Slinn Heather 202104 Phd.Pdf

Community-level consequences of Piper chemistry on tropical herbivores, parasitoids and fungi

Heather Lea Slinn

A Thesis

presented to

The University of Guelph

In partial fulfilment of requirements for the degree of

Doctor of Philosophy

Integrative Biology

Guelph, Ontario, Canada

ABSTRACT

COMMUNITY-LEVEL CONSEQUENCES OF PIPER CHEMISTRY ON TROPICAL HERBIVORES, PARASITOIDS, AND FUNGI

Heather Lea Slinn Advisors:

University of Guelph, 2020 Dr. Jonathan Newman

Dr. Lee Dyer

Plant chemistry is a defining feature of plants as taxa exhibit extensive chemical variation, which influences interactions with other organisms, especially insect and mammalian herbivores and fungi. Plant chemistry can act as an important ecological filter that shapes which fungi can colonize host plant tissue and which insects and mammals form associations with it. The majority of research that has been done on the consequences of plant-insect, and more generally plant-herbivore, interactions has focused on isolating the effects of one or two plant compounds. However, looking at phytochemical profiles as a whole is a better reflection of nature as plants often contain compounds that are rich in structural and functional diversity. In addition, tropical ecosystems are home to extremely diverse fungal communities, which are formed by complex ecological filters. Fungi can largely influence plant chemistry and therefore plant-insect interactions; however, only a small percentage of fungi have been characterized, and scientists understand even less about their functional roles. Here, I contribute to the research addressing these knowledge gaps using Piper, which is a pantropical plant genus with a broad spectrum of phytochemistry, and which hosts

highly specialized insect and mammalian herbivores and diverse fungal communities. I start Chapter One with an introduction to the overarching concepts and theoretical framework of my three research chapters (Chapters Two-Four). In Chapter Two, I investigate how phytochemical diversity across several plant species in two tropical ecosystems vary in their effects on herbivore immune response and herbivore parasitism by parasitoids. Chapter Three begins my research into the microbial world of

Piper, where I examine how heritability of plant chemistry in Piper sancti-felicis and land-use history at the site affect herbivory, herbivore immunity, and fungal endophyte communities. Chapter Four establishes the consequences of bat digestion on the seed fungal communities of P. sancti-felicis and identifies anti-fungal properties in the fruit. I conclude with a summary and overall integration of the themes and findings of my research chapters in Chapter Five. In conclusion, I show how plant chemistry in neotropical Piper is an important determinant of species interactions and community assembly across two different kingdoms of organisms.

ACKNOWLEDGEMENTS

There are many people over the course of my degree that have assisted with my research and development as a scientist. First, I’d like to acknowledge my colleague and good friend, Bernal Matarrita-Carranza, who is a phenomenal scientist and taught me many microbial techniques while I was at La Selva Biological Station. I hope we can continue to work together in the future.

Thanks to Greg Crutsinger, Kerri Crawford, Mariano Rodriguez-Cabal and Matt Barbour who got me started in science. I wouldn’t be here if it weren’t for the time and energy that you put into my development.

To the UNR folks: Thanks to my former advisors Angela Smilanich and Matt Forister, and my committee member Paul Hurtado for your support and investment in me while I was at UNR. I hope I was a beneficial addition to the plant-insect group. Special thanks to my friends and colleagues that in some ways acted as unofficial advisors: Josh Harrison, Josh Jahner and Zach Marion. Sorry I never achieved becoming an asshole.

I’d like to thank my La Selva friends for making my Costa Rica research trips extra memorable: Jan Bechler, Leith Miller, Paul Fellmann, Elodie Moureau, Orlando Vargas Ramirez, Laura Bizzarri, Nada Nikolic, Ping and Diego (and family). I miss our time together and I hope to see you on the other side of the pandemic.

Many thanks go to the numerous funders of my research over the last five years, including: NSERC, Sonoma County Mycological Association, Mycological Association of America and to UoG and UNR. I would not be writing this document if it weren’t for your financial investment.

To my advisors: Lee, thanks for staying with me as I transitioned back to Canada and for putting up with my incessant Canadian jokes and (hockey) trash talk; Jonathan,

thanks for taking me in mid-degree without knowing me. I hope I was a valuable addition to the lab. Thanks also to my committee member Posy Busby who stayed with me across this transition. And to Sally Adamowicz for joining my advisory committee in Guelph and for providing valuable advice in your fields of barcoding and bioinformatics.

Lastly, I’d like to thank my partner Chris for his love and support, especially as I trekked around Latin America looking at plants, insects and fungi for months at a time.

1 Table of Contents Abstract ...... ii

Acknowledgements ...... iv

List of Tables ...... xi

List of Figures ...... xiv

List of Appendices ...... xviii

Chapter Publications ...... xxii

1 General Introduction ...... 1

1.0 Overview ...... 1

1.1 Tropical Ecosystem Significance ...... 2

1.2 Species’ Functional Roles and Interaction Diversity are Central to Communities ...... 3

1.3 Community Assembly and Ecological Filtering of Microbial Communities ...... 6

1.4 Objectives...... 8

1.5 Study Sites ...... 9

2 Across Multiple Species, Phytochemical Diversity and Herbivore Diet Breadth Have Cascading Effects on Herbivore Immunity and Parasitism in a Tropical Model System ...... 10

2.1 Abstract ...... 10

2.2 Introduction ...... 11

vii

2.3 Methods ...... 15

2.3.1 Piper–Eois, Piper–Quadrus System ...... 15

2.3.2 Long-term Rearing Databases ...... 16

2.3.3 Immune Assay ...... 17

2.3.4 Statistical Analyses ...... 18

2.4 Results...... 19

2.5 Discussion ...... 22

2.6 Tables ...... 27

2.7 Figures ...... 32

3 Intraspecific variation in plant chemistry and land-use history in a common garden experiment act as an ecological filter to insect herbivory and fungal endophyte communities 35

3.1 Abstract ...... 35

3.2 Introduction ...... 36

3.2.1 Causes of Phytochemical Variation ...... 36

3.2.2 Consequences of Phytochemical Variation...... 37

3.2.3 Objectives ...... 40

3.3 Materials and Methods ...... 42

3.3.1 Piper–Eois System ...... 42

viii

3.3.2 Common Garden Experiment ...... 44

3.3.3 Metabolomics ...... 44

3.3.4 Herbivory Quantification...... 46

3.3.5 Immune Assay ...... 47

3.3.6 Culture-Independent Identification of Fungal Endophytes ...... 48

3.3.7 Statistical Analyses ...... 53

3.4 Results...... 57

3.4.1 Community Analyses ...... 57

3.4.2 Metabolomics - Heritability ...... 59

3.4.3 Immune Assay ...... 60

3.5 Discussion ...... 60

3.5.1 Summary ...... 60

3.5.2 Parent land-use history influenced fungal endophyte communities but not phytochemical diversity. 61

3.5.3 Lack of heritability in phytochemical diversity, individual secondary metabolites and species interactions but not in latent variables of secondary metabolites ...... 62

3.5.4 Specialist herbivory was influenced by phytochemical diversity and ontogeny but not total herbivory 64

3.6 Conclusion ...... 65

3.7 Tables ...... 66

3.8 Figures ...... 67

4 Digestion by a Specialist Bat Cleans Fungi from Seeds of the Neotropical Shrub, Piper sancti-felicis...... 78

4.1 Abstract ...... 78

4.2 Introduction ...... 79

4.2.1 Dispersal is Critical for Terrestrial Communities ...... 79

4.2.2 Ecological Filters of Fungal and Seed Fungal Communities ...... 80

4.2.3 Objectives ...... 82

4.3 Materials and Methods ...... 83

4.3.1 Piper-Carollia System ...... 83

4.3.2 Experimental Treatments...... 84

4.3.3 Culture-Based Identification of Fungi ...... 85

4.3.4 Microdilution Assay ...... 88

4.3.5 Scanning Electron Microscopy ...... 92

4.4 Results...... 92

4.4.1 Culture-Based Seed fungi ...... 92

4.4.2 Microdilution Assay ...... 94

4.4.3 Scanning Electron Microscopy ...... 94

4.5 Discussion ...... 94

4.5.1 Summary ...... 94

4.5.2 Seeds Were Cleaned of Fungi as They Passed through the Carollia Gut ...... 95

4.5.3 No Important Changes to the Seed Fungal Community ...... 97

4.5.4 Sources of Piper sancti-felicis Seed Fungi ...... 98

4.5.5 Fruit Alkenylphenols have Antifungal Properties ...... 100

4.5.6 Digestion has no Effect on Seed Microstructures or Scarification ...... 102

4.6 Conclusion ...... 103

4.7 Tables ...... 105

4.8 Figures ...... 109

5 Conclusion...... 114

References ...... 120

6 Appendix...... 154

LIST OF TABLES

Table 2.1: Description of the hypotheses and predictions behind each path in our supported structural equation models...... 27

Table 2.2: Eois caterpillars and their host plants collected for immune assays where sample size is indicated by ‘n’. Host plant–caterpillar species information from two multi–year databases includes all collection records, specified in the ‘records’ column, caterpillars that made it to adulthood and parasitism percentage. Unitalicized names represent undescribed species...... 28

Table 2.3: Quadrus cerealis caterpillars and their host plants collected for immune assays where sample size is indicated by ‘n’. Host plant–caterpillar species information from a 19–year database includes all collection records, specified in the ‘records’ column, caterpillars that made it to adulthood and parasitism percentage...... 29

Table 2.4: SEM results from Costa Rica Eois and Q. cerealis study systems. Our hypotheses tested for: I) ‘Phytochemical diversity regulation hypothesis’ – Phytochemical diversity having direct and indirect effects on higher trophic levels and which are mediated by both herbivore immunity and herbivore diet breadth (model fit: Robust test statistic = 0.004, df = 1, P = 0.95, scaling factor = 2.08), II) ‘Diet breadth regulation hypothesis’ – Herbivore diet breadth is the main driver of herbivore immunity which in turn influences herbivore parasitism (model fit: Robust test statistic = 0.81, df = 1, P = 0.37, scaling factor = 1.13). SEM results from Ecuador Eois system. Our hypotheses tested for: I) ‘Phytochemical diversity regulation hypothesis’ – Phytochemical diversity having direct and indirect effects on higher trophic levels and which are mediated by both herbivore immunity and herbivore diet breadth (model fit: Robust test statistic = 0.28, df = 1, P = 0.60, scaling factor = 1.34), II) ‘Diet breadth regulation hypothesis’ – Herbivore diet breadth is the main driver of herbivore immunity

xii

which in turn influences herbivore parasitism (model fit: Robust test statistic = 0.16, df = 1, P = 0.69, scaling factor = 0.60). Asterisks represent significant path coefficients (P < 0.05)...... 30

Table 3.1: Multiple linear regression on specialist herbivory that was Box-Cox 2 transformed (Multiple R = 0.19, F3, 96 = 7.45, P = 0.002). Predictor variables were Z- standardized...... 66

Table 3.2: Multiple linear regression on generalist herbivory that was Box-Cox 2 transformed (Multiple R = 0.022, F3, 96 = 0.73, P = 0.54). Predictor variables were Z- standardized...... 66

Table 4.1: Frequency of fungal taxa by treatment and individual fruit where 4 seeds were collected from each fruit (NS_DIG: unsterilized, digested treatment; ST_DIG: sterilized, digested treatment)...... 105

Table 4.2: Frequency of fungal taxa by treatment (NS_DIG: unsterilized, digested treatment; ST_DIG: sterilized, digested treatment)...... 106

Table 4.3: Taxonomic and guild information of fungal taxa and the frequency of each taxon...... 107

Table 4.4: Coefficients from logistic regression of digestion and sterilization treatments on presence of fungi...... 108

Table 4.5: Beta diversity estimates using the Bray-Curtis dissimilarity index on fungi across treatments. 0 represents no dissimilarity, meaning that communities were identical, where values closer to 1 means that communities are almost completely dissimilar. NS_DIG: unsterilized, digested treatment and ST_DIG: sterilized, digested treatment...... 108

xiii

Table 4.6: Results from a mixed-effects model of an eight-serial dilution that halved the concentration of alkenylphenols on the growth of three different fungi measured as average absorbance in optical density...... 108

xiv

LIST OF FIGURES

Figure 2-1: Meta–model that structured our a priori hypotheses. Letters over paths are associated with hypotheses in Table 2-1...... 32

Figure 2-2: Multi-panel regression plots of Eois ecoimmunological parameters in Costa Rica: A) Relationship between diet breadth, measured as number of host species, and

Eois immune response, measured as total phenoloxidase absorbance per minute (B1 = 2 0.001, R = 0.003, F1,68 = 0.18, P = 0.67). B) Eois immune response and percent Eois 2 parasitism (B1 = –1.5, R = 0.16, F1,68 = 12.95, P < 0.001). C) Phytochemical diversity, 2 measured as NMR binned peak diversity, and Eois percent parasitism (B1 = 1.13, R =

0.48, F1,68 = 63.78, P < 0.001). D) Phytochemical diversity and Eois immune response 2 (B1 = –0.14, R = 0.11, F1,68 = 8.67, P = 0.004)...... 33

Figure 2-3: Multi-panel regression plots of Eois ecoimmunological parameters in Ecuador: A) Relationship between diet breadth, measured as number of host species, and Eois immune response, measured as total phenoloxidase absorbance per minute 2 (B1 = –0.003, R = 0.016, F1,81 = 1.28, P = 0.26). B) Eois immune response and percent 2 Eois parasitism (B1 = 0.083, R = 0.0004, F1,81 = 0.036, P = 0.85). C) Phytochemical diversity, measured as NMR binned peak diversity, and Eois percent parasitism (B1 = 2 1.93, R = 0.23, F1,81 = 23.82, P < 0.001). D) Phytochemical diversity and Eois immune 2 response (B1 = 0.30, R = 0.088, F1,81 = 7.82, P = 0.006)...... 34

Figure 3-1: A meta-model of the interacting variables within the experiments which shows how plant chemistry, its heritability, and land-use history act as predictors in our experiments...... 67

Figure 3-2: A map of the land-use history at La Selva Biological Station created in 2005 from the Organization of Tropical Studies...... 69

Figure 3-3: Taxa accumulation curves, with 999 permutations and where light blue shaded area represents standard error for A: offspring ZOTUs, and B: Parent ZOTUS. Effective number of fungal endophyte species using C: Alpha diversity, and D: Beta diversity. Effective number of species is a metric to understand how many equally abundant taxa are present in a community. ‘q’ refers to the exponent that Hill’s number is raised to. q = 0 is species richness, q = 1 is the exponential of Shannon’s diversity and q = 2 is the inverse of Simpson’s diversity (Gotelli and Ellison 2013). ‘q’ values are positively related to the emphasis on abundant species...... 70

Figure 3-4: A partial distance-based redundancy analysis on the effects of phytochemical diversity and land-use history of parent plants on the fungal endophyte communities (Monte Carlo permutation test: F(8,48) = 1.14, P = 0.06). The constrained variables, which were land-use history and phytochemical diversity, explained 16% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 22% of the constrained variance in the dissimilarity matrix, while canonical analysis of principal coordinates axis 2 (CAP2) explained 21%. Black data points represent fungal endophyte taxa, and larger coloured data points represent site scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by land-use history. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013). . 72

xvi

group scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by ontogeny. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013)...... 73

Figure 3-6: Parent offspring scatter plots and model analysis for variables that we hypothesized might be heritable. Chemistry data points represent chemical shifts collected from LC-MS, which is determined by functional groups and the closest structural components of the molecule. A) shows heritable compound 16 (H2 = 1.31, R2

= 0.70, F(1, 94) = 225, SE = 0.09, P< 2.2e-16). B) displays another heritable compound 2 2 (H = 0.9, R = 0.66, F(1, 94) = 182, SE = 0.004, P< 2.2e-16). C) shows how foliar fungal 2 2 richness is not heritable (H = 0.35, R = 0.0031, SE = 0.81, F1,59 = 0.19, P = 0.67). D) 2 2 demonstrates how total herbivory (%) is not heritable (H = 0.13, R = 0.0012, F1,48 = 0.58, P = 0.45)...... 75

Figure 3-7: Parent-offspring regression for factor scores from minimum residual factor 2 2 analysis with orthogonal (varimax) rotation A: (H = 0.90, R = 0.66, F(1, 94) = 182, SE = 2 2 0.067, P < 2.2e-16). B: (H = 0.87, R =0.66, F(1, 94) = 184, SE = 0.065, P < 2.2e-16). Error band represents standard error...... 76

Figure 3-8: Eois caterpillars feeding on P. sancti-felicis seedlings (n = 13) reduce the immune response (measured as total phenoloxidase) by 1.5x compared to caterpillars feeding on mature plants (n = 13) (t = 4.13, df = 23, P = 0.00041). Upper whiskers represent the third quartile + 1.5x the interquartile range. Lower whiskers represent the first quartile – 1.5x the interquartile range...... 77

Figure 4-1: Meta–model of hypotheses for how bat digestion and alkenylphenols shape the seed fungal community. Dashed arrow indicates mechanism not tested in this experiment, but which has been tested by Baldwin and Whitehead (2015)...... 109

xvii

Figure 4-2: The consequences of seed treatment (bat digestion and seed surface sterilization) on the proportion of P. sancti-felicis seeds that yielded culturable fungi. The number of seeds collected to isolate fungi for each treatment was the following: Unsterilized, Undigested: n = 108; Unsterilized, Digested: n = 84; Sterilized, Digested: n = 84. The error bars represent standard error...... 110

Figure 4-3: A partial distanced based redundancy analysis on the effects of seed treatments (bat digestion (DIG) and seed surface sterilization (NS – non-sterilized, ST - sterilized)) on the community of fungi that were cultured from seeds (Monte Carlo permutation test: F(2,22) = 0.81, P = 0.66). The treatments only explained 7% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 60% of the constrained variance (i.e. effect of treatments only) in the dissimilarity matrix, while canonical analysis of principal coordinates axis 2 (CAP2) explained 40%. Royal blue data points fungal taxa and larger coloured data points represent site scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by treatment...... 111

Figure 4-4: The effects of an 8-serial dilution (factor of 2) of the concentration (mg/ml) of alkenylphenols found in a fruit, on three P. sancti-felicis seed fungi. We measured absorbance (OD) as a proxy for fungal growth with a spectrophotometer and took measurements after 24 and 72 hours...... 112

Figure 4-5: Seed micro-structures of digested (A,B) and undigested (C,D) seeds taken with a scanning electron microscopy...... 113

xviii

LIST OF APPENDICES

Supplementary Information 1 - Species identification methods: ...... 154

Supplementary Table 1: (Chapter Two) Bootstrapping was used to impute missing data to account for differences in sampling effort. Data was bootstrapped to meet the sample size of the largest plant–caterpillar species pairs from our immune assays. We bootstrapped our data using the Hmisc v.4.0–3 (Harrell, 2015) package in R v3.4.2 (R core Team 2017) (Harrell, 2015). We specified a non–linear regression type model with one imputation and 3 knots for all our datasets which allowed for extrapolation from our data. To determine the appropriate number of knots for our data, we ran a series of imputations which varied in the number of knots and chose the number based on the lowest mean and absolute error (Harrell, 2015). After the imputation for each dataset, a R2 value was generated to predict our original measured data from the imputation as a measure of error. High R2 values indicate strong predictions. Our R2 values for the imputations in all 3 datasets were between 0.37 and 0.59...... 155

Supplementary Table 2: (Chapter Two) Bootstrapping was used to impute missing data to account for differences in sampling effort. Data was bootstrapped to meet the sample size of the largest plant–caterpillar species pairs from our immune assays. Here this is a sample size of 19...... 157

xix

‘Interaction hypothesis’, VI) ‘Simple phytochemical diversity hypothesis’, and VII) ‘Immunity does not predict parasitism hypothesis’. Asterisks next to path coefficients indicate statistically significant paths (P < 0.05). Quadrus cerealis data for VI) ‘Simple phytochemical diversity hypothesis’ did not fit the model. Eois data from Ecuador did not fit any of our models...... 158

Supplementary Table 4: (Chapter Four) Fungal taxa classification using the ITS region and the UNITE database (v8.2). Lab ID refers to the ID the taxa are given in our data. UNITE generate estimates of species hypothesis (SH) based on the ITS2 region and SH codes below the species name are digital object identifiers...... 167

Supplementary Figure 1: (Chapter Three) Bayesian t-test on unedited total herbivory measurements of the MIPAR herbivory recipe. Blue histograms represent the credibility of all possible parameter values, which is the posterior distribution. When the distribution of values is roughly symmetric, the mean is given at the top of the figure, if it is skewed, the mode is given. The 95% HDI is the highest density interval and shows the uncertainty in the estimated parameter. The comparison value is 0 since we are interested in whether the differences in the means deviate from this (shown in green). The ROPE represents the region of practical equivalence and is an interval of parameter values that are considered equivalent to our comparison value...... 162

Supplementary Figure 2: (Chapter Three) A partial distanced based redundancy analysis evaluating ontogeny, phytochemical diversity and herbivory on fungal endophyte community incidence data (Monte Carlo permutation test: F(4,90) = 4.15, P = 0.001). Predictor variables explained 16% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 73% of the constrained variance in the (Jaccard) dissimilarity matrix, while canonical analysis of principal

coordinates axis 2 (CAP2) explained 14%. Black data points represent fungal endophyte taxa and larger coloured data points represent site scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by ontogeny. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013)...... 163

Supplementary Figure 3: (Chapter Three) A partial distanced based redundancy analysis on the effects of phytochemical diversity land use history of parent plants on the fungal endophyte community incidence data (Monte Carlo permutation test: F(8,48) = 1.88, P = 0.001). Predictor variables explained 24% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 43% of the constrained variance in the (Jaccard) dissimilarity matrix, while canonical analysis of principal coordinates axis 2 (CAP2) explained 20%. Black data points represent fungal endophyte taxa and larger coloured data points represent site scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by land use history. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013)...... 164

Supplementary Figure 4: (Chapter Three) A variable importance plot, derived from a random forest regression to determine the influence of individual parent chemical compounds on a heritable offspring factor (Factor Two), which was derived from a factor analysis (Random Forest: number of trees grown = 500, number of variables randomly sampled at each split = 26). Essentially, how do individual compounds influence the most heritable latent variable in offspring. %IncMSE represents how much the model accuracy decreases if that compound was left out of the model and stands for % increase in mean quared error. IncNode Purity indicates the increase in how well each node was split while building the random forest at each compound, where 0 represents a pure node, meaning a perfect split and is measured by the change in sum of squares.

xxi

The overall model explained 68% of the variance in the data and had a mean squared residual of 8.9e-08. When the model was evaluated against the test set, the explained variance decreased to 40% and the mean squared error was 1.60e-07. While the overall model is meaningful and explains a large proportion of the data, the importance of any one compound is low. The analysis was derived from the R package and function randomForest v4.6-14 (Liaw and Wiener 2002)...... 165

xxii

CHAPTER PUBLICATIONS

Chapter 2: Across multiple species, phytochemical diversity and herbivore diet breadth have cascading effects on herbivore immunity and parasitism in a tropical model system, was published in Frontiers in Plant Science in 2018. DOI: 10.3389/fpls.2018.00656

The microdilution assay in Chapter 4: Digestion by a bat specialist cleans fungi from seeds of the neotropical shrub, Piper sancti-felicis, was prepared for publication with my colleagues from the Whitehead lab at Virginia Tech, as: Secondary metabolites in a neotropical shrub: spatiotemporal allocation and role in fruit defense and dispersal. The manuscript was published in Ecology in 2020. DOI: 10.1002/ecy.3192

1 General Introduction

1.0 Overview

My research explores the community ecology of tropical herbivores and fungi with a special focus on how plant chemistry is involved in plant-insect and plant-fungi- herbivore interactions. Specifically, I am interested in how: i) plant chemistry shapes interactions across insect herbivores and their parasitoid enemies, ii) plant chemistry may act as a dominant ecological filter in the community assembly of fungal communities (including fungal endophytes), and iii) other ecological filters, such as plant ontogeny, land-use history, herbivory and digestion may contribute to fungal endophyte communities. Fungal endophytes are a cryptic component of biodiversity that colonize the interior tissues of all plants, are typically asymptomatic and exhibit a spectrum of interactions with their host plants (Rodriguez et al. 2009). Historically, fungal endophytes have been challenging to study because of their asymptomatic nature and their enormous diversity, where thousands of species are capable of colonizing a single tropical tree (Fröhlich and Hyde 1999). However, advances in ‘systems biology1’ have helped us open this formerly black box. Due to the large impact that fungal endophytes can have on plant chemistry and plant-insect interactions, it is crucial that we also understand other aspects of their ecology. The overall hypothesis for my dissertation is

1 Systems biology encompasses ‘omics’ techniques such as: genomics, transcriptomics, proteomics, and metabolomics.

that the assembly and interactions of tropical insect, mammalian herbivores and fungal communities is complex given the high species diversity of the ecosystem, but that plant chemistry is critical in shaping each organismal group. The motivation of my body of research is to understand better the ecological consequences of plant chemistry on species interactions and communities, which can help to inform scientists about how species interactions and communities might shift in response to perturbations such as climate and global change or provide potential applications for agriculture.

1.1 Tropical Ecosystem Significance

Tropical ecosystems host extremely diverse ecological communities, which make up the majority of global biological diversity (Dirzo and Raven 2003). These ecosystems provide essential functions and services for the planet (Myers et al. 2000). Tropical insects and fungi are both extremely speciose and have important roles in tropical communities. However, in contrast to temperate species and ecosystems, the bulk of tropical insects and fungi, and the assembly of their communities, are undescribed or unknown (Delabye et al. 2019, Lopez-Vaamonde et al. 2019). For the tropical fungi that are known, we understand the roles of only 5% or less (Arnold and Lutzoni 2007). Meanwhile, overall species diversity and abundance are severely affected by global change drivers (Salcido et al. 2019), such as land-use change (Sala et al. 2000), overexploitation of flora and fauna (Symes et al. 2018) and climate change (Dyer and Letourneau 2013), which are impacting important ecosystem functions (e.g. nutrient cycling) and services (e.g. climate regulation) (Bradshaw et al. 2009). In order to conserve essential services and resources, it is critical to identify insect and fungal species and comprehend the assembly and functioning of these tropical communities.

1.2 Species’ Functional Roles and Interaction Diversity are Central to Communities

In well-studied temperate ecosystems, scientists have uncovered some critical information about the diversity of plants, insects and fungi, including their interaction diversity (i.e. the number of different interactions and their abundance between species in a community; Janzen 1974, Thompson 1996) and the function of these species in their communities. For example, plants are foundational for establishing communities as they create and affect resources for other organisms (Price et al. 1980, Scherber et al. 2010, Hunter 2016), such as pollinators, seed dispersers, herbivores, enemies, decomposers and humans. The majority of species interactions in terrestrial ecosystems occurs between plants, herbivorous insects and their parasitoids (i.e. tri- trophic interactions) (Price 2002). The phytochemical landscape, which includes plant species diversity, genetic variation, nutrition and chemistry, can have cascading effects on herbivores, higher trophic levels and ecosystem processes (Whitham et al. 2006, Hunter 2016), including primary productivity (Tilman et al. 2001, Crutsinger et al. 2006, but see Grace et al. 2007) and nutrient cycling (Tilman et al. 1996).

Plant chemistry is a particularly important plant trait affecting tri-trophic interactions (Fraenkel 1959, Price et al. 1980), especially in the tropics (Coley and Aide 1991, Coley and Barone 1996, Becerra 2015). Over the last 60+ years, community ecologists and evolutionary biologists have focused on the relationship between plant chemistry and herbivores and untangling the origins of their diversity (Ehrlich and Raven 1964, Berenbaum and Zangerl 2008). Both plant chemical dissimilarity and insect herbivore specialization have been important reciprocal drivers of diversification (Becerra 2015).

Perhaps surprisingly, closely related tropical plant species often have disparate chemical profiles (Kursar et al. 2009, Salazar et al. 2016a). In plant communities, unique plant chemistry tends to provide better defence against herbivores compared to communities with similar chemistries (Endara et al. 2017). Furthermore, ingestion of host plants with high plant chemical diversity (or phytochemical diversity) has negative effects on the ability of caterpillars to mount an immune response (Smilanich et al. 2009b, Hansen et al. 2017). Herbivore immune response is a useful index of herbivore fitness (Bukovinsky et al. 2009, Smilanich et al. 2009a, Lampert and Bowers 2015), and the degree to which the immune response is affected is due to herbivore physiology and diet breadth (Smilanich et al. 2009 a,b, Richards et al. 2010). For instance, a weakened specialist herbivore immune response from feeding on toxic plant hosts increases the susceptibility of specialist herbivores to defend themselves against parasitoids (Smilanich et al. 2009b, Hansen et al. 2017) but better defends them against predators (Dyer et al. 2004a). While the studies listed above have greatly advanced our knowledge on the reciprocal interactions between plants and their insect herbivores, future research by tropical scientists that uses analytical chemistry to document a broader spectrum of the plant compounds present can add new information to the many studies that have focused on one or two compounds (as reviewed by Dyer et al. 2018). It has been proposed that the study of phytochemical diversity and investigating synergistic effects of compounds are more relevant and needed to explain patterns in tri-trophic communities (Richards et al. 2010, 2016). Similarly, while plant chemistry alone is an important facet for understanding plant-insect interactions, there are other kingdoms of organisms which also strongly impact host plant chemistry, and therefore plant-insect interactions, that deserve more attention.

Fungi are one example of these other kingdoms that influence plant chemistry and species interactions and, in particular, fungal endophytes which colonize the interior of their host plants (Spiteller et al. 2015). In particular, fungal endophytes produce a wealth of secondary metabolites inside their host plants (Nisa et al. 2015), with some well-known medicines being sourced from them (e.g. Taxol: Stierle et al. 1993). Due to changes in host plant chemistry caused by fungal endophytes, they can negatively affect mammalian herbivores (Ralphs and Stegelmeier 2011) and insect herbivores, depending on the diet breadth of the herbivore (Hartley and Gange 2009, Gange et al. 2019). Fungal endophytes can alter the communities of insects on a plant by altering plant chemistry (Estrada et al. 2014), nutrient uptake (Christian et al. 2019) and pathogen resistance (Busby et al. 2016, Christian et al. 2017). Moreover, fungal endophytes can upregulate the host plant’s immune response and modify the expression of over a hundred different host plant genes, some of which are involved in chemical defence, such as plant secondary metabolites2 (PSMs) and genes involved in the ethylene defence pathway (Mejía et al. 2014). Due to the large impact that fungal endophytes can have on plant chemistry and plant-insect interactions, it is crucial that we also understand other aspects of their ecology.

2 PSMs are products of metabolism that are not used in primary functions such as growth, development, and reproduction.

1.3 Community Assembly and Ecological Filtering of Microbial Communities

In order to understand fully the roles of fungal endophytes in their host plants and to be able to predict the context with which they appear and the how they interact with other organisms, it is critical to identify the factors that structure fungal endophyte communities (i.e. community assembly). By definition, community assembly is the process by which organisms establish in an environment and is influenced by the regional species pool, dispersal, physiology, timing of arrival and biotic interactions (Morin 2011, Kraft et al. 2015). One classic example of the downstream consequences of a fungal endophyte on its community is Epichloë (Clavicipitaceae). Epichloë is found in cool-season grasses and is involved in a variety of symbioses with its host that depend on the context of the environment and reproduction method of the fungus (Newman et al. 2020). For example, when Epichloë acts as a mutualist, it can negatively impact the health of grazing herbivores due to the production of toxic alkaloids (Clay 1988). Thus, understanding the underlying mechanisms that enabled the endophyte to colonize its host is useful to predict changes to the community. In general, the community assembly of microorganisms involves both deterministic and stochastic processes and has been developed for community ecology by borrowing four mechanisms of evolution from population genetics: 1) selection, 2) drift, 3) diversification (in place of mutation), and 4) dispersal (in place of migration) (Vellend 2010, Nemergut et al. 2013, Vacher et al. 2016). Selection is a deterministic process of microbe community assembly whereby abiotic factors such as: precipitation (McGuire et al. 2012), temperature (Zimmerman and Vitousek 2012) and leaf nutrient concentration (Larkin et al. 2012) can affect selection (establishment). Furthermore, selection can act

via plant traits (Kembel and Mueller 2014, Kembel et al. 2014), such as leaf defensive chemistry, which may influence the colonization of fungi in the leaf (Saunders and Kohn 2009, Lau et al. 2013, Rajala et al. 2014). There is also a lot of evidence that demonstrates how plant genotype can shape fungal endophyte communities (Balint et al. 2013, Busby et al. 2013, Wagner et al. 2016), although the environment can sometimes outweigh the importance of plant genotype (Whitaker et al. 2018).

