Microbial Ecology of Sulfur Biogeochemical Cycling at a Mesothermal Hot Spring Atop Northern Himalayas, India

Abstract

Sulfur related prokaryotes residing in hot spring present good opportunity for exploring the limitless possibilities of integral ecosystem processes. Metagenomic analysis further expands the phylogenetic breadth of these extraordinary sulfur (S) metabolizing microorganisms as well as their complex metabolic networks and syntrophic interactions in environmental biosystems. Through this study, we explored and expanded the microbial genetic repertoire with focus on S cycling genes through metagenomic analysis of S contaminated hot spring, located at the Northern Himalayas. The analysis revealed rich diversity of microbial consortia with established roles in S cycling such as Pseudomonas, Thioalkalivibrio, Desulfovibrio, and Desulfobulbaceae (Proteobacteria). The major gene families inferred to be abundant across microbial mat, sediment, and water were assigned to Proteobacteria as reflected from the reads per kilobase (RPKs) categorized into translation and ribosomal structure and biogenesis. An analysis of sequence similarity showed conserved pattern of both dsrAB genes (n = 178) retrieved from all metagenomes while other S disproportionation proteins were diverged due to different structural and chemical substrates. The diversity of S oxidizing bacteria (SOB) and sulfate reducing bacteria (SRB) with conserved (r)dsrAB suggests for it to be an important adaptation for microbial fitness at this site. Here, (i) the oxidative and reductive dsr evolutionary time-scale phylogeny proved that the earliest (but not the first) dsrAB proteins belong to anaerobic Thiobacillus with other (rdsr) oxidizers, also we confirm that (ii) SRBs belongs to δ-Proteobacteria occurring independent lateral gene transfer (LGT) of dsr genes to different and few novel lineages. Further, the structural prediction of unassigned DsrAB proteins confirmed their relatedness with species of Desulfovibrio (TM score = 0.86, 0.98, 0.96) and Archaeoglobus fulgidus (TM score = 0.97, 0.98). We proposed that the genetic repertoire might provide the basis of studying time-scale evolution and horizontal gene transfer of these genes in biogeochemical S cycling.

Keywords: biogeochemical cycle; evolution; hot spring; metagenomics; sulfur spring.

FIGURE 1

(A) Geographical location of Himachal Pradesh, Northern Himalayas. (B) Khirganga hot spring as shown in the maps are located in Parvati valley in Kullu district of Himachal Pradesh, India. (C) The different habitats from where the samples were collected are shown: microbial mat (green), sediment (yellow), and water (blue). Samples were collected in replicates from two outlets located 10-m distance apart. (D) The scanning electron micrographs of different habitat samples shown with arrow demarcating the filamentous (cyanobacteria), cocci-shaped and rod-shaped bacteria (SRB) in pink, yellow, and white, respectively.

FIGURE 2

Relative abundance of phylum and genera. (A) The stacked bar representation shows the dominating phylum in all three habitats. The dendrograms show hierarchical clustering between species and samples. (B) The abundant common genera Pseudomonas (20–60%), Desulfobulbaceae_unclasssifed (15–20%), Burkholderia, Desulfovibrio, and Thioalkalivibrio (1–20%) are shown here relative abundance × log_x scale. (C) Taxa based functional profiles demonstrating the major phylum contributing toward COG subsystems and percentage of proteins annotated within each COG category for the habitat sites. The distribution of the RPKs were mapped in accordance to the member abundance in the habitats. Dots were colored according to the phylum.

FIGURE 3

(A) Reconstruction of top 50 pathways annotated using KEGG automatic annotation server. Heatmap matrix representation and clustering was performed by using “pheatmap” package (Kolde and Kolde, 2015) in R (R Development Core Team, 2011). (B) The sulfate reduction pathway involved a group of reductases, kinases, and transferases with the product chemical structures generated through chemDraw7 and Inkscape v0.9 (Inkscape Project, 2020). (C) The gene copy number of both sulfate reduction and sulfide oxidation pathway that were partitioned in different habitats showed here using ggplot2 in R (R Development Core Team, 2011).

FIGURE 4

Sequence similarity network analyses. (A) Diversity of sulfate reduction genes of both assimilatory and dissimilatory pathways in microbial mat (diamond-shaped), sediment (square), and water (spherical) habitats visualized in cytoscape v3.7.1. Highlighted only the classified taxa, where color cyan belong to SRBs and yellow to SOBs. The network was set at threshold e-value cutoff of 1e-30 and nodes represent sequences connected through edges if the similarity exceeds the cutoff score. Here, the clusters and isolated nodes were showing the conserved pattern and diversified pattern of the proteins significantly playing an important role in sulfate reduction. (B) Topological properties of the similarity networks: degree distribution, average clustering coefficient, average neighborhood connectivity, and closeness centrality are plotted against the number of neighbors. The power law fit curves are shown within each graph. (C) Habitat vs. habitat dN/dS values of all S cycle genes were estimated and plotted using xy-plot in R (R Development Core Team, 2011).

FIGURE 5

Divergence estimation over time. Reconstruction of the phylogenic tree of optimized full length DsrAB and AsrABC subunits in three habitats using PhyloBayes with the CAT-GTR model. The highlighted squares consist of clades with proteins that were remained unclassified through nr database. (A) (i) Among, 78 DsrA nodes that showed here the earliest evolution of the rDsr oxidative proteins occurred in Thiobacillus sp. (ii) 72 nodes of DsrB proteins with similar results. (B) (i) 38 nodes of AsrA and (ii) 21 nodes of AsrB were also compared as a control for branch length shown here.

FIGURE 6

Structural similarity and analogy of unclassified proteins from PDB (Protein data bank) using i-TASSER. (A) DsrA protein subunits: (i) np252939 and (ii); (iii) np500880; np481424 showing Tm-align similarity with PDB ids 3or1 and 2v4j (Desulfovibrio gigas and D. vulgaris), respectively; with SF4 (iron S cluster) ligand binding sites (B) DsrB protein subunits (i) np24977, np24617, np116484 and (ii) np199275, np59533 showing Tm-align similarity with PDB ids 3c7b; with siroheme ligand binding sites.