A fully reversible 25-hydroxy steroid kinase involved in oxygen-independent cholesterol side-chain oxidation

The degradation of cholesterol and related steroids by microbes follows fundamentally different strategies in aerobic and anaerobic environments. In anaerobic bacteria, the primary C26 of the isoprenoid side chain is hydroxylated without oxygen via a three-step cascade: (i) water-dependent hydroxylation at the tertiary C25, (ii) ATP-dependent dehydration to form a subterminal alkene, and (iii) water-dependent hydroxylation at the primary C26 to form an allylic alcohol. However, the enzymes involved in the ATP-dependent dehydration have remained unknown. Here, we isolated an ATP-dependent 25-hydroxy-steroid kinase (25-HSK) from the anaerobic bacterium Sterolibacterium denitrificans. This highly active enzyme preferentially phosphorylated the tertiary C25 of steroid alcohols, including metabolites of cholesterol and sitosterol degradation or 25-OH-vitamin D3. Kinetic data were in agreement with a sequential mechanism via a ternary complex. Remarkably, 25-HSK readily catalyzed the formation of γ-(18O)2-ATP from ADP and the C25-(18O)2-phosphoester. The observed full reversibility of 25-HSK with an equilibrium constant below one can be rationalized by an unusual high phosphoryl transfer potential of tertiary steroid C25-phosphoesters, which is ≈20 kJ mol−1 higher than that of standard sugar phosphoesters and even slightly greater than the β,γ-phosphoanhydride of ATP. In summary, 25-HSK plays an essential role in anaerobic bacterial degradation of zoo- and phytosterols and shows only little similarity to known phosphotransferases.

