Role of Core Microbiome Shifts in Octocoral Litophyton Under Diurnal Temperature Fluctuations
Chien-Yi Wu, Hsien-Yu Cheng, Yen-Chih Lin, Yu-Chien Wang, Yan-Zhen Meng, Yunli Eric Hsieh, An-Chi Liu, Shan-Hua Yang
TL;DR
This study shows how diurnal temperature fluctuations help octocorals like Litophyton cope with heat stress by stabilizing photosynthesis and altering their microbiome.
Contribution
The study reveals that larger temperature fluctuations reduce thermal stress in octocorals and highlights the early role of Endozoicomonas in stress mitigation.
Findings
DTF maintained stable photosynthetic efficiency compared to constant warming.
Endozoicomonas abundance increased before physiological stress signs, suggesting early stress response.
ROS activity differences were only seen in the ± 5 °C group, not the larger ± 7 °C group.
Abstract
Climate change is projected to raise sea surface temperatures and intensify diurnal temperature fluctuations (DTF), threatening the survival of both scleractinian corals and octocorals. Litophyton, a common octocoral in Taiwan’s shallow reefs, is frequently exposed to large DTF and summer heat stress, making it a suitable model to study thermal resilience. Coral-associated bacterial communities are known to shift under thermal stress, and key bacterial taxa may play crucial roles in host acclimation. This study aimed to address two questions: (1) Can higher DTF mitigate cumulative heat stress in octocorals? (2) If so, what physiological and microbial community changes accompany this effect? To answer these questions, we conducted tank experiments under constant warming and two short-term DTF regimes (± 5 °C and ± 7 °C; baseline 25–27.8 °C), along with a no-fluctuation control. We…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10- —https://doi.org/10.13039/100020595National Science and Technology Council
- —National Taiwan University
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsCoral and Marine Ecosystems Studies · Marine Sponges and Natural Products · Microbial Community Ecology and Physiology
Introduction
Diurnal temperature fluctuations (DTF) in seawater have major impacts on coral reef ecology. As global climate change continues to raise sea surface temperatures (SSTs), coral reefs face increasing risks from both long-term warming [1–4] and short-term thermal variability [5–7]. Sea temperature anomalies occur at different frequencies and scales and can greatly affect corals [8, 9]. For instance, a marine heatwave in 2020 triggered mass bleaching events in many coral reefs around the world [10, 11]. If the benthic community of a coral reef changes during bleaching, recovery to pre-bleaching conditions becomes less likely, especially in the face of increased temperature in the hottest month or increased frequency of extremely high temperatures [12, 13]. In addition to long-term temperature variations (seasonal or monthly changes), short-term temperature fluctuations can also greatly impact coral physiology [14, 15], though these effects have received relatively less study. In one example, Safaie et al. [16] systematically assessed 20 environmental variables and 81 bleaching events at reef locations, finding that increased DTF was associated with lower risk of severe bleaching. A further study similarly showed that DTF may mitigate heat stress in two scleractinian coral species, Stylophora pistillata and Pocillopora acuta [17]. Although these studies consistently show that short-term temperature fluctuations can promote scleractinian coral survival, it is still unclear what impacts these conditions may have on octocorals.
Octocorals are crucial components of tropical Indo-Pacific ecosystem. These organisms provide habitat and food to fish, and they form structures such as shallow water reefs, deep seamounts, and submarine canyons [18]. Previous studies have suggested that some octocorals possess a remarkable capacity to tolerate highly variable environmental conditions, potentially enabling them to persist under future climate change scenarios [15, 19, 20]. Taiwan is home to more than 223 octocoral species and is considered a place of high octocoral diversity [21]. Among octocorals, genus Litophyton is widely distributed throughout Taiwan and can be found in shelter reef slopes or the sandy bottom of the shallow sea. Major Litophyton populations are present off the northeastern coast of Taiwan, an area that often exhibits large DTF. Our long-term observations (unpublished) suggest that compared to other shallow octocorals, Litophyton tend to exhibit more obvious bleaching during the summer and recovery during the winter, suggesting high exposure to thermal stress yet potential resilience mechanisms that remain understood.
Corals exist as complex holobionts, composed of the host and diverse associated microorganisms [22]. This close association plays a critical role in coral health, influencing defense against pathogens [23], enhancing cycling and exchange of key chemical substances such as nitrogen [24] or sulfur [25, 26], and even strengthening resistance to environmental stress [27]. Moreover, the bacterial genus Endozoicomonas may be considered an essential beneficial member of the coral-associated microbiome, as it participates in sulfur cycling in scleractinian corals [28, 29] and may also participate in host protection during stress [30]. Numerous studies also indicate that Endozoicomonas is highly associated with some octocoral species, establishing it as a significant dominant bacterial taxon within octocorals [31–33]. While temperature-driven microbiome shifts are well documented in scleractinian corals, little is known about how DTF affects the microbiome dynamics of octocorals. Recent work on octocoral Muricea laxa has shown that octocorals can exhibit distinct physiological responses under dynamic thermal conditions compared with scleractinians [15], highlighting that octocorals may employ different strategies for coping with DTF. However, the study focused primarily on host physiology, leaving the microbial dynamics of the holobiont largely unexplored.
In this study, we used Litophyton sp. to address the following questions: (1) Does higher DTF mitigate cumulative heat stress in octocorals? (2) If so, what changes would occur in the physiological performance and bacterial community composition of Litophyton sp.? To address these questions, this study conducted two tank experiments. Experiment I evaluated octocoral responses under constant warming without short-term temperature fluctuations, to determine whether Litophyton exhibits measurable physiological decline when exposed to continuous heat accumulation. Establishing this baseline response was essential for assessing the potential mitigating role of DTF. Experiment II then tested whether different magnitudes of short-term DTF could alleviate the thermal stress observed under constant warming, examining both host physiology and associated microbial community dynamics.
