The global deterioration of rivers and lakes threatens both ecological stability and human health. Current water quality assessments face challenges in indicator selection and weight determination. This study developed a novel water quality index (BE-WQI) framework that integrated environmental indices (physiochemical factors and emerging contaminants) and multi-taxonomic biological communities (bacteria, archaea, eukaryote, alga, benthos, and fish) based on large-scale environmental DNA (eDNA) sequencing. Leveraging a rigorous three-step selection of core abiotic indicators, we employed an advanced gradient-boosting machine learning algorithm (LightGBM) to quantify the feature importance of core indicators in explaining water quality variability, and to generate LightGBM-based weights (WLGBM). Following absence filtration, range test, and redundancy test, we examined correlation ranks between selected abiotic indicators and core biological indices, which were derived from alpha diversity, taxonomic abundance, and network structure of biological communities, thereby yielding biological response-based weights (WBio). WLGBM and WBio, calculated via the Rank Order Centroid (ROC) method, were integrated through game theory and combined with transformed sub-indices and aggregation functions to construct BE-WQI models. This framework was applied to assess water quality in the eastern route of the South-to-North Water Diversion Project (ER-SNWDP), the world’s largest water transfer project, utilizing our recent monitored eDNA data, 86 physiochemical factors, and 65 emerging contaminants. The BE-WQI classified 46.7% of samples as “slight pollution” and 53.3% as “moderate pollution”, highlighting increased nitrogen pollution downstream and impounded lake impacts. This optimized BE-WQI framework minimized uncertainty and bias in assessment results, providing robust technical support for evaluating water quality.

River ecosystems, crucial components of the global nitrogen cycle, are increasingly affected by antibiotic pollution. However, the mechanistic interplay between nitrogen cycling and antibiotic resistance genes (ARGs) dissemination remains poorly understood, limiting effective ecological risk assessments. Here, we identify nitrate-reducing bacteria (NRBs), key drivers of denitrification and greenhouse gas mitigation, as dual-functional hubs that co-regulate nitrogen turnover and ARG dissemination under antibiotic stress. By integrating 173 metagenomes and 10 metatranscriptomes from the Yangtze River, we reconstruct 4200 metagenome-assembled genomes (MAGs) and find that NRBs harbor ~69% of actively transcribed ARGs in river microbiomes, with antibiotic pressure as the dominant ecological driver. Simulated microcosms exposed to antibiotic gradients reveal a hormetic response, where environmentally relevant concentrations enhanced both NRB-driven denitrification efficiency and ARG dissemination. Multi-omics analyses further reveal antibiotic-driven horizontal gene transfer as the predominant selective force co-shaping ARG and nitrate reduction gene dynamics, accelerating both nitrogen cycling and ARG spread. These findings establish NRBs as central hubs bridging antibiotic resistance and nitrogen metabolism, providing a mechanistic framework for predicting co-selection dynamics and mitigating cascading ecological impacts. Our work highlights the need to integrate microbial co-metabolic functions into pollution control strategies and redefine ecological risk assessments in antibiotic-polluted ecosystems.

Antibiotics and engineered nanomaterials are widely recognized as emerging contaminants in the aquatic environment. While the individual toxic effects of sulfamethoxazole (SMX) and graphene oxide (GO) on denitrifying microorganisms have been well-documented, their combined effects under environmentally relevant concentrations remain poorly understood. This study investigated the joint influence of SMX and GO on aerobic denitrification by the newly isolated strain HA-2. The results showed that while single exposure to either contaminant induced inhibitions of nitrate removal, coexposure to environmental concentrations of SMX and GO unexpectedly enhanced aerobic nitrate reduction efficiency, with this stimulatory effect remaining stable over long-term cultivation. Mechanistically, the transcriptomic repression of apoptotic and necroptotic signaling, coupled with physiologically elevated electron transport activity, switched cells from a passive suppressed state to an active adaptive mode that prioritized energy replenishment and stress resistance over programmed cell death. Active defense involving enhanced extracellular polymeric substance (EPS) production and SMX efflux and a metabolic promotion cascade involving upregulated electron transport activity and energy synthesis drove coordinated adaptation and improved denitrification performance. SMX and GO exhibited synergistic effects on the induction of microbial hormesis, particularly by collectively boosting the energy synthesis. Overall, the findings refine our understanding of microbial responses to combined emerging anthropogenic pollutant stress, offering mechanistic insights into how aerobic denitrifying bacteria cope with co-occurring antibiotic and nanomaterial pollutants.

