ORIGINAL RESEARCH published: 24 July 2019 doi: 10.3389/fmicb.2019.01702 Experimental Inoculation of Coral Recruits With Marine Bacteria Indicates Scope for Microbiome Manipulation in Acropora tenuis and Platygyra daedalea Katarina Damjanovic 1,2* , Madeleine J. H. van Oppen 1,2 , Patricia Menéndez 2,3 and Linda L. Blackall 1 1 School of BioSciences, The University of Melbourne, Parkville, VIC, Australia, 2 Australian Institute of Marine Science, Townsville, QLD, Australia, 3 School of Mathematics and Physics, University of Queensland, Saint Lucia, QLD, Australia Coral-associated microorganisms are essential for maintaining the health of the coral Edited by: Iliana B. Baums, holobiont by participating in nutrient cycling and protecting the coral host from Pennsylvania State University, pathogens. Under stressful conditions, disruption of the coral prokaryotic microbiome United States is linked to increased susceptibility to diseases and mortality. Inoculation of corals Reviewed by: with beneficial microbes could confer enhanced stress tolerance to the host and may Jeremy Brian Weisz, Linfield College, United States be a powerful tool to help corals thrive under challenging environmental conditions. Blake Ushijima, Here, we explored the feasibility of coral early life stage microbiome manipulation by Smithsonian Marine Station (SI), United States repeatedly inoculating coral recruits with a bacterial cocktail generated in the laboratory. *Correspondence: Co-culturing the two species Acropora tenuis and Platygyra daedalea allowed us to Katarina Damjanovic simultaneously investigate the effect of host factors on the coral microbiome. Inoculation kdamjanovic@student.unimelb.edu.au cocktails were regularly prepared from freshly grown pure bacterial cultures, which were Specialty section: hence assumed viable, and characterized via the optical density measurement of each This article was submitted to individual strain put in suspension. Coral early recruits were inoculated seven times Microbial Symbioses, over 3 weeks and sampled once 36 h following the last inoculation event. At this time a section of the journal Frontiers in Microbiology point, the cumulative inoculations with the bacterial cocktails had a strong effect on the Received: 04 February 2019 bacterial community composition in recruits of both coral species. While the location Accepted: 10 July 2019 of bacterial cells within the coral hosts was not assessed, metabarcoding using the Published: 24 July 2019 16S rRNA gene revealed that two and six of the seven bacterial strains administered Citation: Damjanovic K, van Oppen MJH, through the cocktails were significantly enriched in inoculated recruits of A. tenuis and Menéndez P and Blackall LL (2019) P. daedalea, respectively, compared to control recruits. Despite being reared in the Experimental Inoculation of Coral same environment, A. tenuis and P. daedalea established significantly different bacterial Recruits With Marine Bacteria Indicates Scope for Microbiome communities, both in terms of taxonomic composition and diversity measurements. Manipulation in Acropora tenuis These findings indicate that coral host factors as well as the environmental bacterial and Platygyra daedalea. Front. Microbiol. 10:1702. pool play a role in shaping coral-associated bacterial community composition. Host doi: 10.3389/fmicb.2019.01702 factors may include microbe transmission mode (horizontal versus maternal) and host Frontiers in Microbiology | www.frontiersin.org 1 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation specificity. While the long-term stability of taxa included in the bacterial inocula as members of the host-associated microbiome remains to be evaluated, our results provide support for the feasibility of coral microbiome manipulation, at least in a laboratory setting. Keywords: coral microbiome, probiotics, bacterial manipulation, assisted evolution, Acropora tenuis, Platygyra daedalea, host-symbiont specificity INTRODUCTION (Ward, 2018). Climate models predict that most coral reefs in the world will experience similar extreme bleaching annually by the Scleractinian corals are responsible for building the three- end of the century (van Hooidonk et al., 2016). Given their great dimensional structure of coral reefs through the deposition of ecological and cultural importance (Harrison and Booth, 2007; their calcium carbonate skeletons, thereby providing habitat for Blackall et al., 2015) and high economical value (Burke et al., over a quarter of all marine species (Spadling et al., 2001). Most 2011), the loss of reef-forming scleractinian corals would have scleractinian corals are colonial organisms, comprised of large severe consequences for coral reef ecosystems as well as for the numbers of interconnected polyps. Having persisted for over coastal human populations depending on coral reefs. 200 million years (Veron, 1995), scleractinian corals owe this While it is urgent to address the root causes of climate success to their symbiosis with photosynthetic microalgae in change, it is also essential to explore the possibility of augmenting the family Symbiodiniaceae, which supply the host with most coral tolerance and resistance to stress. Since environmental of their energy (Muscatine and Porter, 1977). In addition to degradation may be occurring too fast for corals to adapt through these algal endosymbionts, corals form associations with a huge natural selection (Hoegh-Guldberg, 2004; Hoegh-Guldberg et al., diversity of other microorganisms including bacteria, archaea, 2007), the concept of assisted evolution (AE) (Jones and fungi and viruses (reviewed in Blackall et al., 2015). Altogether, Monaco, 2009) has been proposed as a strategy for coral reef the coral animal and associated microorganisms constitute a conservation (van Oppen et al., 2015). AE aims to accelerate functional entity (analogous to a small-scale ecosystem; Pita et al., the rate of naturally occurring evolutionary processes, in order 2018) called the holobiont (Rohwer et al., 2002). Prokaryotes to develop corals better able to cope with current climate occupy various niches within the host, including intra- and change trajectories (van Oppen et al., 2015). AE encompasses the extracellular spaces in tissues (Work and Aeby, 2014), the manipulation of coral-associated symbionts, including members surface mucus layer, the gastric cavity and the skeleton (Sweet of the Symbiodiniaceae (Chakravarti et al., 2017) and prokaryotes et al., 2010; Ainsworth et al., 2015). These communities exert (Damjanovic et al., 2017; Peixoto et al., 2017; Webster and numerous beneficial functions that are essential for the well- Reusch, 2017; Epstein et al., 2019b). being of the coral animal, such as carbon, nitrogen, sulfur, Microbial inoculations have already been used in plants, and phosphorus cycling (reviewed in Bourne et al., 2016). humans, and a diversity of other host organisms. For example, Bacteria also protect corals from pathogens by occupying entry plant growth promoting rhizobacteria (PGPR) are natural niches and secreting antimicrobial peptides (Bourne et al., 2016). symbionts that colonize the rhizosphere, stimulate plant growth Moreover, evidence from reciprocal transplantation experiments and development, and protect against biotic and abiotic stresses followed by short-term heat stress suggests that coral-associated (reviewed in Singh and Singh, 2013; Gouda et al., 2018). In bacterial communities are linked to intraspecific variation in humans, probiotics are used both as supplementation to improve coral heat tolerance (Ziegler et al., 2017). Maintaining an physiological functions in healthy individuals (Khalesi et al., appropriate microbial community composition is undoubtedly 2018) and administered as treatments to patients suffering key to preserve coral health, as detrimental alterations are usually from various gastrointestinal disorders (Ringel et al., 2012). observed to take place in diseased states and in response to Rumen transfaunation (i.e., transfer of microorganisms from adverse environmental conditions (Jones et al., 2004; Bourne a healthy to a sick ruminant) is commonly conducted to et al., 2008; Littman et al., 2011; Tout et al., 2015). enhance productivity and treat gastrointestinal dysbiosis in Corals are suffering massive declines due to the global impacts livestock (DePeters and George, 2014). Despite challenges of climate change and other anthropogenic disturbances (Hoegh- inherent to employing microbial inocula in aquatic systems, Guldberg, 2011; De’ath et al., 2012). Rising seawater temperature such as establishing a suitable administration method, the is a major cause of coral bleaching (the breakdown of the critical potential of probiotics in aquaculture has been explored for symbiosis between the coral host and its algal endosymbionts) over two decades (Gatesoupe, 1999; Verschuere et al., 2000). and often leads to extensive coral mortality. The most severe Probiotics are now widely used in this industry to promote bleaching on record for the Great Barrier Reef (GBR) occurred animal growth, control disease, ameliorate water quality or during the El Niño event of 2016, when ∼30% of coral present augment stress tolerance (reviewed in Martinez Cruz et al., 2012; was lost (Normile, 2016; Hughes et al., 2017, 2018). Another mass Hai, 2015). Microbiome manipulation in wildlife conservation bleaching event occurred the following year during a summer is currently limited but recent applications offer hope for the heat wave not related to El Niño (Hughes and Kerry, 2017), protection of endangered species. For example, the amphibian resulting in the loss of an additional ∼20% of coral cover cutaneous bacterium Janthinobacterium lividum secretes an Frontiers in Microbiology | www.frontiersin.org 2 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation antifungal metabolite effective against chytrid fungi which cause MATERIALS AND METHODS chytridiomycosis, a disease that has already decimated many frog populations worldwide (Rebollar et al., 2016). Bioaugmentation Coral Spawning and Rearing of antifungal bacteria through the inoculation of amphibian Colonies of the corals A. tenuis and P. daedalea were collected hosts holds promise to prevent the extinction of vulnerable off Falcon Island (S -18◦ 46 E 146◦ 32), Australia, and transported populations. Similar to plants and other animals, exposing corals to the Australian Institute of Marine Science (AIMS) on the 4th to certain microbial communities may trigger a beneficial shift of November 2017. The corals were deployed in 3000 L tanks in the symbiosis and render the holobiont more resilient to containing running 0.4 µm filtered seawater (FSW). On the external pressures. night of spawning (8th and 9th of November 2017 for A. tenuis Although the notion of exogenously adding microbes to and P. daedalea respectively), setting colonies were isolated in corals in a probiotic approach was suggested in 2009 (Teplitski 80 L plastic bins filled with 0.4 µm FSW. Once released, gamete and Ritchie, 2009), the field is still in its infancy but some early bundles were cautiously scooped from the water surface and promising results have been obtained. Fragments of the coral washed over 60 µm plankton mesh to separate eggs from sperm. Mussismilia hartii were inoculated with a bacterial consortium Gametes from six A. tenuis and 12 P. daedalea colonies were able to degrade water-soluble oil fractions, which reduced the mixed separately for each coral species and left to fertilize for negative impact of a simulated oil spill on the experimental 2 h. The embryos were washed again with FSW to remove sperm corals (dos Santos et al., 2015). In another laboratory experiment, and transferred to replicate cylindrical 60 L larval rearing tanks, bacterial strains were selected for putatively beneficial traits which contained flow-through 0.1 µm FSW at 28◦ C and a low including nutrient cycling, antioxidative capacities, and level of aeration. antagonistic activities against