Collection of marine organisms
Marine invertebrates such as Corallium rubrum are ideal organisms to perform controlled experiments and to gather useful information on a variety of environmental conditions74. This species, whose diet is based on small zooplankton captured with the polyp tentacles, has been already used in long-term experiments74,75,76. Coral specimens were collected in March 2017 at ca. 35-m depth in the Marine Protected Area of Portofino (Punta del Faro, 44°17′41.02 N; 9°13′31.30 E) in the Ligurian Sea (North-Western Mediterranean Sea) by scuba divers (using TRIMIX blending).
Experimental design
After recovery, the coral specimens were brought to the laboratory and maintained in a tank (30 L) at in situ temperature (13 ± 0.8 °C) and subjected to the continuous flux of natural seawater filtered onto 0.7-µm pore-size membranes (micro-glass fibre paper, Munktell) by using two submersible pumps (Euronatale, 203 V, 50 Hz, 4 Watt).
Sixty coral branches obtained from different colonies, with similar morphology, and a surface of ~2 cm2 each, were distributed among 12 experimental tanks, in order to have 5 coral branches per tank (12 L glass tanks, containing on average, 274 ± 26.4 coral polyps each). The corals were acclimatised for 20 days in a temperature-controlled room, and dim light conditions, before starting experiments. Each tank, filled with natural seawater, was equipped with a prefiltered (0.2 µm) channelled aeration system combined with motor-driven paddles in order to create convective currents, which allowed the resuspension of the microplastic mixture, thus ensuring as much as possible a homogeneous distribution of the polymers. This experimental system was designed and set up according to Sutherland et al.77. To assess the potential effects of increasing microplastic microparticles L−1 (here defined as low, medium and high concentrations of microplastic particles). We also quantified the exact amount of particles actually interacting with the corals, by discounting the fractions loss due to experimental manipulations (see details in Supplementary Methods). According to the results reported in the Supplementary Results, the systems were responsible for the loss of ca. 40% of the microplastic particles, thus the corals in experimental systems were actually exposed to 60, 300 and 600 microplastic particles per litre (to which we referred the low, medium and high concentrations).
The highest concentrations of microplastics (up to 600 microplastic particles per litre) can reflect future contamination on the basis of estimates obtained by numerical models9, whereas the low and medium concentrations have been selected to represent highly-contaminated marine habitats, including the areas where the corals were collected (Ligurian Sea)78,79. In particular, for the Ligurian Sea, we estimated an average value of 94 microplastic particles L−1, based on the concentrations of microplastic particles (>200 µm) determined by Fossi et al.78,79, and the most cautionary correction factor (105) calculated by Brandon et al.10 for the unaccounted smaller fraction of microplastics (25–75% of the fragments falling approximately in the 20–100 µm dimensional class with median range: 59–116).
Microplastic mixtures were also prepared considering the concentration and composition of dominant polymers in different coastal marine environments, especially in hot spots of microplastic contamination5,7,8.
The microplastic mixture added to the tanks was composed of 76.6% polyethylene, 10.9% polypropylene, 7.3% polystyrene, 3.3% polyvinylchloride and 1.8% polyethylene terephthalate particles. These particles were obtained by milling plastic objects from everyday life (i.e., containers, bottles, cups, pipes) according to Paul-Pont et al.8 (Supplementary Table 5). Plastic milling was carried out under a laminar flow hood in chilled sterilized and 0.02 µm prefiltered milliQ water. All the tools used for handling plastics were pre-treated with 1% sodium hypochlorite in water and rinsed 10 times with sterilized and 0.02 µm prefiltered milliQ water, and then dried under laminar flow hood. Details on the preparation of microplastic mixtures are reported in the Supplementary Methods.
The low, medium and high concentrations of microplastics were added in triplicate tanks (n = 3 for each concentration). Additional systems containing seawater and coral branches without microplastics (n = 3, here defined Controls), and seawater added with microplastics (at the highest concentration) without red corals (n = 3, here define CTRL MPs) were used as controls. Overall, the experimental setup comprised 15 tanks.