The subsurface soil environment, spanning from the vadose zone to the saturated zone, serves as a critical zone where diverse soil processes and biogeochemical interactions take place. However, the specific influence of longterm fertilization on microbial interactions in this zone remains unclear. In this study, we conducted a five-year field experiment, collecting 76 soil samples from three 0-20 m boreholes to investigate the effects of fertilization (N-addition and NPK-addition) on bacterial, fungal and archaeal communities along the soil profile. Long-term fertilization significantly altered the composition, assembly processes, and inter-kingdom co-occurrence patterns of bacterial, archaeal and fungal communities. These changes were particularly pronounced in the saturated soil layer (5-20 m), where the network complexity increased (average degree rising by 5.19 and 8.05 times under Naddition and NPK-addition, respectively) and stability reduced (modularity decreased by 0.47 times and 0.43 times, respectively). Further analysis revealed that pH and nutrients were the main drivers of microbial communities and interactions under both fertilization treatments, explaining 97.20 % of variance in N-addition and 96.70 % in NPK-addition. Moreover, compared to N-addition, the NPK-addition treatment exhibited greater environmental friendliness, with enhanced carbon and nitrogen metabolic potential. This relatively milder impact could be attributed to the enhanced "bridge" role of fungi, acting as stabilizers in the inter-kingdom network. Our findings not only provided new insights into the influence of fertilization on microbial composition and inter-kingdom co-occurrence patterns in the saturated soil layer, but also highlighted the specific role of fungi in mitigating the disruption caused by fertilization.

Reservoirs represent a critical component of greenhouse gas (GHG) emissions, yet the intricacies of how biotic and abiotic factors influence GHG dynamics within reservoirs remain largely unexplored. Herein, we investigated the spatiotemporal patterns of CO2 and N2O emissions and the underlying factors in the Danjiangkou Reservoir, Asia's largest artificial freshwater reservoir. We found that this reservoir was a significant source of GHGs to the atmosphere, with peak CO2 emissions observed in autumn (1544.39 ± 652.53 μatm) and N2O emissions in winter (32.57 ± 8.87 μmol/L). Moreover, we identified crucial bacterial biomarkers that regulate GHG dynamics, and these GHG biomarkers exhibited consistent seasonal patterns with the corresponding GHGs, concerning their abundance and niche breadth. Notably, GHG biomarkers displayed larger effects on the variations of N2O than CO2 emissions, while physiochemical variables were more critical for CO2 dynamics, highlighting the need to consider both biotic and abiotic factors when evaluating GHG emissions from reservoirs. Overall, this study advanced our knowledge of GHG emissions and their driving mechanisms in artificial reservoirs, emphasizing the importance of functional microbes in estimating and managing CO2 and N2O emissions from artificial reservoirs worldwide.

Dissimilatory iron-reducing bacteria (DIRB) have been widely applied to organic matter metabolism in sediments due to their specific biological properties, which are significantly affected by environmental factors. Until now, there has been no systematic study on applying these effects of environmental factors to DIRB enrichment. In this study, we identified the critical environmental factors through RDA, PLS-PM, and correlation analysis and then adjusted the proportion of these factors (C/Fe, C/N, and Fe/S) through single-factor and response surface experiments to enrich DIRB. The performance of the enriched DIRB in carbon metabolism of organic-rich sediments was analyzed by remediation experiment, and the metabolic mechanism was revealed through iron-reducing performance, community structure, and functional genes. The results showed that the carbon, nitrogen, phosphorus, and sulfur were crucial factors influencing the activity of DIRB in river sediments, and the enriched mudbacteria mixture containing a large amount of DIRB (referred to as "DMB") obtained by domestication under C/ Fe = 0.492 and C/N = 13.659 showed the highest iron reduction rate of 55.51% +/- 2.53%. The DMB exhibited better-sustained removal of total organic matter (TOM), which was higher than the control (MB) by 8.81%. Moreover, we discovered that the enriched DIRB: (1) had a better reduction effect on different Fe(III) forms, especially in carbon-limited and iron-limited conditions. (2) increased in abundance and established a beneficial symbiotic relationship with hydrolytic acidifying bacteria. (3) showed an increased abundance of functional pathways associated with organismal systems and signal processing (such as ABC transporters and Translation). These findings revealed the reasons for the improved efficiency of organic matter metabolism.

