Small-molecule active pharmaceutical ingredients of approved cancer therapeutics inhibit human aspartate/asparagine-β-hydroxylase

Graphical abstract


Introduction
Human aspartate/asparagine-β-hydroxylase (AspH, BAH, HAAH) belongs to the family of 2-oxoglutarate (2OG) dependent oxygenases 1 and catalyses the post-translational hydroxylation of specific aspartyland asparaginyl-residues in human epidermal growth factor-like domains (EGFDs) using 2OG and O 2 as co-substrates and Fe(II) as a cofactor ( Fig. 1). 2,3 Hypoxia is reported to regulate the expression levels of human AspH 4,5 and upregulated levels of AspH have been detected on the cell surface of invasive human cancers such as hepatocellular carcinoma, 6,7 breast cancer, 8 and pancreatic cancer. 9 AspH is reported to retain its catalytic activity on the cancer cell surface. 10 The upregulation of AspH and its translocalization from the endoplasmic reticulum (ER) membrane to the cell surface in cancer cells correlates with enhanced cell motility and metastatic spread resulting in a reduced lifespan of the affected patients. 11,12 Studies on naturally occurring mutations of human AspH presumably resulting in inactive AspH (Traboulsi syndrome) [13][14][15][16] and animal models 8,17,18 suggest that AspH might affect cell motility through the notch signalling pathway. Despite the combined evidence suggesting that AspH is a promising medicinal chemistry target for the development of small-moleculebased cancer therapeutics, comparatively few AspH inhibition studies using small-molecules are reported, with most of these relying either on the use of likely non-selective 19,20 or partially selective 21 small-molecules or on the use of L-ascorbic acid (LAA)-derived small-molecules. 18,[22][23][24] This likely reflects the historic lack of simple and reliable assays to monitor recombinant human AspH activity in vitro. Pioneering assays employed (native) partially purified bovine or murine AspH and monitored AspH activity by analysing either 2OG turnover or the amino acid content of EGFD substrate peptides after acidic hydrolysis using mass spectrometry (MS). 20,[25][26][27] Recently, we reported a catalytically active truncated form of recombinant human AspH, His 6 -AspH 315-758 , comprising its catalytic oxygenase domain and the adjacent tetratricopeptide repeat (TPR) domain, which was suitable to perform in vitro turnover assays and crystallographic experiments. 28 We showed that AspH catalyses the Asp/Asn-hydroxylation of EGFDs with a non-canonical disulfide pattern (Cys 1-2, 3-4, 5-6; Supporting Fig. S1a) rather than with the wellcharacterized canonical disulfide pattern (Cys 1-3, 2-4, 5-6; Supporting Fig. S1a), 28 which has been predominantly found in EGFDbearing proteins. 29 Cyclic peptides were designed to mimic the central non-canonical (Cys 3-4) EGFD disulfide macrocycle and were applied as stable AspH substrate analogues in kinetic experiments, the results of which suggest that AspH activity may be unusually susceptible to limiting oxygen availability and thus that AspH might be involved in the physiologically relevant hypoxic response. 30 High-throughput solid phase extraction coupled to mass spectrometry (SPE-MS) AspH inhibition assays using a stable substrate analogue were developed and employed to identify potent AspH inhibitor templates. 21 The development of novel human therapeutics is time-consuming and often compromised by the failure of target and lead candidates to be approved for therapeutic use in humans. 31,32 The repositioning of small-molecule active pharmaceutical ingredients (APIs) of approved human therapeutics, meaning the repurposing of APIs against other than the approved indications, constitutes a potential alternative as multiple phases of the de novo drug discovery process can be bypassed. 31 One aspect of API repositioning involves the identification of previously unrecognized API-protein interactions. 31 Hence, profiling of approved small-molecule APIs against human enzymes other than the reported target enzymes is desirable. 31,33 This also helps to identify inhibitor scaffolds for structure activity studies and to define undesired off-target side effects of approved APIs. We thus aimed to screen APIs against AspH, with a special interest in cancer therapeutics.
We report the effects of 316 small-molecule APIs, which are either components of approved human cancer therapeutics or of human cancer therapeutics under current or previous clinical investigation, on the catalytic activity of AspH, employing SPE-MS inhibition assays. Both natural products and clinical inhibitors of human anti-apoptotic Bcell lymphoma-2 (Bcl-2) 34,35 proteins were identified to efficiently inhibit AspH.