Materials and Methods
In this study, we conducted two independent experiments to assess the effects of thermal stress and DTF on Litophyton sp. Experiment I (constant warming) aims to assess physiological responses under gradually increased stable temperatures. Experiment II aims to evaluate the combined effects of short-term DTF on coral physiology and associated bacterial communities. To assess coral physiological responses under these thermal regimes, we measured the Coral color scores and photosynthetic efficiency (Fv/Fm) in both experiments, and antioxidant enzyme activities, catalase (CAT) and superoxide dismutase (SOD), in Experiment II. In addition, the bacterial community composition of whole corals from Experiment II was analyzed.
Octocoral Collection and Identification
For Experiment I, eight healthy Litophyton specimens (around 15–20 cm^2^) were collected from Kueishan Island (24°50’N, 121°56’E) in November 2023. For Experiment II, nine specimens were collected from the same site in February 2021. Colonies were sampled at a depth of 8–10 m using chisels and hammers, with each colony placed separately in a Ziploc bag. In Experiment I, samples were transported directly to the Institute of Fishery Science (IFS), National Taiwan University, for acclimation. In Experiment II, samples were first maintained overnight in flow-through aquaria under fluorescent light (fluorescent lamps, 150 µmol m⁻² s⁻¹) at the Marine Research Station near Kueishan Island, and then transported to IFS.
At IFS, corals in Experiment I were acclimated at 25 °C for two weeks, while those in Experiment II were acclimated at 25 °C for one week in a closed circulating system. Following acclimation, colonies were fragmented for subsequent temperature treatments (Figure S1).
Morphological identification of octocorals was based on colony appearance and sclerite morphology observed under an optical microscope (BA210, Motic) followed Dai [21]. Tissue samples for appearance and sclerite morphological identification from different parts of each Litophyton colony were preserved in 99% ethanol. For molecular identification, total DNA was extracted from preserved tissue samples with the DNeasy PowerSoil Kit (Qiagen, USA). Single-gene barcoding was performed using cytochrome oxidase I (COI) primers, COII8068F (5’- CCATAACAGGACTAGCAGCATC − 3’) and COIOCTR (5’ - ATCATAGCATAGACCATACC − 3’) following McFadden, Benayahu [34]. The modified PCR protocol consisted of 40 cycles with an initial step of 94 °C for 5 min, 94 °C for 30 s, 47 °C for 30 s, 72 °C for 30 s, and finally 72 °C for 10 min. PCR products were sequenced via Sanger sequencing, and the results were compared against the NCBI database using BLAST. COI sequences are deposited in NCBI GenBank, BioProject ID PRJNA1160454.
Tank Setting and Experimental Design
For both experiments, parent colonies were fragmented into pieces approximately 3 cm² in size using sterilized scissors. In Experiment I, we used eight colonies and each coral colony was divided into 10–12 sub-colonies. Fragments from eight different colonies were evenly and proportionally distributed among the four replicate experimental groups (control and treatment). Each tank thus contained fragments from two colonies, allowing us to balance individual variability across tanks. Due to space limitations in the tanks, the design was adjusted to duplicates to balance tank space feasibility and statistical robustness. A total of eight recirculating tanks were used, each equipped with a filtration chamber and supplemented with live rocks. Among these, one tank was maintained as a backup system to ensure that sample mortality would not arise from unexpected tank conditions; therefore, only three experimental treatments are presented in the final dataset.
In Experiment II, sub-colonies from the same colony were placed into three independent tanks. Each tank was equipped with a Hang-on Back aquarium filter (Figure S1), and the tanks were grouped in sets of three and positioned within three larger cyclic tanks (63.8 cm × 42.5 cm × 32 cm). The water temperature was controlled using the Apex Fusion System (Apex System, Neptune, USA). The actual temperature profiles for the three treatment groups were as follows: Control group : 25–27.8 °C, stable daily; ± 5 °C group: 25 ± 5–27.8 ± 5 °C, fluctuating around a mean of 25–27.8 °C (min : 20 °C ; max : 32.8 °C); ± 7 °C group: 25 ± 7–27.8 ± 7 °C (min : 18 °C ; max : 34.8 °C), fluctuating around a mean of 25–27.8 °C. Temperature cycled daily and elevated 0.1 °C per day. These conditions were based on the SSTs of the Kueishan Island, the data were obtained from the Taiwan Central Weather Administration’s open data platform, covering approximately the five months (Fig S1). The highest daily temperature occurred between 12:00 to 13:12, while the lowest occurred between 00:00 to 01:12. In each larger cyclic tank, the temperature measurements were recorded using a HOBO data logger (HOBO Pendant Temp/Light, 64 K) (Fig. 1c). A light intensity of 150 µmoles m^− 2^ s^− 1^ was provided by an LED aquarium lamp (Optimus Reef NANO) following a diel cycle (12 h:12 h) for four weeks. Both two experiments were conducted using artificial seawater prepared by mixing RO (reverse osmosis) water with commercial sea salt (Aquaforest, Portland). We changed water per week, and calcium, magnesium, KH and salinity levels of the artificial seawater were monitored and recorded daily. In Experiment I, sampling was performed every five days. The first sampling, conducted after the acclimation period, was designated as Day 0, with a total of eight sampling points collected throughout the experiment. In Experiment II, sampling was performed weekly, with four sampling points in total on Days 7, 14, 21, and 28.
