CAN PLANTS SMELL AN IMPENDING ATTACK?
RESPONSE OF HYBRID ASPEN TO VOLATILE ORGANIC COMPOUNDS EMITTED BY PHRATORA VITELLINAE
Kristen Grauer-Gray Master of Science Thesis Department of Environmental and Biological Sciences University of Eastern Finland May 2016
UNIVERSITY OF EASTERN FINLAND, Faculty of Science and Forestry, Department of Environmental and Biological Sciences, Environmental Biology Program (EnvBio)
KRISTEN GRAUER-GRAY: Can Plants Smell An Impending Attack? Response Of Hybrid Aspen To Volatile Organic Compounds Emitted ByPhratora vitellinae
MSc thesis, 59 pages, 2 appendixes (3 pages) Supervisors: James Blande (PhD), Tao Li (PhD) May 2016
___________________________________________________________________________
Keywords: plant-insect communication, hybrid aspen, Phratora vitellinae, VOCs, EFN
ABSTRACT
Induced defenses allow plants to preserve resources while still defending themselves against herbivores. However, these defenses are only effective if they are activated before the plant suffers too much damage. The volatile organic compounds (VOCs) emitted by herbivores are a reliable signal of impending attack, but the question of whether herbivore-derived VOCs induce plant defenses has not been well-studied. This study asks whether VOCs emitted by specialist leaf beetles (Phratora spp.) induce indirect defenses in hybrid aspen (P. tremulaL.
xP. tremuloides Michx.). Hybrid aspens enclosed in glass chambers were exposed to VOCs from Phratora spp. larvae and subsequently damaged by herbivores to test for priming effects. The VOC exposure did not induce either EFN secretion or VOC emissions, and adult Phratora showed no preference between the odors of exposed and unexposed trees.
Following herbivore damage, exposed trees emitted more DMNT and salicylaldehyde than unexposed trees. However, this effect was only marginally significant. For all trees, herbivory induced VOC emissions but did not induce EFN secretion. Overall, this study does not support the hypothesis that herbivore-derived VOCs induce indirect defenses in aspen.
Future studies should verify the observing priming of DMNT and salicylaldehyde emissions, and should also measure the effect ofPhratoraVOCs on the direct defenses of aspen.
ACKNOWLEDGEMENTS
This study was conducted from July to August 2015 at the Department of Environmental and Biological Sciences at the University of Eastern Finland in Kuopio.
My sincerest thanks to my advisors, James Blande and Tao Li, for their patient help and insightful comments throughout the thesis process. Thanks especially to Tao for teaching me the lab procedures used in this research and for patient guidance in data analysis, and to James for making the time to comment on the thesis when my deadline for finishing it changed on short notice. Thanks also to Jarmo Holopainen for reviewing this work.
Many professors and staff members at the Department of Environmental and Biological Sciences provided help with data collection and analysis. Thanks to Jaana Rissanen for preparing needed reagents and to Elina Häikiö for aid in propagating aspen. Thanks to Pasi Yli-Pirilä and Samuel Hartikainen for help with the GC-MS. A big thanks to Daniel Blande for writing Python code that greatly sped up the VOC analysis. Finally, thanks to Minna Kivimäenpää and Jouni Sorvari for helpful statistics discussions.
Thanks to my family for always supporting me no matter how far I’ve traveled from home.
Kiitos paljon to Akiko Kosaka and Ana-Maria Castro for friendship and support, both in stumbling through Finnish grammar and life in general, and thanks especially for the Finnish language lunches and relaxing evenings of cooking and conversation. A big kiitos to my Finnish language partner Lea Karjalainen for taking me into her home once a week for coffee, pastries, and conversation, and for introducing me to the joys of the Finnishmökki.
Many thanks andkiitos to all.
Kristen Grauer-Gray 17 May 2016
ABBREVIATIONS
DMNT (E)-4,8-dimethyl-1,3,7-nonatriene EFN extrafloral nectar
GLV green leaf volatile
TMTT (E,E)-4,8,12-trimethyl-1,3,7,11-tridecatetraene VOC volatile organic compound
TABLE OF CONTENTS
Acknowledgements ... 3
Abbreviations ... 4
Table of Contents ... 5
1. Introduction ... 7
2. Literature Review ... 9
2.1 Indirect defenses ... 9
2.2 Extrafloral nectar (EFN) ... 10
2.2.1 Defensive role of EFN ... 10
2.2.2 Induction of EFN ... 11
2.3 Volatile organic compounds (VOCs) ... 12
2.3.1 Types of VOCs ... 12
2.3.2 Attraction of natural enemies by VOCs ... 14
2.4 Induction of defenses by VOCs ... 15
2.4.1 Induction of defenses by VOCs from neighboring plants... 16
2.4.2 Induction of defenses by herbivore-derived VOCs ... 17
2.5 Study system: hybrid aspen andPhratora vitellinae ... 18
2.5.1 Hybrid aspen as a model plant ... 19
2.5.2Phratora vitellinae as a model herbivore ... 20
3. The Aim of the Work ... 23
4. Materials and Methods ... 24
4.1 Plants and insects... 24
4.2 Overview of experiment ... 24
4.3 Leaf numbering ... 25
4.4 Acclimation and baseline samples ... 25
4.5 VOC exposure ... 26
4.6 Herbivore damage and priming effects ... 27
4.7 Nectar collection and analysis ... 28
4.8 Behavior tests ... 28
4.9 VOC collection and analysis ... 29
4.10 Statistical analyses ... 31
5. Results ... 32
5.1. Olfactometry assays ... 32
5.2. VOC emissions ... 33
5.2.1 Effects of exposure toPhratoraVOCs ... 37
5.2.2 Effects of chamber conditions and herbivore damage ... 38
5.3. EFN secretion ... 39
5.4. Leaf damage ... 40
6. Discussion ... 41
6.1. Activation of defenses by herbivore-derived VOCS ... 41
6.1.1 Effect ofPhratora VOCs on EFN secretion ... 41
6.1.2 Induction of VOC emissions byPhratora larval VOCs ... 42
6.1.3 Priming of VOC emissions: false positive or actual effect? ... 43
6.1.4 Olfactometry assays ... 44
6.2. Induction of indirect defenses by herbivory ... 45
6.2.1 Effect of herbivory on EFN secretion ... 45
6.2.2 Induction of VOC emissions by herbivory ... 46
6.3. Effect of chamber conditions on VOC emissions ... 47
6.4. Fungal infection as a confounding factor ... 47
6.5. Conclusion and future directions ... 48
Conclusions and Summary ... 50
References... 51
d
Appendixes:
Appendix 1: Recipe for WPM-3 medium Appendix 2: VOC emissions data
1. INTRODUCTION
Because plants lack the ability to flee from herbivores, they must be able to protect themselves against attack. Plants may defend themselves directly by producing toxins, digestibility reducers, or physical barriers. They may also defend themselves indirectly by recruiting outside help in the form of the predators and parasitoids of herbivores. Two common strategies of indirect defense are the secretion of extrafloral nectar (EFN) and the emission of volatile organic chemicals (VOCs). EFN provides a food source for predators and parasitoids, while VOCs provide information on the location and identity of attacking herbivores. However, all these defenses require plants to use resources that could otherwise be spent on growth and reproduction. As a result, they may reduce fitness in cases where herbivores are absent.
