Chemicals and reagents: Human AngiotensinII (AngII), Bradykinin (BK), Isoproterenol (ISO), epidermal growth factor (EGF), PD184352, Wortmannin, 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), Doxorubicin hydrochloride, poly-L-lysine hydrobromide and poly-L-ornithine hydrochloride were purchased from Sigma Aldrich. UBO-QIC from Cedarlane. Dulbecco’s modified Eagles medium (DMEM), fetal bovine serum (FBS), Phosphate Buffered Saline (PBS) and gentamicin were purchased from Gibco, Life Technologies. Coelenterazine 400a (DeepBlue C) and Coelenterazine H were purchased from Nanolight Technology. Dimethyl sulfoxide (DMSO), Phenylmethyl-sulfonyl fluoride (PMSF), leupeptin, aprotinin, pepstatin A, NaF and EDTA were from BioShop. Glutathione Sepharose™ 4B was from GE healthcare. Linear polyethylenimine 25-kDa (PEI) was from Polysciences. The phospho-p44/42 MAPK (ERK1/2) (Thr202/Tyr204) (E10) (#9106), p44/42 MAPK (ERK1/2) (#9102), phospho-Akt (Thr308) (#9275), Akt (pan) (C67E7) (#4691), Ras (#3965) and RhoA (#2117) antibodies were purchased from Cell Signaling Technology. The anti-HA-Peroxidase (3F10) (#12013819001), anti-FLAG (#F7425) and anti-c-Myc (clone 9E10) (#M4439) antibodies were purchased from Sigma Aldrich. Anti-mouse and anti-rabbit IgG HRP from BioRad. The β-actin (C4) (#sc-47778), H-Ras (#sc-520) and ARF6 (#sc-7971) antibodies were from Santa Cruz Biotechnology. SOS1 Protein (ExD Exchange Domain, aa564-1049, 6xHis tag) (#CS-GE02) and Mant-GTP exchange buffer (2X) (#EB01 from BK100 kit) were purchased from Cytoskeleton Inc.
Compounds acquisition: Compounds 21 (Rasarfin) (CID: 1396167; Catalogue #001-728-363), 21.1 (CID: 2236635; Catalogue #001-615-578), 21.2 (CID: 2238454; Catalogue #001-009-312), 21.3 (CID: 2968266; Catalogue #001-728-365), 21.4 (CID: 1087127; Catalogue #001-728-361), 21.5 (CID: 1088362; Catalogue #001-629-834), 21.6 (CID: 1088375; Catalogue #001-629-837), 21.7 (CID: 2997077; Catalogue #002-020-863) and 21.8 (CID: 2944643; Catalogue #001-728-355) were purchased from MolPort and solubilized in 100% DMSO at a final stock concentration of 50 mM.
Plasmids and constructs: Plasmids encoding β-arrestin1-RlucII, AT1R-YFP 73, signal peptide-Flag tagged human AT1R (sp-Flag-AT1R) 74, β-arrestin2-YFP, HA-B2R, B2R-YFP 75, AT1R-RlucII, B2R-RlucII, β2AR-RlucII, HA-β2AR, rGFP-CAAX, rGFP-FYVE 76, β-arrestin2-RlucII, Gαi3-RlucII 77, Polycistronic Gαq sensor, Flag-Gβ1 22, GFP10-Gγ1 78, PKC and Rho sensors 31, β-arrestin/AP-2 sensor 26 and GST-Rhotekin-RBD 79 were previously described. FLAG-K-Ras-WT, FLAG-K-Ras-G12V, FLAG-SOS1cat-CAAX were kindly provided by Dr. Matthew Smith (Université de Montréal, Qc). GST-Raf1-RBD, GST-GGA3-PBD, HA-ARF6 and HA-ARF6-T27N were described in 69, 71, 80. Myc-Rac1-WT, Myc-Rac1-Q61L, Myc-Rac1-T17N were kindly provided by Dr. Serge Lemay (McGill University, Qc).
