Structure-Guided Design and In-Cell Target Profiling of a Cell-Active Target Engagement Probe for PARP Inhibitors

Inhibition of the poly(ADP-ribose) polymerase (PARP) family of enzymes has become an attractive therapeutic strategy in oncology and beyond; however, chemical tools to profile PARP engagement in live cells are lacking. Herein, we report the design and application of PARPYnD, the first photoaffinity probe (AfBP) for PARP enzymes based on triple PARP1/2/6 inhibitor AZ9482, which induces multipolar spindle (MPS) formation in breast cancer cells. PARPYnD is a robust tool for profiling PARP1/2 and is used to profile clinical PARP inhibitor olaparib, identifying several novel off-target proteins. Surprisingly, while PARPYnD can enrich recombinant PARP6 spiked into cellular lysates and inhibits PARP6 in cell-free assays, it does not label PARP6 in intact cells. These data highlight an intriguing biomolecular disparity between recombinant and endogenous PARP6. PARPYnD provides a new approach to expand our knowledge of the targets of this class of compounds and the mechanisms of action of PARP inhibitors in cancer.


Materials
Chemicals were purchased from Sigma-Aldrich, Fluorochem, Acros Organics, TCI, Alfa Aesar or Fisher Scientific and used without further purification. AzTB and AzRB were synthesized in-house as previously reported. 1 The minimal clickable photocrosslinkable group 5 was synthesized as previously reported. 2 Phthalazinone core 8, AZ9482 and AZ0108 were provided by AstraZeneca. Olaparib was purchased from VWR International. NeutrAvidin agarose resin and PreScission™ Protease were purchased from Thermo Fisher Scientific. Streptavidin magnetic beads were purchased from New England BioLabs. Recombinant GST-PARP6 protein and pcDNA3.1™ mammalian expression vector containing FLAG-PARP6 were provided by AstraZeneca. 6-biotin-17-NAD + was purchased from Bio-Techne Ltd.

Tissue Culture
All cell culturing was carried out in a sterile tissue culture cabinet sprayed with 70% (v/v) EtOH before and after use. All cell lines were cultured at 37 °C in a 5% CO2 incubator. MDA-MB-468 cells were cultured in Dubecco's Modified Eagle Mediumlow glucose (DMEM) supplemented with 10% (v/v) Fetal Bovine Serum (FBS). Cell harvesting was achieved by washing with Dubecco's PBS and treatment with 0.25% (w/v) Trypsin-EDTA. After 5 min incubation at 37 °C, the trypsin was quenched with DMEM to the appropriate volume for passage and aliquoted into the appropriate number of cell culture plates. Mycoplasma tests were carried out monthly. Passage number was limited to 20-25 and stocks of early passages were frozen at -150 °C containing ~10 6 cells in 1 mL FBS containing 10% (v/v) DMSO.

Multipolar Spindle Assay
This was performed as described previously. 3 Briefly, HeLa cells were plated in 96-well plates at 7,000 cells per well and incubated at 37 °C overnight. The cells were treated with compounds in a dose-dependent manner from 0 to 11 M for 48 h. The cells were fixed by 4% (v/v) formaldehyde at room temperature for 10 min and followed by ice-cold methanol fixation for another 10 min. After washing with PBS four times, the cells were blocked in blocking buffer for 1 h at room temperature. The cells were labelled with primary antibodies, 1:2000 dilution of anti-cyclin B antibody (Thermo Fisher) and 1:4000 dilution of anti-pericentrin antibody (Abcam), for 16 h at 4 °C. After washing with PBS four times, the cells were labelled with secondary antibodies, 1:200 Alexa Fluor 488 anti-rabbit antibody and Alexa Fluor 594 anti-mouse antibody, for 1 h at room temperature. After washing with PBS twice, the nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI) for 10 min at room temperature. The cells were washed twice with PBS and then applied to image acquisition by ImageXpress Micro High Content Screening System (Molecular Devices). The data were analyzed by MetaXpress and accessed by AcuityXpress (Molecular Devices). The 16 fields in each well were acquired by ImageXpress Micro. The cyclin B was labelled for scoring the mitotic cells and pericentrin was labelled for scoring the centrosome number in each mitotic cell. Value output was taken as % mitotic cells with greater than 2 centrosomes.

