Gill-associated bacteria are homogeneously selected in amphibious mangrove crabs to sustain host intertidal adaptation

Abstract

Background: The transition from water to air is a key event in the evolution of many marine organisms to access new food sources, escape water hypoxia, and exploit the higher and temperature-independent oxygen concentration of air. Despite the importance of microorganisms in host adaptation, their contribution to overcoming the challenges posed by the lifestyle changes from water to land is not well understood. To address this, we examined how microbial association with a key multifunctional organ, the gill, is involved in the intertidal adaptation of fiddler crabs, a dual-breathing organism.

Results: Electron microscopy revealed a rod-shaped bacterial layer tightly connected to the gill lamellae of the five crab species sampled across a latitudinal gradient from the central Red Sea to the southern Indian Ocean. The gill bacterial community diversity assessed with 16S rRNA gene amplicon sequencing was consistently low across crab species, and the same actinobacterial group, namely Ilumatobacter, was dominant regardless of the geographic location of the host. Using metagenomics and metatranscriptomics, we detected that these members of actinobacteria are potentially able to convert ammonia to amino acids and may help eliminate toxic sulphur compounds and carbon monoxide to which crabs are constantly exposed.

Conclusions: These results indicate that bacteria selected on gills can play a role in the adaptation of animals in dynamic intertidal ecosystems. Hence, this relationship is likely to be important in the ecological and evolutionary processes of the transition from water to air and deserves further attention, including the ontogenetic onset of this association. Video Abstract.

Keywords: Bimodal breathing; Gill system; Microbiome; Symbiosis; Terrestrialisation.

Fig. 1

Bacterial microbiome diversity associated with fiddler crab gills. a Sampling locations and representative images of crab species sampled (Austruca albimana [AA], Cranuca inversa [CI], Tubuca urvillei [TU], Austruca occidentalis [AO], and Paraleptuca chlorophthalmus [PC]). b Bipartite network shows the relationship of bacterial communities with crab species samples (coloured circles) and sediment (brown circles). Samples clustered based on their shared OTUs (small grey circles). Edges (the lines) indicate if an OTU is present in a certain sample (crabs and sediment circles), and edge colour is associated with the geographical location (blue, Republic of South Africa [ZA]; red, Kenya [KY]; yellow, Kingdom of Saudi Arabia [KSA]). c Principal component analysis of gill bacterial microbiome associated with the selected crab species and sediment from the different geographical locations (ZA, KY, and KSA). d, e Bar plots with the relative abundance of the bacterial phyla retrieved in sediments and gills. The asterisk on “Proteobacteria” indicates that the majority of the OTUs within this phylum belong to Alphaproteobacteria. f-i Alpha diversity indexes of the bacterial community of gill and sediment

Fig. 2

Identification of phylogenetic clades with distinct phylogenetic turnover patterns compared to the rest of the microbiome across all samples. The phylogenetic tree concerns all OTUs from all samples. Clades containing more than 15 OTUs are colour-coded on the phylogenetic tree, and the consensus taxonomy is given for each clade on the left with font size proportional to taxonomic breadth. The total phyloscore of each clade (i.e. the sums of the phyloscores across community pairs) is shown to the right as bars with colours matching the clades’ colours. Negative phyloscores indicate clades that contain OTUs whose closest relatives (across samples) reside at shorter phylogenetic distances than those expected by chance, and vice versa. Thus, negative phyloscores indicate clades having lower-than-expected phylogenetic turnover (i.e. clades with a niche across all crab gills) and vice versa

Fig. 3

Distribution of microbial communities over the gill lamellae of fiddler crabs. a–c Scanning electron microscope (SEM) imaging of a representative gill of the fiddler crabs Tubuca urvillei. Representative SEM images of the other crab species gills are provided in Supplementary Figure S5. a Magnification of the gill lamellae. b Detail of the bacterial layer covering the lamellae of the gill of each species investigated. Asterisks indicate the gill lamellae. c Rod-like bacteria covering the entire lamellae of the gills. d–f Transmission electron microscope imaging of fiddler crab gill (indicated by a “g” letter) of Austruca albimana: white asterisks show the gill cuticles, while orange arrows the bacterial layer. d Section of the gill that shows both side of the gill lamellae covered by bacteria. e Magnification of the bacterial layer. f Gill lamellae surface (indicated by a g letter); arrowheads: electro-dense area where bacteria are attached to the gill; black arrows: bacterial pili. g–p Localisation of bacteria in crab gill by confocal laser scanning microscopy (CLSM) of fluorescence in situ hybridisation (FISH)-stained gill lamellae of T. urvillei (g–j) and A. albimana (k–n) with the different probes (see supplementary figure S5 for the bright field and FISH negative control). o, p IMARIS 3D-structure reconstruction of the gill lamellae and the bacterial layer (frontal and lateral view) of T. urvillei. Note the absence of signal inside the gill lamellae that support the evidence that bacteria live on the surface of the lamellae without entering them. Red arrow indicates the heterogeneous inner morphology of the gill lamellae. “High CG bacteria” indicates cells with high guanine-cytosine content typical of Actinobacteria. The images are meant to be typical of the range of observations made

