Coral colonies included in this study were obtained through accidental by-catch from a collaborating fisher crew that operates with bottom-set gillnets on a small (<12 m hull length) vessel around Cape St. Vincent (Sagres, SW Portugal) (Dias et al., 2020) (Figure 1). Cape St. Vincent is characterized by complex oceanographic conditions, including strong upwelling events, high nutrient availability, and dynamic temperature gradients (Relvas and Barton, 2002; Haynes and Barton, 1990). The local coral community of SW Portugal is mainly dominated by soft coral taxa, such as the malacalcyonacean species Eunicella verrucosa, Paramuricea cf. grayi, Leptogorgia sarmentosa (Esper, 1794), and scleralcyonacean species like Ellisella paraplexauroides Stiasny, 1936, the black coral Antipathella subpinnata (Ellis and Solander, 1786), and the scleractinian corals Dendrophyllia cornigera (Lamarck, 1816) and D. ramea (Linneaus, 1758) (Dias et al., 2020; Nestorowicz et al., 2021).

This study was conducted with two main objectives investigating (1) the “in-situ microbiome” variation of coral garden corals and (2) “ex-situ microbiome” variation under aquarium conditions. The focus of the microbiome analysis was laid on the bacterial domain. In objective (1) “in-situ microbiome,” we focused on investigating the natural bacterial community composition and structure of 13 coral species of two octocoral orders (Malacalcyonacea and Scleralcyonacea), the black coral order (Antipatharia), and the stony coral order (Scleractinia) (Figure 1; Table 1). These corals were directly sampled from the fishing net aboard the fishing vessel (19th and 29th of August 2022) and this objective is defined as “natural microbiome” henceforth. The results of this objective will be presented and referred to as “wild” condition.

In (2) “ex-situ microbiome” we assessed how bacterial communities of two locally dominant gorgonians, E. verrucosa and P. cf. grayi [yellow lineage sensu (Coelho et al., 2023)] of Order Malacalcyonacea, responded to (2.1.) ex-situ conditions within the first 45 days of captivity in the tank system of Ramalhete Marine Station (University of Algarve—CCMAR, Faro, Portugal). Here, the natural microbiome samples of E. verrucosa and P. cf. grayi were included as a reference. This objective is referred to as “short-term captivity.” We also assessed microbial changes in E. verrucosa and P. cf. grayi to (2.2.) ex-situ conditions in two public aquaria for ~3 months (Zoomarine, Albufeira, Portugal) and ~1 year (Oceanário de Lisboa, Lisbon, Portugal). Here, we also included the samples of day 45 from the over-time, as well as the samples from the natural microbiome objective, and refer to this objective as “long-term captivity.” The sampling locations are hereafter referred to as Oceanário, Ramalhete, Zoomarine, and wild.

Coral samples were sprayed with filtered [0.2 μm Millipore Sterivex filters (Merck KGaA, Darmstadt, Germany)] and autoclaved seawater prior to sampling for genetic analysis in order to remove loosely associated microbes. Only specimens considered healthy based on visual inspection (absence of tissue necrosis, or weak coloration) were sampled. About 1.5 cm fragments of soft coral specimen branch tips were collected and 5 cm pieces of scleractinian corals (Supplementary Figure S1). From branching octocoral species, and the black coral, a small, apical branchlet fragment was taken. From the scleractinian corals, a fragment with one to three polyps was taken and from the sea pen, a few polyp leaves were sampled. The samples were preserved with aqueous DESS solution (20% salt-saturated DMSO, 0.25 M EDTA and saturated NaCl, adjusted to pH 8.0) and kept at −20°C until processing. In addition to coral samples, we also took water samples from the tank systems where corals were kept in captivity (Oceanário, Ramalhete, Zoomarine). Water samples were taken by filtering 1 L aquarium water through 0.2 μm Millipore Sterivex filters (Merck KGaA, Darmstadt, Germany) and filters were subsequently flash frozen in liquid nitrogen and stored at −80°C.