In vitro PARP fluorescence anisotropy binding assays Dilution of various PARP proteins and fluorescence anisotropy probe
Recombinant full length 6HIS-tagged PARP1 protein was diluted to 6 nM with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl and incubated for four hours with an equivalent volume of 2 nM fluorescent probe diluted with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl. The final DMSO concentration of the probe was kept below 1% (v/v).
Recombinant full length PARP2 protein was diluted to 6 nM with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl and incubated for four hours with an equivalent volume of 2 nM fluorescent probe diluted with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl. The final DMSO concentration of the probe was kept below 1% (v/v).
Recombinant full length GST-tagged PARP6 protein was diluted to 160 nM with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl and incubated for four hours with an equivalent volume of 6 nM fluorescent probe diluted with 50 mM Tris pH 8, 0.001% (v/v) Triton X-100, 10 mM MgCl2, 150 mM NaCl. The final DMSO concentration of the probe was kept below 1% (v/v).

Experimental protocol
Fluorescence anisotropy of the probe when bound to the proteins was measured using a BMG Pherastar FS © in the presence of test compounds or solvent control and the effect on anisotropy determined. % inhibition values for different test compound concentrations were calculated and fitted to a four parameter logistic plot in order to determine the IC50 value. , Sigma-Aldrich) was prepared (20:1) and 20 L was added to each well. Absorbance was measured 4 times per well at 490 nm and the average absorbance taken. The average of the negative control was subtracted from every value and viability was calculated as a percentage relative to the positive control. EC50 values were calculated by fitting data to the IC50 function using GraphPad Prism 5 software.

Probe labelling assays
Probe incubation was carried out in sterile treated 6-well plates (2 mL working volume) or 10 cm dishes (7 mL working volume). All compound treatments were performed with a preprepared 1000× stock of the desired concentration in DMSO and added directly to the relevant plate/well with mixing. Irradiation was performed with an in-house designed and built UV LED box (C. Saunders) with a monochromatic wavelength of 365 nm. Optimal probe incubation times were determined previously through irradiation and lysis at various time points and selection of the earliest saturating time point.
For each experiment, the plates/dishes were pre-seeded with MDA-MB-468 cells and the experiment carried out when cells had achieved 90-100% confluency. DMEM was replaced and the plates/dishes were incubated at 37 °C for 15 min. For competition experiments only, the relevant plates/wells were first treated with DMSO vehicle (0.1% (v/v)) or varying concentrations of parent compound in DMSO as indicated and incubated at 37 °C for 1 h before treating with probe. For all probe labelling experiments, competition or otherwise, the plates/dishes were treated with either DMSO vehicle (0.1% (v/v)) or varying concentrations of probe as required by the experiment and incubated at 37 °C for 3 h.
For photocrosslinking, the following was performed 1 plate/3 dishes at a time out of the incubator. Each plate/dish had media replaced and was irradiated with UV light for 30 s (365 nm) and placed on ice while irradiation of other samples was performed.
The cells were relieved of media, washed twice with PBS, then lysed with lysis buffer (70 L (6-well plates), 300 L (10 cm dishes); 1% (v/v) Triton X-100, 1% (w/v) sodium dodecyl sulfate (SDS), EDTA-free complete protease inhibitor cocktail (1×, Roche) in PBS) on ice for 10 min. The lysates were scraped and transferred to corresponding Lo-Bind Eppendorfs. Each lysate was probe sonicated (20% amplitude, 20 s (2 s pulse, 3 s rest)) to shear the nuclear DNA. Protein concentration was determined using the DC Protein Assay (Bio-Rad) in a 96-well plate as per manufacturer's instructions.

