Combining Applied and Basic Research

Understanding the Powdery Mildew Disease of the Ornamental Plant Phlox:

Dissertation

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy

in the Graduate School of The Ohio State University

Coralie Farinas

Graduate Program in Plant Pathology

The Ohio State University

2020

Dissertation Committee

Dr. Francesca Peduto Hand, Advisor

Dr. Pablo S. Jourdan

Dr. Thomas K. Mitchell

Dr. Pierce A. Paul

Dr. Jason C. Slot

Copyrighted by

Coralie Farinas

2020

Abstract

The characterization of plant germplasm has tremendous potential to help address the many challenges that the field of plant health is facing, such as climate change continuously modifying the regions of previously known disease occurrence. The worldwide trade of the plant genus Phlox represents an important revenue for the horticultural industry. However, Phlox species are highly susceptible to the fungal disease powdery mildew (PM), and infected materials shipping across borders accelerate the risk of disease spread. Through collaboration with laboratories in the U.S., we investigated the genotypic and phenotypic diversity of a PM population to better understand its capacity to adapt to new environments and new resistant hosts. To do this, we developed tools to grow and study PM pathogens of Phlox in vitro, and then used whole genome comparison and multilocus sequence typing (MLST) analysis to study the genetic structure of the population. Additionally, we explored Phlox germplasm diversity to identify a range of plant responses to PM infection by comparing disease severity progression and length of latency period of spore production across a combination of

Phlox species and PM isolates in vitro. Consistent with the literature, our results suggest that compared to most plant pathogenic fungi in the Ascomycota, Golovinomyces magnicellulatus, causal agent of PM, has larger genomes (about 130 Mb) with high repetitive content (about 40%) and a fewer number of protein-coding genes (about 8000).

We found a lack of population structure and genetic diversity, despite diverse phenotypic responses to Phlox germplasm screening. Interestingly, we identified 7 putative secreted proteins, which are predicted to be involved in the infection process, that are differently distributed between the G. magnicellulatus genomes analyzed. We hypothesize that variation in predicted secreted proteins is at the basis of the differences observed in genetic and phenotypic diversity. Our results also suggest the presence of qualitative and quantitative resistant traits in Phlox germplasm. This research explored the genome variation and evolutionary potential of Phlox PM pathogens to infer durability of host resistance, which are key tools to face plant health’s future challenges.

iii

Acknowledgments

First and foremost, I want to thank my mentor, Dr. Francesca Peduto Hand, whose guidance, trust, and hard work were key to this work and to achieving my goals.

Similarly, I am grateful for the generous expertise and guidance offered by my committee members. Dr. Pablo S. Jourdan, Dr. Thomas K. Mitchell, Dr. Pierce A. Paul, and Dr.

Jason C. Slot, thank you.

I cannot emphasize enough how crucial the help of Justin Morse and Mike Kelly, the greenhouse managers, was to maintain a low thrip populations and low PM disease levels on my Phlox plants collection.

I want to recognize all of my former and present lab members: the students, the technicians, and the post-docs who helped me collect the data and provided me with suggestions and guidance. Thank you, Maria Bellizzi, Isabel Emmanuel, Veena Devi

Ganeshan, Nathan Grifford, Shan Lin, Dana Martin, Cristian Olmos, Caterina Villari, and

Eric Warne. I also want to acknowledge the Ornamental Plant Germplasm Center staff who helped me maintain a micropropagated collection of Phlox species: Eric Renze and

Andy Schenkel.

I express my sincere gratitude to the Plant Pathology Department as a whole for the supportive and cohesive community that elevates one student to a well-rounded

iv professional. The department offers mentoring, opportunities, and true friendship essential to one’s development.

Last but certainly not least I thank my friends and family, and specifically Cécile

Rémignon Randon, my mother, and Kevin M. Simmt, my partner in life. Their support was essential to the accomplishment of this work. While my mother’s support started more than 29 years ago, it still feels new, infinite, and unconditional. Kevin’s wisdom, kind words, and patience carried me through the happiest, and most tempestuous times.

Thank you for building a world of love around my endeavors.

Vita

Education

Doctor of Philosophy The Ohio State University, 2015-2020 Major: Plant Pathology College of Food, Agriculture, and Environmental Sciences (CFAES)

Minor: Public Policy and The John Glenn College of Public Affairs 2018-2019 Management Master of Science The Ohio State University, 2015-2017 Plant Pathology CFAES Bachelor University Montpellier 2, 2012-2013 Functional Plant Biology Faculty of Science, France Associate degree LEGTA Garcia Lorca, France 2009-2011 Horticultural Sciences

Professional Experience

American Phytopathological Society - Public Policy Board, 2019- 2021 United States Early Career Intern The Ohio State University, Department of Plant Pathology, 2015-2020 Columbus, OH Graduate Research Associate (Ph.D.) AmericanHort, Columbus, OH 2019 HortScholars Coordinator Tree Resistance Advocacy Group (TRAG), United States 2018 - Member The Ohio State University, Ornamental Plant Germplasm Center 2014-2015 (OPGC), Columbus, OH Research assistant Experimental Station SERAIL, Department of Agriculture, France 2013 Research assistant vi

National Institute of Agronomy Research (INRA), France 2013 Intern Timbuk Farm, Granville, OH 2012 OSU exchange student

Publications

Peer-reviewed abstracts

Farinas, C, Jourdan, P, Paul, PA, Daughtrey, ML, Peduto Hand, F. 2019. Phlox species

have varied susceptibility to powdery mildew isolates from the Eastern United

States. Phytopathology. Vol. 109:S2.91

Farinas, C, Peduto Hand, F. 2018. Development of laboratory bioassays to study

powdery mildew pathogens of Phlox in vitro. Phytopathology. Vol. 108:S1.27.

Farinas, C, Peduto Hand, F. 2017. Genotypic characterization of isolates of

Golovinomyces magnicellulatus, the biotrophic powdery mildew pathogen of Phlox.

Phytopathology. Vol. 107:S5.122.

Peer-reviewed manuscripts

Farinas, C., Peduto Hand, F. First report of Golovinomyces spadiceus causing powdery

mildew on industrial hemp (Cannabis sativa L.) in Ohio. Submitted to Plant Dis. on

01/29/2020 (Under review).

Farinas, C., Jourdan, P., Paul, P., Slot, J., Daughtrey, M., Devi Ganeshan, V., Baysal-

Gurel, F., Peduto Hand, F. 2020. Phlox species show quantitative and qualitative

resistance to a population of powdery mildew isolates from the eastern United States.

Phytopathology. First Look https://doi.org/10.1094/PHYTO-12-19-0473-R

vii

Baysal-Gurel F., Farinas, C., Peduto Hand, F., Avin, F. 2020. First Report of Powdery

Mildew of Phlox Caused by Golovinomyces magnicellulatus in Tennessee. Plant Dis.

First Look https://apsjournals.apsnet.org/doi/10.1094/PDIS-11-19-2498-PDN

Bonello, P., Campbell, F., Cipollini, D., Conrad, A., Farinas, C., Gandhi, KJ.K., Hain,

F., Parry, D., Showalter, D., Villari, C Wallin, K. F. 2020. Invasive tree pests

devastate ecosystems–A proposed new response framework. Frontiers in Forests and

Global Change. 3, 2.

Farinas, C., Gluck-Thaler, E., Slot, J. C., & Hand, F. P. 2019. Whole-Genome Sequence

of the Phlox Powdery Mildew Pathogen Golovinomyces magnicellulatus Strain

FPH2017-1. Microbiology Resource Announcements, 8(36), e00852-19.

Emanuel, IB, Farinas, C, Lin, S, Pyerzynski, J., Crouch, JA, Peduto Hand, F. 2019.

“Occurrence of boxwood blight caused by Calonectria pseudonaviculata in Ohio

landscapes”. Plant Dis. 103(10): 2670.

Farinas, C., Jourdan, P., Paul, P.A., Peduto Hand, F. 2019. Development and evaluation

of laboratory bioassays to study powdery mildew pathogens of Phlox in vitro. Plant

Dis. 103(7):1536-1543

Farinas, C., Villari, C, Martin, D, Taylor, NJ, Peduto Hand, F. 2016. Magnaporthe

oryzae perennial ryegrass pathotype causes leaf spots and blight on Japanese forest

grass in Ohio. Plant Dis.101(3): 507

viii

Editor-reviewed articles

Farinas, C., Henderson, E., Peduto Hand, N. 2020. Evaluation of foliar fungicides in a

preventative or curative program to control powdery mildew on garden Phlox, 2019.

Plant Dis. Management Reports Vol. 14, OT009.

Farinas, C., Gifford, N., Peduto Hand, N. 2017. Evaluation of foliar fungicide

applications to control powdery mildew on garden phlox. Plant Dis. Management

Reports Vol. 11, OT017.

Fields of Study

Major Field: Plant Pathology

Table of Contents

Abstract ...... ii

Acknowledgments...... iv

Vita ...... vi

Table of Contents ...... x

List of Tables ...... xii

List of Figures ...... xiv

Chapter 1. Introduction ...... 1

1.1 Abstract ...... 1 1.2 The host Phlox ...... 2 1.4 Phlox and PM ...... 10 1.5 Evaluation trials for Phlox resistance to PM...... 12 1.6 Control of Phlox PM ...... 14 1.7 Conclusions ...... 18 1.8 Objectives and hypotheses ...... 19

Chapter 2. Development and Evaluation of Laboratory Bioassays to Study Powdery

Mildew Pathogens of Phlox in vitro1 ...... 34

2.1 Abstract ...... 34 2.2 Introduction ...... 35 2.3 Materials and Methods ...... 37

2.4 Results ...... 43 2.5 Discussion ...... 47

Chapter 3. Whole-Genome Sequence of the Phlox Powdery Mildew pathogen

Golovinomyces magnicellulatus strain FPH2017-1 ...... 63

3.1 Abstract ...... 63 3.2 Data availability ...... 66

Chapter 4: Phlox species show quantitative and qualitative resistance to a population of powdery mildew isolates from the eastern United States3 ...... 68

4.1 Abstract ...... 68 4.2 Introduction ...... 69 4.3 Materials and Methods ...... 73 4.4 Results ...... 78 4.5 Discussion ...... 81

Chapter 5. Conclusions and Future Directions ...... 97

References ...... 103

Appendix A. Supplementary Material for Chapter 2 ...... 116

Appendix B. Supplementary Material for Chapter 4 ...... 118

List of Tables

Table 1.1: Taxonomic classification of eastern U.S. Phlox species. The table shows recognized sections, subsections, species, and subspecies, along with photographs of the most cultivated species...... 21

Table 1.2: Reports of powdery mildew pathogens on Phlox hosts worldwide...... 24

Table 2.1: Summary statistics of the effect of type of medium on the percentage of leaf damage (chlorosis + necrosis) observed on detached Phlox paniculata leaves overtime 53

Table 2.2: Effect of the age of pathogen culture on the size of the fungal colonies and the number of spores per unit of area produced by Golovinomyces magnicellulatus and Podosphaera sp. on detached Phlox paniculata leaves...... 54

Table 2.3: Effect of the phenology of the host tissue and side of the leaf surface on the size of the fungal colonies and the number of spores per unit of area produced by Golovinomyces magnicellulatus and Podosphaera sp. on detached Phlox paniculata leaves...... 55

Table 2.4: Summary statistics of the effect of the powdery mildew (PM) species on the incidence and severity of powdery mildew infection observed on micropropagated Phlox glaberrima plantlets...... 56

Table 3.1: Summary statistics of assembly and annotation of G. magnicellulatus ...... 67

Table 4.1: List of powdery mildew isolates included in this study...... 86

Table 4.2: List of primers designed in this study and used in the MLST analysis ...... 88

Table 4.3: Phlox germplasm maintained at the Ohio Plant Germplasm Center (OPGC) that was used in this study...... 89

Table 4.4: Basic diversity indices calculated in the phylogenetic analysis conducted in this study...... 90

Table 4.5: Summary statistics of the effect of PM isolate (n=10) and Phlox species (n=4) on disease progression, spore production and latency period observed on Phlox plantlets...... 91

xii

Table A.1: Summary statistics of the effect of age of pathogen culture on the incidence of powdery mildew infection, the size of the fungal colonies, and the number of spores per unit of area produced by Golovinomyces magnicellulatus and Podosphaera sp. on detached Phlox paniculata leaves...... 116

Table A.2: Summary statistics of the effect of the phenology of the leaf tissue and side of the leaf surface on the incidence of powdery mildew infection, the size of the fungal colonies, and the number of spores per unit of area produced by G. magnicellulatus and Podosphaera sp. on detached P. paniculata leaves……………………………………..111

Table B.1: Gene sequences used in the MLST analysis and correspondent GenBank accession numbers...... 118

Table B.2: Summary statistics of the effect of PM isolate (n=10) and Phlox species (n=4) on disease progression, spore production and latency period observed on Phlox plantlets...... 120

xiii

List of Figures

Figure 1.1: Pictures of various Phlox cultivars and hybrids. A – P. paniculata ‘Charles Van Delft’. B – P. paniculata group ‘Candy Floss’. C – Tall Phlox hybrid ‘Opening Act Ultra Pink’. D – P. paniculata ‘Shortwood’...... 30

Figure 1.2: Golovinomyces magnicellulatus on Phlox paniculata ‘Starfire’ A- chasmothecia embedded in hyphal mat. B – Asci inside a chasmothecium...... 31

Figure 1.3: Conidia of Golovinomyces magnicellulatus on Phlox paniculata ‘Starfire’. 32

Figure 1.4: Detached leaf bioassay of Phlox paniculata ‘Starfire’ maintained on agar medium showing “green islands” symptoms...... 33

Figure 2.1: Example of different phenological stages of Phlox paniculata cv. Starfire leaves. Highlighted with rectangular boxes (from left to right) are 1st, 3rd, and 5th node leaves...... 57

Figure 2.2: Example of a detached Phlox paniculata cv. Starfire leaf maintained on ½ MS + Benzimidazole medium (medium 2). The number 1-8 represents the eight subsequent places on the leaf on which a single colony was pressed and from which eight new colonies developed ...... 58

Figure 2.3: AULDPC of the percentage of leaf damage (chlorosis+necrosis) recorded every 4 days for 20 days after plating on different media types ( P = 0.042). Media 1: ½ Murashige and Skoog (MS); Media 2: ½ MS + Benzimidazole; Media 3: ½ MS + Tetracycline; Media 4: ½ MS + Benzimidazole + Tetracycline. Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Columns followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05)...... 59

Figure 2.4: Percentage of PM incidence developed on 3rd node leaves of P. paniculata by using 14, 18 and 22-day old cultures of G. magnicellulatus (P = 0.07) and Podosphaera sp. (P = <0.001). Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Columns followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05). Statistical analysis performed on arcsine-transformed data. Untransformed data are displayed...... 60

Figure 2.5: Percentage of PM incidence developed on the abaxial and adaxial side of 1st, 3rd, and 5th node leaves of P. paniculata inoculated with G. magnicellulatus (P = xiv

<0.0001) and Podosphaera sp. (P = <0.0001). Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Columns followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05). Statistical analysis performed on arcsine-transformed data. Untransformed data are displayed...... 61

Figure 2.6: Incidence (P = 0.37) and severity (P = 0.115) of PM recorded on micropropagated plantlets of P. glaberrima inoculated with G. magnicellulatus and Podosphaera sp. Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Statistical analysis performed on arcsine-transformed data. Untransformed data are displayed...... 62

Figure 4.1: Cladogram representation of Maximum Likelihood phylogenetic trees generated from TEF-1α (A) and ITS (B) sequences using 31 Golovinomyces magnicellulatus isolates with 100 bootstrap support...... 92

Figure 4.2: Neighbor-net phylogenetic network computed with SplitTree using concatenated sequences of four genes (TEF-1α, ITS, IGS, TUB; Fit = 97.6%). Nodes in the phylogenetic network represent the sample haplotypes, and the edges represent the evolutionary relationship among isolates...... 93

Figure 4.3: Boxplots of the Area Under the Disease Progress Curve (AUDPC) assessed for 28 days post-inoculation. Box plots followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05). The horizontal line in each boxplot represents the median, the box’s lower and upper line represent the lower and the upper quartile at 0.25 and 0.75. The whiskers represent the minimum, and maximum values. Dots outside of the boxplots represent outliers, and rhombus dots inside the boxplot indicate mean values. Maximum AUDPC value for isolate MI1 (1184.3) is not shown in the chart...... 94

Figure 4.4: Scatter plot chart of days between inoculation and sporulation (i.e. latency period), expressed as days post inoculation (dpi), observed among PM isolates on each Phlox species. Each dot represents the mean of 12 data points. Large empty circles in each species indicate mean values...... 95

Figure 4.5: Linear regression showing the relationship between the amount of spores produced on all infected plantlets from all Phlox species (n=480) 28 days post- inoculation and the latency period (i.e. days between inoculation and sporulation; 4-24 dpi). P <0.001. Intercept: 12.8; slope: -0.17; r2 = 0.99). Rhombus dots correspond to mean values  SE...... 96

Chapter 1. Introduction

1.1 Abstract

Phlox, a diverse genus native to the North American continent, comprises mostly herbaceous and semi-woody perennial species, with the exception of a few annuals such as

P. drumondii. Phlox is particularly appreciated in naturalistic arrangements seen in botanical gardens, urban landscapes, and private yards, as it attracts pollinators. The main limitation to its cultivation is the fungal disease powdery mildew (PM). Hence, cultivars deployed through the years have been selected both for their horticultural traits and for resistance to PM. Despite the plant’s popularity, Phlox production seems to always be burdened by the disease. Though some studies on the Phlox-PM pathosystem have begun to unravel the mechanisms underpinning host-pathogen interactions, numerous questions remain unanswered. To add to the challenge, PM pathogens are obligate biotrophs, making the cultivation of the pathogen in vitro difficult. In addition, the already predominant pathogens are now predicted to be ubiquitous due to the exchange of plants throughout the globe. Here, we review the knowledge accumulated around Phlox and the PM disease, and we identify knowledge gaps that have yet to be addressed.

1.2 The host Phlox

History. The first illustrations of Phlox species date back to 1680 when John Banister, an English missionary-naturalist studying the flora and fauna of colonial Virginia, developed a plant catalog that included drawings of P. pilosa and P. subulata (Locklear,

2011). However, it is only after Linnaeus’ description “floris flammeao igneoque colore”

– flowers the color of glowing flame, found in the Genera Plantarum (1737), that the genus

Phlox was officially established (Locklear, 2011). The earliest records of cultivation of

Phlox species were of P. glaberrima (1725), P. carolina (1728), P. divaricata (1740), P. maculata (1740), P. ovata (1740), P. paniculata (1744), and P. pilosa (1759), while P. drummondii became a favorite annual in the 1830s (Locklear, 2011). Since the first Phlox cultivar was commercially released in 1824, many new cultivars have been developed from breeding programs. Although the earliest crosses arose from P. carolina and P. maculata, these hybrids were not performing well enough in gardens, hence the attention was turned to P. paniculata, which is more adaptable to constructed landscapes. American nurseries started offering P. paniculata selections imported from Europe in the mid-19th century

(Locklear, 2011). In 1917, 584 named selections of P. paniculata were reported by a survey of American nurseries (Coombs, 2017). Nowadays, only a few species are commercialized, including P. paniculata, or garden phlox, P. subulata, or moss phlox, and P. drummondii, annual phlox or Drummond's phlox (Zale, 2014) due to their horticultural abilities and performance in constructed landscapes.

Phlox is a popular inhabitant of gardens in the North American continent and the northern hemisphere worldwide and it is increasingly emphasized in native gardens for its 2 ability to attract pollinators. The genus accounts for over $12 million in U.S. national annual sales (NASS, 2014), and breeding and selection is an ongoing process. Examples of selected cultivars can be seen in Fig. 1.1.

Habitat and taxonomy. Phlox L. is an ornamental plant genus in the family

Polemoniaceae cultivated in temperate regions landscapes. Within the genus are approximately 65 species native to North America. They can be found in highly diverse habitats, including marshes, riverbanks, prairies, deserts, and tundra, from regions in northeastern Mexico and central Texas, to as far north as the Canadian province of Quebec.

Approximately 65% of the species can be found in the west and 35% in the east of North

America (Zale, 2014). Most of the species are perennials and only a few are annuals (e.g.

P. drummondii). Plant height can vary from 7 to 183 cm depending on species habit

(Armitage 2008, Coombs 2017; Locklear 2011; Zale and Jourdan 2015; Zale et al. 2016).

The diversity of the genus is exploited in the landscape, as some species, such as P. paniculata, P. carolina, and P. glaberrima, perform best with full sun exposure, whereas others, such as P. divaricata, and P. stolonifera, do so in the shade (Coombs, 2017).

To this day, the taxonomy of the genus remains challenging and incomplete (Zale and

Jourdan, 2015). Edgard T. Wherry’s 174-page monograph published in 1955, is the most comprehensive taxonomic study of Phlox and it is is still used today (Wherry, 1955). In

2014, Zale reorganized the taxonomic arrangement of eastern U.S. Phlox species (Table

1.1), which comprises most of the cultivated species. In this arrangement, three sections are observed: Annuae, which contain popular species like P. drumondii; Occidentales including P. subulata; and Phlox, including P. paniculata and P. carolina.

1.3 The disease Powdery Mildew

Along with mites, powdery mildew (PM) is the major biotic problem of cultivated

Phlox (Fuchs, 1994; Armitage, 2008; Zale and Jourdan, 2015). PM is a fungal disease that affects all vegetative organs of the plant, including leaves, stems and flowers and it is primarily recognized by the appearance of white powdery patches of fungal growth on the surface of infected tissues. Infected leaves remain on the plant until they turn yellow, then necrotic, and finally fall off (Fuchs, 1994). The disease is particularly seen on P. paniculata

- probably due to the dominance of the species in constructed landscapes - and some cultivars cannot be grown without fungicide treatment (Armitage, 2008). The plant’s susceptibility to PM is the most limiting factor to its widespread use in landscapes (Zale and Jourdan, 2015; Coombs, 2017).

Taxonomy. PMs are diseases caused by fungi in the Phylum Ascomycota, class

Leotiomycetes. They are obligate biotrophs, needing a living host to grow and reproduce, and cannot grow in axenic conditions. Many different genera of PM pathogens have been described. Some of these are able to infect a broad host range, whereas others are limited to a few hosts. In 1955, Schmitt artificially infected several ornamental hosts (e.g. Zinnia,

Phlox) using a collection of PM isolates that he called Erysiphe cichoracearum, which had been isolated from Zinnia, Phlox and a cucurbit plant. The PM isolate from Phlox did not infect any other hosts, indicating that the specimen was specific to Phlox. Historically, the genera reported to infect Phlox are numerous (Table 1.2). To date, the most reported genera causing PM on Phlox are Golovinomyces magnicellatus, for which Phlox paniculata is the holotype, and Podosphaera sp. (Bertetti et al. 2016; Farinas et al. 2017, 2019a and 2020;

Liu et al. 2020; Park et al. 2010). Both species can coinfect the same plant specimen

(Farinas, personal observation), making identification difficult.

