CHAPTER 1: Introduction……………………………………………………………1

The Pennsylvania State University The Graduate School Graduate Program in Biology

INFECTIOUS DISEASE DYNAMICS AND THE DIRECT AND INDIRECT EFFECTS OF AN ESCAPED VIRAL-RESISTANCE TRANSGENE ON PLANT FITNESS IN A WILD SQUASH

A Dissertation in Biology by Miruna Ariadna Sasuclark

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy May 2010

The dissertation of Miruna Ariadna Sasuclark was reviewed and approved* by the following:

Andrew G. Stephenson Professor of Biology Dissertation Adviser

Richard Cyr Professor of Biology Chair of Committee

James Winsor Professor of Biology

Mark Mescher Assistant Professor of Entomology

James Rosenberger Professor of Statistics

Douglas Cavener Professor and Head of Biology

*Signatures are on file in the Graduate School

ABSTRACT

This dissertation consists of six chapters: an introduction, four data chapters, and a conclusion: in the Introduction, I present background and provide general information about Cucurbita pepo ssp. texana (the wild progenitor of cultivated squashes), the major herbivores that feed on this plant, and the pathogens they transmit. Background on the two major statistical methods that I have used in the dissertation, analysis of variance (ANOVA) and path analysis, is also included in this section. Additionally, the introduction contains a review of previous experiments from Dr. Stephenson’s laboratory that led to my dissertation work. The second chapter was motivated by previous field studies in the Stephenson lab. These studies revealed that selfed (S) wild gourd plants (Cucurbita pepo ssp. texana) experience greater cucumber beetle herbivory than outcrossed (X) plants but outcrossed plants experience a higher incidence of bacterial wilt disease. Subsequent studies showed that both types of plants were equally susceptible to this disease when exposure was similar. Given that the pathogen responsible for bacterial wilt disease (Erwinia tracheiphila) is known to be transmitted when cucumber beetles chew leaves and then defecate onto the open wounds, the difference in incidence rates required explanation. It is known that cucumber beetles aggregate in the flowers to mate and the Stephenson lab had data showing that X plants produced more flowers than S plants. Furthermore, the flowers from X plants produced more of the volatile compounds that are known to attract cucumber beetles. This phenomenon could be explained if E. tracheiphila can be transmitted via the floral nectaries. Therefore, I explored the possibility of transmission of this plant pathogen via the floral nectaries by: (1) Inoculating greenhouse-grown plants by placing a solution containing Erwinia tracheiphila onto the nectaries of the flowers (after removing the nectar), and showing that E. tracheiphila is able to infect plants via the nectaries; (2) Transforming E. tracheiphila cells with a GFP-marker and visually showing the progression of the bacteria through the nectary tissue and into the xylem of the pedicel; (3) Collecting cucumber beetle frass from field collected flowers and showing that more than 50% of flowers contain cucumber beetle frass in or around the

iii nectaries, and (4) Developing E. tracheiphila-specific primers to screen cucumber beetle frass collected from flowers in the field and showing that more that 90% of flowers contained cucumber beetle frass contaminated with the bacteria.The results of these experiments provide strong circumstantial evidence that E. tracheiphila can be transmitted via the floral nectaries in C. pepo ssp. texana. Chapter three contains a series of experiments to determine if the nectar of C. pepo ssp. texana has antimicrobial properties.Having showed that E. tracheiphila cells can to traverse the nectary tissue of C. pepo flowers and infect plants, I asked if the nectar present in the flowers may interfere with the growth of the bacteria inside the floral nectaries. Flowers of many species have nectar that inhibits the growth of microbes and the secondary compounds responsible for this inhibition are thought to function in preventing the degradation of nectar (to maintain its quality as a pollinator reward). However, in the case of C. pepo, nectar could also function to prevent infection with E. tracheiphila, a floral transmitted pathogen. In one experiment, I examined the effect of nectar on the growth of Erwinia tracheiphila and Escherichia coli using disk diffusion assays. To accomplish this, I grew lawns of E. tracheiphila and E. coli on Petri dishes and examined the area cleared by a filter paper disk impregnated with C. pepo nectar, 40% glucose or ampicillin. This study showed that C. pepo nectar inhibits the growth of both E. tracheiphila and E. coli compared to glucose, a negative control. Additionally, the antibiotic effect of nectar was as great as ampicillin for 12 hrs. In a second experiment, I inoculated greenhouse-grown plants via flowers with and without nectar. This experiment showed that plants inoculated through flowers without nectar experienced significantly higher incidence of wilt disease. Together, these findings show that antimicrobial compounds in C. pepo nectar inhibit the growth of E. tracheiphila and can function to retard the pathogen’s transmission through the floral tissues, allowing time for an abscission layer to form between the flower and pedicel. Chapter four examines the direct and indirect effects of a virus-resistance transgene (VRT) on fitness during introgression in C. pepo ssp. texana in the presence of the full range of insect herbivores (vectors) and diseases. Virus-resistant transgenic squash (Cucurbita pepo ssp pepo) are grown throughout the United States and much of

iv Mexico and it is likely that the VRT has been introduced into wild populations of Cucurbita repeatedly. A three-year, large-scale field study (using wild gourds, transgenic and non-transgenic introgressive plants of various generations) examined reproductive output, herbivory by the primary herbivore (cucumber beetles), and incidence of both the target diseases (viruses transmitted by aphids) and a non-target disease (bacterial wilt disease) on wild gourds and transgenic and non-transgenic introgressives. These studies revealed that (1) Each year, viral diseases became established in mid-July and spread rapidly through the fields until the end of the season when most non-transgenic (susceptible) plants showed symptoms. This also showed that the introgressed transgene was truly effective against viral diseases. (2) The wild gourds, non-transgenic introgressives and transgenic introgressives did not differ in the amount of beetle damage or wilt disease incidence prior to the spread of viral diseases. (3) Once the viral diseases spread through the fields, cucumber beetles preferred to feed on healthy (transgenic) plants and the transgenic plants experienced increased wilt disease incidence. (4) A series of path analysis models showed that the VRT had a direct beneficial effect on reproductive output (transgenic plants are more fit) but this beneficial effect was mitigated by the indirect effects of concentrated cucumber beetle herbivory and increased wilt disease incidence once viral diseases spread thorough the fields. These indirect effects of the VRT on fitness were only apparent in the full context of the Cucurbita pathosystem: thus we can conclude that all target and non-target components are potentially important when examining ecological phenomena. Since chapter four detailed the direct and indirect effects of the VRT when virus is present in all the fields, the next step in chapter five was to determine the costs and benefits of those effects. Therefore, chapter five focuses on the costs and benefits of the VRT during introgression into the wild gourd in the presence and absence of viral diseases, as well as the costs and benefits of cultivar genes during introgression. In this study we planted four one-acre fields with wild gourds, transgenic and non-transgenic introgressives and recorded reproductive output, herbivory by the primary herbivore (cucumber beetles), and incidence of both the target diseases (viruses transmitted by aphids) and a non-target disease (bacterial wilt disease) on the plants in each field. In two fields we allowed viral diseases to spread, while the other two fields were sprayed with

v Endeavor®, an aphid-specific insecticide that has no effect on cucumber beetles or other chewing insects. The insecticide deterred viral transmission by controlling aphid vectors but had no effect on the vectors of bacterial wilt disease. In the unsprayed (virus) fields the data showed that: (1) Viral diseases became established in mid-July and spread rapidly through the fields and by late August most of the susceptible plants showed virus symptoms. (2) The VRT effectively deters viral disease infection in transgenic introgressives. (3) Cucumber beetle herbivory and wilt disease incidence was higher in transgenic introgressives after the virus spread through the fields. (4) Reproductive output was higher in transgenic introgressives. However, on the sprayed (no-virus) fields the data show that the transgenic introgressives also had greater reproductive output than wild gourds and non-transgenic introgressives but the benefit was not as great as in the unsprayed (virus) fields. The benefit of the VRT in the sprayed fields seems to be due to the introduction of the target virus into the fields in August (end of the season) and a benefit of cultivar genes. The optimal growing conditions for wild gourds are hot and dry and because the summer of 2009 was wet and cool, both the transgenic and non- transgenic introgressives had greater reproductive output than wild gourd plants, regardless of disease pressure. The parental cultivar had been bred to grow and reproduce well in the northeastern US while the wild gourd originated in Texas. These findings also show that there are indirect effects of the VRT on a non-target herbivore (cucumber beetles) and the pathogen it transmits (E. tracheiphila) when viral diseases are allowed to spread through the fields. These findings suggest that yearly variations in environmental variables such as temperature, rainfall and time of virus introduction, can impact the fitness of the VRT during introgression. In addition, it is important to note that our studies were performed north of the native range of wild gourds and that different populations of organisms (insects, predators, pathogens) as well as competition (both inter and intra-specific) among plants may also play roles in the fitness of the VRT during introgression. In summary, my dissertation explores the transmission of an important bacterial pathogen (bacterial wilt disease) through floral tissues of Cucurbita pepo ssp. texana and the effect of floral nectar on the transmission of the pathogen, as well as the direct and indirect effects of a virus-resistance transgene (VRT) on plant fitness, cucumber beetle

vi herbivory, and the incidence of bacterial wilt disease and viral diseases as the VRT introgresses into wild Cucurbita pepo. As a unit, my research suggests that it is important to study the individual components of a pathosystem and their interactions when assessing the costs and benefits of resistance genes (both natural and transgenic).

vii TABLE OF CONTENTS

LIST OF FIGURES……………………………………………………………………x

LIST OF TABLES…………………………………………………………………….xii

ACKNOWLEDGEMENTS…………………………………………………………..xiv

CHAPTER 1: Introduction……………………………………………………………1

General background……………………………………………………………....1 The study organisms……………………………………………………………...3 Statistical methods ……………………………………………………………….6 Analysis of variance (ANOVA)………………………………………….6 Path analysis …………………………………………………………...…7 Previous studies of Cucurbita pepo ssp. texana………………………..…9

CHAPTER 2: Floral Transmission of Erwinia tracheiphila by Cucumber Beetles in a Wild Cucurbita pepo

Abstract………………………………………………………………………….11 Introduction……………………………………………………………………...13 Materials and Methods…………………………………………………………..15 Results…………………………………………………………………………...20 Discussion ………………………………………………………………………22 Acknowledgements……………………………………………………………...25

CHAPTER 3: Antimicrobial Nectar Inhibits a Floral Transmitted Pathogen of a Wild Cucurbita pepo (Cucurbitaceae)

viii Abstract…………………………………………………………………………31 Introduction……………………………………………………………………..32 Materials and Methods………………………………………………………….33 Results…………………………………………………………………………..37 Discussion ………………………………………………………………………38 Acknowledgements…………………………………………………………...…40

CHAPTER 4: Indirect Costs of a Non-Target Pathogen Mitigate the Direct Benefits of a Virus Resistant Transgene in Wild Cucurbita

Abstract………………………………………………………………………….46 Introduction……………………………………………………………………...47 Materials and Methods…………………………………………………………..50 Results…………………………………………………………………………...54 Discussion ……………………………………………………………………….56 Acknowledgements………………………………………………………...…….58

CHAPTER 5: The Costs and Benefits of a Virus Resistance Transgene During Introgression into a Wild Gourd Abstract…………………………………………………………………………..67 Introduction…………………………………………………………………...….68 Materials and Methods…………………………………………………...………71 Results………………………………………………………………………...….76 Discussion……………………………………………………………...... 78 Acknowledgements………………………………………………………...…….83

CHAPTER 6: Conclusions and Future Directions…………………………………...96

REFERENCES………………………………………………………………………...101

ix LIST OF FIGURES

Figure 2.1 Cucumber beetles in Texas gourd flowers and feeding damage to floral tissues ……………………………………………………………………………………27

Figure 2.2 Images of E. tracheiphila expressing GFP inside pedicel vascular bundles of C..pepo ssp. texana (Texas gourd) 24 hours after inoculation at 40x under light microscopy using a GFP filter…………………………………………………...28

Figure 3.1 Schematic of a disk diffusion assay plate……………………………………41

Figure 3.2 Disk diffusion assay inhibition zones (cm2) on lawns of E. coli and E. tracheiphila………………………………………………………………………42

Figure 4.1. The proportion of uninfected, living plants (i.e. susceptible plants) on the 15th of each month showing symptoms of viral disease on the 15th of the following month for the years (A) 2006, (B) 2007, (C) 2008………………………………59

Figure 4.2. Number of male flowers and fruits produced per plant on texana gourds, Non-VRT and VRT introgressives ……………………………………………...60

Figure 4.3. Amount of beetle damage on leaves of Non-VRT (texana and non-VRT introgressives) and VRT plants ………………………………………………….61

Figure 4.4. Proportion of susceptible plants (alive and not infected with virus) that became infected with Erwinia in the interval 15 June-15 July (July) and 15 July- 15 August (August)………………………………………………………………62

Figure 4.5. Final path analysis including the effects of (A) time with virus and (B) time with wilt disease on reproductive output………………………………………..63

x Figure 5.1. The proportion of susceptible plants infected with ZYMV on the two

unsprayed fields in which viral diseases were allowed to spread……………….85

Figure 5.2. Reproductive output (male flowers and fruit) during 2009 field season for Texas gourds, non-transgenic BC4 introgressives (ntBC4) and transgenic BC4 introgressives (tBC4) (Least square means ±SE)……………………………….86

Figure 5.4. Amount of beetle damage in August 2009 on leaves of Texas gourds, non- transgenic BC4 introgressives (ntBC4) and transgenic BC4 introgressives (tBC4) (Least square means ±SE )……………………………………………………….89

LIST OF TABLES

Table 2.1 Percentage of plants with wilt disease and the cumulative incidence of wilt disease on the three dates of cucumber beetle frass collection from male flowers…………………………………………………………………………...29

Table 2.2 Percentage of plants infected with Erwinia tracheiphila via floral nectaries under greenhouse conditions……………………………………………………30

Table 3.1 Results of a repeated measures analysis of variance for the effects of treatment (5% ampicillin, 40% glucose, nectar from lightly damaged plants, and nectar from heavily damaged plants), time, date of the assay, plate nested within date (15 plates per bacteria per date), and the interactions of date by treatment and treatment by time on area of inhibition (cm2) on petri plate lawns of A) Escherichia coli and B) Erwinia tracheiphila…………………………………...31

Table 4.1 Magnitude of all significant paths from plant type to reproductive output in the two path models………………………………………………………………….64

Table 4.2. Effects of Year, Field nested within Year, Generation (texana, and non- transgenic F1, BC1, BC2, BC3, BC4) nested within Year, and Family [random] nested with Year on beetle damage during June and July of 2006- 2008………………………………………………………………………………65

Table 4.3. Least Square Means of beetle damage estimates for texana and non-transgenic F1, BC1, BC2, BC3 and BC4 plants. Model: June and July beetle damage = Year + Field(Year) + Generation(Year) + Family(Year) where Family is random…………………………………………………………………………...66

Table 5.1. Effects of field, family (random), plant type (Texas gourds, non-transgenic

xii BC4 introgressives and transgenic BC4 introgressives) and family x plant type on reproductive output during the 2009 season……………………………………..90

Table 5.2. Effects of Field, Family (random), Plant type (Texas gourds, non-transgenic BC4 introgressives and transgenic BC4 introgressives) and Family x Plant type on beetle damage during August 2009…………………………………….………..91

Table 5.3. Effects of Field, Family (random), Plant type (Texas gourds, non-transgenic BC4 introgressives and transgenic BC4 introgressives) and Family x Plant type on beetle damage during June and July 2009………………………………..….92

xiii ACKNOWLEDGEMENTS

In the past five years, I have been fortunate enough to interact and communicate with several people who have contributed to furthering my knowledge and my overall development as a scholar and as a person. I thank my advisor, Dr. Andrew Stephenson for his guidance and support and sharing his vast knowledge with me for the past five years. I also thank the members of my committee: Dr. Richard Cyr, Dr. Mark Mescher, Dr. Jim Winsor and Dr. James Rosenberger. Each made a significant contribution to specific aspects of my research and general knowledge. Each was always available and willing to give words of advice when I sought guidance and all contributed to the overall quality of my research. I would also like to thank Dr. Matthew Ferrari and Dr. Durland Shumway for their statistics expertise and the members of the Stephenson lab, past and present, for their support, kindness, friendship, and expertise: Dr. Jorge Mena-Ali, Heather Simmons, Rupesh Ram Kariat, Kelly Wall, Franz Lichtner, Sarah Scanlon and Melinda Bothe. Tony Omeis has also provided his expertise and I am vastly indebted to him for their endless assistance with greenhouse plantings and maintenance. In addition I would like to thank Bob Oberheim for all his guidance at the horticultural farm. Without them, my work would not have been possible. Several other people have shared this experience with me, contributed to my personal development and served as my support system. I was lucky to have met some life-long friends and would like to thank them for their endless support and encouragement: Gaelan Ritter, Dr. Kathryn Stamps-Mitchell, Greg Mitchell, Dr. Martha Nelson, Liz Bosak, Jasmine Fledderjohann, Christina Harris and Dr. Irmgard Seidl- Adams. These people were the foundation of my happiness even when times were tough. I thank my family, in particular my mother, Mirela Sasuclark, for always being a pillar of strength and for teaching me to be strong and independent. I thank my grandmother, Maria Sasu, for always believing in me; and my grandfather, Nicolae Sasu, for teaching me that life is a series of memories that become what you make of them. I thank my brother, Alex Sasuclark, for his endless support, and teaching me how to have fun on roller-coasters.

xiv

CHAPTER 1

General background Foliar herbivory is an ubiquitous component of terrestrial communities that has dramatic fitness consequences for plants (McNaughton et al 1989, Louda et al. 1990, Marquis 1982). Because leaves function as the main photosynthetic organs for individual plants, foliar herbivory adversely affects vegetative growth, survival, and reproduction through both female (fruit and seed) and male (pollen production and pollen performance) functions (Mutikainen and Delph 1996, Coley and Barone 1996, Delph et al. 1997, Strauss et al. 2001). Leaf damage also induces chemical and physical defense systems that can be costly in terms of energy and nutrients (see reviews by Bergelson and Purrington 1996, Bazzaz 1997, Karban and Baldwin 1997, Strauss et al. 2002). Pathogen infection, like foliar herbivory, has been shown to impact plant fitness, by decreasing leaf area, disrupting cellular and transport processes, and by inducing various biochemical defense systems (see Burdon 1987a, Ryals et al. 1996, Devadas et al. 2002). Pathogens also reduce growth, survival, and reproductive output in both cultivated and wild species (e.g. Burdon 1987a, Burdon 1987b, Burdon and Leather 1990). Studies of plant-herbivore interactions have demonstrated that the amount and impact of herbivory on fitness often has a genetic basis. For instance, Strauss and Agrawal(1999) demonstrated that there is genetic variation for resistance (ability to reduce the performance of a natural enemy) and tolerance to pathogens(ability to reduce the fitness consequences of a natural enemy) in populations of cultivated species. Other groups have demonstrated similar results in non-cultivated species (Kennedy and Barbour 1992, Tiffin and Rausher 1999, Weinig et al. 2003). As with foliar herbivory, there is a genetic basis for both resistance and tolerance (e.g., Fritz and Simms 1992, Simms and Triplett 1994, Biere and Antonovics 1996, Carlsson-Graner 1997). Moreover, pathogen infection and herbivory are synergistic such that herbivores often facilitate pathogen infections by serving as vectors, weakening plant defense systems, and opening wounds for additional pathogen entry. Therefore, the dynamics of disease transmission can be

understood within a framework of host plant-vector-pathogen interactions (Irwin and Thresh 1990). Inbreeding is common in flowering plants and often results in a significant loss of fitness by reducing heterozygosity, which exposes deleterious recessive alleles to selection and reduces the contribution of overdominance to fitness (e.g., Charlesworth and Charlesworth 1987, Husband, and Schemske. 1996 , Crnokrak, and S Barrett. 2002). Increased homozygosity has the potential to adversely affect genes that are known to be expressed in the inducible biochemical defense pathways (e.g., Maleck et al. 2000, Schenk et al. 2000). In addition, scores of mutations are known that directly alter the speed and/or magnitude of the response of the inducible defense pathways (Martin et al. 2003, McDowell and Woffenden 2003). Inbreeding can also affect physical defenses (such as leaf toughness or trichome density) and the volatile organic chemicals produced by plants that influence resistance, pest performance, and pest preference (see Hare et al. 2003, DeMoraes et al. 2001, Kessler and Baldwin. 2001, Eigenbrode et al. 2002). Furthermore, decreases in general plant vigor associated with inbreeding (e.g., Husband and Schemske. 1996) could indirectly affect plant resistance by decreasing the resources available for defense pathways, volatile compound production, wound repair, and by slowing the rate of growth during vulnerable stages of the life cycle. Investigations of the effects of inbreeding on plant-herbivore and (separately) plant-pathogen interactions (Strauss and. Karban 1994, Matheson et al. 1995, Nunez-Farfan 1996, Ouborg et al. 2000, Carr and Eubanks. 2002 , Ivey and Carr. 2005, Hull-Sanders and Eubanks 2005) indicate that resistance to herbivores/pathogens changes with inbreeding (but see Nunez- Farfan et al. 1996) and that the changes occur in a family/population specific manner. If resistance to herbivores and exposure/resistance to pathogens changes with the level of inbreeding, it would have profound implications for the evolution of mating systems, the establishment and spread of diseases in natural populations, the conservation of small populations, and in natural communities it could have ramifications throughout the food chain.

The study organisms The wild gourd, Cucurbita pepo ssp. texana (sometimes seen as var. ozarkana) is thought to be the wild progenitor of the cultivated squashes, and is native to Northern Mexico, Texas, and the states along the Mississippi River from Southern Illinois southward (Decker-Walters 1990, Lira et al. 1995, Decker-Walters et al. 2002). C. pepo ssp. texana is a free-living, annual, monoecious vine with indeterminate growth and reproduction. When cultivated and wild cucurbits are grown in the same vicinity, they can produce hybrid progeny which are frequently viable and fertile (Kirkpatrick and Wilson 1988, Quesada et al. 1993). Cucurbits (pumpkins, gourds, squashes) are one of the 15 most important agricultural crops in the U.S. and annual cucurbit production is valued at more than $1.4 billion dollars (Cantliffe 2004). In addition, the majority of cucurbits are cultivated throughout southern states such as Texas, Georgia, Arizona and Florida, thus overlapping with the native range of wild squash. After a period of vegetative growth (5-7 nodes), the wild gourd produces one large yellow flower (either staminate [male] or pistillate [female]) in the axil of each leaf. Each flower opens for pollination, especially by squash bees, for only one morning. At the base of each flower, a cup-shaped nectary secretes 40μl-120μl of nectar daily, which is completely reabsorbed by the morning after anthesis (Nepi et al. 2001). Female flowers have a completely uncovered cup-shaped nectary, while in male flowers the nectary is covered by an extension of the androecium and has three slits that allow nectar consumption by pollinators and other insects (Nepi et al. 1996).

The leaves and other organs of the wild gourd produce cucurbitacins (oxygenated tetracyclic triterpenes). Cucurbitacins are among the most bitter compounds known, detectable by humans at levels of 1 part per billion (ppb). Although cucurbitacins are toxic to most herbivores, cucumber beetles, which are found throughout the native ranges of Cucurbita species, not only consume the leaves of Cucurbita but also chemically modify cucurbitacins before sequestering them for their own defense (Tallamy 1985, Metcalf and Rhodes 1990, Robinson and Decker-Walters 1997). The males also transfer some of the modified cucurbitacins to the females in their seminal fluid which is used to

3 chemically protect the eggs (Ferguson et al. 1985, Nishida and Fukami 1990, Tallamy and Krischik 1989, Tallamy and Brown 1999). Not only are cucumber beetles attracted to Cucurbita foliage, but it has been shown that the floral volatiles attract cucumber beetles as well as pollinators over relatively large distances (Anderson and Metcalf 1986, Anderson and Metcalf 1987, Metcalf and Lampman 1991). These beetles create a characteristic pattern of holes (1-1.5 cm in diameter) in the portions of the leaves serviced by the smallest veins that has been shown to substantially reduce yield in cultivated cucurbits (e.g., Tallamy and Krischik 1989), and reproductive output in the wild gourd (Quesada et al. 1995, Stephenson et al. 2004, Du et al. 2008). After feeding on the leaves, the beetles aggregate in the flowers to mate and feed (Darlington 2006).

Cucumber beetles (in particular the two species Acalymma vittatum and Diabrotica udecimpuctata howardii) are also the only known vectors of the bacterial pathogen, Erwinia tracheiphila (Enterobacteriaceae), which causes bacterial wilt disease and is the most economically important disease of cultivated cucurbits (cucumbers, melons, pumpkins and squash) in the Eastern U.S.. The disease and is normally managed using insecticides to control the vector (Fleischer et al. 1999). Wilt disease symptoms typically develop 10-15 days after infection and the disease is nearly always fatal (Yao et al. 1996). E. tracheiphila overwinters in the intestinal tract of cucumber beetles. Prior to my dissertation work it was thought that the sole transmission route of E. tracheiphila was via contaminated cucumber beetle fecal pellets entering into their feeding sites on leaves (Fleischer et al. 1999). The bacteria proliferates in the xylem where they secrete an exopolysaccharide (mucilaginous) matrix, which arrests water movement through the plant and causes wilting symptoms.

The four most common viral diseases of cucurbits (Cucumber mosaic virus [CMV], Papaya ringspot virus [PRSV, formerly WMV-1], Watermelon mosaic virus-2 [WMV-2], and Zucchini yellow mosaic virus [ZYMV]) are transmitted by generalist aphids. These diseases produce symptoms that include foliar blisters, necrotic lesions, branches with short internodes, small highly serrate leaves, and other leaf and fruit deformities which have been shown to reduce agricultural yields up to 94% (Blua and Perring 1989). Although viral infection is easy to identify visually, the symptoms of infection are similar for all four viruses and mixed infections can also occur; thus

4 commercially available (Agdia Inc. Elkhart, IN) immunological (DAS-ELISA) kits are needed to identify the specific viruses that infect wild gourd plants (Blancard et al. 1994, MacNab et al. 1994). In the mid-1990s, the United States Department of Agriculture (USDA) deregulated transgenic squash varieties containing coat protein (CP) based resistance to CMV, WMV-2 and ZYMV (USDA 1994, 1996). By 2004, seeds for transgenic plants became available in retail stores, and since then over ½ million acres of cultivated transgenic squash have been planted annually in the U.S. (Cantliffe et al. 2005). Thus it is highly likely that these transgenes are being repeatedly introduced into wild gourd populations because (1) wild gourds (free-living C. pepo) are able to freely hybridize with squash (2) natural populations of wild gourds grow in close proximity to cultivated squash throughout their natural range and (3) gene flow via pollen transfer regularly occurs between cultivated and wild C. pepo over distances exceeding a kilometer (e.g., Kirkpatrick and Wilson 1988, Decker-Walters et al. 2002). One of the USDA deregulated transgenic varieties, Liberator III, is a yellow crookneck squash cultivar, for which the transgene is hemizygous (single gene copy), and is linked to a reporter gene NPTII. The Stephenson lab has bred (introgressed) the transgene into wild gourds, and as a result of the reporter gene (NPTII) is able to detect the presence of the transgene in hybrid progeny using a DAS-ELISA test. In order to control breeding, Dr. Stephenson has produced these experimental seeds by hand- pollinations, in which natural pollinators have been excluded. In addition, as the experiments were performed far north of the natural range of free-living gourds, the accidental introgression of the transgene into wild gourds was not a concern. Escape of disease-resistance transgenes into wild populations could confer a selective advantage. Recent models and simulations suggest that the fate of resistance genes (both natural and transgenic) in populations depends upon: (1) the fitness benefit of the resistance gene when the pathogen is present; (2) the fitness cost of the resistance gene in the absence of the pathogen; (3) the frequency of the disease within the population; (4) the effect of resistant plants on the incidence of the disease in the population; (5) the dominance of the resistance allele and (6) metapopulation dynamics (see Kniskern, and Rausher 2007 for recent discussion, also Garrett and Mundt 1999).

