Introduction

Coral bleaching occurs when the symbiosis between corals and their endosymbionts (Symbiodinium spp) disintegrates in response to stress (Jokiel and Coles, 1977; Steen and Muscatine, 1987; Hoegh-Guldberg and Smith, 1989; Kleppel et al., 1989; Glynn and D’Croz, 1990; Lesser et al., 1990; Glynn, 1991; Hoegh-Guldberg, 1999; Leggat et al., 2006). The incidence of mass coral bleaching has increased over the past 20 years and is directly correlated with periods of exceptionally warm sea temperatures (Glynn, 1991; Hoegh-Guldberg, 1999). Projections using climate models suggest that sea temperatures may eventually rise above the current thermal tolerances of corals at almost all sites worldwide and severely impact sustainability of coral reefs (Glynn, 1991; Hoegh-Guldberg, 1999; Hughes et al., 2003; Donner et al., 2005). It has been widely documented that during periods of warm sea surface temperatures, symbiont photosynthesis is reduced due to an increased susceptibility to photo-inhibition, which leads directly to active oxygen production and results in the breakdown of the symbiosis (Jokiel and Coles, 1977; Iglesias-Preito et al., 1992). Alternatively, recent studies have proposed that bacterial pathogens are the primary cause of bleaching in reef-building corals (Rosenberg and Ben-Haim, 2002; Rosenberg and Falkovitz, 2004; Rosenberg et al., 2007). The Bacterial Bleaching Hypothesis (Rosenberg and Ben-Haim, 2002; Rosenberg and Falkovitz, 2004) arose from studies in which the bacterium, Vibrio shiloi, was shown to cause the annual bleaching of the Mediterranean invading coral Oculina patagonica during warm summer months (Kushmaro et al., 1996, 1998). This hypothesis was extended to other corals and proposed as an alternative to climate change as the primary driver for the increase in mass coral bleaching over the past two decades (Rosenberg and Falkovitz, 2004).

O. patagonica bleaches annually in precisely mid-June when water temperatures rise above 24°C, bleaching then continues throughout the spring and summer, by which time water temperatures reach 32°C and up to 80% of the population has bleached (Kushmaro et al., 1998, 2001; Israely et al., 2001). V. shiloi has been reported, from laboratory studies, to become virulent at water temperatures of 24°C, penetrating the coral and multiplying within the tissues (Kushmaro et al., 1998; Israely et al., 2001). Studies revealed that V. shiloi adheres to β-D-galactopyranoside receptors in the mucus of O. patagonica prior to penetrating the epithelial cells (Banin et al., 2000). V. shiloi then differentiates into a viable-but-not-culture-able form, multiplying and producing a proline-rich peptide toxin, which inhibits Symbiodinium photosystems (Banin et al., 2001a, 2001b). The conclusions drawn from this O. patagonica/V. shiloi model have been used as the basis to propose bacterial pathogenesis as the cause of global patterns of mass bleaching (Rosenberg and Ben-Haim, 2002; Rosenberg and Falkovitz, 2004; Reshef et al., 2006; Rosenberg et al., 2007). The present study aims to investigate the in situ bacterial involvement, and in situ coral microbial ecology, in ecological patterns of bleaching in O. patagonica across the Isreali coastline.

Materials and methods

O. patagonica was extensively monitored (over 140 samples) along the Mediterranean coast of Israel throughout the annual bleaching event in 2005. Bleached colonies, as well as colonies with no apparent colour loss or bleaching were monitored. During the bleaching season (14 June until 22 August), bleached colonies and control unbleached colonies were sampled every 2 weeks at Sdot Yam. Bleached and unbleached colonies were sampled during bleaching season at other sites along the length of the Israeli Mediterranean coastline including Ashkelon, Bat Yam and Acziv. From each sampled colony, three replicate core samples were taken of each tissue region from each coral colony. Regions were designated either unbleached tissue or bleached tissue, bleached tissue cores were taken around the bleaching lesion to ensure both the bleached tissue and active region of bleaching were sampled. The bleaching lesion refers to the interface between the bleached and normally pigmented tissue. Coral cores were fixed in 4% paraformaldehyde in phosphate-buffered saline for 12 h and then stored at 4 °C in phosphate-buffered saline. Coral cores were decalcified in 20% (w/v) EDTA in phosphate-buffered saline, and tissues were processed sequentially through 70, 80, 95 and 100% ethanol, three xylene and three paraffin washes, each for 40 min, prior to paraffin embedding. Serial tissue sections (4 μm) were collected onto Superfrost Plus slides (Menzel, Germany). Fluorescence in situ hybridization (FISH) was used with bacterial oligonucleotide probes in conventional FISH protocols (Manz et al., 2000; Amann et al., 1990, 2001; Amman et al., 1996) and coupled with spectral imaging (Ainsworth et al., 2006), to explore the identity and structural complexity of microbial communities associated with O. patagonica tissue layers. The technique provided imagery of the active bleaching and microbial community dynamics of the bleaching under field conditions. Photomicrographs of positive control samples of experimentally infected O. patagonica with V. shiloi (Rosenberg and Resef, pers. comm. (2005)) coupled with FISH and control tissues labelled with a V. shiloi-specific monoclonal antibody demonstrate the comparative microbial community dynamics associated with experimental pathogenic bleaching.

