Specificity of Associations between Bacteria and the Coral Pocillopora meandrina during Early Development

Abstract

Relationships between corals and specific bacterial associates are thought to play an important role in coral health. In this study, the specificity of bacteria associating with the coral Pocillopora meandrina was investigated by exposing coral embryos to various strains of cultured marine bacteria, sterile seawater, or raw seawater and examining the identity, density, and location of incorporated cells. The isolates utilized in this experiment included members of the Roseobacter and SAR11 clades of the Alphaproteobacteria, a Pseudoalteromonas species of the Gammaproteobacteria, and a Synechococcus species of the Cyanobacteria phylum. Based on terminal restriction fragment length polymorphism analysis of small-subunit rRNA genes, similarities in bacterial communities associated with 170-h-old planulae were observed regardless of treatment, suggesting that bacteria may have been externally associated from the outset of the experiment. Microscopic examination of P. meandrina planulae by fluorescence in situ hybridization with bacterial and Roseobacter clade-specific oligonucleotide probes revealed differences in the densities and locations of planulae-associated cells. Planulae exposed to either raw seawater or strains of Pseudoalteromonas and Roseobacter harbored the highest densities of internally associated cells, of which 20 to 100% belonged to the Roseobacter clade. Planulae exposed to sterile seawater or strains of the SAR11 clade and Synechococcus did not show evidence of prominent bacterial associations. Additional analysis of the raw-seawater-exposed planulae via electron microscopy confirmed the presence of internally associated prokaryotic cells, as well as virus-like particles. These results suggest that the availability of specific microorganisms may be an important factor in the establishment of coral-bacterial relationships.

INTRODUCTION

Adiverse assemblage of endosymbiotic dinoflagellates (zooxanthellae), endolithic algae, fungi, bacteria, archaea, and viruses reside within the distinctive skeletal, tissue, and mucus habitats of scleractinian corals (7, 9, 47, 48, 52, 63). This complex community of organisms, referred to as the coral holobiont, is hypothesized to work together to maintain the healthy functioning of individual coral colonies (22, 49). A current challenge in coral biological research is to understand how the different components of the coral holobiont function, both independently and collectively, to create a healthy, functioning colony. The study of microorganisms (defined here as bacteria and archaea) associated with corals has been greatly advanced by the use of cultivation-independent approaches, and numerous studies have now highlighted the abundance and diversity of microbes associated with adult corals (48, 54). These microorganisms are hypothesized to play a role in the transfer of potentially sparse nutrients to the coral organism (31, 62) and the production of probiotic compounds (19, 44).

Cultivation-independent studies are often performed on homogenized coral samples, making it difficult to assign specific habitats to the identified microorganisms within the biochemically diverse mucus, tissue, and skeletal microenvironments of the coral (17, 25). These microhabitats provide unique niches that are undoubtedly favorable to particular microbial functions and, therefore, may harbor specific assemblages of microorganisms. The application of fluorescence in situ hybridization (FISH) techniques to examine the proximity of microorganisms to the coral's cellular architecture provides one means to assess how specific microbial taxa physically associate with the coral and offers a framework for examining taxon-specific microbial interactions within the holobiont. To date, only a few studies have localized individual microbial cells within the coral holobiont using FISH techniques (2–5, 13).

Relationships with specific bacterial taxa have now been demonstrated for several species of coral over various geographic ranges (33, 49). For example, the Gammaproteobacteria species PA1 was shown to be widely distributed with the coral Porites astreoides—a relationship maintained over thousands of kilometers (49). However, the mechanisms that corals utilize to preserve specific associations with microorganisms are not well understood. Many invertebrates maternally transmit microbial symbionts to their offspring in order to ensure the longevity of the association, and this mechanism provides opportunities for coevolution and codiversification of the hosts and their symbionts (41). Horizontal transmission of microorganisms from the environment is also a well-utilized mechanism for acquiring symbionts (20, 39, 45). These relationships are also frequently species specific and require the availability of symbionts from the environment and for both partners to be actively involved in the symbiont incorporation process (39). Additionally, the accessibility of certain symbionts from the environment may be reliant on particular conditions.

The onset of association between microorganisms and corals was recently examined throughout the early developmental cycle of the coral Pocillopora meandrina using a variety of molecular techniques, including FISH (6). In that study, microorganisms were not detected internally within the tissues of eggs or embryos up to 51 h old, but bacteria (primarily belonging to the Roseobacter clade of Alphaproteobacteria) were found within the tissues of planulae (free-swimming larval corals possessing a simplified body plan including a mouth and ectoderm [outer] and endoderm [inner] tissues) that were at least 79 h old. Based on the detection of the specific Roseobacter lineage in the surrounding seawater and observations that the number of Roseobacter cells incorporated into P. meandrina tissues increased with the density of microorganisms they were reared with, it was hypothesized that the bacteria originated from seawater (6).

