Can plant–microbe–insect interactions enhance or inhibit the spread of invasive species?


Correspondence author. E-mail:


  1. Invasive species are one of the great challenges facing the world leading to great economic losses. Increasing numbers of species introductions are also increasing the likelihood of new species interactions – particularly between plants, microbes and insects.
  2. Frequently discovered interactions between plants, microbes and insects are giving rise to a new field: plant–microbe–insect (PMI) interactions. This paper focuses on novel PMI interactions created from the introduction of new plant, insect and microbe species. In particular, this paper asks: Do novel PMI interactions promote or inhibit invasive plants, microbes and insects? And can we predict whether novel PMI interactions are likely to become invasive?
  3. While we might predict that novel PMI interactions are likely to be simple additive interactions due to their relatively short period of interaction, instead this review demonstrates that most novel PMI interactions are actually nonadditive. This manuscript shows that there are a great number of instances where invasive species are promoted by novel PMI interactions. By contrast, the studied cases where PMI interactions limit invasive species are predominantly biocontrol PMI interactions.
  4. Future research on novel PMI interactions should focus on predicting future novel PMI interactions that promote invasive species. Given that many novel PMI interactions involve plant pathogens and their insect vectors, this novel PMI interaction deserves more focus. New research should also focus on non-novel PMI interactions that could be manipulated to hinder the spread of invasive plant, microbe and insect species.


Invasive species are one of the greatest challenges for most ecosystems in the world. Alone, invasive fungi cost US agriculture $21 billion USD per year (Rossman 2009). Invasive plant species have similarly strong economic impacts: in 2003, invasive plants were estimated to cost Australia approximately $4 billion AUD per year (Sinden et al. 2004), whereas Vila et al. (2010) estimated that the cost of eradicating only the 30 most common invasive plant species in Europe was greater than €150 million. In addition, European crop losses to invasive insects are estimated at €2·8 billion (Vila et al. 2010). These figures demonstrate that plants, insects and microbes are having dramatic effects on the world economy, and many of these economic impacts are actually due to interactions between plants, insects and microbes.

Human transport of non-native species has likely occurred since the movement of humans across the globe began in Neolithic times (Webb 1985), but was not typically recorded until the age of European exploration (approximately AD 1500) when the introduction of species began to increase (di Castri, Hansen & Debussche 1990). Today, however, as human movement and trade across the globe becomes increasingly common, the introduction of new species has increased in frequency (reviewed in Bradley et al. 2012). Seeds, eggs, spores and other living material are transported both knowingly and unknowingly. While most introductions fail, the increasing number of introductions is leading to an increasing number of successful introductions (reviewed in Bradley et al. 2012). This paper will distinguish between two types of introduced species: ‘invasive’ species that dominate habitats causing changes to ecosystems and economic losses, and the broader category of ‘noninvasive’ introduced species that either have become naturalized in their new environment or have yet to become invasive.

Multiple theories have been proposed to explain what biotic factors influence invasive species success. The vast majority of these theories have focused on or been developed for invasive plants. These theories include the Enemy Release Hypothesis (invasives encounter fewer enemies in new environments; Elton 1958; Keane & Crawley 2002; Liu & Stiling 2006), Evolution of Increased Competitive Ability (invasives evolve to be better competitors in new environments in the absence of enemies; Blossey & Notzold 1995) or the Novel Weapons Hypothesis [in which invasive species (primarily plants) exclude competitors using toxic compounds; Callaway & Ridenour 2004]. A recent review revealed 29 different proposed hypotheses for the success of invasive species, and the vast majority of these hypotheses focused on traits of invasive plant species or two-way biotic interactions (Catford, Jansson & Nilsson 2009). For example, the Enemy Release Hypothesis applies to the enemies of plants, regardless of whether they be insect or microbial, but does not consider the influence of insect–microbe interactions on plants.

Greater numbers of introduced species can also increase the likelihood of novel species interactions (Richardson et al. 2000), particularly between plants, microbes and insects. The effects of microbes, insects and plants on each other can be additive or nonadditive. When effects are additive, the outcome of one interaction is not influenced by a second or third interaction – all influences are independent allowing us to ‘add’ the influence of one organism to the influence of a second organism. Nonadditive interactions between plants, microbes and insects are those in which the influence of multiple organisms is not equal to the sum of the individual influences of each organism (Fournier et al. 2006). Nonadditive interactions can be synergistic (in which the outcome is greater than the additive effects) or antagonistic (in which partners hinder each other and the outcome is less than the additive effects; Fournier et al. 2006). In the new field of plant–microbe–insect (PMI) interactions highlighted in this issue, research has identified additive, synergistic and antagonistic interactions between plants, insects and microbes (Table 1).