To generate RlucII-tagged GGA3 (1-316) domain, the GGA3 (1-316) cDNA was amplified by PCR primers using GST-GGA3(1-316) DNA as a template. The PCR product was subcloned into the NheI/HindIII sites of RlucII containing vector (βarr2-RlucII, 68 using Gibson assembly (New England Biolabs). Partial cDNAs of Raf and PAK were obtained by RT-PCR of HEK293 total RNA. For subcloning of the Ras-binding domain (RBD) of Raf, the RasBD was PCR amplified using the partial cDNA of Raf as a template and assembled into the NheI/HindIII sites of RlucII containing vector of βarr2-RlucII 68 using Gibson assembly. The CRIB domain of PAK1 was PCR amplified using the partial cDNA of PAK1 as a template and subcloned into the KpnI/AgeI sites of RlucII containing vector (PKN-RBD-RlucII, 31) using Gibson assembly. For the Akt sensor, the PH domain of Akt was RT-PCR amplified from HEK293SL cell’s mRNA as a template. The PCR product was reamplified with flanking sequences for the Gibson assembly. The final PCR product was assembled into NheI/HindIII sites of RlucII containing vector of βarr2-RlucII 68.
To generate pGEX-6P-1-H-Ras, the cDNA of H-Ras (full length) was obtained by RT-PCR of HEK293SL cells total RNA. PCR product was then re-amplified with flanking sequences for the Gibson assembly. The final PCR product was subcloned into the pGEX-6P-1 vector, kindly provided by Dr. Matthew Smith (Université de Montréal, Qc), into the BamHI and NotI sites using Gibson assembly.
The S17N, Y32A and Y40A substitutions in FLAG-K-Ras was generated by complementation PCR reaction, whereas the Y32A/Y40A substitution was generated by overlapping PCR amplification using the two PCR products for Y32A and Y40A as templates. All the final PCR products from the amplified complementing fragments were subcloned into XhoI/HindIII sites of FLAG-K-Ras vector using Gibson assembly. For generating the pGEX-6P-1-FLAG-K-Ras-WT and pGEX-6P-1-FLAG-K-Ras-Y32A, K-Ras-WT and K-Ras-Y32A DNA were PCR amplified and assembled into the BamHI/NotI sites of pGEX-6P-1 vector using Gibson assembly. All constructs were verified by DNA sequencing before use (McGill Genome Center).
Cell Culture: HEK293SL cells, characterized in 22, are a subclone derived from regular HEK293 cells (Ad5 transformed) selected in our lab and have been used in all experiments. These cells have a cobblestone appearance and show better adherence as compared with regular HEK293 and HEK293T cells, making them more amenable to microscopy and BRET experiments. The MDA-MD-231 breast cancer cells were previously described 71, 80. These cells were regularly tested for mycoplasma contamination (PCR Mycoplasma Detection kit, abm, BC, Canada) and cultured in DMEM supplemented with 10% FBS and 20 µg/ml of gentamycin. Cells were transiently transfected with conventional calcium phosphate methods or 25-kDa linear PEI (2:1 PEI/DNA ratio) methods.
Live cell imaging/Confocal Microscopy: One day before transfection, cells were seeded in 35-mm glass-bottom dishes (MatTek Corporation) at a density of 1 × 105 cells per dish. For the recordings of receptor internalization, HEK293 SL cells were transfected with 2 µg of AT1R-YFP. For the recordings of β-arrestin-2 recruitment to the receptor, HEK293SL cells were transfected with 50 ng of β-arrestin2-YFP and 250 ng of Flag-AT1R or HA-B2R. Forty-eight hours post-transfection, cells were serum starved, preincubated with DMSO (0.1% final concentration) or Rasarfin (50 µM) for 30 min at 37 °C. AT1R-expressing cells were stimulated with angiotensinII (AngII; 100 nM) and B2R-expressing cells were stimulated with bradykinin (BK; 1 uM) for 30 min (receptor-YFP) or 15 min (β-arrestin2-YFP). Cells were imaged with Zeiss LSM-510 and/or LSM-710 laser scanning confocal microscope. To detect YFP, UV laser was used with 405 nm excitation and BP 505–550 nm emission filter. Images (2048 × 2048) were collected using a 63x oil immersion lens.
