BackgroundObligate bacterial primary (P-) endosymbionts that are maternally inherited and codiverge with hosts are widespread across insect lineages with nutritionally restricted diets. Secondary (S-) endosymbionts are mostly facultative, but in some hosts, they complement P-endosymbiont function and therefore become obligate. Phylogenetic evidence exists for host switching and replacement of S-endosymbionts. The community dynamics that precede endosymbiont replacement and complementation have been little studied across host species, yet they are fundamental to the evolution of endosymbiosis.ResultsWe performed bacterial 16S rRNA gene amplicon sequencing of 25 psyllid species (Hemiptera, Psylloidea) across different developmental stages and ecological niches by focusing on the characterisation of the bacteria other than the universally present P-endosymbiont Carsonella (Gammaproteobacteria). Most species harboured only one dominant representative of diverse gammaproteobacterial S-endosymbionts that was consistently detected across all host individuals and populations (Arsenophonus in eight species, Sodalis or Sodalis-like bacteria in four species, unclassified Enterobacteriaceae in eight species). The identity of this dominant obligate S-endosymbiont varied across closely related host species. Unexpectedly, five psyllid species had two or three co-occurring endosymbiont species other than Carsonella within all host individuals, including a Rickettsiella-like bacterium (Gammaproteobacteria) in one psyllid species. Based on standard and quantitative PCR, all psyllids carried Carsonella, at higher titres than their dominant S-endosymbionts. Some psyllids also had Alphaproteobacteria (Lariskella, Rickettsia, Wolbachia) at varying prevalence. Incidence of other bacteria, including known plant pathogens, was low. Ecological niche of gall-forming, lerp-forming and free-living psyllid species did not impact endosymbiont communities. Two flush-feeding psyllid species had population-specific differences, and this was attributable to the higher endosymbiont diversity in native ranges and the absence of some endosymbionts in invasive ranges.ConclusionsOur data support the hypothesis of strict vertical transmission of minimal core communities of bacteria in psyllids. We also found evidence for S-endosymbiont replacement across closely related psyllid species. Multiple dominant S-endosymbionts present in some host species, including at low titre, constitute potential examples of incipient endosymbiont complementation or replacement. Our multiple comparisons of deep-sequenced minimal insect bacterial communities exposed the dynamics involved in shaping insect endosymbiosis.Electronic supplementary materialThe online version of this article (doi:10.1186/s40168-017-0276-4) contains supplementary material, which is available to authorized users.

Coevolution between insects and bacterial endosymbionts contributes to the success of many insect lineages. For the first time, we tested for phylogenetic codivergence across multiple taxonomic scales, from within genera to superfamily between 36 psyllid species of seven recognised families (Hemiptera: Psylloidea), their exclusive primary endosymbiont Carsonella and more diverse secondary endosymbionts (S-endosymbionts). Within Aphalaridae, we found that Carsonella and S-endosymbionts were fixed in one Glycaspis and 12 Cardiaspina populations. The dominant S-endosymbiont was Arsenophonus, while Sodalis was detected in one Cardiaspina species. We demonstrated vertical transmission for Carsonella and Arsenophonus in three Cardiaspina species. We found strong support for strict cospeciation and validated the informative content of Carsonella as extended host genome for inference of psyllid relationships. However, S-endosymbiont and host phylogenies were incongruent, and displayed signs of host switching and endosymbiont replacement. The high incidence of Arsenophonus in psyllids and other plant sap-feeding Hemiptera may be due to repeated host switching within this group. In two psyllid lineages, Arsenophonus and Sodalis genes exhibited accelerated evolutionary rates and AT-biases characteristic of long-term host associations. Together with strict vertical transmission and 100% prevalence within host populations, our results suggest an obligate, and not facultative, symbiosis between psyllids and some S-endosymbionts.

1 Climate, resource availability and natural enemies are pivotal factors influencing population fluctuations of insects. However, the principal factors underlying insect outbreaks, particularly in natural ecosystems, are still debated and may vary between species. 2 We assessed the role of these factors on psyllid population dynamics during the peak and decline of a Cardiaspina psyllid outbreak in a critically endangered eucalypt woodland in Australia. Initially, this involved describing the field biology and ecology of a newly reported Cardiaspina species on grey box (Eucalyptus moluccana Roxb.). 3 Within 1 year, the psyllid completed four generations. Its biology and parasitoid complex were similar to other Cardiaspina species during outbreaks. 4 Minimum winter temperature was a key driver of psyllid development and density.Natural enemies did not prevent or control this outbreak. 5 The outbreak resulted in area-wide and chronic defoliation of host trees. Resource depletion and summer heat waves impacting critical developmental stages of psyllids were the major factors responsible for the significant reduction of psyllid populations in early 2013. However, ongoing regeneration of trees in the highly fragmented woodlands may allow recolonization of new foliage and chronic infestations to continue.

