Bacterial assemblages differ between compartments within the coral holobiont

It is widely accepted that corals are associated with a diverse and host species-specific microbiota, but how they are organized within their hosts remains poorly understood. Previous sampling techniques (blasted coral tissues, coral swabs and milked mucus) may preferentially sample from different compartments such as mucus, tissue and skeleton, or amalgamate them, making comparisons and generalizations between studies difficult. This study characterized bacterial communities of corals with minimal mechanical disruption and contamination from water, air and sediments from three compartments: surface mucus layer (SML), coral tissue and coral skeleton. A novel apparatus (the 'snot sucker') was used to separate the SML from tissues and skeleton, and these three compartments were compared to swab samples and milked mucus along with adjacent environmental samples (water column and sediments). Bacterial 16S rRNA gene diversity was significantly different between the various coral compartments and environmental samples (PERMANOVA, F = 6.9, df = 8, P = 0.001), the only exceptions being the complete crushed coral samples and the coral skeleton, which were similar, because the skeleton represents a proportionally large volume and supports a relatively rich microflora. Milked mucus differed significantly from the SML collected with the 'snot sucker' and was contaminated with zooxanthellae, suggesting that it may originate at least partially from the gastrovascular cavity rather than the tissue surface. A common method of sampling the SML, surface swabs, produced a bacterial community profile distinct from the SML sampled using our novel apparatus and also showed contamination from coral tissues. Our results indicate that microbial communities are spatially structured within the coral holobiont, and methods used to describe these need to be standardized to allow comparisons between studies.