The Effects of Atmospheric Composition, Climate, and Herbivory On

THE EFFECTS OF ATMOSPHERIC COMPOSITION, CLIMATE, AND HERBIVORY ON

PLANT SECONDARY METABOLISM AND VOLATILE ORGANIC CARBON EMISSIONS by AMY MARIE TROWBRIDGE B.S., University of Illinois at Urbana-Champaign, 2005

A thesis submitted to the

Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirement for the degree of

Doctor of Philosophy Department of Ecology & Evolutionary Biology 2012

This thesis entitled: The effects of atmospheric composition, climate, and herbivory on plant secondary metabolism and volatile organic carbon emissions written by Amy Marie Trowbridge has been approved for the Department of Ecology & Evolutionary Biology

Russell K. Monson

M. Deane Bowers

Tim R. Seastedt

Alan R. Townsend

Dena M. Smith

Date

The final copy of this thesis has been examined by the signatories, and we

Find that both the content and the form meet acceptable presentation standards

Of scholarly work in the above mentioned discipline. iii

Trowbridge, Amy Marie (Ph.D., Ecology & Evolutionary Biology)

The effects of atmospheric composition, climate, and herbivory on plant secondary metabolism and volatile organic carbon emissions

Thesis directed by Professor Russell K. Monson

ABSTRACT

Quantifying the biosynthesis and emission of terpenes by plants is necessary to understand ecological interactions and improve global atmospheric models. Many studies have addressed the importance of individual factors influencing terpene production and release into the atmosphere, but few have investigated complex interactions between multiple variables and the resulting chemistry’s effects on interactions between organisms across multiple trophic levels. To address this gap in our knowledge, I conducted a series of experiments to examine abiotic and biotic controls over terpene synthesis and emission in Poplar x canescens and Pinus edulis, in which terpenes play prominent roles in plant physiological protection (isoprene) and defense

(monoterpenes). Increases in atmospheric CO2 resulted in a smaller contribution of stored extra- chloroplastic carbon towards isoprene biosynthesis and a larger investment from recently assimilated carbon in poplar. In addition to climate change scenarios, it is also critical to understand how abiotic and biotic processes affect terpene concentrations and emissions in situ.

Tiger moth herbivory increased pinyon pine monoterpene fluxes three to six fold during spring feeding, but summer drought decreased emissions while maintaining high levels of foliar concentrations. Following a release from drought stress, previously damaged pinyon pines exhibited significantly higher emission rates, potentially due to a drought delayed stimulation of induced monoterpene synthesis. I then performed a series of artificial diet experiments iv investigating how herbivore-induced monoterpenes observed in the field influences insect performance. While the synergistic effect of all monoterpenes present resulted in a trade-off between investment in immunity and growth, diets containing monoterpene levels mimicking herbivore damage encouraged further herbivory with no increase in growth but enhanced immunity to parasitoid infection. To further isolate temperature versus water status controls on monoterpene dynamics, I conducted a pinyon pine transplant experiment. Drought had little influence over the production of foliar monoterpenes, which is largely under genetic control.

However, observed stomatal control over emission rates suggests that plant ecophysiology plays a much larger role in controlling monoterpene fluxes than previously thought. Together, my research provides novel insights into the underestimated contribution of ecophysiology in understanding the role of terpenes in higher trophic level interactions and coevolutionary processes.

DEDICATION

For the two people who have dedicated their lives to making mine extraordinary. Thank you for

always believing in me, Mom and Dad.

ACKNOWLEDGEMENTS

This work would not have been possible without the contribution, support, and insight of many wonderful people. First and foremost I’d like to thank my mentors, Russ Monson and

Deane Bowers, for not only providing me with the skills to become a successful and independent scientist, but for fueling my enthusiasm and reminding me that I am smarter than I think I am. I’d like to thank my major advisor, Russ, for giving an ambitious, garrulous, goofy, mildly entertaining girl with a strong Chicago accent a chance and accepting me into his lab. Russ has taught me the importance of living in the moment, being patient, and seeing the big picture in both life and research. He has also given me opportunities to see the world and pursue my research passions, all the while pushing me to be better than I ever thought I could be. His contribution to my career and my personal growth has been invaluable, and words cannot express the amount of gratitude I have for his advice, support, and friendship.

I was also extremely lucky be advised by Deane, whose chemical ecology seminar changed my research interests and professional goals, but whose endless support and encouragement has changed my life. Deane has always made me believe that my work is exciting and worth sharing with the world. She is inspiring, and her endless energy and enthusiasm for science and life has made her someone I will continue to try and emulate. I feel so lucky to have had the opportunity to be a part of her lab and to learn from her. I am so grateful for her time, invaluable advice, encouraging words, and friendship.

I’d also like to thank all of my friends and colleagues at CU, without whom this would not have been possible. I would particularly like to thank Yan Linhart and Ken Keefover-

Ring for providing me with invaluable advice, hours of intellectually stimulating conversation, and plenty of French wine. I’d also like to thank my wonderful committee, Tim vii

Seastedt, Alan Townsend, and Dena Smith, for offering me their insights and feedback and for nurturing my curiosity and interest in plant-insect interactions. I’d also like to thank Cesar

Nufio and Kasey Barton, who taught me just about everything I know about insects and I am grateful that they were the ones to introduce me to the amazing world of entomology. I could not have made it through graduate school without the many hugs, stories, and laughs I shared with my peers, especially Susan Whitehead, Mary Jamieson, Carolina Quintero, Nicole

Trahan, Se Jin Song, Katie Driscoll, Joey Hubbard, Maggie Richards, Allyson Eller, Mick

Wilkinson, and Marcus Cohen.

Of course, this research could not have happened without the many lab assistants I have worked with over the years including Jonathan Kleinman, Ryan Weaver, Nate Monson,

Lindsay Young, Jeff Beauregard, Meredith Casciato, Jen Morse, Patricia Kazimier, Gift

Poopat, and Melissa Bernatis. I would also like to offer many thanks to the funding agencies that supported this work, including The Department of Ecology and Evolutionary Biology,

The University of Colorado Graduate School, The John Marr Fund, CIRES, as well as the

UROP and BURST programs.

I would like to express my everlasting gratitude to my family. I am so lucky to be surrounded by such loving, caring, selfless, and supportive people. I am the person I am today only because of you. I cannot thank you enough for all of the sacrifices you made, for answering my 3am phone calls, and reminding me every day that you are proud of me no matter what I choose to do with my life.

Finally, but most of all, I would like to say thank you from the bottom of my heart to my fiancé, Paul Stoy, who was instrumental in almost every single aspect of completing this dissertation. His constant encouragement and use of the collective noun “we” when referring viii to finishing this thesis has served as a reminder that I not only have a partner in science but also in life. Words cannot express how grateful I am for his patience, unconditional love and support, and for always believing in me. His faith in my thoughts, ideas, research, and teaching mean more to me than he will ever know. I know I could not have done this without his gentle encouragements, his creative insights and edits, coffee runs, and hugs. Thank you for loving me through it all. ix

CONTENTS

CHAPTER

I. INTRODUCTION……………………………………………………….1

1.1 Plant Secondary Chemistry………………………………..1 1.2 Environmental functions of isoprene and monoterpenes….1 1.3 The role of climate on isoprene and monoterpene synthesis and emissions…………………………………...4 1.4 Effects of monoterpenes on herbivore performance………6 1.5 Experimental questions…………………………………..10

II. CONTRIBUTION OF VARIOUS CARBON SOURCES TOWARD ISOPRENE BIOSYNTHESIS IN POPLAR LEAVES MEDIATED BY ALTERED ATMOSPHERIC CO2 CONCENTRATIONS………....…..12

Abstract………………………………………………………………....12 Introduction…………………………………………………..…………13 Materials & Methods……………………………………………………17 Results…………………………………………………………………..24 Discussion……………………………………………………….……...38 Conclusion…………………………………………………………...….44

III. HERBIVORY AND CLIMATE INTERACT SERIALLY DURING THE GROWING SEASON TO CONTROL MONOTERPENE EMISSIONS FROM PINYON PINE FORESTS……………………………………….……...46

Abstract...... …...46 Introduction………..………………………………………………….…47 Materials & Methods………………………………………………….…49 Results…………………………………………………………………...58 Discussion………………………………………………….…………....70

IV. HERBIVORY-INDUCED CHANGES IN MONOTERPENE CONCENTRATIONS OF PINYON PINE NEEDLES ENCOURAGE FURTHER HERBIVORY AND DETER PARASITISM……….………76

Abstract……………………………………………………….………....76 Introduction…………………………………………………………..….76 Materials & Methods…………………………………………………….80 Results……………………………………...……………………………91 Discussion…………………………………………………………….....99 x

V. UNTANGLING THE PRIMARY DRIVERS OF PINYON PINE NEEDLE MONOTERPENE PRODUCTION IN THE PRESENCE OF DROUGHT ……………………………………………………………………..104

Abstract……………………………………..………….……………...104 Introduction……………………………………………………………105 Materials & Methods………………………….…………….…………109 Results………….……………………………………….……………..116 Discussion…………………………..………………………….……....125

VI. CONCLUSIONS………………………….………….………..….……130

BIBLIOGRAPHY……………………..……………………………………….…….134

TABLES Table

3.1. Mean monoterpene emission rates [µg C gdw-1 hr-1] (±SEM) from undamaged and herbivore-damaged P. edulis trees on the Western (2009) and Eastern (2008) Slopes of the Rocky Mountains in Southern Colorado...... 64

3.2. Mean foliar monoterpene concentrations of individual compounds [mg gfw-1] (±SEM) from undamaged and herbivore-damaged P. edulis trees on the Western (2009) and Eastern (2008) Slopes of the RockyMountains in Southern Colorado...69

4.1. Mean foliar monoterpene concentrations (mg gfw-1) of individual compounds added to artificial diets mimicking undamaged and herbivore-damaged P. edulis trees during herbivory in May on the Eastern Slope of the Rocky Mountains in Saguache, County, Colorado……………….……………………...………….…………….…..85

5.1. The mean, standard deviation, and maximum percentage of total monoterpene emissions and concentrations observed in P. edulis across all sites and treatments in a transplant experiment in northern Arizona, USA ……………………………………………………………………………..119

xii

FIGURES Figure

2.1. Photosynthesis, intercellular CO2, and stomatal conductance for poplars grown under sub-ambient and elevated CO2 regimes………………………………..…..26

13 + + 2.2. CO2 labeling of M41 and M69 as a function of CO2 concentration and time……………………………………………………………….………..28

13 + + 2.3. CO2 labeling of carbon atoms in M69 and M41 and their isotopomers through time……………………………………………….…………………….….30

13 2.4. Rates of C transference between isotopomers as a function of CO2 concentrations………………………………………………………….…..34

2.5. Mean proportion of isotopomers of the parent isoprene molecule labeled at the conclusion of the experiment……………………………….…...…………35

2.6. Isoprene emission rates for poplars grown under 3 different CO2 regimes. …..………………………………………………………….………………37

2.7. Conceptual model illustrating the flow of carbon contributing to isoprene synthesis……………………………………...………………………..…...43

3.1. Mean monthly temperature and the monthly sum of precipitation from long-term Colorado Climate Center meteorological observations near study sites…..52

3.2. Seasonal change in net CO2 assimilation rates (A) stomatal conductance rates (gc) of undamaged and L. ingens damaged P. edulis at sites on the Western Slope and Eastern Slope of the Rocky Mountains in Southern Colorado……………….…….61

3.3. Seasonal change in total monoterpene emission rates and foliar concentrations of undamaged and L. ingens damaged P. edulis at sites on the Western Slope and Eastern Slope of the Rocky Mountains in Southern Colorado…...………………...62

4.1. Foliar monoterpene concentrations of Pinus edulis with and without herbivore damage observed on the Eastern Slope of the Rocky Mountains in Southern Colorado during caterpillar feeding in May 2008.…………………………………………...86

xiii

4.2 Relative growth rate (RGR) and proportion of melanization for larvae grown on low, medium, and high concentrations of α-pinene, β-pinene, β-myrcene, and (-)-limonene………………………………………...…………………...... 93

4.3 Correlation between relative growth rate (RGR) and proportion of melanization for larvae grown on diets containing low, medium, and high concentrations of individual monoterpene compounds found in undamaged pinyon pine needles in the field ……………………………………………………………………………….94

4.4. Approximate digestibility (AD), efficiency with which digested food is converted to body substance (ECD), consumption index (CI), and amount of frass produced over 24 hours for caterpillars grown on diets mimicking monoterpene levels in undamaged and herbivore-damaged pinyons from the Eastern Slope study site……………..96

4.5. Relationship between relative growth rate (RGR) and proportion of melanization for caterpillars grown on undamaged and damaged artificial diets.………...... 98

5.1. Frequency distributions of total and individual monoterpene emission rates and foliar concentrations from pinyon pine needles at both the desert and pinyon-juniper sites across all three treatments (drought, ambient, watered) and time (May-September) ……………………………………………………………………………....118

5.2. Total and individual monoterpene emission rates (µg C gDW-1 h-1) from pinyon pine needles at both the desert and pinyon-juniper sites from May to September. ……………………………………………………………………...……….121

5.3. Total and individual monoterpene foliar concentrations (mg gFW-1) from pinyon pine needles at both the desert and pinyon-juniper sites from May to September…………………………………………………………………...122

5.4. Total monoterpene emission rates, net CO2 assimilation rates, total monoterpene needle concentrations and stomatal conductance rates as a function of pre-dawn water potential for pinyon pines grown at both pinyon-juniper and desert sites across the May to September growing season……………………………….………………124

CHAPTER I

INTRODUCTION

1.1 Plant Secondary Chemistry

In addition to the precursors and products of primary metabolism, plants contain a wide array of chemicals known as secondary compounds. For years these compounds were considered waste products with little importance for plant function via growth and reproduction (Bennet &

Wallsgrove 1994). However, over the past 50 years, the importance of secondary compounds for plant adaptation, fitness, and survival via protection against both physical and biological stress have been described. Terpenes are one of the most diverse groups of secondary compounds with over 25,000 structures described to date across a wide variety of plant taxa (Huber et al. 2004).

Isoprene (C5), for example, is found across a broad taxonomic range (e.g. mosses, ferns, gymnosperms, and angiosperms) (Sharkey et al. 2008) and monoterpenes (C10) are found in 46 families of flowering plants and all conifers (Lerdau 1991). Although the production of specific terpenes, including isoprene and monoterpenes, is under genetic control, phenology, environmental factors, and/or herbivory can alter constitutive plant carbon allocation patterns

(Heyworth et al. 1998; Kuokkanen et al. 2001; Sallas et al. 2003; Nykanen & Koricheva 2004;

Gols et al. 2007) towards the production of these compounds. Whether constitutive or induced, the mixture of compounds present both within plant tissues or released to the atmosphere have been shown to possess multiple important chemical ecological roles (Huber et al. 2004).

1.2 Environmental functions of isoprene and monoterpenes

Isoprene and monoterpenes are some of the most expensive secondary metabolites to synthesize and result in substantial energy trade-offs with plant growth and reproductive processes (Gershenzon 1994; Keeling & Bohlmann 2006). However, both sets of compounds 2 confer a number of advantages to the species that produce them, likely explaining their wide distribution among plant families (Sharkey et al. 2008). One of the most tested hypotheses put forth about the functional importance of isoprene synthesis is thermotolerance (Sharky &

Singsaas, 1995). A number of studies have shown that isoprene’s ability to protect membrane integrity (Siwko et al., 2007) confers tolerance to short high temperature episodes and ozone/reactive oxygen species (Sharkey & Yeh 2001; Velikova & Loreto 2005). Other potential effects of isoprene include increasing rates of flowering in arabidopsis (Terry et al. 1995) and a

‘safety-valve’ to safely discard unwanted metabolites and excess energy (Rosenstiel et al., 2004;

Magel et al., 2007). While research continues to address the evolutionary advantages conferred to the species that are capable of emitting isoprene, understanding the mechanisms regulating isoprene biosynthesis will offer important insights into the selective forces responsible for its distribution across taxa and functional advantage, which I explore in Chapter Two.

Monoterpenes have been shown to mediate antagonistic and mutualistic interactions among organisms, for example by stimulating oviposition in specialist herbivores and deterring oviposition in generalists (Leather 1987; Binder & Robbins 1997; Shu et al. 2004; Grant et al.

2007). Some defoliators, such as sawflies, exhibit no change in behavior or function with exposure to certain monoterpenes, and instead exploit them for their own defense (Codella &

Raffa 1995; Pasquier-Barr et al. 2001). Monoterpenes can also act as toxins to fungal pathogens

(Paine & Hanlon 1994) and can serve as feeding deterrents to generalist herbivores

(Hummelbrunner & Isman 2001; Abdelgaleil et al. 2009). Conifers, in particular, employ a number of defensive strategies against herbivorous and pathogenic pests, including producing oleoresin comprised of a diverse array of monoterpene compounds (Bohlmann & Croteau 1999;

Phillips & Croteau 1999; Trapp & Croteau 2001). Some studies demonstrate that needles of 3 previously defoliated conifers exhibit higher suitability for subsequent defoliator generations

(Niemela et al. 1991; Lyytikainen 1992; Clancy et al. 2004), while others have demonstrated induced resistance after defoliation (Trewhella et al 1997; Smits & Larsson 1999; Hodar et al.

2004). The susceptibility or resistance of previously defoliated trees depends on a number of variables (i.e. tree age, intensity of defoliation, herbivore species, etc); thus, induced responses do not necessarily result in increased resistance (Mumm & Hilker 2006).

Due to their low-molecular weight, high vapor pressures, structural diversity, and lipophilic properties, monoterpenes are well-suited to convey specific chemical messages among species

(Gershenzon & Dudareva 2007). Many species of insects utilize monoterpenes as sex, aggregation, and alarm pheromones (Hick et al., 1999). Plants are also thought to release monoterpenes in an attempt to lure in pollinators, although proof of concept under natural conditions has been difficult to obtain (Wright et al. 2005). Instead of attracting mutualists, some monoterpene signals are exploited by other herbivore species and parasitic plants to locate their hosts (Carroll et al. 2006; Runyon et al. 2006). Furthermore, constitutive monoterpene levels in the vegetative tissues of a plant can be altered through induction by insect feeding (Byun-McKay et al. 2003; Miller et al. 2005) and/or oviposition (Hilker et al. 2002). Many plant species employ induced volatile monoterpene cues to attract the natural enemies of their associated herbivores

(DeMoraes et al. 1998; Thaler et al. 2002; Kant et al. 2004; Mithofer et al. 2005; Dicke et al.,

1990; Turlings et al. 1990; Kessler and Baldwin 2001), the synthesis of which may be triggered by substances in the oral secretion of herbivores or oviposition fluids (Mattiacci et al. 1994;

Alborn et al. 1997). Despite the wealth of literature on the functions of monoterpenes, the effect of constitutive and induced monoterpenes on herbivore performance appears to be species- specific and many conifer-herbivore systems have yet to be examined. 4

Monoterpenes and isoprene are considered trace atmospheric gases, but serve as strong drivers of regional and global atmospheric processes (Fuentes et al. 2000). Due to their reactivity in the atmosphere, both isoprene and monoterpenes can have an important influence on tropospheric chemistry affecting ozone (O3) and hydroxyl radical (OH) concentrations, accelerating the formation of organic nitrates and secondary organic aerosols, as well as contributing to the formation of organic acids and acid rain (Harley et al., 1998). Isoprene synthase is light-activated (Kuzma & Fall 1993), and isoprene is rapidly volatilized if enough substrate is available through recently assimilated carbon. In contrast, a number of plant species store monoterpenes in specialized structures, and their release into the atmosphere is thought to be driven according to physical properties (i.e. Henry’s Law), independent of physiological processes (Lerdau et al. 1995, 1997). Due to the strong influence of isoprene and monoterpene emissions on atmospheric chemistry and climate feedbacks, via radiative forcing and precipitation regimes, it is crucial to understand the complex and interacting effects of the biological and environmental factors that determine their synthesis and emission rates.

1.3 The role of climate on isoprene and monoterpene synthesis and emission

Terpene concentrations and emissions vary significantly across time and space and by species, which leads to large uncertainties in the models used to estimate global emissions. This is due to the complex and interactive nature of the drivers controlling the synthesis and emission of these compounds including vapor pressure gradients, concentration, volatility, genetics, as well as external abiotic and biotic factors (Penuelas & Llusia, 2002). One environmental condition that has been shown to influence isoprene emission rate, and is expected to rise significantly under global climate model scenarios, is atmospheric CO2 concentration (Monson and Fall 1989; Rosenstiel et al. 2003; Wilkinson et al. 2009). Studies have shown that 5 photosynthesis and isoprene emission rates are negatively correlated when exposed to altered

CO2 concentrations, and thereby predict a decrease in isoprene emissions under anticipated increases in atmospheric [CO2] (e.g. Monson et al. 1989; Loreto et al. 2007). Isoprene production is rarely limited by photosynthetically-assimilated carbon, where net CO2 assimilation fluxes on the scale of µmol m-2 s-1 are more than enough to sustain isoprene emissions typically reported

-2 -1 on the scale of nmol m s . This suggests that isoprene emission rates are affected via CO2 concentrations altering dimethylallyl diphosphate (DMADP) substrate availability and/or isoprene synthase activity or quantity. The biochemical mechanisms responsible for this effect have not been fully resolved, but research suggests that an increase in atmospheric CO2 concentration may shift competition for substrate (phosphoenolpyruvate; PEP) in favor of cytosolic processes (Rosenstiel et al., 2003). Furthermore, recent studies have demonstrated support for the hypothesis that differences in DMADP biosynthesis rates observed among CO2 treatments are due to changes in levels of pyruvate (PYR) and glyceraldehyde-3-phosphate

(G3P) availability for the deoxyxylulose-5-phosphate/2-methyl-3-buten-2-ol (DOXP/MEP) pathway (Possell & Hewitt, 2005). However, there remains much to discover about the biosynthetic kinetics and mechanisms by which G3P and Pyr are utilized for isoprene production and how these substrates will be allocated toward isoprene synthesis under altered CO2 conditions.

Changes in water status also impact plant secondary chemistry; seasonal drought, for example, impacts both primary and secondary plant metabolism (Mattson & Haack 1987). show

Strong seasonal patterns that related to water availability can be seen for both conifer foliar monoterpene levels (Shonwitz et al. 1990; Nerg et al. 1994; Riddle et al. 1996; Owens et al.

1998) and emission rates (Staudt et al. 1997; Staudt et al. 2000; Hakola et al. 2006; Holzke et al. 6

2006). While a number of studies have shown that levels of monoterpene emissions tend to decrease dramatically under drought conditions (Llusia & Penuelas 1998; Staudt et al. 2002;

Lavoir et al. 2009), the internal composition and concentration of monoterpenes in the needles may not reflect the composition of released volatiles (Llusia & Penuelas 1998). Stored monoterpene pools are likely determined primarily through genetics and produced during leaf development such that environmental conditions, including temperature and precipitation, have negligible effects on carbon source-sink relationships and monoterpene synthesis (Penuelas and

Estiarte 1998; Constable et al. 1999). Whereas genetic predisposition is a primary control over monoterpene concentrations, monoterpene emissions tend to be altered by climate variability at multiple time scales.

In the short-term, temperature affects monoterpene emissions in an exponential manner

(Tingey et al. 1991) and is commonly considered to be the driving force over monoterpene emissions. However, water limitation, associated with high temperatures during drought, can decrease the permeability of the cuticle to gas exchange (Bertin & Staudt 1996; Llusia &

Penuelas, 1998, 2000), yet studies investigating the interactive effects of temperature and water status on monoterpene concentrations and emission for terpene-storing species in situ are lacking. Together, these complex interactions result in deviations from standard light and temperature algorithms used to estimate terpene production and emissions (Tingey et al. 1991).

Thus, more research into the abiotic controls over monoterpene synthesis and release, especially the role of water availability, are needed to improve atmospheric models in the face of future climate change.

1.4 Effects of monoterpenes on herbivore performance 7

Insect population outbreaks are one of the most impressive and interesting plant-herbivore relationships, and much research has focused on determining the underlying causes of these periodic population surges (Myers 1988; Liebhold & Kamata 2000; Dwyer et al. 2004). Plant chemistry is thought to have a primary influence over herbivorous insect population dynamics

(Trewhella et al. 2000; Barre et al. 2003) despite the fact that many herbivores have evolved adaptations to cope with the characteristic allelochemicals of their host plants. Constitutive and induced/altered plant chemistry can impact herbivores both from the bottom-up, through their effects on herbivore performance via growth, survival, and fecundity (Cates 1980; Havill &

Raffa 2000; Awmack & Leather 2002), but also from the top-down, via host location and acceptance by the third trophic level (Ode 2006), including both predators and parasitoids.

Variation in host-plant chemistry affects herbivores and their natural enemies simultaneously, but sometimes differentially (Teder & Tammaru 2002). Whether plant chemical effects on individual herbivore and parasitoid host-choice and performance translate into effects on population and community-level dynamics remains a topic of debate (Mattson & Haack 1987;

Coley 1998; Larsson et al. 2000). Thus, a more integrated perspective on how host plant chemistry changes over space and time in response to herbivory and climate and their effects on higher trophic levels is critical for identifying potential mechanisms driving forest population fluctuations.

Significant changes in the quality of foliage during feeding can slow insect growth rates

(Rhoades 1983) most likely the result of detoxifying high levels of defense compounds in plants

(Bowers & Puttick 1988; Appel & Martin 1992). The mechanism by which many herbivores detoxify monoterpenes most likely occurs through polysubstrate cytochrome P-450-dependent monooxygenases (PSMOs) (Berry et al. 1980; Zou & Cates 1997), yet the details of the 8 detoxification processes for many species remains to be fully resolved. Regardless, it is likely that consumption of these compounds, and subsequent detoxification, affects larval metabolic activity (Berenbaum 1983; Rhoades and Cates 1997), thus resulting in energy trade-offs between other growth, reproductive, or defense processes. Even species of herbivores known to sequester secondary metabolites have exhibited energy trade-offs, particularly in terms of defense against parasitism (Smilanich et al. 2009). Thus, herbivore-induced changes in plant chemistry can result in poor nutrition or increases in allelochemicals, the latter incurring physiological costs as a result of detoxification.

The insect innate immune response can be immediate (Godfray 1994) and involves interactions between humoral and cellular responses that ultimately target and eradicate foreign objects in the haemocoel (Bulet et al. 1999; Lavine & Strand 2002; Cerenuius & Soderhall 2004;

Beckage 2008). The primary immune response against endoparasitoids includes encapsulation

(building cell layers around a foreign object) and subsequent melanization of these cell layers once they die due to the production of cytotoxic compounds and asphyxiation (Strand 2008). The costs of detoxifying plant secondary compounds may affect herbivore quality and size and have negative consequences for parasitoid size and survival (Blumberg 1997; Turlings & Benrey

1998, Ode 2006). The physiological cost of detoxification may also cause herbivores to become immunocompromised and increase their vulnerability to parasitism (Schmid-Hempel & Ebert

2003, Smilanich et al. 2009). Because parasitoids have been cited as the primary controls over fluctuations in herbivore populations (Berryman 1996), the relative energetic investment made by herbivores toward detoxification can affect top-down forces including the role played by parasitoids in determining the regularity and severity of insect outbreaks. 9

Furthermore, the relative importance of environmental factors affecting defense prioritization by plants in the context of resource status and herbivory is unclear and rarely tested (Nybakken et al. 2008). Environmental factors are important for determining herbivore-induced plant chemistry in natural systems, and studying these interactions in situ in will improve our understanding of tri-trophic ecological interactions and ecosystem functioning. Studying plant- herbivore-parasitoid interactions in ecologically sensitive ecosystems will lead to an improved understanding of ecological functioning for potential conservation benefits.

Pinyon-juniper woodlands are the third largest vegetation type in the Western United States

(West 1999) and have experienced large-scale mortality due to extreme drought conditions and associated insect outbreaks in the last decade, with subsequent shifts in pinyon-juniper ecosystem structure and function (Hanson & Weltzin 2000;Mueller et al. 2005; McDowell et al.

2008). The magnitude of drought-induced insect outbreaks is affected by complex ecological interactions that occur across multiple trophic levels. While much research has focused on understanding differential susceptibility rates of mortality between pinyon and junipers in terms of physiological constraints and successive bark beetle (Ips confusus) attack (Breshears et al.

2005; Breshears et al. 2008, McDowell et al. 2008; Adams et al. 2009; Floyd et al. 2009), few studies have examined how abiotic stresses and pinyon folivores interact (Cobb et al. 1997;

Huberty et al. 2004). To my knowledge, no field studies have explored how multiple stresses (i.e. environmental change and herbivore feeding) interact and alter pinyon secondary chemistry, particularly monoterpene production and emission. Furthermore, the effects of altered pinyon secondary chemistry on multi-trophic interactions, including herbivore population control via top-down effects, and how these ultimately impact ecosystem functioning remain unknown.

Effective management of pinyon-juniper woodlands in the face of current and future drought- 10 induced mortality and insect attacks is a common management concern, but most management actions are based on stand age, density, or heterogeneity (Floyd et al. 2009) while the effects of plant chemistry are largely ignored despite the potential role played by pinyon chemistry on controlling herbivore populations both directly and indirectly.

1.5 Experimental questions

The state of the knowledge of isoprene and monoterpene production, emission, and ecology must be modernized to better-understand the complex feedbacks that exist among atmospheric composition, hydrologic stress, herbivory, and multi-trophic ecological dynamics. The primary goals of this research are as follows:

i) To examine the effects of altered atmospheric CO2 concentrations on carbon

partitioning toward isoprene biosynthesis in poplar, particularly the contribution

of recently assimilated photosynthate (chapter 2).

ii) To characterize the interactive effects of herbivory and seasonal variation in

controlling monoterpene concentrations and emissions in pinyon pine (chapter

iii) To test the effects of constitutive and herbivore-induced foliar monoterpenes on

growth rate and immunocompetency of the folivorous Southwestern tiger moth

(Lophocampa ingens) (chapter 4).

iv) To tease apart the abiotic controls over monoterpene emissions under drought

stress and the implications for susceptibility to future insect attacks and

atmospheric processes (chapter 5).

The research objectives were addressed with a combination of field and laboratory studies using isoprene-emitting poplar (Poplar x canescens) and monoterpene-producing pinyon pine (Pinus 11 edulis) as model systems. The release of these volatile compounds into the atmosphere can have important implications for tropospheric chemistry, climate feedbacks, plant function, and ecological interactions/signaling. Because these species make up a significant portion of North

American forests, changes in their primary and secondary metabolism can have significant cascading effects both within individual ecosystems and on a regional scale. Therefore, it is critical to understand the abiotic and biotic controls over the synthesis and release of these compounds as well as the impacts this altered chemistry has on the plant function via susceptibility/resistance to herbivory in order to improve forest management in the face of future climate change. This research takes an integrative and multi-scale approach to better understand the complex ecological impacts of two dominant biogenic volatile organic compounds in driving biosphere-atmosphere interactions in response to environmental variation and herbivory.