Objective To observe the expression levels of base excision repair(BER)pathway-related proteins in small cell lung cancer(SCLC)tissues,and analyze their relationship with the prognosis and tumor immune microenvironment.Methods A retrospective cohort study was conducted on 74 patients with limited-stage SCLC undergoing surgical treatment in our medical center from December 2018 to June 2023.Immunohistochemical staining was performed to analyze the protein expression of BER pathway components,apurinic/apyrimidinic endonuclease 1(APE1),8-oxoguanine DNA glycosylase 1(OGG1),DNA polymerase β(POLβ),X-ray repair cross-complementing protein 1(XRCC1),ATP-dependent DNA ligase I(LIGⅠ),and immune cell infiltration markers of CD3⁺ T cells,CD8⁺ T cells,CD68⁺ macrophages in SCLC tissues.Chi-square test was applied to analyze the relationship of BER protein expression and clinicopathological features;Kaplan-Meier survival curve was plotted to evaluate the impacts of BER protein expression and immune cells on disease-free survival(DFS)and overall survival(OS),multivariate Cox regression analysis was utilized to identify DFS prognostic factors,and Spearman correlation analysis was performed to analyze the correlation of BER-immune cell infiltration.In in vitro experiments,transient transfection was applied in H196 cells to overexpress APE1/POLβ/LIGⅠ,respectively.Thus,the cells were divided into negative control(NC,empty vector)and overexpression(OE,target plasmids)groups.CCK-8 and TUNEL assays were employed to determine the effects of OEAPE1,OEPOLβ and OELIGⅠon cell sensitivity to cisplatin.In in vivo experiments,nude mice bearing xenograft tumors were grouped into WT,E3330(APE1 inhibitor),cisplatin,and cisplatin+E3330 groups to determine the effects of the combination therapy on tumor growth.Results There were no significant correlations of the expression levels of key BER pathway proteins with clinicopathological characteristics,including gender,age,smoking history,tumor location,Ki67 index,or TNM stage(all P>0.05).The patients with low expression of APE1,POLβ,and LIGⅠ had obviously higher DFS rates than those with high expression(P<0.05),and the patients with larger proportion of CD3+T cells also had higher DFS rates than those with smaller proportion(P=0.043).Multivariate Cox regression analysis indicated that tumor TNM stage(HR=2.465)and APE1 expression(HR=2.730)were independent risk factors for the prognosis of SCLC patients(P<0.05).Spearman correlation analysis demonstrated a positive correlation between APE1 and CD8+T cell proportion in the SCLC patients(r=0.27,P<0.05).In vitro experiments showed that the overexpression(OE)cells(OEAPE1 and OELIGⅠ)exhibited reduced sensitivity to cisplatin than the NC group(P<0.05).Animal experiments indicated that cisplatin+E3330 significant inhibited xenograft tumor growth,indicating enhanced therapeutic efficacy(P<0.01).Conclusion High expression of APE1,POLβ,and LIGⅠ in the BER pathway indicates poor prognosis and low DFS rate in SCLC patients.High expression of APE1 is positively correlated with CD8+T cells,and can be used as an auxiliary marker for SCLC immunotherapy.

Groundwater ecosystems face increasing threat from declining water quality due to intensified urbanization, agricultural, and industrial activities. Accurately identifying anthropogenic disturbances remains challenging, and their effects on microbial nitrogen cycling are still largely unknown. Here, by collecting 64 groundwater samples from an aquifer beneath the Tanghe sewage reservoir in the North China Plain, we conducted a full-spectrum screening of 228 physiochemical indices, 47 nitrogen cycling genes (NCGs) and 2182 metagenome-assembled genomes (MAGs) harboring NCGs. Unmix model identified antibiotic usage, industrial manufacturing, and agricultural practices as the predominant pollution sources, explaining 49.6-92.2 % (averaged 81.0 %) of the variations in aquifer attributes. These activities were primary drivers governing distributions of groundwater NCGs and NCG-hosts, with fragmented denitrification processes being prevalent. Antibiotic usage and industrial activities were probably associated with suppressed nitrogen cycling, while agriculture had a positive effect. Notably, we observed enhanced mutualistic interactions within NCG-hosts and increased enrichment of NCG-antibiotic resistance gene (ARG), NCG-mental resistance gene (MRG), and NCG-ARG-MRG co-hosts under high anthropogenic stresses, suggesting microbial adaptation to optimize nutrient and energy metabolism. This study provided new insight into how groundwater nitrogen cycling responds to anthropogenic disturbances, offering valuable information for developing groundwater management and pollution control strategies.

