In a controlled experimental investigation of aboveground plant-microbe-insect ecology, synthetic bacterial communities on leaf surfaces significantly influenced the performance of chewing insect herbivores while simultaneously triggering heightened plant defence responses. Using an axenically cultivated model system pairing the model plant Arabidopsis thaliana with the Large White butterfly Pieris brassicae, researchers observed that exposure to a diverse synthetic community composed of twenty leaf-associated bacterial strains reduced larval weight by 17.6 percent compared with mock-treated control plants. Larval survival was unaffected across treatments, remaining between 96 and 99 percent throughout the trials. Simultaneously, plants supporting this maximal community richness accumulated significantly higher levels of the primary anti-herbivore defence hormone jasmonic acid following twenty-four hours of feeding than colonized non-fed or axenic control plants. Conversely, insect herbivory exerted a profound reciprocal effect on the leaf microbiome itself, driving a fourfold increase in total bacterial numbers on leaves and selecting for the extreme dominance of a single opportunistic phyllosphere bacterium, while the larval digestive tract selectively enriched a narrow subset of leaf-associated microbes.
The aerial surfaces of terrestrial vegetation, collectively designated as the phyllosphere, represent one of the most expansive microbial habitats on the planet, with global leaf surface area estimated between 640 million and one billion square kilometres. Foliage hosts dense bacterial populations that frequently reach densities of ten million bacterial cells per gram of fresh vegetative tissue. Despite the ubiquity of these phyllosphere bacterial communities and the widespread presence of insect herbivores that consume foliage across agricultural and natural ecosystems, the tripartite dynamics linking resident nonpathogenic leaf bacteria, host plant immune signalling, and chewing herbivores have remained poorly understood. Most prior investigations into plant-microbe-insect interactions focused almost exclusively on single pathogenic organisms or specialized insect endosymbionts rather than the multispecies nonpathogenic commensal consortia that typically colonize foliage. This experimental gap prompted the research team to investigate whether community richness in the phyllosphere alters herbivore performance and whether the damage caused by insect feeding restructures the resident bacterial community in return.
To rigorously decouple microbial interactions from uncontrolled environmental contamination, the authors developed a functionally gnotobiotic plant-insect experimental framework. Surface-sterilized seeds of Arabidopsis thaliana accession Col-0 were germinated and grown axenically in Magenta plant culture boxes containing zeolite substrate moistened with Murashige and Skoog liquid nutrient medium. Lids were fitted with micropore filtration strips to allow gas exchange while preventing airborne particulate or microbial intrusion. Pieris brassicae, a specialist brassicaceous herbivore whose larvae lack an obligate resident gut microbiome and maintain a virtually sterile gut under clean laboratory rearing conditions, was chosen as the insect model. Butterfly eggs were collected from surface-sterilized Brussels sprout foliage, mechanically detached, and hatched in sterile containers. Neonate larvae were aseptically transferred to the four-week-old axenic plants eight days after the foliage had been spray-inoculated, ensuring that baseline background microbial contamination remained negligible throughout the feeding experiments.
To assess the specific contribution of taxonomic and functional richness, the experimental protocol evaluated three distinct tiers of community complexity designated as SynCom5, SynCom10, and SynCom20, containing five, ten, and twenty distinct strains, respectively. Strains were cultured individually in Reasoner 2A broth, harvested through centrifugation, resuspended in phosphate-buffered saline at uniform optical densities, and combined in equal cellular ratios before foliar application at a controlled pressure of ten pounds per square inch. While SynCom5 and SynCom10 configurations were reconstituted from randomly sampled strains across independent experimental replicates to isolate richness effects from single-strain idiosyncrasies, SynCom20 represented a constant assembly containing the maximal functional diversity and phylogenetic breadth present within the complete reference pool.
Seven days of continuous caterpillar feeding demonstrated that while microbial presence was generally antagonistic to herbivore performance, the statistical severity of the impact depended strictly on community complexity. Larvae feeding on plants carrying the twenty-member consortium attained significantly lower body mass than larvae consuming axenic control plants, with an adjusted significance level of p equals 0.0153. Larval weight also displayed moderate negative correlations with both bacterial taxonomic richness and Faith phylogenetic diversity. While average larval biomass was reduced by 13.1 percent in SynCom5 and 14.4 percent in SynCom10 treatments relative to axenic controls, these intermediate community richness levels did not achieve statistically significant separation from the control cohorts, nor were differences among the three inoculated consortia statistically distinguishable. The authors confirmed that larval mortality was indistinguishable among treatments, indicating that the reduced biomass was not an artifact of acute bacterial pathogenicity or lethal systemic infection.
To investigate whether the reduction in larval biomass was mediated by host physiological defences, the researchers measured foliar phytohormone concentrations using liquid chromatography coupled with tandem mass spectrometry before herbivory and twenty-four hours after the introduction of feeding caterpillars. The quantification targeted jasmonic acid and its bioactive isoleucine conjugate jasmonoyl-isoleucine, which regulate defences against foliar chewing insects, alongside abscisic acid and salicylic acid, the latter typically mediating responses against biotrophic pathogens. While jasmonic acid levels rose across all plants subjected to caterpillar feeding, plants colonised by SynCom20 exhibited significantly greater jasmonic acid accumulation than mock-treated controls experiencing identical feeding pressure. In the absence of caterpillar herbivory, SynCom20 inoculation alone provoked no significant induction of jasmonic acid or salicylic acid at either baseline or the twenty-four-hour sampling time point. Furthermore, salicylic acid concentrations remained unchanged across treatments, indicating that the bacterial consortium modified inducible anti-herbivore defence responses without initiating an antagonistic pathogen-associated immune pathway.
