Hot Spring Metagenome Yields P450 Enzyme Frozen Mid-Breath With Oxygen on Board

Deep beneath the steaming waters of the Innot hot springs in North Queensland, Australia, microbial communities have been quietly evolving enzymes capable of some remarkable chemistry. Now, scientists mining the DNA of these communities have pulled out a cytochrome P450 enzyme with a party trick that has stunned structural biologists: the very first crystal structure of the enzyme was solved with molecular oxygen caught bound to its heme iron, a state so fleeting in most P450 enzymes that it has only ever been crystallised a handful of times, and always under carefully controlled conditions. The enzyme, named CYP108D18, was recovered from a metagenome-assembled genome belonging to a Novosphingobium species and is described in a new study published in Microbial Biotechnology.

Cytochrome P450 enzymes are heme-thiolate monooxygenases, widely regarded as among the most versatile biocatalysts in nature. In bacteria, they participate in both primary and secondary metabolism, helping to build and break down terpenes, steroids, fatty acids and aromatic hydrocarbons. Their broad substrate scope and exceptional regio- and stereoselectivity have made them prized targets for industrial and pharmaceutical applications, from drug metabolite synthesis to the production of high-value fragrance compounds. Yet despite decades of study, the identification and characterisation of P450 enzymes from metagenome-assembled genomes remains surprisingly scarce, which makes the new work a notable addition to the field.

The research team, led by investigators at an Australian university with support from the Australian Research Council, identified a 1308-base-pair open reading frame within a metagenomic dataset assembled from microbial biomass collected at the hot springs, where surface temperatures range from 62 to 71 degrees Celsius and the pH sits between 7.2 and 8.1. Sequence analysis revealed that CYP108D18 shares 68 percent identity with CYP108D1 from Novosphingobium aromaticivorans, placing it firmly within the CYP108D subfamily. The team hypothesised that, like its close relative, the enzyme might be involved in degrading aromatic hydrocarbons or oxidising monoterpenes, the fragrant plant-derived molecules that drift into hot spring streams from surrounding vegetation.

Getting the enzyme into a form suitable for study required careful optimisation of heterologous expression in Escherichia coli. The researchers tested a matrix of induction conditions, varying IPTG concentration, delta-aminolevulinic acid supplementation and post-induction temperature, before settling on a protocol that yielded approximately 640 nanomoles of properly folded P450 per litre of culture. Purification by nickel-affinity chromatography produced protein with a Reinheitszahl value of 1.4, a measure of heme content relative to total protein that indicates high purity. Spectroscopic analysis confirmed the characteristic signatures of an oxidised ferric P450, with a Soret peak at 417 nanometres, and the classic 450-nanometre absorbance maximum of the carbon-monoxide-bound ferrous complex confirmed that the enzyme was correctly folded.

Thermal stability testing revealed a moderate but meaningful adaptation to the enzyme’s geothermal origins. The temperature at which half of the enzyme’s carbon-monoxide-binding capacity was lost after a 15-minute incubation, known as the T50, came in at 47 degrees Celsius, some 7 degrees higher than the mesophilic CYP108A1 from Pseudomonas. More strikingly, after a full hour of heat stress the advantage widened to roughly 14 degrees, with CYP108D18 retaining full activity after prolonged incubation at 40 degrees while CYP108A1 did not. Interestingly, the closely related CYP108D1, previously isolated from a subsurface Cretaceous formation 410 metres below ground in North America, proved slightly more thermostable still, hinting at unexpected conservation of stability across this subfamily despite the vast geographic separation of their source habitats.

Binding studies using UV-visible difference spectroscopy showed that phenylcyclohexane and phenanthrene, two aromatic hydrocarbons, induced the largest heme spin-state shifts of any compound tested, at 96 and 89 percent respectively, and bound with dissociation constants in the sub-micromolar range. Monoterpenes such as p-cymene, limonene and alpha-terpineol bound more weakly. Spectroelectrochemical measurements then uncovered something unusual: the heme iron of CYP108D18 sits at a strikingly low redox potential of minus 451 millivolts versus the normal hydrogen electrode in the substrate-free state, and substrate binding shifted it only modestly, to minus 409 millivolts. In most P450 enzymes, substrate binding produces a much larger positive shift that thermodynamically favours electron delivery only when substrate is present. This muted response, also seen in CYP108D1, appears to be a distinguishing feature of the CYP108D subfamily and may explain why the enzyme’s native redox partners proved so difficult to identify.