5 The cost of resistance to plant fitness includes direct costs (e.g., upregulation of defense pathways) and indirect costs associated with reduced resource accumulation (nutrient uptake and photosynthesis), pleiotropic effects on growth and development, decreased resistance and increased exposure to other natural enemies (e.g., Bergelson and Purrington. 1996, Purrington 2000, Heil and Baldwin 1997, Tian et al. 2003, Gassmann and Futuyma 2005).

Statistical Methods Analysis of Variance (ANOVA) Analysis of Variance (ANOVA) is a widely used statistical tool that allows for partitioning of the variation (variance) that occurs in the component variables of an experimental system (Kuehl 2000, Field 2009, Onofri et al. 2009). The variance (s2) is calculated by the sums of squares of the deviations from the mean divided by the sample size minus one (n-1) (Montgomery 2005). In brief, given a certain sample size n, the variance in a model is a function of the variation of each sample from the mean. The total variance in the model can be further subdivided into within- group variability (SS Error - random error term) and between-group variability (SS Effect-variance due to experimental groups) (Montgomery 2005, Hinkelman and Klaus 2008, Field 2009). To test for significant differences among the means of different experimental groups we use the F test, which determines if the ratio of variances (within- group and between-group) is greater than 1 (Montgomery 2005, Field 2009). If the resulting p value is <0.05, by convention we can say that there exists at least one or more differences among means. Unlike the T-test, the F-ratio The F-ratio produces a single p- value and shows that at least one difference exists among experimental group means, without increasing the likelihood of a Type I error (false positive) (Montgomery 2005). However, ANOVA cannot determine which means are different and therefore it is necessary to apply multiple comparison methods such as Tukey-Kramer and Dunnett procedures. These tests compare all experimental group means while controlling for the overall Type I error (Kuehl 2000). I used the Tukey-Kramer multiple comparisons method, which compares each group mean with every other group mean to control for the

6 experiment-wise error rate. I applied this method to the data in chapter four and chapter five. There are three conceptual classes of ANOVA: fixed effects, random effects and mixed effects models. Fixed effects ANOVA assumes that the groups being compared are only different in their means and that the populations from which the data are derived are normally distributed. Random effects ANOVA assumes that the data being analyzed are from different populations and the differences in means can be attributed to those populations. Mixed effects ANOVA models include both fixed effects and random effects (Kuehl 2000, Onofri 2009). I used both fixed effect models and mixed effect models in this dissertation. The ANOVA assumptions are that the data must be normally distributed, the experimental errors are independent and the variance in groups of data should be equal (homoscedascity) (Onofri et al. 2009). In chapter three I used mixed-model repeated measures ANOVA (observations on the same sample at multiple times) coupled with the Tukey-Kramer multiple comparisons test to show that C. pepo nectar had significant inhibitory properties against E. tracheiphila when compared to a positive (ampicillin) and negative (40% glucose) control at different time intervals. In chapters four and five, I used mixed-model ANOVA and Tukey-Kramer multiple comparison tests to show the direct effects and costs and benefits of a virus-resistance transgene (VRT) on wild gourds, non-transgenic introgressives and transgenic introgressives on reproductive output and leaf herbivory by cucumber beetles. The ANOVA tests were performed using SAS Proc GLM and Proc Mixed (SAS Institute 2002).

Path Analysis Path analysis is a statistical tool that allows researchers to examine both the direct and the indirect effects of a dependent variable (e.g. reproductive output) while holding all other variables constant (Sokal and Rohlf 1995, Byrne 2001). Path analysis is a type of structural equation modeling (SEM), which in turn is a form of multiple regression and is useful to describe directed dependencies among variables in a model (Byrne 2001). Path analysis allows the investigators to make inferences based not only on obvious

7 direct effects but also indirect effects by creating a matrix of directed regressions (Kingsolver and Schemske 1991, Conner et al. 1996, Shipley 2000, Scheiner et al. 2000, 2002, Pigliucci and Kolodynska 2006, Campbell and Snow 2007). These models are typically pictorially represented by a diagram of circles or squares (variables) and arrows that connect them. Each one-way arrow depicts a regression weight (ß-beta) and suggests causality, while two-way arrows imply correlation (Byrne 2001). The regression weights are calculated by considering all of the other variables and using maximum likelihood estimation (MLE). Path analysis includes two types of variables: exogenous and endogenous. Exogenous variables are generally characterized as the independent variables with no explicit cause, while endogenous variables are characterized as dependent variables (Byrne 2001). When the model is specified and all directed regressions have been calculated, indirect effects of variables on each other can be calculated by tracing the paths and multiplying the standardized regression coefficients (Shipley 2000, Byrne 2001). Overall model fit is calculated by comparing the observed correlation matrix (saturated model) to the model being tested and a goodness-of-fit estimation is calculated. Goodness-of-fit is typically approximated by χ2 (Chi-square) approximation, root mean square error approximation (RMSEA) and Aikake information criterion (AIC) (Byrne 2001). The fit of the final model should always be measured using all three estimators because one method may not be entirely appropriate given the context of the model (e.g. χ2 test is overly sensitive when a model includes large sample sizes). When performing path analysis the following assumptions are made: (1) every relationship among variables is linear (2) there are no interaction effects among variables in the model (3) residual (unmeasured) variables are not correlated (4) indicator variables are not correlated (low multicollinearity) (5) disturbance terms (residual errors) are uncorrelated among endogenous variables (6) no important variables are left out of the model (proper specification) (7) no feedback loops created with arrows (proper recursivity) and (8) no values are missing (Bolen 1989, Byrne 2001). Several software packages including SPSS Amos, LISREL and Stata offer structural equation modeling. In chapter four I constructed a series of path analysis models using SPSS Amos to find the indirect effects on the Cucurbita pathosystem

8 including several key ecological players affecting wild gourds, non-transgenic and transgenic introgressive C. pepo ssp. texana plants. Path analysis allowed me to examine the effect of a virus-resistance transgene (VRT) on each variable in the model (reproductive output, family, year, cucumber beetle damage, incidence of wilt disease and viral diseases) while holding constant all other factors that have paths leading to that variable. Subsequently, using path analysis revealed the importance of factors that indirectly affect other variables in the model.

Previous studies of C. pepo ssp. texana

Previous large-scale projects from the Stephenson formed the starting point form my dissertation. The particular projects are those that aimed: (1) to examine the effects of herbivores and pathogens on the magnitude of inbreeding depression in Cucurbita pepo ssp. texana, (2) to determine if resistance and tolerance to herbivory varies with inbreeding, and (3) to determine if the incidence of infection by wilt disease and mosaic viral diseases differ on inbred and outbred plants. From 2002 to 2005 Dr. Stephenson’s lab group performed a series of large-scale field experiments, using selfed (S) and outcrossed (X) plants (and later transgenic and non-transgenic introgressive plants) from five families. These experiments were performed at the Rock Springs Agricultural Station at The Penn. State University. The data collected includes male and female flower production, total annual fruit production, amount of cucumber beetle damage, aphid population sizes on each plant, and incidence of wilt and viral diseases. The analyses of these data revealed that (1) there is broad-sense heritable genetic variation for resistance to herbivores among families (2) inbred plants suffer greater herbivore damage and harbor larger populations of aphids than outcrossed plants (3) both male and female functions are significantly affected by total beetle damage, viral infection, and inbreeding (4) beetle damage, viral infection, and inbreeding depress the female function more than the male function (5) the timing of herbivory is important with respect to the developmental stage of the plant (Hayes et al. 2004, Stephenson et al. 2004, Stephenson et al. 2006, Du et al. 2008)

9 Further analyses revealed that (1) 8-55% of the plants in each field eventually contract wilt disease and die 10-15 days after the first appearance of symptoms (but continue to produce flowers until the plants collapse) (2) 5-50% of the plants in each field contract ZYMV or WMV-2, which significantly decreases the growth rate and flower production but does not kill the plants (3) inbred plants harbor larger aphid populations, and are significantly more likely to be infected with a viral diseases than outbred plants (4) wild gourd introgressives are more likely to contract viral diseases than non- introgressive wild gourds, indicating that wild gourds are more resistant to aphids and/or the pathogens they transmit than the cultivar plants (5) plants that contract viral diseases are significantly less likely to subsequently contract bacterial wilt disease and (6) outbred plants consistently experience less beetle damage than inbred plants but, contrary to expectations, experience significantly higher incidence of wilt disease (Ferrari et al. 2007). The discovery that outcrossed plants experience a higher incidence of wilt disease even though selfed plants are subject to more cucumber beetle damage led to an investigation of the differences in susceptibility to E. tracheiphila among maternal families and breeding (S/X). The resulting experiments revealed that seedlings of inbred and outcrossed plants do not differ in susceptibility to E. tracheiphila when exposure is constant. Given that both types of plants were equally susceptible to wilt, the lab began to examine factors that would differentially influence the exposure rates to E. tracheiphila by examining the volatile profiles produced by the flowers of S and X plants (using gas chromatography coupled with mass spectrometry). This study revealed that flowers from X plants released more volatiles that attract cucumber beetles than S plants (Ferrari et al. 2006). These data suggested that X plants have a higher incidence of wilt disease (despite the fact that S plants experience higher cucumber beetle herbivory) due to increased exposure to the herbivores (and in turn the bacterial pathogen) via floral volatiles. These results, taken in combination with the fact that cucumber beetles aggregate inside the flowers to mate, led me to investigate the possibility that E. tracheiphila may be transmitted via the flowers, which is addressed in chapter two. As the transmission appeared to occur via the flowers it became apparent that nectar may play a role in the

10 transmission of this disease. Thus, I undertook an investigation of the antimicrobial effects of cucurbit nectar which is discussed in chapter three. To gain a deeper understanding of these plant-vector-pathogen interactions I integrated the individual components of the cucurbit system and considered the pathosystem as a whole. Thus, I examined the direct and indirect effects of the virus resistance transgene (VRT) on plant fitness, cucumber beetle herbivory and wilt and viral disease incidence, as discussed in chapter four. In addition, I explored the costs and benefits of the VRT on plant fitness, cucumber beetle herbivory and wilt and viral disease incidence, both in the presence and absence of the target pathogen (viral diseases) as discussed in chapter five.

CHAPTER 2 Floral Transmission of Erwinia tracheiphila by Cucumber Beetles in a Wild Cucurbita pepo

M. A. Sasu,1 I. Seidl-Adams,2 K. Wall,1 J. A. Winsor3 and A. G. Stephenson4

1Department of Biology and Center for Chemical Ecology, 208 Mueller Lab, The Pennsylvania State University, University Park, PA 16802 2Department of Entomology and Center for Chemical Ecology, Chemical Ecology Lab, The Pennsylvania State University, University Park, PA 16801 3Department of Biology, Department of Biology, The Pennsylvania State University, Altoona, PA, 16601, USA 4Department of Biology and Center for Chemical Ecology and Center for Infectious Disease Dynamics, 208 Mueller Lab, The Pennsylvania State University, University Park, PA 16802

Note: Accepted for publication in Environmental Entomologist. In press.

Abstract Cucumber beetles, Acalymma vittatum (F.) and Diabrotica undecipunctata howardi (Barber), are specialist herbivores of cucurbits and the vector of Erwinia tracheiphila (E.F. Smith) Holland, the causative agent of wilt disease. Cucumber beetles transmit E. tracheiphila when infected frass falls onto leaf wounds at the site of beetle feeding. We show that E. tracheiphila also can be transmitted via the floral nectaries of Cucurbita pepo ssp. texana (L) Andres (Texas gourd). Under field conditions, we find that beetles aggregate in flowers in the late morning, that these beetles chew the anther filaments that cover the nectaries in male flowers thereby exposing the nectary, and that beetle frass accumulates on the nectary. We use RealTime PCR to show that most of the

12 flowers produced during the late summer possess beetle frass containing E. tracheiphila. Greenhouse experiments, in which cultures of E. tracheiphila are deposited onto the nectaries of flowers, reveal that Texas gourds can contract wilt disease via the floral nectaries. Finally, we use GFP-transformed E. tracheiphila to document the movement of E. tracheiphila through the nectary into the xylem of the pedicel prior to the abscission of the flower. Together, these data show that E. tracheiphila can be transmitted via infected frass that falls on or near the floral nectaries. We hypothesize that the concentration of frass from many beetles in the flowers increases both exposure to and the concentration of E. tracheiphila and plays a major role in the dynamics of wilt disease in both wild populations and cultivated squash fields.

Introduction The leaves and other organs of plants in the Cucurbitaceae produce cucurbitacins (oxygenated tetracyclic triterpenes). Cucurbitacins are among the most bitter compounds known, detectable by humans at levels of 1ppb, and are toxic to most herbivores (Tallamy 1985, Metcalf and Rhodes 1990). However, cucumber beetles (Diabrotica spp. and Acalymma spp.) are adapted to feed on cucurbitacins in the cotyledons, leaves, flowers and fruits of Cucurbita species and are found throughout the native ranges of Cucurbita species (Tallamy 1985, Metcalf and Rhodes 1990, Eben and Barbercheck 1996). Cucumber beetles are attracted to cucurbitacins in the foliage of Cucurbita and, when flowers are present, it has been shown that floral volatiles not only attract bees (the pollinators) but also cucumber beetles over relatively large distances (Anderson and Metcalf 1986, 1987, Lampman and Metcalf 1988, Metcalf and Lampman 1991). In the mid to late mornings, the beetles aggregate in the flowers of Cucurbita species to feed and mate (Anderson and Metcalf 1987). Foliar feeding by these beetles results in a characteristic pattern of small holes in the portions of the leaves serviced by the smallest veins. Beetle damage has been shown to substantially reduce yield in cultivated cucurbits (cucumbers, melons, squash and pumpkins) and reproductive output in wild Cucurbita (e.g., Tallamy and Krischik 1989, Quesada et al. 1995, Stephenson et al. 2004, Du et al. 2008).

13 In addition to reducing photosynthetic leaf area, herbivory by cucumber beetles on wild and cultivated cucurbits also increases exposure to a variety of pathogens, including Erwinia tracheiphila (E.F. Smith) Holland (Enterobacteriaceae) (Fleischer et. al. 1999, Stephenson et. al. 2004, Du et. al. 2008). Erwinia tracheilphila is the causative agent of bacterial wilt disease which is an economically significant disease vectored by cucumber beetles (Brust 1997c, Fleischer et. al. 1999, Mitchell and Hanks 2009). After entering the plant the bacteria proliferate in the xylem where they secrete an exopolysaccharide matrix that cuts off the water supply resulting in wilting. Wilt symptoms typically develop 7-15 days after infection and the disease is nearly always fatal once symptoms appear (Yao et al. 1996). Death occurs 1-3 weeks after symptoms appear. In the eastern United States, E. tracheiphila overwinters in the digestive track of cucumber beetles (Garcia-Salazar et al. 2000). Serological studies have shown that ~7- 11% of the Acalymma vittatum (F.) that emerged from the soil in the spring tested positive for E. tracheiphila (Fleischer et al. 1999). Transmission is thought to occur when fecal pellets containing E. tracheiphila land on leaf wounds at the sites of feeding damage (Leach 1964). A recent study using an endpoint PCR-based technique revealed that the frass of A. vittatum that had fed on infected plants contained E. tracheiphila and that the frass was capable of causing infection using pin-prick inoculations (Mitchell and Hanks 2009). Fleischer et al. (1999) estimate that the proportion of beetles harboring E. tracheiphila is typically 5X greater than the proportion of beetles that actually transmit disease when an individual beetle is placed into a cage with a cucurbit seedling. Other studies have shown that the efficiency of transmission depends upon the size of the wound, the inoculum dose and the amount of time the infected beetles feed on the plant (Lukezic et al. 1996, Brust 1997a,b) Over the last seven years, we have investigated the interrelationships among inbreeding, herbivory by cucumber beetles (A. vittatum and Diabrotica undecipunctata howardi (Barber), and the incidence of bacterial wilt disease in a wild gourd, Cucurbita pepo ssp. texana (L) Andres (Texas gourd) in a series of large field scale studies (Hayes et al. 2004, Stephenson et al. 2004, Ferrari et al. 2006, 2007, Du et al. 2008). These studies have consistently shown that inbred plants grew more slowly, produced fewer flowers and fruits, and suffered greater levels of leaf damage by cucumber beetles than

14 outbred plants but that outbred plants had a greater incidence of wilt disease. Controlled inoculation studies revealed that there were no significant differences between inbred and outbred plants in resistance to E. tracheiphila and spatially explicit autocorrelation models of disease spread in our fields revealed that differences in the incidence of wilt disease were not due to spatial artifacts (Ferrari et al. 2007). Retrospective analyses of our field data revealed that there were no differences in the incidence of wilt disease on inbred and outbred plants prior to the initiation of flowering (late June) but, once flowering began, those plants that were producing the most flowers per week were the most likely to contract wilt disease. Together, these findings suggested that E. tracheiphila could be transmitted through the flowers when A. vittatum aggregate in the flowers to feed and mate in the mid to late morning. The goal of the research described here is to determine if E. tracheiphila can be transmitted via the nectary of Texas gourd flowers. Specifically, we develop a Real-Time PCR-based technique to identify E. tracheiphila in beetle frass and we use this technique to show that a high proportion of flowers in the field contain E. tracheiphila contaminated frass; we determine that most flowers under field conditions contain frass that has fallen directly onto the nectary; we show that disease progression occurs when E. tracheiphila is placed onto the nectaries of flowers; and we transform E. tracheiphila with green fluorescent protein and track the movement of the bacteria through the nectary and into the xylem.

Materials and Methods Study Species and Field Plots The Texas gourd, Cucurbita pepo subsp. texana is an annual monoecious vine with indeterminate growth and reproduction. It is native to northern Mexico, Texas, and the lower Mississippi River drainage area and is thought to be either the wild progenitor of the cultivated squashes (C. pepo subsp. pepo) or an early escape from cultivation (Decker and Wilson, 1987, Decker-Walters, 1990, Lira et al., 1995, Decker-Walters et al., 2002). After germination and seedling emergence, there is a period of vegetative growth (5–7 nodes). Thereafter, most nodes produce one large yellow flower (either male or female) in the axils of each leaf. The flowers last for only one morning and are pollinated by bees, especially squash bees of the genera Peponapis and Xenoglossa

15 (Winsor et al., 2000, Avila-Sakar et al., 2001). Male flowers are oriented vertically and are displayed above the leaves on long pedicels. The base of the three fused filaments of the androecium surrounds and covers the nectary which is only accessible to the tongues of the bees via three narrow slits. The female flowers are oriented horizontally and the nectary forms a ring around the base of the style (Nepi et al. 1996). The flowers abscise 24-48 hrs after anthesis and then the pedicel begins to senesce (except on pollinated female flowers where the ovary and pedicel persist). Field studies were conducted in four 1 acre fields at The Pennsylvania State University Agriculture Research Farm at Rock Springs, Pennsylvania. In late May 2008 we transplanted 18 Texas gourd plants and 18 backcrossed (backcross 4) plants (Texas gourd X cultivated squash with Texas gourd as the recurrent parent) from each of five maternal families (180 plants per field, 720 total plants). The fields were not sprayed with insecticide and viral and wilt diseases were allowed to occur naturally. The plants in each field were monitored throughout the growing season for incidence of wilt disease (field diagnosis confirmed by isolating Erwinia from diseased plants and using the isolate to infect greenhouse grown plants (see Ferrari et al. 2007 for techniques)). In order to determine the number of beetles that aggregate in the flowers, we counted the number of beetles in one male and one female flower on all healthy (no visible symptoms of viral or wilt disease) plants that produced at least one flower on August 1, 10, and 18 (N = 856 male flowers; N = 397 female flowers). If a plant produced two or more male or female flowers, we randomly selected the male and female flower before looking into the flower to count the beetles. These counts were made between 9 and 11 in the morning and represent the number of beetles at any one time in a flower (i.e., these are not cumulative numbers of beetles that visit a flower over the course of one morning). To determine the effects of flower type (male or female), date (Aug 1, 10, 18) and their interaction on the number of beetles per flower, we performed an analysis of variance (Proc Mixed. SAS Institute 2002). On August 1, 10, and 18 we also surveyed male flowers in our fields and recorded the proportion of flowers in which beetles had damaged (chewed) the base of the anther filaments and exposed the nectary and the number of flowers in which beetle frass was

16 found on the nectary pad and/or in the nectary. Voucher specimens were placed in the Frost Entomological Museum, Pennsylvania State University.

RealTime PCR Method for Identifying E. tracheiphila in Beetle Frass To determine the proportion of flowers that were exposed to beetle frass containing E. tracheiphila we collected a random sample of 16 male flowers on August 1 and 10 and September 9 in the late morning (48 total flowers although one sample was later discarded). We removed the beetle frass from flowers with clean toothpicks, placed it into 1.5 ml microcentrifuge tubes and froze the samples at -20°C. DNA was extracted from these samples in 50 – 100μl extraction buffer according to the recommended protocol with the PREPGEM Bacteria kit (ZyGEM Corp Ltd. Hamilton, New Zealand). In order to break up the frass and release the DNA an initial incubation step at 72°C for 15 minutes was incorporated before adding the lysozyme. We used RealTime PCR to detect E. tracheiphila genomic DNA. We chose to amplify the outer membrane protein (EtOMP), because gene sequences for EtOMP from several different strains of E. tracheiphila are published (NCBI Accessions AF220817- AF220822). Primers were designed to anneal to conserved regions of EtOMP of E. tracheiphila, such that all published strains of E. tracheiphila are indiscriminately detected (Et73: GGCGATCACGACACAGTTGT, Et140: CAGTTTTTGGTCAGGGCATACTC). PCRs were run in 20μl reactions using the 2x SybrGreen Mastermix (Bio-Rad Inc., Hercules, CA), 0.5μM final primer concentration, and 5μl of the DNA preparation as template. The cycling program consisted of an initial denaturation at 95°C for 3 minutes, 40 cycles of 95°C for 15 seconds, 58°C for 15 seconds, 72°C for 30 seconds, and a final extension at 72°C for 5 minutes. Fluorescence reads were taken at the end of each extension step. The identity of the PCR products was verified by cloning and subsequent sequencing. In addition, the specificity of our primers was verified using E. coli, floral tissue and uninfected beetle frass as negative controls and all 5 of our 2006- 2008 field isolates of E. tracheiphila and the E. tracheiphila we transformed with GFP

17 (see below) as positive controls. Every sample was analyzed in triplicate. If the CT - values of all three technical replicates were within one cycle of each other and the melting curves showed a clean peak at the correct melting temperature, the samples were considered to contain E. tracheiphila, given that the three replicates had clear melting curves in all 3 technical replicates. If the CT - values varied by more than one cycle, the samples were considered to contain trace amounts of E. tracheiphila and if the melting curves showed no detectable product the samples were considered negative for E. tracheiphila.

Nectary Inoculations of E. tracheiphila To determine if E. tracheiphila on the nectaries of Texas gourds could result in wilt disease, we grew 250 Texas gourd plants in one gallon pots in Pro-Mix BX with fungicide (Premier Horticulture Inc., Riviere-du-Loup, Quebec, Canada) potting soil in a greenhouse. At ~10 weeks post emergence, we began to inoculate the nectaries of male flowers. On the day that a flower was inoculated, we removed the nectar using a 50μl capillary tube, and placed 100μl of E. tracheiphila inoculum onto the nectary with a blunt 18 gauge needle and 1ml Tuberculin syringe. The inoculum was prepared from E. tracheiphila isolated the previous summer from field grown plants. The isolates were grown in nutrient broth supplemented with extra peptone and placed in a 15% glycerol solution and frozen at -800C (see Ferrari et al. 2007). Samples of the frozen isolates were thawed and streaked onto plates of Difco (Sparks, MI) nutrient agar supplemented with extra agar and peptone (NAP) (see DeMackiewicz et. al 1998) for 5-7 days. For greenhouse inoculations the resulting colonies were dislodged with a sterile L-rod and

transferred into distilled deionized water (diH2O). The average concentration of E. tracheiphila cells in each inoculum was 5.35x108 cells/ml and was determined using a spectrophotometer at OD 600. Each day we also performed 2 control inoculations: one

plant was inoculated with diH20 and another plant was inoculated directly through the vasculature by using a thin needle, piercing leaf and stem vasculature with the E. tracheiphila suspension used that day. Five flowers were inoculated on each plant unless a plant developed symptoms of wilt disease in which case we performed no additional floral inoculations. We then recorded the first day of wilt symptoms and wilt disease

18 development. To be certain that wilting was caused by E. tracheiphila, we re-isolated the bacteria from inoculated plants, confirmed colony morphology, and infected additional plants using vasculature inoculations. A few months later, we repeated the above experiment using female flowers that had been either pollinated or unpollinated. This experiment used a new set of plants and E. tracheiphila that recently had been isolated from field infected plants.

