Nitric oxide production rather than oxidative stress and cell death is associated with the onset of coral bleaching in Pocillopora acuta

Study species and collection site

The scleractinian Pocillopora acuta (which has historically been misidentified as its sister species, P. damicornis, in Hawai‘i and other locations) (Poquita-Du et al., 2017; Johnston, Forsman & Toonen, 2018) was selected for this investigation because it is abundant on the shallow reefs adjacent to the Hawai‘i Institute of Marine Biology (HIMB), broadly distributed in the Pacific, well studied, and an important reef-builder in some parts of its range. Coral branches 3–6 cm in length were collected at 1–2 m depth with bone shears from colonies on the fringing reef adjacent to HIMB under a permit issued by the Department of Land and Natural Resources (Permit Number: SAP 2006-02).

Field investigation

To contextualize the results of our temperature perturbation experiments (described below) we examined NOS activity, oxidative stress and cell death in symbionts freshly isolated from P. acuta collected from the field between June, 2006 and September, 2006. During the summer and early autumn, seawater temperatures around HIMB typically range from 27 to 28 °C, sometimes reach 28–29 °C for up to several weeks, but rarely exceed 29 °C (Jokiel & Brown, 2004; Bahr, Jokiel & Toonen, 2015; Jury & Toonen, 2019). Seawater temperature on the reef during the timeframe of this study was obtained from the National Oceanographic and Atmospheric Administration (NOAA) Tides and Currents website (tidesandcurrents.noaa.gov) for the Moku o Lo‘e climate station located at HIMB.

A single coral branch was collected from each of 1–3 haphazardly selected colonies at each time point. The branches were each immediately placed into ice-cold sea water to prevent mucus build-up and to slow metabolism, preserving the physiological status of the coral until the sample could be processed, as described below. The time between coral collection and sample processing was always <30 min.

Temperature perturbation experiment

One to four coral branches were collected from each of ~30 haphazardly selected colonies (no more than one branch per colony was used in each treatment per trial). Branches were attached to micro-centrifuge tubes using underwater epoxy and placed upright on plastic trays. Coral fragments were allowed to recover from collection and mounting and acclimate to the experimental system for at least 3 days prior to beginning the experiment.

The experimental system consisted of nine aquaria, each eight L volume, placed in a large recirculating water bath maintained at 26 °C with an aquarium chiller. Each aquarium was outfitted with a 189 L h⁻¹ submersible pump, which provided turbulent water flow, and an aquarium heater, which maintained the temperature of each aquarium at the target level. After the acclimation period, corals were randomly assigned to one of three temperature treatments: 28.1, 28.6, or 29.9 °C (Table 1). Although these corals can tolerate temperatures of 28–28.6 °C for at least several weeks, even a few days of exposure at 29.6–30 °C can induce bleaching (Jokiel & Brown, 2004; Jury & Toonen, 2019). Hence, these treatments span the range from typical summertime temperatures to those which induce rapid bleaching during natural thermal stress events (Jokiel & Brown, 2004). Temperature was spot checked and recorded frequently in each aquarium during each experiment trial. The nine aquaria were arranged in a 3 × 3 grid and treatment temperature assigned to each aquarium in a Latin Square design, to control for row and column effects.

At the beginning of each experimental trial, two randomly selected coral branches were placed into each aquarium (ensuring that not more than one branch per colony was assigned to each treatment). After 24 h, one of the two branches was randomly selected for destructive sampling while the other was destructively sampled after 72 h. Three separate experiment trials to examine symbiont responses were performed during each of September, October, and November, 2006. One additional experiment trial was performed in December, 2006 to assess the potential contribution of coral host NOS activity to that measured for the symbionts.