Fig. 4

Metagenomics distinguishes the functional repertoire of the fiddler crab gill microbiome from the neighbouring sediment microbiome. a Grouping of gill and sediment microbiome samples (2–3 independent replicates per species/site) based on the Bray-Curtis dissimilarity index calculated using abundance/coverage of a random set of 100,000 genes in the annotated gene catalogue of both microbiomes. Crab species and site samples are colour-coded by location (same as in Fig. 1), and crab names are labelled according to crab species names: T. urvillei (TU), C. inversa (CI), A. albimana (AA), A. occidentalis (AO), and P. chlorophthalmus (PC); see Supplementary Table S1 for more details. b, c The abundance of the fifty most highly represented functions based on random forest predictions in gill (b) and sediment (c) microbiomes (2–3 replicates per sample). The plots represent independent gene sets that may have a few copies with low coverage in the gill or sediment microbiome, explaining the additional x-axis labels despite these being abundant in either the gill or sediment microbiomes. Boxplots show median coverage as the middle horizontal line and interquartile ranges as boxes (whiskers extend no further than 1.5× the interquartile range). Circular symbols reveal the diversity of enriched genes, with colours reflecting the location of the sample. Mean values are shown as white coloured diamonds. Different lowercase letters at the top of each boxplot denote significance differences based on the two-tailed unpaired Wilcoxon test (p<0.05). RPKM, reads per kilobase per million mapped reads. d Bar graphs show the high proportion of genes encoding transposases among the fifty most enriched annotated KOs in the crab gill microbiomes relative to the sediment microbiome

Fig. 5

Abundance and taxonomic origin of key enzymes encoded by the gill microbiome with relevance to host physiology. a, c, d Boxplots show the normalised abundance of seven different metabolic genes in reads per kilobase per million mapped reads (RPKM) in the microbiome associated with sediment and gills of different crab species. The circular symbols show the mean RPKM value (log-scaled) for each predicted unique gene copy (per enzyme; n = 2–3 replicates) and are coloured according to the sampling location. The mean RPKM value of all genes is shown as a horizontal bar. Different lowercase letters on the x-axis indicate microbiomes with significantly different mean (RPKM) abundances as determined by a two-tailed, unpaired Wilcoxon test (p < 0.05). In all cases, gene family diversity is higher in the sediment microbiome than in the gill microbiome of all crab species. b, e, f Bar graphs show the predicted taxonomic origin (at the phylum level) of genes encoding enzymes in the nitrogen (b), sulphur (e), and carbon (f) cycles in the gill microbiome relative to the sediment microbiome. The percentages in the bar graphs show the relative abundance of each phylum based on the aggregated RPKM of all genes shown in the individual boxplots. Gene abbreviations (and corresponding KO identifiers): AMT, ammonia transporter (K03320); GDH, glutamate dehydrogenase (K15371); GS, glutamine synthetase (K01915); soxC, sulfite oxidase (K00387); SQR, sulphide:quinone oxidoreductase (K17218); psrA, polysulfide reductase subunit A (K08352); and cutL, aerobic carbon monoxide dehydrogenase, large subunit (K03520). Abbreviations of crab species: AA, A. albimana; CI, C. inversa; TU, T. urvillei; AO, A. occidentalis; and PC, P. chlorophthalmus. RPKM, reads per kilobase per million mapped reads

Fig. 6

Expression of ammonia, sulphide, and carbon monoxide detoxification pathways in the gills of wild crabs. a Taxonomic assignment of bacteria transcripts assembled from gill tissue metatranscriptomes (CI_1–CI_4) of the fiddler crab C. inversa collected from the Red Sea coast near KAUST (related to KSA samples). Bar plots show the relative proportion of total transcripts assigned to prokaryotes (details in the “Methods”) in four independent animals. b Normalised mean expression (as RPKM values log-scaled) of key enzymes involved in nitrogen (AMT, GDH, and GS), sulphur (soxC and SQR), and carbon (cutL) metabolism, most of which are derived from Actinobacteria in the case of C. inversa. Gene abbreviations (and corresponding KO identifiers): AMT, ammonia transporter (K03320); GDH, glutamate dehydrogenase (K15371); GS, glutamine synthetase (K01915); soxC, sulphite oxidase (K00387); SQR, sulphide:quinone oxidoreductase (K17218); and cutL, aerobic carbon monoxide dehydrogenase, large subunit (K03520). No psrA, polysulfide reductase subunit A (K08352), was detected. RPKM (reads per kilobase per million mapped reads) denotes normalised expression levels. The red stack of the bar plot represents Actinobacteria and the black stack of all the other bacterial phyla