Research ArticleCell Biology

The subcellular localization and activity of cortactin is regulated by acetylation and interaction with Keap1

See allHide authors and affiliations

Sci. Signal.  24 Nov 2015:
Vol. 8, Issue 404, pp. ra120
DOI: 10.1126/scisignal.aad0667

Acetylation against cell migration

The actin-binding protein cortactin promotes cell migration through cytoskeletal remodeling in the cell cortex and is abundant in certain types of aggressive cancers. Ito et al. found that the cytosolic protein Keap1 promoted the localization of cortactin to the cell cortex and thus cell migration. Cortactin shuttled between the cytoplasm and the nucleus; however, upon acetylation, cortactin no longer bound to Keap1 and became predominantly localized in the nucleus. Thus, increasing the acetylation of cortactin or preventing it from binding to Keap1 may suppress the metastasis of cancer cells.

Abstract

Cortactin is an F-actin–binding protein that localizes to the cell cortex, where the actin remodeling that is required for cell migration occurs. We found that cortactin shuttled between the cytoplasm and the nucleus under basal conditions. We identified Kelch-like ECH-associated protein 1 (Keap1), a cytosolic protein that is involved in oxidant stress responses, as a binding partner of cortactin that promoted the cortical localization of cortactin and cell migration. The ability of cortactin to promote cell migration is regulated by various posttranslational modifications, including acetylation. We showed that the acetylated form of cortactin was mainly localized to the nucleus and that acetylation of cortactin decreased cell migration by inhibiting the binding of cortactin to Keap1. Our findings reveal that Keap1 regulates cell migration by affecting the subcellular localization and activity of cortactin independently of its role in oxidant stress responses.

INTRODUCTION

The Kelch-like ECH-associated protein 1 (Keap1)/nuclear factor (erythroid-derived 2)–like 2 (Nrf2) system is a critical cellular defense mechanism against oxidative and electrophilic stresses (1). Keap1 tethers the basic leucine zipper transcription factor Nrf2 on actin filaments and acts as a ubiquitin E3 ligase for Nrf2. Under basal conditions, Keap1 binds to and constantly ubiquitinates Nrf2, resulting in the proteasome-mediated degradation of Nrf2 in the cytoplasm. Oxidative stress inactivates Keap1, resulting in the stabilization and nuclear translocation of Nrf2, which activates the transcription of genes encoding factors involved in detoxifying oxidants and electrophiles (2). Thus, Keap1 acts as an intracellular sensor for oxidative and electrophilic stresses, thereby regulating Nrf2-mediated gene expression.

The actin-binding protein cortactin is a central regulator of the actin cytoskeleton. Cortactin is targeted to sites of actin polymerization in cells, including lamellipodia in migrating cells, cell-cell junctions in epithelial cells, growth cones of neurons, podosomes of osteoclasts, invadopodia of tumor cells, and sites of actin rearrangement induced by bacteria and viruses (36). In addition, cortactin may be a biomarker for aggressive cancers (7, 8). Amplification of the CTTN gene and increased protein abundance of cortactin have been frequently observed in primary metastatic breast carcinomas, in head and neck squamous cell carcinomas (7, 9), and in invasive cancers including melanoma (10) and colorectal cancer (11). Cortactin is composed of multiple domains including an N-terminal acidic (NTA) domain, cortactin repeats, an α-helical domain, a proline-rich domain, and an Src homology 3 (SH3) domain. The NTA domain nucleates actin polymerization through direct interaction with the Arp2/3 complex (12). The cortactin repeat domain, which binds to filamentous actin (F-actin) (13), consists of six units of 37–amino acid tandem repeating segments and one incomplete segment. The SH3 domain binds effector molecules including regulators of actin polymerization such as WIP and N-WASP, factors involved in endocytosis and vesicle trafficking such as dynamin-2, regulators of guanosine triphosphatases such as FGD1 and BPGAP, and adaptor and/or scaffold proteins such as CD2AP, CortBP1, ZO-1, and Shank (36).

Cortactin is regulated not only by various binding partners but also by several posttranslational modifications. Cortactin is phosphorylated on multiple tyrosine and serine/threonine residues in cells stimulated by numerous factors, including several growth factors (14). Cortactin is also acetylated (15) and deacetylated by various proteins, including histone deacetylase 6 (HDAC6), a zinc-dependent class IIb histone deacetylase (15). In cells stimulated with growth factors, HDAC6 translocates together with cortactin to the cell periphery where HDAC6 deacetylates cortactin, leading to enhanced cortactin binding to F-actin and thereby stimulating cell migration (15). Accordingly, HDAC6 inhibitors reduce cell motility (16). Two nicotinamide adenine dinucleotide–dependent class III protein deacetylases, SIRT1 (sirtuin 1) and SIRT2, can also deacetylate cortactin (1719). However, in contrast to the extensively characterized cortactin deacetylases (15, 1720), the enzymes responsible for cortactin acetylation have not been well characterized.

Here, we found that CBP [CREB (cyclic adenosine monophosphate response element–binding protein)–binding protein] acetylated cortactin in the nucleus. In addition, we identified Keap1 as a binding partner of cortactin. Keap1 tethered cortactin in the cytoplasm and promoted actin rearrangement and cell migration. Furthermore, Keap1 was indispensable for efficient translocation of cortactin to the cell periphery, and cell migration was reduced in cells with Keap1 knockdown, suggesting a role for Keap1 in cortactin-mediated cell migration.

RESULTS

Acetylated cortactin is localized in the nucleus

To determine the subcellular localization of acetylated cortactin, we generated a specific antibody that recognizes acetylated peptides of human cortactin (21). As previously reported (15), immunoblot analysis using an antibody against acetylated lysine (AcLys) revealed that acetylation of wild-type cortactin was detected in cells treated with the HDAC inhibitors trichostatin A (TSA) (22) and nicotinamide (NA) (fig. S1A). We next investigated whether all seven conserved lysines were acetylated sites. The 7KR mutant, in which all the putative lysine residues were replaced with arginine, was not detectably acetylated. By contrast, all of the 6KR mutants, in each of which one of the seven lysine residues remains, were acetylated in the presence of TSA and NA. These results suggested that all seven conserved lysine residues in the repeat sequences of human cortactin could be acetylated and that other potential acetylation sites, if any, were not detectably acetylated (fig. S1A). Because Lys309 was acetylated to the greatest extent, we used a peptide containing acetylated Lys309 as an antigen to generate another antibody, which we confirmed (anti–Ac-cttn) specifically recognized all the acetylated sites of cortactin (fig. S1B). Although it appeared to preferentially react with acetylated Lys309, it was probably because Lys309 was the residue with the greatest acetylation (fig. S1A). Using the anti–Ac-cttn antibody, we found that knockdown of HDAC6 and SIRT2 by small interfering RNA (siRNA) increased the acetylation of cortactin (fig. S2, A and B), consistent with previous studies indicating that these two enzymes act as deacetylases for cortactin (15, 1820). On the other hand, the acetylation of cortactin was only slightly increased in cells with knockdown of SIRT1 (fig. S2B), which can deacetylate cortactin in cells (17). The discrepancy could be due to the variation of the subcellular localization of SIRT1 among cell lines because SIRT1 in the cells used in the previous study is localized in both the nucleus and the cytoplasm, whereas it was mainly localized in the nucleus in the cell lines that we used (fig. S3).