Table 1. Summary of all known (and some hypothesized) effects of plant–microbe–insect (PMI) interactions on invasive plants, insects and microbes
Invasive organismPMI Interactions effect typeFunctional groupExamplesPromote invasion?
  1. The first column lists the category of invasive organisms (plant, insect or microbe), and the second column lists the type of interaction (additive or nonadditive where nonadditive is broken down into synergistic or antagonistic). The remaining columns are divided based on the number of examples within each type of interaction. The third column lists the functional group of the invasive organism. In this table, all plants essentially fall into a common functional group, but insects and microbes can be divided into multiple functional groups based on the organisms with which they are interacting (e.g. plant pathogen vs. entomopathogen). All known (and some hypothesized) examples of PMI interactions involving an invasive organism are listed in column four, and whether these interactions are known to promote invasion of the focal organism (along with references supporting that conclusion) are listed in the final column. In some cases (e.g. plant–endophyte interactions and insect vector and plant pathogen interactions), there were too many references, so examples have been provided instead. ‘Unknown’ listed in the ‘Promote Invasion?’ column refers to a potential interaction for which there are no published examples.

PlantAdditivePlantPlant Pathogens + HerbivoresNo (O'Brien et al. 2010; Rayamajhi et al. 2010)
PlantSoil Plant Mutualists + HerbivoresVariable (Kempel et al., unpublished)
PlantSoil Plant Mutualists + PollinatorsYes (Gange & Smith 2005)
PlantPathogens + PollinatorsNo (Swope & Parker 2010)
PlantEntomopathogen + HerbivoreYes (Smith, de Lillo & Amrine 2010)
SynergisticPlantPathogens + HerbivoresNo (Smith, de Lillo & Amrine 2010)
PlantHerbivore + Herbivore MutualistNo (Wang, Wu & Ding 2010)
AntagonisticPlantEndophyte + HerbivoreYes (Clay, Holah & Rudgers 2005; Rudgers & Clay 2008; Uchitel, Omacini & Chaneton 2011)
PlantSoil Plant Mutualists + HerbivoresVariable (Kempel et al., unpublished)
PlantInsect vector + PathogenNo (Jiu et al. 2007; Guo et al. 2010; Kaiser et al. 2010)
PlantPathogens + HerbivoresUnknown
InsectAdditiveHerbivoreEntomopathogens + Plant HostNo (Hu et al. 2009; Hajek & Delalibera 2010)
SynergisticHerbivoreEndosymbionts + Plant HostsYes (Hansen et al. 2007; Gueguen et al. 2010; Kaiser et al. 2010; Himler et al. 2011; Giron et al., this issue)
Pathogen VectorPathogens + Plant HostsYes (Hubbes 1999; Gomes et al. 2000; Morse & Hoddle 2006; Hummel et al. 2009; Kaiser et al. 2010; Smith, de Lillo & Amrine 2010)
Pathogen vectorPathogen + Endosymbiont + Plant HostYes (Gottlieb et al. 2010)
AntagonisticPathogen VectorPathogens + Plant HostsUnknown
MicrobeAdditiveEntomopathogenInsect Host + Insect Plant HostsUnknown
EndophytePlant Hosts + HerbivoresYes (Clay, Holah & Rudgers 2005; Rudgers & Clay 2008; Uchitel, Omacini & Chaneton 2011)
SynergisticPlant PathogenPlant Hosts + Insect VectorsYes (Karnosky 1979; Hubbes 1999; Gomes et al. 2000; Morse & Hoddle 2006; De Barro et al. 2008; Hummel et al. 2009; Kaiser et al. 2010; Li et al. 2010; Smith, de Lillo & Amrine 2010; Giron et al., this issue)
EntomopathogenInsect Host + Toxic PlantYes (Keesing et al. 2011)
Plant PathogenPlant Hosts + HerbivoreYes (Castlebury, Rossman & Hyten 2006)
Plant Pathogen + EndosymbiontPlant Hosts + Insect VectorYes (Gottlieb et al. 2010)
Plant PathogenHerbivore + Plant HostsUnknown
AntagonisticPathogenHerbivores + Plant HostsYes (Tack, Gripenberg & Roslin 2012)

This paper will focus on novel PMI interactions created from the introduction of new species. We would expect that new interactions among species might be simple (or additive) in nature. However, there are several cases in which these novel interactions have nonadditive effects. For example, the combination of a plant pathogen and a novel insect vector of that pathogen has led to the loss or damage of 1300 acres of grapevines in California (Gomes et al. 2000): the glassy winged sharpshooter (Homalodisca vitripennis), an invasive insect introduced to California in 1989, has become a vector for the Xylella fastidiosa pathogen, which causes Pierce's disease in grapevines. Here, the synergistic interaction between the invasive insect and native pathogen has promoted the range expansion of the pathogen. In a second example of novel synergistic PMI interactions involving three novel interactions, the invasive shrub Rosa multiflora is fed on by an invasive eriophyid mite that vectors a virus (Smith, de Lillo & Amrine 2010). Prior to their introduction to North America, these organisms had never interacted before. Thus, despite their short association, novel PMI interactions containing invasive species can be synergistic and have dramatic consequences for native and agricultural systems.