pathogens (Rosado et al., 2018). Nine days post spawning, approximately 9000 larvae from Inoculation of Pocillopora damicornis nubbins with the resulting each species were distributed across three 60 L tanks (i.e., consortium was able to partially mitigate coral bleaching and three settlement tanks for A. tenuis larvae and three different alleviate pathogenic infection (Rosado et al., 2018). Thus, settlement tanks for P. daedalea larvae) containing flow-through coral bacterial community composition seems to be flexible 0.4 µm FSW at 27.5◦ C. The tanks were set under a 12 h to some extent and adjustable to benefit the host. Finally, light/dark illumination cycle reaching a maximum light intensity a single exposure of coral larvae to the mucus-associated of 70 µmol·m−2 ·s−1 after 5 h ramping. Aragonite settlement microbes of four different coral species resulted in divergent plugs conditioned with a microbial biofilm and crustose coralline prokaryotic communities after 4 months of rearing in filter- algae (CCA) and subsequently autoclaved had been placed in sterilized seawater (Damjanovic et al., 2017). Even though PVC trays on the bottom of the tanks. Six days after A. tenuis the initial inoculum composition was not characterized, this and P. daedalea juveniles were settled on the plugs in their experiment showed that coral-associated microbiomes could respective tanks, the total number of recruits was counted under be influenced to develop in distinct directions following a dissecting microscope and plugs were randomized into 12 microbial dosing. Early coral life stages may be particularly PVC trays. From this point onwards, A. tenuis and P. daedalea suitable for targeted microbial inoculation, as the microbial recruits were therefore co-reared on the same trays. The latter composition and cell density in the surrounding environment were spread across 12 new experimental 60 L tanks, set under strongly influences the microbiome acquired by juvenile corals the same light and temperature conditions, with a flow rate of (Apprill et al., 2009; Damjanovic et al., 2017). Further, the 1 L/min (i.e., each hour the water volume completely changed). bacterial communities associated with early coral life stages Water did not recirculate between these 12 tanks, as each one of tend to be more dynamic as compared to the associations them had different inflow and outlet pipes. In total, an average harbored by adults (Littman et al., 2009a; Lema et al., 2014; of 80 A. tenuis recruits and 47 P. daedalea recruits per replicate Epstein et al., 2019a). tank were available at the start of the inoculation experiment The objective of the present study was firstly to gain (Figure 1A). All recruits were tracked and counted again at the more insights into how associations between young corals and end of the experiment to assess survival rate. bacteria can be manipulated by targeted bacterial inoculations, and secondly, to examine whether different hosts are a determinant of the microbial community development. Coral Preparation of Bacterial Pure Cultures recruits of two taxonomically divergent species, Acropora The inoculation cocktail was made from pure cultures of tenuis and Platygyra daedalea, were co-cultured in the same bacteria belonging to the genera Acinetobacter, Bacterioplanes, aquaria and simultaneously exposed to a cocktail composed Marinobacter, Paracoccus, Pseudoalteromonas, Pseudovibrio, and of pure bacterial cultures. Metabarcoding of the 16S rRNA Vibrio, which had been previously isolated from marine gene was used to assess the bacterial communities of the invertebrates and stored as frozen stocks (Table 1). As the coral recruits following repeated inoculations. While undefined goal of the present study consisted in assessing the potential host factors were shown to play a role in the composition of manipulating the microbiome of juvenile corals, the bacteria of coral recruit-associated bacterial communities, inoculated used were not selected for putative beneficial properties. Rather, corals of both species were significantly enriched for some we aimed to test to what extent the bacteria harbored by of the taxa that comprised the inoculum compared to non- young recruits could be influenced by targeted inoculation as inoculated control corals. further support for the proof-of-concept of coral microbiome Frontiers in Microbiology | www.frontiersin.org 3 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 1 | Overview of the experimental design. (A) Larvae from A. tenuis and P. daedalea were independently settled on plugs pre-conditioned with a microbial biofilm and CCA and subsequently sterilized. Six days post-settlement, recruits were counted and randomized across twelve 60 L tanks containing flow-through FSW. (B) On each inoculation day, bacterial pure cultures grown on MA from the frozen stocks were suspended in FSW and adjusted to similar densities in order to prepare a fresh bacterial cocktail. Recruits from each 60 L tank were placed in containers filled with 1 L of static FSW. The bacterial cocktail was added to six containers using a serological pipette to inject 10 mL of the suspension into the water directly above the recruits. The bacterial density in the inoculated containers was estimated at 105 –106 cells/mL. Recruits from the control treatments were handled in the same manner, except that FSW was added to the containers instead of the bacterial cocktail. After 6 h in static water, recruits were transferred back into the 60 L flow-through tanks. Two days following the last inoculation, all recruits were counted, sampled from the 60 L tanks together with 1 L of surrounding water and snap-frozen for DNA extraction and 16S rRNA gene metabarcoding. Symbols for diagrams were modified from the Courtesy of Integration and Application Network (2018). Frontiers in Microbiology | www.frontiersin.org 4 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation TABLE 1 | Characteristics of the bacterial strains used to generate the inoculation consortium. Bacterial identity Closest relative % identity from BLAST of X nt of GenBank Isolation source the 16S rRNA gene accession Acinetobacter sp. A. baumannii 99.9% from 1466 nt MH744724 Healthy coral, GBR, Australia Oceanospirillales isolate Bacterioplanes sanyensis 99.7% from 1410 nt MH744725 Healthy coral, GBR, Australia Marinobacter sp. M. sediminum 98.6% from 1512 nt MK088251 Laboratory-bred anemone Exaiptasia pallida, Australia Paracoccus marcusii (Identified based on whole – MH744726 Healthy coral, GBR, Australia genome data) Pseudoalteromonas sp. P. shioyasakiensis 99.8% from 1403 nt MK088250 Laboratory-bred anemone Exaiptasia pallida, Australia Pseudovibrio denitrificans (Identified based on whole – KX198136 Healthy coral, GBR, Australia genome data) Vibrio harveyi This is the type strain: DSM – X56578∗ Dead luminescing amphipod No 1963 Talorchestria sp., United States The references for these strains are specified in the main text. nt, nucleotides. ∗ GenBank accession number for the whole genome shotgun sequence: BAOD01000001. manipulation. The use of isolates from various marine organisms transferred into Difco Marine Broth (BDTM , United States) and provides information about a range of bacterial sources that could grown overnight in a shaking incubator at 28◦ C. Liquid cultures be taken up by given coral species and hence how flexible these were re-streaked onto minimal marine agar; the procedure was symbiotic partnerships are. repeated until pure cultures were obtained, which were stored at Pure cultures of Marinobacter sp. and Pseudoalteromonas sp. −80◦ C in 20% glycerol. were isolated from the anemone Exaiptasia pallida maintained Vibrio harveyi was purchased from the Leibniz Institute at the University of Melbourne (Table 1). Single anemones were DSMZ – German Collection of Microorganisms and Cell transferred from the maintenance culture dish into 0.2 µm- Cultures (Table 1). filtered Red Sea Salt water (Red SeaTM ) (FRSSW) and left static for 30 min in the dark to rinse external seawater. Each anemone was then gently transferred to a sterile glass homogenizer and Inoculation of Coral Recruits With homogenized with 1 mL FRSSW. Serial dilutions from 10−1 to Bacteria and Sampling 10−4 of the anemone homogenates in FRSSW were prepared Bacterial inoculation of the coral recruits was repeated seven and 50 µL were spread onto replicate plates of marine agar times at regular intervals between November 27th and (MA - DifcoTM Marine Agar 2216). MA plates were incubated in December 18th 2017 (Figure 1B). A fresh bacterial cocktail the dark at 26◦ C for 1 week when individual, clearly separated was prepared from the pure cultures before each inoculation colonies were 16-streak inoculated to new MA plates and event. Approximately 72 h before each inoculation, the seven incubated in the dark at 26◦ C for 2–3 days. This process was bacterial cultures were revived from frozen stocks, streaked onto repeated until pure cultures were obtained, which were preserved marine agar (DifcoTM Marine Agar 2216, BD) and incubated at −80◦ C on cryobeads (Protect Microorganism Preservation at 28◦ C. On the morning of coral recruit inoculation, sterile System, Thermofisher, Cat# TS/80-MX). inoculation loops were used to collect bacterial colonies from the Pure cultures of Acinetobacter sp., an Oceanospirillales isolate agar plates and cells of each strain were suspended separately in (used in van de Water et al., 2018), Paracoccus marcusii and 9 mL 0.22 µm FSW. Optical density (OD660 ) of the suspensions Pseudovibrio denitrificans (Raina et al., 2016) were obtained from was measured with a NanoDrop1000 spectrophotometer and healthy coral colonies of A. millepora and P. damicornis (Table 1). cellular density was estimated using the calculation reported in The corals were collected from Pelorus Island and Davies Agilent Genomics (2018). Therefore, a cell density-OD660 value Reef respectively (Great Barrier Reef: 18◦ 330 S; 146◦ 290 E and standard was used to infer cellular density. The seven bacterial 18◦ 510 S; 147◦ 410 E), and maintained in aquaria at AIMS. Coral suspensions were adjusted to 107 –108 cells/mL (as estimated fragments (approximately 30 mm in length) were collected from from the measured OD660 : 0.082–0.21), either by dilution with each colony and washed in sterile artificial seawater (ASW) to additional 0.22 µm FSW or by introducing more cells in the remove loosely attached microbes. Coral tissue/microbe slurries suspension using a sterile inoculation loop. Nine mL of each were produced by airbrushing (550 kPa) each coral fragment suspension were then combined in a 100 mL sterile glass bottle. into 5 mL of ASW and a dilution series was spread plate The optical density of the final bacterial cocktail was measured inoculated immediately onto modified minimal marine agar (1% again to verify the collective cellular density. While optical bacteriological agar, 25 g of NaCl, 0.7 g of KCl, 0.05 g of KH2 PO4, density measurements do not distinguish between viable and 1 g of NH4 NO3 , 1 g of MgSO4 ·7H2 O, 0.2 g of MgCl2 ·H2 O, dead bacterial cells, we assumed that the bacteria used to prepare 0.02 g of CaCl2 ·2H2 O, 0.005 g of FeEDTA, 1 g of Tris, 5 g of the inoculation cocktails were alive as they were collected from sodium succinate, 1.35 g of glucose in 1 L of distilled water). freshly grown cultures. All OD660 measurements are summarized After 2 days of incubation at 28◦ C, single bacterial colonies were in Supplementary File 2. Frontiers in Microbiology | www.frontiersin.org 5 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation At 11:00 am on the inoculation days, all trays with recruits on sample were pooled for DNA extraction (i.e., both treatments CCA plugs from each of the twelve 60 L tanks were transferred provided six replicates of ∼30 polyps each), while 60 µL were into twelve 3 L plastic containers filled with 1 L 0.4 µm FSW used for the C. goreaui culture and each bacterial inoculum. and maintained at 27.5◦ C. Ten mL of the bacterial cocktail