A total of 13 coral species were sampled. This included six biological replicates of the selection of 13 representative coral garden taxa following the regional species composition outlined by Dias et al. (2020), of which only one specimen each of Alcyonium sp. Linneaus, 1758 (Order Malacalcyonacea), Antipathella subpinnata (Order Antipatharia), and Callogorgia verticillata [(Pallas, 1766), Order Scleralcyonacea] was included in the natural microbiome objective, due to limited availability. Three of the six biological replicates per coral garden species and the three species with one biological replicate were selected for sequencing (Table 1). Tissue of each coral specimen was sampled aboard the vessel, immediately after reaching the surface and was subsequently preserved in DESS for genetic analysis (see above). The remaining colony or fragment of the colony was tagged and kept as taxonomic vouchers [except the sea pen Pennatula rubra Ellis, 1764 and soft coral Alcyonium sp.]. In total, 33 coral samples from 13 coral garden species were selected for microbiome analysis and sequenced.

Coral microbiomes have gained increasing interest over the last decades, and a high microbial host specificity has been frequently shown for both tropical shallow-water corals (Rosenberg et al., 2007; Ainsworth et al., 2015; Glasl et al., 2017), as well as for scleractinian cold-water corals (Galand et al., 2020; Appah et al., 2021). However, so far only a few representatives of coral garden forming corals were studied in detail and microbial records are available (Ransome et al., 2014; Van De Water et al., 2018; Keller-Costa et al., 2021). Our study aimed to contribute to a baseline for such temperate coral microbiomes.

