Diversity of culturable bacteria in the mucus of the Red Sea coral Fungia scutaria

Authors


  • Editor: Ralf Conrad

Correspondence: Yael Lampert, Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel. Tel.:+972 3 6357355; fax:+972 3 5352355; e-mail: lampery@mail.biu.ac.il

Abstract

Coral reefs are the most biodiverse of all marine ecosystems. Bacteria are known to be abundant and active in seawater around corals, inside coral tissues, and within their surface microlayer. Very little is known, however, about the structure, composition and maintenance of these bacterial communities. In the current study we characterize the culturable bacterial community within the mucus of healthy specimens of the Red Sea solitary coral Fungia scutaria. This was achieved using culture-based methods and molecular techniques for the identification of the bacterial isolates. More than 30% of the isolated bacteria were novel species and a new genus. The culturable heterotrophic bacterial community of the mucus of this coral is composed mainly of the bacterial groups Gammaproteobacteria, Alphaproteobacteria and of Actinobacteria. This study provides the first evidence of actinomycetes isolated from corals.

Introduction

Coral reefs are the most biodiverse of all marine ecosystems; however, very little is known about the prokaryotic diversity in these systems. Our understanding of the role of bacteria in coral reef ecosystems is still evolving.

Mucus is a characteristic product of all corals, forming a coating over their polyps (Meikle et al., 1988). The mucus that is excreted by ectodermal unicellular glands forms, under ordinary circumstances, a thin layer over the external surface of the polyps. When the polyps are stressed, however, they excrete enormous quantities of mucus, giving an almost gelatinous consistency to the water around them (Duerden, 1906). Several studies have dealt with the composition of coral mucus (Coles & Strathmann, 1973; Benson & Muscatine, 1974; Ducklow & Mitchell, 1979a; Meikle et al., 1988). In Fungia scutaria, the composition of the mucus was found to be low in caloric value, and to contain small amounts of lipids, nitrogen, phosphorus, sulphur and amino acids (Krupp, 1981, 1985). The role of mucus produced by corals has been known for many years (Herndl & Velimirov, 1986). Several physiological functions essential to coral growth have been suggested, including a protective mechanism against sedimentation by cleansing processes (Duerden, 1906; Johannes, 1967; Schuhmacher, 1977; Coffroth, 1985; Meikle et al., 1988); a feeding mechanism (Duerden, 1906; Lewis & Price, 1976; Edwards & Davies, 2002; Kushmaro & Kramarsky-Winter, 2004); protection against desiccation (Meikle et al., 1988); a protective layer against changes in temperature and salinity in the water surrounding the coral (Marcus & Thorhaug, 1982; Coffroth, 1985; Kato, 1987); and a shield from UV radiation (Drollet et al., 1993; Lyons et al., 1998). In addition, mucus production may confer an ecological advantage to the coral in retarding epiphytic and epizoic growth (Herndl & Velimirov, 1986), and regulating surface bacterial growth by the self-cleaning behaviour of the host (Ducklow & Mitchell, 1979b).

Bacteria are known to be abundant and active in seawater around corals, in coral tissues, and within their surface microlayer. Very little is known, however, about the structure, composition and maintenance of these bacterial communities. An investigation of the bacterial populations in mucus layers of several species of corals revealed that the bacterium Vibrio alginolyticus is a member of the microbial populations capable of utilizing compounds in the mucus (Ducklow & Mitchell, 1979a). It was suggested that this bacterium may play a role in processing coral mucus for reef detritus feeders. Another study of the production and bacterial colonization of mucus from the soft coral Briarium asbestinum suggested that mucus production and bacterial populations are enhanced with increasing turbidity (Rublee et al., 1980). Investigation of the bacterial utilization of mucus on the brain coral Platygyra lamellina from Aqaba, northern Red Sea, revealed seven genera of bacterial isolates (Pascal & Vacelet, 1981). In this study, Gram-negative strains were more common (73%) than the Gram-positive strains (27%). A comparison of the relative activities of microbial populations of the coral surface microlayer and their overlying seawater determined that coral-associated bacteria grow approximately 10 times faster than those in the overlying water column (Paul et al., 1986). In another investigation of the composition of the bacterial community in the mucus of four species of hard corals from Aqaba, northern Red Sea, seven genera of bacteria were isolated (Wahbeh & Mahasneh, 1988). The isolates included Gram-negative and Gram-positive strains, with Micrococcus sp. found to be the most common. An assessment of the microbial utilization of mucus produced by two hard corals in New Caledonia suggested that only a fraction of the mucus includes proteins and lipids is actually utilized by the bacteria (Vacelet & Thomassin, 1991). Another study found nitrogen-fixing bacteria in the skeleton of the coral Favia favus (Shashar et al., 1994). It was suggested that the interaction between the nitrogen-fixing bacteria and the coral may be of major importance to the nitrogen budget of corals. A study of the carbon-source utilization patterns of the bacterial communities of coral surface microlayers found that bacterial communities associated with corals are largely coral-species-specific (Ritchie & Smith, 1995; Ritchie & Smith, 1997; Ritchie & Smith, 2004). In addition, the bacterial communities appear to reflect the phylogenetic grouping of the corals at the genus level, since bacterial communities are more similar within genera.

