Bark Beetle - Department of Entomology

3 downloads 0 Views 1MB Size Report
Oct 16, 2013 - Brevundimonas vesicularis. Fig. 2 Effects of bacteria associated with mountain pine beetle on concentrations of Pinus monoterpenes and ...
J Chem Ecol (2014) 40:1–20 DOI 10.1007/s10886-013-0368-y

REVIEW ARTICLE

Terpenes Tell Different Tales at Different Scales: Glimpses into the Chemical Ecology of Conifer - Bark Beetle - Microbial Interactions Kenneth F. Raffa

Received: 16 October 2013 / Revised: 9 November 2013 / Accepted: 21 November 2013 / Published online: 13 December 2013 # Springer Science+Business Media New York 2013

Abstract Chemical signaling mediates nearly all aspects of species interactions. Our knowledge of these signals has progressed dramatically, and now includes good characterizations of the bioactivities, modes of action, biosynthesis, and genetic programming of numerous compounds affecting a wide range of species. A major challenge now is to integrate this information so as to better understand actual selective pressures under natural conditions, make meaningful predictions about how organisms and ecosystems will respond to a changing environment, and provide useful guidance to managers who must contend with difficult trade-offs among competing socioeconomic values. One approach is to place stronger emphasis on cross-scale interactions, an understanding of which can help us better connect pattern with process, and improve our ability to make mechanistically grounded predictions over large areas and time frames. The opportunity to achieve such progress has been heightened by the rapid development of new scientific and technological tools. There are significant difficulties, however: Attempts to extend arrays of lower-scale processes into higher scale functioning can generate overly diffuse patterns. Conversely, attempts to infer process from pattern can miss critically important lowerscale drivers in systems where their biological and statistical significance is negated after critical thresholds are breached. Chemical signaling in bark beetle - conifer interactions has been explored for several decades, including by the two pioneers after whom this award is named. The strong knowledge base developed by many researchers, the importance of In Honor of Milt Silverstein and John Simeone Recipient of the 2011 ISCE Silverstein/Simeone award for outstanding research. K. F. Raffa (*) Department of Entomology, University of Wisconsin, 1630 Linden Dr., Madison, WI 53706, USA e-mail: [email protected]

bark beetles in ecosystem functioning, and the socioeconomic challenges they pose, establish these insects as an ideal model for studying chemical signaling within a cross-scale context. This report describes our recent work at three levels of scale: interactions of bacteria with host plant compounds and symbiotic fungi (tree level, biochemical time), relationships among inducible and constitutive defenses, population dynamics, and plastic host-selection behavior (stand level, ecological time), and climate-driven range expansion of a native eruptive species into semi-naïve and potentially naïve habitats (geographical level, evolutionary time). I approach this problem by focusing primarily on one chemical group, terpenes, by emphasizing the curvilinear and threshold-structured basis of most underlying relationships, and by focusing on the system’s feedback structure, which can either buffer or amplify relationships across scales. Keywords Bark beetles . Conifers . Cross-scale interactions . Plant defense . Symbiosis . Terpenes

Bark Beetles as Agents of Ecosystem Function, Socioeconomic Challenge, and Insight into Chemical Signalling Bark beetles that attack living conifers likely exert stronger ecological impacts and cause more socioeconomic challenges, in both natural and managed forest ecosystems, than any insect group throughout the temperate zone (Wood 1982b). Native bark beetle species have interacted with conifer biomes in particular for millions of years, and are natural disturbance agents contributing to vital ecosystem processes. They influence nutrient cycling, succession, forest structure, solar reflectance, soil dynamics, hydrology, fire, and biodiversity (Griffin et al. 2011; Kaiser et al. 2012; Kurz et al. 2008; Romme et al. 1986). Bark beetles also cause significant economic losses.

