|Home | About | Journals | Submit | Contact Us | Français|
Accumulating evidence has implicated Rho GTPases, including Rac1, in many aspects of cancer development. Recent findings suggest that phosphorylation might further contribute to the tight regulation of Rho GTPases. Interestingly, sequence analysis of Rac1 shows that Rac1 T108 within the 106PNTP109 motif is likely an extracellular signal-regulated kinase (ERK) phosphorylation site and that Rac1 also has an ERK docking site, 183KKRKRKCLLL192 (D site), at the C terminus. Indeed, we show here that both transfected and endogenous Rac1 interacts with ERK and that this interaction is mediated by its D site. Green fluorescent protein (GFP)-Rac1 is threonine (T) phosphorylated in response to epidermal growth factor (EGF), and EGF-induced Rac1 threonine phosphorylation is dependent on the activation of ERK. Moreover, mutant Rac1 with the mutation of T108 to alanine (A) is not threonine phosphorylated in response to EGF. In vitro ERK kinase assay further shows that pure active ERK phosphorylates purified Rac1 but not mutant Rac1 T108A. We also show that Rac1 T108 phosphorylation decreases Rac1 activity, partially due to inhibiting its interaction with phospholipase C-γ1 (PLC-γ1). T108 phosphorylation targets Rac1 to the nucleus, which isolates Rac1 from other guanine nucleotide exchange factors (GEFs) and hinders Rac1's role in cell migration. We conclude that Rac1 T108 is phosphorylated by ERK in response to EGF, which plays an important role in regulating Rac1.
The Rho family of small GTPases mediates a plethora of cellular effects, including regulation of cell size, proliferation, apoptosis/survival, cell polarity, cell adhesion, cell motility, and membrane trafficking (1). As a consequence of the large number of key functions assigned to Rho proteins, it is not surprising that they play important roles in many human diseases. Accumulating evidence has implicated Rho GTPases in many aspects of cancer development, especially in cancer cell invasion and metastasis. Deregulated Rho GTPases have been discovered in many human tumors, including colon, breast, lung, myeloma, and head and neck squamous cell carcinomas (2). Rho GTPases and the signal pathways regulated by them have thus been proposed as potential anticancer therapeutic targets (3).
The Rho family of GTPases accounts for as many as 23 candidate members. These include several branches from which the archetypes RhoA, Rac1, and Cdc42 have been the most well characterized (4, 5). Within the subfamily of Rac, Rac1, Rac2, and Rac3 share significant sequence identity (~88%). These three diverge primarily in the C-terminal 15 residues. Rac1 regulates important cellular processes relevant to cancer cell behaviors and transduces signals in a variety of oncogenic pathways. All of the Rac-related proteins regulate actin polymerization and the formation of lamellipodia and membrane ruffles, presumably through interaction with the WAVE complex (6). Endothelium-specific excision of Rac1 results in embryonic lethality in midgestation (around embryonic day 9.5 [E9.5]) (7, 8). The in vivo and in vitro studies in the last decades have firmly established the role of Rac1 in cancer cell invasion and metastasis (9). Rac1 can stimulate matrix metalloproteinase 1 (MMP-1) or membrane type 1 (MT1)-MMP production in lung cancer cell lines and enhance invasion in vitro (10). When adherens junctions are weakened by epidermal growth factor (EGF) or hepatocyte growth factor, Rac is required to promote cell migration and invasion (11, 12).
Like all members of the small GTPase superfamily, the regulatory cycle of Rac1 is exerted by three distinct families of proteins: the activator or guanine nucleotide exchange factors (GEFs) and two families of suppressors, the GTPase-activating proteins (GAPs) and the guanine nucleotide dissociation inhibitors (GDIs). The cycling of Rho proteins between the GTP- and GDP-bound states might be required for effective signal flow through Rho GTPases to elicit downstream biological functions, and this could involve the concerted action of all classes of the regulatory proteins (13, 14).
Prenylation also plays a role in the regulation of Rho GTPase function. It is generally believed that the newly synthesized Rac1 is geranylgeranylated, which increases the hydrophobic character of Rho GTPases and thus targets them to the plasma membrane and promotes their activation by facilitating interaction with GEFs (13). Recent findings suggest that additional regulatory mechanisms such as posttranscriptional regulation by microRNAs (14), ubiquitination (15), palmitoylation (16), and phosphorylation (17) might further contribute to the tight regulation of Rho GTPases. RhoA was the first Rho protein shown to be phosphorylated. The cyclic AMP (cAMP)-dependent protein kinase (PKA) and the cGMP-dependent protein kinase (PKG) phosphorylate RhoA on serine (S) 188 in vitro and in vivo (18, 19). This phosphorylation does not modify its GTPase activity and its interaction with GEFs and GAPs (20). However, phosphorylated RhoA significantly increases its interaction with RhoGDI (20, 21). Subsequently, the other members of the Rho family have been shown to be regulated by serine or tyrosine phosphorylation. Cdc42 is phosphorylated at tyrosine (Y) 64 by Src, and this phosphorylation results in the increased interaction between Cdc42 and GDI (22). RhoE is phosphorylated at S11 by ROCK 1, and this phosphorylation causes the cytosolic relocation and increased stability of RhoE (23). Rac1 is phosphorylated at S71 by Akt (24). The phosphorylation of Rac1 inhibits its GTP binding activity without any significant change in GTPase activity. Both the GTP binding and GTPase activities of the mutant Rac1 S71A are abolished regardless of the activity of Akt (24). Moreover, Rac1 may be phosphorylated at Y64 by focal adhesion kinase (FAK) and Src, which plays a role in the regulation of cell spreading (25).