To identify the enzyme that acetylated cortactin, we overexpressed several major HATs in COS-7 or 293T cells and measured cortactin acetylation by immunoblotting with anti–Ac-cttn (Fig. 1A) or anti-AcLys antibody (fig. S4A). Cortactin acetylation was increased when CBP was overexpressed and slightly increased by Tip60 overexpression but unaffected by overexpression of PCAF (p300/CBP-associated factor) or p300. Immunoblotting using anti–Ac-cttn and anti-AcLys antibodies gave similar results. Overexpression of a CBP mutant lacking catalytic activity (23) failed to increase acetylation, indicating that the intrinsic acetyltransferase activity of CBP was required for cortactin acetylation (fig. S4A). In addition, the acetylation of cortactin was reduced in CBP- but not p300- or PCAF-knockdown cells (Fig. 1B and fig. S4B), and reintroduction of CBP into CBP-knockdown cells restored the cortactin acetylation (fig. S4C). Furthermore, purified CBP directly acetylated cortactin as efficiently as histones in vitro (Fig. 1C). These results indicated that CBP functioned as the major acetyltransferase of cortactin in COS-7 cells, although Tip60 may contribute to the acetylation of cortactin to a lesser extent in cells.

Fig. 1 Cortactin acetylated by CBP is localized in the nucleus.

(A) Effect of CBP overexpression on cortactin acetylation. Proteins were immunoprecipitated (IP) using the indicated antibodies from 293T cells expressing Flag- or hemagglutinin (HA)–tagged histone acetyltransferases (HATs). Immunoprecipitates and total cell lysates were analyzed by immunoblotting (IB) as indicated. n = 4 independent experiments. (B) Reduction of cortactin acetylation by CBP knockdown. COS-7 cells transfected with control (Cont) or siRNA oligonucleotides targeting CBP or p300 mRNA were treated with TSA and NA. Acetylation of cortactin was detected as described in (A). Knockdown efficiency was measured by immunoblotting whole-cell lysates as indicated. n = 2 independent experiments. (C) In vitro acetylation assay. Recombinant Flag-tagged cortactin or histones were incubated with acetyl coenzyme A (acetyl-CoA) and recombinant CBP. The acetylation of cortactin or histone H3 was determined by immunoblotting with anti–Ac-cttn or anti-AcLys, respectively. n = 2 independent experiments. BSA, bovine serum albumin. (D) Subcellular localization of acetylated cortactin. COS-7 cells expressing either Flag-tagged wild-type (WT) cortactin or the 7KR mutant were treated with or without TSA and NA and stained with the indicated antibodies and 4′,6-diamidino-2-phenylindole (DAPI). For the peptide competition assay, 100-fold molar excess of the acetylated peptide (AcLys309) or the control peptide (Lys309) was included in the preparation of cells overexpressing WT cortactin. The fluorescence intensity of acetylated cortactin was quantified. At least 200 cells per experiment were analyzed. Data are means ± SD of three independent experiments. **P < 0.01. Pictures show representative examples. Scale bars, 25 μm. (E) COS-7 cells transfected with control or siRNA oligonucleotides targeting CBP or p300 mRNA were stained with the indicated antibodies and DAPI. Fluorescence intensity of nuclear acetylated cortactin was quantified. At least 900 cells per experiment were analyzed. Data are means ± SD of four independent experiments. **P < 0.01. Pictures show representative examples. Scale bars, 50 μm.

Because CBP was almost exclusively localized in the nucleus in both of these cell types (fig. S5), we asked whether cortactin was acetylated in the nucleus. Immunostaining with the anti–Ac-cttn antibody indicated that consistent with the nuclear localization of CBP, both exogenously expressed and endogenous cortactin proteins were exclusively localized in the nucleus when acetylated, although total cortactin protein was mainly localized in the cytoplasm (Fig. 1D and fig. S6). This nuclear signal was specific because it was lost when the 7KR mutant was expressed instead of wild-type cortactin, when cortactin was knocked down, or when the fixed cells were incubated with the cortactin peptide containing acetylated Lys309 but not nonacetylated peptide as a competitor (Fig. 1D and figs. S6 and S7). The background signal in cells overexpressing the 7KR mutant appeared to be due to the detection of the endogenous acetylated cortactin, because the signal for acetylated cortactin in cells overexpressing the 7KR mutant was almost the same as that of the endogenous cortactin in untransfected cells (Fig. 1D). Consistent with these findings, the nuclear signal of acetylated cortactin was significantly reduced in cells with knockdown of CBP but not that of p300 (Fig. 1E). Treatment with TSA and NA induced an increase in the signal intensity, as well as the cytoplasmic signal of acetylated cortactin (Fig. 1D), suggesting that acetylated cortactin that accumulated in the nucleus was ultimately distributed into the cytoplasm. Together, these results suggested that cortactin was acetylated in the nucleus primarily by the acetyltransferase activity of nuclear CBP.

Cortactin shuttles between the cytoplasm and the nucleus

Because acetylated cortactin was specifically localized in the nucleus, it seemed likely that cortactin normally shuttled between the cytoplasm and the nucleus, although its steady-state localization was primarily cytoplasmic. To test this possibility, we analyzed the effect of leptomycin B (LMB), a selective inhibitor of chromosomal region maintenance 1 (CRM1)–mediated nuclear export of proteins (24, 25), on the localization of endogenous cortactin. Immunostaining using an anti-cortactin antibody revealed that LMB treatment caused nuclear accumulation of cortactin, indicating that cortactin is exported out of the nucleus by CRM1 (Fig. 2A). We next determined the nuclear export signal (NES) in cortactin by testing the subcellular localization of a series of deletion mutants (Fig. 2B). Although a pool of ΔRepeat mutant remained in the cytoplasm, the ΔSH3, N-terminal, and Repeat mutants were localized in the nucleus even in the absence of LMB and no longer changed their localization upon LMB treatment (Fig. 2C), suggesting that the NES is in the SH3 domain. We identified two putative NES-like sequences, NES1 and NES2, in the SH3 region (fig. S8A). The NES1 mutant (NES1-4A), like wild-type, exhibited cytoplasmic localization that was altered by LMB, whereas the NES2 mutant (NES2-4A) accumulated in the nucleus even in the absence of LMB (Fig. 2D). These results indicated that NES2 in the SH3 region was the major functional NES in cortactin.

