Grape Powdery Mildew Management: the Interaction Between Inflorescence Stage and Fungicide Chemistry

AN ABSTRACT OF THE THESIS OF

Brent W. Warneke for the degree of Master of Science in Botany and Plant Pathology presented on March 7, 2018.

Title: Grape Powdery Mildew Management: The Interaction Between Inflorescence Stage and Fungicide Chemistry.

Abstract approved: ______Walt F. Mahaffee

Grape powdery mildew (GPM, causal agent Erysiphe necator) is the most economically important disease of grapevine in the Western U.S. Low levels of GPM infection on clusters (1-5%) can negatively affect wine sensory quality, so intensive fungicide programs are common. This may lead to problems like excessive fungicide use and buildup of resistant E. necator strains. Fungicide redistribution could improve fungicide protection of grape clusters, but has not been investigated for many products.

Many fungicides effectively control GPM cluster infections, but optimal timing regimes for specific products remain poorly characterized. Increased knowledge of fungicide redistribution and application timing may alleviate the need for narrow application windows during bloom.

Bioassays determined that all fungicides tested could redistribute by some mechanism, with xylem and translaminar mechanisms being common. Some fungicides performed as expected, while others redistributed at a lower level or via mechanisms not previously demonstrated, such as fluopyram redistributing via vapors. Vapor

redistribution appeared to protect clusters in the field and three fungicides could translocate through flower caps.

In small plot trials, most fungicide treatments timed to the end of bloom provided significantly better GPM control on clusters than non-treated and wettable sulfur control treatments (P<0.01). The application of two sequential treatments with a 14-day interval of either trifloxystrobin, quinoxyfen, or fluopyram at different bloom stages showed that applications initiated at end of bloom resulted in the lowest respective berry infection probabilities of 0.073, 0.097, and 0.020. Rotations between quinoxyfen and fluopyram initiated at end of bloom resulted in berry infection probabilities of 0.059 when quinoxyfen was applied first, and 0.076 when fluopyram was applied first.

After industry reported GPM control failures when using quinone outside inhibitor (QoI) products, a survey was conducted to determine if QoI resistant strains were present in Oregon viticultural regions. Survey results using a competitive TaqMan qPCR assay showed that 62% of field samples had populations of E. necator carrying the

G143A mutation, and it was present in all three regions sampled. The qPCR assay determined that 87% of isolates contained the G143A mutation and phenotypic characterization validated that isolates containing G143A were resistant and could withstand high concentrations of two QoI products (EC50 > 100µg/ml).

This research showed that an improved understanding of fungicide redistribution could improve disease management recommendations, especially if crop phenology is also considered. Integrating two carefully timed applications of redistributing fungicides initiated at end of bloom into a fungicide program may be an effective strategy for wine grape growers in western Oregon seeking to produce fruit with low GPM infection levels

and minimum synthetic chemical input. This strategy would have the added benefit of reducing selection for fungicide resistant strains of E. necator.

Grape Powdery Mildew Management: The Interaction Between Inflorescence Stage and Fungicide Chemistry

by Brent W. Warneke

A THESIS

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Master of Science

Presented March 7, 2018 Commencement June 2018

Master of Science thesis of Brent W. Warneke presented on March 7, 2018

APPROVED:

Major Professor, representing Botany and Plant Pathology

Head of the Department of Botany and Plant Pathology

Dean of the Graduate School

I understand that my thesis will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my thesis to any reader upon request.

Brent W. Warneke, Author

ACKNOWLEDGEMENTS

I thank Walt Mahaffee for his guidance, encouragement, and for keeping his door open for anything from music recommendations to metaphysical discussions. I could not have found a lab that better fit my personality, personal values, and research interests.

I thank the other members of my MS committee: Amy Dreves for her practical approach to problems and the beneficial discussions about research and agriculture; Jay Pscheidt for his teaching and sharing knowledge of the grape industry; Shawn Mehlenbacher for taking the time to act as my graduate school representative.

I thank Tara Neill for her constantly good attitude and excellent job managing the Foliar

Pathology Laboratory. I would also like to thank Tara for all the guidance and help with vineyard maintenance and her help in gathering data in the phenological timing work.

I thank all other members of the Foliar Pathology laboratory: Bailey Williams, Carly

Allen, Katelynn Thrall, Chris Gorman, Andy Albrecht, Sarah Lowder, for the help in gathering data, general vineyard maintenance, and their general good attitudes that made the Foliar Pathology laboratory culture so great.

I thank Andy Albrecht and Steve Castagnoli for their help in obtaining samples in the

2015 QoI resistance survey.

I thank Francisco Mauro (Paco) for the statistical consultation with R, and the indispensable camaraderie through my time in Corvallis.

I thank my family members for the emotional support throughout my pursuit of an MS degree, and my brother Christopher, for providing consultation on statistics and academic dynamics.

I thank all my friends in Corvallis for enriching my experience at OSU and providing a local support system which has been crucial to upkeep of my morale throughout pursuit of this degree.

Finally, I thank the Oregon Wine Board for funding my research and all the members of the USDA ARS Horticultural Crops Research Lab and OSU Department of Botany and

Plant Pathology who have supported me throughout classes and have helped make a great work environment. I am also grateful to the grape industry cooperators that participated in this research.

CONTRIBUTION OF AUTHORS

Brent Warneke designed, optimized, conducted, analyzed the experiments and wrote the manuscript of Chapter 2. Lindsey Thiessen and Walt Mahaffee conceived of the phenological fungicide experiments, obtained funding, assisted with experiment design and analysis and edited the manuscript of Chapter 2.

Walt Mahaffee, Brent Warneke, and Tara Neill designed and optimized the conidia germination bioassays in Chapter 3. Brent Warneke and Tara Neill wrote the portion of the manuscript presented here (except Table 1). Brent Warneke collected the field samples, maintained the isolates, extracted DNA from them, and fitted models to calculate EC50s. Tim Miles and Jesse Yamagata performed the TaqMan qPCR assay on isolate and field sample DNA from Oregon.

TABLE OF CONTENTS

Page

Chapter 1. General Introduction ...... 1

Fungicide Development and Use ...... 2 Fungicide Resistance ...... 6 Fungicide Redistribution ...... 14 FRAC Groups Used to Manage Grape Powdery Mildew ...... 17

Chapter 2. The Grape Powdery Mildew Conundrum – Fungicide Selection and Application Timing ...... 29

Introduction ...... 30 Materials and Methods ...... 33 Results ...... 45 Discussion ...... 55 References ...... 62

Chapter 3. Detection of Quinone Outside Inhibitor Resistant Isolates of Erysiphe necator in Oregon Vineyards ...... 65

Introduction ...... 67 Materials and Methods ...... 69 Results ...... 75 Discussion ...... 76 References ...... 80

Chapter 4. General Conclusions ...... 83

Bibliography ...... 87

Appendices ...... 99

LIST OF FIGURES

Figure Page

1.1. Technologies and disciplines used in the development of fungicides ………..…4

1.2. Illustration of sites of action of FRAC groups...……………...... ……………..…5

1.3. Examples of mock fungicide labels ………………………....………………..…6

1.4. Fungicide resistance selection illustration…………………………….…....…....7

1.5. Resistance mechanism illustrations.………………………….……..……….…..8

1.6. Graphical illustration of the types of fungicide resistance ………………….….9

1.7. Fungicide adsorption and absorption illustration ………………...…..……..…14

1.8. Illustration of fungicide redistribution mechanisms …………………...…...….16

2.1 Detached leaf setup and data collection illustration. ………………………….....35

2.2. Growth stages selected to investigate in phenological timing experiment ...…..38

2.3. Timeline of approximate fungicide application schedule and data collections in field experiments ………………….…...………………………………...... …41

2.4 Mean proportions of infected berries by treatment in fungicide through calyptra study. …………………………………………………………………………….…..46

2.5 AUDPC values for phenologically timed fungicide treatments in 2015 and 2016……………………………………………………………………….………….48

2.6 Estimated odds ratios of timing comparisons separated by fungicide …….….…49

2.7 Probabilities of berry infection in 2015 and 2016 …………...…………..….…..51

2.8 Probabilities of berry infection in the field redistribution assessment ...….…..…53

2.9 Mean AUDPC values by treatment from the 2017 fungicide rotation experiment …………………...…………………………………………….……..….54

2.10 Probabilities of berry infection in the 2017 rotation experiment ……...... ……..54

3.1 Approximate E. necator sample collection locations in Oregon ……………..…69

LIST OF FIGURES (Continued)

Figure Page

3.2 Two-tier petri dish detached leaf chambers and isolate spore settling tower ………………………………………………………………………...... ……71

3.3 Determination of genotype using the TaqMan assay …………..………….……76

LIST OF TABLES

Table Page

1.1. Fungicides registered for powdery mildew control in OR and their attributes.... 28

2.1. Fungicides and rates used in small plot and all fungicide redistribution assays. ………………………………………………………………………………………………. 34

2.2 Mean areas of inhibition from detached leaf assays ……………….…………... 45

3.1 Primers and probes utilized in the detection of the G143A mutation associated with QoI resistance ………………………………………………………………… 74

3.2 Origin and QoI sensitivity of Oregon Erysiphe necator isolates collected in 2015……………………………………………………………………………….… 75

LIST OF APPENDICES

Appendix Page

Appendix A: 2017 Commercial Phenological Application Trial……………..……100

Appendix B: Minimum Inhibitory Concentration Experiment…………………..…108

Appendix C: Oregon Wine Research Institute Technical Newsletter Article….…..111

LIST OF APPENDIX FIGURES

Figure Page

A1. Maps of commercial trial sites ……………………………..………..……..….100

A2. Inflorescence phenology at the time of experimental applications………….…102

A3. Berry infection proportion comparisons at vineyard 2…………………...... …104

A4. Probabilities of berry infection at collection points in the trial area at vineyard 2………………………………………………………………………….………….105

B1. Fungicide application points and diameter measurements illustration …..……109

C1. Graphics and pictures of grapevine flowering stages …..……………….….…111

C2. Timeline of fungicide timing experiment …..……………………………..…..113

C3. Area under disease progress curves by treatment for 2015 & 2016 …..…...…114

C4. Mean probabilities of berry infection by treatment for 2015 & 2016 …..…..…115

LIST OF APPENDIX TABLES

Table Page

B1. Fungicides and rates used in the minimum inhibitory concentration study.…………………………………………………………...……………..…….108

B2. Coefficients of determination and minimum inhibitory concentration values…110

C1. Selected fungicides, mechanisms and application rates………………………..112

Chapter 1

A Primer on Fungicides Used in Grape Production for Powdery Mildew Management

Brent W. Warneke

Foreword

This introduction was written as a standalone extension-type document providing general information about fungicide resistance and fungicide groups used to control grape powdery mildew (Erysiphe necator). The intended audience of this document was growers, consultants, and participants in the wine grape industry. It is organized into two parts: 1) General sections focusing on [i] fungicide use and development, [ii] fungicide resistance, [iii] fungicide redistribution; 2) An overview of fungicides used to control grape powdery mildew separated by FRAC group. Each FRAC group section discusses the background of the group, mode of action and unique attributes such as redistribution profile, as well as resistance status at the time of this writing.

Fungicide use in grape production is thought to account for 95% of potential yield by protecting against primarily powdery and downy mildew (Gianessi and Reigner,

2005). Thus, grapes are the most fungicide intensive crop produced in the United States, with a combined 45 million pounds of fungicides applied at a cost of $123 million annually (Gianessi and Reigner, 2005). In California alone, 35 million pounds of fungicide are applied annually at a product cost of $70.5 million and an estimated application cost of between $110 – 201 million (Sambucci et al., 2014). Heavy reliance on fungicides in the grape industry has led to pathogenic fungi developing resistance to commonly used fungicide products (Brent and Hollomon, 1995; Gianessi and Reigner,

2005) and puts pressure on pesticide producers and grape growers alike as they seek to control grape fungal diseases. Improved knowledge of fungicide attributes such as mode of action and redistribution profiles can improve their use, and could result in better disease control. Understanding the factors involved in resistance development to fungicides can facilitate their use in a manner that mitigates selection pressures on development of resistant fungal strains and could increase their useful life.

Fungicide development The active ingredients in modern fungicides are typically small synthetic molecules that interfere with one or multiple biochemical processes of the target organisms (Morton and Staub, 2008). In the past, new molecules were discovered through biological screening, but with increasing knowledge from bioinformatics techniques, compounds are being designed to inhibit specific biochemical pathways

(Klausener et al., 2007).

Modern fungicide development has become a highly interdisciplinary process that uses age old techniques and the newest biotechnologies to increase the efficiency of the drug discovery process (Figure 1.1). The many steps involved make this an extremely expensive process: from Figure 1.1 Technologies and disciplines used in the discovery of a single active development of fungicides (adapted from Klausener et al 2007). ingredient to its first market sales is estimated to cost around $286 Million (Philips McDougall, 2016). With such a large initial investment and time commitment (approx. 10 years) required, only fungicides that have the potential to serve large markets and thus, recoup the research and development costs within the remaining patent period after commercialization (Russell, 2006) are pursued. For example, when developing fungicides to control GPM, chemical companies will often target compound development and screening towards larger markets such as wheat powdery mildew. Once the larger market has been established, then other crop registrations are pursued (Russell, 2006). However, the cost of additional regulatory reviews and packaging updates can lead chemical companies to the conclusion that registration is not warranted for smaller markets, even if the product is expected to be effective (Russell, 2006). These constraints lead to fewer new fungicidal products being registered for horticultural markets (such as viticulture) and subsequently, reliance on

older fungicide groups, and the increased potential for resistance development to those groups.

Fungicide groups Most fungicides used on grapes inhibit cellular processes critical to pathogen growth, such as cellular respiration or microtubule synthesis, by binding with an enzyme or other cellular component and interfering with its natural functioning (Figure 1.2; Latin

2011; Morton and Staub, 2008). This requires the fungicide to be present at concentrations high enough to outcompete the natural substrate. The specific process that the chemical inhibits in the fungus is referred to as its mode (or mechanism) of action.

Figure 1.2 Illustration of sites of action in a cell where FRAC groups used to manage grape powdery mildew act. Adapted from Latin (2011). Based on their mode of action, fungicides can be broadly grouped into two categories: single site and multi-site. Single site fungicides inhibit a single cellular process, typically an enzyme. Most modern synthetic fungicides are single site products.

Multi-site fungicides inhibit multiple cellular processes simultaneously. Most multi-site

fungicides are older products that require high application rates and good coverage to be effective.

Chemical companies formed a trade group called the Fungicide Resistance Action

Committee (FRAC) that organizes fungicides into groups based on their mode of action across fungicide products. Mode of action groups (referred to as FRAC groups) are assigned a code which fungicide manufacturers display on their labels (Figure 1.3). These codes make it easy to distinguish between modes of action and facilitate implementing resistance management strategies such as tank mixing or rotating. All of this is done with the goal of slowing the development of fungicide resistance and prolonging the life span of fungicide active ingredients.

Figure 1.3. Examples of mock fungicide labels. FRAC groups are indicated in the black portion of the fungicide group band. Ful Crop is an example of a fungicide with single fungicide group. Folmaster is an example of a combination product incorporating two fungicide groups in one product.

Fungicide resistance

Modern fungicides are prone to fungicide resistance development due to their single site mode of action, high potency, and longer residual time on the crop; all of which contribute to the development of resistant fungal strains (Brent and Hollomon,

1995). Fungicide resistant pathogen populations are the result of natural mutations caused

by imperfections in the DNA replication process or UV radiation (Brent and Hollomon,

1995). These mutations in DNA affect protein form and function that can result in

Figure 1.4. Illustration of selection for resistant individuals with fungicide applications. Each successive fungicide application will inhibit sensitive individuals while not affecting resistant individuals. Circles on leaves represent fungal colonies with blue sensitive individuals and red resistant individuals. Adapted from Bradley et al. (2016). reduced fungicide binding by the protein and resulting loss of efficacy. When a fungicide application is made to a pathogen population, the susceptible individuals will be inhibited, while the resistant individuals will grow and reproduce with little or no inhibition (Figure 1.4; Bradley et al., 2016). As this process repeats, the proportion of resistant individuals increases.

Mechanisms of fungicide resistance

Alteration of fungicide target site. The most common way that fungi become resistant to fungicides is by random genetic mutations occurring in genes which code for the protein affected by the fungicide (Brent and Hollomon, 1995, Figure 1.5A). These mutations result in conformational changes in the protein that eliminates or reduces the ability of the fungicide to interact with the protein.

Detoxification. After uptake of the fungicide into a fungal cell, cellular enzymes can metabolize the fungicide into a less toxic byproduct (Figure 1.5B). This renders the fungicide less effective at its mechanism of action, resulting in decreased control of the

pathogen (Brent and Hollomon, 1995).

Detoxification must occur before the fungicide can

reach the site of its activity in the cell.

Overexpression. Many fungicides compete with the

natural substrate of a fungal enzyme (Figure 1.5C).

There are multiple methods by which fungi can over-

express enzymes and dilute the effect of the

fungicide (Leroux et al., 2002; Ma et al., 2006).

Mobile genetic elements can increase gene copy

numbers encoding for an enzyme or mutations in a

gene promotor can increase the expression of the

enzyme. Higher doses of fungicide may still be

effective, but are typically not practical in field Figure 1.5. Resistance mechanism illustrations. (A) Alteration of binding site, (B) detoxification, (C) conditions because of residue concerns and label overexpression of the target, (D) removal or reduced uptake. restrictions.

Removal or reduced uptake. The fungal cell wall is covered with protein structures called efflux pumps that span cell walls and transport materials into and/or out of cells.

Genetic mutations can affect genes or promotors encoding for these pump proteins which could affect the rate of fungicide uptake or export from cell cytoplasm, resulting in lower fungicide efficacy (Deising et al., 2008, Figure 1.5D).

Types of fungicide resistance

Qualitative. Qualitative fungicide resistance is the most common type of resistance and is typically observed as a rapid loss of disease control in crops that have been treated

(Figure 1.6). A pathogen gains qualitative resistance when a gene encoding the target protein of a fungicide is mutated, altering the fungicide target site and making the resulting protein unaffected by the fungicide (Figure 1.5A, Brent and Hollomon, 1995).

Resistant individuals can typically withstand concentrations higher than can be legally applied. Qualitative resistance can be inherited by the offspring of resistant individuals and can remain in populations for years after fungicide applications are ceased (Deising et al., 2008).

Figure 1.6. Graphical illustration of the types of fungicide resistance. The curves in the graph represent pathogen populations with rightward movement showing a change from a susceptible to a resistant population due to fungicide selecting for resistant individuals to reproduce. Adapted from Brent and Hollomon, (1995).

Quantitative. Quantitative fungicide resistance (also known as partial resistance) is a gradual increase in tolerance to higher and higher concentrations of a fungicide (Figure

1.6). Quantitative resistance usually involves multiple mechanisms contributing to keeping the concentration of fungicide, relative to the target site, in a cell low enough for it to function. Mechanisms often work in tandem and include efflux of fungicide out of a cell, degradation of the fungicide by enzymes, and overexpression of the target protein.

When a fungicide is applied, individuals possessing these fungicide mitigation mechanisms could survive and sub-lethal fungicide stress may further stimulate the development of those mechanisms (Deising et al., 2008). In this way, repeated applications of fungicides in a FRAC group will shift a pathogen population from one that is sensitive to that of one that is resistant (Deising et al., 2008).

Another consideration: Cross-resistance. When a pathogen becomes resistant to a fungicide it typically becomes “cross-resistant” to all fungicides in that FRAC group.

Often, resistance involves alteration of the target site, preventing the entire FRAC group from binding. In resistance cases mediated by other mechanisms, similar concepts apply where the resistance mechanisms accumulated work across products within the FRAC group, although sometimes to varying effect. There are cases where pathogens resistant to one fungicide in a FRAC group can be controlled by another fungicide in the same FRAC group, but this is not the norm (e.g. FRAC 7, Ishii et al., (2011)).

Factors influencing fungicide resistance

Fungicide resistance is a complex arms race between fungi evolving resistances, farmers trying to control the fungi, and chemical manufacturers introducing new products.

Fungicide groups and pathogens vary widely and both have factors that influence the rate at which fungicide resistance develops in this complex system.

Fungicide factors

There are a wide range of fungicidal products on the market, each with its own characteristics. The fungicide site of action, dose, frequency of application, and disease status at the time of application are three factors that have the most influence on the risk of resistance development to a fungicide group (Bradley et al., 2016; Grimmer et al.,

2015).

Fungicides with single-site modes of action are generally at higher risk of resistance development than multi-site products. There is a higher probability that a mutation in a single DNA region will affect the activity of a single-site fungicide, than the simultaneous occurrence of multiple mutations in all regions that are involved in a fungicide with multiple mechanisms of action. As the number of sites a fungicide affects goes up, the probability that an individual will incur enough mutations to confer resistance goes down.