were Samples were placed in 1.5 mL sterile microcentrifuge tubes inoculated to six of the 3 L containers by pipetting the suspension containing 250 µL of extraction buffer (100 mM Tris pH 9.0, directly into the 1 L of water above the recruits. The final bacterial 100 mM EDTA, 1% SDS, 100 mM NaCl). Ten µL of lysozyme density in the inoculated containers was 105 –106 cells/mL at 10 mg/mL were added to all tubes and tubes were incubated (corresponding to a 1:100 dilution of 107 –108 cells/mL present in at 37◦ C for 30 min. About 30 mg of sterile acid-washed glass the cocktail). To each of the remaining six 3 L containers, 10 mL beads (size 710 – 1180 µm, Sigma-Aldrich G1152) and 10 µL of of FSW were pipetted into the water and thus served as controls. Proteinase K at 20 mg/mL were added to the tubes. The samples All recruits were left in the static 3 L containers for 6 h before were bead-beaten at 4 m/s for 20 s and incubated at 55◦ C for 2 h, being returned to the 60 L flow-through tanks. The trays were followed by 65◦ C for 15 min. After this step, 62.5 µL of KOAc handled manually wearing different pairs of ethanol-sterilized at 5 M were pipetted into the tubes and incubated on ice for gloves between each replicate. On the first day of bacterial 30 min. After spinning the tubes at 25,000 g for 15 min at room inoculation, recruits from all tanks were also exposed to a temperature, the supernatant was transferred into new 1.5 mL pure culture of Symbiodiniaceae Cladocopium goreaui (formerly sterile microcentrifuge tubes and 0.8 vol. isopropanol was added known as type C1; LaJeunesse et al., 2018), strain SCF055-01.10 to precipitate DNA. The solutions were left at room temperature (GenBank accession number MK027323) at a density of ∼8,500 for 15 min and centrifuged again at 25,000 g for 15 min. cells/mL. The C. goreaui pure culture was obtained from the After removing the supernatant, the precipitate was washed with AIMS Symbiont Culture Facility. The algal cells had been grown 100 µL of 70% ethanol, centrifuged at 25,000 g for 3 min, in Corning cell culture flasks at 27◦ C using a photoperiod of air-dried and resuspended overnight in 25 µL MilliQ water. 14 h: 10 h light to dark illumination cycle and 60 µmol·m−2 ·s−1 . The same DNA suspensions were used to amplify the 16S Cells were maintained in 0.2 µm FSW and Daigo’s IMK sterile rRNA gene and ITS2 marker. Variable regions V5 to V6 of the culture medium for marine microalgae (Nihon Pharmaceutical 16S rRNA gene were amplified using the forward primer 784F [50 - Co., Ltd.). This Symbiodiniaceae species was chosen because it TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAGGATT commonly associates with A. tenuis on the GBR (LaJeunesse et al., AGATACCCTGGTA-30 ] and reverse primer 1061R [50 - 2004; van Oppen et al., 2005). While P. daedalea more often GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGCRRCA hosts a different species of Symbiodiniaceae in the wild (formerly CGAGCTGACGAC-30 ] (Andersson et al., 2008; Röthig known as type C3) (Fisher et al., 2011), it also naturally associates et al., 2016). The Symbiodiniaceae ITS2 region was with C. goreaui (Abrego et al., 2009). amplified with the specific primer pair ITS2F [50 - Thirty-six hours after the recruits had been placed back in the TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTGAAT flow-through 60 L tanks following the last bacterial inoculation TGCAGAACTCCGTG-30 ] and ITS2R [50 -GTCTCGTGGGC (at Day 40), recruits were counted under the dissecting TCGGAGATGTGTATAAGAGACAGCCTCCGCTTACTTATAT microscope to assess survival rate and were also sampled for GCTT-30 ] (Boulotte et al., 2016). The underlined segments bacterial and Symbiodiniaceae community composition analysis. represent Illumina adapter overhangs (Illumina, San Diego, CA, The microbial communities detected on the coral recruits United States). The PCRs were conducted in 10 µL triplicates therefore accounted for the cumulative effect of all seven using the AmpliTaq Gold 360 Master Mix and 0.4 µM of each inoculations. Degraded cells or residual DNA from the inocula primer. The amplification cycles were: 95◦ C for 10 min; 30 surrounding the recruits were minimized in the samples, because cycles each at 95◦ C for 30 s, 57◦ C for 1 min, 72◦ C for 30 s; the water flow of 1 L/min replaced the entire volume of the 60 L a final extension at 72◦ C for 7 min. The PCR triplicates for tanks 36 times. A. tenuis and P. daedalea recruits from each each template were pooled and sent to Ramaciotti Centre tray were removed from the plugs with a sterile scalpel blade, for Genomics (UNSW, Sydney) for library preparation and rinsed with 0.22 µm FSW and placed into cryovials for snap- sequencing on the Illumina MiSeq system with 2 × 300 bp freezing with liquid nitrogen. Bacterial and Symbiodiniaceae paired-end reads. communities were also assessed in 1 L of water collected from In addition to processing the samples, six blank DNA each recruit maintenance tank and filtered through a 0.22 µm extractions and three no-template PCRs were performed and SterivexTM filter using a peristaltic pump. Samples of the seven sequenced to check for laboratory contamination. bacterial inocula and the C. goreaui pure culture were collected as positive controls. The 16S rRNA genes and the nuclear DNA Bioinformatics ribosomal internal transcribed spacer 2 (ITS2) were targeted in Both the 16S rRNA gene partial sequences and ITS2 region the metabarcoding analyses. sequences were processed using the QIIME 2 pipeline version 2017.10 (Caporaso et al., 2010; QIIME 2 Development Team, DNA Extraction and Amplification 2017a). Plugin demux (QIIME 2 Development Team, 2017b) DNA was extracted from the samples following modifications was used to visualize interactive quality plots and check read of the protocol reported in Wilson et al. (2002). To provide quality. Plugin DADA2 (Callahan et al., 2016) was subsequently enough prokaryote biomass for downstream PCR amplification applied to remove primers, truncate poor-quality bases based on of the 16S rRNA gene, ∼30 polyps for each recruit replicate the interactive plots, dereplicate, identify chimeras and to merge Frontiers in Microbiology | www.frontiersin.org 6 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation paired-end reads. Commands included in the plugin feature-table All statistical analyses were conducted using R version 3.3.2 (McDonald et al., 2012) enabled generation of summary statistics (R Core Team, 2018) and packages phyloseq (McMurdie and of sequences associated with the samples. For 16S rRNA gene Holmes, 2013), vegan (Oksanen et al., 2016), DESeq2 (Love sequences, a Naïve Bayes Classifier was trained with the feature- et al., 2014), tidyverse (Wickham, 2017), randomForest classifier plugin (QIIME 2 Development Team, 2017d) using the (Liaw and Wiener, 2002), RVAideMemoire (Herve, 16S rRNA gene database at 99% similarity of the SILVA 128 2018) and multcomp (Hothorn et al., 2008) and ggplot2 QIIME release (Quast et al., 2013) and based on the 784F/1061R (Wickham, 2009). primer pair. A phylogenetic tree for further downstream analyses was created with the plugins alignment (Katoh and Standley, 2013) and phylogeny (Price et al., 2010). Finally, the taxa RESULTS plugin (QIIME 2 Development Team, 2017c) allowed to filter mitochondria and chloroplast sequences, as well as to visualize Survival of Coral Recruits taxonomic bar plots and generate tables with absolute read counts Following settlement, 962 A. tenuis and 566 P. daedalea recruits of all taxa for each sample. For ITS2 sequences, taxonomic were distributed across the 12 experimental tanks. Thus, 80 assignment was performed using the Arif et al. (2014) ITS2 A. tenuis and 47 P. daedalea recruits were randomized in each database with the method vsearch (Rognes et al., 2016), at 97% tank at the start of the experiment. The average survival of sequence similarity level. The biom tables containing taxonomic the 4-week old recruits was high (AcroCtrl : 99.2%, AcroInoc : counts, the metadata and phylogenetic trees were imported into 98.6%, PlatyCtrl : 95.5%, PlatyInoc : 95.7%, Figure 2A) with no R (R Core Team, 2018) for statistical analyses. significant difference in survival across species or treatment (for all pairwise contrasts: p > 0.9). However, survival appeared Statistical Analyses more variable among tanks for P. daedalea recruits than for Exploratory and statistical analyses were performed at the A. tenuis recruits. Recruits of both coral species were well amplicon sequence variant (ASV) level, a higher-resolution pigmented, meaning that they successfully established symbiosis equivalent of the operational taxonomic unit (OTU) (Callahan with C. goreaui (Figure 2B). et al., 2017). Taxa for which the minimum overall relative abundance was lower than 10−5 were filtered out of the dataset in Sequence Data Statistics and Diversity order to mitigate the generation of spurious sequences (Bokulich After removal of rare sequences (minimum total relative et al., 2013). To account for the variability in the number of abundance < 10−5 ), a total of 1,162,570 high quality 16S rRNA reads per sample, data counts were rarefied to an even depth, gene reads were obtained for 53 samples, representing 2,026 corresponding to the minimum number of reads across samples. ASVs belonging to 196 bacterial families. The number of reads Recruit survival rates were modeled using generalized linear per sample varied between 10,290 and 34,180 (Supplementary models with binomial distribution, and linear contrasts were used Figure S1). Coral recruits and water samples were respectively to test for differences between the treatments. Alpha diversity as associated with a total of 1,362 ASVs and 960 ASVs, sharing a a measure for richness was computed using Chao and Shannon total of 417 ASVs. α-diversity indices (Lande, 1996; Legendre and Legendre, 1998). Samples were rarefied at 10,290 reads to account for the To better understand differences in richness between coral variability in sequencing effort and rarefaction curves confirmed species and treatments, generalized linear models with gamma that this depth was sufficient to reflect the diversity present in distributions were fitted, and linear contrasts were used to test the samples. Rarefaction curves displayed an asymptote before differences in richness. Differences in community composition reaching the set threshold of 10,290 reads (Supplementary (β-diversity, Anderson et al., 2006) were computed using Figure S2). Moreover, in all sample groups after rarefaction, the Bray–Curtis dissimilarity matrices and tested via permutational similarity between the observed and expected richness (Chao multivariate analysis of variance (PERMANOVA, Anderson, index) was greater than 96%. For each sample category, the 2001). Variation in community composition among samples observed and Chao diversity index, evenness and Shannon was visualized using Principal Coordinate Analysis (PCoA) α-diversity index estimates are reported in Table 2 together with (Legendre and Legendre, 1998). A test for homogeneity of their standard deviations (s.d). multivariate dispersions (Anderson, 2006) was used to check Platygyra daedalea recruits displayed significantly higher for homogeneity of variances and pairwise comparisons were bacterial species richness than A. tenuis recruits for both control performed between groups using a single-step (Scheffé, 1953) and inoculated treatments (Figure 3; zCtrl = 2.905, p < 0.01; and the Benjamin and Hochberg corrections for multiple testing zInoc = 3.247, p < 0.001). Within each coral species, the inoculated (Benjamini and Hochberg, 1995). In addition, we performed recruits had significantly lower bacterial species richness than differential pairwise abundance comparisons using unrarefied the control recruits (zAcro = −2.439, p = 0.015; zPlaty = −2.091, data based on generalized linear models with negative binomial p = 0.036). The Shannon α-diversity index was higher for distributions as described in Love et