OUP user menu

Coral-associated bacterial communities on Ningaloo Reef, Western Australia

Janja Ceh, Mike Van Keulen, David G. Bourne
DOI: http://dx.doi.org/10.1111/j.1574-6941.2010.00986.x 134-144 First published online: 1 January 2011

Abstract

Coral-associated microbial communities from three coral species (Pocillopora damicornis, Acropora tenuis and Favites abdita) were examined every 3 months (January, March, June, October) over a period of 1 year on Ningaloo Reef, Western Australia. Tissue from corals was collected throughout the year and additional sampling of coral mucus and seawater samples was performed in January. Tissue samples were also obtained in October from P. damicornis coral colonies on Rottnest Island off Perth, 1200 km south of Ningaloo Reef, to provide comparisons between coral–microbial associates in different locations. The community structures of the coral-associated microorganisms were analysed using phylogenetic analysis of 16S rRNA gene clone libraries, which demonstrated highly diverse microbial profiles among all the coral species sampled. Principal component analysis revealed that samples grouped according to time and not species, indicating that coral–microbial associations may be a result of environmental drivers such as oceanographic characteristics, benthic community structure and temperature. Tissue samples from P. damicornis at Rottnest Island revealed similarities in bacteria to the samples at Ningaloo Reef. This study highlights that coral-associated microbial communities are highly diverse; however, the complex interactions that determine the stability of these associations are not necessarily dependent on coral host specificity.

Keywords
  • bacteria
  • coral
  • Ningaloo Reef
  • 16S rRNA gene

Introduction

Coral reefs worldwide are currently subjected to unprecedented degradation, likely caused by anthropogenic influences including increased sea surface temperatures, coastal degradation, pollution, diseases, ecosystem imbalance and the synergistic effects of multiple stressors (Harvell, 2002; Pandolfi, 2003; Bruno, 2007). These changing anthropogenic stressors often upset the delicate balance between the coral animal and its associated microorganisms, which includes symbiotic dinoflagellates (Zooxanthellae), endolithic algae, fungi, Bacteria, Archaea and viruses, all closely associated with the coral animal and together representing the dynamic assemblage referred to as the coral holobiont (Rohwer, 2002; Rosenberg, 2007; Bourne, 2009). The importance of coral–microbial interactions is increasingly being recognized, with studies demonstrating the active role these associated microbial communities play in maintaining coral health and resilience. Bacteria associated with corals have been studied extensively and found to occupy several microniches within the coral host including the surface mucus layer, the tissue and the calcium carbonate skeleton (reviewed in Rosenberg, 2007). Information on the spatial and temporal composition of bacterial communities associated with corals is also accumulating, with previous studies suggesting that some associations may be relatively stable and species specific (Ritchie & Smith, 1997; Rohwer, 2001, 2002; Littman, 2009). The functional role of these bacteria associated with coral is poorly understood, although reports have indicated they may be involved in the provision of nutrients (Shashar, 1994; Olson, 2009) and the exclusion of potentially pathogenic microorganisms (Ritchie, 2006).

Corals are increasingly faced with changing conditions on both regional and global scales. Understanding microbial communities associated with corals, their functional roles and how they change through time is the key to understanding how these changes will affect the coral holobiont. Shifts in the bacterial community composition may affect coral health and increase the susceptibility of the host to disease (Bourne & Munn, 2005; Ritchie, 2006; Bourne, 2008). The specificity of coral–bacterial associations is still not well understood and there is a requirement for long-term monitoring studies to enhance our understanding of the interactions between the coral host and associated microorganisms, and to assess the potential influence of environmental parameters structuring these communities. If bacteria routinely form species-specific associations with corals, it would be expected that such associations would be maintained over space and time (Rohwer, 2002). However, several investigators have shown that the bacterial population associated with a coral can change as a function of disease or varying environmental conditions (Pantos, 2003; Ritchie & Smith, 2004; Koren & Rosenberg, 2006; Bourne, 2008; Hong, 2009). Based on these and other studies, Reshef (2006) proposed the Coral Probiotic Hypothesis, suggesting that corals can adapt rapidly to changing environmental conditions by altering their population of symbiotic bacteria. To study coral microbial interactions, it is necessary to determine which microorganisms reside on or within corals, whether there is a geographical consistency and how the dynamics of the coral-associated consortia vary in time. Multiple impacts, such as geographical and environmental factors (Koren & Rosenberg, 2006; Bourne, 2008; Hong, 2009; Littman, 2010) as well as coral physiology (Littman, 2010), have been suggested to drive the structure of coral-associated microorganisms.

