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. Under high CO₂, prokaryotes in plastisphere significantly enhanced nitrogen dissimilation and ureolysis, raising the possibility that elevated CO₂ 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 simulated high CO₂ significantly affected some secondary producers based on 18S rDNA gene sequencing. For example, Mayorella amoebae were highly resistant to OA whereas labyrinthulids were sensitive to high CO₂. This shows that OA may alter plastisphere food chains in subtropical eutrophic waters.

Mesocosm setup

We used the Facility for the Study of Ocean Acidification Impacts of Xiamen University (24◦ 31’48” N, 118◦10’47” E), starting October 9th, 2019, for 32 days (Figure 1). Nine cylindrical transparent thermoplastic polyurethane mesocosm bags, each 3 m deep x 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 mm, MU801-4T, Midea, China) with pre-filtration by 5mm nylon net and then pumped simultaneously into the bags, filling them with 3000 L within 36 hours. Bags 1, 3, 5, 7, and 9 were acidified with CO₂-saturated seawater to adjust seawater to 1000 matm CO₂ as a high CO₂ treatment (HC) corresponding to the year 2100 under a high Greenhouse Gas emissions scenario. Bags 2, 4, 6, and 8 were controls (ambient CO₂, AC). AC and HC bags were aerated with ambient air of ∼410 ppmv CO₂ and premixed air-CO₂ of 1000 ppmv CO₂ (5L min⁻¹), respectively. Three plastic bottles per mesocosm bag were attached inside the mesocosms (Figure 1). Each bag was inoculated with 80 L of in situ seawater containing a natural microbe community filtered by a 180 mm mesh for the main investigation of the effects of OA on the plankton community in the mesocosm bags as described in (Huang et al. 2021). Seawater was collected from 0.5 m depth from each bag at 10 a.m. every 1–3 days 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). Samples were filtered through a 0.45 mm cellulose acetate membrane for NO₃⁻, NO₂⁻ , NH₄⁺ and PO₄³⁻ measurements by an auto-analyzer (AA3, Seal, Germany).

DNA extraction, amplification, and sequencing

After 32 days, 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 (100mM Tris, 40mM EDTA, 100mM NaCl, 1% SDS) was used to extract DNA, followed by phenol-chloroform extraction and ethanol precipitation. DNA from samples was used to amplify the 16S V4-V5 region and 18S V9 region. 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 (CCTTCYGCAGGTTCACCTAC)(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. PCR products were purified using an AxyPrepDNA gel extraction kit (Axygen, United States) from 2% agarose gel after electrophoresis. A DNA library was constructed following the MiSeq Reagent Kit guide (Illumina, United States). The sequencing was conducted using an Illumina MiSeq PE300 platform (Majorbio Bio-pharm Technology Co. Ltd., Shanghai, China) 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). 1130309 sequences for 18S V9 and 952158 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 21400 and 41865 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 ANOSIM (analysis of similarities) 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 LDA score was set to 2.0 and P<0.05. FAPROTAX (Functional Annotation of Prokaryotic Taxa) was used to predict the functional profile of bacterial communities (Louca et al. 2016). 

Statistical analyses 

Differences in pHtotal value between HC (High CO₂) and AC (Ambient CO₂) 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.