CHAPTER II

CONTRIBUTION OF VARIOUS CARBON SOURCES TOWARD ISOPRENE BIOSYNTHESIS IN POPLAR LEAVES MEDIATED BY ALTERED ATMOSPHERIC CO2 CONCENTRATIONS1

ABSTRACT

Biogenically released isoprene plays important roles in both tropospheric photochemistry and

13 plant metabolism. We performed a CO2-labeling study using proton-transfer-reaction mass spectrometry (PTR-MS) to examine kinetics of recently assimilated photosynthate into isoprene emitted from poplar (Poplar x canescens) when trees are grown and measured at different atmospheric CO2 concentrations. This is the first study to explicitly consider the effects of altered atmospheric CO2 concentration on carbon partitioning to isoprene biosynthesis. We studied changes in the proportion of carbons labeled as a function of time in two mass fragments,

M41+, which represents, in part, substrate derived from pyruvate, and M69+, which represents the whole unlabeled isoprene molecule. We observed a trend of slower 13C incorporation into isoprene carbon derived from pyruvate, consistent with the previously hypothesized origin of chloroplastic pyruvate from cytosolic phosphenolpyruvate (PEP). Trees grown under sub-

+ + ambient CO2 (190 ppmv) had rates of isoprene emission and rates of labeling of M41 and M69 that were nearly twice those observed in trees grown under elevated CO2 (590 ppmv). However, they also demonstrated the lowest proportion of completely labeled isoprene molecules. These results suggest that under reduced atmospheric CO2 availability, more carbon from stored/older

1 Published in PLoS ONE (2011) 7(2) e32387. doi:10.1371/journal.pone.0032387 13 carbon sources is involved in isoprene biosynthesis, and this carbon most likely enters the isoprene biosynthesis pathway through the pyruvate substrate. We offer direct evidence that extra-chloroplastic rather than chloroplastic carbon sources are mobilized to increase the availability of pyruvate required to up-regulate the isoprene biosynthesis pathway when trees are grown under sub-ambient CO2.

INTRODUCTION

Isoprene (C5H8, 2-methyl-1,3-butadiene) is a principal and highly volatile biogenic hydrocarbon released into the atmosphere predominantly by plants (Fehsenfeld et al. 1992;

Harley et al. 1999; Fuentes et al. 2000; Sharkey & Yeh 2001). Because of isoprene’s reactivity with tropospheric oxidants and large global emission rate, considerable research has gone into identifying the biochemical processes that control isoprene emissions from leaves, including their sensitivities to environmental change and representation in regional and global emission models (Guenther et al. 1991; 1993; 1995; 2006; Monson et al. 2007; Wilkinson et al. 2009;

Heald et al. 2009). Briefly, isoprene is synthesized in leaf chloroplasts from dimethylallyl diphosphate (DMADP), a product of the deoxyxylulose-5-phosphate/2-methyl-3-buten-2-ol

(DOXP/MEP) pathway, which utilizes glyceraldehyde-3-phosphate (G3P) and pyruvate (Pyr) as initial substrates (Rohmer et al. 1993; Lichtenthaler 1999). Both G3P and Pyr are derived from photosynthetically assimilated CO2, with G3P being a direct product of the reductive pentose phosphate pathway in chloroplasts. Most evidence to date indicates that pyruvate, however, is produced from photosynthate that is exported from the chloroplast, converted to phosphoenolpyruvate (PEP) through glycolysis in the cytosol, and then imported back into the chloroplast, likely via a phophoenolpyruvate/phosphate translocator (PPT) (Flugge 1999). While 14

PEP transport into chloroplasts has not been directly observed, observed affinities for PEP and inorganic phosphate (Pi) of the isolated PEP-Pi Transporter (PPT) from chloroplast envelopes, and lack of a glycolytic sequence capable of converting hexose-phosphates into PEP within the chloroplast, have led to the inference that PEP is imported into C3 chloroplasts from the cytosol

(Fischer et al. 1997). Once in the chloroplast, PEP is converted to Pyr by pyruvate kinase (see reviews in Flugge 1999 & Logan et al. 2000). Because of its direct connection to photosynthetic

CO2 assimilation, especially through the use of G3P as a primary substrate, labeling of

13 13 photosynthetic products with CO2 causes the C isotope to appear rapidly in emitted isoprene

(Delwiche & Sharkey 1993). One mystery yet to be fully resolved, however, is why a significant fraction (~20% on average) of emitted isoprene carbons remain unlabeled with 13C even after

13 several hours of exposure to CO2 (Karl et al. 2002). Using poplar leaves, Schnitzler et al.

(2004) showed that alternative sources of carbon substrate exist for isoprene production in poplars, including compounds transported as carbohydrates in the xylem stream (potentially from stored carbon in roots) and starch stored in chloroplasts. Also using poplar leaves, researchers have observed an increase in the fraction of unlabeled isoprene during high-temperature or severe water stress, and attributed this to greater reliance on 'stored' carbon, perhaps from extra- chloroplastic sources, during periods when photosynthesis rates decrease (Funk et al. 2004; Brilli et al. 2007). Thus, there is clear evidence for the use of stored 'older' carbon to support isoprene biosynthesis, and there is also evidence that plants adjust their reliance on this older carbon depending on environmental conditions.

One environmental condition that has been shown to influence isoprene emission rate, especially from poplar leaves, is the atmospheric CO2 concentration at which the plants are grown and measured (e.g. Sanadze 1964; Monson and Fall 1989; Rosenstiel et al. 2003; 15

Wilkinson et al. 2009). These past studies have shown a negative correlation between photosynthetic rate and isoprene emission rate when exposed to altered CO2 concentrations. A number of studies have demonstrated that while photosynthesis rates increase with increasing

[CO2], isoprene emissions decrease (Monson & Fall 1989; Loreto & Sharkey 1990). Isoprene production is rarely limited by photosynthetic assimilated carbon, where net CO2 assimilation fluxes on the scale of µmol m-2 s-1 are more than enough to sustain isoprene emissions typically reported on the scale of nmol m-2 s-1. This suggests that isoprene emission rates are affected via

CO2 concentrations altering DMADP substrate availability and/or isoprene synthase activity or quantity. The biochemical mechanism responsible for this effect has not been fully resolved, but research suggests that competition between cytosolic and chloroplastic processes for available

PEP substrate plays a role, with an increase in atmospheric CO2 concentration shifting competition in favor of cytosolic processes (Rosenstiel et al. 2003, 2004; Loreto et al. 2007).

Supporting this hypothesis, Possell & Hewitt (2011) recently demonstrated an increase in PEP carboxylase activity with increasing CO2 concentrations and a concomitant decline in DMADP content. While the same PEP carboxylase activity levels were exhibited by trees grown under sub-ambient CO2 as those grown under normal, ambient CO2, a decreased demand for cytosolic

PEP under these conditions may still result in a flux of carbon into the chloroplast. Furthermore, they showed that an application of fosmidomycin (a competitive substrate inhibitor of the second enzymatic step in the MEP pathway) to plants grown under sub-ambient and elevated CO2 conditions resulted in isoprene emission rates statistically similar to those grown under ambient

CO2 conditions. Together, these data support the hypothesis that differences in DMADP biosynthesis rate observed among CO2 treatments are due to changes in pyruvate and G3P availability for the MEP pathway. However, there remains much to discover about the 16 biosynthetic kinetics and mechanisms by which G3P and Pyr are utilized for isoprene production.

One way to understand how changes in atmospheric CO2 affect isoprene biosynthesis is to evaluate the contribution of different carbon sources to competing metabolic processes under

13 13 various CO2 regimes. Studies capable of resolving CO2 labeling kinetics, and movement of C through precursor pools to isoprene biosynthesis, were improved considerably by the development of the proton-transfer reaction mass spectrometer (PTR-MS). The PTR-MS approach, which allows for near-continuous measurement of compound masses in air flowing through a leaf gas-exchange chamber (Karl et al. 2002; Schnitzler et al. 2004; Ghirardo et al.

2010), was further improved by the discovery that increasing the electric field within the drift tube led to unique compound fragmentation patterns (for more detail see Materials and methods). These patterns increased the potential to observe changes in 13C labeling of a specific fragment (M41+)that was confirmed to be the 3-carbon methyl-vinyl fragment of the 5-carbon isoprene molecule (Karl et al. 2002). The methyl-vinyl fragment contains two carbons contributed to isoprene biosynthesis from the Pyr substrate and one carbon from G3P. Thus, use of the PTR-MS makes it possible to track the labeling kinetics of not only the whole isoprene molecule, but also the fragment that contributes carbon from the labeled Pyr pool. We used this approach to study 13C labeling dynamics in the leaves of poplar trees grown and measured under different CO2 environments (190 ppmv, 400 ppmv, and 590 ppmv). Our goal was to 1) resolve labeling kinetics in the whole isoprene molecule, the methyl-vinyl fragment, and, by inference, leaf pyruvate pools, to determine if different CO2 growth conditions influence the use of specific carbon sources for isoprene biosynthesis and 2) to elucidate the potential pathway through which 17 the flow of carbon from these various sources is ultimately incorporated into isoprene via pyruvate as opposed to G3P substrate.

MATERIALS AND METHODS

Plant material and growth conditions

We grew hybrid poplar trees (Poplar x canescens; syn. Populus tremula x P. alba) in

Germany from cuttings that were initially grown in small pots with sterile sand. The misting rooms were circulated with ambient air at 24 °C and maintained a photoperiod of 14h at a PPFD of 200 μmol m-2 s-1 and 70% relative humidity (RH). Once roots formed, plants were shipped to the Duke University Phytotron, transferred to pots filled with 1:1:1 (v:v:v) sand:perlite:peat, and placed in one of three growth chambers (Model M-13, Environmental Growth Chambers,

Chagrin Falls, OH). Plants were then grown at 27:23 °C day:night temperatures with 16:8 day:night photoperiods at a PPFD of 700 µmol photons m-2 s-1 at canopy level and 50% relative humidity at peak leaf area. Plants were also grown under one of three CO2 concentrations, 190 ppm, 400 ppm, or 590 ppm, with the diurnal range in concentration being less than 10 ppmv. The

CO2 concentration of chamber air was measured with an infra-red gas analyzer (LiCor 6252,

Lincoln, NE) every 2-5 minutes throughout the growth period. The elevated CO2 environment was created by injection of pure CO2 into the air stream as needed, whereas low CO2 concentrations were maintained by scrubbing the incoming air with soda lime before injecting it with CO2. The trees were cut to just above soil level after growth under the CO2 treatments for 2 months. After growing again for two months, the trees were trimmed several nodes above the soil, and then allowed to re-grow for one month prior to making measurements. We made measurements on a leaf two nodes below the second trim point, one month after trimming. These 18 leaves were estimated to be three months old and were fully expanded. The isoprene emission rates for these leaves generally ranged from 4.5 – 5.5 nmol m-2 s-1 when measured under ambient

CO2 at 30 ˚C. These rates are within the same range of isoprene emission rates observed for leaf

Node 9 in the studies of Behnke et al. (2007) on the same wild-type poplar lines used in this study, and which represented leaves with fully-matured isoprene emission capacity (Behnke et al. 2007). The CO2 treatments were rotated among the three chambers every three weeks to minimize chamber effects, and plants were moved from spot-to-spot within each chamber on a weekly basis to minimize spatial biases on growth. For this study, we used seven trees from each growth chamber, 21 plants in total.

Leaf gas-exchange measurements

Point measurements of baseline gas exchange were made throughout the experiment using a portable photosynthesis system and a standard 6 cm2 leaf chamber equipped with a programmable LED light source (LI-6400, LiCor Inc., Lincoln, NE, USA). Because the Licor

13 12 instrument is differentially sensitive to CO2 and CO2, it was impossible to simultaneously measure physiological responses and isoprene emission rates during labeling. Therefore, one leaf two nodes below the second trim point (one month after trimming) was measured from trees grown under elevated (590 ppm) and sub-ambient CO2 (190 ppm) either prior to or after the labeling experiment. These measurements were made under light-saturating conditions (1,000

-2 -1 µmol photons m s ) at 30 ºC under the growth CO2 conditions of the plant. Sampling occurred at three points during the day (early morning, midday, and late afternoon) to obtain a daily mean for net CO2 assimilation rate (A), stomatal conductance rate (gs), and intercellular CO2 concentration (Ci). These data offer an appropriate framework in which to understand how 19

altered CO2 growth conditions affect the average carbon metabolism of poplars and subsequently, their average isoprene biosynthesis kinetics and emission rates. Separate from the

PTR-MS isoprene measurements (described below), simultaneous isoprene emission rates were measured relative to the gas exchange measurements by diverting a fraction of the outgoing air from the leaf cuvette to a chemi-luminescence based fast isoprene sensor (FIS) (Hills Scientific,

Boulder, CO, USA).

Proton Transfer Reaction-Mass Spectrometry

To determine the kinetic dynamics at which 13C progressively replaced 12C in the isoprene molecule and its fragment, we combined a LiCor 6400 cuvette system (LiCor Inc., Lincoln, NE,

USA) with a proton-transfer reaction mass spectrometer (PTR-MS) (Ionicon GmbH, Innsbruck,

Austria). This unique system was used to determine the isoprene concentration in the outgoing cuvette air and the mass variants (isotopomers) of isoprene and associated fragments, which

12 13 reflected the time-dependent turnover of C after labeling the air with CO2.

One leaf (2 nodes below the trim point) per individual was placed in a LI-6400 portable photosynthesis system leaf chamber as described above (LiCor, Inc., Lincoln, Nebraska). The flow of air through the cuvette was 350 ml min-1, which was sufficient to allow for a reasonable

‘wash-in’ and ‘wash-out’ time during the labeling experiment. Approximately 50 ml min-1 of cuvette air was diverted to the PTR-MS using Teflon tubing. Inlet air to the cuvette was obtained from an air source that was mixed each day using a clean air generator (model 737, Aadco Inc.,

Cleves, OH, USA) with an activated charcoal scrubber on the outlet to ensure air purity, and the

12 13 addition of either CO2 (Airgas, Inc., Durham, NC) or CO2 (Cambridge Isotope Laboratories,

Inc., Andover, MA) through manual injection. The volume of each unlabeled or labeled CO2 20

source gas required to create the appropriate CO2 concentration (v/v) for each treatment was calculated for 190 ppm, 400 ppm, and 590 ppm assuming 90 L of VOC (volatile organic compound)-free air. Depending on the treatment, either 17.1 mL, 36 mL, or 53.1 mL,

12 13 respectively, of source CO2 or CO2 were injected into an empty 100L Tedlar bag (CEL

Scientific Corp, Santa Fe Springs, CA), and VOC-free air was immediately pumped in at 4.5

L/min for 20 minutes. This mixed air source was then stored in the large Tedlar bag, and slowly evacuated as air was needed for the experiment. Positive pressure was maintained in the cuvette to prevent the ingress of contaminants. During the measurement period, leaves in the cuvette were maintained at 30 °C leaf temperature, 1000 μmol m-2 s-1 photosynthetically active radiation

(PAR) and 50-70% relative humidity (RH). In each experiment, leaves in the cuvette were fed air

12 with CO2 at the same concentration as their growth environment until a stable isoprene emission rate was observed. Once a stable point was reached, the air source was switched to

13 CO2 at the same concentration, and time-dependent changes in isoprene mass variants were

13 monitored. Once all labeled isoprene mass variants/signals were stable, the CO2 supply was

12 switched back to CO2. Due to complications with differential sensitivity of the LiCor CO2

13 12 13 sensor to CO2 and CO2, we were not able to measure the CO2 concentration of the air in the

Tedlar bag, which was used as the labeling source. However, it was mixed to specifications to

12 provide 190 ppm, 400 ppm, and 590 ppm CO2, respectively, and we did check the CO2 concentration in the chamber air before and after the labeling to confirm that this was within 20 ppmv of the target values.

The PTR-MS instrument design and underlying principles of operation have been described previously in detail (Lindinger, et al. 1998). For this study, the instrument was operated at an

E/N of 140 Td to induce a high degree of fragmentation. Operating the instrument under these 21 conditions changed the mass spectrum to favor higher production frequencies of the 3-C fragment from isoprene measured at M41+ (32.6%) and M42+ (1.1%), as opposed to lower frequencies observed under normal operating conditions that produce 7.1% and 0.2% M41+ and

M42+, respectively (Karl et al. 2002). The 3-C fragment is known to be derived from the methyl- vinyl component of isoprene, which is originally derived from pyruvate (rather than directly from glyceraldehyde 3-phosphate) as it enters the MEP pathway (Lichtenthaler 1999; Karl et al.

2002). The drift tube pressure, temperature, and voltage were 1.96hPa, 60˚C, and 550V,

+ respectively. The count rate of H3O H2O ions measured before labeling was less than 1% of the

+ 6 -1 count rate of H3O ions, which was 9.2-10.9 x 10 counts s . The PTR-MS was calibrated by generating a standard curve for both M41+ and M69+ after measuring the counts per second (cps) of these masses with different known isoprene concentrations, which were created by dilution of a 1.5 ppmv isoprene standard (Scott-Marrin, Inc., Riverside, CA) with humidified VOC-free air at the beginning of each day’s experiment. The detection limit of the PTR-MS based on the calibration of m/z 69+ was 55 ppt / normalized count per second. Leaf isoprene fluxes were calculated as:

f (Ca Ci ) J A

Where f is the flow rate through the cuvette (ml min-1) and A is the leaf area enclosed within the

2 cuvette (cm ). Ca – Ci is the difference in gas partial pressure between the empty and leaf-filled cuvette, expressed in nmol mol-1.

Calculations and Statistical Analysis

To compare the labeling dynamics of both the 3-carbon fragment, M41+, and the 5-carbon parent molecule, M69+, we measured the time dependent change in the proportion of 13C 22 simultaneously incorporated into each. To do this we 1) calculated the total labeled carbon atoms

(13C) (on a molar basis) in both M41+ and M69+ at each point in time and 2) determined the total isoprene emission rate for each point in time. Total labeled carbons were determined for the methyl-vinyl fragment (M41+) by summing the products of the mass variants’ emission rates

(M42+, M43+, and M44+) and the number of labeled carbons their detection represents (1, 2, and

3, respectively). For example, at any point in time the total labeled carbons in the M41+ 3-C methyl-vinyl fragment was equal to {(M42+) + (M43+ x 2) + (M44+ x 3)}, whereas the total labeled carbons in the 5-C parent molecule was equal to {(M70+) + (M71+ x 2) + (M72+ x 3) +

(M73+ x 4) + (74+ x 5)}. The total labeled carbon was then divided by the total isoprene emission rate at each point during the experiment (before and after labeling) to obtain and compare the number of labeled carbons in both the fragment and parent molecule simultaneously through time. The number of labeled carbons in both the parent molecule and the fragment were graphed together over time (before and after labeling) but separately for each treatment. Results were based on a qualitative assessment of how the lines diverged in time relative to the number of labeled carbons plotted on the y-axis. If labeled carbon (13C) was added to the parent molecule

(M69+) through the methyl-vinyl fragment (M41+), then both lines would increase simultaneously. If labeled carbon was being incorporated into the parent molecule faster than it is appearing in the fragment, a divergence of the M41+ line from the M69+ would result.

To evaluate the effects of different [CO2] on the labeling rate of whole-isoprene and the

M41+ fragment (representing the substrate originating from pyruvate), we had to account for the simultaneous gain and loss of labeled carbons (13C) as they successively moved through the observed mass fractions. Conceptually, this can be accomplished by considering the rates of change of each isotopomer in terms of the “loss” of carbon from the mass preceding it. For 23 example, as M72+ gained a labeled carbon and became M73+, this subsequently caused an equal

“loss” of M72+. Observing the labeling of M72+ graphically over time would then show a positive slope (representing the “gain” of 13C into M71+, thus producing M72+) followed by a peak and a subsequent shorter negative slope (denoting the “loss” or movement of 13C into the higher mass, in this case M73+). With the data expressed in its raw form exhibiting two slopes, as just described, it would be extremely difficult to obtain an accurate rate of labeling for this mass variant. Therefore, to accurately account for movement of 13C between masses, the slopes, or rates of 13C gain and loss for each mass from the start of labeling until the concentration of all masses reached steady state in the air leaving the cuvette, were summed (denoted by an “s” to distinguish from individual masses) according to: sM70+ = sum(M70++ M71++ M72++ M73++

M74+), which represents the rate of labeling of isoprene where isoprene has at least 1 carbon labeled; sM71+ = sum(M71++ M72++ M73++ M74+), representing the rate of labeling of isoprene molecules that have at least 2 carbons labeled; and sM72+= sum(M72++ M73++ M74+), representing the rate of labeling of isoprene molecules that have at least 3 carbons labeled, and so on. The same calculations were applied to the fragment M41+ where sM42+ = sum (M42++

M43++ M44+), representing the rate of labeling of isoprene’s fragment where at least one carbon on this 3-C subunit is labeled; and sM43+ = sum (M43++ M44+), representing the rate of labeling of the 3-C subunit that have at least 2 carbons labeled, and so on. The slopes of each individual signal (expressed as molecules/cycle, where each cycle represents every 30 seconds and a PTR-

MS dwell time of 2 seconds) were calculated using a generalized linear model with an identity link function that provides the relationship between the linear predictor and the mean of the distribution function, which in this case is normal. The effect of the three different growth and 24

measurement CO2 concentrations on the rate of labeling of each analog was then analyzed using a one-way analysis of variance (ANOVA).

To determine the proportion of each analog of both the parent isoprene molecule and the methyl-vinyl fragment that was labeled at the end of the experiment, the emissions for each analog at each point during the last 30 minutes of the total 1.5-2 hour labeling time (once the mass had become stable and reached its maximum labeling) were divided by the total emission rate at that same point in time (sum(M69+ through M74+)). A one-way ANOVA was performed to determine if proportions of labeling (in both the parent and fragment molecules) differed for each analog between treatments. Isoprene emission rates for individuals were calculated by

+ 12 averaging the steady-state emission of M69 during exposure of the leaf to CO2 prior to the

13 CO2 labeling. A one-way ANOVA was used to evaluate differences in isoprene emissions between treatments. A Student’s t test was used to determine differences in net CO2 assimilation rates (A), stomatal conductance rates (gs), and intercellular CO2 concentrations (Ci) between trees grown under sub-ambient and elevated CO2 conditions. Finally, linear regression models were used to evaluate the relationship between isoprene emission rate and each of the physiological parameters described above. All statistical analyses were performed with R

(2.10.1, Vienna, Austria).

RESULTS

Physiological data and total isoprene emission rates

Leaf gas exchange measurements were made concurrently with isoprene emission rates from trees grown under sub-ambient and elevated CO2 conditions. Trees grown under elevated CO2

-2 -1 had significantly higher net CO2 assimilation rates (11.74 ± 0.88 µmol m s ) and intercellular 25

-1 CO2 concentrations (428.11± 11.52 µmol mol ) than trees grown under sub-ambient CO2 conditions (5.88 ± 0.66 µmol m-2 s-1 and 145.5 ± 3.12 µmol mol-1, respectively; n=6 and

P<0.0001) (Figure 2.1A and B). It should be noted that all values, unless otherwise specified, are reported as the mean ± standard error of the mean, sample size, and probability of Type I error, respectively. Furthermore, trees grown under elevated CO2 exhibited significantly lower stomatal

-2 -1 conductance rates (0.162 ± 0.019 mol m s ) compared to those grown under sub-ambient CO2

(0.348 ± 0.162 mol m-2 s-1; n=6, P<0.0001) (Figure 2.1C). We used linear regression models to evaluate whether significant relationships exist between each of the three physiological parameters described above and total isoprene emission rates, which were measured simultaneously. Results indicated a significant positive linear relationship between stomatal conductance and isoprene emission rate (R2=0.58; P<0.0001) and a significant negative linear

2 relationship between intercellular CO2 concentration and isoprene emission rate (R =0.64;

2 P<0.0001) as well as between net CO2 assimilation rate and isoprene emission rate (R =0.18;

P=0.006). Individual regressions were performed on all data for each continuous variable (i.e. after combining values from trees grown under both sub-ambient and elevated CO2 conditions).

Figure 2.1 Photosynthesis, intercellular CO2, and stomatal conductance for poplars grown under sub-ambient and elevated CO2 regimes. Net CO2 assimilation rate (A), intercellular CO2 concentration (B), and stomatal conductance rate (C) for trees grown under sub-ambient (190 13 ppm) and elevated (590 ppm) CO2 conditions prior to CO2 labeling experiment. Error bars represent the standard errors of the mean (SEM) and means with (***) are significantly different (P<0.0001). Trees grown under elevated CO2 exhibited significantly higher intercellular CO2 concentrations and net CO2 assimilation rates relative to poplars grown under sub-ambient CO2 conditions. Conversely, trees exposed to elevated CO2 growth conditions demonstrated significantly lower stomatal conductance rates compared to those grown under sub-ambient CO2. 27

Qualitative analysis of recently assimilated carbon incorporation into isoprene

When comparing the simultaneous change in the number of labeled carbons in both the

+ + fragment (M41 ) and parent molecule (M69 ) among the three CO2 regimes, both lines fall directly on one another, which indicated an immediate labeling of the first carbon (Fig. 2.2). This result was consistent with the hypothesis that the 13C that initially enters the isoprene pool is recovered in the 3-C methyl-vinyl fragment. Keep in mind that M41+ and M69+ were initially

13 13 13 unlabeled compounds with respect to C. As time in the presence of CO2 progressed, if C was added to M69+ through the methyl-vinyl fragment, then both M41+ and M69+ should have increased simultaneously. While Figure 2.2 shows this to be the case for the labeling of the first carbon, as time progressed and a second carbon became labeled on the parent molecule, the

M41+ lines diverged for leaves grown under all three treatments. Because the labeling happened faster for the second carbon in the isoprene molecule than for the second carbon in the methyl- vinyl fragment, it appears that second labeled carbon on the isoprene molecule was not coming from the M41+ subunit. This suggests relatively fast incorporation of MEP substrate derived directly from G3P rather than from pyruvate. Furthermore, the lines first begin to diverge most quickly for leaves from the low and ambient CO2 treatments (after ~1 carbon was labeled), while the line of the high CO2 treatment leaves first begin to diverge later (after ~1.5 carbons were labeled) (Fig. 2.2). This result was consistent with a faster incorporation of 13C into pyruvate in the leaves of plants grown under high CO2.

13 + + Figure 2.2 CO2 labeling of M41 and M69 as a function of CO2 concentration and time. The number of labeled carbons present in both the M41+ fragment and M69+ parent isoprene molecule prior to and after 13C labeling with error bars representing the standard error of the mean (SEM). Change in the lines are evaluated in reference to the number of labeled carbons in both the fragment and parent molecule over time, with lines falling on one another representing labeling occurring simultaneously in both molecules and a divergence representing a faster label + + 13 incorporated into M69 that is not derived from M41 . Before leaves were exposed to CO2 12 labeling at 1000 seconds, plants were exposed to the same CO2 concentrations at which they were grown. As expected, no labeling occurred for either M41+ (closed circles) nor M69+ (open circles) before labeling. Immediately after labeling, one carbon was labeled in both the parent molecule and the fragment (demonstrated by the simultaneous increase in both lines), suggesting that the first carbon used to synthesize isoprene is contributed from the M41+ fragment. However, as time progressed and a second carbon becomes labeled on the parent molecule, the M41+ lines diverged for leaves grown in all three treatments, suggesting that the second labeled carbon on the isoprene molecule is not coming from the M41+ subunit. 29

13 Exposure of poplar leaves to CO2 elucidates the contribution of pyruvate-derived versus different carbon sources to isoprene synthesis over time

To examine the relative contribution of pyruvate-derived carbons for isoprene synthesis, we

13 + analyzed real-time CO2 labeling kinetics for both the parent isoprene molecule (M69 ) and its

+ 3-C fragment (M41 ) for a representative tree grown under ambient CO2 of 400 ppm (Figure

2.3). By examining simultaneous changes between labeling in the parent molecule (Figure 2.3A) with changes in the 3-C fragment (Figure 2.3B), one could obtain a detailed account of the sequence with which labeled carbons are contributed to isoprene synthesis via the 3-C (methyl- vinyl) or 2-C fragments of the fractured isoprene molecule. From Figure 2.3, there was an

+ + 13 immediate and extremely fast increase in the sM42 and sM70 signals following CO2 labeling.

This result confirmed that isoprene emission rate was closely coupled to carbon assimilation and suggested that the addition of labeled carbon to the 3-C fragment had the same consequence for the signal of the parent molecule. This result also supported the conclusions from the labeled carbon data for both M41+ and M69+, which showed that the first labeled carbon transferred into the isoprene pool was recovered as part of the methyl-vinyl subunit. We note that there was a slight increase in the total isoprene emission rate and total amount of 3-C methyl-vinyl isoprene

12 13 fragment detected when we switched from the CO2 source to the CO2 source. This effect was detected in all treatments, and was small in magnitude compared to the differences in total emission rates. It is possible that the CO2 concentration was slightly lower than desired in the

13 CO2 source, but we were careful to prepare this source according to precise calculations. It is also possible that there was a small decrease in flow rate through the chamber when the sources were changed, and that the flow controller we used was differentially biased toward the presence

13 of CO2. 30

13 + + Figure 2.3 CO2 labeling of carbon atoms in M69 and M41 and their isotopomers through 13 time. (A) CO2 labeling of carbon atoms in trees grown and measured in ambient CO2 conditions (400 ppm CO2) in the parent isoprene molecule, as characterized by a decrease in the M69+ signal (orange circles) and simultaneous increase in its isotopomers (denoted as sums) as labeled carbons were successively incorporated through time. Total emission (blue circles), sM70+ (red downward triangles), sM71+ (green triangles), sM72+ (yellow squares), sM73+ (sea + 13 green squares), sM74 (purple diamonds) are represented. (B) CO2 labeling of carbon atoms in trees grown and measured at 30 ºC in ambient CO2 conditions (400 ppm CO2) in the 3-C methyl- vinyl isoprene fragment, characterized by a decrease in the M41+ signal (light orange dotted downward triangles) with a simultaneous increase in its labeled isotopomers (denoted as sums). Total emission (blue dotted squares), sM42+ (pink crossed circles), sM43+ (green hexagons), + 13 sM44 (yellow diamonds) are represented. Before leaves were exposed to CO2 labeling at 1000 12 seconds, plants were exposed to the same CO2 concentrations at which they were grown. The simultaneous labeling of the first carbon in the parent molecule (sM70+) and the fragment (sM42+) suggest that the first carbon contributing to the synthesis of isoprene comes from the M41+ fragment. However, while all of the isoprene molecules show the next two carbons labeled shortly after (sM71+ and sM72+), the next two carbons on the M41+ fragment (sM43+ and sM44+) are never fully labeled and may result from the incomplete labeling of pyruvate. 31

The fast incorporation of recently assimilated carbon into isoprene via the 3-C methyl-vinyl fragment was then followed by an equally fast incorporation of a second labeled carbon, as characterized by the slope exhibited in sM71+ (Fig. 2.3A). Eventually, at least 2 labeled carbons occurred in all isoprene molecules emitted, regardless of CO2 growth conditions (data not shown), clearly demonstrated as the sM71+ signal plateaued at the maximum total emission value. However, the labeling data for the sM43+ fragment (Fig. 2.3B) revealed that, at maximum labeling, not all of the molecules in the 3-C methyl-vinyl fragment had at least two 13C, even though there were at least 2 carbons labeled in all the parent molecules. These observations supported the conclusion that the second isoprene carbon to be labeled came from a source that was not part of the methyl-vinyl fragment containing carbon from pyruvate and presumably originated from the movement of labeled carbon into the MEP pathway through the direct incorporation of G3P substrate. Likewise, the same argument could be used to demonstrate that the third carbon contributed to isoprene biosynthesis did not originate from the methyl-vinyl fragment, indicative of pyruvate. Again, it was evident that all of the emitted isoprene eventually contained at least 3 labeled carbons, as illustrated by the sM72+ signal reaching the total emission plateau (Fig. 2.3A). Yet, not all of the 3-C methyl-vinyl fragments obtained from isoprene were completely labeled, as illustrated in the sM44+ signal not reaching the maximum total value (Fig. 2.3B).