Click reaction
The desired amount of lysed protein from each sample was made up to 0.5-2 mg mL -1 with PBS to a total volume of ≤300 L. The following "click mixture" was prepared separately, preparing 6 L for every 100 L of lysate: The click mixture was vortexed and incubated at rt for 2 min before 6 L of the mixture was added to every 100 L of lysate. The reaction mixtures were shaken at rt for 1 h before being quenched with EDTA (500 mM in H2O) to a final concentration of 5 mM.
A table-top centrifuge was pre-chilled to 4 °C. Proteins were precipitated by adding H2O (1 vol), MeOH (2 vol) and CHCl3 (0.5 vol), vortexing briefly then centrifuging at 17,000 × g for 5 min. The CHCl3 and H2O/MeOH layers were discarded and the middle layer of protein pellet was retained. The pellet was washed with MeOH (300 L), sonicated to break up the pellet then stored at -80 °C for at least 20 min. The proteins were pelleted by centrifugation at 10,000-17000 × g for 5-10 min or until a compact pellet was formed. The MeOH was decanted and the pellet air-dried for 5 min. The pellet was resuspended by completely dissolving in 1% (w/v) SDS in PBS (to 5 mg mL -1 protein) before being made up to 1 mg mL -1 protein with PBS.
In-gel fluorescence 10-or 15-well SDS-polyacrylamide gels with a 12% resolving gel and 4% stacking gel were used for all gel electrophoresis experiments and were prepared using the following recipe (makes 2 gels): Pull down 50 L of Pierce™ NeutrAvidin™ Agarose beads (proteomics) or 300 L of streptavidin coated magnetic beads (western blot) were used per 1 mg of total protein per sample to an absolute minimum of 20 L of agarose beads or 15 L magnetic beads. All bead washes were performed by moderate shaking for 1 min then either briefly pelleting by table-top centrifuge then vacuum aspirating the supernatant with fine-end pipette tips (agarose) or by partitioning the beads using a magnet (magnetic). The beads were pre-washed three times with 0.2% (w/v) SDS in PBS, then protein samples (1 mg mL -1 ) were added over the beads and incubated with moderate shaking at rt for 2 h.
For western blot analysis, the beads were washed three times with 300 L 0.2% (w/v) SDS in PBS and captured proteins were released from the beads by boiling in 14 L 2× sample loading buffer (95 °C, 10 min), briefly centrifuging and the supernatant loaded straight on to an SDS-PAGE gel. For proteomic analysis (agarose beads), the beads were treated as described below.

Chemical Proteomics
Lysates for all proteomics experiments were derived from cells cultured in 10 cm dishes in triplicate for each experimental condition. 600 g of labelled protein clicked to AzRB were enriched on NeutrAvidin agarose, all as described above. All buffers were prepared fresh and filtered (0.2 m) and the work surface cleaned with 70% (v/v) EtOH.
The beads were washed twice with 300 L 1% (w/v) SDS in 50 mM HEPES (pH 8). Proteins were reduced and alkylated with 5 mM TCEP and 10 mM chloroacetamide in 60 L 50 mM HEPES with moderate shaking for 30 min at rt. The beads were washed 3 times with 300 L 50 mM HEPES. Beads were resuspended in 30 L 50 mM HEPES and proteins were digested on-bead by treatment with 1 L trypsin (Promega, 20 g dissolved in 100 L 50 mM HEPES) with vigorous shaking at 37 °C overnight. The samples were briefly centrifuged and 10 L of the supernatant from each sample was TMT-labelled by combining with 10 L of the appropriate TMT10plex™ Isobaric Mass Tag Labelling Reagent (Thermo Scientific) dissolved in acetonitrile (8 mg mL -1 ) with moderate shaking for 2 h at rt (see Extended Data 2 for TMT label used for each sample). TMT-labelling was quenched by the addition of 1.1 L of 5% (w/v) hydroxylamine and the samples from each TMT set were combined into one "plex" solution. These samples were evaporated to dryness.