In parallel, the experimental measurements demonstrated that the ecological pressure exerted by insect herbivory profoundly reorganised the epiphytic bacterial populations inhabiting the leaves. Quantification of viable bacterial colonies recovered from leaf washes showed that total bacterial abundance increased approximately fourfold on plants subjected to seven days of larval feeding compared with undamaged inoculated controls, shifting from an average of 100 million colony-forming units per gram of fresh leaf tissue to 420 million units per gram.
High-throughput amplicon sequencing targeting the V3 to V4 hypervariable regions of the bacterial 16S ribosomal RNA gene revealed that this herbivory-induced population boom was highly selective rather than uniform across taxa. Communities on caterpillar-damaged leaves exhibited a pronounced reduction in Shannon evenness and alpha diversity compared with non-fed controls, primarily driven by the massive proliferation of Pantoea eucalypti strain 299R. Minor increases in relative abundance were observed for Microbacterium and Bacillus isolates, whereas undamaged control leaves maintained a broader, more balanced distribution of phyllosphere commensals, including notably higher relative abundances of Arthrobacter strain Leaf145.
Amplicon sequencing of homogenised caterpillar tissue demonstrated that the insect digestive tract exercised rigorous ecological filtering on the ingested leaf microbiota. Despite continuously ingesting foliage heavily dominated by Pantoea eucalypti 299R, the larval gut communities diverged sharply from the composition of the food source. The insect environment strongly enriched strains belonging to the genera Methylobacterium, specifically Methylobacterium radiotolerans 0-1 and Methylobacterium strain AC433, alongside Microbacterium strain AC026, Curtobacterium strain AC273, and Williamsia strain Leaf354. In contrast, Sphingomonas strains, which were present at low abundance in both fed and non-fed leaves, were strongly underrepresented in larval samples.
The authors discussed two complementary hypotheses to explain the elevated jasmonic acid defence induction and reduced insect biomass observed on plants colonised by the full community. First, prior bacterial colonization may cause physiological priming, whereby resident nonpathogenic microbes sensitise plant cellular surveillance systems to respond more rapidly and intensely when subsequently challenged by herbivore feeding. Second, the response may be density-dependent: the fourfold increase in bacterial biomass triggered by feeding, combined with the physical entry of bacterial cells and microbe-associated molecular patterns into the apoplast via wound sites, may breach detection thresholds required for plasma-membrane pattern recognition receptors to amplify downstream hormonal cascades. Although the co-occurrence of reduced larval growth and stronger jasmonic acid accumulation provides a plausible link to altered plant defences, direct sublethal effects of bacteria on larval physiology, digestion, or food conversion efficiency cannot be ruled out.
The study highlights important methodological considerations and ecological boundaries. Because phytohormone quantification was executed exclusively on the maximal SynCom20 assembly and axenic controls, the current data cannot definitively separate whether the augmented jasmonic acid response was caused by bacterial species richness per se, complex multispecies emergent metabolic synergies, or the unique physiological contribution of individual key strains such as Pantoea eucalypti. Future research tracking individual strains across leaves, larvae, and frass, alongside testing broader richness gradients and alternative community compositions, will help disentangle the mechanisms governing these tripartite interactions. Nonetheless, these findings confirm that commensal leaf microbiota serve as active participants in plant anti-herbivore defence networks, providing a robust, reproducible gnotobiotic foundation for dissecting how aerial plant microbiomes modulate ecosystem functioning.
Subject of Research: Reciprocal ecological interactions between phyllosphere synthetic bacterial communities, plant defences, and herbivorous insects
Article Title: From the leaf to the gut and back again: using synthetic bacterial communities to trace the fate and influence of phyllosphere bacteria in an Arabidopsis–Pieris brassicae system
Article References: Müller, M., Huve, M. A. P., Kunzler, M., Remus-Emsermann, M. N. P., Schlechter, R. O., & Voirol, L. R. P. (2026). From the leaf to the gut and back again: using synthetic bacterial communities to trace the fate and influence of phyllosphere bacteria in an Arabidopsis–Pieris brassicae system. Environmental Microbiome, 21(1), Article 96. https://doi.org/10.1186/s40793-026-00947-y
Image Credits: AI Generated
DOI: 10.1186/s40793-026-00947-y
Keywords: phyllosphere microbiome, synthetic communities, Arabidopsis thaliana, Pieris brassicae, plant herbivory, jasmonic acid, gnotobiotic ecology, microbial filtering, leaf, back, again, synthetic
News Source: Morgan Morrow. (November 10, 2026). Diverse Leaf Bacteria Curb Caterpillar Growth While Feeding Reshapes Microbial Communities. Scienmag.



