The Pennsylvania State University

The Graduate School

Department of Entomology

CHEMICAL ECOLOGY OF CO-OCCURRING LYGUS BUGS: EFFECTS ON HOST

PLANTS AND METHODS FOR SUSTAINABLE CONTROL

A Dissertation in

Entomology

by

Sean T. Halloran

2012 Sean T. Halloran

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

December 2012

The dissertation of Sean T. Halloran was reviewed and approved* by the following:

James H. Tumlinson Ralph O. Mumma Professor of Entomology Dissertation Advisor Chair of Committee

Shelby Fleischer Professor of Entomology

John Tooker Assistant Professor of Entomology

David Mortenson Professor of Weed and Applied Plant Ecology

Gary Felton Professor of Entomology Head of the Department of Entomology

*Signatures are on file in the Graduate School

iii

ABSTRACT Lygus bugs (Hemiptera: Miridae) are considered important pests in many crop systems throughout the United States. Lygus feed by injecting salivary enzymes into plant tissue

(reproductive structures and growing tips) through their piercing-sucking mouthparts, which eventually causes death and abscission of plant structures at the feeding site. Lygus are commonly controlled with pesticides, but are developing resistance to the most commonly used chemical controls. Novel controls, such as resistant cultivars, trap crops, and natural enemies, should be explored as sustainable alternatives to chemical spraying for Lygus. Here we examine basic aspects of plant defense (indirect and direct) against Lygus bugs in an economically relevant crop host (Medicago sativa) and a wild host plant (Melilotus officinalis), and examine chemically- mediated interactions between Lygus, an introduced parasitoid bio-control organism (Peristenus relictus) and a promising trap crop (Erigeron annuus) that could function as a component of sustainable control measures. Our results indicate that Lygus feeding induces volatiles and phytohormones characteristic of both pathogen and herbivore defense pathways, and that these aspects of the plant defense response vary by Lygus species and life stage, as well as among crop hosts. Our data are consistent with there being variation in the enzymes present in the saliva of different species/stages of Lygus, and further indicate that the plant-derived breakdown products of these enzymes may be capable of bypassing traditional herbivore defense pathways (e.g., JA induction) to directly elicit defense responses. We also found that E. annuus releases a complex blend rich in green leaf volatiles, terpenes, and other compounds, and that this blend is highly attractive to both female Lygus bugs as well as to female P. relictus parasitoids relative to VOCs being released from cotton plants. These results indicate that E. annuus has great potential for acting as a trap crop that can pull Lygus off of crop plants in addition to attracting and provisioning Lygus natural enemies (introduced Peristenus wasps).

iv

TABLE OF CONTENTS

Abstract ...... iii

List of Figures...... vi

List of Tables ...... vii

Acknowledgements ...... viii

Chapter 1 - Introduction ...... 1

Plant Defense and Volatile Signaling ...... 1 Regulation of Induced Plant Defenses ...... 5 Biology of Study Organisms ...... 6 Overview of Chapters ...... 10 References...... 16

Chapter 2 - Chemical ecology of two co-occurring Lygus species on shared host plants ...... 23

Abstract ...... 23 Introduction ...... 24 Methods ...... 27 Results ...... 32 Discussion ...... 35 References...... 41

Chapter 3 - Induction of phytohormones in response to feeding by two species of Lygus bugs on alfalfa and sweet clover ...... 55

Abstract ...... 55 Introduction ...... 56 Methods ...... 60 Results ...... 65 Discussion ...... 67 References...... 73

Chapter 4 - Chemical ecology of Erigeron annuus: a weedy host and trap crop of bugs in the genus Lygus ...... 81

Abstract ...... 81 Introduction ...... 82 Methods ...... 85 Results ...... 90 Discussion ...... 92 References...... 97

v

Chapter 5 - Attraction of the parasitoid wasp, Peristenus relictus, to Erigeron annuus, a potential trap crop for Lygus bugs ...... 110

Abstract ...... 110 Introduction ...... 111 Methods ...... 116 Results ...... 121 Discussion ...... 123 References...... 127

Chapter 6 - Conclusions...... 135

References...... 145

vi

LIST OF FIGURES

Figure 2.1. Medicago sativa green leaf volatiles principal component analysis...... 45

Figure 2.2. Medicago sativa green leaf volatiles individual compounds ...... 46

Figure 2.3. Medicago sativa terpene principal component analysis ...... 47

Figure 2.4. Medicago sativa terpene individual compounds...... 48

Figure 2.5. Medicago sativa aromatic volatiles and unidentified volatiles analysis...... 49

Figure 2.6. Melilotus officinalis green leaf volatiles principal component analysis ...... 50

Figure 2.7. Melilotus officinalis green leaf volatiles individual compounds ...... 51

Figure 2.8. Melilotus officinalis terpene principal component analyis...... 52

Figure 2.9. Melilotus officinalis terpene individual compounds...... 53

Figure 2.10. Melilotus officinalis aromatic volatiles and unidentified volatiles analysis...... 54

Figure 3.1. Medicago sativa Feeding Site Analysis...... 77

Figure 3.2. Medicago sativa Systemic Site Analysis...... 78

Figure 3.3. Melilotus officinalis Feeding Site Analysis ...... 79

Figure 3.4. Melilotus officinalis Systemic Site Analysis...... 80

Figure 4.1. Erigeron annuus and Cotton Total Volatiles...... 101

Figure 4.2. Response of Adult Female Lygus Bugs to Erigeron and Cotton Volatiles ...... 102

Figure 4.3. Response of Adult Female Lygus Bugs to Clean Air and Plant Volatiles...... 103

Figure 4.4. Erigeron annuus and Cotton Total Volatiles During Y-tube Assays...... 104

Figure 5.1. Response of Peristenus relictus Females to Plant Volatiles ...... 130

Figure 5.2. Erigeron annuus and Cotton Total Volatiles...... 131

vii

LIST OF TABLES

Table 4.1. Choice tests performed to examine movement of Lygus bugs in response to volatile cues...... 105

Table 4.2. Mean amounts of each compound +/- standard error from undamaged, adult damaged, and nymph damaged Erigeron plants...... 106

Table 4.3. Mean amounts of each compound +/- standard error from undamaged, adult damaged, and nymph damaged cotton plants...... 109

Table 5.1. Mean amounts of each compound +/- standard error from undamaged, and nymph damaged Erigeron and cotton plants...... 132

viii

ACKNOWLEDGEMENTS

I have many people to thank for their help and support in creating this dissertation. I would like to thank my advisor, Dr. James Tumlinson, whose knowledge and experience was invaluable over the years. This work could not have been completed without his support. I would also like to thank my dissertation committee members, Dr. Shelby Fleisher, Dr. John Tooker, and Dr. David Mortensen for helping me along in developing this dissertation. I also want to thank all the hard-working people in the Entomology office: Thelma Brodzina, LuAnn Weatherholtz, Karen Dreibelbis, Roxie Smith, Ellen Johnson, Dave Love, Pamela Murray, and Marcia Kerschner. Without them I doubt anyone could reach graduation. Our department head, Dr. Gary Felton, also deserves special mention for working with me to help me reach this goal. A special thanks to all of the members of the Tumlinson lab and the other denizens of the Chemical Ecology building: Christy Harris, Tracy Conklin, Ezra Schwarzberg, Liz Bosak, Emily Kuhns, Naoko Yoshinaga, Irmgard Seidl-Adams, Jon Lelito, and Dan Schemhl. Bryan Banks deserves appreciation for the construction of the wind tunnel I used and being the person to go to for my engineering needs. Thanks also to Nate McCartney, who keeps everything in the lab running and still has time to help us in developing our experiments. This work could not have been completed without the support of Thomas Dorsey from the Philip Alampi Beneficial Insect Laboratory, New Jersey Department of Agriculture from which I obtained many of the insects used in this dissertation. I would like to thank all my friends who gave me support through the years: Dr. Amanda Bachman, Dr. Lori Shapiro, Dr. Alexis Barbarin, Dr. Fernanda Peñaflor, and Dr. Nina Stanczyk. Finally, none of this would have been possible without the support I received from my family: Gregory Halloran, Desiree Halloran, Siobhan Halloran, and especially my wife Kerry Mauck.

Chapter 1

Introduction

Plant Defense and Volatile Signaling

Plants are subject to attack from a variety of shoot and root-feeding insect herbivores employing feeding strategies that inflict different amounts of damage. For this reason plants have evolved complex defense mechanisms that function to preserve plant tissue and resources in order to maintain some level of fitness in the face of these many biotic attackers. The first level of these defenses comes in the form of constitutive or

‘passive’ defenses that are produced over the course of the plant’s normal growth. These can take the form of physical barriers such as trichomes, resin production, or lignification, and also include a vast array of secondary metabolites such as terpenes, latex, or phenolics (Gatehouse 2002, Taiz and Zeiger 2006). However, when herbivores are able to bypass these constitutive defenses (either through behavioral adaptations, or biochemical methods of detoxifying secondary metabolites) plants are also able to upregulate the production of inducible in-plant direct defenses (e.g. α-amylase inhibitors, lectins, and proteinase inhibitors) as well as indirect defenses such as volatile organic compounds (VOCs)(Taiz and Zeiger 2006). Some VOCs common to many plant species are six carbon aldehydes, alcohols, and esters (frequently referred to as green leaf volatiles), terpenoids such as monoterpenes, sesquiterpenes, and homoterpenes, as well aromatics such as indole and methyl salicylate (Pare and Tumlinson 1999).

2 The VOCs are of particular interest as they seem to not only affect the signaling of the attacked plant (Arimura et. al. 2000, Frost et. al. 2007, Heil and Silva Bueno

2007), but also the signaling of nearby downwind plants (Farmer and Ryan 1990,

Engelberth et. al. 2004) as well as the behavior of insect herbivores and natural enemies

(Turlings et. al. 1990, De Moraes et. al. 1998). Like in-plant chemical defenses, plants can have a constitutive profile of VOCs that they give off even when undamaged with both up-regulation of constitutive compounds and production of novel compounds occurring in response to insect feeding (Dicke and van Loon 2000). Many of these induced VOCs are not produced in response to simple mechanical wounding, but will only appear in response to being attacked by an insect herbivore (Pare and Tumlinson

1997). These novel compounds are produced by the plants in response to specific bioactive compounds, called ‘elicitors’, which are found in the insect’s saliva or regurgitant. The first of these elicitors to be characterized was the fatty acid- amino acid conjugate, volicitin, which is found in the regurgitant of beet army-worm and triggers volatile release when applied to mechanically inflicted wounds in maize (Alborn et. al.

1997). Since then a number of unique protein and non-protein insect salivary or regurgitant components that induce specific volatile responses in plants have been found in different insect taxa (Mattiacci et. al. 1995, Alborn et. al. 1997, Koch et al. 1999,

Alborn et. al. 2007).

The specificity of induced VOCs produced in response to elicitors suggests that plant odors can function as signals that convey information about the attacking herbivore species among parts of the same plant, between different plants, and to other insects. For instance, studies have found that VOCs can be important in mediating interactions

3 between plants and herbivores (Bernasconi et al. 1998, De Moraes et al. 2001, Szendrei and Rodriguez-Saona 2010) and between herbivores and their predators (McCormick et al. 2012). VOCs are also thought to play a role in protection from pathogens (which may be introduced directly by an insect vector, or may gain entry incidentally as a result of herbivore wounding). For example, green leaf volatiles have been shown to have anti- bacterial effects against gram-negative and gram-positive bacteria, as well as fungicidal activity (Matsui 2006). These anti-microbial effects may play a role in protecting the plants from potential pathogen infection at the site of feeding damage.

The mediation of interactions across multiple trophic levels by constitutive and induced VOCs is of particular interest to researchers because it provides insight into the adaptive nature of VOC emissions for plants, and also has practical applications for controlling pests. As previously mentioned, the blend of VOCs given off by a plant can affect the behavior of herbivores that encounter these blends. Many herbivores are attracted to the constitutive volatile blend released from their host plants (Dicke 2000,

Szendrei and Rodriguez-Saona 2010), and some are even attracted to VOC blends released from plants currently being fed upon by conspecifics (Blacker et. al. 2004), perhaps as a way of locating mates or to overwhelm plant defenses such as in many species of bark beetles (reviewed in Raffa 2001). In other cases, volatiles released from already damaged plants can act as feeding or oviposition deterrents that actually repel insect herbivores from visiting the plants (Khan et. al. 2000, De Moraes et. al. 2001,

Kessler and Baldwin 2001), and this knowledge has been used to manipulate agricultural landscapes (e.g., through interplanting of repellant plants) in a way that reduces pest pressure within a vulnerable crop (Khan et al. 2000, Cook et al. 2007). Additionally,

4 predators and parasitic wasps that prey on herbivorous insects have been shown to respond preferentially to plant volatiles that are induced by insect feeding (Turlings et al.

1990, Barbosa et al. 1991, De Moraes et al. 1998). In many cases the preferential responses shown by these natural enemies have been shown to be a result of learning on the part of the predator or parasitoid. Parasitoid wasps in particular will respond positively in the future to VOCs they encountered during either oviposition or feeding

(Lewis and Takasu 1990, Papaj et. al. 1994, Vet et. al. 1995), but there is also evidence that other predators can learn to associate prey with certain plant cues (De Boer et al.

2005). Arthropod behavior can also be influenced by chemical cues from conspecifics and prey organisms (Janssen et al. 1997, Ide et al. 2007).

Knowledge of these interactions has resulted in the development of control strategies that use biological means, such as predators and parasitic wasps, to manage pest populations in agricultural settings (Pickett et al. 1997, Cook et al. 2007, Shennan

2008). Rather than relying on chemical pesticides, these bio-control strategies rely on native and released natural enemies to reduce pests below economically damaging levels.

As organic agriculture increases in acreage and popularity, and as farmers move towards reduced pesticide use and more integrated pest management techniques, knowledge of interactions among host plants, pests, and their natural enemies have become more important. There is a particular need for new research in crop systems that are at the threshold of a transition to pesticide-free management practices, since such information could hasten this transition and significantly reduce pesticide usage among farmers.

5 Regulation of Induced Plant Defenses

Plant defenses against insect herbivores are regulated through a series of signal cascades that have been studied mainly in plant-insect interactions involving chewing herbivores. The process begins when a chewing herbivore damages plant tissue, triggering the release of linolenic acid from the cell membrane of plant cells (Snoeren et al. 2009). Linolenic acid then serves as the precursor for both jasmonic acid (JA) and also for the green leaf volatiles (GLVs) (Gatehouse 2002, Matsui 2006). To make JA, linolenic acid is first converted into (9S, 13S)-oxophytodienoic acid (OPDA) within the chloroplast and OPDA is then moved to the peroxisome where it is converted into JA

(Gatehouse 2002, Taiz and Zeiger 2006). Jasmonic acid is considered the main wound- signaling molecule in plants and is believed to play a role in the up-regulation of herbivore defense genes including those that produce proteinase inhibitors and systemic volatile release (Thaler 1996, Gatehouse 2002, Taiz and Zeiger 2006). However, this response is typically not observed when phloem-feeding piercing-sucking herbivores attack plants. Instead, these piercing-sucking insects have been shown to not only avoid triggering JA production, but instead seem to trigger the production of a different phytohormone, salicylic acid (SA) (Walling 2000, Zarate et. al. 2007).

In contrast to JA, SA production is usually associated with pathogen infection of the plant tissues. SA and its conjugates are thought to be the signaling molecules that trigger pathogenesis related (PR) genes that lead to plant responses such as systemic acquired resistance or other gene products meant to fight off pathogens (Kessman et. al.

1994). Of particular interest, however, is the fact that JA and SA have been shown to have an antagonistic relationship, where the production of one phytohormone suppresses

6 the production of the other (Thaler et. al. 2002, Zarate et. al. 2007, Thaler et. al. 2012).

Researchers have shown that some herbivores will exploit this fact by preferentially laying eggs on plants already infected with a pathogen and thus unable to mount a JA- based response (Cardoza et. al. 2003). Additionally, it has been theorized that phloem- feeding insects may be triggering SA in order to avoid having to overcome their host’s herbivore defense response (Zarate et al. 2007).

A rather substantial amount research has been carried out on the regulation of plant defense in response to chewing herbivores, and on a select number of phloem- feeding insects, but there is still very little data on how plants respond to piercing-sucking insects that do not feed on the phloem. If the different feeding styles of insect herbivores were to be organized on a continuum of how much wounding damage they cause to their host plants, chewing herbivores would be on one end while phloem-feeders would be on the other, and piercing-sucking insects such as Lygus bugs would sit intermediate between the two. Understanding how plants respond to damage from herbivores with a macerate-and-flush feeding style (like Mirids and other plant bugs) would aid our understanding of what types of defenses plants mount against these insects and how these defenses may impact plant vulnerability or resistance to other organisms in a complex landscape.

Biology of Study Organisms

Lygus bugs

Lygus bugs are Hemipteran insects belonging to the family Miridae, with over

160 species found worldwide (Schuh 2012). Several species within the genus are considered important pests in crop systems throughout the United States. One species in

7 particular, L. lineolaris, has a host range of over 300 different plant species (Young

1986), and is a major pest on cotton, alfalfa, and a variety of fruit crops along the Eastern

US (Clancy and Pierce 1966, Tingey and Pillemer 1977, Young 1986, Leigh and

Mathews 1996, Layton 2000, Matos and Obrycki 2004), while a second species, L. rubrosignatus, is not considered a pest, but shares a host range with L. lineolaris (Goeden and Ricker 1974, Kelton 1980, Young 1986, Scudder 1997) and may contribute to crop damage as well.

Lygus bugs feed by injecting salivary enzymes into plant tissue through their piercing-sucking mouthparts, preferring to feed on meristematic tissue such as apical buds and fruiting bodies (Handley and Pollard 1993, Leigh and Mathews 1996, Layton

2000). Damage to plant hosts is primarily the result of the death of tissues in and around the site of feeding and is directly caused by salivary components (Handley and Pollard

1993, Leigh and Mathews 1996, Layton 2000, Rodriguez-Saona et. al. 2002, Shackel et. al. 2005). This type of feeding damage contrasts with the tissue removal and mechanical damage characteristic of caterpillar feeding, which has been well characterized. It also contrasts with the relatively unobtrusive feeding style of phloem feeders, like aphids, which cause very little tissue damage while removing plant nutrients (Walling 2000, Will et al. 2007). Thus, Mirid bugs, and Lygus in particular, represent an understudied group of organisms in terms of plant-insect interactions, even though these insects are major pests and such knowledge would assist in development of Lygus-resistant cultivars. This requires intimate knowledge of Lygus salivary components, their effects on plant tissues, and host defense responses at the genetic and metabolite level.

8 Research has shown that the macerate-and-flush feeding style of bugs in the family Miridae (subfamily Mirinae) is associated with key cell wall degrading enzymes

(CWDEs) called polygalacturonases (PGs). These PGs are present in particularly high amounts in Lygus bugs (Shackel et al. 2005, Frati et al. 2006), and research has shown that injection of PGs from fungi into developing bud tissues produces damage symptoms like those of Lygus feeding (Shackel et al. 2005). PGs degrade pectin in the cell walls of plant hosts, and in the process produce oligogalacturonide fragments (OGs) that have been shown to act as elicitors of plant defenses in interactions between PG-producing fungal and bacterial pathogens and their hosts (D’Ovidio et al. 2004). In fact, much of the previous work on PGs and host responses has focused on pathogen-host plant interactions, while ignoring the role of PGs in interactions among insects and their hosts

(Federici et al. 2006). Characterization of Lygus PGs is limited, but suggests that PG protein structure and amounts in the salivary glands can differ by species and by sex

(Frati et al. 2006, de la Paz Celorio-Mancera et al. 2008). It is hypothesized that PGs in

Lygus saliva may elicit specific plant defenses such as PG inhibiting proteins (PGIPs), the phenylpropanoid pathway, and the production of salicylates (Frati et al. 2006). Links between OGs and the jasmonic acid pathway have also been shown (Song and Nam

2005). However, the role of PGs in elicitation of direct and indirect plant defenses

(volatiles), and the potential differences in elicitation among bug species and life stage have not been investigated.

Lygus pests are currently controlled through the use of insecticides. In cotton, growers have begun to reduce early season spraying for many non-Lygus pests due in part to the success of boll weevil eradication programs as well as the development and

9 widespread use of transgenic cotton that controls lepidopterous pests (Layton 2000,

Hardee et.al. 2001). Additionally, L. lineolaris has shown a consistent ability to develop resistance to different classes of insecticides (Zhu et. al. 2004, Zhu et. al. 2011).

Snodgrass (1996) demonstrated widespread resistance to pyrethroids as well as some organophosphate and cyclodiene insecticides. More recently, varying levels of acephate resistance have been identified in L. lineolaris populations in the areas of Louisiana,

Mississippi, and Arkansas (Snodgrass & Scott 2000). The practice of reduced sprayings and the development of resistance to multiple classes of insecticides indicate a need for more sustainable Lygus controls measures.

Lygus parasitoids

Due to the fact that native natural enemies of Lygus, including several parasitoids, have been shown, alone, to not provide adequate control in crop fields (Clancy and Pierce

1966, Day et.al. 1990, Carignan et.al. 2007), a non-native species of parasitoid wasp,

Peristenus relictus (Hymenoptera: Braconidae), has begun to be released in the US for use as a biocontrol agent of Lygus. Peristenus relictus (known as P. stygicus in older literature) is a multivoltine parasitoid wasp native to the Mediterranean region of Europe.

In its native range, P. relictus attacks the nymphal stages of Lygus rugulipennis (the

European tarnished plant bug) as well as several other Lygus, and non-Lygus mirid species (Mason et. al. 2011). Within the past few years, a rearing program for this wasp has begun and releases have been made in alfalfa in New Jersey and California with over- wintering success recorded in southern regions suitable for cotton growing (Pickett et.al

2007, Hoelmer et.al. 2008). Peristenus relictus is a promising candidate for biological

10 control of Lygus in cotton, since it is native to regions that experience similar weather conditions and temperatures as cotton-growing regions, and is already cleared for release in the U.S. Additionally, rearing programs are established for P. relictus, so production and release is already streamlined. However, all current research on preferences and performance of P. relictus is limited to alfalfa, even though this wasp has the potential to control Lygus in many agricultural systems in the southern U.S.

Overview of Chapters

This thesis integrates studies on Lygus-host interactions at the biochemical and molecular level with behavioral studies that examine the effects of host changes in response to Lygus damage on bug and natural enemy behavior. The goals of this thesis are to better understand the response of host plants to the feeding style of Lygus bugs, and to provide basic information that will aid in developing more sustainable control strategies for Lygus pests. Specifically, my project will focus on the following objectives:

1) Characterize the defense responses of cultivated and weedy legume hosts to feeding damage by different species and life stages of Lygus bugs.

Research has demonstrated that the magnitude and nature of in-plant and VOC responses varies depending on the attacking species, with consequences for both the herbivore performance and attraction of predators (De Moraes et. al. 1998, Agrawal

2000, Walling 2000, Rasmann et. al. 2005). Plant antagonists also respond to these defenses with the evolution of counter-acting strategies, leading to an “arms race” between biotic enemies and their sessile plant hosts. Research on these interactions has

11 focused intensively on characterizing the biochemical and genetic responses of crops in response to feeding by specific pests – namely the herbivorous larval (caterpillar) forms of insects in the order Lepidoptera that feed by chewing and removing plant tissue

(Walling 2000). This research has led to a deeper understanding of plant defense evolution and crop resistance capabilities. However, very little is known about how insects with vastly different feeding strategies, such as the macerate-and-flush feeding style of Lygus bugs, elicit plant defenses at the biochemical and genetic level (de la Paz

Celorio-Mancera et al. 2008). Furthermore, it is unknown whether Lygus-specific volatile responses in hosts vary by species and life stage (factors that are relevant for both bug and natural enemy foraging).

In chapter 2 of this thesis the volatile response of two species of legumes fed upon by Lygus were measured under a series of different damage treatments. The two plant species used were the cultivated Medicago sativa cv.‘charger’ (alfalfa) and the wild legume, Melilotus officinalis (sweet clover). These plants were subjected to feeding damage from Lygus lineolaris, a pest of many agricultural plant species across the

Eastern coast of the US, and L. rubrosignatus, a species not considered a pest but that is found sharing many of the same host plants as L. lineolaris. The adult and nymphal life stages of both of these bug species were allowed to feed on these plants and the volatile organic compounds that were released from the plants were collected onto a chemical adsorbent and then analyzed using a gas chromatograph (GC) and gas chromatograph- mass spectrometer (GC-MS).

The volatile blends were then compared using Principal Component Analysis, as well as other univariate statistics, and we found that that both Lygus lineolaris and L.

12 rubrosignatus induce volatile blends that are distinct both between the two bugs species, but also that are also distinct between adults and nymphs of the same species. These results reinforce previous work by other researchers on the presence of ‘elicitors’ within

Lygus bug salivary components, and also indicate that that these components vary throughout the life of one species in addition to differing between species, which in turn cause variation in elicitation of volatile synthesis and release.

In chapter 3 we examined the in-plant phytohormone response of the two plant species tested in chapter 2 when fed upon by the different life stages and species of Lygus bugs. Adults and nymphs of both L. lineolaris and L. rubrosignatus were confined to a feeding site on the alfalfa or M. officinalis plants and allowed to feed for 24 hours. The next day, both the feeding site and an undamaged but systemically connected trifoliate were excised, weighed, and immediately flash frozen in liquid nitrogen. The tissue was the processed and phytohormones were collected using vapor-phase extraction as described in Schmelz et. al. (2003). The level of phytohormones contained within the tissue was analyzed using a gas chromatograph coupled with a mass spectrometer (GC-

MS), and the resulting amounts were standardized by the weight of tissue used.

We measured indicators of direct defenses, namely the phytohormones jasmonic acid, salicylic acid, auxin, abscisic acid (ABA), linolenic acid (a precursor for JA and some volatiles), and the SA precursor cinnamic acid. Around the feeding sites we found that both JA and SA were induced by feeding from Lygus bugs, and there was also an up- regulation of linolenic acid, but these changes in phytohormone production did not carry over to the systemic site. Bug species also modulated the levels of induction of ABA at the feeding site. Overall, we found that bug feeding elicits responses characteristic of

13 both pathogen attack (SA) and chewing herbivore attack (JA), and that there were species level differences in the pattern of phytohormone induction between the two bug species.

However, we were not easily able to match these differences in phytohormone induction with the differences we previously saw in volatile induction.

2) Determine the suitability of the common edge weed Erigeron annuus for use as a trap crop by characterizing volatile profiles (Erigeron and crop plants) and utilizing these volatiles in choice tests with Lygus adults as well as with females of the Lygus natural enemy, Peristenus relictus.

