Bioinformatic Discovery of a Cambialistic Monooxygenase

Dinuclear monooxygenases mediate challenging C–H bond oxidation reactions throughout nature. Many of these enzymes are presumed to exclusively utilize diiron cofactors. Herein we report the bioinformatic discovery of an orphan dinuclear monooxygenase that preferentially utilizes a heterobimetallic manganese–iron (Mn/Fe) cofactor to mediate an O2-dependent C–H bond hydroxylation reaction. Unlike the structurally similar Mn/Fe-dependent monooxygenase AibH2, the diiron form of this enzyme (SfbO) exhibits a nascent enzymatic activity. This behavior raises the possibility that many other dinuclear monooxygenases may be endowed with the capacity to harness cofactors with a variable metal content.

Purification protocol for ARO proteins.All protein purification steps were performed at 4 °C.Thawed cell pellets were resuspended with 10 mL lysis buffer (200 mM NaCl, 50 mM HEPES pH 8.0) per 1 L culture of cells.Cells were lysed by sonication (15 min total pulse time, 3 seconds on, 3 seconds off).The lysate was centrifuged at 12,000 rpm for 40 min at 4 °C to remove cell debris and the resulting supernatant was loaded onto a column equipped with Ni-NTA resin (Thermo Fisher Scientific).The resin was washed with 200 mL lysis buffer, and 700 mL wash buffer (lysis buffer + 25 mM imidazole), and eluted with 200 mL elution buffer (lysis buffer + 250 mM imidazole).The protein was concentrated to ~40 mL using an Amicon® stirred cell (Millipore Sigma) equipped with a 30 kDa filter and subsequently buffer exchanged into 20 mM HEPES pH 7.5 via dialysis.Protein purity was evaluated via sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE (Fig S7).Protein concentrations were determined from the absorbance at 280 nm calculated based on their molar absorption constant.The calculated molar concentrations are relative to the protein monomers.
For SfbO preparations, the N-terminal His6-tag was removed with TEV protease by stirring the protein in lysis buffer at a final concentration of 1 mg/mL with 2.5 mM dithiothreitol and 1 mg of TEV protease for every 10 mg of SfbO overnight at 4 °C.The resulting solution was filtered and passed down a Ni-NTA affinity column, and the flow-through collected and buffer exchanged to 20 mM HEPES pH 7.5 via dialysis as described above.The resultant protein was used for crystallography and reaction assays.The His6-tag of the other 3 AROs was not removed.
Dithionite titration and oxygen reactivity of Fe-containing ARO proteins.Aerobically-purified protein solutions were rendered anaerobic following repeated evacuation and purging with N2 and subsequently allowed to equilibrate for at least 1 h in a Coy chamber.These samples were transferred to an air-tight cuvette amenable for UV/visible absorption studies.UV/visible spectra of these solutions revealed characteristic Fe III -OHx charge transfer bands (Fig 2B).Subsequent spectra were recorded during the addition of substoichiometric aliquots of sodium dithionite and incubated for 5 min before each successive spectrum was recorded.Absorbance values were normalized to account for dilution.Following the titration endpoint, the cuvette was opened to air and gently mixed via hand shaking to aerate the solution.Subsequent UV-visible spectra were taken over 20 minutes to observe the regeneration of the Fe III -OHx charge transfer bands.