In contrast, drift, a stochastic mechanism, may alter microbe community assembly, particularly when communities are dominated by rare species, because randomness largely affects species presence (Pedros-Alio 2006, Unterseher et al. 2011). This is frequently the pattern in microbial communities (Pedros-Alio 2006, Unterseher et al. 2011). Diversification of genetic variation may occur under either deterministic or stochastic processes, as microorganisms have the capability of undergoing dormancy and may become re-activated under ideal environmental conditions or by a random event (Buerger et al. 2012, Shapiro et al. 2012). Moreover, microbes have short generation times (Schmidt et al. 2007) and high mutation rates and can pass genes through horizontal gene transfer (Ochman et al. 2000). Dispersal may occur via stochastic or deterministic processes depending on the dispersal strategy (Nemergut et al. 2013). Fungi disperse primarily via water and air (Adams et al. 2013), and the diversity of fungi in plants is affected by climate (Arnold and Lutzoni 2007). Moreover, herbivory can increase fungal colonization of plant tissue by exposing the interior of the leaf (Humphrey and Whiteman 2019). Understanding how communities are shaped can lead to a better understanding of how species interact with one another and their environment, shed light on their functional roles and potential human

applications and enable us to make predictions across communities (Zambell and White 2017).

1.4 Objectives

Overall, plant chemistry is a pivotal component of ecological communities, which can be both influenced by other organisms and simultaneously shape the community around them. In my dissertation, I use multiple plant taxa from my study system Piper (Piperaceae) to answer my questions. Piper is a pantropical genus that is extremely species rich, and which exhibits large variation in plant chemical profiles. Piper also hosts an abundance of highly specialized insect and mammalian herbivores that are adapted to feeding on its toxic and well-defended plant organs. For instance, Eois is a lepidopteran genus that exclusively feeds on Piper, and in addition, individual species of Eois tend to be relatively specialized and only feed on 1-3 different Piper species (Slinn et al. 2018).

In Chapter Two, I begin my investigation into the consequences of variation in plant chemistry on insect-herbivore and parasitoid interactions. I sought to understand how variation in plant chemical diversity across different plant species can affect herbivore immunity and their susceptibility to parasitoids in two different tropical ecosystems and between specialist and generalist herbivores. While working on this chapter, I began to wonder about how other symbiotic organisms alter and are altered by plant chemistry.

In Chapters Three and Four, I investigate the ecological filters that influence fungal endophyte communities. In Chapter Three I explore the heritability of the full

phytochemical profile of Piper sancti-felicis in parent-offspring regression design. I wondered whether heritability in plant chemistry influenced heritability in fungal endophyte communities and insect herbivory. I researched how differences in plant chemistry between young and mature Piper influenced herbivory damage. I also sought to identify other ecological filters that shaped the fungal endophyte communities independently of the parent-offspring regression experiment, such as land-use history, herbivory and phytochemical diversity. In my last research chapter, I examine the role of fruit chemistry in protecting fungi from damaging or colonizing seeds and how seed fungal communities and micro-structures are affected by digestion of a specialist bat (Chapter Four).

1.5 Study Sites

My dissertation research took place at two different field stations in the neotropics: 1) La Selva Biological Station, Heredia Province, Costa Rica (100 26’ N 830 59’ W) (all chapters) and 2) Yanayacu Biological station, Napo Province, Ecuador (000 36’ S 770 53’ W) (chapter 2). The La Selva Biological reserve is 1600 ha of lowland rainforest and ranges from 35 to 140m in elevation and is surrounded by a combination of disturbed, agricultural habitat, and natural forest. The mean annual precipitation is approximately 4200mm. Sampling at Yanayacu Biological Station included the 100-ha owned by the station as well as thousands of hectares of surrounding cloud forest on the slopes of the eastern Andes. The elevation at the station is 2100m, and the annual precipitation is approximately 2624mm. Both sites are ideal to address my research questions as they host a high number of Piper species, and La Selva in particular has the necessary equipment for microbial-based work.

2 Across Multiple Species, Phytochemical Diversity and Herbivore Diet Breadth Have Cascading Effects on Herbivore Immunity and Parasitism in a Tropical Model System

2.1 Abstract

Terrestrial tri-trophic interactions account for a large part of biodiversity, with approximately 75% represented in plant-insect-parasitoid interactions. Herbivore diet breadth is an important factor mediating these tri-trophic interactions, as specialization can influence how herbivore fitness is affected by plant traits. We investigated how phytochemistry, herbivore immunity, and herbivore diet breadth mediate plant- caterpillar-parasitoid interactions on the tropical plant genus Piper (Piperaceae) at La Selva Biological station in Costa Rica and at Yanayacu Biological Station in Ecuador. We found that phytochemical diversity was an important predictor for herbivore immunity, herbivore parasitism, and diet breadth for specialist caterpillars, but that the specifics of these relationships differed between sites. Furthermore, phytochemical diversity did not affect herbivore immunity and parasitism for the more generalized herbivore. Results also indicated that herbivore diet breadth is an important factor mediating herbivore immunity and parasitism success for Eois both in Costa Rica and Ecuador. These patterns contribute to a growing body of literature that demonstrate strong cascading effects of phytochemistry on higher trophic levels that are dependent on herbivore specialization and that can vary in space and time. Investigating the

interface between herbivore immunity, plant chemical defence, and parasitoids is an important facet of tri-trophic interactions that can increase biodiversity.

2.2 Introduction

Tri-trophic interactions are important for structuring communities and maintaining the biodiversity that composes a large proportion of terrestrial food webs and mediates ecosystem processes (Price et al., 1980; Agrawal, 2000; Price, 2002; Whitham et al., 2006). For instance, terrestrial plant-insect-enemy interactions may make up approximately 75% of biodiversity (Price, 2002). Ecologists have found that tri-trophic interactions can shape community parameters, such as species diversity, functional diversity, primary productivity, and consumer abundance (Hairston et al., 1960; Ives et al., 2005; Singer and Stireman, 2005; Crutsinger et al., 2006; Johnson, 2008; O’Connor et al., 2016). Many tri-trophic studies have focused on how primary producers affect biotic communities through effects on densities or population dynamics of herbivores, mutualists, and natural enemies (Crutsinger et al., 2006; Crawford et al., 2007; Barbour et al., 2015). Plant chemical defence is one of the most important components of these bottom-up effects, and there is a rich literature documenting how chemistry affects plant-insect interactions (Frankel, 1959; Ehrlich and Raven, 1964; Schoonhoven et al., 2005; Hunter, 2016), via both negative and positive physiological effects on herbivores and their natural enemies (Smilanich et al., 2016). One clear gap in our knowledge of how phytochemistry influences tri-trophic interactions is empirical data that consider the entire suite of plant secondary metabolites in a species instead of focusing on one or two major compounds (Richards et al., 2010; Richards et al., 2016; Smilanich et al., 2016). Given that herbivores are exposed to the full array of compounds during their

larval development, significant consideration should be given to the diversity of secondary metabolites found in plants (Hay et al., 1994; Richards et al., 2015). Here, we use phytochemical diversity as a metric of plant defence to investigate the effects on herbivore performance as measured by immune strength, and whether effects on the immune response cascade to impact parasitism (Smilanich et al., 2009a; Richards et al., 2015; Hansen et al., 2017).

Research on the role of herbivore immunity as a mediator of tri-trophic interactions has been expanding over the last decade (Bukovinsky et al., 2009; Smilanich et al., 2009b; Richards et al., 2012; Singer et al., 2014; Lampert and Bowers, 2015). However, the majority of this work has been performed in temperate systems (but see: Smilanich et al., 2009a; Smilanich and Dyer, 2012; Hanson et al., 2017), where plant chemistry is typically less diverse and compounds are less toxic (Coley and Barone, 1996; Dyer and Coley, 2002). In general, increased concentrations or mixture complexities of plant chemical compounds have a detrimental impact on herbivore immunity (Haviola et al., 2007; Smilanich et al., 2009b; Lampert, 2012; Richards et al. 2010; Richards et al., 2016; Hansen et al., 2017), but these effects can differentially influence the success of predators and parasitoids (Dyer et al., 2004; Bukovinsky et al., 2009; Richards et al., 2015). For instance, research shows that specialist caterpillars (Junonia coenia: Nymphalidae) sequestering high concentrations of secondary metabolites have a weakened immune response (i.e. ‘vulnerable host hypothesis’ Smilanich et al., 2009b; Lampert and Bowers, 2015). Similarly, the ‘safe haven hypothesis’ proposes that parasitoids are more likely to survive in sequestering hosts since these hosts will be protected from predators due to the sequestered compounds (Gentry and Dyer, 2002; Lampert et al., 2010; Lampert and Bowers, 2015). These two

hypotheses combined suggest that phytochemically defended plants may be hosts to herbivores that are immunocompromised and more likely to be attacked by parasitoids (Lampert et al., 2010).

The effect of host plant chemistry on the immune response also depends on the natural history of the organism: herbivores that utilize physiologically expensive strategies to tolerate host plant chemistry may incur physiological costs to eating toxic diets and experience compromised immune systems (Smilanich et al., 2009b). A previous study found that a mixture of plant secondary metabolites from a neotropical shrub in the genus Piper (Piperaceae) affected a naïve generalist noctuid caterpillar (Spodoptera) versus adapted specialist geometrid caterpillars (Eois) differently, with Spodoptera experiencing high mortality through direct toxicity and indirect top-down effects on Eois via increased levels of parasitism (Richards et al., 2010). This study suggests that higher phytochemical diversity may weaken the caterpillar’s immune response, leading to increased parasitoid success – based on the experimental design, there were no direct effects of chemistry on adult parasitoids, since caterpillars were exposed to parasitoids first and then subsequently assigned to feeding treatments. Similarly, iridoid glycosides sequestered by buckeye caterpillars (Junonia coenia) depressed the immune system of these specialists (Smilanich et al., 2009b) but did not affect the immune response of the generalist caterpillar, Grammia incorrupta (Erebidae: Arctiinae) (Smilanich et al., 2011). Overall, there is growing evidence that plant chemistry may mediate parasitoid susceptibility via the herbivore’s immunity, and the strength or direction of this relationship is dependent on the natural history of the plant- herbivore interaction. While these previous studies have included how diet breadth may

affect the ecoimmunology3 of tri-trophic interactions, there are other axes of variation that are likely to be important, including biogeographical differences among sites. For example, plant chemistry and tri-trophic interactions vary across elevations (Rodriguez- Castañeda et al., 2016) and with rainfall intensity (Cunningham et al., 1999); thus, the same herbivore species may be affected differently by host plant chemistry and parasitoids across elevational gradients, due to differences in chemistry and in enemy communities (Rodriguez-Castañeda et al., 2016).

In this study, we used the tropical plant genus, Piper, the associated herbivore genus, Eois (Lepidoptera: Geometridae; all investigated Neotropical species are monophagous on Piper), and a Piper generalist, Quadrus cerealis (Lepidoptera: Hesperiidae), to investigate whether variation in phytochemical diversity influences the strength of the herbivore immune response and associated levels of parasitism (Figure 2.1, Table 2.1). In addition to examining variation across these different herbivore species, we examined these relationships in two distinct ecosystems – a lowland wet forest in Costa Rica (La Selva, Sarapiqui) and a cloud forest in Ecuador (Yanayacu, Napo). Specifically, we designed our study to address the following questions: 1) How does phytochemical diversity influence herbivore immunity and levels of parasitism, and how are these relationships affected by diet breadth? 2) How do these effects vary across different herbivore species and different locations?

3 Ecoimmunology is the study of the ecological consequences caused by variation in the immune systems of organisms.

2.3 Methods

2.3.1 Piper–Eois, Piper–Quadrus System

The plant genus Piper (Piperaceae) is an emerging tropical model system for studying tri-trophic interactions because of the growing knowledge on its evolutionary history, genomics, plant chemistry, distribution, and insect communities (Marquis 1991; Greig 1993a; Dyer and Palmer, 2004; Richards et al., 2015; Glassmire et al., 2016; Salazar et al., 2016b). Currently, there are over 2000 species of Piper that have been described pantropically, with approximately 1300 of these species occurring in the neotropics, 50 species present at the La Selva Biological station, and 20 present at the Yanayacu station. Piper is a phytochemically diverse genus, including compounds from at least 15 classes, and a total of 667 individual compounds have been discovered (Richards et al., 2016). In this study, we used previously published data quantifying phytochemical diversity for multiple Piper species (Richards et al., 2015). Briefly, phytochemical diversity is an effective number of functional groups, transformed from a Simpson’s diversity entropy calculated from proton nuclear magnetic resonance (1H–NMR), which incorporates both mixture complexity and structurally complexity, the two key components of chemical diversity (Richards et al., 2015).

Piper species host diverse lepidopteran herbivore communities that vary in diet breadth (Dyer and Palmer, 2004). Caterpillars in the genus Eois (Lepidoptera: Geometridae) are Piper specialists that feed exclusively on 1–4 different Piper species (Connahs et al., 2009). They are one of the most well studied and abundant genera of caterpillars found on Piper, and over 80% of Eois species are found in the neotropics with others in Africa, Asia and Australia (Rodriguez-Castenada et al., 2010; Brehm et

al., 2011). In contrast, the Piper skipper, Quadrus cerealis (Lepidoptera: Hesperiidae), has been recorded feeding on 23 Piper species; in this paper we categorize this skipper as a Piper generalist (http://www.caterpillars.org; Dyer et al., 2010).

2.3.2 Long-term Rearing Databases

Our group of collaborators have been documenting plant-herbivore-parasitism data in Costa Rica for the last 20 years. We used this collection database from 1996–2015 (where 2015 was the final year of complete data) to determine herbivore diet breadth and parasitism levels (see Richards et al., 2015; Tables 2.2 and 2.3). I personally collected data from September 2015 to December 2015. Data consisted mainly of entries from La Selva Biological Station, but also from other areas nearby such as Braulio Carrillo National Park and the Tirimbina Biological Reserve. Caterpillars were collected year-round in all forest types and reared on the host plant from which they were collected in ambient conditions until they pupated and eclosed into adulthood, or if parasitized prior to collection, until they succumbed to parasitism. In this study, we evaluated herbivore immunity for 4 different Eois species collected from 5 different Piper species (Table 2.2). For these species, we found a total of 2011 records in our database, with 900 caterpillars successfully reaching adulthood (Table 2.2). Additionally, we collected Quadrus cerealis from 10 different Piper species, though we have records of larvae feeding on 23 different Piper species (Table 2.3). We recorded 117 instances of Q. cerealis on these 10 Piper species, with 75 caterpillars successfully reaching adulthood (Table 2.3).

The same data collection procedure was utilized at the Ecuador site, where the database spans 15 years (2001–2015). Here again, I collected data over the course of

five weeks at the end of December 2015 and beginning of January 2016. Larvae were collected in the cloud forest surrounding Yanayacu Biological Station. At this site, we measured the immune response from 8 Eois morphospecies feeding on 3 different Piper species (Table 2.2). We had 2079 records of our Eois morphospecies in our database, with 809 records of parasitism (Table 2.2). We calculated diet breadth and levels of parasitism using the same method at both sites. Diet breadth was calculated as the number of Piper species on which a caterpillar species was found feeding and successfully reared to adult moth or parasitoid. Parasitism was calculated as number of parasitism events for each caterpillar species divided by the total number of successfully reared adults and instances of parasitism of that species (parasitoids/(healthy adults + parasitoids).

2.3.3 Immune Assay

Phenoloxidase (hereafter PO) is an important enzyme for triggering the melanization process, a mechanism of innate immunity involving deposition of pigments on foreign bodies (Beckage, 2008; Gonzalez-Santoyo and Cordoba-Aguilar, 2012). Since active PO can have locally toxic effects (Cerenius et al., 2007), it is derived from a non- activated form called prophenoloxidase (proPO), typically stored in hemolymph cells. Upon infection or natural enemy attack, proPO is converted to the active form, PO, which catalyzes the cascade to produce melanin. PO has been shown to be an important part of the immune response in arthropods, protecting them from bacteria, viruses, and parasitoids (Cerenius et al., 2007). We measured the activity of the PO enzyme as an indicator of the strength of the herbivore immune response (Gonzalez- Santoyo and Cordoba-Aguilar, 2012). We collected early-instar caterpillars and reared them in ambient conditions until they reached 5th instar. To measure PO activity, we

took 2 µL of hemolymph from each Eois caterpillar (Costa Rica: N=70, Ecuador: N=83) and 5 µL from each Q. cerealis caterpillar (Costa Rica: N=65) and divided the volume into two Eppendorf tubes—one for cell-free PO found in the hemolymph at the time the hemolymph is taken (standing PO), and one for cell-bound PO, which was artificially activated by adding a chemical activator (total PO). The aliquots of hemolymph were added to 50 µl of phosphate buffered saline (PBS) for Eois individuals and 100 µl PBS for Q. cerealis individuals. For the total PO in both species, 35 µl of chymotrypsin (1mg/mL) was added to the PBS-bound hemolymph, vortexed for 2 seconds, then incubated at room temperature for 20 minutes. During incubation, the substrate, dopamine, (0.0284g/10 mL distilled water), was prepared. Since this compound is light sensitive, fresh dopamine was prepared daily. For Eois, we added 300 µl of dopamine to each Eppendorf tube, vortexed for 2 sec., then added 25 µl of the dopamine- hemolymph mixture to a well plate. For Q. cerealis, we added 500 µl of dopamine to each Eppendorf tube, vortexed for 2 sec., then added 200 µl of the dopamine- hemolymph mixture to a well plate. We used a spectrophotometer (BIO-RAD: iMark Microplate Absorbance Reader) at a wavelength of 490nm to measure the activity of PO every 30 seconds for 45 minutes. We measured the slope, which was the rate of reaction, from the first 10 minutes because it was a linear increase. PO assays were performed in Costa Rica from January 2013 to December 2015 and in Ecuador from December 2015 to January 2016.

2.3.4 Statistical Analyses

Structural equation models (SEM): We used SEM to evaluate 7 a priori hypotheses, which tested for bottom-up effects of plant chemistry and herbivore diet breadth on herbivore immunity and parasitism success (Figure 2.1, Table 2.1). We used the global

estimation method in the R packages piecewiseSEM v.1.2.1 (Lefcheck, 2016) and lavaan v.0.5–23 (Rosseel et al., 2017) to run our SEMs in R v3.4.2 (R core Team 2017). We were not able to normalize the residuals of our data, so we chose a more robust estimator to account for non-normality and unequal variance instead of the default maximum likelihood method; this method is based on the Satterthwaite approach and is called the ‘maximum likelihood estimation with robust standard errors and a mean and variance adjusted test statistic’ (Rosseel et al., 2017). Lastly, we used the same 7 hypotheses in our Ecuador dataset as we had no reason to believe that our systems would operate differently (Figure 2.1, Table 2.1).

2.4 Results

Average immune response for Eois, as measured by total PO absorbance per minute (∆Abs), was approximately equal across sites (Eois: Costa Rica: 0.03 ± 0.004 ∆Abs; Ecuador: 0.02 ± 0.001 ∆Abs; here and elsewhere, error is 1 SEM), and between specialist Eois and Piper generalist, Q. cerealis (Q. cerealis: 0.02 ± 0.002 ∆Abs); however, average parasitism level was higher for Q. cerealis (0.34 ± 0.03 percent parasitism) compared to Eois at both sites (Costa Rica: 0.12 ± 0.01 percent parasitism; Ecuador: ± 0.01 percent parasitism). Parasitoid families attacking the caterpillars also differed between sites and species. Quadrus cerealis parasitism was entirely tachinid fly parasitoids, while Eois parasitism in Costa Rica was 80% braconid wasps, 8% tachinids, and 12% parasitism by other families. Eois parasitism in Ecuador was 24% tachinids, 41% braconids, and 35% parasitism by other families.

Overall, the best-fit structural equation models supported the hypotheses that both phytochemical diversity and herbivore diet breadth are important factors shaping

herbivore, immunity and parasitism for Eois species in both Ecuador and Costa Rica; however, for some relationships, the directions of the effects were reversed from one site to another (Table 2.4, Figure 2.3, Figure 2.4). Tests of 7 a priori models to explain the relationships of our measured variables (see Supplementary Table 3) yielded the same two best models for both sites: I) The phytochemical diversity regulation hypothesis, and II) The diet breadth regulation hypothesis (Table 2.4).

Costa Rica: The phytochemical diversity regulation hypothesis (model I) included phytochemical diversity as an exogenous variable with direct paths to herbivore immunity, herbivore diet breadth and herbivore parasitism; the model also included effects of herbivore immunity and diet breadth on herbivore parasitism (model fit: Robust test statistic= 0.004, df= 1, P = 0.95, scaling factor = 2.08). This model supported the hypothesis that there is a strong direct positive effect of phytochemical diversity on herbivore parasitism (Figure 2.3C, standardized path coefficient (hereafter, spc) = 0.65, P < 0.01, slope (B1) = 1.13), showing that herbivores feeding on plants with high phytochemical diversity had higher parasitism rates. This model also showed that phytochemical diversity decreases herbivore immunity (Figure 2.3D, spc = –0.34, P <

0.01, B1 = –0.14). It supports the hypothesis that higher herbivore immunity decreases herbivore parasitism frequency (Figure 2.3B, spc = –0.19, P = 0.08, B1 = –1.52). Lastly, this model shows a negative effect of phytochemical diversity on herbivore diet breadth (i.e. Piper species with greater phytochemical diversity consumed by more specialized

Eois species; spc = –0.12, P = 0.03, B1 = –2.66), but herbivore diet breadth has a weak, positive effect on herbivore parasitism (i.e. generalists have higher levels of parasitism; spc = 0.17, P = 0.11, B1 = 0.001). The diet breadth regulation hypothesis (model II) is a simpler model focusing on the effects of herbivore diet breadth on herbivore immunity

and parasitism (model fit: Robust test statistic= 0.81, df = 1, P = 0.37, scaling factor = 1.13). This model shows that consuming a greater number of host plants weakly increases herbivore immune response (Figure 2.3A, spc = 0.05, P = 0.76, B1 = 0.001) and that immune function reduces parasitism success (Figure 2.3B, spc = –0.40, P <

0.01, B1 = –1.52). Both phytochemical diversity and herbivore diet breadth are important determinants of herbivore and parasitoid ecology. Models for Q. cerealis caterpillars in Costa Rica did not fit the data; for example, a model where phytochemical diversity affects herbivore immunity, which in turn influences herbivore parasitism, was a poor fit to the data (model fit: Robust test statistic= 28.37, df= 1, P < 0.01, scaling factor= 0.50). However, a separate regression analysis showed that phytochemical diversity had a negative relationship with Q. cerealis parasitism (B1 = – 4.39, F(1,63) =15.25, P < 0.01).

Ecuador: The same two models were supported by our Ecuador Eois data. The phytochemical diversity regulation hypothesis (model I) was strongly supported by the data; however, the directions of some of the relationships were reversed (model fit: Robust test statistic= 0.28, df= 1, P = 0.60, scaling factor= 1.34). Consistent with the Costa Rica data, phytochemical diversity has a strong positive effect on herbivore immunity (Figure 2.4D, spc = 0.30, P < 0.01, B1 = 0.30) and on herbivore parasitism

(Figure 2.4C, spc = 0.53, P < 0.01, B1 = 1.93). Phytochemical diversity has a negative effect on herbivore diet breadth (spc = –0.21, P < 0.01, B1 = –9.55), and herbivore immunity negatively affects herbivore parasitism (Figure 2.4B, spc = –0.13, P = 0.22, B1 = 0.08). Lastly, diet breath has no effect on herbivore parasitism (spc = 0.06, P = 0.62,

B1 = –0.003). The diet breadth regulation hypothesis was again supported by our data (model II) (model fit: Robust test statistic= 0.16, df= 1, P = 0.69, scaling factor= 0.60), but for this site, diet breadth has a weak negative effect on herbivore immunity (Figure

2.4A, spc = –0.13, P = 0.30, B1 = –0.003), and herbivore immunity has no effect on herbivore parasitism (Figure 2.4B, spc = 0.02, P = 0.84, B1 = 0.08).

2.5 Discussion

Our results corroborate many other studies demonstrating that the chemistry of herbivore host plants, as well as the overall number of host plant species consumed, have strong effects on multiple aspects of herbivore ecology (Berenbaum and Neal, 1985; Haviola et al., 2007; Diamond and Kingsolver, 2011; Lampert and Bowers, 2015), including immunity and parasitism (Smilanich et al., 2009b; Hansen et al., 2017). A focus on the immune response allows for investigation of an important natural history parameter that is directly linked to protection against natural enemies (Smilanich et al., 2009a), putting our results in a strong tri-trophic context. It is also interesting that the relationships between phytochemical diversity, immunity, and parasitism were dependent upon the diet breadth of the specialist herbivores and that relationships varied across herbivore taxa and sites. In Costa Rica, Eois feeding on Piper species with high phytochemical diversity had a weakened immune response, while the immune response of Q. cerealis was unaffected. Furthermore, Eois data in Ecuador fit the same two models as in Costa Rica; however, some relationships were reversed. For example, in Costa Rica, individuals with a strong immune response had lower parasitism frequency (model II); however, in Ecuador herbivore immunity had almost no effect on parasitism frequency. This difference may be due to the differences in parasitoid pressure between the two sites. Compared to Ecuador, the database shows that Eois in Costa Rica have three times more parasitism by a relatively more specialized parasitoid community (Braconidae). Our Ecuador data include plant-caterpillar species pairs that

are not well represented in our historical database and which have 0% parasitism as a result. We ruled out that this was driving our observed patterns by re-running our SEMs without plant-caterpillar species pairs that had low representation in our database, but found the same qualitative result. We therefore included these data points in our final analysis. Other possible particulars of the taxa and sites used for our study, such as degree of specialization and elevation of the site, may also be responsible for these differences, but greater insight into those variables will require further experimentation using carefully selected taxa and locations.

Untangling relationships between plant chemistry, herbivores, and natural enemies has been a focus of insect ecology for decades (Price et al., 1980; Bernays and Graham, 1988; Dyer, 1995; Dyer, 2011), and our results with Eois in Costa Rica are consistent with emerging paradigms of the importance of phytochemistry in mediating multi-trophic interactions. Most notably, we provide further support for the ‘safe haven hypothesis’ (Lampert et al., 2010) and the ‘vulnerable host hypothesis’ (Smilanich et al., 2009b). Briefly, specialized herbivores should be better adapted to diverse mixtures of secondary metabolites in individual host plant species than generalist herbivores, and thus more likely to sequester those compounds. While sequestration can protect specialists from natural enemies (e.g., Dyer, 1995), the energetic costs that accompany sequestration may lead to reallocation of resources such that the immune response suffers, rendering specialists more susceptible to parasitism (Smilanich et al., 2009b). These specialists provide a ‘safe haven’ for parasitoids because they are subjected to a compromised immune system and are less likely to be consumed by other natural enemies, which tend to avoid toxic specialist hosts (Dyer, 1995). Indeed, generalists are often better protected than specialists against parasitoids (Dyer and Gentry, 1999). Eois

data from Costa Rica support all aspects of this ‘safe haven hypothesis’, and data from both sites support the more general concept that changes in chemistry are likely to alter herbivore immunity and parasitism – the positive effects of phytochemical diversity on herbivore immunity in Ecuador are not inconsistent with this hypothesis; they simply require further investigation to determine mechanisms causing this relationship. Furthermore, both SEM models (Table 2.4, Hypotheses I and II) are consistent with the growing body of evidence that the ability of an insect herbivore to mount an immune response is negatively associated with herbivore parasitism (Bukovinsky et al., 2009; Quintero et al., 2014), which is an important component of the safe haven hypothesis, and some have argued that this is the best predictor of parasitism (Smilanich et al., 2009a; Greeney et al., 2012).

Other studies that support the ‘safe haven hypothesis’ (Gentry and Dyer 2002; Lampert et al., 2010) or related hypotheses (i.e. ‘nasty host hypothesis’ Barbosa et al. 1991; Gauld et al. 1992) have focused on detoxification or sequestration of individual compounds or entire classes of compounds and have measured relative concentrations of those compounds (e.g., Haviola et al., 2007; Diamond and Kingsolver 2011; Lampert and Bowers, 2015). We utilize a different approach and consider the fact that phytochemical mixtures are complex, and herbivores may be as susceptible to mixture complexity, synergies, or additive effects rather than just increases in concentrations of individual compounds or classes, such as tannins (Richards et al., 2015). One shortcoming of this approach is that results will require further investigation to get at mechanism. In Costa Rica, the immune responses of Eois species were negatively affected by increases in phytochemical diversity (Table 2.4, Hypothesis I). Another study with Eois on Piper found that changes in mixture complexity are associated with

synergistic effects on parasitoid success (Richards et al, 2010). It is possible that host plants with higher phytochemical diversity are more likely to have synergistic effects on herbivores, impairing immune function, regardless of whether the mixtures are sequestered.

It is interesting to note that the results depended on taxon (Quadrus versus Eois) and site (Ecuador versus Costa Rica). Such variation is expected, and it is worth further investigation to determine conditions that are favorable for these chemically mediated tri-trophic interactions. Site and taxon were treated as random effects in the broader sense and were not statistically compared; nevertheless, it is interesting to consider possibilities for some of the differences across the two taxa and the two sites. Specialist Eois caterpillars in Costa Rica support our predictions, whereas the same genus of caterpillars in Ecuador do not support any of our a priori models. Elevation is one clear difference between these sites, with the cloud forest in Ecuador situated 2,000 m higher than the lowland forest in Costa Rica. It is well known that herbivore development rates, herbivory, levels of predation, and herbivore diversity are lower at higher elevations, while parasitism and parasitoid diversity increase with elevation (Rodríguez-Castañeda et al., 2011; Rodríguez-Castañeda et al., 2016), so it is not surprising that the specifics of chemically mediated tri-tropic interactions would vary with elevation. Reasons for the positive effect of phytochemical diversity on immunity at higher elevation are not obvious, but given the higher levels of parasitism and slow development rates, it is possible that maximized immunity is enhanced with slow development rates since larvae are exposed to parasitoids for longer periods of time. Similarly, there are many differences between the geometrid and hesperiid caterpillars utilized in our study, including diet breadth; however, one large difference is that Quadrus is a concealed

feeder (i.e. folds leaves around itself as it feeds), and concealed feeders are affected less by phytochemical defence (Sandberg and Berenbaum, 1989; Berenbaum 1990) and experience very high levels of parasitism (Gentry and Dyer, 2002). As such, Q. cerealis appeared to be unaffected by changes in chemistry and experienced extremely high levels of parasitism. There are likely unmeasured variables that influence immunity of hesperiids and more generally of concealed feeders, and it is certainly possible that the greater diet breadth played a role in the differences noted here.

In summary, our research builds on previous work investigating the effects of phytochemical diversity and herbivore diet breadth on ecoimmunology and tri-trophic interactions. These results are unique in that variation in phytochemical diversity, rather than concentrations of specific compounds, was a predictor of tri-trophic interactions and herbivore immunity. These patterns are also particularly important for understanding tropical systems due to intense biotic interactions and high levels of diversity (Dyer and Coley, 2002; Novotny et al., 2006). Future work should investigate how much intraspecific phytochemical variation exists within these species and how such variation affects higher trophic levels.

2.6 Tables

Table 2.1: Description of the hypotheses and predictions behind each path in our supported structural equation models.