The taxonomy of PM pathogens has largely been revised based on rDNA sequence evidence and spore morphological characteristics such as conidia produced in chain

(catenescent), presence/absence of fibrosin bodies, and appressorium morphology. For example, Cook and Braun (2009) have tried to redefine conidial germination types to match the diversity of known taxa. Among other findings, they suggested a new species separation for a Reticuloidium type of Golovinomyces with longitubus germ tubes (i.e. long, undifferentiated, negatively hydrotropic germ tubes). One sample of Podosphaera sp. on

P. paniculata was classified as Fibroidium subtype brevitubus syn. Magnicellulatae type

(fuliginea; Cook and Braun, 2009). Both Golovinomyces and Podosphaera subsection

Magnicellulatae infect herbaceous plants and have Asteraceae as the most common host family, but the Magnicellulatae phylogeny within the Asteraceae family is not as strict as for Golovinomyces spp. However, research looking into the phylogeny of the subsection

Magnicellulatae of the Podosphaera genus (Ito and Takamastu, 2010; Hirata et al. 2000) did not include any Phlox specimens. Identification of the Podosphaera sp. infecting Phlox will be the first step towards answering many evolutionary questions, for example: have

Podosphaera sp. and G. magnicellulatus emerged from convergent evolution?

Five major clades exist within the order Ersiphales. The genus Golovinomyces belongs to the tribe Golovinomyceteae, whereas the genus Podosphaera pertains to the tribe

Cystotheceae. Golovinomyces has external mycelium, no specialized hyphae, produces catenescent conidia without fibrosin body, and has multiple asci per chasmothecium, the

5 latter having myceliod appendages (Glawe, 2008). The genus Podosphaera also has external mycelium, no specialized hyphae, and catenescent conidia. However, the conidia possess fibrosin bodies, the chasmothecium produces only one ascum, and its appendages can be myceliod to dichotomously branched (Glawe, 2008).

Matsuda and Takamatsu (2003) used ITS and 28S rDNA to investigate the phylogenetic relationship of species within the genus Golovinomyces. They found G. magnicellulatus to be the pathogen infecting Phlox, which grouped with G. orontii isolated from Physalis. This subclade was grouped within a larger “LAC” group (formerly

Lactuceae tribe now subtribe Lactucinae of the Cichorieae tribe), where isolates from many host families (e.g. Lamiaceae, Solanaceae, Cucurbitaceae) were grouped with some isolates from the subtribe Lactucinae of the Asteraceae family. The authors also suggested that there is a possible co-speciation between Golovinomyces and the Asteraceae family and suggested that two phases in the evolutionary history of the Golovinomyces genus occurred. First, Golovinomyces would have expanded its host range in concert with the radiation of the tribes of the Asteraceae family. Then, the authors hypothesized that the pathogens switched hosts within the Asteraceae and also infected other plant families (host- jumping phase). Interestingly, the authors’ results suggest that this kind of host-jumping may have occurred within and around the LAC group but not in the other groups (Matsuda and Takamatsu, 2003). Surprisingly, in a later study based on rDNA ITS regions and the

28S rDNA sequences, Mastuda and Takamatsu (2013) identified a specimen on Physalis alkekengi as G. magnicellulatus, among other P. paniculata specimens, which grouped together in a lineage. However, the analysis comprised only a single sequence from

Physalis. The authors also suggested that Golovinomyces has expanded its host ranges at a relatively early evolutionary stage since all hosts of a single plant family, except for the family Polemoniaceae, belong to the clade Euasterid I (Matsuda and Takamatsu 2013).

Moreover, they suggested that after the expansion of its host range, the genus has not gone through further host jumping events. The need for evidence concerning the Polemoniacae family is required to verify this hypothesis. Takamatsu et al. (2006) used ITS and 28S rDNA sequence data of Golovinomyces spp. from hosts in the Asteraceae family to formulate the following hypothesis: the genus Golovinomyces arose from the Northern

Hemisphere after the Asteraceae family was introduced from South America.

To date, PM on Phlox has been reported in many countries in all five continents (Table

1.2). With the expansion of the horticultural trade and the lengthening of growing seasons due to climate change, it is likely that PM of Phlox will spread to environments where it has not been previously reported. Research into the genetic diversity of Phlox-PM populations will help identify the routes of disease spread and evolutionary risks of the pathogens.

With the increased accessibility of whole genome sequence data, knowledge about the genetics of PMs have augmented, revealing singular characteristics. Spanu et al. (2010) demonstrated the high presence of transposable elements (TE) in PM genomes, which causes the genomes size to be larger than most fungi (the average genome size of

Ascomycota is 36.91 Mb; Mohanta and Bae 2015) and with higher mutations rates, accounting for the high genetic diversity found among fungal pathogen species. To this date, the genome of two isolates of G. magnicellulatus from Phlox are publicly available.

One isolate was annotated by Farinas et al. (2019b). The size of the assembly was 129.9

Mb, which is within the range of other PM genomes (120-220 Mb), and contained 40% predicted repeat sequences. The authors’ annotation reported 8,172 predicted protein- coding genes, which is higher than other reported fungal genomes (6,000-7,000). Within the predicted protein-coding genes, they also identified 304 secreted proteins.

Disease cycle. Powdery mildews’ (PM) life cycles are associated with their host’s life cycle. Primary infections start from overwintering hyphae in dormant buds, or from a spore landing on the plant tissue. PM pathogens produce spores either asexually (i.e. conidia) or sexually (i.e. ascospores). Ascospores are protected in sac-like structure called asci (Fig.

1.2B), which are found within a melanized, thick-layered fruiting body known as chasmothecium (Fig. 1.2A). Chasmothecia overwinter in the soil or in infected plant debris, and ascopores are released once the increased turgor due to rain provokes a ‘chasm’ in the peridium (Glawe, 2008). Once the ascospore has infected the plant tissue, it produces hyphae on the surface of leaves, stems, and inflorescences. With time, the hyphae grow into an intricate net resembling a white mat (i.e. mycelium) that gives rise to conidiophores on which conidia are produced, either singly or in chains depending on the species. Conidia are produced throughout the growing season as long as conditions are favorable for disease development and are dispersed over long distances via wind and air currents. For example,

G. cichoracearum conidia infecting lettuce travelled over 200 km in California

(Schnathorst 1959). Hermansen et al. (1978), using known gene virulence markers, demonstrated that Blumeria graminis could travel up to 700 km using air currents from the

British Isles to Denmark. Conidia were also reported to be released by electrostatic charges

(Adams et al. 1986). Although possible, water-splash dispersal can be detrimental to conidia spread due to germination inhibition by water (Glawe, 2008).

Once a conidium lands on a plant tissue, it produces a germination tube, which swells into an appressorium. The penetration peg then penetrates the plant tissue through mechanical puncture due to turgor pressure and through enzymatic degradation. G. magnicellulatus may not prefer stomata openings for infection (Fig. 1.3: Farinas, personal observation). After penetration of the host cell, an haustorium is formed, which will invaginate the plant cell plasma membrane without penetrating it (Green et al. 2002), forming a haustorial complex comprised of an extrahaustorial membrane (EHM) and an extrahaustorial matrix (EHMAT). The EHMAT is where molecular signaling events occur to prevent plant recognition and favor nutrient absorption (Glawe, 2008). At the end of the season, chasmothecia are produced embedded in the mycelium, and will be released once mature. Certain types of appendages are thought to help dislodge chasmothecia from the hyphal mat, and travel through air (Glawe, 2008).

In 1962, Bushnell and Allen observed the presence of “green islands” on PM-infected plants, a symptom associated with areas around PM colonies that remain green while the rest of the plant tissue appears senescent. This phenomenon is regularly observed on detached Phlox leaves inoculated with PM (Fig. 1.4: Farinas, personal observation).

Bushnell (1967) advanced the hypothesis that the zone is the result of cytokinins emanating from the fungus or produced by the host in response to fungal invasion. Coghlan and

Walters (1990) also hypothesized that these regions represent areas of retarded senescence where biosynthetic activity is retained, in which polyamines play an important role. Further

9 research in the genes involved in this phenomenon will shed light on the biotrophic lifestyle of Phlox-PM, but also many other biotrophic pathogens.

1.4 Phlox and PM

Economic Importance. PM is ubiquitous in the environment and is the main concern of

Phlox growers (Coombs, 2017). If the infection is severe, plants may be unable to survive the winter (Coombs, 2017). Phlox cultivation heavily relies on the use of ecologically and economically expensive fungicides, which are applied as frequently as every 10 to 14 days throughout the growing season to protect the ornamental value of the plant (Armitage,

2008). Many researchers have tried to understand the relationship of Phlox and its PM pathogens, in order to reduce the use of chemical applications.

Host phenology. Jarosz et al. (1982) investigated the relationship between the thickness of the Phlox leaf cuticle and resistance to powdery mildew, whose causal agent was previously identified as Erysiphe cichoracearum. However, no relationship was found.

Moreover, the authors suggested that the pathogen might penetrate the host by enzymatic degradation rather than by mechanical puncture.

Farinas et al. (2019a) developed a detached leaf bioassay and a micropropagated plantlet bioassay to assess growth in vitro of PM pathogens (G. magnicellulatus and

Podosphaera sp.). Using the leaf bioassay, they found that 3rd node leaves supported more

PM growth than 1st or 5th node leaves, suggesting an ontogenic resistance mechanism in

Phlox. When testing the sides of the leaf, the authors found that the abaxial side supported more PM growth than the adaxial side in certain experiments but not others. They also

10 found that using 18-day-old PM colonies as inoculum generated more PM growth than using 14 and 22-days old colonies, suggesting that the age of the conidia is important in initiating infection. The findings of this research established a foundation for studying

Phlox-PM pathogens in vitro. Questions regarding why PM pathogens may infect preferentially the abaxial side of the leaf, or what are the molecular mechanisms associated with the host ontogenic resistance currently remain unanswered.

Phytohormones. Talieva and Kondrat’eva (2002) evaluated the influence of exogenous treatments with salicylic acid in mildew-susceptible and -resistant Phlox species on the level of phytohormone production associated with plant defense mechanisms. Exogenous applications of salicylic acid increased the content of cytokinins and endogenous salicylic acid. When infected with PM and exposed to exogenous salicylic acid, the level of abscisic and salicylic acids increased in the susceptible species (P. paniculata) whereas cytokinin increased in the resistant species (P. setacea L., syn. P. subulata ssp. setacea). The hormones are thought to be responsible for turning on defense mechanisms upon biotic stress. Salicylic acid is known to reinforce the cell membranes and therefore prevents acute modification in hormonal balance in reaction to stress. The authors argue that salicylic acid is key in the interactions of Phlox and PM (Talieva and Kondrat’eva, 2002).

A different study involving phytohormones tried to unravel their role in the development of conidia of powdery mildew on Phlox (Mishina et al., 2002). Cytokinins increased the percentage of conidia germination whereas abscisic acid decreased it.

However, cytokinins were also responsible for a disturbed differentiation of the appressorium, linking the hormone to host resistance. The study concluded that the effects

11 of phytohormones are ambiguous and relative to different pathogen species. Further research needs to evaluate the roles that salicylic acid plays in Phlox-PM infection, and how cytokinins might be important in Phlox resistance to PM.

1.5 Evaluation trials for Phlox resistance to PM

Resistance to PM is highly examined in Phlox and many field trials have been carried out to evaluate Phlox resistance to the disease. In 1964, Thompson and Svejda evaluated

49 cultivars for their resistance to powdery mildew in the field. Two cultivars out of the 49 tested showed resistance to powdery mildew during the first few years but became susceptible with time. The authors performed 350 crossings from resistant plants. Only 12 seeds germinated, among which, the potentially resistant Phlox plants became heavily infected after the third year of cultivation. The authors argued that susceptibility to the disease might increase with plant age. Lastly, the authors reported P. subulata to be free from infection. In 1981, Jarosz and Levy confirmed previously observed general resistance patterns (Jarosz et al., 1981) of P. carolina, and P. villosissima, and susceptible patterns of the P. drummondii lineage and P. pilosa ssp. detonsa.

The Chicago Botanic Garden ran a four-year evaluation trial (1993-1996) of Phlox species and hybrids selected based on their bloom quality, for resistance to powdery mildew (Hawke, 1999). According to the study, P. carolina ‘Reine du Jour’, Phlox

‘Chattahoochee’, P. paniculata ‘Katherine’ and P. pulchra ‘Morris Berd’ had the highest overall ratings. Whereas ‘Chattahoochee’, P. glaberrima ‘Don Hackenberry’, P. glaberrima ssp. triflora, P. pilosa, P. pulchra ‘Morris Berd’, and P. ‘Spring Delight’ were

12 noted as excellent for their resistance to powdery mildew but not for other horticultural characteristics. The cultivars of garden phlox P. paniculata ‘Bright Eyes’, ‘David’ and

‘Franz Schubert’, were considered promising because of their mildew resistance along with good ornamental qualities. The review stated that the difference in susceptibility and resistance of garden phlox to powdery mildew depends on its geographic location. The article also mentioned regional studies carried out at the University of Vermont, Cornell

University, North Carolina State University, and University of Arkansas, that provided recommendations for mildew-resistant P. paniculata such as ‘Delta Snow’, ‘Robert Poore’,

‘Blue Boy’, ‘David’, ‘Orange Perfection’, ‘Prime Minister’, ‘Red Magic’, Natascha’,

‘Speed Limit 45’, and ‘Eden’s Crush’. From 2001 to 2009, the Chicago Botanical Garden evaluated another 78 Phlox taxa, representing the commercially available phlox of 2000, in full-sun field trials (Hawke, 2011). They identified 27 Phlox with superior horticultural traits, spider mite resistance, and PM resistance. The highest score was given to P. paniculata ‘Shortwood’. Here again, the author noted that resistance seemed to be dependent on environmental conditions. Hence, Phlox species grown in locations possessing distinct environmental conditions are expected to perform differently.

Bir and Conner (2002) were also interested in evaluating garden phlox cultivars resistance to Erysiphe cichoracearum in a 3-year study carried out in North Carolina.

While they did not observe total immunity in any of the P. paniculata cultivars tested, they found a high degree of resistance in 'David', 'Delta Snow', 'Natascha', 'Robert Poore', 'Speed

Limit 45' and in the species P. carolina, which showed less than 11% damage. The Mt.

Cuba Center has evaluated 94 taxa, representing 8 Phlox species from 2015 to 2017

(Coombs, 2017) for horticultural traits, PM resistance, and capacity to attract butterflies.

The top performers with excellent PM resistance were P. amplifolia, P. carolina ‘Bill

Baker’, P. carolina ssp. carolina ‘Kim’, P. ‘Forever Pink’, P. glaberrima ‘Morris Berd’,

P. glaberrima ‘N. Tasahce rvfo hakof’, P. ‘Minnie Pearl’, P. paniculata ‘David’, P. paniculata ‘Delta Snow’, and P. paniculata ‘Jeana’.

Phlox paniculata ‘David’ was observed by several trials cited above as being PM resistant, and was chosen as the 2002 Perennial Plant of the Year (Perennial Plant

Association, 2002). Interestingly, Farinas et al. (2020) collected several isolates of Phlox-

PM from the cultivar David, indicating that the cultivar may have lost its resistance to the disease. However, when the authors tested several Phlox-PM isolates aggressiveness on four Phlox species, the isolates from ‘David’ were not the most aggressive. Scientists have reported conflicting results on the effects that cultivar resistance in the constructed landscape has on pathogen aggressiveness changes (Cowger and Brown 2019; Delmas et al. 2016). Further research on the impact of planting resistant cultivars in constructed landscapes on PM pathogen population will give insight into the durability of host resistance.

1.6 Control of Phlox PM

Host resistance. Understanding the genetic diversity of a pathogen and its evolutionary potential is a critical step in improving management of host resistance genes and fungicides and maximize their durability (McDonald and Linde, 2002). Powdery mildews have a large effective population size and produce long-distance dispersive clones (conidia; Milgroom,

2015), therefore they tend to have large genetic diversity and a high degree of genotype flow (movement of asexual propagules). McDonald and Linde (2002) hypothesized that genotype flow presents a higher risk of faster evolution because it represents a linked package of co-adapted alleles that have already been selected. To identify the evolutionary risks of Phlox-PM, more research is needed to characterize the diversity of powdery mildew pathogens of Phlox.

Of similar importance, characterizing the genetic diversity of Phlox germplasm can help find resistant traits to use in breeding programs. Farinas et al. (2020) aimed to characterize the genetic diversity of a population of G. magnicellulatus isolates from the eastern United States to better understand the pathogen’s capacity to adapt to new environments and new resistant hosts. Using multilocus sequence typing analysis and whole genome comparison, they found a lack of population structure and low genetic diversity within the PM population examined. They also screened 4 species of Phlox

(amoena, subulata, glaberrima, and paniculata) for their resistance to different isolates of

PM. The authors compared the disease severity progression and length of latency period of spore production and found evidence of quantitative and qualitative resistance in Phlox species. Phlox paniculata ‘Dunbar Creek’ was used as an internal control in this study as it had been previously described as resistant in the field (<25% infection; Hawke, 2011).

Indeed, ‘Dunbar Creek’ showed consistent resistance to all PM isolates even though the experimental conditions were highly conducive to disease development (Farinas et al.

2020). The authors found diverse phenotypic responses in the germplasm screening which were in contrast with the genetic diversity observed in the pathogen population, and which

15 could be explained by the diversity of predicted secreted proteins found across the PM genomes analyzed.

While exploring the genetic diversity and ploidy levels of Phlox germplasm, Zale and

Jourdan (2015) noted that triploid cultivars of P. paniculata, specifically ‘Robert Poore’ and ‘John Fanick’, were associated with resistance to PM in the literature. However, other resistant cultivars such as ‘Shortwood’, ‘David’, and ‘Delta Snow’, were found to be diploid. Although anecdotal, these observations might indicate that resistance to PM can have different pathways. The introduction of Phlox germplasm in breeding programs could increase genetic diversity and offer strategies for polyploids, while providing new sources of resistance. Phlox maculata, for example, has recently increased in popularity because it seemed to show some degree of resistance to the disease (Armitage, 2008).

As the genus genetic diversity has not been fully characterized yet, it will be beneficial to relate enhanced horticultural traits to PM resistance traits. Further investigations using quantitative trait locus (QTL) analyses will identify traits to incorporate into breeding programs. Moreover, interspecific hybridization and ploidy levels can give good insights into the potential of germplasm use for PM disease resistance.

Control. Fungicides are heavily used in the control of PM on Phlox. Armitage (2008) recommended spraying fungicide around June 15 and continuing applications every 10 to

14 days. Efficacy of fungicides are evaluated frequently (e.g. IR-4 Project) as new fungicides are developed and pathogens resistance evolves rapidly. For example, the efficacy of the fungicide groups benzimidazoles, hydroxypyrimidine, DMI, and strobilurin was overcome by Podosphaera xanthii, PM pathogen of cucurbit, between 1 and 3 years

(Hollomon and Wheeler, 2002). Sulfur, a protectant fungicide, has proven efficient at controlling Phlox-PM in the greenhouse (author’s personal observation), however, sulfur can interfere with insect predator activity, and become phytotoxic if the temperature rises above 27°C (Belanger and Labbe, 2002).

As an alternative to chemical sprays, Farinas et al. (2019a) used a 20% raw milk solution weekly as preventative and curative treatments in greenhouse-grown P. paniculata

‘Starfire’. When using Phlox plants that had been treated with milk for a long period of time (i.e. 10 months), less disease incidence and lower spore production were observed after the pathogen was inoculated in vitro (Farinas, unpublished), which can be explained by increased systemic resistance of the plant and/or increased production of reactive oxygen species (ROS). The mechanism by which milk sprays protect the plant from powdery mildew is poorly understood (Bettiol 1999; Kamel et al. 2017; Medeiros et al.

2012). However, it is believed that the microorganisms contained in raw milk activate the systemic resistance of the plant (Kamel et al. 2017; Medeiros et al., 2012; Stadnik and

Bettiol 2001; Sudisha et al. 2011). Another hypothesis is that milk treatments may generate the production of oxygen radicals in natural light (Crisp et al., 2006). Further experiments are needed to understand the mechanism of milk control. For example, the characterization of microorganisms found in raw milk, and their isolated effects on PM can inform us on their specific activity.

Another alternative would be the use of Silicone (Si) to control PM. Few studies have evaluated the efficacy of Si to control PM on Phlox. Frantz et al. (2010) supplied 30 to 60 ppm of potassium silicate (K2Si2O5) in fertilizer mixes to Phlox plants and found delays in

PM symptoms development by one to two weeks compared to the control. The authors suggested integrating the use of Si to management program to reduce the use of pesticides.

The mechanism by which Si controls PM is poorly understood. Using a transcriptomic approach, research on Arabidopsis and PM suggested that Si participates to the synchronized plant defense response rather than act as a simple mechanical barrier

(Fauteux et al. 2006). Additional research identifying the optimal dose, delivery method, and the mechanism of action of Si are needed to understand how to best incorporate it in

PM management programs.

Recently, PM management research has focused on the use of artificial UV-light exposure for disease control. UV-B and UV-C light were used on various crops such as cucumber (Patel et al. 2020; Suthaparan et al. 2014), strawberries and rosemary

(Suthaparan et al. 2016) in the greenhouse or the laboratory. Suthaparan et al. (2016) showed that UV-B reduced PM severity by 90% compared to the untreated control, with exposure of as low as 2 minutes once every 3 nights. To date, no research has been done on the use of UV light on Phlox crops.

1.7 Conclusions

Phlox-PM research is advancing, and further research concerning the identification of the specific pathogens infecting the crop, as well as their evolutionary potential are needed.

Indeed, a better understanding of the genetic diversity of PM pathogen populations is needed to achieve long-lasting host resistance. Furthermore, understanding the effectors and enzymes involved in the infection process and essential to biotrophic lifestyles will

18 facilitate investigations related to Phlox resistance. In parallel, research into the host germplasm genetic diversity using QTL analyses should be completed to identify durable resistant traits to incorporate into breeding programs and develop PM resistant polyploids and interspecific hybrids. Further research into non-chemical control options such as milk application and UV radiation will encourage growers to use alternative disease management methods.

1.8 Objectives and hypotheses

The long-term goal of this project was to contribute to a deeper understanding of the mechanisms of resistance to PM pathogens in Phlox, to the identification of phenotypic traits that can be used in breeding programs for obtaining disease resistant plants, and to a better understanding of PM pathogens’ capacity to adapt to new environments and new resistant hosts. The specific objectives of this research and underlining hypotheses were:

Objective 1: To generate reliable laboratory bioassays for conducting in vitro experiments with PM pathogens of Phlox.

