Engineering Plant-Microbiome Interaction Networks for Predictive Soil Bioremediation Under the Stress-Stability Paradox.

Soil pollution poses a profound threat to ecosystem and human health. This review proposes a novel framework centered on engineering biological interaction networks for efficient and sustainable soil decontamination, moving beyond the traditional single-species paradigm. We dissect the architecture and dynamics of key interactions, including plant-plant, plant-microbe, and microbe-microbe interactions, within remediation contexts, elucidating how mechanisms like mutualism, competition, and cross-kingdom signaling govern the fate of heavy metals, organic pollutants, and complex mixtures. Crucially, we explore how these natural networks can be actively engineered through strategies such as targeted bioaugmentation, precision biostimulation, and rational plant community assembly to enhance remediation outcomes. Furthermore, we highlight how cutting-edge multi-omics, synthetic ecology, and computational modeling are transitioning the field from descriptive ecology to predictive network design, enabling the decoding of the soil black box and the rational construction of tailored, resilient remediation consortia. Finally, we discuss the ecological challenges of introducing designed networks and outline a future road map toward precision restoration ecology, where theory-guided interaction network management enables effective, stable, and ecologically sound soil clean-up. This network-centric paradigm represents a fundamental shift from experience-based trial-and-error to a principled design approach for restoring soil health.