Biogeography and ecological functions of underestimated CPR and DPANN in acid mine drainage sediments

Abstract

Recent genomic surveys have uncovered candidate phyla radiation (CPR) bacteria and DPANN archaea as major microbial dark matter lineages in various anoxic habitats. Despite their extraordinary diversity, the biogeographic patterns and ecological implications of these ultra-small and putatively symbiotic microorganisms have remained elusive. Here, we performed metagenomic sequencing on 90 geochemically diverse acid mine drainage sediments sampled across southeast China and recovered 282 CPR and 189 DPANN nonredundant metagenome-assembled genomes, which collectively account for up to 28.6% and 31.2% of the indigenous prokaryotic communities, respectively. We found that, remarkably, geographic distance represents the primary factor driving the large-scale ecological distribution of both CPR and DPANN organisms, followed by pH and Fe. Although both groups might be capable of iron reduction through a flavin-based extracellular electron transfer mechanism, significant differences are found in their metabolic capabilities (with complex carbon degradation and chitin degradation being more prevalent in CPR whereas fermentation and acetate production being enriched in DPANN), indicating potential niche differentiation. Predicted hosts are mainly Acidobacteriota, Bacteroidota, and Proteobacteria for CPR and Thermoplasmatota for DPANN, and extensive, unbalanced metabolic exchanges between these symbionts and putative hosts are displayed. Together, our results provide initial insights into the complex interplays between the two lineages and their physicochemical environments and host populations at a large geographic scale.IMPORTANCECandidate phyla radiation (CPR) bacteria and DPANN archaea constitute a significant fraction of Earth's prokaryotic diversity. Despite their ubiquity and abundance, especially in anoxic habitats, we know little about the community patterns and ecological drivers of these ultra-small, putatively episymbiotic microorganisms across geographic ranges. This study is facilitated by a large collection of CPR and DPANN metagenome-assembled genomes recovered from the metagenomes of 90 sediments sampled from geochemically diverse acid mine drainage (AMD) environments across southeast China. Our comprehensive analyses have allowed first insights into the biogeographic patterns and functional differentiation of these major enigmatic prokaryotic groups in the AMD model system.

Keywords: DPANN archaea; biogeography; host-cell interactions; microbial ecology; ultramicrobacteria.

Fig 1

Overview and abundance of CPR and DPANN in AMD sediments. (A) Size, GC content, completeness, and contamination of CPR (blue) and DPANN (red) genomes. (B) Accumulation curves of CPR and DPANN genomes in the AMD sediments. Dots represent the average numbers of CPR and DPANN genomes for all combinations of a given number of samples, and error bars represent the range. (C) Relative abundances of genomes in the microbial communities, with CPR and DPANN highlighted and legend shared with other subfigures. (D) The abundance ranking curve of the rpS3 genes detected in the genome at each sampling site shows the five sampling sites with the highest relative abundance of CPR and DPANN, respectively.

Fig 2

Distribution of CPR and DPANN in AMD and other environments. Genome relative abundances of class-level lineages within CPR (which was reclassified as a single phylum according to GTDB classification) (A) and DPANN (B) in all mining areas. Unrooted phylogenetic tree of the CPR (C) based on 15 concatenated ribosome proteins and of the DPANN (D) based on 14 concatenated ribosomal proteins (Fig. S3 and S4 show more readable trees). The tree scale is shown in the legend. The pie chart next to each lineage represents the environments from which the genomes in the lineage were recovered. The size of the habitat pie chart for CPR and DPANN, respectively, is proportional to the number of genomes (ln scale).

Fig 3

Biogeographic patterns of CPR and DPANN. VPA of the abundance variation of CPR (A) and DPANN (B) is explained by the comprehensive set containing physicochemical factors (Phy.), geographical distances (Geo.), and climatic variables (Cli.). The number represents the relative effect of each factor or a combination of factors. (C) Pairwise comparisons between various variables based on Pearson’s correlations (top right). The color gradient denotes Pearson’s correlation coefficient, and the star mark denotes statistical significance: *P < 0.05, **P < 0.01, and ***P < 0.001. Influence factors of CPR and DPANN taxonomic composition in partial Mantel tests under controlling geographical distance and climate variables (lower left). The edge width corresponds to Mantel’s r statistic for the corresponding distance correlations, and the edge color denotes the statistical significance. EC, electrical conductivity; Ferrous, ferrous iron; Ferric, ferric iron; TOC, total organic carbon; TN, total nitrogen; TP, total phosphorus; AP, available phosphorus; MAT, mean annual temperature; MAP, mean annual precipitation; Dist., distance from the equator; _ra, relative abundance; _abun, abundance.

Fig 4

Profiles and biomarkers of CPR and DPANN metabolism. (A) Metabolic heatmaps of 282 CPR genomes and 189 DPANN genomes, with columns representing the class-level lineages of CPR and the order-level lineages of DPANN, and rows representing the relative abundance of key genes required for various metabolic and biosynthetic functions in the genomes of each taxonomic unit. (B) Biomarkers of metabolic capacity (gene [g], function [f], and category [c]). LEfSe cladogram showing CPR and DPANN biomarkers.

Fig 5

Co-abundance patterns and host interactions of CPR and DPANN. (A) Microbial community co-occurrence network of CPR (above) and DPANN (below) in the AMD sediment samples. Each node represents a genome (average relative abundance > 0.1%), and each edge represents a strong and significant correlation between two points (Pearson’s |r| > 0.9, P < 0.05). Nodes classified as CPR are marked in red and DPANN in blue. The nodes and edges connected to CPR and DPANN are marked in different colors to represent their phylum-level classification levels, and the colors are the same as in the legend for panel B. (B) Interaction patterns of CPR and DPANN with their putative hosts. Putative hosts are grouped at the phylum level. The number of MAGs in each group is shown in the lower right corner. The black arrow represents compounds exchanged. Specific metabolic exchange information is provided in Table S9. The compounds in the group were derived by SMETANA and verified with Kyoto Encyclopedia of Genes and Genomes (KEGG) Mapper. AAs, amino acids; 4abz, 4-aminobenzoate; acgam1p, N-acetyl-D-glucosamine 1-phosphate; bz, benzoate; CO, carbon monoxide; coa, coenzyme A; f6p, D-fructose 6-phosphate; fald, formaldehyde; gam6p, D-glucosamine 6-phosphate; glyald, D-glyceraldehyde; nh4, ammonium; succ, succinate; urea, urea CH₄N₂O. (C) Comparison of node-level topological characteristics (degree, betweenness centrality, and closeness centrality) of CPR and DPANN and the Wilcoxon rank sum test. ns, 0.05 < P; *, 0.01 < P < 0.05; **, 0.001 < P < 0.01; ***, 0.0001 < P < 0.001; and ****, P < 0.0001.

Fig 6

Analysis of iron reduction-related genes in the CPR and DPANN genomes. (A) Pie charts of the percentages of CPR and DPANN genomes containing the flavin-based EET mechanism genes. Genes not identified in the genomes of both lineages are not shown. (B) Linear regression relationship between the total abundance of CPR and DPANN Ndh2 genes and the concentration of ferrous (top) and ferric (bottom) iron. The statistical test used was two tailed. (C) Maximum-likelihood phylogenetic tree with Ndh2 genes from CPR and DPANN compared to homologs found in eggNOG v5.0.0 database, colored according to the phylum level classification of the origin of the homologs (the color ring). The Clusters of Orthologous Genes (COG) numbers are indicated by a star.