We investigated E. verrucosa, L. sarmentosa, P. cf. grayi, I. elongata, and A. subpinnata, whose microbiome has been studied before in the Atlantic and Mediterranean (La Rivière et al., 2013; Ransome et al., 2014; Keller-Costa et al., 2017; Van De Water et al., 2018). The diverse coral community around Cape St. Vincent (Sagres, Portugal) comprises octocoral, scleractinian, and black coral species (Dias et al., 2020). The findings presented here not only expand the current microbiome dataset for coral garden-forming species but also deliver the first microbiome insights for previously undocumented species, shedding new light on the microbial diversity and host specificity in this ecologically important coral community. While this study provides the first broad assessment of the natural microbiome of coral garden species in the NE Atlantic, it is important to acknowledge limitations related to the sampling approach (via fisheries bycatch). While only visually healthy specimens were sampled, poor health conditions prior to collection may have introduced biases that affect the scalability of our conclusions. Nonetheless, host-specificity was observed across multiple coral taxa and consistently among biological replicates. Few examples of host specificity exist at the anthozoan order level, and our results reveal previously unreported marked differences in the microbiomes of different anthozoan orders (Figure 2). In particular, we identified a distinct microbiome associated with the Malacalcyonacea coral species studied (except Alcyonium sp. represented by one replicate), which displayed significantly lower alpha diversity and a unique microbial community composition compared to the other orders studied (Figure 2). We hypothesize that this relatively low variation and alpha diversity among Malacalcyonacea microbiomes could be linked to a homogeneous fan-like colony shape of the Malacalcyonacea species, and thus to low variation in the microbial niches made available by the host. Within this order, Alcyonium sp., which was only represented by one replicate, possessed a distinct microbiome from the fan-shaped representatives, which could be attributed to its different anatomical structure, and to the partially unresolved place of the genus within the Malacalcyonacea group which requires further revision (McFadden et al., 2022). However, given that this observation is based on a single specimen, it must be interpreted with caution. Limited replication prevents general conclusions, and further sampling will be required to verify whether this microbiome divergence is consistent across Alcyonium populations. Several factors, such as morphology and metabolic needs, can influence microbial host species-specificity by selecting for different microbial communities (McCauley et al., 2016; Camp et al., 2020). Further to coral morphology, depth-related environmental conditions (e.g., temperature, light availability, pressure) were not recorded in this study, but likely contributed to the observed species-specific microbiome variability, as microbial communities are known to vary along environmental gradients (Ransome et al., 2014; Keller-Costa et al., 2021; Reigel and Hellberg, 2023). Therefore, it remains unresolved if the observed microbiome specificity at the order level in Malacalcyonacea is driven by the inclusion of morphologically similar fan-shaped species, environmental conditions, or rather a true taxonomic signal, thus highlighting the importance of considering morphology and ecological traits in microbiome studies (McCauley et al., 2016; Camp et al., 2020). Microbial niches are potentially driven by the development of morphological microhabitats through increased body plan complexity [see McCauley et al., 2023 for a review], such as the development of skeletons in groups like Antipatharia, Helioporacea, and Scleractinia, and this process is suggested to result in increased microbial richness and diversity. Scleractinian corals are known to exhibit a strongly partitioned niche-specific microbiome in their mucus, tissue and skeletons (Rohwer et al., 2002; Rosenberg et al., 2007; van Oppen and Blackall, 2019), with higher microbial diversity in their mucus and skeleton, compared to their tissue (Marchioro et al., 2020; McCauley et al., 2023). In scleractinian corals, the skeletal endolithic community contributes to nitrogen and carbon fixation (Sangsawang et al., 2017; van Oppen and Blackall, 2019), whereas the mucus layer provides pathogen defense by providing a boundary niche for microbial commensals (Glasl et al., 2016; Marchioro et al., 2020). The coral tissue hosts stable, species-specific microbial communities (Rohwer et al., 2002; Glasl et al., 2016; Marchioro et al., 2020; Bergman et al., 2022). A similar case of high niche partitioning has not yet been observed in non-scleractinian coral species. However, distinct structural features are also well described within octocorals (McFadden et al., 2022). The potential presence of specific microbial communities within octocoral skeletal elements, such as calcitic sclerites or proteinaceous, and calcium carbonate axes in Malacalcyonacea and Scleralcyonacea (McFadden et al., 2022), akin to the endolithic communities found in scleractinian corals, remains largely unexplored. Similarly, whether the mucus-associated microbial community in octocoral species performs similar functions as those observed in tropical shallow-water scleractinian corals (Rohwer et al., 2002; Rosenberg et al., 2007; van Oppen and Blackall, 2019) remains to be determined. However, we found that coral garden species with high mucus production, specifically the octocorals I. cf. elongata, P. rubra, and the scleractinian Dendrophyllia spp., exhibited a high abundance of Rhodobacteraceae, a microbial group commonly associated with mucus production (Hall-Spencer et al., 2007; Frade et al., 2020; Marchioro et al., 2020; Keller-Costa et al., 2021). Since our study was conducted on homogenized coral fragments, the presence of multiple microhabitats with distinct microbial communities (Engelen et al., 2018) should be further addressed for coral garden species.

The gammaproteobacterial family Endozoicomonadaceae has been consistently reported as a dominant microbial component in tropical soft and hard corals (Ainsworth et al., 2015; Bourne et al., 2016; Pollock et al., 2018; Frade et al., 2020; Reigel and Hellberg, 2023), as well as in shallow-water populations of several temperate coral species (Bayer et al., 2013; La Rivière et al., 2013; Ransome et al., 2014; Keller-Costa et al., 2022). This study further confirms the consistently high abundance and prevalence of Endozoicomonadaceae across coral garden species (Figure 2; Supplementary Figure S8). The microbial family Endozoicomonadaceae is metabolically versatile, presumably contributing to coral holobiont health through vitamin translocation and chitin degradation, with recent research highlighting their chitinolytic activity as a key mechanism for nutrient acquisition in corals (Da Silva et al., 2023; McCauley et al., 2023).