This study examines coral–microbial associations over time for three coral species, Pocillopora damicornis, Acropora tenuis and Favites abdita, living in close proximity to each other on Ningaloo Reef. Corals were sampled repeatedly throughout the year to investigate whether temporal environmental changes would lead to a natural variation in coral-associated bacterial diversity. To test whether the growth form of corals plays a role in structuring bacterial consortia as has been suggested previously (Sunagawa, 2010), we included branching (P. damicornis, A. tenuis) vs. massive corals (F. abdita) in our study. Pocillopora damicornis samples were collected from two locations on the Western Australian coast, Ningaloo Reef and Rottnest Island, to identify which bacteria might be conserved across geographically distinct locations. This study provides the first analysis and comparison of microbial communities from Western Australian corals. Compared with other reef systems, warm water bleaching events and outbreaks of coral diseases are virtually unknown for Ningaloo Reef, which makes it an excellent reference study place representing a relatively healthy coral reef environment.

Materials and methods

Sampling sites

Ningaloo Reef is the largest fringing coral reef system in Australia (280 km in length) and the only extensive coral reef worldwide fringing the west coast of a continent (Taylor & Pearce, 1999). The main reef line consists of a series of elongated reef segments, which are punctuated by gaps with channels through which the majority of lagoon flushing occurs. The shallow sedimentary lagoon has a main depth of about 2 m, with occasional patch and nearshore platform reefs.

Rottnest Island (32°00′S, 115°31′E) is located 19 km off the coast from Fremantle and is 11 km long and 4.5 km wide at its widest point; its long axis is oriented east–west. During winter, the island is warmed by the Leeuwin Current (Fig. 1), providing a higher minimum temperature (19 °C) than the adjacent mainland coast (15 °C). This temperature regime allows 25 species of 16 genera of corals to occur around Rottnest Island, which makes it the best-developed coral community of the temperate coast (Veron & Marsh, 1988).

1

Location of the study sites (Coral Bay and Rottnest Island) where corals were sampled. The polewards-flowing Leeuwin Current indicated by the arrow is part of a complex interplay current system that moves seawater along the Western Australian coast and is only represented as a simplification in this figure.

Sample collection

Two replicate coral colonies, similar in size, of three coral species, P. damicornis, A. tenuis and F. abdita, were tagged on a reef flat (5–6 m water depth) near Coral Bay (23°07′S, 113°07′E), Ningaloo Reef, Western Australia. Two replicate coral nubbins (approximately 2 cm in size) from each tagged colony were placed in individual zip-lock bags under water, rinsed on the surface twice with 0.2 μm filtered and autoclaved artificial seawater (ASW) to remove loosely attached microorganisms, and then placed on ice. Massive corals were sampled using a hammer and chisel to extract a 2 × 2 cm piece of coral. The coral samples were air brushed with 2 mL of ASW to remove the coral tissue including the associated microorganisms from the coral skeleton. The tissue slurry was aliquoted into cryovials and stored at−80 °C until required for analysis. Two replicate coral mucus samples per coral colony were collected using 50-mL syringes. Seawater samples were collected in sterile plastic bottles (1 L) about 2 m distance from each tagged coral colony. Mucus samples as well as water samples were filtered through Sterivex (0.22 μm) filter columns (Millipore) and stored at −80 °C for later analysis. Samples were processed within 1 h of sampling. Sampling scars healed before resampling and all corals appeared to be healthy throughout the sampling period and 6 months after the last collection. Samples were collected randomly from the coral colonies, regardless of former samplings. Coral tissue, coral mucus and water samples were collected in January to assess bacterial communities between these different environments, which have been shown to differ in previous studies (Frias-Lopez, 2002; Rohwer, 2002; Bourne & Munn, 2005; Pantos & Bythell, 2006). Subsequent sampling in March, June and October 2008 from Ningaloo Reef was of coral tissue only. Seawater temperatures in close proximity to the sampling site on Ningaloo Reef were recorded at 30-min intervals using a temperature logger (Odyssey, Temperature Recorder) for the sampling period from late March to November to assess the possible correlations of changes in coral microbiota with seawater temperature. A logger fault resulted in the loss in temperature data from January till March, with the missing information substituted with sea surface temperature data of Ningaloo Reef from the National Oceanic Atmospheric Administration Satellite and Information service (http://coralreefwatch.noaa.gov/satellite/current/sst_series_ningaloo_cur.html).

