A total of 56 L. cf. striatus colonies were collected during the Red Sea Deep Blue Expedition in October and November 2020 (15 colonies), the Red Sea Decade Expedition from February to June 2022 (32 colonies), and the OceanX Relationships Cultivation Expedition in June 2022 (9 colonies) on board the M/V OceanXplorer along the Saudi Arabian Red Sea coast ( Appendix S1 ). Sampling occurred in regions spanning the whole latitudinal range of the Saudi Arabian coast from the Gulf of Aqaba (GoA) (4 sites) and northern Red Sea (NRS) (8 sites) in NEOM waters to the Al Wajh region (AlW) (4 sites), central Red Sea (CRS) (namely Yanbu and Thuwal) (10 sites) and southern Red Sea (SRS) (4 sites) ( Appendix S2 ). The two distinct sampling regions in NEOM waters were sampled both in Fall 2020 and in Spring 2022, thus allowing us to investigate the Symbiodiniaceae community composition stability across time and between two seasons.

The entire coral colonies, or fragments, were collected between 70 and 127 m water depth using an Argus Mariner XL Remotely Operated Vehicle (ROV) or a Triton 3300/3 submersible with a Schilling T4 hydraulic manipulator. The ROV and the submersible dives were video-recorded, and frame grabs of the colonies were extracted from the videos using the open-source software MPC-HC (Media Player Classic – Home Cinema) and Adobe Premier software PRO™, respectively. The underwater vehicles position was provided by Kongsberg HIPaP 501 USBL (Ultra-Short Baseline), Sonardyne Sprin INS (Inertial Navigation System), and Sonardyne Ranger Pro 2 USBL.

After sampling, a small fragment of each colony, or the whole colony, was preserved in absolute ethanol for molecular analyses. The remaining part of the corallum was bleached in sodium hypochlorite for 48 hours to remove organic tissue parts, rinsed with fresh water, and air-dried for morphological identification. The coral skeletons and tissue samples are deposited at the King Abdullah University of Science and Technology (KAUST, Saudi Arabia).

Temperature, salinity, and depth at the sampling localities were recorded using a RBR Maestro CTD mounted on the ROV. Downcast RBR Maestro CTD data were extracted and visualized with Origin Pro 2020 (Origin Lab) software. Standard deviation (SD) was used as an uncertainty metric. RBR Maestro CTD raw data is available in Appendix S3 .

Symbiodiniaceae genomic DNA was extracted from the coral tissues using the DNeasy® Blood and Tissue kit (Qiagen Inc., Hilden, Germany), following the manufacturer’s protocol. Symbiodiniaceae genotypes were characterized using PCR amplification of the ITS2 region for the Illumina MiSeq platform in the KAUST Bioscience Core Laboratory. The primer sequences were (overhang adapter sequences underlined): 5’ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGAATTGCAGAACTCCGTGAACC 3’ (SYM_VAR_5.8S2) and 5’GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCGGGTTCWCTTGTYTGACTTCATGC 3’ (SYM_VAR_REV) (Hume et al., 2018). PCRs were run with 11 μL of 2X Multiplex PCR kit (Qiagen Inc., Hilden, Germany), 2μL of 10 μM of each primer, and 5 μL of DNA, in a total volume of 25 μL. The following PCR conditions were used: 15 min at 94°C, followed by 30 cycles of 95°C for 30 s, 56°C for 90 s, 72°C or 30 s, and a final extension step of 10 min at 72°C. PCRs success was tested with QIAxcel Advanced System (Qiagen Inc., Hilden, Germany). Amplified samples were cleaned with Agencourt AMPPure CP magnetic bead system (Beckman Coulter, Brea, CA, USA). Nextera XT indexing and sequencing adapters were added via PCR (8 cycles) following the manufacturer’s protocol. Samples were normalized and pooled using SequalPrep™ Normalization Plate Kit 96-well (ThermoFisher Scientific, Waltham, MA, USA). The samples were then checked with Aligen BioAnalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA) and qPCR (ThermoFisher Scientific, Waltham, MA, USA) to check library size and concentration. Libraries were sequenced using the Illumina MiSeq platform (Illumina, San Diego, CA, USA) and kit reagents v3 (2 x 300bp pair-ended reads) at KAUST Bioscience Core Lab, following the manufacturer’s protocol.