In tropical photosymbiotic shallow-water corals, Endozoicomonadaceae have been frequently found in close proximity to the coral’s photosynthetic dinoflagellate endosymbionts (family Symbiodiniaceae). While dinoflagellates of corals produce Dimethylsulphoniopropionate (DMSP), and are thereby expected to provide carbon and sulphur to bacterial symbionts like Endozoicomonadaceae (Yost and Mitchelmore, 2009; Bourne et al., 2013), a recent study has shown that aphotic environments also contribute to the global DMSP production (Zheng et al., 2020). Here, we found microbes that metabolize DMSP, such as Endozoicomonadaceae, Alteromonadaceae, and Pseudomonadaceae (Bourne et al., 2013), to be potential core microbes in coral garden forming corals, similar to tropical shallow-water corals (Bourne et al., 2016; Frade et al., 2020; Haydon et al., 2022). Our data strongly supports the dominance of Endozoicomonadaceae within non-photosymbiotic coral garden species, such as A. subpinnata, C. verticillata, F. quadrangularis, and I. cf. elongata, reinforcing their importance beyond shallow-water symbiotic coral systems. In particular, microbiome observations for Alcyonium sp., A. subpinnata, and C. verticillata are based on a single sample (n = 1), which limits the generalizability of these findings and only indicate potential microbial patterns, which require further scientific support. However, the high dominance of Endozoicomonadaceae align with previous research outlining their association with anthozoans that lack photosymbionts, including a significant proportion of non-scleractinian corals (McCauley et al., 2023). The high presence of these microbial groups in coral garden species (Figure 2; Supplementary Figure S7) is most likely linked to their metabolic and immune-protective role for the coral host (Neave et al., 2016; Pogoreutz et al., 2022). Yet, future research needs to show if the production of DMSP can also be found in non-photosymbiotic corals and their associated microbes, similar to photosymbiotic corals (Raina et al., 2010; Frade et al., 2016). Juvenile apo-symbiotic corals can indeed produce DMSP, implying that non-photosymbiotic corals, including those in deep-sea environments, may have the capacity to produce DMSP independently of photosynthetic symbionts (Raina et al., 2010), offering therefore a potential explanation for the high abundance of Endozoicomonadaceae in the coral species we studied. This could give insights not only into the role of coral gardens in the context of climate regulation, but also into the role of DMSP under stress conditions for non-photosymbiotic corals.

Besides the association between Endozoicomonadaceae and photosymbiosis, these microbes have been generally linked to corals in shallow-water environments (Morrow et al., 2012; Bayer et al., 2013; Apprill et al., 2016; Neave et al., 2016; Frade et al., 2020) and were reported to be rare or absent in deep-sea corals (Kellogg, 2019; Appah et al., 2021). Here, we demonstrate that Endozoicomonadaceae, particularly OTU 50381 (Endozoicomonas), was the most abundant and dominant microbe in specimens of coral species from the order Malacalcyonacea predominantly sampled at depths ranging 60–80 m (e.g., L. sarmentosa and Paramuricea sp.) and 80–100 m (e.g., Alcyonium sp. (n = 1), Eunicella verrucosa, L. sarmentosa, and P. cf. grayi), but much less abundant in species occurring at greater depths of 100–400 (e.g., C. verticillata (n = 1), E. paraplexauroides, I. cf. elongata, F. quadrangularis, P. rubra, and A. subpinnata (n = 1), Figure 1; Supplementary Figure S4). Lower pH and temperatures found in deep-sea environments may limit the growth of Endozoicomonaceae (Kellogg, 2019). Our results indicate a dominance of Endozoicomonadaceae in corals sampled down to ~100 m depth and suggest that the prevalence of Endozoicomonadaceae may be linked to their ecological role in shallower coral habitats, underscoring their potential importance in supporting coral holobiont health within these environments.

