Report

Role of IQGAP1, a Target of the Small GTPases Cdc42 and Rac1, in Regulation of E-Cadherin- Mediated Cell-Cell Adhesion

See allHide authors and affiliations

Science  07 Aug 1998:
Vol. 281, Issue 5378, pp. 832-835
DOI: 10.1126/science.281.5378.832

Abstract

The small guanosine triphosphatases (GTPases) Cdc42 and Rac1 regulate E-cadherin–mediated cell-cell adhesion. IQGAP1, a target of Cdc42 and Rac1, was localized with E-cadherin and β-catenin at sites of cell-cell contact in mouse L fibroblasts expressing E-cadherin (EL cells), and interacted with E-cadherin and β-catenin both in vivo and in vitro. IQGAP1 induced the dissociation of α-catenin from a cadherin-catenin complex in vitro and in vivo. Overexpression of IQGAP1 in EL cells, but not in L cells expressing an E-cadherin–α-catenin chimeric protein, resulted in a decrease in E-cadherin–mediated cell-cell adhesive activity. Thus, IQGAP1, acting downstream of Cdc42 and Rac1, appears to regulate cell-cell adhesion through the cadherin-catenin pathway.

Dynamic rearrangement of cell-cell adhesion is a critical step for various cellular processes, including establishment of epithelial cell polarity and developmental patterning (1). Cell-cell adhesion mediated by a cadherin-catenin complex participates in the initial stages of association of polarized cells (2, 3). Cadherins are Ca2+-dependent adhesion molecules, and catenins such as α- and β-catenins are cadherin-associated cytoplasmic proteins that are required for cadherin-mediated cell-cell adhesion. The small GTPases Cdc42 and Rac1 regulate cellular properties such as cell shape, cell growth, and cell polarity (4). Rac1 (5) and Cdc42 (6), together with their exchange factor Tiam-1 (7), also regulate cadherin-mediated cell-cell adhesion. We now present evidence that IQGAP1 mediates these effects.

IQGAP1, a target of Cdc42 and Rac1 (8, 9), is localized at sites of cell-cell contact in Madin-Darby canine kidney (MDCK) epithelial cells, where α-catenin is also localized (9). We therefore examined whether IQGAP1 contributes to cadherin-mediated cell-cell adhesion. IQGAP1 accumulated at sites of cell-cell contact in mouse L cells stably expressing E-cadherin (EL cells) (3,10) (Fig. 1A), but it did not accumulate at such sites in L cells or in L cells stably expressing an E-cadherin mutant in which the cytoplasmic domain was deleted and replaced by the COOH-terminal domain of α-catenin (nEαCL cells). We confirmed that β-catenin did not accumulate at sites of cell-cell contact in nEαCL cells (3). These observations indicated that IQGAP1 accumulates at sites of cell-cell contact in a manner dependent on E-cadherin, β-catenin, or the NH2-terminal domain of α-catenin.

Figure 1

Localization of IQGAP with E-cadherin at sites of cell-cell contact (A) and coimmunoprecipitation of β-catenin and E-cadherin with IQGAP1 (B). (A) Confluent L cells were fixed and doubly stained with antibodies to IQGAP1 (anti-IQGAP1) and anti–β-catenin, and confluent EL cells and nEαCL cells were fixed and doubly stained with anti-IQGAP1 and anti–E-cadherin (10). (B) Confluent EL cells were incubated in the absence or presence of DSP, and cell lysates were then subjected to immunoprecipitation (IP) with no antibody (−), with anti-IQGAP1, or with control immunoglobulin G (IgG) (12). The immunoprecipitates were separated by centrifugation and subjected to SDS-PAGE and immunoblot analysis with antibodies to the indicated proteins (12). Data are representative of three independent experiments.

Immunoprecipitation of IQGAP1 from EL cells resulted in the coprecipitation of β-catenin and E-cadherin (Fig. 1B) (11,12). The stoichiometries of β-catenin and E-cadherin associated with immunoprecipitated IQGAP1 were ∼0.2 and 0.03, respectively. Exposure of EL cells to a cross-linker, dithiobis(succinimidyl propionate) (DSP), before lysis resulted in an increase in the amounts of β-catenin and E-cadherin coprecipitated with IQGAP1, such that their stoichiometries relative to IQGAP1 became 1.0 and 0.4, respectively. These results suggested that IQGAP1 interacts simultaneously with β-catenin and E-cadherin. α-Catenin was not co-immunoprecipitated with IQGAP1 from EL cells that had been incubated in the absence or presence of DSP. Neither the E-cadherin mutant nor β-catenin was immunoprecipitated with IQGAP from nEαCL cells (13), possibly because of the small amounts of β-catenin present in these cells (3).

