Cell culture, cellular activation and determination of cell numbers

The MDA-MB-231 cell line (ATCC, Teddington, UK) was used due to the established and reliable patterns of TF and PAR2 expression and the response to PAR2 activation [20]. Initial analysis also indicated the presence of MAGI proteins in these cells [21]. The cells were cultured in DMEM (Lonza, Cambridge, UK) and supplemented with foetal calf serum 10% (v/v). Cells were seeded out as described and activated by incubation with PAR2-agonist peptide (PAR2-AP; SLIGKV; 20 µM) for durations stated in the results section. Cell numbers were determined by crystal violet staining as previously described [22] and calculated from a standard curve.

Prediction of the PDZ binding domain within the cytoplasmic domain of TF and identification of potential binding proteins

The analysis of the possible PDZ binding domains within the cytoplasmic domain of TF was carried out using the POW web-site program and the scores for possible binding proteins determined [23].

Duolink Proximity Ligation (PLA) assay

The procedure was carried out using the Duolink reagents (Sigma Chemical Co. Ltd, Poole, UK) and adapted from that previously described in detail [24] with minor modifications. MDA-MB-231 cells (103) were seeded out into 35 mm-glass based μ-dishes (InVitro Scientific/Cellvis, Sunnyvale, USA) and adapted to serum-free medium for 1 h prior to activation. Sets of cells were then incubated with PAR2-AP (20 μM) for up to 40 min and then fixed with 4% (v/v) paraformaldehyde for 15 min. The cells were then washed three times with PBS and permeabilised with Triton X-100 0.1% (v/v) in PBS, for 5 min. All samples were blocked with Duolink blocking buffer for 1 h and incubated overnight with combinations of antibodies at 4 °C. In order to assess the proximity and potential interaction between TF and MAGI1-3, a mouse anti-TF antibody (HTF1; 5 μg/ml; eBioscience/Thermo Scientific, Warrington, UK) was used together with a rabbit anti-MAGI1 antibody (H-70; 2 μg/ml; Santa Cruz Biotechnology, Heidelberg, Germany), a rabbit anti-MAGI2 antibody (2 μg/ml; GenTex/Insight Biotechnologies, Wembley, UK) or a rabbit anti-MAGI3 antibody (2 μg/ml; Novus/R&D Systems, Abingdon, UK). In other experiments, the proximity between TF and the overexpressed PDZ-domains of MAGI1 were analysed using a rabbit anti-HA tag (C2954; Cell Signalling Technologies/New England Biolabs, Hitchin, UK) which was used at 1 µg/ml. The antibodies were diluted in the provided antibody diluent and the samples were blocked with the provided blocking buffer. A positive control was prepared using mouse anti-TF antibody (HTF1; 5 μg/ml) and a rabbit polyclonal anti-TF antibody (FL295; 5 μg/ml; Santa Cruz Biotechnology). Additionally, as negative controls, the antibodies were substituted with rabbit or mouse IgG isotypes (New England Biolabs; 2 μg/ml and 5 µg/ml respectively). The cells were washed three times with PBS and PLA performed according to the manufacturer’s instructions. The cells were then stained with DAPI (2 μg/ml) and Phalloidin-FITC (2 µg/ml). Images were acquired using a Zeiss Axio Vert.A1 inverted fluorescence microscope with a × 40 magnification (Carl Zeiss Ltd, Welwyn Garden City, UK). The number of red fluorescent events and nuclei were determined using ImageJ, in 10 fields of view from each assay [24].