To preserve resources while maintaining a strong defense, plants often express defensive traits only when herbivores are present. Defenses are induced by herbivore damage, with the plant responding to chemicals in herbivore saliva or to compounds released by damaged leaf cells. However, these stimuli come too late: by the time the plant is exposed to herbivore saliva, it has already been damaged. From a fitness standpoint, it is more logical for plants to detect cues that are present before damage occurs. The ideal cue must be reliable, meaning it must indicate a high probability of future damage, and it must be specific to the attacking herbivore species. The VOCs released by herbivores are both reliable and specific, and therefore provide an excellent signal of impending herbivory.
To date, few studies have examined whether plants are able to respond to VOCs from herbivores. However, multiple studies have shown that plants respond to VOCs from other plants. The parasitic plant dodder (Cuscuta pentagona) grows toward a source of VOCs emitted by a suitable host (Runyon et al. 2006). Lima bean, poplar, and maize all upregulate their defenses following exposure to VOCs from damaged plants (Heil & Silva Bueno 2007;
Engelberth et al. 2004; Dolch & Tscharntke 2000). In addition, one plant species has been shown to respond to herbivore-derived VOCs. Helms et al (2013) demonstrated that tall goldenrod (Solidago altissima) plants exposed to VOCs from male Eurosta solidaginis experience less herbivory than unexposed controls and are less attractive to ovipositing female E. solidaginis. These results clearly show that VOCs from insects can induce plant defenses. However, it is not known whether this phenomenon is common.
This study investigates whether herbivore-derived VOCs induce plant defenses in a model
system consisting of hybrid aspen (Populus tremula L. xPopulus tremuloidesMichx.) and a specialist leaf beetle (Phratora spp.). It focuses on two indirect defenses, VOC emission and EFN secretion. The study also examines whether herbivore-derived VOCs prime defenses, meaning that exposed trees respond more quickly or strongly to herbivory than unexposed ones. Finally, the effects of exposure to herbivore-derived VOCs on the host preferences of adult Phratora spp. are measured using olfactometry assays. Overall, this study aims to contribute to a greater understanding of plant-herbivore interactions and to provide insights into plant defenses that can ultimately be used in protecting crops against herbivores.
2. LITERATURE REVIEW
Plants are unable to flee from herbivores, and yet they are constantly under attack. On average, herbivorous insects consume 10% of the biomass that plants produce; this number rises as high as 50% in environments with a high herbivore density (Schoonhoven et al.
2005). Even small amounts of damage reduce plant fitness by decreasing growth rate and seed production (Schoonhoven et al. 2005). To survive in the face of herbivore attack, plants must defend themselves. Plant defensive strategies include direct defenses, such as the production of toxins and physical barriers, and indirect defenses, which involve recruiting the predators and parasitoids of herbivores (Roda & Baldwin 2003; Heil 2008). These strategies often increase fitness in environments with a high herbivore density, but they can be costly and wasteful in environments with a low population of herbivores (Zavala et al. 2004). Plants can preserve resources by activating their defenses only when herbivores are present (Karban
& Baldwin 1997). However, this strategy presents a problem: plants must be able to detect cues that indicate a high probability of future herbivory. Plants do not have eyes to see approaching herbivores, but they do have a “nose” to smell them. Multiple studies have shown that exposure to chemicals released by damaged leaves induces a defensive response in undamaged plants (Dolch & Tscharntke 2000; Karban et al. 2000; Arimura et al. 2000).
Do volatile chemicals released by herbivores also induce plant defenses?
This literature review discusses strategies of indirect defense in plants and asks whether airborne chemicals induce expression of these defenses. The review begins with an overview of common indirect defense strategies. It then discusses the ability of plants to respond to airborne chemicals emitted by other plants, and asks whether plants also respond to chemicals emitted by herbivores. Finally, it introduces the model system used in this study and discusses why the system was chosen.
2.1 INDIRECT DEFENSES
In order to clarify the discussion of plant defenses below, it is first necessary to define the term “defense”. The interactions between plants and herbivores can be viewed from either the plants’ perspective or the herbivores’ perspective. “Defense” refers to the plants’ perspective:
a defensive trait is any trait that decreases the negative effects of herbivory on plant fitness, regardless of the trait’s effect on herbivore fitness (Lindroth & St. Clair 2013; Heil 2004).
“Resistance” refers to the herbivores’ perspective: a trait provides resistance if it reduces herbivore fitness or preference for the plant (Lindroth & St. Clair 2013; Heil 2004). In this
work, defense and resistance are considered together, and the term “defense” is used to describe both types of traits.
Many plants defend themselves directly using physical barriers, toxins, or compounds that reduce digestibility (Roda & Baldwin 2003). However, the production of direct defenses can be metabolically expensive (Zavala et al. 2004), and it is not always effective. Specialist herbivores are adapted to the defenses of their hosts, and are often attracted by toxins that repel generalists (Ali & Agrawal 2012). To overcome this problem, plants also defend themselves indirectly by summoning outside help: the predators and parasitoids of attacking herbivores. Like any mercenary troops, the predators and parasitoids must benefit from the relationship, and plants pay their defenders by providing food, shelter, or information on the location of herbivores (Heil 2008). The following sections discuss two common mechanisms of indirect defense: the provision of food via extrafloral nectar (EFN) and the provision of information via emission of volatile organic compounds (VOCs).
2.2 EXTRAFLORAL NECTAR (EFN)
While floral nectar is secreted by reproductive plant parts, extrafloral nectar is produced by nectaries on vegetative parts (Heil 2015; Weber & Keeler 2013). Extrafloral nectaries are commonly found on leaves, petioles, bracts, or stipules (Weber & Keeler 2013). Like floral nectar, EFN is composed mainly of sugars dissolved in water, and usually contains glucose, fructose, and sucrose (Heil 2015). EFN also contains smaller amounts of amino acids and antimicrobial enzymes (Heil 2015). The composition of EFN varies greatly between plant species, and the variation most likely reflects the preferences of the insects attracted by each species (Grasso et al. 2015).
Extrafloral nectaries have been observed in almost 4000 plant species from over 100 families, most of which are angiosperms (Weber & Keeler 2013; Weber et al. 2015). The Fabaceae, Passifloraceae, and Euphorbiaceae contain the greatest number of species with extrafloral nectaries (Weber et al. 2015), but the ability to produce EFN is also widespread in individual genera of other families. In the genus Populus(Salicaceae), which is studied in this work, at least 26 of the approximately 30 species produce EFN (Cervera et al. 2005; Weber et al.
2015).
2.2.1 Defensive role of EFN
Unlike floral nectar, which attracts pollinators, EFN functions mainly as a mechanism of
defense. By providing an easily accessible food source, plants attract predators and parasitoids, which then attack the plants’ herbivores (Heil 2015; Weber & Keeler 2013). EFN attracts a variety of arthropods, including ants (Ness 2003; Heil et al. 2001; Kost & Heil 2005), wasps (Röse et al. 2006; Kost & Heil 2005), flies (Kost & Heil 2005), ladybirds (Lundgren & Seagraves 2011; Pemberton 1993), spiders (Soren & Chowdhury 2011), and mites (Choh et al. 2006). In addition to attracting beneficial insects, it may also attract herbivores. For example, adult aspen leaf miners (Phyllocnistis populiella) consume EFN from quaking aspen (Populus tremuloides), but larvae of the same species feed on aspen leaves (Doak et al. 2007).