While existing early-warning systems struggle to achieve cross-species cyanobacterial risk prediction with the required synchronicity and accuracy in aquatic ecosystems, our study pioneers a genome architecture-driven monitoring paradigm through decoding 317 cyanobacterial metagenome-assembled genomes from the world's largest phosphorus-limiting water transfer system, the Middle Route of the South-to-North Water Diversion Canal (MR-SNWDC). We found an evolutionary blueprint where genome minimization (<3 Mbp) confers ecological dominance under phosphorus scarcity. These streamlined genomes showed predominance and remarkable seasonal dynamics and demonstrated metabolic specialization in phosphorus turnover, light harvesting, and carbon fixation compared to larger genomes. Importantly, we identified a 3 Mbp genomic threshold distinguishing low-risk cyanobacterial consortia from their toxin-producing counterparts. This genome-proxy system enables preemptive risk mitigation by predicting toxic transitions through genome size tracking, fundamentally advancing algal management from reactive monitoring to proactive regulation in water transfer networks.

Heavy metals (HM) pose a persistent and severe threat to global coastal wetlands, with sediments serving as the primary HM sink within mangrove forests. Although the sources and factors influencing HM accumulation in mangrove sediments are increasingly understood, their relative importance and underlying accumulation mechanisms remain unclear. Spatial patterns of HM concentrations in surface sediments and their quantitative relationships with influencing factors, as well as the detailed accumulation mechanisms were investigated in the contiguous mangroves across Hainan Island, China. At the scale of the entire Hainan Island, lead (Pb) and manganese (Mn) concentrations in mangrove sediments decreased with increasing latitude, while nickel (Ni) concentration increased. Chromium (Cr), Pb, and Mn concentrations decreased with increasing longitude, whereas Cd concentration increased. Arsenic (As), Cr, and Mn exhibited relatively severe contamination, with only As and Cd posed moderate to considerable ecological risks. Among all the collected influencing factors, the sediment physicochemical properties had the highest explanatory power for the spatial variations in sediment HM concentrations of Hainan Island's mangrove forests. Although anthropogenic activities are the primary source, their effects on the HM accumulation in sediments were indirect and secondary. Our findings underscore the critical role of sediment physicochemical properties in determining HM concentrations in mangrove sediments and emphasize the need for integrated management strategies to mitigate HM pollution in these vital coastal ecosystems. This study provides theoretical and technical support for the scientific assessment of HM pollution risks, guides pollution control and ecological restoration efforts in mangrove forests.

Large-scale farms with concentrated animal feeding operations generate significant volumes of dairy livestock wastewater (LWW) that are rich in nutrients and salinity, and its improper management can quickly deteriorate natural environments. Here, we cultivated two microalgae, Chlorella protothecoides and Chlamydomonas reinhardtii, in filtered dairy LWW to achieve nutrient recovery from high-salinity dairy wastewater. Under 1000-7000 lux using 25-100 % LWW, we observed high removal of 86.8-95.3 % dissolved total nitrogen, 75.5-92.0 % dissolved organic nitrogen, 98.3-99.8 % NH4+-N, and 57.2-100 % total phosphorus, with moderate removal of 12.1-67.5 % for dissolved organic carbon and 2.49-33.4 % for total 19 metal(loid)s. Notably, the aromaticity of LWW dramatically reduced under higher light intensity and low rainwater dilution levels, while organic nitrogen utilization was markedly enhanced when light intensity exceeded 3000 lux. We found the highest values of 5.33-6.52 g/L for biomass yields, 37.9-42.0 % for protein contents, and 22.2-27.6 % for lipid contents. Total fatty acids (TFAs) contained 89.5-100 % C16-C18, predominantly C16:0, C18:1, and C18:2. Lipid accumulation was significantly enhanced under higher light intensities, though C. reinhardtii exhibited a greater sensitivity to higher LWW proportions compared to C. protothecoides, which favored lower rainwater dilution levels for increased TFA accumulation. Our study indicated that these two microalgae, when cultivated in filtered dairy LWW rich in nutrients and salinity, are promising feedstock candidates for biodiesel production.

