Coral collection and experimental design: effects of thermal stress on different life stages
In this study, we examined the effects of thermal stress on the following life stages and functions in the coral Acropora tenuis: adults (Experiment 1), adult fecundity (egg number and size) (Experiment 2), larval settlement rates (Experiment 3), and early recruitment (Experiment 4) (Fig. 1). To examine thermal stress and its carryover effects, we used coral fragments from the same colony to ensure a uniform genetic background, (Fig. 1, Supplementary Information 1). Coral fragments were exposed to two temperatures—28 °C (control) and 31 °C (stress)—for 10 d in a flow-through system. In the stress experiment, we placed a heater in each tank, and the temperature was gradually increased from 28 °C to ~31 °C over the course of 1.5 d (~0.5 °C every 6 h). We maintained the temperature at 31 °C until the fragments started to show visible bleaching. During this experiment, the fragments were maintained under 12-h light (160 µmol quanta m–2 s–1) and 12-h dark conditions.
We photographed the fragments to monitor their color and the maximum quantum yield of photosystem II (Fv/Fm) throughout the experiment (see below). We also measured their physiological parameters (including the symbiotic algal density and chlorophyll content) before and after the experiment (see below). Finally, all fragments were transplanted back to the reef until spawning occurred the following year.
In the spring of 2019, we examined the effects of thermal stress on larval stages, larval settlement, and early recruitment in corals (Fig. 1). One week before the anticipated spawning date, fragments from the previous year’s thermal stress experiment were retrieved from the reef. They were transported to an outdoor tank supplied with a flow-through seawater system and kept there until spawning. We set two treatment conditions (thermal stress and control) and used stressed fragments (SF) and controlled fragments (CF) for measurements. First, we collected gamete bundles from one coral fragment per aquarium at 19:30–20:40 hrs on June 12, 2019. We then mixed gametes from the same treatment groups (not from other treatments) for cross-fertilization. For Expriments 2, 3, and 4, we divided the larvae from each treatment into two groups (thermal stress and control). The thermal stress experiments on larvae were followed by the measurement of larval settlement rates and early recruitment under thermal stress (~31 °C) or control (28 °C) conditions. The temperature in the thermal stress treatment was increased at a rate of ~0.5 °C every 6 h and then maintained for 21 d. The water temperature was monitored every 10 min and recorded using a temperature logger (HOBO Water Temp Pro v2, Onset Computer Corporation, Bourne, MA, USA) throughout the experiment. Each treatment included three replicates, and all tanks were filled with seawater and equipped with a heater.
Physiological parameters of coral adults
We evaluated the photosynthetic activity of coral fragments in each treatment group by measuring the maximum quantum efficiency of photosystem II (Fv/Fm), the color of the fragment, and the symbiotic algal density (Supplementary Information 1). Following the thermal stress experiment, the fragments were transferred to the reef, and their recovery status was monitored. The fragments were photographed monthly using an underwater camera (Canon Powershot G10, Canon Inc., Tokyo, Japan). Survival rates were recorded 10 months after the thermal stress experiment.
Carryover effects of thermal stress on oogenesis and subsequent larvae
We counted the number of eggs per bundle and measured the size of eggs to examine the effects of thermal stress on the fecundity of adult corals (Supplementary Information). We measured the longest and shortest axis (length and width, respectively) of each egg using Image J (50)assuming an elliptical shape. Egg volume (EV) was calculated using the formula EV = 4/3 πab2, where a=½ egg length and b=½ egg width. We also measured the amount of lipids in the eggs according to a previously described method (44).
We measured the survival rate, size, and lipid content of the larvae to investigate the carryover effects of thermal stress on the next generation. We set two temperature treatments (control, 28 °C; thermal stress, 31 °C), with five glass containers for each treatment (five replicates × two treatments). We filled the containers with 300 mL of seawater (replaced every day). For survival analysis, we counted the number of live larvae in each treatment daily for 21 d. At 5 days post-fertilization (dpf), we randomly collected 50 larvae from each treatment group and photographed them under a microscope (Olympus BX53, Olympus Corporation, Tokyo, Japan) using a CCD camera (Olympus DP 72) at 10× magnification.
We compared the settlement rates of larvae in each treatment condition (control and stress) and under two temperature treatments (control and thermal stress) in five containers. In addition, we counted metamorphosed planulae as "settled planulae" after 24 and 48 h with a stereomicroscope (Olympus SZ61). We also photographed the spats with a digital camera (Canon G 10) and measured their size using ImageJ.
To evaluate the survival of early recruits in each condition, we maintained early recruits from settled larvae under thermal stress for 21 d. The cells were counted and photographed daily.
Statistical analysis
Before conducting statistical analysis, we examined the normality and homogeneity of the data using the Shapiro–Wilk and Levene's tests, respectively. To assess the differences in brightness and Fv/Fm among adult colonies over the experimental period, we used an LMM. The models were fitted using the 'nlme' package in R statistical software and selected according to Akaike's Information Criterion (51). We examined the symbiotic algal density and chlorophyll content of adult colonies among treatments using an unpaired t-test. In addition, we applied the Kaplan-Meier log-rank test to analyze survivorship in adult colonies, their larvae, and the recruits in different treatment groups. To compare the number of eggs per bundle, egg volume, and larval settlement rates between groups, we used a non-parametric Kruskal-Wallis test or pairwise Wilcoxon test with multiple comparisons. In cases where the assumptions for variance, normality, and homogeneity could not be met, we used a generalized linear model to compare larval sizes and larval settlement rates between treatments. All statistical analyses were performed using R statistical software (v.4.0.4) (51).