For in-situ incubation, research-grade polymer foils made of low-density polyethylene (PE), polyethylene terephthalate (PET) or polystyrene (PS) were used (Goodfellow, Hamburg, DE). Foils were cut into 4 mm × 4 mm × 0.125 mm square snippets using an ethanol-sterilized wire binder (Pavo Sales B.V., Oss, NL) as described previously (Leiser et al., 2021). As containers for the incubation, stainless steel tea strainers (5 cm diameter, Contacto Bander GmbH, Erkrath, DE) were heat-sterilized (200°C, 4 h) and subsequently filled with 100 polymer snippets of a single polymer type. Incubation of the polymer snippets lasted 5 weeks from September 12^th to October 17^th, 2018, at three different sites: Ems river (51°57′06.7″ N, 7°59′56.4″ E), Lake Emssee (51°57′11.9″ N, 8°00′08.9″ E) and the Rieselfelder, an interconnected system of shallow reservoirs which were inundated by a waste water treatment plant effluent (52°01′23.4″ N, 7°39′33.8″ E). The containers were mounted to foamed polystyrene lifting bodies to a depth of about 30 cm in the respective water columns (Supplementary Figure S1). The incubations were weekly sampled by detaching one container per polymer type for measuring chlorophyll fluorescence and biofilm biomass on the snippets (see below). At each sampling site, a sterilized brush was used to remove biomass adhering on the outside of the residual containments for enabling continued sunlight penetration. For transport to the laboratory, the containments were stored at 4°C in heat-sterilized glass beakers. Prior to further processing, snippets were washed three times using sterile phosphate buffered saline (PBS, pH 7.4). After 5 weeks the in-situ incubation was stopped and a fraction of the snippets was used for isolating bacteria (see below). To determine the microbial community of the biofilms, additional 15 particles per plastic material and site were pooled and stored at −78°C until further processing for DNA isolation (see below).

Biofilm-biomass quantification was performed using crystal-violet staining with protocol modifications adapted from Christensen et al. (1985), Stepanović et al. (2000, 2007) and Arias-Andres et al. (2018). After washing three times with PBS, snippets were transferred to a 24-well microplate with one particle per well. After biofilm fixation at 60°C for 1 h, 0.5 to 1 ml crystal violet (0.3% w/v) per particle was added (until they were completely submerged) with subsequent incubation on a rocker shaker for 15 min at room temperature. After removal of the crystal violet solution, particles were washed four times with H₂O_Millipore. Residual supernatants were removed and 1 ml 33% (v/v) acetic acid was added. After 20–25 min incubation at 120 rpm, 900 μl of supernatants were transferred into new 24-well microplates for measuring absorption at 595 nm with covered lid using a Tecan® GENios™ microplate reader (Tecan Group AG, Männedorf, CH).

Chlorophyll autofluorescence was measured with the ChemiDoc™ imager (Bio Rad Laboratories, Hercules, USA). As excitation light source, Green Epi Fluorescence with a wavelength of approximately 550 nm was used (Zecher et al., 2015). Fluorescence emission was detected using the 695/55 filter. Exposure times were manually adjusted to avoid signal overmodulation. Fluorescence intensities were calculated using Image Lab™ Version 4.1 (Bio-Rad Laboratories) and normalized to area and exposure time.

For analysing microbial communities analysis of biofilms from in-situ incubations and from enriched photoautotrophic consortia, DNA was extracted with the DNA Power Soil Pro Kit (Qiagen, Hilden, DE). Biofilm-covered plastic snippets were transferred into PowerBead Pro Tubes containing 800 μl solution CD1 and shaken horizontally in a vortex adapter for 1 h. After addition of 25 μl proteinase K (22 mg/ml) and incubation at 37°C for 1 h the extraction was continued according to the manufacturer’s instructions. Library preparation, sequencing and data analysis were performed by Microsynth AG (Belgach, Switzerland). Extracted DNA was submitted to two-step PCR amplification of the V4-V5 region of the bacterial 16S rRNA gene, using the primer pair 515F-Y and 926R (Parada et al., 2016). PCR-products were sequenced using a v2 500 cycle kit on the Illumina MiSeq platform. Raw data was submitted to the European Nucleotide Archive (ENA) database and were assigned the Project ID PRJEB45856. For determination of relative abundances of bacterial phyla we performed standard statistical analysis and bioinformatics including the program R (R Core Team, 2018).