Symbiont preparation

Coral tissues containing symbionts were removed from the skeleton using a Waterpik and recirculated, ice-cold artificial sea water (ASW; Coralife scientific grade marine salt, 34.34 g L⁻¹). Approximately 400 mL of ASW was used per sample. A portion of the resultant slurry was transferred to a 50 mL centrifuge tube and spun down at 3,100g for 5 min at 4 °C, to pelletize the symbionts. The overlaying supernatant containing coral tissue and ASW was decanted and the symbiont pellet was retained. The centrifuge tube was refilled with more coral slurry, the symbiont pellet resuspended, and spun down again. This process was repeated until all symbionts from each sample had been isolated in a single 50 mL centrifuge tube. The pellet was then rinsed 5–8 times by resuspending it in 30 mL ice-cold ASW, centrifuging it at 2,100–3,100g for 5 min at 4 °C each time, and decanting so as to remove residual coral tissue. After the final rinse the pellet was resuspended in two mL ice-cold ASW and transferred to a two mL microcentrifuge tube. Six 50 µL aliquots were removed from the homogenized, isolated symbionts and placed into microcentrifuge tubes—five for cell staining and one for symbiont density determination. The remaining symbionts were kept on ice and used for the NOS activity assay, as described below.

Symbiont cell densities were assessed by diluting each 50 µL aliquot with 450 µL ASW and counting four 10-µL subsamples using a hemocytometer. The remaining symbiont isolate in the 50 mL centrifuge tube was then centrifuged to pelletize the symbionts, the supernatant discarded, and the pellet resuspended in ~300–500 µL homogenization buffer (50 mM HEPES, 1 mM EDTA, pH = 7.4, in ASW) to yield a density of 50,000 cells µL⁻¹.

NOS activity assay, symbiont density, and normalization

An aliquot of the symbiont sample at the standardized cell density of 50,000 µL⁻¹ was pipetted to a two mL microcentrifuge tube along with glass beads (0.3g beads of diameter 425–600 µm and 0.15g beads of diameter 150–212 µm, Sigma Aldrich, St. Louis, MS, USA) and homogenized with a Qiagen tissuelyser (Retsch, Haan, Germany) at 30 Hz for 2 min. The sample was centrifuged at 13,000g for 2 min to remove the glass beads from the symbiont cell homogenate and the supernatant was pipetted to a 500 µL PCR tube for the NOS assay.

The NOSdetect kit from Stratagene (La Jolla, CA, USA) was used to determine NOS activity following the protocol of Trapido-Rosenthal et al. (2001, 2005. This method relies on the conversion ³H-arginine → ³H-citruline + NO with subsequent binding and removal of unreacted ³H-arginine to assess the rate of NO production via NOS, and assuming a 1:1 stoichiometry. A reaction mix containing final concentrations of 0.8 mM CaCl₂, 1.4 mM NADPH, and 0.028 mCi mL⁻¹³H-arginine was prepared for each sampling period using the reaction buffer provided in the kit and stock solutions of 6 mM CaCl₂, 10 mM NADPH, and 1 mCi mL⁻¹³H-arginine. Subsamples of 14 µL from each symbiont cell homogenate (each containing 700,000 cell equivalents) were transferred to PCR tubes and combined with 36 µL reaction mix. Final concentrations of reagents were 0.6 mM CaCl₂, 1 mM NADPH, and 0.02 mCi mL⁻¹³H-arginine. Blanks containing no NOS activity were produced as a procedural control by combining 14 µL homogenization buffer with 36 µL reaction mix. After 1 h incubation each reaction was halted by the addition of 400 µL STOP buffer (100 mM HEPES, 10 mM EDTA, pH = 5.5). Highly vortexed resin from the NOSdetect kit was then added to each tube (100 µL per tube) and the contents were reflux pipetted 20 times to ensure binding of unreacted ³H-arginine. The contents were centrifuged at 13,000g trapping the ³H-arginine bound to the resin in a filter provided in the NOSdetect kit and separating it from the fluid containing ³H-citruline. The liquid fraction was transferred to a scintillation vial, combined with five mL Universol^R liquid scintillation fluid (MP biomedicals, Solon, OH, USA), capped, shaken vigorously and counts per minute (cpm) read with a Beckman LS 3801 scintillation counter. The blanks were used to provide a baseline correction for the symbiont samples. Counts per minute were converted to attomoles (amol, 10⁻¹⁸ mol) ³H-citruline symbiont cell⁻¹ min⁻¹.