Fig. 1. Experimental design and actual temperature record from Hobo loggers**. a** Experiment I, b Experiment II and c Schematic of experimental tanks and actual temperature record for each group. Experiment I was divided into two groups: a control group maintained at a constant temperature of 25 °C and a treatment group gradually heated from 25 °C to 31 °C over 30 days, followed by a 5-day holding period at 31 °C (total duration: 35 days). Coral samples were collected at Day 0, Day 30, and Day 35, with n = 4 biological replicates per group. All tanks were maintained under a 12 h light : 12 h dark photoperiod. Experiment II was divided into three groups: control groups (constant 25 °C), ± 5 °C DTF groups (25 ± 5 °C), and ± 7 °C DTF groups (25 ± 7 °C). Coral samples were collected at Day 7, Day 14, Day 21 and Day 28, with n = 3 biological replicates per group. All groups were maintained under a 12 h light : 12 h dark photoperiod. All groups started at an initial temperature of 25 °C, with the temperature increasing by 0.1 °C per day for a total duration of 28 days
Physiological Measurements on Litophyton in the Tank Experiment
The Appearance of Litophyton
The appearance of coral samples was checked by the coral health chart (Coral Health Chart, Coral Watch, The University of Queensland, Australia) [35]. In Experiment I, coral colonies were photographed at each sampling time using a Canon EOS-650D camera (Japan) at ISO 400, whereas in Experiment II, photographs were taken using a smartphone camera. Both methods provided sufficient resolution to accurately assess coral coloration and health status. In Experiment I, four replicate of experimental groups (control and treatment) were included at each time point, and eight samples were collected per control/treatment (n = 16), resulting in a total of 95 samples in 6 time points (one fragment was dead) for image analysis; In Experiment II, three treatments were included at each time point, and three samples were collected per treatment (n = 9), resulting in a total of 36 samples in 4 time points for image analysis. Coral coloration was quantified by analyzing photographs that were standardized using the Coral Health Chart for white-balance calibration in ImageJ. For each image, whole regions of interest (ROIs) were selected from the coral tissue and converted to grayscale. The mean gray value of each ROI was measured, and values were matched to a calibrated grayscale reference based on the coral health chart to assign a color score from 1 (pale) to 6 (dark). Scores from multiple ROIs were averaged per sample and reported to one decimal place [36, 37].
Photosynthetic Performance
The photosynthetic performance of the Litophyton was evaluated before been collected by measuring the maximum quantum yield (Fv/Fm) of symbiotic algae using a Mini-PAM (Heinz Walz GmbH, Germany) for Experiment I, and Diving PAM (DIVING-PAM Underwater Fluorometer, Heinz Walz GmbH, Germany) for Experiment II after the Litophyton fragments were acclimated to darkness for 30 min.
Activities of Antioxidant Enzymes
The activities of antioxidant enzymes (SOD, CAT) are widely recognized biomarkers of oxidative stress in cnidarians. Under thermal stress, excess ROS production by Symbiodiniaceae triggers activation of antioxidant enzymes, making SOD and CAT reliable indicators of cellular stress responses [17, 38–40]. SOD and CAT activities were tested during Experiment II following the method of Higuchi et al. [40]. Litophyton tissue homogenate was collected by grinding fragments using a stainless-steel mortar. Tissues were homogenized in 1–2 ml of ice-cold lysis buffer (50 mM phosphate, 0.1 mM EDTA, 10% [v/v] glycerol, pH 7.0). The resulting homogenate was kept on ice and then centrifuged at 2000 g for 5 min at 4 °C. The supernatant was collected and centrifuged again at 16,000 g for 5 min at 4 °C. Afterward, the final supernatant was aliquoted and snap-frozen with liquid nitrogen.
CAT activity was measured by tracking the depletion of H_2_O_2_ at 240 nm using a spectrophotometer (SpectraMax^®^ ABS Plus, Molecular devices). For the blank, 740 µl of potassium phosphate buffer (50 mM, pH 7.0, 0.1 mM EDTA) was mixed with 10 µl of the sample in a quartz cuvette. The reaction was initiated by adding H_2_O_2_ to a final concentration of 20 mM and monitored for 3 min at room temperature. Triplicate measurements were taken for each sample, and CAT activity was calculated using an extinction coefficient for H_2_O_2_ of 43.6 M^− 1^cm^− 1^ [41].
SOD activity was assessed using a SOD Determination Kit (19160, Sigma-Aldrich, US). Protein content was quantified using the PierceTM 660 nm protein assay kit (Thermo Fisher Scientific, US) and a spectrophotometer we used to measure CAT activity, with BSA as the standard.
Enzyme activities were normalized to total soluble protein concentration, quantified via a Bradford assay. SOD activity was expressed as U mg⁻¹ protein, where one unit represents a 50% inhibition of NBT reduction. CAT activity was expressed as µmol H₂O₂ decomposed min⁻¹ mg⁻¹ protein.
Bacterial Community Analysis
DNA Extraction and 16 S rRNA V6V8 Region Amplicon PCR
Litophyton samples from Experiment II (n = 36) were used for bacterial community analysis. The samples were homogenized using a sterilized mortar and pestle to facilitate DNA extraction. Total DNA was extracted with the DNeasy PowerSoil Kit (Qiagen, USA), and the quality of the extracted DNA was assessed with NanoDrop spectrophotometer (Thermo Fisher Scientific, US). PCR was performed using bacterial universal primers, 968F (5’- AGAGTTTGATCMTGGCTCAG-3’) and 1391R (5’-CTGCTGCCTCCCGTAGG-3’), targets the V6-V8 hypervariable region of the bacterial 16 S ribosomal DNA gene [42]. The PCR protocol consisted of 30 cycles with an initial step of 94 °C for 5 min, 94 °C for 30 s, 52 °C for 20 s, 72 °C for 45 s, and finally 72 °C for 10 min [17]. DNA tagging PCR (DT-PCR) was employed to label the PCR products [43] for Illumina sequencing. Amplicon sequencing of the V6-V8 regions was performed using the Illumina MiSeq platform with 300 bp paired-end configuration (Yourgene Health, Taiwan).
