Decoding toxic metal stress: Microbial community shifts as sentinels of contamination gradients

Potentially toxic metal (PTM) contamination in coal-mining subsidence wetlands poses a significant ecological threat, yet microbial responses at the sediment-water interface (SWI) remain poorly understood. Here, we perform a mechanistic assessment of how PTM gradients shape bacterial and fungal communities and evaluate their utility as ecological indicators. We survey a series of subsidence ponds in Huaibei, China, spanning gradients of Cd, As, Cu, Zn, Pb, Cr, Co, and Ni, and combined 16S rRNA and ITS gene sequencing with geochemical analyses. The results reveal multifaceted but complementary responses: bacterial communities were highly sensitive to PTMs, exhibiting significant declines in α-diversity (e.g., Shannon index, p < 0.01) and pronounced compositional turnover, whereas fungal communities displayed relative resilience, maintaining overall diversity through taxonomic substitution. LEfSe identifies robust biomarkers of high contamination-most notably enrichment of Proteobacteria and depletion of functionally important Nitrospirota. Null-model analyses show that the PTM gradient acts as a strong environmental filter, driving a shift from stochastic to deterministic assembly processes; this pattern is more pronounced in bacterial assemblages, with homogeneous selection increasing from 7.1 % to 46.4 % in sediments, and is corroborated by a reorganization of co-occurrence networks from synergistic to highly competitive, "winner-take-all" architectures dominated by PTM-associated hubs. Collectively, we establish a multi-metric indicator framework integrating diversity metrics, key taxonomic biomarkers and community-assembly mechanisms that offers a mechanistic route for ecological risk assessment and monitoring of PTM-contaminated aquatic ecosystems and to guide remediation efforts.