Global metagenomic survey reveals a new bacterial candidate phylum in geothermal springs

Abstract

Analysis of the increasing wealth of metagenomic data collected from diverse environments can lead to the discovery of novel branches on the tree of life. Here we analyse 5.2 Tb of metagenomic data collected globally to discover a novel bacterial phylum ('Candidatus Kryptonia') found exclusively in high-temperature pH-neutral geothermal springs. This lineage had remained hidden as a taxonomic 'blind spot' because of mismatches in the primers commonly used for ribosomal gene surveys. Genome reconstruction from metagenomic data combined with single-cell genomics results in several high-quality genomes representing four genera from the new phylum. Metabolic reconstruction indicates a heterotrophic lifestyle with conspicuous nutritional deficiencies, suggesting the need for metabolic complementarity with other microbes. Co-occurrence patterns identifies a number of putative partners, including an uncultured Armatimonadetes lineage. The discovery of Kryptonia within previously studied geothermal springs underscores the importance of globally sampled metagenomic data in detection of microbial novelty, and highlights the extraordinary diversity of microbial life still awaiting discovery.

Figure 1. New lineage identified using metagenomic and single-cell genomic approaches.

Workflow used to (a) identify novel SSU rRNA gene sequences globally, along with (b) single-cell genomics pipeline to screen and sequence single cells isolated from geothermal springs samples. For the three geothermal spring environments, we sequenced 13, 2 and 3 SAGs, respectively. SSU rRNA gene, small-subunit ribosomal gene; MDA, multiple displacement amplification; QC, quality control; SAG, single-amplified genome. The photograph of Jinze Pool, Tengchong, China, was taken from ref. .

Figure 2. Maximum likelihood concatenated protein phylogeny and cell imaging for ‘Ca. Kryptonia.'

(a) Phylogeny was based on concatenation of 56 conserved marker proteins, where at least 10 marker proteins were used to infer SAG phylogenetic placement (with the exception of JGI-22 with only six marker proteins recovered). Bootstrap support values ⩾50% are shown with small circles on nodes with robust phylogenetic support. The FCB superphylum is shown in the grey shaded region. Expanded phylogenetic tree for ‘Ca. Kryptonia' shows the placement of the proposed four genera represented by GFMs and SAGs, along with the estimated genome completeness shown in parentheses. (b) A ‘Ca. Kryptonia'-specific FISH (fluorescence in situ hybridization) probe was designed and used to visualize cells from Dewar Creek Spring sediment samples. ‘Ca. Kryptonia' cells hybridizing with the probe are green, while other cells are visualized with 4',6-diamidino-2-phenylindole (DAPI; blue). Scale bar, 5 μm.

Figure 3. Limited, yet widely dispersed biogeographic distribution of ‘Ca. Kryptonia' genomes and CRISPR spacers.

All genomic content from the ‘Ca. Kryptonia' GFMs and SAGs was used to comprehensively search the collection of 640 Gb of assembled metagenomic data from 4,290 environmental samples, including 169 samples from geothermal springs and hydrothermal vents denoted by red triangles (temperature ⩾50 °C). Marked circles are as follows: (A) Great Boiling Spring, Nevada; (B) Dewar Creek Spring, Canada; (C) Jinze pool, Yunnan Province, China; and (D) Gongxiaoshe pool, Yunnan Province, China. Significant matches were determined for sequences ⩾250 bp in length and with ⩾75% identity threshold for non-ribosomal genomic regions. For metagenomic contigs mapping to the ‘Ca. Kryptonia' ribosomal operon, a 97% identity threshold was used to capture only high-quality matches to ‘Ca. Kryptonia.' For CRISPR spacers, only significant matches allowing for up to 3 bp mismatch along the entire length of the spacer were considered. The ‘Ca. Kryptonia' genomic hits can be found in Supplementary Data 4 and the manually curated spacer hits can be found in Supplementary Data 3.

Figure 4. Reconstructed metabolic capacity of ‘Ca. Kryptonia.'

Key metabolic predictions and novel features identified in ‘Ca. Kryptonia' GFM and SAGs, with full gene information available in Supplementary Data 6.

Figure 5. Co-occurrence patterns and metabolic complementarity with ‘Ca. Kryptonia.'

(a) Spearman-rank correlation values were calculated based on reconstructed SSU rRNA sequences across 22 geothermal spring metagenomes, and led to the identification of a cluster of highly correlated phylotypes with ‘Ca. Kryptonia.' Armatimonadetes (cluster 3107) had the highest correlation value (ρ=0.82) with ‘Ca. Kryptonia.' (b) Biosynthetic pathways present in the Armatimonadetes genome which complement missing components in ‘Ca. Kryptonia.' Full gene information for the Armatimonadetes genome is available in Supplementary Data 7. Each arrow represents an enzymatic component of the biosynthetic pathways; arrows highlighted in blue are contributed by the Armatimonadetes, while arrows highlighted in dark orange are contributed by ‘Ca. Kryptonia.' Black arrows indicate enzyme was not recovered in either.