EFN is an effective defense mechanism. EFN secretion generally leads to a reduction in herbivory, and the rate of secretion is correlated to the strength of the defense. In the tree Macaranga tanarius, leaves with artificially-induced EFN secretion were visited by a greater number of “defender” arthropods than uninduced leaves, and suffered over ten times less herbivore damage (Heil et al. 2001). In wild lima bean (Phaseolus lunatus), tendrils treated with artificial EFN hosted higher populations of predators and parasitoids than untreated tendrils, and suffered less herbivory (Kost & Heil 2005). However, a greater amount of EFN does not necessarily lead to an increased population of natural enemies. In quaking aspen (Populus tremuloides), trees that were defoliated for two seasons had a greater density of extrafloral nectaries than controls, yet there was no correlation between the density of extrafloral nectaries and the abundance of predatory arthropods (Wooley et al. 2007). This result illustrates the complexity of indirect defense systems. The efficacy of EFN as a defense depends on multiple environmental factors, including the density of attacking herbivores, the species of herbivore, and the abundance of natural enemies. In addition, these factors must be balanced with the costs of producing EFN and the tradeoffs between EFN production and other defense mechanisms.
2.2.2 Induction of EFN
EFN secretion may be constitutive or induced. Many plants constitutively secrete small amounts of EFN, but the secretion of larger amounts is induced by herbivore damage (Heil et al. 2001; Ness 2003; Wäckers 2001). The induction of EFN in response to herbivory has been demonstrated for both herbaceous and woody species, including lima bean (Phaseolus lunatus; Fabaceae), cotton (Gossypium herbaceum; Malvaceae), castor (Ricinus communis;
Euphorbiaceae), Catalpa bignonioides (Bignoniaceae), and Macaranga tanarius
(Euphorbiaceae) (Choh & Takabayashi 2006; Wäckers 2001; Heil et al. 2001; Ness 2003).
The magnitude of induction is usually positively correlated with the level of damage (Heil et al. 2001; Xu et al. 2014). However, the exact level of induction depends on the plant species.
For instance, damage bySpodoptera littoralis leads to a 2.5-fold increase in EFN production in castor (Ricinus communus), but to a 12-fold increase in cotton (Gossypium herbaceum).
The magnitude of EFN induction also depends on the identity of the herbivore. In native populations of tallow tree (Triadica sebifera), damage by both specialists and generalists induces EFN secretion, but induction is stronger when damage is caused by a specialist (Wang et al. 2013).
Herbivory does not always induce EFN secretion. In black cottonwood (Populus trichocarpa), damage by mealy bugs (Hemiptera: Pseudococcidae) induces EFN secretion, but damage by three species of Lepidoptera does not (Escalante-Perez et al. 2012). In hybrid aspen (Populus tremulaxPopulus tremuloides), damage by the generalist caterpillar Epirrita autumnata not only fails to induce EFN secretion, it actually reduces it (Li et al. 2012). This reduction may be due to herbivore damage to nectaries, to a reduction in carbohydrate availability following leaf damage, or to an as yet unverified ability of herbivores to suppress EFN production (Li et al. 2012).
2.3 VOLATILE ORGANIC COMPOUNDS (VOCS)
Wounded plants appear to suffer in silence, but on the chemical level, they are shouting.
Herbivore-damaged plants release large quantities of volatile organic compounds (VOCs) into the surrounding air. These VOCs act as a cry for help, attracting predators and parasitoids that prey on the plant’s herbivores. In some cases, VOCs also repel herbivores; in other cases, they backfire by making the plant easier for herbivores to find. This section gives an overview of the types of VOCs, their ecological functions, their induction following herbivory, and their roles in plant defense.
2.3.1 Types of VOCs
Plants damaged by herbivores emit a complex mixture of VOCs, which can contain more than 200 compounds (Dicke & Van Loon 2000). The volatile phytohormones methyl jasmonate (MeJA), methyl salicylate (MeSA), and ethylene act as signals that mediate plant responses to wounding and upregulate defense genes (Arimura et al. 2005). Most other VOCs fall into one of four categories: green leaf volatiles (GLVs), terpenoids, aromatic compounds,
and nitrogen or sulfur-containing compounds (Clavijo McCormick et al. 2012). Figure 1 shows the chemical structures of some common herbivore-induced VOCs.
Figure 1: Examples of VOCs emitted by plants following herbivore damage. Images are from ChemSpider (http://www.chemspider.com).
Green leaf volatiles (GLVs) are six-carbon aldehydes, alcohols, and esters synthesized from fatty acids (Scala et al. 2013; Heil 2008). They are emitted almost immediately upon wounding and provide a rapid signal of damage (Scala et al. 2013; Turlings et al. 1995).
Plants can sense airborne GLVs, and exposure to GLVs often induces a defensive response (Kost & Heil 2006; Engelberth et al. 2004; Farag & Paré 2002). For example, the GLV(Z)-3- hexenyl acetate induces EFN secretion in lima bean (Kost & Heil 2006) and primes terpenoid emission in maize and poplar (Frost et al. 2008b; Engelberth et al. 2004). Another GLV,(E)- 2-hexenal, induces terpenoid emission in tomatoes (Farag & Paré 2002) and transcription of defense genes inArabidopsis thaliana(Kishimoto et al. 2005).
Terpenoids constitute the largest portion of herbivore-induced VOCs. Stored terpenoids are emitted immediately upon wounding, but most terpenoids must be synthesized de novo and are emitted after a time lag of several hours (Turlings et al. 1990; Turlings et al. 1995). The terpenoids are a structurally diverse family containing over 40,000 volatile and non-volatile compounds (Withers & Keasling 2007), all of which are synthesized from isopentenyl diphosphate (IPP) or dimethylallyl diphosphate (DMAPP) (Holopainen & Blande 2012;
Dudareva et al. 2006). Terpenoids are classified by their number of carbons. The hemiterpene (C5) isoprene is constitutively emitted and protects plants from heat stress (Peñuelas et al.
2005). Most herbivore-induced terpenoids are monoterpenes (C10), sesquiterpenes (C15), or homoterpenes (C11 or C16) (Dudareva et al. 2006). Terpenoids are commonly involved in the attraction of natural enemies to herbivore-infested plants (see “Attraction of natural enemies by VOCs” below).
2.3.2 Functions of VOCs
Plant VOCs serve as both messengers and defenders. They protect plants from abiotic and biotic stresses, including high temperatures (Peñuelas et al. 2005), reactive oxygen species (Loreto & Velikova 2001), pathogens (Kishimoto et al. 2005; Nakamura & Hatanaka 2002), and herbivores (Kessler & Baldwin 2001; Heil 2004). They carry messages between different parts of the plant, and thereby allow plants to overcome the constraints imposed by limited vascular connections between branches (Heil & Silva Bueno 2007; Frost et al. 2007). VOCs emitted by damaged plants are also detected by undamaged neighbors, which sometimes respond by activating their own defenses (Arimura et al. 2000; Karban et al. 2000; Baldwin
& Schultz 1983).
VOCs mediate the interactions between plants, herbivores, and the natural enemies of herbivores. They attract herbivores to suitable hosts (Fernandez & Hilker 2007) and repel them from unsuitable ones (De Moraes et al. 2001; Bernasconi et al. 1998). At the same time, they increase the effectiveness of predators and parasitoids by signaling the location of herbivore-infested plants (Turlings et al. 1990; Kessler & Baldwin 2001; De Moraes et al.
1998). The attraction of natural enemies by VOCs is a common mechanism of indirect defense, and is considered in more detail below.
2.3.2 Attraction of natural enemies by VOCs
Herbivore-induced VOCs attract a wide variety of natural enemies to plants, including predatory mites (Dicke & Sabelis 1987), parasitic wasps (Turlings et al. 1990; Turlings et al.