These novel PMI interactions may be further promoted by climate change which is likely to influence novel PMI interactions by increasing the introduction of novel species (Hellmann et al. 2008), favouring invasive species over native species and promoting the range expansion of invasive species in their new environments (reviewed in Dukes 2011). We know increases in temperature often increase insect activity (reviewed in Bale et al. 2002), while changes in CO2 and O3 alter both insect (Hillstrom & Lindroth 2008) and soil microbial mutualist communities (Andrew & Lilleskov 2009; but see Klironomos et al. 2005) as well as plant pathogens (Eastburn, McElrone & Bilgin 2011). Climate changes have also been predicted to promote species range shifts and the expansion of invasive species globally (reviewed in Hellmann et al. 2008; Dukes 2011). All of these effects could lead to new encounters between plants, microbes and insects. For example, beech bark disease, the result of an association between the beech scale insect (Cryptococcus fagisuga) and a fungus, is moving northward in its introduced range (Kasson & Livingston 2009). In addition, there are suggestions that the invasive insect herbivore whitefly (Bemisia tabaci), a vector of the economically devastating Gemini viruses, is currently expanding its range in Europe in response to warmer climates (E. Bejarano, pers. comm.). It can be difficult to make sweeping predictions about the influence of climate changes on these interactions (Dukes et al. 2009), but these examples also suggest that climate change could increase the proportion and impact of novel PMI interactions.

The examples above highlight that interactions between plants, microbes and insects can create serious invasive species problems. As demonstrated in this issue, PMI interactions have far-reaching consequences in a wide variety of systems. Our increased knowledge of PMI interactions in invasions has identified three major questions that need addressing: Will novel PMI interactions promote invasive plants, insects and microbes? Can we use novel PMI interactions to inhibit new invasions? Can we predict whether novel PMI interactions are likely to become invasive?

The aim of this paper is to address these three questions with our current knowledge of PMI interactions. In particular, this paper will focus on whether PMI interactions promote or inhibit invasive plants, insects or microbes.

Invasive plants

Can PMI interactions promote invasive plants in their new environments?

We might expect that novel PMI interactions with additive or synergistic effects would be most likely to promote invasive species in their new environment (Table 1). However, antagonistic PMI interactions that hinder a pathogen or herbivore may also promote invasive species fitness. This section highlights cases where all three types of interactions promote invasive plants.

Plant–endophyte interactions have been shown to promote the spread of invasive relative to native plant species due to their antagonistic effects on herbivores (Clay, Holah & Rudgers 2005; Rudgers & Clay 2008; Uchitel, Omacini & Chaneton 2011). Endophytes are commonly defined as microbes that live within host plant tissues without negatively impacting their host fitness (Wilson 1995). Endophyte–plant interactions range from commensal through to mutualistic and were first identified in grasses (Bacon et al. 1977) but are now also known to occur in a wide variety of other plant species including several herbaceous (Gange et al. 2007), noncereal crop (Hallmann & Sikora 1994; Hallmann et al. 1997; Raps & Vidal 1998; Sturz, Christie & Nowak 2000; Garbeva et al. 2001; Jallow, Dugassa-Gobena & Vidal 2004; Hashiba & Narisawa 2005) and tree species (Arnold et al. 2000; Vega et al. 2008; Albrectsen et al. 2010; Newcombe, Martin & Kohler 2010). Some endophytes have been shown to increase invasive plant species fitness (Clay 1988; Rudgers, Koslow & Clay 2004; Aschehoug et al. 2012). While endophytes can confer many different benefits, research has frequently shown negative effects of endophytes on mammalian and insect herbivores (Raps & Vidal 1998; Jallow, Dugassa-Gobena & Vidal 2004; Clay, Holah & Rudgers 2005; Rudgers & Clay 2008; Vega et al. 2008; Newcombe, Martin & Kohler 2010; Menjivar et al. 2012; Zhang et al. 2012), although these effects can also be influenced by nutrients and other biotic factors (Saikkonen et al. 2006; Eschen et al. 2010). In some cases, the negative effects of endophytes on herbivores can be passed on to the second generation of herbivores – even if they are not feeding on the endophyte-infected host plant (Jaber & Vidal 2010) – or onto predators or parasitoids of the herbivores (de Sassi, Muller & Krauss 2006; Harri, Krauss & Muller 2008b) or even hyperparasitoids (Harri, Krauss & Muller 2008a). The successful discouragement of herbivores in many grass–endophyte combinations has been shown to be due to the production of highly toxic alkaloid compounds with universal effects against both mammalian and insect herbivores (Bacon et al. 1977). The most striking aspect of these effects, however, is that they can be replicated in a wide variety of geographically spatially distinct locations (Uchitel, Omacini & Chaneton 2011). The lack of spatial specificity in some endophyte–herbivore effects suggests two things: first, endophyte effects are independent of herbivore species, and second, abiotic factors such as soil structure are not likely to hinder endophyte effects on herbivores. As a result, a plant species infected with an effective antiherbivore endophyte could invade and be successful in almost any environment.

