Pathogenic marine microbes influence the effects of climate change on a commercially important tropical bivalve

There is growing evidence that climate change will increase the prevalence of toxic algae and harmful bacteria, which can accumulate in marine bivalves. However, we know little about any possible interactions between exposure to these microorganisms and the effects of climate change on bivalve health, or about how this may affect the bivalve toxin-pathogen load. In mesocosm experiments, mussels, Perna viridis, were subjected to simulated climate change (warming and/or hyposalinity) and exposed to harmful bacteria and/or toxin-producing dinoflagellates. We found significant interactions between climate change and these microbes on metabolic and/or immunobiological function and toxin-pathogen load in mussels. Surprisingly, however, these effects were virtually eliminated when mussels were exposed to both harmful microorganisms simultaneously. This study is the first to examine the effects of climate change on determining mussel toxin-pathogen load in an ecologically relevant, multi-trophic context. The results may have considerable implications for seafood safety.

Where sample size n<16 this was the result of mussel mortality during 14 day exposure to different simulated climate change conditions and microorganisms (Supplementary Table 1). For toxicity determination, samples were pooled (see supplementary methods for details.) First, all intercepts were allowed to differ across groups. Second, each variable was investigated by setting the intercept as equal across groups, which changed the model χ2. The difference in model χ2 indicates whether the intercept is significantly different across groups.

Calculation of toxin-pathogen load score
Paralytic shellfish toxin (PST) content of mussels was measured in µgSTX di-HCl eq/kg shellfish tissue and Vibrio concentration in CFU/g. For each mussel sample, one, or both of these values were normalised (depending on exact microorganism exposure regime) and added together where appropriate, to give an overall value per individual for toxin-pathogen load (see Supplementary Tables 5 and 6). were immediately transferred to the experimental aquarium facility <60 min after collection. Upon arrival mussels were exposed to constant conditions for at least five days to remove any effects of differences in recent environmental history. This was achieved by placing the mussels in a number of aquaria (volume= 200 L) filled with aerated seawater T (°C) = 28, S (PSU) = 35 that had previously been sand filtered and ozonated to remove the microbial community. Stocking density was at a maximum of one mussel per two litres. Mussels were exposed to a 12h:12h L:D cycle and fed the non-toxin producing diatom Thalassiosira weissflogii (Instant Algae TW1200, Reed Mariculture Inc, Campbell, CA, USA) once daily at a concentration of 1000 cells/mL. Every 1-2 d faeces and pseudofaeces were removed from the aquaria and half of the water was exchanged to remove metabolic waste products. Dead mussels were removed daily, however in most cases overall mortality was negligible (see Supplementary Table 1 for details). Temperature and salinity were measured daily and corrected if needed. parahaemolyticus culture was then sub-cultured into fresh TSBS medium and the optical density (OD 600 ) of which was measured every 30 min spectrophotometrically.

Culture of microorganisms
Simultaneously, aliquots were drawn for viable count determination by plating on tryptone soy agar with 1% NaCl (TSAS). The plates were incubated overnight at 30 °C. A growth curve was prepared by plotting the viable count (x-axis) against OD 600 values (y-axis). This graph was then used to determine the viable cell count.

Lysosome membrane stability
Lysosome membrane stability was evaluated in mussel haemocytes using the Neutral Red Retention Assay 3,4 . In healthy mussels lysosomes will retain the acidophilic vital dye neutral red. Cellular toxicity will result in the integrity of the membrane being compromised and subsequent leakage of the dye 5

Determination of gill Na + /K + -ATPase activity
Gill samples were defrosted on ice, pulse centrifuged and the SEI buffer decanted, following which samples were transferred to 2 mL microcentrifuge tubes. Next, 1.2 mL SEI deoxycholate buffer was added (0.1 % Na deoxycholic acid in SEI buffer) and the samples were homogenised with a bead beater (MM300, Retsch, Düsseldorf, Germany) using glass beads (Sigma-Aldrich, Poole, UK). The supernatant was analysed for Na + /K + -ATPase activity 6