Additional replicate tissue samples from two P. damicornis colonies were sampled once from Rottnest Island (32°00′S, 115°31′E), 1200 km south of Coral Bay, to allow for comparisons between associated bacteria from geographically dispersed corals of the same species.

DNA extraction and purification

Coral tissue slurries (2 mL) were centrifuged at 13 000 g to pellet the tissue and the associated microbiota. The supernatant was removed and the pellet was resuspended in 500 μL of buffer (0.75 M sucrose, 40 mM EDTA, 50 mM Tris base, pH 8.3). Total DNA was extracted, following the methods specified in Bourne (2008). For the extraction of DNA from mucus and seawater samples, the Sterivex filter units were filled with 1.6 mL of lysis buffer (0.75 M sucrose, 40 mM EDTA, 50 mM Tris base, pH 8.3) and 0.2 mL lysozyme solution (1 mg mL−1) was added. The filter unit was incubated at 37 °C for 1 h, 0.2 mL Proteinase K/sodium dodecyl sulphate (SDS) solution (0.2 μg mL−1 Proteinase K+1% SDS) was added and the filter column was incubated at 55 °C for 1 h. The lysate was recovered, the Sterivex filter columns were rinsed with an additional 1 mL lysis buffer and the pooled lysates were extracted twice with an equal volume of phenol–chloroform–isoamyl alcohol (25 : 24 : 1, pH 8). The liquid phase was removed and extracted with an equal volume of chloroform–isoamyl alcohol (24 : 1); after removing the aqueous phase, 0.1 mL sodium acetate (3 M, pH 5.2) along with 1 mL 100% isopropanol were added and the DNA were pelleted at 13 000 g at 4 °C for 10 min and washed with 70% ethanol (modified method by Schauer, 2000).

Extracted total DNA of all samples (coral tissue, coral mucus and seawater) were resuspended in 30 μL of sterile Milli-Q water, loaded on a 1.2% low-melting agarose gel and the high-quality DNA (>2 kb) were cut out and purified using a QIAquick gel extraction kit (Qiagen) following the instructions of the manufacturer. The DNA was recovered from the columns in two washes (30 and 20 μL) of sterile Milli-Q water and stored at −20 °C until further processing.

PCR amplification of 16S rRNA genes and clone library construction

Bacterial-specific primers 63f and 1387r (Marchesi, 1998) were used to amplify the 16S rRNA genes from extracted DNA for bacterial clone library construction of coral tissue, coral mucus and seawater. The PCR mixtures (50 μL) contained 0.2 pmol μL−1 of each primer, 200 mM each dNTP, 1 × PCR buffer [Tris-Cl, KCl, (NH4)2SO4, 1.5 mM MgCl2], 0.08% w/v bovine serum albumin and 1.25 U of Taq polymerase (Scientifix, Clayton, Vic., Australia). PCR was performed using an Applied Biosystems 2720 thermocycler and programmed with an initial 4 min step at 94 °C and 30 cycles consisting of 94 °C for 1 min, 55 °C for 1 min and 72 °C for 1.5 min and a final extension for 10 min at 72 °C.

The amplified bacterial DNA of pooled replicates for each extracted sample were ligated into the TOPO-TA cloning vector (Invitrogen, Carlsbad, CA), following the manufacturer's instructions. This resulted in 17 clone libraries with ligations that were then submitted to the Australian Genome Research Facility for transformation, cloning and subsequent sequencing. Ninety-six clones were sequenced from each library using the M13f primer. The nucleotide sequence data of all clones reported in this paper appear in the GenBank nucleotide sequence database under the accession numbers GU184380GU185837.

Sequence analysis and statistical analysis

Sequences were checked for chimera formation using the check_chimera software of the Ribosomal Database Project (Maidak, 1996). Sequence data were aligned to the closest relative using the blast database algorithm (Altschul, 1997). Sequence affiliations were determined by >97% identity to bacterial 16S rRNA gene sequences in the GenBank database. From a total of 1450 16S rRNA gene sequences, bacteria were grouped into operational taxonomic units (OTUs) based on the assumption that bacteria that share>97% sequence identity represent an individual OTU (Ward, 1998).