Co-immunoprecipitation of proteins, SDS-PAGE and western blot analysis

Cells were lysed in PhosphoSafe Extraction Reagent (Sigma) containing a protease inhibitor cocktail and cell debris removed by centrifugation. TF and MAGI1 proteins were immunoprecipitated from cell lysates (2 × 105) using the anti-TF (HTF-1; 4 µg; eBioscience/Thermo Scientific, Warrington, UK) and anti-MAGI1 (H-70; 4 µg; Santa Cruz) antibodies respectively. To ensure specificity, relevant IgG isotypes (4 µg) were also included as well as additional controls without any antibody. All samples were incubated at 4 °C overnight with gentle shaking. Pureproteome protein A-magnetic beads (10 µl) (Merck-Millipore, Watford, UK) was added to all samples and controls and incubated at 4 °C for 90 min. The samples were then placed in a magnetic stand and the supernatant removed, washed five times with PBST (1 ml) and the samples denatured in SDS-PAGE loading buffer (70 µl) (Sigma). The samples were separated by 12% (w/v) SDS-PAGE, transferred onto nitrocellulose membranes and then blocked with TBST (10 mM Tris–HCl pH 7.4, 150 mM NaCl, 0.05% Tween-20). When immunoprecipitating with anti-TF, the membranes were probed using anti-MAGI1 (H-70) and a polyclonal rabbit anti-TF antibody (FL295; Santa Cruz) diluted 1:4000 (v/v) in TBST. When immunoprecipitating with anti-MAGI1, the membranes were probed with an anti-TF antibody (HTF1) and a mouse anti-MAGI1 antibody (SS-5; Santa Cruz). In some experiments, immunoprecipitation was carried out from transfected MDA-MB-231 cells using a rabbit anti-HA antibody (C2954; 4 µg; Cell Signalling Technologies/New England Biolabs) and the presence of TF examined using the anti-TF antibody (HTF1). The membranes were then washed three times with TBST and probed with a goat anti-mouse or goat-anti-rabbit alkaline phosphatase-conjugated antibody (Santa Cruz Biotechnology), diluted 1:4000 (v/v), and incubated for 90 min. Bands were then visualised using the Western Blue stabilised alkaline phosphatase-substrate (Promega Corp., Southampton, UK), recorded and analysed using ImageJ program.

Pull-down assay using immobilised TF peptides

The plate pull-down procedure used was an adaptation of the method described and confirmed previously [25]. Substrate peptides, corresponding to the last 18 amino acids of the cytoplasmic domain of TF were synthesised in biotinylated form (Biomatik, Ontario, Canada). In total 4 peptides were synthesised in non-phosphorylated, single-phosphorylated (phospho-253 or phospho-258) and double-phosphorylated (phospho-253 and 258) forms and used as bait. A scrambled peptide (biotin-SWGNVSKLSAPRQGVNKE) was also included alongside as a negative control. TF peptides (5 µM final concentration) were diluted with PBS and distributed (100 µl per well) into a NeutrAvidin-coated 96-well plate (Thermo Scientific, Warrington, UK) and incubated for 2 h at room temperature to allow binding. The wells were then washed four times with PBST (300 µl). Cell lysates from non-activated and PAR2-activated cells (100 µl from 2 × 105 cells) were incubated in the plates for 1 h at room temperature. The wells were then washed four times and probed with rabbit anti-MAGI1 (H-70; Santa Cruz) or mouse anti-MAGI3 (46; Santa Cruz) antibodies diluted 1:200 (v/v) in PBST. The samples were detected using goat anti-rabbit and goat anti-mouse horseradish peroxidase-conjugated antibodies (Santa Cruz), diluted 1:200 (v/v) and the colour developed with TMB One Solution (100 µl) (Promega). Once the colour was developed the reactions were stopped by the addition of 2 M sulphuric acid (50 µl) and absorptions measured at 450 nm using a plate reader.