1995), thrips (Tatemoto & Shimoda 2008), flies (Hulcr et al. 2005), and entomopathogenic nematodes (Rasmann & Turlings 2007). The use of VOCs to attract natural enemies is a common defense mechanism and has been documented in at least 25 plant families (Mumm
& Dicke 2010).
The VOCs emitted by damaged plants are not just a general shout for help, but rather a specific and detailed signal. Different blends of VOCs are released in response to mechanical and herbivore damage (Turlings et al. 1990), and the composition of the VOC signal emitted after herbivore damage depends on the plant species (De Moraes et al. 1998; Turlings et al.
1995), herbivore species (De Moraes et al. 1998; Ozawa et al. 2000), and developmental stage of the herbivore (Takabayashi et al. 1995; Yoneya et al. 2009). VOC emissions also vary temporally, with different blends emitted during the day and at night (De Moraes et al.
2001). Predators and parasitoids can distinguish between VOCs induced by different herbivore species, and are more strongly attracted to VOCs induced by their hosts or prey (De Moraes et al. 1998).
VOCs can also act as a direct defense by repelling herbivores. Herbivores use VOCs to determine host suitability, and may avoid damaged plants due to a lower nutritional content, greater competition for food, or a greater density of natural enemies (De Moraes et al. 2001).
The aphid Rhopalosiphum maidis prefers the odor of both air and undamaged maize plants over the odor of herbivore-damaged maize plants (Bernasconi et al. 1998). Similarly, female Heliothis virescens moths are repelled by VOCs emitted from damaged tobacco and avoid ovipositing on these plants (De Moraes et al. 2001). However, in other cases herbivores are attracted to VOCs from damaged plants since they represent mating opportunities (Ruther et al. 2002) or provide a strong signal that makes the plant easier to locate (Carroll et al. 2006).
2.4 INDUCTION OF DEFENSES BY VOCS
Indirect defenses are often induced defenses, meaning they are expressed only in response to cues indicating the presence of herbivores (Dicke & Hilker 2003). However, this strategy comes with a detection problem, as the plant must recognize the presence of herbivores.
Chemicals from herbivore saliva are a reliable sign of herbivory and often induce defenses (Arimura et al. 2005). Yet this signal comes too late: by the time the plant is exposed to herbivore saliva, damage has already occurred. Can plants also recognize the presence of herbivores before damage begins? If so, what cues inform them that herbivores are near?
The cues that induce defenses must be reliable, meaning they must indicate a strong probability of herbivory in the near future (Karban et al. 1999). To avoid a false alarm, they must also be specific and be associated only with the attacking herbivore species. One cue that is both reliable and specific is the volatile chemicals released by the herbivores themselves. The ability of plants to respond to volatile chemicals is well-established. The parasitic plant dodder (Cuscuta pentagona) grows toward a source of VOCs from its host plant and distinguishes between VOCs from suitable and unsuitable hosts (Runyon et al.
2006). Further, undamaged plants become more resistant to herbivores after exposure to VOCs from damaged neighbors (Dolch & Tscharntke 2000; Heil & Silva Bueno 2007;
Karban et al. 2000). Herbivore-derived VOCs are a more reliable signal of future herbivory than VOCs from other plants, and it therefore seems likely that they induce defenses as well.
This section discusses the possibility that herbivore-derived VOCs induce plant defenses. The section contains two parts. The first part illustrates the ability of plants to respond to airborne signals by providing examples of plant responses to VOCs from damaged neighbors. The second part asks whether plants are also able to detect VOCs from herbivores and summarizes the studies that exist on this topic.
2.4.1 Induction of defenses by VOCs from neighboring plants
Plants are able to sense and respond to airborne chemicals. In one of the first studies demonstrating this phenomenon, tomato plants were placed in closed chambers with sagebrush, which emits methyl jasmonate. The exposed tomato plants subsequently accumulated larger amounts of proteinase inhibitors than unexposed controls (Farmer &
Ryan 1990). Subsequent experiments demonstrated increased resistance to herbivores in cotton grown near mite-infested conspecifics (Bruin et al. 1992), tobacco grown near mechanically-damaged sagebrush (Karban et al. 2000; Karban 2001), and black poplar grown near defoliated conspecifics (Dolch & Tscharntke 2000). The induction of defenses by VOCs from damaged plants has since been demonstrated in several other species, including lima bean (Arimura et al. 2000; Heil & Silva Bueno 2007), sagebrush (Karban et al. 2006), aspen (Li et al. 2012), willow (Pearse et al. 2013), and maize (Engelberth et al. 2004).
VOCs from damaged plants do not always induce defenses. In many cases, they instead prime defenses, preparing the plant for a stronger or faster response upon herbivore attack (Frost et al. 2008a). Maize exposed to GLVs from damaged conspecifics produces more jasmonic acid than unexposed controls following herbivore damage, and also emits greater
quantities of terpenoids (Engelberth et al. 2004). Hybrid poplar (Populus deltoides x nigra) exposed to the GLV (Z)-3-hexenyl acetate and then damaged by gypsy moth larvae (Lymantria dispar) emits greater quantities of terpenoids, produces more jasmonic acid, and exhibits more rapid induction of protease inhibitor transcription than unexposed controls (Frost et al. 2008b). Priming provides a balance between defense and resource preservation:
it allows the plant to prepare for a strong response while still avoiding the use of resources on a false alarm (Engelberth et al. 2004).
Plants do not only respond to VOCs from damaged neighbors, but also to VOCs from damaged parts of the same plant. Many plants have poor vascular connections between adjacent branches, and VOCs provide an efficient way of sending signals to neighboring branches (Orians 2005; Heil & Karban 2010). In hybrid poplar (Populus deltoides x nigra), exposure to VOCs from damaged leaves primes terpenoid emissions in undamaged leaves on the same branch, even if the damaged and undamaged leaves have no vascular connection to each other (Frost et al. 2007). In highbush blueberry (Vaccinium corymbosum), exposure to VOCs from damaged branches both induces direct defenses and primes VOC emissions in undamaged neighboring branches (Rodriguez-Saona et al. 2009). Unlike the detection of VOCs from other plants, systemic defense induction provides a clear fitness benefit to the emitting plant, and may explain why plants evolved the ability to detect VOCs in the first place (Heil & Karban 2010). The induction of defenses by VOCs from neighboring plants may then have evolved secondarily as plants obtained a fitness benefit from eavesdropping on the internal signals of their neighbors. The question then arises: if plants can eavesdrop on signals sent by other plants, can they also eavesdrop on VOCs emitted by herbivores?
2.4.2 Induction of defenses by herbivore-derived VOCs
Insects use VOCs for communication with conspecifics and for defense against predators (Gullan & Cranston 2005). VOCs serve as sex pheromones that attract potential mates, alarm pheromones that warn of danger, and aggregation pheromones that inform conspecifics of a suitable food source (Gullan & Cranston 2005). Defensive VOCs are often bad-smelling or irritating chemicals that repel predators, such as the salicylaldehyde and iridoid monoterpenes secreted byPhratorabeetle larvae (Termonia & Pasteels 1999).
For plants, the VOCs emitted by herbivores represent a potentially valuable source of information. Herbivore-derived VOCs provide a reliable signal that herbivores are present and therefore indicate a high likelihood of impending attack. The information provided by
VOCs is also specific. Different herbivore species emit different blends of chemicals, and the composition of the VOC blend therefore indicates which species is present. The identity of the herbivore is important information for plants, which often express different defenses in response to attacks by different species (Ali & Agrawal 2012). Herbivore-derived VOCs possess all the characteristics of an ideal cue for inducing plant defenses: they are reliable, they are species-specific, and they are present before the attack begins. From a fitness point of view, it makes sense for herbivore-derived VOCs to induce or prime plant defenses.