The variation of microbial diversity in the clone libraries was investigated using the following indices and models (Magurran, 1988): the Shannon diversity index (H) (Shannon & Weaver, 1963), Simpson's evenness index (D) (Simpson, 1949), Fisher's alpha log series richness index (Fisher, 1943) and coverage (C) values (Good, 1953). A principal component analysis (PCA) was implemented to determine whether bacterial profiles from coral samples grouped according to species or time. Any ribotype constituting 5% or more (arbitrarily assigned as dominant) of each clone library was included in the PCA. PCA statistical analyses were carried out using past statistical software (Ryan, 1995).

Results

Clone library analyses

Bacterial clone libraries derived from three coral species, P. damicornis, A. tenuis and F. abdita collected in January from Ningaloo Reef, were dominated by Alphaproteobacteria- and Gammaproteobacteria-affiliated sequences (Table 1). Alphaproteobacteria sequences were dominant in P. damicornis tissue (73%), although they only represented 15.9% and 17.3% of A. tenuis and F. abdita libraries, respectively. Gammaproteobacteria sequences constituted the majority of clones in A. tenuis (72%) and F. abdita (34.7%) libraries, although they were in low relative abundance in P. damicornis libraries (5.4%). In addition, P. damicornis contained a relatively high number of Flavobacteria (6.5%), whereas 8%Betaproteobacteria- and 8%Actinobacteria-affiliated sequences were obtained from the coral F. abdita. Additional clone libraries constructed from the mucus derived from the same coral colonies displayed a higher bacterial diversity as assessed by diversity indices including the Shannon diversity index (H), Simpson's evenness index (D) and Fisher's alpha log series richness index (Table 2). Similar to the tissue samples, Alphaproteobacteria- and Gammaproteobacteria-affiliated sequences dominated these mucus libraries although Bacteroidetes- and Flavobacteria-affiliated sequences were also more predominant, constituting between 6.5% to 12% and 10.6% to 19.6% of the libraries, respectively. The sequences obtained from the surrounding seawater showed a high resemblance to the mucus libraries, with the relative abundance of retrieved Bacteriodetes- and Flavobacteria-affiliated sequences being similar. The exception was that Alphaproteobacteria-affiliated sequences were more often associated with coral mucus (two- to threefold) than with seawater, whereas the abundance of Gammaproteobacteria was much higher (two- to fourfold) in seawater libraries compared with coral mucus libraries (Table 1).

View this table:
1

Proportions of bacterial taxonomic classes for each clone library

January clone libraries (%)March clone libraries (%)June clone libraries (%)October clone libraries (%)
Bacteria classificationP. damicornisA. tenuisF. abditaP. damicornis MA. tenuis MF. abdita MSeawaterP. damicornisA. tenuisF. abditaP. damicornisA. tenuisF. abditaP. damicornisA. tenuisF. abditaP. damicornis R
Alphaproteobacteria73.1015.9017.3030.4021.3021.709.6025.0011.0023.303.202.1013.7044.4027.6039.1044.10
Betaproteobacteria2.202.308.000.000.001.100.005.7021.908.102.101.108.400.002.301.103.20
Gammaproteobacteria5.4072.7034.709.809.6026.1047.0025.0017.8018.6073.4071.6047.409.9014.9013.0023.70
Deltaproteobacteria0.000.001.302.201.100.000.000.000.002.300.000.000.002.501.101.106.10
Bacteroidetes2.200.001.306.5010.6012.006.004.504.104.700.000.000.003.702.305.403.10
Flavobacteria6.501.102.7018.5010.6019.6016.905.702.708.100.000.000.000.005.706.500.80
Sphingobacteria0.001.100.000.001.101.102.400.000.000.000.000.000.000.000.000.000.00
Actinobacteria0.001.108.000.000.000.000.004.502.700.001.100.000.000.000.000.000.00
Bacilli0.000.002.700.000.001.102.400.000.000.0013.808.4011.600.000.000.000.00
Cyanobacteria0.000.000.002.200.000.000.000.002.700.000.000.001.106.204.603.303.10
Firmicutes0.000.000.000.000.000.000.000.000.001.200.000.000.001.200.000.000.00
Unclassified10.805.7024.0030.4045.7017.4015.7029.5037.0033.706.4016.8017.9032.1041.4030.4015.90
  • M, mucus-derived libraries.

  • R, Rottnest Island sampled corals.

