2-Oxoglutarate derivatives can selectively enhance or inhibit the activity of human oxygenases

2-Oxoglutarate (2OG) oxygenases are validated agrochemical and human drug targets. The potential for modulating their activity with 2OG derivatives has not been explored, possibly due to concerns regarding selectivity. We report proof-of-principle studies demonstrating selective enhancement or inhibition of 2OG oxygenase activity by 2-oxo acids. The human 2OG oxygenases studied, factor inhibiting hypoxia-inducible transcription factor HIF-α (FIH) and aspartate/asparagine-β-hydroxylase (AspH), catalyze C3 hydroxylations of Asp/Asn-residues. Of 35 tested 2OG derivatives, 10 enhance and 17 inhibit FIH activity. Comparison with results for AspH reveals that 2OG derivatives selectively enhance or inhibit FIH or AspH. Comparison of FIH structures complexed with 2OG derivatives to those for AspH provides insight into the basis of the observed selectivity. 2-Oxo acid derivatives have potential as drugs, for use in biomimetic catalysis, and in functional studies. The results suggest that the in vivo activity of 2OG oxygenases may be regulated by natural 2-oxo acids other than 2OG.


Supplementary Figures
Supplementary Figure 1. Synthesis of the 2OG derivatives used in this work. The C3/C4-substituted 2OG derivatives (S5) were prepared according to a reported strategy employing cyanosulfur ylids S3 as intermediates 1 . Cyanosulfur ylids S3 were prepared by the dehydrative coupling of mono-methyl dicarboxylic acid half-esters (S1) with the reported tetrahydrothiophene salt S2 2 . The cyanosulfur ylids (S3) were oxidized using oxone to afford the corresponding dimethyl dicarboxylic acid esters (S4). C3/C4-substituted 2OG derivatives (S5) were obtained via lithium hydroxide-mediated saponification of dimethyl esters S4, in sufficient purity to be employed in biochemical and crystallographic experiments. Note that 2OG derivatives 11 and 22 have not previously been reported.

Supplementary Figure 5. Hydroxylation rates of FIH-catalyzed HIF-1α788-822 hydroxylations used to determine kinetic parameters for 2OG derivatives (continues on the following page). Maximum velocities
(v max app ) and Michaelis constants (K m app ) of FIH were determined in independent triplicates for 2OG and the 2OG derivatives 1, 12, 14, and 22, monitoring FIH-catalyzed hydroxylation of the HIF-1α788-822 4 substrate peptide by SPE-MS as described in the Methods section. Conditions: 0.15 μM FIH, 5.0 μM HIF-1α788-822 4 , 100 μM L-ascorbic acid (LAA), and 20 μM ammonium iron(II) sulfate hexahydrate (FAS, (NH4)2Fe(SO4)2·6H2O) in buffer (50 mM Tris, 50 mM NaCl, pH 7.5, 20° C). Measurement times were normalized to the first sample injection analyzed after the addition of FIH to the Substrate Mixture (t = 0 s), by which time low levels of hydroxylation were manifest. Data are shown as the mean of three independent runs (n = 3; mean ± standard deviation, SD). Source data are provided as a Source Data file.  (NOFD 5 ). SPE-MS was used to monitor the hydroxylation of the HIF-1α788-822 peptide 4 in buffer (50 mM Tris, 50 mM NaCl, pH 7.5, 20 °C) using identical MS configurations as described in the Methods section. Data are shown as the mean of three independent runs (n = 3; mean ± standard deviation, SD (a) Overview of the FIH:1 crystal structure; (b) representative OMIT electron density map (mFo-DFc) contoured to 3σ around (S)-1 of the FIH:1 structure. (S)-1 coordinates to the Zn ion in a bidentate manner and is positioned to interact with the sidechains of FIH active site residues Lys214, Tyr145, Thr196, Asn294, and Asn205 (distances in Å); (c) superimposition of a view from the FIH:1 structure with one from the reported FIH:2OG structure (FIH: pale green, carbon-backbone of 2OG: green, Fe: orange; PDB ID: 1H2N) 9 reveals similar FIH conformations (Cα RMSD = 0.17 Å).
(a) Representative OMIT electron density map (mFo-DFc) contoured to 3σ around (S)-1 modelled in the FIH:1 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around (S)-1 are shown in green and red mesh, respectively; (b) representative OMIT electron density map (mFo-DFc) contoured to 3σ around (R)-1 modelled in the FIH:(R)-1 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around (R)-1 are shown in green and red mesh, respectively. Even though the experimental data do not support the predominant presence of either enantiomer in the FIH:1 structure, (S)-1 was selected for the structure refinement by analogy to the FIH:1:CA20 and FIH:1:TANK2691-710 structures, which show clear evidence for the predominant (at least) presence of (S)-1 in the structures ( Supplementary Figures 14 and 16). This proposal is consistent with the binding mode of N-(carboxycarbonyl)-D-phenylalanine (NOFD) to FIH 5 ; Note, that the nitrogen atom of NOFD affects the Cahn-Ingold-Prelog priority rules resulting in its formal assignment as (R)-enantiomer. OMIT electron density map (mFo-DFc) contoured to 3σ around (R)-3 modelled in the FIH:(R)-3 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around 3σ around (R)-3 are shown in green and red mesh, respectively. Negative densities were observed at the -2σ level around the C3 atom and the alkyl substituent of (R)-3; these negative densities were not observed at the -2σ level for (S)-3 as shown in panel (a). The superimposed image shows that the C3 atom and the alkyl substituent of (S)-3 positioned to compensate for the negative densities observed for (R)-3 in the putative FIH:(R)-3 structure. Thus, the electron density analysis indicates that the (S)-enantiomer of 3 is predominantly (at least) present in the FIH:3 structure. This proposal is consistent with the binding mode of N-(carboxycarbonyl)-D-phenylalanine (NOFD) to FIH 5 ; Note, that the nitrogen atom of NOFD affects the Cahn-Ingold-Prelog priority rules resulting in its formal assignment as (R)enantiomer. Color code: FIH: grey; carbon-backbone of (S)-3-methyl-2OG ((S)-1): yellow; carbon-backbone of (R)-3-methyl-2OG ((R)-1): deep green; Zn: lavender blue; water: red sphere; oxygen: red; nitrogen: blue. Distances are in Å.

