Worldwide exploration of the microbiome harbored by the cnidarian model, Exaiptasia pallida (Agassiz in Verrill, 1864) indicates a lack of bacterial association specificity at a lower taxonomic rank

Introduction

Insights into the microbiome diversity of metazoan hosts have triggered a considerable interest in uncovering the regulatory principles underlying host/microbe interactions across multicellular organisms. Over the last several years, microbial symbionts living with vertebrates have been clearly shown to influence disease, physiological and developmental phenotypes in their host (Blaser et al., 2013; Le Chatelier et al., 2013; Lozupone et al., 2013; Ridaura et al., 2013). In many marine invertebrates, bacteria associated with host epithelium have also been shown to play a pivotal role in host development (McFall-Ngai et al., 2013). For instance, in the bobtail squid, the bioluminescent bacteria, Allivibrio fisheri (Beijerinck, 1889; Urbanczyk et al., 2007) are required symbionts from early host developmental stages so that a functional and healthy light organ can develop (McFall-Ngai, 1994; Nyholm & McFall-Ngai, 2004). Similar profound effects have been documented for other basal metazoans such as cnidarians. In the case of Hydra viridis (Medusozoa: Hydrozoa), induced absence of a microbial community in host polyps causes strong developmental defects and reduces asexual reproduction via budding (Rahat & Dimentman, 1982). This suggests that the evolution of microbes and host interactions dates back to earlier diverging metazoan lineages (i.e., cnidarians), which has triggered an imperative interest to understand whether bacterial cores comprised of specific species have evolved in intimate association with their hosts since the early times of metazoan evolution (Bosch & Miller, 2016).

Despite the simple body plans in cnidarians molecular analyses of the microbiota associated with these early-diverging organisms, predominantly corals (Anthozoa: Scleractinia), have uncovered an unprecedented bacterial diversity (Bourne & Munn, 2005; Rodriguez-Lanetty et al., 2013; Rohwer et al., 2001; Sunagawa et al., 2009). Additionally, the species composition and structure of these microbial partnerships are complex and dynamic. The association between coral host and the consortia of these microorganisms, including bacteria, fungi, viruses and the intracellular microalgae Symbiodinium, has been referred to as the coral holobiont (Rohwer et al., 2001). Several studies demonstrate that certain bacterial groups associate specifically with some coral species (Bayer et al., 2013; Morrow et al., 2012; Rodriguez-Lanetty et al., 2013; Rohwer et al., 2001; Speck & Donachie, 2012), implicating the effects of coevolution between coral lineages and certain bacterial strains. However, other studies have revealed that the dominant bacterial genera differ between geographically-spaced hosts of the same coral species (Klaus et al., 2007; Kvennefors et al., 2010; Littman, Bourne & Willis, 2010) and even locally within reefs (Kvennefors et al., 2010), which suggests that environmental factors are largely responsible in shaping coral-associated microbial community diversity.

The elucidation of the mechanisms that mediate the complex interactions between microbial communities and anthozoans may be facilitated by studying a tractable model system that can be cultured and manipulated in laboratory conditions. For this purpose, the sea anemone (Anthozoa: Actiniaria) Exaiptasia pallida (Agassiz in Verrill, 1864), previously known as Aiptasia pallida (see Grajales & Rodriguez, 2014), has been proposed as a model organism to study various aspects of the cell biology and physiology of anthozoan-Symbiodinium symbiosis (Weis et al., 2008) (Fig. 1). The use of this cnidarian model system has advanced our understanding of the molecular and cellular mechanism underlying anthozan/Symbiodinium regulation (Davy, Allemand & Weis, 2012). Likewise, the E. pallida model system could benefit investigations regarding the influence of the associated microbiota on physiological, developmental, and disease-resistant host cnidarian phenotypes. Recent data have shown the importance of the presence of photosymbionts (Symbiodinium) on the bacterial community assemblages associated with E. pallida (Roethig et al., 2016); however, we still lack baseline knowledge about the bacterial diversity and assemblages associated with this sea anemone over wide geographical distributions. In the current study, we characterized the composition, structure and specificity patterns of microbial communities associated with the sea anemone E. pallida from worldwide populations using samples from the Caribbean Sea, and the Atlantic and the Pacific oceans. We then compared the microbial composition of natural versus specimens reared in the laboratory over different periods of time as means to document the effect of aquarium conditions in the composition and structure of microbial taxa. Furthermore, we analyze the microbial community assemblage variance as a function of the associated-Symbiodinium species.

