Cell lines and reagents
Human colon carcinoma HCT116wt, DLD-1 and HT29 cells were grown in RPMI 1640 medium (PAA Laboratories GmbH, Pasching Austria) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (PANTM Biotech GmbH, Germany) and kept at 37°C in a humidified incubator (95% air, 5%CO2). Cells were treated at 50% confluency with 40 or 60 μM TQ (Sigma-Aldrich) dissolved in DMSO (final concentration less than 0.1%) and collected at different time points. PAK1 inhibition was performed by preincubating the cells with 10 μM IPA-3 (Sigma-Aldrich) for 1 hour.
Peptide Array PepChip Kinomics v2
HCT116wt cells (1.5x105/ml) were treated with 40 μM TQ and collected after 24 hours. Proteins were extracted and shipped to manufacturer (Pepscan, Netherland) according to manufacturer protocol. Briefly, cell pellets were lysed (20 mM Tris–HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM beta-glycercophosphate, 1 mM sodium orthovanadate (Na3VO3), 1 mM NaF, Roche Complete protease inhibitor cocktail). After sonication, the lysates were spin down and snap frozen before being shipped to the manufacturer. The concentration of the samples was determined using Bradford Protein Assay (Bio-Rad, Hercules, CA). Samples were then hybridized on PepChip Kinomics v2 chip and the kinome profiling was performed by the Pepscan Company. The peptide array was done with samples obtained from two independent experiments and every sample was spotted two times on the array.
Array Analysis software Ingenuity pathway analysis (IPA)
Peptide array data was analyzed by Ingenuity pathway analysis software (IPA, version 9.0, Ingenuity © Systems, http://www.ingenuity.com, Mountain View, CA, USA). The data set contained upregulated proteins (fold-change ≥ 2 after 24 hours of 40 μM TQ treatment) identified by the pepscan peptide array and their protein ID number. Statistical significance of the different pathways obtained by IPA was calculated using a right-tailed Fisher’s Exact test . Pathways with a p–value ≤ 0.05 were selected. WebLogo v3 application (http://weblogo.threeplusone.com/) was used to analyze the phosphorylation consensus motifs of all the kinases and proteins identified by the array.
Pathway Annotation Clustering
The probable pathways by which the 50 candidate proteins, with fold-change ≥ 2 after 24 hours of 40 μM TQ treatment, could interfere were predicted using DAVID bioinformatics tool (http://david.abcc.ncifcrf.gov) [35, 36]. For pathway mapping, Uniprot IDs of differentially phosphorylated proteins were submitted to the DAVID web resource with colorectal cancer specific proteins from Human Protein Atlas (http://www.proteinatlas.org/) as background. Pathway predictions were made by including the subset pathway databases: KEGG, Panther, Biocarta, Reactome and BBID followed by pathway annotation clustering.
HCT116wt, DLD-1 and HT29 cells were grown to 50% confluency and treated with 40 μMTQ and harvested over different time points. The cells were then lysed with RIPA lysis buffer (150 mM NaCl, 0.5% DOC, 0.1% SDS, 1% NP40, 50 mM Tris pH8) containing a cocktail of protease inhibitors. The protein concentration of all samples was determined by by DC BioRad protein assay kit (BioRad Laboratories, Hercules, CA) using bovine serum albumin as standard. Total proteins (50 μg /sample) were separated on an 8–12% SDS polyacrylamide gel and transferred to nitrocellulose membrane by blotting. After blocking with 5% non-fat dry milk in TBST buffer, membranes were incubated with primary antibody at 4°C overnight, washed three times, with TBST buffer, and incubated again with the corresponding HRP-conjugated secondary antibody at room temperature for 1 h. The membranes were then washed with TBST buffer and protein bands were detected by enhanced chemiluminescence. PAK1, ERK1/2, pERK1/2 and PARP were purchased from Cell signaling; pPAKThr212 and pPAKSer144 from Abcam and pPAKThr423 from Abgent. Densitometric analysis was performed for all western blots using ImageJ 1.45 s software.
Cells treated with 60 μM TQ for 10, 45, 60 minutes and 24 hours were collected and lysed using RIPA buffer supplemented with proteases and phosphatases inhibitors. Immunoprecipitiation was performed using the Dynabeads Protein G magnetic separation KIT as per manufacturer protocol (Invitrogen). After incubating the beads with 600 μg of proteins, they were incubated with anti-PAK1 and anti-ERK1/2 antibodies. The precipitated lysates were then loaded into SDS-PAGE gels and immunoblotted against total PAK1 and ERK1/2.
Plasmids and siRNA transfections
Plasmids (PAK1wt, K299R, T212E, T212A, T423E) were a gift from Prof. Jonathan Chernoff (Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia). Cells seeded in 6-well plates were transfected at 90% confluency with 1 μg of the different plasmids for 6 hours, using Invitrogen Lipofectamine 2000 according to manufacturer instructions. Afterwards cells were treated with 100 μM TQ (due to the high cellular density) and collected after 24 hours. Cell lysates were then used for Western blotting analysis.
PAK1 and scrambled siRNA reagents were purchased from Thermo Fisher Scientific (Dharmacon RNAi Technologies). Lipofectamine RNAiMAX reagent was used to transfect HCT116 cells with 10 μM siRNA, according to manufacturer’s instructions. The next day the transfected cells were split and 50% confluent plates were treated with 60 μM TQ. Cell lysates were then used for Western blotting analysis.
