Mutualism Stability and Gall Induction in the Fig and Fig Wasp Interaction

Item Type text; Electronic Dissertation

Authors Martinson, Ellen O'Hara

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 28/09/2021 01:14:56

Link to Item http://hdl.handle.net/10150/265556 MUTUALISM STABILITY AND GALL INDUCTION IN THE FIG AND FIG WASP INTERACTION

Ellen O. Martinson

______

A Dissertation Submitted to the Faculty of the

ECOLOGY AND EVOLUTIONARY BIOLOGY

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2012

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Ellen O. Martinson entitled MUTUALISM STABILITY AND GALL INDUCTION IN THE FIG AND FIG WASP INTERACTION and recommend that it be accepted as fulfilling the dissertation requirement for the

Degree of Doctor of Philosophy

______Date: 11/02/12 A. Elizabeth Arnold

______Date: 11/02/12 Jeremiah D. Hackett

______Date: 11/02/12 Carlos A. Machado

______Date: 11/02/12 Rob H. Robichaux

______Date: 11/02/12 Noah K. Whiteman

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: 11/02/12 Dissertation Director: A. Elizabeth Arnold

______Date: 11/02/12 Dissertation Director: Jeremiah D. Hackett 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED: ______Ellen O. Martinson

ACKNOWLEDGEMENTS

I am eternally grateful to my co-advisors A. Elizabeth (Betsy) Arnold and Jeremiah Hackett. Thank you Betsy for letting me join your lab. Your mentorship has left me a better scientist and writer and has taught me the importance of a positive attitude. Thank you Jeremiah for opening your lab to me and always being available to work on a problem. I would like to thank my committee members, Carlos Machado, Noah Whiteman, and Rob Robichaux for their support and feedback. I am especially grateful to Carlos Machado for showing me the fig and fig wasp mutualism first hand and for his support and early advisement. I wish to thank current and former University of Arizona faculty, including David Galbraith, Nancy Moran, Howard Ochman, Michael Nachman, Rich Jorgensen, David Maddison, Robert Steidl, and Peter Reinthal for their advice and excellent instruction. I thank my co-author Allen Herre for his valuable feedback on manuscripts and introducing me to Barro Colorado Island. I am grateful to the Smithsonian Tropical Research Institute for logistical support and the government of Panama for permission to carry out this research. I thank Aldaberto Gomez for teaching me to navigate the Panama Canal and to my fellow researchers and staff on Barro Colorado Island for their feedback and support, especially Egbert Leigh Jr., Oris Acevedo, Robert Horan, Nadia Sitas, Matt McRoy, and Carlos Aguilar. I am grateful for my labmates in the Arnold, Hackett, and Machado labs, including Jen Wisecaver, Will Driscoll, Jana U’Ren, Mary Jane Epps, K. Mali Gunatilaka, Carlos Flores, Wendy Marussich, and Anna Himler for their camaraderie and technical support in the lab. Thank you to the Ecology and Evolutionary Biology staff for their assistance, especially Liz Oxford, Jean Mason, Carol Freeman, Beth Sanchez, and Brian Morton. And I am grateful to the UA IGERT program in genomics, NSF GRFP, and the Smithsonian Institute for funding during my graduate career. For my friends, thank you for your encouragement, friendship, and awesome Halloween costumes. I wish to especially thank Kevin Vogel, Gaelen Burke, Matt Heron, Joel Wertheim, Betsy Wertheim, Joe Deas, Lindsey Sloat, Bryan Helm, Steve Foldi, Erin Kelleher Meisel, and Ming-Min Lee. I am grateful to Bryan Bishop and Greg Hoch for guiding me towards a career in science. I would like to thank my parents Robert and Barbara Suurmeyer, my siblings Nathan, Laura, and Jill, and my in-laws, Joanne and Wynn Martinson, for their incredible support and encouragement throughout my graduate career. Finally, my husband, Vince Martinson, thank you for all of our support, guidance, friendship, and love.

DEDICATION

To Vince and my family for their endless love and support

TABLE OF CONTENTS

ABSTRACT...... 7

INTRODUCTION...... 9

PRESENT STUDY...... 23

REFERENCES...... 27

APPENDIX A: CULTURE-FREE SURVEY REVEALS DIVERSE AND

DISTINCTIVE FUNGAL COMMUNITIES ASSOCIATED WITH DEVELOPING

FIGS (FICUS SPP.) IN PANAMA………...... 33

APPENDIX B: RATIOS OF POISON SAC AREA:EGGS SUGGEST A SIMILAR

IMPORTANCE OF THE POISON SAC IN OVIPOSITION BY POLLINATING AND

NON-POLLINATING FIG WASPS...... 55

APPENDIX C: METATRANSCRIPTOME ANALYSIS OF FIG FLOWERS REVEALS

POSSIBLE MECHANISMS FOR MUTUALISM STABILITY AND GALL

INDUCTION...... 75

ABSTRACT

The interaction between figs (Ficus spp.) and their pollinating wasps (fig wasps;

Chalcidoidea, Hymenoptera) is a classic example of an ancient and apparently stable mutualism. A striking property of this mutualism is that fig wasps consistently oviposit in the inner flowers of the fig syconium (gall flowers, which develop into galls that house developing larvae), but typically do not use the outer ring of flowers (seed flowers, which are pollinated and develop into seeds). This dissertation explores the potential differences between gall and seed flowers that might influence oviposition choices, and the unknown mechanisms underlying gall formation. To identify the microbial community that could influence oviposition choice, I identified fungi in both flower types across six species of

Ficus. I found that whereas fungal communities differed significantly as a function of developmental stages of syconia and lineages of fig trees, communities did not differ significantly between receptive gall and seed flowers. Because secretions from the poison sac that are deposited at oviposition are thought to be important in gall formation by both pollinating fig wasps and non-pollinating, parasitic wasps, I examined poison sac morphology in diverse galling wasps from several species of Ficus in lowland Panama. I found that the size of the poison sac was positively associated with egg number across pollinating and non-pollinating fig wasps. Finally to determine difference in defense and metabolism between gall and seed flowers, and to identify genes involved in galling, I compared gene expression profiles of fig flowers at the time of oviposition choice and early gall development. I found a prominence of flavonoids and defensive genes in both 8 pollinated and receptive gall flowers of Ficus obtusifolia, and revealed detectable differences between gall flowers and seed flowers before oviposition. Several highly expressed genes were also identified that have implications for the mechanism of gall initiation. This dissertation explores previously unstudied aspects of the fig and fig wasp mutualism and provides important molecular tools for future study of this iconic and ecologically important association.

INTRODUCTION

1.1 Literature Review

Current evidence suggests that every living organism is involved in at least one mutualistic interaction (Bronstein et al. 2006; Leigh et al. 1995). Mutualism has been classically defined as an interaction between individuals in different species that is beneficial to both (Boucher et al. 1982). This definition encompasses interactions that range from diffuse mutualisms between several species of pollinators and plants to obligate and highly intimate interactions between endosymbionts and their hosts

(Boucher et al. 1982; Herre et al. 1999). More recently the definition of mutualism has been reframed as a reciprocally exploitative interaction that provides net benefits to both partner species (Herre et al. 1999; Kjellberg et al. 2001). This definition more clearly reflects the issue of maintenance of stability within a mutualism, which has long been a challenge in mutualism biology (Axelrod et al. 1981; Bull et al. 1991; Sachs et al. 2004).

However, mechanisms that maintain mutualisms largely remain a mystery (Herre et al.

2008).

In mutualistic interactions, both partners incur a fitness cost to provide a benefit to another species (Axelrod et al. 1981; Bull et al. 1991; Sachs et al. 2004). A precarious balance results because selection shapes lineages to maximize their individual fitness

(Holland et al. 2004; Neuhauser et al. 2004), thus creating an inherent tension. In vertically transmitted mutualisms (e.g., between an obligate bacterial endosymbiont and its host) these tensions are minimized by the relatively direct alignment of the fitness of the symbiont and the fitness of the host (Herre et al. 1999). However, in mutualisms with 10 horizontal transmission (e.g., pollinator and plant mutualisms), the fitness of each partner is not as closely aligned, suggesting that the cost/benefit ratio may shift over time toward parasitism and thus towards a loss of the mutualistic interaction (Sachs et al. 2006). In this dissertation I use the fig and fig wasp mutualism to explore factors influencing the stability of a horizontally transmitted mutualism and related aspects of that mutualism’s mechanistic interactions and complexity.

Fig and fig wasp mutualism

Over the last 80 million years, fig trees (Ficus, Moraceae) and fig wasps (Chalcidoidea,

Hymenoptera) have formed one of the most interdependent insect-plant mutualisms known (Machado et al. 2001; Rønsted 2005). Currently over 700 species of figs are known, each supporting a unique community of fig wasps (Berg 1989; Molbo et al. 2003;

Wiebes 1982). In each case, a female fig wasp enters a receptive fig through the ostiole, losing her wings and part of her antennae in the process. Once the wasp has entered the enclosed inflorescence, or syconium, she is surrounded by hundreds of flowers. When a flower receives pollen, an egg, and a maternal secretion from the poison sac, a gall is formed that will provide nutrition to and protect the wasp larva, but will not result in seed production. Flowers only receiving pollen develop into a fig seed. Some species of wasps disperse pollen actively, storing pollen in specialized cuticle pockets and fertilizing flowers individually; others pollinate passively, fertilizing flowers as pollen brushes off the wasp’s body (Jousselin et al. 2003). Passive pollination is considered the ancestral condition in Ficus, with ca. 60 million years separating the New World actively- and 11 passively pollinated clades of Ficus (Jousselin et al. 2003). Following development, males emerge first and mate with female wasps still within their galls; then males chew an exit tunnel for the females to emerge from the fig to seek out a new receptive fig tree.

Stability in the fig and fig wasp mutualism

The fig needs both seeds (female fitness) and female wasps to carry pollen (male fitness).

However, wasps do not benefit directly from seed production in the short term, and wasps prevent seed development in individual flowers by ovipositing. Theory suggests that this conflict of interest could lead to the wasp using most or all of the fig’s resources for its offspring, leaving little resources for the fig’s seeds (West et al. 1994). Yet wasps consistently use only about 50 percent of the fig’s flowers to produce wasps, pollinating the rest to develop into fig seeds (Herre et al. 1997). If multiple wasps enter the same fig, wasps may only realize 25% of their reproductive potential to maintain this ratio of wasps to seeds (Herre 1989). This act of apparent self-control goes against the wasp’s objective to produce more offspring, indicating that stabilizing mechanisms may be used by the fig tree to control the fig wasp-to-seed ratio.

A stabilizing mechanism has been identified in another classic seed- predator/pollinator mutualism: that of the yucca (Yucca filamentosa) and yucca moth

(Tegeticula yuccasella) (Cook et al. 2004; Fleming et al. 1998; Pellmyr 1997). In this mutualism, yucca moths pollinate and oviposit within the yucca flower, allowing their young to feed on a subset of the developing seeds. Yuccas can abort flowers that are overburdened with moth larvae (Pellmyr et al. 1994). Whereas some fig trees can abort 12 syconia in which no pollination has occurred (Jandér et al. 2010), no evidence suggests that fig trees preferentially abort syconia based on the gall-to-seed ratio (Galil et al. 1971;

Jousselin et al. 2003; Jousselin et al. 2001).

Hypotheses for stabilizing mechanisms in the fig- and fig-wasp mutualism

Attempts to identify the stabilizing mechanism behind the fig and fig wasp mutualism have produced four main hypotheses. The first states that the foundress wasp simply does not have enough eggs to utilize all of the flowers in a syconium (Nefdt et al. 1996). This hypothesis, however, does not take into account that many fig species host multiple foundress wasps, which collectively have more than enough eggs to oviposit in every flower (Anstett et al. 1996). Moreover, figs species that consistently have higher numbers of foundress wasps do not necessarily have more galls than seeds per syconium (Herre

1989).

The three remaining hypotheses draw from the large variation in the style length of the flowers lining the inside of the syconium. Fig wasps preferentially oviposit in the short-styled flowers (hereafter, gall flowers) that sit closer to the inner cavity of the fig as opposed to long-styled flowers (hereafter, seed flowers) that sit closer to the fig wall

(Herre 1989). One hypothesis suggested that the fig wasp’s ovipositor is too short to oviposit in all the flowers, limiting the number of usable flowers to the layer closest to the open cavity of the fig (Ganeshaiah et al. 1995). However, this hypothesis cannot be generalized to all fig species because many species of fig wasps can reach the majority of flowers (some up to 99%; Bronstein 1988). 13

The next hypothesis states that externally-ovipositing parasitic wasps can better attack pollinating wasp larvae that develop in seed flowers, therefore pressuring the pollinating wasps to only the gall flowers that constitute “enemy-free space” (Dunn et al. 2008). This hypothesis also cannot be generalized to fit all fig species, because not all fig species have wasps that parasitize pollinators, and many species of these parasitic wasps have ovipositors that can reach the gall flowers (West et al. 1994).

The final hypothesis proposes that some flowers block the ovipositing and/or development of fig wasps through biochemical or developmental mechanisms that differ between seed- and gall flowers. This has been referred to as the ‘unbeatable seed’ hypothesis (West et al. 1994). Evidence for the unbeatable seed hypothesis comes from the fact that both pollinating and non-pollinating wasps (see below) only gall flowers that are closer to the inner cavity of the fruit (West et al. 1994). This leaves the row of flowers closest to the fig wall to develop into seeds, even though that outer layer of flowers is more accessible to the non-pollinating wasps that oviposit through the fig wall. This observation led West et al. (1994) to suggest that seed flowers differ fundamentally from gall flowers in a manner that makes them less appealing to, or less functional for, oviposition and/or larval development.

However, West et al. (1994) did not identify clear differences between gall and seed flowers that might explain wasps’ preferences for oviposition. Subsequent studies have not provided strong evidence for important differences between these flower types, although some physical and developmental differences are evident. Whereas gall flowers and seed flowers have the same structural characteristics in their ovary and ovule, they 14 differ in style length and shape. The stigmas of gall flowers form two equal lobes on top of the style, whereas the stigmas of seed flowers form two unequal branches, with one branch being far longer and thinner than the other (Verkerke 1986). However, these traits reflect style morphology, which has more of a role in pollination than gall development: foundress wasps bypass the style and lay their eggs in the tissue between the integument and nucellus of the ovule (Verkerke 1986). Thus the reasons for which particular flowers are not used by fig wasps remain unknown and a focus of this dissertation.

Complexity in the fig- and fig-wasp mutualism

The study of mutualisms traditionally has focused on the interactions between two organisms. However, other associated organisms can alter the balance of costs and benefits to each partner (e.g., Moran 2007; Scott et al. 2008). Ecologists increasingly appreciate that mutualisms should be interpreted in a multipartite context (e.g., Agrawal et al. 2007; Herre 1993; West et al. 1994), which often reveals previously unexplored components of even the most classic two-partner associations (e.g., Currie et al. 2003a;

Hoffman et al. 2010; Oliver et al. 2009; Van Bael et al. 2009). For example, the classic mutualism between leaf-cutting ants and their fungal partner now is described as a tripartite mutualism involving a bacterium (actinomycete) that is carried by the ants to protect fungal gardens from parasites (Currie et al. 2003b).

In addition to pollinating wasps, a diverse community of organisms is associated with the microcosm of the fig syconium (e.g., non-pollinating wasps, nematodes, mites)

(Herre 1995; Wang et al. 2009; West et al. 1994). Studies focused on these organisms 15 have provided insight into the study of coevolution, plant-parasite interactions, and the evolution of virulence (e.g., Herre 1993; Marussich et al. 2007; West et al. 1994).

Perhaps the most well known of these interactions is the competition between pollinating and parasitic, non-pollinating fig wasps. Non-pollinating wasps parasitize the fig and fig wasp mutualism by ovipositing through the fig wall from the outside, thereby avoiding pollination (Boucek 1993).

There are three distinct ecological groups of non-pollinating wasps. The physically largest non-pollinating wasps lay their eggs in the wall of the syconium, initiating a very large, amorphous gall that forms into the lumen of the fig (e.g.,

Aepocerus and Idarnes ‘incerta’). Unlike other fig-associated wasps, these non- pollinators seem to prevent unpollinated fruits from being aborted by the fig tree, such that they are not dependent of the presence of pollinating fig wasps (West et al. 1996).

The second group of non-pollinators (e.g., Physothorax) are the least common of the three main types of non-pollinating wasps (Marussich et al. 2007). They are true parasitoids of the large gallers and do not use fig tissue in their development (West et al.

1994; West et al. 1996).

The third group of parasitic wasps in the fig-fig wasp system is the most common ecological type and has drawn the most attention because they utilize the same flower tissue as pollinating wasps to produce galls that appear identical (Pereira et al. 2005).

Examples include Idarnes or Critogaster spp., depending on the host fig species.

Non-pollinating wasps in this third ecological group (hereafter, non-pollinating wasps) therefore cost both partners in the mutualism not only by initiating gall growth without 16 providing pollination services, but also directly competing with the pollinating wasps for potential gall flowers (West et al. 1996). Although fig trees appear to sanction pollinators by aborting unpollinated fruit (Jander et al. 2010), sanctions against non-pollinating wasps have not yet been identified (West et al. 1994). At one time, pollinating and non- pollinating wasps both were assigned to the family Agaonidae; however, this grouping was found to be paraphyletic and the six subfamilies of non-pollinator wasps were reassigned to families within the superfamily Chalcidoidea (Rasplus et al. 1998).

Communities of non-pollinating wasps and their influence on the fig and fig-wasp mutualism have been studied (Marussich et al. 2007; West et al. 1994; West et al. 1996), as have mites and other metazoan associates of developing figs (Flock et al. 1955; Herre

1995; Wang et al. 2009). However, the microbial communities associated with figs or fig wasps have not yet been explored thoroughly. Microbial communities can have large impacts on plant-insect interactions; for example, endophytic fungi can influence an insect’s choice of host plant (Cardoza et al. 2003; Jallow et al. 2008; Omacini et al. 2001;

Vidal 1996) and bacteria allow some insects to survive on nutrient-poor diets (e.g.,

Moran 2007). Previous culture-based studies have detected a few yeasts and bacteria in cultivated figs, suggesting that the syconium is effectively sterile before the fig wasp enters (e.g., Miller et al. 1962; Mrak et al. 1942). Several surveys of fungal communities have screened leaves and rotting fruits of fig trees (Doster et al. 1996; Suryanarayanan et al. 2001; Wang et al. 2008), yet the fungal communities within the developing syconia of non-domesticated Ficus spp. have not been studied previously. Therefore potential 17 influences of microbes on the interaction between fig wasps and fig trees were not known, providing a second major focus for this work.

Mechanistic interactions

The fig and fig wasp interaction is a distinctive mutualism that involves creation of a gall in a flower (Shorthouse et al. 1992). Unlike many plant-insect relationships in which the insect directly consumes the plant, the fig wasp triggers growth in the plant tissue to form a gall that is consumed by the wasp’s larva. The wasp’s signal to initiate gall growth and the fig’s response to that signal is one of the fundamental mechanistic components of this mutualism; however, the insect’s signal to trigger gall formation is unknown.

A plant gall is an abnormal growth of plant tissue caused by another organism

(Mani 1992). Gall formation is an intimate process in which a galling organism, such as an insect, extends its phenotype to a plant to create a unique structure specific to its needs in its host plant’s tissues (Dawkins 1999; Stern 1995). The capacity to induce galls is widespread across the tree of life, with galling organisms occurring among viruses, bacteria, fungi, nematodes, and arthropods (Mani 1992). To date, gall induction by

>13,000 insect species has been recognized (Shorthouse et al. 1992). Although galls have interested scientists for centuries and galling research has spanned the fields of natural history, ecology, and chemistry, the question of how an insect induces gall formation remains unanswered.

The formal study of galls began in the seventeenth century (Raman et al. 2005).

Since that time many mechanisms have been proposed for gall induction (Mani 1992). 18

Some have focused on mechanical stimuli of the mouthparts or ovipositor, but most have centered on chemical compounds in fluids transferred from the insect to the plant, usually through maternal secretions or saliva (Sopow et al. 2003). McCalla et al. (1962) assayed the maternal secretions of a sawfly that galls willow, and found an array of compounds such as uric acid, glutamic acid, and uridine. When these chemicals were introduced artificially to plant tissue they could not induce galls to form, but when applied to an existing gall with the larva removed, they maintained gall growth (McCalla et al. 1962).

Similarly, indole acetic acid (IAA) and other growth-promoting plant hormones have been found in the salivary glands of insects, and have been suggested as gall inducers

(Hori et al. 1979; Miles 1967; Nuorteva 1956). However, the concentration of IAA in salivary glands of insects is far lower than in the plant before and after gall induction, and experimental application of IAA has had only mixed success in inducing galls (Hori

1992). Recently, it has been suggested that a chemical stimulus alone cannot describe the complexity and diversity of galls across the insect- and plant trees of life: galls range from a simple roll of a leaf to multi-chambered structures with hairs and pigments unlike those in any other plant tissue (Raman et al. 2005). This yields uncertainty regarding how a cocktail of chemicals contained in a very small amount of insect secretion can create such diverse and complex structures.

Several researchers have suggested that a genetic component such as RNA from the insect may be involved in gall formation (Hori 1992; Taylor 1949; Weis et al. 1986).

Galls generally are created specifically to fit the insect’s needs, differ morphologically from surrounding plant tissue, and are shaped more by insect species than by plant 19 species (Raman et al. 2005). Moreover, when gall traits are mapped phylogenetically, more closely related insects share more similar gall characteristics regardless of their host plants (Stone et al. 1998). High concentrations of nucleic acids and protein have been identified within maternal secretions of insects such as sawflies and the saliva of larval cynipid wasps (Hovanitz 1959; Taylor 1949), and when exposed to plant cell cultures these secretions induce aberrant plant tissue growth. However, whether this mechanism is consistent across all galling insects has yet to be established.

In the case of the fig- and fig-wasp interaction, the signal from the insect to trigger gall formation is unknown. However, the development of gall flowers in figs following oviposition and pollination has been well documented by Verkerke (1986).

Flowers that receive pollen, an egg, and several drops of a maternal secretion develop into galls; in contrast, flowers that receive only pollen develop into seeds. For both gall- and seed flowers, pollen germinates on the stigma and the pollen tube starts to form down the style. The pollen tube delivers the pollen to the female gametophyte, fertilizing it and creating a fig embryo and endosperm in the embryo sac. In seed flowers, the embryo sac slowly enlarges and the cotyledons differentiate, the endosperm is gradually reabsorbed, and the ovule enlarges regularly in all directions until the seed reaches maturity. In contrast, gall flowers are characterized by sudden, explosive growth immediately following oviposition, which is presumed to be triggered by the foundress wasp’s maternal secretion. Growth continues until the ovule attains its final size about nine days later. During the period of explosive growth the wasp egg hatches and the larva feeds on the growth of the plant. After the gall reaches its final size, the wasp larva moves into the 20 embryo sac and the fig embryo is aborted. The larva feeds on the remaining cellular endosperm, compressing the surrounding tissue until it fills the entire gall and is fully developed (Verkerke 1986).

The maternal secretion: key to the galling process?

The maternal secretion of the foundress wasp is thought to play an important role in gall initiation, similar to many other galling insects (McCalla et al. 1962; Sopow et al. 2003).

The maternal secretion of fig wasps is stored in the poison gland (Grandi 1929). The poison gland occurs throughout Hymenoptera and varies in its function. For example, in honey bees (Apis mellifera) the poison gland holds the venom associated with painful stings (Haberman 1972). In ants it holds communication pheromones (Hölldobler 1971);

(Vandermeer et al. 1980), and in ichneumonid wasps it stores polydnaviruses that are injected into their prey to compromise their immune system, thus protecting the wasp’s eggs (Espagne et al. 2004).

Several factors indicate that the poison gland plays an important role in ovipositioning and subsequent gall formation in the fig/fig wasp system. The maternal secretion is specific to ovipositing because it is placed only in flowers in which an egg has been laid (Verkerke 1986). The size of the poison gland in female fig wasps also suggests its importance. Slightly smaller than the ovaries, the poison sac is the second largest organ within adult, female fig wasps (Grandi 1929). Fig wasps are long-distance dispersers with a short lifespan (3-5 days), and maintaining a large, fluid-filled gland would appear costly unless its contents were vital for its fitness. As yet, however, the role 21 of this secretion in instigating gall growth is unknown. This motivates the final focus of this dissertation.

1.2 Explanation of dissertation format

This dissertation examines several aspects of the ecological relationship between fig trees and their pollinating wasps. These aspects are tied together under the broad question of stability in the fig- and fig-wasp mutualism, with special attention to (1) identifying the fungal community associated with syconia, (2) comparing poison sac morphology in pollinating and non-pollinating wasps, and (3) taking a metatranscriptome approach to address the unbeatable seed hypothesis and examine, for the first time, the developmental-transcriptomic events in different fig flowers at different points in the mutualistic interaction. As such this dissertation contains three papers, which are included as Appendices A-C. Each has either been published or is intended for publication. The appendices’ formats follow the requirements of the journal to which they have been or will be submitted.

In Appendix A, I used a culture-free approach to examine the diversity and composition of fungal communities associated with fig flowers at four developmental stages in six species of Ficus in lowland Panama. I addressed matters of stability and complexity in the fig and fig wasp mutualism by identifying the fungal community and testing the null hypotheses that fungal communities are consistent among Ficus species, gall- and seed flowers, and developmental stages of syconia. 22

In Appendix B, I compared the size of the poison sac to egg number in actively pollinating, passively pollinating, and non-pollinating fig wasps. I addressed matters of galling and complexity by using phylogenetically independent contrasts to test if the ratio of poison sac area to egg number is consistent among phylogenetically and ecologically diverse wasps, despite differences in body size and symbiotic affiliation with the fig tree.

I proposed the hypothesis that the relationship of poison sac size and egg number suggests a shared importance of the maternal secretion for gall induction in pollinating as well as non-pollinating wasps.

In Appendix C, I used a metatranscriptome approach to analyze eukaryotic gene expression within fig flowers at the time of oviposition choice and early gall development. I addressed matters of stability and galling by comparing the differences between gall and seed flower types that may influence oviposition of wasps and by analyzing gene expression of both plant and wasp at the beginning of gall development.

PRESENT STUDY

The methods, results and conclusions from this research are presented in the appended manuscripts. The following is a summary of the most important findings in this document.

2.1 Culture-free survey reveals diverse and distinctive fungal communities associated with developing figs (Ficus spp.) in Panama

The interaction between fig trees and fig-pollinating wasps is a classic, well-studied mutualism. However the microbial community associated with the mutualism is virtually unexplored. In Appendix A, my co-authors and I conducted a culture-free survey to examine the diversity and composition of fungal communities associated with fig flowers and their pollinating wasps. Fungal communities in syconia and on pollinating wasps were similar, dominated by diverse and previously unknown Saccharomycotina, and distinct from leaf- and stem endophyte communities in the same region. Fungal communities differed significantly in flowers after pollination vs. before pollination, consistent with the introduction of fungi by female wasps. Differences in fungal communities among different developmental stages suggest a possible role of fungi in providing volatile cues that indicate a termination of receptivity for pollinators and an onset of ripeness for frugivorous seed dispersers. Significant differences also were found between anciently diverged lineages of Ficus consistent with introduction and maintenance of different fungal communities by wasps of each clade, which may frequently host-switch within the actively and passively pollinated clades, but not between the two clades. Finally, fungal communities were similar between gall- and seed 24 flowers before pollination, leading us to suggest that fungi are not the drivers of oviposition choice between these flower types and do not appear to contribute to the presence of ‘unbeatable seeds’ in figs.

2.2 The relationship between gall formation and the fig wasp poison sac

Oviposition by mutualistic pollinating fig wasps and parasitic non-pollinating fig wasps produces galls within fig flowers identical in size and structure for the development of their offspring. Fig trees sanction their pollinating species of wasps if they gall without pollinating, but sanctions against species of non-pollinating wasps have not yet been identified. We posited that contents of the poison sac (i.e., the maternal secretion) are important for gall development in both pollinating and non-pollinating species of fig- galling wasps. As part of this framework we anticipated that the poison sac would be highly important in the reproductive biology of fig-gallers. In Appendix B, my co-authors and I compared the size of the poison sac in pollinating and non-pollinating wasps and used phylogenetically independent contrasts to determine the ratio of poison sac area to egg number. We found that even though the size of the poison sac differed between pollinators and non-pollinators, there is a strong positive correlation between the poison sac area and egg number that is consistent among galling wasps associated with figs regardless of their roles as pollinators or parasites. This observation is broadly consistent with the hypothesis that both types of wasp use a similar galling mechanism that prevents the fig tree from differentiating between mutualistic and parasitic associates allowing the non-pollinating wasps to gall the fig flowers unsanctioned. We suggested that maternal 25 secretion in the poison sac is important for gall induction by pollinating and non- pollinating wasps associated with figs.

2.3 Metatranscriptome analysis of fig flowers reveals possible mechanisms for mutualism stability and gall induction

Fig wasps consistently oviposit in the inner flowers of the fig syconium (gall flowers), but typically do not use the flowers closer to the syconium wall (seed flowers). Some authors suggest that this pattern is a stabilizing mechanism used by the fig tree to control the fig wasp-to-seed ratio. In Appendix C, my co-authors and I used a metatranscriptome approach to analyze eukaryotic gene expression within fig flowers at the time of oviposition choice and early gall development to better understand potential differences between gall and seed flowers that might influence oviposition choices, and the unknown mechanisms underlying gall formation. We found an abundance of transcripts assigned to flavonoid production and defense gene orthologs in both pollinated and receptive gall flowers of Ficus obtusifolia. We discovered detectable differences in metabolism and defense between gall flowers and seed flowers before oviposition, which provided support for a stabilizing mechanism. Differences in sugar and carbohydrate metabolism between gall and seed flowers suggested that seed flowers do not have the nutritional quality to support the development of a wasp larva or contain diminished nutrition that would impose a longer developmental time on larvae, or result in a smaller adult size.

The high expression of a wasp transcript annotated as the venom gene icarapin during wasp embryogenesis has implications for the mechanism of gall initiation. Finally, a high 26 level of Ficus transcripts assigned to chalcone synthase also was found at the beginning of gall initiation, similar to other galling organisms.

