Metagenomics-guided analysis of microbial chemolithoautotrophic phosphite oxidation yields evidence of a seventh natural CO2 fixation pathway

Abstract

Dissimilatory phosphite oxidation (DPO), a microbial metabolism by which phosphite (HPO₃^2-) is oxidized to phosphate (PO₄^3-), is the most energetically favorable chemotrophic electron-donating process known. Only one DPO organism has been described to date, and little is known about the environmental relevance of this metabolism. In this study, we used 16S rRNA gene community analysis and genome-resolved metagenomics to characterize anaerobic wastewater treatment sludge enrichments performing DPO coupled to CO₂ reduction. We identified an uncultivated DPO bacterium, Candidatus Phosphitivorax (Ca. P.) anaerolimi strain Phox-21, that belongs to candidate order GW-28 within the Deltaproteobacteria, which has no known cultured isolates. Genes for phosphite oxidation and for CO₂ reduction to formate were found in the genome of Ca. P. anaerolimi, but it appears to lack any of the known natural carbon fixation pathways. These observations led us to propose a metabolic model for autotrophic growth by Ca. P. anaerolimi whereby DPO drives CO₂ reduction to formate, which is then assimilated into biomass via the reductive glycine pathway.

Keywords: carbon fixation; formatotrophic; metagenomics; phosphite oxidation; reductive glycine pathway.

Fig. 1.

(A) Phosphite oxidation by cultures derived from enrichments amended with rumen fluid. Phosphite (closed symbols with solid lines) and phosphate (open symbols with dashed lines) concentrations in cultures containing 9 mM phosphite and 5% rumen fluid (blue circles); 5% rumen fluid only (red squares); 9 mM phosphite only (green triangles); 9 mM phosphite, 5% rumen fluid, and 5 mM molybdate (orange inverted triangles). Data points represent the average of triplicate cultures, with error bars denoting one SD. DNA samples for community analysis were obtained for all data points. (B–E) Taxonomic composition of microbial communities in cultures containing 9 mM phosphite and 5% rumen fluid (B); 5% rumen fluid only (C); 9 mM phosphite only (D); 9 mM phosphite, 5% rumen fluid, and 5 mM molybdate (E). Each OTU with an average normalized abundance of ≥1% of the community under any treatment is labeled according to the lowest taxonomic rank assigned to it: c, class; d, domain; f, family; g, genus; p, phylum. Each dataset represents the average of independently sequenced triplicate cultures.

Fig. 2.

Growth of Ca. Phosphitivorax anaerolimi Phox-21 in enrichment cultures containing either 8 mM phosphite and 5% rumen fluid (blue circles with solid lines) or 5% rumen fluid only (red squares with solid lines). Blue circles with dashed lines indicate phosphite concentrations in phosphite-containing cultures. Phox-21 16S rDNA copy numbers were determined by qPCR using taxon-specific primers and normalized by the total volume of culture sampled for DNA extraction. Data points for 16S rDNA copy numbers represent the geometric mean of triplicate cultures, with error bars denoting one geometric SD. Data points for phosphite concentrations represent the average of triplicate cultures, with error bars denoting one SD.

Fig. 3.

Phylogenetic tree showing the placement of Ca. Phosphitivorax anaerolimi Phox-21 within the Deltaproteobacteria. Selected 16S rRNA sequences were aligned using the Silva aligner, and a maximum likelihood phylogenetic tree was constructed with 1,000 bootstrap resamplings using RAxML-HPC. Members of the Acidobacteria were included as an outgroup. Taxa represented in the Silva reference database were assigned to known orders within the Deltaproteobacteria based on their Greengenes taxonomic assignments. GenBank accession numbers are provided in parentheses. Taxa for which there is experimental evidence of DPO activity are indicated in bold. Internal nodes with bootstrap support of >70% are indicated by closed circles and those with support of >50% by open diamonds. (Scale bar: 0.1 change per nucleotide.)

Fig. 4.

(A) Ptx-ptd gene clusters of Ca. Phosphitivorax anaerolimi Phox-21 and Desulfotignum phosphitoxidans FiPS-3. IMG gene annotations: ptdC, major facilitator superfamily transporter; ptdF, nucleoside-diphosphate-sugar epimerase; ptdG, nucleotide-binding universal stress protein; ptdH, radical SAM superfamily enzyme; ptdI, hypothetical protein; ptxD, phosphite dehydrogenase; ptxE, transcriptional regulator. Genes highlighted in blue are present in APO organisms, genes highlighted in green are present only in FiPS-3 and Phox-21, and genes highlighted in gray are present only in FiPS-3. (B) Phylogenetic tree of the phosphite dehydrogenase PtxD. Protein sequences from selected organisms were aligned using Clustal Omega, and a maximum likelihood tree was constructed with 1,000 bootstrap resamplings using RAxML-HPC. Black branches indicate PtxD sequences while gray branches indicate related outgroup sequences belonging to the d-hydroxyacid dehydrogenase protein family (GDH, glycerate dehydrogenase; LDH, lactate dehydrogenase; PGDH, 3-phosphoglycerate dehydrogenase). Organisms for which there is experimental evidence of DPO are highlighted in green while those for which there is experimental evidence of APO are highlighted in blue. Colored circles indicate the taxonomic affiliations of the organisms harboring each PtxD sequence. Internal nodes with bootstrap support of >70% are indicated by closed circles and those with support of >50% by open diamonds. (Scale bar: 0.1 change per amino acid residue.)

Fig. 5.

Functional overview of high-quality binned genomes. Filled boxes indicate the presence of the metabolic process or bioenergetic complex in the corresponding genome while empty boxes indicate its absence. Only metabolisms and complexes that are present in at least one high-quality binned genome are depicted here. See Table S3 for a list of the key functional genes used to indicate the presence of a metabolism or complex. A metabolism was considered present only if all of the genes encoding at least one full enzymatic pathway capable of carrying out that metabolic process were found in the genome. A bioenergetic complex was considered present only if all of the genes encoding the necessary subunits for a functional protein complex were found in the genome.

Fig. 6.

Genomics-based metabolic model of dissimilatory phosphite oxidation coupled to CO₂ reduction in Ca. Phosphitivorax anaerolimi Phox-21. Dotted lines denote putative mechanisms based on physiological and genomic observations that have yet to be confirmed by direct biochemical evidence. Dissimilatory phosphite oxidation proteins: PtdC, phosphite transporter; PtdF, nucleoside-diphosphate-sugar epimerase; PtdH, radical SAM superfamily enzyme; PtdI, hypothetical protein; PtxD, phosphite dehydrogenase. Ion motive force and reducing equivalents proteins: ATPase, ATP synthase complex; NfnAB, NAD-dependent ferredoxin:NADP oxidoreductase; Rnf, sodium-translocating ferredoxin:NAD oxidoreductase complex. Inorganic carbon assimilation (reductive glycine pathway) proteins: FdhAB, NADP-dependent formate dehydrogenase; Fhs, formate:THF ligase; FolD, methenyl-THF cyclohydrolase/methylene-THF dehydrogenase; GCV, glycine cleavage system (GcvH, lipoate-binding protein; GcvP, glycine dehydrogenase; GcvT, aminomethyltransferase; Lpd, dihydrolipoyl dehydrogenase); GlyA, serine hydroxymethyltransferase; SdaA, serine deaminase.