Determination of metabolite concentration
Mantle levels of ATP, ADP, AMP, glucose and glycogen were determined using standard assays. Beforehand mantle extracts were prepared 7,8 . Briefly, mantle samples were removed from storage at -80 °C and a known mass was transferred to chilled 2 mL microcentrifuge tubes. Next, the samples were homogenised with four parts 0.9 mol L -1 HClO 4 using a bead beater and glass beads, following which the homogenate was centrifuged for 10 min at 20,000 g at 4 ºC. The homogenate was then transferred to a second 2 mL microcentrifuge tube. To neutralise the effects of the acid, four parts K 2 CO 3 , 3.75 mol L -1 for five parts of HClO 4 was added and the tubes were then centrifuged for 10 min at 20,000 g at 4 ºC. The resulting supernatant was removed and used for the following assays.
Concentrations of ATP, ADP, AMP, glucose and glycogen were determined spectrophotometrically in the mantle samples. All assays were undertaken in a microplate format using a Varioskan Flash plate reader. Mantle ATP, ADP and AMP concentrations were determined using NADH linked assays 8

Determination of standard MO 2
After the 14 d feeding and abiotic exposure period, mussels were placed in aerated, clean (sand filtered, ozonated and GF/F filtered (0.45 µm)) seawater in individual 1 L aquaria containing T. weissflogii (1000 cells/mL) for 12 h to allow for defecation of unassimilated A. minutum, V. parahaemolyticus and/or T. weissflogii from the guts.
After this period, standard MO 2 was determined using closed respirometry.
Respirometry chambers (volume = 500 mL) were filled with aerated, clean (sand filtered, ozonated and GF/F filtered) seawater at the respective experimental temperature and salinity and a magnetic flea added. A platform (50 x 50 x 15 mm) above the magnetic flea prevented contact between the mussel and the magnetic flea. A mussel was added to each chamber, which was then sealed while submerged to prevent air bubbles. The chambers were loosely covered with aluminium foil to ensure that disturbance to the mussel was minimised. Chambers were then placed onto magnetic stirrers (Remi Laboratory Instruments, Mumbai, India) to ensure adequate mixing of seawater and to prevent stratification of oxygen within the chamber. Before measurements began mussels were allowed to settle in the chambers for 1 h, which is the minimum time required for establishing resting MO 2 .
Planar optode spots (diameter 0.5 cm; PreSens Precision Sensing GmbH, Regensburg, Germany) were glued to the inside of each chamber. Oxygen levels in the chambers were measured every 5 min for a period of 1 h using a Fibox 4 oxygen

meter (PreSens Precision Sensing GmbH) and PreSens Datamanager software (PreSens Precision Sensing GmbH). All the equipment was located inside rooms
where the appropriate temperature level was maintained. For each chamber the decline in pO 2 was linear over the measurement period and was never allowed to fall to hypoxic levels. Background respiration was taken into account by running blanks, and the average value across a number of blanks was subtracted from the original MO 2 value. MO 2 was expressed as µmol O 2 h -1. g -1 . Upon completion of measurements of MO 2 , mussels were removed from the chambers, gently blotted dry and weighed. Mussel volume was also obtained by displacement.

Toxin extraction and analysis
Analysis of PST was conducted using liquid chromatography with tandem mass

Vibrio uptake quantification by qPCR
Quantification of V. parahaemolyticus uptake by mussels was determined using qPCR. For use in qPCR, initially genomic DNA was extracted from 1 g of homogenised mussel soft tissue using alkaline lysis and treatment with cetyltrimethylammonium bromide-NaCl and phenol-chloroform, followed by ethanol precipitation 12   Each tank contained 16 mussels which were randomly assigned to four groups at the end of the exposure time: Group one (four mussels): i) Haemolymph was taken for lysosomal membrane stability ii) Mantle tissue was taken for ATP, ADP, AMP, glucose and glycogen determination. iii) Gill tissue was taken for determination of gill Na + /K + -ATPase activity.
Group three (four mussels): Toxin and/or bacterial concentration quantification.