Coral co-culture alters bacterial community structure by altering rare bacterial species

Abstract

introduction

Although coral reefs cover less than 0.2% of the ocean worldwide, they are home to more than 25% of marine life and are the foundation of marine ecosystems (Fisher et al. 2015). Corals maintain their activity by forming holobionts comprising symbiotic associations of bacteria, zooxanthellae, archaea, fungi, and viruses (Bourne et al. 2016). Symbiotic bacteria perform various functions, including defense against pathogenic bacteria, production of antibiotics, and supply of nutrients (Bourne et al. 2016; Rosado et al. 2019). Some symbiotic bacteria can improve coral health, and programs are underway to release these bacteria into corals for coral conservation (Rosado et al. 2019; Peixoto et al. 2020).

 Therefore, clarifying the relationship between coral flora and habitat is essential for understanding coral ecology. Coral flora depends on pH (Thurber et al. 2009; Meron et al. 2011), sea temperature (Thurber et al. 2009; Webster et al. 2011; Tracy et al. 2015), salinity (Randle et al. 2020), and eutrophication (Jessen et al. 2013). Additionally, it has been shown that water flow is one of the most critical factors that affects the marine ecosystem. Furthermore, water flow inhibits bleaching in corals exposed to thermal stress (Lee et al. 2017).

Kevin Walsh et al. 2017 showed that corals influence the bacterial composition and function of water columns around coral reefs by forming a bacterial "aurabiome" of the surrounding seawater (Walsh et al. 2017). Laura Weber et al. 2019 also showed a bacterial exchange between corals and seawater by determining the bacterial composition of seawater proximal and distal to the coral (Weber et al. 2019). Additionally, coral symbiotic bacteria are transmitted from the surrounding seawater (Damjanovic et al. 2019), suggesting that water currents and hydrodynamic boundary layers around corals influence nutrient supply to corals (Shashar et al. 1996). 

Acropora tenuis, one of the major coral species in Okinawa Prefecture, Japan, and is found in both inner and outer reefs (Omori, 2011). Our previous studiesas well as those by others have confirmed the abundance of Acropora and Montipora corals in Sesoko Island waters (Cabaitan et al. 2012; Maruyama et al. 2021). With little changes in water flow in the inner reef, Montipora digitata surround A. tenuis. In contrast, the outer reef is covered by Acropora corals and is exposed to continuous water flow. The effect of communication between different coral species and water currents on the bacterial community structure is not well understood. Therefore, in this study, we examined the impact of environmental and biological factors on Acropora tenuis and Montipora digitata, which were selected for our experiments based on the above findings. The findings from this study will provide useful insights into the conservation of coral ecosystems and design of coral plantings for reef restoration.

materials and methods

Coral branch sampling and Rearing

Coral samples were obtained on November 24, 2016; A. tenuis samples were obtained from a colony inhabiting the southern area of Sesoko Island, Okinawa Prefecture (Sesokominami, 26°53.05’ N, 127°85.77’ E), whereas M. digitata samples were obtained from six colonies inhabiting the northern area of Sesoko Island, Okinawa (Ishikawabaru, 26°67.51’ N, 127°87.00’ E). The coral samples were transported to the Sesoko Station, the Tropical Biosphere Research Center, at the University of the Ryukyus, and was allowed to stand for a day in a tank with external seawater inflow. Approval for collecting coral samples was obtained from Okinawa Prefecture by Okinawa Prefecture Fishing Regulation.

Experimental design

To evaluate the effect of environmental factors on the bacterial composition of A. tenuis, we divided them into three groups: a control group, water flow control group, and co-culture group (Fig. 1). The rearing environment simulated the coral reefs around the Sesoko Island in Okinawa Prefecture. A. tenuis occupy the Sesokominami field with the water flow averaged to 15 cm/s in the natural environment. Therefore, it was referred to as a water flow control group in the rearing environment. In contrast, the Ishikawabaru field is occupied by Montipora digitata surrounding A. tenuis, and water currents averaged less than 5 cm/s in the natural environment. Therefore, they were referred to as the co-culture group in the rearing environment (Supplemental figure 1).   As a control group, we also established an environment with only A. tenuis, which was not exposed by water flow. In all three groups, nets were set up in the aquarium, and six coral branches (5–10 cm) of A. tenuis originated from one community, were placed on the nets. All groups were equipped with aerators to supply oxygen, and the inflowing seawater was the same as the seawater outside the experimental site. In the water flow group, a pump was used to generate a water flow of 15 cm/s in the tank based on the average water flow in the sea around Okinawa (Kitada et al. 2006). In the co-culture group, six branches (5–10 cm) of M. digitata originating from several colonies were placed around the branches of A. tenuis. To determine the floral community structure at the beginning of the experiment, we collected A. tenuis from the control group, M. digitata branches from the co-culture group, and seawater from the three tanks. After four weeks of rearing, branches of A. tenuis and M. digitata were also collected from each group to obtain the bacterial flora after the experiment.