All the above-mentioned studies used phenotypic and biochemical characterization to identify the bacterial strains isolated in the coral mucus. Recently, with the advent of molecular methods, 16S rRNA gene sequencing has become the primary tool used to identify bacteria isolated from nature. The first use of such methods in coral-associated bacteria characterized the prokaryotic microbiota associated with the Caribbean coral Montastraea franksi (Rohwer et al., 2001). The cultured bacteria were found to be closely related to previously described bacteria and consisted mostly of representatives of the Gammaproteobacteria. However, the culturable bacteria represent only a small fraction of the coral microbiota.

During the last three decades, most of the studies on coral-associated bacteria have linked the presence of bacteria in corals to several diseases (Mitchell & Chet, 1975; Antonius, 1985; Edmunds, 1991; Kushmaro et al., 1996; Kushmaro et al., 1997; Santavy & Peters, 1997; Ben-Haim et al., 2003; Sutherland et al., 2004). In the current study, we have focused on the culturable natural bacterial community within the mucus of healthy specimens of the Red Sea solitary coral Fungia scutaria. Our primary goal was to characterize the culturable natural bacterial community using culture-based methods and molecular techniques for identifying these isolates.

Materials and methods

Sample collection

Mucus samples of three individuals of the coral Fungia scutaria were collected in situ at 1–2 m depth during June and July 2002 from the coral reef of the Gulf of Eilat, northern Red Sea. The samples were collected in front of the Inter University Institute (IUI) of marine biology at Eilat. The sampled corals were separated by distances of at least 5 m to minimize the likelihood of sampling from single coral clones. All corals appeared to be healthy at the time of sampling. The mucus samples were collected with sterile swabs that were placed in 15 mL sterile tubes. The sterile tubes were inverted and opened underwater adjacent to the coral surface to minimize the amount of seawater entering the tubes. Mucus samples of c. 1 cm2 coral surface area were taken with these swabs, and placed back inside the empty tubes. The tubes were closed tightly underwater. Seawater samples were collected with 50 mL sterile tubes that were opened underwater adjacent to the same corals. All samples were transferred to the laboratory for further work.

Bacterial isolation and enumeration

The mucus swab samples were transferred to sterile tubes with 1 mL autoclave-sterilized artificial seawater (ASW), in a sterile hood. The bacteria were suspended in ASW by vigorous vortexing for 3 min. Representatives of each colony morphotype were isolated using standard serial dilution and plating techniques in triplicate on Bacto Marine Agar 2216 (Difco BD Biosciences, Franklin Lakes, NJ), a medium designed for the isolation and enumeration of heterotrophic marine bacteria. The medium was prepared according to the manufacturer's procedure with the salt concentration adjusted to 3.6%, corresponding to northern Red Sea seawater salinity. All plates were incubated at a temperature of 25°C, corresponding the ambient seawater temperature, for 3–7 days. The pure bacterial strains were cultivated at 25°C, and glycerol stocks were prepared in 30% glycerol in Marine Broth and maintained at −70°C.

