Does plastic type matter? Insights into non-indigenous marine larvae recruitment under controlled conditions

Macrofouling species, larval spawning, and culturing

Crassostrea gigas larvae were cultured in a hatchery under controlled conditions (Rico-Villa, Pouvreau & Robert, 2009, Vignier et al., 2021). Briefly, adult oysters were transferred to the Cawthron Aquaculture Park (CAP; Nelson New Zealand) hatchery for conditioning and fed ad libitum with bulk cultured Isochrysis galbana (8–9 × 10⁶ cells ml⁻¹) and Pavlova lutheri (10–12 × 10⁶ cells ml⁻¹). Fully mature oysters were strip-spawned according to Allen & Bushek (1992) and gametes were collected and fertilized. Embryos were then incubated in static 170-L tanks at 23 °C for 24 h, and D-larvae were transferred to 170-L flow-through rearing systems and continuously fed with a mixed diet of Chaetoceros calcitrans (CS-178) and Tisochrysis lutea (CS-177) throughout rearing. After 17 days, larvae developed into the pediveliger stage and were competent to settle.

Ciona savignyi adults collected from the underside of pontoons at the Nelson Marina (Nelson, New-Zealand–Lat: 41°16′ 14.81″ S; Long: 173° 17′ 2.54″ E) were housed in water lily baskets suspended in 50-L glass aquaria for up to 3 weeks. Aquaria were held at 18 ± 1 °C (mean ± standard error), 34 ± 1 Practical Salinity Unit (psu). Constant full-spectrum fluorescent light prevented premature spawning. Every day the filtration of the aquaria was paused for 3 h while C. savignyi were fed 250 mL of an 8−9 × 10⁶ cells mL⁻¹ Isochrysis galbana culture. Three gravid individuals with densely packed egg and sperm ducts were spawned according to Cahill et al. (2016). Following spawning, larvae were transferred to conical flasks and diluted with reconstituted seawater (RSW; 33 ± 0.5 psu; Red Sea Salt, Red Sea Aquatics, Cheddar, UK) to yield desired larval densities.

Spirobranchus caraniferus adults were collected from Delaware Bay (Nelson, New-Zealand – Lat: 41°09′ 33.6″ S; Long: 173° 28′ 34.1″ E). After collection, spawning was induced by removing the external calcareous tube of the worms and placing them together in a 100-mL glass beaker filled with RSW as described by Brooke et al. (2018). The eggs and sperm were left for 1 h to fertilize and then transferred to a 10-L conical flask filled with UV-sterilized filtered (0.4 µm) seawater (FSW; temperature 18 ± 1 °C; salinity 34 ± 1 ppt; pH 8.0 ± 0.2) and aerated with an aquarium bubbler. Larvae were cultured for 15 days with daily replacement of FSW and fed with a bispecific diet of I. galbana as per C. savignyi. The 15-day-old larvae were made competent to settle by exposure to 10⁻³ M 3-isobutyl-1-methylxanthine for 4 h and were then rinsed five times with RSW before being placed in the bioassay chambers.

Mytilus galloprovincialis adults were collected from submerged ropes at Elaine Bay in the Marlborough Sounds (Lat: 41° 3′ 19″ S; Long: 173° 46′ 9″ E, New Zealand). Mature individuals were induced to spawn by thermal stimulation, as described by His, Seaman & Beiras (1997). After spawning, fertilization and embryo-larval development were carried out according to (ASTM International, 2021). Larvae were kept in conical flasks, each containing 5-L of FSW and one aquarium bubbler stone to promote gentle mixing. Larvae were kept in culture and fed with a mixed diet of C. calcitrans (CS-178) and T. lutea (CS-177), with a daily change of FSW. After 23 days, larvae developed to the pediveliger stage and were competent for settlement.

Bioassay chambers, polymer production and study design

Six flow-through bioassay chambers were fabricated by Nelson Plastic Ltd. (Nelson, New-Zealand) following a design modified from Pansch et al. (2017). Each chamber consists of an inner V-shaped stand in which four panels (i.e., three different types of polymer tokens and a glass control) are placed at a 120° angle facing each other (Fig. 1). Six bioassay chambers (145 mm [L] × 105 mm [W] × 82 [H]) were used during each consecutive microcosm experiment.

Three different types of polymer tokens (50 mm × 76 mm), manufactured by Scion (Rotorua, New Zealand), were used as the ‘Substrate’ treatment in each microcosm experiment (Fig. 1). The tokens were injection-moulded from formulations of each investigated plastic polymer as follows: LLDPE–base LLDPE with Irganox 1076 (CAS 2082-79-3) and 0.25% Irganox B215 (Irganox B215 = 67%) Irgafox® 168 (CAS 31570-04-4) and 33% Irganox® 1010 (CAS 6683-19-8); PLA–Ingeo 3052D–blended with ethylene bis(stearamide) (CAS 110-30-5); Nylon-6–Ultramid B3S with talc and 0.5% Nylostab S-EED (CAS 42774-15-2). The polymer types used in this study (LLDPE, PLA and Nylon-6) contained additives typically included in the manufacture of these products for UV light stability and degradation prevention.

