Mitigation of Vibrio coralliilyticus-induced coral bleaching through bacterial dysbiosis prevention by Ruegeria profundi

ABSTRACT

Vibrio species are prevalent in ocean ecosystems, particularly Vibrio coralliilyticus, and pose a threat to corals and other marine organisms under global warming conditions. While microbiota manipulation is considered for coral disease management, understanding the role of commensal bacteria in stress resilience remains limited. Here, a single bacterial species (Ruegeria profundi) rather than a consortium of native was used to combat pathogenic V. coralliilyticus and protect corals from bleaching. R. profundi showed therapeutic activity in vivo, preventing a significant reduction in bacterial diversity in bleached corals. Notably, the structure of the bacterial community differed significantly among all the groups. In addition, compared with the bleached corals caused by V. coralliilyticus, the network analysis revealed that complex interactions and positive correlations in the bacterial community of the R. profundi protected non-bleached corals, indicating R. profundi’s role in fostering synergistic associations. Many genera of bacteria significantly increased in abundance during V. coralliilyticus infection, including Vibrio, Alteromonas, Amphritea, and Nautella, contributing to the pathogenicity of the bacterial community. However, R. profundi effectively countered the proliferation of these genera, promoting potential probiotic Endozoicomonas and other taxa, while reducing the abundance of betaine lipids and the type VI section system of the bacterial community. These changes ultimately influenced the interactive relationships among symbionts and demonstrated that probiotic R. profundi intervention can modulate coral-associated bacterial community, alleviate pathogenic-induced dysbiosis, and preserve coral health. These findings elucidated the relationship between the behavior of the coral-associated bacterial community and the occurrence of pathological coral bleaching.

IMPORTANCE

Changes in the global climate and marine environment can influence coral host and pathogen repartition which refers to an increased likelihood of pathogen infection in hosts. The risk of Vibrio coralliilyticus-induced coral disease is significantly heightened, primarily due to its thermos-dependent expression of virulent and populations. This study investigates how coral-associated bacterial communities respond to bleaching induced by V. coralliilyticus. Our findings demonstrate that Ruegeria profundi exhibits clear evidence of defense against pathogenic bacterial infection, contributing to the maintenance of host health and symbiont homeostasis. This observation suggests that bacterial pathogens could cause dysbiosis in coral holobionts. Probiotic bacteria display an essential capability in restructuring and manipulating coral-associated bacterial communities. This restructuring effectively reduces bacterial community virulence and enhances the pathogenic resistance of holobionts. The study provides valuable insights into the correlation between the health status of corals and how coral-associated bacterial communities may respond to both pathogens and probiotics.

INTRODUCTION

Recurring bleaching events, both local and global in scale, have exacerbated the increasingly disastrous environmental crisis of coral degradation (1). It is widely recognized that coral reef ecosystems are experiencing unparalleled threats prompted by human pressure, such as global climatic stress, pollution, and overfishing (2), resulting in the loss of coral habitats with profound consequences for the survival of the coral ecosystem (3). Given the backdrop of environmental disturbances and rising exposure to multiple stressors, coral opportunistic pathogens can proliferate, resulting in epidemics characterized by elevated mortality rates (4). Natural recovery processes in corals are exceedingly slow and are often constrained by unknown factors (5), implying that coral reef ecosystems are becoming threatened with extinction (6). Therefore, there is an urgent need to establish effective approaches to protect the coral holobionts against disease.

Vibrio coralliilyticus, as one of the causative agents of epidemic vibriosis in marine organisms, can induce bleaching, tissue loss, and mortality of corals by producing virulence factors like zinc-metalloprotease (7). A recent study has showed that pathogenic bacterial infections induced bleaching in scleractinian corals in the Gulf of Aqaba due to exposure to non-anthropogenic factors, with V. coralliilyticus identified as the disease-causing agent (8). It also has been implicated in coral disease across numerous reefs worldwide, exhibiting an increase in both frequency and intensity, spanning from Caribbean to Mediterranean Sea, Indo-Pacific, Indian Ocean, Great Barrier Reef, Micronesia, Polynesia, and South China Sea (9–14). Ongoing global warming and environmental pressures have led to a high abundance and increased virulence of V. coralliilyticus (15). Moreover, pathogenic V. coralliilyticus infection among corals occurs through various transmission mechanisms, such as direct contact, waterborne transmission, pollution, and infection via other marine organisms (16). Therefore, it is crucial to prevent the proliferation of Vibrio species within corals, curtail the transmission of host-derived pathogens, and mitigate the potential for disease expansion as essential strategies for protecting localized coral reefs.

Coral microbiome has garnered growing attention due to its stability and flexibility, which are intricately linked to host health (17). The disruption of the microbiome within the coral holobionts caused by environmental changes is considered as a mechanism for coral degradation (18). Furthermore, there is evidence indicating that interactions among stressors have the potential to disrupt both the hosts and the coral-associated bacterial community (4). Particularly, the loss of dinoflagellate symbionts, commensal bacteria, and reduced microbiome stability occurs when corals are infected by V. coralliilyticus, leading to the disintegration of the symbiosis relationship between the coral host and its symbionts (19, 20). Despite that several Ruegeria strains have been identified as capable of inhibiting the growth of V. coralliilyticus through inhibition zone methods, it remains uncertain whether these species play a specific role in protecting corals from pathogens. Notably, R. profundi was selected to verify its probiotic function primarily because of the lack of prior investigation on this microorganism.