Maltose-binding protein (MBP)–E-cadherin (the cytoplasmic domain; amino acids 734 to 884), MBP–β-catenin, and MBP–α-catenin fusion proteins were individually incubated with beads coated with either glutathione S-transferase (GST) or a GST- IQGAP1 fusion protein (14). MBP–β-catenin and MBP–E-cadherin specifically interacted with GST-IQGAP1, whereas MBP–α-catenin did not (Fig. 2A). Binding of MBP–β-catenin and MBP–E-cadherin to GST-IQGAP1 was dose dependent and saturable; half-maximal binding was apparent at 20 and 400 nM, respectively. The stoichiometries of the β-catenin and E-cadherin bound to IQGAP1 were each ∼1.0. IQGAP1 interacted with β-catenin, but not with α-catenin, under conditions in which E-cadherin interacted with both catenins (Fig. 2B) (15). IQGAP1 also inhibited the binding of α-catenin to immobilized β-catenin (13).

Figure 2

Direct interaction of IQGAP1 with β-catenin and E-cadherin. (A) Saturable and stoichiometric interaction of MBP–β-catenin and MBP–E-cadherin with GST-IQGAP1. MBP–β-catenin (•), MBP–E-cadherin (○), or MBP–α-catenin (▪) was incubated at the indicated concentrations with beads coated with GST-IQGAP1 (40 pmol), and their interactions with GST-IQGAP1 were examined (14). (B) Preferential binding of GST-IQGAP1 to MBP–β-catenin in the presence of MBP–α-catenin. MBP–β-catenin and MBP–α-catenin were mixed and incubated for 1 hour at 4°C. The mixture was then incubated with beads coated with GST, GST-IQGAP1, or GST–E-cadherin (cytoplasmic domain). Proteins eluted from the beads by glutathione were subjected to SDS-PAGE and immunoblot analysis with anti–β-catenin or anti–α-catenin (15). Data are representative of three independent experiments.

We next examined the interaction of fragments of IQGAP1 with β-catenin or E-cadherin (16). MBP–β-catenin interacted with GST-IQGAP1 (amino acids 1 to 1657) and to a lesser extent with GST–IQGAP1-N (amino acids 1 to 863) and GST–IQGAP1-C (amino acids 764 to 1657) (Fig. 3A). MBP–β-catenin did not interact with GST–IQGAP1-ΔC (amino acids 1 to 1504). MBP–E-cadherin also interacted with GST-IQGAP1 and to a lesser extent with GST–IQGAP1-N and GST– IQGAP1-C, but not with GST–IQGAP1-ΔC. The binding of IQGAP1-N or IQGAP1-C to β-catenin or E-cadherin was dose-dependent (13). These results suggest that the entire structure of IQGAP1 is important for high-affinity binding to β-catenin and E-cadherin. When expressed in amounts smaller than that of the endogenous protein in EL cells, recombinant full-length IQGAP1, but none of the IQGAP1 mutants, accumulated at sites of cell-cell contact (Fig. 3B) (17), suggesting that the interaction of IQGAP1 with β-catenin or E-cadherin (or both) may be essential for its localization at these sites. IQGAP1-N was localized in the nucleus. Wild-type IQGAP1 and IQGAP1-C are able to interact with Cdc42 (9, 13).

Figure 3

Interaction of IQGAP1 mutants with β-catenin and E-cadherin (A) and their localization in EL cells (B). (A) MBP–β-catenin (50 nM) or MBP–E-cadherin (250 nM) was mixed with beads coated with the indicated GST fusion proteins. Proteins eluted from the beads by glutathione were subjected to SDS-PAGE and immunoblot analysis with anti–β-catenin or anti–E-cadherin, as indicated (16). (B) The indicated plasmids were microinjected into the nuclei of EL cells. After 5 hours, the cells were fixed and doubly stained with the indicated antibodies (17). Data are representative of three independent experiments.