H1: A reliable in vitro bioassays for PM pathogens of Phlox can be generated by

using detached Phlox leaves and/or micropropagated Phlox plantlets maintained on

artificial media.

H0: PM pathogens cannot be cultured in vitro on neither detached Phlox leaves nor

micropropagated Phlox plantlets maintained on artificial media.

Objective 2: To evaluate Phlox germplasm for its susceptibility to PM through in vitro

19 screenings using a collection of PM isolates.

H1: There are significant differences in the susceptibility of Phlox species to PM

isolates.

H2: There are significant differences in phenotypic responses of PM isolates to

Phlox species.

H0: Phlox species do not differ in their susceptibility to PM isolates, which do not

vary in their aggressiveness to Phlox species.

Objective 3: To infer the genetic diversity of a G. magnicellulatus population of isolates collected from the eastern United States.

H1: The G. magnicellulatus population infecting Phlox is genetically diverse.

H2: Its genetic diversity is expressed in terms of geographic isolation and/or host

virulence.

H0: The G. magnicellulatus population infecting Phlox is not genetically diverse,

regardless of origin of the isolates and/or virulence.

Table 1.1: Taxonomic classification of eastern U.S. Phlox species. The table shows recognized sections, subsections, species, and subspecies, along with photographs of the most cultivated species.

Eastern U.S.A. Phlox Taxaa,b Section Subsection Species Phlox amoena Sims Phlox drumondii Phlox divaricata L. ssp. divaricata Phlox divaricata L. ssp. laphamii (Wood) Wherry Phlox cuspidata Scheele Phlox drummondii Hooker ssp. drummondii Phlox drummondii var. peregrina Shinners Phlox drummondii ssp. glabriflora Brand, Pflanzer Phlox drummondii ssp. johnstonii (Wherry) Wherry)

21 Phlox drummondii ssp. mccallisteri (Whitehouse) Wherry Phlox drummondii ssp. tharpii (Whitehouse) Wherry Divaricatae Phlox floridana Bentham ssp. floridana Annuae (Wherry) Phlox floridana Bentham ssp. bella Wherry Prather Phlox pattersonii Prather Phlox pilosa L. ssp. pilosa Phlox pilosa L. ssp. deamii Levin Phlox pilosa L. ssp. fulgida (Wherry) Wherry Phlox pilosa L. ssp. longipilosa (Waterfall) Locklear Phlox pilosa L. ssp. ozarkana (Wherry) Wherry Phlox pilosa ssp. sangamonensis Levin Phlox pulcherrima (Lundell) Lundell Phlox roemeriana Scheele Phlox villosissima Turner Continued

Table 1.1 continued Section Subsection Species Subulatae Phlox bifida Beck ssp. bifida Phlox subulata L. (Peter) Phlox bifida Beck ssp. arkansana Marsh Wherry Phlox bifida Beck ssp. stellaria (Gray) Wherry Occidentales Phlox nivalis Loddiges ssp. nivalis Wherry Phlox subulata L. ssp. subulata Phlox subulata L. ssp. brittonii (Small) Wherry Phlox subulata L. ssp. setacea (L.) Locklear (syn. australis Wherry) Phlox Phlox carolina L. ssp. carolina Phlox carolina L. Ferguson Phlox carolina L. ssp. alta Wherry (syn. Phlox carolina L.ssp. angusta Wherry

22 Ovatae Phox glaberrima L. ssp. glaberrima

Wherry) Phlox glaberrima L. ssp. interior Wherry Phlox Phlox glaberrima L. ssp. triflora (Michaux) Wherry Phlox maculata L. Phlox ovata (L.) Locklear Phlox pulchra Wherry

Continued

Table 1.1 continued

Section Subsection Species Paniculatae Wherry Phlox amplifolia Britton Phlox paniculata L. Phlox paniculata L.

Stoloniferae Wherry Phlox stolonifera Sims

Clutanae Wherry Phlox buckleyi Wherry

a Reproduced and modified with permission from Zale (2014). b Taxa are derived mainly from Ferguson et al. (1999), Locklear (2011), Turner (1998), and Wherry (1955).

Table 1.2: Reports of powdery mildew pathogens on Phlox hosts worldwide.

PM species Reference Location Host

Erysiphe Goos 2010 Rhode Island Phlox sp. cichoracearum

Amano 1986 Yugoslavia P. decussata Shaw 1973 Washington P. diffusa Ginns 1986 Canada P. divaricata Gilman et al. 1929 Iowa P. divaricata Anonymous 1960 Indiana P. divaricata Anonymous 1960 New York P. divaricata Anonymous 1961 Ohio P. divaricata Anonymous 1962 Wisconsin P. divaricata Ginns 1986 Canada P. drummondii E. cichoracearum (Golovinomyces Alfieri et al. 1984 Florida P. drummondii cichoracearum) Gilman et al. 1929 Iowa P. drummondii Cho et al. 2004 Korea P. drummondii Amano 1986 USSR P. drummondii Anonymous 1960 Michigan P. maculata French 1989 California Phlox sp. Abbott et al. 1989 Canada Phlox sp. Firman 1972 Fiji Phlox sp. Alfieri et al. 1984 Florida Phlox sp. Gilman et al. 1929 Iowa Phlox sp.

Continued

Table1.2 Continued PM species Reference Location Host

Maneval 1937 Missouri Phlox sp. Grand et al. 1985 North Carolina Phlox sp. Shaw 1973 Oregon Phlox sp. Shaw 1974 Washington Phlox sp. Amano 1986 USSR P. × arendsii

E. communis (E. pisi Ginns 1986 Canada P. paniculata var. pisi) Ginns 1987 Canada Phlox sp.

Fakirova 1991 Bulgaria P. drummondii

Mulenko et al. 2008 Poland P. drummondii

Fakirova 1991 Bulgaria P. paniculata

E. magnicellulata (G. Ali et al. 2000 Germany P. paniculata magnicellulatus var. Dynowska et al. 1999; magnicellulatus) Poland P. paniculata Mulenko et al. 2008

Eriksson 1992 Sweden P. paniculata

Tobias and Russia Phlox sp. Tikhomirova 1998

Braun 1995 Yugoslavia P. decussata

Braun 1996 Bulgaria P. paniculata E. magnicellulata var. Braun 1997 Germany P. paniculata magnicellulata Braun 1998 Hungary P. paniculata

Braun 1999 Poland P. paniculata

Continued

Table 1.2 Continued PM species Reference Location Host

Braun 2000 Romania P. paniculata

Braun 2001 Switzerland P. paniculata

Braun 2002 USSR P. paniculata

Braun 2003 Yugoslavia P. paniculata

Braun 2004 Bulgaria Phlox sp.

Braun 2005 North America Phlox sp.

Braun 2006 United Kingdom Phlox sp.

Braun 2007 USSR Phlox × arendsii

G. cichoracearum Shin 2000 Korea P. subulata

Rusanov and Russia P. arendsii Bulgakov 2008

Girilovich et al. Belarus P. drummondii 2005

Voytyuk et al. 2006 Israel P. drummondii

Rusanov and Russia P. drummondii Bulgakov 2008

G. magnicellulatus Jones and Baker United Kingdom P. drummondii 2007

Bolay 2005 Switzerland P. maculata

Cook et al. 2006 Japan P. paniculata

Rusanov and Russia P. paniculata Bulgakov 2008

Eriksson 2014 Sweden P. paniculata

Dudka et al. 2004 Ukraine P. paniculata

Continued

Table 1.2 Continued PM species Reference Location Host

Kiss et al. 2001; United States P. paniculata Cook et al. 2006

Voytyuk et al. 2006 Israel Phlox sp.

P. paniculata, P. amoena, North America, diffusa, P. Asia, Europe, G. magnicellulatus divaricata, P. Braun and Cook 2012 Australia, Fiji, var. magnicellulatus drumondii, P. Egypt, Israel, maculata, P. Africa procumbens, P. subulata

Simmonds 1966 Australia P. drummondii

Amano 1986 India P. drummondii

Amano 1987 Iran P. drummondii

Amano 1988 Nigeria P. drummondii

Amano 1989 South Africa P. drummondii

Amano 1990 Sudan P. drummondii

Leveillula taurica Amano 1991 Tanzania P. drummondii

Amano 1992 USRR P. drummondii

Whiteside 1966 Zimbabwa P. drummondii

Braun 1995 Romania P. paniculata

Amano 1986 Iran P. pyramidalis

Braun 1995 Italy Phlox sp.

Braun and Cook 2012 NA Phlox sp.

Continued

Table 1.2 Continued PM species Reference Location Host

L. golovinii Braun and Cook 2012 NA P. paniculata

Simmonds 1966 Australia P. drummondii Oidiopsis taurica - Paul and Thakur 2006 India P. drummondii (Leveillula taurica) Paul and Thakur 2007 India Phlox sp. Gorter 1977 South Africa Phlox sp.

Amano 1986 Czechoslovakia P. paniculata Oidium drummondii Amano 1987 Poland P. paniculata (G. magnicellulatus Amano 1988 Romania P. paniculata var. magnicellulatus) Amano 1989 Switzerland P. paniculata Amano 1990 USSR P. paniculata

Sampson and Walker Australia P. drummondii 1982 Alfieri et al. 1984 Florida P. drummondii Deighton 1936 Sierra Leone P. drummondii Oidium sp. Whiteside 1966 Zimbabwe P. drummondii Cho and Shin 2004 Korea P. paniculata Sampson and Walker Australia Phlox sp. 1982

Podosphaera sp. Cook and Braun 2009 Germany P. paniculata

Continued

Table 1.2 Continued PM species Reference Location Host

Cunnington 2003 Australia Phlox sp.

P. gracilis, P. Braun and Cook P. collomiae North America longifolia, P. 2012 paniculata

North America, Braun and Cook South America, P. xanthii Phlox sp. 2012 Asia, Caucasus, Europe

Amano 1986 France P. acuminata Simonyan 1981 Armenia P. drummondii Sphaerotheca fuliginea (P. Amano 1986 USSR P. drummondii fuliginea) Amano 1987 England Phlox sp. Crous et al. 2000 South Africa Phlox sp.

Braun 1995 France P. acuminata Shin 2000 Korea P. drummondii S. fusca (P. fusca) Braun 1995 Hungary Phlox sp. Braun 1996 United Kingdom Phlox sp.

Gilbertson et al. Wyoming P. longifolia 1979 Anonymous 1960 Kansas P. paniculata S. humuli (P. Anonymous 1961 New Hampshire P. paniculata macularis) Anonymous 1962 New York P. paniculata Anonymous 1963 Ohio P. paniculata Anonymous 1964 Washington P. paniculata

S. macularis (P. Amano 1987 Australia Phlox sp. macularis) Amano 1987 Japan P. subulata

A B

C D

A B

Figure 1.2: Golovinomyces magnicellulatus on Phlox paniculata ‘Starfire’ A- chasmothecia embedded in hyphal mat. B – Asci inside a chasmothecium.

Figure 1.3: Conidia of Golovinomyces magnicellulatus on Phlox paniculata ‘Starfire’.

Figure 1.4: Detached leaf of Phlox paniculata ‘Starfire’ maintained on agar medium showing “green islands” symptoms.

Chapter 2. Development and Evaluation of Laboratory Bioassays to Study Powdery Mildew Pathogens of Phlox in vitro1

2.1 Abstract

The genus Phlox consists of approximately 65 species that include some of the most prevalent ornamental plants in the temperate zone. These popular ornamentals are extremely susceptible to powdery mildew (PM) caused by the biotrophic fungi

Golovinomyces magnicellulatus and Podosphaera sp. In this study, we used Phlox paniculata and P. glaberrima to develop a set of laboratory tools to study these pathogens in vitro, including a detached leaf and a micropropagated plantlet bioassay. We assessed pathogen growth under different experimental conditions, which included the use of four different media variations (½ MS medium amended with benzimidazole and tetracycline), three ages of pathogen culture (14, 18 and 22 days), three phenological stages of the host tissue (1st, 3rd and 5th node leaves), placement of inoculum on both leaf surfaces (abaxial and adaxial), and three different inoculation techniques (single spore transfer, colony tapping, colony brushing). Detached P. paniculata leaves were successfully maintained on benzimidazole-amended ½ MS medium for up to 3 weeks. The adaxial side of 3rd node leaves supported statistically significant more fungal growth compared to the adaxial side

1Copyright The American Phytopathological Society. Reproduced, by permission, from Farinas, C., Jourdan, P., Paul, P.A., Peduto Hand, F. 2019. Development and evaluation of laboratory bioassays to study powdery mildew pathogens of Phlox in vitro. Plant Dis. 103(7):1536-1543. 34 of 1st and 5th node leaves. Both pathogens also successfully infected micropropagated plantlets of P. glaberrima. These newly developed tools should facilitate in vitro studies on PM of Phlox and possibly be applicable to other ornamental species attacked by the same fungi.

2.2 Introduction

Phlox L. is a genus of ornamental plants in the family Polemoniaceae comprised of approximately 65 species that are characterized by colorful, long-lasting flowers displaying mostly monochromatic tones of white, blue, violet, and pink. The genus is native to North

America, inhabiting very diverse habitats, from marshes and riverbanks, to prairies, deserts, and tundra (Locklear 2011; Zale and Jourdan 2015; Zale et al. 2016). Almost all the species are herbaceous or semi-woody perennials, only a few are annuals. Plant height varies from several centimeters in species with creeping habit to over one meter in the upright herbaceous species (Coombs 2017; Locklear 2011; Zale and Jourdan 2015; Zale et al. 2016). The genus plays an important role in horticulture as it is widely used in gardens and landscapes, both in formal and naturalistic arrangements. Within the three billion dollar horticultural industry in the United States, Phlox accounts for over $12 million in annual national sales (wholesale and retail combined value; NASS, 2014).

A major constraint to the use of Phlox in landscapes, and a constant burden for nursery growers, is its susceptibility to powdery mildew (PM; Armitage 2008; Coombs

2017; Zale and Jourdan 2015). Under conditions of high relative humidity and warm temperatures, the disease drastically limits plant development and plantings may become

35 unsightly. Affected plants display powdery growth of the fungal colonies on the surface of flowers, leaves and stems that reduce the plant's photosynthetic ability (McGrath and

Shishkoff 2001). Infected leaves become chlorotic, then necrotic, and eventually fall off

(Hermann 1994). Severe infections can cause extensive leaf drop, leading to plant death

(Coombs 2017). Due to the high degree of susceptibility to the disease, Phlox cultivation heavily relies on the use of ecologically and economically expensive fungicides, which are applied as frequently as every 10 to 14 days throughout the growing season to protect the ornamental value of the plant (Armitage 2008).

Powdery mildew pathogens are biotrophic parasitic fungi encompassing different genera and species. These fungi infect susceptible hosts through dormant mycelium in buds or by wind-dispersed conidia, which are continuously produced in chains on the mycelium.

In late summer, chasmothecia are formed from which ascospores are released the following season by a special form of dehiscence known as vertical slits (Braun et al. 2002). Different

PM fungi are known to infect Phlox in the United States. A recent survey conducted in

Ohio showed that Podosphaera sp. and Golovinomyces magnicellulatus (U. Braun) V.P.

Heluta are predominant (Farinas and Peduto Hand 2017).

Since PM fungi are biotrophic organisms, studies involving these pathogens need to be conducted on living plants. However, in vitro bioassays have been developed for many

PM-crop systems (extensively reviewed by Nicot et al. 2002), including rose (Linde and

Debener 2003), rhododendron (Kenyon et al. 1995), barley (Surlan-Momirovic et al. 2016), grape (Peduto et al. 2013) and cucurbits (McGrath and Shishkoff 2001) among others.

These bioassays have facilitated the study of PM pathogens, their biology and

36 epidemiology, and the susceptibility of host populations to the disease they cause. Thus, the availability of these laboratory tools provides an opportunity to study these biotrophic organisms under controlled conditions.

Most of the published literature on Phlox PM has focused on the screening of different species and cultivars for resistance to the disease (Bir and Conner 2002; Hawke 1999;

Hawke 2011; Thompson and Svejda 1964). These studies have been carried out exclusively in field trials, typically relying on natural infection of the plant. In vitro bioassays would be helpful tools for carrying out such studies, as they would provide the benefit of controlled experimental conditions (including infection). However, at present, no record of a laboratory bioassay for the Phlox-PM system has been found in the literature. Therefore, the objectives of this study were to develop (i) a detached leaf bioassay, (ii) a micropropagated plantlet bioassay, and (iii) to optimize the experimental conditions for their use in studying PM pathogens of Phlox in vitro.

2.3 Materials and Methods

Plant material. The plant materials used in this study included both adult potted plants and micropropagated plantlets. Bare root adult plants of Phlox paniculata cv. Starfire were purchased from a commercial grower in May 2017 and February 2018 and transplanted into 3-gal nursery pots containing Metromix® 360 growing medium (Sun Gro

Horticulture, Bellevue, WA) amended with perlite at a ratio of 3:1. Eight grams of

Osmocote® 14-14-14 slow release fertilizer (ICL Specialty Fertilizers, Dublin, OH) were spread on the soil surface of each pot at transplant, and then plants were further fertilized

37 with 150 ppm of Peters® Professional 20-10-20 water-soluble fertilizer (ICL Specialty

Fertilizers, Dublin, OH) as needed. One set of plants was maintained in a greenhouse at a temperature between 20 and 25°C, and supplemental lighting (Plantmax Metal Halide Sky

Blue 600 watt) was used during the winter months to maintain a 14-hour photoperiod.

Another set of plants was maintained in a growth chamber (Conviron PGR-15) at a temperature of 24°C during the day and 22°C at night, with a 14-hour photoperiod. To prevent natural powdery mildew infection in the greenhouse, plants were sprayed weekly to run-off with a 20% raw (i.e. unpasteurized) milk solution using a hand-pump sprayer.

Every 2 months, plants were cut back to allow new vegetative growth.

Pathogen inoculum. Powdery mildew-infected Phlox paniculata plants collected from two botanical gardens, one in Ohio and one in Virginia, were used to produce single spore cultures for this study. Infected leaves were pressed against the surface of an empty sterile

Petri dish to dislodge conidia from the conidiophores. Under a dissecting microscope

(Leica S6D), single conidia were picked up from the dish using an ethanol-disinfected eyelash glued to a wooden stick, and transferred on disinfested P. paniculata leaves maintained inside 100 x 15 mm Petri dishes containing 25 mL of half-strength Murashige and Skoog (i.e. ½ MS) medium (Caisson Labs) and 2.5 mg/L of Gelrite (PlantMedia, bioWorld). Inoculated leaves were incubated on a laboratory shelf under white light

(Philips F54T5/841) with a 14-hour photoperiod at 24±1°C. After two weeks, and every two to three weeks thereafter, each newly developed sporulating colony was transferred onto a new disinfected leaf by cutting the leaf tissue around the colony with a sterile scalpel and then pressing the colony onto the new leaf. Isolate species identification was carried

38 out through rDNA ITS sequence analysis. To this extent, total genomic DNA was extracted from single spore cultures using a Chelex extraction method (Walsh et al. 1991) and a nested PCR was performed using the primer pairs ITS5/P3 and ITS5/ITS4 according to

Matsuda and Takamatsu (2003). Isolates of G. magnicellulatus and Podosphaera sp. were used in separate experiments, for all the experiments described below.

Detached leaf bioassay. In order to develop a detached leaf bioassay, leaves on the second and third nodes from the apical meristem were harvested from the potted plants and transferred to the laboratory where they were initially washed with a 0.067% dish soap solution (Dawn, Procter & Gamble) and rinsed three times with Milli-Q Type 1 Ultrapure

Water (Milli-Q, EMD Millipore). Leaves were then disinfected in a 2% sodium hypochlorite solution for 5 minutes and further rinsed three times with sterile ultrapure water. Following disinfection, leaves were blotted dry using sterile paper towels before being placed on the surface of agar media, inside a laminar flow hood.

Effects of different media on the health of detached leaves overtime. Four different types of media in 100 x 15 mm Petri dishes were tested for their ability to maintain the health of detached Phlox leaves in vitro overtime. One-half strength MS medium containing 2.5 mg/L of Gelrite was used as the base medium (medium 1) and was compared to three different variations: one containing 0.01% of the fungicide benzimidazole

(medium 2); one containing 0.01% of the antibiotic tetracycline (medium 3); and one containing the combination of both (medium 4). Each treatment was tested on six replicated plates, which were placed in stacks of four and arranged in a completely randomized design on a laboratory shelf under white light with a 14-hour photoperiod at 24±1°C. Every day,

39 plates in each stack were rotated, in order to be exposed to the same daily amount of light.

The severity of leaf chlorosis and necrosis (i.e. leaf damage), expressed as percentage of leaf area covered by the symptoms, was visually assessed on each leaf every four days for three weeks, after which the area under the leaf damage progress curve (AULDPC) was calculated (Madden et al. 2007). The experiment was conducted twice.

Development of powdery mildew on detached leaves. To define the optimum conditions for PM growth on a detached leaf bioassay, three factors were considered: age of the culture, phenology of the host tissue, and side of the leaf surface where inoculum is placed.

The first factor was tested in one experiment, and the second two factors were tested in combination in a second experiment, both carried out using medium 2 described above.

Each experiment was conducted at least twice, and each was composed of six replicated plates arranged in a completely randomized design and incubated on a laboratory shelf as previously described for 15 days.

The effect of age of pathogen culture was tested by comparing spores transferred from

22, 18, and 14-day old PM colonies onto the test leaves. The effect of the phenology of the host tissue was tested in a separate experiment by comparing the use of young leaves with obvious anthocyanin presence (i.e. 1st node leaves), middle-aged leaves transitioning to green color (i.e. 3rd node leaves) and older leaves completely green in color (i.e. 5th node leaves; Figure 2.1). In addition, the effect of leaf surface was tested by comparing the placement of inoculum on the abaxial or adaxial side of the leaf. The two treatment factors

(i.e. host phenology and leaf surface) were arranged in a complete factorial experiment consisting of one plate for each factor combination, replicated 6 times.

In all experiments, the pathogen was inoculated by excising a sporulating colony from an infected leaf and pressing the colony onto eight places on the surface of a disinfected healthy leaf (Figure 2.2).

Data collection. Treatment effect on PM development was assessed by calculating the incidence of PM infection, and by measuring the size of the developing colony and the spore production per unit of colony area over time. Because colonies were not circular, colony size was measured by recording the maximum length and width of each colony. To do so, colonies were observed under a dissecting microscope (Leica DM750) and pictures of each colony were taken with a microscope camera (Leica EC4) to subsequently record the aforementioned measurements using the software LAS EZ (Leica Microsystems). The two measurements of each colony were averaged to estimate the size of the colony.

Treatment effect on spore production of each growing colony was assessed by excising 4- mm leaf disks from each sporulating colony on each leaf using a cork borer and by placing them in individual microcentrifuge tubes containing 0.5 ml of a 0.1% Tween-20 water solution. Tubes were vortexed for 20 seconds, then centrifuged for two minutes at 15000 rpm, then vortex again for 5 seconds before spore counts. Spores were counted using a hematocytometer and two counts were averaged. Spore density was derived by dividing counts by colony area according to the following formula: average spore number × haemocytometer constant x volume of spore suspension/leaf disk area.

