Metagenome-assembled genomes provide new insight into the microbial diversity of two thermal pools in Kamchatka, Russia

Abstract

Culture-independent methods have contributed substantially to our understanding of global microbial diversity. Recently developed algorithms to construct whole genomes from environmental samples have further refined, corrected and revolutionized understanding of the tree of life. Here, we assembled draft metagenome-assembled genomes (MAGs) from environmental DNA extracted from two hot springs within an active volcanic ecosystem on the Kamchatka peninsula, Russia. This hydrothermal system has been intensively studied previously with regard to geochemistry, chemoautotrophy, microbial isolation, and microbial diversity. We assembled genomes of bacteria and archaea using DNA that had previously been characterized via 16S rRNA gene clone libraries. We recovered 36 MAGs, 29 of medium to high quality, and inferred their placement in a phylogenetic tree consisting of 3,240 publicly available microbial genomes. We highlight MAGs that were taxonomically assigned to groups previously underrepresented in available genome data. This includes several archaea (Korarchaeota, Bathyarchaeota and Aciduliprofundum) and one potentially new species within the bacterial genus Sulfurihydrogenibium. Putative functions in both pools were compared and are discussed in the context of their diverging geochemistry. This study adds comprehensive information about phylogenetic diversity and functional potential within two hot springs in the caldera of Kamchatka.

Figure 1

Sampling locations in the Uzon Caldera, Kamchatka Russia. DNA had been extracted in 2009 by Burgess et al. from sediment samples of two active thermal pools Arkashin Schurf (b) and Zavarzin Spring (c). Photos were taken by Dr. Russell Neches (ORCID: 0000-0002-2055-8381) during an expedition in 2012 and he granted permission through a CC-BY license 4.0. Maps were plotted in R v. 3.4.0 with the package ‘ggmap’ v. 2.6.1.

Figure 2

Placement of the MAGs into their phylogenetic context. Taxonomy of the MAGs (metagenome-assembled genomes) was refined by placing them into a phylogenetic tree using PhyloSift v. 1.0.1 with its updated markers database for the alignment and RAxML v. 8.2.10 on the CIPRES web server for the tree inference. This tree includes the 36 MAGs (red dots), all taxa previously identified by Burgess et al. (2012) with complete genomes available on NCBI (n = 148), and 3,102 archaeal (yellow) and bacterial (grey) genomes previously used in Hug et al. (2016)^–. The complete tree in Newick format and its alignment of 37 concatenated marker genes can be found on Figshare^,. Branches with MAGs found in Arkashin Schurf (ARK) and Zavarzin Spring (ZAV) are enlarged (orange nodes). Blue: taxa from Burgess et al. (2012), black: taxa from Hug et al. (2016). GCA IDs from NCBI are shown for the closest neighbours of the MAGs. (a) Microbial tree of life, reconstructed with genomes representing taxa reported in Burgess et al. highlighting the placement of MAGs in this study; (b) Dictyoglomales, Thermoanaerobacteriales, Caldisericales, and Mesoaciditogales; (c) Nitrosphaera, Bathyarchaeota, Korarchaeota, and Crenarchaeota; (d) Chloroflexales; (e) Euryarchaeota; and (f) Deferribacteriales, Desulfobacteriales, and Aquificales. ARK-02, ARK-03, ARK-04, ZAV-04, ZAV-08, ZAV-09, and ZAV-15 can be found in Supplementary Figure S1.

Figure 3

Shared phyla between MAGs and different sequencing methods. Venn diagrams depict the number of shared phyla observed between metagenome assembled genomes (MAGs) and different methods of sequencing and taxonomic assignment for (a) Arkashin Schurf (ARK) and (b) Zavarzin Spring (ZAV). The different methods include the Ribosomal Database Project v. 11.5 inferred taxonomy of the 16S rRNA gene Sanger clone libraries prepared by Burgess et al., the Kaiju v. 1.6.2 inferred taxonomy for the Sanger metagenomes prepared by TIGR and the Kaiju v. 1.6.2 inferred taxonomy for the Solexa reads which were later assembled to bin the MAGs. The different circles represent the 16S rRNA genes (blue), the MAGs (yellow), the Sanger metagenomes (orange) and the Solexa reads (magenta).

Figure 4

Pangenomic comparison of shared genera between pools. Desulfurella genera (a) and Sulfurihydrogenibium genera (b) identified in both Arkashin Schurf (ARK) and Zavarzin Spring (ZAV) are visualized respectively in anvi’o against reference genomes downloaded from NCBI. ARK-08 and ZAV-10 were compared to three representative Desulfurella genomes including D. multipotens (GCA_900101285.1), D. acetivorans (GCA_000517565.1) and D. amilsii (GCA_002119425.1), while bins ARK-13 and ZAV-16 were compared to four representative Sulfurihydrogenibium genomes including S. subterraneum (GCA_000619805.1), S. yellowstonense (GCA_000173615.1), S. azorense (GCA_000021545.1) and S. sp. YO3AOP1 (GCA_000020325.1). Genomes are arranged based on a phylogenetic tree of shared single-copy core genes produced in anvi’o using FastTree v. 2.1. Gene clusters have been grouped into categories based on presence/absence including: ‘Single-copy core genes’ (gene clusters representing # genes from Campbell et al.), ‘Universally shared’ (gene clusters present in all genomes), ‘Often Shared’ (gene clusters present in two or more genomes) and ‘Unique’ (gene clusters present in only one genome). Gene calls were annotated in anvi’o using NCBI’s Clusters of Orthologous Groups (COG’s). Gene clusters with an assigned NCBI COG are indicated in black.