cp strain EP155 as positive controls and 15 of the confirmed weakly virulent Cp strain strain SG2,3. EP155 and SG2,3 cultures were obtained from the American Chestnut Foundation under APHIS license. Failed inoculations, or cankers affected by obvious contamination by naturally occurring fungi were deleted from the data set. The mean NCL for each strain and observations concerning the emergence of fruiting bodies were used as proxy measures for fungal virulence. Differences in NCL between knockout strains and the DK80 parent strain were calculated using Student’s t-test (α=0.05). The small stem assay Detached stem assays have been shown to be good predictors of canker size resulting from inoculation in live trees (Elliston 1978, Jacob-Wilk et al 2009). The cambium layer of stem segments preserved in this manner is made up of live cells with intact constitutive and induced defense capabilities. The assay can be completed within one month and carried out at any time of year using frozen material. However, the detached stem assay may be limited by the inability of live cells to draw on distal 36 resources as they might in a live tree. This is the reason a second type of virulence assay was conducted in the bark of small seedlings in their first year of growth, the small stem assay. For the small stem assay, pure American chestnut seeds were obtained from a single pure American parent tree by the Maryland Chapter of the American Chestnut Foundation. These seeds were stratified at 4°C and planted in a greenhouse in late 2017, and again in late 2018. Inoculations were carried out when enough seedlings had stems 6 mm or greater in diameter at the base. Inoculations involved placement of plugs (approximately 1mm x 1 mmx 5 mm) of knockout strain or DK80 mycelia grown on PDA medium into 1 x 5 mm incisions made through the bark, but not into the heartwood of the stems, with one inoculation per stem. The incision was made with a 2 mm cork borer cut at a 45° angle and sharpened with a scalpel. At least five incisions were inoculated with sterile agar medium as negative controls. All tools were sterilized with ethanol and flared between contact with different fungal strains. After placement of the mycelial plugs into the incisions, each inoculation site was wrapped tightly with a 2 cm wide piece of pre-stretched parafilm and twisted in place. The parafilm was left in place for seven days, and measurements of the length of resulting cankers were taken at 14-day intervals. At each measurement, the length of each canker was recorded, and the trees were also assigned a qualitative score, using the decision tree in figure 5. The qualitative score was designed to characterize stages of canker development, and to be used to compare the rate of canker development between weakly virulent strains that do not result in mortality. 37 Is the stem distal to the inoculation site dead? If yes, Score=5 If no, is there a canker that is sunken in the middle compared to the margins, or which has Cp fruiting bodies? If yes to either, Score=4 If no, does the canker fully encircle the stem, or does it extend more than 5 mm from either end of the inoculation site? If yes to either, Score=3 If no, is there any obvious sign of fungal infection around the inoculation site? If yes, Score=2 If no, Score=1 Figure 5. Decision tree for qualitative rankings of small stem assay cankers. Measurements continued for 14 weeks. At the end of the period, mean canker length and the mean number of weeks until stem death were calculated for each fungal strain. Comparisons of canker length, mean days of stem survival (days until the portion of the stem above the inoculation site dies) and qualitative scores (if necessary) were to be made in Excel using t-tests with unequal variance for pairwise comparisons between any given knockout strain and DK80 and negative controls, or by ANOVA if testing multiple knockout strains in one assay. Results and Discussion Profiles of target genes A comparison between the Cp strain EP-155 and Gc strain UCSC1 reference genomes, using the gene selection strategy described in chapter 2, yielded 33 predicted homologous genes with E-values less than 10-6 . Of these, three had been reported to be essential genes (lethal when impaired) in the model ascomycete yeast 38 Sc. Of the remaining 30, six had previously been shown to be up-regulated at least 2- fold in Gc (UCSC1) haustoria, the feeding structures of the pathogen (Wu et al 2018). None of these six genes (TG1-TG6) have previously been characterized in Cp or Gc, and a literature search revealed no reference to the genes in studies of either species. However, BlastP searches of the predicted proteins in Cp produced over 100 homologues for each, and homologues of each of the genes have been characterized or studied in Sc (table 4). Table 4. Cp Target Genes 1 through 6 and their homologues in Gc and Sc. Target Cp strain Cp Cp amino Gc strain S. cerevisiae Gene EP-155 gene acid UCSC1 homologous protein ID 1 length sequence protein ID protein/gene3 length number2 TG1 96843 2136 564 32023 YCR068W/ATG15 TG2 355196 1003 110 210066 YDR382W/RPP2B TG3 242884 926 286 78010 YJL158C/CIS3 TG4 347494 1086 234 120011 YBR171W/SEC66 TG5 320126 1569 302 132010 YKL120W/OAC1 TG6 334581 709 197 197034 YDL046W/NPC2 1 From U.S. Department of Energy Joint Genome Institute (JGI) genome.jgi.doe.gov/Crypa2, 2 From Genbank accession number MCR00000000.1, 3 from www.yeastgenome.org. Though all six target genes were predicted by analysis with SignalP v. 3.0 to encode proteins with an N-terminal signal peptide, and analysis with TMHMM did not predict transmembrane domains in any of the six, an examination of the most similar proteins discovered by BlastP search suggested that TG1, TG2, TG4 and TG5 encode proteins that are predicted to be targeted to intracellular membrane compartments, and that the TG3 protein may be a components of the cell wall. A BlastP search of 39 TG6 homologues provided no clues regarding the subcellular localization of these proteins, suggesting it may be a secreted effector protein. Barakat et al (2009 and 2012) examined the transcriptome of Cp-infected and non- infected Chinese and American chestnut. We accessed the cDNA reads from these studies (available at https://www.hardwoodgenomics.org) and used Hisat2 to map the reads to the Cp genome. However, only 2,924 reads out of a total of 129,508 mapped to the Cp genome rather than to the chestnut genome. There were five reads corresponding to TG2, and none corresponding to any of the others. We believe the RNA extraction methods used by Barakat et al must have been optimized for plant tissue, which was the subject of their study. We also examined two other studies that looked at the differential expression of Cp genes. Kim et al (2012) did a proteomic analysis of Cp, comparing growth of an uninfected strain and an isogenic hypovirulent (infected with hypovirus CHV1) in PDA medium and PDA amended with tannic acid, which has been shown to induce expression of certain pathogenesis-related genes in Cp. Wang et al (2016) also studied the influence of hypovirus on gene expression in Cp. Neither identified any of our target genes among the genes up- or down-regulated under the conditions of their respective studies. 40 TG1: a putative autophagy-related protein A BlastP search of the predicted protein from the Cp TG1 gene generated 92 non- duplicate hits from the NCBI non-redundant protein sequences database, all corresponding to fungi, and with an average amino acid sequence identity of 76%. Seventy percent of the TG1 hits were in plant-pathogenic fungi, with the closest homologues found in fungi of the Sordariomycete class, which also includes Cp, within the phylum Ascomycotina. TG1 homologues are generally annotated as “predicted lipases, or autophagy lipase protein Atg15.” While we have found no reference in available literature to research on the TG1 gene in Cp or any powdery mildew species, its homologue in Sc has been well-studied. Atg15 is one of several highly-conserved autophagy related protein genes found in eukaryotic cells. Autophagy is a process by which damaged or unnecessary cytoplasmic components and toxic aggregates can be degraded within the vacuole (in fungi and plants) or lysosome (in animals) and recycled (Delorme-Axford et al 2018). Atg proteins mediate a process by which specialized structures capture target substrates and deliver them to the vacuole/lysosome, where they are broken down into raw materials that can be exported back to the cytosol for re-use (Epple et al 2001). Autophagy is essential for cell growth and development and occurs at a low level in all cells. It increases significantly during nutrient starvation, pathogen infection or other stress conditions, and helps maintain homeostasis (Delorme-Axford 2018, Ramya et al 2016). 41 Epple (2001) reports that the Sc ATG15 gene is essential for the breakdown of autophagic bodies in the vacuole. Ramya et al (2016) describe Atg15 as the only lipase among Atg proteins in Sc and report that it preferentially hydrolyses the cellular membrane component phosphatidyl serine. Parzych et al (2018) report that one role of Atg15 in the vacuole is to break down liquid droplets, specialized organelles that can store neutral lipids and sequester toxic compounds such as fatty. They further report that yeast cells lacking the Atg15 protein do not entirely lose the ability to break down lipids through autophagy, but that such cells lose viability in nitrogen starvation conditions within six days, sooner than is the case for wild-type cells (Parzych et al 2018). While the exact roles of TG1 in Gc and Cp are not known, the studies of Sc discussed above suggest that it may help plant-pathogenic fungal cells cope with the nutrient- poor conditions that exist within host tissue by recycling nutrients. Also, in the case of Cp, fatty acid molecules produced by host plants have been shown to inhibit fungal growth (Samann et al 1978), and TG1 may form part of the fungal pathway that sequesters, traffics and breaks down these anti-fungal compounds, among other lipids. TG2: a putative 60s ribosomal subunit P2 acidic protein A BlastP search of the predicted protein from the Cp TG2 gene generated 91 non- duplicate hits from the NCBI database, all corresponding to fungi, and with an average amino acid sequence identity of 76%. The closest homologues of the TG2 protein were reported in other ascomycete fungi of the Pezizomycete subphylum, to 42 which Cp also belongs. Compared to the other target genes in this study, TG2 homologues were less concentrated among plant pathogens, with a greater percentage of hits corresponding to saprophytes or animal/insect pathogens. The high amino acid sequence conservation of the TG2 protein in fungi of various classes and lifestyles suggests that the protein is ancient, and that it likely plays an important housekeeping or regulatory function. BlastP hits were generally annotated as “putative 60s ribosomal subunit proteins,” or “P2 acidic proteins.” A literature search produced no previous research into the TG2 gene in powdery mildew or Cp, but the gene has been the subject of several studies in Sc, in which it is characterized as a P2 acidic protein. In Eukaryotic species, P1 and P2 acidic proteins interact with the P0 protein to form the ribosomal stalk, a structure which is involved in translation elongation (Remacha 1995). Remacha reports that there are genes encoding two forms each of the P1 and P2 acidic proteins in Sc, and an analysis of P1/P2 mutants suggested that their absence affected the rate of cell growth, but not cell viability. The absence of different P1/P2 proteins from the ribosome did not affect expression of different metabolic pathways in the same way, and Remacha (1995) hypothesized that the different acidic proteins play different roles in the translation of different mRNAs. Cardenas (2019), studying P1/P2 mutants in Sc found the absence of certain acidic proteins affects the translation of specific mRNAs, and to leads to certain phenotypic traits, such as cold-sensitivity. Cardenas describes P1 and P2 proteins as part of a stalk assembly mechanism that can produce heterogeneous ribosomal stalks. 