Micropropagated plantlet bioassay. Plantlets of P. glaberrima were micropropagated from a collection maintained at the Ornamental Plant Germplasm Center (OPGC) in

Columbus, OH. Stem cuttings were transferred to tubes containing 15 mL of MS medium

41 supplemented with 8.88x10-06 M of 6-Benzyl amino purine (BAP; Sigma-Aldrich) and 4.75 g/L of TC gel (Caisson Labs) to promote shoot production. The plantlets were grown inside an incubator (I-66VL Percival Scientific) under 14 hours light at 25°C for 2 months, then transferred to MS medium without cytokinin supplement to promote root growth and incubated under the same conditions as described above for three additional months.

Five-month-old micropropagated plantlets were inoculated with 18-day old cultures of

Podosphaera sp. and G. magnicellulatus in separate experiments. Plantlets were inoculated with an ethanol-disinfected artist fan paintbrush, which was previously thinned to reduce bristle density. Plantlets were removed from their tubes and a 9-mm colony from an 18- day old PM culture was brushed on the entire plantlet surface. One PM colony was used per plantlet. Then plantlets were returned to the corresponding tubes. Each experiment had

6 replications and was conducted twice. After inoculation, the plantlets were completely randomized on a laboratory shelf under 14 hours photoperiod at 24±1°C. Incidence of PM infection expressed as presence or absence of any PM growth on the entire plantlet, and disease severity expressed as percentage of plantlet area covered by PM growth, were visually assessed weekly for three weeks post-inoculation.

Data analysis. All statistical analyses were conducted using the R package agricolae v. 1.2-8 (De Mendiburu, 2009) and nlme v. 3.1-131 (Box et al. 1994; Davidian and Giltinan

1995; Laird and Ware 1982; Littler et al. 1996; Lindstrom and Bates 1990; Pinheiro and

Bates 1996; Pinheiro and Bates 2000; Venables and Ripley 2002) in RStudio v. 1.1.383.

In all experiments, treatment effects on the AULDPC of severity of leaf damage, disease incidence, disease severity, size of the colony and spore production per unit of area were

42 analyzed by ANOVA, and Tukey HSD test was used to separate means ( ≤ 0.05). All spore production data except for the experiment testing the age of culture using G. magnicellulatus, were log transformed to provide better homogeneity of variance prior to

ANOVA analysis, and then back-transformed to present results. All incidence data were arcsine transformed to provide better homogeneity of variance prior to ANOVA analysis.

The different runs of the experiments were treated as random effects and the treatments as fixed effects. Each experimental run was analyzed separately first, then the data were pooled together. Finally, a Chi-square Test of Independence was used to assess if the sequence of inoculation presses on each leaf (1-8; Figure 2.2) was related to a chance of a colony developing.

2.4 Results

Detached leaf bioassay: For clarity of reporting, results are presented by each individual factor that was tested in the development of the bioassay.

Effects of different media on the health of detached leaves overtime. There was a significant effect of media type on the health of detached leaves overtime in the first experimental run but not in the second run (Table 2.1). For clarity of reporting, data from the two experimental runs were pooled together. Medium 1 showed statistically lower leaf damage than Medium 3 and 4, on which the highest AULDPC was recorded (Figure 2.3).

Medium 1 and 2 were not significantly different from each other.

Effect of age of culture on the development of PM. For G. magnicellulatus, the Chi- square test of independence showed a non-significant chance of colonies developing from

43 each inoculation press [χ2 (7, N=576) = 13.576; P>0.05]. There was a significant effect of age of culture on PM incidence and colony size in the first run of the experiment but not in the second run (Supplemental Table A.1). No age of culture effect was observed on spore production and on colony size.

Data from the two experimental runs were combined for clarity of reporting. The incidence of PM infection did not differ significantly among the different ages of culture tested (Figure 4; P=0.07). High incidence values were recorded in all treatments (range 76-

90%; Figure 2.4). When using 18-day old cultures as inoculum, the colony size was significantly bigger than when using 22-day old cultures, but not significantly different than when using 14-day old cultures. The number of spores produced per unit area was not significantly different among treatments (Table 2.2).

For Podosphaera sp., the Chi-square test of independence also showed a non- significant chance of colonies developing from each inoculation press [χ2 (7, N=576) =

8.7254; P>0.05]. There was a significant effect of age of culture on PM incidence in both experimental runs and on spore production in the first run, but there was no statistically significant effect on colony size in either experimental run (Supplemental Table A.1).

Data from the two experimental runs were combined for clarity of reporting. Incidence of PM infection was significantly higher on leaves inoculated with 18 compared to 14 and

22-day old cultures (Figure 2.4; P=<0.001). The colony size was not significantly different among treatments. When using 18-day old cultures as inoculum, the number of spores produced per colony area was significantly higher than when using 22-day old cultures but not significantly different than when using 14-day old cultures (Table 2.2).

Effect of phenology of the host tissue and leaf surface on the development of PM. For

G. magnicellulatus, there was a significant effect of host phenology (i.e. leaf node) on PM incidence in both experimental runs (Supplemental Table A.2). However, the effect of the side of the leaf surface was only significant in the second run. When looking at the colony size, only the side of the leaf surface had a significant effect in the second run. When looking at the spore production per unit of area, the side of the leaf surface had a statistically significant effect in both runs, while the leaf node had statistically significant effect in the second run only. No significant treatment interactions were observed across datasets and experimental runs (Supplemental Table A.2).

The data were pooled together for clarity of results. Disease incidence on the adaxial side of 3rd node leaves was statistically higher than on 1st and 5th node leaves and was not statistically different than that observed on the abaxial side of all three node leaves (Figure

2.5). Disease incidence on the abaxial side of the leaf was not statistically different across leaf nodes (Figure 2.5). Within 1st node leaves, the abaxial side had statistically higher incidence than its adaxial side (Figure 2.5).

Concerning the size of the colonies, within either the abaxial or adaxial sides, the three node leaves were not statistically different from each other (Table 2.3). However, within leaf nodes, 1st node leaves showed statistical differences between leaf surfaces, with the abaxial having larger colonies than the adaxial side (Table 2.3). Third and 5th node leaves did not show statistical differences within side of the leaf surface (Table 2.3).

Concerning spore production, the abaxial sides of the three different node leaves were not statistically different (Table 2.3). However, colonies growing on the adaxial side of 3rd

45 node leaves produced significantly more spores than on 1st node leaves (Table 2.3). Lastly, on 3rd node leaves, more spores were produced on the abaxial side compared to the adaxial side (Table 2.3).

For Podosphaera sp., there was a significant effect of host phenology (i.e. leaf node) when looking at incidence, colony size, and spore production in the three experimental runs

(except the first run of the spore production data (P=0.06); Supplemental Table A.2). There was not a significant effect of side of the leaf surface when looking at incidence. There was a significant effect of side of the leaf surface in the first run of the colony size data, and in the three experimental runs of the spore production data. No significant treatment interaction was observed when looking at incidence except in the first run (Supplemental

Table A.2).

The results from three experimental runs were combined for clarity of reporting. When looking at disease incidence on the adaxial side of the leaf, 3rd node leaves had statistically higher incidence than 5th node leaves (Figure 2.5). Concerning the abaxial side, no treatment was statistically different (Figure 2.5).

Concerning the size of the colony, 3rd node leaves had statistically larger colonies than

5th node leaves when grown on the adaxial side, however no statistical difference was observed when growing on the abaxial side (Table 2.3). No statistical difference was observed between leaf surfaces within each leaf node (Table 2.3).

Concerning spore production, no statistical difference was observed among treatments, except the abaxial side of 1st node leaves, which significantly produced more spores than the adaxial side of 5th node leaves (Table 2.3).

Micropropagated plantlet bioassay. The species of PM pathogen used in the experiments did not have a significant effect on PM incidence and severity in either experimental run (Table 2.4). Both powdery mildew pathogens successfully infected micropropagated plantlets of P. glaberrima with disease incidence and severity values reaching 83% and 6%, respectively, in plants infected with G. magnicellulatus, and 67% and 15% respectively, in those infected with Podosphaera sp. (Figure 2.6).

2.5 Discussion

This study represents the first attempt to develop a set of laboratory tools to grow, maintain, and study PM pathogens of Phlox in vitro. The experiments carried out in this study identified the best conditions for two PM fungi (Podosphaera sp. and G. magnicellulatus) to grow in vitro on a detached leaf and a micropropagated plantlet bioassay and should facilitate future studies on host-microbe interactions, pathogen biology, and disease epidemiology. Additionally, our results could provide guidelines for working with other PM-ornamental systems as well.

The major limitation to Phlox production in nurseries and to its cultivation in landscapes is the plant's susceptibility to PM (Armitage 2008; Coombs 2017; Zale and

Jourdan 2015). In the development of our detached leaf bioassay, we encountered several challenges as we aimed to maintain clean plants in the greenhouse as our source of plant material. Sulfur-based fungicides have been widely used as protectants against powdery mildews since 1000 BC (Nicot et al. 2002). Previous reports of in vitro PM bioassays in model systems indicated that sulfur could be washed off the leaves before use (Nicot et al.

2002; Peduto et al. 2013; Willocquet et al. 1996). However, the effect of sulfur on Phlox was unknown. In our studies, preliminary experiments showed that while vaporized sulfur applications were very efficient at protecting Phlox plants from PM infection in the greenhouse, they would not allow the pathogen to grow in vitro on leaves detached from those plants even after several water rinses and then undergoing disinfection (data not shown). In search for an alternative method to keep plants clean, we explored the use of a

20% raw milk solution, which we sprayed weekly on the entire plant surface. The mechanism by which milk sprays protect the plant from powdery mildew is poorly understood (Bettiol 1999; Kamel et al. 2017; Medeiros et al. 2012). While this study did not investigate such mechanism, the use of raw milk proved successful at controlling PM in the greenhouse and did not affect the ability of the pathogen to grow in vitro following disinfection of the leaves. Further investigations are needed to explore the mechanism by which raw milk protects Phlox from PM infections.

Inoculation techniques tested in this study encompassed the use of a hand-made tool made of an eyelash glued to a wooden stick; pressing an infected leaf onto a new disinfected leaf; and brushing inoculum using an artist paintbrush. The first method, described by

Peries in 1962 (Nicot et al. 2002), allows for picking single conidia and depositing them in a precise location onto a clean surface, which facilitates generation of single-spore colonies. The technique was described as being more successful than the use of fine needles, due to greater suppleness of the eyelash and possibly to higher static electricity, and was used successfully in this study to generate pure cultures of the two PM isolates.

Pressing a leaf segment bearing one PM culture up to eight different places on a clean leaf

48 did not result in different colony incidence and allowed quick and easy generation of multiple colonies from little inoculum. Finally, an artist paintbrush, described numerous times in the literature (Nicot et al. 2002; Leus et al. 2006; Kenyon et al. 1995) was used to successfully and uniformly inoculate entire micropropagated plantlets. To limit the loss of inoculum, we thinned the number of bristles on each paintbrush and observed under a dissecting microscope that less spores were caught up after plantlet inoculation (data not shown).

When using a leaf-based bioassay, the longer the leaves are maintained in good conditions, the less frequent the transfer of PM on new healthy leaves is needed.

Benzimidazole has been used in numerous leaf bioassay studies to protect cultures from contaminations, and to delay tissue senescence (Arabi and Jahwar 2002; Azmat et al. 2013;

Nicot et al. 2002; Linde and Debener 2003; Surlan-Momirovic et al. 2016). Use of tetracycline has been reported by Bardin et al. (1999) to reduce occurrence of contaminants as well. However, when comparing different media for their ability to maintain leaf quality over time, in this study we did not observe a significant effect of benzimidazole nor tetracycline in the medium when used individually compared to non-amended medium.

However, when used in combination, leaf damage was significantly higher compared to non-amended medium.

To define the best conditions for PM growth in a detached leaf bioassay, we considered the effect of different factors: age of the culture used to inoculate leaves, phenology of the host tissue, and side of the leaf surface onto which inoculum is placed. Regardless of the

PM fungus tested (G. magnicellulatus or Podosphaera sp.), the use of 18-day old cultures

49 resulted in the generation of a higher number of colonies that were larger and on which more spores were produced, compared to the use of 14 or 22-day old cultures. PM conidia are produced in chains on the mycelium. Nicot et al. (2002) described three stages of conidial formation. Among these, mature conidia, which are connected to the chain by a papillum on the rounded end, are only observed in the last stage. In his studies on PM of lettuce, Schnathorst (1959) showed that in G. chicoracearum only the last spore (oldest) is able to infect the host, even if all spores in the chain are able to germinate, suggesting that age of the spores is an important factor in conidial infection. These observations are in accordance with the results of this study in which younger cultures (14-day old) did not infect the host as much as 18-day old cultures. Furthermore, Nicot et al. (2002) mentioned that conidial production ceases with senescence of the host. In our media experiment, a high percentage of leaf necrosis and chlorosis were observed on detached leaves 20 days post plating (data not shown). It is likely that as leaf health degraded, inoculum also started to senesce. On the basis of these observations, 18-day old cultures were subsequently used to test the additional experimental factors.

Phlox paniculata cv. Starfire shows anthocyanin pigment on young leaves, which then disappears as the leaves mature. This factor, as well as the side of the leaf on which the pathogen spores land, might influence PM growth. Ontogenic or age-related resistance of the host has been reported in different PM systems, for example strawberry (Asalf et al.

2014; Sombardier et al. 2009) and grape (Gadoury et al. 2003; Ficke et al. 2002). In strawberries, Asalf et al. (2014) have found that relative disease severity declined with increasing leaf age. Moreover, Sombardier et al. (2009) reported a 10 times higher infection

50 efficiency on the abaxial side of strawberry leaves, compared to the adaxial side. In the grape PM system, host susceptibility has been found to be centered around bloom (Gadoury and Seem 1995; Gadoury et al. 1997; Gadoury et al. 2000), with 4-week-old berries being nearly immune to the disease (Ficke et al. 2002). Ficke et al. (2002) also suggest that resistance in grape berries is unlikely linked to cuticle or cell wall thickness. In Phlox,

Jarosz et al. (1982) did not find a correlation between leaf cuticle thickness and resistance to powdery mildew when comparing different Phlox species with related differences in cuticle thickness, suggesting that it is unlikely that a difference in cuticle thickness between the adaxial and abaxial side of the leaf plays a role in appressorium penetration. Even though PM incidence on the abaxial side of the leaf was higher in some of our experiments, after two weeks of culturing, we observed that the leaf tended to curl up, which makes accessibility to the inoculum and consequent measuring or rating very difficult. Moreover, the side of the leaf surface on which inoculum was placed had an inconsistent effect across the three experimental runs. Consequently, depending on the test’s purpose and type of data assessed, it might be more practical to use the adaxial side of the leaf for further tests on PM of Phlox. When using the adaxial side of the leaf, 3rd node leaves would significantly sustain more powdery mildew growth than 5th node leaves, indicating a possible ontogenic resistance mechanism in Phlox. Further investigations are needed to explore the difference in PM infection between the two sides of the leaf.

In this study we also explored the use of micropropagated Phlox plantlets to carry out in vitro PM studies. Based on our results, the use of micropropagated plantlets could provide several benefits. First, micropropagated plantlets are more likely to be PM-free

51 since grown in vitro, which would result in the avoidance of laborious fungicide sprays and possible contaminations. Second, Phlox is a broad genus that encompasses species with very diverse leaf phenology. The use of plantlets over detached leaves would allow standardizing the bioassay when using different Phlox species. Lastly, using plantlets over detached leaves allows evaluating plant response from an entire plant, compared to a response from an injured, detached leaf. However, micropropagated plantlets are slow to grow (approximately 5 months), and propagation is time consuming. Moreover, the plants are grown in microenvironments that could affect the leaf tissue and the interaction with the pathogen.

In our experiments we successfully infected micropropagated Phlox glaberrima plantlets. While both inocula resulted in high disease incidence, low severity was observed.

This could be due to the length of time left for the pathogens to develop (3 weeks), and the amount of inoculum used to inoculate the plantlets (one 9 mm diameter colony per plantlet). This method should nevertheless prove easier for further studies on interaction between the host and its pathogen.

In conclusion, in this study we successfully developed two laboratory bioassays and optimized the experimental conditions to use these tools to study PM pathogens of Phlox in vitro. Further studies focused on screening perennial Phlox germplasm for susceptibility to PM through these bioassays are underway and should provide additional information on the practical applications of these tools. Moreover, the bioassays developed here could serve as a model for other PM-ornamental crop systems.

Table 2.1: Summary statistics of the effect of type of medium on the percentage of leaf damage (chlorosis + necrosis) observed on detached Phlox paniculata leaves overtime.

AULDPCa Factor tested Experimental Run F P value

Type of mediumb 1 5.249 0.003

2 1.2 0.321 aArea under the leaf damage progress curve bFour different media were tested: ½ MS; ½ MS + Benzimidazole; ½ MS + Tetracycline; ½ MS + Benzimidazole + Tetracycline.

Age of pathogen culture (days) P Pathogen Parameter value 14 18 22

Mean 5.45 ab 6.69 a 4.79 b Colony size 0.008 (mm) SEx 0.63 0.67 0.47 Golovinomyces magnicellulatus Mean 837 1179 782 Spore/mm2 0.086 SE 164 228 165

Mean 2.87 3.25 2.3 Colony size 0.06 (mm) SEx ± 0.46 ± 0.37 ± 0.28

y Podosphaera sp. Mean 83.1 ab 81.5 a 13.5 b

Spore/mm2 0.021 95% 0, 213 0, 142 5, 53 CLz xSE = Standard Error of the mean yANOVA analysis performed on log transformed data. Back-transformed data reported here. zCL = Confidence Limit

st rd th 1 node 3 node 5 node Pathogen Parameter P Abaxial Adaxial Abaxial Adaxial Abaxial Adaxial value

Mean 4.51 a 1.43 b 4.46 a 3.35 ab 3.76 ab 4.36 ab Colony size 0.009 (mm) SEx ±0.58 ±0.24 ±0.4 ±0.56 ±0.33 ±0.46 Golovinomyces magnicellulatus Meany 169 ab 0.66 c 523.2 a 30.57 b 399.4 ab 188.67 ab

Spore/mm2 <0.001 55 z 31.9, 282.9, 5.19, 227.5, 69.9,

95% CL 0.29, 1.5 896.6 967.5 180 701.1 508.9

Colony size Mean 6.19 ab 4.13 bc 6.5 a 5.1 ab 4.6 abc 2.1 c <0.001 (mm) SEx ±0.55 ±0.76 ±0.67 ±0.91 ±0.41 ±0.43 Podosphaera sp. Meany 121.5 a 36.6 ab 66.7 ab 36.6 ab 73.7 ab 6.05 b Spore/mm2 <0.001 38.5, 14.2, 23.4, 15.1, 20.6, 95% CLz 1.5, 24.2 382.7 94.2 189.9 88.5 263.3

wValues with the same letter within each row are not significantly different according to Tukey HSD test (α=0.05) xSE = Standard Error of the mean yANOVA analysis performed on log transformed data. Back-transformed data reported here. zCL = Confidence Limit

Table 2.4: Summary statistics of the effect of the powdery mildew (PM) species on the incidence and severity of powdery mildew infection observed on micropropagated Phlox glaberrima plantlets.

PM Incidenceb PM Severity Factor tested Experimental Run F P value F P value

PM speciesa 1 0.38 0.55 0.96 0.35

2 0.38 0.55 1.62 0.23 aTwo PM species were tested: Golovinomyces magnicellulatus and Podosphaera sp. bANOVA analysis performed on Arcsine-transformed data.

Figure 2.1: Example of different phenological stages of Phlox paniculata cv. Starfire leaves. Highlighted with rectangular boxes (from left to right) are 1st, 3rd, and 5th node leaves.

Figure 2.3: AULDPC of the percentage of leaf damage (chlorosis+necrosis) recorded every 4 days for 20 days after plating on different media types (P = 0.042). Media 1: ½ Murashige and Skoog (MS); Media 2: ½ MS + Benzimidazole; Media 3: ½ MS + Tetracycline; Media 4: ½ MS + Benzimidazole + Tetracycline. Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Columns followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05)

Figure 2.5: Percentage of PM incidence developed on the abaxial and adaxial side of 1st, 3rd, and 5th node leaves of P. paniculata inoculated with G. magnicellulatus (P = <0.0001) and Podosphaera sp. (P = <0.0001). Columns represent the mean value of two runs of the experiment. Bars indicate ± SE. Columns followed by the same letter are not statistically different according to Tukey’s HSD test (α=0.05). Statistical analysis performed on arcsine-transformed data. Untransformed data are displayed.

Chapter 3. Whole-Genome Sequence of the Phlox Powdery Mildew pathogen Golovinomyces magnicellulatus strain FPH2017-12

3.1 Abstract

Powdery mildew (PM) fungi are obligate biotrophs capable of infecting diverse plant hosts, ranging from monocotyledonous agricultural crops to dicotyledonous ornamental crops. The PM lifestyle poses significant challenges for studying these pathogens in isolation from their host. Here, we present the genome of Golovinomyces magnicellulatus, a host-specific PM on Phlox.

Golovinomyces magnicellulatus (Leotiomycetes, Ascomycota) is an obligate, host- specific fungal biotroph that causes powdery mildew (PM) disease on ornamental plants in the Phlox genus (Matsuda and Takamatsu, 2003). Due to difficulties in growing PM fungi under axenic conditions, little is known regarding the genetic and evolutionary bases of their lifestyles, presenting an opportunity to gain insight through a genome-focused approach.

G. magnicellulatus strain FPH2017-1 was isolated from Phlox paniculata in Leipsic,

Ohio, USA. A single spore was isolated on a detached leaf bioassay (Farinas et al. 2019) and grown on P. paniculata ‘Starfire’ plants in a growth chamber. Spores were harvested periodically over one month by rinsing infected leaves with 0.1% Tween solution and then

2Copyright © 2019 Farinas et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license. Reproduced, by permission, from Farinas, C., Gluck-Thaler, E., Slot, J. C., & Hand, F. P. 2019. Whole-Genome Sequence of the Phlox Powdery Mildew Pathogen Golovinomyces magnicellulatus Strain FPH2017-1. Microbiology Resource Announcements, 8(36), e00852- 19. 63

filtering with Miracloth and stabilizing using 10 mM Tris Buffer (pH=7). The solution was centrifuged, and the resulting pellet was immersed in liquid nitrogen and kept at -80°C.

DNA was extracted from the pellet using the DNeasy® Plant Mini Kit (Qiagen).

DNA libraries were prepared using the NEBNext® Ultra™ II DNA Library Prep Kit and sequenced using Illumina PE300 MiSeq. Un-sheared DNA extracts were prepared using a Ligation Sequencing kit (#SQK-LSK109) and sequenced using MinION (Oxford

Nanopore Technologies).