The dose of fungicide received by pathogens affects the rate of resistance development. In general, applying reduced rates of fungicides allows more of a pathogen population to survive, which can lead to some of those individuals becoming less sensitive to a fungicide product (Brent and Hollomon, 1995; Deising et al., 2008). Over

time this could lead to a quantitative resistance response. However, a lower dose may decrease selection for qualitative resistance through larger numbers of sensitive isolates surviving and reproducing than would happen at higher rates (Brent and Hollomon,

1995). However, research and modeling also indicates that lower doses can increase the rate of resistance development (Bosch et al., 2014). At registered label rates, the variability of fungicide deposition across plant tissues in field scenarios results in variable doses of fungicide to pathogen populations, so applying label rates is recommended to generally mitigate selection pressures on both qualitative and quantitative resistance types.

The chance of fungi developing resistance to a fungicide group increases when it is applied more than once in a growing season (Brent and Hollomon, 1995). The risk of resistance development to that a fungicide group is higher still when it is applied to the same area year after year.

Curative fungicide applications are made after the pathogen is established on a crop, with the goal of impacting the fungal life cycle in various ways, such as reducing sporulation rate or increasing the time it takes to sporulate (latent period; Bosch et al.,

2014). Curative applications are typically made when there are relatively large pathogen populations (i.e. later in the season) so due to the larger pathogen population size, there is a greater chance some of those individuals present may contain a mutation making them less sensitive to the fungicide (Bosch et al., 2014). Thus, curative fungicide applications are discouraged because the selection pressure they exert for resistant fungal strains to succeed.

Pathogen factors

Pathogen populations have a wide variety of characteristics that allow them to adapt to environmental stresses. The generation time, genetic diversity of a population, and the fecundity of the organism are characteristics that influence adaptation to stresses such as fungicide applications (Grimmer et al., 2015).

The generation time is how long it takes from when a spore lands on susceptible tissue until it infects and produces another spore. The shorter the generation time, the more likely a pathogen population is to develop fungicide resistance because more individuals will be produced over the course of a season, increasing the chance of mutations occurring that effect fungicide activity. While under selection pressure, having a short generation time speeds up selection for pathogen strains with fungicide resistance.

Variation in genetic characteristics present in a population, referred to as genetic diversity, is important in the resiliency of a species to changes in living conditions. Given a selection pressure such as fungicide applications, pathogen populations with high genetic diversity will be more adaptable than populations with low genetic diversity. A common way that genetic diversity is increased is through sexual reproduction. During sexual reproduction, genes are rearranged and passed on to offspring, resulting in increased genetic diversity in a population. GPM populations on the west coast of the

USA have lower genetic diversity than populations on the east coast of the USA, but populations in both regions regularly reproduce sexually (Brewer and Milgroom, 2010).

Fecundity refers to the degree to which an organism reproduces. When fungal colonies can produce hundreds or thousands of spores every day, this increases the chance that some of those spores will have a natural mutation conferring resistance and

decreases the time it takes for a mutant to become a dominant part of the population

(Brent and Hollomon, 1995).

Fungicide redistribution

Fungicide redistribution in plants is the movement of fungicide from the point of deposition on plant tissue to a different location. It is impossible to achieve 100% plant coverage with field applications of fungicides, so redistribution can aid in fungicide activity (Klittich, 2014). There are multiple mechanisms (such as xylem, apoplast, or by vapors) by which a fungicide can be redistributed (Klittich,

2014). Most fungicidal products that can redistribute will be Figure 1.7. Illustration of Fungicide adsorption and absorption. Red spheres represent fungicide molecules. able to move via multiple mechanisms at the same time. The rate and degree of redistribution is influenced by fungicide formulation, plant characteristics, and environmental factors.

Once fungicides land on plant tissues, they are first adsorbed to the cuticular waxes by intermolecular forces (Figure 1.7). Fungicides like sulfur, copper, and bicarbonate products can only undergo adsorption and thus rely on intermolecular forces to retain their residues on plant tissue. These fungicides can be moved around by precipitation and need to be periodically applied to cover new growth. After adsorption,

many modern synthetic fungicides can be absorbed into tissues by passing through the cuticle and epidermis. Once absorbed, these fungicides are less susceptible to being washed off by rain and can then translocate through plant tissues and be redistributed to other areas of the plant.

Redistribution mechanisms

Xylem. After absorption, fungicides can diffuse into the xylem vessels and be transported acropetally (upward or outward) in leaf and stem tissue from the point of application (Figure 1.8A). Xylem redistribution is the most common form of fungicide

translocation and is commonly referred to as

being “systemic”. This type of redistribution

can help protect growing tissues from

infection when fungicide redistributes into

new tissues, but through this process it can

dilute out the fungicide concentration so that

it is no longer effective.

Translaminar. Translaminar redistribution

occurs when the fungicide moves from one

side of a leaf to the other side by diffusion

(Figure 1.8B). The interior of leaves is

loosely packed with cells (Figure 1.7), with

large intercellular spaces that allow for

Figure 1.8. Illustration of fungicide diffusion of small molecules in water films. redistribution mechanisms: (A) Xylem (B) Translaminar (C) Vapor Translaminar redistribution provides short distance movement of an active ingredient, helping to provide uniform coverage on tissues.

Vapor. Many fungicides are non-polar, meaning they are water insoluble. This allows them to readily adsorb to leaf and berry waxes after application, forming residue deposits across their surface. From these deposits, fungicides can evaporate into a thin boundary layer of air close to plant surfaces and be transported across them (Figure 1.8C). After movement, the fungicide then re-binds to the surface waxes and can provide fungicidal activity at this new location if present in a high enough concentration.

FRAC groups used to manage grape powdery mildew

There is a wide range of vineyard management philosophies that govern the choice of fungicidal products. Synthetic products are widely used in conventional viticulture and in some sustainable certification programs such as Low Input Viticulture and Enology

(LIVE). Other sustainable certifications such as ‘USDA Organic’ and ‘Demeter

Biodynamic’ do not use synthetic products, however these certifications make up a small portion of viticultural production. For example, 2% and 7% of grape production is certified organic in California and Oregon, respectively (Strayer, 2017). Synthetic products will be the focus of this review because of their prevalence of use.

FRAC group 1: Benzimidazoles

Benzimidazole fungicides were among the first fungicides that had broad spectrum and curative activity with systemic redistribution, while still being relatively benign to non-target organisms and host plants. These attributes allowed for adoption of

farming practices that are commonplace today such as extended spray intervals and curative fungicide applications (Morton and Staub, 2008).

Benzimidazoles inhibit cellular division by binding to the β-subunit of fungal tubulin which prevents normal microtubule assembly. They have little or no effect on the initial stages of spore germination and are more effective at inhibiting germ tube elongation and mycelial growth (Fungicide Resistance Action Committee, 2017). This gives the benzimidazoles some curative activity as they inhibit cellular division among actively growing cells (Davidse, 1986). They have also been shown to redistribute in plant tissue in the xylem, accumulating in highly transpiring tissues such as actively growing leaves (Chatrath et al., 1972).

Resistance to benzimidazoles is due to a mutation in the fungal β-tubulin gene which causes decreased fungicide binding to tubulin, and cross resistance to all members of this FRAC group (Fungicide Resistance Action Committee, 2018). There are 6 different mutations known to cause loss of field performance (Fungicide Resistance

Action Committee, 2017). Resistant isolates have been characterized as having decreased fitness, by increasing or decreasing microtubule stability and sensitivity to temperature extremes (Fungicide Resistance Action Committee, 2017).

The benzimidazoles have historically been used to manage powdery mildew and botrytis, but buildup of resistance and the availability of other more efficacious products has decreased their usage. Since the 1980s there has been about 50% of natural populations of Botrytis with stable resistance to benzimidazoles in the Willamette Valley of Oregon found on strawberry, wine grape, blackberry, and snap beans (Johnson et al.,

1994). This fungicide class is not recommended for management of GPM or bunch rot in western Oregon (Fungicide Resistance Action Committee, 2013; Pscheidt, 2017).

FRAC group 3: Sterol biosynthesis inhibitors, demethylation inhibitors

Sterols are important biological molecules that help maintain fluidity of cell membranes over environmental temperature ranges, their production is the target of sterol biosynthesis inhibiting fungicides. There are a wide variety of sterol biosynthesis inhibiting fungicides, with the most significant group being the demethylation inhibitors

(DMIs).

DMIs inhibit sterol biosynthesis through inactivation of the enzyme 14α- demethylase (Kwok and Loeffler, 1993). This causes a build-up of sterol intermediates that interfere with normal cellular growth and can cause cell wall fidelity issues. This mode of action makes DMIs more effective after an infection has taken place, by reducing hyphal growth and sporulation. Spore germination is unhindered by the application of DMI fungicides because conidia already have the sterols needed in their cytoplasm for initial germ-tube formation and infection.

Most DMIs can redistribute via the xylem and as such, they have some degree of translaminar movement; some can also redistribute via vapors (Table 1.1; Ehr et al.,

(2008)).

There are three known mechanisms of fungicide resistance to DMIs: target site mutations in the 14α-demethylase enzyme, overexpression of the 14α-demethylase enzyme, and overexpression of cellular efflux pumps (Frenkel et al., 2014). If a fungal

strain possesses one of these mechanisms, it does not mean that it will express a high level of resistance. Target site mutations typically affect individual or a subset of DMI fungicides, with cross resistance across the whole FRAC group being incomplete.

Overexpression of the target enzyme is not specific to single DMI compounds, so cross resistance through this mechanism is common. However, resistance levels are usually lower than that of target site mutations (Cools et al., 2013) and gradually build up over a long period of repeated exposure, often due to an increase in the number of genetic alleles conferring the trait (copy number variant). Efflux pumps are also non-specific, and can reduce sensitivity to multiple DMI products. However, their role in resistance varies and has not been well characterized in E. necator (Cools et al., 2013; Frenkel et al., 2014).

Resistance levels are determined by varying combinations of these mechanisms and thus, resistance levels vary depending on their expression.

DMI resistance in E. necator has been documented since the 1990’s in the U.S.

(Erickson and Wilcox, 1997; Gubler et al., 1996) and has recently been shown to be present in Oregon and Washington at levels exceeding legal use rates (Mahaffee and

Neill, unpublished).

FRAC group 7: Succinate dehydrogenase inhibitors

The succinate dehydrogenase inhibitor (SDHI) fungicides have been around since the late 1960s and have undergone multiple phases of development. Early use focused on seed disinfection treatments and disease management in rice, with efficacy limited to basidiomycete diseases (Glattli et al., 2011). In the 1990s the discovery of improved

chemical structuring expanded the spectrum of disease control to include ascomycete diseases, such as GPM. This led to development of broad spectrum products focused on pathogens in cereal and specialty crops in the 2000s (Glattli et al., 2011).

Succinate dehydrogenase is a membrane bound protein in mitochondria that couples oxidation of succinate to fumarate with reduction of ubiquinone to ubiquinol, both of which are important processes in cellular energy production. SDHIs inhibit cellular respiration by binding in place of ubiquinone on the succinate dehydrogenase enzyme, halting the reduction of ubiquinone and subsequent cellular respiration (Glattli et al., 2011). In addition, SDHIs can penetrate the cuticle and move through leaf tissue, resulting in local and translaminar movement (Table 1.1).

SDHIs have a high risk of resistance development. Resistance has been documented in isolates of Botrytis cinerea (causal agent of bunch rot and grey mold) in

Chile and Italy (Angelini et al., 2014; Piqueras et al., 2014). E. necator populations resistant to SDHIs were found in Europe and were shown to be the result of four different mutations (Graf, 2018). No resistance to SDHIs has been characterized in E. necator in the USA to date. Cross resistance to FRAC 7 products is incomplete and resistance levels vary depending on the mutation and FRAC 7 fungicide used (Graf, 2018; Ishii et al.,

2011).

FRAC group 11: Quinone outside inhibitors, strobilurins

Quinone outside inhibitors (QoIs) are synthetically-modified versions of naturally-occurring compounds produced by wood rotting fungi (Vincelli, 2002).

Modifications to their chemical structure increased their stability in ultraviolet light,

which made them effective in field applications. QoIs inhibit complex III of the electron transport chain in mitochondria by binding to the quinone outside site of cytochrome b, a protein in the respiration chain of fungi and other eukaryotes. This binding inhibits electron transfer between cytochrome b and cytochrome c1, ceasing energy production in the cell (Bartlett et al., 2002). QoIs inhibit the germination of fungal spores, making them effective preventative fungicides.

All QoI products available for use in viticulture exhibit translaminar movement

(Bartlett et al., 2002; Vincelli, 2002). Some can also move via vapor or systemically

(Table 1.1).

Resistance to QoIs has been reported in E. necator on the east coast of the USA

(Baudoin et al., 2008) and throughout California (Miller and Gubler, 2004), Oregon, and

Washington (Warneke et al., 2016). Resistance has also been found in Michigan (Miles et al., 2012). There is cross-resistance between all FRAC 11 products (Fungicide Resistance

Action Committee, 2018).

FRAC group 13: Quinolines

Quinoline fungicides are chemically related to quinine, a natural alkaloid found in plants used commercially to flavor tonic water. The exact mode of action of quinolines is unknown, but disruption of signals that trigger spore germination and appressoria formation processes has been proposed (Lee et al., 2008; Wheeler et al., 2003).

Quinolines will do little to cure current infections and should only be used preventatively

(Deliere et al., 2010).

The most notable quinoline, quinoxyfen, can redistribute to a small extent on plant tissues through the xylem and via vapor. Its low amount of redistribution makes it more of a surface-active product (Wolters et al., 2004).

Multiple signal pathways are thought to be inhibited by quinoxyfen, resulting in medium risk of resistance development (Fungicide Resistance Action Committee, 2018).

High levels of resistance in E. necator were found in Virginia, where it played a major part in a GPM spray program (Colcol and Baudoin, 2015). Cross resistance to quinoline fungicides is present in E. necator (Fungicide Resistance Action Committee, 2018).

FRAC group M2: Sulfur

Sulfur is one of the, if not the, oldest fungicides known to humans. Sumerians and

Chinese societies used sulfur as early as 2500 BC to combat plant diseases (Oerke, 2006).

The Greek poet Homer referred to sulfur as “pest averting,” and the ancient Greeks used it to control rust diseases on wheat around 1000 BC (Tweedy, 1981). Sulfur has gone in and out of favor as new fungicides are developed and old products are retired. When the

Bordeaux mixture was invented in the 1880s, it quickly displaced many of the uses of elemental sulfur. In the early 1900s, sulfur regained favor as lime sulfur summer sprays were explored. Mid-way through the 20th century, its use decreased with the introduction of organic sulfur fungicides such as Captan and Maneb. The recent rise in producing agricultural crops with lower synthetic chemical input has placed renewed importance on sulfur use. Through the ebbs and flows of its popularity, sulfur has continued to be a cost- effective, reliable product.

The precise mode of action of sulfur has been the subject of numerous investigations, all of which have failed to clearly elucidate a single pathway (Tweedy,

1981). The most widely accepted theory, referred to as the general toxicant theory, revolves around inorganic sulfur reacting with compounds in the phyllosphere and producing fungicidal compounds. Two compounds that are formed when sulfur has been applied are sulfur dioxide and hydrogen sulfide. Sulfur dioxide is formed by oxidation and is directly toxic to microorganisms; hydrogen sulfide is hypothesized to be produced as a result of chemical reactions that are detrimental to fungal growth (Tweedy, 1981).

Sulfur primarily works when it is in contact with the fungus itself, making it a protective fungicide. Sulfur vaporizes on leaf and berry tissue providing a small zone of inhibition around sulfur deposits, but thorough coverage is essential for disease control.

Sulfur is used in dust, micronized, and colloidal forms. When used as a dust sulfur is applied dry and ideally mixes with dew on plant surfaces, helping in its retention and disease prevention activity (Tweedy, 1981). Most modern products are micronized wettable formulations that are applied in aqueous sprays. Micronized sulfurs are predominantly used today because of their lower dust content and the ability to mix them in a spray tank with other products. Modern micronized sulfur formulations spread well during applications due to small-uniform sulfur particle sizes, resulting in better coverage

(Tweedy, 1981). Colloidal sulfurs are similar to micronized formulations, but are produced in a different manner.

Sulfur can cause phytotoxicity if high relative humidity (75%+), and high temperature (30°C +, Bernard et al., 2003) conditions occur shortly (<2 hr) after spraying.

To cause phytotoxicity, sulfur vaporizes from plant tissues at temperatures dependent on

the sulfur particle size, with smaller particles vaporizing easier and at lower temperatures

(Bernard et al., 2003). Vaporized sulfur is oxidized to form sulfur dioxide (SO2) which is

2- then converted to sulfite (SO3 ) in humid air like that close to leaf surfaces. Sulfite can enter leaf tissues through the stomata where it mixes with water to form sulfuric acid

(H2SO4), which burns tissues (Bernard et al., 2003). Guidelines for when to avoid spraying sulfur because of the risk of phytotoxicity vary by region, rates used, and growing conditions. Most guidelines advise caution when temperatures reach 30°C

(86°F) shortly after spraying (Bernard et al., 2003; Cowgill and Rosenberger, 2013;

Wolf, 2005). Temperatures higher than 30°C are typically fine when they occur more than 2 hours after spraying in combination with low humidity conditions (<75%, Bernard et al., 2003). When sulfur sprays are made during the day in high temperature conditions

(28°C+), water droplets can cause “magnifying glass effects” on foliage, which can contribute to phytotoxicity (Bernard et al., 2003). In general, if sulfur sprays are avoided near the onset of high temperature and relative humidity conditions, the risk of phytotoxicity should be low.

FRAC Group U6: Amidoximes Cyflufenamid is the only amidoxime currently registered and was discovered in the early 2000s through chemical substitutions on existing acaricides and herbicides.

Subsequent trials showed its efficacy and toxicological profiles to be favorable for management of powdery mildew diseases (Sano et al., 2007).

The exact mode of action of cyflufenamid is unknown. A study examining its effect on wheat powdery mildew infection processes found that cyflufenamid did not

affect spore germination or infection processes before appressorium formation, but was effective at inhibiting subsequent processes such as haustorium formation, colony growth, and sporulation (Sano et al., 2007). Cyflufenamid has curative activity, providing nearly complete control of cucurbit powdery mildew (Podosphaera xanthii) when applied four days after inoculation (Haramoto et al., 2006). Field tests on GPM showed cyflufenamid to be effective at managing infections on leaves and clusters (Hulin et al.,

2009).

Cyflufenamid is absorbed into plant tissues and can translocate through leaves providing effective translaminar activity, however xylem redistribution is minimal.

Cyflufenamid also redistributes via vapors (Haramoto et al., 2006).

Commercial field resistance to cyflufenamid has not been documented. High levels of resistance were found in research fields where four applications were made per year to control cucurbit powdery mildew (Pirondi et al. 2014).

FRAC group U8: Benzophenones

Benzophenones are some of the newest entrants into the powdery mildew fungicide market. The exact mode of action of benzophenones is not known, but disruption of actin production and assemblage is proposed (Opalski et al., 2006).

Benzophenone compounds absorb into the cuticle, making applications rain-fast. Both compounds available in grapes (metrafenone and pyriofenone) redistribute via vapor, but can also redistribute via the apoplast, resulting in translaminar activity (BASF

Corporation, 2010; Summit Agro USA, 2017; Warneke et al., 2017).

There is no documented resistance to benzophenone fungicides in the US, but resistant E. necator populations that were cross resistant to both metrafenone and pyriofenone were observed in Italy (Kunova et al., 2016). In another study, resistant E. necator isolates were eliminated when application of metrafenone was ceased, indicating instability of the metrafenone resistant phenotype in that study (Stammler et al., 2014).

Table 1.1. Fungicides registered for grape powdery mildew control in Oregon and their attributes. Table adapted from Ehr et al., 2008. FRAC Active ingredient Trade Name(s) Redistribution profileab Attributesc Group Thiophanate- Incognito, Topsin Protectant, 1 Xylem systemic methyl M curative Protectant, Fenarimol Focus, Vintage Xylem systemic curative, eradicant Protectant, Myclobutanil Rally Xylem systemic, vapor curative Curative, Elite, Toledo, Xylem systemic, 3 Tebuconazole protectant, Tebucon, etc. translaminar, vapor eradicant Curative, Xylem systemic, Tetraconazole Mettle protectant, translaminar eradicant Protectant, Triflumizole Procure Xylem systemic curative Boscalid Endura Penetrant Protectant Xylem systemic, 7 Fluopyram Luna Privilege Protectant translaminar, vapor benzovindiflupyr Aprovia No Data No Data Protectant, Azoxystrobin Abound Xylem, translaminar curative 11 Trifloxystrobin Flint Translaminar, vapor Protectant Protectant, Kresoxim-methyl Sovran Translaminar, vapor curative 13 quinoxyfen Quintec Vapor, translaminar Protectant Microthiol Disperss, M2 Sulfur Vapor Protectant Cosavet, Golden Micronized etc. Protectant, U6 Cyflufenamid Torino Vapor, translaminard curatived U8 Metrafenone Vivando Vapor, translaminar Protectant bFrom Ehr et al., (2008) with modifications based on the results from Chapter 2 of this thesis. bXylem systemic fungicides typically exhibit translaminar activity at the same time. cFrom Ehr et al., (2008). d From Haramoto et al., (2006)

Chapter 2

The Grape Powdery Mildew Conundrum – Fungicide Selection and Application Timing

Brent Warneke, Lindsey Thiessen, Walt Mahaffee

Formatted for Plant Disease

American Phytopathological Society

Introduction

Grapes are one of the highest value non-citrus fruit crops in the US with a farm gate value of $6.2 billion (USDA NASS, 2018). Excluding juice grapes, approximately half of the 7.1 Million tons of US production (42%) is for wine, followed by fresh consumption (40%), and raisin production (18%) (International Organization of Vine and

Wine, 2017). In 2017, the total economic impact of wine output on the US economy was

$219.9 billion (John Dunham & Associates, 2017).

