Chemically Mediated Interactions with Macroalgae Negatively Affect Coral Health but Induce Limited Changes in Coral Microbiomes

Allelopathic chemicals facilitated by the direct contact of macroalgae with corals are potentially an important mechanism mediating coral–macroalgal interactions, but only a few studies have explored their impacts on coral health and microbiomes and the coral’s ability to recover. We conducted a field experiment on an equatorial urbanized reef to assess the allelopathic effects of four macroalgal species (Bryopsis sp., Endosiphonia horrida, Hypnea pannosa and Lobophora challengeriae) on the health and microbiomes of three coral species (Merulina ampliata, Montipora stellata and Pocillopora acuta). Following 24 h of exposure, crude extracts of all four macroalgal species caused significant coral tissue bleaching and reduction in effective quantum yield. The corals were able to recover within 72 h of the removal of extracts, except those that were exposed to L. challengeriae. While some macroalgal extracts caused an increase in the alpha diversity of coral microbiomes, there were no significant differences in the composition and variability of coral microbiomes between controls and macroalgal extracts at each sampling time point. Nevertheless, DESeq2 differential abundance analyses showed species-specific responses of coral microbiomes. Overall, our findings provide insights on the limited effect of chemically mediated interactions with macroalgae on coral microbiomes and the capacity of corals to recover quickly from the macroalgal chemicals.


Introduction
Coral reefs are experiencing degradation globally due to the cumulative impacts of multiple local anthropogenic stressors (e.g., overfishing, eutrophication and disease) compounded by rising seawater temperatures and ocean acidification [1][2][3].The declines in live coral cover are often accompanied by the proliferation of macroalgae [4,5], resulting in an increase in the competition between corals and macroalgae for space.Established macroalgal communities may generate positive feedback loops, which further reinforce macroalgal dominance and coral-depauperate state by promoting greater macroalgal recruitment [6], suppressing herbivory [7], reducing coral growth and survivorship [8,9], as well as inhibiting coral fecundity and recruitment [10,11].Understanding the processes that mediate coral-macroalgal interactions is critical to the management and conservation of coral reef ecosystems [12].
The coral microbiome consists of dinoflagellates, bacteria, archaea, fungi and viruses [13].Together, they play important roles in nutrient cycling, coral host immunity against pathogens and mediating environmental stress [13][14][15].Macroalgae can negatively affect corals and their microbiomes through a range of physical, chemical and microbial mechanisms [16][17][18].In addition to space pre-emption, macroalgae may physically damage corals via shading, abrasion and increased localized sedimentation rates [9,19].Several studies have documented that some macroalgae possess allelochemicals, which can cause coral tissue bleaching and mortality [17,20] and reduce coral larval survivorship and settlement rates [21,22].Some macroalgal chemicals have also been shown to possess antibacterial properties that differentially target marine bacteria, resulting in shifts to higher abundances of harmful microbes and/or lower abundances of beneficial microbes [21,23].In addition, macroalgae may destabilize the coral microbiome by transmitting pathogenic microbes to corals upon contact [24] and by releasing a fraction of their photosynthetically fixed carbon as dissolved organic matter (DOM) into the water column [25].Macroalgal DOM is readily consumed by microbes [26], which can lead to an overactivity in the coral microbiome, resulting in hypoxic microenvironments that may favor copiotrophic and potentially pathogenic microbes [25,27].
Previous studies have revealed that chemically mediated interactions between corals and macroalgae are dependent on direct contact [17,20,28].Hydrophobic, lipid-soluble chemicals extracted from surfaces of macroalgae were embedded into gels and placed in contact with corals, causing coral bleaching and reduction in coral photosynthetic efficiency in 79% of the 24 coral-macroalgal interaction pairs assayed [17].Similarly, extracts from eight Lobophora spp.Induced tissue bleaching in Acropora muricata and Stylophora pistillata, but not in Porites cylindrica and Montipora hirsuta [20].The low solubility of lipid-soluble macroalgal extracts in water also suggests that these chemicals are most efficient when they are transferred to corals by direct contact rather than dissolution through water [17,28].However, interactions between corals and macroalgae are complex, with varying allelopathic potential among macroalgal species and differential susceptibility among coral species [16,17,20].Furthermore, it remains unclear how corals and their microbiomes recover from the damage caused by macroalgal chemicals.
To narrow this knowledge gap, a field experiment on an equatorial urbanized reef was conducted to assess the allelopathic effects of macroalgae on coral health and the prokaryotic portion of their microbiomes.Specifically, the effects of lipid-soluble extracts of four macroalgal species (Bryopsis sp., Endosiphonia horrida, Hypnea pannosa and Lobophora challengeriae) were tested on three coral species (Merulina ampliata, Montipora stellata and Pocillopora acuta).The responses of corals after the macroalgal extracts were removed were also examined to assess the coral recovery potential.