Recently we determined that phospholipase C-γ1 (PLC-γ1) is a Rac1 GEF both in vitro and in vivo (26). We showed that the interaction between PLC-γ1 and Rac1 is mediated by the PLC-γ1 Src homology 3 (SH3) domain and the Rac1 proline-rich motif 106PNTP109 (26). Moreover, we showed that an EGF-induced interaction between the PLC-γ1 SH3 domain and Rac1 106PNTP109 motif resulted in the activation of Rac1 and enhanced EGF-induced cytoskeleton reorganization and cell migration. These findings established the proline-rich motif 106PNTP109 as an important regulatory element of Rac1 (26).
Extracellular signal-regulated kinase (ERK) cascades are critical in regulating cell proliferation, survival, and differentiation. Aberrant regulation of ERK cascades contributes to cancer and other human diseases (28). EGF activates ERK through the activation of Ras (29). ERK phosphorylates the serine or threonine in the dipeptide motif S/T-P. There is some preference for proline at position −2 or −3 relative to the phosphoacceptor (30). Moreover, the selectivity of ERK substrate is dependent on ERK docking sites (D sites) with the core consensus motif (K/R)1-3-X1-6-ϕ-X-ϕ (where ϕ is a hydrophobic residue), located on ERK-interacting proteins (31, 32). Thus, it is interesting that Rac1 T108 within the 106PNTP109 motif is likely an ERK phosphorylation site. More interestingly, Rac1 also contains an ERK D site at its C terminus, 183KKRKRKCLLL192 (Fig. 1A). These observations together suggest that it is very likely that in response to EGF, ERK phosphorylates Rac1 at T108. In this communication, we provide evidence to demonstrate that ERK phosphorylates Rac1 T108 in response to EGF stimulation. We show that ERK interacts with Rac1 and that this interaction is mostly dependent on the D site of Rac1. We also show that T108 phosphorylation alters Rac1 activity and Rac1's subcellular localization. Finally, we show that T108 phosphorylation impacts Rac1 function in mediating cell migration.
COS-7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 IU/ml penicillin, and 100 μg/ml streptomycin and were maintained in a 5% CO2 atmosphere at 37°C. For the EGF treatments, COS-7 cells were serum starved for 12 to 16 h, followed by addition of EGF to a final concentration of 50 ng/ml for 15 min or as indicated in the figures. To inhibit ERK activation, cells were pretreated with 5 μM U0126 for 30 min and then incubated with EGF. MCF-7 cells and MDA-MB-231 cells were cultured in DMEM–F-12 (50:50) supplemented with 10% FBS, 100 IU/ml penicillin, and 100 μg/ml streptomycin.
Plasmid DNA for transfection was prepared by using Qiagen midiprep kits according to the manufacturer's instructions. COS-7 cells were grown to 70 to 80% confluence in 6-cm dishes before the transfection. The transfection was performed using the calcium phosphate transfection method with N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (BES) buffer (140 mM NaCl, 0.75 mM sodium phosphate dibasic [Na2HPO4], 25 mM BES, pH 6.95). Cells were typically analyzed at 60 to 72 h posttransfection.
Mouse monoclonal anti-Rac1 antibody was purchased from Cytoskeleton, Inc. (Denver, CO). Mouse monoclonal antithreonine and rabbit anti-phospho-ERK (anti-p-ERK) substrate (PXTP) antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). Mouse monoclonal anti-PLC-γ1 antibody was from Upstate Biotechnology, Inc. (Lake Placid, NY). Rabbit anti-green fluorescent protein (anti-GFP) antibody was from Clontech (Mountain View, CA). Mouse monoclonal anti-p-ERK, antitubulin, anti-E-cadherin, rabbit anti-ERK, and rabbit anti-lamin A antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Purified His-tagged Rac1 protein was from Cytoskeleton, Inc. (Denver, CO). Purified active ERK1 was purchased from SignalChem (Richmond, BC, Canada). Glutathione cross-linked to 4% agarose, goat anti-mouse IgG conjugated with agarose, protein A conjugated with agarose, and amido black staining solution were purchased from Sigma-Aldrich (St. Louis, MO). Mammalian protein extraction reagent (M-PER) was purchased from ThermoFisher Scientific, Inc. (Rockford, IL). Unless otherwise specified, all chemicals were purchased from Sigma-Aldrich.
GFP-tagged wild-type Rac1 (GFP-Rac1) was a gift from Mark R. Philips (School of Medicine, New York University). A glutathione S-transferase (GST) fusion p21-activated kinase (PAK) Rho binding domain (GST-PAK) construct was a gift from Gary Eitzen (University of Alberta). GST-tagged wild-type Rac1 (GST-Rac1) and GST-tagged PLC-γ1 SH3 (GST-PLC-γ1 SH3) were generated previously in the laboratory (27).
All the mutants with point mutations were created with a QuikChange multiple site-directed mutagenesis kit (Stratagene, La Jolla, CA) with GFP-Rac1 or GST-Rac1 as templates. These mutants include a GFP-tagged mutant Rac1 with mutation of threonine 108 to alanine (GFP-Rac1 T108A) and a GFP-tagged mutant Rac1 with mutation of threonine 108 to glutamic acid (GFP-Rac1 T108E). We also created a GFP-tagged mutant Rac1 with the deletion of the ERK D site (GFP-Rac1ΔD), which lacks the 10 amino acid residues from positions 183 to 192. A GFP-tagged mutant Rac1 with mutation of T17 to asparagine (GFP-Rac1 T17N) and a GFP-tagged mutant Rac1 with mutation of Q61 to leucine (GFP-Rac1 Q61L) were generated previously in the laboratory (26). The GST-tagged Rac1 mutants were produced similarly except using GST-Rac1 as the template. Plasmids were sequenced to confirm the presence of the desired mutations.
