Abstract
The role of mycorrhizal associations in controlling forest soil carbon storage remains under debate. This uncertainty is potentially due to an incomplete understanding of their influence on the free-living soil microbiome and its functions. In this study, rhizosphere and non-rhizosphere soils were collected from eight arbuscular mycorrhizal (AM) and seven ectomycorrhizal (ECM) tree species in a temperate forest. We employed high-throughput sequencing and (18)O-H(2)O labeling to analyze the soil microbial community and carbon use efficiency (CUE), respectively. We find microbial respiration rates are higher in rhizosphere than that in non-rhizosphere soils for ECM trees, whereas microbial growth rates show no significant differences. Consequently, microbial CUE is lower in rhizosphere compared to non-rhizosphere soils for ECM trees. In addition, we find that non-rhizosphere soils from ECM trees exhibited higher CUE compared to those from AM trees. Furthermore, we observe that bacterial-fungal co-occurrence networks in ECM soils exhibit greater complexity relative to AM ones. Using random forest and structural equation modeling analyses, we find that microbial stoichiometric carbon/nitrogen imbalance and network complexity are key predictors of soil microbial CUE for AM and ECM trees, respectively. Our findings shed new light on the pivotal role of mycorrhizal associations in shaping free-living microbial communities and their metabolic characteristics in the studied soils. These insights are critical for predicting soil carbon sequestration in response to shifts in ECM and AM species within temperate forest under climate change.