Structural analysis by docking
To understand the interaction between PAK1 and TQ, we modeled the structure of PAK1 and performed molecular docking with TQ using the Schrodinger suite (Maestro, version 9.3, Schrodinger, Inc, New York, NY, 2012). A model of the possible active state conformation of PAK1 was obtained by integrating the available crystal structures of the autoinhibitory domain 78-147(PDB ID: 1F3M)  and the kinase domain 250-542(PDB ID: 3Q52)  of PAK1 and a threaded model of the region corresponding to 148–249 using ITASSER [29, 30]. The possible ligand binding sites in PAK1 was analyzed using the SiteMap module (SiteMap, version 2.6, Schrödinger, Inc., New York, NY, 2012) of the Schrodinger suite. In order to analyze the binding mode of TQ near to Thr212 site, grids were generated focusing on Thr212 site and the interacting residues were identified. To study the interaction details of PAK1 and ERK2, the crystal structure of ERK2 available with PDB ID:2ERK  was docked to the TQ bound and unbound conformations of PAK1 using ClusPro [37, 38]. The energies of the complexes were determined through energy minimization using AMBER11  . Crystal structure of the autoinhibited dimer of PAK1 (PDB ID: 1F3M) was used to understand the binding mode of IPA-3 on PAK1 . The structural coordinates of the residues (416–422), missing in the PDB, belong to the kinase activation segment of PAK1 and were modeled using Modeller9v12  . IPA-3 was then docked to the autoregulatory region of PAK1 (autoinhibited conformation). To analyze the combined effect of TQ and IPA-3 on PAK1, TQ was docked to the IPA-3 bound conformation of PAK1. All the dockings were performed using the extra precision mode of the Glide program of Schrodinger (Glide, version 5.8, Schrödinger, Inc., New York, NY, 2012). The structures of TQ and IPA-3 were obtained from the PubChem database. Protein and ligand structures for docking were prepared using Protein Preparation Wizard and LigPrep utilities of Schrodinger (LigPrep, version 2.5, Schrödinger, Inc., New York, NY, 2012). All renderings were done using Chimera 1.8 .
Kinase Activity assay
Multi-kinase ELISA array was performed as per the manufacturer protocol (Symansis). Briefly, Cells treated with 40 μM TQ were collected after 1, 3, 6, 24 hours and lysed with 1X denaturing cell lysis buffer (Symansis CLB001). 30 μg of proteins were loaded on pre-coated strips with antibodies corresponding to the investigated kinases. After several washing and hybridization steps, absorbance of each well was measured at 450 nm using VICTORTM X3 multilabel reader.
Cell Viability assay
Cellular viability was measured by crystal violet staining. HCT116wt (3.75 × 104/ml) cells were seeded in 96 well plate and treated with 40 μM TQ and/or 1-50 μM IPA-3 for 24 hours. The treated cells were washed once with PBS, fixed for 15 min in a crystal violet solution (0.5% crystal violet in 20% methanol) at room temperature, then washed twice with water and air-dried. The stained cells were solubilized with methanol for 15 min with mild agitation. Absorbance was measured at 595 nm using VICTORTM X3 multilabel reader. IC50 values were calculated using EXCEL 2010. Statistical analysis was performed using SPSS version 19. One tailed Student T-Test was performed by comparing 2 samples assuming that they have equal variances.
Annexin V/Propidium Iodide staining
Apoptosis was measured using Annexin V/PI co-staining. HCT116wt cells (1.5 × 105/ml) were pretreated with IPA-3 (10 μM) for 1 hour then treated with TQ (40 μM) for 24 hours. After collection, cells were centrifuged at 200 g for 5 min, 4°C and washed with 1X PBS. The pellet was resuspended in 100 μl Annexin-V-Fluos labeling solution (10 μl annexin reagent and 10 μl PI solution in 150 μl incubation buffer (according to manufacturer, Roche). The samples were incubated for 7 min in the dark, at room temperature then 100 μl incubation buffer was added. The cellular fluorescence was then measured using a Fluorescence Activated Cell Sorter (FACS) flow cytometer (BDFACS CantoTM II). Each sample was collected as 20,000 ungated events and the different cell populations were determined using FlowJo software (FlowJo7.6.5).
Immunohistochemistry (IHC) was performed on available paraffin fixed tissue blocks from a xenograft experiment  to detect the expression of pERK1/2. Rehydration of tissue sections was performed in descending concentrations solutions of ethanol (96% to 70%). Antigen was retrieved by heating in a pressure cooker (1 mmol/L Tris-EDTA buffer, 120°C, 5 min). The slices were incubated in blocking solution (Dako, Glostrup, Denmark) to prevent nonspecific binding sites. Next pERK1/2 (1:2000) primary antibody was added to the slices and incubated for 30 min at room temperature. The sections were then washed with washing buffer (Dako) and incubated with EnVision + System horseradish peroxidase-linked secondary antibody (goat anti-rabbit, Dako) at room temperature for 30 min. Positive immunoreactivity was detected using diaminobenzidine + (Dako). Positive and negative IHC controls were included in this study. Percentage and intensity of positively stained epithelial cells (cytoplasmic versus nuclei) was quantified and scored by an expert pathologist (T.T.R.). Statistical analysis was performed using SPSS. Two tailed student t-test was performed by comparing 2 samples assuming that they have equal variances.