To purify various GST fusion proteins, pGEX plasmids containing GST alone, GST-Rac1, GST-PAK, and GST-PLC-γ1 SH3 constructs were transformed into Escherichia coli DH5α. Bacteria were grown to an optical density at 600 nm (OD600) of 0.6 to 0.8 at 37°C, induced with 0.2 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and incubated for 4 h at 30°C with shaking. After pelleting, bacterial cells were lysed by sonication in phosphate-buffered saline (PBS) in the presence of protease inhibitors [0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 10 μg/ml aprotinin, and 1 μM pepstatin A]. After sonication, 1% Triton X-100 was added to enhance solubilization. Particulates were removed by centrifugation for 15 min at 10,000 rpm, and the cleared supernatant was incubated with 50:50 glutathione-Sepharose beads (Sigma-Aldrich) in PBS for 2 h at 4°C. The beads were washed three times with ice-cold PBS and stored. The immobilized GST fusion proteins on beads were used for GST pulldown assays, and eluted GST, GST-Rac1, and GST-Rac1 T108A were used for in vitro kinase assays.
COS-7 cells were treated with or without EGF and U0126 and then lysed into BOS buffer (50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 1% Nonidet P-40, 10% glycerol, 10 mM NaF, 2.5 mM MgCl2, and 1 mM EDTA) with protease inhibitors. The lysates were centrifuged at 21,000 × g at 4°C for 15 min. Supernatants were used in the pulldown assay. GST fusion proteins bound to glutathione-Sepharose beads were added to the supernatant and incubated at 4°C for 2 h with shaking. Beads were collected by centrifugation and washed three times with BOS buffer, after which 2× sample loading buffer was added. The pulled down proteins were resolved on SDS-PAGE and analyzed by Western blotting.
To load GST-Rac1 with GTP-γS, GST-Rac1 immobilized on glutathione-Sepharose was loaded with 200 μM GTP-γS in a buffer containing 25 mM Tris-HCl, pH 8.0, 1 mM dithiothreitol (DTT), 40 g/ml bovine serum albumin (BSA), 4.7 mM EDTA, and 0.16 mM MgCl2 for 20 min at 30°C. The reaction was stopped by the addition of MgCl2 to a final concentration of 10 mM, and then the product was mixed with COS-7 cell lysates.
Rac1 activity was determined by using an assay developed by Ren and Schwartz (33) as we described previously (26). The Rac1 binding domain of PAK, a Rac1 effector, was used as a GST fusion protein to pull down active Rac1. Briefly, COS-7 cells with transfections were lysed into GST-PAK buffer (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1% Triton X-100, and 10 mM MgCl2) with protease inhibitors. The lysates were centrifuged at 21,000 × g at 4°C for 15 min. Supernatants were used in the binding assay. GST-PAK fusion proteins bound to glutathione-agarose beads in GST-PAK buffer were added and incubated at 4°C for 2 h. Beads were collected by centrifugation and washed three times with GST-PAK buffer, after which SDS loading buffer was added. The pulled down active Rac1 proteins were resolved on SDS-PAGE and analyzed by Western blotting.
GST-Rac1 or GST-Rac1 T108A was eluted from glutathione-Sepharose beads in glutathione elution buffer (10 mM reduced glutathione and 50 mM Tris-HCl, pH 8.0). Approximately 2 μg of GST, GST-Rac1, or GST-Rac1 T108A and 5 μg of purified His-tagged Rac1 were incubated with 0.1 μg of active ERK1 in kinase buffer (5 mM morpholinepropanesulfonic acid [MOPS], pH 7.2, 2.5 mM β-glycerol-phosphate, 5 mM MgCl2, 1 mM EGTA, 0.4 mM EDTA, 0.05 mM dithiothreitol) in the presence of 200 μM ATP and 5 μCi of [γ-32P]ATP at 30°C for 60 min in a volume of 25 μl. The reaction was stopped by addition of SDS-PAGE sample loading buffer, and the mixture was boiled for 5 min. Samples were then separated by SDS-PAGE (8% gel), transferred to a polyvinylidene difluoride membrane, and subjected to autoradiography.
Immunoprecipitation (IP) experiments were carried out as described previously (34). Briefly, cells were lysed with IP buffer [20 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 0.1% sodium deoxycholate, 100 mm NaF, 5 mM MgCl2, 0.5 mM Na3VO4, 0.02% NaN3, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 10 μg/ml aprotinin, and 1 μM pepstatin A]. Cell lysates were centrifuged at 22,000 × g for 30 min to remove debris. The supernatants, containing approximately 1 mg of total protein, were precleared with the agarose beads and then were used for incubation with 1 μg of specific antibody at 4°C overnight with gentle mixing. Then, goat anti-mouse IgG conjugated with agarose or protein A conjugated with agarose was added to each fraction, and the fractions were incubated for 2 h at 4°C with agitation. Both the agarose beads and the nonprecipitated supernatant were collected by centrifugation. For the controls, mouse or rabbit IgG was used to replace the primary antibodies. The agarose beads were washed three times with IP buffer and then mixed with 2× sample loading buffer. The sample was boiled for 5 min and subjected to a Western blot assay.
The protein content of cell lysates was determined by Bradford analysis, and approximately 20 μg of total protein was used for each sample. Protein samples were resolved by SDS-PAGE and electrophoretically transferred onto nitrocellulose membranes. After being blocked in 3% milk for 60 min, membranes were incubated with primary antibody at 4°C overnight. The primary antibodies were detected with their corresponding horseradish peroxidase-conjugated secondary antibodies, followed by enhanced chemiluminescence development (Pierce Chemical, Rockford, IL) and light detection with Fuji (Tokyo, Japan) Super RX film.