In an effort to expand our previous observations, we sought to further examine the specificity of bacterial associations with developing Pocillopora meandrina. We exposed developing embryos to dilute seawater cultures of marine microorganisms that were isolated from seawater of a coral reef environment and from an adult coral. A rRNA gene-based approach was utilized in conjunction with fluorescence microscopy to determine whether and how the cultured microorganisms associated with 170-h-old planulae. Additionally, electron microscopy was utilized to examine the specific location of microbial cells in relation to the larval coral tissues. The ability to rear P. meandrina with different strains of microorganisms provides a unique system to examine the specificity and localization of microorganisms incorporated into planular tissues.

MATERIALS AND METHODS

Exposure of P. meandrina embryos and larvae to bacterial cultures.

On 21 April 2008, fragments of healthy P. meandrina colonies were removed from the reef flat in Kaneohe Bay, Oahu, HI (21°27.345′N, 157°46.961′W) and held in seawater aquaria for 2 days prior to spawning on the morning of 23 April 2008. Eggs from 10 colonies were combined, fertilized with sperm for 30 min, and subsequently rinsed using seawater sterilized by the method outlined below. A subsample of 4-h-old embryos was rinsed with sterile seawater and preserved for microscopy by fixation with 3.7% paraformaldehyde in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4) for 6 h at 4°C, followed by five washes in phosphate-buffered saline for 5 min each at room temperature. The embryos were subsequently dehydrated in 100% ethanol and frozen to −20°C.

The remaining embryos were placed in individual 100-mm-diameter petri dishes containing raw seawater, sterile seawater, or individual bacterial strains as treatments (Table 1). The bacterial strains utilized in this study are single-strain laboratory cultures that represent some of the major groups of planktonic bacteria present in Kaneohe Bay seawater (M. S. Rappé and M. L. Brandon, unpublished data), as well as an isolate from an adult coral (Table 1). Water for the sterile-seawater treatment was collected from the surface of Kaneohe Bay 4 days prior to spawning and sterilized by tangential flow filtration (TFF) using a Millipore Pellicon 2 mini-TFF unit equipped with a 30-kDa-cutoff regenerated-cellulose filter cassette (Millipore Corp., Billerica, MA). The sterility of this water was confirmed by microscopy daily. The raw (nonsterile)-seawater treatment was prepared by coarsely filtering Kaneohe Bay surface seawater through a 1.6-μm-pore-sized GF/A filter (Whatman International Ltd., Kent, United Kingdom) in order to exclude zooplankton and larger organisms.

Table 1.

Summary of treatments applied to developing P. meandrina embryos and larvae

Treatment	Source	Cell density (cells ml−1) in medium at:
		Each 24-h addition	72 ha
Roseobacter strain HIMB1	Isolated from surface seawater of Kaneohe Bay, Oahu, HI	1.00 × 104	1.43 × 105
SAR11 strain HIMB4	Isolated from surface seawater of Kaneohe Bay, Oahu, HI	5.00 × 104	4.80 × 105
Synechococcus strain HIMB12	Isolated from surface seawater of Kaneohe Bay, Oahu, HI	5.00 × 104	5.18 × 105
Pseudoalteromonas strain HIMB1276	Isolated from tissues of adult colony of the coral Pocillopora damicornis	5.00 × 104	1.89 × 106
Sterile seawater	Filtered (TFF, 30-kDa cutoff) surface seawater from Kaneohe Bay, HI	<102 (detection limit)	9.00 × 102
Raw seawater	Filtered (1.6-μm pore size) surface seawater from Kaneohe Bay, HI	4.30 × 105	1.00 × 107

^aGrowth determined 72 h after start of experiment but 24 h after addition of a fresh axenic bacterial culture or seawater treatment.

The experimental methods are described in detail in the subsequent paragraphs. Briefly, bacterial treatments were established by growing bacterial strains in liquid batch culture and subsequently incubating diluted (1 ×10⁴ or 5 ×10⁴ bacterial cells ml⁻¹) bacterial cells with rinsed coral planulae. After 24 h of incubation, planulae were removed, rinsed, and incubated with the same bacterial strain, freshly diluted. After repeating the treatment again at 48 h and incubating the planulae for another 24-h period (T = 72 h), free-living bacterial cells were counted via microscopy (described below) and a subsample collected for community structure characterization. At 170 h, planulae were rinsed and collected for community structure characterization and microscopy analysis, and subsamples of free-living bacterial cells were again collected for characterization of bacterial community structure. This characterization was performed in order to assess whether the bacterial community was altered following incubation with coral embryos.