Explanatory Response Paths Hypotheses and predictions Citations variables variables Berenbaum and Neal, 1985; Jones Plant quality Herbivore Low plant quality caused by toxic secondary metabolites, and Firn, 1991; Smilanich et al., (Phytochemical fitness A and high phytochemical diversity are more likely to 2009a; Diamond and Kingsolver, diversity) (Immunity) contain compounds that decrease herbivore fitness 2011; Lampert and Bowers, 2015

The immune system provides important protection against Herbivore fitness Herbivore Bukovinsky et al., 2009; Smilanich et B parasitoids, thus as the strength of the immune system (Immunity) parasitism al., 2009b; Quintero et al. 2014 decreases, parasitism increases

Low plant quality caused by toxic secondary metabolites, Plant quality and higher phytochemical diveristy are more likely to Lill et al., 2002; Bukovinszky et al., Herbivore (Phytochemical C weaken herbivores via the presence of bioactive 2009; Richards et al., 2010; Sternberg parasitism diversity) compounds and/or toxic synergies, increasing parasitoid et al., 2012; Hunter, 2016 success

Plant quality Plants with greater diversity of phytochemical compounds Herbivore diet Becerra, 2007; Becerra, 2015; Dyer (Phytochemical D are more likely to host specialized herbivores that have breadth et al., 2007; Richards et al., 2015 diversity) adapted to bioactive compounds and/or toxic synergies

Herbiovre Specialist herbivores are adapted to detoxifying or Herbivore diet Coley et al., 2006; Richards et al., fitness E sequestering toxic plant compounds and will perform breadth 2010; Lampert, 2012 (Immunity) better on their host plants than generalists

Herbivores that feed on a greater number of plants are Barbosa et al., 1991; Carvalheiro et Herbivore diet Herbivore F exposed to a greater variety of toxic plant compounds al., 2010; Lampert et al., 2011; breadth parasitism which weaken herbivores, increasing parasitoid success Reudler et al., 2011

Site Eois spp. Piper spp. Database n records adults % parasitized Costa Rica Eois nympha Piper biseriatum 9 44 7 29 Piper cenocladum 28 921 317 18 Eois apyraria Piper cenocladum 1 328 164 8.4 Piper imperiale 7 616 359 1.4 Eois russearia Piper sancti-felicis 12 48 24 4 Eois mexicaria Piper umbricola 13 54 29 0 Total 70 2011 900 Ecuador Six black two pink spots Piper baezanum 2 6 1 0 Piper kelleii 16 1792 700 14 Piper lancifolium 1 1 0 0 Lime slime Piper baezanum 1 3 1 0 Piper kelleii 7 9 0 0 Two black spots Piper kelleii 27 83 29 3.3 Piper lancifolium 1 1 0 0 Eois viridiflava Dognin Piper baezanum 1 2 0 0 Piper lancifolium 20 36 0 0 Pink spots funk Piper kelleii 3 86 37 8.1 Piper lancifolium 1 1 0 0 Eight black blur Piper baezanum 1 1 9 0 Eois beebei Fletcher Piper kelleii 1 36 19 14 Eois ignefumata Dognin Piper kelleii 1 22 13 19 Total 83 2079 809

Site Piper spp. Database n records adults % parasitized Costa Rica Piper arboreum 3 2 2 0 Piper cenocladum 1 4 3 25 Piper colonense 13 16 13 38 Piper garagaranum 1 3 2 33 Piper imperiale 6 2 1 50 Piper multiplinervium 19 26 26 7.7 Piper pseudobumbratum 1 1 1 0 Piper reticulatum 18 62 26 68 Piper trigonum 2 0 0 0 Piper umbricola 1 1 1 0 Total 65 117 75

2.7 Figures

Figure 2-1: Meta–model that structured our a priori hypotheses. Letters over paths are associated with hypotheses in Table 2-1.

Figure 2-2: Multi-panel regression plots of Eois ecoimmunological parameters in Costa Rica: A) Relationship between diet breadth, measured asEcuador number of host species, and Eois immune response, A B 2 measured as total phenoloxidase absorbance per minute (B1 = 0.001, R = 0.003, F1,68 = 0.18, P = 0.67).

2 B) Eois immune response and percent Eois parasitism (B1 = –1.5, R = 0.16, F1,68 = 12.95, P < 0.001). C)

Phytochemical diversity, measured as NMR binned peak diversity, and Eois percent parasitism (B1 =

2 1.13, R = 0.48, F1,68 = 63.78, P < 0.001). D) Phytochemical diversity and Eois immune response (B1 = –

2 0.14, R = 0.11, F1,68 = 8.67, P = 0.004).

C D

Figure 2-3: Multi-panel regression plots of Eois ecoimmunological parameters in Ecuador: A) Relationship between diet breadth, measured as number of host species, and Eois immune response,

2 measured as total phenoloxidase absorbance per minute (B1 = –0.003, R = 0.016, F1,81 = 1.28, P = 0.26).

2 B) Eois immune response and percent Eois parasitism (B1 = 0.083, R = 0.0004, F1,81 = 0.036, P = 0.85).

C) Phytochemical diversity, measured as NMR binned peak diversity, and Eois percent parasitism (B1 =

2 1.93, R = 0.23, F1,81 = 23.82, P < 0.001). D) Phytochemical diversity and Eois immune response (B1 =

2 0.30, R = 0.088, F1,81 = 7.82, P = 0.006).

3 Intraspecific variation in plant chemistry and land-use history in a common garden experiment act as an ecological filter to insect herbivory and fungal endophyte communities

3.1 Abstract

Plant chemistry is an important mediator of community assembly and species interactions and is a particularly important factor shaping insect herbivores, their enemies and the colonization of fungi. Simultaneously, these taxa exhibit pressure on the plant, which can alter host plant chemistry. Due to the ecological significance of plant chemistry, it is crucial to understand what causes variation in phytochemistry and how these changes will influence species interactions. In this chapter, we sought to quantify the heritability of Piper sancti-felicis chemistry using a common garden experiment. We also estimated heritability for traits such as herbivory and fungal endophyte communities. Lastly, we looked to identify other ecological factors shaping plant-species interactions and to determine whether differences in plant chemistry across different stages of ontogeny influenced herbivore immunity. We found that land- use history had a meaningful effect on fungal endophyte community composition. In addition, parents and offspring differed greatly in their fungal endophyte communities, which was partially attributed to phytochemical diversity and herbivory. In contrast, we evaluated the effect of phytochemical diversity, ontogeny and fungal endophytes on herbivory and found that ontogeny and phytochemical diversity were both important in predicting the percentage of specialist herbivory found on leaves. Our analysis

suggested that there were two compounds that were strongly heritable but that no species interaction measures were heritable across parents and their offspring. A factor analysis showed that there were latent variables extracted from the metabolomics data that were highly heritable. Lastly, we found that there were important chemical differences between parents and offspring and that this had a meaningful impact on the ability of specialist herbivores to mount an immune response. Our work adds to the existing literature showing how phytochemistry is important in mediating species interactions.

3.2 Introduction

3.2.1 Causes of Phytochemical Variation

Plant chemistry is a foundational trait of communities and ecosystems (Hunter 2016), due partially to its role in host plant defence (Fraenkel 1959, Berenbaum et al. 1986, Dyer et al. 2004a, Agrawal and Weber 2015). Variation in the phytochemical landscape is influenced by genetics (Johnson et al. 2009, Barbour et al. 2015), plant ontogeny (Coley and Barone 1996, Boege and Marquis 2005, Barton and Koricheva 2010), fungal endophyte colonization (Clay 1988, Mejía et al. 2014, Pusztahelyi et al. 2015) and the abiotic environment (Cunningham et al. 1999, Rodriguez-Castañeda et al. 2016, Glassmire et al. 2017, 2019). For example, changes in plant chemistry vary nonlinearly with ontogeny; however, the direction and magnitude of these changes depend on the plant life form and the plant compound classes, which are hypothesized to be driven by herbivore pressure and resource allocation and availability (Barton and Koricheva 2010). In the tropics, foliar plant secondary metabolites (PSMs) are an especially important layer of defence for young leaves, which are often better chemically defended

than mature leaves due to the intense pressure of herbivores on expanding leaves (Coley and Barone 1996, Wiggins et al. 2016). Furthermore, the PSM phenotype is influenced by the interaction of its genotype and the environment, where a given genotype may exhibit a variable phenotype across an environmental gradient (i.e. reaction norms). The production of PSMs tends to be a highly heritable physiological plant trait compared to other types of traits (e.g. morphology, developmental phenology, vegetative performance; Geber and Griffen 2003). While high heritability does not necessarily equate to high evolutionary potential for many traits due to positive correlations between additive variance and other contributions to phenotypic variation, this does not seem to be the case for some products of plant secondary metabolism (Hansen et al. 2011, Moore et al. 2014). Thus, a high degree of heritability in plant chemistry may allow plants to adapt quickly to their environment and important selection pressures such as herbivores and pathogens (Agrawal et al. 2002, Dyer and Coley 2002, Maynard et al. 2020), though evolutionary constraints still exist (Geber and Griffen 2003, Johnson et al. 2009). Despite having learned a great deal about the causes of PSMs, this chapter seeks to understand how heritability and other ecological factors contribute to plant chemical phenotypes in a neotropical shrub and the consequences these have for the interactions and symbiont communities which it hosts.

3.2.2 Consequences of Phytochemical Variation

The consequences of plant chemical variation are central in shaping bottom-up effects of species interactions, particularly for insect herbivores and their enemies, and have received an enormous amount of attention over the last 60 years (Fraenkel 1959, Ehrlich and Raven 1964, Hunter 2016, Dyer et al. 2018). One area of research on the ecological interactions of plant chemistry is how heritability of many intrinsic chemical

factors can influence the community assembly of other species. For example, previous research has found a spectrum of effect sizes across plant systems that investigated how heritability in plant chemistry influenced herbivore communities (Johnson et al. 2009, Barbour et al. 2015, Barker et al. 2019). In evening primrose (Oenothera biennis), Johnson et al. (2009) found high heritability in phenolics, ellagitannins, and flavonoids, but relatively low heritability in herbivory damage, although herbivory was still strongly influenced by plant chemistry. Similarly, a meta-analysis on the genotype-by- environment effects in the willow family (Salicaceae) found that heritability of salicinoids, condensed tannins, and insect herbivore performance (which included survival, growth, and fecundity) was high and outweighed phenotypic plasticity (Barker et al. 2019). While PSMs are a critical component explaining insect herbivore community responses (e.g. plant damage, herbivore performance, or composition), other plant traits can outweigh them, and this can sometimes depend on the measure of herbivore response (Carmona et al. 2011, Barker et al. 2019) and the herbivore species.

While untangling the consequences of heritability in phytochemical variation on herbivory and herbivore performance is important, other features of the phytochemical profile and insect herbivores need to be considered. Quantifying the diversity of plant chemical compounds and their functional groups is an ecologically important metric that yields a more realistic understanding of how these compounds may interact with one another, through synergy, to influence herbivore communities (Richards et al. 2015, 2016). For instance, increasing phytochemical diversity in Piper increased herbivore diversity and the abundance of specialized herbivores but decreased overall herbivory (Richards et al. 2015). Simultaneously, given the strong top-down selection pressure of parasitoids on insect herbivores (Hawkins et al. 1997), research on Piper has found that

increased phytochemical diversity increased parasitism, which negatively impacted specialist herbivores (Richards et al. 2015, Hansen et al. 2017, Slinn et al. 2018). Selection pressures on insect herbivores have enabled them to evolve diet-breadths and adaptations that allow them to feed either on plants with particular suites of defences (e.g. specialists) or across plant families (e.g. generalists) (Ehrlich and Raven 1964, Dyer et al. 2007, Forister et al. 2015). Some of the adaptations include detoxification (Zangerl et al. 2008) or sequestration (Bowers 1990) of toxic compounds, which often come at the cost of a reduced immune function (Haviola et al. 2007, Smilanich et al. 2009b, Lampert 2012). In addition, due to the nature of PSM profiles changing with plant ontogeny, the plant developmental stage poses a significant influence on herbivore immunity (Quintero et al. 2014) and herbivory. In response to reduced immune function, parasitoids tend to be more successful in using these herbivores as a host (i.e. vulnerable host hypothesis, Dyer 1995, Bukovinsky et al. 2009, Smilanich et al. 2009b). Meanwhile, predators are more likely to avoid specialist herbivores because of their toxicity due to sequestration and detoxification of plant compounds, further increasing the survival of parasitoids residing in specialists (i.e. safe haven hypothesis; Dyer 1995, Gentry and Dyer 2002, Smilanich et al. 2009b). Thus, the herbivore immune response is an important mediator of tri-trophic interactions and is one of the most important defences against herbivore parasitism and a critical component of herbivore fitness (Smilanich et al. 2009a).

In addition to the direct consequences of the plant chemical profile on insect herbivores and their parasitoid enemies, there are reciprocal interactions between plant chemistry, insect herbivory, and fungal endophyte communities. Host plant chemistry is an important ecological filter that affects the colonization of fungal endophyte

communities since the interior of the leaf is in direct contact with fungal endophytes (Saunders and Kohn 2009, Fernandes et al. 2011, Christian et al. 2020). Reciprocally, the host plant may mount an immune response, either through the identification of unspecific molecules typically created by microorganisms that may or may not be pathogenic, or through pathogen-specific virulence factors (Jones and Dangl 2006). Thus, the mounting of the immune response can occur as a result of fungal endophyte colonization, which can in turn affect the production of PSMs (Mejía et al. 2014) and consequently insect herbivory (Hartley and Gange 2009, Gange et al. 2019). Moreover, herbivory can increase the colonization of fungal endophytes by creating openings for colonization (Humphrey et al. 2014, Humphrey and Whiteman 2020), and these endophytes can reciprocally alter their host plant chemistry through the production of their own secondary metabolites (Stierle et al. 1993, Nisa et al. 2015, Christian et al. 2020). Variation in host plant chemistry, or other factors associated with the host species or host genotype (for e.g. structural traits), can shape heritable variation across interactions of other symbionts (Lamit et al. 2015, Whitham et al. 2020), such as herbivores (Barbour et al. 2015, Barker et al. 2019) and fungal endophytes (Peiffer et al. 2013, Lamit et al. 2014, 2016). For example, we know that certain plant genotypes can recruit particular communities of microbes (Bodenhausen et al. 2014, Wagner et al. 2016, Jones et al. 2019), and this can sometimes be mediated by host-plant chemistry (Simon et al. 2020).

3.2.3 Objectives

Intraspecific variation in plant chemistry is an important driver of species interactions and is simultaneously influenced by genotype and phenotypic plasticity, leading to diverse norms of reaction. To address the heritability of plant chemistry and the

complex, reciprocal interactions this has on herbivores and fungal endophytes, we used the neotropical shrub, Piper sancti-felicis (Piperaceae), and its lepidopteran herbivore genus Eois (Geometridae; all Neotropical species are monophagous on Piper) (Figure 3.1). Piper is a useful system to address this research topic, as it has been shown to exhibit strong intraspecific chemical variation and different reaction norms for PSMs and herbivory (Marquis 1990, Dyer et al. 2004b, Glassmire 2017). We conducted our study at La Selva Biological Station, which is located in the lowland rainforest in Costa Rica. Here, we used a common garden experiment to determine the broad-sense heritability of PSMs, which is the contribution of genetic variation to phenotypic variation (Connor and Hartl 2004, Zimmer and Emlen 2016). It should be noted that P. sancti-felicis has bisexual flowers and it is unclear how much outcrossing occurs. For this reason and because we only have the trait data of the maternal plant, we classify heritability here as broad-sense heritability. We collected parents across an environmental gradient at La Selva Biological Station (Figure 3.2), and offspring were propagated from seed and grown first in the field in a common garden and then they were moved into a shade house. We asked five questions: 1) How do land-use history and phytochemical diversity influence fungal endophyte communities in parent plants? 2) How do phytochemical diversity, parent-offspring group, and fungal endophyte communities influence herbivory? And conversely, 3) How do phytochemical diversity, parent- offspring group, and herbivory influence fungal endophyte communities? 4) How heritable are plant chemistry, fungal endophyte communities, and herbivory? And 5) How does variation in plant chemistry via ontogeny alter herbivore immunity? For Q1, we predicted that given the importance of the environment in filtering microbial communities, especially horizontally transmitted microbes, land-use would be an important determinant of fungal endophyte communities. We also predicted that

phytochemical diversity would be important given the results of previous studies on the consequences of phytochemical diversity on species interactions in Piper (Richards et al., 2015, 2016, Hansen et al. 2017, Slinn et al. 2018). Similarly, for Q2, we predicted that phytochemical diversity, parent-offspring group, and fungal endophyte communities would shape overall herbivory, as fungal endophytes can directly (Nisa et al. 2015) and indirectly (Mejía et al. 2014) influence host plant chemistry. Our prediction for Q3 was that due to reciprocal interactions between fungal endophytes and herbivores, we expected to find that phytochemical diversity, parent-offspring group and herbivory are important ecological filters for fungal endophyte communities. We predicted that for Q4, we would see heritability for the intensity of interactions with at least one of our symbionts, especially herbivory, given the importance of this for Piper (Dyer et al. 2001, Richards et al. 2015). Finally, for Q5 we predicted that the changes in plant chemistry associated with ontogeny would mean that herbivores feeding on young plants would have a more difficult time mounting an immune response because of higher phytochemical defence.

3.3 Materials and Methods

3.3.1 Piper–Eois System

Our research took place at La Selva Biological Station, near the town of Puerto Viejo de Sarapiqui, Heredia province, Costa Rica. La Selva Biological Station is a 1600 ha reserve that consists of a diverse land-use history located in the lowland rainforest of eastern Costa Rica. We collected samples from secondary old-growth forests that had been logged within the last 22–33 years. We also sampled from more selectively logged forests adjacent to primary old-growth rainforest and from a shaded pasture. We

collected from less-forested sites, which included an abandoned plantation, an active plantation, an open pasture, and the developed area around the station buildings. Lastly, we sampled from an experimental area called the successional plots that are periodically clear cut, and natural growth is allowed to return over several years, with different plots at different stages of succession. Across these land-use histories, plant communities and light availability change drastically.

We conducted a series of experiments to address our objectives. First, we collected leaf samples from the parent plants that were existing mature P. sancti-felicis shrubs that were randomly selected across the reserve. This enabled us to correlate the phytochemical diversity of leaves and land-use history with foliar fungal endophyte communities. Next, we set-up a common garden experiment to estimate broad-sense heritability of plant secondary metabolites and species interactions. Finally, we conducted an immune assay using Eois to determine how ontogenetic shifts in plant chemistry affected herbivore immune response.

At La Selva Biological station, there are over 50 Piper species present, with over 2000 total found in the genus pantropically. Piper is perhaps best known for its production of the alkaloids piplartine and piperine (Dyer and Palmer 2004, Salehi et al. 2019), but the genus produces many other alkaloids, amides, lignans, and terpenes as well (Parmar et al. 1997, Dyer and Palmer 2004). Many Piper species, particularly those with amides and propenylphenols, are capable of producing anti-fungal compounds (Parmar et al. 1997, Kato and Furlan 2007), which could have a direct impact on the fungal endophytes able to colonize their tissues. In particular, P. sancti-felicis is known to produce a compound class that is ecologically important and relatively unique to Piper, termed alkenylphenols (Maynard et al. 2020). Alkenylphenols vary across plant

organs and are thought to improve fruit defence, through anti-fungal activity and to aid in seed dispersal (Maynard et al. 2020).

3.3.2 Common Garden Experiment

Humberto Garcia Lopez, Lee Dyer, Angela Smilanich, and Chris Jeffrey set up the experiment in January 2016 by collecting three to five replicate clones and three to five replicate seeds from the mature fruiting plants (also called parent plants) (N = 92) across La Selva Biological Station and mature plants (N = 8) at Tirimbina Biological Reserve. In total, we had 100 parent plants. Piper can be grown clonally, or asexually, by breaking branches at one of their nodes where they are attached to a central trunk and sexually through seed. Humberto Garcia Lopez and Lee Dyer grew parent clones and offspring in a common garden along the edges of the Huertos trail at La Selva Biological station. Unfortunately, our common garden was damaged on the edge of the trail, and we had to move the common garden into a shade house near the research buildings. Humberto Garcia Lopez, Lee Dyer, Dani Salcido, and an Earthwatch team harvested offspring in the summer of 2017 after reaching maturity, where the offspring started producing fruit.

3.3.3 Metabolomics

In collaboration with the Jeffrey lab at the University of Nevada Reno, we conducted liquid chromatography – mass spectrometry (LC-MS) on our offspring and our parent plant samples, which gives information about the structure of compounds in samples and their mixture complexity. The raw chemistry data used in the analyses are values associated with chemical shifts, which is determined by functional groups and the

closest structural components of the molecule. As mentioned in the introduction, both structure of compounds and mixture complexity have important implications for our study system as phytochemical diversity is a functional trait (i.e., a trait that influences plant fitness) and can significantly impact the herbivores, the success of their enemies, and the amount of damage they inflict (Richards et al. 2015, Slinn et al. 2018). Note: The chemical shifts, which could be compounds or functional groups, identified from mass spectrometry data present in my dissertation have not been identified. I prepared samples through weighing and grinding of leaves, and the Jeffrey lab conducted the remainder of the downstream metabolomics work. I conducted the statistical analysis that is presented here.

We prepared our samples by grinding dried, surface-sterilized leaves with a tissue lyser (TissueLyser II, Qiagen) and weighing 100mg of each sample to be used in the chemical analysis. We added 10mL of methanol (Optima grade, Fisher Scientific) to our samples and sonicated the suspensions for 10 minutes before extracting them using mechanical overnight wrist-action shaking. We filtered the suspensions using a piece of cotton, and we concentrated the filtrate in vacuo before weighing the remaining sample.

We re-dissolved the samples in di-methanol (Sigma Aldrich) and diluted an aliquot of the solution to 1:10 using protonated methanol for LC-MS analysis.

We conducted LC-MS using Agilent 1200 analytical HPLC equipped with a binary pump, autosampler, column compartment, and diode array UV detector, coupled to an Agilent 6230 Time-of-flight mass spectrometer via an electrospray ionization source (ESI-TOF; gas temperature: 325 °C, flow: 8 L/m; nebulizer pressure: 35 psig; VCap: 3500 V; fragmentor: 175 V; skimmer: 65 V; octopole: 750 V). We took 0.50 μl extracts and injected them with naringenin internal standard (1.00 μl, 0.1mM, MilliporeSigma)

and then eluted this at 0.400 ml/min through a Kinetex Phenyl-hexyl column (Phenomenex, 2.1 x 100 mm, 2.6 μ, 100 Å) at 40 °C. We used a linear binary gradient, which consisted of buffers A (water containing 0.1 % formic acid, 18 MΩ) and B (acetonitrile containing 0.1 % formic acid, Optima grade, Fisher Scientific), which changed over 24 minutes according to the following instructions: 0 min 20% B, ramp to 40% B at 4 min, ramp to 70% B at 14 min, ramp to 100% B at 16 min, ramp to 0.8 mL/min at 19 min, ramp to 20% B at 0.5 mL/min at 20 min, and ramp to 20% B at 0.4 mL/min at 24 min to re-equilibrate column. We converted the LC-MS data to a mzML format using ProteoWizard (Kessner et al. 2008) before processing using XCMS (Smith et al. 2006) in R. We corrected chromatographic features using retention time and aligned the features using density grouping. We grouped features into pseudospectra by retention time, isotopic distribution, peak shape, and correlation across individuals using the XCMS wrapper CAMERA (Kuhl et al. 2012). We summed all of the peaks within each pseudospectrum before normalizing to dry plant mass.

3.3.4 Herbivory Quantification

Humberto Garcia Lopez and I removed leaves from the common garden P. sancti-felicis plants, took images of them, dried them, and stored them in silica to be passed for microbial and chemical analysis. I quantified herbivory of leaves using the image processing software MIPAR (Sosa et al. 2014). In conjunction with a software specialist, I developed a program within MIPAR that automated herbivory measurements. We structured leaf measurements to estimate total original leaf area, even if significant herbivory damage was made on leaf edges, and to estimate missing leaf area (i.e. herbivory). In addition, the program specified different types of herbivory that are left by different genera based on damage characteristics of the most common Piper insect

herbivores (Dyer et al. 2010). We used this information to classify specialist and generalist herbivory, as well as total herbivory. We were able to make these classifications because of previous work which collected common Piper herbivores from the field, independently fed them Piper leaves, and documented their damage (herbivory identification key: Dyer et al. 2010). In the past, herbivory data were collected using the software ImageJ (Schneider et al. 2012). We compared our results to data already processed using ImageJ and results from MIPAR were within a 5% error threshold, and MIPAR was up to 16 times faster than ImageJ. We conducted a formal Bayesian analysis to determine whether there was a true difference between herbivory measurements of the two software (Supplementary Figure 1).

3.3.5 Immune Assay

We used the same protocol as in Slinn et al. (2018), where we quantified the production of phenoloxidase (PO) in hemolymph from its inactivated form, prophenoloxidase, using a spectrophotometer (BIO-RAD: iMark Microplate Absorbance Reader). An undergraduate student (Quinn Campbell), Humberto Garcia Lopez, and an Earthwatch team collected early instars of the specialist caterpillar genus, Eois, from different Piper species in the wild and reared them to fifth instar. They reared caterpillars on either mature plants (N = 12) from the wild or from seedlings (N = 13) in the common garden experiment; however, we have no reason to believe that this distinction is influential in the interpretation of our results as our question is about the difference in the effects of chemistry associated with ontogeny. They took data from the first 10 minutes because this is the beginning of the activation of the PO enzyme and used the slope of the PO production during this interval to represent immune strength (González-Santoyo and Córdoba-Aguilar 2012). A steeper slope represents that the insect can mount a strong

immune response more quickly. The spectrophotometer is set to take measurements every 30 seconds for 45 minutes so that we can confirm the leveling off of PO production.

3.3.6 Culture-Independent Identification of Fungal Endophytes4

3.3.6.1 Sample Collection

We performed a culture-independent assessment of foliar fungal endophytes from the common garden experiment. To isolate fungal endophytes from the rest of the leaf microbial community, we surface-sterilized leaves in a laminar flow hood using a series of bleach-ethanol washes. We agitated leaf fragments in sterile glass vials in 95% ethanol for 30 seconds, followed by two minutes in 5% bleach, two minutes in 70% ethanol, and finished with a rinse in sterile DI water (Harrison et al. 2018a). We confirmed that our sterilization removed all surface fungi by imprinting leaves on agar plates.

3.3.6.2 Library Preparation and Next Generation Sequencing

We extracted DNA (Qiagen: DNeasy plant mini kit) from parent (N = 67) and offspring plant samples (N = 64). We transported genomic DNA from La Selva Biological Station to the University of Nevada Reno, where the Nevada Genomics Center quantitated our DNA with a fluorometric spectrophotometer (ThermoFisher Scientific: Invitrogen Qubit

4 I conducted all of the culture-independent work, bioinformatics, and statistical analyses.

2.0). We normalized our samples to a maximum of 10 ng/μl (with 10mM of Tris pH 8.5), although some samples had lower concentrations of DNA and used negative controls for each batch of DNA extraction. Following suggestions by Nguyen et al. (2015) about determining the quality of the run, we added a mock community with equal DNA concentrations from eight different fungal species of broad taxonomic breadth. We targeted the ITS1 region, which is approximately 250 base pairs long, using primers ITS1F (5’-CTTGGTCATTTAGAGGAAGTAA-3’) and ITS2 (5’- GCTGCGTTCTTCATCGATGC-3’) (White et al. 1990). We sent our samples for library preparation and next-generation sequencing to the Genomic Sequencing and Analysis Facility (GSAF) at the University of Texas at Austin in April of 2017 (parent samples) and 2018 (offspring samples). We split samples by ontogeny for sequencing runs. We used the Illumina MiSeq platform to generate paired-end reads (2x250). GSAF used two PCR stages for their NGS library preparation: amplicon PCR and Index PCR. For amplicon PCR, GSAF mixed both forward and reverse primers (1μl at 5μM each), with NEBNext 2X Master Mix (New England Biolabs; 10μl), water (7μl), and genomic DNA (1μl). The cycling conditions for this stage of PCR included an initial denaturation of 98oC for 30 seconds. This was followed by 12 cycles of: 98oC for 30 seconds, 62oC for 30 seconds, and 72oC for 30 seconds. The final extension was held at 72oC for five minutes, before entering into a 4oC hold. GSAF added indexed primers for in silico sample assignment in the next PCR stage. Once again, they combined NEBNext 2X Master Mix (15μl) with indexed Illumina primers (2.4μM at 5μl each) and amplicon DNA (10μl) from the first stage of PCR. Cycling conditions were the same as the first cycle, except that the cycling occurred 7 times instead of 12 for the second cycle. GSAF used negative controls for both PCR stages. GSAF conducted PCR clean-up after both amplicon and index PCR in order to remove additional primers and primer dimers. This

purification required the addition of AMPure XP beads at a concentration of 80% compared to the sample (Beckman Coulter). GSAF mixed the beads with the sample, removed the supernatant, and conducted two 80% ethanol washes (200μl). GSAF used qPCR to quantitate the libraries, normalized them, and pooled the samples together. They included a positive control in this PCR. They also used a High Sensitivity Bio- analyzer chip (Agilent) to check for primer adapter contamination. Finally, GSAF demultiplexed our reads, meaning that in silico, they were reassigned to the sample that they originally came from.

3.3.6.3 Sequence Processing

Below we describe our bioinformatics pipeline based on Josh Harrison’s (personal communication) and USEARCH’s v11.0.667 example ITS 2x250 Illumina Miseq pipeline (Edgar 2010). The functions we used came mostly from USEARCH (Edgar 2010) with a few from VSEARCH v2.8.5 (Rognes et al. 2016), as there are small differences in some of the functionality between the two metabarcoding programs. Therefore, we used USEARCH unless otherwise specified. Due to the long size of the 250 bp region, we merged forward and reverse reads using fastq_mergepairs with maximum base pair differences set at 50 and a minimum read alignment at 60% using fastq_pctid. To confirm that all reads were oriented correctly, we used orient against the Warcup database on the first five samples, which is enough of a sample to ensure that all reads are oriented correctly. Using VSEARCH (Rognes et al. 2016), we removed the primer- binding regions (ITS1F: 22 base pairs; ITS2: 20 base pairs) with fastx_truncate and then filtered reads, meaning that we independently checked fastq files for expected errors, rather than relying on their given quality scores (Q scores), which describe the probability of a sequencing error. The higher the Q score, the lower the probability of a

sequencing error for every base pair; it is not an average for the read. We filtered reads with fastx_filter, where we specified reads with an expected error of one or more were removed (fastq_maxee = 1). We also removed duplicate sequences (termed dereplication), using derep_fulllength in VSEARCH. We specified that the Q scores were a maximum of 45 (Probability = 3.16e-05 or 0.00316 %) for the last two functions. To generate exact sequence variants (ESVs) or zero-radius operational taxonomic units (ZOTUs), we denoised reads, which means that we removed errors and chimeras identified from the filtering step mentioned above. We used ESVs over traditional operational taxonomic units, which are defined by arbitrary thresholds because ESVs delineate sequences based on as little as one base pair (Callahan et al. 2017). Moreover, they are replicable across studies, facilitating reproducibility of research, and they are not affected by reference databases that lack their sequence (Callahan et al. 2017). To denoise our sequence data we used the unoise3 algorithm (Edgar 2016a), then we classified ESVs to taxa using the SINTAX algorithm (Edgar 2016b) against the UNITE (Kõljalg et al. 2005) and Warcup (Deshpande et al. 2016) reference databases (accessed April 15th, 2020) using a minimum of 80% confidence threshold for taxonomic classifications.