The degradation of cholesterol and related steroids by microbes follows fundamentally different strategies in aerobic and anaerobic environments. In anaerobic bacteria, the primary C26 of the isoprenoid side chain is hydroxylated without oxygen via a three-step cascade: (i) water-dependent hydroxylation at the tertiary C25, (ii) ATP-dependent dehydration to form a subterminal alkene, and (iii) water-dependent hydroxylation at the primary C26 to form an allylic alcohol. However, the enzymes involved in the ATP-dependent dehydration have remained unknown. Here, we isolated an ATP-dependent 25-hydroxy-steroid kinase (25-HSK) from the anaerobic bacterium Sterolibacterium denitrificans. This highly active enzyme preferentially phosphorylated the tertiary C25 of steroid alcohols, including metabolites of cholesterol and sitosterol degradation or 25-OHvitamin D 3 . Kinetic data were in agreement with a sequential mechanism via a ternary complex. Remarkably, 25-HSK readily catalyzed the formation of γ-( 18 O) 2 -ATP from ADP and the C25-( 18 O) 2 -phosphoester. The observed full reversibility of 25-HSK with an equilibrium constant below one can be rationalized by an unusual high phosphoryl transfer potential of tertiary steroid C25-phosphoesters, which is ≈20 kJ mol −1 higher than that of standard sugar phosphoesters and even slightly greater than the β,γ-phosphoanhydride of ATP. In summary, 25-HSK plays an essential role in anaerobic bacterial degradation of zoo-and phytosterols and shows only little similarity to known phosphotransferases.
Cholesterol and related zoo-, phyto-, and mycosterols share a common architecture comprising a tetracyclic sterane core structure linked to an isoprenoid side chain. They are involved in maintaining membrane fluidity in eukaryotes and serve as precursor molecule for a large number of steroids with various endocrine and other physiological activities (1,2). Microbial transformation/degradation of steroids alters the sex steroid profile of eukaryotic hosts and is essential for bacterial pathogens such as Mycobacterium tuberculosis to persist in macrophages (3)(4)(5)(6). Once released into the environment, their hydrophobicity makes steroids ubiquitous persistent pollutants of global environmental concern. Microbial biodegradation is a major process to eliminate endocrine-disrupting steroids from the environment (7)(8)(9).
In aerobic cholesterol degrading microorganisms, side chain degradation is initiated by oxidation of ring A (10,11) and/or by cytochrome P450 monooxygenase-dependent hydroxylation and oxidation of the unactivated primary C26 to a C26 carboxylate (Fig. 1A), (12,13). After activation of the latter to a CoA ester, the side chain is degraded by modified β-oxidation to acetyl-CoA, propionyl-CoA, and androsta-1,4-diene-3,17dione (ADD), (14). In anaerobic bacteria, the cholesterol side chain is also degraded by β-oxidation via a C26 carboxylate intermediate (15). However, oxidation of primary C26 is achieved by an oxygenase-independent strategy that has been studied mainly in the denitrifying β-proteobacterial model organism Sterolibacterium denitrificans Chol-1S.
Recently, we demonstrated in extracts of S. denitrificans grown with cholesterol and nitrate that the formation of a tertiary alcohol at the steroid side chain initiates an enzymatic reaction cascade that finally results in the formation of a C26-carboxylate. It comprises the ATP-dependent dehydration of 25-OH-CDO to the Δ24 alkene desmost-1,4-diene-3-one (DDO), followed by a second water-dependent hydroxylation at the allylic C26 giving (E)-26-OH-DDO ( Fig. 1) (15). The latter step is catalyzed by steroid C26 dehydrogenase 1 (S26 DH 1 ), a further molybdenumdependent enzyme of the DMSO reductase family that also catalyzed the subsequent dehydrogenation of the allylic alcohol to the corresponding aldehyde. The C-H bond dissociation energy to carbocations and a hydride is significantly lower at the allylic C26 of DDO in comparison to the primary C26 of CDO or cholesterol. In summary, this unprecedented cascade allows for the hydroxylation of an unactivated primary carbon with water at the expense of one ATP (Fig. 1B) (15).
In agreement, S. denitrificans extracts converted the chemically synthesized 25-phospho-CDO to desmost-1,4diene-3-one (DDO) demonstrating that it represents an intermediate during the ATP-dependent dehydration of 25-OH-CDO to DDO (15). It has remained unclear, whether the assumed phosphorylation and phosphate elimination are catalyzed by a single enzyme or by a distinct kinase and lyase. The gene(s) and enzyme(s) involved in the ATP-dependent dehydration of 25-OH-steroids have remained unknown.
In this work, we aimed at identifying the enzyme involved in formation of the assumed 25-phospho-CDO intermediate during anaerobic cholesterol degradation in the denitrifying β-proteobacterium S. denitrificans. For this purpose, a 25-phospho-CDO forming activity was enriched from S. denitrificans to allow for the identification of the encoding gene. After its heterologous production, we identified a highly active and specific 25-hydroxysteroid kinase (25-HSK) that ATP-dependently formed the 25phosphoester of CDO and related 25-OH-steroids with an extremely high phosphoryl group transfer potential.