GFP-Transformed Erwinia tracheiphila and Infection via the Nectary In order to examine the progression of E. tracheiphila through the nectary and into the xylem of the pedicel of flowers, field-collected E. tracheiphila were transformed with a high copy number plasmid (pfdC4Z’-gfp) expressing green fluorescent protein (GFP). The plasmid pfdC4Z’-gfp contains the GFP mutant2 coding sequence with the T7 gene 10 ribosomal binding site just upstream of the ATG start codon and a chloramphenicol resistance gene. The mutant 2 GFP protein has a 19 fold higher fluorescence emission intensity than the wild type GFP protein at 488nm (Cormack et al. 1998). Plasmids for transformation of E. tracheiphila were maintained and multiplied in E.coli (One Shot® Top10 cells from Invitrogen Inc. Carlsbad, CA). Transformants in E.coli were screened for successful transformation by PCR with the following primers: GFPL-502: CTGGGTATCTCGCAAAGCAT and GFPR-670: GGTGATGTTAATGGGCACAA. Electro-competent E. tracheiphila were generated following standard procedures. Briefly, to transfer E. tracheiphila into liquid culture 1ml of liquid nutrient broth (8 grams nutrient broth, 5 grams peptone per liter distilled water) was inoculated with E. tracheiphila which had been propagated on plates. These liquid starter cultures were incubated overnight at 28°C, with vigorous shaking. The following morning 10ml of nutrient broth were inoculated with 250μl of this starter culture and incubated overnight at 28°C, with vigorous shaking. The following morning the overnight cultures were washed and concentrated. For each wash the bacterial suspension was centrifuged at 5000 rpm for 10 minutes at 4°C, the supernatant was discarded and the bacteria were re-suspended, first in 2.5ml, then in 2ml ice cold, sterile, double distilled water, and thirdly in 1ml ice cold, sterile 10% glycerol. Final bacterial pellets were re- suspended in 200μl ice cold, sterile, 10% glycerol. Aliquots of 40μl of this final bacterial

19 suspension were transformed with 200ng of the plasmid pfdC4Z’-gfp by electroporation with voltage set at 2500V for 5 msec (actually delivered voltage was 2470V). Directly after electroporation the bacterial suspension was transferred to a 2ml Eppendorff tube containing 500μl nutrient broth. Bacteria were allowed to recuperate for 1 hour at room temperature shaking horizontally on a rotary shaker. To select for GFP expressing E. tracheiphila 4.5 ml of nutrient broth containing 250μl chloramphenicol (20μg/μl in ethanol) were inoculated with the suspension of recuperated E. tracheiphila, and incubated overnight at 28°C, vigorously shaking. 21 hours later aliquots of this bacterial suspension were visually inspected with a Zeiss Axiover light microscope at 40X for live E. tracheiphila expressing GFP. For further maintenance and propagation of GFP expressing E. tracheiphila, bacteria were streaked on nutrient plates containing Chloramphenicol (final concentration of 1μg/μl). Glycerol stocks were made for long term storage. We grew 30 Texas gourds in the greenhouse and used the same procedure as in the greenhouse inoculation experiments to inoculate through the nectaries of male flowers. We inoculated 1-5 flowers on each plant for a total of 75 flowers. We collected the male flowers and their pedicels at 24 hours after inoculation. Each pedicel ranged from 7.9-12 cm in length. Immediately after collection of the flowers, we used a tool made of 10 razor blades soldered together with a 1mm distance between the blades to make cross-section cuts at 1mm intervals along the pedicel starting at the nectaries to 1cm below the nectary on 71 of the flowers. The tool was sterilized using 70% EtOH after every cut. On the remaining 4 floral pedicels, we used the razor blade tool to longitudinally slice the entire length of the pedicel which resulted in 3-6 longitudinal slices per pedicel. Each longitudinal slice and each 1 mm cross sectioned segment was mounted on a slide in a drop of water (to prevent desiccation) and examined on both sides for presence of E. tracheiphila expressing GFP at 40x under the light microscope with a GFP filter.

Results A mixed effects model analysis of variance revealed that male flowers had significantly more beetles per flower than female flowers (2.4 ± 0.1 vs. 1.9 ± 0.1; Least

20 Square Means ± SE; F = 8.49; df = 1,1247; P < 0.004) (Fig. 2.1A). Our surveys of flowers in the late morning, just before the flowers closed, revealed that the beetles had chewed the base of the anther filaments so that all or part of the nectary was exposed on 68% of the male flowers (342 of 500) (Fig. 2.1B,C). On 52% of the male flowers, frass was visible on the nectary pad or in the nectary (260 of 500) (Fig. 2.1C, D). In some flowers there was evidence that beetles had fed on nectary tissue and in a few flowers the nectary had been chewed so severely that the vascular bundles in the pedicel were visible. Therefore in the majority of the male flowers all the prerequisites are met for E. tracheiphila from the frass to enter the plant through the vascular bundles of the pedicel. From early August until early September the percentage of plants with symptoms of wilt disease increased from 6% to 19% while the cumulative incidence of wilt disease (living plants with symptoms plus plants that died of wilt disease) increased from 18% to 43% (Table 2.1). On August 1, 10 and September 9, we collected frass from 47 male flowers and screened it for presence of E. tracheiphila using RealTime PCR. We detected E. tracheiphila unequivocally in 39 samples, trace amounts in 6 samples, and were unable to detect E. tracheiphila in 2 samples. To verify that E. tracheiphila was present in the frass samples with trace amounts, we ran these samples twice and we were able to duplicate the results. In short, 95% of a randomly sampled set of male flowers in our fields during August and early September contained cucumber beetle frass containing E. tracheiphila. The greenhouse inoculation experiment, in which cultured E. tracheiphila were placed directly onto the nectary after nectar removal, showed that inoculated plants contracted wilt disease through both male and females flowers (Table 2.2). Plants receiving nectary inoculations via pollinated female flowers had a greater probability of contracting wilt disease than plants inoculated via unpollinated female flowers (χ2 = 5.04; df = 1; P < 0.025). Plants that were inoculated via female flowers (either pollinated or unpollinated) had a greater probability of contracting wilt than plants pollinated via the nectaries of male flowers (both χ2 > 18; df = 1; both P < 0.001). It should be noted, however, that the male and female flowers were inoculated in separate greenhouse experiments using different E. tracheiphila isolates from the field.

21 Following nectary inoculations of 75 male flowers with GFP-transformed E. tracheiphila, we found that 2 of the 4 longitudinally sectioned pedicels had GFP- expressing E. tracheiphila in the first 1mm below the nectary but no bacteria were detected any further down the pedicel at 24 hours after inoculation (Fig 2.2A). These findings suggested to us that the movement of GFP transformed E. tracheiphila through the nectaries and into the pedicel was rather frequent and that we should concentrate our examinations on the 1 cm of pedicel directly below the nectary. In the longitudinal sections, however, it was difficult to determine where the E. tracheiphila were actually located in the pedicel (Fig 2.2A) because many of the bacteria were floating in the drop of water (i.e., the tissue in which they had been associated was damaged by the longitudinal section). Consequently, the remaining 71 pedicels were cross-sectioned in 1mm segments from the nectary to 1 cm below the nectary. No GFP-expressing E. tracheiphila were observed more than 5 mm from the nectary at 24 hrs post-inoculation. We found GFP-expressing E. tracheiphila on the surface of 1-4 cross sections (i.e., the top or bottom of one to four 1 mm segment(s)) from 26 different pedicels. In these cross sections, we observed that the bacteria aggregate in small colonies (less than 20μm across) inside the vascular (xylem) tissue (Fig 2.2B-D). Because colony size is small and because we could only visualize superficial colonies of GFP-expressing E. tracheiphila (i.e. colonies located at either end of a 1 mm segment) we suspect that our findings greatly underestimate the number of colonies per pedicel and, possibly, the proportion of pedicels with GFP-expressing E. tracheiphila as well as the distance from the nectary that the bacteria have traveled in 24 hrs. We also attempted to examine the movement of GFP-expressing E. tracheiphila from the surface of the nectary to the pedicel below the nectary. Unfortunately, the tissue of the nectary autofluoresced so brightly that individual GFP-expressing bacterial cells could not be detected against this background fluorescence. In the male flowers of C. pepo an abscission layer forms between the end of the pedicel (bottom of the nectary) and they abscise within 24-48 hours after the flowers close. Our data indicate that E. trachephila are able to move through the nectary tissue and some of the pedicel before the formation of the abscission layer.

22 Discussion Strong circumstantial evidence supports the long held presumption that E. tracheiphila is transmitted by cucumber beetles via foliar feeding when frass containing E. tracheiphila falls onto the open wounds (Leach 1964). For example, caged bioassays indicate that a single beetle can transmit wilt disease (Brust 1997b, Fleischer et al. 1999); A. vittatum is known to harbor E. tracheiphila in its gut (Fleischer et al. 1999, Garcia- Salazar et al. 2000); the frass of A. vittatum contains E. tracheiphila and the frass is capable of causing wilt disease following pin prick inoculation (Mitchell and Hanks 2009). The study presented here provides strong circumstantial evidence that cucumber beetles also transmit E. tracheiphila via the floral nectaries on C. pepo ssp. texana. Specifically, we show, under field conditions, that the beetles aggregate in the flowers in the mid to late morning, that a large proportion of the flowers by late morning contain frass contaminated with E. tracheiphila, that the beetles frequently damage the base of the anther filaments exposing the nectary, and that frass often accumulates on or in the nectaries of male flowers. Moreover, we show that greenhouse grown plants contract wilt disease when E. tracheiphila is deposited onto nectaries of both male and female flowers, and that GFP-transformed E. tracheiphila, when placed onto the nectary of male flowers, enter into the nectary and proceed down the pedicel before floral abscission. Interestingly, Erwinia amylovora (Burr.) and Erwinia pyrifoliae (Kim et al. 1999), the causative agents of fire blight and Asian pear blight, respectively, are known to also to enter their hosts via the nectaries (Thomson 1986, Kim et al. 1999, Buban et al. 2003). Our greenhouse inoculation experiments revealed that E. tracheiphila can be transmitted via the nectaries of male flowers and via unpollinated and pollinated (ovary retained between the pedicel and nectary) female flowers. It should be noted, however, that the difference between male and female flowers that we observed in the probability of contracting wilt disease could be an artifact. The greenhouse inoculations of male and female flowers used different field isolates and the isolates differed in the amount of time that they were cultured and stored in the lab. Our casual observations over the last seven years have indicated that isolates can vary in virulence and that virulence tends to decrease with culturing and storage. However, even if the 3-fold difference between male and female flowers in the probability of contracting wilt disease is not an artifact, we

23 suspect that plants are more likely to contract wilt disease via male flowers under field conditions because a) Texas gourds (and C. pepo in general) produce approximately 7 male flowers for every female flower (Avila et al 2001, Stephenson et al. 2004), b) male flowers are oriented vertically (frass accumulates on the bottom of the flower near the nectaries) whereas female flowers are oriented horizontally; c) male flowers attract significantly more beetles per flower than female flowers which would increase both exposure to and the amount of E. tracheiphila in male flowers, and d) beetles chew the filaments of the anthers (which brings the beetles into proximity with the nectary) but do not chew the stigma/style of female flowers. In Central Pennsylvania, and throughout the Northeastern US, cucumber beetle populations increase through the growing season as the resident beetles undergo successive generations (A. vittatum) and other species (mostly D. undecipunctata howardi) migrate into cucurbit fields (Ferrari et al. 2007). In a three year study, Fleischer et al. (1999) found that 7-10% of newly emerged beetles in the spring tested positive for E. tracheiphila and this number increased to 39-78% during the growing season. Yao et al. (1996) found a strong positive relationship between cucumber beetle density and the incidence of wilt disease over the course of the growing season. In this study, we found 95% of the male flowers in August-early September had frass containing E. tracheiphila and that 52% of the flowers had frass on or in the nectary. Our previous studies have shown that leaf damage by cucumber beetles accumulates on all plants in our fields throughout the growing season and that healthy plants are making multiple male flowers per day in August (Stephenson et al. 2004, Du et al. 2008). Consequently, it is likely that, by August, virtually every plant in this study is exposed to E. tracheiphila on a daily basis via the leaves and/or the flowers. Not surprisingly, the percentage of living plants with wilt disease increased from 6% to 19% from August 1 to September 9 while the cumulative incidence of wilt disease increased from 18% to 43% over the same time period. Although bacterial wilt disease was clearly epidemic in our fields, it also is evident that only a small proportion of the field exposures to E. tracheiphila result in disease progression. Single beetle caged bioassays, in which individual beetles are placed into a cage with a seedling under greenhouse conditions, reveal that the proportion of

24 beetles that test positive for E. tracheiphila is 5X greater than the proportion that transmit wilt disease (Fleischer et al. 1999). Moreover, it is generally agreed that tender greenhouse seedlings, such as those used in caged bioassays, are far more susceptible to wilt disease than field hardened mature plants (e.g., Rand and Endlows 1920). Early in the growing season when A. vittatum populations are small, males release aggregation pheromones when feeding on the leaves of cucurbit seedlings which attract both male and female cucumber beetles (Smyth and Hoffmann 2003). This leads to concentrated feeding on some seedlings. Brust (1997b) observed that wilt disease is more likely to occur on seedlings where concentrated feeding occurs. These findings suggest that the dosage of E. tracheiphila is important for disease transmission under field conditions. As Cucurbita plants grow, however, foliar feeding becomes less concentrated (Du et al. 2008) and, perhaps, the leaves become more resistant to E. tracheiphila invasion. However, cucumber beetles are attracted to the flowers of C. pepo over relatively long distances (Andersen and Metcalf 1986, Lampman et al. 1987, Lampman and Metcalf 1988) where they aggregate to mate and feed. Although our Real-Time PCR analyses of E. tracheiphila on floral tissue provide no quantitative measure of the dose of E. tracheiphila found in flowers, we suspect that the aggregation of cucumber beetles in flowers effectively concentrates the frass (and consequently E. tracheiphila) from many beetles in the vicinity of the nectaries, which, in turn, provides access to the vascular system that supplies the nectaries. If true, this second mode of transmission would play a key role in the maintenance of wilt disease epidemics in fields at a time when the plants are becoming more resistant to foliar transmission of E. tracheiphila. Finally, our previous studies (Stephenson et al. 2004, Ferrari et al. 2007, Du et al. 2008) revealed that inbred Texas gourds experienced greater leaf damage by cucumber beetles than outbred gourds but had a lower incidence of wilt disease--a disease that was only known to be transmitted via foliar feeding. Transmission of E. tracheiphila via floral nectaries resolves this conundrum. The outbred plants produced significantly more flowers than the inbred plants (which would increase the frequency of exposure). Moreover, we have shown that the flowers of the outbred plants produce significantly more of the volatiles that are known to attract cucumber beetles to the flowers (Ferrari et

25 al. 2006) which, in turn, would increase both exposure to and the dose of E. tracheiphila in the flowers.

Acknowledgements We thank Franz Lichtner for field and greenhouse assistance; Jim Tumlinson for use of his RealTime-PCR machine; Richard Cyr for guidance and use of microscopy equipment; Klaus Geider for the GFP plasmids; Bob Oberheim and his staff for use of the Horticulture Farm at the PSU Experimental Farms at Rock Springs, PA; and Tony Omeis for advice, assistance and use of the Biology Greenhouses. This research was supported by NSF grant DEB02-35217 to AGS and JAW.

A B

1d 1c

CDD

Figure 2.1. Cucumber beetles in Texas gourd flowers and feeding damage to floral tissues. A. Male flower with cucumber beetles inside corolla (1A) and feeding damage to the petals (2A). B. Close-up of two beetles inside male flower corolla. C. Anther and anther filament with beetle damage exposing nectary tissue and cucumber beetle frass inside nectary (1C). D. Female flower with floral tissue damage and cucumber beetle feeding on nectary tissue (1D).

27 3 1 1 2 1B

100µm 100µm A B

100µm C 100µm D

Figure 2.2. Images of E. tracheiphila expressing GFP inside pedicel vascular bundles of C. pepo ssp. texana (Texas gourd) 24 hours after inoculation at 40x under light microscopy using a GFP filter. Each image is from a different pedicel, was not cropped, and includes a scale of 100 microns. A. Longitudinal section of the pedicel immediately below the nectary. Nectary would be positioned in the top right-hand corner. 1A. Spiral thickenings of intact xylem column. 2A. Colonies of E. tracheiphila cells inside xylem column. 3A. E. tracheiphila cells released from other xylem columns when the vascular bundle was longitudinally sectioned. These cells are floating freely in water as a result of the longitudinal cut. B. Cross section of vascular bundle 1 mm below nectary. 1B. Colonies of E. tracheiphila cells inside xylem column. C. Cross section of vascular bundle with colony forming at 3 mm below nectary. GFP-expressing E. tracheiphila are fluorescing inside nearly all xylem columns. 1C. Same image with vascular bundle

28 columns circled. D. Cross section of vascular bundle 5mm below nectary with GFP- expressing E. tracheiphila fluorescing inside one xylem column.

Date in 2008 1-Aug 10-Aug 9-Sep Number of Living Plants 627 593 514 Number of Plants that Died from Wilt Disease 93 127 206 Number of Living Plants 40 79 100 with Wilt Disease (%) (6%) (13%) (19%) Cumulative Incidence of Wilt Disease 0.18 0.29 0.43

Table 2.1. Percentage of Plants with Wilt Disease and the Cumulative Incidence of Wilt Disease on the Three Dates of Cucumber Beetle Frass Collection from Male Flowers.

30 Number of Plants Type of Flower with Symptoms after Inoculated on Each Plant N 20 d (%) Pollinated female 62 43 (69%) Unpollinated female 90 46 (51%) Male 143 34 (24%) Total flowers 295 123 (42%) Vascular inoculation 21 21(100%) Inoculum with no E. tracheiphila 21 0 (0%)

Table 2.2. Percentage of Plants Infected with Erwinia tracheiphila via Floral Nectaries under Greenhouse Conditions.

31 CHAPTER 3

Antimicrobial Nectar Inhibits a Floral Transmitted Pathogen of a Wild Cucurbita pepo (Cucurbitaceae)

Miruna A. Sasu2, Kelly L. Wall2 and Andrew G. Stephenson2,3

2Department of Biology, Center for Chemical Ecology, and Center for Infectious Disease Dynamics, The Pennsylvania State University, University Park, PA 16802 3Author for correspondence

Note: Accepted for publication to American Journal of Botany.

Abstract Floral nectars of many species contain antimicrobial chemicals but their function in nectar is subject to debate. Previously, we have shown that Erwinia tracheiphila, the causative agent of bacterial wilt disease in cucurbits, can be transmitted via the floral nectaries. Here, we use a disk diffusion assay (DDA) to determine the antimicrobial effects of nectar from a wild gourd on lawns of E. coli and E. tracheiphila. We also use E. tracheiphila to inoculate flowers of wild gourd plants, with and without nectar. The DDA showed that paper disks saturated with 10 μl of nectar inhibited the growth of E. coli on a larger area of the lawn than 40% glucose but a smaller area than 5% ampicillin for 12 hrs. On lawns of E. tracheiphila, nectar inhibited growth on a larger area than glucose for 24 hrs and there were no significant differences between ampicillin and nectar for12 hrs. A significantly larger proportion of the plants inoculated via flowers without nectar contracted wilt disease than plants with nectar. These findings indicate that nectar reduces E. tracheiphila transmission via the nectaries and reveal the potential for floral transmitted pathogens to influence the evolution of floral traits.

Introduction The showy, often fragrant, flowers of animal pollinated plants function to attract and reward pollinators. However, the same flowers can also serve as an easily exploited resource for a variety of natural enemies (Strauss and Whittall, 2006). Consequently, plants face a dilemma: their flowers must be apparent and readily accessible to the visitors that provide the pollination service while simultaneously discouraging the visitors that would exploit their resources without effecting pollination. For example, the presence of glucose, fructose, sucrose and other simple sugars in floral nectar makes nectar an excellent medium for the growth of microbes (Herrera et al., 2008). Not surprisingly, microbes have been reported in the nectars of many species and, for many species, the sugar concentration of nectar tends to decrease with time after anthesis while the ethanol concentration in nectar tends to increase—a finding that is consistent with microbial degradation of sugar (see Nicolson and Thornburg, 2007; Herrera et al., 2008). In addition to sugars, however, the nectars of many species contain secondary chemicals, such as phenolic compounds, and proteins that are known to have antimicrobial properties (Baker and Baker, 1983a; Adler, 2000; Nicolson and Thornburg, 2007, Hancock, et al., 2008). For example, a protein that generates very high levels of hydrogen peroxide (a powerful antimicrobial agent) was found in the nectar of nine of the 15 species examined (Carter and Thornburg, 2000) and some species are known to produce other proteins that have antifungal and antibacterial properties (Nicolson and Thornburg, 2007). Because microbial degradation of the sugars in nectar would reduce the average nectar reward per flower and perhaps also decrease the efficiency of the pollinators due to the increased concentration of ethanol in nectar, antimicrobial compounds in nectar have been hypothesized to function in the attraction and reward of pollinators (see Adler, 2000; but see Weins et al., 2008). Recently, Carter and Thornburg (2004) have noted that wind, pollinators and other floral visitors can carry microbes, including plant pathogenic microbes, into the reproductive tract of the flowers. They hypothesize that antibiotic compounds in nectar function to protect the ovary from invading microorganisms. Although these hypotheses are neither exhaustive (e.g., the presence of antibiotic

33 chemicals in the nectar could be a pleiotropic by-product of their production elsewhere in the plant (Adler, 2000)) nor mutually exclusive, there are no studies that explicitly examine the effect of antimicrobial nectar on pollination or the incidence of a disease that is transmitted via the floral organs. Erwinia tracheiphila (E.F. Smith) Holland (Enterobacteriaceae) is the causative agent of bacterial wilt disease, an important disease of cultivated cucurbits (cucumbers, melons, and squash) and wild gourds (Cucurbita spp.). E. tracheiphila is vectored by cucumber beetles (Diabrotica spp. and Acalymma spp.). In the eastern United States, E. tracheiphila overwinters in the digestive tract of cucumber beetles (Fleischer et al., 1999; Garcia-Salazar et al., 2000) and transmission occurs when fecal pellets containing E. tracheiphila land on leaf wounds at the sites of feeding damage (Leach, 1964). After entering the plant the bacteria proliferate in the xylem where they secrete an exopolysaccharide matrix that cuts off the water supply resulting in wilting. Wilt symptoms typically develop 7-15 days after infection and, once symptoms appear, death occurs in 1-3 weeks (Yao et al., 1996). After feeding on the leaves, cucumber beetles aggregate in the flowers in the mid morning to feed and mate (Darlington 2006). Our recent studies have shown that E. tracheiphila is also transmitted via the floral nectaries when the E. tracheiphila contaminated cucumber beetle fecal pellets fall onto or near the floral nectaries (Sasu et al., 2010). The goals of the study presented here are to determine a) if the nectar of a wild gourd (Cucurbita pepo ssp texana (L) Andres) has antibiotic properties; b) if the antibiotic properties of the nectar change with the amount of leaf damage by cucumber beetles (because chemical defenses in the leaves of Cucurbita pepo are known to be induced by damage); c) if the strength of the antibiotic properties varies over time (because the flowers abscise 24-48 hrs after anthesis); and d) if the antibiotic properties of the nectar protect the nectaries from invasion by E. tracheiphila.

Materials and Methods Study Species Cucurbita pepo ssp. texana (wild gourd) is an annual monoecious vine with indeterminate growth and reproduction. It is native to northern Mexico, Texas, and the

34 lower Mississippi River drainage area and is thought to be either the wild progenitor of the cultivated squashes (C. pepo ssp. pepo) or an early escape from cultivation (Decker and Wilson, 1987; Decker-Walters, 1990; Lira et al., 1995; Decker-Walters et al., 2002). After germination and seedling emergence, there is a period of vegetative growth (5-7 nodes) after which most nodes produce one large yellow flower (either staminate or pistillate) in the axils of each leaf. The flowers last for only one morning and are pollinated by bees, especially squash bees of the genera Peponapis and Xenoglossa. The fruits are round to oval in shape with a volume of 175-450 ml and typically contain 150- 300 seeds that weigh 20-40 mg (Winsor et al., 2000; Avila-Sakar et al., 2001). The leaves and other organs of this wild gourd produce bitter compounds called cucurbitacins (oxygenated tetracylic triterpenes) that deter most herbivores (Tallamy, 1985; Metcalf and Rhodes, 1990). However, cucumber beetles are adapted to feed on cucurbitacins in the leaves, and are found throughout the native ranges of Cucurbita species (Robinson and Decker-Walters, 1997). Cucumber beetles feed on leaves and flowers and cause a characteristic pattern of holes (typically 1-1.5 cm in diameter) in the portions of the leaves serviced by the smallest veins. Leaf damage over the entire growing season by cucumber beetles has been shown to substantially reduce yield in cultivated squash (e.g., Tallamy and Krischik, 1989) and reduce reproductive output in the wild (free living) gourd (Quesada et al., 1995; Stephenson et al., 2004; Du et al., 2008). Cucumber beetles are also the only known vector of the deadly bacterial pathogen, Erwinia tracheiphila Smith (Yao et al., 1996), which is the causative agent of bacterial wilt disease. Disk diffusion assays To determine if the nectar of the wild gourd has antibiotic properties, we performed disk diffusion assays (DDA) (Gabhainn et al., 2004). In brief, we grew lawns of Escherichia coli and Erwinia tracheiphila on separate sterile culture plates and placed four filter paper disks on each plate. The four disks were saturated with a) 5% ampicillin solution, b) 40% glucose, c) nectar from field grown plants with light beetle damage, and d) nectar from field grown plants with heavy beetle damage. We then measured the area of inhibition around each disk (see Fig 1 for schematic). Field nectar collection

35 During the 2008 field season we collected nectar for the diffusion assays from male flowers on healthy (no visible symptoms of disease), unsprayed wild gourd plants grown at The Pennsylvania State University Agriculture Research Farms at Rock Springs, Pennsylvania. The day before the DDA was performed we identified male flower buds that would open the next day and we tied them shut with a twist tie so that the bees would not have early access to the nectar the following day. We also assessed foliar damage by cucumber beetles to the branch bearing the flower bud using a linear scale from 0-5 where 0 = no visible damage to foliage and 5 = significant visible damage to all leaves with at least one leaf having greater than 50% of the leaf area removed (see Stephenson et al., 2004 for details). Damage levels of 0-2 were considered light beetle damage and damage levels of 3-5 were considered heavy beetle damage. Approximately 20 male flowers each were collected in the early morning from plants with light and 20 from plants with heavy damage each day. The flowers were placed on ice and brought to the lab where the nectar was collected using microcapillary tubes. The nectar from each damage level was pooled and stored on ice during the extraction process, and then used immediately for the DDA. Plate preparation The E. tracheiphila used in the DDA was originally isolated and cultured from our field plants with symptoms of wilt disease. To verify that the cultured bacteria were indeed E. tracheiphila, we inoculated seedlings grown in a greenhouse and recorded wilt symptoms (see Ferrari et al., 2007). The E. coli used in the DDA was obtained from Invitrogen (Carlsbad, CA; One Shot® Top 10 chemically competent E. coli). Both bacteria were stored separately in 15% glycerol solution at -80C. Sterile culture plates (100 x 15 mm) were prepared using nutrient agar supplemented with agar and peptone (NAP). To prepare bacteria for the DDA, we thawed the glycerol stock to room temperature and applied 100 µl inoculum of either E. tracheiphila or E.coli directly onto the NAP plates using a sterile L-rod (see DeMackiewicz et al., 1998). We allowed the surface of the plates to dry and then placed four sterile individual filter paper disks with a diameter of 6 mm equidistant from each other on the surface of the agar medium. To each disk we added 10 μl of one of four treatment solutions: 5% ampicillin solution (50 mg/ml), 40% glucose solution, nectar from lightly damaged plants and nectar from

36 heavily damaged plants (Fig. 1). The DDA was performed on each of 4 days (July 16, 25, August 8, 13). Each day, 15 plates were prepared per bacteria for a total of 30 plates per day (60 plates E. coli and 60 plates E. tracheiphila in total). The plates were incubated at 26C in the dark and the diameter of inhibition zone was measured at 8, 12 and 24 hours. The area of the inhibition zone was later calculated (A = πr2 – area of filter paper disk). Statistical analysis To determine the effects of Treatment (the 4 solutions added to the filter paper), Time (8, 12, or 24 hrs), Date (the 4 dates in which the DDA was performed), and the interaction between Date and Treatment on the area cleared around each filter paper disk on each plate, we performed a repeated measures analysis of variance (Proc. Mixed, SAS Inst., 2002) for each of the two types of bacterial lawns. Date was treated as a fixed effect because preliminary studies indicated that nectar volume per flower varied with soil moisture (recent rainfall) and we felt that the antimicrobial properties of the nectar might vary with nectar volume/dilution. The repeated variable was Time (random) and we used autoregressive covariance structure. To measure the variance among inhibition zones across Time, the repeated variable (Time) was calculated on each Treatment within each Plate on each Date that the assay was performed. We also performed multiple pairwise comparisons among the four Treatments at each point in time (separately for 8, 12, and 24 hrs) while controlling for Type I error rates, to identify differences among the least square means of the four Treatments at each point in time (Tukey-Kramer, SAS Inst., 2002) for each of the two types of bacteria. Nectary Inoculations of E. tracheiphila To determine if floral nectar alters the probability of infection by E. tracheiphila via the nectaries, we grew 120 wild gourd plants in February of 2009 in one gallon pots in Pro-Mix BX with fungicide (Premier Horticulture Inc., Riviere-du-Loup, Quebec, Canada) potting soil in a greenhouse. At ~10 weeks post emergence, we began to inoculate the nectaries of male flowers. On the day that a flower was inoculated, we a) removed the nectar using a 50μl capillary tube, and placed 100μl of E. tracheiphila inoculum onto the nectary with a blunt 18 gauge needle and 1ml Tuberculin syringe; or b) we placed 100μl of E. tracheiphila inoculum onto the nectary with a blunt 18 gauge needle and 1ml Tuberculin syringe without removing the nectar. The inoculum was

37 prepared from E. tracheiphila isolated the previous summer from field grown plants. The isolates were grown in nutrient broth supplemented with extra peptone and placed in a 15 % glycerol solution and frozen at -80C (see Ferrari et al., 2007). Samples of the frozen isolates were thawed and streaked onto NAP plates and incubated at 26C for 5-7 days. The resulting colonies were dislodged with a sterile L-rod and transferred into deionized water (diH2O). The average concentration of E. tracheiphila cells in each inoculum was 5.47x108 cells/ml and was determined using a spectrophotometer at OD 600. Each day that we performed inoculations, we also inoculated 2 control plants: one plant with diH20 and another directly through the stem vasculature with the E. tracheiphila used that day. Five flowers were inoculated on each plant (all with nectar or all without nectar) unless a plant developed symptoms of wilt disease in which case we performed no additional floral inoculations. We recorded the first day of wilt symptoms and we followed disease progression until the plant died. To be certain that wilting was caused by E. tracheiphila, we re-isolated the bacteria from the wilting plants and confirmed colony morphology.

