The microbial profile of a tissue necrosis affecting the Atlantic invasive coral Tubastraea tagusensis

The Southwestern Atlantic rocky reef ecosystems are undergoing significant changes due to sun-corals (Tubastraea tagusensis and T. coccinea) invasion. At Búzios Island, on the northern coast of São Paulo State, where the abundance of T. tagusensis is particularly high, some colonies are displaying tissue necrosis, a phenomenon never reported for this invasive nor any other azooxanthellate coral species. Using next-generation sequencing, we sought to understand the relationship between T. tagusensis tissue necrosis and its microbiota. Thus, through amplicon sequencing, we studied both healthy and diseased coral colonies. Results indicate a wide variety of bacteria associated with healthy colonies and an even higher diversity associated with those corals presenting tissue necrosis, which displayed nearly 25% more microorganisms. Also, as the microbial community associated with the seven healthy colonies did not alter composition significantly, it was possible to verify the microbial succession during different stages of tissue necrosis (i.e., initial, intermediate, and advanced). Comparing the microbiome from healthy corals to those in early tissue necrosis suggests 21 potential pathogens, which might act as the promoters of such disease.


Scientific Reports
| (2021) 11:9828 | https://doi.org/10.1038/s41598-021-89296-z www.nature.com/scientificreports/ may be related to its reproductive characteristics, such as high production of planula 42 ; early reproductive age 43 ; clonality 44,45 ; regeneration capacity 46 ; and quick incubation-all hallmarks of opportunistic species 47 . As a result of asexual planulae production, Southwestern Atlantic invasive T. tagusensis displays a high clonal rate 44 , and therefore a low genetic diversity, a phenomenon previously observed in other invasive species populations 48 . Such a decrease in diversity is caused by the founding effect-few specimens colonizing a new environment, which can reduce the adaptive potential of the species over time 48 . This condition of T. tagusensis may be the downside of this invasive species because, in addition to having a high rate of clonality, it also showed the absence of significant differences in the microbial community along a depth gradient 30 .
In 2014, when several rocky shores of the Búzios Island were already saturated with invasive corals, colonies of T. tagusensis displaying tissue necrosis were observed. Initially, affected colonies were seen in only one small location, but since then, affected colonies have become widespread. Such tissue necrosis is the first report of disease in azooxanthellate scleractinian corals. To better understand this tissue necrosis, here we characterize and compared its bacterial composition during different necrosis stages. Thus, the T. tagusensis microbiome was quali-quantitatively determined during the tissue necrosis progression, but a better understanding of the cause of such lesions requires further studies.

Materials and methods
Eighteen specimens of the invasive coral Tubastraea tagusensis were collected in Búzios Island, on the northern coast of the São Paulo State, between 5 and 7 m deep. Sampling was carried out between August 28th and October 09th, 2017. All samples were readily frozen after collection. Part of the samples represented healthy coral colonies (n = 7), and the remaining specimens were from colonies displaying different stages of tissue necrosis: initial (n = 6) defined as colonies with a small necrotic region at the calicular margin (Fig. 1A, B); intermediate (n = 2) defined as polyps with more than 50% of the calicular margin affected (Fig. 1C); and advanced (n = 3). defined as fully necrotic polyps (Fig. 1D).  www.nature.com/scientificreports/ In all analyzed colonies (healthy and diseased) a fragment of approximately 25 mg (~ 5mm 2 ) of the calicular margin containing tissue and skeleton was used for total genomic DNA extraction, for diseased colonies, the extracted portion was exactly the diseased area. DNA was extracted using the DNeasy Blood & Tissue kit (QIAGEN), following the manufacturer's instructions. The quality and purity of extracted DNA were analyzed through agarose gel electrophoresis (1,5%) and spectrophotometry (NanoDrop), respectively.
Using the universal bacterial primers 27F and 519R (V1-V3 region-LANE 1991; Turner et al. 1999), a fragment of the 16S rDNA with approximately 600 bp was amplified using the Advantage HF 2 PCR kit, following concentrations recommended by the manufacturer and two-stage cycling for diseased colonies, as follows: 94 °C for 1 min followed by 38 cycles at 94 °C for 30 s and 68 °C for 1 min. For healthy colonies, PCR reaction was conducted with the following cycling: initial denaturation step 94 °C for 60 s, followed by 30 cycles at 94 °C for 30 s, 56 °C for 40 s, and 68 °C for 33 s, with a final extension step at 68 °C for 33 s.