43 Cellular expression of various proportions of P1 and P2 proteins appears to function as a regulatory mechanism moderating the relative efficiency of translation of different mRNAs (Cardenas, 2019). Such a regulatory mechanism may play a role in enabling fungi to adapt to changing environmental conditions associated with changes in internal or environmental conditions at different points in their lifecycles. TG3: a putative cell wall mannoprotein The annotation of the TG3 gene changed significantly between the first and second versions of the Cp genome published. A reannotation of the Cp genome was performed at the University of Southern Mississippi in 2017, and included a third annotation for the TG3 gene, but has not yet published (Levine-Dawe personal communication). BlastP searches were carried out on the proteins predicted by all three annotations, and all three produced a highly similar set of hits from the NCBI non-redundant proteins database. After removing duplicate species, we obtained a list of 96 fungal proteins with an average amino acid sequence homology of 52%. Of these, 75 were found in plant, animal or insect pathogens. Hits with the highest similarity scores were all plant pathogens in the fungal order Diaporthales, to which Cp also belongs. Annotations of the putative homologues included “covalently linked cell wall protein,” “cell wall mannoprotein,” “cell wall Cis3 Protein,” “Pir3 protein,” and “Pir5 protein.” All of these refer to a family of glycosylated proteins found in the outer cell walls of certain fungi. (Hsu et al 2015, Klis et al 2006, DeGroot et al 2005). TG3 is the only member of this protein family that has been annotated as such in Cp. 44 A literature search found no previous reference to the TG3 gene, or any genes bearing the same annotations, for Cp or Gc, but there has been extensive research into the localization and possible roles for such glycoproteins in Sc, and some discussion of homologues in filamentous fungi, especially ascomycetes (DeGroot et al 2005). The PIR (proteins with internal repeats) family consists of mannose-containing glycoprotein constituents of the fungal cell wall. DeGroot et al (2005) describe the cell walls of fungi from which such proteins have been isolated as consisting of an internal skeleton of 1-3 beta-glucan chains, surrounded by a denser layer rich in proteins. DeGroot and others speculate that the repeats in PIR proteins allow them to bind to multiple 1-3 beta-glucan molecules and stabilize the otherwise highly flexible and porous cell wall (Hsu et al 2015, Klis et al 2006, DeGroot et al 2005.) PIR proteins appear to have more than a passive reinforcement function, however. The incorporation of PIR proteins and other proteins into the cell wall is tightly regulated, based on location and on what stage of the cell cycle the cell is in. Cell wall protein composition is also influenced by osmotic pressure and physical stress/damage and other environmental conditions, through a variety of signaling pathways (Hsu et al 2015). Possible functions for PIR proteins enumerated by DeGroot et al (2015) include: water retention, maintaining cell wall integrity in response to stress and/or growth and development, adhesion to the host and protection from host defenses. It has been observed that yeast cells lacking multiple PIR proteins swell, and grow slowly. The absence of individual PIR proteins have been shown in yeast to increase sensitivity to 45 plant antifungal defensive chemicals such as osmotin, and antibiotics such as hygromycin (DeGroot et al 2015). For both Cp and powdery mildew species, TG3 may help the fungi adapt to live inside living host tissue by strengthening the cell walls of invasive structures and protecting the fungi from host defensive proteins. It is also possible that the TG3 mannoprotein may migrate into host tissues and play a role there, in which case it would be a bona fide secreted effector protein. Early studies of mannoproteins in Sc found that about 5% of mannonprotein migrated into the growth medium, and that this happened at a constant rate throughout the cell cycle. The researchers believed that the released mannoprotein was either synthesized de-novo or represented mannorproteins that were non-structural in nature (Kratky et al 1975). TG4: a putative pre-protein translocase A BlastP search for homologues of the TG4 protein produced 96 non-duplicate fungal hits with an average amino acid sequence identity of 70%. Homologous proteins in other plant pathogens in the fungal order Diaporthales were especially highly- conserved, with identities over 80%. Analysis of the TG4 sequence with TMHMM 2.0 suggested one transmembrane domain located within the first 30 amino acid residues of the protein, which we initially discounted as coinciding with the N- terminal signal peptide. However, annotations of TG4 homologues, where provided, consistently referred to the Sec66 translocase, a subunit of the Sec62/63 translocation complex, which is an integral membrane protein of the endoplasmic reticulum. A literature search found no references to the TG4 being previously studied in Cp or 46 powdery mildew, but the protein was extensively examined in Sc, and its possible role and functions have been reported in some other yeasts and filamentous fungi. The Sec66 protein , also known in Sc as Sec71 and Kar7, interacts in an auxiliary manner with the Sec62/63 complex, membrane proteins which, in turn, interact with the Sec61 pore in the endoplasmic reticulum (ER) membrane. It is part of a heteromer associated with post-translational translocation, especially of secreted proteins (Rapaport 2007). The role of TG4 may not be entirely restricted to secreted proteins, however. Jung et al (2014) report evidence that the Sec66 protein also helps regulate topogenesis of membrane proteins in eukaryotic cells. The Sec66 protein is not essential for yeast cell growth or survival (Feldheim 1993). Sc sec66-null mutants were found to be viable at 30°C but not at 37°C (Feldheim 1993) but the role of the protein is clearly more than just to stabilize translocation functions at high temperatures. Sec66 has been shown to be important in several disparate cellular functions in yeast, and in other fungi. For example, Nishikawa et al (2008) found that sec66-null mutants were unable to accomplish the karyogamy associated with sexual reproduction in yeast due a failure of outer nuclear envelopes to fuse. Katta et al (2015) found that the absence of a functioning Sec66 gene led to defects in spindle pole body duplication during mitosis in yeast. Both Nishikawa and Katta noted that the defects they observed occurred at moderate temperatures (30°C) as well as at high temperatures (37°C). Lee and Heitman (2012) reported that sec66 was necessary for the completion of opposite sex and unisex mating in the dimorphous yeast Cryptococcus neoformans. Kang and Jiang (2005) found Sec66 to 47 be one of several protein secretion-related genes upregulated during the transition to filamentous growth in dimorphic yeast. It is notable that all of these cellular functions are, in turn, induced by chemical (e.g. mating pheromones) or environmental (temperature, nutrient deprivation) signals. Whether SEC66 affects the various cellular functions described above through its role as a translocator of other proteins, or in a manner entirely separate from its translocation function remains unclear. What the examples provided in the literature have in common, however, is that TG4 homologues come into play when fungal cells are experiencing environmental stress or developmental change, suggesting that this otherwise dispensable protein may be an important regulator of cellular responses necessary to cope with such changes. TG5: a putative mitochondrial carrier protein A BlastP search of the TG5 protein revealed strong amino acid sequence identity among homologues in the highest scoring fungal hits. Sequence identity averaged 84% with coverage ranging from 95 to 100%. The closest homologues to TG5 were found in other plant pathogenic fungi of the order Diaporthales, to which Cp also belongs. Numerous animal and insect pathogens were represented, but relatively few saprophytes. Most hits were annotated as “mitochondrial carrier protein,” “mitochondrial inner membrane protein,” or “mitochondrial oxaloacetate carrier protein.” A literature search found no previous examination of the TG5 gene in Cp or any powdery mildew species. The characterization of the protein as a mitochondrial carrier is based on the discovery of a homologous gene, OAC1 in Sc. 48 The Sc gene OAC1 has been localized to the mitochondrial inner membrane (Palmieri 1999). Analysis of the TG5 protein with TMHMM 2.0 initially predicted no transmembrane domains, but a graphical analysis shows up to five predicted transmembrane domains with varying degrees of probability. Annotations of the Sc OAC1 gene report three transmembrane domains that correspond to repetitions of the carrier protein sequence. OAC1 has been shown to transport oxaloacetate, sulfate and malonate into mitochondria, and to transport α-isopropylmalate (IPM) from the mitochondria to the cytosol. IPM is used in leucine biosynthesis (Palmieri 2016). There is some debate in the literature about the relative importance of transport into versus out of mitochondria, given that other pathways exist in Sc for transporting these compounds in each direction. The OAC1 gene in Sc is not essential for growth or survival. Yeast cells lacking OAC1 showed a slightly reduced growth, due to partial auxotrophy for leucine, which was correctable by the complementation of the OAC1 knockout with a plasmid carrying the OAC1 gene or by growing the yeast in media containing leucine (Marrobio 2008). The OAC1 gene’s expression is downregulated by and inhibited by α-ketoisocaproate (KIC), a precursor and metabolite of leucine produced by the mitochondria (Marrobio 2008). How these regulatory relationships are connected is not clear, but OAC1 does appear to be subject to an overarching regulatory mechanism. The fact that the TG5 homologue in powdery mildew is upregulated in 49 haustoria, and the fact that OAC1 supplies raw materials for leucine biosynthesis when leucine is scarce have a common thread--in both cases the TG5 homologues play a role in helping cells adapt to changes in the availability of nitrogen. TG6: a possible effector protein with an ML domain A BlastP search for homologues of the TG6 protein produced 83 non-duplicate hits in fungi with an average amino acid sequence identity of 27%, far lower than any of the other five target genes. Like TG1 (at 70%), most TG6 hits (60%) were among plant pathogenic fungi. Annotations of homologous proteins included “ML-domain containing protein” and “phosphatidylglycerol /phosphatidylinositol transfer protein (PG/PI-TP).” ML-proteins (MD-2-related lipid recognition proteins) were first characterized by Inohara and Nunez in 2002, who described them as “single-domain proteins predicted to form a β-rich fold containing multiple strands, and to mediate diverse biological functions through interacting with specific lipids.” The highest scoring TG6 hit was a putative PG/PI-TP transfer protein in Gc, but the Cp TG6 protein only shared 28% amino acid sequence identity with this protein. A BlastP search beginning at residue 65 of the Cp TG6 gene, the point where the predicted ML-domain begins, produced a nearly identical set of hits to the whole protein. A separate BlastP search for the first 65 amino acid residues (without the ML-domain) produced only one significant hit, a deltaproteobacterium found in marine sediment with 35% sequence identity. This appears unrelated to TG6. 