城市环境研究所在塑料际抗生素抗性健康风险及驱动机制方面取得新进展----中国科学院城市环境研究所.pdf

Water Research 245 (2023) 120574 Contents lists available at ScienceDirect Water Research journal homepage: www.elsevier.com/locate/watres Quantifying health risks of plastisphere antibiotic resistome and deciphering driving mechanisms in an urbanizing watershed Longji Zhu a, †, Ruilong Li a, b, †, Kai Yang a, Fei Xu a, Chenshuo Lin a, Qinglin Chen c, Dong Zhu c, Qian Sun a, Yong-Guan Zhu a, d, Li Cui a, * a Key Lab of Urban Environment and Health, Fujian Key Laboratory of Watershed Ecology, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China b School of Marine Science, Guangxi University, Nanning 530004, China c Key Laboratory of Urban Environment and Health, Ningbo Urban Environment Observation and Research Station, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen 361021, China d State Key Lab of Urban and Regional Ecology, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing 100085, China A R T I C L E I N F O A B S T R A C T Keywords: Microplastic Antibiotic resistance genes Health risk Biotic factor Abiotic factor Microplastics (MPs) ubiquitous in environments promote the dissemination of antibiotic resistance genes (ARGs), threatening ecosystem safety and human health. However, quantitative assessments of the health risks of ARGs (HRA) in plastisphere and an in-depth exploration of their driving mechanisms are still lacking. Here, the microbiomes, ARGs, and community assembly processes of five types of MPs in an urbanizing watershed were systematically investigated. By fully considering the abundance, clinical availability, human pathogenicity, human accessibility, and mobility of 660 ARGs in plastisphere, the HRA on MPs were quantified and compared. Polyethylene had the highest HRA among the five MP types, and urbanization further increased its risk index. In addition to abiotic factors, more complex biotic factors have been shown to drive HRA in plastisphere. Specif­ ically, dispersal limitation accounted for the increasing diversity and interaction of bacteria that determined HRA in plastisphere. Further analysis of metabolic functions indicated that a higher HRA was accompanied by decreased normal metabolic functions of plastisphere microbiota due to the higher fitness costs of ARGs. This study advances the quantitative surveillance of HRA in plastisphere and understanding of its driving mecha­ nisms. This will be helpful for the management of both MPs and ARGs treatments for human health. 1. Introduction Plastic waste, which is ubiquitous in the environment, is becoming a geological indicator of the Anthropocene era (Zalasiewicz et al., 2016). In particular, microplastics (MPs, plastic debris < 5 mm) are of great concern because of their small size, refractory nature, and high mobility; they can move between different large-scale environments, including terrestrial habitats, water bodies, and air, and even accumulate in food webs (Allen et al., 2019; Wright et al., 2017; Galloway et al., 2016; Zhu et al., 2022). Moreover, MPs can provide a suitable habitat for microbial colonization and biofilm formation, termed the ‘plastisphere’ (Zettler et al., 2013). Many harmful microbial communities are selectively enriched in the plastisphere, including potential human pathogens with or without antibiotic resistance (ABR), which threaten ecosystem safety and human health (Yang et al., 2020; Lamb et al., 2018). Even worse, these harmful microbes can be transported by MPs across rivers and finally into the ocean and possibly become invasive species of fragile ecosystems (Gregory et al., 2009; Lebreton et al., 2017). Risk assessment of the plastisphere is a strategic priority for the management of detrimental microbes in the plastisphere and the asso­ ciated health threats. As one of the main threats, ABR causes nearly 700,000 deaths annually worldwide (May 2021). Although numerous studies have demonstrated the prevalence of antibiotic resistance genes (ARGs) and potential pathogens in plastispheres in water and soil eco­ systems, evaluating the Health Risks of ARGs (HRA) in the plastisphere is more complex and remains rarely quantified owing to the limitations of quantitative assessment approaches (Zhu et al., 2022; Yang et al., 2020). Recently, new frameworks for quantitatively and qualitatively * Corresponding author. E-mail address: lcui@iue.ac.cn (L. Cui). † These authors contributed equally to this work. https://doi.org/10.1016/j.watres.2023.120574 Received 16 June 2023; Received in revised form 24 August 2023; Accepted 2 September 2023 Available online 2 September 2023 0043-1354/© 2023 Elsevier Ltd. All rights reserved. L. Zhu et al. Water Research 245 (2023) 120574 2. Materials and methods assessing HRA in environments based on an integrative consideration of clinical availability, human pathogenicity, human accessibility, and mobility of ARGs and their abundance have been established, providing improved means of evaluating HRA in the plastisphere (Zhang et al., 2022b, 2021). Understanding the driving mechanisms underlying HRA is critical to mitigating the impact of plastisphere on human health. Both biotic and abiotic factors affect the HRA of MPs. Some studies have demonstrated that abiotic factors including urbanization, plastic type, and the physi­ cochemical properties of ecosystems, can affect ARGs on MPs (Wang et al., 2020; Li et al., 2021b). However, in-depth analysis of the biotic factors that directly contribute to the HRA of plastisphere, such as the microbial assembly process, functions of the microbiota on MPs, and their interactions with human activities and environmental factors, is lacking. The microbial community assembly determines the presence and abundance of species in the plastisphere, thereby influencing ecosystem function and microbial risks of the ecosystem (Xun et al., 2019). In addition, the diversity, interactions, and functions of the plastisphere microbiota that may affect the HRA of plastispheres war­ rant a detailed investigation. Rivers are an important link for transporting MPs from inland terrestrial habitats to oceans (Lebreton et al., 2017; Tibbetts et al., 2018). About 70–80% of plastic waste in the ocean was transported by rivers (Wagner et al., 2014). In addition, river watersheds, often accompanied by different urbanization levels, are ideal ecosystems for investigating the HRA of MPs and their driving mechanisms. Human activities introduce MPs into rivers in multiple ways, including sewage effluents, landfill leachate, and leakage from plastic production plants (McCormick et al., 2014; Eerkes-Medrano et al., 2015). Urbanization further increases both MPs and ARGs pollution in urban areas and sur­ rounding rivers (Wang et al., 2020) due to increased human activities (Yonkos et al., 2014), different land-use types (Baldwin et al., 2016), and large-scale wastewater treatment (McCormick et al., 2014). Based on the above background, this study focused on quantifying HRA in riverine MPs and deciphering the underlying driving mecha­ nisms. By utilizing metagenomics, 16S rRNA, and ITS sequencing, combined with advanced quantitative health risk assessment methods of ARGs and in-depth mechanism analysis of community assembly and function, we aimed to 1) explore the shifts of microbial communities and antibiotic resistome with urbanization in different MPs along the river, 2) quantify and compare the HRA of different MPs, and 3) reveal the underlying biotic and abiotic mechanisms driving HRA in plastisphere. 