Metagenomes from high-temperature chemotrophic systems reveal geochemical controls on microbial community structure and function

Abstract

The Yellowstone caldera contains the most numerous and diverse geothermal systems on Earth, yielding an extensive array of unique high-temperature environments that host a variety of deeply-rooted and understudied Archaea, Bacteria and Eukarya. The combination of extreme temperature and chemical conditions encountered in geothermal environments often results in considerably less microbial diversity than other terrestrial habitats and offers a tremendous opportunity for studying the structure and function of indigenous microbial communities and for establishing linkages between putative metabolisms and element cycling. Metagenome sequence (14-15,000 Sanger reads per site) was obtained for five high-temperature (>65 degrees C) chemotrophic microbial communities sampled from geothermal springs (or pools) in Yellowstone National Park (YNP) that exhibit a wide range in geochemistry including pH, dissolved sulfide, dissolved oxygen and ferrous iron. Metagenome data revealed significant differences in the predominant phyla associated with each of these geochemical environments. Novel members of the Sulfolobales are dominant in low pH environments, while other Crenarchaeota including distantly-related Thermoproteales and Desulfurococcales populations dominate in suboxic sulfidic sediments. Several novel archaeal groups are well represented in an acidic (pH 3) Fe-oxyhydroxide mat, where a higher O2 influx is accompanied with an increase in archaeal diversity. The presence or absence of genes and pathways important in S oxidation-reduction, H2-oxidation, and aerobic respiration (terminal oxidation) provide insight regarding the metabolic strategies of indigenous organisms present in geothermal systems. Multiple-pathway and protein-specific functional analysis of metagenome sequence data corroborated results from phylogenetic analyses and clearly demonstrate major differences in metabolic potential across sites. The distribution of functional genes involved in electron transport is consistent with the hypothesis that geochemical parameters (e.g., pH, sulfide, Fe, O2) control microbial community structure and function in YNP geothermal springs.

Figure 1. Habitat context and geothermal site characteristics.

Site photographs and scanning electron micrographs (SEM) of microbial mats and solid phases associated with each geothermal sample used for metagenome sequencing (map of Yellowstone National Park and site locations shown in top left panel). A. Crater Hills (CH, gold); B. Norris Geyser Basin (NGB, red); C. Joseph's Coat Hot Springs (JCHS, blue); D. Mammoth Hot Springs (MHS, green). E. Calcite Springs (CS, violet).

Figure 2. Phylogenetic analysis of metagenome sequence data.

Binning of metagenome sequence reads (left column) from Crater Hills (gold), Norris Geyser Basin (red), Joseph's Coat Springs (blue), Mammoth Hot Springs (green) and Calcite Springs (violet) (with blastn similarity scores (E-values) of <10⁻¹⁰) to closest reference microbial genomes (abbreviations below). Environmental sequence reads were further categorized based on nucleotide identity ranging from 47–100% (shaded from light to dark, legend shown only for MHS). Fragment recruitment (right column) of metagenome sequence reads to reference microbial genomes is plotted across each reference genome (x-axis) at a nucleotide identity ranging from 50–100% (y-axis). Reference genomes: MK1 = Metallosphaera sp. str. MK1 (partial genome sequence) ; AS  = Acidilobus sulfurireducens (partial genome sequence) ; SS  = Sulfolobus solfataricus ; CM  = Caldivirga maquilingensis ; AP  = Aeropyrum pernix ; PA  = Pyrobaculum arsenaticum ; TV  = Thermoplasma volcanium ; NM  = Nitrosopumilus maritimus –; Hyd  = Hydrogenobaculum sp. Y04AAS1 ; GK  = Geobacillus kaustophilus ; TP  = Thermofilum pendens; CS  = Caldicellulosiruptor saccharolyticus ; SY03  = Sulfurihydrogenibium sp. Y03AOP1 , ; SY  = Sulfurihydrogenibium yellowstonensis , ; TA  = Thermus aquaticus Y5.1 MC23; TL  = Thermotoga lettingae .

Figure 3. Nucleotide word frequency plots and phylogenetic analysis of metagenome assemblies.

Nucleotide word frequency principal component analysis (PCA) of assembled metagenome sequence data (contigs>1500 bp) from five chemotrophic geothermal habitats in YNP: A. Metagenome sequence colored by site (Crater Hills  =  gold, Norris Geyser Basin  =  red, Joseph's Coat  =  blue, Mammoth Hot Springs  =  green, Calcite Springs  =  violet). B. Identical PCA orientation of metagenome sequence observed in Panel A, but colors now designate phylogenetic affiliation at the order level (Sulfolobales  =  gold; Desulfurococcales  =  light blue; Thermoproteales  =  dark blue; Aquificales  =  green; Thermales  =  violet; Unassigned  =  black), and C. Identical PCA orientation with phylogenetic classification at the domain-level (Archaea  =  gold, Bacteria  =  violet).

Figure 4. Functional gene analysis.

Two-way clustering of biochemical pathways that contributed most to the variability between sites based on PCA analysis. A. Comparison of metagenomes and relevant reference genomes based on pathway completeness. Reference genomes: Sulfolobus solfataricus P2, Metallosphaera sedula DSM 5348, Pyrobaculum arsenaticum, Thermus aquaticus Y5.1 MC23, Sulfurihydrogenibium sp. YO3AOP1. B. Comparison of metagenomes based on the median number of blast hits to enzymes in a pathway on a log scale. Abbreviations: MEP, methylerythritol phosphate; GGPP, geranylgeranyl-diphosphate; THF, tetrahydrofolate; 3-HP/4-HB, 3-hydroxypropionate/4-hydroxybutyrate; BCKA, branched-chain keto-acid.

Figure 5. Diversity of heme copper oxidases present in metagenome sequence data.

Phylogenetic tree (deduced protein sequences) of heme Cu oxidases and their relationship to nitric oxide (NO) reductases (NorB). Metagenome sequences observed across the five sites are included (Site_Meta). All other entries are from annotated genomes found on NCBI. [notations for heme Cu oxidases: AoxB  = A. pernix ; SoxB, SoxM  =  S. acidocaldarius , ; DoxB  =  A. ambivalens ; FoxA  =  M. sedula , and NorB  =  nitric oxide reductases. Tree  =  distance tree created with MEGA using the neighbor-joining method with 100 bootstraps].