REFERENCES

Agrawal AA, Ackerly DD, Adler F, Arnold AE, Caceres C, et al. 2007. Filling key gaps in population and community ecology. Frontiers in Ecology and the Environment 5:145- 52

Anstett MC, Bronstein JL, HossaertMcKey M. 1996. Resource allocation: A conflict in the fig/fig wasp mutualism? Journal Evolutionary Biology 9:417-28

Axelrod R, Hamilton WD. 1981. The evolution of cooperation. Science 211:1390-6

Berg CC. 1989. Classification and distribution of Ficus. Experientia 45:605-11

Boucek Z. 1993. The genera of Chalcidoid wasps from Ficus fruit in the New-world. Journal of Natural History 27:173-217

Boucher DH, James S, Keeler KH. 1982. The ecology of mutualism. Annual Review of Ecology and Systematics 13:315-47

Bronstein JL. 1988. Mutualism, antagonism, and the fig-pollinator interaction. Ecology 69:1298-302

Bronstein JL, Alarcon R, Geber M. 2006. The evolution of plant-insect mutualisms. New Phytologist 172:412-28

Bull JJ, Rice WR. 1991. Distinguishing mechanisms for the evolution of cooperation. Journal of Theoretical Biology 149:63-74

Cardoza YJ, Teal PEA, Tumlinson JH. 2003. Effect of peanut plant fungal infection on oviposition preference by Spodoptera exigua and on host-searching behavior by Cotesia marginiventris. Environmental Entomology 32:970-6

Cook JM, Bean D, Power SA, Dixon DJ. 2004. Evolution of a complex coevolved trait: active pollination in a genus of fig wasps. Journal Evolutionary Biology 17:238-46

Currie CR, Scott JA, Summerbell RC, Malloch D. 2003a. Fungus-growing ants use antibiotic-producing bacteria to control garden parasites. Nature 423:461-

Currie CR, Wong B, Stuart AE, Schultz TR, Rehner SA, et al. 2003b. Ancient tripartite coevolution in the attine ant-microbe symbiosis. Science 299:386-8

Dawkins R. 1999. The extended phenotype : the long reach of the gene. New York: Oxford University Press

Doster MA, Michailides TJ, Morgan DP. 1996. Aspergillus species and mycotoxins in figs from California orchards. Plant Disease 80:484-9

Dunn DW, Yu DW, Ridley J, Cook JM. 2008. Longevity, early emergence and body size in a pollinating fig wasp - implications for stability in a fig-pollinator mutualism. Journal of Animal Ecology 77:927-35

Espagne E, Dupuy C, Huguet E, Cattolico L, Provost B, et al. 2004. Genome sequence of a polydnavirus: Insights into symbiotic virus evolution. Science 306:286-9

Fleming TH, Holland JN. 1998. The evolution of obligate pollination mutualisms: Senita cactus and Senita moth. Oecologia 114:368-75

Flock R, Wallace J. 1955. Transmission of fig mosaic by the eriophyid mite Aceria ficus. Phytopathology 45:52-4

Galil J, Eisikowitch D. 1971. Studies on mutualistic symbiosis between syconia and sycophilous wasps in monoecious figs. New Phytologist:773-87

Ganeshaiah KN, Kathuria P, Shaanker RU, Vasudeva R. 1995. Evolution of style-length variability in figs and optimization of ovipositor length in their pollinator wasps - a coevolutionary model. Journal of Genetics 74:25-39

Grandi G. 1929. Studio morfologico e biologico della Blastophaga psenes (L.). 2a ediz. riveduta. Bollettino del Laboratorio di Entomologia del R.Istituto Superiore Agrario di Bologna. 2:1-147

Haberman E. 1972. Bee and wasp venoms. Science 177:314-22

Herre EA. 1989. Coevolution of reproducive characteristics in 12 species of new world figs and their pollinator wasps Experientia 45:637-47

Herre EA. 1993. Population structure and the evolution of virulence in nematode parasites of fig wasps. Science 259:1442-5

Herre EA. 1995. Factors affecting the evolution of virulence: Nematode parasites of fig wasps as a case study. Parasitology 111:S179-S91

Herre EA, Knowlton N, Mueller UG, Rehner SA. 1999. The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends in Ecology and Evolution 14:49-53

Herre EA, West SA. 1997. Conflict of interest in a mutualism: documenting the elusive fig wasp-seed trade-off Proceedings of the Royal Society B 264:1501-7 29

Herre EA, Jandér KC, Machado CA. 2008. Evolutionary ecology of figs and their associates: recent progress and outstanding puzzles. Annual Review of Ecology, Evolution, and Systematics 39:439-58

Hoffman MT, Arnold AE. 2010. Diverse bacteria inhabit living hyphae of phylogenetically diverse fungal endophytes. Applied Environmental Microbiology 76:4063-75

Holland JN, DeAngelis DL, Schultz ST. 2004. Evolutionary stability of mutualism: interspecific population regulation as an evolutionarily stable strategy. Proceedings of the Royal Society B 271:1807-14

Hölldobler B. 1971. Sex pheromone in ant Xenomyrmex-floridanus. Journal of Insect Physiology 17:1497-9

Hori K. 1992. Insect Secretions and their effect on plant growth, with special reference to Hemipterans. In Biology of insect-induced galls, ed. aOR J.D. Shorthouse:157–70. New York.: Oxford University Press. pp.157–70

Hori K, Singh DR, Sugitani A. 1979. Metabolism of ingested indole compounds in the gut of three species of Heteroptera. Comparative Biochemistry and Physiology 64C:435- 8

Hovanitz W. 1959. Insects and plant gall. Scientific American:150-62

Jallow MFA, Dugassa-Gobena D, Vidal S. 2008. Influence of an endophytic fungus on host plant selection by a polyphagous moth via volatile spectrum changes. Arthropod- Plant Interactions 2:53-62

Jander KC, Herre EA. 2010. Host sanctions and pollinator cheating in the fig tree-fig wasp mutualism. Proceedings of the Royal Society B 277:1481-8

Jousselin E, Hossaert-McKey M, Herre EA, Kjellberg F. 2003. Why do fig wasps actively pollinate monoecious figs? Oecologia 134:381-7

Jousselin E, Kjellberg F. 2001. The functional implications of active and passive pollination in dioecious figs. Ecology Letters 4:151-8

Kjellberg F, Jousselin E, Bronstein JL, Patel A, Yokoyama J, Rasplus JY. 2001. Pollination mode in fig wasps: the predictive power of correlated traits. Proceedings of the Royal Society B 268:1113-21

Leigh EG, Rowell TE. 1995. The evolution of mutualism and other forms of harmony at various levels of biological organization. Ecologie 26:131-58

Machado CA, Jousselin E, Kjellberg FW, Compton SG, Herre EA. 2001. Phylogenetic relationships, historical biogeography and character evolution of fig-pollinating wasps. Proceedings of the Royal Society B:685-94

Mani MS. 1992. Introduction to Cecidology. In Biology of insect-induced galls, ed. JD Shorthouse, O Rohfritsch:3-8. New York: Oxford University Press. pp. 3-8

Marussich WA, Machado CA. 2007. Host‐specificity and coevolution among pollinating and nonpollinating New World fig wasps. Molecular Ecology 16:1925-46

McCalla DR, Genthe MK, Hovanitz W. 1962. Chemical nature of an insect gall growth-factor. Plant Physiology 37:98-103

Miles PW. 1967. Studies on the salivary physiology of plant-bugs: Transport from haemolymph to saliva. Journal Insect Physiology:1787-801

Miller MW, Phaff HJ. 1962. Successive microbial populations in Calimyrna figs. Applied Microbiology 10:394-400

Molbo D, Machado CA, Sevenster JG, Keller L, Herre EA. 2003. Cryptic species of fig- pollinating wasps: implications for the evolution of the fig-wasp mutualism, sex allocation, and precision of adaptation. Proceedings of the National Academy of Sciences 100:5867-72

Moran NA. 2007. Symbiosis as an adaptive process and source of phenotypic complexity. Proceedings of the National Academy of Sciences 104:8627

Mrak EM, Phaff HJ, Vaughn RH, Hansen HN. 1942. Yeasts occurring in souring figs. Journal of Bacteriology 44:441-50

Nefdt RJC, Compton SG. 1996. Regulation of seed and pollinator production in the fig- fig wasp mutualism. Journal of Animal Ecology:170-82

Neuhauser C, Fargione JE. 2004. A mutualism-parasitism continuum model and its application to plant-mycorrhizae interactions. Ecological Modeling 177:337-52

Nuorteva P. 1956. Studies on the effect of the salivary secretion of some Heteroptera and Homoptera on plant growth. Annals of the Entomological Society of America:102- 24

Oliver KM, Degnan PH, Hunter MS, Moran NA. 2009. Bacteriophages encode factors required for protection in a symbiotic mutualism. Science 325:992-4 31

Omacini M, Chaneton EJ, Ghersa CM, Muller CB. 2001. Symbiotic fungal endophytes control insect host-parasite interaction webs. Nature 409:78-81

Pellmyr O. 1997. Pollinating seed eaters: Why is active pollination so rare? Ecology 78:1655-60

Pellmyr O, Huth CJ. 1994. Evolutionary stability of mutualism between yuccas and yucca moths. Nature 372:257-60

Pereira RAS, do Prado AP. 2005. Non-pollinating wasps distort the sex ratio of pollinating fig wasps. Oikos 110:613-9

Raman A, Schaeffer CW, Withers TM. 2005. Galls and gall-inducing arthropods: An overview of the biology, ecology, and evolution. In Biology, Ecology, and Evolution of Gall-inducing Arthropods., ed. A Raman, CW Schaeffer, TM Withers:1-33. Enfield, NH.: Science Publisher, Inc, . pp. 1-33

Rasplus JY, Kerdelhué C, Le Clainche I, Mondor G. 1998. Molecular phylogeny of fig wasps Agaonidae are not monophyletic. Comptes Rendus de l'Académie des Sciences- Series III-Sciences de la Vie 321:517-27

Rønsted N. 2005. 60 million years of co-divergence in the fig–wasp symbiosis. Proceedings of the Royal Society B 272:2593-9

Sachs JL, Mueller UG, Wilcox TP, Bull JJ. 2004. The evolution of cooperation. The Quarterly Review of Biology 79:135-60

Sachs JL, Simms EL. 2006. Pathways to mutualism breakdown. Trends in Ecology and Evolution 21:585-92

Scott JJ, Oh DC, Yuceer MC, Klepzig KD, Clardy J, Currie CR. 2008. Bacterial protection of beetle-fungus mutualism. Science 322:63

Shorthouse JD, Rohfritsch O. 1992. Biology of insect-induced galls. Oxford University Press

Sopow SL, Shorthouse JD, Strong W, Quiring DT. 2003. Evidence for long-distance, chemical gall induction by an insect. Ecology Letters 6:102-5

Stern DL. 1995. Phylogenetic evidence that aphids, rather than plants, determine gall morphology Proceedings of the Royal Society B 260:85-9

Stone GN, Cook JM. 1998. The structure of cynipid oak galls: patterns in the evolution of an extended phenotype. Proceedings of the Royal Society B 265:979-88

Suryanarayanan TS, Vijaykrishna D. 2001. Fungal endophytes of aerial roots of Ficus benghalensis. Fungal Diversity 8:155-61

Taylor SH. 1949. Initiation and development of the gall of Aylax glechomae on Nepeta hederacea. American Journal of Botany 36:222-30

Van Bael SA, Fernández-Marín H, Valencia MC, Rojas EI, Wcislo WT, Herre EA. 2009. Two fungal symbioses collide: endophytic fungi are not welcome in leaf-cutting ant gardens. Proceedings of the Royal Society B 276:2419-26

Vandermeer RK, Glancey BM, Lofgren CS, Glover A, Tumlinson JH, Rocca J. 1980. The poison sac of red imorted fire ant queens- source of a pheromone attractant Hymenoptera, Formicidae. Annals of the Entomological Society of America 73:609-12

Verkerke W. 1986. Anatomy of Ficus ottoniifolia (Moraceae) syconia and its role in the fig-fig wasp symbiosis. Proceedings der Koninklijke Nederlandse Akademie van Wetenschappen 443-69

Vidal S. 1996. Changes in suitability of tomato for whiteflies mediated by a non- pathogenic endophytic fungus. Entomologia experimentalis et applicata 80:272-4

Wang GQ, Wei SG, Yang D. 2009. Six new eriophyoid mites (Acari: Eriophyoidea) associated with Ficus spp. (Moraceae) from China. Zootaxa 2201:49-62

Wang HK, Hyde KD, Soytong K, Lin FC. 2008. Fungal diversity on fallen leaves of Ficus in northern Thailand. Journal of Zhejiang University-Science B 9:835-41

Weis AE, Abrahamson WG. 1986. Evolution of host-plant manipulation by gall makers - ecological and genetic factors in the Solidago-Eurosta system. American Naturalist 127:681-95

West SA, Herre EA. 1994. The ecology of the new-world fig-parasitizing wasps Idarnes and implications for the evolution of the fig-pollinator mutualism. Proceedings of the Royal Society B 258:67-72

West SA, Herre EA, Windsor DM, Green PRS. 1996. The ecology and evolution of the New World non‐pollinating fig wasp communities. Journal of Biogeography 23:447-58

Wiebes JT. 1982. The phylogeny of Agaonidae (Hymenoptera, Chalcidoidea). Netherlands Journal of Zoology 32:395-411

APPENDIX A:

CULTURE-FREE SURVEY REVEALS DIVERSE AND DISTINCTIVE FUNGAL

COMMUNITIES ASSOCIATED WITH DEVELOPING FIGS (FICUS SPP.) IN

PANAMA

Published in Microbial Ecology (2012) doi 10.1007/s00248-012-0079-x

Rightslink Printable License https://s100.copyright.com/App/PrintableLicenseFrame.jsp?pub...

SPRINGER LICENSE TERMS AND CONDITIONS Oct 08, 2012

This is a License Agreement between Ellen O Martinson ("You") and Springer ("Springer") provided by Copyright Clearance Center ("CCC"). The license consists of your order details, the terms and conditions provided by Springer, and the payment terms and conditions.

All payments must be made in full to CCC. For payment instructions, please see information listed at the bottom of this form.

License Number 3004250214332

License date Oct 08, 2012

Licensed content publisher Springer

Licensed content publication Microbial Ecology

Licensed content title Culture-Free Survey Reveals Diverse and Distinctive Fungal Communities Associated with Developing Figs (Ficus spp.) in Panama

Licensed content author Ellen O. Martinson

Licensed content date Jan 1, 2012

Type of Use Thesis/Dissertation

Portion Full text

Number of copies 1

Author of this Springer Yes and you are the sole author of the new work article

Order reference number

Title of your thesis / Mutualism stability and gall induction in the fig and fig wasp dissertation interaction

Expected completion date Nov 2012

Estimated size(pages) 150

Total 0.00 USD

Terms and Conditions

Introduction The publisher for this copyrighted material is Springer Science + Business Media. By clicking "accept" in connection with completing this licensing transaction, you agree that the following terms and conditions apply to this transaction (along with the Billing and Payment terms and conditions established by Copyright Clearance Center, Inc. ("CCC"), at the time that you opened your Rightslink account and that are available at any time at

1 of 4 10/8/12 10:34 AM 35

Microb Ecol DOI 10.1007/s00248-012-0079-x

HOST MICROBE INTERACTIONS

Culture-Free Survey Reveals Diverse and Distinctive Fungal Communities Associated with Developing Figs (Ficus spp.) in Panama

Ellen O. Martinson & Edward Allen Herre & Carlos A. Machado & A. Elizabeth Arnold

Received: 24 February 2012 /Accepted: 30 May 2012 # Springer Science+Business Media, LLC 2012

Abstract The ancient association of figs (Ficus spp.) and pollination, fungal communities were similar between gall- their pollinating wasps (fig wasps; Chalcidoidea, Hymenop- and seed flowers and among Ficus species. However, fungal tera) is one of the most interdependent plant–insect mutual- communities differed significantly in flowers after pollina- isms known. In addition to pollinating wasps, a diverse tion vs. before pollination, and between anciently diverged community of organisms develops within the microcosm lineages of Ficus with active vs. passive pollination syn- of the fig inflorescence and fruit. To better understand the dromes. Within groups of relatively closely related figs, multipartite context of the fig–fig wasp association, we used there was little evidence for strict-sense host specificity aculture-freeapproachtoexamine fungal communities between figs and particular fungal species. Instead, mixing associated with syconia of six species of Ficus and their of fungal communities among related figs, coupled with pollinating wasps in lowland Panama. Diverse fungi were evidence for possible transfer by pollinating wasps, is con- recovered from surface-sterilized flowers of all Ficus spe- sistent with recent suggestions of pollinator mixing within cies, including gall- and seed flowers at four developmental syconia. In turn, changes in fungal communities during fig stages. Fungal communities in syconia and on pollinating development and ripening suggest an unexplored role of wasps were similar, dominated by diverse and previously yeasts in the context of the fig–pollinator wasp mutualism. unknown Saccharomycotina, and distinct from leaf- and stem endophyte communities in the same region. Before Introduction

Electronic supplementary material The online version of this article (doi:10.1007/s00248-012-0079-x) contains supplementary material, Mutualisms are a feature of every ecosystem and increas- which is available to authorized users. ingly are recognized as a driving force in the diversification of life on earth [14, 40]. Often characterized as bipartite E. O. Martinson Department of Ecology and Evolutionary Biology, exchanges of commodities such as nutrition, protection, or The University of Arizona, enhanced reproductive success [e.g., 11, 20, 41, 48, 66, 86], Tucson, AZ 85721, USA mutualisms exist within communities of species that can shape the currency or rate of exchange between partners E. A. Herre Smithsonian Tropical Research Institute, [15, 62, 72]. Ecologists increasingly appreciate that mutual- Balboa, Ancon, isms should be interpreted in a multipartite context [e.g., 1, Panama City, Republic of Panama 38, 88], which often reveals previously unexplored components of even the most classic two-partner associations [e.g., C. A. Machado Department of Biology, The University of Maryland, 22, 43, 63, 78]. College Park, MD 20742, USA Fig trees (Ficus, Moraceae) and their pollinating wasps (fig wasps; Chalcidoidea, Hymenoptera) share a coevolu- A. E. Arnold (*) tionary history that spans up to 90 million years [56, 57, 71]. School of Plant Sciences, The University of Arizona, Tucson, AZ 85721, USA Their interactions represent some of the most interdependent e-mail: [email protected] plant–insect mutualisms known [49, 54]. With the exception

E. O. Martinson et al.

of parthenocarpic figs used in agriculture, Ficus spp. depend of microbes in gall formation, host plant selection by herbi- solely on fig wasps to transfer their pollen from tree to tree, vores, and plant nutritional quality are well recognized [9, and the larvae of fig wasps can only develop within fig 32, 64, 67, 70]. Previous studies have detected yeasts in flowers. cultivated figs [e.g., 59, 60], and recent work suggests that When a female fig wasp enters a receptive fig (syconi- these fungi influence volatile signatures of mature figs, with um), she encounters hundreds of flowers arranged in two effects on frugivory by bats [73]. However, despite several layers. Flowers that receive pollen yield a fig seed, whereas surveys of fungal communities within leaves and rotting those receiving an egg develop into a gall that provides fruits of fig trees [24, 59, 60, 75, 83], fungal communities nutrition to the wasp’soffspringattheexpenseofthat within developing syconia of non-domesticated Ficus spp. flower’s seed production [41, 87]. Some species of wasps have not been studied previously. pollinate actively, storing pollen in specialized pockets and We used a culture-free approach to examine the diversity fertilizing flowers individually [50]. However, species in the and composition of fungal communities associated with fig basal lineages of fig-pollinating wasps do so passively, flowers at four developmental stages. Sampling encom- fertilizing inflorescences haphazardly as pollen brushes off passed six species of Ficus and their pollinating wasps, the wasp’s body [50]. Passive pollination is considered the including both actively and passively pollinated figs from ancestral condition in Ficus, with ca. 60 million years sep- a lowland, moist tropical forest in Panama. Here we exam- arating the New World actively- and passively pollinated ine fungal communities among Ficus species, gall- and seed clades [47, 50]. flowers, and developmental stages of syconia to ask: (1) do Regardless of pollination syndrome, female fig wasps fungal communities differ among Ficus spp., such that they (foundresses) consistently choose the inner ring of flowers may play a role in pollinator attraction to particular species (hereafter, gall flowers), rather than the flowers closer to the of Ficus? (2) Do communities differ in gall- vs. seed flow- syconium wall (hereafter, seed flowers), for oviposition ers, such that they may influence oviposition by pollinators? [80]. The reason for this preference is not known, but (3) Do communities differ in syconia as a function of de- explanations such as limited ovipositor length and parasitoid velopmental stage, such that they may cue pollinators to avoidance have been refuted [see 13, 25, 30]. The observa- indicate the conclusion of receptivity or frugivores to indi- tion that foundresses consistently oviposit in only ~50 % of cate ripeness? available flowers despite having sufficient eggs to deposit in more, and thus die after realizing only a portion of their reproductive potential [29, see also 41], led West and Herre Materials and Methods [88] to suggest that some flowers may be impervious to ovipositioning and/or gall development. The mechanism In January–April 2010, developing figs from one mature by which these “unbeatable seed flowers” [sensu 88] differ individual of each of six species of Ficus were collected at from gall flowers is not known, but preference against them Barro Colorado National Monument, Panama (BCNM; is strong: with few exceptions, even non-pollinating wasps, 9°9′ N, 79°51′ W; 25 m above sea level; for a full site which oviposit from outside the syconium and do not pol- description see [53]). Focal species represent both actively linate figs, preferentially use gall flowers even though the pollinated species (Ficus costaricana, Ficus obtusifolia, Ficus outer ring of seed flowers is more accessible [88]. Structural popenoei,andFicus triangle;subgenusUrostigma,section features of flowers such as ovary position or style length do Americana) pollinated by Pegoscapus spp., and passively not explain the selective avoidance of seed flowers by pollinated species (Ficus insipida, Ficus maxima;subgenus pollinators or parasitic wasps [13, 80], prompting us to Pharmacosycea,sectionPharmacosycea)pollinatedbyTet- explore alternative explanations. rapus spp. [12, 57]. All surveyed trees were located at the In addition to pollinating wasps, a diverse community of edge of Lake Gatún, where their readily accessible canopies organisms develops within the microcosm of the syconium overhang the water. Trees were separated by a mean of 2.6 km (e.g., non-pollinating wasps, nematodes, and mites) [39, 84, (±1.8 km) (Supplemental Fig. 1). 88]. Microbial communities associated with developing sy- Intact, apparently healthy syconia were collected in conia are an especially unexplored aspect of the fig–wasp four developmental stages. Receptive syconia (hereafter, mutualism with potential implications for oviposition choice receptive or pre-pollination) contained fully developed both at the level of Ficus species and tissue type (gall- vs. flowers but had not yet been entered by a pollinating seed flowers). Female-pollinating wasps use volatile cues to wasp (typically a span of 24 to 72 h after gall- and seed identify receptive figs of the appropriate species and devel- flowers differentiate) [80]. Early post-pollination syconia opmental stage for oviposition [35, 79, 85, 87]. Some plant- (hereafter, early) were collected after a pollinating wasp associated microbes influence oviposition behavior of had entered and oviposited but before larvae pupated insects by altering volatile signals [17, 46, 81], and the roles (a span of 1 to 2 weeks) [80]. Late post-pollination

Fungi in Developing Figs

syconia (hereafter, late) were collected after galls and following program: 94 °C for 3 min; 35 cycles of 94 °C seeds within the fig had developed fully and wasps had for 30 s, 52 °C for 30 s, and 68 °C for 1 min; and 68 °C for pupated (a span of 2 to 4 weeks) [80]. Ripe fruits were 8min.EthidiumbromidewasusedtovisualizeDNAbandson collected after female wasps emerged from the fig but 1.2 % agarose gels. Positive controls containing verified fun- before a fruit dropped from the tree (a span of 1 to gal DNA, and negative controls containing sterile distilled 4days;EOMpersonalobservation)[80]. Collections water in place of DNA template, were run with every PCR. were staggered such that figs at different developmental Any reaction set with a failure in either control was removed stages were harvested at the same time to decouple date from the study. of collection from developmental stage. Positive products were cloned using the Stratagene Stra- Wasps were collected from late post-pollination figs at taClone PCR Cloning Kit (La Jolla, CA) using the manu- BCNM in June–August 2009 (wasps from F. obtusifolia, facturer’s protocol, followed by PCR with primers T3 and F. maxima, and F. popenoei; stored in sterile SDS buffer T7. Up to 15 positive clones per sample were chosen hap- at −20 °C) and April 2005 (wasps from F. insipida, F. hazardly for sequencing. PCR products were cleaned by costaricana,andF. triangle; stored in 70 % ethanol at adding 0.2 μlofNEBcalfintestinalphosphataseand 4 °C). Wasps were collected from different individual 0.2 μl of NEB exonuclease I to each sample, vortexing for trees than those sampled above but from the same area 30 s, and incubating for 15 min at 37 °C followed by 15 min surrounding Lake Gatún (Machado, unpublished data). at 80 °C (J. Stavrinides, personal communication). Products were sequenced bidirectionally at the UAGC DNA Extraction, PCR, and Sequencing sequencing facility at The University of Arizona on an Applied Biosystems 3730xl DNA Analyzer (Foster City, Figs were stored at 4 °C in sealed plastic bags and processed CA). Contigs were assembled and basecalls verified manu- within 24 h after collection. Gall- and seed flowers from ally based on chromatograms in Sequencher v. 4.5 (Gene each syconium were separated with sterile microforceps Codes, Ann Arbor, MI). No chimeric sequences were under a dissecting microscope and stored separately in detected. Sequences have been deposited in GenBank under 70 % ethanol at −20 °C. Flowers were surface-sterilized accession numbers JX174729-JX175042. by sequential immersion in 95 % ethanol (30 s), 10 % bleach (0.5 % NaOCl; 2 min), and 70 % ethanol (2 min) [7] Ecological Analyses followed by three rinses with sterile distilled water. This method removed exogenous DNA that might have contam- Operational taxonomic units (OTU) were defined on the inated samples in the lab from flower surfaces [29]. Wasps basis of 95 % sequence similarity over shared sequence were not surface-sterilized so that fungi on wasps’ cuticles lengths with a criterion of at least 40 % overlap using could be evaluated [28]. Sequencher 4.5 [6], which estimates OTU that are con- Each sample of fig tissue, defined as a 0.2-ml tube gruent with species-level clades of tropical plant- containing gall- or seed flowers collected at the same associated fungi [76]. To select representative clones for time from one to three syconia from the same tree, was phylogenetic analyses, we chose one member of each ground in liquid nitrogen prior to extraction of total group from figs or wasps as defined by 99 % sequence genomic DNA using the Qiagen DNeasy Plant Mini Kit similarity (following [29]). This approach allows for mi- (Germantown, MD; manufacturer’sprotocol).Three nor sequencing errors while still capturing the genotypic wasps per species were pooled prior to DNA extraction diversity of the sample. with the Qiagen Puregene Core Kit A (Germantown, Species accumulation curves, bootstrap estimates, and MD; manufacturer’s protocol). diversity (measured as Fisher’s α, which is robust to varia- The largely fungal-specific primer ITS1F and nonselec- tion in sample size [27]) were inferred in EstimateS v. 8.2.0 tive primer LR3 (CTTGGTCAT TTAGAGGAAGTAA and (http://viceroy.eeb.uconn.edu/estimates). Similarity among GGTCCGTGTTTCAAGAC, respectively) [31, 82]were partitions of the fungal community was assessed in PAST used to PCR-amplify the fungal nuclear ribosomal v. 2.06 [37] or EstimateS v. 8.2.0 using OTU (based on 95 % internal-transcribed spacers and 5.8S gene (ITS; ca. sequence similarity, as above) that were found more than 600 bp) and an adjacent portion of the nuclear ribosomal once (i.e., non-singletons). Similarity values were calculated large subunit (LSU; ca. 500 bp) as a single fragment. Each using Jaccard’sindex(JGR,basedonpresence/absence 25 μl reaction mixture contained 12.5 μl GoTaq® Green data) and the Morisita index (MGR, based on incidence). Master Mix (Madison, WI), 1 μl of each primer (5 μM), Indices were compared statistically using analysis of simi- 2 μl of DNA template, and 8.5 μl of PCR-quality H2O. larity (ANOSIM; [19]) with visualization by non-metric Reactions were run on an Eppendorf Mastercycler ep multidimensional scaling in PAST v. 2.06 [37] or Wilcoxon gradient S thermocycler (Hamburg, Germany) with the tests in JMP v. 8.0.1 (www.jmp.com).