Amplicon Sequence Analysis
16 S rDNA amplicon sequences were processed with the QIIME 2 pipeline (version 2019.10) [44]. Initially, all raw reads were demultiplexed using cutadapt (version 1.15) [45]. Subsequently, the demultiplexed sequences were processed with the DADA2 algorithm, and chimeras were removed [46]. This procedure generated amplicon sequence variants (ASVs) with an average sequencing depth of approximately 20,000 reads per sample, and no samples were removed due to insufficient read depth. Taxonomic assignment of ASVs was performed using the classifier-consensus-vsearch plugin [47] with the SILVA 138 NR99 database [48, 49]. For quality control, ASVs identified as mitochondria or chloroplast were removed; no additional abundance or prevalence filtering was applied. Although rarefaction was conducted during preliminary data exploration, all downstream community and differential abundance analyses were performed using the unrarefied ASV table, following current best-practice recommendations [50, 51].
Statistical Analysis
Normality of each dataset was checked using a Shapiro-Wilk test in R to determine the appropriate statistical approach. Color scores and Fv/Fm in Experiment II met normality assumptions (p > 0.05), whereas color scores and Fv/Fm in Experiment I did not. Similarly, all antioxidant enzyme measurements (SOD and CAT) and microbial community data also violated normality assumptions. Accordingly, parametric datasets (Experiment II color score and Fv/Fm) were analyzed using One-way ANOVA with Tukey’s HSD post-hoc tests, while non-parametric datasets (Experiment I color score, Experiment I Fv/Fm, SOD, CAT, and microbial diversity indices) were analyzed using Kruskal–Wallis tests followed by Dunn’s post-hoc tests. Statistical significance was defined as p ≤ 0.05. All tests were performed with the R package ‘stats’ [52]. Alpha diversity of the microbial community was assessed using four indices: Observed, Chao1, Shannon, and Simpson. Statistical differences were evaluated using the Kruskal–Wallis test, followed by Dunn’s post-hoc test. Dissimilarity in bacterial composition among samples was calculated by the Bray-Curtis distance for principal coordinates analysis (PCoA), and differences among experimental groups were further tested using analysis of similarities (ANOSIM) to assess whether group membership explained community variation. All calculations were conducted using the R package ‘vegan’ [53]. Bacterial community analyses were managed and analyzed using the R package ‘phyloseq’ [54], allowing integration of taxonomic, abundance, and sample metadata. Core microbiome were identified and filtered using the R package ‘microbiome’ [55], and statistical differences were evaluated using the Kruskal-Wallis test, followed by Wilcox Rank-Sum post-hoc test. All figures were visualized using R package ‘ggplot2’ [56].
Phylogenetic Analysis
A phylogenetic tree was constructed using 16 S rRNA gene sequences of Endozoicomonas in MEGA (version 12) [57]. Sequence alignment was performed with MUSCLE [58]. The alignment was manually trimmed to remove gaps and ensure all sequences started and ended at the same positions. The tree was generated using the Maximum Likelihood method with 1,000 bootstrap replicates. Because the aligned 16 S rRNA region was short (408 bp), most nodes exhibited low bootstrap support, a common outcome for short hypervariable regions [59]. Top 6 Endozoicomonas ASVs in the core microbiome were selected. Additionally, four previously reported Endozoicomonas genome sequences from Litophyton [60] were included to construct the phylogenetic tree. The reference Endozoicomonas sequences were obtained from the NCBI database.
Results
Coral Host Species Identification
According to sequencing of the single-gene barcoding marker COI, all collected octocoral colonies were identified as Litophyton acuticonica, with percentage identities above 98% (Table S1), suggesting that all samples represent the same Litophyton species and were therefore treated as Litophyton sp. in subsequent analyses.
Physiological Responses Under Different Temperature Treatments
In Experiment I (non-fluctuating experiment) (Fig. 1a), although the color scores in both the control and treatment groups slightly declined after the 30-day experiment (control: from 5.38 ± 0.41 to 4.88 ± 0.36; treatment: from 5.28 ± 0.36 to 4.73 ± 0.37), the differences were not statistically significant within each group (Kruskal-Wallis test, p > 0.05). Moreover, no significant differences were observed between the control and treatment groups either before or after the experiment (Fig. 2a) (Dunn’s post-hoc test, p > 0.05, Table S2). However, the Fv/Fm results revealed a significant difference between groups after the experiment (Dunn’s post-hoc test, p > 0.05, Table S3). On the Day 30, corals in the treatment group exhibited significantly lower Fv/Fm values compared to those in the control group, indicating a decline in photosynthetic efficiency likely caused by heat stress induced by elevated temperatures (Fig. 2c).
Fig. 2. Temporal changes in color score and maximum dark-adapted quantum yields of Litophyton sp. holobiont in two independent experiments. Corals were exposed to different gradually increasing temperature with/without different DTF. Color score of Litophyton sp. in a experiment I and b experiment II. Maximum dark-adapted quantum yields (Fv/Fm) of Litophyton sp. in c experiment I and d experiment II. The error bars represent the standard deviation. Statistical differences are indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.001)
In Experiment II (fluctuating experiment) (Fig. 1b, c), no significant within-group or between-group differences in color scores were observed between the early and late phases of the experiment (Day 7 and Day 28) (Tukey’s post hoc test, p > 0.05, Table S4). A minor decline was noted only in the ± 7 °C group, where the color score decreased from 5.30 ± 0.10 on Day 7 to 4.97 ± 0.25 on Day 28 (Fig. 2b). Interestingly, the Fv/Fm results showed a similar pattern, with no significant intra- or inter-group differences across time (Tukey’s post-hoc test, p > 0.05, Table S5). The mean Fv/Fm values remained generally above 0.6 in all groups, except for the ± 7 °C group on early phase. From early phase to late phase, Fv/Fm values remained relatively stable across treatments: control group (0.627 ± 0.002 to 0.686 ± 0.007), ± 5 °C group (0.623 ± 0.026 to 0.672 ± 0.008), and ± 7 °C group (0.594 ± 0.041 to 0.687 ± 0.022), indicating consistent photosynthetic quantum yields (Fig. 2d). Compared to the Fv/Fm results from Experiment I, the findings suggest that short-term DTF may help mitigate the adverse effects of thermal stress on the photosynthetic efficiency of coral symbiotic algae, even when temperatures exceed 31 °C.