Bacterial counts that represent the number of culturable bacteria were recorded as CFUs from each swab and expressed as CFU per 1 cm2 of coral surface area or 1 mL of seawater sample. Total bacterial counts were determined by epifluorescent microscopy as described previously (Porter & Feig, 1980). Briefly, cell suspensions from the coral mucus or seawater were preserved in a final concentration of 2% filter-sterilized formaldehyde and stored in the dark at 4°C until processing. Tween 80 (10 μg mL−1) was added to each sample and samples were sonicated for 30 s to minimize bacterial aggregates (Kepner & Pratt, 1994). The samples were then stained with 4′6-diamidino-2-phenylindole (DAPI, Sigma), filtered onto 25 mm black polycarbonate 0.22 μm filters (Poretics, GE Osmonics Inc., Trevose, PA), and enumerated by fluorescent microscopy. Slides were kept at −20°C until they were counted. Approximately 40 random fields per slide were counted in an Olympus epifluorescence microscope, and the number of bacteria per coral surface area was calculated.

Bacterial identification by 16S rRNA gene sequence analysis

The bacterial strains were identified by 16S rRNA gene sequence analysis. The bacterial isolates were grown overnight in 2 mL Marine Broth. Total DNA was extracted with the UltraClean Microbial DNA Isolation Kit (Mo Bio Laboratories Inc., Carlsbad, CA) using the manufacturer's procedure. Eubacterial-specific primers [forward primer 8-27: 5′-AGAGTTTGATCCTGGCTCAG-3′ (Weisburg et al., 1991) and reverse primer 1492: 5′-GGTTACCTTGTTACGACTT-3′ (Reysenbach et al., 1992)] were used to amplify 16S rRNA genes. PCR fragments were purified using a QIAquick Gel Extraction Kit (Qiagen, Valencia, CA), and sequenced on an ABI 377 automated sequencer using the PRISM Ready Reaction Kit (Applied BioSystems, Foster City, CA). Sequence data were analysed by comparison with 16S rRNA genes in the GenBank database. The nearest relatives of each organism were obtained by BLAST searches (Altschul et al., 1990). The sequences were manually compiled and aligned to Escherichia coli using the PHYDIT software (Chun, 2001). Phylogenetic trees were generated using the neighbour-joining (Saitou & Nei, 1987), Fitch–Margoliash (Fitch & Margoliash, 1967), and maximum parsimony (Kluge & Farris, 1969) algorithms in the PHYLIP package (Felsenstein, 1993). Evolutionary distance matrices for the neighbour-joining and Fitch–Margoliash methods (Jukes & Cantor, 1969) were generated. Tree topologies were evaluated after 1000 bootstrap re-samplings of the neighbour-joining data.

Bacterial characterization

The bacterial strains, excluding the Actinobacteria, were characterized by 22 biochemical tests (oxidase, indole production, reduction of nitrates to nitrites, reduction of nitrates to nitrogen, acidification of glucose, arginine dihydrolyase, urease, esculine hydrolysis, protease, β-galactosidase, and assimilation of glucose, arabinose, mannose, mannitol, N-acetyl-glucosamine, maltose, gluconate, caprate, adipate, malate, citrate, phenyl-acetate) using api-20 NE (micromethod tests for the identification of non-enteric Gram-negative bacteria; bioMérieux SA, Marcy-l'Étoile, France). For the Actinobacteria we used the api-20 Coryne (micromethod tests for the identification of Actinobacteria; bioMérieux SA). This kit characterizes 21 biochemical tests (oxidase, nitrate reduction, pyrazinamidase, pyrrolidonyl arylamidase, alkaline phosphatase, β-glucuronidase, β-galactosidase, α-glucosidase, N-acetyl-β-glucosaminidase, esculine hydrolysis, urease, gelatine hydrolysis, catalase, and fermentation of glucose, ribose, xylose, mannitol, maltose, lactose, sucrose, glycogen). The standard api-20NE and api-20 coryne methods were used with the media adjusted to 3.6% NaCl.

Sensitivity to antibiotics (10 U penicillin-G, 30 μg chloramphenicol, 30 μg novobiocin, 30 μg tetracycline, 300 units polymixin B and 2 μg clindamycin), each applied to a paper disc, was determined after incubation for 24–48 h at 30°C on Marine Agar.