The intervention groups were divided into control, flow control, and co-culture groups. All the tanks were well aerated and supplied with seawater. The water flow was adjusted to 15 cm/s in the flow control using a pump. The intervention experiment lasted four weeks, and sampling for 16S rDNA sequencing was performed three times: when the experimental corals were obtained from the natural environment, at the beginning of the experiment, and at the end of the experiment.

DNA extraction and library preparation from coral branch

For 16S rDNA gene amplicon sequencing, coral branches were transferred into a 5 mL tube (Eppendorf Ltd., Hamburg, Germany) containing 4 mL RNAlater (Thermo Fisher Scientific, Waltham, MA, USA) and then frozen with liquid nitrogen and stored in a -80°C deep freezer until analysis. The frozen samples were thawed, and the coral branches were then picked with tweezers and washed with artificial seawater. Next, in a sterilized zipper bag, the coral tissue was blown off using a Waterpik (EW-DJ61-W, Panasonic Corp., Japan), and the suspensions were collected into a 50 mL tube (Nippon Genetics Co., Ltd., Tokyo, Japan). After centrifugation at 10,000 × g, 4 °C for 30 min, the supernatant was removed, and the pellet was resuspended in 500 μL of artificial seawater. The suspension was transferred into a 1.5 mL tube (SSIBio, Lodi, CA, USA) and centrifuged at 15,000 × g, 4 °C for 15 min. After the supernatant was removed, the pellet was stored at -80 °C until DNA extraction. DNeasy Plant Mini Kit (69104; QIAGEN, Hilden, Germany) was used for DNA extraction according to the manufacturer’s instruction, with a minor modification. The thawed pellet was mixed with 400 μL of Buffer AP1 and transferred into 2 mL screw cap tubes with stabilized 0.1 mm zirconia beads (Yasui Kikai Corp., Osaka, Japan). The tubes were homogenized three times at 2,500 rpm for 60 s at 60 s intervals. After centrifugation, 4 μL of RNase A was added, and the sample was incubated at 65°C for 10 min before DNA extraction.

DNA extraction from seawater

One liter of seawater was collected from each tank and filtered through a 1.6 μm filter. The filtered seawater was filtered again using a 0.22 μm filter. The 0.22 μm filtrates after filtration were stored in RNA later (Thermo Fisher Scientific, USA). The filter was then cut with scissors, and DNA extraction was carried out through the same process as that for coral DNA extraction.

16S rDNA gene amplicon sequencing and bioinformatic analysis

The bacterial community was identified by targeting the variable region V1-V2 of the 16S rDNA gene using the primer set 27F and 338R. PCR amplification of the 16S rDNA gene amplicon (ca. 300 bp) was performed in a 25 μL mixture, and the amplicons were sequenced using Ion Torrent PGM with 318 Chip v2 (Thermo Fisher Scientific).

16S rDNA gene analysis was performed using QIIME2 2020.6 (Bolyen et al. 2019). For quality control, amplicon sequence variants (ASVs) were constructed using DADA2 (via q2-dada2) with default parameters (Callahan et al. 2016). All ASVs were aligned using mafft (q2-alignment) with default parameters (Katoh et al. 2002). fasttree2 (via q2-mask and q2-phylogeny) was used to construct a phylogenetic tree based on the masked ASV alignments with default parameters (Price et al. 2010). Taxonomic assignments were performed with 99% OTU full-length sequences of the silva database (version 138) using the q2-feature-classifier (Bokulich et al. 2018). Downstream analysis (α-diversity and β-diveristy analysis) was performed using R Bioconductor phyloseq (McMurdie and Holmes 2013) and MicrobiotaProcess package in R version 3.5.2 (Xu and Yu 2022). PERANOVA test was performed using a vegan package.

Results

Sequence data statistics and microbial diversity in corals

We conducted a 4-week rearing experiment using A. tenuis (collected from the natural environment) and compared the changes in symbiotic bacterial community structure before and after the experiment in the water flow control group and the co-culture group with the control group (no treatment). In all three groups (water flow control, co-culture, and control), no coral mortality or bleaching was observed, suggesting that fluctuations in the rearing environment did not cause fatal damage to the corals (data not shown).

