Oncogenic Activity of Ect2 Is Regulated through Protein Kinase Cι-mediated Phosphorylation

Abstract

The Rho GTPase guanine nucleotide exchange factor Ect2 is genetically and biochemically linked to the PKCι oncogene in non-small cell lung cancer (NSCLC). Ect2 is overexpressed and mislocalized to the cytoplasm of NSCLC cells where it binds the oncogenic PKCι-Par6 complex, leading to activation of the Rac1 small GTPase. Here, we identify a previously uncharacterized phosphorylation site on Ect2, threonine 328, that serves to regulate the oncogenic activity of Ect2 in NSCLC cells. PKCι directly phosphorylates Ect2 at Thr-328 in vitro, and RNAi-mediated knockdown of either PKCι or Par6 leads to a decrease in phospho-Thr-328 Ect2, indicating that PKCι regulates Thr-328 Ect2 phosphorylation in NSCLC cells. Both wild-type Ect2 and a phosphomimetic T328D Ect2 mutant bind the PKCι-Par6 complex, activate Rac1, and restore transformed growth and invasion when expressed in NSCLC cells made deficient in endogenous Ect2 by RNAi-mediated knockdown. In contrast, a phosphorylation-deficient T328A Ect2 mutant fails to bind the PKCι-Par6 complex, activate Rac1, or restore transformation. Our data support a model in which PKCι-mediated phosphorylation regulates Ect2 binding to the oncogenic PKCι-Par6 complex thereby activating Rac1 activity and driving transformed growth and invasion.

Introduction

Epithelial cell transforming sequence 2 (Ect2) is a guanine nucleotide exchange factor (GEF)³ for Rho family small GTPases (RhoA, Rac1, and Cdc42), and Ect2 function is essential in the regulation of cytokinesis (1–5). Ect2 is composed of a C-terminal GEF catalytic domain and an N-terminal regulatory domain that modulates the GEF activity and intracellular localization of Ect2 (Fig. 1A) (2). In nontransformed cells, Ect2 is localized almost exclusively to the interphase nucleus in a hypophosphorylated state, where it is thought to be inactive. During mitosis, Ect2 becomes hyperphosphorylated and is distributed throughout the cytoplasm, and during cytokinesis it is localized to the mitotic spindle at the cleavage furrow where it plays a critical role in furrow scission through regulation of RhoA (3).

FIGURE 1.

Identification of a unique phosphorylation site on Ect2. A, schematic diagram showing the domain structure of Ect2 and the unique Thr-328 phosphorylation site detected in this study. BRCT, BRCA-like C-terminal domain; DH, Dbl homology RhoGEF catalytic domain. Shown are the full-length Ect2 protein and the original ΔN-Ect2 transforming Ect2 clone (7). B, Ect2 is associated with the PKCι-Par6 complex in NSCLC cells. Immunoprecipitation of FLAG-tagged Par6 from cytoplasmic extracts of H1703 NSCLC cells co-precipitates PKCι and Ect2. The identity of the bands on the silver-stained gel was determined by mass spectrometry.

As suggested by its name, Ect2 has also been implicated in cellular transformation (6–13). Ect2 was originally identified by expression library screening for transforming genes expressed in human epithelial tumor cells (7). Subsequent analysis demonstrated that the original transforming Ect2 clone was generated as an artifact of the DNA isolation and transfection procedure and corresponded to an N-terminally truncated protein that had lost the autoinhibitory BRCT domains as well as the tandem nuclear localization signals (NLS) but retained the C-terminal GEF catalytic domain (ΔN-Ect2) (Fig. 1A). Interestingly, ectopic expression of ΔN-Ect2 but not full-length Ect2 caused transformation in mouse fibroblasts (7, 8). However, introduction of missense mutations that disrupted NLS function causes both mislocalization to the cytoplasm and activation of full-length Ect2 transforming activity, suggesting that mislocalization to the cytoplasm can promote Ect2 oncogenic potential (8). Expression analysis demonstrates that full-length Ect2 is overexpressed in several human tumor types, suggesting a role for Ect2 in these tumors (9–11, 14). However, these studies also suggest that unique mechanisms operate to regulate oncogenic Ect2 function in human cancer cells (15). We recently demonstrated that Ect2 is a potential oncogene in non-small cell lung cancer (NSCLC) (6). Specifically, we found that Ect2 mRNA and protein is highly overexpressed in human NSCLC cell lines and primary NSCLC tumors and that the Ect2 gene ECT2 is a target for frequent tumor-specific gene amplifications as part of the chromosome 3q26 amplicon (6). In NSCLC tumors and cell lines, Ect2 protein is mislocalized to the cytoplasm where it binds to the oncogenic PKCι-Par6 complex (6). Functional analysis demonstrated that Ect2 regulates Rac1 activity downstream of the PKCι-Par6 complex. Interestingly, our analysis also demonstrated that the PKCι-Par6 complex regulates the cytoplasmic mislocalization of Ect2, that the transforming activity of Ect2 requires its interaction with PKCι-Par6, and that the role of Ect2 in transformation is distinct from its role in cytokinesis (6).

In this study, we investigate the mechanism by which the PKCι-Par6 complex regulates Ect2 function in NSCLC cells. We find that Ect2 isolated from NSCLC cells is highly phosphorylated at the novel and previously uncharacterized site Thr-328 within its hinge-like domain. We further show that PKCι directly phosphorylates Thr-328 in vitro and that PKCι and Par6 regulate Thr-328 phosphorylation in intact NSCLC cells. Functionally, we demonstrate that Thr-328 phosphorylation is important for the oncogenic activity of Ect2. Our data indicate that Thr-328 phosphorylation is important for the ability of Ect2 to support transformation by facilitating Ect2 binding to the PKCι-Par6 complex. Taken together, our data suggest a model for Ect2 regulation in which PKCι directly phosphorylates Ect2, an event that favors its interaction with the PKCι-Par6 complex and facilitates its GEF activity toward Rac1, a critical downstream effector of PKCι-Par6-Ect2-dependent transformation.