Results

Isolation and enumeration of bacteria

The number of bacteria in the mucus of the Red Sea stony coral F. scutaria and in adjacent seawater was investigated by two methods. The first, enumeration of culturable bacteria is presented as CFUs per swab that sampled c.1 cm2 of coral surface area, or CFUs 1 mL−1 of seawater. The mean number of culturable bacteria in the mucus of F. scutaria was 4.8 × 103±2.2 × 103 CFU cm−2. In comparison, the mean number of culturable bacteria in seawater adjacent to the corals was 5.9 × 102±3.1 × 102 CFU mL−1. The second method enumerated the total number of bacteria by epifluorescence microscopy, which includes the culturable and the unculturable bacterial consortium measured as number of cells per swab that sampled approximately 1 cm2 of coral surface area or number of cells 1 mL−1 of seawater. The mean number of total bacteria in the mucus of F. scutaria was 1.4 × 106±3.1 × 105 cells cm−2. In comparison, the mean number of total bacteria in seawater adjacent to the corals was 1.9 × 105±3.1 × 104 cells mL−1.

Twenty-two bacterial strains were isolated from the solitary coral F. scutaria. Identification of the strains by 16S rRNA gene sequence analysis revealed that 35% belonged to the Gammaproteobacteria, 32% to the Alphaproteobacteria, 23% to the Actinobateria, 5% to the Bacillales and 5% to the Cytophaga–Flavobacter group (Table 1). Most of the isolates were closely related to previously described bacterial species, with an average identity of 97% over the first 550 bp of the 16S rRNA gene.

Table 1.   16S ribosomal RNA gene sequencing analysis of bacteria isolated from the Red Sea coral Fungia scutaria, based on BLAST analysis
Bacterial strainGenBank accession numberMost closely related hit in GenBankIdentities over
550 bp* (%)
Bacterial group
  • *

    bp, number of base pairs sequenced.

  • CFB, Cytophaga–Flavobacter/Flexibacter–Bacteroides.

Fun-106DQ107384Ruegeria sp. AS-36 (AJ391197)99Alphaproteobacteria
Fun-112DQ107385Alphaproteobacterium MBIC1876 (AB026194)99Alphaproteobacteria
Fun-118aDQ107386Marine Alphaproteobacterium AS-26 (AJ391187)94Alphaproteobacteria
Fun-118bDQ107387Marine Alphaproteobacterium AS-26 (AJ391187)96Alphaproteobacteria
Fun-120DQ107388Erythrobacter sp. MB-16 (AF325446)99Alphaproteobacteria
Fun-123DQ107389Silicibacter lacuscaerulensis (U77644)97Alphaproteobacteria
Fun-125DQ107390Alphaproteobacterium NW4327 (AF384141)97Alphaproteobacteria
Fun-103DQ107391Marine bacterium NBF22 (AF343943)98Gammaproteobacteria
Fun-107DQ107392Vibrio pelagius (AJ293802)98Gammaproteobacteria
Fun-110DQ107393Leucothrix mucor (X87277)98Gammaproteobacteria
Fun-111DQ107394Vibrio sp. PH1 (AF513461)95Gammaproteobacteria
Fun-115DQ107395Alteromonas sp. (X86459)96Gammaproteobacteria
Fun-119DQ107396Shewanella fidelia KMM3582T (AF420312)96Gammaproteobacteria
Fun-127DQ107397Alteromonas sp. MED100 (AY136117)98Gammaproteobacteria
Fun-128DQ107398Pseudoalteromonas sp. S9 (U80834)99Gammaproteobacteria
Fun-101DQ107399Kocuria sp. KMM 3905 (AY211385)99Actinobacteria
Fun-102DQ107400Kocuria sp. KMM 3905 (AY211385)99Actinobacteria
Fun-105DQ107401Dermatophilus congolensis (AJ243918)93Actinobacteria
Fun-109DQ107402Micrococcus luteus (AJ409096)99Actinobacteria
Fun-121DQ107403Kytococcus sedentarius (X87755)97Actinobacteria
Fun-124DQ107404Staphylococcus pasteuri (AJ717376)100Firmicutes
Fun-113DQ107405Cytophaga sp. I-2029 (AB073566)92CFB

Phylogenetic analysis

The bacterial strains Fun-106, Fun-112, Fun-118a, Fun-118b, Fun-120, Fun-123 and Fun-125 were found to be members of the Alphaproteobacteria by 16S rRNA gene sequence analysis (Fig. 1). BLAST analysis of bacterial strains Fun-106, Fun-112, Fun-123 and Fun-125 revealed that all these strains are close relatives, with 97–99% similarity, of the strain Alphaproteobacterium MBIC1876, which was isolated from a marine sponge (GenBank accession number AB026194). In addition, all these strains are close relatives, with 96–99% similarity, of the strains Alphaproteobacterium SOGA14 and Alphaproteobacterium SOGA1, which are bacterial symbionts, isolated from the accessory nidamental glands of the cuttlefish Sepia officinalis (GenBank accession numbers AJ244791 and AJ244780, respectively). The two strains Fun-123 and Fun-125 are close relatives, with 97% similarity, of the strain Alphaproteobacterium NW4327, which was isolated from a diseased sponge from the Great Barrier Reef (GenBank accession number AF384141).