The data consisted of 50 samples (seawater: 3 samples, coral: 44 samples (5 conditions, two coral species)), containing a total of 1,412,100 sequences (average length 294 bp). After exclusion of chimeric and host-derived sequences, 1,268,392 sequences were annotated to bacteria. A total of 7154 ASVs were constructed in total by 16S rDNA sequencing analysis, of which 5528 ASVs were detected in corals, and others were detected in seawater. Phylogenetic annotation showed that most (9.44–83.2%) of the bacterial composition in coral tissue belonged to the order Oceanspilaris (Endozoicomonaceae), followed by the orders Rhodobacteales, Flavobacterales, and Propiobacteriales (Fig. 2). In the following analysis, the data volume of all samples was adjusted to 10,000 reads by Rarefaction.            

To measure the diversity of symbiotic bacteria in each coral, α-diversity (Observed ASVs, Shannon index, Simpson index) was calculated in A. tenuis and M. digitata after the experiment (Fig. 3). The number of ASVs detected in A. tenuis (observed ASVs) averaged 91.5 (SD: ±23.3) at the beginning of the experiment, compared to that observed in the control (316.5 (SD: ±109.1)), water flow control (396.8 (SD: ±181.2)), and co-culture (315.3 (SD: ±124.8)) groups. There was a significant increase in diversity in all experimental groups after the experiment (Fig. 3A). However, the Shannon index, which indicates the uniformity of α-diversity, showed a significant variation in the control group (p = 0.041) and the water flow group (p = 0.0043). In contrast, no significant variation was observed in the co-culture group using the Simpson's index (p = 0.82, Fig. 3A). Simpson index did not show significant variation in any sample, and M. digitata in the co-culture group did not show significant variation the alpha diversity indices of all samples (Fig. 3B).

Variation in bacterial composition between the co-culture group and water flow control group

We compared the β-diversity among the experimental groups using weighted Unifrac and unweighted Unifrac distances to analyze the variation in bacterial flora structure among the groups (Fig. 4). In A. tenuis, a comparison between the three experimental groups at the beginning of the experiment and after four weeks of intervention showed a significant change in the bacterial composition (PERANOVA; p = 0.0027, F = 2.6474, supplemental fig. 2). Next, we analyzed β-diversity among the three groups after the experiment, and hierarchical clustering confirmed that the co-culture groups formed clusters at both weighted Unifrac and unweighted Unifrac distances (Fig. 4A, D). The bacterial flora of the control group and the two experimental groups were compared after the rearing experiment. The results showed that the intragroup distance of the co-culture group was significantly reduced compared with that of the control group (weighted Unifrac: p = 2.7e-07; unweighted Unifrac: p = 7.8e-05, Fig. 4C, F). Additionally, the intra-county distance of the co-culture group was significantly decreased by weighted Unifrac compared with that of the control group (weighted Unifrac: p = 0.0086; unweighted Unifrac: p = 0.081, Fig. 4C, F). These results suggest that the rearing environment in the co-culture group affects the structural variation of the bacterial flora of A. tenuis, resulting in the convergence of the flora structure. In contrast, the intra- and inter-county distances of both weighted Unifrac and unweighted Unifrac in the water flow control group did not change significantly compared to those of the control group (Fig. 4B, E). Therefore, the water flow adjustment did not have a substantial effect on the bacterial flora of A. tenuis.

Variation in rare bacteria between experimental groups

Next, we analyzed the presence and absence of ASVs in the bacterial flora of each experimental group to identify the bacteria involved in bacterial flora variation (Fig. 5, Table. 1). Of the 5528 ASVs in the three groups, at the beginning and end of the experiment, 30 ASVs (core ASVs) were commonly detected at the beginning of the investigation in all groups, with average relative abundance of 25.6% (0.118–36.3%) (Fig. 5A; upset groups: green). Most bacteria were members of the Endozoicomonaceae family (Endozoicomonas sp.) (Fig. 5B; upset groups: green). However, there were 192 ASVs commonly detected in the three experimental groups except at the beginning of the experiment, with an average relative abundance of 24.7% (0.047–78.1%) and the average relative abundance of 2.61% (0.00813.4%) for each ASV (Fig. 5A; red). Although the number of ASVs detected specifically in each group at the end of the experiment was high, they accounted for a low percentage of the total composition (Table. 1). Next, when we focused on the presence or absence of ASVs shared among the groups, we found 237 ASVs that were not present in the co-culture group exclusively (Fig. 5B; blue; total ASVs: 5.74 ± 4.63%, each ASV: 0.192 ± 0.381%). These ASVs had more than twice as much the ASVs absent in the control group (total ASV: 2.15 ± 1.85%, each ASV: 0.145 ± 0.620%) and the water flow control group (Fig. 5B; yellow; total ASV: 1.48 ± 1.06%, each ASV: 0.123 ± 0.178%). The results suggest that co-localization with M. digitata inhibits the establishment of many bacteria on A. tenuis compared to water flow and control conditions.