2.1. Study area and MPs selection The Beilun River (21◦ 32′–22◦ 45′N, 107◦ 32′–108◦ 03′E) was selected as the study area. It is a typical watershed with different urbanization levels flowing through rural, peri‑urban, and urban areas. It is 109 km long and forms part of the Sino-Vietnamese border (Fig. 1a). The river eventually flows into the Beibu Gulf, an important seafood breeding base, with a great risk of MPs spreading along the river into the sea and being devoured by fish or humans. Five types of commonly used MPs, namely polyethylene (PE), polypropylene (PP), polystyrene (PS), PEfiber (PF), and PE-fiber-PE (PFP) were purchased from Aladdin Chem­ icals (Shanghai, China). PFP is a sandwich structure of composite plas­ tics with three layers of PE, fiber, and PE, and widely used in the packages of our daily products such as milk carton, instant noodle box and so on. It is different from the composite PF with two layers (PE and fiber). The physicochemical properties (e.g., size, elemental composi­ tion, specific surface area, hydrophobicity, zeta potential, and Raman spectra) of the MPs are described in detail in our previous work (Li et al., 2021b, 2022). 2.2. Experimental design A 30-day in situ incubation experiment was conducted to study the riverine microbiome and antibiotic resistome in the plastisphere at 14 sites with different urbanization levels. The MPs were placed in a specially designed cylindrical barrel with a filter, that allowed adequate water exchange but prevented MPs from escaping from the device. The materials of both the cylindrical barrel and filter were stainless steel and did not contain any fiber or plastic. To avoid miscellaneous microbial contamination, the MPs were sterilized with 70% ethanol and rinsed three times with sterile water before incubation in the river. According to the level of socio-economy (e.g., population size, medical level, and domestic sewage) and land-use types around each sampling site, the 14 sampling sites were artificially divided into rural (sites 1–5), peri‑urban (sites 6–9), and urban (sites 10–14) areas. Detailed information about the sampling sites and devices can be found in the previous study (Li et al., 2021b). After 30 days of incubation, 84 samples were collected from 14 sampling sites, including five MP types and water samples. Three replicates were analyzed for each sample. MPs and water samples were divided into two parts: one part was stored at 4 ◦ C for physico­ chemical analysis and the other was at − 20 ◦ C for DNA extraction. Ur­ banization and environmental property analyses are described in the Supporting Information (Text S1). Fig. 1. Spatial distribution of bacterial and fungal diversity and their relationships with urbanization level. a) Sampling sites of MPs and water along the river. b) Regression analysis between the Shannon-Weiner index and sampling sites. The level of urbanization gradually increases from sites 1 to 14. 2 L. Zhu et al. Water Research 245 (2023) 120574 (4) 2.3. Genome DNA extraction and amplicon sequencing RI = HA × MO × HP × CA Genomic DNA of both microorganisms in the MPs and water samples was directly extracted using the FastDNA Spin Kit for Soil (MP Bio­ medicals, Santa Ana, CA, USA) according to the manufacturer’s in­ struction (Zhu et al., 2017). The concentration and quality of the extracted DNA were inspected using Nanodrop ND-1000 (Thermo Fisher Scientific, Wilmington, USA) and Qubit 3.0 Fluorometer (Thermo Fisher Scientific, Inc., Waltham, USA). The procedure steps for library con­ struction, 16S rRNA and ITS amplicon sequencing, and bioinformatics analysis are detailed in the Supporting Information (Text S2). Although these four indicators may have different proportions for the RI of ARGs, there is currently no evidence to determine which indicator is the most important (or how important it is); thus, a ‘non-discrimina­ tory’ approach without adding coefficients to each indicator was adopted. The RI value for each sample containing different ARGs was calcu­ lated as RIsample = Σni=1 Abundancei ×RIi where Abundancei and RIi are the abundance and RI values of each type of ARGi in the sample, respectively. 2.4. Metagenomic sequencing and bioinformatic analysis A paired-end metagenomic library was constructed using NEXTFLEX Rapid DNA-Seq after the DNA extracts were fragmented to approxi­ mately 400 bp using Covaris M220. The paired-end library was sequenced using Illumina NovaSeq 6000 (Illumina Inc., San Diego, CA, USA) with NovaSeq Reagent Kits (Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China). After sequencing, fastp was used to trim adaptors and remove low quality reads (i.e., length < 50 bp, quality value < 20, and N bases) (Chen et al., 2018). High-quality reads were assembled using MEGAHIT (Li et al., 2015). Contigs > 300 bp were used for taxonomic and functional annotations. Open reading frames (ORFs) were predicted from assembled contigs using MetaGene (Noguchi et al., 2006). ORFs longer than 100 bp were retrieved and translated in to amino acid sequences. CD-HIT (Fu et al., 2012) and SOAPaligner (Li et al., 2008) were used to construct non-redundant gene catalogs based on 90% sequence identity and coverage and to calculate gene abundance based on 95% identity. Taxonomic and functional annotations of representative sequences of non-redundant gene catalogs were performed using Diamond (Buchfink et al., 2015) against the National Center for Biotechnology Information (NCBI), Non-Redundant Protein Sequence Database (NR), Kyoto Ency­ clopedia of Genes and Genomes (KEGG), and the Comprehensive Anti­ biotic Research Database (CARD) with an e-value cutoff of 1e− 5. The raw sequencing data for metagenomic, 16S rRNA, and ITS sequencing were submitted to the NCBI Sequence Read Archive under accession number PRJNA946910. 