Results The repeated measures analysis of variance revealed that the Treatment applied to the filter paper disks, the Date in which the disk diffusion assay was performed, and the interaction between Treatment and Date all had significant effects on the growth of E. coli (Table 1A). The significant Treatment by Date interaction was due primarily to the strong effects of the ampicillin treatment and the relatively weak effects of the two types of nectar on the second Date (although both types of nectar had larger zones of inhibition than the glucose treatment). The analysis also showed that the area of each inhibition zone varied over Time and that Time had a significant effect on the growth of E. coli on each Treatment within each Plate on each Date the assay was performed (Table 1A). At 8, 12 and 24 hrs, the ampicillin treatment had cleared a significantly larger area of the E. coli lawn than did the glucose treatment or either of the two types of nectar (from plants lightly and heavily damaged by cucumber beetles) (Fig. 2). At 8 and 12 hrs, the nectar from lightly damaged plants had cleared a significantly larger area of the E. coli lawn than did the glucose treatment but by 24 hrs there were no significant differences in the

38 area cleared by nectar and glucose (Fig. 2). At no time point did the two types of nectar differ in the area they cleared on the lawns of E. coli. The repeated measures analysis of variance also revealed that Treatment and the interaction among Treatment and Date had significant effects, but that Date did not have a significant effect on the growth of E. tracheiphila (Table 1B). The Treatment by Date interaction was due primarily to the strong effects of both types of nectar on the first Date of the experiment (which exceeded the ampicillin treatment) and the complete lack of inhibition by the glucose treatment on the fourth Date of the experiment. The analysis also showed that the area of each inhibition zone varied over Time and that Time had a significant effect on the growth of E. tracheiphila on each Treatment within each Plate on each Date the assay was performed (Table 1B). At 8 and 12 hrs, the ampicillin and both types of nectar had cleared a significantly larger area of the E. tracheiphila lawn than did the glucose treatment (Fig. 2). There were, however, no significant differences in the area of the lawn cleared by the ampicillin or either of the two types of nectar. At 24 hrs, the ampicillin had cleared a larger area of the E. tracheiphila lawn than the glucose or either type of nectar. At 24 hrs, the nectar from the lightly damaged plants had cleared a significantly larger area of the E. tracheiphila lawn than did the glucose treatment while the nectar from the heavily damaged plants had cleared a nearly significantly larger area than did the glucose treatment (p = 0.06) (Fig. 2). At no time point, did the two types of nectar differ significantly in their effect on the growth of E. tracheiphila. In the greenhouse inoculation experiment, we found that the probability that a plant will contract wilt disease following inoculation of the nectary with E. tracheiphila is dependent upon the presence of nectar at the time of inoculation (χ2 = 15.4; df = 1; p < 0.001). Twenty-four of the 50 plants in which their male flowers were inoculated with E. tracheiphila after nectar removal contracted wilt disease. In contrast only 6 of the 50 plants in which their male flowers were inoculated without nectar removal contracted wilt disease. All ten of the plants that received stem inoculations (positive control for E. tracheiphila virulence) contracted wilt disease while none of the plants whose flowers

were inoculated with diH2O contracted wilt disease (negative control).

39 Discussion The floral nectar of many species contains secondary chemical compounds and/or proteins that are known to have antimicrobial properties (Martini et al., 1990; Adler, 2000: Thornburg et al., 2003; Nicholson and Thornburg, 2007; Junior et al., 2008). The nectars of other species are known to have antimicrobial properties but the specific chemicals responsible for these properties have not been investigated (Adler, 2000; Nicholson and Thornburg, 2007) In this study, we show that the nectar produced by wild gourd plants has a short term (12 hr) effect on the growth of E. coli in culture and a pronounced effect on the growth of E. tracheiphila in culture for 24 hrs. In fact, wild gourd nectar is as effective at inhibiting E. tracheiphila growth as 5% ampicillin for 12 hrs. Moreover, our greenhouse inoculation study showed that when the nectar is removed from flowers, the plants are 4 times more likely to contract wilt disease than plants whose nectar was not removed prior to inoculation. Although we exercised great care when removing the nectar from the flowers in the greenhouse, we cannot rule out the possibility that the process of nectar removal may, perhaps by damaging floral tissue, have facilitated infection by E. tracheiphila. It appears, however, that the nectar was able to inhibit E. tracheiphila infection even though the volume (100 ul) of the inoculum and the concentration of the E. tracheiphila in the inoculum (5.47x108 cells/ml) were far greater than what is likely to be encountered in nature. These findings suggest that the anti- E. tracheiphila properties of the nectar play a role in the transmission of wilt disease via the floral nectaries of Cucurbita species. This study focused on the antimicrobial properties of nectar from male flowers and transmission in the greenhouse via the inoculation of male flowers because male flowers attract more beetles per flower than female flowers under field conditions (Sasu et al., 2009), because male flowers are oriented vertically and the fecal pellets of the beetles accumulate at the base of the flower near the nectary whereas female flowers are oriented horizontally, and because wild gourd plants make approximately 7 times more male flowers than female flowers and, consequently, there are greater opportunities for exposure to E. tracheiphila via male flowers (Sasu et al., 2010). Previous greenhouse inoculation studies, however, have shown that E. tracheiphila can also be transmitted via

40 the female flowers (Sasu et al., 2010) and a pilot study that we conducted suggested that the nectar of female flowers also has antimicrobial properties (data not shown). Both male and females flowers of Cucurbita species open at sunrise and close in the late morning. Nectar secretion begins a few hours before anthesis and is produced continuously until the flowers close (Nepi et. al., 1996). In our fields, total nectar secretion by the late morning ranges from 10-100 μl in unvisited (bagged) flowers and seems to vary with soil moisture. However, the flowers are visited by pollinators repeatedly throughout the morning: first by squash bees, shortly after anthesis, which forage for both nectar and pollen, and then by squash bees, bumble bees and honey bees which forage for the newly secreted nectar throughout the remainder of the morning (Winsor et al., 2000). In the mid to late morning, cucumber beetles aggregate in the flowers to mate and feed. At the same field site during the same time of the growing season in which the flowers were collected for the disk diffusion assay, we have found, using real-time PCR with E. tracheiphila specific primers, that >50% of the flowers have cucumber beetle frass on or in the nectary (Sasu et al., 2010). Moreover, >90% of the flowers by 11:00 AM contain cucumber beetle frass contaminated with E. tracheiphila (Sasu et al., 2010). In the late morning, the flowers close and within 24-48 hrs a complete abscission layer is formed between the flower and the pedicel and the flower falls to the ground. In order for a plant to become infected, it is necessary for the E. tracheiphila to traverse the nectary and move into the xylem of the pedicel before the flower abscises. Infection by E. tracheiphila may be facilitated by repeated visitation by nectar foraging bees throughout the morning. In addition, Cucurbita flowers are known to resorb the nectar left in the flowers before abscission (Nepi et al., 1996) which may also facilitate bacterial transport through the nectary tissue. At our field site, a typical wild gourd plant produces 90-150 male flowers during July and August (Stephenson et al., 2004; Du et al., 2008; Sasu et al., 2009). Given the very high exposure rates of the flowers to E. tracheiphila that we have observed in our fields (> 90% with E. tracheiphila contaminated frass) (Sasu et al., 2010), it is surprising that only 5-40% of the plants per year contracted wilt disease over a 7 year period of time with 720 plants (4 fields with 180 plants) per year (Ferrari et al., 2007; Du et al., 2008;

41 Sasu et al., 2009). We suspect that the low incidence of wilt disease in our fields is due to the anti- E. tracheiphila properties of the nectar that sufficiently slows the growth of E. tracheiphila until floral abscission occurs. These findings support the hypothesis proposed by Carter and Thornburg (2004) that the antibiotic properties of nectar function to prevent pathogens from gaining access to the plants via the nectaries. The production of defensive chemicals (cucurbitacins) in the leaves and other organs of C. pepo (squash and wild gourds) are known to have both a constitutive component and a component that is inducible by cucumber beetle herbivory (Tallamy, 1985; Metcalf and Rhodes, 1990). Adler et al. (2006) found that nectar alkaloids increased with herbivory in Nicotiana tobacum. However, some floral nectars contain antimicrobial proteins, hydrogen peroxide and other compounds that may not be upregulated as a response to herbivory (Martini et al., 1990: Thornburg et al., 2003; Nicolson and Thornburg, 2007; Junior et al., 2008; Hancock et al., 2008). We found that the antibiotic properties of the wild gourd nectar did not differ significantly between plants with light and heavy damage by cucumber beetles. This finding suggests that the components of the nectar that are responsible for the antibiotic properties are either not inducible by cucumber beetle herbivory or that both light and heavy damage by cucumber beetles are sufficient to induce the antibiotic compounds. Historically, studies of the evolution of floral traits (such as their color, odor, size, shape, longevity, and quantity and composition of nectar) have focused on the roles of these traits in attracting and rewarding pollinators, disseminating pollen and fruit/seed production. However, there is a growing realization that floral traits can also evolve in response to selective pressures imposed by natural enemies (e.g., Stephenson, 1981; 1982; Galen, 1983; Strauss et al., 1996; Baldwin et al., 1997; Adler, 2000; Ashmann, 2002; Irwin et al., 2004). Although most of these studies focus on herbivores, recent studies have argued that the evolution of floral longevity can be influenced by exposure rates to floral transmitted pathogens (Shykoff et al., 1996; Kaltz and Shykoff, 2001; Valdivia et al., 2006). The data presented in this paper (showing that the antimicrobial nectar of a wild gourd reduces E. tracheiphila transmission via the floral nectaries) suggest the possibility that floral transmitted pathogens can also influence the evolution of nectar composition.

42 Acknowledgements We thank J. Rosenberger and D. Shumway for statistical advice, J Winsor for E. tracheiphila isolation, S. Scanlon for field, greenhouse and lab assistance, T. Omeis for use of the Biology Greenhouse, and R. Oberheim and his staff for use of the Horticulture Farm at the PSU Agriculture Experiment Station at Rock Springs, PA. This work was supported by NSF Grant DEB02-35217 and USDA Grant 2008-35302-04577.

Glucose Ampicillin control control

Light damage Heavy damage. nectar nectar

Figure 3.1. Schematic of a disk diffusion assay plate. Each 100 x 15 mm sterile culture plate was inoculated with either E. coli or E. tracheiphila and four 10 μl treatments were applied to the four 6 mm in diameter paper disks on each plate (5% ampicillin, 40% glucose, nectar from lightly damaged plants and nectar from heavily damaged plants). The area bacteria cleared by each paper disk was calculated at each of three times (8, 12, and 24 hours).

A. a

b b,c x x x c

B. a

b x x x b,c c y

C. a x

z y*z b b b y

E.coli E. tracheiphila

45 Fig. 3.2. Disk diffusion assay inhibition zones (cm2) on lawns of E. coli and E. tracheiphila. Graphs A, B and C show the least square means ± standard errors of the inhibition zones at 8, 12 and 24 hours for each of the four treatments. Least square means and standard errors were obtained from a general linear model analysis of variance including the following effects: date (4 dates), treatment (5% ampicillin, 40% glucose, nectar from lightly damaged plants, nectar from severely damaged plants), plate nested within date and the interaction of treatment by date. Lower case letters (a, b and c for E. coli and x, y and z for E. tracheiphila) show significance differences resulting from multiple pairwise comparisons performed at each time point (8, 12, and 24 hours) (Tukey-Kramer SAS Inst. 2002). y* represents a nearly significant (p>0.06) difference between the glucose and heavy damage treatments at 24 hours.

A. E. coli Effect Num Den F P DF DF Treatment 3 224 94.00 <0.0001 Date 3 224 3.62 0.0140 Date x 9 224 3.06 0.0018 Treatment Covariate Estimate SE Z P Time on 0.44 0.04 11.60 <0.0001 each Plate (Date x Treatment) B. E. tracheiphila Effect Num DF Den F P DF Treatment 3 224 20.30 <0.0001 Date 3 224 0.97 0.4057 Date x 9 224 2.97 0.0027 Treatment Covariate Estimate SE Z P Time on 0.41 0.04 9.98 <0.0001 each Plate(Date x Treatment)

Table 3.1. Results of a repeated measures analysis of variance for the effects of treatment, (5% ampicillin, 40% glucose, nectar from lightly damaged plants, and nectar from heavily damaged plants), date, and the interaction of date and treatment on area of inhibition (cm2) on sterile culture plate lawns of A) Escherichia coli and B) Erwinia tracheiphila. In this model time was a repeated measures variable; it was treated as a random variable, and it was specified for each treatment within each plate on each date experiment was performed. Autoregressive covariance structure was specified (Proc Mixed SAS Institute, 2002).

CHAPTER 4 Indirect Costs of a Non-Target Pathogen Mitigate the Direct Benefits of a Virus Resistant Transgene in Wild Cucurbita

Miruna A. Sasua, Matthew J. Ferraria, Daolin Dua,b, James A. Winsorc, Andrew G. Stephensona aDepartment of Biology, Center for Infectious Disease Dynamics, and Center for Chemical Ecology, 208 Mueller Lab, The Pennsylvania State University, University Park, PA 16802; a,bSchool of Environment, The Jiangsu University No.301, Xuefu Road, Zhenjiang, Jiangsu, 212013, China and cDepartment of Biology, The Pennsylvania State University, Altoona, PA 16601

Note: Published in Proceedings of the National Academy of Sciences 106 (45):19067- 19071. Accepted for publication September 16, 2009.

48 Abstract Virus resistant transgenic squash are grown throughout the USA and much of Mexico and it is likely that the virus resistant transgene (VRT) has been introduced to wild populations repeatedly. The evolutionary fate of any resistance gene in wild populations and its environmental impacts depend upon trade-offs between the costs and benefits of the resistance gene. In a 3-year field study using a wild gourd and transgenic and non-transgenic introgressives, we measured the effects of the transgene on fitness, on herbivory by cucumber beetles, on the incidence of mosaic viruses, and on the incidence of bacterial wilt disease (a fatal disease vectored by cucumber beetles). In each year, the first incidence of zucchini yellow mosaic virus occurred in mid-July and spread rapidly through the susceptible plants. We found that the transgenic plants had greater reproduction through both male and female function than the susceptible plants, indicating that the VRT has a direct fitness benefit for wild gourds under the conditions of our study. Moreover, the VRT had no effect on resistance to cucumber beetles or the incidence of wilt disease prior to the spread of the virus. However, as the virus spread through the fields, the cucumber beetles became increasingly concentrated upon the healthy (mostly transgenic) plants which increased exposure to and the incidence of wilt disease on the transgenic plants. This indirect cost of the VRT (mediated by a non-target herbivore and pathogen) mitigated the overall beneficial effect of the VRT on fitness.

49 Introduction Gene flow between crops and their wild relatives is common and difficult to contain (Ellstrand a, b, 2003) and spontaneous hybridization between transgenic cultivars and wild relatives occurs for 12 out of the world’s 13 most important crops (Ellstrand et al. 1999). Consequently, there are concerns that crop transgenes conferring resistance to herbivores or pathogens could escape and enhance the fitness and weediness of wild species, impact non-target species such as pollinators, herbivores, predators, or soil fauna, or alter the biodiversity within communities (Ellstrand a, b, 2003, Pilson and Prenderville 2004, Fuchs and Gonsalves 2007, NRC 2002). A variety of models suggest that the evolutionary fate of resistance genes (both natural and transgenic) in natural populations often depends upon trade-offs between the benefits and costs of the resistance gene in the presence and absence of the natural enemy (Mitchell and Bradley 1996, Boots and Bowers 1999, Kniskern and Rausher 2007). Fitness trade-offs can be either direct, (genetic resistance causes a reduction in one of the components of fitness), or indirect (pleiotropy results in ecological trade-offs that affect growth and development or resistance to other natural enemies (Bergelson and Purrington 1996, Heil and Baldwin 1997, Purrington 2000, Tian et al. 2003, Gassman and Futuyma 2005)). Direct trade-offs (e.g., in flower production) have been well-studied and characterized for many systems (Strauss et al. 2002, Leimu and Koricheva 2006). Ecological trade-offs, however, have been frequently hypothesized (Bergleson and Purrington 1996, Leimu and Kroicheva 2006, Puustinen et al. 2004) but less often demonstrated. In most cases these trade-offs and costs of resistance have been studied using relatively simple experimental settings in which plants are challenged with individual natural enemies. However, the predictions from pairwise challenges will not necessarily hold at the community scale when host plants are subject to attack from multiple enemies, and natural enemies must compete for hosts (Strauss et al. 2002, Leimu and Koricheva 1996, Burdon 1987, Irwin and Thresh 1990, Maddox and Root 1990). As such, it is unclear how ecological trade-offs affect the spread of resistance alleles at the scale of natural communities. For vectored pathogens, transmission depends on the exposure of the host to the pathogen, as mediated through the foraging behavior of the vector, and on the resistance

50 of the host to the pathogen. Unlike pathogen resistance, which can often be thought of as a fixed characteristic, conditional on the genotype or phenotype of the host individual, pathogen exposure is highly plastic. Vector foraging behavior, and the resultant patterns of pathogen exposure, can be highly dependent on the spatial arrangement and distribution of plant traits within the population (Ferrari et al. 2005, Ferrari et al. 2007) and the dynamics of the vector. Further, as vector foraging may be affected by disease or other natural enemies (e.g., Irwin and Thresh 1990, Maddox and Root 1990), the distribution of pathogen exposure can be dynamic in time. These complexities are difficult to capture within the framework of pairwise pathogen challenges that lack the full host-vector-pathogen community context. Gene flow from cultivated squash (Cucurbita pepo) to free-living taxa of Cucurbita is common and well-documented ( Kirckpatrick and Wilson 1988, Wilson et al. 1994, Decker-Walters et al. 1993). In 1996, the USDA deregulated a transgenic cultivar (CZW-3) with coat protein (CP)-based resistance to watermelon mosaic virus (WMV), zucchini yellow mosaic virus (ZYMV) and cucumber mosaic virus (CMV) (USDA 1996). In the transgenic cultivar a marker gene, neomycin phosphotransferase II (NPTII) was co-transferred. By the late 1990s, several cultivars with the transgene were developed, marketed, and grown throughout the USA (Fuchs and Gonsalves 2007). In 2004, the growth of transgenic squash was deregulated in several Mexican states (Senasica 2004). Because transgenic cultivars are often grown near wild populations, it is likely that the virus resistant transgene (VRT) has been introduced to natural populations repeatedly. The Cucurbita pathosystem that we study consists of the interactions among Cucurbita pepo ssp. texana (a wild C. pepo, texana gourd), its primary herbivores (cucumber beetles and aphids) and the bacterial and viral pathogens these insects transmit (Ferrari et al. 2007, Stephenson et al. 2004, Ferrari et al. 2006, Du et al. 2008). Cucumber beetles (Diabrotica spp. and Acalymma spp.) are specialist herbivores of the Cucurbitaceae. The leaves and other organs of the Cucurbitaceae produce bitter compounds called cucurbitacins (oxygenated tetracyclic triterpenes) that are toxic to most herbivores (Tallamy 1985). Cucumber beetles, however, are attracted to cucurbitacins in the foliage of Cucurbita (which they use both for defense and in mating) and to floral

51 volatiles that also attract bee pollinators ( Metcalf and Rhodes 1990, Anderson and Metcalf 1986, 1987, Lampman and Metcalf 1988, Metcalf and Lampman 1991). Cucumber beetles transmit Erwinia tracheiphila (Enterobacteriaceae), the causative agent of bacterial wilt disease, when fecal pellets containing Erwinia fall onto the open wounds as beetles feed on the leaves and when the fecal pellets fall in the vicinity of the nectaries when the beetles aggregate in the flowers to mate (Fleischer et al. 1999). Cucumber beetles tend to forage selectively on large plants with many flowers, resulting in biased exposure and disease mortality in larger plants (Ferrari et al. 2007). A variety of generalist aphids transmit CMV, WMV, and ZYMV. Wilt disease is always fatal once visible symptoms appear while the mosaic viruses slow growth and reproduction but generally do not kill the plant. In the Cucurbita pathosystem, the VRT should convey a fitness advantage in the face of a viral epidemic. However, when both pathogens are present, avoidance of the smaller, viral infected plants by cucumber beetles may make VRT plants relatively more attractive to the beetles, resulting in increased exposure to the lethal, non-target pathogen. We examined the fitness (flower and fruit production and proportion of progeny sired) of the VRT during introgression into the texana gourd and the effects of the VRT on herbivory by cucumber beetles, the incidence of the three mosaic viruses, and the incidence of wilt disease in a 3-year field scale study within the full host-vector-pathogen community. We created F1, Backcross (BC1), BC2, BC3 and BC4 by crossing a transgenic squash cultivar, Liberator III (hemizygous for the VRT with resistance to 3 viruses), to texana gourds with the texana gourd as the recurrent parent. In 2006, we transplanted 18 texana plants, 3 F1, 3 BC1, and 3 BC2 transgenic plants and 3 F1, 3 BC1, 3 BC2 non-transgenic siblings from each of 5 families (180 total plants, 25% were transgenic) into each of four 0.4 ha fields. In 2007 we planted two fields as above and two fields with texana, transgenic BC3 (25% of total plants) and non-transgenic BC3 from each of 5 families. In 2008 we planted two fields using the same design as the BC3 fields except that we planted texana and BC4 (with and without the VRT) from each of 5 families. We monitored flower and fruit production and the incidence of viral and wilt diseases throughout the growing season and recorded beetle damage to the leaves of new growth on 15 June, 15 July, and 15 August of each year (see Materials and Methods).

52 Our fields were located at the Penn State University Experimental Farms at Rock Springs, PA.

Materials and Methods Study System The texana gourd, Cucurbita pepo ssp. texana (Cucurbitaceae) is an annual monoecious vine with indeterminate growth and reproduction. It is native to Texas and states along the lower Mississippi River. It is completely inter-fertile with the cultivated pumpkins and squashes (C. pepo ssp. pepo and ssp. ovifera) and several annual Cucurbita taxa from Mexico (Decker-Walters 1993, Lira et al. 1995, Arriaga 2006). After a period of vegetative growth (5-7 nodes), texana gourds produce one large yellow flower (either male or female) in the axil of each leaf. The flowers last for only one morning and are pollinated by bees. The fruits of the wild gourd are round and typically contain 150-300 seeds (Avila-Sakar 2001). Cucumber beetles (Diabrotica spp. and Acalymma spp.) are specialist herbivores on the Cucurbitaceae and are the only known vector of Erwinia tracheiphila (Enterobacteriaceae), the causative agent of bacterial wilt disease, which overwinters in the guts of cucumber beetles. Wilt symptoms typically develop 10-15 days after infection and the disease is fatal once symptoms appear. However, fruits that are >10 days old when the first symptoms appear on the plant will mature (Du et al. 2008). WMV and ZYMV are the two most common viral diseases of cucurbits at our field site in central Pennsylvania. Both are single-stranded, positive-sense RNA viruses of the family Potyviridae and are transmitted via aphids. These diseases are rarely fatal but do depress reproductive output and produce symptoms that include leaf blisters, necrotic lesions, and branch deformities.