The activation of plant defenses by herbivore-derived VOCs is a new area of research, and few studies have examined the phenomenon to date. However, a recent series of studies by Helms and colleagues provides intriguing evidence that herbivore-derived VOCs do indeed induce plant defenses. These studies examine a model system consisting of a d plant and herbivore: tall goldenrod (Solidago altissima) and a specialist gall-inducing fly (Eurosta solidaginis). WhenS. altissima is exposed to VOCs emitted by male E. solidaginis, exposed plants experience less herbivory than unexposed controls and are less attractive to ovipositing femaleE. solidaginis (Helms et al. 2013). This effect is specific to S. altissima and does not occur in plants that have no known association withE. solidaginis, such as maize (Zea mays), squash (Cucurbita pepo var.texana), and calico aster (Symphyotrichum lateriflorum) (Helms et al. 2014; Helms et al. 2013). The increased herbivore resistance of VOC-exposed S.
altissima may be due to the priming of defenses. In comparison with unexposed plants, exposed plants produce more jasmonic acid and emit greater quantities of terpenoids following herbivore damage (Helms et al. 2014; Helms et al. 2013).
While the studies by Helms and colleagues clearly prove that herbivore-derived VOCs are able to induce plant defenses, it is not known whether this phenomenon is common. This study aims to add to the literature on defense induction by herbivore-derived VOCs through the examination of a different model system: hybrid aspen (P. tremula L. x P. tremuloides Michx.) and the brassy willow leaf beetle (Phratora vitellinaeL.).
2.5 STUDY SYSTEM: HYBRID ASPEN AND PHRATORA VITELLINAE
This section describes the model system used in the study and the reason why it was chosen.
It begins by describing the characteristics of aspen that make it ideal for studies of defense induction by VOCs, namely its ability to respond to airborne chemicals and its large variety of defense strategies. It then summarizes the biology of P. vitellinae and describes the traits that make it likely to induce a defense response in aspen: the fact that it causes a high level of
damage to aspen and the fact that it emits large quantities of VOCs.
2.5.1 Hybrid aspen as a model plant
The hybrid aspen Populus tremula L. x Populus tremuloides Michx. is a cross between Eurasian aspen (Populus tremula L.), which is native to northern Europe and Asia, and quaking aspen (Populus tremuloides Michx.), which is native to Canada and the United States (Tullus et al. 2012). Hybrid aspen belongs to the genus Populus of the Salicaceae family (Tullus et al. 2012). This genus also includes poplars as well as black cottonwood (Populus trichocarpa), the first tree to be genetically sequenced (Jokipii et al. 2004).
Hybrid aspen is grown on plantations in Finland and other parts of northern Europe (Tullus et al. 2012). Its rapid growth makes it useful as a biofuel, and the light color of its wood makes it valuable for paper production (Tullus et al. 2012; Jokipii et al. 2004). In addition to its economic uses, hybrid aspen is often used as a model plant because it grows quickly, is easy to propagate, and displays a large amount of genetic variation (Jokipii et al. 2004).
Hybrid aspen was chosen for this study because it responds to airborne chemicals. Both aspen and its close relative, poplar, activate their defenses in response to VOCs from damaged leaves. In hybrid aspen, exposure to VOCs from herbivore-damaged conspecifics induces EFN secretion and primes terpenoid emissions in undamaged plants (Li et al. 2012).
In hybrid poplar (Populus deltoides x nigra), exposure to VOCs from damaged leaves primes terpenoid emissions in undamaged leaves on the same plant (Frost et al. 2007). These results show that aspen possesses the receptors or other biochemical processes needed to detect airborne chemicals, as well as the signaling pathways required to respond to these chemicals.
It is therefore a likely candidate for detecting and responding to herbivore-derived VOCs.
Aspen was also chosen for this study due to its wide range of defensive strategies. Aspen leaves contain condensed tannins, which reduce nutrient value, and phenolic glycosides, which are toxic to many generalist herbivores (Lindroth & St. Clair 2013; Schoonhoven et al.
2005). Following herbivore damage, aspen expresses defensive proteins, including chitinases, polyphenol oxidases, and protease inhibitors (Lindroth & St. Clair 2013). In addition to these direct defenses, aspen defends itself indirectly by secreting EFN and emitting VOCs.
Herbivore damage induces terpenoid emissions in hybrid aspen, with increased emissions specifically observed for linalool, DMNT, and (E,E)- -farnesene (Blande et al. 2007; Li et al.
2012). It seems likely that these VOCs serve as an indirect defense. While no studies have
been done on the attraction of natural enemies to VOCs from aspen, herbivore-induced VOCs from black poplar (Populus nigra) have been shown to attract parasitic wasps (Havill &
Raffa 2000; Clavijo McCormick et al. 2014a).
In addition to emitting VOCs, aspen provides food for the natural enemies of herbivores by secreting EFN. Aspen leaves have a pair of extrafloral nectaries at the base, which secrete EFN constitutively (Figure2) (Escalante-Perez et al. 2012). Nectaries are generally found on all young leaves, but only on some older leaves (Escalante-Perez et al. 2012). Herbivore damage can alter EFN secretion rate, but does not necessarily increase it: in a recent study, damage by the generalist herbivoreEpirrita autumnata led to a decrease in EFN secretion (Li et al. 2012). While the exact effect of herbivore damage on the indirect defenses of aspen is still unclear, it is clear that their expression is greatly altered by herbivory. For this reason, the indirect defenses of aspen were chosen as a model system for studying whether herbivore-derived VOCs induce plant defenses.
Figure 2: Extrafloral nectaries of hybrid aspen (Populus tremula L. x Populus tremuloides Michx.). Photo courtesy of Jarmo Holopainen.
2.5.2 Phratora vitellinae as a model herbivore
The brassy willow leaf beetle, Phratora vitellinae (Coleoptera: Chrysomelidae), was used as a model herbivore species.P. vitellinae is a specialist that feeds onSalix andPopulus, and it causes significant damage to aspen plantations (Palokangas & Neuvonen 1992; Köpf et al.
1997; Gruppe et al. 1999). It is well-adapted to the chemical defenses of its hosts: not only is it not harmed by the phenolic glycosides in aspen leaves, it actually uses them to synthesize defensive chemicals (Termonia & Pasteels 1999; Rank et al. 1998). P. vitellinaewas chosen as a model herbivore for two reasons: it constitutively emits VOCs, and it causes significant damage to aspen.
Figure 3. (a) P. vitellinae larva showing droplets of defensive secretion. (b) Aggregation of feeding P. vitellinaelarvae on aspen leaf. Photo (a) is courtesy of Jarmo Holopainen; photo (b) is author’s own photo.
