Bacterial and fungal assemblages of the gut and carapace of loggerhead sea turtles

doi:10.21203/rs.3.rs-3574161/v1

Download PDF

Research Article

Bacterial and fungal assemblages of the gut and carapace of loggerhead sea turtles

https://doi.org/10.21203/rs.3.rs-3574161/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

The research on microbial communities of marine megafauna species represents a valuable source of knowledge that could be used for species conservation programs. Loggerhead sea turtles are marine flagship species with a currently thriving population in the Mediterranean Sea due to strong conservation efforts. In our study we aimed to further understand their biology in relation to their associated microorganisms. We investigated bacterial and fungal community structure of cloacal, oral swabs and carapace biofilm samples obtained from 18 juvenile, subadult and adult turtles as well as 8 respective enclosures, during a period of 3 years, by amplicon sequencing of 16S rRNA gene and ITS2 region of nuclear ribosomal gene.

Results

Our results reveal a trend of decreasing diversity of distal gut bacterial communities with the age of turtles. Notably, Tenacibaculum genus shows higher relative abundance in juveniles than in adults. Differential abundance of genera like Tenacibaculum, Moraxellaceae, Cardiobacteriaceae, and Campylobacter was observed in both cloacal and oral samples in addition to having distinct microbial compositions such as Halioglobus genera in oral samples. Fungal communities in loggerheads' cloaca were diverse and varied significantly among individuals, differing from those of tank water. However, we could not identify discernible patterns related to age, sex, or clinical status due to study limitations.

Conclusion

This study provides a comprehensive view of the loggerhead sea turtle bacterial microbiota and marks the first report of distal gut fungal communities studied via amplicon sequencing. Our findings expand the known microbial diversity repertoire of loggerheads, highlighting interesting taxa specific to individual body sites. This contributes to establishing a baseline understanding of the loggerhead bacterial and fungal microbiome.

Microbiota

microbiome

fungal diversity

reptiles

Adriatic Sea

conservation

Microbial communities associated with animal hosts play crucial roles in various aspects of the host’s development, physiology, immune response, metabolism, and reproduction, and may have an impact on the host’s evolutionary potential [1, 2]. While the importance of epibiosis of sea turtles with macro-epibionts (> 1 mm) such as barnacles has long been recognized [3], research on the microbial communities inhabiting both the external (epibionts) and internal (endobionts) environments of sea turtles has only recently gained attention [4–6]. Sea turtles hold a unique ecological role as keystone species, connecting terrestrial and coastal habitats, but they are also highly vulnerable to anthropogenic threats, such as climate change, disruption of feeding and breeding habitats, egg poaching, and accidental bycatch [7–9]. To lessen some of the pressure sea turtles face today, global conservation efforts have focused on safeguarding female turtles and nesting areas, and rehabilitating injured turtles [8]. Motivated by the aim of enhancing the rehabilitation and conservation of wild animals and their associated microbiomes, the studies of sea turtle gut, skin, egg, and nest microbiomes has become a forefront in sea turtle conservation research, building upon cultivation-based and pathogen-oriented research in the veterinary domain [10, 11].

Loggerhead sea turtles (Caretta caretta) are a widely distributed species and are classified as vulnerable by the IUCN Red List of threatened species [12]. However, the Mediterranean subpopulation of loggerhead sea turtles is considered to be of least concern due to successful conservation efforts [8, 12]. Currently, loggerhead sea turtles’ microbiota is the second most studied, preceded only by green turtles [6]. Previous studies that used culture-dependent approaches have identified most common pathogens associated with mucosal surfaces, skin lesions (such as bacterial families Aeromonadaceae, Pseudomonadaceae, Enterobacteriaceae), and hatchling failure (Fusarium spp.), with the presence of antibiotic resistance genes indicating loggerheads as sentinels of antibiotic pollution in the Mediterranean [11, 13–21]. Recent investigations using next-generation sequencing approaches to study the loggerhead microbiota shed light on the bacterial community structure and composition of the gastrointestinal tract [5, 22, 23], the impact of rehabilitation on mucosal bacteriomes [23, 24], the effects of plastic pollution on the gut bacteria [25], variations in microbial communities driven by localities [26, 27] or turtle anatomy [27, 28], and host-microbial coevolution patterns [29]. On the other hand, the fungal communities associated with marine reptiles, including sea turtles, have received limited attention using cultivation-independent approaches, despite the vulnerability of sea turtles to infections caused by Fusarium spp. during early development [17]. Recent work by Guo et al. [30] provided initial insights into the fungal communities found on carapace (healthy and ulcerated), in faeces, and in the seawater of green turtle juveniles undergoing rehabilitation, however, comprehensive surveys of endobiotic fungal communities in loggerhead sea turtles have not yet been conducted. Given the ecological significance of loggerhead sea turtles in the Mediterranean basin ecosystem, their role as sentinels for pollution, and their potential to act as vectors for zoonotic diseases, a comprehensive approach including eukaryotic microorganisms is necessary to understand the loggerhead sea turtle microbiota. This knowledge will contribute to the advancement of current conservation practices and future microbial stewardship efforts [31].

In this study, our objective was to examine the bacterial and fungal communities associated with loggerhead sea turtles incidentally caught or found, and sampled during non-winter months in the period 2019–2021. We analysed the bacterial communities in the cloacal, oral, and tank water samples using amplicon sequencing targeting the V3-V4 (V34) region of 16S rRNA gene. Additionally, we aimed to provide an overview of the fungal communities of cloacal and tank water samples by amplicon sequencing of the ITS2 region of nuclear ribosomal genes. Furthermore, to gain insight into loggerhead epizoic and endozoic bacterial community dynamics, we investigated corresponding carapace samples out of which some were analysed as a part of our earlier studies [27]. By combining these datasets, we provide a comprehensive overview of the environmental, surface, and internal microbiota of the loggerhead turtles, establishing a baseline for future holobiont approaches to studying the loggerhead sea turtles.

Endozoic and tank water bacterial communities

Altogether, 50 endozoic and water samples (plus one negative control) were sequenced, and 7,110,067 high quality sequences were obtained (median frequency per sample was 107,522, min. 2, max. 911,443). Denoising yielded a total of 11,105 ASVs. After filtering mitochondrial and chloroplast sequences, the number of ASVs decreased to 10,946. Forty-three samples yielded enough high-quality reads (at minimum sequencing depth above 10,000) for downstream analyses (19/21 cloacal, 16/20 oral, and 8/9 tank water samples).

The highest number of observed features (OF) and phylogenetic diversity (Faith’s PD) was found in tank water samples (median OF = 603, IQR = 450; median Faith’s PD = 42.8, IQR = 53.9), followed by oral samples (median OF 473, IQR = 428; median Faith’s PD 40.2, IQR = 26.2), and cloacal samples (median OF = 308, IQR = 235; median Faith’s PD = 28.4, IQR = 20) (Fig. 1a). ASV richness and evenness (Shannon’s index) was highest in oral samples (median = 6.28, IQR = 1.33), followed by cloacal (median = 5.62, IQR = 1.66), and tank water samples (median = 5.44, IQR = 1.67) (Fig. 1b). Alpha diversity showed significant differences among sample sites (Kruskal-Wallis rank sum test) for OF and Faith’s PD (p-value < 0.01), while pairwise sample site comparisons (Wilcoxon rank sum test) showed differences between cloaca vs. oral samples and cloaca vs. tank water (q-value < 0.05). Shannon’s index showed weak statistical difference among samples sites (p-value = 0.04). Pearson correlation on alpha diversity indices against size of the turtle (CCL) in individual sample site groups showed strong negative correlation between cloacal samples and OF and Faith’s PD (R = -0.67, p-value = 0.002 and R = -0.59, p-value = 0.007, respectively) (Fig. 1c), while Shannon’s index showed weaker negative correlation with CCL (R = -0.53, p-value = 0.021). Oral and tank water alpha diversity did not show any correlation effects. Further, within cloacal samples, differences were detected (Kruskal-Wallis p-value < 0.05) in age range (OF, pairwise Wilcox: adult vs. juvenile q-value = 0.024; Faith’s PD, pairwise: adult vs. juvenile q-value = 0.024) and sex of the turtles (OF, pairwise Wilcox: male vs. ND q-value = 0.004; Faith’s PD, pairwise: male vs. ND q-value = 0.018), which are directly related to turtles’ size.

Bacterial communities among samples sites differed based on PERMANOVA for Bray-Curtis (R² = 0.09, Pr(> F) = 0.001), Jaccard (R² = 0.08, Pr(> F) = 0.001), unweighted and weighted UniFrac (R² = 0.12, Pr(> F) = 0.001; R² = 0.13, Pr(> F) = 0.001, respectively), and robust Aitchison distances (R² = 0.24, Pr(> F) = 0.002). Pairwise PERMANOVA detected differences between all sample site pairs for most beta diversity metrics except robust Aitchison where only cloaca vs. tank water and tank water vs. oral samples pairs were significantly different (Pr(> F) ≤ 0.002). Highly ranked ASVs impacting the distribution of samples on rPCA biplot were assigned to Gammaproteobacteria, order Oceanospirillales, Shewanella algae, Vibrio sp., and NS3a marine group (Fig. 1d). Within cloacal samples, significant differences in bacterial communities were detected between adults vs. juveniles (Bray-Curtis, Jaccard, and unweighted UniFrac Pr(> F) < 0.05) and ND vs. males (all distances except robust Aitchison Pr(> F) < 0.05) or females (Jaccard and weighted UniFrac Pr(> F) < 0.05). In oral samples there were differences between early vs. late hospitalization duration (all distances except robust Aitchison Pr(> F) ≤ 0.05) and early vs. mid (weighted UniFrac Pr(> F) < 0.05) hospitalization durations.

