Setup of the marine system
A microplastic incubation experiment was performed at the aquaculture facility of the Center of Excellence in Marine Science (CEMarin) at the Justus Liebig University, Giessen, Germany. Three independent 80-liter tanks (biological replicates) were filled with artificial seawater (ATI-Aquaristik, Coral Ocean plus, Premium Quality Reef Salt; containing in mg L− 1, Ca2+: 410, Mg2+: 1230, PO43− <0.03, salinity of 34). Alkalinity was maintained at 2.52 mmol L− 1 by the addition of NaHCO3. Tanks were equipped with two pumps to generate horizontal currents and 300 W heaters that maintained the water temperature at 26°C using a feedback controlled regulation (Profilux 3, GHL Advanced Technology GmbH & Co. KG, Germany), and a 10:14 light:dark photoperiod using T5 tubes (Aquablue Special, ATI-Aquaristik, Germany). Water parameters were kept constant during the experiment (twelve weeks). All experimental tanks were connected to a 4,000 L tropical seawater system that harbor marine biota, which emulated a realistic coral reef environment. The experimental tanks had an exchange rate with the main system of 120 L water day− 1 (equivalent to 150% of the tank volume). Coral fragments of the species Acropora muricata, Pocillopora verrucosa, Porites lutea, and Heliopora coerulea were placed randomly in the tanks. The system was acclimatized over two weeks before addition of surface-sterilized MP and sediment particles. The plastic polymer was black low-density polyethylene (LD-PE) powder, a self-adhesive thermoplastic powder used for indoor and outdoor coatings, as corrosion protection, electrostatic and heavy-duty coatings (Novoplastik, Hockenheim, Germany). MP sizes varied between 37 and 163 µm with a mean diameter of 112.7 ± 11.1 µm (mean ± SD) and a density of 0.95 g cm− 3. Their planar surface area ranged from 819 to 32,487 µm2 with a median of 4,477 µm2. The irregularly shaped particles exhibit a rough surface structure with a specific surface area of 0.0204 m2 g− 1, determined in a Mastersizer 2000 (Malvern, Worcestershire, UK) (Reichert et al. 2018). MP and sandy sediment particles were surface-sterilized for 24 hours in 70% (v/v) ethanol and rinsed with filtered (0.22 µm pore-size Sterivex-GP filtration units, Merck Millipore, Burlington, Massachusetts, USA) and autoclaved ambient water (collected from the tanks). The sterility of the particles was tested by incubation of some particles on marine agar which was incubated for 4 days at 25°C. No bacterial growth was obtained. A concentration of 2.5 mg of MP particles were added per 1 L water (200 mg per tank); this resulted in approximately 10% of particles floating in the water column (= 0.25 mg L− 1) and a count of 200 particles L− 1. In parallel, 90 g of sandy sediments were added to each tank. The concentration of MP was controlled weekly by manual counting under a stereo microscope and further MP were added when the amount was below the initial amount. Pooled MP were studied instead of single particles to minimize the bias, which may have caused by younger biofilms of the few particles added later. The concentration of bacterial cells in the water column was monitored by SYBR Green I (SG-I) staining as described previously (Glaeser et al. 2010) using a DM5000 B epifluorescence microscope with a DFC3000 G camera system, and the LAS X software (Leica Microsystems, Wetzlar, Germany) for cell counting. The concertation of bacterial cells was in the range of 105 to 106 cells mL− 1 tank water, which is typical for aquatic marine environments. An overview of the experimental system is given in Fig. S1.
Sample collection
After twelve weeks of incubation, samples were randomly collected for each particle type from four locations per tank (intra-tank replicates) from three independent tanks (independent biological replicates). MP particles distributed in the water column were collected with a sterile micropipette and sterile pipette tips. Sediments and detritus particles were collected from random areas of the tank bottom using sterile glass pipettes. Samples for molecular analyses were placed in sterile 2 mL tubes containing 1 mL of 0.22 µm-filer sterilized and autoclaved artificial seawater (ASW) and stored at -20°C until DNA extraction. Two liters of water per tank were collected from the water column (20 cm below water surface) using 60 mL syringes and filtered immediately. Water samples were serially filtered through (i) 5 µm Minisart syringe filters (Sartorius, Göttingen, Germany), to collect particle-attached (PA) bacterial communities and (ii) through sterile 0.22 µm Sterivex-GP filter units (Millipore) to collect the free-living (FL) bacterioplankton. Remaining water was completely removed from filter units, which were stored immediately at -20°C until DNA extraction.
