Impact of coastal deoxygenation on antibiotic resistance gene profiles in size-fractionated bacterial communities

Oxygen loss disrupts marine ecosystems, threatening biodiversity and causing mass mortality of marine life. Antibiotic resistance genes (ARGs) pose a significant threat to human health by promoting the spread of resistant pathogens, making infections harder to treat and increasing mortality risks. However, the interplay between deoxygenation and ARG dynamics remains poorly understood. In this study, we employed time-series metagenomics to investigate the responses of ARG profiles in free-living (FL) and particle-associated (PA) fraction to oxygen loss during a 22-day summer deoxygenation event in the East China Sea. In total, we identified 1,186 ARG subtypes and 2,279 mobile genetic element (MGE) subtypes. The most dominant resistance classes of antibiotics were multidrug (23.5%), followed by tetracycline (15%), macrolide-lincosamide-streptogramin (13.4%), peptide (10.3%), glycopeptide (8.7%), aminoglycoside (7.3%), and beta-lactam (4.9%). We found that ARG richness in FL fraction increased with declining oxygen levels, particularly for beta-lactam and multidrug class, while no significant relationship was observed in the PA fraction. Although the total relative abundance of ARGs in both fraction showed no significant oxygen dependence, beta-lactam and multidrug resistance genes in FL fraction significantly increased with oxygen loss. Co-occurrence network analysis revealed stronger positive associations between ARGs and MGEs in the FL fraction, suggesting enhanced gene transfer among environmental bacteria. Furthermore, neutral community model analysis indicated that stochastic processes also played an interactive role in shaping ARG composition dynamics in both bacterial fractions. Our findings provide evidence that coastal deoxygenation preferentially enriches high-risk ARGs (e.g., beta-lactamase genes) in FL bacteria through MGE-mediated transfer, highlighting escalating antibiotic resistance risks that threaten both ecosystem and human health under climate warming. This study offers a framework for size-fractionated ARG monitoring and targeted mitigation strategies in coastal ecosystems.