Multiple integrated metabolic strategies allow foraminiferan protists to thrive in anoxic marine sediments

Abstract

Oceanic deoxygenation is increasingly affecting marine ecosystems; many taxa will be severely challenged, yet certain nominally aerobic foraminifera (rhizarian protists) thrive in oxygen-depleted to anoxic, sometimes sulfidic, sediments uninhabitable to most eukaryotes. Gene expression analyses of foraminifera common to severely hypoxic or anoxic sediments identified metabolic strategies used by this abundant taxon. In field-collected and laboratory-incubated samples, foraminifera expressed denitrification genes regardless of oxygen regime with a putative nitric oxide dismutase, a characteristic enzyme of oxygenic denitrification. A pyruvate:ferredoxin oxidoreductase was highly expressed, indicating the capability for anaerobic energy generation during exposure to hypoxia and anoxia. Near-complete expression of a diatom's plastid genome in one foraminiferal species suggests kleptoplasty or sequestration of functional plastids, conferring a metabolic advantage despite the host living far below the euphotic zone. Through a unique integration of functions largely unrecognized among "typical" eukaryotes, benthic foraminifera represent winning microeukaryotes in the face of ongoing oceanic deoxygenation.

Fig. 1. Expression of plastids and plastid-related genes from N. stella and B. argentea metatranscriptomes.

(A) Mapping of the N. stella and B. argentea metatranscripts onto the chloroplast genomes of four diatoms: C. simplex, O. sinensis, T. oceanica, and S. pseudocostatum. Rings in dark purple and dark blue show the maximum detection of each gene in any sample of N. stella and B. argentea, respectively. Rings in pink and lighter blue indicate the mean of each gene’s log₁₀-transformed coverage across all N. stella and B. argentea samples. The outermost ring displays gene category for select genes: RP, ribosomal proteins (gray); PS, photosystem proteins (orange); RuBisCo (green). (B) Expression of transcripts for plastid-encoded genes, rbcL and rbcS, and nuclear-encoded GAPDH, PGK, and FCP in N. stella and B. argentea, across the different oxygen treatments. Circle size is proportional to the abundance of the genes and displayed in counts per million (CPM) (log₂). In each circle, the pie chart colors divide each gene’s total coverage according to the transcripts’ high-level taxonomic categorization.

Fig. 2. Key metabolic pathways and their expression levels in N. stella and B. argentea.

(A) Schematic representation of metabolic pathways in the foraminifera N. stella and B. argentea based on gene expression data from the host-enriched transcriptome. Identified proteins are displayed in grey boxes. N. stella has two proteins associated with ammonium assimilation: GS, glutamine synthetase and NADH-GOGAT, NADH-dependent glutamate synthase. B. argentea has glutamine dehydrogenase (GDH), which can be used for ammonium recycling or assimilation to glutamate. Enzymes in the denitrification pathway are NRT2; NNP, nitrate/nitrite porter; pNR; NirK, nitrite reductase; Nod, nitric oxide dismutase; NorZ, nitric oxide reductase. The identified proteins in H₂-producing anaerobic energy metabolism are PFOR and Fe-hydrogenase, [FeFe]-hydrogenase. (B) Heatmap showing expression of select transcript clusters associated with functions illustrated in (A). Each cell’s color shows the average log₂-tranformed CPM of each cluster (columns) in the different treatments (rows). “Amended” samples were treatments resupplied with nitrate and/or peroxide to mimic in situ conditions (see Materials and Methods); “no amendment” samples lacked replenishment.

Fig. 3. Phylogenetic and functional domain analysis of proteins involved in denitrification in N. stella and B. argentea.

(A) The domain structure of N. stella and B. argentea putative nitrate reductase protein as identified by Pfam domain search, with Cyt-b5, cytochrome b5-like heme/steroid binding domain (green) present in the N terminus followed by MOSC N-terminal beta barrel domain (blue), central MOSC domain, molybdenum cofactor binding domain (yellow), and then FAD and nicotinamide adenine dinucleotide (NAD) oxidoreductase binding domains in C terminus (purple and red, respectively). (B) The domain structure of canonical eukaryotic nitrate reductase protein, with Oxidored_molyb being the oxidoreductase molybdopterin binding domain (orange); Mo-co_dimer, Mo-co oxidoreductase dimerization (light blue); central Cyt-b5 (green); and then FAD and NAD oxidoreductase binding domains in C terminus (purple and red, respectively). (C) Maximum likelihood tree showing the placement of putative foraminiferal Nod-like sequences in the context of other Nor–related functions. Leaves corresponding to the putative foraminiferal sequences are bolded. Branch numbers show bootstrap support above 50. The tree is rooted with bacterial NorB. Scale bar, 0.1% sequence divergence. (D) A multiple sequence alignment showing the quinol binding sites and catalytic sites of NorZ, Nod, and the putative foraminiferal Nod-like proteins. Positions critical for the function of NorZ are highlighted in blue, and positions of conserved mutations from the canonical NorZ function are highlighted in red.

Fig. 4. Phylogenetic and functional domain analysis of the anaerobic metabolic protein pyruvate ferredoxin/flavodoxin oxidoreductase.

(A) The domain structure of N. stella and B. argentea PFOR by Pfam domain search, with PFOR-N, Pyruvate ferredoxin/flavodoxin oxidoreductase thiamine diP-bdg (green); PFOR, pyruvate ferredoxin/flavodoxin oxidoreductase core II (blue); 4Fe-4S, an NADH hydrogenase (purple); TTP, thiamine pyrophosphate enzyme, C-terminal (dark red); Flavodoxin (yellow); and then FAD and NAD oxidoreductase binding domains in C terminus (orange and red, respectively). (B) Phylogenetic tree of eukaryotic and bacterial PFOR protein inferred from maximum likelihood analysis. The tree illustrates how PFOR protein sequences from N. stella and B. argentea are included in the eukaryotic clade and closely related to Blastocystis (protist) and C. teleta (metazoan). The numbers along the branches represent bootstrap values. Nodes with statistical support of >50 are shown. Tree rooted with eukaryotic mitochondrial pyruvate dehydrogenase. Scale bar, 0.8% sequence divergence.