ATP-dependent dehydration of 25-OH-CDO involves a 25-hydroxy-steroid kinase
To identify the unknown enzyme(s) involved in the ATPdependent conversion of 25-OH-CDO to DDO, we opted for an enrichment from soluble extracts of S. denitrificans cells grown with cholesterol and nitrate as carbon and energy sources. Activity was monitored in an ultraperformance liquid chromatography (UPLC) based assay following the ATPdependent consumption of 25-OH-CDO and formation of product(s) (Fig. S1). After a first DEAE anion exchange chromatography enrichment step, protein fractions were obtained that converted 25-OH-CDO (0.5 mM) to 25-phospho-CDO as the only product in the presence of 2-hydroxypropyl-βcyclodextrin (HPCD, solubility enhancer) and MgATP (5 mM) (Fig. S1). This finding indicates that the ATP-dependent dehydration of 25-OH-CDO involves two distinct activities, a kinase activity henceforth referred to as 25-hydroxy-steroid kinase (25-HSK) and a phosphate eliminating C-O lyase. Further enrichment of the kinase activity via hydrophobic interaction chromatography and size-exclusion chromatography yielded two major protein bands migrating at 30 and 40 kDa in SDS-polyacrylamide gels and a minor one migrating at 50 kDa ( Fig. 2A). The final 25-HSK activity enrichment factor was 117 with a yield of 8% (Table 1).
Mass spectrometric analyses of tryptic peptides from the three dominant enriched protein bands identified proteins from S. denitrificans with highest scores to acyl-CoA synthetases (60 kDa band), phytoene dehydrogenases (50 kDa), phosphotransferases (40 kDa band), and luciferases (30 kDa band, Table S1). Thus, the 40 kDa protein is the most likely 25-hydroxy-steroid kinase candidate responsible for 25-OH steroid phosphorylation. In a previous differential proteome analysis, the corresponding gene product (SDENChol_21286, S. denitrificans genome annotation, A0A7Z7HU23) was upregulated in cells grown with cholesterol versus cells grown with testosterone that lacks the isoprenoid side chain (21). Protein blast analysis of the protein with the NCBI database (refseq_protein) showed highest similarities to gene products from the 17β-estradioldegrading Denitratisoma sp. DHT3 (55% and 51% amino acid sequence identity, respectively, accession numbers WP_186453974.1 and WP_145841310.1) (Fig. 2C, Table S2). High similarities (≤46% identity) were also found to putative phosphotransferase family gene products from Actinobacteria.
Heterologous expression of the encoding gene yields a highly active 25-HSK The putative gene encoding 25-HSK (WP_154717265.1) was cloned and expressed in E. coli BL21 as C-terminal Streptagged fusion protein. Extracts from E. coli BL21 expressing this gene catalyzed 25-OH-CDO phosphorylation with a specific activity of 0.9 ± 0.05 μmol min −1 mg −1 , whereas cells containing the same plasmid without the gene showed no such activity (<0.001 μmol min −1 mg −1 ). After enrichment by Strep-Tactin affinity chromatography, SDS-PAGE analysis of the active protein pool revealed a single protein band with an apparent molecular mass of approximately 40 kDa (Fig. 2B) fitting to the mass (43.3 kDa) deduced from the amino acid sequence. The native molecular mass of recombinant 25-HSK was determined by gel filtration and was 45 ± 5 kDa, indicating a monomeric architecture. The 25-HSK preparation catalyzed the time-dependent transformation of 1 mM 25-OH-CDO to 25-phospho-CDO in the presence of 5 mM ATP and 10 mM MgCl 2 at a specific activity of 19.5 ± 0.5 μmol min −1 mg −1 (Fig. S2). The enzyme was used for all kinetic experiments and was stored in 50 mM MOPS/KOH pH 7 buffer for several months at 4 C without a significant loss of activity.

25-HSK prefers tertiary C25 alcohols of steroids
For testing the substrate preference of 25-HSK, a photometric enzyme assay was used that coupled ATP hydrolysis to the reduction of NADH to NAD + in the presence of 25-hydroxy-steroid kinase phosphoenolpyruvate, pyruvate kinase, and lactate dehydrogenase. The highest specific activity was observed with 25-OH-CDO (0.5 mM) at 16 ± 0.6 μmol min −1 mg −1 fitting well to the activity determined in UPLC-based assays (Fig. 3). Only slightly lower values were determined with other C25 alcohols of steroids with modifications in ring A or in the isoprenoid side chain. 25-OH-steroids with other modifications were converted at lower rates such as 25-hydroxy-vitamin D 3 (VitD 3 ). Nonsteroidal tertiary alcohols were converted albeit only at <0.1% of the activity of 25-OH-CDO. Notably, low but significant activities were also detected for primary 26-OHsteroids such as (25R)-26-OH-CDO and (25R)-26-OH-CEO.

