Globally, reef-building corals are the most prolific producers of dimethylsulphoniopropionate (DMSP), a central molecule in the marine sulphur cycle and precursor of the climate-active gas dimethylsulphide. At present, DMSP production by corals is attributed entirely to their algal endosymbiont, Symbiodinium. Combining chemical, genomic and molecular approaches, we show that coral juveniles produce DMSP in the absence of algal symbionts. DMSP levels increased up to 54% over time in newly settled coral juveniles lacking algal endosymbionts, and further increases, up to 76%, were recorded when juveniles were subjected to thermal stress. We uncovered coral orthologues of two algal genes recently identified in DMSP biosynthesis, strongly indicating that corals possess the enzymatic machinery necessary for DMSP production. Our results overturn the paradigm that photosynthetic organisms are the sole biological source of DMSP, and highlight the double jeopardy represented by worldwide declining coral cover, as the potential to alleviate thermal stress through coral-produced DMSP declines correspondingly.

The majority of marine invertebrates produce dispersive larvae which, in order to complete their life cycles, must attach and metamorphose into benthic forms. This process, collectively referred to as settlement, is often guided by habitat-specific cues. While the sources of such cues are well known, the links between their biological activity, chemical identity, presence and quantificationin situare largely missing. Previous work on coral larval settlementin vitrohas shown widespread induction by crustose coralline algae (CCA) and in particular their associated bacteria. However, we found that bacterial biofilms on CCA did not initiate ecologically realistic settlement responses in larvae of 11 hard coral species from Australia, Guam, Singapore and Japan. We instead found that algal chemical cues induce identical behavioral responses of larvae as per live CCA. We identified two classes of CCA cell wall-associated compounds – glycoglycerolipids and polysaccharides – as the main constituents of settlement inducing fractions. These algae-derived fractions induce settlement and metamorphosis at equivalent concentrations as present in CCA, both in small scale laboratory assays and under flow-through conditions, suggesting their ability to act in an ecologically relevant fashion to steer larval settlement of corals. Both compound classes were readily detected in natural samples.

The induction of larval attachment and metamorphosis of benthic marine invertebrates is widely considered to rely on habitat specific cues. While microbial biofilms on marine hard substrates have received considerable attention as specific signals for a wide and phylogenetically diverse array of marine invertebrates, the presumed chemical settlement signals produced by the bacteria have to date not been characterized. Here we isolated and fully characterized the first chemical signal from bacteria that induced larval metamorphosis of acroporid coral larvae (Acropora millepora). The metamorphic cue was identified as tetrabromopyrrole (TBP) in four bacterial Pseudoalteromonas strains among a culture library of 225 isolates obtained from the crustose coralline algae Neogoniolithon fosliei and Hydrolithon onkodes. Coral planulae transformed into fully developed polyps within 6 h, but only a small proportion of these polyps attached to the substratum. The biofilm cell density of the four bacterial strains had no influence on the ratio of attached vs. non-attached polyps. Larval bioassays with ethanolic extracts of the bacterial isolates, as well as synthetic TBP resulted in consistent responses of coral planulae to various doses of TBP. The lowest bacterial density of one of the Pseudoalteromonas strains which induced metamorphosis was 7,000 cells mm⁻² in laboratory assays, which is on the order of 0.1 –1% of the total numbers of bacteria typically found on such surfaces. These results, in which an actual cue from bacteria has been characterized for the first time, contribute significantly towards understanding the complex process of acroporid coral larval settlement mediated through epibiotic microbial biofilms on crustose coralline algae.

The crown-of-thorns starfish (COTS, the Acanthaster planci species group) is a highly fecund predator of reef-building corals throughout the Indo-Pacific region¹. COTS population outbreaks cause substantial loss of coral cover, diminishing the integrity and resilience of reef ecosystems^2,3,4,5,6. Here we sequenced genomes of COTS from the Great Barrier Reef, Australia and Okinawa, Japan to identify gene products that underlie species-specific communication and could potentially be used in biocontrol strategies. We focused on water-borne chemical plumes released from aggregating COTS, which make the normally sedentary starfish become highly active. Peptide sequences detected in these plumes by mass spectrometry are encoded in the COTS genome and expressed in external tissues. The exoproteome released by aggregating COTS consists largely of signalling factors and hydrolytic enzymes, and includes an expanded and rapidly evolving set of starfish-specific ependymin-related proteins. These secreted proteins may be detected by members of a large family of olfactory-receptor-like G-protein-coupled receptors that are expressed externally, sometimes in a sex-specific manner. This study provides insights into COTS-specific communication that may guide the generation of peptide mimetics for use on reefs with COTS outbreaks.