Demultiplexed forward and reverse fastq.gz files were submitted online to the SymPortal framework (https://symportal.org; Hume et al., 2019). A standardized quality control (QC) of sequences was conducted as part of the submissions using mothur 1.39.5 (Schloss et al., 2009), the BLAST+ suite of executables (Camacho et al., 2009), and Minimum Entropy Decomposition (MED; Eren et al., 2015; Hume et al., 2019). Then, existing sets of ITS2 sequences on the database were used to find and assign ITS2 profiles to samples. The Symbiodiniaceae genotypes are represented as proxies by the ITS2 profile predictions in the SymPortal outputs. The online framework was used to download the post-MED ITS2 sequence relative abundances ( Appendix S4 ), the ITS2 profiles relative abundances ( Appendix S5 ), and the corresponding absolute counts between sample distances and between profile distances. We created stacked bar plots to relate the Symbiodiniaceae genotypes to different categorical levels (i.e., body of water, sampling depth) with the R package ggplot2 (Wickham et al., 2016).

The Symbiodiniaceae type profiles of Porites lutea Milne Edwards & Haime, 1851, and Porites columnaris Klunzinger, 1879, from Terraneo et al. (2019a), obtained through SymPortal submission, were taken into consideration for this study. In particular, patterns of specificity and generalism and comparison with algal communities of L. cf. striatus (this study) and P. lutea and P. columnaris (Terraneo et al., 2019a) were visualized using Venn Diagrams (Heberle et al., 2015).

The scleractinian coral species studied in this work ( Figure 1 ) belongs to the family Agariciidae in the complex clade (Kitahara et al., 2016). To date, it has not been studied from a molecular point of view and is currently synonymized with Leptoseris hawaiiensis Vaughan, 1907 (Hoeksema and Cairns, 2023). However, a morphological study of the type material of both species (illustrated in Dinesen, 1980) has shown consistent differences in corallite shape, size, and number of radial elements. Previous studies in the Red Sea addressing the physiology of this mesophotic specialist have identified it as Leptoseris fragilis (Schlichter et al., 1985, 1986, 1994, 1997; Fricke et al., 1987; Schlichter and Fricke, 1991; Kaiser et al., 1993; Ferrier-Pagès et al., 2022). However, Benzoni (2022) recently re-described the type material of L. fragilis and provided its first detailed description based on which all diagnostic characters based on skeletal morphology differ from the material we examined.

In Fall 2020, considering this study depth range of 70-130 m, the GoA and the NRS were comparable in terms of water temperature and salinity (25.9 ± 0.7 °C, 40.7 ± 0.05, respectively, for the GoA, and 25.5 ± 0.6 °C, 40.7 ± 0.04, respectively, for the NRS) ( Figure 2 ; Appendix S3 ). In Spring 2022, the coolest and most saline basin was the GoA (22.9 ± 0.3 °C, 40.33 ± 0.04), followed by the NRS (23.05 ± 0.3 °C, 40.3 ± 0.02), AlW (23.1 ± 0.3 °C, 40.2 ± 0.09), Yanbu (23.8 ± 0.3 °C, 39.9 ± 0.1), Thuwal (25.2 ± 1.4 °C, 39.6 ± 0.5), and finally, SRS (25.8 ± 1.9 °C, 39.4 ± 0.6) ( Figure 2 ; Appendix S3 ), with the temperature increasing from the North to the South, and the salinity decreasing along the North to South gradient, in agreement with the already known water seasonal and latitudinal gradient previously described from the basin (Sofianos and Johns, 2003; Raitsos et al., 2013; Rowlands et al., 2014, 2016; Chaidez et al., 2017; Berumen et al., 2019a; Manasrah et al., 2019).