2

Intermittent landscape-scale outbreaks can convert sustainable resource-based economies to boom-and-bust cycles, with resulting hardships falling disproportionately on rural communities. Economic losses intensify along boundaries between managed and unmanaged forests, and between forests and human dwellings, where killed trees can damage electric wires and raise fire hazard, and increase personal liability and litigation. Native bark beetles have always undergone outbreaks, but in recent years the frequency, extent, range, and inter-specific synchronicity of outbreaks have intensified (Bentz et al. 2010; Raffa et al. 2008). Bark beetles also have served as valuable models for advancing the discipline of Chemical Ecology. In fact, both namesakes of this award, Milt Silverstein and John Simeone, performed some of the pioneering work on chemical signaling by these insects. Several recipients of ISCE’s Silver Medal Award have likewise contributed greatly to our understanding of bark beetle semiochemistry. My paper will not attempt a comprehensive literature review or synthesis, as several outstanding summaries of bark beetle semiochemistry, host relations, population dynamics, symbioses, and genetics already are available. The Silverstein-Simeone Lecture explicitly calls on recipients to describe recent and ongoing work in their own laboratories. I will try to do so in a way, however, that helps our discipline further contribute to a theme receiving increased attention throughout ecology and natural resource management, cross-scale interactions. Our work emphasizes how landscape-scale events can arise from micro-scale processes, and conversely how higher-scale factors affect whether lower-scale processes are buffered or amplified (Raffa et al. 2008). A major challenge to delineating cross-scale interactions is linking pattern with process. I approach this by illustrating how just one class of organic compounds, in this case terpenes, can have important ramifications at all levels of spatial and temporal scale. I recognize that this approach necessitates giving inadequate attention to other important chemical constituents such as phenolics and phenyl propanoids (Bois et al. 1999; Faccoli and Schlyter 2007; Kelsey et al. 2001; Klepzig et al. 1996), important physical factors such as the exudation and supporting duct structure of resin (Kane and Kolb 2010), histological responses such as traumatic duct formation and autonecrosis (Franceschi et al. 2005; Raffa et al. 2005; Schmidt et al. 2011), bark texture and reflectance (Strom et al. 1999), non-host foliar compounds that can influence beetle orientation (Zhang et al. 2000; Zhang and Schlyter 2003), and critical environmental factors such as temperature that affect beetles directly (Hicke et al. 2006; Powell and Bentz 2009; Régnière and Bentz 2007). Hopefully it illustrates the value of incorporating cross-scale interactions into our understanding of chemical signaling, and conversely how an understanding of chemical signaling can sharpen our ability to delineate common drivers across complex heterogeneous systems.

J Chem Ecol (2014) 40:1–20

Terpenes Inject Both Flavor and Venom into Bark Beetle - Conifer Affairs At first glance, the life cycle of bark beetles appears deceptively simple: Adults emerge from trees in which they developed, fly to new trees, tunnel through the bark into the phloem, mate and oviposit, and the brood develop to adulthood. However, several foundational relationships of coniferbark beetle interactions set the stage for dynamics that can either be buffered or amplified across spatiotemporal scales. These foundational relationships yield opposing rate reactions, and critical thresholds (Martinson et al. 2013; Mawby et al. 1989; Raffa et al. 2008; Smith et al. 2011) that are quantitatively determined yet yield qualitatively distinct outcomes. They also confront us with systems that can be quite difficult to approach empirically, and so are prone to both spurious correlations and overlooked drivers. First, the old pharmacological dictum ‘The dose makes the poison’ (Philippus Aureolus Theophrastus Bombastus von Hohenheim:1493–1541) couldn’t be more operative. Low concentrations of monoterpenes often benefit bark beetles while high concentrations inhibit them (Raffa et al. 2005). So it is less helpful to think of particular compounds as falling into strict categories than to quantify their dose–response manifolds. Second, these manifolds are rarely linear. Instead, they are more often curvilinear and exhibit thresholds. Third, a tree’s subcortical tissues can only support a fixed number of brood. This carrying capacity is directly related to tree diameter and phloem thickness (Amman 1972), which in turn are influenced by endogenous drivers such as tree age and exogenous drivers, e.g., other herbivores and pathogens, fireinjury, and drought. Fourth, successful brood development exhausts a tree’s suitability for subsequent brood development, thereby subtracting from the stand-level resource and requiring each generation to locate and colonize new suitable hosts. This introduces an important source of negative feedback (Berryman 1979). Terpenes Enable Bark Beetles to Locate and Colonize Host Trees … but also Pose Formidable Barriers to These Herbivores and Their Symbionts Almost all (if not all) of the major life history processes of conifer bark beetles benefit from specific concentrations of some terpenoids (Raffa et al. 2005; Seybold et al. 2006). For example, beetles exploit volatiles during host location that improves orientation toward host trees, they exploit low concentrations of some monoterpenes as post-landing host recognition cues and feeding stimulants, utilize host- and beetlederived oxygenated terpenes for pheromonal attraction of mates, and undergo a number of physiological reactions driven by terpenoid juvenile hormones during various life stages (Blomquist et al. 2010; Borden 1985; Pureswaran et al. 2000;