The bacterial communities of all samples were primarily represented [average relative abundance (RA) > 1% across sample site groups] by phyla Proteobacteria and Bacteroidota, Firmicutes and Campilobacterota in cloacal and oral samples, Verrucomicrobiota in cloacal and tank water samples, Bdellovibrionota in oral and tank water samples, Fusobacteriota and Spirochaetota in cloaca, Actinobacteriota in oral cavity, and Patescibacteria in tank water (Table S3). Proteobacteria consisted of 17 families (and corresponding taxa) with average RA above 1% in at least one sample site group: Cardiobacteriaceae (Cardiobacterium spp., Suttonella spp., and uncultured representatives), Shewanellaceae (Shewanella algae), Rhodobacteraceae (unclassified), Moraxellaceae (Acinetobacter spp., Psychrobacter spp.), Vibrionaceae (Vibrio spp., Vibrio fluvialis at 30% RA in 16S0074O), Pasteurellaceae (unclassified at 24% RA in 16S0074C), Neisseriaceae (unclassified at 2–10% RA in 16S0074C, 16S0092C, 16S0141C, 16S0117O; Stenoxybacter spp. at ~ 7% RA in 16S0087C, 16S0089C, 16S0089O), Pseudoalteromonadaceae (Pseudoalteromonas spp.), Saccharospirillaceae (Littoribacillus spp., Oceanobacter spp., and uncultured), Spongiibacteraceae, Sedimenticolaceae, Alteromonadaceae (Glaciecola spp. in tank water), Pseudomonadaceae (Pseudomonas at 8% RA in 16S0088O), Colwelliaceae (Colwellia spp.), Nitrincolaceae (Marinobacterium spp. and M. marisflavi), Oleiphilaceae, and Marinomonadaceae (uncultured Marinomonas spp.). Bacteroidota consisted of seven families (> 1% average RA): Flavobacteriaceae (unclassified, Flavirhabdus spp., Tenacibaculum spp.), Cryomorphaceae (uncultured), Saprospiraceae (Phaeodactylibacter spp.), NS9 marine group (uncultured Flavobacterium), Weeksellaceae (Moheibacter sp.), Bacteroidaceae (Bacteroides spp.), and Marinifilaceae (uncultured). Firmicutes consisted of two families: Lachnospiraceae, and Exiguobacteraceae at 33% RA in 16S0088O. Campilobacterota consisted of three families: Arcobacteraceae, Campylobacteraceae (Campylobacter spp.), Helicobacteraceae (Helicobacter sp. at 19% RA in 16S0092C). Fusobacteriota consisted of two families: Fusobacteriaceae and Leptotrichiaceae, and Spirochaetota had the Leptospiraceae family with average RA above 1%. Verrucomicrobiota had no families above 1% average RA but in 16S0087C Akkermansiaceae were present at 8% RA. Bdellovibrionota, Actinobacteriota, and Patescibacteria had no families above 1% average RA across at least one sample site group.

Cloacal bacterial communities differed between age groups based on PERMANOVA for Bray-Curtis (R² = 0.15, Pr(> F) = 0.018), Jaccard (R² = 0.14, Pr(> F) = 0.008) and unweighted UniFrac distance (R² = 0.16, Pr(> F) = 0.013). Pairwise PERMANOVA showed significant differences only between adults and juveniles for all above-mentioned diversity measures (Pr(> F) ≤ 0.009). The differences between adults and juveniles are observed in community composition and structure as well: phylum Verrucomicrobiota appears more often in juveniles, juveniles also have higher RA of Rhodobacterales (Alphaproteobacteria), Cardiobacterales (Gammaproteobacteria), Kineosporiales (Actinobacteriota), Oligoflexales (Bdellovibrionota), Arcobacteraceae (Campylobacterales), and more Flavobacteriales (Bacteroidota) than adults. On the other hand, adults carry more Bacteroidales (Bacteroidota), Pasteurellales (Gammaproteobacteria), Helicobacteraceae and Campylobacteraceae (Campylobacterales), and Leptotrichiaceae (Fusobacteriales) (Table S3, visualizations available at Github). The low number of samples (≤ 10) in each age group prevented us from further differential abundance analyses.

Out of pathogenic bacteria reported in sea turtles as reviewed in [11], we detected Cardiobacterium spp. and Suttonella spp. (Cardiobacteriaceae) in cloacal and oral samples, Vibrio spp. mostly in oral and tank water samples, Citrobacter spp. (Enterobacteriaceae) at 2–5% RA in several cloacal and one oral sample, and Morganella spp. (Morganellaceae) at 12% RA in one cloacal sample (16S0073C). Other pathogenic bacteria were detected, albeit sparsely and at low RA (rarely reaching 1–3% RA in individual samples).

Differential abundance analysis detected 33 significantly DA features (post adjusting for multiple testing) out of which two were differentially abundant in cloaca (ASV467 Rhodobacteraceae and ASV4007 Shewanella algae), and three in oral samples (ASV9794 Truepera sp., ASV6166 and ASV467 both belonging to Rhodobacteraceae) (Fig. 2a). Tank water had 28 DA ASVs when tested against cloaca or oral samples, most of which belonged to typical marine taxa (Cryomorphaceae, NS3a marine group, SAR406 clade, Rhodobacteraceae, Nitrincolaceae, etc.). When collapsed to species level, 75 significantly DA taxa were detected, out of which 13 in cloaca and 15 in oral samples (Fig. 2b). Depending on the tested sample site pair several features would be DA in both oral and cloacal samples when compared to tank water e.g., genera Marinifilum, Tenacibaculum, and Labrenzia, members of Cardiobacteriaceae and Comamonadaceae families (Fig. 2b). (p-value ≤ 0.05), while double asterisk (**) indicates significance after adjusting for multiple testing (q-value ≤ 0.05).

Negative control sample yielded 478 reads (23 ASVs) out of which 46% belonged to genus Cupriavidus (ASV8329). The rest of the reads were assigned to Phyllobacterium (9%, ASV6495), Sphingomonas (5%, ASV7577), Archaea Nitrosopumilaceae (4%, ASV14), Lactobacillus (3%, ASV4269), Lachnoclostridium (3%, ASV4451), Pseudomonas (3%, ASV10106), Archaea ANEM 2a-2b (3%, ASV26), Caulobacteraceae family (3%, ASV6031), Bradyrhizobium (2%, ASV6541), Methylobacterium jeotgali (2%, ASV6409), and Archaea Nitrosopumilaceae family (ASV18), Paracoccus solventivorans (ASV6950), Stenotrophomonas (ASV10578), Blastocatellaceae JGI 0001001-H03 metagenome (ASV131), Lactobacillus iners (ASV4273), Archaea order Methanomassiliicoccales (ASV62), Rikenellaceae RC9 gut group (ASV713), family Microscillaceae (ASV1470), family Chitinophagaceae (ASV778), Sphingorhabdus (ASV7595), Cloacibacterium (ASV2845), Rhodomicrobium (ASV6512) (all below 1%). Fourteen ASVs from negative control sample were detected in other samples as well. Low read samples (up to 5000 reads and all excluded from analyses) had 45% of total reads assigned to negative control Cupriaviudus ASV in comparison to high read samples that had 0.007%, followed by Caulobacteraceae and Phyllobacterium at 7% and 3%, respectively (compared to high read samples that had < 0.0001% reads assigned to those ASVs). The rest of negative control ASVs followed a similar pattern where they were comprising a substantial proportion of reads in low read vs. high read samples (Table S3). Top 3 ASVs in negative control (Cupriavidus, Phyllobacterium, Sphinogomonas) comprised 61% of total reads in negative control and 53% of total reads in low read samples (only 0.008% of total reads in high read samples).

Endozoic and tank water fungal communities

Overall, 29 samples (plus one negative control) were sequenced, and 2,208,729 high quality sequences were obtained (median frequency per sample was 69,203, min. 11,722, max. 150,142). Denoising yielded a total of 9,547 ASVs. Based on alpha rarefaction curves, 27 samples yielded enough high-quality reads (at minimum sequencing depth at 25,000 reads) for downstream statistical analyses requiring rarefied data (19/20 cloacal, and 8/9 tank water samples). All samples except negative control were used in compositional data analyses.

