Systematic Assessment of Fragment Identification for Multitarget Drug Design

Abstract Designed multitarget ligands are a popular approach to generating efficient and safe drugs, and fragment‐based strategies have been postulated as a versatile avenue to discover multitarget ligand leads. To systematically probe the potential of fragment‐based multiple ligand discovery, we have employed a large fragment library for comprehensive screening on five targets chosen from proteins for which multitarget ligands have been successfully developed previously (soluble epoxide hydrolase, leukotriene A4 hydrolase, 5‐lipoxygenase, retinoid X receptor, farnesoid X receptor). Differential scanning fluorimetry served as primary screening method before fragments hitting at least two targets were validated in orthogonal assays. Thereby, we obtained valuable fragment leads with dual‐target engagement for six out of ten target combinations. Our results demonstrate the applicability of fragment‐based approaches to identify starting points for polypharmacological compound development with certain limitations.

The FXR LBD construct (aa244-472) was cloned using the pFA-CMV-hFXR-LBD 2 as PCR template (forward primer 5'-CCCCGGATCCGAACTGACCCCAGATCAACAG-3' and reverse primer 5'-CCCCCTCGAGTCATTATCACTGCACGTCCCAGATTTC-3). The PCR product as well as a modified pET29-vector were linearized with the restriction enzymes BamHI and XhoI and purified using NucleoSpin® Gel and PCR Clean-up (MACHERY-NAGEL) according to supplier protocol. The purified products were ligated using T4-ligase for 1h at 37 °C. The mixture was then used to transform E. coli (DH5α) by standard heat shock transformation protocol. Single clones were selected from agarose plates, containing a final concentration of 100 µg/ml Kanamycin and multiplied plasmids were extracted using NucleoSpin® Plasmid EasyPure (MACHERY-NAGEL) according to supplier protocol. The sequence of the insert was verified by Sanger sequencing.  Figure S1: Alignment of the His-TEV-FXR(aa244-472) construct and the sequencing data (T7term primer (TGCTAGTTATTGCTCAGCGG) from Microsynth Seqlab) using the standard settings of "map to reference" with the software Geneious 11.0.5.

RXRα cloning:
The wildtype RXRα LBD construct (aa225-462) was first amplified via PCR (forward primer 5'-CCGAAAACCTGTACTTCCAGGGATCCACCTCTTCCGCCAATGAAGATATGCCGG-3' and reverse primer 5'-CTGCAGGTCGACCTCGAGTCATTAGGTCATCTGATGCGGAGCTTCC-3') from a tag free RXRα LBD construct and carried an N-terminal TEV recognition and cleavage side. The PCR product was again cloned into a pMAL vector with an ampicillin resistance using PCR (forward primer 5'-ATGACTCGAGGTCGACCTGCAGGCAAGCTTGG-3' and reverse primer 5'-CTGCAGGTCGACCTCGAGTCATTAGGTCATCTGATGCGGAGCTTCC-3'). The final construct carried an N-terminal maltose binding protein (MBP) tag followed by a poly-N-tag, a His8-tag and a TEV site. Protein expression and purification FXR, LTA4H, sEH-H and sEH-Fl: The LTA4H was purified as previously described 3 and FXR was purified adapted from Merk et. al. 4 . In brief, for the expression of LTA4H and FXR transformed E.coli BL21 (DE3) cells were grown in LB Broth supplemented with 100 µg/ml Kanamycin at a temperature of 37 °C and shaking speed of 180 rpm until an OD600 of 0.8 was reached. Then the cultures were induced by the addition of isopropyl β-D-thiogalactopyranosid (IPTG), with a final concentration of 400 µM in case of the LTA4H and 100 µM for FXR. After induction the incubation temperature was lowered to 21 °C for LTA4H and 18 °C in case of FXR. Cultures were harvested after ~ 18 h by centrifugation at 4 °C with 5500 x g for 20 min (centrifuge: Sorvall LYNX 4000 super speed centrifuge; rotor: F12-6X500-LEX fiberlite). For the expression of sEH-H and sEH-Fl transformed E.coli BL21 (DE3) cells were grown in 5052 media supplemented with 100 µg/ml kanamycin at 37 °C and 180 rpm as described by Hahn et al. 5 and Lukin et al. 6 . Temperature was reduced to 16 °C after 3 h and cells were harvested by centrifugation after additional 42 h of incubation. The cell pellets were either stored at -20 °C or directly suspended in buffer A supplemented with DNAse I and protease inhibitor mix (cOmplete™, EDTA-free Protease Inhibitor Cocktail, Roche). Cell suspension was lysed by a cell homogenizer (Invensys APV-1000 Homogenizer, Denmark). The