Illumina sequencing generated 17,742,739 reads (35-300bp in length). Reads were trimmed using Trimmomatic v.0.36 (Bolger et al. 2014) with options “ILLUMINACLIP,

TruSeq3-PE.fa:2:30:10, CROP:290, SLIDINGWINDOW:10:25, HEADCROP:10,

MINLEN:100” (Andrews 2010)). Nanopore sequencing generated 427,831 reads (46-

42,472bp in length). Reads were quality filtered using Albacore v.2.3.1 (Pomerantz et al.

2018). Iterative BLASTn searches against the NCBI nucleotide database (last accessed

2/11/19) in conjunction with BBSplit (Bushnell 2014) were used to identify and remove contaminant reads of non-fungal origin. To do so, we inserted the taxonomic lineage feature in BLAST to all hits to identify all non-fungal contaminants, and only selected

Ascomycota and Basidiomycota contigs. Subsequently, using only long blasts coverage

(more than 2000 nt in length) to prevent false positives, we created a list of all contaminants found. This allowed us to download identified genomes and use them to filter our reads using BBSplit. We then performed de novo hybrid genome assembly using SPAdes v.3.12.0 (Antipov et al 2016) and identified known and de novo repeat elements using

RepeatModeler v.1.0.11 (Bao et al 2015).

We annotated the assembly using three iterations of a MAKER v.2.31.9 (Holt and

Yandell 2011) pipeline. In the first iteration, we provided MAKER with RNA-seq data of

Golovinomyces cichoracearum (http://genome.jgi.doe.gov/Golci1, last accessed: 2/12/19) and 10 protein datasets from other Leotiomycetes species (Blumeria graminis f. sp. hordei

DH14, B. graminis f. sp. hordei Race1, B. graminis f. sp. tritici 96224, Erysiphe necator,

G. cichoracearum, Amorphotheca resinae, Meliniomyces variabilis, Sclerotinia sclerotiorum, Rhizoscyphus ericae, Botrytis cinerea). For the second iteration, we provided

MAKER with additional evidence from the ab initio gene predictors SNAP v2013-02-16

(Korf 2004) (trained using high quality predictions from round 1) and Augustus v3.3

(Stanke et al 2008) (trained using BUSCO v3.0.1 (Watehouse et al 2018)). For the final iteration, we provided MAKER with updated evidence from SNAP and Augustus (both re- trained using high quality predictions from round 2) and set the option “keep_preds” to 1.

Another filtering step was performed to remove contaminant proteins. We blasted our predicted proteins and removed all proteins with more than 50% coverage and 75% identity with non-fungal hits.

Many PM genomes are estimated to be large, ranging from 120-220 Mb (Wu et al

2018), due in part to high repeat content. Conversely, PMs generally possess fewer protein- coding genes (6000-7000) compared to other fungi (Sonah et al 2016; Wu et al 2018). Our genome falls within the reported PM genome size while our annotation process recovered more protein-coding genes than generally reported (Table 3.1), which we attribute to the multiple lines of ab initio evidence used in the annotation process.

3.2 Data availability

This Whole Genome Shotgun project has been deposited at DDBJ/ENA/GenBank under the accession VCMJ00000000 and PRJNA540711 (SRA database).

Table 3.1: Summary statistics of assembly and annotation of G. magnicellulatus

Parameter Value Assembly Genome size (Mb) 129.9 Avg coverage (X) (no. of Illumina reads)a 46 (97) Avg coverage (X) (no. of Nanopore reads)a 4 (60) No. of contigs 84,604 N50 (bp) 4,118 Longest scaffolds (kbp) 197 GC content (%) 44 BUSCOb (% recovered) 88.2

Annotation Total no. of protein-coding genes 8,172 Avg gene length (bp) 1,764 No. of coding sequencesc 8 No. of repeat sequencesc 40 No. of proteins with at least one Pfam domain 6,396 No. of secreted proteinse 304 a Percentage of reference bases covered, estimated using BBMap v.37.93 (Bushnell 2014). b Sordariomyceta data set. c Percent assembly size. d Identified using InterProScan v.5.25-64 (Jones et al 2014). e Identified using SignalP-5.0 (Armenteros et al 2019) and TMHMM v.2.0 (http://www.cbs.dtu.dk/services/TMHMM/) to exclude transmembrane proteins.

Chapter 4: Phlox species show quantitative and qualitative resistance to a population of powdery mildew isolates from the eastern United States3

4.1 Abstract

Ornamental plants in the genus Phlox are extensively planted in landscapes and home gardens around the world. A major limitation to a more widespread use of these plants is their susceptibility to powdery mildew (PM). In this study, we used multilocus sequence typing (MLST) analysis to gain insights into the population diversity of 32 Phlox PM pathogen (Golovinomyces magnicellulatus and Podosphaera sp.) isolates collected from the eastern U.S. and relate it to the ability to overcome host resistance. Low genetic diversity, and a lack of structure, were found within our population. Whole genome comparison of two isolates was used to support low genetic diversity evidence found with the MLST analysis. Recombination was suggested by the incongruences observed in the six phylogenetic trees generated from the housekeeping genes TEF-1, CSI, ITS, IGS, H3, and TUB. Contrasting with low genetic diversity, we found high phenotypic diversity when using ten of the 32 isolates to evaluate host resistance in four different Phlox species (P. paniculata ‘Dunbar Creek’, P. amoena OPGC 3598, P. glaberrima OPGC 3594,

3Copyright The American Phytopathological Society. Reproduced, by permission, from Farinas, C., Jourdan, P., Paul, P., Slot, J., Daughtrey, M., Devi Ganeshan, V., Baysal-Gurel, F., Peduto Hand, F. Phlox species show quantitative and qualitative resistance to a population of powdery mildew isolates from the eastern United States. Phytopathology. First Look https://doi.org/10.1094/PHYTO-12-19-0473-R

and P. subulata OPGC 4185) using in vitro bioassays. We observed quantitative and qualitative resistance in all Phlox species and a consistent low disease severity in our control Phlox paniculata ‘Dunbar Creek’. Taken together, the results generated in this study constitute a robust screening of popular Phlox germplasm, that can be incorporated into breeding programs for PM resistance, and it also provides significant information on the evolution of PM pathogens.

4.2 Introduction

Durable disease resistance is the ultimate goal of plant pathology. It is the most environmentally and economically friendly method to manage diseases as it leads to reduced chemical use. “Durable resistance” comprises the plant’s capability to fend off a pathogen’s infection, and also the failure of the pathogen to overcome plant defense mechanisms, both over the long-term (McDonald and Linde, 2002). There are different types of resistance. Qualitative, also known as gene-for-gene resistance, can provide complete immunity and is usually controlled by a dominant gene. Quantitative resistance is controlled by multiple genes, each contributing a relatively small effect, and it is thought to be more durable than qualitative resistance (Cowger and Brown 2019; Mundt 2014).

Several resistance deployment strategies for achieving durable resistance have been described, including gene rotation and gene pyramiding, in which several identified resistance genes are used either in replacement or simultaneously, to prevent the pathogen from losing avirulence genes (Mundt 2014). Plant germplasm is often a resource for finding resistance. For example, when breeding for resistance to powdery mildew (PM), many different studies characterized germplasm to find durable resistance in sunflower, barley, 69

wheat and pea (Ji et al. 2008; Reddy et al. 2013; Surlan-Momirovic et al. 2016; Vaz Patto et al. 2007) among other crops.

Ornamental plants in the genus Phlox are highly popular in landscapes worldwide but also highly susceptible to PM (Armitage 2008; Combs 2017; Zale and Jourdan 2015). PM diseases are caused by obligate biotrophic pathogens, able to generate large populations through the abundant production of asexual spores, which are capable of long-distance dispersal and are responsible for population clonality (Milgroom 2015, p.186-213).

Sexually derived chasmothecia primarily function in dormancy, and release ascospores to infect new hosts in the spring. PM pathogens most recovered from Phlox are the host- specific fungus Golovinomyces magnicellulatus (U. Braun) V.P. Heluta as well as species of Podosphaera. Takamatsu et al. (2006) proposed that the genus Golovinomyces has its center of origin in North America and that the interaction with its hosts has evolved through time, resulting in tremendous diversity of the genus. Therefore, this suggests that

Golovinomyces species have large genetic diversity and a high degree of genotype flow

(=movement of asexual propagules). McDonald and Linde (2002) hypothesized that genotype flow presents a high evolutionary potential that is likely to overcome genetic resistance because it represents a linked package of co-adapted alleles that have already been selected for over time. The genome of Golovinomyces magnicellulatus has been sequenced and annotated (Farinas et al. 2019b) using its haploid stage (asexual spores).

PM pathogens become diploid during sexual reproduction, when sexual recombination is likely to occur. In temperate climates, sexual reproduction is expected to happen at the end of every growing season. Blumera graminis, the PM pathogen of wheat, is thought to have a maximum of one sexual cycle per year (Wicker et al. 2013). As with other PM pathogen 70

genomes, G. magnicellulatus has a large genome (130 Mb) and contains 40% repeats. The high repeat content is mostly composed of transposons (Spanu et al. 2010) and is associated with extensive gene reshuffling.

Due to the plant’s historical popularity in landscapes and flower arrangements, many outdoor trials have been conducted to assess new Phlox cultivars for resistance to PM

(Coombs 2017; Robbins and Cartwright 2001). Thompson and Svejda (1964) were unsuccessful at identifying resistance in P. paniculata L. While two of the 49 cultivars included in the study initially showed resistance in field trials, they tended to lose it with time. The Chicago Botanic Garden (Glencoe, IL) has run trials of Phlox species and hybrids to evaluate horticultural characteristics such as bloom quality, and also their resistance to

PM (Hawke 1999; Hawke 2011). The 1999 review stated that the difference in susceptibility and resistance of garden Phlox to PM depended on its geographic location

(Hawke 1999), suggesting that different populations of PM can be found throughout the

US. Zale and Jourdan (2015) noted that triploid cultivars of P. paniculata, specifically

‘Robert Poore’ and ‘John Fanick’, were also associated with resistance to powdery mildew in the literature. However, other resistant cultivars such as ‘Shortwood’, ‘David’, and

‘Delta Snow’, were found to be diploid, which indicates that resistance to powdery mildew could have several different pathways. The authors advanced the hypothesis that germplasm introduction of wild-collected Phlox could increase genetic diversity and offer strategies for polyploid breeding. Indeed, exploring the genus diversity could help identify new sources of resistance (Armitage, 2008). It is important to realize that breeding for resistance has an impact on pathogen populations. Artificial selection happens every time a new resistant cultivar is developed and cultivated, resulting in a decreased number of 71

pathogen genotypes. The Phlox-PM pathosystem is likely subject to genetic selection, because of the instant popularity of less susceptible cultivars.

Additionally, the horticulture crop system is different from other crop systems, and in the context of exploring durable resistance, several aspects need to be taken in consideration. Compared to typical row crop production, the horticultural crop system differs in its global trade, heterogeneity, and disease tolerance (Parke and Grünwald 2012).

Methods to prevent the movement of pathogens through domestic and global trade are currently not sufficient: infected plant materials are often shipped across borders, as highlighted by Goss et al. (2009) who studied the movement of Phytophthora ramorum through U.S. nursery plants using genetic markers. Moreover, a typical ornamental greenhouse operation grows a large diversity of plants in a relatively small, enclosed, environment, and nurseries likewise produce many more different plant species and cultivars than the average farmer. Because of the high value of each individual plant and the market demand for highly attractive foliage and flowers, growers of ornamentals have a very low threshold for disease tolerance, resulting in extensive use of pesticides that contributes to small population sizes and limits genetic diversity (McDonald and Linde

2002).

Hence, this study had the following objectives: (i) to examine the genotypic diversity of a PM pathogen population of isolates collected in the eastern United States (U.S.) to better understand PM pathogen capacity to adapt to new environments and new resistant hosts; and (ii) to identify the range of plant responses to PM infection by using a combination of Phlox species and PM pathogen isolates.

4.3 Materials and Methods

Plant material and pathogen isolation. Powdery mildew-infected Phlox plants were collected from botanical gardens, residential properties, and garden centers across the eastern U.S. during spring, summer and fall of 2016-2019 (Table 4.1). Plants were removed from their original location, repotted, boxed and shipped with an appropriate U.S.

Department of Agriculture, Animal and Plant Health Inspection Service PPQ 526 permit to the Ornamental Pathology Laboratory at the Ohio State University (Columbus, OH).

Once received, each individual plant was enclosed in a plastic bag and placed in a growth chamber (Conviron PGR-15, Controlled Environments Inc., Pembina, ND) at a temperature of 24°C during the day and 22°C at night with a 14-hour photoperiod to induce sporulation. Single spore isolates were then cultured on detached P. paniculata cv. Starfire leaves maintained in vitro following the protocol described by Farinas et al. (2019a). From each sampling location, one or more single spore isolates were obtained from different leaves of the same plant sample (Table 4.1). Between 2016 and 2019 we obtained 32 isolates from infected Phlox plants collected across Eastern U.S. states. All isolates but one were from P. paniculata cultivars, and all isolates but one were identified as G. magnicellulatus (Table 4.1).

Genotype characterization. Molecular Genotyping. Total genomic DNA was extracted from single spore cultures using a Chelex extraction method (Walsh et al. 1991).

PCR was performed using primer pair ITS4/ITS5 (White et al. 1990) following the protocol of Matsuda and Takamatsu (2003) to identify isolates to the genus level. Five additional primer sets (Table 4.2) targeting the nuclear loci translation elongation factor alpha (TEF-

1α), chitin synthase I (CSI), histone (H3), β-tubulin (TUB), and partial intergenic spacer

(IGS), were designed (Primer-BLAST, NCBI) by aligning publicly available sequences

(KR815572.1, KR815611.1, KR815689.1, KR815650.1, KR815805.1 respectively) of

Golovinomyces orontii (Castagne) V.P. Heluta (Pirondi et al. 2016) to the whole genome sequence of G. magnicellulatus strain FPH2017-1 (Farinas et al. 2019b; isolate referred to in this study as OH1). All PCR reactions were performed according to the protocol of

Pirondi et al. (2016) adapted to a total volume of 25 µL. All PCR reagents were purchased from Thermo Fisher Scientific (Waltham, MA) and all oligonucleotides were synthesized by Eurofins MWG Operon (Eurofins Genomics LLC., Louisville, KY). Each reaction contained 2.5 µL of 10x DreamTaq Buffer, 2 µL of 25 mM MgCl2, 1 µL of 10 mM dNTPs,

1 µL of 10 µM of each primer, 0.25 µL DreamTaq DNA polymerase (5 U/µL), and 1 µL of extracted DNA. The PCR program was as follows: 95°C for 3 min, 35 cycles of 95°C for 30 s, 52-58°C (Table 4.2) depending on primer set for 30 s, and 72°C for 1 min, with a final extension of 72°C for 5 min. Amplicons were visualized on 1% agarose gels stained with GelRed® (Biotium Inc., Fremont, CA) following manufacturer’s instructions, to confirm presence and size of PCR products. These were then purified using ExoSAP-IT

(Affymetrix Inc., Santa Clara, CA), and sequenced in both directions using sanger sequencing (OSUCCC Genomics Shared Resource, Columbus, OH). Forward and reverse nucleotide sequences were aligned, checked for sequencing errors and edited manually, then aligned with MAFFT v.7.407 (Katoh and Standley 2013) and trimmed again using trimal v.1.4 (Capella-Gutierrez et al. 2009). Consensus sequences were then deposited in the GenBank database (Supplementary Table B.1).

Polymorphism and phylogenetic analyses. DnaSP v5 (Librado and Rozas 2009) was used to calculate the following parameters: sequence diversity, including the number of

SNPs; departure from an equilibrium-neutral model of evolution (Tajima’s D; Tajima

1989); number of haplotypes (h); haplotype diversity (Hd; Nei 1987); average number of pairwise nucleotide differences (k; Tajima 1983); pairwise nucleotide diversity (π; Nei

1987), and nucleotide polymorphism (θw; Watterson 1975; Table 4). Maximum likelihood phylogenetic trees were computed using IQ-TREE v.1.6.9 (Chernomor et al. 2016), with the option -m MFP+MERGE (Lanfear et al. 2012) and 100 bootstraps, and visualized using

FigTree v.1.4.4 (Rambaut et al. 2018). Because there was bootstrap support for incongruence between the generated trees, a neighbor-net phylogenetic network was also built to observe the evolutionary relationship of our population using SplitTree v.4.14.8

(Huson and Bryant 2006).

To verify if the number of G. magnicellulatus isolates analyzed in this study could be representative of the genetic diversity of eastern U.S. PM populations, we compared the whole genome sequence of two distantly related isolates (OH1 and OH2). The genome of isolate OH1 was previously sequenced and assembled (Farinas et al. 2019b) and the same protocol for DNA extraction was used for isolate OH2. Whole genome sequencing was executed using the Illumina HiSeq PE150 platform (Illumina Inc., San Diego, CA) with

100x coverage. OH2 reads (NCBI Biosample No. SAMN13608638) were mapped to the reference genome OH1 using Bowtie2 v.2.2.9 (Langmead et al. 2018). SAMTools v.1.3.1

(Li et al. 2009) and VCFTools v.0.1.14 (Danecek et al. 2011) were used to assess the number of single nucleotide polymorphisms (SNPs) and their location on the genome, the percentage of synonymous vs. non-synonymous SNPs, and small insertions and deletions 75

of 1-26 bp (INDEL). Lastly, secreted proteins identified in OH1 were compared to an OH2 assembly (tblastn) using BLAST v.2.4.0+ to test for gene polymorphism

(presence/absence) to further screen for genetic factors underpinning phenotypic diversity.

A threshold of 50% query coverage and an E-value of 1.2 were used to identify positive matches.

Phenotype characterization. Phlox germplasm screening. Plantlets of P. glaberrima

OPGC 3594, P. paniculata ‘Dunbar Creek’, P. subulata OPGC 4185, and P. amoena

OPGC 3598 were micropropagated from a collection maintained at the Ornamental Plant

Germplasm Center (OPGC) in Columbus, OH (Table 4.3) following the protocol described by Farinas et al. (2019a). Phlox paniculata ‘Dunbar Creek’ is a commercially available cultivar that is thought to be resistant to PM when planted outdoors (<25% infection;

Hawke 2011). It was here used as an internal control.

Five-month-old micropropagated plantlets maintained in tubes containing 15 mL of

MS medium were inoculated with nine isolates of G. magnicellulatus and one isolate of

Podosphaera sp. in separate experiments as described in Farinas et al. (2019a; Table 4.1).

Each treatment had six single plantlet replications and each experiment was conducted twice. After inoculation, the plantlets were completely randomized on a laboratory shelf at

24±1°C, with a 14-hour photoperiod. Latency period (i.e. days between inoculation and sporulation) was visually assessed using a dissecting microscope (Leica S6D, Leica

Microsystems Inc., Buffalo Grove, IL) every four days, up to 24 days post-inoculation

(dpi). To measure the latency period, we recorded presence or absence of spores at each evaluation and used the average of all replications (n=12). Disease severity was also visually assessed using a dissecting microscope every four days, up to 28 dpi. At each 76

rating, the percentage of leaf area covered by PM growth was visually assessed on each individual leaf; values for all leaves were then summed and divided by the number of leaves on each plantlet. Following disease severity ratings, the plantlets were repositioned in the trial. The Area Under the Disease Progress Curve (AUDPC) was calculated using disease severity ratings at 4-day intervals (Madden et al. 2007). After the final disease severity evaluation, spore production was evaluated using six infected leaves excised from each plantlet (n=12) and placed in individual microcentrifuge tubes containing 0.5 ml of a 0.1% sterilized Tween-20 water solution to retrieve spores. Tubes were vortexed for 20 seconds, centrifuged for 2 minutes at 21,130 × g, then vortexed again for 5 seconds. Spores were counted using a haemocytometer and two counts from each tube were averaged. Estimated spore production on each Phlox species was derived by dividing spore counts by the corresponding standardized plant dry weight according to the following formula: average spore number × haemocytometer constant × volume of spore suspension in tube/plant dry weight. The dry weight of each Phlox species was assessed by averaging the weight of 12 plantlets (leaves and stems) after 12 h in an oven at 60°C.

Data analysis. All statistical analyses were conducted using the R packages agricolae v. 1.2.8 (De Mendiburu, 2017) and nlme v. 3.1.131 (Pinheiro et al. 2019) in RStudio v.1.1.383 (RStudio Team 2016). In all experiments, treatment effects on the AUDPC, latency period, and spore production, were analyzed by ANOVA. Tukey’s HSD test (P <

0.05) was used to test for differences among means. All spore production data were log transformed to provide better homogeneity of variance prior to ANOVA analysis. The different runs of the experiments were treated as random effects and the treatments as fixed effects. Main effect and interaction of each experiment run was analyzed separately first 77

(Table B.2), then the data were pooled together. Regression analysis using mean values was used to analyze the relationship between the spore production data and latency period data collected between 4 and 24 dpi (Chambers 1992).

4.4 Results

Genotype characterization. Polymorphism analysis. In total, we sequenced 3920 nucleotides per pathogen isolate using six genetic markers. Eleven single nucleotide polymorphisms (SNPs) were observed among isolates in all genes except CSI and H3, which were therefore omitted from further analysis, resulting in 2785 total nucleotides per isolate. Total haplotype diversity (Hd) was 0.89. From the concatenated alignment of the four genes, 11 SNPs were identified, or 3.9 SNPs/kb (Table 4.4). TEF-1α showed the highest number of SNPs (0.46% of variable sites) but ITS had the highest number of haplotypes (n=4) and haplotype diversity (0.73; Table 4.4). Based on measurement of

Tajima’s D, none of the genes deviated from neutral evolution (Table 4.4). TEF-1α had the greatest nucleotide polymorphism (1.25), and the greatest pairwise nucleotide diversity with ITS (0.002 and 0.0022, respectively). Overall, low genetic diversity was detected in our population.

A total of 221,982 SNPs was found in isolate OH2 when mapped to the reference genome OH1, which corresponds to 1.7 SNPs/kb. Of the total number of SNPs, 77.4% were located in the intergenic regions, 2.7% were located in exons, and 0.4% were located in introns, which represent 1.3, 0.046 and 0.007 SNPs/kb, respectively. The ratio of non- synonymous vs synonymous substitutions (dN/dS) was 0.822. Additionally, 69,550 insertions and 805,503 deletions were found. A tBLASTn comparison of OH1 secreted 78

proteins against the OH2 draft assembly recovered only 285 out of 292 proteins. The seven proteins missing from OH2 have predicted functions such as putative oxidoreductase

(KND86368.1), glycoside hydrolase (OAA42022.1), putative effector protein

(RKF82013.1), SGNH hydrolase-type (OAA72930.1), chitinase (OAQ87645.1), cell wall galactomannoprotein (OAA35810), and biotrophy-associated secreted protein

(XP008096078.1).