Four consecutive bioassay microcosm trials were undertaken between December 2020 and March 2021. The first bioassay was performed with C. gigas larvae (A1, 7-16 December 2020) followed by C. savignyi (A2, 11-20 January 2021), S. caraniferus (A3, 19-28 February 2021) and M. galloprovincialis (A4, 15-24 March 2021). During each trial (i.e., for each species), six bioassay chambers were supplied with 2.5 L of filtered (0.2 µm), seawater h⁻¹ as part of a 1,000-L recirculating system held at 18 ± 1 °C, and 34 ± 1 psu for the duration of the experiment. The seawater was not replaced between assays. Filtration of the recirculating seawater system consisted of a Dacron screen and a trickling biofilm filter filled with ~0.125 m³ of generic ‘bioball’ filter media. Bioassay chambers were kept under a 12:12 light to dark regime.

Each experiment consisted of: (i) Control assay: three bioassay chambers were kept in the recirculating system with constant seawater supply for 7 days to allow biofilm growth; (ii) Settlement assay: the remaining three bioassay chambers were kept in the same recirculating system with constant seawater supply for 7 days. After 7 days, 500 competent larvae of the respective model macrofouling species were placed in each of the three settlement assay aquaria for 48 h to investigate preferential settlement on each substrate (Fig. 2).

After 9 days (7 days of biofilm formation and 2 days of larval settlement) polymer and glass tokens were collected individually. For each experiment, there were 24 samples in total (six samples of LLDPE, PLA, Nylon-6 and glass controls). The 12 tokens from the settlement assays were used for microscopic analysis of larval settlement and were kept in a glass container with RSW for up to 1 h. The 12 tokens from the control assays were used for bacterial characterization were individually placed in Fisherbrand™ sterile sampling bags and kept on ice until processing within 2 h (Fig. 3).

Larvae counts and visualization of settlement location

Counts of larval settlement were performed with a dissecting microscope at 20× magnification (RS PRO Microscope, Norman King, Beauvais Cedex). The total number of settled larvae was recorded in each instance, along with specific location of larvae on the tokens as a proxy for larval aggregation. The location of the larvae on the different substrates was assessed visually with a dissecting microscope.

Metabarcoding of bacterial communities

Each step of the following molecular analysis (Fig. 3) was conducted in separate sterile laboratories with sequential workflow to eliminate cross-contamination. Rooms dedicated to DNA extraction, amplification set-up and template addition were equipped with laminar flow cabinets with HEPA filtration and room-wide ultra-violet sterilization which was switched on for >15 min before and after each use (Pearman et al., 2020). Aerosol’s barrier tips (Axygen, San Francisco, CA, USA) were used throughout.

Within 2 h, the sterile sampling bags containing the polymer tokens from the control assays were filled with 30 mL of ice-cold Tris-EDTA Buffer solution (Tris 1 mM, EDTA 1 mM; prepared from sterile, Ultrapure water, Ultrapure Tris pH 8.0 and Ultrapure EDTA pH 8.0) and sonicated for 2 min at 50 Hz in an ice-cold ultrasonic water bath to recover the attached biofilm fraction (Bandelin Sonorex RF 100H, 50–60 Hz, Sigma-Aldrich, USA). Following sonication, each homogenate was poured into separate sterile 50 mL Falcon tubes (Cat No. 227–261, Greiner Cellstar®, Sigma-Aldrich New Zealand). The sonicate solution was centrifuged (4,500 × g, 10 min, 4 °C). Supernatants were gently decanted and discarded, followed by an additional 5 min centrifugation step and removal of the remaining supernatant with a pipette (Wallbank et al., 2022).

Microbial DNA was extracted individually from each pelleted biofilm sample using the PowerSoil® DNA Isolation Kit (QIAGEN, MOBIO, Carlsbad, USA) following the manufacturer’s instructions. DNA was extracted from a total of 60 samples which was comprised of 12 samples for each of the bioassay trials, plus three procedural control samples per bioassay trial (one TAE buffer control, one seawater control before adding the larvae, one seawater control after adding the invasive larvae), along with extraction kit control blanks. Each sample was eluted in a final volume of 50 µL of elution buffer.

