Proteins Rewired Iron-Binding Sites as Earth’s Oxygen Rose

Iron has powered life’s central chemistry for billions of years, shuttling electrons and catalyzing reactions in organisms that breathe oxygen and in those that cannot tolerate it. A new open-access study reports that as Earth’s surface became oxygenated, proteins did not abandon this versatile metal. Instead, they rewired the local chemistry around it. Analyzing hundreds of experimentally determined protein structures, researchers found that proteins operating in oxygen-rich metabolic contexts hold their iron atoms in markedly more hydrophilic surroundings and use the amino acid cysteine far less often near the binding site than proteins from oxygen-poor contexts. The pattern, the authors argue, records an evolutionary compromise: life kept iron as its core redox metal while reconfiguring the residues around it to reduce vulnerability to oxidative damage.

The study, published in Discover Life on 17 July 2026 by Benjamin I. Jelen, Beth Christensen, and Eli K. Moore, introduces a computational workflow called Coordination Sphere Analysis and Comparison, or CSAC. Unlike sequence alignments or structural comparisons that assess overall protein folds, CSAC focuses exclusively on the immediate three-dimensional chemical environment of a metal center. The researchers provided the tool with a curated list of Protein Data Bank entries, each tagged with pathway metadata and an oxygen category, and the software parsed the deposited atomic coordinates to identify every iron atom, compute the geometric center of its ligand cluster, and catalog the residues lying within a defined local shell. For lone iron ions, that shell extends 5 angstroms from the metal; for larger ligands such as heme groups, it extends 2 angstroms beyond the ligand’s outer surface. The result is a standardized snapshot of what the protein presents to its iron cofactor, independent of evolutionary homology.

The dataset behind the findings was assembled with unusual care. Rather than sweeping every iron-containing structure from the Protein Data Bank, the team built pathway-based bins spanning major biogeochemical modules. Aerobic bins included aerobic respiration, reactive oxygen species detoxification, methane oxidation, nitrification, aerobic aromatic-ring oxygenation, and sulfur oxidation. Anaerobic bins covered anaerobic terminal reductases, extracellular electron transfer, electron bifurcation, hydrogen metabolism, sulfur and sulfate respiration, anaerobic nitrogen cycling, Wood-Ljungdahl acetogenesis, anaerobic aromatic degradation, and methanogenesis. Oxygen assignment followed pathway-level reaction logic: pathways whose overall physiology requires or couples to molecular oxygen were labeled aerobic, while those that proceed without it were labeled anaerobic, and ambiguous mixed cases were excluded from the primary contrast. After quality-control filtering, which excluded computed-only models, flagged mutant or engineered entries, and applied resolution thresholds of 3.0 angstroms for X-ray structures and 3.5 angstroms for cryo-electron microscopy, the primary input contained 108 structures, 67 aerobic and 41 anaerobic.

CSAC detected iron centers in 94 of those structures, identifying 567 iron-binding centers in total. Most belonged to chemically explicit cofactors: heme c accounted for 32.8 percent of all centers, the Fe4S4 iron-sulfur cluster for 27.2 percent, and heme b for 12.3 percent. Of the 567 centers, 63, or 11.1 percent, were lone ions, a class the authors scrutinized manually because deposited metal labels in the Protein Data Bank can occasionally reflect purification artifacts or annotation errors rather than physiological cofactors. Most lone-ion centers corresponded to expected functional mononuclear or diiron cofactors in non-heme oxygenases, rubredoxin, hydrogenase, and related contexts. After excluding 11 centers with no capturable nearby residues, 556 centers entered the residue-composition and hydropathy analyses.

The hydropathy result is striking. Using the Kyte-Doolittle scale, in which higher values indicate more hydrophobic character, the mean average hydropathy of residues surrounding iron centers was −0.40 for aerobic centers, based on 273 centers, versus 1.31 for anaerobic centers, based on 283. Lower hydropathy means a more hydrophilic, water-friendly local environment. Because multiple iron centers within a single protein, such as the heme arrays of multiheme cytochromes, are not independent evolutionary observations, the team’s primary statistical layer aggregated all centers within each protein to a single value. Even with this conservative approach, the aerobic-anaerobic hydropathy difference remained highly significant, with a Welch t-test p-value of 7.03 × 10⁻⁷ and a Mann-Whitney p-value of 3.86 × 10⁻⁵, corresponding to an effect size of −1.11. A complementary mixed-effects model accounting for repeated centers within entries and pathway structure supported the same direction and significance.