BRET measurements: HEK293SL cells were seeded at a density of 7.5 × 105 cells per 100-mm dish and 24 hours later, transiently transfected as such. For receptor internalization experiments, cells were transfected with 0.12 µg AT1R-RlucII, B2R-RlucII or β2AR-RlucII and 0.48 µg of either rGFP-CAAX or rGFP-FYVE. For β-arrestin recruitment assays, cells were transfected with 0.48 µg of receptor-YFP along with 0.12 µg of β-arrestin-RlucII. For β-arrestin/AP-2 binding experiments, cells were transfected with 1 µg Flag-AT1R, 1 µg β2-Adaptin-YFP and 0.12 µg β-arrestin2-RlucII. For G protein activation, cells were transfected with 3 µg of sp-Flag-AT1R along with either 4.5 µg of the Gαq-polycistronic BRET sensor or 0.24 µg of the Gαi3-RlucII and 0.6 µg of GFP10-Gγ2 and Gβ1 sensors or 0.12 µg PKN-RBD-RlucII and 0.48 µg of rGFP-CAAX. For PKC activation, cells were transfected with 3 µg of sp-Flag-AT1R and 0.18 µg of the PKC sensor. For GTPase activation experiments, cells were transfected with 1 µg Flag-AT1R, 0.48 µg of rGFP-CAAX and either 0.12 µg of GGA3-PBD-RlucII, PAK-CRIB-RlucII or Raf-RBD-RlucII. For biosensor validation experiments, cells were additionally transfected with 500 ng FLAG-K-Ras-WT, Flag-KRas-G12V, Flag-KRas-SOS-CAAX, HA-ARF6-WT, HA-ARF6-T27N, Myc-Rac1-WT, Myc-Rac1-Q61L or Myc-Rac1-T17N. For PI3K/Akt activation experiments, cells were transfected with 0.48 µg of rGFP-CAAX and 0.12 µg of Akt (PH)-RlucII. After 18 h of transfection, the media was replaced and cells were divided for subsequent experiments. Cells were detached and seeded onto poly-L-ornithine-coated 96-well flat white bottom plates (BrandTech Scientific) at a density of 2.5 × 104 cells per well in media. The next day, cells were washed once with Tyrode’s buffer (140 mM NaCl, 2.7 mM KCl, 1 mM CaCl2, 12 mM NaHCO3, 5.6 mM D-glucose, 0.5 mM MgCl2, 0.37 mM NaH2PO4, 25 mM HEPES, pH 7.4) and left in Tyrode’s buffer. For kinetics of β-arrestin binding to receptor and AP-2 experiments, cells were serum starved, pretreated with 21 (50 µM) or Barbadin (100 µM) for 30 min, stimulated with 100 nM AngII and BRET signals were monitored at indicated times using a Victor X Light plate reader (PerkinElmer). Coelenterazine H (final concentrations of 5 µM) was added 3–5 min prior to BRET measurements. Filter set was 460/80 nm and 535/30 nm for detecting the RlucII Renilla luciferase (donor) and YFP (acceptor) light emissions, respectively. The BRET ratio was determined by calculating the ratio of the light emitted by YFP over the light emitted by the RlucII. For G protein activation, cells were serum starved, pretreated with compounds for 30 min, stimulated with 100 nM AngII for 2 min (Gαq, Gαi3 and Rho sensors) or 5 min (PKC sensor). For the kinetics of GTPase and PI3K/Akt activation, cells were serum starved, pretreated with 21 (50 µM) or W for 30 min, stimulated with ligand (100 nM AngII or 100 ng/ml EGF) and BRET signals were monitored at indicated times. For concentration-response curves, cells were serum-starved, pretreated with various concentrations of compounds for 30 min and stimulated with ligand (100 nM AngII, 1 µM BK, 10 µM Isoproterenol (Iso) or 100 ng/ml EGF) in Tyrode’s buffer for 30 min for receptor internalization, 10 min for ARF6, 2 min for Rac and Rho activation and 5 min for Ras activation and PI3K/Akt activation. BRET signals were monitored using a Synergy2 (BioTek) microplate reader and coelenterazine 400a (final concentrations of 5 µM) added 3–5 min prior to BRET measurements. Filter set was 410/80 nm and 515/30 nm for detecting the RlucII Renilla luciferase (donor) and rGFP (acceptor) light emissions, respectively. The BRET ratio was determined by calculating the ratio of the light emitted by rGFP over the light emitted by the RlucII.