To obtain an enriched population of EL cells expressing recombinant IQGAP1, we transfected EL cells with plasmids encoding wild-type or mutant IQGAP1 together with a plasmid encoding tAic2A, an interleukin-3β1 receptor lacking its cytoplasmic domain. The IQGAP1-expressing cells were collected immunologically with an antibody to Aic2 (18, 19). The amount of recombinant IQGAP1 expressed in these cells was two to three times that of endogenous IQGAP1 (13). In cell dissociation assays (3, 19), the IQGAP1-expressing cells were largely dissociated, with many single cells apparent, whereas control cells or IQGAP1-ΔC–expressing cells remained as aggregates (Fig. 4A). Thus, the adhesive activity of IQGAP1-expressing cells appeared reduced relative to that of control cells or of cells expressing IQGAP1-ΔC. Overexpression of neither IQGAP1-N nor IQGAP1-C altered the adhesive properties of EL cells (13). When these experiments were repeated with nEαCL cells, the IQGAP1-expressing cells showed an adhesive activity similar to that of control cells. Therefore, the effect of IQGAP1 on cell adhesion appeared to be mediated by β-catenin or E-cadherin.

Figure 4

Effects of IQGAP1 mutants on cell-cell adhesion. (A) Effect of recombinant IQGAP1 on cell-cell adhesion. EL cells (upper panels) or nEαCL cells (lower panels) were transfected with the indicated plasmids and with pME18-tAic2A. After incubation for 20 hours, cells expressing the interleukin-3β receptor were collected with anti-Aic2 (18,19). The collected cells (5 × 105 cells per well) were seeded into 48-well plates and incubated for 24 hours. A dissociation assay was then performed after treatment of the cells with trypsin in the presence of either Ca2+ (TC treatment) or EGTA (TE treatment) as described (3,19). The total particle numbers after TC or TE treatment were designated N TC andN TE, respectively. The cell dissociation index (N TC/N TE) was scored and is shown below the panels. (B) Effect of Cdc42 on cell-cell adhesion. EL cells were transfected with pEF-BOS-Myc-IQGAP1 and pME18-tAic2A plasmids together with either pEF-BOS-HA or pEF-BOS-HA-Cdc42Val12. The cells were collected and subjected to the dissociation assay (3, 19). (C) Dissociation of α-catenin from a cadherin-catenin complex by IQGAP1 in vivo. EL cells were transfected with pEM18-tAic2A and either pEF-BOS-Myc, pEF-BOS-Myc-IQGAP1, or pEF-BOS-Myc–IQGAP1-ΔC (19,20). The cells were isolated and transferred to new plates. After incubation for 24 hours, the cells were treated with DSP, lysed, and subjected to centrifugation (12, 20). The resulting supernatant was subjected to immunoprecipitation (IP) with either anti–E-cadherin or control IgG, and the immune complexes were isolated by centrifugation and subjected to SDS-PAGE and immunoblot analysis with antibodies to the indicated proteins. Data are representative of three independent experiments.

When the dissociation assay was performed with EL cells expressing both IQGAP1 and Cdc42Val12, a Cdc42 mutant that is defective in GTPase activity and is thought to exist constitutively in the GTP-bound form in cells, the cells did not dissociate, remaining as aggregates (Fig. 4B). Expression of IQGAPI with either Cdc42Asn17, a mutant that preferentially binds GDP rather than GTP and is thought to exist constitutively in the GDP-bound form in cells, or RhoAVal14 did not inhibit cell dissociation (13). Thus, Cdc42 inhibits IQGAP1 function, possibly through direct interaction with IQGAP1.

The amount of α-catenin, but not that of β-catenin, associated with E-cadherin was reduced when IQGAP1 (but not IQGAP1-ΔC) was overexpressed in EL cells (Fig. 4C) (20). Overexpression of IQGAP1 or IQGAP1-ΔC did not affect the amounts of E-cadherin, α-catenin, or β-catenin expressed in these cells, and the amounts of recombinant IQGAP1 and IQGAP1-ΔC were similar (13). Thus, overexpression of IQGAP1 appeared to induce dissociation of α-catenin from the cadherin-catenin complex. The dissociation of α-catenin from β-catenin may be responsible for the in vivo action of IQGAP1.

Treatment of cells with pervanadate results in the dissociation of α-catenin from the cadherin-catenin complex, and the dissociation of α-catenin from the complex correlates with the decrease in cadherin activity (21). Dissociation of α-catenin from the cadherin-catenin complex also occurs during the passage of human breast epithelial cells (22). Our data suggest that IQGAP1 might participate in these processes.

REFERENCES AND NOTES

View Abstract

Navigate This Article