50 ML proteins are found in numerous animal, plant and fungal genomes. They show no sequence homology to non-specific lipid interacting proteins, leading Inohara and Nunez (2002) to the hypothesis that they interact only with specific lipids with a diverse range of biological functions. Subsequent research has produced results consistent with this hypothesis. For example, the TG6 homologue in Aspergillis oryzae, previously characterized as a membrane-targeted lipid transfer protein with a specific affinity for phosphatidylglycerol /phosphatidylinositol, is an ML-protein of unknown function. In animals, the MD-1 and MD-2 proteins, from which the ML- domain was first characterized, are co-factors with Toll-like receptors in lipopolysaccharide signaling-based anti-bacterial immune responses (Inohara and Nunez, 2002). Berger et al (2005) observed a strong homology between the Sc gene NPC2 and the human hNPC2 gene, defects of which are implicated in the hereditary lipid storage/cholesterol metabolism disorder Niemann-Pick disease type C. Berger et al (2005) were able to restore normal cholesterol transport in hNPC2-null mutant animal cells by complementing them with NPC2 from yeast, and speculated that the gene’s function in yeast is to maintain lipid homeostasis. Menardo et al found six ML proteins among suspected effector proteins produced by the barley PM fungus Blumeria graminis (Menardo et al 2017). Research on arbuscular mychorrizal (Zeng et al 2006) and ectomycorrhizal (Sebastiana, 2017) fungi suggests that ML proteins play a role in lipid signaling in the host-symbiont interaction. 51 Information available about Cp TG6 is not sufficient to suggest where this protein localizes in the cell after entering the ER lumen, or whether it is secreted, or whether it interacts with cellular or extracellular lipids. However, TG6 appears to be the most likely of the six target genes that may encode an effector protein. With less than 300 amino acid residues, it meets criteria commonly used to screen for candidate effector proteins, including short protein length, and lack of close homologues or homology to proteins with known functions (Kim et al 2016, Sperschneider 2018). With respect to cysteine-richness, the TG6 protein is at the extreme low end (2%) of the 2%-20% range considered typical of fungal effector proteins (Lu and Edwards 2016). When we analyzed each of the target gene proteins using EffectorP v2.0 (effectorp.csiro.au/software.html), a machine learning-based platform for predicting effector proteins (Sperschneider 2018), it gave TG6 a 0.828 probability of being an effector protein. TG4 received a 0.531 probability, but its homologues are thought to be integral proteins of the ER. The other target gene proteins were assigned probabilities close to zero. The Generation and Characterization of Knockout Strains Multiple attempts at transformation Transformation was carried out on three occasions, in February 2018 (TG1-TG4 only), August 2018 (all six target genes) and November 2018 (all six target genes). At each attempt at transformation, one sample of spheroplasts was also put through 52 the transformation process without any gene disruption constructs as a positive control, in each case confirming that the spheroplasts were viable. During the February 2018 round of transformations, we were able to recover heterokaryon mycelia for each attempted knockout. These could grow on PDA + hyg medium, but development of fruiting bodies was very delayed. Each isolate was also grown on plain PDA until it produced oozing fruiting bodies, and the spores were collected in droplets of water with a 10 μl pipette. Ten μl of spore suspension was diluted in 1 mL of distilled water, and serial dilutions, ranging from 2 μl to 200 μl of this suspension were spread on PDA+hyg medium. We were only able to recover monokaryon knockout colonies for the TG4 knockout. Spores from TG1, TG2, and TG3 heterokaryon strains did not germinate on PDA+hyg medium. We attempted hyphal tip cultures for TG1, TG2 and TG3 in case spore germination was suppressed more than vegetative growth by the hygromycin. Heterokaryon mycelia from these isolates were grown in plates with WA+hyg medium, where they formed sparse and branchy mycelia, from which we excised hyphal tips under a dissecting scope and transferred them to PDA+hyg plates. While some of these hyphal tip cultures subsequently grew well, PCR visualization of the DNA segments corresponding to the target genes showed that all of the hyphal tip cultures that survived on the PDA+hyg plates were still heterokaryons. We found it not feasible with the equipment available to cut hyphal tips that did not contain multiple nuclei. The failure to generate monokaryon colonies by either single spore or hyphal tip cultures raised the question of whether the deleted wild-type genes were essential. 53 However, we also noted that the TG4 knockout was the only strain to be developed during the first transformation with a final PCR-amplification step. The TG1-TG3 knockouts were made with non-linearized plasmid DNA. When the transformation and regeneration process was repeated in August 2018, we added a final PCR amplification step for all gene disruption constructs to ensure that all of the DNA used in the transformation was linearized. The August round of transformations also added TG5 and TG6, for which gene disruption constructs had not been available in February. We also used a lower dosage of hygromycin (30 mg/mL instead of the 50 mg/mL recommended by Churchill et al) in the regeneration and selection media, in case the failure of TG1- TG3 spores to germinate was due to a naturally lower expression of those genes compared to TG4, resulting in less production of the hph enzyme. However, at 30 mg/mL, we observed that untransformed mycelia could survive and grow on solid (but not in liquid) media, making it difficult to screen out wild-type colonies. We obtained heterokaryon colonies for TG1 and TG6 at this dosage, as confirmed by PCR measurement of DNA fragments corresponding to the target genes, but we could not isolate monokaryon colonies from single spore inoculation of PDA or EPC medium amended with 30 mg/mL of hygromycin. The third round of transformations, beginning in November 2018, repeated the transformation and regeneration process for TG1, TG2, TG3, TG5 and TG6, at a hygromycin dosage of 40 mg/mL of hygromycin, based on the previous successful 54 practices of other researchers (Levine-Zhang personal communication). However, we continued to observe untransformed, wild-type colonies growing on selection medium. Use of a novel growth medium to induce spore germination The failure of spores from the TG1, TG2, TG3 and TG5 knockout strains to germinate on solid media at any hygromycin dosage, including dosages on which heterokaryon and wild-type mycelia could grow, was puzzling. We did not consider it likely that all of these genes were essential for cell survival, as none of the homologues of the target genes were found to be essential in yeast, and one possible explanation may lie in what has been observed about how some of these genes are regulated in yeast. For example, the ATG15 gene in yeast (the TG1 homologue), was found to be up- regulated during starvation-induced autophagy (Delorme-Axford 2018). The OAC1 gene (the TG5 homologue) is down-regulated by leucine (Marrobio 2008), a downstream byproduct of the OAC1 gene’s activity. The Gc homologues of all six genes are also up-regulated in the haustoria of Gc (Wu et al 2018). All of this is consistent with the hypothesis that the target genes are induced by the conditions that prevail in live plant tissue, and enable the fungus to survive there. The hygromycin resistance marker cassette we introduced includes its own constitutive fungal promoter, but some or all of the target gene loci may be subject to higher-level regulation, and the genes may only be expressed when induced by specific stimuli. For this reason, we decided to attempt a final round of single spore 55 inoculations, using heterokaryon cultures created in previous transformation rounds. Table 5 indicates the heterokaryon strains we had available. None were available for TG5. We collected asexual spores from these isolates, and spread them on two types of selection medium-- EPC medium and a chestnut induction medium (CIM)--each dosed with 50 mg/mL of hygromycin. CIM contains the trace ingredients and malt extract as EPC, but is amended with a water extract of American chestnut bark, uses sucrose in place of glucose to force the induction of digestive enzymes (Griffin 1994, pg. 135), and omits yeast extract to deprive the fungus of an exogenous source of nitrogen. (See Appendix I for the formula for CIM.) Table 5. Heterokaryon and monokaryon isolates produced in attempts to knock out TG1-TG6 by homologous gene replacement. Target Gene Heterokaryon isolates Monokaryon isolates TG1 1C, 1BB None TG2 2A None TG3 3A, 3C None TG4 4A, 4B, 4C, 4D 4A-5,4A-8,4B-1, 4B-2, 4B-3, 4B-4 TG5 None None TG6 6P, 6Q, 6R None In the final round of single spore inoculations, only spores from the TG6 knockout germinated and produced mycelia on EPC+Hyg50 medium, but fewer and more slowly than on CIM+Hyg50. The spores of all other knockout strains completely failed to germinate on EPC+Hyg50, but produced abundant mycelia on CIM+Hyg50. TG1 knockout colonies grew more slowly than the other knockout strains, taking about twice as long to produce equivalent biomass. TG3 knockouts were also 56 notable, producing significantly fewer single spore colonies than the other knockout strains. PCR visualization of the target gene segments of isolates grown from spores that germinated on CIM, however, showed that they were all either heterokaryon (likely growing from hyphal fragments picked up with the spores) or wild-type colonies. A subsequent test found that untransformed DK80 could also grow on CIM+Hyg50, albeit somewhat impaired. Evidently some property or component of CIM neutralizes the effect of hygromycin. At the end of three rounds of transformation, we were only successful in knocking out TG4. Observations of the TG4 knockout strain, in vitro. The TG4 knockout strain generated four heterokaryon isolates, TG4A-D. Eight single spore colonies grown from isolates TG4A and TG4B were cultured on PDA+hyg40 medium, and all showed normal growth rates and high resistance to hygromycin. Hyphae from eight of these single spore colonies, and from one sample of the wild- type EP155 strain of Cp (the wild-type parent strain of DK80) were transferred into separate flasks containing liquid EPC medium and allowed to incubate for several days. DNA was extracted, and used as template DNA in PCR reactions with TG4 57 flanking sequence primers. All TG4 knockouts showed bands corresponding to the predicted longer length of the TG4 gene disruption construct, while the EP155 sample showed a band corresponding the predicted length of the wild-type target gene plus flanking sequences (figure 6). Figure 6. As predicted, the segment amplified between TG4 flanking sequence primers is longer than the wild-type segment in isolate EP155 (identical to DK80). Isolate dTG4A-8 (“d” indicates deletion) was chosen as a representative sample of the TG4 knockout strain, and grown on EPC, CBA and WA media, as described in chapter 2. In vitro, dTG4A-8 exhibited consistently different growth rates and phenotype than DK80. Radial growth measures were taken twice on CBA, EPC and WA media, once directly on media in March 2018, and once on cellophane over media in March 2019. The most notable difference between dTG4A-8 and DK80 was the former’s relatively fast but sparse radial growth on synthetic media. Comparisons by Student’s t-test on both occasions, each of which involved three replicates of each fungus type on all three types of media, showed that the diameter of dTG4A-8 mycelium was significantly larger (α=0.05) than that of DK80 on nutrient rich (EPC) and nutrient poor (WA) media (table 6), but not on CBA medium. 