The experiments for assessing the impact of microplastics on red corals started immediately after the microplastic mixture addition (time 0). During the experiment, seawater temperature (range: 13.10 ± 0.01–13.13 ± 0.05 °C), salinity (range: 38.35 ± 0.18–38.65 ± 0.18) and oxygen levels (7.10 ± 0.08–7.36 ± 0.2 mg L−1) were monitored daily in all tanks using a probe (YSI Professional Plus, USA) and corals were fed three times a week with 103 Artemia salina nauplii L−1.
After ten days the condition of corals that were exposed to microplastics was deteriorating, so we collected one coral branch from each tank for molecular analyses (i.e., associated microbiome, gene expression and DNA damage). After 14 days, the experiment was stopped because the coral branches that were exposed to the medium and high microplastic particle concentrations were completely wrapped in mucus, with a large portion of damaged tissue and without polyp activity, therefore corals were defined dead (overall 12 branches, see ‘Results’ section for details). Coral branches in the controls showed no visible signs of necrosis or other macroscopic stress. The tissue remained intact, and the colour unchanged until the end of the experiment.
Effects of microplastic ingestion
Coral feeding activity
To assess the impact of microplastics on feeding activity of C. rubrum, analyses based on the use of Artemia salina were performed after 2 and 10 days from the start of the experiment (t0) in replicate systems (n = 3 for each treatment, n = 3 for the controls) according to standard international protocols80. The nauplii of Artemia salina were reared in the laboratory, incubating 0.5 g of cysts (Ocean Nutrition) in 1 L of seawater filtered onto 0.2-µm filter in a separatory funnel, 2 days before the analysis of feeding rate. At hatching, nauplii were counted and maintained in vials to obtain the concentration of 1000 nauplii L−1. To avoid stress, corals (one branch for each tank) were transferred underwater to beakers along with 1 L seawater of each tank. After addition of live A. salina nauplii (1000 nauplii L−1) to the 1 L beakers containing the coral branches and to the controls, three aliquots of 10 ml seawater were collected after ~30 s from the start of the experiment (t0) and after 2 and 4 h. The remaining nauplii present in each seawater aliquot were counted under a stereomicroscope at ×3.2 magnification (Zeiss Stami 2000). Mean ingestion rates (nauplii removed h−1) were determined by linear regression analysis.
Accumulation of microplastics by C. rubrum
To investigate the accumulation of plastic polymers by C. rubrum polyps, the number of microplastic particles ingested by coral polyps was evaluated after 14 days of exposure to microplastic mixture, by dissolving polyps and skeleton of the corals (one for each tank at the concentration of 1000 microplastic particles L−1) using an acid/base digestion protocol36 with some modifications.
To exclude biases on the estimate of the number of microplastic particles actually accumulated within the polyps, coral branches were accurately rinsed with milliQ water and checked under stereomicroscope (at ×50 magnification) for the potential presence of microplastic particles adherent to the coral tissue. Coral branches were then soaked in 5 ml of 4.5% sodium hypochlorite (NaClO) for 24 h and dissolved in 5 ml of 37% HCl for 30 min. Particulate material was retained on a 0.2-μm filter in a vacuum filtration system, and microplastic particles were counted under a stereomicroscope at ×50 magnification. The chemical composition of the polymers ingested by corals was confirmed by FT-IR analyses (Perkin Elmer, software Packages Spectrum 5.3.1). To evaluate possible damage to plastic polymers due to the use of acid/base solutions, we exposed polypropylene, polyethylene, polystyrene, polyvinylchloride and polyethylene terephthalate at the same volume and concentration of NaClO and HCl during digestion of the coral.