Based on the measured color score and Fv/Fm results from the previous experiment, we further assessed whether Litophyton sp. experienced thermal stress at the cellular level. To do so, we measured ROS-related enzyme activities in the collected samples. The mean values and standard deviations were calculated for each group at the different time-points.
For SOD activity, the ± 5 °C group exhibited the highest mean SOD activity levels and largest standard deviations at Day 7 and Day 14 (27.83 ± 18.87 U/mg and 43.45 ± 25.04 U/mg, respectively), whereas lower levels were observed in the control and ± 7 °C groups. Notably, at Day 21, although the SOD activity across all treatments declined, the SOD activity in the ± 5 °C group was significantly higher than that of the control group (Dunn’s post-hoc test, p < 0.05, Table S6). Throughout the experiment, while the ± 5 °C group consistently exhibited higher SOD activity than the control and the ± 7 °C groups, no other significant differences were observed between treatments (Dunn’s post-hoc test, p > 0.05, Table S6) (Fig. 3a).
Fig. 3. The activities of superoxide dismutase and catalase in Litophyton sp. holobiont of Experiment II. Corals were exposed to different gradually increasing temperature with different DTF for four weeks. a Superoxide dismutase (SOD) activity of Litophyton sp. b Catalase (CAT) activity of Litophyton sp. Coral samples was collected at each time point, with n = 3 biological replicates per group. The error bars represent the standard deviation. Statistical analysis was conducted using Kruskal-Wallis and Dunn’s post-hoc test. Statistical differences are indicated by asterisks (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.001)
For CAT activity, significant variations were observed among treatments (Kruskal–Wallis test, p < 0.05). At Day 7, the CAT activity in the ± 5 °C group was significantly elevated compared to the ± 5 °C group (Dunn’s post-hoc test, p < 0.05, Table S6). The ± 5 °C group continued to show high activity levels through Day 21, where it reached its peak (nearly 500 U/mg). At this time point (Day 21), the CAT activity of the ± 5 °C group was significantly higher than that of the control group (Dunn’s post-hoc test, p < 0.05, Table S6). However, by Day 28, CAT activity in the ± 5 °C group declined to levels lower than those of the other two groups, showing no significant differences (Dunn’s post-hoc test, p > 0.05, Table S6). When considering all time points together, CAT activity displayed a trend similar to that observed for SOD activity: the ± 5 °C group consistently exhibited higher CAT activity than both the control and the ± 7 °C groups (Fig. 3b).
In summary, although both SOD and CAT activities exhibited certain significant fluctuations, particularly in the ± 5 °C group, no corresponding decline was observed in color scores or Fv/Fm values, regardless of the duration of exposure or the magnitude of DTF. Furthermore, neither SOD nor CAT activity showed a sustained increase over time or a positive correlation with larger fluctuation magnitudes (± 7 °C vs. ± 5 °C). These results indicate that the observed enzymatic fluctuations remained within the physiological tolerance range of Litophyton sp. The stability of color scores and photosynthetic efficiency suggests that the corals were able to maintain their physiological performance despite these transient shifts in antioxidant enzyme activities. However, at the cellular level, the elevated activities of SOD and CAT suggest that the ± 5 °C treatment imposed a significantly higher degree of heat stress on Litophyton sp. compared to the control group. Detailed statistical results for all time points and treatments are provided in Supplementary Table 2.
Bacterial Dynamics in Litophyton sp. Under Different Magnitudes of DTF
In Experiment II, we analyzed n = 3 fragments per temperature group, with a total of 36 samples collected across four time points. Throughout Experiment II, the fluctuations groups exhibited distinct dynamics from the control group in terms of bacterial diversity (Fig. S2). By Day 14, all four indices in the control group had decreased, but they rose by Day 21 and remained relatively high through Day 28. In contrast, the two fluctuating groups showed an increase on Day 14, followed by a decline on Day 21, and a slight recovery by Day 28. But when comparing across all treatments, no significant differences were observed in the four diversity indices (Observed, Chao1, Shannon and Simpson) during the experimental period (Dunn’s post-hoc test, p > 0.05, Table S7).
When the three temperature conditions were analyzed collectively, disregarding the effect of time, the Bray-Curtis dissimilarity showed significant differences in bacterial community composition among the three groups (Fig. S3, ANOSIM, R = 0.069, p < 0.05). Comparing the different treatments at each week further revealed temporal changes in bacterial community composition. No significant differences were detected among the three groups at either Day 7 or Day 14 (Fig. 4a–b; ANOSIM: Day 7, R = 0.136, p > 0.05; Day 14, R = 0.012, p > 0.05). However, the divergence became more pronounced in Day 21, with both treatment groups showing significant difference from the control group (Fig. 4c, ANOSIM, R = 0.325, p < 0.05). By Day 28, the bacterial community compositions in the treatment groups with daily fluctuations again began to resemble that of the control group (Fig. 4d, ANOSIM, R = 0.119, p > 0.05).