Erigeron annuus and other Erigeron species are remarkably attractive to Lygus lineolaris, which is a serious pest in many crop systems (e.g. strawberries, alfalfa, and cotton). Previous studies demonstrating that L. lineolaris will leave crop plants in order to arrest and feed on Erigeron indicates that this plant shows real potential as a trap crop

(Fleischer and Gaylor 1987, Fleischer et al. 1988). Determining the mechanism of this attraction is the first step towards developing a strategy in using E. annuus as part of a trap cropping system to control Lygus in agriculturally important crop systems. It may also be possible to combine the trap crop of Erigeron annuus with a targeted release of the introduced wasp, Peristenus relictus. Lygus bugs would be drawn to the Erigeron, and then attacked by the P. relictus within the trap crop patches. Even if Erigeron plants alone are not an ideal trap crop, their presence may increase or decrease the effectiveness of P. relictus. It is important to determine the ability of P. relictus to use Erigeron volatile cues to locate hosts, as well as the ability of P. relictus wasps to successfully parasitize and develop within hosts that are actively feeding on Erigeron.

14 Chapter 4 of this thesis characterizes the volatile blend released from Erigeron annuus plants when both undamaged and when damaged by adult and nymphal Lygus bugs, and compares this blend with that released from cotton plants with similar Lygus feeding damage. Volatile collections were carried out on both undamaged and Lygus adult and nymph damaged E. annuus and cotton plants and the resulting blends were analyzed using GC and GC-MS in a similar manner to chapter 2. Our analysis found that

E. annuus produces a highly complex volatile blend consisting of over 60 distinct compounds at amounts far greater than those released from similarly damaged cotton plants. When E. annuus is fed upon by Lygus, novel compounds are induced and constitutive compounds are up-regulated. Hypothesizing that this rich blend of volatiles play a very important role in the attractiveness of Lygus to E. annuus, we ran several Y- tube bioassays with L. rubrosignatus females in order to test the attractiveness of these volatiles relative to those coming from cotton plants. We found that female Lygus bugs are highly attracted to E. annuus volatiles over those of cotton in the Y-tube, and that this attraction is not simply due to the volume of volatiles released, but appears to be based on qualitative aspects of the Erigeron blend. These results indicate that Erigeron annuus may be able to act as a trap crop by attracting and concentrating Lygus bugs into a small area – adjacent to or separate from crops – that could then be subject to application of chemical or biological controls.

Chapter 5 of this thesis examines how the host finding ability of the parasitoid wasp, Peristenus relictus, is affected by the presence of Erigeron annuus. To determine this we ran bioassays in a wind tunnel with female P. relictus wasps that previously were given experience ovipositing on Lygus nymphs in presence of volatiles from a Lygus-

15 damaged cotton plant. After receiving this oviposition experience the wasps were placed in the wind tunnel downwind from an E. annuus plant and a cotton plant. In some assays both plants were undamaged, while in others one of the two plants was being actively fed upon by Lygus nymphs. To which plant the wasps chose to fly was observed and recorded, and we found that cotton-experienced wasps overwhelmingly preferred the odors of Erigeron plants over cotton plants in most combinations, particularly Erigeron plants that had received nymph damage.

We then attempted to closely compare the individual compounds released from both E. annuus and cotton plants and found that Erigeron releases almost all of the same compounds as cotton plants, both in the undamaged and damaged condition, with emissions of shared compounds from Erigeron plants often being enhanced relative to amounts emitted from cotton. Furthermore, a large number of additional compounds not present in the cotton blends (ether undamaged or damaged) are also released from

Erigeron. Overall, this led us to believe that Erigeron annuus would serve well as a trap crop in not only pulling Lygus out of a nearby crop field, but also by attracting and retaining natural enemies of the pest.

16 References

Alborn, H.T, Turlings, T.C., Jones, T.H., Stenhagen G., Loughrin J.H., Tumlinson, J.H. 1997. An elicitor of plant volatiles from beet armyworm oral secretion. Science. 276: 945-949.

Alborn, H.T., Hansen, T.V., Jones, T.H., Bennett, D.C., Tumlinson, J.H., Schmelz, E.A., Teal, P.E. 2007. Disulfooxy fatty acids from the American bird grasshopper Schistocera americana, elicitors of plant volatiles. Proceedings of the National Academy of Sciences USA. 104:12976-12981.

Arimura, G., Ozwa, R., Shimoda, T., Nishioka, T., Boland, W., Takabayashi, J. 2000. Herbivory-induced volatile elicit defence genes in lima bean leaves. Nature. 406:512- 515.

Agrawal, A.A. 2000. Specificity of induced resistance in wild radish: causes and consequences for two specialist and two generalist caterpillars. Oikos. 89:493-500.

Barbosa, P., Gross, P., Kemper, J. 1991. Influence of plant allelochemicals on the tobacco hornworm and its parasitoid Cotesia congregata. Ecology. 72:567-1575.

Blackmer, J.L., Rodriguez-Saona, C., Byers, J.A., Shope, K.L., Smith, J.P. 2004. Behavioral response of Lygus hesperus to conspecifics and headspace volatiles of alfalfa in a Y-tube olfactometer. J. Chem. Ecol. 30:1547-1564.

Bernasconi, M.L., Turlings, T.C.J., Ambrosetti, L., Bassetti, P., Dorn, S. 1998 Herbivore induced emissions of maize volatiles repel the corn leaf aphid, Rhopalosiphum maidis. Entomologia Experimentalis et Applicata. 87:133-142.

Cardoza, Y.J., Teal, P.E.A., Tumlinson, J.H. 2003. Effect of peanut plant fungal infection on oviposition preference by Spodoptera exigua and on host-searching behavior by Cotesia marginiventris. Environmental Entomology. 32:970-976.

Carignan, S., Stewart, R.K., Godin, C., Bolvin, G. 2007. Parasitism activity of Peristenus spp. (Hymenoptera: Braconidae) on Lygus lineolaris (Hemiptera: Miridae) nymphs prior to the establishment of Peristenus digoneutis in southwestern in Quebec. Biocontrol Science and Technology. 17:623-633.

Clancy, D.W., Pierce H.D. 1966. Natural Enemies of Some Lygus Bugs. J. Econ. Entomol. 59: 853-858.

Cook, S.M., Khan, Z.R., Pickett, J.A. 2007. The use of push-pull strategies in integrated pest management. Annu. Rev. Entomol. 52:375-400.

17 Day, W.H., Hedlund, R.C., Saunders, L.B., Coutinot, D. 1990. Establishment of Peristenus digoneutis (Hymenoptera: Braconidae), a parasite of the Tarnished Plant Bug (Hemiptera: Miridae), in the United States. Ann. Entomol. Soc. Am. 85:482-488.

De Boer, J., Snoeren, T., Dicke, M. 2005. Predatory mites learn to discriminate between plant volatiles induced by prey and nonprey herbivores. Animal Behaviour. 69:869-879. de la Paz Celorio-Mancera, M., Allen, M.L., Powell, A.L., Ahmadi, H., Salemi, M.R., Phinney, B.S., Shackel, K.A., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2008. Polgalacturonase causes lygus-like damage on plants: cloning and identification of western tarnished plant bug (Lygus hesperus) polygalacturonases secreted during feeding. Arthropod-Plant Interactions. 2:215-225.

De Moraes, C.M., Lewis, W.J., Paré, P.W., Tumlinson, J.H. 1998. Herbivore infested plants selectively attract parasitoids. Nature. 393:570-574.

De Moraes, C.M., Mescher, M.C., Tumlinson, J.H. 2001. Caterpillar-induced nocturnal plant volatiles repel conspecifics females. Nature. 410:577-580.

Dicke, M. 2000. Chemical ecology of host-plant selection by herbivorous arthropods: a multitrophic perspective. Biochemical systematics and ecology 28:601-617

Dicke, M., van Loon, J.J.A. 2000 Multitrophic effects of herbivore-induced plant volatiles in an evolutionary context. Entomologia Experimentalis et Applicata. 97:237- 249

D’Ovidio, R., Mattei, B., Roberti, S., Bellincampi, D. 2004. Polygalacturonases, polygalacturonase-inhibiting proteins and pectic oligomers in plant-pathogen interactions. Biochimica et Biophysica Acta. 1696:237-244.

Engelberth, J., Alborn, H.T., Schmelz, E.A., Tumlinson, J.H. 2004. Airborne signals prime plants against insect herbivore attack. Proceedings of the National Academy of Science USA. 101:1781-1785.

Farmer, E.E., Ryan, C.A. 1990. Interplant communication: Airborne methyl jasmonate induced synthesis of proteinase inhibitors in plant leaves. Proceedings of the National Academy of Science USA. 87:7713-7716.

Federici, L., Di Matteo, A., Fernandez-Recio, J., Tsernoglou, D., Cervone, F. 2006. Polygalacturonase inhibiting proteins: players in plant innate immunity? TRENDS in Plant Science. 11:65-70.

Fleischer, S.J., Gaylor, M.J. 1987. Seasonal abundance of Lygus lineolaris (Heteroptera: Miridae) and selected predators in early season uncultivated hosts: Implications for managing movement into cotton. Environ. Entomol. 16:379-389.

18

Fleischer, S.J., Gaylor, M.J., Hue, N.V. 1988. Dispersal of Lygus lineolaris (Heteroptera: Miridae) adults through cotton following nursery host destruction. Environ. Entomol. 17:533-541.

Frati, F., Galletti, R., De Lorenzo, G., Salerno, G., Conti, E. 2006. Activity of endo- polygalacturonases in mired bugs (Heteroptera: Miridae) and their inhibition by plant cell wall proteins (PGIPs). European Journal of Entomology 103:515-522.

Frost, C.J., Appel, H.M., Carlson, J.E., De Moraes, C.M., Mescher, M.C., Schultz, J.C. 2007. Within-plant signaling via volatiles overcomes vascular constraints on systemic signaling and primes responses against herbivores. Ecology Letters. 10:490-498.

Gatehouse, J.A. 2002. Plant resistance towards insect herbivores: a dynamic interaction. New Phytologist. 156:145-169.

Goeden, R.D., and Ricker, D.W. 1974. The phytophagous insect fauna of the ragweed, Ambrosia chamissonis, in Southern California. Environ. Entomol. 3:835-839.

Handley, D. T., and J. E. Pollard. 1993. Microscopic examination of tarnished plant bug (Heteroptera: Miridae) feeding damage to strawberry. J. Econ. Entomol. 86:505–510.

Heil, M., Silva Bueno, J.C. 2007. Within-plant signaling by volatiles leads to induction and priming of an indirect plant defense in nature. Proceedings of the National Academy of Science USA. 104:5467-5472.

Hoelmer, K., Fuester, R., Taylor, P., Coutinot, D., Dorsey, T., Hudson, W., Robbins, G., Peiffer, R., Barczewski, R. 2008. Introduction of a European parasitoid, Peristenus relictus (=P. stygicus), against tarnished plant bug in the mid-Atlantic states. Journal of Insect Science. 8:49.

Ide, T, Suzuki, N, Katayama, N. 2007. The use of honeydew in foraging for aphids by larvae of the ladybird beetle, Coccinella septempunctata L. (Coleoptera: Coccinellidae). Ecological Entomology. 32:455-460.

Janssen, A., Bruin, J., Jacobs, G., Schraag, R., Sabelis, M.W. 1997. Predators use volatiles to avoid prey patches with conspecifics. The Journal of Animal Ecology. 66:223-232.

Kelton, L.A. “The Plant Bugs of the Prairie Provinces of Canada.” The Insects and Arachnids of Canada: Part 8. Hull: Canadian Government Publishing Centre. 1980. 118 & 133. Print.

Kessler, A., Baldwin, I.T. 2001. Defensive function of herbivore-induced plant volatile emissions in nature. Science. 291:2141-2144.

19

Kessmann, H., Staub, T., Hofmann, C., Maetzke, T., Herzog, J., Ward, E., Uknes, S., Ryals, J. 1994. Induction of systemic acquired disease resistance in plants by chemicals. Annual Review of Phytopathology 32:439-459.

Khan, Z.R., Pickett, J.A., van den Berg, J., Wadhams, L.J., Woodcock, C.M. 2000. Exploiting chemical ecology and species diversity: stem borer and striga control for maize and sorghum in Africa. Pest Management Science. 56:957-962.

Koch T., Krumm T., Jung V., Engelberth J., Boland W. 1999. Differential induction of plant volatile biosynthesis in the lima bean by early and late intermediates of the octadecanoid-signaling pathway. Plant Physiology. 121:153-162.

Layton, M.B. 2000. Biology and Damage of the Tarnished Plant Bug, Lygus lineolaris, In Cotton”. Southwest. Entomol. Suppl. 23:7-20.

Leigh, T., Matthews, G.A. “Lygus (Hemiptera: Miridae) and Other Hemiptera.” Insect Pests of Cotton. Ed. G.A. Matthews and J.P. Tunstall. Oxfordshire, UK: CAB International, 1996. 367-379.

Lewis, W.J., Takasu, K. 1990.Use of learned odours by a parasitic wasp in accordance with host and food needs. Nature. 348: 635-636.

Mason, P.G., Broadbent, A.B., Whistlecraft, J.W., Gillespie, D.R. 2011. Interpreting the host range of Peristenus digoneutis and Peristenus relictus (Hymenoptera: Braconidae) biological control agents of Lygus spp. (Hemiptera: Miridae) in North America. Biological Control. 57:94-102.

Matos, B. and Obrycki, J.J. 2004. Abundance and parasitism of Lygus lineolaris in alfalfa and strawberry fields. Journal of the Kansas Entomological Society. 77:69-79.

Matsui, K. 2006. Green leaf volatiles: hydroperoxide lyase pathway of oxylipin metabolism. Current Opinions in Plant Biology. 9:274-280.

Mattiacci, L., Dicke, M., Postumus, M.A. 1995. β-Glucosidase: an elicitor of herbivore- induced plant odor that attracts host-searching parasitic wasps. Proceedings of the National Academy of Sciences USA. 92:2036-2040.

McCormick, A.C., Unsicker, S.B., Gershenzon, J. 2012. The specificity of herbivore- induced plant volatiles in attracting herbivore enemies. Trends in Plant Science 17:303- 310

Papaj, D.R., Snellen, H., Swaans, K., Vet, L.E.M. 1994. Unrewarding experiences and their effect on foraging in the parasitic wasp Leptopilina heterotoma (Hymenoptera: Eucoilidae). Journal of Insect Behavior. 7:465-481.

20

Pare, P.W., Tumlinson J.H. 1997. De novo biosynthesis of volatiles induced by insect herbivory in cotton plants. Plant Physiology. 114:1161-1167.

Pare, P.W., Tumlinson, J.H. 1999. Plant volatiles as a defense against insect herbivores. Plant Physiology. 121:325-331.

Pickett, J.A., Wadhams, L.J., Woodcock, C.M. 1997. Developing sustainable pest control from chemical ecology. Agriculture, Ecosystems and Environment. 64:149-156.

Pickett, C.H., Rodriguez, R., Brown, J., Coutinot, D., Hoelmer, K.A., Kuhlmann, U., Goulet, H., Schwartz, M.D., Goodell, P.B. 2007. Establishment of Peristenus digoneutis and P. relictus (Hymenoptera: Braconidae) in California for the control of Lygus spp. (Heteroptera: Miridae). Biocontrol Science and Technology. 17:261-272.

Raffa, K. 2001. Mixed messages across multiple trophic levels: the ecology of bark beetle chemical communication systems. Chemoecology. 11:49-65.

Rasmann, S., Köllner, T.G., Degenhardt, D., Hiltpold, I., Toepfer, S., Kuhlmann, U., Gershenzon, J., Turlings, T.C.J. 2005. Recruitment of entomopathogenic nematodes by insect-damaged maize roots. Nature. 434:732-737.

Rodriguez-Saona, C., Crafts-Brandner, S.J., Williams III, L., Paré, P.W. 2002. Lygus hesperus feeding and salivary gland extracts induce volatile emissions in plants. Journal of Chemical Ecology. 28:1733-1747.

Shackel, K.A., Celorio-Mancera, M.P., Ahmadia, H., Greve, L.C., Teuber, L.R., Backus, E.A., Labavitch, J.M. 2005. Micro-injection of Lygus salivary gland proteins to simulate feeding damage in alfalfa and cotton flowers. Archives of Insect Biochemistry and Physiology. 58:69-83.

Schmelz E.A., Engelberth, J., Alborn, H.T., O’Donnell, P., Sammons, M., Toshima, H., Tumlinson, J.H. 2003. Simultaneous analysis of phytohormones, phytotoxins, and volatile organic compounds in plants. Proceedings of the National Academy of Sciences. 100:10552-10557.

Schuh, R.T. 2012. On-line Systematic Catalog of Plant Bugs (Insecta: Heteroptera: Miridae). http://research.amnh.org/pbi/catalog/

Scudder, G.G.E. “True Bugs (Heteroptera) of the Yukon.” Insects of the Yukon. Eds. H.V. Danks and J.A. Downes. Ottawa: Biological Survey of Canada, 1997. 241-336. Print.

21 Shennan, C. 2008 Biotic interactions, ecological knowledge and agriculture. Philosophical transactions of the Royal Society of London Series B, Biological Sciences. 363:717-739

Snodgrass, G.L. 1996. Pyrethroid resistance in field populations of the tarnished plant bug (Heteroptera: Miridae) in cotton in the Mississippi Delta. J. Econ. Entomol. 89:783-790.

Snodgrass, G.L., Scott, W.P. 2000. Seasonal changes in pyrethroid resistance in tarnished plant bug (Heteroptera: Miridae) populations during a three-year period in the delta area of Arkansas, Louisiana, and Mississippi. J. Econ. Entomol. 93:441-446.

Snoeren T.A.L., Van Poecke R.M.P., Dicke M. 2009. Multidisciplinary approach to unraveling the relative contribution of different oxylipins in indirect defense of Arabidopsis thaliana. Journal of Chemical Ecology 35:1021-1031.

Song, K., and Nam, Y. 2005. Genomic organization and differential expression of two polygalacturonase-inhibiting protein genes from Medicago truncatula. Journal of Plant Biology. 48(4):467-478.

Szendrei, Z., Rodriguez-Saona, C. 2010. A meta-analysis of insect pest behavioral manipulation with plant volatiles. Entomologia Experimentalis et Applicata 134:201-210

Taiz, L., Zeiger, E. “Secondary Metabolites and Plant Defense.” Plant Physiology. Sunderland, MA: Sinauer Associates Inc., 2006. 315-344.

Thaler, J.S., Stout, M.J., Karban, R., Duffey, S.S. 1996. Exogenus jasmonates simulate insect wounding in tomato plants (Lycopersicon esculentum) in the laboratory and field. Journal of Chemical Ecology. 22:1767-1781.

Thaler, J.S., Karban, R., Ullman, D.E., Boege, K., Bostock, R.M. 2002. Cross-talk between jasmonate and salicylates plant defense pathways: effects on several plant parasites. Oecologia. 131:227-235.

Thaler, J.S., Humphrey, P.T., Whiteman, N.K. 2012. Evolution of jasmonate and salicylate signal crosstalk. Trends in Plant Science. 17:260-270.

Tingey, W.M. and Pillemer, E.A. 1977. Lygus bugs: crop resistance and physiological nature of feeding injury. Bull. Entomol. Soc. Am. 23:227-287.

Turlings, T. C. J., Tumlinson, J. H., and Lewis, W. J. 1990. Exploitation of herbivore- induced plant odors by host-seeking parasitic wasps. Science. 250:1251–1253.

22 Vet, L.E.M., Lewis, W.J., Cardé, R.T. “Parasitoid Foraging and Learning.” Chemical Ecology of Insects 2. Ed. Cardé, R.T., and Bell, W.J. New York: Chapman & Hall, 1995. 65-104.

Walling, L.L. 2000. The myriad plant responses to herbivores. J. Plant Growth Regul. 19:195-216.

Will, T., Tjallingii, W.F., Thönnessen, A., van Bel, A.J.E. 2007. Molecular sabotage of plant defense by aphid saliva. Proceedings of the National Academy of Sciences. 104:10536-10541.

Young, O.P. 1986. Host plants of the tarnished plant bug, Lygus lineolaris (Heteroptera: Miridae). Ann. Entomol. Soc. Am. 79:747-762.

Zarate, S.I., Kempema, L.A., Walling L.L. 2007. Silverleaf whitefly induces salicylic acid defenses and suppresses effectual jasmonic acid defenses. Plant Physiology. 143:866-875.

Zhu, Y.C., Snodgrass, G.L., Chen, M.S. 2004. Enhanced esterase gene expression and activity in a malathion-resistant strain of the tarnished plant bug, Lygus lineolaris. Insect Biochemistry and Molecular Biology. 34:1175-1186.

Zhu, Y.C., West, S., Snodgrass, G., Luttrell, R. 2011. Variability in resistance-related enzyme activities in field population of the tarnished plant bug, Lygus lineolaris. Pesticide Biochemistry and Physiology. 99:265-273.

Chapter 2

Chemical ecology of two co-occurring Lygus species on shared host plants

Abstract

The chemical ecology of generalist and Hemipteran insects is understudied. A greater understanding of chemical cues that mediate interactions among plants, Hemipterans, and their natural enemies would contribute to a basic understanding of plant-insect interactions, and would benefit biological control efforts. Here we took a comparative approach to examine interactions among two species of generalist Lygus bugs

(Hemiptera: Miridae), and common agricultural and weedy host plants. One species,

Lygus lineolaris, has a broad range throughout the eastern U.S. and is a serious agricultural pest on alfalfa, cotton, and many fruit crops. The second species, Lygus rubrosignatus, co-occurs with L. lineolaris on many hosts, but is not considered a target of pest control measures. Volatiles were collected from Medicago sativa and Melilotus officinalis with and without the damage from Lygus bug feeding using a push-pull collection system. Volatile blends from the two species of plants both uninjured as well as those induced from feeding bythe two species of insects at different life stages were compared using Principle Component Analysis and univariate statistics. Our experiments demonstrate that both Lygus lineolaris and Lygus rubrosignatus induce distinct and characteristic volatile blends when feeding on the two different legume host plant species. Additionally, we found that nymphs of both of these Lygus species also induce volatile blends that are distinct from adult blends of conspecifics and heterospecifics.

24 These different treatments (adults and nymphs of the two species each represented on the two legume hosts) cluster with other members of that treatment in a PCA, and cluster distinctly from members of other treatments and controls. Our results indicate that Lygus salivary components may cause changes in the plant that result in elicitation of volatile synthesis and release, and that variation in these components may vary throughout the life of one species in addition to differing between species.

Introduction

Lygus bugs are Hemipteran insects belonging to the family Miridae, and polyphagous species within this genus are considered important pests of many cropsthroughout the United States. Damage to plant hosts is primarily the result of the death of tissues in and around the site of feeding and is directly caused by salivary components (Handley and Pollard 1993, Leigh and Mathews 1996, Layton 2000,

Rodriguez et. al. 2002, Shackel et. al. 2005). This type of feeding damage contrasts with the tissue removal and mechanical damage characteristic of feeding by other major crop pests classified as chewing herbivores, particularly those in the orders Lepidoptera and

Coleoptera (De Moraes et. al. 1998, Agrawal 2000, Rasmann et. al. 2005). Furthermore, chewing herbivores have been well characterized in terms of their effects on the induction of plant defenses, including volatile emissions that act as cues for natural enemies while foraging for hosts or prey (Dicke et al. 2009). For instance, De Moraes et al. (1998) found that distinct, but closely related caterpillar species induce different volatile blends in the same species of plant (tobacco), and that parasitoids can use these cues to locate the caterpillar species that is its preferred host. However, very little is

25 known about how insects with vastly different feeding strategies, such as the macerate- and-flush feeding style of Lygus bugs, elicit plant defenses at the biochemical level

(Walling 2000, de la Paz Celorio-Mancera et al. 2008). Furthermore, it is unknown whether Lygus-specific volatile responses in hosts vary by species and life stage (factors that are relevant for both bug and natural enemy foraging). In this study we address this lack of knowledge by examining the induction of indirect defenses in two species of legumes by multiple life stages of two different bugs in the genus Lygus.

Forage legumes (alfalfa and clovers) are common hosts for Lygus bugs and other

Mirids, and Lygus bugs are considered significant pests of alfalfa (Medicago sativa) being grown for seed since they feed on growing tips and reduce seed set (Kelton 1980,

Day 1999, Day et. al. 2003). Medicago fields grown for forage can also serve as reservoirs from which Lygus disperse after mowing (Day 1999, Day et. al. 2003). Due to their attractiveness as hosts for Lygus, forage legumes, particularly M. sativa, have been used as Lygus bug trap crops inter-planted within cotton fields (Sevacharian and Stern

1974), lettuce (Rämert et. al. 2001, Accinelli et. al. 2005), and late-season strawberry

(Easterbrook and Tooley 1999). Additionally, research has shown that both late-instar nymphs and females respond positively to the odors of M. sativa plants damaged by conspecifics (Blackmer et al. 2004), which suggests that volatiles play a role in the effectiveness of M. sativa as a trap crop, and that the specific blend of induced volatiles may be particularly important for mediating interactions among conspecific Lygus bugs and legume plant hosts. Therefore, in addition to filling in basic gaps in our knowledge of

Lygus effects on host biochemistry, studies that examine Lygus-induced plant volatiles in forage legumes will aid in our understanding factors mediating Lygus movement in the

26 field and how such signaling could be used to manipulate (e.g., through use of attractant compounds) Lygus movement and feeding behavior.

The induced volatiles produced in response to Lygus damage are likely the result of the unique macerate-and-flush feeding style of these insects (also called “lacerate-and- flush” in some sources), which is distinct both from the feeding style of chewing herbivores (where large areas of tissue are removed and digested within the herbivore) and the feeding style of phloem feeders, like aphids and whiteflies, where plant damage is extremely minimal and the insects ingest sap from vascular tissues without causing cell death (Walling 2000). Lygus bugs feed on forage legumes and other host plants (e.g., unopened flower buds, developing fruits, and the apical meristems of various weedy hosts) by injecting salivary enzymes into plant tissue through their piercing-sucking mouthparts and ingesting the macerated plant tissue that is broken down in the host plant as a result of these enzymes (Handley and Pollard 1993, Leigh and Mathews 1996,

Layton 2000). Research has shown that the macerate-and-flush feeding style of bugs in the family Miridae is associated with cell wall degrading enzymes called polygalacturonases (PGs). These PGs are present in particularly high amounts in Lygus bugs (Shackel et al. 2005, Frati et al. 2006). PGs degrade pectin in the cell walls of plant hosts, and in the process produce oligalacturonide fragments that have been shown to act as elicitors of plant defenses in interactions between PG-producing fungal and bacterial pathogens and their hosts (D’Ovidio et al. 2004). However, the role of insect-produced

PGs in elicitation of direct and indirect plant defenses (volatiles), and the potential differences in elicitation among bug species and life stage have not been investigated.