We analyzed ESVs using both rarefied incidence data and normalized relative abundance data (as suggested by Nguyen et al. 2015; see Supplementary Figures 2 & 3 for rarefied incidence data analysis). While rarefying has historically been used in sequencing to account for variation in library sizes (number of reads per sample), given the nature of data generated by amplicon sequencing and the lack of accuracy in read abundance as a representation of taxa abundance (Amend et al. 2010), rarefying raw reads can inaccurately skew downstream community analyses (McMurdie and Holmes

2014). Moreover, rarefying throws away valuable taxa (McMurdie and Holmes 2014). Therefore, it is recommended to rarefy incidence data, which yields information about the community as a whole, and uses this in conjunction with normalized relative abundance libraries giving more information about the most abundant species (Nguyen et al. 2015). To rarefy reads, we used rrarefy in vegan v2.5.6 (Oksanen et al. 2019) and rarefied to 1000. To normalize reads, we used the calcNormFactors function in EdgeR v3.11 (Robinson et al. 2010) and normalized reads using the trimmed mean of the M-values (TMM) (Robinson and Oshlack 2010). Normalization corrects for technical errors, variation in sequencing depth, and compositionality generated by NGS, by creating relative abundance data, also known as differential abundance (McMurdie and Holmes 2014). calcNormFactors generates a sample pairing across all samples and calculates each pair’s factor to scale their read counts relative to one another (Robinson et al. 2010). This scaling accounts for technical errors. A limited number of reads are generated each run across samples, and samples vary in their read counts, which are not always correlated with taxon abundance (Robinson et al. 2010). We generated the scaling factors TMM, which trims the mean by 30% on each tail and then weights the mean based on its variance (Robinson and Oshlack 2010). TMM assumes that over half of the samples are not differentially expressed; however, this method is still robust to violations up to 30% below this assumption (Robinson and Oshlack 2010).

Given that microbial communities are dominated by rare species, which we assume are not differentially expressed, this is an appropriate method to normalize reads. Additionally, due to NGS’s potential error associated with rare species with low read counts, we subset our data by removing ESVs with less than 500 reads and re-ran analyses using the normalized relative abundance approach (Weiss et al. 2015). Across

these three analyses, we found that our qualitative results remained the same. We will refer to the analysis on the whole ESV dataset, where normalized, relative abundance data was used, for the rest of our work. We archived sequences at the National Center for Biotechnology Information’s (NCBI) GenBank under BioProject: PRJNA694161.

3.3.7 Statistical Analyses

3.3.7.1 Community Analyses

We performed all our statistical analyses in R v3.6.3 (R Core Team 2020). To test whether we adequately sampled fungal endophyte communities in leaves of parents and offspring, we generated species accumulation curves using specaccum in vegan with 999 permutations (Oksanen et al. 2019). We then calculated the alpha and beta diversity of fungal endophytes within the leaves of parents and offspring using Hill numbers (Hill 1973), including entropy diversity values transformed to effective number of species (Jost 2006, 2007, Gotelli and Ellison 2013). The effective number of species is a metric that explains diversity as equally common species counts (Hill 1973, Jost 2006, Jost 2007, Gotelli and Ellison 2013). We calculated alpha and beta diversity for q- values 0 through 3 using d in vegetarian v1.2 (Charney and Record 2012). Increasing q-values correspond with the down weighting of rare species; thus, q = 0 is equivalent to species richness, q = 1 to the exponential of the Shannon diversity index, and q = 2 to the inverse of Simpson’s diversity (Gotelli and Ellison 2013). The benefit of using effective numbers of species is that it has units that are easy to interpret, it’s intuitive to compare across samples (and datasets), and it retains its doubling property (Gotelli and Ellison 2013).

To test how predictor variables affected fungal endophyte communities, we Hellinger standardized our data (Legendre and Gallagher 2001) and created dissimilarity matrices using the Bray-Curtis index for normalized, relative abundance data (see Appendix for incidence data: Supplementary Figures 2 & 3). We conducted a partial distance-based redundancy analysis (dbRDA) using the function capscale in vegan (Oksanen et al. 2019). We used this constrained ordination twice, once to determine the effect of phytochemical diversity and land-use history on the fungal endophyte communities of parent plants, and again to determine the effect of phytochemical diversity and herbivory on fungal endophyte communities across parent-offspring pairs. We tested the importance of the constrained variables using a Monte Carlo permutation test using anova.cca in vegan (Oksanen et al. 2019) with 999 permutations. To account for the influence of spatial autocorrelation of our parent plants on fungal richness and phytochemical diversity, we generated Bray-Curtis distance matrices with vegdist in vegan (Oksanen et al. 2019) followed by a Mantel test with 9999 permutations using mantel.rtest in the package ade4 v1.7-15 (Dray and Dufour 2007).

We tested for the normality of residuals using a Shapiro-Wilk test with shapiro.test in R base and evaluated residual diagnostic plots (R Core Team 2020). Because residuals were not normally distributed and variance was unequal, we used a Box-Cox transformation on each response variable, total herbivory, and fungal richness. This transformation resolved regression assumption violations. To understand whether fungal endophyte richness, ontogeny, and phytochemical diversity influenced herbivory, we fit two multiple linear regressions using lm (R Core Team 2020): one for total percent herbivory and the other for specialist percent herbivory. We Z-standardized the predictor variables before running the regression using scale (R Core Team 2020),

which centers the mean to 0 and the variance to 1 (Gotelli and Ellison 2013). We checked for multicollinearity in predictor variables using ggpairs in GGally (Schloerke et al. 2020), and they were not collinear (r = 0.085).

3.3.7.2 Metabolomics – Heritability

To determine what compounds might be candidates for high heritability, we created a correlation matrix for all pairwise comparisons of compounds and ran a Random Forest analysis (see Supplementary Figure 5). We generated a correlation matrix with cor (R Core Team 2020), to identify potential heritable compounds. We selected a few compounds that were strongly correlated for parent-offspring regression analysis. Regressions with slopes close to 1 represent pairs of compounds where all (or almost all) of the phenotypic variation is attributable to genetic variance (Connor and Hartl 2004)5. We then ran a simple linear regression on the identified compounds, using lm (R Core Team 2020). Again, we tested for regression assumptions violations the same way we mentioned in the previous paragraph and found that residuals were not normally distributed, and variance was unequal. Despite these violations, other analyses and transformations that could rectify these violations, such as generalized linear models and Box-Cox transformations, would not provide relevant parameter estimates for broad-sense heritability; therefore, we left our regressions untransformed.

5 This is based on the equation for broad-sense heritability, where Vg represents the contribution of genetic variance to phenotypic variation and where Ve is the contribution of environmental variance to phenotypic variation: 퐻2 = 푉푔 푉푔+푉푒

Moreover, because we suspected that heritability in plant chemistry may shape species interactions, we ran simple linear regressions on fungal endophyte richness and herbivory across parents and offspring, using the same methodology listed above.

We suspected that with hundreds of plant chemical compounds, where many may share similar biochemical pathways, the compounds could likely be reduced to several underlying latent variables. To assess this, we standardized each compound using Z-scoring with scale (R Core Team 2020). Standardization is important for factor analysis where variables differ in magnitude, which may artificially increase the importance of a variable (Gotelli and Ellison 2013). To determine how many factors to use for the factor analysis, we used a Very Simple Structure, vss in the psych package v1.9.12 (Revelle 2019). We then ran an exploratory factor analysis, using fa in psych with an orthogonal (varimax) rotation and a minimum residual factor analysis algorithm (Revelle 2019). This algorithm specifies that factors are uncorrelated with one another. We extracted important factors from the parent and offspring factor analyses and ran a series of parent-offspring regressions.

3.3.7.3 Immune Assay

We ran a two-sample t-test on parent plants from the parent-offspring regression experiment and seedlings that were not in the experiment. We used t.test in the base R package (R core team 2020). The data met the statistical test assumptions as residuals were homoscedastic and normally distributed.

3.4 Results

3.4.1 Community Analyses

We found that pooled parent and offspring sequences yielded a combined total of 3,454,085 unfiltered reads, with 3,250,062 merged reads passing filtering. Moreover, we identified 125,688 unique sequences and classified 3,194,812 reads to 3130 ESVs. We removed five chimeras. Neither of the offspring or parent species accumulation curves reached their asymptotes, which indicated inadequate sampling; however, our parent fungal endophyte communities were close to the asymptote (Figure 3.3A, B). Fungal endophyte alpha diversity was six times higher for offspring richness (q0 – richness: mean = 270, SE = 17.9, q1 – Shannon diversity index: mean = 62.4, SE = 3.63, q2 – Simpson diversity index: mean = 25.9, SE = 1.56; ; q3: mean = 17.8, SE = 1.09, Figure 3.3C), than for parents (q0 – richness: mean = 44.2, SE = 2.82; q1 – Shannon diversity index: mean = 20.2, SE = 1.29; q2 – Simpson diversity index: mean = 12.8, SE = 0.9, Figure 3.3C; q3: mean = 10.1, SE = 0.73, Figure 3.3C). Meanwhile, fungal endophyte beta diversity was 1.3X higher for offspring than parents (q0 – offspring: mean = 9.24, parents: mean = 7.33); however, offspring richness was highly influenced by rare taxa, and when these were down-weighted using effective numbers with higher q-values, beta diversity estimates were higher for parents (q1 – offspring: mean = 3.96, parents: mean = 6.19; q2 – offspring: mean = 4.00, parents: mean = 6.68; q3 – offspring: mean = 4.97, parents: mean = 7.98). The majority (53%) of all ESVs (i.e. parents and offspring) were from the division Ascomycota, followed by Basidiomycota at 6% and Chytridiomycota at 1%. The greatest number of ESVs came from the class Sordariomycetes (29%), Dothideomycetes (4%), Agaricomycetes (3%), Eurotiomycetes

(2%), and Leotiomycetes (1%). The top three most abundant ESVs for offspring were taxa from the genera Xylaria (Xylariaceae) and two from Preussia (Sporormiaceae). For parents, the three most abundant fungi were taxa from the genera Pseudocercospora (Mycosphaerellaceae), an undetermined taxon from Basidiomycota, and Glomerella (Glomerellaceae).

To determine the effects of phytochemical diversity and land-use history on the fungal endophyte communities of parent plants, we conducted a partial distanced-based redundancy analysis (Figure 3.4). We found that land-use history had a marginal effect on fungal endophyte communities and explained 16% of the variance in the data (Monte

Carlo permutation test: F(8,48) = 1.14, P = 0.06). Canonical analysis of principal coordinates axis 1 (CAP1) explained 22% of the constrained variance, while CAP2 explained 21%. However, when we removed phytochemical diversity from the model, the model strength improved (Monte Carlo permutation test: F(7,49) = 1.16, P = 0.04), and the model explained 14% of the variation. A Mantel test showed that spatial autocorrelation of parents had an important effect on the phytochemical diversity of the plants (r = 0.28, P = 0.033) but not for fungal richness (r = -0.001, P = 0.34). To assess the consequences of ontogeny, phytochemical diversity, and herbivory on fungal endophyte communities, we performed a second constrained ordination (Figure 3.5). The three predictor variables explained 20% of the variance in the data (Monte Carlo permutation test: F(4,95) = 5.93, P = 0.001). CAP1 explained 87% of the constrained variance in the data, and CAP2 explained 5%.

To assess the heritability of species interaction variables, we ran a series of parent-offspring regressions. We found that total herbivory and fungal richness were not 2 2 heritable (total herbivory: R = 0.010, H = 0.017, F(1,48) = 0.48, SE = 0.025, P = 0.491,

2 2 Figure 3.6D; fungal richness: R = 0.0035, H = 0.023, F(1,59) = 0.21, SE = 0.050, P = 0.649, Figure 3.6C). In addition, we analyzed whether fungal endophytes influenced herbivory as the relationship between fungal endophytes and herbivores can be bidirectional. In a model that evaluated how ontogeny, phytochemical diversity, and fungal endophyte richness influenced herbivory, we found that ontogeny and chemical diversity were the most important predictors influencing specialist herbivory (%) 2 (Multiple R = 0.19, F(3,96) = 7.45, P = 0.0002; Table 3.1), but that this was not the case 2 for total herbivory (%) (Multiple R = 0.02, F(3,96) = 0.73, P = 0.54; Table 3.2).

3.4.2 Metabolomics - Heritability

We also quantified the heritability of plant chemical compounds and sought to identify any heritable latent variables in the plant chemical profile. We found a compound 2 2 (compound 16) to be heritable (H = 1.31, R = 0.70, F(1, 94) = 225, SE = 0.09, P< 2.2e- 16, Figure 3.6A). Another compound identified in our initial pairwise Pearson correlations across all compounds that seemed highly heritable (r = 0.81) was in fact 2 2 strongly heritable as it had a H = 0.9 (R = 0.66, F(1, 94) = 182, SE = 0.004, P< 2.2e-16, Figure 3.6B). A factor analysis indicated 11 latent variables in the metabolomics data for both parents and offspring. From these factors, we identified two sets of latent variables 2 2 that were highly heritable between parents and offspring (H = 0.90, R = 0.66, F(1, 94) = 2 2 182, SE = 0.067, P < 2.2e-16, Figure 3.7A; H = 0.87, R =0.66, F(1, 94) = 184, SE = 0.065, P < 2.2e-16, Figure 3.7B).

3.4.3 Immune Assay

To assess indirectly how changes in plant chemistry due to ontogeny indirectly influence herbivores, we tested the immune response of Eois feeding on seedlings and mature P. sancti-felicis plants. We found that seedlings (mean = 33.89 absorbance/min, SE = 2.62) had an immune response that was 1.5x weaker than Eois reared on mature plants (mean = 50.41 absorbance/min, SE = 3.06) (t = 4.13, df = 23, P = 0.00041) (Figure 3.8).

3.5 Discussion

3.5.1 Summary

We found that offspring alpha diversity of endophytes was much higher than for P. sancti-felicis parents. Beta diversity was greater for species richness (q=0) of offspring compared to parents, but smaller for effective numbers with q-values greater than 0 (Figure 3.3). Land-use history was an important predictor of fungal endophyte communities for parents (Figure 3.4), and fungal endophyte communities differed greatly between parents and offspring due partially to differences in phytochemical diversity and herbivory (Figure 3.5). We also sought to quantify how fungal endophytes and plant chemistry influence herbivory as plant chemistry and fungal endophytes can reciprocally influence one another. We found that fungal endophyte richness did not influence total or specialist herbivory, but that phytochemical diversity and ontogeny influenced specialist herbivory (Tables 3.1, 3.2 & 3.3). It seems that total herbivory and fungal endophyte richness were not heritable across parents and offspring (Figure 3.6). However, two independent compounds (Figure 3.6) and two sets of latent variables in the metabolomics data identified from a factor analysis were strongly heritable (Figure

3.7). Finally, we found that differences in plant chemistry due to different plant ages had a strong influence on specialist herbivores’ ability to mount an immune response (Figure 3.8).

3.5.2 Parent land-use history influenced fungal endophyte communities but not phytochemical diversity.

Contrary to our predictions, we found that the two hypothesized ecological filters (i.e., phytochemical diversity and land-use history) only weakly influenced fungal endophyte communities. Also, phytochemical diversity and land-use history were spatially autocorrelated. Spatial autocorrelation means that parent plants in the same land-use history locations had more similar phytochemical diversities than plants in different locations. These variables together did not have a strong effect on the fungal endophyte communities. Once we removed phytochemical diversity from the model, the importance of land-use history as a constrained variable improved the model’s strength. Thus, land-use history, and not phytochemical diversity, seems to be an important ecological filter in parent plants, even though plant chemistry more generally is an important ecological filter for foliar fungal communities in other plant systems (Arnold et al. 2003a; Saunders and Kohn 2009, Christian et al. 2020). Individual plant compounds or compound classes may be more influential in the community assembly of fungal endophytes, such as antifungal compound classes like amides. We included phytochemical diversity in our model as we know that it is an important metric in other species interactions (Richards et al. 2015, Slinn et al. 2018). However, no other research has looked at how phytochemical diversity influences fungal endophyte communities. Our final results are consistent with what other research has found, particularly in crop-affiliated microorganisms. In an agricultural setting, land-use history

is typically defined as a farming management strategy, which includes the spectrum between organic and conventional farming, and that this can have a small (aboveground: Gdanetz and Trail 2017) to large (belowground: Schlatter et al. 2020) effect on crop microbial communities. For example, both the diversity and abundance of fungi and bacteria found in the wheat rhizosphere were influenced by land-use history, although bacteria were more strongly influenced than fungi (Schlatter et al. 2020). Despite these changes to the microbial community, there was still a core microbiome present across the land-use history treatments (Schlatter et al. 2020).

Meanwhile, the importance of land-use history on species interactions is seen in agricultural systems. Estendorfer et al. (2017) found that in orchardgrass (Dactylis glomerata), land-use history, measured by the intensity of fertilization, mowing, and grazing, had a large effect on bulk and rhizosphere soil microorganisms, but less so on root endophyte communities. Furthermore, they found that this disparity between the microbial communities in the different types of samples became more similar as land- use intensity decreased (Estendorfer et al. 2017). While our results are not in complete agreement with our predictions, we find that land-use history is an important ecological filter for foliar fungal endophyte communities in our system, which is valuable information in understanding how anthropogenic changes to our environments, such as land-use change, alter the microbiome of plants (Li et al. 2019).

3.5.3 Lack of heritability in species interactions but not in secondary metabolites

Given the strong previous evidence for heritability in plant secondary metabolites in other systems (Geber and Griffen 2003, Hansen et al. 2011, Moore et al. 2014), we expected to find the same results in P. sancti-felicis, at least for some individual

compounds. And this is indeed what we found. Moreover, we also found that latent variables of secondary metabolites, identified from a factor analysis, showed heritable variation. As we have yet to identify the compounds in our data, we cannot say whether those latent variables are the same groups of compounds that are exhibiting heritable variation across parents and offspring. If, for instance, compound classes demonstrate heritable variation, then this would be particularly interesting for the evolutionary defense implications for P. sancti-felicis. For example, it could be that the factors we identify correspond to a metabolic process, for instance the production of an enzyme, that is involved upstream in the metabolic pathway. The support of heritability in individual compounds is supported by previous research on Piper that shows just how variable reaction norms in plant chemistry are (Dyer et al. 2004b, Glassmire 2017).

We expected to find that measures of interaction intensity would have high heritability, particularly with herbivory, as heritability in herbivore response has been found in other plant systems where plant chemistry also exhibits strong influences on herbivorous insects (Barbour et al. 2015). And we know that phytochemical diversity is important in shaping herbivore communities and herbivory in Piper (Richards et al. 2015). However, we found no evidence for heritable variation in the amount of herbivory plants received, and this may suggest that phenotypic plasticity in PSMs is more important than genetic variance (Barbour et al. 2019). Barbour et al. (2019) showed that phenotypic plasticity generated by environmental variation in plant traits, including plant chemistry, was better at explaining arthropod species richness. It’s possible that the measure of herbivore response that we chose (herbivory: total, generalist and specialist) did not adequately reveal heritable variation in herbivore response. For example, perhaps feeding guilds would have been a better measure of herbivore

community response (such as Barbour et al. 2015, 2016) rather than grouping herbivores into generalists and specialists. Or it may be that in P. sancti-felicis, other traits are more important at shaping herbivore communities. A meta-analysis showed that in contrast to commonly accepted knowledge about PSMs, secondary metabolites are not among the most important predictors of insect herbivores. Instead, it is life history traits and then morphological traits (Carmona et al. 2011). For microbes, we were not sure what we would find for heritable variation in fungal endophyte richness or diversity, as very little research exists on the fungal endophyte communities of Piper. Still, we know that host-plant genotype can play a big role in the community assembly of plant-associated microbes like fungal endophytes (Peiffer et al. 2013, Wagner et al. 2016). While we didn’t find heritable variation in fungal endophyte richness, other microbial parameters may be at play here that we did not measure, such as microbial abundance (Walters et al. 2018).

3.5.4 Specialist herbivory was influenced by phytochemical diversity and ontogeny but not total herbivory

Interestingly, we found that specialist herbivory was influenced by plant chemistry and parent-offspring group, but not total herbivory, which included generalists’ damage. This result was not a surprise, as I noticed that offspring were dominated by damage from specialists while we harvested the experiment (personal observation). This result is in line with what we would expect given that the offspring, despite having taken these measurements at maturity, were likely much younger than the parent plants. Leaves appeared younger still, and less dark (personal observation). Perhaps, this result can be explained by the plants’ age, with specialists better able to consume young, highly defended plants in comparison to generalists, which may avoid the plants until leaves

are more mature. Other research on Piper supports this result. Increasing phytochemical diversity decreases overall herbivory, but increases the damage done by specialists, which suggests that specialists are better adapted to deal with the phytochemical profile (Richards et al. 2015). This is further supported by the immune assay that we conducted where younger leaves had a larger negative effect on Eois immune response than did the mature leaves.

3.6 Conclusion

Overall, our study contributes meaningfully to the literature on the chemical ecology of Piper and to our understanding of the heritability of an important functional trait and species interactions in a diverse tropical genus that is dominant in understory communities (Dyer and Palmer 2004, Dyer et al. 2018). Future elucidation of the compounds in our study will help us interpret the heritability we found in the latent variables of our metabolomics data. It’s essential to untangle the contributions of genetics and the environment to phenotypic plasticity and reaction norms to understand the evolution of these species and their symbionts and to predict how adaptable these communities will be in the face of global change (Balint et al. 2015, Vacher et al. 2016).

3.7 Tables

Table 3.1: Multiple linear regression on specialist herbivory that was Box-Cox transformed (Multiple R2 =

0.19, F3, 96 = 7.45, P = 0.002). Predictor variables were Z-standardized.

Table 3.2: Multiple linear regression on generalist herbivory that was Box-Cox transformed (Multiple R2 =

0.022, F3, 96 = 0.73, P = 0.54). Predictor variables were Z-standardized.

3.8 Figures

Figure 3-1: A meta-model of the interacting variables within the experiments which shows how plant chemistry, its heritability, and land-use history act as predictors in our experiments.

Figure 3-2: A map of the land-use history at La Selva Biological Station created in 2005 from the Organization of Tropical Studies.

that Hill’s number is raised to. q = 0 is species richness, q = 1 is the exponential of Shannon’s diversity and q = 2 is the inverse of Simpson’s diversity (Gotelli and Ellison 2013). ‘q’ values are positively related to the emphasis on abundant species.

developed_area

0.5 Land use history abandoned plantation shaded_pasture ) current plantation

% phytochemical_diversity logged_forest

4 developed area

. 0

2 open pasture ( 0.0

2 secondary forests 10−21 years P

A current_plantation

C select logged forests secondary_forest shaded pasture open_pasture successional plots −0.5

successional_plots

−1.0 −0.5 0.0 0.5 CAP1 (21.91%)

generalist_herbivory specialist_herbivory 0.5

) 0.0 parent %

7 Groups

. 5

( offspring

P parent

A C

−0.5

−1.0 phytochemical_diversity

−1.0 −0.5 0.0 0.5 CAP1 (86.98%)

Figure 3-5: A partial distance-based redundancy analysis evaluating ontogeny, phytochemical diversity, and herbivory on fungal endophyte communities (Monte Carlo permutation test: F(4,95) = 5.93, P = 0.001). Predictor variables explained 20% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 87% of the constrained variance in the dissimilarity matrix, while canonical analysis of principal coordinates axis 2 (CAP2) explained 5%. Black data points represent fungal endophyte taxa, and larger coloured data points represent ontogenetic group scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by ontogeny. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013).

2 2 16 (H = 1.31, R = 0.70, F(1, 94) = 225, SE = 0.09, P< 2.2e-16). B) displays another heritable compound

2 2 (H = 0.9, R = 0.66, F(1, 94) = 182, SE = 0.004, P< 2.2e-16). C) shows how foliar fungal richness is not

2 2 heritable (H = 0.35, R = 0.0031, SE = 0.81, F1,59 = 0.19, P = 0.67). D) demonstrates how total herbivory

2 2 (%) is not heritable (H = 0.13, R = 0.0012, F1,48 = 0.58, P = 0.45).

0.0010

s 0.0008 e

r 0.0006 o

c 0.0004

0.0002

r 0.0000

o t

−0.0002 c

a −0.0004

−0.0006

g n

i −0.0008 r

p −0.0010

s f

f −0.0012 O −0.0014 −0.0014 −0.0012 −0.0010 −0.0008 −0.0006 −0.0004 −0.0002 0.0000 0.0002 0.0004 0.0006 Parent Factor 10 scores B

s 0.0025

r o

c 0.0020

0.0015

t c

0.0010

g 0.0005

r p

0.0000

f O −0.0005

−0.0004 −0.0002 0.0000 0.0002 0.0004 0.0006 0.0008 0.0010 0.0012 0.0014 0.0016 0.0018 0.0020 Parent Factor 3 scores

Figure 3-7: Parent-offspring regression for factor scores from minimum residual factor analysis with

2 2 2 orthogonal (varimax) rotation A: (H = 0.90, R = 0.66, F(1, 94) = 182, SE = 0.067, P < 2.2e-16). B: (H =

2 0.87, R =0.66, F(1, 94) = 184, SE = 0.065, P < 2.2e-16). Error band represents standard error.

)

(

3 0 2 mature seedling Plant life stage

4 Digestion by a Specialist Bat Cleans Fungi from Seeds of the Neotropical Shrub, Piper sancti-felicis

4.1 Abstract

In the tropics, endozoochory is the dominant dispersal strategy for plants. Moreover, fungi are predominantly limited by dispersal, and the co-dispersal of fungi with diaspores through endozoochory, found in recent research, has important implications for plant fitness and future seedling success. In our study, we sought to understand how digestion by a frugivorous bat, Carollia, which specializes on Piper sancti-felicis, modified the fungal community of the seed and whether scarification occurred. We investigated whether the dominant infructescence compounds, termed alkenylphenols, contained antifungal properties. We found that digestion cleaned the seeds of fungi, and that scarification did not appear to occur, but fungi still managed to colonize the interior of the seed coat. Despite these results, the community composition of fungi was not affected by the treatment. Finally, we found that the alkenylphenols isolated from ripe infructescences contained antifungal properties and reduced, but did not completely eliminate, the presence of fungi on seeds within the fruit. Our research provides an interesting and novel insight into the roles that fruit chemistry and endozoochory play in plant defence against potential fungal pathogens.

4.2 Introduction

4.2.1 Dispersal is Critical for Terrestrial Communities

Understanding the natural history of mutualistic interactions between plants and their dispersers is critical for terrestrial communities (Howe and Smallwood 1982, Schupp et al. 2010), as they can have profound impacts on gene flow (Ouborg et al. 1999, Jordano et al. 2007), plant fitness, and plant community composition (Traveset et al. 2014). Plants have evolved a variety of strategies to disperse their offspring, using either abiotic (wind: anemochory; water: hydrochory) or biotic factors (through facilitation with animals: zoochory; Bewley et al. 2013). Some plants have adapted improved efficiency of dispersal by providing fruit, which is high in nutrition, to birds and other mammals (i.e. endozoochory), who digest and defecate seeds across the landscape (Bewley et al. 2013). In the tropics, endozoochory is the dominant dispersal strategy, where ≥50% of shrubs and trees have adaptations to promote animal-mediated dispersal (Fleming 1987). This type of dispersal, particularly by small birds and mammals, helps minimize intra-specific competition for shared resources and consequently high densities of their natural enemies (Janzen 1970, Connell 1971). While biotic interactions are an important factor determining seedling establishment and success, abiotic conditions (i.e. environmental filters), such as humidity, temperature, and light availability where the diaspore is deposited, determine whether diaspores are able to germinate and establish in that location (Jordano et al. 2007, Kraft et al. 2015). Taken together, whether a diaspore is successful depends on a complex set of interactions between abiotic and biotic variables.

4.2.2 Ecological Filters of Fungal and Seed Fungal Communities

When abiotic and biotic factors are both investigated to determine what affects the establishment of an organism in a community, they are collectively described as ecological filters (Keddy 1992, Myers and Harms 2009). One particularly important ecological filter for fungi is dispersal limitation (Peay et al. 2016, Bruns 2019). Even research studying the seed microbiome showed that the seed fungi were strongly limited by dispersal (Rezki et al. 2018). Moreover, endozoochory is a type of ecological filter, which influences the fungi that are able to colonize diaspores. For instance, arbuscular mycorrhizal fungi can be transported with their associated diaspores through the gut of a bird disperser (Correia et al. 2019). In addition, ecological filters that affect the diaspore fungal community may improve germination and seedling establishment (White et al. 2019). Digestion may alter the fungal community of diaspores by promoting the colonization of the interior of the seed coat through seed scarification (White et al. 2019). Scarification damages the seed coat, which allows water and air to pass through, and is known to increase germination in some species, but it may also allow for the establishment of fungi (White et al. 2019). The colonization of the interior of seeds by fungal endophytes has been associated with a co-evolved mutualism between plants and fungi. For example, in some cool-season grasses (Clay 1988, Newman et al. 2020) and legumes (Panaccione et al. 2014), chemically mediated mutualistic interactions occur where fungi produce defensive compounds for the host plant and are passed from parent to offspring during seed development. Factors that alter fungal community assemblages can affect plant fitness by providing protection against antagonists, by removing pathogens and increasing germination (Nelson 2018).

In addition to endozoochory and dispersal, fungal communities, and in particular endophytes, can be strongly affected by the chemical environment of the host plant (Arnold et al. 2003a, Schulz and Boyle 2005, Saunders and Kohn 2009, Van Bael et al. 2017). Many plants produce defensive compounds to protect themselves against enemies, such as herbivores and fungal pathogens (Fraenkel 1959, Levin 1976), especially in the tropics (Coley and Barone 1996). Antifungal compounds produced to protect the plant from fungal pathogens are expected also to filter the fungal endophyte community as fungal endophytes are closely related to fungal pathogens (Arnold et al. 2009). For example, work on spotted locoweed (Astragalus lentiginosus) found that the concentration of swainsonine, a toxic bioactive compound produced by a systemic fungal endophyte Alternaria fulva, reduced the richness of the fungal endophyte community (Harrison et al. 2018a). Moreover, fungal endophytes often trigger the plant’s immune response when they colonize, which causes immediate chemical changes to the interior of the leaf (Mejía et al. 2014, Busby et al. 2015). Fungal endophytes can both shape and be shaped by host plant chemistry.

Understanding ecological filters (Antwis et al. 2017), including animal digestion (Nelson 2018, Hardoim 2019) and plant chemistry (Van Bael et al. 2017), is an ongoing and important area of research in microbial ecology as it helps researchers to understand the roles and interactions of fungi in the community. Ecological filters like plant genotype (Balint et al. 2013, Busby et al. 2013, Wagner et al. 2016) are well known to influence fungal endophyte communities in temperate plants. However, it’s important to understand how endozoochory (Nelson 2018, Shahzad et al. 2018, Hardoim 2019) and plant chemistry (Van Bael et al. 2017) affect seeds and their fungal communities because of the impacts they have on plant fitness. Simultaneously,

research on fungal endophytes in the tropics, where plant and fungal endophyte diversity are highest, has focused on characterizing fungal endophyte communities with fewer experimental manipulations (but see, Arnold et al. 2003a, Christian et al. 2017, Sarmiento et al. 2017).

4.2.3 Objectives

There is a fascinating potential co-evolutionary relationship between Piper (Piperaceae) and its specialist bat seed disperser Carollia (Phyllostomidae). Carollia become active at dusk, at the same time that Piper fruit ripen each day (Fleming 2004). By morning, nearly all of the ripe fruit have been removed, and fruit that ripened and remain on the plant are rotting and have fungal growth (Fleming 2004, personal observation). Interestingly, there is a brief window in time, where P. sancti-felicis fruit are healthy and ripe, and this coincides with the exact time Carollia begin their evening foraging.

We used the neotropical plant, Piper sancti-felicis, and three species of Carollia at a lowland rainforest site in Costa Rica (La Selva, Sarapiqui). We were interested in how digestion by a Piper specialist would influence putative seed success variables and how plant chemistry may act as an ecological filter (Figure 4.1). We sought to document changes to the seed fungal community and the seed coat, as digestion in some study systems increases germination through scarification (Traveset and Verdú 2002, Naranjo et al. 2003, Lobova et al. 2003). Moreover, we were interested in whether fungi colonized the interior of seeds and included a treatment where seeds sat in guano for one week before being surface sterilized. However, we suspect that being secreted in guano is only one way that fungi may colonize the interior of seeds; scarification may allow for colonization as well (White et al. 2019). We asked three questions: 1) How

does digestion of a seed by Carollia alter the seed fungal community of both the seed surface and the interior of the seed? 2) Does the fruit of P. sancti-felicis exhibit antifungal properties? And 3) How does digestion alter seed micro-structures? For Q1, we predicted that: a) digestion by Carollia would increase the abundance and richness of the fungal community because the gut and guano are rich in fungi and other microbes, and b) (Q3) damage to the seed coat that occurs from digestion (i.e. scarification) would enable the interior colonization of the seed by fungi. We also predicted (Q2), that according to optimal defence theory, fruit would contain anti-fungal properties to protect seeds.

