Proposed authors:
Scott J. Pollack, Kim S. Beyer, Christopher Lock, Ilka Müller, David Sheppard, Mike Lipkin, David Hardick, Peter Blurton, Philip M. Leonard, Paul A. Hubbard, Daniel Todd, Chris Richardson, Thomas Ahrens, Manuel Baader, Doris O. Hafenbradl, Kate Hilyard, Roland W. Bürli
A comparative study of fragment screening methods on the p38α kinase: new methods, new insights
Affiliations
S. J. Pollack · C. Lock · I. Müller · D. Sheppard · M. Lipkin · D. Hardick · P. Blurton · P. M. Leonard · P. A. Hubbard · D. Todd · C. Richardson · K. Hilyard · R. W. Bürli (corresponding author)
BioFocus, Chesterford Research Park
Saffron Walden, Essex, CB10 1XL
United Kingdom
T: +44 1799 533547
F: +44 1799 531495
K. Beyer · Thomas Ahrens · Manuel Baader · D. Hafenbradl
BioFocus, Switzerland
Gewerbestrasse 16
4123 Allschwil
Switzerland
Supplementary Material
SPR screen
The fragment library was screened for p38a affinity using a BiacoreTM T100 (GE Healthcare). 6xHis-tagged p38a was first immobilized by capture onto a nickel-containing nitrilotriacetic acid (NTA)-derivatized sensor chip (GE Healthcare), preceeded by activation of the carboxymethylated dextran surface for mild amine coupling. 6xHis-tagged p38a capture was carried out on flow cell 2 of an NTA chip preconditioned as per the manufacturer (GE)’s instructions at 25 oC on a Biacore T100 with HBS-P as the running buffer (10 mM HEPES, pH 7.4 with 150 mM NaCl, 50 µM EDTA and 0.05% Tween20). The surface was first charged with nickel (500 µM NiCl2 in HBS-P buffer for 300 s at a flow rate of 5 µl/min). The dextran surface was then activated with a 1:1 mixture of Nethyl-N-dimethylaminopropylcarbodiimide (EDC) and N-hydroxysuccinimide (NHS) on line for final concentrations of EDC and NHS of 37.5 mg/ml and 5.8 mg/ml, respectively, for 240 s at a flow rate of 5 µl/min. p38a (200 nM in HBS-P buffer in the presence of 1 µM of the ATP site binding compound SB203580 to protect the active site) was then captured for 900 s at a flow rate of 20 µl/min followed by a final blocking injection of 1 M ethanolamine, pH 8.5, for 300 s at a flow rate of 5 µl/min. A freshly immobilized surface was used for each phase (primary screening, confirmation screening, affinity studies). Net immobilization levels of 4537 to 4912 resonance units (RU) were obtained. For some experiments, 6xHis-tagged carbonic anhydrase (CA; R&D Systems), MAP kinase kinase-6 (MKK6) or non-activated p38a were immobilized using the same method on parallel flow cells. Net immobilization levels of 6046 to 6059, 6337 and 4811 to 5197 RU were obtained for these proteins, respectively.
Binding measurements were carried out in a running buffer consisting of 50 mM Tris, pH 7.6 with 150 mM NaCl, 10mM MgCl2, 1 mM MnCl2, 0.05% (v/v) Tween-20 and 5% (v/v) DMSO. For affinity measurements, compound solutions were prepared using ten two-fold serial dilutions in running buffer from a 3 mM top concentration (prepared from DMSO stocks) along with a zero (running buffer) control. A Biacore method program was used that included a series of ten start up injections (running buffer), zero control (running buffer) and 500 µM ATP injections every 20 cycles as calibration standards, and cycles of solvent correction samples every 40 cycles consisting of eight buffer solutions varying from 3.75 to 6.5% (v/v) DMSO in order to correct for DMSO bulk responses between reference and target flow cells. A flow rate of 30 µl/min was used with a 30 s contact time (using high performance injection parameters) followed by a 120 s dissociation phase. An extra tubing wash step after each injection with 50% DMSO was included. Results were analyzed by subtracting the signals of the reference surface from the signals for the protein-bound surfaces and performing a solvent correction. Compound binding responses were normalized for molecular weights and for the calibration standard (ATP and blank) responses throughout the screen. For affinity measurements, the percentage of theoretical Rmax (% TRmax) and Kd values were determined using the Biacore affinity analysis curve fitting algorithm. For competition experiments, each injection series included (1) a blank (running buffer) for background subtraction; (2) the ATP site binding compound SB203580 (at a saturating concentration of 1 µM); (3) the test fragment at 500 µM; and finally (4) a mixture of SB203580 (1 µM) and the test compound (500 µM). Two independent experiments were performed.
