Optimization of a “bump-and-hole” approach to allele-selective BET bromodomain inhibition† †Electronic supplementary information (ESI) available: Supplemental results (Fig. S1–S13 and Tables S1–S10). Crystallography data collection and refinement statistics. Experimental section. Detailed compound synthesis and characterisation. Supplemental references. See DOI: 10.1039/c7sc02536j

Allele-specific chemical genetics enables selective inhibition within families of highly-conserved proteins.


Figure S1 -BET Bromodomain Sequence alignment
Sequence alignment of the eight human BET bromodomains, with positions of α-helices and associated loops. Conserved residues highlighted in green, with mutated leucine in red. Conserved residues making direct contact with I-BET shown with black dot. Ruler numbering based on BRD4 BD1 sequence.

Table S2 -Effect of L/V Mutation on BRD:Peptide Affinity
Results of ITC titrations of di-and tetra-acetylated H4 peptides titrated into WT and L/V BET bromodomains.

Figure S2 -Effect of L/V Mutation on Binding Energies of BRD:Peptide Interactions
Thermodynamic measurements of tetra-acetylated H4 peptide titrated into WT and mutant BET bromodomains.

Figure S3 -BLI Screen of WT & L/V BET Bromodomains against Acetylated Histone Library
BLI binding profiles of WT and L/V BET bromodomains tested against acetylated histone peptide library. Average response of two biocytin negative controls subtracted, to account for non-specific binding. Responses color-coded to maximum and minimum response of each construct.

Figure S4 -Crystallographic Analysis of ZA Loop during Binding
Aligned X-ray crystal structures of BRD2 BD2 L/A apo (green) (4QEU) and bound to ET (brown) (4QEW), as well as BRD2 BD2 WT bound to an acetylated H4 peptide (blue) (2E3K). Bound peptide not shown. Mutated leucine/alanine side-chain highlighted.

S7
Bio-JQ1  Results are means of three consistent ITC titrations of ligand into WT and L/V BET bromodomains. For some weak interactions N was fixed to 1 S10 Figure S9 -Effects of SAHA treatment on FRAP assay window.
Recovery times of GFP-labelled BRD4 WT/WT in FRAP assay, following 0.5s laser bleach event, at a range of SAHA concentrations. Each bar is mean and SE of ~30 U2OS cells.

-Cloning & Mutagenesis
Plasmids pNIC28-Bsa4 KanR containing the 8 single BET bromodomain constructs BRD2 BD1 (protein start and stop positions K71-K176), BRD2 BD2 G344-D455, BRD3 BD1 P24-E144, BRD3 BD2 G306-P416, BRD4 BD1 N44-E168, BRD4 BD2 K333-E460, BRDT BD1 N21-E137 and BRDT BD2 S257-M361 were provided by the Oxford Structural Genomics Consortium (SGC) for expression with an N-terminal His6-tag and a TEV protease cleavage site (UniProt accession number, BRD2: P25440; BRD3: Q15059; BRD4: O60885). For tandem construct of BRD2, cDNA encoding both bromodomains and the linker region (G73-D455) was cloned from full-length cDNA clone purchased from DNASU Plasmid Repository at the Arizona State University into a pCri (11b)  Single point mutations were introduced using QuickChange II Site directed Mutagenesis Kit (Agilent). Primers were designed following the recommendations from manufacturer. The polymerase chain reaction (PCR) was performed on a 2720 Thermal Cycler (Applied Biosystems®). Upon digestion of the parental DNA strands by DpnI restriction enzyme, the PCR product was transformed into competent E. coli DH5α cells and grown on lysogeny broth (LB) agar plates containing corresponding selection antibiotics at 37°C for 16h. Single colonies were then picked from the agar plates and grown for 12h in 10 mL of LB medium containing selection antibiotics. Extracted and purified plasmid DNA were then sequenced to confirm the presence of the desired mutation.