Axenic cultures of each strain were grown in specific media: Roseobacter strain HIMB1 and SAR11 strain HIMB4 were grown in low-nutrient medium (sterile seawater amended with ammonia and phosphate), Synechococcus strain HIMB12 was grown in low-nutrient medium amended with dilute carbon additions (d-glucose, d-ribose, pyruvate, succinate, ethanol, glycerol and N-acetylglucosamine, each at 0.001% final concentration), and Pseudoalteromonas strain HIMB1276 was grown in R2A medium with seawater as the base (Sigma-Aldrich, St. Louis MO). Cells were grown in 12-h light/dark cycles at 30°C. Every 24 h, 30-ml treatments of each bacterial strain were prepared by diluting growing cultures 10× to 1,000× (e.g., adding 1 to 0.004 ml of culture to 30 ml total) into sterile seawater (Table 1). Cellular abundances in axenic bacterial cultures were monitored daily by fixing aliquots (1 ml) with filtered, buffered formaldehyde to a final concentration of ca. 2% for 15 min in the dark at room temperature. The cells were subsequently filtered onto 25-mm-diameter, 0.2-μm polycarbonate membranes (GE Osmonics, Inc., Minnetonka, MN) and stained with the DNA-binding dye DAPI (4′,6′-diamidino-2-phenylindole) at a final concentration of 5 μg ml⁻¹ in 15% formamide solution (0.9 M NaCl, 20 mM Tris-HCl [pH 7.4], 15% formamide, 0.01% sodium dodecyl sulfate [SDS]) for 10 min at room temperature, followed by two 5-min washes in 15% formamide solution. The membranes were air dried and mounted on microscope slides, and cells were enumerated using a Leica DM5000B phase contrast/fluorescence microscope (Leica Microsystems, Wetzlar, Germany) with a 100× objective under UV light.

Planulae were rinsed daily using 40-μm-mesh-size cell strainers (Fisher Scientific, Pittsburg, PA) and sterile seawater and placed into new petri dishes containing fresh treatments. Treatments were initially replicated in triplicate, but after high embryo mortality during the first 24 h, replicates were subsequently combined to ensure that adequate material was available for the duration of the experiment. After 72 h (i.e., following three 24-h treatments), the abundance of microbial cells present in the treatment water exposed to the planulae was determined by microscopic enumeration as described above. Additionally, after 72 and 170 h (i.e., following three and seven 24-h treatments, respectively), subsamples of 30 to 50 ml of treatment water were filtered onto 13-mm-diameter, 0.2-μm-pore-size polyethersulfone membrane filters (Supor 200; Pall Gelman, Inc., Ann Arbor, MI). Filters were stored at −80°C in DNA lysis buffer (20 mM Tris-HCl [pH 8.0], 2 mM EDTA [pH 8.0], 1.2% [vol/vol] Triton X-100) for subsequent DNA extraction (56).

After 170 h, 100 to 200 planulae per treatment were preserved for microscopy and frozen to −20°C as described above, and 50 to 100 planulae per treatment were preserved in 250 μl of lysis buffer and frozen to −80°C for subsequent DNA extraction.

T-RFLP of bacterial SSU rRNA genes.