27 In the present study we examine the volatile emissions induced by Lygus bugs in cultivated alfalfa (Medicago sativa) and sweet clover (Melilotus officinalis), a weedy pasture legume, both of which share a similar phenology and growth habit. To determine whether induced volatiles are unique to different Lygus species we use two species of bugs in our experiments: Lygus lineolaris and Lygus rubrosignatus. Due to the wide host range of L. lineolaris (Young 1986) both of these species of bugs co-occur within the

United States and overlap in their use of host plants (Goeden and Ricker 1974, Kelton

1980, Young 1986, Scudder 1997). Although only Lygus lineolaris is commonly listed as a crop pest, it is likely that these two species are both having an impact on shared crop hosts. To further examine the specificity of induction by these two species, we look at both adult (mixed male and female groups) and immature life stages (3rd and 4th instar nymphs).

Materials and Methods

Plant Maintenance

Two plant species were used in these experiments: Medicago sativa cv. ‘charger’ and a wild legume, Melilotus officinalis (sweet clover). Medicago sativa seeds were provided by Scott Smiles, farm manager for the Russell E. Larson Agricultural Research

Farm in Rock Springs, the M. officinalis seeds were obtained from Johny’s Selected

Seeds (Waterville, MA). ( Plants were germinated from seed in Metro-Mix potting soil with 5g of Osmocote Plus slow release fertilizer (Scotts) mixed in at the time of potting in 4-inch diameter pots and maintained at 25±1°C. They were grown in a pest-free greenhouse with a light:dark regimen of 16:8 hours and a relative humidity of 50±10%.

28 Plants were bottom-watered daily with a hose. Medicago sativa and M. officinalis were allowed to grow until early budding when small buds began to form in the top leaf axils

(approximately 4 weeks) at which point they were used in experiments.

Insect Rearing and Maintenance

Lygus rubrosignatus adults and nymphs were part of a laboratory colony originally obtained from The Phillip Alampi Beneficial Insect Laboratory, West Trenton,

NJ. Lygus lineolaris bugs were collected from M. sativa and Erigeron annuus plants growing on the Penn State Agricultural Research Farm, Rock Springs, PA. The L. rubrosignatus colony was kept in a rearing room maintained at 22±2°C, 60±10% RH, and a light-dark regimen of 14:10 hours (L:D). Individuals were reared on packets of

Lygus artificial diet purchased from Bio-serv, and adult females were allowed to oviposit on similar packets full of Carrageenan Gelcarin GP812 (PhytoTechnology Laboratories,

Shawnee Mission, KS). Diet packs were changed out every other day, while oviposition packs were removed when they became full of eggs. Oviposition packs were then placed into new containers where the nymphs were allowed to hatch and grow. Lygus lineolaris bugs (colony started from locally collected individuals) were reared with a similar procedure, but with a diet supplemented by fresh green beans grown in our own greenhouse, organic broccoli florets purchased from a local grocery store and using green beans as an oviposition substrate instead of Gelcarin.

29 Volatile Collections

To quantify and identify VOCs released by the study plants, the headspace around the plants was collected using a push/pull design. Collections were run during the day over a 12-hour period starting at 9:30 and ending at 21:30. Supplemental lighting over the collection system was provided by ten high pressure metal halide light fixtures, and collections were always performed on sunny days without heavy cloud cover. While some collections were performed during winter and early spring months (see below), we did not observe an effect of slight changes in day length due to external weather conditions on the quantities of volatiles released. An automatic volatile collection system built by Analytical Research Systems (Gainesville, FL) was used to control sampling periods. The system was capable of controlling simultaneous collections from 12 individual treatment plants at any chosen set of intervals 24 hours a day. A 30cm portion of each non-excised plant was enclosed in a 7L glass bell jar with a metal guillotine-type base. Charcoal-purified air was pumped through the top of the bell jars at a rate of 4L min-1and allowed to pass over the plant before being pulled out through the bottom of the jars at a rate of 1L min-1. This air was pulled through filters containing 45mg of Super-Q adsorbent (80/100 mesh, Alltech, Deerfield, IL). Clean cotton-balls were packed around the junction of the stem and base to prevent air from lower portions of the plants or the soil from entering the bell jars.

The portion of the plants being collected from was roughly 30cm of the apical portion of early budding M. sativa and M. officinalis (stems of the two species are comparable in size and architecture). For each of the two Lygus species plants were divided into three treatments during the collections (1) control (undamaged), (2) damage

30 from Lygus adult feeding, and (3) damage from Lygus nymph feeding. For treatments involving insect damage, either 12 adult (1:1 sex ratio) or 16 nymph Lygus bugs, were placed in the bell jars an hour before collections began(4 extra nymphs were added to account for the fragility of this life stage during collection and movement). All insects remained in the bell jars and were allowed to constantly feed over the whole duration of the collections. The collections were checked several times a day and any dead insects were aspirated out and replaced with new individuals. Sample sizes are between 4-8 plants per species x damage treatment. Collections from M. sativa were carried out in the

April and June of 2009 and September of 2010. For M. officinalis collections were carried out in July and December of 2009, and February and March of 2010. These day collections were divided between 3 Super-Q filters that each collected for 4 hours. This prevented breakthrough or loss of small molecular weight compounds, and allowed for analysis of different time points if necessary. An 8-hour night collection was also run from 21:30 to 5:30 the next morning. All collections occurred within a pest-free greenhouse maintained at 25°C, 50%RH and a light:dark cycle of 16:8 hours.

Additionally, a volatile collection was carried out with Lygus bugs alone in order to identify and exclude any volatiles being produced by the insects and not the plants.

Samples were analyzed by first eluting the compounds off of each filter using

125μL of 1:1 dichloromethane:hexanes (Burdick & Jackson High Purity, and J.T. Baker

95% Purity respectively) and then adding 200ng n-octane and 400ng nonyl acetate to each eluted sample to act as internal standards. The eluted samples were then run on an

Agilent 6890 analytical GC equipped with an FID detector, with a splitless injector, and a

HP-1 column (15m x 0.25mm x 0.25μm, Agilent). In order to identify compounds,

31 selected samples were run on an Agilent 6890N GC equipped with an Agilent 5973N mass selective detector configured for electron impact mode and a HP-1MS column (30m x 0.25mm x 0.25μm, Agilent). Mass spectra were compared to spectra for standards available in the National Institute of Standards and Technology (NIST) library as well as known standards from the lab. The Kovats Index (KI) of each compound was also determined and used to identify some compounds by comparing the unknown sample KIs to those of known standards run on the same type of GC column (HP-1).

Statistical Analysis

Total volatile emissions over the daytime period were calculated for each sample by adding the amounts of each recorded compound (amounts calculated relative to the internal standard nonyl acetate based on peak area). Since the same amount of plant tissue was enclosed within the bell jars during the collections, we did not correct volatile amounts for plant weight. Compounds were then split into three categories based upon different bio-synthetic pathways: green leaf volatiles (GLVs), terpenes, and other

(aromatics from the shikimic acid pathway & unidentified compounds). The volatiles were then analyzed using a similar technique as in Delphia et. al. 2009. For each plant species (M. sativa and M. officinalis) the GLVs category and terpenes category were analyzed using principal component analysis to reduce the volatile blend from multiple variables (volatiles) to a smaller number of principal components (Minitab v. 14). Scores for each replicate included in each PCA were generated for the first two principal components, and ANOVA with post-hoc Tukey tests were performed on these scores to compare the different treatments to each other and controls (Minitab v. 14).The purpose

32 of the PCA was to determine if there are broad-scale differences among treatments for these key biosynthetic pathways capable of responding differently to different organisms.

Additional examination of individual compounds for all three categories of volatiles

(comparison of mean and standard error overlap as in Eigenbrode et al. 2002, Mauck et al. 2010) was performed to provide a more detailed analysis of compounds that contribute the most to explaining variation due to treatments observed in the PCA.

Results

Medicago sativa volatiles

The first two components of the PCA generated from the subset of volatiles identified as alcohols, aldehydes, and acetates (GLVs) collectively explained 67.6% of the variation in GLV emissions (Fig. 2.1). PC1, which explains over 50% of the variation, demonstrates that induction of GLVs by L. rubrosignatus nymphs is weak

(groups with the undamaged control treatment), but that all other treatments group separately from undamaged controls indicating induction in response to these damage treatments (PC1 scores are significant predictors of treatment effects in ANOVA, df=4,

26, F=46.77, P=0.000) (Fig. 2.1B). Lygus lineolaris adults and nymphs group together, while L. rubrosignatus adults group separately from these two treatments (Fig. 2.1B).

PC2 explains about 15% of the variation in GLV emissions and there is no significant separation of treatments along this axis (not a significant predictor of treatment effects in

ANOVA, P=0.59), although the pattern does follow that also seen along the PC1 axis. An examination of individual compounds indicates that drivers of separation between L.

33 rubrosignatus nymphs and all other treatments are KI 777, Z-3-hexen-1-ol, KI 857, and

Z-3-hexenyl acetate (Fig. 2.2). Additionally, L. rubrosignatus adults show slightly higher emissions of Z-3-hexen-1-ol and a hexyl acetate compound relative to both L. lineolaris nymphs and adults, as well as induction of compounds that are absent in one or both of the L. lineolaris treatments (KI 824, KI 883) (Fig. 2.2).

Examination of terpene emissions suggests that these compounds are overall less predictive of variation in volatile emissions among treatments, with PC1 and PC2 explaining a combined 58.2% of variation (PC1 scores are a significant predictor of treatment effects [df=4, 26, F=4.87, P=0.005], PC2 scores are not [df=4, 26, F=0.44,

P=0.782]) (Fig. 2.3B&C). Only the L. lineolaris nymph treatment is differentiated from the control plants and the other treatments according to the blend of terpene emissions

(Fig. 2.3B&C). This variation is explained by the higher emissions of E-beta ocimene and alpha farnesene by plants damaged by L. lineolaris nymphs (Fig. 2.4). Alpha farnesene in particular is unique to this treatment and does not appear in any other damage treatments or in undamaged plants (Fig. 2.4). L. lineolaris nymphs also separate significantly from all other damage treatments by their unique induction of high levels of the aromatic volatile, methyl salicylate (Fig. 2.5). L. rubrosignatus nymphs and adults also show minor differences in emissions relative to L. lineolaris nymphs and adults based on two unidentified compounds (KI 1172 and KI 1497) (Fig. 2.5).

Melilotus officinalis volatiles

The first two components of the PCA generated from the set of variables including GLVs explained 67.2% of the variation in GLV emissions (Fig. 2.6). Both PC1

34 and PC2 are significant predictors of variation when an ANOVA is run on the scores generated for each sample relative to each axis (PC1 df=4,29, F=3.09, P=0.031; PC2 df=4, 29, F=23.90, P=0.000). Treatments do not show significant separation along PC1, but differences along PC2 demonstrate that the two species of nymph treatments separates from the two species of adult treatments and from each other, while the adult treatments do not separate (Fig. 2.6C). The analysis of individual compounds indicates that species-level differences are driven largely by emissions of Z-3-hexenyl acetate, with differences between L. lineolaris nymph and adult treatments driven by the few minor compounds where adults showed stronger induction relative to nymphs (Fig. 2.7).

Additionally, hexyl acetate is induced only by adult treatments (Fig.2.7).

Terpene induction plays a large role in separating damage based on species (Fig.

2.8). The PC1 axis is the strongest predictor explaining 38.8% of variation in emissions

(ANOVA on scores for PC1 significant, df=4, 29, F=17.05, P=0.000). Along this axis all

L. lineolaris treatments separate from L. rubrosignatus treatments (Fig. 2.8B). The PC2 axis is also a significant predictor of treatment level differences (df=4, 29, F=7.19,

P=0.000) but does not explain separation by species or life stage, and mainly explains variation in the differences between each treatment and the undamaged control (Fig.

2.8C). Species level differences observed along PC1 are largely explained by the differential induction of Z-beta ocimene, KI 1419, and beta-farnesene by L. rubrosignatus nymphs and adults (and near complete absence of these compounds in L. lineolaris treatments), as well as the induction of limonene, linalool, KI 1166, and caryophyllene by

L. lineolaris nymphs and adults (with absence in L. rubrosignatus treatments) (Fig. 2.9).

Species level differences among individual compounds are also apparent for the aromatic

35 methyl salicylate, and several unknown compounds (notably KI 1497) which show induction only by adults and nymphs of one species (Fig. 2.10).

Discussion

Both species of Lygus bugs induced changes in green leaf volatiles, terpenes, and other volatile compounds when feeding on the two legumes species in this study. Insect species level differences were evident in alfalfa (Medicago sativa), where L. lineolaris adults differed from L. rubrosignatus adults, and L. lineolaris nymphs differed from L. rubrosignatus nymphs, both along the PC1 axis of the GLV analysis (Fig. 2.1B). Most insect damage treatments resulted in similar induction of terpenes and aromatic compounds in M. sativa, however, the L. lineolaris nymph treatment resulted in significantly higher induction of E-beta ocimene relative to control and adult L. lineolaris treatments, and the compounds alpha-farnesene and methyl salicylate were both only induced by L. lineolaris nymphs (Figs. 2.4 & 2.5). In contrast to species level differences being evident among GLV volatile induction in M. sativa, in sweet clover (Melilotus officinalis) species-level differences were more pronounced for terpene emissions (Figs.

2.8, 2.9). Several terpenes were only induced by members of one species, with few differences between life stages (Fig. 2.9). Additionally, methyl salicylate was only induced by L. lineolaris in both plant hosts tested, and several unknown compounds were only induced by L. rubrosignatus.

These results demonstrate that there are differences both within Lygus species

(based on life stage) and among Lygus species, and that these differences are not uniform but vary depending on the host plant being attacked. Our results for M. sativa show

36 similarities with those obtained when volatile induction due to feeding by 30-40 Lygus hesperus individuals (nymphs or adults) was quantified for pre-flowering M. sativa, most notably in the induction of terpenes (E- and Z- ocimenes, alpha-farnesene, tridecatetraene) (Blackmer et al. 2004). However, unlike the results in Blackmer et al.

2004, we also saw significant induction of GLVs by several damage treatments, across both bug species (Fig. 2.2), and saw significant differences in induction between nymphs and adults of both species (L. rubrosignatus Fig. 2.2, L. lineolaris Fig. 2.4 and Fig. 2.5).

Thus, in combination with the few previous examinations of other Lygus species feeding on M. sativa, our results demonstrate that there is likely to be significant variation in volatile blends induced by different Lygus species, and that blend composition (presence or absence of compounds) and emission levels are also likely to vary by life stage.

There are two main hypotheses to explain the variation in volatile blends induced by different Lygus species and life stages. First, there may be subtle differences in the way bugs of different species and life stages feed throughout the day. Bugs that induce higher levels of GLVs, for instance, may engage in more mechanical wounding of the plants (e.g., through increased frequency of probing). However, more often it has been found that variation in volatile emissions from herbivore-damaged plants is due to differential perception of the salivary or regurgitant components of different closely related herbivore species (Takabayashi et. al. 1995, De Moraes et. al. 1998), and this hypothesis seems more likely given what is already known about Lygus salivary components. . Lygus bugs and other bugs in the family Miridae contain one or more variants of a polygalacturonase enzyme in their saliva (Frati et. al. 2006, Allen and

Mertens 2008, de la Paz Celorio-Mancera et. al. 2008, de la Paz Celorio-Mancera et. al.

37 2009, Walker and Allen 2010), which break down different polysaccharides in the plant cell wall (e.g., homogalacturonan, the main component of pectin [Di Matteo et al. 2006]) and are produced by both insects and plant-infecting fungi (Walling 2000, Habibi et. al.

2001, D’Ovidio et. al. 2004, Frati et. al. 2006). By breaking down cell wall components,

PGs increase access to other cellular components (e.g., starch) which can be further broken down and consumed (Baptist 1941, Walling 2000, Di Matteo et. al. 2006).

Additionally, PG activity on cell walls produces a variety of oligogalacturonide acid

(OGA) fragments, which can be perceived by the plant in a similar manner as insect- derived salivary elicitors (Aziz et. al. 2004, D’Ovidio et. al. 2004, Federici et. al. 2006,

Frati et. al. 2006). Different PGs will produce different OGAs, and OGAs have been shown to activate a number of different plant defense pathways that have bearing on volatile production, most notably the octadecanoid pathway (Leitner et. al. 2008) and the phenylpropanoid pathway (Ridley et. al. 2001). Activity of these OGA elicitors can also be enhanced if the plants possess PG inhibitor proteins (PGIPs) that are capable of acting on the PGs, as this will prevent the further breakdown or consumption of OGAs, allowing them to remain intact for a longer period of time, increasing the chance of their eliciting the above-mentioned pathways (D’Ovidio et. al. 2004, Di et. al. 2006, Federici et. al.

2006).

Although PGs from Mirid bug saliva are just beginning to be explored in detail, studies have already shown that multiple PG genes can be found in Lygus lineolaris

(Allen and Mertens 2008). Additionally, a recent study showed that at least five distinct

PGs can be purified from Lygus hesperus saliva, and that they generate different breakdown products (OGAs) when tested in vitro (de la Paz Celorio-Mancera et al.

38 2009). These studies indicate that variation in the number and activity of PGs present in

Lygus saliva could be one factor involved in species-specific volatile responses in host plants, since different PGs will generate different patterns of OGA elicitors (and may be more or less susceptible to PGIPs activated by the plant in response to salivary injection).

The multi-genic nature of these enzymes suggests that within a given species, PG gene expression is likely to vary depending on the life stage, and possibly even the sex of the bug. Indeed, a preliminary study which performed a semi-quantitative PCR analysis of

PG expression in L. lineolaris did find variation in expression between life stages

(nymphs vs. adults) and adult sexes for multiple PGs (Allen and Mertens 2008). In another study, Frati et al. (2006) tested the activity of several PGIPs (derived from

Phaseolus vulgaris, Arabidopsis thaliana, and Glycine max) against the collective activity of all PGs in the saliva from L. rugulipennis and L. pratensis. Results showed that some of the PGIPs failed to inhibit the activity of bug PGs, while others inhibited PG activity, but did so to different extents depending on the species (e.g., one PGIP from P. vulgaris inhibited L. rugulipennis PG activity nearly twice as effectively as it inhibited L. pratensis PG activity) (Frati et al. 2006). Thus, the plant’s ability to inhibit PG activity

(and prolong the life of OGAs as elicitors) may also contribute to variation in volatile response to a given bug species if PGs are indeed involved in eliciting volatile responses in host plants.

In our study system, the different volatile responses observed in the two legume host plants are consistent with there being variation in PG expression by life stage, PG identity by bug species, and in the plant’s relative ability to respond to saliva injection. In

M. sativa, L. lineolaris nymphs strongly induce GLVs, terpenes, and the aromatic volatile

39 methyl salicylate. These results may be indicative of enhanced activity, quantity, or diversity of PGs and production of high levels of OGAs that are then perceived by the plant as elicitors of indirect and direct defense responses. In contrast, L. rubrosignatus nymphs do not show the same induction of the above mentioned volatiles on M. sativa, which may indicate reduced PG levels or reduced diversity of PGs. On M. officinalis, most damage treatments significantly induce both GLVs and terpenes (with terpene induction being in a species-specific manner), suggesting a fundamentally different interaction between bug feeding and M. officinalis relative to M. sativa. Furthermore, the specific induction of methyl salicylate by both nymphs and adults of L. lineolaris suggests that only this species is capable of eliciting the shikimic acid and phenylpropanoid pathways, which produce precursors of salicylic acid and methyl salicylate, as well as lignin (to reinforce cell walls as a defense mechanism) (Lee et al.

1995, Fraser and Chapple 2011). Overall, the fact that there are significant differences in production of de novo synthesized volatiles in response to bug feeding suggests the involvement of salivary components, and argues against differences in bug feeding behavior (e.g., the frequency of punctures) being entirely explanatory of the observed patterns.

These data have important implications for development of plants with resistance to Lygus bugs, as it suggests that a focus on PGIPs (e.g., through creation of transgenic plants expressing PGIPs with activity against PGs from several Lygus species) may lead to cultivars that experience reduced damage due to Lygus PGs, and which preserve the longevity of OGAs produced in response to PG activity. Preservation of OGAs can lead to induced systemic resistance and the elicitation of an enhanced volatile response as the

40 plant has an opportunity to perceive these effective elicitor molecules. Induced volatiles are important in mediating interactions between Lygus bugs, their hosts, and conspecifics, as well as between beneficial insects (e.g., Lygus parasitoids) and their hosts (Turlings et. al. 1991, Tumlinson et. al. 1993, Blackmer et al. 2004, Williams et. al. 2008). Future experiments focusing on Lygus salivary components and their effects on host plants should further elucidate the in-plant biochemical responses to damage by different bugs

(e.g., through measurement of phytohormones and their precursors), and should also explore the possibility that Lygus expression of PGs is adaptable within a given life stage, e.g., in response to feeding on a certain host plant. Besides providing insight into plant responses to Lygus attack with relevance to resistance, our results also suggest that differences between life stages may allow beneficial parasitoids to discriminate plants with susceptible hosts (nymphs) from plants with unsusceptible hosts (adults), and that differences between species may allow further discrimination. Parasitoids may also learn the volatile profiles associated with one species if they receive positive oviposition experiences in the presence of species-specific volatiles. Future experiments should further explore the implications of species and life-stage specific volatile blends for host- finding by natural enemies that show promise as biological controls for Lygus.

41 References

Accinelli, G., Lanzoni, A., Ramilli, F., Dradi, D., Burgio, G. 2005. Trap crop: an agroecological approach to the management of Lygus rugulipennis on lettuce. Bulletin of Insectology. 58:9-14.

Agrawal, A.A. 2000. Specificity of induced resistance in wild radish: causes and consequences for two specialist and two generalist caterpillars. Oikos. 89:493-500.

Allen, M., Mertens, J. 2008. Molecular cloning and expression of three polygalacturonase cDNAs from the tarnished plant bug, Lygus lineolaris. Journal of Insect Science. 8.

Aziz, A., Heyraud, A., Lambert B. 2004. Oligogalacturonide signal transduction, induction of defense-related responses and protection of grapevine against Botrytis cinerea. Planta. 218:767-774.

Baptist, B.A. 1941. The morphology and physiology of the salivary glands of Hemiptera- Heteroptera. Quarterly journal of microscopical science. 329:91-139

Blackmer, J.L., Rodriguez-Saona, C., Byers, J.A., Shope, K.L., Smith, J.P. 2004. Behavioral response of Lygus hesperus to conspecifics and headspace volatiles of alfalfa in a Y-tube olfactometer. J. Chem. Ecol. 30:1547-1564.

Day, W.H. 1999. Host preferences of introduced and native parasites (Hymenoptera: Braconidae) of phytophagous plant bugs (Hemiptera: Miridae) in alfalfa-grass fields in the northeastern USA. Biocontrol. 44:249-261.

Day, W.H., Eaton, A.T., Romig, R.F., Tilmon, K.J., Mayer, M., Dorsey, T. 2003. Peristenus digoneutis (Hymenoptera: Braconidae), a parasite of Lygus lineolaris (Hemiptera: Miridae) in northeastern United States alfalfa, and the need for research on other crops. Entomological News. 114:105-111. de la Paz Celorio-Mancera, M., Allen, M.L., Powell, A.L., Ahmadi, H., Salemi, M.R., Phinney, B.S., Shackel, K.A., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2008. Polgalacturonase causes lygus-like damage on plants: cloning and identification of western tarnished plant bug (Lygus hesperus) polygalacturonases secreted during feeding. Arthropod-Plant Interactions 2:215-225. de la Paz Celorio-Mancera, M., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2009. Identification of endo- and exo-polygalacturonase activity in Lygus hesperus (Knight) salivary glands. Archives of Insect Biochemistry and Physiology. 70:122-135.

De Moraes, C.M., Lewis, W.J., Paré, P.W., Tumlinson, J.H. 1998. Herbivore infested plants selectively attract parasitoids. Nature. 393:570-574.

42 Delphia, C.M., Rohr, J.R., Sthephenson, A.G., De Moraes, C.M., Mescher, M.C. 2009. Effects of genetic variation and inbreeding on volatile production in a field population of horsenettle. Int. J. Plant Sci. 170:12-20.

Di, C., Zhang, M., Xu, S., Cheng, T., An, L. 2006. Role of poly-galacturonase inhibiting protein in plant defense. Critical Reviews in Microbiology. 32:91-100.

Di Matteo, A., Bonivento, D., Tsernoglou, D., Federici, L., Cervone, F. 2006. Polygalacturonase-inhibiting protein (PGIP) in plant defence: a structural view. Phytochemistry. 67:528-533.

Dicke, M., van Loon, J.J.A., Soler, R. 2009. Chemical complexity of volatiles from plants induced by multiple attack. Nature. 5:317-324.

D’Ovidio, R., Mattei, B., Roberti, S., Bellincampi, D. 2004. Polygalacturonases, polygalacturonase-inhibiting proteins and pectic oligomers in plant-pathogen interactions. Biochimica et Biophysica Acta 1696:237-244.

Easterbrook, M.A., Tooley, J.A. 1999. Assessment of trap plants to regulate numbers of the European tarnished plant bug, Lygus rugulipennis, on late-season strawberries. Entomologia Experimentalis et. Applicata. 92:119-125.

Eigenbrode, S.D., Ding, H., Shiel, P., Berger, P. 2002. Volatiles from potato plants infected with potato leafroll virus attract and arrest the virus vector, Myzus persicae (Homoptera: Aphididae). Proc. R. Soc. Lond. B. 269:455-460.

Federici, L., Di Matteo, A., Fernandez-Recio, J., Tsernoglou, D., Cervone, F. 2006. Polygalacturonase inhibiting proteins: players in plant innate immunity? TRENDS in Plant Science. 11:65-70.

Frati, F., Galletti, R., De Lorenzo, G., Salerno, G., Conti, E. 2006. Activity of endo- polygalacturonases in mired bugs (Heteroptera: Miridae) and their inhibition by plant cell wall proteins (PGIPs). European Journal of Entomology 103:515-522.

Goeden, R.D., and Ricker, D.W. 1974. The phytophagous insect fauna of the ragweed, Ambrosia chamissonis, in Southern California. Environ. Entomol. 3:835-839.

Habibi, J., Backus, E.A., Coudron, T.A., Brandt, S.L. 2001. Effect of different host substrates on hemipteran salivary protein profiles. Entomologia Experimentalis et Applicata. 98:369-375.

Handley, D. T., and J. E. Pollard. 1993. Microscopic examination of tarnished plant bug (Heteroptera: Miridae) feeding damage to strawberry. J. Econ. Entomol. 86:505–510.

43 Kelton, L.A. “The Plant Bugs of the Prairie Provinces of Canada.” The Insects and Arachnids of Canada: Part 8. Hull: Canadian Government Publishing Centre. 1980. 118 & 133. Print.

Layton, M.B. 2000. Biology and Damage of the Tarnished Plant Bug, Lygus lineolaris, In Cotton. Southwest. Entomol. Suppl. 23:7-20.

Leigh, T., Matthews, G.A. “Lygus (Hemiptera: Miridae) and Other Hemiptera.” Insect Pests of Cotton. Ed. G.A. Matthews and J.P. Tunstall. Oxfordshire, UK: CAB International, 1996. 367-379. Print.