Tank water samples had significantly higher number of observed features and phylogenetic diversity (median = 674, IQR = 97.8; Faith’s PD median = 98.0, IQR = 10.5) than cloacal samples (median = 509, IQR = 372, Faith’s PD median = 66.6, IQR = 51.7) (Fig. 3a) based on Kruskal-Wallis rank sum test (OF p-value = 0.008; Faith’s PD p-value = 0.007). Shannon’s diversity was not detected as significantly different between sample sites (tank water median = 8.66, IQR = 1.31; cloaca median = 6.76, IQR = 2.58) (Fig. 3b). There were no significant differences in alpha diversity values within cloacal or tank water samples when tested for age range, sex, or rehabilitation duration. The only beta diversity metric that showed significant differences between tank water and cloaca was unweighted UniFrac (R² = 0.06, Pr(> F) = 0.042). Unweighted UniFrac PCoA showed unclear sample groupings, however, there is a separation of cloacal and corresponding tank water samples on PCA1 axis (Fig. S1a, corresponding samples connected with dashed lines). Compositional data analysis rPCA did not show clear separation of samples sites, and top ranked ASVs belonged to Preussia flanaganii, genus Tetracladium, order Xylariales, genus Rhizoctonia, and family Nectriaceae (Fig. S1b).When collapsed to species level, cloacal and tank water samples each had five differentially abundant taxa detected. In cloacal samples DA taxa were Leptodiscella, Elaphomyces, Sclerocleista, Gliomastix, Thelephora; and in tank water they were Cladosporium, unidentified Leotiomycetes, unidentified Sordariomycetes (Chaetosphaerilaes), Nectria, and Humicola - although none of them were statistically significant after correction for multiple testing (Fig. 3c). Cloacal and tank water fungal communities were represented mostly by phyla Ascomycota (average RA ± standard deviation in cloaca and tank water = 57 ± 15% and 53 ± 9%, respectively), Basidiomycota (19 ± 9% and 22 ± 3%), Glomeromycota (7 ± 6% and 10 ± 12%), Mortierellomycota (2 ± 2% and 3 ± 2%) unidentified Fungi (13 ± 20% and 11 ± 7%) (Fig. 3d). Within Ascomycota, most classes were found in cloaca and/or tank water above 1% average RA: Sordariomycetes (family Nectriaceae in cloaca and tank water, order Xylariales in cloaca), Pezizomycotina cls Incertae sedis (Pezizomycotina fam Incertae sedis), Leotiomycetes (family Helotiaceae and unclassified members of order Helotiales), Dothideomycetes (families Phaeosphaeriaceae in cloaca, Sporormiaceae in both, and Lophiostomataceae in tank water), Pezizomycetes (family Pyronemataceae), Eurotiomycetes, Saccharomycetes (cloaca). In Basidiomycota these were classes Agaricomycetes (families Glomeraceae, Serendipitaceae, Russulaceae in both, Sebacinaceae and Inocybaceae only in tank water), and Tremellomycetes (Piskurozymaceae in cloaca). In Glomeromycota and Mortierellomycota, classes Glomeromycetes (family Glomeraceae) and Mortierellomycetes (family Mortierellaceae) were present above 1% average RA in both cloaca and tank water, respectively (Table S4). Out of all fungal ASVs, 18% could not be identified below phylum level (Table S4).

Clinically relevant taxa such as Fusarium spp. (family Nectriaceae), Cladosporium spp. (family Davidiellaceae), and Penicillium spp. (family Trichocomaceae) were detected sporadically across samples at or around 1–2% RA, with Fusarium spp. reaching 8% RA in ITS0094C sample. Two ASVs assigned to Nectriaceae and Solicoccozyma aeria were shared across 90% of all samples (26/29), while 90% of cloacal samples (18/20) shared an ASV assigned to Nectriaceae, and 100% of tank water samples (9/9) shared eight ASVs assigned to Nectriaceae, Bartaliniaceae, Serendipitaceae, Clonostachys sp., Tetracladium sp., Solicoccozyma aeria, Ilyonectria robusta, and Pseudaleuria sp. (Table S4).

Negative control sample yielded 60,329 ITS2 reads (with 1,131 fungal ASVs) most of which belonged to Ascomycota (90%) (Table S4). Out of all negative control fungal ASVs 424 were detected in cloacal and tank water samples making up 71% of total reads in negative control, 36% in cloacal and 32% of total reads in tank water samples. Maximum RA of an individual ASV in negative control reached 2% (assigned to Inocybe maculata), with top 14 ASVs at 1–2% RA.

Epizoic and endozoic bacterial communities

After trimming epizoic, endozoic, and tank water sequences to V4 region of 16S rRNA gene (65 samples in total) and denoising, 10,072,866 high quality sequences were merged across all sequencing events (median frequency per sample was 126,629, min. 0, max. 946,183) and yielded 13,263 ASVs. Due to reduced resolution after trimming the longer V34 region to shorter V4 region, the number of ASVs for cloacal, oral and tank water samples was expectedly lower (7,725 in V4 vs. 11,105 in V34 sequences).

Carapace samples had the highest median species richness and diversity (OF median 732, IQR = 420, Faith’s PD median = 45.7, IQR = 25.2; median Shannon’s index = 5.94, IQR = 2.22), which was followed by tank water, oral and cloacal samples (Fig. 4a, 4b). The differences between sample sites’ alpha diversity were detected as statistically significant (Kruskal-Wallis rank sum test), however pairwise comparisons showed differences only in Faith’s phylogenetic diversity and OF: between cloaca and all other sample sites (except tank water based on OF), oral cavity and carapace. All beta diversity metrics showed significant differences between body sites for Bray-Curtis (R² = 0.13, Pr(> F) = 0.001), Jaccard (R2 = 0.10, Pr(> F) = 0.001), unweighted and weighted UniFrac (R² = 0.16, Pr(> F) = 0.001; R² = 0.18, Pr(> F) = 0.001, respectively), and robust Aitchison distances (R² = 0.09, Pr(> F) = 0.001). Pairwise PERMANOVA detected differences between all sample site pairs (Pr(> F) < 0.01). Highly ranked ASVs impacting the distribution of samples on rPCA biplot belonged to Pseudoalteromonas sp., Vibrio sp., Shewanella sp., uncultured Saccharospirillaceae, uncultured Marinifilum, uncultured Cardiobacteriaceae (Fig. 4c).

The taxonomic composition of endozoic samples and tank water resembled the one detected with V34 sequences (Table S5). The carapace samples mostly consisted of families Pseudoalteromonadaceae (Pseudoalteromonas spp.), Rhodobacteraceae (unclassified and Paracoccus spp.), Flavobacteriaceae (unclassified and Tenacibaculum spp.), Vibrionaceae (Vibrio spp.), Alteromonadaceae (Alteromonas spp.), Moraxellaceae (Acinetobacter spp., Psychrobacter spp.), Saccharospirillaceae (unclassified and Oceaniserpentilla spp.), Arcobacteraceae, Sphingomonadaceae, Colwelliaceae (Colwellia spp.), Cyclobacteriaceae, Shewanellaceae (Shewanella spp.), Spongiibacteraceae (BD1-7 clade), Cellvibrionaceae, JGI 0000069-P22, and Pseudomonadaceae (Pseudomonas spp.) (Table S5).

Differential abundance analysis on ASVs (structural zeros excluded) detected 17 significantly DA ASVs across all sample sites (Fig. 5). A member of Rhodobacteraceae family (v4ASV8861) and Halioglobus sp. (v4ASV11062) were consistently DA in oral samples while the same goes for Cardiobacterium sp. (v4ASV10814) and a member of Enterobacteriaceae family (v4ASV11383) in cloaca. In carapace samples a member of Flavobacteriaceae family (v4ASV2446), cyanobacteria Leptolyngbya sp. (v4ASV4472), Ahrensia sp. (v4ASV8480), Sphingomonadaceae (v4ASV9910), Erythrobacter sp. (v4ASV9994), Gammaproteobacteria (v4ASV10111), Alteromonas sp. (v4ASV10327), Arenicella sp. (v4ASV10634), Psychrobacter sp. (v4ASV12017) and Vibrio sp. (v4ASV12307) were DA relative to other sample sites, while Marinomonas sp. (v4ASV11704) was DA abundant in carapace tested against endozoic samples, but not tank water. In tank water a member of Flavobacteriaceae (v4ASV2446) and NS3a marine group (v4ASV2817) were consistently DA (Fig. 5).

Differential abundance analysis on taxa collapsed to species level (structural zeros excluded) detected 58 DA taxa (Fig. S2). Most taxa were found to be DA in carapace (Erythrobacter spp., Thalassospira spp., Leptolynbya sp., etc.) or tank water samples (e.g., Cryomorphaceae spp., NS3a marine group, Marinomonas spp.). Taxa collapsed to Tenacibaculum, Moraxellaceae, Cardiobacteriaceae, Marinifilum, Campylobacter, and Bacteroidales, were consistently DA in both cloaca and oral cavity in comparison to carapace and tank water. Bacteroides, WCHB1.41, Rikenellaceae RC9 gut group, Cardiobacterium, Shewanella were consistent in cloacal samples, while Halioglobus and Truepera were DA in oral samples (Fig. S2).

This study provides an overview of the bacterial and fungal communities in loggerhead sea turtle juveniles, subadults, and adults primarily found stranded, floating, or caught as bycatch in the Adriatic Sea. Additionally, we aimed to investigate bacterial communities in carapace samples in relation to the gastrointestinal tract and tank water to understand loggerhead microbiota dynamics. The carapace exhibited the highest bacterial diversity, followed by oral samples, influenced by the tank water environment, and then cloacal samples. Each sampling site had distinct microbial communities and cloacal bacterial diversity negatively correlated with turtle size and age. Conversely, fungal communities in the cloaca were distinct from tank water and showed high heterogeneity among individual turtles, with no discernible patterns related to age or sex categories.