We performed two different subcellular fractionations. For cells transfected with GFP-tagged Rac1 and mutants, we separated the cell homogenates into two fractions: a nuclear faction and a nonnuclear fraction. This subcellular fractionation was performed as we described previously (35). Briefly, transfected COS-7 cells were treated with or without EGF (50 ng/ml) and scraped into homogenization buffer [0.25 M sucrose, 20 mM Tris-HCl, pH 7.0, 1 mM MgCl2, 4 mM NaF, 0.5 mM Na3VO4, 0.1 mM 4-(2-aminoethyl)-benzenesulfonyl fluoride, 10 μg/ml aprotinin, and 1 μM pepstatin A]. The cells were homogenized with a Dounce homogenizer, and then the lysate was passed through a 25-gauge needle 10 times. The nuclei were then separated from the homogenate by being spun down at 200 × g for 10 min two times. The supernatant was then centrifuged at 14,000 × g for 10 min to spin down contaminating nuclei and cell debris. The supernatant, which contained cytoplasm and cell membrane, was then saved. The pellet of the first centrifugation was suspended in homogenization buffer and was then centrifuged at 200 × g for 10 min at least three times to remove cytoplasmic contaminations. The pellets were then suspended in M-PER and used as nuclear extracts. The loading volumes of both fractions were adjusted based on equal cell equivalents, and fractions were analyzed by SDS-PAGE and Western blotting.
To locate the endogenous Rac1, the homogenates of COS-7 cells were subcellularly fractioned to the nuclear, total membrane, and cytosolic fractions. Briefly, COS-7 cells were treated with EGF (50 ng/ml) for the times indicated in the figures, and the nuclear fraction was obtained as described above. The postnuclear supernatant was centrifuged at 100,000 × g for 30 min to yield a supernatant which was collected as the cytosolic fraction, and the pellet was resuspended in 2× SDS-PAGE loading buffer and collected as the total membrane fraction.
Cells were grown on glass coverslips. After transfection for 48 h, cells were serum starved for 12 h and treated with 50 ng/ml EGF for 15 min, and then the cells were fixed by immersion in 4% paraformaldehyde for 5 min and washed three times with PBS. The cells were examined for GFP-tagged Rac1 with an inverted fluorescence microscope (Axiovert 200; Carl Zeiss, Inc., Germany) with a Plan-Apochromat 63× (numerical aperture [NA], 1.40) oil immersion objective equipped with a digital charge-coupled-device (CCD) camera and Northern Eclipse software (Empix Imaging, Inc., Canada). The nuclear localization of Rac1 is easily identifiable with the naked eye. To quantitate the percentage of the cells with GFP-Rac1-positive nuclei, we counted at least 20 transfected cells for each experiment, and each datum is the average of three experiments with more than 60 transfected cells.
COS-7 cells were grown on 24-well plates until 70 to 80% confluent and transfected with different GFP-tagged Rac1 constructs. After 48 h, the cells reached 100% confluence and were serum starved for 12 h and treated with 50 ng/ml EGF for 15 min or left untreated. A wound was created by scraping the cell monolayers with a pipette tip. A nearby reference point was created by a needle. The cells were observed under a fluorescence microscope to ensure that the leading edge of the wound with many positively transfected cells was selected for image acquisition. Both phase-contrast and fluorescence images were acquired by matching the reference point with an inverted microscope (Axiovert 200; Carl Zeiss, Inc.) with a Plan-Neofluar 10×/0.30 NA dry objective. To calculate the rate of migration of the transfected cells, we measured the distance traveled toward the center of the wound after 8 h. We performed wound-healing assays only at 8 h, which can limit the effect of the DNA synthesis. At least five randomly chosen areas were quantified using ImageJ software (NIH, Bethesda, MD). Experiments were repeated three times, and an individual photograph was chosen as the example. The relative distance to the reference line was calculated and normalized to that of the nontransfected cells, and the data are expressed as means ± standard errors (SE) of the percentage of the cells transfected with GFP vector.
Results are presented as the means ± standard errors (SE). Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Tukey's post hoc test in SPSS, version 17.0 (SPSS Inc., Chicago, IL). Differences were considered significant at a P value of <0.05.
Rac1 T108 within the 106PNTP109 motif is likely an ERK phosphorylation site, and Rac1 also contains an ERK D site at its C terminus, 183KKRKRKCLLL192 (Fig. 1A). It is very likely that in response to EGF, ERK phosphorylates Rac1 T108. We first determined whether Rac1 is directly associated with ERK in response to EGF. We transfected COS-7 cells with GFP-Rac1. The GFP-Rac1 was then immunoprecipitated by rabbit anti-GFP antibody from COS-7 cells, with or without EGF stimulation. Immunoblotting of the immunoprecipitates showed that Rac1 was associated with ERK (total) but unexpectedly in an EGF-independent manner (Fig. 1B). However, the association of Rac1 with activated ERK was dependent on EGF; the level was at the minimum without EGF, reached a peak at 15 min of EGF stimulation, and decreased to the minimum at 30 min of EGF stimulation (Fig. 1B). To determine whether this was simply due to the change in ERK phosphorylation levels in response to EGF or EGF-dependent binding of activated ERK to Rac1, we examined the pattern of EGF-induced ERK phosphorylation. Interestingly, the change in activated ERK levels followed a very similar pattern as the association of activated ERK with Rac1 (Fig. 1C). These results suggest that the seemingly EGF-induced association of Rac1 and activated ERK is, in fact, due to the increased level of activated ERK in response to EGF.
We next examined whether endogenous Rac1 and ERK interact with each other. The interaction between Rac1 and ERK was examined by co-IP experiments in the COS-7, MCF-7 and MDA-MB-231 cell lines. As shown in Fig. 1D, when Rac1 was immunoprecipitated with anti-Rac1 antibody, ERK was coimmunoprecipitated in all three cell lines even without EGF stimulation. These data suggest that the interaction between ERK and Rac1 occurred under physiological conditions. It is interesting that the breast cancer cell lines MCF-7 and MDA-MB-231 have higher ERK and Rac1 expression levels than COS-7 cells, which may be responsible for the observed strong co-IP of ERK with Rac1 in these breast cancer cells (Fig. 1D).