Interactions between soil-dwelling plant mutualists, plants and insects are also a frequently studied combination of PMI interaction involving invasive plants, but the effects are less consistent and can be either additive or nonadditive. These plant mutualists include nitrogen-fixing bacteria that fix atmospheric nitrogen for their host plants and mycorrhizal fungi that increase the uptake of limiting nutrients (e.g. N and P). All research in the area of PMI influences on invasive plants has focused on arbuscular mycorrhizal (AM) fungi. Recently, Kempel and colleagues examined the defence responses to herbivory of invasive and noninvasive introduced plant species associated with AM fungi (Kempel et al., 2013) and found no overall interaction between invasive status and mycorrhizal status on plant defence against herbivory. Instead, the influence of AM fungi on induced defences was both plant and herbivore species specific. In an antagonistic example, in the invasive Bidens frondosa AM fungi promoted induced defences over constitutive defences for one, but not a second, herbivore. In other cases, AM fungi suppressed induced defences against herbivores (Bennett, Bever & Bowers 2009; Kempel et al., 2013). Thus, mutualistic soil microbes can influence the fitness of herbivores of introduced plant species by manipulating plant defence systems, although we are a long way from making consistent predictions about how soil mutualistic microbes are likely to influence plant defence in invasive species.

Soil mutualistic microbes may also have an additive influence on pollinating insects, although significantly less is known about the outcome of these pollinator interactions with invasive plant species in general (Stout & Morales 2009). AM fungi increased pollinator visitation rates to two noninvasive introduced species (Tagetes eracta and Tagetes patula; Gange & Smith 2005). Some authors have proposed that if a plant species (invasive or otherwise) responds positively to a mutualistic soil microbe, then that microbe will likely promote plant–pollinator interactions (Cahill et al. 2008), thereby promoting invasion. To date, all studies of pollinator PMI interactions and invasive plants have utilized AM fungi, and thus, the effects of other soil mutualistic microbes on pollination of invasive plant species is still an open question.

As demonstrated above, novel PMI interactions can promote invasive plant species, but there are also several examples of invasive plant species becoming successful because they reduce (or avoid) interactions with microbes or insects. Invasive species often have reduced responses to their soil communities (Kulmatiski 2006) and herbivores, although these effects may not be correlated (Morrien, Engelkes & van der Putten 2011). This could be due to the frequent ability of invasive plant species to alter their soil community (reviewed in van der Putten, Klironomos & Wardle 2007; Raizada, Raghubanshi & Singh 2008; Sanon et al. 2009) and disrupt or diminish the mycorrhizal fungal–plant mutualism (Vogelsang & Bever 2009). In addition, the Enemy Release Hypothesis suggests that invasive plant species are successful due to a reduction in interactions with microbial pathogens and insect herbivores in their new environment (Elton 1958; Keane & Crawley 2002; Liu & Stiling 2006), and there are several examples demonstrating this strategy (reviewed in Liu & Stiling 2006). There are also several examples of invasive species that use allelopathy to limit both microbes and insects (Weidenhamer & Callaway 2010; Inderjit et al. 2011). For example, Vincetoxicum rossicum (Mogg et al. 2008), Alliaria petiolata (Burke 2008; Wolfe et al. 2008; Barto et al. 2011; Cantor et al. 2011; Keesing et al. 2011) and Centaurea maculosa (Mummey & Rillig 2006; Broz, Manter & Vivanco 2007) have been shown to limit microbial population growth using allelopathic chemicals and, in the case of both A. petiolata (Stinson et al. 2006) and C. maculosa (reviewed in Callaway & Ridenour 2004), have been shown to limit microbial interactions in neighbouring plants. These influences on soil microbial communities can directly (Keesing et al. 2011) or indirectly influence insect–plant interactions. As a result, we must consider that in some cases the success of invasive species may be due to reduced novel interactions with insects and microbes.

Can PMI interactions inhibit invasive plants in their new environments?

Most research efforts on understanding the inhibitory effects of PMI interactions on invasive plants have been dedicated to finding herbivores and pathogens that could act as biocontrol agents. In addition, most biocontrol efforts have not identified microbial–insect interactions that inhibit invasive plants and then released them, but instead have found microbial–insect interactions that have formed after introduction. In addition, some biocontrol efforts introduced multiple insect and microbial agents with the expectation that they act additively [e.g. biocontrol herbivores (Eustenopus villosus and Chaetorellia succinea) and a fungal rust (Puccinia jaceae var. solstitialis) on yellow starthistle (Centaurea solstitialis)]. However, there are several possible PMI interactions that could inhibit invasive plant species that have rarely been explored, such as interactions between pollinators and pathogens (Swope & Parker 2010; Table 1).