Because we have not accounted for the 2 remaining carbons on the 3-C methyl-vinyl fragment, several lines of reasoning led us to deduce that these carbons were labeled last in the sequence of 13C transfers into the isoprene pool, with this 'slow' transfer having occurred through the pyruvate substrate pool. First, we know that two of the carbons on the 3-carbon fragment came from pyruvate. Additionally, as discussed above, the one carbon atom in this fragment that 32 was derived from G3P was also, based on knowledge of photosynthetic metabolism, the first

13 carbon within the G3P pool to have been labeled after the assimilation of CO2. Second, at the end of the experiment (after ~2 hours), the labeling data showed that the sM43+ and sM44+ fragments never reached the pre-labeling maximum exhibited by M41+, despite the fact that the abundance of M41+ eventually went to zero. This meant that some of the 3-C methyl-vinyl fragments only had one or two 13C-labeled carbons, not three (Fig. 2.3B). On the basis of this evidence, we concluded that C-1 of the G3P molecule carried the 13C label through the MEP pathway and into isoprene first, followed by incorporation of 13C through the C-2 and C-3 carbons of G3P substrate and that any 13C that entered through pyruvate came later. Furthermore, it appeared that the fraction of emitted isoprene carbon that remained unlabeled, even after

13 several hours in the presence of CO2, most likely originated from pyruvate carbon.

13 Rates of CO2 labeling and the proportion of labeling at steady state between CO2 treatments

Because on-line PTR-MS can distinguish individually labeled isoprene species during 13C labeling, we measured the rates at which each mass variant appeared and reached steady state.

This, in turn, allowed us to estimate the rates of 13C transferred from 13C-labeled photosynthate into isoprene as the rate of mass loss from the M69+, M70+, M71+, M72+ and M73+ signals. The rate of transfer of 13C into isoprene was ~2 times faster for the first four masses in the leaves of poplars grown and measured under sub-ambient CO2 conditions, compared to those grown and measured under ambient and elevated CO2 conditions (Fig. 2.4). The rate of transfer showed the same trend in the loss of M73+, compared to the other masses, but the trend was not statistically significant. Similarly, the rate of mass loss for M41+, M42+ and M43+ was approximately twice as fast for the leaves grown under sub-ambient CO2, compared to the other two treatments (Fig. 33

2.4 inset). Results above demonstrate that trees grown under sub-ambient CO2 exhibited net CO2 assimilation rates ~2 times lower than trees grown under elevated CO2 (Fig. 2.1A) despite having

13 higher stomatal conductance rates (Fig. 2.1C). Thus, recently assimilated CO2 was transferred at a greater rate into isoprene in leaves grown under sub-ambient CO2 compared to leaves grown under elevated or ambient CO2, despite having lower net CO2 assimilation rates.

The slope data illustrate how quickly pools contributing carbon to isoprene production became labeled (Fig. 2.4); however, it is also necessary to consider the proportion of each mass that becomes completely labeled (Fig. 2.5). While carbon pools that contributed to isoprene production in trees grown under sub-ambient CO2 exhibited the fastest initial rates of labeling and reached steady state more quickly than trees grown under higher CO2, sub-ambient CO2 trees also had the lowest proportion of total isoprene molecules completely labeled (0.413 ±

+ 0.026, n=7, Fig. 2.5 M74 ), a value only two-thirds that for trees grown under elevated CO2

+ (0.635 ± 0.014, n=7, P<0.0001, Fig. 2.5 M74 ). Furthermore, trees grown under ambient CO2 had significantly less isoprene completely labeled (0.474 ± 0.040, n=5, P<0.01, Fig. 2.5 M74+) compared to elevated CO2 trees, but were not significantly different from trees grown under sub- ambient CO2 (P=0.302). A very similar pattern of labeling was also demonstrated by the methyl- vinyl fragment where trees grown under sub-ambient CO2 had the lowest proportion of total fragment molecules completely labeled (0.587 ± 0.022, n=7), and this proportion was about three-fourths that for trees grown under elevated CO2 (0.767 ± 0.029, n=7). Also similar to labeling occurring in the parent molecule, trees grown under ambient CO2 had significantly less isoprene completely labeled (0.634 ± 0.031, n=7, P<0.05) compared to elevated CO2 trees, but were not significantly different from trees grown under sub-ambient CO2 (P=0.474).

13 Figure 2.4 Rates of C transference between isotopomers as a function of CO2 concentrations. Mean rates of loss of the labeled isotopomers in units of molecules/cycle (cycle = detection every 30 seconds with a PTR-MS dwell time of 2 seconds) for both the parent molecule M69+ + and its fragment M41 (inset graph) among individuals grown at three different CO2 concentrations (sub-ambient = 190 ppm (black triangles; dashed line); ambient = 400 ppm (dark gray circles; dashed line); elevated = 590 ppm (light gray squares; solid line)). In general, the photosynthetic pools of the leaves grown in sub-ambient CO2 were labeled faster than leaves grown at ambient or elevated CO2.

Figure 2.5 The mean proportion of the isotopomers of the parent isoprene molecule labeled at the conclusion of the experiment. Values were taken at stabilized conditions after 2 hr. Leaves 13 grown in sub-ambient CO2 demonstrated significantly lower proportions of total C labeling (M74+) compared to the high proportion of total labeled isoprene molecules from leaves grown in elevated CO2.

Total isoprene emission rates

Before the labeling treatment, we measured isoprene emissions of poplar leaves from trees grown under all three CO2 regimes. While the total isoprene emission rates between trees grown under ambient and elevated CO2 were not significantly different (n=7, P=0.946), the trees grown

-2 under sub-ambient CO2 had significantly higher isoprene emission rates (9.41 ± 0.39 nmol m s-1) than rates exhibited by trees grown under ambient (5.04 ± 0.35nmol m-2 s-1; n=5, P<0.01) or

-2 -1 elevated (4.86 ± 0.46 nmol m s ; n=7, P<0.001) CO2 (Fig. 2.6).

Figure 2.6 Mean isoprene emission rates measured as the total of M69+ prior to the labeling experiment for poplars grown under 3 different CO2 regimes (sub-ambient = 190 ppm (black bar); ambient = 400 ppm (dark gray bar); elevated = 590 ppm (light gray bar)). Error bars represent the standard errors of the mean (SEM) and means with the same letter are not significantly different (P<0.05). Trees grown in sub-ambient CO2 demonstrated significantly higher isoprene emission rates compared to trees exposed to ambient and elevated CO2 concentrations.

DISCUSSION

The goal of our study was to provide insight into the regulatory mechanisms controlling isoprene production, particularly the contribution of carbon from recently-assimilated CO2. The tracking of recently-assimilated CO2 into isoprene biosynthesis has been accomplished in past studies using gas chromatography-mass spectrometry (GC-MS) (Loreto et al. 1996), as well as the PTR-MS approach (Karl et al. 2002; Brilli et al. 2007). However, our study is the first to apply the PTR-MS approach to the issue of how carbon allocation to isoprene emission changes when trees are grown at different atmospheric CO2, including a level expected to be reached in the next few decades due to continued fossil fuel burning. Furthermore, we aimed to clarify the potential roles of pyruvate compared with G3P as substrates that control the response of isoprene biosynthesis in different growth and measurement CO2 concentrations. We highlight two major conclusions: (1) When the rate of photosynthetic assimilation of atmospheric CO2 decreases, due to limited availability of CO2, emitted isoprene molecules show more evidence of biosynthetic construction from stored (older) carbon sources than from recently-produced photosynthate. This is consistent with the results from Funk et al. (2004), though in that case net CO2 assimilation was limited by severe water and temperature stress (Funk et al. 2004). (2) The flow of carbon from alternative, older sources most likely enters the MEP pathway through the pyruvate substrate, rather than the G3P substrate. This latter conclusion provides insight into the central role of cytosolic PEP as a control point for the channeling of carbon from different sources into isoprene biosynthesis. In the next few paragraphs, we expand on each of these two points.

Poplar leaves grown under elevated CO2 emitted a significantly higher fraction of isoprene molecules completely labeled with recently assimilated 13C than did leaves grown under ambient or sub-ambient CO2. This is consistent with the significantly higher internal CO2 concentrations 39

and carbon assimilation rates found in poplars grown under elevated CO2 conditions.

Furthermore, carbon pools contributing to isoprene biosynthesis were labeled ~2 times slower in trees grown under elevated CO2 conditions, so that overall, they displayed the lowest isoprene emission rates and slowest initial labeling but had the largest proportion of isoprene completely labeled at steady-state. We interpret these results as indicating that the rate of carbon utilization for isoprene production was relatively low in the trees grown at elevated CO2, allowing the

13 availability of photosynthate produced from recently-assimilated CO2 to be closer to the margin required to support that low utilization rate, compared to trees grown under sub-ambient CO2. In trees grown under sub-ambient CO2 where the rate of carbon utilization for isoprene was low, the

13 availability of recently assimilated CO2 was likely significantly below the margin required to support isoprene emission. This would have forced greater reliance on older, stored carbon substrate.

An alternative explanation for the decrease in 13C labeling of emitted isoprene from leaves grown at sub-ambient CO2 is that there is always some level of incomplete labeling of photosynthetic intermediates in the pentose phosphate pathway. It has been known for many years that intermediate compounds in the photosynthetic carbon reduction cycle can approach an asymptote of approximately 90% labeling when exposed to labeled atmospheric CO2, and this was shown most recently in the study by Hasunuma et al. (2010) using Nicotiana tabacum

13 leaves. Past studies using nuclear magnetic resonance, accompanied with CO2 labeling of soybean leaves, have shown that dilution of the 13C-labeled PGA pool by unlabeled glycerate from photorespiration causes a delay in full 13C labeling of photosynthetic metabolites (Cegelski

& Schaefer 2006). However, this study also showed that the lag was brief, and that by 8 minutes

12 13 after switching from CO2 to CO2, photosynthetic metabolites were 95% labeled. In our 40

present study, only 41 % of the emitted isoprene from leaves grown under sub-ambient CO2 was

13 completely labeled after two hours of exposure to CO2. While it is possible that this lack of complete labeling in isoprene could be due to unlabeled G3P which originates from unlabeled intermediates of the photosynthetic carbon reduction cycle and subsequently enters the MEP pathway, this explanation would not account for the fact that the greatest fraction of incomplete labeling in the isoprene molecule occurs in the fragment derived from pyruvate. Logic would lead us to expect that a carry-over of incomplete labeling from the photosynthetic carbon reduction cycle should show up in the G3P derived fragment of isoprene first. This is not the case. The simplest explanation is that there exists a constant channeling of unlabeled, older stored carbon into isoprene biosynthesis, through the pyruvate substrate, which dilutes the 13C- labeling of isoprene to a steady-state value that varies as a function of CO2 availability. The source of this stored carbon remains to be identified.

Our results suggest that the C-1 carbon of G3P, which would be the first to be labeled after

13 assimilation of CO2, is transferred quickly to isoprene and thus appears quickly as C-1 of the methyl-vinyl fragment. This carbon presumably enters isoprene from chloroplastic G3P that moves directly into the MEP pathway. Following the entry of 13C through the C-1 of G3P, the label appears to enter isoprene through the C-2 and C-3 of G3P, as evidence by the progressive divergence of the M41+ line from the M69+ line. We interpret the slow and incomplete labeling of the methyl-vinyl fragment to indicate the carbons derived from pyruvate carry labeled carbon into isoprene even more slowly than the carbons of G3P, and that some fraction of these carbons are perpetually derived from older, unlabeled stored carbon sources. Our observation of an inverse relationship between atmospheric CO2 growth and measurement conditions and isoprene synthesis in poplar leaves is in agreement with a number of other studies, particularly for trees 41

grown at elevated and sub-ambient CO2 concentrations (e.g., Monson & Fall 1989; Rosenstiel et al. 2003; Loreto et al. 2007; Wilkinson et al. 2009). Our results showed that isoprene emission rates observed from poplars grown under ambient and elevated CO2 were statistically equivalent.

It is likely that our experiment suffered from an inadequate sample size required to resolve differences beyond sample-to-sample variance, though the trends are consistent with what we might expect. The cause of this anti-correlation has been suggested to be the down-regulation of isoprene synthase activity when trees are grown in the presence of elevated CO2 (Scholefield et al. 2004) and/or the up-regulation of cytosolic PEP carboxylase in the case of growth at elevated

CO2 or increase in activity of PEP carboxylase in the case of short-term measurements at elevated CO2 (Rosenstiel et al. 2003; Loreto et al. 2007).

Given our conclusion that the isoprene carbon derived from older, stored reserves is channeled through the Pyr substrate, we can begin to piece together a conceptual model based on our hypothesis for how the differential control of isoprene emission by CO2 availability might occur. In our model, low CO2 availability compared to higher CO2 availability would generally result in substrate limitations to isoprene biosynthesis, particularly DMADP, because of slower rates of G3P production and reduced/negligible availability of stored starch (Fig. 2.7A and 2.7B for low- and high-CO2 availability, respectively). Furthermore, no starch is thought to be contributing to isoprene synthesis under elevated CO2 conditions as the simultaneous breakdown and synthesis of these storage carbohydrate structures remains undocumented in poplar

(Schnitzler et al. 2004). Assuming that the demand for chloroplastic pyruvate remains relatively high in the face of these substrate limitations, and that the flux of G3P through glycolysis is regulated to be nearly constant (Rontein et al. 2002), then the gap between the availability of 42 recent photosynthate/starch and substrate demands of the MEP pathway may be closed by the mobilization of extra-chloroplastic carbohydrate reserves.

Homeostatic maintenance of the glycolytic flux in the face of reduced sugar availability has been demonstrated in tomato cell cultures (Rontein et al. 2002), and in the case of our model would be required to maintain the production of pyruvate substrate from mobilized, extra- chloroplastic carbon sources. This type of regulation on the supply side of PEP production may be augmented by changes in the demand for pyruvate in the chloroplast due to up-regulation of

MEP pathway or isoprene synthase gene expression when plants are grown under sub-ambient

CO2. Up-regulation may occur if end products of the MEP pathway (e.g., carotenoids, abscisic acid (ABA), or isoprene itself) are needed to enhance tolerance of the stresses imposed by a constrained net CO2 assimilation rate. In that case, the mobilization of alternative carbon sources may be triggered by increased demand for pyruvate substrate to drive amplified MEP pathway activity, and that increased demand could be met with extra-chloroplastic or chloroplastic sources of stored carbohydrate. The gap between demand for Pyr substrate to synthesize isoprene, and what can be provided through cytosolic processing of recently assimilated photosynthate, was clearly observed in our experiments in the comparison between the leaves measured at sub-ambient CO2 and those measured at elevated CO2 (i.e., the extremes of the treatments). In that case, leaves from the sub-ambient treatment exhibited both lower fractions of total isoprene and three-carbon fragment that were labeled with 13C, and higher isoprene emission rates, compared to leaves measured at elevated CO2. The leaves measured at ambient

CO2 were not clearly distinguishable from leaves measured in the sub-ambient and elevated CO2 treatments, in a consistent manner.

Figure 2.7 Conceptual model illustrating the flow of carbon contributing to isoprene synthesis from both recently assimilated carbon in the form of glyceraldehyde-3-phosphate (G3P), chloroplastic carbon sources (starch), and extra-chloroplastic carbon (hexose phosphate) via glycolysis and the production of phosphoenolpyruvate (PEP) under sub-ambient (A) and elevated (B) CO2 growth and measurement conditions. Arrow thickness designates rates of production or transport. Low CO2 availability (panel A), compared to higher CO2 availability (panel B) result in substrate limitations to isoprene synthesis because of slower rates of G3P production and reduced/negligible availability of stored starch. Furthermore, no starch is thought to be contributing to isoprene synthesis under elevated CO2 conditions as simultaneous breakdown and synthesis is yet to be shown in poplar and the stored carbon utilized for isoprene synthesis is thought to come from extra-chloroplastic sources. RuBP = ribulose 1,5- bisphosphate; OAA = oxaloacetate; Pyr = pyruvate; TCA = tricarboxylic acid; DMADP= dimethylallyl pyrophosphate; PEPc = phosphoenolpyruvate carboxylase; PK = pyruvate kinase; IspS = isoprene synthase.

Notably, Rasulov and co-workers have explained the CO2 response of isoprene emission in terms of limitations of chloroplastic ATP, rather than the import of cytosolic PEP (Rasulov et al.

2009a,b). If the limitation to isoprene biosynthesis rate were solely due to chloroplastic ATP at elevated CO2, rather than availability of Pyr substrate, the differences in the labeling kinetics we observed between treatments simply cannot be explained. Limited ATP availability at elevated

CO2 imposed by reduced inorganic phosphate (Pi) could indeed explain reduced isoprene emission rates. However, if ATP availability was the ultimate control over the CO2-sensitivity of isoprene emission, then the proportion of 13C label in the isoprene emitted from leaves measured at sub-ambient CO2 would be similar to the isoprene emitted from leaves at elevated CO2. That is not what we observed.

CONCLUSIONS

Poplar trees grown under sub-ambient CO2 exhibited higher isoprene emission rates with a higher proportion of isoprene incompletely labeled. Across all CO2 treatments, the first carbon that contributed to isoprene synthesis appears to be derived from a rapidly labeled G3P pool, while the last two carbons come from a more slowly labeled pyruvate source. The fact that all treatments showed some level of incomplete labeling suggests that the carbon that goes into making pyruvate comes at least partly from older carbon sources within the plant. Overall, we conclude: 1) that trees experiencing low photosynthetic rates due to reduced atmospheric CO2 availability have a higher percentage of carbon from stored/older carbon sources for isoprene biosynthesis, 2) that carbon most likely enters the MEP pathway through the pyruvate substrate, and 3) that extra-chloroplastic rather than chloroplastic carbon sources are most likely mobilized to increase the availability of pyruvate required to support an up-regulation of the MEP pathway. 45

Our study shows that trees grown under conditions that limit CO2 assimilation rely more heavily on extra-chloroplastic carbon sources, most likely via the pyruvate substrate, for isoprene biosynthesis. However, the identities of these alternative carbon sources, their relative importance under long-term exposure to altered atmospheric CO2 concentrations, and the specific role of pyruvate, remain unknown. Starch is often suggested as a possible carbon source for isoprene synthesis, particularly under elevated CO2 conditions. To discriminate between starch degradation and alternative extra-chloroplastic sources in providing carbon for isoprene production, long-term CO2 studies using isotopic labels coupled with PTR-MS methodologies are needed. Starch accumulation could be quantified over time in trees grown under various CO2 regimes and the relative contribution of those labeled carbons toward isoprene synthesis would be assessed using methods similar to those described above. While starch does not appear to play a role in short-term carbon contribution towards isoprene synthesis—particularly in our study— this relationship may change under long-term exposure to altered atmospheric CO2 with subsequent changes in carbon allocation dynamics as the plants acclimate. In addition to long- term studies, future work should also consider following isotopes through other potential carbon contributors toward isoprene synthesis to identify these elusive carbon sources, although this may prove difficult considering the exchange rate of chloroplastic and extra-chloroplastic carbon compounds. Large sample sizes and considerable investment are likely required for a more comprehensive analysis of carbon allocation and isoprene biosynthesis.

CHAPTER III

HERBIVORY AND CLIMATE INTERACT SERIALLY DURING THE GROWING SEASON TO CONTROL MONOTERPENE EMISSIONS FROM PINYON PINE FORESTS

ABSTRACT

The escape of volatile monoterpenes from coniferous trees has been linked to changes in the oxidative state of the troposphere and multi-trophic signaling between plants and animals. Past laboratory studies have revealed that climate anomalies and herbivory alter the rate of tree monoterpene emissions. However, no studies to date have been conducted to test these relations in situ. We conducted a two-year field experiment at two semi-arid sites dominated by pinyon pine (Pinus edulis) during outbreaks of the specialist herbivore, the southwestern tiger moth

(Lophocampa ingens: Arctiidae). In typical years, this herbivore feeds gregariously during late spring, and can cause substantial tissue damage during outbreaks. During outbreak years, we discovered that in early spring, when herbivory rates were highest, monoterpene emission rates were approximately three to six times higher from herbivore-damaged trees compared to undamaged trees, and this increase in emissions was due to α-pinene, β-pinene, and camphene at both sites. During the growing season, as herbivory rates decreased but the mid-summer drought progressed, emission rates decreased but were not significantly different between previously damaged and undamaged trees. However, mid-summer tissue monoterpene concentrations were significantly lower in the needles of damaged trees. During the late summer, with the onset of monsoon rains, emission rates from previously damaged trees increased again, to levels higher 47 than those of undamaged trees, despite no herbivore activity. We conclude that: (1) herbivory by tiger moth larvae systemically increases the flux of terpenes to the atmosphere during the spring,

(2) drought overrides the effect of past herbivory as the primary control over emissions during the mid-summer, and (3) a release from drought and the onset of late-summer rains is correlated with a secondary increase in emissions, particularly from herbivore-damaged trees, possibly due to a drought-delayed stimulation of induced monoterpene synthesis.

INTRODUCTION

Conifers employ a number of defensive strategies against herbivorous pests, including producing oleoresin containing of a diverse array of terpenoid compounds (Bohlmann & Croteau

1999; Phillips & Croteau 1999; Trapp & Croteau 2001). Monoterpenes (C10) are the dominant secondary compound in conifer resin, and while their volatility results in a fraction of the needle terpene pool escaping constitutively, insect damage can also affect emissions through induction and/or exposure of the internal needle structures to the atmosphere (Litvak et al. 1998). Although the specific monoterpenes that are produced within different plant tissues is under genetic control, factors such as phenology (Shonwitz et al. 1990; Nerg et al. 1994; Owens et al. 1998), temperature (Tingey et al. 1980; Guenther et al. 1991; Lerdau et al. 1994; Constable et al. 1999), drought (Mattson & Haack 1987), and herbivory (Litvak et al. 1998; Byun-McKay et al. 2003;

Miller et al. 2005) can alter biosynthesis patterns and emissions to the atmosphere (Heyworth et al. 1998; Kuokkanen et al. 2001; Sallas et al. 2003; Nykanen & Koricheva 2004; Gols et al.

2007). The concentration of monoterpenes within conifer needle tissues reflects a balance between the rates of biosynthesis and emissions. Two of the most important influences on this balance are past herbivory, which can trigger biochemical reactions that induce tissues to 48 increase the rate of biosynthesis, and abiotic stresses, which can slow carbon assimilation rates and rates of terpene biosynthesis as well as increase the rate of emissions (e.g., for the case of high temperature). Past laboratory and single-factor field studies have demonstrated how biotic and abiotic controls influence monoterpene tissue concentration and emission (Litvak et al. 1998;

Staudt et al. 2002; Lavoir et al. 2009), but few field studies have investigated the interactive effects of biotic and abiotic factors on monoterpene dynamics in conifer forests.

The relative importance of environmental factors affecting defense prioritization by plants in the context of herbivory is unclear (Nybakken et al. 2008), yet may be important in determining herbivore-induced plant chemistry in natural systems. Seasonal variability has been shown to significantly affect conifer foliar monoterpene levels (Shonwitz et al. 1990; Nerg et al. 1994;

Riddle et al. 1996; Owens et al. 1998) and emission rates (Staudt et al. 1997; Staudt et al. 2000;

Holzke et al. 2006), which may impact or be altered by previous or current herbivory (Hakola et al. 2006). One specific environmental factor that has demonstrated a strong temporal correlation with the occurrence of insect outbreaks is drought. Since Mattson & Haack (1987), a mounting body of evidence has linked water stress to the susceptibility of conifers to insect herbivory due to a significantly lower investment toward resin defenses (Waring & Cobb 1992; Cobb et al.

1997). While studies have shown monoterpene emission rates to decrease dramatically under drought conditions (Llusia & Penuelas 1998; Staudt et al. 2002; Lavoir et al. 2009), foliar concentrations of monoterpenes may not reflect the composition of emissions into the atmosphere (Llusia & Penuelas 1998). However, plant responses to drought are highly variable and species-specific, and our knowledge of drought-mediated changes in plant secondary compounds and the influence of altered chemistry on herbivores is lacking (Gutbrodt et al.,

2011). Thus, understanding the abiotic and biotic controls over monoterpene production and 49 emissions can offer mechanistic explanations for herbivore population dynamics and ecosystem function, and understanding these dynamics in ecologically-sensitive ecosystems can ultimately improve management efforts.

Pinyon-juniper woodlands are the third largest vegetation type in the western United States

2005). Long-term research conducted in the pinyon-juniper ecosystem provides insight into the complex dynamics that exist between climate, pinyon fitness, and some of the pinyon’s associated insects, fungi, and microbes associated with pinyon pine (Brown et al. 2001). This study addresses how pinyon pines in the field respond both physiologically and chemically to environmental variability and herbivory. We hypothesize that the complex chemical responses of pinyon pine to variation in herbivory and drought over time lead to shifts in the primary control over monoterpene synthesis and emissions throughout the season with important implications for atmospheric processes and higher trophic level interactions.

MATERIALS AND METHODS

Study system: Pinyon pine and a specialist herbivore

Pinyon pine (Pinus edulis), are found primarily in the southern Rocky Mountains of

Colorado, Utah, Arizona, and New Mexico (Lanner et al. 1998; Powell & Opler 2009) forming extensive forests on mountain ranges up to 2100 m. The 2-4 cm rigid needles of pinyon pines are normally found in 2-leaved clusters, are marked by numerous rows of stomata, and drop irregularly between the fourth and ninth years. Pinyon pines dominate a semi-arid habitat type 50 that covers 20 million hectares, thus, the interaction of environmental stress and herbivory on pinyon pine function will have important consequences for the rest of the community. One specialist folivore that feeds on native pines is the univoltine, phytophagous Southwestern tiger moth (Lophocampa ingens: Arctiidae).

In typical years, L. ingens exhibits extensive feeding in April and May, confined to one or two branches per infested tree, and new growth can be seen later in the season (personal observation). Damage can become more extensive during outbreaks and extend to 25-100% of the tree, and in some cases lead to tree mortality (personal observation). In semi-arid ecosystems, breeding activity is synchronized with the availability of water and metabolically-active foliage; thus, the emergence of adults is dependent on local temperature and precipitation regimes

(Wagner 2009). After emergence, adult moths are in flight from late July through mid-August, during which time they will mate and search for suitable host plants on which to oviposit.

Caterpillars hatch and the larvae will begin feeding gregariously during the early fall, simultaneously building silken tents in which to overwinter. While the specific factors contributing to outbreak events are unknown, L. ingens populations are generally controlled by cold winter temperatures, pathogens, and the presence of their natural endoparasitoid enemy

(Meteorus euschausiae: Bracondidae), which parasitize early instars of L. ingens early in the fall

(Cranshaw 2006).

Field study sites

The two experimental sites were located in the pinyon-juniper woodlands of Southern

Colorado on both sides of the Continental Divide during the 2008 and 2009 growing seasons. In

2008, we sampled from eastern slope forests on Bureau of Land Management Land in Saguache 51

County, CO (37.83° N 106.30° W, elevation 2195 m), hereafter ‘Eastern Slope; in 2009 we collected samples from the western slope on private land in Montrose County, CO (38.30° N

107.83° W, elevation 2130 m), hereafter ‘Western Slope’. Both sites are characterized as semi- arid, and experience hot summers and relatively cold winters, intense solar radiation due to a high percentage of clear days, and characteristically windy conditions (Western Regional

Climate Center, Reno, NV). The Western Slope experiences more evenly distributed precipitation throughout the year, and the Eastern Slope experiences a pronounced summer drought and receives the majority of its annual precipitation during the late July-August monsoon (Fig. 3.1). Air temperature and relative humidity were measured throughout each sampling period using two Temperature/Relative Humidity Data Loggers (Onset HOBO Model-

U12, Cape Cod, MA) and cumulative precipitation and average maximum temperatures for each site were obtained from the closest NOAA weather station located to site GPS points in both counties (http://www.ncdc.noaa.gov/oa/climate/stationlocator.html).

Figure 3.1 Mean monthly temperature and the monthly sum of precipitation from long-term Colorado Climate Center meteorological observations near Saguache and Montrose Counties in Colorado. (A) The mean (black line) and variance (dashed line) of available monthly temperature from 1885-2011 at Montrose meteorological station #2 with mean monthly temperature of the study year (2009) plotted in red. (B) The mean (black line) and standard deviation (dashed line) of the sum of monthly precipitation from 1885-2011 at Montrose meteorological station #2 with the monthly sum of precipitation during study year (2009) plotted in red. (C) Same as (A) but for the 1894-2009 period at Saguache with the 2008 study year plotted in red. (D) Same as (C) but for precipitation in Saguache County.

Observational experimental design

To assess the effects of herbivore damage on pinyon pine physiology and chemistry throughout the season, we implemented a paired observational design, selecting only cone- bearing pinyon pines (~80 years old) approximately 2-2.5 m tall. Each “Damaged” pinyon pine, characterized by ~25-30% needle damage resulting from tiger moth feeding in May, were paired with an “undamaged” pinyon pine, characterized by no visual signs of defoliation by tiger moth larvae or other herbivore species. Despite potential for the next generation of tiger moths to feed on the same damaged trees in the fall, we did not observe repeated herbivory, thus, any responses measured during the end of the growing season were only due to early season herbivory and herein the phrase “previously damaged” refers only to trees damaged in May of the current year.

The paired trees were of similar age and size, experienced similar micrometeorological conditions (aspect, sunlight, wind, etc.), and were within 10 m of one another. Measurements were conducted on 14-19 May, 17-19 June, 17-20 July, 15-19 August and 22-25 September 2008 in Saguache County and on 20-22 May, 16-19 July, 17-19 August and 17-19 September 2009 in

Montrose County. No data were collected in June of 2009 from Montrose County. All samples and measurements were taken approximately 1.5 m aboveground from the distal 10-15 cm of south-facing sun-exposed branches, which included the previous year and current-year growth.

All measurements were carried out on clear sunny days from 0900-1500 h with damaged and undamaged pairs randomly chosen in both space and time for sampling within and between months. Gas exchange and volatile emissions were measured on three branches from the same trees across the study period (n=8).

Field measurements: Leaf gas exchange and monoterpene volatile emissions sampling 54

Leaf-level gas exchange was measured using a portable photosynthesis system with a conifer chamber (LI-6400, LiCor Inc., Lincoln, NE, USA) to obtain rates of net CO2 assimilation (A) and stomatal conductance (gs) rates. Repeated measures were performed on the same trees each month by placing ~4 cm of the terminal growth of each study branch in the conifer chamber and sealing around the branch with silly putty to prevent any leaks. After volatile samples were taken, branches were harvested and stored in a labeled paper bag to obtain needle dry weight and leaf area, using methods for volume displacement.

Following gas exchange measurements, a dynamic headspace branch-level enclosure technique was used to measure the flux of monoterpenes emitted from pinyon pines (see Ortega et al., 2008). Branch enclosure headspace chambers were created from cut teflon bags (1570 cm3) with one end placed over a custom made 0.635 cm (¼”) thick 25 cm diameter Teflon back plate fitted with 0.635 cm inlet and outlet bulkhead fittings. A new oven bag was used for each branch sampled to ensure no memory effects or off-gassing from previous samples. Each back plate was equipped with an 8 cm2 2 cm fan enclosed in a custom Teflon casing (Jameco

Electronics, Belmont, CA,) and a 2-channel air temperature and humidity data logger (HOBO

U12 Temperature/Relative Humidity Data Logger, Onset, Cape Cod, MA). Ambient air and needle temperature were measured just prior to sampling using the LI-6400. While the pumps were running, the end of the teflon bag not secured to the back plate was placed over the same branch sections measured for gas exchange and fastened with a plastic zip tie around a portion of the branch free of needles. Once enclosed, we allowed the chamber to equilibrate for 10 minutes at a flow rate of 1L min-1. This relatively short equilibrium time was due to the experimental limitations of battery power in a remote field site as well as unrealistically high temperatures that could be reached in the chamber at longer sampling times. Air entering the chamber through a 55 double-headed diaphragm pump was scrubbed of volatile organic compounds (VOCs) and ozone using activated charcoal (Alltech Associates Inc., Deerfield, IL) and MnO2 coated net ozone scrubbers (O.B.E. Corporation, Fredericksburg, TX), respectively.