The V3-V4 regions of the bacterial 16S ribosomal RNA (16S rRNA) gene was amplified by Polymerase Chain Reaction (PCR), using the bacterial specific primers 341F: 5′-CCT ACG GGN GGC WGC AG-3′ and 805R: 5′-GAC TAC HVG GGT ATC TAA TCC-3′ (Herlemann et al., 2011; Klindworth et al., 2013). Both primer sets contained an Illumina overhang adapter for NEXTERA indexing, as described by Pochon et al. (2019). PCR reactions were undertaken in an Eppendorf Mastercycler (Eppendorf, Hamburg, Germany) in a total volume of 50 µL using MyFiTM PCR Master Mix (Bioline Meridian Bioscience, Memphis, Tennessee, USA), including 2 µL of each primer (10 mM stock) and 2 µL of template DNA. The PCR cycles for the 16S rRNA gene amplification were as follows: 94 °C for 3 min followed by 35 cycles of 94 °C (20 s), 52 °C (20 s) and 72 °C (30 s) with a final extension at 72 °C for 5 min. Negative (no-template) PCR controls were included in each PCR run (Audrézet et al., 2022). Amplicon PCR products were purified using AMPure XP PCR Purification beads (Agencourt, Beverly, MA, USA), quantified using a Qubit Fluorometer (Life Technologies, Carlsbad, CA, USA) and diluted to 3 ng µL⁻¹. An additional water control was added to test for potential contamination during the sequencing workflow. Normalized PCR products and controls (n = 63; 54 samples, five extraction blanks, and four PCR blanks) were sent for library preparation and sequencing on an Illumina MiseqTM platform at Auckland Genomics, University of Auckland, New Zealand following the Illumina16S rRNA metagenomics library preparation manual (D’Amore et al., 2016). Sequencing adapters and sample-specific indices were added to each amplicon via a second round of PCR using a Nextera Index kit. After that, 5 μL of each indexed sample was pooled, and a single clean-up of pooled PCR products was undertaken, as previously described by Audrézet et al. (2022). A bioanalyzer was used to check the quality of the library which was then diluted to 4 nM and denatured. The library was diluted to a final loading concentration of 7 ρM with a 15% spike of PhiX. Paired-end sequences (2 bp × 250 bp) were generated on an Illumina MiSeq instrument. Raw sequences were deposited in the NCBI short read archive under accession: PRJNA836386.

Bioinformatics and statistical analyses

For larval settlement data, one-way analysis of variance (ANOVA) was performed on the larval count results for the factor ‘Substrate’ using the ‘Vegan’ package in RStudio (Oksanen et al., 2013; R Development Core Team, 2013). Differences at p ≤ 0.05 were deemed statistically significant. Following ANOVA analysis, Tukey’s honestly significant difference test (Tukey’s HSD) was performed in RStudio to test differences between substrate types with the R package agricolae (De Mendiburu, 2014).

For metabarcoding data, raw sequence reads (with Illumina adapter sequences removed by the sequencing instrument) were trimmed using cutadapt v2.10 to remove primer DNA sequences (Martin, 2011), with no primer mismatch allowed. Quality filtering, denoising, merging pair-end sequences, and calling amplicon sequence variants (ASVs) were performed using the DADA2 version 1.20.0 package, implemented in R version 4.0.5 (Callahan et al., 2016). Following exploration of the DNA sequence quality plots, sequences were trimmed at a length of 220 for both forward and reverse reads, two or six errors were allowed for forward or reverse reads respectively. Reads were truncated at a quality score less than 2, and the maximum number of ambiguous nucleotides was set to zero. Singleton ASVs data (i.e., isolated sequences that were observed only once in the dataset) were removed to overcome sequencing errors (Tedersoo et al., 2010, Caporaso et al., 2011, Edgar, 2013). The remaining paired-end reads were merged with a minimum overlap of 25 bp and one mismatch allowed in the overlap region. Chimaera removal was performed using the default (consensus) method and the de-noised ASVs taxonomically classified against the SILVA 132 database for 16S rRNA (Quast et al., 2012) using DADA2 “assign Taxonomy” command, based on the RDP classifier (Wang et al., 2007).

The 16S rRNA dataset was filtered to exclude any ASVs classified as Eukaryota in the rank Kingdom, Chloroplast in the rank Class, and Mitochondria in the rank Family using the “subset_taxa” command implemented within the R package phyloseq (McMurdie & Holmes, 2013). The maximum number of ASVs found across negative controls was subtracted from the corresponding ASVs to offset potential contamination noise (Bell et al., 2019; Clark et al., 2020). Rarefaction curves were plotted using the ‘ggrare’ function in R (package ranacapa; Kandlikar et al. (2018)). The 16S rRNA rarefaction curves indicated that the sequencing depth attained per sample adequately captured biodiversity (i.e., the curves have reached a plateau). However, 10 samples yielded an extremely low post-filtering number of reads (<2,000) and were removed from further analysis (Table S1).