E. O. Martinson et al.

Comparison with Non-Syconia Endophyte Communities Fig. 2). The UniFrac metric determines the fraction of the total branch length that is unique or shared by two communi- To assess the distinctiveness of syconia-associated fungi ties, with statistical support determined by 1,000 permutations relative to fungi occurring in symbiosis with other aerial [55]. UniFrac scores were assessed using both unweighted plant parts, one representative of each non-singleton OTU (based on presence/absence) and weighted (based on relative obtained from figs was compared against a database of abundance) analyses. 5,010 ITS-partial LSU sequences representing 581 OTU of To determine the placement of sequences obtained leaf and stem endophytes from central Panama [3–5, 7, 8, from wasps, 80 sequences were integrated into the 42, del Olmo, unpublished data; Arnold, unpublished data]. alignment described above using MAFFT v. 6 [51]. Sequence data represented isolates obtained in culture and Phylogenetic relationships were inferred using ML in sequences obtained by cloning, as mentioned above, from RAxML [74]usingGTR+I+γ,asmentionedabove. healthy tissues of 258 species representing 190 genera and Topological support was evaluated by 1,000 ML boot- 28 families of vascular plants (including Ficus)insites strap replicates. throughout central Panama in the wet and dry seasons of 1999–2010. Of these, 4,061 sequences represented fungi that were isolated (3,671 isolates) or directly sequenced Results (390 clones) from diverse plants at BCNM. Data from figs and wasps were compared against the endophyte data in Fungi were detected in every sample of syconia tissue from Sequencher as described above to determine groups with six species of Ficus in lowland Panama, and from all sam- 95 and 100 % sequence similarity. ples of wasps associated with these fig species (Table 1). A total of 234 ITS-partial LSU sequences representing 26 Phylogenetic Analyses samples of fig flower tissue yielded 23 OTU (based on 95 % sequence similarity; Fisher’s α08.72; 30.4 % single- Fungal ITS-partial LSU sequences from figs and wasps tons) and 81 genotypes (based on 99 % sequence similarity). were compared to the NCBI non-redundant database by A total of 80 sequences from six samples of wasps yielded 9 the basic local alignment search tool (BLASTn) [2]to OTU (Fisher’s α010.02; 5.0 % singletons) and 19 geno- estimate taxonomic placement at the class level and above types. Comparison of the bootstrap estimate of total species and to establish taxon sampling for phylogenetic analyses. richness with the 95 % confidence interval around observed The 5.8S and LSU portion of one representative sequence of richness indicated that our sampling effort was statistically each unique genotype obtained from each sample of fig sufficient to capture the total estimated OTU richness for flowers (defined by 99 % overall sequence similarity; N081 each fig species, both flower types, each developmental sequences) was aligned using MAFFT v. 6 [51]with37 stage, and the wasps evaluated here (Fig. 1; Supplemental reference sequences selected from the top BLASTn hits Figs. 3 and 4), providing a robust basis for community obtained from GenBank. The alignment was adjusted manu- comparisons. ally and ambiguously aligned regions were excluded in Mes- quite v. 2.74 [58]. The alignment is accessioned at TreeBase Community Structure Inferred from Fungal OTU under accession 12698. Phylogenetic relationships were inferred using maxi- Relative to fungal communities found in living stems and mum likelihood (ML) in RAxML [74]andBayesian leaves of vascular plants of the region, fungal communities MCMCMC in Mr. Bayes v. 3.1.2 (seven million gener- in figs and pollinating fig wasps were highly distinct. No ations, two chains, each initiated with random trees, and sequences of fungi found in syconia were 100 % identical to sampling every 1,000th tree) [45] using GTR+I+γ, de- a previously recorded leaf- or stem endophyte. None of the termined to be the best-fitting model of evolution based Saccharomycotina OTU found in our surveys was detected on comparisons of the Akaike information criterion in previously as an endophyte using culture-based- or culture- ModelTest 3.7 [68]. Topological support was evaluated free methods. Five of 27 OTU from syconia and wasps were further by 1,000 ML bootstrap replicates. Output from 95 % similar to leaf- or stem endophytes (OTU 4, 6, 8, 10, MrBayes was filtered to remove the burn-in, defined as and 11); all were clones with top BLAST matches for the sample of the posterior for which the standard devi- Pezizomycotina, which made up <10 % of sequences found ation of the split frequencies was >0.01, and a majority in the present survey (see Fig. 2). rule consensus was constructed from 5.2 million trees in Fungal communities from figs strongly resembled Mesquite. those recovered from pollinating wasps both in terms Phylogenetic diversity of fungi was assessed with UniFrac of presence/absence and relative abundance of fungal [55]usingtheuncollapsed,mostlikelytree(Supplemental OTU (JGR: p00.9228; MGR: p00.8875). Overall, the

Fungi in Developing Figs

Table 1 Fungi obtained via cloning from syconia at four developmental stages, representing six species of Ficus as well as their pollinating wasps in lowland Panama

Sequences Species OTU Bootstrap Fisher's alpha (samples)represented (95 % CI)estimate (SD)

Ficus flower survey Fig species Passively pollinated F. insipida 63 (6) – 11 (7.5–14.5) 12.2 3.86 (0.80) F. maxima 19 (4) – 6 (4.7–7.3) 6.6 3.02 (1.10) Actively pollinated F. costaricana 20 (2) – 4 (2.7–5.3) 4.5 1.50 (0.55) F. obtusifolia 45 (5) – 12 (9.5–14.5) 14.0 5.35 (1.27) F. popenoei 24 (4) – 5 (4.1–5.9) 5.6 1.92 (0.63) F. triangle 63 (6) – 10 (7.5–12.5) 10.9 3.35 (0.71) Life stage Receptive 61 (8) 5 13 (10.4–15.6) 14.0 5.06 (1.04) Early 74 (8) 5 14 (10.6–17.4) 15.8 5.11 (0.97) Late 87 (8) 4 14 (9.6–18.4) 15.7 4.72 (0.85) Ripe 12 (2) 2 1 (1.0–1.0) 1.0 – Flower type Gall 127 (13) 6 22 (20.2–23.8) 23.8 7.68 (1.13) Seed 107 (13) 6 21 (15.1–26.9) 24.3 7.81 (1.23) Overall 234 (26) 6 29 (25.7–32.3) 32.8 8.72 (1.00) Pollinating fig wasp survey Overall 80 (6) 6 8 (4.7–11.3) 9.3 2.22 (0.46)

Columns indicate the number of sequences obtained; the number of species represented in each pool (as relevant); fungal OTU richness; bootstrap estimate of fungal OTU richness; and diversity (Fisher’s alpha) for each flower type, life stage, fig species, and survey

combined fig and wasp data sets included 27 OTU; of these, limited to draw conclusions). However, communities of 12 were found in both fig- and fig–wasp surveys despite the fungi differed as a function of the developmental stage of different timing of sampling and the collection of material syconia in two ways. First, communities in receptive from different individual trees (Fig. 2; Supplemental Fig. 5). flowers differed significantly from those of pollinated Fungal community composition did not differ significant- flowers (Fig. 3; ripe figs excluded and early- and late ly among species of Ficus (range of JGR00.1995–1.00, p> pollinated flowers pooled for analysis because they were 0.05 in all cases; MGR00.1545–1.00; p>0.05 in all cases; highly similar: JGR: p00.3512; MGR: p00.3777). Sec- data not shown). Fungal diversity, defined as Fisher’salpha, ond, actively and passively pollinated fig species, which

was similar among fig species overall (F5,2200.67, p00.6536; differ in their pollinating wasps, had highly similar communi- Table 1). Fungal communities were especially similar among ties prior to entrance by pollinators (JGR: p00.8550; MGR: syconia of different species in the receptive phase (e.g., Fig. 3, p00.5713) but differed significantly after pollination (JGR: Table 2) but differed markedly after pollination as described p00.0303; MGR: p00.0185; Fig. 3b). below. Composition of fungal communities did not differ signifi- Community Structure Inferred from Phylogenetic Analyses cantly between gall- and seed flowers overall (JGR00.2685, p0 0.2671; MGR00.5571, p00.5605; Supplemental Fig. 6), and Phylogenetic analyses of fungal communities from figs and diversity was similar between flowers of each type (t220−0.33, fig wasps (Fig. 2; see also Supplementary Fig. 2) corrobo- p00.7442). However, even though fungal communities were rate OTU analyses in six ways. First, taxonomic placement especially similar between gall- and seed flowers in the recep- of syconia- and wasp-associated fungi reveals their distinctive phase, they diverged markedly after pollination (Table 3). tiveness relative to foliar- and stem endophyte communities Diversity of fungi was similar among receptive, early, in central Panama. The 27 OTU recovered here encompass and late stages of syconium development (F2,2200.43, at least five classes of fungi, including Basidiomycota p00.6536; sequencing success from ripe figs was too (Microtyromycetes and Tremellomycetes, two OTU, and

E. O. Martinson et al.

Figure 1 Accumulation of A B operational taxonomic units 40 20 All Sequences Receptive Flowers (OTU) of fungi from Ficus

syconia, 95 % confidence 30 15 intervals, and bootstrap estimates of richness based on ITS-LSU OTU (defined by 20 10 95 % sequence similarity) for a combined sequences from the 10 5 entire survey (234 sequences); b receptive flowers; c early 0 0 stage pollinated flowers; d late 0 25 50 75 100 125 150 175 200 225 0 20 40 60 80 stage pollinated flowers; e gall flowers (from both receptive C D 20 20 and pollinated syconia); and f Pollinated Flowers (Early) Pollinated Flowers (Late) seed flowers (from both receptive and pollinated syconia) of 15 15 six species of Ficus in lowland Panama 10 10 OTU

5 5

0 0 0 20 40 60 80 0 20 40 60 80

E F 30 All Gall Flowers 30 All Seed Flowers

20 20

10 10 Observed richness 95% confidence intervals Bootstrap estimate 0 0 0 25 50 75 100 125 0 25 50 75 100 125 Sequences

seven sequences) and Ascomycota (primarily Saccharomy- syconia-associated yeasts (Fig. 2,SupplementaryFig.5 cotina (Saccharomycetes), encompassing 15 OTU and 281 clade containing OTU 19–23). This lineage encompasses sequences overall and, to a lesser extent, Pezizomycotina, the most common OTU in our surveys (172 sequences) including Dothideomycetes (6 OTU, 20 sequences), Sordar- and includes no sequence with >81 % affinity for any iomycetes (two OTU, three sequences), Eurotiomycetes sequence data available through GenBank. (one OTU, one sequence), and one OTU of uncertain affin- Third, syconia of all species of Ficus harbored phyloge- ity (three OTU, two sequences)). In contrast, leaf- and stem netically similar fungal communities (Unifrac analysis of endophyte communities are strongly dominated by Pezizo- uncollapsed tree containing samples from receptive and mycotina in lowland Panamanian forests (>98 % of sequen- pollinated figs: p00.15–1.00 in presence/absence analyses; ces in all surveys to date, with particular dominance by p00.30–1.00 in weighted analyses; Supplementary Fig. 2). Sordariomycetes, followed by Dothideomycetes and Euro- No strict-sense host specificity was observed at either the tiomycetes [3–8, 42, Arnold, unpublished data]). None of community level or within individual OTU. For example, the common clades of yeasts recovered from syconia has the most common phylotypes (OTU 21 and 23, Saccharo- been found in surveys of foliar endophytes in Panama. mycotina) occurred in every species, flower type, and de- Second, all non-singletons obtained from the pollinat- velopmental stage and also were found in samples from ing wasp survey grouped with phylotypes of fungi wasps (Fig. 2, Supplemental Fig. 2). Wasps were frequently known to date only from syconia (Fig. 2,Supplemental found with fungi that were not recovered from their natal Fig. 5). The four singletons from wasps were placed species of fig and instead were known from other fig species with strong support within a distinctive lineage of (Fig. 2).

Fungi in Developing Figs

Leucosporidium antarcticum AF444529 Sequences Rhodotorula glutinis FJ345357 in OTU Wasps Gall Seed ReceptivePollinated F. insipidaF. maximaF. costaricanaF. obtusifoliaF. popenoeiF. triangle 71 Rhodotorula mucilaginosa AB193175 75 100 100 Rhodotorula sp SY100 AB026010 OTU 1 6 100 Cryptococcus victoriae AM160647 100 Cryptococcus tephrensis BQ000318 1 OTU 2 Taphrina deformans DQ470973

75 Schizosaccharomyces pombe Z19136 100 OTU 3 2 OTU 4 2 100 98 100 100 100 100 Fusarium brachygibbosum GQ505450 100 100 OTU 5 1 -- 99 OTU 6 2 T 100 Coniozyma leucospermi EU552113 100 OTU 7 1 72 95 85 Phoma medicanginis EU167575 90 OTU 8 2 100 OTU 9 100 1

99 OTU 10 4 100 100 100 Bipolaris spicifera GU183125 3 100 OTU 11 100 Alternaria tenuissima AY154712 100 1 100 OTU 12 100 Saccharomyces cerevisiae EU649673 9 100 OTU 13 Ogataea cecidiorum FJ897742 98 100 Debaryomyces hansenii FN667855

Candida quercitrusa AM160627 -- 98 93 100 Candida sp HA 1671 AM160629

Candida tropicalis FJ515227 100 100 Lodderomyces elongisporus FJ515234

Williopsis saturnus FN868262

Pichia sp UFMG EU580140 83 99 Pichia anomala M10 FJ865436 100 95 100 OTU 14 3 100 100 P 100 OTU 15 4 Candida sp A564 HM364288

99 OTU 16 100 99 7 100 100 100 OTU 17 19 Candida picinguabensis DQ377638

Metschnikowia agaves HM189300

Candida gelsemii DQ988045 90 99 Metschnikowia bicuspidata EU809439

100 -- Candida picachoensis FJ614653 93 99 Unidentified Metschnikowia AM161112 98 100 Metschnikowia sp XY201 DQ367882 -- 93 Metschnikowia sp XY103 DQ367881 100 IMOP 100 OTU 18 24 Candida flosculorum EF137918 92 100 Candida akabanensis EU100744

Candida sp 1A EF565862

Candida sp C107DX Y2 FJ865216

Clavispora lusitaniae AY174089 77 91 Candida oregonensis FN667841 100 100 OTU 19 7 98 100 100 100 OTU 20 2 99 T 100 100 OTU 21 45 100100 OTU 22 1 100 99 CO 100 OTU 23 87

0.3

Figure 2 Most likely tree resulting from RAxML analyses of 5.8S and number of sequences in each OTU, followed by darkened circles that partial LSU sequence data, including results of maximum likelihood indicate presence of the OTU in the survey of pollinating wasps, each bootstrap (≥70 %; above branches) and Bayesian posterior probabili- flower type, broad developmental stage, or fig species. Superscripts in ties (≥90 %, below branches), revealing the phylogenetic placement of the wasp column indicate from which Ficus species the pollinating fungal OTU recovered from figs in six species of Ficus in Panama. wasps were collected: C0F. costaricana, O0F. obtusifolia, P0F. pope- Leucosporidium antarcticum was included as an outgroup. Accession noei, T0F. triangle, I0F. insipida, and M0F. maxima. Gray area numbers are listed for all sequences obtained from GenBank. Black indicates passively pollinated fig species; all others are actively boxes indicate 95 % sequence similarity groups. Columns indicate the pollinated

E. O. Martinson et al.

ab 1.00 Early vs late 1.00 Receptive flowers: pollinated flowers actively vs passively pollinated Receptive vs Pollinated flowers: 0.75 pollinated flowers 0.75 actively vs passively pollinated

0.50 0.50

Similarity Value Similarity 0.25 0.25 * * * 0 0 Jaccard Morisita Jaccard Morisita

Figure 3 Similarity of fungal communities between a early vs. late (alpha00.05). Analyses are based on non-singleton OTU; significance was pollinated flowers, and receptive vs. pollinated flowers, and b actively vs. assessed by ANOSIM of Jaccard’sindex(basedonpresence/absenceonly) passively pollinated flowers before (receptive) and after (pollinated) visita- and the Morisita index (based on abundance) tion by wasps. Asterisks indicate significant differences in communities

Fourth, communities in gall- and seed flowers did not Sixth, pollinated flowers of actively vs. passively pollinat- differ significantly in overall phylogenetic composition ed figs differed significantly in the phylogenetic structure of (Unifrac analysis of uncollapsed tree containing receptive their fungal assemblages (Table 2). Although communities in and pollinated figs: p00.61 in presence/absence analyses; figs of each type were not mutually exclusive, they differed in p00.99 in weighted analyses). Overall, only one non- relative abundance of several OTU (Supplementary Fig. 2). singleton phylotype was found uniquely in seed flowers The clade containing OTU 14–17 was found only in pollinat- (OTU 1, Basidiomycota). All of the Saccharomycotina phy- ed flowers of actively pollinated species, where it occurred in lotypes found more than once were found in both seed and both gall- and seed flowers. gall flowers (N09; Fig. 2). Most of the non-singleton Pezi- zomycotina were found only in gall flowers, including OTU that were found both before and after pollination (Fig. 2). Discussion However, Pezizomycotina were relatively rare. Fifth, receptive and pollinated flowers differed signifi- We used a culture-free approach to characterize fungal com- cantly in the phylogenetic structure of their fungal assemb- munities associated with pollinating fig wasps and figs in lages (Table 2). Across the entire dataset, 11 OTU were found six species of Ficus at four developmental stages in Panama, only in pollinated flowers (Fig. 2). Seven of these OTU were with special attention to evaluating community structure recovered more than once (OTU 6 and 11, Dothideomycetes; among fig species, between gall- and seed flowers, and as OTU 14–17 and 19, Saccharomycotina). All members of the well-sampled clade containing OTU 14, 15, 16, and 17 Table 3 Fungal communities are more similar between gall and seed (affinity for Candida sp. 564; Fig. 2) were found in syconia flowers in receptive figs than in figs after pollination. Data represent only after pollination (Fig. 2). This clade was recovered only the mean and standard error for all pairwise comparisons among fig from actively pollinated figs and from wasps associated with species at the receptive stage and after pollination (including early and F. popenoei,anactivelypollinatedspecies. late post-pollination stages). N indicating the number of comparisons from which means were computed. p values are based on (1) direct comparisons using nonparametric statistics (pall) as shown below and (2) comparisons based on permutations to equalize sample sizes in Table 2 Fungal community structure in fig syconia differed as a which 16 similarity values were drawn at random from the pollinated function of pollination status and, after pollination, between actively fig data set, and means compared against the observed values from vs. passively pollinated species. Bold indicates p≤0.05. Data are p receptive figs. Bold font highlights significant values values from weighted and unweighted analyses of phylogenetic structure in UniFrac, based on the tree shown in Supplemental Fig. 5 Receptive (N) Pollinated (N) pall prand Unweighted Weighted Jaccard 0.22±0.02 (16) 0.11±0.03 (70) 0.0009a 0.0604b for abundance for abundance Morisita 0.60±0.09 (16) 0.19±0.04 (70) 0.0001c 0.0001d Pollinated vs. receptive flowers 0.0100 0.0880 a X2 011.1, df01 Actively vs. passively pollinated flowers: b t01.9, df0999 Receptive flowers 0.5300 0.8030 c X2 014.6, df01 Pollinated flowers 0.0030 0.0260 d t05.1, df0999

Fungi in Developing Figs

a function of developmental stage. To our knowledge this is the same yeast was present in different life stages of a focal the first examination of fungal communities associated with insect [e.g., 33]. In future work, we will assess whether non-agricultural figs within their natural range and the first members of clades that appear to be closely related based on in a tropical forest. the analyses presented here differ functionally among fig Our survey revealed that fungi were common and taxo- species. nomically diverse in apparently healthy syconia of six spe- Similarly, our surveys revealed that gall- and seed flow- cies of Ficus (Table 1). In contrast to Miller and Phaff’s[59] ers had similar communities before pollination, leading us to conclusion that figs such as Calimyrna have essentially suggest that fungi likely are not the drivers of oviposition sterile internal tissues prior to visitation by pollinators, we choice between these flower types (Table 3 and Supplemen- found that fungi were present both before and after visits by tal Fig. 6). The occurrence of Pezizomycotina preferentially pollinators. Many syconia-associated fungi were found in or in gall flowers is intriguing, but in general these fungi were on fig wasps (Fig. 2, Supplemental Fig. 2). These fungi found at low abundances, such that it is difficult to distin- never were observed in culture-based surveys of the same guish rarity from apparent specificity. At present, we con- fig tissue (data not shown), which may suggest specialized clude that the presence, absence, and composition of fungal nutrient requirements and/or growth conditions. assemblages studied here do not appear to contribute to Fungi associated with syconia and their wasps were “unbeatable seeds” in figs [sensu 88]. distinct at the community level relative to those recorded By decoupling timing from developmental stage, our in foliage and stems of diverse vascular plants in lowland study reveals that fungal communities differ markedly Panama, and in some cases, have been found to date only in among developmental stages of figs, with three results of association with figs. Fig- and wasp-associated fungal com- note. First, communities differed significantly in receptive munities were especially rich in several clades of Saccha- vs. pollinated figs (Fig. 3). These observations are consistent romycotina that are unique relative to sequenced strains with a possible role of fungi in providing volatile cues that available through GenBank (see clades containing OTU indicate a termination of receptivity for pollinators and an 14-23) and have not been recorded previously in aerial plant onset of ripeness for frugivorous seed dispersers. parts. The most common yeasts observed here were present Second, communities in gall- and seed flowers differed in all focal species of Ficus, both flower types, and all pre- following pollination, consistent with the introduction of ripening developmental stages, but we observed marked fungi by female wasps (Table 3). Fig-pollinating wasps live differences at the community level as a function of receptive an average of 2 to 3 days after leaving the natal fig [26, 52], vs. pollinated status and, after pollination, as a function of spending most of that time in the airstream above the forest pollination type (active vs. passive; Figs. 2 and 3). canopy in search of compatible fig species in a receptive Our OTU-level and phylogenetic analyses reveal that state. They do not ingest any plant material during this time fungal communities did not differ significantly among [21]. Thus fungi recovered from pollinating wasps and Ficus species prior to pollination, suggesting that fungal pollinated fig flowers may be transmitted by fig wasps from communities likely do not play a decisive role in host– their natal syconia. We found that wasps frequently carried species identification or selection by foundresses (Fig. 3 fungi that had been found in the syconia of species that were and Table 2). This contrasts with the yeast/cactophilic not their natal species but were present in other Ficus Drosophila system [10], wherein volatiles from individual surveyed. This could reflect undersampling of the fungal yeast species selectively attract particular species of flies. community in the syconia; however, our sampling reached Fig wasps require volatile signals produced by the fig statistical sufficiency for the figs evaluated here. Thus our tree to identify their host species at the correct develop- observations are consistent with pollinator mixing among mental stage [44, 85]. Similar communities of yeasts in closely related figs that share the same pollination syndrome different species of Ficus, which change markedly after pol- [57, 61]andsuggesthost–species generalism of many lination, may provide a non-species-specific indicator to fig yeasts recovered here. wasps that syconia are receptive, perhaps via amplification of Third, fungal communities differ in pollinated flowers the fig’svolatilerepertoire[asinthecaseofnectar-inhabiting of passively and actively pollinated fig species, which yeasts and some foliar endophytes; 23, 34, 46, 69]. The source have been separated by at least 60 million years in the of the pre-pollination fungal community is currently un- New World clades (Fig 3)[56, 71]. This is consistent known. It is possible that fungi of the same genotype or with introduction and maintenance of different fungal OTU differ in functional traits or could be distinguished at communities by wasps of each clade, which may fre- loci that evolve more quickly than this portion of the ribosom- quently host-switch within theactivelyandpassively al repeat [16, 65]. However, previous studies indicate that pollinated clades but not between the two clades [57, variation in the ITS region can distinguish yeast species 61,Machado,unpublished].However,somefungido [e.g., 18] and some have used only LSU data to conclude that occur in figs of both clades; mechanisms by which they

E. O. Martinson et al.

may occur in both, even when pollinators do not mix, 10. Barker JSF (1992) Genetic variation in cactophilic Drosophila remain to be resolved. Notably, some fungi found in for oviposition on natural yeast substrates. Evolution 46:1070– 1083 receptive flowers were allied phylogenetically with the 11. Baumann P, Baumann L et al (1995) Genetics, physiology, and Basidiomycetous yeast Rhodotorula mucilaginosa, which evolutionary relationships of the genus Buchnera—intracellular was cultured from a cultivated fig in Japan (Ficus carica symbionts of aphids. Annu Rev Microbiol 49:55–94 cv. Masui-Dofin and Horaishi) [36]. This suggests a 12. Berg CC (1989) Classification and distribution of Ficus. Experi- entia 45:605–611 geographically and taxonomically broad association with 13. Bronstein JL (1988) Mutualism, antagonism, and the fig–pollina- figs that merits further study. tor interaction. Ecology 69:1298–1302 Our study provides a first evaluation of the diversity and 14. Bronstein JL, Alarcon R et al (2006) The evolution of plant–insect affiliations of microfungi in developing figs and their pollinat- mutualisms. New Phytol 172:412–428 15. Bronstein JL, Barbosa P (2002) Multi-trophic/multi-species mutu- ing wasps. Although strong evidence was not obtained to alistic interactions: the role of non-mutualists in shaping and medi- indicate a fungal role in host identification or oviposition pref- ating mutualisms. In: Tscharntke T, Hawkins BA (eds) Multitrophic erence, the differences observed here for receptive and post- level interactions. Cambridge University Press, Cambridge, pp 44– pollination figs, and figs with different pollination syndromes, 66 16. Cairney JWG (1999) Intraspecific physiological variation: impli- set the stage for exploring their interactions with the iconic fig- cations for understanding functional diversity in ectomycorrhizal and fig–wasp partnership. Increasingly the study of mutualisms fungi. Mycorrhiza 9:125–135 has been expanded to include additional participating members, 17. Cardoza YJ, Teal PEA et al (2003) Effect of peanut plant fungal rather than the bipartite interactions of mutualistic partners infection on oviposition preference by Spodoptera exigua and on host-searching behavior by Cotesia marginiventris. Environ Ento- alone [8, 22, 77]. The classic fig–fig wasp mutualism operates mol 32:970–976 in the context of the microbial associations of each partner, 18. Chen YC, Eisner JD et al (2001) Polymorphic internal transcribed which may play yet unexplored but important roles in spacer region 1 DNA sequences identify medically important this otherwise well-studied association. yeasts. J Clin Microbiol 39:4042–4051 19. Clarke KR, Green RH (1988) Statistical design and analysis for a biological effects study. Mar Ecol Prog Ser 46:213–226 Acknowledgments We gratefully acknowledge the National Science 20. Clay, K. (1988) Fungal endophytes of grasses—a defensive mutu- Foundation for supporting this research (IOB-062492 to AEA and an alism between plants and fungi. Ecology 6910-16 NSF Graduate Research Fellowship to EOM) as well as the Smithsonian 21. Compton SG, Ellwood MDF et al (2000) The flight heights of Institute (Predoctoral Fellowship to EOM). We thank the Smithsonian chalcid wasps (Hymenoptera, Chalcidoidea) in a lowland Bornean Tropical Research Institute for logistical support and the government of rain forest: fig wasps are the high fliers. Biotropica 32:515–522 Panama for permission to carry out this research. We are grateful to J. 22. Currie CR, Scott JA et al (2003) Fungus-growing ants use Hackett and A. Gomez for technical assistance, W. Marussich for collec- antibiotic-producing bacteria to control garden parasites. Nature tion of fig wasps, and J. U’Ren and V. Martinson for helpful discussion. 423:461–461 23. Daisy BH, Strobel GA et al (2002) Naphthalene, an insect repel- lent, is produced by Muscodor vitigenus,anovelendophytic References fungus. Microbiology 148:3737–3741 24. Doster MA, Michailides TJ et al (1996) Aspergillus species and mycotoxins in figs from California orchards. Plant Dis 80:484–489 1. Agrawal AA, Ackerly DD et al (2007) Filling key gaps in popu- 25. Dunn DW, Segar ST et al (2008) A role for parasites in stabilizing lation and community ecology. Front Ecol Environ 5:145–152 the fig–pollinator mutualism. PLoS Biol 6:490–496 2. Altschul SF, Gish W et al (1990) Basic local alignment search tool. 26. Dunn DW, Yu DW et al (2008) Longevity, early emergence and J Mol Biol 215:403–410 body size in a pollinating fig wasp—implications for stability in a 3. Arnold AE, Maynard Z et al (2000) Are tropical fungal endophytes fig–pollinator mutualism. J Anim Ecol 77(5):927–935 hyperdiverse? Ecol Lett 3:267–274 27. Fisher RA, Corbet AS et al (1943) The relation between the 4. Arnold AE, Maynard Z et al (2001) Fungal endophytes in dicoty- number of species and the number of individuals in a random ledonous neotropical trees: patterns of abundance and diversity. sample of an animal population. J Anim Ecol 12:42–58 Mycol Res 105:1502–1507 28. Feldman TS, O’Brien H et al (2008) Moth dispersal of mycopar- 5. Arnold AE, Herre EA (2003) Canopy cover and leaf age affect asites and endophytes associated with Claviceps paspali and the colonization by tropical fungal endophytes: ecological pattern and grass Paspalum (Poaceae). Microb Ecol 56:742–750 process in Theobroma cacao (Malvaceae). Mycologia 95:388–398 29. Gallery RE, Dalling JW et al (2007) Diversity, host affinity, and 6. Arnold AE, Henk DA et al (2007) Diversity and phylogenetic distribution of seed-infecting fungi: a case study with Cecropia. affinities of foliar fungal endophytes in loblolly pine inferred by Ecology 88:582–588 culturing and environmental PCR. Mycologia 99:185–206 30. Ganeshaiah KN, Kathuria P et al (1995) Evolution of style- 7. Arnold AE, Lutzoni F (2007) Diversity and host range of foliar length variability in figs and optimization of ovipositor length fungal endophytes: are tropical leaves biodiversity hotspots? Ecol- in their pollinator wasps—acoevolutionarymodel.JGenet ogy 88:541–549 74:25–39 8. Arnold AE, Miadlikowska J et al (2009) A phylogenetic estimation of 31. Gardes M, Bruns TD (1993) ITS primers with enhanced specificity trophic transition networks for ascomycetous fungi: are lichens cradles for basidomycetes: application to the identification of mycorrhizae of symbiotrophic fungal diversification? Syst Biol 58:283–297 and rusts. Mol Ecol 2:113–118 9. Barash I, Manulis-Sasson S (2009) Recent evolution of bacterial 32. Gehring C, Bennett A (2009) Mycorrhizal fungal–plant–insect pathogens: the gall-forming Pantoea agglomerans case. Annu Rev interactions: the importance of a community approach. Environ Phytopathol 47:133–152 Entomol 38:93–102