Fig. 2 Cortactin shuttles between the cytoplasm and the nucleus.

(A) Nuclear accumulation of cortactin by LMB treatment. LMB-treated HeLa cells were stained with the anti-cttn antibody and DAPI. The fluorescence intensities of nuclear and cytoplasmic cortactin were quantified. At least 8000 cells per experiment were analyzed. Data are means ± SD of six independent experiments. **P < 0.01. Pictures show representative examples. Scale bars, 25 μm. (B) Schematic representation of the cortactin mutants. N-ter, N-terminal; C-ter, C-terminal. (C and D) Subcellular localization of cortactin mutants (C) and identification of the functional NES in cortactin (D). HeLa cells expressing Flag-tagged WT cortactin or the mutants described in (B) or fig. S7A were treated with LMB and immunostained with the anti-Flag antibody. The fluorescence intensities of nuclear and cytoplasmic cortactin were quantified. At least 600 cells per experiment were analyzed for both (C) and (D). Data are means ± SD of three independent experiments. **P < 0.01. Pictures show representative examples. Scale bar, 25 μm. (E) Increased acetylation of cortactin after treatment with TSA/NA and LMB. A549 cells treated with various combinations of TSA/NA and LMB were analyzed by immunoblotting. n = 3 independent experiments. (F) Increased acetylation of the NES2 mutant. HeLa cells expressing Flag-tagged WT cortactin, NES1-4A, or NES2-4A were treated with the indicated combinations of TSA, NA, and LMB. Proteins immunoprecipitated with anti-Flag antibody were analyzed by immunoblotting. n = 2 independent experiments. (G) Effect of HDAC6 or SIRT1 knockdown on LMB-induced cortactin acetylation. A549 cells transfected with control or siRNA oligonucleotides targeting SIRT1 or HDAC6 mRNA were treated with LMB. The cell lysates were immunoblotted with the indicated antibodies. n = 4 independent experiments.

Nuclear accumulation enhanced the acetylation of cortactin

Next, we tested whether the acetylation level of cortactin is affected by LMB-induced nuclear localization. The acetylation of endogenous cortactin was increased by LMB in A549 or COS-7 cells (Fig. 2E), suggesting that acetylation occurred in the nucleus and that nuclear import was sufficient to induce acetylation of cortactin. Treatment with TSA and NA caused an increase in the acetylation of cortactin in the presence of LMB (fig. S8B), suggesting that deacetylation of cortactin also occurs in the nucleus. Next, we investigated the acetylation of the cortactin NES2 mutant, which was constitutively localized in the nucleus (Fig. 2D). The acetylation of the NES2 mutant in TSA- and NA-treated cells was high even in the absence of LMB and was not increased by LMB treatment. By contrast, acetylation of wild-type or the NES1 mutant increased further by the combination of LMB, TSA, and NA (Fig. 2F). We hypothesized that deacetylation of cortactin in the nucleus is mediated by SIRT1, which deacetylated cortactin (fig. S3) (17). In COS-7 cells, knockdown of SIRT1 but not HDAC6 resulted in a significant increase in the acetylation of cortactin upon treatment with LMB (Fig. 2G). These results suggested that the acetylation of nuclear cortactin was regulated by CBP and SIRT1.

Keap1 binds cortactin

To investigate the mechanism by which cortactin shuttles between the nucleus and the cytoplasm, we conducted a proteomic screen for cortactin-interacting proteins using Flag-tagged cortactin, which identified Keap1 as a protein associated with cortactin. Keap1 is an inhibitor of Nrf2, a transcription factor involved in oxidative stress responses (26). Reciprocal immunoprecipitation assays revealed that cortactin interacted with Keap1 (Fig. 3, A and B). Pull-down assays using purified recombinant glutathione S-transferase (GST)–fused full-length Keap1 and cortactin revealed that cortactin bound directly to Keap1 (fig. S9A). To determine the domains of cortactin and Keap1 responsible for their interaction with each other, we tested the binding affinities of a series of deletion mutants of cortactin (Fig. 2B) and Keap1 (fig. S9B) (27). Among the deletion mutants of cortactin we tested, the C-terminal and ΔRepeat mutants, both of which lack the repeat domain, did not bind to Keap1 (fig. S9C). On the other hand, the ability of Keap1 to bind cortactin was almost lost in NTR (N-terminal region) + BTB (broad complex, tramtrack, and bric-a-brac) and IVR (intervening region), and strongly reduced in ΔDC mutants, all of which lack the DC domain and modestly reduced in the ΔBTB and ΔIVR mutants (fig. S9D). These observations suggested that Keap1 and cortactin interacted with each other mainly through the DC domain and the repeat domain, respectively, although NTR, BTB, and IVR domains in Keap1 may partially contribute to the binding to cortactin.

Fig. 3 Keap1 interacts with cortactin and tethers cortactin in the cytoplasm.

(A and B) Interaction between cortactin and Keap1. Proteins immunoprecipitated with anti-Flag antibody from 293T cells expressing Flag-tagged cortactin (A) or Keap1 (B) were analyzed by immunoblotting. Whole-cell lysates were also analyzed by immunoblotting to confirm comparable abundance of each protein. n = 2 and 4 independent experiments, respectively. (C) Nuclear accumulation of cortactin in Keap1 knockout (KO) mouse embryonic fibroblasts (MEFs). The subcellular localization of cortactin in WT and Keap1 KO MEFs, and Keap1 KO MEFs, which had been stably infected with an empty vector (KO-vec) or a vector for HA-tagged Keap1 (KO-Keap1), was determined by immunofluorescence. The fluorescence intensities of nuclear and cytoplasmic cortactin were quantified. At least 1000 cells per experiment were analyzed. Data are means ± SD of five independent experiments for WT and KO MEFs and four independent experiments for KO-vec and KO-Keap1 cells. *P < 0.05. Pictures show representative examples. Scale bars, 25 μm. (D) Cortactin abundance in WT, KO, KO-vec, and KO-Keap1 MEFs. Protein abundance of Keap1 and cortactin were analyzed by immunoblotting. An anti–α-tubulin antibody was used as an internal control. n = 2 independent experiments. (E) Increases in TSA- and NA-induced acetylation of cortactin in KO MEFs. WT or KO MEFs treated with or without TSA and NA were immunoblotted with the indicated antibodies. n = 2 independent experiments.