Bacterial species not detected at the beginning of the experiment but were commonly detected in the three groups at the end mainly belonged to the Rhodobactraceae family (Fig. 5B; red). The Paraspirulinaceae (genus unknown) and Phycisphaeraceae (Phycisphaera genus) families were specifically detected in the water flow control group. Bacteria that were not detected in the co-culture group included Endozoicomonaceae (Endozoicomonas genus), Rhodobacteraceae (Limibaculum, Roseibacterium, and Silicimonas genus), and Piscirickettsiaceae (Piscirickettsia genus) (Fig. 5B blue). In particular, Piscirickettsia was the only bacterium that disappeared under co-culture conditions while it was detected under other conditions.

Table. 1 Number of ASVs detected among each group

Discussion

The movement of corals to captive aquaria reconstructs bacterial composition

The bacterial flora in A. tenuis fluctuated wildly in all experimental conditions, throughout the 4-week period. Additionally, the alpha diversity (observed ASV) of the A. tenuis flora increased in all intervention groups. An increase in bacterial diversity has been reported in the bacterial flora of diseased corals (Morrow et al. 2012; Welsh et al. 2017). These results suggest that the bacterial composition restructuring in the captive environment of this experiment may have caused the corals to be exposed to stress. Additionally, this intervention increased the number of bacteril members in the A. tenuis bacterial flora, including members of the Rhodobacteraceae family, which increases during stress in bleaching, and the Flavobacteriaceae family includes pathogenic taxa. The abundance of these bacteria has also increased during previous A. hemprichii transplantation experiments (Ziegler et al. 2019) and has been reported to increase the bacteria in corals injured (Clements et al. 2020).

Co-localization with different coral species converges bacterial composition in corals

Surprisingly, the bacterial composition of A. tenuis co-culture with M. digitata converged to a unique structure compared to the composition in the control group (Fig. 4C). This suggests that co-culture with M. digitata had a substantial effect on the bacterial composition of A. tenuis. Additionally, alpha diversity results showed that the co-culture group had a lower mean Shannon index than others. Unweighted Unifrac analysis showed that the bacterial composition of the co-culture groups was similar. These results suggest that the co-culture group formed a unique bacterial composition structure because certain bacterial species detected in the control and water-flow groups were not detected in the co-culture group due to the influence of Montipora corals. Among the bacteria that were not detected in the bacterial presence/absence comparison results, were those belonging to the genus Piscirickettsia, that has been reported as a pathogenic bacterium of a salmon but is present in the coral and seawater (Mauel and Fryer 2001; Sweet et al. 2017). It has been suggested that the bacterium amplifies its abundance in diseased corals (Sweet et al. 2017). These results indicate that co-cultured Montipora corals may reduce the establishment of pathogenic bacteria in the ocean.

A possible reason for the low alpha diversity and unique structure of coral flora, in the co-culture groups, is the influence of the ecosphere formed by Montipora corals. Communication between organisms via seawater in coral reefs has been reported between corals and other organisms such as seaweeds. The organisms that make up coral reefs release their microorganisms into the surrounding environment, creating an environment with a unique bacterial structure called the ecosphere. It has been reported that genes associated with resistance to antibiotics and toxins are enriched in seawater in the vicinity of M. digitata and M. braziliensis, a member of the same genus (Walsh et al. 2017). In our experiments, M. digitata may have inhibited the establishment of pathogenic bacteria in A. tenuis through its function in the ecosphere.

Impact of coral co-localization on coral conservation

Coral reefs are composed of various corals and are classified as inner and outer reefs, which differ in the environment and bacterial composition of the corals (Roder et al. 2015). Acropora tenuis is one of the important coral species in Okinawa, Japan, and is found in both inner and outer reefs (Omori 2011). Inner reefs have slight water flow variation and are surrounded by Montipora digitata. There is little change in water flow in the inner reef and corals surrounding A. tenuis (Supplemental figure 1).