In the present study, the microbiome of P. rubra, I. cf. elongata, and F. quadrangularis (Scleralcyonacea) showed a substantially high presence of Spongiibacteraceae (Figure 2; Supplementary Figure S7). This group is known to degrade hydrocarbons that probably facilitate energy acquisition and thus provides additional energy resources for corals in deep nutrient-poor environments (Quintanilla et al., 2022). Interestingly, Spongiibacteraceae also displayed increased abundance in species occurring in shallower-water within the Malacalcyonacea family, particularly in species such as Eunicella verrucosa and Leptogorgia sarmentosa (Cúrdia et al., 2013). Rhodobacteraceae and Flavobacteriaceae, which are often observed in response to elevated nutrient levels (Frade et al., 2020; Keller-Costa et al., 2021), were also found to be highly abundant, prevalent, as well as indicative of coral garden species (Figures 2, 5; Supplementary Figure S7). These microbes are also associated with amino acid biosynthesis, nitrogen fixation, and sulfur metabolism in healthy corals (Hall-Spencer et al., 2007; Connelly et al., 2023). Contradicting to the presence of Spongiibacteraceae, which support corals in nutrient poor environments, this high presence of Rhodobacteraceae and Flavobacteriaceae may be linked to elevated nutrient levels associated with the nutrient-rich upwelling region around Cape St. Vincent in SW Portugal, the submarine canyon (Haynes and Barton, 1990; Relvas and Barton, 2002; Nestorowicz et al., 2021), and the predominant hydrology (Relvas and Barton, 2002; Rautenbach et al., 2024). Altogether, these nutrient-associated microbes could reflect natural variability in nutrient availability, as well as adaptive responses to environmental conditions, with potential links to terrestrial runoff and estuarine input in the region.

Our study has also identified Vibrionaceae as a predominant and persistent member of the core coral microbiome (Figure 2; Supplementary Figure S7). Vibrionaceae are frequently identified as coral pathogens and linked to disease outbreaks in tropical and temperate shallow-water corals and mortality events (Hall-Spencer et al., 2007; Glasl et al., 2017; Frade et al., 2020; Keller-Costa et al., 2021; Connelly et al., 2023), and have been observed to increase in abundance under stressful conditions and during disease onset in temperate corals (Hall-Spencer et al., 2007; Corinaldesi et al., 2022). However, recent research indicates a more complex role. For example, Keller-Costa et al. (2021) found no significant increase in Vibrio abundance in the microbiomes of necrotic Eunicella gazella tissue compared to healthy tissue. This suggests that the presence of Vibrionaceae in our study, particularly OTU 50351, may simply reflect a natural component of the coral microbiome rather than an indication of pathological conditions. In contrast to other temperate coral species studied to date, in which the family Spirochaetaceae is usually found as a regular member (Van De Water et al., 2016, 2017; Keller-Costa et al., 2022; McCauley et al., 2023), we did not find this group to be part of the core microbiome of the species analyzed, with one exception. One OTU from this family (OTU 46622, Spirochaetaceae) was abundant in Alcyonium sp. (n = 1) (Malacalyconaceae) and in some members of Scleralcyonacea [C. verticillata (n = 1), P. rubra, I. cf. elongata]. The identified abundance of Spirochaetaceae in Alcyonium sp. (Malacalcyonacea) and C. verticillata (Scleralcyonacea) only serves as an indication for a high occurrence in these species, due to the limited replicate number (n = 1). Spirochaetaceae are generally associated with healthy octocorals, support nutrient cycling and development, and have been suggested to even influence the color of the precious red coral Corallium rubrum (Van De Water et al., 2016; Van De Water et al., 2018; Bonacolta et al., 2021; McCauley et al., 2023). However, despite the high abundance in Scleralcyonacea the precise functions of Spirochaetaceae in the coral microbiome remain unclear and require further investigation. Pseudomonadaceae emerged as the second most prevalent and abundant microbial group in our study (Figure 2; Supplementary Figure S7). Pseudomonadales, including Pseudoalteromonas, have frequently been associated with healthy corals and are known for their role in the biosynthesis of amino acids and B vitamins (Keller-Costa et al., 2022; McCauley et al., 2023) and were found to be abundant in the coral tissue, skeleton, and mucus, highlighting their functional importance within and across different coral compartments (Keller-Costa et al., 2022; McCauley et al., 2023). Although we did not analyze specific microhabitats within the corals sampled, our findings provide a basic understanding of the microbial composition in coral garden species and highlight the inclusion of Pseudoalteromonadaceae as an integral component that should be further investigated in future research.