lysed cells were centrifuged with 43992 x g at 4 °C for 1 h (centrifuge: Sorval RC 5B Plus; rotor: HFA22.50). The supernatant was filtrated using a 0.45 mm syringe filter, before the filtrate was further purified by using an ÄKTA purifier (GE Healthcare) with a 5 ml HisTrap HP (GE Healthcare) equilibrated and run in buffer A. The HisTrap was eluted using buffer B and the protein concentration of the elution fractions was measured. The fractions containing protein were pooled and if necessary the volume was reduced to 5 ml by using an Amicon® Ultra-15 Centrifugal Filter Unit with a 3 kDa cutoff and centrifuging at 4°C with 2500 x g (centrifuge: 5810R; rotor: F-34-6-38). In case of FXR, LTA4H and sEH-H the protein was filtrated through a 0.22 mm filter and the 5 ml were further purified by using the ÄKTA purifier (GE Healthcare) running in buffer C and a HiLoad 16/600 Superdex 75 pg TM (GE Healthcare) for FXR and a HiLoad 16/600 Superdex 200 pg TM (GE Healthcare) in case of LTA4H and sEH-H. sEH-Fl was instead dialyzed two times against ~ 100x volume of buffer C for about 8h each. The purity of the protein was evaluated by SDS page using a 14% polyacrylamide gel followed by Coomassi staining. The amount of protein was determined by measuring the sample with a NanoDrop TM 2000c. The fractions containing pure protein were pooled. They were either used for ITC experiments or in case of sEH-H and sEH-Fl the pooled fractions were supplemented with glycerol to a final concentration of 20%. The pooled fractions were then frozen in liquid nitrogen and then stored at -80 °C.
For the ITC experiments FXR was expressed as described above but after the purification using a 5 ml HisTrap HP (GE Healthcare) the fractions containing the protein were pooled and a self-produced TEV protease containing a 6xHis tag was added with a molar ratio of 1:10. The mixture was dialyzed overnight using buffer A and a dialysis membrane with a 3 kDa MWCO. Then a second run using a 5 ml HisTrap HP (GE Healthcare) was performed. Proteins without His-tag were in the flow through. These fractions were concentrated using an Amicon® Ultra-15 Centrifugal Filter Unit with a 10 kDa cutoff centrifuged at 4 °C with 2500 x g (centrifuge: 5810R; rotor: F-34-6-38) until a final volume of 5 ml was reached. Concentrated sample was further purified by gel filtration in an analog fashion to the normal FXR purification. From the fractions containing the highest amount of protein an SDS page was performed and the purest fractions were pooled and dialyzed two times for 12h in 1 L buffer C. The protein was frozen in liquid nitrogen and stored at -80 °C. The dialysis buffer was stored at -20 °C.

5-LOX:
The 5-LOX protein was expressed and purified as previously published 7,8 . In short the transformed E.coli BL21 (DE3) cells were grown in LB Broth supplemented with 50 µg/ml ampicillin at a temperature of 37 °C and shaking speed of 180 rpm for 5 h. Then the temperature was reduced to 21 °C and the cultures were induced by adding IPTG to a final concentration of 200 µM. After growing overnight, the cells were harvested by centrifugation (centrifuge: Sorval RC 5B Plus; rotor: HFA12.500). The cells were suspended using 20 ml buffer A per Liter of expression culture and chilled for 20 min on ice. The suspended cells were lysed by sonification (Bandolin MS73) 3x 1 min with 90% intensity. Cell homogenate was centrifuged (centrifuge: Sorval® RC-5B Refrigerated Superspeed Centrifuge; rotor: 5534) for 15 min at 4 °C with 10,000 x g, before the supernatant was further centrifuged for 70 min at 4 °C with 100,000 x g (centrifuge: Optima™ LE-80K Ultracentrifuge; rotor: Beckman fixed angle rotor 55.2 Ti). The supernatant was applied on a custom-made ATP-column with a column volume of ~ 3.5 ml followed by a ResourceQ 1 ml column. Purification steps were performed by means of an ÄKTA Xpress system (GE Healthcare) using buffer B. The 5-LOX was eluted using buffer B supplemented with 500mM NaCl and then a gradient from 0% up to 100% of buffer B supplemented with 20 mM ATP. For the DSF experiments the buffer was changed to buffer B using an ÄKTA purifier (GE Healthcare) with three subsequentially arranged 5 ml HiTrap® Desalting Columns (GE Healtcare). The concentration was determined by performing a Bradford assay according to the Bio-Rad instruction manual.

