Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2013 Aug 10.
Published in final edited form as: Mol Cell. 2012 Aug 10;47(3):469–483. doi: 10.1016/j.molcel.2012.06.037

The signaling adapter Gab1 regulates cell polarity by acting as a PAR protein scaffold

Ziqiang Yang 1, Bin Xue 2,3, Masataka Umitsu 1, Mitsuhiko Ikura 1,4, Senthil K Muthuswamy 1,2,3,4, Benjamin G Neel 1,4
PMCID: PMC3462001  NIHMSID: NIHMS399681  PMID: 22883624

Summary

Cell polarity plays a key role in development and is disrupted in tumors, yet the molecules and mechanisms that regulate polarity remain poorly defined. We found that the scaffolding adaptor GAB1 interacts with two polarity proteins, PAR1 and PAR3. GAB1 binds PAR1 and enhances its kinase activity. GAB1 brings PAR1 and PAR3 into a transient complex, stimulating PAR3 phosphorylation by PAR1. GAB1 and PAR6 bind the PAR3 PDZ1 domain and thereby compete for PAR3 binding. Consequently, GAB1 depletion causes PAR3 hypo-phosphorylation and increases PAR3/PAR6 complex formation, resulting in accelerated and enhanced tight junction formation, increased trans-epithelial resistance and lateral domain shortening. Conversely, GAB1 over-expression, in a PAR1/PAR3-dependent manner, disrupts epithelial apical-basal polarity, promotes multi-lumen cyst formation, and enhances growth factor-induced epithelial cell scattering. Our results identify GAB1 as a novel negative regulator of epithelial cell polarity that functions as a scaffold for modulating PAR protein complexes on the lateral membrane.

Introduction

Precise control of cell polarity is essential for many biological processes, including asymmetric division of stem/progenitor cells, epithelial morphogenesis, epithelial/mesenchymal transformation (EMT) and cell migration. Disrupted epithelial cell polarity is a hallmark of cancer and contributes to other diseases. Progress has been made in delineating the polarity machinery, but how signaling pathways regulate cell polarity is poorly understood.

Initially identified as the products of C. elegans genes whose mutation disrupts asymmetric cell division (Kemphues, 2000), the PAR proteins are essential, highly conserved components of the polarity machinery. Worms have six par genes (par1-par6): par1 and par4 encode protein kinases, par3 and par6 encode scaffolding molecules, par5 encodes a 14-3-3 homolog and par2 encodes an E3-ubiquitin ligase-like protein (Macara, 2004b). All of the PAR proteins except PAR2 are found in mammals, where they are indispensable for epithelial development and polarization. Par3 deletion in mice causes abnormal epicardial development (Hirose et al., 2006). PAR3 knockdown in MDCK II cells causes the formation of multi-lumen cysts (Bryant et al., 2010; Hao et al., 2010), delays tight junction (TJ) formation and decreases trans-epithelial electrical resistance (TER) in “calcium switch” assays, although TJs eventually do form normally (Chen and Macara, 2005). Although Par1a and Par1b single knock-out mice are viable, compound deletion results in embryonic lethality (Hurov et al., 2007; Lennerz et al., 2010). At the cellular level, PAR1 depletion causes shortening of the lateral membrane (Cohen et al., 2004; Suzuki et al., 2004), and PAR1 knockdown cells show increased TER compared with parental cells during polarization (Suzuki et al., 2004). Conversely, but similar to the effects of PAR3 depletion, PAR1b over-expression causes multi-lumen cysts (Cohen et al., 2011).

The effects of PAR proteins are accompanied, and presumably caused, by their asymmetrical distribution along the apical-basal axis of polarized epithelial cells. PAR3, PAR6 and aPKC are targeted to TJs, where aPKC is activated by GTP-bound CDC42. Similar to the effects of PAR3 deficiency, depletion of aPKC also impairs TJ formation (Suzuki et al., 2004). PAR6 and aPKC also localize to the apical domain through interactions with the Crumbs complex (Morais-de-Sa et al., 2010). By contrast, PAR1 resides on the lateral membrane, whereas PAR4 and PAR5 are found mainly in the cytosol. The PAR proteins are responsible, at least in part, for these differences in localization. For example, the PAR3/PAR6/aPKC complex at TJs phosphorylates PAR1, promoting its release into the cytosol and preventing PAR1 from “invading” the TJ from its location on the lateral domain. Conversely, PAR3 is excluded from the lateral domain by PAR1-catalyzed phosphorylation and PAR5-mediated cytosolic translocation (Macara, 2004a, b; Suzuki and Ohno, 2006).

As master regulators of cell polarity, PAR proteins must integrate signals from multiple stimuli. Although the detailed mechanisms remain unclear, the tumor suppressor ASPP2, the receptor tyrosine kinase (RTK) ErbB2 and the Eph family RTK ligand Ephrin B1 modulate polarity via PAR proteins (Aranda et al., 2006; Cong et al., 2010; Lee et al., 2008; Sottocornola et al., 2010). Furthermore, CagA (cytotoxin-associated antigen A), a key virulence determinant of Helicobacter pylori, the primary cause of gastric ulcers/carcinoma (Hatakeyama, 2004), binds to and inhibits PAR1 and promotes scattering (“hummingbird phenotype”) of gastric epithelial cells. This interaction, together with CagA binding and activation of the protein-tyrosine phosphatase SHP2, is thought to contribute to H. pylori-induced oncogenesis (Saadat et al., 2007).

CagA is functionally similar to the GAB family (GAB1, GAB2 and GAB3) of scaffolding adaptors (Hatakeyama, 2003), which are critical components of multiple growth factor and cytokine receptor signaling pathways (Gu and Neel, 2003; Wohrle et al., 2009). Upon agonist stimulation, GABs are recruited to receptors, where are phosphorylated on multiple tyrosyl residues, creating binding sites for signal relay molecules. For example, tyrosyl phosphorylated GAB1, like CagA, recruits SHP2 and promotes ERK activation. Gab1 deletion in mice causes embryonic lethality, accompanied by defective skin, liver and placental development (Itoh et al., 2000; Sachs et al., 2000), some aspects of which could reflect aberrant cell polarity. In MDCK II cells, GAB1 is required for HGF-induced scattering and tubulogenesis (Rosario and Birchmeier, 2003). GAB1 binds to the HGF receptor (Met), and the Met-binding domain of GAB1 and its SHP2 binding sites are essential for HGF-induced scattering and tubulogenesis. GAB1 localizes to epithelial cell-cell contacts via its N-terminal PH domain, and proper sub-cellular localization is crucial for its role in HGF-induced morphogenesis (Maroun et al., 1999; Maroun et al., 2000).

GAB function also is co-opted in several neoplasms. For example, Gab2 is required for BCR/ABL-evoked myeloid transformation (Sattler et al., 2002), whereas GAB2 over-expression accentuates and Gab2 deficiency impairs Erbb2/Neu-evoked mammary carcinogenesis (Bentires-Alj et al., 2006; Ke et al., 2007). Notably, combining GAB2 over-expression with HER2 (ERBB2) activation in MCF-10A cells results in disruption of their normal gland-like architecture and an invasive “multi-acinar” phenotype (Bentires-Alj et al., 2006; Brummer et al., 2006).

Given the similar actions of CagA and GAB1, we asked if GABs also regulate cell polarity via PAR proteins. Indeed, we found that GAB1, like CagA, affects PAR1, but in an entirely different way. Our results identify GAB proteins as key PAR protein scaffolds, which control the dimensions of the lateral membrane of polarized epithelia, the kinetics and extent of TJ formation, and their dissolution during HGF-induced cell scattering.