2.6. Statistical analysis The alpha diversity of the bacterial and fungal communities was calculated using QIIME 1.9.1. Nonmetric multidimensional scaling (NMDS) based on the Bray-Curtis distance was performed using the vegan package (Dixon, 2003). Network analysis was performed based on Spearman’s correlation (r > 0.9, P < 0.01) and the topological properties of the networks were calculated using Gephi 0.9.2. Community assem­ bly mechanisms were inferred using phylogenetic bin-based null model analysis (iCAMP) to partition the relative importance of community assembly processes in plastisphere (Ning et al., 2020). Structural equa­ tion modeling (SEM) was conducted using AMOS 21.0 (Wu et al., 2019a; Bahram et al., 2018). 3. Results 3.1. Diversity and abundance of the microbial community in five riverine MPs with urbanization We conducted an in -situ incubation experiment by placing five types of MP along a river with significant urbanization gradients for 30 days. A total of 28,097 bacterial and 15,114 fungal OTUs were identified across all 70 MP samples (five polymer types × 14 sites) and 14 water samples along the river from rural to urban areas (Fig. 1a). The rarefaction curves reflected sufficient depth and accuracy of the 16S rRNA gene and ITS sequencing (Fig. S1). There were significant differences in the commu­ nity composition between the plastisphere and water for both bacteria and fungi (Adonis, P = 0.001; Figs. S2 and S3). Bacterial alpha diversity significantly increased in the plastisphere and water with urbanization (P < 0.01; Fig. 1b and Fig. S4). The slope of the fitted curve of the plastisphere (0.19) was lower than that of water (0.25), suggesting that bacterial alpha diversity in the plastisphere was less affected by ur­ banization than that in the water. However, fungal alpha diversity significantly increased with urbanization only in the water (P < 0.01), rather than in the plastisphere (P > 0.05). In addition, the number of shared OTUs between the plastisphere and water significantly increased with urbanization, indicating that microorganisms in water are more likely to colonize MPs with urbanization (Text S3, Fig. S5). There were no significant differences in microbial diversity among the different MP types (P > 0.05). However, when considering the different geographical areas separately (Fig. S6), bacterial diversity in the PP plastisphere was found to be the lowest in rural areas, but was the highest among the five MP types in peri‑urban and urban areas. For fungal communities, the diversity of PE plastisphere was significantly higher than that of the PP plastisphere in rural areas, whereas the opposite was observed in peri‑urban areas. These results suggest that, in addition to the urbanization level, polymer type may be another factor affecting microbial diversity in the plastisphere. To further explore the detailed differences in microbial community structure in the plastisphere between rural and urban areas, we calcu­ lated the significantly enriched OTUs in different areas. The number of 2.5. Calculation of the HRA in plastisphere The HRA in plastisphere was calculated based on metagenomic data by considering not only the gene abundance but also four important indicators: human accessibility (HA, the ability of ARGs to transfer from the environment to humans), mobility (MO, the ability of ARGs to transfer by horizontal gene transfer), human pathogenicity (HP, the ability of ARGs to transfer from nonpathogenic bacteria to pathogenic bacteria), and clinical availability (CA, the consumption of commonly used antibiotics associated with ARGs in clinics) as previously reported (Zhang et al., 2022b). These indicators were calculated as follows: HA = Average abundancehuman × Prevalencehuman (1) HP = Numberpathogenic/Numberall (2) CA = Σni=1 Usage of antibiotici (3) (5) where Average abundancehuman and Prevalencehuman are the average abundance and prevalence of ARGs in human-associated habitats, respectively. Numberpathogenic and Numberall are the number of patho­ genic hosts and the total number of hosts of ARGs, respectively. The ‘n’ in CA represents the number of antibiotics that a ARG can resist. Addi­ tionally, the MO of the ARGs was determined based on the number of related MGEs in the completed genomes. Considering that the health risk index (RI) is proportional to each indicator, the RI of each ARG type was obtained by multiplying these indicators as follows: 3 L. Zhu et al. Water Research 245 (2023) 120574 Notably, in contrast to microbial diversity, which showed an increasing trend with urbanization, only the abundance of ARGs in the PP plasti­ sphere increased with urbanization levels, indicating that there may be different colonization behaviors between bacteria with and without ARGs. For example, bacteria carrying the blaPER-1 gene showed higher cell adhesiveness and biofilm formation than bacteria without this ARG (Zhang et al., 2019). We further analyzed the mobility of ARGs on different MPs by calculating the abundance of MGEs located on the same contig of ARGs (Fig. 2b). The mobility of ARGs in PE, PP, and PS was significantly higher than that in PF (Krusal-Wails, P = 0.007). The ARGs that cooccurred with MGEs were mainly located on plasmids rather than chromosomes (Fig. 2c), representing a potentially higher transmission risk. Furthermore, the dominant predicted hosts of the ARGs were Proteobacteria, Actinobacteria, and Chloroflexi at the phylum level. By comparing ARGs hosts (at the species level) with the Human Pathogenic Bacteria (HPB) database, the composition of HPB hosts showed signifi­ cant differences between the PF and other plastispheres (Fig. 2d). Many dangerous HPB with high abundances in PE, PP, PS, and PFP were not detected in PF, such as Acinetobacter baumannli, Pseudomonas koreensis, and Acinetobacter johnsonii. Based on the abundance, mobility, and HPB hosts of the ARGs, different HRA can be anticipated in different plastispheres. To fully quantify HRA in the plastisphere, all major risk factors, including clinical availability, human accessibility, human pathoge­ nicity, and mobility were considered to calcualte the health risk index (RI) according to a recently developed risk assessment method (Fig. 2f, significantly enriched bacterial OTUs was much higher in urban areas than in rural areas for all five MP types (Fig. S7a). Compared to the rural plastisphere, the PFP and PP plastispheres in urban areas enriched the highest number of bacterial OTUs (284 and 51, respectively), which was consistent with the results of functional prediction about biofilm for­ mation (Fig. S8). Fungal OTUs were significantly enriched in most urban plastispheres compared to those in rural plastispheres, except for PE (Fig. S7b). The significant enrichment of different microbial taxa reflects the heterogeneity of different MP types and geographical areas, sug­ gesting that urbanization and MP type collectively affect microbial di­ versity and abundance in the plastisphere. 