Macroalgal Crude Extracts
The allelopathic potency of lipid-soluble extracts from four common macroalgal species in Singapore were examined, namely the green alga Bryopsis sp., the brown alga Lobophora challengeriae and the red algae Endosiphonia horrida and Hypnea pannosa.These macroalgal species were selected because they interacted frequently with corals on Singapore's reefs [29], and their chemical effects on adult corals and their microbiomes have not been explored previously.Samples were collected from three shallow fringing reefs in Singapore at depths of 2-3 m (Figure S1) and cleaned of epiphytes.Hydrophobic, lipid-soluble metabolites were extracted from the macroalgal surfaces, including their surface-associated microbiome, following the hexane dip technique [30].Samples were first spun in a salad spinner to remove excess water and thereafter were extracted in 100% hexane for 30 s while being vortexed vigorously [30,31].Extracts were dried in vacuo and stored at −20 • C. To prepare the extract gels, extracts were resuspended in 0.5 mL methanol and added to auto-claved molten Phytagel medium (Sigma-Aldrich, St. Louis, MI, USA) at natural volumetric concentrations (e.g., 10 mL volume of gel was prepared using 10 mL displacement volume of algae).The extract gel was poured in a mold fitted with plastic screen net, allowed to cool down and cut into 2 cm × 2 cm squares.Control gels were made in the same manner using only methanol without the addition of macroalgal extracts.Gels were stored at −20 • C for two weeks until deployment in the experiment.

Experimental Procedures
The effects of algal extracts were tested on the colonies of Merulina ampliata, Montipora stellata and Pocillopora acuta growing at depths of 2-3 m on Pulau Satumu, Singapore (Figure S1).These corals were selected as they are common on Singapore's reefs [32] and within genera that have been shown to vary in susceptibility to macroalgal competition [33].They also have branching morphologies, which made it easier to attach the gels.Fifteen colonies of each coral species (>15 cm diameter), which had no signs of bleaching, were tagged.On each colony, five extract gels (one control and each of the four macroalgal species) were secured to different coral branches using cable ties (following Ref. [28]).While a blocked experimental design may result in algal treatments having synergistic effects on coral colonies, previous studies have demonstrated that the effects of macroalgae and their extracts were localized to the area of direct contact [17,28,31].The use of a blocked design also minimized the effect of inter-colony variation, especially in coral microbiomes, among algal treatments.Finally, the design substantially reduced the number of coral colonies that had to be sacrificed-from 75 colonies per coral species when using a complete independent design to 15 colonies per coral species when using a blocked design.
After 24 h, the gels were removed from all colonies, and the impacts of macroalgal extracts on coral health were assessed.The coral tissue underneath the area occupied by each gel was photographed to quantify the percentage of coral tissue bleached using ImageJ v1.52n [34] (Figure S2).The effective quantum yield of each coral branch where gels were applied was also measured using a portable diving pulse amplitude modulated (PAM) fluorometer (Heinz Walz GmbH, Effeltrich, Germany).Cable ties, which were used to secure the extract gels, were left at the base of the coral branches to mark the location of the coral branches where treatments were applied.
To assess the recovery potential of corals, the percentage of coral tissue bleached and effective quantum yield were measured 24 h and 72 h after gels were removed.At each time point (i.e., 24 h after gels were applied, 'day 0'; 24 h after gels were removed, 'day 1'; and 72 h after gels were removed, 'day 3'), the coral branches that were exposed to macroalgal treatments were also collected from five colonies of each coral species for microbial analysis (hence, the number of available replicates for the percentage of coral tissue bleached and effective quantum yield was reduced by five at each time point).All PAM measurements were taken between 11:30 and 14:00.Coral fragments were put in individual sterile sampling bags and transported on ice to the laboratory.Samples were stored in a −80 • C freezer until further processing.

