Elasmobranch fishes, of the Chondrichthyes clade and teleost fishes, of the Osteichthyes clade, diverged approximately 420 mya , resulting in morphological and physiological differences and here we show they extend to the relationship with the skin microbiome. We found host-species to harbor unique microbial symbiotic communities, both taxonomically and at the functional gene level. Species-specific microbiomes are a common pattern in nature, occurring in many host organisms, including, coral reef fishes, nasonia wasps, mosquitos, mice, and drosophila [10, 24]. Species specificity of the functional gene composition of the microbiome communities is described for a few marine organisms, including the skin of the common thresher shark  and an algal species, Ulva australis . Here we show the microbial functional genes are specific to a further eight marine host species.
Microbiome community similarity is predicted to decrease with the increasing evolutionary divergence of host organisms . Therefore, we predicted that host species microbiomes would be more similar among clades than across clades. We found that teleost fish exhibited a lower microbiome phylogenetic distance relative to the inter-clade comparison, but this was not the case for Elasmobranchs. However, some species within each clade had microbiomes that were more similar to species across clade boundaries, thus a non-significant comparison. For functional genes, elasmobranchs exhibited a lower microbiome functional dissimilarity within clade compared with between clades, but this was not the case for teleost fish microbiomes. Most phylosymbiosis studies have not compared across clades [10, 17], suggesting that eco-environmental conditions may be reducing the effect of host phylogeny on the relationship of host and microbiomes.
We tested for the effects of host evolution on the microbiome, predicting that species with a more recent common ancestor have a microbiome that is more similar in composition. With the methods used here, we observed that elasmobranch fishes exhibited increased microbiome divergence with increasing host difference. Teleost fishes, however, exhibited no relationship in microbiome divergence and host difference. In fact, for teleost, the slope trended in the opposite direction relative to predictions (albeit a non-significant slope was reported). The phylogenetic assessment of the elasmobranch species based on the COX1 gene, suggests they have a more recent common ancestor relative to the teleost fish species. However, we note caution in the interpretation of the results based on the COX1 gene, as the mitochondrial DNA among Chondrichthyes fish species accumulates nucleotide differences at rates much slower than their Osteichthyes fish counter-parts . Therefore, we suggest that the elasmobranch species emerged earlier than the teleost species and that the microbiome and elasmobranchs have evolved a phylosymbiosis relationship.
Teleost fish, however, lack a consistent phylosymbiosis relationship, which may be the result of convergent evolution of traits in skin features selecting for specific microbiome inhabitants. The increasing microbiome similarity (for both phylogeny and functional genes of the microbiome) with increasing host distance suggests convergent evolution for traits that fish use to select and maintain a microbiome [44, 45]. Similarly, Chiarello et al (2018) found weak support for phylosymbiosis in coral reef fishes but did not analyze the microbial functions. Teleost fishes are covered in mucus of varying chemical compounds and thickness [28, 46], which influences microbiome composition depending on the presence of host immunological factors and mucus chain sugar residues . In addition, the epidermal mucus from teleost harbor anti-microbial properties [48, 49]. Thus, the microbiome requires similar functional genes to utilize the mucus and evade the anti-microbial properties, possibly leading to the convergent evolution of the microbiome on teleost fish.
The lack of a pattern for phylosymbiosis in the gene required to live on elasmobranch fishes may occur because mucus is not a selective mechanism. There are low amounts of mucus excreted onto the skin surface , except for stingrays . The microbes on the elasmobranchs are not utilizing mucus but using the skin surface as a habitat. In this case, the microbes require unique traits to attach and establish a biofilm on each of the elasmobranchs. Therefore, the microbial species and elasmobranch host may have evolved together.
The microbial relationship patterns may occur because common ancestor of elasmobranchs and teleost fishes have maintained symbiotic interactions with specific microbial groups, which remained conserved among some species, while in others, these symbionts were lost. Such processes have been hypothesized for the convergence observed in some teleost fish and mammal gut microbiome , in which teleost fishes formed distinct symbiotic relationships, which remained conserved as mammals radiated from the bony fish clade. Similarly, human and old-world monkey gut microbiomes did not show phylosymbiosis as a result of host adaption for an omnivorous diet . The diet adaptive process results in the acquisition of microbiome symbionts, which evolved before the evolution of the host organism. Groussin et al. (2017) showed that host organisms which share a common ancestor more recently, had stronger patterns of phylosymbiosis while increasing time since shared ancestry corresponded with a decrease in phylosymbiosis. They attribute this relationship to dietary switching, which has led to acquiring microbial symbionts that evolved independently of the host organism, therefore host species with common ancestors that share dietary constraints have more similar gut microbial communities. Similar results were found by Muegge et al. (2011) which reported that the host diet was a better predictor of microbiome composition than was phylogenetic placement. By mapping conserved gene sequences on the tree of life, we observed conserved and specific microbial classes across the fish clades. The divergent microbial species suggest a possible co-evolutionary interaction between microbial species and the host.
The lack of consistency in the relationship of microbiomes across clades could also be the result of a geographic effect. Capture sites of blacktip reefs sharks (Carcharhinus melanopterus) accounted for high variation in microbiome composition . However, the observed patterns are not consistent with the location of sampling in our study. For instance, the stingrays were collected in the San Diego region, as were the leopard sharks; however, the stingray microbiomes were more similar to the whale shark microbiomes, which were collected in La Paz, Mexico. If the environment is a stronger driver of microbiome composition than phylogeny, we would expect all teleost fish and the stingray to have a similar microbiome, as all samples were collected in San Diego, USA. The trophic position of the host influences the microbiome structure . However, whale sharks, which are filter feeders, and stingrays, which are benthic carnivores, had similar microbiomes. Whereas, stingrays and leopard sharks both consume benthic invertebrates, but their microbiomes were dissimilar. The similarity of thresher shark and killifish microbiomes further contradicts the trophic hypothesis. Therefore, in our study suggests that the taxonomy of the elasmobranch microbiome follows the phylosymbiosis model, while the teleost microbiomes appears to be converging.