View this table:
2

Diversity indices calculated from OTUs of 16S rRNA gene clones derived from all clone libraries

January clone librariesMarch clone librariesJune clone librariesOctober clone libraries
Bacteria classificationP. damicornisA. tenuisF. abditaP. damicornis MA. tenuis MF. abdita MSeawaterP. damicornisA. tenuisF. abditaP. damicornisA. tenuisF. abditaP. damicornisA. tenuisF. abditaP. damicornis R
No. of clones analysed9388759294928388738694959581879265
No. of OTU groups (ribotypes)878668791083466988
Coverage of clone libraries (%)35.4880.684832.6129.7958.760.2434.0939.7340.787.2382.1167.3728.432.1829.3526.15
Shannon diversity (H)1.471.081.451.471.511.711.582.132.051.990.550.511.131.731.91.871.99
Fisher's alpha (α)3.361.933.192.262.342.62.224.365.43.240.610.891.622.644.593.845.9
Simpson's evenness (D)0.660.510.660.710.740.780.760.880.840.850.310.250.550.810.810.830.79
  • M, mucus-derived libraries.

  • R, Rottnest Island sampled corals.

The bacterial profiles associated with the coral tissue changed over time as the same coral colonies were repeatedly sampled in January, March, June and October. Although Alphaproteobacteria- and Gammaproteobacteria-affiliated sequences still dominated the libraries, their relative proportions changed through the sampling period. For example, compared with tissue samples in January, tissues in March demonstrated a decrease in both Alphaproteobacteria- and Gammaproteobacteria-affiliated sequences, while ribotypes affiliated with Betaproteobacteria, Bacteroidetes and Flavobacteria were more prominent in March (Table 1). Coral microbial profiles derived from June samples demonstrated the most drastic shift, with Gammaproteobacteria sequences dominating in all libraries (between 47.4% and 73.4%). Alphaproteobacterial sequences only constituted between 2.1% and 13.4% of the libraries, while Bacillales-affiliated sequences represented 8.4–13.8% of the libraries. Profiles of October-derived samples again shifted, and like January and March samples, displayed a higher relative abundance of Alphaproteobacteria-affiliated sequences (39.1–44.4%), while the relative abundance of Gammaproteobacteria sequences declined (9.9–14.9%). No Bacillales-affiliated sequences were retrieved while small numbers of Bacteroidetes (2.3–5.4%), Flavobacteria (0–6.5%) and Cyanobacteria (3.3–9.2%) were retrieved from all coral species (Table 1).

Libraries constructed from the tissue of P. damicornis coral sampled from geographical distant colonies at Rottnest Island were pooled because the relative abundance of ribotypes was consistent. Coral-associated bacterial profiles at each site displayed a strong similarity in diversity and community structure (Fig. 2, Table 2). Both Ningaloo and Rottnest libraries were dominated by Alphaproteobacteria (44.4% and 44.1%) and Gammaproteobacteria (9.9% and 23.7%), respectively (Fig. 2). In addition, sequences related to Deltaproteobacteria (2.5% and 6.1%), Bacteroidetes bacteria (3.7% and 3.1%) and Cyanobacteria (6.2% and 3.1%) were retrieved. The Ningaloo clone library additionally contained a small proportion of Firmicutes-related sequences, while Betaproteobacteria and Flavobacteria sequences were retrieved from tissue of corals from Rottnest Island.

2

Dominant bacterial 16S rRNA gene sequence affiliations for Pocillopora damicornis clone libraries in October. Sequences were grouped into dominant ribotypes (>5% of clone libraries) at the phylum and class level.

Sequence affiliations at the lower taxonomic levels of genus and family were incorporated into a PCA, which, similar to the grouping at the class and phylum level, demonstrated that libraries derived from the corals grouped according to the sampling time and not the species type (Fig. 3). For example, January samples were correlated with an increased relative abundance of Alcanivorax sp.-, Methylarcula sp.- and Salinivibrio sp.-affiliated sequences. In comparison, March coral tissue libraries correlated with Achromobacter sp., Acinetobacter sp., Bacteroidetes and Brevundimonas sp. June libraries grouped based on strong correlations to Bacillales sp. and Vibrio sp., whereas the clone libraries derived from October sampling associated with Cyanobacteria, Erythrobacter sp. and Rhodobacter sp. sequences. Interestingly, the October clone library of P. damicornis collected from Rottnest Island grouped closely to the clone libraries derived from the coral samples collected in October on Ningaloo Reef, showing a strong consistency within seasonal bacterial communities regardless of geographical separation (Fig. 3). Coral mucus samples collected in January grouped together based on correlations with Flavobacteria, Litoricola sp., Methylarcula sp. and Pseudoalteromonas sp. sequences, although this was also consistent with the seawater clone library.