For isolation of bacteria from in-situ grown biofilm, 15 washed snippets of each polymer type were pooled in 2 ml microcentrifuge tubes with 10 sterile glass beads (2.7 mm diameter, Carl Roth GmbH + Co. KG, Karlsruhe, DE). After adding 1 ml of sterile PBS, the tube was vortexed at low speed for 30 s. The resulting supernatant was diluted and plated on solid medium B (Jagmann et al., 2010) supplemented with 0.1% (v/v) 7-vitamins solution (Pfennig, 1978), 0.01 mM ATP (Bruns et al., 2003) and triple concentrated mixed carbon sources (Cho and Giovannoni, 2004). From these plates, colonies with different morphologies were selected for purifying bacterial strains by repeated transfers on YPG-agar medium [0.075% (w/v) yeast extract, 0.05% (w/v) peptone, 0.075% (w/v) d-glucose, 1.2% (w/v) Bacto™ Agar (Becton Dickinson GmbH, Heidelberg, DE); adapted from Bruns et al. (2003)]. All cultivation steps were performed at room temperature.

For enrichment and cultivation of photoautotrophic algal-bacterial consortia from in-situ grown PE biofilms, modified Diatom Medium (DM; pH 6.75/HCl; adapted from Cohn and Pickett-Heaps, 1988 and Cohn et al., 2003) was used in which soil extract, FeSO₄ and MnCl₂ was replaced by 0.1% (v/v) f/2 trace element solution (Guillard and Ryther, 1962). After autoclaving, 0.1 mM NaHCO₃ and DM vitamin solution were added to the medium. For enrichment cultures, 15 washed PE snippets were pooled in a 2 ml microcentrifuge tube with 10 sterile glass beads (see above) in 1 ml of sterile PBS and vortexed at low speed for 30 s (Supplementary Figure S2). After removal of the supernatant, the snippets were washed three times in PBS. Single snippets were then used to inoculate 1 ml of DM in a 24-well microplate containing a sterile pristine PE snippet. These microtiter plates were incubated for 12 days at 21°C with 180 rpm in a light incubator (Phytobiochamber Model EGCS 701, EQUiTEC; light source: Lumilux Warmwhite, Osram, DE; photon flux density: approx. 90 μE m⁻² s⁻¹; light/dark cycle: 14: 10 h). For transferring the enrichment cultures, 20 μl of the suspended fraction were used to inoculate 980 μl DM containing two pristine PE snippets in a microtiter plate. After 12 days of incubation, when the pristine PE snippets were covered with a chlorophyll-containing biofilm, this transfer procedure was repeated. After a total of three transfers, six biofilm-covered polymer particles were streaked onto solid DM and cultivated in the light incubator. The resulting photoautotrophic algal-bacterial consortia were re-plated onto new solid DM every 1 to 2 weeks.

For isolation of heterotrophic bacteria from algal-bacterial consortia, cell material from the consortia was streaked onto solid medium B supplemented with 2 mM of each d-glucose, N-acetyl-d-glucosamine and sodium glycolate (MB3G). Plates were incubated in the dark. For isolation of photoautotrophic microalgae from consortia, cell material was streaked onto solid DM and transferred with weekly changing antibiotics [in chronological order: 260 μg/ml disodium carbenicillin and 60 μg/ml monosodium ampicillin, 50 μg/ml streptomycin sulfate, 10 μg/ml gentamycin sulfate, 50 μg/ml kanamycin sulfate; adapted from Cottrell and Suttle (1993)]. Plates were incubated in the light as described above for the enrichment cultures. Axenity of the algal isolates was tested by 4′,6-diamidino-2-phenylindole (DAPI) staining and by cultivation on MB3G in the dark. The resulting axenic algal isolates were re-plated onto new solid DM with 260 μg/ml disodium carbenicillin and 60 μg/ml monosodium ampicillin every 2 to 4 weeks for maintenance.

For taxonomical classification of isolated bacterial strains, DNA was isolated with the Gentra Puregene Yeast/Bact. Kit (Qiagen, Hilden, DE) according to the manufacturer’s instructions. Purified genomic DNA was amplified using the primer pair 16S_27_fw (5’-AGAGTTTGATCATGGCTCA-3′) and 16S_1492_rev (5’-TACGGTTACCTTGTTACGACTT-3′, adapted from Weisburg et al., 1991). PCR-amplified DNA was purified with the GeneJET PCR Purification Kit (Thermo Fisher Scientific, Waltham, USA) and sequenced by Eurofins Genomics (Ebersberg, DE) with the Mix2Seq Kit.

For taxonomical classification of isolated algal strains, DNA isolation was performed according to the protocol described by Jagielski et al. (2017). For cell lysis, glass beads (d: 400–600 μm; Sigma Aldrich, St. Louis, USA), Mikro-Dismembrator S (Sartorius AG, Göttingen, DE), 300 μl of 5 M NaCl and 240 μl CTAB buffer were used. After adding the phenol/chloroform/isoamylalcohol solution, additional 10 s vortexing and subsequent centrifuging at 16,699 × g for 1 min were implemented.

After DNA precipitation, centrifugation was performed with 18,407 × g. Purified DNA was PCR-amplified using the primer pairs CHLORO_fw (5′-TGGCCTATCTTGTTGGTCTGC-3′)/CHLORO_rev (5′-GAATCAACCTGACAAGGCAAC-3′; Gumbi et al., 2017), ITS1_fw (5′-AGGAGAAGTCGTAACAAGGT-3′)/ITS4_rev (5′-TCCTCCGCTTATTGATATGC-3′; Hadi et al., 2016), rbcL_192_fw (5′-GGTACTTGGACAACWGTWTGGAC-3′)/rbcL_657_rev (5′-GAAACGGTCTCKCCARCGCAT-3′; Hadi et al., 2016), rbcL_375_fw (5′-TTTGGTTTCAAAGCIYTWCGTGC-3′)/rbcL_1089_rev (5′-ATACCACGRCTACGRTCTTT-3′; Hadi et al., 2016), tufA_fw (5′-TGAAACAGAAMAWCGTCATTATGC-3′)/tufA_rev (5′-CCTTCNCGAATMGCRAAWCGC-3′; Hall et al., 2010), tufA_50_fw (5′-TGGATGGTGCTATTYTAGTTG-3′)/tufA_870_rv (5′-ATAGTGTCRCCTGGCATAGC-3′; Hall et al., 2010). PCR-amplified DNA was purified and sequenced as described above.

For phylogenetic affiliation analysis, the BLASTn suite search tool (Altschul et al., 1990) was used.

Growth experiments with photoautotrophic consortia were performed for re-colonization experiments and for grazing experiments. In both cases, pre-cultures were set up by resuspending cell material from consortia growing on solid DM in liquid DM and incubated for 7 days.

For the re-colonization experiments, the photoautotrophic consortia Co_1 to Co_6 were used. Pre-cultures were inoculated from agar plates in 20 ml DM in 100 ml Erlenmeyer flasks at 175 rpm and 21°C in the light incubator (EQUiTEC). For main cultures, 800 μl DM were inoculated with 200 μl pre-culture in 24-well plates (Thermo Scientific™ Nunc™ Cell-Culture Treated Multidishes [Thermo Fisher Scientific Inc., Waltham, USA]) containing PE, PET or PS snippets, which had prior been sterilized by bathing in in ethanol (70% v/v) for 15 min followed by drying and 5 min UV-irradiation in a laminar-flow sterile bench before use. After 7 days incubation, chlorophyll fluorescence and biofilm biomass on the snippets were determined as described above.

For grazing experiments, consortium, Co_3 was used. Pre-cultures were grown in 125 ml DM in 500 ml Erlenmeyer flasks without shaking at room temperature and daylight. As biofilm surfaces, slides of 2.5 cm × 8 cm were cut from PE-, PET- and PS foils (Goodfellow, thickness: 0.125 mm), which were roughened on one side using 60 grid sandpaper. Before use, polymer slides were sterilized in 70% (v/v) ethanol for 30 min, washed in sterile H₂O_demin and stored in sterile H₂O_demin at 4°C. As control material, fully frosted glass microscope slides were used which had prior been rinsed with H₂O_demin and autoclaved. Before inoculation of main cultures, pre-cultures were washed via centrifugation of 50 ml at 6,000 rpm for 30 min. Supernatants were carefully discarded, and pellets were resuspended in sterile DM-medium to a volume with a chlorophyll concentration of 1 μg/ml using the corresponding chlorophyll fluorescence intensities determined for the diatom Phaeodactylum tricornutum as described by Zecher et al. (2015).