J Chem Ecol (2014) 40:1–20

Seybold et al. 2006; Wood 1982a, b). Some ‘parasitic’ species such as Dendroctonus micans (Kugelann) may even gain partial escape from competitors and other natural enemies from their terpene-rich environments (Everaerts et al. 1988). Because bark beetles most strongly capture attention during outbreaks, it is easy to think of trees as passive substrate, an assumption that can become inadvertently imbedded in overly general management policies or overly descriptive scientific approaches. Nothing could be farther from the truth: Most species of bark beetles cannot colonize live trees, most of those that can attack live trees are always confined to a small subset of highly stressed individuals, most of those that can sometimes attack less stressed trees do not undergo region-wide outbreaks, and the few remaining ‘outbreak species’ spend most generations in lengthy endemic periods during which they are limited to stressed trees (Lindgren and Raffa 2013). This overwhelming concentration of species and populations within dead or stressed trees is testimony to the potent defenses of conifers. From an evolutionary perspective, this should not be surprising: destruction of subcortical tissue during brood development usually results in tree death, imposing strong selection pressures to prevent colonization. Conifer defenses against subcortical invasion are multifaceted, including chemical, histological, and morphological components. A variety of chemical groups are involved in these defenses, but terpenes are particularly important (Bohlmann et al. 2000; Franceschi et al. 2005; Kane and Kolb 2010; Keeling and Bohlmann 2006a, b; Lorio and Sommers 1986; Raffa and Berryman 1983; Schmitt et al. 1988; Strom et al. 2002; Tisdale et al. 2003). At high concentrations, monoterpenes are repellant, adulticidal, ovicidal, larvicidal, and moderately fungicidal (Everaerts et al. 1988). Diterpene acids are highly fungicidal, and may have some insecticidal activity. We know little about how different terpenes interact, but synergisms are likely, and their diversity alone poses physiological challenges to insects. Both monoterpenes and diterpene acids are highly inducible, undergoing rapid localized biosynthesis at the attack site, achieving concentrations toxic to the insects and microbial symbionts within just a few days. Because bark beetle brood require months to years (depending on species and weather) to develop, they are highly vulnerable to these chemical changes. Thus, the relevant time scale is a full beetle generation, because all life stages must be able to survive the host environment for reproduction to succeed. This places strong selective pressures on adults to not enter trees that can undergo lethal defensive reactions, especially since bark beetles invest all, or most, of their egg clutch in one tree. The available evidence suggests that most adults never find a tree they are willing to enter, and die before doing so. Yet, paradoxically this critical aspect of the conifer-bark beetle interaction leaves no ‘signature’, so it is often omitted from analyses and models.

3

Terpenes Enable Bark Beetles to Exhaust Tree Defenses by Mass Attack … but also Attract Their Predators and Competitors Bark beetles can sometimes overcome the sophisticated defenses of conifers through cooperative behavior, which is mediated by aggregation pheromones that are frequently oxidized terpenes (Blomquist et al. 2010; Gitau et al. 2013; Wood 1982a, b). The induction, attractiveness and biosynthesis of these pheromones are likewise influenced by host monoterpenes (Pureswaran and Borden 2005). The roles of host compounds in different phases of pheromone synthesis vary among beetle species. When beetles succeed in killing the tree, tree defenses are exhausted and so exert little to no effect on brood survival. That is, whatever a tree’s potential resistance is prior to attack, if the beetles’ arrival rate and density are sufficient, the concentrations of terpenes needed to kill the beetles, brood, and symbionts are not achieved. However, if the arrival of cohorts is inadequate, the attack fails, and the tunneling beetle either exits or is killed. Because beetles are repelled by high terpene concentrations (Erbilgin et al. 2006; Zhao et al. 2011), abandonment is probably more common. Again, this behavioral dimension has the paradoxical effect that actual beetle mortality caused by tree defense is quite low, and so classical life table approaches do not capture the role of tree defense. Natural enemies often are attracted to plant volatiles induced by herbivores, and exploit these signals as host-finding cues (Delphia et al. 2006; Dicke 2009; Felton and Tumlinson 2008; Zhang et al. 2013). These responses are categorized as ‘indirect defenses’, although the extent to which they contribute to plant fitness in natural ecosystems is uncertain. In the case of bark beetle predators, host tree compounds are most attractive when combined with the herbivore’s pheromones, which orients them efficiently toward prey (Wood 1982a). Although predator attraction to this combination of pheromones and induced tree chemistry may reduce the herbivore population, it does not benefit the plant emitting the signals. For example Thanasimus (Coleoptera: Cleridae) adults feed on bark beetle adults, then oviposit at the beetles’ entrance galleries, and the larvae feed on the bark beetle brood. So without tree death, the food source for Thanasimus larvae would not become available. In such cases, it is more appropriate to interpret tritrophic signaling as ‘eavesdropping’ by natural enemies than as plants ‘calling in’ the herbivore’s enemies. The complexity of terpenoid chemistry provides bark beetles with options, however. The enormous number of permutations of stereoisomers, concentrations, and mixtures allows beetles to modify their signaling in ecological time in manners that maintain intraspecific functionality while gaining at least partial escape from eavesdroppers (Raffa et al. 2007).