Fig. 4. Principal coordinates analysis (PCoA) of Litophyton-associated bacterial communities across three temperature exposure groups. Bray-Curtis dissimilarity was calculated based on the relative abundance of bacterial ASVs. Panels (a) to (d) correspond to Day 7, 14, 21 and 28, respectively. Statistical analysis was performed using ANOSIM
Comparison of Bacterial Compositions in Litophyton sp. Under Different DTF
The bacterial compositions of Litophyton sp. under different temperature conditions were compared. Analysis of the top 20 abundant bacterial genera revealed that Endozoicomonas was overwhelmingly dominant, followed by the BD1-7 clade (Cellvibrionales), Terasakiellaceae (family level), Tenacibaculum, and others (Fig. 5). Notably, the ± 7 °C group exhibited a higher relative abundance of Endozoicomonas compared to both the control and ± 5 °C groups. To identify which bacteria stably associate with Litophyton sp. under different DTF regimes, we analyzed the core microbiome (defined as ASVs present in at least 80% of samples within each treatment). We identified 227,564 reads of core ASVs in the control group, 211,678 reads in the ± 5 °C group, and 170,985 reads in the ± 7 °C group, among which Endozoicomonas and the BD1-7 clade together accounted for 70–80% of the core bacterial community (Fig. 6, Table S8). Specifically, Endozoicomonas was 29.4% more significantly abundant in the ± 7 °C group (136,403 reads) compared to the control (114,743 reads) (Wilcox pairwise post-hoc test, p < 0.05, Table S9). It was also 24.4% more abundant than in the ± 5 °C group (117,360 reads), although this difference was not statistically significant (Wilcox pairwise post-hoc test, p > 0.05, Table S9). In contrast, the BD1-7 clade was most abundant in the ± 5 °C group (73,014 reads), showing a 13.5% significant increase over the control (47,857 reads) (Wilcox pairwise post-hoc test, p < 0.05, Table S10). It was also 20.2% higher than in the ± 7 °C group (24,530 reads), although this difference was not statistically significant (Wilcox pairwise post-hoc test, p > 0.05, Table S10, Fig. 6). These findings suggest that Endozoicomonas and the BD1-7 clade are stable associated bacteria of Litophyton sp., under both fluctuating diurnal temperature, with Endozoicomonas representing the most abundant core genus, potentially playing a key role in maintaining the microbial stability of the host.
Fig. 5. Relative abundance profiles of the top 20 bacterial genera in three different temperature exposure groups. Genus-level bacterial composition associated with Litophyton sp. in Experiment II. Each bar represents the mean relative abundance of three samples per group at each time point. ASVs representing less than 1 % of total relative abundance were grouped as “Others”. Colors represent different bacterial genera
Fig. 6. Taxonomic composition of the core bacterial genera identified in three different temperature exposure groups. Bacterial composition at the genus level in Litophyton sp. across all time points of Experiment II is shown. ASVs were selected if present in at least 80 % of the samples from each treatment. Colors represent different bacterial genera
Phylogenetic Comparison of Endozoicomonas Associated with Litophyton sp. Under Varying DTF regimes
The core microbiome was predominantly composed of the genus Endozoicomonas, which accounted for approximately 50–80% of its relative abundance. Within this genus, ASV5 was highly abundant in the core microbiome across all three experimental groups (control: 35.09%; ± 5 °C: 42.68%; ± 7 °C: 50.89%), with its relative abundance increasing as the magnitude of temperature fluctuations grew. We constructed a phylogenetic tree to determine the taxonomic placement of the Endozoicomonas ASVs identified in our samples, using previously isolated Endozoicomonas strains from various marine hosts as references. Due to the short length of the aligned region (408 bp), all internal nodes showed low bootstrap support. The analysis revealed that ASV5 is more closely related to strains isolated from Scleractinia (scleractinian corals), whereas the ASV2, 3 and 6, the other ASVs which were found to be abundant in ± 7 °C treatment group, cluster more closely with strains derived from octocorals (Fig. 7).
Fig. 7. Maximum likelihood Phylogenetic tree of Endozoicomonas ASVs associated with Litophyton sp. under different temperature fluctuation treatments. ASV1-6 indicates sequences from three temperature treatments and selected from the core microbiome of three temperature treatments. Different color indicated the group in which each ASV exhibits its highest relative abundance. Statistical analysis was conducted by Maximum Likelihood (ML) method, and the tree is rooted using sequences from the V6-V8 region of Marinobacterium nitratireducens and Simiduia agarivorans
Discussion
Large Diurnal Temperature Fluctuations may Influence the Physiological Stability of the Octocoral Litophyton sp.
Corals inhabiting intertidal zones in coastal and estuarine environments are regularly exposed to highly variable environmental conditions, particularly in temperature [61–63], and have consequently developed the capacity to tolerate substantial thermal fluctuations [64–66]. Previous studies suggest that such fluctuations may have important ecological implications. For instance, both Safaie et al. [16] and Hsieh et al. [17] reported that DTF can mitigate coral heat accumulation in both field and controlled experiments. Similarly, corals inhabiting dynamic environments may exhibit enhanced thermal tolerance [14, 67–69]. Keshavmurthy et al. [70] further demonstrated that corals living in high-latitude, non-reef environments are able to adapt to substantial temperature fluctuations, showing no signs of thermal stress such as reduced photochemical efficiency, decreased Symbiodiniaceae density, or loss of host soluble protein under conditions typically associated with bleaching.
In this study, we investigated Litophyton sp., a shallow-water octocoral collected from Kueishan Island in northern Taiwan, a high-latitude, non-reef environment characterized by pronounced seasonal and large temperature fluctuation throughout the year. By assessing physiological parameters, including coral color scores, photosynthetic efficiency of symbiotic algae (Fv/Fm), and holobiont reactive oxygen species (ROS) activity, our results demonstrated that DTF significantly influence the physiological state of Litophyton sp., as evidenced by shifts in antioxidant enzyme activities, yet this impact is confined to a specific range of thermal variance. Specifically, the observed heat stress was more pronounced under moderate fluctuations (e.g., ± 5 °C), whereas larger DTF appeared to alleviate thermal stress, further indicating that Litophyton sp. may possess adaptive strategies enabling its successful persistence in dynamic environments.
The observed alleviation of thermal stress under large DTF could potentially be attributed to a process of interrupted heat accumulation and subsequent physiological recovery. It is hypothesized that the lower temperatures during the nighttime phase in a large DTF cycle may help to provide a critical recovery that offsets the ROS accumulation incurred during peak temperatures at midday [71]. This nighttime cooling might effectively neutralize ROS, preventing the cumulative cellular damage that typically occurs under stable elevated temperatures [71–73]. Despite these clear trends in Litophyton sp., the broader physiological responses of octocorals to dynamic thermal conditions remain complex and often species-specific [71, 74]. While our study highlights the benefits of DTF in alleviating thermal stress, other researchers suggested that these outcomes are highly dependent on the environmental context. For example, a study on the octocoral Muricea laxa found that prolonged thermal fluctuations around an elevated mean temperature significantly increased mortality and metabolic suppression, suggesting that the timing and duration of the dynamic stress are critical factors [15]. Furthermore, the interpretation of stress biomarkers in octocorals is complicated by internal physiological cycles. Studies on octocoral Virgularia cynomorium, for instance, found that the seasonal decrease in antioxidant biomarkers (e.g., CAT, SOD, LPO) during warmer months was primarily linked to energy reallocation for spawning and fluctuations in food availability, rather than acute thermal stress [19]. These findings emphasize the necessity of considering the species’ annual physiological cycle and energy budget when evaluating thermal stress responses based on molecular biomarkers in octocorals.
Endozoicomonas Associated Litophyton sp. Under Large Diurnal Temperature Fluctuations
The coral-associated bacteria play a crucial role in holobiont health, metabolism and stress resilience [75–78]. Monitoring changes in these bacterial communities can yield valuable insights into the physiological responses of octocoral holobionts under varying environmental conditions [79–81]. In this study, beta diversity analyses by Day 21 revealed distinct microbial shifts, which coincided temporally with significant changes in host antioxidant activities (SOD and CAT). This temporal alignment suggests that microbial community changes may precede or act as an early warning mechanism before significant host physiological stress is observed. This is a critical finding, suggesting a potential link between shifts in microbial composition and antioxidant properties [81], although a direct causal relationship between bacterial functions and ROS levels remains to be determined.
The core microbiome of octocorals comprises bacterial taxa stably associated with the host, likely involved in essential holobiont functions [32, 76]. However, current available data are mainly focused on temperate gorgonian species inhabiting the Mediterranean Sea. Here, we expand this knowledge base by characterizing the microbial community of the shallow-water Alcyonacea Litophyton sp. under large DTF. Our results show that Endozoicomonas is a persistent and dominant member of the Litophyton holobiont, maintaining high relative abundance across treatments. This stable presence suggests that Endozoicomonas plays a functional role in sustaining holobiont stability.
Endozoicomonas is broadly distributed in both scleractinian and octocoral symbioses and it is believed to have co-evolved with its coral hosts [82–84]. often contributing to health and functional coral health and functional stabilization [82, 85]. One well-characterized trait reported in some lineages is the ability to degrade dimethylsulfoniopropionate (DMSP) into dimethylsulfide (DMS), which can act as an effective antioxidant and has been proposed to mitigate excessive ROS in corals [86, 87]. Although we did not conduct functional or genomic analyses in this study, the phylogenetic affiliation of the Endozoicomonas ASVs detected here places them close to lineages previously reported to carry DMSP-degradation genes, suggesting that they may possess similar metabolic potential. In our experiment, the ± 7 °C group exhibited both a higher relative abundance of Endozoicomonas and lower SOD and CAT activities compared with the ± 5 °C group. While this pattern is correlative and does not establish causation, it raises the possibility that Endozoicomonas may contribute (directly or indirectly) to modulate oxidative stress in Litophyton sp. under fluctuating temperatures. Further metagenomic or transcriptomic analyses are needed to determine whether the functional capacities documented in other hosts are conserved in the Litophyton-associated lineages observed here.
Potential Function of the Core Microbiome Associated to Litophyton sp.
Litophyton sp. may modulate its reliance on heterotrophic versus autotrophic nutrition in response to environmental conditions, exhibiting high trophic plasticity utilizing carbon from diverse sources with varying δ¹³C values [88]. Enhanced heterotrophic feeding may prevent carbon limitation under heat stress, thereby improving thermal resistance [89–91]. A recent study revealed that Endozoicomonas possess endo-chitinases and other genes involved in chitin degradation, suggesting that this bacterial group may play an important role in chitin processing in octocorals [92]. Chitin is primarily derived from marine zooplankton, a well-established food source for octocorals [93–95]. Hsu et al. [88] monitored δ^13^C and δ^13^N levels in Litophyton from the same region and showed that Litophyton favor larger zooplankton.
The second most abundant bacterial group was Cellvibrionales BD1-7, a member from oligotrophic marine Gammaproteobacteria (OMG) group [96]. This clade has also been reported as a dominant member of Mediterranean gorgonians [32, 76] and Alcyonacea Litophyton in this study. BD1-7 bacteria are believed to be capable of using light to produce ATP via proteorhodopsin proton pumps, serving as an alternative energy source for mixotrophic growth [97, 98]. Although their specific role within the host remains unclear, these traits may explain why they are the second most abundant core bacterial group in the shallow-water, mixotrophic octocoral Litophyton sp.