The observation that there is a physical association between Rac1 and ERK is confirmed by GST pulldown assay. As shown in Fig. 1E, GST-Rac1 strongly pulled down ERK from the total lysates of COS-7 cells, independent of EGF stimulation. We next examined whether the activation status of Rac1 affects its interaction with ERK. We added the nonhydrolyzable GTP analogue GTP-γS to the cell lysate in the presence of GST-Rac1, which increases the amount of GTP-loaded Rac1. Interestingly, the interaction between ERK and Rac1 is not dependent on Rac1 GTP loading (Fig. 1F).
It has been well established that the substrate selectivity of ERK is dependent on the ERK D site (31, 32), and the Rac1 C terminus includes the consensus sequence for the ERK D site (183KKRKRKCLLL192). To determine whether this putative ERK D site on Rac1 mediates the interaction between ERK and Rac1, we performed GST pulldown experiments. We constructed a GST fusion with Rac1 with a deletion of the D site (GST-Rac1ΔD). Lysates of COS-7 cells with or without EGF stimulation were incubated with GST-Rac1ΔD or GST-Rac1. Following the incubation, the proteins associated with GST beads were immunoblotted with antibodies to ERK. As shown in Fig. 2A, while a significant amount of ERK was pulled down by GST-Rac1, no ERK was pulled down by GST-Rac1ΔD. This result indicated that deletion of the D site eliminated the binding between Rac1 and ERK. Our results also confirmed that the interaction between ERK and Rac1 is not EGF dependent (Fig. 2A).
To confirm the role of the Rac1 ERK D site in mediation of the interaction, we performed co-IP experiments. We constructed a GFP-tagged mutant Rac1 with the deletion of the D site (GFP-Rac1ΔD). We then expressed GFP-Rac1 or GFP-Rac1ΔD in COS-7 cells by transient transfection and examined the interaction between ERK and GFP-Rac1ΔD by co-IP. GFP-Rac1 or GFP-Rac1ΔD was immunoprecipitated with antibody to GFP, and the immunoprecipitates were immunoblotted with antibodies to both GFP and ERK. As shown in Fig. 2A, compared with GFP-Rac1, the association between GFP-Rac1ΔD and ERK was significantly decreased, indicating that the direct interaction between Rac1 and ERK is mostly mediated through the ERK D site at the Rac1 C terminus (Fig. 2B).
To determine whether the interaction between Rac1 and ERK results in Rac1 T108 phosphorylation by ERK, we examined whether Rac1 is threonine phosphorylated in vivo in response to EGF. We transfected COS-7 cells with GFP-Rac1. With or without EGF stimulation, GFP-Rac1 was immunoprecipitated by rabbit anti-GFP antibody. Immunoblotting with antibody to phospho-threonine (pT) showed that Rac1 is threonine phosphorylated in response to EGF (Fig. 3A). Moreover, we examined whether EGF-induced Rac1 threonine phosphorylation is mediated by ERK. We showed that inhibition of ERK by U0126 blocked EGF-induced Rac1 threonine phosphorylation, which suggests that EGF-induced Rac1 threonine phosphorylation is likely mediated by ERK (Fig. 3A).
As there are several threonine residues in Rac1 and as only one is within the PXTP motif, to examine whether the threonine within the PNTP motif is the one phosphorylated, we employed an antibody specific to the phospho-PXTP motif (pPXTP) and performed a reciprocal IP experiment. We immunoprecipitated all of the proteins that contained the pPXTP motif from the lysates of COS-7 cells transfected with GFP-Rac1, with or without EGF stimulation. The immunoprecipitates were then immunoblotted with antibody to GFP. We showed that GFP-Rac1 was immunoprecipitated by antibody to pPXTP (Fig. 3B), which suggests that T108 within the PNTP motif is likely phosphorylated.
To confirm that T108 is phosphorylated in response to EGF, we mutated T108 to alanine (A) (T108A) and tagged this mutant with GFP (GFP-Rac1 T108A). COS-7 cells were transfected with GFP-Rac1 T108A and then lysed. The cell lysates were incubated with antibody to pPXTP to immunoprecipitate all of the proteins that contained the pPXTP motif. Immunoblotting of the immunoprecipitates with antibody to GFP indicated that GFP-Rac1 T108A was not immunoprecipitated, which indicates that GFP-Rac1 T108A was not threonine phosphorylated (Fig. 3B). Together, these data strongly indicated that EGF stimulates the phosphorylation of Rac1 at T108.
We showed above that inhibition of ERK blocked EGF-induced threonine phosphorylation of Rac1. To confirm this and to determine whether activated ERK directly phosphorylates Rac1, we did an in vitro ERK kinase assay using purified activated ERK1 and recombinant Rac1 (Fig. 4A and andB).B). We first examined the ability of activated ERK to phosphorylate His-Rac1 by using an ERK kinase assay kit in the presence of [γ-32P]ATP. Elk1 was used as a positive control. As shown in Fig. 4A, His-Rac1 was strongly phosphorylated by ERK at a level similar to that of the positive control Elk1. The finding that purified active ERK was able to phosphorylate the purified Rac1 in vitro in the absence of any other kinases indicates that ERK is a direct kinase of Rac1.