Fungi in Developing Figs

33. Gibson C, Hunter M (2009) Inherited fungal and bacterial endo- 57. Machado CA, Robbins N et al (2005) Critical review of host symbionts of a parasitic wasp and its cockroach host. Microb Ecol specificity and its coevolutionary implications in the fig/fig wasp 57:542–549 mutualism. Proc Natl Acad Sci U S A 102:6558–6565 34. Goodrich KR, Zjhra ML et al (2006) When flowers smell 58. Maddison WP, Maddison DR (2009) Mesquite: a modular system fermented: the chemistry and ontogeny of yeasty floral scent for evolutionary analysis in pawpaw (Asimina triloba:Annonaceae).IntJPlantSci 59. Miller MW, Phaff HJ (1962) Successive microbial populations in 167:33–46 Calimyrna figs. Appl Microbiol 10(5):394–400 35. Grison-Pige L, Hossaert-McKey M et al (2002) Fig volatile com- 60. Mrak EM, Phaff HJ et al (1942) Yeasts occurring in souring figs. J pounds—a first comparative study. Phytochemistry 61:61–71 Bacteriol 44:441–450 36. Hamanaka D, Norimura N et al (2010) Surface decontamination of 61. Marussich WA, Machado CA (2007) Host-specificity and coevo- fig fruit by combination of infrared radiation heating with ultravi- lution among pollinating and nonpollinating New World fig wasps. olet irradiation. Food Control 22:375–380 Mol Ecol 16:1925–1946 37. Hammer O, Harper DAT et al (2001) PAST: Paleontological Sta- 62. Mooney KA, Mandal K (2010) Competition hierarchies among tistics software package for education and data analysis. Palaeontol ants and predation by birds jointly determine the strength of multi- Electron 4:1–9 species ant–aphid mutualisms. Oikos 119:874–882 38. Herre EA (1993) Population structure and the evolution of viru- 63. Oliver KM, Degnan PH et al (2009) Bacteriophages encode factors lence in nematode parasites of fig wasps. Science 259:1442–1445 required for protection in a symbiotic mutualism. Science 325:992– 39. Herre EA (1995) Factors affecting the evolution of virulence: 994 nematode parasites of fig wasps as a case study. Parasitology 64. Omacini M, Chaneton EJ et al (2001) Symbiotic fungal endo- 111:S179–S191 phytes control insect host–parasite interaction webs. Nature 40. Herre EA, Knowlton N et al (1999) The evolution of mutualisms: 409:78–81 exploring the paths between conflict and cooperation. Trends Ecol 65. Parlade J, Hortal S et al (2011) Intraspecific variability of Lactar- Evol 14:49–53 ius deliciosus isolates: colonization ability and survival after cold 41. Herre EA, West SA (1997) Conflict of interest in a mutualism: storage. Mycorrhiza 21:393–401 documenting the elusive fig wasp–seed trade-off. Proc R Soc Lond 66. Pellmyr O, Huth CJ (1994) Evolutionary stability of mutualism B Biol Sci 264:1501–1507 between yuccas and yucca moths. Nature 372:257–260 42. Higgins KL, Coley PD et al (2011) Culturing and direct PCR 67. Pitzschke A, Hirt H (2010) New insights into an old story: Agro- suggest prevalent host generalism among diverse fungal endo- bacterium-induced tumour formation in plants by plant transfor- phytes of tropical forest grasses. Mycologia 103:247–260 mation. EMBO J 29:1021–1032 43. Hoffman MT, Arnold AE (2010) Diverse bacteria inhabit living 68. Posada D (2006) ModelTest Server: a web-based tool for the hyphae of phylogenetically diverse fungal endophytes. Appl En- statistical selection of models of nucleotide substitution online. viron Microbiol 76:4063–4075 Nucleic Acids Res 34:W700–W703 44. Hossaert-Mckey M, Gibernau M et al (1994) Chemosensory at- 69. Raguso RA (2004) Why are some floral nectars scented? Ecology traction of fig wasps to substances produced by receptive figs. 85:1486–1494 Entomol Exp Appl 70:185–191 70. Rohfritsch O (2008) Plants, gall midges, and fungi: a three- 45. Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian infer- component system. Entomol Exp Appl 128:208–216 ence of phylogenetic trees. Bioinformatics 17:754–755 71. Rønsted N, Weiblen G et al (2005) 60 million years of co- 46. Jallow MFA, Dugassa-Gobena D et al (2008) Influence of an divergence in the fig–wasp symbiosis. Proc R Soc Lond B Biol endophytic fungus on host plant selection by a polyphagous moth Sci 272:2593–2599 via volatile spectrum changes. Arthropod-Plant Interactions 2:53– 72. Rudgers JA, Gardener MC (2004) Extrafloral nectar as a re- 62 source mediating multispecies interactions. Ecology 85:1495– 47. Jander KC, Herre EA (2010) Host sanctions and pollinator cheat- 1502 ing in the fig tree–fig wasp mutualism. Proc R Soc Lond B Biol Sci 73. Sanchez F, Korine C et al (2006) Ethanol and methanol as possible 277:1481–1488 odor cues for Egyptian fruit bats (Rousettus aegyptiacus). J Chem 48. Janzen DH (1966) Coevolution of mutualism between ants and Ecol 32:1289–1300 Acacias in Central America. Evolution 20:249 74. Stamatakis A (2006) RAxML-VI-HPC: Maximum likelihood- 49. Janzen DH (1979) How to be a fig. Annu Rev Ecol Syst 10:13–51 based phylogenetic analyses with thousands of taxa and mixed 50. Jousselin E, Hossaert-McKey M et al (2003) Why do fig wasps models. Bioinformatics 22:2688–2690 actively pollinate monoecious figs? Oecologia 134:381–387 75. Suryanarayanan TS, Vijaykrishna D (2001) Fungal endophytes of 51. Katoh K, Misawa K et al (2002) MAFFT: a novel method for rapid aerial roots of Ficus benghalensis. Fungal Divers 8:155–161 multiple sequence alignment based on fast Fourier transform. 76. U'Ren JM, Dalling JW et al (2009) Diversity and evolutionary Nucleic Acids Res 30:3059–3066 origins of fungi associated with seeds of a neotropical pioneer tree: 52. Kjellberg F, Doumesche B et al (1988) Longevity of a fig wasp a case study for analysing fungal environmental samples. Mycol (Blastophaga-psenes). Proc K Ned Akad Wet C 91:117–122 Res 113:432–449 53. Leigh EG, Rand AS et al (1996) The ecology of a tropical forest: 77. U'Ren JM, Lutzoni F et al (2010) Community analysis reveals seasonal rhythms and long term changes. Smithsonian Institute, close affinities between endophytic and endolichenic fungi in Washington mosses and lichens. Microb Ecol 60:340–353 54. Lopez-Vaamonde C, Winkström N et al (2009) Molecular dating 78. Van Bael SA, Fernandez-Marin H et al (2009) Two fungal symbi- and biogeography of fig-pollinating wasps. Mol Phylogenet Evol oses collide: endophytic fungi are not welcome in leaf-cutting ant 52:715–726 gardens. Proc R Soc Lond B Biol Sci 276:2419–2426 55. Lozupone C, Knight R (2005) UniFrac: a new phylogenetic meth- 79. van Noort S, Ware AB et al (1989) Pollinator-specific volatile od for comparing microbial communities. Appl Environ Microbiol attractants released from the figs of Ficus Burtt-davyi. S Afr J 71:8228–8235 Sci 85:323–324 56. Machado CA, Jousselin E et al (2001) Phylogenetic relationships, 80. Verkerke W (1986) Anatomy of Ficus ottoniifolia (Moraceae) historical biogeography and character evolution of fig-pollinating syconia and its role in the fig–fig wasp symbiosis. Proc K Ned wasps. Proc R Soc Lond B Biol Sci 268:685–694 Akad Wet 89:443–469

E. O. Martinson et al.

81. Vidal S (1996) Changes in suitability of tomato for whiteflies 85. Ware AB, Kaye PT et al (1993) Fig volatiles—their role in attract- mediated by a non-pathogenic endophytic fungus. Entomol Exp ing pollinators and maintaining pollinator specificity. Plant Syst Appl 80:272–274 Evol 186:147–156 82. Vilgalys R, Hester M (1990) Rapid genetic identification and 86. Way MJ (1963) Mutualism between ants and honeydew-producing mapping of enzymatically amplified ribosomal DNA from several Homoptera. Annu Rev Entomol 8:307–344 Cryptococcus species. J Bacteriol 172:4238–4246 87. Weiblen GD (2002) How to be a fig wasp. Annu Rev Entomol 83. Wang HK, Hyde KD et al (2008) Fungal diversity on fallen leaves 47:299–330 of Ficus in northern Thailand. J Zhejiang Univ Sci B 9:835–841 88. West SA, Herre EA (1994) The ecology of the New-World fig- 84. Wang GQ, Wei SG et al (2009) Six new eriophyoid mites (Acari: parasitizing wasps Idarnes and implications for the evolution of Eriophyoidea) associated with Ficus spp. (Moraceae) from China. the fig–pollinator mutualism. Proc R Soc Lond B Biol Sci 258:67– Zootaxa 2201:49–62 72

3 4 1 2

4.32 km

Supplemental Figure 1 - Locations of trees sampled in Barro Colorado National

Monument, Panama. 1 F. insipida, 2 F. maxima, 3 F. triangle, 4 F. costaricana, 5 F. popenoei, and 6 F. obtusifolia. Physically proximate trees harbored communities that were actually significantly less similar (comparison among trees 1–4; averaged

JGR = 0.6891) than trees at further distances (trees 5–6 compared to trees 1–4; averaged

JGR = 0.9804)

Leucosporidium antarcticum AF444529 Rhodotorula glutinis FJ345357 7 1 Rhodotorula mucilaginosa AB1931 75 100 -- 75 97 Rhodotorula sp SY100 AB026010 100 -- 99 F. insipida Recep Seed 3 -- 91 F. obtusifolia Early Seed 2-5 Cryptococcus victoriae AM160647 100 100 Cryptococcus tephrensis DQ000318 F. triangle Early Seed 1 Taphrina deformans DQ470973 Schizosaccharomyces pombe Z19136 7 5 -- F. obtusifolia Recep Gall 6-2 F. obtusifolia Recep Gall 2-2 100 9 8 100 100 Fusarium brachygibbosum GQ505450 100 100 F. triangle Early Gall 14 100 100 -- F. obtusifolia Early Seed 4-2 99 100 F. obtusifolia Recep Gall 5 100 Coniozyma leucospermi EU552113

7 2 8 5 F. insipida Recep Seed 1 95 90 F. obtusifolia Recep Gall 8 9 4 100 Phoma medicaginis EU167575 100 100 F. insipida Early Seed 3

9 9 F. insipida Recep Gall 9-3 100 F. popenoei Recep Gall 8

100 Bipolaris spicifera GU183125 100 100 F. triangle Early Gall 6-3 100 Alternaria tenuissima AY154712 100 100 F. obtusifolia Recep Gall 1 F. insipida Early Seed 1-2 100 F. insipida Early Seed 8-3 100 7 4 -- F. maxima Early Seed 1 -- 100 F. obtusifolia Early Seed 3 F. insipida Recep Gall 10 7 0 9 8 -- F. obtusifolia Late Seed 8 100 Saccharomyces cerevisiae EU649673 Ogataea cecidiorum FJ897742 Debaryomyces hansenii FN667855 Candida quercitrusa AM160627 -- 9 8 9 3 100 Candida sp HA 1671 AM160629

100 Candida tropicalis FJ515227 100 Lodderomyces elongisporus FJ515234 Williopsis saturnus FN868262 Pichia sp UFMG EU580140 8 3 Pichia anomala FJ865436 99 100 F. costaricana Late Seed 6 100 F. costaricana Late Gall 2-2 F. triangle Late Gall 10 100 100 F. triangle Late Gall 6 9 5 F. triangle Late Seed 5-2 100 Candida sp A564 HM364288 F. costaricana Late Seed 1-5 9 9 9 9 F. costaricana Late Gall 1 9 0 100 -- F. costaricana Late Gall 6 100 F. obtusifolia Late Gall 12-8 100 F. popenoei Early Seed 11-2 100 100 100 F. popenoei Early Gall 11-4 F. popenoei Early Seed 10-3 F. popenoei Early Gall 2-2 Candida picinguabensis DQ377638 Metschnikowia agaves HM189300

9 0 Candida gelsemii DQ988045 9 9 Metschnikowia bicuspidata EU809439 Candida picachoensis FJ614653 9 3 100 9 9 Unidentified Metschnikowia AM161112 -- 9 8 100 Metschnikowia sp XY201 DQ367882 -- 9 3 Metschnikowia sp XY103 DQ367881 F. insipida Recep Seed 5-3 100 100 F. insipida Late Gall 11-8 F. insipida Late Seed 4-7 F. insipida Late Seed 5-2 F. insipida Late Gall 1-4 Candida flosculorum EF137918 9 2 100 Candida akabanensis EU100744 Candida sp 1A EF565862 Candida sp C107DX-Y2 FJ865216

7 7 Clavispora lusitaniae AY174089 9 1 Candida oregonensis FN667841

100 F. costaricana Late Gall 3-3 100 F. costaricana Late Seed 13-4

100 F. maxima Recep Seed 6 100 F. maxima Recep Gall 1 9 8 100 F. triangle Late Gall 1 100 F. triangle Late Seed 1 100 F. triangle Early Seed 3-11 100 9 9 100 F. popenoei Recep Gall 1 100 F. triangle Early Gall 12 F. maxima Recep Gall 3-2 F. triangle Early Gall 15-6 F. triangle Ripe Gall 1-11 100 100 F. triangle Late Seed 8-4 F. triangle Late Gall 3-4 F. triangle Late Gall 8-4 F. costaricana Late Seed 8-2 F. triangle Recep Seed 2 F. insipida Recep Gall 11 2 9 9 F. popenoei Recep Seed 2 100 7 9 F. popenoei Recep Gall 4 100 F. insipida Recep Seed 7-2 F. obtusifolia Late Gall 15-8 F. insipida Early Seed 2 9 1 100 F. popenoei Early Seed 6 F. obtusifolia Late Seed 1-3 F. maxima Recep Gall 5 F. maxima Recep Seed 8-6 F. insipida Early Gall 3-10 8 5 9 0 F. insipida Late Seed 2-3 F. obtusifolia Early Seed 12-5 7 5 100 F. triangle Early Gall 16-2 9 3 100 F. obtusifolia Late Seed 2-3 F. insipida Recep Gall 7-6 F. triangle Early Gall 4-3 8 7 100 F. popenoei Recep Gall 5-5 100 100 F. triangle Recep Seed 1-4 7 9 100 F. popenoei Recep Seed 4-3 F. maxima Recep Gall 6-5 F. costaricana Late Seed 9 F. maxima Recep Seed 1 9 6 9 1 F. insipida Recep Gall 6 F. insipida Early Seed 6 0.3 49

Supplemental Figure 2 - Most likely tree resulting from RAxML analyses of the 5.8S and partial LSU, including results of maximum likelihood bootstrap (≥70 %; above branches) and Bayesian posterior probabilities (≥90 %, below branches), revealing the phylogenetic placement of fungal OTU recovered from figs in six species of Ficus in Panama. Tree was used for UniFrac analyses; clades are not collapsed, such that the position of each sequence can be examined. Leucosporidium antarcticum was included as an outgroup. Black bars indicate OTU based on 95 % sequence similarity

16 F. insipida 16 F. maxima Observed richness 12 12 95% confidence intervals Bootstrap estimate 8 8

4 4

0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 16 F. costaricana 16 F. popenoei

12 12

8 8 OTU

4 4

0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60

16 F. obtusifolia 16 F. triangle

12 12

8 8

4 4

0 0 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Sequences

Supplemental Figure 3 - OTU accumulation, 95 % confidence intervals, and bootstrap estimates of richness based on ITS-LSU OTU (95 % sequence similarity) for fungal communities sampled from each of six Ficus species

12 Pollinating Fig Wasp Survey

6 OTU

Observed richness 2 95% confidence intervals Bootstrap estimate 0 0 10 20 30 40 50 60 70 80 Sequences

Supplemental Figure 4 - OTU accumulation for fungi recovered from pollinating fig wasps, 95 % confidence intervals, and bootstrap estimates of richness based on the survey of six species of wasps in lowland Panama

Leucosporidium antarcticum AF444529 9 5 Rhodotorula glutinis FJ345357 Rhodotorula mucilaginosa AB193175 Rhodotorula sp SY100 AB026010 8 3 F. obtusifolia Early Seed 2-5 F. insipida Recep Seed 3 100 Cryptococcus tephrensis DQ000318 9 7 Cryptococcus victoriae AM160647 F. triangle Early Seed 1 Taphrina deformans DQ470973 Schizosaccharomyces pombe Z19136 F. obtusifolia Recep Gall 6-2 100 F. obtusifolia RecepvGall 2-2 100 F. triangle Early Gall 14 100 Fusarium brachygibbosum GQ505450 100 Coniozyma leucospermi EU552113 100 F. obtusifolia Recep Gall 5 Wasp F. triangle 13 100 Wasp F. triangle 12 Wasp F. triangle 1 Wasp F. triangle 11 F. obtusifolia Early Seed 4-2 Wasp F. triangle 14 Wasp F. triangle 3 Wasp F. triangle 2 7 0 F. insipida Recep Seed 1 9 3 Phoma medicaginis EU167575 100 F. obtusifolia Recep Gall 8 F. insipida Early Seed 3 100 F. insipida Recep Gall 9-3 F. popenoei Recep Gall 8 100 F. triangle Early Gall 6-3 100 Bipolaris spicifera GU183125 Alternaria tenuissima AY154712 100 F. obtusifolia Recep Gall 1 100 Saccharomyces cerevisiae EU649673 F. obtusifolia Late Seed 8 F. insipida Recep Gall 10 F. obtusifolia Early Seed 3 F. maxima Early Seed 1 8 4 9 5 F. insipida Early Seed 1-2 F. insipida Early Seed 8-3 Ogataea cecidiorum FJ897742 Debaryomyces hansenii FN667855 8 1 9 9 Candida quercitrus AM160627 Candida sp HA 1671 AM160629 100 Lodderomyces elongisporus FJ515234 Candida tropicalis FJ515227 Williopsis saturnus FN868262 Pichia anomala FJ865436 8 0 Pichia sp UFMG EU580140 Candida sp A564 HM364288 9 9 F. costaricana Late Seed 1-5 8 8 F. costaricana Late Gall 6 100 F. costaricana Late Gall 1 100 F. obtusifolia Late Gall 12-8 F. popenoei Early Seed 11-2 F. popenoei Early Gall 11-4 F. popenoei Early Gall 2-2 F. popenoei Early Seed 10-3 100 F. costaricana Late Gall 2-2 F. costaricana Late Seed 6 100 Wasp F. popenoei 3 Wasp F. popenoei 15 F. triangle Late Gall 10 Wasp F. popenoei 11 9 8 F. triangle Late Gall 6 Wasp F. popenoei 7 8 7 Wasp F. popenoei 8 F. triangle Late Seed 5-2 Wasp F. popenoei 9 Wasp F. popenoei 10 Wasp F. popenoei 13 Wasp F. popenoei 4 Wasp F. popenoei 5 Wasp F. popenoei 1 Wasp F. popenoei 14 9 0 Candida flosculorum EF137918 Candida sp 1A EF565862 Candida akabanensis EU100744 Candida picinguabensis DQ377638 Metschnikowia agaves HM189300 8 6 Candida gelsemii DQ988045 Metschnikowia bicuspidata EU809439 8 9 Candida picachoensis FJ614653 9 1 Unidentified Metschnikowia AM161112 Metschnikowia sp XY201 DQ367882 Metschnikowia sp XY103 DQ367881 Wasp F. insipida 5 Wasp F. insipida 14 9 2 Wasp F. insipida 1 100 Wasp F. insipdia 8 F. insipida Late Seed 5-2 F. insipida Late Gall 1-4 F. insipida Late Seed 4-7 F. insipida Late Gall 11-8 Wasp F. insipida 3 Wasp F. insipida 7 100 7 5 Wasp F. insipida 9 Wasp F. insipida 2 7 4 Wasp F. insipida 13 Wasp F. insipida 10 F. insipida Recep Seed 5-3 Wasp F. insipida 4 Wasp F. insipida 11 7 4 Wasp F. maxima 8 Wasp F. maxima 2 Wasp F. maxima 14 Wasp F. maxima 7 Wasp F. maxima 4 Wasp F. obtusifolia 10 Wasp F. maxima 6 Wasp F. obtusifolia 7 Wasp F. maxima 9 Wasp F. maxima 12 Wasp F. maxima 15 Wasp F. maxima 13 Wasp F. maxima 1 Wasp F. popenoei 6 Wasp F. maxima 5 Wasp F. obtusifolia 14 Wasp F. maxima 3 Wasp F. maxima 11 Wasp F. maxima 10 Candida sp C107DX-Y2 FJ865216 Candida oregonensis FN667841 Clavispora lusitaniae AY174089 Wasp F. popenoei 12 9 4 9 6 F. costaricana Late Seed 13-4 F. costaricana Late Gall 3-3 Wasp F. costaricana 9 100 F. maxima Recep Gall 1 F. maxima Recep Seed 6 Wasp F. costaricana 8 9 5 F. triangle Early Gall 4-3 Wasp F. costaricana 1 Wasp F. costaricana 3 8 4 Wasp F. costaricana 11 F. popenoei Recep Seed 4-3 F. popenoei Recep Gall 5-5 F. triangle Recep Seed 1-4 Wasp F. costaricana 14 Wasp F. costaricana 7 Wasp F. costaricana 5 Wasp F. costaricana 13 Wasp F. costaricana 12 F. costaricana Late Seed 9 9 8 F. maxima Recep Seed 1 F. insipida Early Seed 6 F. insipida Recep Gall 6 F. maxima Recep Gall 6-5 Wasp F. obtusifolia 1 F. insipida Recep Gall 7-6 F. obtusifolia Early Seed 12-5 7 4 F. triangle Early Gall 16-2 8 6 F. obtusifolia Late Seed 2-3 F. maxima Recep Seed 8-6 F. insipida Late Seed 2-3 7 5 F. insipida Early Gall 3-10 Wasp F. obtusifolia 4 8 1 F. obtusifolia Late Gall 15-8 F. maxima Recep Gall 5 F. insipida Early Seed 2 7 4 F. popenoei Early Seed 6 F. obtusifolia Late Seed 1-3 Wasp F. obtusifolia 12 Wasp F. obtusifolia 8 F. triangle Recep Seed 2 F. insipida Recep Gall 11-2 F. insipida Recep Seed 7-2 F. popenoei Recep Seed 2 F. popenoei Recep Gall 4 Wasp F. obtusifolia 15 F. costaricana Late Seed 8-2 Wasp F. obtusifolia 5 Wasp F. obtusifolia 6 Wasp F. obtusifolia 2 Wasp F. obtusifolia 11 Wasp F. obtusifolia 9 Wasp F. obtusifolia 3 8 1 F. triangle Late Gall 1 9 8 Wasp F. triangle 5 9 9 Wasp F. triangle 9 Wasp F. triangle 7 Wasp F. triangle 10 Wasp F. triangle 8 9 9 Wasp F. triangle 6 Wasp F. triangle 15 F. triangle Late Seed 1 F. triangle Early Seed 3-11 F. popenoei Recep Gall 1 F. triangle Early Gall 12 F. maxima Recep Gall 3-2 F. triangle Late Gall 8-4 F. triangle Ripe Gall 1-11 F. triangle Late Gall 3-4 F. triangle Early Gall 15-6 F. triangle Late Seed 8-4 Wasp F. triangle 4 0.2 53

Supplemental Figure 5 Most likely tree resulting from RAxML analyses of the 5.8S and partial LSU, including results of maximum likelihood bootstrap (≥70 %; above branches), revealing the phylogenetic placement of fungal OTU recovered from figs in six species of Ficus and their associated pollinating fig wasps in Panama. L. antarcticum was included as an outgroup. Black bars indicate OTU based on 95 % sequence similarity for Ficus sequences and red bars indicate sequences from their pollinating fig wasps

ANOSIM p = 0.5552 Coordinate 2 Coordinate

Seed Flowers Gall Flowers

Coordinate 1

Supplemental Figure 6 Non-metric multidimensional scaling (NMDS) analysis of gall- and seed-flower fungal communities. Additional analysis shown in upper right corner was assessed by ANOSIM with the Morisita index based on non-singleton

OTU

APPENDIX B:

RATIOS OF POISON SAC AREA:EGGS SUGGEST A SIMILAR

IMPORTANCE OF THE POISON SAC IN OVIPOSITION BY

POLLINATING AND NON-POLLINATING FIG WASPS

To be published in Acta Oecologica.

Ratios of poison sac area:eggs suggest a similar importance of the

poison sac in oviposition by pollinating and non-pollinating fig wasps

Ellen O. Martinson1*, Carlos A. Machado2, E. Allen Herre3

1: Department of Ecology and Evolutionary Biology, The University of Arizona, Tucson,

AZ 85721 USA

2: Department of Biology, The University of Maryland, College Park, MD 20742 USA

3: Smithsonian Tropical Research Institute, Balboa, Ancon, Republic of Panama

Abstract. In the context of the fig and fig wasp mutualism, pollinating fig wasps

(Agaonidae) compete with phylogenetically diverse, parasitic, non-pollinating wasps for oviposition sites within fig syconia. Oviposition by both types of wasps produces galls identical in size and structure for the development of their offspring. Fig trees sanction their pollinating species of wasps if they gall without pollinating, but sanctions against species of non-pollinating wasps have not yet been identified. It has been proposed that pollinating and non-pollinating wasps use the same mechanism to induce gall formation, such that the tree cannot differentiate between the two types of wasps. In pollinating wasps, a maternal secretion stored in the poison sac apparently plays a critical role in gall production; however, this has never been studied in non-pollinating wasps. We posit that maternal secretion in the poison sac is important for gall development in both pollinating and non-pollinating wasps. If products of the poison sac are used similarly to initiate galls at oviposition, we anticipate that galling wasps (regardless of their roles in pollination) should have a similar relationship of poison sac size to egg number. To evaluate this prediction, we measured egg number and poison sac area as a function of body size in diverse pollinating and non-pollinating wasps affiliated with non-domesticated Ficus spp. in Panama. Even though relative area of the poison sac differs among species and between pollinators and non-pollinators, phylogenetically independent contrasts reveal a strong positive correlation between poison sac area and egg number that is consistent across pollinating and non-pollinating fig wasps. This observation is consistent with the 58 hypothesis that both types of wasp use a similar galling mechanism that prevents the fig tree from differentiating between mutualistic and parasitic associates, and sets the stage for evaluating the subsequent hypothesis that contents of the poison sac are similar in each group.

Keywords: Ficus, fig wasps, galling, poison sac, egg load, mutualism, pollination, sanctions 59

Introduction

Over the last 60-80 million years, fig trees (Ficus, Moraceae) and the fig wasps that pollinate their flowers (Chalcidoidea, Hymenoptera) have formed one of the most interdependent insect-plant mutualisms known (Machado 2001; Machado 2005; Rønsted

2005). When a female pollinating fig wasp enters a receptive fig, she is surrounded by hundreds of flowers in the specialized inflorescence known as a syconium (Janzen 1979).

In some flowers she deposits an egg then several drops of a maternal secretion. These flowers develop into galls in which wasp larvae develop and feed (Verkerke 1986). Other flowers that only receive pollen produce a viable fig seed (Janzen 1979). Some species of wasps pollinate actively, storing pollen in specialized pockets and fertilizing flowers individually (Jousselin et al. 2003). However, earlier arising lineages of pollinating wasps do so passively, with pollen brushing off their bodies during oviposition (Jousselin et al.

2003).

Several factors suggest that the poison sac of pollinating wasps plays an important role in galling in the fig/fig wasp system. The maternal secretion, stored in the poison sac, is placed only in flowers in which an egg has been laid (Verkerke 1986; Jansen-González

2012). Nucellar epidermal cells of the fig flower develop rapidly after the maternal secretion has been deposited (Verkerke 1988). This growth occurs before the egg hatches or the pollen tube reaches the fig ovule and is attributed solely to the secretion (Verkerke

1986). However, the mechanism by which this secretion initiates growth is unknown.

Similarly, the size of the poison sac in females of the pollinating fig wasps also suggests its importance. Slightly smaller than the ovaries, the poison sac is the second largest 60 organ within the fig wasp. It is connected directly to the ovipositor and used exclusively during oviposition (Grandi 1929). Pollinating fig wasps are long-distance dispersers with a short lifespan (3-5 days), such that maintaining a large, fluid-filled sac might be costly unless its contents were vital for fitness.

A variety of parasitic, non-pollinating wasps also use the syconium for reproduction by ovipositing through the fig wall from the outside (Boucek 1993). At one time, pollinating and non-pollinating wasps both were assigned to the family Agaonidae; however, this grouping was found to be paraphyletic and the six subfamilies of parasitic, non-pollinating wasps were reassigned to diverse families within the Chalcidoidea

(Rasplus et al. 1998). Some of these non-pollinating fig wasps produce galls that appear identical to those produced by pollinating wasps and utilize the same type of flower tissue

(Pereira and do Prado 2005). Non-pollinating wasps therefore impose a cost on the fig tree not only by initiating gall growth without providing pollination services, but also by directly competing with the pollinating wasps for oviposition sites (West et al. 1996). Fig trees sanction their pollinating species of wasps if they do not pollinate, aborting whole figs or reducing the number of successful wasp offspring per fig (Jander and Herre 2010).

However, sanctions against non-pollinating species of wasps have not yet been identified

(West and Herre 1994). One hypothesis is that pollinating and non-pollinating wasps use the same mechanism to induce gall formation in figs, such that the tree cannot differentiate between the two types of wasps.

We posit that the maternal secretion in the poison sac is important for gall induction by pollinating and non-pollinating wasps associated with figs. Under this 61 scenario, we propose that diverse species of fig-galling wasps, regardless of their roles as pollinators or non-pollinators, should display similar investment in the poison sac relative to the number of eggs produced.

To evaluate this prediction, we measured poison sac area as a function of body size in diverse pollinating and non-pollinating wasps affiliated with non-domesticated

Ficus spp. in Panama. Even though size of the poison sac differs among species and between pollinators and non-pollinators, phylogenetically independent contrasts reveal a strong positive correlation between poison sac area and egg number that is consistent across pollinating and non-pollinating fig wasps. This observation is consistent with the hypothesis that both types of wasp use a similar galling mechanism that prevents the fig tree from differentiating between mutualistic and parasitic associates, and sets the stage for evaluating the subsequent hypothesis that contents of the poison sac are similar in each group.

Methods

Mature figs representing seven species of Ficus, including both actively and passively pollinated species, were collected at the Barro Colorado National Monument, Panama in

January-April 2010. Fruits were collected after male wasps had emerged from their galls and started to mate with female wasps. Within 5 hours of collection, fruits were opened in the lab and placed in sealed Petri dishes with 2-3 figs from conspecific trees, allowing female wasps to emerge. Seven species of pollinating wasps (Pegoscapus spp., 5 species; and Tetrapus spp., 2 species) and two species of non-pollinating wasps (Idarnes sp. and 62

Critogaster sp.) were obtained (Table 1).

Measurements

A total of 19-28 wasps per fig species and pollinator type was examined from 2-6 syconia collected from the same tree. Emerging wasps were dissected with microforceps under a dissecting microscope. Mesothorax length (hereafter thorax length), head width, poison sac area, and number of eggs were measured for each individual. Thorax length was measured with the wasp on its side along a line through the coxae. Poison sac size was standardized by body size by dividing the area of the sac by the thorax length of the wasp. Ovaries were removed, allowed to soak in water for two minutes, and gently spread with a cover slip. Eggs in each ovary were photographed and counted under a compound light microscope, and the number of eggs was averaged for the two ovaries per individual.

Species of wasp were grouped as actively pollinating, passively pollinating, or non-pollinating (Table 1). Mean area of the poison sac (relative to body size) was compared among active pollinators, passive pollinators, and non-pollinating wasps by

ANOVA. Mean poison sac area for each species was then compared using a Tukey-

Kramer HSD test. A simple linear regression was used to determine the relationship between poison sac area and average egg number per ovary for each species of wasp. All analyses were performed in JMP v. 8.0.1. (www.jmp.com).

Independent Contrasts 63

Phylogenetic relationships among the sampled wasps were inferred from a 1.5 KB sequence of the cytochrome oxidase subunit I (COI) gene (Machado 2001; Molbo et al.

2003; Marussich and Machado 2007). Sequences were aligned with MUSCLE (Edgar

2004) with further manual adjustments in MacClade (Maddison and Maddison 2005).

Phylogenetic relationships for members of each class were inferred using Bayesian

Metropolis-coupled Markov chain Monte Carlo (MCMCMC) analyses. Modeltest 3.7

(Posada 2006) was used to select the GTR+I+G model using the Akaike information criterion. Analyses were executed in MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001) for two runs of up to 4.6 million generations each, sampling every 1,000th generation.