Potential transfer of microplastics by zooplankton
While testing the exposure of the red corals to microplastics, we also determined the rate of microplastic ingestion by A. salina nauplii used to feed the red corals, in order to assess their role as potential vectors of microplastics. To do this, additional tanks (n = 3) were added with 0.2 µm prefiltered 12 L natural seawater, 1000 nauplii L−1 of A. salina and the same microplastic mixture used for the experiment on the red corals (at the highest concentration). Three other tanks were used as controls containing 0.2 µm prefiltered 12 L natural seawater and 1000 nauplii L−1 of A. salina.
Microplastic ingestion by A. salina was determined after 2 and 10 days of experiment following the enzymatic digestion protocol previously developed81 with some modifications. Such a procedure degrades biological tissues without affecting shape, colour and composition of plastic fragments. Gut contents of 100 individuals of A. salina (n = 5) were assessed under a stereomicroscope (Leica MZ125) and light microscope (Zeiss Axiovert 200) and photographed with a Zeiss Axiocam digital camera. Afterwards, nauplii were processed immediately according to the modified enzymatic digestion protocol. Nauplii were dried in an oven for 3 h at 60 °C, transferred to glass jars containing a buffer homogenizing solution (400 mM Tris-HCl pH 8, 60 mM EDTA pH 8, 5 M NaCl, SDS 1%) incubated at 50 °C for 15 min and exposed to Proteinase K (1 mg ml−1). Then, samples were dried for 2 h at 50 °C, homogenized and re-incubated at 60 °C for 20 min and sonicated on ice (1–2 min) three times. After digestion, the microplastic-containing suspensions were placed in Utermöhl chambers and the microplastics were examined at the inverted light microscope (Leica DMI3000-Bat ×200 magnification) and counted. Microplastics obtained from nauplii digested after 10 days of incubation were also measured and categorized by colours and shape to evaluate the numbers and the size spectra of microplastics ingested by A. salina during the experiment.
Physical impact on coral coenenchyma
Scanning electron microscopy (SEM) analyses
To investigate the physical damage of the microplastic mixture on the coral tissues, samples (one branch from each tank including the control) were collected before the start of the experiment (t0), after 7 days and at the end of the experiment and prepared for SEM analyses according to standard protocols82 with some modifications. Coral branches were stored in 0.7 µm prefiltered seawater with 4% buffered formalin. After 24 h, samples were washed with 0.7 µm prefiltered seawater and dehydrated for 3 h in 20% ethanol. After 3 h they were washed in the same way and dehydrated in ethanol 50%. After 3 h, samples were stored in 70% ethanol. Samples were stored at +4 °C. We dehydrated samples using different gradients of ethanol solutions (70–80%, 80–90%, 90–95%, 95–99% in 2 days)82. Then, samples were dried using HMDS (Hexamethyldisilazane, Aldrich 440191)83. Dried samples were mounted on aluminium stubs using Leit-C glue (conductive carbon cement, Neubauer Chemikalien) and sputter-coated with gold. Samples were examined with a Scanning Electron Microscope (Zeiss SUPRA 40). In addition, the tissue damage percentage was assessed on SEM micrographs at ×200 of magnification by using PhotoQuad v1.4 software84. Such a software for advanced image processing of 2D photographic quadrat samples, dedicated to ecological applications, was used for the analysis of three randomly selected areas from the apex to the base of each coral rotating it on three sides (n = 9). Additional analyses through random SEM observations (n = 20) at 3.00KX to 17.00KX of magnification were carried out to determine prokaryotic cell abundances around lesions of corals (n = 3) exposed to high concentrations of microplastic particles. Data were standardised to the coral surface analysed.