al. (2014), to identify P. daedalea, although the difference was only significant in taxa for which there was a significant logarithmic fold change the inoculated treatment (Figure 3; zCtrl = −1.497, p = 0.134; (LFC) in abundance between groups. In addition, random Forest zInoc = −2.521, p = 0.012); Shannon α-diversity index was classification was used to understand feature importance of the significantly lower in inoculated A. tenuis compared to inoculated bacterial community (Breiman, 2001). P. daedalea (zAcro = 2.422, p = 0.015; zPlaty = 1.398, p = 0.162). Frontiers in Microbiology | www.frontiersin.org 7 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 2 | (A) Survival of control and inoculated A. tenuis and P. daedalea recruits at the end of the experiment (42 days post-spawning, after seven inoculations) as compared to the start of the experiment (16 days post-spawning, before any inoculation). (B) Representative pictures of A. tenuis (top) and P. daedalea recruits (bottom) at four weeks post-settlement. Black scale bar is 1 mm. TABLE 2 | Overview of the number of samples and corresponding diversity indices for 16S rRNA sequences (average ± s.d.). Sampling group No of samples Observed Richness Chao diversity index estimate Evenness Shannon α-diversity index A. tenuis Control 6 307 ± 44 314 ± 47 0.81 ± 0.03 4.61 ± 0.21 Inoculated 6 254 ± 51 256 ± 52 0.77 ± 0.04 4.27 ± 0.38 P. daedalea Control 6 391 ± 32 401 ± 35 0.81 ± 0.02 4.83 ± 0.14 Inoculated 6 333 ± 34 336 ± 37 0.79 ± 0.02 4.62 ± 0.16 Water Control 6 316 ± 21 330 ± 20 0.72 ± 0.02 4.13 ± 0.12 Inoculated 6 285 ± 20 293 ± 21 0.74 ± 0.02 4.17 ± 0.10 C. goreaui culture 1 71 69 0.67 2.86 Bacterial cocktail (7 inoculations) 7 8±1 8±1 0.82 ± 0.04 1.75 ± 0.05 Blank DNA extractions 6 47 ± 36 47 ± 36 0.58 ± 0.07 2.16 ± 0.36 No template PCRs 3 18 ± 7 18 ± 7 0.57 ± 0.02 1.65 ± 0.29 The Symbiodiniaceae composition of coral recruits was the aquaria, and negative controls (i.e., blank extractions and analyzed and compared with the pure culture that was inoculated no-template PCR products), were subjected to PERMANOVA to the recruits at the start of the experiment (Supplementary analysis, which confirmed that corals and water samples hosted Figure S3). Both A. tenuis and P. daedalea juveniles exclusively significantly different bacterial communities from negative harbored C. goreaui (formerly known as clade C; LaJeunesse controls. Bacterial communities in water and coral also differed et al., 2018) sequence types that were also recovered from significantly from one another (Supplementary Table S1 and the monoclonal cultures used for inoculation. Since more than Supplementary Figure S5). one C. goreaui sequence variant was found, we conclude that A total of 237 out of the 2,026 detected ASVs were present intragenomic variation exists within the inoculated C. goreaui in the negative control samples, with four ASVs belonging culture. Intragenomic variation is common and widespread to the genera Burkholderia-Paraburkholderia, Ralstonia and an within the Symbiodiniaceae (Wilkinson et al., 2015, 2018). unidentified Oxalobacteraceae dominating the blank (73.2%) and The C. goreaui culture also comprised 74 bacterial ASVs, no-template samples (73.4%), (Supplementary Table S2). As largely dominated by Rhodobacteraceae and Flammeovirgaceae these four ASVs had a 10–100 fold lower abundance in the test (Supplementary Figure S4). samples, it is reasonable to assume that their presence was due to laboratory or reagent contamination. As recommended (Lee Sequencing Control Samples et al., 2015), these four ASVs were removed from the dataset prior DNA contamination originating from laboratory reagents to further analyses. and sample handling, as well as biases occurring during The use of mock communities has been highly recommended PCR amplification and sequencing have often been reported to assist in estimating biases in the overall metabarcoding process to distort the results obtained via metabarcoding (Salter (e.g., PCR, sequencing and bioinformatics) (Yeh et al., 2018). In et al., 2014; Fouhy et al., 2016). The bacterial community this study, the seven inocula comprising seven different bacterial composition data from coral recruits, water collected from species were used as mock communities. The cultures were Frontiers in Microbiology | www.frontiersin.org 8 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 3 | Boxplots representing richness based on Chao and Shannon α-diversity indices in coral recruits at the end of the experiment. Different letters denote groups with significantly different richness or Shannon α-diversity index (p < 0.05). Within each graph, groups with the same letter do not have a statistically significant difference in α-diversity. FIGURE 4 | PCoA visualization using the Bray–Curtis dissimilarity measurement separating samples by treatment and host species. Samples were collected at the end of the experiment (day 42) following the seven bacterial inoculations. quantified by OD660 and from this, their relative abundances (in terms of cells/mL) were estimated. These estimates were similar, although not identical, to the organismal relative abundances obtained by metabarcoding data analysis (Supplementary Figure p = 0.001). After checking for homogeneity of variances among S6). The ASVs with 100% sequence identity to the seven bacterial all four groups (PERMDISP with 999 permutations: F = 2.77, cultures represented more than 99% of the counts in each p = 0.066), pairwise comparisons with correction for multiple inoculum. We thus conclude there were minimal biases in testing confirmed that the treatment had a significant effect the metabarcoding process. These seven ASVs will henceforth within each coral species (A. tenuis Inoc vs. Ctrl: p = 0.002, be referred to as “ASV1” to “ASV7” (being Acinetobacter, P. daedalea Inoc vs. Ctrl: p = 0.002). Bacterioplanes, Marinobacter, Paracoccus, Pseudoalteromonas, The relative abundances of two and six bacterial strains Pseudovibrio, Vibrio, respectively). used in the inocula were significantly greater in the inoculated compared to the control recruits for A. tenuis and P. daedalea, respectively (Figure 5). For each coral species, a differential Influence of Bacterial Inoculation and abundance analysis was performed on the non-rarefied dataset to Host Species on Coral-Associated identify ASVs for which there is a significant LFC in abundance Bacterial Communities between inoculated and control recruits. Using a significance In all treatments, A. tenuis and P. daedalea recruits were level of α = 0.05, only bacteria contained in the inocula were dominated by the bacterial class Alphaproteobacteria, followed reported to have a significantly different LFC between treatments by Gammaproteobacteria (Supplementary Figure S7). PCoA (two for A. tenuis and six for P. daedalea, Table 3). Random based on ASV data showed that coral samples clustered Forest analyses performed on the microbiome data of each according to host species and treatment (inoculation vs. control; coral species corroborated the results obtained with differential Figure 4). On inspection of the multivariate homogeneity of abundance analysis. In A. tenuis and P. daedalea, ASVs matching group variances, A. tenuis recruits were characterized by a the seven inoculated ones were classified as the most important higher dispersion than P. daedalea recruits (PERMDISP with 999 predictors for the discrimination of inoculated and control permutations: F = 14.72, p = 0.001). recruits (Figure 6). In this analysis, several bacterial taxa (such The bacterial microbiomes harbored by A. tenuis and as members of Rhodobacteraceae and Rhodospirillaceae) not P. daedalea recruits were significantly different from one present in the inocula were also classified as important predictors another (for both treatments), and the difference in community for separating control and inoculated recruits in each coral composition between inoculated and control recruits of the same species (Figure 6). species was also statistically significant (PERMANOVA with 999 Despite being reared in a common environment, A. tenuis permutations: F Treatment = 3.6279, p = 0.001; F Species = 8.48, and P. daedalea recruits overall developed distinct bacterial Frontiers in Microbiology | www.frontiersin.org 9 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 5 | Bubble plot representing the relative abundance within each sample of the seven bacterial ASVs used for inoculation (i.e., having a 100% sequence identity to the present in the inocula). The asterisks next to the bacterial names indicate that this ASV had a significantly different LFC in abundance between control and inoculated recruits. Coral samples were collected at the end of the experiment (day 42) following the seven bacterial inoculations. Red refers to A. tenuis while blue refers to P. daedalea. Rel ab, relative abundance. TABLE 3 | Bacteria having a significantly different LFC in abundance between genus Endozoicomonas), Comamonadaceae, Colwelliaceae inoculated and control recruits. and Fibrobacteraceae were detected in A. tenuis but not Coral host ASV Bacterial genus Log2 Fold Change pAdj in P. daedalea, while Nitrosomonadaceae, Polyangiaceae, Parachlamydiaceae, Oxalobacteraceae were detected in A. tenuis ASV4 Paracoccus 5.69 2.4·10−4 P. daedalea and absent from A. tenuis. ASV6 Pseudovibrio 8.49 1.57·10−9 P. daedalea ASV2 Bacterioplanes 7.70 8.85·10−6 ASV3 Marinobacter 5.84 6.59·10−3 ASV4 Paracoccus 8.03 5.25·10−6 DISCUSSION ASV5 Pseudoalteromonas 5.48 2.75·10−4 ASV6 Pseudovibrio 8.19 4.29·10−6 Coral Host Factors Influence the ASV7 Vibrio 5.11 1.40·10−2 Microbiome Results of the differential abundance analysis are reported for the Wald test While some studies report that coral-associated bacterial and parametric fit. communities depend on environmental factors and geographical location (Littman et al., 2009b; Lee et al., 2012; Pantos et al., communities. The most abundant bacterial families across coral 2015; Hester et al., 2016), other reports indicate that corals samples showed some degree of variability between A. tenuis and harbor species-specific bacterial assemblages (Rohwer et al., P. daedalea samples, as well as between inoculated and control 2002; Rohwer and Kelley, 2004; Bourne et al., 2008; Littman recruits (Figure 7). For instance, family Rhodobacteraceae (to et al., 2009b; Tremblay et al., 2010; Apprill et al., 2012; Morrow which the inocula strains Paracoccus and Pseudovibrio belong) et al., 2012; Ainsworth et al., 2015; Zhang et al., 2015; Chu was highly abundant in A. tenuis and particularly so in the and Vollmer, 2016). The observed differences in bacterial inoculated recruits. Family Alteromonadaceae was also more communities retrieved from corals versus those from the abundant in A. tenuis compared to P. daedalea (Figure 7). surrounding water column are compelling evidence that A differential abundance analysis was performed on the coral coral-bacteria associations are non-random and subject to recruits within each treatment to identify the ASVs for which selective mechanisms (Sunagawa et al., 2010; Sweet et al., the relative abundance significantly differed between A. tenuis 2010). Emerging data indicate both environmental and host and P. daedalea (Figure 8). ASVs belonging to Alteromonadaceae factors are drivers of coral-associated microbial community (two different Alteromonas ASVs) were found significantly more composition (Kvennefors et al., 2010; Kelly et al., 2014; abundant in A. tenuis for both control and inoculated corals. Zhang et al., 2015; Frade et al., 2016), and our study provides Rhodobacteraceae [including Paracoccus (ASV4), Pseudovibrio strong support for a role of host factors in shaping the (ASV6) and a Roseovarius (ASV)] were significantly more composition of the microbiome. At the sampling time abundant in inoculated A. tenuis compared to inoculated point, bacterial communities of the two coral host species P. daedalea. In the control treatments, Vibrionaceae (three Vibrio cultured