EXPERIMENTAL PROCEDURES

Antibody Reagents and Cell Lines

The following antibodies were used in these studies: Ect2 and RhoA (Santa Cruz Biotechnology, Santa Cruz, CA); PKCι, Cdc42, and Rac1 (BD Transduction Laboratories, San Jose, CA); β-actin, MEK1, and lamin A/C (Cell Signaling, Danvers, MA); FLAG epitope (Sigma). The S-peptide monoclonal antibody was a kind gift from Dr. S. Kaufmann, Mayo Clinic. The A427, A549, H1703, and MDCK cell lines were obtained from the American Type Culture Collection (Manassas, VA) and maintained in low passage culture as recommended by the supplier.

Mass Spectrometry Analysis of Ect2 Phosphorylation

Ect2 was immunoprecipitated from cytosolic extracts of H1703 cells as described previously (6). Immunoprecipitated Ect2 was resolved by SDS-PAGE, and the band corresponding to Ect2 was excised and submitted to the Mayo Clinic Cancer Center Protein Chemistry and Proteomics Shared Resource for proteolytic cleavage and phosphorylation site analysis by mass spectrometry. The SDS-polyacrylamide gel bands were prepared for mass spectrometry analysis using the following procedures. Silver-stained gel bands were destained with 15 mm potassium ferricyanide and 50 mm sodium thioisulfate in water until clear and then rinsed with water several times to remove all color (16). The bands were then reduced with 50 mm tris(2-carboxyethyl)phosphine, 50 mm Tris, pH 8.1, at 55 °C for 40 min and alkylated with 40 mm iodoacetamide at room temperature for 40 min in the dark. Proteins were digested in situ with either 30 μl (0.005 μg/μl) of trypsin (Promega Corp., Madison, WI), chymotrypsin, or Lys-C (Roche Diagnostics) in 20 mm Tris, pH 8.1, 0.0002% Zwittergent 3-16, at 37 °C for 4 h to overnight, followed by peptide extraction with 20 μl of 2% trifluoroacetic acid and then 60 μl of acetonitrile. The pooled extracts were concentrated to less than 5 μl on a SpeedVac spinning concentrator (Savant Instruments, Holbrook, NY) and then brought to 0.15% formic acid, 0.05% trifluoroacetic acid for protein identification by nano-flow liquid chromatography electrospray tandem mass spectrometry (nanoLC-ESI-MS/MS) using a ThermoFinnigan LTQ Orbitrap Hybrid Mass Spectrometer (ThermoElectron Bremen, Germany) coupled to an Eksigent nanoLC-two-dimensional HPLC system (Eksigent, Dublin, CA). The peptide digest was back-loaded onto a 250-nl OPTI-PAK trap (Optimize Technologies, Oregon City, OR) custom-packed with Michrom Magic C8 solid phase (Michrom Bioresources, Auburn, CA). Chromatography was performed using 0.2% formic acid in both the A solvent (98% water, 2% acetonitrile) and B solvent (80% acetonitrile, 10% isopropyl alcohol, 10% water), and running a 2% B to 50% B gradient over 60 min at 325 nl/min through a Michrom Magic C18 (75 μm × 200 mm) packed tip capillary column. The LTQ Orbitrap mass spectrometer experiment was set to perform an Fourier transform full scan from 375 to 1600 m/z with resolution set at 60,000 (at 400 m/z), followed by linear ion trap MS/MS scans on the top five [M + 2H]²⁺ or [M + 3H]³⁺ ions. Dynamic exclusion was set to one repeat of the same ion, which is then placed on an exclusion list for 15 s. The lock-mass option was enabled for the FT full scans using the ambient air polydimethylcyclosiloxane ion of m/z = 445.120024 or a common phthalate ion m/z = 391.284286 for real time internal calibration (17). This gave <2 ppm mass tolerances for precursor masses. The MS/MS raw data were converted to DTA files using extract_msn.exe from Bioworks 3.2 and correlated to theoretical fragmentation patterns of tryptic peptide sequences from the SwissProt data base using both SEQUEST^TM (ThermoElectron, San Jose, CA) and Mascot^TM (Matrix Sciences London, UK) search algorithms. All searches were conducted with fixed modification of carbamidomethyl-cysteine and variable modifications allowing for oxidation of methionines, formyl-Lys, phosphorylated Ser, Thr, and Tyr, along with loss or H₃PO₄ and protein N-terminal acetylation. The search was restricted to full trypsin, chymotrypsin, or Lys-C generated peptides as appropriate, allowing for two missed cleavages, and was left open to all species. Peptide mass search tolerances were set to 10 ppm, and fragment mass tolerances were set to ± 0.8 daltons. All protein identifications were considered when individual peptide scores were above the 95% percentile for probability and rank number one of all the hits for the respective MS/MS spectra. The identified phosphorylation site on phosphorylated peptides was manually validated.

Lentiviral RNAi Constructs, Cell Transduction, and Immunoblot Analysis

Lentiviral RNAi against human Ect2, PKCι, and Par6α was obtained from the Sigma Mission short hairpin RNA library, packaged into recombinant lentiviruses, and characterized for target gene knockdown (KD) as described previously (18). A nontarget lentiviral RNAi (NT-RNAi) that does not recognize any human genes was used as a negative control. RNAi target sequences and characterization of the specificity of RNAi reagents were published previously (6). Stably transduced NSCLC cell populations were generated as described previously (18). Ect2, PKCι, and Par6α RNAi constructs were analyzed for efficiency of target gene KD by quantitative PCR as described previously (6). Cell lysates from NT, Ect2, and PKCι KD cells were prepared in RIPA buffer (50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1 mm EDTA, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS) supplemented with protease inhibitor mixture (Roche Applied Science) and phosphatase inhibitor mixture I and II (Sigma) and subjected to immunoblot analysis as described previously (19).