A total of 6,876,798 sequences were produced using Illumina MiSeq and then submitted to the SymPortal. The post-MED output included 75 different ITS2 sequences associated with L. cf. striatus. Of these, 12 belonged to the genus Symbiodinium, 60 to the genus Cladocopium, and three to the genus Durusdinium. Durusdinium was only recorded during Spring 2022 in the North and South Red Sea, while Symbiodinium and Cladocopium were found all along the Saudi Arabian Red Sea coast. Overall, in the 56 coral samples examined, the genus Symbiodinium represented 2.6% of the Symbiodiniaceae community, the genus Cladocopium 97.1%, and the genus Durusdinium 0.02% ( Appendices S5 , S6 ).

A total of 17 ITS2 profiles were recovered ( Appendix S4 ). The ITS2 type profile composition per specimen was visualized using stacked bar charts to compare the relative abundance at each locality ( Figure 3 ). Overall, the type profiles were not equally distributed among the samples, with 50/56 specimens interacting with two newly discovered ITS2 profiles (C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C1aq and C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C89b), and 53/56 specimens associating with the C1 ITS2 majority sequences ( Appendix S5 ). The A1 radiation was found in 14/56 specimens. The remaining profiles, found associated with single specimens, belonged to A11, C3/C3u, C39/C1, C116/C116f, C39, C3/C1, C3, and D4 radiations.

Overall, during Fall 2020, three ITS2 profiles were recorded, specifically C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C1aq (9 colonies), C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C89b (6 colonies), and A1/A11g (1 colony) ( Figure 3 ). During Spring 2022, the ITS2 profiles C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C1aq and C1-C1cu-C1me-C42dr-C1mf-C1db-C1mg-C89b were recovered in association with L. cf. striatus at all five examined regions, indicating high levels of specificity ( Figure 3 ). The second most abundant type profile, retrieved from specimens collected in Spring 2022, A1-A1du-A1bw-A1bf, was associated with five colonies at four sampling localities (namely, GoA, NRS, CRS, and SRS) ( Figure 3 ). The ITS2 profiles C116/C116f, C3/C3u/C1-C115, and C3 were found only in the GoA during the spring sampling collection ( Figure 3 ). The type profiles A1-A1mp and A1-A1du-A1bw-A1bf-A1bx were recovered in the NEOM area (GoA and NRS) and only in the GoA, respectively, in Spring 2022 ( Figure 3 ). Finally, the A11 type profile was associated with one Yanbu (CRS) sample in May 2022 ( Figure 3 ). Considering the genus Durusdinium, during spring 2022, we recorded only two type profiles (D4-D4i-4k-D4ak-D6v and D4-D4i) associated with L. cf. striatus specimens ( Figure 3 ). The first was associated with two samples in the NRS and SRS, while the latter was associated with one colony in the NRS ( Figure 3 ).

In the different geographical regions that we identified, we also recovered colony-specific profiles, suggesting latitudinal and seasonal signatures ( Figure 3 ). In the GoA, we recovered three specific ITS2 profiles belonging to the A1, C116, and C3 radiations ( Figure 3 ). In the NRS, we found four ITS2 profiles, one found during Fall 2020 (A1/A1g) ( Figure 3 ), and the remaining recovered only during Spring 2022, belonging respectively to A1, D4, C1, and C39 radiations ( Figure 3 ). In the CRS (Yanbu and Thuwal), only one host-specific ITS2 profile was recorded, namely A11 ( Figure 3 ). Finally, in the SRS, we found two ITS2 profiles belonging to the C3 and C39 radiations ( Figure 3 ).