Phytoplankton-bacteria interactions are critical but often overlooked in assessing the impacts of pollutants on ecosystems. Herein, we used a coculture consisting of the green alga Chlamydomonas reinhardtii and river bacteria to investigate their responses to antibiotic stress. Both partners exhibited hormesis in cocultures but were inhibited in monocultures under exposure to 10 different antibiotics, especially azithromycin (AZM). Notably, mutualistic cooperation between the partners shifted the effect of AZM from inhibition in monocultures to promotion in cocultures. C. reinhardtii alleviates AZM stress on bacteria by providing organic carbon and efficiently removing antibiotics. In turn, the altered phycospheric bacteriome supplied ammonia, phosphate, vitamin B12, and indole-3-acetic acid to promote C. reinhardtii growth. The antibiotic-induced growth promotion was also observed in natural phytoplankton-bacteria communities. Our findings challenge the reliability of ecotoxicity assessment that is typically based on single-species tests, emphasizing the importance of cross-kingdom interactions in assessing pollutant effects.

Water diversion projects effectively mitigate the uneven distribution of water resources but can also influence aquatic biodiversity and ecosystem functions. Despite their importance, the impacts of such projects on multi-domain microbial community dynamics and the underlying mechanisms remain poorly understood. Utilizing high-throughput sequencing, we investigated bacterial, archaeal, and fungal community dynamics along the eastern route of the South-to-North water diversion project during both non-water diversion period (NWDP) and water diversion period (WDP). Our findings revealed competitive exclusion effects among bacterial and archaeal communities during the WDP, characterized by decreased species richness and increased biomass, while fungal biomass significantly declined. Distance-decay relationships suggested microbial homogenization during the WDP. Robustness analyses revealed reduced community stability during the WDP, with water diversion primarily influencing bacterial stability, while environmental factors had a greater impact on archaeal and fungal communities. Stochastic processes, primarily homogenizing dispersal and drift, intensified for bacterial and fungal communities during the WDP. Notably, only bacterial functional diversity decreased during the WDP, with increased relative abundance of chemoheterotrophic and organic compound catabolic bacteria and declined photoautotrophic bacteria. PLS-PM indicated that water diversion primarily shaped bacterial assembly processes and functional guilds, whereas environmental factors had a greater influence on archaeal communities. This study enhances our understanding of microbial dynamics during the WDP and underscores the importance of assessing both direct impacts and resulting environmental fluctuations.

The Yangtze River Estuary, a critical ecological and economic zone in China, exhibits complex interactions between biogenic substances and heavy metals that define its eutrophication status and ecosystem health. This study investigated their spatiotemporal dynamics through comprehensive sampling at 12 sites (3 in the Xuliujing (XLJ) state-controlled section; 9 at estuarine outlets). Biogenic substances (COD > TOC > DOC; TN > NO2--N > NH3-N; TP > SRP) peaked in late spring/summer within the XLJ section, due to seasonal runoff inputs. Spatially, concentrations in Chongming Island's Northern Branch exceeded those in the Southern Branch, attributed to preferential "south-to-north" drainage and limited hydrological connectivity. Heavy metals (As, Cd, Co, Cr, Cu, Ni, Pb, Sb, Zn) remained below China's Class I surface water standards (GB 3838-2002). Temporal maxima occurred in spring (XLJ section), while estuarine distributions followed a characteristic "decrease-increase-decrease" spatial pattern, with peaks near Chongming branch origins and the East China Sea confluence. Principal component analysis (PCA) showed anthropogenic sources (transportation, industrial emissions, and agricultural activities) and natural sources as the main potential contributors to heavy metal pollution. Notably, Chlorophyll-alpha demonstrated the strongest correlation with heavy metals among conventional parameters, whereas COD, TOC, DOC, and TP showed significant associations with multiple metals. A Random Forest model effectively predicted Co (R = 0.95) and Cu concentrations using physicochemical and biogenic inputs but underperformed for Sb and Zn. Conductivity, Chlorophyll-alpha, and dissolved oxygen were identified as key predictors for heavy metal concentrations. This integrated assessment provides a scientific foundation for ecological management strategies.

Although the terrestrial subsurface harbors a substantial fraction of Earth's microbial biomass, the genomic diversity of groundwater microbiomes and their potential for bioprospecting remain poorly characterized. Here, we recovered 44,320 bacterial and archaeal genomes from in-house and publicly available metagenomic datasets, establishing a large-scale groundwater microbiota catalog (GWMC) spanning 167 phyla, including four candidate phyla and over 12,000 previously uncharacterized species. This unprecedented phylogenetic diversity was accompanied by a bimodal genome size distribution (0.3-12.8 Mbp), revealing divergent strategies of genomic allocation. By mining extensive genomic resources, we found that small genomes prioritized molecular defense and redox regulation, whereas large genomes frequently harbored greater biosynthetic potential. Notably, we establish the largest selenoprotein catalog to date and highlight groundwater as an overlooked hotspot of microbial selenium metabolism. Overall, this work advances our understanding of microbial diversity in aquifers and uncovers underexplored genomic resources with potential for biotechnology and biomedicine.