Supplementary
(a) Representative OMIT electron density map (mFo-DFc) contoured to 3σ around (S)-1 modelled in the FIH:1:CA1-20 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around (S)-1 are shown in green and red mesh, respectively; (b) representative OMIT electron density map (mFo-DFc) contoured to 3σ around (R)-1 modelled in the FIH:1:CA1-20 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around (R)-1 are shown in green and red mesh, respectively. The methyl carbon atom of (R)-1 is proximate to the Leu188 sidechain (2.9 Å); the van der Waals overlap is ~0.6 Å. This clash with the Leu188 sidechain supports the predominant presence of (S)-enantiomer of 1 in the FIH:1:CA1-20 structure, as both C6 and C3 carbon atoms of (S)-1 are more than 4.1 Å distant from the Leu188 sidechain. Moreover, the electron density analysis in the FIH:1:TANK2691-710 structure supports the predominant binding of (S)-1 in the FIH structures (Supplementary Excel.  (14) in buffer (50 mM Tris, 50 mM NaCl, pH 7.5, 20° C). Measurement times were normalized to the first sample injection analyzed after the addition of the 2OG oxygenases to the Substrate Mixture (t = 0 s), by which time low levels of hydroxylation were manifest. 14 is a reported inhibitor of human AspH (Supplementary  Table 1) 1 , and a cosubstrate for FIH (Table 1).

Supplementary Figure 27. 4-Ethyl-2OG (14) selectively inhibits AspH in the presence of FIH in a dose
Varying the concentration of 14 in the presence of 100 μM 2OG reveals that 14 selectively inhibits AspH (black circles) in a dose dependent manner, while no substantial inhibitory effect on FIH activity (orange squares) was Representative OMIT electron density map (mFo-DFc) contoured to 3σ around (R)-18 modelled in the FIH:18 structure (blue mesh). Electron density maps (Fo-Fc) contoured to 2σ and -2σ around (R)-18 are shown in green and red mesh, respectively. Negative density at the -2σ level was observed around the C4 and C6 atoms of (R)-18 modelled in the FIH:18 structure; negative density for these atoms of (S)-18 was not observed at the -2σ level as shown in panel (a). The superimposed image shows that the C4 and C6 atoms of (S)-18 are positioned to avoid the negative densities observed for (R)-18 in the putative FIH:(R)-18 structure. Thus, the analysis of the electron density maps indicates that the (S)-enantiomer of 18 is predominantly (at least) present in the FIH:18 structure.
This proposal is consistent with the binding mode of N-(carboxycarbonyl)-D-phenylalanine (NOFD) to FIH 5 ; Note, that the nitrogen atom of NOFD affects the Cahn-Ingold-Prelog priority rules resulting in its formal assignment as the (R)-enantiomer. NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Figure 35. 1 H and 13 C NMR spectra of compound 38.
NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Figure 36. 1 H and 13 C NMR spectra of compound 39.
NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1. Figure 37. 1 H and 13 C NMR spectra of compound 11. NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Figure 38. 1 H and 13 C NMR spectra of compound 40.
NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Figure 39. 1 H and 13 C NMR spectra of compound 41.
NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Figure 40. 1 H and 13 C NMR spectra of compound 22.
NMR spectra were recorded in CDCl3 using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1). Spectra were analyzed and processed using Bruker TopSpin 3.6.1.

Supplementary Tables
Supplementary Table 1 AspH and 50 mM Tris, 50 mM NaCl, pH 7.5 for FIH), which might, at least in part, reflect the observed differences of the synthetic 2OG derivatives to inhibit the two 2OG oxygenases. Low levels of selective FIH inhibition was observed for 3-propyl-2OG (3), which appears to inhibit AspH with twofold reduced potency than FIH (Entry 4).