Mucus release and trapped microplastics and prokaryotic cells
To evaluate the first symptoms of coral stress, a photographic report was conducted daily. The abundance of microplastic particles trapped in coral mucus was estimated using an enzymatic digestion protocol81 with some modifications. Mucus produced by corals exposed to higher microplastics concentrations was dried in oven at 60 °C for 12 h. After 12 h, five ml of homogenizing solution was added to the samples and incubated at 50 °C for 15 min. Proteinase K (1 mg mL−1) was added to the samples, which subsequently were incubated at 50 °C for 2 h. Then, samples were homogenized and incubated again at 60 °C for 20 min, after that samples were sonicated three times (three 1-min treatments using a Branson Sonifier 2200; 60 W). After digestion, microplastics-containing suspension was filtered on 0.2-μm filters in a vacuum filtration system (Whatman, Nuclepore). Filters were analysed at stereomicroscope at ×50 magnification (Zeiss Stemi 2000).
Stress signals at the molecular level
RNA extraction, cDNA synthesis and gene expression level by qPCR
To assess potential changes in the gene expression pattern of C. rubrum due to microplastics, total RNA was extracted from ca. 20 mg of tissue (wet weight) from one coral branch randomly collected from each treatment (n = 3) and control (n = 3) after 10 days of experiment by using Quick-RNA™ MiniPrep (Zymo Research, Freiburg, Germany) according to the manufacturer’s instructions. Total RNA was also extracted from additional samples of coral branches collected randomly at the beginning of the experiment. Once scraped by surgical disposable scalpels (Braun), coral tissues were placed in new 2 ml sterile tubes and washed three times with phosphate-buffered saline (PBS 1×). Samples were centrifuged at 1800 rpm for 10 min in an Eppendorf® 5810r refrigerated centrifuge using a swing-out rotor at 4 °C and, after removing the supernatant, were homogenized for 5 min with a RNase-free sterile glass stick in RNA lysis buffer. Contaminating DNA was degraded by treating each sample with DNase dissolved in RNase-free water included in the kit. For each sample, 250 ng of total RNA extracted was retrotranscribed with an iScript™ cDNA Synthesis kit (Bio-Rad, Milan, Italy), following the manufacturer’s instructions. The reaction was performed on the Veriti™ 96-Well Thermal Cycler (Applied Biosystem, Monza, Italy). To evaluate the efficiency of cDNA synthesis, a PCR was performed with primers of the reference gene, cytochrome oxidase I (COI, Supplementary Table 6). The reaction was carried out using MyTaq™ HS DNA Polymerase (Bioline, Luckenwalde, Germany) on the Veriti™ 96-Well Thermal Cycler (Applied Biosystem, Monza, Italy). The PCR programme consisted of a denaturation step at 95 °C for 1 min, 35 cycles at 95 °C for 45 s, 60 °C for 45 s, and 72 °C for 45 s and a final extension step at 72 °C for 10 min.
The expression levels of the six genes of hsp70, hsp60, MnSOD, mtMutS, EF1 and cytb, involved in a broad range of functional responses, such as stress, detoxification processes, and DNA repair, were followed by real-time qPCR to identify potential stress of corals exposed to microplastics61. For the cytb, target-specific primer pairs were designed with the Primer 3 software (http://primer3.ut.ee85) using nucleotide sequences retrieved from the GenBank database for C. rubrum as template (https://www.ncbi.nlm.nih.gov/genbank/; Supplementary Table 6). SensiFAST™ SYBR® & Fluorescein mix (Bioline, Luckenwalde, Germany) were used for measuring the levels of mRNAs on CFX Connect™ Real-Time PCR detection system (Biorad, Milan, Italy). Fluorescence was measured using CFX Manager™ software (Biorad, Milan, Italy). All genes tested by qPCR in this study were amplified with primers purchased from Life Technologies/Thermo Fisher Scientific (Milan, Italy). The fold change in target gene mRNA expression of corals exposed to microplastics compared with the control was calculated using the comparative CT method using the 2−ΔΔCt equation86. COI was used as reference gene for normalising the gene expression analyses.