Likelihoods converged to a stable range for each data set, and all trees prior to convergence were discarded as burn-in (Supplemental figure 1). Independent contrasts were determined using molecular branch lengths in the PDAP:PDTREE module of

Mesquite (Midford et al. 2005; Maddison and Maddison 2009).

Results and Discussion

The maternal secretion of pollinating fig wasps is stored in the poison sac, which is common throughout Hymenoptera and varied in its function. For example, in honeybees

(Apis mellifera) the poison sac holds the venom associated with painful stings (Haberman

1972). In ants it holds communication pheromones (Holldobl 1971; Vandermeer et al.

1980), and in ichneumonid wasps it stores polydnaviruses that are injected into prey to compromise their immune system, thus protecting the wasp’s eggs (Espagne et al. 2004).

A previous study in cynipoid wasps showed that galling wasps have significantly larger 64 poison sacs than related parasitoid wasps, suggesting that the size of the poison sac may be relevant to its function (Vårdal 2006). Although the function of the poison sac in pollinating and non-pollinating fig wasps is currently unknown, its size and role in oviposition suggests an important function for pollinating wasps (Grandi 1929; Verkerke

1986). Poison sacs in non-pollinating wasps that parasitize figs have not been studied previously.

We found that the poison sac, corrected for differences in wasp body size, differed significantly in size among functional groups based on pollination activity

(active pollinator, passive pollinator, and non-pollinator) (Figure 1). Actively pollinating species generally had the largest poison sacs corrected for body size by thorax length, with the exception of P. gemellus from Ficus popenoei (Figure 1). Non-pollinating wasps had significantly smaller poison sacs compared to actively or passively pollinating wasps

(t89 = 11.09, p < 0.001 and t82 = 7.12, p < 0.001 respectively) (Figure 1).

The significantly smaller poison sac of non-pollinating wasps relative to pollinators is consistent with the reproductive biology of pollinators and non-pollinators.

Pollinating fig wasps live for ca. 3 days and do not eat anything as adults, depending solely on the nutrition gained as larva (Compton et al. 2000). They lay their eggs in a single syconium within a few hours; therefore, a female wasp's lifetime supply of eggs and maternal secretion is present when they emerge from their natal fig as they need to be prepared to ovipostit in the first receptive fig they find. In contrast, non-pollinating species of wasps live for ca. 2 weeks and can oviposit in multiple syconia (Compton

1994; West and Herre 1994), such that they produce eggs and maternal secretion for a 65 much longer period. Therefore, the number of developed eggs should be taken into account with the measurement of the poison sac to determine its importance in oviposition for the available eggs.

The average number of eggs per ovary ranged from 28.2 (a non-pollinator of F. popenoei, Idarnes sp.) to 116.2 eggs (a passive pollinator of F. maxima, Tetrapus americanus) (Table 1). Strikingly, egg number was strongly and positively correlated with poison sac size across species of actively pollinating, passively pollinating, and non- pollinating wasps (R2 = 0.806, p = 0.0006) (Figure 2). If phylogeny is taken into account through independent contrasts, the positive relationship between poison sac size and average egg number remains significant (p = 0.001).

Pollinating fig wasps compete with certain non-pollinating wasps for oviposition sites within a receptive syconium (West and Herre 1994). The galls produced by many pollinators and non-pollinators of figs are indistinguishable in size and structure (Pereira and do Prado 2005). Although it has been presumed that the maternal secretion in pollinating wasps plays a role in gall production, the function of the secretion has never been studied in non-pollinating species of fig-associated wasps.

Here, we show that even though size of the poison sac differs among species and between pollinators and non-pollinators, there is a strong positive correlation between the poison sac area and egg number that is consistent among galling wasps associated with figs regardless of their roles as pollinators or parasites. This observation is consistent with the hypothesis that both types of wasp use a similar galling mechanism that prevents the fig tree from differentiating between mutualistic and parasitic associates, and sets the 66 stage for evaluating the subsequent hypothesis that contents of the poison sac are similar in each group. To test that hypothesis directly, the exact function of the poison sac must be known in both types of wasps. The first steps have been taken recently to describe this function (e.g., Appendix C, this volume). Although preliminary, the present study is the first to suggest a similar importance of the poison sac in non-pollinating as well as pollinating fig wasps. If both types of wasp use the same mechanism for inducing galls within fig flowers, the fig tree might be unable to distinguish between pollinating and parasitic non-pollinating fig wasps, and thus not be able to differentially impose sanctions.

Acknowledgments

We gratefully acknowledge the National Science Foundation for supporting this research

(NSF Graduate Research Fellowship to EOM), and the Smithsonian Institute (Predoctoral

Fellowship to EOM). We thank the Smithsonian Tropical Research Institute for logistical support, A. E. Arnold for helpful comments on the manuscript, and the government of

Panama for permission to carry out this research. We are grateful to A. Gomez for technical assistance and V. Martinson for helpful discussion.

References Boucek Z. 1993. The genera of Chalcidoid wasps from Ficus fruit in the New-world. Journal of Natural History. 27:173-217.

Compton SG, Rasplus JY, Ware A. 1994. African fig wasp parasitoid communities. Parasitoid community ecology. 343-68

Compton SG, Ellwood MDF, Davis AJ, and Welch K. 2000. The flight heights of chalcid wasps (Hymenoptera, Chalcidoidea) in a lowland bornean rain forest: Fig wasps are the high fliers. Biotropica. 32:515-522.

Edgar RC. 2004. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics. 5:1-19.

Espagne E, Dupuy C, Huguet E, Cattolico L, Provost B, Martins N, Poirie M, Periquet G, and Drezen JM. 2004. Genome sequence of a polydnavirus: Insights into symbiotic virus evolution. Science. 306:286-289.

Grandi G. 1929. Studio morfologico e biologico della Blastophaga psenes (L.). 2a ediz. riveduta. Bollettino del Laboratorio di Entomologia del R.Istituto Superiore Agrario di Bologna. 2:1-147.

Haberman E. 1972. Bee and wasp venoms. Science. 177:314-322.

Hölldobler B. 1971. Sex pheromone in ant Xenomyrmex-floridanus. Journal of Insect Physiology. 17:1497-1499.

Huelsenbeck JP and Ronquist F. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics. 17:754-755.

Jander KC and Herre EA. 2010. Host sanctions and pollinator cheating in the fig tree-fig wasp mutualism. Proceedings of the Royal Society B-Biological Sciences. 277:1481- 1488.

Janzen DH. 1979. How to be a fig. Annual Review of Ecology and Systematics. 10:13-51.

Jansen-González S, Teixeira SP, Pereira RAS. 2012. Mutualism from the inside: coordinated development of plant and insect in an active pollinating fig wasp. Arthropod- Plant Interactions. DOI: 10.1007/s11829-012-9203-6

Jousselin E, Hossaert-McKey M, Herre EA, and Kjellberg FW. 2003. Why do fig wasps actively pollinate monoecious figs? Oecologia. 134:381-387.

Machado CA, et al. 2001. Phylogenetic relationships, historical biogeography and character evolution of fig-pollinating wasps. Proceedings of the Royal Society B. 685- 694.

Machado, C. A., N. Robbins, et al. (2005). Critical review of host specificity and its coevolutionary implications in the fig/fig-wasp mutualism. Proceedings of the National Academy of Sciences of the United States of America 102: 6558-6565.

Maddison, D. R. and W. P. Maddison, 2005. MacClade 4: Analysis of phylogeny and character evolution. Version 4.08a. http://macclade.org.

Maddison WP, and Maddison DR (2009). Mesquite: a modular system for evolutionary analysis.

Marussich WA, and Machado CA. 2007. Host-specificity and coevolution among pollinating and nonpollinating New World fig wasps. Molecular Ecology. 16:1925-1946.

Midford PE, Garland Jr. T, and Maddison WP (2005). PDAP Package of Mesquite.

Molbo D, Machado CA, Sevenster JG, Keller L, and Herre EA. 2003. Cryptic species of fig-pollinating wasps: implications for the evolution of the fig-wasp mutualism, sex allocation, and precision of adaptation. Proceedings of the National Academy of Sciences of the United States of America. 5867-5872.

Pereira RAS and do Prado AP. 2005. Non-pollinating wasps distort the sex ratio of pollinating fig wasps. Oikos. 110:613-619.

Posada D. 2006. ModelTest Server: a web-based tool for the statistical selection of models of nucleotide substitution online. Nucleic Acids Research. 34:700-703.

Rasplus JY, Kerdelhue C, Le Clainche I, and Mondor G. 1998. Molecular phylogeny of fig wasps Agaonidae are not monophyletic. Comptes Rendus De L Academie Des Sciences Serie Iii-Sciences De La Vie-Life Sciences. 321:517-527.

Rønsted N. 2005. 60 million years of co-divergence in the fig–wasp symbiosis. Proceedings of the Royal Society B. 2593-2599.

Vandermeer RK, Glancey BM, Lofgren CS, Glover A, Tumlinson JH, and Rocca J. 1980. The poison sac of red imorted fire ant queens- source of a pheromone attractant Hymenoptera, Formicidae. Annals of the Entomological Society of America. 73:609-612.

Vårdal H. 2006. Venom gland and reservoir morphology in cynipoid wasps. Arthropod structure & development 35:127-36. 69

Verkerke W. 1986. Anatomy of Ficus ottoniifolia (Moraceae) syconia and its role in the fig-fig wasp symbiosis. Proceedings der Koninklijke Nederlandse Akademie van Wetenschappen. 443-469.

Verkerke W. 1988. Sycone morphology and its influence on the flower structure of Ficus sur (Moraceae). Proceedings der Koninklijke Nederlandse Akademie van Wetenschappen. 443-469.

West SA, and Herre EA. 1994. The ecology of the new-world fig-parasitizing wasps Idarnes and implications for the evolution of the fig-pollinator mutualism. Proceedings of the Royal Society of London Series B-Biological Sciences. 258:67-72.

West SA, Herre EA, Windsor DM, and Green PRS. 1996. The ecology and evolution of the New World non-pollinating fig wasp communities. Journal of Biogeography. 23:447- 458.

0.12 A Active Pollinator Passive Pollinator 2 B Non-pollinator 0.09 BC CDE CD

DE E

0.06 F F

0.03 Poison sac area/thorax length (mm ) sac area/thorax Poison

0 F. obtusifolia F. triangle F. citrafolia F. costaricana F. maxima F. insipida F. popenoei F. maxima F. popenoei

Figure 1: Poison sac area corrected for wasp body size (i.e., by thorax length) separated by wasp ecology: active pollinators (white), passive pollinators (light gray), and non- pollinators (dark gray). Letters denote significant differences (Tukey HSD, alpha <

0.05).

70 p = 0.0006 R 2 = 0.806

2 56

42 Poison sac area ( μ m ) 29 Active Pollinator Passive Pollinator Nonpollinator 15 20 40 60 80 100 120 Average eggs per ovary

Figure 2a: Correlation of the average number of eggs per ovary of nine species fig associated wasps with the average area of their poison sac (n = 9, r2 = 0.806, p = 0.0006).

100

F. citrifolia F. obtusifolia F. costaricana F. popenoei F. insipida F. popenoei non-pollinator F. maxima F. triangle 80 F. maxima non-pollinator ) 2 60

40 Poison Sac Area ( μ m Sac Area Poison

0 0 80 40 60 80 100 120 140 160

Eggs per ovary

Figure 2b: Correlation of the average number of eggs per ovary with the average area of their poison sac for every individual wasp measured (n = 213, r2 = 0.553, p > 0.0001).

Colors represent different species of pollinating and non-pollinating fig wasps listed by their host tree.

Table 1. Fig wasp species collected from seven species of Ficus in lowland Panama. Columns

indicate the host Ficus species; the number of individuals collected; the averaged area of the

poison sac; the poison sac area corrected for body size by thorax length, and the average number

of eggs per ovary for each species. Standard error is listed in parentheses.

Poison sac Poison sac/thorax Average egg Wasp Species Fig host Individuals area (µm2) length per ovary Active pollinator

Pegoscapus tonduzi F. citrifolia 28 37.8 (1.5) 0.085 (0.0030) 65.5 (2.25)

Pegoscapus estherae F. costaricana 25 32.2 (1.7) 0.074 (0.0034) 61.2 (2.30)

Pegoscapus hoffmeyerii F. obtusifolia 22 67.2 (1.7) 0.113 (0.0034) 109.2 (2.51)

Pegoscapus gemellus F. popenoei 27 25.4 (1.6) 0.067 (0.0031) 52.0 (2.21)

Pegoscapus lopesi F. triangle 27 43.4 (1.6) 0.097(0.0031) 103.1 (2.30)

Passive pollinator

Tetrapus costaricanus F. insipida 26 46.2 (1.6) 0.067(0.0031) 93.5 (2.26)

Tetrapus americanus F. maxima 28 57.7 (1.6) 0.081(0.0030) 116.2 (2.17)

Non-pollinator

Idarnes sp. F. popenoei 24 21.8 (1.6) 0.045 (0.0032) 28.2 (2.30)

Critogaster sp. F. maxima 19 29.7 (1.9) 0.051 (0.0038) 30.8 (3.19)

F. insipida

100

F. maxima

100 F. triangle

F. costaricana 100 93

F. citrifolia

F. obtusifolia

F. popenoei

Nonpoll F. popenoei

100

Nonpoll F. maxima

0.07

Supplemental Figure 1: Most likely tree resulting from Bayesian Metropolis-coupled

Markov chain Monte Carlo (MCMCMC) analysis on the cytochrome oxidase subunit I

(COI) gene from nine species of fig wasp used in determining phylogenetic independent contrasts. Bayesian posterior probabilities > 0.90 are shown above the branches.

APPENDIX C

METATRANSCRIPTOME ANALYSIS OF FIG FLOWERS

REVEALS POSSIBLE MECHANISMS FOR MUTUALISM

STABILITY AND GALL INDUCTION

To be published in an ecological genomics journal.

Metatranscriptome analysis of fig flowers reveals possible mechanisms

for mutualism stability and gall induction

Ellen O. Martinson1*, Jeremiah Hackett1, A. Elizabeth Arnold2

1: Department of Ecology and Evolutionary Biology, The University of Arizona, Tucson,

AZ 85721 USA

2: School of Plant Sciences, The University of Arizona, Tucson, AZ 85721 USA

*Author for correspondence: [email protected]

Abstract: The interaction between figs (Ficus spp.) and their pollinating wasps (fig wasps; Chalcidoidea, Hymenoptera) is a classic example of an ancient and apparently stable mutualism. A striking property of this mutualism is that pollinating fig wasps consistently oviposit in the inner flowers of the fig syconium (gall flowers, which develop into galls that house developing larvae), but typically do not use the outer ring of flowers (seed flowers, which are pollinated and develop into seeds). To better understand potential differences between gall and seed flowers that might influence oviposition choices, and the unknown mechanisms underlying gall formation, we used a metatranscriptomic approach to analyze eukaryotic gene expression within fig flowers at the time of oviposition choice and early gall development. In a lowland forest in Panama, we introduced individual pollinating wasps (Pegoscapus hoffmeyerii) into receptive syconia of Ficus obtusifolia and evaluated the metatranscriptome of gall and seed flowers before and after pollination and oviposition. Consistent with the unbeatable seed hypothesis, which posits that gall- and seed flowers differ in a manner that influences oviposition choice or success, we found significant differences in gene expression assigned to defense and metabolism between the two flower types in receptive syconia.

Transcripts assigned to flavonoids and defensive genes were prevalent in receptive flowers more likely to become galls, and carbohydrate metabolism was significantly up- regulated. There was no evidence that the fig tree protects a subset of flowers from oviposition with known defensive genes. High expression of the venom gene icarapin during wasp embryogenesis within galled flowers distinguishes it as a candidate gene for gall initiation. In response to galling, the fig significantly up-regulates the expression of 78 chalcone synthase, which previously has been connected to formation of other plant galls.

This study is the first to simultaneously sequence the gene expression profile of both mutualistic partners in a plant-insect mutualism and provides evidence for a mechanism of stability for the mutualism.

Keywords: mutualism, metatranscriptome, Ficus, chalcone synthase, unbeatable seed hypothesis 79

Introduction

Mutualisms exist at all levels of every ecosystem and often provide critical nutrition, protection, metabolic enhancement, or reproductive aid to participating partners

(Bronstein et al. 2006; Leigh et al. 1995). However, the mechanisms that maintain mutualisms are not generally clear (Anstett 2001; Herre et al. 1999). The intricate and obligate mutualism between figs (Ficus spp.) and fig wasps (Chalcidoidea, Agaonidae) has evolved over the past 80 million years (Cruaud et al. 2012; Machado et al. 2001;

Machado et al. 2005; Rønsted 2005). Within the enclosed inflorescence (syconium) of the fig, individual flowers that receive the egg of a fig wasp develop into galls in which wasp larvae develop and complete metamorphosis, but flowers receiving only pollen develop into seeds. Therefore, a single flower can produce either a fig wasp or a fig seed

(Herre et al. 1997), creating tension in the mutualism as the interests of the fig and wasp do not fully align: the fig requires both seeds (female fitness) and wasps to carry pollen to the next fig (male fitness); however, the fig wasp does not have an individual advantage to leaving any flowers for seeds (Herre et al. 1997).

Researchers have long noted that wasps consistently oviposit in only about half of the flowers available in a fig syconium (Herre et al. 1997). This behavior cannot be attributed to physical limitations of fig wasps (i.e. ovipositor length, egg number) (Anstett et al.

1996; Bronstein 1988; Herre et al. 1999; Nefdt et al. 1996), leading some authors to suggest that stabilizing mechanisms are used by the fig tree to control the fig wasp-to- seed ratio (West et al, 1994). The exact mechanism is unknown, but previous studies 80 have shown that flowers within a given syconium are not equally attractive to fig wasps

(Verkerke 1986; West et al. 1994). Flowers located closer to the lumen of the syconium

(hereafter gall flowers) are strongly preferred by fig wasps for oviposition, whereas those near the fig wall (hereafter seed flowers) are preferentially pollinated and develop into seeds of the fig (Anstett 2001) (Figure 1). West and Herre (1994) proposed an explanation for the stability in which some flowers block the ovipositing and/or development of fig wasps through biochemical or developmental mechanisms that differ between seed- and gall flowers, thereby creating an ‘unbeatable seed’ (West et al. 1994).

This hypothesis is supported by the fact that regardless of how many foundress wasps enter the fig, the number of offspring emerging remains constant across syconia of the same species (Herre 1989). However, this hypothesis has yet to be tested directly, and the reasons for which particular flowers are not used by fig wasps remains unknown.

In addition to oviposition choice, the ability of the ovipositing fig wasp (i.e., foundress wasp) to successfully initiate gall formation in a fig flower is another critical yet unknown mechanism in this mutualism. Although >13,000 insect species induce galls in plants, the mechanism of how an insect induces a gall to form remains unknown (Mani

1992). Galls are an extended phenotype of the insect, so it is the insect that directs the plant to create a unique structure specific to its needs in the host plant tissue (Stern 1995).

The structure of the fig wasp gall is simple and confined to a single flower providing protection and nutrition for a single wasp larva (Weiblen 2002). It is a rare example of a galling insect involved in a mutualism with its host (Shorthouse 1992). Neither the signal 81 by which the wasp initiates the gall, nor the response of the fig to that signal, has been identified.

In this study, we sequenced the mRNA of F. obtusifolia flowers and all other eukaryotic organisms present within the flower tissue, creating a metatranscriptome of the fig and fig wasp mutualism. We compared metatranscriptomes of seed and gall flowers in figs harvested before and after entrance by fig wasps to determine (1) biochemical and metabolic differences between these two flower types that may influence oviposition by wasps, and (2) gene expression of both plant and wasp at the beginning of gall development. These two elements (oviposition choice and galling), which are essential to the fig and fig wasp interaction, may provide clues to the evolutionary origin and stability of this ancient and classic mutualism.

Materials and Methods

Flowers of Ficus obtusifolia (subgenus Urostigma, section Americana) were collected in

January-February 2010 from Barro Colorado National Monument, Panama (BCNM;

9°9’N, 79°51’W; 25m above sea level; for a full site description see Leigh et al. 1996).

Dozens of trees were surveyed to identify two trees with compatible timing, defined as one releasing pollinating wasps (i.e., source tree) and the other having nearly receptive syconia (i.e., recipient tree). Selected trees were mature individuals located ca. 8.6 km apart, which is within the flight range of foundresses (Nason et al. 1996).

Syconia of the recipient tree were covered individually with fine mesh bags to exclude pollinating and non-pollinating fig wasps. When syconia on the recipient tree became receptive, one pollinating fig wasp (Pegoscapus hoffmeyerii), obtained from syconia on the source tree, was introduced to each fig. Each recipient syconium was then re-bagged for another 24 hours to ensure pollination and galling by a single individual.

RNA isolation and sequencing

A ‘receptive’ sample of fig syconia was collected from the recipient tree prior to the introduction of a pollinating fig wasp, and a ‘pollinated’ sample of fig syconia was collected 24 hours after wasp introduction. Syconia were submerged immediately in

RNAlater (Qiagen) and held at -20°C until gall- and seed flowers were separated based solely on the location within the syconium (Figure 1) with sterile microforceps and stored separately in 0.2ml tubes in RNAlater at -80°C. Samples of each type (i.e., receptive gall flower, receptive seed flower, pollinated gall flower, and pollinated seed flower) flowers were pooled from 2-3 syconia from the single recipient tree.

Before extraction, each sample was soaked in RNase-free water for 1 minute to remove residual RNAlater and then ground in liquid nitrogen. Total RNA was extracted from receptive gall flowers, receptive seed flowers, pollinated gall flowers, and pollinated seed flowers separately using the Spectrum Plant Total RNA Kit (following manufacturer's instructions; Sigma-Aldrich, St. Louis, MO) and visualized on an Agilent 2100

Bioanalyzer (Santa Clara, CA) to assess quality. mRNA was purified using a poly(T) 83 oligo, selecting for all eukaryotic sequences in the sample. cDNA library creation and sequencing was performed at the University of Arizona Genetics Core (UAGC) using the

Illumina RNA-Seq TruSeq protocol on a HiSeq 2000 (San Diego, CA) with a 300 bp average insert length. The resulting four libraries were pooled and run on two-thirds of one lane.

Assembly and mapping

After assessing read quality in FastQC Version 0.10.0

(www.bioinformatics.bbsrc.ac.uk/projects/fastqc), we removed the first 10 base pairs of every read because they contained a skewed GC content relative to the rest of the read. Further trimming was performed in Trimmomatic Version 0.22

(http://www.usadellab.org/cms/index.php?page=trimmomatic) where low quality reads

(phred scores <15) were identified from both ends and across the entire read by a 15 bp sliding window. After trimming, reads shorter than 75 bp were removed.

De novo assembly was performed using Oases version 0.2.08 (Schulz et al) on each sample, followed by an additional assembly using the total combined reads to form a non-redundant reference metatranscriptome. Six different builds were created per sample using a range of hash lengths from 31 to 51 kmers with the minimum contig length set to twice the kmer length. The final build was constructed using the largest hash length (51 kmer) and included the five additional builds as long sequences enabling the Oases

“conserve long” option (Schulz et al). Oases clusters highly similar sequences into 84 groups likely derived from the same loci. If multiple transcripts were listed in one of these groups, the longest read in the group was used for subsequent analyses to prevent redundancy in mapping back to the reference metatranscriptome.

Sequence reads from each sample were mapped to the non-redundant reference metatranscriptome to quantify the abundance of each assembled transcript using the default parameters in CLC Genomics Workbench software (CLC bio, Denmark). The coverage of each transcript was determined in terms of the number of reads per kilobase per million (RPKM).

Annotation and taxonomic placement

For functional annotation of each transcript, all sequences were searched against the non- redundant protein database (nr) using BLASTx with a cutoff e-value = 1 x 10-5. Gene ontology (GO) terms were then assigned by BLAST2GO v2.5.0 (Conesa et al. 2005). GO terms were subsequently assigned to metabolic pathways according to KEGG mapping

(Kanehisa et al. 2000; Kanehisa et al. 2012).

Whereas the majority of the tissue sequenced was plant tissue, a diverse community of organisms was identified within the microcosm of fig syconia (e.g., wasps, fungi, bacteria). The taxonomic assignments of transcripts were predicted using MEGAN4 v4.62.7 (Huson et al. 2011) using the BLASTx results mentioned above. Each taxonomic assignment was performed using the LCA algorithm, which assigns each transcript to the 85 lowest common ancestor in the NCBI taxonomy from a subset of the best scoring matches in the BLAST result. Transcripts assigned to Hymenoptera were compared to the

Nasonia vitripennis official peptide set using BLASTx (Werren et al. 2010).

Statistical analyses

Transcript abundance levels from CLC Genomic Workbench were analyzed for significant differences in expression levels based on RPKM. Pollinated gall flowers were compared separately to pollinated seed flowers and receptive gall flowers by using the proportion-based test of Baggerly et al. (2003) with a p-value corrected for false discovery rate (FDR). Expression levels were also compared between receptive seed flowers and receptive gall flowers using the Kal test (Kal et al. 1999) with a p-value corrected for FDR.

Overrepresented gene ontology (GO) categories were determined using BiNGO in

Cytoscape (Maere et al. 2005). Subsets of GO terms were compared to all the GO terms identified in the reference metatranscriptome. Subsets included GO terms from up- and down-regulated genes in pollinated gall flowers relative to pollinated seed and receptive gall flowers, as well as up- and down-regulated genes in receptive gall flowers relative to receptive seed flowers.

Results and Discussion

After trimming low quality base pairs and reads, ca. 204.7 million clean paired-end reads remained for assembly. Samples averaged 51.2 million reads (range: 17.7 to 83.5 million reads) (Table 1). De novo assembly of clean reads with Oases resulted in an average of

36,092 (SE = 3,179) transcripts per sample after selecting for the longest transcript per cluster.

Pollinated gall flowers had the greatest number of assembled transcripts partly because they contained eggs of pollinating fig wasps as well as other organisms introduced to the fig by the wasp in addition to the plant tissue (Table 1). Because all samples were collected from the same individual, a non-redundant metatranscriptome was assembled using the combined cleaned reads from all four samples to be used as a reference metatranscriptome for mapping. The overall metatranscriptome was only 16% larger than the largest single assembly (54,550 contigs with an n50 of 1995 base pairs), demonstrating that the majority of transcripts were expressed in more than one sample

(Supplemental figure 1).

Mapping

The cleaned, short-sequence reads were mapped for each sample using the overall metatranscriptome as a reference. Over half of the contigs (51.2%, 27,929 contigs) were present in all four samples. An average of 2,612 reads covered each contig, but this was highly variable (SD = 6,772). On average 69.7% of the short sequence reads mapped onto 87 the reference metatranscriptome, with only 1.3% of the reads mapping to multiple loci

(Table 2). A total of 30,166 of 54,550 contigs (55.3%) had similarity to a known protein in the NCBI nr database, with a cutoff of e < 1 x 10-5. BLAST2GO assigned a functional annotation to 19,566 contigs (35.9% of the total assembly).

MEGAN analysis

MEGAN was used to sort and provide taxonomic assignments for contigs. Taxonomic assignment was given to 72.4% of the contigs with an annotation from BLAST (Table 3).

The vast majority of the assignments (85.9%) in pollinated seed flowers were to

Viridiplantae, with very few assignments to fungi (1.0%) and other organisms (13.0%).

The majority of the assignments for pollinated gall flowers also were to Viridiplantae

(60.8%); however, nearly a quarter of the reads were assigned to Hymenoptera (21.3%).

Pollinated gall- and seed flowers were collected from the same fig syconia ca. 24 hours after ovipositioning and pollination. At this time there were no visible indication of galling, such that the two flower types were collected based solely upon their position within each syconium. Our samples provide further evidence for the strong the preference of the fig wasp to oviposit in the inner ring gall flowers.

Expression differences between receptive gall and seed flowers

Fig trees and wasps have maintained a stable mutualism for over 80 million years

(Machado et al. 2001; Machado et al. 2005; Rønsted 2005). While the exact mechanism behind this stability is unknown, it has been shown that both pollinating and non- 88 pollinating fig wasps strongly prefer to oviposit in the inner ring of flowers within the fig syconium (Verkerke 1986; West et al. 1994) (Figure 1). This preference has been interpreted as the fig tree limiting the fig wasp to a subset of its flowers allowing for development of both wasp pollinators and fig seeds. The unbeatable seed hypothesis posits that this limitation is due to an inherent difference between gall flowers and seed flowers prior to oviposition or pollination (West et al. 1994). Because there are no visible structural differences in the ovule shape, it is presumed to be a biochemical or defensive difference (Verkerke 1986). To determine if there is any mechanism present in seed flowers to hinder oviposition, thereby supporting the unbeatable seed hypothesis, we compared gene expression of gall and seed flowers from a fig that was receptive to oviposition and pollination, but before a fig wasp has entered the syconium (i.e., receptive gall flower, receptive seed flower). We found 4,707 loci differentially expressed between the two flower types. This is the first study to provide evidence that there are biochemical differences between gall and seed flowers in agreement with predictions from the unbeatable seed hypothesis proposed by West and Herre (1994).

Of the 4,707 differentially expressed loci, 2,848 were up-regulated in gall flowers.

Among the up-regulated genes, 778 loci were assigned GO terms, of which 127 were significantly overrepresented in the receptive gall flower compared to the reference metatranscriptome (corrected p-value 0.00471 to 7.01x10-44) (Supplemental Table. 1).

Transcripts assigned to L-phenylalanine metabolism, specifically chalcone synthase, were 89 the most common, present in 11 overrepresented GO terms and making up 30% of the top

100 expressed transcripts in gall flowers (Supplemental Table 2).

Chalcone synthase is the first step in the flavonoid pathway, which serves many distinct functions in plants (pigment production, defense compounds, and UV protection e.g.,

Ferret et al. 1999; Ryder et al. 1987; Strid et al. 1994)). The roles that flavonoids play in plant-insect interactions also are varied (reviewed in Simmonds 2001; 2003). Flavonoids can be a feeding stimulant or deterrent to both polyphagous and oligophagous insects

(Abou-Zaid et al. 1993; Barbehenn 2002; Beninger et al. 1997), can be sequestered in the cuticle to increase fitness (Bernays et al. 2000; Burghardt et al. 2001; Tamura et al.