4.3 Materials and Methods

4.3.1 Piper-Carollia System

P. sancti-felicis is a shrub and a low-land rainforest, monoecious species, found commonly in disturbed areas with abundant sunlight or on the edge of forests and in some canopy gaps. Its fruit are called infructescences (referred hereafter as fruit) because they are made up of over a thousand small bisexual flowers on a single spike- shaped inflorescence, where every flower produces a seed (Greig 2004). P. sancti- felicis is pollinated by bees and flies, and is also able to reproduce clonally, through the cutting and planting of stems (Greig 1993b). At La Selva Biological Station, this is a common species which can produce a lot of fruit on a single shrub when it is located in bright areas (Greig 1993b) and where the fruit take approximately 6-8 months to reach maturity (Fleming 2004). P. sancti-felicis produces fruit continuously throughout the year (Thies and Kalko 2004).

In our study, we captured6 three different Carollia species (C. perspicillata: n=10, C. castanea: n=11, and C. sowelli: n=1). All three species are specialists of neotropical Piper, but the diet of the three species varies between ~45% to 80% Piper, depending on the species of Carollia (Fleming 1991, Lopez and Vaughan 2007). Carollia disperse Piper seeds across the landscape as they disperse seeds during flight as well as near their roosts (Fleming 2004). Dispersal from Carollia have been shown to increase Piper species diversity around their roosts, therefore altering community composition in these habitats and reducing overall herbivory of leaves, which may improve plant fitness (Salazar et al. 2013). To conduct this experiment, we caught Carollia adults with mist nests and fed them one half of a ripe fruit in the field in June 2018. The other half of the infructescence was used as the control. Bat digestion occurs quickly (Fleming 2004), and bats excreted seeds 5-10 minutes after feeding. After guano was excreted, bats were released. We rinsed seeds with sterile deionized (DI) water to remove fruit pulp and guano before plating them on malt extract agar (4 seeds/plate from the same fruit).

4.3.2 Experimental Treatments

We manipulated seed digestion, seed surface sterilization and time seed remained in guano. We used a surface-sterilization treatment to assess whether fungi were colonizing the interior of seeds and, if so, how the interior seed fungal community varied from the surface community across the digestion treatment. We collected undigested

6 Lauren Maynard caught the bats for the experiment. She had the following permit: Costa Rica Ministry of the Environment, Energy and Technology, permit number 012-2018-ACC-PI.

control seeds directly from mature plants and plated them. We then used two treatments: 1) digested, unsterilized seeds, which we collected from guano and plated immediately, and 2) digested, surface-sterilized seeds which we let remain in the guano for one week before we plated them. Our pilot experiment showed that undigested, sterilized seeds carried no fungi within the seed coat, and we did not further pursue this treatment in the full experiment (unpublished data, Heather Slinn). We conducted a series of surface sterilization in the following washes: 10 seconds in 95% ethanol, 2 minutes in 0.7% bleach; 2 minutes in 70% ethanol; rinsed with sterile DI water (Sarmiento et al. 2017). We rinsed tweezers in 95% ethanol and then flamed them in candle in between samples to avoid contamination. We imprinted seeds on malt extract agar plates to check that our surface sterilization worked and waited one week at 26oC.

4.3.3 Culture-Based Identification of Fungi

In order to isolate fungi, we plated four seeds from each fruit on 2% malt extract agar (ThermoFisher Scientific Oxoid Malt Extract) at 26oC for 6-7 days and then sub-cultured them to ensure pure cultures. Because the majority of seeds yielded either one or no fungi, the number of sequences generated per seed were either one or zero. We performed DNA extractions (QIAGEN: DNeasy Plant Mini Kit) and Sanger sequenced the internal transcribed spacer (ITS) region (ITS1, 5.8S, ITS2) between the ribosomal RNA genes using primers ITS1F (5’-CTTGGTCATTTAGAGGAAGTAA-3’) and ITS4 (5’- TCCTCCGCTTATTGATATGC-3’) (University of Guelph: Laboratory Services – Oligo Synthesis and Sequencing) (White et al. 1990, Gardes and Bruns 1993). In August of 2018, we performed DNA extractions at La Selva Biological Station and ran PCR. To extract DNA, we scraped the mycelia from the surface of one plate, which was then weighed and grinded using a tissue lyser for 2 minutes. During the final step of DNA

extraction, we eluted genomic DNA to 100μl. We used negative controls (buffer with PCR mix) with every PCR run. We made a master mix for 50μl reactions with 3μl of genomic DNA using 5μl of 10x DreamTaq Buffer (Thermo Fisher Scientific), 1μl of 10mM dNTPs (University of Guelph: Laboratory Services – Oligo Synthesis and Sequencing), 2.5μl at 10 μM each of forward and reverse primers, 0.25μl of 5U/μl DreamTaq Polymerase (ThermoFisher Scientific), and 35.75μl of PCR water. Stage 1 of PCR included 94oC for 1 minute; Stage 2 was 94oC for 1 minute, 51oC for 1 minute, then 72oC for 1 minute. We repeated this stage 35 times. Stage 3 consisted of 72oC for 8 minutes and then a hold at 4oC. We stored amplicons at -20oC until they were transported to the University of Guelph in March of 2019.

In preparation for Sanger Sequencing, we purified amplicons (Thermo Fisher Scientific: Invitrogen PureLink PCR Purification Kit) and quantitated DNA with a spectrophotometer (Thermo Fisher Scientific: NanoDrop 2000). We performed the next round of PCR and Sanger sequencing at the University of Guelph’s Advanced Analysis Centre Genomics Facility with Big Dye Terminator v3.1 (Thermo Fisher Scientific) on the Applied Biosystems 3730 DNA analyzer (Thermo Fisher Scientific). We used ITS1F and ITS4 primers with 1μl at 10pmol/μl generating both forward and reverse reads. We created a master mix for approximately 12μl reactions with 1μl of BigDye, 2μl of 5X SeqBuffer, and 9μl of PCR water. Based on DNA quantitation mentioned above, we calculated and standardized volumes of amplicon to be 28ng/1kb. PCR cycling conditions included: Stage 1 at 96oC for 2 minutes, Stage 2 at 96oC for 30 seconds, 45oC for 15 seconds, and 60oC for 4 minutes. We repeated stage 2 29 times before moving to stage 3, which we held at 10oC. We then purified sequences with Sephadex columns (Sigma Aldrich) and passed them to electrophoresis. We performed basecalls

with the v5.2 KB Basecaller (Thermo Fisher Scientific) and assigned ambiguous bases to any base with a QV (phred score) of 12 or less, yielding a minimum confidence of 95% for base calls. We trimmed primers and generated consensus sequences combining the forward and reverse reads for each sample using CodonCode Aligner v8.0.2 (CodonCode Aligner Company).

We pooled Sanger sequences from cultures and assigned them to exact sequence variants (ESV) using the derep_fulllength function in USEARCH v11.0.667 (Edgar and Flyvbjerg 2015). We did this because some fungal cultures exhibited minor variation in morphological features but yielded the same ESV. Due to the nature of Sanger sequences having a read depth of 1, dereplicating sequences was all that was needed to generate ESVs. Following dereplication, we used the Basic Local Alignment Search tool (BLAST, Altschul et al. 1990) to align our sequences with the UNITE (v8.2) reference database (Kõljalg et al. 2005, accessed March 13thth, 2020), rather than directly through other international sequence databases, such as the National Center for Biotechnology Information’s (NCBI) GenBank, because of the high percentage of errors found in them (Nilsson et al. 2006). UNITE adds additional layers of quality control, where it checks for low-quality reads and is curated by fungal taxonomists, bioinformaticians, and ecologists (Nilsson et al. 2019). UNITE is based on the GenBank sequence database, where it is updated regularly with new sequences (Nilsson et al. 2019). Our Michrodochium lycopodinum sequences aligned with a sequence from a type material based on a curated fungal collection, and therefore we are highly confident about the genus classification despite a 96-98% identity match. We refrained from classifying our taxa to species using UNITE’s species hypothesis algorithm because it uses the ITS2 region to approximate species classification. Moreover, even

the full ITS region does not always distinguish equally well across different species (Schoch et al, 2012). However, we have included the species hypotheses and their digital object identifiers in Supplementary Table 4 as they match genus-level identification and are linked to ecological data on the taxa and other probable sequence alignments. Finally, we deposited our sequences at GenBank under accession numbers MT093652 – MT093654 (Supplementary Table 4).

To determine whether treatments (digestion and sterilization) differed in the number of seeds that produced fungi (0 = no fungi present or 1 = fungal growth present), we used one logistic regression on presence and absence of ESVs using glm in the base package of R v3.6.0 (R Core Team 2019). We calculated beta diversity using vegdist in vegan (Oksanen et al. 2019) using the Bray-Curtis dissimilarity matrix. We performed a community analysis in order to determine how taxa changed across treatments. We ran a partial distance-based redundancy analysis with capscale in vegan (Oksanen et al. 2019) using a Bray-Curtis dissimilarity matrix. We used a Monte Carlo permutation test to assess whether treatment affected fungal communities using the anova.cca function in vegan (Oksanen et al. 2019). We used the FUNGuild database (accessed March 16th, 2020) to assign taxa to ecological guild (Nguyen et al. 2016).

4.3.4 Microdilution Assay

We used a microdilution assay to assess antifungal properties of alkenylphenols isolated from P. sancti-felicis fruit. We modified the Zgoda and Porter (2001) protocol for our research question. Details on the modifications are in the following paragraphs. We conducted the microdilution assay in September 2018.

We used the three most common fungi to make a liquid media culture, in order to promote the production of blastopores (asexual spores), which enables the standardization of fungal concentration. We harvested mycelia from agar plates made in the culture-based portion of the experiment by adding 1ml of sterile water and probing the culture with the end of a tip to dislodge fungi. We stored fungi in a sterile Eppendorf tube at 4oC until needed.

We made 2% malt extract broth using 100ml dH2O and 2g of malt extract (BD Bacto malt extract). We mixed the solution for a few minutes and then autoclaved it at 120oC for 20 minutes. Once the malt extract broth cooled, we added 100μl of the inoculum to 75ml of the broth and incubated it on a shaker at 200 rpm for 4 days at room temperature. We filtered hyphae using sterile cheese cloth and counted blastospores using a hemocytometer to standardize fungi to 1 million spores/ml.

To extract alkenylphenols, we collected P. sancti-felicis fruits locally, oven dried them at 60oC, and then ground and extracted the alkenylphenols in 100% ethanol. We removed the solvent on a rotary evaporator, resuspended the extract in a water and ethanol (3:1) solution and then performed multiple ethanol partitions. We took the ethanol layers from the partitions and evaporated them using the rotary evaporator. We resuspended the extracts in 100% ethanol and transferred them to scintillation vials. Because these dried extracts were not pure alkenylphenols, we ran an internal standard with a known concentration against our extract, to determine the concentration of alkenylphenols found in a single infructescence.

To determine changes of fungal concentration, we used a spectrophotometer (BIO-RAD: iMark Microplate Absorbance Reader) at 450nm with 96-well plates and measured absorbance (in optical density). We ran a preliminary trial with the different wavelengths available on the microplate reader to select the optimal wavelength for the experiment by generating a series of absorbance curves (415nm, 450nm, 490nm, 520nm, 595nm, 655nm, 750nm). We selected 450nm as it was the best at detecting changes in inoculum concentration for our experiment. Due to minor variation in microplate reader measurements, we took two absorbance measurements, one right after the other, and averaged them to attempt to control this source of variance. We took measurements at 0 hours, 24 hours, and 72 hours. In between measurements, we stored 96-well plates in sterile ziplock bags with damp sterile filter paper to maintain humidity.

The microdilution assay included 8-serial dilutions. Each well received: half of the amount of alkenylphenols compared to the previous well, fungal inoculum suspended in sterile DI water, 2% malt extract to provide nutrients for fungal growth, and sterile DI water. The total volume of all wells was 200μl. First, each of the eight wells received 100μl of sterile DI water. Next, well one received the highest amount of extract with 5μl of 73.15 mg/ml alkenylphenols and an additional 195μl of water. We mixed the water and extract by pipetting and then 100μl of this 200μl solution was aliquoted into well two. We continued this process of mixing newly aliquoted extract into 100μl of water in subsequent wells, across the wells, until the eighth well where the 100μl taken was discarded. Once the wells had the appropriate gradient of alkenylphenols, each well received 80μl of 2% malt extract and 20μl of fungal inoculum at 1 million spores/ml. The final concentration of our extract at the highest concentration of the serial-dilution was

0.91 mg/ml, which is approximately equivalent to 6.2% of the average alkenylphenols concentration found in a single ripe infructescence. We included a sterile control with no fungi and no alkenylphenol extract, but it did include 95μl of water, 80μl of 2% malt extract, and 25μl of 100% ethanol. We also included a negative control where 20μl of fungal inoculum was added but alkenylphenols were not. These wells then received 95μl of water, 80μl of 2% malt extract, 5μl of 100% ethanol in addition to the inoculum.

We analyzed the microdilution assay in R v3.6.0 (R Core Team 2019) using a linear mixed-effects model with lme4 v1.1-21 (Bates et al. 2015) where the fungal taxa were used as fixed effects because we were interested in how the result differed across taxa (Harrison et al. 2018b). Our response variable was absorbance (units: optical density (OD)), which reflects the concentration of fungal hyphae suspended in solution in the well. We used concentration of alkenylphenols as a fixed effect because it was our main variable of interest. We ran our model with an interaction between fungal taxa and alkenylphenols concentration. We fully crossed fungal taxa, meaning that each taxon was present across two time periods (24h, 72h). We used time as a random slope as we were interested in extrapolating our findings beyond these two times (Harrison et al. 2018b). Moreover, we used the random slope over the random intercept model because the slopes and intercepts of absorbance varied across times. A Shapiro-Wilk’s test indicated that residuals were normally distributed. Predictors shared low values of collinearity (r < 0.5) which were not a concern for the precision of coefficient estimation (Harrison et al. 2018b).

4.3.5 Scanning Electron Microscopy

We took a separate bat sample with two of the same Carollia species (C. perspicillata: n=3, and C. sowelii: n = 1) (digested seeds: n = 18, undigested seeds: n = 16), in November 2019 to assess how digestion affected seed micro-structures. We used scanning electron microscopy (SEM) on digested and undigested seeds to determine if scarification occurred. We removed fruit pulp, rinsed seeds with sterile water, and then dried them in a laminar flow hood for three hours. We transferred seeds to a hermetic jar containing silica gel desiccant and then dried them again in silica for 48 hours. We mounted seeds on aluminum bases using double-sided carbon tape and covered them with gold in an IB3 ionic cover (Eiko Engineering Co. LTD., Japan). Lastly, we used the electron microscope Hitachi S-3700 to obtain digital images of the micro-structures of the seed coats.

4.4 Results

4.4.1 Culture-Based Seed fungi

The majority of seeds hosted either one or no culturable fungi (Table 4.1). Only two seeds yielded two fungi each. We found that out of all the seeds we plated for culturing, undigested, unsterilized seeds produced the greatest proportion of fungi, with 32% of the seeds in this treatment culturing fungi; digested, unsterilized seeds yielded 12%, and digested, sterilized treatment yielded 2% (Figure 4.2, Table 4.2). Digested, unsterilized seeds had ~70% fewer fungi than control seeds, and digested, sterilized seeds had ~94% fewer fungi than control seeds. The majority of seeds in each treatment produced no culturable fungi. Across treatments, there were 15 different taxa

that were isolated (Table 4.2). Eleven of these taxa occurred only once and made up 22% of all isolated fungi (Table 4.2). Nearly all fungi were classified to Ascomycota, with one taxon from Basidiomycota and one from Mucoromycota. The most dominant fungus that was isolated was Microdochium lycopodinum (Ascomycota), which is listed as an endophytic fungi and plant pathogen (Tedersoo et al. 2014, Nguyen et al. 2016) (Table 4.3); it occurred in 19 different seeds, which made up 43% of all seeds that produced fungi (Table 4.2). This taxon occurred at least once across all treatments (Table 4.2). The next most abundant taxon was Fusarium (Ascomycota), which occurred 10 times and which consisted of 23% of the isolated fungi (Table 4.2). Fusarium is a common animal and plant pathogen, endophyte (Blacutt et al. 2018), and soil and wood saprotroph (Nguyen et al. 2016) (Table 4.3). It is also a relatively common seed endophyte and seed pathogen (Rodrigues and Menezes 2005).

In order to assess whether treatment effects varied in the number of seeds that produced fungi, we used a logistic regression. We found that there was a large difference (3.6–18 times) between treatments in the number of fungal taxa they produced (Table 4.4). To evaluate whether taxa changed across treatments, we calculated beta diversity estimates, which were large (B = 0.78-0.95; Table 4.5), and we used a partial distance-based redundancy analysis and a Monte Carlo permutation test. However, the constrained ordination showed that our treatments did not affect the identity of taxa, where our two constrained axes (CAP1 and CAP2) only explained 7% of the variation in the data (Figure 4.3, Monte Carlo permutation test: F(2,22)= 0.81, P = 0.66). Additionally, CAP1 explained 60% of the constrained portion of variance, and CAP2 explained 40% (Figure 4.3).

4.4.2 Microdilution Assay

To determine how alkenylphenols affected the growth of fungi over time, we ran a linear mixed-effects model. We found that the percent dry weight of alkenylphenols strongly decreased fungal growth (coef = -0.17, SE = 0.07, P = 0.02), but that this depended on the fungi (alkenylphenols concentration x Fusarium A: coef = -0.19, SE = 0.09, P = 0.04, alkenylphenols concentration x M. lycopodinum: coef = -0.23, SE = 0.09, P = 0.01) (Figure 4.4, Table 4.6). The estimated variance of the random effect, which was time, was 0.00093 OD.

4.4.3 Scanning Electron Microscopy

Based on a qualitative assessment, we found no difference in the intactness of crystal structures, which we suspect are calcium oxalate (CaOx), and no difference in other morphological features between digested and undigested seeds (Figure 4.5) (Ilarslan et al. 2001, Nakata 2002).

4.5 Discussion

4.5.1 Summary

We found that 32% of seeds in the control (undigested, unsterilized group) produced fungal taxa, which was more than both of the treatment groups combined (Figure 4.2, Table 4.2). Digested, unsterilized seeds yielded the next largest percentage of fungal taxa at 12%, while the digested, sterilized group yielded 2% (Figure 4.2). Digestion by Carollia removed ~70% of fungi in the digested, unsterilized treatment from the control seeds and ~94% in the digested, sterilized treatment. Microdochium lycopodinum was

the most abundant fungus and was present across all three treatments; however, the majority of taxa occurred once. Moreover, treatment did not affect the community composition of fungi (Figure 4.3), despite large beta diversity estimates (Table 4.5). We found that the alkenylphenols exhibited strong antifungal properties when fungi were exposed in a serial dilution of alkenylphenols, but that this result depended on the fungal taxa (Figure 4.4, Table 4.6). Finally, we discovered that there was no effect of bat digestion on seed micro-structures or scarification (Figure 4.5).

4.5.2 Seeds Were Cleaned of Fungi as They Passed through the Carollia Gut

We expected the greatest number of fungal taxa to be found in the digested, unsterilized treatment because animal guts are hyper-diverse microbial environments (Hardoim 2019), and Piper fruit often contain antimicrobial compounds (Orjala et al. 1998, Valdivia et al. 2008, Yang et al. 2013). In contrast to our prediction, we found that the undigested, unsterilized seeds (control) had the greatest number of fungal taxa, which is consistent with the hypothesis that digestion by animals is detrimental to microbes (Figure 4.2, Table 4.4; Janzen 1977). Fricke et al. (2013) found the same result in Capsicum (Capsiceae), where digestion of seeds by a disperser reduced the number of seed fungal pathogens. Interestingly, Fricke et al. (2013) showed that changes to seed condition, which included seed fungal pathogens and chemical volatiles, following digestion were more important in determining seed success against predators and pathogens than dispersal away from conspecifics as posited by the Janzen-Connell hypothesis. This is consistent with common knowledge that seeds and fruits are frequently targeted by fungal pathogens, especially in the tropics, because of their high nutrient value, environmental conditions which favour microbial growth, and the intense pressure from enemies. Another study on the frugivorous interactions

between Ficus colubrinae (Moraceae) fruit and a phyllostomid bat (Ectophylla alba) found that fungal pathogens decreased after digestion and fungal mutualists increased (Chaverri and Chaverri 2019). The incongruency between our initial hypothesis (i.e. control seeds would have fewer fungi than digested seeds) and our result (i.e. digested seeds had fewer fungi than control seeds) illustrates that fruit alkenylphenols did not prevent, or at least did not fully prevent, fungal colonization of control seeds. The causes controlling seed-colonizing fungi appear to be complex and dependent on a number of different variables.

In addition to the effects of digestion directly reducing the presence of fungi, if alkenylphenols and other compound classes (e.g. amides) in Piper fruit have strong antifungal and other antimicrobial properties (Orjala et al. 1998, Valdivia et al. 2008, Yang et al. 2013, Whitehead and Bowers 2014), then we would also expect fruit chemistry to restrict fungal colonization of seeds. However, it is possible that some specialist fungi have adapted to alkenylphenols found in P. sancti-felicis fruit and can circumvent their antifungal effects through the use of alternative enzymes (Kerscher et al. 1999, Marcet-Houben et al. 2009, O’Donnell et al. 2011, Adams et al. 2020). Regardless of fruit chemistry, it is likely that defensive compounds will not entirely eliminate fungi, and it is not unusual to find seed-colonizing fungi in the fruit of plants (e.g. Cecropia: Gallery et al. 2007, review: Nelson 2018). For example, previous work has found that the colonization of fruit by some fungi can increase desirability of the fruit to mammal and bird dispersers through the production of volatiles (Peris et al. 2017). Irrespective of the potential mechanism at play, our results highlight that there may be a benefit to seeds digested by bats, as passage through the bat gut removes fungi, which may include fungal pathogens that colonize the seed within the fruit (Table 4.3). More

research is required to assess plant fitness consequences of seed fungal colonization, such as germination and seed-predator feeding trials in our study system. Lastly, our culture-based experiment is undoubtably missing unculturable fungi, which may have important interactions with the seeds and their frugivores.

4.5.3 No Important Changes to the Seed Fungal Community

Our results showed that both digestion and sterilization treatments had no effect on the community composition of fungi from seeds, but that our treatment only explained 6.8% of the variation in the data (Figure 4.3), even though beta diversity estimates were large (Table 4.5). This discrepancy is likely due to the difference in culturable fungi across treatments, with control seeds producing 35 fungal cultures, the digested, unsterilized treatment yielding ten, and the digested, sterilized treatment yielding two. The lack of change in the community composition calculated by the constrained ordination is an unusual result because seed fungi are generally sensitive to environmental and biotic changes (Nelson 2018, Hardoim 2019). In order to determine whether this pattern was primarily driven by the dominant fungus, M. lycopodinum, we re-analyzed our ordination with this taxon removed, but found that this did not qualitatively influence our original result. One pattern that we did see in our experiment was that seeds often produced a binary pattern of fungi, meaning that each seed either yielded one or no fungi. This is consistent with the hypothesized seed endophyte bottleneck, where this binary pattern seems to be common across studies when culturing fungi from seeds (Newcombe et al. 2018). In our experiment, there were only two instances where a seed yielded two species of fungi each. It’s possible that this pattern is generated by using a culture- based technique and that a culture-independent technique may yield a better understanding of the community of seed-associated fungi. However, the bottleneck is

still expected for culture-independent techniques because seeds are more heavily defended than other parts of the plant, which is suspected to restrict seed fungal diversity (Newcombe et al. 2018). Indeed, the taxon richness of fungi found in seeds is far lower than what we have found in P. sancti-felicis leaves (see Chapter Three), but the number and concentration of alkenylphenols in P. sancti-felicis is low in seeds and leaves and high in fruit (Maynard et al. 2020). We did try to metabarcode seed fungal communities but were not able to amplify the DNA that was extracted, which is another common problem for seed fungi because of their size and interference from chemical compounds.

Furthermore, the seed fungi that we isolated are consistent with groups isolated from other culture-based and culture-independent studies (Nelson 2018, Hardoim 2019). For instance, our fungal taxa were predominantly from the division Ascomycota (Hodgson et al. 2014, Barret et al. 2015, Rezki et al. 2018), with only one taxon from Basidiomycota and one from Mucoromycota (Table 4.3). Within Ascomycota, almost all of our taxa were grouped within the Sordariomycetes, which contains Fusarium (Hodgson et al. 2014). The other classes we isolated, such as Dothideomycetes and Eurotiomycetes, are also frequently found in seeds (Barret et al. 2015).

4.5.4 Sources of Piper sancti-felicis Seed Fungi

We found that the most abundant species of fungus, M. lycopodinum, was present across all three treatments and that other fungal taxa were shared between the control and the digested, unsterilized treatment (Table 4.2). The presence of M. lycopodinum across the treatments and control indicates that it may have colonized the seed from the fruit environment and remained associated with the seed into the digested, sterilized

treatment, suggesting that it internally colonized the seed. Similarly, an experiment investigating bird digestion on the presence of arbuscular mycorrhizal fungi (AMF) in Rubus (Rosaceae) seeds found that AMF remained associated with their seeds after digestion (Correia et al. 2019). Meanwhile, additional research has shown that other types of animals, particularly arthropods, molluscs, and rodents, are able to disperse mycorrhizal fungi, such as AMF and ectomycorrhizal fungi (ECM) (Vašutová et al. 2019). These organisms can disperse AMF and ECM independently of their host plant through both endo and ectozoochory, to re-form successful plant symbioses (Vašutová et al. 2019). Alternatively, M. lycopodinum could be an abundant fungus that independently colonized the seed from the fruit and then lost its association with the seed after digestion, only to be recolonized by other hyphae of the same taxa in the bat gut. Future work should investigate the microbiome of the fruit and the digestive tract and guano of Carollia. In our experiment, we included the sterilized, digested treatment where we incubated seeds in guano for one week, because we were interested in how guano may increase the colonization of fungi to the interior of seeds. Given that some plant species have dominant and ecologically important fungi that are passed vertically from mother to seed (Clay 1988, Cook et al. 2014), we wondered if fungal colonization of the interior of the seed may be aided by animal digestion. In a pilot experiment, we had already determined that fungi were essentially not vertically transmitted in P. sancti- felicis (unpublished data, Heather Slinn). Similar to other studies on fungal microbiomes, we saw a clear dominant fungal species present across our treatments and control.

In addition to the presence of M. lycopodinum across treatments, we found two other fungal taxa, Fusarium and Neopestalotiopsis clavispora, that were shared between control seeds and digested, unsterilized seeds (Table 4.2). Again, this indicates that fungi present in the fruit could continue their association with seeds after dispersal, or it may be that they were newly colonized in the bat gut. Only two taxa, M. lycopodinum and Chaetothyriales, were cultured from seeds in the one-week guano treatment, where seeds were then surface sterilized. Regardless of whether fungi colonized the interior or the surface of the seed, they may still provide important functional roles to the seed and developing plant through colonization of the seedling (Shahzad et al. 2018, Nelson 2018), and this would be an important avenue of research.

4.5.5 Fruit Alkenylphenols have Antifungal Properties

Interestingly, we found that alkenylphenols exhibited strong antifungal properties for two out of three fungi used in our assay (Figure 4.4, Table 4.6), similar to other studies on antimicrobials of Piper alkenylphenols (Orjala et al. 1998, Valdivia et al. 2008, Yang et al. 2013). Given that fungal pathogens are a high source of mortality for seeds and seedlings, our finding that alkenylphenols can protect seeds against some fungi has important implications for plant fitness. Moreover, this matches what we would expect based on optimal defence theory (McKey 1974), which posits that plants will invest heavily in plant parts (e.g. reproductive structures) where the cost of losing those parts would be high. Studies have shown that this is often the case with herbivory damage in strongly defended new leaves compared to older leaves (Coley and Barone 1996, McCall and Fordyce 2010) and across different plant organs, such as roots, leaves, and reproductive structures, where reproductive structures are the best constitutively

100

defended structures (Zangerl and Rutledge 1996). Furthermore, another study on P. sancti-felicis alkenylphenols found that the fruit and developing flowers contained the greatest number of alkenylphenols compounds, and fruit had the highest concentration of alkenylphenols compared to the rest of the plant organs (Maynard et al. 2020). In addition, the concentration of alkenylphenols peaked across inflorescence development when fruit ripened, further supporting that these alkenylphenols serve a defensive purpose (Maynard et al. 2020).

While we sought to determine the antifungal properties of alkenylphenols, the fungal taxa in our assay are all listed as potential endophytes and plant pathogens, based on prior evidence (FunGuild: Nguyen et al. 2016, Fusarium: Tedersoo et al. 2014, Blacutt et al. 2018, Microdochium: Hernandez-Restrepo et al. 2016). Fusarium is well known for producing plant pathogens across many types of ecosystems and in different crops (Booth 1971, Summerell 2019). For example, one of the most common Fusarium pathogens, Fusarium graminearum, damages the reproductive structures of multiple different species of cereals and other crops in both temperate and semi-tropical habitats (Goswami and Kistler 2004). Other species of Fusarium are known pathogens of Piper species: P. betle and P. nigrum (Shahnazi et al. 2012, Edward et al. 2013). While Microdochium is a common pathogen of grasses (Hernandez-Restrepo et al. 2016), it can also act as a dark septate endophyte, which colonizes the roots of grasses (Mandyam et al. 2012). Interestingly, in an experimental inoculation with Microdochium as a dark septate endophyte, colonization increased plant biomass or had no effect, depending on the species of grass (Mandyam et al. 2012). These studies suggest that Microdochium can be beneficial, antagonistic, or have no effect on the host plant, depending on the context. Overall, while the taxa in our microdilution assay are likely

101

pathogens under certain contexts, future work should test identified Piper pathogens against their fruit alkenylphenols.

Lastly, given that one of our fungal taxa (Fusarium sp. B) did not demonstrate a decline in growth with concentration of alkenylphenols (Fig. 4.4), it is possible that it has evolved greater tolerance to plant secondary metabolites that interfere with metabolism, which may be achieved through the use of alternative energy production enzymes (Kershser et al. 1999, Marcet-Houben et al. 2009, O’Donnell et al. 2011). For instance, in the wild chili pepper (Capsicum chacoense), fungal tolerance to dihydrocapsaicin was found in 16 fungal pathogens from 4 distantly related genera (Alternaria, Colletotrichum, Fusarium, and Phomopsis), which had alternative enzymes at different stages of the electron transport chain, such as for NADH dehydrogenase and oxidase (Adams et al. 2020). Due to the phylogenetic breadth of this study, it suggests that alternative mechanisms which allow fungal pathogens to tolerate capsaicinoids may be common (Adams et al. 2020). Furthermore, it’s interesting that one Fusarium taxon experienced a decline in growth, while the other taxon did not. This might be caused by variations in alternative metabolism enzymes mentioned above, as this is a large genus which infects many different plant species. However, more pathogenic fungi from different orders and from closely related species should be tested against alkenylphenols to assess the generality of this potential conclusion.

4.5.6 Digestion has no Effect on Seed Microstructures or Scarification

Qualitatively, we found that bat digestion did not alter seed-microstructures, such as CaOx, or scarify seeds (Figure 4.5). Calcium oxalate is used by plants for growth, herbivore defence, and calcium regulation (Korth et al. 2006, Nataka 2012). These

102

crystals are also found on the seed coat in Piper aduncum (Dousseau et al. 2011) and in other vegetative organs of Piper (Raman et al. 2012). Scarification can be an important factor that improves seed germination and success in other plant systems (Traveset and Verdú 2002) by facilitating the passage of air and water across the seed coat; however, our SEM images indicate that it does not appear to happen here. While there is no other research on whether scarification happens in Piper, there are mixed reports about whether germination success and germination rate of various Piper species are influenced by bat digestion (Fleming 2004, Baldwin and Whitehead 2015, Saldana-Vasquez et al. 2019). For instance, one study showed that gut retention time in Carollia perspicillata increased germination success of Piper peltatum, but not for four other Piper species (Baldwin and Whitehead 2015). Furthermore, a recent meta- analysis found that bat digestion improved germination in three different Piper species, compared to Baldwin and Whitehead (2015) (Saldana-Vasquez et al. 2019). This positive effect is consistent with a larger meta-analysis that looked across different types of frugivores and plants, where tropical plants and plants that had fleshy fruit both independently had higher germination rates after digestion by a frugivore (Traveset and Verdú 2002). However, future work should investigate how digestion and seed microbial communities alter seed germination variables.

