2 h1652019)14g79 et aL.2015:Martinez et al.2015:Van Boeckel et al.2015)via 45 min in 100%ethanol).The samples were then dried by Polarior mobile genetic (MGE)such as plasmids ith gold Given that some such as et al.2016:K i et al. the SEM. ARGs from environmental b cteria and travel with microplasti 2.3.Flow cytometry(FCM)measurement of path a pote indic ehre2aa”Yn6 part were ic tre (B So used to nick et al..20 eckmann et al 2014 ice included a gene rator So onopuls D3200 catio Here,we investigated the diffe nces be (ro the ed with microp astics and stained by SYBR Green I(10000×di d the features of the microbial ity on microplast ences.Us for 10m e signal excited d by the blue laser a tance genes v the freshwater ecosystem n.red fluo ence 2.Materials and method green after data we ed usin 2.1.Experimental design and water quality parameters described previously (Wen) 2.4.Sampling and DNA extraction ent in Northern China,flowing through ere imme ed in 100 mL of inteecop 30s o detached article -5)wa 14 China kit (M Bio Laborato plan Cark bad. USA)The ktonic meshes to e ugh ed ce membra s and pla d at 20C befor action. on,agaros redat-90C0rheoncentr particle type tothe replicates of th 2.5.165 rRNA gene sequencing and data processing May 2nd.2018). eeks(from April 18th to (Su 2.Scanning Electron Microscope(SEM)imaging GCAG-3) and 806R (5'-GGAC The formati of biofilms de sequence we PCR ollowed by uL dNTP.10 uM et al., 2015; Martinez et al., 2015; Van Boeckel et al., 2015) via mobile genetic elements (MGEs) such as plasmids, transposons, bacteriophages, insertion sequences and integrons (Stokes and Gillings, 2011). Given that some opportunistic pathogens, such as Vibrio spp. have been discovered in microplastic biofilm (Foulon et al., 2016; Keswani et al., 2016; Kirstein et al., 2016), it is possible that specific pathogens in microplastic biofilm will acquire ARGs from environmental bacteria and travel with microplastics to reach remote environments. The resistance to antibiotics of path￾ogens harbouring ARGs make them hard to be killed by therapeu￾tics, which poses a potential worldwide threat to ecosystem and human health. Previous studies have indicated that the biofilm communities on plastic substrates were distinctive from those in water columns or sediment (Amaral-Zettler et al., 2015; De Tender et al., 2015; McCormick et al., 2014; Oberbeckmann et al., 2014; Zettler et al., 2013). However, the link between the structure and function of the biofilm communities on microplastic is still not fully understood. Here, we investigated the differences between biofilms on microplastics and two natural substrates (rock and leaf) by comparing the microbial community structure, ARG profiles and ARG bacterial hosts of the biofilm associated with microplastics and two natural substrates. This approach allowed us to better under￾stand the features of the microbial community on microplastics, and to provide insight into the possibility of various surfaces, both anthropogenic and naturally occurring, to spread antibiotic resis￾tance genes via biofilms in the freshwater ecosystem. 2. Materials and methods 2.1. Experimental design and water quality parameters In order to test the effects of substrate type (anthropogenic or natural) on the associated biofilms, we used river water to culture the biofilm in bioreactor (BioFlo CelliGen 115, New Brunswick, Eppendorf, USA). The river water was collected in the Haihe River, the largest river catchment in Northern China, flowing through several cities and finally into the sea. Polyvinyl chloride (PVC) microplastic pellets (density 1.35e1.45 g cm
3 , ø 3 mm) were purchased from Aladdin Biochemical Technology Co. Ltd. (Shanghai, China). Rock (quartz) was purchased from a flower shop and leaves (Platanus acerifolia) were cut into small pieces. Rocks and leaves were sieved with stainless steel laboratory grade meshes to ensure they were within the size range of 2e4 mm. All sieved particle (microplastic, rock, and leaf) were rinsed with deionized water three times and placed in the dark until they were dried at room temperature. Prior to the experiment, the bioreactor was rinsed with deionized water three times and sterilized by autoclaving. River water was continuously pumped into the bioreactor. All treated particles (microplastic, rock, and leaf) were wrapped with sterilized gauze and each type was divided into 5 independent aggregates. The 5 aggregates or each particle type correspond to the 5 replicates of the 16S rRNA gene amplicon sequencing in the following analysis. All 15 aggregates (n ¼ 5 replicates 3 types of particles) were incubated in a biore￾actor with 5 L of working volume for 2 weeks (from April 18th to May 2nd, 2018). 