The community structure analyses were performed on the unrarefied dataset transformed into proportional read abundance. An ASV table generated by the bioinformatic pipeline was uploaded into the Plymouth Routines in Multivariate Ecological Research (PRIMER 7) v7.0.13 software (Anderson, 2001; Clarke & Gorley, 2015). Square root transformed data was used to construct Bray-Curtis similarity matrix (at ASV level) (Bukin et al., 2019), which was used to analyse bacterial community structure for the experimental factors ‘Substrate’ and ‘Assay’ (i.e., different polymer types and glass control, and temporal evolution of the bacterial community structure at the end of each microcosm experiment) with PERMANOVA (Permutational Multivariate Analysis of Variance). Before that, Levene’s and Shapiro-Wilk’s tests were performed to confirm that assumptions for normality and heterogeneity are met. The relative abundance of the 10 most abundant bacterial families related to the different assays (A1-A4) and between substrates was visualized using bar plots generated with the phyloseq and ggplot2 packages in Rstudio. In parallel, differences in bacterial community composition between substrates and among assays were visualized through a Principal Coordinate Analysis (PCoA), based on weighted unifrac distances.

Substrate-specific larvae recruitment

The model macrofouling organisms displayed different settlement patterns on each tested substrate (Fig. 4). Crassostrea gigas and C. savignyi preferentially settled on the lower part of the tested substrates (away from the water surface), with larvae clustering together (Figure S1). In contrast, S. cariniferus larvae were densely aggregated in the upper part of the substrates (near the water surface) (Fig. 4). The pediveligers of M. galloprovincialis clustered together on all substrates, aggregating in the center of the tokens (Fig. 4).

The number of settled C. savignyi larvae differed significantly among tested substrates (One-way ANOVA p ≤ 0.001), with maximum recruitment observed on nylon (44 ± 2 larvae per nylon token), and minimum recruitment on LLDPE and PLA (11 ± 5, and 13 ± 3 larvae per token, respectively). A similar pattern was observed for S. caraniferus, with significant differences among substrates (One-way ANOVA p ≤ 0.001). The maximum settlement was associated with PLA and Nylon (38 ± 4 larvae, and 36 ± 2 larvae per token, respectively) and minimum settlement on Glass and LLDPE (11 ± 5, and 16 ± 3 larvae per token, respectively). No preferential recruitment was detected for the two bivalve species: M. galloprovincialis (p = 0.219) and C. gigas (p = 0.534) (Fig. 5).

Biofilm community composition on polymers

There were no significant differences in the bacterial community composition between the four polymer types for each of the sequential control assays (A1-A4), (PERMANOVA, p = 0.17; Fig. 6, Figure S2). However, there was a progressive shift in the composition of the bacterial communities over the four sampling periods (Fig. 6), as revealed by the PERMANOVA analysis (p = 0.001). Biofilm communities went from being ostensibly dominated by Alteromonadaceae in Assay 1, to becoming increasingly diverse so that by Assay 4 there were roughly equal proportions of Alteromonadaceae, Cellvibrionaceae, Solimonadaceae, Hyphomonadaceae, and Rhodobacteraceae.

Although the differences within assays were insignificant, some trends between substrate types were observed. During the first assay there was a higher contribution of Alteromonadaceae on nylon and glass (75% and 80%, respectively), compared with LLDPE and PLA (70% and 65%, respectively). In contrast, the relative contribution of Flavobacteriaceae was higher on LLDPE and PLA (17% and 18%, respectively), compared with nylon and glass (12% and 10% respectively). In the second assay, differences in bacterial communities’ abundance were mainly associated with Nylon, where the relative contribution of Cellvibrionaceae, and Nisaeaceae were lower than on any other substrates, and Saccharospirillaceae’ relative contribution was two times greater than on LLDPE, and glass control (Fig. 6). A similar observation was made during assay 3, where the relative contribution of Bradymonadales was at least twice as high on Nylon compared to any other substrates. In addition, differences in bacterial family contribution were observed on LLDPE in A3. Hyphomonadaceae contributed 20% of the overall diversity on LLDPE, whereas its abundance was on average at 11% on the other substrates. Cellvibrionaceae abundance was lower in comparison with other substrates (12%, 24%, 21%, and 22% on LLDPE, PLA, nylon, and glass, respectively). The bacterial community was more homogeneous in the final assay (A4). Alteromonadaceae, Cellvibrionaceae, Solimonadaceae, Hyphomonadaceae, and Rhodobacteraceae contributed to more than 90% of the overall bacterial diversity across polymer types and glass control.