Fig 1.
Multivariate analysis of metagenomes microbial taxonomical and functional features.
A) Unsupervised hierarchical clustering of taxonomical (TAX) and functional (FUNC) features. Clades formed by gill samples and intestine plus digestive gland samples are highlighted respectively in purple and dark green. The pvclust AU (Approximately Unbiased) and BP (Bootstrap Probability) p-values are shown at clades nodes in red and light green respectively. B) Relative abundances of metagenome bacterial classes and functions respectively signed with RefSeq and COG (Level 2) databases. C) Multiple dimensional scale (MDS) plot of N. reynei metagenomics samples using distances and the fifteen most important bacteria classes (RefSeq database) and functions (COG database at Level 2) vectors as calculated from unsupervised random forest. D) Digestive gland, intestine and gill samples are shown as green, brown and purple dots respectively. Clusters of Orthologous Groups are classified into functional categories as follow: A—RNA processing and modification; E—Amino acid transport and metabolism; F—Nucleotide transport and metabolism; H—Coenzyme transport and metabolism; I—Lipid transport and metabolism; K–Transcription; L—Replication, recombination and repair; N—Cell motility; O—Post-translational modification, protein turnover, and chaperones; Q—Secondary metabolites biosynthesis, transport, and catabolism; S—Function Unknown; T—Signal transduction mechanisms; U—Intracellular trafficking, secretion, and vesicular transport; W—Extracellular structures; Z–Cytoskeleton.
Table 1.
N. reynei gill symbiotic genome bins metrics and features.
Fig 2.
Phylogeny of two high quality Teredinibacter bins based on a concatenated alignment of 43 proteins.
RAxML maximum likelihood tree based on the CAT model of rate heterogeneity and LG amino acid substitution matrix, with 100 rapid bootstraps. To improve readability, bootstraps for the Teredinibacter turnerae sub-clades I and II (with bootstrap values of 87 and 97 respectively) are omitted. The tree is rooted in between the Cellvibrionaceae branch and representatives of Halieaceae, Porticoccaceae, and Spongiibacteraceae. Color-coding of the families is similar to the one used by Spring and co-authors (2015). The scale bar represents 0.05 nucleotide substitutions per site.
Fig 3.
Gill genome bins and un-binned contigs for carbohydrate active enzymes (CAZymes).
A) Proportion of GH domains according to the substrate specificity at gills.bin1 and gills.bin 4 binned genomes and at other un-binned contigs from gills metagenome dataset. Substrate specificities as previously defined [9]: dark green = cellulose/xylan GH families (5, 6, 8, 9, 10, 11, 12, 44, 45, 51, 52, 62, and 74); dark blue = GH families with other or not-unique specificities (1, 2, 3, 13, 15, 23, 27, 30, 31, 35, 39, 43, 73, 77, 78, 79, 94, 95, 103, 108, 109, 115, 128, 130); light grey = laminarin GH families (16 and 81); yellow = other plant cell wall polysaccharides GH families (26, 53, and 67); light blue = chitin GH families (18, 19, and 20); light green = peptidoglycan GH families (28 and 105). B) Novel multicatalytic CAZymes configurations detected on Teredinibacter sp. gills.bin.4 genome bin (labeled in black), and on un-binned contigs from gill-symbiotic gammaproteobacterial community (labeled in red). Abbreviations key: glycoside hydrolases (GH), carbohydrate esterases (CE), carbohydrate binding modules (CBM). Catalytic domains and binding modules are color coded as black squares and gray circles.
Fig 4.
Gill genome bins contigs for Biosynthetic Gene Clusters (BGCs).
A) Secondary metabolomes. BGCs/contigs detected in Teredinibacter turnerae (T7901 and gills.bin.1), Teredinibacter sp.(gills.bin.4) and on contigs derived from N. reynei gill assembled dataset. Clusters are grouped, and color coded according to the compound class. B) Novel putative trans-AT PKS biosynthetic gene clusters detected on Teredinibacter sp. gills.bin.4 genome and the main predicted catalytic domains of their open reading frames (ORFs). Biosynthetic, transport-related, regulatory, other β-branching-related genes are color coded in dark-blue, light-blue, green, gray and orange respectively. antiSMASH gene clusters abbreviations (not cited on the main text body): PUFA—Polyunsaturated fatty acids; hrslactone–homoserine lactone. Domains keys: thioesterase (TE), aminotransferase (AMT), adenylation (A), ketosynthase (KS), ketoreductase (KR), dehydrogenase (DH), thiolation (T), enoyl-CoA dehydratases (ECH), 2-Nitropropane dioxygenase (2NPD), polyketide synthase cyclase (PKS CY), 4'-phosphopantetheinyl transferase (4PPT). Putative inter-proteins bimodules are present in yellow, PKS-encoded ECH catalytic domains are present in orange.
Fig 5.
A) resistome PCA using the fifteen most important genes based on mean decrease accuracy of Random Forest analysis. Teredinibacter forms a major group influenced by polyamine, isoniazid, and PK-NRP-derived antibiotics. B) Relative abundance of resistance gene markers, BGCs, and NRPS, PKS and hybrid BGCs. C) Host associated bacteria present a significantly larger metabolome and abundance of NRPS, PKS and their hybrid metabolic paths than free living bacteria. One asterisk represents Kruskal-Wallis test p-values <0.001 and two represent <0.0001 both adjusted with Bonferroni method.