RXRα:
For the RXRα protein expression the pMal plasmid coding for the wildtype RXRα LBD construct was transformed into competent E.coli T7-express cells which already held a pGro7 co-plasmid, coding for the GroEL/ES chaperone, with a chloramphenicol resistance (from the Chaperone Plasmid Set Cat. #3340, TAKARA Bio Inc. (Japan)). Positive clones were selected on LB -plates supplemented with 100 µg/ml ampicillin and 34 mg/ml chloramphenicol. These clones where then used to inoculate a pre-culture of 250 ml LB-medium supplemented with 100 µg/ml ampicillin and 34 mg/ml chloramphenicol. After incubation of 1 h (37 °C, 180 rpm), 40 ml of the pre-culture were used to inoculate the expression culture of 1 l LB-medium without antibiotics. The expression culture was incubated at 37 °C with 180 rpm. At an OD600 of approximately 0.6-0.7 the GroEL/ES expression was induced using 1 mg Arabinose. The flasks were then moved into a shaker at 20 °C and 120 rpm. 30 minutes later at an OD600 of 0.9-1.0 the target protein expression was induced using 0.5 mM IPTG. The expression cultures were incubated overnight and then harvested at 6000 rpm at 4 °C for 15 min (centrifuge: Sorval® Lynx 6000; rotor: F10-4x1000). For the purification a pellet of 2 l E. coli culture was suspended in 50 ml lysis buffer A supplemented with protease inhibitor, DNAse, RNAse and lysozyme. Incubation was performed for at least 30 min and the mechanically lysis was done with a cell homogenizer (Invensys APV-1000 Homogenizer, Denmark). In order to force the ATP dependent GroEL/ES release the lysate was then incubated with 1 mM ATP for at least 30 min on ice. The cell debris were removed via centrifugation (16500 x g, 20 min, 4 °C, centrifuge: Eppendorf 5810 R; rotor: F-34-6-38) and the supernatant was applied onto a 15 ml Ni-IDA Sepharose 6 fast-flow column. The target protein eluted at 300 mM imidazole. The N-terminal MBP used as a solubility-tag was removed with the help of a recombinant MBP-TEV protease during dialysis in 3 l buffer B. The TEV protease was applied with a 1:50 molar ratio. In order to remove the MBP-TEV protease, the cleaved N-Terminal MBP-tag and undigested target protein from the protein solution an amylose affinity chromatography (AAC) step was performed. An Amylose High Flow resin (NEB) on a gravity flow column was used. The MBP-TEV digest was applied onto the column and the flow through containing the cleaved protein was collected. The flow through of the AAC was then concentrated and purified in a final step with a 26/60 Superdex 75 column on an ÄKTA purifier system in assay buffer C. The elution fractions were pooled, shock frozen with liquid nitrogen and stored at -80 °C.
For the ITC measurements RXRα was dialyzed, using a dialysis membrane with a 3 kDa MWCO and placed in 2 l fresh buffer C, overnight. Then concentrated using an Amicon® Ultra-15 Centrifugal Filter Unit with a 3 kDa cutoff and centrifuging at 4°C with 1500 x g (centrifuge: Eppendorf 5810R; rotor: F-34-6-38) until a concentration higher than 50.5 µM was reached.  The measurements were performed using a iCycler iQ™ Single-Color Real Time PCR device with a heat rate of 1 °C per min, an excitation wavelength of 490 nm and an emission wavelength of 570 nm. The first derivation calculated by MyiQ Optical Software 1.0 was analyzed using Microsoft Office Excel 2013 by comparing the maxima. A shift of 1 °C (in case of 5-LOX 0.9°C) or more compared to the DMSO control were assumed as potentially binding of the fragment.