Results

GAB proteins interact with PAR1 family members

We first asked if, like CagA (Saadat et al., 2007), GAB proteins bind PAR1. There are three mammalian GABs (GAB 1, 2 and 3) and four PAR1 isoforms (PAR1a–d). Expression constructs for HA-tagged mouse Gab1 or Gab2 and Flag-tagged human PAR1b were co-transfected into 293T cells, and after 48 hours, lysates were prepared and immunoprecipitated with anti-HA antibodies. Notably, PAR1b was detected in Gab1 and Gab2 immune complexes (Fig. 1A), and Gab1 (or Gab2) were recovered in PAR1b (anti-Flag) immunoprecipitates (Fig. 1B and data not shown). Gab1 or Gab2 also co-immunoprecipated with PAR1a, PAR1c and PAR1d (data not shown). Moreover, Gab1 (or Gab2) and PAR1b co-immunoprecipitated from lysates of co-transfected MDCK II cells, which (unlike 293 cells) undergo regulated changes in cell polarity (Fig. 1C). Using available anti-PAR1 antibodies, we could not detect endogenous PAR1b (or other PAR1 family members) in GAB1 immunoprecipitates from MDCK II cell lysates, so we generated an MDCK II line stably expressing Flag-tagged PAR1b. As expected (Hurov et al., 2001), two endogenous PAR1b isoforms (~ 75 and 80 kDa, respectively) were detected following immunoprecipitation and immunoblotting with anti-PAR1 antibodies, and Flag-tagged PAR1b was expressed at a similar level to these isoforms (Fig. 1D, left panel). Using these cells, we found that GAB1 and PAR1b form a complex when both are expressed at physiological levels (Fig. 1D, right panel). Endogenous GAB1/PAR1b interaction also was detected in Caco-2 cells (Fig. S1A).

Figure 1. GAB proteins associate with PAR1b.

Figure 1

Open in a new tab

(A–C) The indicated expression constructs were co-transfected into 293T (A, B) or MDCK-II (C) cells. Lysates were prepared, and immunoprecipitated and immunoblotted with the indicated antibodies. (D) Left: Lysates from WT MDCK II cells or MDCK II cells stably expressing PAR1b-Flag (PAR1b-MDCK II) were immunoprecipitated and immunoblotted with anti-PAR1b antibodies. Arrows indicate endogenous PAR1b isoforms (Mr ~ 75 and 80 kDa). Note comparable levels of endogenous PAR1b and expressed PAR1b-Flag. Right: PAR1b-MDCK II lysates were immunoprecipitated with anti-GAB1 antibodies or control IgG, followed by immunoblotting. (E) Gab1 mutants and their interaction with PAR1b; (+) indicates binding; (−) no interaction. (F) The indicated expression constructs were co-transfected into 293T cells, and lysates were immunoprecipitated with anti-HA antibodies and immunoblotted, as indicated. (G) Diagram of PAR1b mutants and their interaction with Gab1. (H) The indicated expression constructs were co-transfected into 293T cells, and lysates were immunoprecipitated with M2 beads and immunoblotted, as indicated. (I) His-Gab1-152–250 immobilized on nickel agarose was incubated with purified GST-KA1, and bound proteins were eluted, resolved by SDS-PAGE and stained with Coomassie Blue. Also, see Fig. S1.

Gab1 binds the PAR1b KA1 domain

Using deletion mutants, we defined the region in Gab1 required for interaction with PAR1b (Fig. 1E). Expression constructs for HA-tagged versions of these mutants were co-transfected with a PAR1b expression construct, and Gab1/PAR1b association was assessed by anti-HA immunoprecipitation, followed by anti-PAR1b immunoblotting (Fig. S1B–D). A Gab1 mutant containing only its N-terminal 300 amino acids (Gab1 1–300) interacted with PAR1b (Fig. 1E and Fig. S1B), while mutants lacking their N-terminal 300 amino acids failed to bind (Fig. 1E and Fig. S1B). Deleting the PH domain (Gab1 116–695) diminished, but did not abrogate, Gab1/Par1b interaction (Fig. 1E and Fig. S1B). Thus, PH domain-dependent targeting to an appropriate intracellular location probably is important for Gab1/PAR1b interaction, but amino acids 117–300 of Gab1 contain a second key region for PAR1b interaction. To further define this region, we tested additional Gab1 mutants. Gab1 152–695 bound PAR1b comparably to Gab1 116–695 (Fig. 1E and Fig. S1C), whereas Gab1 201–695 and Gab1 251–695 lost nearly all binding (Fig. 1E and Fig. S1C). In parallel, we found that Gab1 1–250 was sufficient for binding PAR1b (Fig. S1D). These results suggested that the critical PAR1b binding region lies between Gab1 amino acids 152–250, and indeed, deleting this region (Gab1Δ152–250) markedly reduced Gab1/PAR1b interaction (Fig. 1F).

Next, we mapped the Gab1-binding region on PAR1b. PAR1b contains an N-terminal kinase domain, followed by a UBA (ubiquitin-associated) domain, a long spacer region with aPKC phosphorylation sites, and a C-terminal KA1 (kinase-associated 1) domain. The kinase and UBA domains (PAR1b 1–329) had almost no ability to interact with Gab1 (Fig. 1G and Fig. S1E), but Par1b constructs containing amino acids 330–755 or 551–755 bound Gab1 even better than WT PAR1b (Fig. 1G and Fig. S1E), as did a mutant (PAR1b 651–755) containing only the KA1 domain (Fig. 1G and Fig. S1F). Thus, the KA1 domain is the major binding site for Gab1 and, as expected, a PAR1b mutant lacking its KA1 domain had minimal Gab1 binding (Fig. 1H). Notably, the Gab1 binding site on PAR1b (KA1 domain) is distinct from its binding site for CagA (kinase domain; see Discussion). Also, the increased binding of N-terminal truncations of PAR1b to Gab1 suggested that the Gab1 binding site in (full length) PAR1b might be obscured by interaction between the KA1 and N-terminal domains, a notion confirmed subsequently (see below).

Recently, Pak4 was found to associate with Gab1 (Paliouras et al., 2009) via Gab1 amino acids 116–234, which overlap the PAR1b-binding site (Fig. 1E). To ask if Gab1 interacts directly with PAR1b, we carried out in vitro binding assays. Bacterially expressed, His-tagged Gab1 152–250, immobilized on nickel agarose, was incubated with recombinant GST-tagged PAR1b KA1 domain. Bound proteins were recovered by centrifugation, resolved by SDS-PAGE and detected with Coomassie Blue. The KA1 domain of PAR1b bound to Gab1 152–250, indicating that Gab1 and PAR1b can interact directly (Fig. 1I).

Gab1 also binds PAR3/aPKC

We next asked if Gab1 interacts with other PAR proteins. Indeed, transiently expressed HA-Gab1 co-immunoprecipitated with transiently expressed Myc-PAR3 and Myc-aPKC; by contrast, there was minimal HA-Gab1 association with Myc-PAR6, though the latter was expressed at higher levels than PAR3 (Fig. 2A). Co-transfection of PAR3 significantly enhanced aPKC interaction with Gab1 (Fig. S2A), suggesting that aPKC binds Gab1 at least in part via PAR3 (see below).

Figure 2. Gab1 binds to the PDZ1 domain of PAR3.

Figure 2

Open in a new tab

(A) The indicated expression constructs were co-transfected into 293T cells, and lysates were immunoprecipitated with anti-Myc antibodies and immunoblotted, as indicated. (B) Gab1 mutants and their interactions with PAR3; (+) indicates binding; (−) no interaction. (C–D) The indicated constructs were co-transfected into 293T cells, and lysates were immunoprecipitated with anti-HA antibodies and immunoblotted, as indicated. * indicates IgG heavy chain. (E) The indicated constructs were co-transfected into 293T cells, and lysates were immunoprecipitated with anti-Myc antibodies and immunoblotted, as indicated. (F) MDCKII/ErbB2 cells, which express ErbB2 fused to FKBP (Aranda et al., 2006), were stimulated with AP1510 (1µM) as indicated, and anti-GST (Control Ab) or anti-GAB1 immunoprecipitates were prepared and immunoblotted. Arrows indicate PAR3 isoforms (150 kD and 180 kD) that co-precipitate with GAB1. Also, see Fig. S2.

