Molecular characterization of Bathymodiolus mussels and gill symbionts associated with chemosynthetic habitats from the U.S. Atlantic margin

Abstract

Mussels of the genus Bathymodiolus are among the most widespread colonizers of hydrothermal vent and cold seep environments, sustained by endosymbiosis with chemosynthetic bacteria. Presumed species of Bathymodiolus are abundant at newly discovered cold seeps on the Mid-Atlantic continental slope, however morphological taxonomy is challenging, and their phylogenetic affinities remain unestablished. Here we used mitochondrial sequence to classify species found at three seep sites (Baltimore Canyon seep (BCS; ~400m); Norfolk Canyon seep (NCS; ~1520m); and Chincoteague Island seep (CTS; ~1000m)). Mitochondrial COI (N = 162) and ND4 (N = 39) data suggest that Bathymodiolus childressi predominates at these sites, although single B. mauritanicus and B. heckerae individuals were detected. As previous work had suggested that methanotrophic and thiotrophic interactions can both occur at a site, and within an individual mussel, we investigated the symbiont communities in gill tissues of a subset of mussels from BCS and NCS. We constructed metabarcode libraries with four different primer sets spanning the 16S gene. A methanotrophic phylotype dominated all gill microbial samples from BCS, but sulfur-oxidizing Campylobacterota were represented by a notable minority of sequences from NCS. The methanotroph phylotype shared a clade with globally distributed Bathymodiolus spp. symbionts from methane seeps and hydrothermal vents. Two distinct Campylobacterota phylotypes were prevalent in NCS samples, one of which shares a clade with Campylobacterota associated with B. childressi from the Gulf of Mexico and the other with Campylobacterota associated with other deep-sea fauna. Variation in chemosynthetic symbiont communities among sites and individuals has important ecological and geochemical implications and suggests shifting reliance on methanotrophy. Continued characterization of symbionts from cold seeps will provide a greater understanding of the ecology of these unique environments as well and their geochemical footprint in elemental cycling and energy flux.

Fig 1. Map of sampling locations of Bathymodiolus spp. in this study.

Fig 2. Bayesian phylogeny constructed from mitochondrial COI+ND4 including 21 mussels collected from Mid-Atlantic seep sites (MAS).

Posterior probabilities above 0.90 = *; above 0.95 = **. S3 Table contains Genbank accession numbers of individuals not collected in this study.

Fig 3

Minimum spanning networks created from A) COI and B) ND4 sequences generated from Mid-Atlantic seep mussels. Each circle represents a unique haplotype. Size is proportional to the number of mussels sharing the haplotype. Sample sizes ≥ 10 are reported inside the circles. Hash marks are mutational steps between haplotypes. GOM = B. childressi from several Gulf of Mexico sample sites, BCS = Baltimore Canyon Seep, NCS = Norfolk Canyon Seep, CTS = Chincoteague Seep, GOC = B. mauritanicus from Gulf of Cadiz, WAF = B. cf. mauritanicus from West Africa, BAP = B. sp B from the Barbados Accretionary Prism. See Table 1, S1, and S4 Tables for sample information.

Fig 4. 16S Bayesian phylogram based upon 16S sequences from known endosymbionts from bathymodiolins and other deep-sea hosts.

The nodes are labeled with the ML probabilities based on 500 bootstrap replicates before the slash and Bayesian posterior probabilities after the slash. If the node placement did not agree between the two trees, a “-” is indicated before the slash. The branch tips are labeled with the name of the host species. If more than one sequence from that host is represented in that clade, the sample size is in parentheses after the name. A Gammaproteobacteria thiotroph from a hydrocarbon seep tubeworm, Escarpia sp. (JF969172), was used as an outgroup. Our consensus sequence, M, is in bold. S8 Table contains Genbank accession numbers of all individuals in the tree.

Fig 5. 16S Bayesian phylogram based upon 16S sequences for Campylobacterota.

The nodes are labeled with maximum likelihood bootstrap probabilities based on 500 bootstrap replicates before the slash and Bayesian posterior probabilities after the slash. If the node placement did not agree between the two trees, a “-” is indicated before the slash. The branch tips are labeled with the name of the host species. If more than one sequence from that host is represented, the samples size is in parentheses after the name. In cases where the host is not a marine organism, the host and symbiont are both listed, separated by a semi-colon. Sulfurovum lithotrophicum, Arcobacter marinus, and Sulfurospirillum deleyianum were collected from sediment. S8 Table contains Genbank accession numbers of individuals in the trees. Phylotypes C1 and C2 are in bold. *The clade including C1 also includes symbionts isolated from Bathymodiolus azoricus, B. childressi, B. manuensis, B. mauritanicus, B. sp. 9 South, and B. sp. Pakistan.

Fig 6. Phylum-level diversity per sample recovered by four primer sets.

Phyla abundances as assigned with Mothur and the SILVA v132 reference database. Each panel represents a different primer set (ps1-ps4). The y-axis represents number of OTUs x1000. 1 = MAS100, 2 = MAS109, 3 = MASM22, 4 = MASM30, 5 = MASM34, 6 = MASM36, 7 = MASM45, 8 = MASM5, 9 = MAS538, 10 = MAS562. NCS samples are shown in bold on the x axis.

Fig 7. Hierarchical distribution of bacterial diversity at each site.

The top circle represents Baltimore Canyon site (BCS) and the bottom circle represents Norfolk Canyon site (NCS). The taxonomic hierarchy proceeds outward. Primer set 1 (ps1) is shown. The results from ps2 and ps4 were similar, but ps3 lacked Campylobacterota due to substantial primer mismatches.