A recurring “rule of thumb” is that aquarium conditions induce a rapid decline in alpha diversity within tropical coral holobionts but diversity often recovers upon reintroduction to the natural environments (Vega Thurber et al., 2009; Glasl et al., 2016). Our study revealed that despite the generally homogeneous microbiome of the Malacalcyonacean studied, E. verrucosa and P. cf. grayi exhibit distinct microbial responses under similar ex-situ conditions. We observed that E. verrucosa exhibited rapid microbial changes, particularly within the first 12 h and continuing over 45 days. This response of E. verrucosa aligns with the concept of “microbiome conformers” (Ziegler et al., 2019; Van De Water et al., 2020), which are species with a high degree of microbial flexibility, as opposed to “microbiome regulators” (Ziegler et al., 2019; Van De Water et al., 2020), which maintain a more stable holobiont structure and composition, as was observed for P. cf. grayi in our study. However, the distinction between microbial conformers and microbial regulators appears to be a continuum, with coral species displaying varying degrees of microbial regulation or conformity (Keller-Costa et al., 2021). Our results align with previous studies showing that E. verrucosa exhibits an immediate microbial response to stress, a reaction similarly observed in other temperate coral species such as L. sarmentosa (Ransome et al., 2014; Keller-Costa et al., 2022). It is also possible that the specific conditions under which the corals are maintained, including the stressors they encounter, may better suit some species than others, resulting in the different responses shown by P. cf. grayi and E. verrucosa. While P. cf. grayi showed greater microbiome stability under short-term captivity, the physiological basis for this remains unclear, as no coral health metrics or stress indicators were collected in this study. Future research combining microbiome data with host physiology could help elucidate mechanisms of resilience.

Closely related species like Eunicella cavolini have been identified as microbiome regulators due to their low microbial variation under temporally and spatially changing environmental conditions (Van De Water et al., 2020). Differences in the microbial responsiveness within the genus Eunicella, might be attributed to the geographic distribution of E. cavolini, extending from the Mediterranean Sea into the Atlantic (Ransome et al., 2014; Aurelle et al., 2017), which may have driven the development of a more flexible microbiome in E. cavolini, compared to E. verrucosa which shows a geographically wide distribution within the Atlantic. Future studies are needed to conduct a distribution wide comparative analysis on species within the Eunicella genus and across other coral garden taxa, specifically to identify microbial response patterns to changing conditions to develop targeted restoration and conservation approaches for coral garden forming species.

The observed significant decrease in alpha diversity of E. verrucosa kept over 45 days in the aquarium compared to wild-collected samples could be linked to a lower regulative capacity of the host, resulting in a higher transmissibility of microbes. Yet, microbial changes in E. verrucosa occurred independently of the water community, which aligns with earlier studies emphasizing the different microbial composition of corals compared to the surrounding water communities (Ransome et al., 2014; Glasl et al., 2016). While the transient nature of microbiomes is presumably beneficial to enable the host to rapidly adapt to dynamic environments (Frade et al., 2020), severe environmental stress can weaken the host’s immune defense, and pathogens and opportunistic microbes can invade the coral and proliferate, ultimately leading to the formation of a pathobiome (Stévenne et al., 2021; Voolstra et al., 2024). Fast acclimation processes take place in rare prokaryotes, while significant community-level shifts in core microbial communities tend to occur after long-term environmental changes (Ransome et al., 2014; Hernandez-Agreda et al., 2016; Haydon et al., 2022; Reigel and Hellberg, 2023). Here, we also found that core microbes remained stable in E. verrucosa and P. cf. grayi over short-term aquarium maintenance (Figure 5).