4

J Chem Ecol (2014) 40:1–20

Some competitors also show increased attraction to bark beetle pheromones in the presence of host monoterpenes. Again, their effect is to reduce beetle population growth, not increase the likelihood of tree survival. In some cases ‘secondary beetle’ species are attracted to host and beetle volatiles emanating from trees killed by ‘primary’ species, then produce their own mating pheromones, and these pheromones attract additional predators that then feed on larvae of the primary beetle (Boone et al. 2008). In contrast to such injustice, tritrophic signaling in conifer-bark beetles systems also may reinforce fair-play social behavior in tree-killing beetles. Researchers have long wondered if there might be some benefit to a ‘cheating’ strategy in bark beetles, i.e., only responding to signals of success by other beetles rather than initiating risky pioneering attacks. One cost to that strategy, however, is that late-arriving adult beetles may experience higher predation than early-arriving beetles, due to the relative time course of volatile emissions, beetle entry, and predator arrival (Aukema and Raffa 2004). Some Emerging Areas That Could Improve Our Understanding of Cross-Scale Interactions in Chemical Signaling The dose-dependent, curvilinear, and opposing effects of terpenes on various aspects of conifer-bark beetle-microbialpredator relationships set the stage for important cross-scale interactions. Some are well characterized, so I will not revisit those here. Instead I will focus on three areas that we are currently researching, and particular components that we are just beginning to understand: Micro-Scale—Interactions between terpenes and bacteria within bark beetle galleries. Meso-Scale—Interactions among terpenoid-based tree defenses, stand-level beetle population density, and exogenous stress agents. Macro-Scale—Comparative defense chemistries of native vs. semi-naïve and naïve hosts being made increasingly accessible by changing climate and other anthropogenic forces.

The Micro Scale: Interactions Among Beetles, Bacteria, and Terpenes Within Galleries Until recently, almost everything we knew about bark beetle—microbial interactions was limited to fungi. This was due partly to the more visible signs fungi leave (such as blue stain), the clearly important roles that fungi play (Ayres et al. 2000; Bleiker and Six 2007), and the difficulties inherent in bacterial identification (many species are not readily culturable, molecular methods were poorly developed). Recent conceptual

advances in the often multipartite nature of symbioses, and technical advances in the capabilities and accessibilities of molecular tools, have changed that paradigm (MoralezJimenez et al. 2009; 2012). Bacteria appear to play important roles in how bark beetles acquire host tree nutrients and defend their resource from opportunistic fungi (Cardoza et al. 2006; Scott et al. 2008). We’ve begun to ask two additional questions: Do bacteria associated with bark beetles detoxify host terpenes, and if so to what extent? Do bacteria influence fungi that mediate conifer-bark beetle interactions, and if so are these processes affected by host terpenes? In the future we hope to extend these results to more complete and realistic combinations, such as in vivo bioassays with mountain pine beetles (Dendroctonus ponderosae Hopkins), bacterial associates, and symbiotic and opportunistic fungi. Do Bacterial Symbionts Detoxify Host Terpenes? We conducted a three-pronged approach: a metagenomics analysis characterizing the major groups of bacteria associated with mountain pine beetle, and an evaluation of whether they have genes encoding for enzymes that are reported to metabolize terpenes; in vitro bioassays to determine whether selected bacteria could reduce concentrations of terpenes in media; and toxicity assays to evaluate effects of various terpenes on beetle—associated bacteria. 1. Bacterial communities associated with mountain pine beetles and their hosts are enriched with genes involved in terpene metabolism Our metagenomic analyses, and also denaturing grade gel electrophoresis (DGGE), indicate that Gammaproteobacteria, especially Serratia, Pseudomonas, Stenotrophomonas, and Erwinia , are prevalent in mountain pine beetles, galleries, and host species (Adams et al. 2013). This community contains numerous bacterial genes associated with degradation of monoterpenes and diterpenes (Table 1), including wellrepresented KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways for pinene and limonene degradation, and many genes homologous to the dit gene cluster of Table 1 Bacterial symbionts of mountain pine beetle have genes putatively involved in monoterpene degradation (Adams et al. 2013) Terpene