Figure 1.

 Neighbour-joining phylogenetic tree from analysis of >500 bp of 16S rRNA gene sequence of bacterial isolates from the Red Sea coral Fungia scutaria clustering within the Alphaproteobacteria. ‘f’ and ‘p’ indicate branches that were also found using the Fitch–Margoliash or maximum parsimony methods, respectively. The numbers at the nodes are percentages indicating the levels of bootstrap support, based on a neighbour-joining analysis of 1000 resampled data sets. Only values of >50% are shown. The scale bar represents 0.1 substitutions per nucleotide position.

The bacterial strains Fun-103, Fun-107, Fun-110, Fun-111, Fun-115, Fun-119, Fun-127 and Fun-128 were found to be members of the Gammaproteobacteria by 16S rRNA gene sequence analysis (Fig. 2). BLAST analysis of bacterial strain Fun-107 revealed that this strain is a close relative, with 98% similarity, of the strain Vibrio fortis LMG 21557, which has been isolated from aquatic animals and the marine environment (GenBank accession number AJ514916).

Figure 2.

 Neighbour-joining phylogenetic tree from analysis of >500 bp of 16S rRNA gene sequence of bacterial isolates from the Red Sea coral Fungia scutaria clustering within the Gammaproteobacteria. ‘f’ and ‘p’ indicate branches that were also found using the Fitch–Margoliash or maximum parsimony methods, respectively. The numbers at the nodes are percentages indicating the levels of bootstrap support, based on a neighbour-joining analysis of 1000 resampled data sets. Only values of >50% are shown. The scale bar represents 0.1 substitutions per nucleotide position.

The bacterial strains Fun-101, Fun-102, Fun-105, Fun-109 and Fun-121 were found to be members of the Actinobacteria by 16S rRNA gene sequence analysis (Fig. 3). BLAST analysis of bacterial strains Fun-101 and Fun-102 revealed that these strains are close relatives, with 99% similarity, of the strain Kocuria marina KMM 3905, a novel Actinobacterium that was isolated from marine sediment of the East Siberian Sea (GenBank accession number AY211385). Actinobacterium strain Fun-105 (Fig. 3) is a close relative, with 92% similarity, of the strain Arsenicicoccus bolidensis CCUG 47306, a novel genus of Actinobacteria that was isolated from contaminated lake sediment (GenBank accession number AJ558133). Strain Fun-109 was also found to be a member of the Actinobacteria (Fig. 3), closely related, with 99% similarity, to the strain Antarctic bacterium R-9183, which was isolated from microbial mats from Antarctic lakes (GenBank accession number AJ441006), and with 99% similarity to the strains Micrococcus sp. 98TH11321 and Micrococcus sp. 98TH11318, which are biofilm-forming marine bacteria from Korea (GenBank accession numbers AY159888 and AY159885, respectively). The closest relatives, with 96–97% similarity, of strain Fun-121 are strains in the genus Kytococcus, previously Micrococcus.

Figure 3.

 Neighbour-joining phylogenetic tree from analysis of >500 bp of 16S rRNA gene sequence of bacterial isolates from the Red Sea coral Fungia scutaria clustering within the Actinobacteria. ‘f’ and ‘p’ indicate branches that were also found using the Fitch–Margoliash or maximum parsimony methods, respectively. The numbers at the nodes are percentages indicating the levels of bootstrap support, based on a neighbour-joining analysis of 1000 resampled data sets. Only values of >50% are shown. The scale bar represents 0.01 substitutions per nucleotide position.