2002), or, in several species of butterfly, can stimulate oviposition (Feeny et al. 1988;

Honda 1990; Nishida et al. 1987). Whereas the exact structures and functions of the flavonoids within the gall flower cannot be determined by our metatranscriptome analysis, previous research has demonstrated that insects can detect flavonoids in plants and use this information for host selection (Roessingh et al. 1991). The up-regulation of flavonoids in gall flowers would create a strong signal, detectable to the fig wasp, which could direct foundress wasps to oviposit in gall flowers. We propose this as a hypothesis for further study.

Receptive gall flowers have 108 transcripts that are unique, having no expression in seed flowers (Supplemental Table 3). This contrasts with the observation of only five unique transcripts in seed flowers. Transcripts assigned to disease resistance were the most 90 prevalent group of unique transcripts in gall flowers, making up 55% of the loci assigned an annotation (30 loci). Over half of these transcripts were specifically assigned to orthologs that encode nucleotide-binding site-leucine-rich repeat (NBS-LRR) proteins, which are used to detect pathogen-related molecules in plants (DeYoung et al. 2006). No putative NBS-LRR transcripts were observed in seed flowers. If NBS-LRR proteins are also involved in defense in figs, this observation raises the question: why would fig trees preferentially protect receptive gall flowers over seed flowers before anything has entered the syconium as both are equally important to the fitness of the tree?

A possible explanation could come from differences in nutritional quality of the two flower types, which may lead to unequal infection rates of pathogens. In the domesticated fig F. carica, concentrations of sucrose, galactose and fructose vary throughout fig development. The highest concentrations occur when the fruit is ripe; however, a secondary peak in sugar content occurs when the fig is receptive (i.e., at the end of flower development (Ersoy et al. 2007; Yoshioka 1995)). A large set of up-regulated transcripts in gall flowers, present in 25 (19.7%) overrepresented GO terms, dealt with sugar and carbohydrate metabolism (Supplemental Table 1 & 4). On average transcripts within these GO terms had a 5.1-fold increase of expression from seed to gall flowers. Several transcripts assigned to sucrose, galactose and fructose metabolism (e.g., sucrose synthase,

UDP-D-glucuronate carboxy-lyase, alpha-galactosidase and fructokinase-2) had high expression in the gall flower. This significantly increased expression suggests that gall flowers may have higher sugar content than seed flowers. The higher nutritional quality 91 of the tissue could not only lead to a higher infection rate by bacteria and fungi present in the syconia (e.g., Martinson et al. 2012), but may also play a role in oviposition choice for the wasps (details below).

Avoidance of seed flowers

Transcripts up-regulated in receptive gall flowers might reveal factors that attract foundress wasps to that set of flowers. In contrast, transcripts up-regulated in receptive seed flowers may provide insights into how fig trees restrict oviposition in the outer ring of flowers.

Of the 4707 transcripts that were differentially regulated between gall and seed flowers,

1859 were up-regulated in seed flowers. Within the up-regulated transcript set, 201 loci were assigned GO terms, of which 23 were significantly overrepresented in the receptive gall flower compared to the reference metatranscriptome (corrected p-value 0.04921 to

1.03x10-21) (Supplemental Table 5). Unfortunately more information for gene annotation is needed before any clear assessment of the up-regulated seed genes can be determined.

In the top 100 expressed transcripts of seed flowers, 73% either had no putative annotation or were assigned an annotation of unknown/hypothetical protein (compared to

40% in gall flowers). Overrepresented GO terms were very general (e.g., translation, gene expression, protein metabolic process) and gave no indication to any specific mechanism that might influence oviposition (Supplemental Table 5).

The preference for ovipositing in gall flowers is very strong for foundresses. Fig wasps from the same genera (Pegoscapus spp.) have been shown to die with up to 75% of their eggs unlaid, yet seed flowers in the same syconium remained without eggs (Herre 1989).

Additionally, non-pollinating wasps, which exhibit different ecologies (competitors, parasitoids, etc) and oviposit from outside the syconium, preferentially use gall flowers even though the outer ring of seed flowers is more accessible (West et al. 1994). In this study, we provide further evidence for seed-flower avoidance, as we did not find any evidence of wasp eggs in seed flowers (see Megan analysis above). As the knowledge of plant functional genomics increases, a mechanism by which seed flowers actively inhibit wasp oviposition may come forth. However, the difference in sugar and carbohydrate metabolism between gall and seed flowers presents a new hypothesis: seed flowers do not have the nutritional quality to support the development of a wasp larva or contain diminished nutrition that would severely decrease larval fitness with a longer developmental time or smaller adult size. The hypothesis has some support from previous studies that have shown wasps that develop closer to the fig wall are significantly smaller and are less likely to be mated and released by male wasps (Anstett

2001; Dunn et al. 2008).

Theory has suggested that positive feedback between the fitness of mutualistic partners, also known as partner fidelity feedback, is sufficient to prevent cheating and can create a more stable mutualism than punitive and costly host sanctions (Weyl et al. 2010). By providing a nutrient-rich subset of flowers for wasp development, the fig tree could increase its female fitness by attaining seeds in the less nutritious seed flowers, and 93 increase its male fitness by fostering more robust wasps that can carry more pollen greater distances.

Gall initiation and expression differences between pollinated gall and seed flowers

Currently, little is known about gall initiation in pollinating fig wasps. It has been shown that several drops of a maternal secretion are deposited in every flower in which a foundress wasp lays an egg (Verkerke 1986). This secretion is produced by the poison gland (also known as the venom or acid gland) and has been associated with the initiation of gall growth within the fig flower (Grandi 1929; Verkerke 1986; 1988). However, the mechanism by which this secretion instigates gall growth is unknown. This study is the first look at gene expression dealing with the interaction between the fig flower and fig wasp at the initiation of gall formation.

Fig wasp expression

The 7,045 contigs found in pollinated gall flowers that were assigned to homologs in

Hymenoptera had hits to 4,401 genes in the Nasonia vitripennis official peptide set using

BLASTx (e = 1.0 x 10-7). Nasonia vitripennis is a parasitic chalcid wasp within the same superfamily as pollinating fig wasps. The contigs covered 34.3% of the total number of genes in N. vitripennis that have been validated by expression data (Werren et al. 2010).

The difference in the number of genes could be explained by differential expression of a

24-hour-old egg compared to an adult wasp or differences in gene content between the fig wasp and Nasonia. The average number of reads for a Hymenoptera transcript was 140.9, 94 demonstrating good coverage of the transcripts that were sequenced. Additionally, central metabolic pathways are relatively complete compared to Nasonia’s KEGG pathways, such as the oxidative phosphorylation pathway (80.0% complete with 68/85 genes) or the citrate cycle pathway (91.3% complete with 21/23 genes).

The hymenopteran transcript with the highest level of expression was assigned to the gene Giant (gt) with 17,121 unique reads (RPKM = 9238.3 and ~3500 reads more than the next highest transcript). Knockouts of gt in N. vitripennis result in the loss of the head and thorax, as well as fused A6 and A7 abdominal segments (Brent et al. 2007). Giant is a maternally expressed gene localized in the anterior region of the wasp embryo to repress the expression of the trunk gap genes so that the head and thorax can form properly (Brent et al. 2007). Additional transcripts with annotations to embryonic development genes were present among the highest expressed genes. Two other transcripts, totaling 5813 reads, were assigned to the predicted protein lethal(2)essential for life [l(2)efl] in N. vitripennis. Mutations of l(2)efl are lethal in the embryonic stage and l(2)efl is highly expressed throughout embryogenesis (Kurzik-Dumke et al. 1995).

The high level of expression of these genes suggests that the fig wasp was still undergoing embryogenesis when the gall flowers were collected.

Potential gall initiation genes

To explore potential roles of the maternal secretion in gall formation, assembled transcripts were compared to ESTs of the N. vitripennis venom gland (where the maternal 95 secretion is produced) as well as other known venom genes to determine if there was any expression of these genes in the embryonic wasp (De Graaf et al. 2010; Werren et al.

2010) (Supplemental Table 6). Two transcripts, aspartylglucosaminidase (RPKM 189 with a putative function in host paralysis) and cysteine-rich/KU venom protein (98

RMKM and a proteinase inhibitor), had significant matches to genes expressed in the N. vitripennis poison gland (De Graaf et al. 2010). An additional seven transcripts had significant matches to other known venom genes: carboxylesterase-6 (30 RPKM), dipeptidyl peptidase 4 (three transcripts totaling 219.4 RPKM), serine carboxypeptidase

(two transcripts totaling 157.9 RPKM), and icarapin. The transcript matching to icarapin is especially interesting because it is the 79th top expressed gene, with a RPKM of 624.2.

Icarapin has previously been found in honey bees and plays an active role in honey bee venom producing an icarapin-specific IgE-response in humans (Peiren et al. 2006).

Although its exact function is unknown in this case, the high level of expression and association with the venom gland suggests that this gene may have a role in gall initiation and warrants further study.

The origin of the icarapin transcript is also unknown. It may be deposited as part of the maternal secretion of the foundress wasp, but the survival of an unprotected, foreign mRNA transcript in the fig is unlikely. Alternatively, it may arise from maternally acting genes within the fig wasp egg/embryo deposited by the foundress wasp similar to the

Giant gene. A third possibility is that the embryo itself could be expressing this gene in 96 early development before the formation of the poison gland to initiate and maintain the gall.

Response of the fig to oviposition

To analyze gene expression involved with the interaction between the fig flower and the fig wasp embryo, transcripts dealing with pollination and normal tissue function needed to be removed from further analyses. Transcriptomes were compared and all non- significantly differentially expressed transcripts (FDR corrected p < 0.05) were removed.

Expression profiles from pollinated gall flowers were compared to expression profiles from pollinated seed flowers to identify and remove genes involved in pollination, because both pollinated gall and seed flowers typically receive a pollen grain from the foundress wasp (Jousselin et al. 2003). Additionally, pollinated gall flowers also were compared to receptive gall flower expression to identify and remove genes unique to gall flowers, but not involved in the interaction with the fig wasp.

In the pollinated gall flowers, 910 transcripts were differentially expressed relative to both pollinated seed and receptive gall expression. With transcripts involved in normal seed and gall flower expression removed from the analysis, these 910 transcripts were identified as being potentially involved with the interaction between the fig flower and the pollinated wasp embryo at the beginning of gall formation. Relative to both of the comparison samples, the majority of the differentially expressed, pollinated gall transcripts were up-regulated (508/910), only 23% of the transcripts down-regulated 97

(209/910), and the remainder had mixed expression (i.e. up-regulated compared to the pollinated seed, but down-regulated compared to the receptive gall).

From the 910 differentially expressed transcripts, 460 GO terms were assigned. Within the up-regulated transcripts of pollinated gall flowers, 44 GO terms were significantly overrepresented (corrected p-value 0.00359 to 2.97x10-9) (Supplemental Table 7).

Similar to receptive galls, transcripts assigned to genes involved in chalcone synthase activity were the most common, present in 15 overrepresented GO terms (34%).

Chalcone synthase and isomerase also made up 20% of the top 100 expressed transcripts.

The role of flavonoids in gall formations has recently been discovered: previous studies have found that high levels of flavonoids (specifically chalcone synthase) are present at the beginning stages of gall formation by Agrobacterium tumefaciens, clubroot

(Plasmodiophora brassicae), and root-knot nematodes, and in the nodules formed by

Rhizobia (Hutangura et al. 1999; Päsold et al. 2010; Schwalm et al. 2003; Wasson et al.

2006). Within gall formation, flavonoids act as a transport inhibitor to auxin, a plant growth hormone, which allows its concentration to build in the galled tissue (Hutangura et al. 1999; Päsold et al. 2010; Schwalm et al. 2003; Wasson et al. 2006). To the best of our knowledge, this is the first study to demonstrate a high expression level of transcripts assigned to chalcone synthase in an insect-induced gall, which provides additional evidence that chalcone synthase may play a role in a universal mechanism of gall induction. The rise in concentration of flavonoids occurs before any changes in the expression of auxins in previously studies plants (Hutangura et al. 1999), which suggests 98 an explanation for the lack of growth factors in the overrepresented GO terms in the pollinated gall flowers. Whether the exact composition of flavonoids remains the same in receptive to pollinated gall flower can not be determined; however, it is a possibility that fig wasps may be directed to oviposit in gall flowers by an integral part of the gall initiation mechanism.

Transcripts assigned to class III peroxidases were also common, present in 14 overrepresented GO terms (32%) (Supplemental Table 7). Class III peroxidases are present in large multigene families in all land plants (Duroux et al. 2003; Tognolli et al.

2002) and are involved in numerous physiological processes throughout the plant life cycle (Cosio et al. 2009), which makes it difficult to determine the function of the peroxidases found here. Previous studies have found high levels of peroxidase in the early stages of gall development in other plant species (Hori et al. 1997; Päsold et al.

2010), where they were first thought to break down auxins with flavonoids preventing this reaction. However, this was shown not be the case as flavonoids present in gall development actually increased the degradation of auxins with certain preoxidases

(Mathesius 2001; Päsold et al. 2010). Why the plant would simultaneously sequester and degrade auxins was not addressed in that work, and remains unclear.

Alternatively, peroxidases could be performing a different function. Previous studies have shown a significant increase in expression of class III peroxidases as a form of defense against the Hessian fly in wheat and rice (Liu et al. 2010). Peroxidases form 99 reactive oxygen species (ROS) that can be directly toxic to pathogens or trigger other defensive mechanisms (Apel et al. 2004). They are an effective defense against microbial pathogens, but studies of their effectiveness against insect pathogens are still preliminary

(Almagro et al. 2009). A defensive function is supported by the presence of other transcripts assigned to defense-induction genes (i.e. beta-galactosidase, chitinase, and beta-1,3-glucanase) in three overrepresented GO terms (roles in plant defense reviewed in

Grover 2012; Ketudat Cairns et al. 2010). If they have the same function in figs, the presence of transcripts assigned to defensive response gene orthologs prompts the question: why would the pollinated fig flowers attack the mutualistic pollinator? The fig wasp, however, is not the only organism to enter the syconium. Fig wasps are hosts to many organisms (i.e. bacteria, fungi, nematodes), some of which are plant specific (Herre

1995; Martinson et al. 2012). This non-specific defensive response could be protecting the developing gall from potential pathogens the wasp carries from tree to tree. In turn, fig wasps, which have been co-evolving with fig trees for over 80 million years, would have most likely developed a tolerance for these non-specific defense mechanisms.

Conclusion

This study is the first to simultaneously describe and quantify the gene expression profile of both mutualistic partners in a plant-insect mutualism. Analyses of these metatranscriptomes have shown the prominence of transcripts assigned to flavonoids and defensive orthologs in both pollinated and receptive gall flowers of Ficus obtusifolia, and have revealed for the first time detectable differences in metabolism and defense between 100 gall flowers and seed flowers before oviposition, which provides support for the unbeatable seed hypothesis. The high expression of a transcript annotated as the venom gene icarapin during wasp embryogenesis has implications for the mechanism of gall initiation. Further, our new hypothesis for the mechanism of mutualism stability in the fig and fig wasp system suggests that the choice between nutrient-rich and nutrient-poor flowers may explain the balance of the seed-to-wasp ratio.

Acknowledgments

We would like to thank C.A. Machado and E.A. Herre for conversations about the fig and fig wasp mutualism that directed and influenced this study. We are grateful to J.

Wisecaver for technical assistance and V. Martinson for helpful discussion. We gratefully acknowledge the National Science Foundation for supporting this research (IOB-062492 to AEA and an NSF Graduate Research Fellowship to EOM), as well as the Smithsonian

Institute (Predoctoral Fellowship to EOM). We thank the Smithsonian Tropical Research

Institute for logistical support and the government of Panama for permission to carry out this research.

101

References: Abou-Zaid MM, Beninger CW, Arnason JT, Nozzolillo C. 1993. The effect of one flavone, two catechins and four flavonols on mortality and growth of the European corn borer (Ostrinia nubilalis Hubner). Biochemical Systematics and Ecology 21:415-20

Almagro L, Ros LVG, Belchi-Navarro S, Bru R, Barceló AR, Pedreño M. 2009. Class III peroxidases in plant defence reactions. Journal of Experimental Botany 60:377-90

Anstett MC. 2001. Unbeatable strategy, constraint and coevolution, or how to resolve evolutionary conflicts: the case of the fig/wasp mutualism. Oikos 95:476-84

Anstett MC, Bronstein J, HossaertMcKey M. 1996. Resource allocation: a conflict in the fig/fig wasp mutualism? Journal of Evolutionary Biology 9:417-28

Apel K, Hirt H. 2004. Reactive oxygen species: metabolism, oxidative stress, and signal transduction. Annual Review of Plant Biology 55:373-99

Baggerly KA, Deng L, Morris JS, Aldaz CM. 2003. Differential expression in SAGE: accounting for normal between-library variation. Bioinformatics 19:1477-83

Barbehenn RV. 2002. Gut-based antioxidant enzymes in a polyphagous and a graminivorous grasshopper. Journal of chemical ecology 28:1329-47

Beninger CW, Abou-Zaid MM. 1997. Flavonol glycosides from four pine species that inhibit early instar gypsy moth (Lepidoptera: Lymantriidae) development. Biochemical Systematics and Ecology 25:505-12

Bernays EA, Chapman RF. 2000. Plant secondary compounds and grasshoppers: beyond plant defenses. Journal of chemical ecology 26:1773-94

Brent AE, Yucel G, Small S, Desplan C. 2007. Permissive and instructive anterior patterning rely on mRNA localization in the wasp embryo. Science's STKE 315:1841

Bronstein JL. 1988. Mutualism, antagonism, and the fig-pollinator interaction Ecology 69:1298-302

Bronstein JL, Alarcon R, Geber M. 2006. The evolution of plant-insect mutualisms. New Phytologist 172:412-28

Burghardt F, Proksch P, Fiedler K. 2001. Flavonoid sequestration by the common blue butterfly Polyommatus icarus: quantitative intraspecific variation in relation to larval hostplant, sex and body size. Biochemical Systematics and Ecology 29:875-89

102

Conesa A, Gotz S, Garcia-Gomez JM, Terol J, Talon M, Robles M. 2005. Blast2GO: a universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics 21:3674-6

Cosio C, Dunand C. 2009. Specific functions of individual class III peroxidase genes. Journal of Experimental Botany 60:391-408

Cruaud A, Rønsted N, Chantarasuwan B, Chou LS, Clement WL, et al. 2012. An Extreme Case of Plant-Insect Co-Diversification: Figs and Fig-Pollinating Wasps. Systematic Biology doi: 10.1093/sysbio/sys068

De Graaf DC, Aerts M, Brunain M, Desjardins CA, Jacobs FJ, et al. 2010. Insights into the venom composition of the ectoparasitoid wasp Nasonia vitripennis from bioinformatic and proteomic studies. Insect molecular biology 19:11-26

DeYoung BJ, Innes RW. 2006. Plant NBS-LRR proteins in pathogen sensing and host defense. Nature immunology 7:1243-9

Dunn DW, Yu DW, Ridley J, Cook JM. 2008. Longevity, early emergence and body size in a pollinating fig wasp–implications for stability in a fig–pollinator mutualism. Journal of Animal Ecology 77:927-35

Duroux L, Welinder KG. 2003. The peroxidase gene family in plants: a phylogenetic overview. Journal of molecular evolution 57:397-407

Ersoy N, Sadiye G, Lami K. 2007. Changes in sugar contents of fig fruit (Ficus carica l. Cv. Bursa Siyahi) during development. Suleyman Demirel Universitesi Ziraat Fakultesi Dergisi 2:22-6

Feeny P, Sachdev K, Rosenberry L, Carter M. 1988. Luteolin 7-O-(6''-O-malonyl)-[beta]- d-glucoside and trans-chlorogenic acid: Oviposition stimulants for the black swallowtail butterfly. Phytochemistry 27:3439-48

Ferret J, Jez JM, Bowman ME, Dixon RA, Noel JP. 1999. Structure of chalcone synthase and the molecular basis of plant polyketide biosynthesis. Nature structural biology 6:775- 84

Grandi G. 1929. Studio morfologico e biologico della Blastophaga psenes (L.). 2a ediz. riveduta. Bollettino del Laboratorio di Entomologia del R.Istituto Superiore Agrario di Bologna. 2:1-147

Grover A. 2012. Plant Chitinases: Genetic Diversity and Physiological Roles. Critical Reviews in Plant Sciences 31:57-73

103

Herre EA. 1989. Coevolution of reproducive characteristics in 12 species of new world figs and their pollinator wasps Experientia 45:637-47

Herre EA. 1995. Factors affecting the evolution of virulence: Nematode parasites of fig wasps as a case study. Parasitology 111:S179-S91

Herre EA, Knowlton N, Mueller UG, Rehner SA. 1999. The evolution of mutualisms: exploring the paths between conflict and cooperation. Trends in Ecology and Evolution 14:49-53

Herre EA, West SA. 1997. Conflict of interest in a mutualism: documenting the elusive fig wasp-seed trade-off Proceedings of the Royal Society B 264:1501-7

Honda K. 1990. Identification of host-plant chemicals stimulating oviposition by swallowtail butterfly, Papilio protenor. Journal of chemical ecology 16:325-37

Hori K, Wada A, Shibuta T. 1997. Changes in phenoloxidase activities of the galls on leaves of Ulmus davidana formed by Tetraneura fuslformis (Homoptera: Eriosomatidae). Applied entomology and zoology 32:365-71

Huson DH, Mitra S, Ruscheweyh HJ, Weber N, Schuster SC. 2011. Integrative analysis of environmental sequences using MEGAN4. Genome Research 21:1552-60

Hutangura P, Mathesius U, Jones MGK, Rolfe BG. 1999. Auxin induction is a trigger for root gall formation caused by root-knot nematodes in white clover and is associated with the activation of the flavonoid pathway. Functional Plant Biology 26:221-31

Jousselin E, Hossaert-McKey M, Herre EA, Kjellberg FW. 2003. Why do fig wasps actively pollinate monoecious figs? Oecologia 134:381-7

Kal AJ, Van Zonneveld AJ, Benes V, Van Den Berg M, Koerkamp MG, et al. 1999. Dynamics of gene expression revealed by comparison of serial analysis of gene expression transcript profiles from yeast grown on two different carbon sources. Molecular biology of the cell 10:1859-72

Kanehisa M, Goto S. 2000. KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Research 28:27-30

Kanehisa M, Goto S, Sato Y, Furumichi M, Tanabe M. 2012. KEGG for integration and interpretation of large-scale molecular data sets. Nucleic Acids Research 40:D109-D14

Ketudat Cairns JR, Esen A. 2010. β-Glucosidases. Cellular and molecular life sciences 67:3389-405

104

Kurzik-Dumke U, Lohmann E. 1995. Sequence of the new Drosophila melanogaster small heat-shock-related gene, lethal (2) essential for life [ l(2)efl ], at locus 59F4, 5. Gene 154:171-5

Leigh EG, Rand AS. 1996. The ecology of a tropical forest: seasonal rhythms and long term changes. Washington, DC: Smithsonian Institute Press

Leigh EG, Rowell TE. 1995. The evolution of mutualism and other forms of harmony at various levels of biological organization. Ecologie 26:131-58

Liu X, Williams CE, Nemacheck JA, Wang H, Subramanyam S, et al. 2010. Reactive oxygen species are involved in plant defense against a gall midge. Plant physiology 152:985-99

Machado CA, Robbins N, Gilbert MTP, Herre EA. 2005. Critical review of host specificity and its coevolutionary implications in the fig/fig-wasp mutualism. Proceedings of the National Academy of Sciences of the United States of America 102:6558-65

Maere S, Heymans K, Kuiper M. 2005. BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 21:3448-9

Mani MS. 1992. Introduction to Cecidology. In Biology of insect-induced galls, ed. JD Shorthouse, O Rohfritsch:3-8. New York: Oxford University Press. Number of 3-8 pp.

Martinson EO, Herre EA, Machado CA, Arnold AE. 2012. Culture-Free Survey Reveals Diverse and Distinctive Fungal Communities Associated with Developing Figs (Ficus spp.) in Panama. Microbial Ecology: doi 10.1007/s00248-012-0079-x

Mathesius U. 2001. Flavonoids induced in cells undergoing nodule organogenesis in white clover are regulators of auxin breakdown by peroxidase. Journal of Experimental Botany 52:419-26

Nason JD, Herre EA, Hamrick JL. 1996. Paternity analysis of the breeding structure of strangler fig populations: evidence for substantial long‐distance wasp dispersal. Journal of Biogeography 23:501-12

Nefdt RJC, Compton SG. 1996. Regulation of seed and pollinator production in the fig- fig wasp mutualism. Journal of Animal Ecology 65:170-82 105

Nishida R, Ohsugi T, Kokubo S, Fukami H. 1987. Oviposition stimulants of a Citrus- feeding swallowtail butterfly, Papilio xuthus L. Cellular and molecular life sciences 43:342-4

Päsold S, Siegel I, Seidel C, Ludwig-Müller J. 2010. Flavonoid accumulation in Arabidopsis thaliana root galls caused by the obligate biotrophic pathogen Plasmodiophora brassicae. Molecular plant pathology 11:545-62

Peiren N, de Graaf DC, Brunain M, Bridts CH, Ebo DG, et al. 2006. Molecular cloning and expression of icarapin, a novel IgE-binding bee venom protein. FEBS letters 580:4895-9

Roessingh P, Stadler E, Schoni R, Feeny P. 1991. Tarsal contact chemoreceptors of the black swallowtail butterfly Papilio polyxenes: responses to phytochemicals from host- and non-host plants. Physiological entomology 16:485-95

Rønsted N. 2005. 60 million years of co-divergence in the fig-wasp symbiosis. Proceedings of the Royal Society B 272:2593-9

Ryder TB, Hedrick SA, Bell JN, Liang X, Clouse SD, Lamb CJ. 1987. Organization and differential activation of a gene family encoding the plant defense enzyme chalcone synthase in Phaseolus vulgaris. Molecular and General Genetics MGG 210:219-33

Schulz MH, Zerbino DR, Vingron M, Birney E. 2012. Oases: robust de novo RNA-seq assembly across the dynamic range of expression levels. Bioinformatics 28:1086-92

Schwalm K, Aloni R, Langhans M, Heller W, Stich S, Ullrich CI. 2003. Flavonoid- related regulation of auxin accumulation in Agrobacterium tumefaciens-induced plant tumors. Planta 218:163-78

Shorthouse JD, Rohfritsch O. 1992. Biology of insect-induced galls. New York: Oxford University Press

Simmonds MSJ. 2001. Importance of flavonoids in insect-plant interactions: feeding and oviposition. Phytochemistry 56:245-52

Simmonds MSJ. 2003. Flavonoid-insect interactions: recent advances in our knowledge. Phytochemistry 64:21-30

Stern DL. 1995. Phylogenetic evidence that aphids, rather than plants, determine gall morphology Proceedings of the Royal Society B 260:85-9

106

Strid Å, Chow WS, Anderson JM. 1994. UV-B damage and protection at the molecular level in plants. Photosynthesis Research 39:475-89

Tamura Y, Nakajima K, Nagayasu K, Takabayashi C. 2002. Flavonoid 5-glucosides from the cocoon shell of the silkworm, Bombyx mori. Phytochemistry 59:275-8

Tognolli M, Penel C, Greppin H, Simon P. 2002. Analysis and expression of the class III peroxidase large gene family in Arabidopsis thaliana. Gene 288:129-38

Verkerke W. 1986. Anatomy of Ficus ottoniifolia (Moraceae) syconia and its role in the fig-fig wasp symbiosis. Proceedings der Koninklijke Nederlandse Akademie van Wetenschappen 443-69

Verkerke W. 1988. Sycone morphology and its influence on the flower structure of Ficus sur (Moraceae). Proceedings der Koninklijke Nederlandse Akademie van Wetenschappen:319-44

Wasson AP, Pellerone FI, Mathesius U. 2006. Silencing the flavonoid pathway in Medicago truncatula inhibits root nodule formation and prevents auxin transport regulation by rhizobia. The Plant Cell Online 18:1617-29

Weiblen GD. 2002. How to be a fig wasp. Annual Review of Entomology 47:299-330

Werren JH, Richards S, Desjardins CA, Niehuis O, Gadau J, Colbourne JK. 2010. Functional and evolutionary insights from the genomes of three parasitoid Nasonia species. Science 327:343-8

West SA, Herre EA. 1994. The ecology of the new-world fig-parasitizing wasps Idarnes and implications for the evolution of the fig-pollinator mutualism. Proceedings of the Royal Society B 258:67-72

Weyl EG, Frederickson ME, Douglas WY, Pierce NE. 2010. Economic contract theory tests models of mutualism. Proceedings of the National Academy of Sciences 107:15712- 6

Yoshioka K. 1995. Changes in Saccharides and Enzyme Activities Related to Carbohydrate Metabolism during Growth and Development of Fig Fruit (Ficus carica L.) on the Tree. Journal of the Japanese society for food science and technology 42:219-24

107

"g wall

lumen

seed !owers gall !owers

2 cm

Figure 1 – Diagram of a cross section of a receptive Ficus obtusifolia synconium

108

638.0 956.3 965.8 1159.2 1059.0 Average length s indicate

total number of base n50 2424 2218 1782 1996 1995

32394 32711 33669 45595 54550 Contigs

20667729 37917150 32199308 48283560 52687202 Base pairs aining after quality trimming,

17220340 72519250 31455052 83553656 204748298 After trimming

23729734 96584232 48698380 110933944 279946290 Raw reads

obtained from the original run, reads rem

Table 1: Sequencing and assembly statistics for each flower type all samples combined. Column pairs from all quality reads, number of contigs that were assembled, the n50 assembled contigs, and average contig length. Receptive Seed Receptive Gall Pollinated Seed Pollinated Gall All Samples Combined number of raw reads

109

mapped

ome, reads t ip of reads that

262587 (1.61) 973360 (1.63) 360204 (1.91) 946752 (2.27) specifically - (% counted reads)

Non

d the number

(98.37) mapped to the reference transcr 11321482 (98.39) 50089645 21780903 (98.09) 58022725 (97.73) Uniquely mapped (% counted reads)

(% total)

Counted reads

11584069 (67.27) 51063005 (70.41) 22141107 (70.39) 58969477 (70.58) parentheses

mapped, the number of reads that mapped to only one contig, an

(% total)

5636271 (32.73) 9313945 (29.10) of total reads shown in 21456245 (29.59) 24584179 (29.42) Uncounted reads

Percent

Total reads 17220340 72519250 31455052 83553656

be mapped, reads that could

that could not Table 2: Mapping statistics for each flower type. Columns indicate the total number of reads Receptive Seed Receptive Gall Pollinated Seed Pollinated Gall mapped to multiple contigs.