4.6 Conclusion

Overall, there are clear consequences of digestion by specialist Carollia bats on the seed fungi of P. sancti-felicis. Digestion removed fungi from seeds (~70% for digested, unsterilized and ~94% for digested, sterilized) compared to the control, but it also created an opportunity for fungi to colonize the interior of the seed coat. Seed fungal

103

community composition did not differ across digestion and sterilization treatments. Fruit alkenylphenols exhibited strong antifungal properties for two out of three fungi, and digestion did not scarify seeds or alter their surface micro-structures. While our results are intriguing, it has not escaped our notice that non-specialist dispersers of P. sancti- felicis, such as birds, may influence seeds and their fungi in different ways (Maynard et al. 2020).

In conclusion, our study is important because the seed mycobiome is an overlooked area of seed dispersal (Nelson 2018, Shahzad et al. 2018, Hardoim 2019), and fungi are known to play important functional roles in plants in various organs (Rodriguez et al. 2009, Mejía et al. 2014, Christian et al. 2019). These fungi may be mediating crucial interactions at the seed stage that we have yet to identify. Additionally, understanding the role of fungi on plant fitness could have important implications for our understanding of Piper and its associated interactions with other organisms and the environment. Factors that shape these interactions may be applied in an agricultural context for sustainable pest management, especially those that are chemically mediated (Kauppinen et al. 2016). For example, Epichloë, a fungal endophyte commonly found in grasses, is being used in some countries as a method of biocontrol in forage grasses and turf grasses as it can reduce herbivory and improve plant resistance to environmental stress (Kauppinen et al. 2016).

104

4.7 Tables

Table 4.1: Frequency of fungal taxa by treatment and individual fruit where 4 seeds were collected from each fruit (NS_DIG: unsterilized, digested treatment; ST_DIG: sterilized, digested treatment).

105

Table 4.2: Frequency of fungal taxa by treatment (NS_DIG: unsterilized, digested treatment; ST_DIG: sterilized, digested treatment).

106

Table 4.3: Taxonomic and guild information of fungal taxa and the frequency of each taxon.

107

Table 4.4: Coefficients from logistic regression of digestion and sterilization treatments on presence of fungi.

108

4.8 Figures

109

0.3

d e

e treatment

s 0.2

control f

o NS_DIG

n ST_DIG

p o

r 0.1 P

0.0

Unsterilized, Undigested Unsterilized, Digested Sterilized, Digested Treatment

110

0 ST_DIG

) Treatment

% 0

4 control

(

2 NS_DIG P

A ST_DIG C

−1

NS_DIG

−2

−0.5 0.0 0.5 1.0 1.5 CAP1 (60%)

111

) 1.00 D

●

O (

● ●

e ● ● Time c

● 24 hours n ● ● a ● 72 hours

b ● ● r 0.75 ● o ● Fungi

s ● ● ● b ● ● ● Fusarium A

● A

● ● Fusarium B ● e ● Microdochium lycopodium

g ●

a ● r 0.50

e ●

v A ●

0.25

0.00 0.25 0.50 0.75 Total alkenylphenols (mg/ml)

112

Figure 4-5: Seed micro-structures of digested (A,B) and undigested (C,D) seeds taken with a scanning electron microscopy.

113

5 Conclusion

Overall, my dissertation research shows how important plant chemistry and other biotic and abiotic variables are in shaping fungal, insect, and mammalian interactions associated with the tropical shrub, Piper (Slinn et al. 2018, Maynard et al. 2020). Chemically mediated plant-insect interactions has been a large and important field that has grown over the last 50 years since the foundational works of Fraenkel (1959) and Ehrlich and Raven (1964). Since this time, chemical ecologists have progressed in our understanding of the importance plant chemistry in shaping its associated communities (Hunter 2016, Dyer et al. 2018). More recently, research has sought to understand whether heritable variation exists in plant chemistry, among other important plant functional traits, and how existing heritable variation scales up to influence heritable variation in species interactions such as herbivory (Geber and Griffen 2003, Johnson et al. 2009, Barbour et al. 2015). Understanding the causes of plant chemistry is important for evolutionary plant theory as well as the ecological consequences of it. However, one field of research which has gained much more interest in recent years is how microorganisms are mediating the relationships between plant chemistry and other animal interactions with host plants, as well as how they are shaped by it. Studies are especially lacking for tropical study systems as well as seed and dispersal-related microbial processes (Nelson 2018, Shahzad et al. 2018, Hardoim 2019). What we do know is that plant chemistry can be shaped both directly and indirectly by the colonization of different kingdoms of microbes such as bacteria, fungi, and archaea and that this can have important consequences for other organisms (Hardoim et al. 2015). And research in this area has greatly improved as high-throughput sequencing techniques have dropped dramatically in price over the last 13 years since the human

114

genome project in 2007. My dissertation research contributes to this knowledge gap by investigating the ecological filters of seed fungal communities pre- and post-dispersal as well as how plant chemistry, herbivory, and land-use history shape the foliar fungal endophyte communities.

The fields of chemical ecology and ecoimmunology are two important disciplines that shape my research, most especially in Chapter Two (Slinn et al. 2018). In this chapter, I demonstrated how interspecific differences in plant chemistry of Piper species alter the ability of herbivores to mount an immune response and that this response depends on the diet breadth of the herbivore (Slinn et al. 2018). Here, I also showed how the effects of plant chemistry on herbivores can mediate parasitoid success at two different tropical sites (Slinn et al. 2018). While it is possible that the differences we see in herbivore response at the two different sites may be a product of the differences in responses among the species comprising the different herbivore communities, I am not concerned about how this influences the interpretation of our results as biotic communities are always an attribute of the site they inhabit. Furthermore, there was no overlap in herbivore species between the sites, so we cannot compare any one species found in both sites. Past research has looked at how caterpillar density may influence how frequently they are parasitized and has shown that it is not a predictor of parasitism rate for any of the Eois species that we studied in this chapter; rather, plant density (Richards et al. 2015) and plant chemical diversity are more important (Glassmire 2017).

My work in Chapters Two (Slinn et al. 2018) and Three are particularly interesting as they have been built on over 20 years of chemical and ecological knowledge and observation at La Selva Biological Station in Costa Rica, and approximately 15 years at

115

Yanayacu Biological Station in Ecuador. In terms of chemical knowledge, we have identified important classes of compounds found in Piper, and we know that variation in chemical profiles and phytochemical diversity varies more across species of Piper than within a species (Fincher et al. 2008, Richards et al. 2015, Salazar et al. 2016a). To identify species boundaries in Piper, our work uses traditional taxonomy, which focuses on the reproductive parts, as well as phylogenomics (Uckele et al. 2020). Similarly, to identify the lepidopteran specialist, Eois, our research group has collected a matrix of over 120 characters, which is based mostly on genitalia, but also includes wing, leg, and antennal characteristics (Dyer et al. 2010). We have confirmed these characters using genomic data, but not with the COI locus, which is not a useful locus for this genus. The lab group at the University of Nevada Reno is also working on phylogenomic analyses for Eois as well and has found that there are cryptic species in Ecuador, but that these cryptic species were already identified with genitalic examination (manuscript in prep, Lee Dyer). For the labeling of Eois as a specialist, we have two decades worth of evidence to suggest this (summarized in Chapter Two), which shows that these caterpillars typically feed on one to two plant species (Table 2.2), and often just one leaf. For our generalist lepidopteran species, Quadrus cerealis, which feeds on 23 Piper species (Table 2.3), it is possible that that there are cryptic Quadrus species that we have yet to identify. While we know that the COI locus is not useful for Eois, we are unsure about its utility for Quadrus. Regardless, Janzen et al. (2011) have shown that specimens for Quadrus cerealis in Costa Rica are only one species using the COI locus. The chemical ecology group at the University of Nevada Reno is currently doing genomics work on Quadrus along with morphology of genitalia to examine whether there is evidence for potential cryptic species (unpublished data, Lee Dyer). Similarly to

116

Eois, Quadrus cerealis tends to stick to feeding on one leaf, which is partially due to the fact that it is a shelter builder.

There are a number of other limitations associated with my research chapters investigating the Piper microbiome (Chapters Three and Four). As with any studies that investigate microbial communities colonizing hosts, the limitations of current high- throughput technologies can take our interpretations only so far. For instance, read counts do not equate to abundance data of microorganisms when Next-Generation Sequencing (NGS) is used (Amend et al. 2010, Nguyen et al. 2015), although there are solutions to this being developed via internal standards (Harrison et al. 2021) or through qPCR when particular microbial taxa are of interest. Here, I addressed this shortcoming by normalizing my data to account for potential sequencing errors (McMurdie and Holmes 2014, Robinson et al. 2010) and analyzing my sequence data in a variety of ways. For example, I removed ESVs with low read counts, to account for potential spurious ESVs (Weiss et al. 2015), and I also modified my community matrix to create presence/absence data, thus removing the relative abundance data generated by my normalization. Across all three of these modifications, I analyzed my sequence data and found that the results were all qualitatively the same, and quantitatively similar. As a result, I am not worried about these issues influencing our interpretations, although I would still not put stock into normalized read counts as abundance. In contrast, in Chapter Four, I was not able to use NGS technology on seed fungal communities, likely due to seed chemistry interfering with DNA extraction, and I relied instead on culture- based techniques to identify the fungal community. In this experiment, we are likely underrepresenting the fungal community of seeds as some fungi are difficult to culture. Despite this, there was a large effect size of digestion and sterilization on the proportion

117

of seed fungi (Figure 4.2) we were able to culture, and it is likely still true that these treatments are important for fungal abundance. Perhaps with high-throughput technologies, we will be better equipped to determine whether there is any meaningful change in seed fungi.

In conclusion, through my research, I have sought to link together the world of plant-insect and chemical ecology with microbial ecology in a tropical ecosystem. I have shown across my research chapters that plant chemistry is an important ecological filter of seed fungal communities and foliar fungal endophyte communities, in some contexts, and plays an important role in modifying tri-trophic interactions. In Chapter Two, I demonstrated how important plant chemical diversity is at influencing the ability of an herbivore to mount an immune response and how likely they are to defend themselves against parasitoids in two different tropical ecosystems (Slinn et al. 2018). I then went on to quantify heritability in plant chemistry and other species interactions and found that classes of compounds and chemical shifts seem to be strongly heritable (Chapter Three). In this chapter I also found that land-use history was an important ecological filter of foliar fungal endophyte communities, as was plant ontogeny, which was aided by phytochemical diversity and herbivory. In turn, specialist herbivory was influenced by host plant ontogeny and phytochemical diversity but not fungal richness. My last research chapter (Chapter Four) sought to understand how fruit chemistry affected the ability of fungi to colonize seeds (Maynard et al. 2020) and whether digestion modified seed communities. Here, I found that a compound class, termed alkenylphenols, negatively affected fungal growth in two of three fungal candidates and that the abundance of fungal taxa was greatly reduced in digested treatments compared to undigested seeds. My work will be useful as a steppingstone for future research on this

118

interdisciplinary topic across these three disciplines in Piper, where there is little to no research conducted on the microbes that colonize it (with the exception of pathogens in Piper crops). Given the ecological importance of Piper across the new and old-world tropics, both in agricultural and natural communities, I hope that future work continues to seek out answers about how microbial communities are shaped and how they can facilitate and change the complex network of species interactions which Piper hosts.

119

REFERENCES

Adams, R. I., Miletto, M., Taylor, J. W., & Bruns, T. D. (2013). Dispersal in microbes: Fungi in indoor air are dominated by outdoor air and show dispersal limitation at short distances. ISME Journal, 7(7), 1262–1273.

Adams, C. A., Zimmerman, K., Fenstermacher, K., Thompson, M. G., Skyrud, W., Behie, S., & Pringle, A. (2020). Fungal seed pathogens of wild chili peppers possess multiple mechanisms to tolerate capsaicinoids. Applied and Environmental Microbiology, 86(3), e01697-19.

Agrawal, A. A. (2000). Mechanisms, ecological consequences and agricultural implications of tri-trophic interactions. Current Opinion in Plant Biology, 3(4), 329– 335.

Agrawal, A. a., & Weber, M. G. (2015). On the study of plant defence and herbivory using comparative approaches: how important are secondary plant compounds. Ecology Letters, 18(10), 985–991.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W., & Lipman, D. J. (1990). Basic Local Alignment Search Tool. Journal of Molecular Biology, 215(3), 403–410.

Amend, A. S., Seifert, K. A., & Bruns, T. D. (2010). Quantifying microbial communities with 454 pyrosequencing: Does read abundance count? Molecular Ecology, 19(24), 5555–5565.

Antwis, R. E., Griffiths, S. M., Harrison, X. A., Aranega-Bou, P., Arce, A., Bettridge, A. S., … Sutherland, W. J. (2017). Fifty important research questions in microbial ecology. FEMS Microbiology Ecology, 93(5), 1–10.

Arnold, A. E. (2008). Endophytic fungi: hidden components of tropical community ecology. In W. Carson & S. A. Schnitzer (Eds.), Tropical forest community ecology (pp.178–188). Oxford, UK: Wiley-Blackwell.

Arnold, A. E., & Lutzoni, F. (2007). Diversity and Host Range of Foliar Fungal Endophytes: Are Tropical Leaves Biodiversity Hotspots? Ecology, 88(3), 541–549.

120

Arnold, A. E., Mejía, L. C., Kyllo, D., Rojas, E. I., Maynard, Z., Robbins, N., & Herre, E. A. (2003a). Fungal endophytes limit pathogen damage in a tropical tree. Proceedings of the National Academy of Sciences, 100(26), 15649–15654.

Arnold, A. E., Miadlikowska, J., Higgins, K. L., Sarvate, S. D., Gugger, P., Way, A., … Lutzoni, F. (2009). A phylogenetic estimation of trophic transition networks for ascomycetous Fungi: Are lichens cradles of symbiotrophic fungal diversification? Systematic Biology, 58(3), 283–297.

Baldwin, J. W., & Whitehead, S. R. (2015). Fruit secondary compounds mediate the retention time of seeds in the guts of Neotropical fruit bats. Oecologia, 177(2), 453– 466.

Bálint, M., Tiffin, P., Hallström, B., O’Hara, R. B., Olson, M. S., Fankhauser, J. D., … Schmitt, I. (2013). Host Genotype Shapes the Foliar Fungal Microbiome of Balsam Poplar (narro). PLoS ONE, 8(1), e53987.

Bálint, M., Bartha, L., O’Hara, R. B., Olson, M. S., Otte, J., Pfenninger, M., … Schmitt, I. (2015). Relocation, high-latitude warming and host genetic identity shape the foliar fungal microbiome of poplars. Molecular Ecology, 24(1), 235–248.

Barbosa, P., Gross, P., & Kemper, J. (1991). Influence of plant allelochemicals on the Tobacco Hornworm and its parasitoid, Cotesia congregata. Ecology, 72(5), 1567– 1575.

Barbour, M. A., Rodriguez-Cabal, M. A., Wu, E. T., Julkunen-Tiitto, R., Ritland, C. E., Miscampbell, A. E., Jules, E. S., & Crutsinger, G. M. (2015). Multiple plant traits shape the genetic basis of herbivore community assembly. Functional Ecology, 29(8), 995–1006.

Barbour, M. A., Fortuna, M. A., Bascompte, J., Nicholson, J. R., & Julkunen-Tiitto, R. (2016). Genetic specificity of a plant – insect food web: Implications for linking genetic variation to network complexity. Proceedings of the National Academy of Sciences USA, 113(8), 2128–2133.

121

Barbour, M. A., Erlandson, S., Peay, K., Locke, B., Jules, E. S., & Crutsinger, G. M. (2019). Trait plasticity is more important than genetic variation in determining species richness of associated communities. Journal of Ecology, 107(1), 350–360.

Barker, H. L., Holeski, L. M., & Lindroth, R. L. (2019). Independent and interactive effects of plant genotype and environment on plant traits and insect herbivore performance: A meta-analysis with Salicaceae. Functional Ecology, 33(3), 422– 435.

Barret, M., Briand, M., Bonneau, S., Préveaux, A., Valière, S., Bouchez, O., … Jacques, M. A. (2015). Emergence Shapes the Structure of the Seed Microbiota. Applied and Environmental Microbiology, 81(4), 1257–1266.

Barton, K. E., & Koricheva, J. (2010). The ontogeny of plant defense and herbivory: Characterizing general patterns using meta-analysis. American Naturalist, 175(4), 481–493.

Bates, D., Martin M., Bolker, B., Walker, S. (2015). Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 67(1), 1–48.

Becerra, J. X. (2007). The impact of herbivore-plant coevolution on plant community structure. Proceedings of the National Academy of Sciences, 104(18), 7483–7488.

Becerra, J. X. (2015). On the factors that promote the diversity of herbivorous insects and plants in tropical forests. Proceedings of the National Academy of Sciences, 112(19), 6098–6103.

Beckage, N. E. (2008). Insect immunology. Academic Press.

Berenbaum, M. (1990). Evolution of specialization in insect-umbellifer associations. Annual Review of Entomology, 35(1), 319–343.

Berenbaum, M., & Neal, J. J. (1985). Synergism between myristicin and xanthotoxin, a naturally cooccurring plant toxicant. Journal of Chemical Ecology, 11(10), 1349– 1358.

122

Berenbaum, M. R., Zangerl, A. R., & Nitao, J. K. (1986). Constraints on Chemical Coevolution: Wild Parsnips and the Parsnip Webworm. Evolution, 40(6), 1215– 1228.

Berenbaum, M. R., & Zangerl, A. R. (2008). Facing the Future of Plant-Insect Interaction Research: Le Retour a la “Raison d’Etre.” Plant Physiology, 146(3), 804–811.

Bernays, E. A., & Graham, M. (1988). On the Evolution of Host Specificity in Phytophagous Arthropods. Ecology, 69(4), 886–892.

Bewley, J. D., Bradford, K. J., Hillhorst, H. W. M. and Nonogaki, H. (2013). Seeds: Physiology of development, germination and dormancy. 3rd Edition, Springer, New York.

Blacutt, A. A., Gold, S. E., Voss, K. A., Gao, M., & Glenn, A. E. (2018). Fusarium verticillioides: Advancements in understanding the toxicity, virulence, and niche adaptations of a model mycotoxigenic pathogen of maize. Phytopathology, 108(3), 312–326.

Bodenhausen, N., Bortfeld-Miller, M., Ackermann, M., & Vorholt, J. A. (2014). A Synthetic Community Approach Reveals Plant Genotypes Affecting the Phyllosphere Microbiota. PLoS Genetics, 10(4), e1004283.

Boege, K., & Marquis, R. J. (2005). Facing herbivory as you grow up: the ontogeny of resistance in plants. Trends in Ecology & Evolution, 20(8), 441–448.

Booth, C. (1971). The genus fusarium. Commonwealth Mycological Institute, Kew, U.K.

Bowers, M.D. (1990). Recycling plant natural products for insect defense. In: Evans, D.L. & Schmidt, J.O. (eds). Insect Defenses. SUNY Press, Albany, NY, pp. 353– 386.

Bradshaw, C. J. A., Sodhi, N. S., & Brook, B. W. (2009). Tropical turmoil: A biodiversity tragedy in progress. Frontiers in Ecology and the Environment, 7(2), 79–87.

Brehm, G., Bodner, F., Strutzenberger, P., Hünefeld, F., & Fiedler, K. (2011). Neotropical Eois (Lepidoptera: Geometridae): Checklist, biogeography, diversity,

123

and description patterns. Annals of the Entomological Society of America, 104(6), 1091–1107.

Bruns, T. D. (2019). The developing relationship between the study of fungal communities and community ecology theory. Fungal Ecology, 39, 393–402.

Buerger, S., Spoering, A., Gavrish, E., Leslin, C., Ling, L., & Epstein, S. S. (2012). Microbial Scout Hypothesis and Microbial Discovery. Applied and Environmental Microbiology, 78(9), 3229–3233.

Bukovinszky, T., Poelman, E. H., Gols, R., Prekatsakis, G., Vet, L. E. M., Harvey, J. A., & Dicke, M. (2009). Consequences of constitutive and induced variation in plant nutritional quality for immune defence of a herbivore against parasitism. Oecologia, 160(2), 299–308.

Busby, P. E., Lamit, L. J., Keith, A. R., Newcombe, G., Gehring, C. a., Whitham, T. G., & Dirzo, R. (2015). Genetics-based interactions among plants, pathogens, and herbivores define arthropod community structure. Ecology, 96(7), 1974–1984.

Busby, P. E., Newcombe, G., Dirzo, R., & Whitham, T. G. (2013). Genetic basis of pathogen community structure for foundation tree species in a common garden and in the wild. Journal of Ecology, 101(4), 867–877.

Busby, P. E., Peay, K. G., & Newcombe, G. (2016). Common foliar fungi of Populus trichocarpa modify Melampsora rust disease severity. New Phytologist, 209, 1681– 1692.

Callahan, B. J., McMurdie, P. J., & Holmes, S. P. (2017). Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME Journal, 11(12), 2639–2643.

Carmona, D., Lajeunesse, M. J., & Johnson, M. T. J. (2011). Plant traits that predict resistance to herbivores. Functional Ecology, 25(2), 358–367.

Carvalheiro, L. G., Buckley, Y. M., & Memmott, J. (2010). Diet breadth influences how the impact of invasive plants is propagated through food webs. Ecology, 91(4), 1063–1074.

124

Cerenius, L., Lee, B. L., & Söderhäll, K. (2008). The proPO-system: pros and cons for its role in invertebrate immunity. Trends in Immunology, 29(6), 263–271.

Charney, N. & Record, S. (2012). vegetarian: Jost Diversity Measures for Community Data. R package version 1.2. https://CRAN.R-project.org/package=vegetarian

Chaverri, P., & Chaverri, G. (2019). A missing link in mutualistic networks: symbiotic fungi in plant-animal interactions. BioRxiv, 761270.

Christian, N., Herre, E. A., & Clay, K. (2019). Foliar endophytic fungi alter patterns of nitrogen uptake and distribution in Theobroma cacao. New Phytologist, 222(3), 1573–1583.

Christian, N., Herre, E. A., Mejia, L. C., & Clay, K. (2017). Exposure to the leaf litter microbiome of healthy adults protects seedlings from pathogen damage. Proc. R. Soc. B, 284, 20170641.

Christian, N., Sedio, B. E., Florez-Buitrago, X., Ramírez-Camejo, L. A., Rojas, E. I., Mejía, L. C., … Herre, E. A. (2020). Host affinity of endophytic fungi and the potential for reciprocal interactions involving host secondary chemistry. American Journal of Botany, 107(2), 1–10.

Clay, K. (1988). Fungal Endophytes of Grasses: A Defensive Mutualism between Plants and Fungi. Ecology, 69(1), 10–16.

Coley, P. D., & Barone, J. A. (1996). Herbivory and plant defenses in tropical forests. Annual Review of Ecology and Systematics, 27, 305–335.

Coley, P. D., Bateman, M. L., & Kursar, T. A. (2006). The effects of plant quality on caterpillar growth and defense against natural enemies. Oikos, 115(2), 219–228.

Connahs, H., Rodríguez-Castañeda, G., Walters, T., Walla, T., & Dyer, L. (2009). Geographic variation in host-specificity and parasitoid pressure of an herbivore (Geometridae) associated with the tropical genus Piper (Piperaceae). Journal of Insect Science, 9(28), 1–11.

125

Connell, J. H. (1971). On the role of natural enemies in preventing competitive exclusion in some marine animals and in rain forest trees. Dynamics of Populations, (298), 312.

Connor, J. K. & Hartl, D. H. (2004). A primer of ecological genetics. Sunderland, MA.

Cook, D., Gardner, D. R., & Pfister, J. A. (2014). Swainsonine-containing plants and their relationship to endophytic fungi. Journal of Agricultural and Food Chemistry, 62(30), 7326–7334.

Correia, M., Heleno, R., da Silva, L. P., Costa, J. M., & Rodríguez-Echeverría, S. (2019). First evidence for the joint dispersal of mycorrhizal fungi and plant diaspores by birds. New Phytologist, 222(2), 1054–1060.

Crawford, K. M., Crutsinger, G. M., & Sanders, N. J. (2007). Host-plant genotypic diversity mediates the distribution of an ecosystem engineer. Ecology, 88(8), 2114– 2120.

Crutsinger, G. M., Collins, M. D., Fordyce, J. a, Gompert, Z., Nice, C. C., & Sanders, N. J. (2006). Plant genotypic diversity predicts community structure and governs an ecosystem process. Science, 313, 966–968.

Cunningham, S. A., Summerhayes, B., & Westoby, M. (1999). Evolutionary divergences in leaf structure and chemistry, comparing rainfall and soil nutrient gradients. Ecology, 69(4), 569–588.

Delabye, S., Rougerie, R., Bayendi, S., Andeime-Eyene, M., Zakharov, E. V., DeWaard, J. R., … Decaëns, T. (2019). Characterization and comparison of poorly known moth communities through DNA barcoding in two Afrotropical environments in Gabon. Genome, 62(3), 96–107.

Deshpande, V., Wang, Q., Greenfield, P., Charleston, M., Porras-Alfaro, A., Kuske, C. R., … Midgley, D. J. (2016). Fungal identification using a Bayesian classifier and the Warcup training set of internal transcribed spacer sequences training set of internal transcribed spacer sequences. Mycologia, 108(1), 1–5.

126

Diamond, S. E., & Kingsolver, J. G. (2011). Host plant quality, selection history and trade-offs shape the immune responses of Manduca sexta. Proceedings of the Royal Society of London B: Biological Sciences, 278(1703), 289–297.

Dirzo, R., & Raven, P. H. (2003). Global State of Biodiversity and Loss. Annual Review of Environment and Resources, 28(1), 137–167.

Donnell, A. O., Harvey, L. M., & Mcneil, B. (2011). The roles of the alternative NADH dehydrogenases during oxidative stress in cultures of the filamentous fungus Aspergillus niger. Fungal Biology, 115(4–5), 359–369.

Dousseau, S., Alvarenga, A. A. D., Alves, E., Chaves, I. D. S., Souza, E. D. S., & Alves, J. D. S. (2011). Physiological, morphological and biochemical characteristics of the sexual propagation of Piper aduncum (Piperaceae). Brazilian Journal of Botany, 34(3), 297–305.

Dray, S., & Dufour, A. B. (2007). The ade4 package: Implementing the duality diagram for ecologists. Journal of Statistical Software, 22(4), 1–20.

Dyer, L.A. & T. Floyd. (1993). Determinants of predation on phytophagous insects: the importance of diet breadth. Oecologia, 96:575–582.

Dyer, L. A. (1995). Tasty generalists and nasty specialists? Antipredator mechanisms in tropical Lepidopteran larvae. Ecology, 76(5), 1483–1496.

Dyer, L. A. (2011). New Synthesis-Back to the Future: New Approaches and Directions in Chemical Studies of Coevolution. Journal of Chemical Ecology, 37(7), 669.

Dyer, L. A., & Coley, P. D. (2002). Tri-trophic interactions in tropical versus temperate communities. In: Tscharntke T, Hawkins B (eds). Multitrophic Level Interactions. Cambridge University Press, pp 67–88.

Dyer, L. A., C. D. Dodson, J. O. Stireman, A. M. Smilanich, R. M. Fincher, & Letourneau, D. K. (2003). Synergistic effects of three Piper amides on generalist and specialist herbivores. Journal of Chemical Ecology, 29(11), 2499–2514.

127

Dyer, L. A., & Gentry, G. (1999). Predicting natural-enemy responses to herbivores in natural and managed systems. Ecological Applications, 9(2), 402–408.

Dyer L. A., Gentry, G. L., Greeney, H. F., & Walla T. W. (2010). Caterpillars and parasitoids of the Eastern Andes in Ecuador. [WWW document] URL http://www.caterpillars.org [accessed 16 January 2018].

Dyer, L. A., & Letourneau, D. K. (2013). Can climate change trigger massive diversity cascades in terrestrial ecosystems? Diversity, 5(3), 479–504.

Dyer, L. A., Letourneau, D. K., Chavarria, G. V., & Amoretti, D. S. (2010). Herbivores on a dominant understory shrub increase local plant diversity in rain forest communities. Ecology, 91(12), 3707–3718.

Dyer, L. A., Letourneau, D. K., Dodson, C. D., Tobler, M. A., Stireman, J. O., & Hsu, A. (2004a). Ecological causes and consequences of variation in defensive chemistry of a neotropical shrub. Ecology, 85(10), 2795–2803.

Dyer, L. A. & Palmer, A. N. (2004). Piper: A model genus for studies of phytochemistry, ecology and evolution. New York, NY, USA: Kluwer Academic Publishers.

Dyer, L. A., Dodson, C. D., & Richards, J. (2004). In L. A. Dyer & A. N. Palmer (Eds.), Piper: A model genus for studies of phytochemistry, ecology and evolution (pp. 117–139). New York, NY: Kluwer Academic Publishers.

Dyer, L. A., Dodson, C. D., Beihoffer, J., & Letourneau, D. K. (2001). Trade-offs in antiherbivore defenses in Piper cenocladum: Ant mutualists versus plant secondary metabolites. Journal of Chemical Ecology, 27(3), 581–592.

Dyer, L. A., Philbin, C. S., Ochsenrider, K. M., Richards, L. A., Massad, T. J., Smilanich, A. M., … Jeffrey, C. S. (2018). Modern approaches to study plant–insect interactions in chemical ecology. Nature Reviews Chemistry, 2(6), 50–64.

Dyer, L. A., Singer, M. S., Lill, J. T., Stireman, J. O., Gentry, G. L., Marquis, R. J., Ricklefs, R. E., Greeney, H. F., Wagner, D. L., Morais, H. C., Diniz, I. R., Kursar, T. A., & Coley, P. D. (2007). Host specificity of Lepidoptera in tropical and temperate forests. Nature, 448(7154), 696–699.

128

Edgar, R. C. (2010). Search and clustering orders of magnitude faster than BLAST. Bioinformatics, 26(19), 2460–2461.

Edgar, R. C. (2016a). UNOISE2: improved error-correction for Illumina 16S and ITS amplicon sequencing. BioRxiv, 081257.

Edgar, R. (2016b). SINTAX: a simple non-Bayesian taxonomy classifier for 16S and ITS sequences. BioRxiv, 074161.

Edgar, R. C., & Flyvbjerg, H. (2015). Error filtering, pair assembly and error correction for next-generation sequencing reads. Bioinformatics, 31(21), 3476–3482.

Edward, E. J., King, W. S., Teck, S. L. C., Jiwan, M., Aziz, Z. F. A., Kundat, F. R., … Majid, N. M. A. (2013). Antagonistic activities of endophytic bacteria against fusarium wilt of black pepper (Piper nigrum). International Journal of Agriculture and Biology, 15(2), 291–296.

Ehrlich, P. R., & Raven, P. H. (1964). Butterflies and Plants: A Study in Coevolution. Evolution, 18(4), 586–608.

Endara, M., Coley, P. D., Ghabash, G., Nicholls, J. A., Dexter, K. G., Donoso, D. A., … Kursar, T. A. (2017). Coevolutionary arms race versus host defense chase in a tropical herbivore – plant system. Proceedings of the National Academy of Sciences, 114(36), E7499–E7505.

Estrada, C., Rojas, E. I., Wcislo, W. T., & Van Bael, S. A. (2014). Fungal endophyte effects on leaf chemistry alter the in vitro growth rates of leaf-cutting ants’ fungal mutualist, Leucocoprinus gongylophorus. Fungal Ecology, 8(1), 37–45.

Estendorfer, J., Stempfhuber, B., Haury, P., Vestergaard, G., Rillig, M. C., Joshi, J., … Schloter, M. (2017). The influence of land use intensity on the plant-associated microbiome of Dactylis glomerata L. Frontiers in Plant Science, 8(930).

