Sulfur cycling and host-virus interactions in Aquificales-dominated biofilms from Yellowstone's hottest ecosystems

Abstract

Modern linkages among magmatic, geochemical, and geobiological processes provide clues about the importance of thermophiles in the origin of biogeochemical cycles. The aim of this study was to identify the primary chemoautotrophs and host-virus interactions involved in microbial colonization and biogeochemical cycling at sublacustrine, vapor-dominated vents that represent the hottest measured ecosystems in Yellowstone National Park (~140 °C). Filamentous microbial communities exposed to extreme thermal and geochemical gradients were sampled using a remotely operated vehicle and subjected to random metagenome sequencing and microscopic analyses. Sulfurihydrogenibium (phylum Aquificae) was the predominant lineage (up to 84% relative abundance) detected at vents that discharged high levels of dissolved H₂, H₂S, and CO₂. Metabolic analyses indicated carbon fixation by Sulfurihydrogenibium spp. was powered by the oxidation of reduced sulfur and H₂, which provides organic carbon for heterotrophic community members. Highly variable Sulfurihydrogenibium genomes suggested the importance of intra-population diversity under extreme environmental and viral pressures. Numerous lytic viruses (primarily unclassified taxa) were associated with diverse archaea and bacteria in the vent community. Five circular dsDNA uncultivated virus genomes (UViGs) of ~40 kbp length were linked to the Sulfurihydrogenibium metagenome-assembled genome (MAG) by CRISPR spacer matches. Four UViGs contained consistent genome architecture and formed a monophyletic cluster with the recently proposed Pyrovirus genus within the Caudovirales. Sulfurihydrogenibium spp. also contained CRISPR arrays linked to plasmid DNA with genes for a novel type IV filament system and a highly expressed β-barrel porin. A diverse suite of transcribed secretion systems was consistent with direct microscopic analyses, which revealed an extensive extracellular matrix likely critical to community structure and function. We hypothesize these attributes are fundamental to the establishment and survival of microbial communities in highly turbulent, extreme-gradient environments.

Fig. 1. Microbial streamer communities collected from sulfidic hydrothermal vents (Stevenson Island Deep Hole, Yellowstone Lake, WY) contain morphologically diverse cells and accumulate elemental sulfur.

A Bathymetric map of Yellowstone Lake relative to the boundary (red dashed line) of the most recent caldera-forming eruption. B Bathymetric map of three sampling sites within the Deep Hole east of Stevenson Island. C A thermophilic sulfur-rich biofilm sampled at one hydrothermal vent (Sample 2016_B02). D Scanning electron microscopy demonstrates morphological diversity of different microorganisms. E Elemental analysis of sulfur (blue) within the largest (~3 µm diameter) filamentous cells of the vent biofilm (more detail in Supplementary Figure 5).

Fig. 2. Community composition and key pathways of Sulfurihydrogenibium-dominated hydrothermal streamers.

Relative abundance is shown for biofilm metagenome from vent sample 2016_B01, which was the cleanest unmanipulated sample retrieved (the relative abundance of other samples is provided in Supplementary Table 1). Concentrations of H₂, H₂S, and CO₂ were previously determined for YLake vents in the Deep Hole [33, 35, 39]. Arrows either represent (i) autotrophic pathways, (ii) enzymatic conversions between sulfur species, or (iii) hydrogenases. Arrow color corresponds to the microbial lineage encoding that particular enzyme; arrow thickness is weighted by mean coverage for the contig containing the relevant gene/s. [apr = adenylylsulfate reductase; sat = sulfate adenylyltransferase; cys = adenylylsulfate kinase; phs = polysulfide reductase; dsr = dissimilatory sulfite reductase; sdo = sulfur dioxygenase; sre = sulfur reductase; sqr = sulfide:quinone oxidoreductase].

Fig. 3. Virus–host diversity and genomic characteristics of Sulfurihydrogenibium viruses.

A Family-level community structure of recovered viral sequences in the black and white chart is related to microbial host at the class level in colored pie charts. B Maximum-likelihood phylogenetic tree of deduced capsid protein sequences from Sulfurihydrogenibium viruses showing a monophyletic relationship with the newly proposed “Ca. Pyrovirus” genus within the Caudovirales [66]. Entries with red circles represent circularized viral genomes; entries with blue circles had CRISPR-resolved links to a Sulfurihydrogenibium host. Vibrio phage corresponds to the 1.232.O._10N.261.51.E11 identifier; Escherichia phage corresponds to the vB_EcoP_PTXU04 identifier. Contig IDs corresponding to Sulfurihydrogenibium viruses are listed in Supplementary Table 16. C Genome architecture and annotations of circular UViGs that had CRISPR links to a Sulfurihydrogenibium host.

Fig. 4. Taxonomic identity and sequence character of contigs within the Sulfurihydrogenibium MAG.

Contig sequences were clustered based on mean coverage values (shown as nX) and tetranucleotide frequencies. Sulfurihydrogenibium contigs (n = 680) exhibited consistent G + C content (%) with a wide range of coverage values from 32 to 4500X (white asterisk indicates a group of five contigs with coverage values >4500X). 35 putative viruses contained within the same tetranucleotide cluster are indicated, and line up with positive BLASTn hits to CRISPR spacer sequences (i.e., “CRISPR Spacer Hits”; Supplementary Table 3). Six complete CRISPR arrays detected in the assembled metagenome (Supplementary Table 3) are also indicated [SNV density = single nucleotide variant density, per kilobase].

Fig. 5. Early evolved filament system in Sulfurihydrogenibium spp. is encoded and transcribed from CRISPR-targeted plasmid.

A Plasmid sequences from the Deep Hole and Liberty Cap are compared and RNA expression is reported as reads per kilobase per megabase (RPKM) for the Liberty Cap sequence. Hash marks in the beta-barrel porin bar indicate RNA expression of 3.6, beyond the limits of the y-axis. B Four CRISPR target sites are indicated along the plasmid sequence from the Deep Hole. C A conserved gene neighborhood in both plasmid sequences encodes a complete TFF system [OM outer membrane, IM inner membrane, IMP inner membrane platform]. D, E Scanning electron micrographs of biofilm matrices at (D) the Deep Hole and (E) Liberty Cap. F Bayesian phylogeny of deduced secretin protein sequences from Sulfurihydrogenibium plasmids detected in three regions of YNP. Posterior probabilities are indicated by circles at nodes. Asterisk indicates the clade containing secretin genes from the genomic (non-plasmid) DNA of Sulfurihydrogenibium.