Cloacal bacterial diversity and structure changes with the turtle age

Similarly to previous studies on cloacal microbiota in Adriatic loggerhead sea turtles, we observed changes in bacterial richness, diversity, and structure, negatively correlating with loggerheads' CCL and, consequently, their age [23]. According to research on vertebrate gut microbiomes, it is expected that the bacterial diversity increases with body size in animals with complex digestive systems, like ruminants, and decreases in animals with simple guts, often omnivores or carnivores [32]. Loggerhead sea turtles are omnivores and exhibit an ontogenetic shift from oceanic-pelagic habitats as juveniles to neritic-benthic feeding grounds as adults. However, in the Mediterranean Sea they employ an "amphi-habitat strategy", where juveniles, subadults and adults share feeding grounds and similar feeding behaviours [33, 34]. Differences in diet, jaw size, bite force, and diving ability between juveniles and adults may result in distinct diets and digestive physiologies, potentially reflected in cloacal bacterial communities as increased richness. Studies on loggerhead diet show no significant differences in stomach contents among adults, subadults, and juveniles, but it's worth noting that softer prey like tunicates or jellyfish may not be as detectable due to easier digestion and lack of hard remains in morphology-based studies [33].

In this study, juveniles exhibited higher abundance and more frequent presence of Flavobacteriaceae and Tenacibaculum spp. in oral and cloacal samples than subadult and adult turtles. Tenacibaculum spp. are marine pathogens possibly carried by cnidarians and ctenophores as vectors [35, 36], suggesting a higher proportion of soft prey in juveniles' diet. This study was conducted during non-winter months when jellyfish populations increase in warmer seasons potentially influencing prey availability. Adults likely have access to these prey types but also consume more challenging prey inaccessible to juveniles. Notably, in other studies Tenacibaculum spp. were highly abundant on loggerhead skin as well [27], indicating their propensity to inhabit marine vertebrate surfaces and gastrointestinal tracts, irrespective of animal’s diet. While the Mogibacteraceae family detected in faeces correlated with CCL in previous research [23], we did not detect it in our data, possibly due to different taxonomy databases being utilized (Greengenes vs. SILVA in our study). BLAST analysis showed ASVs assigned as Peptostreptococcales-Tissieralles closely related to Mogibacterium kristiansenii but it still did not exhibit a CCL correlation.

In contrast to our previous study [24], the loggerheads in this research exhibited greater bacterial diversity in the gastrointestinal tract. The prior study involved mostly winter-trawled individuals, showing a higher relative abundance of Verrucomicrobiota, (previously assigned to Kiritimatiellaeota), in cloacal samples. However, in this study, only two individuals were caught by trawling in winter, and Verrucomicrobiota were more prevalent in oral samples, accounting for less than 10% of the relative abundance per sample. The increased bacterial diversity in the non-trawled, non-winter group can be attributed to a higher number of juveniles included in the study and seasonal behavioural shifts characterized by heightened turtle activity and dietary variety in spring, summer, and autumn compared to winter [24, 37].

Loggerhead cloacal communities are a promising source of novel Campilobacterota

Members of the Campilobacterota phyla, particularly Arcobacteriaceae, Campylobacteraceae, and Helicobacteraceae, have recently been found to form unique, cold-adapted communities in ectothermic reptiles [38]. As expected, and based on their physiology and lifestyle, the Arcobacter genus, which can thrive at atmospheric oxygen levels and prefers lower temperatures, was present in all sample sites. However, it was more abundant in oral than cloacal samples and was rarely detected in tank water. The Helicobacter and Campylobacter genera, which are vertebrate-associated, were observed in cloacal and occasionally oral samples. Notably, one adult male (ID057) showed a higher prevalence of Helicobacter, reaching up to 20% relative abundance in the cloaca, represented by a single ASV assigned as Helicobacter sp. and not found in any other sample. This suggests a potential overgrowth or infection despite the relatively good clinical status of the individual observed at the time. Therefore, like other reptiles, loggerhead sea turtles could serve as a valuable source of previously undiscovered cold-adapted Campylobacter and Helicobacter species, as well as mucosal Arcobacter species.

Oral bacterial communities harbour distinct taxa and could reflect recent diet

In oral samples, the genera Truepera and Halioglobus showed differential abundance. Trueperaceae, a relatively new family, includes one isolate Truepera radiovictrix, known for radiation resistance and thriving in extreme conditions like other Deinococci members [39]. Truepera was also a dominant part of the oral microbiota in splendid japalure lizards [40], but its role and functions in reptile oral microbiomes remain unclear. Tolerance to extreme environments and ability to use diverse carbon sources may allow Truepera spp. to outcompete other taxa on reptilian oral mucosa. The genus Halioglobus is typically found in seawater, marine sediments, and has been associated with dinoflagellate blooms and starving green-lipped mussels [41, 42]. We have previously reported Halioglobus in oral samples of loggerhead sea turtles [24]. Halioglobus bacteria likely derive from loggerhead prey such as mussels or oysters rather than being an intrinsic property of sea turtle mucosal surfaces, given the limited information on host-associated Halioglobus. Similarly, the high relative abundance of the genus Exiguobacterium in oral and cloacal samples of one juvenile loggerhead (ID069) sampled upon admission may indicate recent feeding, as some Exiguobacterium species are commonly found on shrimp or algae [43–45].

Carapace microbiota is distinct, yet can harbour taxa specific to oral or cloacal microbiota

The carapace and skin of loggerhead sea turtles harbour diverse microbial biofilms, rich in prokaryotes, microeukaryotes, and macroeukaryotes like barnacles and algae [3, 27]. These surface microbial communities vary with the turtle's geographical location and the sampled anatomical site [27, 28]. They can be considered as microbial reservoirs and “diversity hotspots” in otherwise scarce environments [46, 47]. In our study, we found expectedly distinct microbial communities in the carapace, oral mucosa, and cloaca, although many microbial taxa were present across all body sites, including potential zoonotic pathogens primarily from the cloaca. Oral and cloacal samples shared several differentially abundant microbial taxa (Tenacibaculum, Cardiobacteraceae, Campylobacter), suggesting co-inhabitation of the gastrointestinal tract. The carapace microbiota differed significantly from tank water, with indication of transfer of certain taxa from carapaces to tank water or vice versa. Taxa like Alteromonadaceae and Colwelliaceae (Thalassotalea), differentially abundant on carapaces in this study, were prominent in enclosure tank water in prior studies that did not examine carapace bacterial communities [23], possibly originating from captive turtles' carapaces. The implications of the interplay between microbes from different body sites on loggerhead sea turtles in their natural habitats and during captivity are not yet clear. However, it is essential to consider surface microbiota and potential opportunistic pathogens, especially when rehabilitating severely injured and possibly immunocompromised individuals.

Fungal communities of loggerhead cloaca are highly heterogeneous and diverse

Fungal communities in loggerhead sea turtles' cloaca and tank water exhibit high variability among individuals, with greater diversity observed in tank water. In this study, we couldn't attribute differences in the composition and structure of cloacal mycobiota to the turtles' age, sex, or hospitalization status, possibly due to the limited sample size per sampling site and condition assessed. Conversely, captive juvenile green turtles displayed more consistent mycobiota richness and diversity across various sampling sites and health conditions, unaffected by environmental fungi [30]. The taxonomic composition of cloacal and tank water fungal phyla in our study aligns with previously reported marine fungi groups found in green sea turtle faeces [30], as well as marine algicolous fungi, sediments, and sponges [48]. Due to the lower resolution of the ITS2 gene marker and many unassigned fungal ASVs beyond the family level, this study offers just a general overview and serves as a foundation for further exploration of specific groups of interest.

We sporadically detected pathogenic fungal genera [17], with nine pathogenic genera found in both cloaca and tank water. Among these, Fusarium species (family Nectriaceae) are recognized sea turtle pathogens linked to reduced hatchling success [49]. Our study identified Fusarium oxysporum (a known pathogen), Fusarium neocosmosporellium, and Fusarium waltergamsii, not previously reported as sea turtle pathogens but related to known pathogenic species in the Fusarium solani species complex [17, 50]. Fusarium ASVs were relatively less abundant, except in one sample where they reached 8%. However, some ASVs assigned to the Nectriaceae genus may belong to Fusarium species as ITS region can be insufficient in detecting Fusarium species [51], potentially affecting our reported abundance of pathogenic fungi in cloacal and tank water samples.

The origin of these fungi—whether they are metabolically active and intrinsic to the sampled turtle population or introduced from the environment as spores or through food—is unclear. The impact of these fungi on the host, the interaction with the turtle's immune system and physiology, and their potential host association remain unknown. Many fungal ASVs detected in this study belong to primarily terrestrial taxa, suggesting possible terrestrial sources during rehabilitation or a lack of marine fungal sequences in taxonomy databases. Additionally, numerous reads were assigned only as "Fungi," indicating either undiscovered fungal taxa, lack of representative sequences or possible turtle host origin.

Despite being underexplored, marine fungi, especially those associated with hosts, offer new avenues for biotechnological exploration. Fungi adapted to marine vertebrate hosts may prove valuable in expanding our toolbox for treating increasingly drug-resistant fungal pathogens in human health and as biocontrol agents in agriculture [52, 53].