Orthogonal assay systems Fragments: The 10 fragments showing dual activity in the DSF experiments were validated by functional assay screening with a fixed concentration of 500 µM in case of sEH-H and LTA4H, 100 µM in case of 5-LOX and 30 µM, 10 µM and 5 µM in case of FXR and RXRα. The FXR and RXRα activity assays didn't show any results, because the concentration of used fragments was toxic for the cells or didn't show any activity. The compounds showing an inhibition greater than 50% (sEH and LTA4H), greater than 70 % (5-LOX) as well as the hits determined in the DSF experiments were further measured with concentrations between 1000 µM and 0.1 µM (sEH-H and LTA4H) and between 300 µM and 1 µM (5-LOX). The sEH and LTA4H activity assays were performed as triplicate and repeated 3 times. The 5-LOX activity assay was performed with one point per concentration and repeated 3 times Raw data was first analyzed in Microsoft Office Excel (2013), before IC50 curves were plotted and further analyzed using GraphPad Prism 7 and GraphPad Prism 8 for the shown graphs.
Approved drugs: The eight approved drugs were measured in the orthogonal assay systems analogous to the fragments. The approved drugs were tested on the targets, where the fragment structure they contain, showed activity. For the sEH-H and LTA4H, they were measured as triplicate in concentrations between 1000 µM and 1 µM and each measurement was repeated three times. For the 5-LOX they were measured as single point measurement in three different experiments with concentrations between 300 µM and 1 µM. For FXR and RXRα a single hybrid reporter gene measurement with three concentrations between 30 µM and 3 µM was performed. Raw data was processed according to the fragments.
sEH-H activity assay: The sEH-H activity assay was performed adapted from Hahn et al. 5 and Lukin et al. 6 . In short 1 µl of compound dilution in DMSO or pure DMSO (for the positive and negative control) was placed in a black 96-well plate (Thermo Scientific™ Nunc™ F96 MicroWell™ Black Polystyrene Plate) and then 89 µl of protein-mix or buffer mix (without protein as negative control) were added and the mixture was incubated for 30 min at room temperature. 10 µl of 0.6 mM 3-phenyl-cyano(6-methoxy-2naphthalenyl)methylester-2-oxiraneacetic acid (PHOME) were added to every well. The Plate was measured every minute for 45 min. Final concentration per well was 3 nM sEH-Fl, 0.001 % Triton X-100, 25 mM Bis-tris, 0.1 mg/ml BSA, pH = 7.0, 1.6% DMSO and 60 µM PHOME.
LTA4H activity assay: The LTA4H activity assay was performed adapted from Hiesinger et al. 3 .As in the sEH-H measurement 1 µl of compound dilution in DMSO or pure DMSO (for the positive and negative control) was placed in a black 96-well plate (Thermo Scientific™ Nunc™ F96 MicroWell™ Black Polystyrene Plate) and then 89 µl of protein-mix or buffer mix (without protein as negative control) were added and incubated for 30 min at room temperature. Then the reaction was started by the addition of 10 µl 1.82 mM L-arginin-7-amido-4-methylcumarine (Arg-AMC) to every well. After the addition the plate was measured every minute for 45 min. The final concentration per well was 100 nM LTA4H, 0.001% Triton X-100, 48.7 mM Tris, 48.7 mM NaCl, pH = 8, 1.4% DMSO and 182 µM Arg-AMC.

5-LOX activity assay:
The 5-LOX-activity assay was performed adapted from Wertz et al. 9 and Kretschmer et al. 7 . In short, the purified 5-LOX protein was used in a final concentration of 3 µg/ml in assay buffer (1x PBS, 1 mM EDTA, pH = 7.4) and preincubated with 10 µl compound in DMSO or pure DMSO as positive control. Then 1 mM ATP (Roth) was added. The mixture was preincubated for 15 min on ice and then for 30 s at 37 °C using a water bath. The reaction was started by addition of 2 mM CaCl2 and 20 µM arachidonic acid (Cayman). After 10 min of incubation at 37 °C the reaction was stopped by adding 1 ml of chilled methanol and the reaction vessel was placed on ice. 30 µl 1N HCl, 200 ng of Prostaglandin B1 (as internal standard) and 500 µL PBS were added. The Clean-UP C-18 solid phase extraction columns (UCT) were conditioned using 1 ml of methanol and 1 ml of water. Then the samples were given on the columns and washed with 1 ml water and 1 ml 25% methanol. The extraction was performed by the addition of 300 µl 100% methanol. The extract was diluted by adding 120 µl water and 100 µl of this dilution were analyzed by HPLC using the method and setup as described from Steinhilber et al. 10 .
