Altered acrylic acid concentrations in hard and soft corals exposed to deteriorating water conditions

The corals examined in this study were housed in a long established (>10 year) 180 gallon glass aquarium equipped with 14 000 k metal halide (Plusrite, Ontario, California, USA) and VHO actinic (UV Lighting Company International, Brook Park, Ohio, USA) lights, protein skimmer, and ReeFlo Dart pump (ReeFlo Pumps, Colorado Springs, Colorado, USA). The aquarium community comprised various species of hard corals (Acropora sp., Caulastrea sp., Euphyllia sp., Galaxea sp., Hydnophora sp., Montipora sp., Seriatopora sp.), soft corals (Cladiella sp., Discosoma sp., Lobophytum sp., Plexaurella sp., Protopalythoa sp., Ricordea sp., Sarcophyton sp., Sinularia sp., Zoanthus sp.), and fishes (Zebrasoma velifer, Zebrasoma flavescens, Centropyge loriculus, Amphiprion frenatus, Chromis viridis). Fishes and corals were fed a mixed diet of green and red algae (Ocean Nutrition, Newark, California, USA) and commercial frozen fish/coral food (San Francisco Bay Brand, Inc., Newark, California, USA). Water quality parameters such as temperature, salinity by specific gravity, pH, ammonia (salicylate test), nitrite, nitrate, calcium (precipitation/titration test), phosphate (orthophosphate test), and alkalinity were routinely monitored on a weekly basis, as well as at the times of sample collection for metabolomic assessment using commercial colorimetric test kits commonly used in the management of artificial sea water systems (API, Chalfont, Pennsylvania, USA). Sudden changes in three water quality parameters taken in aggregate characterized a decline in water quality that was recognized as needing immediate remediation to minimize impending impacts on coral health (Delbeek and Sprung 1994; Borneman 2001). The accuracy of colorimetric rapid test kits relative to standard validated bench chemistry methods varies substantially by test, but they generally provide sufficient precision to support water management decisions in aquaculture, and their use is a common practice in small- to medium-sized systems to support water quality management decisions made by biologists and farmers (Naigaga et al. 2016). To optimize accuracy and precision, it is important to use kits that are in date and properly stored, and to follow the directions carefully (Naigaga et al. 2016), as was done for the system in this study. Quality control of API commercial test kits to assure agreement with laboratory prepared standards is conducted by lot prior to distribution to consumers by the manufacturer, Mars Fishcare (GJ. Jones, personal communication, 2017). A limited number of individuals conducted the routine water testing of the aquarium system we examined, using in-date kits properly stored away from light and heat. Unexpected changes in test results were repeated with a fresh kit by multiple aquarists. The sudden simultaneous aggregate change in readings of three separate water quality parameters based on different underlying chemical reactions further strengthened our confidence that a change in water quality had occurred. The limits of detection for the three water parameter test kits that showed changes due to water quality decline are 0 mg/L for ammonia, 20 mg/L for calcium, and 0 mg/L for phosphate. After the collection of coral samples used in this study to represent corals in declining water quality, subsequent monitoring of the system water and coral health confirmed our interpretation. The three water quality parameters continued to worsen, as did other parameters, reaching conditions where frank coral loss occurred before the extensive mitigation efforts of water changes and system reconstruction returned the system to stability. Ammonia in a healthy reef system is rapidly removed through bacterial reduction to nitrite, and concentrations detectable with colorimetric tests provide warning of an inability of biofilter organisms to sufficiently reduce bio-wastes. Detectable inorganic phosphate concentrations can indicate an unexpected phosphate source or be an indication of protein skimmer inefficiency or loss of phosphate fixing organisms in biofilters. Elevated phosphates can support excessive algal growth and the depletion of water calcium concentrations necessary for coral skeletal growth. Water calcium concentrations are monitored for depletions of the mineral critical for coral skeleton growth. Parameters considered optimal for this aquarium were 0 ppm ammonia, 0 ppm phosphate, and ≥420 ppm calcium. A summary of water quality testing over the period of the metabolomics investigations is provided in Table 1, showing the stability of the system prior to the sudden water quality deterioration event due to unknown cause(s).

Coral samples were collected at two time points. Multiple small samples (ranging from 219.8 to 453.7 mg) were collected from the distal tips of a colony of stony small polyp coral (Acropora sp.) and soft coral (Lobophytum sp.) maintained under water quality conditions that had supported vigorous coral growth. One sample was processed for metabolomics analysis as described below, and the rest were stored at −80 °C for future work.

The same coral colonies were sampled for the second time 445 d later in a similar manner, 11 d after an unplanned acute aquarium water quality decline characterized by moderately increased ammonia (0.25 ppm) and phosphate (0.15 ppm) and moderately decreased calcium (360 ppm). At this time, the sampled corals showed clinical signs of distress with bleaching and algal growth where corals had previously exhibited active growth. Coral samples obtained from each collection were immediately frozen at −79 °C to quench enzymatic activity and minimize the oxidative destruction of susceptible metabolites. This was done by placing tissue samples directly in ∼1 cm diameter × ∼2.5 cm deep wells drilled into a dry ice block. Snap frozen coral samples were transferred to individual cryogenic tubes and then to a −80 °C freezer for storage prior to extraction with coral from deteriorated water quality conditions. Flash-frozen deteriorated water quality coral samples were processed without storage within 30 min of collection. Individual, similarly sized coral samples were selected from each deep frozen cohort for each type of NMR experiment and for the zooxanthellae-dissociated polyp experiments.