Leitner, M., Kaiser, R., Rasmussen, M.O., Driguez, H. 2008. Microbial oligosaccharides differentially induce volatiles and signaling components in Medicago truncatula. Phytochemistry. 69:2029-2040.

Mauck, K.E., De Moraes, C.M., Mescher, M.C. 2010. Deceptive chemical signals induced by a plant virus attract insect vectors to inferior hosts. Proceedings of the National Academy of Science. 107:3600-3605.

Rämert, B., Hellqvist, S., Ekbom, B., Banks, J.E. 2001. Assessment of trap crops for Lygus spp. in lettuce. International Journal of Pest Management. 47:273-276.

Rasmann, S., Köllner, T.G., Degenhardt, D., Hiltpold, I., Toepfer, S., Kuhlmann, U., Gershenzon, J., Turlings, T.C.J. 2005. Recruitment of entomopathogenic nematodes by insect-damaged maize roots. Nature. 434:732-737.

Ridley, B.L., O’Neill, M.A., Mohnen, D. 2001. Pectins: structure, biosynthesis, and oligogalacturonide-related signaling. Phytochemistry. 57:929-967.

Rodriguez-Saona, C., Crafts-Brandner, S.J., Williams III, L., Paré, P.W. 2002. Lygus hesperus feeding and salivary gland extracts induce volatile emissions in plants. J. Chem. Ecol. 28:1733-1747.

Scudder, G.G.E. “True Bugs (Heteroptera) of the Yukon.” Insects of the Yukon. Eds. H.V. Danks and J.A. Downes. Ottawa: Biological Survey of Canada, 1997. 241-336. Print.

Shackel, K.A., Celorio-Mancera, M.P., Ahmadia, H., Greve, L.C., Teuber, L.R., Backus, E.A., Labavitch, J.M. 2005. Micro-injection of Lygus salivary gland proteins to simulate feeding damage in alfalfa and cotton flowers. Archives of Insect Biochemistry and Physiology. 58:69-83.

Sevacherian, V., Stern, V.M. 1974. Host plant preferences of Lygus bugs in alfalfa- interplanted cotton fields. Environ. Entomol. 3:761-766.

44 Takabayashi, J., Takahashi, S., Dicke, M., Posthumus. 1995. Developmental stage of herbivore Pseudaletia separata affects production of herbivore-induced synomone by corn plants. Journal of Chemical Ecology. 21:273-287.

Tumlinson, J.H., Lewis, W.J., and Vet, L.E.M. 1993. How parasitic wasps find their hosts. Sci. Am. 268:100–106.

Turlings, T.C.J., Tumlinson, J.H., Heath, R.R., Proveaux A.T., and Doolittle, R.E. 1991. Isolation and identification of allelochemicals that attract the larval parasitoid, Cotesia marginiventris (Cresson), to the microhabitat of one of its hosts. J.Chem. Ecol. 17:2235– 2251.

Walker, W., and Allen, M. 2010. Expression and RNA interference of salivary polygalacturonase genes in the tarnished plant bug, Lygus lineolaris. Journal of Insect Science. 10.

Walling, L.L. 2000. The myriad plant responses to herbivores. J. Plant Growth Regul. 19:195-216.

Williams, L., Rodriguez-Saona, C., Castle, S.C., Zhu, S. 2008. EAG-Active herbivore- induced plant volatiles modify behavioral responses and host attack by an egg parasitoid. J. Chem. Ecol. 34:1190-1201.

Young, O.P. 1986. Host plants of the tarnished plant bug, Lygus lineolaris (Heteroptera: Miridae). Annals of the Entomological Society of America. 79:747-762.

45

Figure 2.1: Medicago sativa green leaf volatiles principal component analysis. [A]Principal component analysis (PCA) of compounds identified as alcohols, aldehydes, and acetates. Mean scores for principal component 1 [B] and principal component 2 [C] generated from the PCA. Letters beside bars indicate significant differences among treatments (P<0.05) by Tukey Test. Percentage values in the Y-axis label indicate the percent of variation explained by each component.

46

Figure 2.2: Medicago sativa green leaf volatiles individual compounds. Mean and standard error values for each individual compound that contributed to the PCA, by each treatment. Red boxes highlight compounds that are induced in all damage treatments relative to controls, but that differ among two or more damage treatments (separation of >2 standard errors), and black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

47

Figure 2.3: Medicago sativa terpene principal component analysis. [A]Principal component analysis (PCA) of compounds identified as mono-, sesqui-, and homo-terpenes. Mean scores for principal component 1 [B] and principal component 2 [C] generated from the PCA. Letters beside bars indicate significant differences among treatments (P<0.05) by Tukey Test. Percentage values in the Y-axis label indicate the percent of variation explained by each component.

48

Figure 2.4: Medicago sativa terpene individual compounds. Mean and standard error values for each individual compound that contributed to the PCA, by each treatment. Red boxes highlight compounds that are induced in all damage treatments relative to controls, but that differ among two or more damage treatments (separation of >2 standard errors), and black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

49

Figure 2.5: Medicago sativa aromatic volatiles and unidentified volatiles analysis. Mean and standard error values for individual compounds that are aromatics (named compounds) or unknowns that could not be identified. Black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

50

Figure 2.6: Melilotus officinalis green leaf volatiles principal component analysis. [A]Principal component analysis (PCA) of compounds identified as alcohols, aldehydes, and acetates. Mean scores for principal component 1 [B] and principal component 2 [C] generated from the PCA. Letters beside bars indicate significant differences among treatments (P<0.05) by Tukey Test. Percentage values in the Y-axis label indicate the percent of variation explained by each component.

51

Figure 2.7: Melilotus officinalis green leaf volatiles individual compounds. Mean and standard error values for each individual compound that contributed to the PCA, by each treatment. Red boxes highlight compounds that are induced in all damage treatments relative to controls, but that differ among two or more damage treatments (separation of >2 standard errors), and black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

52

Figure 2.8: Melilotus officinalis terpene principal component analysis. [A]Principal component analysis (PCA) of compounds identified as mono-, sesqui-, and homo-terpenes. Mean scores for principal component 1 [B] and principal component 2 [C] generated from the PCA. Letters beside bars indicate significant differences among treatments (P<0.05) by Tukey Test. Percentage values in the Y-axis label indicate the percent of variation explained by each component.

53

Figure 2.9: Melilotus officinalis terpene individual compounds. Mean and standard error values for each individual compound that contributed to the PCA, by each treatment (legend in graph A). Black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

54

Figure 2.10: Melilotus officinalis aromatic volatiles and unidentified volatiles analysis. Mean and standard error values for individual compounds that are aromatics (named compounds) or unknowns that could not be identified. Black boxes indicate compounds that are induced by one or more damage treatments relative to controls (>2 standard errors relative to control or vs. absence in the control), but which are absent in one or more other damage treatments.

Chapter 3

Induction of phytohormones in response to feeding by two species of Lygus bugs on alfalfa and sweet clover

Abstract

Defensive responses of plants to feeding by chewing insect herbivores have been extensively studied, but far less research has been carried out examining how plants respond to other types of insect herbivory. While chewing herbivores most often induce jasmonic acid (JA), recent research suggests that phloem feeding insects may instead induce production of the phytohormone salicylic acid (SA) and other pathogenesis defense genes, which may in turn cause the down-regulation of JA and allow for the herbivores to feed without having to overcome the plant defenses activated by JA.

However, whether this pattern holds true for other piercing-sucking insects, such as bugs in the family Miridae, that cause an intermediate amount of damage between chewing and phloem feeding herbivores is not yet known. In this study we explored defense patterns in response to this type of insect by measuring the phytohormone induction response of two legume species Medicago sativa (alfalfa) and Melilotus officinalis (sweet clover) to feeding by the adult and nymphal life-stages of two Lygus bug species, L. lineolaris and L. rubrosignatus (Hemiptera: Miridae). Previous research has shown that both bug species and life stage can have an effect on indirect plant defenses (volatile organic compounds) during feeding. In this study we measured indicators of direct defenses, namely the phytohormones jasmonic acid, salicylic acid, auxin, abscisic acid

56 (ABA), linolenic acid (a precursor for JA and some volatiles), and the SA precursor cinnamic acid. We found that in both plant species feeding by Lygus bugs induced an increase in the production of both JA and SA at the site of feeding damage, as well as an increase in the amount of linolenic acid, while there was little change in the production of these compounds at systemic sites. Bug species also modulated the levels of induction of

JA and SA, as well as induction of ABA (a hormone associated with drought stress responses). These results indicate that bug feeding elicits responses characteristic of both pathogen attack (SA) and chewing herbivore attack (JA), and that components of the bug saliva differ between the two insect species tested. However, we did not see a clear relationship between species and life-stage specific induction of phytohormones and the previously characterized induction of volatile emissions for these same bug-plant combinations, so bug feeding may be having effects on other pathways, or modulating down-stream expression of relevant defense genes.

Introduction

Plants possess complex defense mechanisms that function to preserve plant tissue and overall fitness in the face of a multitude of biotic attackers and abiotic stressors.

Many plants have a vast array of constitutive (e.g. terpenes, phenolics, and nitrogen- containing secondary metabolites) and inducible (e.g. α-amylase inhibitors, lectin, and proteinase inhibitors) in-plant defenses (Taiz and Zeiger 2006). However, plants also defend themselves by sending out long-distance signals, via volatile organic compounds

(VOCs) that communicate information about plant status to other organisms, such as natural enemies of the herbivores (Turlings and Tumlinson 1992, Turlings et. al. 1993).

57 These “indirect defenses” can have important fitness effects for the plant if they help to deter herbivores (e.g., by suggesting the plant is well defended, or that there are competitors present) (De Moraes et. al. 1998) or if they effectively recruit predators and parasitoids that control herbivores already present on the plant (Kessler and Baldwin

2001). Given the nature of the information being conveyed by VOCs, and especially induced VOCs, it is necessary that these signals be a fairly specific response to a given antagonist. Several studies suggest that VOC induction does vary depending on the attacking organism, and that the consumers of VOC signals can discriminate amongst

VOCs induced by different attackers (De Moraes et. al. 1998, Rodriguez-Saona et. al.

2003, De Boer et. al. 2005, Dicke et. al. 2009). These results suggest that along with the attacker-specific variation in VOC blends, there is corresponding variation in within- plant changes in the regulators of volatile emissions (e.g., phytohormones and their precursors).

Recently, we documented significant differences in the induced volatile blends emitted by two species of legume, Medicago sativa and Melilotus officinalis, in response to feeding by different life stages (nymph vs. adult) of two species of Mirid bugs in the genus Lygus (L. rubrosignatus and L. lineolaris) (Halloran et al. 2012). These differences suggest that these life stages and species of bugs may differ in identity or activity of enzyme components of the saliva that they inject into plants during feeding in order to perform pre-digestion of tissue prior to consumption via a sucking mechanism (Frati et. al. 2006, Allen and Mertens 2008, de la Paz Celorio-Mancera et. al. 2008, de la Paz

Celorio-Mancera et. al. 2009, Walker and Allen 2010). Other studies have also found that purified enzymes involved in plant cell wall degradation (e.g., cellulases) can influence

58 volatile emissions (Piel et al. 1997, Koch et al. 1999) and that unique protein and non- protein insect salivary or regurgitant components induce specific volatile responses in plants that are distinct from mechanical wounding alone (Mattiacci et. al. 1995, Alborn et. al. 1997, Koch et al. 1999, Alborn et. al. 2007). Studies examining other plant bug species have also shown that there are species-level differences in the activities and identities of key cell wall degrading enzymes called polygalacturonases (PGs) that are present in the saliva of bugs in the family Miridae (Allen and Mertens 2008, de la Paz

Celorio-Mancera et al. 2009), which are thought to be the main cause of economic damage to crops by these species (wilting, stunting, and abscission of damaged tissue)

(Strong 1970, Shackel et al. 2005).

PGs break down pectin into oligogalacuronide (OG) fragments, with different

PGs producing fragments of different lengths (D’Ovidio et. al. 2004). PGs are common components of the cocktails of enzymes employed by necrotrophic fungal pathogens to invade plant hosts via degradation of cell walls, and studies aimed at these fungal PGs have shown that plants respond to PG activity with enzyme inhibitors called polygalacturonase inhibiting proteins (PGIPs) (D’Ovidio et. al. 2004, Di et. al. 2006,

Federici et. al. 2006). PGIPs both inhibit the activity of PGs and prevent the complete breakdown of OGs. This second effect of PGIP mobilization is relevant to subsequent plant defense responses to PG presence, since OG fragments with longer polymerization

(e.g., 9-16 units, Mathieu et al. 1991) are perceived as elicitors by many plants, and the longer they persist the more effectively the plant can respond to the attacker (Ridley et al.

2001). These OG elicitors can initiate induction of the phenylpropanoid pathway, leading to the production of the phytohormone signaling molecule, salicylic acid, salicylate-

59 derived volatiles, and lignification of cell walls and production of secondary metabolites

(reviewed in D’Ovidio et al. 2004 and Ridley et al. 2001). OG elicitors can also stimulate the production of jasmonic acid via the octadecanoid pathway (which potentiates the production of a number of volatile organic terpenes) (Ridley et al. 2001). Thus, variation in PG activity and identity among plant bugs is likely to lead to differential induction of these defense responses, and subsequently, differential emission of phytohormone- regulated VOCs.

Here we investigate basic induced defenses in response to Lygus bug feeding, as well as the mechanisms underlying observed differences in volatile emissions in M. sativa and Melilotus officinalis in response to attack by nymphal and adult life stages of

L. rubrosignatus and L. lineolaris (Halloran et al. 2012). We examined induction of several phytohormones that give insight into the pathways induced by Lygus feeding (and thus plant exposure to PGs). To examine induction of the phenylpropanoid pathway we examined induction of the pathway intermediate, cinnamic acid, and the key signaling hormone, salicylic acid (Taiz and Zeiger 2006). To examine induction of the octadecanoid pathway, we measured levels of the first compound used in this pathway

(linolenic acid) and the end product of this pathway (the signaling hormone, jasmonic acid) (Taiz and Zeiger 2006). Because damage to tissues by Lygus bugs is stressful and disruption of cell walls also has the potential to disrupt turgor pressure (inducing a type of water stress), we also measured levels of abscisic acid. These regulators of plant defense were measured in tissue that had been directly damaged by Lygus bugs of each species and life stage, as well as in a systemic leaf section that was adjacent to the damage site to determine if signals were being translocated to other areas of the plant.

60 Materials and Methods

Plant Maintenance

The two plant species used in these experiments were Medicago sativa cv.

‘Charger’ and the wild legume Melilotus officinalis (sweet clover). Medicago sativa seeds were provided by Scott Smiles, farm manager for the Russell E. Larson

Agricultural Research Farm in Rock Springs, the M. officinalis seeds were obtained from

Johny’s Selected Seeds (Waterville, MA). All seeds were grown in Metro-Mix potting soil with 5g of Osmocote Plus slow release fertilizer (Scotts) mixed in at the time of potting in 4-inch diameter pots and maintained at 25±1°C. They were grown in a pest- free greenhouse with a light:dark regimen of 16:8 hours and a relative humidity of

50±10%. Plants were bottom-watered daily with a hose. Medicago sativa and M. officinalis were allowed to grow until early budding when small buds began to form in the top leaf axils (approximately 4 weeks) at which point they were used in experiments.

Insect Rearing and Maintenance

Two species of Lygus bugs were used in these experiments, Lygus rubrosignatus and L. lineolaris. Lygus rubrosignatus adults and nymphs were part of a laboratory colony originally obtained from The Phillip Alampi Beneficial Insect Laboratory, West

Trenton, NJ. Lygus lineolaris bugs were collected from M. sativa and E. annuus plants growing on the Penn State Agricultural Research Farm, Rock Springs, PA. The L. rubrosignatus colony was kept in a rearing room maintained at 22±2°C, 60±10% RH, and a light-dark regimen of 14:10 hours (L:D). Individuals were reared on packets of

Lygus artificial diet purchased from Bio-serv, while adult females were allowed to

61 oviposit on similar packets filled with Gelcarin GP812 (PhytoTechnology Laboratories,

Shawnee Mission, KS). Diet packs were changed out every other day, while oviposition packs were removed when they became full of eggs. Oviposition packs were then placed into new containers where the nymphs were allowed to hatch and grow. Lygus lineolaris bugs (colony started from locally collected individuals) were reared with a similar procedure, but with a diet supplemented by green beans and broccoli, and using green beans as an oviposition substrate instead of Gelcarin.

Measurement of Phytohormones

All of the experiments were carried out in an insect-free greenhouse with supplemental lighting on a 16:8 light-dark photoperiod. On the first day of the experiments, the top leaf axils of four-week-old Medicago sativa and M. officinalis plants were covered with a small clip cage and damage treatments were applied to the plants.

Control plants were left with an empty clip cage, while plants in the damaged treatment had either four Lygus adults (1:1 sex ratio) or four 3rd instar nymphs added into the clip cages. A damage treatment for each Lygus species and life stage were applied to both of the plant species, making four damage treatments and one control treatment for each plant species. The clip cages and Lygus bugs were added to the plants at 3:00 p.m.

The Lygus bugs were allowed to feed within the clip cages for 24 hours before both the bugs and the clip cages were removed from the plants. Immediately after, the

Lygus-damaged portion of each plant (1-2 trifoliates at the apical tip) was excised, weighed, and immediately flash frozen in liquid nitrogen. Directly after collecting this portion of the plant, the next trifoliate down the shoot (un-exposed to damage but

62 systemically connected to the apical tip) was excised, weighed, and frozen in a separate tube. Experiments with M. sativa were performed across two blocks the first being carried out in Feburary of 2012 and the second in August 2012. Within each block there were 5 replicate plants for each damage treatment. For M. officinalis there was single experiment, carried out in September 2012, with 8 replicate plants for each damage treatment.

All tissue samples were processed and analyzed according to Schmelz et. al.

(2003) using gas chromatography and mass spectrometry. The protocol, in brief, consisted of adding an internal isotopic or isomeric standard to each tissue sample (100 ng of each standard), and then grinding the samples in a propanol/water buffer with

50mM HCL using a Fast-Prep machine. The standards used were dihydro-jasmonic acid,

2 2 [ H6] salicylic acid, gamma-linolenic acid, and [ H6] absiscic acid. Dichloromethane was then immediately added to the samples and they were shaken again in the Fast Prep machine, and then centrifuged at 11000 G for 1 minute in order to pellet the leaf material and separate the dichloromethane and aqueous solvents. The dichloromethane layer

(containing compounds of interest) was then pipetted to a new glass vial and dried. The samples were then derivatized by adding a mixture of 1:9 methanol:diethyl ether solvent and 2.3μL of 2.0M trimethylsilyldiazomethane to the dried residue and allowing them to incubate in the dark at roughly 21°C for 24 minutes. This allowed for the phytohormones to be derivatized into methyl esters. The reaction was then stopped by adding to each sample 2.3μL of a 12:88 mixture of glacial acetic acid:hexanes and then allowing them to incubate as before. Again the solvents were evaporated off and methylated compounds were then collected using vapor-phase extraction, which consisted of heating the vials to

63 200°C and pulling air at 1L/min through a filter containing 40mg of Super-Q adsorbent that was inserted into the vials.

Compounds were then eluted off of these filters using dichloromethane and analyzed on an Agilent 7890 gas chromatograph (GC) fitted with a mass spectrometer

(MS) set to chemical ionization mode (using isobutane gas). The GC-MS was fitted with a HP-5 column with dimensions of 30m x 250μm x 0.1μm with a carrier gas of helium and a flow level of 0.927 mL/min. An injection volume of 2 μL was injected into a splitless injector heated to 250°C and an oven program was run starting at 40°C for 1 minute and then ramping up 15°C/min to 300°C where it held for 7 additional minutes allowing for a total run time of just over 25 minutes.

For this protocol, we included standards that permitted the measurement of jasmonic acid (JA), salicylic acid (SA), auxin, cinnamic acid, fatty acids (linolenic [LNA] and linoleic acids), and abscisic acid (ABA). However, the only phytohormones present in the plants, out of this list, were JA, SA, ABA, and linolenic acid (linoleic acid was also detected, but we focused on linolenic acid since it is the primary precursor for JA biosynthesis). Peak areas for each compound and its synthetic standard were quantified and the amount of each compound was calculated based on the quantity of standard added (100ng) and the fresh weight of the tissue determined at collection. A defect in the isotopic standard for ABA resulted in no methylation of this compound, so endogenous

ABA was instead quantified according to the SA isotopic standard in order to correct for slight differences in methylation efficiency between samples.

64 Statistical analysis

We analyzed differences among treatments in induction of JA, SA, ABA, and the precursor for the octadecanoid pathway in legumes, the fatty acid LNA. First, we confirmed that induction occurred for each phytohormone separately, within each tissue type (feeding site and systemic site) and each plant species, by performing ANOVA with treatment as the main factor (control, L. lineolaris adult-damaged, L. lineolaris nymph- damaged, L. rubrosignatus adult damaged and L. rubrosignatus nymph- damaged).Melitous officinalis was analyzed separately from M. sativa and we did not include a block effect since only one experiment was used for this species. Block was included in the model for all analyses with Medicago sativa and was never significant.

Within this ANOVA, we also performed Dunnett’s test of comparison with a control, to determine whether the means of each damage treatment differed from the control

(indicating induction or suppression). If at least one treatment differed from the control, we then performed a factorial ANOVA which excluded the control treatment, allowing us to include species, life stage, and the interaction term of species by life stage as factors in the model. Tukey’s test was performed for the species by life stage interaction in order to identify significant differences among the four damage treatments. All analyses were performed using Minitab v. 14. Where necessary, data were log or square root transformed prior to analysis in order to give the data a more normal distribution. Box plots were also used to identify any outliers in the data set prior to analysis.

65 Results

Across all plant species, we detected measurable levels of SA, JA, ABA, and linolenic acid. Cinnamic acid and auxin were not detected in any treatment. At the site of feeding damage on Medicago sativa the amount of JA induced was significantly different from the control plants for each of the Lygus damage treatments (Fig. 3.1A). In the factorial ANOVA, significant differences were seen on the ‘species’ level (df=1,48,

F=23.20, p=0.00), but there was no significant effect of ‘life stage’ or the interaction of

‘species’ and ‘life stage’ (Fig. 1A). The induction of JA was higher for the combined mean of both Lygus rubrosignatus damaged treatments relative to the mean of the Lygus lineolaris treatments. All damage treatments also showed a significantly higher induction of SA relative to the controls, but there were no significant differences in the amount of

SA between the different damage treatments (Fig. 3.1B). Measuring linolenic acid, all damage treatments except for L. lineolaris adult damaged M. sativa showed higher induction relative to the controls (Fig. 3.1C). In the factorial ANOVA, significant differences were seen on the ‘species’ level (df=1,48, F=4.04, p<0.05), but not at any other level (Fig. 3.1C). The combined mean of both L. rubrosignatus damaged treatments again showed a significantly higher induction of linolenic acid relative to the mean of the

L. lineolaris damage treatments. Finally, neither of the L. lineolaris damage treatments showed a significant induction of ABA over the control treatment, while both of the L. rubrosignatus treatments showed a higher induction of ABA relative to controls (Fig.

3.1D). Lygus rubrosignatus adult damaged and L. rubrosignatus nymph damage both induced higher amounts of ABA relative to the two L. lineolaris damage treatments (Fig.

3.1D).

66 Jasmonic acid levels at the systemic site on M. sativa were significantly lower relative to controls for both L. rubrosignatus treatments while the L. lineolaris damage treatments were not significantly different from controls (Fig. 2.2A). There was a significant effect of ‘species’ on the amount of JA (df=1,48, F=44.39, p=0.00) with the two L. rubrosignatus treatments showing a significantly lower induction of JA relative to the other two damage treatments. The remaining phytohormones (SA, linolenic acid,

ABA) showed no significant difference in induction systemically due to Lygus damage relative to the controls (Fig. 3.2B-C).

The levels of JA at the feeding site on Melilotus officinalis were all significantly greater relative to controls except for the L. lineolaris nymph treatment (Fig. 3.3A). In the factorial ANOVA, significant differences were seen on the ‘species’ level (df=1,38,

F=6.33, p<0.02), as well as on the ‘life stage’ level (df=1,38, F=8.94, p<0.01), but there was no effect of the interaction between ‘species’ and ‘life stage’. Overall, L. rubrosignatus adult damage induced significantly more JA than either of the L. lineolaris damage treatments, but this amount was not significantly different from the L. rubrosignatus nymph damage treatment (Fig. 3.3A). None of the remaining three treatments were significantly different from each other. All four Lygus damage treatments induced significantly higher amounts of SA relative to the control (Fig. 3.3B), but none of these damage treatments were significantly different from each other. Linolenic acid was induced in significantly higher amounts relative to controls in the two L. lineolaris treatments, but not in the L. rubrosignatus treatments (Fig. 3.3C). The factorial ANOVA showed no significant differences between the different damage treatments, but there was a slight trend towards a ‘species’ level effect (df=1,38, F=3.24, p=0.08). The two L.

67 rubrosignatus treatments trended towards having less linolenic acid induced than the L. lineolaris treatments (Fig. 3.3C). Measuring the levels of ABA, none of the damage treatments, except for L. rubrosignatus adults, are significantly different from the levels found in controls (Fig. 3.3D). In the factorial ANOVA, there is a slight trend towards there being an effect on the ‘species’ level (df=1,38, F=3.35, p=0.07), but it is not significant. Overall, L. rubrosignatus adult damaged plants are producing significantly more ABA than the plants in the two L. lineolaris damaged treatments, but not significantly more than the L. rubrosignatus treatment plants (Fig. 3.3D).

At the systemic site on M. officinalis only ABA showed any difference in induction relative to controls in response to feeding damage by Lygus bugs (Fig. 3.4). For

ABA, only the L. lineolaris adult damaged plants showed a higher induction of ABA relative to the controls. The factorial analysis showed that there was a significant

‘species’ by ‘life stage’ interaction for levels of ABA, with L. lineolaris adult-damaged plants being higher than L. rubrosignatus adult-damaged plants (Fig. 3.4D).