Volatiles were sampled onto custom made glass adsorbent cartridges (7.6 cm long, 0.635 cm

OD, borosilicate glass, Allen Scientific Inc., Boulder, CO) packed with ~0.25 g Tenax® GR adsorbent (20/35 mesh, Alltech Associates Inc., Deerfield, IL) between plugs of quartz wool at a sample flow rate of 150 mL min-1 for 10 minutes. All flow rates were set and controlled using mass flow controllers (MKS Instruments Inc., Methuen, MA). Immediately following sampling, the cartridge samples were capped with Swagelok fittings, transported at 0 ºC back to the lab, and stored in a -20 ºC freezer for chemical analysis. Sample branches were immediately cut and brought back to the lab to determine total needle area following Chen et al.’s (1999) volume displacement method, assuming a semi-cylindrical needle. Briefly, a container of water (with 5% detergent added as a surfactant) was placed on a top-loading electronic balance and tared. The detached shoot was submersed in the water using tweezers, and the volume of the needles plus the twig was calculated from the mass of water displaced by the shoot. The shoot was then dried with paper towels and the needles were removed, counted, and their average length calculated using digital calipers (Fisherbrand Traceable Digital Calipers 8”, Fisher Scientific Inc.,

Pittsburgh, PA). The volume of the twig without needles was then measured, and the needle volume was calculated by difference. Needle surface area was computed from measurements of needle volume, average needle length, and hemi-cylindrical needle cross section using:

4.10 Vnl 56 where V is the volume displacement by the needles, n is the number of needles, and l is the average length of the needles. Following volume displacement measurements, dry weight was determined after drying the needles at 60 ºC for 48 hours.

Volatile chemical analysis: thermal desorption

Cartridge samples were analyzed for monoterpene emission rates within 7-21 days after sample collection by thermal desorption (Perkin-Elmer ATD400) gas chromatography-flame ionization detection/mass spectrometry (GC-FID/MS; Hewlet-Packard 5890/5970, Wilmington,

DE) following Helmig et al. (2004). The GC was fitted with a DB-1 column (0.32 mm x 60 m,

1mm film thickness, J & W Scientific, Folsom, CA) and compound separation was achieved through the following program: initial temperature of 40 ºC held for 5 minutes with a ramp of 6

ºC min-1 up to 200 ºC and held for 5 minutes at this final temperature. A helium carrier gas swept the analyte from the thermal desorber to the GC at a flow rate of 2.1 ml min-1, split between FID and MS detectors at a 4:1 split ratio. The instrument was calibrated with an n-alkane reference

-1 standard (NOAA, Boulder, CO) to determine a response factor (RF, peak area mL sample volume ppbC-1) used to calculate monoterpene concentrations. Individual compounds were identified by using authentic standards where available as well as the comparison of GC retention index (RI) and mass spectra scans to those reported by Adams (1989). Chromatograms were integrated using PeakSimple software (SRI Instruments, Menlo Park, CA). Emission rates were calculated using:

ER = [(Cout – Cin) × Q ] / mdry

-1 where Cout and Cin are the outlet and inlet concentrations of the compound of interest (µg C L ),

-1 Q is the flow rate of the bag purge air (1000 ml min ), and mdry is the dried mass (g) of enclosed 57

needles. Cin was set to zero because it consisted of VOC-free air. To determine the concentration of the compounds of interest we used the following equation:

PeakArea MWC C out RF ECN

Where MWC is the molecular weight of monoterpenes and ECN is the effective carbon number for each monoterpene, used to compensate for a compounds molecular bonding structure and relative FID response. This value was then corrected for enclosure temperature and pressure, sample volume and needle dry weight. Final emission rates are reported in units of µg C gram dry weight (gdw)-1 hour-1.

Foliar samples and monoterpene chemical analysis

Needles were removed from the sample branches after all volatile and physiological measurements were taken and immediately flash frozen, stored in liquid nitrogen, transported to lab, and placed in a -80 ºC freezer. We used the monoterpene extraction method of Keefover-

Ring and Linhart (2010) and, through method development, modified it for processing conifer needles. Needles were ground in liquid nitrogen using a chilled mortar and pestle to minimize monoterpene loss. Frozen needle powder was then weighed into 2 dram glass vials, exact weights (between 0.4 and 0.6 g) were recorded, and 2 mL of gas chromatograph (GC)-grade n- hexane (Fisher Scientific, Pittsburgh, USA) containing 0.1 µL mL-1 (+)-fenchone (Sigma-

Aldrich, St. Louis, USA) as an internal standard was added to the vial. Vials were immediately closed with teflon lined caps, mixed with a vortexer, and allowed to soak for one week at ambient temperature. After the seven-day soaking period, 100 µL aliquots from each sample were taken for chemical analysis. 58

One µL of each sample was then injected on an Agilent Technologies 6890 gas chromatograph-5975 mass spectrometer (GC-MS) fitted with a Cyclodex-B chiral column (30 m

250 µm 0.25 µm; J&W Scientific). Helium was used as the carrier gas at a flow rate of 1.0 mL min-1 with a split flow ratio of 25:1. Injector temperature was set at 230 ºC. The oven profile consisted of an initial temperature of 60 ºC followed by a ramp of 5 ºC min-1 to 200 ºC, then a second ramp of 25 ºC min-1 to 230 ºC. Monoterpene enantiomers were identified by comparing retention times of known standards and mass spectra using the NIST library and MSD

Chemstation software. Concentrations for each compound were calculated using 4-point calibration curves with injections of known amounts of pure standards and the internal standard, fenchone. All standards were purchased from Sigma Aldrich (Saint Louis, MO) with the exception of β-phellandrene (Glidco Organics, Jacksonville, FL).

Statistical analyses

All statistical analyses were performed using SAS statistical software (SAS Institute, Cary,

NC, 2010) and data from each field site were assessed separately. Gas exchange data from both sites were normally distributed and analyzed using a mixed model, repeated measures ANOVA

(proc mixed) with a Bonferroni correction. Total and individual monoterpene emission data from both sites were log10 transformed while foliar concentrations met the assumptions of normality.

Both emission and concentration data were analyzed using repeated measures ANOVA, as described above. Relative humidity was included as a covariate when analyzing the emission data.

RESULTS 59

We first present the results the leaf gas exchange, monoterpene volatile emissions, and monoterpene foliar concentration analyses for the Western Slope followed by those performed on the Eastern Slope.

Western Slope: Leaf gas exchange and monoterpene volatile emissions

Pinus edulis net CO2 assimilation (A) and stomatal conductance (gs) rates on the Western

Slope indicated seasonal physiological activity (Fig. 3.2A and 3.2C; A: F3, 100=36.32, gs: F3,

100=30.53, P < 0.0001), with trees exhibiting the highest photosynthesis rates during both May and September, which had relatively cool temperatures; whereas the lowest rates were observed during the middle of the summer (July and August) when temperatures exceeded 30 ºC.

Herbivory had no significant effects on pinyon pine photosynthesis rates or stomatal conductance on the Western Slope (Fig. 3.2A and 3.2C; A: F1, 18=1.62, gs: F1, 18=0.80, P >0.05), and there was no interactive effect of time and herbivory on photosynthesis or stomatal conductance (A: F3,

100=1.28, P = 0.28; gs: F3, 100=2.03, P = 0.12).

Total monoterpene emission rates of pinyon pine on the Western Slope were significantly affected by time (Fig. 3.3A; F3, 114=34.22, P<0.0001), likely driven by the significant increase in constitutive emissions during summer and decrease in the fall. While herbivory alone did not have a significant effect on emission rates (F1, 15=1.22, P=0.29), this was likely due to similar emissions between damaged and undamaged trees throughout most of the growing season, there was an interactive effect of herbivory and time (F3, 114=7.31, P<0.0005). Concurrent with herbivore feeding during May, damaged pinyon pines exhibited total monoterpene emission rates

~3 times higher than their undamaged counterparts (damaged: 1.05 ± 0.10; undamaged: 0.34 ±

0.11 µg C gdw-1 h-1, P<0.0005). While emission rates from damaged trees did not change 60 significantly into July (1.06 ± 0.22 µg C gdw-1 h-1), undamaged trees exhibited significantly higher total emission rates (1.14 ± 0.23 µg C gdw-1 h-1, P<0.005) under the high temperatures and relatively high levels of precipitation experienced by damaged trees. These high emission rates were then followed by a gradual decrease in total monoterpene emissions from both damaged and undamaged pinyon pines, with the lowest emission rates observed in September

(damaged: 0.14 ± 0.05 µg C gdw-1 h-1; undamaged: 0.17 ± 0.05 µg C gdw-1 h-1).

Figure 3.2 Seasonal change in net CO2 assimilation rates (A) stomatal conductance rates (gc) of undamaged and L. ingens damaged P. edulis at sites on the Western Slope (A and C) and Eastern Slope (B and C) of the Rocky Mountains in Southern Colorado. Herbivore feeding, pupation, and mating/oviposition periods are noted at across the top, and while herbivory does occur in the fall, we did not observe repeated herbivory during this time in our study. Precipitation values represent cumulative precipitation two weeks prior to sampling and temperature values represent the mean maximum daily temperature at each site during sampling. Values are means ± standard error of the mean (SEM) (n=8). No physiology data (ND) were taken during June from the Western Slope site. 62

Figure 3.3 Seasonal change in total monoterpene emission rates and foliar concentrations of undamaged and L. ingens damaged P. edulis at sites on the Western Slope (A and C) and Eastern Slope (B and C) of the Rocky Mountains in Southern Colorado. Herbivore feeding, pupation, and mating/oviposition periods are noted at across the top, and while herbivory does occur in the fall, we did not observe repeated herbivory during this time in our study. Total monoterpene emissions represent the sum of emissions from α-pinene, β-pinene, β-myrcene, β-phellandrene, camphene, δ-carene, and limonene. Total foliar concentrations represent the sum of monoterpene enantiomers (+)-α-pinene, (-)-α-pinene, β-pinene, β-myrcene, β-phellandrene, camphene, and (-)- limonene. Precipitation and temperature values follow Fig. 3.1. Lophocampa. ingens larvae fed primarily during May, pupated in June, mated in-flight and oviposited from July-August. ** Indicates a significant difference between treatments (P<0.05) and values are means ± SEM (n=8). No emission samples (ND) were taken during June from the Western Slope site.

To explore the impacts of herbivory and time on individual monoterpene emissions, the

relative concentrations of seven individual monoterpenes were compared between undamaged

and herbivore-damaged pinyon pines for each month throughout the season (Table 3.1). Early in

the season, during caterpillar feeding, damaged trees on the Western Slope emitted significantly

higher levels of α-pinene, β-pinene, β-phellandrene, and camphene. Damaged trees on the

Eastern Slope had elevated emission rates in α-pinene, β-pinene, β-myrcene, camphene, and

limonene. In July, when adults are potentially searching for oviposition sites, trees on the

Western Slope that were previously damaged early in the season emitted higher levels of β-

phellandrene (damaged: 0.06 ± 0.01, undamaged: 0.02 ± 0.01, P<0.05), although no difference

Fig. 3. could be detected in the total monoterpene emission rates between treatments during mid- Emission rates of summer (Fig. 3.3A). Previously damaged trees on the Eastern Slope emitted higher levels of α- individual monoterpene pinene (damaged: 1.04 ± 0.17, undamaged: 0.50 ± 0.18, P<0.05) and δ-carene (damaged: 0.15 ± compounds from 0.02, undamaged: 0.05 ± 0.02, P<0.05) during the flight of adult moths in August. undamaged and damaged pinyon pines during adult tiger moth flight (August) from the eastern (a) and western (b) slopes. * Indicates a significant difference between treatments (P<0.05). Error bars represent SEM.

Table 3.1 Mean monoterpene emission rates [µg C gdw-1 hr-1] (±SEM) from undamaged (left) and herbivore-damaged (right) P. edulis trees throughout the 2009 growing season on the Western Slope (top) and 2008 growing season on the Eastern Slope (bottom) of the Rocky Mountains in Southern Colorado. Letters represent significant differences between treatments (P<0.05, repeated measures ANOVA). ND = No Data.

May June July August September

0.21±0.07a 0.63±0.07b ND 0.73±0.16a 0.71±0.16a 0.53±0.19a 0.49±0.19a 0.10±0.03a 0.10±0.03a α-Pinene 0.17±0.21a 1.20±0.21b 0.28±0.07a 0.31±0.07a 0.18±0.04a 0.12±0.04a 0.50±0.18a 1.04±0.17b 0.27±0.09a 0.51±0.09a

0.02±0.02a 0.14±0.02b ND 0.06±0.02a 0.06±0.01a 0.04±0.02a 0.05±0.02a 0.01±0.004a 0.01±0.004a β-Pinene 0.10±0.09a 0.49±0.09b 0.06±0.01a 0.05±0.01a 0.03±0.01a 0.03±0.01a 0.06±0.02a 0.08±0.02a 0.05±0.01a 0.05±0.01a

0.04±0.01a 0.06±0.01a ND 0.16±0.04a 0.13±0.04a 0.08±0.11a 0.21±0.11a 0.02±0.01a 0.02±0.01a β-myrcene 0.02± 0.05a 0.32±0.05b 0.06±0.02a 0.06±0.02a 0.04±0.01a 0.02±0.01a 0.06±0.02a 0.08±0.02a 0.05±0.01a 0.04±0.01a

0.02±0.01a 0.08±0.01b ND 0.06±0.01a 0.02±0.01b 0.03±0.01a 0.03±0.01a 0.02±0.004a 0.01±0.004a β-phellandrene 0.04±0.07a 0.24±0.07a 0.04±0.01a 0.02±0.01a 0.02±0.01a 0.02±0.01a 0.04±0.01a 0.04±0.01a 0.04±0.01a 0.03±0.01a

0.01±0.01a 0.05±0.01b ND 0.03±0.01a 0.03±0.01a 0.02±0.01a 0.03±0.01a 0.01±0.001a 0.01±0.001a camphene 0.06±0.06a 0.26±0.06b 0.02±0.007a 0.02±0.007a 0.01±0.003a 0.008±0.003a 0.02±0.006a 0.03±0.006a 0.02±0.003a 0.02±0.003a

0.03±0.02a 0.05±0.02a ND 0.07±0.02a 0.06±0.02a 0.04±0.02a 0.05±0.02a 0.01±0.01a 0.01±0.01a Δ-carene 0.002±0.03a 0.09±0.03a 0.07±0.04a 0.05±0.04a 0.02±0.01a 0.01±0.01a 0.05±0.02a 0.15±0.02b 0.05±0.02a 0.09±0.02a

0.01±0.003a 0.02±0.002a ND 0.03±0.01a 0.05±0.01b 0.02±0.01a 0.03±0.01a 0.005±0.001a 0.005±0.001a limonene 0.02±0.02a 0.14±0.02b 0.03±0.02a 0.07±0.02a 0.008±0.002a 0.008±0.002 0.03±0.01a 0.05±0.009a 0.01±0.002a 0.02±0.002a

Western Slope: Monoterpene foliar concentrations

Total monoterpene foliar concentration of pinyon pines on the Western Slope were significantly affected by time (F3, 102=26.71, P<0.0001) but not herbivory (F1, 14=0.04, P=0.84), and there was a significant interaction between herbivory and time (F3, 102=18.38, P<0.0001 (Fig.

3.3C). While L. ingens feeding in May resulted in significantly higher total monoterpene emissions (Fig. 3.3A), there was no difference in foliar concentrations between herbivore- damaged and undamaged pinyon pines (Fig. 3.3C). Total monoterpene concentrations in undamaged pinyon pines during July (7.34 ± 0.74 mg gfw-1) were unchanged from those measured in May (8.17 ± 0.73 mg gfw-1; Fig. 3.3C), yet pinyon pines that had been previously damaged and exhibited relatively high monoterpene emission rates in May experienced ~40% decrease in total foliar monoterpene concentration over the same time period (May: 8.23 ± 0.73 mg gfw-1, July: 4.94 ± 0.75 mg gfw-1, P<0.0001), which were significantly lower concentrations than observed in undamaged pinyon pines (P<0.05). Foliar concentrations of undamaged pinyon pines decreased in August (5.08 ± 0.77 mg gfw-1, P<0.0001) with a significant increase to mid- summer levels in September (6.8 ± 0.74 mg gfw-1, P<0.001), while previously damaged trees experienced a steady increase in foliar monoterpenes into the fall (August: 6.57 ± 0.75 mg gfw-1,

September: 7.4 ± 0.77 mg gfw-1, P<0.01) with a concurrent decrease in monoterpene emissions

(Fig. 3.3A).

All seven quantified monoterepene enantiomers changed significantly over time on the

Western Slope (F3, 102 = 8.15, P<0.0001) and were affected by the interaction of herbivory and time (F3, 102 = 4.61, P<0.0001), with the exception of camphene, which only changed over the season (Table 3.2, F3, 102 = 12.44, P<0.0001). While δ-carene was detected in some of the samples, it exhibited very similar elution times with β-pinene, often causing the less abundant, 66 almost negligible, δ-carene (RT 8.54) to elute as a small shoulder to the much larger β-pinene peak (RT 8.41) and become indistinguishable from the baseline. Therefore, it was only possible to accurately integrate the β-pinene peak and no values for δ-carene are subsequently reported.

There was no difference in individual compound concentrations between damaged and undamaged pinyon pines throughout the growing season except for in July, where previously damaged trees had significantly lower levels of β-myrcene (damaged: 0.48 ±0.11 mg gfw-1, undamaged: 0.90 ±0.11 mg gfw-1, P<0.05) and β-phellandrene (damaged: 0.07 ±0.02 mg gfw-1, undamaged: 0.13 ±0.02 mg gfw-1, P<0.05).

Eastern Slope: Leaf gas exchange and monoterpene volatile emissions

On the Eastern Slope time significantly affected A and gs in pinyon pines (Fig. 3.2B and Fig.

3.2D; A: F4, 162=19.23, gs: F4, 162=18.72, P <0.0001), although the seasonal pattern was different than that exhibited on the Western Slope. Specifically, photosynthesis rates in undamaged pinyon pines decreased from May to July, concurrent with consistently high temperatures and minimal precipitation, while the highest rates were observed in August after the onset of the late summer monsoons (Fig. 3.2B). This sharp increase was then followed by a decrease in A such that rates in May and September were similar (Fig. 3.2B). Rather than decreasing from spring to summer, photosynthesis rates in herbivore-damaged pinyon pines remained at the same level until the late summer monsoons, at which point they followed the same pattern as undamaged pinyon pines. Eastern Slope pinyon pine photosynthesis and stomatal conductance rates were also not significantly affected by herbivore damage (Fig. 3.2B and 3.2D, A: F1, 16=1.32, P =

0.27; gs: F1, 16=3.05, P = 0.10), but there was a significant interactive effect of time and herbivory on physiology (A: F4,162=2.99, P < 0.05; gs: F4,162=2.56, P <0.05). 67

Total monoterpene emission rates from Eastern Slope pinyon pines were significantly affected by herbivory (Fig. 3.3B, F1, 18=10.76, P<0.005), time (F4, 150=14.78, P<0.0001), and the interaction between these two main effects (F4, 150=4.29, P<0.005). In May, trees that were experiencing herbivore damage emitted monoterpenes at a rate ~6 times greater than undamaged pinyon pines (damaged: 2.74 ± 0.45 µg C gdw-1 h-1; undamaged: 0.43 ± 0.43 µg C gdw-1 h-1,

P<0.05, Fig. 3.3B). From June through July, there was no difference in total monoterpene emission rates between damaged and undamaged trees, with the lowest emissions observed during the mid-summer drought in July (damaged: 0.26 ± 0.07 µg C gdw-1 h-1; undamaged: 0.35

± 0.07 µg C gdw-1 h-1). With the onset of the late summer monsoons in August and a release from drought stress, Eastern Slope pinyon pines emitted higher levels of total monoterpenes

(P<0.0001), and trees that had been damaged earlier in the season had significantly higher emission rates than undamaged pinyon pines (damaged: 1.53 ± 0.22 µg C gdw-1 h-1; undamaged:

0.82 ± 0.24 µg C gdw-1 h-1, P<0.05). This burst in monoterpene emissions was followed by a decrease in September, and monoterpene emissions between damaged and undamaged pinions were not significantly different (damaged: 0.76 ± 0.13 µg C gdw-1 h-1, undamaged: 0.57 ± 0.13

µg C gdw-1 h-1, P=0.10).

Eastern Slope: Monoterpene foliar concentrations

Pinyon pines on the Eastern Slope exhibited changes in monoterpene concentrations over time (F4, 66=11.54, P<0.0001) yet did not experience significantly altered monoterpene foliar content as a result of herbivory (F1, 14=2.66, P=0.13) or the interaction between herbivory and time (Fig. 3.3D, F4, 66=1.61, P=0.18). Total monoterpene concentrations in undamaged pinyon pines were highest in July, and were significantly greater than previously damaged pinyon pines 68

(damaged: 3.28 ± 0.67 mg gfw-1; undamaged: 5.04 ± 0.68 mg gfw-1, P<0.05). A spike in monoterpene emissions in both damaged and undamaged pinyon pines with the onset of August monsoon rains (Fig. 3.3B), resulted in a significant decrease in monoterpene concentrations in both sets of trees (Fig. 3.3D, damaged: 2.34 ± 0.67 mg gfw-1; undamaged: 3.73 ± 0.66 mg gfw-1,

P<0.05). Undamaged pinyon pines maintained similar total levels of foliar monoterpene concentrations into September (4.03 ± 0.65 mg gfw-1), but foliar monoterpene concentrations in previously damaged trees increased significantly (4.05 ± 0.67 mg gfw-1, P<0.0005) reaching levels observed in undamaged trees.

There was no difference in foliar concentrations in Eastern Slope pinyon pines of either enantiomer of α-pinene, β-pinene, β-phellandrene, or camphene between undamaged and damaged trees throughout the season; and only α-pinene and β-phellandrene were significantly affected by time (Table 3.2, F4, 66 = 13.52, F4, 66 = 9.91, respectively P<0.0001). Concentrations of β-myrcene changed both over the season (F1, 15 = 14.26, P<0.005) and in response to herbivory (F4, 66 = 15.19, P<0.0001). Herbivore-damaged pinyon pines sustained significantly lower levels of β-myrcene during and after caterpillar feeding, from late spring to early fall

(P<0.05). While this difference was consistent throughout the year, it only translated into changes in total monoterpene concentration during mid-summer. Levels of (-)-limonene in damaged pinyon pines (0.06 ± 0.02 mg gfw-1) were significantly lower than those of undamaged trees (0.13 ± 0.02 mg gfw-1) during caterpillar feeding in May (P<0.05).

Table 3.2 Mean foliar monoterpene concentrations of individual compounds [mg gfw-1] (±SEM) from undamaged (left) and herbivore-damaged (right) P. edulis trees throughout the 2009 growing season on the Western Slope (top) and 2008 growing season on the Eastern Slope (bottom) of the Rocky Mountains in Southern Colorado. Different letters represent a significant difference between undamaged and herbivore-damaged trees (P<0.05, repeated measures ANOVA).

May June July August September

1.47±0.28a 1.79±0.28a ND 1.60±0.28a 1.30±0.28a 1.29±0.29a 1.66±0.28a 1.50±0.28a 1.62±0.28a (-)-α-Pinene 0.78±0.10a 0.55±0.10a 0.95±0.11a 0.77±0.11a 0.78±0.10a 0.55±0.10a 0.81±0.11a 0.53±0.11a 1.02±0.11a 0.81±0.11a

3.97±0.62a 3.42±0.61a ND 3.26±0.62a 2.10±0.63a 2.17±0.64a 2.95±0.63a 2.80±0.62a 3.32±0.63a (+)-α-Pinene 1.31±0.26a 0.90±0.26a 1.59±0.28a 1.42±0.29a 1.78±0.29a 1.30±0.29a 1.28±0.29a 0.74±0.29a 1.35±0.28a 1.42±0.29a

1.35±0.20a 1.89±0.20a ND 1.27±0.21a 0.84±0.21a 0.69±0.21a 1.02±0.21a 1.18±0.21a 1.62±0.21a β-Pinene 1.05±0.25a 0.80±0.25a 1.35±0.28a 0.90±0.29a 1.24±0.30a 0.64±0.29a 0.89±0.29a 0.63±0.29a 0.81±0.28a 1.19±0.29a

1.00±0.11a 0.77±0.11a ND 0.90±0.11a 0.48±0.11b 0.70±0.11a 0.67±0.11a 0.95±0.11a 0.59±0.11a β-myrcene 0.57±0.06a 0.34±0.05b 0.67±0.06a 0.44±0.06b 0.74±0.06a 0.43±0.06b 0.55±0.06a 0.30±0.06b 0.68±0.06a 0.45±0.06b

0.14±0.02a 0.13±0.02a ND 0.13±0.02a 0.07±0.02b 0.08±0.02a 0.08±0.02a 0.14±0.02a 0.10±0.02a β-phellandrene 0.08±0.01a 0.05±0.01a 0.09±0.01a 0.07±0.01a 0.11±0.01a 0.07±0.01a 0.08±0.01a 0.04±0.01a 0.09±0.01a 0.06±0.01a

0.07±0.01a 0.07±0.01a ND 0.07±0.01a 0.05±0.01a 0.06±0.01a 0.06±0.01a 0.07±0.01a 0.07±0.01a camphene 0.07±0.01a 0.05±0.01a 0.07±0.01a 0.06±0.01a 0.07±0.01a 0.05±0.01a 0.06±0.01a 0.04±0.10a 0.06±0.10a 0.04±0.10a

0.17±0.02a 0.18±0.02a ND 0.15±0.02a 0.12±0.02a 0.10±0.02a 0.14±0.02a 0.17±0.02a 0.12±0.02a (-)-limonene 0.13±0.02a 0.06±0.02b 0.07±0.02a 0.09±0.03a 0.10±0.03a 0.08±0.03a 0.07±0.03a 0.05±0.03a 0.06±0.03a 0.08±0.03a

DISCUSSION

While a number of studies have explored the effects of abiotic stress on pinyon pine physiology and architecture in the context of herbivore-susceptibility (Cobb et al. 1997; Brown et al. 2001), our study is the first to consider the interactive effects of natural environmental variation and folivory on monoterpene concentrations and emissions from pinyon pines in situ.

Our results offer a clearer understanding of the trade-offs that occur between primary controls over plant chemistry in a species adapted to semi-arid ecosystems. The results of this study provide us three major conclusions: (1) early spring herbivory by tiger moth larvae increases the flux of monoterpenes to the atmosphere, (2) drought overrides the effect of past herbivory as the primary control over emissions during the mid-summer, and (3) that the end of that drought by late-summer rains is correlated with a secondary increase in emissions, particularly in herbivore- damaged trees, possibly due to a drought-delayed stimulation of induced monoterpene synthesis.

We observed sharp increases in total monoterpene emission rates during periods of herbivore feeding at both sites. Because we measured undamaged branches that shared vascular connections to damaged branches, we assume these responses to be systemic and that they cannot be explained via decreased diffusive resistance resulting from mechanical exposure of oleoresin to the atmosphere (Litvak et al. 1999; Loreto et al. 2000). Total monoterpene emission rates from Eastern Slope pinyon pines increased ~6 fold compared to Western Slope pinyon emission rates, which were only ~3 times higher that their undamaged counterparts (Fig. 3.3A and B), despite the fact that Western Slope pinyon pines exhibited monoterpene foliar concentrations ~2 times greater than Eastern Slope pinyon pines (Fig. 3.3C and 3.3D). These findings are consistent with studies showing that Pinus species with high constitutive levels of monoterpene cyclase activity exhibita lower wound-induced response (Lewinsohn & Croteau, 71

1991). It is difficult to assess whether emissions were due to de novo production of monoterpenes or if compounds were transported from other plant tissues (Johnson & Croteau

1987). However, the significant increases in herbivore-induced emission rates at both sites occurred despite similar foliar monoterpene concentrations between undamaged and herbivore- damaged pinyon pines (Fig 3.3C and 3.3D), similar to P. ponderosa that exhibited an increase in monoterpene synthase activity resulting in high emission rates but no net effect on foliar monoterpene pool size (Litvak et al. 1998).

Total monoterpene emission rates from pinyon pines on the Western Slope followed the site’s characteristic temperature and precipitation dynamics (Fig. 3.3A), while a pronounced mid-summer drought resulted in the low emission rates from both undamaged and previously damaged pinyon pines on the Eastern Slope (Fig. 3.3B). My data suggest that high herbivore- induced emissions early in the season may explain the significantly lower total monoterpene foliar concentrations observed in previously damaged Eastern Slope pinyon pines during mid- summer (Fig. 3.3D). Despite potential herbivore-induced enzymatic up-regulation earlier in the season, this effect may have been event-specific and not reflected throughout the remainder of the growing season. Instead, low foliar concentrations in previously damaged pinyon pines during mid-summer reflects the early season losses, and while monoterpene synthesis may be upregulated to keep pace with previous evaporative losses (Constable et al. 1999), more carbon may have been allocated toward replacing lost leaf tissue as opposed to an investment in defense

(Bryant et al. 1983). Furthermore, these seasonally low emission rates occurred when temperatures were high, despite the well-established short-term exponential relationship that exists between temperature and monoterpene emissions (Tingey et al. 1980; Guenther et al. 72

1993), demonstrating the important role of seasonal water availability in controlling monoterpene emissions.

My results are relevant to the development of vegetation emission models, which have become an important component of regional-to-global atmospheric chemistry modeling

(Fehsenfeld et al. 1992; Guenther et al. 2000; Liao et al. 2007; Peñulas & Staudt 2010). While some atmospheric models consider the influence of conifer herbivory on tropospheric chemistry, algorithms are based on short-term effects and simulated herbivory. My data suggest that herbivory will episodically increase the total monoterpene emission rates from pinyon, likely leading to an enhanced oxidative capacity of the local troposphere (Litvak et al. 1999).

Furthermore, these higher emission rates are driven by specific compounds (Table 3.1), which

- should all react in a similar pathway given the same conditions (OH , NOx, etc.), but have different reactivity rates with various atmospheric constituents, resulting in diverse secondary products with important feedbacks on regional chemistry. Thus, my results highlight the importance of taking into account longer-term seasonal dynamics and field condition herbivory when developing future flux models.

A number of studies have stressed the importance of water availability on monoterpene emissions (Kesselmeir & Staudt 1998; Bertin & Staudt 1996; Staudt et al. 2000), which may be especially important on the Eastern Slope of the Southern Rocky Mountain where the dominant precipitation events occur during late-summer when evaporative demand is high. Monoterpene emissions not only increased significantly with the onset of Eastern Slope monsoon rains, but previously damaged pinyon pines had significantly higher emission rates compared to undamaged pinyon pines, suggesting an interactive effect between previous herbivory and release from drought stress, as this cannot be explained by higher average temperature or foliar 73 concentrations as both decreased in August. This difference in emissions between previously damaged and undamaged pinyons also cannot be explained by changes in carbon metabolism alone, as both net CO2 assimilation (A) and stomatal conductance rates (gs) significantly increased from July to August in both damaged and undamaged pinyons (Fig. 3.2B and 3.2D).

This lack of relationship is consistent with studies showing that: 1) the relatively large storage pool in the resin ducts of conifers does not strongly depend on recent photoassimlates to support monoterpene biosynthesis (Lerdau & Gray 2003), and 2) a lack of control by stomatal conductance in driving monoterpene emissions (Lerdau et al. 1994, 1997), although monthly- averaged conductance follows similar trends with monthly-averaged monoterpene emissions here. Thus, water relations and stomatal conductance may help explain this emission pattern, but donot fully describe the mechanism underlying the significantly higher emissions observed from previously damaged trees, suggesting a more biochemical explanation through a potentially delayed-induced response.