25-HSK reaction is fully reversible
Although the ATP concentration was five times higher than that of 25-OH-CDO in the assays, the reaction always levelled off when 70-80% 25-OH-CDO was converted (Fig. S2). To test the possibility whether this observation was due to the establishment of a thermodynamic equilibrium, we followed the reaction at equimolar ATP and 25-OH-CDO concentrations (1 mM each) and observed that the reaction levelled off when 50% of the substrates were converted into the products (Fig. 4, A and B). A similar result was obtained when the reaction was run in the reverse reaction, thereby following the ADP (1 mM) and 25-phospho-CDO (1 mM) dependent synthesis of ATP and 25-OH-CDO (Fig. 4, C and D).
To verify that the γ-phosphoryl group of the ATP formed in the reverse reaction derives indeed from the 25-phosphoester, we enzymatically synthesized ( 18 O   Kinetic data indicate a ternary complex mechanism and an equilibrium constant near 1 The stoichiometry and kinetics of the 25-HSK catalyzed phosphorylation were determined using the UPLC-based enzymatic assay. In the presence of saturating concentrations of 25-OH-CDO and ATP, a stoichiometry of 1.1 ± 0.1 mol ATP hydrolyzed to ADP per mol 25-OH-CDO formed was determined. In agreement, 25-HSK exhibited only marginal ATPase activity in the absence of 25-OH-CDO (<0.1% of 25-OH-CDO dependent ATPase activity). The kinetic data obtained were evaluated by a fit to Michaelis-Menten curves (Fig. S3), and K m values of 0.3 ± 0.01 mM for 25-OH-CDO and 0.8 ± 0.05 mM for ATP were calculated for the forward reaction (for a full list of kinetic parameters determined, see Table S3). The reaction was also followed in the reverse direction in the presence of 5 mM ADP and 1 mM 25-phospho-CDO giving K m values of 8 ± 0.7 mM for 25-phospho-CDO and 35 ± 3 μM for ADP (Fig. S3). Notably, the high K m value for 25-phospho-CDO has to be taken as an apparent value because only 25-phospho-CDO concentrations up to 4 mM (stock concentration: 40 mM in 80% [v/v] 2-propanol/H 2 O) could be tested. At higher concentrations, the added solvent 2propanol inhibited 25-HSK significantly (at concentrations above 10% [v/v]). As a result, no hyperbolic saturation was reached for the reverse reaction.
The determination of approximately equal equilibrium concentrations of substrate and products (Fig. 4) suggested an equilibrium constant (K eq ) of 25-HSK catalyzed reaction near 1. Using the kinetic parameters obtained (Table S3), we  25-hydroxy-steroid kinase determined K eq according to the Haldane equation (V f , K mf = forward reaction; V r , K mr = reverse reaction).
A value of 0.34 was determined, which is in line with the equilibrium concentrations of products and educts determined (Fig. 4A).
Reactions catalyzed by kinases usually follow a sequential or a ping-pong mechanism (22). In the former, phosphoryl transfer occurs within ternary complexes between the enzyme and its substrates. Both scenarios can be distinguished for 25-HSK by analyzing rates at varying ATP and 25-OH-CDO concentrations. Double-reciprocal plots of 25-HSK activity versus 25-OH-CDO concentrations at four different ATP concentrations yielded clearly intersecting lines, which is consistent with a ternary-complex mechanism (Fig. 6). This

25-hydroxy-steroid kinase
result is consistent with the inability to observe a phosphorylated 25-HSK intermediate on SDS-gels that had been preincubated with γ-32 P-ATP (not shown).