Evaluation of small-molecule cancer APIs for AspH inhibition
Initially, a compound library composed of 316 small-molecules (the Approved Oncology Drugs Set of the National Cancer Institute/the National Institutes of Health Developmental Therapeutics Program combined with the TDI Expanded Oncology Drug Set of the Target Discovery Institute, University of Oxford), which are either APIs of approved human cancer therapeutics or of human cancer therapeutics under current or previous clinical investigation, was investigated for AspH inhibition under the previously established AspH inhibition assay conditions. 36 AspH substrate-(hFX-CP 101-119 ; Supporting Fig. S1b), 2OG-, and Fe(II)-concentrations close to their Michaelis (K m ) constants 30 were employed and substrate depletion/product formation (+16 Da) was monitored by SPE-MS to identify potent AspH inhibitors (Supporting Fig. S2 and Supporting Data Sheet). At a fixed inhibitor concentration (20 μM), eleven compounds inhibited > 80% AspH activity with respect to DMSO and pyridine-2,4-dicarboxylic acid (2,4-PDCA; Table 1, entry 1) 19,20,36 negative and positive controls, respectively (Table 1).
Two out of the remaining seven identified AspH inhibitors were histone deacetylase (HDAC) inhibitors, both bearing a hydroxamate functional group, which inhibited > 90% AspH activity at 20 μM: the HDAC6-selective inhibitor tubacin 48 (Table 1, entry 7) and the broadspectrum HDAC inhibitor belinostat (PXD101) 45 (Table 1, entry 5), which is an approved API for the treatment of relapsed or refractory peripheral T-cell lymphoma. 59 The two small-molecule APIs midostaurin 46 Vemurafenib is reported to inhibit selectively the human serine/ threonine kinase BRAF bearing an activating V600E mutation over the wild-type enzyme. It is approved for the treatment of metastatic melanoma and seems to impose a modest inhibitory effect on AspH activity (~20% AspH activity remaining at 20 μM; Table 1, entry 12).

Determination of AspH inhibitory concentrations
Fresh DMSO solutions of the ten identified AspH inhibitors were prepared from separately obtained commercial solids and used to determine half-maximum inhibitory concentrations (IC 50 ) using substrate (hFX-CP 101-119 ; Supporting Fig. S1b), 2OG-, and Fe(II)-concentrations close to their Michaelis (K m ) constants (Table 1, Fig. 2). 30 The Abbottdeveloped Bcl-2 inhibitor ABT-737 55 (Table 1, entry 13) was added to the panel due to its structural similarity with navitoclax and venetoclax. The AspH inhibition assays were of good quality which high signal-tonoise (S/N) ratios and high Z'-factors 56 (> 0.75 for each plate) manifest (Supporting Fig. S3). This is in agreement with our previous AspH inhibition assays and highlights the power of the SPE-MS highthroughput screening platform to perform AspH inhibition studies. 36 The IC 50 -values are in good agreement with the initial screening data at a fixed inhibitor concentration: (R)-Gossypol, bleomycin A 2 , and venetoclax all inhibited AspH with IC 50 -values of ~1 μM or below (Table 1). With the exception of navitoclax (IC 50 ~ 1.0 μM; Table 1, entry 7), which inhibits AspH with a similar potency as venetoclax, all the other identified APIs displayed only modest potencies (~4.1 to ~13.2 μM; Table 1). ABT-737 was about three fold less potent than its derivatives navitoclax and venetoclax (IC 50 ~ 3.4 μM; Table 1, entry 13). (R)-Gossypol inhibits AspH with similar potency as its Bcl-2 validated targets (i.e. IC 50 (Bcl-X L ) ~ 0.5 μM) 37 while the Abbott-developed APIs are significantly more potent against their original Bcl-2 protein

Mechanistic studies
To investigate whether the mechanisms by which the identified eleven small-molecule APIs inhibit AspH involve competition with 2OG or substrate for AspH binding or non-specific Fe(II)-chelation, their IC 50 -values were determined at elevated 2OG (~330·K m ; K m = 0.6 μM 30 ), elevated substrate peptide (hFX-CP 101-119 ; ~8·K m ; K m = 1.2 μM 30 ) or elevated Fe(II) (~14·K m ; K m = 1.4 μM 30 ) assay concentrations, respectively ( Table 2). We initially studied the potent but non-selective AspH inhibitor 2,4-PDCA (Table 1, entry 1) as a positive control. 19,20,36 These analyses demonstrated that 2,4-PDCA competes with 2OG for binding to the AspH active site: at an elevated 2OG concentration, 2,4-PDCA is about five times less potent than at the AspH K m -concentration for 2OG, while varying the AspH substrate peptide or Fe(II) concentrations did not affect its IC 50 -value (Table 2; entry 1). This observation is in agreement with prior crystallographic studies and biophysical experiments, 36 indicating that this SPE-MS AspH inhibition assay set-up is useful to probe the mechanism by which the small-molecule APIs inhibit AspH.