3.2. Antibiotic resistome and their quantitative health risks in the plastisphere To further assess HRA in the plastisphere, metagenomics was used to quantify the ARGs. The coverage of metagenomic data was 90%, indi­ cating sufficient reliability. Metagenomic sequencing identified 660 ARGs that conferred resistance to 21 major classes of antibiotics in five different types of MP plastispheres. Among these, 77 ARGs (11.67%) conferred resistance to multidrug classes, with an average relative abundance of up to 57.8% in the plastisphere (Fig. 2a). Additionally, 394 of the 660 detected ARGs (60%) conferred resistance to beta-lactam antibiotics, which are widely used in clinics (Fig. S9) (World Health Organization, 2018). Regarding the relative abundance of ARGs in different plastispheres, PE displayed a significantly higher abundance than the other MP types (P < 0.05), while the PF had the lowest. Fig. 2. Distribution patterns of ARGs and their health risk in the plastisphere. a) ARGs abundance in five types of MPs in rural (R), peri‑urban (P) and urban (U) areas. b) Abundance of MGEs located on the same contig with ARGs. c) Genetic location of contigs carrying both ARGs and MGEs. d) Potential human pathogenic bacteria (HPB) hosts of ARGs. Only the top 12 most abundant species are shown. e) Health risk index of different plastisphere. f) Framework for assessing the HRA in the plastisphere by considering clinical availability, human pathogenicity, human accessibility, and mobility. g) HRA for each class in different plastisphere. **, P < 0.01; ***, P < 0.001. 4 L. Zhu et al. Water Research 245 (2023) 120574 Table S1) (Zhang et al., 2022b). The HRA of PE was significantly higher than those of the other four types of MPs (P < 0.01; Fig. 2e). Notably, unlike the abundance pattern of ARGs, HRA increased significantly from rural to urban plastispheres for PE, PP, and PS (P < 0.05), suggesting that abundance did not directly reflect health risk. One possible reason for this is that the high use of antibiotics in urban areas increases the clinical availability of ARGs and potentially increases the health risks to humans (Bungau et al., 2021). The HRA in the polymers of plastics and fibers (i.e., PF and PFP) slightly decreased with urbanization. In addi­ tion, the RI of all ARG classes in the PF plastisphere were lower than those in the PE plastisphere (Fig. 2g). Multidrug resistance was the highest-risk class among the various ARG classes. These results suggest that polymer type is a major determinant of HRA in plastisphere, and urbanization may further increase their HRA. 0.96, P = 0.009) and PS (r = 0.89, P = 0.04) plastispheres, which were the two highest-risk MPs (Fig. 3c and Fig. 2e). In addition, MP type was the second most important abiotic factor affecting HRA. Among the properties of the different MPs, the specific surface area showed a sig­ nificant positive correlation with the health risk index (Fig. 3d, r = 0.35, P = 0.004). In addition, biotic factors, such as microbial community assembly, diversity, and metabolic function had more direct effects on HRA in the plastisphere (Fig. 3a, b). The following two sections analyze these biotic factors and their association with HRA in detail. 3.4. Microbial metabolic functions of plastisphere and their association with HRA To further explain the underlying biotic mechanisms driving HRA in the plastisphere, we analyzed the metabolic functions in the plastisphere and their association with HRA. Perfect microbial metabolic functions are usually recognized as helpful in protecting against pathogenic mi­ croorganisms. However, microbial metabolic functions have long been overlooked as factors affecting HRA in the plastisphere. The primary patterns of variation in microbial metabolic functions in the plastisphere were separated according to the type of MPs (NMDS, stress = 0.072; Fig. 4a). Notably, microbial metabolic functions in PE and PF were completely separated in the NMDS plots, similar to their health risk patterns. To explore the underlying factors, significant differences in the metabolic functions between PE and PF were compared (Fig. 4b). Con­ trary to the higher HRA of PE, 13/15 of the most abundant metabolic functions were enriched in the PF plastisphere, including many impor­ tant functions associated with basic microbial life activities, such as carbon metabolism, oxidative phosphorylation, pyruvate metabolism, TCA cycle, and amino acid metabolism. Moreover, when considering the five MP types together, the metabolic functions were differentiated by the urbanization level and most (9/10) significantly decreased in abundance from rural to urban areas (Fig. 4c). These results indicate that the basic metabolic function of microbes showed an opposite pattern to that of HRA in the plastisphere. This was further confirmed through the correlation analysis between HRA and 50 major metabolic functions in the plastisphere (Fig. S10). Most metabolic functions were significantly negatively correlated with HRA, especially functions associated with replication, recombination, and repair of genes, such as DNA replica­ tion, mismatch repair, and homologous recombination (P < 0.001). 