Oxidative metabolisms catalyzed Earth's oxygenation

The paper addresses why and how Earth's atmosphere transitioned from O2-poor to O2-rich, focusing on the Great Oxidation Event (GOE, ~2400–2300 Ma). While O2 accumulates when production exceeds consumption, traditional models emphasize shifts in global redox equilibrium or nutrient-driven productivity and burial. The authors propose an alternative dynamic: multiple equilibria with positive feedbacks enabling transitions between stable states. They hypothesize that oxidative metabolisms producing partially oxidized organic matter (POOM) enhanced mineral-associated preservation and burial in low-O2 sediments, thereby promoting atmospheric O2 accumulation despite O2-consuming metabolisms. They introduce a mechanistic model for burial efficiency under variable oxygen exposure and test temporal consistency by reconstructing the evolutionary history of a representative POOM-producing enzyme family (flavin-dependent Baeyer–Villiger monooxygenases, BVMOs).

Background work links the GOE to changes in volcanic degassing, mantle/crust oxidation, hydrogen escape, and biogenic methane dynamics, as well as nutrient (especially phosphorus) controls on productivity and organic burial. Prior sedimentary organic matter studies show mineral-association enhances preservation, and oxygen-exposure time reduces burial efficiency. Oxygenases are known to insert oxygen into organic substrates, generating oxygen-containing functional groups (carboxyls, hydroxyls) characteristic of recalcitrant DOM. SAR202 bacteria are prevalent in deep oceans and implicated in oxidizing recalcitrant DOM using BVMOs. Flavin cofactors (FAD, FMN) likely predate the GOE, suggesting early potential for oxidative metabolism. However, BVMO evolutionary timing and its link to Earth’s oxygenation had not been thoroughly explored.

Two-part approach: (1) Mechanistic modeling of burial efficiency with POOM formation; (2) Phylogenetic and molecular clock analyses of BVMO gene family and hosts.

• Burial model: Organic matter comprises labile g1 and recalcitrant g2. In oxic sediments, g1 degrades at k1; g2 degrades at k2 (k2 < k1). Oxygen-exposure time is tox. Initial fractions: g1(0)=(1−α)g0, g2(0)=αg0. The model is extended to include conversion of g1 to POOM (g2) at rate k12 during 0 < t ≤ tox: dg1/dt = −k1 g1 − k12 g1; dg2/dt = k12 g1 − k2 g2; with k12=0 when t>tox. Burial efficiency is g2(tox)/g0. Analytical solution (Methods) yields conditions for a positive feedback: burial efficiency increases with tox when k12 > α k2/(1−α); a maximum occurs at a critical exposure time t* determined by rate constants (see Eqs. 7–9 in Methods). Conceptually, partial oxidation increases reactive oxygen-containing functional groups, enhancing sorption to minerals and steric protection against enzymes.
• Phylogenetics: Identified 298 taxa for a species tree using 30 ribosomal proteins; 330 BVMO protein sequences for the gene tree (31 from SAR202). Sequences aligned with MAFFT; poorly aligned regions trimmed; species tree and gene tree inferred by IQ-TREE with ModelFinder-selected models (species: LG+I+R9; gene: LG+R6). Species tree rooted based on established bacterial rooting; BVMO gene tree rooted with MAD. Molecular dating by PhyloBayes: species tree calibrated with a normal root prior (mean 3.9 Ga, SD 0.25 Ga; bounds informed by 4.4 Ga zircon and 3.4 Ga stromatolites) and multiple broad secondary calibrations within Chloroflexi; gene tree calibrated using five SAR202 node CIs from the species tree and a flat root prior 3.2–1.2 Ga. Chronograms generated from posterior samples (n=1023). Horizontal gene transfers (HGTs) inferred via RANGER-DTL 2.0 using default D/T/L costs, mapping the rooted gene tree to the rooted species tree; events with bootstrap ≥80% retained. Constructed weighted distributions for older and younger bounds of 68 BVMO HGT acquisition timings. Diversification rates for SAR202 BVMOs computed as per-gene rates r in 100 Myr bins using posterior node counts; compared against a null model of random branching at the mean rate, assessing temporal correlation structure via power spectra and Ljung-Box tests.