Material and Methods

Results

Sequencing efforts produced a total 679,061 reads for the V1–V3 16S rRNA from 49 anemone samples (3–4 replicates per population site) and an average of 13,581 reads per samples. After quality-based filtering a total of 507,882 reads were obtained representing an average of 10,157 reads per sample. To assess differences in community assemblages across samples sequences were clustered into operational taxonomic units (OTUs), based on a similarity level of 97%. Rarefaction analyses suggested that the sequencing effort per population site seemed sufficient for the estimation of OTU diversity for most of the sampled sites; however, the maximum expected species richness was not reached in most of the studied sites (Fig. S1).

A high richness of bacterial species (a total of 12,585 OTUs) was revealed to engage in association with the sea anemones across ocean basins (Table S1). The highest average bacterial OTU richness were obtained from the E. pallida samples obtained from a commercial pet store (1,671 ± 144) and the CC7 clonal population samples reared in the lab (1,358 ± 225), which were approximately four and three times higher respectively than anemones with the lowest OTU richness from a natural population in Hawaii (409 ± 227). Interestingly, no significant differences in alpha diversity based on Shannon-Wiener index were detected across populations (One-way ANOVA and Tukey’s HDS p > 0.1, Table S1).

Classification of sequences on the phylum level revealed that bacterial communities were dominated by Proteobacteria and Bacteroides followed by the phyla Firmicutes and Cyanobacteria (Fig. 2). Proteobacteria dominated the communities (>50%) associated with samples collected from all wild populations (Pacific and Atlantic oceans, Caribbean Sea). However, the bacterial communities associated with some anemones from the Northeastern Atlantic and from an outdoor flow-through sea water system in KML (Florida Keys) were dominated by Firmicutes (nearly 70%; Phyla Indicator Analysis, p < 0.001). Besides these four major represented phyla across samples, most of the other rare phyla were not indicators of a specific association with a particular geographical location (Phyla Indicator Analyses, p > 0.05).

Anemones collected from the KML outdoor sea water system were transported and maintained in re-circulating indoor aquaria using artificial seawater during a period of four months (population referred to as shortlab). During this short period the microbial community underwent a considerable shift of bacterial phyla. Firmicutes decreased from ∼70% to abundances of less than 1%, and Cyanobacteria and Proteobacteria increased in relative abundances, ∼42% and 44% respectively. The bacterial communities associated with anemones reared in the laboratory for six years (clone CC7, from an unknown Florida Keys population), using Instant Ocean Water, were dominated by Proteobacteria (68%), with similar relative abundances to those detected in many other wild populations.

Within Proteobacteria, the class Alphaproteobacteria was most commonly dominant in anemones from the North Pacific (51%), Caribbean (53%), Atlantic (47%), four-month lab reared/shortlab (58%), and six-year lab reared/CC7 (78%) populations (Fig. S2). Gammaproteobacteria and Alphaproteobacteria were equally common in the Eastern Pacific (44% and 40%) and the commercial pet store (48% and 44%) anemone populations. Gammaproteobacteria was found to be most common in Central Pacific anemone samples (53%). The anemones from the outdoor sea water system at the KML that were dominated by Firmicutes also had a very distinct group of Proteobacteria, represented mainly by Deltaproteobacteria (45%).

By examining the composition and structure of the bacterial community associated with E. pallida based on OTUs, we detected considerable differences among populations and geographical locations. The multivariate ordination of the bacterial communities did not exhibit clear grouping of the samples based on geographical origin (nMDS, Fig. 3). However, anemone samples reared in captivity showed less variability and each of the three captivity groups (4-month versus 6-year versus pet store) clustered in its own ordination grouping.