58 Table 6. In vitro comparison of DK80 and dTG4A-8 by radial growth and weight at final day of measurement (6-7 dpi). Bold indicates the isolate with greater measurement. Medium Diameter on medium (mm) Diam. on cellophane (mm) Mycelium weight (mg) DK80 dTG4A-8 P-value DK80 dTG4A-8 P-value DK80 dTG4A-8 P-value CBA 80.33 84.0 P=0.2756 62.7 58.5 P=0.0247 10.0 7.6 P=0.007 EPC 73.7 81.3 P=0.0203 57.0 65.0 P=3.0x10-4 107.3 80.5 P=8.5x10-5 WA 54.0 60.7 P=5.0x10-4 46.5 51.7 P=5.1x10-5 2.8 1.2 P=0.026 The rapid radial growth of dTG4A-8 appears to reflect poor fitness rather than vigor. Rapid sparse mycelial expansion is a normal response of filamentous to poor nutrition (Jennings and Lysek, 1999, pg. 13). The less vigorous growth of dTG4A-8 was substantiated by comparing the weights of dTG4A-8 and DK80 mycelia grown on layers of cellophane over the same three types of media, which showed that despite its rapid radial growth (figure 7), dTG4A-8 accumulated significantly less biomass than DK80 (figure 8). Notably, dTG4A-8 showed both less radial growth and less biomass accumulation than DK80 on CBA medium. 70 60 50 40 DK80 30 dTG4A-8 20 10 0 CBA EPC WA Figure 7. Comparison of mycelial diameter of DK80 and dTG4A-8 on three types of media, at seven dpi. Error bars indicate standard error. 59 mycelial diameter at 7 dpi (mm) 1000 900 800 700 600 500 DK80 400 dTG4A-8 300 200 100 0 CBA EPC WA Figure 8. Comparison of accumulated weight of DK80 and dTG4A-8 mycelia grown on cellophane over three types of media, at seven dpi. Error bars indicate standard error. In addition to different growth rates, there were also observable phenotypic differences between dTG4A-8 and DK80 (figure 9), including: • dTG4A-8 reached each stage of development (e.g. appearance of pigment and fruiting bodies) later than DK80; • dTG4 A-8 hyphae grew in disorderly, meandering manner compared to DK80 (figure 10); • dTG4 A-8 produced less pigment than DK80 and the pigment was more tan/less orange than DK80 (figure 10); • dTG4A-8 had fewer fruiting bodies, which appeared randomly within the mycelium, while DK80’s more numerous fruiting bodies appeared mainly in concentric rings; fruiting bodies of both strains produced numerous, normal looking, viable conidiospores; 60 mycelial weight (mg) • dTG4A-8 showed no zonal growth, while zonal patterns corresponding to light cycle were distinct in DK80. Figure 9. DK80 (upper row) and dTG4A-8 (lower row) on three media, left to right: EPC, CBA and WA. Figure 10. Side by side comparison of dTG4A-8 (left) and DK80 (right) inoculated on the same day, as seen from underneath shows that DK80 is denser, more pigmented and organized more distinctly into zones than dTG4A-8. 61 Light generally plays a role in regulating zonal growth and conidiation in filamentous fungi (Griffin 1994, pg. 345), so the difference between dTG4A-8 and DK80 in zonal growth may be due to a difference in light sensitivity. To test whether this was the case, we prepared six petri dishes with PDA medium, inoculated three with each Cp strain, and incubated them for eight days at room temperature in a dark box. When removed for observation, neither strain had yet developed fruiting bodies, which normally appear as early as day two in bench-top cultures. The darkness also abolished zonal growth in DK80, but dTG4A-8 grown in darkness appeared roughly identical to dTG4A-8 grown on the benchtop, with the exception of having almost no pigment (figure 10). Figure 11. Growth in darkness abolished zonal growth in DK80 but did not change the morphology of dTG4A-8 mycelium. 62 Detached stem assay of the TG4 knockout strain The dTG4A-8 strain was less virulent than DK80 against live American chestnut stem tissue in a detached stem assay conducted in March 2019. The assay used live, dormant American chestnut branch sections 2-5 cm in diameter, which were collected in January 2019 from orchards operated by the Maryland Chapter of the American Chestnut Foundation. Fifteen inoculations with 5 mm plugs of each of the following Cp strains, grown on PDA medium, were used in the assay: • EP155: a standard virulent strain, used as a positive control; • DK80: a virulent mutant derivative of EP155, and the parent strain of dTG4A-8; • dTG4A-8: the TG4 knockout strain produced in this research; • SG2,3: a standard weakly virulent strain of Cp; and • Sterile PDA: a negative control. Cankers produced by dTG4A-8 were significantly smaller (P=0.03) at 24 days post- inoculation (dpi) than those produced by wild-type DK80. Figure 12 below shows photographs of DK80, dTG4A-8 and control inoculations on the same stem at 24 dpi. Figure 13, below, shows mean NCL for each type of inoculum. The Cp strains clustered into two groups, with the letter B representing the two statistically similar virulent strains (EP155 and DK80) and C representing the statistically similar weak strains (SG2,3 and dTG4A-8). 63 Figure 12. Photographs of DK80 (left), dTG4A-8 (center) and control inoculations (right) on the same stem, 24 dpi. Figure 13. Mean normalized canker lengths for each type of inoculum used in the detached stem assay Small stem assay of the TG4 knockout strain A small stem assay was conducted in summer 2018 on 80 American first-year chestnut seedlings from the same mother tree, using fungal strains available at that time. These included dTG4A-8, and four heterokaryon knockout strains, TG1C(h), 64 TG2A(h), TG3C(h) and TG6P(h). DK80 was used as a positive control and sterile agar as a negative control. Unfortunately, the DK80 culture used for this inoculation had been sub-cultured too many times and had lost virulence. Only one out of 12 seedlings produced a canker. The small stem assay will be repeated with new seedlings and fresh inoculum in summer 2019. Nevertheless, data collected from the 2018 small stem assay produced meaningful results. The 11 seedlings inoculated with dTG4A-8 showed no stem mortality after 98 days of observation, the same result observed for the sterile agar control. All of the heterokaryon strains, which were inoculated into 10-12 seedlings each, produced mortality in at least three. Data for mean days of survival is shown in figure 14. In terms of canker length, dTGA-8 produced significantly) longer cankers (t-test, P=0.009) than the sterile control (the cankers of which were basically healed scars), but significantly shorter (t-test, P=7x10-5) than TG3C(h), the strain that produced the next shortest mean canker length. Figure 14. Virulence as measured by days of survival post-inoculation, for the portion of the stem distal to the inoculation site. 65 Chapter 3: Conclusions, Reflection and Future Directions General Conclusions The objectives of this thesis research are laid out in Chapter 1. Had all six target genes been as easy to disrupt in Cp as the TG4 gene, these objectives could have been accomplished over the past two and a half years. However, unexpected challenges, particularly, in isolating monokaryon mutants from heterokaryon cultures, slowed the progress of the project. So far, we were only completely successful in knocking out TG4 and analyzing impact of TG4 deletion on fungal phenotype and virulence. As of April 2019, efforts to isolate additional Cp knockout strains continue. In addition, ectopic expression and silencing (via HIGS) of these TG homologs from PM fungi (i.e. Gc) in Arabidopsis is also underway, but there are no results to report at this stage. Nevertheless, the four objectives of this ambitious project have partially been accomplished, setting the stage for more thorough research in the future. The use of Cp as a surrogate for genetic study of conserved genes in biotrophic PM fungi The methods we used to select genes of interest that have homologues in Cp and Gc did yield six interesting candidate genes, which, based on detailed literature research and analysis, appear to potentially play roles in pathogenesis, either by acting on the plant host, or by enabling fungal adaptation to the host. The genetic data we obtained for the one gene were successfully knocked out, TG4, demonstrates that the gene plays a role in Cp’s virulence against chestnut, as shown by reduced canker size caused by the TG4 knockout strain in the in planta assays (figures 12-14). 66 Whether the remaining five Cp target genes contribute to virulence remains unknown due to the lack of corresponding monokaryon Cp mutants. Nor is it clear whether the TG4 homologue in Gc plays a similar role in virulence. If the results from the HIGS experiment underway are positive, it will provide more evidence that the “surrogate” approach taken in this research is useful for functional characterization of conserved fungal pathogenicity genes in genetically intractable biotrophic fungi such as PM. Identification of pathogenicity genes in plant-pathogenic fungi Although homologues of TG1-TG6 have been studied in yeast, they might have undergone functional diversification and play distinct and important roles in host colonization in plant pathogenic fungi. TG1-TG5 all appear to serve as auxiliary regulatory genes that help pathogenic fungi adjust to conditions in a live host, while the exact function of TG6 remains a mystery. If pathogenicity is broadly defined as the ability of a pathogen to overcome the defenses of its host, establish itself and obtain nutrients, and reproduce there, then the approach we have taken may help shed light on previously unstudied aspects of the host-pathogen relationship. In this sense, TG1-TG6 warrant further study in other plant-fungal pathosystems. Improving our understanding of the Arabidopsis-Gc pathosystem As of April 2019, efforts were underway to overexpress and silence Gc homologues of target genes using genetically modified Arabidopsis plants. Results, when available, may reveal new information about mechanisms of pathogenicity in Gc. 67 The nature of the six target genes, as genes that help the fungus respond to host- induced stress, has improved our understanding of the types of genes involved in colonization of the host, and changes the scope of what we could define as a pathogenicity gene. Identifying new targets for HIGS in American chestnut This research has added six new potential targets in the Cp genome for silencing by means of HIGS in transgenic American chestnut. TG4 apparently contributes to Cp virulence in chestnut, and the possibility that TG5 may be essential in Cp suggests that it may be an excellent target for silencing as well. If results from ongoing HIGS experiments demonstrate that transgenic Arabidopsis plants expressing siRNA targeting any of the Gc TG1 to TG6 genes exhibit resistance to Gc, it may be worthwhile to direct future efforts towards engineering Cp resistance in American chestnut by targeting these target genes by HIGS. Specific conclusions concerning the role of TG4 in Cp The TG4 knockout strain showed a difference in phenotype compared to the wild- type DK80 parent strain in vitro, as well as reduced virulence in planta. The research conducted within the scope of this study does not point to an exact mechanism for TG4 protein’s activity, but these results, combined with the previous characterization of the TG4 homologue in Sc, SEC66, offer some clues. 