Crystallization of AROs.
Single crystals of all proteins were obtained by sitting-drop vapor diffusion at room temperature.Each reservoir was filled with the condition solution (200 μL), and 2 μL of this solution was added to the well containing 2 μL of the protein sample.All proteins (except the Mn2-metallated form of SfbO) were crystallized in the presence of a two-fold excess of Fe II to ensure a diiron active site constitution (initial protein samples were obtained via expression in M9 minimal medium with 0.125 mM Fe II ).All crystals were harvested and frozen in liquid N2.The specific crystallization conditions employed are listed below SfbO (8 mg/mL initial protein concentration) was crystallized with a condition solution consisting of 21% PEG 5000 MME, 0.1 M MES pH 6.0, 0.2 M Ammonium sulfate, and 10% v/v glycerol.Fe2-SfbO was crystallized aerobically.Mn II 2-SfbO was crystallized in the presence of a 2-fold molar excess of Mn 2+ (MnCl2•4H2O) anaerobically using the same crystallization condition.B6JH23 (11 mg/mL initial protein concentration) was crystallized with a condition solution consisting of 4.2 M 1,6-Hexanediol, 0.1 M BISTRIS pH 6.5, 0.015 M magnesium sulfate, 2% v/v glycerol.Protein expressed in M9 medium supplemented with additional Fe II was used for crystallization.W7L2Y2 (9 mg/mL initial protein concentration) was crystallized with a condition solution of 0.07 M Ammonium acetate pH 5.0, 5.6% PEG 4000, and 20% v/v glycerol.Protein expressed in M9 medium supplemented with additional Fe II was used for crystallization.D3EI84 (17 mg/mL initial protein concentration) was crystallized with a condition solution of 3.6 M sodium formate, and 10% v/v glycerol.Protein expressed in M9 medium supplemented with additional Fe II was used for crystallization.
Data processing and structure solution.XRD data were collected at the Advanced Light Source (ALS) and the Stanford Synchrotron Radiation Light source (SSRL).Diffraction data were processed using XDS (3) or iMosfilm(4) followed by scaling and merging using Aimless.(5)Phasing was performed using Phaser-MR via molecular replacement with the homology model generated by SWISS-MODEL(6) based on AibH2 subunit of PDB ID: 6M2I as the initial search model.The models were refined with phenix.refineand manual modeling was performed with COOT.The coordinates, occupancies and b-factors of the metal ions were freely refined.ChimeraX (7) and PyMol were used for visualization.The pdb codes for the described structures are listed below for convenience.SAD experiments were performed on Mn2-and Fe2-SfbO at the respective wavelengths (Table S4) to determine the anomalous dispersion response.SfbO: diiron (8SM6), dimanganese (8SMA) B6JH23: 8SM8 W7L2Y2: 8SM9 D3EI84: 8SM7 Metalation of SfbO.Homogenously metalated forms of aerobic Fe2-SfbO preparations (Table S2) were obtained via rapid dialysis of freshly purified samples against 20 mM HEPES pH 7.5 with 10 mM EDTA for 1 h.While the resulting iron content was not significantly diminished by this procedure, the contaminating metals were found to be removed from the as-isolated Fe-SfbO, as judged by ICP-OES.
Access to alternative metalation states of SfbO required the preparation of apo-SfbO.Such preparations employed protein expressed in M9 medium supplemented with additional Mn II .The metals found in the resultant, crude Mn-SfbO preparations are entirely removed upon treatment with 5 mM sodium dithionite and 10 mM EDTA for 1 h and subsequent buffer exchange by concentrating the protein sample to ~5 mL using an Amicon® stirred cell and filling with 20 mM HEPES pH 7.5 up to 100 mL five times.This apo-SfbO was reconstituted with 2 equivalents of Mn II to furnish Mn-SfbO (Table S2).Overnight incubation of Mn-SfbO with one equivalent of Fe II and subsequent buffer exchange furnished samples of Mn/Fe-SfbO that contain equimolar amounts of each ion (Table S2).
SfbO reaction assay protocol.Enzymatic assays were performed at pH 7.25, and all substrate stocks (1 M or 500 mM) were adjusted to pH = 7.25 with hydrochloric acid or sodium hydroxide.MES, PIPES, HEPES, and TES have good buffering capacity at pH 7.25, so no additional buffer was needed for the reaction assays.Substrate stocks of taurine, isethionate, and CHES were prepared with 5 mM MOPS added for pH adjustment, and those reactions were carried out in 10 mM MOPS pH 7.25.MOPS is not detectably converted by SfbO to generate either sulfoacetaldehye (SA) or sulfopropionaldehyde. Manganese(II) chloride tetrahydrate and ammonium iron(II) sulfate hexahydrate were used to prepare stocks of the supplementary Mn II and Fe II metal ions, respectively.
Standard enzymatic assays (50 μL total volume) contained the listed SfbO preparation (0.5 μM), sodium ascorbate (2 mM), supplementary Fe II or Mn II metal ions (0.25 μM for Fig 3, 1 μM for Fig S5A and  B), sulfonate substrate (200 mM), and 0.05 mg/ml catalase where stated.The protein-free controls replaced SfbO with 1 μM each of free Fe II and Mn II .The reagents were combined aerobically and incubated with shaking at 300 rpm and 20 o C for 3 h.Assays were quenched by the addition of a 5 mM DNPH solution in 2 M hydrochloric acid (12 μl), following existing protocols (11).To ensure complete derivatization of the sulfoacetaldehyde product, these mixtures were shaken at 300 rpm at 30 o C for 30 min.The mixture was then neutralized by addition of 2 M NaOH (12 μl), and centrifugated at 16 000 g for 10 min.The yellow solution was then analyzed by analytical HPLC or LCMS.To ascertain the Michaelis-Menton parameters for PIPES hydroxylation, 1 μM of SfbO was employed and reactions were quenched after 25 minutes while under steady-state conditions (Fig S5G ).
Synthesis of Sulfoacetaldehyde (SA) Standard.Sodium isethionate (0.125 g, 0.844 mmol) was dissolved in DMSO (1 mL) and combined with 1.5 equivalents of Dess-Martin periodinane (0.537 g, 1.27 mmol) and stirred at 37 o C for 2 h.The resulting mixture was heated to 50 o C and evaporated to dryness in vacuo on a Schlenk line.The resultant oily residue was washed with DCM, and the remaining solid was redissolved in water and centrifuged.The clear, colorless supernatant was decanted and dried in vacuo overnight at room temperature.The resultant crude sulfoacetaldehyde product was redissolved in water as a stock solution and stored at -80 o C. NMR spectroscopy indicates that the sulfoacetaldehyde is composed of a mixture of the aldehyde and the hydrated geminal-diol.Aldehyde: 1 H NMR (400 MHz, D2O): δ 9.65 (t, J = 2.9 Hz, 1H), 4.04 (d, J = 2.9 Hz, 2H).Hydrate: 1 H NMR (400 MHz, D2O): δ 5.38 (t, J = 5.2 Hz, 1H), 3.17 (d, J = 5.2 Hz, 2H). 13C NMR (126 MHz, D2O): δ 87.85, 58.08.These parameters closely match those found previously.(8) Typical crude yields of 53% do not adequately account for extensive hydration of the product.Analytical NMR quantification suggested that spectroscopic yields are typically in the range of 15%.Sulfoacetaldehyde concentrations were quantified by 1 H NMR by addition of a known concentration of phenol standard and calculated based on the integral ratio of the aldehyde-(δ 9.65) and geminal-diol hydrogen (δ 5.38) resonances in sulfoacetaldehyde and the aromatic hydrogen resonances of phenol (Fig S10 ) (Integrals of the α-Hs (δ 4.04, 3.17) are not suitable for quantification due to their acidity and H/D exchange).The pre-scan delay (D1) was set to 50 s, ensuring a quantitative relationship between the 1 H NMR integrals.Mass spectroscopy indicates that two major ions matching with the aldehyde (m/z = 122.95,exact mass = 122.98)and the hydrate (m/z = 140.97,exact mass = 140.99).The sulfoacetaldehyde quantification resulting from 1 H NMR is corroborated by the ESI mass spectrum of a mixture of 5.0 mM sulfoacetaldehyde (aldehyde + hydrate) and 5.0 mM sodium allylsulfonate (98±1% consistent with 1 H NMR quantification) (Fig S12).The sulfoacetaldehyde standard was derivatized with 2,4-dinitrophenylhydrazine (DNPH) using the same protocol as the reaction assays and used for calibration curve construction.