6-layer fractionation
Each stage tip was prepared by cutting 3× polystyrene-divinylbenzene copolymer modified with sulfonic acid (SCX) disks (3M) and using these to plug a p200 pipette tip. The tip was activated with 150 L MeCN by centrifugation (3000 × g, 3 min) then equilibrated with 150 L H2O by centrifugation (3000 × g, 3 min). Samples were dissolved in 1% (v/v) aqueous trifluoroacetic acid and each sample transferred to a stage tip. Peptides were loaded onto the SCX column by centrifugation as above. Peptides were desalted by centrifugation with 3× 60 L of 0.2% (v/v) trifluoroacetic acid. Peptides were then liberated from the column sequentially by centrifugation with 60 L of each of the following buffers into separate Lo-Bind Eppendorfs: All fractions of each sample were evaporated to dryness and stored at -80 °C.

LC-MS/MS methodology
Samples were rehydrated in 0.5% (v/v) formic acid, 2% (v/v) UPLC grade MeCN in Optima™ LC/MS H2O (Fisher Scientific) and dissolved completely by vortexing and sonication. Samples were filtered through 3x Durapore® membrane filters (Millipore) plugged into a p20 pipette tip by centrifuging the samples through the filters (4000 × g, 5 min) into a mass spectrometry vial. Samples were stored at 4 °C until ready for analysis.
Peptides were separated on an EASY-Spray™ Acclaim PepMap C18 column (50 cm × 75 m inner diameter, Thermo Fisher Scientific) using a 3-hour linear gradient separation of 0-100% solvent B (80% MeCN supplemented with 0.1% formic acid): solvent A (2% MeCN supplemented with 0.1% formic acid) at a flow rate of 250 nL min -1 . The liquid chromatography was coupled to a QExactive mass spectrometer via an easy-spray source (Thermo Fisher Scientific) which operated in data-dependent mode with survey scans acquired at a resolution of 70,000 at m/z 200. Scans were acquired from 350 to 1800 m/z. Up to 10 of the most abundant isotope patterns with charge +2 or higher from the survey scan were selected with an isolation window of 1.6 m/z and fragmented by HCD with normalized collision energy of 25. The maximum ion injection times for the survey scan and the MS/MS scans (acquired with a resolution of 35,000 at m/z 200) were 20 and 120 ms, respectively. The ion target value for MS was set to 10 6 and for MS/MS to 10 5 , and the intensity threshold was set to 8.3 × 10 2 .

Data analysis
Peptide searches were performed in MaxQuant version 1.6.0.2. Under group-specific parameters and type, reporter ion MS2 was selected, and the appropriate TMT10plex™ isobaric labels selected for both lysines and N-termini. For all experiments, oxidation (M) and acetyl (protein N-term) were set as variable modifications, carbamidomethyl (C) was set as a fixed modification, trypsin/P was set as the digestion mode, re-quantify and match between runs were selected, and up to date UniProt FASTA files for the human proteome and contaminants databases were used.
Data analysis was performed in Perseus version 1.6.7.0. Reporter intensity corrected values were loaded into the matrix. Data was filtered by removing rows based on "only identified by site", "reverse", and "potential contaminant" columns. Data were log2 transformed and filtered by row, retaining those that had 2 valid values in each triplicate condition. TMT data were normalized by subtracting the mean of each row within each TMT "plex" (if appropriate) and subtracting the median of each column. Volcano plots were generated using a pairwise Student's T-Test and the cut-offs generated using the false discovery rate (FDR) and S0 values indicated.