Microbes and insects introduced as biocontrol agents can influence each other in additive (Paynter & Hennecke 2001), synergistic or antagonistic ways. In an example of an additive inhibitive effect on an invasive species, the biocontrol weevil Oxyops vitosa, psyllid Boreioglycaspis melaleucae and rust fungus Puccinia psidii all negatively influence the invasive tea tree (Melaleuca quinquenervia) in Florida. When combined, the negative effect of each organism on tea trees is increased by the negative effect of each additional insect or microbe (Rayamajhi et al. 2010). The effects of both insect herbivores (E. villosus and C. succinea) and a fungal rust, P. jaceae var. solstitialis, on yellow starthistle (C. solstitialis) have also been shown to be additive as there is no interaction between the presence of insects and fungi on starthistle fitness (O'Brien et al. 2010). In a study in the same system, P. jaceae increased adult feeding but reduced larval feeding by E. villosus, resulting in little net impact and no influence on visiting insect pollinators (Swope & Parker 2010).

There are also cases where the presence of both insects and microbes has nonadditive effects that limit invasive species. For example, fungal root pathogens (Fusarium and Rhizoctonia species) had the greatest effects on leafy spurge (Euphorbia esula-virgata) when roots experienced insect herbivory (Kremer, Caesar & Souissi 2006). The negative effects of a leafhopper (Erythroneurini spp.) and the fungal pathogen Puccinia myrsiphylli on the invasive Asparagus asparagoides are significantly greater than if their individual effects are added (Turner et al. 2010). Also, Rosa multiflora is best controlled by the combination of an eriophyid mite that transmits a pathogenic virus (reviewed in Smith, de Lillo & Amrine 2010). A future goal of biocontrol PMI interactions will be to identify insect–microbe interactions with additive and nonadditive effects like those above before release.

There are some biocontrol candidates currently being explored that rely heavily on PMI interactions. One such case involves Japanese knotweed (Fallopia japonica). A series of surveys have identified a weevil (Euops chinesis) with high specificity for F. japonica, and this specificity appears to be due to a particular fungus that occurs in the leaf rolls produced by E. chinesis when feeding on F. japonica. This fungus does not occur in the leaf rolls of E. chinesis on any other host plant, and E. chinesis cannot survive on alternative host plants (Wang, Wu & Ding 2010). Thus, incorporating PMI interactions is opening a whole new field of biocontrol exploration that may lead to the discovery of biocontrol agents with greater host specificity and efficacy.

Invasive insects

Can PMI interactions promote invasive insects in their new environments?

Just like invasive plants, invasive insects can be subject to additive, synergistic and antagonistic interactions with plants and microbes. Several examples of interactions between invasive insects and plants and microbes have been heavily explored (e.g. insects acting as vectors, endosymbionts), but other interactions have never been examined (e.g. influence of plant pathogens on invasive insect herbivore fitness; Table 1).

A potential synergistic PMI interaction involves bacterial endosymbionts which can help insects adapt to new environments quickly and can contribute to the invasive ability of insects. One of the major discoveries of the last few decades is that a number of insects, not just aphids, host bacterial endosymbionts that confer traits that are not present in the insect genome (reviewed in Feldhaar 2011). These endosymbionts fall into two categories: obligate (often called primary) and facultative (often called secondary). Insects cannot survive without obligate/primary endosymbionts that typically perform a vital integrative function within the insect. The most well-known example of a primary endosymbiont is the Buchnera aphidicola of aphids that synthesize tryptophan and other essential amino acids for their hosts (reviewed in Shigenobu & Wilson 2011; McCutcheon & Moran 2012). In an extension of PMI interactions, B. aphidicola have been shown to produce a protein, GroEL, that binds to the economically damaging plant pathogens luteo- and poleroviruses as they pass through the aphid and prevent the viruses from being degraded (Ishikawa, Yamaji & Hashimoto 1985; Baumann, Baumann & Clark 1996; Humphreys & Douglas 1997; Douglas 1998; Bouvaine, Boonham & Douglas 2011), thereby increasing the negative influence of the virus on the host plant fed on by the aphid.

By contrast, secondary symbionts are not obligate and can confer benefits in certain environments (Oliver et al. 2010). Secondary endosymbionts and their rapid evolutionary responses to environmental change have been suggested to contribute to the world-wide invasion of B. tabaci (Gueguen et al. 2010; Himler et al. 2011). Another example of endosymbionts contributing to insect invasions involves a biocontrol parasitoid, Psyllaphaegus bliteus, introduced to California to control the invasive herbivore, Glycaspis brimblecombei, of the invasive Eucalyptus camaldulensis tree. A survey of the endosymbionts of G. brimblecombei revealed that endosymbionts that conferred resistance to P. bliteus were present in higher frequencies in populations where P. bliteus was present, therefore reducing the efficacy of the biocontrol agent (Hansen et al. 2007). Given that researchers are just beginning to identify the presence of secondary endosymbionts in many organisms and the function of many of these endosymbionts has yet to be determined, it seems likely that future research will reveal a large role of endosymbionts in novel synergistic PMI interactions.

In addition to carrying adaptive endosymbionts, herbivores can also host plant-attacking microbes as endosymbionts that can have synergistic effects and increase herbivore fitness. For example, the leafmining moth, Phyllonorycter blancardella, appears to host a Wolbachia bacterium that manipulates plant cytokinin levels (Kaiser et al. 2010). The manipulation of host plant physiology increases P. blancardella fitness by preventing plants from senescing attacked leaves and forcing the plant to continue to deliver nutrients to areas where P. blancardella is feeding (Kaiser et al. 2010, Giron et al., 2013). While this is so far a unique mechanism of action for an insect endosymbiont, it seems likely that continued explorations of insect endosymbionts and plant pathogens will reveal similar cases.

Antagonistic novel PMI interactions can also promote invasive insects. While it has been shown, as discussed above, that toxic and allelopathic plant compounds can negatively impact both insects and microbes, in some cases herbivores can benefit from these compounds due to a reduced antagonistic microbial load. This could be due to sequestration of toxic compounds [e.g. Junonia coenia herbivores sequester toxins from a novel host plant P. lanceolata (Bowers 1984)], the consumption of toxic compounds that have stronger effects on microbes than on the host insect, or due to the influence of the toxic compounds in the medium (e.g. soil) in which the insect lives. For example, allelochemicals from the invasive Allaria petiolata plant released into the soil negatively influence the generalist entomopathogenic fungus Beauveria bassiana, thereby reducing the load of B. bassiana hosted by waxworms and potentially promoting waxworm fitness in the presence of Allaria petiolata (Keesing et al. 2011). This type of positive influence on insects has rarely been explored in the context of invasions, and it may be likely that in some cases insects benefit from associating with toxic plants (allelopathic, endophytic or otherwise) which may promote their invasive ability in environments where other insects cannot survive.

One of the most disastrous synergistic effects invasive insects can have on an ecosystem is to introduce new microbial pathogens or to increase the spread of native microbial pathogens. In some cases, invasive insect species act as partially co-evolved vectors, whereas in other cases invasive insects pick up new associations or incidentally transfer pathogens. There are several cases of invasive insect vectors. For example, the introduction of the glassy winged sharpshooter (H. vitripennis) is increasing the spread of the native bacterial pathogen Pierce's disease (X. fastidiosa) among grapevines in California (Gomes et al. 2000). The invasive panicle rice mite has been spreading through the Caribbean and Central America where it devastates rice crops. However, much of the damage caused by panicle rice mite may actually be due to the interaction between the mite and two rice pathogens Sarocladium oryzae and Burkholderia glumae (reviewed in Hummel et al. 2009). Thrips (insects of the order Thysanoptera) vector a wide variety of viral pathogens, are common invaders in agricultural systems and have received a great deal of attention due to the limited mechanisms available for controlling them or the diseases they spread (reviewed in Morse & Hoddle 2006). The combination of invasive insects and their microbial associates can therefore cause significant economic damage to agricultural systems.

Like agricultural systems, forest systems are being hit hard by combinations of insect pests and their associated microbes (Hulcr & Dunn 2011). Novel insect–microbe combinations are changing the composition of forests in Europe and North America by eliminating trees which often act as keystone species (Loo 2009), and altering ecosystem properties such as carbon cycling (Hicke et al. 2012). The effects of some of these novel insect–microbe interactions are predicted to become more severe with the onset of climate change (Dukes et al. 2009). For example, the Asian ascomycete pathogenic fungi, Ophiostoma ulmi, Ophiostoma himal-ulmi and Ophiostoma novo-ulmi (Harrington et al. 2001), the cause of Dutch elm disease in Europe and North America, have led to significant damage. The spread of Dutch elm disease in North America was likely increased due to the introduction of European elm bark beetle (Scolytus multistriatus), which along with the native elm bark beetle (Hylurgopinus rufipes) act as vectors for Dutch elm disease (Karnosky 1979; Hubbes 1999). In this case, an invasive insect spreads an invasive pathogen which the insect had never previously encountered before introduction. Beech bark disease is the result of another novel insect–microbe combination causing havoc in forest systems. The introduction of the beech scale (C. fagisuga) into Europe and North America has led to the spread of beech bark disease. Unlike S. multistriatus, which acts as a vector for fungal pathogens, beech scale does not vector the North American (Neonectria faginata) and European (Neonectria galligena) fungi (Castlebury, Rossman & Hyten 2006) that produce beech bark disease. Instead, it simply creates the wounds in beech bark that are easily colonized by the fungi (Houston 1994). Although not a tree, R. multiflora is a common invader of forest and other habitats. It also suffers from a novel invasive pathogen–insect combination involving the introduced eriophyid mite (Phyllocoptes fructiphilus) that transmits a pathogenic virus to R. multiflora while feeding. None of these organisms, plant, mite or virus, had previously encountered one another before introduction; however, the combination is negatively impacting the invasive R. multiflora (reviewed in Smith, de Lillo & Amrine 2010).

