Novel, Deep-Branching Heterotrophic Bacterial Populations Recovered from Thermal Spring Metagenomes

Abstract

Thermal spring ecosystems are a valuable resource for the discovery of novel hyperthermophilic Bacteria and Archaea, and harbor deeply-branching lineages that provide insight regarding the nature of early microbial life. We characterized bacterial populations in two circumneutral (pH ~8) Yellowstone National Park thermal (T ~80°C) spring filamentous "streamer" communities using random metagenomic DNA sequence to investigate the metabolic potential of these novel populations. Four de novo assemblies representing three abundant, deeply-branching bacterial phylotypes were recovered. Analysis of conserved phylogenetic marker genes indicated that two of the phylotypes represent separate groups of an uncharacterized phylum (for which we propose the candidate phylum name "Pyropristinus"). The third new phylotype falls within the proposed Calescamantes phylum. Metabolic reconstructions of the "Pyropristinus" and Calescamantes populations showed that these organisms appear to be chemoorganoheterotrophs and have the genomic potential for aerobic respiration and oxidative phosphorylation via archaeal-like V-type, and bacterial F-type ATPases, respectively. A survey of similar phylotypes (>97% nt identity) within 16S rRNA gene datasets suggest that the newly described organisms are restricted to terrestrial thermal springs ranging from 70 to 90°C and pH values of ~7-9. The characterization of these lineages is important for understanding the diversity of deeply-branching bacterial phyla, and their functional role in high-temperature circumneutral "streamer" communities.

Keywords: Aquificales; Calescamantes; Pyropristinus; Thermotogae; Yellowstone National Park; hot springs; hyperthermophiles.

Figure 1

Nucleotide word frequency PCA plots of metagenome assemblies from two Aquificales “streamer” communities in YNP. (A) Data colored by site: Octopus Spring = red; Bechler spring = white. (B) Identical PCA orientation with phylogenetic analysis and assignment (dashed-white circles): “Pyropristinus” Type 1-r01 = red; “Pyropristinus” Type 1-r02 = light-red; “Pyropristinus” Type 2-r01 = pink; Calescamantes-like = orange; Firm_T1-r01 = yellow; Thermocrinis-r01 = dark-blue; Thermocrinis-r02 = light-blue; Pyrobaculum spp. = green; Aigarchaeota_T1-r01 = purple.

Figure 2

Frequency plots of the G+C content (%) of random shotgun sequence reads (Sanger) from filamentous “streamer” communities at Octopus Spring (OCT_11) and Bechler springs (BCH_13). Taxonomic (phylogenetic) assignment of each sequence read was performed using BLASTn (>90% nt ID) against curated de novo assemblies generated from these sites (i.e., Figure 1): [light-gray = total reads, red = “Pyropristinus” T1-r1 (G+C = 44%), light-red = “Pyropristinus” T1-r02 (G+C = 44%), pink = “Pyropristinus” T2-r01 (G+C = 29%), orange = Calescamantes-like (G+C = 35%), blue = Thermocrinis-like r01 (G+C = 45.5%), light-blue = Thermocrinis-like r02 (G+C = 45%), yellow = Firmicutes (G+C = 53%), green = Pyrobaculum-like (G+C = 57–58%), purple = Aigarchaeota (G+C = 60%)].

Figure 3

Distribution plots of amino acid identity % (AAI) of protein-coding genes between pairwise comparisons of novel bacterial assemblies. (A) Amino acid identity of “Pyropristinus” Type 1 from Octopus Spring (T1.1) vs. “Pyropristinus” Type 1 from Bechler Spring (T1.2) (black; mean = 94.2 ± 10.1%), and “Pyropristinus” Type 2 (T2.1 vs. T1.2) (white; mean = 46.6 ± 12.3%). (B) Amino acid identity of the Octopus Spring (OS) Calescamantes population vs. Ca. Calescibacterium nevadense (gray; mean = 78.0 ± 18.1%), “Pyropristinus” Type 1 (T1.2) (black; mean = 42.7 ± 10.0%), and “Pyropristinus” Type 2 (T2.1) (white; mean = 41.8 ± 9.0%).

Figure 4

Phylogenomic analysis of “Pyropristinus” and Calescamantes lineages. Maximum-likelihood tree based on genomic analysis of 13 bacterial-specific and 5 universal housekeeping genes (total of 18 genes coding for 8928 amino acid positions). Twenty-seven archaeal references were used as an outgroup. Phyla with more than one reference were collapsed and the number of genomes per group are given in parentheses. Bootstrap values (1000 replicates) are given at the nodes where ≥50%. Scale shows expected substitutions per site.