Discussion

The vast majority of research on plant responses to herbivore damage has been carried out using insects with a chewing or tearing style of feeding, while relatively less work has been done on insects utilizing assorted styles of piercing-sucking feeding, such as bugs in the genus Lygus. Our analysis of Lygus bugs, insects with a unique form of the piercing-sucking feeding mechanism, indicates that they induce jasmonic acid (JA) and salicylic acid (SA) production at the site of damage on two important legume species

(Figs. 3.1 & 3.3). Additionally, in most cases, Lygus feeding also induced the release of

68 linolenic acid, a fatty acid molecule that is the precursor for synthesis of bioactive molecules (such as volatile six carbon alcohols and aldehydes and the in-plant signal, JA) through the action of lypoxygenases, hydroperoxide lyase, and enzymes in the octadecanoid pathway (Matsui 2006) (Figs. 3.1 & 3.3). Jasmonic acid has been shown in many plant systems to be induced by herbivore attack, particularly chewing herbivores

(Bonaventure 2012), and salicylic acid is more commonly induced in response to pathogen infection and is known to regulate a number of downstream pathogen resistance responses (Walling 2000, Thaler et. al. 2012). Other research has shown that salicylic acid is also commonly induced in response to the piercing sucking mechanism used by aphids and whiteflies, which feed (with very little damage) on the phloem of intact plants

(Walling 2000, Zarate et. al. 2007). JA and SA have been shown to have an antagonistic relationship with the upregulation of one often leading to the suppression of the other

(Thaler et. al. 2002, Zarate et. al. 2007, Thaler et. al. 2012). However, this generality does not seem to be the case with damage caused by either of these two Lygus species. At the feeding site JA and SA are both strongly induced relative to control levels, indicating that plants respond in multiple ways to Lygus damage and do not follow the previously described patterns for either chewing herbivores or piercing-sucking herbivores.

This result may be due to the nature of the potential elicitors produced in response to the external digestive action of Lygus salivary components, particularly the polygalacturonases (PGs) which have been shown to exist in all Lygus species so far examined (Laurema et. al. 1985, Frati et. al. 2006, Celorio-Mancera et. al. 2008, de la

Paz Celorio-Mancera et. al. 2009). Interestingly, PGs are also present in many species of fungi that infect and utilize plant tissues through digestion of cell walls, and have been

69 shown to activate both salicylic acid and jasmonic acid pathways in plants (D’Ovidio et. al. 2004). Polygalacturonases from fungi are known to produce bio-active cell wall fragments during the process of breaking down cells (Federici et. al. 2006). Cell wall fragments are perceived by the plant as elicitors if they are not immediately digested into smaller fragments by the action of PGs. Therefore, if Lygus-induced defenses are partially mediated by PGs, they may induce both the pathogen and herbivore directed defense response pathways.

Our analysis also indicates that there are important species level differences in the induction of these phytohormones. Jasmonic acid in particular was higher in L. rubrosignatus treatments relative to L. lineolaris treatments for the site of feeding on

Medicago sativa, and the precursor for JA production, linolenic acid, was also simultaneously higher for L. rubrosignatus treatments on this plant (Fig. 3.1A&C).

Abscisic acid also followed a strong species-specific induction pattern in Medicago, with only L. rubrosignatus feeding inducing this compound above control levels (Fig. 3.1D).

Overall, on M. sativa, L. rubrosignatus damage treatments caused a higher induction of defense responses at the site of feeding, but caused a reduction in levels of JA in the systemically sampled leaf tissue (Fig. 3.2). Melilotus officinalis plants showed a slightly different response to the damage treatments at the feeding site, with higher JA and ABA induction being associated with only L. rubrosignatus adults instead of L. rubrosignatus treatments in general (Fig. 3.3). Additionally, there were very few systemic changes in phytohormone levels in this plant in response to any treatments, with only a slight elevation in ABA levels in response to L. lineolaris adult feeding.

70 The species-level differences we observed, particularly in M. sativa responses to feeding damage, support the hypothesis that bugs differ in the composition of their salivary components, particularly in those that are capable of inducing plant defenses, like PGs.. If Lygus salivary compositions differ, then each mixture may be more or less efficiently inhibited by counter-acting enzymes (for instance, PG inhibiting proteins) present in the plant, or in some cases, not inhibited at all (Frati et. al. 2006). Our data indicate that M. sativa, and to some extent, M. officinalis, have an enhanced defense response to L. rubrosignatus damage relative to L. lineolaris damage in terms of herbivore-specific hormone signals. The high levels of ABA also induced in M. sativa only by L. rubrosignatus may suggest that this species is the more disruptive feeder of the two, since ABA induction is characteristic of water stress in the plant (Taiz and Zeiger

2006). Water stress-like symptoms may be induced if L. rubrosignatus salivary enzymes weaken or destroy cell walls, disrupting turgor pressure within the feeding site. Thus, the more robust response of M. sativa plants to L. rubrosignatus feeding, by either life stage, and of M. officinalis to L. rubrosignatus adult feeding, may also be partially due to the fact that the saliva of this species is more disruptive to plant tissues. Induction of ABA has also been shown to be linked to the herbivore-wounding response in some plant species. In tomato ABA-deficient mutants were less able to defend themselves from herbivore attack than wild-type plants (Thaler and Bostock 2004), and also had a much higher salicylate response and better protection against the plant pathogen Pseudomonas syringae. Additionally, ABA may be able to upregulate the production of the herbivore defense protein proteinase inhibitor II through pathways that are independent of JA

(Damman et. al. 1997).

71 Interestingly, despite the enhanced local response to L. rubrosignatus feeding, the systemic, undamaged tissues of Medicago had suppressed JA production relative to L. lineolaris systemic leaves (Fig. 3.2A). Systemic suppression of JA has been noted in previous experiments, particularly when systemic acquired resistance (SAR) is induced through the SA pathway (e.g., by pathogen attack) (Felton et al. 1999, Thaler et al. 2012).

Salicylic acid was induced by all Lygus feeding treatments, but it is possible that the induction of SAR only occurred in response to L. rubrosignatus feeding, leading to suppression of JA production in other parts of the plant. These results could have important implications in a field context, where L. rubrosignatus feeding could reduce overall plant resistance to chewing herbivores.

While very specific species-level responses to Lygus bug feeding are evident in our analysis of phytohormone induction, these differences do not correlate with the emissions of volatiles previously measured for these same bug species by life stage combinations damaging the same two plant species (Halloran et al. 2012). The previous study found that there were significant differences in emissions of green leaf volatiles

(GLVs) (C6 alcohols, aldehydes and acetates) on a species by life stage basis in M. sativa

(adults of each species differed from each other, and nymphs of each species differed from each other). Additionally, L. lineolaris nymphs induced significantly higher levels of several terpenes and the aromatic methyl salicylate, relative to all other treatments.

However, these differences appear un-related to the very clear patterns of phytohormone induction observed here. For instance, L. rubrosignatus induced higher levels of JA and linolenic acid, but this did not translate into higher emissions of terpenes (expected with higher JA induction) (Schmelz et. al. 2003) or higher emissions of GLVs. In fact, L.

72 rubrosignatus nymphs were shown to have lower emissions of several common GLVs relative to all other damage treatments (Halloran et al. 2012).

A similar lack of correlation between phytohormone patterns and volatile emissions was also found for M. officinalis. On this plant, L. rubrosignatus adults and nymphs induced somewhat higher levels of Z-3-hexen-1-ol and Z-3-hexenyl acetate relative to L. lineolaris, but had levels of linolenic acid that were equal to or slightly less than those for L. lineolaris. Additionally, both L. rubrosignatus adults and nymphs induced emission of three terpenes (Z-beta ocimene, an unknown terpene, and beta- farnesene) that were not induced by L. lineolaris adults or nymphs (which both induced a different set of terpenes) (Halloran et al. 2012). However, only L. rubrosignatus adults induced high levels of JA in Melilotus, with L. lineolaris nymphs not inducing JA at all, even though JA-based emissions (terpenes) were clearly induced by this treatment when only volatiles were measured (Halloran et al. 2012). Additionally, on both plants, emissions of methyl salicylate were observed in response to L. lineolaris nymph damage, but we did not see any indication of differences in SA that might mediate these emissions

(e.g., a drop in SA in L. lineolaris treatments as it was used for production of methyl SA).

This lack of correlation between phytohormone results and the previously observed volatile results indicates that there may be more complex signaling differences occurring in response to feeding by these two species of Lygus bugs that cannot be fully characterized by simply measuring phytohormone levels. An analysis of downstream products (e.g. genes regulated by JA and SA pathways) would provide a mechanistic explanation for the differences observed in indirect defenses induced by Lygus feeding.

73 References:

Alborn, H.T, Turlings, T.C., Jones, T.H., Stenhagen G., Loughrin J.H., Tumlinson, J.H. 1997. An elicitor of plant volatiles from beet armyworm oral secretion. Science. 276: 945-949.

Alborn, H.T., Hansen, T.V., Jones, T.H., Bennett, D.C., Tumlinson, J.H., Schmelz, E.A., Teal, P.E. 2007. Disulfooxy fatty acids from the American bird grasshopper Schistocera americana, elicitors of plant volatiles. Proceedings of the National Academy of Sciences USA. 104:12976-12981.

Allen, M., Mertens, J. 2008. Molecular cloning and expression of three polygalacturonase cDNAs from the tarnished plant bug, Lygus lineolaris. Journal of Insect Science. 8.

Bonaventure, G. 2012. Perception of insect feeding by plants. Plant Biology. doi:10.1111/j.1438-8677.2012.00650.x

Damman, C., Rojo, E., Sánchez-Serrano, J.J. 1997. Abscisic acid and jasmonic acid activate wound-inducible genes in potato through separate, organ-specific signal transduction pathways. The Plant Journal. 11:773-782.

De Boer, J.G., Snoeren, T.A., Dicke, M. 2005. Predatory mites learn to discriminate between plant volatiles induced by prey and nonprey herbivores. 69:869-879. de la Paz Celorio-Mancera, M., Allen, M.L., Powell, A.L., Ahmadi, H., Salemi, M.R., Phinney, B.S., Shackel, K.A., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2008. Polgalacturonase causes lygus-like damage on plants: cloning and identification of western tarnished plant bug (Lygus hesperus) polygalacturonases secreted during feeding. Arthropod-Plant Interactions 2:215-225. de la Paz Celorio-Mancera, M., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2009. Identification of endo- and exo-polygalacturonase activity in Lygus hesperus (Knight) salivary glands. Archives of Insect Biochemistry and Physiology. 70:122-135.

De Moraes, C.M., Lewis, W.J., Paré, P.W., Tumlinson, J.H. 1998. Herbivore infested plants selectively attract parasitoids. Nature. 393:570-574.

Di, C., Zhang, M., Xu, S., Cheng, T., An, L. 2006. Role of poly-galacturonase inhibiting protein in plant defense. Critical Reviews in Microbiology. 32:91-100.

Dicke, M., van Loon, J.J.A., Soler, Roxina, S. 2009. Chemical complexity of volatiles from plants induced by multiple attack. Nature Chemical Biology. 5:317-324.

74 D’Ovidio, R., Mattei, B., Roberti, S., Bellincampi, D. 2004. Polygalacturonases, polygalacturonase-inhibiting proteins and pectic oligomers in plant-pathogen interactions. Biochimica et Biophysica Acta 1696:237-244.

Federici, L., Di Matteo, A., Fernandez-Recio, J., Tsernoglou, D., Cervone, F. 2006. Polygalacturonase inhibiting proteins: players in plant innate immunity? TRENDS in Plant Science. 11:65-70.

Felton, G.W. Korth, K.L., Bi, J.L., Wesley, S.V., Huhman, D.V., Mathews, M.C., Murphy, J.B., Lamb, C., Dixon, R.A. 1999. Inverse relationship between systemic resistance of plants to microorganisms and to insect herbivory. Current Biology. 9:317- 320.

Frati, F., Galletti, R., De Lorenzo, G., Salerno, G., Conti, E. 2006. Activity of endo- polygalacturonases in mired bugs (Heteroptera: Miridae) and their inhibition by plant cell wall proteins (PGIPs). European Journal of Entomology 103:515-522.

Halloran, S.T., Mauck, K.E., Tumlinson, J.H. 2012 Chemical ecology of two co- occurring Lygus species on shared host plants. See chapter 2. pages 23-49, this document.

Kessler, A. and Baldwin, I. 2001. Defensive function of herbivore-induced plant volatile emissions in nature. Science. 291:2141-2144.

Koch T., Krumm T., Jung V., Engelberth J., Boland W. 1999. Differential induction of plant volatile biosynthesis in the lima bean by early and late intermediates of the octadecanoid-signaling pathway. Plant Physiology. 121:153-162.

Laurema, S., Varis, A.L., Miettinen, H. 1985. Studies on enzymes in the salivary glands of Lygus rugulipennis (Hemiptera, Miridae). Insect Biochemistry. 15:211-224.

Mathieu, Y., Kurkdjian A., Xia, H., Guern, J., Koller, A., Spiro, M.D., O’Neill, M., Albersheim, P., Darvill A. 1991. Membrane responses induced by oligogalacturonides in suspension-cultured tobacco cells. The Plant Journal 1:333-343.

Matsui, K. 2006. Green leaf volatiles: hydroperoxide lyase pathway of oxylipin metabolism. Current Opinions in Plant Biology. 9:274-280.

Mattiacci, L., Dicke, M., Postumus, M.A. 1995. β-Glucosidase: an elicitor of herbivore- induced plant odor that attracts host-searching parasitic wasps. Proceedings of the National Academy of Sciences USA. 92:2036-2040.

Piel J., Atzorn R., Gäbler R., Kühnemann F., Boland W. 1997. Cellulysin from the plant parasitic fungus Trichoderma viride elicits volatile biosynthesis in higher plants via the octadecanoid signaling cascade. FEBS Letters. 416:143-148.

75 Ridley, B.L., O’Neill, M.A., Mohnen, D. 2001. Pectins: structure, biosynthesis, and oligogalacturonide-related signaling. Phytochemistry. 57:929-967.

Rodriguez-Saona, C., Crafts-Brandner, S.J., Cañas, L.A. 2003. Volatile emissions triggered by multiple herbivore damage: Beet armyworm and whitefly feeding on cotton plants. Journal of Chemical Ecology. 29:2539-2550.

Schmelz, E.A., Alborn, H.T., Banchio, E., Tumlinson, J.H. 2003. Quantitative relationships between induced jasmonic acid levels and volatile emissions in Zea mays during Spodoptera exigua herbivory. Planta. 216:665-673.

Schmelz E.A., Engelberth, J., Alborn, H.T., O’Donnell, P., Sammons, M., Toshima, H., Tumlinson, J.H. 2003. Simultaneous analysis of phytohormones, phytotoxins, and volatile organic compounds in plants. Proceedings of the National Academy of Sciences. 100:10552-10557.

Shackel, K.A., Celorio-Mancera, M.P., Ahmadia, H., Greve, L.C., Teuber, L.R., Backus, E.A., Labavitch, J.M. 2005. Micro-injection of Lygus salivary gland proteins to simulate feeding damage in alfalfa and cotton flowers. Archives of Insect Biochemistry and Physiology. 58:69-83.

Strong F.E. 1970. Physiology of injury caused by Lygus hesperus. Journal of Economic Entomology 63:808-814.

Taiz, L., Zeiger, E. “Secondary Metabolites and Plant Defense.” Plant Physiology. Sunderland, MA: Sinauer Associates Inc., 2006. 315-344.

Thaler, J.S., Bostock, R.M. 2004. Interactions between abscisic-acid-mediated responses and plant resistance to pathogen and insects. Ecology. 85:48-58

Thaler, J.S., Karban, R., Ullman, D.E., Boege, K., Bostock, R. 2002. Cross-talk between jasmonate and salicylate plant defense pathways: effects on several plant parasites. Oecologia. 131:227-235.

Thaler, J.S., Humphrey, P.T., Whiteman, N.K. 2012. Evolution of jasmonate and salicylate signal crosstalk. Trends in Plant Science. 17:260-270.

Turlings, T.C., Tumlinson, J.H. 1992. Systemic release of chemical signals by herbivore- injured corn. Proceedings of the National Academy of Science USA. 89:8399-8402.

Turlings, T.C., McCall, P.J., Alborn, H.T., Tumlinson, J.H. 1993. An elicitor in caterpillar oral secretions that induces corn seedlings to emit chemical signals attractive to parasitic wasps. Journal of Chemical Ecology. 19:411-425.

76 Walker, W., and Allen, M. 2010. Expression and RNA interference of salivary polygalacturonase genes in the tarnished plant bug, Lygus lineolaris. Journal of Insect Science. 10.

Walling, L.L. 2000. The myriad plant responses to herbivores. J. Plant Growth Regul. 19:195-216.

Zarate, S.I., Kempema, L.A., Walling L.L. 2007. Silverleaf whitefly induces salicylic acid defenses and suppresses effectual jasmonic acid defenses. Plant Physiology. 143:866-875.

77

Figure 3.1: Medicago sativa Feeding Site Analysis Phytohormones detected in leaf tissue samples from the feeding site of damaged M. sativa. Abbreviations on the x-axis: LLA = Lygus lineolaris adult damage, LLN = L. lineolaris nymph damage, LRA = L. rubrosignatus adult damage, LRN = L. rubrosignatus nymph damage. In figures, dotted line indicates the mean for the control treatment, [+] indicates that a damage treatment is significantly different from the control mean at P≤0.05, [*] indicates that the species level means are significantly different at P≤0.05, and letters indicate significant differences among individual species x life stage treatments at P≤0.05.

78

Figure 3.2: Medicago sativa Systemic Site Analysis Phytohormones detected in leaf tissue samples from the systemic site of damaged M. sativa. Abbreviations on the x-axis: LLA = Lygus lineolaris adult damage, LLN = L. lineolaris nymph damage, LRA = L. rubrosignatus adult damage, LRN = L. rubrosignatus nymph damage. In figures, dotted line indicates the mean for the control treatment, [+] indicates that a damage treatment is significantly different from the control mean at P≤0.05, [*] indicates that the species level means are significantly different at P≤0.05, and letters indicate significant differences among individual species x life stage treatments at P≤0.05.

79

Figure 3.3: Melilotus officinalis Feeding Site Analysis Phytohormones detected in leaf tissue samples from the feeding site of damaged M. officinalis. Abbreviations on the x-axis: LLA = Lygus lineolaris adult damage, LLN = L. lineolaris nymph damage, LRA = L. rubrosignatus adult damage, LRN = L. rubrosignatus nymph damage. In figures, dotted line indicates the mean for the control treatment, [+] indicates that a damage treatment is significantly different from the control mean at P≤0.05, [*] indicates that the species level means are significantly different at P≤0.05, and letters indicate significant differences among individual species x life stage treatments at P≤0.05.

80

Figure 3.4: Melilotus officinalis Systemic Site Analysis Phytohormones detected in leaf tissue samples from the systemic site of damaged M. officinalis. Abbreviations on the x-axis: LLA = Lygus lineolaris adult damage, LLN = L. lineolaris nymph damage, LRA = L. rubrosignatus adult damage, LRN = L. rubrosignatus nymph damage. In figures, dotted line indicates the mean for the control treatment, [+] indicates that a damage treatment is significantly different from the control mean at P≤0.05, [*] indicates that the species level means are significantly different at P≤0.05, and letters indicate significant differences among individual species x life stage treatments at P≤0.05.

Chapter 4

Chemical ecology of Erigeron annuus: a weedy host and trap crop of bugs in the genus Lygus

Abstract

Insects belonging to the genus Lygus (Hemiptera: Miridae) are major pests on cotton throughout the United States. Currently Lygus is controlled almost entirely through spraying of pesticides, but developing resistance within Lygus populations necessitates the introduction of more sustainable long-term management techniques. A common field edge weed, Erigeron annuus, has been shown to be remarkably attractive to the tarnished plant bug, Lygus lineolaris. These attractive cues function over long distances and result in both attraction and arrestment of L. lineolaris. However, the mechanism of this attraction is currently unknown. Understanding the mechanisms by which E. annuus attracts and arrests L. lineolaris would facilitate use of Erigeron to direct the movements of Lygus in the field. To understand how E. annuus is attractive, we explored the blend of volatile organic compounds emitted by the plants with and without the presence of Lygus bug damage. We also tested the attractiveness of these volatiles to Lygus females in a series of Y-tube bioassays. Volatile collections from E. annuus demonstrate that this species produces a highly complex volatile blend consisting of over 60 distinct compounds at amounts far greater than similarly damaged cotton plants. Furthermore, when fed upon by Lygus, novel compounds are induced and constitutive compounds up- regulated. Additionally, we found that female Lygus bugs are highly attracted to these

82 volatiles over those of cotton in the Y-tube, and that this attraction is not simply due to the volume of volatiles released, but appears to be based on qualitative aspects of the

Erigeron blend. These results indicate that L. lineolaris will preferentially visit and aggregate on this weedy host. Erigeron may therefore be useful to concentrate Lygus into a small area – adjacent to or separate from crops – that could then be subject to application of chemical or biological controls.

Introduction

Lygus bugs are currently one of the few groups of major cotton pests still controlled by chemical spraying. Due to advances in the development of non-pesticide control strategies for management of key pests, such as the boll-weevil, over the past 40 years, cotton can now be described as a low-spray crop (Hardee et.al. 2001, Layton

2000). In recent years growers have begun to reduce early season spraying for many non-

Lygus pests due in part to the success of boll weevil eradication programs as well as the development and widespread use of transgenic cotton that controls lepidopterous pests

(Hardee et.al. 2001, Layton 2000). Additionally, some Lygus species have shown a consistent ability to develop resistance to different classes of insecticides. Snodgrass

(1996) demonstrated widespread resistance of Lygus lineolaris to pyrethroids as well as some organophosphate and cyclodiene insecticides. More recently, varying levels of acephate resistance have been identified in L. lineolaris populations in the areas of

Louisiana, Mississippi, and Arkansas (Snodgrass & Scott 2000).

A renewed focus on integrated pest management of Lygus bugs is necessary to prolong the usefulness of existing chemical controls, and to move towards an overall

83 more sustainable Lygus control strategy. Knowledge of the ecology of Lygus bugs, including their host plant preferences and associated behaviors, is useful in this regard because strong host plant preferences can be used to pull herbivorous pests out of a crop and into a small area of preferred hosts – a “trap crop” (Hokkanen 1991, Cook et al.

2007, Pickett et al. 1997). Furthermore, a thorough understanding of the mechanisms underlying the use of a particular trap crop will facilitate more precise recommendations on use.

Herbivorous insects – particularly highly mobile, strong flyers like Lygus bugs

(MacCreary 1965) – often use plant-produced chemical cues such as volatile organic compounds (VOCs) in order to locate their plant hosts (Bernays and Chapman 1994,

Bruce et. al. 2005). Exploiting this basic behavior through the use of volatile cues can serve as a means of controlling economically important insect pests (Cook et. al. 2007,

De Camelo et. al. 2007). Lygus bugs have been shown to respond to a number of plant- produced compounds by orienting towards the source of those compounds, indicating that

VOC blends are likely important in their host-finding in nature (Blackmer et al. 2004,

Blackmer & Cañas 2005, Williams III et. al. 2010). Additionally, feeding by herbivorous insects can also change the volatile blends released by their host plants both at the site of feeding damage as well as systemically in other parts of the plant (Röse et. al. 1996, Paré and Tumlinson 1997, Dicke and Van Loon 2003). These induced changes in the volatile blend can then make the hosts either more attractive (Blackmer et. al. 2004) or less attractive to conspecifics (De Moraes et. al. 2001).

Feeding and oviposition by Lygus hesperus has been shown to alter and upregulate volatile release in cotton (Rodriguez-Saona et. al. 2002, Williams III et. al.

84 2005), as well as alfalfa (Blackmer et. al. 2004) relative to undamaged plants.

Furthermore, L. hesperus adults have been shown to be attracted to plants either recently damaged or currently being damaged by conspecifics (Blackmer et. al. 2004, Blackmer and Cañas 2005). Here we examine VOC-mediated interactions between Lygus and a very promising potential trap crop, Erigeron annuus, an easy to grow weed that is consistently preferred by Lygus in nature over other weeds and many target crops.

Previous research has shown that Erigeron annuus edge weeds in highway right- of-ways harbor significantly more adult Lygus lineolaris bugs than other co-occurring weeds, which indicates that adults are preferentially choosing pre-flowering and flowering Erigeron as a feeding and oviposition site (Fleischer and Gaylor 1987). One study also used rubidium-marked individuals to examine the movement of Lygus in relation to patches of E. annuus planted within and adjacent to a cotton field (Fleischer et. al. 1988). When pre-infested plots of E. annuus within the cotton field were destroyed, more marked individuals were recovered from the trap plots of E. annuus on the edge of the cotton field than within the crop itself (Fleischer et. al. 1988). Cage studies also showed that when E. annuus was available, Lygus will leave cotton and migrate onto E. annuus (Fleischer et. al. 1988). Other studies indicate that Erigeron species, including E. annuus, are highly preferred as a good-quality food source and oviposition site for a native parasitoid of Lygus bugs (Peristenus pseudopallipes) (Streams et al. 1968) and that increased densities of Erigeron result in increased parasitism rates within Erigeron patches (Shahjahan and Streams 1973). Thus, E. annuus may fulfill a dual role in attracting and retaining Lygus bugs and attracting and provisioning Lygus natural enemies.

85 In this study we examine the role of VOCs in mediating interactions among

Lygus, E. annuus, and cotton using a combination of volatile-based behavioral assays and chemical analytical techniques that examine and compare volatile emissions from these major hosts. By undertaking this study we hope to provide a mechanistic basis for Lygus movement in the field that will aid in the implementation of weed-based trap crops like

Erigeron species and efforts to combine this strategy with natural enemy biological controls.

Materials and Methods

Insect Rearing and Maintenance

Lygus rubrosignatus adults and nymphs were part of a laboratory colony originally obtained from The Phillip Alampi Beneficial Insect Laboratory, West Trenton,

NJ. The colony was kept in a rearing room maintained at 22±2°C, 60±10% RH, and a light-dark regimen of 14:10 hours (L:D). Individuals were reared on packets of Lygus artificial diet purchased from Bio-serv, while adult females were allowed to oviposit on similar packets full of Carrageenan Gelcarin GP812 (PhytoTechnology Laboratories,

Shawnee Mission, KS). Diet packs were changed out every other day, while oviposition packs were removed when they became full of eggs. Oviposition packs were then placed into new containers where the nymphs were allowed to hatch and grow.

Plant Maintenance

All plant species were grown in a pest-free greenhouse with a light:dark regimen of 16:8 hours and a relative humidity of 50±10%. Plants were grown in Metro-Mix

86 potting soil with 5g of Osmocote Plus slow release fertilizer (Scotts) mixed in at the time of potting. All plants were bottom-watered daily with a hose.

Erigeron annuus was grown in 6-inch diameter pots while Gossypium hirsutum

(cultivar DPL90) was grown in 4-inch diameter pots. Both species were maintained at

27±2°C. E. annuus was allowed to grow through its rosette stage and into its bolting stage (approximately 5-6 weeks). Bolting E. annuus (buds present but not yet flowering) were used for the volatile collections and behavioral assays. Gossypium hirsutum plants were grown until early squares began to appear (approximately 4 weeks) at which point the plants were used in experiments.

Volatile collection and analysis

To quantify and identify VOCs released by the study plants, the headspace around the plants was collected using a push/pull design. An automatic volatile collection system built by Analytical Research Systems (Gainsville, FL) was used to control sampling periods. The system was capable of controlling simultaneous collections from 12 individual treatment plants at any chosen set of intervals 24 hours a day.