The same general procedure was used to assess coral host NOS activity except that coral tissues were removed from the skeleton via Waterpik using five mL ice-cold homogenization buffer, rather than ASW, and the NOS assay was performed with a 14 µL aliquot of supernatant containing host tissues rather than freshly isolated symbionts.

Coral skeletal displacement was measured in distilled water in a 100 mL graduated cylinder. Symbiont density in each sample was normalized to skeletal displacement, as was coral host NOS activity (Benayahu & Loya, 1986; Pupier, Bednarz & Ferrier-Pagès, 2018).

NO production assay

We intended to directly examine rates of NO production using the fluorescent dyes DAF-FM and DAF-FM DA (4-amino-5-methyamino-2′-7′-difluoreoscein diacetate, Molecular Probes, Eugene, OR, USA), which have been used to quantify NO production (Perez & Weis, 2006; Bouchard & Yamasaki, 2008). However, we encountered difficulty with this application. DAF-FM DA is a cell-permeant dye but after entering a cell, native esterases cleave off the diacetate group, making the resultant DAF-FM cell-impermeant and retaining it within the cell. DAF-FM is weakly fluorescent in the absence of NO but reacts with NO to produce a benzotriazole derivative which is 160× more fluorescent than DAF-FM.

After incubation with 9 µM DAF-FM DA or DAF-FM and intact cells or cell homogenates, respectively, fluorescence was read using a SpectraMax M2 96 well plate reader (Molecular Devices). To optimize our fluorescence measurements, during pilot experiments we prepared symbiont isolates and symbiont cell homogenates as described above and examined dilution series of 100–10,000 cells or cell equivalents µL⁻¹, using a sample volume of 100 µL. Contrary to expectations, we found that fluorescence (measured as relative fluorescence units, RFU) decreased in proportion to cell or cell equivalent density, suggesting that the symbionts were somehow quenching the fluorescence signal (Fig. S1). Calibration curves of NO production using the NO donor SNAP (S-nitroso-N-acetylpenicillamine, Sigma Aldrich, St. Louis, MO, USA) and mouse isolated NOS (Sigma Aldrich, St. Louis, MO, USA) showed the expected behavior of increasing fluorescence with increasing SNAP concentration using the same solutions and procedure as with the symbionts (Fig. S1). Reactions with mouse NOS were performed using the recommended master mix (final concentrations: 0.4 unit mouse NOS; 1 mM arginine; 12 µM tetrabiodiopterin; 170 µM dithiothreitol; 5 µM oxyhemoglobin; 0.1 mM NADPH; 1 mM magnesium acetate). With mouse NOS, a SNAP concentration of 325 µM produced easily detectable levels of DAF-FM fluorescence, so symbiont samples were incubated with 325 µM SNAP over a cell dilution series to determine the level of detection of NO production using our methods. Unfortunately, rates of NO production in all symbiont isolates were below the level of detection at cell or cell equivalent densities <500 µL⁻¹, which were required to reduce symbiont quenching sufficiently to allow any hope of representative fluorescence measurements. As a result, this assay was abandoned during experiment trials. To the best of our knowledge, the apparent fluorescent quenching we observed has not been previously reported, and we advise caution when applying fluorescent techniques to organisms which contain photoactive compounds.

Symbiont oxidative stress and mortality

We used two cell stains to assess symbiont oxidative stress and cell death in response to high temperature exposure. Fluorescein-derived 2′, 7′-dichlorodihydroflorescein diacetate (H₂DCFDA, Molecular Probes, Eugene, OR, USA) was used to assess oxidative stress. Like DAF-FM, it is a cell-permeant dye but after entering a cell native esterases cleave off the diacetate group, making the resultant H₂DCF cell-impermeant and retaining it within the cell. H₂DCF is essentially non-fluorescent in the absence of oxidants but reacts with hydrogen peroxide (H₂O₂) and hydroxyl radicals (·OH) to produce the fluorescent DCF. Symbiont death was assessed with the mortal stain Sytox green (Molecular Probes, Eugene, OR, USA) which is cell-impermeant for live cells but readily enters the compromised membranes of dead cells and stains nucleic acids green. A stock solution of 5 mM H₂DCFDA was prepared as needed in molecular grade DMSO, which was stored in the dark at −20 °C and was discarded after 2 days. Sytox green (5 mM in DMSO) was allowed to thaw in the dark prior to use.