in the same aquaria since the early recruit stage ASVs including ASV7) were more abundant in A. tenuis than differed in terms of α-diversity (bacterial communities in in P. daedalea. P. daedalea were richer than those in A. tenuis, Figure 2), Some assignable families exclusively appeared in one β-diversity (there was more variability across samples in coral species (5 in A. tenuis and 10 in P. daedalea) or in the A. tenuis than in P. daedalea, Figure 3), and in the relative water (39 families) (Supplementary Figure S8). For example, abundance (Figure 7) and presence/absence of particular members of the families Hahellaceae (all belonging to the bacterial taxa (Supplementary Figure S8). In addition to Frontiers in Microbiology | www.frontiersin.org 10 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 6 | Representative output of the random Forest classification with 100 iterations to identify the most important predictor ASVs in A. tenuis (orange) and P. daedalea (blue). The genus or family (when known) of each ASV is specified. Names written in bold represent the inoculated ASVs. harboring distinct bacterial communities, the two corals as both species were inoculated with the same monoclonal responded differently to the inoculations in that the degree C. goreaui culture. of shift away from the bacterial community harbored by the control recruits was larger in A. tenuis than in P. daedalea (Figure 3). The intrinsically higher microbial diversity in A. tenuis and P. daedalea Recruits P. daedalea might have buffered the perturbation introduced Responded Differently to the Inocula by the inocula, as empirical studies and mathematical models Some of the ASVs corresponding to the inoculated bacterial support the notion that diverse ecosystems are more resistant strains were more significantly abundant in the inoculated to invasion than systems with lower diversity (Levine and compared to the control coral recruits (Figure 5). The two corals, D’Antonio, 1999). The bacterial profiles of control A. tenuis A. tenuis and P. daedalea, may exhibit specific preferences for and P. daedalea recruits are challenging to compare with certain bacterial taxa and possess different selection mechanisms already published data. Previous studies on A. tenuis also to shape their bacterial communities. Potential probiotics for report dominance of Alpha- and Gammaproteobacteria corals might therefore need to be tailored to particular bacterial (Supplementary Figure S7), but with variable proportions species in order to achieve optimal uptake. Interestingly, the same (Littman et al., 2009b; Ceh et al., 2011). Only limited two ASVs significantly enriched in inoculated A. tenuis (ASV4 – information exists on prokaryotic communities associated Paracoccus and ASV6 – Pseudovibrio) were also characterized by with P. daedalea; one study focused on culturable Actinobacteria the highest LFC in P. daedalea (Table 3). ASV4 and ASV6 both (Mahmoud and Kalendar, 2016), while another work reported belong to class Alphaproteobacteria, which was the dominant a P. daedalea sample with a high abundance of Alpha- and bacterial class in all coral recruits in this study, while the five other Gammaproteobacteria, mostly represented by Endozoicomonas ASVs used in the inocula are Gammaproteobacteria. (Liang et al., 2017). Different species of adult corals have been Acropora tenuis is a fast growing and branching coral, reported to maintain distinct microbiomes when reared while P. daedalea is slow growing, massive coral with a thick in identical environments (Sweet et al., 2013), and our tissue layer, which provides more shading to the in hospite study shows that host factors already play a role from early algal symbionts and energy reserves for the host (Loya et al., ontogeny. Importantly, our data were not confounded by 2001; Putnam et al., 2017). Such contrasting morphological and the Symbiodiniaceae species, which can influence bacterial physiological properties might create microhabitats suitable for community composition in young corals (Littman et al., 2009a), different types of microbes and also dictate requirements for Frontiers in Microbiology | www.frontiersin.org 11 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation possible that some bacteria were inherited from parental coral colonies and/or horizontally acquired from the environment. The latter includes bacteria present in the FSW entering the rearing tanks, as well as the bacteria found in the C. goreaui culture, which was administered to all recruits (Supplementary Figures S2, S4). However, knowledge of this baseline microbiome was not necessary to evaluate the effect of bacterial inoculation, since inoculated recruits were compared to control counterparts, which were treated in the same manner. By exposing coral recruits to a chosen bacterial consortium, we were able to significantly modify their microbiome (Figure 3). Several bacterial strains in the inocula (i.e., two in the case of A. tenuis and six for P. daedalea) were statistically significantly enriched in the inoculated recruits (Figure 5). Bacteria used in the inocula were also among the major ASVs driving the separation of bacterial communities associated with control and inoculated recruits (Table 3). Effective inoculations were thus not precluded by using bacteria isolated from non-coral organisms, which demonstrates a degree of flexibility in the coral microbiome. Altogether, this study supports the proof-of-concept for the feasibility of manipulating coral-associated prokaryotes (dos Santos et al., 2015; Rosado et al., 2018). Knowledge Gaps, Experimental FIGURE 7 | Relative abundance of the 10 most prevalent bacterial families Shortcomings and Directions for Future across all coral samples collected at the end of the experiment (day 42). Research To gain more insights into the efficacy of probiotics for coral reef restoration and conservation, future research should target specific microbial functions. It is therefore not surprising that several aspects not covered in the present study. First, an A. tenuis and P. daedalea established associations with different optimal inoculation regime should be developed through the microbial communities over the course of this experiment. determination of suitable inoculation frequency and bacterial Field surveys have reported that branching and massive corals cell density. In the present work, the seven inoculations at naturally tend to associate with distinct bacterial communities, 105 –106 cells/mL arbitrarily followed a 3-4-3-4 days pattern, with higher α-diversities generally observed in massive species as no standard procedure has yet been established. Alternative (Liang et al., 2017). strategies have also yielded effective outcomes, such taking It is possible the young A. tenuis and P. daedalea had already coral fragments out of the water and inoculating them with acquired a subset of bacteria from the water surrounding parental 1 mL (107 cells/mL) twice 5 days apart (Rosado et al., 2018). colonies following gamete release and/or from the water in Nonetheless, systematic experiments are required to determine which larvae and recruits were reared. Despite being filtered with optimal inoculation protocols, by using the same bacterial taxa, 0.4 µm pore-sized membranes, the flow-through water in the coral species and culture conditions while varying bacterial cell experimental system contained some bacteria (as shown by the density and/or frequency of inoculum administration. 16S rRNA gene sequences found in water samples). However, This experiment was conducted over a short timescale the water-associated microbial communities were considerably (4 weeks), which did not allow the long-term stability of coral- divergent from the coral microbiota (Supplementary Figure bacterial associations to be evaluated. It is not clear whether S5), which emphasizes the ability of corals to select certain the dosed bacteria would be retained over time, especially bacteria from the environment. Alternatively, adult A. tenuis when no further bacteria are delivered. Also, shifts in the coral and P. daedalea may have vertically transmitted some bacterial microbiome following transfer between field and aquaria settings symbionts to their offspring prior to spawning. Evidence to have been reported (Pratte et al., 2015; Röthig et al., 2017), which date indicates that the bacterial transmission mode in broadcast could challenge the implementation of microbiome manipulation spawning corals is variable (Sharp et al., 2010; Lema et al., 2014; strategies if the goal is to achieve long-term changes in the Neave et al., 2016; Leite et al., 2017; Zhou et al., 2017). coral-associated microbiome. The coral microbiome is dynamic (Mouchka et al., 2010; Sweet and Bulling, 2017), especially Effect of Bacterial Inoculation on Coral in early life stages (Littman et al., 2009a; Zhou et al., 2017; Recruit Microbiome Epstein et al., 2019a). As recruits develop and grow, the complex Bacterial community composition was not assessed at the start microbial communities associated with juveniles shift toward less of the study due to a limited number of recruits available. It is diverse adult microbiomes. Due to these natural processes, the Frontiers in Microbiology | www.frontiersin.org 12 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation FIGURE 8 | Results from the differential abundance analysis with significance level α = 0.01, expressed as Log2 Fold Change for the comparison between A. tenuis and P. daedalea recruits within each treatment: (A) control recruits and (B) inoculated recruits. Each dot represents one ASV (including ASVs 4, 6, and 7 from the inocula), here identified at the genus level when possible. Positive values represent ASVs proportionally more abundant in P. daedalea, while negative values are ASVs proportionally more abundant in A. tenuis. impact of bacterial inoculations during early life stages might Assessing the location of the inoculated bacteria within the diminish over time, but this needs to be tested. In addition host will also be informative. In this study, the enrichment of to life stage, numerous other biotic and abiotic factors may the inoculated bacteria in the coral recruits could originate from influence the coral microbiome, such as abundance and changes internalized communities or from organisms adhering to the in Symbiodiniaceae species, coral disease, and water parameters surface of the recruits. Bacteria within the mucus layer or the of salinity, temperature, pH and nutrient levels (reviewed in gastric cavity of polyps might be transient due the high variability Hernandez-Agreda et al., 2016b). The capacity for sustained that usually characterizes these microhabitats (Sweet et al., 2010; bacteria-coral associations following inoculation needs to be Thompson et al., 2015; Glasl et al., 2016). Moreover, it needs to be studied beyond laboratory-controlled conditions to understand verified that inoculated bacteria are not simply taken up as food the impact of these fluctuating factors. source, but rather become part of the coral microbiome. Stable An additional important piece of knowledge lacking from our partnerships have been suggested to occur between the coral study is the growth and viability of the bacterial cells used to host and several bacterial taxa occupying intracellular spaces prepare the consortium prior to coral inoculation. The possibility (Ainsworth et al., 2015). These bacteria have been hypothesized that some of the 16S rRNA gene sequences recovered from to comprise the coral core microbiome, which is conserved the recruits originated from dead organisms or residual DNA across time and geographical location (Hernandez-Agreda et al., adhering to the polyps can thus not be excluded. However, 2016a). Visualizing sections of inoculated corals via fluorescence our results support a true change in the coral microbiomes in situ hybridization with probes targeting particular bacteria rather than residual DNA on the samples being responsible