We then mapped the region on Gab1 required for PAR3 association. Deleting the PH domain (Gab1 116–695) abrogated Gab1 interaction with PAR3 (Fig. 2B, C). The N-terminal 300 amino acids of Gab1 (1–300), which is sufficient for PAR1b binding, did not bind PAR3, indicating that the PAR1 and PAR3 binding sites are distinct. However, extending this fragment to include the N-terminal 400 amino acids (Gab1 1–400) restored PAR3 interaction (Fig. 2B, C). Hence, a second key region for PAR3 binding (in addition to the Gab1 PH domain) lies between amino acids 300–400 (Fig. 2B, C). Gab1 lacking this region (Gab1Δ301–400) did not co-immunoprecipitate with PAR3, while internal deletions within this region decreased, but did not eliminate Gab1/PAR3 interaction (Fig. 2D). We therefore define amino acids 301–400 of Gab1 as the PAR3-binding site. Consistent with the results of the PAR3/aPKC co-transfection experiments (Fig. S2A), Gab1Δ301–400 and Gab1Δ301–340 also showed decreased interaction with aPKC (Fig. S2B).

PAR3 contains 3 PDZ domains through which it interacts with multiple proteins. PAR3 fragments containing the PDZ1 domain (1–373, 1–653 and 84–417) interacted with Gab1; conversely, fragments lacking this domain failed to interact (Fig. 2E and Fig. S2C). Moreover, recombinant PAR3 PDZ1 domain (expressed as a GST fusion protein) interacted with transiently expressed HA-Gab1 (Fig. S2D), indicating that the PAR3 PDZ1 domain is necessary and sufficient for interaction with Gab1. GAB1 also co-immunoprecipitated with PAR3 in lysates from MDCK II or MCF-10A mammary epithelial cells (Fig. 2F and Fig. S2E), and this interaction increased following activation of co-expressed rat Erbb2/Neu (see Discussion).

Gab1 acts as a scaffold to facilitate PAR3 phosphorylation by PAR1

Because Gab1 uses different regions to interact with PAR1 and PAR3, respectively, it might function as a scaffold to bring together these two polarity regulators. To test this possibility, we co-transfected 293T cells with expression constructs for Myc-tagged PAR3, Flag-tagged PAR1b and/or HA-tagged Gab1. After 48 hrs, lysates were prepared, PAR3 was immunoprecipitated with anti-Myc antibodies, and co-precipitating PAR1b or Gab1 was detected with anti-PAR1b or anti-HA antibodies, respectively. Indeed, Gab1 co-expression increased the amount of PAR1b recovered in PAR3 immunoprecipitates (Fig. 3A).

Figure 3. Gab1 acts as a scaffold to facilitate PAR3 phosphorylation by PAR1.

Figure 3

Open in a new tab

(A) Gab1 promotes PAR3/PAR1 interaction. The indicated constructs were co-transfected into 293T cells, and Myc-PAR3/PAR1-Flag interaction was assessed. (B) PAR1b, via its kinase activity, disrupts PAR3 interaction with Gab1. The indicated constructs were co-transfected into 293T cells, and Myc-PAR3/HA-Gab1 interaction was assessed. (C) PAR1b disrupts Gab1/PAR3 interaction by phosphorylating PAR3. The indicated constructs were co-transfected into 293T cells. The effect of PAR1b-Flag expression on HA-Gab1/Myc-PAR3 (or Myc-PAR3 (SSAA)) interaction was compared by immunoblotting. (D) Model of Gab1/PAR protein interactions (see text for details). (E) MDCK II cells were transiently transfected with the indicated constructs, and PAR3 phosphorylation analyzed by immunoprecipitation/immunoblotting with anti-p-14-3-3 motif antibodies. Note specificity of the latter. (F) PAR3 phosphorylation by PAR1 in Control or GAB1-KD MDCK II cells undergoing calcium switch. (G) Gab1 enhances PAR1b kinase activity. HA-TAU alone or with HA-Gab1 was transfected into GAB1-KD MDCK II cells stably expressing PAR1b. TAU phosphorylation was assessed by immunoblotting. (H) PAR6 inhibits Gab1/PAR3 interaction. The indicated constructs were co-transfected into 293T cells, and Myc-PAR3/HA-Gab1 interaction was assessed. Red arrows indicate HA-PAR6; the band above is non-specific. (I) Control or GAB1-KD MDCK II lysates were immunoprecipitated with anti-PAR3 antibodies, and co-precipitating PAR6 and aPKC (arrows) were detected by immunoblotting. Also, see Fig. S3.

We noticed that less Gab1 co-immunoprecipitated with PAR3 when Gab1, PAR1 and PAR3 were co-expressed (Fig. 3A) than when Gab1 and PAR3 were co-transfected (Fig. 2A). These findings suggested that PAR1b might inhibit the interaction of Gab1 with PAR3. We compared the interaction of Gab1 and PAR3 with or without co-expressed PAR1b, and indeed, less Gab1/PAR3 complex was detected in the presence of PAR1b. By contrast, a PAR1b point mutant without kinase activity (PAR1b(D160N)) did not decrease the Gab1/PAR3 complex (Fig. 3B). When PAR3 diffuses from TJs into the lateral membranes of polarized epithelial cells, it is phosphorylated by PAR1 and then released into the cytosol; thus, PAR1-catalyzed PAR3 phosphorylation acts to restrict PAR3 to TJs (Benton and St Johnston, 2003; Hurd et al., 2003). We therefore asked whether PAR1b affected the interaction of Gab1 with a phosphorylation-resistant mutant of PAR3 (PAR3 S144A/S873A). Notably, Gab1 co-immunoprecipitation with phosphorylation-resistant PAR3 (SSAA) was unaffected by PAR1 (Fig. 3C), Indicating that PAR1b disrupts Gab1/PAR3 interaction, at least in part, by phosphorylating PAR3.

Taken together, these results suggested that Gab1 brings together PAR3 and PAR1 to facilitate PAR3 phosphorylation (Fig. 3D). To test this hypothesis further, we depleted GAB1 in MDCK II cells (Fig. S3A), and monitored PAR3 phosphorylation during a calcium switch assay by immunoblotting PAR3 immunoprecipitates with an anti-phospho-14-3-3 motif antibody. Notably, this antibody fails to recognize a PAR3 mutant lacking its PAR1b phosphorylation sites, S144/873 (Fig. 3E). As predicted, PAR3 phosphorylation was decreased in GAB1-knockdown (GAB1-KD) cells compared with controls (Fig. 3F). Hence, GAB1, by facilitating PAR1/PAR3 interaction, maintains normal levels of PAR3 phosphorylation during cell polarization. Consistent with this idea, re-expressing Gab1 in GAB1-KD cells (by using a mouse Gab1 cDNA resistant to the canine shRNA) enhanced PAR3 phosphorylation. By contrast, Gab1 mutants lacking their PAR1b-binding or PAR3 binding sites were less efficient in restoring PAR3 phosphorylation during a calcium switch (Fig. S3B).

Gab1 enhances PAR1 kinase activity

The kinase activity of the yeast PAR1 homologs, Kin1p/Kin2p, is auto-inhibited by intramolecular interaction of their C-termini and N-terminal kinase domains (Elbert et al., 2005), so we asked if the C-terminal KA1 domain of PAR1b also interacted with its N-terminus. Lysates of 293T cells transiently expressing PAR1b (1–650) were incubated with purified, recombinant GST-tagged KA1 domain. Bound proteins were recovered on glutathione-agarose and analyzed by immunoblotting (Fig. S3C). As reported for Kin1p/Kin2p, the N-terminal part of human PAR1b, containing the kinase domain, interacted (in trans) with the C-terminal KA1 domain. This interaction was inhibited by adding increasing amounts of lysate from 293T cells transiently expressing Gab1, but not by parental 293T lysate (Fig. S3C). These findings suggested that Gab1, by binding to the PAR1 KA1 domain and disrupting its interaction with the kinase domain, might stimulate PAR1; indeed, TAU phosphorylation by PAR1b was enhanced by Gab1 over-expression (Fig. 3G). Thus, Gab1 promotes PAR1 phosphorylation of PAR3 in two ways: it facilitates PAR1/PAR3 co-localization and increases PAR1 kinase activity by relieving auto-inhibition.

