Invasive weeds cause substantial ecological, economical, and social problems, and are currently being controlled by herbicide applications. However, how herbicides affect other ecological interactions of invasive weeds, including their symbiosis with arbuscular mycorrhizal fungi (AMF), remains poorly understood. In this study, we therefore conducted field investigation to understand how the herbicide glyphosate affects the AMF diversity in the rhizosphere of the invasive weed Solidago canadensis. We also performed a greenhouse experiment to study if AMF can contribute to herbicide resistance. The results showed that the AMF colonization rate was significantly higher in S. canadensis when exposed to glyphosate in the field or in greenhouse settings. AMF diversity was also found to be higher in the rhizosphere soil after glyphosate application in the field. AMF colonization in greenhouse experiments also positively correlated with plant growth and reduced amounts of damaged leaves and the plant's content of the stress markers flavonol and anthocyanin. Chlorophyll content was significantly enhanced by AMF colonization, regardless of glyphosate application. These results indicate that herbicide can promote AMF colonization and diversity, and that AMF can enhance the herbicide resistance of S. canadensis. These findings suggest that herbicide application may promote the spread of S. canadensis through enhanced microbial interactions, posing new eco-environmental risks.

Plant-associated microbiomes are structured by environmental conditions and plant associates, both of which are being altered by climate change. The future structure of plant microbiomes will depend on the, largely unknown, relative importance of each. This uncertainty is particularly relevant for arctic peatlands, which are undergoing large shifts in plant communities and soil microbiomes as permafrost thaws, and are potentially appreciable sources of climate change feedbacks due to their soil carbon (C) storage. We characterized phyllosphere and rhizosphere microbiomes of six plant species, and bulk peat, across a permafrost thaw progression (from intact permafrost, to partially- and fully-thawed stages) via 16S rRNA gene amplicon sequencing. We tested the hypothesis that the relative influence of biotic versus environmental filtering (the role of plant species versus thaw-defined habitat) in structuring microbial communities would differ among phyllosphere, rhizosphere, and bulk peat. Using both abundance- and phylogenetic-based approaches, we found that phyllosphere microbial composition was more strongly explained by plant associate, with little influence of habitat, whereas in the rhizosphere, plant and habitat had similar influence. Network-based community analyses showed that keystone taxa exhibited similar patterns with stronger responses to drivers. However, plant associates appeared to have a larger influence on organisms belonging to families associated with methane-cycling than the bulk community. Putative methanogens were more strongly influenced by plant than habitat in the rhizosphere, and in the phyllosphere putative methanotrophs were more strongly influenced by plant than was the community at large. We conclude that biotic effects can be stronger than environmental filtering, but their relative importance varies among microbial groups. For most microbes in this system, biotic filtering was stronger aboveground than belowground. However, for putative methane-cyclers, plant associations have a stronger influence on community composition than environment despite major hydrological changes with thaw. This suggests that plant successional dynamics may be as important as hydrological changes in determining microbial relevance to C-cycling climate feedbacks. By partitioning the degree that plant versus environmental filtering drives microbiome composition and function we can improve our ability to predict the consequences of warming for C-cycling in other arctic areas undergoing similar permafrost thaw transitions.