In the last years the microbial communities (microbiota) associated with the digestive tract of animals have been subjected to wide research interest (Ley et al., 2008; Zhu et al., 2010; Huttenhower et al., 2012). The presence of functional relationship between the host and the associated microbiome (the genes and genomes of the microbiota) has been highlighted, and the new term of hologenome has been proposed to refer to the set of functions (genes) of host and microorganisms associated with it (Zilber-Rosenberg and Rosenberg, 2008). The study of model animals has revealed roles for the microbiome in adaptive immunity development and in host physiology, ranging from mate selection to skeletal biology and lipid metabolism (Ley et al., 2008; Kostic et al., 2013; Du Toit, 2016). For vertebrates, most of the studies on gut microbiota and microbiome have been performed in mammals (i.e., mouse, rat and humans) and in fishes (as the model Danio rerio) (Huttenhower et al., 2012; Kostic et al., 2013). Recently, microbiotas and microbiomes of non-model organisms have started to be investigated with the aim to shed light on animal-associated microbial diversity (Keenan et al., 2013; Mengoni et al., 2013; Cahill et al., 2016) and to potentially discover new biotechnologically important microbial strains (Papaleo et al., 2012; Sanchez et al., 2012).

Sea turtles (Testudines, Reptilia) occur in oceanic and neritic habitats, from the tropics to subarctic waters, and venture onto terrestrial habitats to nest or bask in tropical and temperate latitudes. Sea turtle populations around the world have dwindled and, in many places, continue to decline (Wallace et al., 2010). Caretta caretta L. (Loggerhead Turtle) is distributed throughout the subtropical and temperate regions of the Mediterranean Sea and Pacific, Indian, and Atlantic Oceans. Loggerhead Turtle is classified as Vulnerable A2b in the IUCN Red List (http://www.iucnredlist.org/details/3897/0). The Loggerhead Turtle plays important roles in maintaining marine ecosystem (Bjorndal and Jackson, 2002; Bolten and Witherington, 2003). These roles range from maintaining productive coral reef ecosystems to transporting essential nutrients from the oceans to beaches and coastal dunes. However, in spite of the considerable importance for the study of vertebrates, few studies only are present on microbial communities associated with sea turtles (Ferronato et al., 2009; Sarmiento-Ramírez et al., 2014; Yuan et al., 2015) and no reports on gut microbial communities.

The aim of this work is the characterization, for the first time, of the gut microbiota of the sea turtle C. caretta, to shed a preliminary light on its features with respect to other reptiles and to marine vertebrates. Both feces and intestine samples were taken to have the wider overview of gut microbiota taxonomic composition.

The sequences dataset (Table 1) was deposited in the GenBank database (URL: http://www.ncbi.nlm.nih.gov/bioproject/; Bioproject PRJNA314462, Biosample accessions SAMN04508196-SAMN04508205). Users can download and use the data freely for research purpose only with acknowledgment to us and quoting this paper as reference to the data.

A total 1882390 reads of all samples of C. caretta passed quality filtering sequences (92.8% of total reads) (Table S1). After OTU assignment (Table S1) rarefaction curves obtained reached or nearly reached a plateau, indicating a satisfactory level of diversity sampling (Figure S1).

Concerning the taxonomic composition (Figure 1) feces samples were dominated by members of phyla Firmicutes (66%), Proteobacteria (23%), Bacteroidetes (6.2%). Within the phylum Firmicutes the class Clostridia was the most abundant (63.20%). The intestine samples were dominated by phyla Firmicutes (87%), Proteobacteria (4.2%) and Bacteroidetes (3.4%). Firmicutes were represented by member of the classes Clostridia (43%) and Bacilli (42.5%). This latter was entirely represented (100%) by order Lactobacillales (Table S1). While the most represented bacterial genera among intestine samples were Vagococcus with 42.3%, and among feces were Clostridium XI 21.3%, and Clostridium sensu strict 14.6% (Table S1). Principal Component Analysis on OTU representation (Figure S2) showed that most of the sample were very similar each other. However, notably the two samples of feces from the same individual (T1 and T3, taken in different times) were separated from the rest of the samples. In particular, for T1 and T3, OTU 3, OTU 4, and OTU 5 (all attributed to Clostridiales) collectively contributed for more than 30% of total variance in the differentiation from the other samples (Table S2). Indeed, T1 and T3 were taken from a young female after few days of hospitalization in the recovery center and may mirror the microbiota of a relatively healthy individual in the wild, while the other samples mainly were from animals kept in the recovery centers for longer times. However, we cannot exclude that T1 and T3 microbiota may represent a phase of rapid changes in gut microbiota due to the change in diet (i.e., artificial feeding in the recovery center), which then may bring to a more stable and homogenous microbiota (present in the other samples) after more days. Sampling of more individuals (healthy) would be needed to clarify this issue.

Finally, we inspected which taxa of the microbiota mostly contribute to differentiate feces vs. intestine. Results obtained after Simper analysis showed that the genera mostly contributing to differences were Vagococcus (Bacilli, Enteroccoaceae) with 11.92%, Robinsoniella (Class Clostridia) with 6.29%, this latter represented more in intestine samples, Clostridium XI (Class Clostridia) with 7.37 % and represented more in feces samples (Table S2).

This first investigation on the gut microbiota of C. caretta showed a pattern of taxa which include well know members colonizing vertebrate guts. In particular the most abundant phyla found (Firmicutes and Bacteroidetes) are also abundant in the human gut (Ley et al., 2008) as well as in other land vertebrates and reptiles (Costello et al., 2010; Keenan et al., 2013). However, especially in the feces samples, Gammaproteobacteria were particularly present (more than 15% of total reads) including member of Oceanospirillales, Alteromonadaceae, Pseudomonadaceae, Enterobacteriaceae. Moreover, as suggested by T1 and T3 samples, quite important differences in the microbiota could be detected, which may be related to the influence of hospitalization in most of the sampled animals.

The presented data could be used for comparative analyses of vertebrate gut microbiotas.

KA performed the experiments. GB helped in data analysis. FS, CM performed sampling. AU conceived the work. AM coordinated the work and drafted the manuscript.