-Protein Expression & Purification
BL21 E. coli cells were transformed with bromodomain-expressing pNIC plasmids using heat shock. Cell cultures were grown at 37°C and 200 rpm in LB media with 50 µg/ml kanamycin. A starter culture was incubated until saturation, then diluted 100-fold in fresh media and grown until reaching an optical density of 2.0 (OD600). The temperature of the culture was decreased to 18°C, and protein expression was induced overnight with 0.4 mM IPTG. Cells were harvested the next day by centrifugation (4000 rpm for 20 minutes at 4°C, JS-4.2 rotor on a Beckman J6-MC centrifuge) and stored as pellets at -20°C.
Cell pellets were resuspended in lysis buffer (50 mM HEPES, 500 mM NaCl, 10 mM imidazole & 2 mM βmercaptoethanol pH7.5) and treated with one complete protease inhibitor (Roche) tablet. Cells were lysed using a Stanstead pressure cell homogenizer and the lysate centrifuged at 20 000 rpm for 1 hour at 4°C (JA-25.50 rotor in Beckman Avanti J-25 centrifuge) and the supernatant transferred and passed through a 0.45 um filter.
The lysate was purified using metal ion affinity chromatography, with a His Trap 5ml Ni sepharose column (GE Healthcare) on an AKTApure™ system (GE Healthcare). The column was washed with 30 ml of lysis buffer, and the protein was eluted using 30 ml of elution buffer (50 mM HEPES, 500 mM NaCl, 250 mM imidazole and 2 mM βmercaptoethanol at pH 7.5). The Ni column elution was concentrated to 2 ml using a Vivaspin 20, 10 000 MWCO (Sartorius) before further purification using size exclusion chromatography. Concentrated Ni column elution was passed through a Superdex 75 16/60 Hiload column (GE healthcare) on an AKTApure™ system, using gel filtration buffer (20 mM HEPES, 150 mM NaCl and 1 mM DTT at pH 7.5). Desired fractions were pooled, concentrated and aliquoted before being flash frozen and stored at -80°C.

-Differential Scanning Fluorimetry (DSF) / Thermal Shift
DSF experiments were performed on a BioRad CFXconnect machine, using clear 96-well plates. Protein constructs were tested at 6 µM, with 2.5X SYPRO orange (Invitrogen), in 40 µl of buffer (10 mM HEPES, 100 mM NaCl, pH7.5). Samples underwent a heat cycle from 20°C to 95°C, heated at a rate of 1°C every minute. Plates were S23 read at 1 minute intervals. Each sample was tested in triplicate. Data was analyzed as recommended by Niesen et al 2 , using GraphPad Prism 6 and the SGC's DSF Analysis 3.02 spreadsheet.

-X-ray Crystallography
Purified BRD2 BD2 L/V protein at 8 mg/mL was mixed with excess amount of a bumped compound (2-4 mM) to form a protein-compound complex. Drops of the complex were mixed 2:1 with precipitant solution in sitting-drop vapour diffusion format. Crystallization condition of protein complexes varies with different bumped compounds, the condition ranges from 0.1 M Tris pH 7.5 -9.0, 45 -60 % pentaerythiol propoxylate (5/4 PO/OH), with or without 0.2 M imidazole as additive. Crystals appeared within hours and were fully grown after 2-3 d. Diffraction data were either collected at in-house Rigaku M007HF X-ray generator equipped with Varimax Cu-VHF optics using a Saturn 944HG+ CCD detector at a wavelength of 1.5418 Å or Diamond Light Source beamline I04-1 using a Pilatus 6M-F detector at a wavelength of 0.92819 Å at temperature 100K.
Indexing and integration of reflections was either performed using XDS with the XDSGUI interface 3 or MOSFLM 4 , and scaling and merging with AIMLESS 5 in CCP4i 6 . To solve the phase problem the molecular replacement method was used with the programs MOLREP 7 and Phaser 8 using search models derived from the coordinates of BRD2 BD2 WT (PDB entry 2DVV). The initial model was refined iteratively using PHENIX 9, 10 and COOT 11 . Ligand structures and restraints were generated using eLBOW 12 , REEL 13 and PRODRG 14 . The MOLPROBITY server 15 was used to validate the geometry and steric clashes in the structures. The structure models have been deposited in the protein data bank (PDB) and data collection and refinement statistics are presented in supplementary information. All figures were generated using PyMOL 16 .