Amplicons were purified with magnetic beads (Agencourt AMPure XP) following the manufacturer's instructions and eluted in 50 µl of 1X TE buffer (10 mM Tris-HCl, pH 8, 1 mM EDTA). Final sample concentrations were measured using the Qubit dsDNA BR Assay Kit. Libraries were assembled with the NEBNext Ultra II FS DNA Library Prep Kit and their respective concentrations and size distributions were verified using the Qubit dsDNA HS (High Sensitivity) Assay Kit and the Bioanalyzer High Sensitivity DNA Chip, respectively. Sequencing was performed on the MiSeq platform (Illumina) using the MiSeq Nano v2 kit (500 cycles), at the Facility Center for Research from the University of São Paulo (CEFAP-USP). Sequences were deposited in the SRA database (PRJNA637639 and PRJNA675612).
Low-quality strands of each sequence, as well as short reads (< 50pb), were removed using SolexaQA++ software 49 . Identical sequences were grouped using the Swarm software with d = 1 50 and then classified on the Mothur platform with a bootstrap cutoff of 80 51 using the database 16S Silva-v.132 (SDB) 52 . Statistical analyses were performed using the STAMP-Statistical Analysis of Metagenomic Profiles software 53 , after the removal of eukaryotic, chloroplast, mitochondria, and unknown sequences. Statistical tests were performed for multiple groups through variance analyses (ANOVA) and post-hoc Turkey-Kramer tests and Benjamini-Hochberg FDR correction of multiple tests, being 0.01 the adopted p-value filter to identify the relative frequency of OTUs for each necrosis stage and their and their differences. The Whites' two-sided nonparametric t-test, CI method (DP bootstrap, Benjamini-Hochberg FDR multiple test correction, and p-value filter of 0.01) was applied to compare the microbiota between healthy colonies and those at the initial necrosis stage. The diversity parameters were derived from the classification tables (phylum, class, order, family, and genera), such as the wealth estimators Shannon, Neff Shannon, Simpson, and Neff Simpson (Simpson's inverse index). All p-values taken into account were the adjusted p-value after correction.

Results
Tubastraea tagusensis tissue necrosis generally starts at the tissue from the calicular margin like a small brown dot (Fig. 1A, B) that later expands through the polyp (Fig. 1C), sometimes leading to the death of several polyps (Fig. 1D). When colonies with tissue necrosis were manipulated, the skeleton of the diseased region was significantly fragilized and brittle. From the tissue and skeleton of healthy and diseased corals, we obtained 738,462 classified sequences (averaging 220 bp) representing bacteria, chloroplasts, mitochondria, unknown, and eukaryotes. On average, the number of obtained taxonomic classifications from diseased corals (58,564) was four times higher than that from healthy corals (13,465). Sequences from eukaryotes, chloroplasts, mitochondria, and unclassified were removed from statistical analyses. The number of bacterial reads associated with target samples was, on average, 19,268 per coral colony displaying tissue necrosis, and ~ 1538 per healthy coral colony ( Table 1).
The richness index as well as the diversity and the effective number of species in both Shannon and Simpson indexes are higher in those colonies presenting tissue necrosis than that from healthy ones. Richness varies considerably between necrotic colonies (261-537), although there was no observable pattern between different necrosis stages. The greatest richness was found in advanced (514-sample Q) and initial (537-the sample I) stages of the disease. Regarding diversity, there is, on average, a fourfold increase in the number of effective genera from healthy to diseased colonies (Shannon Neff from 8.07 to 34.28 and Simpson Neff Simpson from 3.48 to 14.06).
Among Bacteria, 1,021 OTUs were detected, of which 213 occurred in both healthy and diseased colonies. These 213 common OTUs represent 96.8% of the microbiota abundance associated with healthy corals and an average of 83.6% of the abundance of microorganisms associated with diseased colonies. From the 1,021 identified OTUs, only 187 were more abundant in healthy corals, whereas the remaining predominated in diseased ones. In total, 742 genera were exclusively associated with diseased corals, representing 14.86% of their OTUs abundance, while 66 genera were exclusive to healthy colonies and represent 3.19% of their bacterial abundance (Fig. 2). Among the shared OTUs between healthy and diseased coral colonies, 47 displayed significantly different relative frequencies (< 0,01), and, of these, 25 were more frequent in healthy corals.