Phylogenetic analysis. Phylogenetic incongruence was observed among the ML trees obtained from the different genes (Fig. 4.1 and Supplementary Fig. B.1). Only phylogenetic trees based on ITS and TEF-1α provided any phylogenetic resolution (Fig. 4.1). Gene tree topologies conflicted with regards to the placement of OH4, OH5, OH2, IA1, and WI1 isolates.

According to neighbor-net phylogenetic network analysis (Fig. 4.2), where the nodes represent the samples haplotypes and the edges represent the evolutionary relationship, a closed loop, or reticulation, was observed. Isolates with dot numbers (i.e. isolates retrieved from different leaves of the same plant sample) grouped together (i.e. OH4.1-3), indicating clonality among the isolates. Isolates from Mississippi and Tennessee, representing hardiness zone 7 and 8, grouped together, whereas all isolates in hardiness zone 5 and 6 have no apparent population structure.

Phenotype characterization. Phlox germplasm screening. Across both experimental runs, the main and interaction effects of PM isolate (n=10) and Phlox species (n=4) on

AUDPC were highly significant (P<0.001; Supplementary Table B.2). The data from the two experimental runs were pooled (Table 4.5), and the AUDPC for all PM isolates was assessed by Phlox species (Fig. 4.3). Phlox paniculata showed consistently low AUDPC 79

across all isolates with MI1 isolate having a significantly higher mean value (P=0.001). P. amoena also showed overall low AUDPC values for all but three G. magnicellulatus isolates, (MI1, OH1, and OH4.1, corresponding to 16, 14, and 7.3% final disease severity, respectively; P<0.001). In P. subulata, AUDPC values were overall slightly higher compared to P. paniculata and P. amoena and there was a significant difference between isolate OH1 (25.4% final disease severity) and isolates IA1 (4.3%), MI1 (2.9%), OH2

(3.6%), as well as Podosphaera sp. isolate VA1 (1.1%; P=0.005). P. glaberrima showed a varying responses; low AUDPC was recorded when infected with G. magnicellulatus isolates IA1 (3.3% final disease severity), MS1.1 (2.3%), NY1 (0.2%) and OH3.1 (3.1%), as well as Podosphaera sp. isolate VA1 (3.3%), while a very high value was recorded when the species was infected with isolate MI1 (35.2%; P<0.001).

When comparing latency periods, the main and interaction effects of PM isolate (n=10) and Phlox species (n=4) were highly significant across both experimental runs (P<0.001;

Supplementary Table B.2). The data from the two experimental runs were pooled (Table

4.5), and the latency period for all PM isolates was assessed by Phlox species (Fig. 4.4).

Across all four Phlox species, P. amoena and P. paniculata had a significantly longer latency period than P. glaberrima and P. subulata (P<0.001; Fig. 4.4).

Finally, when comparing spore production, the interaction effect of PM isolate (n=10) and Phlox species (n=4) was only significant in the second experimental run (P<0.001;

Table B.2). The data from the two experimental runs were pooled (Table 4.5). A negative relationship (r2=0.99) was observed between the average number of spores produced per mg of dry weight and the latency period (4 to 24 dpi), indicating that a lower number of spores was observed when it took longer for PM to sporulate (Fig. 4.5). 80

4.5 Discussion

This study is the first to characterize Phlox germplasm in vitro by identifying resistant traits in four Phlox species using several PM pathogen isolates. Moreover, it is the first to identify the diversity of the pathogen to infer durability of host resistance.

The population sampling used in this study was both spatial, encompassing eight U.S. states including four different hardiness zones (5 through 8), and temporal, as it was performed throughout four growing seasons (from 2016 to 2019). Six genetic markers were used in this study (TEF-1α, CSI, TUB, IGS, H3 and ITS), among which TEF-1α and ITS were the most informative. In our study, several isolates were sometimes retrieved from the same plant sample. Those isolates did not show genetic variation from one another, suggesting that they may have been clonally reproduced.

Gene trees did not recover consistent evolutionary relationships among certain G. magnicellulatus isolates and were generally poorly resolved. We therefore used a neighbor- net phylogenetic network analysis to further visualize our population structure. Neighbor- net phylogenetic network analyses are thought to deal with systematic errors that tree-based phylogenetic analyses are not protected from, such as mistakes in assumptions of a model

(Huson and Bryant 2006). Closed loops, or reticulations in the network demonstrate phylogenetic disagreement among polymorphic loci and could be the result of recombination, recurrent and/or reverse mutations (Milgroom, 2015, p.203-204), or incomplete lineage sorting of alleles. The availability of fungal tissue, genetic distance, contrasting phenotypic differences, and sympatric diversification were taken into consideration when choosing isolates OH1 and OH2 to provide a more in-depth analysis

of overall SNP diversity among isolates in our population. Lower genetic diversity was observed between these two genotypes (1.6 SNPs/kb) compared to what was observed in the whole population with the MLST analysis using all six loci (2.8 SNPs/kb). Other genomic comparison studies have found an average of 1 SNP/kb between three barley PM pathogen isolates (Hacquard et al. 2013), and between 0.1 to 1.1 SNPs/kb depending on the region of the wheat PM pathogen genome (Wicker et al. 2013). Wicker et al. (2013) suggested that the three isolates used in their study represent 2 haplogroups.

Discordant phylogenetic trees, along with a reticulated SNP network, suggest that the population is recombining. The low SNP density across the North East U.S. is consistent with recent dispersal among states via horticultural trade. Indeed, Phlox plants are grown in production facilities and shipped across and between states to retailer facilities. Because of the disease latency period, sometimes there are few or no signs of the pathogen on even infected plants. Unknowingly shipping infected plant material can therefore contribute to the spread of the pathogen. Sexual recombination and the genotypic diversity that results may allow populations to adapt to changing environments and to new resistant cultivars.

Indeed, according to the risk model proposed by McDonald and Linde (2002), pathogens undergoing recombination events have a higher risk of evolution, and recombined genotypes may result in new avirulence genes. However, in stable environments, recombinant genotypes would be less fit than their parents, reducing the advantage of sexual recombination (Milgroom 2015, p.183).

Strict clonality, as seen with some isolates recovered from the same plant sample (i.e.

OH3.1-2), is at the opposite end of the spectrum from random mating. PMs are well known for producing numerous asexual spores during the growing season, which results in large, 82

clonal populations. Hence, it appears that our Phlox PM population is navigating the continuum between random mating and strict clonality. New questions arise from our results; specifically, has our population recently evolved from an ancestral population, and, could there be a population structure reflecting different hardiness zones?

In this study, P. paniculata ‘Dunbar Creek’, which we used as a control in our germplasm screening, consistently showed low AUDPC when inoculated with different

PM isolates in a highly disease conducive environment. This indicates that this cultivar was the most resistant host among four different Phlox species tested. Based on the small

AUDPC value variation between isolates (some degree of resistance within all PM isolates), P. subulata showed quantitative resistance, P. amoena showed qualitative resistance due to its drastic AUDPC value differences between PM isolates (resistant or susceptible), and P. glaberrima seemed to have both quantitative and qualitative resistance to some PM pathogen isolates.

It should be noted that isolate OH4.1 was originally retrieved from P. paniculata

‘David’, a cultivar chosen as the 2002 Perennial Plant of the Year (Perennial Plant

Association, 2002). This cultivar was initially marketed as highly resistant to PM but has lost its resistance over the years since cultivar introduction. Changes in aggressiveness of pathogens selected by quantitative resistance is a highly studied phenomenon with conflicting results (Cowger and Brown 2019; Delmas et al. 2016). Interestingly, in this study, OH4.1 was not the most aggressive isolate. This study suggests that there is a genetic basis of resistance in Phlox germplasm, emphasizing the advantage of using available germplasm when breeding for resistance. The species P. amoena and P. glaberrima are rarely, if at all, used when breeding for new cultivars of garden Phlox. Combinations of 83

genes for resistance, or gene pyramids, are often used to obtain durable resistance (Mundt

2014), and individuals of these species could supply desirable genes. With our germplasm screening experiment, we also identified high phenotypic diversity among PM isolates, despite the low genetic diversity found with our MLST analysis and a pairwise whole genome SNP comparison. However, in a whole genome comparison of two sympatric PM pathogen isolates (OH1 and OH2), we identified 7 of 292 predicted secreted proteins in

OH1 that are absent from an assembly of OH2, which could in part explain the observed phenotypic diversity among isolates. Indeed, the predicted functions of these secreted proteins are consistent with effectors and enzymes that may facilitate the infection process or impact the success of quantitative plant defense compounds. Such compositional diversity among isolates has been shown to facilitate rapid evolution in populations of plant pathogens undergoing sexual and parasexual recombination (Plissonneau et al 2018).

When looking at the latency period data, we saw that P. paniculata and P. amoena show higher resistance compared to P. glaberrima and P. subulata. Latency period and number of spores produced are important factors in predicting epidemics. Longer latency periods, and lower number of spores produced can slow down an epidemic’s rate of progress and confer better host performance in the landscape (Delmas et al. 2016).

Epidemic frequency as well as mixed reproductive models mentioned above are crucial characteristics to consider when assessing the durability of quantitative resistance (Cowger and Brown 2019). This study is the first to use an in vitro tool to characterize Phlox germplasm for its resistance to PM and to identify resistance patterns to inform future breeding programs. Moreover, this study laid a foundation for characterizing the G. magnicellulatus population present in the eastern U.S. Further studies with higher number 84

of PM isolates and Phlox species as well as herbarium specimens infected with PM could help understand the movement and origin of G. magnicellulatus. The results of this study, taken together, set a strong foundation that will contribute to further research tackling the major challenges in PM pathogen genomics, identified by Bindschedler et al. (2016), including the understanding and analysis of genome variation and evolutionary potential of PM pathogens.

Table 4.1: List of powdery mildew isolates included in this study.

Isolate w Sampling site HZ x Year Host Study y P. paniculata ‘Robert VA1z Botanical garden 6 2016 G, P Poore’ P. paniculata ‘Tequila WI1 Botanical garden 5 2016 G sunrise’ MI1 Commercial nursery 5 2017 P. paniculata ‘Nicky' G, P

MI2 Botanical garden 5 2017 P. paniculata ‘David' G

MI3 Botanical garden 5 2017 P. paniculata ‘David' G

IA1 Residential property 5 2017 P. paniculata G, P

NY1 Residential property 5 2018 P. paniculata G, P

NY2 Residential property 5 2018 P. paniculata G

NY3.1 Residential property 5 2019 P. paniculata G

NY3.2 Residential property 5 2019 P. paniculata G

NY4.1 Residential property 5 2019 P. paniculata G

NY4.2 Residential property 5 2019 P. paniculata G P. × arendsii ‘Baby NY5.1 Commercial nursery 5 2019 G face’ P. × arendsii ‘Baby NY5.2 Commercial nursery 5 2019 G face’ P. paniculata ‘Robert NY6.1 Garden center 5 2019 G Poore’ P. paniculata ‘Robert NY6.2 Garden center 5 2019 G Poore’ MS1.1 Residential property 8 2018 P. paniculata G, P

Continued

Table 4.1 continued Study y Isolate w Sampling site HZ x Year Host

MS1.2 Residential property 8 2018 P. paniculata G

OH1 Residential property 6 2017 P. paniculata G, P P. paniculata OH2 Arboretum 6 2017 G, P ‘Shortwood' P. paniculata PZ11- OH3.1 Germplasm center 6 2018 G, P 044 P. paniculata PZ11- OH3.2 Germplasm center 6 2018 G 044 OH4.1 Botanical garden 6 2018 P. paniculata ‘David' G, P

OH4.2 Botanical garden 6 2018 P. paniculata ‘David' G

OH4.3 Botanical garden 6 2018 P. paniculata ‘David' G

OH5 Wild 5 2018 P. paniculata G, P

OH6.1 Residential property 6 2019 P. subulata G

OH6.2 Residential property 6 2019 P. subulata G P. paniculata ‘Flame TN1.1 Commercial nursery 7 2019 G Purple’ P. paniculata ‘Flame TN1.2 Commercial nursery 7 2019 G Purple’ TN2.1 Residential property 7 2019 P. paniculata ‘David’ G

TN2.2 Residential property 7 2019 P. paniculata ‘David’ G w Isolates are identified by the US state where they were found (VA= Virginia; WI= Wisconsin; MI= Michigan; IA= Iowa; NY= New York; MS= Massachusetts; OH= Ohio; TN= Tennessee) along with a sequential number. Isolate numbers separated by a dot indicate that they were retrieved from different leaves of the same plant sample. x HZ= Hardiness Zone. Retrieved from USDA Plant Hardiness Zone Map based on GPS coordinates of sampling sites. y Isolate was used for genotype (G) and/or phenotype (P) characterization experiments. z Isolate VA1 was identified as Podosphaera sp. All other isolates in the table are Golovinomyces magnicellulatus.

Table 4.2: List of primers designed in this study and used in the MLST analysis

Frag Annealing Primer ment Locus a Sequence Temp. name size (°C) (bp) tefgmF GAACCCGCAAGTAGAAAATGGG 57.7 1307 TEF-1α tefgmR GACGTTATCACCAGGAAGACCT

cs1gmF AGCTGTTTATGCTTTGGCTCATTT 55.0 622 CSI cs1gmR AGGATGTTGATGTAAGTTGGGACC

h3gmF ACCGCTCGCAAGGTATGTATT 52.8 611 H3 h3gmR AGACGACGAGCGAGCTGAAT

tub2gmF ATGTTCATCAAACTCCTGCTTCAAC 52.0 800 TUB tub2gmR CACCTTCTGTGTAATGTCCTTTGG

igsgmF GGGATGTATGGCGCACAGAA 56.2 513 IGS igsgmR AGGCCATGCGATTCGTTAAG a TEF-1α = translation elongation factor alpha; CSI= chitin synthase I; H3 = histone; TUB = β-tubulin; IGS = partial intergenic spacer.

Table 4.3: Phlox germplasm maintained at the Ohio Plant Germplasm Center (OPGC) that was used in this study.

OPGC Phlox Year and origin of Plant species Accession No. subsection collection

P. paniculata ‘Dunbar 4088 Paniculatae 1989, Fayette Co., PA Creek’

4185.7 P. subulata Subulatae 2013, Vinton Co., OH

3594 3.2 P. glaberrima Phlox 2011, Bullitt Co., KY

3598.6 P. amoena Divaricatae 2011, McCreary Co., KY

Table 4.4: Basic diversity indices calculated in the phylogenetic analysis conducted in this study.

Sequence Diversity Index b Lengt Locus a h (bp) SNPs Tajima’s h Hd k π θ (%) D w

TEF-1α 1080 5 (0.46) 1.79 3 0.54 2.08 0.002 1.25

TUB 668 2 (0.29) -1.26 3 0.32 0.19 0.00028 0.5

ITS 560 3 (0.5) 1.6 4 0.73 1.24 0.0022 0.73 0.06 IGS 477 1 (0.2) -1.15 2 0.06 0.00014 0.25 5 Concatenated 2785 11 (0.4) 0.91 12 0.89 3.5 0.001 2.75 a TEF-1α = translation elongation factor alpha; TUB = β-tubulin; ITS = internal transcriber spacer; IGS = partial intergenic spacer. b SNP: number of single nucleotide polymorphisms (excluding gaps); Tajima’s D: departure from an equilibrium-neutral model of evolution; h: number of haplotypes; Hd: haplotype diversity; k: average number of nucleotide differences; π: pairwise nucleotide diversity; θw: Watterson’s theta, nucleotide polymorphism.

Table 4.5: Summary statistics of the effect of PM isolate (n=10) and Phlox species (n=4) on disease progression, spore production and latency period observed on Phlox plantlets.

AUDPC a Spore production b Latency period c Factor F value P value F value P value F value P value

PM isolate 12.7 < 0.001 9.6 <0.001 8.9 < 0.001

Phlox species 14.7 < 0.001 17.4 < 0.001 25.9 < 0.001

PM isolate × Phlox 4.3 < 0.001 3.4 <0.001 4.2 < 0.001 species aAUDPC = Area under the disease progress curve. Values were calculated from disease severity data collected for 28 days post-inoculation. b Amount of spore produced per mg of plant dry weight 28 days post inoculation. Data were log-transformed prior to ANOVA analysis. c Latency period corresponds to the number of days between inoculation and sporulation.

Figure 4.5: Linear regression showing the relationship between the amount of spores produced on all infected plantlets from all Phlox species (n=480) 28 days post-inoculation and the latency period (i.e. days between inoculation and sporulation; 4-24 dpi). P <0.001. Intercept: 12.8; slope: -0.17; r2 = 0.99). Rhombus dots correspond to mean values  SE.

Chapter 5. Conclusions and Future Directions

Obligate biotrophic pathogens make for very challenging study systems as they cannot be grown in axenic conditions and need a healthy living host to survive. Nicot et al. (2002) extensively reviewed experimental methods that have allowed collecting information on the biology and epidemiology of PM pathogens across different crop systems using a variety of plant tissues maintained on artificial media (e.g. cotyledons, leaves, or plantlets).

While many studies have been conducted on PM pathogens of agricultural crops, very little research has been done in ornamentals. The biotrophic characteristics of PM pathogens present challenges also in the execution of genomic studies, for example rendering genome assembly difficult to perform. Indeed, many contaminants are collected alongside the PM pathogen during the fungal tissue collection process, requiring the availability of a robust bioinformatic pipeline to remove contaminants reads prior to genome assembly.

The ultimate goal of this research was to assess the resistance potential of Phlox germplasm to the disease PM. To do this, we generated two reliable bioassays to conduct in vitro experiments with PM pathogens of Phlox (objective 1). The leaf bioassay was used to generate single-spore PM isolates and to maintain isolates alive, while the micropropagated plantlet bioassay was used to screen Phlox species for their susceptibility to PM pathogens. The two bioassays were complementary and both necessary to complete the tasks. Indeed, PM inoculum must be free of other contaminants when used in inoculation experiments in order to prevent contamination of the plantlets or the growing 97

medium inside the micropropagation tube, where the presence of high relative humidity generates a highly conducive environment for disease development. Hence it is first necessary to generate clean PM inoculum on detached leaves in vitro and then use this inoculum to infect Phlox plantlets. Here, the detached leaf bioassay was also used to generate single spore isolates from each infected plant sample to account for the diversity of PM pathogens present on the plant.

For the purpose of the leaf bioassay, healthy, PM-free plants must be available at all times. As demonstrated in Chapter 2, preliminary experiments showed that vaporized sulfur applications to protect Phlox plants from PM infection in the greenhouse would not allow the pathogen to grow in vitro on leaves detached from those plants, even after several water rinses and undergoing disinfection. In search for an alternative method to keep plants clean, we explored the use of a 20% raw cow milk solution, which we sprayed weekly on the entire plant surface. The technique was efficient at controlling PM in the greenhouse, however, when using Phlox plants that had been treated with milk for a period of approximately 10 months, less disease incidence and lower spore production were observed after the pathogen was inoculated in vitro (data not shown). It is believed that the microorganisms contained in raw milk activate the systemic resistance of the plant (Bettiol

1999; Kamel et al. 2017; Stadnik and Bettiol 2001; Sudisha et al. 2011). Another hypothesis is that milk treatments may generate the production of Reactive Oxygen Species

(ROS; Medeiros et al. 2012). To overcome the challenge of needing plant material that was highly susceptible to infection, while also maintaining PM-free plants in the greenhouse, in this study we used plants that were sprayed with raw milk for a maximum of six months before being replaced. In our study, raw milk was frozen between each application to 98

lengthen its shelf-life, which did not seem to impact its efficacy in controlling disease.

Subsequent studies should focus on identifying if and which specific organisms are associated with the curative effect of raw milk and their potential use as biological control.

An alternative to disease management in the greenhouse, is to maintain clean plants inside growth chambers using strict sanitary precautions to avoid introducing PM inoculum. In our study, the use of a growth chamber proved efficient at maintaining plants

PM-free without the use of raw milk or sulfur applications. Our sanitary precautions included the use of a clean lab coat and gloves anytime we would visit the plants. If PM growth was to be seen on the plants, a complete sanitation of the growth chamber and starting with new Phlox stock was necessary. If feasible, this technique is less demanding than maintaining Phlox plants in the greenhouse with weekly raw milk applications.

Concerning the use of a micropropagated plantlet collection, in our study we standardized experiments by using 5-month-old plantlets of each Phlox species we screened. However, certain species had a faster growth rate than others and reached the top of the tube within 4 months. If new projects were to use the plantlet bioassays protocol, it is recommended to reduce the plantlets growth to 2 months in MS media with cytokinin, and 2 months without before proceeding with inoculation. This would be slightly quicker and would not impact the results of the experiments.

Our second research objective was to screen different Phlox species for their susceptibility to PM. Our goal was to generate information on the resistance potential of

Phlox germplasm, which contains about 65 species. At the same time, we aimed to provide a proof of concept about the usefulness of using the bioassays we had developed to screen a large quantity of plants while removing environmental variation. Particularly, the 99

plantlets bioassay was key to using minimal space and standardizing environmental conditions. However, this method demands specialized technical skills and is high maintenance. In our study, we wanted to examine the aggressiveness of each PM isolate separately, hence, we grew and inoculated each isolate in separate experiments. However, if using the developed protocol for a different purpose, such as whole germplasm screening, it would be easier to use multiple PM isolates combined and maintained artificially on a leaf bioassay, instead of several PM isolates in separate experiments.

Based on the results of our germplasm screening, breeders should incorporate Phlox paniculata ‘Dunbar Creek’ in their breeding programs. However, it will be useful to continue screening the germplasm for resistance. Indeed, continuous breeding with various

Phlox species that possess different horticultural attributes and resistant traits as well will satisfy the particularities of green industry consumers and their appeal for constant change.

To fulfill our third study objective we ran into the challenge of developing a bioinformatic pipeline for whole genome sequence analysis that would remove most non-

PM contaminants from our assembly. Iterative runs of filtering steps, as explained in

Chapter 3, were successful at removing most sequences of contaminants from bacteria, oomycetes, insects, and plants. The visualization of the filtering steps with the software

Vizbin (data not shown) was key in confirming that we kept most of PM genome reads but removed the rest. However, our filtering parameters were somewhat conservative, and our final draft assembly still contains some contaminants. Unfiltered raw reads are publicly available (NCBI Biosample No. SAMN13608638) and could be used by other teams when more bioinformatic tools are developed, which could help future assemblies.

