Evolutionary strategies of viruses, bacteria and archaea in hydrothermal vent ecosystems revealed through metagenomics

Abstract

The deep-sea hydrothermal vent habitat hosts a diverse community of archaea and bacteria that withstand extreme fluctuations in environmental conditions. Abundant viruses in these systems, a high proportion of which are lysogenic, must also withstand these environmental extremes. Here, we explore the evolutionary strategies of both microorganisms and viruses in hydrothermal systems through comparative analysis of a cellular and viral metagenome, collected by size fractionation of high temperature fluids from a diffuse flow hydrothermal vent. We detected a high enrichment of mobile elements and proviruses in the cellular fraction relative to microorganisms in other environments. We observed a relatively high abundance of genes related to energy metabolism as well as cofactors and vitamins in the viral fraction compared to the cellular fraction, which suggest encoding of auxiliary metabolic genes on viral genomes. Moreover, the observation of stronger purifying selection in the viral versus cellular gene pool suggests viral strategies that promote prolonged host integration. Our results demonstrate that there is great potential for hydrothermal vent viruses to integrate into hosts, facilitate horizontal gene transfer, and express or transfer genes that manipulate the hosts' functional capabilities.

Figure 1. Pie charts showing breakdown of read classification for the cellular metagenome (A) and the viral metagenome (B) according to annotation by the M5NR database.

Reads were annotated with a minimum e-value cutoff of 1e-05.

Figure 2. Recruitment plot of metagenomic reads to Caminibacter mediatlanticus TB-2.

Cellular metagenomic reads were mapped to the longest contig of the draft genome of C. mediatlanticus TB-2, with percent similarity on the y-axis and base pair numbers on the x-axis (A). Coverage plot of read recruitment is shown per base pair, with blue line showing actual coverage and green line showing a convolution function of the coverage plot using a weighting of 50000 (B). Percent GC plot for the same contig is shown on the same scale, with base pair numbers marked below (C), and are annotated with CRISPR loci and recombinases or integrases found on the contig. Orange shading shows the location of CRISPR loci on the genome; green shading shows the location of two metagenomic islands.

Figure 3. Functional comparisons of the hydrothermal vent cellular and viral subset metagenomes according to the SEED subsystems and Clusters of Orthologous Groups (COG) databases.

Metagenomes were annotated in MG-RAST with a minimum e-value of 1e-03 and a minimum identity cutoff of 60%. A single asterisk indicates a significant difference in abundance between the viral subset and the cellular metagenome. A) Matches to the SEED subsystems database; B) matches to the COG database.

Figure 4. Functional comparisons of the hydrothermal vent cellular and viral subset metagenomes according to the KEGG Orthology annotation system.

Metagenomes were annotated in MG-RAST with a minimum e-value of 1e-03 and a minimum identity cutoff of 60%. A single asterisk indicates a significant difference in abundance between the viral subset and the cellular metagenome. A) Matches to the KEGG Orthology database; B) Matches to the energy metabolism category of the KEGG Orthology database.

Figure 5. Histogram of dN/dS ratios for each metagenome.

A total of 863 genes were included for the cellular metagenome calculation, 191 for the viral metagenome, and 64 for the viral subset. Values are shown only for genes that had a minimum depth coverage of 5 and minimum nucleotide coverage of 100. Frequency values are normalized by percent. Bins are scaled in increments of 0.1 until 1, and then in increments of 0.5. Inset shows mean and 95% confidence intervals for calculated dN/dS for all three data sets, indicating that the average cellular dN/dS is significantly greater than the average dN/dS for the viral metagenome.