HPLC and LC-MS analysis.
Liquid chromatography for analytical HPLC was performed using a mixture of water and acetonitrile (MeCN) spiked with 0.1% trifluoroacetic acid.The chromatographic sequence involved a 1 min solvent gradient from 0 -20% MeCN, a 10 min gradient from 20 -40% MeCN, a final 1 min gradient from 40 -100% MeCN and a holding phase at 100% MeCN for 5 min at a constant flow rate of 1 mL min -1 .Liquid chromatography for LC-MS was performed using a 10 min solvent gradient from 0 -100% MeCN and holding at 100% MeCN for 5 min at a flow rate of 1 ml min -1 .The usage of this LC-MS data was limited to the detection of DNPH-SA ions (m/z = 303 for DNPH-SA, m/z = 304 for DNPH-d1-SA), and all product quantifications were based on the integrated absorbance at 360 nm from the HPLC analysis.
Electron Paramagnetic Resonance Spectroscopy.EPR studies were performed in the CalEPR center in the University of California, Davis.X-band continuous-wave EPR spectra were collected on a Bruker Biospin EleXsys E500 spectrometer equipped with a super high Q resonator (ER4122SHQE) in the perpendicular mode.Cryogenic temperatures were controlled by an ESR900 liquid helium cryostat with a temperature controller (Oxford Instrument ITC503).Spectrometer settings were: conversion time, 60 ms; modulation amplitude, 0.8 mT; modulation frequency, 100 kHz.All data was collected at 15 K and 0.6325 mW power.Small amounts of adventitious Mn II signal were subtracted using the corresponding spectrum obtained on an aqueous solution of manganese(II) chloride with the same experimental parameters.EPR spectra were simulated in Matlab with Easyspin 5.2.28 toolbox.(9) The Mn/ 57 Fe-SfbO was reconstituted with the same protocol as Mn/ NA Fe except that 57 FeCl2 was used in place of natural abundance Fe(II) sources.