No. enzymes in Bacteria detected from beetles degradation pathway Enzymes Genes

(-)-(S)-limonene 10 (+)-(R)-limonene 8 α–pinene 10

5 2 8

179 103 283

J Chem Ecol (2014) 40:1–20

5

conditions as above, to safeguard against insect or plant fragments having inadvertently provided a source of carbon. We selected the most promising candidates for detoxification assays. These consisted of amending media with various terpenes and then quantifying the concentrations of amended compounds that remained after exposure to bacteria vs. controls. Chemical analyses were performed by gas liquid chromatography and high-pressure liquid chromatography. The predominant isolates we obtained were Pseudomonas spp. and Serratia spp. The most abundant bacteria, and best able to tolerate terpenes and apparently use them as a carbon source, were Pseudomonas (Adams et al. 2013; Boone et al. 2013). In bioassays, four species of bacteria associated with mountain pine beetle adults, galleries, and hosts reduced concentrations of monoterpenes (Fig. 2) (Boone et al. 2013). Their closest matches to species in Genebank are to Serratia marcescens, Pseudomonas mandelii, Pseudomonas migulae, and Rahnella aquatilis. Serratia marcescens had the strongest effect, reducing concentrations of all monoterpenes except (+)- and (-)-α-pinene. For example, it reduced mean concentrations of 3-carene and (-)-β-pinene by 49–79 %. Pseudomonas mandelii reduced concentrations of all monoterpenes by 15– 24 %. Rahnella aquatilis lowered concentrations of all monoterpenes, and caused the greatest decrease in (-)-α-pinene (38 %) and (+)-α-pinene (46 %). Not all bacteria reduced monoterpene concentrations. For example, Pseudomonas brenneri and Pantoea agglomerans had no effect. Two bacteria greatly reduced concentrations of the diterpene abietic acid (Fig. 2) (Boone et al. 2013). These effects were dose-dependent. The bacteria whose closest matches were to S. marcescens and Brevundimonas vesicularis caused 100 % reduction at low concentrations of abietic acid, but this ability diminished with increasing concentrations. The

Pseudomonas abietophila that contributes to diterpene degradation. We found a significantly higher proportion of such genes in these metagenomes than in other plant biomassassociated microbial communities. Most of the bacterial genes involved in monoterpene and diterpene degradation are from Pseudomonas and Rahnella. The genus Pseudomonas contains many species known to degrade a broad range of plant toxins, xenobiotics, and pollutants (Foght and Westlake 1988; Kanaly and Harayama 2000; Martin et al. 1999; Whyte et al. 1997). These results are consistent with the hypothesis that bacterial symbionts contribute to D. ponderosae’s ability to attack live pines. These do not appear to be strict vector relationships, however. Rather some of these bacteria may be pine residents that utilize host terpenes as a carbon source and become especially active when entry by beetle-fungal complex elicits resinosis. 2. Some bacterial associates of mountain pine beetles can grow on minimal diet enriched with terpenes, and decrease concentrations of monoterpenes and diterpene acids Insects encounter many bacteria, many of them transients, so we needed a screening strategy to guide more focused experiments. Our approach is illustrated in Fig. 1. We collected beetles and gallery samples in the field, and cultured bacterial isolates. We tested whether beetle-associated bacteria could tolerate conifer terpenes, whether they could grow with terpenes as their only or primary source of carbon, and whether they were frequent or incidental associates. We first plated bacteria on minimal media (water and vitamins) amended with terpenes. We obtained pure cultures, and separated morphotypes representing putatively different species. We then made new cultures of these morphotypes under the same Fig. 1 Flow diagram of isolation, culture, and screening for subsequent terpene detoxification assays of bacteria associated with mountain pine beetles and host trees. Insert map shows locations of collection sites in Alberta and British Columbia

Analyze community metagenome: Screen for genes known to encode for terpene-metabolizing enzymes (DNA extraction, sequencing, BLAST, KEGG)

Sample bacteria from beetles, galleries, trees

Tolerate terpenes?

Grow with terpenes as only C-source?

Frequent associate? (Isolation; 16S rRNA; DGGE)

Does bacterium reduce terpene concentrations? (Amend medium, Apply bacteria or controls, GLC, HPLC)

6

J Chem Ecol (2014) 40:1–20

Pseudomonas mandelii

-45 -60

Pseudomonas brenneri

-75

Rahnella aquatilis

Abietic acid 0.87%

Abietic acid 0.43%

Abietic acid 0.22%

Abietic acid 0.11%

(S)-limonene

(R)-limonene

(S)-B-pinene

(S)-a-pinene

(R)-a-pinene

3-carene

-90

Brevundimonas vesicularis

Fig. 2 Effects of bacteria associated with mountain pine beetle on concentrations of Pinus monoterpenes and diterpene acids in amended media. For each terpene, the percent change relative to controls (media not amended with bacteria) is shown for a battery of bacteria assayed (From Boone et al. 2013, statistical analyses therein, and unpublished data). Note-names are based on nearest match in Genebank, so all species names are approximations

