Stable and sporadic symbiotic communities of coral and algal holobionts

Abstract

Coral and algal holobionts are assemblages of macroorganisms and microorganisms, including viruses, Bacteria, Archaea, protists and fungi. Despite a decade of research, it remains unclear whether these associations are spatial-temporally stable or species-specific. We hypothesized that conflicting interpretations of the data arise from high noise associated with sporadic microbial symbionts overwhelming signatures of stable holobiont members. To test this hypothesis, the bacterial communities associated with three coral species (Acropora rosaria, Acropora hyacinthus and Porites lutea) and two algal guilds (crustose coralline algae and turf algae) from 131 samples were analyzed using a novel statistical approach termed the Abundance-Ubiquity (AU) test. The AU test determines whether a given bacterial species would be present given additional sampling effort (that is, stable) versus those species that are sporadically associated with a sample. Using the AU test, we show that coral and algal holobionts have a high-diversity group of stable symbionts. Stable symbionts are not exclusive to one species of coral or algae. No single bacterial species was ubiquitously associated with one host, showing that there is not strict heredity of the microbiome. In addition to the stable symbionts, there was a low-diversity community of sporadic symbionts whose abundance varied widely across individual holobionts of the same species. Identification of these two symbiont communities supports the holobiont model and calls into question the hologenome theory of evolution.

Figure 1

Bacterial communities associated with coral and algal holobionts. Each host's associated community was assessed at the OTU level with the Bray–Curtis dissimilarity and those dissimilarities were reduced to principle coordinates, embedding them in Euclidean space (a). The variation of each community was evaluated by measuring the Euclidean distance for each sample associated with a host to the host centroid (b). The bar represents the median distance to centroid and the boxes represent upper and lower 25% quartiles. The whiskers represent minimum and maximum values.

Figure 2

AU method for determining significant deviations in an OTU's distribution across the samples studied. The AU method was utilized to distinguish between OTUs (97% similar) whose community relative abundance did not vary across each community studied and those that did. OTUs that fall above the 0.01 confidence interval (black line) are found in fewer samples than expected given their mean community relative abundance (that is, OTUs whose distribution in the data set is restricted and are therefore classified as sporadic). Panels a–e represent bacterial communities associated with Acropora hyacinthus (a), Acropora rosaria (b), Porites lutea (c), CCA (d) and Turf (e). Panel f is a mock plot illustrating the different classifications of OTUs based on the AU test. The different color points indicate sporadic (red), stable predicted (blue) and stable observed symbionts (green).

Figure 3

Comparison of the Holobiont model (Rohwer et al. 2002; Knowlton and Rohwer, 2003) and Hologenomic Theory of Evolution (Rosenberg et al., 2007). The primary difference between the two hypotheses is the importance of heredity. For the illustration, the coral animal is represented by hexagons, the colored circles represent Symbiodinium and oval shapes represent Bacteria. In actuality, the symbionts also include viruses, Archaea, fungi, protists, other macrobes and so on. In the Holobiont model, (a) heritability is not required and the relationships among the bionts are best described by ecological dynamics. Membership in the holobiont may be stable or sporadic as predicted by the AU test (blue ovals in figure) or obligate (for example, plastids or mitochondria). Stable and sporadic symbionts are recruited from other holobionts or from the environment (different color ovals). Changing environmental conditions lead to symbiont switching. Symbionts remain independent entities, may be found outside the holobiont and are sometimes absent from any individual holoboint of the same species (that is, not ubiquitous, similar to PA1). Even in a stable environment, members of the holobiont may be replaced by bionts with similar functions. (b) The Hologenome is evolutionary and requires heredity, otherwise it incorporates the same premises laid forth under the Holobiont model and is indistinguishable.