2-Oxoglutarate regulates binding of hydroxylated hypoxia-inducible factor to prolyl hydroxylase domain 2† †Electronic supplementary information (ESI) available: Details of experimental procedures and additional experiments. See DOI: 10.1039/c8cc00387d

The binding of prolyl-hydroxylated HIF-α to PHD2 is hindered by prior 2OG binding; likely, leading to the inhibition of HIF-α degradation under limiting 2OG conditions.


Non-Denaturing MS Experiments
For non-denaturing MS measurements, spectra were obtained using a Waters Q-TOF

NMR Experiments
Nuclear Magnetic Resonance (NMR) spectra were recorded using a Bruker AVIII 700 MHz NMR spectrometer equipped with a 5-mm inverse cryoprobe using 5 mm diameter NMR tubes (Norell) or 3 mm MATCH NMR tubes (Cortectnet). Data were processed with Bruker 3.1 software.

C-2OG and C-NODD/CODD Displacement Experiments
The CLIP-HSQC sequence was used for 1D 13 C HSQC experiments (without 13 C decoupling). 12 The relaxation delay was 2 s and the 1 J CH was set to 160 Hz. A 6.8 ms Q3 180degree pulse was used for selective 13 15 complexes as templates using COOT. 16 All models were conjugate energy minimized to ensure their geometric quality without applying any external energy term.

MD Simulation Studies
Two structures of tPHD2.Mn.CODD were studied; one with Hyp564 in the C4 cis conformation and one with Hyp564 in the C4 trans conformation. For each conformation, two different co-substrate/product binding modes were considered, i.e. bidentate binding with 2-oxoglutarate (2OG) and monodentate binding with succinate. In total, four systems were prepared, as indicated in Table S1. The protein was solvated using the Solvate plug-in of VMD, 17 resulting in a box size of (82x82x82) Å 3 . The system was ionised to be electrically neutral using a 0.15 mM background ionic concentration of NaCl. Protonation states were chosen corresponding to pH 7.0, with δ-His313 and δ-His374 protonation states in the active site.
For each system, we first tested the validity of the initial PDB derived model of tPHD2, using a substrate conformation obtained based on semi-automated (knowledge-based) docking. The initial MD simulation system setup was based on the structure of tPHD2.Mn.CODD.2OG 11 with Mn(II) exchanged for Fe(II). The MD simulation (10 ns) was performed for the initial modelled complex with backbone restraints of tPHD2 chain A, as well as the active site residues H313, D315, H374, 2OG or succinate, Fe(II) replacing the Mn(II) density, and H564 from the substrate peptide. Following this, MD production runs in the NPT ensemble were performed for 40 ns for each system, totalling 200 ns, keeping only the active site restrained, using a system temperature of 300 K and 1 bar pressure.
The protein conformation was simulated using the CHARMM 36 force-field (FF), with the CMAP correction; 18 a TIP3P water model was employed. 19 CHARMM FF parameters for the 2OG ligand compatible with version 2b7 of the CHARMM General Force-Field (CGenFF) 20-23 were generated and validated using the Force-Field toolkit plugin in VMD, 24 after initial atom typing and assignment of parameters and charges using version 0.9.7 of the CGenFF program at paramchem.org. 21,23
Details of the systems studied are summarized in     -tPHD2 and (B)    analyses imply the potential for a clash is the greatest with 2OG and the C4-exo hyCODD prolyl conformation, consistent with the experimental results ( Fig. 2 and S6)