Plasmids, Transfections, and Immunoprecipitations

The cDNA of full-length human RNAi-resistant Ect2 described previously (6) was cloned into the pTrAP plasmid vector that contains an S-peptide tag, His tag, and streptavidin-binding tag (a kind gift from Dr. S. Kaufmann, Mayo Clinic). Ect2 cDNA sequences were mutagenized to encode alanine substitutions at Thr-328, Thr-341, and Thr-412 to remove these phosphorylation sites; in addition, Thr-328 was mutagenized to encode an aspartic acid substitution to mimic phosphorylation at this site. Ect2 RNAi cells were transfected with the empty pTrAP vector, or ones encoding wild-type Ect2, or Ect2 phosphorylation mutants using Lipofectamine 2000 (Invitrogen). For co-immunoprecipitation experiments, Ect2 RNAi cells were co-transfected with FLAG-tagged human Par6α plus either S-peptide-tagged Ect2 or empty vector. Cell lysates were subjected to immunoprecipitation and immunoblot analysis as described previously (6, 18).

Soft Agar Growth and Cellular Invasion Assay

Anchorage-independent growth and invasion were assayed as described previously (6, 18, 19). Matrigel-coated plates were obtained from BD Biosciences. Experiments were independently conducted in triplicate at least three times.

Rho GTPase Activity Assays

The level of activated Rac1-GTP, Cdc42-GTP, and RhoA-GTP was determined by affinity pulldown as described previously (6, 18, 20). Densitometry was performed using Kodak Molecular imaging software (Carestream Molecular Imaging, New Haven, CT). Results were normalized to total Rac1, Cdc42 or RhoA.

Phospho-specific Thr-328 Ect2 Antibody Production and Validation

Rabbit polyclonal antibodies to phosphorylated peptides corresponding to phosphorylated Thr-328 Ect2 were generated by 21st Century Laboratories (Montgomery, TX). Immunogenic phosphopeptides were as follows: Ac-³²¹CYLYEKANpTPELKKSV³³⁵-amide and Ac-³²¹YLYEKANpTPELKKSVC³³⁵-amide. Rabbits were injected subcutaneously with the phosphopeptides conjugated to keyhole limpet hematoxylin, and sera were collected and analyzed for antibody titer by ELISA. The sera were immunodepleted of antibodies recognizing the unphosphorylated forms of the above peptides, and phospho-specific antibodies were affinity-purified with the corresponding antigen. To validate antibody specificity, the purified peptides Ac-³²¹YLYEKANpTPELKKSVC³³⁵-amide and Ac-³²¹CYLYEKANpTPELKKSV³³⁵-amide were conjugated to bovine serum albumin (BSA), and dilutions of the purified BSA conjugate were spotted onto nitrocellulose sheets and subjected to immunoblot analysis. In addition, H1703 NSCLC cells were transiently transfected with wild-type, T328A, or T328D S-peptide-tagged Ect2 expression plasmids using Lipofectamine 2000 in accordance with the manufacturer's instructions. Lysates were subjected to immunoblot analysis as described below. The antibody was further validated by in vitro λ phosphatase treatment. Briefly, wild-type S-peptide-tagged Ect2 was expressed and immunoprecipitated from H1703 NSCLC cells. Immunoprecipitated Ect2 was incubated in the presence or absence of 200 ng of λ phosphatase in 1× phosphatase buffer for 30 min at 30 °C. Reactions were terminated by the addition of SDS sample buffer and boiling for 5 min, and the samples were subjected to SDS-PAGE and immunoblot analysis using affinity-purified phospho-Thr-328 Ect2 antibody or total Ect2 antibody. In some experiments, total Ect2 was immunoprecipitated from H1703 NT, PKCι KD, and Par6α KD cells and subjected to SDS-PAGE and immunoblot analysis using affinity-purified phospho-Thr-328 Ect2 antibody or total Ect2 antibody as described above. Primary lung tumor and adjacent matched normal lung tissues were obtained from surgical resections of NSCLC cancer patients under advised consent. Fresh tissues were homogenized and lysed in RIPA buffer supplemented with protease inhibitor mixture (Roche Applied Science), phosphatase inhibitor mixture I and II (Sigma). Total Ect2 was immunoprecipitated from lysates and subjected to SDS-PAGE and immunoblot analysis using affinity-purified phospho-Thr-328 Ect2 antibody or total Ect2 antibody as described above.

Subcellular Fractionation

Cells were fractionated using NE-PER nuclear and cytoplasmic extraction reagents (Pierce). Nuclear and cytoplasmic fractions were subjected to immunoblot analysis using antibodies to Thr(P)-328 Ect2, total Ect2, lamin A/C, and MEK1 (all at 1:1000), and HRP-labeled secondary antibodies and ECL detection as described previously (6).

In Vitro Kinase Assays

PKCι in vitro kinase assays were performed as described previously (21). Briefly, purified recombinant bacterially expressed full-length human Ect2 was incubated in the presence or absence of 100 ng of recombinant human PKCι (Millipore, Danvers, MA) in reaction buffer (20 mm Tris-HCl, pH 7.5, 10 mm CaCl₂, 10 mm MgCl₂, and 40 μg/ml phosphatidic acid) supplemented with 200 μm ATP for 30 min at 25 °C. Reactions were stopped by addition of 2× Laemmli buffer, and phosphorylation was determined by immunoblot analysis for Thr(P)-328 Ect2 and total Ect2 as described above.