The DC domain of Keap1 and the cortactin repeat domain are both involved in F-actin binding (28, 29). To determine whether F-actin mediates the interaction between cortactin and Keap1 in cells, we performed pull-downs after disassembling F-actin with latrunculin A, an actin-depolymerizing agent. Even in the absence of F-actin, however, Keap1-cortactin binding was maintained (fig. S9, E and F). Together, these results suggested that cortactin and Keap1 directly interacted through their repeat and DC domains, respectively.

Keap1 regulates the nucleocytoplasmic transport of cortactin

To determine whether Keap1 regulates the nucleocytoplasmic transport of cortactin, we used Keap1 KO MEFs to investigate the role of Keap1 in cortactin subcellular localization. Immunostaining revealed that the nuclear accumulation of cortactin was slightly but significantly increased in Keap1 KO cells (Fig. 3C). Although cortactin abundance was unchanged in the KO cells (Fig. 3, D and E), the extent of acetylation induced by HDAC inhibitors was greater than in wild-type cells (Fig. 3E), similar to the effect of LMB (fig. S8B) or the NES mutation (Fig. 2F), further supporting the idea that the nuclear transport of cortactin was promoted in the absence of Keap1. Reintroduction of the Keap1 gene into the KO cells restored the cytoplasmic localization of cortactin (Fig. 3C). Together, these results suggested that Keap1 functioned to sequester cortactin in the cytoplasm. Notably, both the C-terminal and ΔRepeat mutants of cortactin appeared to show mixed localization in the cytoplasm and the nucleus, although both mutants have the NES2 (Fig. 2, B and C). Because these mutants could not bind to Keap1 (fig. S9C), the defect in tethering of cortactin in the cytoplasm by Keap1 might be responsible for their weak cytoplasmic localization.

Keap1 regulates cell migration by controlling the cortical localization of cortactin

Because cortactin plays a pivotal role in cancer cell migration and invasion (30), we asked whether Keap1 affected cell migration by regulating cortactin activity. To this end, we measured the cell migration of human lung cancer A549 cells with Keap1 knockdown (Fig. 4A) with a transmembrane assay (31). Knockdown of both cortactin and Keap1 significantly reduced cell migration (Fig. 4B) without affecting cell growth (fig. S10). In addition, wound-healing assays revealed that cell migration in Keap1 KO MEFs was decreased compared with that in parental MEFs (Fig. 4, C and D). Reintroduction of Keap1 into the Keap1 KO cells (Fig. 3D) restored cell migration (Fig. 4, E and F) as well as the cytoplasmic localization of cortactin (Fig. 3C). These observations suggested that Keap1 promoted cell motility by suppressing the nuclear import of cortactin.

Fig. 4 Keap1 promotes cell migration.

(A) Knockdown efficiency of Keap1 or cortactin siRNA in A549 cells. (B) Decrease in cell migration in Keap1-knockdown cells. Cell migration activity of A549 cells transfected with control, Keap1, or cortactin siRNAs was measured by transmembrane assay. Data are means ± SD of four independent experiments performed in duplicate, normalized to control values. *P < 0.05. Pictures show representative examples. Scale bar, 500 μm. (C and D) Decrease in cell migration in Keap1 KO MEFs. The cell migration activity of WT and Keap1 KO MEFs after scratch wounding was monitored. The migration area was calculated by subtracting the wound area at 4, 8, 12, or 24 hours from the wound area at 0 hour. Data are means ± SD of five independent experiments. **P < 0.01. Pictures show representative examples. (E and F) Effect of reintroduction of Keap1 into Keap1 KO cells on cell migration. Cell migration activity of Keap1 KO MEFs that were stably transfected with an empty retrovirus expression vector (KO-vec) or a retrovirus encoding HA-tagged Keap1 (KO-Keap1) was analyzed as described in (C) and (D). Data are means ± SD of nine independent experiments. **P < 0.01. Pictures show representative examples.

Upon stimulation by growth factors, cortactin is translocated from the cytosol to the cell cortex, where it promotes actin polymerization and cell migration (12, 32). Indeed, when we stimulated cells with phorbol 12-myristate 13-acetate (PMA), which activates the protein kinase C pathway, cortactin translocated to the cell periphery (Fig. 5A). Because Keap1 bound to cortactin and regulated its subcellular localization (Fig. 3), we tested whether Keap1 also regulated the cortical translocation of cortactin induced by PMA. In response to PMA, exogenously expressed Keap1 was translocated to the cell periphery and colocalized with cortactin in the cell cortex (Fig. 5A). Furthermore, PMA treatment enhanced the interaction between endogenous cortactin and Keap1 (Fig. 5B). PMA-mediated cortical localization of cortactin was significantly reduced in Keap1-knockdown cells (Fig. 5, C and D, and fig. S11). Circular dorsal ruffles are membrane protrusions composed of actin-rich structures formed on the apical surfaces of cells and involved in internalization and cell motility, with which cortactin colocalizes (33). The localization of cortactin to structures resembling circular dorsal ruffles was also reduced in Keap1-knockdown cells (fig. S12). Reintroduction of wild-type Keap1 containing silent mutations that conferred resistance to the Keap1 siRNA restored efficient translocation of cortactin to the leading edges (Fig. 5, E and F). Moreover, PMA-induced dynamic motion of green fluorescent protein (GFP)–fused cortactin to the cell periphery in living cells was decreased by the Keap1 knockdown (Fig. 5G and movies S1 and S2). These results suggest that upon stimulation of cell growth, Keap1 recruits cortactin to the cortical region, where dynamic assembly of F-actin occurs in response to growth signaling.

Fig. 5 Keap1 is indispensable for the cortical localization of cortactin upon PMA stimulation in A549 cells.