Experimental results suggest an additional interplay between herbivory and ecohydrology; herbivore-damaged pinyon pines on the Eastern Slope did not demonstrate a significant decline in photosynthesis or stomatal conductance from May to July (Nabity et al. 2009), suggesting that the moderate loss of leaf area from herbivory may have conferred whole-tree water savings

(Pataki et al. 1998) and resulted in improved plant function relative to undamaged pinyon pines.

Under drought conditions, this increase in water savings may have enabled a stronger metabolic response with release from stress, including increased emissions later in the season as a result of delayed herbivore-induction. This mechanism has been proposed in various species over different time scales (Tuomi et al. 1990; Haukioja et al. 1991; Ruohomaki et al., 1996) and is supported by the observed increase in monoterpene emissions in the absence of herbivory or 74 changes in foliar monoterpene concentration later in the season. Higher emission rates emitted from previously damaged pinyon pines may then be a combination of drought-delayed stimulation of induced monoterpene synthesis and improved water relations. In other words, pinyon pines are potentially “conditioned” to express defenses later in the season, perhaps through the up-regulation of biosynthetic processes during stress that are expressed upon release from drought (Niinemets 2009; Walters 2009).

Understanding how monoterpene emissions and concentrations change throughout the growing season and in the context of herbivore life cycles will offer insight into top-down and bottom-up herbivore population controls mediated by plant chemistry in this ecosystem. The interactive effect of herbivory with seasonal temperature/precipitation events on the Eastern

Slope led to significant increases in α-pinene and δ-carene during August, which may play a role in conveying host quality information to L. ingens moths searching for suitable oviposition sites during that time. However, these compounds could also function as pinyon defense compounds if exploited as host location cues by the parasitoids that attack L. ingens, which prefer to parasitize early larval instars (personal observation). The functional role of these individual volatiles has yet to be determined in this system. Many studies have shown herbivory to result in increased plant resistance, yet this is not always the case (Williams and Myers 1984). Contrary to most studies investigating the effects of herbivory on secondary compounds, we observed significantly higher levels of β-myrcene and (-)-limonene in undamaged pinyon pines during herbivore feeding on the Eastern Slope. If these compounds are toxic to L. ingens, the increased concentrations present in undamaged pinyon pines may lead to greater herbivory on damaged trees, thus increasing herbivore performance and densities (i.e. outbreaks), and in turn, pinyon susceptibility to mortality. Unfortunately, little is known about how monoterpene structure 75 affects toxicity and immunocompetency in insects and the role of monoterpenes as conifer defense compounds against folivores remains highly debated due to differential species-specific effects (Mumm & Hilker, 2006).

This study provides a new perspective that contrasts with the traditional framework within which studies investigate various controls over monoterpene production. My results suggest that abiotic and biotic controls over pinyon monoterpene emissions, and thus foliar concentrations, show notable interactive effects and trade-offs in relative importance throughout the season. Data suggest a strong influence of water availability and herbivory on pinyon chemistry, which may offer a more mechanistic view of how insects exacerbate drought-induced pinyon mortality through delayed induced plant chemical responses. In addition to offering insight into the complex feedbacks that exist between multiple drivers in altering foliar and volatile monoterpenes during outbreaks, our results have potentially important consequences for atmospheric processes and higher trophic level interactions in pinyon-juniper ecosystems. Future studies examining the effects of herbivore-induced pinyon chemistry on insect preference and performance will aid in our explanation of herbivore population outbreaks in this system.

CHAPTER IV

EFFECTS OF HERBIVORY ON PINYON PINE MONOTERPENE CHEMISTRY: CONSEQUENCES FOR FUTURE HERBIVORY AND PARASITISM

ABSTRACT

Insect herbivory has been weakening pinyon pine trees across the Southwestern U.S., and enhancing climate-induced mortality. To examine the effects of pinyon pine chemistry on the growth and immune response of the specialist Southwestern tiger moth (Lophocampa ingens), we created artificial diets mimicking constitutive and herbivore-induced monoterpene concentrations observed in situ. While the combination of all monoterpenes, regardless of previous herbivory, results in a trade-off between investment in melanization

(and thereby defense against parasitism) and growth, herbivore-damaged pinyon pines with lower overall tissue monoterpene concentrations, particularly β-myrcene and (-)-limonene, encouraged further herbivory with no increase in growth but enhanced immunity to parasitoid infection. Thus, as an outbreak progresses, herbivores exhibit higher consumption rates and enhanced immunity against parasitoid infection. Such responses are likely to have important implications for forest-pest dynamics.

INTRODUCTION

One of the primary defense strategies employed by conifers against pathogenic pests and herbivores is the presence of a diverse array of oleoresin monoterpenes (Bohlmann & 77

Croteau 1999; Phillips & Croteau 1999; Trapp & Croteau 2001). Although monoterpenes

(C10) have been shown to function as oviposition stimulants to specialist herbivores, they also serve as feeding deterrents to generalists (Leather 1987; Binder & Robbins 1997;

Hummelbrunner & Isman 2001; Shu et al. 2004; Grant et al. 2007; Abdelgaleil et al. 2009).

The inhibited growth that is evident in many herbivores that ingest monoterpenes is presumably due to the energy costs required to detoxify the compounds (Bowers & Puttick

1988; Appel & Martin 1992). However, some species, such as sawflies, exhibit no change in behavior, function or growth rate when consuming tissues with high monoterpene concentration, and instead exploit them for their own defense (Codella & Raffa 1995;

Pasquier-Barr et al. 2001). The variable responses by herbivores to foliar monoterpenes have resulted in a lack of consensus concerning the role of these compounds in plant defense,

(Linhart 1991, Mumm & Hilker 2006). This may help explain why the foliage of some tree species exhibit induced resistance following extensive herbivore feeding (Trewhella et al.

1997; Smits & Larsson 1999; Hodar et al. 2004), while others do not (Niemela et al. 1991;

Lyytikainen 1992; Clancy et al. 2004).

Constitutive and induced plant secondary chemistry can impact herbivores from the bottom-up, through their effects on herbivore preference, growth, survival, and fecundity

(Cates 1980; Havill & Raffa 2000; Awmack & Leather 2002), but also from the top-down, by facilitating host location by predators and parasitoids (Ode 2006). Suitable host nutritive value and/or evasion of parasitism can lead to increases in herbivore densities, which have consequences for plant community composition. Although the direct and indirect effects of induced defense chemistry on herbivore performance have been investigated separately, little 78 is known about cumulative effects via simultaneous changes in herbivore growth and development and parasitoid performance.

In response to herbivore feeding, a number of tree species have been shown to up- regulate the synthesis of secondary defense compounds (Karban and Myers 1989), the presence of which can mitigate future herbivory. While not all induced plant responses increase resistance through changes in defense chemistry (e.g. physiological or structural changes), insects that specialize on conifers have adapted mechanisms to effectively detoxify relatively high concentrations of foliar secondary compounds induced by initial feeding

(Litvak et al. 1998). However, the energy required to metabolize secondary compounds incurs a cost resulting in less available energy available for other biological functions including reproduction (Bernays and Chapman 1994; Shoonhooven et al. 2005), growth

(Awmack & Leather 2002), and immunocompetency (Muller et al. 2003; Kraaijeveld et al.

2001). The physiological cost of detoxification affects the insect innate immune response— encapsulation (building cell layers around a foreign object) and melanization (production of cytotoxic compounds)—toward endoparasitoid eggs and increases an herbivore’s vulnerability to parasitism (Schmid-Hempel & Ebert 2003; Smilanich et al. 2009). Because parasitoids have been repeatedly cited as the primary controls over fluctuations in herbivore populations (Berryman 1996), the relative energetic investment made by herbivores toward detoxification can have important implications for host-parasitoid interactions (Bukovinszky et al. 2009; Kapari et al. 2006).

In the last decade, the severity of drought-associated insect outbreaks have negatively impacted pinyon pines (Pinus edulis Engelm.) in the Western United States resulting in large-scale mortality events and significant shifts in pinyon-juniper ecosystem structure and 79 function (Hanson & Weltzin 2000; Breshears et al. 2005; Mueller et al. 2005; McDowell et al. 2008). Effective management of this ecosystem depends on our knowledge of chemically mediated plant-insect interactions. Studies have shown that the rapid increase in herbivore populations during outbreaks tends to be correlated with low rates of parasitism, but it is unclear if increased parasitoid resistance and/or inefficient plant defense chemistry is responsible (Kapari et al. 2006).

In this study, we used field observations of pinyon pine needle chemistry to create artificial diets mimicking constitutive and herbivore-induced levels of monoterpenes. We demonstrate that diets simulating herbivore-damaged needles, containing relatively low monoterpene concentrations, increase a specialist herbivore’s ability to actively defend against parasitism through enhanced consumption and unaltered growth rates. While individual monoterpenes have varying effects on herbivore relative growth rate and immunity across concentrations, when confronted with the additive effects of all the monoterpenes present in both constitutive and herbivore-induced diets, larvae show a significant trade-off between growth and immune response. The energetic trade-off forcing this response is likely due to the high metabolic costs of detoxification as we observed no evidence of sequestration. Thus, as an outbreak progresses and herbivore-damaged trees become more abundant, the southwestern tiger moth is exposed to lower levels of monoterpenes resulting in higher biomass consumption providing extra energy to mount a more effective immune response. This chemically-mediated evasion of parasitism may offer insight into the biological mechanism responsible for increased herbivore population densities in this ecologically sensitive landscape.

MATERIALS AND METHODS

Study system: Pinyon pine and a specialist herbivore

We studied one commonly associated pinyon pine herbivore, the univoltine southwestern tiger moth (Lophocampa ingens Edwards: Arctiidae) (Powell & Opler 2009). These caterpillars hatch from eggs in late summer or early fall and begin feeding gregariously while simultaneously building silken tents in which to overwinter. Larvae do not undergo complete diapause, but may feed during mild periods of the winter months and early spring (Duncan

2006). Larvae resume feeding in April and May before pupating in June. Damage can become more extensive during pandemic outbreaks, which last approximately 1-3 years.

Past research has shown that L. ingens populations are primarily controlled by the presence of their natural endoparasitic enemy (Meteorus euschausiae Muesebeck:

Braconidae) (Cranshaw 2006). Limited published literature exists describing the life history of this particular species of parasitoid, but M. euschausiae likely has a univoltine life cycle similar to their host, parasitizing 1st or 2nd instar L. ingens larvae in early fall and overwintering in its host until mid-April, when the caterpillar larvae engage in more active feeding with warmer temperatures (personal observation).

Pinyon-juniper study site

Pinyon pine trees were monitored and measured at a Bureau of Land Management site on the eastern side of the Continental Divide (38.00° N 106.25° W; elevation 2195 m) within the pinyon-juniper woodlands of southern Colorado. Pinus edulis trees at this site were experiencing varying degrees of herbivore damage during the second year of an outbreak of

L. ingens. To assess the effects of herbivore damage on pinyon pine needle chemistry, we 81 implemented a paired observational design selecting only cone-bearing trees (~80 years old) approximately 2-2.5 m tall in which a “damaged” pinyon, characterized by ~30% needle damage resulting from tiger moth feeding, was paired with an “undamaged” pinyon characterized by no visual signs of defoliation by tiger moth larvae or other folivore species

(n=8). The paired trees were of similar age and size, experienced similar micrometeorological conditions (aspect, sunlight, wind, etc.), and were within 10 m of one another.

Pinyon pine needle collections and analysis

Needles were collected 14-19 May, 2008, from 0900-1500 h from three south-facing sun- exposed branches per tree, all located approximately 1.5 m aboveground. We cut the distal

10-15 cm of each branch, which included the previous year and current year growth, and immediately placed the needles in liquid nitrogen, transported them back to the laboratory, and stored them in a -80˚C freezer for foliar terpene analysis.

We used the monoterpene extraction method of Keefover-Ring and Linhart (2010) and, through method development, modified it for processing conifer needles. Needles were ground in liquid nitrogen using a chilled mortar and pestle to minimize monoterpene loss.

Frozen needle powder was then weighed into 2 dram glass vials, exact weights (between 0.4 and 0.6 g) were recorded, and 2 mL of gas chromatograph (GC)-grade n-hexane (Fisher

Scientific, Pittsburgh, USA) containing 0.1 µL mL-1 (+)-fenchone (Sigma-Aldrich, St. Louis,

USA) as an internal standard was added to the vial. Vials were immediately closed with teflon lined caps, mixed with a vortexer, and allowed to soak for one week at ambient 82 temperature. After the seven-day soaking period, 100 µL aliquots from each sample were taken for chemical analysis.

To determine foliar monoterpene concentrations, one µL of each sample was injected on an Agilent Technologies 6890 gas chromatograph-5975 mass spectrometer (GC-MS) fitted with a Cyclodex-B chiral column (30 m 250 µm 0.25 µm; J&W Scientific, Santa Clara,

USA). Helium was used as the carrier gas at a flow rate of 1.0 mL min-1 with a split flow ratio of 25:1. Injector temperature was set at 230 ºC while the oven profile consisted of an initial temperature of 60 ºC followed by a ramp of 5 ºC min-1 to 200 ºC, then a second ramp of 25 ºC min-1 to 230 ºC with no temperature hold. Monoterpene enantiomers were identified by comparing retention times and mass spectra to authentic standards using MSD

Chemstation software version D.02.00.275 (Agilent Technologies, Santa Clara, USA).

Concentrations for each compound were calculated using 4-point calibration curves with injections of known amounts of pure standards and fenchone as an internal standard. All standards were purchased from Sigma Aldrich with the exception of β-phellandrene, which was from Glidco Organics (Jacksonville, USA).

To determine whether foliar concentrations of monoterpenes were different between undamaged and herbivore damaged pinyon pine trees, we analyzed these data using ANOVA

(Proc Mixed; statistical analyses were performed using SAS, Version 9.1; SAS Institute, Inc.,

Cary, USA).

Tiger moth collection and artificial diet studies

Lophocampa ingens tents (n=5, ~1,000 caterpillars total) were collected from natural pinyon pine populations on 11 April 2011 at an outbreak near Tijeras, New Mexico (35.08 ºN 83

106.38 ºW, elevation 6396 ft) on the Eastern Slope of the Continental Divide. Second and third instar larvae were transported back to Boulder, Colorado and kept in growth chambers with a photoperiod of 14 hour days and 10 hour nights with and day:night temperatures of

22 ºC: 15 ºC.

To determine the effects of monoterpene quality and quantity typical of undamaged and herbivore damaged pinyon pine foliage on L. ingens growth and immune response, measured foliar concentrations (described above) were mimicked using artificial diets. Of the seven monoterpenes and their enantiomers that were identified and quantified, four were chosen based on their abundance, the availability of standards, and altered concentration with herbivory. Both α-pinene [(-)-α-pinene and (+)-α-pinene] and β-pinene were chosen as they are the primary constituents of pinyon resin, uncertainties exist regarding their defensive role against folivores, and they are important solvents for the other monoterpenes and resin acids.

The compound β-myrcene was selected on the basis of its significantly higher concentrations in undamaged pinyons and its demonstrated anti-herbivore function of increasing detoxifying enzyme activity (mid-gut esterases) in lepidopteran species (Redak & Cates 1984; Mukherjee

2010). The compound (-)-limonene was also found in significantly higher concentrations in undamaged pinyons and has been shown to reduce insect activity in a variety of insect orders and families [Lepidoptera (Peterson et al. 1994; Sadof & Grant 1997; Petrakis et al. 2005),

Coleoptera (Smith 1975; Alfaro et al. 1980; Regnault-Roger et al. 1993), Homoptera

(Valterova et al. 1997)].

Artificial diets were prepared using tobacco budworm diet (Bio-Serv, Frenchtown, USA) as a protein and carbohydrate source, and the appropriate monterpene concentrations were added and mixed thoroughly after the diet had cooled to avoid monoterpene loss. The diets 84 were then refrigerated immediately until needed. Target concentrations were confirmed by extracting 0.6 g of each diet in 2 mL n-hexane containing 0.1 µL mL-1 (+)-fenchone as an internal standard and identifying and quantifying the compounds using the GC-MS methods described above. Diets were fed to the larvae in excess, and were changed every other day to ensure that monoterpene loss via evaporation was minimal.

Using a total of 23 artificial diets, including a control, created based on monoterpene concentrations and proportions observed in the field (Table 4.1 and Fig. 4.1), we designed two experiments to address the following objectives: 1) Determine the individual effect of each monoterpene on herbivore growth and immune response across biologically relevant concentrations, and 2) Investigate whether the synergistic effects of the monoterpenes present in undamaged and herbivore-damaged needles affect herbivore growth and immune response differentially across biologically relevant concentrations. The volume of each monoterpene standard needed to mimic field foliage concentrations (mL monoterpene L-1 agar) was determined by dividing the average foliar concentration at which each compound was found in both herbivore-damaged and undamaged pinyons (mg gfw-1) by the density of each compound (g mL-1) and multiplying by mass and volume conversion factors.

Table 4.1. Mean foliar monoterpene concentrations (mg gfw-1) of individual compounds added to artificial diets mimicking undamaged (top) and herbivore-damaged (bottom) P. edulis trees during herbivory in May on the Eastern Slope of the Rocky Mountains in Saguache, County, Colorado. Levels of each compound added were based on concentrations observed in the field (M: medium) then subsequently doubled (H: high), quadrupled (UH: ultra-high), halved (L: low), and quartered (UL: ultra-low). Diets created to investigate effects of individual compounds were based on levels observed in undamaged trees and only included three concentration gradients: L, M, and H (i.e. a total of 12 diets marked with *), and diets created to investigate synergistic effects of combined monoterpenes are marked with § (i.e. a total of 10 treatments plus the control (artificial diet with no monoterpenes).

Undamaged UL L M H UH 54% α-pinene 0.52 1.04* 2.09* 4.17* 8.34 28% β-pinene 0.26 0.52* 1.05* 2.10* 4.20 15% β-myrcene 0.14 0.29* 0.57* 1.15* 2.30 3% (-)-limonene 0.03 0.06* 0.13* 0.25* 0.50 Total 0.96§ 1.92§ 3.83§ 7.67§ 15.34§

Damaged UL L M H UH 55% α-pinene 0.36 0.73 1.46 2.92 5.83 30% β-pinene 0.20 0.40 0.80 1.60 3.20 13% β-myrcene 0.09 0.17 0.34 0.68 1.35 2% (-)-limonene 0.01 0.03 0.06 0.12 0.24 Total 0.66§ 1.33§ 2.66§ 5.31§ 10.63§

Fig. 4.1 Foliar monoterpene concentrations of Pinus edulis with or without herbivore damage observed on the Eastern Slope of the Rocky Mountains in Southern Colorado during caterpillar feeding in May 2008. * Indicates a significant difference between treatments (P<0.05) and values are means ± SEM (n=8). (+)ap=(+)-α-pinene; (-)ap=(-)-α-pinene; bp=β-pinene; bmyr=β- myrcene; bphel=β-phellandrene; cam=camphene; and lim=(-)-limonene.

To address objective 1 and examine how L. ingens larvae respond to individual monoterpenes, a set of diets were prepared that contained each individual compounds (α- pinene [(-)-α-pinene and (+)-α-pinene)], β-pinene, β-myrcene, and (-)-limonene) mimicking constitutive levels observed in undamaged pinyon needles in the field (“medium”) and two other concentrations, twice the amount found naturally (“high”) and half the constitutive level (“low”).

In order to assess the additive effects of the monoterpenes present in both undamaged and herbivore-damaged foliage on L. ingens growth rate and immune response (objective 2), two stock solutions, hereafter ‘damaged’ and ‘undamaged’, were created. Each solution contained all four monoterpenes (see above) representative of their proper relative proportions to one another observed in the field. These were then added to artificial diets at five concentrations representing the average levels found in the field (“medium”) as well as two higher and two lower concentrations. The two high concentrations were double (“high”) and quadruple

(“ultra high”) the average amount observed in the field and the two low concentrations were half (“low”) and quarter (“ultra low”) the average levels found naturally (Table 1; Fig. 1); however, all concentrations were well within the biologically relevant range measured from natural populations.

For both experiments, three third instar larvae were chosen from the five tents collected from the field to minimize the confounding variable of herbivore genetic family, and maintained on each treatment (n=15 larvae per diet). Larvae were placed individually in 9 oz. opaque plastic cups (Solo, Lake Forest, USA), which contained an excess of fresh artificial diet that matched the treatment group to which it was randomly assigned and their initial weights and population (i.e. the tent from which they came) were recorded. Larvae were 88 reared under the growth chamber conditions described above. Artificial diets were replaced and the cups were cleaned of frass every other day. Molting events were noted and dead larvae were discarded.

The role of monoterpenes in affecting caterpillar relative growth rates and immunocompetency assay

To investigate the influence of monoterpenes on caterpillar growth, larvae from the two previously described experiments (all 23 diets) were weighed every other day until they reached the fifth instar (approximately 2 weeks) to calculate relative growth rates (RGR).

The RGR for each larva was calculated as biomass gained in 14 days (mg) / initial weight

(mg) (Karban et al. 2010). To compare the effects of undamaged and herbivore-damaged diets at different concentrations on RGR, data were square root transformed and analyzed using ANOVA, with caterpillar population as a random effect (Proc mixed; SAS). Because caterpillar initial weight was already included in the calculation of RGR, it was not included in the model statement.

To estimate larval nutritional indices, once larvae molted to the 5th instar they were starved for 24 hours and their initial weight and weight of the artificial diet was recorded, they were allowed to continue feeding. After 24 hours the larvae, frass, and remaining food were weighed again to calculate nutritional indices following the Waldbauer (1968) method.

The following indices were calculated for each caterpillar: consumption index, CI [food consumed/average weight gained]; approximate digestibility, AD [(food ingested – frass)/

(mass of food ingested)]; and efficiency of conversion of digested food, ECD [(larval mass gain)/ (food ingested – frass)]. For each of the 22 diets, we determined if differences in 89 concentration resulted in altered AD, ECD, and CI by using one-way ANOVA (Proc mixed;

SAS). Student’s t-tests were used to compare average AD, ECD, CI, and frass production between larvae grown on undamaged and damaged diets. A Pearson’s product-moment coefficient (Proc corr; SAS) was used to measure the correlation between RGR and ECD and assess the influence of post-digestive processes in determining growth.

After data for nutritional indices were collected, caterpillars were injected with silica beads as proxies for parasitoid eggs to quantify the immune response following methods in

Smilanich et al. (2009). This method provides a general assessment of immune function in insects (Rantala and Roff 2007) and controls for the presence of unwanted viruses and fluids that can be injected with the use of parasitic wasps (Smilanich et al. 2009). Briefly, silica beads (DEAE Sephadex A25, 40-120 µm in size; Sigma-Aldrich) were dyed using a 0.1% mixture of Congo red dye and allowed to dry. The beads were then suspended in Ringer’s solution and stored in the refrigerator until injections were performed. After larvae were briefly anesthetized using CO2, a 10 µL micropipette (Fischer Scientific) was used to inject each larva with approximately 10-20 beads at the base of the third proleg. Caterpillars were returned to their food for 24 hours and then freeze-killed. Beads were retrieved by dissecting the caterpillars in Ringer’s solution and photographed at 60X magnification using a camera mounted on a dissection microscope (AxioCam software; Carl Zeiss, Oberkochen Baden-

Wurttemburg, Germany). Using Adobe Photoshop version 6.0 (Adobe, San Jose, USA), the red value (R value) of each bead was measured and a mean R value calculated for each individual caterpillar. Because the beads were originally dyed red, this technique assesses levels of encapsulation and melanization based on darker values, with more melanization indicating a greater immune response. The mean R value was also calculated for control 90 beads that were dyed together with those used for caterpillars but not injected. The relative change in R value between control and injected beads [1- (R value caterpillar / R value control)] was used as a measure of the proportion of melanization for each caterpillar.

To compare the effects of individual compounds on immunocompetence, proportional melanization data were arcsine square root transformed and each diet was analyzed separately using ANOVA (Proc mixed; SAS) with concentration as a fixed effect. The effect of undamaged and damaged diets on immune response was analyzed with arcsine square root transformed melanization data using ANOVA (Proc mixed; SAS). Diet and concentration were input as fixed effects and the number of beads recovered, as well as the population of origin, were listed in the model statement. The existence of a trade-off between RGR and the proportion of melanization was also determined using a Pearson’s product moment coefficient (Proc corr; SAS).

Assessing potential for monoterpene sequestration and metabolism

The potential for L. ingens to sequester monoterpenes and/or detoxify these compounds may influence the allocation of available energy toward other biological processes, i.e. growth and immunity. Therefore, we measured monoterpene concentrations in intact larvae and frass via

GC-MS. After dissection for bead retrieval and the removal of their digestive system, each caterpillar was transferred into 15 mL test tubes containing 4 mL of 0.1µL fenchone/mL hexane solution, ground using sand and glass rods, and left at room temperature for one week. Frass from caterpillars fed on each diet was collected during the fourth instar and examined for monoterpenes to determine whether tiger moth larvae excrete the ingested monoterpenes rather than pass them across the gut into the hemocoel. Frass was collected every day and stored in 91 glass vials in a -20 ˚C freezer. For sample preparation, approximately 0.4 g was ground with liquid nitrogen into a fine powder using a chilled mortar and pestle and extracted in 0.1µL fenchone/mL hexane solution for a week. Both insect and frass samples were prepared for GC-

MS analysis as described above.

RESULTS

Before we investigated how altered foliar monoterpene concentrations impact L. ingens via growth and immune responses, we determined how monoterpene quality and quantity differed between undamaged and herbivore-damaged pinyon pine needles from a field population. While undamaged trees, and therefore the respective artificial diet, exhibited total monoterpene concentrations ~1.5 times higher than damaged trees (undamaged: 3.83 mg gfw-1; damaged: 2.66 mg gfw-1; Table 4.1), only two minor constituents, β-myrcene

(undamaged: 0.57±0.06 mg gfw-1, damaged: 0.34±0.05 mg gfw-1, P<0.005) and (-)-limonene

(undamaged: 0.13±0.02 mg gfw-1, damaged: 0.06±0.02 mg gfw-1, P<0.05), were found in significantly higher concentrations in undamaged trees (Fig. 4.1).

The quality and quantity of monoterpenes presented in the artificial diets differentially affected L. ingens RGR. Caterpillars fed on α-pinene exhibited a trend towards decreased relative growth rates (F2, 11 = 1.72, P=0.22) with increasing concentrations (Fig. 4.2A). When larvae were fed diets of varying concentrations of β-pinene, β-myrcene, or (-)-limonene, there were non- significant trends towards higher relative growth rates when grown on “medium” diets, which represented the monoterpene concentrations of undamaged trees observed in the field (Fig. 4.2C,

E, and G; F2, 14 = 0.32, P=0.73; F2, 12 = 2.08, P=0.16, F2, 13 = 1.07, P=0.37, respectively). There 92 was no significant difference in AD, ECD, or CI for caterpillars grown on varying concentrations of α-pinene, β-pinene, β-myrcene, or (-)-limonene.

Similar to RGR, caterpillars fed on α-pinene exhibited a trend towards decreased levels of melanization with increasing concentrations (Fig. 4.2B; F2, 6 = 4.00, P=0.07), resulting in a positive linear relationship between RGR and proportion melanization (r = 0.96, P=0.19) (Fig.

4.3A). For caterpillars grown on β-pinene and (-)-limonene, the greatest immune response occurred in larvae feeding on the extreme concentrations of each compounds and the lowest proportions of melanization were exhibited when caterpillars were grown on the “medium” diet

(Fig. 4.2D and H). As a result, significant negative linear relationships were observed between

RGR and immune response (β-pinene: r = -0.98, P<0.05; (-)-limonene: r = -0.99, P<0.05) (Fig.

4.3B and D).

Fig. 4.2 Relative growth rate (RGR) and proportion of melanization for larvae grown on low, medium, and high concentrations of α-pinene (panels A, B), β-pinene (panels C, D), β-myrcene (panels E, F), and (-)-limonene (panels G, H). There were no significant differences in relative growth rate or proportion of melanization between concentrations for any compound. Values are means ± SEM (n=15).

Fig. 4.3 Correlation between relative growth rate (RGR) and proportion of melanization for larvae grown on diets containing low, medium, and high concentrations of individual monoterpene compounds found in undamaged pinyon pine needles in the fieldIndividual points represent mean values ± SEM (n=15).

There was no significant difference in RGR between caterpillars grown on varying concentrations of the diets mimicking undamaged (F4, 21=0.09, P>0.05) or damaged pinyon pine trees (F4, 21=0.32, P>0.05). Due to the lack of significant differences between concentrations, data were combined to compare RGR between the two diets, and no difference was observed in RGR between caterpillars grown on damaged or undamaged diets

(F1, 70=1.07, P=0.30). Larvae grown on diets mimicking both undamaged and herbivore- damaged pinyon pine trees exhibited similar AD and ECD across all monoterpene concentrations and after pooling the data (Fig. 4.4A, AD: 0.94 ± 0.005 and Fig. 4.4B, ECD:

0.17 ± 0.02, respectively). However, larvae grown on damaged diets had a significantly higher consumption index (CI) than those grown on undamaged diets (Fig. 4.4C, t = -3.91, df

= 68, P < 0.0005), and also tended to produce less frass, although not significantly so (Fig.

4.4D). In short, L. ingens grown on the damaged diet had greater CI despite similar ECD and

RGR exhibited by larvae grown on undamaged diets.

Fig. 4.4 (A) approximate digestibility (AD), (B) efficiency with which digested food is converted to body substance (ECD), (C) consumption index (CI), and (D) amount of frass produced over 24 hours for caterpillars grown on diets mimicking monoterpene levels in undamaged (grey bars) and herbivore-damaged (black bars) pinyons in Eastern Slope study site. ** Indicates a significant difference between treatments (P<0.05) and bars represent means ± SEM (n=50).

There was no significant difference in melanization between caterpillars grown on varying concentrations of the diet mimicking undamaged pinyon pine trees (F4, 13 =0.47,

P>0.05) or damaged pinyon pine trees (F4, 25=0.38, P>0.05). Due to the lack of significant differences between concentrations, data were combined to compare proportions of melanization between damaged and undamaged diets. Larvae exposed to damaged diets exhibited a significantly greater overall melanization response (F1, 52=4.03, P<0.05). In addition there was a significant negative relationship between proportion of melanization and

RGR for both the undamaged and damaged diets (Fig. 4.5, r = -0.61, P<0.05).

Regardless of diet or concentration, none of the four monoterpene compounds were detected or recovered from either intact fifth instar L. ingens tissues (no sequestration) or fourth instar frass (complete breakdown of compounds).

Fig. 4.5 Correlation between relative growth rate (RGR) and proportion of melanization for caterpillars grown on undamaged (open circles) and damaged (closed circles) artificial diets. The solid line represents the linear relationship between the two variables. Values are means ± SEM (n=15).

DISCUSSION

Despite the temporal link between drought-stress pinyon pines and increased susceptibility to herbivory, a potential mechanism explaining the increase phase of L.ingens outbreaks in this semi-arid system has not been acknowledged. Our field studies show that the unique feeding pattern of L. ingens on pinyon pine results in lower overall foliar concentrations, with significantly lower β-myrcene and (-)-limonene levels, as opposed to the increase in monoterpene concentrations observed in other studies (Lewinsohn et al. 1990;

Litvak et al. 1998; Byun-McKay et al. 2003). From our field study we could not rule out the possibility of herbivore-induced enzymatic up-regulation, but the level of damage and/or high emission rates (unpublished data) most likely resulted in depleted foliar monoterpene pools (Litvak et al. 1998). It is also possible that undamaged trees simply exhibited higher constitutive monoterpene levels. Regardless, using artificial diets, we mimicked these field levels of constitutive and herbivore-induced monoterpene and demonstrated that the chemical make-up of damaged pinyon pines has important implications for individual consumption, growth, and immunity.