-Bio-Layer Interferometry (BLI)
All BLI experiments were carried out on a OctetRed 384 instrument (ForteBio), at 25°C and in 20 mM HEPES, 150 mM NaCl pH7.5 buffer. A histone peptide library, possessing both single and multiple lysine acetylation marks, was obtained from Alta Bioscience Ltd. Peptides were 20 residues long and were biotinylated on the C-terminal (with aminohexanoic linker). Streptavidin-coated BLI tips (ForteBio) were loaded in 100 µl 5 µM peptide (or 10 µg/ml biocytin) over 10 minutes. For the assay bromodomain constructs were kept at 20 µM in 100 µl buffer in black 384well plates, agitated at 1000rpm. Each assay involved exposing peptide-loaded BLI tips to buffer for 120s to determine baseline signal, 240s in protein to measure association and finally 120s in buffer to measure dissociation. Raw data was then analyzed using the ForteBio software, to account for background signal and non-specific binding.

-Isothermal Titration Calorimetry (ITC)
ITC titrations were performed on an ITC200 instrument (MicroCal™, GE Healthcare). Proteins, peptides and compounds were all dissolved in ITC buffer (20 mM HEPES, 100 mM NaCl, pH7.5), with protein samples bufferexchanged through dialysis using D-tubes (Millipore). Each ITC titrations consisted of 20 titrations: 1 initial injection of 0.4 µl over 0.8s, followed by 19 injections of 2 µl over 4s, at 2 minute intervals. Data was analyzed using the Microcal LLC ITC200 Origin data analysis software, using a single binding site model, to determine binding values such as Kd and enthalpy of binding.
Each assay well had a final volume of 25 µl. First 5 µl of 5X ligand was mixed with 10 µl of a 2.5X mix of bromodomains and Bio-JQ1 (prepared en-masse, aliquoted, flash-frozen and stored at -80°C) and incubated for 1 hour at room temperature. The assay plate was then moved to a dark room and 5 µl of 5X acceptor beads were added and incubated for 1 hour. Then (still in darkness) 5 µl of 5X donor beads were added, the plate was incubated for 1 more hour before being read. The direct titrations of Bio-JQ1 against bromodomains follows the same procedure, but in the first step of the assay 5 µl 5X Bio-JQ1 was mixed with 10 µl 5X bromodomain. Bio-JQ1 followed an 11-step 3fold serial dilution, starting at 10 µM (final concentration).

-DMPK
Plasma Half-life: 50 µM test compound incubated in mouse BALB/c plasma (pre-warmed to 37°C and buffered to pH7.4 in ratio of 70:30 plasma:buffer). At 0, 30, 60, 120 and 180 minutes a 50 µl aliquot of incubation mixture were removed and mixed with 200ul acetonitrile, containing 50 ng/ml Donepezil as the internal standard, to stop the reaction. The samples were centrifuged to sediment any precipitated protein and the microplates sealed prior to UPLC-MS/MS analysis using a Quattro Premier XE (Waters Corporation). XLfit (IDBS) was used to calculate the exponential decay and hence the rate constant (K) using the ratio of the peak areas of the test compound to the internal standard at each time point. The half-life was calculated for each test compound using the formula: t1/2 = 0.693/K. Intrinsic Clearance: 0.5 µM test compound was incubated with BAB/c female CD1 mouse liver microsomes (Xenotech LLC TM; 0.5 mg/ml 50 mM potassium phosphate buffer, pH7.4) and the reaction started with addition of excess NADPH (8 mg/ml 50 mM potassium phosphate buffer, pH7.4). At 0, 3,6,9,15 and 30 minutes a 50 µl aliquot of the incubation mixture was removed and mixed with 100 µl acetonitrile to stop the reaction. Internal standard was added to all samples, which were centrifuged to sediment precipitated protein and the microplates were then sealed prior to UPLC MS/MS analysis using a Quattro Premier XE. XLfit was used to calculate the exponential decay and hence the rate constant (K), based on the ratio of the peak areas of test compound to internal standard at each time point. The rate of intrinsic clearance (CLi) was then calculated using the following formula: CLi (ml/min/g liver) = K x V x microsomal protein yield Where V (ml/mg protein) is the incubation volume/mg protein added, and microsomal protein yield is taken as 52.5 mg protein/g liver. 0.5 µM Verapamil used as a positive control to confirm assay performance.