Some herbivores can modify the physiology of plant modules to meet their nutritional requirements. Induction of premature leaf senescence could benefit herbivores since it is associated with the mobilisation of nutrients. We compared the effects of nymphal feeding by Cardiaspina near densitexta on Eucalyptus moluccana with endogenous processes associated with senescence to assess the relative merits of an insect manipulation or plant defence interpretation of responses. Evidence supporting insect manipulation included increased size of fourth and fifth instar nymphs (in the latter the effect was restricted to forewing pad length of females) on leaves supporting high numbers of conspecifics and feeding preventing leaf necrosis. Intra-specific competition negated greater performance at very high densities. High and very high abundances of nymphs were associated with increased concentrations of amino acid N but only very high abundances of nymphs tended to be associated with increased concentrations of six essential amino acids. Contrary to the insect manipulation interpretation, feeding by very high abundances of nymphs was associated with significant reductions in chlorophyll, carotenoids and anthocyanins. Evidence supporting plant defence included the severity of chlorosis increasing with the abundance of nymphs. Leaf reddening did not develop because ambient conditions associated with photoinhibition (high irradiance and low temperature) were not experienced by leaves with chlorotic lesions. Leaf reddening (from anthocyanins) alone is not expected to adversely affect nymphal survival; only leaf necrosis would kill nymphs. For senescence-inducing psyllids, nutritional enhancement does not fit neatly into either an insect manipulation or plant defence interpretation.

Frequency and severity of insect outbreaks in forest ecosystems are predicted to increase with climate change. How this will impact canopy leaf area in future climates is rarely tested. Here, we document function of insect outbreaks that fortuitously and rapidly occurred in an ecosystem under free-air CO enrichment. Over the first 2 years of CO fumigation of a naturally established mature Eucalyptus woodland, we continuously assessed population responses of three sap-feeding insect species of the psyllid genera Cardiaspina, Glycaspis and Spondyliaspis for up to ten consecutive generations. Concurrently, we quantified changes in the canopy leaf area index (LAI). Large and rapid shifts in psyllid community composition were recorded between species with either flush (Glycaspis) or senescence-inducing (Cardiaspina, Spondyliaspis) feeding strategies. Within the second year, two psyllid species experienced significant and rapid population build-up resulting in two consecutive outbreaks: first, rainfall stimulated Eucalyptus leaf production increasing LAI, which supported population growth of flush-feeding Glycaspis without impacting LAI. Glycaspis numbers then crashed and were followed by the outbreak of senescence-feeding Cardiaspina fiscella that led to significant defoliation and reduction in LAI. For all three psyllid species, the abundance of lerps, protective coverings excreted by the sessile nymphs, decreased at e[CO ]. Higher lerp weight at e[CO ] for Glycaspis but not the other psyllid species provided evidence for compensatory feeding by the flush feeder but not the two senescence feeders. Our study demonstrates that rainfall drives leaf phenology, facilitating the rapid boom-and-bust succession of psyllid species, eventually leading to significant defoliation due to the second but not the first outbreaking psyllid species. In contrast, e[CO ] may impact psyllid abundance and feeding behaviour, with psyllid species-specific outcomes for defoliation severity, nutrient transfer and trophic cascades. Psyllid populations feeding on Eucalyptus experience rapid boom-and-bust cycles depending on availability of suitable foliage driven by rainfall patterns and leaf phenology.