As microbial regulators or conformers, corals possess different ways to maintain a beneficial host microbiome. One mechanism is periodic mucus shedding (Glasl et al., 2016). Another way to control and/or prevent a microbial shift can be through biochemical pathways that favor certain groups of microbes (Raina et al., 2013; Jones and King, 2015; Ochsenkühn et al., 2018). For instance, photosymbiotic scleractinian corals can expel Symbiodiniaceae symbionts under stress or adjust their symbiont composition by switching to more temperature- and light-resistant strains (Teplitski et al., 2016; LaJeunesse et al., 2018; Voolstra et al., 2024). The existence of similar mechanisms in non-photosymbiotic corals to “switch” to beneficial microbes and to selectively expel unwanted microbes (Teplitski et al., 2016) remain largely unexplored. However, “microbial switching” with exchanged OTUs within the same bacterial genus has recently been documented in other gorgonian species under stressed or altered conditions, both in symbiotic and non-symbiotic taxa (Van De Water et al., 2018; McCauley et al., 2020; Montaño-Salazar et al., 2023), supporting the observation that ex-situ maintenance can lead to significant holobiont restructuring. Drivers for distinct microbial changes in the two corals E. verrucosa and P. cf. grayi (and corals in general), whether mediated molecularly, behaviourally or by immunity, need to be further investigated in order to improve our understanding of the different levels of active and passive microbiome alterations in corals.

In contrast to the typical short-term responses of the microbial conformer E. verrucosa and the microbial regulator P. cf. grayi in the different aquaria, both coral species exhibited significant microbial shifts with increased alpha and beta diversity and notable changes in microbial community composition after 1 year at the Oceanário. Galand et al. (2018) highlighted that coral colonies typically require 1 year to fully acclimate to aquarium conditions. Whether our findings are primarily due to the length of time in captivity, or are influenced by specific aquarium conditions, such as water quality, nutrition, and other factors, remains to be explored. The water quality and nutrient content was not controlled in this study, but by comparing microbial communities in this study we found that changes in the microbiome of E. verrucosa or P. cf. grayi did not correlate with the microbial community in the water samples, which concurs with earlier publications (Ransome et al., 2014; Glasl et al., 2016). Water samples from Oceanário displayed differences in the microbial composition (Supplementary Figure S17) and a distinct separation (PCoA, Supplementary Figure S18) between water samples, but due to limited sample size, the observed difference requires further investigation and statistical support. In particular, higher beta diversity (high dispersion among samples) and a distinct microbial community structure have often been associated with a reduced regulatory capacity of the host and thus often considered to be the result of a dysbiotic state in corals, typically reflecting stress or disease of the host (Thurber et al., 2017; Zaneveld et al., 2017; Keller-Costa et al., 2021; Voolstra et al., 2024). In fact, E. verrucosa and P. grayi showed a greater interspecific dissimilarity in the Oceanário samples than in the specimens kept at the Zoomarine public aquarium, as well as at Ramalhete and in the natural environment (Figure 3; Supplementary Figure S19). It has been shown that destabilization of the typical microbial community under stress is strongly associated with a decline in essential Gammaproteobacteria, such as Endozoicomonadaceae, which commonly characterize a healthy microbiome in corals (Ransome et al., 2014; Keller-Costa et al., 2021; Voolstra et al., 2024). Consistent with this, we observed a reduction in Endozoicomonas (OTU 50381) abundance in E. verrucosa and P. cf. grayi samples from Oceanário, potentially indicating a shift toward a dysbiotic state under these aquarium conditions for 1 year (Figures 4, 5C). Yet, this decline was not significant (Figure 5) and Oceanário has been recognized by its success in gorgonian long-term maintenance, where coral colonies consistently appeared healthy and in good condition. At the same time, we observed a rise in opportunistic microbes such as Rhodobacteraceae, Sphingomonadaceae, and Saccharospirillaceae, along with a significant increase in Vibrionaceae in P. cf. grayi samples at the Oceanário (Figures 4, 5C), which could indicate a dysbiotic state of the corals kept for 1 year in the aquarium. Other microbial groups in samples of E. verrucosa and P. cf. grayi at Oceanário, including Saccharospirillaceae, Rhodobacteraceae, and Sphingomonadaceae also showed an increased abundance (Figure 5). Saccharospirillaceae are known for their adaptability to fluctuating environmental conditions through aerobic and anaerobic respiration (Campbell et al., 2022). Rhodobacteraceae are associated with corals (La Rivière et al., 2013; Keller-Costa et al., 2021; Luo et al., 2021) and linked to dysbiosis under changing environmental conditions or coral disease in both temperate (Keller-Costa et al., 2021; Corinaldesi et al., 2022) and tropical regions (Frade et al., 2020). The increased presence of these families, particularly in relation to nutrient cycling (Appah et al., 2021), may reflect an enrichment of specific nutrients in the aquarium environment, potentially in response to feeding processes under human care, promoting the growth of opportunistic species and indicating a shift in the microbial community structure under different trophic conditions (Pogoreutz and Ziegler, 2024). Common aquarium maintenance practices, such as tank cleaning and water changes, as well as environmental parameters (light, temperature, flow) can lead to shifts in microbial assemblages, affecting nutrient cycling in aquarium systems, and thus affecting the reared organisms (Bik et al., 2019). However, not all environmental parameters (e.g., light intensity or spectrum) were controlled or recorded here, potentially impairing direct ex-situ comparisons. This especially applies for mesophotic species, as the observed microbiome shifts could reflect both the effects of captivity and mismatches in the naturally prevailing environmental parameters. Overall, while opportunistic microbes increased, the relatively stable presence and prevalence of core microbes with a potentially beneficial role, such as Endozoicomonadaceae and Pseudomonadaceae, may indicate that the pathogen defense role provided by these groups (Ransome et al., 2014; Keller-Costa et al., 2021) may still be maintained under long term aquarium conditions. Therefore, whether this microbial response of the coral colonies after 1 year at Oceanário is indicative of a dysbiotic state or an altered microbiome under human care needs further investigation.