By contrast with the results for 2,4-PDCA, the IC 50 -values of the small-molecule APIs do not differ notably at elevated 2OG, substrate peptide or Fe(II) concentrations, within the experimental error ( Table 2). These results suggest that the APIs do not seem to inhibit AspH in a 2OG or substrate competitive manner or through lowering the available Fe(II) concentration. However, this analysis is not conclusive, as for some of the AspH inhibition curves, i.e. with bleomycin A 2 , venetoclax, belinostat, mithramycin A (plicamycin), obatoclax, and verumafenib, comparably low Hill coefficients 57 (< -2.0) were observed, with mithramycin displaying the most pronounced effect (Fig. 2). Low Hill coefficients, which diverge notably from the 'ideal' value (~−1), can be observed if enzymes contain multiple inhibitor binding sites. 60,61 Low Hill coefficients can also be a result of inhibitor aggregation, 62,63 which is a well-established phenomenon in inhibitor identification from high-throughput compound library screenings; 63-65 the small-molecules can aggregate to form colloidal particles 66 which then can associate with the enzyme and trigger denaturation. 67 Vemurafenib is reported to form colloidal particles in aqueous solutions and in cell-cultures, 68,69 indicating that aggregation might account for the observed inhibitory effect of some of the identified smallmolecule AspH inhibitors. The AspH inhibition assays were thus performed in the presence of detergents 70 with the aim of suppressing small-molecule aggregation. However, both the tested zwitterionic detergent CHAPS 71 and the non-ionic detergent triton-X100 72,73 suppressed AspH substrate peptide ionization at concentrations typically used to suppress aggregation; the use of detergents was hence not compatible with the SPE-MS AspH inhibition assay. An orthogonal binding assay was therefore applied to elucidate if the identified AspH inhibitors bind AspH. AspH melting temperatures (T m ) were assayed in the presence of the small-molecules using differential scanning fluorimetry (Supporting Fig. S4). Several identified AspH inhibitors were incompatible with the assay conditions due to intrinsic compound fluorescence interfering with the excitation/emission wavelengths of SYPRO orange; all the other inhibitors did not show a substantial effect on the AspH T m when compared to the reported 2OG competitor 2,4-  . S2). b Mean of two independent runs (n = 2; mean ± standard deviation, SD) using 50 nM His 6 -AspH 315-758 , 1 μM hFX-CP 101-119 (Supporting Fig. S1b  In the case of mithramycin A (plicamycin), Fe(II)-chelation could account for some of its observed inhibitory effect, because no inhibition (i.e. IC 50 > 50 μM) of AspH was observed at higher Fe(II) assay concentrations (Table 2, entry 9). This proposal is consistent with the chemical structure of mithramycin A, which bears functional groups that can engage in metal chelation (Table 1, entry 9), with its mode of action that requires divalent metal cations to form inhibitor-metal cation-DNA ternary complexes, 50 and with its negligible effect on the AspH T m (ΔT m = 0.4 ± 0.2 °C; Supporting Fig. S4).