A portion of each non-excised plant was enclosed in a 7L glass bell jar with a metal guillotine-type base. Clean cotton-balls were packed around the junction of the stem and base to prevent air from lower portions of the plants or the soil from entering the bell jars. The portion of the plant being collected from was a single stem with buds for E. annuus and the apical portion of the main stem with three developing squares for

G. hirsutum. Charcoal-purified air was pumped through the top of the bell jars at an average rate of 4L min-1and allowed to pass over the plant before being pulled out

87 through the bottom of the jars at a rate of 1L min-1. This air was pulled through filters containing 45mg of Super-Q adsorbent (80/100 mesh, Alltech).

The plants were divided into three treatments during the collections (1) control

(undamaged), (2) damage from L. rubrosignatus adult feeding, and (3) damage from L. rubrosignatus nymph feeding. For treatments involving insect damage, either 12 adult

(1:1 sex ratio) or 16 nymph L. rubrosignatus, were placed in the bell jars an hour before collections began (4 extra nymphs were added to account for the fragility of this life stage during collection and movement). All insects remained in the bell jars and were allowed to constantly feed over the whole duration of the collections. The collections were checked several times a day and any dead insects were aspirated out and replaced with new individuals. Sample sizes are between 4-8 plants per species x damage treatment.

The collections were run during the day over a 12-hour period starting at 9:30 and ending at 21:30. These day collections were divided between 3 Super-Q filters that each collected for 4 hours each. This prevented breakthrough or loss of small molecular weight compounds, and allowed for analysis of different time points if differences were observed. Collections occurred between December of 2008 and March of 2009. All collections occurred within a pest-free greenhouse maintained at 25°C, 50%RH and a light:dark cycle of 16:8 hours.

Samples were analyzed by first eluting the compounds off of each filter using

125μL of 1:1 dichloromethane:hexanes (Burdick & Jackson) and then adding 200ng n- octane and 400ng nonyl acetate to each eluted sample to act as internal standards. The eluted samples were then analyzed on an Agilent 6890 analytical gas chromatograph

(GC) equipped with an FID detector, with a splitless injector, and a HP-1 column (15m x

88 0.25mm x 0.25μm, Agilent). In order to identify compounds, selected samples were run on an Agilent 6890N GC equipped with an Agilent 5973N mass selective detector configured for electron impact mode and a HP-1MS column (30m x 0.25mm x 0.25μm,

Agilent). Mass spectra were compared to spectra for standards available in the National

Institute of Standards and Technology (NIST) library as well as known standards from the lab. The Kovats Index (KI) of each compound was also determined (Kováts 1965) and used to identify some compounds by comparing the unknown sample KIs to those of known standards run on the same type of GC column (HP-1). These indices were also used to ensure consistency of identification between experiments over time.

Behavioral assays

To determine up-wind, volatile based orientation preferences of Lygus bugs for E. annuus vs. crop plants under different damage conditions (as described previously) a vertical Y-tube olfactometer was used (2.5cm diameter, base 18cm, arms 11cm). The choice tests performed are outlined in Table 4.1 and all tests assessed the behavior of adult females, as these are the most relevant life stage due to their mobility and the fact that they are searching for feeding and oviposition sites (the resulting flightless nymphs being the main source of economic losses in crops). Plant odors were derived from intact portions of whole plants enclosed in 7L glass bell jar (1-2 stalks of pre-flowering E. annuus and the apical squaring portion of cotton plants). Clean, humidified air was delivered to each bell jar at a rate of 4L min-1 and a Teflon tube connected the headspace within each dome to one of the Y-tube arms using a ground-glass connector. To perform each test, a single adult female bug was placed in a ground-glass connecting chamber at the base of the Y-tube. The connector was immediately fitted into the Y-tube base and a

89 vacuum (1L min-1) pulled air through the Y-tube from each of the odor sources via the connecting Teflon tubes. The bug was given five minutes to walk up the Y-tube and enter one of the arms (with the aid of traction provided by a Y-shaped stick that had been positioned inside the tube). If the bug entered an arm and walked the entire length to the endpoint then that was considered a choice. Bugs that did not choose within 5 minutes were excluded from the analysis. The Y-tube set up was tested with clean air vs. a squaring cotton plant to ensure that bugs could locate and orient towards cotton volatiles even in the absence of another choice. The bioassay materials were cleaned with acetone and the treatment input arms switched every 4 bugs. Volatile samples were collected from several of the odor sources simultaneous to running choice tests to verify that the cues being delivered were similar to those we detected during the more detailed collections performed independent of the choice tests.

Additionally, after the first round of data collection, an additional choice combination of undamaged E. annuus vs. cotton damaged by 40 Lygus nymphs was added to the treatment list to attempt to correct for the large difference in total volatiles released from E. annuus and cotton plants when the same damage treatment is applied to both plants.

Statistical analysis

Total volatile emissions over the daytime period were calculated for each replicate plant by summing the amounts of each recorded compound across the three samples collected over 12 hours (amounts calculated relative to the internal standard nonyl acetate based on peak area). Total volatiles emitted by cotton and Erigeron were

90 compared within each bug damage treatment using GLM (Minitab) with plant species as the main factor. Compound means and standard errors were also calculated and each damage treatment was compared to the control to determine if a compound was novel, up-regulated, or down-regulated. Behavioral tests were analyzed using chi-square tests and volatiles collected during choice tests were analyzed as for the larger collections.

Results

Volatile collection and analysis

Erigeron annuus released significantly more total volatiles than cotton both in the presence and absence of different types of Lygus damage (Fig. 4.1). This was despite the much larger size and total leaf mass of the cotton plants relative to E. annuus. Induction was also observed for both plant species in the presence of Lygus bug feeding (Fig. 4.1,

Tables 4.2&4.3). Table 4.2 displays the compounds emitted by E. annuus across the three damage treatments, as well as the mean and standard error for each compound within each treatment. In addition to an increase in the number of compounds (from 93 to 114), many compounds that are released constitutively in low amounts are up-regulated when

E. annuus is damaged by adult Lygus bugs. Nymphs also induce novel compounds and cause up-regulation of constitutive compounds, but several constitutive compounds are also absent from the nymph-induced blend (Table 4.2). Differences in the blends among

E. annuus damaged by adults and nymphs are also apparent, with some constitutively released compounds up-regulated differently between the two treatments (Table 4.2).

Overall, both constitutive and induced blends released from Erigeron are rich in terpenes, with induced blends showing considerable up-regulation of constitutive terpenes as well

91 as induction of novel terpenes. Cotton plants also showed induction of new compounds in response to both adult and nymph feeding, with more novel compounds being induced by nymph feeding than adult feeding (Table 4.3). However, cotton plants, even under damage, released less than half the number of compounds relative to Erigeron (Table

4.3). There is some overlap with Erigeron treatments in the identities of terpenes induced by damage (Tables 4.2 & 4.3), indicating that in both instances feeding may be inducing similar pathways. Additionally, both species emit more alcohols, aldehydes and acetates

(“green leaf volatiles” or “GLVs) in response to feeding. Collectively, these results indicate that E. annuus releases a highly complex blend of compounds in large amounts, both constitutively and when attacked by Lygus bugs, and that volatile emissions from E. annuus are significantly higher (sometimes up to 10-fold) than emissions from the co- occurring cotton crop plant.

Behavioral assays

In all choice combinations where both plants had the same treatment applied

(undamaged, nymph damaged, female adult damaged, male adult damaged) a significant percent of the female Lygus adults chose to move towards the E. annuus odor source

(Fig. 4.2). However, when cotton damaged by female Lygus bugs was tested against undamaged E. annuus there was no significant difference between the two choices (Fig.

4.3), while the combination of cotton plants damaged with 40 nymphs vs. undamaged E. annuus plants showed a significant percent of the females choosing the Erigeron odor

(Fig. 4.3). In combinations involving a cotton plant vs. clean air, a significant percentage of the females moved towards the cotton (Fig. 4.3).

92 Total volatiles released from both the nymph damaged cotton and the undamaged

E. annuus used in the choice tests are not significantly different (Fig. 4.4). Total volatiles released from the undamaged E. annuus vs. female damaged cotton used in the choice tests do show a significant difference in total emissions (Fig. 4.4). These data demonstrate that a quantitative difference in volatile production does not fully explain the remarkable attractiveness of Erigeron.

Discussion

Our data demonstrates that Erigeron annuus produces a complex blend of volatile organic compounds both when the plants are undamaged or damaged by Lygus feeding.

Between damaged and undamaged Erigeron plants the blends differ in both the quantity of total volatiles as well as in the identity of the compounds that make up the blend

(Table 4.2). Relative to cotton, E. annuus plants release both more compounds and greater amounts of the same compounds overall. Our behavioral data demonstrate that this volatile blend plays a strong role in mediating the attraction of Lygus bugs to

Erigeron, which has previously been observed in the field (Fleisher et. al. 1988), since female adult Lygus bugs showed an attraction to E. annuus in Y-tube bio-assays that prevented the insects from using visual or gustatory stimuli.

Despite the overwhelming volume of volatile compounds that it emits, the attractiveness of E. annuus to Lygus bugs cannot be explained by quantity of volatiles alone. When the same damage treatment was applied to both cotton and Erigeron plants the Erigeron plants did emit a much larger total volume of volatiles relative to the cotton

(Fig.4.1), and were more attractive in all cases (Fig. 4.2). However, when different

93 damage treatments were applied to the two choice plants (nymph damaged cotton vs. undamaged Erigeron), total volatiles were similar (Fig. 4.4), but female adult Lygus bugs were still attracted to Erigeron over the cotton treatment (Fig. 4.3). Similarly, when bugs were presented with a choice between undamaged Erigeron and female adult damaged cotton, they did not display a preference (Fig. 4.3) even though Erigeron released more total volatiles than the damaged cotton (Fig. 4.4). These data indicate that specific qualitative aspects of the blend of Erigeron mediate the increased attraction (possibly the large diversity and volume of terpenes emitted from both undamaged and damaged

Erigeron plants [Table 4.2]).

Previous studies into Lygus attraction towards specific plant volatiles has shown that several green leaf volatiles, terpenes, and aromatic compounds are detected by Lygus bugs and elicit a variety of responses. Blackmer et. al. (2004) demonstrated that female adult L. hesperus were attracted to Medicago sativa plants that had either been damaged by conspecifics for several days or recently had conspecifics added onto them.

Furthermore it was shown that the compounds (Z)-β -ocimene, (E)-β-ocimene, β- caryophyllene, and α-farnesene were upregulated by L. hesperus damage, while β-pinene, myrcene, methyl salicylate, and tridecatetraene were all novel compounds induced by bug feeding.

While Blackmer et. al. (2004) did not test these compounds individually,

Williams III et. al. (2010) carried out both electroantennogram (EAG) tests and Y-tube olfactometer bioassays that tested the response of L. hesperus to individual compounds.

They found that there was strong antennal response to the GLVs (E)-3-hexen-1-ol, (Z)-3- hexen-1-ol, 1-hexenol, (E)-2-hexenyl acetate, and (E)-2-hexenal and a moderate response

94 to the terpenes (E)-ß-ocimene, α-farnesene, and linalool and the aromatic methyl salicylate. All of these compounds, as well as many other GLVs and monoterpenes, appear in the volatile blend from E. annuus in both undamaged and Lygus damaged treatments (Table 2). However, the Y-tube tests carried out by Williams III et. al. (2010) were less consistent with male and female L. hesperus showing no positive response to several of these individual compounds and in some cases even repellant responses. This may be due to the fact that insects are attracted to specific blends of volatiles rather than any one volatile within a blend (Ngumbi et. al. 2007).

While the large total amount of volatiles produced by E. annuus may not play a direct role in making the plant attractive to Lygus, this characteristic may make the plant ideal for use as a trap crop. A single Erigeron plant emits a far greater amount of volatiles than a similarly sized G. hirsutum plant, which means that a travelling Lygus bug may be able to locate and preferentially move to a small patch of Erigeron even if it’s located in a large cotton field. In fact, this seems to be the case as seen in field studies carried out by Fleisher et. al. (1988) where L. lineolaris adults were shown to disperse through a 4.7 ha field of cotton into 4 6m by 6m plots of E. annuus located at the edges of the field.

Even more importantly, our data show that Erigeron remains an attractive host plant for

Lygus females relative to cotton under a wide variety of damage treatments, so even during the portion of the growing season when some Lygus have migrated onto cotton plants the patches of Erigeron will still be more likely to attract the majority of new females migrating into the field.

Simply ‘pulling’ Lygus bugs off of cotton may not be sufficient alone to make

Erigeron a successful trap crop for use in controlling the pest. A higher density of Lygus

95 on the trap plant species relative to the economically important crop does not always lead to a decrease in the pest population on the crop or a decrease in the damage caused to the crop. This can be seen in attempts to manage L. rugulipennis in strawberry using

Matricaria recutita and Medicago sativa as trap crops (Easterbrook and Tooley 1999).

Additional management steps may be necessary, such as mowing of the trap crop, or spraying of pesticides into the trap crop patches. Some successful attempts at controlling

Lygus species using trap crops have used these methods. Accinelli et. al. (2005) found that plots of M. sativa grown alongside lettuce plots did not reduce the damage on lettuce by L. rugulipennis feeding alone, but when combined with spraying of pesticides in the

M. sativa plots damage was significantly reduced on the lettuce. Godfrey and Leigh

(1994) also found some success in controlling L. hesperus numbers in cotton that was interplanted with strips of M. sativa that were then cut over the course of the growing season (see also Stern et. al. 1969). However, Godfrey and Leigh (1994) also noted that increased frequency of mowing in the M. sativa patches led to decreases in populations of natural enemies of Lygus. Natural enemies are also a concern when using a pesticide as they are often more susceptible to insecticide than the pests themselves (Ruberson et al.,

1998). In fact, parasitoids of Lygus have been found to be particularly sensitive to insecticide regimes (Tillmon and Hoffman 2003, Williams III et. al. 2003).

The effects of management practices on natural enemies of Lygus are important, since biological control using natural enemies could function as a logical addition to a trap crop system. However, native natural enemies of L. lineolaris, including several parasitoids, have been shown to be inadequate for control in crop systems (Clancy and

Pierce 1966, Day et.al. 1990, Carignan et.al. 2007,). As a result, over the last several

96 decades the United States Department of Agriculture – Agriculture Research Service

(USDA-ARS) started a program for the mass-rearing and release of two nymph-attacking parasitoid species, Peristenus digoneutis and Peristenus relictus (Hymenoptera:

Braconidae), both native to Europe. The first parasitoid, P. digoneutis, showed considerable success in establishing and parasitizing Lygus populations in the Northeast

United States and southern Canada (Day 1996), but was unable to establish in warmer climates. The more recent parasitoid P. relictus, however, has already shown some success in establishing farther south than P. digoneutis (Pickett et.al 2007, Hoelmer et.al.

2008) and is a promising candidate for controlling Lygus populations in cotton. The presence or absence of different weedy hosts can greatly increase or undermine the effectiveness of different natural enemies (reviewed in Bottrell and Barbosa 1998).

Further research is necessary to determine how the presence of Erigeron could affect the host finding ability of P. relictus, as well as its ability to control Lygus in cotton.

However, previous research demonstrating the compatibility of Erigeron hosts with other

(native) Lygus parasitoids in the genus Peristenus (Streams et al. 1968) indicates that combining Erigeron trap patches with releases of introduced Peristenus could provide a long-term, sustainable solution for controlling Lygus in cotton, and potentially in other crops.

97 References

Accinelli, G., Lanzoni, A., Ramilli, F., Dradi, D., Burgio, G. 2005. Trap crop: an agroecological approach to the management of Lygus rugulipennis on lettuce. Bulletin of Insectology. 58:9-14.

Bernays, E.A., and Chapman, R.F. Host-plant Selection by Phytophagous Insects. New York: Chapman & Hall, 1994. Print

Blackmer, J.L., Rodriguez-Saona, C., Byers, J.A., Shope, K.L., Smith, J.P. 2004. Behavioral response of Lygus hesperus to conspecifics and headspace volatiles of alfalfa in a Y-tube olfactometer. J. Chem. Ecol. 30:1547-1564.

Blackmer, J.L., Cañas, L.A. 2005. Visual cues enhance the response of Lygus hesperus (Heteroptera: Miridae) to volatiles from host plants. Environmental Entomology. 34:1524-1533.

Bottrell, D.G., Barbosa, P. 1998. Manipulating natural enemies by plant variety selection and modification: a realistic strategy? Annu. Rev. Entomol. 43:347-367.

Bruce, T.J.A., Wadhams, J., Woodcock, C.M. 2005. Insect host location: a volatile situation. TRENDS in Plant Science. 10:69-274.

Carignan, S., Stewart, R.K., Godin, C., Bolvin, G. 2007. Parasitism activity of Peristenus spp. (Hymenoptera: Braconidae) on Lygus lineolaris (Hemiptera: Miridae) nymphs prior to the establishment of P. digoneutis in southwestern in Quebec. Biocontrol Science and Technology 17:623-633.

Clancy, D.W., Pierce H.D. 1966. Natural Enemies of Some Lygus Bugs. J. Econ. Entomol. 59:853-858.

Cook, S.M., Khan, Z.R., Pickett, J.A. 2007. The use of push-pull strategies in integrated pest management. Annu. Rev. Entomol. 52:375-400.

Day, W.H., Hedlund, R.C., Saunders, L.B., Coutinot, D. 1990. Establishment of Peristenus digoneutis (Hymenoptera: Braconidae), a parasite of the Tarnished Plant Bug (Hemiptera: Miridae), in the United States. Ann. Entomol. Soc. Am. 85:482-488.

Day, W.H. 1996. Evaluation of biological control of the Tarnished Plant Bug (Hemiptera: Miridae) in alfalfa by the introduced parasite Peristenus digoneutis (Hymenoptera: Braconidae). Environ. Entomol. 25:512-518.

De Camelo, L.A., Landolt, P.J., Zack, R.S. 2007. A kairomone based attract-and-kill system effective against alfalfa looper (Lepidoptera: Noctuidae). J. Econ.Entomol. 100:366-374.

98

De Moraes, C.M. Mescher, M.C., Tumlinson, J.H. 2001. Caterpillar-induced nocturnal plant volatiles repel conspecifics females. Nature. 410:577-580.

Dicke, M., Van Loon, J.J.A. 2003. Multitrophic effects of herbivore-induced plant volatiles in an evolutionary context. Entomol. Exp. Appl. 97:237–249

Easterbrook, M.A., Tooley, J.A. 1999. Assessment of trap plants to regulate numbers of the European tarnished plant bug, Lygus rugulipennis, on late-season strawberries. Entomologia Experimentalis et Applicata. 92:119-125.

Fleischer, S.J., Gaylor, M.J. 1987. Seasonal abundance of Lygus lineolaris (Heteroptera: Miridae) and selected predators in early season uncultivated hosts: Implications for managing movement into cotton. Environmental Entomology. 16:379-389.

Fleischer, S.J., Gaylor, M.J., Hue, N.V. 1988. Dispersal of Lygus lineolaris (Heteroptera: Miridae) adults through cotton following nursery host destruction. Environmental Entomology. 17:533-541.

Godfrey, L.D., Leigh, T.F. 1994. Alfalfa harvest strategy effect on Lygus bug (Hemiptera: Miridae) and insect predator population density: implications for use as trap crop in cotton. Environmental Entomology. 23:1106-1118.

Hardee, D.D., Van Duyn, J.W., Layton, M.B., Bagwell, R.D. 2001. Bt Cotton & Management of the Tobacco Budworm-Bollworm Complex. U.S. Department of Agriculture, Agricultural Research Service, ARS–154. 40 pp.

Hoelmer, K., Fuester, R., Taylor, P., Coutinot, D., Dorsey, T., Hudson, W., Robbins, G., Peiffer, R., Barczewski, R. 2008. Introduction of a European parasitoid, Peristenus relictus (=P. stygicus), against tarnished plant bug in the mid-Atlantic states. Journal of Insect Science. 8:49.

Hokkanen, H.M.T. 1991. Trap cropping in pest management. Annu. Rev. Entomol. 36:119-138.

Kováts, E. “Gas chromatographic characterization of organic substances in the retention index system.” Advances in Chromatography, Vol. 1. Ed. Giddings, J.C., Keller, R.A. New York. Marcel Dekker, 1965. 229-247.

Layton, M.B. 2000. Biology and Damage of the Tarnished Plant Bug, Lygus lineolaris, in Cotton. Southwest. Entomol. Suppl. 23:7-20.

MacCreary, D. 1965. Flight range observations on Lygus lineolaris and certain other Hemiptera. J. Econ. Entomol. 58:1004-1005.

99

Ngumbi, E., Eigenbrode, S.D., Bosque-Pérez, N.A., Ding, H., Rodriguez, A. 2007. Myzus persicae is arrested more by blends than by individual compounds elevated in headspace of PLRV-infected potato. J. Chem. Ecol. 33:1733-1747.

Paré, P.W., Tumlinson, J.H. 1997. De novo biosynthesis of volatiles induced by insect herbivory in cotton plants. Plant Physiol. 114:1161-1167.

Pickett, C.H., Rodriguez, R., Brown, J. Coutinot, D., Hoelmer, K.A., Kuhlmann, U., Goulet, H., Schwartz, M.D., Goodell, P.B. 2007. Establishment of Peristenus digoneutis and P. relictus (Hymenoptera: Braconidae) in California for the control of Lygus spp. (Heteroptera: Miridae). Biocontrol Science and Technology. 17:261-272.

Pickett, J.A., Wadhams, L.J., Woodcock, C.M. 1997. Developing sustainable pest control from chemical ecology. Agriculture, Ecosystems and Environment. 64:149-156.

Rodriguez-Saona, C., Crafts-Brandner, S.J., Williams III, L., Paré, P.W. 2002. Lygus hesperus feeding and salivary gland extracts induce volatile emissions in plants. J. Chem. Ecol. 28:1733-1747.

Röse, U.S.R., Manukian, A., Heath, R.R., Tumlinson, J.H. 1996. A systemic response of living plants to caterpillar damage. Plant Physiol. 111:487-495.

Ruberson, J.R., Nemoto, H., Hirose, Y. 1998. Pesticides and conservation of natural enemies in pest management. In: Barbosa, P. (Ed.), Conservation Biological Control. Academic Press, London.

Shahjahan, M., Streams, F.A. 1973. Plant Effects on host-finding by Leiophron pseudopallipes (Hymenoptera: Braconidae), a parasitoid of the tarnished plant bug. Environmental Entomology. 2:921-925.

Snodgrass, G.L. 1996. Pyrethroid resistance in field populations of the tarnished plant bug (Heteroptera: Miridae) in cotton in the Mississippi Delta. J. Econ. Entomol. 89:783- 790.

Snodgrass, G.L., Scott, W.P. 2000. Seasonal changes in pyrethroid resistance in tarnished plant bug (Heteroptera: Miridae) populations during a three-year period in the delta area of Arkansas, Louisiana, and Mississippi. J. Econ. Entomol. 93:441-446.

Stern, V.M., Mueller, A., Sevacherian, V. Way, M. 1969. Lygus bug control in cotton through alfalfa interplanting. California Agriculture, 23:8-10.

Streams, F.A., Shahjahan, M., LeMasurier, H.G. 1968. Influence of plants on the parasitization of the tarnished plant bug by Leiophron pallipes. J. Econ. Entomol. 61:996- 999.

100

Tilmon, K.J., Hoffmann, M.P. 2003. Biological control of Lygus lineolaris by Peristenus spp. in strawberry. Biological Control. 26:287-292.

Williams III, L., Price, L.D., Manrique, V. 2003. Toxicity of field-weathered insecticide residues to Anaphes iole (Hymenoptera: Mymaridae), an egg parasitoid of Lygus lineolaris (Heteroptera: Miridae), and implications for inundative biology control in cotton. Biological Control. 26:217-223.

Williams III, L., Rodriguez-Saona, C., Paré, P.W., Crafts-Brandner, S.J. 2005. The piercing-sucking herbivores Lygus hesperus and Nezara viridula induce volatile emissions in plants. Archives of Insect Biochemistry and Physiology. 58:84-96.

Williams III, L., Blackmer, J.L., Rodriguez-Saona, C., Zhu, S. 2010. Plant volatiles influence electrophysiological and behavioral responses of Lygus hesperus. J. Chem. Ecol. 36:467-478.

101

Figure 4.1: Erigeron annuus and Cotton Total Volatiles Mean total volatiles +/- standard error for Erigeron and cotton under different damage treatments. [A] both plant species undamaged, [B] both plant species damaged by a mixture of male and female adults, [C] both plant species damaged by 3-4th instar nymphs. In graphs * indicates significant difference between the two treatments at P<0.05.

102

Figure 4.2: Response of Adult Female Lygus Bugs to Erigeron and Cotton Volatiles Female adult Lygus behavioral responses to odor cues from Erigeron or cotton plants with the same damage treatments. Analysis by chi-square tests determined if the distribution of bug choices deviated from 50:50 (* indicates significant difference at P<0.05). The number of females used for each set of choice tests (and the number of plant pairs) was: adult damaged (female) n=112 (6 pairs), adult damaged (male) n=59 (4 pairs), nymph damaged n=43 (4 pairs), and undamaged n=75 (6 pairs).

103

Figure 4.3: Response of Adult Female Lygus Bugs to Clean Air and Plant Volatiles Female adult Lygus behavioral responses to odor cues from cotton vs. clean air and undamaged Erigeron vs. cotton with different damage treatments. Analysis by chi-square tests determined if the distribution of bug choices deviated from 50:50 (* indicates significant difference at P<0.05). The number of females used for each set of choice tests (and the number of plant pairs) was: undamaged cotton vs. clean air n=59 (2 plants), undamaged Erigeron vs. adult (F) damaged cotton n=65 (3 pairs), undamaged Erigeron vs. nymph damaged cotton n=37 (2 pairs).

104

Figure 4.4: Erigeron annuus and Cotton Total Volatiles During Y-tube Assays Mean total volatiles +/- standard error for Erigeron and cotton plants under different damage treatments during the Y-tube choice tests. [A] choice test using 40 actively feeding nymphs on a cotton plant vs. an undamaged Erigeron plant, [B] choice test using 12 actively feeding and ovipositing female Lygus bugs vs. an undamaged Erigeron plant (* indicates significant difference at P=0.05).