One of the most important diseases of wine grapes (Vitis vinifera) the world over is Grape Powdery Mildew (GPM), caused by the obligate biotrophic pathogen, Erysiphe necator (Gadoury et al., 2012). Due to a lack of constitutive resistance in V. vinifera, management of GPM relies on regular fungicide applications that can amount to as much as 37% gross value of production (Sambucci et al., 2014).

Conventional GPM management programs begin with early season fungicide applications, with the goal of preventing infections as long as possible (Lybbert et al.,

2016; Sambucci et al., 2017). After initiation, applications typically continue on a 7 – 21 day schedule, depending on the products used (Sambucci et al., 2014). Even low levels

(1-5%) of E. necator infection on clusters can drastically decrease wine quality (Stummer et al., 2003), so viticulturists use tightly spaced applications to produce fruit with low disease levels. Furthermore, wine producers have been known to reject grape crops with > 3% infected berries (Bettiga et al., 2013; Hellman, 2003). In 2011, California statewide costs of GPM management were estimated to be $189 million and 74% of all restricted use pesticide applications were made targeting GPM (Sambucci et al., 2014).

Grape tissues gain ontogenic resistance to E. necator infection as they age that is strongly expressed in berries (Gadoury et al., 2003). Initially, V. vinifera inflorescences and young berries are highly susceptible to infection until berries are 3-4mm in diameter, when they become nearly immune to new infection (Ficke et al., 2002; Gadoury et al.,

2003). Thus, fungicide applications before this period are important to maintain disease- free fruit. Timing synthetic fungicide applications from bloom to when berries are 3 mm in diameter has shown potential to reduce the number of applications needed to obtain disease-free fruit. Gadoury et al. (2003) showed that two sequential applications of kresoxim-methyl applied to individual grape inflorescences at bloom and 10-days later provided similar disease control to eight fungicide applications beginning at 15 cm shoot growth (Gadoury et al., 2003). Similarly, in management of hop powdery mildew on hop cones, ontogenic resistance has been shown when fungicide programs ceasing at stage II of cone development had the same disease control as season-long programs (Gent et al.,

2014; Nelson et al., 2014). Integrating knowledge of ontogenic resistance in berries into fungicide spray plans may improve the efficiency of chemical use in vineyards; further improvement may be obtained by optimizing fungicide selection during grape bloom.

Redistribution of active ingredient around plant tissues is an important characteristic of many modern synthetic fungicides that aids in their efficacy and reliability (Klittich, 2014). There are multiple mechanisms of redistribution (e.g. via xylem, translaminar and vapor), with each mechanism advantageous in its own way to enhance fungicidal coverage of plant tissues. The degree of redistribution and the resulting efficacy differences have not been well investigated for many fungicidal products. For example, when quinoxyfen was applied during early hop flowering,

powdery mildew control was improved by 32% to 52% compared to treatment regimens that did not use quinoxyfen (Nelson et al., 2014). Quinoxyfen redistributes via xylem and vapors, so fungicide redistribution could explain some of the increased efficacy of applications during hop flowering (Henry, 2003). In vineyard canopies, which are typically large and dense, redistribution of fungicide may be important in protecting inflorescences and clusters, which are difficult to adequately cover during fungicide applications.

Timing fungicide applications to specific phenological stages provides a way to standardize GPM management programs over years and may elucidate optimal application timing regimes that take advantage of the unique chemical attributes of redistributing fungicides. Kast and Bleyer (2011) showed that three consecutive sprays, one before flowering, one during flowering, and one when the berries were 2mm in diameter (BBCH 55, 65, 73, respectively), had 90% of the effect at reducing GPM on clusters as a program containing seven sprays. Hoffman et al., (2004) observed that when myclobutanil was applied for three two-week interval sprays starting at immediate pre- bloom, nearly complete control of grape black rot (causal agent Guignardia bidwellii) was achieved on clusters. Interestingly, two applications initiated at the same time provided similar control (Hoffman et al., 2004). Timing applications of specific redistributing fungicides to grape flowering stages may improve management of GPM berry infections by taking advantage of fungicide redistribution.

The objectives of this study were to examine the redistribution properties of fungicides used to control grapevine powdery mildew and to determine the application

timing in relation to grape bloom that most effectively reduces the proportion of infected grape berries for five commonly-used fungicides with varying redistribution profiles.

Materials and Methods

Standard methods for detached leaf redistribution bioassays

To determine the mechanism(s) and extent of redistribution of fungicidal products, detached leaf bioassays with E. necator growth as a bioindicator were developed to examine xylem, translaminar, and vapor mechanisms (Table 2.1). Pinot noir leaves were surface disinfested with 0.6% sodium hypochlorite and placed in double- decker Petri dishes with tap-water in the bottom (Quinn and Powell Jr, 1982). Inside the double decker Petri dishes, the relative humidity was circa 70%, due to leaf transpiration and evaporation of water from the bottom petri dish. Commercial formulations of fungicidal products were suspended in non-sterile tap water. Six-millimeter diameter discs made from Whatman #54 filter paper were submerged in fungicide suspensions, or nonsterile tap water and each treatment was shaken in an orbital shaker at 80 rpm for 10 min. In addition, six-millimeter diameter discs made of virgin polytetrafluoroethylene

(PTFE, 0.5mm thickness) were applied to a small piece of wax paper with a portion of vacuum grease. Fungicides or nonsterile tap water were applied to PTFE discs by holding a Preval sprayer (Chicago Aerosol, Coal City, IL) 25 cm from the disc at an approximate angle of incidence of 30 - 45° and applying a ~ ½ sec burst of spray and then allowed to dry. Fungicide and water treated filter paper or PTFE discs were placed on the left and right side of the main leaf vein, respectively, at the intersection of the first two veins in

the first main lobe of the leaf (Figure 2.1). Each leaf had a fungicide treatment and water control disc. Treated leaves for all assays were incubated at 22 °C, light cycle of 16:8

(day:night), and relative humidity of 55% ± 3% for 48 hours prior to an E. necator inoculation and for the duration of each experiment.

Table 2.1. Fungicides and rates used in small plot and all fungicide redistribution assays. Fungicide Name Fungicide Mechanism Application Rate FRAC Redistribution Tradea Technical Mode of Action (formulated Group Properties product/ha)b Unknown; Inhibition of cell signaling and Quintecc Quinoxyfen 13 Xylem, vapor 292 ml appressorium development proposed

Translaminar, Toledoc Tebuconazole 0.28 kg Sterol xylem 3 demethylation inhibition Translaminar, Rhyme Flutriafol 292 ml xylem, vapor

Succinate Luna Translaminar, Fluopyram 7 dehydrogenase 292 ml Privilegec xylem, vapor inhibition Q I inhibitor of o Translaminar, Flintc Trifloxystrobin 11 mitochondrial bc1 0.14 kg vapor complex Microthiol Unknown multi- Sulfur M2 Vapor 3.36 kg Disperss c site activity Unknown; Disruption of Translaminar, Vivando Metrafenone U8 804 ml actin assemblage vapor proposed aAll fungicides listed here were used in detached leaf mobility assays. bFungicides for all experiments were mixed at the listed rate applied in a volume of 467 L of water per hectare at full canopy (50gal/A). cFungicides that were used in the small plot phenological experiment and fungicide through calyptra study.

An E. necator population originally collected from Oregon vineyards was maintained on Chardonnay seedlings in a growth chamber and used as inoculum for all

detached leaf experiments. This population had no known resistances to fungicides. After incubation of the experimental detached leaves, leaves were inoculated with E. necator conidia in a 55 cm x 55 cm x 100 cm (LxWxH) settling tower. Three to five Chardonnay seedlings heavily infected with E. necator were held 25 cm below the top of the tower, sprayed with three ~ ½ sec bursts of compressed air, then allowed to settle for 5 min after placing a lid on the tower. Leaves were then incubated again until even E. necator growth was observed across the leaf surface (~7-10 days) then assessed by measuring the area of disease inhibition (Figure 2.1).

Figure 2.1. Detached leaf setup and data collection illustration. (A) Leaf with treated disc (red) and untreated water control disc (blue). (B) Mildew-infected leaf with discs and visual showing measurement of infected area, measurements indicated with black arrows and approximate areas of disease inhibition indicated by triangles. All assays were arranged in a completely randomized design with four replicate leaves per treatment. Mean areas of inhibition (mm2) were used as the response to fit general linear models with fungicide as the explanatory variable using the base lm function in R 3.3.2 (R Core Team, 2016). Treatment comparisons were conducted using the LSD.test function (least significant difference) from the agricolae package

(Mendiburu, 2016). P values were adjusted with false discovery rate methods to account for multiple comparisons (Benjamini and Hochberg, 1995).

Xylem and translaminar redistribution assay

Whatman filter paper fungicide and water-impregnated discs were applied to the adaxial surface of a leaf (described above) for the xylem redistribution assay; for the translaminar assay, filter discs were applied to the abaxial leaf surface. After resting on the leaf for 10 min, discs were removed and discarded and the leaves were incubated for

48 hours prior to inoculation as described above. The area of disease inhibition was measured using an optical micrometer calibrated to millimeters in a Leica MKZ stereomicroscope with epiluminescence at 27 × magnification. Three measurements starting at the center of each filter paper location were conducted as follows: one acropetally along the main vein that the fungicide impregnated disc was applied to, one basipetally along the same vein, and one perpendicular to the vein extending through vein intersection where the fungicide impregnated disc was applied (Figure 2.1). An area of E. necator growth inhibition was calculated using the formula below.

퐴푟푒푎 표푓 퐼푛ℎ𝑖푏𝑖푡표푛 (푚푚2)

= ((0.5 ∗ 푝푒푟푝푒푛푑𝑖푐푢푙푎푟 푣푒𝑖푛 푚푒푎푠푢푟푒푚푒푛푡)

∗ 푎푐푟표푝푒푡푎푙 푚푒푎푠푢푟푒푚푒푛푡)

+ ((0.5 ∗ 푝푒푟푝푒푛푑𝑖푐푢푙푎푟 푣푒𝑖푛 푚푒푎푠푢푟푒푚푒푛푡)

∗ 푏푎푠𝑖푝푒푡푎푙 푚푒푎푠푢푟푒푚푒푛푡)

The criterion used to determine where to stop the measurements was where mycelia came within one millimeter of the measurement line in any direction.

Vapor redistribution assay PTFE discs coated with fungicide or tap water were applied to the adaxial surface of leaves and remained there for the duration of the assay. If the leaf surface at the vein intersection described above was not level, the discs were applied to a different spot on the leaf that was level. Leaves were inoculated as above. After 7 to 10 days, inhibition of

E. necator was exemplified as a torus around the PTFE disc that was not always uniform.

To account for these irregularities, the maximum and minimum distances from the edge of the PTFE disc to E. necator mycelia was averaged and was used to calculate an area

(mm2) of inhibition that excluded that of the PTFE disc using the equation below.

퐴푟푒푎 표푓 퐼푛ℎ𝑖푏𝑖푡𝑖표푛 (푚푚2)

= 휋(푎푣푒푟푎푔푒 푟푎푑𝑖푢푠 표푓 𝑖푛ℎ𝑖푏𝑖푡𝑖표푛 + 푑𝑖푠푘 푟푎푑𝑖푢푠)2 − 푑𝑖푠푘 푎푟푒푎

Fungicide translocation through calyptras

To determine if fungicide could translocate through calyptras and inhibit growth of E. necator on berries after calyptras detach, a greenhouse experiment was designed using potted grape vines. Pixie grape plants in 3.8 L pots were maintained powdery mildew free in a greenhouse through nightly vaporization of sulfur. Inflorescences at

BBCH 55-57 (flowers appressed – flowers separating) were marked and one leaf above and below the inflorescence were removed to prevent fungicide deposition on them. The experiment was arranged in a randomized complete block design with four replicates.

Blocks contained six treatments: Five fungicides and a tap water control were applied to inflorescences with the experimental unit as one inflorescence per plant.

Fungicides (Table 2.1) were suspended in unsterilized tap water in glass Preval spray jars (Chicago Aerosol, Coal City, IL). A one second spray burst was applied to each side of the inflorescence, then plants were placed in a greenhouse without sulfur volatilization for 7 days to allow the calyptras to detach. After calyptras detached, inflorescences were inoculated with a conidial suspension of 20,000 conidia × ml-1 in

Figure 2.2. Growth stages selected to investigate phenological timing of fungicide applications. Photos by Brent Warneke. sterilized distilled water containing 0.05% TWEEN 20. A Nalgene hand pump sprayer was used to deliver a one second burst of spray to each side of the inflorescence. After inoculation, inflorescences were incubated in the same greenhouse for 7 to 10 days before microscopic examination. Greenhouse day:night temperatures were 24 and 16 °C, a light cycle of 16:8 light:dark, and the average relative humidity was 58% ± standard deviation of 13%. The number of infected and uninfected berries/flowers was recorded.

The proportion of infected berries (0-1 value) was used as the response to fit a general linear model with fungicide as the explanatory variable using the base lm function in R 3.3.2 (R Core Team, 2016). Treatment comparisons were conducted using

Tukey’s honest significant difference (P < 0.05) from the agricolae package (Mendiburu,

2016).

Research vineyard A 19-year-old vineyard of Pinot noir (clone 2A on 420A root stock) located at the

Oregon State University Botany and Plant Pathology Field Laboratory in Corvallis,

Oregon was used to conduct the small plot phenological application timing experiment, field mobility assessment, and the 2017 fungicide rotation experiment. Vines were arranged with 1.5 m x 1.8 m spacing, trained in a bilateral Guyot system, and cane- pruned with vertical shoot positioning. After bud-break, shoots were thinned to one per node and hedged as needed when shoots grew above the top of the trellis (~1.8 m). The vineyard was scouted for GPM flag shoots after bud-break and any found were removed.

There was no fruiting zone leaf removal, irrigation, or fertilization applied over the course of the season.

Small plot phenological experiment Five fungicides were chosen due to their widespread use in the industry and varied redistribution profiles (Table 2.1). Three growth stages (Figure 2.2) during grapevine flowering were chosen to time fungicide applications due to the high susceptibility of inflorescences to infection, and the influence of inflorescence architecture on fungal spore deposition (Gee et al., 2008; Lorenz et al., 1995; Shavrukov et al., 2004). The inflorescence stages and fungicides were respectively used in a 3 x 5 factorial design and arranged in a randomized complete block design. There were 17

plots of five grapevines in every block, with each plot representing the experimental unit.

The phenologically-timed treatments were compared to two controls, a non-treated control where water was applied, and a wettable sulfur (WS) control applied during each

14-day calendar fungicide application.

All treatments except the non-treated received WS (Microthiol Disperss, UPI,

King of Prussia, PA) applications every 14 days beginning when shoots were 20 – 30 cm

(8” – 12”) long until véraison (Figure 2.3). Applications were made at a concentration of

7.2 g/L formulated product (5.76 g/L active ingredient) and applied at a rate of 467 L/ha

(50 gal/A) at full canopy. An estimated 14 days prior to ~50% of the vineyard reaching each phenological stage, WS applications in phenological plots were ceased. When ~50% of the vineyard reached each phenological stage, fungicides (Table 2.1) were applied to plots specified to receive fungicide at that phenological stage. A second application of the same fungicide was made to those same plots 14 days later for a total of two sequential applications of each fungicidal product for all phenologically timed treatments. After the two phenologically initiated applications, sulfur applications resumed at 14-day intervals and were continued until véraison. The experiment was conducted in the same location during the 2015 and 2016 growing seasons with randomization of treatments within blocks at the beginning of each season.

All WS and phenological fungicide treatments (Figure 2.3) were applied using a custom ducted over-the-row air-blast sprayer with venturi nozzles (Rears Manufacturing,

Coburg, Oregon). Nozzles were fitted with TeeJet CP4916-30 spray discs (Spraying

Systems Co., Glendale Heights, IL) and operated at 207kPa (30psi), each nozzle yielding an output of 6.25e-6 m3/s (0.099 gpm). Early season applications started with four nozzles in use (two on each side of the sprayer) at an approximate output of 203L × ha-1

Figure 2.3. Timeline of approximate fungicide application schedule and data collections in field experiments. Photos by Brent Warneke. (21.5 gal/A), and one nozzle per side was added as the canopy grew until the maximum of 10 nozzles was reached. Every time a nozzle was added to each side of the sprayer, total sprayer output increased by 102 L/ × ha-1 (10.73 gal/A).

All data was collected from the middle three vines in each plot to minimize plot- plot interference. GPM leaf incidence assessments were conducted weekly starting at approximately 30 cm shoot growth and continued until véraison. Leaf incidence was assessed visually by examining 10 arbitrarily selected leaves per plant between four and

seven leaves down from the topmost unfurled leaf on a shoot. To assess cluster infection levels, ten clusters per plot were arbitrarily selected just prior to véraison, collected individually in a zip sandwich bag, and frozen at – 20 °C until assessment. For cluster disease assessments, frozen berries were removed from the rachis of each cluster individually and 25 berries per cluster were randomly selected for microscopic examination. Berry disease incidence was assessed at 27 × magnification using a Leica

MKZ stereomicroscope with epiluminescence, enabling detection of diffuse E. necator infections (Gadoury et al., 2007). The presence of hyphae and/or 2-3 groupings of necrotized epidermal cells were used as the indicator of infection. The number of infected berries out of 25 on each cluster was recorded.

Leaf incidence data was used to calculate absolute area under disease progress curves (AUDPC) using the agricolae package (Mendiburu, 2016). Leaf AUDPCs were used to fit linear mixed effects models using the lme function from the nlme package, with block as a random effect (Mendiburu, 2016; Pinheiro et al., 2016; R Core Team,

2016; Warnes et al., 2015). Year was shown to have a significant effect on model fit

(Likelihood ratio = 17.2, 1 df, P<0.001) so analysis of individual years was done to avoid homogenization of variance between years. Student’s t-tests were used to compare

AUDPCs to non-treated water and WS controls with P-values adjusted using the

Bonferroni correction to account for 16 comparisons. The number of infected berries per experimental unit (n/250) was used as a binomially distributed probability of berry infection and modeled using a generalized linear mixed model (GLMM). There was a significant interaction between fungicide and application timing (Drop-in- deviance test,

χ2 = 33.1, df = 8, P<0.001). A factorial model excluding the controls was fitted to

examine the interaction between fungicide and application timing. Controls were excluded from the interaction model because they did not have a phenological timing parameter associated with their application. For the factorial GLMM, year was included as a fixed effect to account for the incremental difference in disease levels between years.

A simple effects model was used to compare disease levels among all treatments. For simple effects GLMMs separate models were fit to each year of data to avoid homogenization of variance and to observe trends between years. In both simple and interaction models, block was included as a random effect. Treatment contrasts in simple and factorial models were conducted with lsmeans package and fit of all GLMMs were checked with the DHARMa package and an overdispersion function (Bolker et al., 2009;

Hartig, 2016; Lenth, 2016). Uncertainty was estimated using asymptotic 95% confidence intervals in both simple and factorial GLMMs after comparison of profile, bootstrap, and asymptotic methods. Any overdispersion due to extra-binomial variation was corrected for using an observational level random effect (Harrison, 2014).

Field redistribution assessment

Fungicide redistribution to clusters from leaves was investigated concurrently in a field setting during the 2015 & 2016 phenological application timing experiment. A set of

10 clusters was marked in every growth stage × fungicide treatment plot and covered with plastic zip top sandwich bags (Bi-Mart, Eugene, OR) during the phenologically timed fungicide applications to prevent any fungicide deposition directly onto the clusters. Plastic bags were removed from marked inflorescences immediately following

fungicide applications. Clusters were harvested and assessed for disease as described above. Statistical analysis was conducted using GLMM methods as described above.

2017 fungicide rotation experiment

To investigate the efficacy of rotations around quinoxyfen and fluopyram applied at BBCH 69 (end of bloom), a trial was conducted in the same research vineyard in 2017.