Hybrid reporter gene assay FXR and RXRα: The assay was performed as described by Pollinger et al. 2 . In brief HEK293T cells were cultured at 37 °C and 5% CO2 using a mixture of DMEM high glucose, 10% FCS, 1 mM sodium pyruvate, 100 U/ml penicillin and 100 µg/ml streptomycin. The cells were seeded in a 96-well plate with a density of 3*10 4 cells/well and after a growth time of 24 h the medium was changed to Opti-MEM. For the transfection Lipofectamin LTX reagent (Invitrogen™) with the respective plasmids was used as described in the manufacturer's protocol. For FXR testing the plasmids pFR-Luc (Stratagene), pRL-SV40 (Promega) and pFA-CMV-hFXR-LBD were used. For RXRα the plasmids pFR-Luc (Stratagene), pRL-SV40 (Promega) and pFA-CMV-hRXRα-LBD. 5 h after transfection the medium was changed to a mixture of Opti-MEM, 100 U/ml penicillin, 100 µg/ml streptomycin and 0.1% DMSO with or without compound (untreated control). After incubation for 14-16 h the Dual-Glo™ Luciferase Assay System (Promega) was performed as described by the manufacturer. The luminescence was measured using a Spark 10M luminometer (Tecan Deutschland GmbH) device. The measured firefly luciferase data were divided by the associated renilla luciferase data and multiplied by 1000 leading to a normalized result as relative light unit (RLU). The fold activation was then obtained by dividing the RLUs of the compound by the mean of the untreated control. Every plate also contained reference compounds (GW4064 for FXR and bexarotene for RXRα) as positive control.
ITC measurements for sEH-H, LTA4H, FXR, and RXRα: ITC experiments were performed in normal mode using an "Affinity ITC" (TA-Instruments) at 25 °C. FXR was used with a concentration of 50 µM in Tris buffer (10 mM Tris,100 mM NaCl,5 mM DTT,pH = 8.3), supplemented with 1% DMSO and 0.5 % CHAPS. RXRα was also used in a concentration of 50 µM in a HEPES buffer (containing 150 mM potassium fluoride, 25 mM HEPES, 5 mM DTT, 10% Glycerol, pH = 7.5) supplemented with 1% DMSO and 0.5% CHAPS. LTA4H was used with a concentration of 20 µM (for compound 10) and 40 µM (for compound 9) in a Tris buffer (50 mM Tris, 50 mM NaCl, pH = 8) supplemented with 1% of DMSO. sEH-H was used with a concentration of 20 µM (for compound 10) and 40 µM (for compound 9) in a Tris buffer (50 mM Tris, 500 mM NaCl, pH = 8) supplemented with 1% of DMSO. 300 µL of the protein solution were placed in the cell. For blank experiments the dialysis buffer supplemented with 1% DMSO and 0.5% CHAPS was used. A 250 µM or a 100 µM dilution of compound in buffer containing final concentrations of 1% DMSO (as well as 0.5 % CHAPS, in case of FXR and RXRα) was placed in the syringe. The titrations were performed with 1 injection of 0.5 µL and 30 injections with 2 µL or 35 injections with 1.5 µL as well as a spacing time between the injections of 240 or 300 seconds. One measurement with protein in the cell and blank buffer (respective buffer supplemented with 1% DMSO (as well as 0.5% CHAPS, in case of FXR and RXRα)) in the syringe was performed as well as two times for every compound measurement. A second blank experiment was performed where compound was titrated in the respective blank buffer. The results were analyzed using the program "NanoAnalyze Data Analysis" (Version 3.10.0 from TA Instruments) by subtracting a fix heat, determined by the blank experiments and using the independent option for modeling.
Results of the fragment testing DSF experiments: sEH-H: Graph S1: ITC titration of 40 µM sEH-H in cell with 250 µM compound 9 in syringe. One initial step with 0.5 µL followed by 35 steps with 1.5 µl volume. The measurement was performed at 25 °C and between every step there was a spacing of 240 s. steps with 2 µl volume. The measurement was performed at 25 °C and between every step there was a spacing of 300 s.
Results of the tested approved drugs sEH-H activity: