Ocean acidification has a strong effect on communities living on plastic in mesocosms

We conducted a mesocosm experiment to examine how ocean acidification (OA) affects communities of prokaryotes and eukaryotes growing on single‐use drinking bottles in subtropical eutrophic waters of the East China Sea. Based on 16S rDNA gene sequencing, simulated high CO2 significantly altered the prokaryotic community, with the relative abundance of the phylum Planctomycetota increasing by 49%. Under high CO2, prokaryotes in the plastisphere had enhanced nitrogen dissimilation and ureolysis, raising the possibility that OA may modify nutrient cycling in subtropical eutrophic waters. The relative abundance of pathogenic and animal parasite bacteria also increased under simulated high CO2. Our results show that elevated CO2 levels significantly affected several animal taxa based on 18S rDNA gene sequencing. For example, Mayorella amoebae were highly resistant, whereas Labyrinthula were sensitive to OA. Thus, OA may alter plastisphere food chains in subtropical eutrophic waters.

may modify nutrient cycling in subtropical eutrophic waters. The relative abundance of pathogenic and animal parasite bacteria also increased under simulated high CO 2 . Our results show that elevated CO 2 levels significantly affected several animal taxa based on 18S rDNA gene sequencing. For example, Mayorella amoebae were highly resistant, whereas Labyrinthula were sensitive to OA. Thus, OA may alter plastisphere food chains in subtropical eutrophic waters.
Plastic pollution has become one of the most concerning environmental issues due to its persistence in the environment. Van Sebille et al. (2015) estimated that around 236,000 metric tons of plastics are dispersed throughout the ocean. These plastics often entangle or are ingested by marine animals of all sizes, from microbes to whales. Microplastic fragments can also be transferred through food webs along with chemicals and pathogens (Napper and Thompson 2020).
Plastics provide substrata for microbial attachment and are now a novel habitat for a variety of organisms. The term "plastisphere" was first used by Zettler et al. (2013) to describe an ecosystem of microbes living on floating plastic with communities that were distinct from the surrounding water. Plastics can carry organisms horizontally and vertically from the sea surface to the seafloor (Amaral- Zettler et al. 2020). Plastic types, environmental conditions, and biogeography all play significant roles in the development and diversity of the plastisphere (Amaral- Zettler et al. 2015;Toyofuku et al. 2016;Napper and Thompson 2020;Seeley et al. 2020). Core members of the plastisphere vary depending on the type of plastic as well as location (Dudek et al. 2020;Wu et al. 2020;Du et al. 2022). Nitrogen cycling can be affected by the plastisphere in sediments (Seeley et al. 2020) and different types of plastic may also influence biogeochemical cycles (Romera-Castillo et al. 2018;Sanz-L azaro et al. 2021;Wang et al. 2021;Shen et al. 2022). Yet few have considered the effects of anthropogenic change (such as warming, acidification and eutrophication) on plastisphere composition and function.
Considering the high durability of plastics in the ocean, it is prudent to investigate the effects of future projected environmental conditions on plastisphere community composition and function. One of the most widespread anthropogenic changes on Earth is ocean acidification (OA) which happens as CO 2 from the atmosphere dissolves in seawater to produce carbonic acid, which lowers surface ocean pH. Since the beginning of the Industrial Revolution, surface ocean H + ion concentration has increased by 26% due to the increasing atmospheric CO 2. Under a high CO 2 emissions scenario (SSP3-7.0), average surface seawater pH is projected to drop another 0.3 units during this century (IPCC, 2021). There is a severe lack of knowledge about how the plastisphere will be affected by OA as only two studies have investigated this; both used drinking bottles placed along natural gradients in CO 2 at marine volcanic seeps in Japan (Harvey et al. 2020;Kerfahi et al. 2022). In these studies, OA greatly altered the composition of bacterial assemblages on plastic and decreased the prevalence of genes associated with cell-to-cell interactions and antibiotic resistance.
In this study, a mesocosm platform was used to investigate both eukaryotic and prokaryotic community composition and function on polyethylene terephthalate (PET) drinking bottles in subtropical coastal seawater in Wuyuan Bay, East China Sea. Seawater acidification is more severe in this bay than in the open ocean because nutrient inputs have caused eutrophication and stimulated microbial degradation of organic matter, lowering seawater pH (Cai et al. 2011;Wallace et al. 2014). Mesocosm experiments bridge the gap between laboratory experiments and CO 2 seep studies by simulating effects on pelagic ecosystems in miniature. We sequenced 16S rDNA and 18S rDNA to explore the effects of OA on the prokaryotic and eukaryotic community composition and predicted functions of the plastisphere.

Mesocosm setup
We used the Facility for the Study of Ocean Acidification Impacts of Xiamen University (24 31 0 48 00 N, 118 10 0 47 00 E), starting 09 October 2019, for 32 d (Fig. 1). Nine cylindrical transparent thermoplastic polyurethane mesocosm bags, each 3 m deep Â 1.5 m diameter, were fixed in steel frames and covered by cone lids. In situ seawater around the platform was filtered (pore size of 0.01 μm, MU801-4T, Midea, China) with prefiltration by 5-mm nylon net and then pumped simultaneously into the bags, filling them with 3000 liters within 36 h. Bags 1, 3, 5, 7, and 9 were acidified with CO 2saturated seawater to adjust seawater to 1000 μatm CO 2 as a high CO 2 (HC) treatment corresponding to the year 2100 under a high Greenhouse Gas emissions scenario. Bags 2, 4, 6, and 8 were controls (ambient CO 2 , AC). AC and HC bags were aerated with ambient air of $ 410 ppmv CO 2 and premixed air-CO 2 of 1000 ppmv CO 2 (5 L min À1 ), respectively. Three plastic bottles per mesocosm bag were attached inside the mesocosms (Fig. 1). Each bag was inoculated with 80 liters of in situ seawater containing a natural microbe community filtered by a 180-μm mesh for the main investigation of the effects of OA on the plankton community in the mesocosm bags as described by Huang et al. (2021). Seawater was collected from 0.5 m depth from each bag at 10 a.m. every 1-3 d to measure pHtotal using an Environmental Water Analyzer (iSEA, Ma et al. 2018) and total alkalinity measured using an Automated Spectrophotometric Analyzer (Li et al. 2013).

DNA extraction, amplification, and sequencing
After 32 d, the plastic bottles were removed. For the following analysis, two plastic bottles were chosen randomly from a pool of three bottles for each bag. Bottles were scratched for DNA extraction. A 70 C preheated lysis buffer (100 mM Tris, 40 mM ethylene diamine teraacetic acid, 100 mM NaCl, 1% sodiumdodecyl sulfate was used to extract DNA, followed by phenol-chloroform extraction and ethanol precipitation. DNA from samples was used to amplify the 16S V4-V5 and 18S V9 regions. The16S V4-V5 region was amplified using the primers 515AF (GTGYCAGCMGCCGCGGTAA) and 926R (CCGYCAATTYMTTTRAGTTT; Parada et al. 2016), while the 18S V9 region was amplified using the primer 1389F (TTGTACACACCGCCC) and 1510R (CCTTCYGCAGGTT CACCTAC; Amaral- Zettler et al. 2009). The amplification conditions were: initial denaturation at 95 C for 3 min, 29 cycles of denaturation for 16S V4-V5 and 30 cycles for 18S V9 at 95 C for 30s, annealing at 53 C for 30 s extension at 72 C for 45 s, and final extension at 72 C for 10 min. Polymerase chain reaction (PCR) products were purified using an AxyPrepDNA gel extraction kit (Axygen) from 2% agarose gel after electrophoresis. A DNA library was constructed following the MiSeq Reagent Kit guide (Illumina). The sequencing was conducted using an Illumina MiSeq PE300 platform (Majorbio Bio-pharm Technology Co. Ltd.) after the purification and quantification of PCR products.

Sequence assignment and data analysis
Our sequencing data have been uploaded to NCBI (project ID: PRJNA895187). Raw fastq sequences were adapters removed and quality filtered by fastp (v0.19.6) and merged by FLASH (v1.2.7) before analysis. The filtered reads were imported into QIIME 2, and DADA2 was used to de-noise sequences, resulting in high-resolution amplicon sequence variants (ASVs). A number of 1,130,309 sequences for 18S V9 and 952,158 sequences for 16S V4-V5 were obtained. For taxonomic classification, we used the SILVA 138 database and the TARA 18S V9 database (http://taraoceans.sb-roscoff.fr/EukDiv/ index.html) for 16S V4-V5 and 18S V9 sequencing data, respectively. All samples were standardized by random subsampling using the "sub.sample" command in Mothur. The prokaryotic and eukaryotic sequences were rarified to 21,400 and 41,865 reads per sample, respectively. Alpha diversity was estimated using Mothur 1.30. Beta diversity was analyzed with QIIME 2. Alpha diversity describes the species diversity within a community. Beta diversity describes the species diversity between communities. Beta diversities were calculated by Bray-Curtis matrixes and visualized by non-metric multidimensional scaling (NMDS) analysis. We performed analysis of similarities (ANOSIM) for the significant difference test. We performed linear discriminant analysis effect size (LEfSe) analysis to identify taxa that were differentially abundant in different samples. The default setting of the Linear Discriminant Analysis (LDA) score was set to 2.0 and p < 0.05. Functional Annotation of Prokaryotic Taxa (FAPROTAX) was used to predict the functional profile of bacterial communities (Louca et al. 2016).

Statistical analyses
Differences in pH total value between HC and AC at different time points were tested by a one-way ANOVA test in SPSS. We performed the non-parametric Kruskal-Wallis rank tests to detect statistical differences in alpha diversity between treatments. The Wilcoxon rank-sum test was used to test for significant differences in relative abundance of taxa and predicted functions of prokaryotic community between treatments. Statistical significance was determined at p < 0.05.

Seawater in mesocosms
The in situ pHtotal in Wuyan Bay was 7.78, which is representative of eutrophic coastal seawater in the region. The pH total varied in the AC and HC treatments in response to a growing then declining phytoplankton population. Under HC, pH decreased from 7.70 AE 0.02 on Day 0 (after CO 2saturated water addition) to 7.60 AE 0.02 on Day 2, subsequently increased to a peak of 8.06 AE 0.05 under HC on Day 10. The pH value under AC decreased from 7.77 AE 0.01 on Day 0 to 7.67 AE 0.01 on Day 2, then peaked on Day 8 at 8.28 AE 0.12. The pH values under HC and AC then gradually declined (7.67 AE 0.05 under HC and 7.82 AE 0.06 under AC). From Day 0 to Day 2, an increase in bacteria concentration was responsible for the decrease in pH under HC and AC. As intended, the pH total of the HC treatments was always lower than the AC treatment ( p < 0.05; Fig. 2). On Day 0, the concentration of NO À 3 þ NO À 2 was 29.23 AE 1.21 μmol L À1 under HC and 26.05 AE 6.48 μmol L À1 under AC; NH þ 4 was 11.58 AE 1.47 μmol L À1 under HC and 12.62 AE 1.23 μmol L À1 under AC; and PO À 4 was 1.29 AE 0.11 μmol L À1 under HC and 1.24 AE 0.24 μmol L À1 under AC. The seawater was potentially moderately phosphorus-limited eutrophic (Guo et al. 1998).

Discussion
In just 32 d, 25 prokaryotic Phyla and 26 eukaryotic Phyla had colonized on single-use plastic drinking bottles, demonstrating that plastics can quickly support highly complex communities with a proliferation of protists and metazoan groups that are seldom studied (Du et al. 2022).
Our results demonstrate that the prokaryotic community was more significantly affected by HC compared to the eukaryotic community (Fig. 3). In our study, Planctomycetota which have unusual features, such as intracellular compartmentalization and a lack of peptidoglycan in their protein cell wall (Fuerst and Sagulenko 2011), had significantly higher relative abundance under HC (Fig. 4). It would be interesting to investigate whether the response of Planctomycetota to HC is consistent in other locations. At the late stage of the phytoplankton bloom in mesocosm bags, dinoflagellate concentrations were higher under HC (Wang et al. in prep.) which may result in varied dissolved organic matter in quality and quantity. This might explain the different prokaryotic compositions of the plastisphere under different CO 2 treatments (Buchan et al. 2014;Xu et al. 2022).
The dominant primary producers were Stramenopiles (mainly diatoms) and there were comparatively few Cyanobacteria in the plastisphere. Diatoms are widely found in the plastisphere in the photic zone due to their ability to secrete sticky exopolysaccharides (Amaral-Zettler et al. 2020). In the mesocosm bags, the high concentration of diatoms easily attached to plastic surfaces. Studies along CO 2 seep gradients in Italy and Japan have shown that elevated CO 2 significantly increases diatom abundance on plastic surfaces (Johnson et al. 2013;Kerfahi et al. 2022). However, no plastisphere primary producers were significantly affected by our CO 2 treatments in our study (Fig. 4), perhaps due to lower light intensity although a combination of many factors may be the cause such as differences in phytoplankton biogeography. In future, amplicon sequencing combined with microscope observation should be used to better understand the effects of OA on the plastisphere, because amplicon sequencing does not always distinguish statistically significant differences between the communities (Dudek et al. 2020).
Plastics are providing a novel and globally multiplying habitat for marine animals. We found a high proportion of Metazoa and Ciliphora in the plastisphere community ( Supplementary Fig. 1). Some secondary producers with relatively low abundance were clearly affected by increases in CO 2 , indicating that the organisms in higher trophic levels are sensitive to high CO 2 (Fig. 4). For example, Mayorella proliferated in simulated high CO 2 whereas Labyrinthula were more prevalent in ambient CO 2 conditions. Mayorella can engulf and digest microalgae (Laybourn-Parry et al. 1987). The Labyrinthulids feed saprotrophically, especially on marine algae (Finlay and Esteban 2019). Such changes may affect food chain dynamics.
In our study, stimulated high CO 2 significantly increased the relative abundance of bacterial taxa involved in nitrogen and nitrate respiration, as well as ureolysis (Fig. 5). Thus, OA may promote denitrification and organic nitrogen utilization in the plastisphere in subtropical eutrophic seawater. Until now, very few studies on the effects of climate change on nitrogen cycling in the plastisphere have been reported (Shen et al. 2022). Our findings are consistent with a recent study conducted in eutrophic seawater in Xiamen, which illustrated that the plastisphere had higher denitrifying activity and N 2 O production compared to the surrounding seawater (Su et al. 2022).
Perhaps of most concern was our observation of enrichment of human pathogenic microorganisms and animal parasites in the plastisphere under simulated high CO 2 . Plastic pollution is a transport vehicle that may accelerate the spread of infectious diseases (Zettler et al. 2013;Laverty et al. 2020;Du et al. 2022). Mariculture usually lowers seawater pH due to intense respiration (Gao et al. 2022) and produces plastic wastes from lost culture gear and packaging so that some seafood growing areas have a severe plastic pollution problem (Chen et al. 2018;Feng et al. 2020). The intersection of global change, plastic pollution and altered microbial communities is a major unstudied risk in mariculture. The higher relative abundances of the functions of prokaryotic communities under simulated high CO 2 in this study were based on FAPROTAX predictions, not direct measurements. We recommend that metagenomics and metatranscriptomics be used to identify modifications to plastisphere community functions induced by global change.
In summary, this first study of the effects of simulated high CO 2 on the plastisphere in eutrophic coastal conditions showed significant biological changes. To better prepare for the interactive effects of climate change and plastic pollution, investigations should target different types of plastics, regions, and seasons, especially in coastal areas that are relied upon to provide human food.