Trifloxystrobin was excluded from the 2017 trial due to widespread FRAC 11 fungicide resistance in Oregon grape growing regions (Warneke et al., 2016). The experiment was arranged as a randomized complete block with treatments each applied to a row of the vineyard, amounting to four replications. Early season WS applications were followed by treatments consisting of single applications of fluopyram or quinoxyfen (Table 2.1) and rotations between the products, all initiated when ≥ 50% of the vineyard reached end of bloom (Figure 2.3) for a total of four different fungicide treatments. Sulfur was resumed after synthetic applications and continued until véraison.

Fungicide treatments were compared to non-treated water and calendar sulfur controls both applied as described above. One plot at alternating ends of each calendar sulfur row was used as a non-treated control.

Leaf incidence (as above) was monitored on the second and fourth vine within the three middle plots of each row (6 vines per row), and the middle three vines of each non- treated control plot. Just prior to véraison, one cluster was collected from each monitored vine (6 per experimental unit) into individual bags and frozen until assessed as above.

Leaf incidence data was used to calculate absolute area under disease progress curves as described above. GLMM fitting and treatment contrasts were conducted as described above. A simple effects model containing all treatments was used to compare disease levels among the treatments.

Results

Detached leaf fungicide redistribution Flutriafol and fluopyram had significantly larger areas of inhibition than all other treatments when fungicides were applied to the adaxial leaf surface (P<0.05, Table 2.2).

All other fungicides applied to the adaxial surface had a zone of inhibition that was not significantly different from the control (P>0.05).

Table 2.2. Mean areas of inhibition (mm2) from detached leaf assays. Detached leaf assaysv Fungicidew Xylemx Translaminary Vaporz Flutriafol 578 a 289 a 148 a Fluopyram 256 b 269 a 69 b Tebuconazole 124 bc 116 b 55 b Metrafenone 107 bc 74 bc 54 b Wettable 98 bc 0 c 29 bc Sulfur Trifloxystrobin 81 bc 79 bc 75 b Quinoxyfen 56 c 99 b 25 bc Control 12 c 0 c 0 c vMean values with dissimilar letters are significantly different at P<0.05. wFungicides used, trade names and rates are listed in Table 2.1. xAreas of inhibition from adaxial fungicide application. yAreas of inhibition from abaxial fungicide application. zAreas of inhibition surrounding PTFE fungicide treated discs.

Flutriafol and fluopyram had significantly larger translaminar areas of inhibition than all other fungicides, indicating translaminar activity in combination with xylem mobility (P<0.05). Quinoxyfen and tebuconazole had significantly larger translaminar areas of inhibition than the control (P<0.05), but significantly smaller areas than flutriafol or fluopyram (P<0.05), indicating they could translocate through the leaf but with little additional xylem mobility. Metrafenone and trifloxystrobin showed inhibition of E. necator growth, but not significantly greater than the control (P>0.05, Table 2.2).

Most fungicides had significantly larger vapor areas of inhibition than the control

(P<0.05). Although WS and quinoxyfen had similar areas of inhibition to the control, they visually inhibited E. necator growth, indicating a small amount of redistribution was occuring.

Fungicide translocation through calyptras Fluopyram, trifloxystrobin, and tebuconazole resulted in significantly lower proportions of infected berries compared to the tap water control (P<0.05, Figure 2.4). WS and

Figure 2.4. Mean proportions of infected berries by treatment in fungicide through calyptra study. Treatments with similar letters do not differ significantly from each other at P<0.05.

quinoxyfen treatments resulted in lower but not significantly different proportions of infected berries to the tap water control.

Phenologically timed application experiment Across both years, 50% bloom and end of bloom timings were more likely to result in significantly lower AUDPC values than the calendar sulfur control (P≤0.05).

Fluopyram treatments tended to have lower mean AUDPC values as application time progressed, with the end of bloom application resulting in the lowest values (Figure 2.5).

The end of bloom trifloxystrobin treatment resulted in significantly lower AUDPC values than the calendar sulfur in both years (P≤0.05). In 2016, quinoxyfen displayed a similar trend as fluopyram, whereby there was a trend towards lower AUDPC values when quinoxyfen was applied later in flowering. Overall, AUDPC values for all phenologically timed fungicide treatments were similar to or lower than that of the calendar sulfur control, and all were significantly lower than the non-treated control (P<0.01).

icates a significant difference from from significant difference a icates

SE. Double and single asterisks above bars ind above asterisks single and SE.Double

Area under disease progress curve values for phenologically timed in treatments 2015 phenologically values curve progress for fungicide disease under Area

Figure 2.5. Figure bars are Error 2016. and sulfur respectively. P<0.05, controland P<0.01 at calendar the

In the factorial analysis, the mean proportion of infected berries was significantly influenced by the interaction between fungicide and application timing (Drop-in-deviance test, χ2 = 33.1, df = 8, one sided P<0.001). For fluopyram, quinoxyfen, and trifloxystrobin mean odds ratios indicated that at inflorescence elongation berries had 11.45, 8.04, and Figure 2.6. Estimated odds ratios of timing comparisons separated by fungicide. Dots are mean odds ratios and bars are 4.84 times higher odds of 95%confidence intervals. The horizontal line in each plot at 1 indicates where the odds of berry infection between the two treatments compared is equal. infection than when applied at end of bloom (P<0.01, Figure 2.6). For fluopyram and quinoxyfen inflorescence elongation and 50% bloom comparisons, odds ratios resulted in a similar inference as end of bloom comparisons (P<0.01). For trifloxystrobin the mean odds ratio comparing inflorescence elongation and 50% bloom resulted in moderate evidence that application at 50% bloom was significantly more efficacious than at inflorescence elongation (P =

0.01). Mean odds ratios were close to one for all comparisons in WS and tebuconazole groups, indicating that application timing had no significant effect on the mean odds of

berry infection in these treatments (Figure 2.6). Probabilities of berry infection varied among treatments both within and between fungicides (Figure 2.7). Fluopyram and trifloxystrobin end of bloom applications resulted in significantly lower mean proportions of infected berries compared to calendar application of sulfur in both years (P<0.05).

Quinoxyfen end of bloom applications were the most effective application timing in both years, but only resulted in significantly lower mean proportions of infected berries compared to the calendar sulfur control in 2016 (P<0.05). All application timings for tebuconazole and WS resulted in relatively high mean proportions of infected berries that had confidence intervals overlapping with each other and the calendar sulfur control, indicating statistical similarity between those treatments (P>0.05, Figure 2.7).

Probabilities of berry infection in 2015 and 2016. Points are the mean probability of berry infection and bars infection and are berry of probability 95% infection 2015 in Probabilitiesmean the Points 2016. berry and of are

nfidence intervals. intervals. nfidence

Figure 2.7 Figure co

Field redistribution assessment The majority of treatments had significantly lower disease levels on clusters covered with plastic bags compared to the non-treated control in both 2015 and 2016 (Figure 2.8).

Treatments in 2015 had variable probabilities across fungicides with the majority of phenological treatments having statistically similar (P>0.05) mean probabilities of infection to the calendar sulfur control indicated by overlapping confidence intervals

(Figure 2.8). Fluopyram, quinoxyfen and trifloxystrobin showed similar trends to their un-bagged counterparts where 50% bloom and end of bloom application timings resulted in the lowest mean probabilities of berry infection (Figure 2.8).

dar sulfur control were not covered with bags and are included for comparison. included are covered not for and were sulfur bags with dar control

Probabilities of berry infection in the field redistribution assessment. Dots represent mean infection field represent Dots the in Probabilitiesassessment. redistribution berry of

Figure 2.8. Figure Note water the that intervals. 95% confidence bars infection represent and asymptotic berry of probabilities calen and control

Figure 2.9. Mean leaf incidence area under disease progress curve values by treatment from the 2017 fungicide rotation experiment. Letters indicate statistically different treatments at P<0.05. 2017 Fungicide rotation experiment All fungicide treatments resulted in significantly less leaf and fruit disease compared to the non-treated control (P<0.05, Figure 2.9 and 2.10, respectively). The fluopyram- quinoxyfen rotations resulted in significantly lower disease levels than other treatments in mean AUDPC values and berry incidence (P<0.05).

Figure 2.10. Probabilities of berry infection in the 2017 rotation experiment. Dots and bars represent estimated mean probabilities of berry infection and asymptotic 95% confidence intervals, respectively.

Discussion

The results from this study indicate that timing redistributing fungicide applications to the time between late flowering and early berry development (50% bloom and end of bloom) significantly improves GPM management on grape berries. Targeting redistributing fungicide applications to periods when they will be most efficient at disease management such as late bloom and early berry development may allow their overall use to be decreased. Reducing use of synthetic products while maintaining fruit with low disease levels would reduce fungicide cost since sulfur or comparable products are less expensive (Sambucci et al., 2014). In addition, synthetic products are prone to resistance development by phytopathogenic fungi and there is increasing concern about the presence of pesticides in wine (Brent and Hollomon, 1995; Doulia et al., 2017; Strayer,

2017). Reducing the number of applications of fungicides with single site modes of activity reduces selection pressures for fungicide resistant E. necator strains, and could also result in lower fungicide residue levels on harvested grapes (Brent and Hollomon,

1995).

Bloom and early berry development are well-known as critical times to make effective fungicide applications targeted at GPM (Gadoury et al., 2003; Kast and Bleyer,

2011). Our results agree with a similar experiment conducted by Gadoury et al. (2003) where two sequential applications of kresoxim-methyl with a 10-day interval to individual shoots starting at grape bloom provided as much control of GPM on grape berries as a full season fungicide program of eight sprays. In this study, applications of fluopyram, quinoxyfen, and trifloxystrobin initiated at bloom resulted in low mean proportions of infected berries, however applications initiated at end of bloom were more

effective. Integrating two applications of these chemistries initiated at end of bloom into a fungicide program may cover the period of peak berry susceptibility more adequately than sprays initiated at the beginning or middle of bloom. Individual V. vinifera grape berries are most susceptible to infection by E. necator from early bloom until three to four weeks post bloom, when they become nearly immune to infection (Gadoury et al.,

2003). When applied at end of bloom and two weeks later, fungicide residual activity would be longer than that of sprays initiated at mid-bloom and could potentially remain until around when the berries acquire ontogenic resistance.

The lower disease levels on berries in the end of bloom treatments could be related to fungicide absorption. At the end of bloom, berries are anatomically predisposed to efficiently receive and absorb active ingredient based on the large cuticular surface area present and lack of cuticular waxes. The grape berry cuticle starts to be formed one week prior to anthesis, and by full bloom cuticular ridges are present and tightly appressed (Casado and Heredia, 2001). At fruit set, the grape berry cuticle is characterized by numerous ridges that are spread apart, increasing the cuticular surface area (Casado and Heredia, 2001; Rosenquist and Morrison, 1988) which could increase the amount of fungicide absorbed. At fruit set there is also a low amount of epicuticular wax accumulated on berries (Rosenquist and Morrison, 1988). Epicuticular wax increases the hydrophobicity of grape berries, making it more difficult for fungicide suspensions to adhere and be absorbed. At six days post-bloom, rudimentary wax crystals start to form and by 18 days post-bloom, cuticular ridges are still present but wax platelets have started to cover the grape berry (KyoungHee et al., 2010). An application at 10 to 14 days post- bloom, as in the present study, should allow for fungicide to contact the berry cuticle

before berries are completely covered with epicuticular wax. In addition, when applications were made 10 to 14 days post-bloom, berries were physically larger, allowing for more fungicidal spray interception. Thus, in relation to grape flowering, fungicide deposition and absorption would likely be maximized at the end of bloom

(Rosenquist and Morrison, 1988). The absorption of fluopyram, trifloxystrobin, and quinoxyfen into grape berries and subsequent redistribution may have enhanced coverage of clusters enough to where they effectively protected a higher proportion of berries from

E. necator infection than sulfur. Clusters are difficult to adequately cover with fungicide applications due to their location within the canopy (Austin et al., 2011; Wise et al.,

2010). In this study, fluopyram, trifloxystrobin, and quinoxyfen all could translocate through leaves, and of the three, all but quinoxyfen could translocate through calyptras.

While quinoxyfen did not exhibit as much redistribution as fluopyram or trifloxystrobin, this could be due to the conditions in the redistribution experiments, which were not reflective of a field scenario.

In our study tebuconazole (FRAC 3) applied at bloom did not effectively control

GPM on grape berries. DMI fungicides exhibit fungistatic effects on established E. necator colonies, halting their growth (Gadoury et al., 2007). If berries had not yet acquired ontogenic resistance once the activity of tebuconazole had subsided, continued colony growth would have resulted in formation of diffuse infections on berry surfaces

(Gadoury et al., 2007). Microscopic evaluation of berries in this study enabled detection of these macroscopically inconspicuous colonies. Prevention of diffuse colonies requires extending protection of berries through the point when they are susceptible to diffuse colonization, which is after the fruit gain ontogenic resistance (Gadoury et al., 2003,

2007). The two applications of tebuconazole in this study may not have provided the residual efficacy needed to prevent formation of diffuse colonies on berries. In addition, some of the lower efficacy of tebuconazole on grape berries could be due to suspected shifts to lower DMI sensitivity in E. necator populations at the field site used due to their historic use in adjacent plots.

The higher efficacy of fluopyram, quinoxyfen, and trifloxystrobin on berries could be due to their inhibition of the initial stages of E. necator infection. All three have a mode of action that is effective at stopping spores in the initial stages of E. necator host tissue colonization (Bartlett et al., 2002; Glattli et al., 2011; Wheeler et al., 2003). By stopping the initial stages of infection, they may be more effective at preventing formation of diffuse E. necator colonies on berries. Redistribution around the cluster could have provided additional efficacy against diffuse E. necator infections.

The lower levels of infection seen in clusters in the phenological timing experiment could be in part due to vapor redistribution from adjacent tissues that were treated directly. In Pearson et al. (1994), clusters and leaves did not receive any direct fungicide application and formulated FRAC 3 fungicides were applied to cheesecloth sheets which were suspended at various distances from clusters at bloom. The greatest disease reduction on clusters was observed in a 15 cm band around the treated cheesecloth (Pearson et al., 1994). In this study, leaves and tissues positioned near clusters that were directly treated with fungicides could have released active ingredient to the fruiting zone via vaporization. ‘Pinot noir’ is a tight bunch variety, which results in less airflow through the bunch (Shavrukov et al., 2004). Reduced airflow increases the diameter of the boundary layer of air close to plant tissues, which may facilitate higher

retention of vaporized active ingredient in tight bunched clusters (Nobel, 1975). While fungicide mobility likely aided in protecting clusters, prophylactic sulfur spray applications were important in keeping disease pressure in check until phenological fungicide applications.

The lower levels of berry infection in the phenological treated plots could also have been related to reduced foliar infection from fungicide treatment of leaves. In

Oregon, disease pressure prior to flowering is variable among years, but initial disease detection occurs prior to bloom (Thiessen et al., 2017). In 2015, there were few fungicide treatments that had significantly lower AUDPCs compared with the calendar sulfur control but in 2016 there were many (Figure 2.5). Bagged clusters followed a markedly similar trend to leaves where in 2015 many had similar incidence levels to the calendar sulfur control. Similarly, in 2016, most bagged clusters had significantly lower proportions of infected berries (Figure 2.8), indicating leaf infection levels may have played a role in cluster infection levels. Foliar E. necator infections are an inoculum source for clusters (Gadoury et al., 2001). When managing GPM on clusters, prophylactic sprays up until mid-bloom and use of synthetic chemistries at mid to late bloom until berries are 3 mm in diameter is a viable strategy to get efficacious GPM control on clusters, even in high disease pressure situations (Gadoury et al., 2003; Kast and Bleyer,

2011).

necator infection and growth. All fungicides investigated showed a torus of E. necator growth inhibition in the vapor assay, and most clusters covered during fungicide applications had reduced proportions of infected berries. Vapor redistribution of fungicide could play a role in the field efficacy of many fungicides, and concentrations in air close to vines may be higher than previously thought. Since there is limited data on the effect of fungicide redistribution on the growth of E. necator, this study expands our understanding of fungicide redistribution impacts on infection and growth of E. necator.

Adaptation of the detached leaf bioassay could provide a cost-effective way to investigate other aspects of fungicide activity, such as the curative and residual effects of fungicide redistribution. Alteration of the environmental parameters during the detached leaf assay could provide information on the redistribution of fungicides more relevant to field scenarios.

The phenologically timed WS treatments may have been less effective at reducing berry infection due to the rate and interval used, as well as a lack of absorption and residual activity. WS was applied at a rate of 3.36 kg/ha at an interval of 14 days, the lowest label rate and longest interval recommended by the manufacturer (Microthiol

Disperss label, UPI). While the end of bloom application may have covered the inflorescence effectively, rapid berry growth in the following weeks and lack of redistribution may have left portions of berries susceptible to infection. In addition, reliance on sulfur adsorption to berry surfaces may have left berries susceptible to infection if dew washed deposits off (Müller and Riederer, 2005). If a shorter application interval and/or a higher application rate were used, sulfur treatments may have been more effective at controlling GPM on grape berries.

Of the three application timings tested (inflorescence elongation, 50% bloom and end of bloom) fungicide applications at end of bloom were the most effective to reduce berry infection. In addition, fungicides (e.g. fluopyram, quinoxyfen) that have more activity in preventing spore germination and infection resulted in significantly lower leaf and berry infection levels. If viticulturists are using fluopyram, quinoxyfen, or trifloxystrobin, their effectiveness will likely be maximized if they are applied during end of bloom and one interval thereafter. In combination with cultural practices and early season management, two carefully timed synthetic fungicide applications from late bloom until 3-4mm berries may provide an effective strategy for wine grape growers in western Oregon seeking to produce fruit with low disease levels and minimum synthetic chemical inputs. Reduced synthetic inputs would have the added benefit of a reduction in selection pressure for fungicide resistant E. necator strains. Timing different fungicides to specific phenological growth stages could be an important aspect to improving disease management in grape and other production systems. Understanding how crop phenology and fungicide redistribution interact to impact disease development could help improve the efficacy of other fungicides used for GPM management and similar concepts explored herein may apply to other pathosystems.

References

Austin, C.N., Grove, G.G., Meyers, J.M., Wilcox, W.F., 2011. Powdery mildew severity as a function of canopy density: Associated impacts on sunlight penetration and spray coverage. Am. J. Enol. Vitic. 62, 23–31. Bartlett, D., Clough, J., Godwin, J., Hall, A., Hamer, M., Parr-Dobrzanski, B., 2002. The strobilurin fungicides. Pest Manag. Sci. 58, 649–662. Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc., B (Methodological) 57, 289–300. Bettiga, L.J., Gubler, W.D., Leavitt, G.M., 2013. Powdery Mildew, in: Bettiga, L.J. (Ed.), Grape Pest Management. University of California, pp. 137–149. Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H., White, J.-S.S., 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24, 127–135. Brent, K.J., Hollomon, D.W., 1995. Fungicide resistance in crop pathogens: How can it be managed?, 1st ed. International Group of National Associations of Agrochemical Manufacturers, Brussels. Casado, C.G., Heredia, A., 2001. Ultrastructure of the cuticle during growth of the grape berry (Vitis vinifera L.). Physiol. Plant. 111, 220–224. Doulia, D.S., Anagnos, E.K., Liapis, K.S., Klimentzos, D.A., 2017. Effect of clarification process on the removal of pesticide residues in white wine. Food Control 72, 134– 144. Ficke, A., Gadoury, D.M., Seem, R.C., 2002. Ontogenic resistance and plant disease management: A case study of grape powdery mildew. Phytopathology 92, 671– 675. Gadoury, D.M., Cadle-Davidson, L., Wilcox, W.F., Dry, I.B., Seem, R.C., Milgroom, M.G., 2012. Grapevine powdery mildew (Erysiphe necator): a fascinating system for the study of the biology, ecology and epidemiology of an obligate biotroph. Mol. Plant Pathol. 13, 1–16. Gadoury, D.M., Robert Seem, Andrea Ficke, Wayne Wilcox, 2003. Ontogenic resistance to powdery mildew in grape berries. Phytopathology 93, 547–555. Gadoury, D.M., Seem, R.C., Ficke, A., Wilcox, W.F., 2001. The epidemiology of powdery mildew on Concord grapes. Phytopathology 91, 948–955. Gadoury, D.M., Seem, R.C., Wilcox, W.F., Henick-Kling, T., Conterno, L., Day, A., Ficke, A., 2007. Effects of diffuse colonization of grape berries by Uncinula necator on bunch rots, berry microflora, and juice and wine quality. Phytopathology 97, 1356–1365. Gee, C.T., Gadoury, D.M., Cadle-Davidson, L., 2008. Ontogenic resistance to Uncinula necator varies by genotype and tissue type in a diverse collection of Vitis spp. Plant Dis. 92, 1067–1073. Gent, D.H., Grove, G.G., Nelson, M.E., Wolfenbarger, S.N., Woods, J.L., 2014. Crop damage caused by powdery mildew on hop and its relationship to late season management. Plant Pathol. 63, 625–639.