Biochemical assays
For both biochemical assay formats employed in this case study, LC3000 Caliper mobility shift assay (MSA) and fluorescence lifetime (FLT), full assay development was carried out including determination of optimal enzyme concentration, ATP KM, substrate KM, DMSO tolerance, sensitivity to reference inhibitors and automation. The following final conditions were employed for screening (with all reagent amounts given as final concentrations):
For the LC3000 assay 1 µl of compound or sample in DMSO (8.3%) were pre-incubated with 5.5 μl p38a (10 nM) in reaction buffer (50 mM HEPES, pH 7.4, 10 mM MgCl2, 1 mg/ml BSA, 1 mM DTT) in Greiner LowVolume 384-well plates for 30 min. The reaction was started by addition of Kinase ProfilerPro Peptide 08 (Caliper LifeSciences) (1 μM) and ATP (10 μM) in 5.5 μl reaction buffer. The reaction was stopped after 2 h reaction time by addition of 12 μl stop solution (100 mM HEPES, pH 7.4, 40 mM EDTA, 0.015% Brij-35, 0.2% coating reagent #3). Phosphorylated and non-phosphorylated peptides were separated on a Caliper LC3000 device equipped with a 12-sipper chip. Separation buffer and conditions were as described in Caliper LifeSciences Application note 210 [1]. Percent conversion values as determined by the WellAnalyzer Software (Caliper LifeSciences) were used as raw data for the analysis of activity data.
For the FLT assay in 384-well, black, flat bottom plates (Greiner 781076), a FLEXYTE™ peptide substrate (Almac) tailored for p38a and incorporating a long lifetime fluorophore (9-aminoacridine)[2] was used in all subsequent experiments. Compound solution (5 µl, 250 µM in 2% DMSO) was treated with p38a (10 µl, 10 nM) and p38a peptide substrate (1.5 µM) in reaction buffer (100 mM HEPES pH 7.4, 10 mM MgCl2, 1 mM DTT, 0.015% Brij 35). The kinase reaction was performed at ambient temperature for 180 min in the dark; it was started by addition of ATP (10 µl, 150 µM) and stopped by addition of 25 µl FLEXYTE™ stop solution comprising an iron (III) based small molecule chelate which specifically complexes the phosphate group of the product peptide, resulting in a reduction in fluorescence lifetime of the 9-aminoacridine fluorophore of up to 5 ns. The amount of substrate phosphorylation and hence enzyme activity was reported through changes in the measured fluorescence lifetime. FLT was measured using the NanoTaurus FLT plate reader (Edinburgh Instruments Ltd.) employing time-correlated single photon counting (TCSPC). The excitation source was a semiconductor laser at 405 nm, producing picosecond light pulses at a repetition rate of 5 MHz. Emission was collected through a 438 nm bandpass filter (Semrock) with 20 nm bandwidth. Data was collected to 3000 counts in the peak channel and average lifetimes calculated using a biexponential decay model.
Microscale thermophoresis (MST)
The Microscale thermophoresis (MST) assay was developed at NanoTemper and validated with the well characterized p38a inhibitors BIRB796 and SB202190 (data not shown). First, p38a was labelled with a proprietary red fluorescent dye. Free dye was removed by purification on a Sephadex G-25 column and the labelled protein was resuspended in reaction buffer containing 50 mM Tris, pH 7.5, 150 mM NaCl, 10 mM MgCl2, 1 mM MnCl2, 0.1% Tween-20, 0,1% BSA. For Kd determinations, the concentration of p38a was kept constant at 50 nM and increasing amounts of fragments were added. Thermophoresis was analyzed on a proprietary instrument after different incubation times.
X-ray structures
Recombinant N-terminally His-tagged human p38a comprising residues 1-360 was expressed in E. coli BL21 cells and purified by Ni-NTA affinity chromatography, followed by ion exchange and size exclusion chromatography. Purified protein in 25 mM Tris, pH 7.5, 200 mM NaCl and 1 mM tris(hydroxypropyl)phosphine (THP) was concentrated to 8 mg/ml, with 1 mM adenosine added prior to crystallization. Crystallization was carried out at 20 °C by hanging drop vapor diffusion, mixing the protein with an equal amount of reservoir solution containing 100 mM BisTris, pH 6.5, 200 mM NaCl and 20% PEG3350. Crystals grew overnight to a typical size of 0.4 x 0.3 x 0.05 mM. These crystals were used for soaking with the compounds identified by SRP and Caliper LC3000 assay. The soaking solution contained 100 mM BisTris pH 6.5, 200 mM NaCl and 25% PEG3350 and was supplemented with variable amounts of ligand in 100% DMSO, and crystals were incubated for several days (Table 1).