Using constitutive levels of monoterpenes, we found individual compound type and concentration to have differential effects on L. ingens larval growth and immune response.

None of the monoterpenes used in this study were sequestered or detected in the larval frass.

Thus, the ability of L. ingens to metabolize monoterpenes is similar to other lepidopteran species (Raffa and Powell, 2004) and this energetically costly metabolic process is important to consider when evaluating trade-offs in growth, and immunocompetency. Caterpillars grown on both β-pinene and (-)-limonene exhibited a trade-off between relative growth rate and melanization across varying concentrations, exhibiting relatively high RGR but poor 100 protection against endoparasitism when exposed to average constitutive field concentrations

(Fig. 4.2C, D, G, H). While our findings are consistent with previous studies the detrimental effects of high monoterpene concentrations—particularly limonene—on growth across a spectrum of herbivore species ([Lepidoptera (Peterson et al. 1994; Sadof & Grant 1997;

Petrakis et al. 2005), Coleoptera (Smith 1975; Alfaro et al. 1980; Regnault-Roger et al.

1993), Homoptera (Valterova et al. 1997)]), the low immune response to endoparasitism and relatively high RGR when exposed to average constitutive field concentrations of (-)- limonene is notable.

While (-)-limonene may be considered a minor constituent of the total terpenes making up pinyon pine resin (~3%), studies have also shown high concentrations of the dominant monoterpene in this species, α-pinene (54%), to effectively repel herbivores (e.g. spruce beetle Wallin & Raffa 2004). Consistent with previous studies, we found increases in α- pinene concentration to reduce not only RGR but also immunocompetence. These results suggest that α-pinene causes L. ingens to invest more energy into catabolic processes, particularly at high concentrations, at the expense of allocating resources to growth or immune responses, including encapsulation and melanization. Furthermore, the low levels of melanization (~20%) exhibited by larvae reared on field concentrations of both α-pinene and

(-)-limonene may help explain the effectiveness of parasitoids in regulating L. ingens populations under typical non-outbreak field conditions. While it is possible that differences in individual monoterpene chemical structures account for a large percent of the variation in herbivore physiological responses, their effects at the molecular level remain unclear

(Gershenzon & Dudareva 2007). Specialists, such as L. ingens, have likely adapted to detoxify these compounds, but the efficiency with which this occurs is a function of the 101 quality and quantity of detoxifying enzymes needed, which is dependent on both monoterpene structure and concentration and is likely to be species-specific (Yu 1987;

Jackson et al. 2008).

Evaluating larval responses to individual monoterpenes may elucidate the biological activity of different compounds, but does not offer a realistic assessment of how L. ingens respond to altered plant chemistry in the field, nor does it take into account the possibility of antagonistic or synergistic effects (Hay et al. 1994; Nelson and Kursar 1999; Dyer et al.

2003). We observed a significant trade-off between relative growth rates and immune response across diets created to mimic monoterpene concentrations of four compounds in L. ingens-damaged or undamaged pinyon needles (Fig. 4.5). Thus, while better defended against endoparasitoids, caterpillars feeding on multiple monoterpenes present in the field may be so at the expense of growth. Neither pre-digestive (i.e. AD) nor post-digestive (i.e.

ECD) processes differed among larvae grown on the damaged and undamaged diets.

Furthermore, growth on diets mimicking damaged or undamaged pinyon pines did not result in a difference in RGR. This is in contrast to larvae that were exposed to individual compounds, where differences in both compound type and concentration led to differing

RGR. Together, these data suggest that the synergistic effects of monoterpenes result in negligible effects on ECD or growth.

While the differences in monoterpene ratios and concentrations between the damaged and undamaged diets did not influence RGR, larvae reared on diets mimicking damaged pinyon pines did exhibit, on average, significantly higher CI (Fig. 4.4C) and higher melanization. In other words, L. ingens grown on damaged plants tended to eat more biomass and exhibited greater immune responses for a given growth rate. One explanation for the increased 102 consumption on damaged diets could simply be the lower total concentration of monoterpenes present, resulting in lower toxicity, relatively higher nutritive quality, and thus, higher overall suitability. An alternative hypothesis could be that the subtle, yet specific, ratios of compounds in damaged tissue serve as a feeding stimulant during outbreaks to provide more overall energy available for defense due to lower concentrations of monoterpenes and higher nutritive quality. Lophocampa ingens larvae have chemical receptor cells on the maxillae, and variation in monoterpenes may induce receptor changes that could potentially alter feeding habits (Schoonhoven 1969). Monoterpene quality and concentration can also stimulate feeding in specialists, particularly limonene (Alfaro et al.

1980). In our study, (-)-limonene was one of two compounds that was found in significantly lower concentrations in damaged pinyon pine needles, and larvae grown on the (-)-limonene diets also demonstrated a trade-off between relative growth and melanization rates. Thus, the low levels of limonene present could potentially make the diet less toxic overall and allow for a greater caloric intake to fuel higher melanization rates. We must also consider the possibility of synergistic and antagonistic interactions, driven primarily by the two minor constituents, β-myrcene and (-)-limonene, as they were the only compounds that differed between the diets.

Our results have important ecological implications for L. ingens population dynamics in pinyon-juniper woodlands. This higher immunocompetency for larvae feeding on damaged pinyon pines may aid in explaining the increase phase of L. ingens outbreak cycle, especially considering the higher proportion of damaged trees present during an outbreak. As L. ingens population densities increase during an outbreak, the proportion of damaged trees also increases. Similar to other studies, we show that herbivore-induced chemistry conveys an 103 increase in immunocompetence during the increase phase of an outbreak (Kapari et al. 2006).

However, our results are the first to show an herbivore-induced decrease in defense compounds can occur under natural field conditions, which can result in a lower incidence of successful parasitism, perpetuating insect outbreaks. While we cannot rule out the possibility that high monoterpene emission rates may increase parasitoid perception and host-location resulting in higher overall parasitism rates, we do show that once caterpillars are parasitized they would have a better chance of surviving parasitism events if feeding on damaged trees.

Furthermore, enhanced consumption of diets mimicking damaged needles likely increases the severity of outbreaks and pinyon pine tree mortality currently observed in this ecosystem.

To our knowledge, this is the first study to investigate how herbivore-induced monoterpenes influence relative growth and immunocompetency of pinyon defoliators in natural populations. Together, our observations emphasize the importance of monoterpene concentration in mediating relationships between pinyon pines, defoliating insects, and the third trophic level, offering a new perspective in how insect attack may exacerbate mortality in this species.

104

CHAPTER V

UNTANGLING THE PRIMARY DRIVERS OF PINYON PINE NEEDLE MONOTERPENE PRODUCTION IN THE PRESENCE OF DROUGHT

ABSTRACT

Monoterpenes play a variety of ecological roles, serving as both carbon-based plant defense compounds as well as contributing to both regional and global atmospheric processes.

Environmental variation has been shown to have important consequences for monoterpene biosynthesis and emissions, yet we lack an understanding of the ecophysiological mechanisms driving these changes in semi-arid forest species and the potential effects on this ecologically sensitive habitat. Thus, we examined the effects of temperature and water availability on pinyon pine monoterpene production and emissions, by transplanting trees from their native habitat, a cooler and wetter site at a moderate elevation, to a hotter and drier habitat in a desert site at a lower elevation. We hypothesized that increased drought would increase tissue concentrations in accordance with the carbon-nutrient balance hypothesis (CNBH). Furthermore, we predicted that higher temperatures and lower water availability together would influence monoterpene emission rates, but only through shifts in tissue concentrations. Our results showed that neither total needle monoterpene concentrations nor emission rates differed between sites, and furthermore, foliar concentrations were not a significant predictor of emissions. Rather, stomatal conductance was significantly and positively correlated with monoterpene emission rates below 1 µg C gDW-1 h-1, explaining 10% of the variation in the data when measured across the growing season, but only in the pinyon pine native habitat. However, we also observed periods of emission rates greater 105 than 1 µg C gDW-1 h-1at both sites during the month of June, which were not controlled by stomatal conductance or tissue monoterpene concentration. Thus, there appear to be yet-to-be- identified phenological factors that exert control over these extremely high mid-season emission rates. Our results suggest that drought has no significant influence over tissue monoterpene concentrations in this system, which are likely under tight genetic control. However, drought and phenology may exert dominant controls over the emission of monoterpenes through influences on stomatal conductance and other plant processes. Our results demonstrate a greater role than previously recognized for ecophysiological processes, uncoupled from tissue monoterpene concentrations, in controlling monoterpene fluxes from semi-arid conifer species.

INTRODUCTION

Pinyon-juniper woodlands are a dominant ecosystem in the Southwestern U.S., (ca. 15 million hectares), yet severe droughts within the last decade have resulted in high levels of pinyon pine mortality causing significant community shifts with consequences for regional ecology (Mueller et al., 2005). Even under conservative climate models, the proximate factors causing pinyon pine mortality (extreme heat and water stress) are predicted to increase in both frequency and severity (IPCC 2007). Studies have focused on understanding the physiological differences between pinyon pines and junipers, the latter being less susceptible to extreme droughts (Breshears et al. 1997; Linton et al. 1998). Pinyon pine is isohydric, which means it adjusts its stomatal conductance continuously downward during drought to maintain a leaf water potential of ca. -2 MPa, and consequently experiences low rates of net carbon gain during drought, potentially leading to death by carbon starvation (Lajtha & Barnes 1991; Williams &

Ehleringer 2000; West et al. 2008; McDowell et al. 2008). In addition, drought-induced 106 susceptibility to insect herbivory has been cited as a stress contributing to pinyon pine mortality

(McDowell et al. 2008), likely due to weakened resin flow (Mopper et al. 1991; Cobb et al.

1997) and chemical defenses, such as monoterpenes. Despite monoterpenes being the dominant constituent of pinyon pine resin (Smith 2000) and the prevalence of drought conditions in the

Southwest, there exists a gap in our knowledge as to the sensitivity of monoterpene biosynthesis and emissions to altered temperature and precipitation in arid systems. Thus, identifying how environmental drivers affect pinyon pine physiological and monoterpene production will offer a foundation for assessing the subsequent consequences for plant-insect and biosphere-atmosphere relationships.

In conifers, monoterpenes are known to play important roles as plant defense compounds by acting as oviposition and feeding deterrents (Hummelbrunner & Isman 2001; Abdelgaleil et al.

2009), negatively affecting larval performance and survival through toxicity (Lerdau et al. 1994), decreasing insect immunocompetency (Trowbridge et al., unpublished), and serving as volatile host location cues for parasitoids (DeMoraes et al. 1998; Thaler et al. 2002; Kant et al. 2004;

Mithofer et al. 2005). One potential hypothesis for drought-related mortality of pinyon pines is that warmer minimum temperatures result in increased insect survival, which subsequently exacerbates the weakened physiological condition of the tree (McDowell et al. 2008). However, whether this drought-induced weakening of plant function and resistance is due to changes in secondary defense compounds, such as monoterpenes, remains uncertain and is likely insect species-dependent. Furthermore, studies have shown that needle concentrations and emissions are tightly coupled, with monoterpene emission rates being positively correlated to foliar concentrations (Lerdau et al. 1997). Thus, understanding how environmental variability alters both monoterpene storage as well as emission rates will offer insight into the generalization of 107 this relationship across species, ecological controls over production versus emissions, and potential bottom-up (toxicity) and top-down (signaling) consequences for herbivores.

In addition to drought influencing monoterpene-mediated multi-trophic interactions, and potentially ecosystem structure and function, altered monoterpene emission rates can have important implications for regional atmospheric composition. Once released into the atmosphere, monoterpenes are involved in the production of ozone and carbon monoxide as well as organic nitrates and secondary organic aerosols (Jacob & Wofsy 1988; Fehsenfeld et al. 1992). Volatile monoterpenes can also react with hydroxyl radicals (OH), decreasing its oxidative availability and increasing the lifetimes and concentrations of other compounds such as CH4 and CO

(Fehsenfeld et al. 1992). The volatility and emission rates of monoterpenes are controlled by the vapor pressure of these compounds within plant tissues, which is in turn controlled by temperature and the rate of biosynthesis (Lerdau et al. 1994). Resource allocation towards monoterpene production can be predicted based on the Carbon-Nutrient Balance Hypothesis

(GDBH), with states that under resource limiting conditions, such as drought, plants should allocate more of their resources towards these defensive compounds. Whether pinyon pines truly increase monoterpene foliar concentrations under drought conditions and this increased production results in a subsequent increase in emissions remains unclear.

The composition and concentration of monoterpenes in pine oleoresin is dynamic and can be significantly altered by both abiotic and biotic factors (Zulak & Bohlmann, 2010). The exponential relationship of monoterpene emissions with temperature is well-established (Tingey et al. 1980), resulting from an enhancement of enzymatic activities of synthesis, increased vapor pressure, and decreased resistance of diffusion (Tingey et al. 1991). The loss of these compounds to the atmosphere can result in altered foliar pool concentrations (Litvak et al. 1998). However, 108 while most monoterpenes show strong temperature dependence, with emissions fitting reasonably well to atmospheric models (see Guenther et al. 1993; Bertin et al. 1997), it is important to note that some monoterpenes are not sensitive to temperature (e.g., cis- and β- ocimene) (Loreto et al. 1998). Thus, with minor exceptions, monoterpene synthesis is thought to be under strong enzymatic control. Yet, this temperature response is short-term and many studies are performed in the laboratory, thus there is a lack of understanding of how emissions respond to temperature increases in the field throughout entire growing seasons (Penuelas & Staudt,

2010).

Many regions are expected to become not only hotter under predicted climate change, but also drier (IPCC 2007). Decreases in overall water availability may affect protein and substrate levels resulting in significantly lower monoterpene concentrations and thus, also negatively impacting emission rates (Lerdau et al. 1997). While a number of laboratory and field studies have shown that levels of monoterpene emissions tend to decrease during drought (Llusia &

Penuelas 1998, Staudt et al. 2002; Lavoir et al. 2009), the internal composition and concentration of monoterpenes in the needles may not always reflect the composition or levels of released volatiles. Furthermore, foliar concentrations were, in some cases, found to increase under severe drought (Llusia & Penuelas 1998), or no reduction in monoterpene synthesis was observed

(Lavoir et al. 2009). Furthermore, the effect of drought on monoterpene production and emissions depends on the severity and frequency of these events, such that some studies have shown mild drought stress to result in an increase in overall emission rates (e.g., Staudt et al.

2008). Together, the complex interactions between water availability and temperature have led to conflicting results in terms of their effects on secondary metabolism; and while both factors significantly influence monoterpene concentrations and emissions differentially, there is a lack of 109 a basic understanding of how these two variables interact to affect monoterpene production and emissions, particularly in semi-arid ecosystems.

Here, we used a transplant experiment to quantify pinyon pine monoterpene foliar concentrations and emissions under different temperature and moisture conditions. Temperature was manipulated by transplanting pinyon pines from a cooler (on average) pinyon juniper woodland to a hotter desert. Water status was altered by either watering trees or adding an impervious barrier to prevent water from entering the soil as precipitation and leaving it through evaporation. Combined, a range of temperature and moisture conditions were established that enabled us to study variation in pinyon pine monoterpene emissions and concentrations across the main pinyon pine growing season in response to seasonality and drought. We hypothesize that drought will increase tissue concentrations in accordance with plant resource allocation theory, but only affect emission rates through shifts in tissue concentrations, rather than through plant ecophysiological responses.

MATERIALS AND METHODS

Field study sites

This study took place during the 2010 growing season (May through September) at two experimental sites in Northern Arizona (a pinyon-juniper woodland— 35.49 N 111.85 W and a

Sonoran desert site— 35.45 N 111.50W). In October 2008, pinyon pine trees of similar size and age were randomly selected from the pinyon-juniper population and, using a 229 cm tree spade, were transplanted into open areas at the local pinyon-juniper site (hereafter PJ) and ~50 kilometers east at the lower elevation desert site (hereafter desert). The mean annual temperature at the desert site is ca. 2-4 ºC warmer than the PJ site. All selected trees had nearest neighbor 110 canopy-to-canopy distances of at least 1 m and were transplanted in a grid of squares with a spacing of at least 10 m. Eighteen trees at each site were randomly assigned to 3 moisture treatment groups: watered (3 trees), ambient (6 trees), and barrier (9 trees). Ambient trees were left exposed to natural conditions, watered trees received 25 gallons of water one week prior to sampling on a monthly basis, and barrier trees were fitted with a 4.3 m 4.3 m below-canopy rain-out shelter consisting of a UV resistant tarp elevated above the soil surface on a PVC frame that prevented direct vertical exchanges of moistures between the soil and atmosphere of the transplanted trees.

Measurements at both sites were conducted during five campaigns on 18-21 May, 24-27

June, 23-26 July, 20-23 August and 23-26 September 2008. Air temperature and relative humidity were measured throughout each sampling period (HOBO U12 Temperature/Relative

Humidity Data Logger, Onset, Cape Cod, MA). Pre-dawn water potential in experimental trees was measured using a pressure bomb to assess the degree of drought stress among treatment groups. All gas exchange and monoterpene emission measurements were taken ca. 1.5 m aboveground from the distal 10-15 cm of south-facing sun-exposed branches, which included the previous year and current-year growth. All measurements were carried out on clear days from

0800-1200 h with each site and tree randomly chosen for sampling between and within months, respectively. Immediately following gas exchange measurements and sampling for volatile emissions, a sub-sample of needles from the measured branches were collected and stored in liquid nitrogen, transported back to the laboratory, and stored in a -80˚C freezer until later foliar terpene analysis.

Field measurements: Gas exchange 111

Branch gas exchange was measured in the field using a portable photosynthesis system with a conifer chamber (LI-6400, LiCor Inc., Lincoln, NE, USA) to obtain rates of net CO2 assimilation (A) and stomatal conductance (gs). All measurements were taken with the condition

-1 of 400 ppm CO2 at a flow rate of 500 µmol s . Repeated measures were performed on the same trees each month by placing ~4 cm of the terminal portion of each study branch in the conifer chamber which was sealed with silly putty to minimize leaks. Because there was no controllable light source, measurements were only made once the light, CO2, and humidity measurements within the chamber were stable. Photosynthetically active radiation (PAR) and leaf and air temperature outside of the chamber were also recorded. After gas exchange measurements and volatile samples were taken, the remaining parts of the branches were harvested and stored in a labeled paper bag to obtain needle dry weight and leaf area using methods for volume displacement.

Volume displacement

We used the volume displacement approach of Chen et al. (1999) to measure needle area.

Briefly, a container of water (with 5% detergent added as a surfactant) was placed on a top- loading electronic balance and tared. The detached shoot was submersed in the water using tweezers, and the volume of the needles plus the twig was calculated from the mass of water displaced by the shoot. The shoot was then dried with paper towels and the needles were removed, counted, and their average length calculated using digital calipers (Fisherbrand

Traceable Digital Calipers 8”, Fisher Scientific Inc., Pittsburgh, PA). The volume of the twig without needles was then measured, and the needle volume was calculated by difference. Needle 112 surface area was computed from measurements of needle volume, average needle length, and hemi-cylindrical needle cross section using:

4.10 Vnl where V is the volume displacement by the needles, n is the number of needles, and l is the average length of the needles.

Monoterpene volatile emissions sampling

The dynamic headspace branch-level enclosure technique was used to measure the flux of monoterpenes emitted from pinyon pines (see Ortega et al. 2008). Two headspace chambers were created from cut Teflon bags (48 cm 60 cm) with one end placed over a custom made

0.635 cm (¼”) thick 25 cm diameter Teflon back plate fitted with 0.635 cm inlet and outlet bulkhead fittings as well as a groove around the edge to secure the bag with a large rubber band.

A new bag was used for each branch sampled to ensure no memory effects or off-gassing from previous samples. Each back plate was equipped with an 8 cm2 2 cm fan (Jameco Electronics,

Belmont, CA) powered with a 12V battery and a 2-channel temperature and humidity data logger

(HOBO U12 Temperature/Relative Humidity Data Logger, Onset, Cape Cod, MA). The end of the bag not secured to the back plate was placed over the same branch sections measured for gas exchange and fastened with a plastic zip tie around a portion of the twig free of needles. The branch enclosure was purged with clean air at a rate of 1 L min-1 for 10 minutes by scrubbing ambient air with activated charcoal (Alltech Associates Inc., Deerfield, IL) and MnO2 coated net ozone scrubbers (O.B.E. Corporation, Fredericksburg, TX).

Monoterpenes were sampled onto custom-made glass adsorbent cartridges (7.6 cm long,

0.635 cm OD, Allen Scientific Inc., Boulder, CO) packed with ~0.25 g Tenax® GR adsorbent 113

(20/35 mesh, Alltech Associates Inc., Deerfield, IL) between plugs of quartz wool at a flow rate of 150 mL min-1 for 10 minutes. VOC-free air was delivered to the chambers using a double- headed diaphragm pump and volatiles were collected by pulling the chamber air over the sample tubes using a single-headed diaphragm pump (Brailsford & Company, Antrim, NH). Air was transported through the system using 0.635 cm ID Teflon tubing connected with stainless steel

Swagelok fittings. All flow rates were set and controlled using two 5000 sccm and two 500 sccm mass flow controllers and a 4-channel power supply readout box (MKS Instruments Inc.,

Methuen, MA). This two chamber set-up allowed for volatile sampling of two branches simultaneously to minimize variation in emissions due to light, temperature, or time of day.

Immediately after sampling, the glass tube was disconnected from the outlet and both ends were capped with Swagelok nuts and caps, cooled to 0 ºC, transported back to the lab, and stored in a -20 ºC freezer until later chemical analysis. Sample branches were then immediately cut, placed in labeled paper bags, and brought back to the lab to determine total needle area. Dry weight was measured after drying the needles at 60 ºC for 48 hours.

Volatile chemical analysis: thermal desorption

Sample tubes were analyzed for monoterpene emission rates within 7-21 days using thermal desorption (Perkin-Elmer ATD400) GC-FID/MS (Hewlet-Packard 5890/5970, Wilmington, DE) following Helmig et al. (2004). The thermal desorption automated cartridge desorber was set at a hydrogen purge flow rate of 50 mL min-1 and an outlet-split flow rate of 10.2 mL min-1. Volatiles were desorbed at 300 ºC for 30 minutes. The micro cold-trap was set at -20 ºC with a desorption temperature of 325 ºC. The GC transfer temperature was 275 ºC and the transfer flow rate through the column was 2.1 mL min-1. The GC was fitted with a DB-1 column (30m length, 114

0.320 ID, 0.025 film thickness, J & W Scientific, Folsom, CA) and compound separation was achieved through the following program: initial temperature of 40 ºC held for 5 minutes with a ramp of 6 ºC min-1 up to 200 ºC and held for 5 minutes at this final temperature. The carrier gas was hydrogen and was added at a column flow rate of 2.1 mL min-1 and a MS split flow rate of

0.4 mL min-1 and FID spit flow rate of 1.7 mL min-1. The FID air flow rate was 245 mL min-1 with a supplemental hydrogen flow rate 21.6 mL min-1 and was calibrated with an n-alkane standard before every set of runs to determine the response factor (RF, peak area mL-1 ppbC-1) needed to convert results into true concentrations. Individual compounds were identified by using authentic standards as well as the comparison of retention index (RI) and mass spectra scans to those reported by Adams (1989). Chromatograms were integrated using PeakSimple software (SRI Instruments, Menlo Park, CA). Emission rates were calculated using:

ER = [(Cout – Cin) × Q ] / mdry

-1 where Cout and Cin are the outlet and inlet concentrations of the compound of interest (µg C l ),

-1 Q is the flow rate of the bag purge air (1000 ml min ) and mdry is the dried mass (g) of enclosed needles. Cin was set to zero because it consisted of VOC-free air. To determine the concentration of the compounds of interest we first accounted for the RF correction. Because pure materials were not used for detector calibration, we used:

PeakArea MWC C out RF ECN where MWC is the molecular weight of monoterpenes and ECN is the effective carbon number for each monoterpene. This value was then corrected for temperature and pressure. Unit conversions for time and mass were then used to obtain final emission rates in µg C gram dry weight (gdw)-1 hour-1.

115

Foliar samples and monoterpene chemical analysis

At the time of collection, needles were flash frozen and stored in liquid nitrogen, transported to the lab, and placed in a -80 ºC freezer. We used the monoterpene extraction method of

Keefover-Ring and Linhart (2010) and, through method development, modified it for processing conifer needles. Needles were ground in liquid nitrogen using a chilled mortar and pestle to minimize monoterpene loss. Frozen needle powder was then weighed into 2 dram glass vials, exact weights (between 0.4 and 0.6 g) were recorded, and 2 mL of gas chromatograph (GC)- grade n-hexane (Fisher Scientific, Pittsburgh, USA) containing 0.1 µL mL-1 (+)-fenchone

(Sigma-Aldrich, St. Louis, USA) as an internal standard was added to the vial. Vials were immediately closed with teflon lined caps, mixed with a vortexer, and allowed to soak for one week at ambient temperature. After the seven-day soaking period, 100 µL aliquots from each sample were taken for chemical analysis.

One µL of each sample was injected onto an Agilent Technologies 6890 gas chromatograph-

5975 mass spectrometer (GC-MS) fitted with a Cyclodex-B chiral column (30 m 250 µm

0.25 µm; J&W Scientific). Helium was used as the carrier gas at a flow rate of 1.0 mL min-1 with a split flow ratio of 25:1. Injector temperature was set at 230 ºC. The oven profile consisted of an initial temperature of 60 ºC followed by a ramp of 5 ºC min-1 to 200 ºC, then a second ramp of 25

ºC min-1 to 230 ºC. Monoterpene enantiomers were identified by comparing retention times of known standards and mass spectra using the NIST library and MSD Chemstation software.

Concentrations for each compound were calculated using 4-point calibration curves with injections of known amounts of pure standards and the internal standard, fenchone. All standards were purchased from Sigma Aldrich (Saint Louis, MO) with the exception of β-phellandrene, which was from Glidco Organics (Jacksonville, FL). 116

Statistical analyses

All statistical analyses were performed using SAS statistical software (SAS Institute, 2010).

To compare net CO2 assimilation rates (A) and stomatal conductance rates (gs) between sites and treatments over time, we used a mixed model (proc mixed procedure) repeated measures

ANOVA with site and water treatment as the fixed effect and a Bonferroni correction. Total and individual monoterpene emission data were log transformed to meet assumptions of normality.

To determine the effect of site and water treatment on emissions and foliar concentrations over time, total and individual monoterpene emission and concentration data from the two sites were analyzed separately using a repeated measured ANOVA as above, with a Bonferroni correction and relative humidity as a covariate for emission data. Pearson’s correlation coefficients and significance values were used to determine significant correlations among environmental drivers and monoterpene emissions and concentrations. The distributions of individual and total monoterpene emissions and concentrations were computed using kernel density estimation for visualization.

RESULTS

Distributions of observed monoterpene emissions and concentrations

Monoterpene emissions were comprised largely of α-pinene (Fig. 5.1A), which makes up

30% of total monoterpene emissions on average (Table 5.1). The distributions of α-pinene, β- pinene and limonene have fatter tails on the high emission side of the emission rate distributions than do β-phellandrene, camphene, and δ-carene. β-myrcene, β-phellandrene, limonene and δ- carene comprise 13-14% of monoterpene emissions on average, and β-pinene and camphene 117 comprise 7-8% of monoterpene emissions on average (Table 5.1). The compound α-pinene comprised on average 48%, but sometimes up to 66%, of the observed total monoterpene foliar concentration (Fig. 5.1B; Table 5.1). β-pinene comprised on average 19% of the total monoterpene foliar concentration, and follows a bimodal distribution with a secondary peak at a higher concentration than the mode. β-myrcene and limonene comprised 13% of total monoterpene concentrations on average, β-phellandrene and camphene comprised 2-4% of total monoterpene concentrations on average, and δ-carene fell below the GC-MS measurement detection limit in all instances.

118

Fig. 5.1 Frequency distributions of total and individual monoterpene emission rates (A) and foliar concentrations (B) from pinyon pine needles at both the desert and pinyon-juniper sites across all three treatments (drought, ambient, watered) and time (May-September). Data have been log transformed and the total is represented as the sum of seven compounds (α-pinene, β- pinene, β-myrcene, β-phellandrene, limonene, camphene, and δ-carene) measured here.

119

Table 5.1 The mean, standard deviation, and maximum (in parentheses) of the percentage of total monoterpene emissions and needle concentrations observed in pinyon pine across all sites and treatments in a transplant experiment in northern Arizona, USA.

Percent Total Percent Total Emission Concentration α-pinene 30 ± 17 (75) 48 ± 9 (67) β-pinene 7 ± 5 (35) 19 ± 9 (45) β-myrcene 13 ± 6 (35) 13 ± 4 (30) β-phellandrene 14 ± 9 (57) 4 ± 2 (9) Limonene 13 ± 8 (51) 13 ± 6 (36) Camphene 8 ± 4 (25) 2 ± 2 (10) δ-Carene 14 ± 7 (39) N/A

120

Seasonal patterns of monoterpene emissions and concentrations

The water treatments did not work as intended, in that there were no significant differences in the means for water potential or gas exchange parameters among the treatments. Thus, we combined all treatments for further analysis. On average, the desert site was warmer (28.7 ºC versus 27.2 ºC) and drier (-1.80 MPa versus -1.54 MPa) than the PJ site, but total monoterpene emissions were not significantly different between sites; emissions were only significantly affected by time (F4, 257=16.39, P<0.0001). Total monoterpene emission rates peaked significantly at both sites from May to June (P<0.01) and were explained by transient increases in α-pinene, β-pinene, β-myrcene, and limonene (Fig. 5.2). Monoterpene emission rates of > 1

µg C gDW-1 h-1 comprised 25% of the observations in June, but only 4.5% of the entire observation record. The relative contribution of different monoterpene compounds changed throughout the growing season as each compound followed different seasonal patterns.

Time was significantly related to total foliar monoterpene concentrations (F4,232= 8.88;

P<0.0001), but no difference between sites was observed. Thus, the difference in average seasonal temperature difference between the PJ and desert sites did not differentially affect total foliar monoterpenes. Furthermore, we found no relationship between foliar concentrations and photosynthesis or stomatal conductance, suggesting that monoterpene synthesis and storage in pinyon needles are not strongly coupled to net carbon uptake rates or water status. Unlike total monoterpene emissions, total monoterpene concentrations did not experience a pronounced increase during June (Fig. 5.3) and were not significantly correlated with emissions. Yet, similar to monoterpene emissions, concentrations of individual compounds in the needles showed different patterns of foliar production across time and between sites such that the mixture within the needles varied throughout the growing season. 121

Figure 5.2. Total and individual monoterpene emission rates (µg C gDW-1 h-1) from pinyon pine needles at both the desert and pinyon-juniper sites from May to September. Values were determined using repeated measures ANOVA and error bars represent ± SEM.

122

Figure 5.3. Total and individual monoterpene foliar concentrations (mg gFW-1) from pinyon pine needles at both the desert and pinyon-juniper sites from May to September. Values were determined using repeated measures ANOVA and error bars represent ± SEM.