3.3. Multiple abiotic and biotic factors driving the HRA in plastisphere Multiple biotic and abiotic factors potentially affecting HRA in plastisphere were considered to fully investigate the underlying driving mechanisms. Abiotic factors include urbanization levels (population size, medical level, and domestic sewage production), MP type, anti­ biotic residuals in water (fluoroquinolones, tetracycline, macrolides, sulfonamides, and diaminopyrimidines), and the physicochemical properties of water (flow rates, pH, dissolved oxygen, DOC, total N, total P, NH3-N BOD5, and COD). Biotic factors in the plastisphere include microbial assembly processes, diversity, and metabolic functions. SEM showed that 65% of the total variance in HRA could be explained by the above biotic and abiotic factors. Among these factors, urbanization and MP type were the two most critical abiotic factors determining HRA in the plastisphere (Fig. 3a, b). Urbanization had a positive direct effect on HRA in the plastisphere (λ = 0.48, P = 0.001). Nevertheless, MP type had indirect effects on HRA, mediated by microbial metabolic function (λ1 = − 0.69, P < 0.001; λ2 = − 0.50, P < 0.01). In addition, the com­ munity assembly process had indirect effects on HRA in the plastisphere by influencing bacterial diversity (λ1 = 0.48, P < 0.001; λ2 = − 0.49, P < 0.05). As indicated by the standard total effects from the SEM, urbanization was the most important abiotic factor affecting HRA (Fig. 3b). Among the factors associated with urbanization, the amount of sewage pro­ duction had a significant positive correlation with HRA in the PE (r = Fig. 3. Multiple factors driving the HRA in the plastisphere. a) Sankey diagram showing the comprehensive relationships among urbanization, MP type, antibiotic and physicochemical properties of water, microbial diversity, assembly process, metabolic function, and HRA in the plastisphere. Data were obtained from the results of the structural equation model (SEM). The thicker the line, the greater the impact. The gray line represents non-significant effects (P > 0.05) and all other color lines represent significant effects (P < 0.05). +, positive effect; -, negative effect. Model χ2 = 15.42, d.f. = 10, P = 0.12; b) Standard total effects (direct plus indirect of significant effects) on the HRA in the plastisphere derived from SEM. Relationships between HRA in plastisphere and c) sewage production and d) specific surface area of MPs via linear regression analysis. 5 L. Zhu et al. Water Research 245 (2023) 120574 Fig. 4. Microbial metabolic functions and their relationships with HRA in the plastisphere. a) Non-metric multidimensional scaling (NMDS) plot showing functional variation among samples implied that primary clustering is by MP types of PE, PP, PS, PF, and PFP; b) Metabolic functions (level 3 from KEGG database) with significant differences between PE and PF plastisphere. Only the metabolic functions with the top fifteen relative abundances of significantly different functions are shown. c) Metabolic functions (level KO from KEGG database) with significant differences between different urbanization levels by considering five types of MPs together. Only the metabolic functions with the top ten relative abundances of significantly different functions are shown. Metabolic function data are derived from metagenomic sequencing. *, P < 0.05; **, P < 0.01. These results demonstrated that microbial metabolic functions are important factors in the suppression and regulation of HRA in plastisphere. significantly enriched OTUs from the urban plastisphere, the size, con­ nectivity, and degree of the network decreased, whereas the density increased. This suggests that the enriched OTUs in the urban plasti­ sphere play an important role in the stability of the microbial ecological network. A significantly larger proportion of bacterial OTUs were contained in networks than that of fungal OTUs (Fig. 5e), suggesting the dominant role of bacterial communities in the formation of biofilms in MPs, and the proportion further increased from rural to urban plastispheres. Keystone nodes that were highly connected to other members of the network were identified. For example, Anaeromyxobacter, which was connected to 191 different nodes and was enriched in PFP plastispheres in urban areas, was identified as playing a core role in the co-occurrence network of urban plastispheres (Fig. 5c). Anaeromyxobacter has been confirmed to be involved in nitrogen fixation in many previous studies (Masuda et al., 2020; Knief et al., 2012) and may thus provide a nitrogen source for the connected members in the plastisphere. Notably, most of the core anaerobic bacteria (e.g., Anaeromyxobacter and Anaerolinea) were hosts of ARGs based on host prediction from metagenomic data. 3.5. Microbial community assembly of plastisphere and their association with HRA To further understand ecological mechanisms controlling the aboverevealed microbial diversity and metabolic function in different plasti­ sphere, the relative importances of microbial community assembly processes were quantified using ‘iCAMP’ (Ning et al., 2020). The results showed that dispersal limitation, homogeneous selection, and drift were more important than other processes in bacterial and fungal assembly processes, with average relative importances of 34.7–48.9%, 26.6–32.8%, and 19.7–25.4%, respectively (Fig. 5a). Overall, stochastic processes dominated the microbial community assembly process in the plastisphere, with a relative importance exceeding 64.3%. The relative importance of different assembly processes in the plastisphere changed with increasing urbanization. For bacterial community assembly, the ratio of dispersal limitation increased significantly with urbanization (P < 0.001; Fig. 5a). This may cause a dramatic increase in bacterial di­ versity in the plastisphere (Fig. 1b) and further increase HRA in PE, PP, and PS (Fig. 2e). Interestingly, the variation was almost the opposite for the fungal community assembly, in which the ratio of drift increased and the ratio of dispersal limitation decreased with urbanization (P < 0.001, Fig. 5a). A comparison of the relative importance of community as­ sembly processes in different types of MPs showed significant differ­ ences (Fig. 5b). Collectively, our results indicated that although stochastic processes mainly govern community assembly processes in the plastisphere, urbanization and MP type are two deterministic factors affecting the relative importance of each ecological process. The assembly process of microbial communities may also involve interactions between species, which were explored using co-occurrence network analysis (Fig. 5c). The microbial networks were larger in size (total nodes), with greater connectivity (total number of edges) and higher degree and density in the peri‑urban and urban plastispheres than in rural plastisphere (Fig. 5d). From rural to urban areas, the mi­ crobial network exhibited more significant small-world characteristics, with a decreased average path length (from 6.09 to 4.06) and an increased average clustering coefficient (from 0.419 to 0.441). Never­ theless, modularity decreased from rural to urban areas, suggesting decreased stability of the network (Olesen et al., 2007). This unstable network structure may further increase the stochastic assembly of mi­ crobial communities in the plastisphere (Fig. 5a). After removing the 4. Discussion MPs have been reported as ‘hotspots’ and ‘rafts’ of ARGs to facilitate long-distance transport of ABR in environments (Li et al., 2021a; Amaral-Zettler et al., 2020; Wright et al., 2021; Wu et al., 2019b). However, only the abundance or diversity of ARGs in plastisphere has been analyzed in previous studies, which is not sufficient to assess their health risks (Xu et al., 2022; Martínez et al., 2015). This study quantified for the first time the HRA of five types of MPs in a large-scale river watershed by integrating clinical availability, human pathogenicity, human accessibility, mobility, and the abundance of ARGs for holistic risk assessment. Among the five MP types, special attention should be paid to the PE with the highest HRA, which increased with specific surface area and was further exacerbated by urbanization (Figs. 2e, 3c, d). Notably, PE was particularly abundant among the different polymer types because of its desirable product performance, easy manufacturing, and low cost. For PP and PS, although their RI values were much lower than those of PE, there was still an increasing trend for RI from rural to urban plastispheres. For PF and PFP (composites of plastic and fiber), the RI values were lower than those of traditional MPs (i.e., PE, PP, and PS) and did not increase with urbanization. This finding indicates that these plastic and fiber composites (especially PF) may be alternatives to traditional plastics in terms of reducing their HRA. For the future management of MP pollution, we should systematically and 6 L. Zhu et al. Water Research 245 (2023) 120574 Fig. 5. Microbial assembly processes and potential interactions in the plastisphere. a) Bacterial and fungal community assembly processes in rural (inner ring) and urban (outer ring) plastisphere. a) Comprehensive results from five types of MPs, i.e., PE, PP, PS, PF, and PFP. b) Relative importance of different assembly processes for bacterial and fungal communities in different plastisphere. c) Visualization of constructed microbial ecological networks from rural to urban plastisphere. The final network indicates the removal of all enriched OTUs (Fig. S7) from the microbial network of the urban plastisphere. d) Changes of network topology from rural to urban plastisphere. URO, urban plastisphere after removing enriched OTUs. e) Relative contributions of bacteria and fungi in microbial networks from rural to urban plastisphere. *, P < 0.05. quantitatively evaluate the health risks, instead of simply focusing on the abundance and type of microbiome and ARGs in the plastisphere. We further explored the external and intrinsic drivers of HRA in the plastisphere by considering multiple biotic and abiotic factors. Urbani­ zation level and polymer type are abiotic factors that influence HRA in the plastisphere. Of note, sewage production, a representative of ur­ banization level, was positively correlated with HRA in the PE and PS plastispheres, suggesting that human activities contribute to increasing HRA, especially in these two MPs (Fig. 3c). A previous study also showed that human-associated habitats posed a much higher HRA in the envi­ ronment on a global scale (May 2021). Additionally, the specific surface area of MPs is an important abiotic factor because of its positive corre­ lation with HRA in the plastisphere (Fig. 3d). Therefore, reducing sewage discharge and monitoring the specific surface area of MPs are two potential measures for controlling and reducing HRA in the plastisphere. In addition to abiotic factors, biotic factors directly affect HRA in the plastisphere. First, the community assembly process is an important biotic factor that influences bacterial diversity and determines HRA in the plastisphere (Fig. 3a). Similar to previous studies on estuarine and marine MPs (Sun et al., 2021; Zhang et al., 2022a), dispersal limitation played the most important role in the community assembly of riverine plastispheres in this study. This may be because MP particles are usually separated from each other in water; thus, the migration of microbes from one MP particle to another, mediated by the surrounding water, is limited (Zhang et al., 2019b). Although stochastic processes are the main dirvers of the plastisphere, community assembly is partially influenced by deterministic processes. Allochthonous inputs from terrestrial sources, sewage, sludge, and humans may contribute to increasing microbial diversity and ARGs in urban rivers (Huang et al., 2020) and further contribute to their colonization on MPs. Increasing bacterial diversity and community complexity in plastispheres with urbanization provides more potential hosts for ARGs transmission, potentially further increasing HRA in the plastisphere (i.e., PE, PP, and PS, Fig. 2e). Second, microbial interactions may be another important factor influencing HRA in the plastisphere. During the microbial community assembly processes, microbial interactions are very complex to adapt to changing environments (Yuan et al., 2021). The small-world charac­ teristics of microbial networks in urban plastispheres allow the effects of perturbation to distribute rapidly through the entire network, rendering the whole system efficient and the network more sensitive (i.e., unsta­ ble) to external changes (Berry et al., 2014). Closer connections in mi­ crobial communities may provide more opportunities for the spread of ARGs among species, further increasing HRA in the plastisphere. The increased positive interactions among bacteria in the network with ur­ banization (Fig. 5c) may indicate increased limitation due to shared environmental conditions and limited resources (Hesse et al., 2021; Fuhrman et al., 2009), which explains the increased dispersal limitation in the urban plastispheres (Fig. 5a). This phenomenon also fits well with 7 L. Zhu et al. Water Research 245 (2023) 120574 the stress gradient hypothesis that interactions between species become more positive under abiotic stress (Dunphy et al., 2019; Hammarlund et al., 2019), such as the urbanization gradient in this study. Third, the microbial metabolic function is another critical but longterm overlooked biotic factor that influences HRA in the plastisphere. A negative relationship between the metabolic function and HRA in the plastisphere was observed. A possible reason for this is the higher fitness costs of antibiotic resistant strains, which may display decreased meta­ bolic activity relative to the corresponding sensitive strains. This is consistent with a previous culture-based study showing that lab-evolved strains with ABR mostly show decreased metabolic activity relative to the unevolved ancestor (Björkman et al., 2000). To adapt to the increasing environmental stress due to urbanization, microorganisms carrying ABR must give up some of their basic metabolic activities. Moreover, metabolic functions related to DNA replication, homologous recombination, and mismatch repair showed significant negative cor­ relations with HRA in the plastisphere, suggesting a critical role of these functions in reducing HRA in the plastisphere. References Allen, S., Allen, D., Phoenix, V.R., Le Roux, G., Durántez Jiménez, P., Simonneau, A., Binet, S., Galop, D., 2019. Atmospheric transport and deposition of microplastics in a remote mountain catchment. Nat. Geosci. 