CDC42-IQGAP Interactions Scrutinized: New Insights into the Binding Properties of the GAP-Related Domain

Abstract

The IQ motif-containing GTPase-activating protein (IQGAP) family composes of three highly-related and evolutionarily conserved paralogs (IQGAP1, IQGAP2 and IQGAP3), which fine tune as scaffolding proteins numerous fundamental cellular processes. IQGAP1 is described as an effector of CDC42, although its effector function yet re-mains unclear. Biophysical, biochemical and molecular dynamic simulation studies have proposed that IQGAP RASGAP-related domains (GRDs) bind to the switch regions and the insert helix of CDC42 in a GTP-dependent manner. Our kinetic and equilibrium studies have shown that IQGAP1 GRD binds, in contrast to its C-terminal 794 amino acids (called C794), CDC42 in a nucleotide-independent manner indicating a binding outside the switch regions. To resolve this discrepancy and move beyond the one-sided view of GRD, we carried out affinity measurements and a systematic mutational analysis of the interfacing residues between GRD and CDC42 based on the crystal structure of the IQGAP2 GRD-CDC42^Q61L GTP complex. We determined a 100-fold lower affinity of the GRD1 of IQGAP1 and of GRD2 of IQGAP2 for CDC42 mGppNHp in comparison to C794/C795 proteins. Moreover, partial and major mutation of CDC42 switch regions substantially affected C794/C795 binding but only a little GRD1 and remarkably not at all the GRD2 binding. However, we clearly showed that GRD2 contributes to the overall affinity of C795 by using a 11 amino acid mutated GRD variant. Furthermore, the GRD1 binding to the CDC42 was abolished using specific point mutations within the insert helix of CDC42 clearly supporting the notion that CDC42 binding site(s) of IQGAP GRD lies outside the switch regions among others in the insert helix. Collectively, this study provides further evidence for a mechanistic framework model that is based on a multi-step binding process, in which IQGAP GRD might act as a 'scaffolding domain' by binding CDC42 irrespective of its nucleotide-bound forms, followed by other IQGAP domains downstream of GRD that act as an effector domain and is in charge for a GTP-dependent interaction with CDC42.

Keywords: CDC42; GAP; GAP-related domain; GRD; GTPase activating protein; IQGAP; RASGAP; RHO GTPases; nucleotide-independent binding; scaffold protein; scaffolding protein; switch regions.

Figure 1

IQGAP GRD binding is nucleotide independent. (A) Domain organization of the IQGAP paralogs along with their GRDs and C-terminal fragments assessed in this study (see text for more details). (B) Fluorescence polarization analysis of IQGAP1 and IQGAP2 proteins with mGppNHp- and mGDP-bound CDC42. (C) Pull-down of endogenous IQGAP1 FL from HEK293 lysates with GppNHp- or GDP-bound GST-CDC42 and GST-RAC1, respectively. Densitometry evaluation of relative IQGAP1 binding to GST-CDC42 proteins (a. u., arbitrary unit) from a triplicate experiment is shown as bar charts.

Figure 2

Analysis of CDC42 switch region and IQGAP1 GRD mutants. (A) The selection of GRD2 and CDC42 mutations is based on the GRD2/CDC42^Q61L structure (PDB: 5CJP). Interacting residues colored on both proteins were selected for mutational analysis. For more details see also Table S1. (B) Multiple sequence alignments of switch regions of RHO GTPases and IQGAPs highlight identical or homologous interacting residues that have been replaced in this study by alanine for analyzing their impact on IQGAP binding. Conserved residues are shaded in grey. Mutations in CDC42 switch regions include 2xSW (bolded residues) and 8xSW (all eight residues, as indicated), and 11xGRD in IQGAP2 C795. (C) Fluorescence polarization measurements of mGppNHp-bound CDC42 WT, 2xSW and 8xSW with IQGAP1 GRD1 or IQGAP2 GRD2. (D) The K_d values for the interactions of IQGAP1 C794, IQGAP2 C795 and C795 11xGRD with the CDC42 variants in mGppNHp- and mGDP-bound form were determined using fluorescence polarization. n.b. stands for no binding observed. C794 and C795 CDC42 WT measurements are included from Figure 1B for simple comparison. (E) Observed rate constants (k_obs) for the IQGAPs association with mGppNHp-bound CDC42 WT, 2xSW and 8xSW were measured using stopped-flow fluorimetry. (F) Pull-down of endogenous IQGAP1 FL from HEK293 lysates with GST-CDC42 in GppNHp-bound or GDP-bound state. Cell lysate was used as an input control. Densitometry evaluation of relative IQGAP1 binding to GST-CDC42 proteins (a. u., arbitrary unit) from triplicate experiments is shown as bar charts.