Hierarchical cluster analyses and similarity profile test (SIMPROF) were performed to detect bacterial community structure among the samples independent of their geographical origin. These statistical analyses revealed that very distinct microbial communities characterized most of the samples. While the 49 anemone specimens were collected from ten and four wild and captive populations respectively, SIMPROF analyses detected 32 significant bacterial assemblages (SIMPROF, p < 0.05; Fig. 4). Out of these 32 groupings, fifteen groups were conformed by two samples and only one group was conformed by three samples. The remaining 16 samples were not clustered in any group indicating their unique bacterial assemblage. These clustering analyses produced similar results to the previous nMDS analysis and revealed that the associated bacterial community in wild anemones did not show geographical patterns. On the other hand, bacterial communities from captive E. pallida were more similar to each other. Out of 12,585 OTUs, no single bacterial OTU was shared among all anemones regardless of the geographic origin. The most prevalent OTU (OTU9148: Vibrio tubiashii) was found in 75% of the samples and only 92 OTUs (less than 1% of the total discovered OTUs) were shared by one quarter of all samples (Fig. S3). Based on taxonomic classification, species within the genera Vibrio, Nautella, Ruegeria, Marinobacter, Lentisphaera, and Flaviobacterium were common representatives within the microbial community associated with E. pallida.

Regarding the Symbiodinium composition associated with E. pallida, we detected five Symbiodinium types associated with the studied Exaiptasia samples (Table S2). The majority of the anemones hosted single symbiont species represented mostly by Symbiodinium B1. Population sites, including the Florida KML populations, Puerto Morelos (Mexico) and the CC7 indoor aquarium population, hosted Symbiodinium A4. Two populations, including Bermuda and another Florida population, hosted a mixture of Symbiodinium species composed of B1/C1 and A4/B2, respectively. Interestingly, the two anemones obtained from the local pet store each hosted a different Symbiodinium type (C3 andA4). Comparison of bacterial community and Symbiodinium composition is discussed below.

Discussion

This study revealed the previously uncharacterized bacterial community associated with the cnidarian sea anemone model Exaiptasia pallida at different locations throughout the Northern Hemisphere. This examination of anemone/bacterial association is the first global-scale study of its kind. The results show that a complex and diverse microbial community colonizes the anemone and varies considerably both among and within sampling locations. This indicates a lack of a bacterial core community evolving in intimate association with this cnidarian host. The pattern of community assemblages across sampling locations and geographies does not correlate with the composition of Symbiodinium symbionts associated with the anemones (Table S2). Anemones from seven populations across different oceans (including North and Central Pacific, Caribbean and Atlantic) harbored the same genetic species of Symbiodinium B1 (Grajales, Rodríguez & Thornhill, 2016; Table S2); however, their microbiomes were different as indicated by the multivariate test analyses (Figs. 3 and 4). Furthermore, in some cases anemones from same sampling sites depicted very similar microbiomes but showed to harbor distinct Symbiodinum composition. For instance, two anemones collected from the local petstore each hosted different Symbiodinium species (C3 and A4), however their microbial communities were closer to each other than microbiomes from other sampling sites. This indicates the lack of specific bacterial species associated to a specific photosymbiont type, at least at the level of resolution our study explored. It also highlights the fact that the anemone does not exert high selectivity in shaping the associated microbial community. Although it has been shown that coral reef invertebrate microbiomes correlate with the presence of photosymbionts (Bourne et al., 2013; Roethig et al., 2016), our data indicate that the type/species of photosymbiont (in such case Symbiodinium) does not seem to explain natural variability observed in the microbial community associated with E. pallida across distinct geographical locations. It is expected that the metabolic contribution from Symbiodinium to the anemone host has an effect structuring the associated bacterial community compared with aposymbiotic anemones lacking Symbiodinium, but differential Symbiodinium species composition may not drive further changes of anemone microbiomes.

When examined at the phylum level, bacterial diversity and abundance associated with the cosmopolitan sea anemone E. pallida are broadly conserved across geographic localities with samples containing largely Proteobacteria and Bacteroides. These two phyla have been documented to represent the most abundant bacteria associated with scleractinian coral, a close relative of sea anemones (Ainsworth et al., 2015; Chu & Vollmer, 2016; Rodriguez-Lanetty et al., 2013; Rohwer et al., 2002; Sunagawa et al., 2009). However, the species-level makeup within these phyla differs drastically across space suggesting a high-level core microbiome with local adaptation of the constituents. There was no a single bacterial OTU ubiquitously found in all anemones samples. This finding differs from the postulation, based on a study conducted in the class Hydrozoa (Fraune & Bosch, 2007), that cnidarian hosts actively select, regardless of environmental conditions, and maintain species-specific microbial communities (Fraune & Bosch, 2007). Our results indicate that differences in global and local environmental factors might play important roles sorting the composition of bacterial species that associates with the anemone E. pallida. This paradigm is supported by a recent study which also demonstrated that environmental parameters such as salinity, dissolved oxygen, and ammonium are key drivers in the regulation of the composition and structure of bacterial communities associated with scleractinian corals (Lee et al., 2012). Moreover, changes in the microbial community structure were revealed after rearing specimens of E. pallida in captivity within a period of just four months, supporting the idea that differences in aquatic environments have a strong effect on shaping associated bacterial assemblages at the species level.