Field Experiments The Liberator III crookneck squash cultivar is a commercially available transgenic squash with coat protein (CP) based resistance to ZYMV, WMV, and CMV. In this cultivar, the virus resistant transgene (VRT) is hemizygous and, importantly, the NPTII gene conferring resistance to neomycin has not been deactivated and is still tightly

53 linked to the CP genes of the three viruses. Consequently, we have been able to introgress the transgene (CP genes and NPTII) into texana gourds (using texana gourd as the recurrent parent) because the presence of the transgene in hybrid progeny can be identified using DAS-ELISA (kit available from Agdia Inc. Elkhart IN) to detect the NPTII protein (half of the progeny from each cross are transgenic). No permit is required as this VRT has been deregulated for all subspecies and cultivars of C. pepo. Extensive seed germination and screening for NPTII in the F1 (1410 seeds), BC1 (1410 seeds), BC2 (1410 seeds), BC3 (940 seeds), BC4 (940 seeds) as part of the field studies described below revealed no transmission bias via pollen with respect to the VRT (6100 seeds; 50.3% transgenic). In 2006, 2007, and 2008, we germinated in a greenhouse (in mid May) and transplanted (late May) 18 texana plants, 9 transgenic introgressives and 9 non-transgenic (full siblings) introgressives from each of 5 families (180 total plants per field, 25% were transgenic) into each of four (2006-07) or two (2008) 0.4 ha fields as described in the Introduction. The fields were not sprayed with insecticide and viral and wilt diseases were allowed to occur naturally. The plants in each field during each year were monitored throughout the growing season for incidence of CMV, WMV and ZYMV (field diagnosis confirmed by DAS-ELISA tests; Agdia Inc., Elkhart, IN) and wilt disease (field diagnosis confirmed by isolating Erwinia from diseased plants and using the isolate to infect greenhouse grown plants (see 22 for techniques)). We counted male and female flowers weekly (an unbiased estimate of total flower production because flowers last for only one morning) and we counted total fruit production per plant at the end of the growing season. We non-destructively estimated beetle damage to leaves of new growth on 15 June, 15 July, 15 August of each year using a 0-5 index in which 0 = most leaves with no beetle damage and no leaf with more than 5% of the leaf area removed, and 5 = all leaves damaged and at least one leaf with > 50% of the leaf area removed. Three people, who were blind with respect to plant family and type of plant, simultaneously and independently evaluated damage on each plant. If two or three of the evaluators agreed on the score, we recorded that value. If the evaluators differed on their assessments, we recorded the middle score (~5% of the cases). In order to determine whether healthy and diseased plants differ in the number of beetles that aggregate in the flowers to mate, we counted the number of beetles in one

54 randomly chosen male and one female flower on all plants that produced at least one flower on 3 dates in August in each of 3 years (N = 3093 male flowers; N = 1486 female flowers). In order to determine whether the frequency of the VRT changed in the progeny generation, we harvested 2 mature fruits from each non-VRT introgressive and texana plant at the end of each growing season, removed and pooled their seeds and scored a random sample (475 seeds per field per year) for the presence of the VRT (NPTII protein) using DAS-ELISA.

Statistical Analysis. We used a mixed effects model analysis of variance (ANOVA) (SAS Institute 2008) to determine the effects of Plant Type (texana gourd, non-VRT introgressives, and VRT-introgressives), Year, Family (random) nested within Year, and Field nested within Year (a blocking variable that was dropped from the final analysis because it was not significant) on male flower and fruit production. To determine whether the cultivar genes affect the amount of beetle damage during introgression, we examined the effects of Year, Generation (texana, non-VRT F1, BC1, BC2, BC3, BC4) nested within Year, Family (random) nested within Year, and Field nested within Year on the amount of beetle damage prior to 15 July. For those plants that survived until the 15 July assessment of beetle damage, we used the mean of the 15 June and 15 July assessments. For those plants that died between 15 June and 15 July we used the 15 June assessment. Those plants that died prior to 15 June were excluded from the analysis. This analysis revealed that Generation nested within Year had no significant affect on beetle damage (Tables 4.2,4.3) (i.e., the texana and non-VRT introgressives did not differ in their resistance to cucumber beetles and were subsequently pooled in our analyses). To determine the effect of the VRT on resistance to cucumber beetles, we used a mixed effects model ANOVA to determine the effects of Plant Type (VRT-introgressives and non-VRT plants [texana and non-VRT introgressives]), Year, and Family (random) nested within Year on Beetle Damage prior to 15 July. We again used a mixed effects model ANOVA to determine the effects of Plant Type (VRT-introgressives and non-VRT plants), Year, and Family (random) nested within Year on Beetle Damage on 15 August. In this analysis, we used the 15 August assessment of Beetle Damage. Plants that died prior to 15 August were

55 treated as above. We then repeated the analysis of the 15 August Beetle Damage assessment but we included the Time with Virus (log days with virus + 1 from first visible symptoms until 1 September) as a covariate to determine the effect of viral disease on beetle herbivory. Finally, we performed a fixed effects model ANOVA to determine the effects of Disease Status (healthy or virus infected), Plant Type (VRT or non-VRT plants), Flower Type (male or female), Year, Family nested within Year, and Date nested within Year on the number of beetles per flower. We used a logistic regression to assess the effect of plant type (non-VRT or VRT), month, and year on the likelihood of infection with Erwinia. We assumed that those plants infected during 15 June to 15 July, and 15 July to 15 August were conditionally independent. Thus, we analyzed the plants that were infected during 15 July to 15 August as coming from the pool of plants that remained alive and uninfected on July 15. We included an interaction between plant type and month in the model to account for changing odds of infection throughout the season. In order to assess the direct and indirect effects of Plant Type (VRT-introgressives and non-VRT plants), Year, Beetle Damage, Viral and Wilt diseases on Reproductive Output, we used structural equation modeling/path analysis (Gassman and Futuyma 2005, Scheiner et al. 2000, Shipley 2000, Pigliucci and Kolodynska 2006). (Reproductive Output = Fruit number per plant + (male flowers per plant ÷ 7) because across all plants there were 7X more male flowers than fruits; see Fig 4.2), With path analysis it is possible to examine the effect of the VRT on each character in the model while holding constant all other factors that have paths leading to that character (Sokal and Rohlf 1995). We performed the path analysis using AMOS 16 for Windows (SPSS AMOS 1999). We coded the VRT introgressives as 1 and non-VRT plants as a 0 so that paths proceeding from Plant Type reveal the relationship with the VRT plants. The initial models included Family nested within Year but this term was dropped from the final models because no significant paths proceeded from it. The final models were assessed by: (a) goodness of fit (χ2), (b) root mean square error approximation (RMSEA), and (c) Akaike’s Information criterion (AIC) calculated for the final model tested and a saturated model (Pigliucci and Kolodynska 2006).

Results In all years, the VRT effectively deterred viral infections in the introgressives (Fig. 4.1 ). In 2006 and 2008, the first symptoms of viral disease (ZYMV) occurred in mid July and spread rapidly through the fields so that by September most of the remaining texana and non-VRT introgressives (those that had not died of wilt disease) had become infected with ZYMV (Fig. 4.1). In 2007, the first incidence of a viral disease (WMV) occurred in mid-June and spread slowly until mid July when we recorded the first incidence of ZYMV which spread rapidly through the remaining non-VRT plants (texana and introgressives). In each year, a few VRT plants became infected with ZYMV after 15 August but the symptoms were very mild (Fig. 4.1). There was no significant difference in the proportion of texana and non-VRT introgressives that contracted a viral disease over the 3 years of this study (χ2 = 0.36; df = 1; p > 0.10). In each year, flowering commenced in late June and peak flower and fruit production occurred from late July to late August (when ZYMV was spreading through the fields). The VRT introgressives produced significantly more flowers and mature fruits than did the virus susceptible types and there were no significant differences in flower number and fruit number between non-VRT introgressives and texana gourds (Fig 4.2). If fertilization is random with respect to the presence of the VRT in the populations that we planted, we expected 12.5% of the seeds on the non-VRT introgressives and texana gourds to have the VRT (25% of the plants in each field were hemizygous for the VRT). We found that the VRT plants sired a higher proportion of the seeds on the virus susceptible plants than is expected by chance alone (27% in 2006; 26.9% in 2007; 29.4% in 2008; all χ2 > 175; df =1; all p <0.0001). We found that the non-VRT plants and the VRT plants did not differ in the amount of beetle damage they experienced prior to 15 July (before viral diseases were prevalent in our fields) (F1,1821 = 0.56; p = 0.45) but that the VRT plants experienced significantly greater beetle damage during August (when viral diseases were prevalent in the non-VRT plants) (Fig. 4.3A). However, if the time with virus is included as a covariate in the model, there is no significant difference between the amount of damage

57 sustained by the VRT and non-VRT plants in August (Fig. 4.3B) suggesting that the beetles prefer to feed on healthy plants, rather than VRT plants per se. Moreover, we found that significantly more beetles aggregated in the flowers of healthy plants (mostly VRT introgressives) than in the flowers of viral infected plants during August (LSMean ±

SE; Healthy = 3.1 ± 0.1; Viral infected = 2.6 ± 0.2; F1,4553 = 8.6; p < 0.005). In each of the 3 years, the first symptoms of bacterial wilt disease appeared in our fields within 2 weeks after transplanting. There were, however, no differences in the incidence of wilt disease on VRT (5.6%) and non-VRT (5.8%) plants prior to mid-July indicating that the VRT does not alter the resistance of plants to the non-target pathogen, Erwinia. However, we found that the incidence of wilt disease on the transgenic plants from mid July until 1 September was significantly higher (17.5%) than the incidence of wilt disease on the non-transgenic plants (10.9%) over the 3 years of this study (χ2 = 10.9; df = 1; p < 0.001). Moreover, a logistic regression model revealed a significant positive interaction between month and the presence of the transgene (p=0.02) indicating that the odds of developing wilt disease increased for transgenic plants after mid-July in all three years (Fig. 4.4). In 3 years, only 2 viral infected plants (0.4% of the viral infected plants) contracted wilt disease and there was no significant difference in the percentage of texana and non-VRT introgressives that contracted wilt disease (χ2 = 2.25; df = 1; p > 0.10). To determine both the direct and indirect effects of the VRT on fitness, we performed a series of path analyses. In these analyses, non-VRT plants are coded as 0 and VRT plants as 1: thus, a path proceeding from Plant Type reveals the relationship to the VRT plants. Because wilt infected plants die and cannot later contract a viral disease, we are unable to include time with wilt disease and time with viral disease in the same model. The first analysis, using only those plants that survived to the end of August during 2006-2008 (i.e., no plants that contracted wilt disease), revealed that the VRT has no direct effect on reproductive output (see Materials and Methods for calculation) or beetle damage (Fig. 4.5A). Rather, the VRT has an overall positive effect on reproductive output via its strong negative relationship with time with viral disease which, in turn, has a negative effect on reproductive output. The second analysis, using all plants but ignoring time with viral disease, reveals that the VRT has a positive and direct effect on reproductive output, beetle damage, and time with wilt disease (Fig. 4.5B). From Fig.

58 4.5A, we know that the positive relationship between the VRT and reproductive output is due to viral resistance and the negative effects of viral diseases on reproductive output. However, the overall positive effect of the VRT on reproductive output is reduced because the time with wilt disease is negatively related to reproductive output. Moreover, beetle damage has a positive effect on the duration of infection with wilt disease (beetles transmit wilt disease while feeding on leaves), which, in turn, is negatively related to reproductive output (Table 4.1).

Discussion In this 3-year study, viral diseases (mostly ZYMV) colonized and spread rapidly through our fields during the period of peak gourd reproduction. The VRT plants (25% of the population) produced 30.6% of the fruits and 38% of the male flowers (but only half of the pollen grains in these flowers have the VRT). These data clearly demonstrate that the squash VRT would have a fitness advantage through both the male and female functions during the initial stages of introgression into a population of wild texana gourds under the conditions of our study. Moreover, our data show that the VRT plants sired even more of the progeny than would be predicted based on male flower production. Assuming all male flowers in the population produce the same number of pollen grains and that pollen removal and deposition are random with respect to the VRT, we would expect that 19% of the seeds on non-VRT plants would contain the transgene. However, 26-29% of the seeds were sired by pollen with the transgene, suggesting that the viral free VRT plants attract more pollinators, produce and disseminate more pollen per flower, and/or the pollen from VRT plants are competitively superior to pollen from the non- VRT plants. We have also shown that the fitness advantage of the VRT plants comes with a cost due to the interactions with the herbivore and pathogen community. This study revealed that the VRT did not affect resistance to cucumber beetles or bacterial wilt disease prior to the establishment and spread of viral pathogens in the susceptible plants. Infection with a viral pathogen, however, reduced the attractiveness of the non-VRT plants as both a food source and as a mating location for the cucumber beetles. The VRT plants experienced higher levels of leaf herbivory in August than the non-VRT plants;

59 viral infection reduced flower production on non-VRT plants compared to VRT plants; and the VRT plants attracted more beetles per flower. In short, the presence of viral pathogens tended to concentrate the cucumber beetles onto healthy (mostly VRT) plants resulting in increased Erwinia exposure on VRT plants through both modes of transmission (foliar feeding and floral transmission). As viral infection spread through the susceptible plants in the population, the incidence of wilt disease increased on the VRT plants. Our path analysis indicates that the VRT has an indirect effect on the feeding preference of a non-target herbivore (cucumber beetles) and the pathogen it transmits (Erwinia) that mitigates the overall fitness of the transgene during introgression. Thus, there is an indirect, ecological cost associated with the VRT when the non-target pathogen and its vector are also present in the population. Texana gourds have long been recognized as an important weed in soybean and cotton fields (Oliver et al. 1983). Whether introgression of the VRT into wild populations of Cucurbita will result in a more problematic weed or if it poses a threat to natural biodiversity depends upon the fitness of the VRT introgressives. Our data indicate that the VRT could increase in frequency in wild Cucurbita. However, the indirect costs due to Erwinia exposure may mitigate the reproductive benefits of the VRT. The fitness of the VRT may depend upon (a) the arrival times and transmission rate of both the target and the non-target diseases in the population, (b) the number of cucumber beetles in the population, and (c) the proportion of plants in the population with the VRT (which will determine the increase in beetle concentration onto VRT plants as non-VRT plants become infected with virus and perhaps also affect the transmission rate of the virus (Garrett and Mundt 1999, Klas et al. 2006)). A large body of theory (but limited empirical data) suggests that ecological trade- offs between the costs and benefits of a resistance gene, such as the Cucurbita VRT, play important roles in determining the gene’s effects on fitness (Mitchell-Olds and Bradley 1996, Boots and Bowers 1999, Bergleson and Purrington 1996, Heil and Baldwin 1997, Purrington 2000, Tian et al. 2003, Gassman and Futuyma 2005, Strauss 2002). Most studies of the fitness of herbivore or pathogen resistant transgenes in crop-wild plant hybrids (and their backcross progeny) have tended to examine direct trade-offs in pairwise comparisons (e.g., with and without the target herbivore/pathogen) that lack the

60 full host-natural enemy-community complex. Typically, these studies have found that seed production of hybrids carrying resistance to herbivores or pathogens increases in the presence of their target herbivores or pathogens (Burke and Rieseberg 2003, Snow et al. 2003, Fuchs et al. 2004, Chen et al. 2006). However, upon escape into natural populations, resistance transgenes face a complex suite of direct and indirect costs. As we have shown here, a full understanding of the combined effect of these forces on the fitness of an escaped transgene may not be apparent without the context of the complete ecological community.

Acknowledgements

We thank A. Guirand, J. Verbano, R. Kariyat, K. Wall, M.Bothe, C. Dryoff, S. Scanlon and G. Stephenson for field, greenhouse and lab assistance; T. Omeis for the use of the Biology greenhouses and R. Oberheim for use of the Horticulture Farm at The PSU Agriculture Experiment Station at Rock Springs, PA. This research was supported by NSF grant DEB02-35217 to AGS and JAW.

Figure 4.2. Number of male flowers and fruits produced per plant on texana gourds, Non- VRT and VRT introgressives. LSmeans ± SE bars. Model: Male Flowers (or Fruits) per

plant = Year+Family(Year)+Plant type [texana, Non-VRT, VRT]. Plant type F2,2155 =

12.15 for Male Flowers and F2,2155 = 15.23 for fruits; both p < 0.0001.

Figure 4.3. Amount of beetle damage on leaves of Non-VRT (texana and non-VRT introgressives) and VRT plants. LSmeans ± SE bars. Model: Beetle damage = Year+Family(Year)+Plant type[VRT vs. non-VRT]. A. Beetle damage on 15 August.

Plant type F1,1821 = 5.63; p = 0.017. B. Beetle damage on 15 August with time with viral disease in the model. Plant type F1,1820 = 0.05; p = 0.83.

Figure 4.4. Proportion of susceptible plants (alive and not infected with virus) that became infected with Erwinia in the interval 15 June-15 July (July) and 15 July-15 August (August). Error bars indicate ± SE. Asterisks indicate significant differences within month (p<0.05).

Figure 4.5. Final path analysis including the effects of A. time with virus and B. time with wilt disease on reproductive output. Values of standardized regression coefficients are given by line weight and dotted vs solid lines (see inset legend). Final models include only significant paths. Significance levels of correlations are denoted by **p<0.01; ***p<0.001. (A) χ2 = 7.0, df = 3, p = 0.07, RMSEA = 0.03, AIC = 41.08. (B) χ2 = 0.03, df=1, p = 0.85, RMSEA = 0, AIC = 38.03.

Effects Pathway Magnitude Model including time with virus 1 p(plant type, days with virus) x p(days 0.0492 with virus, reproductive output ) Model including time with wilt 1 p(plant type, beetle damage) x p(beetle -0.0016 damage, time with wilt ) x p(time with wilt, reproductive output ) 2 p(plant type, time with wilt) x p(time -0.0059 with wilt, reproductive output ) 3 p(plant type, reproductive output) 0.1060 Total effects in model 0.0985 including time with wilt (1+2+3)

Table 4.1. Magnitude of all significant paths from plant type to reproductive output in the two path models.

Effect Num DF Den DF F p Year 2 1166 84.75 <0.0001 Field(Year) 7 1166 20.75 <0.0001 Generation(Year) 8 1166 0.90 0.52 Family(Year) 12 1166 1.17 0.29

Table 4.2. Effects of Year, Field nested within Year, Generation (texana, and non- transgenic F1, BC1, BC2, BC3, BC4) nested within Year, and Family (random) nested with Year on beetle damage during June and July of 2006-2008.

Generation Year LSMeans ± SE

texana 2006 1.50±0.04 F1 Non-VRT 2006 1.45±0.09 BC1 Non-VRT 2006 1.46±0.10 BC2 Non-VRT 2006 1.33±0.09 texana 2007 0.77±0.04 F1 Non-VRT 2007 0.81±0.13 BC1 Non-VRT 2007 0.83±0.13 BC2 Non-VRT 2007 1.05±0.14 BC3 Non-VRT 2007 0.70±0.08 texana 2008 0.70±0.07 BC4 Non-VRT 2008 0.70±0.05

CHAPTER 5 The Costs and Benefits of a Virus Resistance Transgene During Introgression into a Wild Gourd

Miruna A. Sasu, Matthew J. Ferrari and Andrew G. Stephenson1 Department of Biology and Center for Infectious Disease Dynamics, The Pennsylvania State University, 208 Mueller Lab, University Park, PA 16802

Abstract Virus resistant transgenic squash are grown throughout the United States and much of Mexico where they are inter-fertile with several co-occurring wild taxa of Cucurbita. The virus resistance transgene (VRT) is likely to escape into wild populations. The environmental impacts of the VRT will depend upon its fitness during introgression into the wild populations. In a large field study, we examined the effects of the VRT on plant fitness, herbivory by the primary herbivores (cucumber beetles), and the incidence of a non-target pathogen in the presence and absence of the viral pathogens during introgression into wild Cucurbita pepo. We found that in the presence of the virus, the introgressive plants with the VRT had greater reproductive output, more herbivory by cucumber beetles, and a higher incidence of bacterial wilt disease (vectored by the beetles). In the absence of the virus, plants with the VRT also had greater reproductive output (but the increase was not as great as in the virus fields) but they did not have greater herbivory or a higher incidence of wilt disease. These findings suggest that the fitness of pathogen resistance transgenes in crop-wild plant hybrids can only be determined within the context of the full pathosystem.

70 Introduction Foliar herbivory is an ubiquitous component of terrestrial communities (McNaughton et al. 1989; Louda et al. 1990; Marquis 1992). For individual plants, leaves function as the main photosynthetic organs and are also essential for the acquisition of mineral nutrients. Leaf damage induces chemical and physical defenses that can also be costly in terms of energy and nutrients (see reviews by Bergelson and Purrington 1996; Bazzaz 1997; Karban and Baldwin 1997; Strauss et al. 2002). Therefore, it is not surprising that foliar herbivory has been shown to adversely affect vegetative growth, survival, and reproduction through both the female (fruit and seed) and male (pollen production and pollen performance) functions (Marquis 1992; Coley and Barone 1996; Delph et al. 1997; Mutikainen and Delph 1996; Strauss et al. 2001). Similarly, pathogens are also ubiquitous in natural systems (e.g., MacClement and Richards 1956; Burdon 1987a; Raybould et al. 1999) and they can also reduce the pool of resources within individual plants by decreasing leaf area, disrupting cellular and transport processes, and by inducing various biochemical defense systems that are costly in terms of energy and nutrients (see Burdon 1987a; Conner et al. 1996; Ryals et al. 1996). In addition, pathogens are also known to reduce growth, survival, and reproductive output in both cultivated and wild species (e.g. Burdon 1987ab; Burdon and Leather 1990, Maskell et al. 1999). It is also important to note that herbivores often facilitate infection by pathogens by serving as vectors for the transmission of pathogens, weakening plant defense systems, and by opening wounds where pathogens can enter a plant. Consequently, host plants, pathogens, and insect herbivores/vectors may best be viewed as interdependent components of a complex pathosystem that determine the transmission dynamics of diseases within populations (Irwin and Thresh 1990). Because of the impact of herbivores and pathogens on plant fitness, genes conferring resistance are expected to evolve in natural populations. Theory suggests that the fate of a resistance gene in populations (loss, polymorphism, fixation) depends upon the fitness benefit of the resistance gene when the target enemy (herbivore or pathogen) is present, the fitness cost of the resistance gene in the absence of the target enemy, and the frequency/abundance of the target enemy within the population (Mitchell-Olds and Bradley 1996; Boots and Bowers 1999; Kniskern and Rausher 2007). The cost of

71 resistance can be direct and indirect. Direct costs include costs to survival and reproductive output such as the costs of up-regulation of defense pathways, pleiotropic effects of the resistance gene on growth and development, and reduced resource accumulation. Indirect (ecological) costs include decreased competitive ability, decreased resistance to other natural enemies, or increased exposure to pathogens (e.g., Bergelson and Purrington 1996; Heil and Baldwin 1997; Purrington 2000; Tian et al. 2003; Gassmann and Futuyma 2005). Direct costs of resistance genes have been studied and characterized for several systems in the presence/absence of the target enemy (E.g., Strauss et al. 2002; Leimu and Koricheva 2006). Although frequently hypothesized (Bergelson and Purrington 1996; Puustinen et al. 2004; Leimu and Koricheva 2006), ecological costs are rarely demonstrated because it is necessary to know both the genetic basis of resistance and the interactions among the individual components (target enemy and non-target herbivores and pathogens) that comprise the host plant pathosystem. Transgenic squash (Cucurbita pepo ssp. ovifera) with coat protein based resistance to three of their most common viruses are grown throughout the United States and much of Mexico (USDA 1996; Feber 1999). These squash are completely inter- fertile with several co-occurring annual taxa of wild Cucurbita in both the United States and Mexico and gene flow between cultivated and wild Cucurbita is both common and well-documented (Kirkpatrick and Wilson 1988; Decker-Walters et al. 1993; Wilson et al. 1994; Arriaga et al. 2006). Therefore, it is likely that the transgene has been introduced into wild populations repeatedly. Compared to wild gourds, cultivated squash typically have large seeds, short internodes and large fruits with a soft pericarp with sweet (not bitter) flesh. Normally, such agricultural traits are at a selective disadvantage in the wild and tend to be rapidly purged from wild populations (Decker-Walters et al. 1993). However, there are concerns that the escaped virus resistance transgene will enhance the fitness of wild plants during introgression, increase the weediness of wild populations, impact non-target species such as pollinators, herbivores, predators, or soil fauna, or pose a threat to the biodiversity of communities (NRC 2002; Pilson and Prendeville 2004; Fuchs and Gonsalves 2007). Consequently, it is imperative to examine the direct and indirect effects of the transgene on fitness as it introgresses into wild Cucurbita. The introgression of the transgene into wild populations also presents an

72 opportunity to examine the evolutionary fate of a resistance gene within the context of the Cucurbita pathosystem. Two studies have examined the direct effects of the squash virus resistant transgene on fitness during the initial stages of introgression into the Texas gourd, Cucurbita pepo ssp. texana (Fuchs et al. 2004; Laughlin et al. 2009). These studies revealed that the introgressed transgene effectively deters infection by the target viruses; that the plants with the transgene had a tremendous fitness advantage over the susceptible plants in experimental populations in which the susceptible plants were infected with a target virus as seedlings; and that there were no direct costs of the transgene in the relative absence of the target viruses (isolated or insecticide sprayed fields). Recently, we (Sasu et al. 2009) examined both the direct effects of the transgene during introgression into the Texas gourd on plant fitness and the indirect costs of the transgene mediated by cucumber beetle herbivory and the bacterial (non-target) pathogen that it transmits. In this 3 year study, the target viruses were allowed to naturally colonize and spread through the fields over the course of the growing season and there were no attempts to alter the composition of the host plant pathosystem (via isolation, insecticide, or removal of plants infected with non-target diseases). This study also found that the introgessed transgene effectively deters the target viruses and increases fitness in the presence of the target viruses, but the fitness benefit of the transgene was partially offset by increased cucumber beetle herbivory on the transgenic plants and an increased incidence of the deadly bacterial disease vectored by the cucumber beetles on the transgenic introgressives. We speculated that the preference of the cucumber beetles for the transgenic plants was due to a preference of the beetles for healthy (not viral infected) plants rather than transgenic plants. The large scale field study presented here measures the direct and ecological effects of the virus resistant transgene on plant fitness, herbivory by the primary herbivores, and the incidence of a non-target pathogen (bacterial wilt disease) in the presence and absence of the viral pathogens after extensive introgression into wild Cucurbita pepo ssp. texana. To prevent infection by the aphid transmitted target viruses while maintaining the other components of the Cucurbita pathosystem, we used an aphid specific pesticide that is ineffective against chewing insects and compared the fitness of

73 Texas gourds, non-transgenic introgressives, and transgenic introgressives in sprayed (no virus) and unsprayed (natural virus infections) fields. Specifically, we ask: 1) What is the fitness advantage of the transgene when one of the target viral diseases naturally colonizes and spreads through the population? 2) What are the costs of the transgene in the absence of the target pathogens? 3) Are there indirect (ecological) costs associated with the transgene in either the presence or the absence of the target pathogens? 4) Does the transgene affect the foraging behavior of the primary herbivores and the incidence of the non-target pathogen that they vector? This large scale field study with open pollination is possible because the growth of transgenic C. pepo with coat protein based viral resistance has been federally deregulated for all cultivars and subspecies of C. pepo (Accord 1996; USDA 1996). To avoid the risk of escape transgene to wild populations, we performed the study in Pennsylvania which is outside the range of the natural populations. However, because cultivated squash are grown throughout the Eastern US, the other components (herbivores and non-target pathogens) are present in Pennsylvania.