In this study, we hypothesized that R. profundi may act as a guard that can protect coral health by manipulating coral-associated bacterial community and mitigating the effects of pathogens. To confirm this hypothesis, R. profundi and V. coralliilyticus strains were isolated separately from corals, using Acropora sp. as an animal model. The selection of Acropora sp. was driven by their endangered survival status, with disease cited as a vital driver for their heightened risk of extinction (21, 22). To the best of our knowledge, this study is the first to demonstrate the potential protective efficacy of probiotic R. profundi against pathogenic V. coralliilyticus. 16S rRNA-sequencing was employed to examine the alterations in bacterial community diversity and structure, dominant and core bacteria, and co-occurrence relationships of among different treatments. These analyses were aimed at elucidating the diverse responses of coral-associated bacterial communities to pathogen-induced stress and the interactions between pathogen and probiotic. Considering the significance of the microbiome in normal and altered host physiological states (23), functional pathways and genes related to bacterial community pathogenicity were characterized through metagenomic sequencing. This study contributes to our understanding of the concrete characterization of bacterial dysbiosis in bleached corals to reveal the potential mechanisms of actions for probiotics in addressing coral-associated bacterial imbalance. It also provides a theoretical basis for managing pathogenic invasions and enhancing stress tolerance in corals.

MATERIALS AND METHODS

Bacterial isolation, culture and identification

The healthy Acropora carduus fragments employed in this study were collected from Luhuitou fringing reef, located in Sanya, southern coast of Hainan Island (18°12′N, 109°28′E). To isolate the coral-associated bacteria, 3 g of coral samples was homogenized by mechanical grinding with a sterile porcelain pestle and mortar. Coral tissue homogenate was then added to 20 mL of sterile seawater in a test tube, where it was vigorously shaken for 1 min. The sample was serially diluted (up to 10⁴ dilution) with sterile seawater. Then, 20 µL of the diluted sample was spread onto a 2216E agar plate (containing 0.5% yeast extract, 0.1% tryptone, 2% agar, and 0.001% FePO₄·4H₂O in seawater). The plates were incubated at 28℃ for 72 h. Colonies were randomly selected using a numbered grid, sub-cultured on 2216E plates, and incubated for 48 h at 28°C.

For bacterial identification, the 16S rRNA genes were amplified using colony PCR with the universal bacterial primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) for all bacterial colonies. PCR was performed in a C1000 Touch thermal cycler under the following conditions: an initial denaturing step of 1 cycle at 94°C for 2 min; 32 cycles of 94°C for 20 s, 58°C for 30 s, and 72°C for 45 s; and a final extension at 72°C for 10 min. After verification of PCR amplification by 1.2% agarose gel electrophoresis, PCR products were sequenced by Aiji Biotechnology company (Guangzhou, China). The obtained sequence information was subjected to comparison with known sequences in the NCBI Database using BLAST analysis (http://www.ncbi.nlm.nih.gov/BLAST/). All purified strains were stored in a 20% (vol/vol) glycerol solution at –80°C. Rejuvenate strains were subsequently reinoculated on 2216E media at 28°C overnight.

Preparation of bacterial strains

The possible probiotic R. profundi and pathogenic V. coralliilyticus were incubated for 48 h in 2216E media at 28°C with an oscillating speed of 185 rpm. Aliquots of bacterial cultures were collected at 0, 3, 6, 12, 18, 24, 30, 36, and 48 h to assess cell growth by measuring the optical density (OD) at 600 nm using a UV spectrophotometer. Simultaneously, 10 mL of bacterial solution at these specific time points underwent 10-fold serial dilution and was inoculated onto 2216E agar plates for viable count analysis, yielding colony-forming unit values (c.f.u.mL⁻¹).

The two bacterial strains were cultured in 250 mL of 2216E medium at 28°C. When they reached the stationary growth phase, the bacteria were centrifuged at 2,000 g for 4 min. Based on the infection trails in previous studies (24–26) using V. coralliilyticus strain, it was determined that the concentration range of 10⁶–10⁸ c.f.u.mL⁻¹ could lead to tissue loss and bleaching. Therefore, preliminary experiments were conducted to ascertain the bacterial concentration for incorporation. The bacterial cells were collected, resuspended in sterile seawater, subjected to an additional round of centrifugation to remove any residual culture medium, and finally resuspended in 1 mL of sterile seawater, resulting in a final concentration of 1 × 10⁷ c.f.u.mL⁻¹, ready for coral infection experiments.

Coral infection experimental design

A. carduus fragments were transported to the laboratory in aerated plastic bags and initially placed in an 800 L bucket with seawater at 24°C for 4 h for acclimation. Coral branches (approximately 4 cm in length) were cut using a cutting machine and maintained in 18-L tanks with 12 L of seawater prepared from Instant Ocean Sea Salt (Instant Ocean, Blacksburg, VA) to achieve a salinity of 35‰. The room temperature and the seawater temperature were kept at 21°C and 24°C, respectively. During the 7-day acclimatization period, 15% of the seawater in the tanks was refreshed prior to the formal experiment. All fragments received illumination from a blue LED light source, following natural day-night light cycles based on the local sunrise (6:30–6:45 a.m.) and sunset (5:40–6:02 p.m.) times.

The coral branches were randomly distributed in the 12 tanks with each tank hosting six branches. Additionally, three parallel controls were established. One control group consisted of volumetric seawater (C group), while the other two were exposed to 1 mL of V. coralliilyticus (Y group) and R. profundi (D group), respectively. A third group was inoculated with V. coralliilyticus for 3.5 h, after which R. profundi was inoculated into (DY group). The bacterial solution was carefully administered to ensure a final concentration of 1 × 10⁷ c.f.u.mL⁻¹. The bacterial solution was gently introduced to the surrounding water of coral fragments using 5-mL sterile syringes in each treatment group. The health state and color of the coral tissues were assessed from images captured for each fragment on 6 consecutive days following injection.

DNA extraction and sequencing analysis

As the conclusion of the bacterial infection experiment, six coral fragments were sampled from each group, yielding a total of 24 samples collected from both treatment and control groups. All coral fragments were rapidly pulverized in a sterile mortar and pestle using liquid nitrogen. Subsequently, total DNA was extracted from the coral fragments using the Dneasy PowerSoil Pro Kit (Qiagen, Hilden, Germany) for the construction of 16S rRNA gene amplicon sequence libraries and metagenomic analysis.