105 Table 4.1: Choice tests performed to examine movement of Lygus bugs in response to volatile cues.

106 Table 4.2: Mean amounts of each compound +/- standard error from undamaged, adult damaged, and nymph damaged Erigeron plants. Amounts are nanograms of volatiles/12 hour collection period. Yellow indicates compounds that are novel in a damage treatment relative to the undamaged treatment. Orange indicates compounds that are up-regulated (non-overlap of standard errors) relative to undamaged plant emissions. Blue highlights instances where an undamaged compound is absent or down-regulated in a damage treatment. Number of compounds: undamaged=93, adult damaged=114, nymph damaged=90

107 Table 4.2: Continued

108 Table 4.2: Continued

109

Table 4.3: Mean amounts of each compound +/- standard error from undamaged, adult damaged, and nymph damaged cotton plants. Yellow indicates compounds that are novel in a damage treatment relative to the undamaged treatment. Orange indicates compounds that are up-regulated (non-overlap of standard errors) relative to undamaged plant emissions. Blue highlights instances where an undamaged compound is absent or down-regulated in a damage treatment. Number of compounds: undamaged=30, adult damaged=34, nymph damaged=44

Chapter 5

Attraction of the parasitoid wasp, Peristenus relictus, to Erigeron annuus, a potential trap crop for Lygus bugs

Abstract

The field-edge weed Erigeron annuus is remarkably attractive to insects belonging to the genus Lygus, which in turn are a serious pest in many crop systems (e.g. strawberries, alfalfa, and cotton). Previous studies have demonstrated that Lygus will leave crop plants in order to arrest and feed on E. annuus, and that a compound-rich volatile blend released from the plants plays a strong role in this attraction. Therefore E. annuus shows potential as a trap crop for managing Lygus populations in the field, but the effects this plant has on natural enemies of Lygus bugs has not been fully explored.

Recently, a focus has been placed on the mass rearing of the non-native nymph parasitoid of Lygus, Peristenus relictus, and several successful trials utilizing this wasp to control

Lygus in alfalfa suggest that it may be effective in controlling Lygus bugs in other crops.

To understand how the presence of Erigeron annuus affects the host finding ability of

Peristenus relictus we carried out two-choice bioassays within a wind tunnel using E. annuus and cotton (Gossypium hirsutum), a crop host on which Lygus is a major pest.

Our study focused on comparing P. relictus responses to plants with and without damage when wasps had been given previous experience in ovipositing on nymph-damaged cotton plants. We found that cotton-experienced wasps overwhelmingly preferred the odors of Erigeron plants over cotton plants in most combinations, particularly Erigeron

111 plants that had received nymph damage. Examination of the volatile blends released by these plants showed that Erigeron releases almost all of the same compounds as cotton plants, both in the undamaged and damaged condition, with emissions of shared compounds from Erigeron plants often being enhanced relative to amounts emitted from cotton. Furthermore, a large number of additional compounds not present in the cotton blends (ether undamaged or damaged) are also released from Erigeron. The combination of compounds present in the nymph-damaged cotton blend (which wasps have learned) with the novel Erigeron compounds is highly attractive to P. relictus females and suggests that P. relictus will actively search out and visit patches of Erigeron, where they are likely to encounter high densities of hosts.

Introduction

Lygus bugs (Hemiptera: Miridae) are polyphagous insects that are considered important pests in many crop systems throughout the United States. Lygus pests are currently controlled primarily through the use of insecticides. In cotton (Gossypium hirsutum), growers have recently begun to reduce early season spraying for many non-

Lygus pests due in part to the success of boll weevil eradication programs as well as the development and widespread use of transgenic cotton that controls lepidopterous pests

(Layton 2000, Hardee et.al. 2001). Resistance to chemical controls has also begun to be reported in different Lygus populations. Lygus lineolaris has shown consistent ability to develop resistance to several different classes of insecticides (Zhu et. al. 2004, Zhu et. al.

2011). Widespread resistance to pyrethroids, organophosphate, and cyclodiene has also been reported (Snodgrass 1996), and varying levels of acephate resistance have been

112 identified in L. lineolaris populations in the area of Louisiana, Mississippi, and Arkansas

(Snodgrass & Scott 2000). The practice of reduced sprayings and the development of resistance to multiple classes of insecticides indicate a need for more sustainable Lygus controls measures.

By learning more about the ecology of Lygus and their behaviors associated with host plant location, a process that exploits these behaviors can be developed. Specifically, these behaviors can be used to develop a “trap crop” system (Pickett et al. 1997, Cook et al. 2007) that would pull Lygus out of an economically important crop field and into small patches of a preferred host. Furthermore, a thorough understanding of the mechanisms underlying the use of a particular trap crop will facilitate more precise recommendations on use. Ideally, an appropriate trap crop would also act to attract natural enemies into the area (Hokkanen 1991).

Previous work has demonstrated that the early sucessional Erigeron annuus is highly attractive to Lygus bugs. Researchers have shown that E. annuus plants in highway right-of-ways harbor significantly more adult Lygus lineolaris bugs than other co-occurring weeds, which indicates that adults are preferentially choosing pre-flowering and flowering Erigeron as a feeding and oviposition site (Fleischer and Gaylor 1987).

Another study used rubidium-marked individuals to observe the movement of Lygus through a cotton field and patches of E. annuus adjacent to the cotton (Fleischer et. al.

1988). When plots of E. annuus or mustard, pre-infested with marked L. lineolaris adults, located within the cotton field were destroyed, more marked individuals were recovered from the trap plots of E. annuus on the edge of the cotton field than within the crop itself

(Fleischer et. al. 1988). A cage study also demonstrated that when E. annuus was

113 available, Lygus bugs would leave cotton and migrate onto E. annuus (Fleischer et. al.

1988). Additionally, it has been shown that adult female Lygus rubrosignatus bugs in Y- tube bioassays will choose to move towards E. annuus plants over cotton plants under a variety of different damage regimes (Halloran et. al. 2012). Erigeron annuus shows strong potential as a trap crop that can ‘pull’ Lygus bugs off of adjacent field crops, however how it affects the host finding ability and effectiveness of natural enemies has not yet been explored.

Utilizing natural enemies alongside a trap crop system is not a new idea in the realm of IPM. Often times a trap crop will not only serve to pull pests out of a crop system, but will also confer some kind of benefit to natural enemies that leads to an increase in natural enemy population or increased predation/parasitization of the pest species (Hokkanen 1991). One of the most well known of these systems is the one developed for subsistence cereal farmers in sub-Saharan Africa (Pickett et. al. 1997,

Khan et. al. 2000). In this system maize or sorghum is intercropped with another plant

(such as Desmodium unicinatum, or Melinis minutiflora) that produces semiochemicals that are repellant to lepidopterous pests. Adjacent to this intercrop field a trap crop of

Napier grass (Pennisetum purpureum) or Sudan grass (Sorghum sudanensis) is planted.

These two plants are highly attractive to the pests and serve to pull them out of the cereal field. Additionally, the M. minutiflora also leads to an increase of parasitism of the

Lepidopteran pests (Khan et. al. 2000) The presence or absence of different weedy hosts can greatly increase or undermine the effectiveness of different natural enemies

(reviewed in Bottrell and Barbosa 1998). In Brazil, a trap crop system in soybean was paired with the inundative release of the egg parasitoid Trissolcus basalis (Wollaston)

114 (Hymenoptera: Scelionidae) leading to a nearly 60% decrease in the stink bug populations in the soybean fields (Corrêa-Ferreira et. al. 1996). Weedy or flowering plants grown adjacent to or intercropped alongside economically important plants can also serve as a reservoir for natural enemies to accumulate and then disperse into the crop fields. The trap crops can serve as either a refuge for the natural enemies during mowing or harvesting of the crop plants and also as a food or nectar source that the crop plant alone cannot provide (Lys et. al. 1994, Hickman and Wratten 1996). Previous work has shown that Erigeron species, including E. annuus, are highly preferred as a good-quality food source and oviposition site for a native parasitoid of Lygus bugs (Peristenus pseudopallipes) (Streams et al. 1968) and that increased densities of Erigeron result in increased parasitism rates within Erigeron patches (Shahjahan and Streams 1973). Thus,

E. annuus may fulfill a dual role in attracting and retaining Lygus bugs and attracting and provisioning Lygus natural enemies.

Biological control of Lygus bugs through the use of natural enemies has long been considered a viable method of reducing plant bug populations due to its durability in the face of resistance and compatibility with other low-spray control measures. However, native natural enemies of Lygus, including several parasitoids, have been shown, alone, to not provide adequate control in crop fields (Clancy and Pierce 1966, Day et.al. 1990,

Carignan et.al. 2007). Because of this, non-native parasitoids of tarnished plant bug nymphs were considered for inundative release with the goal of reducing Lygus populations in alfalfa fields (Day et.al. 1990, Day 1996). One of the most successful parasitoids chosen, Peristenus relictus, is a nymphal parasite of Lygus bugs that is native to Mediterranean regions of Europe. Within the past few years, a rearing program for this

115 wasp has begun and releases have been made in alfalfa in New Jersey and California with over-wintering success recorded in southern regions suitable for cotton growing (Pickett et.al 2007, Hoelmer et.al. 2008). Peristenus relictus is a promising candidate for biological control of Lygus in cotton, since it is native to regions that experience similar weather conditions and temperatures, and is already cleared for release in the U.S.

Additionally, rearing programs are established for P. relictus, so production and release is already streamlined. However, all current research on preferences and performance of P. relictus is limited to alfalfa, even though this wasp has the potential to control Lygus in many agricultural systems in the southern U.S.

It may be possible to combine the trap crop of Erigeron annuus with a targeted release of the introduced wasp, Peristenus relictus. Lygus bugs would be drawn to the

Erigeron, and then attacked by the P. relictus within the trap crop patches. Erigeron has been shown to release volatile blends that are highly complex relative to those emitted by crop plants such as cotton (sometimes more than 100 distinct compounds) (Halloran et al.

2012). This blend has been shown to be highly attractive to Lygus bugs alone and versus cotton. However, responses of Peristenus relictus to Erigeron plant volatiles have not been examined. Here we perform a series of choice test experiments to determine the relative attractiveness of Erigeron to P. relictus when presented as a choice against cotton, and examine differences in the volatile blends released from damaged and undamaged choice plants, in order to determine if the use of P. relictus is compatible with use of Erigeron as a trap crop.

116 Materials and Methods

Insect Rearing and Maintenance

Lygus rubrosignatus adults and nymphs were part of a laboratory colony originally obtained from The Phillip Alampi Beneficial Insect Laboratory, West Trenton,

NJ. The colony was kept in plastic tubs with shredded paper floors inside a rearing room maintained at 22±2°C, 60±10% RH, and a light-dark regimen of 14:10 hours (L:D). L. rubrosignatus were reared on parafilm packets containing Lygus artificial diet (Bio-serv), while adult females were allowed to oviposit on similar packets full of Carrageenan

Gelcarin GP812 (PhytoTechnology Laboratories, Shawnee Mission, KS). Diet packs were changed out every other day, while oviposition packs were removed when they became full of eggs. Oviposition packs were then placed into new tubs where the nymphs were allowed to hatch and grow.

Peristenus relictus wasps were obtained as adults from the Phillip Alampi

Beneficial Insect Research Lab (New Jersey Department of Agriculture) in West Trenton,

New Jersey where they are normally reared for commercial release. Upon arrival, wasps were maintained in a large screen cage with a dilute wasp diet consisting of honey and other essential nutrients (yeast extract) in the same temperature and humidity conditions as for the L. rubrosignatus bugs.

Plant Maintenance

Erigeron annuus seeds were collected from plants taken from edge-habitats around Centre County, Pennsylvania. Gossypium hirsutum (var. DPL90) seeds were

117 taken from a stock kept in the lab. All E. annuus and G. hirsutum plants used in the experiments were grown from seed in Metro-Mix potting soil with 5g of Osmocote Plus slow release fertilizer (Scotts) mixed in at the time of potting in 4-inch diameter pots and maintained at 25±1°C. They were grown in a pest-free greenhouse with a light:dark regimen of 16:8 hours and a relative humidity of 50±10%. Plants were bottom-watered daily with a hose. E. annuus plants were used for bioassays when they began to bolt, but before the buds opened while G. hirsutum plants were used once they began to develop squares.

Twenty-four hours before the start of the behavioral assays with the wasps, 3rd- instar L. rubrosignatus nymphs were placed in “clip cages” on the squares of G. hirsutum or the buds of E. annuus plants that would be used during the next day as choice plants.

These plants were cut stems placed in vials of water in order for them to fit into the wind tunnel. Additionally, to give the wasps the experience of stinging nymphs on G. hirsutum plants, a second G. hirsutum plant was also infested with 3rd-instar nymphs. The next morning the damaged portions of one of the G. hirsutum plants was cut off and placed in a large glass Petri dish with the nymphs still feeding. Female P. relictus wasps were then placed inside the dish and allowed to parasitize 2-5 nymphs in the presence of nymph damaged G. hirsutum plant volatiles. After 10 minutes of resting time, each female wasp was then flown in the wind-tunnel assay.

118 Parasitoid Host Choice Assay

Bioassays to determine up-wind, volatile-based orientation preferences of P. relictus females for E. annuus vs. G. hirsutum plants under different Lygus damage conditions were performed in a wind tunnel (Table 5.1). The wind tunnel was a .61 x .61 x 1.83m "push-pull" acrylic glass wind tunnel with a charcoal-filtered air input and a wind velocity of 0.5 m/s. The wind tunnel was maintained at 25-29°C and 50% relative humidity during the bioassays. Bioassays were carried out during the months of

September through October of 2011 between noon and 5:00 p.m.

A single E. annuus and G. hirsutum plant were placed equidistant from the side- walls and each other as choice plants in the upwind position, while 1m downwind a starting G. hirsutum plant infested with a single Lygus nymph was placed. At the beginning of the assay a single female P. relictus was placed on the starting G. hirsutum plant and was allowed to patrol the area and orient upwind. Where the female flew to was then recorded, and they were considered to have made a choice if they landed on either of the two plants upwind. Some individuals landed on the walls or ceiling before re- orienting and then flew to the choice plants, and these were also recorded as choices. If a wasp flew to the wall or ceiling and then did not reorient and fly upwind within 30 seconds then it was recorded as a ‘no choice’. The wasps were allowed 10 minutes to take flight before they were removed. Wasps removed from the starting location in this manner were not recorded as ‘no choice’ as it usually meant that the conditions of the wind tunnel were not conducive to flight on that day.

In the first set of trials the E. annuus plants were undamaged while the G. hirsutum plants were damaged by 16 3rd-instar L. rubrosignatus nymphs. In a second trial

119 the damage treatments was reversed, with the G. hirsutum plant being left undamaged, and the E. annuus plant being damaged by Lygus nymphs, followed by both plant species being Lygus damaged, and finally neither plant species being Lygus damaged.

Volatile Collections and Analysis

To quantify and identify VOCs released by the study plants, the headspace around the plants was collected using a push/pull design. An automatic volatile collection system built by Analytical Research Systems (Gainsville, FL) was used to control sampling periods. The system was capable of controlling simultaneous collections from 12 individual treatment plants at any chosen set of intervals 24 hours a day.

A portion of each plant was excised with clean scissors and placed in a flask of water. The top portions of these excised plants were enclosed in a 7L glass bell jar with a metal guillotine-type base. Clean cotton-balls were packed around the junction of the stem and base to prevent air from lower portions of the plants entering the bell jars. The portion of the plant being collected from was a single stem with buds for E. annuus and the apical portion of the main stem with three developing squares for G. hirsutum.

Charcoal-purified air was pumped through the top of the bell jars at a rate of 4L min-1and allowed to pass over the plant before being pulled out through the bottom of the jars at a rate of 1L min-1. This air was pulled through filters containing approximately 45mg of

Super-Q adsorbent (80/100 mesh, Alltech).

The plants (6 Erigeron and 6 cotton) were divided into two treatments during the collections (1) control (undamaged), and (2) damage from L. rubrosignatus nymph

120 feeding. For treatments involving insect damage 16 L. rubrosignatus nymphs were placed in the bell jars 24 hours before the start of the collection. All insects remained in the bell jars and were allowed to constantly feed over the whole duration of the collections. The collections were checked several times a day and any dead insects were aspirated out and replaced with new individuals.

The collections were sampled for six hours from 12:00 to 18:00 using one super-

Q filter per bell jar. This corresponds with the period of time that behavioral assays with wasps were conducted. All collections occurred within a pest-free greenhouse maintained at 25°C, 50%RH and a light:dark cycle of 16:8 hours during the month of August in

2012.

Samples were analyzed by first eluting the compounds off of each filter using

125μL of 1:1 dichloromethane:hexanes (Burdick & Jackson High Purity, and J.T. Baker

95% Purity respectively) and then adding 200ng n-octane and 400ng nonyl acetate to each eluted sample to act as internal standards. The eluted samples were then analyzed on an Agilent 6890 analytical gas chromatograph (GC) equipped with a flame ionization detector, with a splitless injector, and a HP-1 column (15m x 0.25mm x 0.25μm,

Agilent). In order to identify compounds, selected samples were analyzed on an Agilent

6890N GC equipped with an Agilent 5973N mass selective detector configured for electron impact mode and a HP-1MS column (30m x 0.25mm x 0.25μm, Agilent). Mass spectra were compared to spectra for standards available in the National Institute of

Standards and Technology (NIST) library as well as known standards from the lab. The

Kovats Index (KI) of each compound was also determined (Kováts 1965) and used to

121 identify some compounds by comparing the unknown sample KIs to those of known standards run on the same type of GC column (HP-1).

Statistical analysis

Total volatile emissions over the daytime period were calculated for each replicate plant by summing the amounts of each recorded compound across the three samples collected over 6 hours (amounts calculated relative to the internal standard nonyl acetate based on peak area). Total volatiles emitted by cotton and Erigeron were compared within each bug damage treatment using GLM (Minitab) with plant species as the main factor. Compound means and standard errors were also calculated and each damage treatment was compared to the same treatment for the other species to determine if a compound was novel, up-regulated, or down-regulated. Behavioral tests were analyzed using chi-square tests.

Results

Parasitoid Host Choice Assay

Despite only receiving oviposition experience on nymph damaged G. hirsutum plants, female P. relictus wasps still flew upwind towards E. annuus plants in several of the damage treatments (Fig. 5.1). When both choice plants were left undamaged a little over 70 percent of the parasitoids flew upwind towards Erigeron. Although this percentage was trending towards significance it was not significantly different from those that went to the undamaged cotton plants (Chi squared equals 2.882, df = 1, p = 0.089).

However, when both plants were damaged by Lygus nymphs, parasitoids were

122 significantly more attracted to the E. annuus plants (Chi squared equals 8.048, df = 1, p =

0.005), and when nymph damaged Erigeron was compared to undamaged cotton plants

100% of the females wasps flew to the E. annuus plants (Chi squared equals 14.000, df =

1, p = 0.0002). In contrast, when only the cotton plants were damaged by L. rubrosignatus nymphs there was no significant difference between the number of wasps that flew to either choice plant (Chi squared equals 0.429, df = 1, p = 0.5127).

Volatile Collections and Analysis

Both cotton and Erigeron annuus release a blend of compounds constitutively when the plants are undamaged and release a greater amount of compounds when damaged by Lygus nymph feeding (Fig. 5.2). In both damage treatments E. annuus plants release a far greater total amount of volatiles than the cotton plants (Fig. 5.2) as well as a greater amount of unique compounds (Table 5.1). Table 5.1 displays the compounds emitted by cotton and E. annuus across the undamaged and nymph damaged treatments, as well as the mean and standard error for each compound within each treatment. In addition to an increase in the number of compounds (from 53 to 119), many compounds that are released constitutively in low amounts are up-regulated when E. annuus is damaged by Lygus nymphs. In contrast, most of the increase in total volatiles being produced by damaged cotton plants is due to the induction of new compounds not found in the undamaged treatment. In the undamaged treatment, all but two compounds

(unknown KI 888, and 1-methyl-2-(1-methylethyl)-benzene KI 1011) that are produced by undamaged cotton are also produced by undamaged E. annuus plants. When damaged by nymphs, E. annuus releases the same compounds as damaged cotton (missing only

123 unknown KI 823, unknown KI 888, and 1-methyl-2-(1-methylethyl)-benzene KI 1011) and of those that are shared, E. annuus releases many in higher amounts than in cotton.

Comparing undamaged E. annuus to damaged cotton, many compounds produced by the damaged cotton (except for the three compounds mentioned above) are shared by the undamaged E. annuus plants. Overall, the volatile blend produced by both undamaged and nymph damaged Erigeron annuus plants matches the blends produced by nymph damaged cotton plants, and also produce additional unique compounds not found in the cotton blend.

Discussion

The results of the wind tunnel bioassays demonstrate that Erigeron annuus is highly attractive to female Peristenus relictus wasps. In every treatment the E. annuus plants attracted a large proportion of the responding female wasps, even when the

Erigeron plant was undamaged and the cotton plant was damaged by nymphs (Fig. 5.1).

This was the case despite the fact that these wasps had been allowed previous oviposition experience only on Lygus nymph damaged cotton. A large body of previous literature

(reviewed in Vet et. al. 1995) has indicated that parasitoid wasps are capable of associative learning. When a wasp finds a suitable host and oviposits in that host in the presence of certain cues those cues will be associated with hosts and wasps will respond positively to those cues while foraging (Lewis and Tumlinson 1988, Papaj et. al. 1994,

Vet et. al. 1995). These cues often include odors that the wasp encounters during the oviposition event (Lewis and Tumlinson 1988, Lewis and Takasu 1990). These odors could be produced by their host, or by the surrounding plant material. Parasitoid wasps

124 have been shown to be capable of developing these long term memories after only a single experience (Smid et. al. 2007), and can maintain different odor memories for locating food sources or suitable hosts (Lewis and Takasu 1990).

We chose to focus on the odors produced by the plants themselves within each damage treatment to understand why E. annuus showed such attractiveness to wasps given experience on a completely different plant. Erigeron plants damaged by L. rubrosignatus nymphs were highly attractive and a significant percentage of P. relictus wasps chose to move to these plants over cotton, even when the cotton plants were also nymph damaged (Fig 5.1.). Examining the volatile blends of the plants from this treatment (Table 5.1) shows that damaged E. annuus plants release many of the same compounds as damaged cotton, and release a large number of additional compounds, some of which are in very high amounts. The addition of these other compounds to the blend already released by cotton (those compounds shared between the two treatments) does not seem to interfere with the ability of the wasps to respond to the blend it has learned in the presence of hosts, and may enhance the attractiveness of E. annuus relative to cotton.

In the choice tests, undamaged E. annuus plants were also attractive to the P. relictus females, as roughly half of the wasps flew to undamaged E. annuus even in the presence of a nymph infested cotton plant. Comparing undamaged E. annuus to damaged cotton, many compounds found in the damaged cotton (except unknown compound KI

823, unknown compound KI 888, 1-methyl-2-(1-methylethyl)-benzene KI 1011, and unknown compound KI 1500) are shared by the undamaged Erigeron. However, the undamaged blend of E. annuus has fewer compounds overall than the damaged E. annuus

125 blend, which may indicate that these extra compounds are part of what makes the damaged Erigeron annuus plants so attractive. Overall, it seems as if the same compounds the parasitoids are encountering on the damaged cotton when they are receiving their oviposition experience are also being produced by Erigeron annuus plants, sometimes in even greater amounts and even when the plant is undamaged (Table

5.1).

The situation we have observed here is similar to what was reported by Khan et. al. (2000) regarding the plant, Melinis minutiflora. In this case the constitutive volatiles produced by the M. minutiflora were similar to the induced volatiles produced by pest- damaged maize. Khan et. al. (2000) theorized that this compound overlap (which had already been shown to be attractive to parasitoids) would act to repel lepidopterous pests, and this theory was proven correct as Melinis minutiflora intercropped amongst cereal fields does reduce the population of stem borers in those fields. The repellant effect of these odors towards female Lepidoptera searching for oviposition sites could have been due to the females attempting to avoid areas that are being foraged by parasitoids, or that may already have high amounts of colonization by conspecific larvae. Volatiles from plants damaged by herbivores repelling ovipositing female Lepidoptera has been shown in other cases as well, such as the night-time volatiles from Heliothis virescens damaged tobacco plants being repellent to females searching for oviposition sites (De Moraes et. al. 2001).

Unlike these examples, though, Lygus bugs seem to be attracted to plants containing conspecifics, or which were recently damaged by conspecifics. Blackmer et. al. (2004) demonstrated in Y-tube bioassays that L. hesperus adults are attracted to alfalfa

126 plants that either contained conspecifics, or had been recently damaged by conspecifics that were then removed before the trial. Additionally, in another Y-tube study it was found that L. rubrosignatus adult females will move towards Erigeron annuus and cotton plants that are damaged by nymph and adult stage L. rubrosignatus bugs (Halloran et. al.

2012). In fact, this same study showed results with a 50/50 split between female Lygus bugs visiting undamaged E. annuus and adult damaged cotton (Halloran et. al. 2012); lending further evidence that undamaged E. annuus may be producing a volatile signal that is similar to that produced by Lygus damaged cotton.

Previous research has shown that Erigeron is a preferred host that arrests Lygus bugs (Fleischer and Gaylor 1987, Fleischer et. al. 1988). This attraction may be due to the disproportionately large and diverse volatile signal being released from the plant, which may in turn resemble the volatile blend produced by Lygus damaged plants. Our research indicates that female P. relictus wasps also show a strong preference for

Erigeron over similarly damaged cotton plants, and will show this preference even after receiving oviposition experience on cotton. These results, considered alongside previous work showing that E. annuus serves as a high quality nectar source for other, native,

Peristenus species (Streams et. al. 1968), indicates that E. annuus may fulfill a dual role by attracting and retaining Lygus bugs outside of crop fields, and attracting and provisioning P. relictus wasps released into the area.

127 References

Blackmer, J.L., Rodriguez-Saona, C., Byers, J.A., Shope, K.L., Smith, J.P. 2004. Behavioral response of Lygus hesperus to conspecifics and headspace volatiles of alfalfa in a Y-tube olfactometer. Journal of Chemical Ecology. 30:1547-1564.

Bottrell, D.G., Barbosa, P. 1998. Manipulating natural enemies by plant variety selection and modification: a realistic strategy? Annu. Rev. Entomol. 43:347-367.

Carignan, S., Stewart, R.K., Godin, C., Bolvin, G. 2007. Parasitism activity of Peristenus spp. (Hymenoptera: Braconidae) on Lygus lineolaris (Hemiptera: Miridae) nymphs prior to the establishment of Peristenus digoneutis in southwestern in Quebec. Biocontrol Science and Technology. 17:623-633.

Clancy, D.W., Pierce H.D. 1966. Natural Enemies of Some Lygus Bugs. J. Econ. Entomol. 59: 853-858.

Cook, S.M., Khan, Z.R., Pickett, J.A. 2007. The use of push-pull strategies in integrated pest management. Ann. Rev. Entomol. 52:375-400.

Corrêa-Ferreira, B.S., Moscardi, F. 1996. Biological control of soybean stink bugs by inoculative releases of Trissolcus basalis. Entomologia Experimentalis et. Applicata. 79:1-7.

Day, W.H., Hedlund, R.C., Saunders, L.B., Coutinot, D. 1990. Establishment of Peristenus digoneutis (Hymenoptera: Braconidae), a parasite of the Tarnished Plant Bug (Hemiptera: Miridae), in the United States. Ann. Entomol. Soc. Am. 85:482-488.

Day, W.H. 1996. Evaluation of biological control of the Tarnished Plant Bug (Hemiptera: Miridae) in alfalfa by the introduced parasite Peristenus digoneutis (Hymenoptera: Braconidae). Environ. Entomol. 25:512-518.