Discussion
In this work, we identified a highly active kinase that catalyzes the ATP-dependent conversion of 25-OH-to 25phospho-steroids. This unprecedented phosphorylation of a hydroxyl functionality at a steroid side chain represents a crucial step in the oxygen-independent degradation of zooand phytosterols. The identification and isolation of 25-HSK leave little doubt that ATP-dependent dehydration of tertiary 25-OH-steroids to the corresponding Δ24-alkenes is a twostep process involving 25-HSK and a so far unknown C-O lyase that catalyzes the subsequent elimination of phosphate from 25-phospho-CDO. This system differs from the only other known ATP-dependent dehydration system involved in the dehydration of hydrated forms of NAD(P)H (23). The latter reaction plays a role in nicotinamide nucleotide repair and is catalyzed by a single enzyme. Results from kinetic analyses in this work clearly indicate a sequential phosphoryl transfer mechanism of 25-HSK via a ternary complex composed of the enzyme, 25-OH-steroid, and MgATP. Such a mechanism has been reported for the majority of kinases (22).
The anaerobic degradation of zoo-and phytosterols in the model organism S. denitrificans involves specific Modependent hydroxylases acting either on tertiary C25 or at allylic C26 (13, 17) (Fig. 7). In contrast, there is only a single gene encoding 25-HSK in the genome of S. denitrificans. This finding explains the promiscuity of 25-HSK as evidenced by its ability to phosphorylate 25-OH-steroids with side chain modifications (25-OH-CDO or 25-OH-sitost-1,4-diene-3-one) or with modifications in the sterane skeleton (25-OH-CEO or 25-OH-VitD 3 ). Most likely, a promiscuous 25-HSK/25phosphosteroid lyase tandem accomplishes the ATPdependent dehydration of various 25-OH-steroids (Fig. 7). The gene encoding the proposed 25-phospho-steroid lyase is unknown, and there is no candidate gene in vicinity of the gene encoding 25-HSK. The phosphorylation of primary 26-OHsteroids is a further promiscuity of 25-HSK, which appears to be rather problematic for the cell because the corresponding 26-phospho-steroids formed are no intermediates during steroid side chain degradation. However, the observed ATPconsuming 26-OH-steroid phosphorylation will be largely prevented in vivo by the immediate oxidation of 26-OH-steroid intermediates to the corresponding aldehydes by the 26-OH-steroid forming S26 DHs (15). In contrast to the observed broad substrate spectrum with regard to modification in the isoprenoid side chain or ring A of the sterane skeleton, low-molecular tertiary alcohols such as tert-butanol were merely converted (<0.1% of the rate). This finding indicates that the sterane skeleton is essential for specific binding of the tertiary alcohol. Putative gene products with high similarities to 25-HSK from S. denitrificans were identified in the denitrifying, estradiol-degrading strain Denitratisoma sp. DHT3, though its ability to use side-chain-containing steroids was not reported, yet (Fig. 2) (24). This finding motivated the search for genes encoding putative S25 DHs and S26 DHs, and we indeed identified the corresponding candidate genes in the genome (Table S2). Thus, we propose that Denitratisoma sp. DHT3 can degrade side-chain-containing steroids, thereby  25-hydroxy-steroid kinase using the same pathway for steroid side chain degradation as established for S. denitrificans. In other denitrifying bacteria that are restricted to the degradation of steroids without side chains such as estradiol (Denitratisoma oestradiolicus) (25,26) or testosterone (S. denitrificans) (27,28), no candidate genes for 25-HSK are present in the genome. The 25-phosphoester intermediate of anaerobic steroid side chain degradation shows an extremely high phosphoryl group transfer potential, which is in the range of the β,γ-phosphoanhydride of ATP as evidenced by: (i) the establishment of an equilibrium at almost equal substrate/product concentrations, (ii) the establishment of this equilibrium in the forward and reverse reaction, (iii) a K eq value of 0.34 as calculated by the Haldane equation using the kinetic parameters determined in this work, and (iv) the formation of γ-( 18 O) 2 -ATP from ADP and the ( 18 O) 2 -C25-phosphoester. The standard free Gibbs energy of ATP hydrolysis is around 32 kJ mol −1 (29). Our data indicate that 25-phospho-CDO hydrolysis will be similar if not even slightly more negative, around 20 kJ mol −1 more negative than standard sugar-phosphate esters (-13 kJ mol −1 ) (29). However, although 25-phosphosteroids can be regarded as an energy-rich cellular molecule with the potential for ATP synthesis, such a reaction will unlikely occur in the cell. During cholesterol degradation, 25-phospho-steroids accumulate only to a very low steady-state concentrations as the subsequent phosphate elimination to the corresponding alkene is considered essentially irreversible (15). The rational for the extremely high phosphoryl group transfer potential of the 25phosphoesters is likely the highly hydrophobic environment of the double negatively charged phosphoryl-group (at pH 7) that is surrounded by three alkyl functionalities whereas in ATP it is fully hydrated. Other examples of an extremely energy-rich phosphoester in biology are phosphenolpyruvate and poly-phopshorylated inositols. In the former, dephosphorylation allows the unfavorable enol to tautomerize to the much more stable ketone, which renders pyruvate kinase essentially irreversible in the direction of ATP synthesis (30). In poly-phosphorylated inositols, the high electrostatic repulsion increases the Gibbs free energy of hydrolysis in comparison to sugar-monophosphates. Indeed, reversibility of inositol phosphate (hydroxyl) kinases has been experimentally shown (31), albeit with K eq values of 14 being clearly higher than that of 25-HSK (32).