The five symbiont samples collected from each coral for cell staining were used to produce (1) an unstained negative control, (2) an H₂DCFDA-stained sample, (3) a Sytox green-stained sample, (4) an H₂DCFDA-stained positive control, and (5) a Sytox green-stained positive control. The H₂DCFDA-stained positive control was incubated concurrently with 10 mM exogenous H₂O₂, to ensure a highly oxidizing environment was present within the algal cells. For the Sytox green-stained positive control, the symbionts were first killed by placing them in a 55 °C water bath for 10 min. Each sample was then incubated in the dark at a dye concentration of 5 µM (no dye in the negative control) for at least 25 min at room temperature. After incubation, the total symbiont density (stained and unstained) in each sample was determined by counting with a hemacytometer, as described above. The cells were then excited with blue light (excitation 450–490 nm, emission 515–565 nm) to determine the proportion of positively stained cells, using the positive and negative staining controls to set upper and lower bounds, respectively. When imaging the H₂DCFDA stain, a band pass filter was used to exclude excitation light with a wavelength shorter than 495 nm because chlorophyll autofluorescence interfered with imaging.

Statistical analysis

For the temperature perturbation experiment, the results of the symbiont NOS activity assay were averaged for each trial, with trial representing a statistical replicate. However, the NOS activity assay could not be completed for several samples during the October, 24 h sampling time, so these data were excluded from the analysis (24 h NOS, n = 2; 72 h NOS, n = 3). The staining data were analyzed using each nubbin as a statistical replicate. However, the samples for H₂DCF staining were lost during the September 72 h sampling time (72 h H₂DCF staining, n = 6; all other staining, n = 9).

The data were analyzed by two-way ANOVA with temperature and sampling time as factors, followed by a Tukey HSD as a post hoc in R v.3.5.2 (R Core Team, 2020). The staining data were arcsine square root transformed prior to analysis.

Field investigation

Seawater temperature recorded by the NOAA weather station at HIMB varied from ~26.5–29 °C during the summer of 2006 (Fig. 1). No coral bleaching was observed at any time during the field study (B Boeing, 2016, personal observations). NOS activity in freshly isolated symbionts ranged from 0.2 to 19.6 amol citrulline cell⁻¹ min⁻¹, whereas the proportion of oxidatively stressed cells (from H₂DCF staining) ranged from 13% to 33% and the proportion of dead cells (from Sytox staining) ranged from 13% to 35% (Fig. 1). These values may represent the normal range of symbiont NOS activity, oxidative stress, and cell death in P. acuta under typical summertime conditions in Hawai‘i and provide a baseline for our thermal perturbation experiments below.

Temperature perturbation experiment

Overall symbiont cell density (normalized to coral skeletal displacement) averaged 3.9 ± 2.5 × 10⁶ cells skeletal mL⁻¹ and did not differ significantly by treatment. Hence, we successfully sampled the corals during the early stages of heat stress, preceding significant reductions in symbiont density due to bleaching.

NOS activity increased according to temperature treatment as well as time of exposure (Table S1). NOS activity was similar in all treatments to that measured during the field investigation, except for the values measured under the 29.9 °C, 72 h treatment (Table 1), which were substantially higher than those measured in the field. Coral host NOS activity was examined to determine if residual host tissue could have biased the symbiont measurements, but was 3–6 orders of magnitude lower than those from the symbionts (see Supplemental Data File). Alternatively, it is possible that the host fraction contained minor residual contamination from the symbionts and the very low levels of NOS activity measured was not derived from the coral hosts. Regardless, these results gave us confidence that the symbionts were the primary site of NOS activity.

Symbiont oxidative stress increased with temperature, whereas cell death increased with both elevated temperature and time of exposure (Table S1). These values were similar to or somewhat higher than those obtained from the field-collected corals (Figs. 1 and 2). Unlike NOS activity, these two response variables (oxidative stress and cell death) tended to show maximum values in the intermediate temperature treatment (Fig. 2).