Gab1 regulates the amount of the PAR3-PAR6-aPKC complex

When epithelial cells polarize, PAR3, PAR6 and aPKC form a complex. PAR3 interaction with PAR6 is mediated by the PAR3 PDZ1 domain (Macara, 2004b). Because Gab1 also binds to this domain (Fig. 2E), and Gab1 forms a complex with PAR3, but not with PAR6 (Fig. 2A), we suspected that Gab1 and PAR6 might compete for binding to PAR3. Indeed, co-expression of PAR6 inhibited the interaction of transiently expressed PAR3 and Gab1 (Fig. 3H). Furthermore, more endogenous PAR6 co-immunoprecipitated with PAR3 from lysates of GAB1-KD MDCK II cells than from controls. We also detected more PAR3-aPKC complex in GAB1-KD cells, consistent with increased PAR3-PAR6 interaction in these cells (Fig. 3I, arrows), and these cells formed more PAR3/PAR6/aPKC complex during a calcium switch (Fig. S3D). Therefore, GAB1 also controls the amount of PAR3/aPKC/PAR6 complex in two ways: first, it competes directly with PAR6 for binding to PAR3 (Fig. 3D) and second, it promotes PAR1 phosphorylation of PAR3, releasing PAR3 from the membrane and limiting its access to PAR6.

GAB1 regulates TJ formation and function

Given the ability of Gab1 to regulate PAR1 activity, PAR3 phosphorylation, and PAR6/PAR3 interaction, we hypothesized that GAB1, like the PAR proteins, is important for cell polarity control. To explore this possibility, we first compared TJ formation in control and GAB1-KD MDCK II cells subjected to a calcium switch assay. As expected, TJs, as assessed by immunofluorescence staining for ZO-1, were disrupted in control and GAB1-KD MDCK II cells placed in calcium-free media (Fig. 4A). Upon re-addition of calcium, TJs began to form at cell-cell contacts. In GAB1-KD cells, ZO-1 began to accumulate as early as 15 minutes after calcium repletion, and formed a complete, honeycomb-like structure by 1 hour. By contrast, in controls, interruptions in the pattern of ZO-1 staining persisted for > 1 hour (Fig. 4A). In addition, PAR3 accumulated much faster at cell-cell contacts in GAB1-KD cells (Fig. S4A), and GAB1 depletion in human Caco-2 colonic epithelial cells also enhanced TJ formation (Fig. S4B). Gab1 distributed throughout the cell in calcium-depleted cells, but accumulated rapidly at cell-cell contacts upon calcium repletion (Fig. 4C), placing it in an appropriate location to regulate TJ formation. A key function of GAB1 is to bind SHP2 and promote RAS/ERK activation in response to growth factor stimulation. However, addition of the MEK inhibitor U0126 did not affect TJ formation in control or GAB1-KD MDCK II cells (data not shown).

Figure 4. GAB1 regulates TJ formation/function.

Figure 4

Open in a new tab

(A) Calcium switch assays. Control and GAB1-KD MDCK II cells were fixed at the indicated times after Ca2+ repletion and immunostained for ZO-1. Nuclei were stained with DAPI. Representative fluorescent images are shown. (B) TER (mean ± SD) in control and GAB1-KD MDCK II cells following calcium switch (*** p<0.001). (C) MDCK II cells stably expressing HA-Gab1 undergoing calcium switch were stained with anti-HA antibody to visualize Gab1. Note that Gab1 resides in the cytosol when cells are in Ca2+-free medium, but accumulates at cell-cell contacts upon calcium switch. (D) MDCK II-HA-Gab1 cells grown on filters for 24 h were immunostained with anti-HA (green) and anti-ZO-1 (red). Gab1 resides on the lateral membrane, whereas ZO-1 is on tight junctions. Representative x–y, x–z, and y–z slices are shown. Arrow indicates position of x–y plane in the x–z axis. Also, see Fig. S4.

To assess the integrity of the TJs in GAB1-KD cells (and to provide independent, quantitative validation of the effects of GAB1 deficiency on TJ formation), we measured TER. In calcium-free medium, TER was close to zero in control and GAB1-KD cells, indicating disrupted TJs. When calcium was repleted, TER increased in control and GAB1-KD cells, but GAB1-KD cells attained a higher TER at a faster rate than did controls (Fig. 4B), consistent with their more rapid TJ formation (Fig. 4A).

Gab1 regulates TJ function by binding to PAR proteins

To confirm that precocious TJ formation in GAB1-KD cells reflected GAB1 deficiency rather than off-target effects, we re-expressed Gab1 (Fig. 5A). Gab1 (pSp-shGab1+pBabe-Gab1) decreased the rate of TJ formation, compared with control vector-reconstituted cells (pSp-shGab1+pBabe) (Fig. 5B). Unfortunately, using this vector system, Gab1 was substantially over-expressed, and we were unable to find an expression system that resulted in re-expression of Gab1 to near-endogenous levels. Nevertheless, reconstituting GAB1-KD MDCK II cells with Gab1 mutants lacking their PAR1b (pSp-shGab1+pBabe-Gab1Δ152–250) or PAR3 (pSp-shGab1+pBabe-Gab1Δ301–400) binding sites did not affect TJ formation. By contrast, a Gab1 mutant lacking its CRK-binding sites, two of which overlap the PAR3 binding site, was as competent as WT (wild-type) Gab1 in delaying TJ formation (Fig. 5B).

Figure 5. GAB1 regulates TJ formation via the PAR proteins.

Figure 5

Open in a new tab

(A) Lysates from GAB1-KD MDCK II cells reconstituted with WT Gab1 or Gab1 mutants were immunoblotted, as indicated. Arrow indicates endogenous GAB1. * indicates Gab1Δ152–250-PAR1b fusion protein; note its low expression compared with endogenous GAB1. (B) WT Gab1 and Gab1 mutant-reconstituted cells undergoing calcium switch were fixed 1 h following Ca2+ repletion, and stained with anti-ZO-1 antibodies. Representative confocal images are shown. (C) TER (mean ± SD) was measured during calcium switch. “Controls” are parental MDCK II cells infected with retroviruses packaged with empty vectors of pBabe and pSuperior.retro.puro. ** p<0.01, *** p<0.001, KD + V (or KD + Δ152–250/KD + Δ301–400) versus KD + Gab1. (D) GAB1-KD MDCK II cells stably expressing Gab1Δ152–250-FKBP and PAR1b-F*RB-HA were plated onto filters in S-MEM containing AP21697 or EtOH (vehicle), and subjected to calcium switch. One hour later, cells were fixed and immunostained with anti-ZO-1 antibodies. Representative confocal images are shown. (E) Calcium switch assays on GAB1-KD MDCK cells stably expressing Gab1Δ152–250-FKBP and PAR1b-F*RB-HA grown on filters in S-MEM in the presence of AP21697 or EtOH (vehicle) and TER (mean ± SD, *** p<0.001) was monitored. Also, see Fig. S5.

Because the Pak4-binding site on Gab1 (Paliouras et al., 2009) overlaps the PAR1b-binding site, we took two approaches to assess the effects of Gab1/PAR1 complex formation directly. First, we fused Gab1Δ152–250 to PAR1b via a flexible linker (Fig. S5A). We also generated a Gab1 mutant that can associate inducibly with PAR1b by fusing the Gab1Δ152–250 C-terminus to FKBP and the PAR1b C-terminus to a modified FRB (F*RB) peptide (Fig. S5B). The direct fusion of Gab1Δ152–250 and PAR1b (pSp-shGab1+pBabe-Gab1Δ152–250-PAR1b) was as effective as WT Gab1 in delaying TJ formation (Fig. 5B), even though it was expressed at much lower (and near endogenous GAB1) levels (Fig. 5A). Similarly, inducing Gab1Δ152–250-FKBP association with PAR1b-F*RB in GAB1-KD MDCK II cells, by treatment with the rapamycin analog AP21697 (dimerizer), delayed TJ formation (Fig. 5D).

We also monitored TER development in GAB1-KD- and -reconstituted MDCK II cells. Consistent with their effects on TJ formation, expressing WT Gab1 (KD+Gab1) or the direct Gab1Δ152–250-PAR1b fusion protein (KD+fusion) lowered TER in GAB1-KD MDCK-II cells. By contrast, the control vector (KD+V), PAR1b-binding mutant (KD+Δ152–250) and PAR3-binding mutant (KD+Δ301–400) were defective in this assay (Fig. 5C). AP21697-induced dimerization of Gab1Δ152–250-FKBP and PAR1b-F*RB also lowered TER (Fig. 5E). We conclude that precocious TJ formation and TER establishment in GAB1-KD MDCK II cells is due to loss of GAB1 and, more specifically, its interactions with PAR1b and PAR3.