Fincher, R. M., Dyer, L. a., Dodson, C. D., Richards, J. L., Tobler, M. a., Searcy, J., … Pidcock, W. (2008). Inter- and intraspecific comparisons of antiherbivore defenses in three species of rainforest understory shrubs. Journal of Chemical Ecology, 34(4), 558–574.

129

Fleming, T. H. (1987). Fruit bats: prime movers of tropical seeds. Bats, 5(3), 3-5.

Fleming, T. H. (1991). The Relationship between Body Size, Diet, and Habitat Use in Frugivorous Bats, Genus Carollia (Phyllostomidae). Journal of Mammalogy, 72(3), 493–501.

Fleming, T. H. (2004). Dispersal ecology of neotropical Piper shrubs and treelets. In L. A. Dyer & N. A. Palmer (Eds.), Piper: A model genus for studies of phytochemistry, ecology and evolution (pp. 58–77). New York, NY: Kluwer Academic Publishers.

Forister, M. L., Novotny, V., Panorska, A. K., Baje, L., Basset, Y., Butterill, P. T., … Dyer, L. A. (2015). The global distribution of diet breadth in insect herbivores. Proceedings of the National Academy of Sciences of the United States of America, 112(2), 442–447.

Fraenkel, G. S. (1959). The Raison d’Etre of Secondary Plant Substances. Science, 129(3361), 1466–1470.

Fricke, E. C., Simon, M. J., Reagan, K. M., Levey, D. J., Riffell, J. A., Carlo, T. A., & Tewksbury, J. J. (2013). When condition trumps location: seed consumption by fruit- eating birds removes pathogens and predator attractants. Ecology Letters, 16(8), 1031–1036.

Fröhlich, J., & Hyde, K. D. (1999). Biodiversity of palm fungi in the tropics: Are global fungal diversity estimates realistic? Biodiversity and Conservation, 8(7), 977–1004.

Gallery, R. E., Dalling, J. W., & Arnold, E. A. (2007). Diversity, Host Affinity, and Distribution of Seed-Infecting Fungi: a Case Study With Cecropia. Ecology, 88(3), 582–588.

130

Gange, A. C., Koricheva, J., Currie, A. F., Jaber, L. R., & Vidal, S. (2019). Meta‐analysis of the role of entomopathogenic and unspecialised fungal endophytes as plant bodyguards. New Phytologist, 223, 2002–2010.

Gardes, M., & Bruns, T. D. (1993). ITS primers with enhanced specificity for basidiomycetes ‐ application to the identification of mycorrhizae and rusts. Molecular Ecology, 2(2), 113–118.

Gauld I. D., Gaston K. J., Janzen D. H., 1992. Plant allelochemicals, tritrophic interactions and the anomalous diversity of tropical parasitoids: the "nasty" host hypothesis. Oikos 65:353-357.

Gdanetz, K., & Trail, F. (2017). The wheat microbiome under four management strategies, and potential for endophytes in disease protection. Phytobiomes Journal, 1(3), 158–168.

Geber, M. A., & Griffen, L. R. (2003). Inheritance and natural selection on functional traits. International Journal of Plant Sciences, 164(S3), S21–S24.

Gentry, G. L., & Dyer, L. A. (2002). On the conditional nature of neotropical caterpillar defenses against their natural enemies. Ecology, 83(11), 3108–3119.

Glassmire, A. E., Jeffrey, C. S., Forister, M. L., Parchman, T. L., Nice, C. C., Jahner, J. P., Wilson, J. S., Walla, T. R., Richards, L. A., Smilanich, A. M., Leonard, M. D., Morrison, C. R., Simbana, W., Salagaje, L. A., Dodson, C. D., Miller, J. S., Tepe, E. J., Villamarin-Cortez, S., & Dyer, L. A. (2016). Intraspecific phytochemical variation shapes community and population structure for specialist caterpillars. New Phytologist, 212(1), 208–219.

Glassmire, A. E. (2017). Intraspecific phytochemical variation as mosaics of defense in tropical forests. [Doctoral dissertation, University of Nevada Reno]. ScholarWorks.

Glassmire, A. E., Philbin, C., Richards, L. A., Jeffrey, C. S., Snook, J. S., & Dyer, L. A. (2019). Proximity to canopy mediates changes in the defensive chemistry and herbivore loads of an understory tropical shrub, Piper kelleyi. Ecology Letters, 22(2), 332–341.

131

González-Santoyo, I., & Córdoba-Aguilar, A. (2012). Phenoloxidase: A key component of the insect immune system. Entomologia Experimentalis et Applicata, 142(1), 1– 16.

Goswami, R. S., & Kistler, H. C. (2004). Heading for disaster: Fusarium graminearum on cereal crops. Molecular Plant Pathology, 5(6), 515–525.

Gotelli, N. J. & Ellison, A. M. (2013). A primer of ecological statistics (2nd edition). Sunderland, MA: Sinauer Associates Inc.

Grace, J. B., Michael Anderson, T., Smith, M. D., Seabloom, E., Andelman, S. J., Meche, G., ... & Knops, J. (2007). Does species diversity limit productivity in natural grassland communities?. Ecology Letters, 10(8), 680-689.

Greeney, H. F., Dyer, L., & Smilanich, M. (2012). Feeding by lepidopteran larvae is dangerous: A review of caterpillars’ chemical, physiological, morphological, and behavioral defenses against natural enemies. Invertebrate Survival Journal, 9(1), 7–34.

Greig, N. (1993a). Predispersal Seed Predation on 5 Piper Species in Tropical Rain- Forest. Oecologia, 93(3), 412–420.

Greig, N. (1993b). Regeneration Mode in Neotropical Piper: Habitat and Species Comparisons. Ecology, 74(7), 2125–2135.

Greig, N. (2004). Introduction. In L. A. Dyer & N. A. Palmer (Eds.), Piper: A model genus for studies of phytochemistry, ecology and evolution (pp. 1–4). New York, NY: Kluwer Academic Publishers.

Hairston, N. G., Smith, F. E., & Slobodkin, L. B. (1960). Community structure, population control and competition. The American Naturalist, 94(879), 421–425.

Hansen, A. C., Glassmire, A. E., Dyer, L. A., & Smilanich, A. M. (2017). Patterns in parasitism frequency explained by diet and immunity. Ecography, 40(7), 803–805.

Hansen, T. F., Pélabon, C., & Houle, D. (2011). Heritability is not Evolvability. Evolutionary Biology, 38(3), 258–277. https://doi.org/10.1007/s11692-011-9127-6

132

Hardoim, P. R., van Overbeek, L. S., Berg, G., Pirttilä, A. M., Compant, S., Campisano, A., … Sessitsch, A. (2015). The Hidden World within Plants: Ecological and Evolutionary Considerations for Defining Functioning of Microbial Endophytes. Microbiology and Molecular Biology Reviews, 79(3), 293–320.

Hardoim (2019). The ecology of seed microbiota. In S. K. Verma & J. F. White Jr (Eds.), Seed Endophytes (pp. 103–125). Cham, Switzerland: Springer.

Harrison, J. G., Parchman, T. L., Cook, D., Gardner, D. R., & Forister, M. L. (2018a). A heritable symbiont and host-associated factors shape fungal endophyte communities across spatial scales. Journal of Ecology, 106(6), 2274–2286.

Harrison, X. A., Donaldson, L., Correa-Cano, M. E., Evans, J., Fisher, D. N., Goodwin, C. E. D., … Inger, R. (2018b). A brief introduction to mixed effects modelling and multi-model inference in ecology. PeerJ, 6, e4794.

Harrison, J. G., John Calder, W., Shuman, B., & Alex Buerkle, C. (2021). The quest for absolute abundance: The use of internal standards for DNA-based community ecology. Molecular Ecology Resources, 21, 30–43.

Hartley, S. E., & Gange, A. C. (2009). Impacts of Plant Symbiotic Fungi on Insect Herbivores: Mutualism in a Multitrophic Context. Annual Review of Entomology, 54(1), 323–342.

Haviola, S., Kapari, L., Ossipov, V., Rantala, M. J., Ruuhola, T., & Haukioja, E. (2007). Foliar phenolics are differently associated with Epirrita autumnata growth and immunocompetence. Journal of Chemical Ecology, 33(5), 1013–1023.

Hawkins, B. A., Cornell, H. V., & Hochberg, M. E. (1997). Predators, parasitoids, and pathogens as mortality agents in phytophagous insect populations. Ecology, 78(7), 2145–2152.

Hay, M.E., Kappel, Q.E., & Fenical, W. (1994). Synergisms in plant defenses against herbivores - interactions of chemistry, calcification, and plant-quality. Ecology, 75(6), 1714-1726.

133

Hernandez-Restrepo, M., Groenewald, J. Z., & Crous, P. W. (2016). Taxonomic and phylogenetic re-evaluation of Microdochium, Monographella and Idriella. Persoonia: Molecular Phylogeny and Evolution of Fungi, 36, 57–82.

Hill, M. O. (1973). Diversity and Evenness: A Unifying Notation and Its Consequences. Ecology, 54(2), 427–432.

Hodgson, S., de Cates, C., Hodgson, J., Morley, N. J., Sutton, B. C., & Gange, A. C. (2014). Vertical transmission of fungal endophytes is widespread in forbs. Ecology and Evolution, 4(8), 1199–1208.

Howe, H. F., & Smallwood, J. (1982). Ecology of Seed Dispersal. Annual Review of Ecology and Systematics, 13(1), 201–228.

Humphrey, P. T., & Whiteman, N. K. (2019). Insect herbivory reshapes a native leaf microbiome. Nature Ecology and Evolution, 4(2), 221–229.

Hunter, M. D. (2016). The phytochemical landscape: linking trophic interactions and nutrient dynamics. Princeton University Press.

Hunter, M. D., & Price, P. W. (1992). Playing chutes and ladders: Heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology, 73(3), 724–732.

Ilarslan, H., Palmer, R. G., & Horner, H. T. (2001). Calcium Oxalate Crystals in Developing Seeds of Soybean. Annals of Botany, 88(2), 243–257.

Ives, A. R., Cardinale, B. J., & Snyder, W. E. (2005). A synthesis of subdisciplines: Predator-prey interactions, and biodiversity and ecosystem functioning. Ecology Letters, 8, 102–116.

Janzen, D. H. (1970). Herbivores and the Number of Tree Species in Tropical Forests. The American Naturalist, 104(940), 501–528.

Janzen, D. H. (1977). Why fruits rot, seeds mold and meat spoils. The American Naturalist, 111(980), 691–713.

134

Janzen, D. H., Hallwachs, W., Burns, J. M., Hajibabaei, M., Bertrand, C., & Hebert, P. D. N. (2011). Reading the complex skipper butterfly fauna of one tropical place. PLoS ONE, 6(8), e19874.

Johnson, M. T. J. (2008). Bottom-up effects of plant genotype on aphids, ants, and predators. Ecology, 89(1), 145–154.

Johnson, M. T. J., Agrawal, A. A., Maron, J. L., & Salminen, J. P. (2009). Heritability, covariation and natural selection on 24 traits of common evening primrose (Oenothera biennis) from a field experiment. Journal of Evolutionary Biology, 22(6), 1295–1307.

Jones, J. D. G., & Dangl, J. L. (2006). The plant immune system. Nature, 444(7117), 323–329.

Jones, C. G., & Firn, R. (1991). On the evolution of plant secondary chemical diversity. Philosophical Transactions of the Royal Society B: Biological Sciences, 333, 273– 280.

Jones, P., Garcia, B. J., Furches, A., Tuskan, G. A., & Jacobson, D. (2019). Plant Host- Associated Mechanisms for Microbial Selection. Frontiers in Plant Science, 10, 862.

Jordano, P., Garcia, C., Godoy, J. A., & Garcia-Castano, J. L. (2007). Differential contribution of frugivores to complex seed dispersal patterns. Proceedings of the National Academy of Sciences, 104(9), 3278–3282.

Jost, L. (2006). Entropy and diversity. Oikos, 113(2), 363–375.

Jost, L. (2007). Partitioning diversity into independent alpha and beta components. Ecology, 88(10), 2427–2439.

Karban, R., & Baldwin, I. T., (1997). Induced Responses to Herbivory. Chicago, IL: The University Chicago Press.

Kato, M. J., & Furlan, M. (2007). Chemistry and evolution of the Piperaceae. Pure and Applied Chemistry, 79(4), 529–538.

135

Keddy, P. A. (1992). Assembly and response rules: two goals for predictive community ecology. Journal of Vegetation Science, 3(2), 157–164.

Kembel, S. W., & Mueller, R. C. (2014). Plant traits and taxonomy drive host associations in tropical phyllosphere fungal communities. Botany, 92(4), 303–311.

Kembel, S. W., O’Connor, T. K., Arnold, H. K., Hubbell, S. P., Wright, S. J., & Green, J. L. (2014). Relationships between phyllosphere bacterial communities and plant functional traits in a neotropical forest. Proceedings of the National Academy of Sciences, 111(38), 13715–13720.

Kerscher, S. J., Okun, J. G., & Brandt, U. (1999). A single external enzyme confers alternative NADH: ubiquinone oxidoreductase activity in Yarrowia lipolytica. Journal of Cell Science, 112(14), 2347–2354.

Kessler, A., & Kalske, A. (2018). Plant Secondary Metabolite Diversity and Species Interactions. Annual Review of Ecology, Evolution, and Systematics, 49(1), 115– 138.

Kessner, D., Chambers, M., Burke, R., Agus, D. & Mallick, P. (2008) ProteoWizard: open source software for rapid proteomics tools development. Bioinformatics, 24, 2534-2536.

Kõljalg, U., Larsson, K.-H., Abarenkov, K., Nilsson, R. H., Alexander, I. J., Eberhardt, U., … Ursing, B. M. (2005). UNITE: a database providing web-based methods for the molecular identification of ectomycorrhizal fungi. New Phytologist, 166(3), 1063–1068.

Korth, K. L., Doege, S. J., Park, S.-H., Goggin, F. L., Wang, Q., Gomez, S. K., … Nakata, P. A. (2006). Medicago truncatula Mutants Demonstrate the Role of Plant Calcium Oxalate Crystals as an Effective Defense against Chewing Insects. Plant Physiology, 141(1), 188–195.

Kraft, N. J. B., Adler, P. B., Godoy, O., James, E. C., Fuller, S., & Levine, J. M. (2015). Community assembly, coexistence and the environmental filtering metaphor. Functional Ecology, 29(5), 592–599.

136

Kuhl, C., Tautenhahn, R., Bottcher, C., Larson, T. R. & Neumann, S. (2012) CAMERA: an integrated strategy for compound spectra extraction and annotation of liquid chromatography/mass spectrometry data sets. Analytical chemistry, 84, 283-289.

Kursar, T. a, Dexter, K. G., Lokvam, J., Pennington, R. T., Richardson, J. E., Weber, M. G., … Coley, P. D. (2009). The evolution of antiherbivore defenses and their contribution to species coexistence in the tropical tree genus Inga. Proceedings of the National Academy of Sciences, 106(43), 18073–18078.

Lamit, L. J., Lau, M. K., Sthultz, C. M., Wooley, S. C., Whitham, T. G., & Gehring, C. A. (2014). Tree genotype and genetically based growth traits structure twig endophyte communities. American Journal of Botany, 101(3), 467–478.

Lamit, L. J., Busby, P. E., Lau, M. K., Compson, Z. G., Wojtowicz, T., Keith, A. R., … Whitham, T. G. (2015). Tree genotype mediates covariance among communities from microbes to lichens and arthropods. Journal of Ecology, 103(4), 840–850.

Lamit, L. J., Holeski, L. M., Flores-Rentería, L., Whitham, T. G., & Gehring, C. A. (2016). Tree genotype influences ectomycorrhizal fungal community structure: Ecological and evolutionary implications. Fungal Ecology, 24, 124–134.

Lampert, E. (2012). Influences of plant traits on immune responses of specialist and generalist herbivores. Insects, 3(2), 573–592.

Lampert, E. C., & Bowers, M. D. (2015). Incompatibility between plant-derived defensive chemistry and immune response of two Sphingid herbivores. Journal of Chemical Ecology, 41(1), 85–92.

Lampert, E. C., Dyer, L. A., & Bowers, M. D. (2010). Caterpillar chemical defense and parasitoid success: Cotesia congregata parasitism of Ceratomia catalpae. Journal of Chemical Ecology, 36(9), 992–998.

Lampert, E. C., Zangerl, A. R., Berenbaum, M. R., & Ode, P. J. (2011). Generalist and specialist host-parasitoid associations respond differently to wild parsnip (Pastinaca sativa) defensive chemistry. Ecological Entomology, 36(1), 52–61.

137

Larkin, B. G., Hunt, L. S., & Ramsey, P. W. (2012). Foliar nutrients shape fungal endophyte communities in Western white pine (Pinus monticola) with implications for white-tailed deer herbivory. Fungal Ecology, 5(2), 252–260.

Lau, M. K., Arnold, A. E., & Johnson, N. C. (2013). Factors influencing communities of foliar fungal endophytes in riparian woody plants. Fungal Ecology, 6(5), 365–378.

Lefcheck, J. S. (2016). piecewiseSEM: Piecewise structural equation modelling in r for ecology, evolution, and systematics. Methods in Ecology and Evolution, 7(5), 573– 579.

Legendre, P., & Gallagher, E. D. (2001). Ecologically meaningful transformations for ordination of species data. Oecologia, 129(2), 271–280.

Levin, D. A. (1976). The Chemical Defenses of Plants to Pathogens and Herbivores. Annual Review of Ecology and Systematics, 7(61), 121–159.

Li, X., Jousset, A., de Boer, W., Carrión, V. J., Zhang, T., Wang, X., & Kuramae, E. E. (2019). Legacy of land use history determines reprogramming of plant physiology by soil microbiome. ISME Journal, 13(3), 738–751.

Liaw, A. & Wiener, M. (2002). Classification and Regression by randomForest. R News 2(3), 18–22.

Lill, J. T., Marquis, R. J., & Ricklefs, R. E. (2002). Host plants influence parasitism of forest caterpillar. Nature, 417(6885), 170–173.

Lobova, T. A., Mori, S. A., Blanchard, F., Peckham, H., & Charles-Dominique, P. (2003). Cecropia as a food resource for bats in French Guiana and the significance of fruit structure in seed dispersal and longevity. American Journal of Botany, 90(3), 388– 403.

Lopez, J. E., & Vaughan, C. (2007). Food niche overlap among neotropical frugivorous bats in Costa Rica. Revista de Biología Tropical, 55(1), 301–313.

138

Lopez-Vaamonde, C., Sire, L., Rasmussen, B., Rougerie, R., Wieser, C., Allaoui, A. A., … Lees, D. C. (2019). DNA barcodes reveal deeply neglected diversity and numerous invasions of micromoths in Madagascar. Genome, 62(3), 108–121.

Mandyam, K., Fox, C., & Jumpponen, A. (2012). Septate endophyte colonization and host responses of grasses and forbs native to a tallgrass prairie. Mycorrhiza, 22(2), 109–119.

Marcet-Houben, M., Marceddu, G., & Gabaldón, T. (2009). Phylogenomics of the oxidative phosphorylation in fungi reveals extensive gene duplication followed by functional divergence. BMC Evolutionary Biology, 9(1), 295.

Marquis, R. (1990). Genotypic Variation in Leaf Damage in Piper arieianum (Piperaceae) by a Multispecies Assemblage of Herbivores. Evolution, 44(1), 104– 120.

Marquis, R. J. (1991). Herbivore fauna of Piper (Piperaceae) in a Costa Rican wet forest: Diversity, specificity, and impact. In P. W. Price, T. M. Lewinsohn, G. W. Fernandes, and W. W. Benson (eds). Plant–Animal Interactions: Evolutionary Ecology in Tropical and Temperate Regions (pp. 179–208). New York, NY: John Wiley & Sons.

Maynard, L. D., Slinn, H. L., Glassmire, A. E., Matarrita-Carranza, B., Dodson, C. D., Nguyen, T. T., Burroughs, M. J., Dyer, L. A., Jeffrey, C. S., & Whitehead, S. R. 2020. Secondary metabolites in a neotropical shrub: spatiotemporal allocation and role in fruit defence and dispersal. Ecology e03192.

McCall, A. C., & Fordyce, J. A. (2010). Can optimal defence theory be used to predict the distribution of plant chemical defences? Journal of Ecology, 98(5), 985–992.

Mcguire, K. L., Fierer, N., Bateman, C., Treseder, K. K., & Turner, B. L. (2012). Fungal Community Composition in Neotropical Rain Forests: the Influence of Tree Diversity and Precipitation. Microbial Ecology, 63(4), 804–812.

McKey, D. (1974). Adaptive Patterns in Alkaloid Physiology. The American Naturalist, 108(961), 305–320.

139

McMurdie, P. J., & Holmes, S. (2014). Waste Not, Want Not: Why Rarefying Microbiome Data Is Inadmissible. PLoS Computational Biology, 10(4), e1003531.

Mejía, L. C., Herre, E. A., Sparks, J. P., Winter, K., García, M. N., Van Bael, S. A., … Maximova, S. N. (2014). Pervasive effects of a dominant foliar endophytic fungus on host genetic and phenotypic expression in a tropical tree. Frontiers in Microbiology, 5, 479.

Moore, B. D., Andrew, R. L., Külheim, C., & Foley, W. J. (2014). Explaining intraspecific diversity in plant secondary metabolites in an ecological context. New Phytologist, 201(3), 733–750.

Morin, P. J. (2011). Community Ecology. 2nd ed. Wiley-Blackwell, Hoboken, NJ.

Myers, J. A., & Harms, K. E. (2009). Seed arrival, ecological filters, and plant species richness: a meta-analysis. Ecology Letters, 12(11), 1250–1260.

Myers, N., Mittermeier, R. A., Mittermeier, C. G., Fonseca, G. A. B., & Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nature, 403(6772), 853–858.

Nakata, P. A. (2002). Calcium oxalate crystal morphology. Trends in Plant Science, 7(7), 324.

Nakata, P. A. (2012). Plant calcium oxalate crystal formation, function, and its impact on human health. Frontiers in Biology, 7(3), 254–266.

Naranjo, M. E., Rengifo, C., & Soriano, P. J. (2003). Effect of ingestion by bats and birds on seed germination of Stenocereus griseus and Subpilocereus repandus (Cactaceae). Journal of Tropical Ecology, 19(1), 19–25.

Nelson, E. B. (2018). The seed microbiome: Origins, interactions, and impacts. Plant and Soil, 422(1–2), 7–34.

Nemergut, D. R., Schmidt, S. K., Fukami, T., O’Neill, S. P., Bilinski, T. M., Stanish, L. F., … Ferrenberg, S. (2013). Patterns and Processes of Microbial Community Assembly. Microbiology and Molecular Biology Reviews, 77(3), 342–356.

140

Newcombe, G., Harding, A., Ridout, M., & Busby, P. E. (2018). A Hypothetical Bottleneck in the Plant Microbiome. Frontiers in Microbiology, 9, 1645.

Newman, J. A., Gillis, S., & Hager, H. A. (unpublished). A Note on the Use and Abuse of the ‘Mutualism – Parasitism Continuum’ Concept for Epichloe Fungi.

Nisa, H., Kamili, A. N., Nawchoo, I. A., Shafi, S., Shameem, N., & Bandh, S. A. (2015). Fungal endophytes as prolific source of phytochemicals and other bioactive natural products: A review. Microbial Pathogenesis, 82, 50–59.

Nilsson, R. H., Ryberg, M., Kristiansson, E., Abarenkov, K., Larsson, K. H., & Köljalg, U. (2006). Taxonomic reliability of DNA sequences in public sequences databases: A fungal perspective. PLoS ONE, 1(1), e59.

Nilsson, R. H., Larsson, K. H., Taylor, A. F. S., Bengtsson-Palme, J., Jeppesen, T. S., Schigel, D., … Abarenkov, K. (2019). The UNITE database for molecular identification of fungi: Handling dark taxa and parallel taxonomic classifications. Nucleic Acids Research, 47(D1), D259–D264.

Nguyen, N. H., Smith, D., Peay, K., & Kennedy, P. (2015). Parsing ecological signal from noise in next generation amplicon sequencing. New Phytologist, 205(4), 1389–1393.

Nguyen, N. H., Song, Z., Bates, S. T., Branco, S., Tedersoo, L., Menke, J., … Kennedy, P. G. (2016). FUNGuild: An open annotation tool for parsing fungal community datasets by ecological guild. Fungal Ecology, 20, 241–248.

Novotny, V., Drozd, P., Miller, S. E., Kulfan, M., Janda, M., Basset, Y., & Weiblen, G. D. (2006). Why are there so many species of herbivorous insects in tropical rainforests? Science, 313(5790), 1115–1118.

Ochman, H., Lawrence, J. G., & Groisman, E. A. (2000). Lateral gene transfer and the nature of bacterial innovation. Nature, 405(6784), 299–304.

O’Connor, M. I., Gonzalez, A., Byrnes, J. E. K., Cardinale, B. J., Duffy, J. E., Gamfeldt, L., Griffin, J. N., Hooper, D., Hungate. B. A., Paquette, A., Thompson, P. L., Dee, L.

141

E., & Dolan, K. L. (2016). A general biodiversity–function relationship is mediated by trophic level. Oikos, 126(1), 18–31.

O’Donnell, A., Harvey, L. M., & McNeil, B. (2011). The roles of the alternative NADH dehydrogenases during oxidative stress in cultures of the filamentous fungus Aspergillus niger. Fungal Biology, 115(4–5), 359–369.

Oksanen, J. F., Blanchet, G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P. R., O'Hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H., Szoecs, E. & Wagner, H. (2019). vegan: Community Ecology Package. R package version 2.5-5. https://CRAN.R-project.org/package=vegan

Orjala, J., Mian, P., Rali, T., & Sticher, O. (1998). Gibbilimbols A-D, cytotoxic and antibacterial alkenylphenols from Piper gibbilimbum. Journal of Natural Products, 61(7), 939–941.

Ouborg, N. J., Piquot, Y., & Van Groenendael, J. M. (1999). Population genetics, molecular markers and the study of dispersal in plants. Journal of Ecology, 87(4), 551–568.

Panaccione, D. G., Beaulieu, W. T., & Cook, D. (2014). Bioactive alkaloids in vertically transmitted fungal endophytes. Functional Ecology, 28(2), 299–314.

Parmar, V. S., Jain, S. C., Bisht, K. S., Jain, R., Taneja, P., Jha, A., … Boll, P. M. (1997). Phytochemistry of the genus Piper. Phytochemistry, 46(4), 597–673.

Peay, K. G., Kennedy, P. G., & Talbot, J. M. (2016). Dimensions of biodiversity in the Earth mycobiome. Nature Reviews Microbiology, 14(7), 434–447.

Pedros-Alio, C. (2006). Marine microbial diversity: can it be determined? Trends in Microbiology, 14(6), 257–263.

Peiffer, J. A., Spor, A., Koren, O., Jin, Z., Tringe, S. G., Dangl, J. L., … Ley, R. E. (2013). Diversity and heritability of the maize rhizosphere microbiome under field conditions. Proceedings of the National Academy of Sciences of the United States of America, 110(16), 6548–6553.

142

Peris, J. E., Rodríguez, A., Penã, L., & Fedriani, J. M. (2017). Fungal infestation boosts fruit aroma and fruit removal by mammals and birds. Scientific Reports, 7(5646).

Price, P. W., Bouton, C. E., Gross, P., McPheron, B. A., Thompson, J. N., & Weis, A. E. (1980). Interactions among three trophic levels: Influence of plants on interactions between insect herbivores and natural enemies. Annual Review of Ecology and Systematics, 11, 41–65.

Price, P. W. (2002). Resource-driven terrestrial interaction webs. Ecological Research, 17(2), 241–247.

Pusztahelyi, T., Holb, I. J., & Pócsi, I. (2015). Secondary metabolites in fungus-plant interactions. Frontiers in Plant Science, 6, 573.

Quintero, C., Lampert, E. C., & Bowers, M. D. (2014). Time is of the essence: direct and indirect effects of plant ontogenetic trajectories on higher trophic levels. Ecology, 95(9), 2589–2602.

R Core Team. (2019). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. www.R-project.org/

R Core Team. (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. Vienna, Austria. www.R-project.org/

Rajala, T., Velmala, S. M., Vesala, R., Smolander, A., & Pennanen, T. (2014). The community of needle endophytes reflects the current physiological state of Norway spruce. Fungal Biology, 118(3), 309–315.

Ralphs, M. H., & Stegelmeier, B. L. (2011). Locoweed Toxicity, Ecology, Control, and Management. International Journal of Poisoning Plant Research, 1, 47–64.

Raman, V., Galal, A. M., & Khan, I. A. (2012). An Investigation of the Vegetative Anatomy of Piper sarmentosum, and a Comparison with the Anatomy of Piper betle (Piperaceae). American Journal of Plant Sciences, 3(8), 1135–1144.

143

Reudler, J. H., Biere, A., Harvey, J. A., & van Nouhuys, S. (2011). Differential performance of a specialist and two generalist herbivores and their parasitoids on Plantago lanceolata. Journal of Chemical Ecology, 37(7), 765–778.

Revelle W (2019). psych: Procedures for Psychological, Psychometric, and Personality Research. Northwestern University, Evanston, Illinois. R package version 1.9.12, https://CRAN.R-project.org/package=psych.

Rezki, S., Campion, C., Simoneau, P., Jacques, M. A., Shade, A., & Barret, M. (2018). Assembly of seed-associated microbial communities within and across successive plant generations. Plant and Soil, 422(1–2), 67–79.

Richards, L. A., Dyer, L. A., Smilanich, A. M., & Dodson, C. D. (2010). Synergistic effects of amides from two Piper species on generalist and specialist herbivores. Journal of Chemical Ecology, 36(10), 1105–1113.

Richards, L. A., Dyer, L. A., Forister, M. L., Smilanich, A. M., Dodson, C. D., Leonard, M. D., & Jeffrey, C. S. (2015). Phytochemical diversity drives plant–insect community diversity. Proceedings of the National Academy of Sciences, 112(35), 10973–10978.

Richards, L. A., Glassmire, A. E., Ochsenrider, K. M., Smilanich, A. M., Dodson, C. D., Jeffrey, C. S., & Dyer, L. A. (2016). Phytochemical diversity and synergistic effects on herbivores. Phytochemistry Reviews, 15(6), 1153–1166.

Richards, L. a., Lampert, E. C., Bowers, M. D., Dodson, C. D., Smilanich, A. M., & Dyer, L. a. (2012). Synergistic effects of iridoid glycosides on the survival, development and immune response of a specialist caterpillar, Junonia coenia (Nymphalidae). Journal of Chemical Ecology, 38(10), 1276–1284.

Robinson, M. D., McCarthy, D. J., & Smyth, G. K. (2010). edgeR: A Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics, 26(1), 139–140.

Robinson, M. D., & Oshlack, A. (2010). A scaling normalization method for differential expression analysis of RNA-seq data. Genome Biology, 11(3), R25.

144

Rodrigues, A. A. C., & Menezes, M. (2005). Identiﬁcation and pathogenic characterization of endophytic Fusarium species from cowpea seeds. Mycopathologia, 159, 79–85.

Rodriguez, R. J., White, J. F., Arnold, A. E., & Redman, R. S. (2009). Fungal endophytes: Diversity and functional roles. New Phytologist, 182(2), 314–330.

Rodríguez-Castañeda, G., Dyer, L. A., Brehm, G., Connahs, H., Forkner, R. E., & Walla, T. R. (2010). Tropical forests are not flat: How mountains affect herbivore diversity. Ecology Letters, 13(11), 1348–1357.

Rodríguez-Castañeda, G., Forkner, R. E., Tepe, E. J., Gentry, G. L., & Dyer, L. A. (2011). Weighing defensive and nutritive roles of ant mutualists across a tropical altitudinal gradient. Biotropica, 43(3), 343–350.

Rodríguez-Castañeda, G., Brehm, G., Fiedler, K., & Dyer, L. (2016). Ant predation on herbivores through a multitrophic lens: How effects of ants on plant herbivore defense and natural enemies vary along temperature gradients. Current Opinion in Insect Science, 14, 73–80.

Rognes, T., Flouri, T., Nichols, B., Quince, C., & Mahé, F. (2016). VSEARCH: A versatile open source tool for metagenomics. PeerJ, 4, e2584.

