Selection Is a Significant Driver of Gene Gain and Loss in the Pangenome of the Bacterial Genus Sulfurovum in Geographically Distinct Deep-Sea Hydrothermal Vents

Abstract

Microbial genomes have highly variable gene content, and the evolutionary history of microbial populations is shaped by gene gain and loss mediated by horizontal gene transfer and selection. To evaluate the influence of selection on gene content variation in hydrothermal vent microbial populations, we examined 22 metagenome-assembled genomes (MAGs) (70 to 97% complete) from the ubiquitous vent Epsilonbacteraeota genus Sulfurovum that were recovered from two deep-sea hydrothermal vent regions, Axial Seamount in the northeastern Pacific Ocean (13 MAGs) and the Mid-Cayman Rise in the Caribbean Sea (9 MAGs). Genes involved in housekeeping functions were highly conserved across Sulfurovum lineages. However, genes involved in environment-specific functions, and in particular phosphate regulation, were found mostly in Sulfurovum genomes from the Mid-Cayman Rise in the low-phosphate Atlantic Ocean environment, suggesting that nutrient limitation is an important selective pressure for these bacteria. Furthermore, genes that were rare within the pangenome were more likely to undergo positive selection than genes that were highly conserved in the pangenome, and they also appeared to have experienced gene-specific sweeps. Our results suggest that selection is a significant driver of gene gain and loss for dominant microbial lineages in hydrothermal vents and highlight the importance of factors like nutrient limitation in driving microbial adaptation and evolution.IMPORTANCE Microbes can alter their gene content through the gain and loss of genes. However, there is some debate as to whether natural selection or neutral processes play a stronger role in molding the gene content of microbial genomes. In this study, we examined variation in gene content for the Epsilonbacteraeota genus Sulfurovum from deep-sea hydrothermal vents, which are dynamic habitats known for extensive horizontal gene transfer within microbial populations. Our results show that natural selection is a strong driver of Sulfurovum gene content and that nutrient limitation in particular has shaped the Sulfurovum genome, leading to differences in gene content between ocean basins. Our results also suggest that recently acquired genes undergo stronger selection than genes that were acquired in the more distant past. Overall, our results highlight the importance of natural selection in driving the evolution of microbial populations in these dynamic habitats.

Keywords: hydrothermal vents; metagenomics; pangenome.

FIG 1

Pangenomic structure of Sulfurovum populations in the two vent fields. (A) Anvi’o image of each Sulfurovum metagenome-assembled genome (MAG) represented by a gray ring in the circle, ordered from MAGs with the most (outermost) to the least (innermost) gene clusters. The black boxes in each genome represent open reading frames (ORFs) recovered in that MAG. Boxes that align create a gene cluster for that ORF. The outermost ring in blue represents the number of MAGs containing the corresponding ORF. Other properties are shown in the upper right quadrant, where “num singletons” represents the number of genes found in only one MAG and “num gene clusters” represents the number of total gene clusters found in that MAG. The most common gene clusters (green) and phosphate-related gene clusters (blue) are highlighted. (B) Proportion of gene clusters present in one Sulfurovum MAG compared to every other Sulfurovum MAG. MAGs are hierarchically clustered by these proportions. Gray branches show Axial Seamount lineages, whereas golden branches show Mid-Cayman Rise lineages.

FIG 2

ORF annotations across the Sulfurovum metagenome-assembled genomes (MAGs) as a function of gene cluster frequency. (A) Each bar represents a cluster of orthologous groups (COG) category for the gene cluster of interest in the Sulfurovum MAGs. Each color corresponds to each COG category. Gene clusters with unknown annotations are excluded. N/A, no COG category assigned. (B through E) The proportion of ORFs that are in each of four COG categories is shown as a function of the gene cluster frequency (number of MAGs that contains the gene cluster). The x axis represents the number of MAGs sharing a specific gene cluster. The y axis represents the proportion of gene clusters that fall within a specific COG category out of all gene clusters shared among that number of MAGs. These four COG categories were included because they exhibited the strongest differences in abundance between high-frequency and low-frequency genes.

FIG 3

The cumulative distribution function (CDF) value for the proportion of genes present in Axial Seamount metagenome-assembled genomes (MAGs) relative to Mid-Cayman Rise MAGs across different clusters of orthologous groups (COG) categories. A higher value means that the open reading frame (ORF) is found in more Axial MAGs than expected, and vice versa. Some COG categories are not shown because they were not represented in enough MAGs (see Results). The blue dashed line represents a CDF value of 0.95, and a statistical significance cutoff for ORFs represented more in Axial MAGs (P < 0.05). The red dashed line represents the cutoff for ORFs represented more in Mid-Cayman Rise MAGs. Phosphate-related genes are shown in black under “Inorganic ion transport and metabolism.”

FIG 4

(A) The estimated pN/pS ratio of each identified open reading frame (ORF) within each gene cluster as a function of gene cluster frequency in the pangenome. The x axis represents the number of metagenome-assembled genomes (MAGs) sharing a specific gene cluster. The inset in panel A shows the trend of the mean of pN/pS ratios from lowest to highest frequency. The x axis is the same as in the larger figure. The dashed red line represents the value of pN/pS = 1. A similar plot is shown in panel B, where all ORFs with certain frequencies are grouped together (N), showing that ORFs within core gene clusters have lower pN/pS ratios.

FIG 5

Gene-specific sweep signatures based on single nucleotide variants (SNVs) and P values. (A) The P value for each contig is plotted relative to the number of actual SNVs in the contig. Each point represents a single contig. The plots are separated according to the number of metagenome-assembled genomes (MAGs) in which that specific gene cluster was found. (B) P values as a function of gene cluster frequency in the pangenome. The x axis represents the number of MAGs sharing a specific gene cluster. (C) The SNV density for each MAG at either Axial Seamount of Mid-Cayman Rise. The MAG numbering is indicated in Table S2 in the supplemental material.