Glattli, A., Grote, T., Stammler, G., 2011. SDH-inhibitors: History, biological performance and molecular mode of action, in: Dehne, H.W., Deising, H.B., Gisi, U., Kuck, K.H., Russell, P.E., Lyr, H. (Eds.), Modern Fungicides and Antifungal Compounds VI. Presented at the 16th International Reinhardsbrunn Symposium, Deutsche Phytomedizinische Gesellschaft eV Selbstverlag, Friedrichroda, Germany, pp. 159–170. Harrison, X.A., 2014. Using observation-level random effects to model overdispersion in count data in ecology and evolution. PeerJ 2, e616. Hartig, F., 2016. DHARMa: Residual diagnostics for hierarchical (multi-level/mixed) regression models. Hellman, E.W., 2003. Oregon Viticulture. Oregon State University Press. Henry, M.J., 2003. Fungicides, quinoxyfen. Encycl. Agrochem. 872–873. Hoffman, L.E., Wilcox, W.F., Gadoury, D.M., Seem, R.C., Riegel, D.G., 2004. Integrated control of grape black rot: Influence of host phenology, inoculum availability, sanitation, and spray timing. Phytopathology 94, 641–650. International Organization of Vine and Wine, 2017. OIV Statistical Report on World Viticulture. John Dunham & Associates, 2017. The wine industry boosts the American economy by $219.9 billion in 2017 [WWW Document]. URL http://wineamerica.org/wp- content/themes/wineamerica/pdfs/impact/United-States-Report.pdf Kast, W.K., Bleyer, K., 2011. Efficacy of sprays applied against powdery mildew (Erysiphe necator) during a critical period for infections of clusters of grapevines (Vitis vinifera). J. Plant Pathol. 93, S29–S32. Klittich, C.J., 2014. Fungicide mobility and the influence of physical properties, in: Myung, K., Satchivi, N.M., Kingston, C.K. (Eds.), Retention, Uptake, and Translocation of Agrochemicals in Plants. American Chemical Society, pp. 95– 109. KyoungHee, S., InMyung, C., KyuHwan, C., JaeMan, L., HeeSeung, P., 2010. Morphological development and chemical composition of epicuticular wax crystals in “Campbell Early” grape. Hortic. Environ. Biotechnol. 51, 253–256. Lenth, R.V., 2016. Least-Squares Means: The R Package lsmeans. J. Stat. Softw. 69, 1– 33. Lorenz, D. h., Eichhorn, K. w., Bleiholder, H., Klose, R., Meier, U., Weber, E., 1995. Growth stages of the grapevine: Phenological growth stages of the grapevine (Vitis vinifera L. ssp. vinifera)—Codes and descriptions according to the extended BBCH scale. Aust. J. Grape Wine Res. 1, 100–103. Lybbert, T.J., Magnan, N., Gubler, W.D., 2016. Multidimensional responses to disease information: How do winegrape growers react to powdery mildew forecasts and to what environmental effect? Am. J. Agric. Econ. 98, 383–405. https://doi.org/10.1093/ajae/aav097 Mendiburu, F., 2016. agricolae: Statistical Procedures for Agricultural Research. Müller, C., Riederer, M., 2005. Plant surface properties in chemical ecology. J. Chem. Ecol. 31, 2621–2651. Nelson, M.E., Gent, D.H., Grove, G.G., 2014. Meta-analysis reveals a critical period for management of powdery mildew on hop cones. Plant Dis. 99, 632–640.

Nobel, P.S., 1975. Effective thickness and resistance of the air boundary layer adjacent to spherical plant parts. J. Exp. Bot. 26, 120–130. Pearson, R.C., Riegel, D.G., Gadoury, D.M., 1994. Control of powdery mildew in vineyards using single-application vapor-action treatments of triazole fungicides. Plant Dis. 78, 164–168. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Core Team, 2016. nlme: Linear and Nonlinear Mixed Effects Models. Quinn, J.A., Powell Jr, C.C., 1982. Effects of temperature, light, and relative humidity on powdery mildew of Begonia. Phytopathology 72, 480–484. R Core Team, 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rosenquist, J.K., Morrison, J.C., 1988. The development of the cuticle and epicuticular wax of the grape berry. Vitis 27, 63–70. Sambucci, O., Alston, J., Fuller, K., 2014. The costs of powdery mildew management in grapes and the value of resistant varieties: Evidence from California. Robert Mondavi Inst. Cent. Wine Econ. Working Paper 1402. Sambucci, O., Lawell, C.-Y.L., Lybbert, T.J., 2017. The spraying decisions of grape growers in response to disease forecasting information: A dynamic structural econometric model. Working paper, University of California at Davis. Shavrukov, Y.N., Dry, I.B., Thomas, M.R., 2004. Inflorescence and bunch architecture development in Vitis vinifera L. Aust. J. Grape Wine Res. 10, 116–124. https://doi.org/10.1111/j.1755-0238.2004.tb00014.x Strayer, P., 2017. The Organic Opportunity: Will the U.S. Wine Industry Miss Out? [WWW Document]. Wines Vines. URL http://www.winesandvines.com/columns_article/178258 (accessed 12.13.17). Stummer, B.E., Francis, I.L., Markides, A.J., Scott, E.S., 2003. The effect of powdery mildew infection of grape berries on juice and wine composition and on sensory properties of Chardonnay wines. Aust. J. Grape Wine Res. 9, 28–39. Thiessen, L.D., Neill, T.M., Mahaffee, W.F., 2017. Timing fungicide application intervals based on airborne Erysiphe necator concentrations. Plant Dis. 101, 1246–1252. https://doi.org/10.1094/PDIS-12-16-1727-RE USDA NASS, 2018. NASS Quick Stats [WWW Document]. URL https://quickstats.nass.usda.gov/ Warneke, B., Yamagata, J., Neill, T., Miles, T., Mahaffee, W., 2016. Detection of quinone outside inhibitor resistant isolates of Erysiphe necator in Oregon vineyards. Phytopathology, p. S71. Warnes, G., Bolker, B., Lumley, T., Johnson, R., 2015. gmodels: Various R Programming Tools for Model Fitting. Wheeler, I.E., Hollomon, D.W., Gustafson, G., Mitchell, J.C., Longhurst, C., Zhang, Z., Gurr, S.J., 2003. Quinoxyfen perturbs signal transduction in barley powdery mildew (Blumeria graminis f.sp. hordei). Mol. Plant Pathol. 4, 177–186. Wise, J.C., Jenkins, P.E., Schilder, A.M.C., Vandervoort, C., Isaacs, R., 2010. Sprayer type and water volume influence pesticide deposition and control of insect pests and diseases in juice grapes. Crop Prot. 29, 378–385.

Chapter 3

Detection of Quinone Outside Inhibitor Resistant Isolates of Erysiphe necator in Oregon Vineyards

Brent Warneke, Jesse Yamagata, Tara Neill, Timothy Miles, Walt Mahaffee

Full Manuscript Name and Authors:

TAQMAN AND DIGITAL DROPLET PCR TECHNIQUES TO DETECT

RESISTANT E. NECATOR ISOLATES TO QUIONONE OUTSIDE INHIBITOR

FUNGICIDES IN LEAF AND AIR SAMPLES

Timothy D. Miles1, Jesse S. Yamagata1, Brent Warneke2, Sharifa G. Crandall1, Tara

Neill3, and Walt F. Mahaffee3

1California State University, Monterey Bay, 5108 Fourth Avenue, Marina, CA 93933

2Department of Botany and Plant Pathology, Oregon State University, Corvallis, OR 97330

3USDA ARS HCRL, 3420 NW Orchard Avenue Corvallis, OR 97330

Below is a presentation of the components of this manuscript that were conducted by Brent Warneke.

Introduction

Wine grapes (Vitis vinifera), an important agricultural commodity throughout the world, are susceptible to a wide array of pathogens; of which, one of the most consistent and devastating diseases is grape powdery mildew (GPM), caused by Erysiphe necator

(Gadoury et al., 2015). The mild climates that favor high quality wine grape production also favor the reproductive biology of E. necator (Gadoury et al., 2012). V. vinifera has no constitutive resistance to E. necator, with all green tissues highly susceptible to infection (Brewer and Milgroom, 2010; Gadoury et al., 2003). Even low levels (1-5%) of

GPM infection on clusters can drastically decrease wine quality (Stummer et al., 2003).

Thus, wine producers have been known to reject grape crops with > 3% infected berries

(Bettiga et al., 2013; Hellman, 2003). These factors result in growers using intensive fungicide programs designed to keep GPM cluster infections to a minimum.

Conventional GPM management programs begin with early season fungicide applications, with the goal of preventing infections as long as possible (Lybbert et al.,

2016; Sambucci et al., 2017). After initiation, applications typically continue on a 7 – 21 day schedule, depending on the products used (Sambucci et al., 2014). While most fungicide programs rotate between applications of multi-site protectant fungicides (e.g. sulfur) and synthetic fungicides [e.g. quinone outside inhibitors (QoI) or sterol demethylation inhibitors (DMI)], there is still a heavy reliance on synthetic fungicides because of their rainfastness, higher efficacy at lower doses, and longer intervals between sprays (Morton and Staub, 2008). However, synthetic fungicides are susceptible to resistance development; particularly single mode of action fungicides like the QoIs

(Fungicide Resistance Action Committee, 2012a; Gisi et al., 2002).

The QoIs were first introduced in 1996 and quickly became one of the most widely used group of fungicides, registered for use on over 400 host/disease systems in

72 countries (Bartlett et al., 2002). QoI fungicides have been one of the most popular choices for viticulturists due to their broad-spectrum activity, making them effective management tools to control GPM and grape downy mildew, caused by Plasmopara viticola (Ash, 2000; Skinkis et al., 2017). While viticulturists use QoI products in rotation with other modes of action (Sambucci et al., 2014), they tend to use them multiple times per season and when GPM is already present in the vineyard, which increases the chance of resistance development (Brent and Hollomon, 1995).

Resistance to QoI products emerged shortly after their introduction in wheat powdery mildew (causal agent Erysiphe graminis f. sp. tritici) and grape downy mildew pathosystems (Heaney et al., 2000; Sierotzki et al., 2000). In both cases resistance was shown to be the result of a point mutation in the cytochrome b gene that caused an alanine to be substituted for glycine at the 143-position in the cytochrome b protein, termed “G143A.” The amino acid swap changed the conformation of the quinone outside binding site of cytochrome b just enough to result in ineffective active ingredient binding.

This resulted in a qualitative resistance response, where resistant individuals could withstand high concentrations of fungicide. Resistance to QoIs in E. necator populations mediated by the G143A mutation has been found on the east coast of the USA and in

Michigan; QoI resistance has also been found in California, although that study did not test for the presence of G143A (Baudoin et al., 2008; Miles et al., 2012; Miller and

Gubler, 2004). QoI resistance has not been previously documented in Oregon E. necator

populations, although QoI resistance was found in the Willamette Valley among Septoria tritici blotch (Zymoseptoria tritici) isolates in 2012 (Estep et al., 2013).

In the summer of 2015 there were multiple reports of inadequate control of GPM in the Willamette Valley of Oregon. Many of the vineyards in question had been using

QoI products, so an E. necator isolate collection survey was undertaken in response to grower concerns. The objectives included, 1) evaluate Oregon populations of E. necator from multiple growing regions to determine if QoI resistance was present; 2) characterize the sensitivity of isolates to QoI fungicides; and 3) validate a competitive TaqMan quantitative PCR assay to monitor for QoI resistance.

Materials and Methods

Field sample collection. Grape plant material infected with E. necator was collected from commercial vineyards reporting issues managing GPM throughout the Willamette

Figure 3.1. Approximate E. necator sample collection locations in Oregon. Note: only one point was placed in CG because the provenance of the samples received from that region was unknown.

Valley (WV, Figure 3.1), the Columbia River Gorge (CG, Figure 3.1), and Southern

Oregon (SO, Figure 3.1) wine grape growing regions between August 11 and October 23,

2015. In the Willamette Valley, 19 commercial vineyards were sampled, generating 57 field samples; six commercial vineyards in Southern Oregon yielded 13 field samples; and 10 field samples from 10 vineyards were received from a collaborator, Dr. Steve

Castagnoli (Extension Horticulturist, Hood River County), in the Columbia River Gorge region. In addition, two isolates were obtained from a research greenhouse that had never been treated with QoI fungicides. Single conidial chain isolates were generated from each field sample (described below) and the remaining fungal material (hyphae, conidia, cleistothecia) was collected onto cellophane tape (Scotch Tape, 3M, Saint Paul, MN,

USA) similar to the methods of (Brewer and Milgroom 2010). A small roll (~5 mm diameter) of tape was rolled over all visible colonies on the plant tissue, then transferred to sterile 1.5 ml microcentrifuge tubes, and stored at -20 ˚C until DNA extraction.

Genomic DNA was extracted from the tape samples using the UltraClean Microbial DNA

Isolation Kit (MoBio Laboratories, Inc., Carlsbad, CA, USA) using the manufacturer’s protocol and stored at -20 ˚C until analysis.

Isolate generation, maintenance, and DNA extractions. An eyelash glued to a Pasteur pipet was surface disinfested in 70% ethanol for one minute then used to transfer individual conidial chains from infected plant material onto surface disinfested Carignan or Chardonnay leaves maintained in a double petri-dish isolation chamber and incubated at 21°C with a 16 h photoperiod [(Quinn and Powell Jr, 1982) Figure 3.2A, B].

Disinfested leaves were generated by harvesting leaves at the 4th or 5th node from the

growing tip of disease-free source plants maintained in a greenhouse in which sulfur was vaporized nightly. Leaves were submerged in 10% bleach (0.6% sodium hypochlorite,

Clorox, Oakland CA), vortexed 2 minutes at max speed on a platform adaptor, and

repeatedly rinsed with 1L sterile deionized

water. Isolates were transferred every 2 to 3

weeks to fresh disinfested leaves in duplicate

using autoclaved fine paintbrushes. After at

least a 10-day incubation, E. necator fungal

material was collected with tape rolls

(described above) from sporulating colonies

of each isolate. Isolate DNA was extracted

using a method modified from Brewer and

Milgroom (2010). Following freezing, 200ul

of 5% w/v Chelex 100 (Sigma-Aldrich, Saint

Louis, MO, USA) in molecular-grade water

was added to each 2 ml tube using a large-

orifice pipette tip. Tubes were vortexed

horizontally at max speed for 5 minutes,

Figure 3.2. Two-tier petri dish detached leaf briefly centrifuged at 10,000 × g, incubated at chambers of Carignan (A, B) leaves used for E. necator maintenance and a homemade 95 °C for 10 min, then vortexed for 5 sec and spore settling tower (C) used for inoculations. briefly centrifuged before another 10 min incubation at 95°C. The tubes were then allowed to cool to room temperature, centrifuged

at 10,000 × g for 2 min, and the supernatant was transferred to a sterile labeled 1.5ml microcentrifuge tube and stored at -20˚C until analysis.

QoI fungicide sensitivity bioassays. 24-well plate bioassay. A 24-well plate conidia germination bioassay was used to calculate the EC50 for each sensitive isolate as determined by the G143A qPCR. EC50 was determined as the concentration at which conidial germination for each isolate was reduced by 50% in comparison with the control. Water agar (1.5%) amended with formulated trifloxystrobin or kresoxim-methyl (Flint 50WG and Sovran WG, respectively) at rates of 0, 0.01, 0.1, 1, 10, and 100µg/ml each including 100µg/ml salicylhydroxamic acid (SHAM, 99%, Sigma-Aldrich, Saint Louis, MO, USA) were dispensed, 1.5ml/well in Costar 24-well TC-treated multiwall plates (Corning Inc.,

Corning, NY, USA), 4 wells per fungicide concentration. Plates were inoculated using settling towers constructed from 50L plastic garbage cans (Figure 3.2C, Sterilite Corp.,

Townsend, MA, USA). One 1-second burst of 483 kPa compressed air was blown onto detached leaves with 10-14 day old E. necator sporulating lesions, then allowed to settle onto the 24-well plates for 5 minutes before covering the plates and incubating for 24h at ambient temperature (21˚ to 23˚C) under fluorescent light (Miles et al., 2012). Each plate was microscopically examined and 50 conidia per well assessed as germinated or not based on if it contained an appressorium and/or a germ tube ≥ the width of the conidium.

The four well subsamples per treatment were averaged and percent germination calculated for each fungicide concentration. EC50 values were calculated by fitting a broken-stick model to spore germination percentage data in R version 3.2.2 using the

log10 of the fungicide active ingredient concentration as the explanatory variable (R Core

Team, 2016). A broken stick model was used to minimize overestimation of EC50.

Discriminatory dose bioassay. Upon determination of EC50 ranges from sensitive isolates in the 24-well bioassay, a discriminatory dose assay was developed to more quickly screen other isolates. Petri dishes (60mm x 15mm) with 1.5% water agar amended with 0.1µg/ml formulated trifloxystrobin or kresoxim-methyl and 100µg/ml

SHAM were inoculated with E. necator conidia from 10-14 day old colonies by tapping a conidia-covered fine paintbrush over each plate in a laminar flow hood. Water agar amended with 100µg/ml SHAM were included as germination controls. Plates were incubated as described above. The number of germinated conidia was enumerated as described above. An isolate was determined to be sensitive to trifloxystrobin or kresoxim-methyl if it exhibited less than 50% germination on fungicide-amended agar compared to the SHAM control.

Table 3.1. Primers and probes utilized in the detection of the G143A mutation associated with QoI resistance. Name Sequence (5’-3’) Purpose Source Primers Baudoin ARMS-SYBR forward CCTTGGTGACAAATGAGTTTTTGGAG G-143 detection WTa Baudoin ARMS-SYBR Baudoin et forward CCTTGGTGACAAATGAGTTTTTGGAC A-143 detection al. (2008) MTa Baudoin SYBR common common CAACTTCTTTTCCAATTAATGGGATAG reverse primer reverse Amplification 83F CGCTACAGACTGGGTCACTG of cytβ This Amplification manuscript 517R AGTCTCTTAGGGCCCCCATT of cytβ

Probes WT probea [FAM] AGCCTATGGGGTGCAACCGT [BHQ1] G-143 detection This MT probea [HEX] AGCCTATGGGCTGCAACCGT [BHQ1] A-143 detection manuscript aBold and underlined nucleotides note the binding location for the single nucleotide polymorphism Molecular determination of resistance. A competitive TaqMan quantitative PCR (qPCR) assay that targeted the G143A mutation associated with QoI resistance (Miles et al.,

2012) was developed by collaborator Tim Miles using mitochondrial DNA sequences from the National Center for Biotechnological Information and five complete E. necator sequences (Jones et al., 2014). To concurrently detect resistant and sensitive alleles, primers were paired with TaqMan probes, HEX to the resistant (A-143), and FAM to the sensitive (G-143, Table 3.1). To validate that this molecular tool was applicable to test field samples, samples from the Oregon E. necator survey were used. Samples were tested with the TaqMan assay and the phenotypes were compared to the genotyping results.

Results

E. necator single spore isolates showed two distinct responses to QoI fungicides. The isolates segregated into two distinct categories: resistant (n=41) and sensitive (n=6) to both trifloxystrobin and kresoxim-methyl (Table 3.2). Those that were resistant had EC50 values of ≥ 100µg/ml, the highest concentration tested, while susceptible isolates all exhibited an EC50 ≤ 1 µg/ml. Conidia germination on media amended with fungicide

Table 3.2. Origin and QoI sensitivity of Oregon Erysiphe necator isolates collected in 2015. Vineyard Region QoI QoI Trifloxystrobin Kresoxim-methyl Resistant Sensitive EC50 range (µg/ml) EC50 range (µg/ml) Sensitive Resistant Sensitive Resistant Willamette Valley 27 6 0.003- >100 0.018- >100 0.012 0.050 Southern Oregona 11 NA NA >100 NA >100 Columbia Gorgea 3 NA NA >100 NA >100 Total 41 6 aSensitive is not applicable (NA) because only resistant isolates were obtained from these regions concentrations up to 2mg/ml was examined for a subset of resistant isolates with no significant difference in conidia germination at any concentration (data not shown).

Molecular determination. The TaqMan qPCR assay and sequencing results were in

100% agreement with the phenotypic determination of resistance from the conidia germination bioassays. Both wild type and the G143A resistant allele were detected in all growing regions that were sampled (Figure 3.3, left) with >70% of isolates and samples testing positive for the G143A mutation. The bulk field samples from the Willamette

Figure 3.3. Determination of genotype using the TaqMan assay for field samples (left) and isolates (right). Red indicates resistant samples, green indicates sensitive samples, and blue indicates mixed samples as determined by the TaqMan assay.

