Materials
Recombinant human LDHA and LDHB were purchased from R&D (Minneapolis, MN, USA). NAD+ was purchased from MP Biomedicals (Santa Ana, CA, USA). DTT, NaCl, and Triton X-100 were obtained from Sigma–Aldrich (St. Louis, MO, USA). Dihexylammonium acetate (DHAA) was obtained from Tokyo Kasei Kogyo (Tokyo, Japan). NADH, Pyruvate, bovine serum albumin (BSA), and other reagents were purchased from Fujifilm Wako (Osaka, Japan).
AXKO-0046 was synthesised by the Takeda Pharmaceutical Co. Ltd., whereas AXKO-0058, AXKO-0060, AXKO-0067, and AXKO-0077 were synthesised by Axcelead Drug Discovery Partners, Inc.
LDHB and LDHA Assays for Primary Screening
We performed a screening campaign for compounds (final concentration: 10 μM) within pooled libraries consisting of more than 345,000 compounds. Each compound in DMSO (1 mM) was dispensed to 384-well V base plates (100 nl per well) using an Echo 555 acoustic dispenser (Labcyte, Sunnyvale, CA, USA). LDHB (5 μl was diluted in the assay buffer (20 mM Tris at pH 7.5, containing 2 mM DTT, 0.005% Triton X100, and 0.005% BSA) and dispensed to the assay plates (final concentration: 0.25 nM). The plates were centrifuged (200×g for 1 min), incubated at room temperature for 30 min, and added with 5 μl of NADH and pyruvate in the assay buffer (final concentration: 75 μM). After incubation for 30 min at room temperature, the plates were added with acetonitrile (10 μl) and water (50 μl) to quench the reaction. All additions were performed using a Multidrop Combi (Thermo Fisher Scientific, Waltham, MA, USA). Assay plates were sealed and centrifuged (2,700×g for 10 min) prior to storage at −20 °C.
The conditions for LDHA assay were as described above, except for substrate concentrations, i.e., 400 µM pyruvate and 100 µM NADH). These concentrations were chosen based on fitting to the Michaelis–Menten model with Km value (Supplementary Figure S2).
High-Throughput Mass Spectrometry Assay
For RapidFire mass assay (Agilent, Wakefield, MA, USA), the enzyme reaction solution (5 μl) was aspirated directly from the quenched assay plates and loaded onto a C8 solid-phase extraction cartridge (Agilent) with a mobile phase of water containing DHAA (5 mM) for 2,500 ms at a flow rate of 1.25 ml/min. Then, the analytes were coeluted into the mass spectrometer using water:acetonitrile:acetone (2:1:1) containing ammonium acetate (5 mM) for 5,000 ms at a flow rate of 1.0 ml/min. NADH and NAD+ were detected using multiple reaction monitoring with QI/Q3 transitions at m/z 664.2 to 397.0 and m/z 662.1 to 540.1, respectively, on Sciex API4000 triple quadrupole mass spectrometer (Applied Biosystems, Foster City, CA, USA) in the negative electrospray ionization mode. Extracted ion chromatograms for each transition were integrated and processed using RapidFire Integrator (Agilent). The data for each well were normalised using the monitoring product conversion with the ratio of AUCproduct / (AUCproduct+ AUCsubstrate).
To determine the compound activities, percent inhibition data normalised to 0% (DMSO only) and 100% inhibition (no-enzyme control) wells were calculated. Data analysis was performed using Tibco Spotfire (Boston, MA, USA) packages.
Substrate-Competition Assay
To evaluate the mechanism of inhibition of the selected compounds, their initial reaction velocities were measured using 5 concentrations of NADH or pyruvate (10, 30, 50, 100, and 200 μM) incubated with the 100 μM pyruvate for NADH titration or 100 μM NADH for pyruvate titration. Then, the solution of 7 concentrations of AXKO-0046 (0.00001, 0.0001, 0.001, 0.01, 0.1, 1, 10, and 100 μM) with LDHB of 0.25 nM were added. The enzymatic reactions were performed at room temperature for 15 min and conducted in triplicate. At each inhibitor concentration, the dependence of the initial reaction velocities on substrate concentration was fitted to nonlinear fits of Michaelis–Menten model to obtain the Km and Vmax values using GraphPad Prism. Lineweaver–Burk plot was generated by superimposing the data of 1/ [velocity] and 1/ [substrate] and the line corresponding to the Michaelis–Menten nonlinear fits.