Figure S6. (A) Anaerobic reaction assays of 1 μM
Mn/Fe-and Fe2-SfbO with 200 mM PIPES in the presence of 2 mM H2O2.These assays were set up in an anaerobic chamber and incubated at room temperature for 3 h.No significant production SA was detected.(B) Reaction assays of Mn/Fe-SfbO and (C) Fe2-SfbO with 200 mM PIPES in the presence of catalase were performed with or without addition of 0.1 mg/ml bovine Cu/Zn superoxide dismutase (SOD).No significant activity alteration was observed with the addition of SOD.(D) Dependence of the catalytic activity of SfbO on Mn II and Fe II at a total metal equivalency of 2 per monomer.Premixed metal stocks containing different ratios of Mn 2+ and Fe 2+ were added to Apo-SfbO and incubated at room temperature anaerobically for 1 h before the reaction assays were performed.The results indicate that the heterometallic Mn/Fe form is the most active cofactor for this substrate.S5).Weblogo 3 was used for visualization (13).The range of sequence identity is 25 -100%.Table S5.Uniprot IDs of candidate AROs that contains small PF00355 protein within their genome neighborhood.

Figure S1 .
Figure S1.(A-D) Macromolecular structures of selected AROs.Cartoon representations of the structures of SfbO (A), B6JH23 (B), W7L2Y2 (C), and D3EI84 (D).One monomer of each of the homodimeric protein complexes is colored in front.The helices and beta sheets corresponding to the canonical TIM barrel fold are explicitly numbered for clarity.Metal atoms are rendered as orange red spheres.The diiron-metalated active site structures of SfbO (E), B6JH23 (F), W7L2Y2 (G), D3EI84 (H).

Figure S2 .
Figure S2.Structures of PtmU3 and AibH1H2.(A) Dimanganese form of PtmU3 (6OMP).(B) Diiron form of AibH1H2 (8FUM).A molecule of Tris coordinates the Site 2 metal via its amino and hydroxyl groups.(C) Amino acid residues that constitute a substrate binding pocket (highlighted in beige) in PtmU3 bound with PTM substrate (PDB 6OMQ).(D) Amino acid residues that constitute a substrate binding pocket (highlighted in dark purple) of AibH1H2 (PDB 8FUM).

Figure S3 .
Figure S3.UV/visible traces demonstrating the formation of the Fe III -OHx-Fe III LMCT bands (complementary to Fig 2B) upon exposure to air.Spectra were taken at intervals of 30 seconds until saturation.