threshold for S. marcescens was at approximately 0.43 % abietic acid, and the threshold of B. vesicularis was approximately 0.87 %. Several interpretations emerge from these results. First, multiple bacterial species associated with one beetle species appear to have complementary roles in terms of which plant defense compounds they decrease. Second, bacteria-tree interactions appear to experience dynamic and opposing rate reactions, similar to beetle-tree and fungal-tree interactions. That is, based on our knowledge of tree chemistry, these bacteria appear able to detoxify concentrations present in constitutive but not induced host tree phloem tissue. Third, there appears to be some redundancy and complementarity among various species of bacteria, which may be of ecological importance given the substantial between-individual and between-population variability in bacterial communities associated with a bark beetle species (Adams et al. 2010). Further, bacteria likely add to the overall combination of physical and metabolic methods by which beetles (Cano-Ramírez et al. 2013; Sandstrom et al. 2006) and fungi (Davis and Hofstetter 2011; DiGuistini et al. 2011; Hammerbacher et al. 2013; Lieutier et al. 2009; Wang et al. 2013) lower terpene concentrations, rather than accounting for such a critical function by themselves. 3. Effects of monoterpenes on bacterial symbionts of bark beetles range from total inhibition to stimulation. Bacterial responses appear to reflect beetle life history Exposure of the full range of bacteria associated with mountain pine beetle to a broader range of monoterpenes

100 80

D. ponderosae

60 40 20 0 100

Stimulation No Effect

80

Partial Inhibition 60

Strong Inhibition

40

Total Inhibition

20 0

A-Pinene 1% A-Pinene 5% B-Pinene 1% B-Pinene 5% Limonene 1% Limonene 5% Myrcene 1% Myrcene 5% 3-Carene 1% 3-Carene 5% B-Phelland 1% B Phelland 5%

Serratia marcescens

-30

% of Bacteria

% Change

-15

revealed several trends (Fig. 3 upper) (Adams et al. 2011): First, β-phellandrene, the predominant terpene of mountain pine beetle’s primary host lodgepole pine, showed very low toxicity. Bacteria appear well-adapted to this compound. In contrast, α -pinene was highly toxic, and β-pinene showed a strong threshold relationship. Secondly, myrcene had either low effect on or actually enhanced the growth of mountain pine beetle-associated bacteria. This relationship is noteworthy because myrcene synergizes attraction of mountain pine beetles to their aggregation pheromone trans-verbenol. Third, 3-carene was the most inhibitory compound to bacteria. Interestingly, 3-carene is one of the least toxic monoterpenes to bark beetles and their fungal symbionts in laboratory assays, yet has sometimes been associated with the degree of lodgepole pine populations’ historical exposure to mountain pine beetle. Toxicity to bacteria might partially explain that otherwise paradoxical pattern. Fourth, bacteria showed an overall high tolerance to limonene: This is likewise noteworthy because limonene is often the most insecticidal and fungicidal monoterpene in conifers. Fifth, there was high variation among different bacteria associated with mountain pine beetle: For example, the effect of 5 % β-pinene ranged from complete inhibition to no effect, depending on species. Bacteria associated with the red turpentine beetle, Dendroctonus valens LeConte, were much more tolerant of

% of Bacteria

0

D. valens

Fig. 3 Effects of Pinus monoterpenes on bacteria associated with mountain pine beetle and red turpentine beetle. For each monoterpene concentration, the percentage of bacteria that showed total through partial inhibition, no effect, or stimulation is shown (From Adams et al. 2011 statistical analyses therein)

J Chem Ecol (2014) 40:1–20

7

host monoterpenes than were those associated with mountain pine beetle (Fig. 3 lower) (Adams et al. 2011). None of these were strongly or totally inhibited. This likely relates to differences in lifestyle: Whereas mountain pine beetle uses pheromone-mediated mass attacks to deplete tree chemistry, D. valens usually conducts solitary attacks in the Lake States region. It reproduces in live trees, a highly terpene-rich environment. These bacterial associates appear particularly well adapted to the life-styles of their host beetles, in terms of their ability to tolerate terpenes. Do Bacterial Symbionts of Bark Beetles Facilitate Their Fungal Symbionts? Symbiotic fungi play critically important roles in the ecology and reproductive success of bark beetles (Ayres et al. 2000; Klepzig et al. 2009; Lee et al. 2006; Six and Klepzig 2004). Bacteria associated with bark beetles can stimulate the mycelial growth, production of conidia, and production of conidiophores by these fungal symbionts (Adams et al. 2009). However, these relationships appear to be strongly mediated by host compounds, and in a variety of manners. Some examples showing the complexity arising from just a single monoterpene and two fungi are illustrated in Table 2. In some cases, neither the bacterium nor the host compound stimulates