100

To investigate the genetic diversity of a G. magnicellulatus population of isolates collected from the eastern U.S. we used MLST analysis paired with whole genome comparison analysis of two genetically distant isolates. In the MLST analysis we used four housekeeping genes (CSI, H3, TUB, TEF-1α), plus ITS and IGS, which are non-coding genes. We recovered 2.8 SNPs/kb, a higher SNPs content than what observed when using whole genome comparison analysis, in which we found 1.6 SNPs/kb. Considering our goal to identify genetic diversity among our isolates, our MLST analysis was sufficient at indicating the low genetic diversity in our population. However, using whole genome comparison, we were able to look at specific coding regions predicted to be involved in the infection process or impacting quantitative plant defense compounds. Specifically, we looked at the distribution of more than 200 predicted secreted proteins. Measuring that type of diversity with an MLST analysis would not be possible. Hence, our MLST analysis was sufficient to answer questions about population structure and genetic diversity, whereas whole genome comparison was necessary to explain phenotypic diversity by identifying specific predicted genes as differently arranged between genomes. In future studies, whole genome sequencing should be used to compare the presence/absence of predicted secreted proteins in many PM pathogen isolates. The identification of specific effectors involved in the infection process is important to target resistant genes and engineer Phlox plants, as well as other ornamental plants.

This study represents the first attempt to develop laboratory bioassays for the ornamental genus Phlox. It provides effective and useful tools for phytopathological studies on the biology and epidemiology of Phlox-PM pathogens, as well as information on PM-resistance mechanisms in Phlox. In addition, it provides breeders with useful tools 101

for screening resistant lines. The bioinformatic pipeline developed in this study can serve many other projects looking to remove contaminants from their sequenced reads. Aside from the technical tools, this study revealed insights into the genetic diversity and population structure of G. magnicellulatus and informed us of the role of the horticultural trade in the spread of pathogens. Because PM is a critical disease of other popular ornamental genera such as Gerbera, Monarda, and Rudbeckia, the tools developed in this study can serve as a model for other important PM-ornamental systems as well.

102

References

Abbott, S.P., and Currah, R.S. 1989. The larger cup fungi and other Ascomycetes of Alberta, an annotated checklist. Devonian Botanic Garden, University of Alberta, Edmonton, 96 pages; Ginns, J.H. 1986. Compendium of plant disease and decay fungi in Canada 1960-1980. Res. Br. Can. Agric. Publ. 1813: 416. Azmat, M.A., A.A. Khan, H.M.N. Cheema, M. Ashraf and S. Niaz, 2013. Detached leaf assay coupled with microscopic conidial quantification: an efficient screening method for powdery mildew resistance in pea. Int. J. Agric. Biol., 15: 957–962 Adams, G.C. Jr., Gottwald, T.R., Leach, C.M. 1986. Environmental factors initiating liberation of conidia of powdery mildews. Phytopathology 76:1239–45 Alfieri Jr., S.A., Langdon, K.R., Wehlburg, C., and Kimbrough, J.W. 1984. Index of Plant Diseases in Florida (Revised). Florida Dept. Agric. and Consumer Serv., Div. Plant Ind. Bull. 11: 1-389 Ali, N., Otto, P., and Jage, H. 2000. Beitrage zur Kenntnis phytoparasitischer Pilze im Stadtgebiet von Leipzig (Sachsen). Boletus 23: 103-118. Amano, K. (Hirata) 1986. Host range and geographical distribution of the powdery mildew fungi. Japan Sci. Soc. Press, Tokyo, 741 pages Andrews, S. 2010. FastQC: a quality control tool for high throughput sequence data. Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc Anonymous 1960. Index of Plant Diseases in the United States. U.S.D.A. Agric. Handb. 165: 1-531 Antipov, D., Korobeynikov, A., McLean, J. S., & Pevzner, P. A. 2015. hybridSPAdes: an algorithm for hybrid assembly of short and long reads. Bioinformatics 32(7):1009- 1015. Arabi, M.I.E, Jawhar, M. 2002. The ability of barley powdery mildew to grow in vitro. J. Phytopathology. 150: 304-307. Armenteros, J. J. A., Tsirigos, K. D., Sønderby, C. K., Petersen, T. N., Winther, O., Brunak, S., von Heijne, G., Nielsen, H. 2019. SignalP 5.0 improves signal peptide predictions using deep neural networks. Nature biotechnology 37:420-423 Armitage, A. M. 2008. Herbaceous perennial plants. A treatise on their identification, culture, and garden attributes. Third ed. Stipes Pub LLC. pp. 787-800 Asalf, B., Gadoury, D.M., Tronsmo, A.M., Seem, R.C., Dobson, A., Peres, N.A., and Stensvand, A. 2014. Ontogenic resistance of leaves and fruit, and how leaf folding 103

influences the distribution of powdery mildew on strawberry plants colonized by Podosphaera aphanis. Phytopathology. 104:954-963. Bao, W., Kojima, K.K., Kohany, O. 2015. Repbase Update, a database of repetitive elements in eukaryotic genomes. Mob DNA 6(1):11 Bardin, M., Carlier, J., Nicot, P.C. 1999. Genetic differentiation in the French population of Erysiphe cichoracearum, a causal agent of powdery mildew of cucurbits. Plant Pathol. 48: 531-540. Belanger, R.R, and Labbe C. 2002. Control of powdery mildews without chemicals: prophylactic and biological alternatives for horticultural crops. In: Be ́langer RR, Bushnell WR, Dik AJ, Carver TLW (eds), The Powdery Mildews: A Comprehensive Treatise. APS Press, St Paul, MN, pp. 256-267. Bertetti, D., Ortega, S. F., Gullino, M. L., & Garibaldi, A. 2016. First report of powdery mildew caused by Golovinomyces magnicellulatus on Phlox paniculata grown in Italy. Protezione delle Colture, (3), 18-20. Bolay, A. 2005. [Powdery mildews of Switzerland (Erysiphaceae)]. Cryptog. Helv. 20: 1-176. Bettiol, W. 1999. Effectiveness of cow's milk against zucchini squash powdery mildew (Sphaerotheca fuliginea) in greenhouse conditions. Crop Prot., 18:489-492. Bindschedler, L. V., Panstruga, R., and Spanu, P. D. 2016. Mildew-omics: how global analyses aid the understanding of life and evolution of powdery mildews. Frontiers in Plant Science, 7: 123. Bir, R.E, and Conner, J.L. 2002. Mildew resistant garden Phlox. N.C state University. http://www.ces.ncsu.edu/fletcher/staff/rbir/Phloxmildew.html Bolger, A. M., Lohse, M., & Usadel, B. (2014). Trimmomatic: A flexible trimmer for Illumina Sequence Data. Bioinformatics 30(15):2114-2120 Box, G.E.P., Jenkins, G.M., and Reinsel G.C. 1994. Time series analysis: forecasting and control. 3rd Edition, Holden-Day. Braun U., Cook, R.T.A., Inman, A.J., Shin, H.-D. 2002. The taxonomy of the powdery mildew fungi. In: Be ́langer RR, Bushnell WR, Dik AJ, Carver TLW (eds), The Powdery Mildews: A Comprehensive Treatise. APS Press, St Paul, MN, pp. 13–55. Braun, U. 1995. The Powdery mildews (Erysiphales) of Europe. Gustav Fischer Verlag, 337 pages Braun, U., Ale-Agha, N., Bolay, A., Boyle, H. , Brielmaier-Liebetanz, U., Emgenbroich, D., Kruse, J., and Kummer, V. 2009. New records of powdery mildew fungi (Erysiphaceae). Schlechtendalia 19: 39-46. Braun, U., and Cook, R.T.A. 2012. Taxonomy Manual of the Erysiphales (Powdery Mildews). CBS Biodivers. Ser. 11: 703. Bushnell, B. 2014. BBMap: a fast, accurate, splice-aware aligner (No. LBNL-7065E). Lawrence Berkeley National Lab. (LBNL), Berkeley, CA (United States)

104

Bushnell, W. R. 1967. Symptom development in mildewed and rusted tissues. Pages 21- 39: The dynamic role of molecular constituents in Plant-Parasite interaction. C.J. Mirocha and I. Uritani, eds. American Phytopathology Society, St. Paul Bushnell, W. R., & Allen, P. J. 1962. Induction of disease symptoms in barley by powdery mildew. Plant Physiology, 37(1), 50. Cantarel, B. L., Korf, I., Robb, S. M., Parra, G., Ross, E., Moore, B., Holt, C., Sanchez Alvarado, A., and Yandell, M. 2008. MAKER: an easy-to-use annotation pipeline designed for emerging model organism genomes. Genome Res. 18:188-196. Capella-Gutierrez, S., Silla-Martinez, J. M., Gabaldon, T. 2009. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics 25: 1972-1973. Cartwright, J. Robbins, G. Klingaman. 2001. Evaluation of Phlox paniculata cultivars for susceptibility to powdery mildew. Horticultural Studies 2001, University of Arkansas: 95-96. Chambers, J. M. 1992. Linear models. Chapter 4 of Statistical Models in S eds J. M. Chambers and T. J. Hastie, Wadsworth & Brooks/Cole. Chernomor, O., von Haeseler, A., and Minh, B. Q. 2016. Terrace aware data structure for phylogenomic inference from supermatrices. Syst. Biol., 65:997-1008 Cho, W.D., and Shin, H.D., Eds. 2004. List of plant diseases in Korea. Fourth edition. Korean Society of Plant Pathology, 779 pages Coghlan SE, Walters DR. 1990. Polyamine metabolism in ‘green-islands’ on powdery mildew-infected barley leaves: possible interactions with senescence. New Phytol. 116:417–24 Cook, R.T.A, and Braun, U. 2009. Conidial germination patterns in powdery mildews. Mycological Res. 113: 616-636 Cook, R.T.A., Henricot, B., Henrici, A., and Beales, P. 2006. Morphological and phylogenetic comparisons amongst powdery mildews on Catalpa in the UK. Mycol. Res. 110: 672-685 Coombs, G. 2017. Phlox for the mid-atlantic region. Research Report. Mt. Cuba center Cowger, C., & Brown, J. K. 2019. Durability of quantitative resistance in crops: greater than we know? Annual Review of Phytopathology, 57: 253-277. Crisp, P., Wicks, T. J., Troup, G., & Scott, E. S. 2006. Mode of action of milk and whey in the control of grapevine powdery mildew. Australasian Plant Pathology, 35(5), 487- 493. Crous, P.W., Phillips, A.J.L., and Baxter, A.P. 2000. Phytopathogenic Fungi from South Africa. University of Stellenbosch, Department of Plant Pathology Press, 358 pages

105

Cunnington, J. 2003. Pathogenic fungi on introduced plants in Victoria. A host list and literature guide for their identification. Department of Primary Industries, Research Victoria : 57. Danecek, P., Auton, A., Abecasis, G., Albers, C. A., Banks, E., DePristo, M. A., Handshaker, R.E., Lunter, G., Marth, G.T.,Sherry, S.T.,McVean, G., Durbin, R. 2011. The variant call format and VCF tools. Bioinformatics, 27(15), 2156-2158. Davidian, M., and Giltinan, D.M. 1995. Nonlinear mixed effects models for repeated measurement data. Chapman and Hall. De Mendiburu, F. 2017. Agricolae: Statistical procedures for agricultural rResearch. R package version 1.2-8. https://CRAN.R-project.org/package=agricolae De Mendiburu, Felipe. 2009. Una herramienta de analisis estadistico para la investigacion agricola. Tesis. Universidad Nacional de Ingenieria (UNI-PERU). Deighton, F.C. 1936. Preliminary list of fungi and diseases of plants in Sierra Leone and list of fungi collected in Sierra Leone. Bull. Misc. Inform. Kew 1936: --. Delmas, C. E., Fabre, F., Jolivet, J., Mazet, I. D., Richart Cervera, S., Deliere, L., & Delmotte, F. 2016. Adaptation of a plant pathogen to partial host resistance: selection for greater aggressiveness in grapevine downy mildew. Evolutionary Applications, 9(5), 709-725. Dudka, I.O., Heluta, V.P., Tykhonenko, Y.Y., Andrianova, T.V., Hayova, V.P., Prydiuk, M.P., Dzhagan, V.V., and Isikov, V.P. 2004. [Fungi of the Crimean Peninsula] (Translated from Russian). M.G. Kholodny Institute of Botany, National Academy of Sciences of Ukraine, 452 pages. Dynowska, M., Fiedorowicz, G., and Kubiak, D. 1999. Contributions to the distribution of Erysiphales in Poland. Acta Mycol. 34: 79-88; Mulenko, W., Majewski, T., and Ruszkiewicz-Michalska, M. 2008. A preliminary checklist of Micromycetes in Poland. W. Szafer Institute of Botany, Polish Academy of Sciences 9: 752. Eriksson, O.E. 1992. The non-lichenized pyrenomycetes of Sweden. Btjtryck, Lund, Sweden, 208 pages. Eriksson, O.E. 2014. Checklist of the non-lichenized ascomycetes of Sweden. Acta Univ. Upsal., Symb. Bot. Upsal. 36: 499. Fakirova, V.I. 1991. Fungi Bulgaricae, 1 tomus, ordo Erysiphales. In Aedibus Academiae Scientiarum Bulgaricae, Serdicae, 154 pages Fauteux, F., Chain, F., Belzile, F., Menzies, J. G., & Bélanger, R. R. 2006. The protective role of silicon in the Arabidopsis–powdery mildew pathosystem. Proceedings of the National Academy of Sciences, 103(46), 17554-17559.

Farinas, C., and Peduto Hand, F. 2017. Genotypic characterization of isolates of Golovinomyces magnicellulatus, the biotrophic powdery mildew pathogen of Phlox. Phytopathology. 107(12S):S5.122. 106

Farinas, C., Gluck-Thaler, E., Slot, J. C., Peduto Hand, F. 2019b. Whole-genome sequence of the Phlox powdery mildew pathogen Golovinomyces magnicellulatus strain FPH2017-1. Microbiology Resource Announcements 8(36) e00852-19. Farinas, C., Hand, F. 2017. Genotypic characterization of isolates of Golovinomyces magnicellulatus, the biotrophic powdery mildew pathogen of Phlox. Phytopathology. 107(12S): S5.122. Farinas, C., Jourdan, P., Paul, P. A., & Peduto Hand, F. 2019a. Development and Evaluation of Laboratory Bioassays to Study Powdery Mildew Pathogens of Phlox In Vitro. Plant disease, PDIS-01. Farinas, C., Jourdan, P., Paul, P., Slot, J., Daughtrey, M., Devi Ganeshan, V., Baysal-Gurel, F., Peduto Hand, F. Phlox species show quantitative and qualitative resistance to a population of powdery mildew isolates from the eastern United States. Phytopathology. First Look https://doi.org/10.1094/PHYTO-12-19-0473-R Ferguson, C. J., F. Krämer and R. K. Jansen. 1999. Relationships of eastern North American Phlox (Polemoniaceae) based on ITS sequence data. Systematic Bot. 24: 616-631. Ficke, A., Gadoury, D.M, and Seem, R.C. 2002. Ontogenic resistance and plant disease management: a case study of grape powdery mildew. Phytopathology. 92:671-675. Firman, I.D. 1972. A list of fungi and plant parasitic bacteria, viruses and nematodes in Fiji. Phytopathol. Pap. 15: 1-36; Frantz, J. M., Locke, J. C., and Mattson, N. 2010. Research Update: Does Silicon Have a Role in Ornamental Crop Production? Ofa Research. Number 924. French, A.M. 1989. California Plant Disease Host Index. Calif. Dept. Food Agric., Sacramento, 394 pages. Fuchs, H. 1994. Phlox, Stauden-und PolsterPhlox. Eugen Ulmer GmbH & Co, Stuttgart, Germany, pp. 198-199. Gadoury, D. M., and Seem, R. C. 1995. Development of ontogenic resistance to powdery mildew (Uncinula necator) in fruit of Concord grapevines. (Abstr.) Phytopathology 85:1149. Gadoury, D. M., Seem, R. C., and Wilcox, W. F. 1997. Early ontogenic resistance to powdery mildew in Chardonnay and Riesling grapes. (Abstr.) Phytopathology 87(suppl.):S31. Gadoury, D. M., Seem, R. C., and Wilcox, W. F. 2000. The early development of ontogenic resistance to powdery mildew in fruit of Vitis labruscana and Vitis vinifera grapevines. Proc. Int. Workshop Grapevine Downy and Powdery Mildew, 3rd. SARDI Res. Rep. Ser. 50:111-112. Gadoury, D. M., Seem, R. C., Ficke, A., and Wilcox, W. F. 2003. Ontogenic resistance to powdery mildew in grape berries. Phytopathology. 93:547-555.

107

Gilbertson, R.L., Cummins, G.B., and Darnall, E.D. 1979. Indexes to W.G. Solheim's Mycoflora Saximontanensis Exsiccata. Mycotaxon 10: 49-92. Gilman, J.C., and Archer, W.A. 1929. The Fungi of Iowa Parasitic on Plants. Iowa State Coll. J. Sci. 3: 299-507 Ginns, J.H. 1986. Compendium of plant disease and decay fungi in Canada 1960-1980. Res. Br. Can. Agric. Publ. 1813: 416. Girilovich, I.S., Gulis, V.I., Khramtsov, A.K., and Poliksenova, V.D. 2005. [Micromycetes of state national park of republik Belarus II. Powdery mildew fungi]. Mikol. Fitopatol. 39: 24-30. Glawe, D.A. 2008. The powdery mildews: a review of the world’s most familiar (yet poorly known) plant pathogens. An. Review of Phyto. 46:27-51. Goos, R.D. 2010. The mycota of Rhode Island: a checklist of the fungi recorded in Rhode Island (including lichens and myxomycetes) Vol. 4. Rhode Island Natural History Survey, 222 pages Gorter, G.J.M.A. 1977. Index of plant pathogens and the diseases they cause in cultivated plants in South Africa. Republic South Africa Dept. Agric. Techn. Serv. Pl. Protect. Res. Inst. Sci. Bull. 392: 1-177. Goss, E. M., Larsen, M., Chastagner, G. A., Givens, D. R., and Grünwald, N. J. 2009. Population genetic analysis infers migration pathways of Phytophthora ramorum in US nurseries. PLoS Pathogens, 5(9), e1000583. Grand, L.F., Ed. 1985. North Carolina Plant Disease Index. North Carolina Agric. Res. Serv. Techn. Bull. 240: 1-157. Green, J.R.; Carver, T.L.W.; Gurr, S.J. 2002. The formation and function of infection and feeding structures. In: Be ́langer RR, Bushnell WR, Dik AJ, Carver TLW (eds), The Powdery Mildews: A Comprehensive Treatise. APS Press, St Paul, MN, pp. 66–82. Grenon P., Smith B. 2004. SNAP and SPAN: Towards dynamic spatial ontology. Spat. Cognit. Comput. 4:69–104 Hacquard, S., Kracher, B., Maekawa, T., Vernaldi, S., Schulze-Lefert, P., and van Themaat, E. V. L. 2013. Mosaic genome structure of the barley powdery mildew pathogen and conservation of transcriptional programs in divergent hosts. Proceedings of the National Academy of Sciences, 110(24), E2219-E2228. Hawke, R. 1999. Plant evaluation notes. An evaluation report of selected Phlox species and hybrids. Chicago Botanic Garden. Issue 13. Hawke, R. 2001. Plant evaluation notes. A comparative study of Phlox paniculata cultivars. Chicago Botanic Garden. Issue 35. Hermansen JE, Torp U, Prahm LP. 1978. Studies of transport of cereal mildew and rust fungi across the North Sea. Grana 17:41–46

108

Hirata, T., Cunnington, J.H., Paksiri, U., Limkaisang, S., Shishkoff, N., Grigaliunaite, B., Sato, Y., Takamatsu, S. 2000. Evolutionary analysis of subsection Magnicellulatae of Podosphaera section Sphaerotheca (Erysiphales) based on the rDNA internal transcribed spacer sequences with special reference to host plants. Can. J. Bot. 78: 1521-1530 Hollomon, D.W. and Wheeler, I.E. 2002. Controlling powdery mildews with chemistry. In: Be ́langer RR, Bushnell WR, Dik AJ, Carver TLW (eds), The Powdery Mildews: A Comprehensive Treatise. APS Press, St Paul, MN, pp. 256-267. Huson, D. H., Bryant, D. 2006. Application of phylogenetic networks in evolutionary studies, Mol. Biol. Evol., 23(2):254-267. Ito, M., Takamatsu, S. 2010. Molecular phylogeny and evolution of subsection Magnicellulatae (Erysiphaceae: Podosphaera) with special reference to host plants. Mycoscience. 51: 34-43 Jarosz, A., and Levy, M. 1981. Resistance patterns and evidence of specialization in a feral host-obligate pathogen interaction: Phlox-erysiphe cichoracearum. Bot. Soc. Amer. Misc. Publ. 160: 30. Jarosz, A.M., Sheets, M., Levy, M. 1982. Cuticle thickness in Phlox and resistance to powdery mildew: an unreliable line of defense. Amer. J. Bot. 69:824-828. Ji, X., Xie, C., Ni, Z., Yang, T., Nevo, E., Fahima, T., Liu, Z.,Sun, Q. 2008. Identification and genetic mapping of a powdery mildew resistance gene in wild emmer (Triticum dicoccoides) accession IW72 from Israel. Euphytica, 159(3), 385-390. Jones, D.R., and Baker, R.H.A. 2007. Introductions of non-native plant pathogens into Great Britain, 1970-2004. Pl. Pathol. 56: 891-910. Jones, P., Binns, D., Chang, H. Y., Fraser, M., Li, W., McAnulla, C., McWilliam, H., Maslen, J., Mitchell, A., Nuka, G., Pesseat, S., Quinn, A.F., Sangrador-Vegas, A., Scheremetjew, M., Yong, S-Y., Lopez, R., Hunter, S. 2014. InterProScan 5: genome- scale protein function classification. Bioinformatics 30(9):1236-1240 Kamel, S.M., Ketta, H.A., and Emeran, A.A. 2017. Efficacy of raw milk and whey against cucumber powdery mildew disease caused by Sphaerotheca fuliginea (Schlecht.) Pollacci under plastic house conditions. Eg. J. of Bio. Pest Con. 27:135-142. Kenyon, D.M., Dixon, G.R., Helfer, S. 1995. Culture in vitro of Phododendron and Erysiphe sp. Plant Pathology. 44: 350-354. Kiss, L., Cook, R.T.A., Saenz, G.S., Cunnington, J.H., Takamatsu, S., Pascoe, I.G., Bardin, M., Nicot, P.C., Sato, Y., and Rossman, A.Y. 2001. Identification of two powdery mildew fungi, Oidium neolycopersici sp. nov. and O. lycopersici, infecting tomato in different parts of the world. Mycol. Res. 105: 684-697; Laird, N.M., and Ware, J.H. 1982. Random-effects models for longitudinal data. Biometrics. 38: 963-974.