DNA was extracted using the DNeasy tissue kit (Qiagen, Inc., Valencia, CA) with modifications (8) and quantified using the PicoGreen fluorescent assay (Invitrogen Corp., Carlsbad, CA) on a SpectraMax M2 plate reader (Molecular Devices Corp., Sunnyvale, CA). For terminal restriction fragment length polymorphism (T-RFLP) analysis (34), bacterial small-subunit (SSU) rRNA genes were amplified via PCR using oligonucleotide primers 27F-B-FAM (5′-AGRGTTYGATYMTGGCTCAG-3′) and 519R-VIC (5′-GWATTACCGCGGCKGCTG-3′), with “FAM” and “VIC” indicating 5′-end labeling with FAM or VIC fluorochromes, respectively. Each 50-μl PCR mixture contained 2 U of Sahara enzyme (Bioline USA, Inc., Taunton, MA), 1× Sahara reaction buffer, 2 mM Sahara MgCl₂, 200 μM each deoxynucleoside triphosphates (dNTPs), 200 nM each primer, and 100 ng of template genomic DNA (up to 1 μg for samples that did not amplify at lower concentrations). After an initial denaturation step at 95°C for 5 min, the reaction conditions were as follows: 29 cycles of 95°C denaturation for 30 s, 55°C annealing for 1 min, and 72°C extension for 2 min, concluding with an extension at 72°C for 20 min. The reactions were performed in a MyCycler personal thermal cycler (Bio-Rad Laboratories, Hercules, CA). Products were purified using the QIAquick PCR purification kit (Qiagen, Inc.) and subsequently restricted in a 10-μl reaction mixture containing 100 ng of purified amplification product, 2 μg of bovine serum albumin (BSA), 1× enzymatic reaction buffer, and 5 units of HaeIII restriction endonuclease (10 units μl⁻¹; Promega, Madison, WI) for 7 h at 37°C. Restriction digests were purified using the QIAquick nucleotide removal kit (Qiagen, Inc.), and 30 ng μl⁻¹ of each product was subsequently electrophoresed on an ABI 3100 genetic analyzer (Applied Biosystems, Foster City, CA). Operational taxonomic units (OTUs) were identified as individual T-RF peaks from both the forward and reverse labeled primer for fragments between 33 and 550 bp in length. To account for small differences in the amount of DNA loaded on the ABI 3100, data were normalized by excluding peaks that contributed <0.05% of the total peak area for each sample (46).

FISH.

Planulae were examined using fluorescence in situ hybridization (FISH) with an oligonucleotide probe suite targeting the SSU rRNA of all bacteria (40) and an oligonucleotide probe targeting the SSU rRNA of members of the Roseobacter clade (10). In order to prepare samples for hybridization, planulae were fixed and subsequently permeabilized and secondarily fixed for hybridization as previously described (6). For hybridization reactions using the general bacterial probe suite, stored samples were first washed for 10 min at room temperature in 15% formamide hybridization solution (0.9 M NaCl, 20 mM Tris-HCl [pH 7.4], 15% formamide, 0.01% SDS), followed by incubation in fresh formamide hybridization solution for 30 min at 37°C. A suite of five Cy3-labeled probes (EUB-27R, EUB-338Rpl, EUB-700R, EUB-700Ral, and EUB-1522R) (40) were each added at 2 ng μl⁻¹ in 15% formamide hybridization solution and hybridized to the fixed planulae at 37°C for 14 h in the dark. Negative-control samples were incubated with nonsense oligonucleotide 338F-Cy3 (40), added at 2 and 10 ng μl⁻¹, as well as a no-probe control, under the same conditions. After hybridization, the samples were washed two times for 10 min at 50°C in 0.15 M NaCl hybridization wash (150 mM NaCl, 20 mM Tris-HCl [pH 7.4], 6 mM EDTA, 0.01% SDS). Planulae were also examined for the presence of members of the Roseobacter clade using the Cy3-labeled oligonucleotide probe Roseo536R (10). Sample preparation and hybridization reactions were performed as described above with the following exceptions: hybridizations were performed in a 35% formamide hybridization solution (0.9 M NaCl, 20 mM Tris-HCl [pH 7.4], 35% formamide, 0.01% SDS) at 37°C, and the posthybridization washes were performed at 52°C in a 0.07 M NaCl hybridization wash (70 mM NaCl, 20 mM Tris-HCl [pH 7.4], 5 mM EDTA, 0.01% SDS). Control samples lacking oligonucleotide probe were prepared using the same conditions.

After being washed, samples were dehydrated in 100% isopropanol for 5 min at room temperature and placed in a solution containing one part benzyl alcohol to two parts benzyl benzoate. The planulae were mounted on slides in this solution. Sample visualization was performed with a Zeiss LSM 510 confocal laser scanning microscope (Carl Zeiss, Inc., Jena, Germany) and a 63× objective. Ten specimens were examined from each treatment, and representative specimens were imaged into optical stacks and processed using Zeiss LSM Image Browser version 4.2.0.121 (Carl Zeiss, Inc.). The maximal densities of probe-hybridized cells were computed from images by counting the number of cells visible in a 10- by 10- by 1-μm area in the x, y, and z dimensions, respectively. Due to rapid fading of probe-conferred fluorescence, only small regions were targeted for counting. Multiple slices within an optical stack were counted if they were at least 2 μm apart. Bacterial cells were considered internalized within the coral if surrounded by the autofluorescent coral tissues under magnification.

Postprocessing of confocal stack images was conducted using Volocity software (Improvision, Coventry, United Kingdom). An intensity rendering was uniformly applied to all images.