the fungus, but jointly they do. An example is the combined effect of Pseudomonas and racemic α-pinene on the mycelial growth of Grosmannia clavigera. In some other cases, the stimulatory effect of the bacterium is synergized by the host compound. An example includes how racemic α-pinene synergizes the stimulatory effect of Pseudomonas on conidiophore production by Leptographium procerum . In some cases, the ability of the bacterium to stimulate the fungus is negated by the host compound. For example, Pantoea stimulates the mycelial growth of Leptographium procerum, but not in an environment containing racemic α-pinene. In addition to enhancing some stages of symbiotic fungi, beetle-associated bacteria can also inhibit antagonistic fungi (Cardoza et al. 2006; Scott et al. 2008). In a comparative study, Cardoza et al. (2006) found that bacteria in the oral egestions of D. rufipennis were more inhibitory to antagonistic Trichoderma and Aspergillus species than to the symbiotic Leptographium abietinum . Despite these relationships observed under controlled laboratory conditions, however, we still know little about how fungal-bacterial-conifer interactions function in vivo, where factors such as timing and microsite can be critical. Hence, the extent to which these relationships quantified under controlled conditions scale up to tree-level interactions under natural conditions is unknown.

Table 2 Effects of bacterial associates, plant terpenes, and their interaction on fungal symbionts of bark beetles (Adams et al. 2009) Fungus

Bacteria

Fungal performance

Grosmannia clavigera

Pseudomonas sp.

Mycelial growth None Conidiophores

Pantoea sp.

Effect of bacteria

Effect of racemic Effect of bacteria and Interaction between plant terpene and α-pinene racemic α-pinene beetle-associated bacteria on fungal symbiont of beetle None

Stimulation

Stimulation High stimulation Stimulation

Mycelial growth Stimulation Stimulation

High stimulation

Conidiophores Stimulation Stimulation Pectobacterium sp. Mycelial growth None Stimulation

Stimulation High stimulation

Conidiophores Leptographium Pseudomonas sp. procerum

Pantoea sp.

Stimulation Stimulation

High stimulation

Mycelial growth Stimulation None

Low stimulation

Conidiophores

High stimulation

Stimulation None

Mycelial growth Stimulation None

None

Conidiophores

Stimulation Stimulation

High stimulation

Pectobacterium sp. Mycelial growth Stimulation Stimulation

High stimulation

Conidiophores

Stimulation Stimulation

High stimulation

Terpene interacts with bacteria to stimulate fungus Bacteria reduce ability of terpene to stimulate fungus Terpene and bacteria interact to synergize fungus No interaction Bacteria increase terpene stimulation of fungus Terpene and bacteria synergize their stimulation of fungus Terpene reduces bacterial stimulation of fungus Terpene increases bacterial stimulation of fungus Terpene inhibits bacterial stimulation of fungus Terpene and bacteria synergize their stimulation of fungus Terpene and bacteria synergize their stimulation of fungus Terpene and bacteria synergize their stimulation of fungus

8

J Chem Ecol (2014) 40:1–20

'Ménage à Quatre'—A Conceptual Model of Conifer- BeetleFungal- Bacterial Interactions Figure 4 provides a conceptual model that integrates previous work, the relationships shown here, and postulated mechanisms into an overall framework of bark beetle-conifermicrobial interactions. Its major purpose is to guide future work as we attempt to better frame bacteria within the overall system. Initial beetle entry is elicited by low concentrations of monoterpenes (Raffa et al. 2005), which also serve as precursors and synergists of aggregation pheromones (Blomquist et al. 2010). Each beetle in a resulting mass attack severs resin ducts (Berryman 1972), and metabolizes terpenes (Sandstrom et al. 2006). Trees respond by rapidly inducing high concentrations of insecticidal monoterpenes and fungicidal diterpene acids (Erbilgin et al. 2006; Huber et al. 2004; Raffa et al. 2005). Outcomes are strongly influenced by higher-scale factors such as forest structure, environmental stress, stand-level beetle population size, and weather (Raffa et al. 2008), and lower-scale processes such as those involving microorganisms. Specifically, Serratia, Pseudomonas and Rahnella reduce monoterpenes, and Brevundimonas and Serratia reduce diterpene acids. Some ophiostomatoid fungi may also metabolize monoterpenes (DiGuistini et al. 2011; Wang et al. 2013) and stilbenes (Hammerbacher

et al. 2013), but their metabolism of diterpene acids seems unlikely (Fig. 4) (Kopper et al. 2005). Ogataea yeasts may also alter monoterpene composition (Davis and Hofstetter 2011). If terpenoid induction by trees (central boxes) is very rapid and pronounced, however, microbial metabolic abilities may be exhausted (Adams et al. 2011), resulting in toxicity to the attacking agents and tree survival. Some bacteria apparently are introduced into trees by beetles, but some also reside within trees and may be activated by terpene flow induced by beetle attack (Adams et al. 2010). Individual components of this model have been validated under controlled laboratory conditions, but it remains to be seen how such interactions function in nature. We cannot yet predict the extent to which consequences of specific microbial interactions are amplified or buffered at higher scales.