Figure 3

CDC42 IH mutations decrease binding affinity. (A) Amino acid alignment of the insert helix of selected members of the RHO GTPase family. CDC42 mutations (red) to RAC1 were introduced outside of the conserved regions (grey). (B) Fluorescence polarization data for the interaction of GRD1 and C794 with the CDC42 IH variants. (C) Pull-down of endogenous IQGAP1 FL from HEK293 lysates with GST-CDC42 IH variants in both GppNHp-bound and GDP-bound forms. Cell lysate was used as an input control. The pull-down data for GST-CDC42^wt is shown in Figure 2F as all pull-down experiments were conducted under the same conditions. Densitometry evaluation of relative IQGAP1 binding to GST-CDC42 proteins (a. u., arbitrary unit) from a triplicate experiment is shown as bar charts.

Figure 4

Comparative analysis of IQGAP1 GRD1 interaction with CDC42^Q61L and CDC42^wt. (A) Structural overlay of CDC42^wt GDP AlF₃ p50GAP (green; PDB: 1GRN) on CDC42^Q61L GTP IQGAP2 (blue; PDB: 5CJP) with the focus on Q61 hydrogen bonds (red dashed lines). GDP AlF₃ mimics the transition state of the GTP hydrolysis reaction and is coordinated with the magnesium ion (Mg²⁺) and the nucleophilic water molecule (w) and the arginine finger (R282) of p50GAP. Aluminum trifluoride (AlF₃) mimics the γ-phosphate of GTP in the transition state. In contrast to L61, Q61 is critical for the catalysis of the GTP hydrolysis reaction through three hydrogen bonds (see text). (B) Fluorescence polarization data of IQGAP GRD1 with CDC42 mGppNHp and CDC42 mGDP. (C–F) IQGAP GRD differently forms complexes with CDC42^WT and CDC42^Q61L, respectively, when applied on an analytical SEC. For this purpose, CDC42^WT GppNHp (C) or CDC42^Q61L GppNHp (D) were mixed with IQGAP1 GRD1 and SEC was performed on a Superdex 200 10/300 column using an ÄKTA purifier (flow rate of 0.5 mL/min, fraction volume of 0.5 mL) and a buffer, containing 30 mM Tris/HCl, pH 7.5, 150 mM NaCl, and 5 mM MgCl₂. The elution profiles represented one peak for the respective CDC42 proteins (#1), two peaks for the respective mixtures of respective CDC42 proteins with GRD (#2 and #3) and one peak for the GRD1 (#4). (E) Coomassie brilliant blue staining of the corresponding elution volumes indicated that only peaks #2 contain GRD1 complexes with CDC42^WT or CDC42^Q61L, respectively. Peaks #3 only contain the CDC42 proteins as compared to the peaks #1 and #4. (F) The SEC profiles of CDC42^WT and CDC42^Q61L are summarized for each peak regarding the elution volume, the molecular weight (MW) and the stoichiometry. M stands for monomeric and D for dimeric. The theoretical MWs of CDC42 (21.2 kDa) and GRD (43 kDa) were calculated using the Expasy Protparam tool. The presented MWs for each peak was calculated based on the calibration curve (aldolase 158 kDa and ovalbumin 44 kDa, respectively) and partition coefficient plot (Kav = Ve − V0/Vc − V0) versus the logarithm of MWs; Ve: elution volume number; V0: void volume (8 mL); Vc: geometric column volume (24 mL)). Accordingly, peaks #2 correspond to a heterotetrameric complex between CDC42^WT GppNHp and GRD1 with a MW of 130 kDa, and a heterotrimeric complex of GRD and CDC42^Q61L GppNHp with a MW of 85 kDa.

Figure 5

Deviation in RAS-binding residues in GRDs cause lack of RASGAP activity. (A) Sequence alignment of human RASGAPs p120, neurofibromin (NF1) and the three IQGAP paralogs highlights distinctive deviations in three signature motifs (grey boxes): the finger loop, FLR region and a7/variable loop. RAS-binding residues are shown in red and conserved residues in blue. The catalytic arginine (green) is substituted by threonine in IQGAPs. The numbers correspond to the amino acids of the respective proteins. (B) GTP hydrolysis of various RAS family GTPases was measured in the absence (blue) and in the presence of p120 GAP domain (red; positive control, where no GTP detected) or GRD1^wt (orange) and GRD1^T1046R (green). The GTP hydrolysis of the RAS proteins (10 µM) was measured via HPLC and the GTP content normalized to 100% before adding p120 or GRD1, respectively, at 100 µM concentrations and 1 min incubation time.