Overall, two Symbiodiniaceae type profiles associated with L. cf. striatus were the most abundant and found at each analyzed site ( Figures 3 , 4 ). Of the remaining 15 type profiles, four were shared between pairs of regions (in particular, one between CRS and SRS; one between GoA and NRS; one between NRS and SRS; finally, one found in all four regions mentioned before) ( Figures 3 , 4 ), while 11 were only found associated at a specific region ( Figures 3 , 4 ).

Zooxanthellate corals’ growth and survival are influenced by light availability, with consequent limitations on their depth distribution (Done, 2011; Muir et al., 2015). Understanding how zooxanthellate corals can survive in extreme environments, such as mesophotic waters, is still highly debated and is receiving increasing attention as technology improvements allow the exploration of previously inaccessible depths (Armstrong et al., 2019; Turner et al., 2019). Structural adaptation in macro and micro-morphological characters are commonly encountered in mesophotic corals in response to low-light availability, e.g., less self-shading (Ow and Todd, 2010), lower calical relief, or lower corallite density (Wijsman-Best, 1974; Willis, 1985; Carricart-Ganivet and Beltrán-Torres, 1993; Muko et al., 2000; Klaus et al., 2007). Moreover, the coral holobiont, particularly the associated Symbiodiniaceae community, can influence the host’s adaptability to such extreme environments. In the Red Sea, L. cf. striatus occurs in the lower mesophotic zone (70-145 m) (Fricke and Schuhmacher, 1983; Fricke and Knauer, 1986; Kaiser et al., 1993; Tamir et al., 2019; Ferrier-Pagès et al., 2022). Despite being considered a model species for research on coral physiology and ecology in mesophotic waters, its symbionts remain largely uncharacterized. Here, we provide the first assessment of the L. cf. striatus Symbiodiniaceae community. In agreement with previous works on different scleractinian taxa (Ezzat et al., 2017; Osman et al., 2020; Terraneo et al., 2023), we found in L. cf. striatus a Cladocopium-dominated community of zooxanthellae also consistent with Ziegler et al. (2017) who hypothesized a strong selection for this Symbiodiniaceae genus throughout the entire Arabian region (Osman et al., 2020).

Ziegler et al. (2015) collected different coral species between 1 and 60 meters along the Red Sea to understand the mechanisms behind the coral holobiont photoacclimatization at mesophotic depths. Their data suggested that the Symbiodiniaceae ITS2 type profiles play a role in the vertical distribution of corals. In particular, analysis of the zooxanthellae associated with coral samples revealed five distinct ITS2 sequences of known type profiles (C1, C3, C15, C39, D1a) (Ziegler et al., 2015). Similarly, our study reported an association between L. cf. striatus and the known type profiles A1, A11, C1, C116, C3, C39, and D4 radiations, with the Symbiodiniaceae community dominated by the C1 genotypes. Symbiodiniaceae type profile C1 was found to be associated with coral hosts ascribed to the genera Leptoseris and Seriatopora Lamarck, 1816, over different mesophotic depth ranges in other locations outside the Red Sea, e.g., Hawaii and Western Australia (Chan et al., 2009; Cooper et al., 2011b). The prevalence of Cladocopium genotypes in mesophotic species could be related to the ability of the genus to fix carbon under low-light environments (Ezzat et al., 2017). At the same time, different studies reported that the genus Durusdinium could be found in higher proportions during thermal stress than in normal environmental conditions (Stat et al., 2013; Cunning et al., 2015; Poquita-Du et al., 2020; Dilworth et al., 2021). We found Durusdinium genotypes associated with L. cf. striatus in the Northern and the Southern Red Sea in June and April 2022, during a season characterized by cooler sea temperatures. Different Symbiodiniaceae genotypes associated with stony corals are reported to cope differently with various environmental stressors (Silverstein et al., 2011, 2012; Rouzé et al., 2019). Certain corals can change symbiont proportions and Symbiodiniaceae species in response to different disturbances (Rowan, 2004; Berkelmans and Van Oppen, 2006). This ability has been shown to directly correlate with Symbiodiniaceae diversity within the host (Cunning et al., 2015; Rouzé et al., 2019). It is still unclear whether these adaptations similarly apply to L. cf. striatus. However, we hypothesized that the appearance of presumed thermal taxa in the genus Durusdinium in association with L. cf. striatus could be linked to environmental perturbation, such as exposure to cooler sea temperatures, a pattern already found in some colonies of Leptoria phrygia (Ellis and Solander, 1786) in Taiwan (Huang et al., 2020).

