Could Phage Therapy be Effective Against Coral Disease?

The dramatic, global deterioration of tropical coral reefs has been connected to anthropogenic stressors such as poor water quality (including sedimentation, nutrification and pollution), direct physical damage (e.g. dynamite fishing), overexploitation of keystone species and elevated sea surface temperatures (SSTs).  Such factors are correlated to increasingly widespread and severe incidences of coral disease.  Vibrio spp. have been linked to temperature dependant bleaching and tissue necrosis in many species of tropical corals.  For example, the putative coral pathogen Vibrio coralliilyticus has been implicated in the coral disease ‘White Syndrome’ (WS), in which tissue necrosis is preceded by the loss of Symbiodinium.  Previously, it has been shown that V. coralliilyticus causes inactivation of Symbiodinium photosystem (PS) II via production of the proteolytic enzyme Zinc-metalloprotease, leading to inhibition of photosynthetic activity (and subsequent loss of Symbiodinium).  Such virulence is thought to be multifactorial, dependant on the general structure of the host microbiota (e.g. Symbiodinium susceptibility varies via clade) and temperature (which contributes to host density and Zn-metalloprotease production).

                Cohen et al. (2013) were interested in the potential for phage therapy treatment of V. coralliilyticus-based disease outbreaks on the Great Barrier Reef (GBR), and began by isolating a bacteriophage infectious to the V. coralliilyticus strain P1 (previously isolated from diseased corals).  The phage (named bacteriophage YC) was isolated from the waters around Magnetic Island on the GBR, where strain P1 was also originally isolated.  A series of enrichments and plaque assays was conducted, producing a high titre, pure phage stock.  The morphology of phage YC was observed as consistent with the Myoviridae family (indicated via Electron micrograph images), and the rate of phage absorption on to strain P1 was determined to be rapid (ascertained by adsorption kinetics experiments).  In order to test how phage YC might influence the affect of strain P1, the authors grew up various concentrations of strain P1, both without the addition of phage YC, and with phage added after 0, 2, 8 and 18 h (after the beginning of bacterial growth), and treated Symbiodinium cells and coral (Acropora millepora) juveniles with filtered and centrifuged ‘P1 supernatant’.

                The authors observed that Symbiodinium treated with P1 supernatant derived from bacterial cultures incubated with YC phage demonstrate significantly less photosynthetic inhibition (reduced quantum yield values), compared to Symbiodinium treated with P1 supernatant derived from bacterial cultures to which YC phage had not been added (>90% reduction quantum yield), suggesting higher PSII inactivation in this treatment.  Similar results were observed from the coral juveniles.  Expulsion of Symbiodinium, tissue lesions, and eventual total tissue loss coincided with quantum yield measurements of zero, 9 h after treatment with P1 supernatant produced without the addition of YC phage.  Conversely, in juveniles with P1 supernatant derived from bacterial cultures with YC phage administered after 2 h of growth, no change in appearance or signs of Symbiodinium expulsion were observed, and no decline in quantum yields was recorded.  However, juveniles treated with P1 supernatant derived from bacterial cultures incubated with YC phage added after 8, and 18 h of growth suffered a very similar fate to juvenilles treated with P1 supernatant produced without the addition of YC phage (i.e. total tissue loss etc).

This investigation is unusual because: a) the authors applied their isolated phage to the putative pathogen before applying it to the Symbiodinium/coral juveniles; b) they actually removed bacterial cells before applying the P1 supernatant to the Symbiodinium/juveniles.  Previous phage therapy investigations have generally added both the putative bacteria and the phage to the corals, either together (Efrony et al., 2006), or with an interval before adding the phage (Efrony et al., 2009).  Significantly, the high densities of V. coralliilyticus employed in this investigation have been recorded in the coral host, rather than in seawater, yet in this investigation the role of the host in bacteria-bacteriophage interaction has been excluded.  Additionally, it is hard to imagine a scenario where the coral host would experience only the secondary metabolites of a pathogen, rather than the bacterium itself. For these reasons, it appears that the relevance of these findings to what may occur in the field may be questionable.    

Cohen et al. (2013) conclude that as Zn-metalloprotease production is known to occur at high V. coralliilyticus densities, and because the addition of phage was observed to cause bacterial lysis, the results they observed were due to the presence/absence of Zn-metalloprotease expression in consequence of the presence/absence of phage activity.  Perhaps the removal of bacterial cells in this investigation was undertaken in order to provide evidence for the role of phage in reducing Zn-metalloprotease production.  This might be investigated further with the application of some sort of Zn-metalloprotease assay.  It appears that, although the authors have shown that the phage they isolated was effective against strain P1, more work needs to be undertaken before the relevance of this finding to the prevention of coral disease on the GBR can be ascertained.

Cohen, Y., Joseph Pollock, F., Rosenberg, E. & Bourne, D. G. (2013) Phage therapy treatment of the coral pathogen Vibrio coralliilyticus. MicrobiologyOpen 2: 64–74.