123

The effect of pre-dawn water potential on pinyon pine primary and secondary metabolism

Neither total foliar monoterpene emissions nor concentrations were significantly correlated with Ψ when all observations were pooled (Figure 5.4A and 5.4C). Rather, the distributions of both foliar terpene emissions and concentrations exhibited a Gaussian-type distribution when plotted against Ψ, with a peak (in the case of emissions) or minimum (in the case of concentrations) at ca. -1.5 MPa. Thus, at the extreme ends of the Ψ range, we find low relative emissions but high relative concentrations. We found that the data points demonstrating less negative Ψ and low emissions were from pinyon pine trees measured in September. Stomatal conductances (gs) were significantly correlated with lower rates of monoterpene emissions (< 1

µg C gDW-1 h-1, P < 0.05) (Figure 5.4D), which were, in turn, significantly correlated with pre- dawn water potential (P < 0.01), along with photosynthesis (A; Fig. 5.4B) as described by the exponential relationships

A = 2.1exp(0.97Ψ)

gs = 0.040exp(1.3Ψ) 124

-1 -1 Fig.5.4. Total monoterpene emission rates (µg C gDW h ) (A), net CO2 assimilation rates (B), -1 total monoterpene needle concentrations (mg gFW ) (C), and stomatal conductance (gs) rates (D) as a function of pre-dawn water potential (Ψ) for pinyon pines grown at both pinyon-juniper and desert sites across the May to September growing season. A and gs are significantly related to Ψ (P < 0.001).

125

DISCUSSION

To our knowledge this study is the first to explore the consequences of altered hydrology and temperature on monterpene emissions and concentrations in situ in such an ecologically sensitive system.We show that pinyon pine needle monoterpene emissions are under significant stomatal control, which is itself controlled by predawn water potential, in contrast to foliar concentrations driving emissions in this species. This observation is different from other studies and atmospheric models that have suggested or derived algorithms based solely on the effects of foliar monoterpene pool size and temperature for estimating future monoterpene emissions

(Guenther et al. 1993; Lerdau et al. 1997). Furthermore, we show that total emissions are not correlated with at relatively low and high water potentials; there is a definite optimum in the distribution of this relationship. Finally, we demonstrate that, on average, total monoterpene concentrations are high and emission rates are low when pinyon pines exhibit water potentials below the zero net assimilation threshold (-2 MPa).

In pinyon pine, α-pinene was the primary constituent of total monoterpene concentrations and emissions, in agreement with past assessments of xylem chemistry of this species from populations across the Southwestern U.S. (e.g. Smith 2000), and drove many of the patterns that we observed for both total emissions and concentrations. Although many individual compounds exhibit similar concentration and/or emission patterns to α-pinene (e.g. β-pinene and β-myrcene), the magnitude and percent change over time, between sites, and in response to water status was often different. Furthermore, some compounds show little or no change in either concentration or emissions (e.g. β-phellandrene), perhaps due to lower relative volatility and/or deterministic gene expression patterns. In conifers, prenyl diphosphates are produced as substrates via the MEP pathway for terpene synthases (TPS) and cyclases (Croteau 1987; Keeling & Bohlmann 2006), 126 which can be upregulated in response to both biotic and abiotic stress (Litvak et al. 1998; Zulak

& Bohlman, 2010). These enzymes provide the vast diversity of structures possessed by monoterpenes, and are capable of producing single or multiple products. Differential upregulation of individual enzymes in response to temperature, water, or seasonal variability may help explain the patterns we observed, especially with regard to the observation that foliar concentrations are less affected by changes in environment than are emissions (discussed below).

While some monoterpenes have been shown to have dosage dependent effects on herbivores

(e.g. Langenheim 1994), synergistic effects and the ratio of compounds to one another can influence herbivores directly by affecting growth and immunocompetence (Dyer et al. 2003) but also indirectly by attracting parasitoids (e.g. DeMoraes et al. 1998, Havill & Raffa 2000). Thus, the concentrations and ratios found in this study offer a foundation for future work to determine the effects of this altered chemistry on specific herbivores shown to significantly affect pinyon pine function.

Because drought involves both increased temperatures and lower overall water availability, it can be difficult to determine the relative contribution of each of these drivers when evaluating changes in plant chemistry. Temperature has a well-established exponential relationship with monoterpene emission rates, which increase 2-3 fold for every 10 ºC increase in temperature

(Lerdau 1994, 1997), and this relationship is used to construct global atmospheric models

(Guenther 1991, 1993). However, water availability is also a significant determinant of monoterpene emissions in the semi-arid pinyon pine habitat as water status can strongly constrain both primary and secondary metabolic processes as well as stomatal aperature. In this study, we did not observe a significant difference in total monoterpene emission rates between sites despite a 1.5 ºC average temperature increase at the desert site, at which predawn water 127 potential was on average 0.3 MPa lower. Thus, our results do not suggest that an instantaneous change in pinyon pine monoterpene emissions with altered temperatures does not or cannot occur at these sites; rather it is likely that other environmental variables (i.e. water status, phenology) have a stronger control over average seasonal emission patterns.

Past studies, in addition to citing temperature as a primary driver of monoterpene emission rates from conifers, have also shown emissions to be a function of foliar monoterpene concentrations, released through a Henry’s Law relationship (Lerdau et al. 1994, 1997). In addition to our inability to find a significant temperature-emissions relationship in this study, we also failed to verify a relationship between monoterpene pool concentrations and emissions (data not shown), which has been supported by other studies (e.g. Constable et al. 1999). Total monoterpene concentrations were not related to either water potential or temperature, and different compounds followed different seasonal patterns, each maintaining a relatively stable pool size despite changes in emissions. Thus, foliar monoterpene pool sizes appear to be controlled by genetically determined synthesis rates and are less sensitive to environmental changes than emissions.

The relationship between monoterpene emissions and water potential (Figure 5.5A) can be explained by responses to stomatal conductance (emissions < 1 µg C gDW-1 h-1, including values on the ends of the curve) and phenological or seasonally-induced production of resin balls, which cause significant increases in emission rates and create the maximum on the curve. Likely due to stomatal closing in response to low water potentials, monoterpene emissions in excess of 1 µg C gDW-1 h-1 were not observed when water potential was below ca. -2 MPa, the approximate zero- assimilation point of pinyon pine (Lajtha and Barnes 1991; Breshears et al. 2009; Adams et al.

2009) below which, over time, pinyon pines become carbon starved. While monoterpene 128 emission rates were low at the end of the growing season when water potential was favorable and relatively high (less negative), this was also driven by stomatal conductance as trees were likely beginning to downregulate primary and secondary metabolism. The large, stochastic monoterpene pulse events, frequently observed during June, were potentially due to the exposure of monoterpene-containing resin during periods of budburst and rapid phenological change. The limited range of environmental conditions experienced during June precluded simple environmental or ecophysiological explanations for the large pulses, which were not related by stomatal conductance or any other measured environmental variable. When pure resin is exposed to the atmosphere, as occurs with the formation of resin balls, emission rates may be orders of magnitude higher than levels emitted from conifer needles. The timing and mechanism for why conifers seemingly spontaneously release resin is yet to be determined; however, we suspect that phenology, particularly bud burst, results in altered hydrological forcings which may make the tree more likely to secrete resin from its buds early in the season. Together, these findings suggest that monoterpene emissions are controlled deterministically by stomatal opening, but also stochastically, which may be explained either through periods of rapid phenological change and/or the presence of exposed resin to the atmosphere, particularly in June.

Our study represents an initial step to add secondary chemistry to a comprehensive understanding of pinyon pine response to drought and temperature stress with a strong emphasis on ecophysiolgical responses. In addition to describing the physiological controls over emissions via changes in water status, we find that more severe drought conditions (Ψ< -2 MPa) result in lower overall total monoterpene emission rates but relatively high maintenance of foliar concentrations. While we predicted a relatively high carbon investment toward monoterpenes under drought conditions, we also expected subsequent increases in emission rates due to the 129 positive correlation shown in other studies, but this was not the case. These findings suggest that plant ecophysiological controls play a larger role in determining plant volatile defense compounds than initially considered, offering insight into the feedbacks that exist between environmental variability, plant biophysical responses to stress, and plant defenses.

Initially, high levels of monoterpenes in pinyon pine resin may confer protection against some species of herbivores via direct toxicity and altered immunity but also a lower potential for indirect defense resulting from decreases in overall volatile signaling for effective parasitoid host location. However, over time and under more severe and consistent drought stress (Ψ < -3 MPa observed in our single season study), we may expect to see a shift in monoterpene emissions and concentrations as pinyon pines become more carbon limited, which may increase susceptibility of these pines to a suite of insect pests. Because the quantity and quality of monoterpenes can be differentially detrimental to herbivores depending on species, future studies should address how altered chemistry under initial and prolonged drought stress affects growth and fitness of the most damaging herbivores in this forest ecosystem (e.g. bark beetles, tip moths, etc.). Future studies should address how both foliar chemistry and emissions impact the natural enemies of the herbivores to understand how altered immunocompetency and host location cues affect herbivore population dynamics from both a bottom-up and top-down perspective. Furthermore, our results suggest that plant water status plays an important role in monoterpene emissions and should be incorporated into regional and global models of monoterpene efflux. Notably, research should focus on the underlying phenological mechanisms that are responsible for large bursts of monoterpene emissions (e.g. budburst, production of resin exposed to atmosphere, etc.) to improve our inventories of emissions under predicted global change.

130

CHAPTER VI

CONCLUSIONS

The objective of this conclusions chapter is to highlight the novelty of my research and discuss how my experimental findings can advance the field of plant-insect interactions by integrating chemical ecology and plant ecophysiology. In addition, I will discuss the importance of this work toward understanding the complex interactions affecting forest ecosystems in the context of future climate change.

Of the new findings presented here, four emerge as starting points for future research initiatives. All build from a strong basis in the literature, have been advanced by the experimental findings discussed in previous chapters, and are grounded in biological principles.

These findings include: i) the role of recent photoassimilate in isoprene synthesis under elevated

CO2; ii) the interactive effects of drought and herbivory on monoterpene emissions; iii) the trade- off between caterpillar growth and defense when presented with diets differing in monoterpene amount and proportion; and iv) stomatal versus stochastic controls on monoterpene emissions in conifers. Elements of each can be combined to develop experiments to better understand the impacts of future climate change on the chemical ecology of natural systems.

13 Coupling CO2-labeling with PTR-MS technology, we revealed a higher relative contribution from older/stored carbon sources toward isoprene synthesis under pre-industrial

CO2 conditions, and the incorporation of more recently assimilated carbon under elevated CO2 conditions. I framed our results in the context of plant physiological responses and carbon metabolism, and offered a mechanistic explanation for patterns of labeling kinetics observed under these varying environmental conditions. While isoprene emissions are strongly controlled 131 by instantaneous light and temperature controls, and these terms feature prominently in global and regional emission models, my work has demonstrated that for monoterpenes (particularly for species that store monoterpenes), these instantaneous controls may be overwhelmed by other environmental or biotic factors that occur over longer time scales, including drought and herbivory. Models should be updated to include hydrological and biological processes, especially for semi-arid ecosystems studied here.

Monoterpenes are synthesized during needle development and subsequently stored in resin ducts, and work here demonstrates that their emission is under stronger environmental control than often thought. More concrete links must be established between stored carbon resources, primary metabolism, monoterpene synthesis, and monoterpene emission to better understand the environmental effects of plant secondary chemistry. My results lay the foundation for developing hypotheses focused on the quantification of carbon contribution toward secondary metabolism under various combinations of abiotic and biotic stress, including how elevated CO2, drought, and herbivory may interact to determine monoterpene production and emission. While studies have failed to show a significant relationship between photosynthesis and monoterpene production and emissions in monoterpene storing species, recently assimilated carbon may play an elusive but important role in replenishing stored pools following significant losses via volatilization (e.g. after insect feeding) and should be investigated. Furthermore, results of such studies may improve carbon allocation theories (e.g. the growth differentiation balance hypothesis and/or the carbon nutrient balance hypothesis) that intend to link primary and secondary metabolism under varying stresses, and illustrate the proper adjustments needed for theoretical improvements. Although monoterpene synthesis is under genetic control, it is possible that ecophysiology has a significantly larger influence over production of plant defense 132 compounds – particularly those that are also volatile – than currently considered and such a finding may alter our thinking when considering controls over and evolution of secondary defense.

I found that herbivory caused high monoterpene emissions during feeding and after the drought-breaking rains in late summer, the latter suggesting an interaction between previous herbivory and latent emissions with release from drought. Monoterpene emissions are often thought to rely on needle concentrations (e.g. Lerdau et al. 1997), but my results suggest that foliar chemistry during insect feeding may instead be controlled by emission rates. Furthermore, my research has shown that this resulting foliar chemistry can cause trade-offs between growth and immunity in herbivores, resulting from energy investments toward detoxification. Therefore, if emissions impact foliar composition, and ecophysiological processes play a relatively large role in determining emission rates (e.g. via stomatal conductance), this can have important implications for both atmospheric modeling and plant-insect interactions. Detailing these under- studied (e.g. water status, herbivory, etc) ecological controls over monoterpene emissions in conifers will improve mechanistic representation in regional and global models. In addition, herbivores, such as tiger moths, may choose to oviposit on suitable host plants by using volatile chemical cues that also affect the foliar concentrations that their offspring will be exposed to.

Thus, future research should examine herbivore preference given monterpene signals released as a result of previous herbivory to determine if feedbacks exist in this system. Because climate has such a large impact on emissions, studies should also investigate how, or if, climate mediates these signaling feedbacks. Similar questions should also be addressed with the parasitoid of the southwestern tiger moth and other folivorous species to understand the effects volatiles have as host location cues but also how the resulting foliar chemistry affects the herbivore’s ability to 133 defend itself against parasitism. Approaching these complex chemically-mediated interactions from the perspective of plant ecophysiology will lead to a greater understanding of the ecological mechanisms that underlie plant-insect interactions.

In the absence of herbivory, plants respond to multiple environmental stresses. The majority of atmospheric models use simple light and temperature relationships to predict both isoprene and monoterpene emission rates (Guenther et al. 1993). However, arid ecosystems are consistently exposed to varying degrees of high temperatures and low water availability, or drought, and water status may be a primary driver of monoterpene synthesis and emissions despite short-term temperature dependence. My transplant study showed a relationship between stomatal conductance and monoterpene emissions, largely driven by pinyon pine water potentials, with the exception of large spikes in emissions seen in early summer. These stochastic bursts in emissions may be the result of resin exposure to the atmosphere via budburst or some other phenological or physiological process. These emission responses should be examined in other tree species and experiments aimed at understanding the causes of “pitching out” should be investigated. Detailing the mechanisms responsible for these seemingly random high emission rates would not only improve emission inventories but change people’s perspective on how plants emit secondary compounds.

Overall, this research contributes to the body of knowledge on the controls over plant secondary chemistry with important implications for plant-insect-natural enemy interactions and ecosystem function in the face of a changing climate.

134

BIBLIOGRAPHY

Abdelgaleil, S. A. M., M. I. E. Mohammed, M. E. I. Badawy, and S. A. A. El-Arami. 2009. Fumigant and contact toxicities of monoterpenes to Sitophilus oryzae (L.) and Tribolium castaneum (Herbst) and their inhibitory effects on acetylcholinesterase activity. Journal of Chemical Ecology 35:518-525.

Adams, R., P., 1989. Identification of essential oils by ion trap mass spectrometry. Academic Press, San Diego. Adams, H. D., M. Guardiola-Claramonte, G. A. Barron-Gafford, J. C. Villegas, D. D. Breshears, C. B. Zou, P. A. Trock, and T. E. Huxman. 2009. Temperature sensitivity of drought-induced tree mortality portends increased regional die-off under global- change-type drought. Proceedings of the National Academy of Sciences 106:7063- 7066. Alborn, H. T., T. C. J. Turlings, T. H. Jones, G. Stenhagen, J. H. Loughrin, and J. H. Tumlinson. 1997. An elicitor of plant volatiles from beet armyworm oral secretion. Science 276:945-949. Alfaro, R. E., H. D. Pierce Jr., H. H. Borden, and A. C. Oehlschlager. 1980. Role of volatile and nonvolatile components of Sitka spruce bark as feeding stimulants for Pissodes strobe Peck (Coleoptera: Curculionidae). Canadian Journal of Zoology 58:626-632. Allen, C. D., and D. D. Breshears. 1998. Drought-induced shift of a forest-woodland ecotone: Rapid landscape response to climate variation. Proceedings of the National Academy of Sciences 95:14839-14842. Appel, H.M, and M. M. Martin. 1992. Significance of metabolic load in the evolution of host specificity in Manduca sexta. Ecology 73:216-228. Awmack, C. S., and S. R. Leather. 2002. Host plant quality and fecundity in herbivorous insects. Annual Review of Entomology 47:817-844. Banthorpe, D., and B. Charlwood. 1980. The terpenoids. In Encyclopedia of Plant Physiology, Vol. 12. E. Bell and B. Charlwood (eds.) Springer-Verlag, Berlin, 185- 219. Barre, F., F. Goussard, and C. Géri. 2003. Variation in the suitability of Pinus sylvestris to feeding by two defoliators, Diprion pini (Hym., Diprionidae) and Graellsia isabellae galliaegloria (Lep., Attacidae). Journal of Applied Entomology 127:249-257. Beckage, N. E. 2008. Insect Immunology. Academic Press, Oxford. Behnke K, Ehlting B, Teuber M, Bauerfeind M, Louis S, et al. .2007. Transgenic, non- isoprene emitting poplars don’t like it hot. The Plant Journal 51: 485–499. Bennett R.N. and R.M. Wallsgrove. 1994. Secondary metabolites in plant defense mechanisms. New Phytologist 127: 617-633. 135

Berenbaum, M. 1983. Coumarins and caterpillars: a case for coevolution. Evolution 37:163- 179. Bernays, E. A.and R. F. Chapman, 1994. Host Plant Selection by Phytophagous Insects. Chapman and Hall, New York. Berry, R.E., Yu, S.J., Terriere, L.C. 1980. Influence of host plants on insecticide metabolism and management of variegated cutworm. Journal of Economic Entomology 73: 771-774. Berryman, A. A. 1996. What causes population cycles of forest Lepidoptera? Trends in Ecology and Evolution 11:28-32. Bertin, N. and M. Staudt. 1996. Effect of water stress on monoterpene emissions from young potted holm oak (Quercus ilex L.) trees. Oecologia 107: 456-462. Bertin, N., M. Staudt, U. Hansen, G. Seufert, P. Ciccioli, P. Foster, J.L. Fugit, and L. Torres. 1997. Diurnal and seasonal course of monoterpene emissions from Quercus ilex (L.) under natural conditions application of light and temperature algorithms. Atmospheric Environmen 31: 135-144. Binder, B. F. & J. C. Robbins. 1997. Effect of terpenoids and related compounds on the oviposition behavior of the European corn borer Ostrinia nubilalis (Lepidoptera: Pyralidae). Journal of Agricultural and Food Chemistry 45:980-984. Blumberg, D. 1997. Parasitoid encapsulation as a defense mechanism in the Coccoidea (Homoptera) and its importance in biological control. Biological Control 8:225-236. Bohlmann J., and R. Croteau. 1999. Diversity and variability of terpenoid defenses in conifers: molecular genetics, biochemistry and evolution of the terpene synthase gene family in grand fir (Abies grandis). In D. J. Chadwick, and J. A. Goode, eds, Insect Plant Interactions and Induced Plant Defense. John Wiley and Sons Ltd., West Sussex, UK, pp 132–146 Bowers, M. D., and G. M. Puttick. 1988. Response of generalist and specialist insects to qualitative allelochemical variation. Journal of Chemical Ecology 14:319-334. Breshears, D.D., O.B. Myers, S.R. Johnson, C.W. Meyer, and S.N. Martens. 1997. Differential use of spatially heterogeneous soil moisture by two semiarid woody species: Pinus edulis and Juniperus ,onosperma. Journal of Ecology 85: 289-299 Breshears, D. D., N. S. Cobb, P. M. Rich, K. P. Price, C. D. Allen, R. G. Balice, W. H. Romme, J. H. Kastens, M. L. Floyd, J. Belnap, J. J. Anderson, O. B. Myers, and C. W. Meyer. 2005. Regional vegetation die-off in response to global-change-type drought. Proceedings of the National Academy of Sciences 102:15144-15148. Breshears, D.D., O.B. Myers., C.W. Meyer, F.J. Barnes , C.B. Zou, C.D. Allen, N.G. McDowell, and W.T. Pockman. 2009. Tree die-off in response to global change-type drought: Mortality insights from a decade of plant water potential measurements. Forntiers in Ecology and the Environment 7: 185-189. 136

Brilli F, Barta C, Fortunati A, Lerdau M, Loreto F, et al. 2007. Response of isoprene emission and carbon metabolism to drought in white poplar (Populus alba) saplings. New Phytologist 175: 244–254. Brown, J.H., T.G. Whitham, S.K. Morgan Ernest, and C.A. Gehring. 2001. Complex species interactions and the dynamics of ecological systems: Long-term experiments. Science 293: 643-650. Bryant, J.P., F.S. Chapin, and D.R. Klein. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate herbivory. Oikos 40: 357-368 Bukovinszky, T., E.H. Poelman, R. Gols, G. Prekatsakis, L.E. Vet, J.A. Harvey, and M. Dicke. 2009.Consequences of constitutive and induced variation in plant nutritional quality for immune defence of a herbivore against parasitism. Oecologia 160: 299–308. Bulet, P., C. Hetru, J. Dimarcq, and D. Hoffmann. 1999. Antimicrobial peptides in insects; structure and function. Developmental and Comparative Immunology 23:329-344. Byun-McKay, A., K.-A. Godard, M. Toudefallah, D. M. Martin, R. Alfaro, J. King, J. Bohlmann, and A. L. Plant. 2006. Wound-induced terpene synthase gene expression in Sitka Spruce that exhibit resistance or susceptibility to attack by the White Pine Weevil. Plant Physiology 140:1009-1021. Carroll, M.J., Schmelz, E.A., Meagher, R.L. & Teal, P.E.A. 2006. Attraction of Spodoptera frugiperda larvae to volatiles from herbivore-damaged maize seedlings. J. Chem. Ecol. 32, 1911–1924. Cates, R. G. 1980. Feeding patterns of monophagous, oligophagous, and polyphagous insect herbivores: The effect of resource abundance and plant chemistry. Oecologia 46:22- 31. Cegelski L, Schaefer J. 2006. NMR determination of photorespiration in intact leaves using 13 in vivo CO2 labeling. Journal of Magnetic Resonance 178: 1–10. Cerenius, L., and K. Söderhäll. 2004. The prophenoloxidase-activating system in invertebrates. Immunological Reviews 198:116-126. Chen, J.M, P.M. Rich, S.T. Gower, J.M. Norman, and S. Plummer. 1997. Leaf area index of boreal forests: Theory, techniques, and measurements. Journal of Geophysical Research 102: 29429-29443. Clancy, K. M., Z. Chen, and T. E. Kolb. 2004. Foliar nutrients and induced susceptibility: genetic mechanisms of Douglas-fir resistance to western spruce budworm defoliation. Canadian Journal of Forest Research 34:939-949. Cobb, N. S., S. Mopper, C. A. Gehring, M. Caouette, K. M. Christensen, and T. G. Whitham. 1997. Increased moth herbivory associated with environmental stress of pinyon pine at local and regional levels. Oecologia 109: 389-397. 137

Codella, S. G., and K. F. Raffa. 1995 Contributions of female oviposition patterns and larval behavior to group defense in conifer sawflies (Hymenoptera: Diprionidae). Oecologia 103:24-33. Coley, P. D. 1998. The effects of climate change on plant-herbivore interactions in moist tropical rainforests. Climate Change 39:455-472. Constable, J.V.H., A.B. Guenther, D.S. Schimel, and R.K. Monson. 1999. Modelling changes in VOC emission in response to climate change in the continental United States. Global Change Biology 5: 791-806. Croteau, R. 1987. Biosynthesis and catabolism of monoterpenes.Chemical Reviews 87: 929-954. Delwiche CF, and T.D. Sharkey. 1993. Rapid appearance of 13C in biogenic isoprene when 13 (CO2)- C is fed to intact leaves. Plant, Cell and Environment 16: 587–591. DeMoraes C.M., Lewis W.J., Pare P.W., Alborn H.T., Tumlinson J.H. 1998. Herbivore-infested plants selectively attract parasitoids. Nature 393:570-573. DeMoraes, C. M., M. C. Mescher, and J. H. Tumlinson. 2001. Caterpillar-induced nocturnal plant volatiles repel conspecific females. Nature 410:577-580. Dicke M., Sabelis M.W., Takabayashi J., Bruin J., Posthumus M.A. 1990. Plant strategies of manipulating predator-prey interactions through allelochemicals: Prospects for application in pest control. Journal of Chemical Ecology 16: 3091-3118. Duncan, R.W. 2006. Conifer Defoliators of British Columbia. Victoria, BC: Natural Resources Canada, Canadian Forest Service, and Pacific Forestry Centre. Dwyer, G., J. Dushoff, and S. H. Yee. 2004. The combined effects of pathogens and predators on insect outbreaks. Nature 430:341-345. Dyer, L.A., C. D. Dodson, J. O. Stireman, M. A. Tobler, A. M. Smilanich, R. M. Fincher and D. K. Letourneau. 2003. Synergistic effects of three Piper amides on generalist and specialist herbivores. Journal of Chemical Ecology 29: 2499-2519. Dyer, L.A., T. R. Walla, H. F. Greeney, J. O Stireman III, and R. F. Hazen. 2010. Diversity of interactions: A metric for studies of biodiversity. Biotropica 42:281-289. Fehsenfeld F.C., J. Clavert, R. Fall, P. Goldan, A.B. Guenther, et al. 1992. Emissions of volatile organic compounds from vegetation and the implications for atmospheric chemistry. Global Biogeochemical Cycles 6: 389–430. Fischer K., B. Kammerer, M. Gutensohn, B. Arbinger, A. Weber, et al. 1997. A new class of plastidic phosphate translocators: A putative link between primary and secondary metabolism by the phosphoenolpyruvate/phosphate antiporter. Plant Cell 9: 453–462. Floyd, M. L., M. Clifford, N. S. Cobb, D. Hanna, R. Delph, P. Ford, and D. Turner. 2009. Relationship of stand characteristics to drought-induced mortality in three Southwestern piñon-juniper woodlands. Ecological Applications 19:1223-1230. 138

Flugge U.I. 1999. Phosphate translocators in plastids. Annual Review of Plant Physiology and Plant Molecular Biology 50: 27–45. Fuentes J.D., M. Lerdau, R. Atkinson, D.D. Baldocchi, J.W. Bottenheim, et al. 2000. Biogenic hydrocarbons in the atmospheric boundary layer: A review. Bulletin of the American Meteorological Society 81: 1537–1575. Funk J.L., J.E. Mak, and M.T. Lerdau. 2004. Stress-induced changes in carbon sources for isoprene production in Populus deltoides. Plant, Cell & Environment 27: 747–755. Gershenzon, J. 1994. Metabolic costs of terpenoid accumulation in higher plants. Journal of Chemical Ecology 20:1281-1382. Gershenzon, J. and N. Dudareva. 2007. The function of terpene natural products in the natural world. Nature Chemical Biology. 3: 408-414. Ghirardo A, K. Koch, R. Taipale, I. Zimmer, J.P. Schnitzler, et al. 2010. Determination of de 13 novo and pool emissions of terpenes from four common boreal/alpine trees by CO2 labeling and PTR-MS analysis. Plant, Cell and Environment 33: 781–792. Godfray H.C.J. (1993) Parasitoids: Behavioral and Evolutionary Ecology. Princeton University Press, Princeton, NJ. 473 pages. Gols, R., C. E. Raaijmakers, N. M. van Dam, M. Dicke, T. Bukovinszky, and J. A. Harvey. 2007. Temporal changes affect plant chemistry and tritrophic interactions. Basic and Applied Ecology 8:421-433. Gouinguene, S.P. and T.C.J. Turlings. 2002. The effects of abiotic factors on induced volatile emissions in corn plants. Plant Physiology 129: 1296-1307. Grant, G.G., J. Guo, L. MacDonald, and M. D. Coppens. 2007. Oviposition response of spruce budworm (Lepidoptera: Tortricidae) to host terpenes and green-leaf volatiles. The Canadian Entomologist 139:564-575. Guenther A.B., R.K. Monson, and R. Fall. 1991. Isoprene and monoterpene emission rate variability: Observations with eucalyptus and emission rate algorithm development. Journal of Geophysical Research 96: 10,799–10,808. Guenther A.B., P.R. Zimmerman, P.C. Harley, R.K. Monson, and R. Fall. 1993. Isoprene and monoterpene emission rate variability: Model evaluations and sensitivity analyses. Journal of Geophysical Research 98: 12,609–12,617. Guenther, A.B., C. N. Hewitt, D. Erickson, R. Fall, C. Geron, T. Graedel, P. Harley, L. Klinger, M. Lerdau, W. McKay, T. Pierce, B. Scholes, R. Steinbrecher, R. Tallamraju, J. Taylor, and P. Zimmerman. 1995. A global model of natural volatile organic compound emissions, Journal of Geophysical Research 100:8873-8892. Guenther, A.B., C. Geron, T. Pierce, B. Lamb, P. Harley, and R. Fall. 2000. Natural emissions of non-methane volatile organic compounds, carbon monoxide, and oxides of nitrogen from North America. Atmospheric Environment 34:2205-2230. 139

Guenther A.B., T. Karl, P. Harley, C. Wiedinmyer, P.I. Palmer, et al. 2006. Estimates of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and Aerosols from Nature). Atmospheric Chemistry and Physics Discussions 6: 107– 173. Gutbrodt, B., K. Mody, and S. Dorn. 2011. Drought changes plant chemistry and causes contrasting responses in lepidopteran herbivores. Oikos 120: 1732-1740. Hakola, H., V. Tarvainen, J. Bäck, H. Ranta, B. Bonn, J. Rinne, and M. Kulmala. 2006. Seasonal variation of mono- and sesquiterpene emission rates of Scots pine. Biogeosciences 3:93-101. Hanson, P. J., and J. F. Weltzin. 2000. Drought disturbance from climate change: response of United States forests. Science of the Total Environment 262:205-220. Hanula, J. L., G. L. DeBarr, and C. W. Berisford. 1985. Adult activity of Dioryctria amatella (Lepidoptera: Pyralidae) in relation to development of immature stages in loblolly pine cones. Environmental Entomology 14:842-845. Harley P., Fridd-Stroud V., Greenburg J., Guenther A., Vasconcellos P. 1998. Emission of 2- methyl-3-buten-2-ol by pines: A potentially large natural source of reactive carbon to the atmosphere. Journal of Geophysical Research 103: 25479-25486. Harley P.C., R.K. Monson, M.T. Lerdau. 1999. Ecological and evolutionary aspects of isoprene emission from plants. Oecologia 118: 109–123. Harvey, J.A., N.M. Van Dam, L.M.A Witjes, R. Soler, and R. Gols. 2007. Effects of dietary nicotine on the development of an insect herbivore, its parasitoid and secondary hyperparasitoid over four trophic levels. Ecological Entomology 32:15-23. Harvey, J.A, T. Bukovinszky, and W. H. van der Putten. 2010. Interactions between invasive plants and insect herbivores: A plea for a multitrophic perspective. Biological Conservation 143:2251-2259. Hasunuma T., K. Harada, S. Miyazawa, A. Kondo, E. Fukusaki, et al. 2010. Metabolic 13 turnover analysis by a combination of in vivo 13C-labelling from CO2 and metabolic profiling with CE-MS/MS reveals rate-limiting steps of the C3 photosynthetic pathway in Nicotiana tabacum leaves. Journal of Experimental Botany 61: 1041–1051. Haukioja, E. 1991. Induction of defenses in trees. Annual Review of Entomology 36:25-42. Havill, N. P., and K. F. Raffa. 2000. Compound effects of induced plant responses on insect herbivores and parasitoids: implications for tritrophic interactions. Ecological Entomology 25:171-179. Hay, M.E., Q.E. Kappel, and W. Fenical. 1994. Synergisms in plant defenses against herbivores: Interactions of chemistry, calcification, and plant quality. Ecology 75: 1714-1726. 140