At the genus level, even though most microorganisms have a relatively low frequency, it was possible to infer that some constitute most of the microbiome and markedly differ between coral conditions. In healthy colonies, 20 genera (Fig. 3A) represent 89.2% of the microbiota, with Rubrobacter, Marine_Methylotrophic, and Idiomarina being exclusive to them. The remaining OTUs are significantly abundant in diseased corals, although in a lower proportion (61.1%) of its associated community. Of the 20 most abundant OTUs in diseased colonies (Fig. 3B) (which encompass 74.1% of the detected microbiota), a single one was exclusive to diseased states (Cyclobacte-riaceae_unclassified), while the remaining represent 82% of the microbiome associated with healthy colonies.
Comparisons of the microbiome from healthy and the observed stages of coral tissue necrosis suggest that 15 phyla compose nearly all the microbiota associated with T. tagusensis. Those same 15 phyla represent the majority of taxa, regardless of the disease condition and despite the presence of additional 22 Table 2). When monitoring the microbial succession in the analyzed stages of tissue necrosis, it is clear that most Bacterial groups present in the microbiome of healthy colonies are also in diseased ones. When the initial signs of tissue necrosis are observed (initial phase), the richness and total reads have a significant increase, but 85.4% (200 OTUs) of microorganisms identified in healthy colonies are also identified in this stage. In the intermediate stage of infection, the number of OTUs increases if compared to healthy colonies (Fig. 4) but decreases 39% concerning the initial stage of the disease. In advanced necrosis stages, a rise in microorganism abundance is observed when     The microbial core (persistent microorganisms found in all specimens of a particular holobiont species) of T. tagusensis is comprised of 8 genera (Zanotti et. al 2020). Of these, Rubrobacter is absent in all necrotic colonies, Acinetobacter is absent in the intermediate stage, Enhydrobacter is absent in the intermediate and initial stages, and Hydrogenophilus is absent in intermediate and advanced tissue necrosis stages. The microbial core is responsible for only a part of the microbial community of healthy colonies (6.68%), but it is significantly lower in comparison to that from diseased colonies, representing only 2.98%, 2.22%, and 1.39% in colonies at initial, intermediate, and advanced stages respectively. Ruegeria was more frequent in diseased colonies, representing    (Table 3).
To identify potential bacterial triggers of tissue necrosis in T. tagusensis, special attention was given to the community of microorganisms associated exclusively with the initial stage of the disease, and its comparison to those from healthy coral colonies. Among them, 65 significant differences were observed (p value between 0 and 0.0096), but genera that were not linked to disease were excluded, such as those more frequent in healthy colonies (4) and those that were not detected in all diseased colonies in the initial stage (30). Therefore, 31 OTUs fitted these criteria, of which only 13 were classified to the genus level. To improve robustness, genera with less than 10 reads per colony were excluded from further characterization as disease potential triggers. Finally, besides being identified in all colonies during the initial stage of necrosis, these 21 OTUs (Table 4) were also present in all subsequent stages of the disease.