BackgroundParasitoids are hyperdiverse and can contain morphologically and functionally cryptic species, making them challenging to study. Parasitoid speciation can arise from specialisation on niches or diverging hosts. However, which process dominates is unclear because cospeciation across multiple parasitoid and host species has rarely been tested. Host specificity and trophic interactions of the parasitoids of psyllids (Hemiptera) remain mostly unknown, but these factors are fundamentally important for understanding of species diversity, and have important applied implications for biological control.ResultsWe sampled diverse parasitoid communities from eight Eucalyptus-feeding psyllid species in the genera Cardiaspina and Spondyliaspis, and characterised their phylogenetic and trophic relationships using a novel approach that forensically linked emerging parasitoids with the presence of their DNA in post-emergence insect mummies. We also tested whether parasitoids have cospeciated with their psyllid hosts. The parasitoid communities included three Psyllaephagus morphospecies (two primary and, unexpectedly, one heteronomous hyperparasitoid that uses different host species for male and female development), and the hyperparasitoid, Coccidoctonus psyllae. However, the number of genetically delimited Psyllaephagus species was three times higher than the number of recognisable morphospecies, while the hyperparasitoid formed a single generalist species. In spite of this, cophylogenetic analysis revealed unprecedented codivergence of this hyperparasitoid with its primary parasitoid host, suggesting that this single hyperparasitoid species is possibly diverging into host-specific species. Overall, parasitoid and hyperparasitoid diversification was characterised by functional conservation of morphospecies, high host specificity and some host switching between sympatric psyllid hosts.ConclusionsWe conclude that host specialisation, host codivergence and host switching are important factors driving the species diversity of endoparasitoid communities of specialist host herbivores. Specialisation in parasitoids can also result in heteronomous life histories that may be more common than appreciated. A host generalist strategy may be rare in endoparasitoids of specialist herbivores despite the high conservation of morphology and trophic roles, and endoparasitoid species richness is likely to be much higher than previously estimated. This also implies that the success of biological control requires detailed investigation to enable accurate identification of parasitoid-host interactions before candidate parasitoid species are selected as biological control agents for target pests.Electronic supplementary materialThe online version of this article (doi:10.1186/s12862-017-0959-2) contains supplementary material, which is available to authorized users.

Summary Global climate and atmospheric change are widely predicted to affect many ecosystems. Herbivorous insects account for 25% of the planet's species so their responses to environmental change are pivotal to how future ecosystems will function. Atmospheric change affects feeding guilds differently, however, with sap‐feeding herbivores consistently identified as net beneficiaries of predicted increases in atmospheric carbon dioxide concentrations (eCO2). The mechanistic basis for these effects remains largely unknown, and our understanding about how multiple environmental changes, acting in tandem, shape plant–insect interactions is incomplete. This study investigated how increases in temperature (eT) and eCO2 affected the performance of the pea aphid (Acyrthosiphon pisum) via changes in amino acid concentrations in the model legume, lucerne (Medicago sativa). Aphid performance increased under eCO2 at ambient temperatures, whereby aphid fecundity, longevity, colonization success and rm increased by 42%, 30%, 25% and 21%, respectively. eT negated the positive effects of eCO2 on both fecundity and rm, however, and performance was similar to when aphids were reared at ambient CO2. We identified discrete functional groups of amino acids that underpinned the effects of climate and atmospheric change, in addition to plant genotype, on aphid performance. Effects of eT and eCO2 held true across five M. sativa genotypes, demonstrating the generality of their effects. Combining this knowledge with amino acid profiles of existing cultivars raises the possibility of predicting future susceptibility to aphids and preventing outbreaks of a global pest. Moreover, environmentally induced changes in the nutritional ecology of aphids have the capacity to change life‐history strategies of aphids and their direct and indirect interactions with many other organisms, including mutualists and antagonists. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.12715/suppinfo is available for this article.

Insect herbivore outbreaks frequently occur and this may be due to factors that restrict top-down control by parasitoids, for example, host-parasitoid asynchrony, hyperparasitization, resource limitation and climate. Few studies have examined host-parasitoid density relationships during an insect herbivore outbreak in a natural ecosystem with diverse parasitoids. We studied parasitization patterns of Cardiaspina psyllids during an outbreak in a Eucalyptus woodland. First, we established the trophic roles of the parasitoids through a species-specific multiplex PCR approach on mummies from which parasitoids emerged. Then, we assessed host-parasitoid density relationships across three spatial scales (leaf, tree and site) over one year. We detected four endoparasitoid species of the family Encyrtidae (Hymenoptera); two primary parasitoid and one heteronomous hyperparasitoid Psyllaephagus species (the latter with female development as a primary parasitoid and male development as a hyperparasitoid), and the hyperparasitoid Coccidoctonus psyllae. Parasitoid development was host-synchronized, although synchrony between sites appeared constrained during winter (due to temperature differences). Parasitization was predominantly driven by one primary parasitoid species and was mostly inversely host-density dependent across the spatial scales. Hyperparasitization by C. psyllae was psyllid-density dependent at the site scale, however, this only impacted the rarer primary parasitoid. High larval parasitoid mortality due to density-dependent nymphal psyllid mortality (a consequence of resource limitation) compounded by a summer heat wave was incorporated in the assessment and resulted in density independence of host-parasitoid relationships. As such, high larval parasitoid mortality during insect herbivore outbreaks may contribute to the absence of host density-dependent parasitization during outbreak events.