68 Morphological differences between the TG4 knockout and DK80 in vitro, suggest that the gene is involved in multiple cellular processes. For example, the lack of zonal growth in the TG4 knockout indicates that deletion of TG4 impairs Cp’s ability to sense or respond to light signals. Given its localization in yeast at the ER membrane, it is more likely that TG4 is involved in the delivery of light-sensing proteins to the cell membrane or extracellular space than that it is involved in light sensing itself. The relative lack of pigment in the TG4 knockout may also result from impaired light sensing, or simply an inability to deliver pigments or the enzymes that produce them to the right cellular locations. The disorderly growth of dTG4A-8 hyphae may also be due to sensory impairment of some kind. In addition, the rapid but sparse radial growth of TG4 knockout mycelia compared to the wild-type is typical of how fungi respond to low-nutrient environments (Jennings and Lysek 1999, pg.13), suggesting that deletion of TG4 impairs Cp’s ability to sense or take up nutrients. These disparate phenotypic changes may all be explained by the disruption of delivery of multiple specialized enzymes with a variety of functions. The reduction in fungal virulence against chestnut tissue resulting from the deletion of the TG4 gene may be the result of impairment in delivery of the same proteins that are responsible for the changes in phenotype, or in the delivery of other proteins. The fact that dTG4A-8 could not even match DK80 in radial growth on CBA medium as it did on synthetic media is notable because it indicates that absence of the TG4 gene impairs the fungus’s ability to overcome the effects of water-soluble, growth- 69 inhibiting constituents of chestnut bark. This is consistent with the idea that TG4 is important for adaptation to the host. Further study will be required to determine which TG4-interacting proteins are responsible for the loss of virulence. However, the results of in vitro and in planta observations of the TG4 knockout strain suggest that TG4 would be a good target for engineering novel forms of resistance using HIGS in transgenic American chestnut. Impairment of the gene results in a reduction in virulence but does not fundamentally impair the fungus’s ability to grow as a saprophyte or to produce asexual spores. HIGS -based resistance targeting TG4 could help artificially establish a host-pathogen equilibrium between American chestnut and Cp. Further research and future directions Closing the loop on work underway The results of the research described in this thesis demonstrate that the method used to screen for candidate virulence-related genes in two related fungi was sound. Additional small stem assays to compare the TG4-knockout Cp strain to the parent strain, DK80, will be done in the coming summer. Identification of the role of TG4 in virulence in Cp provides an opportunity to develop resistance against Cp in transgenic chestnut trees using HIGS. These promising results encourage continued efforts to obtain monokaryon knockout strains for the remaining target genes, followed by in vitro and in planta assays. 70 Determining the role of TG4 and other target gene homologs in Gc We have made DNA constructs to overexpress Gc TG4 and GcTG6, fused with green fluorescent protein, and introduced them into Arabidopsis via Agrobactgerium- mediated transformation. We will observe the transformed plants using fluorescent microscopy to pinpoint the cellular location of Gc TG4 and Gc TG6 proteins, and to observe whether the proteins have a toxic effect on the plant host. We have also prepared another construct to test the HIGS efficacy against Gc TG5. A Gc TG5 gene fragment has been cloned in the binary vector pK7WlWG2(I) which is designed for RNAi applications such as HIGS. HIGS transgenic Arabidopsis plants and isogenic wild-type plants will be inoculated with Gc spores to see if silencing Gc TG5 homologue results in resistance. Exploring differential expression of Cp target genes While our selection criteria included upregulation by two-fold or greater in haustoria for the Gc homologues of the target genes, we do not know whether or how expression of the Cp target genes changes when the fungus is growing on its host. This could be accomplished by quantitative RT-PCR or RNA-seq. This information would help clarify whether the target genes perform basic housekeeping functions, or whether they are specifically involved in adapting and colonizing the host. 71 Characterizing subcellular localization of target proteins in Cp Once we have successfully knocked out target genes, and know the right parameters for repeating the process, we can study the Cp target protein in situ. For example, the best means to study the subcellular localization of a target Cp protein is to replace them with GFP-tagged versions and observe where they accumulate under a fluorescence microscope. Using the same techniques, we can also replace the Cp target genes with their homologues from Gc and assess whether and to what degree the Gc genes can functionally complement the loss of their Cp homologues. On to transgenic chestnut A long-term goal of this project is to engineer Cp resistance via HIGS of key Cp pathogenicity or virulence genes in transgenic chestnut. The potential advantages of HIGS over the use of exotic transgenes in chestnut include the fact that, by targeting genetic sequences specific to the pathogen, there is a lower risk of off-target effects in the host or in the ecosystem. In addition, multiple dsRNA constructs targeting different pathogen genes could be included in a single dsRNA-encoding gene, greatly reducing the possibility that the pathogen could adapt to overcome the resistance (Ghag et al, 2014; Nowara et al, 2010; and Weiberg et al, 2014). The State University of New York’s College of Environmental Science and Forestry (SUNY-ESF) has developed methods to genetically transform American chestnut, but they are complex and time-consuming (Newhouse et al, 2014; Welch et al, 2007). To produce a single seedling from an embryonic chestnut cell, takes multiple simultaneous attempts over a period of 12-18 months (Bruce Levine – Linda 72 McGuigan personal communication, September 2018). It is desirable to find ways to identify and test transgene constructs before undertaking such laborious efforts. The research described in this thesis has helped identify one promising gene for silencing in Cp, TG4. Planned work to silence the Gc homologues of these genes by HIGS in Arabidopsis may provide confirmation of other promising targets. The next step would be to test if Cp is susceptible to siRNA-mediated gene silencing, which has yet to be demonstrated by any laboratory. A technique for exposing fungi in vitro to siRNAs is described in Ghag et al (2014), as a “fungal inhibition assay.” Ghag et al designed dsRNA constructs based on exonic regions of genes of interest in the banana pathogen Fusarium oxysporum f.sp. cubensis. The researchers had the dsRNA commercially synthesized, and then incubated Fusarium spores with the dsRNA in specialized buffers. Mycelia grown from these spores were then examined in vitro, and showed the morphological and growth defects predicted. This screening process allowed researchers to confirm that their genes of interest could be silenced by RNAi before going to the effort of developing transgenic plants. It would be advisable to do a similar screening with Cp prior to developing HIGS transgenic chestnut tissue or trees. Cp cultures silenced in this manner could be tested in vitro and in chestnut using the same assays used for knockout strains. If in vitro silencing of specific Cp target genes results in reduced virulence, then it would make sense to try HIGS in transgenic Chestnut trees for developing a novel form of resistance to Cp. 73 Reflections on methodology The conclusions described in this chapter are based on results achieved over two and a half years. It was relatively straightforward to delete and replace the TG4 gene in Cp and observe interesting and informative changes in phenotype and virulence. For various reasons, we have not yet isolated pure knockout strains of Cp for any of the other target genes. The methods used are described in chapter 2 for the benefit of future researchers who may wish to follow up on this work or adapt it to some other purpose. We have attempted to describe one or more approaches to achieve success at each step of both the genetic transformation of Cp and of the assays used to observe changes in fungal phenotype and virulence. However, the reader will also see that the process was not always straightforward. For that reason, this section highlights some of the pitfalls encountered so that future researchers will be best prepared for unexpected obstacles. DNA Amplification methods will vary by target gene The target genes and associated DNA fragments were idiosyncratic in their behavior. It was difficult to optimize PCR methods to create gene disruption constructs, and required significant, time-consuming trial and error. The efficiency of amplification varied wildly depending on which target gene was being manipulated. There was no universal PCR protocol that was effective for overlapping PCR or for final PCR amplification of the full-length gene disruption constructs for all target genes. In fact, variation in results for different target genes seemed related to the genetic sequence 74 itself, as it was not possible to achieve uniform results simply by changing DNA concentration, temperature, reaction times or other parts of the PCR protocol. For those gene disruption constructs which we chose to amplify as plasmid DNA in E. coli bacteria (i.e. all except TG5), we also achieved markedly different results using different combinations of vectors and competent cells for different target genes. The amplification methods described in table 3 represent what worked for us, but further experimentation may produce better results. Selection methods for knockout strains must be optimized for the target gene It is unclear why the selection methods described in Churchill et al (1990) worked so readily for TG4 knockouts but failed for all others. In all cases (except TG5, which we believe may be an essential gene in Cp spheroplasts), we remain optimistic that it is possible to produce monokaryon knockout mycelia, given enough time. The spheroplast preparation, transformation and regeneration methods described in chapter 2 worked well, and readily produced heterokaryon mycelia, containing both wild-type and knockout nuclei, for all but TG5. For TG1, TG2, TG3 and TG6, however, we got bogged down in attempts to recover monokaryon mycelia from single spore or hyphal tip cultures generated from the heterokaryon mycelia. This final stage of the homologous gene replacement process, the isolation and culture of monokaryon knockout colonies, is the most challenging. Serial dilutions are important for both spheroplasts in regeneration medium, and spore suspension in the generation of single spore colonies. Cells that produce the hph enzyme will create a hygromycin-free zone in the medium immediately around them 75 into which hyphae from neighboring spheroplasts or spores can grow. It is important to find a dilution that will put spheroplasts far enough apart that wild-type hyphae will not find hygromycin-free zones. We found that transferring mycelial plugs from solid medium to liquid media, in which hygromycin can diffuse better, enabled us to screen out wild-type mycelia, but we could not use it to distinguish between heterokaryon and monokaryon knockouts. Collecting spores for the production of single spore colonies also proved challenging. It is necessary to wait at least two weeks for Cp grown in vitro to produce fruiting bodies that ooze asexual spores. Once they do, one can use a pipette to deliver a water droplet to the tips of oozing fruiting bodies and then transfer the droplet to a larger volume of water for dilution. However, we recovered heterokaryon DNA from numerous TG2 and TG6 knockout colonies that had grown from such diluted spore suspension. As Cp asexual spores carry only one nucleus, this should not have happened – unless either dilution was not sufficient, and wild-type spores were germinating in the hygromycin-free zones surrounding mutant spores and then fusing with them by anastamosis, or we were picking up small multinucleate hyphal fragments along with the spores, and these were growing into heterokaryon colonies. In either case, we would have expected to recover at least some monokaryon