Awareness of the ability of introduced and invasive insects to spread disease has had impacts on the field of biocontrol. Biocontrol insect candidates are now often being screened for their ability to spread native and invasive microbial pathogens, and rejected if they show the potential to do so (Lennox et al. 2009). Almost all of the examples above are from agricultural or forestry systems for which invasive insects spreading microbial pathogens cause significant economic impacts. However, it seems likely that invasive insects and their associated microbes are impacting other systems as well; however, more research attention will need to be dedicated to determining those effects.

Can PMI interactions inhibit invasive insects in their new environments?

Research into PMI interactions that can limit invasive insect plant pests has also focused on identifying biocontrol agents from the original habitat of the invasive insect that have additive or antagonistic interactions. In particular, entomopathogenic microbes are often explored as potential biocontrol agents of insect plant pests (Cory & Ericsson 2010; Hajek & Delalibera 2010). For example, the invasion of a forest tree pest, the Asian longhorned beetle (Anoplophora glabripennis), in North America and Europe is partially controlled by entomopathogenic fungi (Hu et al. 2009). It has been proposed that a combination of entomopathogenic fungi and manipulation of tree communities to include preferred and nonpreferred host trees would provide the best control of A. glabripennis (Hu et al. 2009). As a result, entomopathogens, particularly entomopathogenic fungi, are gaining increasing attention as potential biocontrol agents of invasive insects. Researchers are currently exploring the efficacy of entomopathogenic fungi for controlling the spread of B. tabaci in the UK (Cuthbertson et al. 2011). However, research also shows that nonadditive interactions with the plant can determine the efficacy of entomopathogenic fungi as plant chemistry can influence the success of fungal attack of insects (Cory & Ericsson 2010). This further demonstrates the need to incorporate all three organisms (plants, insects and microbes) into biocontrol studies.

Invasive microbes

Can PMI interactions promote invasive microbes in their new environments?

There has been less research identifying novel additive, synergistic and antagonistic interactions that influence the promotion of invasive microbes (Table 1). In particular, there are no recorded cases of insect interactions helping invasive microbes overcome the influence of negative plant interactions, and there are no known examples of positive plant interactions allowing invasive microbes to overcome a negative interaction with an insect (Table 1). As a result, there is plenty of opportunity for exploring these interactions in future.

Instead, the vast majority of our understanding of how invasive microbes interact with insects and plants derives from the synergistic interaction of insect vectors of microbes. In the previous section, we discussed several examples of invasive microbes vectored by insects, including panicle rice mite increasing the spread or effects of the pathogens S. oryzae and B. glumae (reviewed in Hummel et al. 2009), vectoring of the three invasive fungal Ophiostoma spp. that cause Dutch elm disease (Karnosky 1979; Hubbes 1999; Harrington et al. 2001), and the spread of a number of microbial pathogens by thrips (reviewed in Morse & Hoddle 2006). Another concerning example is the case of the whitefly (B. tabaci) and the spread of Geminiviruses. The invasive agricultural pest B. tabaci is spreading around the world, including sites as disparate as Indonesia (De Barro et al. 2008), Costa Rica (Guevara-Coto et al. 2011) and La Réunion (Delatte et al. 2009). B. tabaci is an efficient vector of Geminiviruses, particularly begomoviruses (De Barro et al. 2008; Li et al. 2010), so wherever B. tabaci appears, diseases produced by Geminiviruses, such as lettuce infectious yellows virus, tobacco curly shoot virus, tomato yellow leaf curl virus and African cassava mosaic virus, appear or increase shortly afterwards. In some cases, the interaction between viruses such as tomato yellow leaf curl virus and B. tabaci increases B. tabaci fecundity and fitness (Jiu et al. 2007; Guo et al. 2010). The transmission of viruses by B. tabaci to plants requires a protein created by an endosymbiotic bacteria (Gottlieb et al. 2010), thereby increasing the complexity of this novel synergistic PMI interaction. The introduction of B. tabaci and its associated begomoviruses has caused complete crop losses, severe economic losses, and contributed to famines in some areas of the world (Thompson 2011).