Coral Microbiome Analysis
Tissues from the coral fragments were extracted using the WaterPik method.DNA extraction was carried out using the Qiagen DNeasy PowerBiofilm Kit and Zymo Clean & Concentrator-25 kit, resulting in 30 µL of eluted DNA for each coral tissue sample.PCR was carried out to amplify the 16S rRNA V4 region with the 515F-806R primer set [35][36][37].The PCR was run as follows: 1 cycle of 95  C for 10 min, followed by rest at 4 • C. Each PCR reaction mix consisted of 10 µL of Qiagen HotStar Taq Plus Master Mix, 1 µL of 10 µM 515F and 806R primers each, 1 µL of 400 µM Bovine Serum Albumin, 5 µL Ambion nuclease-free water (Carlsbad, CA, USA) and 2 µL of 5 ng µL −1 template DNA.
For M. ampliata and M. stellata, PCR products were subjected to magnetic bead cleanup with AMPure magnetic beads and were eluted in 50 µL TE buffer.For P. acuta, PCR products were run on 2% agarose gel, excised with the Invitrogen PureLink Quick Gel Extraction Kit and eluted in 50 µL kit Elution Buffer (TE buffer).All PCR products were quantified with Invitrogen Qubit DNA High Sensitivity Kit, followed by visualization using the BioAnalyzer with Agilent D1000 Screentapes.The purified PCR products were submitted to the sequencing facility at the Singapore Centre for Life Sciences Engineering for amplicon sequencing using the Illumina MiSeq300 platform.

Statistical Analyses
To evaluate the allelopathic effects of macroalgal crude extracts on corals, linear mixedeffect (LME) models were performed using the lme function from the nlme package [38] in R v4.2.1 [39].The percentage of tissue bleaching and effective quantum yield of corals were fitted as response variables, while macroalgal treatment, coral species and their interactions were fitted as fixed factors.Colony identity was included as a random effect to account for the blocked experimental design.Separate analyses were conducted for each time point due to unequal number of samples (i.e., n = 15 for day 0, n = 10 for day 1 and n = 5 for day 3).Significance tests of fixed factors were determined based on likelihood ratio tests and changes in Akaike's information criterion (AIC).Post hoc Dunnett's tests for comparison of each treatment against control were conducted using the emmeans package [40].Where necessary, the residual variance was allowed to differ among coral species and macroalgal treatment to control for heteroscedasticity.Assumptions of normality and homoscedasticity were validated by visual inspection of the models' residual plots.
Coral microbial data were processed following DADA2 pipeline v. 1.16 [41] to generate amplicon sequence variants (ASVs).In summary, the sequences were trimmed with the trimLeft and truncLen parameters set to (52,52) and (190,190), respectively, followed by dereplication, merging of the paired reads and removal of chimeric sequences.Taxonomic data were assigned to the sequences using dada2-formatted reference files developed from the Silva-ARB version 138.1 database with the assignTaxonomy and addSpecies functions.Subsequent analyses of coral microbiome data were performed separately for each coral species, as they had distinct microbial communities (PERMANOVA: F = 38.7,p = 0.001; Figure S3).
Three alpha diversity indices (chao1, Shannon-Wiener, inverse Simpson) were generated using phyloseq [42].LME models were fitted to determine whether the alpha diversity indices were significantly different among macroalgal treatments and time points.Non-metric multi-dimensional scaling (NMDS) plots using Bray-Curtis dissimilarity of square-root-transformed data were generated to examine the patterns in the beta diversity of coral microbiomes.Permutational multivariate analysis of variance (PERMANOVA) and permutational analysis of multivariate dispersions (PERMDISP) were carried out to compare the differences in the average community composition of coral microbiomes and their variability among macroalgal treatments and time points.The DESeq2 package [43] was also used to determine which ASVs had significant difference in abundance when the microbiomes between the control and treated microbiomes were compared for each experimental time point.