Residue composition told a consistent story. Each iron center was encoded as a 20-element vector of normalized amino-acid frequencies in its neighborhood, and principal component analysis revealed that the first component, explaining 78.3 percent of total variance, was dominated almost entirely by cysteine, with a loading of 0.967. Aerobic centers clustered at the cysteine-poor end of this axis, while anaerobic centers extended toward the cysteine-rich end. In the per-protein statistical layer, cysteine differences were significant with a Welch p-value of 9.48 × 10⁻⁹ and a Mann-Whitney p-value of 2.99 × 10⁻¹¹, and compositional permutation tests confirmed the signal with a p-value of 0.000333. Aerobic neighborhoods also showed higher histidine representation, though the authors caution this should not be read as a simple one-for-one cysteine-to-histidine swap; cysteine depletion is the dominant pattern, with histidine enrichment a secondary feature.

Cofactor composition contributes part of the cysteine difference but not all of it. Heme-like centers are inherently low in cysteine and are more common in the aerobic set, and a simple heme versus non-heme decomposition suggests this accounts for roughly one-third to one-half of the center-level cysteine difference. The remainder reflects genuine shifts in local residue choice. The authors emphasize that the patterns represent changes in relative representation rather than absolute exclusions: anaerobic contexts in the dataset still contain many heme-like centers, including multiheme cytochromes used in extracellular electron transfer and anaerobic respiration, while aerobic contexts retain iron-sulfur clusters in enzymes such as Rieske oxygenases. Oxygenation, in this view, did not eliminate iron-sulfur chemistry or confine heme to aerobic metabolism; it shifted the relative use, structural protection, and local residue environments of iron cofactors across metabolic contexts.

Why would oxygen favor hydrophilic, cysteine-poor iron sites? The authors lay out a mechanistic chain. Reactive oxygen species such as superoxide and hydrogen peroxide, which arise in oxygenated settings, can oxidize cysteine thiols and damage protein active sites. Oxygen and these reactive species can also destabilize iron-sulfur clusters directly by oxidizing cluster sulfide or the cysteine ligands that anchor them, inactivating the enzyme. Furthermore, iron-sulfur cofactor biogenesis depends on dedicated assembly and repair systems, and oxidative stress increases the demand for cluster protection and reassembly. Because iron-sulfur and heme redox potentials are strongly tuned by the surrounding protein environment, oxygenation may also have selected for local residue changes that adjust cofactor shielding, solvent exposure, and electron-transfer thermodynamics. In anaerobic settings, by contrast, extensive cysteine use carries little oxidative risk, allowing the cysteine-rich environments of iron-sulfur clusters to persist.

The authors situate these findings within Earth’s oxygenation history. Geochemical evidence indicates that marine oxygenation before the Great Oxidation Event was spatially and temporally dynamic, with transient oxygenation of shelf bottom waters by roughly 2.65 billion years ago, consistent with episodic oxidative pressures that could have shaped metalloprotein evolution. Iron-sulfur clusters illustrate iron’s continuity across this transition: they were central to early anaerobic metabolism and remain important in many aerobic systems today. The study also notes a modern parallel: in the compartments of E. coli cells today, cysteine usage is reduced in proteins in the outer, relatively oxidized compartments compared with the more reducing cytoplasm, echoing the ancient pattern at a cellular scale.

The work carries caveats. Iron centers within the same protein share evolutionary history and cannot be treated as fully independent observations, which is why protein-level aggregation served as the primary inferential comparison. The Protein Data Bank itself carries sampling biases toward well-studied, experimentally tractable proteins from model organisms, and deposited metal identities may include annotation errors, a concern most relevant to isolated ions. Excluding the single most ambiguous lone-ion center had negligible effect on the protein-level results. The authors suggest that future studies could refine oxygen assignments using oxidation-reduction potential, extend CSAC to secondary coordination spheres and other metals such as copper, and compare local coordination descriptors across protein, mineral, and meteoritic datasets to probe links between biological and geological metal environments. For now, the study offers a molecular record of planetary change: as oxygen rose, life held onto iron and rebuilt the chemistry around it.

Subject of Research: How iron coordination chemistry in protein structures differs between aerobic and anaerobic metabolic contexts

Article Title: Structural evolution of iron coordination in proteins across earth’s oxygenation history

Article References: Jelen, B. I., Christensen, B., & Moore, E. K. (2026). Structural evolution of iron coordination in proteins across earth’s oxygenation history. Discover Life, 56(1), Article 29. https://doi.org/10.1007/s11084-026-09751-7

Image Credits: AI Generated

DOI: 10.1007/s11084-026-09751-7

Keywords: iron-binding proteins, coordination sphere, Earth oxygenation, metalloproteins, Fe-S clusters, heme, cysteine, hydropathy, Protein Data Bank, oxidative stress, CSAC workflow, protein evolution