Electron microscopy.

One hundred 70-h-old planulae exposed to the raw-seawater treatment were analyzed for the presence of microbial cells via transmission electron microscopy. Samples were fixed in 2.5% glutaraldehyde in 0.1 M sodium cacodylate with 0.32 M sucrose (pH 7.4) for 48 h at 4°C, washed with 0.1 M sodium cacodylate with 0.44 M sucrose, and postfixed with 1% osmium tetroxide in 0.1 M sodium cacodylate. After dehydration in 100% ethanol, samples were embedded in LX-112 epoxy resin (Ladd Research Industries, Burlington, VT). Ultrathin sections (60 to 80 nm) were cut on an Ultracut E ultramicrotome (Reichert-Jung/Leica Microsystems, Wetzlar, Germany), mounted onto copper grids, and viewed at 100 kV accelerating voltage on a LEO 912 energy-filtered transmission electron microscope (Zeiss, Oberkochen, Germany). Images were taken with a slow-scan, fast-transfer, charge-coupled-device camera (ProScan, Impala Park, South Africa).

RESULTS

Microbes in treatment media.

An increase in free-living microbial cells was observed in all treatments during exposure to developing P. meandrina (Table 1). After 72 h of exposure to the developing coral, (but only 24 h after the addition of a fresh strain), cultured strain treatments primarily contained cell morphologies consistent with those of the respective treatment microorganisms; however, other cell morphologies were also observed via microscopy (data not shown). After 72 h, sterile-seawater treatments also contained free-living microbial cells possessing a variety of cellular morphologies (Table 1).

Bacterial community structure analysis.

Using SSU rRNA gene T-RFLP, a comparison of bacterial communities associated with 170-h-old P. meandrina planulae incubated with different bacterial strains revealed that the planulae were not solely associated with the specific forward and reverse T-RF pair of the bacterial strain they were exposed to but were instead associated with a mixed bacterial community (Fig. 1). Bacterial communities associated with planulae exposed to axenic strains of Roseobacter clade strain HIMB1 and Pseudoalteromonas strain HIMB1276 possessed the signature T-RF patterns of each isolate (Fig. 1), strongly suggesting that these bacteria physically associated with or attached to the planulae. However, the planulae-associated bacterial communities resulting from these two treatments also contained several other T-RFs that did not originate from the axenic bacterial strains used as treatments (Fig. 1). In contrast, the bacterial communities associated with planulae exposed to axenic strains of SAR11 strain HIMB4 and Synechococcus strain HIMB12 did not possess the T-RFs characteristic of their respective treatment strains (limit of detection, >0.5% relative abundance), indicating that these isolates did not physically associate with P. meandrina planulae (Fig. 1). Instead, the planulae-associated bacterial communities resulting from these two treatments were comprised of several different T-RFs (Fig. 1).

Fig 1.

Heat map depicting the relative abundance of bacterial T-RFs from 170-h-old P. meandrina planulae and the medium treatment water after two different 24-h bacterial culture exposures (72 and 170 h). Percentages in the T-RFLP profiles, shown in the color key, are from combined forward and reverse profiles. Also shown are the T-RFLP profiles for all bacterial culture and seawater treatments at T = 0.

Prior to the addition of planulae, the sterile-seawater treatment alone failed to yield an SSU rRNA gene PCR amplicon. However, the 170-h-old planulae reared in sterile seawater were associated with several T-RFs, including F-34, F-191, and R-91 (where F and R indicate forward and reverse and the number is the fragment length in base pairs) (Fig. 1). Planulae reared in raw seawater were associated with the T-RFs F-191, R-91, and R-144 (among others), which were not abundant members of the raw-seawater treatment prior to the addition of planulae (Fig. 1).

Some bacterial community members were found to be ubiquitously associated with 170-h-old planulae regardless of treatment (Fig. 1). In particular, bacteria represented by T-RFs F-34, F-191, F-225, R-91, and R-115 were shared between the planulae from all treatments. These T-RF signatures include cells previously identified as belonging to the Roseobacter clade of Alphaproteobacteria (F-34 and R-91) and Gammaproteobacteria Pseudoalteromonas spp. (F-34 and R-115) (6). Planktonic microbial communities growing in the media collected following 72 or 170 h of planulae exposure also harbored many of these T-RFs (Fig. 1).

Localization of Bacteria and Roseobacter clade cells.