DNA oxidative damage
For evaluating oxidative DNA damage potentially due to microplastic exposure on C. rubrum, the content of 8-hydroxydeoxyguanosine (8-OHdG) was analysed. DNA was extracted from 20 mg (wet weight) of tissue randomly collected from one coral branch for each treatment (n = 3) and control (n = 3) after 10 days of experiment using DNeasy Blood & Tissue Kits (Qiagen, Valencia, CA) and following the manufacturer’s protocol. Finally, samples were kept at −20 °C before subsequent analyses. Nucleic acids extracted (2 μg) were transferred into new 2-ml tubes and incubated for 5 min at 95 °C, then rapidly chilled on ice. Samples were digested to nucleosides by incubating the denatured DNA in sodium acetate 20 mM, pH 5.2 with 2 μl of nuclease P1 (6 U/μl; Merck KGaA, Darmstadt, Germany) for 2 h at 37 °C. Each sample was then incubated with 5 μl alkaline phosphatase (1 U/μl; Roche, Mannheim, Germany) in Tris-HCl 100 mM, pH 7.5 for 1 h at 37 °C. The reaction mixtures were then centrifuged for 5 min at 6000 × g and the supernatants tested for DNA oxidation with an OxiSelect™ Oxidative DNA Damage ELISA Kit (8-OHdG Quantitation; Cell Biolabs, CA, USA). As positive control, Escherichia coli genomic DNA (2 μg) was incubated in a final concentration of 50 and 100 mM H2O2 overnight at 37 °C, and subsequently tested.
Prokaryotic abundance in coral mucus and surrounding seawater
To highlight possible effects in terms of prokaryotic contamination associated with the exposure of the corals to microplastics, we determined prokaryotic abundances in the mucus released by C. rubrum and the surrounding seawater.
Prokaryotic abundances in the coral mucus collected from each tank (except for the control where coral mucus was not released) after 14 days of the experiment, were analysed by epifluorescence microscopy. The extraction of prokaryotic cells from the mucus (ca. 1 mL for each tank) was performed using pyrophosphate (final concentration, 5 mM) and ultrasound treatment (three 1-min treatments using a Branson Sonifier 2200; 60 W)87. Then, samples were diluted from 50- to 100-fold with sterile water filtered onto 0.2-μm pore-size filters (Anodisc filters; black-stained polycarbonate). The filters were stained using SYBR Green I (10,000× in anhydrous dimethyl sulfoxide, Molecular Probes-Invitrogen) diluted 1:20 in prefiltered TE buffer (pH 7.5) and incubated in the dark for 20 min; a drop (20 µl) of antifade solution (composed of 50% 6.7 mmol L−1 phosphate buffer at pH 7.8 and 50% glycerol with the addition of 0.5% ascorbic acid) was laid both on a glass slide and on the filter mounted on it. Prokaryotic counts were performed under epifluorescence microscopy (magnification, ×1000; Zeiss filter set #09, 488009-9901-000, excitation BP 450–490 nm, beam splitter FT 515, emission LP 520), by examining at least 20 fields per slide and counting at least 400 cells per filter.
For the determination of prokaryote abundance in seawater surrounding corals, three replicates of 10 ml of seawater were collected from each tank. Total prokaryotic abundance was determined according to Danovaro87. Samples were filtered onto 0.2-μm pore-size filters (Anodisc black-stained polycarbonate filters, Whatman) into a funnel with vacuum pressure no greater than 20 kPa (or 150 mmHg) to avoid cell damage. When the sample had passed through, filters were stained with 20 µl of SYBR Green I (10,000× in anhydrous dimethyl sulfoxide, Molecular Probes-Invitrogen) diluted 1:20 in prefiltered TE buffer (pH 7.5) and incubated in the dark for 20 min. Then, to remove the excess stain, filters were washed three times using 3 ml of Milli-Q water; a drop (20 µl) of antifading solution (composed of 50% 6.7 mmol L−1 phosphate buffer at pH 7.8 and 50% glycerol with the addition of 0.5% ascorbic acid) was laid both on a glass slide and on the filter mounted on it. Prokaryotic counts were carried out as described above.