(A) Colocalization of Keap1 and cortactin at the cortical regions in response to PMA. Serum-starved A549 cells expressing HA-tagged Keap1 were treated with PMA and stained with the indicated antibodies and DAPI. n = 3 independent experiments. Scale bars, 25 μm. (B) Enhanced interaction between endogenous cortactin and Keap1 upon PMA treatment. Serum-starved A549 cells were treated or not with PMA. Immunoprecipitated proteins were analyzed by immunoblotting. n = 2 independent experiments. WCL, whole-cell lysate. (C and D) Effect of Keap1 knockdown on PMA-induced cortical localization of cortactin. A549 cells transfected with control (siCont) or Keap1 siRNA (siKeap1 #1 or #2) were serum-starved and immunoblotted to assess the efficiency of Keap1 knockdown and abundance of Nrf2 and γ-glutamylcysteine synthetase catalytic subunit (γ-GCSc). n = 3 independent experiments. Serum-starved A549 cells transfected with control (siCont) or Keap1 siRNA (siKeap1 #1) were treated with PMA, and cortactin localization was determined by staining with anti-cttn antibody and DAPI. Two representative pictures taken at different magnifications were shown. Cortical localization of cortactin, indicated by arrows, was quantified. At least 800 cells per experiment were analyzed. Data are means ± SD of three independent experiments. *P < 0.05. Pictures show representative examples. Scale bars, 25 μm. (E and F) Effect of reintroduction of WT Keap1 and the G430C mutant into Keap1-knockdown cells on the cortical localization of cortactin. A549 cells were stably transfected with an empty retrovirus expression vector or a retrovirus encoding HA-tagged Keap1 WT or G430C mutant with silent mutations resistant to the siRNA (Keap1), then transfected with control (siCont) or Keap1 siRNA (siKeap1). Cell lysates were analyzed by immunoblotting. Cortical localization of cortactin was analyzed as described in (D) and (F). Cortical localization of cortactin, indicated by arrows, was quantified. At least 400 cells per experiment were analyzed. Data are means ± SD of three independent experiments. **P < 0.01. Pictures show representative examples. Scale bars, 25 μm. (G) Live-cell imaging of PMA-induced cortical localization of cortactin. A549 cells were transfected with control (siCont) or Keap1 siRNA (siKeap1) and GFP-fused cortactin. GFP fluorescence was imaged every minute after stimulation with PMA in serum-starved cells. n = 5 independent experiments. Scale bars, 20 μm.

Nrf2 accumulates in the nucleus in the absence of functional Keap1. Deficiency of Nrf2 in cancer cells or vascular smooth muscle cells enhances cell migration (34, 35), suggesting that Nrf2 inhibits cell migration. However, it is unlikely that Keap1 promotes cell migration by suppressing Nrf2 activity because Nrf2 is constitutively activated in A549 cells (36, 37). Somatic mutations of the DC domain of Keap1 have been frequently found in non–small cell lung cancer cells including A549 cells (36, 37), and these mutations impair Nrf2-binding activity (36, 38, 39). Indeed, in A549 cells, Keap1 knockdown did not significantly increase the protein abundance of Nrf2 or γ-GCS, an Nrf2 target gene product (Fig. 5C). These results are consistent with previous observations that the somatic mutations in the DC domain of Keap1 in A549 cells abolish the ability of Keap1 to repress Nrf2 (37, 40). To further confirm that Keap1 promotes cell migration independently of Nrf2, we introduced a Keap1 mutant (G430C), which cannot repress (fig. S14A) (36, 38, 41), into the Keap1-knockdown A549 cells and found that not only wild-type but also mutant Keap1 restored cortactin translocation to the cell periphery in Keap1-knockdown cells (Fig. 5F). Indeed, the G430C mutant still retained the ability to bind to cortactin (fig. S14B). These observations indicate that Keap1-mediated recruitment of cortactin to the cell cortex is a main contributor to PMA-induced cell migration.

Acetylation of cortactin regulates Keap1-dependent cell migration

Acetylation of cortactin reduces cell motility (15). Furthermore, most, if not all, acetylatable lysine residues of cortactin were located in its repeat domain (fig. S1A), which was essential for its binding to Keap1 (fig. S9C), an interaction that played a vital role in cortactin-meditated cell migration (Figs. 4 and 5). These observations prompted us to investigate whether acetylation regulated cortactin-mediated cell migration by affecting the interaction with Keap1. To test this hypothesis, we examined the effects of the 7KR mutation, which mimics hypoacetylation, and the 7KQ mutation, which mimics hyperacetylation, on the binding of cortactin to Keap1 in cells. The ability of the 7KQ mutant to bind Keap1 was significantly lower than that of wild-type cortactin, whereas the 7KR mutant exhibited slightly enhanced Keap1 binding (Fig. 6, A and B). Consistent with this, in vitro binding assay using purified recombinant proteins showed that the ability of cortactin to directly bind to GST-fused full-length Keap1 was reduced by the 7KQ mutation (Fig. 6C). Furthermore, the binding between cortactin and Keap1 was significantly impaired in the presence of TSA and NA, which cause an increase in the acetylation of cortactin. TSA and NA failed to reduce the binding when the 7KR mutant was used instead of wild-type cortactin (Fig. 6, D and E). Consistent with this finding, cortical localization of cortactin was significantly reduced in cells treated with TSA and NA or expressing the cortactin 7KQ mutant but not in cells expressing the cortactin 7KR mutant (fig. S13). TSA and NA reduced the motion of wild-type cortactin to the cell periphery but not that of the 7KR mutant (Fig. 6F and movies S3 to S5). By contrast, the 7KQ mutant did not move to the cell periphery (Fig. 6F and movie S6). Together, these results suggested that acetylation of the repeat domain of cortactin diminishes its ability to bind Keap1, resulting in attenuation of cortactin-mediated cell migration.

Fig. 6 Acetylation of cortactin reduces its binding to Keap1.

(A to C) Effect of mutations in the acetylation sites on cortactin binding to Keap1. Flag immunoprecipitates from 293T cells expressing Flag-tagged WT, 7KR, or 7KQ cortactin were analyzed by immunoblotting (A), and band intensity was quantitated (B). Keap1 bound to cortactin was quantified on the basis of the ratio of the signal for immunoprecipitated Keap1 to that for IP Flag. Data are means ± SD of four independent experiments. **P < 0.01. (C) Pictures show representative examples. In vitro GST pull-down assay was performed as described in Fig. 3C. The nonspecific bands were indicated by “*.” n = 3 independent experiments. (D and E) Effect of HDAC inhibitors on the interaction between cortactin and Keap1. Flag immunoprecipitates from 293T cells transfected as in (A) and treated with TSA and NA were analyzed by immunoblotting. The relative amount of Keap1 bound to cortactin was quantified as described in (B) and (E). Data are means ± SD of eight independent experiments. *P < 0.05. (F) Effect of acetylation on the dynamic motion of cortactin in living cells. GFP fluorescence was imaged in serum-starved A549 cells expressing GFP-tagged cortactin WT, 7KQ, or 7KR every minute after pretreatment with TSA and NA and stimulation with PMA. n = 4 independent experiments. Scale bars, 20 μm. DMSO, dimethyl sulfoxide.