For the main cultures, in which biofilm formation should occur, four square petri dishes per material type were filled with 50 ml of the washed cell suspension. In each petri dish, four polymer or glass slides were placed with the rough side facing upwards. The petri dishes were incubated for 7 days at room temperature and daylight. Biofilm formation was evaluated via chlorophyll fluorescence as described above. For analysing shifts in the bacterial community compositions via amplicon sequencing, cell material was isolated from Co_3 on solid DM media from three transfers (January 2020, August 2020, and March 2021), 7-day-old biofilms on different plastic types and from the respective surrounding supernatants before grazing started. DNA-extraction and 16S rDNA amplicon sequencing were performed as described above. To confirm that the isolated algae Alg_3.1 was originally present in Co_3 its 18S rRNA sequence was aligned against the R2 mate samples of the amplicon sequencing data of Co_3 from the three transfers using bowtie2 using the “sensitive-local” algorithm (version 2.5.1, https://github.com/BenLangmead/bowtie2).

Plastic and glass slides with biofilms of Co_3, which were produced in the growth experiments described above, were placed in 1,000 ml glass beakers with 500 ml synthetically reconstituted surface freshwater (Osterauer et al., 2010) and 4 individuals of the gastropod Physa fontinalis. Snails obtained from a commercial distributor for aquarist supplies, who kept the snails in quarantine had a length of 3.8 ± 0.5 mm (mean ± standard deviation). The slides were replaced every 3 ½ days with slides from the same batch stored at 4°C in the dark. Experiments lasted for 3 weeks in a climate-controlled room at 20°C with a 16:8 h cycle. The physiological parameters of P. fontinalis were determined in weekly intervals, including: size measured by a digital caliper as well as a digital microscope (VHX-5000, Keyence Corp.), mortality, faeces dry mass and the numbers of eggs and egg packages. Biofilms were analysed by measuring chlorophyll fluorescence before and after the grazing experiments.

The microplastic snippets of all three materials (PE, PET, PS) were colonized with chlorophyll-containing biofilms over time on all three incubation sites (Figure 1). The chlorophyll intensities were generally lower with the Emssee and Rieselfelder samples compared to the Ems samples. Microscopic examination revealed the presence of microalgae (various diatoms and Chlorella- and Chlamydomonas-like morphotypes) and of prokaryotic cells (not shown).

From all three incubation sites, strains of heterotrophic bacteria could be isolated by direct plating of sheared biofilm material (Supplementary Table S1). Most isolated strains belonged to the Alpha-, Beta- and Gammaproteobacteria including genera of characteristic freshwater bacteria such as Gemmobacter, Mitsuaria, and Aeromonas, respectively. Typical freshwater Bacteroidetes, such as Flavobacterium spp., could also be isolated.

The cultivation-independent analysis of the biofilm communities exhibited a much greater diversity than the cultivation analysis (Figure 2). While Alpha-, Beta- and Gammaproteobacteria and Bacteroidetes were also found to be abundant, the amplicon sequencing revealed a large proportion of Planctomycetia that were not retrieved by cultivation (Figure 2B). Furthermore, the Rieselfelder samples contained a large percentage of members of the genus Nitrospira, which were not detected in the biofilms from the other incubation sites. Statistical analysis of the cultivation-independent analysis showed that the bacterial communities clustered according to the incubation sites rather than the plastic material, on which the biofilm had formed (Figure 2A).

For isolating microorganisms from these in-situ enrichments that could be used for producing defined biofilms for grazing experiments we set up laboratory enrichments under photoautotrophic conditions in the next step.

These laboratory enrichments were inoculated with plastic snippets from which loosely attached microorganisms had been detached by gentle shearing forces (Supplementary Figure S2). In the following, we selected for microorganisms that were able to colonize pristine plastic snippets under photoautotrophic conditions. We restricted the enrichment to PE snippets because biofilms on this material showed the highest biofilm-biomass in the in-situ incubations (Figure 1). This procedure led to the enrichment of communities that reproducibly re-colonized PE-surfaces. After 3 transfers, plastic snippets with biofilms were used to inoculate solid media. By this procedure, we obtained photoautotrophic microbial communities that could be maintained on agar plates by regular transfer about every fortnight since November 2018. In total, we enriched six consortia: two originating from Rieselfelder (Co_1 and Co_2) and four from Ems river PE-particles (Co_3 to Co_6; Figure 3A). In comparison to in-situ grown biofilms the consortia had originated from, the bacterial communities within the consortia displayed a smaller diversity (Figure 3B). The phylum of Planctomycetes, which exhibited large proportions in the in-situ grown biofilms (Figure 2B) was only represented in Co_6 and Co_2 at low abundances. The phylum of Nitrospira was not present in the algal-bacterial consortia derived from the Rieselfelder while it was abundant in the in-situ sample. Cyanobacteria, represented in comparably small abundances within in-situ grown biofilms in the Ems river, dominated the algal-bacterial consortia Co_4 and Co_5, derived from PE biofilms grown in the Ems river. In Co_3 and Co_1, we also observed the yet uncultivated candidate phylum WPS-2. The presence of microalgae in the consortia was verified by microscopy (not shown).

For investigating whether the consortia were able to re-colonize plastic surfaces from liquid culture after being maintained on agar plates and for identifying the consortium with the highest biofilm production, re-colonization experiments were performed with all three plastic polymers (PE, PET and PS) originally used in the in situ incubation (Figure 4). The consortia showed large differences in biofilm formation when exposed to plastic surfaces. Co_3, which was derived from a biofilm on PE snippets from the Ems river, showed the highest chlorophyll fluorescence values and highest biofilm biomass. While all consortia showed at least some biofilm biomass, there was no chlorophyll fluorescence detected in Co_1 and 5, and also only low values in other consortia.

For further characterizing the composition and stability of Co_3, we repeatedly submitted it to amplicon sequencing. These analyses showed a reduction of the diversity on class and genus levels with time. The candidate genus WPS-2 disappeared within about 8 months while members of genera with cultivated representatives remained (Figure 5A).

As Co_3 should contain microorganisms capable of re-colonizing a plastic surface from a culture with suspended cells, we isolated algal and bacterial strains from this consortium. For this, cell material from the consortia plates were transferred to solid media and growth conditions that favor the growth of axenic microalgae and heterotrophic bacteria, respectively. The procedure for obtaining axenic microalgae led to the isolation of strain Alg_3.1, which was classified as Chlamydomonas sp. Alignment of the 18S rRNA of the isolate to the amplicon sequencing data from Co_3 confirmed that Alg_3.1 was present in the original Co_3 consortium and in the transfers on solid DM medium (alignment rates of 4.40, 2.78, and 6.42% in January 2020, August 2020, and March 2021, respectively). The procedure for obtaining heterotrophic bacteria, which was also applied to the other consortia, led to the isolation of several strains of Alphaproteobacteria of the genus Gemmobacter and Betaproteobacteria of the genus Acidovorax (Supplementary Table S2). Apart from further proteobacterial strains, a member of the genus Flectobacillus (Bacteroidetes) was isolated. Pre-test with a number of these bacterial strains in co-culture with Chlamydomonas sp. strain Alg_3.1 revealed biofilm formation of several isolates in co-culture with Chlamydomonas sp. Alg_3.1 on PET (not shown). For further colonization experiments, Gemmobacter sp. strain O was then chosen because it showed reliable biofilm formation on different plastic surfaces. Biofilm formation and algal growth could be largely reduced when co-cultures were supplied with succinate as growth substrate for the bacterium indicating that biofilm formation was optimal when the co-culture relied on photoautotrophic conditions (Supplementary Figure S3). However, these dual-species biofilms were not grazed by our model invertebrate Physa fontinalis (not shown).

Due to the stability of the community composition together with the recolonization capacity biofilms of Co_3 were tested as a substrate for the grazer P. fontinalis. For this, we set up biofilm experiments with Co_3 by offering three types of plastic (PE, PET, PS) and glass as surfaces as control. These experiments showed that Co_3 reproducibly formed biofilms for the test period of 8 weeks (Supplementary Figure S4). During this time, P. fontinalis showed growth and activity for a period of 3 weeks (Table 1). Snails grazed on the biofilms leading to small but continuous growth rates, reproduction and faeces excretion. There were no significant differences in snail performance among the different substrates.

For further assessing the suitability of using Co_3 as model community we analyzed the community composition directly before the snails were added. Statistical analysis of the communities within the biofilm and the suspended fraction by amplicon sequencing revealed that the composition clustered according to the individual suspended and biofilm fractions rather than according to surface material (Figures 5A,B) as was also previously observed in the in-situ incubations. A heat-map analysis of the community structure indicated that the largest difference to the stock consortium of Co_3 on solid medium was the strong reduction of Sediminibacterium sp. and Hydrogenophaga sp.in the biofilm and the suspended fraction as well as a strong increase of Rhodobacter sp. in all suspended fractions (Figure 5C).