colonies, but we did not, leaving an open possibility that all of the target genes aside from TG4 are essential in Cp. 76 It is also possible that if a given target gene is not constitutively expressed, a knockout strain with the hph gene stably integrated into its genome may still not produce enough hph enzyme to grow in hyg+ medium unless the medium also contains the factors necessary to induce gene expression. While the hph marker cassette contains its own constitutive promoter, the location of the target gene may also determine whether or not the whole cassette is actively transcribed. Our limited efforts to experiment with induction media were inconclusive—the medium itself clearly inhibited hygromycin, and it was not clear whether the medium also induced spore germination better than EPC medium. In any case, it may not always be feasible even to guess whether a particular target gene is constitutively expressed or needs to be induced, and if so, what induces it. Hygromycin may not be the ideal selective agent Other antibiotics or other selection methods may be less problematic than hygromycin. We learned during the course of this research, that Cp naturally has some degree of tolerance for hygromycin. We discovered this when we recovered TG5 wild-type mycelia from media containing 30 mg/mL of hygromycin. Raising the hygromycin dosage to 50 mg/mL reduced the number of wild-type colonies but did not eliminate them. The TG4 knockout, which grew perfectly well on EPC medium with 100 mg/mL of hygromycin when started from mycelial plugs, showed very delayed germination when inoculated in the form of spores. No other Cp strain would grow on EPC+hyg100 at all. This suggests that spores are more sensitive to hygromycin than hyphae. Our only successful knockout used a dosage of 50 mg/mL, 77 but this may not be the right dosage for all potential targets. As noted above, the selective dosage may depend on the expression level of the target gene. Cp tolerance for hygromycin also varies by strain. Figure 15 shows several strains of Cp growing on medium with 50 mg/mL of hygromycin. The weakly virulent standard Cp strain SG2,3 shows greater tolerance for hygromycin than DK80 or DK80’s parent strain EP155. Figure 15. Four isolates of Cp growing on Chestnut Induction Medium with 50 mg/ml of hygromycin. The dTG4A-8 isolate carries an introduced hygromycin resistance gene, while the others do not. Wild type SG2,3 nevertheless shows an ability to survive and grow in the presence of hygromycin. Contamination can mimic hygromycin resistance Even following standard laboratory procedures and working under a laminar flow hood, hygromycin-resistant airborne bacterial and fungal contaminants frequently appeared in our cultures. We were able to identify them by culturing them, extracting 78 genomic DNA, amplifying the ITS and RPB2 regions and searching the sequence ID by BlastN. We identified Aspergillis nigra, Penicillium restrictum, Sporothrix and Lasiodiplodia species, and were unable to identify several other fungal or any bacterial contaminants. Some contaminants, including the bacteria and the Penicillium fungi, were not just insensitive to hygromycin, but capable of neutralizing hygromycin in the media around them and allowing non-transformed Cp colonies to grow. We spent considerable time attempting to separate Cp colonies from contaminants, culturing them, extracting DNA and performing genomic analysis via PCR only to learn that they were wild-type DK80 colonies. It was also quite difficult to separate some contaminants from Cp. Several Penicillium isolates grew parasitically on Cp hyphae, rather than on the medium (figure 16), though they could also be cultured on nutrient-rich medium. These fungi were not always visible until we attempted to grow single spore colonies, at which time we would discover that most of the spores which germinated were contaminants. 79 Figure 16. Dark hyphae of Penicillium c.f. restrictum fungi growing parasitically on the larger clear hyphae of Cp. To manage the risk of contamination, we recommend growing isolates of all important strains developed during the course of research on filter paper, and storing the filter papers dry in a -20°C freezer, in case it is necessary to go back and make new inoculum from an uncontaminated source. Standard cultures stored at 4°C are not reliable after more than 2-3 months, as they continue to grow, lose virulence and can get contaminated in storage. For heterokaryon strains, we observed that the ratio of mutant to wild-type nuclei tended to decline over time in storage, as reflected in the relative brightness of bands appearing in gel electrophoresis. All inoculum should be the same age The morphology and virulence of Cp mycelia will change over time, and with successive subcultures. During the course of this research, we learned that DK80, and presumably any daughter strains developed from it, ages and loses virulence 80 unusually fast. Whether this is due to the impaired non-homologous end joining pathway that makes DK80 so efficient for transformation, or to some other factor is unknown. DK80 begins to exhibit abnormal phenotype (wavy margins, intense pigmentation and slow growth – figure 17) only 2 – 3 subcultures after being started from spores. Our use of an older DK80 subculture resulted in a near complete failure to produce cankers in a small stem assay attempted in summer 2018. For all in vitro and in planta observations, Cp isolates should be started from spores on PDA or EPC medium, and subcultured 1-2 times before use in inoculation assays. Figure 17. Two DK80 subcultures from the same parent culture, growing on PDA medium, inoculated on the same day. The one on left was started with a mycelial plug, while the one on the right was started from spores collected from the parent colony. 81 Appendix I Fungal Growth Media Used in this Study (all recipes given for 1 liter volume) Water Agar 20 g Difco Bacto Agar 1 liter dH20 Autoclave at 120°C for 20 mins. Potato Dextrose Broth/Agar (PDB/PDA) 24 g Difco Potato Dextrose Broth 15 g Difco Bacto Agar (omit for broth) 1 liter dH20 Autoclave at 120°C for 20 mins. (To make acidified PDA, add 4.5 ml 25% lactic acid after medium cools to 50°C) Endothia trace elements solution: 60 mg H3B03 140 mg MnCl2 x 4H20 400 mg ZnCl2 40 mg Na2MoO4 x 2H20 100 mg FeCl2 x 6H20 400 mg CuSO4 x 5H20 Autoclave at 120°C for 20 mins (FeCl2 will precipitate from solution. Stir before using.) Endothia salt solution: 24 g NH4N03 16 g KH2PO4 4 g Na2SO4 8 g KCl 2 g MgSO4 x 7H20 1g CaCl2 Add dH20 to volume of 1 liter 82 Autoclave at 120°C for 20 mins Endothia parasitica minimal (EPM) broth/agar: 54.5 ml Endothia Salt Solution 8 ml Endothia Trace Elements Solution 10 g Glucose 15 g Difco Bacto Agar (omit for broth) Autoclave at 120°C for 20 mins 2 mg Thiamine hydrocholoride (added as a filter-sterilize aqueous solution after autoclaving) Endothia parasitica complete (EPC) broth/agar: 54.5 ml Endothia Salt Solution 8 ml Endothia Trace Elements Solution 10 g Glucose 25 g Difco yeast extract 75 g Difco malt extract 15 g Difco Bacto Agar (omit for broth) Autoclave at 120°C for 20 mins 2 mg Thiamine hydrochloride (added as a filter-sterilize aqueous solution after autoclaving) Chestnut bark extract (CBE): 100 g dried bark strips, approximately 5 mm wide and 50mm long, stripped from washed, surface disinfested chestnut stems, 3-10 cm in diameter. (Note: fungus grows better on winter harvested bark, and Cp grows better on bark from American chestnut than Chinese chestnut) 1 liter dH20 Allow bark strips to steep at room temperature for 24 hours, then allow another 24-48 hours at 4°C to prevent fermentation. Filter through paper filter, then autoclave at 120°C for 20 mins (can be pH adjusted with NaOH, HCl or lactic acid) Chestnut Bark Agar (CBA): Prepare 500 ml dH20 20 g Difco Bacto Agar And separately prepare 83 Chestnut Bark Extract Autoclave in separate flasks at 120°C for 20 mins. Allow media to cool under hood to 50°C, then mix thoroughly by swirling and pour. (Note: the tannic acid in CBE will hydrolyze the agar if autoclaved together, resulting in a soft, semi-solid medium.) Cp Induction Medium: Prepare 500 ml dH20 54.5 ml Endothia Salt Solution 8 ml Endothia Trace Elements Solution 5 g Sucrose (do not use glucose, which can inhibit induction of some genes) 3.9 g Difco Malt Extract 20 g Difco Bacto Agar and separately prepare Chestnut Bark Extract Autoclave in separate flasks at 120°C for 20 mins. Allow media to cool under hood to 50°C, then mix thoroughly by swirling. Then add: 2 mg Thiamine hydrochloride (added as a filter-sterilize aqueous solution after autoclaving) Pour before it solidifies. (Note: thiamine is temperature sensitive and will break down if media is reheated in the microwave. Also, the tannic acid in CBE will hydrolyze the agar if mixed and then brought to a high temperature, resulting in a soft, semi-solid medium.) 84 Appendix II Spheroplast Preparation, Transformation and Regeneration A. Buffers and media needed Osmotic Medium (for 500 ml): 1.2M MgSO4 (use MgSO4•7H2O [148g /500ml] which dissolves easily in water, combining anhydrous MgSO4 with water is very exothermic) in 10mM NaH2PO4 (use NaH2PO4•H2O, 0.69g /500ml) ; adjust to pH 5.8 with 1M Na2HPO4 (0.5M Na2HPO4 is easier to prepare, add the chemical slowly to water). Note that a precipitate forms while adjusting the pH so add the 0.5M Na2HPO4 slowly; 50-100ml may be necessary to reach pH 5.8. Filter sterilize and store at 4°C. Digestion Buffer (for 50 ml) (Note: Prepare fresh for each use): 50 ml osmotic medium 500 mg bovine serum albumen 100 mg lysing enzyme from Trichoderma harziana 1,000 mg vintastepro enzyme (beta D glucanase) 500 ul of beta-glucuronidase Add ingredients in order listed, and allow BSA to dissolve thoroughly before adding the next ingredients. Trapping Buffer 0.4 M sorbitol (36.4 g/ 500ml, or 200ml of 1M sorbitol/500ml) in 100mM Tris-HCl, pH 7.0 (50ml of a 1M solution); autoclave and store at 4°C. STC 1M sorbitol (91.1g/500ml) in 100mM Tris-HCl, pH 8.0 (50ml of a 1M solution) and 100mM CaCl2- dihydrate (7.35g/500ml); autoclave and store at 4°C. PTC 40% polyethylene glycol 4000MW (80g / 200ml), 100mM Tris-Hcl, pH8.0 (20ml of a 1M solution), 100mM CaCl2-dihydrate (2.94g /200ml); autoclave and store at room temperature. 85 Spheroplast Storage Buffer 4 parts STC, 1 part PTC, 0.05 parts DMSO Regeneration Medium (400 ml) 390 ml dH20 171 g sucrose 0.5 g yeast extract 0.5 g casein hydrolysate 8.9 g Difco Bacto agar Dissolve the sucrose in the dH20, add the other ingredients, with agar added last. Autoclave at 120°C for 20 minutes.. B. Cryphonectria parasitica spheroplast preparation: 1) Use a pipette tip to “cotton ball” Cp hyphae from colonies grown on PDA medium. Transfer to 1.5 ml centrifuge tubes. Wash with dH20 and grind with plastic micro-pestle (about 50 twists of the pestle will create a suspension 86 of finely fragmented hyphal tissue). Inoculate 100 ml PDB medium (1 good cotton ball from one fresh PDA culture should be enough). Incubate the culture on bench top over 3 days. You should have white clouds of mycelia. If it coats the bottom of the flask, or turns orange, try again starting with the youngest hyphae you can get. 2) When ready for spheroplast preparation, make fresh Digestion Buffer When preparing, add osmotic medium to tube, add dry ingredients starting with BSA, and allow it to dissolve before adding the other dry ingredients. Add beta-glucuronidase last. Mix by inversion. 3) Place mycelium suspension in sterile, 50 ml disposable centrifuge tubes and spin down at room temperature at 4000 rpm for 5 mins. Pour off supernatant. Combine contents so that mycelia from 100 ml of media are in one 50 ml tube. Wash by gently adding ~25 ml of autoclaved dH20. Spin down at 4000 rpm for 3 min. Repeat wash and spin. If necessary to remove moisture, you can pour the mycelia onto autoclaved filter paper and blot. 4) Re-suspend pellet in 25-30 ml of Digestion Buffer, screw the lid on tight, and incubate horizontally at 30C overnight at a very low speed, e.g. 50 rpm. At the end of about 16 hours, you should have homogenous cloudy suspension. If you have a lot of lumps or the mycelium appears unchanged, it has not digested properly. 