Materials and Methods The Study System The Texas gourd, Cucurbita pepo ssp. texana (Cucurbitaceae) is an annual monoecious vine with indeterminate growth and reproduction. It is native to Texas, and to states along the Mississippi River from Southern Illinois southward (sometimes called ssp. ozarkana). It is completely inter-fertile with the cultivated pumpkins and squashes (C. pepo ssp. pepo and ssp. ovifera) and several annual Cucurbita taxa from Mexico (C. pepo ssp. fraterna, C. moschata, C. ficifolia, C. agyrosperma ssp sororia and ssp agyrosperma) (Arriaga et al. 2006; Decker-Walters et al. 1993). After a period of vegetative growth (5-7 nodes), wild gourds produce one large yellow flower (either male or female) in the axil of each leaf. The flowers last for only one morning and are pollinated by bees, especially squash bees. The fruits of the wild gourd are round to oval with a volume of 175-450 ml and typically contain 150-300 seeds that weigh 20-40 mg (Winsor et al. 2000; Avila-Sakar et al. 2001). The leaves and other organs of the wild gourd produce cucurbitacins (oxygenated tetracyclic triterpenes). Cucurbitacins are among the most bitter compounds known, and

74 are toxic to most herbivores (Metcalf and Rhodes 1990; Tallamy 1985). However, cucumber beetles (Diabrotica spp. and Acalymma spp.) are adapted to feed on the leaves of Cucurbita species and are found throughout the native ranges of wild Cucurbita species (Robinson and Decker-Walters 1997). The cucumber beetles chemically modify the cucurbitacins, use them for their own protection, and the males also transfer some of the modified cucurbitacins to the females in their seminal fluid, which is used to chemically protect the eggs (Ferguson et al. 1985; Nishida and Fukami 1990; Tallamy and Krischik 1989; Tallamy and Brown 1999). Cucumber beetles are attracted to cucurbitacins in the foliage of Cucurbita and, when flowers are present, it has been shown that floral volatiles not only attract bees (the pollinators) but also cucumber beetles over relatively large distances (Anderson and Metcalf 1986;1987; Metcalf and Lampman 1991; Lampman and Metcalf 1988). These beetles cause a characteristic pattern of holes (1-1.5 cm in diameter) in the portions of the leaves serviced by the smallest veins and beetle damage has been shown to substantially reduce yield in cultivated cucurbits (e.g., Tallamy and Krischik 1989) and reproductive output in the wild gourd (Quesada et al. 1995; Stephenson et al. 2004; Du et al. 2008). After feeding on the leaves, the beetles aggregate in the flowers to feed and mate. The full impact of herbivory by cucumber beetles on wild gourds, however, also includes increased exposure to a pathogen, Erwinia tracheiphila (Enterobacteriaceae)— the causative agent of bacterial wilt disease. Cucumber beetles are the only known vector of Erwinia, which is known to overwinter in the alimentary canal of cucumber beetles (Garcia-Salazar et al. 2000). Transmission occurs when fecal pellets containing Erwinia contact leaf wounds at the sites of feeding damage (Fleischer et al. 1999). Recently, we have shown that transmission also occurs via the floral nectaries when the beetles aggregate in the flowers to mate and their Erwinia contaminated fecal pellets fall onto the floral nectaries (Sasu et al. in press). After entering the plants via the leaf wounds or the floral nectaries, the bacteria proliferate in the xylem where they secrete an exopolysaccharide (mucilaginous) matrix that cuts off the water supply resulting in wilting. Wilt symptoms typically develop 10-15 days after infection and the disease is nearly always fatal once symptoms appear (Yao et al. 1996). Bacterial wilt disease is the most economically important disease of cultivated cucurbits (cucumbers, melons,

75 pumpkins and squash) in the Eastern US (Fleischer et al. 1999) and is typically controlled using insecticides that target cucumber beetles. Several generalist aphids are known to feed on Cucurbita. These aphids also vector the four most common viral diseases of cucurbits (Cucumber mosaic [CMV], Papaya ringspot [PRSV], Watermelon mosaic [WMV], and Zucchini yellow mosaic [ZYMV]). ZYMV and WMV are the two most common viral diseases of cucurbits at our field sites in Central Pennsylvania (Sasu et al. 2009). Both are single-stranded, positive- sense RNA viruses of the family Potyviridae and are transmitted via aphids in a non- persistent manner. These viral diseases produce symptoms that include blisters, necrotic lesions, branches with short internodes and small highly serrate leaves, and other leaf and fruit deformities. In 1994, the USDA deregulated a transgenic yellow crookneck squash (Asgrow, ZW-20) that was engineered to express a dual coat protein (CP) construct of WMV and ZYMV that conferred resistance to those viruses (USDA 1994; Tricoli et al. 1995; Fuchs and Gonsalves 1995). In 1996, a second transgenic cultivar (CZW-3) with CP-based resistance to WMV, ZYMV, and CMV was deregulated (USDA 1996). In both transgenic cultivars a marker gene, neomycin phosphotransferase II (NPTII) was co- transferred. By the late 1990s, many cultivars with the transgene were developed, marketed, and grown commercially throughout the United States but especially in the Southeastern US where it greatly reduced pesticide use and increased yield (Feber 1999; Fuchs and Gonsalves 2007). In 2004, the Mexican Ministry of Agriculture also deregulated the growth of transgenic squash in several Mexican States.

Introgression of the Virus Resistant Transgene and Field Experimental Design The Liberator III crookneck squash cultivar is one of many commercially available transgenic squash derived from CZW3 with coat protein (CP) based resistance to ZYMV, WMV, and CMV. In this cultivar, the virus resistant transgene (VRT) is hemizygous and, importantly, the NPTII gene conferring resistance to neomycin has not been deactivated and is still tightly linked to the CP genes of the three viruses. Consequently, we have been able to introgress the transgene (CP genes and NPTII) into 20 families of Texas gourds (using the wild gourd as the recurrent parent) because the

76 presence of the NPTII protein in hybrid progeny can be detected serologically (DAS- ELISA kits obtained from Agdia Inc., Elkhart IN)--half of the progeny from each cross are transgenic. Extensive seed germination and screening for NPTII studies of F1 (1410 seeds), backcross 1 (BC1: 1410 seeds), BC2 (1410 seeds), BC3 (940 seeds) as part of an earlier study (Sasu et al. 2009), and BC4 (940 seeds) for this study (see below) revealed no transmission bias via pollen with respect to the transgene (6100 seeds; 50.3% transgenic). In order to determine the effectiveness of the introgressed VRT at deterring viral infections and to determine the impacts of the transgene on plant fitness, herbivory by the major herbivores, and the incidence of a non-target pathogen in the presence and absence of the viral pathogens, we germinated seeds in a greenhouse and, in late May, transplanted 18 Texas gourd plants, 9 BC4 transgenic plants (tBC4) and 9 BC4 non- transgenic plants (ntBC4) from each of 5 families (180 total plants per field, 25% were transgenic) into each of four 0.4ha fields at the Horticulture Farm of The Pennsylvania State University Agriculture Experiment Station at Rock Springs, PA. Two of the fields (located 100m apart) were sprayed every 3 weeks (late June, mid July, early August) with Pymetrozine (Endeavor 50 WDG®), a selective biocide against aphids, whiteflies and plant hoppers (sucking insects). However, this biocide is ineffective against chewing insects such as cucumber beetles (e.g. Harrewijn and Kayser 1997; Wyss and Bolsinger 1997). The other two fields were located 1.5 km downwind (of the prevailing direction of the wind) and not sprayed with insecticide. These fields were separated by 20m of grass and permitted to become naturally infected with mosaic viruses. In all 4 fields, the plants were the progeny from the same five randomly selected families (the Texas gourd progeny and the ntBC4 and tBC4 shared the same maternal parents). The tBC4 and ntBC4 were full siblings as were all of the Texas gourd plants from a given maternal parent. The plants in each of the 4 fields were monitored throughout the growing season for incidence of CMV, WMV and ZYMV (field diagnosis confirmed by DAS-ELISA tests; Agdia Inc. Elkhart, IN) and wilt disease (field diagnosis confirmed by isolating Erwinia from diseased plants and using the isolate to infect greenhouse grown plants (see Ferrari et al. 2007 for techniques)). We non-destructively estimated beetle damage to the

77 leaves of new growth on 15 June, 15 July, 15 August using a 0-5 index in which 0 = most leaves with no beetle damage and no leaf with more than 5% of the leaf area removed, and 5 = all leaves damaged and at least one leaf with > 50% of the leaf area removed. Three people, who were blind with respect to plant family and type of plant, simultaneously and independently evaluated damage on each plant. If two or three of the evaluators agreed on the score, we recorded that value. If the evaluators differed on their assessments, we recorded the middle score (only a few cases). We assessed reproductive output by counting male and female flowers weekly (an unbiased estimate of total flower production because the flowers last for only one morning) until the end of August and by counting total mature fruit production after the first frost.

Statistical Analyses Because the fields were not randomly assigned to treatment (Pymetrozine spray, no spray) due to our desire to isolate the treatments from cross contamination of aphids and the viral diseases they vector, we performed separate analyses for each treatment. We used a mixed effect model analysis of variance (ANOVA; Proc. Mixed, SAS Inst. 2007) to determine the effects of Plant Type (Texas gourd, ntBC4, tBC4), Family (random), Field, and the 2 way interaction of Family and Plant Type on Reproductive Output in the non-sprayed fields. The reproductive output of each plant is the number of fruits produced + [the number of male flowers on that plant X (the mean number of fruits produced on all plants in that field divided by the Mean number of male flowers produced on all plants in that field). Thus the male contribution to reproductive output equals the female contribution for the plants in each field. Differences in the reproductive output from the three Plant Types were compared using Tukey pairwise comparisons controlling for the overall probability of Type 1 error (alpha = 0.05). To determine the cost of the transgene in the absence of the target viruses on Reproductive Output, we performed an identical analysis of variance for the plants in the two fields sprayed with Pymetrozine. We also used a mixed effects ANOVA (Proc. Mixed) to determine the effects of Plant Type, Family (random), Field and the 2 way interaction Plant Type and Family on Mean Beetle Damage per plant during June and July in non-sprayed and (separately) on

78 the sprayed fields. Mean Beetle Damage is the average of our two estimates on the new growth (15 June, 15 July). If a plant died before the 15 June estimate of damage, the plant was omitted from the analysis. If a plant died between 15 June and 15 July, we used the 15 June estimate only. Differences in beetle damage on the three Plant Types were compared using Tukey pairwise comparisons To determine the effects of widespread virus infection on beetle damage to the new growth, we performed similar mixed effects ANOVAs using the 15 August estimates only on plants in non-sprayed and (separately) on the sprayed fields. Finally, we performed chi-square tests of independence to determine the effects of Plant Type on the incidence of bacterial wilt disease before and after the spread of virus infection in the non-sprayed and (separately) sprayed fields (before and after 15 July). Finally, to determine if Treatment (spray/no spray) had an effect on the amount of beetle damage before 15 July, we also used a mixed effects ANOVA (Proc. Mixed). The model was applied to all plants in all fields and included Plant Type, Family (random), Field nested within Treatment and the 2 way interactions involving Treatment, Plant Type and Family on Mean Beetle Damage per plant during June and July.

Results The first visible symptoms of viral disease (revealed by DAS-ELISA tests to be ZYMV; Agdia Inc. Elkhart, IN) occurred in mid to late July in the two unsprayed fields. ZYMV then spread rapidly through the two unsprayed fields so that by the end of August the incidence of ZYMV among the Texas gourds and the ntBC4 was very high (Fig. 5.1). In the sprayed fields, 9 plants showed symptoms of ZYMV infection (confirmed by DAS-ELISA tests) in mid to late August and these plants were removed from the fields to prevent secondary spread of the disease. Neither CMV nor WMV occurred in either the sprayed or unsprayed fields in 2009. In contrast, the VRT effectively deterred viral infections in the tBC4 plants in the unsprayed fields although a few plants exhibited mild symptoms and tested positive for ZYMV in the unsprayed fields in late August (Fig. 5.1). In both the sprayed and unsprayed fields, flowering commenced in early July and peak flower and fruit production occurred from late July to late August. In the sprayed (no virus) fields, only Plant Type significantly affected reproductive output (Table 5.1A).

79 The tBC4 had significantly greater reproductive output than the Texas gourds while the reproductive output of the ntBC4 did not differ significantly from either the Texas gourds or the tBC4 plants (Fig. 5.2). In the unsprayed (virus) fields, both Plant Type and Field had a significant effect on reproductive output (Table 5.1B). In these fields, the tBC4 had significantly greater reproductive output than both the ntBC4 and Texas gourd plants which did not differ significantly in their reproductive output (Fig. 5.2). In the sprayed (no virus) fields, Field, Plant Type and Family had significant effects on mean beetle damage in June and July (Table 5.2A). The Texas gourd plants had significantly more beetle damage than the ntBC4 plants which, in turn, did not differ significantly from the tBC4 plants (Fig. 5.3A). In the unsprayed (virus) fields, Plant Type and the Family by Plant Type interaction significantly affected mean beetle damage in June and July and Family was marginally non-significant (p = 0.058) (Table 5.2B). The Texas gourd plants in these fields had significantly more beetle damage during June and July than both the ntBC4 and the tBC4 plants. There were no significant differences in the amount of beetle damage on the ntBC4 or tBC4 plants (Fig. 5.3B). However, when we examined the August 15 estimate of beetle damage in sprayed fields we found no significant effects of Field or Plant type but we did find significant effects of Family and the interaction between Family and Plant type on beetle damage (Table 5.3A; Fig. 5.4A). In the unsprayed fields we found significant effects of Field, Plant type, Family and the interaction between Field and Family (Table 5.3B).The tBC4 had significantly more beetle damage than the Texas gourds while the ntBC4 were not significantly different from either the Texas gourds (p = 0.11) or the tBC4 (p = 0.51) (Fig. 5.4B). There was no significant effect of treatment on beetle damage during June-July (F1,685 =0.05; p = 0.82) (compare Fig. 5.4A and 5.4B). In the sprayed (no virus) fields, the probability that a plant would contract bacterial wilt disease was independent of Plant Type (Texas gourd, ntBC4, BC4) both before 15 July (chi-square = 0.82; df =2; p > 0.10) and after 15 July (chi-square = 1.49; df = 2; p > 0.10). In the unsprayed (virus) fields, the probability that a plant would contract bacterial wilt disease was also independent of Plant Type before 15 July (chi-square = 2.27; df = 2; p > 0.10). However, after 15 July (after the virus spread through the susceptible plants in the unsprayed fields) the probability that a plant would contract wilt

80 disease was not independent of Plant Type (chi-square = 29.32; df = 2; p << 0.001). Twenty-one percent of the tBC4 contracted wilt disease while only 2.9% of the ntBC4 and 1.6% of the Texas gourds contracted wilt disease. The probability that a plant would contract wilt disease prior to 15 July was independent of the spray treatment (chi-square = 1.15; df =1; p > 0.10).

Discussion This large scale field study examined the impact of a virus resistant transgene (VRT) on reproductive output, herbivory by the primary herbivore, and the incidence of a non-target pathogen (transmitted by the primary herbivore) in the presence (unsprayed fields) and absence (Pymetrozine sprayed fields) of the target viral pathogens. In the unsprayed fields, the VRT target pathogens (viruses) were permitted to naturally establish and spread through the fields. As with all previous studies of transgenic squash-wild gourd hybrids (Fuchs et al. 2004; Laughlin et al. 2009; Sasu et al. 2009), our study shows that the VRT effectively deters a target virus during introgression from cultivated squash to a wild Cucurbita pepo (Texas gourd). As has been previously observed in both transgenic squash and in transgenic squash-wild gourd hybrids (see Fuchs et al. 2004, Sasu et al. 2009), a few of the tBC4 plants in this study had mild symptoms and tested positive for ZYMV at the end of the growing season in unsprayed fields in which ZYMV was epidemic. We used the aphid specific insecticide, Pymetrozine, to control the aphid vectors of the target viral diseases while permitting herbivory by cucumber beetles and exposure to the deadly pathogen, E. tracheiphila, which is transmitted by the cucumber beetles. Our data show that the plants in the sprayed and unsprayed fields have similar amounts of beetle damage prior to July 15 and a similar incidence of wilt disease prior to establishment and spread of ZYMV in the unsprayed fields indicating that Pymetrozine did not impact these components of the pathosystem. However, the Pymetrozine only delayed the entry of ZYMV into our fields by approximately one month. We removed nine ZYMV infected plants from our fields in August, to discourage secondary spread, until we finished collecting flower production data. Thereafter, ZYMV spread through the sprayed fields.

81 Effects of the VRT on Fitness As with all previous studies of the effects of the squash VRT on fitness during introgression into the Texas gourd (Fuchs et al. 2004; Laughlin et al. 2009; Sasu et al. 2009), this study revealed that plants with the VRT enjoy a significant fitness advantage when the virus is present in the fields. The reproductive output of the Texas gourd plants in our unsprayed fields was only 37% of the tBC4 plants while the reproductive output of the ntBC4 plants was 60% of that of the tBC4 plants. However, the fitness advantage of the transgenic plants in this study was not as great as that observed in the Fuchs et al. (2004) and Laughlin et al. (2009) studies in which the non-transgenic plants were all infected with one of the target viruses early in the seedling/pre-reproductive phases of their life cycles. For example, Laughlin et al. (2009) hand-inoculated all of the susceptible seedlings with ZYMV and found that they produced only 6% of the fruits and 3% of the seeds of the transgenic introgressives in the same fields. However, unlike previous studies, our study revealed that the transgenic plants enjoyed a reproductive advantage over the Texas gourds even in the sprayed fields, although the reproductive advantage was not as great as in the unsprayed (virus) fields. The reproductive output of the Texas gourds was only 72% of that of the tBC4 in the sprayed fields while the ntBC4 had 87% of the reproductive output of the tBC4. We suspect that the VRT had a large fitness advantage in the sprayed fields because of a late introduction of viral diseases despite the insecticide treatment. Because it takes 5-10 days for hand inoculated seedlings in a greenhouse to exhibit symptoms of ZYMV, we suspect that aphid transmitted ZYMV to adult healthy plants takes even longer to exhibit symptoms. Consequently, it is likely that by the middle of August a large portion of the plants in the sprayed fields were pre-symptomatic but infected with ZYMV. If pre- symptomatic ZYMV infection reduces the growth rate of Cucurbita plants (one flower is produced per node) or the developmental rate of the flowers that have already differentiated then the sprayed and unsprayed fields only differ in the time and intensity of viral pressure in the fields rather than the presence and absence of viral disease. Although this study used BC4 plants in order to reduce the detrimental effects of traditional crop genes on reproductive output that have been observed in other studies of crop-wild gourd hybrids (Spencer and Snow 2001; Fuchs et al. 2004), there is some

82 indication in our data that conventional crop genes actually had a beneficial impact on reproductive output. In both the sprayed and unsprayed fields the reproductive output of the ntBC4 plants was greater than the Texas gourds. However, this difference was not significant in our multiple pairwise comparisons (Fig.5.2). The Texas gourds used in this study are derived from seeds that were originally collected from a population along a dried river bank in Texas where the growing season is hot and dry (Quesada et al. 1993) while this study took place in agricultural fields in central Pennsylvania, during a relatively wet and cool summer. Therefore, it is possible that the remaining cultivar traits in the BC4 plants provided some advantage in wetter agricultural conditions that would not occur in the native habitats of Texas gourd. Interestingly, Laughlin et al. (2009) found that BC2 plants had higher reproductive output than C. pepo ssp. texana under virus pressure in a study conducted in the state of New York. The lower reproduction of the Texas gourds on both the sprayed and unsprayed plots could also be due to the higher levels of beetle damage that we observed on the Texas gourds compared to both BC4 types prior to 15 July. Because cucumber beetles are attracted to and preferentially feed on leaves containing cucurbitacins (Metcalf et al. 1980; Metcalf and Rhodes 1990), it is possible that conventional breeders inadvertently selected for reduced foliar cucurbitacins. Our previous studies of Texas gourds have repeatedly shown that reproductive output decreases with the amount of beetle damage and that there is broad sense heritable variation for resistance to beetle damage (Stephenson et al. 2004; Du et al. 2008). In this study, we found that Family significantly influences the amount of beetle damage both before and after 15 July in the sprayed fields and before 15 July on the unsprayed fields (after 15 July the effect of Family is marginally insignificant (p = 0.058)). These data suggest that there is broad sense heritable variation for resistance to beetle damage among the five families used in this study.

Effects of the VRT on Cucumber Beetle Foraging and the Incidence of Wilt Disease Our data clearly show that there is an ecological cost of the VRT that is mediated by changes in the foraging behavior cucumber beetles in the presence of a target viral disease. Although the Texas gourds experience greater beetle damage prior to 15 July

83 than either of the introgressives, the beetles preferentially foraged on the tBC4 plants after ZYMV spread through the unsprayed fields. Because the beetles did not prefer the tBC4 plants either before or after 15 July on the sprayed fields nor did they prefer the tBC4 plants before the spread of viral disease among the susceptible plants on the unsprayed fields, it is likely that cucumber beetles prefer healthy plants, rather than the VRT plants per se. Moreover, infection decreased flower production on both susceptible types compared to the tBC4 plants. Consequently, the tBC4 plants had greater exposure to E. tracheiphila via both modes of transmission (foliar feeding and floral transmission) after ZYMV spread through the unsprayed fields. In the unsprayed fields, our data show that there were no significant differences in the incidence of wilt disease among Texas gourds, ntBC4, and tBC4 plants before 15 July. However, after 15 July (when ZYMV was present in the fields) the tBC4 plants experienced a far greater incidence of bacterial wilt disease than did either of the virus susceptible types. In the sprayed fields, there was no effect of Plant Type on the incidence of wilt disease either before or after 15 July. Because beetle damage that occurs during the peak period of reproduction is known to adversely affect reproductive output, in the Texas gourds (Stephenson et al. 2004; Du et al. 2008), it is likely that the greater beetle damage on the tBC4 depressed their reproductive output. In addition, bacterial wilt disease kills the plants over a 10-14d period of time and fruits that are younger than 15d at the time of death fail to mature (i.e., they rotted and were not counted at the end of the season). Consequently, wilt disease both robs plants of future reproduction and it discounts recent fruit production. These findings may account for the two-fold difference in reproductive output that we observed between the tBC4 plants on the unsprayed vs. sprayed fields (see Fig. 5.2). However, it should be noted that the fields were not randomly assigned (hence differences in soil fertility, prior land use history, etc. are confounded with treatment) and that Pymetrozine undoubtedly reduced herbivory by aphids and perhaps other sucking insects on the sprayed fields.

84 Risks to the Environment Transgenic squash and cross compatible taxa of annual Cucurbita are now grown in close proximity in both the southern United States and a large part of Mexico (Arriaga et al. 2006; Decker 1988; Decker-Walters et al. 1993; 2002). Moreover, gene flow between cultivated squash and wild Cucurbita is common and well-documented (Kirkpatrick and Wilson 1988; Decker-Walters et al. 1993; Arriaga et al. 2006). Consequently, it is likely that the VRT has escaped repeatedly into wild populations. Whether the escaped VRT poses an environmental threat largely depends on its fitness advantages during introgression in wild Cucurbita. This study is the fourth to assess the fitness of the VRT during introgression into the Texas gourd (Fuchs et al. 2004; Laughlin et al. 2009; Sasu et al. 2009). These studies differed in the date at which the target virus was introduced into the population, the specific target virus that was used, the techniques used to control non-target herbivores and pathogens (e.g., insecticides, plant removal), and the stage of introgession that was examined (F1 to BC4). All four studies show that the VRT effectively deters infection by the target pathogens (viruses) during introgression; that the VRT has a direct beneficial effect on fitness in the presence of a target virus; that there is no detected direct cost of the transgene in the absence of the virus; and that the costs/benefits of conventional crop traits tend to vary by stage of introgression and year. Together, these studies suggest that the VRT is likely to become established in wild populations and that the rate of increase in the population would depend upon the frequency of a target viral disease in the population and the timing of its introduction. Because surveys in the US conservatively estimate that 38% of wild gourd populations have at least one individual with one of the target viruses (USDA 1996; Quemada et al. 2002; see Laughlin et al. 2009), we can assume that the VRT would provide a fitness advantage and become established in many wild populations if the presence/absence of the virus is the only determinant of VRT establishment. All of these studies of the fitness of VRT plants during introgression, however, have been performed north of the native ranges of wild annual Cucurbita. This study, and our previous study (Sasu et al. 2009) reveal indirect/ecological costs that mitigate the fitness of the VRT during introgression. These indirect costs are mediated by the primary herbivores and the interaction of a target and a non-target pathogen. These findings

85 suggest that the effects of the VRT on fitness depend not only on the frequency and timing of the establishment and spread of a target disease, but also on the population size of the primary herbivores (cucumber beetle herbivory becomes concentrated onto the healthy plants as virus spreads through the population) and the frequency and timing of the establishment and spread of the non-target pathogen that the beetles transmit. Moreover, upon escape into populations of wild Cucurbita in the southern US and Mexico, the VRT is likely to face an even more complex suite of direct and indirect effects that influence the costs/benefits of the VRT, such as competition with other species and, perhaps, a more diverse suite of herbivores and pathogens. Therefore, studies conducted in agricultural settings outside of the natural ranges of wild Cucurbita are limited in their ability to predict the indirect effects of the VRT on fitness. Indirect effects could either enhance the fitness of the VRT in the presence of the virus, or as our data show, reduce the fitness of the VRT in the presence of the virus. To date, studies of crop-wild plant hybrids with resistance transgenes have focused on the likelihood of hybridization, the fitness of the transgenic hybrids in the absence of the target herbivore or pathogen, and occasionally on the fitness of the transgenic hybrids in the presence of the target herbivore/pathogen (see Snow et al. 2003; Burke and Rieseberg 2003; Pilson and Prendeville 2004; Fuchs and Gonsalves 2007 for discussion). This study shows that an escaped transgenic resistance gene can have ecological costs that are not apparent from pairwise challenges (e.g., with and without the target pathogen). Unfortunately, ecological costs (or benefits) of herbivore and pathogen resistance transgenes can only be determined within the context of the full pathosystem in natural habitats and, thus, are unlikely to be studied prior to the deregulation of the transgene.

Acknowledgments We thank J.A. Winsor, S. Scanlon, M. Bothe, and T. Deveney for field, greenhouse and lab assistance, T. Omeis for use of the Biology Greenhouse, and R. Oberheim and his staff for use of the Horticulture Farm at the PSU Agriculture Experiment Station at Rock

86 Springs, PA. This project is supported by the Biotechnology Risk Assessment Program Grant no. 2009-33120-20093 from the USDA National Institute of Food and Agriculture.

ntBC4 wild gourds tBC Cumulative incidence of viral diseases

June July August

Figure 5.1. The proportion of susceptible plants infected with ZYMV on the two unsprayed fields in which viral diseases were allowed to spread.