PAMPA:
PAMPA was performed using a 96-well pre-coated BD Gentest™ PAMPA plate (BD Biosciences). Each well was divided into two chambers; donor and acceptor, separated by a lipid-oil-lipid tri-layer constructed in a porous filter. The effective permeability, Pe, of the compound was measured at pH 7.4. Stock solutions (5 mM) of the compound were prepared in DMSO. The compound was then further diluted to 10 µM in phosphate buffered saline at pH 7.4. The final DMSO concentration did not exceed 5% v/v. The compound dissolved in phosphate buffered saline was then added to the donor side of the membrane and phosphate buffered saline without compound was added to the acceptor side. The PAMPA plate was left at room temperature for 5 h. After which time, an aliquot (100 µl) was then removed from both acceptor and donor compartments and mixed with acetonitrile (80 µl) containing an internal standard. The samples were centrifuged (10 min, 5⁰C, 3270 g)  Recovery of compound from donor and acceptor wells was calculated and data was only accepted when recovery exceeded 70 %.

-Fluorescence Recovery after Photobleaching (FRAP)
FRAP experiments were performed in human osteosarcoma U2OS cells transfected with pcDNA5 FRT/TO plasmids encoding wild-type and mutant GFP chimeras of BET proteins. Cells were seeded into glass-bottom, 35x10mm dishes (WillCo) at ~200 000 cells in 2 ml media per dish ~66 hours before FRAP. Cells were transfected using FuGENE HD (Promega) and 1 µg of plasmid DNA ~42 hours before FRAP. 18 hours before FRAP cells were treated with test compounds and SAHA (Sigma-Aldrich) to a final concentration of 1 µM and 2 µM, respectively, giving a final DMSO concentration of 0.03%.
FRAP experiments were carried out on a Deltavision Elite imaging system (GE Healthcare) running Resolve 3D (SoftWoRx) kept at 37°C and using a FITC filter set (488 nm excitation, 525 nm emission). Cells were imaged using a 60X objective lens (Olympus), with an exposure time of 0.05s and using 2x2 binning to give a 512x512 pixel image. Cells were excluded for FRAP experiments if they displayed aberrant morphology and signs of cell-death. Cells providing a signal of <1000 fluorescence units were excluded as they were too close to the background signal for photobleaching (requiring reduction to ~50% fluorescence) to be observed. Cells fluorescing over 3500 units were also excluded due to the risk of detector saturation.
Cells were photobleached using a quantifiable laser module (QLM), set to a wavelength of 488 nm and 100% power, for 0.5s, covering an area with a 0.5 µm 2 radius. For each cell 5 images were captured pre-bleach (over 5 seconds) and 32 post-bleach, using a CoolSNAP HQ camera (Photometrics). Post-bleach imaging was usually spread over 60s and distributed to best measure a t½ of ~2s, although this was altered in some cases. FRAP experiments were analyzed in SoftWorx (GE Healthcare) using the PK analysis function, set to analyze a bleach event with 0.5 µm 2 radius. Calculated t½ values were extracted and analyzed in Prism 6 (GraphPad). Each experiment was run twice on separate days, with ~20 cells tested on each day.