PARP6 auto-ADP ribosylation assay
This protocol was adapted from Hutin, Grimaldi and Matthews. 4 Briefly, reaction tubes containing 1× assay buffer (20×: 1 M Tris-HCl, pH 8.0, 4 mM DTT, 80 mM MgCl2), 25 M NAD + -biotin (added last), 0.3 M GST-PARP6, appropriate concentration of inhibitor (from 10× stock (1% (v/v) DMSO)) were made up to final volume with H2O, substituting for appropriate controls. The reactions were shaken on ice for 30 min, quenched with 4× sample loading buffer and boiled at 95 °C for 5 min. The samples were separated on to two SDS-PAGE gels and transferred to nitrocellulose. Each membrane was blotted separately using NeutrAvidin-HRP and anti-PARP6 (total protein).

Plasmid production
All microbiology work was carried out in a work area sterilized with 70% (v/v) EtOH and in the presence of an open flame. All equipment and media was either bought sterile or sterilized by autoclave. DNA concentration was measured using a NanoDrop OneC (Thermo Scientific). Sequencing was performed by Genewiz.

Mammalian cell uptake
Transfections were performed with Lipofectamine® LTX with Plus™ Reagent according to standard manufacturer's protocol. Briefly, MDA-MB-468 cells were seeded in sterile treated 6well plates and grown to confluency. Transfections were performed at 1:3 DNA:Lipofectamine with 2 g DNA, incubating for 24 h. Cells were then treated and irradiated as described above and analyzed by in-gel fluorescence and western blot.

Lysate-based photocrosslinking assay
Native MDA-MB-468 lysates were generated by trypsination of one 75 cm 2 flask of cells as described above. Trypsin was quenched and removed by aspiration after centrifugation (200 × g, 5 min). The cells were washed similarly in PBS and resuspended in cold PBS (250 L). Cells were lysed on ice by probe sonication (20% amplitude, 18 s (3 s pulse, 3 s rest)) and centrifuged to remove cell debris. Protein concentration was determined as above and the lysate snap frozen in liquid N2 and stored at -80 °C. Lysates were thawed on ice before use.
In a clear 96-well plate, wells containing 2 mg mL -1 native MDA-MB-468 lysate, 0.1 M GST-PARP6 (or blank storage buffer), appropriate concentration of PARPYnD/AZ0108 (from 100× stock (10% (v/v) DMSO)added last) were made up to final volume with PBS and incubated on ice for 30 min in the dark. The plate was irradiated with UV light for 5 min (365 nm monochromatic) and each sample transferred to Eppendorfs. The proteins were precipitated as described above to remove Tris (in GST-PARP6 storage buffer) which would otherwise inhibit the click reaction. Proteins were clicked to AzTB and enriched as described above, and the results analyzed by gel and western blot.

Variation with Prescission™ Protease
When also treating the samples with PreScission™ Preotease, samples were made up with the cleavage buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, pH 7.0) rather than PBS, and GST-PARP6 was added to a concentration of 0.3 M. Cleavage was performed using 1 U of enzyme overnight with moderate shaking at 4 °C. Samples were irradiated and prepared for gel-based analysis as described above.

Supporting Schemes
Scheme S1 -Synthesis of PARPYnD. Figure S1. PARP in vitro binding assays. Dose-response curves for PARPYnD tested in fluorescence anisotropy competition assay with various recombinant PARP enzymes. N, number of biological replicatesdata displayed ±SEM of these replicates, fit to 4 parameter dose-response function. PARP1 IC50 = 0.038 M, PARP2 IC50 = 0.006 M, PARP6 IC50 = 0.230 M. Raw data can be found in Extended Data 1.       Labelled proteins were enriched, identified and compared pairwise as previously described (Figure 3).

Supporting Figures
(E) PARPYnD (1 M) +UV vs. DMSO +UV (S0 = 1, FDR = 0.01). In-gel fluorescence of PARPYnD labelled lysate with PARP6 ±cleavage with PreScission™ Protease. GST tag was found to have no effect on binding and/or labelling. Cleaving the tag off after labelling (lanes 4-6) shows that the 71 kDa PARP6 portion of the protein is labelled (*). Cleaving the tag off before labelling (lanes 7-9) shows that the GST tag does not bias the recombinant protein towards compound binding.