would allow localizing them when administered to specific for the observed changes. Firstly, A. tenuis and P. daedalea host niches (Ainsworth et al., 2006; Wada et al., 2016). Even exhibited distinct enrichments of ASVs corresponding to the though a stable association seems more likely when microbes bacterial strains used in the inocula. Secondly, not all strains are endosymbiotic, consistently detecting the inoculated bacteria were significantly enriched in inoculated recruits (two in the over time in any coral compartment would support the notion case of A. tenuis and six in the case of P. daedalea – Figure 5). of a lasting partnership. Stably labeling cells with fluorescent Differential enrichment according to host species and failure to proteins through genetic engineering as suggested by Pollock detect a statistically significant higher abundance for all seven et al. (2015) would allow to precisely track the probiotic bacteria, strains in inoculated recruits indicate a preferential uptake of as well as the offspring generated from their division. bacterial groups by the corals. Lastly, based on random Forest classification, ASVs other than the ones present in the inocula Practical Considerations were identified as important predictors to separate control and The application of probiotics to wild corals (or to captive corals inoculated recruits (Figure 6). These patterns suggest a change in aimed to be deployed to the field for reef restoration) requires the host bacterial community composition that is promoted by further research into the possibility of scaling up such efforts. the inoculation. While industrial infrastructures to produce cultured bacterial Frontiers in Microbiology | www.frontiersin.org 13 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation consortia already exist, the latter need to be suitably delivered to and stability of inoculated bacteria in corals still need to be corals. Inocula could be prepared by encapsulating bacteria into characterized. By demonstrating the feasibility to manipulate the microscopic feed particles (Peixoto et al., 2017). Alternatively, coral microbiome in a laboratory setting, our results provide probiotic bacteria could be administered through a heterotrophic hope for the application of probiotics in coral reef conservation food source such as the brine shrimp Artemia, which has already and restoration. proven successful in spiny lobster larvae (Goulden et al., 2012). As the delivery of probiotics to entire reef systems could be unmanageable due to their large size as compared to available AUTHOR CONTRIBUTIONS resources, efforts could be prioritized to most vulnerable or ecologically relevant sites. KD, MO, and LB developed the study design. KD conducted Importantly, potential environmental impacts should also be the experiment and collected the data. PM and KD performed strictly evaluated. It is indeed argued that coral reef ecosystems the statistical analyses. KD wrote the manuscript with might be inadvertently harmed through manipulations that we contributions from all authors. All authors read and approved cannot entirely predict or control (Sweet et al., 2017). For the final manuscript. instance, disease agents such as pathogens or parasites could be transferred from captive systems to the natural environment and impact the native fauna (Sweet et al., 2017). Moreover, potential FUNDING probiotic bacteria (such as Vibrio) could be converted into pathogens if they acquired the right virulence genes (Bruto et al., This work was funded by an Australian Research Council 2017). This would be particularly problematic if probiotics are Discovery Project grant (DP160101468 to MO and LB) and administered at large cell numbers. In general, an overabundance AIMS. KD acknowledges the receipt of the International Research of certain bacteria in the wild could have unintended effects Scholarship and Australian Post-graduate Award, as well as on the ecosystem. Therefore, future research should focus on the Holsworth Endowment offered by the Ecological Society of understanding the risk of probiotics before releasing microbes Australia. Symbols for diagrams were modified from the ones into open reefs (National Academies of Sciences Engineering, made available by the Integration and Application Network and Medicine, 2019). It has already been advocated that rigorous (http://ian.umces.edu/symbols). MO acknowledges Australian scientific trials and risk/benefit analyses should be carried out Research Council Laureate Fellowship FL180100036. prior to introducing any foreign microbial communities into the ocean (van Oppen et al., 2017). ACKNOWLEDGMENTS CONCLUSION The authors are particularly grateful to Dr. Jean-Baptiste Raina for providing the Paracoccus, Pseudovibrio, Acinetobacter, and This study involved the co-culturing of two taxonomically Bacterioplanes isolates and to Miss Ashley Dungan for providing divergent coral species and inoculating them with a bacterial the Pseudoalteromonas and Marinobacter isolates. The authors cocktail generated in the laboratory. Despite sharing the same thank the team of the National Sea Simulator at AIMS for environment since a very early life stage, A. tenuis and P. daedalea technical support in conducting the experiment and Mr. Carlos recruits had distinct bacterial communities at the time of Alvarez-Roa for providing the Symbiodiniaceae culture. KD sampling, which exhibited different responses to the inocula. thanks Dr. Nicole Webster and Dr. J. B. Raina for discussions and In the absence of confounding factors such as environmental for providing advice, as well as Dr. Sara Bell, Lesa Peplow, and Dr. parameters and algal symbiont type, our findings highlight that Patrick Laffy for technical support. host factors play a noticeable role in shaping coral bacterial community composition. As the microbiomes of A. tenuis and P. daedalea changed in response to the inoculum, the bacteria SUPPLEMENTARY MATERIAL present in the cocktail were identified as the ones driving the main differences between inoculated and control corals. The Supplementary Material for this article can be found The bacterial pool surrounding these young recruits therefore online at: https://www.frontiersin.org/articles/10.3389/fmicb. also influenced their microbiomes. The long-term maintenance 2019.01702/full#supplementary-material REFERENCES Ainsworth, T. D., Fine, M., Blackall, L. L., and Hoegh-Guldberg, O. (2006). Fluorescence in situ hybridization and spectral imaging of coral-associated Abrego, D., VAN Oppen, M. J., and Willis, B. L. (2009). Highly infectious bacterial communities. Appl. Environ. Microbiol. 72, 3016–3020. doi: 10.1128/ symbiont dominates initial uptake in coral juveniles. Mol. Ecol. 18, aem.72.4.3016-3020.2006 3518–3531. doi: 10.1111/j.1365-294x.2009.04275.x Ainsworth, T. D., Krause, L., Bridge, T., Torda, G., Raina, J. B., Zakrzewski, Agilent Genomics (2018). E. coli Cell Culture Concentration from OD600 M., et al. (2015). The coral core microbiome identifies rare bacterial taxa Calculator. Available at: http://www.genomics.agilent.com/biocalculators/ as ubiquitous endosymbionts. ISME J. 9, 2261–2274. doi: 10.1038/ismej. calcODBacterial.jsp (accessed 2018). 2015.39 Frontiers in Microbiology | www.frontiersin.org 14 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation Anderson, M. J. (2001). A new method for non-parametric multivariate analysis of Damjanovic, K., Blackall, L. L., Webster, N. S., and van Oppen, M. J. H. variance. Austral Ecol. 26, 32–46. doi: 10.1111/j.1442-9993.2001.01070.pp.x (2017). The contribution of microbial biotechnology to mitigating coral Anderson, M. J. (2006). Distance-based tests for homogeneity of multivariate reef degradation. Microb. Biotechnol. 10, 1236–1243. doi: 10.1111/1751-7915. dispersions. Biometrics 62, 245–253. doi: 10.1111/j.1541-0420.2005.00440.x 12769 Anderson, M. J., Ellingsen, K. E., and McArdle, B. H. (2006). Multivariate De’ath, G., Fabricius, K. E., Sweatman, H., and Puotinen, M. (2012). The 27-year dispersion as a measure of beta diversity. Ecol. Lett. 9, 683–693. doi: 10.1111/j. decline of coral cover on the Great Barrier Reef and its causes. Proc. Natl. Acad. 1461-0248.2006.00926.x Sci. U.S.A. 109, 17995–17999. doi: 10.1073/pnas.1208909109 Andersson, A. F., Lindberg, M., Jakobsson, H., Bäckhed, F., Nyrén, P., and DePeters, E. J., and George, L. W. (2014). Rumen transfaunation. Immunol. Lett. Engstrand, L. (2008). Comparative analysis of human gut microbiota by 162, 69–76. doi: 10.1016/j.imlet.2014.05.009 barcoded pyrosequencing. PLoS One 3:e2836. doi: 10.1371/journal.pone. dos Santos, H. F., Duarte, G. A., Rachid, C. T., Chaloub, R. M., Calderon, 0002836 E. N., Marangoni, L. F., et al. (2015). Impact of oil spills on coral reefs Apprill, A., Marlow, H. Q., Martindale, M. Q., and Rappe, M. S. (2009). The onset of can be reduced by bioremediation using probiotic microbiota. Sci. Rep. 5: microbial associations in the coral Pocillopora meandrina. ISME J. 3, 685–699. 18268. doi: 10.1038/ismej.2009.3 Epstein, H., Smith, H., Torda, G., and van Oppen, M. (2019b). Microbiome Apprill, A., Marlow, H. Q., Martindale, M. Q., and Rappe, M. S. (2012). Specificity engineering: enhancing climate resilience in corals. Front. Ecol. Environ. 17, of associations between bacteria and the coral Pocillopora meandrina during 100–108. doi: 10.1002/fee.2001 early development. Appl. Environ. Microbiol. 78, 7467–7475. doi: 10.1128/aem. Epstein, H., Torda, G., Munday, P. L., and van Oppen, M. J. H. (2019a). 01232-12 Parental and early life stage environments drive establishment of bacterial Arif, C., Daniels, C., Bayer, T., Banguera-Hinestroza, E., Barbrook, A., Howe, C. J., and dinoflagellate communities in a common coral. ISME J. 13, 1635–1638. et al. (2014). Assessing Symbiodinium diversity in scleractinian corals via next- doi: 10.1038/s41396-019-0358-3 generation sequencing-based genotyping of the ITS2 rDNA region. Mol. Ecol. Fisher, P. L., Malme, M. K., and Dove, S. (2011). The effect of temperature stress 23, 4418–4433. doi: 10.1111/mec.12869 on coral–Symbiodinium associations containing distinct symbiont types. Coral Benjamini, Y., and Hochberg, Y. (1995). Controlling the false discovery rate: a Reefs 31, 473–485. doi: 10.1007/s00338-011-0853-0 practical and powerful approach to multiple testing. J. R. Stat. Soc. 57, 289–300. Fouhy, F., Clooney, A. G., Stanton, C., Claesson, M. J., and Cotter, P. D. (2016). doi: 10.1111/j.2517-6161.1995.tb02031.x 16S rRNA gene sequencing of mock microbial populations- impact of DNA Blackall, L. L., Wilson, B., and van Oppen, M. J. (2015). Coral - The world’s most extraction method, primer choice and sequencing platform. BMC Microbiol. diverse symbiotic ecosystem. Mol. Ecol. 24, 5330–5347. doi: 10.1111/mec.13400 16:123. Bokulich, N. A., Subramanian, S., Faith, J. J., Gevers, D., Gordon, J. I., Knight, R., Frade, P. R., Roll, K., Bergauer, K., and Herndl, G. J. (2016). Archaeal and bacterial et al. (2013). Quality-filtering vastly improves diversity estimates from Illumina communities associated with the surface mucus of Caribbean corals differ in amplicon sequencing. Nat. Methods 10, 57–59. doi: 10.1038/nmeth.2276 their degree of host specificity and community turnover over reefs. PLoS One Boulotte, N. M., Dalton, S. J., Carroll, A. G., Harrison, P. L., Putnam, H. M., Peplow, 11:e0144702. doi: 10.1371/journal.pone.0144702 L. M., et al. (2016). Exploring the Symbiodinium rare biosphere provides Gatesoupe, F. J. (1999). The use of probiotics in aquaculture. Aquaculture 180, evidence for symbiont switching in reef-building corals. ISME J. 10, 2693–2701. 