Microbiome of corals exposed to microplastics
The coral microbiome was analysed immediately before the start of the experiment (before the addition of microplastics) and after 10 days of the experiment, both in replicated coral branches exposed to microplastics and in unexposed corals (Control t10). For the analysis of the microbiome, ca. 20 mg of tissue (wet weight) from one coral branch randomly collected from two tanks of each treatment and control was scraped from the skeleton by using surgical disposable scalpels (Braun) and DNA extraction was performed using the QIAGEN DNeasy Blood & Tissue Kit. Briefly, samples were digested with proteinase K at 56 °C overnight or until the tissue was completely lysed, then samples were processed following the manufacturer’s protocol. Finally, samples were held at −20 °C before PCR amplification and sequencing. The molecular size of the DNA extracts was analysed by agarose gel electrophoresis (1%) and the amount and purity of DNA were determined by Nanodrop spectrophotometer (ND-1000). For PCR amplification of the 16S V3 region, the Bacteria-specific primer pair 805R/341F was chosen with Illumina-specific adapters and barcodes. Sequencing was performed on an Illumina MiSeq platform by LGC Genomics GmbH (Berlin, Germany).
Raw sequencing paired-end reads were first joined using the bbmerge tool from the BBMap suite88 in a two-step process: reads that did not merge in a first step were quality-trimmed to remove low-quality bases (Q < 10) prior to re-joining to increase the number of merged sequences. Subsequently, joined sequences were submitted to the QIIME2 pipeline, as described by Corinaldesi et al.89 Amplicon sequence variants (ASVs) were identified through the DADA2 strategy90. The SILVA database v13891 was used as reference database for taxonomic affiliation of sequences; briefly, reference 16S sequences contained in the database were trimmed within QIIME2 to the region amplified by sequencing primers and representative ASVs were analysed using the classify-consensus-vsearch approach (consensus over 51% of at most 5 best hits) for taxonomic affiliation. Representative sequences were subsequently aligned using the MAFFT aligner within the QIIME2 suite92. The alignment was masked and was utilized for the construction of a midpoint-rooted phylogenetic tree using the FastTree software92. The ASV abundance table was randomly subsampled to 20,000 sequences after removal of ASVs represented by <10 sequences (to remove potentially spurious or non-informative sequences) and used, together with the rooted phylogenetic tree, to carry out statistical analyses to compare samples89.
Statistics and reproducibility
Differences in feeding rates, prokaryotic abundances in seawater and polymers ingested by corals were analysed with general linear models in R, using two-sided tests of one-way or two-way (Time, Treatment) experimental designs. Each experimental tank was treated as a replicate. Homogeneity of variance and normality was checked using residual and quantile–quantile plots. Where significant interactions occurred, post hoc Tukey multiple comparisons were carried out with Bonferroni adjustments to P values. Prokaryote abundance in mucus samples was taken in a balanced fashion across tanks, but one tank returned lower values across all three treatments. For this analysis, among-treatment effects were tested using a mixed model using lme4 in R93, treating “tank” as a random effect. Among-treatment differences in the extent of physical damage to corals were assessed using logistic regression which is a particular case of a generalized linear model, where the assumed distribution is binomial and uses a logit link function. The effect sizes among treatments were expressed as percentage differences converted from log-scale parameter estimates. Thus, estimates of 95% confidence limits are asymmetric on the arithmetic scale. Microbiome composition in corals was investigated after pre-treating the subsampled ASV table using the Hellinger transformation89 for normalizing the contribution of sequences to ASVs (after subsampling the ASV table to 20,000 sequences). The same treatment was applied to the contribution of sequences to prokaryotic families after subsampling to 20,000 sequences. Compositional and structural differences among treatments in the prokaryotic assemblage were tested using permutational multivariate analysis of variance (PERMANOVA)94 and pair-wise comparisons. Significant differences in mRNA expression 8-OHdG levels were determined using one-way ANOVA.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Source: Ecology - nature.com