Low biomass of samples is a potential limitation to interpreting results

Microbial composition results should be interpreted with caution due to low biomass collected with swabs and fewer fungal cells compared to bacterial cells in vertebrate guts [54]. Negative control (sterile water) sequenced with bacterial primers showed a small number of reads that were also detected in low-read samples, indicating potential contamination during DNA extraction and sequencing specifically for samples with low biomass (e.g., kitome) [55, 56]. Surprisingly, the negative control sequenced with fungal primers had significantly more reads than the bacterial negative control, containing numerous ASVs also found in cloacal and tank water samples. Although the community structures of cloacal and tank water samples differed from the negative control, indicating genuine fungal communities rather than random contaminants, the high number of reads from seemingly low-biomass samples or the "empty" negative control raises concerns about our sample handling and sequencing approach. To our knowledge, no studies addressed fungal contamination in DNA extraction kits as they do for prokaryotes, making it challenging to pinpoint the exact source of this fungal DNA. For future studies we suggest including additional negative control samples at various sampling and processing stages when investigating as of yet undescribed gut-associated fungal communities.

Considerations and focusing future efforts in microbiome management

Sea turtles face growing threats from climate change, habitat loss, and human activities. In addition to traditional conservation efforts such as rescue and expanding protected areas, studying their microbiomes can enhance our understanding of sea turtle biology and support conservation. As global shifts in microbiomes gain attention, investigating endangered species' microbiomes is crucial [57]. Sea turtles, including loggerheads, can act as carriers for zoonotic diseases, raising concerns as their behaviour shifts and interactions with humans increase. Handlers of loggerhead sea turtles must take precautions to prevent exposure to opportunistic pathogens sea turtles may carry. Moreover, as microbial sequencing becomes more accessible, establishing the baseline state of microbial health of turtles is essential. This information will aid in developing probiotics, prebiotics, or environmental/food enrichments to potentially mitigate pollutant ingestion (e.g., plastics) and promote sea turtle recovery and healthy development.

Loggerhead sea turtle-associated microbial communities are crucial for understanding the biology of these endangered reptiles and supporting conservation. In this study, we examined bacterial and fungal communities in juvenile, subadult, and adult loggerhead sea turtles. Our findings revealed distinct microbial communities in the carapace, cloaca, and oral mucosa, with characteristic taxa for each site and shared taxa between cloacal and oral samples (e.g., Tenacibaculum, Moraxellaceae, Cardiobacteriaceae, and Campylobacter). Cloacal bacterial communities exhibited decreasing diversity and changing composition with turtle age, likely due to shifts in diet as juveniles develop stronger bite force and diving capabilities. Microbial exchange with the environment appears to occur, particularly from turtles to tank water, especially from the carapace. Fungal communities in cloaca and tank water displayed high heterogeneity across individuals, with no age or clinical patterns, possibly due to limited samples and low fungal biomass in cloacal samples. Loggerhead sea turtles host complex microbial communities, including potential bacterial and fungal pathogens, which pose risks to handlers or general public as changing turtle behaviour leads to increased human-turtle interactions. Despite growing research on loggerhead microbiomes, defining a healthy microbiome beyond bacteria remains challenging. Future research should prioritize establishing a description of healthy loggerhead microbiomes at different developmental stages (from eggs and hatchlings to juveniles and adults) to enhance conservation practices and explore potential probiotics and prebiotics for addressing their current and future needs.

Loggerhead sea turtle sampling

Loggerhead sea turtles were sampled from 2019 to 2021 in spring, summer, and autumn season (Table S1). The turtles were found strained, injured, or were unintentionally caught by fishermen in the Adriatic Sea area. Subsequently, the turtles were transported to local veterinary clinics or rescue centres, namely the Sea Turtle Clinic (STC) of the Department of Veterinary Medicine of University of Bari “Aldo Moro” in Italy and the Marine Turtle Rescue Center Aquarium Pula in Croatia (Fig. 6a) [24]. Upon arrival, all turtles underwent examination for injuries and relevant information on their clinical status was collected. Most turtles were found floating or stranded, exhibiting visible injuries caused by boat propellers or entanglement in fishing nets. Some turtles were in good clinical condition but unable to dive (Table S1, Table S2). Sampling was conducted by trained personnel in accordance with the 1975 Declaration of Helsinki, as revised in 2013 and the applicable national laws immediately upon the turtles’ arrival to the rehabilitation centres or during/after rehabilitation, and prior to release (Table 1). Rehabilitation or hospitalisation time period were categorized in three groups: “early” (0–2 days), “mid” (3–10 days), and “late” (> 10 days) (Table S1). The turtles were kept in individual enclosures with artificial saltwater at STC or recirculating and filtered local seawater in Aquarium Pula [24]. A total of 18 loggerhead sea turtles and 8 respective enclosures were included in the sampling, with three turtles sampled twice during their rehabilitation period: ID057 during the early stage, ID093 during both early and late stages, and ID096 in the late stage (Table 1, Table S1). Turtle age range categories were established based on size measurements (curved carapace length) grouped according to [33] as follows: juveniles ≤ 59.9 cm, subadults 60-69.9 cm, adults ≥ 70 cm. Sex determination was based on observable physical characteristics when possible (Fig. 6b). Out of the turtles sampled, ten turtles were classified as juveniles, four as subadult (including two females), and four as adults (one female and three males) (Table 1, Fig. 6).

Juvenile turtles had a median weight of 8 kg [interquartile range (IQR) = 9.6] and a median CCL of 40.4 cm (IQR = 16.2). Subadult turtles had a median weight of 31.6 kg (IQR = 5.3) and median CCL of 66.1 cm (IQR = 3.3), adult turtles had a median weight of 50.8 kg (IQR = 5.3) and median CCL of 75.5 cm (IQR = 4.9) (Table 1).

The cloaca and oral cavity were sampled in triplicates with sterile synthetic swabs (Aptaca Nuova) as described in Filek et al. [24]. Each swab tip was stored individually in 2 ml sterile centrifuge tubes (Eppendorf) in 1 ml of 96% EtOH. When available (during or after rehabilitation, Table S1), 400 ml of tank water was collected in sterile containers and kept cool until further processing in the laboratory. Once in the lab, 250 ml of the tank water was vacuum filtered using a 45 mm in diameter, 0.2 micrometre pore size sterile Whatmann polycarbonate membrane filters (Sigma-Aldrich). Each filter was then folded and stored in 2 ml tubes in 1 ml of 96% EtOH. All samples were stored at -20°C until DNA extraction. Additionally, corresponding epizoic carapace samples were obtained for most of the turtles (Table 1). Briefly, the entire carapace was randomly and gently brushed using a sterile toothbrush, and the collected material was resuspended in 96% EtOH for preservation until further processing [27, 58].

Table 1

Information about loggerhead sea turtles and their corresponding endozoic (16S rRNA gene and fungal ITS2 region) and epizoic (16S rRNA gene) samples obtained from Kanjer et al. (V4 – [27], V34 – this study). The turtles were retrieved at various locations in the Adriatic Sea and were admitted to Marine Turtle Rescue Center Aquarium Pula (Croatia) unless indicated otherwise at the “Turtle name” column in super script (^T Tyrrhenian Sea; ^STC The Sea Turtle Clinic at University of Bari, Italy). The abbreviations are as follows: Clo – cloaca, Orl – oral cavity, TW – tank water, Car – carapace, NA – not available, ND – not determined.
Turtle ID	Turtle name	Sampling event	Endozoic 16S/ITS2 sample presence (+) or absence (-) ^*			Epizoic 16S sample ID (region)	Sex	Age range
Turtle ID	Turtle name	Sampling event	Clo	Orl	TW	Car	Sex	Age range
ID010	Merry Fisher	0084	+/+	-/-	+/+	TB139 (V4)	Female	Subadult
ID047	Žal	0064	+/+	+/-	+/+	TB115 (V4)	ND	Juvenile
ID056	Samba	0073	+/+	+/-	+/+	TB117 (V4)	Female	Adult
ID057	Angelo^{T, STC}	0074	+/+	+/-	-/-	NA	Male	Adult
ID057	Angelo^{T, STC}	0092	+/+	+/-	-/-	TB119 (V4)	Male	Adult
ID068	Kanooh^STC	0087	+/+	+/-	-/-	TB145 (V4)	ND	Juvenile
ID069	Kanfus^STC	0088	+/+	+/-	-/-	NA	ND	Juvenile
ID070	Futon^STC	0089	+/+	+/-	-/-	TB149 (V4)	ND	Juvenile
ID071	Cosmyn^STC	0090	+/+	+/-	-/-	TB151 (V4)	ND	Subadult
ID073	Marvin	0093	+/+	+/-	+/+	TB155 (V4)	ND	Juvenile
ID074	Ryan	0094	+/+	+/-	-/-	TB157 (V4)	ND	Juvenile
ID093	Ella Ravka	0113	+/+	+/-	-/-	TB159 (V34)	ND	Subadult
ID093	Ella Ravka	0119	+/+	+/-	+/+	NA	ND	Subadult
ID096	Maro	0117	+/+	+/-	+/+	TB167 (V34)	Female	Subadult
ID096	Maro	0118	+/+	+/-	+/+	NA	Female	Subadult
ID097	Freewings	0120	+/+	+/-	+/+	NA	ND	Juvenile
ID098	Maksimus	0123	+/+	+/-	+/+	TB175 (V34)^**	ND	Juvenile
ID117	Karlo Albano	0141	+/+	+/-	-/-	TB215 (V34)	Male	Adult
ID118	Oliver Raul	0142	+/+	+/-	-/-	TB217 (V34)	ND	Juvenile
ID119	Martin	0143	+/+	+/-	-/-	TB219 (V34)	ND	Juvenile
ID122	Luka Amadeo	0146	+/-	+/-	-/-	NA	Male	Adult
^* Each endozoic sequencing sample ID has a 16S or ITS prefix, sampling event number, and suffix corresponding to sampling site: C - cloaca, O - oral cavity, W - tank water; for example, sample ID ITS0084C represents cloacal sample of fungal ITS2 sequences for sampling event 0084.
^** Carapace sample was taken at a different sampling event (a month prior to endozoic sampling).