GAB1/PAR complexes regulate apical-basal polarity

In polarized epithelia, the cell membrane is divided into apical and lateral domains. GAB1, like PAR1, localized to the lateral domain of MDCK II cells (Fig. 4D). Previous studies showed that PAR1 depletion (by RNAi) results in shortening of the lateral membrane (Cohen et al., 2004; Suzuki et al., 2004). Because GAB1 and PAR1 form a complex and GAB1 enhances PAR1 activity, we assessed the effect of GAB1 on apical-basal polarity. Control and GAB1-KD MDCK II cells were cultured on porous filters for 3 days, and then immunostained with anti-ZO-1, anti-E-cadherin and anti-gp135 antibodies. Notably, the lateral membrane (E-cadherin stain) was shortened in GAB1-KD MDCK II cells, compared with controls (Fig. 6A and Fig. S6B), providing further evidence of the functional importance of GAB1/PAR1 interaction. GAB1 depletion also increased the thickness of TJs (i.e., the fraction of the membrane occupied by TJs), without affecting overall cell polarization, as shown by normal apical marker staining (Fig. 6B).

Figure 6. Gab1 over-expression disrupts polarity.

Figure 6

Open in a new tab

(A, B) Control and GAB1-KD MDCK II cells were cultured on filters for 3 days, then co-stained with anti-ZO-1 (red) and anti-E-cadherin (green) antibodies (A), or anti-ZO-1 (red) and anti-gp135 (yellow) antibodies (B). (C) GAB1-KD MDCK II cells reconstituted with WT Gab1 or Gab1 mutants were cultured on filters for 3 days, and then fixed and co-stained with anti-ZO-1 (red) and anti-E-cadherin (green) antibodies. Note disrupted polarity in WT Gab1-expressing cells. (D) GAB1-KD MDCK II cells reconstituted with WT Gab1 or Gab1 mutants were cultured in Matrigel for 3 days. The resultant structures were co-stained with anti-ZO-1 (red) and anti-β-catenin (green) antibodies. Note that WT Gab1 (over)-expressing cells form multi-lumen cysts in Matrigel. Scale bar, 10µm. Also, see Fig. S6.

We next analyzed cells re-expressing WT Gab1 or its PAR1- or PAR3-binding deficient mutants. As noted above (Fig. 5), these Gab proteins were over-expressed in reconstituted cells, and Gab1 over-expression caused a severe loss of polarity: ZO-1 diffused to the basolateral domain and cells lost their typical cobblestone morphology (Fig. 6C). By contrast, ZO-1 and E-cadherin remained well-segregated in cells expressing the PAR1b- or PAR3-binding site mutants (Fig. 6C and Fig. S6C). Cells (over)-expressing WT Gab1 also displayed a wavy apical surface as shown by gp135 staining, while PAR1b- or PAR3-binding deficient mutant-expressing cells retained an even apical surface (Fig. S6A). Thus, GAB1 over-expression disrupts apical-basal polarity in a PAR1b- and PAR3-dependent manner.

MDCK II cells form single lumen cysts when cultured in Matrigel, and PAR3 depletion causes multi-lumen cysts due to abnormal spindle orientation (Bryant et al., 2010; Hao et al., 2010). Our data indicated that Gab1 over-expression negatively regulates PAR3 by increasing its phosphorylation by PAR1. Indeed, MDCK II cells expressing WT Gab1 formed cysts with multiple lumens (Fig. 6D and Fig. S6D). Moreover, cells within these cysts had disrupted apical-basal polarity (as shown by the presence of multiple ZO-1-positive domains), although most cells on the outer surface of the cysts retained normal apical-basal polarity with distinct, properly ordered TJs (ZO-1 stain) and lateral membranes (β-catenin stain) (Fig. S6E). By contrast, cells expressing Gab1 mutants lacking PAR1b- or PAR3-binding sites formed single lumen cysts, providing further evidence that GAB1 controls cell polarity by regulating PAR1 and PAR3.

GAB1/PAR complexes regulate cell scattering

Growth factors can promote the disruption of cell junctions and the conversion of polarized epithelia to motile cells with a mesenchymal phenotype, a process termed “cell scattering.” Gab1 over-expression sensitizes MDCK II cells to HGF-induced scattering via Gab1/SHP2 and Gab1/Pak4 interactions (Paliouras et al., 2009). We wondered whether Gab1/PAR interactions also might be important. Indeed, over-expression of WT Gab1 promoted scattering even in the absence of HGF, whereas scattering in response to an HGF dose (1 ng/ml) that does not promote scattering of WT cells (Fig. 7A, leftmost panels) was potentiated by Gab1 expression (Fig. 7A, pSp-shGab1+pBabe-Gab1, and Fig. S7A). By contrast, vector control (pSp-shGab1+pBabe)-, Gab1Δ152–250 (pSp-shGab1+pBabe-Gab1Δ152–250)- or Gab1Δ301–400 (pSp-shGab1+pBabe-GAB1Δ301–400)-reconstituted cells largely remained in tight epithelioid colonies (Fig. 7A and Fig. S7A). These data indicate that the ability of Gab1 to promote cell scattering is impaired by deletion of its PAR1b- or PAR3-binding sites. Finally, forced dimerization of Gab1Δ152–250-FKBP and PAR1b-F*RB in GAB1-KD MDCK II cells restored HGF-evoked scattering (Fig. 7B and Fig. S7B), confirming that Gab1/PAR1b interaction is required for this process.

Figure 7. GAB1 induces cell scattering via PAR1b/PAR3.

Figure 7

Open in a new tab

(A) GAB1-KD MDCK II cells reconstituted with control vector or the indicated Gab1 constructs were starved in 0.2% FBS, and then left untreated or stimulated with HGF (1 ng/ml). Note that WT Gab1-, but not Gab1Δ152–250- or Gab1Δ301–400-reconstituted, cells dissociated and adopted an elongated morphology. (B) GAB1-KD MDCK II cells stably expressing Gab1Δ152–250-FKBP and PAR1b-F*RB-HA were starved in 0.2% FBS containing AP21697 or EtOH (vehicle), and then left untreated or stimulated with 1 ng/ml HGF. Note that induced dimerization of Gab1Δ152–250-FKBP and PAR1b-F*RB sensitized cells to HGF-induced scattering. (C–E) Role of GAB1 in regulating epithelial cell polarity: (C) In polarized epithelia, PAR3/PAR6/aPKC complex resides at the apical side of TJs (PAR6 and aPKC also localize to the apical domain), while PAR1 and Gab1 form a complex along the lateral membrane. Localization of these complexes is mutually exclusive. (D) PAR3 that “strays” from TJs and invades the lateral domain is “captured” by Gab1 and brought to PAR1 for phosphorylation. (E) Phosphorylated PAR3 is released from the Gab1/PAR1 complex into cytosol, where it can bind PAR5 (not shown). Also, see Fig. S7.

Discussion

During development, epithelial cells polarize and assemble into complex structures, directed, in part, by growth factor and cytokine signaling. These processes are disrupted in many diseases, including cancer, yet how signaling pathways impact the polarity machinery has largely remained elusive. We uncovered a role for GAB proteins, conserved scaffolding adaptors that play essential roles in growth factor/cytokine signaling from worms to humans, in regulating polarity via PAR proteins. GAB proteins regulate epithelial cell scattering and morphogenesis (Rosario and Birchmeier, 2003), but heretofore, these effects have been attributed to their ability to recruit and activate SH2 domain-containing signal relay molecules (e.g., SHP2, PI-3K). The finding that H. pylori CagA, which co-opts GAB protein functions, binds and inhibits PAR1 (Saadat et al., 2007), raised the possibility that GAB proteins have similar effects. Indeed, GAB1 interacts with the polarity machinery, but in an entirely different way. Our studies reveal a novel and general role for GAB1 (and probably other GAB proteins) as a protein scaffold that facilitates mutual exclusion of PAR complexes from TJs and the lateral membranes, respectively, of polarized epithelial cells. As GAB proteins are implicated in several malignancies, our findings also have implications for understanding how polarity is disrupted in cancer.