12 (5), 339–344. Amaral-Zettler, L.A., Zettler, E.R., Mincer, T.J., 2020. Ecology of the plastisphere. Nat. Rev. Microbiol. 18 (3), 139–151. Bahram, M., Hildebrand, F., Forslund, S.K., Anderson, J.L., Soudzilovskaia, N.A., Bodegom, P.M., Bengtsson-Palme, J., Anslan, S., Coelho, L.P., Harend, H., HuertaCepas, J., Medema, M.H., Maltz, M.R., Mundra, S., Olsson, P.A., Pent, M., Põlme, S., Sunagawa, S., Ryberg, M., Tedersoo, L., Bork, P., 2018. Structure and function of the global topsoil microbiome. Nature 560 (7717), 233–237. Baldwin, A.K., Corsi, S.R., Mason, S.A., 2016. Plastic debris in 29 great lakes tributaries: relations to watershed attributes and hydrology. Environ. Sci. Technol. 50 (19), 10377–10385. Berry, D., Widder, S., 2014. Deciphering microbial interactions and detecting keystone species with co-occurrence networks. Front. Microbiol. 5, 219. Björkman, J., Nagaev, I., Berg, O.G., Hughes, D., Andersson, D.I., 2000. Effects of environment on compensatory mutations to ameliorate costs of antibiotic resistance. Science 287 (5457), 1479–1482. Buchfink, B., Xie, C., Huson, D.H., 2015. Fast and sensitive protein alignment using DIAMOND. Nat. Methods 12 (1), 59–60. Bungau, S., Tit, D.M., Behl, T., Aleya, L., Zaha, D.C., 2021. Aspects of excessive antibiotic consumption and environmental influences correlated with the occurrence of resistance to antimicrobial agents. Curr. Opin. Environ. Sci. Health 19, 100224. Chen, S., Zhou, Y., Chen, Y., Gu, J., 2018. Fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34 (17), i884–i890. Dixon, P., 2003. VEGAN, a package of R functions for community ecology. J. Veg. Sci. 14 (6), 927–930. Dunphy, L.J., Yen, P., Papin, J.A., 2019. Integrated experimental and computational analyses reveal differential metabolic functionality in antibiotic-resistant pseudomonas aeruginosa. Cell Syst. 8 (1), 3–14. Eerkes-Medrano, D., Thompson, R.C., Aldridge, D.C., 2015. Microplastics in freshwater systems: a review of the emerging threats, identification of knowledge gaps and prioritisation of research needs. Water Res. 75, 63–82. Fu, L.M., Niu, B.F., Zhu, Z.W., Wu, S.T., Li, W.Z., 2012. CD-HIT: accelerated for clustering the next-generation sequencing data. Bioinformatics 28 (23), 3150–3152. Fuhrman, J.A., 2009. Microbial community structure and its functional implications. Nature 459 (7244), 193–199. Galloway, T.S., Lewis, C.N., 2016. Marine microplastics spell big problems for future generations. Proc. Natl. Acad. Sci. 113 (9), 2331–2333. Gregory, M.R., 2009. Environmental implications of plastic debris in marine settingsentanglement, ingestion, smothering, hangers-on, hitch-hiking and alien invasions. Philos. Trans. R. Soc. B-Biol. Sci. 364 (1526), 2013–2025. Hammarlund, S.P., Harcombe, W.R., 2019. Refining the stress gradient hypothesis in a microbial community. Proc. Natl. Acad. Sci. 116 (32), 15760–15762. Hesse, E., O’Brien, S., Luján, A.M., Sanders, D., Bayer, F., van Veen, E.M., Hodgson, D.J., 2021. Buckling, A. Stress causes interspecific facilitation within a compost community. Ecol. Lett. 24 (10), 2169–2177. Huang, L., Bai, J., Wen, X., Zhang, G., Zhang, C., Cui, B., Liu, X., 2020. Microbial resistance and resilience in response to environmental changes under the higher intensity of human activities than global average level. Glob. Change Biol. 26 (4), 2377–2389. Knief, C., Delmotte, N., Chaffron, S., Stark, M., Innerebner, G., Wassmann, R., von Mering, C., Vorholt, J.A., 2012. Metaproteogenomic analysis of microbial communities in the phyllosphere and rhizosphere of rice. ISME J. 6 (7), 1378–1390. Lamb, J.B., Willis, B.L., Fiorenza, E.A., Couch, C.S., Howard, R., Rader, D.N., True, J.D., Kelly, L.A., Ahmad, A., Jompa, J., Harvell, C.D., 2018. Plastic waste associated with disease on coral reefs. Science 359 (6374), 460–462. Lebreton, L.C.M., Zwet, J.v.d., Damsteeg, J.W., Slat, B., Andrady, A., Reisser, J., 2017. River plastic emissions to the world’s oceans. Nat. Commun. 8 (1), 15611. Li, R., Li, Y., Kristiansen, K., Wang, J., 2008. SOAP: short oligonucleotide alignment program. Bioinformatics 24 (5), 713–714. Li, D., Liu, C.M., Luo, R., Sadakane, K., Lam, T.W., 2015. MEGAHIT: an ultra-fast singlenode solution for large and complex metagenomics assembly via Succinct de Bruijn graph. Bioinformatics 31 (10), 1674–1676. Li, C., Wang, L., Ji, S., Chang, M., Wang, L., Gan, Y., Liu, J., 2021a. The ecology of the plastisphere: microbial composition, function, assembly, and network in the freshwater and seawater ecosystems. Water Res. 202, 117428. Li, R., Zhu, L., Yang, K., Li, H., Zhu, Y.G., Cui, L., 2021b. Impact of urbanization on antibiotic resistome in different microplastics: evidence from a large-scale whole river analysis. Environ. Sci. Technol. 55 (13), 8760–8770. Li, R., Zhu, L., Cui, L., Zhu, Y.G., 2022. Viral diversity and potential environmental risk in microplastic at watershed scale: evidence from metagenomic analysis of plastisphere. Environ. Int. 161, 107146. Martínez, J.L., Coque, T.M., Baquero, F., 2015. What is a resistance gene? Ranking risk in resistomes. Nat. Rev. Microbiol. 13 (2), 116–123. Masuda, Y., Yamanaka, H., Xu, Z.X., Shiratori, Y., Aono, T., Amachi, S., Senoo, K., Itoh, H., 2020. Diazotrophic anaeromyxobacter isolates from soils. Appl. Environ. Microbiol. 86 (16) e00956-20. May, M., 2021. Tomorrow’s biggest microbial threats. Nat. Med. 27 (3), 358–359. McCormick, A., Hoellein, T.J., Mason, S.A., Schluep, J., Kelly, J.J., 2014. Microplastic is an abundant and distinct microbial habitat in an urban river. Environ. Sci. Technol. 