To determine whether T108 is the amino acid residue that is phosphorylated by ERK, we examine whether active ERK can phosphorylate mutant GST-Rac1 T108A. An ERK kinase assay showed that ERK strongly phosphorylated GST-Rac1 but only slightly phosphorylated GST-Rac1 T108A (Fig. 4B and andC).C). As a negative control, GST was not phosphorylated by ERK. Our data indicated that ERK phosphorylates Rac1 mostly on residue T108. The minor phosphorylation observed for GST-Rac1 T108A may be due to either nonspecific phosphorylation or weak phosphorylation of other Rac1 threonine or serine residues by ERK.
We showed that the association of Rac1 with ERK was independent of EGF stimulation (Fig. 1B to toEE and and2A).2A). We wanted to determine whether EGF-induced phosphorylation of both ERK and Rac1 T108 affects the interaction between ERK and Rac1. We expressed GFP-Rac1 and GFP-Rac1 T108A in COS-7 cells by transient transfection. Following stimulation with EGF for the times indicated in Fig. 5A, GFP-Rac1 and GFP-Rac1 T108A were immunoprecipitated with antibody to GFP. Immunoblotting with antibodies to both ERK and p-ERK showed that ERK was associated with both wild-type Rac1 and mutant Rac1 T108A, with or without EGF stimulation (Fig. 5A). We further showed that inhibition of ERK activation by U0126 did not affect the association between ERK and Rac1 (Fig. 5B). These data indicated that the phosphorylation of Rac1 or ERK did not affect their interaction.
We next examined whether T108 phosphorylation alters the activity of Rac1. We constructed a GFP-tagged mutant Rac1 with the mutation of T108 to glutamic acid (E) (GFP-Rac1 T108E) that mimics the phosphorylated T108. We examined Rac1 activity by measuring its GTP loading status. We expressed GFP-Rac1, GFP-Rac1ΔD, GFP-Rac1 T108A, and GFP-Rac1 T108E in COS-7 cells by transient transfection. Following EGF stimulation to activate Rac1, the activated Rac1 was pulled down using the GST-tagged Rac-binding domain of PAK (GST-PAK). As shown in Fig. 6, the GST-PAK protein was able to pull down a significant amount of wild-type Rac1. The activation level of GFP-Rac1ΔD was similar to that of the wild-type Rac1. The mutant GFP-Rac1 T108A that cannot be phosphorylated at T108 by ERK showed higher activity than wild-type Rac1. However, the activation level of GFP-Rac1 T108E was significantly reduced. We also expressed GFP-tagged mutant Rac1 Q61L (constitutive active) and Rac1 T17N (dominant negative) in COS-7 cells as positive and negative controls. Our data indicated that phosphorylation of Rac1 at T108 inhibited EGF-induced activation of Rac1, which suggests that phosphorylation of Rac1 T108 by ERK served as a negative regulatory mechanism.
Next, we determined how T108 phosphorylation hinders the activation of Rac1. We showed previously that the PLC-γ1 SH3 domain interacts with Rac1 106PNTP109 and serves as its GEF to activate Rac1 (26). We proposed that phosphorylation of T108 may hinder the interaction between Rac1 and PLC-γ1 and thus block the activation of Rac1 by PLC-γ1. To test this hypothesis, we examined the interaction between PLC-γ1 and various Rac1 mutants by GST pulldown assay. We first used a GST-tagged PLC-γ1 SH3 domain to pull down Rac1 and Rac1 mutants. GFP-tagged Rac1, Rac1 T108A, and Rac1 T108E were expressed in COS-7 cells by transient transfection. Following EGF stimulation, the cell lysates were incubated with the GST-tagged PLC-γ1 SH3 domain. As shown in Fig. 7A, the PLC-γ1 SH3 domain pulled down both Rac1 and Rac1 T108A but not Rac1 T108E. As Rac1 T108E is a T108 phosphorylation mimic, our data suggest that phosphorylation of Rac1 on T108 blocks its interaction with PLC-γ1. To confirm this result, we used GST-tagged Rac1, Rac1 T108A, and Rac1 T108E to pull down PLC-γ1 from the lysates of COS-7 cells treated with EGF. We showed that PLC-γ1 interacts with wild-type Rac1 and Rac1 T108A but not with Rac1 T108E (Fig. 7B). These data further suggest that phosphorylation of Rac1 T108 disrupted its interaction with PLC-γ1 and thus blocked the activation of Rac1 by PLC-γ1.
The activity and function of Rac1 are closely related to its subcellular localization. Rac1 has been shown to localize to various subcellular compartments, including the cytosol, the plasma membrane, and the nucleus, and various mechanisms have been proposed for the translocation of Rac1 between various subcellular compartments (13, 36, 37). Therefore, we wanted to see whether T108 phosphorylation of Rac1 affects its subcellular localization. We first examined the subcellular localization of Rac1 by fluorescence microscopy. We expressed GFP-Rac1, GFP-Rac1 T108E, and GFP-Rac1 T108A in COS-7 cells by transient transfection, and the localization of Rac1 and the mutants was determined by the intrinsic fluorescence of GFP (Fig. 8A and andB).B). We showed that EGF induced a significant amount of GFP-Rac1 to translocate from the cytosol to the nucleus (Fig. 8A and andB).B). Strikingly, the T108 phosphorylation mimic GFP-Rac1 T108E was almost exclusively localized to the nucleus, with or without EGF stimulation (Fig. 8A and andB).B). On the other hand, the mutant GFP-Rac1 T108A that is unable to be phosphorylated showed little nuclear translocation in response to EGF. These data suggest that T108 phosphorylation stimulates the translocation of Rac1 into the nucleus.
To confirm this finding, we examined the subcellular localization of Rac1 and the mutants by subcellular fractionation (Fig. 8C and andD).D). Similarly, our data showed that EGF stimulated the translocation of wild-type Rac1 to the nucleus. GFP-Rac1 T108E is mostly localized to the nucleus, with or without EGF stimulation. GFP-Rac1 T108A showed little nuclear localization with or without EGF stimulation.