Western Blot analysis: HEK293SL cells (105 cells per well) were seeded in a poly-L-lysine-coated 6-well plate and transiently transfected with 3 µg Flag-AT1R, HA-B2R or HA-β2AR and/or 500 ng Flag-MEK1-WT, Flag-MEK1-DD, Flag-BRAF-WT, Flag-BRAF-V600E, Flag-K-Ras-WT, Flag-K-Ras-G12V, HA-ARF6-WT or HA-ARF6-T27N. Forty-eight hours post transfection, in a 37 °C water bath, cells were serum-starved for 30 min, pretreated with DMSO or Rasarfin (50 µM or at indicated concentrations) for 30 min, then stimulated or not with the indicated ligand [AngII (1 µM), BK (1 µM), Iso (10 µM) or EGF ( 100 ng/ml)] at indicated times. Cells were put on ice, washed with PBS and solubilized in 2x laemmli buffer (250 mM Tris–HCl pH 6.8, 2% SDS (w/v), 10% glycerol (v/v), 0.01% bromophenol blue (w/v) and 5% b-mercaptoethanol (v/v)) by heating at 65 °C for 15 min. Lysates were resolved on 10% or 14% SDS-PAGE, transferred to nitrocellulose membranes and immunoblotted for p-ERK1/2, ERK1/2, p-Akt, Akt, HA or FLAG. ImageLab 5.2 software was used to quantify the digital blots as fold of the phosphorylated protein over total protein, which were then normalized as indicated in the figure legends.
Purification of recombinant proteins: GST, GST-tagged Golgi Associated, Gamma Adaptin Ear Containing, ARF Binding Protein 3 Binding Domain (GST-GGA3-PBD), Raf1-Ras Binding Domain (GST-Raf1-RBD) and Rhotekin-Rho Binding Domain (GST-Rhotekin-RBD), as well as Ras proteins (pGEX-6P-1-H-Ras, pGEX-6P-1-K-Ras-WT and pGEX-6P-1-K-Ras-Y32A) were expressed in E. coli BL21 cells grown in LB medium and induced with isopropyl β-D-1-thiogalactopyranoside (IPTG) under the respective conditions: 1 mM for 1 hour at 37 °C (GST and GST-Raf1-BD), 0.6 mM for 3–4 hours at 30 °C (GST-Rhotekin-RBD and GST-GGA3-PBD) or 0.12 mM for 16 hours at 15 °C (pGEX-6P-1-H-Ras, pGEX-6P-1-K-Ras-WT and pGEX-6P-1-K-Ras-Y32A). The GST fusion proteins were purified using Glutathione Sepharose 4B as previously described by 81.