Chemicals and bacterial strains
The chemicals used in this work were of analytic grade. S. denitrificans Chol-1S (DSMZ 13999) was obtained from the Deutsche Sammlung für Mikroorganismen und Zellkulturen (DSMZ). E. coli BL21 (DE3) and E. coli 5α were obtained from New England Biolabs (NEB).

Culture conditions and preparation of cell extracts
S. denitrificans was cultivated under denitrifying conditions in a phosphate-buffered medium (3 g L −1 NaH 2 PO 4 × 2 H 2 O, 4 g L −1 K 2 HPO 4 , 0.54 g L −1 NH 4 Cl at pH 6.9) at 30 C with 2 mM cholesterol as carbon source in a 200-L-fermenter. Cells were harvested in the late exponential phase (8000g, 20 min, and 4 C) and frozen in liquid nitrogen. Frozen cells were lysed by a French pressure cell at 137 MPa in the presence of two volumes of lysis buffer containing 50 mM MOPS/KOH (pH 7.0) and 0.1 mg DNase I.
The ADP-dependent reverse reaction was followed at 30 C (750 rpm) in the presence of 50 mM Tris/HCl (pH 7.8), 1 mM 25-phospho-CDO, 5 mM ADP, 10 mM MgCl 2 , 9% (w/v) HCPD, and 5-50 μg ml −1 25-HSK unless otherwise stated. The enzyme assays were prepared for separation via UPLC as described above. The kinetic parameters were determined via UPLC-based enzymatic assays by varying one substrate concentration at a fixed saturating concentration of the other. For 25-OH-CDO, 5 mM ATP and for 25-phospho-CDO, 5 mM ADP was added to determine kinetic parameters. For ATP or ADP, 1 mM 25-OH-CDO or 1 mM 25-phospho-CDO was added, respectively.

Enrichment of 25-HSK from cholesterol grown S. denitrificans
After ultracentrifugation (150,000g, 1.5 h, and 4 C) of crude extracts from S. denitrificans, the soluble protein fraction was applied to a DEAE-Sepharose column (60 ml, GE Healthcare) at 5 ml min −1 with buffer A 1 (20 mM  Active fractions were concentrated, diluted in ten volumes buffer C (20 mM HEPES/KOH, pH 7.6, and 150 mM KCl), and applied to a Superdex 200 Increase 10/300 GL column at a flow rate of 0.5 ml min −1 .

Protein identification by mass spectrometry of peptides
Proteins were identified by excising the bands of interest from a SDS-PAGE gel. The bands were washed 3× with 80 μl 10 mM NH 4 HCO 3 (10 min) and 80 μl 50% EtOH in 5 mM NH 4 HCO 3 (10 min). The gel bands were incubated with 80 μl 10 mM DTT in 10 mM NH 4 HCO 3 (30 min, 56 C), the supernatant was removed, and 80 μl 50 mM iodoacetamide in 10 mM NH 4 HCO 3 was added (30 min, 25 C). The supernatant was removed and the bands were washed 3× with 10 mM NH 4 HCO 3 (10 min) and 100% EtOH, and then dried with a Speedvac Vacuum Concentrator (60 min, 40 C). The dried gel bands were incubated with 20 μl trypsin at 37 C (0.6 μg). The proteins were released from the gel bands by adding 75 μl 0.05% TFA in 50% acetonitrile. The samples were sonicated for 10 min and afterward dried via a Speedvac Vaccum Concentration (120 min, 45 C). The dried products were dissolved in 20 μl 0.1% TFA and applied to an Acquity UPLC I-class system (Waters) using an Acquity UPLC peptide CSH C18 column (1.7 μm, 2.1 mm × 150 mm). For separation, a gradient of 40% acetonitrile + aqueous 0.1% formic acid at a flowrate of 40 μl min −1 was used. Proteins were identified using a Synapt G2-Si high-resolution ESI-Q-TOF mass spectrometry system (Waters) operating in positive mode at 3 kV capillary voltage, 350 C desolvation temperature, and a desolvation gas flow of 800 L h −1 . The resulting spectra were analyzed with Pro-teinLynx Global Server 3.0.3 (Waters) by matching with the UniProt database of S. denitrificans (3020 protein-coding sequences). The following parameters were used for protein identification via ProteinLynx Global Server: Mass tolerance for precursor ions = 5 ppm, mass tolerance for fragment ions = 13 ppm, minimal fragments per peptide = 3, minimal peptides per protein = 1, minimal fragments per protein = 7, false discovery rate = 4%.