Irrespective of treatment, bacterial cells were found to be associated with 60% of the P. meandrina planulae when visualized with the bacterial probe suite via FISH after 170 h of treatment exposure. Planulae exposed to Roseobacter strain HIMB1, Pseudoalteromonas strain HIMB1276, and raw seawater possessed the most consistent associations and the highest densities of bacteria (Table 2 and Fig. 2). Planulae exposed to Pseudoalteromonas strain HIMB1276 harbored bacterial cells that were primarily localized to the external edge and outer ectoderm (Fig. 2c). Bacterial cells were also associated with the ectodermal tissues of planulae exposed to Roseobacter strain HIMB1 (Fig. 2a) and with planulae held in the raw-seawater treatments (Fig. 2e). Only 2/10 of planulae reared in the presence of Synechococcus strain HIMB12, SAR11 strain HIMB4, or the sterile-water treatment contained visibly associated bacterial cells; those that did harbored a low density of cells in ectodermal tissues (Table 2; also see Fig. S1 in the supplemental material).

Table 2.

Summary of FISH results on 170-h-old P. meandrina planulae

Treatment	Bacterial probe suite			Roseobacter clade probe
	Probe-positive planulaea	Locationb	Maximum density (cells μm−3)c	Probe-positive planulaea	Locationb	Maximum density (cells μm−3)c
Roseobacter strain HIMB1	10/10	Ectoderm	0.06–0.10	10/10	Ectoderm	0.02–0.05
SAR11 strain HIMB4	2/10	Ectoderm	0–0.03	2/10	Ectoderm	0.01–0.02
Synechococcus strain HIMB12	2/10	Ectoderm	0.02–0.03	2/10	Ectoderm	0–0.01
Pseudoalteromonas strain HIMB1276	10/10	External, ectoderm	0.04–0.07	2/10	Ectoderm	0.01–0.02
Sterile seawater	2/10	Ectoderm	0–0.03	1/10	Ectoderm	0–0.04
Raw seawater	10/10	Ectoderm, endoderm	0.04–0.06	10/10	Ectoderm	0.04–0.05

^aProportion of planulae that positively hybridized with each probe assay out of the total number inspected.

^bMost commonly observed location of positively hybridized cells within the planulae.

^cMaximum density of positively hybridized cells within the planulae.

Fig 2.

Confocal epifluorescence micrographs of 170-h-old P. meandrina planulae reared in the presence of Roseobacter clade strain HIMB1 (a, b), Pseudoalteromonas strain HIMB1276 (c, d), and raw seawater (e, f). Bacterial cells were observed in planulae hybridized with the bacterial probe suite (indicated by arrows in panels a, c, and e) but not in control samples exposed to a nonsense control probe (b, d, f). Coral tissues and zooxanthellae (representative cells indicated with “z”) are illuminated due to autofluorescence in all images. Gland cells and/or nematocysts (g/n) are prominent in samples due to nonspecific hybridization. The ectodermal (ec) and endodermal (en) differentiation is drawn for reference.

Hybridization experiments with a probe specific for the Roseobacter clade of Alphaproteobacteria revealed that cells of this phylogenetic lineage appeared to be most consistently and abundantly associated with 170-h-old P. meandrina planulae exposed to either Roseobacter strain HIMB1 or raw seawater (Table 2 and Fig. 3). Planulae exposed to SAR11 strain HIMB4, Synechococcus strain HIMB12, and Pseudoalteromonas strain HIMB1276 were occasionally (2 of 10 planula each) associated with a low density of Roseobacter clade cells, which were located in the ectoderm (Table 2; also see Fig. S1 in the supplemental material). Nine of 10 planulae exposed to the sterile-seawater treatments contained no visible Roseobacter clade cells (Table 2).

Fig 3.

Confocal epifluorescence micrographs of 170-h-old P. meandrina planulae hybridized with a probe specific for cells of the Roseobacter clade. Planulae reared in the presence of Roseobacter clade strain HIMB1 (a) and raw seawater (b) are shown. Cells of the Roseobacter clade are indicated with arrows. The line differentiates the ectodermal (ec) and endodermal (en) tissues.

No internally or externally associated microbial cells could be detected in FISH assays targeting all Bacteria and members of the Roseobacter clade in 4-h-old P. meandrina embryos (see Fig. S2 in the supplemental material).

Detection of planula-associated microorganisms in electron micrographs.