Glutathione S-transferase (GST) pull-down assays: Activation of ARF1, ARF6, Ras, and Rho were assessed by using GST pull-down assays. HEK293SL cells were transiently transfected with 3 µg AT1R-flag only (Ras and Rho) or along with 500 ng HA-ARF6 (ARF6) or HA-ARF1 (ARF1). Forty-eight hours later, the HEK293 or MDA-MB-231 cells were serum-starved for 4 hours with DMEM containing 20 mM HEPES then pretreated with DMSO or Rasarfin (50 µM) for 30 min. Cells were then stimulated with 1 µM AngII or 100 ng/ml EGF for 0 and 2 min. Cells were then washed once with ice-cold PBS and lysed for 30 min at 4 °C in 300 µL of lysis buffer (pH 7.5, 50 mM Tris-HCl, 140 mM NaCl, 5 mM MgCl2, 10% glycerol, 1% Nonidet P-40, 1 mM dithiothreitol) supplemented with protease inhibitors (1 mM phenylmethyl-sulfonyl fluoride (PMSF), 10 µg/ml leupeptin, 5 µg/ml aprotinin, 1 µg/ml pepstatin A) and phosphatase inhibitors (20 mM NaF, 0.025 mM sodium pervanadate). The samples were cleared by centrifugation and 30 µL (cell lysates) was kept for assesing total protein contents. The remaining was transferred to fresh tubes with 20 µg of either GST, GST-GGA3-PBD, GST-Raf1-RBD or GST-Rhotekin-RBD coupled to glutathione resin and rotated for 1–2 h at 4 °C. Beads were washed twice with lysis buffer and proteins were eluted in 25 µl 2x laemmli buffer by heating at 65 °C for 15 min. Proteins were resolved on 14% SDS-PAGE, transferred to nitrocellulose membranes and immunoblotted for HA, Ras, RhoA, H-Ras or ARF6. ImageLab 5.2 software was used to quantify the digital blots as fold of the amount of pulled down protein over total protein.
GTP Exchange Factor (GEF) assay: Ras activity was assessed in a 384-well black bottom plate using the mant-GTP exchange factor assay that measures the uptake of the fluorescent nucleotide analog N-methylanthraniloyl-GTP (mant-GTP) into GTPases. For Ras activation, 2X exchange buffer, 1.66 µg of purified pGEX-6P-1-H-Ras or pGEX-6P-1-K-Ras-WT or pGEX-6P-1-K-Ras-Y32A were added per well in presence of DMSO, different concentrations of Rasarfin (as indicated) or 50 µM of Compounds 21.4, 21.7 or 21.8. Using the Infinite 200 Pro plate reader (Tecan) with filters set at 360 nm and 440 nm for detecting the excitation and emissions, respectively, and temperature set at 20 °C, 5 readings were recorded every 5 sec before the addition of either H2O, 40 mM EDTA or 0.66 µg purified SOS1. The fluoresence of mant-GTP uptake was measured every 30 sec for 30 min and quantified as the delta relative florescence units (RFU), which was calculated as the RFU post-addition minus the 5 averaged RFU pre-addition, per condition.
Computational Site Finding and Docking of Rasarfin and Compound 21.4: SOS1 (chain S) and water molecules were deleted from the SOS-Ras X-ray structure (PDB 1BKD). The Ras-structure (chain R) was protonated and charged accordingly using Structure Preparation in Molecular Operating Environment (MOE) 2019.01.04 (Chemical Computing Group ULC, Montreal, QC, Canada, 2019). The Site Finder Tool in MOE suggests a binding site of 47 side chain contact atoms on the protein. The selected binding site has a PLB (propensity of ligand binding) 82 of 1.77 and displays 25 hydrophobic contact atoms on the protein. Dummy atoms were used as a binding site reference for the flexible protein docking procedure. Triangle Placement was carried out using Triangle Matcher, which allows 300 seconds for ligand placement and a maximum of 1000 poses returned for the compound. London dG was used for the Scoring as it estimates the free Energy of binding for each pose. Post-placement refinement was carried out employing Induced Fit of the Receptor, and the Force Field based Scoring Function GBVI/WSA dG for final Scoring, retrieving 5 docking poses for each compound. We evaluated the docking poses according to their docking scores and visual inspection of ligand-protein interactions. We selected the highest scoring pose which resembled best the feature similarity to the protein-protein interactions of SOS and Ras.