Indeed, the ArR-Arx electron transfer system from N. aromaticivorans, which efficiently supports many other P450s from that organism, showed negligible activity with CYP108D18. After screening a panel of alternative redox partner combinations, the researchers found that spinach ferredoxin paired with ferredoxin reductase was the only system able to drive catalysis, delivering a NADPH consumption rate of approximately 18 micromoles per minute per micromole of enzyme with phenylcyclohexane as substrate. The match makes thermodynamic sense: spinach ferredoxin has a midpoint potential of around minus 430 millivolts, close enough to the enzyme’s highly reducing substrate-bound heme to permit electron transfer. With this system in place, gas chromatography-mass spectrometry confirmed that CYP108D18 hydroxylates phenylcyclohexane selectively at the C4 position, producing trans-4-phenylcyclohexanol as the sole detectable product. No other substrate tested yielded any product.

The crystallographic work produced the study’s most eye-catching result. The substrate-free structure, solved at 2.06 angstrom resolution, revealed the canonical P450 fold of twelve alpha-helices, six 310-helices and nine beta-sheets, but also showed an elongated electron density above the heme that was too large for a water molecule yet perfectly sized for dioxygen. The bound species was modelled with an oxygen-oxygen bond length of 1.2 angstroms, positioned 1.8 angstroms from the heme iron at an S-Fe-O angle of 173.4 degrees, closely matching the geometry of the ferrous-oxy complex of the archetypal enzyme CYP101A1. Because dioxygen normally binds only after the heme iron has been reduced from ferric to ferrous, and CYP108D18’s redox potential is far too low for spontaneous reduction under crystallisation conditions, the researchers attribute the observation to X-ray-induced photoreduction during data collection. The crystal, in effect, captured the enzyme mid-reaction, with oxygen stabilised at roughly 38 percent occupancy in a state best described as a resonance hybrid between Fe(III)-superoxide and Fe(II)-dioxygen.

The phenylcyclohexane-bound structure, determined at 2.25 angstrom resolution, showed the substrate poised above the heme with the C4 hydroxylation site 4.5 angstroms from the iron. Remarkably, the enzyme did not undergo the open-to-closed conformational transition typical of most P450s upon substrate binding; the active-site cavity remained fully solvent-accessible, with an overall root-mean-square deviation of just 0.15 angstroms between the two structures. Molecular dynamics simulations confirmed that the enzyme stays open throughout, with the bulky substrate itself acting as a lid that shields the heme from bulk water while leaving the pocket exposed. The researchers propose that this persistent heme exposure to solvent may explain the unusually low redox potentials of CYP108D18 and CYP108D1, since greater water access to heme is empirically linked to more negative Fe(III/II) potentials in heme proteins.

Computational work tied the structural observations to the enzyme’s selectivity. Molecular docking showed that the open, hydrophobic active-site cavity could also accommodate larger plant-derived terpenes, including the sesquiterpene valencene, which produced a measurable spin-state shift in binding assays, hinting at a possible physiological role in plant-metabolite oxidation. ONIOM quantum mechanics/molecular mechanics calculations on the phenylcyclohexane complex revealed that abstraction of the C4 hydrogen carries a transition-state barrier 4 kilocalories per mole lower than abstraction at C3, with the difference arising not from the quantum region but from steric clashes between the substrate and residues Ser274 and Val322 in the molecular mechanics layer. This prediction matched the experimentally observed trans-4-phenylcyclohexanol product exactly. CYP108D18 now stands as the third structurally characterised member of the CYP108 family, the first CYP108D enzyme solved with a substrate bound, and the first CYP108 ever recovered from a metagenome-assembled genome of a geothermal environment, opening a window on how nature’s most versatile oxidising machines adapt to life at the edge of the thermal habitable world.

Subject of Research: Structural and functional characterisation of the metagenome-derived cytochrome P450 enzyme CYP108D18 from a geothermal microbial community

Article Title: Structural and Functional Characterisation of a Metagenome‐Derived CYP108D18 Reveals an Unusual Dioxygen‐Bound State and Selective Hydrocarbon Hydroxylation

Article References: Kundral, S., Giang, P. D., Kirk, A. M., Buczynski, J. B., Arachchige, K. S. A., Khare, S. K., Bernhardt, P. V., Evans, P. N., Guddat, L. W., Bell, S. G., & De Voss, J. J. (2026). Structural and Functional Characterisation of a Metagenome‐Derived CYP108D18 Reveals an Unusual Dioxygen‐Bound State and Selective Hydrocarbon Hydroxylation. Microbial Biotechnology, 19(10), Article e70451. https://doi.org/10.1111/1751-7915.70451

Image Credits: AI Generated

DOI: 10.1111/1751-7915.70451

Keywords: cytochrome P450, metagenomics, hot springs, CYP108D18, crystallography, dioxygen-bound state, hydrocarbon hydroxylation, Novosphingobium, redox potential, biocatalysis, terpenes, molecular dynamics