Rosseel Y., Oberski, D., Byrnes, J., Vanbrabant, L., Savalei, V., Merkle, E., Hallquist, M., Rhemtulla, M., Katsikatsou, M., Barendse, M., & Chow, M. (2017). Package ‘lavaan’. Retrieved on January 16, 2018. https://CRAN.R- project.org/package=lavaan

Sala, O. E., Chapin, F. S., Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo, R., ... & Leemans, R. (2000). Global biodiversity scenarios for the year 2100. Science, 287(5459), 1770-1774.

Salazar, D., Jaramillo, M. A., & Marquis, R. J. (2016a). Chemical similarity and local community assembly in the species rich tropical genus Piper. Ecology, 97(11), 3176–3183.

145

Salazar, D., Jaramillo, A., & Marquis, R. J. (2016b). The impact of plant chemical diversity on plant – herbivore interactions at the community level. Oecologia, 181(4), 1199–1208.

Salcido, D. M., Forister, M. L., Lopez, H. G., & Dyer, L. A. (2020). Loss of dominant caterpillar genera in a protected tropical forest. Scientific Reports, 10(422), 1–10.

Saldaña-Vázquez, R. A., Castaño, J. H., Baldwin, J., & Perez-Torres, J. (2019). Does seed ingestion by bats enhance germination? A new analysis 15 years later. Mammal Review, 49(3), 201–209.

Salehi, B., Zakaria, Z. A., Gyawali, R., Ibrahim, S. A., Rajkovic, J., Shinwari, Z. K., … Setzer, W. N. (2019). Piper Species: A Comprehensive Review on Their Phytochemistry, Biological Activities and Applications. Molecules, 24(7), 1364.

Sandberg, S. L., & Berenbaum, M. R. (1989). Leaf-tying by tortricid larvae as an adaptation for feeding on phototoxic Hypericum perforatum. Journal of Chemical Ecology, 15(3), 875–885.

Sarmiento, C., Zalamea, P.-C., Dalling, J. W., Davis, A. S., Stump, S. M., U’Ren, J. M., & Arnold, A. E. (2017). Soilborne fungi have host affinity and host-specific effects on seed germination and survival in a lowland tropical forest. Proceedings of the National Academy of Sciences, 114(43), 11458–11463.

Saunders, M., & Kohn, L. M. (2009). Evidence for alteration of fungal endophyte community assembly by host defense compounds. New Phytologist, 182(1), 229– 238.

Scherber, C., Eisenhauer, N., Weisser, W. W., Schmid, B., Voigt, W., Fischer, M., … Tscharntke, T. (2010). Bottom-up effects of plant diversity on multitrophic interactions in a biodiversity experiment. Nature, 468(7323), 553–556.

Schlatter, D. C., Yin, C., Hulbert, S., & Paulitz, T. C. (2020). Core rhizosphere microbiomes of dryland wheat are influenced by location and land-use history. Applied and Environmental Microbiology, 86(5), e02135-19.

146

Schloerke, B., Cook, D., Larmarange, J., Briatte, F., Marbach, M., Thoen, E., Elberg, A., & Crowley, J. (2020). GGally: Extension to 'ggplot2'. R package version 2.0.0. https://CRAN.R-project.org/package=GGally

Schmidt, S. K., Costello, E. K., Nemergut, D. R., Cleveland, C. C., Reed, S. C., Weintraub, M. N., … Martin, A. M. (2007). Biogeochemical consequences of rapid microbial turnover and seasonal succession in soil. Ecology, 88(6), 1379–1385.

Schneider, C. A., Rasband, W. S., & Eliceiri, K. W. (2012). 10 NIH Image to ImageJ 25 years of image analysis. Nature Methods, 9(7), 671–676.

Schoch, C. L., Seifert, K. A., Huhndorf, S., Robert, V., Spouge, J. L., Levesque, C. A., … Schindel, D. (2012). Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proceedings of the National Academy of Sciences, 109(16), 6241–6246.

Schoonhoven, L. M., Van Loon, J. J., & Dicke, M. (2005). Insect-plant biology. Oxford University Press on Demand.

Schulz, B., & Boyle, C. (2005). The endophytic continuum. Mycological Research, 109(6), 661–686.

Schupp, E. W., Jordano, P., & Gómez, J. M. (2010). Seed dispersal effectiveness revisited: A conceptual review. New Phytologist, 188(2), 333–353.

Shahnazi, S., Meon, S., Vadamalai, G., Ahmad, K., & Nejat, N. (2012). Morphological and molecular characterization of Fusarium spp. associated with yellowing disease of black pepper (Piper nigrum L.) in Malaysia. Journal of General Plant Pathology, 78(3), 160–169.

Shahzad, R., Khan, A. L., Bilal, S., Asaf, S., & Lee, I.-J. (2018). What Is There in Seeds? Vertically Transmitted Endophytic Resources for Sustainable Improvement in Plant Growth. Frontiers in Plant Science, 9, 24.

Shapiro, B. J., Friedman, J., Cordero, O. X., Preheim, S. P., Timberlake, S. C., Szabó, G., … Alm, E. J. (2012). Population Genomics of Early Events in the Ecological Differentiation of Bacteria. Science, 336(6077), 48–51.

147

Simon, S. J., Tschaplinski, T. J., M. LeBoldus, J., Keefover-Ring, K., Azeem, M., Chen, J. G., … DiFazio, S. P. (2020). Host plant genetic control of associated fungal and insect species in a Populus hybrid cross. Ecology and Evolution, 10, 5119–5134.

Singer, M. S., Mason, P. A., & Smilanich, A. M. (2014). Ecological immunology mediated by diet in herbivorous insects. Integrative and Comparative Biology, 54(5), 913–921.

Singer, M. S., & Stireman, J. O. (2005). The tri-trophic niche concept and adaptive radiation of phytophagous insects. Ecology Letters, 8, 1247–1255.

Slinn, H. L., Richards, L. A., Dyer, L. A., Hurtado, P. J., & Smilanich, A. M. (2018). Across Multiple Species, Phytochemical Diversity and Herbivore Diet Breadth Have Cascading Effects on Herbivore Immunity and Parasitism in a Tropical Model System. Frontiers in Plant Science, 9, 656.

Smilanich, A. M., & Dyer, L. A. (2012). Effects of banana plantation pesticides on the immune response of lepidopteran larvae and their parasitoid natural enemies. Insects, 3(3), 616–628.

Smilanich, A. M., Dyer, L. A., & Gentry, G. L. (2009a). The insect immune response and other putative defenses as effective predictors of parasitism. Ecology, 90(6), 1434– 1440.

Smilanich, A. M., Dyer, L. A., Chambers, J. Q., & Bowers, M. D. (2009b). Immunological cost of chemical defence and the evolution of herbivore diet breadth. Ecology Letters, 12(7), 612–621.

Smilanich, A. M., Fincher, R. M., & Dyer, L. A. (2016). Does plant apparency matter? Thirty years of data provide limited support but reveal clear patterns of the effects of plant chemistry on herbivores. New Phytologist, 210(3), 1044–1057.

Smilanich, A. M., Vargas, J., Dyer, L. A., & Bowers, M. D. (2011). Effects of ingested secondary metabolites on the immune response of a polyphagous caterpillar Grammia incorrupta. Journal of Chemical Ecology, 37(3), 239–245.

148

Smith, C. A., Want, E. J., O'Maille, G., Abagyan, R. & Siuzdak, G. (2006) XCMS: processing mass spectrometry data for metabolite profiling using nonlinear peak alignment, matching, and identification. Analytical chemistry, 78, 779-787.

Sosa, J. M., Huber, D. E., Welk, B., & Fraser, H. L. (2014). Development and application of MIPARTM: a novel software package for two- and three-dimensional microstructural characterization. Integrating Materials and Manufacturing Innovation, 3, 123–140.

Spiteller, P. (2015). Chemical ecology of fungi. Natural Product Reports, 32(7), 971– 993.

Sternberg, E. D., Lef, T., Li, J., Lopez, C., Castillejo, F. De, Li, H., Hunter, M. D., & Roode, J. C. De. (2012). Food plant-derived disease tolerance and resistance in a natural butterfly-plant-parasite interactions. Evolution, 66(11), 3367–3377.

Stierle, A., Strobel, G., & Stierle, D. (1993). Taxol and Taxane Production by Taxomyces andreanae, an Endophytic Fungus of Pacific Yew. Science, 260(5105), 214–216.

Summerell, B. A. (2019). Resolving Fusarium: Current Status of the Genus . Annual Review of Phytopathology, 57(1), 323–339.

Symes, W. S., Edwards, D. P., Miettinen, J., Rheindt, F. E., & Carrasco, L. R. (2018). Combined impacts of deforestation and wildlife trade on tropical biodiversity are severely underestimated. Nature communications, 9(1), 4052.

Tedersoo, L., Bahram, M., Põlme, S., Kõljalg, U., Yorou, N. S., Wijesundera, R., ... & Smith, M. E. (2014). Global diversity and geography of soil fungi. Science, 346(6213), 1256688.

Thies, W., & Kalko, E. K. V. (2004). Phenology of neotropical pepper plants (Piperaceae) and their association with their main dispersers, two short-tailed fruit bats, Carollia perspicillata and C. castanea (Phyllostomidae). Oikos, 104(2), 362– 376.

149

Tilman, D., Reich, P. B., Knops, J., Wedin, D., Mielke, T., & Lehman, C. (2001). Diversity and productivity in a long-term grassland experiment. Science, 294(5543), 843–845.

Tilman, D., Wedin, D., & Knops, J. (1996). Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature, 379(6567), 718–720.

Traveset, A., Heleno, R., & Nogales, M. (2014). The ecology of seed dispersal. In Robert Gallagher (Ed.) Seeds: the ecology of regeneration in plant communities, (3rd ed., pp. 62–93). Boston, MA: CABI.

Traveset, A., & Verdú, M. (2002). A Meta-analysis of the Effect of Gut Treatment on Seed Germination. In D. J. Levey, W. R. Silva & M. Galetti (Eds.) Seed Dispersal and Frugivory: Ecology, evolution, and conservation, (pp.339–350). New York, NY: CABI Publishing.

Uckele, K. A., Jahner, J. P., Tepe, E. J., Richards, L. A., Dyer, L. A., Ochsenrider, K. M., … Parchman, T. L. (2020). Evidence for both phylogenetic conservatism and lability in the evolution of secondary chemistry in a tropical angiosperm radiation. BioRxiv, 11.30.404855.

Unterseher, M., Jumpponen, A., Öpik, M., Tedersoo, L., Moora, M., Dormann, C. F., & Schnittler, M. (2011). Species abundance distributions and richness estimations in fungal metagenomics - Lessons learned from community ecology. Molecular Ecology, 20(2), 275–285.

Vacher, C., Hampe, A., Port, A. J., Sauer, U., & Morris, C. E. (2016). The Phyllosphere: Microbial Jungle at the Plant – Climate Interface. Annual Review of Ecology and Systematics, 47, 1–24.

Van Bael, S., Estrada, C., & Arnold, A. E. (2017). Foliar endophyte communities and leaf traits in tropical trees. In J. Dighton & J. F. White (Eds.) The Fungal Community: Its Organization and Role in the Ecosystem The fungal community: its organization and role in the ecosystem, (4th ed., pp.79–94). Boca Raton, FL: CRC Press.

150

Valdivia, C., Marquez, N., Eriksson, J., Vilaseca, A., Muñoz, E., & Sterner, O. (2008). Bioactive alkenylphenols from Piper obliquum. Bioorganic and Medicinal Chemistry, 16(7), 4120–4126.

Vašutová, M., Mleczko, P., López-García, A., Maček, I., Boros, G., Ševčík, J., … Kjøller, R. (2019). Taxi drivers: the role of animals in transporting mycorrhizal fungi. Mycorrhiza, 29(5), 413–434.

Vellend, M. (2010). Conceptual synthesis in community ecology. The Quarterly Review of Biology, 85(2), 183–206.

Wagner, M. R., Lundberg, D. S., del Rio, T. G., Tringe, S. G., Dangl, J. L., & Mitchell- Olds, T. (2016). Host genotype and age shape the leaf and root microbiomes of a wild perennial plant. Nature Communications, 7, 12151.

Walters, W. A., Jin, Z., Youngblut, N., Wallace, J. G., Sutter, J., Zhang, W., … Ley, R. E. (2018). Large-scale replicated field study of maize rhizosphere identifies heritable microbes. Proceedings of the National Academy of Sciences of the United States of America, 115(28), 7368–7373.

White, T. J., Bruns, T., Lee, S., & Taylor, J. (1990). Amplification and Direct Sequencing of Fungal Ribosomal Rna Genes for Phylogenetics. PCR Protocols: A Guide to Methods and Applications, 18(1), 315–322.

White, J. F., Kingsley, K. L., Butterworth, S., Brindisi, L., Gatei, J. W., Elmore, M. T., ... & Kowalski, K. P. (2019). Seed-vectored microbes: Their roles in improving seedling fitness and competitor plant suppression. In S. K. Verma & J. F. White Jr (Eds.), Seed Endophytes (pp. 3–20). Cham, Switzerland: Springer.

Whitaker, B. K., Reynolds, H. L., & Clay, K. (2019). Foliar fungal endophyte communities are structured by environment but not host ecotype of Panicum virgatum (switchgrass). Ecology, 99(12), 2703–2711.

Whitehead, S. R., & Bowers, M. D. (2014). Chemical ecology of fruit defence: Synergistic and antagonistic interactions among amides from Piper. Functional Ecology, 28, 1094–1106.

151

Whitham, T. G., Bailey, J. K., Schweitzer, J. A., Shuster, S. M., Bangert, R. K., LeRoy, C. J., Lonsdorf, E. V., Allan, G. J., DiFazio, S. P., Potts, B. M., Fischer, D. G., Gehring, C. A., Lindroth, R. L., Marks, J. C., Hart, S. C., Wimp, G. M., & Wooley, S. C. (2006). A framework for community and ecosystem genetics: from genes to ecosystems. Nature Reviews Genetics, 7(7), 510–523.

Whitham, T. G., Allan, G. J., Cooper, H. F., & Shuster, S. M. (2020). Intraspecific Genetic Variation and Species Interactions Contribute to Community Evolution. Annual Review of Ecology, Evolution, and Systematics, 51(1), 587–612.

Wiggins, N. L., Forrister, D. L., Endara, M.-J., Coley, P. D., & Kursar, T. A. (2016). Quantitative and qualitative shifts in defensive metabolites define chemical defense investment during leaf development in Inga, a genus of tropical trees. Ecology and Evolution, 6(2), 478–492.

Yang, S. X., Sun, Q. Y., Yang, F. M., Hu, G. W., Luo, J. F., Wang, Y. H., & Long, C. L. (2013). Sarmentosumols A to F, new mono- and dimeric alkenylphenols from Piper sarmentosum. Planta Medica, 79(8), 693–696.

Zambell, C. B., & White, J. F. (2017). Community Assembly of Phyllosphere Endophytes. In J. Dighton & J. F. White (Eds.) The Fungal Community: Its Organization and Role in the Ecosystem (4th ed., pp. 95–108). Boca Raton, FL: CRC Press.

Zangerl, A. R., & Rutledge, C. E. (1996). The Probability of Attack and Patterns of Constitutive and Induced Defense: A Test of Optimal Defense Theory. The American Naturalist, 147(4), 599–608.

Zangerl, A. R., Stanley, M. C., & Berenbaum, M. R. (2008). Selection for chemical trait remixing in an invasive weed after reassociation with a coevolved specialist. Proceedings of the National Academy of Sciences of the United States of America, 105(12), 4547–4552.

Zgoda, J. R., & Porter, J. R. (2001). A Convenient Microdilution Method for Screening Natural Products Against Bacteria and Fungi. Pharmaceutical Biology, 39(3), 221– 225.

152

Zimmer, C. & Emlen, D. J. (2016). Evolution: Making sense of life. 2nd ed. W. H. Freeman and company, New York, NY.

Zimmerman, N. B., & Vitousek, P. M. (2012). Fungal endophyte communities reflect environmental structuring across a Hawaiian landscape. Proceedings of the National Academy of Sciences, 109(32), 13022–13027.

153

6 Appendix

Supplementary Information 1 - Species identification methods:

(Chapter Two) In order to identify our lepidopterans, we pin and curate adults using standard techniques (Forister et al. 2015 – supplemental material). For parasitoids, we prepare them according to the suggestions of the taxonomists that identified them, which is typically to either pin or place specimens in vials of alcohol, and then pinned specimens are properly deposited at museums with vouchers. Once vouchers have been collected for identification and description purposes, a series of additional specimens are preserved in 95% ethanol for future molecular systematic work. Each specimen is labeled with the following information: country, province, collection site, site elevation, longitude and latitude, month and year of collection, and affiliation of collecting site. The voucher code which links the specimen to its event- based record in the database is added and another label is added when appropriate, identifying the specimen to family, genus and species. Similarly, specimens in liquid media are given labels which include all the above standard information. Every collection event, which is the act of finding a caterpillar, receives a unique voucher code. Every 1-3 years, specimens that are fully identified and curated are deposited to the appropriate museums in Costa Rica, Ecuador, Brazil and the United States. Voucher specimens examined by the taxonomists, including newly established types and undescribed species are housed in collections of the taxonomists’ preference to facilitate further description and systematic studies of the material, unless our permits specify otherwise. All Lepidoptera reared in this chapter are provisionally identified by the rearing and curation crews at the sites and then via a network of collaborating taxonomists (see www.caterpillars.org).

154

Supplementary Table 1: (Chapter Two) Bootstrapping was used to impute missing data to account for differences in sampling effort. Data was bootstrapped to meet the sample size of the largest plant– caterpillar species pairs from our immune assays. We bootstrapped our data using the Hmisc v.4.0–3 (Harrell, 2015) package in R v3.4.2 (R core Team 2017) (Harrell, 2015). We specified a non–linear regression type model with one imputation and 3 knots for all our datasets which allowed for extrapolation from our data. To determine the appropriate number of knots for our data, we ran a series of imputations which varied in the number of knots and chose the number based on the lowest mean and absolute error (Harrell, 2015). After the imputation for each dataset, a R2 value was generated to predict our original measured data from the imputation as a measure of error. High R2 values indicate strong predictions. Our R2 values for the imputations in all 3 datasets were between 0.37 and 0.59.

155

Site Eois spp. Piper spp. Bootstrapping n bootstrapped total n Costa Rica Eois nympha Piper biseriatum 19 28 Piper cenocladum 0 28 Eois apyraria Piper cenocladum 27 28 Piper imperiale 21 28 Eois russearia Piper sancti-felicis 16 28 Eois mexicaria Piper umbricola 15 28 Total 98 168 Ecuador Six black two pink spots Piper baezanum 25 27 Piper kelleii 11 27 Piper lancifolium 26 27 Lime slime Piper baezanum 26 27 Piper kelleii 20 27 Two black spots Piper kelleii 0 27 Piper lancifolium 26 27 Eois viridiflava Dognin Piper baezanum 26 27 Piper lancifolium 7 27 Pink spots funk Piper kelleii 24 27 Piper lancifolium 26 27 Eight black blur Piper baezanum 26 27 Eois beebei Fletcher Piper kelleii 26 27 Eois ignefumata Dognin Piper kelleii 26 27 Total 295 378

156

Supplementary Table 2: (Chapter Two) Bootstrapping was used to impute missing data to account for differences in sampling effort. Data was bootstrapped to meet the sample size of the largest plant– caterpillar species pairs from our immune assays. Here this is a sample size of 19.

Site Piper spp. Bootstrapping n bootstrapped total n Costa Rica Piper arboreum 16 19 Piper cenocladum 18 19 Piper colonense 6 19 Piper garagaranum 18 19 Piper imperiale 13 19 Piper multiplinervium 0 19 Piper pseudobumbratum 18 19 Piper reticulatum 1 19 Piper trigonum 17 19 Piper umbricola 18 19 Total 125 190

157

Supplementary Table 3: (Chapter Two) SEM results of bootstrapped data from Costa Rica Eois data. We generated 7 a priori hypotheses to explain the relationships between our variables based on previous work on Piper. Our hypotheses tested for: I) ‘Herbivore mediation hypothesis’, II) ‘Diet breadth regulation hypothesis’, III) ‘Phytochemical diversity regulation hypothesis’, IV) ‘Combination hypothesis’, V) ‘Interaction hypothesis’, VI) ‘Simple phytochemical diversity hypothesis’, and VII) ‘Immunity does not predict parasitism hypothesis’. Asterisks next to path coefficients indicate statistically significant paths (P < 0.05). Quadrus cerealis data for VI) ‘Simple phytochemical diversity hypothesis’ did not fit the model. Eois data from Ecuador did not fit any of our models.

158

159

References:

Harrell, F. E. (2015). Package ‘Hmisc’. Retrieved on February 5, 2018.

160

Group 1 Mean Data Group 1 w. Post. Pred.

N1 = 82

mean = 3.12 5

)

(

0 . 95% HDI 0

2.06 4.23 0

. 0 1 2 3 4 5 6 0 10 20 30 40 50 m1 y Group 2 Mean Data Group 2 w. Post. Pred.

N2 = 82

mean = 2.39 5

)

(

0 . 95% HDI 0

1.15 3.73 0

. 0 1 2 3 4 5 6 0 10 20 30 40 50 m2 y Group 1 Std. Dev. Difference of Means mode = 2.97 mean = 0.731

16.1% < 0 < 83.9% 6% in ROPE

95% HDI 95% HDI 2.07 4.14 −0.718 2.22

1 2 3 4 5 6 7 8 −1 0 1 2 s1 m1 - m2 Group 2 Std. Dev. Difference of Std. Dev.s mode = 3.21 mode = −0.26

95% HDI 95% HDI 2.03 4.96 −1.83 1.1

1 2 3 4 5 6 7 8 −2 −1 0 1 s2 s1 - s2 Normality Effect Size mode = 0.148 mode = 0.203

16.1% < 0 < 83.9%

95% HDI 95% HDI −0.02 0.35 −0.245 0.699

0.0 0.1 0.2 0.3 0.4 −0.4 −0.2 0.0 0.2 0.4 0.6 0.8 log10(n) (m - m ) (s2 + s2) 2 1 2 1 2 161

162

generalist_herbivory

1.0

) 0.5 %

5 Groups

. 3

1 offspring

(

2 parent

P specialist_herbivory

A C 0.0 parent

phytochemical_diversity −0.5

−1.0 −0.5 0.0 0.5 CAP1 (73.37%)

Supplementary Figure 2: (Chapter Three) A partial distanced based redundancy analysis evaluating ontogeny, phytochemical diversity and herbivory on fungal endophyte community incidence data (Monte

Carlo permutation test: F(4,90) = 4.15, P = 0.001). Predictor variables explained 16% of the variance in the data. Canonical analysis of principal coordinates axis 1 (CAP1) explained 73% of the constrained variance in the (Jaccard) dissimilarity matrix, while canonical analysis of principal coordinates axis 2 (CAP2) explained 14%. Black data points represent fungal endophyte taxa and larger coloured data points represent site scores, which are the weighted sum of species. Ellipses correspond to a 95% confidence value by ontogeny. Arrow lengths do not represent the effect size of the variables or treatment levels but their directionality and loading values (Gotelli and Ellison 2013).

163

developed_area 1.0

0.5 Land use history abandoned plantation

) current plantation

% phytochemical_diversity

4 developed area

. 0

2 current_plantation open pasture

( 0.0 shaded_pasture

2 logged_forest secondary forests 10−21 years

P A

C select logged forests

open_pasture shaded pasture secondary_forest successional plots −0.5

−1.0

−0.5 0.0 0.5 1.0 CAP1 (43.22%)

164

model1

X6 X6 X15 X15 X11 X11 X12 X4 X77 X12 X24 X417 X4 X77 X417 X7 X23 X23 X504 X24 X214 X452 X7 X214 X2 X2 X499 X109 X82 X8 X209 X46 X452 X82 X513 X399 X8 X147 X63 X101 X498 X499 X154 X209 X46 X140 X147 X138 X138 X97 X109 X136 X140 X498 X101 X111 X544 X81 X136 X117

3.0 3.5 4.0 4.5 5.0 5.5 0e+00 4e−07 8e−07 %IncMSE IncNodePurity

165

accuracy decreases if that compound was left out of the model and stands for % increase in mean quared error. IncNode Purity indicates the increase in how well each node was split while building the random forest at each compound, where 0 represents a pure node, meaning a perfect split and is measured by the change in sum of squares. The overall model explained 68% of the variance in the data and had a mean squared residual of 8.9e-08. When the model was evaluated against the test set, the explained variance decreased to 40% and the mean squared error was 1.60e-07. While the overall model is meaningful and explains a large proportion of the data, the importance of any one compound is low. The analysis was derived from the R package and function randomForest v4.6-14 (Liaw and Wiener 2002).

166

A B C D E F 1 Lab ID Accession # Taxa Score (bits) E-value % Identity 2 18 MT093632 Paraphaeosphaeria sp. 1092 0 99.83 3 91 MT093633 Microdochium lycopodinum 974 0 96.61 4 92 MT093634 Fusarium sp. 1077 0 99.17 5 93 MT093635 Phomopsis sp. 1055 0 97.57 6 94 MT093636 Microdochium lycopodinum 979 0 97.24 7 1 MT093637 Microdochium lycopodinum 972 0 97.06 8 10 MT093638 Fusarium sp. 1000 0 100 9 11 MT093639 Microdochium lycopodinum 983 0 97.73 10 12 MT093640 Microdochium lycopodinum 961 0 97.35 11 13 MT093641 Fusarium sp. 1000 0 100 12 14 MT093642 Microdochium lycopodinum 963 0 97.42 13 15 MT093643 Fusarium solani 1061 0 99.66 14 17 MT093644 Fusarium solani 1064 0 100 15 19 MT093645 Fusarium sp. 1064 0 99.66 16 2 MT093646 Coniella sp. 1085 0 99.5 17 20 MT093647 Microdochium lycopodinum 963 0 97.52 18 21 MT093648 Microdochium lycopodinum 966 0 97.86 19 22 MT093649 Fusarium sp. 1068 0 99.83 20 23 MT093650 Fusarium sp. 1040 0 99.82 21 25 MT093651 Microdochium lycopodinum 963 0 97.52 22 26 MT093652 Fusarium sp. 1070 0 100 23 27 MT093653 Fusarium sp. 1026 0 100 24 28 MT093654 Microdochium lycopodinum 990 0 97.58 25 29 MT093655 Microdochium lycopodinum 1003 0 97.61 26 3 MT093656 Microdochium lycopodinum 992 0 97.53 27 30 MT093657 Neopestalotiopsis clavispora 1053 0 99.65 28 31 MT093658 Pestalotiopsis sydowiana 1042 0 100 29 33 MT093659 Mucor sp. 1110 0 99.83 30 34 MT093660 Agaricomycotina 562 4.00E-158 80

167

A B C D E F 31 35 MT093661 Neopestalotiopsis clavispora 1050 0 99.83 32 36 MT093662 Lasiodiplodia theobromae 1064 0 99.83 33 37 MT093663 Chaetothyriales 1031 0 98.3 34 38 MT093664 Pseudopestalotiopsis theae 1083 0 100 35 39 MT093665 Hypocreales 994 0 98.44 36 4 MT093666 Microdochium lycopodinum 974 0 97.38 37 41 MT093667 Cladosporium sp. 1041 0 99.32 38 42 MT093668 Fusarium sp. 1070 0 100 39 43 MT093669 Microdochium lycopodinum 977 0 97.23 40 45 MT093670 Microdochium lycopodinum 989 0 97.9 41 46 MT093671 Microdochium lycopodinum 979 0 97.4 42 5 MT093672 Microdochium lycopodinum 949 0 97.51 43 6 MT093673 Microdochium lycopodinum 959 0 97.67 44 7 MT093674 Microdochium lycopodinum 949 0 97.51 45 9 MT093675 Fusarium sp. 1059 0 100

168

G 1 Closest UNITE SH (<0.5% similarity) 2 TH008125: Didymosphaeriaceae | SH2177394.08FU 3 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 4 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 5 TH025193: Phomopsis (Sacc.) Bubák | SH2209054.08FU 6 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 7 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 8 TH020880: Fusarium concolor Reinking | SH2456062.08FU 9 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 10 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 11 TH020880: Fusarium concolor Reinking | SH2456062.08FU 12 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 13 TH020880: Fusarium solani (Mart.) Sacc. | SH2947023.08FU 14 TH006226: Hypocreales | SH2228351.08FU 15 TH020880: Fusarium concentricum Nirenberg & O'Donnell | SH2456034.08FU 16 TH024622: Coniella Höhn. | SH2255849.08FU 17 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 18 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 19 TH020880: Fusarium concentricum Nirenberg & O'Donnell | SH2456034.08FU 20 TH020880: Fusarium proliferatum (Matsush.) Nirenberg ex Gerlach & Nirenberg | SH2456081.08FU 21 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 22 TH020880: Fusarium proliferatum (Matsush.) Nirenberg ex Gerlach & Nirenberg | SH2456081.08FU 23 TH006760: Nectriaceae | SH2456050.08FU, TH012113: Gibberella circinata Nirenberg & O'Donnell ex Britz, T.A. Cout., M.J. Wingf. & Marasas | SH2456044.08FU 24 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 25 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 26 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 27 TH019623: Neopestalotiopsis rosae Maharachch., K.D. Hyde & Crous | SH2251711.08FU 28 TH019623: Neopestalotiopsis foedans (Sacc. & Ellis) Maharachch., K.D. Hyde & Crous | SH2251708.08FU 29 TH009719: Mucor nidicola Madden, Stchigel, Guarro, Deanna A. Sutton & Starks | SH2365264.08FU 30 TH005300: Agaricomycotina | SH2455436.08FU

169

G 31 TH019623: Neopestalotiopsis rosae Maharachch., K.D. Hyde & Crous | SH2251711.08FU 32 TH015964: Lasiodiplodia theobromae (Pat.) Griffon & Maubl. | SH2131006.08FU 33 TH005271: Pezizomycotina | SH2574952.08FU 34 TH012234: Pseudopestalotiopsis simitheae (Y. Song, N. Tangthirasunun, K.D. Hyde & Y. Wang) Maharachchikumbura & K.D. Hyde | SH2251713.08FU 35 TH006226: Hypocreales | SH2282772.08FU 36 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 37 TH008870: Cladosporium halotolerans Zalar, de Hoog & Gunde-Cim. | SH2320205.08FU 38 TH020880: Fusarium proliferatum (Matsush.) Nirenberg ex Gerlach & Nirenberg | SH2456081.08FU 39 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 40 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 41 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 42 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 43 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 44 TH017257: Microdochium lycopodinum (Jaklitsch, Siepe & Voglmayr) M. Hern.-Restr. & Crous | SH2261955.08FU 45 TH020880: Fusarium proliferatum (Matsush.) Nirenberg ex Gerlach & Nirenberg | SH2456081.08FU

170

H I J K L M 1 Notes 2 3 type material: culture from holotype of Microdochium lycopodinum 4 type material: culture from holotype of Microdochium lycopodinum 5 6 type material: culture from holotype of Microdochium lycopodinum 7 type material: culture from holotype of Microdochium lycopodinum 8 9 type material: culture from holotype of Microdochium lycopodinum 10 type material: culture from holotype of Microdochium lycopodinum 11 12 type material: culture from holotype of Microdochium lycopodinum 13 14 15 16 17 type material: culture from holotype of Microdochium lycopodinum 18 type material: culture from holotype of Microdochium lycopodinum 19 20 21 type material: culture from holotype of Microdochium lycopodinum 22 23 24 25 type material: culture from holotype of Microdochium lycopodinum 26 type material: culture from holotype of Microdochium lycopodinum 27 28 29 30

171

H I J K L M 31 32 33 34 35 36 type material: culture from holotype of Microdochium lycopodinum 37 38 39 type material: culture from holotype of Microdochium lycopodinum 40 type material: culture from holotype of Microdochium lycopodinum 41 type material: culture from holotype of Microdochium lycopodinum 42 type material: culture from holotype of Microdochium lycopodinum 43 type material: culture from holotype of Microdochium lycopodinum 44 type material: culture from holotype of Microdochium lycopodinum 45

172