The potency of bleomycin A 2 did not depend on the Fe(II) assay concentration ( Table 2, entry 3), an observation which is notable given that bleomycin A 2 is reported to require activation by Fe(II) and O 2 in order to exert its medicinal function as DNA-cleavage agent. 42,74-75 Hill coefficient analysis for bleomycin A 2 AspH inhibition curves indicates that compound aggregation alone is unlikely to account for the observed AspH inhibition; to our knowledge, bleomycin A 2 has not been reported to form colloidal aggregates in aqueous solutions. One possible explanation for its inhibitory effect could be that activated bleomycin A 2 , which is a known oxidant, 42 alters the assay redox equilibrium in the assay mixture, and thus indirectly inhibits AspH, which is known to be sensitive towards subtle changes in the redox environment. 30 Although further work is required, this possibility is in agreement with the results that altered 2OG, substrate peptide or Fe(II) assay concentrations did not affect bleomycin A 2 potency (Table 2, entry 3) and that bleomycin A 2 did not seem to bind AspH efficiently (ΔT m = −0.8 ± 0.4 °C; Supporting Fig. S4).
The efficiencies of the four most potent small-molecule AspH inhibitors identified apart from bleomycin A 2 (i.e. (R)-gossypol, venetoclax, navitoclax, and ABT-737), which are all reported inhibitors of Bcl-2 proteins, did not seem to depend on 2OG, substrate peptide or Fe(II) assay concentrations (Table 2). Even though the structures of these AspH inhibitor resemble those of reported small-molecule aggregators, 76,77 the Hill coefficient analysis for their AspH inhibition curves (Fig. 2) together with their slight effect on the AspH T m (Supporting Fig. S4) indicates that mechanisms other than aggregation might account for their observed inhibition of AspH. As co-crystallization experiments with AspH with the Bcl2-inhibitors have been unsuccessful, molecular docking studies were performed to predict the binding of the four reported Bcl-2 inhibitors to AspH. The results indicate that binding of the four small-molecules to defined AspH residues is feasible, at least in principle (Supporting Figs. S5-S8).

Discussion
Small-molecule AspH probes for use in vivo should possess drug-like properties 78 to facilitate cell-based biochemical as well as animal model studies. A straight-forward way of identifying drug-like AspH probes is to investigate small-molecule APIs of approved human therapeutics for AspH inhibition, as drug repositioning constitutes a successful general strategy to develop novel enzyme inhibitors. 31,33 Therefore, we investigated the inhibitory effect of 316 small-molecules, which are either APIs of approved human cancer therapeutics or of human cancer therapeutics under current or previous clinical investigation, against AspH using our previously reported high-throughput SPE-MS AspH inhibition assay (Supporting Fig. S2 and Supporting Data Sheet). 36 The focus was on APIs of cancer therapeutics as the identification of potential side-effects of these APIs through inhibiting AspH could render alternative explanations of their biological effects.
The high-throughput screen revealed that eleven small-molecule APIs, categorised into reported inhibitors of Bcl-2 family proteins, hydroxamate-containing HDAC inhibitors, kinase inhibitors, and antitumor antibiotics, effected AspH activity (Table 1). Hydroxamates, including HDAC inhibitors, have been reported to inhibit 2OG oxygenases. [79][80][81] However, efficient AspH inhibition was only observed for the antitumor antibiotic bleomycin A 2 40-42 and for four Bcl-2 inhibitors, i.e. (R)-gossypol, 37-39 venetoclax, 43,44 navitoclax, 49 and ABT-737, 55 while the remaining identified AspH inhibitors displayed only moderate potency (Table 1). These five small-molecules efficiently inhibited AspH at different 2OG assay concentrations (3 μM and 200 μM 2OG; Table 2) which are in the range of reported physiological 2OG levels in human plasma (9-12 μM 2OG) 82 and cells (> 1mM 2OG). [83][84][85] Although bleomycin A 2 efficiently inhibits AspH (IC 50 ~ 1.5 μM; Table 2, entry 3), its precise mechanism of inhibition requires further work. Some evidence suggests that bleomycin A 2 aggregates in solution (Fig. 2), which might cause AspH inhibition. Alternatively, reactive oxygen species (ROS) generated by bleomycin A 2 might account for its observed inhibitory effect. AspH activity has been previously shown to decrease in the presence of ROS 30 and bleomycin A 2 is reported to form ROS in the presence of Fe(II) and O 2 . 