• Mechanistic result: In O2-limited settings where unoxidized OM dominates inputs (α < 1), if POOM formation rate exceeds modified degradation (k12 > α k2/(1−α)), burial efficiency increases with oxygen-exposure time up to a maximum, creating a positive feedback: more O2 exposure → more POOM → stronger mineral protection → more burial → higher atmospheric O2. At higher O2 (long tox), negative feedback dominates as all OM degrades, stabilizing at a higher O2 steady state.
• HGT timing: The earliest inferred BVMO acquisition (HGT #1) on the branch between stem and crown SAR202 likely occurred 2600–1540 Ma (95% interval), overlapping the GOE and the Lomagundi carbon isotope excursion. Mean older bound (stem) 2350 Ma (95% CI: 2056–2598 Ma); mean younger bound (crown) 1830 Ma (95% CI: 1535–2110 Ma). 85% and 97% of the donor–recipient branch interval probability density intersect the GOE and Lomagundi intervals, respectively.
• Diversification bursts: SAR202 BVMO per-gene diversification rates show three major peaks: ~2500 Ma (Neoarchean/Paleoproterozoic; GOE), ~1200 Ma (Middle/Late Mesoproterozoic; diversification of eukaryotic marine algae), and ~300–200 Ma (Late Paleozoic/Early Mesozoic; Permo-Carboniferous O2 maximum). A sharp decline occurs near ~700 Ma, coincident with Neoproterozoic global glaciations.
• Statistical support: Temporal structure of diversification rates departs from white-noise expectations of a randomly branching null model; power spectrum peak significance p < 0.01; Ljung-Box tests (25 lags) p-values > 0.5 in the null, indicating the empirical series has significant temporal correlation.
• Geobiological linkage: Extensive BVMO HGTs among Chloroflexi, Actinobacteria, and Proteobacteria span the Proterozoic–Phanerozoic, with apparent increase from the Late Neoproterozoic. Mineral evolution (Fe(II) → Fe(III), growth of clay and biomineral abundance) would have increased mineral protection, amplifying POOM preservation and O2 accumulation. In the Phanerozoic, BVMO diversification broadly tracks reconstructed atmospheric O2 levels.

The findings support a POOM-mediated positive feedback in low-O2 sediments: oxidative metabolisms produce oxygenated functional groups that increase organic-mineral associations, reducing enzymatic accessibility and enhancing burial. Early O2 production in the Late Archean could have spurred the emergence and spread (via HGT) of POOM-producing enzymes like BVMOs, with initial SAR202 acquisition temporally coincident with the GOE and Lomagundi event, consistent with enhanced burial. As atmospheric O2 rose, iron oxidation and the proliferation of Fe(III) oxides, clays, and biominerals further strengthened mineral protection, promoting additional O2 accumulation. Diversification peaks align with known redox transitions (GOE; Mesoproterozoic algal expansion; Permo-Carboniferous O2 high), while declines correspond to low-productivity, anoxic intervals (Neoproterozoic glaciations). The analysis suggests that environmental oxygenation both drove and was driven by the diversification of oxidative metabolisms, producing an autocatalytic dynamic coupling biological innovation with evolving surface mineralogy. The mechanism emphasizes burial efficiency increases rather than primary productivity increases, and is consistent with modern sediment preservation processes.

The study proposes and supports the POOM hypothesis: oxidative metabolisms generating partially oxidized organic matter enhance mineral-protected burial under low-O2 conditions, creating a positive feedback that catalyzed Earth's oxygenation. A simple kinetic model identifies the conditions for this feedback; phylogenetic and molecular clock analyses of BVMO enzymes show temporal alignment of acquisition and diversification with major oxygenation intervals (GOE, Lomagundi, Mesoproterozoic algal rise, Permo-Carboniferous O2 peak). Together, these results indicate Earth's oxygenation was an autocatalytic transition arising from synergistic biological and geological changes. Future research should: (1) analyze additional POOM-producing enzyme families to refine temporal resolution; (2) perform laboratory and field studies quantifying POOM preservation under ancient-like low-O2 conditions; (3) seek sedimentary records of mineral-associated POOM abundance across oxygenation and oceanic anoxic events (e.g., via ramped pyrolysis/oxidation); and (4) evaluate historical roles of Actinobacteria and Proteobacteria in POOM production, alongside improvements in molecular clock precision.

• Molecular dating uncertainties are substantial, especially for Proterozoic–Phanerozoic HGT timing; broad priors and secondary calibrations limit precision. Single-gene family analyses constrain temporal resolution and may not capture the full diversity of oxidative pathways.
• The gene tree root is uncertain; reconciliations and HGT inferences depend on tree topology and rooting.
• The mechanistic model is simplified (two-pool kinetics, first-order rates, constant parameters) and does not capture spatial heterogeneity, variable mineralogy, or complex sediment diagenesis.
• The analysis focuses on burial efficiency rather than explicitly modeling nutrient (P) dynamics; assumes constant organic-bound P burial and C/P ratios.
• The model demonstrates conditions for positive feedback but does not fully address the stability of the initial low-O2 state or quantify transition thresholds at the Earth system scale.