Figure S4 .
Figure S4.Discovery of SfbO reactivity on MES.(A) Reactivity trials for Fe-SfbO on small molecules containing alkene group at different positions, and isethionate.The trial assays of styrene sulfonate were derivatized via a colorimetric NBP assay for epoxide detection (12) and analyzed with HPLC.The trial assays of allylsulfonate, vinylsulfonate and isethionate were analyzed by LCMS.A product eluting at 1.7 mL exhibited an m/z = 123 was observed for enzymatic reactions of vinylsulfonate and isethionate.(B) Single turnover assays were performed on Fe-SfbO and analyzed with HPLC.The air-oxidized protein was reduced with 10 mM dithionite anaerobically, and the excessive dithionite was removed by buffer exchange.The assay mixtures were prepared anaerobically first and exposed to air for 20 min before quenching by the DNPH derivatization method.(i) A reaction containing 200 μM reduced Fe-SfbO, 200 μM supplementary Fe II , 400 μM sodium isethionate, and 50 mM MES at pH = 6.5.(ii) Assay mixtures with the Fe-SfbO excluded.(iii) Assay mixtures with the supplementary Fe II excluded.(iv) Assay mixtures with the sodium isethionate excluded.(v) Assay mixtures with the MES excluded.The asterisk peak (*) at retention time of 9.1 min is the DNPH-SA product.The peak at 10 min comes from a consistently-observed, unknown DNPHderived impurity.(C) and (D) The potential alpha C-H bond hydroxylation product is sulfite, which was not observed in the reaction of Fe-SfbO or Fe-B6JH23 with MES or taurine.(C) The calibration curve used for sulfite quantification by using Ellman's reagent.(10) (D) Reactions containing 50 μM enzymes, 5 mM taurine, 50 mM MES pH = 6.5, 100 uM supplementary Fe II , and 5 mM sodium ascorbate.The enzyme in the negative control assay is replaced by 100 uM Fe II .The assays were allowed to incubate at room temperature for 3 h before being analyzed by the Ellman's reagent.

Figure S5 .
Figure S5.(A) Reactivity of as-isolated Fe-SfbO and Mn-SfbO (0.5 μM) supplemented with additional Fe II (1 μM) towards the oxidative conversion of MES to 2-sulfoacetaldehyde.The presence of catalase was found to markedly influence the reaction outcome.Reactions were performed in triplicate.(B) Substrate scope tests of the reactivity of as-isolated Fe-SfbO and Mn-SfbO with additional Fe II with the listed sulfonate-containing substrates (200 mM if not labeled specifically).Negative controls lack SfbO and contain 1 μM free Fe II and Mn II .SA was not detected from the reaction of SfbO with 20 mM MOPS that was used as buffer.(C) Representative LC-MS traces of enzymatic assays.The DNPH-SA product ion (m/z = 303) was observed only when all reaction components are included.(D) Assays of Mn/Fe-SfbO (0.5 μM), Fe2-SfbO (1 μM), and Apo-SfbO (1 μM) with 200 mM PIPES incubated without shaking at room temperature anaerobically or aerobically.All reaction components were degassed for at least one hour.No significant product formation was detected in the anaerobic reactions or the reactions of Apo-SfbO.(E) Schematic rationalizing the formation of d 1 -sulfoacetaldehyde (d 1 -SA) in assays of Mn/Fe-SfbO with d 18 -PIPES.The initially-formed d 3 -sulfoacetaldehyde product is rapidly converted to d 1 -SA via hydrogen-deuterium exchange under the acidic workup conditions required for DNPA derivatization.(F) DNPH-functionalized d 1 -SA (m/z = 304) is the sole product detected in the reaction of Mn/Fe-SfbO with d 18 -PIPES.No DNPH-SA (m/z = 303) was detected, confirming that SA generation is derived from the PIPES substrate.(G) The formation rate of SA by Mn/Fe-SfbO is linear for the first 30 minutes of the assay (R 2 = 0.996).

Figure S12 .
Figure S12.ESI Mass spectrum resulting from a neutral solution containing 5.0 mM sulfoacetaldehyde (aldehyde + hydrate) and 5.0 mM sodium allylsulfonate.The relative abundance of allylsulfonate ion (m/z = 120.96) is set to 100% for comparison with sulfoacetaldehyde ions.Experiment was performed in triplicate (black, red, blue traces).

Table S1 .
Details and coloring scheme of the most conserved genes within three coding regions of the small Rieske proteins present in Fig 1B.

Table S2 .
Observed metal content of differentially-metalated SfbO samples.Values smaller than the detection limit (0.01 ppm) of the ICP-OES instrument are reported as zero.Samples of as-iso Fe-SfbO were obtained following crude protein purification were not subject to any chelation.Each sample was measured by ICP-OES at least twice.

Table S3 .
Pairwise sequence identity between the 4 featured AROs.

Table S4 .
X-ray crystallography data collection and refinement statistics.Numbers in parenthesis correspond to statistics of the highest resolution shell.