48 (20), 11863–11871. Ning, D., Yuan, M., Wu, L., Zhang, Y., Guo, X., Zhou, X., Yang, Y., Arkin, A.P., Firestone, M.K., Zhou, J., 2020. A quantitative framework reveals ecological drivers 5. Conclusion We comprehensively evaluated the HRA of five plastispheres in a river watershed with increasing urbanization levels and revealed both abiotic and biotic mechanisms driving the HRA. Among the five types of MPs, PE, one of the most commonly used types of plastic worldwide, had the highest HRA. Thus, the prevalence of PE in the environment should be strictly controlled. Urbanization increased HRA in the plastisphere, especially for PE, PP, and PS, suggesting that strict control of sewage discharge may contribute to reducing HRA in the plastisphere. Mor­ evoer, decreasing dispersal limitation and microbial interaction and increasing metabolic function may further decrease HRA in the plasti­ sphere. This study provides important insights into the use of these abiotic and biotic factors as regulators of HRA in riverine plastispheres. These findings extend our knowledge of HRA in the plastisphere and provide a theoretical basis for controlling MP pollution. Author contributions LC, LZ, and RL conceived and designed the research; RL and YK performed experiments; FX and CL analyzed data, LZ and RL wrote the draft; QC, DZ, QS, YGZ, and LC edited and reviewed the draft. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. Acknowledgments This work was financially supported by the National Natural Science Foundation of China (22241603, 32100083), the Key Collaborative Research Program of the Alliance of International Science Organizations (ANSO–CR-KP-2020-03), and the CAS Youth Interdisciplinary Team (JCTD-2021-13). Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.watres.2023.120574. 8 L. Zhu et al. Water Research 245 (2023) 120574 Yang, K., Chen, Q.L., Chen, M.L., Li, H.Z., Liao, H., Pu, Q., Zhu, Y.G., Cui, L., 2020. Temporal dynamics of antibiotic resistome in the plastisphere during microbial colonization. Environ. Sci. Technol. 54 (18), 11322–11332. Yonkos, L.T., Friedel, E.A., Perez-Reyes, A.C., Ghosal, S., Arthur, C.D., 2014. Microplastics in four estuarine rivers in the Chesapeake Bay USA. Environ. Sci. Technol 48 (24), 14195–14202. Yuan, M.M., Guo, X., Wu, L., Zhang, Y., Xiao, N., Ning, D., Shi, Z., Zhou, X., Wu, L., Yang, Y., Tiedje, J.M., Zhou, J., 2021. Climate warming enhances microbial network complexity and stability. Nat. Clim. Change 11 (4), 343–348. Zalasiewicz, J., Waters, C.N., Ivar do Sul, J.A., Corcoran, P.L., Barnosky, A.D., Cearreta, A., Edgeworth, M., Galuszka, A., Jeandel, C., Leinfelder, R., McNeill, J.R., Steffen, W., Summerhayes, C., Wagreich, M., Williams, M., Wolfe, A.P., Yonan, Y., 2016. The geological cycle of plastics and their use as a stratigraphic indicator of the anthropocene. Anthropocene 13, 4–17. Zettler, E.R., Mincer, T.J., Amaral-Zettler, L.A., 2013. Life in the “plastisphere”: microbial communities on plastic marine debris. Environ. Sci. Technol. 47 (13), 7137–7146. Zhang, M., Wang, L., Xu, M., Zhou, H., Wang, S., Wang, Y., Bai, M., Zhang, C., 2019. Selective antibiotic resistance genes in multiphase samples during biofilm growth in a simulated drinking water distribution system: occurrence, correlation and lowpressure ultraviolet removal. Sci. Total Environ. 649, 146–155. Zhang, A.N., Gaston, J.M., Dai, C.L., Zhao, S., Poyet, M., Groussin, M., Yin, X., Li, L.G., van Loosdrecht, M.C.M., Topp, E., Gillings, M.R., Hanage, W.P., Tiedje, J.M., Moniz, K., Alm, E.J., Zhang, T., 2021. An omics-based framework for assessing the health risk of antimicrobial resistance genes. Nat. Commun. 12, 4765. Zhang, S.J., Zeng, Y.H., Zhu, J.M., Cai, Z.H., Zhou, J., 2022a. The structure and assembly mechanisms of plastisphere microbial community in natural marine environment. J. Hazard. Mater. 13. Zhang, Z., Zhang, Q., Wang, T., Xu, N., Lu, T., Hong, W., Penuelas, J., Gillings, M., Wang, M., Gao, W., Qian, H., 2022b. Assessment of global health risk of antibiotic resistance genes. Nat. Commun. 13, 1553. Zhu, Y.G., Zhao, Y., Li, B., Huang, C.L., Zhang, S.Y., Yu, S., Chen, Y.S., Zhang, T., Gillings, M.R., Su, J.Q., 2017. Continental-scale pollution of estuaries with antibiotic resistance genes. Nat. Microbiol. 2, 16270. Zhu, D., Ma, J., Li, G., Rillig, M.C., Zhu, Y.G., 2022. Soil plastispheres as hotspots of antibiotic resistance genes and potential pathogens. ISME J. 16 (2), 521–532. of grassland microbial community assembly in response to warming. Nat. Commun. 11, 4717. Noguchi, H., Park, J., Takagi, T., 2006. MetaGene: prokaryotic gene finding from environmental genome shotgun sequences. Nucleic Acids Res 34 (19), 5623–5630. Olesen, J.M., Bascompte, J., Dupont, Y.L., Jordano, P., 2007. The modularity of pollination networks. Proc. Natl. Acad. Sci. 104 (50), 19891–19896. Sun, Y., Zhang, M., Duan, C., Cao, N., Jia, W., Zhao, Z., Ding, C., Huang, Y., Wang, J., 2021. Contribution of stochastic processes to the microbial community assembly on field-collected microplastics. Environ. Microbiol. 23 (11), 6707–6720. Tibbetts, J., Krause, S., Lynch, I., Sambrook Smith, G., 2018. Abundance, distribution, and drivers of microplastic contamination in urban river environments. Water (Basel) 10 (11), 1597. Wagner, M., Scherer, C., Alvarez-Muoz, D., Brennholt, N., Bourrain, X., Buchinger, S., Fries, E., Grosbois, C., Klasmeier, J., Marti, T., Rodriguez-Mozaz, S., Urbatzka, R., Vethaak, A.D., Winther-Nielsen, M., Reifferscheid, G., 2014. Microplastics in freshwater ecosystems: what we know and what we need to know. Environ. Sci. Eur. 26, 12. Wang, J., Qin, X., Guo, J., Jia, W., Wang, Q., Zhang, M., Huang, Y., 2020. Evidence of selective enrichment of bacterial assemblages and antibiotic resistant genes by microplastics in urban rivers. Water Res. 183, 116113. World Health Organization. WHO report on surveillance of antibiotic consumption: 2016-2018 early implementation. 2018. Wright, S.L., Kelly, F.J., 2017. Plastic and human health: a micro issue? Environ. Sci. Technol. 51 (12), 6634–6647. Wright, R.J., Langille, M.G.I., Walker, T.R., 2021. Food or just a free ride? A metaanalysis reveals the global diversity of the plastisphere. ISME J. 15 (3), 789–806. Wu, L., Ning, D., Zhang, B., Li, Y., Zhang, P., Shan, X., et al., 2019a. Global diversity and biogeography of bacterial communities in wastewater treatment plants. Nat. Microbiol. 4 (7), 1183–1195 a. Wu, X., Pan, J., Li, M., Li, Y., Bartlam, M., Wang, Y., 2019b. Selective enrichment of bacterial pathogens by microplastic biofilm. Water Res. 165 (4), 114979. Xu, C., Lu, J., Shen, C., Wang, J., Li, F., 2022. Deciphering the mechanisms shaping the plastisphere antibiotic resistome on riverine microplastics. Water Res. 225, 119192. Xun, W., Li, W., Xiong, W., Ren, Y., Liu, Y., Miao, Y., Xu, Z., Zhang, N., Shen, Q., Zhang, R., 2019. Diversity-triggered deterministic bacterial assembly constrains community functions. Nat. Commun. 10, 3833. 9