Results
Coral tissue health and effective quantum yield differed significantly among the macroalgal gel treatments (Table 1, Figure 1).At day 0, compared to the control gels, coral branches that were exposed to the crude extracts of Bryopsis sp., E. horrida, H. pannosa and L. challengeriae had a significantly higher percentage of tissue bleached and lower effective quantum yield (Figure 1).Twenty-four hours after the extracts were removed (i.e., day 1), corals that were previously exposed to Bryopsis sp., H. pannosa and L. challengeriae extract gels had significant tissue bleaching (Table 1, Figure 1).In contrast, 72 h after the extracts were removed (i.e., day 3), only corals that were treated with L. challengeriae extract gels had significant tissue bleaching and lower effective quantum yield compared to the control gels (Table 1, Figure 1).While the percentage of tissue bleached and effective quantum yield differed significantly among corals, there were no significant interactions between coral and macroalgal treatments (Table 1).quantum yield (Figure 1).Twenty-four hours after the extracts were removed (i.e., day 1), corals that were previously exposed to Bryopsis sp., H. pannosa and L. challengeriae extract gels had significant tissue bleaching (Table 1, Figure 1).In contrast, 72 h after the extracts were removed (i.e., day 3), only corals that were treated with L. challengeriae extract gels had significant tissue bleaching and lower effective quantum yield compared to the control gels (Table 1, Figure 1).While the percentage of tissue bleached and effective quantum yield differed significantly among corals, there were no significant interactions between coral and macroalgal treatments (Table 1).Coral microbiomes were dominated by members of the Phyla Proteobacteria, Bacteroidota and Firmicutes, while other Phyla, such as Bdellovibrionota, Cyanobacteria, Coral microbiomes were dominated by members of the Phyla Proteobacteria, Bacteroidota and Firmicutes, while other Phyla, such as Bdellovibrionota, Cyanobacteria, Planctomycetota and Verrucomicrobiota, were present in lower abundances (Figure S4).Macroalgal extracts caused significant changes in the alpha diversity indices of coral microbiomes, but the changes were variable among coral species (Figure 2).For M. ampliata, significant differences in alpha diversity indices were observed on day 1: the E. horrida treatment had higher Shannon-Wiener and inverse Simpson indices than the controls, while H. pannosa and L. challengeriae treatments had a higher inverse Simpson index than the controls (Figure 2).For P. acuta, significantly higher Shannon-Wiener and inverse Simpson indices were observed in corals exposed to L. challengeriae extracts compared to controls on day 0. In contrast, the diversity of M. stellata microbiomes remained relatively stable among treatments at each time point (Figure 2).
Planctomycetota and Verrucomicrobiota, were present in lower abundances (Figure S4).Macroalgal extracts caused significant changes in the alpha diversity indices of coral microbiomes, but the changes were variable among coral species (Figure 2).For M. ampliata, significant differences in alpha diversity indices were observed on day 1: the E. horrida treatment had higher Shannon-Wiener and inverse Simpson indices than the controls, while H. pannosa and L. challengeriae treatments had a higher inverse Simpson index than the controls (Figure 2).For P. acuta, significantly higher Shannon-Wiener and inverse Simpson indices were observed in corals exposed to L. challengeriae extracts compared to controls on day 0. In contrast, the diversity of M. stellata microbiomes remained relatively stable among treatments at each time point (Figure 2).There were shifts in the coral microbial community compositions across time points, including the control gels (Figure S5).Therefore, to assess the specific response of the coral microbiomes to macroalgal treatments, we focused on the comparisons between coral fragments that received algal extracts and control gels at each time point.PERMANOVA and PERMDISP showed that on day 0, the coral microbiomes did not vary significantly among macroalgal treatments following 24 h of extract exposure (Table 2, Figure 3).During the recovery period (day 1 and 3), there were also no significant differences in the average composition and variability of coral microbial communities between control gels and macroalgal extracts (Table 2, Figure 3).However, DESeq2 analysis revealed that there were many ASVs that had significant changes in abundance (Table 2).In M. stellata, the greatest ASV changes were observed on day 0, while the highest number of ASV changes in M. ampliata and P. acuta occurred on day 1.Notably, among the three coral species, M. stellata had the greatest number of ASVs that differed in abundance across the time points.There were shifts in the coral microbial community compositions across time points, including the control gels (Figure S5).Therefore, to assess the specific response of the coral microbiomes to macroalgal treatments, we focused on the comparisons between coral fragments that received algal extracts and control gels at each time point.PERMANOVA and PERMDISP showed that on day 0, the coral microbiomes did not vary significantly among macroalgal treatments following 24 h of extract exposure (Table 2, Figure 3).During the recovery period (day 1 and 3), there were also no significant differences in the average composition and variability of coral microbial communities between control gels and macroalgal extracts (Table 2, Figure 3).However, DESeq2 analysis revealed that there were many ASVs that had significant changes in abundance (Table 2).In M. stellata, the greatest ASV changes were observed on day 0, while the highest number of ASV changes in M. ampliata and P. acuta occurred on day 1.Notably, among the three coral species, M. stellata had the greatest number of ASVs that differed in abundance across the time points.The ASVs were mostly members of the Phyla Proteobacteria, Firmicutes, Bacteroidota, Actinobacteriota, Bdellovibrionota and Verrucomicrobiota (Table S1).The ASVs were mostly members of the Phyla Proteobacteria, Firmicutes, Bacteroidota, Actinobacteriota, Bdellovibrionota and Verrucomicrobiota (Table S1).