Intracellular bacterial proliferation was also examined via transmission electron microscopy

Cores of bleached and healthy O. patagonica colonies (n=6) were collected between 2–5 m depth from reefs adjacent in Sdot Yam (latitude 32.4765°N, longitude 34.8864°E) on the Israeli coast in July 2005. Small fragments (2 cm2) from each colony were immediately fixed in 3% gluteraldehyde in 0.1 M cacodylate buffer. Coral cores were decalcified in 20% EDTA at 4 °C (Ainsworth et al., 2006), and ultra thin 1 μm tissue sections were stained using 1% Toluidine blue (Sigma-Aldrich Pty Ltd, St Louis, MO, USA, catalogue no. 89640). To investigate general tissue structure and endosymbiont location samples were photographed using standard light microscopy, following which sample grids with ultra-thin sections were viewed in a transmission electron microscope (JEOL 1010) at acceleration voltage 90 kV and images taken using the Megaview III Soft Imaging system.

Results

V. shiloi is not associated with O. patagonica bleaching

Bleaching of O. patagonica occurred along the Mediterranean coastline to the same extent as reported previously. Bleaching was observed throughout the summer period within coral colonies from Northern, mid and southern regions of the Isreali coastline. Bleaching was observed to begin in mid-June 2005 and continued throughout the summer months. However, the annual mass bleaching of O. patagonica along the coast was found to occur without the involvement of V. shiloi, Vibrio species or any bacterial penetration and proliferation within its tissue layers.

No evidence of bacterial populations was found associated with any of the 48 samples of bleached corals, 48 samples of unbleached tissues of corals undergoing bleaching or 48 samples healthy samples taken during this study (Table 1). Each replicate sample was individually probed with general bacterial probe mix, γ-proteobacterial probe and Vibrio sp probe, with a resulting 144 FISH experiments conducted on each region type. Each in situ was also photographed at replicated regions within the tissue sample at 20, 40 and 63 times magnification. Samples were also arbitrarily selected and FISH analysis repeated. Bacterial populations penetrating and/or multiplying were not found within either the epithelium or gastrodermis of the bleached regions or bleaching lesions of O. patagonica tissues using FISH (Figure 1). Samples were investigated with a range of bacterial probes, including EUB338mix (Manz et al., 2000; Amann et al., 1990, 2001; Amman et al., 1996), GAM 42a (Manz et al., 2000) and a marine Vibrio probe (Moreno et al., 1999), none of which labelled any bacterial populations within the tissue layers of bleached coral colonies (Figure 1). Significant intracellular bacterial populations were absent from all normal unbleached colony tissues and unbleached tissue regions of bleached colonies investigated within this study (Figure 1). Electron microscopy of bleached tissues also demonstrates a lack of intracellular bacterial proliferation (Figure 1).

Table 1 Number of coral colonies and replicate samples investigated of O. patagonica on the Israeli Eastern Mediterranean coastline
Figure 1
figure 1

Bleached (a, b) and healthy (c, d) O. patagonica colonies from the Israeli Mediterranean coastline during the summer of 2005. Bleached colonies are evident by the white lesions (b) and areas of low density of endosymbiotic dinoflagellates (e, f). Large populations of endolithic microbial communities were found in both bleached (i, j) and healthy corals, (h, k, l) but no Vibrio sp as evident by the lack of probe staining in either bleached (i, j) or healthy colonies (k, l), as also seen using transmission electron microscopy of bleached (m, n) and healthy tissues (o, p). Scale, 1 cm (ad), 50 μm (el), 1 μm (mp). BL, bleaching lesion; H, healthy tissue; Zx, endosymbiotic dinoflagellates; En, endolithic microbial population; Po, Polyp.; Ga, gastroderm.