2.2. Scanning Electron Microscope (SEM) imaging The formation of biofilms on different substrates was investi￾gated after seven days using a field-emission scanning microscope (JEOL JSM 7800, Japan). The samples were rinsed with PBS buffer and post-fixed with 2% osmium tetroxide. Dehydrated by graded ethanol series (15 min each in 35%, 50%, 75%, 90%, followed by 45 min in 100% ethanol). The samples were then dried by Polarion E3000 Critical Point Dryer overnight. Sputter was coated with gold layer at 25 mA under Argon (Ar) atmosphere at 0.3 MPa, the sam￾ples were transferred to the conductive carbon tape mounted on the sample holder, and the morphology was characterized under the SEM. 2.3. Flow cytometry (FCM) measurement In brief, 1 g particles (microplastic, rock and leaf) were sampled and rinsed with sterile PBS. The particles were immersed in 10 mL of sterilized PBS buffer and ultrasonic treatment (B Sonopuls HD 3200, Bandelin Sonorex, Rangendingen, Germany) was used to detach the bacteria associated with the particles. The ultrasonic device included a generator (Sonopuls HD3200), ultrasonication energy transfer unit (UW 2200), booster horn (SH 213G), and needle (MS72). The settings used were: amplitude: 302 mm; cycle duration: 30 s; pulse level: 50%; power: 50%. Flow cytometry analysis was used to determine the biomass of the biofilm and the planktonic bacteria concentration every two days, based on previously described methods (Wen et al., 2015). 1 mL of sample was stained by SYBR Green I (10000 diluted, Invitrogen). Flow cytometry analysis was performed using a BD Accuri C6 Plus instrument (BD Biosciences, USA). After being mixed thoroughly with vortex and incubated in the dark for 10 min at 37 C, the emitting fluorescence signal excited by the blue laser at 488 nm was selected on FITC-PerCP tunnel (flow rate: 66 mL/min, green fluorescence tunnel: 533 nm, red fluorescence tunnel: >670 nm) and the total cell concentration (TCC) could be measured after data were processed using BD Accuri C6 Plus software as described previously (Wen et al., 2015). 2.4. Sampling and DNA extraction On day 14, the same type of particles (microplastic, wood and rock) were recovered and rinsed three times with sterilized PBS. Particles were immersed in 100 mL of sterilized PBS and treated by 30 s of ultrasonication. The detached biofilm from microplastic, rock, leaf particles (particle-associated part fraction, n ¼ 5) was collected by centrifugation (14,000 g, 10 min) and DNA was extracted using the Mobio PowerBiofilm® DNA isolation kit (Mo Bio Laboratories, Carlsbad, CA, USA). The planktonic bacteria in river water (planktonic part fraction, n ¼ 5) were collected by filtration through a sterilized mixed cellulose esters membrane with a pore size of 0.1 mm membrane (Millipore, USA). The mem￾branes were stored at
20 C before DNA extraction. DNA was isolated using the Mobio PowerWater® DNA isolation kit (Mo Bio Laboratories, Carlsbad, CA, USA). The extraction processes followed the manufacturer's instructions. Following the extraction, agarose gel electrophoresis (2.0%) and a Qubit 2.0 Fluorometer (Invitrogen) were used to check the concentration of DNA samples, which were stored at
80 C for further study. 2.5. 16S rRNA gene sequencing and data processing Two-step PCR was conducted to amplify the 16S rRNA gene (Sutton et al., 2013). With this approach, tags and adapters were added in a second round of PCR amplification. To amplify the V3eV4 hypervariable regions of 16S rRNA gene, the primer set 338F (50 -ACTCCTACGGGAGGCAGCAG-30 ) and 806R (50 -GGAC￾TACHVGGGTWTCTAAT-30 ) combined with adapter sequences and barcode sequences were used (Chu et al., 2015). Triplicate PCR re￾actions were performed in 50 mL reaction mixtures, which con￾tained 10 mL GoTaq buffer, 0.2 mL Q5 High-Fidelity DNA Polymerase, 10 mL High GC Enhancer, 1 mL dNTP, 10 mM of each primer, 60 ng 2 X. Wu et al. / Water Research 165 (2019) 114979