The V4 hypervariable region of the 16S rRNA gene was amplified using the primers 515FomdF (5′-GTGYCAGCMGCCGCGGTAA-3′) and 806RmodR (5′-GGACTACNVGGGTWTCTAAT-3′) (27, 28). The amplification products were purified using the E.Z.N.A. Gel Extraction Kit (Omega, USA) and subsequently sequenced using the Illumina MiSeq2500 platform at a commercial sequencing company (Majorbio Bio-Pharm Technology Co. Ltd., Shanghai, China). The high-quality 16S rRNA gene sequences underwent demultiplexing and quality filtering using QIIME (version 1.9.1). Taxonomic classification was performed using the Silva (version 138) 16S rRNA gene database with a confidence threshold of 70, and then, the Uparse 11 algorithm was utilized to cluster the filtered sequences into operational taxonomic units (OTUs) with a 97% similarity cutoff (29).

The Illumina sequencing libraries were constructed using the NEXTFLEX Rapid DNA-Seq Kit and Illumina NovaSeq Reagent Kits. The metagenomic sequencing dataset-pair of each experimental group was co-assembled into contigs and analyzed with MEGAHIT software (version 1.1.2). Raw reads data were trimmed to remove low-quality sequences with Fastp (version 0.20.0). Gene prediction was executed using Prodigal (version 2.6.3) (https://github.com/hyattpd/Prodigal). After the compilation of all predicted genes from metagenomic scaffolds, a non-redundant gene catalog was generated through clustering with CD-HIT (version 4.7) with a 90% sequence identity threshold. The relative abundance of specific genes in the corresponding samples was determined with a 95% identity threshold using SOAPaligner (version 2.21) (30). Taxonomic sources of the non-redundant gene were annotated by aligning high-quality bacterial reads against the NCBI non-redundant protein sequences (NR) database using DIAMOND with an E-value < 1e⁻⁵. Furthermore, functional annotation of genes was based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. From the annotated sequences, genes related to betaine lipids and Type VI section system (T6SS) were identified through the search for KEGG Orthology identifiers (KO IDs).

Statistical analysis

To describe the effects of the pathogen and potential probiotic bacteria on the coral-associated bacterial community pattern, a series of analyses were conducted. Alpha-diversity index analysis was performed on the sequencing data to assess community richness (Chao index) and diversity (Shannon index). Beta-diversity index was conducted through Bray-Curtis distance-based nonmetric multidimensional scaling analysis (NMDS) and partial least-squares discriminant analysis (PLS-DA) based on OTUs, utilizing the “vegan” and “mixOmicis” packages in R software (version 3.3.1). The coral-associated bacterial Venn diagram and column chart were generated using the “VennDiagram” package in R. For each treatment group, two stacked bar plots were generated at class and order taxonomic levels. The visual circular diagram was drawn using Circos-0.67-7 (http://circos.ca/) to illustrate the corresponding relationship between treatment groups and genera, offering a visual representation of the dominant bacterial genera within each group. Linear discriminant analysis effect size (LEfSe) was employed to estimate differences in the effect size of genera among all groups, applying a threshold LAD score > 4 and 3.

To assess significant differences in the relative abundances of bacterial genera, betalain biosynthesis, and bacterial secretion system metabolic pathways among groups, the nonparametric Kruskal-Wallis (KW) H test and Tukey-Kramer post hoc tests with false discovery rate (FDR) correction for multiple comparisons were employed. A significance level of P = 0.05 and a confidence level of 95% were considered statistically significant. Co-occurrence networks among the top 100 abundant bacterial genera were constructed using Networkx software (version1.11) with a pairwise Spearman correlation coefficient > 0.5 and P < 0.05 values as the threshold. These networks were visualized using Gephi software (version 0.9.2). Genes related to the choline/glycine/proline betaine transport protein (betaine lipids) and T6SS were identified by searching KO IDs and visualized as a Z-score heatmap with row-normalized values.

RESULTS

Enhancing coral survival with R. profundi against V. coralliilyticus-induced bleaching

To test the effectiveness of R. profundi within coral holobionts under conditions of elevated V. coralliilyticus in the hosts’ surrounding environment (seawater), a diffusion-based coral infection experiment was conducted. For a more tangible illustration, Fig. 1A and B depicted the experimental design and procedure. Prior to injection, coral fragments were acclimated for 3 days at 24°C in a sterilized seawater with a salinity of 35‰ and until no coral bleaching was observed in any aquarium. The coral health status is depicted in Fig. 1C. Half of the coral fragments exhibited acute rapid tissue necrosis, resulting in bare skeletons within 2 days after inoculation with V. coralliilyticus. Over time, prolonged exposure to V. coralliilyticus led to a visible sign of extensive tissue sloughing and bleaching in the coral fragments. In contrast, the other groups, including those injected with seawater (blank group as the control), R. profundi, and a combination of R. profundi and V. coralliilyticus, did not exhibit impaired polyp tissue integrity. These findings suggested that R. profundi played a significant role in supporting the host’s defense against pathogenic V. coralliilyticus infection.

Fig 1.

The experimental design overview and representative responses of A. carduus to bacterial introduction. (A) Schematic of the bacterial infection setup and the experimental design. (B) Bacteria were inoculated into the surrounding of A. carduus using the injection syringe at 0 days. Photographs were taken daily on 6 consecutive days using natural light. (C) Photos of the healthy states and tissue discoloration in A. carduus across the various treatment groups. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

Coral-associated bacterial community profiles regulated by R. profundi and V. coralliilyticus

To further confirm our hypothesis, a 16S rRNA gene amplicon sequencing was conducted to evaluate and compare bacterial communities within corals from four experimental setups. A total of 1,486,334 optimized sequence reads were obtained from the 24 samples, comprising 376,024,991 bases with an average read length of 252 bases. These aligned sequences were then clustered into 1,712 OTUs. The DY group had the highest number of OTUs (1,265), followed by the C group (1,102 OTUs) and the D group (1,078 OTUs), whereas the Y group exhibited the lowest number of OTUs (827).