42,[74][75] This proposal for the mode of action of bleomycin A 2 in inhibiting AspH is consistent with the result that bleomycin A 2 does not seem to bind AspH efficiently (ΔT m ~ −0.8 °C; Supporting Fig. S4). This proposal would potentially complement previously determined mechanisms how natural products inhibit AspH, i.e. 2OG-competition for binding the AspH active site (e.g. N-oxalylglycine, 28,36,86 pyridine-2,3-dicarboxylic acid 21,87 ) and Fe(II)chelation (e.g. L-mimosine 36,88 ). Studies on the inhibition of the activities of other 2OG oxygenases by bleomycin A 2 , both in cells and with   Out of the four Bcl-2 inhibitors found to inhibit AspH, the (R)-enantiomer of the natural product gossypol (AT-101; Table 1, entry 2), [37][38][39] which was originally obtained from cottonseed extracts and which has been investigated as a male contraceptive agent, 94,95 inhibits AspH with similar efficiency (IC 50 (AspH) ~ 0.3 μM; Table 2, entry 2) as its validated Bcl-2 target enzymes, i.e. Bcl-2, Bcl-X L , and Bcl-w (IC 50 (Bcl-X L ) ~ 0.5 μM). 37 However, care should be taken when comparing the absolute IC 50 -values as these were obtained using different assay techniques, i.e. SPE-MS for AspH (50 nM enzyme concentration) and fluorescence polarisation for Bcl-X L (30 nM enzyme concentration). 37 Although molecular docking studies suggest that this natural product may inhibit AspH by directly binding to the enzyme (Supporting Figs. S5 and S8), it has been reported that gossypol induces the formation of reactive oxygen species, which, as for bleomycin A 2 , might also account for some of its observed bioactivity against AspH. 96 The clinical use of gossypol is associated with numerous side-effects, which may have hampered its progression from therapeutic studies into clinical use as an anticancer agent and suggest that this natural product is a nonspecific inhibitor acting against different enzyme-classes apart from Bcl-2 proteins. 97 AspH selectivity counter-screens could thus be performed in order to develop gossypol-derived small-molecule Bcl-2 inhibitors, such as TW-37 98 and sabutoclax 99,100 , into safer medicines.
The Abbott-developed small-molecule Bcl-2 inhibitors venetoclax 43,44 (IC 50 ~ 1.4 μM; Table 2, entry 4) and navitoclax 49 (IC 50 ~ 1.0 μM; Table 2, entry 8) were efficient AspH inhibitors, while ABT-737 55 was slightly less potent (IC 50 ~ 3.4 μM; Table 2, entry 12). Biophysical binding assays and molecular docking studies indicate that the mechanism by which the three small-molecule Bcl-2 inhibitors impede AspH activity may involve binding to the enzyme (Supporting Figs. S4 and S6-S8), which could for example occur in an allosteric manner. Optimizing the interactions of the Abbott-developed smallmolecules with AspH while minimizing their reported interactions with the Bcl-2 family proteins should foster the design of AspH-selective inhibitors as part of a detailed structure-activity relationship study.
The results are also of relevance to the clinical use of venetoclax, 43,44 which is an approved API for the treatment of chronic lymphocytic leukaemia and small lymphocytic lymphoma. 58 AspH inhibition could to some extent account for the observed anti-cancer effects of venetoclax, even though this API inhibits its validated target enzyme, Bcl-2, significantly more efficient than AspH (~1000 fold selectivity). 34 Hill coefficient 57 analysis of AspH inhibition curves suggests that venetoclax might aggregate in solution and form colloidal particles, which could account for its observed inhibitory effect (Table 2). 62 This raises the question if, and to what extent, this small-molecule API approved for clinical applications may form aggregates in buffered aqueous solutions or human body fluids. It has been shown that other APIs of approved therapeutics form colloidal aggregates in a fluid-dependent manner, which can even be stable in serum over a prolonged period of time. 68,101 Aggregate formation can result in reduced cellular API uptake, thus compromising inhibitor efficiency. 68,[102][103][104][105] AspH levels are upregulated on the cell surface of cancer cells, [10][11][12] while, in order to induce cellular apoptosis through the mitochondrial pathway by inhibiting Bcl-2, venetoclax has to penetrate the cell wall. 34,35 An accumulation of venetoclax outside the cell might compensate for its comparably lower efficiency to inhibit AspH.