Scanning electron microscopy (SEM) of different particle types
Collected MP, sediment and detritus particles were washed with sterile ASW and pre-fixed for 1 hour at 4°C in a solution of 0.1 M sodium cacodylate buffer (CB, pH 7.4) containing 8% sucrose, 1.5% paraformaldehyde, and 2.5% glutaraldehyde. Samples were rinsed in CB buffer at 4°C, and immersed in the fixative overnight at 4°C. After several washes in buffer, samples were incubated in 2% OsO4, washed again in buffer, and subsequently in ddH2O. Samples were dehydrated in increasing ethanol series [30, 50, 70, 80, 90, 96, 100% (v/v)] (20 minutes each) after osmium fixation, critical point dried, mounted on SEM holders and gold sputtered. Samples were analyzed using a Zeiss DSM982 field emission scanning electron microscope (FESEM; Carl Zeiss AG, Oberkochen, Germany) at 3–5 kV. Images were taken using a secondary electron-detector with the voltage of the collector grid biased to + 300 V in order to improve the signal-to-noise ratio and to reveal optimal topographical contrast. For element analysis, energy-dispersive X-ray microanalysis was done at 15 kV acceleration voltage using a 10 mm2 Si(Li) detector (Oxford Instruments, Abingdon, UK).
DNA extraction from particle and water samples
For molecular analyses of particle- and water-associated bacterial communities, total community DNA was extracted from particles and filter material. DNA extraction was performed according to Bižic-Ionescu et al. (2015), with few modifications. Particles (20–30 MP, 10 sediment, or detritus particles) were directly added to 1 ml lysis buffer (100 mM Tris-HCl, 50 mM EDTA, 100 mM NaCl, and 1% SDS at pH 8.0). Filters were removed from cartridges of Minisart and Sterivex-GP filter units after being pre-frozen in liquid nitrogen and broken using a hammer. Filters were cut with a sterile scalpel in three pieces, which were used as intra-tank replicates. For filter pieces, a mixture of 0.1 and 0.5 mm zirconia/silica beads (0.3 g each, Carl Roth, Karlsruhe, Germany) was added to the lysis buffer prior first incubation and tubes were bead beaten twice for 1 minute in 30 seconds intervals. All samples were incubated for 30 min at 100°C in lysis buffer followed by a 15 min incubation at 65°C after addition of phenol:chloroform:isolamyl alcohol 25:24:1 (1 mL). Samples were centrifuged for 15 min at 4°C and aqueous phases were transferred into new tubes. An identical volume of chloroform:isoamyl alcohol 24:1 was added, samples were mixed and centrifuged again for 15 min at 4°C; the aqueous phase was again transferred into a new tube. DNA was precipitated for 3 hours at room temperature (RT) after addition of ¼ volume 7.5 M filter sterilized ammonium acetate and 1 volume of 99% isopropanol and centrifugation at 17,000 g at 4°C for 40 min. DNA pellets were washed twice in ice-cold 70% (v/v) ethanol (10 min centrifugation), tubes were drained upside down, and dried in a Speed Vac at RT for 5 min. DNA pellets were dissolved in 100 µL molecular grade water (Carl Roth). To ensure the complete dissolution of the DNA, samples were incubated for 10 min at 37°C. The DNA solution was stored at -20°C for further analyses. Due to the presence of co-extracted inhibitory compounds, undiluted, 1:10, 1:20, 1:50, and 1:100 diluted DNA extracts were tested as PCR template. Finally, 1:10 dilutions were selected for further analyses.
Microbial community analyses
Bacterial community analyses were performed by 16S rRNA gene sequence-based bacterial community fingerprinting using PCR-based denaturing gradient gel electrophoresis (PCR-DGGE) and 16S rRNA gene amplicon based Illumina MiSeq sequencing.