Cloning, transfection and expression of PDZ domains of MAGI1

The cloned cDNA for MAGI1 (pDONR223-MAGI1) was obtained from Addgene (Cambridge, UK) and the regions corresponding to PDZ1-5 were separately amplified using the forward and reverse primers as shown in Supplementary Table 1. Following digestion with BamHI and HindIII, the cDNA was sub-cloned into the FLAG-HA-pcDNA3.1 (Addgene) and positive colonies confirmed by sequencing (Eurofin-MWG, Wolverhampton, UK). In addition, two DNA regions encompassing the N-terminal of MAGI1, including or excluding PDZ1, were also cloned as above, using the primers shown in Supplementary Table 1. The sequences of the cloned DNA plasmid constructs were then verified as above. MDA-MB-231 cells (2 × 105) were seeded out into 48-well plates and transfected with 0.5 µg of the plasmids to express the individual PDZ domains, or the two longer N-terminal-spanning regions. Transfection of the cells was carried out using TransIT-2020 (Geneflow, Lichfield, UK) according to the manufacturer's instructions. Cells were permitted to express the proteins for 48 h and the expression of the proteins were confirmed by western blot using rabbit anti-HA antibody. The associations of the overexpressed MAGI1-PDZ domains, or the MAGI-N-terminal domains with the endogenous TF were then analysed in situ, by PLA. An attempt to sub-clone the entire MAGI1 was unsuccessful due to the large size of the insert DNA.

Additionally, in an attempt to suppress the expression of MAGI1, the cells were transfected with a specific siRNA for MAGI1 (100 nM; SilencerSelect siRNA, ThermoFisher UK, Loughborough, UK) using TransIT-2020 reagent, according to the manufacturer’s instructions. The cells were incubated up to 48 h but were not viable for experimentation.

Finally, in some experiments, cells were transfected with a pCMV-AC-TF-tGFP plasmid (OriGene, Rockville, USA) which encodes for turbo-GFP (tGFP) in tandem, on the C-terminal of TF which was mutated to substitute aspartate instead of serine-253 [26]. Cells (2 × 105) were transfected with 1 μg of plasmid DNA as above and permitted to express the protein over 48 h. The cells were then lysed and the TF-tGFP immunoprecipitated with anti-tGFP-magnetic beads (clone 2H8) (25 μl) according to the manufacturer's protocol (OriGene) at 4 °C. The samples were then examined as above for the presence of MAGI1.

Cell-based thrombin-generation assay

Cell surface TF-fVIIa activity was measured by modification of a previously described procedure [25]. In some experiments, to confirm that the activity was attributed to TF, the cells were pre-incubated with an inhibitory anti-TF antibody (HTF1; 20 µg/ml) before analysis. All cells (105) were washed with the reaction buffer (Tris-buffered saline (TBS) pH 7.4), containing 1% (w/v) bovine serum albumin (BSA) and then incubated for 20 min at 37 °C with a mixture of barium sulphate absorbable proteins (2 mg/ml; Sigma Chemical Company Ltd, Poole, UK) and 5 mM CaCl2 in the reaction buffer (total volume 150 µl). Aliquots of reaction (100 µl) were then transferred to 96-well plates containing 100 µl of thrombin substrate CS-01(38) (0.2 µM H–D-Phe-Pip-Arg-pNA; Hyphen BioMed/Quadratech, Epsom, UK) and incubated for a further 40 min at 37 °C. The reactions were stopped by the addition of 2% (v/v) acetic acid (50 µl) and the absorption of the sample measured at 410 nm on a plate reader. TF activity was determined from a standard curve prepared alongside.

Fluorescence microscopic examination of cellular TF and MAGI1

MDA-MB-231 cells were analysed by fluorescence microscopy, before and after activation with PAR2-AP. The cells were fixed and probed for TF using an FITC-conjugated anti-TF antibody (HTF1-FITC; Miltenyi Biotec, Woking, UK). The cells were also probed with a rabbit anti-MAGI1 antibody (H-70) and then labelled with a goat anti-rabbit IgG-AlexaFluor 594 antibody (Life Technologies, Paisley, UK). The samples were examined by fluorescence microscopy and the co-localisation overlap coefficient values determined.