The Meso Scale: Interface Between Tree Defense Physiology and Stand-Level Beetle Populations Despite the enormous amount of information on how various terpenes influence specific features of conifer - bark beetle microbial interactions, we still find ourselves unable to fully answer a rather basic question: How do tree defenses influence bark beetle population dynamics?

Beetle Entry Bark Surface

Serratia Pseudomonas Rahnella

Monoterpenes Detoxification

Induction

Diterpene Acids Induction

Detoxification

Brevundimonas Serratia

Grosmannia Ogataea

Insect Repellency, Mortality

Fig. 4 Proposed model of integrated bark beetle—microbial depletion of terpene-based defenses. Processes that benefit beetle are indicated with dashed lines, those benefitting tree with solid lines. Initial beetle entry is elicited by low concentrations of monoterpenes (Wallin and Raffa 2000). Beetles also exploit monoterpenes as precursors and synergists of aggregation pheromones (Blomquist et al. 2010). Trees respond by rapidly producing high concentrations of insecticidal monoterpenes and fungicidal diterpene acids (Huber et al. 2004; Raffa et al. 2005; Erbilgin et al. 2006; Bohlmann and Gershenzon 2009). Serratia, Pseudomonas and Rahnella reduce concentrations of monoterpenes (Boone et al. 2013), as do Grosmannia, Ogataea, and beetles (Sandstrom et al. 2006; Davis

Fungal Inhibition

and Hofstetter 2011; DiGuistini et al. 2011). Brevundimonas and Serratia degrade diterpene acids, which inhibit Grosmannia (Fig. 3). Bacteria are introduced by beetles, but some may reside within trees or arrive by other means (arrows), and are activated when beetle attack induces terpenes (Adams et al. 2013). Bacillus may convert monoterpenes to aggregation pheromones (Brand et al. 1975), which would enhance the pathway of tree monoterpenes eliciting more beetle entry. High levels of terpenes, such as those achieved during successful induction (central boxes) are toxic to bacteria (Adams et al. 2011) and impede their ability to break down toxins, thus favoring tree survival

J Chem Ecol (2014) 40:1–20

9

How Do Tree Defenses Scale Up to Influence Bark Beetle Population Dynamics? We’ve recently completed some studies with mountain pine beetle that address two elements of this broader question: Specifically, we asked 1) How do tree-level defense and stand-level beetle population density interact? 2) How do exogenous stress agents affect mechanisms of tree defenses?

100 80 60 40 20 0 0

40

80

120

100 80

1. The role of tree defense physiology in mountain pine beetle reproductive success ranges from critical to inconsequential, depending on stand-level beetle population density Our first study had two main components: Long-term censusing of natural attacks and their outcomes, and relating various physiological attributes of trees to their subsequent likelihoods of being attacked by natural beetle populations. Six 12–18 ha stands were established in British Columbia, and all lodgepole pines within them were censused monthly. Over the course of six years, beetle populations transitioned across endemic, intermittent, and eruptive stages, both among and within sites. In the second component, trees were assayed for a variety of measures of putative defense just prior to the onset of mountain pine beetle flight. These measures included constitutive and induced resin flow, constitutive and induced phloem monoterpene concentrations, and the length of induced autonecrotic lesions. In all cases, induction was elicited by simulating a mountain pine beetle attack by administering a mechanical wound coupled with injection of the beetle’s primary fungal symbiont G. clavigera. The proportion of trees in which beetles were able to elicit mass attacks after entering them, and thus kill the host and breed, varied widely. The primary explanatory driver was stand-level beetle population density (Fig. 5) (Boone et al. 2011). When beetle populations were low, their likelihood of success after entering a tree was likewise extremely low. When beetle populations were high, their likelihood of success after entering a tree was nearly 100 %. This was not simply a matter of some stands being more conducive to attack than others (although that is surely true), because the same trend occurs within individual stands (Fig. 5, insert). Nor was it simply a matter of some years being better for mountain pine beetles than others (again, surely true), because this relationship occurs within individual years. Whether a beetle can generate enough arrivals needed to overcome a tree-level threshold of resistance depends on whether there are enough beetles to exceed a stand-level threshold of resistance. 2. Induced phloem terpene concentrations elicited by simulated beetle-fungal attack provides the best predictor of natural attack. High terpenes and resin defend trees when beetle populations are low. These defenses are overcome

60 40 20 0 0

10

20

30

40

Fig. 5 Effect of stand-level population size of mountain pine beetle on percentage of entered lodgepole pines that become foci of mass attack and are successfully colonized (From Boone et al. 2011). Results from all stands and years are shown in the full graph as fit in Boone et al. (2011) (y =0.21+0.72(1-e−0.084x); R 2 =0.67,F2,27 =27.77, P