PCR-DGGE was performed with Bacteria 16S rRNA gene targeting primers according to Schellenberg et al. (2020). Illumina amplicon sequencing was performed with the Bacteria 16S rRNA gene targeting primer system 341F (5´-CCT ACG GGN GGC WGC AG-3´) and 785R (5´-GAC TAC HVG GGT ATC TAA KCC-3´) (Klindworth et al. 2013). PCRs, product quantification and purification, and Illumina 300 bp paired-end read sequencing using an Illumina MiSeq V3 system was performed by LGC Genomics (Berlin, Germany) and analyzed using the automated Silva NGS pipeline (https://www.arb-silva.de/ngs/; Yilmaz et al. 2014) as described by Aydogan et al. (2016). Due to the limited phylogenetic resolution of the 16S rRNA gene below the genus level, the analysis was performed based on a stable clustering obtained by the phylogenetic calculations performed in SILVAngs. OTUs were assigned to phylogenetic groups, which had a genus-level resolution. Since several of the determined phylogenetic groups had no cultured representatives, phylogenetic groups were used instead of genera all over the text. If sequences were assigned to a described genus it was named accordingly.
Amplicon sequence data were deposit in the sequence read archive (SRA) of NCBI as SRA project with accession number SRP194562 (experiments SRX5781870 to SRX5781884) assigned to the BioProject PRJNA540740 and BioSample SAMN11554495, respectively.
Quantification of Vibrio sp. 16S rRNA gene targets in total community DNA samples
Vibrio sp. 16S rRNA gene copies per ng total community DNA were quantified with two primer systems: 567F (5’-GGC GTA AAG CGC ATG CAG GT-3’)/680R (5’-GAA ATT CTA CCC CCC TCT ACA G-3’) (Thompson et al. 2004) and Vibrio-744F (5’-CAG ATA CTG ACA CTC AGA TG-3’)/Vibrio-849R (5’-CGG CTC AAG GCC ACA ACC T-3’) (this study). In parallel 16 rRNA gene copies of total Bacteria were quantified using a universal 16S rRNA gene targeting primer system (Universal-F 5’-GTG STG CAY GGY TGT CGT CA-3’/ Universal-R 5’-ACG TCR TCC MCA CCT TCC TC-3’; Maeda et al. 2003) to determine the ratio of Vibrio to total bacterial 16S rRNA gene copies. Quantitative PCRs (qPCRs) were performed in a total volume of 10 µL using the Sso Fast EVA Green Supermix and a CFX96 Real-Time System (both Bio-Rad, Feldkirchen, Germany). All intra-tank replicates were analyzed separately in technical replicates. Detailed information of primer design, qPCR conditions, standards used for quantification, cloning and sequencing of qPCR-products are available in the Supplementary Information.
Statistical analyses of bacterial community profiles
Statistical analyses of bacterial community patterns (relative DNA band abundance in DGGE band patterns and relative sequence abundance in amplicon data) were performed in PAST version 3.11 (Hammer et al. 2001) if not indicated otherwise. Non-metric multidimensional scaling (NMDS) plots based on a Bray-Curtis similarity index were used to display differences of bacterial communities between the relative abundance patterns of DNA bands (DGGE) or phylogenetic groups (amplicon sequencing) of bacterial assemblages present on particles and in water samples. Due to the low number of sample replicates, pairwise multivariate analysis of variance (PERMANOVA) was performed combined with the Monte Carlo correction to improve the accuracy of p-values. Analysis was performed in PRIMER 7 with PERMANOVA+ (https://www.primer-e.com) and based on 999 permutations and the sums of squares type: type III (partial). One-way ANOSIM and PERMANOVA analyses were based on a Bray-Curtis similarity matrix. The alpha-diversity of bacterial assemblages was analyzed with Chao 1, Shannon, evenness, and dominance indices. A ternary plot was calculated to illustrate the occurrence of abundant taxa (relative abundance ≥ 1.0%) in the different bacterial assemblages. SigmaPLOT v12.5 (Systat Software, Erkrath, Germany) was used to generate BOX-plots and to determine significant differences among samples by one-way analysis of variance (ANOVA) using the Tukey’s pairwise multiple comparison test. (Post Hoc test), the Shapiro Wilk normality test, and the Brown Forsythe equal variance test. If the normality distribution failed an ANOVA on Ranks analysis was performed.