Differential Rac1 signalling by guanine nucleotide exchange factors implicates FLII in regulating Rac1-driven cell migration

The small GTPase Rac1 has been implicated in the formation and dissemination of tumours. Upon activation by guanine nucleotide exchange factors (GEFs), Rac1 associates with a variety of proteins in the cell thereby regulating various functions, including cell migration. However, activation of Rac1 can lead to opposing migratory phenotypes raising the possibility of exacerbating tumour progression when targeting Rac1 in a clinical setting. This calls for the identification of factors that influence Rac1-driven cell motility. Here we show that Tiam1 and P-Rex1, two Rac GEFs, promote Rac1 anti- and pro-migratory signalling cascades, respectively, through regulating the Rac1 interactome. In particular, we demonstrate that P-Rex1 stimulates migration through enhancing the interaction between Rac1 and the actin-remodelling protein flightless-1 homologue, to modulate cell contraction in a RhoA-ROCK-independent manner.

Hepatocyte Growth Factor (+ HGF) for the indicated times following serum starvation for 18-24 hours. Yellow boxes are used to highlight the differential effects of Tiam1 WT and P-Rex1 WT on cell scattering at indicated time points. Scale bar= 100 µm. (f) Quantification of cell scattering of + dox treated MDCKII cells described in (e). Graph represents percent of scattered or unscattered colonies from a total of 50 colonies per GEF construct from three independent experiments. (g) Lysates from MDCKII cells following + dox treatment for 24 hours to induce expression of indicated GEF constructs were subjected to PAK-CRIB pulldown. Levels of active and total Rac1 were detected by Western blot analysis. α-Tubulin was used as a loading control. Representative Western blot from three independent experiments.  Tandem Affinity Purification (SF-TAP). NIH3T3 cells transduced with the doxycycline (dox)-inducible expression system for SF-Rac1 alone (control) or together with Tiam1/P-Rex1 WT (Set1) or GEF* (Set2) were cultured in media containing Lysine (K) and Arginine (R) isotopes as outlined. In reverse SILAC (r) media labels were swapped as indicated. Following dox induction, lysates were subjected to SF-TAP and mixed eluates were analysed by mass spectrometry to identify Rac1 interactors and their abundance calculated as SILAC ratios relative to control cells. A cut off of ± 1.3 fold-change was utilised to infer differential Rac1 binding. SF-TAP of SF-Rac1 from MCF7 cells treated with 1 µg ml -1 dox to induce expression of SF Rac1 alone (control) or together with indicated GEFs. Co-precipitated endogenous FLII and RhoGDI1 were detected by Western blot analysis. α-Tubulin was used as a loading control. (g) Quantification of endogenous SF-Rac1-bound FLII in MCF7 cells described in (f) normalised to control cells ± s.e.m. from three independent experiments. In (e) and (g) Student's t-test was used to assess significance as indicated on graph. p-values indicated above each bar are relative to + dox treated control cells. NS= non-significant; *= p ≤ 0.05; **= p ≤ 0.01. SHG signal intensity from fibroblast-collagen matrices treated as described in (c). Graph represents the intensity of SHG signal produced by fibrillar collagen ± s.e.m. from three independent experiments. For (b), (c), (e), and (g) Student's t-test was used to assess significance as indicated on graph. p-values indicated above each bar are relative to + dox treated EV cells (b), -dox treated EV control (c, g) or to the NT control (e). NS= non-significant; *= p ≤ 0.05; **= p ≤ 0.01. Diaphanous-related formin-1 13 Kinesin-1 heavy chain 14 Myosin heavy chain 10 15 Myosin heavy chain 14 15 Myosin heavy chain 9 15 Myosin light chain 3 15 Myosin light polypeptide 6 15 Myosin regulatory light chain 12B 15 Myosin-Ic 15 Myosin-Id 15 Protein flightless-1 homolog 16 Ras-related C3 botulinum toxin substrate 3 17 Zyxin 18   27 and Mohammad et al. 28 .  3 L-Arginine monohydrochloride (Sigma-Aldrich, A5131) R6 stock solution 3 L-Arginine HCL 13 C (Silantes, 201204102) R10 stock solution 3 L-Arginine HCL 13

Supplementary Methods
Analysis of cell morphology. For NIH3T3 cells, given their mesenchymal phenotype, the three groups used for assessing GEF-induced changes in cell morphology were as follows: 1) mesenchymal, 2) epithelial-like and 3)

SILAC labelling
Amino acid stock solutions used to prepare the SILAC labelling media are outlined in Supplementary Table 4.
The different SILAC labelling media used in the different experiments are detailed in Supplementary Table 5.

SILAC amino acid incorporation check
After six doubling rounds a small batch of cells were lysed and amino acid incorporation was determined by examining the ratio of lysine and arginine labelled protein peptides versus the unlabelled portion using mass spectrometry. The amino acid incorporation rate was then calculated using the equation outlined below: 1 (Ratio H/L) + 1 1 -H= Labelled protein peptides L= Unlabelled protein peptides

Sample preparation
NIH3T3 cells transduced with the dox inducible system for expressing SF-Rac1 alone or together with the different GEF constructs were plated at 2 x 10 7 cells per 500 cm 2 plate with two plates per condition. After 48 hours cells were treated with 1 µg ml -1 doxycycline (+ dox) and incubated for an additional 48 hours. Cells were then subjected to SF-TAP. Following the FLAG® peptide elution step, the different SILAC labelled samples were mixed together and concentrated using a 4 ml Amicon filter unit (EMD Millipore, UFC800308). Concentrated mixed SF-TAP eluates were then subjected to mass spectrometry. For this, a third of each sample was resolved on 12 % NuPAGE®Novex® Bis-Tris pre-cast gels for 9 minutes. Gels were then stained with Coomassie Blue to visualise bands and cut as one section for processing.

In-Gel digestion
In-gel digestion was performed as described by Krug et al. 30 with minor modifications. In brief, 10 mM ammonium bicarbonate and acetonitrile (1:1 v/v) were used to destain gel pieces by washing three times while agitating. Proteins' reduction was then achieved by incubating with 10 mM dithiothreitol (DTT) in 20 mM ammonium bicarbonate for 45 minutes at 56 °C. This was followed by incubation with 55 mM iodoacetamide in 20 mM ammonium bicarbonate for 30 minutes at room temperature in the dark to alkylate the protein samples.
Proteins were then washed twice with 5 mM ammonium bicarbonate followed by one acetonitrile wash and dehydrated using a vacuum centrifuge. Proteins were then pre-digested with Lys-C (Wako) (12.5 ng µl -1 in 20 mM ammonium bicarbonate) and digested with trypsin (Promega). Resultant peptides were then extracted first in three consecutive steps using the following solutions in the stated order: 1) 3 % TFA in 30% acetonitrile; 2) 0.5 % acetic acid in 80 % acetonitrile; 3) 100% acetonitrile. Upon evaporation, the peptide fractions were desalted using StageTips.

Nano-LC-MS/MS analysis
The nano-LC-MS/MS analysis was performed as described by Krug et al. 30 with a few adjustments. In brief, all peptide fractions were measured using the EASY-nLC II nano-LC coupled to an Orbitrab Velos mass spectrometer. Chromatographic separation was conducted on a 15 cm PicoTip fused silica emitter with 75 µM inner diameter and 8 µM Tip inner diameter with reversed-phase ReproSil-Pur C18-AQ 3 µM resin. Together with 0.5 % acetic acid, peptides were injected into the column at 700 nL minute -1 using a maximum pressure of