88 b

B. y

Texas ntBC4 tBC4 gourds

Figure 5.2. Reproductive output (male flowers and fruit) during 2009 field season for Texas gourds, non-transgenic BC4 introgressives (ntBC4) and transgenic BC4 introgressives (tBC4) (LSmeans ± SE). Model: Reproductive output= Field + Family + Plant type (Texas gourds, ntBC4, tBC4) + Family x Plant type A. Reproductive output

89 on sprayed fields Plant type F 2, 344=4.71; P=0.01 B. Reproductive output on unsprayed fields Plant type F 2,334 =10.82; P<0.001.

a b

B. x

y y

Texas ntBC4 tBC4 gourds

Figure 5.3. Amount of beetle damage in June and July 2009 on leaves of Texas gourds, non-transgenic BC4 introgressives (ntBC4) and transgenic BC4 introgressives (tBC4)( Least square means ±SE). Model: Beetle damage= Field + Family + Plant type (Texas gourds, ntBC4, tBC4) + Family x Plant type A. Beetle damage in June and July on sprayed fields Plant type F 2,344=4.88; P=0.008 B. Beetle damage in June and July on unsprayed fields Plant type F 2,333=2.51; P=0.01.

a a

B. y

Texas ntBC4 tBC4 gourds

Figure 5.4. Amount of beetle damage in August 2009 on leaves of Texas gourds, non- transgenic BC4 introgressives (ntBC4) and transgenic BC4 introgressives (tBC4) (Least square means ±SE ). Model: Beetle damage= Field + Family + Plant type (Texas gourds, ntBC4, tBC4) + Family x Plant type. A. Beetle damage in August on sprayed fields Plant

92 type F 2, 265= 1.3; P=0.274 B. Beetle damage in August on unsprayed fields Plant type F 2,258= 6.18; P=0.002.

A. Effect Num Den F p DF DF Field 1 344 0.90 0.34

Family 4 344 1.80 0.13

Plant type 2 344 4.71 0.01

Family x 8 344 1.98 0.06 Plant type

B. Effect Num Den F p DF DF Field 1 344 34.31 <0.0001

Family 4 344 1.88 0.11

Plant type 2 344 10.82 <0.0001

Family x 8 344 0.55 0.82 Plant type

Table 5.1. Effects of field, family (random), plant type (Texas gourds, non-transgenic BC4 introgressives and transgenic BC4 introgressives) and family x plant type on reproductive output during the 2009 season. A. Results of ANOVA on reproductive output for sprayed fields B. Results of ANOVA on reproductive output on unsprayed fields.

A. Effect Num Den F p DF DF Field 1 258 0.00 0.97

Family 4 258 3.16 0.01

Plant type 2 258 1.73 0.18

Family x 8 258 2.77 0.006 Plant type B. Effect Num Den F p DF DF Field 1 251 6.72 0.01

Family 4 251 2.73 0.03

Plant type 2 251 6.00 0.003

Family x 8 251 1.94 0.06 Plant type

Table 5.2. Effects of Field, Family (random), Plant type (Texas gourds, non-transgenic BC4 introgressives and transgenic BC4 introgressives) and Family x Plant type on beetle damage during August 2009. A. Results of ANOVA on reproductive output for sprayed fields B. Results of ANOVA on reproductive output on unsprayed fields.

A. Effect Num Den F p DF DF Field 1 344 73.77 <0.001

Family 4 344 3.87 0.004

Plant type 2 344 4.88 0.008

Family x 8 344 1.50 0.15 Plant type

B. Effect Num Den F p DF DF Field 1 333 1.17 0.28

Family 4 333 2.31 0.06

Plant type 2 333 7.45 0.007

Family x 8 333 2.51 0.01 Plant type

CHAPTER 6 Conclusions and Future Directions

The series of experiments I performed in chapter two demonstrated not only that E. tracheiphila is the causative agent of bacterial wilt disease, but that it can be transmitted through the flowers of Cucurbia pepo ssp. texana (Texas gourd). Prior to this work, it was thought that the sole transmission route of E. tracheiphila was via contaminated cucumber beetle fecal pellets entering into feeding sites on leaves. The greenhouse experiment in which plants were inoculated through flowers without nectar showed that the bacteria are able to pass through the floral tissue and infect the plants. Another greenhouse experiment using GFP-transformed E. tracheiphila to inoculate plants through their flowers showed the progression of E. tracheiphila through the nectaries and into the vascular system of the pedicel below the nectaries. Cross sections of the floral nectaries and pedicels revealed that E. tracheiphila can pass through the floral tissues into the vasculature of the plant within 24 hours after anthesis. To show that plants are exposed to E. tracheiphila via the floral nectaries under field conditions, I counted the proportion of flowers in the field that contained beetle frass (E. tracheiphila is vectored through the frass of cucumber beetles), and found that more than 50% of the flowers in the field contained cucumber beetle frass in or around the nectary tissue. I then developed a Real-time PCR technique to screen cucumber beetle frass for the presence of E. tracheiphila, and found that more than 90% of the contained E. tracheiphila contaminated cucumber beetle frass. Thus it appears that the majority of the plants in the field experience daily exposure to E. tracheiphila. Together, these experiments provide evidence that E. tracheiphila can be transmitted through the flowers. My research showed that E. tracheiphila is transmitted via floral tissues of C. pepo ssp. texana in the greenhouse, but further research is needed to show that floral transmission can also occur in the field. One approach would be to place potted plants that have been subjected to two treatments, one with exposed flowers and one with bagged flowers, in an experimental gourd field. In this manner, these plants would be exposed to naturally occurring populations of cucumber beetles, and the incidence of wilt

96 disease could then be measured. We would predict that, if transmission of E. tracheiphila occurs through the floral nectaries in the field, the plants with flowers would have a higher incidence of bacterial wilt disease than plants with bagged flowers. In chapter three I investigated the antimicrobial properties of C. pepo ssp. texana floral nectar. I first performed disk diffusion assays (DDA) which revealed that floral nectar from field grown Texas gourds has the capacity to inhibit the growth of both E. coli and E. tracheiphila; however, the inhibitory effect of nectar was greater and lasted longer on E. tracheiphila than on E. coli. To investigate the effect of the nectar on the transmission rate of E. tracheiphila through floral tissue of C. pepo, I performed a greenhouse experiment in which there were two treatments (nectar present or absent), and these plants were inoculated with E. tracheiphila through the flowers. I found that plants inoculated through the flowers without nectar showed a greater incidence of wilt disease than plants inoculated through flowers with nectar. Thus it would appear that C. pepo nectar can reduce the rate of transmission of E. tracheiphila through the floral tissue. These findings help to explain the large differences between the exposure rate under field conditions (Chapter 2) and the low overall incidence of wilt disease in the field. These data also suggest that nectar removed by the pollinators (bees) may play an important role in the transmission dynamics of wilt disease in the field. Although we have evidence that Texas gourd nectar can inhibit or slow the growth of E. tracheiphila, the individual chemical components responsible for this inhibition remain unidentified. In order to understand this inhibitory effect, additional studies are needed to isolate potential antimicrobial compounds present in C. pepo nectar because they could have important applications for pest management and disease control in agricultural settings. C. pepo floral nectar should be tested for known antimicrobial compounds such as hydrogen peroxide, alkaloids, and cucurbitacins, which could be achieved using high-performance liquid chromatography (HPLC). Next, volatile profiles of the nectar could be obtained using gas chromatography coupled with mass spectrometry (GC/MS), to identify volatile compounds that are known to have antimicrobial properties. The resulting chemical compounds could then be used to perform individual DDAs on lawns of E. tracheiphila. In addition, since honey is composed of nectar (and honey bees are often employed to service cucurbit fields), the

97 identification of such compounds might also provide insight as to how honey could be used for medicinal purposes, which could have important applications for human health. In chapters four and five I measured the direct and indirect effects of a virus- resistance transgene (VRT) during introgression into wild gourds on fitness, cucumber beetle herbivory, and incidence of wilt disease and mosaic viruses. I performed large- scale field experiments using wild gourds, non-transgenic introgressives and transgenic introgressives. I measured flower and fruit numbers, cucumber beetle damage, and incidence of bacterial wilt disease and mosaic viruses on each individual plant. In chapter four, these experiments were performed using plants that were exposed to the target pathogen (mosaic viruses) and in chapter five, two of the fields were exposed to mosaic viruses while the other two fields were sprayed to prevent viral spread. In chapter five, in addition to measuring the costs and benefits of the VRT during introgression, I measured the costs and benefits of the cultivar genes during introgression. The data show that the complex ecological interactions among plants, vectors/herbivores, and pathogens cannot be predicted by experiments using only pairwise challenges. The data also show that one cannot predict the evolutionary fate (fixation, polymorphism or purging) or the environmental impact of the VRT in a population without measuring the individual components of the pathosystem and the interactions among them. For instance, we had predicted that the fitness of the transgenic plants would be higher than the non-transgenic plants, however, when we measured all individual components and interactions we found that cucumber beetles and bacterial wilt disease indirectly lower the fitness of transgenic plants. To further study the effects of the VRT during introgression on the entire pathosystem, future experiments should focus on varying different components in the system. One example is to design a large-scale field experiment using wild gourds, non- transgenic introgressives and transgenic introgressives, where the viral diseases are manually introduced in the fields as soon as plants are transplanted. Because we have previously found that bacterial wilt disease symptoms appear in the field shortly after the plants are transplanted, introducing the viral diseases early would allow us to examine cucumber beetle behavior in the presence of both viral diseases and wilt disease early in the life cycle. Because we have found that cucumber beetles prefer to feed on the healthy

98 (transgenic) plants, this study would determine if incidence of wilt disease would increase on transgenic plants early in the season, before the transgenic plants have an opportunity to reproduce. We predict that if viral diseases and wilt disease occurred simultaneously early in the season, the cucumber beetles would concentrate on the transgenic plants and in turn this would greatly decrease the fitness of these plants. If this phenomenon occurs, management strategies could be implemented to control the spread of the VRT in the southern U.S. and Mexico, where the transgene has almost certainly escaped into wild gourd populations. It is imperative that we understand the introgression of resistance transgenes into wild populations of plants because these transgenes, if acquired, may alter selection pressures in natural populations. Wild gourds are considered a noxious weed in cultivated fields and the elimination of viral pressures, in particular the mosaic viruses, may create superweeds that affect agricultural practices as well as genetic variation in wild populations. An additional study could vary the proportion of VRT plants in a large-scale experiment using wild gourds, non-transgenic introgressives and transgenic introgressives. This study would decrease the proportion of transgenic plants and allow both viral diseases and bacterial wilt disease to spread through the fields. We would predict that, as a result, cucumber beetles would concentrate their feeding on the fewer (healthy) transgenic plants which in turn would increase the exposure to and incidence of bacterila wilt disease in transgenic plants. As this study would utilize a low frequency of the VRT in the initial population it would more closely resemble an initial transgene escape into wild populations and thus would provide additional information on how the fitness of the VRT would be modulated in natural populations. All of the studies in my dissertation were performed north of the range of wild gourd populations thus the presence of the VRT should be assessed in the native range (southern U.S. and Mexico). In addition, cucumber beetle populations, incidence of viral diseases, and bacterial wilt disease should be measured in these areas. These studies would yield a more accurate estimation of the introgression rate of the VRT into wild populations. My finding of a new transmission mode for E. tracheiphila combined with my discovery that the nectar of C. pepo inhibits the growth and transmission rate of E.

99 tracheiphila, will aid in developing new management strategies for combating bacterial wilt disease in agricultural settings. My finding that the fitness benefit of the VRT was minimized by cucumber beetle herbivory and wilt disease incidence (non-target pathogen) was contrary to the current literature and provides new insight into how transgene fitness is affected by multiple (non-target) selective pressures. The greatest contribution that my dissertation makes to science is that I have shown that individual components of a pathosystem cannot be considered in isolation but must be examined as a part of a larger whole.

100 REFERENCES Accord, B. R. 1996. Asgrow Seed Co. Availability of determination of nonregulated status for a squash line genetically engineered for virus resistance. Federal Register 61:33484-33485. Adler, L. S. 2000. The ecological significance of toxic nectar. Oikos 91:409-420. Adler L. S., M. Wink, M. Distl, and A. J. Lentz. 2006. Leaf herbivory and nutrients increase nectar alkaloids. Ecology Letters 9:960-967. Anderson, J. F., and R. L. Metcalf. 1986. Identification of a volatile attractant for Diabrotica (Coleoptera, Chrysomelidae) and Acalymma (Coleoptera, Chrysomelidae) spp. from blossoms of Cucurbita maxima duchesne. Journal of Chemical Ecology 12:687-699. Anderson, J. F., and R. L. Metcalf. 1987. Factors influencing distribution of Diabrotica spp. (Coleoptera, Chrysomelidae) in blossoms of cultivated Cucurbita spp. Journal of Chemical Ecology 13:681-699. Antonovics, J. 2005. Plant venereal diseases: insights from a messy metaphor. New Phytologist 165:71-80. Arriaga, L., E. Huerta, R. Lira-Saade, E. Moreno, and J. Alarcón. 2006. Assessing the risk of releasing transgenic Cucurbita spp. in Mexico. Agric Ecosyst & Environ 112:291-299. Ashman, T- L. 2002. The role of herbivores in the evolution of separate sexes from hermaphroditism. Ecology 83:1175-1184. Avila-Sakar, G., G. A. Krupnick and A. G. Stephenson. 2001. Resource allocation to growth and reproduction in Cucurbita pepo ssp. texana: Effects of fruit removal. International Journal of Plant Sciences 162:1089-1095. Baker, H. G., and I. Baker. 1973. Amino acids in nectar and their evolutionary significance. Nature 241:543-545. Baldwin, I. T., C. Preston, M. Euler, and D. Gorham. 1997. Patterns and consequences of benzyl acetone floral emissions from Nicotiana attenuata plants. Journal of Chemical Ecology 23:1175-1184.

101 Barrett, S. C. H., and C. G. Eckert. 1990. Variation and evolution of mating systems in seed plants. Pages 229-254 in D. Kawano, ed. Biological approaches and evolutionary trends in plants. Academic Press, New York, NY, USA. Bazzaz, F. A. 1997. Allocation of resources in plants: state of the science and critical questions. Pages 1-38 in A. Bazzaz and J. Grace, eds. Plant resource allocation. Academic Press, San Diego, CA, USA. Bergelson, J., and C. B. Purrington. 1996. Surveying patterns in the cost of resistance in plants. American Naturalist 148:536-558. Biere, A., and J. Antonovics. 1996. Sex specific costs of resistance to the fungal pathogen Ustilago violacea (Microbotryum violaceum) in Silene alba. Evolution 50:1098- 1110. Blancard, D., H. Lecoq, and M. Pitrat. 1994. A colour atlas of cucurbit diseases: observation, identification & control. Halstead Press. New York, NY, USA. Blua, M. J., and T. M. Perring. Effect of Zucchini yellow mosaic virus on development and yield of cantaloupe Cucumis melo. Plant Disease 73(4): 317-320. Bollen, K. A. 1989. Structural equations with latent variables. Wiley, New York, NY, USA. Boots, M., and R. G. Bowers. 1999. Three mechanisms to host resistance to macroparasites—avoidance, recovery and tolerance—show different evolutionary dynamics. Journal of Theoretical Biology 201:13-23. Brust, G. E. 1997a. Interaction of Erwinia tracheiphila and muskmelon plants. Environmental Entomology 26:849-854. Brust, G. E. 1997b. Seasonal variation in percentage of striped cucumber beetles (Coleoptera: Chrysomelidae) that vector Erwinia tracheiphila. Environmental Entomology 26:580-584. Brust, G. E. 1997c. Seasonal variation in percentage of striped cucumber beetles (Coleoptera: Chrysomelidae) that vector Erwinia tracheiphila. Environmental Entomology 26:580-584. Burdon, J. J. 1987a. Diseases and plant population biology. Cambridge University Press, Cambridge, U.K.

102 Burdon, J. J. 1987b. Phenotypic and genetic patterns of resistance to the pathogen Phakopsora pachyrhizi in populations of Glycine canescens. Oecologia 73:257- 267. Bruton, B. D., U. Melcher, T. Zitter, S. D. Pair, J. Fletcher, and F. Mitchell. 1999. Polymerase chain reaction for detection of Erwinia tracheiphila in cucurbits [abstract]. Phytopathology 89:s10 Bubán, T., Zs. Orosz-Kovács and Á. Farkas. 2003. The nectary as the primary site of infection by Erwinia amylovora (Burr.) Winslow et al.: a mini review. Plant Systematics and Evolution 238:183-194. Burdon, J. J. 1987a. Diseases and plant population biology. Cambridge University Press, Cambridge, U.K. Burdon, J. J. 1987b. Phenotypic and genetic patterns of resistance to the pathogen Phakopsora pachyrhizi in populations of Glycine canescens. Oecologia 73:257- 267. Burdon, J. J., and S. R. Leather. 1990. Pests, pathogens and plant communities. Blackwell Scientific, Oxford, U. K. Burke, J. M., and L. H. Rieseberg. 2003. Fitness effects of transgenic disease resistance in sunflowers. Science 300:1250-1252. Byrne, B. 2001. Structural equation modeling with AMOS: Basic concepts, applications and programming. Laurence Irlbaum Associates, New Jersey, USA. Campbell, L. G., and A. A. Snow. 2007. Competition alters life history and increases the relative fecundity of crop-wild radish hybrids (Raphanus spp.). New Phytologist 173:648-660. Cantliffe, D., N. Shaw, and P. Stoffella. 2005. Current trends in Cucurbit production in the U. S. National Statistics Service, U. S. Department of Agriculture. Vegetable Summary 1994 and 2004. Carr, D. E., and M. D. Eubanks. 2002. Inbreeding alters resistance to insect herbivory and host plant quality in Mimulus guttatus (Scrophulariaceae). Evolution 56:22-30. Carter, C. and R. W. Thornburg 2000. Tobacco nectarin I: purification and characterization as a germin-like, manganese superoxide dismutase implicated in

103 the defense of floral reproductive tissues. Journal of Biological Chemistry 275:36726-36733. Carter, C, and R. W. Thornburg. 2004. Is the nectar redox cycle a floral defense against microbial attack? Trends in Plant Science (9) 7:320-324. Carlsson-Graner., U. 1997. Anther-smut disease in Silene dioica: variation in susceptibility among genotypes and populations, and patterns of disease within populations. Evolution 51:1416-1426. Charlesworth, D., and B. Charlesworth. 1987. Inbreeding depression and its evolutionary consequences. Annual Review of Evolution and Systematics 18:237-268. Chen, L.-Y., A. A. Snow, F. Wang, and L. Boa-Rong. 2006. Effects of insect-resistance transgenes on fecundity in rice (Oryza sativa, Poaceae): a test for underlying costs. American Journal of Botany 93:94-101. Coley, P. D., and J. A. Barone. 1996. Herbivory and plant defenses in tropical forests. Annual Review of Ecology and Systematics 27:305-335. Conner, J. K., S. Rush and P. Jennetten. 1996. Measurements of natural selection on floral traits in wild radish (Raphanus raphanistrum). I. Selection through lifetime female fitness. Evolution 50:1127-1136. Cormack, R. S, I. E Somssich, and K. Hahlbrock. 1998. Isolation of putative plant transcriptional coactivators using a modified two-hybrid system incorporating a GFP reporter gene. Plant Journal 14:685-692. Crnokrak, P., and S. C. H. Barrett. 2002. Perspective: purging the genetic load: a review of the experimental evidence. Evolution 56: 2347-2358. Darlington, M. 2006. Sexual selection in the spotted cucumber beetle, Diabrotica undecimpunctata howardi (Coleoptera: Chrysomelidae): The benefits of female choice? Ph.D. dissertation, Biology Program, The Pennsylvania State University, University Park, PA, USA. Decker, D. S., and H. D. Wilson. 1987. Allozyme variation in Cucurbita pepo complex C. pepovar. overifera vs. C. texana. Systemic Botany 12:263-273. Decker-Walters, D. S. 1990. Evidence for multiple domestication of Cucurbita pepo. Evidence for multiple domestication of Cucurbita pepo. Pages 96-101 in D. M.

104 Bates, R. W. Robinson, and C. Jeffrey, eds. Biology and utilization of the Cucurbitaceae. Cornell University Press, Ithaca, NY, USA.. Decker-Walters, D. S., T. W. Walters, C. W. Cowan, and B. D. Smith. 1993. Isozymic characterization of wild populations of Cucurbita pepo. Journal of Ethnobiology 13:55-72. Decker-Walters, D. S., J. E. Straub, S-M. Chung, E. Nakata and H. D. Quemada. 2002. Diversity in free-living populations of Cucurbita pepo (Cucurbitaceae) as assessed by random amplified polymorphic DNA. Systematic Botany 27:19-28. Delph, L. F., M. H. Jóhannsson, and A. G. Stephenson. 1997. How environmental factors affect pollen performance: ecological and evolutionary perspectives. Ecology 78:1632-1639. De Mackiewicz, M. Blua, S. J. Fleischer, F. L. Lukezic, and F. E. Gildow. 1998. Herbaceous weeds are not important reservoirs of Erwinia tracheiphila. Plant Disease 82:521-529. De Moraes, C. M., W. J. Lewis, P. W. Parés, H. T. Alborn, and J. H. Tumlinson. 2001. Herbivore-infested plants selectively attract parasitoids. Nature 393:570-573. Devadas, S. K., A. Enyedi, and R. Raina. 2002. The Arabidopsis hrl1 mutation reveals novel overlapping roles for salicylic acid, jasmonic acid and ethylene signaling in cell death and defence against pathogens. Plant Journal 30:467-480. Du, D., J. A. Winsor, M. Smith, A. DeNicco, and A. G. Stephenson. 2008. Resistance and tolerance to herbivory changes with inbreeding and ontogeny in a wild gourd (Cucurbitaceae). American Journal of Botany 95:84-92. Dudash, M. R., D. E. Carr, and C. B. Fenster. 1997. Five generations of enforced selfing and outcrossing in Mimulus guttatus: inbreeding depression variation at the population and family level. Evolution 51:54-65. Eben, A., and M. E. Barbercheck. 1996. Field observations on host plant associations and natural enemies of Diabroticite beetles (Chrysomelidae: Luperini) in Veracruz, Mexico. Acta Zoologica Mexicana 67:47-65. Eben, A., and A. Espinosa de los Monteros. 2004. Ideas on the systematics of the genus Diabrotica Wilcox and other related leaf beetles. Pages 59-73 in P. Jolivet, J. A.

105 Santiago-Blay, and M. Schmitt, eds. New developments in the biology of Chrysomelidae. SPB Academic Publishing, The Hague, The Netherlands. Ehrlén, J. S., J. Kack, and J. Ågren. 2002. Pollen limitation, seed predation and scape length in Primula farinosa. Oikos 97:45-51. Eigenbrode, S. D., H. Ding, P. Shiel, and P. H. Berger. 2002. Volatiles from potato plants infected with potato leafroll virus attract and arrest the virus vector, Myzus persicae (Homoptera: Aphididae). Proceedings of the Royal Society of London B. 269:455-460. Ellers-Kirk, D. D. 1996. Population dynamics of Acalymma vittata (Fab.) (Coleoptera:Crysomelidae) and incorporation of entomopathogenic nematodes for integrated pest management of diabroticites in fresh-market cucurbits. MS thesis, The Pennsylvania State University, University Park, PA. Ellstrand, N. C. 2003a. Current knowledge of gene flow in plants: implications for transgene flow. Philosophical Transactions of the Royal Society of London B. 358:1163-1170. Ellstrand, N. C. 2003b. Dangerous liaisons? When cultivated plants mate with their wild relatives. The John Hopkins University Press, Baltimore, MD, USA. Ellstrand, N. C, H. C. Prentice, J. C. Hancock. 1999. Gene flow from domesticated plants Into their wild relatives. Annual Review of Evolution and Systematics 30:539-563. Feber, D. 1999. Genetically modified crops in the crosshairs. Science 286:1662-1665. Fenster, C. B. and M. R. Dudash. 1994. Genetic considerations for plant population restoration and conservation. Pages 34-62 in M. L. Bowles and C. J. Whalen, eds. Restoration of endangered species. Cambridge University Press, Cambridge, U K. Ferguson, J. E., R. L. Metcalf and D. C. Fischer. 1985. Disposition and fate of cucurbitacin B in five species of diabroticites. Journal of Chemical Ecology 11:1307-1321. Ferrari, M. J., O. N. Bjornstad, J. L. Partain, and J. Antonovics. 2005. A gravity model for the spread of a pollinator-borne plant pathogen. American Naturalist 168:294- 303.

106 Ferrari, M. J., C. M. De Moraes, A. G. Stephenson, and M. C. Mescher. 2006. Inbreeding effects on blossom volatiles in Cucurbita pepo ssp. texana. American Journal of Botany 93:1768-1774. Ferrari, M., J. A. Winsor, D. Du, and A.G. Stephenson. 2007. Inbreeding alters host plant quality and incidence of an insect borne pathogen in Cucurbita pepo ssp. texana. International Journal of Plant Sciences 168:603-610. Field, A. P. 2009. Discovering statistics using SPSS: and sex and drugs and rock ‘n’roll. Third Edition. London: Sage. Fleischer, S. J., D. de Mackiewicz, F. E. Gildow, and F. L. Lukezic 1999. Serological estimates of the seasonal dynamics of Erwinia tracheiphila in Acalymma vittata (Coleoptera : Chrysomelidae). Environmental Entomology. 28:470-476. Fritz, R. S., and E. L. Simms. 1992. Plant resistance to herbivores and pathogens: ecology, evolution and genetics. The University of Chicago Press, Chicago, IL, USA. Fuchs, M., and D. Gonsalves. 1995. Resistance of transgenic squash Pavo ZW-20 expressing the coat protein genes of zucchini yellow mosaic virus and watermelon mosaic virus 2 to mixed infections by both potyviruses. Bio/Technol 13:1466- 1473. Fuchs, M., E. M. Chirco, J. R. McFerson, and D. Gonsalves. 2004. Comparative fitness of a wild squash species and three generations of hybrids between wild X virus- resistant transgenic squash. Environmental Biosafety Research 3:17-28. Fuchs, M., and D. Gonsalves. 2007. Safety of virus-resistant transgenic plants two Decades after their introduction: Lessons from realistic field risk assessment studies. Annual Review of Phytopathology 45:173-202. Garcia-Salazar, C., F. E. Gildow, S. J. Fleischer, D. Cox-Foster and F. L. Lukezic. 2000. Alimentary canal of adult Acalymma vittata (Coleoptera: Chrysomelidae): Morphology and potential role in survival of Erwinia tracheiphyla (Enterobacteriaceae). Canadian Entomologist 132:1-13. Galen, C. 1983. The effects of nectar thieving ants on seed set in floral scent morphs of Polemonium viscosum. Oikos 41:245-249. Garrett, K. A., and C. C. Mundt. 1999. Epidemiology in mixed host populations.