Studies on tropical and cold-water coral species have shown that microbiomes respond rapidly and species-specifically to captivity, underlining the relevance of these responses for conservation efforts. For example, Barreto et al. (2021) emphasized the difficulty of simulating natural environmental conditions in captivity and the resulting implications for microbial community dynamics. Rapid shifts in the coral surface mucus layer microbiome were observed after 1 day in captivity in tropical species such as Fungia granulosa (Kooperman et al., 2007), Porites astreoides (Glasl et al., 2016), and Siderastrea siderea (Pratte et al., 2015). In contrast, microbiome stability has been observed in Madrepora oculata, which maintained a stable microbial community over 6 months (Galand et al., 2018), and in Acropora loripes over 4 weeks (Hill and Ralph, 2006; Damjanovic et al., 2020). Our findings align with these studies indicating species-specific differences in response to aquarium conditions: Eunicella verrucosa showed rapid microbiome shifts under short-term captivity, similar to Lophelia pertusa (Galand et al., 2018), while Paramuricea cf. grayi maintained greater microbiome stability. Our study supports recent findings on microbial taxa of potential interest for microbiome stewardship approaches (Peixoto et al., 2022), such as Endozoicomonadaceae, Pseudomonadaceae, and Spongiibacteraceae, yet their functional roles in coral garden species resilience remain to be further experimentally validated. Ongoing work is investigating whether coral garden species show post-reintroduction variation in survival after microbiome shifts under captivity, as demonstrated for tropical corals (Glasl et al., 2016), and will assess how microbial consortia aid restoration outcomes for coral garden species.