Figure 5

Phylogenetic analysis using near full-length 16S rRNA genes. 16S rRNA genes from the “Pyropristinus” Types T1 and T2 assemblies are indicated in bold (T1.1 and T2.1). The Calescamantes-OS assembly did not contain a full-length 16S rRNA gene and were thus omitted from this analysis). OSClone_YNP11_11_4, produced from a 16S rRNA gene library of the same Octopus Spring sample (also in bold) is nearly identical to the Calescamantes population from OS. Groups with multiple entries are collapsed as triangles. Bootstrap values (1000 replicates) are given at the nodes where ≥50%.

Figure 6

Non-metric multidimensional scaling ordination plot of COG distribution among “Pyropristinus,” Calescamantes and closely related lineages. NMDS plots (stress 0.08) were constructed from presence/absence Euclidean-distance matrices of COG groups present in the “Pyropristinus” T1 and T2 assemblies and Calescamantes-OS in addition to Ca. Calescibacterium nevadense (dark red), a subset of Aquificae (bright red), Thermodesulfobacteria (blue) and Thermotogae (purple) genomes, which comprised the closest related lineages to the “Pyropristinus” and Calescamantes lineages.

Figure 7

Metabolic reconstruction based on annotation and manual curation of “Pyropristinus” Type 1 de novo assemblies. ABC-type transporters are represented as blue multi-component transmembrane proteins, other transporters as red transmembrane proteins, and the single antiporter as a blue rectangle. Complexes and enzymes used in aerobic respiration/ATP synthesis are identified in green. Gene abbreviations: bglB, phospho-β-glucosidase; amyA, α-amylase; LPS, Lipopolysaccharides; nadA, quinolinate synthase, nadB, L-aspartate oxidase; nadC, quinolinate phosphoribosyltransferase; nadD, nicotinate-mononucleotide adenylyltransferase; nadE, NAD synthetase; pntA, pyridine nucleotide transhydrogenase (α subunit); pgi, phosphoglucose isomerase; pfk, 6-phosphofructokinase I; fba/p, fructose bisphosphate aldolase/phosphatase; tpi, triose phosphate isomerase; gapA, glyceraldehyde 3-phosphate dehydrogenase-A complex; pgk, phosphoglycerate kinase; gpmA, 2,3-bisphosphoglycerate-dependent phosphoglycerate mutase; eno, enolase; pyk, pyruvate kinase; pdhA, pyruvate dehydrogenase (lipoamide); pdhC, pyruvate dehydrogenase E2 component; gltA, citrate synthase; acn, aconitate hydratase; icd, isocitrate dehydrogenase; kor, 2-oxoglutarate:ferredoxin oxidoreductase; suc, succinyl-CoA synthetase, sdh, succinate dehydrogenase; frdB, fumarate reductase iron-sulfur protein; fumC, fumarase C; mdh, malate dehydrogenase; amtB, ammonium transporter; gdh, glutamate dehydrogenase; acs, acetyl-CoA synthetase; fadD, fatty acyl-CoA synthetase; fadJ, 3-hydroxyacyl-CoA dehydrogenase; ACADM, acyl-CoA dehydrogenase (C-4 to C-12); ACADS, acyl-CoA dehydrogenase (C-2 to C-3); atoB, acetyl-CoA acetyltransferase; QP, quinone pool; nuo, NADH:ubiquinone oxidoreductose complex; HCO, Heme-Cu oxidase; pgi/pmi, glucose/mannose-6-phosphate isomerase; tkt, transketolase; rpe, ribulose-5-phosphate 3-epimerase; rpiB, ribose-5-phosphate isomerase B; prsA, ribose-phosphate pyrophosphokinase; FMN, flavin mononucleotide; FAD, flavin adenine dinucleotide. Question marks indicate genes not identified in either Type 1 assembly (list of identified protein-coding genes in Supplementary Table 4).

Figure 8

Temperature and pH distribution of the “Pyropristinus” and Calescamantes phylotypes detected across different geothermal habitats. Sequences (16S rRNA gene) sharing >97% nucleotide identity to the “Pyropristinus” and Calascamantes-OS population were identified from prior 16S rRNA gene surveys of YNP and publically available databases [red = “Pyropristinus” Type 1 only (n = 24); blue = “Pyropristinus” Type 2 only (n = 3); purple = Calescamantes-OS only (n = 3); cyan = 2-3 phylotypes present (n = 8); open circles = none of the three phyla detected]. Sites not containing these lineages are only shown for datasets that extensively surveyed YNP hot springs with universal bacterial PCR primers.