The relationship between polycyclic aromatic hydrocarbons (PAHs) exposure and female infertility (FI) remains unclear, particularly regarding mixed PAH exposures. This study aimed to investigate the association between individual and mixed PAH exposures and FI by integrating epidemiological and network toxicological approaches. A case-control study was conducted involving 83 infertile patients and 272 non-infertile controls in Guangdong Province, China. Three statistical models were applied to assess the effects of eight urinary hydroxylated-PAHs (OH-PAHs) metabolites on FI. Network toxicology analysis was utilized to identify common genes, potential pathways, and key targets. After adjusting for potential confounders, 3-OHFLU and 2-OHPHE were significantly associated with an increased risk of FI, with 3-OHFLU identified as the predominant risk factor (OR [95 % CI]: 1.92 [1.42, 2.69]). Mixture analysis revealed a positive association between mixed OH-PAHs and FI risk, with 3-OHFLU contributing the most. Based on network toxicology analysis, we propose a potential mechanism: PAHs exposure may activate the TNF signaling pathway, which could trigger an inflammatory response, potentially altering AKT1 expression via the PI3K-Akt pathway. This might lead to reduced cell survival, promoted oocyte apoptosis, and ultimately contribute to FI. These findings indicate that mixed PAHs exposure increases the risk of FI, with 3-OHFLU identified as the predominant contributing factor and the TNF signaling pathway serving as a potential underlying mechanism. Further longitudinal studies are needed to validate these results.

Soil is the basis of bamboo growth and quality formation of bamboo shoots and has an important contribution to the sustainable development of agriculture. To this end, We studied the soil properties and microbial communities of Dendrocalamus brandisii by collecting twenty-one soil samples from its seven typical geographic provenances in Yunnan Province, China. Bacterial 16S rRNA gene amplicons were used to detect soil bacteria and predict bacterial functions using Tax4Fun. The results indicated that the soil bacterial diversity indices (ACE, Chao1, Simpson, and Shannon) were significantly different among different geographical provenances. The dominant bacterial groups at the phylum level in all seven regions were Proteobacteria (19.78~29.06%), Actinobacteria (13.53~30.01%), Chloroflexi (8.03~31.47%), and Acidobacteria (7.12~19.17%), with markedly different constitution proportions. Total phosphorus, available potassium, and pH were the main environmental factors affecting soil bacterial communities. There were significant differences in the secondary metabolic pathways and phenotypes of soil bacterial functions, exhibiting a diversity of functions. The geographical variables of the soil bacterial community in D. brandisii varied with spatial scales. Environmental factors such as available potassium (AK), pH, and total nitrogen (TN) have an impact on soil bacterial communities.

The accumulation of antibiotics in the natural environment can disrupt microbial population dynamics. However, our understanding of how microbial communities adapt to the antibiotic stress in groundwater ecosystems remains limited. By recovering 2675 metagenome-assembled genomes (MAGs) from 66 groundwater samples, we explored the effect of antibiotics on bacterial, archaeal, and fungal communities, and revealed the pivotal microbes and their mechanisms in coping with antibiotic stress. The results indicated that antibiotics had the most significant influence on bacterial and archaeal communities, while the impact on the fungal community was minimal. Analysis of co-occurrence networks between antibiotics and microbes revealed the critical roles of Candidate Phyla Radiation (CPR) bacteria and DPANN archaea, two representative microbial groups in groundwater ecosystem, in coping with antibiotic resistance and enhancing network connectivity and complexity. Further genomic analysis demonstrated that CPR bacteria carried approximately 6 % of the identified antibiotic resistance genes (ARGs), indicating their potential to withstand antibiotics on their own. Meanwhile, the genomes of CPR bacteria and DPANN archaea were found to encode diverse biosynthetic gene clusters (BGCs) responsible for producing antimicrobial metabolites, which could not only assist CPR and DPANN organisms but also benefit the surrounding microbes in combating antibiotic stress. These findings underscore the significant impact of antibiotics on prokaryotic microbial communities in groundwater, and highlight the importance of CPR bacteria and DPANN archaea in enhancing the overall resilience and functionality of the microbial community in the face of antibiotic stress.