Discussion
Changes in the microorganism community associated with the colonies of the invasive coral Tubastraea tagusensis are significant from the first signs of tissue necrosis. Such changes lead to increased microbial diversity, richness, and the effective number of genera (e.g. bacteria), which is compatible with previous studies on coral diseases [54][55][56][57] . However, most of the microbiota associated with diseased corals (~ 83.6%; 213 OTUs) is present Table 3. Mean relative frequencies of genera that comprise the microbial core in healthy and at the initial (DI), intermediate (DM), and advanced (DA) stages of the disease in Tubastraea tagusensis.  Table 4. Classification of the 21 OTUs found in all colonies presenting necrosis and with a tendency of increasing relative frequency as the disease progresses in the coral Tubastraea tagusensis. www.nature.com/scientificreports/ in healthy colonies. Thus, despite the detection of nearly 750 OTUs exclusive in diseased colonies, they did not replace the microbiota associated with healthy hosts but instead reduced it as they grew. Despite the community shift, the relative frequency of bacterial OTUs found exclusively in diseased colonies is low, with the highest disease-specific genus composing 5.1% of the total microbiota in the intermediate stage (unclassified Cyclobacteriaceae). Such low frequency may indicate that the majority of the microbiota in diseased corals is composed of opportunistic species and/or secondary colonizers 55,56 , which survived in the coral due to the imbalance of the microbial community caused by the primary infection. However, the opposite has also been detected (e.g., 63 OTUs present only in healthy colonies, all with low relative frequencies [an average of ~ 0.097%, representing 3.19 of the total abundance]). We also observed that the necrotic affected area is not directly related to the diversity of associated microorganisms. In the initial stage, in which it has less than 0.25 cm 2 , we identified a greater diversity of microorganisms. Also, this is a stage of tissue necrosis that has the most exclusive diversity, suggesting candidate triggers of the disease. In the intermediate stage, which displayed an enlargement of the area affected by the necrosis (Fig. 1C), the number of identified OTUs is lower compared to that from the onset of the disease, possibly indicating a holobiont response. At the advanced stage, the number of identified genera is higher than that in the previous stage, but such increment is not significant compared to the beginning of the infection. Such a variation in the microbiome indicates a rapid destabilization of the symbiont community during the initial stages of T. tagusensis infection, and subsequent colonization by opportunistic microorganisms, as observed in other diseases from zooxanthellate counterparts 55,58 .
In parallel to the microbiome related to necrosis, attention was given to the microbial core. Among the eight genera considered to be part of the microbial core of T. tagusensis (Zanotti et al. 2020), four were not found in colonies presenting tissue necrosis (Table 3). Such a difference might be a result of the microbiome imbalance caused by the disease. Nonetheless, the relative frequency of four microbial core genera was significantly lower in polyps presenting necrosis. However, Ruegeria had different patterns once it was the most abundant genera in diseased colonies. This genus is known to be associated with several healthy 14,59-61 and diseased zooxanthellate corals 10,11,14,[22][23][24] . Previous studies suggested that Ruegeria inhibits/controls the growth of other bacteria genera through tropodithietic acid 62 . Also, Ruegeria inhibited a widely known opportunistic coral pathogen, Vibrio coralliilyticus 63 .
Statistical analyses of the microbiome also indicate that within the bacterial OTUs identified, 21 (Table 4) might represent pathogenic ones. Of these, only five were classified to genus level, one of them being Ruegeria, which as mentioned has no pathogenic profile. Within the remainder, only Arenicella has been associated with coral disease (skeletal growth anomalies in Platygyra carnosa 23 ), although the herein detected Sphingomonadaceae has also been described as a putative pathogen associated with Acropora cervicornis and A. palmata disease 64 . Besides that, although there is no previous report of Epibacterium in relationship to scleractinian corals (a group described associated with seaweed surfaces 65 ), its detection and increased abundance during T. tagusensis tissue necrosis may be related to the exposure of the coral skeleton to the environment. As the necrosis advances, the exposed skeleton becomes a potential substrate for several organisms, like filamentous algae.
Apart from the aforementioned OTUs, several remained unclassified at the genus level. Among them, the Flavobacteriaceae and Rhodobacteriaceae were found in high abundance in T. tagusensis tissue necrosis and were previously associated with the following scleractinian diseases: White band 55 ; White plague 7,66 ; Yellow-band 14 ; White syndrome 67 ; White-spot syndrome in Porites 11 ; White Plague 66 ; stony coral tissue loss disease 68 ; and Black band 10 . Additionally, the skeleton fragility in those colonies affected by the tissue necrosis may be related to a higher frequency of Mastigocoleus, a bacterial genus known to have bioerosion capabilities 69 .
Although concerning as another disease is reported to a scleractinian coral, the described tissue necrosis brings a glimpse of hope in the face of T. tagusensis unprecedented bioinvasion and spreading in the Southwestern Atlantic. Thus, despite its invasion capacity, the founding effect 44,70 , and the associated microbial community without major significant differences might be a disadvantage for this species, making it highly susceptible to diseases as the tissue necrosis reported herein.