Ecogenomics sheds light on diverse lifestyle strategies in freshwater CPR

Abstract

Background: The increased use of metagenomics and single-cell genomics led to the discovery of organisms from phyla with no cultivated representatives and proposed new microbial lineages such as the candidate phyla radiation (CPR or Patescibacteria). These bacteria have peculiar ribosomal structures, reduced metabolic capacities, small genome, and cell sizes, and a general host-associated lifestyle was proposed for the radiation. So far, most CPR genomes were obtained from groundwaters; however, their diversity, abundance, and role in surface freshwaters is largely unexplored. Here, we attempt to close these knowledge gaps by deep metagenomic sequencing of 119 samples of 17 different freshwater lakes located in Europe and Asia. Moreover, we applied Fluorescence in situ Hybridization followed by Catalyzed Reporter Deposition (CARD-FISH) for a first visualization of distinct CPR lineages in freshwater samples.

Results: A total of 174 dereplicated metagenome-assembled genomes (MAGs) of diverse CPR lineages were recovered from the investigated lakes, with a higher prevalence from hypolimnion samples (162 MAGs). They have reduced genomes (median size 1 Mbp) and were generally found in low abundances (0.02-14.36 coverage/Gb) and with estimated slow replication rates. The analysis of genomic traits and CARD-FISH results showed that the radiation is an eclectic group in terms of metabolic capabilities and potential lifestyles, ranging from what appear to be free-living lineages to host- or particle-associated groups. Although some complexes of the electron transport chain were present in the CPR MAGs, together with ion-pumping rhodopsins and heliorhodopsins, we believe that they most probably adopt a fermentative metabolism. Terminal oxidases might function in O₂ scavenging, while heliorhodopsins could be involved in mitigation against oxidative stress.

Conclusions: A high diversity of CPR MAGs was recovered, and distinct CPR lineages did not seem to be limited to lakes with specific trophic states. Their reduced metabolic capacities resemble the ones described for genomes in groundwater and animal-associated samples, apart from Gracilibacteria that possesses more complete metabolic pathways. Even though this radiation is mostly host-associated, we also observed organisms from different clades (ABY1, Paceibacteria, Saccharimonadia) that appear to be unattached to any other organisms or were associated with 'lake snow' particles (ABY1, Gracilibacteria), suggesting a broad range of potential life-strategies in this phylum. Video Abstract.

Keywords: CARD-FISH; CPR; Freshwater lakes; Genome reduction; Lifestyle; Metabolism; Metagenomics; Patescibacteria.

Fig. 1

General genome characteristics for CPR classes compared to free-living bacteria and known parasites or symbionts. All representative CPR genomes from GTDB r89 were selected for this purpose together to the 282 MAGs assembled in this study. RefSeq 81 database was manually curated, and the genomes were classified in different life-strategies categories

Fig. 2

Maximum likelihood (LG + R10, general matrix and FreeRate model with 10 categories for amino acid substitution; 1000 ultrafast bootstraps) phylogeny for the CPR radiation based on 38 concatenated SCGs (Supplementary Table 2). The lake origin of each freshwater MAG obtained from this study is marked by a dot with a different color at the end of the tips. CPR classes are shown with distinct colors in the inner circle. Following annotations starting from the inner circles represent (1) the isolation source of the genome; (2) the estimated genome size; (3) the trophic state of the lake; (4) the relative abundance of each MAG in metagenomic read recruitment expressed as coverage per Gb of metagenome; (5) the GRiD values for estimation of bacterial replication rates. Probe targets for CARD-FISH visualization are indicated by different colors and the name of probes marked with light blue

Fig. 3

A Boxplot of GRiD values for freshwater CPRs according to their classes. B Growth rate estimations for freshwater CPRs obtained in this study and for our genome collections of free-living organisms and symbionts. Doubling time was predicted using gRodon and the median value for each group is represented by a vertical line

Fig. 4

CARD-FISH imaging of different CPR clades. The panels show an overlap of the probe (green), DAPI (blue) and autofluorescence (red) signals. a, b ABY1 members from the GWF2-40-263 and UBA9934 families stained with 2 distinct probes (probe ABY1a-193 and ABY1b-1343). c–f Paceibacteria proposed genus GWA1-54-10 visualized with 2 probes (adl1-132 and adl2-134). h, i Bacteria affiliated to Gracilibacteria proposed genus 2-02-FULL-48-14 are observed using the Pgri-99 probe. j Organisms of the family level group LOWO2-01-FULL-3 (Gracilibacteria) stained with the Pgri-124 probe. k Members of the Saccharimonadia uncultivated family UBA10212 were observed using the SacA-77 probe. l, m Paceibacteria family level group UBA11359 was detected at the surface of other prokaryotes using the ZE-1429 probe

Fig. 5

Metabolic map reconstruction for the Gracilibacteria class. Abbreviations for transporters: ABC—ATP-binding cassette, APC—amino acid-polyamine-organocation, MIT—metal ion transporter, MFS—major facilitator superfamily, MscL—large conductance mechanosensitive ion channel, PSTE—polysaccharide transporter. Abbreviations for compounds: 1,3BPG—1,3-bisphosphoglycerate, 2PG—2-phosphoglycerate, 3PG—3-phosphoglycerate, 5RP—ribulose 5-phosphate, 6PG—6-phosphogluconate, 6PGL—6-phoshoglucono lactone, DAHP—2-Dehydro-3-deoxy-D-arabino-heptonate 7-phosphate, DHAP—dihydroxyacetone phosphate, DHQ—3-Dehydroquinate, dTDP—deoxythymidine diphosphate, dTDP-DXH—dTDP-6-deoxy-D-xylo-4-hexulose, EPSP—5-enolpyruvylshikimate-3-phosphate, F1,6P2—fructose 1,6-bisphosphate, F6P—fructose 6-phosphate, FAD—flavin adenine dinucleotide, FMN—flavin mononucleotide, G3P—glyceraldehyde 3-phosphate, G6P—glucose 6-phosphate, HTPA—(2S,4S)-4-Hydroxy-2,3,4,5-tetrahydrodipicolinic acid, IMP—inosine monophosphate, LPS—lipopolysaccharides, NAD⁺—nicotinamide adenine dinucleotide, P—phosphate, PEP—phosphoenolpyruvate, PPi—pyrophosphate, PPRP—phosphoribosyl pyrophosphate, TDO—dTDP-4-oxo-L-rhamnose, THF—tetrahydrofolate, TXN—thioredoxin, TXN-S-S—thioredoxin disulfide, XMP—xanthosine monophosphate. Pathways or structural complexes: ETC—Electric transport chain, TCA—Tricarboxylic acid

Fig. 6

Capacity to perform certain types of fermentation in freshwater CPR. Colors mark different classes and the proportions of MAGs in each of them encoding specific enzymes is depicted by symbol size. Abbreviations for enzymes: AcK—acetate kinase, ADH—alcohol dehydrogenase, ALDH—aldehyde dehydrogenase, LDH—lactate dehydrogenase, PFOR—pyruvate:ferredoxin oxidoreductase, Pta—phosphotransacetylase

Fig. 7

Maximum likelihood (LG + F + G4, general matrix with empirical codon frequencies counted from data and discrete Gamma model with 4 rate categories) phylogeny for rhodopsins. CPR rhodopsins are marked with a red dot at the end of the tips. A total of 511 sequences were used to generate the tree (alignment length 792), including a database of 392 proteins from known rhodopsin families