147–165. doi: 10.1016/s0044-8486(99)00187-8 doi: 10.1038/ismej.2016.54 Glasl, B., Herndl, G. J., and Frade, P. R. (2016). The microbiome of coral Bourne, D., Iida, Y., Uthicke, S., and Smith-Keune, C. (2008). Changes in coral- surface mucus has a key role in mediating holobiont health and survival upon associated microbial communities during a bleaching event. ISME J. 2, 350–363. disturbance. ISME J. 10, 2280–2292. doi: 10.1038/ismej.2016.9 doi: 10.1038/ismej.2007.112 Gouda, S., Kerry, R. G., Das, G., Paramithiotis, S., Shin, H. S., and Patra, J. K. Bourne, D. G., Morrow, K. M., and Webster, N. S. (2016). Insights into the coral (2018). Revitalization of plant growth promoting rhizobacteria for sustainable microbiome: underpinning the health and resilience of reef ecosystems. Annu. development in agriculture. Microbiol. Res. 206, 131–140. doi: 10.1016/j.micres. Rev. Microbiol. 70, 317–340. doi: 10.1146/annurev-micro-102215-095440 2017.08.016 Breiman, L. (2001). Random forests. Mach. Learn. 45, 5–32. Goulden, E. F., Hall, M. R., Bourne, D. G., Pereg, L. L., and Hoj, L. (2012). Bruto, M., James, A., Petton, B., Labreuche, Y., Chenivesse, S., Alunno-Bruscia, Pathogenicity and infection cycle of Vibrio owensii in larviculture of the ornate M., et al. (2017). Vibrio crassostreae, a benign oyster colonizer turned into a spiny lobster (Panulirus ornatus). Appl. Environ. Microbiol. 78, 2841–2849. pathogen after plasmid acquisition. ISME J. 11, 1043–1052. doi: 10.1038/ismej. doi: 10.1128/aem.07274-11 2016.162 Hai, N. V. (2015). The use of probiotics in aquaculture. J. Appl. Microbiol. 119, Burke, L., Reytar, K., and Spaulding, M. A. P. (2011). Reefs at Risk Revisisted. 917–935. doi: 10.1111/jam.12886 Washington, DC: World Resources Institute. Harrison, P. L., and Booth, D. J. (2007). “Coral reefs: naturally dynamic and Callahan, B. J., McMurdie, P. J., and Holmes, S. P. (2017). Exact sequence variants increasingly disturbed ecosystems,” in Marine Ecology, eds S. D. Connell and should replace operational taxonomic units in marker-gene data analysis. ISME B. M. Gillanders (Melbourne: Oxford University Press), 316–377. J. 11, 2639–2643. doi: 10.1038/ismej.2017.119 Hernandez-Agreda, A., Gates, R. D., and Ainsworth, T. D. (2016a). Defining the Callahan, B. J., McMurdie, P. J., Rosen, M. J., Han, A. W., Johnson, A. J., and core microbiome in corals’ microbial soup. Trends Microbiol. 25, 125–140. Holmes, S. P. (2016). DADA2: high-resolution sample inference from Illumina doi: 10.1016/j.tim.2016.11.003 amplicon data. Nat. Methods 13, 581–583. doi: 10.1038/nmeth.3869 Hernandez-Agreda, A., Leggat, W., Bongaerts, P., and Ainsworth, T. D. (2016b). Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D., The microbial signature provides insight into the mechanistic basis of coral Costello, E. K., et al. (2010). QIIME allows analysis of high- throughput success across reef habitats. mBio 7:e00560-16. community sequencing data. Nat. Methods 7, 335–336. Herve, M. (2018). RVAideMemoire: Testing and Plotting Procedures for Biostatistics. Ceh, J., Van Keulen, M., and Bourne, D. G. (2011). Coral-associated bacterial R package version 0.9-69. communities on Ningaloo Reef, Western Australia. FEMS Microbiol. Ecol. 75, Hester, E. R., Barott, K. L., Nulton, J., Vermeij, M. J., and Rohwer, F. L. (2016). 134–144. doi: 10.1111/j.1574-6941.2010.00986.x Stable and sporadic symbiotic communities of coral and algal holobionts. ISME Chakravarti, L. J., Beltran, V. H., and van Oppen, M. J. H. (2017). Rapid thermal J. 10, 1157–1169. doi: 10.1038/ismej.2015.190 adaptation in photosymbionts of reef-building corals. Glob. Change Biol. 23, Hoegh-Guldberg, O. (2004). Coral reefs in a century of rapid environmental 4675–4688. doi: 10.1111/gcb.13702 change. Symbiosis 37, 1–31. Chu, N. D., and Vollmer, S. V. (2016). Caribbean corals house shared and host- Hoegh-Guldberg, O. (2011). “The impact of climate change on coral reef specific microbial symbionts over time and space. Environ. Microbiol. Rep. 8, ecosystems,” in Coral Reefs: An Ecosystem in Transition, eds Z. Dubinsky and 493–500. doi: 10.1111/1758-2229.12412 N. Stambler (Berlin: Springer), 391–403. doi: 10.1007/978-94-007-0114-4_22 Courtesy of Integration and Application Network (2018). IAN Symbol Libraries Hoegh-Guldberg, O., Mumby, P. J., Hooten, A. J., Steneck, R. S., Greenfield, P., University of Maryland Center for Environmental Science. Available at: Gomez, E., et al. (2007). Coral reefs under rapid climate change and ocean ian.umces.edu/symbols/ acidification. Science 318, 1737–1742. Frontiers in Microbiology | www.frontiersin.org 15 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation Hothorn, T., Bretz, F., and Westfall, P. (2008). Simultaneous inference in general Littman, R., Willis, B. L., and Bourne, D. G. (2011). Metagenomic analysis of the parametric models. Biom. J. 50, 346–363. doi: 10.1002/bimj.200810425 coral holobiont during a natural bleaching event on the Great Barrier Reef. Hughes, T. P., and Kerry, J. T. (2017). Back-to-Back Bleaching Has Now Hit Two- Environ. Microbiol. Rep. 3, 651–660. doi: 10.1111/j.1758-2229.2010.00234.x Thirds of the Great Barrier Reef The Conversation. Available at: https://theco Littman, R. A., Willis, B. L., and Bourne, D. G. (2009a). Bacterial communities nversation.com/back-to-back-bleaching-has-now-hit-two-thirds-of-the-great- of juvenile corals infected with different Symbiodinium (dinoflagellate) clades. barrier-reef-76092 (accessed 2018). Mar. Ecol. Prog. Ser. 389, 45–59. doi: 10.3354/meps08180 Hughes, T. P., Kerry, J. T., Alvarez-Noriega, M., Alvarez-Romero, J. G., Anderson, Littman, R. A., Willis, B. L., Pfeffer, C., and Bourne, D. G. (2009b). Diversities K. D., Baird, A. H., et al. (2017). Global warming and recurrent mass bleaching of coral-associated bacteria differ with location, but not species, for three of corals. Nature 543, 373–377. acroporid corals on the Great Barrier Reef. FEMS Microbiol. Ecol. 68, 152–163. Hughes, T. P., Kerry, J. T., Baird, A. H., Connolly, S. R., Dietzel, A., Eakin, C. M., doi: 10.1111/j.1574-6941.2009.00666.x et al. (2018). Global warming transforms coral reef assemblages. Nature 556, Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change 492–496. doi: 10.1038/s41586-018-0041-2 and dispersion for RNA-seq data with DESeq2. Genome Biol. 15:550. Jones, R. J., Browyer, J., Hoegh-Guldberg, O., and Blackall, L. L. (2004). Dynamics Loya, Y., Sakai, K., Yamazato, K., Nakano, Y., Sambali, H., and van Woesik, of a temperature-related coral disease outbreak. Mar. Ecol. Prog. Ser. 281, 63–77. R. (2001). Coral bleaching: the winners and the losers. Ecol. Lett. 4, doi: 10.3354/meps281063 122–131. doi: 10.1046/j.1461-0248.2001.00203.x Jones, T. A., and Monaco, T. A. (2009). A role for assisted evolution in designing Mahmoud, H. M., and Kalendar, A. A. (2016). Coral-associated actinobacteria: native plant materials for domesticated landscapes. Front. Ecol. Environ. 7, diversity, abundance, and biotechnological potentials. Front. Microbiol. 7:204. 541–547. doi: 10.1890/080028 Martinez Cruz, P., Ibanez, A. L., Monroy Hermosillo, O. A., and Ramirez Saad, Katoh, K., and Standley, D. M. (2013). MAFFT multiple sequence alignment H. C. (2012). Use of probiotics in aquaculture. ISRN Microbiol. 2012:916845. software version 7: improvements in performance and usability. Mol. Biol. Evol. McDonald, D., Clemente, J. C., Kuczynski, J., Rideout, J. R., Stombaugh, J., Wendel, 30, 772–780. doi: 10.1093/molbev/mst010 D., et al. (2012). The biological observation matrix (BIOM) format or: how I Kelly, L. W., Williams, G. J., Barott, K. L., Carlson, C. A., Dinsdale, E. A., learned to stop worrying and love the ome-ome. Gigascience 1:7. Edwards, R. A., et al. (2014). Local genomic adaptation of coral reef-associated McMurdie, P. J., and Holmes, S. (2013). phyloseq: an R package for reproducible microbiomes to gradients of natural variability and anthropogenic stressors. interactive analysis and graphics of microbiome census data. PLoS One Proc. Natl. Acad. Sci. U.S.A. 111, 10227–10232. doi: 10.1073/pnas.140331 8:e61217. doi: 10.1371/journal.pone.0061217 9111 Morrow, K. M., Moss, A. G., Chadwick, N. E., and Liles, M. R. (2012). Bacterial Khalesi, S., Bellissimo, N., Vandelanotte, C., Williams, S., Stanley, D., and Irwin, associates of two Caribbean coral species reveal species-specific distribution and C. (2018). A review of probiotic supplementation in healthy adults: helpful or geographic variability. Appl. Environ. Microbiol. 78, 6438–6449. doi: 10.1128/ hype? Eur. J. Clin. Nutr. 73, 24–37. doi: 10.1038/s41430-018-0135-9 aem.01162-12 Kvennefors, E. C. E., Sampayo, E., Ridgway, T., Barnes, A. C., and Hoegh-Guldberg, Mouchka, M. E., Hewson, I., and Harvell, C. D. (2010). Coral-associated bacterial O. (2010). Bacterial communities of two ubiquitous Great Barrier Reef corals assemblages: current knowledge and the potential for climate-driven impacts. reveals both site- and species-specificity of common bacterial associates. PLoS Integr. Comp. Biol. 50, 662–674. doi: 10.1093/icb/icq061 One 5:e10401. doi: 10.1371/journal.pone.0010401 Muscatine, L., and Porter, J. W. (1977). Reef corals—mutualistic symbioses adapted LaJeunesse, T. C., Bhagooli, R., Hidaka, M., deVantier, L., Done, T., Schmidt, to nutrient-poor environments. Bioscience 27, 454–460. doi: 10.2307/1297526 G. W., et al. (2004). Closely related Symbiodinium spp. differ in relative National Academies of Sciences Engineering, and Medicine (2019). Research dominance in coral reef host communities across environmental, latitudinal Review of Interventions to Increase the Persistence and Resilience of Coral Reefs. and biogeographic gradients. Mar. Ecol. Prog. Ser. 284, 147–161. doi: 10.3354/ Washington, DC: The National Academies Press. meps284147 Neave, M. J., Rachmawati, R., Xun, L., Michell, C. T., Bourne, D. G., Apprill, A., LaJeunesse, T. C., Parkinson, J. E., Gabrielson, P. W., Jeong, H. J., Reimer, et al. (2016). Differential specificity between closely related corals and abundant J. D., Voolstra, C. R., et al. (2018). Systematic revision of symbiodiniaceae Endozoicomonas endosymbionts across global scales. ISME J.11, 186–200. doi: highlights the antiquity and diversity of coral endosymbionts. Curr. Biol. 28, 10.1038/ismej.2016.95 2570–2580.e6. doi: 10.1016/j.cub.2018.07.008 Normile, D. (2016). Massive bleaching killed 35% of the coral on the northern end Lande, R. (1996). Statistics and partitioning of species diversity, and similarity of the Great Barrier Reef. Science. Available at: http://www.sciencemag.org/ne among multiple communities. Oikos 76, 5–13. ws/2016/05/massive-bleaching-killed-35-coral-northern-end-great-barrier-reef Lee, M. D., Walworth, N. G., Sylvan, J. B., Edwards, K. J., and Orcutt, (accessed 2017). B. N. (2015). Microbial communities on seafloor basalts at dorado outcrop Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., et reflect level of alteration and highlight global lithic clades. Front. Microbiol. al. (2016). vegan: Community Ecology Package. R package version 2.4-1. 6:1470. Pantos, O., Bongaerts, P., Dennis, P. G., Tyson, G. W., and Hoegh-Guldberg, Lee, O. O., Yang, J., Bougouffa, S., Wang, Y., Batang, Z., Tian, R., et al. (2012). O. (2015). Habitat-specific environmental conditions primarily control the Spatial and species variations in bacterial communities associated with corals microbiomes of the coral Seriatopora hystrix. ISME J. 9, 1916–1927. doi: 10. from the Red Sea as revealed by pyrosequencing. Appl. Environ. Microbiol. 78, 1038/ismej.2015.3 7173–7184. doi: 10.1128/aem.01111-12 Peixoto, R. S., Rosado, P., Leite, D., Rosado, A. S., and Bourne, D. (2017). Beneficial Legendre, P., and Legendre, L. (1998). Numerical Ecology (Developments in microorganisms for corals (BMC): proposed mechanisms for coral health and Environmental Modelling). Amsterdam: Elsevier Science. resilience. Front. Microbiol. 8:341. Leite, D. C. A., Leao, P., Garrido, A. G., Lins, U., dos Santos, H. F., Pires, D. O., et Pita, L., Rix, L., Slaby, B. M., Franke, A., and Hentschel, U. (2018). The sponge al. (2017). Broadcast spawning coral Mussismilia hispida can vertically transfer holobiont in a changing ocean: from microbes to ecosystems. Microbiome 6:46. its associated bacterial core. Front. Microbiol. 