Table 1 Summary of soaking conditions
Compound # / ligand (mM) / DMSO content v/v / soaking time (days)3 / 8 / 10 % / 6
4 / 50 / 10 % / 2
5 / 8 / 10 % / 5
6 / 50 / 10 % / 2
13 / 20 / 4 % / 5
Data were collected at 100 K on either a Rigaku R-AXIS IV++ image plate or Rigaku Saturn 944+ CCD detector, with data indexed, integrated and scaled using HKL2000, or MOSFLM and SCALA (CCP4).[3-6] The structures were solved by molecular replacement with either PHASER[7] or MolRep (CCP4) using the structure of apo-p38a 1-352 as a search model. Ligand coordinate files and library descriptions were created in Sketcher (CCP4) and the ligand was manually fitted into the electron density using Coot.[8] Structure refinement was carried out in Refmac5 (CCP4).[9] The electron density maps covering the binding sites of the five p38α/ligand complexes are illustrated in Figure 1 and the data collection and refinement statistics are listed in Table 2.
Figure 1 Initial Fo-Fc electron density map around ligand binding site of the five p38α/ligand complexes.
Table 2 Data collection and refinement statistics
p38a + 3 / p38a + 4 / p38a + 5 / p38a + 6 / p38a + 13Data collection
Space group / P212121 / P212121 / P212121 / P212121 / P212121
Cell dimensions (Å) / 45.4 85.3 125.3 / 45.3 85.9 126.2 / 45.3 86.7 125.7 / 45.2 85.9 126.3 / 45.1 85.3 126.0
Resolution range (Å)a / 30.00-2.20 (2.26-2.20) / 40.7-2.3
(2.42-2.30) / 30-2.15
(2.23-2.15) / 30-2.4
(2.49-2.40) / 70.65-2.55 (2.69-2.55)
Rmerge (I)b (%) / 12.4 (31.2) / 9.0 (51.7) / 14.0 (41.6) / 13.7 (54.7) / 9.1 (41.6)
I/sigma (I) / 7.0 (2.3) / 9.4 (1.9) / 8.3 (1.8) / 7.1 (1.7) / 8.0 (2.2)
Completenessa (%) / 94.5 (84.9) / 94.5 (76.0) / 98.3 (84.7) / 99.9 (99.7) / 93.9 (77.9)
Refinement statisticsc
Rwork (%) / 21.3 / 22.6 / 21.7 / 24.3 / 21.9
Rfree (%) / 25.8 / 27.6 / 24.3 / 28.6 / 27.7
Ramachandran plotd
Favoured (%) / 96.8 / 96.7 / 96.2 / 96.7 / 95.3
Additional (%) / 2.6 / 3.0 / 2.9 / 3.3 / 4.7
Outlier (%) / 0.6 / 0.3 / 0.9 / 0 / 0
Refined B-factors
Protein / 30.6 / 27.9 / 29.0 / 32.9 / 32.9
Ligande / 47.9 / 47.1 / 31.3 / 40.1 / 26.1
aValues in parentheses are for the highest-resolution shell
bRmerge (I) = ΣΣj|Ihj-<Ih>|/ΣΣj|Ihj|], where j is the number of reflection h
cRefinement with REFMAC5[8]
dRamachandran diagram has been calculated with RAMPAGE[10]
eLigand occupancy set to 1.0
References
1 Caliper LifeSciences Application note 210 http://www.caliperls.com/assets/009/5902.pdf Accessed 7 Apr 2011
2 Maltman BA, Dunsmore CJ, Couturier SCM, Tirnaveanu AE, Delbederi Z, McMordie RAS, Naredo G, Ramage R, Cotton G Chem Commun (2010) 46: 6929-6931
3 Otwinowski Z, Minor W (1997) in Methods in Enzymology, Volume 276: Macromolecular Crystallography, part A, p.307-326, Carter CW, Sweet RM (ed), Academic Press (New York)
4 Leslie AGW (2006) Acta Cryst D62: 48-57
5 Evans P (2006) Acta Cryst D62: 72-82
6 Collaborative Computational Project, Number 4 (1994) Acta Cryst D50: 760-763
7 McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC, Read RJ (2007) J Appl Cryst 40: 658-674
8 Emsley P, Lohkamp B, Scott WG, Cowtan, K (2010) Acta Cryst D66: 486-501
9 Murshudov GN, Vagin AA, Dodson EJ (1997) Acta Cryst D53: 240-255
10 Lovell SC, Davis IW, Arendall III WB, de Bakker PIW, Word JM, Prisant MG, Richardson JS, Richardson DC (2003) Proteins: Struct. Funct. Genet. 50: 437-450
6