3

Biplot PCA of clone library 16S rRNA gene sequences (family and genus level). Black lines show vectors representative of dominant OTUs (>5% of one or more libraries) driving the differences between clone libraries. Sequence affiliations are included in the analysis and represented by the numbers on the vector coordinates.

A comparison between clone libraries showed that none of the dominant OTU groups (ribotypes>5% within each library) were consistently found in the libraries derived from the same coral species at all sampling time points. For example, sequences grouped in OTU 1, which affiliated with Achromobacter sp., were only found in January and June clone libraries. In addition, OTU groups that did occur at more than one time point were generally detected in all coral species.

Comparison of the diversity indices between all libraries revealed a similar bacterial diversity for different coral species for most sampling times (Table 2). Based on the total number of OTUs, the highest diversity in bacteria was found in March and October, which represent the months with the highest and the lowest water temperature throughout the sampling period. The water temperatures on the actual sampling days were (28.8 °C) and (21.8 °C), respectively (Supporting Information, Fig. S1). Finally, no differences in the bacterial community structure or diversity were detected between branching and massive corals.

Discussion

Previous studies using molecular techniques to profile the bacterial 16S rRNA genes associated with corals have demonstrated highly diverse and abundant microbial communities (Rohwer, 2001, 2002; Cooney, 2002; Frias-Lopez, 2002; Bourne & Munn, 2005). For example, Rohwer (2001) statistically estimated as many as 6000 different ribotypes to be associated with corals. Despite this high diversity in the coral microbiota, a conserved microbial community has been suggested in some coral species (Ritchie & Smith, 1997; Frias-Lopez, 2002; Rohwer, 2002; Bourne, 2008) and bacterial populations from the same coral species have been shown to be consistent even if coral individuals are separated spatially or temporally (Rowher, 2002). The results from this study similarly demonstrate high bacterial diversity associated with corals from the Western Australian coast. However, sequence affiliations at the taxonomic levels of class and phylum, as well as at the genus and family level, demonstrated that libraries derived from three physically adjacent coral species sampled at four different time points throughout the year grouped according to sampling time and not coral species. These conclusions were drawn from consistent observations in dominant bacterial ribotypes (>5% of libraries) recovered from clone libraries of each coral species at each sampling time and suggest that while different coral species may harbour similar bacterial communities, there is low species specificity in the coral–bacterial associations.

Corals and their associated bacteria exist in a delicate balance that is crucial to their survival, and the maintenance of homeostasis has been suggested as a critical factor to the persistence of a mutualistic symbiosis (Kline, 2004). In this study, bacterial profiles for continuously sampled corals were highly different for each sampling time, which strongly indicates temporal shifts in the bacterial communities within these coral species at this site. The results do not suggest major long-term homeostasis between the coral host and its associated bacteria. A recent study by Hong (2009) investigated the species specificity of bacteria associated with the coral Stylophora pistillata and suggested that multiple dynamic factors including seasonal and geographic features were drivers for bacterial diversity in individual coral colonies, supporting the results in the current study. Hong (2009) showed little bacterial species specificity for the Stylophora colonies under investigation although they speculated that such species specificity may be variable for different corals. Their analysis of previous studies on coral-associated bacteria also showed that spatial and temporal factors were important in shaping these associations, although distinct species-specific bacterial profiles were identified in some Caribbean corals. For example, the Caribbean coral species Montastraea annularis and Diploria strigosa were found to harbour similar microbial consortia, with over half of the associated bacteria being present in both coral species (Klaus, 2005). However, a study by Rohwer (2002) noted that bacterial communities associated with healthy colonies of D. strigosa differed substantially from those reported from the same coral species in a study by Frias-Lopez (2002) and highlighted a different methodology as a possible explanation for these differences. A comparison of two studies, both of which investigated the diversity of bacteria associated with the coral Montastraea franksi (Rohwer, 2001, 2002), indicated no essential overlap between retrieved 16S rRNA gene sequences either (Rohwer, 2002). All these studies together (including the current results) indicate ubiquitous and generally conserved microbial–coral consortia, but do not necessarily suggest that different coral species harbour distinct bacterial communities.