DISCUSSION

Keap1 anchors Nrf2 to the actin cytoskeleton, where it targets Nrf2 for ubiquitin-mediated degradation (28). Keap1 is similar to Kelch in Drosophila, which is essential for the formation of actin-rich intracellular bridges (42). The DC domain in Keap1, which comprises six repeats of the Kelch motif and the C-terminal region, is important for both F-actin and Nrf2 binding. Here, we identified cortactin as another Keap1 binding partner and uncovered a role of Keap1 in cortactin-mediated cell motility.

Keap1 has been reported to localize in focal adhesions, which are dynamic protein complexes through which the cytoskeleton of a cell connects to the extracellular matrix (43). We found that Keap1 bound to cortactin through its DC domain to regulate cortactin subcellular localization. Our data demonstrate two distinct functions of Keap1: inhibition of nuclear import of cortactin and recruitment to the cell cortex upon growth stimulation. Furthermore, we showed that cortactin shuttled between the cytoplasm and the nucleus, and that acetylation by the acetyltransferase CBP occurred in the nucleus. Cortactin localization in the cytosol was maintained through two mechanisms: active nuclear export of cortactin, which was mediated by its own NES, and suppression of cortactin nuclear import by Keap1. Because Keap1 played a critical role in cell migration by promoting the cortical translocation of cortactin upon PMA stimulation (Figs. 4 and 5), this second role of Keap1 in cortactin regulation may contribute to cancer cell metastasis. At present, it is unclear whether the interactions of Keap1 with F-actin and cortactin through the DC domain are mutually exclusive. However, because Keap1 forms a dimer through its BTB domain, it seems possible that Keap1 could bind both cortactin and F-actin simultaneously (44), and that Keap1 could bind cortactin in the cytosol and transfer it to F-actin at the cell cortex. Because Keap1 is involved in proteasomal degradation of Nrf2, it is also possible that Keap1 acts as a ubiquitin E3 ligase for cortactin as it does for Nrf2. However, this seems unlikely because the protein abundance of cortactin in Keap1 KO MEFs was comparable to that of wild-type MEFs (Fig. 3, D and E).

Cortactin is acetylated in the cortactin repeat domain and deacetylated upon stimulation by growth factors (15). Acetylation of the conserved lysine residues has been proposed to prevent efficient binding to F-actin (15). Structural analysis suggests that the repeat domain is dynamic and unstructured in solution, and that the protein may undergo a transition from a disordered state (like the “molten globule”) to a more stably folded conformation upon F-actin binding (45). Thus, acetylation of the disordered repeat domain may prevent efficient folding upon F-actin binding. The repeat domain was responsible for Keap1 binding (fig. S9C), and acetylation in the repeat domain of cortactin compromised its ability to bind Keap1 and to translocate to the cortical region (Fig. 6, fig. S13, and movies S3 to S6). Thus, acetylation of cortactin may reduce the affinity to both Keap1 in the cytoplasm and F-actin in the cell cortex.

In conclusion, the results of this study revealed a role for Keap1 in cell migration. Cortactin has been associated with tumor cell invasion, and its abundance is increased in aggressive, metastatic tumors. Because acetylation suppresses cortactin-mediated cancer cell motility, the molecules involved in its deacetylation, such as HDAC6 and SIRT2, represent potential targets for drug discovery. In addition, other means to increase acetylation, such as inhibition of cortactin nuclear export or blockage of Keap1-cortactin interaction, may also have therapeutic potential. Our findings also raise questions that will need to be addressed in future studies regarding how Keap1 promotes cortactin function in the cell cortex and whether cortactin regulates transcription or other nuclear function in the nucleus.

MATERIALS AND METHODS

Antibodies, materials, and cell lines

Monoclonal mouse anti-Flag (clone M2) and anti–α-tubulin (clone B-5-1-2) antibodies were purchased from Sigma-Aldrich. Monoclonal anti-cortactin antibody (clone 4F11), anti-HDAC2 antibody, anti-CBP antibody (clone AC238), anti-PCAF antibody, and anti-p300 antibody (clone RW128) were obtained from Millipore, Sigma, Chemicon, Cell Signaling, and Upstate, respectively. An anti-SIRT2 antibody was purchased from Santa Cruz Biotechnology or Abgent. A polyclonal rabbit antibody against acetylated lysine (anti-AcLys) and an anti-HA antibody were purchased from Cell Signaling Technology and Santa Cruz Biotechnology, respectively. Polyclonal goat anti-GST antibody was obtained from GE Healthcare. Polyclonal goat and rabbit anti-Keap1 antibodies were purchased from Santa Cruz Biotechnology and Proteintech, respectively. TSA (22) and LMB (46) were prepared as described previously. NA was purchased from Sigma-Aldrich. PMA was obtained from LC Laboratories. Calcein AM Fluorescent Dye was purchased from BD Biosciences. COS-7, HeLa, 293T, and A549 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Wako) containing 10% heat-inactivated fetal bovine serum (FBS) (Invitrogen). Wild-type and Keap1 KO MEFs (47) were cultured in DMEM, low glucose (Wako) containing 10% FBS. The polyclonal antibody against cortactin acetylated on Lys309 (anti–Ac-cttn) was generated using a synthetic peptide with the 14–amino acid sequence corresponding to residues 303-SKGFGGK(Ac)YGVQKDC-316 of human cortactin (acetylated at Lys309) as an immunogen. The antibody titer of the obtained antiserum was checked by enzyme-linked immunosorbent assay, using free peptide containing acetylated lysine. The purified antibody was obtained by affinity purification using a column containing acetylated peptide.