General information
Unless otherwise stated, all reagents were from commercial sources (Sigma-Aldrich, Inc.; Fluorochem Ltd; Alfa Aesar). Oxone, obtained from Alfa Aesar, was used to oxidize cyanosulfur ylids according to a literature protocol 2 .
With the exception of 2OG derivatives 11 and 22, all 2OG derivatives used in this study were synthesized according to a reported procedure 1 . All chiral compounds were prepared as racemates, except for the reported FIH inhibitor N-oxalyl-D-phenylalanine 5 (NOFD), which was prepared as a single enantiomer 14 . Anhydrous solvents were from Sigma-Aldrich, Inc. and kept under an atmosphere of nitrogen. Solvents, liquids, and solutions were transferred using nitrogen-flushed stainless steel needles and syringes. All reactions were carried out under an atmosphere of nitrogen unless stated otherwise. Milli-Q ® Ultrapure (MQ-grade) water was used for buffers; LCMS grade solvents (Merck) were used for solid phase extraction coupled to mass spectrometry (SPE-MS).
Purifications were performed using an automated Biotage Isolera One purification machine ( Tensor-27 Fourier transform infrared (FT-IR) spectrometer. High-resolution mass spectrometry (HRMS) was performed using electro-spray ionization (ESI) mass spectrometry (MS) in the positive or negative ionization modes employing a Thermo Scientific Exactive mass spectrometer (ThermoFisher Scientific); data are presented as a mass-to-charge ratio (m/z).
Nuclear magnetic resonance (NMR) spectroscopy was performed using a Bruker AVANCE AVIIIHD 600 machine equipped with a 5 mm BB-F/1H Prodigy N2 cryoprobe operated using Bruker TopSpin software (version 3.6.1).
Spectra were analyzed and processed using Bruker TopSpin 3.6. in brackets indicate close signals that can be differentiated considering second respectively third decimal numbers.

General Procedure A
To a solution of mono-methyl dicarboxylic acid half-ester S1 (1.0 equiv.) and 1-(cyanomethyl)tetrahydro-1Hthiophen-1-ium bromide S2 2 (1.4 equiv.) in anhydrous dichloromethane (0.2 M) were sequentially added redistilled N,N-diisopropylethylamine (4.0 equiv.) and T3P (50%w/w in ethyl acetate, 1.4 equiv.) dropwise at 0° C under an atmosphere of N2 gas. The reaction mixture was stirred and allowed to slowly warm to ambient temperature overnight (10 -12 h). The reaction mixture was diluted with saturated aqueous NaHCO3 solution and extracted three times with dichloromethane. The combined organic extracts were dried over anhydrous Na2SO4, filtered, and evaporated. The crude residue was purified using column chromatography to afford sulfur ylide S3 which was used in the next reaction following General Procedure B.

General Procedure B
To a solution of the sulfur ylide S3 (1.0 equiv.) in a 2:1 mixture (0.08 M final concentration) of methanol (HPLC grade) and Milli-Q ® ultrapure water was added Oxone (2.0 equiv., obtained from Alfa Aesar as recommended by Bode et al. 2 ) under an ambient atmosphere at 0° C. The reaction mixture was stirred vigorously at the same temperature for 5 min, then at ambient temperature for 1 h. It was then filtered and the methanol removed under reduced pressure. The remaining aqueous solution was carefully diluted with saturated aqueous NaHCO3 solution and the mixture extracted three times with dichloromethane. The combined organic extracts were dried over anhydrous Na2SO4, filtered, and evaporated. The crude residue was purified using column chromatography to afford dimethyl dicarboxylic acid ester S4 which was used in the next reaction following General Procedure C.

General Procedure C
To a solution of dimethyl dicarboxylic acid ester S4 (1.0 equiv.) in methanol (0.2 M, HPLC grade) was added an aqueous solution of lithium hydroxide (0.4 M, 2.8 equiv.) under an ambient atmosphere at 0° C. The reaction mixture was allowed to slowly warm to ambient temperature overnight (14 -18 h). The methanol was then removed under reduced pressure. The aqueous reaction mixture was extracted three times with dichloromethane (the organic extracts were discarded) and the aqueous phase was acidified (pH ≈ 7.0 to 7.7) using Dowex ® 50XW8 (H + -form, mesh 200-400). The mixture was filtered and lyophilized to afford the solid dicarboxylic acid S5. The crude product was sufficiently pure as judged by 1 H and 13 C NMR for use in biological assays. pKa-values for the 2OG derivatives were not determined, thus, some might have actually been isolated as the corresponding monoor dilithium salts. The chemical shift data of the 2OG derivatives are likely pH-dependent. In the 1 H NMR spectrum of 2OG derivative 22 (recorded in D2O), evidence for H/D-exchange at the methylene group α to the ketone was observed.