RESULTS

Identification of a Novel Phosphorylation Site on Ect2 in NSCLC Cells

We recently identified Ect2 as a potential oncoprotein in NSCLC (6). Ect2 protein is overexpressed in NSCLC cell lines and primary NSCLC tumors and is mis-localized to the cytoplasm in NSCLC cells where it binds to the oncogenic PKCι-Par6 complex (6). Through its interaction with the PKCι-Par6 complex, Ect2 regulates the activity of Rac1, which in turn is necessary for the transformed growth and invasion of NSCLC cells (6). Because Ect2 activity and intracellular localization have been shown to be regulated by site-specific phosphorylation events (1, 3, 5), we assessed the phosphorylation status of endogenous Ect2 co-immunoprecipitated with the PKCι-Par6 complex from the cytoplasm of H1703 NSCLC cells (Fig. 1B). Tryptic digestion and mass spectrometric (MS) analysis using high performance liquid chromatography (HPLC) interfaced to electrospray ionization on tandem mass spectrometers (LCQ-XP or LTQ-FT, Thermo Electron) detected a single, previously uncharacterized phosphorylation site within the Ect2 protein, threonine 328 (Fig. 1A). Subsequent MS analyses using a combination of trypsin, chymotrypsin, and Lys-C proteases detected phospho-Thr-328 as the only phosphorylation site on Ect2. This combination of proteases provided >90% coverage of the coding sequence of Ect2, including essentially all predicted potential Ser/Thr phosphorylation sites as analyzed by SiteScan (supplemental Fig. 1). MS analysis of tryptic digests of Ect2 isolated from A549 and A427 NSCLC cells also identified Thr-328 as the sole detected phosphorylation site, indicating that Thr-328 phosphorylation is not restricted to H1703 cells. These data indicate that Thr-328 is a major site of phosphorylation on Ect2 in human NSCLC cells, although we cannot exclude the possibility that other phosphorylation sites on Ect2 are also phosphorylated that were not detected in our analysis. Interestingly, Thr-328 is positioned immediately N-terminal to two consensus NLS previously implicated in the regulation of Ect2 GEF activity and intracellular localization (1, 8).

Phosphorylation of Ect2 at Thr-328 Is Functionally Important for Anchorage-independent Growth and Invasion of NSCLC Cells

We previously demonstrated that Ect2 is functionally important for anchorage-independent growth and invasion of NSCLC cells in vitro (6). Therefore, we assessed whether phosphorylation at Thr-328 is important for the oncogenic activity of Ect2 in NSCLC cells. Thr-328 on Ect2 was mutagenized to alanine to produce an Ect2 mutant that is no longer capable of being phosphorylated at this site (T328A) or to aspartic acid to mimic Thr-328 phosphorylation (T328D). cDNA sequences encoding these mutations were introduced into pTrAP-N1-hEct2, a eukaryotic expression vector encoding an S-peptide tag fused to the N terminus of human Ect2 encoded by a cDNA sequence to which silent mutations had been made to render it resistant to lentiviral RNAi-mediated Ect2 KD as described previously (6). H1703 NSCLC cells were stably transfected with empty control vector, wild-type (WT) Ect2, T328A-Ect2, or T328D-Ect2 (6). Ect2 transfectants were then transduced with recombinant lentivirus containing either a nontarget (NT) control or Ect2 RNAi construct, and stable populations of transduced cells were isolated by puromycin selection as described previously (6). Immunoblot analysis demonstrated the efficient KD of endogenous Ect2 expression, and the stable expression of exogenous WT and Ect2 phospho-mutants in Ect2 RNAi cells to levels approaching those of endogenous Ect2 in NT cells (Fig. 2A). As expected, Ect2 KD cells expressing empty vector exhibited impaired anchorage-independent growth and invasion when compared with NT control cells (Fig. 2, B and C), consistent with our published results (6). Interestingly, expression of either exogenous WT Ect2 or the T328D-Ect2 phospho-mutant significantly restored anchorage-independent growth and invasion to Ect2 KD cells, whereas expression of the T328A-Ect2 mutant did not (Fig. 2, B and C). These results indicate that phosphorylation of Ect2 at Thr-328 is functionally important for both the anchorage-independent growth and invasion of H1703 NSCLC cells.

FIGURE 2.

Effect of Thr-328-Ect2 phospho-mutants on anchorage-independent growth and invasion of NSCLC cells. A, immunoblot analysis of H1703 NSCLC cells stably transduced with either nontarget (NT) or Ect2 RNAi lentivirus and stably transfected with either empty vector (V) or one encoding WT-Ect2 (WT), T328A-Ect2 (T328A), or T328D-Ect2 (T328D) mutants. B, effect of WT and Thr-328-Ect2 mutants on anchorage-independent growth in soft agar. Expression of either WT or T328D-Ect2 restores soft agar colony growth to Ect2-RNAi cells, whereas expression of T328A-Ect2 does not. C, effect of WT and Thr-328-Ect2 mutants on cellular invasion through Matrigel. Expression of either WT or T328D-Ect2 restores cellular invasion to Ect2-RNAi cells, whereas expression of T328A-Ect2 does not. Data in B and C represent the mean ± the S.E., n = 4, and are presented as % NT control; * indicates significantly different from NT cells; ** indicates significantly different from Ect2 RNAi cells expressing control empty vector, p < 0.05. Data shown are representative of three independent experiments.

The inability of the T328A mutant to support transformation may result from this mutant possessing dominant-negative properties. Therefore, we assessed anchorage-independent growth and invasion in H1703 cells expressing vector control, WT Ect2, T328A-Ect2, or T328D-Ect2. Expression of WT Ect2, phospho-deficient, or phospho-mimic mutants of Thr-328 had no significant effect on the invasiveness or anchorage-independent growth of H1703 cells compared with cells expressing vector control (supplemental Fig. 2). Thus, the inability of the T328A-Ect2 mutant to rescue Ect2 KD was not due to a dominant-negative effect of this mutant.

Ect2 is phosphorylated in a cell cycle-dependent fashion. Two major mitotic Ect2 phosphorylation sites for cyclin-dependent kinase 1 (Cdk1) (Thr-341 and Thr-412) have been identified within the central hinge region that are important in regulating Ect2 localization and GEF activity during cytokinesis (1, 4, 5). Therefore, we determined whether phosphorylation at these sites was also important for the oncogenic activity of Ect2. For this purpose, we assessed the effect of expressing WT and Thr to Ala Ect2 mutants in Ect2 RNAi cells to restore transformation in NSCLC cells (Fig. 3A). Expression of either a T341A or a T412A mutant of Ect2 in H1703 Ect2 RNAi cells restored anchorage-independent growth and invasion to a level comparable with that restored by WT Ect2, whereas the T328A Ect2 mutant failed to do so (Fig. 3, B and C). These data indicate that phosphorylation of Ect2 at Thr-328 is required for the function of Ect2 in transformation, whereas phosphorylation of Thr-341 or Thr-412 appears to be dispensable for this aspect of Ect2 function. Interestingly, expression of the T328A, T341A, and T412A Ect2 mutants had no appreciable effect on the adherent growth rate of H1703 cells and failed to induce accumulation of multinucleated cells indicative of a cytokinesis defect (data not shown). These data are consistent with our recent observation that Ect2 KD in NSCLC cells does not induce a cytokinesis defect and further confirm that the role of Ect2 in transformation is distinct from its role in cytokinesis (6). These results are also consistent with our previous observation that NSCLC cells, including H1703 cells, undergo cytokinesis in an Ect2-independent fashion (6) and similar to what has been observed in HT1080 fibrosarcoma cells (22).

FIGURE 3.

Mitotic Ect2 phosphorylation sites are not necessary for the oncogenic activity of Ect2. A, immunoblot analysis demonstrating expression of S-tagged wild-type Ect2 (WT), T328A Ect2 (T328A), T341A Ect2 (T341A), and T412A Ect2 (T412A) in Ect2 RNAi NSCLC cells. B, effect of WT, T328A, T341A, and T412A Ect2 mutants on anchorage-independent growth in soft agar. Only the T328A Ect2 mutant fails to restore soft agar colony growth to Ect2-RNAi cells. C, effect of WT, T328A, T341A, and T412A Ect2 mutants on cellular invasion through Matrigel. Only the T328A Ect2 mutant fails to restore soft agar colony growth to Ect2-RNAi cells. Data in B and C represent the mean ± S.E., n = 4, and are presented as % NT control; * indicates significantly different from NT cells; ** indicates significantly different from Ect2 RNAi cells expressing control empty vector; p < 0.05. Data shown are representative of three independent experiments.

Phosphorylation of Thr-328 Regulates Rac1 Activity in NSCLC Cells

We recently demonstrated that Ect2 selectively regulates Rac1, but not RhoA or Cdc42, activity in NSCLC cells and that Rac1 is an important effector of Ect2-mediated transformation in NSCLC (6). Therefore, we assessed Rac1, RhoA, and Cdc42 activity in H1703 Ect2 KD cells stably transfected with empty vector or WT, T328A, and T328D Ect2 (Fig. 4A). As we observed previously (6), H1703 Ect2 KD cells expressing an empty vector control plasmid exhibited a statistically significant decrease in Rac1 activity compared with H1703 NT cells (Fig. 4B). Expression of either WT or T328D Ect2 in Ect2 KD cells significantly restores Rac1 activation, whereas the T328A mutant does not. In contrast, no significant changes in Cdc42 or RhoA activity were observed in these cells. These results are consistent with a role for Thr-328 Ect2 phosphorylation in Ect2-mediated activation of Rac1 in NSCLC cells.

FIGURE 4.

Effect of Thr-328 Ect2 phospho-mutants on Rac1, Cdc42, and RhoA activity in NSCLC cells. A, Rac1-GTP (Active Rac1), RhoA-GTP (Active RhoA), and Cdc42-GTP (Active Cdc42) and total Rac1, RhoA, and Cdc42 levels in H1703 NSCLC cells expressing NT RNAi or cells expressing Ect2 RNAi and either empty control vector (V), wild-type Ect2 (WT), T328A Ect2 (T328A), or T328D Ect2 (T328D). B, quantitative analysis of Rac1-GTP levels. Data represent the mean ± S.E. from three independent determinations. * indicates significantly different from NT control; ** indicates significantly different from empty vector; p < 0.05. Data shown are representative of three independent experiments.