P. vitellinae larvae produce secretions consisting mostly of the volatile compound salicylaldehyde (Termonia & Pasteels 1999; Rank et al. 1998). These secretions repel predators, protect the larvae from microbes, and reduce competition by repelling both adult P. vitellinae and other Phratora species (Gross et al. 2008; Palokangas & Neuvonen 1992;
Rank et al. 1998; Fernandez & Hilker 2007; Rowell-Rahier & Pasteels 1986). When the larvae are disturbed, they evert the reservoirs on their thorax and abdomen, exposing droplets of secretion (Figure3) (Rowell-Rahier & Pasteels 1986). After a few seconds, the secretions are returned to the reservoirs (Gross et al. 2008; Rahfeld et al. 2015). However, the reservoirs have openings which constitutively emit salicylaldehyde, even when droplets of secretion are not exposed (Gross et al. 2008). Salicylaldehyde is therefore a reliable and consistent signal thatP. vitellinae larvae are present. Further, this signal is expected to be strong: P. vitellinae aggregate during feeding, with 10 to 30 larvae on the same leaf (Figure3) (Gross et al. 2008).
The presence of many larvae in one place enhances the strength of the salicylaldehyde signal and increases the chance that aspens are able to detect it.
P. vitellinae was not only chosen because of its VOC emissions, but also because it causes significant damage to aspen. Plants must balance the costs and benefits of a defensive response; the benefits are far more likely to outweigh the costs for herbivores that cause a large amount of damage.P. vitellinaelarvae aggregate when feeding, and tend to consume all the leaf tissue on an entire leaf, leaving behind only veins and petioles (Kristen Grauer-Gray, personal observation). The larvae then move en masse to an adjacent leaf and repeat the
(b) (a)
process. Because of this feeding habit, P. vitellinae larvae usually destroy multiple leaves on the same branch. Replacing these leaves is expected to be expensive for the plant, and it therefore makes sense that aspen would mount a strong, pre-emptive response. As a cause of significant damage which constitutively emits VOCs, P. vitellinae has a high chance of inducing aspen defenses, and is therefore an ideal model species for studying the ability of aspen to respond to herbivore-derived VOCs.
3. THE AIM OF THE WORK
The aim of this study is to determine whether exposure to VOCs from Phratora vitellinae larvae induces a defense response in hybrid aspen. The study also asks whether exposure to P. vitellinae VOCs primes defenses, meaning that exposed aspens have a stronger or faster response to herbivory than unexposed controls. Specifically, the study asks four questions:
1) Does exposure to VOCs from P. vitellinae larvae alter EFN secretion rate and/or composition in hybrid aspen?
2) Does exposure to VOCs from P. vitellinae larvae alter the quantity and/or composition of VOC emissions in hybrid aspen?
3) Do P. vitellinae adults show a preference between hybrid aspen exposed to VOCs fromP. vitellinae larvae and unexposed controls?
4) Does exposure to VOCs from P. vitellinae larvae prime EFN secretion and/or VOC emissions in hybrid aspen?
This study is limited in scope and only measures indirect defenses. Direct defenses, such as the production of tannins and protease inhibitors, are not measured. However, measurements of leaf area consumed byPhratora larvae and olfactometry assays comparing the preferences of Phratora adults for exposed and unexposed trees provide a partial measure of whether other defenses have been induced. IfPhratorahave a significant preference between exposed and unexposed trees, then it can be assumed that another, unexamined defense has been induced.
4. MATERIALS AND METHODS
4.1 PLANTS AND INSECTS
Hybrid aspen (P. tremula L. x P. tremuloides Michx.) of clone 55 were provided by Elina Häikiö of the Department of Environmental Science at the University of Eastern Finland (Häikiö et al., 2007). Aspens were micropropagated on WPM-3 medium (see Appendix 1) and grown in an indoor cultivation room at 22ºC with a 19:5 h photoperiod. Seedlings were dipped in 1 mM indole-3-butyric acid solution to induce root growth, then potted in trays on a mixture of 70% peat, 15% sand, and 15% vermiculite. Trays were covered with a clear plastic cover to maintain humidity and kept indoors for 2-3 weeks at ~25ºC with a 16:8 h photoperiod. Seedlings were then transferred to individual plastic pots and kept indoors under the same conditions for an additional 2-3 weeks. All plants were moved to a greenhouse at least two weeks before the start of the experiment and grown under natural light with supplementary artificial light on cloudy days. Plants were watered daily and fertilized once a week with 1 dL of 0.1% Taimi Superex (Kekkilä, Vantaa, Finland).
All experiments used a mixture ofPhratora laticollis andPhratora vitellinae, which will be referred to as Phratora spp. throughout this work. Of these two species, only P. vitellinae produces salicylaldehyde (Termonia & Pasteels 1999), and is therefore of more interest for the purpose of this experiment. Adult Phratora spp. were collected from the Ruohoniemi research garden at the University of Eastern Finland in Kuopio, Finland. The Phratora population at Ruohoniemi is most likely descended fromPhratorapreviously gathered at the Finnish Forest Research Institute in Suonenjoki, Finland. Phratoralarvae were obtained both from Ruohoniemi and by breeding Phratora in the lab. AllPhratora were reared indoors on hybrid aspen leaves from the same clone used in experiments.
4.2 OVERVIEW OF EXPERIMENT
Figure 4 summarizes the timeline of the experiment. There were two distinct stages: exposure of aspens to VOCs from Phratora larvae (Day 1 evening to Day 5 morning) and herbivore damage (Day 5 evening to Day 6 morning). In addition, sample collection continued for two days after herbivore damage to test for priming effects (Days 6 and 7). Olfactometry assays were conducted on Day 5 after larval VOC exposure ended. EFN and VOCs from individual plants were collected on Day 1 before the start of the exposure, on Day 5 after the end of the exposure, and on Days 6 and 7 after herbivore damage. VOCs from chambers were collected
at least once per day throughout the experiment. Each stage of the experiment is described in more detail in the sections below.
Figure 4: Overview of experimental timeline. Aspens were acclimated to chambers for 2 days preceding the start of the exposure period. Exposure lasted 85-96 h and was carried out by attaching a jar of Phratora larvae to the treatment chamber. Herbivore damage was done usingPhratora larvae as described in the text.
Experiments were carried out in July and August 2015. Three replicates were done, with 1-2 weeks between replicates. After each replicate, chambers were cleaned with 70% ethanol and flushed with filtered ambient air at a flow rate of 4 L/min for at least 17 h. In addition, the treatment and control chambers were switched after each replicate.
4.3 LEAF NUMBERING
Leaves were numbered from the top of the tree moving down. The topmost leaf with a length of 4 cm at the start of the experiment was labeled as leaf 1; smaller leaves above leaf 1 were not numbered. With the exception of VOC collection from chambers, only leaves near the top of the tree were used in sample collection. The olfactometry assays and VOC collection employed the same set of leaves, but the exact leaves used varied between replicates. In the first replicate, VOCs were collected from leaves 9 and above for most of the experiment, but from the entire plant including roots and soil on Day 1. VOCs were collected from leaves 7 and above and leaves 8 and above in the second and third replicates respectively. The small leaves above leaf 1 were included in both VOC collection and olfactometry assays. For all replicates of the experiment, EFN was collected from leaves 1-6.
4.4 ACCLIMATION AND BASELINE SAMPLES
Experiments were conducted in glass chambers (74 cm height, 41 cm diameter, ~98 L volume) with one chamber used for larval VOC exposure and a second chamber used for unexposed controls. Three aspens with a height of 40-60 cm were placed in each chamber at
approximately equal distance apart. Before the trees were placed in the chambers, their leaves were rinsed 3x with tap water to remove EFN, then allowed to dry.
Trees were allowed to acclimate to the chambers for ~48 h before the start of the exposure period. Chambers were ventilated with filtered ambient air at a flow rate of ~4 L/min during the acclimation period. During both the acclimation period and the experiment, trees were illuminated by artificial lights with a photoperiod of 16:8 h.
On Day 1 of the experiment, baseline samples of VOCs from chambers, VOCs from individual trees, and EFN were collected as described below. Trees were watered and returned to the chambers after sample collection. Throughout the experiment, the position of individual trees within each chamber was rotated whenever trees were removed for sample collection and then returned to the chambers.
4.5 VOC EXPOSURE
Trees were exposed to VOCs fromPhratora spp. larvae for ~85 h, from the evening of Day 1 to the morning of Day 5. Exposure was carried out as shown in Figure 5. A glass jar containing 30-35 Phratora larvae in the second to fourth instars was connected to the treatment chamber, while an empty jar was connected to the control chamber. Filtered ambient air was pumped through each jar into the respective chamber at ~4 L/min.
In order to maintain a high concentration of salicylaldehyde in the treatment chamber, larvae were changed four times per day at three hour intervals, with fresh larvae attached at 9 am, 12 pm, 3 pm, and 6 pm. Based on samples collected during the experiment, fresh jars of larvae emitted salicylaldehyde at a rate of approximately 30.5 ± 5.6 µg/h. Larvae were changed every 3 h because pilot experiments showed that emissions decreased over time, with very low levels after 3 h. Salicylaldehyde exposure was minimal during the night as the larvae were not changed until the morning. Emissions from jars of larvae attached at 6 pm averaged only 0.09±0.05 µg/h at 9 am the next morning.
Larvae were reused during the experiment but were allowed to feed for at least 18 h before reuse. In order to maintain consistent conditions in both chambers, the empty jar connected to the control chamber was changed whenever a new jar of larvae was attached to the treatment chamber. Chambers were not opened during the exposure period. However, one tree per chamber was removed on the morning of Day 5 for olfactometry assays. The remaining trees were left in the chambers until the afternoon of Day 5 and had a total exposure time of ~93 h.
Figure 5: Experimental setup used to expose aspens to VOCs from Phratora larvae. The glass jar contained 30-35 Phratora larvae, which were changed 4x/day as described in the text. The control chamber used the same setup but with an empty jar attached to the chamber.
4.6 HERBIVORE DAMAGE AND PRIMING EFFECTS
After the end of the exposure period, all trees were damaged withPhratoralarvae to test for priming effects. Twenty larvae of the second to fourth instar were placed on each plant on the evening of Day 5, with four larvae each on leaves 1, 3, 5, 7, and 9. Larvae varied in size, but an effort was made to place a similar mixture of larvae on each tree. Larvae were able to move freely on each plant and were also able to move between the different plants in the chamber.
After the addition of larvae, the trees were returned to the chambers and the chambers were ventilated with filtered ambient air at ~4 L/min. Herbivore damage was carried out overnight and lasted ~14 h. In the morning, the chambers were opened and larvae were removed from the plants. Leaves were brushed to remove frass, but some frass was impossible to remove without causing damage and remained on the leaves. VOCs and EFN were collected on the two days following herbivore damage as described below. Plants were watered and returned to the chambers after each sample collection, and the chambers were ventilated with filtered ambient air at ~4 L/min.
After the final VOC collection on the evening of Day 7, leaves were removed from trees and herbivore-damaged leaves were scanned. Leaves were then dried for 3 d at 60ºC and dry
mass was recorded. Herbivore damage to leaves was not considered in determining dry mass and was assumed to have a negligible effect.
4.7 NECTAR COLLECTION AND ANALYSIS
Extrafloral nectar (EFN) was collected from leaves 1-6 of each plant before VOC exposure on Day 1, after exposure on Day 5, and after herbivore damage on Days 6 and 7. Nectar was collected as described in Li et al 2012. Briefly, 7 µL of Milli-Q water were pipetted onto each nectary with a 10 µL micropipette. The water was mixed with the nectar by gently sucking up the liquid and pipetting it back onto the nectary ~3 times, then the liquid was collected with the micropipette. This procedure was done twice on each leaf. Nectar was stored at -20ºC for a maximum of 7 wk before analysis.
The sugar content of EFN samples was measured by HPLC (Agilent 1100 Series, Waldbronn, Germany). Depending on the concentration of the sample, a 20 µL, 25 µL, or 40 µL injection volume was used. Sugars were separated on an Agilent Zorbax Carbohydrate Analysis column (150 x 4.6 mm; film thickness 5 µm) and eluted with 0.00125 M H2SO4 at a flow rate of 1 mL/min. Eluted sugars were detected with an hp1037A refractive index detector (Hewlett-Packard, Wilmington, Delaware, USA).
Peaks for sucrose, glucose, and fructose were identified and quantified by comparison with pure standards. Sugar concentrations were converted to secretion rates expressed in units of µg sugar/g leaf DW/day. EFN volume was found by weighing collection tubes before and after collection to obtain mass, then converting mass to volume using a density of 1 mg/L.
4.8 BEHAVIOR TESTS
Olfactometry assays were conducted after the end of VOC exposure but before herbivore damage. Assays were carried out in Y-tube olfactometers (12 cm base arm length, 8 cm side arm length, 1.6 cm interior diameter) using adult Phratora beetles starved for at least 17 h.
~30 beetles were used in each set of assays, with an equal number of male and female beetles used. To determine gender, beetles were captured during mating and the top beetle was assumed to be male. For the third replicate of the experiment, males and females were also differentiated by size, with the smallest beetles identified as male and the largest ones as female. Medium-sized beetles were not used in this replicate since their gender was impossible to determine. Each beetle was used for only one trial, and beetles were not reused between replicates.
Each set of behavior tests used one tree exposed to PhratoraVOCs and one unexposed tree.
VOCs from the trees were directed into the olfactometer as follows. A polyethylene terephthalate bag (Rainbow, Helsinki, Finland) was placed over the top ~8 leaves (the exact number of leaves varied; see leaf numbering section). A hole was cut in the top corner of the bag, a Teflon tube was inserted, and filtered ambient air was pumped into the bag until it inflated. A hole was then cut in the remaining top corner, and a second Teflon tube was inserted and connected to one arm of the olfactometer. The second tree was connected to the other arm of the olfactometer by the same procedure. Air was pumped through both bags and into the olfactometer at ~500 mL/min. However, airflow varied by ~50 mL/min between the two sides of the olfactometer. To account for this difference, the positions of the trees were switched after half the trials were finished. Trees were illuminated by artificial growth lights throughout the olfactometry assays.
Trials lasted 5 min. At the start of each trial, a beetle was introduced to the olfactometer by placing it in the cap and inserting the cap into the olfactometer. The time was started when the beetle left the cap and entered the base arm. Beetles were recorded as having made a choice after walking at least two-thirds of the way down one side arm and remaining there for at least 30 s. Beetles that did not make a choice within 5 min were recorded as “no choice”.
Some trials were stopped early after beetles continued beyond the end of the side arm and entered the tube carrying air into the olfactometer. In this case, beetles were recorded as having chosen the tree connected to that side arm.
After each trial, olfactometers were rinsed with 70% ethanol to remove beetle pheromones.
Olfactometers were then dried at 120ºC for 20 min before reuse.
4.9 VOC COLLECTION AND ANALYSIS
VOCs were collected from chambers and individual plants following the schedule in Table 1.
Briefly, VOCs were collected from chambers daily before and during the exposure period (Days 1-5) and twice daily after herbivore damage (Days 6 and 7). VOCs were collected from individual plants once before the start of exposure (Day 1), once at the end of exposure (Day 5), and twice daily after herbivore damage (Days 6 and 7). VOCs were collected in stainless steel tubes containing Tenax TA and Carbopack B (150 mg of each, mesh 60/80) (Markes International, Llantrisant, UK). Collection lasted 30 min in all cases except collections from chambers on Days 6 and 7, which lasted 10 min.
Table 1: Collection times for VOCs from chambers and individual plants. Collection times are approximate and varied by up to 2 h from the time listed.
Day of experiment Stage of experiment VOC collection time (chambers)
VOC collection time (plants)
1 Before larval VOC exposure 10 am 5 pm
2-4 During larval VOC exposure 8 am —
5 At end of exposure period 8 am 5 pm
6-7 After herbivore damage 10 am, 4 pm 11 am, 6 pm
During the exposure period, additional VOC samples with a 5 min collection time were taken to determine the salicylaldehyde concentration in the chamber headspace. To determine the minimum salicylaldehyde concentration, VOCs were collected in the morning before replacing the larvae from the previous night. To determine the maximum concentration, VOCs were collected 20 min after connecting a fresh jar of larvae.
VOCs were collected from chambers using a T-junction made of Teflon tubing. One arm of the T-junction was connected to the chamber outlet and a VOC collection tube was inserted into the perpendicular arm. The remaining arm was left open to allow excess air to flow out.
Air was pulled through the collection tube with a vacuum pump at ~200 mL/min. Inflowing air came from the chamber and had a flow rate of ~4000 mL/min. VOCs were collected simultaneously from both the treatment and control chambers.
VOC collection from individual plants was carried out as described in Li et al 2012. Briefly, plants were removed from chambers and a polyethylene terephthalate bag (Rainbow, Helsinki, Finland) was closed over the top ~8 leaves of each plant (the exact number of leaves varied; see leaf numbering section). An opening was cut in the top corner of each bag and a Teflon tube was inserted. Filtered ambient air was pumped into the bags at a flow rate of ~500 mL/min, either for 30 min or until the bags inflated. The flow rate was then reduced to ~250 mL/min. An opening was cut at the remaining top corner of the bag and a collection tube was inserted. Headspace was pulled through the collection tube with a vacuum pump at
~200 mL/min for 30 min. Headspace was also collected from an empty bag to control for VOCs present in the filtered air.
VOC samples were analyzed by GC-MS (GC: Agilent 7890A with HP-5 capillary column, column dimensions 50 m x 0.2 mm, film thickness 0.33µm; MS: Agilent 5975C VL MSD, 70 eV; Agilent Technologies, Santa Clara, California, USA). Helium was used for the carrier gas. VOCs were desorbed by heating sampling tubes to 280ºC for 10 min with an automated
thermal desorber (TD-100; Markes International, Llantrisant, UK), then cryofocused at -10ºC and transferred to the GC column. The column flow was set to 1.2 mL/min. The following oven conditions were used: 40ºC for 1 min, an increase of 5ºC min-1 to 210ºC, then an increase of 20ºC min-1 to 250ºC. VOCs were tentatively identified by comparing mass spectra to the spectra of known compounds in the Wiley and NIST libraries, then verified with pure standards when available.
4.10 STATISTICAL ANALYSES
Statistical analyses were carried out in SPSS 22 (IBM Corp, Armonk, New York, USA).
Choice data from olfactometry assays was analyzed using a binomial test, while the lengths of time beetles spent in each olfactometer arm were compared using a Wilcoxon rank-sum test. EFN and VOC data was analyzed by a repeated measures ANOVA with treatment as a fixed factor and time as a repeated factor. Non-normal data was log or square-root transformed before analysis. Greenhouse-Geisser corrections were used in cases where p<0.05 for Mauchly's sphericity test. Post hoc tests comparing different timepoints used a Sidak correction.
Percent damage to leaves was measured using ImageJ (US National Institutes of Health, Bethesda, Maryland, USA). The scale of the image was found by measuring the length of a ruler scanned together with the damaged leaves. The damaged area from all leaves on each plant was summed to obtain the total damaged area per plant. Mean damaged areas for control and treatment groups were compared by an independent samples t-test using log- transformed data.
Throughout this work, all measurements are reported as mean±SE unless otherwise noted.
5. RESULTS
5.1. OLFACTOMETRY ASSAYS
Only data from beetles that made a choice was used in analyses of olfactometry assays. 13%
of females and 18% of males made no choice and were excluded. In all, data was analyzed from 40 females and 37 males. All trials lasted 5 minutes except for 13 trials (6 with females, 7 with males) that ended early after the beetle entered the tube attached to the end of the olfactometer arm. In order to standardize trial length for time analyses, it was assumed that the beetle would have stayed in the chosen arm. The amount of time spent on that side was therefore increased to give a total trial length of 5 minutes.
Figure 6: Choice of adult Phratora spp. beetles between control aspens and aspens exposed to larval VOCs in olfactometry assays. Beetles were considered to have made a choice after walking at least two-thirds of the way down one arm of the olfactometer and remaining there for at least 30 s. Results for both males and females are not statistically significant (p>0.05).
Adult male Phratora had no preference between odors from aspens exposed to larval VOCs and odors from unexposed trees (Figure 6). 49% of males chose the exposed tree, while 51%
chose the control tree (binomial test; p=1.000). Male beetles also spent approximately the same amount of time in each olfactometer arm (Table 2).
Adult female Phratora also had no preference between odors from aspens exposed to larval VOCs and odors from unexposed trees (Figure 6). 60% of females chose the exposed tree, while 40% chose the control tree. This difference is not statistically significant (binomial test;
25 15 5 5 15 25
Number of beetles Male
Female
19 18
16 24
Control Larval VOC exposed No
choice
6 8
p=0.266). Females spent an average of 146 s in the olfactometer arm with odors from the exposed tree, and only 89 s in the arm with odors from the control tree. However, the difference between the length of time spent in each arm is not significant (Table 2).
Table 2: Average length of time Phratora beetles spent in each olfactometer arm in assays comparing aspens exposed to VOCs from Phratora larvae and unexposed controls. Most trials lasted 300 s; data for the remaining trials has been standardized to 300 s as described in the text. Times are expressed as mean ± SE. The Wilcoxon rank-sum test was used for statistical analysis.
Gender n Control (s) Larval VOC exposed (s)
Z p
Female 40 89.0±17.4 146.0±16.9 -1.215 0.224 Male 37 114.2±19.5 107.0±19.3 -0.023 0.982
5.2. VOC EMISSIONS
Emissions were analyzed for 41 VOCs collected from aspens, including 24 terpenoids, 7 aromatics, 5 nitrogen-containing compounds, and 5 GLVs. Other VOCs were also detected but were not analyzed. The chosen compounds were selected because they were present consistently in samples, were induced by herbivore damage, or were found by other studies to play a role in plant defense.
Figure 7 summarizes the total emissions for each category of VOCs before the start of the exposure period, at the end of the exposure period, and after herbivore damage. Complete data on VOC emissions for all measured compounds is given in Appendix 2.
Tables 3 and 4 summarize the results of a repeated measures ANOVA comparing VOC emissions in treatment and control plants. The results of post hoc tests for time effects are included to show whether changes in emission over time are due to herbivory or to changes during the exposure period. The treatment and time effects are discussed in separate sections below.
Figure 7: VOC emissions before and after herbivore damage from aspens exposed to VOCs fromPhratoralarvae (LVOC) and unexposed controls. VOC collection time was 30 min.
Pre=before exposure to larval VOCs, Post=after exposure. Bars show mean±SE.