Protein preparation for structural analysis
Human LDHB (Ala2−Leu334) (NCBI Reference Sequence: NM_001174097.2) DNA was synthesised and ligated into a pET21a vector (Merck Millipore, Darmstadt, Germany) with a N-terminal His-Avi tag, followed by the cleavage site of TEV protease using VectorBuilder (Chicago, IL, USA). The expression plasmid was transfected into E. coli BL21 (DE3) (Nippon Gene, Toyama, Japan), and the cells were grown in Lysogeny broth (LB) medium containing ampicillin (100 mg ml-1). Protein expression was induced with Isopropyl β-d-1-thiogalactopyranoside (IPTG) (0.2 mM) and cultured for 16 h at 16 °C. Harvested cells were lysed using sonication in lysis buffer (50 mM Tris, pH 8.0), NaCl (150 mM), nuclease (5 U mL-1) and centrifuged at 15,000×g for 10 min at 4 °C. The clarified supernatant was loaded onto Ni-NTA Cartridge (FUJIFILM Wako, Osaka, Japan), and the eluted fraction was purified on HiLoad 26/60 Superdex 200 pg column (GE Healthcare, Piscataway, NJ, USA). The His−Avi tag was digested using TEV protease (Sigma-Aldrich, St. Louis, MO, USA), and the digested solution was subsequently passed through a Ni-NTA column to remove TEV protease and the uncleaved protein.
The protein was further purified using anion-exchange chromatography (monoQ, GE Healthcare) with a linear gradient of NaCl. All purification steps were assessed using SDS–PAGE and Coomassie Blue staining (Sigma-Aldrich, St. Louis, MO, USA). The final yield of protein was 46 mg for every 1 litre of culture. For crystallization, the purified protein was buffer-exchanged to the final buffer (50 mM Tris, pH 7.6 with 150 mM NaCl), concentrated (20 mg ml-1) using ultrafiltration (AMICON-ULTRA 10K, Millipore, Bedford, MA, USA), and stored at −80°C.
X-ray crystallography
The complexes of LDHB with NADH or NADH, oxamate, and AXKO-0046 were generated by incubating 3-fold molar excesses of ligands on ice for 2–3 h before crystallization experiments. Both complexes were crystallised from a reservoir solution containing HEPES (0.1 M, pH 7.5), potassium formate or ammonium acetate (0.2 M), and PEG 3350 (20% v/v) at 20°C via the sitting-drop vapour diffusion method. Prior to data collection, crystals were immersed in the reservoir solution containing ethylene glycol (30%) as a cryoprotectant and flash-frozen in liquid nitrogen. Diffraction data were collected from a single crystal using the DECTRIS Pilatus3-S6M PAD detector (Baden-Daettwil, Switzerland) with a BL45XU beamline (SPring-8, Hyogo, Japan) under a 100 K nitrogen cryostream. The diffraction data were reduced and scaled using HKL200028.
The structure was solved according to the molecular replacement method using Phaser29 from the CCP4 software suite30 and the LDHB structure (PDB code 1I0Z18) as a search model. Refinement was performed using REFMAC531 with individual isotropic restrained B factors. Some data (5%) were set aside for cross-validation before refinement, and progress was monitored using Rfree. For TLS refinement, the tetramer of the protein and ligands was set as a single rigid body32. Interactive model building was performed using COOT33. The final models were validated using Molprobity34. All graphical figures were generated using PyMOL (Schrödinger LLC, Cambridge, MA, USA). Crystallographic processing and refinement statistics are summarised in Supplementary Table S1.
Data Analysis
For the kinetic parameters, the initial rates of LDH activity were determined by incubating LDHA or LDHB with various concentrations of NADH and pyruvate in the assay buffer at room temperature. The reaction product was measured using RapidFire mass assay. To estimate the Km and Vmax values, the initial rates were fit to the Michaelis–Menten equation. IC50 and Hill slope calculation and curve fitting using four-parameter fits were performed using GraphPad Prism v6.07 (GraphPad Software, San Diego, CA) or XLfit (IDBS, Guildford, UK) with nonlinear regression analysis, wherein the IC50 value equals the concentration at which the inflection point of the fitted model is reached. The quality and robustness of the screening campaign were determined by analysing the Z’ factor.