DNA constructs, transfection, and drug treatment

Full-length human cortactin complementary DNA (cDNA) containing six and a half repeats was obtained by inserting the sequence corresponding to the sixth repeat (Lys266 to Tyr302) into human cortactin cDNA containing five and a half repeats (clone ID 4824206, Open Biosystems). The resulting full-length cDNA sequence was identical to that of human cortactin (GenBank accession number NM_005231). For construction of the plasmid for N-terminally 6×His- and 2×Flag-tagged human cortactin (pcDNA3.1-HFF-cortactin), recombination-based cloning (Gateway, Invitrogen) was used. The full-length cortactin cDNA was integrated enzymatically into the pcDNA3.1-HFF-ccdB vector, which was constructed by inserting the ccdB cassette, the 6×His cassette, and the 2×Flag cassette into pcDNA3.1 (48, 49). Point and deletion mutants of these plasmids were generated by polymerase chain reaction (PCR) using the QuikChange Site-Directed Mutagenesis Kit (Stratagene). Full-length cortactin was amplified by PCR from pcDNA3.1-HFF-cortactin and subcloned into pCold TF DNA vector (Takara) for the expression of recombinant proteins in Escherichia coli. Flag- and HA-tagged HATs were described previously (49). Human Keap1 cDNA was amplified by PCR and then subcloned into the Eco RV–Xho I sites in the pcDNA3-Flag or pcDNA3-HA vector. Keap1 deletion mutants were constructed in a pIRES vector after PCR amplification with specific primers. Cells were transiently transfected with each plasmid using the Lipofectamine 2000 Reagent (Invitrogen), FuGENE HD, or X-tremeGENE 9 (Roche), according to the manufacturers’ protocol. For rescue of Keap1 KO cells, Keap1 was subcloned into the Not I–Bam HI site of vector pQCXIP (Clontech) from pcDNA3-HA-Keap1 after PCR amplification (Keap1 rescue construct: pQCXIP-HA-Keap1). Keap1 KO cells were infected with the retrovirus harboring pQCXIP empty vector or pQCXIP-HA-Keap1 along with helper plasmid pCL-10A1, and stable transformants were selected with puromycin (1 μg/ml) for 3 weeks (KO-vec and KO-Keap1). For stable overexpression, a silent mutation (1439-GGACAAACCGCCTTAATTC-1457) was introduced into pQCXIP-HA-Keap1 by PCR-based site-directed mutagenesis (pQCXIP-HA-Keap1-si1439). A549 cells were infected with retrovirus harboring this plasmid or pQCXIP empty vector along with helper plasmid pCL-10A1, and stable transformants were selected with puromycin (1 μg/ml) for 3 weeks. For drug treatment, cells were treated with 1 to 3 μM TSA and 5 mM NA for 12 or 16 hours and LMB (10 ng/ml) for 3 to 12 hours. For PMA treatment, cells were cultured in serum-free medium for 24 hours before treatment with 100 nM PMA for 20 min.

siRNA transfection

Nontargeting control and siRNA oligonucleotides were purchased from GE Dharmacon. Each siRNA was described as follows: negative control siRNA pool: siGENOME Non-Targeting siRNA pool (D-001210-02); CBP, p300, PCAF, SIRT1, and HDAC6 siRNA pools containing four oligos each: siGENOME SMARTpool Human CBP (M-003477-02), Human p300 (M-003486-04), Human PCAF (M-005055-00), Human SIRT1 (M-003450-01), and Human HDAC6 (M-003499-00). Cortactin and Keap1 siRNA were obtained from Nippon Gene, and AllStars Negative Control siRNA (Qiagen) was used as a control. The cortactin siRNA target sequence was 5′-GGACAAAGUGGAUAAGAGCTT-3′. Two different Keap1 siRNA target sequences were used: Keap1 siRNA #1, 5′-GGACAAACCGCCUUAAUUCTT-3′ and Keap1 siRNA #2, 5′-CGAAUGAUCACAGCAAUGATT-3′. Cells were transfected with siRNA oligos using the DharmaFECT 1 transfection reagent (GE Dharmacon) or the Lipofectamine RNAiMAX transfection reagent (Invitrogen).

Immunoprecipitation and immunoblotting

Cells were harvested in ice-cold immunoprecipitation buffer containing 50 mM tris-HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 0.1% NP-40, and protease inhibitor cocktail (Roche), and then sonicated twice for 10 s. The resultant lysates were centrifuged at 15,000 rpm for 10 min at 4°C. Supernatants were incubated with the indicated primary antibody for 1 hour at 4°C with gentle agitation and then incubated further with protein A/G agarose beads (Santa Cruz Biotechnology) for another 1 hour. The agarose beads were then washed three times with immunoprecipitation buffer, and the bound proteins were extracted with SDS–polyacrylamide gel electrophoresis (SDS-PAGE) loading buffer by heating at 95°C for 5 min. Proteins were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane (Millipore) by electroblotting. After the membranes were incubated with primary and secondary antibodies, the immune complexes were detected with an ECL Western blotting kit (Amersham) or an Immobilon Western Chemiluminescent HRP Substrate (Millipore), and the luminescence was analyzed with a LAS-3000 image analyzer (Fujifilm). For quantitation of Keap1, the pixel sizes of Keap1 and Flag-cortactin were adjusted using Photoshop (Adobe), before the band intensity of Keap1 was normalized with ImageJ software, to the intensity of the Flag-cortactin.

In vitro acetylation assay

In vitro acetylation assays were performed as described previously with some modifications (23). Recombinant CBP protein (1 μg) was incubated with Flag-tagged cortactin proteins purified from 293T cells transfected with pcDNA3.1-HFF-cortactin or histone H3 proteins in the presence of 0.5 mM acetyl-CoA in 50 μl of reaction buffer [50 mM tris-HCl (pH 8.0), 10% glycerol, 1 mM dithiothreitol, 100 μM EDTA, 1 mM phenylmethylsulfonyl fluoride] for 1 hour at 30°C. Acetylation was analyzed by SDS-PAGE followed by immunoblotting with anti–Ac-cttn or anti-AcLys.

Immunofluorescence

Exponentially growing cells were plated onto 18-mm coverslips (for analyses using a fluorescence microscope) or black clear-bottom 96-well plates (for analyses using the IN Cell Analyzer 2000, GE Healthcare) and incubated overnight. After transfection and drug treatment, cells were washed in phosphate-buffered saline (PBS) and fixed with 3.5% formaldehyde/PBS for 10 min at room temperature. Cells were then rinsed three times with PBS and permeabilized with 1% Triton X-100/PBS for 10 min at room temperature. Cells were blocked in 5% calf serum/PBS for 30 min and then incubated with primary antibody for 1 hour at room temperature or overnight at 4°C. After the incubation, cells were washed three times with PBS, followed by the addition of secondary Alexa Fluor–conjugated anti-mouse or anti-rabbit antibody (Molecular Probes). Cells were incubated with secondary antibodies for 1 hour and then washed three times with PBS. For observation using a fluorescence microscope (DeltaVision, Applied Precision), the coverslips were mounted with DAPI (Vector Laboratory). For quantitative analyses, cells were stained with Hoechst 33342 (10 μg/ml; Molecular Probes) after incubation with secondary antibodies. Plates were imaged using an IN Cell Analyzer 2000 cellular imaging system with at least 16 fields of view per well. Images were analyzed using the IN Cell Developer software (GE Healthcare). Total cell numbers were determined by counting Hoechst-stained nuclei. To quantify the cortactin acetylation per cell, fluorescence intensity of the cortactin acetylation in a given field was divided by the number of cells in the field, which was normalized against values in control cells. To quantify the ratio of nuclear to cytoplasmic cortactin, the fluorescence intensity of nuclear cortactin per cell was divided by that of cytoplasmic cortactin per cell. To quantify the cortical localization of cortactin per cell in a given field, the fluorescence intensity at the peripheral region multiplied by the area of the peripheral region was divided by the number of cells in the field, which was normalized against the values in control cells.