Most studies to date investigating coral–bacterial associations have used clone libraries to assess bacterial diversity, which, due to inherent limitations, undersample the overall diversity of bacteria associated with corals and may contribute to the lack of overlap in sequences retrieved within and between studies. Developing sequencing technologies that are able to provide an in-depth analysis of bacterial diversity can overcome these limitations. A recent study using deep sequencing and analysis of>350 000 16S rRNA gene tags demonstrated that bacterial community composition displayed similar profiles among closely related coral in the same genus or family, but not at higher coral taxonomic levels (Sunagawa, 2010). Applying these developing in-depth sequencing technologies to analyse temporal changes and anthropogenic effects will further enhance our understanding of the stability of coral bacterial associations. Sunagawa (2010) also highlighted the possibility that morphology may play a role in structuring coral microbiota. Our current study investigated three coral species from three different families (Acroporidae, Pocilloporidae and Faviidae), and although all coral demonstrated similar coral-associated profiles at the same time, this was independent of the coral morphology, which included both branching and massive forms.

Given the assumption that microbial communities live in beneficial relationships with their coral host and that corals are exposed to various environmental factors throughout a reproductive cycle, it makes sense that corals harbour different bacterial types at different times according to the benefit these bacteria can provide to the present requirements of their host. Koren & Rosenberg (2006) reported dynamic bacterial communities associated with healthy Oculina patagonica corals and identified temperature as one factor causing microbial shifts between seasons. While Vibrio splendidus-affiliated sequences were dominant (35% of sequenced clones) and appeared to be stable both in summer and in winter, the next 10 most abundant clusters of bacteria differed between seasons. The results from our study detected neither a dominant group of bacteria throughout the year nor a bacteria class to be dominant at all times. Corals in January (with the exception of P. damicornis) were dominated by Gammaproteobacteria, while the relative abundance of Alphaproteobacteria and Gammaproteobacteria sequences was similar in March samples. A shift towards Gammaproteobacteria (47.4–73.4%) was observed in all June samples, with sequences related to Vibrio species being largely responsible for this change representing most of the Gammaproteobacteria-retrieved sequences in these samples. Bacillales-related sequences also increased in the June samples and represented between 8.4% and 13.8% of the sequences in the clone libraries (Table 1). At the final sampling time (October), Alphaproteobacteria-related sequences were again dominant. Some classes of bacteria were mainly or exclusively found in particular months. For example, Bacillales-affiliated sequences were highly abundant in June from all coral-derived libraries, whereas Bacteroidetes as well as Flavobacteria sequences were detected in most coral libraries throughout the sampling period, except in June. Deltaproteobacteria and Cyanobacteria were abundant in October. Despite Vibrio species being implicated previously in bleaching of some coral species, the shift in June (Winter) to a Vibrio-dominated community did not visibly compromise coral health, which supports previous findings that members of this group may form a natural part of the microbial population of healthy corals (Bourne & Munn, 2005). These bacteria are likely involved in important roles within the coral host including fixation of nitrogen (Olson, 2009), which may be particularly important in times of nutrient limitations such as winter months in oligotrophic environments.

Mucus-associated microbial communities are taxonomically and functionally diverse (Ritchie, 2006; Bourne & Munn, 2005) and differ from the bacterial populations in coral tissue and the surrounding seawater (Koren & Rosenberg, 2006). In this study, mucus from three different coral species contained similar bacterial communities and all mucus samples clustered closely on a PCA plot along with the seawater sample. Despite clustering of mucus and water samples, Alphaproteobacteria-affiliated sequences were more abundant in coral mucus than in seawater, whereas Gammaproteobacteria were higher in seawater libraries compared with coral mucus. The number of Bacteroidetes- and Flavobacteria-affiliated sequences, however, was found to be very similar in both mucus and seawater libraries. This potentially indicates that coral mucus selects for Alphaproteobacteria rather than Gammaproteobacteria when bacteria of both classes are abundant and available in the seawater column. The structure of bacterial communities within coral mucus is controlled by specific receptors that bind bacteria on the coral surface mucus (Kvennefors, 2008) and by signals and biocides found in coral mucus (Brown & Bythell, 2005;Ritchie, 2006; Shnit-Orland & Kushmaro, 2009). Given the fact that the coral mucus layer provides a substrate for microbial growth (Ducklow & Mitchell, 1979b; Rublee, 1980; Herndl & Velimirov, 1986) and interactions between the coral host and bacteria are initiated during the colonization of this layer, the mucus might serve as a selection barrier for bacteria that enter into close and stable associations within the coral host tissue. Bacteria that do not colonize this niche potentially only remain in the coral mucus temporarily and are removed by the cleansing mechanism of sloughing, subsequently initiating a new selection cycle.