Discussion
Coral-macroalgal interactions are common on coral reefs; yet, the mechanisms mediating them remain poorly understood.In this study, we found that lipid-soluble extracts of Bryopsis sp., E. horrida, H. pannosa and L. challengeriae caused significant tissue bleaching and reduction in effective quantum yield across the three coral species tested.However, corals were able to recover within 72 h of the removal of macroalgal extracts, except those that were exposed to L. challengeriae.The effects of the macroalgal extracts on the coral

Discussion
Coral-macroalgal interactions are common on coral reefs; yet, the mechanisms mediating them remain poorly understood.In this study, we found that lipid-soluble extracts of Bryopsis sp., E. horrida, H. pannosa and L. challengeriae caused significant tissue bleaching and reduction in effective quantum yield across the three coral species tested.However, corals were able to recover within 72 h of the removal of macroalgal extracts, except those that were exposed to L. challengeriae.The effects of the macroalgal extracts on the coral microbiomes were less pronounced.Some extracts caused an increase in the microbial alpha diversity, but there were no significant differences in the average composition and variability of coral microbiomes among macroalgal treatments.However, marked changes in the abundance of individual ASVs were observed, particularly in M. ampliata microbiomes.Overall, our findings suggest that chemically mediated interactions with macroalgae have limited impacts on coral microbiomes at the time scales we tested, and corals are able to recover quickly from the damage by macroalgal chemicals.
Hydrophobic surface extracts from all four macroalgal species caused significant tissue bleaching and decline in effective quantum yield.Extracts from E. horrida, H. pannosa and L. challengeriae also caused an increase in the alpha diversity of coral microbiomes.While we did not identify and isolate the active allelopathic compounds responsible for affecting the corals, our results corroborate findings from past research, which demonstrated that these macroalgae-potentially with their surface-associated microbiomes-contained allelochemicals.An active anti-predatory depsipeptide compound known as kahalalide F was found in Bryopsis sp. from Hawaii [44].Three halogenated terpenoids isolated from H. pannosa have been shown to possess antimicrobial properties [45,46].The genus Lobophora is known to be chemically rich [20,47,48].Three polyunsaturated alcohols from L. rosacea were identified to cause significant tissue bleaching in Acropora muricata [20].Similarly, a compound isolated from L. variegata known as SQDG was responsible for inducing tissue bleaching in Montastraea cavernosa [48].Aqueous crude extracts of Lobophora sp. were also found to exhibit broad-spectrum antibacterial activity and caused significant changes in coral microbiomes [21,49].The results from the present study were largely consistent with previous bioassays, which demonstrated that Bryopsis sp., Endosiphonia horrida and Lobophora sp.crude extracts were chemically potent [22].However, while H. pannosa extracts did not exert any effect on coral larvae [22], we found that they negatively affected coral health and increased the inverse Simpson index in M. ampliata microbiomes.These differences were likely due to the different extraction techniques used.The crude extracts of H. pannosa in Ref. [22] were obtained through whole-cell extraction; therefore, the extracts likely contained a range of both non-polar and polar compounds originating from both internal and external macroalgal cells.In contrast, the current study employed the hexane dip extraction technique to obtain lipid-soluble chemicals present on the surfaces of the macroalgae because chemical interactions between macroalgae and adult corals are most frequently mediated through direct contact [17].It is also possible that the allelochemicals of H. pannosa act primarily on coral adults instead of coral larvae.