The finding of Vibrio tubiashii as the most ubiquitous bacteria associated with Exaiptasia pallida (found in 75% of all samples) was interesting, as this bacterium was recognized as a significant pathogen of several species of bivalve larvae in the 60’s (Tubiash, Colwell & Sakazaki, 1970). The bacteria was also shown to re-emerge and cause vibriosis in shellfish hatcheries on the west coast of North America causing decline in larval oyster production of up to ∼59% in one hatchery (Elston et al., 2008). It has also been demonstrated that strains of this species can also cause disease in massive corals in the Indian Ocean (Sere et al., 2015). Although the functional and ecological significance of the association between this bacteria species and Exaiptasia pallida remains to be discovered, our findings indicate that these widely distributed anemones could be reservoirs of the pathogen Vibrio tubiashii. Potential evidence supporting this hypothesis is the interesting fact that oyster farms have been one of the plausible means responsible for the spread of E. pallida worldwide (Thornhill et al., 2013), due to the close contact between bivalves and sea anemones. This association has also been observed in different natural environments, such as mangrove roots, where E. pallida is often found growing on top of different bivalves (A Grajales, E Rodriguez & M Rodriguez-Lanetty, pers. obs., 2016). However, the mode of transmission of the pathogenic bacteria from the anemones to susceptible host organisms requires investigation.

Unlike in hydrozoans, our findings suggest a lack of coevolution between a sister lineage within Cnidaria (Anthozoa: Actiniaria) and specific bacteria. Within anthozoans, most of the studies exploring the bacterial diversity via culture-independent approaches have been done within the subclass Hexacorallia, more specifically within the order Scleractinia (i.e., stony corals). In this group a number of coral-associated microbial exploratory studies have shown that corals harbor some of the most highly diverse and abundant microbial communities in marine invertebrates after Porifera (Bourne & Webster, 2013; Mouchka, Hewson & Harvell, 2010; Rodriguez-Lanetty et al., 2013; Sunagawa et al., 2009). Evidence supporting a clear co-evolution or co-diversification pattern between prokaryotes and corals is however absent. While some studies have shown species-specific patterns of bacteria/host associations (Littman et al., 2009; Sunagawa, Woodley & Medina, 2010), recent studies using high throughput 16S rRNA gene sequencing have shown that microbial communities associated with scleractinian corals are not species specific (Hester et al., 2016; Meistertzheim et al., 2016; Zhang et al., 2015) and are controlled primarily by external environmental conditions rather than the coral holobiont (Pantos et al., 2015). Although there may be little support of co-evolutionary patterns, it seems that some core bacteria groups might have broadly specialized to associate with scleractinian corals regardless the host species lineage (Ainsworth et al., 2015).

Based on our findings, we propose that more than maintaining a specific species composition of bacteria, E. pallida, and perhaps many other anthozoans, associate with a wide range of bacterial species as long as they provide the same physiological benefits towards the maintenance of a healthy host. To certain extent this explanation is supported by the fact that at higher taxonomic level we detected more similarities across populations in the host-associated microbial structure. The particular bacterial assemblage that may engage in symbiosis with the anemone host will then depend on the existing pool of bacterial species filtered by the environmental conditions of the host habitat, provided that these bacteria belong to a preferred bacterial group with similar ecological functions. It is interesting to note that current global distribution of E. pallida seems the result of recent invasion events based on the lack of host population genetic structure (Grajales & Rodriguez, 2016; Thornhill et al., 2013), and yet we were able to detect a complete turnover of the bacterial community at the species level associated with this invasive host anemone across global scale. Our study highlights a potential role of the environment to delineate the patterns of host/microbial symbiont associations.

Supplemental Information