DNA extraction and sequencing

Total DNA from swabs and filters was extracted using the DNeasy PowerSoil kit (Qiagen). Nuclease-free water (W4502 Sigma-Aldrich) was used as the negative control for the DNA extraction step and was processed using the DNA extraction kit in parallel to cloacal, oral and tank water samples. The swabs were transferred to PowerBead tubes with added two 3 mm sterile glass beads. The tank water samples’ filters were first cut in half and then added to the tubes as described above. The remaining ethanol in the sample tubes was centrifuged at the highest speed for 5 minutes and the resulting pellet was then transferred to the respective PowerBead tube. The samples were processed following the manufacturer’s instructions with several modifications: (1) the samples were incubated in C1 solution at 65°C for one hour, (2) instead of beadbeating PowerBead Tubes were vortexed horizontally for 10 min at maximum speed, and (3) all downstream incubation times at 2–8°C were increased to 15 min. The DNA was quantified by BioSpec-nano (Shimadzu).

The methods used for epizoic carapace samples that were sequenced for V4 region (ID010 – ID074) are described in Kanjer et al. [27]. The scrapings of epizoic carapace biofilms (ID093 – ID022) were firstly pelleted by centrifugation in conical 15 ml tubes at 1400 rfc for 10 minutes after which the ethanol supernatant was removed and 250 mg of material was used for the extraction. DNA was extracted via DNeasy PowerLyzer PowerSoil extraction kit (Qiagen) following the manufacturer’s instructions and modified as follows: 1) after addition of C1 solution the samples were incubated at 70°C for 10 minutes; 2) bead beating was performed at 30 Hz for 1 minute in TissueLyzer Retsch Qiagen; 3) 50 ul of C6 solution was used for DNA elution and incubated for 5 minutes RT prior to centrifugation. The quality and quantity of extracted DNA was evaluated by BioSpec-nano (Shimadzu). The extracted DNA from both endozoic and epizoic samples was stored at -20°C and sent for Illumina MiSeq v3 300x2 bp paired-end sequencing (Microsynth, Switzerland). Equal aliquots of extracted DNA reaching a minimum concentration of 5 ng/µL were sent as a template for 16S rRNA gene V34 region and ITS2 region of fungal nuclear ribosomal gene sequencing for bacteria and fungi, respectively. Primers used for sequencing the V34 region of 16S rRNA gene were 341F and 805R [59], and primers for fungal ITS2 region of the nuclear ribosomal gene were ITS3 and ITS4 [60].

Bioinformatic and statistical analyses

The obtained endozoic and tank water sequences had non-biological sequences trimmed by the sequencing facility and checked for quality with FastQC [61]. The corresponding carapace samples’ sequences first published by Kanjer et al. [27] were obtained from the European Nucleotide Archive (ENA; see section below). The sequencing data were imported and analysed in Qiime 2 [62]. For ITS2 region the Qiime 2 version 2021.8 was used, and for 16S rRNA gene sequences it was version 2023.2, with the same plugins used for initial processing of the data albeit in versions corresponding to version used. Each sequencing dataset was processed individually due to intrinsic properties of distinct sequencing events and gene markers used. The code for analyses and intermediate files pertaining to each sequencing event are available at GitHub and ZENODO (see section below).

Statistical analyses were performed within Qiime 2 environment and with R. The obtained sequences were demultiplexed by q2-demux and denoised by q2-dada2, producing amplicon sequence variants (ASVs, 100% operating taxonomic units). Sequences were then aligned by MAFFT [63] and FastTree2 in q2-phylogeny to construct a phylogenetic tree [64]. Taxonomy was assigned to ASVs through q2-feature-classifier [65, 66] classify-sklearn naive Bayes taxonomy classifier. For 16S rRNA gene sequences we used SILVA v. 138 [67] and for ITS2 reads we used UNITE v. 8.3 dynamic classifier [68]. Reads identified as chloroplasts and mitochondria were removed from the 16S rRNA gene sequencing data before further processing and statistical analyses. Negative controls and samples below 10 000 reads were excluded from all further statistical analyses. Additionally, to compare endozoic and epizoic 16S rRNA gene sequences, the sequences resulting from individual sequencing events were first trimmed to V4 region by cutadapt v.2023.2.0 using primers 515F and 806R [69, 70] followed by denoising. After denoising, the V4-trimmed epizoic, endozoic and tank water sequences were merged and analysed together.

Alpha diversity indices, including Shannon’s entropy, Pielou’s evenness, Faith’s phylogenetic diversity, and observed ASVs, were calculated via q2-diversity plugin. The Kruskal-Wallis rank sum test was employed to determine differences between selected groups, followed by post-hoc pairwise comparisons using Wilcoxon rank sum exact test. For beta diversity analyses, rarefied data (sequencing depth determined by alpha rarefaction curves) were explored using the q2-diversity plugin with Bray-Curtis, Jaccard, unweighted UniFrac, weighted UniFrac distances [71]. Compositional data analysis on non-rarefied datasets was performed by calculating robust Aitchison distances using the q2-deicode plugin or the R package vegan v.2.6-4 [72–74]. Principal coordinates analysis (PCoA) was conducted on Bray-Curtis, Jaccard, and all UniFrac distances, while principal components analysis was performed for robust Aitchison (rPCA) using q2-diversity and q2-deicode, respectively. To assess the relative impact of factors (age range, sex, and duration of rehabilitation) on microbial communities, a multi-way permutational multivariate analysis of variance (Adonis2 PERMANOVA) with 9999 permutations was employed (Anderson, 2001) in R by using vegan v.2.6-4 and pairwiseAdonis v.0.4.1 packages [75]. Resulting p-values from all pairwise tests were adjusted using the Benjamini-Hochberg false-discovery rate (FDR) correction for multiple comparisons (reported as q-values). Differential abundance analysis was used to identify differentially abundant (DA) features in sample site pairs by using the ANCOMBC package “ancom-bc2” function in R [76, 77]. Default parameters were used and pairwise testing was enabled (with Holm’s method for adjusting p-values, reported as q-values), except for DA testing in V4-trimmed sequences where the “struc_zero” was set to “TRUE” to exclude structural zeros. Log fold change (LFC) indicates the scale of differential abundance between differentially abundant features. Features with p-value ≤ 0.05 were reported as differentially abundant, and q-value ≤ 0.05 as significantly differentially abundant.

Data exploration and visualizations were conducted by using R v.4.3.0 in RStudio (R Core Team 2023) with packages listed above and qiime2R v.0.99.6 [78], tidyverse v2.0.0 [79], ggplot2 v3.4.2 [80], Microsoft Excel, and Adobe Illustrator.

Data and code availability

Sequencing data with non-biological sequences removed is available at EMBL ENA at accessions PRJEB62752, and PRJEB68216 for 16S rRNA gene and PRJEB62762 for ITS2 region sequences. Complete code and instructions for processing of the sequencing data and subsequent statistical analyses, results, and data visualizations are available at Github (https://github.com/kl-fil/2023-Filek_et_al._TBIOME_project) and ZENODO data depository (doi:10.5281/zenodo.8054926). Epizoic samples’ sequencing data obtained from Kanjer et al. [27] and used in this study can be found at ENA under accession PRJEB51458.

ASV

amplicon sequence variants

CCL

curved carapace length

DNA

deoxyribonucleic acid

IUCN

International Union for Conservation of Nature

PCoA

Principal coordinate analysis

rPCA

robust Aitchison Principal component analysis

rRNA

ribosomal ribonucleic acid

STC

Sea Turtle Clinic

Ethical approval and Consent to participate

Sampling was performed in accordance with the 1975 Declaration of Helsinki, as revised in 2013 and the applicable national laws. The sampling at the Sea Turtle Clinic (Bari, Italy) was conducted with the permission of the Department of Veterinary Medicine Animal Ethic Committee (Authorization # 4/19), while sampling in Croatia was done in accordance with the authorization of the Marine Turtle Rescue Center by the Ministry of Environment and Energy of the Republic of Croatia.

Consent for publication

Not applicable.

Data availability

The datasets supporting the conclusions of this article are included within the article and …

Competing interests

The authors declare that there are no competing interests.

Funding

This work has been fully supported by the Croatian Science Foundation under the project number UIP-2017-05-5635. The work of doctoral student K. Filek has been fully supported by the “Young Researchers’ Career Development Project – Training of Doctoral Students” of the Croatian Science Foundation funded by the European Union from the European Social Fund.

Contributions

KF and SB concepted and designed the study; AT, MC and SB collected the samples; MŽ, LK and KF carried out the laboratory work; KF and BV conducted the bioinformatics, statistical analyses, and data visualization and interpretation; KF wrote the first draft of the manuscript; all authors revised the paper and approved the final version of the manuscript.