We next examined EGF-induced subcellular localization of endogenous Rac1. COS-7 cells either not treated with EGF or stimulated with EGF for 15 and 60 min were homogenized and then separated into total membrane, nuclear, and cytosolic fractions. The Rac1 level in each fraction was determined by immunoblotting. As shown in Fig. 9, without EGF stimulation Rac1 was present in all three fractions but mostly in the cytosol. The relative amounts of Rac1 in cytosolic, membrane, and nuclear fractions were 81.9%, 7.9%, and 10.2%, respectively. Following EGF stimulation for 15 min, a significant amount of Rac1 was translocated to both the membrane and nucleus. The relative amounts of Rac1 in cytosolic, membrane, and nuclear fractions were 68.5%, 16.1%, and 15.5%, respectively. Following EGF stimulation for 60 min, the Rac1 in nuclear fraction remained stable (16.6%), but the membrane Rac1 that was translocated from the cytosol following EGF stimulation for 15 min was translocated back to the cytosol. The relative amounts of Rac1 in cytosolic and membrane fractions were 76.2% and 7.2%, respectively (Fig. 9).
Together, our results indicate that EGF-induced Rac1 T108 phosphorylation served as a mechanism to target a small portion of Rac1 to the nucleus. Nuclear localization of Rac1 has been shown to promote cell division (38).
We finally conducted a wound-healing assay to examine the effects of Rac1 T108 phosphorylation on cell migration. We overexpressed GFP-Rac1, GFP-Rac1 T108A, and GFP-Rac1 T108E in COS-7 cells by transient transfection. We showed that overexpression of GFP-Rac1 and GFP-Rac1 T108A enhanced EGF-induced cell migration. However, overexpression of GFP-Rac1 T108E did not enhance EGF-induced cell migration (Fig. 10A and andB).B). As a control, we showed that in the absence of EGF, expression of wild-type or mutant Rac1 did not significantly affect cell migration (Fig. 10A and andB).B). These results suggested that overexpression of T108-phosphorylated Rac1 was unable to affect, either by enhancing or hindering, EGF-induced cell migration.
Since the first report of the phosphorylation of RhoA on the serine at position 188 in vitro and in vivo by PKA and the PKG (18, 19), many small GTPases have been shown to be phosphorylated on either serine or tyrosine residues. Cdc42 is phosphorylated at Y64 by Src (22), RhoE is phosphorylated at S11 by ROCK 1, and Rac1 is phosphorylated at S71 by Akt (24) and at Y64 by FAK and Src (25). Here, we report the first threonine phosphorylation of small GTPases. We show that Rac1 is phosphorylated at T108 in response to EGF stimulation and that this phosphorylation is likely catalyzed by ERK. This finding is supported by the following evidence. First, we show that ERK interacts with Rac1 by both co-IP and GST pulldown experiments (Fig. 1 and and2).2). Second, we show that EGF stimulates the threonine phosphorylation of wild-type Rac1 but not mutant Rac1 T108A (Fig. 3). Third, we show that inhibition of ERK activation by U0126 blocks EGF-induced threonine phosphorylation of Rac1 (Fig. 3). Finally, we show that purified active ERK is able to phosphorylate purified Rac1, i.e., both His-Rac1 and GST-Rac1, in vitro. However, mutation of T108 to A largely abolished Rac1 phosphorylation by active ERK in vitro (Fig. 4). Together, these data suggest that ERK may be a Rac1 kinase in vivo. This finding is not surprising as T108 is within the motif 106PNTP108 that confers the preferred phosphorylation site for ERK (30). Moreover, Rac1 also contains an ERK D site at its C terminus (183KKRKRKCLLL192) (Fig. 1A). The core consensus motif of ERK D sites is (K/R)1-3-X1-6-ϕ-X-ϕ (31, 32). Indeed, the D site is critical for the interaction between Rac1 and ERK. Deletion of the D site significantly reduced the interaction between Rac1 and ERK (Fig. 2). However, the weak binding between ERK and Rac1ΔD (Fig. 2B) and the weak phosphorylation of Rac1 T108A by ERK in vitro (Fig. 4B) are noteworthy. These data suggest the possibility of the presence of other ERK binding and phosphorylation sites in Rac1. For example, 86SP87 and 135TP136 could be the candidates for weak ERK binding and phosphorylation sites.
It is interesting that the Rac1 D site is no longer present following its geranylgeranylation. Thus, the interaction between Rac1 and ERK should occur before the geranylgeranylation. It is generally believed that the newly synthesized Rho GTPases such as Rac1 first interact with cytosolic prenyltransferases that add a geranylgeranyl isoprenoid or farnesyl isoprenoid to the cysteine in the C-terminal CAAX (where A is an aliphatic amino acid) motif of the Rho GTPases (36). The addition of a geranylgeranyl group increases the hydrophobic character of Rho GTPases and thus targets them to the plasma membrane and promotes their activation by facilitating interaction with GEFs (13). The immediate geranylgeranylation of Rac1 will limit its chance to interact with ERK. However, a recent finding shows that the entrance and passage of Rho GTPases through the geranylgeranylation pathway may be neither immediate nor automatic. Rather, this process is regulated by two splice variants of SmgGDS (36). This new finding suggests that newly synthesized Rac1 could well interact with other proteins such as ERK before it is geranylgeranylated. In fact, we show that while the phosphorylation of Rac1 by ERK requires EGF-induced activation of ERK, the binding between ERK and Rac1 is not dependent on EGF stimulation and ERK activation. These data suggest that the newly synthesized Rac1 may actively interact with ERK. The formation of Rac1-ERK complex will protect Rac1 from being geranylgeranylated as the Rac1 D site, including the CAAX motif, is occupied by ERK.