We found that the impacts of hydrophobic, lipid-soluble macroalgal extracts on coral microbiomes were limited.While some measures of alpha diversity of M. ampliata and P. acuta microbiomes were higher following exposure to the extracts compared to control gels, the average composition and variability of coral microbiomes were not altered by contact with them.While multiple studies have documented the shifts in coral microbiomes when corals were in contact with live macroalgae [49][50][51][52] but see [53], the impacts of macroalgal extracts on coral microbiomes were variable depending on their polarity and the species tested [21,23].Aqueous, hydrophilic extracts of Lobophora sp. in the Caribbean altered the microbiomes of Porites astreoides and Orbicella faveolata, while the non-polar, lipophilic extracts of Lobophora sp. had contrasting effects on the coral microbiomes depending on the site and coral species tested [23].Similar to our findings, macroalgal extracts were also found to damage coral tissue without altering coral microbial communities and vice versa [23].Both hydrophilic and lipophilic crude extracts of Lobophora sp. in Australia were found to inhibit growth of bacterial isolates, but only hydrophilic extracts of Lobophora sp.induced shifts to Vibrio sp.dominance in Porites cylindrica microbiomes [21].In this study, we only tested lipid-soluble crude extracts because for allelopathy to be effective, defensive chemicals are usually hydrophobic and transferred by contact [28,31].Our findings add to a growing body of evidence that lipid-soluble extracts of macroalgae have limited impacts on coral microbiomes, suggesting that changes in coral microbiomes when in contact with live macroalgae are likely mediated by the transmission of macroalgal pathogens upon contact [24] or from the increase in the availability of DOM released by macroalgae [25,26].
It should be noted that our findings do not rule out the potential for macroalgae to alter coral microbiomes via polar chemicals mediated by indirect, water-mediated interactions, but further studies will be needed to determine the extent to which the polar fractions of macroalgal allelochemicals can influence coral microbiomes in field conditions.
Our results, however, did reveal a number of microbial ASVs with significant differences in abundance among the macroalgal treatments.These differences were influenced strongly by coral species.In M. stellata microbiomes, the number of differing ASVs ranged from 39 to 61 across the experimental sampling time points, with many ASVs in M. stellata microbiomes (22-37 ASVs) consistently having lower abundance following exposure to the macroalgal extracts.In contrast, there were between 3 and 39 ASVs that showed significant changes in abundance in M. ampliata and P. acuta microbiomes, with the largest changes observed at day 1.Certain ASVs, such as those associated with Rhodobacteraceae (e.g., ASV66) and Alteromonas sp.(e.g., ASV276), increased in abundance in M. stellata and P. acuta microbiomes that were treated with Bryopsis sp. and L. challengeriae extracts.Although these bacterial groups have been associated with the coral core microbiome [54], they have the potential to become opportunistic pathogens when conditions promoting dysbiosis occur [55,56].Species-specific increases in ASVs associated with functions beneficial to coral health were also found 24 h after the removal of algal extracts, including ASVs associated with antimicrobial activity (e.g., Erythrobacter sp.(ASV47) and Paracoccus sp.(ASV69)) in M. ampliata and M. stellata microbiomes [57].The relative abundance of Peredibacter sp.(ASVs 146, 285), which belongs to a group of predatory bacteria [58,59], was also higher in P. acuta microbiomes that were treated with Bryopsis sp. and E. horrida extracts, which likely served as a form of pathogenic bacterial control.On day 3 (i.e., 72 h after the removal of the extracts), the abundance of the ASVs associated with known pathogenic microbial groups, including Vibrionaceae (ASV157) and Sphingomonas sp.(ASV244) [60], had decreased across the macroalgal-treated microbiomes compared to control microbiomes in all three coral species.These changes in the abundances of ASVs suggest that the coral microbiomes were actively involved in maintaining and restoring coral health following sub-lethal stress caused by macroalgae.Future studies should concurrently assess coral holobiont transcriptome and microbiome responses to better understand how coral-associated microbes may influence the coral holobiont's response to the competitive interactions with macroalgae.