In contrast, the outer reef is occupied by Acropora corals and is exposed to persistent water flow. Water flow and co-occurrence with different coral species have not been focused on factors affecting bacterial colonization in the reconstruction of coral bacterial composition. Several studies have suggested that corals under stress rebuild their bacterial flora. This study provides insights into coral conservation and stress tolerance by clarifying the influence of habitat on the rebuilding of the coral bacterial flora.

Conclusion

Previous studies on inter-coral communication have been limited to reconstructing the bacterial composition of coral neonates and the ecosphere. In this study, we used rearing experiments to replicate previously unnoticed factors of inner and outer reefs in the coral reef waters of Okinawa, Japan, to evaluate the effects of these factors on the bacterial composition of A. tenuis. This study suggests the importance of inter-coral communication in coral habitats and maintaining coral diversity within reefs for coral conservation.

Declarations

Data availability

The raw read data of 16S rDNA are available under BioProject number PRJNA821479.

Author Contributions

Y.Na., M.I., H.F. and H.T. designed the experiments. K.I., Y.Ni., and H.T. wrote the manuscript. M.I. and Y.Na. collected the data. K.I. performed all the bioinformatics analyses. All authors have read and approved the final manuscript.

Competing interests

There are no conflicts of interest to declare.

Acknowledgment

This work was supported by the “Construction of the environmental risk mathematical model by the meta-omics analyses of marine unculturable microbes based on single cell genome information” grant from JST-CREST (JPMJCR12A4) and the Collaborative Research of Tropical Biosphere Research Center, University of the Ryukyus.

The super-computing resource was provided by the Human Genome Center (University of Tokyo).

References

Bokulich NA, Kaehler BD, Rideout JR, et al (2018) Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier plugin, Microbiome 6:90. https://doi.org/10.1186/s40168-018-0470-z

Bourne DG, Morrow KM, Webster NS (2016) Insights into the coral microbiome: Underpinning the health and resilience of reef ecosystems. Annu Rev Microbiol 70:317–340. https://doi.org/10.1146/annurev-micro-102215-095440

Bolyen E, Rideout JR, Dillon MR, et al (2019) Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2, Nat Biotechnol 37: 852–857. https://doi.org/10.1038/s41587-019-0209-9

Cabaitan PC, Yamamoto H, Sakai K (2012) Recovery of corals a decade after a thermal stress event at Motobu, Okinawa, Japan: Spatial variability in winners and losers. Galaxea J Coral Reef Stud 14:27–40. https://doi.org/10.3755/galaxea.14.1_27

Clements CS, Burns AS, Stewart FJ, Hay ME (2020) Parasite-host ecology: The limited impacts of an intimate enemy on host microbiomes. Anim Microbiome 2:42. https://doi.org/10.1186/s42523-020-00061-5

Damjanovic K, Oppen van MJH, Menéndez P, Blackall LL (2019) Experimental inoculation of coral recruits with marine bacteria indicates scope for microbiome manipulation in Acropora tenuis and Platygyra Daedalea. Front Microbiol 10:1702. https://doi.org/10.3389/fmicb.2019.01702

Fisher R, O’Leary RA, Low-Choy S, et al (2015) Species richness on coral reefs and the pursuit of convergent global estimates. Curr Biol 25:500–505. https://doi.org/10.1016/j.cub.2014.12.022

Jessen C, Villa Lizcano JFV, Bayer T, et al (2013) In-situ effects of eutrophication and overfishing on physiology and bacterial diversity of the Red Sea coral Acropora hemprichii. PLOS ONE 8:e62091. https://doi.org/10.1371/journal.pone.0062091

Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res. 30:3059-3066. https://doi.org/10.1093/nar/gkf436

Kitada Y, Fujimura H, Tokeshi R, Oomori T (2006) Air-sea CO2 flux and gas exchange coefficient at the Sesoko coral reefs, Okinawa, Japan. J Jpn Coral Reef Soc 8:51–60. https://doi.org/10.3755/jcrs.8.51

Lee STM, Davy SK, Tang SL, Kench PS (2017) Water flow buffers shifts in bacterial community structure in heat-stressed Acropora muricata. Sci rep-uk. Sci Rep 7:43600. https://doi.org/10.1038/srep43600

Maruyama T, Ito M, Wakaoji S, et al (2021) Multi-omics analysis highlights cross-organism interactions in coral holobiont. biorvix. https://doi.org/10.1101/2021.10.25.465660