Coral reefs are known for low nutrient concentrations (Muscatine, 1980; Rahav, 1989; Szmant, 1990; Gast, 1998, 1999; Gili & Coma, 1998); increased algal abundance can cause elevated levels of dissolved organic carbon (DOC) due to excess photosynthate released into the water column by the algal communities (Smith, 2006). Once algae become highly abundant and algal tissue begins decomposing, even larger amounts of organic carbon can be released into the water column and utilized by heterotrophic bacteria (Dinsdale, 2008). Previous studies have shown that elevated DOC levels can disrupt the balance between the coral and its associated microbiota by accelerating the growth rate of bacteria living in the coral mucus layer (Kline, 2004; Smith, 2006). Algal outbreaks with a dominant abundance of Sargassum and Turbinaria occur seasonally on Ningaloo Reef (M. van Keulen, pers. commun.), starting in October/November, peaking around March and dying off in May/June. The seasonal pattern of high productivity and abundance of algae on Ningaloo Reef may have an influence on coral bacterial shifts, particularly in June, when algal detritus increases. For example, sequences related to heterotrophic Vibrio species represented up to 70% of clones derived from libraries in June, potentially indicative of a response to available nutrients released by algal die-off. Diversity indices calculated consistent values of associated bacteria between coral species and sampling times and revealed the highest diversity of microbial ribotypes in March and October, when the average daily temperature was highest and the lowest, respectively, throughout the sampling period. Similar to previous studies (Rosenberg, 2007; Bourne, 2008), these results further indicate that temperature may directly or indirectly have an effect on the diversity of coral-associated bacterial assemblages.

A study carried out along the Northern Line Islands in the central Pacific Ocean demonstrated that the ecology of microbial communities in seawater responds to regional oceanographic differences such as local upwelling and benthic community structure (Dinsdale, 2008). Similarly, these factors may play a role in bacterial communities along Ningaloo Reef. With its central section being located only 1 km off the continental shelf (Hearn & Parker, 1988), the coastal current system along Ningaloo can generate transient upwelling in summer (D'Adamo & Simpson, 2001; Woo, 2006). Environmental parameters have been shown to correlate with shifts in coral–microbial associations and we suggest that changing bacterial communities in the water column driven by regional oceanographic characteristics may contribute to these shifts. Sample collection from P. damicornis in October displayed strong similarities in bacterial diversity and community structure between Rottnest Island and Ningaloo Reef, despite the great distance between the sites and their difference in ecosystem structure. All coral tissue samples collected in October grouped closely on the PCA plot. The Leeuwin Current, which meanders polewards along the Western Australian coast (see Fig. 1), transports warm water along with marine larvae and picoplankton from more northern latitudes towards the south (Paterson, 2008); this could include bacterial communities in the water column. Assuming that coral-associated microbial bacteria are influenced by the microbial communities present in the water column, the Leeuwin Current, as part of the oceanographic regime of the Western Australian coast, could partly explain the conserved coral-associated microbial consortia among coral species from Ningaloo Reef and geographically distant corals from Rottnest Island.

The present study provides the first overview on the coral-associated microbial community structure and diversity of Western Australian corals. Our findings suggest seasonal changes to be involved in driving the microbial consortia in the three coral species, P. damicornis, A. tenuis and F. abdita, rather than coral species and spatial separation. Neither a distinct consortium of bacteria nor coral bacterial species specificity was detected in the coral species investigated; therefore, long-term homeostasis between the coral hosts and their associated bacteria was not confirmed. Temporal shifts in coral–bacterial associations are potentially driven by the bacterial community present in the water column, which may again be influenced by oceanographic characteristics, benthic community structure and temperature. Currently, it remains uncertain whether the coral host facilitates particular bacteria with a favourable habitat to allow for a mutual, beneficial relationship or whether it merely provides suitable living conditions for invading opportunistic microbial consortia that regulate the coral-associated microbial community structure by competing for the same ecological niche. A better understanding of these coral–microbial relationships is important for the conservation, restoration and management of coral reefs worldwide.

Supporting Information

Fig. S1. Daily average sea temperature for the Ningaloo Reef sampling site in 2008.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Acknowledgements

We would like to thank the Australian Institute of Marine Science, the Western Australian Marine Science Institution and Murdoch University for their contributions to this research. We also thank Frazer McGregor, Joetta Perrett, Samuel Wells and Kim Mars for their help in the field, Tim Simmonds is thanked for help in preparing Fig. 1 and Jean-Baptiste Raina for critical comments on the manuscript.

References

View Abstract