Thr-328 Is Highly Phosphorylated in NSCLC Cell Lines and Primary NSCLC Tumors

To further characterize Thr-328 phosphorylation in Ect2-mediated NSCLC transformation, we generated a rabbit polyclonal antibody that specifically recognizes Ect2 phosphorylated at Thr-328. Affinity-purified Thr(P)-328 Ect2 antibody was assessed for specificity by immunoblot analysis of purified phospho-Ect2 peptide conjugated to BSA using the corresponding unphosphorylated Ect2 peptide conjugated to BSA as a negative control (Fig. 5A). The Thr(P)-328 Ect2 antibody specifically recognizes the phospho-Thr-328 peptide conjugate but not the unphosphorylated peptide conjugate (Fig. 5A). The Thr(P)-328 Ect2 antibody also recognizes S-peptide-tagged Wt Ect2 but not S-peptide-tagged T328A or T328D Ect2 mutants expressed in Ect2 RNAi cells (Fig. 5B). We also characterized the specificity of the Thr(P)-328 Ect2 antibody by incubating total Ect2 immunoprecipitated from H1703 cells in the presence or absence of λ phosphatase. Immunoblot analysis reveals that phosphatase treatment completely abolishes the immunoreactivity of Thr(P)-328 Ect2 antibody for Ect2 (Fig. 5C). Taken together, these results demonstrate that our Thr(P)-328 Ect2 antibody is highly specific for Ect2 phosphorylated at Thr-328. Immunoblot analysis of Ect2 immunoprecipitated from nontransformed MDCK cells, and the NSCLC cell lines H1703 and A549 for Thr(P)-328-Ect2 demonstrate that both NSCLC cell lines express much higher levels of Thr(P)-328-Ect2 than MDCK cells, which express relatively little Thr(P)-328-Ect2 (Fig. 5D). We next assessed phospho-Ect2 in a set of four primary NSCLC tumors and matched adjacent lung tissues obtained from surgically resected NSCLC tumors. Because Ect2 is overexpressed in NSCLC tumors, we first conducted immunoblot analysis to determine the relative abundance of total Ect2 in these samples (Fig. 5E, upper panel). These results confirmed our previous finding (6) that NSCLC tumors overexpress Ect2 compared with their matched normal tissues. Densitometric analysis revealed a significant 2.3-fold increase in total Ect2 in these tumors compared with matched normal (n = 4, p < 0.05). To assess the relative abundance of phospho-Thr-328-Ect2, we immunoprecipitated total Ect2 from an appropriate amount of lysate from each sample to yield an equivalent amount of total Ect2. These immunoprecipitates were then subjected to immunoblot analysis for phospho-Ect2 and total Ect2 (Fig. 5E, lower panel). Immunoblot analysis for total Ect2 confirmed that equal amounts of total Ect2 were present in each immunoprecipitate. Furthermore, immunoblotting for Thr(P)-328-Ect2 revealed that the level of Thr(P)-328-Ect2 was higher in the tumor samples when compared with matched normal lung tissue. Densitometric analysis for Thr(P)-328-Ect2 revealed a significant 2.7-fold increase in Thr(P)-328-Ect2 in tumors compared with matched normal tissue (n = 4; p < 0.05). These results not only confirm that total Ect2 levels are elevated in primary NSCLC tumors, but they also demonstrate that Thr-328 Ect2 phosphorylation is increased in these tumors (Fig. 5E).

FIGURE 5.

Characterization of a phospho-specific Thr(P)-328 Ect2 antibody. A, dot blot analysis of a Thr(P)-328 Ect2 rabbit antibody. The indicated amount of an Ect2 peptide-bovine serum albumin conjugate containing either Thr-328 or phospho-Thr-328 (T328-BSA or Thr(P)-328-BSA) was slotted onto a nitrocellulose filter and subjected to immunoblot analysis with affinity-purified Thr(P)-328 Ect2 antibody as described under “Experimental Procedures.” B, immunoblot analysis of Ect2 immunoprecipitates from Ect2 RNAi cells transfected with empty vector, WT, T328A, or T328D Ect2 mutants with Thr(P)-328-Ect2 and pan-Ect2 antibodies. C, λ phosphatase treatment destroys the immunoreactivity of Thr(P)-328 Ect2 antibody against WT Ect2. Total Ect2 immunoblot serves as a loading control in B and C. D, immunoblot analysis for Thr(P)-328 Ect2 and total Ect2 from MDCK, H1703, and A549 cells. Lysates from each cell line were subjected to immunoprecipitation with a total Ect2 antibody followed by immunoblot analysis using Thr(P)-328 Ect2 and total Ect2 antibodies. H1703 and A549 cells express high levels of phospho-Thr-328 Ect2 when compared with MDCK cells. E, total lysates (30 μg of protein) from NSCLC tumors (T) and matched normal tissue (N) were subjected to immunoblot analysis to determine the relative levels of total Ect2 (upper panel). Ect2 was immunoprecipitated from extracts of primary NSCLC tumors (T) and matched adjacent normal lung tissues (N) normalized to total Ect2 content, and subjected to immunoblot analysis for Thr(P)-328 Ect2 and total Ect2 (lower panel). In each case (P1–P4), Thr(P)-328-Ect2 levels are higher in tumor tissue than in matched normal lung tissue.

We previously demonstrated that NSCLC cells exhibit mis-localization of Ect2 to the cytoplasm where it binds to the oncogenic PKCι-Par6 complex (6). Therefore, we wished to assess whether Thr(P)-328-Ect2 was preferentially localized to either the nucleus or cytoplasm of NSCLC cells. For this purpose, NSCLC cells were fractionated into nuclear and cytoplasmic fractions as described previously (6), and each fraction was subjected to immunoblot analysis for Thr(P)-328-Ect2 and total Ect2 (Fig. 6A). Whereas total Ect2 distributed roughly equally between the nuclear and cytoplasmic fractions of NSCLC cells, densitometry indicated that the Thr(P)-328-Ect2/total Ect2 ratio was 2.3-fold higher (n = 3; p < 0.05) in the cytoplasmic fraction than in the nucleus (Fig. 6A). The purity of our cellular fractions was confirmed by immunoblot analysis for the nuclear A/C lamins and MEK1, markers of the nucleus and cytoplasm, respectively. Thus, cytoplasmic Ect2 is highly phosphorylated at Thr-328 when compared with nuclear Ect2 in NSCLC cells.

FIGURE 6.

PKCι directly phosphorylates Ect2 on Thr-328 and regulates Ect2 binding to the PKCι-Par6 complex. A, nuclear and cytoplasmic extracts from H1703 cells were subjected to immunoprecipitation for total Ect2 followed by immunoblot analysis for Thr(P)-328-Ect2 and total Ect2. Immunoblot analysis of lamin A/C and MEK1 serve as markers of nucleus and cytoplasm, respectively. B, RNAi-mediated knockdown of PKCι and Par6 leads to a reduction in cellular Thr(P)-328 Ect2 abundance. H1703 cells were transduced with either NT, PKCι, or Par6α RNAi as described previously (6). Lysates were subjected to Ect2 immunoprecipitation (IP) followed by immunoblot analysis for Thr(P)-328-Ect2 and total Ect2. C, PKCι directly phosphorylates Ect2 at Thr-328. Recombinant human PKCι and recombinant human Ect2 were combined in kinase buffer in the presence or absence of phosphatidic acid (PA) as described under “Experimental Procedures.” Samples were then subjected to immunoblot analysis for Thr(P)-328-Ect2 and total Ect2. D, binding of WT Ect2 and Ect2 mutants to the PKCι-Par6 complex. H1703 NSCLC cells expressing FLAG-Par6α were transduced with either NT or Ect2 RNAi. Ect2 RNAi cells were then transfected with either empty vector (Vector), or S-peptide-tagged WT Ect2, (WT), T328A Ect2 (T328A), or T328D Ect2 (T328D). Lysates were subjected to immunoprecipitation with S-peptide antibody followed by immunoblot analysis for Ect2, PKCι, and FLAG-tagged Par6. WT and T328D Ect2 efficiently co-immunoprecipitates PKCι and Par6, whereas T328A Ect2 does not. Data shown are representative of three independent experiments.

PKCι Phosphorylates Ect2 at Thr-328 in Vitro and Regulates Thr(P)-328 Ect2 Levels in NSCLC Cells

NSCLC cell transformation requires formation of the PKCι-Par6α complex and PKCι kinase activity (18, 19). We therefore reasoned that substrates important for PKCι-mediated transformation might associate with the PKCι-Par6α complex. We recently identified Ect2 as a component of the PKCι-Par6α complex in NSCLC cells and demonstrated that Ect2 becomes mislocalized to the cytoplasm as a result of its interaction with the PKCι-Par6α complex (6). Therefore, we assessed whether the phosphorylation status of Thr(P)-328-Ect2 is regulated by expression of the PKCι-Par6 complex. For this purpose, we assessed the level of Thr(P)-328-Ect2 in cells made deficient in PKCι or Par6 using lentiviral RNAi as described previously (18). Both PKCι- and Par6-RNAi cells exhibited a decrease in Thr(P)-328-Ect2 when compared with NT RNAi cells but showed no change in total Ect2 expression (Fig. 6B), indicating that the PKCι-Par6 complex is involved in regulating Thr-328-Ect2 phosphorylation. Given the preferential localization of Thr(P)-328-Ect2 to the cytoplasm and the decrease in Thr(P)-328 Ect2 phosphorylation in PKCι- and Par6-RNAi cells, we assessed whether Ect2 serves as a direct substrate for PKCι. Analysis of Thr-328 using the ScanSite phosphorylation site prediction program suggested that Thr-328 might represent a PKCι phosphorylation site. To directly assess whether PKCι is a Thr-328-Ect2 kinase, we determined the ability of purified recombinant PKCι to phosphorylate recombinant bacterially expressed Ect2. When Ect2 is incubated with PKCι, we observe Thr(P)-328-Ect2 phosphorylation that is not observed in recombinant Ect2 in the absence of PKCι (Fig. 6C). Furthermore, the level of Thr(P)-328-Ect2 is greatly increased in the presence of phosphatidic acid, a known activator of atypical PKCs (23–25). These data demonstrate that PKCι directly phosphorylates Ect2 at Thr-328 in vitro.

Phosphorylation of Ect2 at Thr-328 Regulates Ect2 Binding to the PKCι-Par6 Complex

To assess the role of Thr-328 phosphorylation in Ect2 binding to the PKCι-Par6 complex, we determined the ability of WT-Ect2, T328A-Ect2, and T328D-Ect2 to bind to the PKCι-Par6 complex. For this purpose, we immunoprecipitated S-peptide-tagged Ect2 mutants from Ect2 RNAi NSCLC cells (Fig. 6D). Both FLAG-tagged Par6 and endogenous PKCι were efficiently co-immunoprecipitated with WT-Ect2 and the T328D-Ect2 mutant, whereas little or no Par6 or PKCι co-immunoprecipitated with the T328A-Ect2 mutant. These data indicate that PKCι-mediated Thr-328 Ect2 phosphorylation regulates the association of Ect2 with the PKCι-Par6 complex in NSCLC cells.

DISCUSSION

Ect2 was initially identified as a proto-oncogene capable of transforming NIH/3T3 fibroblasts (7). However, the molecular mechanism(s) by which the oncogenic function of Ect2 is regulated are not well understood. Analysis of the original oncogenic Ect2 clone revealed that it encodes an N-terminally truncated protein that retains the C-terminal RhoGEF catalytic domain. This Ect2 truncation mutant mis-localized to the cytoplasm, possessed constitutive GEF activity, and could transform fibroblasts in vitro (8). In contrast, full-length Ect2 localized almost exclusively to the nucleus and was not capable of transforming fibroblasts (8), consistent with the observation that sequences within the N-terminal region of Ect2 serve to regulate Ect2 localization and RhoGEF activity. Interestingly, deletion of the C-terminal region of this truncated Ect2 significantly reduced Ect2 transforming activity (8, 13) and led to a selective loss of GEF activity toward Rac1 (13). Thus, an Ect2 mutant consisting of the isolated DH/PH domains (Ect2-DH/PH) was able to activate only RhoA in vitro and in vivo, whereas an Ect2 mutant consisting of the DH, PH, and C domains (Ect2-DH/PH/C) activated RhoA, Rac1, and Cdc42 in vitro. Expression of the Ect2-DH/PH mutant enhances actin stress fiber formation indicative of RhoA activation, whereas expression of the Ect2-DH/PH/C mutant caused lamellipodia formation characteristic of Rac1 activation (13). These data provided circumstantial evidence that Rac1 is a critical effector of Ect2-dependent transformation and is consistent with our subsequent findings in NSCLC (6). These data also indicate that the C domain of Ect2 is important for conferring Rac1 GEF activity upon Ect2.