Valley and Columbia Gorge had mixed and pure populations of both alleles, while the bulk field samples from southern Oregon had either resistant or mixed populations of alleles. Almost all isolates were determined to carry the G143A mutation by the TaqMan qPCR assay (Figure 3.3, right) which was confirmed by sequencing (data not shown).

Discussion

The single nucleotide polymorphism ‘G143A’ mutation was found in E. necator from all three of the main grape growing regions in Oregon. The seemingly ubiquitous presence of the G143A mutation could be through independent natural mutations in E. necator populations in each region. Spatially disparate natural mutations have been shown to be the origin of widespread QoI resistance in two diseases with similar

epidemics, Septoria leaf blotch of wheat and grapevine downy mildew [causal agents

Mycosphaerella graminicola and Plasmopara viticola, respectively (Chen et al., 2007;

Torriani et al., 2009)]. The widespread presence of the G143A mutation could also be the result of introduction through plant material infected with resistant E. necator populations. In Oregon, over the last decade there has been an average annual growth rate of 5.8% in vineyard area (USDA NASS, 2018). Due to the absence of a robust, certified grape nursery industry in Oregon, much of the material for new plantings comes from out-of-state nurseries. The interstate trade of plant material is a thriving industry and could have played a role in the appearance of QoI resistance in all three Oregon viticultural regions in 2015. Phylogenetic analyses on populations of E. necator in

Oregon and adjacent states could determine if this was the case.

The TaqMan assay was designed to detect the G143A mutation and did so in complete agreement with the phenotypic responses from the conidia germination bioassays. This assay detected low amounts (e.g.<5%) of wild type or mutant alleles in mixed samples. This could make it useful in understanding the dynamics of QoI resistance in E. necator field populations. In addition, this assay facilitates rapid testing of E. necator samples for the presence of G143A, circumventing the variability and tedious nature of biological based assays.

The application of QoI products exerts strong selection pressure on pathogen strains containing the G143A mutation to survive and reproduce, which could be the reason for the seemingly rapid appearance of field resistance to these products in Oregon.

In the presence of QoI products, mitochondria lacking G143A are likely killed quickly while mutated mitochondria remain, making selection for individuals with high

proportions of mutated mitochondria rapid (Gisi et al., 2002). The location of this mutation in the mitochondrial genome, which has a high mutation rate, and strong selection pressure in the presence of QoI products could explain the appearance of QoI resistance in numerous pathogens on a wide variety of crops across the world (Fungicide

Resistance Action Committee, 2012b; Gisi et al., 2002). While the dynamics of individual resistance mutations and selection thereof can be effectively parsed, in a field situation where mixtures and rotations of fungicides are used, the dynamics become cloudier.

In E. necator populations in Virginia, the Y136F mutation that contributes to resistance to demethylation inhibitor (DMI, FRAC 3) products was found in all QoI resistant isolates containing the G143A mutation (Rallos et al., 2014; Rallos and

Baudoin, 2016). Many of the vineyards surveyed in this project used DMI products in rotation with QoI products. Although we did not test for the presence of the Y136F mutation in this study, subsequent analysis on the isolates obtained in this survey determined that DMI resistance was present at levels exceeding legal use rates (Mahaffee and Neill, unpublished). The results from this study, and the information obtained from vineyard personnel suggest that QoI and DMI fungicide resistance may have played a role in the GPM management issues observed in some vineyards in the 2015 growing season.

QoI resistance is widespread in Oregon viticultural regions, and previous studies indicate that it is reasonable to expect resistant individuals to remain in the population, even in the absence of QoI selection pressure (Rallos et al., 2014). If QoI use is continued even with conventional resistance management strategies such as tank mixing, the

resistant allele may still be selected for (Genet et al., 2006). The qualitative nature of this resistance makes it such that relying on sensitivity shifts in E. necator in the absence of

QoI use is not likely to be a practical method increasing the sensitivity among resistant isolates (Rallos et al., 2014). More research into the dynamics of QoI resistance may elucidate management tactics that may enable continued use of this FRAC group. At present the most practical method to avoid GPM management problems is to refrain from using QoI products where resistant E. necator populations reside.

References

Ash, G., 2000. Downy mildew of grape [WWW Document]. Plant Health Instr. URL www.apsnet.org/edcenter/intropp/lessons/fungi/Oomycetes/Pages/DownyMildew Grape.aspx (accessed 1.29.18). Bartlett, D., Clough, J., Godwin, J., Hall, A., Hamer, M., Parr-Dobrzanski, B., 2002. The strobilurin fungicides. Pest Manag. Sci. 58, 649–662. Baudoin, A., Olaya, G., Delmotte, F., Colcol, J.F., Sierotzki, H., 2008. QoI resistance of Plasmopara viticola and Erysiphe necator in the Mid-Atlantic United States. Plant Health Prog. 22, 22–24. Bettiga, L.J., Gubler, W.D., Leavitt, G.M., 2013. Powdery Mildew, in: Bettiga, L.J. (Ed.), Grape Pest Management. University of California, pp. 137–149. Brent, K.J., Hollomon, D.W., 1995. Fungicide resistance in crop pathogens: How can it be managed?, 1st ed. International Group of National Associations of Agrochemical Manufacturers, Brussels. Brewer, M.T., Milgroom, M.G., 2010. Phylogeography and population structure of the grape powdery mildew fungus, Erysiphe necator, from diverse Vitis species. BMC Evol. Biol. 10, 268. Chen, W.-J., Delmotte, F., Cervera, S.R., Douence, L., Greif, C., Corio-Costet, M.-F., 2007. At least two origins of fungicide resistance in grapevine downy mildew populations. Appl. Environ. Microbiol. 73, 5162–5172. Estep, L.K., Zala, M., Anderson, N.P., Sackett, K.E., Flowers, M., McDonald, B.A., Mundt, C.C., 2013. First report of resistance to QoI fungicides in North American populations of Zymoseptoria tritici, causal agent of Septoria tritici blotch of wheat. Plant Dis. 97, 1511. Fungicide Resistance Action Committee, 2012a. QoI fungicides general information [WWW Document]. FRAC.info. URL http://www.frac.info/working-group/qol- fungicides Fungicide Resistance Action Committee, 2012b. List of pathogens with field resistance towards QoI fungicides [WWW Document]. FRAC.info. URL http://www.frac.info/docs/default-source/qoi-wg/qoi-quick-references/species- with-qo-resistance-(updated-2012).pdf?sfvrsn=c8db449a_4 Gadoury, D.M., Cadle-Davidson, L., Wilcox, W.F., Dry, I.B., Seem, R.C., Milgroom, M.G., 2012. Grapevine powdery mildew (Erysiphe necator): a fascinating system for the study of the biology, ecology and epidemiology of an obligate biotroph. Mol. Plant Pathol. 13, 1–16. Gadoury, D.M., Robert Seem, Andrea Ficke, Wayne Wilcox, 2003. Ontogenic resistance to powdery mildew in grape berries. Phytopathology 93, 547–555. Gadoury, D.M., Wilcox, W.F., Rumbolz, J., Gubler, W.D., 2015. Powdery Mildew, in: Wilcox, W.F., Gubler, W.D., Uyemoto, J.K. (Eds.), Compendium of Grape Diseases, Disorders, and Pests. APS Press, pp. 75–83. Genet, J.-L., Jaworska, G., Deparis, F., 2006. Effect of dose rate and mixtures of fungicides on selection for QoI resistance in populations of Plasmopara viticola. Pest Manag. Sci. 62, 188–194.

Gisi, U., Sierotzki, H., Cook, A., McCaffery, A., 2002. Mechanisms influencing the evolution of resistance to Qo inhibitor fungicides. Pest Manag. Sci. 58, 859–867. Heaney, S.P., Hall, A.A., Davies, S.A., Olaya, G., 2000. Resistance to fungicides in the Qol-STAR cross-resistance group: current perspectives., in: Proceedings of an International Conference Held at the Brighton Hilton Metropole Hotel, Brighton, UK, 13-16 November 2000. Presented at the The BCPC Conference: Pests and diseases, Volume 2, British Crop Protection Council, Brighton, U.K., pp. 755– 762. Hellman, E.W., 2003. Oregon Viticulture. Oregon State University Press. International Organization of Vine and Wine, 2017. OIV Statistical Report on World Viticulture. Jones, L., Riaz, S., Morales-Cruz, A., Amrine, K.C., McGuire, B., Gubler, W.D., Walker, M.A., Cantu, D., 2014. Adaptive genomic structural variation in the grape powdery mildew pathogen, Erysiphe necator. BMC Genomics 15, 1081. Lybbert, T.J., Magnan, N., Gubler, W.D., 2016. Multidimensional responses to disease information: How do winegrape growers react to powdery mildew forecasts and to what environmental effect? Am. J. Agric. Econ. 98, 383–405. https://doi.org/10.1093/ajae/aav097 Miles, L.A., Miles, T.D., Kirk, W.W., Schilder, A.M.C., 2012. Strobilurin (QoI) resistance in populations of Erysiphe necator on grapes in Michigan. Plant Dis. 96, 1621–1628. Miller, T.C., Gubler, W.D., 2004. Sensitivity of California isolates of Uncinula necator to trifloxystrobin and spiroxamine, and update on triadimefon sensitivity. Plant Dis. 88, 1205–1212. Morton, V., Staub, T., 2008. A Short History of Fungicides [WWW Document]. APSnet Featur. URL http://www.apsnet.org/publications/apsnetfeatures/Pages/Fungicides.aspx Quinn, J.A., Powell Jr, C.C., 1982. Effects of temperature, light, and relative humidity on powdery mildew of Begonia. Phytopathology 72, 480–484. R Core Team, 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Rallos, L.E.E., Baudoin, A.B., 2016. Co-occurrence of two allelic variants of CYP51 in Erysiphe necator and their correlation with over-expression for DMI resistance. PLoS ONE 11, 1–18. Rallos, L.E.E., Johnson, N.G., Schmale, D.G., Prussin, A.J., Baudoin, A.B., 2014. Fitness of Erysiphe necator with G143A-based resistance to quinone outside inhibitors. Plant Dis. 98, 1494–1502. Sambucci, O., Alston, J., Fuller, K., 2014. The costs of powdery mildew management in grapes and the value of resistant varieties: Evidence from California. Robert Mondavi Inst. Cent. Wine Econ. Working Paper 1402. Sambucci, O., Lawell, C.-Y.L., Lybbert, T.J., 2017. The spraying decisions of grape growers in response to disease forecasting information: A dynamic structural econometric model. Working paper, University of California at Davis.

Sierotzki, H., Wullschleger, J., Gisi, U., 2000. Point mutation in cytochrome b gene conferring resistance to strobilurin fungicides in Erysiphe graminis f. sp. tritici field isolates. Pestic. Biochem. Physiol. 68, 107–112. Skinkis, P.A., Pscheidt, J.W., Peachey, E., Dreves, A.J., Walton, V.M., Kaiser, C., 2017. 2017 Pest Management Guide for Wine Grapes in Oregon. Oregon State Extension Service. Stummer, B.E., Francis, I.L., Markides, A.J., Scott, E.S., 2003. The effect of powdery mildew infection of grape berries on juice and wine composition and on sensory properties of Chardonnay wines. Aust. J. Grape Wine Res. 9, 28–39. Torriani, S.F., Brunner, P.C., McDonald, B.A., Sierotzki, H., 2009. QoI resistance emerged independently at least 4 times in European populations of Mycosphaerella graminicola. Pest Manag. Sci. 65, 155–162. USDA NASS, 2018. NASS Quick Stats [WWW Document]. URL https://quickstats.nass.usda.gov/

Chapter 4

General Conclusions

Brent W. Warneke

This research sought to increase the efficiency of wine grape production in western Oregon by improving fungicidal product selection and application timing during bloom to manage Grape Powdery Mildew (GPM, causal agent Erysiphe necator). To accomplish this, field experiments examined fungicide application timing to manage cluster infections, and lab and field experiments examined the effect of fungicide redistribution on the development of E. necator. In addition, Oregon E. necator populations were examined for resistance to QoI fungicides.

Fungicide redistribution experiments showed that most fungicides investigated could redistribute via at least one mechanism, and some were shown to redistribute via mechanisms not previously demonstrated. Some fungicides reputed to redistribute via specific mechanisms (e.g quinoxyfen volatilization) had relatively low inhibition on E. necator infection and growth. All fungicides investigated showed a torus of E. necator growth inhibition in the vapor assay indicating redistribution, and most clusters covered during fungicide applications had reduced proportions of infected berries. Vapor redistribution of fungicides plays a role in the field efficacy of many fungicides, and amounts translocated via vapors may be higher than previously thought. Since there is limited data on the effect of fungicide redistribution on the growth of E. necator, this study expands our understanding of fungicide redistribution impacts on infection and growth of E. necator. Adaptation of the detached leaf bioassay could provide a cost- effective way to investigate other aspects of fungicide activity, such as the curative and residual effects of fungicide redistribution. Alteration of the environmental parameters during the detached leaf assay could provide information on the redistribution of fungicides more relevant to field scenarios.

Of the three application timings tested (inflorescence elongation, 50% bloom and end of bloom) fungicide applications at end of bloom were the most effective to reduce berry infection. In field experiments, probabilities of berry infection (i.e. proportions of infected berries) for applications of fluopyram, quinoxyfen, or trifloxystrobin initiated at end of bloom ranged from 0.020 to 0.14, and consisted of inconspicuous diffuse E. necator colonies. If viticulturists are using fluopyram, quinoxyfen, or trifloxystrobin, their effectiveness will likely be maximized if they are applied near the end of bloom and one interval thereafter. In combination with cultural practices and early season management, making two carefully timed synthetic fungicide applications from late bloom until 3-4 mm berries may provide an effective strategy for wine grape growers in western Oregon seeking to produce fruit with low disease levels and minimum synthetic chemical input. Reduced synthetic inputs would have the added benefit of a reduction in selection pressure for fungicide resistant E. necator strains and reduced fungicide pecuniary costs. Timing different fungicides to specific phenological growth stages could be an important aspect to improving disease management in grape and other production systems. Using phenological product efficacy information in combination with decision support models could improve fungicidal product selection through grapevine development and could further improve the efficiency of GPM management.

Quinone outside inhibitor (QoI) fungicide resistance was found to be widespread in Oregon viticultural regions in the fall of 2015. The results from the TaqMan qPCR assay designed to detect the G143A mutation known to confer resistance to QoI fungicides was in 100% agreement with the phenotypic determination of resistance from the conidia germination bioassays for all isolates tested. The qPCR assay and phenotypic

characterization of single chain isolates collected throughout Oregon respectively demonstrated that 87% of isolates contained the G143A mutation and could withstand high concentrations of two QoI products (EC50>100µg/ml). The qPCR assay could be further tested for use in combination with spore sampling technologies to monitor QoI resistant populations on scales from fields to growing regions. More research into the dynamics of QoI resistance may elucidate management tactics that may enable continued use of this FRAC group. At present the most practical method to avoid GPM management problems is to refrain from using QoI products where resistant E. necator populations reside.

Bibliography

Angelini, R.M.D.M., Masiello, M., Rotolo, C., Pollastro, S., Faretra, F., 2014. Molecular characterisation and detection of resistance to succinate dehydrogenase inhibitor fungicides in Botryotinia fuckeliana (Botrytis cinerea). Pest Manag. Sci. 70, 1884–1893.

Ash, G., 2000. Downy mildew of grape [WWW Document]. Plant Health Instr. URL www.apsnet.org/edcenter/intropp/lessons/fungi/Oomycetes/Pages/DownyMildew Grape.aspx (accessed 1.29.18).

Bartlett, D., Clough, J., Godwin, J., Hall, A., Hamer, M., Parr-Dobrzanski, B., 2002. The strobilurin fungicides. Pest Manag. Sci. 58, 649–662.

BASF Corporation, 2010. Vivando technical information bulletin [WWW Document]. URL http://agproducts.basf.us/products/vivandofungicide-.html?utm_source=vivandofungicide&utm_medium=offline&utm_cam paign=vivando

Baudoin, A., Olaya, G., Delmotte, F., Colcol, J.F., Sierotzki, H., 2008. QoI resistance of Plasmopara viticola and Erysiphe necator in the Mid-Atlantic United States. Plant Health Prog. 22, 22–24.

Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc., B (Methodological) 57, 289–300.

Bernard, M., Carew, M., Clarke, K., Davies, K., Deland, L., Emmett, B., Glenn, D., Hawtin, J., Hitch, C., Hoffman, A., Horne, P., Hurst, P., Jaeger, C., Magarey, C., Magarey, P., Reynolds, J., Rozario, S., Thomson, L., Wicks, T., 2003. Strategic use of sulphur in integrated pest and disease management (IPM) programs for grapevines (Final Report).

Bettiga, L.J., Gubler, W.D., Leavitt, G.M., 2013. Powdery Mildew, in: Bettiga, L.J. (Ed.), Grape Pest Management. University of California, pp. 137–149.

Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H., White, J.-S.S., 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol. 24, 127–135.

Bosch, F. van den, Oliver, R., Berg, F. van den, Paveley, N., 2014. Governing Principles Can Guide Fungicide-Resistance Management Tactics. Annu. Rev. Phytopathol. 52, 175–195.

Bradley, C., Hollier, C., Kelly, H., 2016. Principles of Fungicide Resistance [WWW Document]. Plant Manag. Netw. URL https://www.plantmanagementnetwork.org/hub/soyfungicideresistance/files/Fungi cideResistance.pdf

Brent, K.J., Hollomon, D.W., 1995. Fungicide resistance in crop pathogens: How can it be managed?, 1st ed. International Group of National Associations of Agrochemical Manufacturers, Brussels.

Brewer, M.T., Milgroom, M.G., 2010. Phylogeography and population structure of the grape powdery mildew fungus, Erysiphe necator, from diverse Vitis species. BMC Evol. Biol. 10, 268.

Casado, C.G., Heredia, A., 2001. Ultrastructure of the cuticle during growth of the grape berry (Vitis vinifera L.). Physiol. Plant. 111, 220–224.

Chatrath, M.S., Lyda, S.D., Lauchli, A., 1972. Translocation of 2-(4-Thiazolyl) benzimidazole in maturing cotton plants. Phytopathology 62, 1410–1414.

Chen, W.-J., Delmotte, F., Cervera, S.R., Douence, L., Greif, C., Corio-Costet, M.-F., 2007. At least two origins of fungicide resistance in grapevine downy mildew populations. Appl. Environ. Microbiol. 73, 5162–5172.

Colcol, J.F., Baudoin, A.B., 2015. Sensitivity of Erysiphe necator and Plasmopara viticola in Virginia to QoI Fungicides, Boscalid, Quinoxyfen, Thiophanate Methyl, and Mefenoxam. Plant Dis. 100, 337–344.

Cools, H.J., Hawkins, N.J., Fraaije, B.A., 2013. Constraints on the evolution of azole resistance in plant pathogenic fungi. Plant Pathol. 62, 36–42.

Cowgill, W., Rosenberger, D., 2013. Not understanding phytotoxicity can damage your bottom line [WWW Document]. UMass Fruit Advis. URL http://umassfruitnotes.com/v78n3/a5.pdf

Davidse, L.C., 1986. Benzimidazole fungicides: Mechanism of action and biological impact. Annu. Rev. Phytopathol. 24, 43–65.

Deising, H.B., Reimann, S., Pascholati, S.F., 2008. Mechanisms and significance of fungicide resistance. Braz. J. Microbiol. 39, 286–295.

Deliere, L., Miclot, A.S., Sauris, P., Rey, P., Calonnec, A., 2010. Efficacy of fungicides with various modes of action in controlling the early stages of an Erysiphe necator-induced epidemic. Pest Manag. Sci. 66, 1367–1373.

Doulia, D.S., Anagnos, E.K., Liapis, K.S., Klimentzos, D.A., 2017. Effect of clarification process on the removal of pesticide residues in white wine. Food Control 72, 134– 144.

Edgington, L.V., 1981. Structural Requirements of Systemic Fungicides. Annu. Rev. Phytopathol. 19, 107–124.

Ehr, R.J., Kemmitt, G., Ouimette, D., Guemas, J.C., Costales, M., Brewster, K., Adelfinskaya, Y., 2008. Periodic Table of the Fungicides, 3rd ed. Dow Agrosciences.

Erickson, E.O., Wilcox, W.F., 1997. Distributions of sensitivities to three sterol demethylation inhibitor fungicides among populations of Uncinula necator sensitive and resistant to triadimefon. Phytopathology 87, 784–791.

Estep, L.K., Zala, M., Anderson, N.P., Sackett, K.E., Flowers, M., McDonald, B.A., Mundt, C.C., 2013. First report of resistance to QoI fungicides in North American populations of Zymoseptoria tritici, causal agent of Septoria tritici blotch of wheat. Plant Dis. 97, 1511.

Ficke, A., Gadoury, D.M., Seem, R.C., 2002. Ontogenic resistance and plant disease management: A case study of grape powdery mildew. Phytopathology 92, 671– 675.

