Functional metagenomic investigations of microbial communities in a shallow-sea hydrothermal system

Abstract

Little is known about the functional capability of microbial communities in shallow-sea hydrothermal systems (water depth of <200 m). This study analyzed two high-throughput pyrosequencing metagenomic datasets from the vent and the surface water in the shallow-sea hydrothermal system offshore NE Taiwan. This system exhibited distinct geochemical parameters. Metagenomic data revealed that the vent and the surface water were predominated by Epsilonproteobacteria (Nautiliales-like organisms) and Gammaproteobacteria (Thiomicrospira-like organisms), respectively. A significant difference in microbial carbon fixation and sulfur metabolism was found between the vent and the surface water. The chemoautotrophic microorganisms in the vent and in the surface water might possess the reverse tricarboxylic acid cycle and the Calvin-Bassham-Benson cycle for carbon fixation in response to carbon dioxide highly enriched in the environment, which is possibly fueled by geochemical energy with sulfur and hydrogen. Comparative analyses of metagenomes showed that the shallow-sea metagenomes contained some genes similar to those present in other extreme environments. This study may serve as a basis for deeply understanding the genetic network and functional capability of the microbial members of shallow-sea hydrothermal systems.

Figure 1. Comparison of the taxonomic profiles at the genus level of two samples from the shallow-sea hydrothermal system.

The taxonomic profiles for the vent (G1 colored blue) and surface water immediately above the vent (G2 colored orange) metagenomic datasets were computed using MG-RAST and STAMP v2.0. Corrected P-values (q-values) were calculated based on Fisher’s exact test using Storey’s FDR approach. Dots on either side of the dashed trend line were enriched in one of the two samples. Labeled dots at greater distances from the dashed trend line indicate that these subsystems had greater proportional differences (%) between two metagenomes. A filter was applied to remove features with q value >0.05.

Figure 2. Representative open reading frames encoding carbon fixation, and sulfur, nitrogen and phosphorus utilization-associated functions present in the G1 and the G2 contigs.

The % identities of the homologs to the genes from the reference genomes ( N . profundicola or C . mediatlanticus as references for the G1 dataset analysis and T . crunogena as a reference for the G1 dataset analysis) are listed in the colored boxes. #N/A indicates that no best matches to T . crunogena were found.

Figure 3. Metagenomic profile comparisons of genes associated with sulfur metabolic pathways determined using STAMP analysis.

Positive differences between proportions denote greater abundances in the G1 dataset (blue), whereas negative differences between proportions show greater abundances in the G2 dataset (orange) for the given genes. Corrected P-values (q-values) were calculated based on Fisher’s exact test using Storey’s FDR approach. Features with q value <0.05 were considered significant and were thus retained.

Figure 4. Comparison of genes encoding for key enzymes in the carbon fixation pathways determined using STAMP analysis, including (A) the reverse tricarboxylic acid cycle, (B) the Calvin-Bassham-Benson cycle, (C) the 3-hydroxypropionate cycle, and (D) the reductive acetyl-CoA pathway.

Genes encoding enzymes for the G1 (blue) and G2 (orange) datasets were identified based on KEGG functions within the MG-RAST system.

Figure 5. Multidimensional scaling (MDS) plots of samples using Bray–Curtis similarity according to (A) SEED subsystem and (B) Clusters of Orthologous Groups of protein functional annotations.

Color represents different sampling areas and each habitat label type (n) is indicated [G: shallow-sea hydrothermal field (red), H: deep-sea hydrothermal field (black), O: open sea (blue), C: coastal and estuary (blue), U: other habitats in common ocean (blue), B: biofilm (orange), W: whale biofilm (green). For details, see Supplementary Table S2]. Samples from each of the respective environments clustered together based on their functional profile. The stress values are reported in the top right corner of the figures and represent the goodness-of-fit.

Figure 6. Selected COGs in the order of their contribution to the difference between two samples as assessed using similarity percentage analysis.

Each row represents the relative frequency of genes among samples (A: G1, B: G2, C: H1, D: H2). The value of normalized abundance of genes (relative to the single-copy gene RecA) is assigned with a color relative to the maximum value among all comparisons of each COG, from white to red. The colors represent 0 (white) to maximum (MAX, red) increments of MAX/10. The higher values indicate greater gene abundance. COG descriptions are listed along the row. Among these COG categories, the maximum RecA-normalized gene abundance is shown in a bracket.