109

Lanfear, R., Calcott, B., Ho, S. Y., and Guindon, S. 2012. Partition Finder: Combined selection of partitioning schemes and substitution models for phylogenetic analyses. Molecular Biology and Evolution, 29(6), 1695-1701. Langmead, B., Wilks, C., Antonescu, V., and Charles, R. 2018. Scaling read aligners to hundreds of threads on general-purpose processors. Bioinformatics, 35(3), 421-432. Lebeda, A., Mieslerova, B. 2011. Taxonomy, distribution and biology of lettuce powdery mildew (Golovinomyces cichoracearum sensu stricto). Plant Path. 60:400-415. Li, H., Handsaker, B., Wysoker, A., Fennell T., Ruan J., Homer N., Marth G., Abecasis G., Durbin R. and 1000 Genome Project Data Processing Subgroup. 2009. The Sequence alignment/map (SAM) format and SAMtools. Bioinformatics, 25, 2078-9. Librado, P., Rozas, J., 2009. DNASP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25:1451–2. Linde, M. and Debener, T. 2003. Isolation and identification of eight races of powdery mildew of roses (Podosphaera pannosa) (Wallr.: Fr.) de Bary and the genetic analysis of the resistance gene Rpp1. Theor. Appl. Genet. 107: 256-262. Lindstrom, M.J., and Bates, D.M. 1990. Nonlinear mixed effects models for repeated measures data, Biometrics. 46:673-687. Littel, R.C., Milliken, G.A., Stroup, W.W., and Wolfinger, R.D. 1996. SAS systems for mixed models. SAS Institute. Liu, D., Liu, Q., Zhang, P., Tan, K., & Chen, J. 2020. First Report of Powdery Mildew Caused by Golovinomyces magnicellulatus on Phlox drummondii in China. Plant Dis. 104:5, 1547-1547. Locklear, J. H. 2011. Phlox. A natural history and gardener’s guide. Timber Press, Inc. Madden, L. V., Hughes, G., and van den Bosch, F. 2007. The study of plant disease epidemics. The American Phytopathological Society, APS Press St. Paul, Minnesota. Maneval, W.E. 1937. A List of the Missouri Fungi. Univ. Missouri Stud., Sci. Ser. 12: --. Matsuda, S., Takamatsu, S. 2003. Evolution of host–parasite relationships of Golovinomyces (Ascomycete: Erysiphaceae) inferred from nuclear rDNA sequences. Mol Phylo and Evol. 27:314-327 McDonald, B. A., and Linde, C. 2002. Pathogen population genetics, evolutionary potential, and durable resistance. Annual review of phytopathology, 40(1), 349-379. McGrath, M.T., Shishkoff, N. 2001. Resistance to triadimefon and benomyl: dynamics and impact on managing cucurbit powdery mildew. Plant. Dis. 85: 147-154. Medeiros, F.H.V., W. Bettiol, R. M. Souza, E. Alves, Z. V. Pinto and R Lost. 2012. Microorganisms, application timing and fractions as players of the milk-mediated powdery mildew management. Crop Prot., 40: 8-15. Milgroom, M. G. 2015. Population biology of plant pathogens: genetics, ecology, and evolution. APS Press. 110

Mishina, G.N., Talieva, M.N., Babosha, A.V., Serezhkina, G.V., and Andreev, L.N. 2002. Influence of phytohormones on development of conidial inoculum of causative agents of the phlox and barley powdery mildew. Biology bulletin. 29:55-62. Mohanta, T. K., and Bae, H. 2015. The diversity of fungal genome. Biological procedures online, 17(1): 8. Mulenko, W., Majewski, T., and Ruszkiewicz-Michalska, M. 2008. A preliminary checklist of Micromycetes in Poland. W. Szafer Institute of Botany, Polish Academy of Sciences 9: 752. Mundt, C. C. 2014. Durable resistance: a key to sustainable management of pathogens and pests. Infection, Genetics and Evolution, 27, 446-455. NASS, 2014. https://quickstats.nass.usda.gov/#CD852985-E809-3D46-88F1- 37DA2C7134B8 Nei, M. 1987. Molecular evolutionary genetics. Columbia University Press, New York, NY, USA. Nicot PC, Bardin M, Dik AJ, 2002. Basic methods for epidemiological studies of powdery mildews: culture and preservation of isolates, production and delivery of inoculum, and disease assessment. In: Be ́langer RR, Bushnell WR, Dik AJ, Carver TLW (eds), The Powdery Mildews: A Comprehensive Treatise. APS Press, St Paul, MN, pp. 83– 99. Park, M-J., Park, J-H., Lee, S-G., Shin, H-D. 2010. First confirmed report on powdery mildew of Phlox paniculata caused by Golovinomyces magnicellulatus in korea. Plant Pathol. J. 26(3), 295. Parke, J., and Grünwald, N. J. 2012. A systems approach for management of pests and pathogens of nursery crops. Plant Disease 96(9): 1236-1244. Patel, J. S., Radetsky, L. C., Nagare, R., & Rea, M. S. (2020). Nighttime Application of UV-C to Control Cucumber Powdery Mildew. Plant Health Progress, 21(1), 40-46. Patto, M. V., Fernández-Aparicio, M., Moral, A., & Rubiales, D. 2007. Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genetic Resources and Crop Evolution, 54(7), 1517-1521. Paul, Y.S., and Thakur, V.K. 2006. Indian Erysiphaceae. Scientific Publishers (India), Jodhpur, 134 pages. Peduto, F., Backup, P., Hand, E.K., Janousek, C.N., Gubler, W.D. 2013. Effect of high temperature and exposure time on Erysiphe necator growth and reproduction: revisions to the UC Davis powdery mildew risk index. Plant Dis. 97: 1438-1447. Perennial Plant Association. 2002. Perennial Plant of the year award. https://perennialplant.org/page/PastWinners Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. 2019. R Core Team. nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1-141, https://CRAN.R- project.org/package=nlme. 111

Pirondi, A., Kitner, M., Iotti, M., Sedláková, B., Lebeda, A., and Collina, M. 2016. Genetic structure and phylogeny of Italian and Czech populations of the cucurbit powdery mildew fungus Golovinomyces orontii inferred by multilocus sequence typing. Plant Pathology, 65(6), 959-967. Plissonneau, C., Hartmann, F. E., & Croll, D. 2018. Pangenome analyses of the wheat pathogen Zymoseptoria tritici reveal the structural basis of a highly plastic eukaryotic genome. BMC Biology, 16(1), 5. Pomerantz, A., Peñafiel, N., Arteaga, A., Bustamante, L., Pichardo, F., Coloma, L. A., Barrio-Amoros, C.L., Salazar-Valenzuela, D., Prost, S. 2018. Real-time DNA barcoding in a rainforest using nanopore sequencing: opportunities for rapid biodiversity assessments and local capacity building. GigaScience 7(4), giy033 RStudio Team. 2016. RStudio: Integrated Development for R. RStudio, Inc., Boston, MA. http://www.rstudio.com/. Rambaut, A., Drummond, A., Xie, D., Baele, G., and Suchard, M. 2018. Posterior summarization in bayesian phylogenetics using tracer 1.7. Systematic Biology,67(5), 901-904. Reddy, K. P., Rao, S. C., Kirti, P. B., and Sujatha, M. 2013. Development of a scoring scale for powdery mildew (Golovinomyces cichoracearum (DC.) VP Heluta) disease and identification of resistance sources in cultivated and wild sunflowers. Euphytica, 190(3), 385-399. Rusanov, V.A., and Bulgakov, T.S. 2008. Powdery mildew fungi of Rostov region. Mikol. Fitopatol. 42: 314-322. Sampson, P.J., and Walker, J. 1982. An Annotated List of Plant Diseases in Tasmania. Department of Agriculture Tasmania, 121 pages. Schmitt, J. A. 1955. The host specialization of Erysiphe cichoracearum from Zinnia, Phlox and cucurbits. Mycologia. 47: 688-701. Schnathorst WC. 1959. Spread and life cycle of the lettuce powdery mildew fungus. Phytopathology 49:464–68 Shaw, C.G. 1973. Host fungus index for the Pacific Northwest - I. Hosts. Washington State Univ. Agric. Exp. Sta. Bull. 765: 1-121. Shin, H.D. 2000. Erysiphaceae of Korea. National Institute of Agricultural Science and Technology, Suwon, Korea, 320 pages; Cho, W.D., and Shin, H.D., Eds. 2004. List of plant diseases in Korea. Fourth edition. Korean Society of Plant Pathology, 779 pages. Simmonds, J.H. 1966. Host index of plant diseases in Queensland. Queensland Department of Primary Industries, Brisbane : 111. Simonyan, S.A. 1981. [Mycoflora of Botanical gardens and arboretums of the Armenian S.S.R.] (Translated from Russian). Publ. House Acad. Sci. Armenian S.S.R., 234 pages;

112

Sombardier, A., Savary, S., Blancard, D., Jolivet, J., and Willocquet, L. 2009. Effects of leaf surface and temperature on monocyclic processes in Podosphaera aphanis, causing powdery mildew of strawberry. Can. J. Plant Pathol. 31:439-448. Sonah, H., Deshmukh, R.K, Belanger, R.R. 2016. Computational prediction of effector proteins in fungi: opportunities and challenges. Front Plant Sci. 7:126 Spanu, P. D., Abbott, J. C., Amselem, J., Burgis, T. A., Soanes, D. M., Stüber, K., van Themaat, A.V.L., Brown, J.K.M., Butcher, S.A., Gurr, S.J., Lebrun, M.H., Ridout, C.J., Schulze-Lefert, P., Talbot, N.J., Ahmadinejad, N., Ametz, C., Barton, G.R., Benjdia, M., Bidzinski, P., Bindschedler, L.V., M., Brewer, M., Cadle-Davidson, L., Cadle-Davidson, M.M.,Collemare, J.,Cramer, R., Frenkel, O., Godfrey, D., Harriman, J., Hoede, C., King, B., Klages, S., Kleemann, J., Knoll, D., Koti, P.S., Kreplak, J., López-Ruiz, F.J., Lu, X., Maekawa, T., Mahanil, S., Micali, C., Milgroom, M.G., Montana, G., Noir, S., O’Connell, R.J., Oberhaensli, S., Parlange, F., Pedersen, C., Quesneville, H., Reinhardt, R., Rott, M., Sacristán, S., Schmidt, S. M., Schön4, M., Skamnioti, P., Sommer, H., Stephens, A., Takahara, H., Thordal-Christensen, H., Vigouroux, M., Weßling, R., Wicker, T., Panstruga, R. 2010. Genome expansion and gene loss in powdery mildew fungi reveal tradeoffs in extreme parasitism. Science, 330(6010), 1543-1546. Stadnik, M.J. and W. Bettiol. 2001. Pulverizačáo com leite estimula a microflora do filoplano e reduz a severidade do oídio do pepino. Summa Phytopathol., 27(1): 109. Standley, K. 2013. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Molecular Biology and Evolution 30:772-780 Stanke, M., Diekhans, M., Baertsch, R., Haussler, D. 2008. Using native and syntenically mapped cDNA alignments to improve de novo gene finding. Bioinformatics 24(5): 637-644 Sudisha, J., A. Kumar, K. N. Amruthesh, S. R.Niranjana and H.S.Shetty. 2011. Elicitation of resistance and defense related enzymes by raw cow milk and amino acids in pearl millet against downy mildew disease caused by Sclerospora graminicola. Crop Prot., 3 Surlan-Momirovic, G., Flath, K., Silvar, C., Brankovic, G., Kopahnke, D., Knezevic, D., Schliephake, E., Ordon, F., Perovic, D. 2016. Exploring the Serbian genbank barley (Hordeum valugare L. subsp. Vulgare) collection for powdery mildew resistance. Gen. Ress. And Crop Ev. 69(2):275-287. Suthaparan, A., Solhaug, K. A., Bjugstad, N., Gislerød, H. R., Gadoury, D. M., & Stensvand, A. 2016. Suppression of powdery mildews by UV-B: application frequency and timing, dose, reflectance, and automation. Plant disease, 100(8), 1643- 1650. Tajima, F. 1983. Evolutionary relationship of DNA sequences in finite populations. Genetics, 105: 437-460.

113

Tajima, F. 1989. Statistical methods to test for nucleotide mutation hypothesis by DNA polymorphism. Genetics, 123:585-595. Takamatsu, S. Matsuda, S. 2013. Comprehensive phylogenetic analysis of the genus Golovinomyces (Ascomycota: Erysiphales) reveals close evolutionary relationships with its host plants. Mycologia. 105: 1135-1152. Takamatsu, S., Matsuda, S., Niinomi, S., Havrylenko, M. 2006. Molecular phylogeny supports a Northern Hemisphere of Golovinomyces (Ascomycota: Erysiphales). Mycological Research: 110: 1093-1101. Talieva, M.N, Kondrat’eva, V.V. 2002. Influence of exogenous salicylic acid on the level of phytohormones in tissues of Phlox paniculata and Phlox setacea leaves with special reference to resistance against the powdery mildew causative agent Erysiphe cichoracearum DC. f. phlogis Jacz. Biology Bulletin. 29: 551-554. Thompson, H. S, Svejda, F. J. 1964. Resistance to powdery mildew in perennial Phlox. Can. J. Plant Sci. 45: 258-263. Tobias, A.V., and Tikhomirova, I.N. 1998. Powdery mildew in St. Petersburg. Mikol. Fitopatol. 32(2): 45-47 Turner, B.L. 1998. Atlas of the Texas species of Phlox (Polemoniaceae). Phytologia 85: 309-326. Vaz Patto, M. V., Fern´andez-Aparicio, M., Moral, A., and Rubiales, D. 2007. Resistance reaction to powdery mildew (Erysiphe pisi) in a germplasm collection of Lathyrus cicera from Iberian origin. Genet. Resour. Crop Evol. 54:1517-1521. Venables, W.N., and Ripley, B.D. 2002. Modern applied statistics with S. 4th Edition. Springer-Verlag. Voytyuk, S.O., Heluta, V.P., Wasser, S.P., and Nevo, E. 2006. Genus Golovinomyces (Erysiphales) in Israel: species composition, host range, and distribution. Mycol. Balcan. 3: 131-142. Walsh, P. S., Metzger, D. A., and Higuchi, R. 1991. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques, 10(4): 506-513. Waterhouse, R. M., Seppey, M., Simão, F. A., Manni, M., Ioannidis, P., Klioutchnikov, G., Kriventseva, V., Zdobnov, E. M. 2017. BUSCO applications from quality assessments to gene prediction and phylogenomics. Molecular biology and evolution 35(3):543-548. Watterson, G. A. 1975. On the number of segregating sites in genetical models without recombination. Theoretical Population Biology, 7(2), 256-276. Wherry, E.T. 1955. The Genus Phlox. Morris Aboretum Monographs Philadelphia, PA. White, T.J., Bruns, T.D., Lee, S., Taylor, J. 1990. Amplification and direct sequencing of fungal ribosomal genes for phylogenetics. In: Innis, M.A., Gelfand, D.H., Sninsky,

114

J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Applications. Academic Press, San Diego, CA, pp. 315–322. Whiteside, J.O. 1966. A revised list of plant diseases in Rhodesia. Kirkia 5: 87-196. Wicker, T., Oberhaensli, S., Parlange, F., Buchmann, J. P., Shatalina, M., Roffler, S., Ben- David, R., Doležel, J., Šimková, H., Schulze-Lefert, P., Spanu, P.D., Bruggmann, R., Amselem, J., Quesneville, H., van Themaat, E.V.L., Paape, T., Shimizu, K.K, Keller, B.2013. The wheat powdery mildew genome shows the unique evolution of an obligate biotroph. Nature Genetics, 45(9), 1092. Willocquet, L., Colombet, D., Rougier, M., Fargues, J., Clerjeau, M. 1996. Effects of radiation, especially ultraviolet B, on conidial germination and mycelial growth of grape powdery mildew. Europ. J. of Plant Pathol. 102: 441-449. Wu, Y., Ma, X., Pan, Z., Kale, Y.S., Song, Y., King, H., Zhang, Q., Presley, C., Deng, X., Wei, C-I and Xia, S. 2018. Comparative genome analyses reveal sequence features reflecting distinct modes of host-adaptation between dicot and monocot powdery mildew. BMC 19:705 Zale, P., and Jourdan, P. 2015. Genome size and ploidy and Phlox paniculata and related germplasm in subsections Paniculatae and Phlox. J. Amer. Soc. Hort. Sci. 140(5):436- 448. Zale, P., Robarts, D.H.W., Jourdan, P. 2016. Genome size and ploidy levels of creeping phlox and related germplasm of mat-forming taxa from eastern and western North America. Sc. Hort. 203: 53-61 Zale, P.J. 2014. Germplasm collection, characterization, and enhancement of eastern Phlox species. PhD Dissertation, The Ohio State University. Zale, P.J., Robarts, D.W.H., Jourdan, P. 2016. Genome size and ploidy levels of creeping phlox and relatedgermplasm of mat-forming taxa from eastern and western NorthAmerica. Sc. Hort. 203: 53-61.

115

Appendix A. Supplementary Material for Chapter 2

Supplementary table A.1: Summary statistics of the effect of age of pathogen culture on the incidence of powdery mildew infection, the size of the fungal colonies, and the number of spores per unit of area produced by Golovinomyces magnicellulatus and Podosphaera sp. on detached Phlox paniculata leaves.

Colony size Incidencea (mm) Spore/mm2 b Factor Exp. Pathogen P tested run P F P value F F valu value e

Golovinomyces 1 3.95 0.03 3.465 0.651 2.39 0.11 magnicellulatus 2 0.45 0.45 0.3 0.05 0.435 0.65 Age of c 0.00 culture 1 21.12 <0.001 1.9 0.17 6.25 Podosphaera 6 sp. 2 18.54 <0.001 1.004 0.38 0.152 0.86

aANOVA analysis performed on Arcsine-transformed data. bANOVA analysis for Podosphaera sp. performed on log-transformed data cAge of pathogen culture tested were 14, 18 and 22-day old.

116

Supplementary table A.2: Summary statistics of the effect of the phenology of the leaf tissue and side of the leaf surface on the incidence of powdery mildew infection, the size of the fungal colonies, and the number of spores per unit of area produced by G. magnicellulatus and Podosphaera sp. on detached P. paniculata leaves.

Colony size Incidenceb (mm) Spore/mm2 c

Exp. F P F P F P Pathogen Run Factora value value value

NODE 4.45 0.02 0.371 0.69 0.99 0.39 1 SIDE 11.32 0.002 2.55 0.13 6.1 0.02 G. NODE x SIDE 3.27 0.05 1.45 0.26 0.8 0.46 magnicell ulatus NODE 8.46 0.001 2.77 0.09 4.36 0.03 2 SIDE 3.31 0.08 5.05 0.04 20.02 0.003 NODE x SIDE 0.1 0.9 2.1 0.15 2.26 0.13

NODE 7.027 0.003 5.1 0.01 3.19 0.06

1 SIDE 0.15 0.7 42.35 <0.001 22.37 <0.001

NODE x SIDE 5.644 0.008 1.51 0.24 1.76 0.19

NODE 5.21 0.01 6.5 0.008 4.16 0.045 Podosph 2 SIDE 3.54 0.07 3.65 0.07 8.65 0.01 aera sp. NODE x SIDE 0.74 0.49 0.19 0.83 2.76 0.11

NODE 15.26 <0.001 5.78 0.009 5.45 0.01

3 SIDE 0.92 0.35 1.27 0.27 5.39 0.03

NODE x SIDE 0.61 0.55 0.1 0.9 0.5 0.61

aFactors tested: three phenological stages of leaf development (1st, 3rd, 5th node leaves) and two surfaces of the leaf (abaxial and adaxial). bANOVA analysis performed on Arcsine-transformed data. cANOVA analysis performed on log-transformed data.

117

Appendix B. Supplementary Material for Chapter 4

Supplementary table B.1: Gene sequences used in the MLST analysis and correspondent GenBank accession numbers.

Locus* TUB ITS IGS TEF-1α H3** CSI** Isolate MN830 MN83048 MS1.1 MN822586 MN830815 MN830483 MN830514 481 2

MS1.2 MN822587 MN830816 MN830484 MN830515 OH1 MN822588 MN830817 MN830485 MN830516 OH2 MN822589 MN830818 MN830486 MN830517 OH3.1 MN822590 MN830819 MN830487 MN830518 OH3.2 MN822591 MN830820 MN830488 MN830519 OH4.1 MN822592 MN830821 MN830489 MN830520 OH4.2 MN822593 MN830822 MN830490 MN830521 OH4.3 MN822594 MN830823 MN830491 MN830522 OH5 MN822595 MN830824 MN830492 MN830523 OH6.1 MN822596 MN830825 MN830493 MN830524 OH6.2 MN822597 MN830826 MN830494 MN830525 NY1 MN822598 MN830827 MN830495 MN830526 NY2 MN822599 MN830828 MN830496 MN830527 NY3.1 MN822600 MN830829 MN830497 MN830528 NY3.2 MN822601 MN830830 MN830498 MN830529 NY4.1 MN822602 MN830831 MN830499 MN830530 NY4.2 MN822603 MN830832 MN830500 MN830531

Continued

118

Table B.1 Continued Locus* Isolate TUB ITS IGS TEF-1α H3** CSI** NY5.1 MN822604 MN830833 MN830501 MN830532 NY5.2 MN822605 MN830834 MN830502 MN830533 NY6.1 MN822606 MN830835 MN830503 MN830534 NY6.2 MN822607 MN830836 MN830504 MN830535 MI1 MN822608 MN830837 MN830505 MN830536 MI2 MN822609 MN830838 MN830506 MN830537 MI3 MN822610 MN830839 MN830507 MN830538 IA1 MN822611 MN830840 MN830508 MN830539 WI1 MN822612 MN830841 MN830509 MN830540 TN1.1 MN822613 MN830842 MN830510 MN830541 TN1.2 MN822614 MN830843 MN830511 MN830542

* TEF-1α = translation elongation factor alpha; TUB = β-tubulin; ITS = internal transcriber spacer; IGS = partial intergenic spacer; CSI= chitin synthase I; H3 = histone. **Due to 100% sequence identity among all isolates, only one representative sequence for loci CSI and H3 were deposited in GenBank.

119

Supplementary table B.2: Summary statistics of the effect of PM isolate (n=10) and Phlox species (n=4) on disease progression, spore production and latency period observed on Phlox plantlets.

AUDPCa Spore productionb Latency periodc Exp. Factor Run F value P value F value P value F value P value

PM isolate 12.67 <0.001 5.9 <0.001 8.3 <0.001

Phlox species 14.6 <0.001 13.1 <0.001 13.6 <0.001 1 PM isolate × 4.2 <0.001 1.2 0.25 3.2 <0.001 Phlox species

PM isolate 12.7 <0.001 5.2 <0.001 3.8 <0.001

Phlox species 14.6 <0.001 6.5 <0.001 13.1 <0.001 2 PM isolate × 4.3 <0.001 3.0 <0.001 3.04 <0.001 Phlox species

a AUDPC = Area under the disease progress curve. Values were calculated from disease severity data collected for 28 days post-inoculation. b Amount of spore produced per mg of plant dry weight 28 days post inoculation. Data were log-transformed prior to ANOVA analysis. c Latency period corresponds to the number of days between inoculation and sporulation.

120