5) Very carefully add cold, sterile Trapping Buffer to overlay the spheroplast suspension (about 25 ml, or whatever amount is required to fill the tube). Be careful not to disrupt the spheroplast layer, which should look like a cloudy area of trapping buffer that is densest closer to the margin with the digestion buffer..) Use a 25 ml pipette to gently add the trapping buffer to the side of the tilted 50 ml tube. Spin down again at room temperature at about ~4000 rpm) 6) Collect the spheroplasts (the cloudy layer) at the interface. Transfer to a new 50 ml tube, and dilute with 2 vols of 1M sorbitol. Gently mix thoroughly. Pellet the spheroplasts in rotor at 6,000 rpm, 4°C, 5 min. 7) Remove supernatant by aspiration. Suspend the spheroplast pellet in 100-200 ul of STC. If you have multiple tubes, you can use the same suspension over, transferring from tube to tube until everything is suspended and combined in one tube. You can clean the last spheroplasts from the other tubes with additional fresh STC and transfer that to your main tube. Pellet everything down again, as in step 4. 87 8) Remove supernatant by aspiration. Suspend spheroplasts in STC on ice. Use enough STC (normally 200-1000 ul) so that you have a slightly cloudy homogeneous solution. If very cloudy, dilute more. 9) Observe up to 10 ul of suspension under dissecting scope with a hemacytometer. Spheroplasts are spherical and about 5 um across, and will look under the scope like craters on the moon. See picture below. If you have very high concentrations, you can dilute until you have a final concentration of 2 x 108 cells/ml. 10) Dilute your spheroplast suspension with the following solution: 4 parts STC, 1 part PTC and 0.05 parts DMSO. For example: 4ml STC, 1ml PTC and 50µl DMSO. Freeze in 50 ul aliquots in cryovials and store at-80°C. C. Spheroplast transformation and regeneration 1) If spheroplasts are frozen, quickly thaw in a 37º C water bath until a 2mm diameter ice crystal remains, then place on ice. Spheroplasts should be at a concentration of ~ 2 x 107 cells/ml. 88 2) Pre-cool 1.5 ml tubes on ice, one for each transformation, plus a control. Add 5-10µg of DNA in a volume of 10µl to each tube and 10µl of TE buffer to one other for a control. 3) Add 50µl of spheroplasts to each tube, mix gently, and chill on ice for 30 min. 4) Add 500 ul of PTC to each tube, mix gently and incubate at room temperature for 25 min. THIS IS THE TRANFORMATION STEP. 5) Add 1ml of STC to each tube, mix gently. 6) Evenly distribute the reaction mixtures by droplets onto empty, labeled petri dishes. Make serial dilutions to ensure the separation of individual transformants. A suggested series is 2µl, 20µl and 200µl. Pipet 11 ml of Regeneration Medium at 48º C onto each plate and swirl to mix. Incubate on bench top for 16-18 hours, then add 11 ml of Regeneration Medium containing 100µg/ml Hygromycin B (or other antibiotic) at 48º C as a top layer. Do not add the top layer to the cells mixed with the TE buffer unless you wish to test the antibiotic. (The previous two steps can also be combined with one stem involving 22 mlRegeneration Medium+antibiotic.) Note that the regeneration medium is sticky and difficult to handle. Keep it at 48-50 C in a water bath near where you are doing the inoculations, to prevent it from congealing. Especially if you are doing the two-layer method, gaps or bubbles can allow non-transformed fungus to grow around the selective medium layer. 7) Incubate on bench top for 3-5 days until hyphae can be seen growing through the top layer. Pick individual putative transformants to PDA plates containing antibiotic. Also transfer hyphae from TE control plates to PDA without antibiotic to compare the phenotype to that of the putative transformants. 89 90 Bibliography Anagnostakis, S. L. (1987). Chestnut Blight: The Classical Problem of an Introduced Pathogen. Mycologia, 79(1), 23–37. https://doi.org/10.2307/3807741 Barakat, A., DiLoreto, D. S., Zhang, Y., Smith, C., Baier, K., Powell, W. A., … Carlson, J. E. (2009). Comparison of the transcriptomes of American chestnut (Castanea dentata) and Chinese chestnut (Castanea mollissima) in response to the chestnut blight infection. BMC Plant Biology, 9, 51. https://doi.org/10.1186/1471-2229-9-51 Barakat, A., Staton, M., Cheng, C.-H., Park, J., Yassin, N. M. B., Ficklin, S., … Sederoff, R. (2012). Chestnut resistance to the blight disease: insights from transcriptome analysis. BMC Plant Biology, 12, 38. https://doi.org/10.1186/1471-2229-12-38 Baulcombe, D. C. (2015). VIGS, HIGS and FIGS: small RNA silencing in the interactions of viruses or filamentous organisms with their plant hosts. Current Opinion in Plant Biology, 26, 141–146. https://doi.org/10.1016/j.pbi.2015.06.007 Bent, Andrew. (2018). Plant diseases and strategies for their control. In Plants, Genes and Agriculture: Sustainability through biotechnology (pp. 374–402). Oxford Univeristy Press. Berger, A. C., Vanderford, T. H., Gernert, K. M., Nichols, J. W., Faundez, V., & Corbett, A. H. (2005). Saccharomyces cerevisiae Npc2p is a functionally conserved homologue of the human Niemann-Pick disease type C 2 protein, hNPC2. Eukaryotic Cell, 4(11), 1851–1862. https://doi.org/10.1128/EC.4.11.1851-1862.2005 91 Bigeard, J., Colcombet, J., & Hirt, H. (2015). Signaling Mechanisms in Pattern-Triggered Immunity (PTI). Molecular Plant, 8(4), 521–539. https://doi.org/10.1016/j.molp.2014.12.022 Boutrot, F., & Zipfel, C. (2017). Function, Discovery, and Exploitation of Plant Pattern Recognition Receptors for Broad-Spectrum Disease Resistance. Annual Review of Phytopathology, 55(1), 257–286. https://doi.org/10.1146/annurev-phyto-080614- 120106 Büschges, R., Hollricher, K., Panstruga, R., Simons, G., Wolter, M., Frijters, A., … Schulze-Lefert, P. (1997). The Barley Mlo Gene: A Novel Control Element of Plant Pathogen Resistance. Cell, 88(5), 695–705. https://doi.org/10.1016/S0092- 8674(00)81912-1 Cai, Q., Qiao, L., Wang, M., He, B., Lin, F.-M., Palmquist, J., … Jin, H. (2018). Plants send small RNAs in extracellular vesicles to fungal pathogen to silence virulence genes. Science, 360(6393), 1126–1129. https://doi.org/10.1126/science.aar4142 Cárdenas, D., Revuelta-Cervantes, J., Jiménez-Díaz, A., Camargo, H., Remacha, M., & Ballesta, J. P. G. (2012). P1 and P2 protein heterodimer binding to the P0 protein of Saccharomyces cerevisiae is relatively non-specific and a source of ribosomal heterogeneity. Nucleic Acids Research, 40(10), 4520–4529. https://doi.org/10.1093/nar/gks036 Churchill, A. C. L., Ciuffetti, L. M., Hansen, D. R., Etten, H. D. V., & Alfen, N. K. V. (1990). Transformation of the fungal pathogen Cryphonectria parasitica with a variety of heterologous plasmids. Current Genetics, 17(1), 25–31. https://doi.org/10.1007/BF00313245 92 Chrispeels, M. J. (2017). Sustainable Food Production in the 21st Century. In Plants, Genes and Agriculture: Sustainability through Biotechnology (pp. 584–603). New York: Oxford Univeristy Press. Cline, W. R. (2007). Global Warming and Agriculture: Impact Estimates by Country. Washington, DC: Peterson Institute for International Economics. Consonni, C., Humphry, M. E., Hartmann, H. A., Livaja, M., Durner, J., Westphal, L., … Panstruga, R. (2006). Conserved requirement for a plant host cell protein in powdery mildew pathogenesis. Nature Genetics, 38(6), 716–720. https://doi.org/10.1038/ng1806 Dawe, A. L., Segers, G. C., Allen, T. D., McMains, V. C., & Nuss, D. L. (2004). Microarray analysis of Cryphonectria parasitica Gα- and Gβγ-signalling pathways reveals extensive modulation by hypovirus infection. Microbiology, 150(12), 4033– 4043. https://doi.org/10.1099/mic.0.27339-0 De Groot, P. W. J., Ram, A. F., & Klis, F. M. (2005). Features and functions of covalently linked proteins in fungal cell walls. Fungal Genetics and Biology, 42(8), 657–675. https://doi.org/10.1016/j.fgb.2005.04.002 Delorme-Axford, E., & Klionsky, D. J. (2018). Transcriptional and post-transcriptional regulation of autophagy in the yeast Saccharomyces cerevisiae. Journal of Biological Chemistry, 293(15), 5396–5403. https://doi.org/10.1074/jbc.R117.804641 Dutton, M. V., & Evans, C. S. (1996). Oxalate production by fungi: its role in pathogenicity and ecology in the soil environment. Canadian Journal of Microbiology, 42(9), 881–895. https://doi.org/10.1139/m96-114 93 Elliston, J. (1978). Pathogenicity and Sporulation of Normal and Diseased Strains of Endothia parasitica in American Chestnut. Proceedings of the American Chestnut Symposium, West Virginia University, Jan. 4-5, 1978, 95–100. Epple, U. D., Suriapranata, I., Eskelinen, E.-L., & Thumm, M. (2001). Aut5/Cvt17p, a Putative Lipase Essential for Disintegration of Autophagic Bodies inside the Vacuole. Journal of Bacteriology, 183(20), 5942–5955. https://doi.org/10.1128/JB.183.20.5942-5955.2001 Feldheim, D., Yoshimura, K., Admon, A., & Schekman, R. (1993). Structural and functional characterization of Sec66p, a new subunit of the polypeptide translocation apparatus in the yeast endoplasmic reticulum. Molecular Biology of the Cell, 4(9), 931–939. Gao, S., & Shain, L. (1995). Activity of polygalacturonase produced by Cryphonectria parasitica in chestnut bark and its inhibition by extracts from American and Chinese chestnut. Physiological and Molecular Plant Pathology, 46(3), 199–213. https://doi.org/10.1006/pmpp.1995.1016 Ghag, S. B., Shekhawat, U. K. S., & Ganapathi, T. R. (2014). Host-induced post- transcriptional hairpin RNA-mediated gene silencing of vital fungal genes confers efficient resistance against Fusarium wilt in banana. Plant Biotechnology Journal, 12(5), 541–553. https://doi.org/10.1111/pbi.12158 Gonsalves, D. (1998). Control of Papaya Ringspot Virus in Papaya: A Case Study. Annual Review of Phytopathology, 36(1), 415–437. https://doi.org/10.1146/annurev.phyto.36.1.415 Griffin, D. (n.d.). Fungal Physiology (2nd Edition). New York: Wiley-Liss. 94 Griffin, G. J., Elkins, J. R., McCurdy, D., & Griffin, S. (2006). Integrated use of resistance, hypovirulence, and forest management to control blight in American chestnut. United States National Park Service Natural Resources Report NPS/NCR/CUE/NRR- 2006/001, 97–107. Grünwald, N. J. (2012). Novel insights into the emergence of pathogens: the case of chestnut blight. Molecular Ecology, 21(16), 3896–3897. https://doi.org/10.1111/j.1365-294X.2012.05597.x Han, L., & Luan, Y.-S. (2015). Horizontal Transfer of Small RNAs to and from Plants. Frontiers in Plant Science, 6. https://doi.org/10.3389/fpls.2015.01113 Havir, E. A., & Anagnostakis, S.L. (1983). Oxalate production by virulent but not by hypovirulent strains of Endothia parasitica. Physiol. Mol Plant Pathology, 23, 369– 376. Hebard, F. V., & Fi. (2014). The breeding program of the American chestnut foundation. Acta Horticulturae. Retrieved from http://www.actahort.org/books/1019/1019_20.htm Hebard, F. V. (2005). The Backcross Breeding Program of the American Chestnut Foundation. The Journal of the American Chestnut Foundation, 19(2), 55–78. Hebard, F. V., Griffin, G. J., & Elkins, J. R. (1984). Developmental Histopathology of Cankers Induced by Hypovirulent and Virulent Isolates of Endothia parasitica on Susceptible and Resistant Chestnut Trees. Phytopathology, 74, 140–149. Hsu, P.-H., Chiang, P.-C., Liu, C.-H., & Chang, Y.