Conclusions
The present study validates the robustness of SPE-MS AspH inhibition assays in combination with biophysical techniques to identify druglike small-molecules as AspH probes and novel structures for inhibitor development. Natural products and approved APIs for human therapy have been identified that inhibit AspH. Although further work is required to investigate their mode of action, the observation of AspH inhibition by bleomycin A 2 , venetoclax, and (R)-gossypol (AT-101) could be of utility in the design of selective small-molecule AspH inhibitors for use in investigating its biological roles and its utility as a medicinal chemistry target. Studies on the role of bleomycin A 2 and related drugs on 2OG oxygenases involved in DNA damage repair are of interest.

AspH substrate synthesis
The synthetic thioether linked cyclic peptide hFX-CP 101-119 28 (Supporting Fig. S1b) was used as an AspH substrate; it was designed based on 19 EGFD1 amino acid residues of the sequence of human coagulation factor X (hFX amino acids 101-119), which is a reported cellular AspH substrate. 106 hFX-CP 101-119 was synthesized as reported by an intramolecular thioetherification cyclization reaction from the corresponding linear peptide (D-Ala replacing Cys101 hFX and Ser replacing Cys112 hFX ); 30 it was prepared with a C-terminal amide.

AspH inhibition assays
AspH inhibition assays were performed at 2OG, Fe(II), and hFX-CP 101-119 concentrations close to the relevant K m -values as described. 36 For the most active AspH inhibitors identified, inhibition assays were also performed at elevated 2OG (330·K m ; K m = 0.6 μM 30 ), substrate peptide (hFX-CP 101 Solutions of the bioactive small-molecules (100% DMSO) were dry dispensed across 384-well polypropylene assay plates (Greiner) using an ECHO 550 acoustic dispenser (Labcyte). DMSO and 2,4-PDCA 19,20,36 were used as negative and positive inhibition controls, respectively. The aqueous DMSO concentration was kept constant at 0.5% v/v throughout all experiments (using the DMSO backfill option of the acoustic dispenser). Initial screening of the cancer API compound library was performed at a fixed compound concentration (20 μM). For dose-response experiments, the AspH inhibitors were dry dispensed in a 3-fold and 11-point dilution series using an acoustic dispenser (100 μM top concentration). 2,4-PDCA was used as a positive control; 36 its IC 50value was determined on each assay plate to confirm the assay quality. Each reaction was performed in technical duplicates in adjacent wells on the assay plates; additionally, assays were performed in two independent duplicates on different days using fresh reagents. An Enzyme Mixture (25 μL/well), containing 0.1 μM His 6 -AspH 315-758 in 50 mM HEPES buffer (pH 7.5), was dispensed across the inhibitor-containing 384-well assay plates with a multidrop dispenser (ThermoFischer Scientific) at 20 °C under ambient atmosphere. The plates were subsequently centrifuged (1000 rpm, 30 s) and incubated for 15 min at 20 °C. A Substrate Mixture (25 μL/well), containing 2.0 μM hFX-CP 101-119 , 200 μM LAA, 6.0 μM 2OG, and 4.0 μM FAS in 50 mM HEPES buffer (pH 7.5), was added using the multidrop dispenser. Note: The multidrop dispenser ensured proper mixing of the Enzyme and the Substrate Mixtures which was essential for assay reproducibility. The plates were centrifuged (1000 rpm, 30 s) and after incubating for 7 min, the enzyme reaction was stopped by the addition of 10% v/v aqueous formic acid (5 μL/well). The plates were centrifuged (1000 rpm, 60 s) and analysed by MS.
AspH inhibition assays at elevated 2OG (330·K m : 400 μM 2OG in the Substrate Mixture), substrate peptide (8·K m : 20 μM hFX-CP 101-119 and 20 μM 2OG in the Substrate Mixture) or Fe(II) (14·K m : 40 μM FAS in the Substrate Mixture) assay concentrations were performed in a similar manner. However, the enzyme reaction was stopped after incubating for 6 rather than 7 min when the reaction was performed at an elevated 2OG or Fe(II) concentration, and after incubating for 35 min when the reaction was performed at an elevated AspH substrate peptide concentration.