Measurement of ERK1/2 phosphorylation, Akt phosphorylation, Cyclin D mRNA expression and cell proliferation

In order to determine the outcome of expression of the MAGI1 peptides, MDA-MB-231 cells were transfected to express the N-terminal regions with, or without the PDZ1 domain and the influence on various indicators of cell proliferation and survival were examined. In the first experiment, the cells were lysed and the phosphorylation of Akt was compared in the two transfected samples. The cells (105) were lysed in Laemmli buffer (Sigma) and separated on denaturing 12% (w/v) polyacrylamide gels. Proteins were transferred onto nitrocellulose membranes and blocked with TBST (10 mM Tris–HCl pH 7.4, 150 mM NaCl, 0.05% Tween-20) for 1 h. The membranes were probed with a goat anti-human Akt1/2 (N-19) polyclonal antibody and a rabbit polyclonal anti-human Akt1 (phospho-S473). The membranes were then washed three times with TBST and probed with a donkey anti-goat, or goat anti-rabbit alkaline phosphatase-conjugated antibody (Santa Cruz), diluted 1:4000 (v/v), and incubated for 90 min. Bands were then visualised using the Western Blue stabilised alkaline phosphatase-substrate (Promega), recorded and analysed using ImageJ program.

Next, cells (105) were transfected as above and one set was pre-incubated with a monoclonal antibody (SAM11; 20 µg/ml; Santa Cruz Biotechnology, Heidelberg, Germany) throughout culturing, in order to block PAR2 activation by TF activity. The cells were then lysed and proteins separated by SDS-PAGE as above. Western blot analysis of ERK1/2 phosphorylation in the samples was carried out using an anti-phosphoT202/185-phosphoY204/187-ERK1/2 antibody (Cell Signalling Technology/New England Biolabs, Hitching, UK) and total ERK1/2 was detected using an anti-ERK1/2 antibody (Cell Signalling/NEB) diluted 1:3000 (v/v) in TBST. GAPDH was also detected using a rabbit anti-GAPDH polyclonal antibody (V-18; Santa Cruz Biotechnology, Heidelberg, Germany) diluted 1:5000 (v/v) in TBST. The membranes were then probed with a goat anti-rabbit, or donkey anti-goat alkaline phosphatase-conjugated antibody (Santa Cruz), diluted 1:5000 (v/v) in TBST. Bands were then visualised using the Western Blue stabilised alkaline phosphatase-substrate, recorded and analysed using ImageJ program.

To further demonstrate the influence and the effectiveness of inclusion of PDZ1 within the N-terminal of MAGI1, cells were transfected as above and pre-incubated with a monoclonal antibody (SAM11; 20 µg/ml) capable of blocking PAR2 activation, or a mouse monoclonal antibody to inhibit the protease activity of TF-fVIIa complex (HTF1; 20 µg/ml; eBioscience/Thermo Scientific, Warrington, UK) [27] or a mouse IgG isotypes (New England Biolabs). Total RNA was isolated using the Monarch total RNA extraction kit (New England Biolabs, Hitchin, UK) from 105 cells. Samples (100 ng) were amplified using the primers 5’- CCG TCC ATG CGG AAG ATC -3’ (forward) and 5’- ATG GCC AGC GGG AAG AC -3’ (reverse) [28]. The reaction was carried out at an annealing temperature of 60 °C for 1 min using the GoTaq® 1-Step RT-qPCR System (Promega Corporation Ltd, Southampton, UK) on an iCycler thermal cycler (Bio-Rad, Hemel Hempstead, UK) for 40 cycles. Following amplification, the relative amounts of target mRNA were determined using the 2−ΔΔCT method [29]. Finally, cell numbers were determined using the crystal violet measurement as described above.

Statistical analysis

Presented data include the calculated mean values ± the calculated standard error of the mean. Statistical analysis was carried out using the GraphPad Prism version 9.0 (GraphPad Software, Boston, Massachusetts USA). Significance was determined using one-way ANOVA (analysis of variance) and Tukey’s honesty significance test or where appropriate, by paired t-test and p values of equal or less than 0.05 were deemed to be significant.