FISH probes used on tissues of O. patagonica experimentally infected with V. shiloi (provided by Rosenberg and Resef, pers. comm. (2005)) demonstrated that bacterial proliferation could be identified within O. patagonica tissues pathogen-bleached experimentally (Figure 2).

Figure 2
figure 2

FISH of O. patagonica infected with cultured V. shiloi in a laboratory based infection experiment. V. shiloi is evident within the tissue layers using a V. shiloi-specific monoclonal antibody (ac) and also using EUBmix 338 (df) universal bacterial probe. Scale, 5 μm. Vib, bacteria identified by V. shiloi Mab (red); Bac, bacteria identified by EUB probe (red); Zx, endosymbiotic dinoflagellates (green); Ct, coral tissue (blue).

The only microbial communities interacting with and in close association with the tissue layers of field bleached corals were members of the endolithic community (Figures 1, 3 and 4). Endolithic communities were found underlying not only bleached lesions but also healthy coral tissues and healthy corals prior to bleaching. A shift in endolithic community dominance was observed associated with the coral bleaching, with communities dominated by cyanobacterial members (based on binding of EUBmix probes and presence of chlorophyll pigment spectral signature) in corals prior to bleaching (Figure 4) being replaced by communities dominated by green algae such as Ostreobium spp in bleached regions (Figure 3).

Figure 3
figure 3

FISH and subtractive spectral imaging of filamentous endolithic communities associated with O. patagonica. Subtractive spectral imaging removes the innate chlorophyll-conferred fluorescence (spectral emittance region 575–625 nm) (a) and indicates positive probe binding of the cy-3 fluorescence (spectral emission at 543 nm) of EUBmix probe (b, c) and identification as bacterial (potentially cyanobacterial communities). Scale, 50 μm. BL, bleaching lesion; H, healthy tissue; Zx, endosymbiotic dinoflagellates; EL, endolithic microbial population.

Figure 4
figure 4

Bleaching of O. patagonica tissues in summer of 2005 where endosymbiotic dinoflagellates (Zx) are in low densities in bleached tissue regions (a, c), compared to high densities in unbleached tissue regions adjacent to bleached tissues (b), a change in endolithics dominance to algal communities associated with bleaching lesions, evident by the large size of the algal dominant endolithic population (EL). Scale, 50 μm. BL, bleaching lesion; H healthy tissue; Zx Endosymbiotic dinoflagellates; En, endolithic microbial population. Arrowheads indicate different sizes of endolithic community members.

Discussion

It can be stated unambiguously that V. shiloi is not involved in, nor is currently the primary cause of, the continued annual bleaching of O. patagonica of the Eastern Mediterranean. Not only did bleaching occur without the involvement of V. shiloi in 2005, but the subsequent death of bleached coral colonies took place without Vibrio sp being present. A lack of vibrio-driven bleaching in ecological scenarios is also supported by the recent lack of bacterial bleaching found with in experimental conditions (Reshef et al., 2006). We therefore suggest that there is no evidence to support a primary role of bacteria in causing coral bleaching in the system used, as basis, for the Bacterial Bleaching Hypothesis, or evidence to suggest microbial control techniques could be effective to treat or prevent mass coral bleaching.

The only detectable microbial communities interacting with and in close association with the tissue layers were members of the endolithic community. Here, endolithic communities are found underlying not only bleached lesions but also healthy coral tissues and healthy corals prior to bleaching. Interestingly, a shift in endolithic community dominance occurs associated with the coral bleaching, where communities dominated by cyanobacterial members in corals prior to bleaching are replaced by communities dominated by green algae during the coral bleaching. Koren and Rosenberg (2006) describe high diversity of bacterial communities associated within coral tissues of O. patagonica. Here, it can be seen that the skeleton of the coral may in fact be a region where the high bacterial diversity of O. patagonica is located. This shift in the microbial community associated with O. patagonica highlights the potential importance of the diverse and complex interactions between the many associated components of this coral holobiont and its unique qualities when compared to reef-building corals.