Acknowledgments

For the Croatian sample collection, we are thankful to Milena Mičić, Karin Gobić Medica and the rest of the staff from the Marine Turtle Rescue Center (Aquarium Pula).

McFall-Ngai M, Hadfield MG, Bosch TCG, Carey HV, Domazet-Lošo T, Douglas AE, et al. Animals in a bacterial world, a new imperative for the life sciences. Proc Natl Acad Sci. 2013;110:3229.
Henry LP, Bruijning M, Forsberg SKG, Ayroles JF. The microbiome extends host evolutionary potential. Nat Commun. 2021;12:5141.
Robinson NJ, Pfaller JB. Sea Turtle Epibiosis: Global Patterns and Knowledge Gaps. Front Ecol Evol. 2022;10:844021.
Robinson NJ, Majewska R, Lazo-Wasem EA, Nel R, Paladino FV, Rojas L, et al. Epibiotic diatoms are universally present on all sea turtle species. PLoS ONE. 2016;11:e0157011–1.
Abdelrhman KFA, Bacci G, Mancusi C, Mengoni A, Serena F, Ugolini A. A First Insight into the Gut Microbiota of the Sea Turtle Caretta caretta. Front Microbiol. 2016;7:1–5.
Kuschke SG. What lives on and in the sea turtle? A literature review of sea turtle bacterial microbiota. Anim Microbiome. 2022;4:52.
Stanford CB, Iverson JB, Rhodin AGJ, van Paul P, Mittermeier RA, Kuchling G, et al. Turtles and Tortoises Are in Trouble. Curr Biol. 2020;30:R721–35.
Mazaris AD, Schofield G, Gkazinou C, Almpanidou V, Hays GC. Global sea turtle conservation successes. Sci Adv. 2017;3:e1600730–0.
Mazaris AD, Dimitriadis C, Papazekou M, Schofield G, Doxa A, Chatzimentor A, et al. Priorities for Mediterranean marine turtle conservation and management in the face of climate change. J Environ Manage. 2023;339:117805.
Hird SM. Evolutionary Biology Needs Wild Microbiomes. Front Microbiol. 2017;8:725.
Ebani VV. Bacterial Infections in Sea Turtles. Vet Sci. 2023;10:333.
IUCN. Caretta caretta (Mediterranean subpopulation): Casale, P.: The IUCN Red List of Threatened Species 2015: e.T83644804A83646294. 2015.
Blasi MF, Migliore L, Mattei D, Rotini A, Thaller MC, Alduina R. Antibiotic Resistance of Gram-Negative Bacteria from Wild Captured Loggerhead Sea Turtles. Antibiotics. 2020;9:162–2.
Pace A, Dipineto L, Fioretti A, Hochscheid S. Loggerhead sea turtles as sentinels in the western Mediterranean: antibiotic resistance and environment-related modifications of Gram-negative bacteria. Mar Pollut Bull. 2019;149:110575–5.
Pace A, Rinaldi L, Ianniello D, Borrelli L, Cringoli G, Fioretti A, et al. Gastrointestinal investigation of parasites and Enterobacteriaceae in loggerhead sea turtles from Italian coasts. BMC Vet Res. 2019;15:1–9.
Capri FC, Prazzi E, Casamento G, Gambino D, Cassata G, Alduina R. Correlation Between Microbial Community and Hatching Failure in Loggerhead Sea Turtle Caretta caretta. Microb Ecol. 2023. https://doi.org/10.1007/s00248-023-02197-8.
Gleason FH, Allerstorfer M, Lilje O. Newly emerging diseases of marine turtles, especially sea turtle egg fusariosis (SEFT), caused by species in the Fusarium solani complex (FSSC). Mycology. 2020;11:184–94.
Trotta A, Cirilli M, Marinaro M, Bosak S, Diakoudi G, Ciccarelli S, et al. Detection of multi-drug resistance and AmpC β-lactamase/extended-spectrum β-lactamase genes in bacterial isolates of loggerhead sea turtles (Caretta caretta) from the Mediterranean Sea. Mar Pollut Bull. 2021;164:112015.
Trotta A, Marinaro M, Sposato A, Galgano M, Ciccarelli S, Paci S, et al. Antimicrobial Resistance in Loggerhead Sea Turtles (Caretta caretta): A Comparison between Clinical and Commensal Bacterial Isolates. Animals. 2021;11:2435.
Alduina R, Gambino D, Presentato A, Gentile A, Sucato A, Savoca D, et al. Is Caretta Caretta a Carrier of Antibiotic Resistance in the Mediterranean Sea? Antibiotics. 2020;9:116–6.
Gambino D, Persichetti MF, Gentile A, Arculeo M, Visconti G, Currò V, et al. First data on microflora of loggerhead sea turtle (Caretta caretta) nests from the coastlines of Sicily. Biol Open. 2020;9:bio045252–2.
Arizza V, Vecchioni L, Caracappa S, Sciurba G, Berlinghieri F, Gentile A, et al. New insights into the gut microbiome in loggerhead sea turtles Caretta caretta stranded on the Mediterranean coast. PLoS ONE. 2019;14:e0220329–9.
Biagi E, D’Amico F, Soverini M, Angelini V, Barone M, Turroni S, et al. Faecal bacterial communities from Mediterranean loggerhead sea turtles (Caretta caretta). Environ Microbiol Rep. 2019;11:361–71.
Filek K, Trotta A, Gračan R, Di Bello A, Corrente M, Bosak S. Characterization of oral and cloacal microbial communities of wild and rehabilitated loggerhead sea turtles (Caretta caretta). Anim Microbiome. 2021;3:59.
Biagi E, Musella M, Palladino G, Angelini V, Pari S, Roncari C, et al. Impact of Plastic Debris on the Gut Microbiota of Caretta caretta From Northwestern Adriatic Sea. Front Mar Sci. 2021;8:637030.
Scheelings TF, Moore RJ, Van TTH, Klaassen M, Reina RD. The gut bacterial microbiota of sea turtles differs between geographically distinct populations. Endanger Species Res. 2020;42:95–108.
Kanjer L, Filek K, Mucko M, Majewska R, Gračan R, Trotta A, et al. Surface microbiota of Mediterranean loggerhead sea turtles unraveled by 16S and 18S amplicon sequencing. Front Ecol Evol. 2022;10:907368.
Blasi MF, Rotini A, Bacci T, Targusi M, Ferraro GB, Vecchioni L, et al. On Caretta caretta ’s shell: first spatial analysis of micro- and macro-epibionts on the Mediterranean loggerhead sea turtle carapace. Mar Biol Res. 2021;17:762–74.
Scheelings TF, Moore RJ, Van TTH, Klaassen M, Reina RD. Microbial symbiosis and coevolution of an entire clade of ancient vertebrates: the gut microbiota of sea turtles and its relationship to their phylogenetic history. Anim Microbiome. 2020;2:17–7.
Guo Y, Chen H, Liu P, Wang F, Li L, Ye M et al. Microbial composition of carapace, feces, and water column in captive juvenile green sea turtles with carapacial ulcers. Front Vet Sci. 2022;9.
Trevelline BK, Fontaine SS, Hartup BK, Kohl KD. Conservation biology needs a microbial renaissance: a call for the consideration of host-associated microbiota in wildlife management practices. Proc R Soc B Biol Sci. 2019;286:20182448.
Reese AT, Dunn RR. Drivers of Microbiome Biodiversity: A Review of General Rules, Feces, and Ignorance. mBio. 2018;9. 10.1128/mbio.01294-18.
Mariani G, Bellucci F, Cocumelli C, Raso C, Hochscheid S, Roncari C, et al. Dietary Preferences of Loggerhead Sea Turtles (Caretta caretta) in Two Mediterranean Feeding Grounds: Does Prey Selection Change with Habitat Use throughout Their Life Cycle? Anim Open Access J MDPI. 2023;13:654.
Casale P, Abbate G, Freggi D, Conte N, Oliverio M, Argano R. Foraging ecology of loggerhead sea turtles Caretta caretta in the central Mediterranean Sea: evidence for a relaxed life history model. Mar Ecol Prog Ser. 2008;372:265–76.
Hao W, Gerdts G, Peplies J, Wichels A. Bacterial communities associated with four ctenophore genera from the German Bight (North Sea). FEMS Microbiol Ecol. 2015;91:1–11.
Mabrok M, Algammal AM, Sivaramasamy E, Hetta HF, Atwah B, Alghamdi S, et al. Tenacibaculosis caused by Tenacibaculum maritimum: Updated knowledge of this marine bacterial fish pathogen. Front Cell Infect Microbiol. 2023;12:1068000.
Luschi P, Casale P. Movement patterns of marine turtles in the Mediterranean Sea: a review. Ital J Zool. 2014;81:478–95.
Gilbert MJ, Duim B, Zomer AL, Wagenaar JA. Living in Cold Blood: Arcobacter, Campylobacter, and Helicobacter in Reptiles. Front Microbiol. 2019;10.
Albuquerque L, Simões C, Nobre MF, Pino NM, Battista JR, Silva MT, et al. Truepera radiovictrix gen. nov., sp. nov., a new radiation resistant species and the proposal of Trueperaceae fam. nov. FEMS Microbiol Lett. 2005;247:161–9.
Tian Z, Pu H, Cai D, Luo G, Zhao L, Li K, et al. Characterization of the bacterial microbiota in different gut and oral compartments of splendid japalure (Japalura sensu lato). BMC Vet Res. 2022;18:205.
Hattenrath-Lehmann TK, Jankowiak J, Koch F, Gobler CJ. Prokaryotic and eukaryotic microbiomes associated with blooms of the ichthyotoxic dinoflagellate Cochlodinium (Margalefidinium) polykrikoides in New York, USA, estuaries. PLoS ONE. 2019;14:e0223067.
Li S, Young T, Archer S, Lee K, Alfaro AC. Gut microbiome resilience of green-lipped mussels, Perna canaliculus, to starvation. Int Microbiol. 2023. https://doi.org/10.1007/s10123-023-00397-3.
Kasana RC, Pandey CB. Exiguobacterium: an overview of a versatile genus with potential in industry and agriculture. Crit Rev Biotechnol. 2018;38:141–56.
Liu F, Li Y, He W, Wang W, Zheng J, Zhang D. Exiguobacterium algae sp. nov. and Exiguobacterium qingdaonense sp. nov., two novel moderately halotolerant bacteria isolated from the coastal algae. Antonie Van Leeuwenhoek. 2021;114:1399–406.
Cong M, Jiang Q, Xu X, Huang L, Su Y, Yan Q. The complete genome sequence of Exiguobacterium arabatum W-01 reveals potential probiotic functions. MicrobiologyOpen. 2017;6:e00496.
Keller AG, Apprill A, Lebaron P, Robbins J, Romano TA, Overton E et al. Characterizing the culturable surface microbiomes of diverse marine animals. FEMS Microbiol Ecol. 2021;97.
Filek K, Lebbe L, Willems A, Chaerle P, Vyverman W, Žižek M, et al. More than just hitchhikers: a survey of bacterial communities associated with diatoms originating from sea turtles. FEMS Microbiol Ecol. 2022;98:fiac104.
Jones EBG, Pang K-L, Abdel-Wahab MA, Scholz B, Hyde KD, Boekhout T, et al. An online resource for marine fungi. Fungal Divers. 2019;96:347–433.
Cafarchia C, Paradies R, Figueredo LA, Iatta R, Desantis S, Di Bello AVF, et al. Fusarium spp. in Loggerhead Sea Turtles (Caretta caretta): From Colonization to Infection. Vet Pathol. 2020;57:139–46.
Geiser DM, Al-Hatmi AMS, Aoki T, Arie T, Balmas V, Barnes I, et al. Phylogenomic Analysis of a 55.1-kb 19-Gene Dataset Resolves a Monophyletic Fusarium that Includes the Fusarium solani Species Complex. Phytopathology®. 2021;111:1064–79.
Hoh DZ, Lin Y-F, Liu W-A, Sidique SNM, Tsai IJ. Nest microbiota and pathogen abundance in sea turtle hatcheries. Fungal Ecol. 2020;47:100964.
Vargas-Gastélum L, Riquelme M. The Mycobiota of the Deep Sea: What Omics Can Offer. Life. 2020;10:292.
Debbab A, Aly AH, Lin WH, Proksch P. Bioactive Compounds from Marine Bacteria and Fungi. Microb Biotechnol. 2010;3:544–63.
Lavrinienko A, Scholier T, Bates ST, Miller AN, Watts PC. Defining gut mycobiota for wild animals: a need for caution in assigning authentic resident fungal taxa. Anim Microbiome. 2021;3:75.
Salter SJ, Cox MJ, Turek EM, Calus ST, Cookson WO, Moffatt MF, et al. Reagent and laboratory contamination can critically impact sequence-based microbiome analyses. BMC Biol. 2014;12:87.
Grahn N, Olofsson M, Ellnebo-Svedlund K, Monstein H-J, Jonasson J. Identification of mixed bacterial DNA contamination in broad-range PCR amplification of 16S rDNA V1 and V3 variable regions by pyrosequencing of cloned amplicons. FEMS Microbiol Lett. 2003;219:87–91.
Peixoto RS, Harkins DM, Nelson KE. Advances in Microbiome Research for Animal Health. Annu Rev Anim Biosci. 2021;9:289–311.
Pinou T, Domenech F, Majewska R, Pfaller JB, Zardus JD, Robinson NJ. Standardizing Sea Turtle Epibiont Sampling: Outcomes of the Epibiont Workshop at the 37 th International Sea Turtle Symposium. Mar Turt Newsl., April. https://doi.org/10.13140/RG.2.2.31843.81440.
Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res. 2013;41:1–11.
White TJ, Bruns T, Lee S, Taylor J, AMPLIFICATION, AND DIRECT SEQUENCING OF FUNGAL RIBOSOMAL RNA GENES FOR PHYLOGENETICS. PCR Protocols. Elsevier; 1990. pp. 315–22.
Andrews S. FastQC: a quality control tool for high throughput sequencing data. 2010.
Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37:852–7.
Katoh K. MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Res. 2002;30:3059–66.
Price MN, Dehal PS, Arkin AP. FastTree 2 - Approximately maximum-likelihood trees for large alignments. PLoS ONE. 2010;5.
Bokulich NA, Kaehler BD, Rideout JR, Dillon M, Bolyen E, Knight R, et al. Optimizing taxonomic classification of marker-gene amplicon sequences with QIIME 2’s q2-feature-classifier plugin. Microbiome. 2018;6:90–0.
Robeson MS, O’Rourke DR, Kaehler BD, Ziemski M, Dillon MR, Foster JT et al. RESCRIPt: Reproducible sequence taxonomy reference database management for the masses. 2020;:2020.10.05.326504.
Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, Peplies J, et al. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 2007;35:7188–96.
Nilsson RH, Larsson K-H, Taylor AFS, Bengtsson-Palme J, Jeppesen TS, Schigel D, et al. The UNITE database for molecular identification of fungi: handling dark taxa and parallel taxonomic classifications. Nucleic Acids Res. 2019;47:D259–64.
Parada AE, Needham DM, Fuhrman JA. Every base matters: assessing small subunit rRNA primers for marine microbiomes with mock communities, time series and global field samples. Environ Microbiol. 2016;18:1403–14.
Apprill A, McNally S, Parsons R, Weber L. Minor revision to V4 region SSU rRNA 806R gene primer greatly increases detection of SAR11 bacterioplankton. Aquat Microb Ecol. 2015;75:129–37.
Lozupone C, Lladser ME, Knights D, Stombaugh J, Knight R, UniFrac. An effective distance metric for microbial community comparison. ISME J. 2011;5:169–72.
Gloor GB, Macklaim JM, Pawlowsky-Glahn V, Egozcue JJ. Microbiome datasets are compositional: And this is not optional. Front Microbiol. 2017;8:NOV.
Martino C, Morton JT, Marotz CA, Thompson LR, Tripathi A, Knight R, et al. A Novel Sparse Compositional Technique Reveals Microbial Perturbations. mSystems. 2019;4:1–13.
Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D et al. vegan: Community Ecology Package. 2020.
Arbizu PM. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis. 2017.
Peddada S, Lin H. Multi-group Analysis of Compositions of Microbiomes with Covariate Adjustments and Repeated Measures. preprint. In Review; 2023.
Lin H, Peddada SD. Analysis of compositions of microbiomes with bias correction. Nat Commun. 2020;11:3514.
Bisanz JE. qiime2R: Importing QIIME2 artifacts and associated data into R sessions. 2018.
Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, et al. Welcome to the tidyverse. J Open Source Softw. 2019;4:1686.
Wickham H. ggplot2: Elegant Graphics for Data Analysis. New York: Springer-Verlag; 2016.