LC-ESI-MS analysis of steroid compounds
Metabolites were analyzed by an Acquity I-class UPLC system (Waters) using an Acquity UPLC CSH C18 (1.7 μm, 2.1 × 100 mm) column coupled to a Synapt G2-Si HRMS ESI-Q-TOF device. For separation, an aqueous 10 mM ammonium 25-hydroxy-steroid kinase acetate/acetonitrile gradient was used as described above (UPLC). Samples were measured in positive mode with a capillary voltage of 2 kV, 100 C source temperature, 450 C desolvation temperature, 1000 L min −1 N 2 desolvation gas flow, and 30 L min −1 N 2 cone gas flow. The evaluation of the data was performed using the MassLynx 4.1 (Waters) software. Metabolites were identified by their retention times and m/z values.
Heterologous gene expression and production of 25-HSK Strep in E. coli BL21 The gene encoding 25-HSK (SDENChol_21286) was amplified with a C-terminal Strep-tag by PCR from S. denitrificans genomic DNA as the template using the following primers: 21286_for (GTCATGGCTAGCATGTCT GATACACCGGATCTG) and 21286_rev (CTAGTCAAG CTTTTACTTTTCGAACTGCGGGTGGCTCCATCCTCCG AAATCCCAGGGATCTTCGC). The resulting 1.2 kb DNA fragment was NheI/HindIII double digested, cloned into pASK-IBA15plus, and transformed into E. coli BL21 according to protocols from New England Biolabs (NEB). Induction of 25-HSK was carried out in 2× YT medium (18 g L −1 tryptone, 10 g L −1 yeast extracts, and 5 g L −1 NaCl) supplemented with 100 μg ml −1 ampicillin. When optical density at 600 nm reached values above 0.5, 20 μg ml −1 anhydrotetracycline (AHT) was added. Induction was carried out at 20 C while shaking at 180 rpm. Cells were harvested in the late exponential phase (>16 h induction) by centrifugation (8000g, 20 min, and 4 C), and frozen in liquid nitrogen.
Synthesis γ-( 18 O) 2 -ATP γ-( 18 O) 2 -ATP was synthesized as described previously (34) (for a scheme of synthesis and structures of chemicals used, see Fig. S4). In brief, a P-amidite 3 was generated by hydrolysis of benzylbromide 1 with 18 O labeled water (isotopic purity >98%). Benzylic alcohol 2 was reacted with 0.5 equivalents of diisopropylphosphoramidous dichloride giving P-amidite 3. The P-amidite 3 was used for the introduction of two terminal 18 O labels by reaction with ADP (tetrabutylammonium salt in DMF as solvent), followed by oxidation of the P(III)-P(V) mixed anhydride with mCPBA. The crude material was precipitated from solution by addition of diethylether and the benzyl protecting groups were removed by hydrogenation over Pd/C at elevated pressure (5 bar). The crude material was purified by chromatography over DEAE Sepharose (elution with 30-100 mM aq. NH 4 HCO 3 ). Excess buffer was removed by repeated freeze drying giving pure 18 O labeled ATP as ammonium salt. The ammonium salt was passed through a Dowex-H + column and aq. NaOH solution (2.0 eq.) was added. After lyophilization, the desired heavy ATP was obtained as a colorless solid. Mass spectrometry revealed high isotopic enrichment in the product (>97%).

Data availability
All data are available upon request. The protein mass spectrometry data have been deposited to the ProteomeXchange Consortium via the PRIDE (https://doi.org/10.1093/nar/gky11 06) partner repository with the dataset identifier PXD027031 and 10.6019/PXD027031. Supporting information-This article contains supporting information (34).