The Red Sea is characterized by environmental gradients (Sofianos and Johns, 2003; Raitsos et al., 2013; Rowlands et al., 2014, 2016; Chaidez et al., 2017; Manasrah et al., 2019; Berumen et al., 2019a), which can shape the distribution of the Symbiodiniaceae along the Saudi Arabian coast in shallow (< 30 m) zooxanthellate corals (Berumen et al., 2019a). In fact, previous studies focusing on shallow-water corals reported a zooxanthellae community composition break along the Red Sea latitudinal gradient. For example, Terraneo et al. (2019a) characterized the Symbiodiniaceae community associated with different Porites species along the entire Saudi Arabian Red Sea. Their data showed a shift from the genus Cladocopium to the genus Durusdinium, going from the North to the South of the Red Sea, with the central Red Sea Porites harboring both genera. Similarly, Sawall et al. (2014) and Arrigoni et al. (2016) reported a shift from Cladocopium-dominated to Symbiodinium-dominated communities along the same Red Sea latitudinal gradient in the hard coral genus Pocillopora Lamarck, 1816 and Stylophora, respectively. Together, these results point to a relationship between the Red Sea environmental gradients and the Symbiodiniaceae biogeographical patterns in shallow water taxa.

In this study, when considering zooxanthellae patterns at the genus level, we did not find latitudinal differences in association with the strictly mesophotic L. cf. striatus, as the coral species presented Cladocopium-dominated communities all along the Saudi Arabian Red Sea coast, despite the presence of modest temperature and salinity gradients in the mesophotic layer, compared to stronger gradients in surface waters. Similar results were found by Terraneo et al. (2023) for the corals Coscinaraea monile (Forskål, 1775), Blastomussa merleti (Wells, 1961), Psammocora profundacella Gardiner, 1898, and Craterastrea levis Head, 1983, The authors, however, only compared colonies from the Gulf of Aqaba and the Northern Red Sea, thus a smaller latitudinal gradient than the one examined in this study, and focused on zooxanthellate species commonly mainly occurring at shallow depths with the notable exception of C. levis. This species, known to be exclusively living in low-light conditions (Benzoni et al., 2012), showed a lower Symbiodiniaceae diversity at mesophotic depths than the others and was associated with only one profile belonging to the C1 radiation in both the Gulf of Aqaba and the Northern Red Sea (Terraneo et al., 2023). Other studies reported Cladocopium-dominated communities in the mesophotic waters of the Gulf of Aqaba (Ezzat et al., 2017; Osman et al., 2020). In fact, there is no evidence of environmental gradients, which we confirm here for temperature and salinity along latitude, shaping symbionts distribution in the Red Sea mesophotic waters.