Our findings indicate that corals can-within the time frame of exposure in the current study-mostly recover from the allelopathic impacts of macroalgae within a short period of time.Merulina ampliata, M. stellata and P. acuta recovered within 24 h of exposure to Bryopsis sp. and E. horrida extracts and within 72 h of exposure to H. pannosa extracts.Only corals that were exposed to L. challengeriae extracts did not recover within 72 h of the extracts being removed, suggesting that L. challengeriae is the most chemically potent among the macroalgal extracts tested.A previous study showed that six months after the removal of Sargassum, the growth rates of Acropora millepora and Porites cylindrica fragments that were previously surrounded by Sargassum were comparable to the growth rates of control fragments [61].Another study reported that tissue damage on Porites lobata colonies caused by contact with Galaxuara filamentosa recovered within two weeks of the macroalgae being removed [33].Our results, together with those from Refs.[33,61], demonstrate that corals are able to recover from macroalgal impacts.However, this might be dependent on whether the interaction frequency between corals and macroalgae can be limited temporally, e.g., by herbivory and macroalgal seasonality.
Previous studies have documented among-species variation in coral susceptibility to macroalgal allelopathic effects [17,33].For instance, Pocillopora damicornis was reported to be more susceptible to macroalgal allelopathic damage than Montipora digitata [17,33].Extracts of Halimeda tuna caused shifts in the bacterial communities of Porites astreoides but not Orbicella faveolata, while extracts of Lobophora sp.induced shifts in both corals [23].In contrast, we did not find significant interaction effects of coral species and macroalgal treatments on the percentage of coral tissue bleached and effective quantum yield, implying that M. ampliata, M. stellata and P. acuta were equally susceptible to chemical damage by macroalgae.Further, the limited impacts of macroalgal extracts on the overall coral microbial compositions were consistent among the three coral species.These results align with a previous study, which documented the physiology and microbiomes of M. ampliata, M. stellata and P. acuta being similarly compromised by direct contact with Lobophora sp. and H. pannosa [51].Our findings and those of Ref.
[51] provide strong evidence that M. ampliata, M. stellata and P. acuta on Singapore's reefs are equally susceptible to macroalgal competition.
In summary, our experiment showed that lipid-soluble chemicals found on surfaces of common macroalgae in Singapore can damage a range of coral species.However, the impacts of these chemicals on coral microbiomes were limited to changes in abundances of individual ASVs but not the overall microbial community compositions.Most corals recovered within short periods of time after the chemicals were removed, suggesting resistance to macroalgal damage.Determining the ability of corals to recover from the impacts of macroalgae is critical to understanding reef resilience and potential reversal of degraded reefs to coral-dominated reefs.Management strategies controlling macroalgal proliferation, such as herbivore protection or active macroalgal removal, may be necessary to facilitate coral recovery on reefs that are becoming algal dominated.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms11092261/s1, Figure S1: Map of the Southern Islands of Singapore showing the collection sites for macroalgae (E.horrida collected from Pulau Hantu, L. challengeriae and Bryopsis sp. from Pulau Subar Darat, and H. pannosa from Kusu Island) and Pulau Satumu, where the in situ experiment was conducted (A).Gray shaded areas represent fringing reef areas.Photos of macroalgal species (B-E) and coral species (F-H) tested in this study;  S1: Summary of the DESeq2 analyses showing the ASV differences in M. ampliata, M. stellata and P. acuta microbiomes between the methanol control and macroalgal extracts at different time points.