107 Phytopathology 89:984-990. Gassman, A. J., and D. J. Futuyma. 2005. Consequence of herbivory for fitness cost of herbicide resistance: photosynthetic variation in the context of plant-herbivore interactions. Journal of Evolutionary Biology 18:447-454. Hayes, C. N., J. A. Winsor, and A. G. Stephenson. 2004. Inbreeding influences herbivory in Cucurbita pepo ssp texana (Cucurbitaceae). Oecologia 140: 601-608. Hancock, R. D., J. Chudek, P. Walker, S. D. A. Pont, and R. Viola 2008. Ascorbic acid isolated from the phloem of Cucurbitaceae. Phytochemistry 69:1850-1858. Hare, J. D., E. Elle, and N. M. van Dam. 2003. Costs of glandular trichomes in Datura wrightii: a three-year study. Evolution 57:793-805. Harrewijn, P., and H. Kayser. 1997. Pymetrozine, a fast-acting and selective inhibitor of aphid feeding. In-situ studies with electronic monitoring of feeding behaviour. Pesticide Science 49:185-212. Herrera, C. M., I. M. Garcia, and R. Perez. 2008. Invisible floral larcenies: microbial communities degrade floral nectar of bumble-bee pollinated plants. Ecology 89:2369-2376. Heil, M., and I. T. Baldwin. 1997. Fitness costs of induced resistance: emerging experimental support for a slippery theory. Trends in Plant Science 7:61-67. Herbert, J., Isham, V. 2001. A study of the roll of the mechanism of transmission in macroparasite aggregation. Journal of Applied Probability 38A:249-262. Hinkelmann, K, and K. Oscar. 2008. Design and analysis of experiments, Volume I: Introduction to experimental design. (Second Edition). Wiley, New York, NY, USA. Hull-Sanders, H. M., and M. D. Eubanks. 2005. Plant defense theory provides insight into interactions involving inbred plants and insect herbivores. Ecology 86:897- 904. Husband, B. C., and D. W. Schemske. 1996. Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50:54-70. Irwin, M. E., and J. M. Thresh. 1990. Epidemiology of barley yellow dwarf; a study in ecological complexity. Annual Review of Phytopathology 28:393-424.

108 Irwin, R. E., L. S. Adler and A. K. Brody. 2004. The dual role of floral traits: Pollinator attraction and plant defense. Ecology 85:1503-1511. Ivey, C. T., and D. E. Carr. 2005. Effects of herbivory and inbreeding on the pollinators and mating system of Mimulus guttatus (Phrymaceae). American Journal of Botany 92:1641-1649. Jiménez Martínez, E. S., N. A. Bosque-Perez, P. H. Berger, R. S. Zemetra, H. Ding, and S. D. Eigenbrode. 2004. Volatile cues influence the response of Rhopalosiphum padi (Homoptera: Aphididae) to barley yellow dwarf virus-infected transgenic and nontransformed wheat. Environmental Entomology 33:1207-1216. Junior, V. A., V. R. Celloto, L. G. E. Vieira, J. E. Gonsalves, R. A. C. Gonsalves, and A. J. B. de Oliviera. 2008. Floral nectar chemical composition of floral nectar in conventionalandtransgenic sweet orange, Citrus sinensis (L.) Osbeck, expressing an antibacterial peptide. Plant Systematics and Evolution 1-2:1-7. Kaltz, O. and J. A. Shykoff. 2001. Male and female Silene latifolia plants differ in per- contact risk of infection by a sexually-transmitted disease. Journal of Ecology 89: 99-109. Karban, R., and I. T. Baldwin. 1997. Induced responses to herbivory. The University of Chicago Press, Chicago, IL, USA. Kessler, A., and I. T. Baldwin. 2001. Defensive function of herbivore-induced plant volatile emissions in nature. Science 291:2141-2144 Kennedy, G. G., and J. D. Barbour. 1992. Resistance variation in natural and managed systems. Pages 13-41 in R. S. Fritz and E. L. Simms, eds. Plant resistance to herbivores and pathogens: ecology, evolution and genetics. The University of Chicago Press, Chicago, IL, USA. Kim, W.S., L. Gardan, L. Rhim, S.L. and K. Geider. 1999. Erwinia pyrifoliae sp. Nov., a novel pathogen that affects Asian pear trees (Pyrus pyrifolia Nakai). International Journal of Bacteriology 49:899-905. Kingsolver, J. G. and D. W. Schemske. 1991. Path analysis of selection. Trends in Ecology and Evolution 6:276-280. Kirkpatrick, K. J., and H. D. Wilson. 1988. Interspecific gene flow in Cucurbita: C. texana vs. C. pepo. American Journal of Botany 75:519-527.

109 Klas, F. E., M. Fuchs, and D. Gonsalves. 2006. Comparative spatial spread over time of Zucchini yellow mosaic virus (ZYMV) and Watermelon mosaic virus (WMV) in fields of transgenic squash expressing the coat protein genes of ZYMV and WMV, and in fields of nontransgenic squash. Transgenic Resistance 15:527-541. Kniskern, J. M. and M. D. Rausher. 2007. Natural selection on a polymorphic disease- resistance locus in Ipomoea purpurea. Evolution 61:377-387. Koella, J. C., P. Agnew, and Y. Michalakis. 1998. Coevolutionary interactions between host life histories and parasite life cycles. Parasitology 116 Suppl.:S47-55. Kuehl R. O. 2000. Design of Experiments: Statistical Principles of Research Design and Analysis. Second Edition. Duxbury Press, Pacific Grove, Ca, USA. Lacroix, R., W. R. Mukabana, L. C. Gouagna, and J. C. Koella. 2005. Malaria infection increases attractiveness of humans to mosquitoes. Public Library of Science Biology 3(9): e298. Lampman, R. L., and R. L. Metcalf. 1988. The Comparative Response of Diabrotica Species (Coleoptera, Chrysomelidae) to Volatile Attractants. Environmental Entomology 17:644-648. Lampman, R. L., R. L. Metcalf, and J. F. Andersen. 1987. Semiochemical attractants of Diabrotica undecimpunctata Howardi Barber, Southern corn rootworm, and Diabrotica virgifera virgifera Leconte, the Western corn rootworm (Coleoptera, Chrysomelidae). Journal of Chemical Ecology 13:959-975. Laughlin, K. D., A. G. Power, A.A. Snow, and L.J. Spencer. 2009. Risk assessment of genetically engineered crops: fitness effects of virus-resistance transgenes in wild Cucurbita pepo. Ecological Applications 19:1091-1101. Leach, J. G. 1964. Observations on cucumber beetles as vectors of cucurbit wilt. Phytopathology 54:606-607. Leimu R., and J. Koricheva. 2006. A meta-analysis of genetic correlations between plant resistances to multiple enemies. American Naturalist 168:E15-E37. LeFevre, T., J. C. Koella, F. Renaud, H. Hurd, D. G. Biron and F. Thomas. 2006. New prospects for research on manipulation of insect vectors by pathogens. Public Library of Science Pathogens 2(7) e72.

110 Lira, R., T. C. Andres, and M. Nee. 1995. Cucurbita. Pages 1-115 in R. Lira, (ed). Systematic and ecogeographic studies on crop genepools. Volume 9. International Plant Genetic Resources Institute. Mexico City and Rome. Lloyd-Smith, J. O., S. J. Schreiber, P. E. Kopp, and W. M. Getz. 2005. Superspreading and the effect of individual variation on disease emergence. Nature 438:355-359. Louda, S. M., K. H. Keeler, and R. D. Holt. 1990. Herbivore influences on plant performance and competitive interactions. Pages 413-444 in Grace, J., and D. Tilman, eds. Perspectives on plant competition. Academic Press, San Diego, CA, USA. Lukezic, F. L., W. M. Sackett, S. J. Fleischer, M. D. Orzolek and F. E. Gildow. 1996. Influence of concentration of Erwinia tracheiphila cells on the development of wilt symptoms in field-grown cucumbers and cantaloupe plants. Phytopathology abstracts 86:S123. MacClement, W., and M. Richards. 1956. Virus in wild plants. Canadian Journal of Botany. 34:793. MacNab, A. A., A. F. Sherf, and J. K. Springer. 1994. Identifying diseases of vegetables. The Pennsylvania College of Agricultural Sciences, University Park, PA, USA. Maddox, G. D., and R. B. Root. 1990. Structure of the encounter between goldenrod

(Solidago altissima) and its diverse insect fauna. Ecology 71:2115-2124.

Maleck, K., A. Levine, T. Eulgem, A. Morgan, J. Schmid, K. A. Lawton, J. L. Dangl, and R. A. Dietrich. 2000. The transcriptome of Arabidopsis thaliana during systemic acquired resistance. Nature Genetics 26:403-410. Manson, J. S., M. C. Otterstatter, and J. D. Thompson. 2010. Consumption of a nectar alkaloid reduces pathogen load in bumble bees. Oecologia 162:81-89. Martin, G. B., A. J. Bogdanove, and G. Sessa. 2003. Understanding the functions of plant disease resistance proteins. Annual Review of Plant Biology 54:23-61. Martini M., A. Schmid, and D. Hess. 1990. Antibiotics sugars and amino acids in nectar Of Rhododendron spp. and Piptanthus spp. from Nepal. Botanica Acta 103: 343-348.

111 Marquis, R. J. 1992. The selective impact of herbivores. Pages 301-325 in R. S. Fritz and E. L. Simms, eds. Plant resistance to herbivores and pathogens: ecology, evolution and genetics. The University of Chicago Press, Chicago, IL, USA. Maskell, L. C., A. F. Raybould, J. I. Cooper, M. L. Edwards, and A. J. Gray. 1999. Effects of turnip mosaic virus on the survival, growth and reproduction of wild cabbage (Brassica oleracea). Annals Applied Biology 135:401-407. Matheson, A. C., T. L. White, and G. R. Powell. 1995. Effects of inbreeding on growth, stem form and rust resistance in Pinus elliottii. Silvae Genetica 44:37-46. May R. M. and R. M. Anderson. 1987. Transmission Dynamics of Hiv- Infection. Nature 326:137-142. McDowell, J. M., and B. J. Woffenden. 2003. Plant disease resistance genes: recent insights and potential applications. Trends in Biotechnology 21: 178-183. McElhany, P., L. A. Real and A. G. Power. 1995. Vector preference and disease dynamics: a study of barley yellow dwarf virus. Ecology 76:444-457. McNaughton, S. J., D. A. Oesterheld, D. A. Frank, and K. J. Williams. 1989. Ecosystem- level patterns of primary productivity and herbivory of terrestrial ecosystems. Nature 341:142-144. Metcalf , R.L., R. A. Metcalf, and A. M. Rhodes. 1980. Cucurbitacins as kairomones for diabroticite beetles. Proceedings of the National Academy of Sciences 77:3769- 3772. Metcalf, R. L., and R. L. Lampman. 1991. Evolution of diabroticite rootworm beetle (Chrysomelidae) receptors for Cucurbita blossom volatiles. Proceedings of the National Academy of Sciences 88: 1869-1872. Metcalf, R. L., and A. M. Rhodes. 1990. Coevolution of Cucurbitaceae and Luperini (Coleoptera: Chrysomelidae): basic and applied aspects. Pages 167-182 in D. M. Bates, R. W. Robinson, and C. Jeffrey, eds. Biology and utilizaton of the Cucurbitaceae. Cornell University Press, Ithaca, NY, USA. Mitchell, R. F. and L. M. Hanks. 2009. Insect frass as a pathway for transmission of bacterial wilt of cucurbits. Environmental Entomology 38: 395-403. Mitchell-Olds, T., and D. Bradley. 1996. Genetics of Brassica rapa. 3. Costs of disease resistance to three fungal pathogens. Evolution 50:1859-1865.

112 Molyneux, D. H. and D. Jeffries. 1986. Feeding behaviour of pathogen-infected vectors. Parasitology 92:721-736. Mooney, H. A. and N. R. Chiariello. 1984. The study of plant function: the plant as a balanced system. Pages 305-323 in R. Dirzo and J. H. Sarukhán, eds. Perspectives on plant population ecology. Sinauer Associates, Inc., Sunderland, MA, USA. Moran, P. J. 2001. The effects of wilt symptom development and peroxidase induction on interactions between vascular wilt bacteria and cucumber beetles. Entomologia Experimentalis et Applicata 98:149-156. Muller, K. E. and B. A. Fetterman. 2002. Regression and ANOVA: An Integrated Approach Using SAS Software. SAS Publishing, SAS Institute, Cary, NC, USA. Mutikainen, P., and L. F. Delph. 1996. Effects of herbivory on male reproductive success in plants. Oikos 75:353-358. National Research Council. 2002. Environmental effects of transgenic plants: The scope and adequacy of regulation. National Academy Press, Washington, DC, USA. Nepi, M., F. Ciampolini, and E. Pacini 1996. Development and ultrastructure of Cucurbita pepo nectaries of male flowers. Annals of Botany 78:95-104. Nepi, M., M. Guarneri, and E. Pacini E. 2001. Nectar secretion, and sugar composition in male and female flowers of Cucurbita pepo. International Journal of Plant Sciences. 162:353-358. Nicolson, S., and R. W. Thornburg 2007. Nectar Chemistry. Ch 5: 215-263 in Nectaries and Nectar: A Modern Treatise. Ch. 5:215-263 in Pacini, E. Nepi, M., and S. Nicolson, eds. Springer, Dordrecht, The Netherlands. Nishida R., and H. Fukami. 1990. Sequestration of distasteful compounds by some pharmacophagous insects. Journal of Chemical Ecology 16:151-164. Nunez-Farfan, J., R. A. Cabrales-Vargas, and R. Dirzo. 1996. Mating system consequences on resistance to herbivory and life history traits in Datura stramonium. American Journal of Botany 83:1041-1049. Oliver, L, S. Harrison, and M. McClelland. 1983. Germination of Texas gourd (Cucurbita texana) and its control in soybeans (Glycine max). Weed Science 31:700-706.

113 Onofri A., E. A. Carbonell, H. P. Piepho, A. M. Mortimer, and R. D. Cousens. 2009. Current statistical issues in weed research. Weed Research 50:5-24. Ouborg, N. J., A. Biere, and C. L. Mudde. 2000. Inbreeding effects on resistance and transmission-related traits in the Silene-Microbotryum pathosystem. Ecology 81:520-531. Park, Y. H., S. Sensoy, C. Wye, R. Antonise, J. Peleman, and M. J. Havey. 2000. A genetic map of cucumber composed of RAPDs, RFLPs, AFLPs, and loci conditioning resistance to papaya ringspot and zucchini yellow mosaic viruses. Genome 43:1003-1010. Pilson, D., and H. R. Prendeville. 2004. Ecological effects of transgenic crops and the escape of transgenes into wild populations. Annual Review of Ecology, Evolution and Systematics 35:149-174. Pigliucci, M., and A. Kolodynska. 2006. Phenotypic integration and response to stress in Arabidopsis thaliana: a path analytical approach. Evolutionary Ecology Research 8:415-433. Purrington, C. B. 2000. Costs of resistance. Current Opinion in Plant Biology 3:305-308. Puustinen, S., T. Koskela, and P. Mutikainen. 2004. Direct and ecological costs of resistance and tolerance in the stinging nettle. Oecologia 139:76-82. Quemada, H., D. S. Decker-Walters, and J. E. Staub. 2002. Case study: gene flow from commercial transgenic Cucurbita pepo to “wild” C. pepo populations. Pages 65- 70 in Proceedings of the Scientific Methods Workshop: Ecological and Agronomic Consequences of Gene Flow from Transgenic Crops to Wild Relatives, March 5-6, Columbus, Ohio. Quesada, M. R., J. A. Winsor, and A. G. Stephenson. 1993. Effects of pollen competition on progeny performance in a heterozygous cucurbit. American Naturalist 142:694-706. Quesada, M. R., K. Bollman, and A. G. Stephenson. 1995. Leaf damage decreases pollen production and hinders pollen performance in Cucurbita texana. Ecology 76:437- 443. Rand, F. V. and E. M. A. Endlows. 1920. Bacterial wilt of cucurbits. U. S. Dept. Agric. Bull. 828.

114 Raybould, A. F., L. C. Maskell, M. L. Edwards, J. I. Cooper, and A. J. Gray. 1999. The prevalence and spatial distribution of viruses in natural populations of Brassica oleracea. New Phytologist 141:265-275. Robinson, R. W., and D. S. Decker-Walters. 1997. Cucurbits. Oxford University Press, New York, NY, USA. Ryals, J. A., U. H. Neuenschwander, M. G. Willits, A. Molina, H.-Y. Steiner, and M. D. Hunt. 1996. Systemic acquired resistance. The Plant Cell 8:1809-1819. Saoirse, N. G., O. Bergh, B. Dixon, L. Donachie, J. Carson, R. Coyne, J. Curtin, and I. Dalsgaard. 2004. The precision and robustness of published protocols for disk diffusion assays of antimicrobial agent susceptibility; an inter-laboratory study. Aquaculture 240:1-18. Sasu, M. A, I. Seidl-Adams, K. Wall, J. A. Winsor, and A. G. Stephenson. Floral Transmission of Erwinia tracheiphila by Cucumber Beetles in a Wild Cucurbita pepo. Environmental Entomology. In press. Sasu, M. A, M. J. Ferrari, D. Du, J. A. Winsor, and A. G. Stephenson 2009. Indirect costs of a non-target pathogen mitigate the direct benefits of a virus resistant transgene in wild cucurbita. Proceedings of the National Academy of Sciences 106:19067- 19071. Saccheri, I., M. Kuussaari, M. Kankare, P. Vikman, W. Fortelius, and I. Hanski. 1998. Inbreeding and extinction in a butterfly metapopulation. Nature 292:491-494. SAS Institute. 2002. SAS User’s Guide. SAS Institute, Cary, NC, USA. Scheiner, S. M., R. J. Mitchell, and H. S. Callahan. 2000. Using path analysis to measure natural selection. Journal of Evolutionary Biology 13:423-433. Scheiner, S. M., K. Donohue, L. A. Dorn, S. J. Mazer and L. M. Wolfe. 2002. Reducing environmental bias when measuring natural selection. Evolution 56:2156-2167. Schenk, P. M., K. Kazan, I. Wilson, J. P. Anderson, T. Richmond, S. C. Somerville, and J. M. Manners. 2000. Coordinated plant defense responses in Arabidopsis revealed by microarray analysis. Proceedings of the National Academy of Sciences 97:11655-11660. Senasica. 2004. Regulations for genetically modified organisms in agriculture. Ministry of Agriculture, Livestock, Rural Development, Fisheries and Food. Mexico (In

115 Spanish). http://www.web2.senasica.sagarpa.gob.mx/xportal/inocd/trser/Doc403/ Shipley, B. 2000. Cause and correlation in biology: a user’s guide to path analysis, structural equations and causal inference. Cambridge University Press, Cambridge, U. K. Shykoff, J. A., E. Bucheli, and O. Kaltz. 1996. Flower lifespan and disease risk. Nature 379:779. Simms, E. L. and J. Triplett. 1994. Costs and benefits of plant resistance to disease: resistance and tolerance. Evolution 48:1973-1985. Smyth, R. R. and M. P. Hoffmann. 2003. A male-produced aggregation pheromone facilitating Acalymma vittatum F. (Coleoptera: Chrysomelidae) early-season host plant colonization. Journal of Insect Behavior 16:347-359. Snow, A. A., D. Pilson, L. H. Rieseberg, M. J. Paulsen, N. Pleskac, M. R. Reagon, D. E. Wolf, and S. M. Selbo. 2003. A Bt transgene reduces herbivory and enhances fecundity in wild sunflowers. Ecological Applications 13:279- 286. Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman and Co., New York, NY, USA. Spencer L, J., and A. A. Snow. 2001. Fecundity of transgenic wild-crop hybrids and Cucurbita pepo (Cucurbitaceae): implications for crop-to-wild gene flow. Heredity 86:694- 702. Spinelli, F., F. Ciampolini, M. Cresti, K. Geider and G. Costa. 2005. Influence of stigmatic morphology on flower colonization by Erwinia amylovora and Pantoea agglomerans. European Journal of Plant Pathology 113:395-405. Steets, J. A., T. M. Knight and T-L. Ashman. 2007. The interactive effects of herbivory and mixed mating for population dynamics of Impatiens capensis. American Naturalist 170:113-127. Stephenson, A. G. 1981. Toxic nectar deters nectar thieves of Catalpa speciosa. American Midland Naturalist. 105:381-383. Stephenson, A. G. 1982. Iridoid glycosides in the nectar of Catalpa speciosa are unpalatable to nectar thieves. Journal of Chemical Ecology 8:1025-1034.

116 Stephenson, A. G., B. Leyshon, S. E. Travers, C. N. Hayes, and J. A. Winsor. 2004. Interrelationships among inbreeding, herbivory, and disease on reproduction in a wild gourd. Ecology 85:3023-3034. Stephenson, A. G., J. A. Winsor, D. Du, A. DeNicco and M. Smith. 2006. Herbivory by cucumber beetles affects pollen production and pollen performance in a wild gourd. Pages 301-308 in Cucurbitaceae 2006. G. Holmes, ed. North Carolina. State University Press, Raleigh, NC, USA. Strauss, S. Y., and R. Karban. 1994. The significance of outcrossing in an intimate plant-herbivore relationship. 1. Does outcrossing provide an escape from herbivores to the parent plant? Evolution 48:454-464. Strauss S., J. Conner, and S. Rush. 1996. Foliar herbivory affects floral characters and plant attractiveness to pollinators: implications for male and female plant fitness. American Naturalist 17:1098. Strauss, S. Y., and A. A. Agrawal. 1999. The ecology and evolution of plant tolerance to herbivory. Trends in Ecology and Evolution 14:179-185. Strauss, S. Y., J. K. Conner, and K. P. Lehtilä. 2001. Effects of foliar herbivory by insects on the fitness of Raphanus raphanistrum: damage can increase male fitness. American Naturalist 158:496-504. Strauss, S. Y., J. A. Rudgers, J. A. Lau, and R. E. Irwin. 2002. Direct and ecological costs of resistance to herbivory. Trends in Ecology and Evolution 17:278-285. Tallamy, D.W. 1985. Squash beetle feeding behavior: an adaptation against induced cucurbit defenses. Ecology 66:574-1579 Tallamy, D. W., and V. A. Krischik. 1989. Variation and function of cucurbitacins in Cucurbita: an examination of current hypotheses. American Naturalist 133:766- 786. Tallamy, D. W. and W. P. Brown. 1999. Semelparity and the evolution of maternal care in insects. Animal Behaviour 57:727-730. Tallamy, D. W., D. P. Whittington, F. Defurio, D. A. Fontaine, P. M. Gorski, and D. P. W. Gothro. 1998. Sequestered cucurbitacins and pathogenicity of Metarhizium anisopliae (Moniliales : Moniliaceae) on spotted cucumber beetle eggs and larvae (Coleoptera : Chrysomelidae). Environmental Entomology 27:366-372.

117 Theis, N., M. Lerdau and R.A. Raguso. 2007. The challenge of attracting pollinators while evading floral herbivores: Patterns of fragrance emission in Cirsium arvense and Cirsium repandum (Asteraceae). Intenational Journal of Plant Science 168:587-601. Thomson, S. V. 1986. The role of the stigma in fire blight infection. Phytopathology 76: 476-482. Thornburg, R. W., C. Carter, A. Powell, R. Mittler, L. Rizhisky, and H. T.Horner. 2003. A major function of the tobacco floral nectary is defense and microbial attack. Plant Systematics and Evolution 238:211-218. Tian, D., M. B. Traw, J. Q. Chen, M. Kreitman and J. Bergelson. 2003. Fitness costs of R-gene-mediated resistance in Arabidopsis thaliana. Nature 423:74-77. Tiffin, P. and M. D. Rausher. 1999. Genetic constraints and selection acting on tolerance to herbivory in the common morning glory Ipomoea purpurea. American Naturalist 154:700-716. U. S. Department of Agriculture (USDA). 1994. Environmental assessment for Upjohn Company/Asgrow Seed Company petition for determination of non-regulated status for ZW-20 squash. USDA, Washington, D. C., USA. U. S. Department of Agriculture (USDA). 1996. Environmental assessment for Upjohn Company/Asgrow Seed Company petition for determination of non-regulated status for CZW-3 squash. USDA, Washington, D. C., USA. Uyenoyama, M., K. E. Holsinger, and D. M. Waller. 1993. Ecological and genetic factors directing the evolution of self-fertilisation. Oxford Surveys in Evolutionary Biology 9:327-381. Valdivia, E. R., D. J. Cosgrove, and A. G. Stephenson. 2006. Role of accelerated style senescence in pathogen defense. American Journal of Botany 93:1725-1729 Vogler, D. W., K. Filmore, and A. G. Stephenson. 1999. A comparison of inbreeding depression in plants derived from strong and weak self-incompatibility phenotypes of Campanula rapunculoides L. Journal of Evolutionary Biology 12:483-494.

118 Weinig, C., J. R. Stinchcombe and J. Schmitt. 2003. Evolutionary genetics of resistance and tolerance to natural herbivory in Arabidopsis thaliana. Evolution 57:1270- 1280. Wiens F., A. Zitzmann, M-A. Lachance, M. Yegles, F. Pragst, F. M Wurst, and D. von

Holst. 2008. Chronic intake of fermented floral nectar by wild treeshrews.

Proceedings of the National Academy of Sciences (USA) 105:10426-10431.

Wen, X., M. Gleason, J. C. Batzer, G. A. Beattie, and G. Lewis. 2006. Assessment of a specific primer for Erwinia tracheiphila, causal agent of curcurbit bacterial wilt. [abstract]. Phytopathology 96:s122. Wilson, H. D., R. Lira, I. Rodríguez. 1994. Crop/weed gene flow: Cucurbita Agyrosperma Huber and C. fraterna LH Bailey. Economic Botany 48:293-300. Winsor, J. A., S. Peretz, and A. G. Stephenson. 2000. Pollen competition in a natural population of Cucurbita foetidissima (Cucurbitaceae). American Journal of Botany 87:527-532. Wyss, P., and W. Bolsinger. 1997. Translocation of pymetrozine in plants. Pesticide Science 50:195-202. Yao, C. B., G. Zehnder, E. Bauske, and J. Klopper 1996. Relationship between cucumber beetle (Coleoptera: Chrysomelidae) density and incidence of bacterial wilt of cucurbits. Journal of Economic Entomology 89:510-514.

119 Curriculum Vitae Miruna Sasuclark EDUCATION: 2005-May 2010 -The Pennsylvania State University State College, PA Ph.D Biology and PhD minor Statistics

2001-2005-Colby College Waterville, ME B.A. Biology and Spanish

SKILLS: PCR: Endpoint and Real-Time Culture: Bacterial isolation, Bacterial culture, Diffusion assays Immunological: Enzyme-Linked Immunosorbent Assay (ELISA) Imaging: Spectrophotometry, Zeiss Light Microscopes and working knowledge of Confocal microscopes

WORK EXPERIENCE: 2005-2010- Graduate Research Fellow at Penn State University 2001-2005- Undergraduate Research Assistant Colby College

TEACHING EXPERIENCE: 2009- Teaching Assistant - Honors Introductory Biology 1 section (BIO110H) 2007 and 2008-Teaching Assistant- Introductory Biology 2 sections per semester (BIO110) 2006 and 2007-Teaching Assitant-Tropical Ecology in Costa Rica (BIO 499)

GRANTS, AWARDS AND FELLOWSHIPS: 2009- Popp Scholarship for outstanding research 2006, 2008, 2010-Braddock Research Award for productivity and publication record 2007-Hill Graduate Award for research excellence 2005- Margaret Chase Smith Scholarship for influential women in science

VOLUNTEER SERVICE Centre County Paws 2005 – 2008-Centre County PAWS-Volunteered as kennel technician and fostered animals. 2003 – 2005-Waterville Humane Society-Volunteered as kennel technician and was appointed veterinary technician.