These early results indicated that Ect2 is theoretically capable of causing cellular transformation and suggested a connection between cytoplasmic localization of Ect2, Rac1 GEF activity, and oncogenic potential. However, these studies did not specifically address the role of Ect2 in human tumor cell transformation or the mechanisms underlying the cytoplasmic localization of oncogenic Ect2. We recently demonstrated that Ect2 is genetically and biochemically linked to the PKCι oncogene in human NSCLC cells and primary NSCLC tumors (6). Specifically, we found that PKCι and Ect2 are co-amplified and overexpressed as part of the 3q26 amplicon in NSCLC tumors. Furthermore, we demonstrated that Ect2 is mis-localized to the cytoplasm of NSCLC cells as a consequence of its binding to the oncogenic PKCι-Par6 complex (6). Binding of Ect2 to the PKCι-Par6 complex is associated with activation of Rac1, which is essential for the oncogenic activity of Ect2 (6). These results provided a fundamental mechanistic understanding of how Ect2 functions in transformation and the molecular basis for its cytoplasmic mis-localization in tumor cells.

Our current results identify a molecular mechanism by which the PKCι-Par6 complex can activate the oncogenic potential of Ect2. Here, we demonstrate that Ect2 isolated from NSCLC cell lines and primary NSCLC tumors is highly phosphorylated at a previously uncharacterized site, Thr-328. Thr-328 is located within the central hinge region of Ect2 between the N-terminal regulatory domain and the C-terminal GEF domain. Interestingly, earlier studies demonstrated that deletion of this region results in an Ect2 mutant that exhibits transforming activity, albeit weaker than an N-terminal truncation mutant (8). Our data are consistent with a role for this region as a negative regulator of Ect2 oncogenic activity. Our data also demonstrate that phosphorylation at Thr-328 is important for the oncogenic activity of Ect2 and that Thr-328 phosphorylation regulates the binding of Ect2 to the PKCι-Par6 complex.

Our results are consistent with previous studies demonstrating a role for phosphorylation in the regulation of Ect2 function (1, 3–5). Ect2 has been shown to be phosphorylated transiently at two major sites within the hinge region during mitosis, Thr-341 and Thr-412 (1, 3). These sites are phosphorylated by Cdk1 in a cell cycle-dependent manner and are important for regulating Ect2 localization and activity during cytokinesis. Interestingly, our results indicate that the Thr-341 and Thr-412 phosphorylation sites are not involved in the oncogenic activity of Ect2 in NSCLC cells. These data support our previous finding that the oncogenic activity of Ect2 is distinct from its role in cytokinesis (6) and provide further evidence that the oncogenic activity of Ect2 is unrelated to cytokinesis, at least in NSCLC cells. Our data indicate that the three identified phosphorylation sites on Ect2, Thr-328, Thr-341, and Thr-412, play distinct roles in Ect2 function. The fact that ectopic expression of alanine mutants at each of these sites in NSCLC cells depleted of endogenous Ect2 did not result in a cytokinesis defect indicates that NSCLC cells have evolved mechanisms for executing cytokinesis that are either independent of Ect2 or that require extremely low levels of Ect2 expression. Such an Ect2-independent cytokinesis mechanism has also been described in HT1080 fibrosarcoma cells (22).

Our current data provide novel insight into the biochemical and functional relationship between PKCι and Ect2. We previously demonstrated that these genes are genetically linked through coordinate amplification and overexpression in NSCLC tumors and biochemically linked through their association in a cytoplasmic oncogenic complex with Par6 (6). Our current data demonstrate that in addition to being a binding partner for PKCι, Ect2 also serves as a critical PKCι substrate in transformed cells and primary NSCLC tumors. Recombinant PKCι phosphorylates Ect2 at Thr-328 in vitro, and genetic knockdown of PKCι, or its binding partner Par6, leads to a decrease in cellular Thr(P)-328 Ect2 phosphorylation. The simplest interpretation of these data is that PKCι directly phosphorylates Ect2 at Thr-328 in NSCLC cells, although the involvement of an intervening kinase in vivo cannot be formally ruled out. Our finding that Thr-328 is highly phosphorylated in both NSCLC cell lines and primary tumors provides evidence that PKCι-mediated phosphorylation of Ect2 is important in NSCLC tumor biology in vivo. Expression of Thr-328 Ect2 mutants indicates that Thr-328 phosphorylation regulates the binding of Ect2 to the PKCι-Par 6 complex and affects the GEF activity of Ect2 for its critical downstream effector, Rac1. It remains to be determined whether Thr-328 phosphorylation functions primarily to activate the intrinsic GEF activity of Ect2 toward Rac1, induce, and/or stabilize the association of Ect2 with the PKCι-Par6 complex which serves to bring Ect2 to its substrate Rac1, or both. Our finding that a T328D-Ect2 mutant does not induce Rac1 activity or restore transformation in NSCLC cells lacking PKCι (as a result of RNAi-mediated knockdown) indicates that the interaction with the PKCι-Par6 complex is a critical feature of the oncogenic function of Ect2. Because Rac1 is a well characterized component of the PKCι-Par6 complex via interactions with the CRIB-like domain on Par6 (26), we favor the interpretation that Thr-328 phosphorylation functions primarily to induce and/or stabilize the interaction of Ect2 with the PKCι-Par6 complex where it can efficiently activate Rac1. In conclusion, our present data identify a novel PKCι phosphorylation site on Ect2 that functions to regulate binding of Ect2 to the PKCι-Par6 complex, thereby inducing Rac1 activation and cellular transformation.