Figure 1 .
Figure 1.Effects of macroalgal crude extracts (estimated marginal means and 95% confidence intervals) on the percentage of coral tissue bleached (A-C) and effective quantum yield (D-F).Asterisks indicate significant differences between control and macroalgal gel treatments for all three coral species based on Dunnett's post hoc comparisons.

Figure 1 .
Figure 1.Effects of macroalgal crude extracts (estimated marginal means and 95% confidence intervals) on the percentage of coral tissue bleached (A-C) and effective quantum yield (D-F).Asterisks indicate significant differences between control and macroalgal gel treatments for all three coral species based on Dunnett's post hoc comparisons.

Figure 2 .
Figure 2. Variations in the alpha diversity indices of the coral microbiomes (mean ± SE) when in contact with various macroalgal extracts sampled at different time points.The colors indicate the type of macroalgal extract, while the asterisks indicate significant differences compared to the control.

Figure 2 .
Figure 2. Variations in the alpha diversity indices of the coral microbiomes (mean ± SE) when in contact with various macroalgal extracts sampled at different time points.The colors indicate the type of macroalgal extract, while the asterisks indicate significant differences compared to the control.

Figure 3 .
Figure 3. Non-metric multi-dimensional scaling (NMDS) plots showing the community composition of coral microbiomes at each sampling time point for M. ampliata, M. stellata and P. acuta.The stress values of the nMDS plots were <0.2.

Figure 3 .
Figure 3. Non-metric multi-dimensional scaling (NMDS) plots showing the community composition of coral microbiomes at each sampling time point for M. ampliata, M. stellata and P. acuta.The stress values of the nMDS plots were <0.2.

Figure S2 :
Photos of coral branches (left: Merulina ampliata; right: Pocillopora acuta) suffering from tissue bleaching due to exposure to macroalgal extracts (left: Lobophora challengeriae; right: Hypnea pannosa); Figure S3: Non-metric multi-dimensional scaling (NMDS) plots based on Bray-Curtis distances comparing the coral microbiomes of the three coral species.The different colors indicate the three coral species used in this study; Figure S4: Relative abundances of the coral microbiomes (Phylum-level) of the three coral species that were in contact with macroalgal extracts across different time points; Figure S5: Non-metric multi-dimensional scaling (NMDS) plots showing the overall temporal shifts in the microbiomes of M. ampliata, M. stellata and P. acuta; Table

Table 1 .
Summary of LME models on the effects of macroalgal treatment and coral species on the percentage of tissue bleached and effective quantum yield.

Table 1 .
Summary of LME models on the effects of macroalgal treatment and coral species on the percentage of tissue bleached and effective quantum yield.

Table 2 .
Summary of the PERMANOVA, PERMDISP and DESeq2 results of the coral microbiome composition and changes in the number of ASVs between the control and treated samples for each experimental time point.Details of the ASV changes can be found in TableS1.

Table 2 .
Summary of the PERMANOVA, PERMDISP and DESeq2 results of the coral microbiome composition and changes in the number of ASVs between the control and treated samples for each experimental time point.Details of the ASV changes can be found in TableS1.