There is the potential for antagonistic PMI interactions involving apparent trade-offs between plant defensive systems (such as the JA and SA pathways) to promote invasive pathogens. Although not a PMI interaction involving an invasive species, on the Barbarea vulgaris plant in Denmark white rust is rarely observed – unless the plant has received herbivory (van Molken, unpublished). This suggests, as has been previously shown, that some plant defensive activities can limit the activity of other plant defensive activities (reviewed in Beckers & Spoel 2006). This is a potential mechanism for insect and microbe influence that has rarely been explored in the context of invasive insects and microbes.

It may also be possible that invasive microbes can act antagonistically towards plants and microbes. In Finland, the presence of an invasive mildew on host trees decreases both herbivore number and herbivore species number on host plants (Tack, Gripenberg & Roslin 2012). In this case, there may be unobserved direct effects between the mildew and insects, or the mildew may be able to manipulate host defences to its advantage. Also, as discussed above, the spread of endophyte-infected host plants increases the spread of their endophytes with potentially negative influences on insect herbivore communities (Rudgers & Clay 2008). As a result, there are still several potential unexplored mechanisms describing the interactions between invasive microbes, insects and plants.

Can PMI interactions inhibit invasive microbes in their new environments?

Relatively little research has identified examples where PMI interactions inhibit invasive microbes within their new environments (Table 1). There are several possible avenues by which PMI interactions are likely to inhibit invading microbes. For example, microbes that rely heavily on a vector are only likely to increase their spread if their vector increases their range or is introduced. As a result, microbes that rely on vectors not present in the new environment could cause extensive plant diseases but are limited by the lack of a synergistic interaction. As a result, implementing measures that limit the spread of vectors should also limit vectored invasive microbes. In addition, it seems likely that in some cases, antagonistic interactions like interactions between defensive strategies within host plants described above may move in the opposite direction – herbivory may suppress microbial infection. The lack of research in this area opens up great opportunities for exploring how PMI interactions could suppress invasive microbes.

What is the future of the study of invasive plants, microbes and insects involved in PMI interactions?

Novel PMI interactions are likely to increase in future following increased introductions of species and the influence of climate changes. In addition, the majority of documented successful novel PMI interactions are not additive, but instead nonadditive (Table 1). As a result, the future of the study of novel PMI interactions will need to involve both predictive and preventative components if we are to counter these new invasions.

An obvious place to begin research on predicting invasive novel PMI interactions is with insect vectors and the microbial pathogens they promote. In this review, insects and the microbes they vector were the most common novel PMI interaction and the group of novel PMI interactions causing the most economic damage. While their documentation and economic costs may not be independent factors, their large impact does suggest they are among the most frequent novel PMI interactions. In one pathogen–insect vector system, a secondary endosymbiont is credited with improving insect and pathogen matching (Gottlieb et al. 2010). Therefore, utilizing techniques such as next-generation DNA sequencing, genomics, transcriptomics and metabolomics combined with traditional ecological experiments will likely identify whether endosymbionts are responsible for novel pathogen–insect vector interactions or whether other mechanisms promote these novel interactions in different systems. Researching the factors that allow insects to vector a wide variety of microbes and the factors that allow pathogenic microbes to utilize novel vectors is a first step to developing a predictive framework for identifying novel potentially invasive PMI interactions.

Utilizing multiple synergistic and antagonistic interactions to limit invasive species should also be a priority in the field of PMI interactions. In addition to identifying potential biocontrol PMI interactions (Wang, Wu & Ding 2010), there are common PMI interactions (e.g. soil mutualistic microbes associating with plants eaten by herbivores, endophyte-infected plants that suppress herbivory) that occur in most systems that could possibly be manipulated to limit invasive species. In several cases, microbial plant mutualists have been shown to increase plant defences against herbivores (Clay, Holah & Rudgers 2005; Rudgers & Clay 2008; Gehring & Bennett 2009; Uchitel, Omacini & Chaneton 2011). A second means of targeting invasive insects could be to target their endosymbionts. Eliminating endosymbionts could limit invasive insect adaptation to new host plants and even kill host insects. There is currently a developing field of research focused on community sequencing of pest insects (invasive and otherwise) aimed at identifying endosymbionts (Lefevre et al. 2004; Andersen et al. 2010; Bai et al. 2010; Hail et al. 2011; Hail, Dowd & Bextine 2012). Identification of endosymbionts is the first step to their elimination and the potential elimination of invasive insects. Thus, there are exciting future opportunities for utilizing common and novel synergistic PMI interactions to target invasive species.


The author would like to thank James Hourston, Pete Goddard, Alison Karley, Adrian Newton, Arjen Biere and two anonymous reviewers for comments on the manuscript. This work was funded by the Rural and Environment Science and Analytical Services Division Workpackage Themes 1·1 and 3·3.