No competing interests reported.

FigS1.png
Fig. S1 Fungal communities’ diversity . unweighted UniFrac PCoA plot and rAitchison PCA biplot
FigS2.png
Fig. S2 Relative abundance of differentially abundant taxa for epizoic, endozoic and tank water V4 reads
TableS1loggerheadinfoallsamples.xlsx
Table S1Extended metadata on all samples in this study
TableS2clinicalconsiderations.xlsx
Table S2 Clinical considerations for turtles in this study
TableS3v34rawcounts.xlsx
Table S3 Raw read counts for endozoic and tank water samples V34
TableS4ITS2rawcounts.xlsx
Table S4 Raw read counts for cloacal and tank water samples ITS2
TableS5v4rawcounts.xlsx
Table S5 Raw read counts for epizoic, endozoic, and tank water samples V4

Download PDF

Version 1

posted

You are reading this latest preprint version

Bacterial and fungal assemblages of the gut and carapace of loggerhead sea turtles

Status:

Version 1

Abstract

Background

Results

Conclusion

Figures

Background

Results

Endozoic and tank water bacterial communities

Endozoic and tank water fungal communities

Epizoic and endozoic bacterial communities

Discussion

Cloacal bacterial diversity and structure changes with the turtle age

Loggerhead cloacal communities are a promising source of novel Campilobacterota

Oral bacterial communities harbour distinct taxa and could reflect recent diet

Carapace microbiota is distinct, yet can harbour taxa specific to oral or cloacal microbiota

Fungal communities of loggerhead cloaca are highly heterogeneous and diverse

Low biomass of samples is a potential limitation to interpreting results

Considerations and focusing future efforts in microbiome management

Conclusions

Methods

Loggerhead sea turtle sampling

DNA extraction and sequencing

Bioinformatic and statistical analyses

Data and code availability

Abbreviations

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1