Gab1, like CagA, binds to PAR1b (and other PAR1 isoforms). But whereas CagA binds and inhibits the PAR1 kinase domain (Saadat et al., 2007), Gab1 binds the KA1 domain (Fig. 1G, H and Fig. S1E, F). In yeast PAR1 orthologs, the KA1 domain, via an intramolecular interaction, inhibits kinase activity (Elbert et al., 2005). Similarly, Gab1 binding to PAR1 disrupts KA1/kinase domain interactions in mammalian cells (Fig. S3C). Gab1 expression also enhances the phosphorylation of the PAR1 substrate TAU (Fig. 3G), indicating that Gab1 activates PAR1 in vivo.

Also unlike CagA, we found that Gab1 interacts with PAR3 (Fig. 2A). The PAR1 and PAR3 binding sites on Gab1 map to distinct regions (Fig. 1E, F and Fig. 2B–D), allowing it to function as a scaffold that co-localizes these proteins and enhances PAR3 phosphorylation by PAR1 (Fig. 3D). In addition, aPKC is brought into this Gab1 complex (Fig. 2A). Gab1/aPKC binding is mediated largely via PAR3 (Fig. S2A, B), although we cannot exclude an additional, weak direct interaction with Gab1. By contrast, Gab1/PAR6 interaction is barely detectable (Fig. 2A), because Gab1 and PAR6 each bind, in a mutually exclusive fashion, to the PAR3 PDZ1 domain (Fig. 2E, Fig. S2C and Fig. 3H).

Our results lead to a model in which GAB1 regulates the PAR3/aPKC/PAR6 complex in two ways (Fig. 3D and Fig. 7C–E). First, PAR1-catalyzed phosphorylation of PAR3 releases PAR3 from membranes, whereupon it is sequestered in the cytosol by 14-3-3 (PAR5) (McCaffrey and Macara, 2009). By enhancing PAR1 phosphorylation of PAR3 (Fig. 3F), GAB1 promotes PAR3 release. GAB1/PAR3/aPKC interaction also prevents PAR3/aPKC/PAR6 complex formation, because GAB1 and PAR6 compete for PAR3 binding (Fig. 3H). Hence, GAB1 facilitates mutual exclusion of PAR complexes ((Macara, 2004a, b); see Introduction), negatively regulating TJ assembly and stability and apical-basal polarity (Fig. 4–6). Conversely, in response to growth factor (e.g., HGF) stimulation, GAB1 accelerates junction breakdown, enhancing cell scattering (Fig. 7A, B).

Multiple lines of evidence support this model. First, lowering GAB1 levels in MDCK II cells leads to decreased phosphorylation of PAR3 (on its PAR1 sites) and increased levels of PAR3-PAR6-aPKC complex (Fig. 3F, I). PAR3-KD MDCK II cells show delayed TJ formation in a calcium switch assay (Chen and Macara, 2005), whereas GAB1-KD cells, as predicted by our model, have the converse phenotype: accelerated recruitment of PAR3 to incipient TJs and precocious TJ formation (Fig. 4A, B and Fig. S4A). Re-expressing WT Gab1, but not Gab1 mutants lacking PAR1 or PAR3 binding, reverses this phenotype (Fig. 5B, C). Furthermore, direct fusion of Gab1 lacking its PAR1 binding domain (Gab1Δ152–250) to PAR1b, or ligand-induced binding of PAR1b to such a mutant, restores Gab1 function (Fig. 5B–E). Finally, Gab1 localizes to lateral membranes, perfectly positioned to perform these functions (Fig. 4D).

GAB1 also affects other aspects of polarity via PAR1. Lowering PAR1 levels or inhibiting its kinase activity causes shortening of the lateral membranes of MDCK II cells (Cohen et al., 2004; Suzuki et al., 2004). Consistent with our finding that Gab1 enhances PAR1 activity (Fig. 3G), GAB1-KD MDCK II cells also have shortened lateral membranes (Fig. 6A and Fig. S6B). PAR1 has multiple substrates besides PAR3, including cytoskeleton-regulating proteins such as IRSp53, tau, MAP2, and MAP4 (Benton and St Johnston, 2003; Cohen et al., 2011; Marx et al., 2010); which of these contribute(s) to determining the height of the lateral membrane remains to be elucidated. Conversely, GAB1 over-expression disrupts apical-basal polarity and causes multi-lumen cysts in 3D cultures (Fig. 6). Notably, PAR3 depletion or PAR1b over-expression also leads to multi-lumen cysts (Bryant et al., 2010; Cohen et al., 2011; Hao et al., 2010). Finally, GAB1 enhances growth factor (e.g., HGF)-induced cell scattering. Previous studies showed that GAB1 must interact with SHP2 and Pak4 to mediate scattering (Paliouras et al., 2009); our studies now implicate GAB1/PAR1 and GAB1/PAR3 interactions in this process as well (Fig. 7). Reconstituting GAB1-KD MDCK II cells with WT Gab1, but not mutants lacking either PAR1 (Gab1Δ152–250)- or PAR3 (Gab1Δ301–400)- binding sites, sensitizes MDCK II cells to HGF-induced scattering.. Moreover, forced dimerization of Gab1Δ152–250 and PAR1 directly promotes scattering, implicating PAR1 as the key binding protein within the deleted region (152–250).

At first glance, our results appear to conflict with the ability of CagA (which inhibits PAR1) to evoke the “hummingbird phenotype” (which is essentially cell scattering) in gastric epithelial cells, and its ability to promote spreading of TJ markers to the lateral membrane of MDCK-II cells (Saadat et al., 2007). One possible explanation for this discrepancy could be that over-expression of CagA completely inhibits PAR1 family members; by contrast, GAB1 depletion decreases, but does not abolish, PAR1 activity. In addition, the CagA-induced “hummingbird phenotype” reportedly requires simultaneous inhibition of PAR1 and activation of SHP2 by CagA. Conversely, in our shRNA-mediated depletion and reconstitution experiments, we found that the Gab1/PAR1 complex positively regulates cell scattering. Saadat et al. transiently over-expressed PAR1b to explore its effects on CagA-induced scattering, which could have led to non-physiological effects. By contrast, we studied stable MDCK II cell pools expressing PAR1b at endogenous levels (Fig. 1D, left panel). Unlike Saadat et al., we found that induced interaction of a Gab1 mutant lacking its PAR1-binding site (Gab1Δ152–250) with PAR1b enhances HGF-induced scattering, whereas stable expression of PAR1 in WT, but not GAB1-depleted, cells delays TJ formation (Fig. S4C). These data clearly indicate that PAR1 regulates TJ formation rate in a GAB1-dependent manner. Saadat et al. also found that PAR1b promotes CagA oligomerization, which is required to potentiate CagA binding to SHP2. Conceivably, PAR1b activity must be inhibited to allow CagA oligomerization, in which case, CagA could have evolved to inhibit PAR1 kinase activity.

Although our results show that they act differently than CagA, previous studies nonetheless implicate GAB proteins in oncogenesis. For example, Gab1 is required for Met-induced transformation (Fixman et al., 1997; Mood et al., 2006). GAB2 amplification/over-expression is implicated in melanoma cell migration and invasion (Horst et al., 2009), and Gab2 is required for BCR/ABL-induced transformation (Sattler et al., 2002) and leukemogenesis (W. Chan, G. Mohi, S. Gu, S Li, B.G. Neel and R.V. Etten, manuscript in revision). GAB2 over-expression promotes the proliferation and migration of MCF-10A breast epithelial cells, whereas co-expression of GAB2 and ErbB2/Neu causes multi-acinar structures and invasion (Bentires-Alj et al., 2006; Brummer et al., 2006). Transgenic GAB2 over-expression enhances, whereas Gab2 deletion impairs, Neu-evoked tumorigenesis, and GAB2 is amplified in 10–15% of human breast tumors (Bentires-Alj et al., 2006; Ke et al., 2007). The latter effects have been attributed to GAB2 effects on the RAS/ERK and RHOA pathways. However, ErbB2/Neu activation also causes disruption of the PAR3/PAR6/aPKC complex (Aranda et al., 2006), although the detailed mechanism has remained unclear. Our findings that ErbB2/Neu activation induces GAB1/PAR3 complex formation (Fig. S2E) and that GAB1 negatively regulates TJ formation strongly implicate GAB proteins as key intermediaries in ErbB2/Neu effects on the polarity machinery. Ongoing studies should clarify the extent to which GAB/PAR protein complexes contribute to the oncogenic activity of GAB proteins.