Molecular Dynamics Simulations: The evaluated output utput of the docking was the starting conformation for the MD simulation. A cubic water box was used with a margin of 12 and the protein-compound system centered in that box. 12027 water molecules and 8 Na ions were added for Rasarfin simulations, while 12046 water molecules and 8 Na ions were added for the MD simulation of 21.4. NAMD 2.12 was employed for the Molecular Dynamics simulation, AMBER10 force field was employed for the protein and ETH was utilized for the small molecule (Rasarfin and 21.4). Checkpoints were set to every 500 ps, Sample time to 0.25 ps, Time step: 2 fs. Particle Mesh Ewald implementation was enabled for periodic electrostatic. Cell periodicity wrap all atoms was applied. System was minimized for 100 ps, equilibration was run for 200 ps at 300 K. The production run was run at 300 K. Rasarfin was simulated for a total of 200 ns. One simulation of 100 ns was carried out and furthermore, ten 10 ns simulations were calculated. Five independent 10 ns simulations were computed for 21.4.
MD Pharmacophore docking analysis: The starting coordinates and trajectories of the protein and the respective compound bound to the protein were imported into LigandScout 4.4.1 (Inte:Ligand GmbH) 83 using a stride of 50. The “MD Pharmacophores” tool was employed to create dynamic pharmacophores to show protein-compound interaction patterns and to furthermore calculate their frequency. Calculated interactions were based on the following distances and angles: Hydrophobic interactions: 1–5 Å. H-bond donor/acceptor: 2.5–3.8 Å; Angle tolerance of 180° for sp3 hybridized atoms is an ideal hydrogen bond, which is broken when the angle difference exceeds 34° in either direction around the central position (angle tolerance of 50° is allowed for sp2 hybridized atoms); Angle tolerance of 60° for pi-cation interactions; Angle tolerance of 20° for orthogonal pi-pi interactions and 20° for parallel pi-pi interactions; Aromatic interactions: 0.0–2.0 Å orthogonal/parallel center deviation (minimum and maximum distance of two orthogonal or parallel plane feature center points). Once docked, the final output displays the number of unique pharmacophores, appearance frequency and Feature Timeline of computed interactions between the compound and the protein.
Cell proliferation assay: Cell proliferation was measured by a label-free, non-invasive cellular confluence assay using IncuCyte Live-Cell Imaging Systems (Essen Bioscience, Ann Arbor, MI, USA). MBA-MD-231 breast cancer cells (1,500 cells in 150 µl/well) were seeded overnight on a clear 96-well plate. 24 h later, the cells were treated with 0.1% DMSO, or different concentrations of Rasarfin (as indicated), or 10 µM of 21.4 or 21.8 for 96 hours total. The plate was placed in an XL-3 incubation chamber maintained at 37 °C and the cells were photographed using a 4x objective, every 3 hours for 4 days. Cell confluence was calculated using IncuCyte S3 2019 software and cell proliferation was expressed as an increase in percentage of confluence. Experiments were done in triplicate and repeated three times.
Cell viability assays: MBA-MD-231 breast cancer cells (1,500 cells in 150 µl/well) were seeded overnight on a clear 96-well plate. 24 h later, the cells were treated with 0.1% DMSO, or different concentrations of Rasarfin (as indicated), or 10 µM of 21.4, 21.8 or Doxorubicin for the indicated times at 37 °C. At each time point, 10 µl of 5 mg/ml MTT was added to each well and the plates were incubated for an additional 4 h at 37 °C. The absorbance at 590 nm was measured by plate reader. Experiments were done in triplicate and repeated three times.
Data analysis: Statistical analyses were performed using GraphPad Prism 6 software (GraphPad Software Inc.; La Jolla, CA) using either Student’s t-tests, two-way ANOVAs, or Dunnett’s comparisons tests, when appropriate and as indicted in the figure legends. Curves presented throughout this study were generated using GraphPad Prism software and represent the best fits, from which IC50s were calculated. P values < 0.05 were considered significant.