Mauel MJ, Fryer JL (2001) Amplification of a Piscirickettsia salmonis‐like 16S rDNA Product from bacterioplankton DNA Collected from the Coastal Waters of Oregon, USA. J Aquat Anim Health 13:280–284. https://doi.org/10.1577/1548-8667(2001)013<0280:AOAPSL>2.0.CO;2

McMurdie PJ, Holmes S (2013) phyloseq: An R Package for Reproducible Interactive Analysis and Graphics of Microbiome Census Data. Plos One 8(4):e61217. https://doi.org/10.1371/journal.pone.0061217

Meron D, Atias E, Iasur Kruh LI, et al (2011) The impact of reduced pH on the microbial community of the coral Acropora Eurystoma. ISME J 5:51–60. https://doi.org/10.1038/ismej.2010.102

Morrow KM, Moss AG, Chadwick NE, Liles MR (2012) Bacterial Associates of two Caribbean coral species reveal species-specific distribution and geographic variability. Appl Environ Microbiol 78:6438–6449. https://doi.org/10.1128/AEM.01162-12

Omori M (2011) Degradation and restoration of coral reefs: Experience in Okinawa, Japan. Mar Biol Res 7:3–12. https://doi.org/10.1080/17451001003642317

Peixoto RS, Sweet M, Villela HDM, et al (2020) Coral probiotics: Premise, promise, prospects. Annu Rev Anim Biosci 9: 265–288. https://doi.org/10.1146/annurev-animal-090120-115444

Randle JL, Cárdenas A, Gegner HM, et al (2020) Salinity-conveyed thermotolerance in the coral model Aiptasia is accompanied by distinct changes of the bacterial microbiome. Front Mar Sci 7. https://doi.org/10.3389/fmars.2020.573635, 573635

Rosado PM, Leite DCA, Duarte GAS, et al (2019) Marine probiotics: Increasing coral resistance to bleaching through microbiome manipulation. ISME J 13:921–936. https://doi.org/10.1038/s41396-018-0323-6

Shashar N, Kinane S, Jokiel PL, Patterson MR (1996) Hydromechanical boundary layers over a coral reef. J Exp Mar Biol Ecol 199:17–28. https://doi.org/10.1016/0022-0981(95)00156-5

Sweet MJ, Brown BE, Dunne RP, et al (2017) Evidence for rapid, tide-related shifts in the microbiome of the coral Coelastrea aspera. Coral Reefs 36:815–828. https://doi.org/10.1007/s00338-017-1572-y

Tracy AM, Koren O, Douglas N, et al (2015) Persistent shifts in Caribbean coral microbiota are linked to the 2010 warm thermal anomaly. Environ Microbiol Rep 7:471–479. https://doi.org/10.1111/1758-2229.12274

Thurber RV, Willner‐Hall D, Rodriguez‐Mueller B, et al (2009) Metagenomic analysis of stressed coral holobionts. Environ Microbiol 11:2148–2163. https://doi.org/10.1111/j.1462-2920.2009.01935.x

Walsh K, Haggerty JM, Doane MP, et al (2017) Aura-biomes are present in the water layer above coral reef benthic macro-organisms. PeerJ 5:e3666. https://doi.org/10.7717/peerj.3666

Weber L, Gonzalez‐Díaz P, Armenteros M, Apprill A (2019) The coral ecosphere: A unique coral reef habitat that fosters coral–microbial interactions. Limnol Oceanogr 64:2373–2388. https://doi.org/10.1002/lno.11190

Webster NS, Soo R, Cobb R, Negri AP (2011) Elevated seawater temperature causes a microbial shift on crustose coralline algae with implications for the recruitment of coral larvae. ISME J 5:759–770. https://doi.org/10.1038/ismej.2010.152

Welsh RM, Rosales SM, Zaneveld JR, et al (2017) Alien vs. predator: Bacterial challenge alters coral microbiomes unless controlled by Halobacteriovorax predators. PeerJ 5:e3315. https://doi.org/10.7717/peerj.3315

Xu S, Yu G (2022) MicrobiotaProcess: an R package for analysis, visualization and biomarker discovery of microbiome. R package version 1.6.5, https://github.com/YuLab-SMU/MicrobiotaProcess/

Ziegler M, Grupstra CGB, Barreto MM, et al (2019) Coral bacterial community structure responds to environmental change in a host-specific manner. Nat Commun 10:3092. https://doi.org/10.1038/s41467-019-10969-5