Planulae exposed to raw seawater as the treatment were examined via transmission electron microscopy, which revealed spherical cells in the outer ectodermal tissue and surrounding the mucus glands that ranged in diameter from 0.4 to 0.7 μm and were consistent with the size and shape of prokaryotic cells (Fig. 4a to c). In addition, aggregations of virus-like particles (VLPs) were also detected in the outer ectoderm. The VLPs ranged in size from 40 to 70 nm in diameter (Fig. 4d to f). While the prokaryotic cells and VLPs were both located in the ectoderm, they were not in close proximity to one another and, thus, did not appear specifically associated.

Fig 4.

Transmission electron micrographs of prokaryotic cells and virus-like particles (VLPs) within the ectodermal tissues of 170-h-old P. meandrina planulae. (a to c) Single prokaryotic cells sized 0.4 to 0.7 μm in diameter are located in the ectodermal tissue near the exterior (ex) edge of the planulae. (d) A cluster of VLPs (each VLP sized 40 to 70 nm) near the exterior edge. (e) VLPs near a nucleus (n). (f) Multiple clusters of VLPs near the exterior. Scale bars are 500 nm for panels a to d and 1 μm for panels e and f.

DISCUSSION

Specificity in developing coral-bacterial relationships.

The results of this study provide experimental evidence supporting the hypothesis that taxon specificity plays a role in the onset of relationships between bacteria and developing corals. Differences were visible in the location and density of bacterial cells incorporated into P. meandrina planulae that were exposed to bacteria that differ broadly in phylogenetic affiliation. Incubation with raw seawater, Roseobacter clade strain HIMB1, and Pseudoalteromonas strain HIMB1276 resulted in planulae containing prominent, internally associated microbial cells, while planulae incubated with sterile seawater, Synechococcus strain HIMB12, and SAR11 strain HIMB4 did not. These results suggest that the establishment of this intimate association is not passive and that the planula, bacterium, or both play an active role. Specific host-microbial relationships have been documented for a variety of organisms (38, 39), including associations between corals and Symbiodinium (1, 29, 50, 53). Corals possess an innate immune system which functions primarily in self- or nonself-recognition (36). Thus, some mechanism of recognition must exist in order to determine which bacterial cells can cross the coral epithelium. Proteins involved in pattern recognition may play a role in facilitating specific bacterial cells to cross into the epithelium of P. meandrina (26, 27), and further investigation of these mechanisms is important for understanding the initiation of any coral-microbial relationship.

Planulae exposed to Roseobacter clade strain HIMB1 formed the most prominent internal bacterial associations, which was somewhat surprising due to the planktonic lifestyle of this isolate. However, marine roseobacters represent a diverse group of bacteria that possess a variety of physiologies and have been previously found in internal associations with developing P. meandrina (6) and Porites astreoides (51). Members of the Roseobacter clade also frequently form relationships with adult corals (14, 16, 21, 23, 24, 30, 33, 48, 54, 55), but it is unclear how this specificity translates into homogeneous or heterogeneous roles within the coral holobiont. Members of this important group of marine bacteria are known to perform a variety of distinct functions (12, 37), including the consumption of the organic sulfur metabolite dimethylsulfoniopropionate, which is abundantly secreted by coral zooxanthellae (11, 18), and the production of probiotic compounds (33, 49).

Pseudoalteromonas strain HIMB1276 also formed a prominent association with P. meandrina planulae, although its T-RF pattern was never identified in planulae incubated in raw seawater. Interestingly, it was isolated from the coral Pocillopora damicornis, which belongs to the same genus of coral as utilized in this study. Thus, this isolate appears to harbor a host-associated lifestyle. Planulae exposed to Pseudoalteromonas strain HIMB1276 primarily harbored cells in the outermost edge of the ectoderm, and it is possible that these cells played a direct or indirect role in the settlement or adhesion processes of the planulae. Microbial biofilms on benthic substrates can attract and/or promote settlement of coral larvae, presumably through the production of chemical signals (42, 60, 61). In fact, a strain of Pseudoalteromonas appears to be involved in the metamorphosis of an acroporid coral (42). During the course of our experiment, planulae exposed to Pseudoalteromonas strain HIMB1276 tended to settle onto the surface of the petri dish (data not shown). Future studies are necessary to investigate the specific role of Pseudoalteromonas strain HIMB1276 and other members of the Pseudoalteromonas genus with developing corals and whether or not these associations persist in adults. Investigations examining their role in the settlement process, such as their ability to produce or detect chemical signals, may be particularly worthwhile.