The role of the endolithic community of O. patagonica, particularly during increasing seawater temperatures and environmental stress, and that of the community shift during bleaching evoke questions as to the relationship the microbial community has in the stability of the coral holobiont, as well as the potential sources of opportunistic pathogens. Complex biofilms, such as microbial mats, represent rich, structured and productive microbial ecosystems (Guerrero et al., 2002). Phototrophic biofilms have previously been shown to colonize hard substrate in aquatic and terrestrial ecosystems and contribute large quantities of total fixed carbon to the ecosystem (Guerrero et al., 2002; de la Torre et al., 2003). The role of the endolithic community of the coral O. patagonica and the overall community dynamics associated with stress, bleaching and disease remains to be determined. At this point, it is clear that this layer is far from passive, as demonstrated by the shift occurring during bleaching, and may play a role in a number of key processes.

While microbial populations play important roles governing coral stasis, as evident by the coral holobiont model (Rohwer et al., 2002) and within coral diseases, the importance of opportunistic pathogenesis and the susceptibility of corals related to primary environmental stress is an important question that requires further study. The role of opportunistic pathogens should also be considered while suggesting the application of bacteriophages or microbial treatments of coral reefs during mass bleaching events and disease. If bacteria do not play a primary but rather a secondary role during coral bleaching or some diseases (in that corals are susceptible to microbial attack during stress), it needs to be determined if the use of microbial remedies at local or regional scales could reduce the impact of disease events in what are highly diverse, interconnected and complex microbial environments. This may further be supported by the discovery of a widely distributed and proposed coral pathogen Vibrio coralliilyticus, which causes the lysis of tissues and whitening of tropical corals (presumably via loss of tissue and exposure of the coral skeleton) during increased seawater temperatures (Ben-Haim and Rosenberg, 2002; Ben-Haim et al., 2003a, 2003b). Bourne and Munn (2005) propose the wide distribution of this bacterium and its association with healthy coral may highlight its opportunistic rather than primary pathogenic role. In proposals of bacterial pathogenesis related to coral bleaching a clear and accurate differentiation of the cellular level changes associated with the coral bleaching in ecologically relevant scenarios is important. Differentiation of the cellular causes of whitening of coral colonies (exposure of the coral skeleton) as opposed to coral bleaching (loss of endosymbiotic algae or its associated pigments from host tissue) and the cellular mechanisms underlying these visual changes is imperative in accurately determining the cause of coral death. This is needed if we have to understand the role or microbial communities in the patterns of tissue loss and bleaching.

The findings of this study on the lack of pathogenic V. shiloi or bacterial interaction associated with the bleaching of O. patagonica are important in that they raise questions on hypotheses of bacterial mass coral bleaching (Rosenberg and Ben-Haim, 2002; Rosenberg and Falkovitz, 2004; Reshef et al., 2006) and coral hologenome evolution (Rosenberg et al., 2007). The probiotic hypothesis (Reshef et al., 2006) and the coral hologenome theory of evolution (Rosenberg et al., 2007) state that corals have the ability for rapid evolution in a changing environment by altering their microbial partners to overcome adverse environmental conditions and disease. However, the continued active annual bleaching of O. patagonica in the absence of V. shiloi, in the same manner as previous reports, suggests that while the microbial population has changed, there is no apparent microbial influence over patterns of coral bleaching. As has been the case for the past decade, bleaching of O. patagonica along the coast of Israel continues, beginning precisely in mid-June and continuing through the summer months until coral colony mortalities occur in mid-August.

As such, our study also suggests that caution is required for laboratory studies investigating bacterial pathogenesis with aquaria maintained or pre-stressed corals not held under normal environmental conditions. Rosenberg et al. (2007) report on a newly occurring short-term adherence and penetration of V. shiloi cultures to aquaria-maintained O. patagonica; however, our field-based survey was unable to detect any Vibrio spp adherence to O. patagonica. The study of Kline et al. (2006) shows that bacterial cultures from healthy corals when applied to corals in aquaria experiments caused a doubling of the normal bacterial load, which leads to coral mortality. Documenting the natural microbial community variability that exists in reef systems and how this may vary in contained systems is crucial in determining the impact and role of coral-associated communities not only in coral health but also during stress and disease.