Preparation of recombinant proteins

His-tagged cortactin proteins were expressed in E. coli BL21 by treating bacteria containing expression plasmids with 0.2 mM isopropyl-β-d-thiogalactopyranoside at 18°C for 24 hours. The cells were harvested and then sonicated in a binding buffer [20 mM sodium phosphate, 300 mM NaCl (pH 7.5)] containing 1% Triton X-100, and the His-tagged proteins were purified using Ni-NTA Agarose (Qiagen). Eluted His-tagged cortactin proteins were digested with HRV 3C protease (Novagen) overnight at 40°C to remove the N-terminal His tag. Proteins were then further purified using a HiTrap Q HP anion exchange column (GE Healthcare) with an ÄKTAexplorer 10S system (GE Healthcare).

In vitro pull-down assay

Cortactin and GST-fused Keap1 proteins were incubated in NETN buffer [50 mM tris-HCl (pH 7.7), 150 mM NaCl, 5 mM EDTA, 0.1% NP-40] containing 20% BSA at 4°C for 2 hours. Glutathione-Sepharose beads (GE Healthcare) were then added, and the reaction mixture was incubated at 4°C overnight. After beads were washed three times with NETN buffer, bound proteins were extracted with the SDS-PAGE loading buffer by heating at 95°C for 5 min. Proteins were separated by SDS-PAGE followed by immunoblot analysis.

Cell migration assay

Cell migration was measured in the BD Falcon FluoroBlok 24-Multiwell Insert System (31). A549 cells were trypsinized, resuspended in serum-free DMEM, and plated on the upper side of the filters (5 × 104 cells per well). The lower chambers were filled with DMEM containing 10% FBS as an attractant. Migrated cells were stained with calcein-AM (5 μg/ml) for 1 hour. To quantitate cell migration, fluorescence was measured (excitation, 485 nm; emission, 530 nm) using a SpectraMAX M2e microplate spectrophotometer (Molecular Devices). Data are means ± SD of three independent experiments performed in duplicate, normalized against control values. Cells were also observed under a fluorescence microscope (Olympus IX71).

Wound-healing assay

MEFs were seeded in 12-well culture plates at 2.5 × 105 cells per well and incubated for 24 hours to form a confluent monolayer. The cells were wounded by scratching with a thin pipette tip and monitored at a series of time points under a phase-contrast microscope (Olympus IX71) with a 4× objective. The wound area and area into which cells migrated were measured in pixels using the ImageJ software.

Live-cell imaging

Glass-bottom 35-mm dishes were used for live-cell imaging. After transfection and serum starvation for 24 hours, the cells were examined at 37°C in 5% CO2 using an Olympus IX81 microscope equipped with a UIC-QE cooled charge-coupled device camera (Molecular Devices), a BP460-480HQ excitation filter, a DM485 dichroic mirror, and a BA495-540HQ emission filter. Images were collected every 1 min using the MetaMorph software (Universal Imaging).

Statistical analysis

To determine significance of differences between two groups, the Brown-Forsythe test was first performed to check the equality of variances between the two groups. Next, the Student’s t test was applied to all experiments except for one in Fig. 4D because the pair had unequal variances (P < 0.05). Welch’s t test was performed for this one. Notably, we took the logarithm of the ratio data used for the statistical calculations in Figs. 2 (A, C, and D), 3C, and 6 (B and E). We performed all statistical calculations with the R language adding the “car” package for the Brown-Forsythe test (specifying “center = median” in Levene test command).

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/8/404/ra120/DC1

Fig. S1. Acetylation occurs on all seven conserved lysine residues in the repeat domains of human cortactin.

Fig. S2. HDAC6 and SIRT2 are the major deacetylases for cortactin in cells.

Fig. S3. SIRT1 is localized in the nucleus in A549 and COS-7 cells.

Fig. S4. CBP is a major acetylase for cortactin in cells.

Fig. S5. CBP is localized in the nucleus in A549 and COS-7 cells.

Fig. S6. Specificity of the antibodies for acetylated cortactin.

Fig. S7. Specificity of an anti–Ac-cttn antibody for immunostaining.

Fig. S8. Consensus NES sequences in cortactin and increased acetylation of cortactin upon LMB treatment.

Fig. S9. Keap1 binds cortactin independently of actin.

Fig. S10. Effect of knockdown of Keap1 or cortactin on cell proliferation.

Fig. S11. Keap1 knockdown decreases PMA-induced localization of cortactin to a structure resembling circular dorsal ruffles.

Fig. S12. Keap1 knockdown decreases PMA-induced localization of cortactin to structures resembling circular dorsal ruffles.

Fig. S13. Acetylation reduces cortical localization of cortactin.

Fig. S14. Effects of the Keap1 G430C mutant on repressing the Nrf2 activity and binding to cortactin.

Movie S1. Dynamic movement of cortactin in A549 cells stimulated by PMA.

Movie S2. Effect of Keap1 knockdown on PMA-induced dynamic movement of cortactin.

Movie S3. PMA-induced dynamic movement of cortactin in A549 cells.

Movie S4. Effect of TSA and NA on PMA-induced dynamic movement of cortactin.

Movie S5. The 7KR mutation rescues PMA-induced dynamic movement of cortactin from inhibition by TSA and NA.

Movie S6. Effect of the 7KQ mutation on PMA-induced dynamic movement of cortactin.

REFERENCES AND NOTES

Acknowledgments: We are grateful to T. P. Yao for critically reading the manuscript and his advice and to M. Matsuura for statistical support. Funding: This study was supported in part by the CREST Research Project, the Japan Agency for Medical Research and Development, and by Grants-in-Aid for Scientific Research (S) and on Innovative Area “Cancer” from the Ministry of Education, Culture, Sports, Science and Technology of Japan. Author contributions: A.I., T. Shimazu, and M. Yoshida designed this study. A.I., T. Shimazu, S.M., and A.A.S. performed all experiments with the help of T. Suzuki, H.M., and M. Yamamoto, under the supervision of M. Yoshida. S.-i.I. and T.N. performed cortactin complex analysis. T.T. performed all statistical analyses. A.I., T. Shimazu, S.M., and M. Yoshida wrote this manuscript, and A.A.S., T.T., S.-i.I., T.N., T. Suzuki, H.M., and M. Yamamoto critically read the manuscript and gave advice. All authors discussed the results and commented on the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Navigate This Article