Experimental Procedures

Calcium Switch Assay

Calcium switch assays were performed as described (Cohen and Musch, 2003). Briefly, cells were plated at high density onto 0.4 µm pore size Transwell filters (Corning, NY) in Ca2+-free Suspension-MEM (S-MEM, Sigma, MO) with 10% dialyzed fetal bovine serum (FBS, GibcoBRL, NE) and cultured for 20 hrs. Calcium switch was initiated by changing the media to S-MEM containing 1.8mM Ca2+.

Microscopy

Images were acquired using a Leica DFC 420 fluorescence microscope or an Olympus FluoView 1000 IX-81 laser scanning confocal microscope.

TER Measurements

Triplicate wells of MDCK II cells were plated at high density onto Transwell filters (12mm, 0.4µm pore size) in S-MEM. At various times after adding back Ca2+, resistance across the monolayer was measured with a Millicell ERS Volt-Ohm meter (Millipore, MA).

Cell Scattering

Cells were plated in MEM plus 10% FBS onto 6-well plates (15, 000 cells/well), cultured for 2 days, and then starved in MEM containing 0.2% FBS with or without the AP21697 dimerizer for 6 hours, followed by treatment with 1 ng/ml HGF (PeproTech, NJ) for 18 hours. Images were acquired on a Leica DMI 3000 B microscope. Scattering was quantified by calculating the percentage of cells that dissociated and acquired elongated morphology.

Statistics

Data are presented as mean ± SD. Statistical significance was assessed using two-way ANOVA. Detailed Experimental Procedures are provided in the Supplementary Information.

Supplementary Material

01

Highlights.

  • GAB1 binds to and activates PAR1

  • GAB1 attenuates PAR3-PAR6-aPKC complex assembly

  • GAB1 controls tight junctions, epithelial resistance and lateral domain height.