8:176. Pollock, F. J., Krediet, C. J., Garren, M., Stocker, R., Winn, K., Wilson, B., et al. Lema, K. A., Bourne, D. G., and Willis, B. L. (2014). Onset and establishment of (2015). Visualization of coral host–pathogen interactions using a stable GFP- diazotrophs and other bacterial associates in the early life history stages of the labeled Vibrio coralliilyticus strain. Coral Reefs 34, 655–662. doi: 10.1007/ coral Acropora millepora. Mol. Ecol. 23, 4682–4695. doi: 10.1111/mec.12899 s00338-015-1273-3 Levine, J. M., and D’Antonio, C. M. (1999). Elton revisited: a review of evidence Pratte, Z. A., Richardson, L. L., and Mills, D. K. (2015). Microbiota shifts in the linking diversity and invasibility. Oikos 87, 15–26. surface mucopolysaccharide layer of corals transferred from natural to aquaria Liang, J., Yu, K., Wang, Y., Huang, X., Huang, W., Qin, Z., et al. (2017). Distinct settings. J. Invertebr. Pathol. 125, 42–44. doi: 10.1016/j.jip.2014.12.009 bacterial communities associated with massive and branching scleractinian Price, M. N., Dehal, P. S., and Arkin, A. P. (2010). FastTree 2 – approximately corals and potential linkages to coral susceptibility to thermal or cold stress. maximum-likelihood trees for large alignments. PLoS One 5:e9490. doi: 10. Front. Microbiol. 8:979. doi: 10.3389/fmicb.2017.00979 1371/journal.pone.0009490 Liaw, A., and Wiener, M. (2002). Classification and regression by randomForest. R Putnam, H. M., Barott, K. L., Ainsworth, T. D., and Gates, R. D. (2017). The News 2, 18–22. vulnerability and resilience of reef-building corals. Curr. Biol. 27, R528–R540. Frontiers in Microbiology | www.frontiersin.org 16 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation QIIME 2 Development Team (2017a). QIIME 2. Available at: http://qiime2.org Sweet, M. J., Smith, D., Bythell, J. C., and Craggs, J. (2013). Changes in microbial (accessed 2018). diversity associated with two coral species recovering from a stressed state in a QIIME 2 Development Team (2017b). q2-demux. Available at: https://github.com/ public aquarium system. J. Zoo Aquar. Res. 1, 52–60. qiime2/q2-demux (accessed 2018). Teplitski, M., and Ritchie, K. (2009). How feasible is the biological control of coral QIIME 2 Development Team (2017c). q2-taxa. Available at: https://github.com/ diseases? Trends Ecol. Evol. 24, 378–385. doi: 10.1016/j.tree.2009.02.008 qiime2/q2-taxa (accessed 2018). Thompson, J. R., Rivera, H. E., Closek, C. J., and Medina, M. (2015). Microbes in QIIME 2 Development Team (2017d). q2-feature-classifier. Available at: https: the coral holobiont: partners through evolution, development, and ecological //github.com/qiime2/q2-feature-classifier (accessed 2018). interactions. Front. Cell. Infect. Microbiol. 4:176. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., et al. (2013). Tout, J., Siboni, N., Messer, L. F., Garren, M., Stocker, R., Webster, N. S., et al. The SILVA ribosomal RNA gene database project: improved data processing (2015). Increased seawater temperature increases the abundance and alters the and web-based tools. Nucleic Acids Res. 41, D590–D596. structure of natural Vibrio populations associated with the coral Pocillopora R Core Team (2018). R: A Language and Environment for Statistical Computing. damicornis. Front. Microbiol. 6:432. Vienna: R Foundation for Statistical Computing. Tremblay, P., Weinbauer, M. G., Rottier, C., Guérardel, Y., Nozais, C., and Ferrier- Raina, J. B., Tapiolas, D., Motti, C. A., Foret, S., Seemann, T., Tebben, J., et al. Pagès, C. (2010). Mucus composition and bacterial communities associated (2016). Isolation of an antimicrobial compound produced by bacteria associated with the tissue and skeleton of three scleractinian corals maintained under with reef-building corals. PeerJ 4:e2275. doi: 10.7717/peerj.2275 culture conditions. J. Mar. Biol. Assoc. U.K. 91, 649–657. doi: 10.1017/ Rebollar, E. A., Simonetti, S. J., Shoemaker, W. R., and Harris, R. N. (2016). Direct s002531541000130x and indirect horizontal transmission of the antifungal probiotic bacterium van de Water, J., Chaib De Mares, M., Dixon, G. B., Raina, J. B., Willis, B. L., Bourne, janthinobacterium lividum on green frog (Lithobates clamitans) tadpoles. Appl. D. G., et al. (2018). Antimicrobial and stress responses to increased temperature Environ. Microbiol. 82, 2457–2466. doi: 10.1128/aem.04147-15 and bacterial pathogen challenge in the holobiont of a reef-building coral. Mol. Ringel, Y., Quigley, E. M. M., and Lin, H. C. (2012). Using probiotics in Ecol. 27, 1065–1080. doi: 10.1111/mec.14489 gastrointestinal disorders. Am. J. Gastroenterol. Suppl. 1, 34–40. van Hooidonk, R., Maynard, J., Tamelander, J., Gove, J., Ahmadia, G., Raymundo, Rognes, T., Flouri, T., Nichols, B., Quince, C., and Mahe, F. (2016). VSEARCH: a L., et al. (2016). Local-scale projections of coral reef futures and implications of versatile open source tool for metagenomics. PeerJ 4:e2584. doi: 10.7717/peerj. the Paris agreement. Sci. Rep. 6:39666. 2584 van Oppen, M. J., Gates, R. D., Blackall, L. L., Cantin, N., Chakravarti, L. J., Chan, Rohwer, F., and Kelley, S. (2004). “Culture independent analyses of coral-associated W. Y., et al. (2017). Shifting paradigms in restoration of the world’s coral reefs. microbes,” in Coral Health and Disease, eds E. Rosenberg and Y. Loya Glob. Change Biol. 23, 3437–3448. (New York, NY: Springer-Verlag), 265–277. doi: 10.1007/978-3-662-06414-6_ van Oppen, M. J. H., Mahiny, A., and Done, T. (2005). Geographic patterns of 14 zooxanthella types in three coral species on the Great Barrier Reef sampled after Rohwer, F., Seguritan, V., Azam, F., and Knowlton, N. (2002). Diversity and the 2002 bleaching event. Coral Reefs 24, 482–487. doi: 10.1007/s00338-005- distribution of coral-associated bacteria. Mar. Ecol. Prog. Ser. 243, 1–10. doi: 0487-1 10.3354/meps243001 van Oppen, M. J. H., Oliver, J. K., Putnam, H. M., and Gates, R. D. (2015). Building Rosado, P. M., Leite, D. C. A., Duarte, G. A. S., Chaloub, R. M., Jospin, G., Nunes coral reef resilience through assisted evolution. Proc. Natl. Acad. Sci. U.S.A. 112, da Rocha, U., et al. (2018). Marine probiotics: increasing coral resistance to 2307–2313. doi: 10.1073/pnas.1422301112 bleaching through microbiome manipulation. ISME J. 13, 921–936. doi: 10. Veron, J. E. N. (1995). Corals in Space & Time. The Biogeography and Evolution of 1038/s41396-018-0323-6 the Scleractinia. Ithaca, NY: Cornell University Press. Röthig, T., Ochsenkuhn, M. A., Roik, A., van der Merwe, R., and Voolstra, C. R. Verschuere, L., Rombaut, G., Sorgeloos, P., and Verstraete, W. (2000). Probiotic (2016). Long-term salinity tolerance is accompanied by major restructuring of bacteria as biological control agents in aquaculture. Microbiol. Mol. Biol. Rev. the coral bacterial microbiome. Mol. Ecol. 25, 1308–1323. doi: 10.1111/mec. 64, 655–671. doi: 10.1128/mmbr.64.4.655-671.2000 13567 Wada, N., Pollock, F. J., Willis, B. L., Ainsworth, T., Mano, N., and Bourne, D. G. Röthig, T., Roik, A., Yum, L. K., and Voolstra, C. R. (2017). Distinct bacterial (2016). In situ visualization of bacterial populations in coral tissues: pitfalls and microbiomes associate with the deep-sea coral eguchipsammia fistula from the solutions. PeerJ 4:e2424. doi: 10.7717/peerj.2424 red sea and from aquaria settings. Front. Mar. Sci. 4:259. Ward, S. (2018). How the 2016 Bleaching Altered the Shape of the Northern Salter, S. S., Cox, M. J., Turek, E. M., Calus, S. T., Cookson, W. O., Moffatt, Great Barrier Reef. Available at: http://theconversation.com/how-the-2016- M. F., et al. (2014). Reagent and laboratory contamination can critically impact bleaching-altered-the-shape-of-the-northern-great-barrier-reef-95142 sequence-based microbiome analyses. BMC Biol. 12:87. (accessed 2018). Scheffé, H. (1953). A method for judging all contrasts in the analysis of variance. Webster, N. S., and Reusch, T. (2017). Microbial contributions to the persistence of Biometrika 40, 87–104. coral reefs. ISME J. 11, 2167–2174. doi: 10.1038/ismej.2017.66 Sharp, K. H., Ritchie, K. B., Schupp, P. J., Ritson-Williams, R., and Paul, V. J. (2010). Wickham, H. (2009). ggplot2: Elegant Graphics for Data Analysis. New York, NY: Bacterial acquisition in juveniles of several broadcast spawning coral species. Springer-Verlag. PLoS One 5:e10898. doi: 10.1371/journal.pone.0010898 Wickham, H. (2017). tidyverse: Easily Install and Load the ‘Tidyverse’. R package Singh, J. S., and Singh, D. P. (2013). “Plant growth promoting rhizobacteria version 1.2.1. (PGPR): microbes in sustainable agriculture,” in Management of Microbial Wilkinson, S. P., Brian, J. I., Pontasch, S., Fisher, P. L., and Davy, S. K. (2018). Intra- Resources in the Environment, eds A. Malik, E. Grohmann, and M. Alves genomic variation in Symbiodinium correlates negatively with photosynthetic (Dordrecht: Springer), 361–385. doi: 10.1007/978-94-007-5931-2_14 efficiency and coral host performance. Coral Reefs 37, 691–701. doi: 10.1007/ Spadling, M. D., Ravilious, C., and Green, E. P. (2001). World Atlas of Coral Reefs. s00338-018-1693-y Berkeley, CA:University of California Press. Wilkinson, S. P., Fisher, P. L., van Oppen, M. J. H., and Davy, S. K. (2015). Sunagawa, S., Woodley, C. M., and Medina, M. (2010). Threatened corals provide Intra-genomic variation in symbiotic dinoflagellates: recent divergence or underexplored microbial habitats. PLoS One 5:e9554. doi: 10.1371/journal. recombination between lineages? BMC Evol. Biol. 15:46. pone.0009554 Wilson, K. J., Li, Y., Whan, V., Lehnert, S. A., Byrne, K., Moore, S. S., et Sweet, M., Ramsey, A., and Bulling, M. (2017). Designer reefs and coral probiotics: al. (2002). Genetic mapping of the black tiger shrimp Penaeus monodon great concepts but are they good practice? Biodiversity 18, 19–22. doi: 10.1080/ with Amplified Fragment Length Polymorphisms. Aquaculture 204, 14888386.2017.1307786 297–309. doi: 10.1016/s0044-8486(01)00842-0 Sweet, M. J., and Bulling, M. T. (2017). On the importance of the microbiome and Work, T. M., and Aeby, G. S. (2014). Microbial aggregates within tissues infect a pathobiome in coral health and disease. Front. Mar. Sci. 4:9. diversity of corals throughout the Indo-Pacific. Mar. Ecol. Prog. Ser. 500, 1–9. Sweet, M. J., Croquer, A., and Bythell, J. C. (2010). Bacterial assemblages differ doi: 10.3354/meps10698 between compartments within the coral holobiont. Coral Reefs 30, 39–52. doi: Yeh, Y. C., Needham, D. M., Sieradzki, E. T., and Fuhrman, J. A. (2018). 10.1007/s00338-010-0695-1 Taxon disappearance from microbiome analysis reinforces the value of Frontiers in Microbiology | www.frontiersin.org 17 July 2019 | Volume 10 | Article 1702 Damjanovic et al. Coral Co-culture and Microbiome Manipulation mock communities as a standard in every sequencing run. mSystems 3: Conflict of Interest Statement: The authors declare that the research was e00023-18. conducted in the absence of any commercial or financial relationships that could Zhang, Y., Ling, J., Yang, Q., Wen, C., Yan, Q., Sun, H., et al. (2015). The be construed as a potential conflict of interest. functional gene composition and metabolic potential of coral-associated microbial communities. Sci. Rep. 5:16191. Copyright © 2019 Damjanovic, van Oppen, Menéndez and Blackall. This is an Zhou, G., Cai, L., Yuan, T., Tian, R., Tong, H., Zhang, W., et al. (2017). Microbiome open-access article distributed under the terms of the Creative Commons Attribution dynamics in early life stages of the scleractinian coral Acropora gemmifera in License (CC BY). The use, distribution or reproduction in other forums is permitted, response to elevated pCO2 . Environ. Microbiol. 19, 3342–3352. provided the original author(s) and the copyright owner(s) are credited and that the Ziegler, M., Seneca, F. O., Yum, L. K., Palumbi, S. R., and Voolstra, C. R. (2017). original publication in this journal is cited, in accordance with accepted academic Bacterial community dynamics are linked to patterns of coral heat tolerance. practice. No use, distribution or reproduction is permitted which does not comply Nat. Commun. 8:14213. with these terms. Frontiers in Microbiology | www.frontiersin.org 18 July 2019 | Volume 10 | Article 1702