Frenkel, O., Cadle-Davidson, L., Wilcox, W.F., Milgroom, M.G., 2014. Mechanisms of resistance to an azole fungicide in the grapevine powdery mildew fungus, Erysiphe necator. Phytopathology 105, 370–377.

Fungicide Resistance Action Committee, 2018. FRAC Code List 2018: Fungicides sorted by mode of action [WWW Document]. FRAC.info. URL http://www.frac.info/docs/default-source/publications/frac-code- list/frac_code_list_2018-final.pdf?sfvrsn=6144b9a_2

Fungicide Resistance Action Committee, 2017. Benzimidazole MOA and Mechanism(s) of Resistance [WWW Document]. FRAC.info. URL http://www.frac.info/expert- fora/benzimidazoles/soa-and-mechanism(s)-of-resistance

Fungicide Resistance Action Committee, 2013. Benzimidazole resistance risk and current status [WWW Document]. FRAC.info. URL http://www.frac.info/expert- fora/benzimidazoles/resistance-risk-and-current-status

Fungicide Resistance Action Committee, 2012a. QoI fungicides general information [WWW Document]. FRAC.info. URL http://www.frac.info/working-group/qol- fungicides

Fungicide Resistance Action Committee, 2012b. List of pathogens with field resistance towards QoI fungicides [WWW Document]. FRAC.info. URL http://www.frac.info/docs/default-source/qoi-wg/qoi-quick-references/species- with-qo-resistance-(updated-2012).pdf?sfvrsn=c8db449a_4

Gadoury, D.M., Cadle-Davidson, L., Wilcox, W.F., Dry, I.B., Seem, R.C., Milgroom, M.G., 2012. Grapevine powdery mildew (Erysiphe necator): a fascinating system for the study of the biology, ecology and epidemiology of an obligate biotroph. Mol. Plant Pathol. 13, 1–16.

Gadoury, D.M., Robert Seem, Andrea Ficke, Wayne Wilcox, 2003. Ontogenic resistance to powdery mildew in grape berries. Phytopathology 93, 547–555.

Gadoury, D.M., Seem, R.C., Ficke, A., Wilcox, W.F., 2001. The epidemiology of powdery mildew on Concord grapes. Phytopathology 91, 948–955.

Gadoury, D.M., Seem, R.C., Wilcox, W.F., Henick-Kling, T., Conterno, L., Day, A., Ficke, A., 2007. Effects of diffuse colonization of grape berries by Uncinula necator on bunch rots, berry microflora, and juice and wine quality. Phytopathology 97, 1356–1365.

Gadoury, D.M., Wilcox, W.F., Rumbolz, J., Gubler, W.D., 2015. Powdery Mildew, in: Wilcox, W.F., Gubler, W.D., Uyemoto, J.K. (Eds.), Compendium of Grape Diseases, Disorders, and Pests. APS Press, pp. 75–83.

Gee, C.T., Gadoury, D.M., Cadle-Davidson, L., 2008. Ontogenic resistance to Uncinula necator varies by genotype and tissue type in a diverse collection of Vitis spp. Plant Dis. 92, 1067–1073.

Genet, J.-L., Jaworska, G., Deparis, F., 2006. Effect of dose rate and mixtures of fungicides on selection for QoI resistance in populations of Plasmopara viticola. Pest Manag. Sci. 62, 188–194.

Gent, D.H., Grove, G.G., Nelson, M.E., Wolfenbarger, S.N., Woods, J.L., 2014. Crop damage caused by powdery mildew on hop and its relationship to late season management. Plant Pathol. 63, 625–639.

Gianessi, L.P., Reigner, N., 2005. The value of fungicides in U.S. crop production. Croplife Foundation, Washington, D. C.

Gisi, U., Sierotzki, H., Cook, A., McCaffery, A., 2002. Mechanisms influencing the evolution of resistance to Qo inhibitor fungicides. Pest Manag. Sci. 58, 859–867.

Glattli, A., Grote, T., Stammler, G., 2011. SDH-inhibitors: History, biological performance and molecular mode of action, in: Dehne, H.W., Deising, H.B., Gisi,

U., Kuck, K.H., Russell, P.E., Lyr, H. (Eds.), Modern Fungicides and Antifungal Compounds VI. Presented at the 16th International Reinhardsbrunn Symposium, Deutsche Phytomedizinische Gesellschaft eV Selbstverlag, Friedrichroda, Germany, pp. 159–170.

Graf, S., 2018. Characterisation of metrafenone and succinate dehydrogenase inhibitor resistant isolates of the grapevine powdery mildew Erysiphe necator.

Grimmer, M.K., van den Bosch, F., Powers, S.J., Paveley, N.D., 2015. Fungicide resistance risk assessment based on traits associated with the rate of pathogen evolution. Pest Manag. Sci. 71, 207–215.

Gubler, W.D., Ypema, H.L., Ouimette, D.G., Bettiga, L.J., 1996. Occurrence of resistance in Uncinula necator to triadimefon, myclobutanil, and fenarimol in California grapevines. Plant Dis. 80, 902.

Haramoto, M., Yamanaka, H., Hosokawa, H., Sano, H., Sano, S., Otani, H., 2006. Control efficacy of cyflufenamid in the field and its fungicidal properties. J. Pestic. Sci. 31, 116–122.

Harrison, X.A., 2014. Using observation-level random effects to model overdispersion in count data in ecology and evolution. PeerJ 2, e616.

Hartig, F., 2016. DHARMa: Residual diagnostics for hierarchical (multi-level/mixed) regression models.

Heaney, S.P., Hall, A.A., Davies, S.A., Olaya, G., 2000. Resistance to fungicides in the Qol-STAR cross-resistance group: current perspectives., in: Proceedings of an International Conference Held at the Brighton Hilton Metropole Hotel, Brighton, UK, 13-16 November 2000. Presented at the The BCPC Conference: Pests and diseases, Volume 2, British Crop Protection Council, Brighton, U.K., pp. 755– 762.

Hellman, E.W., 2003. Oregon Viticulture. Oregon State University Press.

Henry, M.J., 2003. Fungicides, quinoxyfen. Encycl. Agrochem. 872–873.

Hoffman, L.E., Wilcox, W.F., Gadoury, D.M., Seem, R.C., Riegel, D.G., 2004. Integrated control of grape black rot: Influence of host phenology, inoculum availability, sanitation, and spray timing. Phytopathology 94, 641–650.

Hulin, L., Rouas, G., Buschhaus, H., 2009. Cyflufenamid, a new compound for control of powdery mildew in grape., in: Association Française de Protection Des Plantes, 9ème Conférence International Sur Les Maladies Des Plantes, Tours, France, 8 et

9 Décembre 2009. Association Française de Protection des Plantes (AFPP), pp. 791–795.

International Organization of Vine and Wine, 2017. OIV Statistical Report on World Viticulture.

Ishii, H., Miyamoto, T., Ushio, S., Kakishima, M., 2011. Lack of cross-resistance to a novel succinate dehydrogenase inhibitor, fluopyram, in highly boscalid-resistant isolates of Corynespora cassiicola and Podosphaera xanthii. Pest Manag. Sci. 67, 474–482.

John Dunham & Associates, 2017. The wine industry boosts the American economy by $219.9 billion in 2017 [WWW Document]. URL http://wineamerica.org/wp- content/themes/wineamerica/pdfs/impact/United-States-Report.pdf

Johnson, K.B., Sawyer, T.L., Powelson, M.L., 1994. Frequency of benzimidazole- and dicarboximide-resistant strains of Botrytis cinerea in western Oregon small fruit and snap bean plantings. Plant Dis. 78, 572.

Jones, L., Riaz, S., Morales-Cruz, A., Amrine, K.C., McGuire, B., Gubler, W.D., Walker, M.A., Cantu, D., 2014. Adaptive genomic structural variation in the grape powdery mildew pathogen, Erysiphe necator. BMC Genomics 15, 1081.

Kast, W.K., Bleyer, K., 2011. Efficacy of sprays applied against powdery mildew (Erysiphe necator) during a critical period for infections of clusters of grapevines (Vitis vinifera). J. Plant Pathol. 93, S29–S32.

Klausener, A., Raming, K., Stenzel, K., 2007. Modern tools for drug discovery in agricultural research, in: Ohkawa, H., Miyagawa, H., Lee, P.W. (Eds.), Pesticide Chemistry: Crop Protection, Public Health, Environmental Safety. John Wiley & Sons, pp. 55–63.

Klittich, C.J., 2014. Fungicide mobility and the influence of physical properties, in: Myung, K., Satchivi, N.M., Kingston, C.K. (Eds.), Retention, Uptake, and Translocation of Agrochemicals in Plants. American Chemical Society, pp. 95– 109.

Klittich, C.J.R., Ray, S.L., 2013. Effects of physical properties on the translaminar activity of fungicides. Pestic. Biochem. Physiol. 107, 351–359.

Koblet, W., Candolfi-Vasconcelos, M.C., Zweifel, W., Howell, G.S., 1994. Influence of leaf removal, rootstock, and training system on yield and fruit composition of Pinot noir grapevines. Am. J. Enol. Vitic. 45, 181–187.

Kunova, A., Pizzatti, C., Bonaldi, M., Cortesi, P., 2016. Metrafenone resistance in a population of Erysiphe necator in northern Italy. Pest Manag. Sci. 72, 398–404.

Kwok, I.M.-Y., Loeffler, R.T., 1993. The biochemical mode of action of some newer azole fungicides. Pestic. Sci. 39, 1–11.

KyoungHee, S., InMyung, C., KyuHwan, C., JaeMan, L., HeeSeung, P., 2010. Morphological development and chemical composition of epicuticular wax crystals in “Campbell Early” grape. Hortic. Environ. Biotechnol. 51, 253–256.

Latin, R., 2011. A practical guide to turfgrass fungicides. American Phytopathological Society Press.

Lee, S., Gustafson, G., Skamnioti, P., Baloch, R., Gurr, S., 2008. Host perception and signal transduction studies in wild-type Blumeria graminis f. sp. hordei and a quinoxyfen-resistant mutant implicate quinoxyfen in the inhibition of serine esterase activity. Pest Manag. Sci. 64, 544–555.

Lenth, R.V., 2016. Least-Squares Means: The R Package lsmeans. J. Stat. Softw. 69, 1– 33.

Leroux, P., Fritz, R., Debieu, D., Albertini, C., Lanen, C., Bach, J., Gredt, M., Chapeland, F., 2002. Mechanisms of resistance to fungicides in field strains of Botrytis cinerea. Pest Manag. Sci. 58, 876–888.

Lorenz, D. h., Eichhorn, K. w., Bleiholder, H., Klose, R., Meier, U., Weber, E., 1995. Growth stages of the grapevine: Phenological growth stages of the grapevine (Vitis vinifera L. ssp. vinifera)—Codes and descriptions according to the extended BBCH scale. Aust. J. Grape Wine Res. 1, 100–103.

Lybbert, T.J., Magnan, N., Gubler, W.D., 2016. Multidimensional responses to disease information: How do winegrape growers react to powdery mildew forecasts and to what environmental effect? Am. J. Agric. Econ. 98, 383–405. https://doi.org/10.1093/ajae/aav097

Ma, Z., Proffer, T.J., Jacobs, J.L., Sundin, G.W., 2006. Overexpression of the 14α- demethylase target gene (CYP51) mediates fungicide resistance in Blumeriella jaapii. Appl. Environ. Microbiol. 72, 2581–2585.

Mendiburu, F., 2016. agricolae: Statistical Procedures for Agricultural Research.

Miles, L.A., Miles, T.D., Kirk, W.W., Schilder, A.M.C., 2012. Strobilurin (QoI) resistance in populations of Erysiphe necator on grapes in Michigan. Plant Dis. 96, 1621–1628.

Miller, T.C., Gubler, W.D., 2004. Sensitivity of California isolates of Uncinula necator to trifloxystrobin and spiroxamine, and update on triadimefon sensitivity. Plant Dis. 88, 1205–1212.

Morton, V., Staub, T., 2008. A Short History of Fungicides [WWW Document]. APSnet Featur. URL http://www.apsnet.org/publications/apsnetfeatures/Pages/Fungicides.aspx

Moyer, M., Grove, G., 2012. Powdery mildew in eastern Washington commercial grape production: Biology and disease management [WWW Document]. Wash. State Univ. Ext. URL http://extension.wsu.edu/publications/wp- content/uploads/sites/54/publications/em058e.pdf

Müller, C., Riederer, M., 2005. Plant surface properties in chemical ecology. J. Chem. Ecol. 31, 2621–2651.

Nelson, M.E., Gent, D.H., Grove, G.G., 2014. Meta-analysis reveals a critical period for management of powdery mildew on hop cones. Plant Dis. 99, 632–640.

Nobel, P.S., 1975. Effective thickness and resistance of the air boundary layer adjacent to spherical plant parts. J. Exp. Bot. 26, 120–130.

Oerke, E.-C., 2006. Crop losses to pests. J. Agric. Sci. 144, 31–43.

Opalski, K.S., Tresch, S., Kogel, K.-H., Grossmann, K., Köhle, H., Hückelhoven, R., 2006. Metrafenone: studies on the mode of action of a novel cereal powdery mildew fungicide. Pest Manag. Sci. 62, 393–401.

Pearson, R.C., Riegel, D.G., Gadoury, D.M., 1994. Control of powdery mildew in vineyards using single-application vapor-action treatments of triazole fungicides. Plant Dis. 78, 164–168.

Philips McDougall, 2016. The Cost of New Agrochemical Product Discovery, Development and Registration in 1995, 2000, 2005-8 and 2010 to 2014.

Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., R Core Team, 2016. nlme: Linear and Nonlinear Mixed Effects Models.

Piqueras, C.M., Herrera, D., Latorre, B.A., 2014. First report of high boscalid resistance in Botrytis cinerea associated with the H272L mutation in grapevine in Chile. Plant Dis. 98, 1441–1441.

Pirondi, A., Nanni, I.M., Brunelli, A., Collina, M., 2014. First report of resistance to cyflufenamid in Podosphaera xanthii, causal agent of powdery mildew, from

melon and zucchini fields in Italy. Plant Dis. 98, 1581–1581. https://doi.org/10.1094/PDIS-02-14-0210-PDN

Pscheidt, J., 2017. Fungicide resistance and fungicide groups, in: Pscheidt, J.W., Ocamb, C.M. (Eds.), Pacific Northwest Plant Disease Management Handbook [Online]. Corvallis, OR: Oregon State University.

Quinn, J.A., Powell Jr, C.C., 1982. Effects of temperature, light, and relative humidity on powdery mildew of Begonia. Phytopathology 72, 480–484.

R Core Team, 2016. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

Rallos, L.E.E., Baudoin, A.B., 2016. Co-occurrence of two allelic variants of CYP51 in Erysiphe necator and their correlation with over-expression for DMI resistance. PLoS ONE 11, 1–18.

Rallos, L.E.E., Johnson, N.G., Schmale, D.G., Prussin, A.J., Baudoin, A.B., 2014. Fitness of Erysiphe necator with G143A-based resistance to quinone outside inhibitors. Plant Dis. 98, 1494–1502.

Rosenquist, J.K., Morrison, J.C., 1988. The development of the cuticle and epicuticular wax of the grape berry. Vitis 27, 63–70.

Russell, P.E., 2006. The development of commercial disease control. Plant Pathol. 55, 585–594.

Sambucci, O., Alston, J., Fuller, K., 2014. The costs of powdery mildew management in grapes and the value of resistant varieties: Evidence from California. Robert Mondavi Inst. Cent. Wine Econ. Working Paper 1402.

Sambucci, O., Lawell, C.-Y.L., Lybbert, T.J., 2017. The spraying decisions of grape growers in response to disease forecasting information: A dynamic structural econometric model. Working paper, University of California at Davis.

Sano, S., Kasahara, I., Yamanaka, H., 2007. Development of a novel fungicide, cyflufenamid. J. Pestic. Sci. 32, 137–138.

Shavrukov, Y.N., Dry, I.B., Thomas, M.R., 2004. Inflorescence and bunch architecture development in Vitis vinifera L. Aust. J. Grape Wine Res. 10, 116–124. https://doi.org/10.1111/j.1755-0238.2004.tb00014.x

Skinkis, P.A., Pscheidt, J.W., Peachey, E., Dreves, A.J., Walton, V.M., Kaiser, C., 2017. 2017 Pest Management Guide for Wine Grapes in Oregon. Oregon State Extension Service.

Stammler, G., Semar, M., Strobel, D., 2014. Resistance management of metrafenone in powdery mildews., in: Dehne, H.W., Deising, H.B., Fraaije, B., Gisi, U., Hermann, D., Mehl, A., Oerke, E.C., Russell, P.E., Stammler, G., Kuck, K.H., Lyr, H. (Eds.), Modern Fungicides and Antifungal Compounds VII. Presented at the 17th International Reinhardsbrunn Symposium, Deutsche Phytomedizinische Gesellschaft eV Selbstverlag, Friedrichroda, Germany, pp. 179–184.

Strayer, P., 2017. The Organic Opportunity: Will the U.S. Wine Industry Miss Out? [WWW Document]. Wines Vines. URL http://www.winesandvines.com/columns_article/178258 (accessed 12.13.17).

Stummer, B.E., Francis, I.L., Markides, A.J., Scott, E.S., 2003. The effect of powdery mildew infection of grape berries on juice and wine composition and on sensory properties of Chardonnay wines. Aust. J. Grape Wine Res. 9, 28–39.

Summit Agro USA, 2017. Prolivo Sell Sheet [WWW Document]. URL http://www.summitagro- usa.com/sites/555def9f2c1cc41147001d6d/content_entry59b846102c1cc47bd41d d23e/59b846942c1cc47bd41dd553/files/PROLIVO_Sell_Sheet.pdf?1518014400

Thiessen, L.D., Neill, T.M., Mahaffee, W.F., 2017. Timing fungicide application intervals based on airborne Erysiphe necator concentrations. Plant Dis. 101, 1246–1252. https://doi.org/10.1094/PDIS-12-16-1727-RE

Torriani, S.F., Brunner, P.C., McDonald, B.A., Sierotzki, H., 2009. QoI resistance emerged independently at least 4 times in European populations of Mycosphaerella graminicola. Pest Manag. Sci. 65, 155–162.

Tweedy, B.G., 1981. Inorganic sulfur as a fungicide, in: Residue Reviews. Springer, New York, NY, pp. 43–68.

USDA NASS, 2018. NASS Quick Stats [WWW Document]. URL https://quickstats.nass.usda.gov/

Vincelli, P., 2002. QoI (Strobilurin) Fungicides: Benefits and Risks. [WWW Document]. Plant Health Instr. URL http://www.apsnet.org/edcenter/advanced/topics/Pages/StrobilurinFungicides.asp x

Warneke, B., Thiessen, L., Neill, T., Mahaffee, W., 2017. The Grape Powdery Mildew Conundrum: Fungicide Selection and Timing, in: Moyer, M., Mahaffee, W.F. (Eds.), Proceedings of the 8th International Workshop on Grapevine Downy and Powdery Mildew. Presented at the 8th International Workshop on Grapevine Downy and Powdery Mildew, Corvallis, OR.

Warneke, B., Yamagata, J., Neill, T., Miles, T., Mahaffee, W., 2016. Detection of quinone outside inhibitor resistant isolates of Erysiphe necator in Oregon vineyards. Phytopathology, p. S71.

Warnes, G., Bolker, B., Lumley, T., Johnson, R., 2015. gmodels: Various R Programming Tools for Model Fitting.

Wheeler, I.E., Hollomon, D.W., Gustafson, G., Mitchell, J.C., Longhurst, C., Zhang, Z., Gurr, S.J., 2003. Quinoxyfen perturbs signal transduction in barley powdery mildew (Blumeria graminis f.sp. hordei). Mol. Plant Pathol. 4, 177–186.

Wilcox, W.F., Riegel, D.G., 2010. Efficacy of new fungicides for control of powdery mildew (Erysiphe necator) and downy mildew (Plasmopara viticola) of grapes. Phytopathology, p. S136.

Wise, J.C., Jenkins, P.E., Schilder, A.M.C., Vandervoort, C., Isaacs, R., 2010. Sprayer type and water volume influence pesticide deposition and control of insect pests and diseases in juice grapes. Crop Prot. 29, 378–385.

Wolf, T., 2005. Viticulture Notes [WWW Document]. Va. Coop. Ext. URL http://www.sites.ext.vt.edu/newsletter- archive/viticulture/05julyaugust/05july.html

Wolters, A., Leistra, M., Linnemann, V., Klein, M., Schäffer, A., Vereecken, H., 2004. Pesticide volatilization from plants: Improvement of the PEC model PELMO based on a boundary-layer concept. Environ. Sci. Technol. 38, 2885–2893.

Appendices

100

Appendix A:

2017 Commercial Phenological Application Trial

Objective: To test the commercial implementation of Quintec and Luna Privilege fungicide applications applied at end of bloom into a commercial grape powdery mildew

(GPM) spray plan.