MS-analyses were performed using a RapidFire RF 365 highthroughput sampling robot (Agilent) attached to an iFunnel Agilent 6550 accurate mass quadrupole time-of-flight (Q-TOF) mass spectrometer operated in the positive ionization mode. Assay samples were aspirated under vacuum for 0.4 s and loaded onto a C4 SPE cartridge. After loading, the C4 SPE cartridge was washed with 0.1% v/v aqueous formic acid to remove non-volatile buffer salts (5 s, 1.5 mL/min). The peptide was eluted from the SPE cartridge with 0.1% v/v aqueous formic acid in 85/15 v/v acetonitrile/water into the mass spectrometer (5 s, 1.25 mL/min) and the SPE cartridge re-equilibrated with 0.1% v/v aqueous formic acid (1 s, 1.25 mL/min). The mass spectrometer parameters were: capillary voltage (4000 V), nozzle voltage (1000 V), fragmentor voltage (365 V), gas temperature (280 °C), gas flow (13 L/ min), sheath gas temperature (350 °C), sheath gas flow (12 L/min). The m/z + 2 charge states of the cyclic peptide (substrate) and the hydroxylated cyclic peptide (product) were used to extract ion chromatogram data, peak areas were integrated using RapidFire Integrator software (Agilent). The data were exported into Microsoft Excel and used to calculate the % conversion of the hydroxylation reaction using the equation: % conversion = 100 × (integral product cyclic peptide)/ (integral substrate cyclic peptide + integral product cyclic peptide). Normalized dose-response curves (2,4-PDCA and DMSO controls) were obtained from the raw data by non-linear regression (GraphPad Prism 5) and used to determine IC 50 -values. The standard deviation (SD) of two independent IC 50 determinations (n = 2) was calculated using GraphPad Prism 5. Z'-factors and S/N ratios were calculated according to the cited literature using Microsoft Excel (Supporting Fig. S3). 56

AspH differential scanning fluorimetry (DSF) assays
A mixture of SYPRO orange (1‰ v/v ) and His 6 -AspH 315-758 (4 μM) in 50 mM aqueous HEPES buffer (pH 7.5, 150 mM NaCl, 50 μM NiCl 2 ) was carefully pipetted into a 96-well VWR PCR-plate (38 μL per well). An inhibitor solution (0.8 mM in DMSO) or a negative control sample (pure DMSO) was added to each well (2 μL per well, final inhibitor concentration of 40 μM) and the resulting solutions were gently mixed using a pipette. The plate was sealed with optical tape (Bio-Rag, iCycler iQ), centrifuged (5 sec, 1000 rpm) and placed into a C1000Touch Thermal Cycler equipped with a CFX96 Real-Time System (Bio-Rad). The instrument was heated with a rate of 1 °C per cycle (starting at an initial temperature of 25 °C; 70 cycles total). Data were analyzed using Microsoft Excel and GraphPad Prism according to the literature. 107 DSF assays were performed in independent duplicates, each composed of technical duplicates (Supporting Fig. S4).

Molecular docking studies
Docking studies were performed using four Bcl-2 inhibitors as potential AspH ligands: (R)-gossypol (AT-101), 37-39 venetoclax (ABT-199), 43,44 navitoclax (ABT-263), 49 and ABT-737. 55 The structure files of AspH and the four ligands were prepared using the BIOVIA Discovery Studio 2019 (Dassault Systèmes BIOVIA) software suite; the protein structure was based on a deposited AspH crystal structure (PDB ID: 5JZA) 28 for which all water molecules and the N-oxalylglycine (NOG) ligand binding the AspH active site were deleted. AspH residues around the Mn(II) atom (radius = 20 Å) were defined as possible binding sites for ligand docking. Molecular docking studies were then performed by docking the Bcl-2 inhibitors into the defined AspH binding sites using the protein ligand docking software GOLD 5.1 (Cambridge Structural Database) 108 with a default setting of 100 genetic algorithm (GA) runs for each ligand. For each GA run, a maximum of 125,000 operations was performed; when the top ten solutions possessed root-mean-square deviation (RMSD) values within 1.5 Å, the docking process was terminated. ChemPLP 109 was used as scoring function. The docking pose for each ligand was analyzed using PyMOL (Supporting Figs. S5-S8).

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.