We compared the variability of L. cf. striatus zooxanthellae communities ( Figure 4A ) to the one previously published from the shallow water P lutea and P. columnaris sampled along the Red Sea latitudinal gradient (Terraneo et al., 2019a) ( Figures 4B, C ). In particular, the previously examined cases of the two shallow water Porites species showed a total of 16 different Symbiodiniaceae type profiles ( Figure 4 ) and presented a latitudinal pattern, with only one type profile, C15, shared between the GoA and AlW in P. columnaris ( Figure 4 ) (Terraneo et al., 2019a). Hence, the ITS2 type profiles associated with the Porites species in the euphotic zone were mostly host-generalist and associated with specific sampling localities. Moreover, both species showed a shift in the Symbiodiniaceae community from Cladocopium in the North to Durusdinium in the South. As a result, no type profile was shared among the different sampling areas in either species ( Figures 4B, C ). Conversely, the genotypes recovered in L. cf. striatus were mainly represented by the genus Cladocopium at all sites. Furthermore, our results show only five genotypes belonging to Symbiodinium and two belonging to Durusdinium, both occurring together with Cladocopium. As a result, L. cf. striatus shared the same two type profiles found in all regions along the Red Sea latitudinal gradient ( Figure 4A ). This suggests that for this mesophotic specialist coral, the Symbiodiniaceae community variability is not shaped by the same environmental gradients as in the shallow reef dwelling species.

The five Saudi Arabian Red Sea regions we investigated were recently identified as four distinct bioregions based on phytoplankton phenology (Kheireddine et al., 2021). The Gulf of Aqaba is commonly considered a coral refuge and resilience area, with the coral communities able to persist through different disturbances and thermal stressors (Fine et al., 2013, 2019). This bioregion (3), together with the Northern and Central Red Sea (4) and the Southern Red Sea (2) bioregions, present a phytoplankton bloom during winter (October to March) and a recurrent summer phytoplankton boom peak (Raitsos et al., 2013; Kheireddine et al., 2021). Conversely, the Southern Red Sea bioregion (1, including the Farasan Banks and the Dahlak Archipelago) is characterized by a summer phytoplankton bloom starting at the end of May (Kheireddine et al., 2021). We sampled the Gulf of Aqaba and the Northern Red Sea in two different years, both supposedly concurrently with the reported phytoplankton bloom (October/November 2020 and June 2022). Our results did not reveal the presence of presumed thermal tolerant genotypes such as Durusdinium radiations during fall, when the sea water was warmer and less saline than in June 2022. Conversely, Symbiodinium and Durusdinium genotypes appeared in June 2022, when the seawater was cooler and more saline sea, indicating that some ecological variables related to temporal environmental changes could shape coral-zooxanthellae association at mesophotic depths. During Spring 2022, we reported a North-South gradient in water temperature and salinity in the mesophotic layer, less pronounced than those reported for surface waters (see Raitsos et al., 2013; Chaidez et al., 2017; Berumen et al., 2019a), in agreement with previous studies (Sofianos and Johns, 2003; Manasrah et al., 2019). In particular, the temperature increases from North to South while the salinity decreases, thus mirroring the shallow water gradient. However, all the remaining regions analyzed during Spring 2022 (namely Al Wajh, Central Red Sea, and Southern Red Sea) presented Cladocopium-dominated communities with more presumed thermally tolerant Symbiodiniaceae genotypes in one sample in the southern Red Sea, and not a shift on the Symbiodiniaceae community as shown in other Red Sea shallow water studies (Arrigoni et al., 2016; Terraneo et al., 2019a), indicating that the algal community of this mesophotic coral is consistent across the Red Sea latitudes. Moreover, different studies reported site and seasonal-specific acclimation signatures in corals (e.g., Brown et al., 1999; Fitt et al., 2001; Edmunds, 2009; Anthony et al., 2022; Sawall et al., 2022). However, processes other than seasonality (e.g., environmental stressors, thermal adaptation, host-symbiont specificity, and anthropogenic disturbances) (see, for example, Silverstein et al., 2012; Cunning et al., 2015; Terraneo et al., 2019a; Howe-Kerr et al., 2020; Claar et al., 2022); could drive the algal community shift in symbiotic corals. It is still unclear which processes made the algal community of L. cf. striatus change across 2 sampling time points. Hence, future studies encompassing longer sampling duration and sampling the same individual over time would be required to address whether seasonality or other processes play a role in the Symbiodiniaceae community composition of L. cf. striatus.