All samples were kept on wet ice throughout processing. Each sample was weighed, then 2:1 (v/w) of amphibian Ringer’s (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA) was added. A 1.4 mm stainless steel bead blend (Next Advance, Inc., Averill Park, New York, USA) was added at a ratio of 1:1 (v/v), and the samples were placed in a Bullet Blender (Next Advance, Inc., Averill Park, New York, USA) on speed 8 for 3 min. The homogenized samples were centrifuged at 13 000g for 25 min. The supernatants of all samples were collected and re-frozen at −80 °C and stored for 17 d prior to NMR analysis.

Extracts were thawed on wet ice and pipetted into a 96-well MultiScreen filter plate with Ultracel-10 membrane (Millipore Ltd., Carrigtwohill, Ireland) that had previously been prepared by soaking and multiple washes with ultrapure water. The plate was centrifuged at 3000g for 60 min at 21 °C, then 100 μL of Milli-Q water (Merck KGaA, Darmstadt, Germany) was added to each well, and centrifugation was repeated. The resulting filtrates were lyophilized (Savant ISS110 SpeedVac Concentrator, Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA), resuspended in 50 μL 100% deuterium oxide with 20 mmol/L phosphate buffer, 0.1 mmol/L TSP, and 1 mmol/L formate (Sigma-Aldrich, St. Louis, Missouri, USA), and placed in 5 mm NMR tubes (Wilmad-LabGlass, Vineland, New Jersey, USA). ¹H spectra were acquired on a 14.1 T Varian INOVA 600 MHz spectrometer at 25 °C. A one-pulse sequence with pre-saturation of the water resonance using a 90° flip angle and a total repetition time of 12.65 s was used to obtain the ¹H spectra.

A combination of homonuclear ¹H, heteronuclear ¹H, and ¹³C correlated NMR methods were applied to confirm the structure and identity of acrylic acid. These 1D and 2D pulsed field NMR experiments were performed on a Bruker AVANCE 700 MHz spectrometer with cryoprobe CP10K at ¹H frequency 700.17 MHz and ¹³C frequency 176.05 MHz at 25 °C using Bruker Topspin 3.1 software (Bruker, Karlsruhe, Germany).

Data were processed with ACD Labs 12.0 1D NMR Processor (Advanced Chemistry Development, Toronto, Ontario, Canada). Data processing included zero filling to 16 000 points, Fourier transformation, phase correction, baseline correction, and alignment of the reference TSP signal. For quantitative metabolite analysis, regions containing residual water and those upfield of TSP or downfield of 9.0 ppm were excluded. Intelligent bucket integration was used with a bin width of 0.04 ppm. The common integrals were normalized by dividing each bin by the sum of all the integrals per sample. To refine quantitative comparison of spectra across water conditions, for each compound of interest, the metabolite concentrations (the area under the curve of the summation of major peaks uncomplicated by peak overlap) were adjusted by dividing by the gross weight of the original coral sample. Chenomx NMR Suite 7.63 (Chenomx, Edmonton, Alberta, Canada), the Human Metabolome Database (Wishart et al. 2013), the Yeast Metabolome Database (Jewison et al. 2012), and the Spectral Database for Organic Compounds (National Institute of Advanced Industrial Science and Technology 2016) were used for peak identification.

The identification of acrylic acid as a key potential marker of coral response to water quality deterioration raised the question of whether this was a coral response or a response of the symbiotic zooxanthellae. To assess this, we isolated coral polyp tissues from zooxanthellae using methods modified from Banaszak (2007). Frozen samples of Acropora sp. and Lobophytum sp. were weighed (Acropora sp. 791.3 mg and Lobophytum sp. 207.8 mg). Coral tissue was scraped off the underlying structure with a sterile scalpel blade and placed in an Eppendorf tube with 1 mL amphibian Ringer’s solution. Tissue was homogenized at 20 000 rpm for 20 s with an Omni Prep Multi-Sample Homogenizer (OMNI International: The Homogenizer Company, Kennesaw, Georgia, USA). Coral tissue homogenates were centrifuged at 16 000g for 10 min; the coral supernatant was saved, and the zooxanthellae pellet was discarded. The supernatant was centrifuged twice more at 16 000g for 10 min, discarding the zooxanthellae pellet each time. Ten microliters of the final isolated coral animal tissue supernatant was placed in a Neubauer improved hemocytometer (INCYTO, Chungnam-do, Korea) to count zooxanthellae cells and assess for contamination. Zooxanthellae contamination of coral animal tissue isolates was further tested with evaluation for chlorophyll a absorbance spectrophotometrically (wavelengths 430 and 662 nm) (Nanodrop 1000, Thermo Fisher Scientific, Inc., Waltham, Massachusetts, USA). Coral tissue isolate samples were frozen at −80 °C, lyophilized at −85 °C (FreeZone Plus 2.5, Labconco, Kansas City, Missouri, USA), and rehydrated with 400 μL 10% deuterium oxide containing 0.1 mmol/L TSP. ¹H spectra were acquired on a Bruker AVANCE 500 MHz spectrometer with Oxford Narrow Bore Magnet, HP XW 4200 Host Workstation, using Topspin 3.1 software version (Bruker, Karlsruhe, Germany) to test for the presence of acrylic acid in coral tissue isolates.