Heald C.L., M.J. Wilkinson, R.K. Monson, C.A. Alo, G. Wang, et al. 2009. Response of isoprene emission to ambient CO2 changes and implications for global budgets. Global Change Biology 15: 1127–1140. Helmig, D., F. Bocquet, J. Pollmann, and T. Revermann. 2004. Analytical techniques for sesquiterpene emission rate studies in vegetation enclosure experiments. Atmospheric Environment 38: 557-572. Heyworth, C.J., G. R. Iason, V. Temperton, P. G. Jarvis, and A. J. Duncan. 1998. The effect of elevated CO2 concentration and nutrient supply on carbon based plant secondary metabolites in Pinus sylvestris L. Oecologia 115:344-350. Hick A.J., Luszniak M.C., Pickett J.A. 1999. Volatile isoprenoids that control insectbehavior and development. Natural Product Reports 16: 39-54. Hilker M., Kobs C., Varama M., Schrank K. 2002. Insect egg deposition induces Pinus sylvestris to attract egg parasitoids. Experimental Biology 205: 455-461. Hódar, J.A., R. Zamora, J. Castro, and E. Baraza. 2004. Feast and famine: previous defoliation limiting survival of pine processionary caterpillar Thaumetopoea pityocampa in Scots pine Pinus sylvestris. Acta Oecologia 26: 203-210. Holopainen, J.K. 2004. Multiple functions of inducible plant volatiles. Trends in Plant Science 9: 529-533. Honda, K. 1995. Chemical basis of differential oviposition by lepidopteran insects. Insect Biochemistry and Physiology 20: 1-23. Holzke, C., T. Hoffmann, L. Jaeger, R. Koppmann, and W. Zimmer. 2006. Diurnal and seasonal variation of monoterpene and sesquiterpene emissions from Scots pine (Pinus sylvestris L.). Atmospheric Environment 40:374-3185. Huber, D.P.W, S. Ralph, and J. Bohlmann. 2004. Genomic hardwiring and phenotypic plasticity of terpenoid-based defenses in conifers. Journal of Chemical Ecology 30: 2399-2418. Huberty, A. F., and R. F. Denno. 2004. Plant water stress and its consequences for herbivorous insects: a new synthesis. Ecology 85:1383-1398. Hummelbrunner, L. A., and M. B. Isman. 2001. Acute, sublethal, antifeedant, and synergistic effects of monoterpenoid essential oil compounds on the tobacco cutworm, Spodoptera litura (Lep., Noctuidae). Journal of Agricultural and Food Chemistry. 49:715-720. IPCC Climate Change 2007: Impacts, Adaptation and Vulnerability, eds. Parry, M. L., O.F. Canziani, J.P. Palutikof, P.J. van der Linden, and C.E. Hanson. 2007. Jacob, D.J. and S.C. Wofsy. 1988. Budgets of reactive nitrogen, hydrocarbons, and ozone over the amazon forest during the wet season. Journal of Geophysical Research 95: 16,737-16,754. 141

Jactel, H., M. Kleinhentz, A. Marpeau-Bezard, F. Marion-Poll, P. Menassieu, and C. Burban. 1996. Terpene variations in maritime pine constitutive oleoresin related to host tree selection by Dioryctria sylvestrella Ratz. Journal of Chemical Ecology 22:1037– 1050. Jermy, T. 1983. Evolution of insect/host plant relationships. The American Naturalist 124:609-630. Johnson, M.A. and R. Croteau. 1987. Biochemistry of conifer resistance to bark beetles and their fungal symbionts in Ecology and Metabolism of Plant Lipids vol 325 pp. 76-92. Kant, M. R., K. Ament, M. W. Sabelis, M. A. Haring, and R. C. Schuurink. 2004. Differential timing of spider mite-induced direct and indirect defenses in tomato plants. Plant Physiology 135:483-495. Kapari, L., E. Haukioja, M.J. Rantala, and T. Ruuhola. 2006 Defoliating insect immune defense interacts with induced plant defense during a populations outbreak. Ecology 87: 291-296. Karban, R. and J.H. Myers. 1989. Induced plant responses to herbivory. Annual Review of Ecology and Systematics 20: 331-348. Karban, R. 2010. The ecology and evolution of induced resistance against herbivores. Functional Ecology 25: 339-347. Karl T., R. Fall, T.N. Rosenstiel, P. Prazeller, B. Larsen, et al. 2002. On-line analysis of the 13 CO2 labeling of leaf isoprene suggests multiple subcellular origins of isoprene precursors. Planta 215: 894–905. Keefover-Ring, K. and Y.B. Linhart. 2010. Variable chemistry and herbivory of ponderosa pine cones. International Journal of Plant Sciences 171: 293-302. Keeling, C. I., and J. Bohlmann. 2006. Genes, enzymes and chemicals of terpenoid diversity in the constitutive and induced defense of conifers against insects and pathogens. New Phytologist 170:657-675. Kessler, A., and I. T. Baldwin. 2001. Defensive function of herbivore-induced plant volatile emissions in nature. Science 291: 2141-2144. Kraaijeveld, A.R., K.A. Hutcheson, E.C. Limentani, H.C.J. Godfray. 2001. Costs of counterdefenses to host resistance in a parasitoid of drosphila. Evolution 55: 1815-1821. Kuokkanen, K., R. Julkunen-Tiitto, M. Keinänen, P. Nimelä, and J. Tahvanainen. 2001. The effect of elevated CO2 and temperature on the secondary chemistry of Betula pendula seedlings. Trees – Structure and Function 15:378-384. Kuzma, J. and R. Fall. 1993. Leaf isoprene emission rate is dependent on leaf development and the level of isoprene synthase. Plant Physiology 101:435-440. 142

Lajtha, K. and F.J. Barnes. 1991. Carbon gain and water use in pinyon pine-juniper woodlands of northern New Mexico: field versus phytotron chamber measurements. Tree Physiology 9: 59-67. Langenheim, J.H. 1994. Higher plant terpenoids: A phytocentric overview of their ecological roles. Journal of Chemical Ecology 20: 1223-1280. Lanner, R.M and T.R.Van Devender. 1998. The recent history of pinyon pines in the American Southwest In Ecology and Biogeography of Pinus, ed. D.M. Richardson. Cambridge University Press pp. 171-181. Larsson, S., B. Ekbom, and C. Björkman. 2000. Influence of plant quality on pine sawfly population dynamics. Oikos 89:440-450. Lavine, M. D., and M. R. Strand. 2002. Insect hemocytes and their role in immunity. Insect Biochemistry and Molecular Biology. 32:1295-1309. Lavoir, A. V., M. Staudt, J. P. Schnitzler, D. Landais, F. Massol, A. Rocheteau, R. Rodriguez, I. Zimmer, and S. Rambal. 2009. Drought reduced monoterpene emissions from the evergreen Mediterranean oak Quercus ilex: results from a throughfall displacement experiment. Biogeosciences 6: 1167-1180. Leather, S. R. 1987. Pine monoterpenes stimulate oviposition in the pine beauty moth, Panolis flammea. Entomologia Experimentalis et Applicata 43:295-303. Lerdau, M., M. Litvak, and R.K. Monson. 1994. Plant chemical defense: Monoterpenes and the growth-differentiation balance hypothesis. Tends in Ecology and Evolution 9: 58-61. Lerdau, M., P. Matson, R. Fall, and R. Monson. 1995. Ecological controls over monoterpene emissions from Douglas-Fir (Pseudotsuga menziesii). Ecology 76:2640-2647. Lerdau, M., M. Litvak, P. Palmer, and R. K. Monson. 1997. Controls over monoterpene emissions from boreal forest conifers. Tree Physiology 17:563-569. Lewinsohn, E., M. Gijzen, and R. Croteau. 1991. Defense mechanisms of conifers. Plant Physiology 96: 44-49. Liao, J., Henze, D., Seinfeld, J. H., Wu, S., Mickley, L. J., 2007. Biogenic secondary organic aerosol over the United States: Comparison of climatological simulations with observations. Journal of Geophysical Research 112, D06201. Lichtenthaler, H.K. 1999. The 1-deoxy-D-xylulose-5-phosphate pathway of isoprenoid biosynthesis in plants. Annual Review of Plant Physiology and Plant Molecular Biology 50: 47–65. Liebhold, A., and N. Kamata. 2000. Introduction: Population dynamics of forest-defoliating insects. Are population cycles and spatial synchrony a universal characteristic of forest insect populations? Population Ecology 42:205-209. 143

Lindinger W., A. Hansel, and A. Jordan. 1998. Proton-transfer reaction mass spectrometry (PTR-MS): on-line monitoring of volatile organic compounds at pptv levels. Chemical Society Reviews 27: 347–354. Linhart, Y.B. and M.C. Grant. 1991. Evolutionary significance of local genetic differentiation in plants. Annual Review of Ecological Systematics 27: 237-277. Linton, M.J.,J.S. Sperry, and D.G. Williams. 1998. Limits to water transport inJuniperus osteosperma and Pinus edulis: implications for drought tolerance and regulation of transpiration. Functional Ecology12: 906–911. Litvak, M. E., and R. K. Monson. 1998. Patterns of induced and constitutive monoterpene production in conifer needles in relation to insect herbivory. Oecologia 114:531-540. Litvak, M.E., S. Madronich, and R.K. Monson. 1999. Herbivore-induced monoterpene emissions from coniferous forests: Potential impact on local tropospheric chemistry. Ecological Applications 9: 1147-1159. Llusià, J. and Peñuelas, J., 1998. Changes in terpene content and emission in potted Mediterranean woody plants under severe drought. Canadian Journal of Botany 76:1366- 1373. Llusià, J., and J. Peñuelas. 2000. Seasonal patterns of terpene content and emission from seven Mediterranean woody species in field conditions. American Journal of Botany 87:133-140. Logan B.A., R.K. Monson, and M.J. Potosnak. 2000. Biochemistry and physiology of foliar isoprene production. Trends in Plant Science 5: 477–481. Loreto F., and T.D. Sharkey. 1990. A gas-exchange study of photosynthesis and isoprene emission in Quercus rubra L. Planta 182: 523–531. Loreto F, P. Ciccioli, A. Cecinato, E. Brancaleoni, M. Frattoni, et al. 1996. Evidence of the photosynthetic origin of monoterpenes emitted by Quercus ilex L. leaves. Plant Physiology 110: 1317–1322. Loreto, F., A. Forster, M. Durr, O. Csiky, and G. Seufert. 1998. On the monoterpene emission under heat stress and on the increased thermotolerance of leaves of Quercus ilex L. fumigated with selected monoterpenes. Plant, Cell, and Environment 21: 101- 107. Loreto, F., P. Nascetti, A. Graverini, and M. Mannozzi. 2000. Emission and content of monoterpenes in intake and wounded needles of the Mediterranean Pine, Pinus pinea. Functional Ecology 14:589-595. Loreto F., M. Centritto, C. Barta, C. Calfapietra, S. Fares, et al. 2007. The relationship between isoprene emission rate and dark respiration rate in white poplar (Populus alba L.) leaves. Plant, Cell and Environment 30: 662–669. 144

Lyytikainen, P. 1992. Comparison of the effects of artificial and natural defoliation on the growth of diprionid sawflies on Scots pine foliage. Journal of Applied Entomology 114:57-66. Mattiaci, L., M. Dicke, and M. A. Posthumus. 1994. Induction of parasitoid attracting synomone in Brussels sprout plants by feeding of Pieris brassicae larvae: Role of mechanical damage and herbivore elicitor. Journal of Chemical Ecology 20:1573- 1561. Mattson, W. J., and R. A. Haack. 1987. The role of drought in outbreaks of plant-eating insects. BioScience 37:110-118. McDowell, N., W. T. Pockman, C. D. Allen, D. D. Breshears, N. Cobb, T. Kolb, J. Plaut, J. Sperry, A. West, D. G. Williams, E. A. Yepez. 2008. Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought? New Phytologist 178:719-739. Miller, B., L. L. Madilao, S. Ralph, and J. Bohlmann. 2005. Insect-induced conifer defense. White pine weevil and methyl jasmonate induce traumatic resinosis, de novo formed volatile emissions, and accumulation of terpenoid synthase and putative octadecanoid pathway transcripts in Sitka Spruce. Plant Physiology 137:369-382. Mithofer, A., G. Wanner, and W. Boland. 2005. Effects of feeding Spodoptera littoralis on lima bean leaves. II. Continuous mechanical wounding resembling insect feeding is sufficient to elicit herbivory-related volatile emission. Plant Physiology 137:1160- 1168. Monson R.K., and R. Fall. 1989. Isoprene emission from aspen leaves: Influence of environment and relation to photosynthesis and photorespiration. Plant Physiology 90: 267–274. Monson R.Kl, Trahan N., Rosenstiel T.N., Veres P., Moore D.J.P., et al. 2007. Isoprene emission from terrestrial ecosystems in response to global change: Minding the gap between models and observations. Philosophical Transactions of the Royal Society of London 365: 1677–1695. Mopper, S., J.B. Mitton, T.G. Whitham, N.S. Cobb, and K.M. Christensen. 1991. Genetic differentiation and heterozygosity in pinyon pine associated with resistance to herbivory and environmental stress. Evolution 45: 989-999. Mueller, R. C. C. M. Scudder, M. E. Porter, R. T. Trotter III, C. A. Gehring, T. G. Whitham. 2005. Differential tree mortality in response to severe drought: Evidence for long- term vegetation shifts. Journal of Ecology 93:1085-1093. Mumm, R., and M. Hilker. 2006. Direct and indirect chemical defence of pine against folivorous insects. Trends in Plant Science 11:351-358. Nabity, P. D., J. A. Zavala, and E. H. DeLucia. 2009. Indirect suppression of photosynthesis on individual leaves by arthropod herbivory. Annals of Botany 103:655-663. 145

Nelson, A.C. and T.A. Kursar. 1999. Interactions among plant defense compounds: a method for analysis. Chemoecology 9: 81-92. Nerg, A., P. Kainulainen, M. Vuorinen, M. Hanso, J. K. Holopainen, and T. Kurkela. 1994. Seasonal and geographical variation of terpenes, resin acids and total phenolics in nursery grown seedlings of Scots pine (Pinus sylvestris L.). New Phytologist 128:703-713. Niemela, P., J. Tuomi, and T. Lojander. 1991. Defoliation of the Scots pine and performance of diprionid sawflies. Journal of Animal Ecology 60:683-692. Niinemets, U. 2009. Mild versus severe stress and BVOCs: thresholds, priming and consequences. Trends in Plant Science 15: 145-153. Nykänen, H., and J. Koricheva. 2004. Damage-induced changes in woody plants and their effects on insect herbivore performance: a meta-analysis. Oikos 104:247-268. Nybakken, L., K. Klanderud, and O. Totland. 2008. Simulated environmental change has contrasting effect on defensive compound concentration in three alpine plant species. Arctic, Antarctic, and Alpine Research 40:709-715. Ode, P. J. 2006. Plant chemistry and natural enemy fitness: Effects on herbivore and natural enemy interactions. Annual Review of Entomology 51:163-185. Ortega, J., and D. Helmig. 2008. Approaches for quantifying reactive and low-volatility biogenic organic compound emissions by vegetation enclosure techniques—Part A. Chemosphere 72:343-364. Owen, S. M., and J. Peñuelas. 2005. Opportunistic emissions of volatile isoprenoids. Trends in Plant Science 10:420-426. Owens, M. K., C.-D. Lin, C. A. Taylor Jr., and S. G. Whisenant. 1998. Seasonal patterns of plant flammability and monoterpenoid content in Juniperus ashei. Journal of Chemical Ecology 24:2115:2129. Paine, T.D., and C. C. Hanlon. 1994. Influence of oleoresin constituents from Pinus ponderosa and Pinus jeffreyi on growth of mycangial fungi from Dendroctonus ponderosae and Dendroctonus jeffreyi. Journal of Chemical Ecology 20:2551-2563. Pataki D. E., R. Oren, and N. Phillips. 1998. Responses of sap flux and stomatal conductance of Pinus taeda L. trees to stepwise reductions in leaf area. Journal of Experimental Botany 49:871-878. Pasquier-Barre, F., C. Palasse, F. Goussard, M.-A. Auger-Rozenberg, and C. Géri. 2001. Relationship of Scots pine clone characteristics and water stress to hatching and larval performance of the sawfly Diprion pini (Hymenoptera: Diprionidae). Environmental Entomology 30:1-6. Peñuelas, J. and M. Estiarte. 1998. Can elevated CO2 affect secondary metabolism and ecosystem function? Trends in Ecology & Evolution 13:20-24. 146

Peñuelas, J. and J. Llusià. 2002. Linking photorespiration, monoterpenes and thermotolerance in Quercus. New Phytologist 155:227-237. Peñuelas, J., Staudt, M., 2010. BVOCs and global change. Trends in Plants Science 15, 133- 144. Peterson, J.K., R.J. Horvat, and K.D. Elsey. 1994. Squash leaf glandular trichome volatiles: Identifcation and influence on behavior of female pickleworm moth. Journal of Chemical Ecology 20: 2099-2109. Petrakis, P.V., V. Roussis, D. Papadimitriou, C. Vagias, and C. Tsitsimpikou. 2005. The effect of terpenoid extracts from 15 pine species on the feeding behavioral sequence of the late instars of the pine processionary caterpillar Thaumetopoea pityocampa. Behavioural Processes 69: 303-322. Phillips, M. A., and R. B. Croteau. 1999. Resin-based defenses in conifers. Trends in Plant Science 4:184-190. Possell M., and N. Hewitt. 2011. Isoprene emissions from plants are mediated by atmospheric CO2 concentrations. Global Change Biology 17: 1595–1610. Powell, J. A., and P. A. Opler. 2009. Moths of Western North America. University of California Press, Berkeley, CA. Rasulov B, L. Copolovici, A. Laisk, and U. Niinemets. 2009. Postillumination isoprene emission: In vivo measurements of dimethylallyldiphosphate pool size and isoprene synthase kinetics in aspen leaves. Plant Physiology 149: 1609–1618. Rasulov B, K. Huve, M. Valbe, A. Laisk, and U. Niinemets. 2009. Evidence that light, carbon dioxide, and oxygen dependencies of leaf isoprene emission are driven by energy status in hybrid aspen. Plant Physiology 151: 448–460. Rantala, M. J., and D. A. Roff. 2007. Inbreeding and extreme outbreeding cause sex differences in immune defence and life history traits in Epirrita autumnata. Heredity 98:329-336. Redak, R. A., and R. G. Cates. 1984. Douglas-fir (Pseudotsuga menziesii)—spruce budworm (Choristoneura occidentalis) interactions: the effect of nutrition, chemical defenses, tissue phenology, and tree physical parameters on budworm success. Oecologia 62:61-67. Regnault-Roger, C., A. Hamraoui, M. Holeman, E. Theron, and R. Pinel. 1993. Insecticidal effect of essential oils from Mediterranean plants upon Acanthoscelides Obtectus Say (Coleoptera, Bruchidae), a pest of kidney bean (Phaseolus vulgaris L.). Journal of Chemical Ecology 19:1233-1244. Rhoades, D.F. and R.G. Cates.1976.Toward a general theory of plant antiherbivore chemistry.Recent Advances in Phytochemistry 10: 168-213. 147

Rhoades, D. F. 1983. Responses of alder and willow to attack by tent caterpillars and webworms: evidence for pheromonal sensitivity of willows. In Plant Resistance to Insects Vol 208. ACS Publications, 55-68. Rhoades, D.F. 1985. Offensive-defensive interactions between herbivores and plants: their relevance in herbivore population dynamics and ecological theory. The American Naturalist 125: 205-238. Riddle, R. R., C. A. Taylor Jr., M. M. Kothmann, and J. E. Huston. 1996. Volatile oil contents of Ashe and redberry juniper and its relationship to preference by Angora and Spanish goats. Journal of Rangeland Management 49:35-41. Rohmer M, Knani M, Simonin P, Sutter B, Sahm H (1993) Isoprenoid biosynthesis in bacteria: a novel pathway for the early steps leading to isopentenyl diphosphate. Biochemical Journal 295: 517–524. Rontein D, M. Dieuaide-Noubhani, E.J. Dufoure, P. Raymond, and D. Rolin, 2002. The metabolic architecture of plant cells. The Journal of Biological Chemistry 277: 43,948–43,960.

Rosenstiel T.N., M.J. Potosnak, K.L. Griffin, R. Fall, and R.K. Monson. 2003. Increased CO2 uncouples growth from isoprene emission in an agriforest ecosystem. Nature 421: 256–259. Rosenstiel T.N., A.L. Ebbets, W.C. Khatri, R. Fall, and R.K. Monson. 2004. Induction of poplar leaf nitrate reductase: A test of extra-chloroplastic control of isoprene emission rate. Plant Biology 6: 12–21. Runyon, J.B., M.C. Mescher, and C.M. De Moraes. 2006. Volatile chemical cues guide host location and host selection by parasitic plants. Science 313, 1964–1967. Ruohomaki, K, F.S. Chapin, E. Haukioja, S. Neuvonen, and J. Suomela. 1996. Delayed inducible resistanec in mountain birch in response to fertilization and shade. Ecology 77: 2302-2311. Sadof, C. S., and G. G. Grant. 1997. Monoterpene composition of Pinus sylvestris varieties resistant and susceptible to Dioryctria zimmermani. Journal of Chemical Ecology 23:1917:1927. Sallas, L., E.-M. Luomala, J. Utriainen, P. Kainulainen, and J. K. Holopainen. 2003. Contrasting effects of elevated carbon dioxide concentration and temperature on Rubisco activity, chlorophyll fluorescence, needle ultrastructure and secondary metabolites in conifer seedlings. Tree Physiology 23:97-108. Sanadze G.A. 1964. Light-dependent excretion of isoprene by plants. Photosynthesis Research 2: 701–707. Schmid-Hempel, P., and D. Ebert. 2003. On the evolutionary ecology of specific immune defence. Trends in Ecology and Evolution 18: 27-32. 148

Schnitzler J.P., M. Graus, J. Kreuzwieser, U. Heizmann, H. Rennenberg, et al. 2004. Contribution of different carbon sources to isoprene biosynthesis in poplar leaves. Plant Physiology 135: 152–160. Scholefield P.A., K.J. Doick, B.M.J. Herbert, C.N.S. Hewitt, J.P. Schnitzler, et al. 2004. Impact of rising CO2 on emissions of volatile organic compounds: Isoprene emission from Phragmites australis growing at elevated CO2 in a natural carbon dioxide spring. Plant, Cell and Environment 27: 393–401. Schönwitz, R., K. Lohwasser, M. Kloos, and H. Ziegler. 1990. Seasonal variation in the monoterpenes in needles of Picea abies (L.) Karst. Trees-Structure and Function. 4:34-40. Schoonhoven, L.M., T. Jermy, and J.J.A van Loon. 1998. Insect-Plant Biology: From Physiology to Evolution. London: Chapman & Hall.

Sharkey, T. D. and E. L. Singsaas. 1995. Why plants emit isoprene. Nature 374:769. Sharkey T.D. and S.S. Yeh. 2001. Isoprene emission from plants. Annual Review of Plant Physiology and Plant Molecular Biology 52: 407–436. Shu, S., G. G. Grant, D. Langevin, D. A. Lombardo, and L. MacDonald. 2004. Oviposition and electroantennogram responses of Dioryctria abietivorella (Lepidoptera: Pyralidae) elicited by monoterpenes and enantiomers from Eastern White Pine. Journal of Chemical Ecology 23:1573-1561. Siwko, M. E., S. J. Marrink, A. H. de Vries, A. Kozubek, A. J. M. Schoot Uiterkamp, and A. E. Mark. 2007. Does isoprene protect plant membranes from thermal shock? A molecular dynamics study. Biochimica et Biophysica Acta (BBA) - Biomembranes 1768:198-206. Smilanich, A. M., L. A. Dyer, J. Q. Chambers, and M. D. Bowers. 2009. Immunological cost of chemical defence and the evolution of herbivore diet breadth. Ecology Letters 12:612-621. Smith, R.H. 1975. Formula for describing effect of insect and host tree factors on resistance to western pine beetle attack. Journal of Economic Entomology 68: 841-844. Smith, R.H. 2000. Pinus edulis In Xylem monoterpenes of pines: distribution, variation, genetics, function.Forest Service, U.S. Department of Agriculture pp. 48-50. Smits, A., and S. Larsson. 1999. Effects of previous defoliation on pine looper larval performance. Agricultural and Forest Entomology 1:19-26. Staudt, M., N. Bertin, U. Hansen, G. Seufert, P. Ciccioli, P. Foster, B. Frenzel, and J.-L. Fugit. 1997. Seasonal and diurnal patterns of monoterpene emissions from Pinus pinea (L.) under field conditions. Atmospheric Environment 31:145-156. Staudt, M., N. Bertin, B. Frenzel, and G. Seufert. 2000. Seasonal variation amount and composition of monoterpenes emitted by young Pinus pinea trees – Implications for emission modeling. Journal of Atmospheric Chemstry 35:77-99. 149

Staudt, M., S. Rambal, and R. Joffre. 2002. Impact of drought on seasonal monoterpene emissions from Quercus ilex in southern France. Journal of Geophysical Research 107:4602. doi: 10.1029/2001JD002043. Staudt, M., E. Amel, M. Florent, and J. Richard. 2008. Do volatile organic compound emissions of Tunisian cork oak populations originating from contrasting climatic conditions differe in their responses to summer drought? Canadian Journal of Forest Research 38: 2965-2975. Teder, T., and T. Tammaru. 2002. Cascading effects of variation in plant vigour on the relative performance of insect herbivores and their parasitoids. Ecological Entomology 27:94-104. Thaler, J. S., M. A. Farag, P. W. Pare, and M. Dicke. 2002. Jasmonate-deficient plants have reduced direct and indirect defences against herbivores. Ecology Letters 5:764-774. Tiberi, R., A. Niccoli, M. Curini, F. Epifano, M.C. Marcotullio, and O. Rosati. 1999. The role of the monoterpene composition in Pinus spp. needles in host selection by the pine processionary caterpillar, Thaumetopoea pityocampa. Phytoparasitica 27: 263-272. Tingey, D. T., M. Manning, L. C. Grothaus, and W. F. Burns. 1980. Influence of light and temperature on monoterpene emission rates from slash pine. Plant Physiology 65:797-801. Tingey, D. T., D. P. Turner, and J. A. Weber. 1991. Factors controlling the emissions of monoterpenes and other volatile organic compounds. In Sharkey, T. D., E. A. Holland, and H. A. Mooney (eds.), Trace Gas Emissions from Plants. San Diego: Academic Press, 93-119. Trapp, S., and R. Croteau. 2001. Defensive resin biosynthesis in conifers. Annual Review of Plant Physiology and Plant Molecular Biology 52:689-724. Trewhella, K. E., S. R. Leather, and K. R. Day. 1997. Insect induced resistance to Lodgepole pine: effects on two pine feeding insects. Journal of Applied Entomology 121:129- 136. Trewhella, K. E., S. R. Leather, and K. R. Day. 2000. Variation in the suitability of Pinus contorta (lodgepole pine) to feeding by three pine defoliators, Panolis flammea, Neodiprion sertifer, and Zeiraphea diniana. Journal of Applied Entomology 124:11- 17. Tuomi, J., P. Niemelä, and S. Sirén. 1990. The Panglossian Paradigm and Delayed Inducible Accumulation of Foliar Phenolics in Mountain Birch. Oikos 59: 399-410. Turlings, T. C. J., and B. Benrey. 1998. Effects of plant metabolites on the behavior and development of parasitic wasps. Ecoscience 5:321-333. Turlings, T. C. J., J. H. Tumlinson, and W. J. Lewis. 1990. Exploitation of herbivore-induced plant odors by host-seeking parasitic wasps. Science 250:1251-1253. 150

Valterová, I., G. Nehlin, and A.-K. Borg-Karlson. 1997. Host plant chemistry and preferences in egg-laying Trioza apicalis (Homoptera, Psylloidea). Biochemical Systematics and Ecology 25:477-491. Velikova, V. and F. Loreto. 2005. On the relationship between isoprene emission and thermotolerance in Phragmites australis leaves exposed to high temperatures and during the recovery from a heat stress. Plant, Cell & Environment 28:318-327. Wagner, D.L. 2009. The immature stages: Structure, function, behavior, and ecology. In Tiger Moths and Woolly Bears. Conner,W.E. (ed.) Oxford University Press, New York, 31-54. Waldbauer, G. P. 1968. The consumption and utilization of food by insects. Advances in Insect Physiology 5:229-288. Wallin, K.F. and K.F. Raffa. 2004. Feedback between individual host selection behavior and population dynamics in an eruptive herbivore. Ecological Monographs 74: 101-116. Waring, G.L. and N.S. Cobb.1992. The impacts of plant stress on herbivore population dynamics. In:Bernays, E.A., ed.Plant-insect interactions, Vol. 4. Boca Raton, FL, USA: CRC,167–226. West, N.E. 1999. Distribution, composition, and classification of current juniper-pinyon woodland and savannas across western North America. In S.B. Monsen and R. Stevens, eds, Proceedings: ecology and management of pinyon-juniper communities within the Interior West. U.S. Department of Agriculture Forest Service Proceedings, pp 20-23. West, A. G., K. R. Hultine, K. G. Burtch, and J. R. Ehleringer. 2007. Seasonal variations in moisture use in a piñon-juniper woodland. Oecologia 153:787-798. West, A. G., K. R. Hultine, J. S. Sperry, S. E. Bush, and J. R. Ehleringer. 2008. Transpiration and hydraulic strategies in a piñon-juniper woodland. Ecological Applications 18:911-927. Wilkinson M.J., R.K. Monson, N. Trahan, S. Lee, E. Brown, et al. (2009) Leaf isoprene emission rate as a function of atmospheric CO2 concentration. Global Change Biology 15: 1189–1200. Williams, D. G., and J. R. Ehleringer. 2000. Intra- and interspecific variation for summer precipitation use in pinyon-juniper woodlands. Ecological Monographs 70:517-537. Williams, K.S. and J.H. Myers. 1984. Previous herbivore attack of red alder may improve food quality for fall webworm larvae. Oecologia 63: 166-170. Wright, G.A., Lutmerding, A., Dudareva, N. & Smith, B.H. 2005. Intensity and the ratios of compounds in the scent of snapdragon flowers affect scent discrimination by honeybees (Apis mellifera). J. Comp. Physiol. [A] 191, 105–114. 151

Yu, S.J. 1987. Microsomal oxidation of allelochemicals in generalist (Spodoptera frugiperda) and semispecialist (Anticarsia gemmatalis) insect. Journal of Chemical Ecology13:423-436. Zitani, N. M. 2003. The evolution and adaptive significance of silk use in the Meteorinae (Hymenoptera, Braconidae). Ph.D. Dissertation, University of Wyoming, August 2003, 126 pp. Zitani, N. M., and S. R. Shaw. 2002. From meteors to death stars: variations on a silk thread (Hymenoptera: Braconidae: Meteorinae). American Entomologist, 48: 228–235. Zou, J., and R. G. Cates. 1997. Effects of terpenes and phenolic and flavonoid glycosides from Douglas Fir on Western Spruce Budworm larval growth, pupal weight, and adult weight. Journal of Chemical Ecology 23:2313-2326. Zulak, K.G. and J. Bohlmann. 2010. Terpenoid biosynthesis and specialized vascular cells of conifer defense. Journal of Integrative Plant Biology 52: 86-97.