Phylogenetic and Ecological Significance of Endozoicomonas Shifts
Due to the various beneficial roles of different Endozoicomonas, members of this genus exhibit host specificity and gene complexity across different environmental scales [76, 83, 99]. Our phylogenetic analysis of Endozoicomonas ASVs associated with Litohyton sp. suggested that different Endozoicomonas members may reflect the physiological status of their host.
Among the Endozoicomonas ASVs identified in our analysis, ASV1 and ASV5 strains were abundant in samples from the control and the ± 7 °C group. These ASVs formed a clade with E. acroporae Acr-14, which was isolated from the scleractinian coral Acropora sp [100]. E. acroporae Acr-14 has been functionally characterized and is capable of metabolizing DMSP into DMS [28]. Furthermore, certain bacterial operational taxonomic units (OTUs) related to E. acroporae and isolated from different sites in Penghu, Taiwan, exhibited a high potential to scavenge ROS and greater flexibility in adapting to environment changes [101]. Based on these studies, E. acroporae and its related bacteria may be considered probiotics for corals. Our findings corroborate this idea, potentially explaining why ASVs in this clade were abundant in the groups with lower SOD and CAT activity than the ± 5 °C group.
ASV2, 3 and 6 were found to be abundant in ± 7 °C treatment group. These ASVs formed a clade with ASVs that were phylogenetically close to E. euniceicola and E. gorgoniicola. E. euniceicola and E. gorgoniicola, which were isolated from the octocoral Eunicea fusca and Plexaura sp [102]. Moreover, these species were the first to yield genome sequences of a culturable octocoral-isolated Endozoicomonaceae strains [103]. The ASVs were also phylogenetically close to Endozoicomonas spp. isolated from Litophyton [104]. According to a mini-review published by Neave et al. and other studies, the suggested functions of Endozoicomonas found on octocoral gorgonians are related to host health [76], like protein, carbohydrate transport and cycling [105], or disease [106]. Although the specific roles of the ASVs of Endozoicomonas remain unclear, we propose that each may represent an octocoral-specific Endozoicomonas species. Consequently, this study is the first to show that DTF can affect bacteria similar to E. euniceicola and E. gorgoniicola.
In summary, the phylogenetic analysis provides valuable insights into the functional roles of different clades of Endozoicomonas in octocorals, and also allows us to detect subtle, fine-scale differences in host physiological states that may not be captured by other physiological approaches.
Conclusion
In this study, we investigated the holobiont of the octocoral Litophyton sp. through controlled experiments and confirmed that Litophyton sp. exhibit physiological stress responses under specific short-term temperature fluctuations. Notably, its bacterial community, particularly the core microbiome Endozoicomonas, underwent compositional shifts that aligned with detectable physiological changes in the host. We propose that variations in Endozoicomonas may support the host in coping with early thermal stress. This study not only helps explain the persistence of Litophyton sp. in shallow-water environments with high environmental variability but also deepens our understanding of the phylogenetic diversity of the octocoral core microbiome and its role in influencing holobiont metabolic stability. Furthermore, our findings suggest that the diversity of Endozoicomonas species may provide some potential functional redundancy that helps maintain the health status of the Litophyton sp. holobiont, particularly under thermal stress.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary File 1 (JPG 3.25 MB)
Supplementary File 2 (JPG 7.06 MB)
Supplementary File 3 (JPG 815 KB)
Supplementary File 4 (DOCX 17.0 KB)
Supplementary File 5 (DOCX 49.1 KB)
Supplementary File 6 (DOCX 19.4 KB)
Supplementary File 7 (XLSX 20.4 KB)
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Quigley KM, Baird AH (2024) Future climate warming threatens coral reef function on world heritage reefs. Glob Change Biol 30(7):e 17407. 10.1111/gcb.1740710.1111/gcb.1740739011806 · doi ↗ · pubmed ↗
- 2Johnson JV et al (2022) The relative influence of sea surface temperature anomalies on the benthic composition of an Indo-Pacific and Caribbean coral reef over the last decade. Ecol Evol 12(9):e 9263. 10.1002/ece 3.926310.1002/ece 3.9263 PMC 944896536091340 · doi ↗ · pubmed ↗
- 3Liberman R et al (2025) Octocoral dynamics over a decade on Florida’s coral reef. Sci Rep 15(1):37635. 10.1038/s 41598-025-21610-510.1038/s 41598-025-21610-5PMC 1256901841152329 · doi ↗ · pubmed ↗
- 4Prioux C et al (2023) Unveiling Microbiome changes in mediterranean octocorals during the 2022 marine heatwaves: quantifying key bacterial symbionts and potential pathogens. Microbiome 11:271. 10.1186/s 40168-023-01711-x 10.1186/s 40168-023-01711-x PMC 1069676538053218 · doi ↗ · pubmed ↗
- 5Jones R et al (2020) Responses of corals to chronic turbidity. Sci Rep 10:4762. 10.1038/s 41598-020-61712-w 10.1038/s 41598-020-61712-w PMC 707592232179846 · doi ↗ · pubmed ↗
- 6Willis C, Desai D, La Roche J (2019) Influence of 16S r RNA variable region on perceived diversity of marine microbial communities of the Northern North Atlantic. FEMS Microbiol Lett 366(13)10.1093/femsle/fnz 152PMC 667376931344223 · doi ↗ · pubmed ↗
- 7Mc Murdie PJ, Holmes S (2014) Waste Not, want not: why rarefying Microbiome data is inadmissible. P Lo S Comput Biol 10(4):e 1003531. 10.1371/journal.pcbi.100353110.1371/journal.pcbi.1003531 PMC 397464224699258 · doi ↗ · pubmed ↗
- 8Weiss S et al (2017) Normalization and microbial differential abundance strategies depend upon data characteristics. Microbiome 5:27. 10.1186/s 40168-017-0237-y 10.1186/s 40168-017-0237-y PMC 533549628253908 · doi ↗ · pubmed ↗