  • GAB/PAR complexes regulate epithelial cell scattering

Acknowledgements

This work was supported by NIH grant R37 49152 to B.G.N., a grant from the Canadian Cancer Society Research Institute to M.I., grants CA098830 and BC075024, Era of Hope Scholar award from DOD Breast Cancer Research Program to S.K.M., and funding from the Princess Margaret Hospital Foundation and the Ontario Ministry of Health and Long Term Care. B.G.N. and M.I. are Canada Research Chairs, Tier 1, Z. Y. is the recipient of a post-doctoral fellowship from The Terry Fox Foundation, and M. U. received fellowships from The Uehara Memorial Foundation, The Naito Foundation and the EIRR21st Research Training Program. We thank Drs. T. Pawson (Samuel Lunenfeld Research Institute), M. Park (McGill University), I. Macara (University of Virginia) and H. Piwnica-Worms (Washington University Medical School) for constructs and Dr. G. Keller (Ontario Cancer Institute) for helpful comments.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Aranda V, Haire T, Nolan ME, Calarco JP, Rosenberg AZ, Fawcett JP, Pawson T, Muthuswamy SK. Par6-aPKC uncouples ErbB2 induced disruption of polarized epithelial organization from proliferation control. Nat Cell Biol. 2006;8:1235–1245. doi: 10.1038/ncb1485. [DOI] [PubMed] [Google Scholar]
  2. Bentires-Alj M, Gil SG, Chan R, Wang ZC, Wang Y, Imanaka N, Harris LN, Richardson A, Neel BG, Gu H. A role for the scaffolding adapter GAB2 in breast cancer. Nat Med. 2006;12:114–121. doi: 10.1038/nm1341. [DOI] [PubMed] [Google Scholar]
  3. Benton R, St Johnston D. Drosophila PAR-1 and 14-3-3 inhibit Bazooka/PAR-3 to establish complementary cortical domains in polarized cells. Cell. 2003;115:691–704. doi: 10.1016/s0092-8674(03)00938-3. [DOI] [PubMed] [Google Scholar]
  4. Brummer T, Schramek D, Hayes VM, Bennett HL, Caldon CE, Musgrove EA, Daly RJ. Increased proliferation and altered growth factor dependence of human mammary epithelial cells overexpressing the Gab2 docking protein. J Biol Chem. 2006;281:626–637. doi: 10.1074/jbc.M509567200. [DOI] [PubMed] [Google Scholar]
  5. Bryant DM, Datta A, Rodriguez-Fraticelli AE, Peranen J, Martin-Belmonte F, Mostov KE. A molecular network for de novo generation of the apical surface and lumen. Nat Cell Biol. 2010;12:1035–1045. doi: 10.1038/ncb2106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chen X, Macara IG. Par-3 controls tight junction assembly through the Rac exchange factor Tiam1. Nat Cell Biol. 2005;7:262–269. doi: 10.1038/ncb1226. [DOI] [PubMed] [Google Scholar]
  7. Cohen D, Brennwald PJ, Rodriguez-Boulan E, Musch A. Mammalian PAR-1 determines epithelial lumen polarity by organizing the microtubule cytoskeleton. J Cell Biol. 2004;164:717–727. doi: 10.1083/jcb.200308104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen D, Fernandez D, Lazaro-Dieguez F, Musch A. The serine/threonine kinase Par1b regulates epithelial lumen polarity via IRSp53-mediated cell-ECM signaling. J Cell Biol. 2011;192:525–540. doi: 10.1083/jcb.201007002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cohen D, Musch A. Apical surface formation in MDCK cells: regulation by the serine/threonine kinase EMK1. Methods. 2003;30:269–276. doi: 10.1016/s1046-2023(03)00033-1. [DOI] [PubMed] [Google Scholar]
  10. Cong W, Hirose T, Harita Y, Yamashita A, Mizuno K, Hirano H, Ohno S. ASPP2 regulates epithelial cell polarity through the PAR complex. Curr Biol. 2010;20:1408–1414. doi: 10.1016/j.cub.2010.06.024. [DOI] [PubMed] [Google Scholar]
  11. Elbert M, Rossi G, Brennwald P. The yeast par-1 homologs kin1 and kin2 show genetic and physical interactions with components of the exocytic machinery. Mol Biol Cell. 2005;16:532–549. doi: 10.1091/mbc.E04-07-0549. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fixman ED, Holgado-Madruga M, Nguyen L, Kamikura DM, Fournier TM, Wong AJ, Park M. Efficient cellular transformation by the Met oncoprotein requires a functional Grb2 binding site and correlates with phosphorylation of the Grb2-associated proteins, Cbl and Gab1. J Biol Chem. 1997;272:20167–20172. doi: 10.1074/jbc.272.32.20167. [DOI] [PubMed] [Google Scholar]
  13. Gu H, Neel BG. The "Gab" in signal transduction. Trends Cell Biol. 2003;13:122–130. doi: 10.1016/s0962-8924(03)00002-3. [DOI] [PubMed] [Google Scholar]
  14. Hao Y, Du Q, Chen X, Zheng Z, Balsbaugh JL, Maitra S, Shabanowitz J, Hunt DF, Macara IG. Par3 controls epithelial spindle orientation by aPKC-mediated phosphorylation of apical Pins. Curr Biol. 2010;20:1809–1818. doi: 10.1016/j.cub.2010.09.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hatakeyama M. Helicobacter pylori CagA--a potential bacterial oncoprotein that functionally mimics the mammalian Gab family of adaptor proteins. Microbes Infect. 2003;5:143–150. doi: 10.1016/s1286-4579(02)00085-0. [DOI] [PubMed] [Google Scholar]
  16. Hatakeyama M. Oncogenic mechanisms of the Helicobacter pylori CagA protein. Nat Rev Cancer. 2004;4:688–694. doi: 10.1038/nrc1433. [DOI] [PubMed] [Google Scholar]
  17. Hirose T, Karasawa M, Sugitani Y, Fujisawa M, Akimoto K, Ohno S, Noda T. PAR3 is essential for cyst-mediated epicardial development by establishing apical cortical domains. Development (Cambridge, England) 2006;133:1389–1398. doi: 10.1242/dev.02294. [DOI] [PubMed] [Google Scholar]
  18. Horst B, Gruvberger-Saal SK, Hopkins BD, Bordone L, Yang Y, Chernoff KA, Uzoma I, Schwipper V, Liebau J, Nowak NJ, et al. Gab2-mediated signaling promotes melanoma metastasis. Am J Pathol. 2009;174:1524–1533. doi: 10.2353/ajpath.2009.080543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hurd TW, Fan S, Liu CJ, Kweon HK, Hakansson K, Margolis B. Phosphorylation-dependent binding of 14-3-3 to the polarity protein Par3 regulates cell polarity in mammalian epithelia. Curr Biol. 2003;13:2082–2090. doi: 10.1016/j.cub.2003.11.020. [DOI] [PubMed] [Google Scholar]
  20. Hurov JB, Huang M, White LS, Lennerz J, Choi CS, Cho YR, Kim HJ, Prior JL, Piwnica-Worms D, Cantley LC, et al. Loss of the Par-1b/MARK2 polarity kinase leads to increased metabolic rate, decreased adiposity, and insulin hypersensitivity in vivo. Proc Natl Acad Sci U S A. 2007;104:5680–5685. doi: 10.1073/pnas.0701179104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hurov JB, Stappenbeck TS, Zmasek CM, White LS, Ranganath SH, Russell JH, Chan AC, Murphy KM, Piwnica-Worms H. Immune system dysfunction and autoimmune disease in mice lacking Emk (Par-1) protein kinase. Mol Cell Biol. 2001;21:3206–3219. doi: 10.1128/MCB.21.9.3206-3219.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Itoh M, Yoshida Y, Nishida K, Narimatsu M, Hibi M, Hirano T. Role of Gab1 in heart, placenta, and skin development and growth factor- and cytokine-induced extracellular signal-regulated kinase mitogen-activated protein kinase activation. Mol Cell Biol. 2000;20:3695–3704. doi: 10.1128/mcb.20.10.3695-3704.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Ke Y, Wu D, Princen F, Nguyen T, Pang Y, Lesperance J, Muller WJ, Oshima RG, Feng GS. Role of Gab2 in mammary tumorigenesis and metastasis. Oncogene. 2007;26:4951–4960. doi: 10.1038/sj.onc.1210315. [DOI] [PubMed] [Google Scholar]
  24. Kemphues K. PARsing embryonic polarity. Cell. 2000;101:345–348. doi: 10.1016/s0092-8674(00)80844-2. [DOI] [PubMed] [Google Scholar]
  25. Lee HS, Nishanian TG, Mood K, Bong YS, Daar IO. EphrinB1 controls cell-cell junctions through the Par polarity complex. Nat Cell Biol. 2008;10:979–986. doi: 10.1038/ncb1758. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lennerz JK, Hurov JB, White LS, Lewandowski KT, Prior JL, Planer GJ, Gereau RWt, Piwnica-Worms D, Schmidt RE, Piwnica-Worms H. Loss of Par-1a/MARK3/C-TAK1 kinase leads to reduced adiposity, resistance to hepatic steatosis, and defective gluconeogenesis. Mol Cell Biol. 2010;30:5043–5056. doi: 10.1128/MCB.01472-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Macara IG. Par proteins: partners in polarization. Curr Biol. 2004a;14:R160–R162. [PubMed] [Google Scholar]
  28. Macara IG. Parsing the polarity code. Nat Rev Mol Cell Biol. 2004b;5:220–231. doi: 10.1038/nrm1332. [DOI] [PubMed] [Google Scholar]
  29. Maroun CR, Holgado-Madruga M, Royal I, Naujokas MA, Fournier TM, Wong AJ, Park M. The Gab1 PH domain is required for localization of Gab1 at sites of cell-cell contact and epithelial morphogenesis downstream from the met receptor tyrosine kinase. Mol Cell Biol. 1999;19:1784–1799. doi: 10.1128/mcb.19.3.1784. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Maroun CR, Naujokas MA, Holgado-Madruga M, Wong AJ, Park M. The tyrosine phosphatase SHP-2 is required for sustained activation of extracellular signal-regulated kinase and epithelial morphogenesis downstream from the met receptor tyrosine kinase. Mol Cell Biol. 2000;20:8513–>8525. doi: 10.1128/mcb.20.22.8513-8525.2000. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Marx A, Nugoor C, Panneerselvam S, Mandelkow E. Structure and function of polarity-inducing kinase family MARK/Par-1 within the branch of AMPK/Snf1-related kinases. Faseb J. 2010;24:1637–1648. doi: 10.1096/fj.09-148064. [DOI] [PubMed] [Google Scholar]
  32. McCaffrey LM, Macara IG. Widely conserved signaling pathways in the establishment of cell polarity. Cold Spring Harb Perspect Biol. 2009;1:a001370. doi: 10.1101/cshperspect.a001370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mood K, Saucier C, Bong YS, Lee HS, Park M, Daar IO. Gab1 is required for cell cycle transition, cell proliferation, and transformation induced by an oncogenic met receptor. Mol Biol Cell. 2006;17:3717–3728. doi: 10.1091/mbc.E06-03-0244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Morais-de-Sa E, Mirouse V, St Johnston D. aPKC phosphorylation of Bazooka defines the apical/lateral border in Drosophila epithelial cells. Cell. 2010;141:509–523. doi: 10.1016/j.cell.2010.02.040. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Paliouras GN, Naujokas MA, Park M. Pak4, a novel Gab1 binding partner, modulates cell migration and invasion by the Met receptor. Mol Cell Biol. 2009;29:3018–3032. doi: 10.1128/MCB.01286-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rosario M, Birchmeier W. How to make tubes: signaling by the Met receptor tyrosine kinase. Trends Cell Biol. 2003;13:328–335. doi: 10.1016/s0962-8924(03)00104-1. [DOI] [PubMed] [Google Scholar]
  37. Saadat I, Higashi H, Obuse C, Umeda M, Murata-Kamiya N, Saito Y, Lu H, Ohnishi N, Azuma T, Suzuki A, et al. Helicobacter pylori CagA targets PAR1/MARK kinase to disrupt epithelial cell polarity. Nature. 2007;447:330–333. doi: 10.1038/nature05765. [DOI] [PubMed] [Google Scholar]
  38. Sachs M, Brohmann H, Zechner D, Muller T, Hulsken J, Walther I, Schaeper U, Birchmeier C, Birchmeier W. Essential role of Gab1 for signaling by the c-Met receptor in vivo. J Cell Biol. 2000;150:1375–1384. doi: 10.1083/jcb.150.6.1375. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sattler M, Mohi MG, Pride YB, Quinnan LR, Malouf NA, Podar K, Gesbert F, Iwasaki H, Li S, Van Etten RA, et al. Critical role for Gab2 in transformation by BCR/ABL. Cancer Cell. 2002;1:479–492. doi: 10.1016/s1535-6108(02)00074-0. [DOI] [PubMed] [Google Scholar]
  40. Sottocornola R, Royer C, Vives V, Tordella L, Zhong S, Wang Y, Ratnayaka I, Shipman M, Cheung A, Gaston-Massuet C, et al. ASPP2 binds Par-3 and controls the polarity and proliferation of neural progenitors during CNS development. Dev Cell. 2010;19:126–137. doi: 10.1016/j.devcel.2010.06.003. [DOI] [PubMed] [Google Scholar]
  41. Suzuki A, Hirata M, Kamimura K, Maniwa R, Yamanaka T, Mizuno K, Kishikawa M, Hirose H, Amano Y, Izumi N, et al. aPKC acts upstream of PAR-1b in both the establishment and maintenance of mammalian epithelial polarity. Curr Biol. 2004;14:1425–1435. doi: 10.1016/j.cub.2004.08.021. [DOI] [PubMed] [Google Scholar]
  42. Suzuki A, Ohno S. The PAR-aPKC system: lessons in polarity. J Cell Sci. 2006;119:979–987. doi: 10.1242/jcs.02898. [DOI] [PubMed] [Google Scholar]
  43. Wohrle FU, Daly RJ, Brummer T. Function, regulation and pathological roles of the Gab/DOS docking proteins. Cell Commun Signal. 2009;7:22. doi: 10.1186/1478-811X-7-22. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS399681-supplement-01.doc (5MB, doc)

RESOURCES