110

3078 3368 3176 5652 15274 Other

20814 19354 20956 20068 81192

Viridiplantae

17 26 25 7045 7113

Hymenoptera

182 202 239 266 889 Fungi

24091 22950 24396 33031 104468 Assigned contigs

32394 32711 33669 45595 144369 Total contigs

assigned to fungi, Hymenoptera, Viridiplantae, and other.

Table 3: Taxonomic placement of assembled contigs using MEGAN for each flower type and total assembly. Columns indicate total number of contigs analyzed, that were assigned taxonomically, and the Receptive Seed Receptive Gall Pollinated Seed Pollinated Gall Total reads 111

Pollinated Gall Receptive Gall

11,851 245 Pollinated Seed 2,792 Receptive Seed 694 208

7,524 1,105 67 88

358 27,939 188 205 304

Supplemental Figure 1 – Venn diagram displaying the number of assembled transcripts shared between samples.

112

Supplemental*Table*1*.*Overrepresented*GO*term*of*up.regulated*genes*in*recetpive*gall*flowers*compared*to* receptive*seed*flowers Description GO-ID Loci0in0GO0term p-value corr0p-value Phenylpropanoid-metabolism phenylpropanoid*metabolic*process 9698 72 5.72E.47 7.01E.44 cellular*aromatic*compound*metabolic*process 6725 91 3.29E.41 2.01E.38 aromatic*compound*biosynthetic*process 19438 72 4.33E.40 1.77E.37 phenylpropanoid*biosynthetic*process 9699 58 7.24E.39 2.22E.36 flavonoid*metabolic*process 9812 50 6.41E.37 1.12E.34 flavonoid*biosynthetic*process 9813 49 2.16E.36 3.30E.34 aromatic*amino*acid*family*metabolic*process 9072 18 2.63E.11 1.61E.09 L.phenylalanine*metabolic*process 6558 5 2.31E.04 3.33E.03 aromatic*compound*catabolic*process 19439 7 6.57E.04 7.89E.03 L.phenylalanine*catabolic*process 6559 4 3.12E.04 4.15E.03 aromatic*amino*acid*family*biosynthetic*process 9073 13 1.09E.08 4.96E.07 aromatic*amino*acid*family*catabolic*process 9074 4 3.12E.04 4.15E.03 Carbohydrate-metbolism carbohydrate*metabolic*process 5975 86 1.22E.11 7.89E.10 polysaccharide*metabolic*process 5976 31 1.01E.08 4.96E.07 carbohydrate*catabolic*process 16052 28 1.61E.07 6.80E.06 monosaccharide*metabolic*process 5996 29 1.86E.07 7.36E.06 polysaccharide*biosynthetic*process 271 18 1.36E.05 3.20E.04 cellulose*metabolic*process 30243 10 2.68E.05 5.49E.04 cellular*carbohydrate*catabolic*process 44275 20 4.98E.05 8.92E.04 rhamnose*biosynthetic*process 19300 4 7.22E.05 1.20E.03 rhamnose*metabolic*process 19299 4 7.22E.05 1.20E.03 UDP.rhamnose*biosynthetic*process 10253 3 1.09E.04 1.76E.03 UDP.rhamnose*metabolic*process 33478 3 1.09E.04 1.76E.03 carbohydrate*biosynthetic*process 16051 25 1.26E.04 2.00E.03 cellular*polysaccharide*metabolic*process 44264 23 1.12E.06 3.91E.05 cellular*carbohydrate*biosynthetic*process 34637 23 1.35E.05 3.20E.04 hexose*metabolic*process 19318 23 1.63E.05 3.64E.04 nucleotide.sugar*metabolic*process 9225 7 6.96E.05 1.19E.03 nucleotide.sugar*biosynthetic*process 9226 5 5.09E.05 8.92E.04 pentose*metabolic*process 19321 7 1.65E.04 2.46E.03 cellulose*biosynthetic*process 30244 8 2.96E.04 4.03E.03 monosaccharide*catabolic*process 46365 14 5.66E.04 6.87E.03 glucose*metabolic*process 6006 18 1.34E.04 2.09E.03 hexose*biosynthetic*process 19319 7 1.16E.03 1.37E.02 glucose*catabolic*process 6007 13 1.35E.03 1.58E.02 polysaccharide*catabolic*process 272 9 1.83E.03 2.02E.02 monosaccharide*biosynthetic*process 46364 7 1.92E.03 2.10E.02 extracellular*polysaccharide*biosynthetic*process 45226 3 4.21E.04 5.38E.03 extracellular*polysaccharide*metabolic*process 46379 3 4.21E.04 5.38E.03 glucan*biosynthetic*process 9250 13 3.71E.05 7.00E.04 glucan*metabolic*process 44042 19 4.48E.06 1.25E.04

113

hexose&catabolic&process 19320 14 5.14E804 6.37E803 cellular&polysaccharide&biosynthetic&process 33692 18 3.83E806 1.12E804 cellular&carbohydrate&metabolic&process 44262 63 2.31E812 1.67E810 cellular&glucan&metabolic&process 6073 17 3.18E805 6.18E804 Cell$wall$ cell&wall¯omolecule&catabolic&process 16998 5 4.92E803 4.71E802 cell&wall¯omolecule&metabolic&process 44036 7 4.55E803 4.39E802 cell&wall&organization 71555 17 1.67E804 2.47E803 cell&wall&organization&or&biogenesis 71554 26 2.77E806 8.50E805 cellular&cell&wall&organization 7047 12 5.70E806 1.49E804 cellular&cell&wall&organization&or&biogenesis 70882 16 2.25E805 4.84E804 primary&cell&wall&biogenesis 9833 2 2.29E803 2.40E802 Amino$acid$metabolism alanine&metabolic&process 6522 2 2.29E803 2.40E802 amine&biosynthetic&process 9309 25 1.24E805 3.03E804 amine&metabolic&process 9308 46 1.44E804 2.18E803 aspartate&metabolic&process 6531 3 4.21E804 5.38E803 cellular&aldehyde&metabolic&process 6081 8 2.96E804 4.03E803 cellular&amine&metabolic&process 44106 42 1.36E804 2.09E803 cellular&amino&acid&and&derivative&metabolic&process 6519 113 2.49E831 3.05E829 cellular&amino&acid&biosynthetic&process 8652 25 2.55E806 8.00E805 cellular&amino&acid&derivative&biosynthetic&process 42398 60 1.19E830 1.33E828 cellular&amino&acid&derivative&metabolic&process 6575 77 4.62E836 6.29E834 cellular&amino&acid&metabolic&process 6520 42 2.87E805 5.68E804 cellular&biosynthetic&process 44249 180 3.54E805 6.78E804 glutamine&biosynthetic&process 6542 5 1.23E805 3.03E804 glutamine&family&amino&acid&biosynthetic&process 9084 7 1.63E803 1.84E802 glutamine&family&amino&acid&metabolic&process 9064 9 3.25E803 3.24E802 glutamine&metabolic&process 6541 7 6.96E805 1.19E803 Other acetyl8CoA&biosynthetic&process 6085 3 3.30E803 3.24E802 alcohol&biosynthetic&process 46165 8 2.71E803 2.77E802 alcohol&catabolic&process 46164 18 2.69E805 5.49E804 alcohol&metabolic&process 6066 36 5.06E807 1.82E805 biosynthetic&process 9058 194 4.49E806 1.25E804 calcium&ion&transport 6816 6 3.66E803 3.56E802 carboxylic&acid&biosynthetic&process 46394 46 5.75E810 3.07E808 carboxylic&acid&catabolic&process 46395 12 2.24E803 2.40E802 carboxylic&acid&metabolic&process 19752 84 1.42E812 1.16E810 catabolic&process 9056 79 2.02E807 7.76E806 cellular&catabolic&process 44248 51 8.70E804 1.04E802 cellular&ketone&metabolic&process 42180 84 4.22E812 2.88E810 cellular&metabolic&process 44237 350 1.45E803 1.68E802 cellular&nitrogen&compound&biosynthetic&process 44271 41 2.55E803 2.65E802 cellular&response&to&hydrogen&peroxide 70301 6 2.55E804 3.60E803 cellular&response&to&oxidative&stress 34599 6 1.79E803 2.00E802

114

cellular'response'to'reactive'oxygen'species 34614 6 1.47E903 1.68E902 chorismate'biosynthetic'process 9423 5 1.43E906 4.60E905 chorismate'metabolic'process 46417 13 1.09E908 4.96E907 cutin'biosynthetic'process 10143 5 5.09E905 8.92E904 dicarboxylic'acid'biosynthetic'process 43650 5 1.43E906 4.60E905 dicarboxylic'acid'metabolic'process 43648 24 6.77E914 6.39E912 external'encapsulating'structure'organization 45229 13 2.73E905 5.49E904 fatty'acid'beta9oxidation 6635 7 1.63E903 1.84E902 fatty'acid'biosynthetic'process 6633 19 1.63E905 3.64E904 fatty'acid'catabolic'process 9062 7 2.61E903 2.69E902 fatty'acid'metabolic'process 6631 25 5.01E906 1.34E904 fatty'acid'oxidation 19395 8 4.33E904 5.42E903 hydrogen'peroxide'catabolic'process 42744 6 1.33E904 2.09E903 hydrogen'peroxide'metabolic'process 42743 6 1.86E904 2.72E903 lignin'biosynthetic'process 9809 9 1.27E906 4.34E905 lignin'metabolic'process 9808 11 3.04E906 9.10E905 lipid'metabolic'process 6629 48 2.85E904 3.97E903 lipid'oxidation 34440 8 4.33E904 5.42E903 malate'metabolic'process 6108 10 3.81E908 1.67E906 metabolic'process 8152 473 5.66E909 2.89E907 monocarboxylic'acid'metabolic'process 32787 37 1.75E907 7.16E906 nitrogen'fixation 9399 5 4.79E906 1.31E904 organic'acid'biosynthetic'process 16053 46 5.75E910 3.07E908 organic'acid'catabolic'process 16054 12 2.24E903 2.40E902 organic'acid'metabolic'process 6082 84 1.52E912 1.17E910 oxidation'reduction 55114 106 8.31E911 4.85E909 oxoacid'metabolic'process 43436 84 1.42E912 1.16E910 oxygen'and'reactive'oxygen'species'metabolic'process 6800 7 3.02E903 3.03E902 primary'metabolic'process 44238 348 3.51E904 4.62E903 primary'shoot'apical'meristem'specification 10072 3 3.30E903 3.24E902 pyruvate'family'amino'acid'metabolic'process 9078 2 2.29E903 2.40E902 response'to'bacterium 9617 15 2.87E903 2.90E902 response'to'cadmium'ion 46686 23 5.61E904 6.87E903 response'to'hydrogen'peroxide 42542 10 4.09E905 7.49E904 response'to'inorganic'substance 10035 37 8.45E906 2.16E904 response'to'metal'ion 10038 27 2.53E904 3.60E903 response'to'oxidative'stress 6979 25 1.74E905 3.81E904 response'to'reactive'oxygen'species 302 12 2.62E905 5.49E904 secondary'metabolic'process 19748 78 2.63E937 5.37E935 small'molecule'biosynthetic'process 44283 124 2.56E930 2.61E928 small'molecule'catabolic'process 44282 39 4.47E907 1.66E905 small'molecule'metabolic'process 44281 212 2.83E938 6.93E936 wax'biosynthetic'process 10025 6 1.51E905 3.49E904 wax'metabolic'process 10166 6 4.05E905 7.49E904

115 ence r 2.80 2.24 0.61 8.53 1.58 3.14 1.44 0.85 1.48 3.44 0.79 4.24 3.13 2.57 3.04 7.44 8.93 4.69 6.27 1.86 6.81 6.61 1.91 1.22 7.30 3.53 6.29 3.87 3.98 1.01 6.06 5.41 3.33 1.73 2.63 1.89 2.66 1.18 0.89 3.75 2.01 5.29 3.90 9.98 3.56 4.35 3.16 10.51 10.76 14.35 31.15 13.31 13.54 40.10 10.68 59.83 10.23 14.45 RG/RS Fold Diffe 1 1 1.03 5.89 7.41 7.21 2.98 9.95 1.85 8.21 2.53 7.04 8.26 4.62 10.14 38. 1 21.31 27.49 21.82 14.90 35.31 14.56 34.12 30.58 23.91 16.86 30.42 31.18 13.93 34.72 37.70 17.32 57.06 37.97 41.91 79.91 44.60 18.86 65.91 26.34 63.32 37.93 54.45 96.12 40.67 1 421.89 234.00 135.15 385.18 448.62 176.52 396.83 636.52 182.97 200.40 185.42 213.64 370.43 149.17 130.57 RPKM Receptive Seed 1 1 14.75 10.41 18.60 15.93 16.50 85.40 68.49 12.71 52.39 79.70 50.09 26.99 30.57 98.51 42.71 65.73 37.12 25.04 51.66 27.08 98.54 70.72 43.65 47.26 181.98 155.44 224. 1 1 1 1 258.82 234.36 369.74 194.44 345.25 988.88 227.47 828.07 744.56 227.54 185.33 349.51 122.66 145.95 796.63 104.91 186.70 421.01 151.09 176.22 132.17 102.75 631.68 134.86 236.68 304.22 587.78 RPKM 1 1 1362.84 2487.43 1233.05 1522.40 Receptive Gall ence r 0.77 1.79 1.87 2.27 2.55 1.89 1.03 2.59 1.65 0.95 1.03 1.89 4.15 3.56 2.94 2.56 4.82 1.63 2.04 2.95 3.24 4.92 3.97 3.83 2.43 2.90 2.76 1.65 2.44 8.15 1.43 5.99 7.95 2.65 2.34 2.05 3.62 4.95 4.04 2.68 4.73 3.03 8.16 1.53 1.94 2.25 1.46 1.06 1.76 1.45 5.38 2.15 2.95 7.29 2.31 2.36 3.05 2.23 PG/PS Fold Diffe 1 1 6.23 3.23 6.19 6.25 12.21 13.27 25. 54.18 67.72 36.49 14.53 74.47 51.99 41.93 40.32 14.96 65.13 45.87 34.30 79.73 70.25 43.89 35.02 52.77 16.87 22.23 32.89 55.16 77.57 15.71 85.39 61.93 23.93 75.88 29.77 37.21 24.98 81.95 22.98 1 1 718.25 174.25 223.75 109.80 359.38 313.31 302.67 403.68 406.76 196.14 261.85 122.84 177.41 278.02 237.17 124.55 174.77 101.75 136.93 RPKM Pollinated Seed 1.66 1 86.42 97.17 69.01 37.56 49.20 43.24 76.03 62.14 74.99 70.13 20.17 85.60 26.28 72.87 75.58 77.10 22.42 25.36 43.23 33.63 80.15 51.25 123.82 122.86 153.80 444.12 230.60 122.75 390.76 167.01 892.19 316.43 170.90 121.01 324.26 101.09 176.75 694.33 251.66 235.37 878.15 147.97 128.17 130.74 120.25 280.80 164.89 133.23 515.16 182.04 189.04 240.22 417.95 RPKM 15 1 1 1346.21 1057.37 1984.47 1557.68 Pollinated Gall - e r 1 1 13 1 1 1 149 1 1.601 1 77.4 538. 248.44 87.8 85. 8 316.235 291.197 155.221 80.1073 950.658 234.572 132.494 680.248 72.4034 120.553 170.244 70.4774 67.0106 82.4185 88.1965 105.916 216.083 148.288 144.436 73.1738 332.798 157.532 903.664 481.871 143.665 82.4185 593.964 50.0618 130.568 459.914 428.713 2180.22 315.464 1838.54 68.5514 73.1738 79.7221 108.997 Bitsco 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 evalue 2.62E- 9.81E- 1.50E- 1.51E- 9.55E-84 1.25E-35 4.80E-13 2.01E-59 1.60E-28 8.22E-26 3.61E-58 5.96E-16 1.08E-09 8.15E-13 2.50E-14 4.91E-16 2.74E-21 7.41E-59 1.89E-33 2.27E-32 6.03E-37 3.31E-32 2.47E-14 3.92E-15 1.97E-28 2.15E-84 6.95E-22 1.61E-13 2.52E-22 2.60E-64 2.00E-105 1.61E-150 2.81E-126 4.12E-139 2.72E-167 6.69E-127 0.00013636 149.1 1 1335.1 1 Q4AE12 Q9FUB7 CBI25555.3 CBI21935.3 BAJ17668.1 AEP17004.1 BAF98296.1 BAF45153.1 BAB87838.1 ADZ54781.1 AEO36936.1 AEO36936.1 ADZ54780.1 AEO36935.1 ADZ54781.1 AEN55616.1 ADB AAB69320.1 ABK93002.1 ACA64737.1 GenBank ID ADO22709.1 ACM69363.1 ACM69363.1 XP_00354 XP_002263617.1 XP_002300066.1 XP_003635302.1 XP_002277375.1 XP_002337752.1 XP_002531676.1 XP_002531676.1 XP_002337752.1 XP_002285634.1 XP_002319582.1 XP_002279162.2 XP_002282867.1 XP_002297654.1 XP_002328073.1 XP_002513485.1 XP_002308562.1 XP_002337752.1 XP_003562680.1 XP_003544220.1 XP_002515664.1 itis vinifera] V itis vinifera] V itis vinifera] V itis vinifera] V itis vinifera] V itis vinifera] itis vinifera] orenia hybrida] V V T hydrolase-like protein 5-like [ A itis vinifera] V iola stagnina] V AltName: Full=Naringenin-chalcone synthase

ligase [Humulus lupulus] ligase [Humulus lupulus] itis vinifera] A A V O-methyltransferase [ A TE transporter [Malus x domestica] A feoyl-Co f Protein SRG1 [ Predicted protein [Populus trichocarpa] Plastidic 3-deoxy-D-arabino-heptulosonate 7-phosphate synthase 2 [Petroselinum crispum] Phospho-2-dehydro-3-deoxyheptonate aldolase 1, chloroplastic [ 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase [Hevea brasiliensis] Probable cinnamyl alcohol dehydrogenase 1-like [ Unnamed protein product [ Calreticulin-3-like [Glycine max] Chalcone synthase [Populus trichocarpa] Flavonoid 3'-hydroxylase [ Chalcone isomerase [Prunus avium] 4-coumarate:Co CHI [Canarium album] CHI [Canarium album] 4-coumarate:Co Naringenin-chalcone synthase [Prunus avium] Chalcone synthase 3 [ BLASTx annotation Phospho-2-dehydro-3-deoxyheptonate aldolase 1 [Ricinus communis] Phospho-2-dehydro-3-deoxyheptonate aldolase 1 [Ricinus communis] phi class glutathione transferase GSTF7 [Populus trichocarpa] Chalcone synthase [Populus trichocarpa] TT12-1 M unknown [Populus trichocarpa] Dihydroflavonol 4-reductase [Humulus lupulus] Flavonoid 3' hydroxylase [Gynura bicolor] Probable rhamnose biosynthetic enzyme 1 [ Dehydroquinate dehydratase/ shikimate dehydrogenase [Populus trichocarpa] 3-hydroxyisobutyryl-Co Chalcone synthase [ Chalcone synthase; F3H [Canarium album] Chalcone isomerase [Prunus avium] Ca Unnamed protein product [ 10-formyltetrahydrofolate synthetase [Populus trichocarpa] Arogenate/prephenate dehydratase [Populus trichocarpa] Xanthine dehydrogenase, putative [Ricinus communis] Chs2 [Morus alba] Precursor of carboxylase p-protein 1 [Populus trichocarpa] Chalcone--flavonone isomerase 2; Short=Chalcone 2 Chalcone synthase [Populus trichocarpa] Phospho-2-dehydro-3-deoxyheptonate aldolase 2 [Brachypodium distachyon] Phospho-2-dehydro-3-deoxyheptonate aldolase 1 [Glycine max] Chitinase, putative [Ricinus communis] able 2: Loci assigned to overrepresented GO terms in receptive and pollinated gall flowers dealing with chalcone synthase T

1 17 11 1 134 136 1 1554 1777 1412 1924 1056 11 1 1 1 1 1 Locus 2057 Locus 20900 Locus 20995 Locus 20932 Locus 21005 Locus 2104 Locus 21042 Locus 2 Locus 21346 Locus 2166 Locus 22064 Locus 22416 Locus 2261 Locus 22631 Locus 22685 Locus 2275 Locus 28403 Locus 23325 Locus 23369 Locus 23847 Locus 24923 Locus 28354 Locus 24375 Locus 26800 Supplemental Locus Locus 1065 Locus Locus Locus Locus Locus Locus Locus 1246 Locus 1263 Locus 13079 Locus 1343 Locus 15569 Locus 1529 Locus 16368 Locus 14128 Locus 1491 Locus 1517 Locus 15251 Locus 16823 Locus 16895 Locus 17 Locus 17579 Locus 17756 Locus 17877 Locus 18547 Locus 1857 Locus 19 Locus 1927 Locus 19337 Locus 19473 Locus 19609 Locus 20082 Locus 20250 Locus 20492 116 5.23 5.25 2.37 0.40 5.03 4.39 5.44 1.24 1.64 9.73 9.64 2.05 3.79 4.07 1.66 5.10 3.86 1.69 1.54 0.44 2.12 5.13 4.69 5.36 6.85 1.12 1.85 3.81 1.27 2.20 3.90 3.49 3.77 3.40 2.83 1.81 2.14 2.05 2.50 9.09 2.94 2.76 2.66 1.90 2.49 0.71 7.42 6.24 1.26 5.34 2.62 8.32 3.50 36.72 45.66 13.96 56.34 49.69 17.32 28.26 10.93 1.15 1.77 1.37 3.90 3.41 2.09 6.20 8.69 0.51 7.97 3.56 18.55 13.38 1 1 1 26.51 20.10 21.71 22.33 26.86 34.67 79.71 19.29 53.16 21.81 32.85 81.20 84.48 12.57 13.47 32.05 18.24 66.72 32.92 49.24 18.26 28.78 29.04 70.23 1 1 109.54 585.34 283.84 506.20 558.01 528.55 123.06 493.29 105.40 264.57 595.47 181.92 299.12 101.27 846.98 238.70 951.43 517.33 717.90 122.02 567.37 781.17 124.26 1844.33 1 1 1.95 7.51 0.65 12.54 88.29 37.59 57.67 19.94 29.63 47.85 40.80 63.69 93.17 53.64 36.14 45.98 54.24 76.10 66.28 11 1 557. 138.62 622.68 259.58 232.19 797.20 121.47 630.17 917.10 108.44 466.67 307.80 262.85 852.94 176.14 555.91 334.61 186.95 669.47 251.28 385.24 432.77 100.82 103.32 325.12 153.55 100.55 245.95 1428.82 1226.37 1085.69 2006.91 1470.82 1348.03 4759.66 3303.22 2034.46 1291.81 1978.81 3510.50 1413.57 1357.65 1.59 4.72 3.04 3.73 1.26 4.54 2.58 3.36 1.43 1.34 0.72 1.72 4.74 2.90 2.76 2.95 3.33 6.70 3.32 1.45 8.46 2.25 1.32 1.34 2.00 3.09 2.61 2.64 1.39 1.81 3.93 0.85 3.85 2.22 5.01 2.82 0.94 2.42 4.20 0.98 2.62 2.90 6.01 3.40 2.42 1.84 4.67 5.03 3.47 4.28 1.44 1.36 1.37 1.35 4.45 1.36 8.18 2.56 4.51 1.93 1 48.73 4.95 6.98 5.27 0.50 6.25 37.42 79.43 32.51 73.93 36.25 84.64 24.18 25.81 74.31 39.70 68.39 83.21 33.01 62.40 44.53 34.65 74.69 87.65 15.39 26.58 15.89 88.29 16.55 21.30 33.14 17.02 49.97 23.42 139.50 201.02 432.15 237.39 108.55 447.62 936.65 423.92 203.40 674.44 232.39 399.99 137.22 399.42 196.99 173.87 325.26 459.37 123.02 145.42 856.20 341.03 464.60 324.99 769.47 199.14 126.36 1 1239.81 1.80 1 83.82 62.23 23.47 59.04 58.14 99.85 79.33 27.92 98.99 26.38 32.61 86.24 92.45 64.41 29.22 23.77 29.12 44.90 24.28 75.70 67.86 51.13 59.93 159.28 6 176.73 296.12 544.54 365.24 105.94 599.26 673.44 981.30 560.45 773.85 162.10 198.34 528.12 608.57 178.78 459.88 452.00 150.53 240.45 180.57 516.33 380.27 443.84 898.07 243.77 1 1077.57 1229.75 1989.21 1326.13 1807.18 3497.22 2483.49 2171.61 3949.71 1392.39 1048.89 1 1 77.4 306.99 52.373 73.559 74.3294 76.6406 73.1738 684.871 107.457 980.704 150.599 55.4546 107.842 84.7297 64.3142 101.293 69.3218 196.823 508.834 97.4413 417.927 56.9954 83.1889 70.0922 747.273 87.0409 1097.42 89.7373 610.527 55.0694 72.4034 74.7146 64.6994 488.419 70.4774 588.571 1207.97 96.2857 1063.52 90.5077 490.345 385.571 843.573 1 1 1 1 1 14 1 1 1 1 1 1 0 0 0 0 0 0 0 1.49E- 3.03E- 2.57E- 9.75E- 1.15E- 6.73E-12 1.36E-12 8.07E-13 7.24E-22 7.39E-35 2.03E-81 6.78E-23 5.00E-15 7.12E-09 5.19E-20 2.17E-10 9.16E-49 7.62E-19 2.74E-05 3.14E-06 1.49E-14 1.29E-10 1.01E-15 1.56E-16 4.27E-06 5.17E-12 7.73E-12 1.65E-18 9.16E-17 3.69E- 2.34E-141 3.10E-172 2.45E-135 1.99E-165 2.45E-136 8.30E-105 1339.1 1 1243.1 1 T75320.2 A AEF32085.1 AEE81750.1 A ADF59061.1 ADP37945.1 ABE03772.1 AEO36936.1 BAC66467.1 AEO96985.1 ADZ54781.1 AEO36936.1 CAC86996.1 AEX32790.1 ADZ54781.1 AEN55617.1 AEN55617.1 ABB83366.1 AAC60566.1 ACV72638.1 ACN54324.1 ADD74169.1 ACM17224.1 ACM69363.1 ACM69363.1 ACM69363.1 ABW03086.1 ADW XP_0036 XP_002337752.1 XP_002328402.1 XP_003543185.1 XP_002520564.1 XP_002882452.1 XP_002298158.1 XP_002312686.1 XP_002314377.1 XP_003615153.1 XP_002512862.1 XP_003635302.1 XP_002337752.1 XP_003538625.1 XP_002337752.1 XP_002324426.1 itis vinifera] V . multicaulis] r [Medicago truncatula] oxidase [Malus x domestica] itis vinifera] A A V oxidase 1 A ligase [Humulus lupulus] ligase [Humulus lupulus] ligase [Humulus lupulus] A A A 3-O-methyltransferase [Boehmeria nivea] synthase subunit C, putative [Ricinus communis] A P T A

citrate lyase b-subunit [Lupinus albus] feoyl-Co P f T acuolar ABC transporter D family member 1-like isoform 2 [Glycine max] Chalcone synthase [Populus trichocarpa] CHI [Canarium album] Naringenin-chalcone synthase [Humulus lupulus] Chalcone synthase [Rosa hybrid cultivar 'Kardinal'] Dehydroquinate dehydratase/ shikimate dehydrogenase [Populus trichocarpa] Phenylalanine ammonia lyase [Morus alba va Probable rhamnose biosynthetic enzyme 1-like [Glycine max] Ca Proline-rich SAC51 [Brassica napus] Fumarylacetoacetate hydrolase, putative [Ricinus communis] Chalcone synthase [Nelumbo nucifera] Chalcone synthase [Gossypium hirsutum] Chalcone isomerase [Prunus avium] Phenylalanine ammonia-lyase 1 precursor [Pyrus x bretschneideri] CHI [Canarium album] 4-coumarate:Co TMS membrane family protein [Arabidopsis lyrata subsp. lyrata] A Phenylalanine ammonia-lyase [ Predicted protein [Populus trichocarpa] Chalcone isomerase [Fragaria chiloensis] Predicted protein [Populus trichocarpa] Chalcone synthase 2 [Camellia chekiangoleosa] Ent-kaurenoic acid oxidase [Castanea mollissima] Phenylalanine ammonia-lyase [Arnebia euchroma] Predicted protein [Populus trichocarpa] 2-dehydro-3-deoxyphosphoheptonate aldolase [Medicago truncatula] V Chalcone isomerase [Prunus avium] Chs3 [Morus alba] Chs3 [Morus alba] Phospho-2-dehydro-3-deoxyheptonate aldolase 1, chloroplastic [ 1-aminocyclopropane 1-carboxylate synthase 5 [Prunus salicina] Chalcone synthase [Populus trichocarpa] Peroxisomal acyl-Co 4-coumarate:Co Chalcone synthase [Populus trichocarpa] Predicted protein [Populus trichocarpa] 4-coumarate:Co Putative medium chain acyl-Co Chalcone synthase 1 [Gossypium hirsutum] NADPH:cytochrome P450 reductase [Gossypium hirsutum] 13 145 1 1

Locus 28447 Locus 28691 Locus 29140 Locus 29380 Locus 29475 Locus 3035 Locus 30420 Locus 3057 Locus 30631 Locus 30985 Locus 31238 Locus 31949 Locus 32054 Locus 32138 Locus 32158 Locus 32226 Locus 33084 Locus 33478 Locus 33610 Locus 338 Locus 34501 Locus 3528 Locus 35336 Locus 3593 Locus 35981 Locus 36926 Locus 3788 Locus 38283 Locus 40863 Locus 41275 Locus 42519 Locus 428 Locus 43476 Locus 4367 Locus 4457 Locus 45660 Locus 4569 Locus 47098 Locus 47175 Locus 4781 Locus 48787 Locus 48803 Locus 49241 Locus 49319 Locus 49650 Locus 4984 Locus 49990 Locus 50535 Locus 50635 Locus 51016 Locus 5 Locus 5 Locus 5167 Locus 51955 Locus 5303 Locus 53457 Locus 5371 Locus 53822 Locus 5410 Locus 604 Locus 621 117