De Moraes, C.M., Mescher, M.C., Tumlinson, J.H. 2001. Caterpillar-induced nocturnal plant volatiles repel conspecifics females. Nature. 410:577-580.

Fleischer, S.J., Gaylor, M.J. 1987. Seasonal abundance of Lygus lineolaris (Heteroptera: Miridae) and selected predators in early season uncultivated hosts: Implications for managing movement into cotton. Environmental Entomology. 16:379-389.

Fleischer, S.J., Gaylor, M.J., Hue, N.V. 1988. Dispersal of Lygus lineolaris (Heteroptera: Miridae) adults through cotton following nursery host destruction. Environmental Entomology. 17:533-541.

128 Halloran, S.T., Mauck, K.E., Tumlinson, J.H. 2012. Chemical ecology of Erigeron annuus: a weedy host and trap crop of bugs in the genus Lygus. See chapter 4. pages 83- 106, this document

Hardee, D.D., Van Duyn, J.W., Layton, M.B., Bagwell, R.D. 2001. Bt Cotton & Management of the Tobacco Budworm-Bollworm Complex. U.S. Department of Agriculture, Agricultural Research Service, ARS–154. 40 pp.

Hickman, J. and Wratten, S. 1996. Use of Phacelia tanacetifolia strips to enhance biological control of aphids by hoverfly larvae in cereal fields. Journal of Economic Entomology. 89:832-840.

Hoelmer, K., Fuester, R., Taylor, P., Coutinot, D., Dorsey, T., Hudson, W., Robbins, G., Peiffer, R., Barczewski, R. 2008. Introduction of a European parasitoid, Peristenus relictus (=P. stygicus), against tarnished plant bug in the mid-Atlantic states. Journal of Insect Science. 8:49.

Hokkanen, H.M.T. 1991. Trap cropping in pest management. Annu. Rev. Entomol. 36:119-138.

Khan, Z.R., Pickett, J.A., van den Berg, J., Wadhams, L.J., Woodcock, C.M. 2000. Exploiting chemical ecology and species diversity: stem borer and striga control for maize and sorghum in Africa. Pest Management Science. 56:957-962.

Kováts, E. “Gas chromatographic characterization of organic substances in the retention index system.” Advances in Chromatography, Vol. 1. Ed. Giddings, J.C., Keller, R.A. New York. Marcel Dekker, 1965. 229-247.

Layton, M.B. 2000. Biology and Damage of the Tarnished Plant Bug, Lygus lineolaris, In Cotton. Southwest. Entomol. Suppl. 23:7-20.

Lewis, W.J., Takasu, K. 1990.Use of learned odours by a parasitic wasp in accordance with host and food needs. Nature. 348: 635-636.

Lewis, W.J., Tumlinson, J.H. 1988. Host detection by chemically mediated associative learning in a parasitic wasp. Nature. 331:257-259.

Lys, J., Zimmermann, M., Nentwig, W. 1994. Increase in activity density and species number of carabid beetles in cereals as a result of strip-management. Entomologia Experimentalis et. Applicata. 73: 1-9.

Papaj, D.R., Snellen, H., Swaans, K., Vet, L.E.M. 1994. Unrewarding experiences and their effect on foraging in the parasitic wasp Leptopilina heterotoma (Hymenoptera: Eucoilidae). Journal of Insect Behavior. 7:465-481.

129 Pickett, C.H., Rodriguez, R., Brown, J., Coutinot, D., Hoelmer, K.A., Kuhlmann, U., Goulet, H., Schwartz, M.D., Goodell, P.B. 2007. Establishment of Peristenus digoneutis and P. relictus (Hymenoptera: Braconidae) in California for the control of Lygus spp. (Heteroptera: Miridae). Biocontrol Science and Technology. 17:261-272.

Pickett, J.A., Wadhams, L.J., Woodcock, C.M. 1997. Developing sustainable pest control from chemical ecology. Agriculture, Ecosystems and Environment. 64:149-156.

Shahjahan, M., Streams, F.A. 1973. Plant effects on host-finding by Leiophron pseudopallipes (Hymenoptera: Braconidae), a parasitoid of the tarnished plant bug. Environ. Entomol. 2:921-925.

Smid, H.M., Wang, G., Bukovinszky, T., Steidle, J.L., Bleeker, M.A.K., van Loon, J.J.A., Vet, L.E.M. 2007. Species-specific acquisition and consolidation of long-term memory in parasitic wasps. Proc. R. Soc. B. 274:1539-1546.

Snodgrass, G.L. 1996. Pyrethroid resistance in field populations of the tarnished plant bug (Heteroptera: Miridae) in cotton in the Mississippi Delta. J. Econ. Entomol.89:783- 790.

Snodgrass, G.L., Scott, W.P. 2000. Seasonal changes in pyrethroid resistance in tarnished plant bug (Heteroptera: Miridae) populations during a three-year period in the delta area of Arkansas, Louisiana, and Mississippi. J. Econ. Entomol. 93:441-446.

Streams, F.A., Shahjahan, M., LeMasurier, H.G. 1968. Influence of plants on the parasitization of the tarnished plant bug by Leiophron pallipes. J. Econ. Entomol. 61:996- 999.

Vet, L.E.M., Lewis, W.J., Cardé, R.T. “Parasitoid Foraging and Learning.” Chemical Ecology of Insects 2. Ed. Cardé, R.T., and Bell, W.J. New York: Chapman & Hall, 1995. 65-104.

Zhu, Y.C., Snodgrass, G.L., Chen, M.S. 2004. Enhanced esterase gene expression and activity in a malathion-resistant strain of the tarnished plant bug, Lygus lineolaris. Insect Biochemistry and Molecular Biology. 34:1175-1186.

Zhu, Y.C., West, S., Snodgrass, G., Luttrell, R. 2011. Variability in resistance-related enzyme activities in field population of the tarnished plant bug, Lygus lineolaris. Pesticide Biochemistry and Physiology. 99:265-273.

130

Figure 5.1: Response of Peristenus relictus Females to Plant Volatiles Percent response of female Peristenus relictus wasps to volatiles from undamaged and Lygus nymph damaged Erigeron annuus (left) and Gossypium hirsutum (right) plants. All wasps were given experience only on nymph damaged cotton and were started on a cotton leaf downwind of the choice plants. The * symbol indicates a significant preference at P<0.05 (Chi-squared test). The number of female wasps used for each set of choice tests was: undamaged E. annuus vs. undamaged cotton n=17, nymph-damaged E. annuus vs. nymph-damaged cotton n=21, undamaged E. annuus vs. nymph-damaged cotton n=29, nymph-damaged E. annuus vs. undamaged cotton n=22.

131

Figure 5.2: Erigeron annuus and Cotton Total Volatiles Mean total volatiles +/- standard error for Erigeron and cotton under different damage treatments. [A] both species undamaged, [B] both species damaged by 16 Lygus rubrosignatus nymphs. In graphs * indicates significant difference between the two treatments at P<0.05

132 Table 5.1: Mean amounts of each compound +/- standard error from undamaged, and nymph damaged Erigeron and cotton plants. Volatiles (ng/6hr) emitted from the different cut plant treatments used in the choice tests. Pairs of treatments are matched based on the color of the shading in the columns. Within each pair, bold values represent compounds that are novel in that treatment relative to its pair, and bold values in red represent compounds that are shared but higher in that treatment relative to its pair (based on non-overlap of two times each standard error surrounding the means). Compounds with a (*) beside them have had identities confirmed using standards.

133 Table 5.1: Continued

134 Table 5.1: Continued

.

Chapter 6

Conclusions

The main goal of this dissertation was to increase our knowledge of the chemical ecology of Lygus-host interactions to apply this knowledge to improving Lygus control in various agricultural contexts. To accomplish this goal two specific objectives were developed. The progress, conclusions, and future directions for each objective are outlined below.

Objective 1) Characterize the defense responses of cultivated and weedy legume hosts to feeding damage by different species and life stages of Lygus bugs.

We hypothesized that different Lygus species (L. rubrosignatus and L. lineolaris) and life stages (nymph vs. adult) would induce qualitatively or quantitatively different responses in Medicago sativa (alfalfa) and Melilotus officinalis (sweet clover). This was due to previous research demonstrating that different species of Lygus have multiple genes coding for polygalacturonase enzymes (PGs) that are distinct between species

(Allen and Mertens 2008, de la Paz Celorio-Mancera et. al. 2008, de la Paz Celorio-

Mancera et. al. 2009), and which are thought to be important in determining the magnitude and nature of the plant’s response to feeding (D’Ovidio et al. 2004, Di et. al.

2006). Additionally, a preliminary study which performed a semi-quantitative PCR analysis of PG expression in L. lineolaris did find variation in expression between life stages (nymphs vs. adults) and adult sexes for multiple PGs (Allen and Mertens 2008).

136 Unlike with chewing herbivores, plants may recognize and respond to attack from piercing-sucking herbivores at a gene-for-gene level similar to how plants recognize pathogen infection, so these differences in salivary enzymes may trigger different defense responses by the plants. Alternatively, defense responses of each host could be similar, reflecting the induction of similar legume biochemical pathways in response to these salivary elicitors. To test this hypothesis, we collected and analyzed the volatile organic compounds released from M. sativa and M. officinalis plants either undamaged or damaged by adults or nymphs of both species of Lygus bug. We also measured several phytohormones associated with plant responses to wounding/herbivory (jasmonic acid), pathogen attack (salicylic acid), or abiotic stresses (abscisic acid) under these same damage treatments.

Overall, we did see differences in the plant response to feeding by adults and nymphs of these two species of Lygus. In alfalfa, L. lineolaris nymphs strongly induce

GLVs, terpenes, and the aromatic volatile methyl salicylate (Figs. 2.1-2.3) while L. rubrosignatus nymph-damaged plants do not show the same induction of the above mentioned volatiles. The L. lineolaris nymph treatment showed significantly higher induction of E-beta ocimene relative to control and adult L. lineolaris treatments, and the compounds alpha-farnesene and methyl salicylate were both only induced by L. lineolaris nymphs (Figs. 2.2&2.3). Species level differences were also evident where L. lineolaris adults differed from L. rubrosignatus adults, and L. lineolaris nymphs differed from L. rubrosignatus nymphs, both along the PC1 axis of the GLV analysis (Fig. 2.1A). When looking at the phytohormone levels in alfalfa, we also saw species level differences, although for the most part these differences were only seen in tissue collected at the

137 feeding site. Jasmonic acid was higher in L. rubrosignatus treatments relative to L. lineolaris treatments for the site of feeding on M. sativa, and the precursor for JA production, linolenic acid, was also higher for L. rubrosignatus treatments on this plant

(Fig. 3.1A&C). In contrast, abscisic acid was only induced to high levels in L. rubrosignatus damaged alfalfa (Fig. 3.1D).

Initially it appears that the volatile and phytohormone results conflict with each other. For example, L. lineolaris nymphs induced significantly higher levels of several terpenes and the aromatic methyl salicylate, relative to all other treatments but induced lower levels of the phytohormones JA and linolenic acid relative to L. rubrosignatus bugs. Lygus rubrosignatus nymphs also induced lower emissions of several common

GLVs and terpenes relative to all other damage treatments despite the high amounts of JA and linolenic that were found in the tissue of alfalfa plants damaged by them. This is a conflict since higher JA induction is usually associated with higher emissions of terpenes

(Schmelz et. al. 2003). However, one possible explanation for these results is the high levels of ABA induced by L. rubrosignatus feeding. ABA is a phytohormone that is produced by plants in response to water stress in order to mediate closing of the stomata.

There is some evidence that plants may release terpenes through open stomata, and that closure of the stomata may lead to a decrease in the amount of volatile terpenes released into the surrounding headspace (personal communication, Irmgard Seidl-Adams). The high levels of ABA induced by L. rubrosignatus feeding may be causing stomata closure in M. sativa, which in turn is preventing the release of the terpenes being induced by the high JA levels. A way to test this theory is through extracting terpenes directly from the

138 plant tissue. This would show whether there was a build-up of terpenes inside the alfalfa tissue that was not being measured in the volatile collections.

In M. officinalis, species-level differences were more pronounced for terpene emissions (Fig. 2.5). Several terpenes were only induced by L. rubrosignatus, with few differences between life stages (Fig. 2.5C). Additionally, methyl salicylate once again was only induced by L. lineolaris and several unknown compounds were only induced by

L. rubrosignatus (Fig. 2.6C). Looking at the phytohormone levels, M. officinalis had higher JA and ABA induction being associated with only L. rubrosignatus adults instead of L. rubrosignatus treatments in general (Fig. 3.3), but only L. lineolaris damaged plants showed an upregulation of linolenic acid. Lygus lineolaris adults induced the production of JA, but the nymphs showed no induction of the phytohormone relative to control plants.

Again, the volatile and phytohormone results seem to be in conflict. For example,

L. lineolaris nymphs did not induce JA at all, even though JA-based volatiles (terpenes) were clearly induced by this treatment. Lygus rubrosignatus also showed no induction of linolenic acid (a precursor to both JA and GLVs), yet adults and nymphs induced somewhat higher levels of Z-3-hexen-1-ol and Z-3-hexenyl acetate relative to L. lineolaris.

This lack of correlation between phytohormone results and the previously observed volatile results indicates that there may be more complex signaling differences occurring in response to feeding by these two species of Lygus bugs that cannot be fully characterized by simply measuring phytohormone levels. An analysis of downstream products (e.g. genes regulated by JA and SA pathways, including genes for the synthesis

139 of terpenes and in-plant defenses such as saponins) could provide an explanation for the differences observed in indirect defenses induced by Lygus feeding. Use of the model legume, Medicago truncatula for these studies would enhance the genetic resources available for examining downstream defense gene induction. One possible way that these genes could be up-regulated in the absence of JA is through the action of the PGs present in the saliva of the different species and life stages, which may be generating oligogalacturonide fragments within plant tissue that are direct elicitors of later portions of defense pathways. For instance, previous studies have shown that some of the downstream products of JA up-regulation are plant-produced oligogalacturonide fragments that participate in activation of other defenses (Gatehouse 2002). These OG fragments have been shown to activate downstream defense genes, and if they are applied to plants exogenously they trigger these defenses without an upregulation of either JA or

SA (Leitner et. al. 2008). Lygus feeding in some cases may generate similar OG fragments that may cause activation of volatiles or other defenses in the absence of a JA burst. This possibility could be investigated by combining the above-mentioned downstream gene expression study with an analysis of Lygus damaged tissue that examines what OG fragments are different between species and relative to controls (e.g., using methods like those in Schols et. al. 2000).

140 Objective 2) Determine the suitability of the common edge weed Erigeron annuus for use as a trap crop by characterizing volatile profiles (Erigeron and crop plants) and utilizing these volatiles in choice tests with Lygus adults as well as with females of the Lygus natural enemy, Peristenus relictus.

We hypothesized that the preference of Lygus bugs for E. annuus in the field could be explained, at least in part, by volatile cues being given off by the plant. To test this hypothesis the volatile blends from undamaged E. annuus plants and plants subjected to a number of different Lygus feeding damage regimes were collected and analyzed.

These collections showed that Erigeron annuus releases a highly complex blend of volatile organic compounds constitutively while Lygus feeding induced both an upregulation of these compounds as well as the production of novel compounds (Table

4.2). We then tested whether these volatiles were attractive to Lygus by carrying out Y- tube bioassays in which the only stimulus the bugs had access to was the plant volatile blend.

In every case, a significant number of Lygus females chose E. annuus odors over the odors of a similarly damaged cotton plant (Fig. 4.2). Additionally, by manipulating the damage levels of the E. annuus and cotton plants, we were able to demonstrate that this attraction is not just a product of the pure volume of volatiles released by E. annuus

(which is often quite high compared to crop hosts), but rather is due to some qualitative features within the blend (Fig. 4.3). Further work within this portion of the project could include examining which aspects of the E. annuus blend are most attractive to Lygus bugs. However, for herbivores, many volatile attractants are not made up of an individual compound, but rather combinations of compounds in key ratios. The highly complex

141 blend released from E. annuus (sometimes over 100 compounds, with many compounds being released at high levels) could make this task extremely difficult. One possible approach would be to examine Lygus antennal responses to the plant-collected Erigeron annuus blend through the use of coupled gas chromatography-electroantennographic detection (GC-EAD), which may narrow down which compounds within the blend are being detected by Lygus bugs.

The second question we asked was whether the attractiveness of Erigeron annuus could be used as a means of creating a trap crop that would pull Lygus bugs away from economically important crops. To test this we ran duel-choice bioassays using adult female Lygus rubrosignatus (Chapter 4) or adult female Peristenus relictus wasps

(Chapter 5) to understand their orientation behavior in response to volatiles from both E. annuus and cotton plants. Lygus females were specifically used because it is the first generation of nymphs that cause economic damage on crop plants, so the plants that females choose for oviposition largely determine where nymph feeding will be concentrated. Lygus females showed a strong preference for the volatiles of E. annuus plants across all damage treatments when given the choice between E. annuus or cotton.

In the bioassays involving P. relictus females we found similar attraction to E. annuus by the wasps (Fig. 5.1). When both cotton and E. annuus were damaged by Lygus nymphs, the wasps were more attracted to the E. annuus, and the same was true when only E. annuus was damaged by nymphs. Even when both plants were left undamaged roughly

70 percent of the P. relictus females flew toward the E. annuus plant over the cotton plant, and this trend was very near significance, suggesting that this plant on its own may attract natural enemies of Lygus even before nymphs appear. Finally, when E. annuus

142 was left undamaged and cotton plants were nymph damaged there was no significant difference in the attraction of the parasitoids to either plant. The results of these wind tunnel bioassays demonstrate that Erigeron annuus is highly attractive to female

Peristenus relictus wasps. In every treatment the E. annuus plants attracted a large proportion of the female wasps, even when the Erigeron plant was undamaged and the cotton plant was damaged by nymphs.

To better understand what the attractant quality of E. annuus might be, we compared the volatile blends being released from undamaged and Lygus damaged

Erigeron annuus with those being released from similarly damaged cotton. This was done with whole plants in Chapter 4 (discussed above), and again with plants cut off at the stem in Chapter 5 (since cut plants were necessary for wind-tunnel bioassays). We found in both cases that, relative to cotton, E. annuus plants release both more compounds and greater amounts of the same compounds, and that E. annuus releases between 5-10 times more volatiles than a cotton plant undergoing the same treatment

(even though due to their growth habit, Erigeron plants have a smaller surface area of leaves than the large-leafed cotton plants) (Figs. 4.1 & 5.2). As discussed in chapter 5, an examination of the compounds within these volatile blends shows that most of the compounds being released by damaged cotton plants are also released constitutively by

E. annuus (see Table 5.1). Previous research on Lygus host choice has shown that Lygus bugs are attracted to Medicago sativa plants that contain conspecifics or that have been recently fed upon by conspecifics (Blacker et. al. 2004). Many of the compounds that

Blackmer et al. (2004) mentioned as being attractive to Lygus hesperus in their study are released constitutively by E. annuus plants, and this may explain in part the attractive

143 nature of the plant. The parasitoid wasps, when foraging for oviposition hosts, are likely to be attracted to volatiles associated with their host organism. In chapter 5, these wasps were given oviposition experience in the presence of volatiles from Lygus nymph- damaged cotton plants. These experiences may have created an association between

Lygus nymphs and the odors released from damaged cotton that are also released constitutively from E. annuus. This may explain the wasps being equally attracted to undamaged E. annuus as they were to nymph damaged cotton. However, this does not fully explain the greater attractiveness of nymph-damaged E. annuus, which releases many novel compounds that are not released by damaged cotton plants. While the greater attraction to E. annuus plants may be partially explained by these plants releasing greater amounts of the compounds already present in the damaged cotton blend, our results suggest that it may also be advisable to test naïve wasps, which have not had oviposition experience in the presence of any volatile cues, in order to see if there is an innate attractive feature of E. annuus over cotton.

Overall, chapters 4 &5 demonstrate that Erigeron annuus is a highly attractive plant for both Lygus bugs and their introduced natural enemy Peristenus relictus. The large amount of total volatiles produced by the plants suggests that could be very useful as a trap crop planted in small patches that can act as potent point-sources for continuous volatile release (similar to the synthetic volatile emitters currently used in other odor- based insect controls). Even relatively small patches of E. annuus located near larger cotton fields could still be detected and located by Lygus bugs or parasitoids. The plant also seems to remain highly attractive across a wide variety of damage treatments, so even during the portion of the growing season when some Lygus have migrated onto

144 cotton plants the patches of Erigeron will still be more likely to attract the majority of new female Lygus migrating into the field. Additionally, cutting of the plant after the initial flower bolt results in growth of many new flowering shoots from axillary positions, and continues throughout the season when plants are grown outdoors (personal observation). These results alongside previous work showing that E. annuus flowers can serve as an excellent food source for native Peristenus wasps (Streams et. al. 1968), indicates that E. annuus may fulfill a dual role by attracting and retaining Lygus bugs outside of crop fields, and attracting and provisioning P. relictus wasps released into the area. Future work should examine this possibility with field studies examining the density of Lygus populations in cotton fields that are either adjacent or not adjacent to patches of

E. annuus, as well as incidence of Lygus in E. annuus patches. Additionally, as mass rearing programs of P. relictus ramp up, studies examining the recovery rate of the wasps after releases as well as parasitism rates on Lygus bugs could be carried out in crop fields with and without E. annuus patches, and within the E. annuus patches themselves.

145 References

Allen, M., Mertens, J. 2008. Molecular cloning and expression of three polygalacturonase cDNAs from the tarnished plant bug, Lygus lineolaris. Journal of Insect Science. 8.

Blackmer, J.L., Rodriguez-Saona, C., Byers, J.A., Shope, K.L., Smith, J.P. 2004. Behavioral response of Lygus hesperus to conspecifics and headspace volatiles of alfalfa in a Y-tube olfactometer. J. Chem. Ecol. 30:1547-1564. de la Paz Celorio-Mancera, M., Allen, M.L., Powell, A.L., Ahmadi, H., Salemi, M.R., Phinney,B.S., Shackel, K.A., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2008. Polgalacturonase causes lygus-like damage on plants: cloning and identification of western tarnished plant bug (Lygus hesperus) polygalacturonases secreted during feeding. Arthropod-Plant Interactions 2:215-225. de la Paz Celorio-Mancera, M., Greve, L.C., Teuber, L.R., Labavitch, J.M. 2009. Identification of endo- and exo-polygalacturonase activity in Lygus hesperus (Knight) salivary glands. Archives of Insect Biochemistry and Physiology. 70:122-135.

Di, C., Zhang, M., Xu, S., Cheng, T., An, L. 2006. Role of poly-galacturonase inhibiting protein in plant defense. Critical Reviews in Microbiology. 32:91-100.

D’Ovidio, R., Mattei, B., Roberti, S., Bellincampi, D. 2004. Polygalacturonases, polygalacturonase-inhibiting proteins and pectic oligomers in plant-pathogen interactions. Biochimica et Biophysica Acta 1696:237-244.

Gatehouse, J.A. 2002. Plant resistance towards insect herbivores: a dynamic interaction. New Phytologist. 156:145-169.

Leitner, M., Kaiser, R., Rasmussen, M.O., Driguez, H., Boland, W., Mithöfer, A. 2008. Microbial oligosaccharides differentially induce volatiles and signaling components in Medicago truncatula. Phytochemistry. 69:2029-2040.

Schmelz, E.A., Alborn, H.T., Banchio, E., Tumlinson, J.H. 2003. Quantitative relationships between induced jasmonic acid levels and volatile emissions in Zea mays during Spodoptera exigua herbivory. Planta. 216:665-673.

Schols, H., Kabel, M., Bakx, E., Daas, P., van Alebeek, G.J., Voragen, F. 2000. HPLC of oligosaccharides: New developments in detection and peak identification. AVH Association – 7th Symposium. 39-45.

VITA Sean T. Halloran

EDUCATIONAL BACKGROUND The Pennsylvania State University (2006-2012) Doctoral Program in Entomology, Advisor: Dr. Jim Tumlinson The College of New Jersey (2001-2005) Bachelor of Science in Biology

PROFESSIONAL AND TEACHING EXPERIENCE RELEVANT TO DISSERTATION Teaching Assistant, Entomology 497 – Insect Biodiversity and Evolution The Pennsylvania State University, University Park, PA September 2012-December 2012 Adjunct Faculty, Biology 417 – Invertebrate Zoology The Pennsylvania State University, Altoona, PA August 2008 – December 2008 Graduate Instructor, Entomology 316 – Field Crop Entomology The Pennsylvania State University, University Park, PA March 2008-May 2008 Teaching Assistant, Entomology 202 – Insect Connections The Pennsylvania State University, University Park, PA September 2007 – December 2007 Teaching Assistant, Entomology 313 – Introduction to Entomology The Pennsylvania State University, University Park, PA January 2007 – March 2007

PRESENTATIONS RELEVANT TO DISSERTATION “Interactions between Lygus bugs and Erigeron annuus: Applications toward a trap crop system for the tarnished plant bug.” Entomological Society of America annual meeting. November 2011, Reno, NV. “Chemical ecology of two co-occuring Lygus species on shared host plants.” Poster presentation, Plant-Insect Interactions Section, Ecological Society of America meeting, Summer 2010, Pittsburgh, PA. “Chemical ecology of two co-occurring Lygus species on shared host plants.” Student poster competition, Plant-Insect Ecosystems Section, Entomological Society of America annual meeting, December 2009, Indianapolis, IN. Second place in the Student Competition for the President’s Prize. “Interactions among Lygus rubrosignatus and its host plants: A study of volatile induction and implications for biological control of Lygus pests.” Student poster competition, Plant-Insect Ecosystems Section, Entomological Society of America annual meeting, November 2008, Reno, NV. “Interactions among Lygus lineolaris and its host plants: Biological control of target pests using volatile information.” Student Poster Competition: International Society for Chemical Ecology annual meeting, August 2008, University Park, PA.

OUTREACH ACTIVITIES The Great Insect Fair, University Park, PA Fall 2006, 2007, 2008, 2009, 2010 Frost Museum Tours, University Park, PA 2007, 2008, 2009 Bug Camp for Kids, University Park, PA Summer 2008, 2009, 2010 Library Presentation, St. Mary’s Library, St. Mary’s, PA Summer 2008 Library Presentation, Tyrone, PA Summer 2008 Lock Haven University Earth Day Celebration, Lock Haven, PA Summer 2008 PA Governor’s School Selection Committee Spring 2008 Penn State Bio Days (now Exploration Days) Festival, University Park, PA Spring 2007

HONORS AND AWARDS College of Agricultural Sciences Competitive Grant (2011), ESA Student Competition for the President’s Prize, Second place (2009), Michael E. Duke Memorial Scholarship in Entomology (2009), William Yendol Travel Award, The Pennsylvania State University (2008, 2009, 2011), University Fellowship(2006, 2007)

PROFESSIONAL ASSOCIATIONS Beta Beta Beta Biological Honor Society, Entomological Society of America, Ecological Society of America, International Society for Chemical Ecology