Biochemical characterization and antibiotic sensitivity of the cultured bacteria

The bacterial strains isolated in this study had different biochemical profiles (Tables 2 and 3), and exhibited variable sensitivity to six different commercial antibiotics (Table 4). These results confirm that all the isolates are different bacterial strains. Isolates that are close relatives according to the phylogenetic tree exhibited different biochemical profiles and antibiotic sensitivity, indicating phenotypic diversity in strains that were very closely related on the basis of 16S rRNA gene sequence analysis. For example, Fun-123 and Fun-125 in the Alphaproteobacteria tree fall closely together (Fig. 1), as do Fun-101 and Fun-102 in the Actinobacteria tree (Fig. 3) but their biochemical profiles and antibiotic sensitivities show that they are different bacterial strains (Tables 2–4).

Table 2.   Biochemical characterization of the bacteria isolated from the Red Sea coral Fungia scutaria*
ReactionFun-103Fun-106Fun-112Fun-113Fun-119Fun-120Fun-123Fun-125Fun-127Fun-128
  • *

    All the bacteria are non-enteric and analysed with api-20NE. The bacterial strains that are not presented here are Actinobacteria or strains that failed to be cultured from the cryo-preserved glycerol-stocks.

Nitrate reduction+++
Indole production
Acidification of glucose
Arginine dihydrolyase
Oxidase++++++++
Urease
β-glucosidase hydrolysis+++++++
Protease hydrolysis+++++
β-galactosidase++++++
Utilization of carbon source:
 Glucose+++++
 Arabinose
 Mannose++++
 Mannitol+++
 N-acetyl-glucosamine++++++
 Maltose++++
 Gluconate++
 Caprate
 Adipate
 Malate+++
 Citrate
 Phenyl-acetate
Table 3.   Biochemical characterization of the Actinobacteria isolated from the Red Sea coral Fungia scutaria*
ReactionFun-101Fun-102Fun-109Fun-121
  • *

    The Actinobacteria isolate Fun-105 was not analysed because it failed to be cultured from the cryo-preserved glycerol-stocks.

Nitrate reduction++
Pyrazinamidase+++
Pyrrolidonyl arylamidase+
Alkaline phosphatase++++
β-glucuronidase
β-galactosidase++
α-glucosidase++++
N-acetyl-β-glucosaminidase
Oxidase+
β-glucosidase hydrolysis+
Urease++
Protease hydrolysis++
Catalase++
Utilization of carbon source:
 Control
 Glucose++++
 Ribose+++
 Xylose++++
 Mannitol++
 Maltose++++
 Lactose
 Sucrose+++
 Glycogen
Table 4.   Antibiotic sensitivity of the cultured bacterial strains isolated from the Red Sea coral Fungia scutaria
Bacterial
strain
Penicillin G
(10)
Tetracycline
(30)
Clindamycin
(2)
Polymyxin B
(300)
Chloramphenicol
(30)
Novobiocin
(30)
  • *

    R, resistant to the antibiotic.

  • The applied concentrations of the antibiotics are noted in parentheses. All concentrations are in μg, except for penicillin G and polymyxin B, which are in units. Sensitivity is expressed as inhibition zones (diameter in mm) around each antibiotic disc, excluding the antibiotic disc diameter

Fun-10124117R*RR
Fun-102>30RRRRR
Fun-103RRR105R
Fun-106RRRRRR
Fun-109RRR1447
Fun-110RRR29R
Fun-112416431621
Fun-1131015R416R
Fun-119RRR716R
Fun-120R10R12277
Fun-1213R7R1916
Fun-123RRR15RR
Fun-125R8R17RR
Fun-127RRR11RR
Fun-128RRR41411