1.61 8.09 2.44 2.17 3.09 3.03 9.48 2.69 68.17 198.50 3.70 0.63 0.18 6.14 72.77 18.54 53.36 73.76 59.28 136.89 1.95 17.08 34.83 58.20 1 150.05 252.22 130.25 160.24 179.45 368.18 2.34 3.28 3.99 1.67 6.86 1.62 3.89 3.02 1.25 13.87 7.29 4.20 72.92 38.99 14.39 44.19 83.54 75.25 14.38 155.00 49.99 16.35 43.48 170.35 127.71 199.64 176.18 139.83 121.82 193.64 17.087 103.99 1 397.512 468.389 417.157 100.908 985.712 168.703 14 1 0 1E-39 1 3.16E-20 7.86E-21 2.02E-24 4.63E- 4.66E-120 3.55E-129 3. T37172.1 A A ABZ01817.1 ADD74168.1 NP_973996.1 AAZ29732.1 XP_002520028.1 XP_002529571.1 XP_002520425.1 , putative [Ricinus communis] r rifolium pratense] T ligase, putative [Ricinus communis] A dependent Co P feoyl-CoA-O-methyltransferase [Broussonetia papyrifera] f Putative 5-enolpyruvylshikimate-3-phosphate synthase [Arabidopsis thaliana] Chalcone synthase [Nelumbo nucifera] AM Chorismate synthase, chloroplast precurso Ca Phenylalanine ammonia lyase [ Laccase, putative [Ricinus communis] Cinnamyl alcohol dehydrogenase [Gossypium hirsutum]

Locus 6937 Locus 74 Locus 7415 Locus 769 Locus 869 Locus 878 Locus 9207 Locus 93 Locus 9536 Locus 9537

118 1.20 6.73 9.20 9.16 9.82 8.27 9.59 7.74 8.35 7.55 9.31 6.74 6.74 6.83 9.58 6.98 9.57 9.74 6.86 7.39 6.94 7.67 6.81 7.61 6.79 8.29 8.36 9.60 1 12.35 13.82 14.23 21.74 10.08 12.23 12.02 12.67 17.17 16.37 14.37 21.70 10.47 16.67 49.63 19.07 RPKM eads r 16.00 18.00 17.00 98.00 71.00 96.00 86.00 75.00 65.00 91.00 87.00 65.00 68.00 97.00 47.00 64.00 47.00 68.00 72.00 1 1 1 205.00 122.00 187.00 159.00 151.00 161.00 423.00 595.00 159.00 108.00 174.00 136.00 142.00 260.00 234.00 339.00 196.00 205.00 694.00 171.00 314.00 370.00 184.00 914.00 126.00 otal gene T eads r 13.00 15.00 88.00 71.00 91.00 86.00 66.00 65.00 91.00 85.00 65.00 68.00 94.00 47.00 64.00 47.00 67.00 72.00 1 1 176.00 121.00 148.00 124.00 142.00 151.00 363.00 562.00 105.00 155.00 108.00 149.00 135.00 142.00 236.00 220.00 335.00 183.00 205.00 594.00 170.00 305.00 355.00 180.00 914.00 109.00 Unique gene 325 355 398 340 214 321 582 536 232 237 138 243 189 373 176 223 269 286 218 189 162 746 186 244 133 280 462 394 194 185 257 320 120 369 184 146 121 189 196 170 274 257 1959 1598 Gene length e r 1 1 17.86 12.08 85. 52.37 52.37 52.76 90.51 66.24 73.94 73.94 48.91 62.77 50.83 82.42 78.18 49.29 90.51 63.93 65.47 76.64 62.77 96.67 73.56 57.77 73.17 63.16 80.88 62.00 49.68 49.29 78.95 48.91 62.39 55.84 1 1 130.95 157.92 105.53 109.38 102.83 104.76 126.72 544.27 386.73 128.64 Bitsco 1 1 1 1 0 e value 1.18E- 1.54E- 2.79E-05 6.14E-29 4.75E-37 2.73E-05 2.10E-05 9.30E-17 8.14E-25 2.18E-09 8.96E-12 8.90E-12 2.04E-08 3.92E-15 2.79E-21 7.99E-05 2.54E-14 1.92E-22 1.38E-21 4.64E-21 4.79E-13 5.75E-29 9.40E-09 3.14E-09 1.36E-12 2.03E-08 1.27E-18 1.18E-27 6.74E-07 2.99E-23 1.58E-08 7.18E-14 3.57E-08 2.82E-13 2.67E-08 2.50E-06 3.09E-28 1.08E-136 0.000303787 0.000236332 0.000183805 0.000236189 0.000308989 itis vinifera) V itis vinifera) itis vinifera) V V itis vinifera) itis vinifera) V V itis vinifera) V itis vinifera) itis vinifera) itis vinifera) itis vinifera) itis vinifera) V V V V V 19 ( At3g14460 ( At3g14460-like (Glycine max) At3g14460-like ( At3g14460-like ( itis vinifera) itis vinifera) 1

V V -like serine/threonine-protein kinase ( r -like protein kinase 26-like (Glycine max) r VITISV_003723 ( VITISV_025268 ( VITISV_027 VITISV_036507 (

-nbs-lrr resistance protein (Populus trichocarpa) -nbs-lrr resistance protein (Populus trichocarpa) r r resistance protein N-like ( resistance protein N-like ( V V -nbs-lrr resistance protein (Populus trichocarpa) r i T TM TM AF402744_1 NBS/LRR resistance protein-like protein (Theobroma cacao) BLASTx annotation BED finge BED finge Beta-glucosidase 12-like (Glycine max) Cc-nbs resistance protein (Medicago truncatula) Cc-nbs resistance protein (Medicago truncatula) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Cc-nbs-lrr resistance protein (Populus trichocarpa) Disease resistance protein I-2 (Medicago truncatula) Disease resistance protein RGA2, putative (Ricinus communis) Disease resistance protein RGA2, putative (Ricinus communis) Disease resistance protein RGA2, putative (Ricinus communis) Disease resistance protein RPM1-like ( NBS-containing resistance-like protein (Medicago truncatula) NBS-LRR resistance protein-like protein (Gossypium hirsutum) Putative CC-NBS-LRR resistance protein (Malus x domestica) Putative disease resistance protein Putative disease resistance protein Putative disease resistance protein Putative disease resistance protein Putative disease resistance RPP13-like protein 1-like ( Putative disease resistance RPP13-like protein 1-like ( Resistance protein (Medicago truncatula) Conserved hypothetical protein (Ricinus communis) Conserved hypothetical protein (Ricinus communis) Conserved hypothetical protein (Ricinus communis) Conserved hypothetical protein (Ricinus communis) Cysteine-rich recepto Dihydrofolate reductase-like (Brachypodium distachyon) Hypothetical protein RCOM_0335820 (Ricinus communis) Hypothetical protein Hypothetical protein Hypothetical protein Hypothetical protein Predicted protein (Populus trichocarpa) Probable LRR recepto able 3: Unique genes to receptive gall flowers T

14 1 Supplemental Locus Locus 32568 Locus 32340 Locus 22870 Locus 22480 Locus 51850 Locus 20709 Locus 23510 Locus 28697 Locus 35780 Locus 42599 Locus 43641 Locus 44710 Locus 48348 Locus 40546 Locus 49176 Locus 50817 Locus 38914 Locus 35628 Locus 46640 Locus 50208 Locus 47824 Locus 15429 Locus 25069 Locus 49416 Locus 45797 Locus 48034 Locus 41981 Locus 15027 Locus 47665 Locus 38616 Locus 46403 Locus 34387 Locus 45571 Locus 51739 Locus 5806 Locus 36945 Locus 4679 Locus 49038 Locus 50422 Locus 40056 Locus 10 Locus 52191 Locus 43590 Locus 9133

119 1 1 1.00 6.77 7.93 7.60 9.73 7.50 9.71 7.48 7.57 9.57 7.05 9.54 6.73 9.50 8.15 7.69 9.55 7.85 7.01 6.76 8.15 7.94 7.93 9.40 9.97 1 10. 24.29 20.87 16.00 13.87 10.60 28.36 10.23 25.24 19.58 13.40 10.65 27.34 10.01 12.08 89.07 13.47 16.37 22.62 25.69 13.86 1.00 17.00 19.00 75.00 90.00 72.00 68.00 75.00 98.00 96.00 72.00 83.00 60.00 98.00 11 1 1 537.00 134.00 179.00 286.00 403.00 157.00 320.00 158.00 129.00 227.00 442.00 105.00 239.00 146.00 143.00 212.00 521.00 204.00 476.00 121.00 160.00 124.00 896.00 129.00 106.00 124.00 234.00 216.00 261.00 286.00 138.00 143.00 17.00 14.00 19.00 75.00 80.00 71.00 68.00 75.00 98.00 86.00 72.00 83.00 57.00 98.00 1 1 1 537.00 125.00 174.00 276.00 403.00 143.00 320.00 158.00 124.00 227.00 442.00 104.00 191.00 139.00 142.00 212.00 521.00 204.00 476.00 121.00 142.00 104.00 896.00 106.00 123.00 234.00 216.00 235.00 105.00 286.00 109.00 139.00 433 217 331 168 350 569 232 290 221 318 188 247 458 343 275 239 176 260 209 434 375 154 285 198 173 341 229 308 328 201 197 322 173 232 307 298 280 148 242 187 199 215 596 271 202 1447 91.28 69.71 52.37 52.37 78.95 65.47 71.63 108.61 100.91 146.75 170.63 1 1 4.50E- 3.25E-22 5.30E-17 1.52E-10 2.77E-05 2.80E-05 9.62E-20 8.00E-17 1.10E-33 3.23E-09 7.10E-41 itis vinifera) itis vinifera) V V TFIID subunit 5-like isoform 2 (

itis vinifera) itis vinifera) itis vinifera) V V V -like protein (Malus x domestica) r ranscription initiation factor ype II inositol-1,4,5-trisphosphate 5-phosphatase FRA3-like (Glycine max) ype II inositol-1,4,5-trisphosphate 5-phosphatase FRA3-like (Glycine max) T T T Protein SUPPRESSOR OF npr1-1, CONSTITUTIVE 1-like ( Recepto RGC2-like protein (Helianthus annuus) Rpp4 candidate 3 (Glycine max) Rpp4C5 (Phaseolus vulgaris) Unnamed protein product ( Unnamed protein product ( Unnamed protein product ( No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit 1 1 Locus 40324 Locus 45217 Locus 18017 Locus 43400 Locus 7337 Locus 7533 Locus 45508 Locus 49630 Locus 33848 Locus 50339 Locus 51257 Locus 142 Locus 14582 Locus 18779 Locus 21553 Locus 22438 Locus 23164 Locus 23366 Locus 23401 Locus 23513 Locus 24497 Locus 28355 Locus 30069 Locus 30587 Locus 32152 Locus 32575 Locus 35136 Locus 35285 Locus 35746 Locus 36064 Locus 36839 Locus 37108 Locus 38137 Locus 41808 Locus 41809 Locus 42081 Locus 42635 Locus 43298 Locus 43516 Locus 4447 Locus 45024 Locus 45309 Locus 45517 Locus 45598 Locus 45599 Locus 47122

120 9.17 8.96 7.00 9.33 7.42 9.36 8.81 8.84 9.12 10.53 16.34 13.50 19.73 10.08 12.13 13.81 49.49 13.75 18.00 70.00 74.00 97.00 87.00 72.00 51.00 73.00 1 121.00 136.00 153.00 140.00 120.00 173.00 135.00 122.00 122.00 417.00 19.00 18.00 70.00 74.00 97.00 87.00 69.00 51.00 73.00 1 1 108.00 121.00 136.00 120.00 140.00 120.00 173.00 131.00 417.00 13 1 252 225 163 153 222 207 139 252 256 182 336 218 160 173 262 165 104 No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit No hit

Locus 47190 Locus 47568 Locus 47712 Locus 47793 Locus 48733 Locus 48805 Locus 48807 Locus 49003 Locus 49323 Locus 49529 Locus 50107 Locus 50562 Locus 50802 Locus 52037 Locus 52336 Locus 52656 Locus 52753 Locus 54587 121 1.86 6.06 4.02 1.86 2.73 3.23 3.88 5.13 5.44 1.90 7.04 2.86 2.72 7.68 2.55 1.60 2.03 3.95 2.41 1.59 1.53 2.62 8.82 4.10 1.58 3.64 3.56 1.45 2.34 1.50 4.36 1.85 1.42 6.24 1.74 2.17 5.47 1.44 3.14 1 12.33 16.63 14.45 20.79 12.15 Fold change RPKM 1.05 r 1 9.76 8.68 12.80 17.32 24.52 39.73 57.28 36.15 28.98 21.81 22.33 62.30 17.75 42.78 49.42 17.46 12.08 55.14 87.16 37.32 59.19 99.84 75.53 61.00 40.67 49.71 90.38 19.14 32.62 43.71 50.21 47.26 38.99 1 4 192.66 379.93 212.48 279.55 101.27 242.62 158.33 105.85 135.15 259.58 flowe Receptive seed RPKM 1.95 r 12.35 18.15 16.77 98.57 11 1 1 1 104.91 108.58 106.26 121.47 120.37 125.06 122.27 134.06 134.42 144.44 143.27 140.40 139.55 147.27 142.77 158.92 153.18 172.41 159.81 790.73 587.78 176.98 797.06 651.23 550.63 497.36 418.89 397.78 396.38 275.54 229.71 194.44 190.56 186.95 182.76 345.67 295.12 213.29 816.31 Receptive gall flowe e r 12.46 97.44 71.25 83.19 1 903.66 518.85 427.17 603.59 921.77 720.69 980.70 291.58 290.43 647.89 454.14 185.65 769.23 438.73 280.03 491.12 567.77 492.66 829.71 155.99 156.38 248.44 548.90 127.49 294.66 280.41 558.53 912.91 171.01 360.15 234.57 150.60 196.05 223.79 744.96 624.78 337.04 2881.28 1097.03 1539.63 Bitsco 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 evalue 1545.1 1 V70659.1 P26518 P50218 P42896 Q9SSV4 A CBI39460.3 BAF42036.1 A CAC86996.1 CAB96173.1 ACB97677.1 AAR88248.1 ABV32545.1 ABG76000.1 CAN69517.1 AAO17294.1 ADO24300.1 ADN33958.1 ADY68846.1 GenBank ID XP_0023 XP_002273962.2 XP_002285634.1 XP_003525377.1 XP_002283298.2 XP_003543185.1 XP_002509687.1 XP_002326407.1 XP_002302264.1 XP_002519229.1 XP_002464260.1 XP_002520852.1 XP_002282575.2 XP_002306099.1 XP_003543564.1 XP_002270550.1 XP_003524918.1 XP_002515664.1 XP_003540213.1 XP_003546526.1 XP_002306893.1 XP_002319953.1 XP_002277375.1 XP_002312686.1 XP_002315071.1 XP_002297881.1 itis vinifera] V ghum bicolor] r itis vinifera] V itis vinifera] itis vinifera] V V ArnA-like [Glycine max]

itis vinifera] V itis vinifera] V itis vinifera] V itis vinifera] V VITISV_024155 [ catalytic subunit 1 [UDP-forming]-like [

A AltName: Full=2-phospho-D-glycerate hydro-lyase

Annotation

citrate lyase b-subunit [Lupinus albus] P T TP-citrate synthase, putative [Ricinus communis] Galacturonosyltransferase 8-like [ Probable rhamnose biosynthetic enzyme 1 [ A Probable polygalacturonase-like [Glycine max] Chitinase [Ficus carica] Cellulase1 [Pyrus communis] Mitochondrial citrate synthase precursor [Citrus junos] Callose synthase 3-like [ Probable rhamnose biosynthetic enzyme 1-like [Glycine max] NAD dependent epimerase/dehydratase, putative [Ricinus communis] Xyloglucan endotransglucosylase/hydrolase [Gossypium hirsutum] Predicted protein [Populus trichocarpa] Predicted protein [Populus trichocarpa] A Hypothetical protein SORBIDRAFT_01g015090 [So Alpha-galactosidase/alpha-n-acetylgalactosaminidase, putative [Ricinus communis] Predicted protein [Populus trichocarpa] GRP-like protein 2 [Gossypium hirsutum] Beta-galactosidase protein 2 [Prunus persica] Cellulose synthase Predicted protein [Populus trichocarpa] Fructokinase-2-like [Glycine max] Putative 3,4-dihydroxy-2-butanone kinase [ Brassinosteroid-regulated protein BRU1-like [Glycine max] Glyceraldehyde-3-phosphate dehydrogenase, cytosolic Enolase [Spinacia oleracea] Chitinase, putative [Ricinus communis] Sucrose synthase-like [Glycine max] Predicted protein [Populus trichocarpa] Bifunctional polymyxin resistance protein Glyceraldehyde-3-phosphate dehydrogenase [Cucumis melo subsp. melo] Inositol-3-phosphate synthase Isocitrate dehydrogenase [NADP] Glyceraldehyde-3-phosphate dehydrogenase [Musa acuminata] Predicted protein [Populus trichocarpa] Granule bound starch synthase Ia precursor [Malus x domestica] Probable cinnamyl alcohol dehydrogenase 1-like [ Unnamed protein product [ Predicted protein [Populus trichocarpa] Predicted protein [Populus trichocarpa] Predicted protein [Populus trichocarpa] Sucrose synthase [Gossypium herbaceum subsp. africanum] Blastx Enolase; Hypothetical protein able 4: Loci assigned to overrepresented GO terms in receptive gall flowers dealing with carbohydrate metabolism T

173 1 Locus 318 Locus 1343 Locus 1438 Locus 35981 Locus 2720 Locus 926 Locus 2125 Locus 7808 Locus 539 Locus 3057 Locus 1655 Locus 1780 Locus 135 Locus 252 Locus 20842 Locus 35027 Locus 1609 Locus 776 Locus 124 Locus 3771 Locus 867 Locus 36 Locus 200 Locus 24930 Locus 22561 Locus 21023 Locus 20250 Locus 31061 Locus 14129 Locus 23522 Locus 8139 Locus Locus 41275 Locus 40144 Locus 17937 Locus 49832 Locus 588 Locus 299 Locus 21042 Locus 16739 Locus 12269 Locus 5249 Locus 17818 Supplemental Locus Locus 1616

122 3.27 3.01 3.12 2.24 3.49 7.54 2.55 4.45 2.18 2.12 2.49 2.53 2.15 2.65 3.33 7.70 6.55 5.21 2.46 3.14 2.77 6.29 3.63 2.45 4.86 4.54 5.82 5.17 5.09 7.01 2.95 3.64 3.70 4.17 4.22 3.78 6.42 7.14 9.16 18.40 10.83 13.40 10.01 15.45 1.52 1.13 1.43 1.28 1.05 5.34 7.43 9.47 7.87 3.52 4.21 6.22 5.36 3.91 2.21 1 1 1 1 1 28.90 30.80 29.29 40.51 26.03 33.85 19.39 39.56 40.32 31.73 31.12 36.19 29.34 23.18 13.53 28.23 21.82 24.49 10.32 17.76 26.26 12.91 13.80 10.35 18.53 13.73 13.48 10.52 10.30 1 1 34. 98.17 94.58 92.60 91.26 90.87 90.74 86.86 86.44 86.32 86.15 85.67 80.46 78.91 78.73 77.77 77.66 77.16 72.93 72.88 70.47 69.31 68.49 67.77 64.89 64.45 64.32 62.75 62.66 60.21 59.06 57.41 55.22 54.62 49.94 49.84 47.15 43.85 43.52 42.15 41.75 39.98 38.23 35.86 1 1 1.89 1 73.17 77.80 172.53 132. 396.74 103.61 444.12 201.45 603.98 539.27 260.38 749.58 363.61 441.43 800.43 790.80 474.94 654.06 463.77 950.66 339.73 526.55 760.37 222.25 759.21 825.47 608.22 142.90 161.00 269.63 158.69 323.55 336.65 202.60 1 15 3147.07 1009.21 1853.95 1497.26 1547.33 1509.97 1029.62 1513.05 1262.28 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 P26518 CBI16442.3 ACJ38665.1 ACF04280.1 ACF04279.1 BAF98296.1 ABL86685.1 AAP97437.1 BAE72075.1 AAP51059.1 ABR45722.1 CAN66037.1 CAN76048.1 ABV32548.1 ABV32547.1 ADG27841.1 AAK51690.1 XP_002523086.1 XP_003525238.1 XP_002301488.1 XP_002528124.1 XP_002282083.1 XP_002270163.1 XP_003526507.1 XP_002317160.1 XP_002302169.1 XP_002307145.1 XP_002325849.1 XP_002323270.1 XP_002520066.1 XP_002269610.2 XP_003549495.1 XP_002283274.1 XP_002302349.1 XP_002280915.2 XP_002298432.1 XP_002280521.1 XP_002312098.1 XP_002531372.1 XP_003538061.1 XP_002325031.1 XP_002514769.1 XP_002326337.1 XP_003539381.1 itis vinifera] V itis vinifera] V itis vinifera] V itis vinifera] itis vinifera] V V 1 [ 1 itis vinifera] V itis vinifera] V itis vinifera] itis vinifera] V V VITISV_030300 [ VITISV_0377 catalytic subunit 8 [UDP-forming] [

A reductase, putative [Ricinus communis] A T3 [Gossypium barbadense] L ransferase, transferring glycosyl groups, putative [Ricinus communis] riosephosphate isomerase [Gossypium hirsutum] Dtdp-glucose 4-6-dehydratase, putative [Ricinus communis] Glyceraldehyde-3-phosphate dehydrogenase, cytosolic Hypothetical protein C Phosphoglucomutase, cytoplasmic-like [Glycine max] Myo-inositol oxygenase [Eucalyptus grandis] Predicted protein [Populus trichocarpa] Hypothetical protein T Xylosyltransferase 1 [ UDP-N-acetylglucosamine--peptide N-acetylglucosaminyltransferase SEC [ Endo-1,4-beta-mannosidase protein 2 [Prunus persica] UDP-D-glucuronate carboxy-lyase [Eucalyptus grandis] Alpha-L-arabinofuranosidase [Malus x domestica] 6-phosphofructokinase 3-like [Glycine max] Predicted protein [Populus trichocarpa] Cellulose synthase [Populus trichocarpa] Hypothetical protein POPTRDRAFT_760228 [Populus trichocarpa] Unnamed protein product [ Predicted protein [Populus trichocarpa] T 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase [Hevea brasiliensis] Endo-1,4-beta-mannosidase protein 1 [Prunus persica] G6PD1 [Actinidia chinensis] Predicted protein [Populus trichocarpa] Cellulose synthase [Betula luminifera] Cinnamoyl-Co Cellulose synthase Xylose isomerase-like [Glycine max] 6-phosphofructokinase 3-like [ Predicted protein [Populus trichocarpa] L-arabinokinase-like [ Cellulose synthase [Populus trichocarpa] Dolichyl-diphosphooligosaccharide--protein glycosyltransferase subunit STT3 [ Predicted protein [Populus trichocarpa] Dtdp-glucose 4-6-dehydratase, putative [Ricinus communis] Beta-glucosidase 42-like [Glycine max] Pear beta-galactosidase3 [Pyrus communis] Latex cyanogenic beta glucosidase [Hevea brasiliensis] Predicted protein [Populus trichocarpa] Beta-hexosaminidase, putative [Ricinus communis] AF307144_1 cytosolic 6-phosphogluconate dehydrogenase [Spinacia oleracea] Glycosyl transferase [Populus trichocarpa] Beta-glucosidase 12-like [Glycine max] 1 1 16 15 1 1

Locus 17188 Locus 9703 Locus 21490 Locus 6953 Locus 5074 Locus 546 Locus 1514 Locus 10983 Locus 14864 Locus 557 Locus 3500 Locus 1842 Locus 27847 Locus 17 Locus 2600 Locus 21422 Locus 3293 Locus 3008 Locus Locus 2162 Locus 14369 Locus 2104 Locus 1885 Locus 2755 Locus 579 Locus 21913 Locus 6707 Locus 4604 Locus 368 Locus 3426 Locus 785 Locus 236 Locus 2151 Locus 4400 Locus 4343 Locus 33808 Locus 20970 Locus 3612 Locus 559 Locus 35868 Locus 13397 Locus 690 Locus 6028 Locus 123

Supplemental Table 5 - Overrepresented GO term of up-regulated genes in receptive seed flowers compared to receptive gall flowers Description GO-ID Loci in GO term p-value corr p-value translation 6412 49 2.19E-24 1.03E-21 gene expression 10467 55 2.20E-15 4.04E-13 cellular macromolecule biosynthetic process 34645 55 2.57E-15 4.04E-13 macromolecule biosynthetic process 9059 55 3.50E-15 4.12E-13 cellular protein metabolic process 44267 70 3.50E-12 3.30E-10 protein metabolic process 19538 72 2.77E-10 2.18E-08 nucleosome assembly 6334 9 8.67E-09 5.83E-07 chromatin assembly 31497 9 1.50E-08 8.85E-07 nucleosome organization 34728 9 2.13E-08 1.11E-06 protein-DNA complex assembly 65004 9 2.96E-08 1.40E-06 translational elongation 6414 8 1.33E-07 5.69E-06 DNA packaging 6323 9 1.93E-07 7.57E-06 chromatin assembly or disassembly 6333 10 2.30E-07 8.33E-06 DNA conformation change 71103 9 2.76E-06 9.30E-05 cellular macromolecule metabolic process 44260 79 3.49E-06 1.10E-04 cellular biosynthetic process 44249 61 8.59E-06 2.53E-04 biosynthetic process 9058 62 2.66E-05 7.38E-04 macromolecule metabolic process 43170 81 4.68E-05 1.23E-03 chromatin organization 6325 11 1.48E-04 3.68E-03 cellular macromolecular complex assembly 34622 10 2.23E-04 5.26E-03 cellular macromolecular complex subunit organization 34621 10 6.30E-04 1.41E-02 protein peptidyl-prolyl isomerization 413 3 1.40E-03 3.00E-02 chromosome organization 51276 11 2.40E-03 4.92E-02

124 155157.1 154998.1 1 1 Nasonia gene NP_00 NP_00 e r 85.50 92.80 543.00 639.00 206.00 459.00 182.00 416.00 530.00 428.00 Bitsco 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 E value 98.83 29.63 35.18 51.14 55.64 61.96 77.57 96.00 164.71 624.16 RPKM eads r 1.00 5.00 1 43.00 19.00 50.00 74.00 89.00 197.00 189.00 233.00 1 otal gene T 156 170 315 847 344 813 961 934 1 2375 1932 Gene length Annotation asp transcripts with hits to known venom genes.

enom carboxylesterase-6 isoform 1 enom dipeptidyl peptidase 4-like [Nasonia vitripennis] enom dipeptidyl peptidase 4-like [Nasonia vitripennis] enom dipeptidyl peptidase 4-like [Nasonia vitripennis] enom serine carboxypeptidase-like isoform 2 [Nasonia vitripennis] enom dipeptidyl peptidase 4-like [Nasonia vitripennis] enom serine carboxypeptidase-like isoform 2 [Nasonia vitripennis] Blastx Aspartylglucosaminidase Cysteine-rich/KU venom protein Icarapin-like [Nasonia vitripennis] V V V V V V V able 6: T

14 1

Supplemental Loci Locus 30672 Locus 7456 Locus 17558 Locus 13766 Locus 34842 Locus 22959 Locus 41657 Locus 15 Locus 16732 Locus 19749 125

Supplemental*Table*7*.*Overrepresented*GO*term*of*up.regulated*genes*in*pollinated*gall*flowers*compared*to*receptive*gall *flowers*and*pollinated*seed*flowers Description GO-ID Loci0in0GO0term p-value corr0p-value Chalcone)synthase acyltransferase*activity 8415 10 4.41E.04 8.01E.03 aromatic*compound*biosynthetic*process 19438 16 1.91E.10 3.55E.08 cellular*amino*acid*and*derivative*metabolic*process 6519 19 1.72E.05 6.13E.04 cellular*amino*acid*derivative*biosynthetic*process 42398 15 6.02E.10 9.31E.08 cellular*amino*acid*derivative*metabolic*process 6575 16 8.99E.09 1.04E.06 cellular*aromatic*compound*metabolic*process 6725 17 2.53E.08 2.35E.06 flavonoid*biosynthetic*process 9813 13 1.13E.11 3.50E.09 flavonoid*metabolic*process 9812 13 1.62E.11 3.75E.09 naringenin.chalcone*synthase*activity 16210 8 3.15E.08 2.65E.06 phenylpropanoid*biosynthetic*process 9699 15 3.01E.12 2.79E.09 phenylpropanoid*metabolic*process 9698 16 9.55E.12 3.50E.09 secondary*metabolic*process 19748 17 1.03E.09 1.37E.07 small*molecule*biosynthetic*process 44283 24 5.50E.07 3.92E.05 small*molecule*metabolic*process 44281 31 3.27E.04 6.19E.03 transferase*activity,*transferring*acyl*groups 16746 11 4.59E.04 8.11E.03 transferase*activity,*transferring*acyl*groups*other*than*amino.acyl*groups 16747 11 2.29E.04 4.89E.03 Peroxidase antioxidant*activity 16209 7 1.37E.04 3.27E.03 catabolic*process 9056 19 5.15E.04 8.67E.03 electron*carrier*activity 9055 13 1.27E.05 5.62E.04 external*encapsulating*structure 30312 12 4.32E.04 8.00E.03 heme*binding 20037 12 9.18E.07 6.08E.05 iron*ion*binding 5506 14 2.92E.05 8.73E.04 oxidation*reduction 55114 26 1.49E.05 6.08E.04 oxidoreductase*activity 16491 34 4.18E.06 2.04E.04 oxidoreductase*activity,*acting*on*peroxide*as*acceptor 16684 7 2.47E.05 7.88E.04 peroxidase*activity 4601 7 2.47E.05 7.88E.04 response*to*oxidative*stress 6979 11 2.30E.06 1.24E.04 response*to*stimulus 50896 44 1.54E.05 6.08E.04 tetrapyrrole*binding 46906 12 2.41E.06 1.24E.04 triacylglycerol*lipase,*putative*[Ricinus 70887 9 2.33E.03 3.48E.02 Beta-galactosidase, chitinase, and beta-1,3-glucanase carbohydrate*metabolic*process 5975 22 5.63E.06 2.61E.04 defense*response 6952 12 3.06E.04 5.92E.03 hydrolase*activity,*acting*on*glycosyl*bonds 16798 18 1.11E.08 1.15E.06 hydrolase*activity,*hydrolyzing*O.glycosyl*compounds 4553 14 1.83E.06 1.06E.04 response*to*biotic*stimulus 9607 10 2.44E.03 3.59E.02 response*to*stress 6950 33 1.36E.06 8.38E.05 response*to*wounding 9611 7 5.55E.05 1.56E.03 Other apoplast 48046 8 2.85E.04 5.74E.03 biological_process 8150 134 1.46E.03 2.30E.02 catalytic*activity 3824 109 1.79E.04 4.15E.03 cell*wall 5618 12 3.00E.04 5.91E.03 enzyme*inhibitor*activity 4857 6 7.67E.05 2.09E.03 extracellular*region 5576 15 5.04E.07 3.89E.05 vacuole 5773 12 2.26E.03 3.43E.02