The Venn diagram in Fig. 2A showed the distribution of unique OTUs in each treatment group and the degree of overlap among groups. In this context, 529 (30.33%) OTUs were shared among all groups. Notably, bleached coral samples (Y group) contained fewer unique OTUs (42, 2.45%) compared with non-bleached coral samples (D and DY groups), which possessed more unique OTUs (159, 9.29% and 206, 12.04%), as well as the blank samples (C group: 91, 5.32%). This analysis from the Venn diagram indicated that V. coralliilyticus can reduce the richness of the coral-associated bacterial community and this impact from V. coralliilyticus can be mitigated by R. profundi.

Fig 2.

Characterization of the alpha-diversity and structure of coral-associated bacterial community in V. coralliilyticus-infected corals treated with R. profundi. (A) Venn diagram comparison of the shared and unique number of OTUs among the C, D, Y, and DY groups. The numbers and percentages displayed above the Venn diagrams representing the number of overlapping and distinct OTUs. (B) Chao 1 index was calculated for the determination of bacterial community richness. (C) Shannon index was calculated for the determination of bacterial community diversity. (D–G) NMDS plot of the bacterial community from all samples among four treatments using the Bray-Curtis dissimilarity matrix. (H) Classification of coral-associated bacterial communities using PLS-DA. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

As shown in Fig. 2B and C, when evaluating bacterial species evenness and species richness among all treatment groups, the alpha-diversity of coral-associated bacteria, as determined by the Chao 1 and Shannon indices, significantly decreased in the bleached corals compared with the control group. No significant difference was found between the other two treatment groups and the control group. These results suggested that coral-associated bacterial species evenness and species richness decreased after V. coralliilyticus infection, but this alteration was mitigated when R. profundi was introduced into coral hosts’ surrounding water. However, OTU-based NMDS analysis (Fig. 2D through G) revealed that the variation observed in the coral-associated bacterial community structure of three treatment groups significantly differed from that of the control groups (P < 0.01). Additionally, PLS-DA (Fig. 2H) was conducted at the OTU level. The analysis results revealed that the distribution characteristics of all groups were relatively homogeneous within each group but distinct among groups. Three distinct clusters were formed, suggesting that the coral-associated bacteria of the D, Y, and DY groups were clearly separated into three relatively independent clusters. In contrast, the bacteria from non-bleached coral groups (including C, D, and DY groups) clustered together. These results indicated that bacterial infection had a significant impact on the coral-associated bacteria by altering the bacterial community structure at OTU level.

After obtaining the abundance and annotation data for OTUs, these data were normalized based on the minimum sample sequence number. The sequence reads were categorized into 2 domains (archaea and bacteria) spanning 2 kingdoms, 35 phyla, 76 classes, 184 orders, 296 families, 513 genera, and 806 species to determine the community composition within each group. The most dominant phyla at the top five levels were Proteobacteria, Verrucomicrobiota, Bacteroidota, Campilobacterota, and Planctomycetota. The classes with the highest relative abundance across all samples were Gammaproteobacteria and Alphaproteobacteria, both belonging to the Proteobacteria phylum (Fig. 3A). These two classes were increased in both Y and DY groups, while the Verrucomicrobiae class decreased in these groups compared with the control group, which is contrast to the D group. As show in Fig. 3B, there was a shift in the dominated order in Y and D groups, unlike in C and DY groups. Oceanospirillales dominated in both C (23.18%) and DY (37.05%) groups, while Y and D groups were dominated by Alteromonadales (31.03%) and Verrucomicrobiales (26.92%), respectively. The relative abundance of the Rhodobacterales order increased in three groups (D: 17.73%, Y: 24.40%, and DY: 15.02%) in comparison with C group (12.13%).

Fig 3.

Changes in bacterial community composition of four treatment corals. (A and B) Mean relative abundances of the bacteria taxa, at least 1% on class level and order level, separately. (C) Circos plot showing the distribution of the most abundant bacteria in the different groups of coral-associated bacteria at genus level and the proportion of each dominant bacteria. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

The Circos graph (Fig. 3C) related to coral bacterial symbionts depicted the proportional distribution of dominant species among the treatments and their distribution within each treatment at the genus level. Endozoicomonas, Rubritalea, and Ruegeria were dominant in the C, D, and DY groups, while Nautella (16%) and Thalassotalea (15%) dominated in the Y group. The relative abundance of Endozoicomonas in the DY group nearly doubled, reaching 32.86%. The dominant Rubritalea genus in the D group increased from 10.20% to 25.47%. In the Y group, the relative abundance of Vibrio reached 10.09%, becoming the predominant genus, while its relative abundance ranged from 0.1% to 0.3% in the other three groups. Both Nautella and Alteromonas increased by approximately 5–5.6-fold, from 3.20% and 1.82% to 15.65% and 10.24%, respectively. The relative abundance of Ruegeria was higher in three treatment groups compared with the control groups, especially in the D and DY groups. In summary, these results suggested that the introduction of cultured bacteria can alter the composition of the coral-associated bacterial community by influencing the abundance of core bacteria.

Effects of V. coralliilyticus and R. profundi treatments on the biomarker genus

To identify distinct coral-associated bacteria (marker species) in the four groups, LEfSe multi-level species difference discriminant analysis was employed (with a linear discriminant analysis (LDA) score threshold of 2 at the phylum-to-genus classification level, as shown in Fig. 4A). A histogram displaying the distribution of LDA values was used to emphasize substantial differences between groups (LDA score = 4). In Fig. 4B, it can be observed that there were 14, 12, 8, and 8 significantly different bacterial taxonomic clades in the C, D, Y, and DY groups, respectively. The primary species observed in the four groups included Cyanobacteria, Proteobacteria, Bacteroidota, and Verrucomicrobiae. At order level, the prominent biomarker for the C group were Rhizobiales, Nitrosococcales, Chloroplast, and Burkholderiales. In the D group, the most prominent biomarkers were Verrucomicrobiales and Flavobacteriales. Among the bleached corals, Alteromonadales and Vibrionales stood out as the prominent biomarkers in the Y group. In the DY group, Oceanospirillales and Cellvibrionales were the prominent biomarkers. Furthermore, at the genus level, distinct evolutionary clusters were apparent, and prominent biomarkers underwent significant changes across all treatments. Due to the invasion of V. coralliilyticus, Nautella, Vibrio, Alteromonas, and Amphritea were the most prominent bacterial genera associated with the microbiome of bleached corals. Following exposure to V. coralliilyticus and R. profundi, Endozoicomonas and Spongiibacter emerged as the more prominent genera associated with the microbiome’s defense against pathogens. In comparison to the control group, the genus enrichment in the D group also exhibited changes. Overall, these results demonstrated that non-bleached corals had relatively distinct marker species compared with bleached corals, allowing for the differentiation of bacterial infection types across all treatments.

Fig 4.

Results of LDA combined with effect size Linear discriminant analysis Effect Size (LEfEs) to detect enriched bacterial taxa with significant differences from four treatment groups. (A) Cladograms of the phylogenetic distribution of the bacteria with significant differences among all groups. (B) Histogram of distribution of LAD values for LEfSe analysis of coral-associated bacteria based on nonparametric KW sum-rank test with P value < 0.05. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

Response of specific bacteria to the exogenous addition of bacteria and host phenotype

The objective was to achieve a deep understanding of the response exhibited by essential bacterial symbionts associated with coral by ascertaining the specific characteristics of notable shifts in bacterial abundance following the introduction of both a native bacterium and a bacterial pathogen. The taxonomic distribution of coral-associated bacteria at the genus, along with their respective relative abundances, varied among all the groups. Both in the non-bleached and bleached corals, the bacteria in the three groups exhibited similar patter. In order to test the potential impact of the pathogen V. coralliilyticus and native commensal bacteria (R. profundi) on host health, at first. The abundance of Vibrio, Ruegeria, and other genera were analyzed and presented in Fig. 5. Ruegeria exhibited no significant change in relative abundance across all groups, whereas the relative abundance of Vibrio significantly (P < 0.01) increased in the Y group compared with the other three non-bleached coral groups. In the Y group, Alteromonas, Amphritea, and Nautella showed significantly (P < 0.05) higher abundances, while Pseudohongiella exhibited significantly (P < 0.05) lower abundance compared with other groups. Therefore, the specific infection by pathogenic V. coralliilyticus may cause the alteration in the bacterial taxa.

Fig 5.

Effects of V. coralliilyticus and R. profundi on coral-associated specific taxa on the level of the genus among four treatments. Statistically significant differences are denoted by asterisks which represent significance levels of *P < 0.05, **P < 0.01, and ***P < 0.001, as determined by the nonparametric Kruskal-Wallis H test. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

BD1-7_clade, Coraliomargarita, Candidatus_Tenderia, Shimia, Portibacter, HTCC5015, Candidatus_Berkiella, and Methylotenera significantly (P < 0.05) decreased in the D, Y, and DY groups compared with the C group. These results indicated that exogenously added bacterial exposure caused significant changes in those bacterial abundance. In the DY group, subjected to a 3.5-h exposure to V. coralliilyticus stress, R. profundi was subsequently inoculated into the surrounding seawater of the coral host. A total of 9 bacteria were observed to significantly (P < 0.05) increase in the DY group compared with the Y group. Notably, these bacteria included Endozoicomonas, Terasakiella, Aestuariibacter, Phaeocystidibacter, Pirellula, Francisella, Magnetospira, Kiloniella, and Coxiella. These changes at the genus level among different treatment groups demonstrated that R. profundi might enhance the flexibility of coral-associated bacteria to respond to pathogenic stress by mobilizing more native bacterial symbionts.

The co-occurrence relationship was more complex after inoculation of the R. profundi strain

The top 100 most abundant genera were selected to analyze with each treatment group. The bacterial community interaction network was constructed using Spearman correlations (P < 0.05) and an absolute correlation coefficient value ≥ 0.5, as shown in Fig. 6. The Networkx (version1.11) software was performed to calculate topological parameters that characterize the interaction network. These parameters included the average degree, network diameter, graph density, modularity, average clustering coefficient, nodes, and edges (Table 1). Analyzing the topological parameters, it was found that the Y group had the lowest number of edges (414), average degree (8.364), graph density (0.085), and clustering coefficient (0.251) among all treatments. These demonstrated that the group exposed to V. coralliilyticus infection had the least complexity network. In contrast, some of the topological characteristic indices of the bacterial network in the DY group were the highest, such as the number of edges (673), average weighted degree (6.938), average degree (13.875), graph density (0.145), and clustering coefficient (0.307). A bacterial correlation network in the DY group with a higher proportion of positive correlations was observed. This highlighted the enhanced complexity of the relationship within the stable symbiotic bacterial community. Network analysis of the D group showed that the addition of R. profundi increased the interaction degree with the coral-associated bacteria, resulting in a more complex interaction pattern within the network.

Fig 6.

Co-occurrence patterns of coral-associated bacterial communities for each treatment group. Each network was conducted based on the top 100 relative abundance genera. The nodes indicated different bacterial genus, and their colors showed the specific modules that the nodes belonged to. The red lines denoted positive correlations and green lines denoted negative correlations between genus. (A) control; (B) R. profundi inoculation; (C) V. coralliilyticus infection; (D) following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

TABLE 1.

Key topological features of bacterial networks in coral-associated bacterial community under different inoculation treatments

Parameters	Control	R. profundi	V. coralliilyticus	V. coralliilyticus + R. profundi
Average degree	10.000	11.160	8.364	13.876
Average weighted degree	5.000	5.580	4.182	6.938
Network diameter	8	7	10	6
Graph density	0.103	0.113	0.085	0.145
Modularity	0.545	0.493	0.571	0.503
Average clustering coefficient	0.283	0.288	0.251	0.307
Positive edges	306	438	242	615
Negative edges	184	120	172	58
Total nodes	98	100	99	97

Effect of bacterial introduction on the betaine lipids and type VI section system pathways of the bacterial community

Furthermore, it is imperative to conduct experimental research to assess the second hypothesis, which examines whether R. profundi effectively functions as a guardian, preserving the normal operation of the coral-associated bacterial community to protect the host from pathogenic threats. Simultaneously, this study aims to investigate the influence of both V. coralliilyticus and R. profundi on the metabolic functions within the symbiotic bacterial community. Consequently, the predicted metagenomes were categorized at KEGG level 3, and significant (P < 0.05) enrichment of genes related to betalain biosynthesis and the bacterial secretion system were identified in the Y group compared with the other treatment groups based on the KW rank sum test (Fig. 7A and B). Specifically, genes associated with betalain biosynthesis and the bacterial secretion system exhibited a significant decrease in abundance within the DY group.

Fig 7.

The distribution of the genes related to betaine lipids and bacterial secretion system across the four treatment groups. (A and B) Plot showing differences in the abundance of the betalain biosynthesis and bacterial secretion system pathway, respectively, using Tukey-Kramer post hoc tests with FDR correction. (C and D) Relative abundance changes of genes involved microbial metabolic pathways summarized by KEGG Ortholog (KO) annotations, including choline/glycine/proline betaine transport protein and T6SS protein. C, control; D, R. profundi inoculation; Y, V. coralliilyticus infection; DY, following a 3.5-h exposure to V. coralliilyticus infection, R. profundi was inoculated.

To assess the contribution of coral-associated bacteria to these two metabolic pathways during coral bleaching, heatmaps were plotted to illustrate the distribution of KO associated with betaine lipids and T6SS in different treatment groups (Fig. 7C and D). Two KOs related to betaine lipids were annotated to K02168 (betT, betS, and choline/glycine/proline betaine transport protein) and K05020 (opuD, betL, and glycine betaine transporter), and these genes were harbored by four symbiotic bacteria, including Alteromonas, Amphritea, Endozoicomonas, and Vibrio. Fourteen KOs related to T6SS were annotated to K11902 (impA, T6SS protein ImpA), K11901 (impB, T6SS protein ImpB), K11900 (impC, T6SS protein ImpC), K11899 (impD, T6SS protein ImpD), K11896 (impG, T6SS protein ImpG), K11895 (impH, T6SS protein ImpH), K11894 (impI, T6SS protein ImpI), K11893 (impJ, T6SS protein ImpJ), K11892 (impK, T6SS protein ImpK), K11891 (impL, T6SS protein ImpL), K11906 (vasD, T6SS protein VasD), K11907 (vasG, T6SS protein VasG), K11903 (hcp, T6SS secreted protein Hcp), and K11904 (vgrG, T6SS-secreted protein VgrG). Vibrio possessed genes related to the T6SS in all groups. The abundances of predicted genes in KOs belonged to betaine lipids and T6SS among all groups, with the highest abundance in the Y group. These results suggested that betaine lipids and T6SS may be considered as metagenomic signatures of the coral microbiome during bleaching diseases.

DISCUSSION

Intensive investigations have documented an increase in coral bleaching and mortality events driven by heatwaves, pollution, overfishing, and other global anthropogenic pressures (31–35). Despite the numerous precautions and actions aimed at protecting coral reef ecosystems, such as implementing local management approaches, the average live hard coral cover has halved over the last century and has not yet recovered (36), signifying an unprecedented global deterioration of coral reefs (37). Therefore, more effective conservation strategies are required to promote coral reef resilience and recovery in the face of global threats such as bleaching. Here, we attempt to explore a probiotic bacterial strategy that may assist in enhancing coral disease resilience and investigate the connections between pathogenic stressors, the timing of the probiotic bacterial inoculation, and the behavioral response of the coral-associated bacterial community.

The inoculation of R. profundi to corals prevents V. coralliilyticus infection

In this work, we demonstrated the prophylactic efficacy of a coral symbiotic microorganism in vivo against coral pathogen infection. In addition to the R. profundi strain used in this study, other strains belonging to Ruegeria genus have shown potential for protecting corals against diseases. These strains have demonstrated in vitro growth-inhibitory activity against pathogenic bacteria including V. coralliilyticus, as reported in the literature (38). The purpose of our experimental design was to investigate whether this protective activity occurs within coral holobionts. As reported, the inoculation of the healthy coral colony with V. coralliilyticus at a density of 10⁶ c.f.u.mL⁻¹ resulted in substantial infection of coral tissue within appropriately 3.5 h (20), leading to coral bleaching after 48 h (39). Similarly, our experiments showed that the inoculation of A. carduus fragments with V. coralliilyticus at a density of 10⁶ c.f.u.mL⁻¹ led to local bleaching or tissue integrity within 48 h to 96 h. Notably, when R. profundi was inoculated after V. coralliilyticus infection, almost all coral fragments always remained non-bleached (in a healthy state) for 2 weeks.

The proportion of Ruegeria and Vibrio genera was obtained by analyzing 16S rRNA gene high-throughput sequencing data. The results suggested an increase in Vibrios when the host was infected and diseased, while no significant changes in both Vibrios and Ruegerias were detected when coral fragments were healthy. This may be due to the important probiotic role of R. profundi in controlling the proliferation of etiological agents (Vibrios). There is consistent evidence showing a survival advantage for Vibrio proliferation during disease processes. Existing studies have assessed the proportion of Vibrio, finding a significant increase in bleached coral samples compared with non-bleached samples (40). Moreover, a high prevalence and concentration of Vibrios were detected in the surrounding water, suggesting more acute disturbances and coral mucus aging cycles (41). Thus, our findings highlight the close associations between microbes and the development of diseases in coral holobionts, adopting a perspective centered on pathogen reduction. However, previous studies have not provided an explanation as to whether the probiotic consortiums colonize the host (42). Our investigation has indicated that the inoculated probiotic R. profundi does not colonize the host at higher abundance. However, it plays a role in preventing opportunistic and potentially pathogenic organisms, consequently reducing disease susceptibility in A. carduus. Therefore, it is plausible to speculate that R. profundi selects for the coral-associated microbiome, thus contributing to the host’s sustainability (43).

R. profundi avoid the disturbance to the diversity and structure of coral-associated bacterial community caused by pathogen V. coralliilyticus

The richness and diversity of coral-associated microorganisms typically increase during stressful events, as the functioning holobionts are disrupted by the growing total microbial community (44). We previously reported that bleached corals sampled across the South China Sea harbored more OTUs and exhibited higher alpha-diversity compared with non-bleached corals (45), a finding consistent with numerous other studies (39, 46, 47). In the current treatment experiment, there was an exception: the invasion of the pathogenic V. coralliilyticus triggered significant changes in the microbiome compositions of corals, resulting in a loss of bacterial diversity. Unlike tanks that are closed systems, ocean sediment and natural seawater harbor a wide range of microorganisms. This is why the laboratory environment differs from in situ conditions.

In line with previous suggestions, pathogens can weaken the holobionts’ immune system and break the coral mucus layer. This allows opportunistic and exogenous microbiome infections from the surrounding environments or pathogens to easily occupy the dominant ecological niche, thereby reducing the host’s ability to fend off pathogens (48). In bleached corals, the bacteria of Vibrio genus occupied specific niches. However, R. profundi may confer protection against V. coralliilyticus and do no significant benefit to pathogenic Vibrio. While the diversity of the bacterial community did not significantly change in non-bleached corals, the NMDS analysis based on the weighted UniFrac distance showed that samples in the C group, D group, and DY group were closer in distance compared with those in the Y group, indicating that the degree of bacterial response among non-bleached groups was similar and the bacterial community structure among non-bleached groups was alike. Therefore, R. profundi prevented shifts in the diversity and structure of the coral-associated bacterial community by the pathogen, thus maintaining holobionts’ homeostasis and promoting host health.

Although there is no clear consensus on whether the diversity and structural changes in coral-associated bacterial communities are the cause of coral bleaching or the result of compromised coral health (49–52), our findings suggest that under the pathogen infection, concurrent shifts in the behavior of the bacterial community alongside alterations in host health are synchronized, mutually influencing the development of both processes. Understanding the changes in bacterial communities is of increasing importance for evaluating and maintaining the health of individuals (53), providing a theoretical basis for protecting coral reef ecosystems. Recent studies have reported the use of antibiotics to treat coral bleaching (54–56). However, others believe that it is inappropriate due to the uncontrollable effects of broad-spectrum antibiotics on coral native microbiota. It poses significant problems related to the potential selection and spread of antibiotic-resistant bacteria, which is one of the largest threats to global public health due to antibiotic overuse (57). Additionally, the use of antibiotics has been found to decrease the diversity of coral-associated bacterial community (58), potentially reducing coral resistance to bleaching under environmental stressors. Furthermore, evidence supports that antibiotic-treated coral becomes more susceptible to V. shiloi colonization and disease (59). Consequently, current strategies involving native symbiotic beneficial bacteria are highly desirable for protecting coral hosts by establishing a more stable bacterial community and preventing pathogenic bacterial invasions.

Microbiome restructuring after inoculation of R. profundi and V. coralliilyticus

Bacterial treatments can elicit specific responses from coral-associated bacteria, which depend on coral physiology (60). Our observation, rooted in the analysis of shifts in the abundance of specific bacterial species, supports the notion that the pathogen V. coralliilyticus invades the host by competitively eliminating commensal bacteria, which leads to a decline in Endozoicomonas species within the Y group (44, 61). Several studies have identified Alteromonas and Nautella as pathogens responsible for inducing bleaching in the crustose coralline alga Porolithon onkodes, often characterized by their higher abundance (62). Consequently, it is speculated that Alteromonas, Amphritea, and Nautella may be potential pathogens capable of promoting coral bleaching and disease. Furthermore, R. profundi mobilized a significant increase in Endozoicomonas and other certain bacteria that have been identified as core bacterial genera in many coral species, suggesting their co-evolution as probiotic commensal bacteria with host holobionts (63–65). Shifts in the bacterial abundances associated with coral due to R. profundi reflected a putatively more beneficial microbiome. This tempts us to surmise that changes in certain symbiotic bacteria, driven by the modulation of R. profundi, enhance the microbiome’s capability for disease resistance, aiding the host in resisting the impact of pathogens. Enhanced defense coincided with reduced microbiome variability and an increase in a putative beneficial bacterium from the family Endozoicomonadaceae. Nevertheless, it remains unclear how coral defenses adapt to rapid shifts in reef states.

The restructuring of bacterial communities has been regarded as a crucial mechanism of coral host plasticity and adaptation to environment stressors (66–68). Similar to the approach reported by Chevrette et al. (69), our controlled experiments contributed to uncover the association between the symbiotic bacterial processes and the health state of the holobionts. Furthermore, the structure of the coral-associated bacterial community is shaped by interactions among even bacterial symbionts, which can be either antagonistic or commensal, collectively influencing the equilibrium within the symbionts and host holobionts (70). R. profundi drove an advantageous restructuring of bacterial community by improving disease resistance and pathogen defense, resulting in novel complex interactions established among the coral symbiotic bacteria. Concretely, R. profundi increased both positive correlations and the strength of interactions among bacteria, leading to a denser microbial interaction network with a more complex structure, which facilitated the stability and equilibrium of the bacterial community. Consequently, the mutualistic symbiosis relationship between symbiotic bacteria and the host becomes more sustainable and resilient.

Bacterial community function manipulations affect the coral health

Metabolic and functional roles of coral-associated bacteria that contribute to holobiont health have been identified, including their involvement in pathogen defense, metabolic cycling, and support for host resilience against multiple stressors (71). The metabolic and functional roles of coral-associated bacteria contributing to holobiont health have been identified. These roles encompass participation in pathogen defense, metabolic cycling, and the provision of support for host resilience against various stressors. To further investigate the effects on the capability of bacterial pathogenic infection and toxin production after inoculating R. profundi and V. coralliilyticus among bleached and non-bleached samples, metagenomic analysis was used to give insights into potential functional traits by profiling the relative abundances of genes within the bacterial community (72). It was found that the abundances of the metabiotic pathways related to betaine lipids and bacterial secretion systems increased in the Y group, while the two pathways significantly decreased in the DY groups. The changes in betaine lipids in bleached corals are consistent with the findings documented by Roach et al. (73), whose research highlighted that betaine lipids are a notable metabolic hallmark in coral exposed to bleaching events. Betaine lipids are secreted by corals and their microbial symbionts, contributing to the regulation of the cellular osmotic pressure (74, 75). Recent studies have also found higher concentrations of glycine betaine in coral tissues compared with ambient seawater (74). Recent studies have revealed elevated concentrations of glycine betaine in coral tissues as opposed to the surrounding ambient seawater. The coral reef ecosystem and scleractinian corals serve as sinks for betaine lipids, a major reservoir of nitrogen, in spite of the general scarcity of nitrogen in the pelagic sea (76). This could be a survival route of the bacterial community by utilizing host resources under stress, potentially increasing disease susceptibility and mortality. Taken together, these observations indicated that microbiome employs a potential mechanism to regulate the bacterial community nitrogen budgets under bleaching and disease stressors.

Marine bacteria possess a set of secretion systems that play a crucial role in host-pathogen interactions, serving as major determinants of pathogenicity and functioning at various stages of the bacterial pathogen infection pathway. These stages involve the export of virulence proteins, cell invasion, in vivo colonization, and competition, particularly in the case of Vibrio coral pathogens (77–79). Specially, the T6SS is one of the essential secretion systems for pathogenesis, directly delivering virulence factors to both eukaryotic and prokaryotic host cells, contributing to competition and exploitation (80). This was corroborated by Kanwal et al. (81), whose research suggested that the presence of T6SS affords the bacteria with a significant survival advantage. Likewise, we observed that Vibrio possesses their own T6SS gene cluster when V. coralliilyticus infected the coral hosts, allowing them to occupy predominant niches. Furthermore, V. coralliilyticus also has a functional T6SS that serves as a competition sensing mechanism during infection coral, enhancing interspecies toxicity and favoring the survival of Vibrio species over competing bacteria (78). Consistent with the findings of field experiments in our previous investigation, bleached corals exhibited a similar metagenomic signature, characterized by reduced anti-pathogenicity and increased virulence factors, among three scleractinia species during a bleaching event (82).

The secretion systems and betaine transport system can be tightly regulated by LuxR, which activates quorum sensing (QS) gene expression (83) and serves as a key regulator of QS genes in Vibrios. Many Vibrios utilize small molecules called autoinducers to govern the expression of bacterial virulence genes when infecting and colonizing corals (84). In view of the fact that QS engages a role in bacterial pathogenesis, research has proved that inhibiting QS can halt the development of coral bleaching and related microbial communities (85). Furthermore, a recent study has suggested that Ruegeria strains carry genes related to the nitric oxide and nitrous oxide reductases, enabling it to assimilate dimethylsulfoniopropionate (DMSP) and nitric oxide, which could mitigate heat stress on photosynthetic Symbiodiniaceae (86). Additionally, Ruegeria can supply vitamin B₁₂ to stimulate the photosynthetic process and the growth of the corals, as corals are unable to produce vitamin B₁₂ themselves (87–89). The present study highlights that the potential of R. profundi as a probiotic strain for assisting in the treatment of V. coralliilyticus infection in corals. However, the specific probiotic properties of R. profundi warrant further investigation through studies testing its functionality. Fully excavating the potential probiotic properties of a strain will provide useful information to elucidate mechanisms and strategies for microbiome therapeutics.

R. profundi safeguards the homeostasis of coral-associated bacteria community

The overall dissimilarity among treatment outcomes was assessed by examining alpha- and beta-diversity, composition, structure, and certain metabiotic functions of bacterial communities to characterize microbial dysbiosis and balance. After infection with pathogenic V. coralliilyticus, the bacterial community in bleached fragments underwent significant changes. In contrast, when corals were cultured with both V. coralliilyticus and R. profundi, they maintained a healthy state with no alteration in alpha-diversity and bacterial community function, except for the structural changes. Previous studies have indicated that analyzing taxon relative abundance and the composition and structure of the bacterial community does not offer sufficient evidence for explaining microbial dysbiosis in bleached corals (65, 90). Therefore, our findings contribute to a deeper understanding of the relationship between bacterial community and the coral host. To confirm coral holobiont dysbiosis, it is imperative that certain prerequisites are fulfilled: (i) significant changes occur in the alpha-diversity, composition, and structure of the coral-associated bacterial community during a bleaching event; (ii) in agreement with previous findings, dysbiosis is characterized by the proliferation of bacterial pathogens or opportunists and the reduction of coral commensal bacteria (61); and (iii) stress-induced changes in the diversity and population of coral microbiota can alter bacterial metabiotic function, leading to a significant increase in virulence factors, tissue necrosis, and disease phenotypes. R. profundi is aimed at addressing these three aspects to restrain microbial dysbiosis by activating a natural defense mechanism against invading bacterial pathogens and manipulating the abundant and beneficial taxa to enhance overall community stability.

Conclusion

This study highlighted the capacity of the probiotic R. profundi to suppress the proliferation of pathogenic V. coralliilyticus, reduce pathogenicity, enhance bacterial community homeostasis, and increase the tolerance of the host holobiont to pathogenic stress. Both the betaine lipids and T6SS pathways showed significant increases in bleached corals caused by V. coralliilyticus, implying that these pathways could potentially serve as signature markers for distinguishing microbial dysbiosis and assessing the effectiveness of probiotic bacterial therapy. While a single strain was successfully in partially arresting bleaching caused by pathogenic bacteria, further efforts should delve into the beneficial mechanisms of R. profundi. The manipulation of the coral microbiome to field applications should be aimed at the selection of the most effective indigenous microorganisms, and this endeavor also involves assessing its impact on other marine organisms.

DATA AVAILABILITY

16S rRNA sequencing and metagenomic sequencing data were submitted to the NCBI’s Sequence Read Archive (SRA) database and are available under accession numbers PRJNA1034259 and PRJNA1034677.