Discussion

The culturable heterotrophic bacterial community of the mucus of the Red Sea stony coral F. scutaria is composed mainly of the bacterial groups Alphaproteobacteria, Gammaproteobacteria and of Actinobacteria (Fig. 4). The fact that we found that a significant proportion of the bacterial community is Actinobacteria is surprising, because previous studies that characterized the culturable bacterial communities in several other corals did not find any members of this important bacterial group (Ritchie & Smith, 1995; Rohwer et al., 2001). These earlier studies used glycerol artificial seawater medium (GASW) as an isolation medium. Both Bacto Marine Agar 2216 used in the present study and GASW are nonselective media. Therefore, the medium used is probably not the reason why a different composition of heterotrophic bacteria was detected. When we compared the composition of bacterial isolates of the coral Montastraea franksi (Rohwer et al., 2001) to the present study (Fig. 4), we found that, while in M. franksi the isolates are dominated mainly by Gammaproteobacteria, in F. scutaria there is almost an equal distribution of the bacterial groups Alphaproteobacteria, Gammaproteobacteria and of Actinobacteria. The recently proposed class of Actinobacteria (Stackebrandt et al., 1997) is composed of high-G+C-content Gram-positive bacteria and includes the actinomycetes (order Actinomycetales). Most of the actinomycetes were isolated from soil and are best known as a source of diverse bioactive natural products. The distribution of actinomycetes in the sea remains largely undescribed. Several studies have reported the isolation of novel marine actinomycetes (Jensen et al., 2005a, b). A sponge from the Great Barrier Reef has been reported to harbour novel actinomycetes, of which only a few were culturable (Webster et al., 2001), and another sponge contained novel Actinobacteria in the subclass Acidimicrobidae (Montalvo et al., 2005). Recently, a major new taxon of obligate marine actinomycetes was discovered from ocean sediments (Mincer et al., 2002). To the best of our knowledge, the current study is the first evidence of actinomycetes isolated from corals.

Figure 4.

 Differences between coral-associated bacterial communities of (a) isolates from the mucus of the Red Sea coral Fungia scutaria (this study), and (b) isolates from the whole tissue of the Caribbean coral Montastraea franksi (Rohwer et al., 2001). CFB, Cytophaga–Flavobacter/Flexibacter–Bacteroides; BC, Firmicutes (including Bacillus–Clostridium strains).

The bacterial strains found in this study were isolated from the mucus layer covering the coral, as opposed to the whole tissue used in previous studies (Rohwer et al., 2001, 2002). The coral mucus is a rich substrate for microorganisms and therefore we focused on this ecological niche. We can compare this microcosm to the well-described ‘marine snow’ in coastal and open ocean environments. Marine snow is composed of macroscopic organic aggregates in the pelagic environment, which were found to be enriched with nutrients and microbial biomass in comparison with the surrounding water (Rath et al., 1998; Azam & Long, 2001). These microcosms are important in the vertical transport and biogeochemical transformation of particulate organic carbon in the marine environment. Several studies have reported that bacterial species inhabiting marine snow particles may differ greatly from those commonly found living free in the surrounding water column (DeLong et al., 1993; Bidle & Fletcher, 1995; Rath et al., 1998; Azam & Long, 2001). The phenomenon that unique microbial species inhabit a nutrient-rich microcosm seems to be similar in coral mucus.

The commonly used criterion for proposing novel genera of bacteria is <93% identity of the sequences to previously published 16S rRNA gene sequences in the GenBank, and <97% for proposing novel species (Rohwer et al., 2002). Based on these criteria, ∼32% of the bacteria isolated in the current study were novel, one at the genus level and the remainder at the species level. The Cytophaga–Flavobacterium sp., Fun-113, is most likely a novel genus, and 27% of the isolates are considered novel species. These results correspond to a previous study of the Caribbean coral Montastraea franksi (Rohwer et al., 2001), where 20% of the isolates were considered novel species and one isolate was also considered a novel genus. It is interesting to note that the strain Fun-113 is a close relative of other bacterial strains isolated from the mucus of the coral Oculina patagonica (unpublished, GenBank accession numbers AY654759, AY654756, AY654778, AY654825), indicating that this strain is most likely a coral-mucus-associated bacterium.

The culturable microbial community of the Red Sea coral F. scutaria comprises only c. 1% of the total bacteria found in the mucus. The culture-based approach taken in the current study enabled us to isolate, for the first time, Actinobacteria from corals. However, in order to understand better the nature of microbial communities associated with coral mucus it is necessary to use cloning and sequencing of 16S rRNA genes to describe the complete microbial community composition. Clone libraries of bacteria from F. scutaria, as well as from other Red Sea corals, were constructed and will be reported in a separate study.

Acknowledgements

We thank the staff of the Inter University Institute of Marine Biology at Eilat for their hospitality and facilities. The comments by two anonymous reviewers are greatly appreciated. This work was supported in part by the US National Science Foundation under Microbial Observatories Grant No. 0238515 to R.T.H. This research was also funded by a Rieger Foundation award to Y.L. This is contribution no. 05–132 from the Center of Marine Biotechnology. This manuscript is part of Y.L.'s dissertation for her PhD at Bar Ilan University.

Ancillary