The SAR11 strain HIMB4 and Synechococcus strain HIMB12 formed less prominent internal associations with P. meandrina planulae and probably possess properties that are not suitable for a host environment. For example, the SAR11 strain HIMB4 is a planktonic organism which is particularly prominent in oligotrophic waters (40). In contrast, Synechococcus is a diverse group of bacteria containing species with planktonic and host-associated lifestyles. Cyanobacteria, but not specifically Synechococcus, have been known to internally associate with corals (28, 32). The Synechococcus strain utilized in this study (HIMB12) is planktonic, was isolated from a coastal ecosystem, and is not known to be host associated.

Widespread Roseobacter associations with developing P. meandrina organisms.

After 170 h of development, P. meandrina planulae exposed to all treatments were associated with cells of the Roseobacter clade, suggesting that roseobacters may be particularly adept at forming associations with developing P. meandrina organisms. Because the associations were more prominent when roseobacters were concentrated in the seawater or media, it is possible that quorum sensing via microbe-secreted autoinducer molecules plays a role in the establishment of these relationships. This mechanism is utilized by microbes in other host associations, including pathogens (43, 65). The origin of the Roseobacter clade cells in other treatments besides that of Roseobacter strain HIMB1 is not yet known but is of considerable interest. These cells potentially originate from surface (mucus) associations with the embryos and are later incorporated into the ectodermal tissues during development. While Roseobacter and other bacterial cells were not observed to be externally or internally associated with 4-h embryos, it is possible that external associations may be disrupted while preparing samples for FISH. In addition, it is also feasible that the Roseobacter clade cells originated from the free-living bacterioplankton community and were not removed during the rinsing process prior to establishing the treatments. Despite daily water changes of sterile-seawater medium, the sterile-seawater treatment showed evidence of a small community of microbial cells after 72 h (with sterile water changes every 24 h) of exposure to the planulae. Both of these scenarios may explain why, following planula exposure, the free-living bacterial communities often contained multiple community members that were similar despite differences in the treatments. The planulae were not treated with antibiotics prior to incubation with the various microbial strains and seawater treatments, which might have been useful to remove loosely associated bacterial cells and obtain barren embryos.

Roseobacters were highly associated with the raw-seawater-exposed P. meandrina planulae in a previous study, but the cells in that study were primarily members of the Jannaschia genus (6). While the SSU rRNA gene T-RF pattern representing the genus Jannaschia (F-191 and R-178) was detected in the planulae reared with Roseobacter clade strain HIMB1 and Pseudoalteromonas strain HIMB1276, it was not detected within the bacterial community associated with planulae reared in the raw-seawater treatment. However, the T-RF pattern broadly associated with the majority of members of the Roseobacter clade (F-34 and R-91/-92) was prominently associated. Thus, this study provides support for the hypothesis that members of the Roseobacter clade are important for developing corals. Because the T-RFLP analysis method used here provides relatively low taxonomic resolution and is not able to differentiate between most subclades within the Roseobacter clade, additional analysis of the identity of the specific community members is needed to further examine the specificity of the Roseobacter clade cells associating with P. meandrina planulae. While it should be noted that the potential for cross-contamination was present due to the employment of a member of the Roseobacter clade possessing the T-RF pattern F-34 and R-91 as a treatment (HIMB1), this scenario is highly unlikely because the experimental setup limited the opportunity for contaminating exposure.

Virus-like particles associated with treatment-exposed planulae.

The electron micrographs provide further evidence for prokaryotic cells in P. meandrina planulae and present the first observations of VLPs associated with coral planulae. The planulae-associated VLPs were similar in size and shape to those uncovered in the tissues of healthy adult corals (47). Metagenomic analyses of corals have also validated the association of phage and eukaryotic viruses with healthy corals (35, 59, 62). Viruses are thought to play a role in disease manifestation (15, 57, 64), but they may also benefit corals (58). The role of viruses in healthy adult and developing corals requires further research to better understand their potential beneficial effects or threats for the coral holobiont.

Summary.

In this study, we found that the taxonomic composition of free-living, planktonic bacteria influences the frequency and nature of bacterial cell incorporation by P. meandrina planulae. We provide additional evidence that the Roseobacter clade of marine Alphaproteobacteria appear particularly adept at forming intimate relationships with developing P. meandrina, making it important to understand how these particular bacteria influence the development, settlement, or health of the coral. Field studies investigating the nature of microorganisms associating with developing corals will undoubtedly be useful to verify laboratory-based results. Additionally, assessing the persistence of association between microorganisms and developing corals with time is necessary to address whether the relationship is opportunistic or more intimate in nature. Finally, we anticipate that continued investigation of how developing P. meandrina organisms respond to altered communities of bacterioplankton will lead to important insights regarding the ability of corals to develop, metamorphose, and endure as adult colonies in a changing marine environment.