-WT Cell Cytotoxicity
Compound cytotoxicity was measured using the CellTiter-Glo assay (Promega). Compounds were serially diluted in a sterile, white, clear-bottom 384-well cell culture microplate (Greiner Bio-one), at 2X concentration and a volume of 25 µl. 25 µl of 2X cell suspension was then added. Both cells and compounds were diluted in RPMI medium. After 48hr incubation 25 µl of CellTiter-Glo reagent was added to each cells. Following 15 minute incubation the luminescence signal was read on a Pherastar FS. The final concentration of assay components are as follows: 3x10 5 cells/ml, 0.05% DMSO, 5 µM and below compound. Compounds were tested in triplicate, against the BETdependent AML cell-lines MV4-11 and HL60.

-NF-B Luciferase Assay
Luciferase experiments were performed in human HEK293T cells seeded in a 24-well plate at 100 000 cells in 500 µl media, and transfected with pcDNA5 FRT/TO plasmids encoding WT and mutant GFP-BRD4 chimeras, and pBABE NF-B-RE/luc2P reporter plasmid. Cells were transfected using Lipofectamine™ LTX (Invitrogen) with PLUS™ reagent, with each well receiving 0.5 µg of each plasmid, 2.5 µl LTX and 0.5 µl PLUS reagent. The next day cells were treated with compounds, giving a final concentration of 1 µM compound and 0.01% DMSO. 6 hours later cells were washed with PBS, lysed with passive lysis buffer (Promega) and stored at -20°C.
Luciferase activity was measured suing the Promega luciferase assay. In a black 384-well plate 3µl of clarified lysate was mixed with 15µl luciferase assay reagent and after 15 minutes luminescence was measured in a Pherastar FS. Differences in cell numbers were controlled for by measuring the protein content of each lysate using the Pierce™ Coomassie (Bradford) assay. Each experiment was run twice, on separate days, with each experiment containing 3 technical replicates.

Chemical Synthesis General information:
NMR spectra were recorded on a Bruker 500 Ultrashield or a Bruker Ascend 400. Chemical shifts are quoted in ppm and referenced to the residual solvent signals: 1 H δ = 7.26 (CDCl3), 13 C δ = 77.16 (CDCl3). High Resolution Mass Spectra (HRMS) were recorded on a Bruker micrOTOF. All chemicals, unless otherwise stated are commercially available and used without further purification. Micro wave reactions were performed in Biotage Initiator. Flash column chromatography was performed using a Teledyne Isco Combiflash Rf or Rf200i. Prepacked columns RediSep Rf Normal Phase Disposable Columns were used. Preparative HPLC was performed on a Waters mass directed HPLC with a Waters X-Bridge C18 column (100 mm x 19 mm; 5 μm particle size). Separation of the diastereomers was achieved with a gradient of 5 % to 95 % acetonitrile in water with 0.1 % formic acid in the aqueous phase.
Compounds 1, 2 I-Bet762, 2 and 2 3 were prepared according to literature procedures. 3, 4 and 4 4 were prepared according to previous work in our group.
General procedure for the alkylation in α-positon: Compound 1 (200 mg, 487 μmol, 1 eq.) or 2 (200 mg, 487 μmol, 1 eq.) was dissolved in anhydrous tetrahydrofuran (5 ml in the case of 1 and 10 ml in the case of 2). This solution was then added drop wise to a solution of potassium hexamethyldisilazane (1.17 ml of a 0.5 M solution in toluene, 584 μmol, 1.2 eq.) in tetrahydrofuran (5 ml) at -80 °C under an atmosphere of nitrogen. After 1 h at this temperature the corresponding alkyl iodide (584 μmol, 1.2 eq.) was added drop wise. The reaction mixture was warmed to 25 °C over 18 h and a few drops of acetic acid were then added to quench the reaction. The solvent was removed and the residue purified by flash column chromatography using a linear gradient from 10 % to 60 % acetone in heptane. For isomerization, the alkylated compound, together with freshly prepared sodium methoxide (10 eq.), was dissolved in methanol (2 ml) and heated to 120 °C for 40 min in a microwave reactor. The reaction mixture was acidified with aqueous hydrochloric acid (1 M), diluted with water and extracted three times with dichloromethane. The combined organic phases were dried over magnesium sulphate and evaporated to dryness. Separation of the diastereoisomers was achieved as described above. (5).