-W. (2015). Characterization of Cell Wall Proteins in Saccharomyces cerevisiae Clinical Isolates Elucidates Hsp150p in Virulence. PLoS ONE, 10(8). https://doi.org/10.1371/journal.pone.0135174 95 Inohara, N., & Nuñez, G. (2002). ML -- a conserved domain involved in innate immunity and lipid metabolism. Trends in Biochemical Sciences, 27(5), 219–221. Jacob-Wilk, D., Moretti, M., Turina, M., Kazmierczak, P., & Van Alfen, N. K. (2012). Differential expression of the putative Kex2 processed and secreted aspartic proteinase gene family of Cryphonectria parasitica. Fungal Biology, 116(3), 363–378. https://doi.org/10.1016/j.funbio.2011.12.006 Jennings, D. H., & Lysek, G. (1999). Fungal Biology: Understanding the Fungal Lifestyle (2nd Edition). Oxford, UK: BIOS Scientific Publishers Ltd. Jones, J. D. G., Vance, R. E., & Dangl, J. L. (2016). Intracellular innate immune surveillance devices in plants and animals. Science, 354(6316), aaf6395–aaf6395. https://doi.org/10.1126/science.aaf6395 Jung, S., Kim, J. E. H., Reithinger, J. H., & Kim, H. (2014). The Sec62–Sec63 translocon facilitates translocation of the C-terminus of membrane proteins. J Cell Sci, 127(19), 4270–4278. https://doi.org/10.1242/jcs.153650 Jung, U. S., & Levin, D. E. (1999). Genome-wide analysis of gene expression regulated by the yeast cell wall integrity signalling pathway. Molecular Microbiology, 34(5), 1049–1057. https://doi.org/10.1046/j.1365-2958.1999.01667.x Kang, C. M., & Jiang, Y. W. (2005). Genome-wide survey of non-essential genes required for slowed DNA synthesis-induced filamentous growth in yeast. Yeast, 22(2), 79–90. https://doi.org/10.1002/yea.1195 Katta, S. S., Chen, J., Gardner, J. M., Friederichs, J. M., Smith, S. E., Gogol, M., … Jaspersen, S. L. (2015). Sec66-Dependent Regulation of Yeast Spindle-Pole Body 96 Duplication Through Pom152. Genetics, 201(4), 1479–1495. https://doi.org/10.1534/genetics.115.178012 Kim, J.-M., Park, J.-A., & Kim, D.-H. (2012). Comparative proteomic analysis of chestnut blight fungus, Cryphonectria parasitica , under tannic-acid-inducing and hypovirus- regulating conditions. Canadian Journal of Microbiology, 58(7), 863–871. https://doi.org/10.1139/w2012-065 Kim, K.-T., Jeon, J., Choi, J., Cheong, K., Song, H., Choi, G., … Lee, Y.-H. (2016). Kingdom-Wide Analysis of Fungal Small Secreted Proteins (SSPs) Reveals their Potential Role in Host Association. Frontiers in Plant Science, 7. Klis, F. M., Boorsma, A., & Groot, P. W. J. D. (2006). Cell wall construction in Saccharomyces cerevisiae. Yeast, 23(3), 185–202. https://doi.org/10.1002/yea.1349 Klis, F. M., Mol, P., Hellingwerf, K., & Brul, S. (2002). Dynamics of cell wall structure in Saccharomyces cerevisiae. FEMS Microbiology Reviews, 26(3), 239–256. https://doi.org/10.1111/j.1574-6976.2002.tb00613.x Koch, A., Kumar, N., Weber, L., Keller, H., Imani, J., & Kogel, K.-H. (2013). Host- induced gene silencing of cytochrome P450 lanosterol C14α-demethylase–encoding genes confers strong resistance to Fusarium species. Proceedings of the National Academy of Sciences, 110(48), 19324–19329. https://doi.org/10.1073/pnas.1306373110 Kuhn, H., Kwaaitaal, M., Kusch, S., Acevedo-Garcia, J., Wu, H., & Panstruga, R. (2016). Biotrophy at Its Best: Novel Findings and Unsolved Mysteries of the Arabidopsis- Powdery Mildew Pathosystem. The Arabidopsis Book, 14, e0184. https://doi.org/10.1199/tab.0184 97 Kratky, Z., Biely, P., & Bauer, S. (1975). Wall mannan of Saccharomyces cerevisiae. Metabolic stability and release into growth medium. Biochim. Biophys. Acta, 404, 1– 6. Lan, X., Yao, Z., Zhou, Y., Shang, J., Lin, H., Nuss, D. L., & Chen, B. (2008). Deletion of the cpku80 gene in the chestnut blight fungus, Cryphonectria parasitica, enhances gene disruption efficiency. Current Genetics, 53(1), 59–66. https://doi.org/10.1007/s00294-007-0162-x Lee, S. C., & Heitman, J. (2012). Function of Cryptococcus neoformans KAR7 (SEC66) in Karyogamy during Unisexual and Opposite-Sex Mating. Eukaryotic Cell, 11(6), 783– 794. https://doi.org/10.1128/EC.00066-12 Lo Presti, L., Lanver, D., Schweizer, G., Tanaka, S., Liang, L., Tollot, M., … Kahmann, R. (2015). Fungal Effectors and Plant Susceptibility. Annual Review of Plant Biology, 66(1), 513–545. https://doi.org/10.1146/annurev-arplant-043014-114623 Marobbio, C. M. T., Giannuzzi, G., Paradies, E., Pierri, C. L., & Palmieri, F. (2008). alpha-Isopropylmalate, a leucine biosynthesis intermediate in yeast, is transported by the mitochondrial oxalacetate carrier. The Journal of Biological Chemistry, 283(42), 28445–28453. https://doi.org/10.1074/jbc.M804637200 McCarroll, D. R., & Thor, E. (1978). The role of oxalic acid in the pathogenesis of Endothia parasitica. Proceedings of the American Chestnut Symposium, West Virginia Univeristy, 122. Menardo, F., Praz, C. R., Wicker, T., & Keller, B. (2017). Rapid turnover of effectors in grass powdery mildew (Blumeria graminis). BMC Evolutionary Biology, 17(1). https://doi.org/10.1186/s12862-017-1064-2 98 Milgroom, M. G., & Cortesi, P. (2004). Biological Control of Chestnut Blight with Hypovirulence: A Critical Analysis. Annual Review of Phytopathology, 42(1), 311– 338. Newhouse, A. E., Polin-McGuigan, L. D., Baier, K. A., Valletta, K. E. R., Rottmann, W. H., Tschaplinski, T. J., … Powell, W. A. (2014). Transgenic American chestnuts show enhanced blight resistance and transmit the trait to T1 progeny. Plant Science, 228, 88–97. https://doi.org/10.1016/j.plantsci.2014.04.004 Nishikawa, S., Hirata, A., & Endo, T. (2008). Nuclear inner membrane fusion facilitated by yeast Jem1p is required for spindle pole body fusion but not for the first mitotic nuclear division during yeast mating. Genes to Cells, 13(11), 1185–1195. Nowara, D., Gay, A., Lacomme, C., Shaw, J., Ridout, C., Douchkov, D., … Schweizer, P. (2010). HIGS: Host-Induced Gene Silencing in the Obligate Biotrophic Fungal Pathogen Blumeria graminis. Plant Cell, 22(9), 3130–3141. https://doi.org/10.1105/tpc.110.077040https://doi.org/10.1111/j.1365- 2443.2008.01236.x Nuss, D. L. (2011). Mycoviruses, RNA Silencing, and Viral RNA Recombination. In K. Maramorosch, A. J. Shatkin, & F. A. Murphy (Eds.), Advances in Virus Research, Vol 80 (Vol. 80, pp. 25–48). San Diego: Elsevier Academic Press Inc. Palmieri, L., Vozza, A., Agrimi, G., De Marco, V., Runswick, M. J., Palmieri, F., & Walker, J. E. (1999). Identification of the yeast mitochondrial transporter for oxaloacetate and sulfate. The Journal of Biological Chemistry, 274(32), 22184– 22190. 99 Parzych, K. R., Ariosa, A., Mari, M., & Klionsky, D. J. (2018). A newly characterized vacuolar serine carboxypeptidase, Atg42/Ybr139w, is required for normal vacuole function and the terminal steps of autophagy in the yeast Saccharomyces cerevisiae. Molecular Biology of the Cell, 29(9), 1089–1099. https://doi.org/10.1091/mbc.E17- 08-0516 Piffanelli, P., Ramsay, L., Waugh, R., Benabdelmouna, A., D’Hont, A., Hollricher, K., … Panstruga, R. (2004). A barley cultivation-associated polymorphism conveys resistance to powdery mildew. Nature, 430(7002), 887. https://doi.org/10.1038/nature02781 Prospero, S., & Rigling, D. (2016). Using molecular markers to assess the establishment and spread of a mycovirus applied as a biological control agent against chestnut blight. Biocontrol, 61(3), 313–323. https://doi.org/10.1007/s10526-015-9713-0 Ramya, V., & Rajasekharan, R. (2016). ATG15 encodes a phospholipase and is transcriptionally regulated by YAP1 in Saccharomyces cerevisiae. FEBS Letters, 590(18), 3155–3167. https://doi.org/10.1002/1873-3468.12369 Rapoport, T. A. (2007). Protein translocation across the eukaryotic endoplasmic reticulum and bacterial plasma membranes. Nature, 450(7170), 663–669. https://doi.org/10.1038/nature06384 Rapoport, T. A., Li, L., & Park, E. (2017). Structural and Mechanistic Insights into Protein Translocation. Annual Review of Cell and Developmental Biology, 33(1), 369–390. https://doi.org/10.1146/annurev-cellbio-100616-060439 Remacha, M., Jiménez-Díaz, A., Santos, C., Briones, E., Zambrano, R., Rodriguez Gabriel, M. L., … Ballesta, J. P. G. (1995). Proteins P1, P2 and P0, components of 100 the eukaryotic ribosome stalk. New structural and functional aspects. Biochem. Cell Biol., 73, 959–968. Rigling, D., & Prospero, S. (2018). Cryphonectria parasitica, the causal agent of chestnut blight: invasion history, population biology and disease control. Molecular Plant Pathology, 19(1), 7–20. https://doi.org/10.1111/mpp.12542 Roane, M. K., Griffin, G. J., & Elkins, J. R. (1986). Chestnut blight, other Endothia diseases, and the genus Endothia. In APS Monograph Series. St. Paul, Minn: APS Press, American Phytopathological Society. Saielli, T. M., & Levine, B. J. (2019). Small Stem Assays may be a Reliable Screening Tool for Testing American Chestnut Resistance to Cryphonectria parasitica. Chestnut, 33(1). Samman, S., Schell, F., & Thor, E. (1979). Effect of Some Compounds of American and Chinese Chestnut Inner Bark on the Growth of Endothia Parasitica. Proceedings of the American Chestnut Symposium, West Virginia University, Jan. 4-5, 1978, 70–72. Schlarbaum, S., Anagnostakis, S. L., & Morton, M. C. (1992). Evaluation of experimental chestnut plantings in eastern North America. Proceedings of the International Chestnut Conference, MOrgantown, WV. Sebastiana, M., Martins, J., Figueiredo, A., Monteiro, F., Sardans, J., Peñuelas, J., … Coelho, A. V. (2017). Oak protein profile alterations upon root colonization by an ectomycorrhizal fungus. Mycorrhiza, 27(2), 109–128. https://doi.org/10.1007/s00572- 016-0734-z Steiner, K. C., Westbrook, J. W., Hebard, F. V., Georgi, L. L., Powell, W. A., & Fitzsimmons, S. F. (2017). Rescue of American chestnut with extraspecific genes 101 following its destruction by a naturalized pathogen. New Forests, 48(2), 317–336. https://doi.org/10.1007/s11056-016-9561-5 United Nations, Economic and Social Affairs, Population Division. (2017). World Population Prospects: the 2017 Revision, Key Findings and Advance Tables. (No. ESA/P/WP/248). New York. Wang, J., Shi, L., He, X., Lu, L., Li, X., & Chen, B. (2016). Comparative Secretome Analysis Reveals Perturbation of Host Secretion Pathways by a Hypovirus. Scientific Reports, 6, 34308. https://doi.org/10.1038/srep34308 Weiberg, A., Wang, M., Bellinger, M., & Jin, H. (2014). Small RNAs: A New Paradigm in Plant-Microbe Interactions. Annual Review of Phytopathology, 52(1), 495–516. https://doi.org/10.1146/annurev-phyto-102313-045933 Welch, A. J., Stipanovic, A. J., Maynard, C. A., & Powell, W. A. (2007). The effects of oxalic acid on transgenic Castanea dentata callus tissue expressing oxalate oxidase. Plant Science, 172(3), 488–496. https://doi.org/10.1016/j.plantsci.2006.10.015 Westbrook, J. W. (2018). Merging backcross breeding and transgenic blight resistance to accelerate restoration of the American chestnut: The American Chestnut Foundation’s breeding and selection plan 2015-2025. Retrieved from American Chestnut Foundation website: https://www.acf.org/wp- content/uploads/2018/03/TACF_2015- 2025_BreedingSciencePlan_updated2_9_18.pdf Wu, Y., Ma, X., Pan, Z., Kale, S. D., Song, Y., King, H., … Xiao, S. (2018). Comparative genome analyses reveal sequence features reflecting distinct modes of host-adaptation 102 between dicot and monocot powdery mildew. BMC Genomics, 19(1), 705. https://doi.org/10.1186/s12864-018-5069-z Xiao, S., Ellwood, S., Calis, O., Patrick, E., Li, T., Coleman, M., & Turner, J. G. (2001). Broad-Spectrum Mildew Resistance in Arabidopsis thaliana Mediated by RPW8. Science, 291(5501), 118–120. https://doi.org/10.1126/science.291.5501.118 Xiao, S., Emerson, B., Ratanasut, K., Patrick, E., O’Neill, C., Bancroft, I., & Turner, J. G. (2004). Origin and Maintenance of a Broad-Spectrum Disease Resistance Locus in Arabidopsis. Molecular Biology and Evolution, 21(9), 1661–1672. https://doi.org/10.1093/molbev/msh165 Zeng, T., Holmer, R., Hontelez, J., Lintel‐Hekkert, B. te, Marufu, L., Zeeuw, T. de, … Limpens, E. (2018). Host- and stage-dependent secretome of the arbuscular mycorrhizal fungus Rhizophagus irregularis. The Plant Journal, 94(3), 411–425. https://doi.org/10.1111/tpj.13908 Zhang, R., Ren, D., & Zhu, B. (1998). A study of chestnut blight pathogenicity and discussion of the pathogen as a hybrid system. Journal of Zhejiang Forest Science and Technology, 18(No. 3), 8–12. 103 104