Figure A1: Maps of commercial trial sites at vineyard 1 (left) and vineyard 2 (right).

Rationale: Bloom is known to be a critical time to manage grape powdery mildew (Ficke et al., 2002; Gadoury et al., 2003). Recommendations are to apply the most efficacious products at the tightest application intervals during bloom (Moyer and Grove, 2012;

Skinkis et al., 2017). Vineyard managers typically make sprays during the window of

101

bloom, but the phenological timing of these applications varies. Small plot research in

2015 and 2016 determined that phenological timing Quintec and Luna privilege applications to end of bloom was an effective strategy to reduce berry infection.

Phenologically timing fungicide applications to end of bloom could help standardize bloom applications with potential of increasing the reliability of fungicide programs across growing seasons. Phenologically timing applications to a time when the chemistry is most effective could provide the most efficient means of disease control on grape clusters, potentially alleviating the need for tight spray intervals during bloom.

Methods

Replicated trails were established at two commercial vineyards in the Willamette

Valley as shown in Figure A1. Vineyard 1 was 12 acres of Chardonnay trained to Geneva double curtain with bilateral cane pruning. Vineyard 2 was 15 acres of Chardonnay with

Guyot training with bilateral cane pruning with vertical shoot positioning; the northwest corner of the trial area was bilateral cordon trained and spur pruned.

102

Quintec (active ingredient: quinoxyfen, Dow Agrosciences, Indianapolis, IN) and

Luna Privilege (active ingredient: fluopyram, Bayer Cropscience,

Research Triangle Park, NC) treatments at end of bloom were applied by vineyard personnel using their standard equipment. Figure A2. Chardonnay inflorescence phenology at the time of experimental applications. Left: vineyard one, Vineyard 1 made the end of bloom photo taken 6/30/17. Right: vineyard 2, photo taken 6/29/17. application on 6/26/17 when the trial area was at BBCH 68 (80% of flower-hoods fallen, Figure A2). Vineyard 2 made the end of bloom application on 6/26/17 when the trial area was at BBCH 73 (berries groat sized, Figure A2). In vineyard 1, applications of Luna Privilege and Quintec were made at 292 ml/ha. In vineyard 2 applications of Luna Privilege and Quintec were made at 497 ml and 438 ml/ha, respectively. At vineyard 2, Microthiol (sulfur), Scala (pyrimethanil), and Spreader 90 (alkylphenol ethoxylate, isopropanol, fatty acids) were included in the end of bloom applications at 2.24kg, 1.26kg, and 1.17L per hectare, respectively, in addition to Luna Privilege or Quintec. In vineyard 2, by choice of the vineyard manager, the same spray mixture applied to Quintec blocks was applied to the control block at the time of experimental applications.

Disease data was collected from five groups of 10 vines evenly space long the length of each replication (Figure A1). Leaf incidence was determined weekly prior to bloom and bi-weekly after bloom by examining 10 arbitrarily selected leaves four to seven nodes from the growing tip per plant. To determine berry incidence, one cluster

103

was collected from each vine in a section just as véraison was beginning and frozen at -20

C until rated for disease. Berries were removed from each cluster and 25 berries per cluster were arbitrarily selected to examine microscopically for disease incidence.

Leaf incidence values were used to calculate area under disease progress curves

(AUDPC) with the R package agricolae (Mendiburu, 2016). The number of infected berries in each cluster was divided by the number of berries examined to calculate a proportion of infected berries. AUDPC values and cluster infection proportions were compared with a Kruskal-Wallis test or Welch t-test if three or two treatments were compared, respectively. The control block was excluded from analysis in vineyard 2 to allow for a more direct comparison between Luna Privilege and Quintec treatments. Two separate comparisons were conducted in vineyard 2 to account for a disease hotspot observed in the area trained to bilateral cordon: one with all four of the treatment blocks included in their entirety and one with all four of the treatment blocks but with the bilateral cordon portion excluded.

Results

In vineyard 1, there were no significant differences among treatments for both leaf and fruit data. Out of 250 clusters collected only six had any GPM observed.

104

When all four of the treatment blocks were included in totality in vineyard 2, there was moderate evidence for a difference in mean AUDPC values between the Luna

Privilege and Quintec treatments (Welch t- test on 156.5 df, P =

0.01). When the bilateral cordon portion of the Luna 1 block was removed from the comparison there was no Figure A3. Berry infection proportion comparisons and mean berry evidence for a infection probabilities for the Luna and Quintec treatment areas at vineyard 2. Dots represent the mean difference in berry infection between the Luna treatment and the Quintec treatment, calculated as difference in mean Luna P – Quintec P and error bars represent 95% confidence intervals for the difference. Dashed lines indicate probabilities of berry infection AUDPC values for Luna (w/ codon (upper) and w/o cordon(lower)) and dotted line indicates probability of berry infection for Quintec treatment. between the Luna and Quintec treatments (Welch t-test on 128.4 df, P = 0.40). When all four of the treatment blocks were included in totality, there was moderate evidence for a difference in mean probability of berry infection between the Luna and Quintec treatments (Welch t-test on 139.62 DF, P = 0.02, Figure A3). The difference in mean berry infection probability was 0.05, signifying that on average berries had a 5% higher probability of infection in the Luna Privilege treatment area than in the Quintec treatment area (Figure

A3). When the bilateral cordon portion of the Luna 1 block was removed from the comparison there was no evidence for a difference in mean probability of berry infection

105

between the Luna and Quintec treatments (Welch t-test on 126.76 DF, P = 0.71, Figure

A3).

Discussion:

Implementing end of bloom applications of two commercially available fungicides as the only modification to two season-long commercial spray programs did not result in differences in disease levels among treatments in this study. There were generally low levels of disease observed in this study, making the interpretation of statistical comparisons on disease levels uninformative to the efficacy of different fungicide treatments.

The similarities in fungicide treatment efficacy and application programs could explain some of the similarly low disease levels observed among treatments in this study. Quintec and

Luna Privilege have both Figure A4. Probabilities of berry infection at collection points in the trial area at vineyard 2. Numbers are probabilities of berry been shown to be effective at infection and are positioned approximately over the location where 10 clusters were collected to generate that probability. controlling GPM, and both Yellow outline indicates the area trained to bilateral cordon. have similar redistribution profiles (Colcol and Baudoin, 2015; Ehr et al., 2008; Henry,

106

2003; Wilcox and Riegel, 2010). In addition, the timing of all end of bloom applications were similar. At both vineyards end of bloom was at the time of a calendar spray application so the control blocks received a spray at end of bloom like the treatment blocks. In a 12-year study, three sprays timed to inflorescence elongation, mid-bloom, and 2mm diameter berries always had similar disease levels to season long spray programs containing seven sprays (Kast and Bleyer, 2011). In both vineyards, applications before the experimental end of bloom application were circa inflorescence elongation and applications after were circa 2-4mm diameter berries and all blocks had season-long fungicide programs, which likely contributed to the low levels of GPM observed. In addition to the characteristics of the spray programs applied, the viticultural characteristics of the vineyards appeared to influence disease levels.

The training systems used likely played a role in the patterns of disease observed.

Cleistothecia collect on bark and trunk areas as they fall or get washed off tissues at the end of the season (Gadoury et al., 2012). Cordon trained systems have a larger trunk area than cane pruned systems resulting in a higher probability of intercepting cleistothecia from leaves at the end of the growing season (Koblet et al., 1994). In vineyard 2, the area trained to bilateral cordon was the first place that disease was observed and had the highest proportion of infected berries; in the cane pruned areas of both vineyards, disease was observed sporadically, with no apparent hotspots, except that adjacent to the cordon area in vineyard 2 (Figure A4). Higher rates of cleistothecia deposition and early season ascosporic infections in cordon trained areas likely explain some of the elevated levels of disease in the bilateral cordon area of vineyard 2 (Gadoury et al., 2012).

107

Previous studies have indicated that phenological applications may provide similar GPM control to season-long programs and could allow for reductions in the number of applications (Gadoury et al., 2003; Kast and Bleyer, 2011). Further testing of synthetic fungicide applications applied in the window of end of bloom and early berry development on a commercial scale could determine to what level application frequency could be reduced to produce fruit with negligible disease levels.

108

Appendix B:

Minimum Inhibitory Concentration Experiment

Objective: To determine the lowest concentration of fungicide needed to inhibit grape

powdery mildew (GPM) from colonizing grape foliar tissue. This concentration was

termed the minimum inhibitory concentration (MIC).

Rationale: Determining the lowest amount of fungicide needed to inhibit GPM from

infecting could enable calculations of the amount of redistribution for a given amount of

fungicide. This could potentially be used to approximate the amount of active ingredient

that translocated in the detached leaf fungicide mobility experiments, or what quantity of

fungicide is needed to achieve optimal coverage given spray and canopy characteristics.

Methods: Table B1. Fungicides and rates used in the minimum inhibitory concentration study. Formulated fungicides were Quantity Active Field AI Trade Name formulated Ingredient (AI) concentrationb product/haa mixed at the field rate Flint Trifloxystrobin 0.14kg 150 Luna (Table B1) and 10-fold Fluopyram 292ml 311 Privilege Quintec Quinoxyfen 292ml 156 serial diluted in nonsterile Toledo Tebuconazole 0.28kg 270 Microthiol tap water, then 5μl drops to Sulfur 3.36kg 5752 Disperss aRecommended quantities determined from fungicide labels. leaves with two bThe field concentration reflects the formulated quantity/ hectare mixed in 467L water (50gal/A) in mg/L. concentrations were applied

to every leaf (Figure B1). Fungicides were applied in areas without major veins to

109

minimize active ingredient translocation. There were three replicate leaves per treatment. Leaves were inoculated as in the detached leaf redistribution experiments (Chapter 2) shortly after fungicide applications had been applied to minimize the effect of fungicide redistribution

Figure B1. Fungicide application points on data collection. After inoculation, leaves were and diameter measurements illustration. Circles represent fungicide application incubated for 7-10 days until even GPM growth points and arrows represent diameter measurements. Not to scale. was observed across the leaf surface. An optical micrometer calibrated to millimeters was used to take one diameter measurement going through the center of each point of fungicide application. The diameter measurement was taken perpendicular to the main lobe vein closest to the application point (Figure B1). The measurement was taken as the maximum distance between the lawn of E. necator mycelia on each side of the fungicide drop. Linear regression models were fit using log10 transformed fungicide concentrations as the explanatory variable and the diameter (mm) as the dependent variable. The transformation employed was: a + log10([fungicide] + b) where a=3 to center the regression at the origin and b = 0.001 was added to include the fungicide-free tap water controls in the regression due to their “0” fungicide concentration. The MIC value was calculated as the concentration needed to obtain a diameter of GPM inhibition of 2mm, which was the diameter of the 5μl drop applied (Klittich and Ray, 2013). MIC values were interpolated from regressions using model coefficients and back transformed to obtain a value in mg/L.

110

Results

MIC values segregated into two distinct groups. Fluopyram, tebuconazole and trifloxystrobin had low MIC values on the order of μg/L, while quinoxyfen and sulfur had much higher values on the order of mg/L (Table B2).

Discussion Table B2. Coefficients of determination and minimum inhibitory concentration values. This study used a Fungicide Model MIC AI in 5ul Adjusted R2 (mg/L) drop (mg) cost-effective method to Fluopyram 0.70 0.071 3.57e-7 determine the lowest Quinoxyfen 0.69 2.027 1.01e-5 concentration of five Sulfur 0.59 8.68 4.34e-5 fungicides applied to the Tebuconazole 0.76 0.026 1.29e-7 adaxial surface of a leaf Trifloxystrobin 0.74 0.078 3.9e-7 that would result in inhibition of GPM infection and growth. These MIC values are a manifestation of the inherent activity of these compounds against E. necator, and could be used to guide other experiments investigating fungicide efficacy and the effects of fungicide redistribution on E. necator.

Using the MIC value, the amount of redistribution from a given amount of fungicide could potentially be approximated using predictive modeling techniques. Using spray characteristics (e.g. volume of application, droplet mean diameter, droplet size spectrum, surfactant additions) and crop characteristics (e.g. leaf area, trunk area, amount of deposition) in combination with the MIC could potentially determine fungicide concentrations in sprays that would achieve optimal crop coverage.

111

Appendix C: Oregon Wine Research Institute Fall Technical Newsletter Article

Grape Powdery Mildew Fungicide Application Timing: The Interaction Between Inflorescence Stage and Fungicide Chemistry Brent Warneke, Graduate Research Assistant, Oregon State University Dr. Lindsey Thiessen, Assistant Professor, North Carolina State University Tara Neill, Biological Science Technician, USDA-ARS Dr. Walt Mahaffee, Research Plant Pathologist, USDA-ARS

Introduction Grape Powdery Mildew (GPM) is present in Oregon vineyards every year and can wreak havoc when the right environmental conditions are met. Thus, numerous fungicide applications are made, typically on a calendar basis, to keep cluster infection to a minimum. However, recent marketing and consumer preference trends emphasize sustainably-produced products which require optimized fungicide use to produce disease- free grapes. Whatever the vineyard management philosophy, making the most of each fungicide application is critical. Flowering is an important time to manage GPM, and general recommendations are to apply the most efficacious materials at the shortest spray intervals during this period. Reasons for intensive GPM management around grape bloom include the high susceptibility of inflorescences to infection, and the potential for inflorescence architecture to increase fungal spore deposition, resulting in cluster infections. Optimization of fungicide selection and application timing during bloom could increase the efficiency of GPM management. The use of fungicides that redistribute around plant tissues after application could provide more uniform coverage on intricate tissues, such as a grape inflorescence, resulting in better GPM control. To investigate fungicide application timing during bloom and the influence of fungicide redistribution on disease control, a small plot experiment was conducted on 19- year old Pinot Noir vines trained to bilateral cane VSP on 5’ vine by 6’ row spacing at the OSU Botany and Plant Figure C1. Top: Graphics showing the progression of grapevine flowering Pathology Farm in as denoted by the BBCH scale. Bottom: Three flowering growth stages to Corvallis during the which fungicide applications were applied and their corresponding BBCH numbers. Photos by Brent Warneke, graphics by Javier Tabima.

112

2015 and 2016 growing seasons.

Methods Redistributing fungicides (Table 1) were applied to plots when > 50% of the research vineyard had reached one of three flowering stages based on inflorescence architecture (Figure 1). Five fungicides were chosen based upon their widespread use in industry and varied redistribution profiles (Table 1). All fungicides were applied on a 14- day schedule (Figure 2). Sulfur was applied prior to initiation of redistributing fungicide

Table C1. Selected fungicides, mechanisms, and application rates used. Fungicide Name Fungicide Mechanism Application FRAC Redistribution Rate Trade Technical Mode of Action Group Properties (units/Acre) Inhibition of cell signaling and Xylem systemic, Quintec Quinoxyfen 13 4 fl oz appressorium vapor phase development Sterol Toledo Tebuconazole 3 demethylation Xylem mobile 4 oz inhibition Succinate Translaminar, Luna Fluopyram 7 dehydrogenase xylem, and vapor 4 fl oz Privilege inhibition phase Q I inhibitor of o Translaminar, Flint Trifloxystrobin 11 mitochondrial bc1 2 oz vapor phase complex Unknown multi-site Microthiol Sulfur M2 Vapor phase 3 lb activity

× growth stage treatments. Beginning at each growth stage, a redistributing fungicide was applied twice on the 14-day schedule, and then sulfur was applied until véraison. This resulted in 3 separate treatments for each redistributing fungicide for a total of 15 treatments (3 growth stages × 5 fungicides). These growth stage × fungicide treatments were compared to two different control treatments, a non-treated control that had water applied during each fungicide application, and a calendar-based sulfur program that had 3 lb/A of sulfur applied every 14 days at a volume of 50 gal/A at full canopy. The 17 total treatments were replicated 4 times and arranged in a randomized complete block design. To concurrently investigate fungicide redistribution under field conditions, a subset of clusters from every growth stage × fungicide treatment were individually covered with plastic bags to prevent fungicide deposition during fungicide applications. These clusters did not receive any direct fungicide exposure, so any disease control was hypothesized to be a result of fungicide redistribution.

113

Figure C2. Timeline of the fungicide timing experiment. Photos by Brent Warneke.

Leaf incidence of GPM was assessed weekly starting at 12” of shoot growth and ending at véraison. To quantify berry disease incidence, 10 green clusters per treatment were collected prior to véraison and 25 berries per cluster were examined under magnification for the presence of GPM.

Results The 2015 and 2016 growing seasons were warm and started early. Bud break occurred at the BPP research vineyard in late March 2015 and early April 2016. This early start led to rapid vegetative growth as both springs were warm with summer-like weather. Flowering lasted 14 days in 2015 (5/29 – 6/12) and 8 days in 2016 (5/29 – 6/6). Throughout both summers, there were multiple days over 100ºF with average high temperatures consistently above historical averages. Disease pressure was higher in 2016 than in 2015 as evidenced by leaf disease levels (Figure 3).

Leaf GPM infection To compare leaf disease levels between treatments, area under disease progress curves were calculated using leaf incidence data. Some fungicides showed variable disease levels among the three application timings while others had a trend towards less disease when fungicides were applied later in inflorescence development (e.g. Luna Privilege, Fig. 3). Overall leaf disease levels of the fungicide timing treatments were similar to or lower than the calendar application of sulfur.

Berry GPM infection The mean number of berries infected was used to compare the probability of berry infection for each treatment. Berry infection probabilities from wettable sulfur and Toledo treatments showed no consistent trend related to application timing in both 2015 and 2016, with probabilities similar to or greater than the calendar sulfur control. Conversely, the bloom and berry set applications of Quintec, Flint, and Luna Privilege

114

Figure C3. Area under the disease progress curve (AUDPC) by treatment in 2015 (left) and 2016 (right). The colors show the fungicide used, and the shape within each bar indicates the timing of the fungicide application. Bars and whiskers represent the mean ± one standard error, respectively. had consistently lower disease incidence than the calendar sulfur control and their respective inflorescence elongation timing application, resulting in average berry infection probabilities of less than 0.15 (Figure 4). Flint, Luna Privilege and Quintec bloom and berry set applications were the most effective at reducing berry infection in both years, often with statistically significant reductions in berry incidence compared to calendar application of sulfur (P<0.05).

Fungicide Redistribution The majority of clusters that were covered with plastic bags during fungicide treatments had significantly reduced disease levels compared to the non-treated control. This result suggests that there was redistribution of the fungicides from treated tissues surrounding the bagged cluster, but there could also be an effect of the reduced leaf infection observed in these plots. Regardless of the mechanism for reduced berry incidence, these data indicate that applying redistributing chemistries during bloom and berry set significantly reduces disease development on berries.

Discussion Similar trends in berry infection probabilities among fungicide products were observed during the two years of this study with varying GPM pressure, which indicates that late in grape bloom is an opportune time to apply redistributing fungicides. All three fungicides that effectively reduced the proportion of berries infected when applied mid-bloom to

115

Figure C4. Probabilities of berry infection from 2015 and 2016. Points are the mean probability of berry infection, bars are 95% confidence intervals. berry set (Flint, Quintec, Luna Privilege) exhibit varying degrees of translocation and vapor phase redistribution (Warneke et al. 2017). The efficacy of these late-bloom applications may be due to the fungicide penetrating the developing berry and/or rachis, the increased surface area available for fungicide deposition in the later applications, the timing of the fungicide application reducing leaf disease at an opportune time, or any combination therein. During early bloom, fungicide applied to the developing inflorescence may have adhered to the cap and was then lost when the cap fell off during bloom, while in later treatments the fungicide may have penetrated the developing berry and/or rachis. This could be due to epicuticular wax accumulating as berries develop, facilitating fungicide retention and redistribution in the cluster (Edgington 1981; Klittich 2014; Rosenquist and Morrison 1988). Furthermore, the surface area available to receive fungicide in a cluster increases as berries set and begin to swell, allowing for more fungicide deposition. Fungicide application at bloom and berry set may also slow leaf infection at a time when the developing clusters are most susceptible to infection, delaying infection pressure to a time when the berries are more developed and thus, less susceptible to infection (Gee et al. 2008). Further testing is needed to determine the mechanisms responsible for reducing GPM infection when applying redistributing fungicides late in bloom. In 2017, demonstration trials were conducted in research and commercial vineyards to assess the applicability of timing redistributing fungicide applications to bloom stages, with results forthcoming.

116

Acknowledgements This work was supported by the Oregon Wine Board (Project Number 1826) and USDA- ARS CRIS 5358-22000-039-00D. The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. Such use does not constitute an official endorsement or approval by the United States Department of Agriculture, the Agricultural Research Service, or Oregon State University of any product or service to the exclusion of others that may be suitable. The authors thank Andy Albrecht, Chris Gorman, Bailey Williams, and Carly Allen for their technical assistance throughout the project.

Online version: https://owri.oregonstate.edu/sites/agscid7/files/owri/owri_tech_newsletter_fall_2017_fin al.pdf