The Rac1 D site also overlaps the Rac1 C-terminal polybasic region (PBR). The PBR is comprised of a series of adjacent lysines or arginines, which often immediately precede the C-terminal CAAX motif. In Rac1 the PBR is 183KKRKRK188 (Fig. 1A). The PBR controls diverse functions of Rho GTPases. It promotes the interactions of Rho GTPases with the GEF SmgGDS. It also regulates other important functions of Rho GTPases, including promoting their association with membrane and their interactions with other proteins (37, 39–42). It has also been shown that the Rac1 PBR has a functional nuclear localization signal (NLS) sequence, which acts as the NLS for protein complexes containing Rac1 (42). Here, we show that as part of the Rac1 D site, the Rac1 PBR also mediates the interaction between Rac1 and ERK.
It is critical to determine whether and how the phosphorylation of Rac1 affects its function. We first examined whether the T108 phosphorylation of Rac1 affects its activity. We showed that Rac1 T108 phosphorylation hinders Rac1 activation by EGF (Fig. 6), which suggests that Rac1 T108 phosphorylation serves as a negative regulator of Rac1 activity. Previous research has shown that phosphorylation of RhoA on S188 does not modify its interaction with GEFs and GAPs. Therefore, it does not affect GDP/GTP exchange and activation (20). However, S188 phosphorylation significantly increases the interaction with GDI, independent of the nucleotide loaded on the protein, and enhances the ability of GDI to extract RhoA from membrane (20, 21). Phosphorylation of Cdc42 at Y64 by Src results in the increased interaction between Cdc42 and GDI (22). Phosphorylation of RhoE at S11 by ROCK increases its activity in inducing stress fiber disruption and inhibiting Ras-induced transformation (43). The phosphorylation of Rac1 at S71 by Akt inhibits its GTP binding activity without any significant change in GTPase activity (24). Thus, phosphorylation of Rac1 by either Akt or ERK negatively regulates Rac1 activity.
We further showed that the inhibition of Rac1 activation by T108 phosphorylation is partially due to the effects on the interaction between Rac1 and PLC-γ1. We have shown that the PLC-γ1 SH3 domain interacts with the Rac1 106PNTP109 motif to act as a GEF for Rac1 (26). We showed here that T108 phosphorylation inhibited the interaction between Rac1 and the PLC-γ1 SH3 domain. Thus, T108 phosphorylation effectively blocks Rac1 activation through the PLC-γ1 pathway. However, several other GEFs have been shown to activate Rac1, possibly through a mechanism independent of Rac1 phosphorylation. These GEFs include Vav proteins and Tiam1 (44–47). Indeed, we showed that Rac1 T108 still retains low activity and that the wild-type Rac1 has high activity (Fig. 6).
Another important finding of this research is that T108 phosphorylation targets Rac1 to the nucleus (Fig. 8 and and9).9). We showed that EGF stimulates the phosphorylation and nuclear localization of wild-type Rac1. The phospho-T108-mimicking mutant T108E is mostly localized to the nucleus with or without EGF stimulation (Fig. 8). Targeting Rac1 to different subcellular compartments is an important mechanism to regulate Rac1 activity. It has been shown that the Rac1 PBR has NLS activity to target Rac1 and associated proteins such as SmgGDS to the nucleus (42). However, as PBR is part of the Rac1 D site that is essential for the interaction between Rac1 and ERK and the subsequent phosphorylation of Rac1 T108 by ERK, the role of PBR in targeting Rac1 to the nucleus may be due not only to its NLS activity but also to its role in mediating Rac1 T108 phosphorylation by ERK. We showed that Rac1 T108E is almost exclusively localized into nucleus, which suggests that T108 phosphorylation is a very strong signal to target Rac1 into nucleus. It was recently shown that the nuclear import of Rac1 is mediated by karyopherin α2 through a direct interaction (48).
While the significance of targeting Rac1 to the nucleus is not clear and needs further research, it could serve as a very important mechanism regulating the activity and function of Rac1. Localization to the nucleus will isolate Rac1 from its regulatory proteins localized in the cytoplasm and the plasma membrane. For example, most of the known Rac1 GEFs, including PLC-γ1, Vav proteins, and Tiam1, do not localize to the nucleus (44–47) and thus will not be able to activate the nuclear Rac1. The observed lower activity of T108-phosphorylated Rac1 may be due not only to its inability to interact with PLC-γ1 but also to its isolation from other GEFs. On the other hand, targeting Rac1 to the nucleus will allow its interaction with a very different set of molecules and fulfill different functions. Several reports have shown that Rac1 regulates the transcription of genes by its interaction with STAT3 and STAT5 (49, 50). Rac1 also affects the activity of nuclear factor-κB (NF-κB) (51). Recently, it was reported that Rac1 is localized to the nucleus during the G2 phase of the cell cycle and promotes cell division (38). Thus, Rac1 T108 phosphorylation could be a mechanism to regulate important cell functions such as cell division. It would be interesting to examine in the future whether the nuclear Rac1 during G2 phase is heavily phosphorylated at T108 and whether Rac1 T108E promotes cell division.
Targeting Rac1 to the nucleus will also reduce the amount of Rac1 in the plasma membrane and, thus, may hinder its ability to regulate cytoskeleton remodeling and cell migration. Indeed, we showed here that transfection of Rac1 T108E in COS-7 cells did not enhance EGF-induced cell migration (Fig. 10), which suggests that T108-phosphorylated Rac1 is unable to regulate cell migration due to its nuclear localization and low activity.
Based on our data and previously published data, it is possible that a pool of newly synthesized Rac1 binds to ERK in the cytosol through its D site, which protects this portion of Rac1 from geranylgeranylation and retains this portion of Rac1 in the cytosol. Following EGF stimulation, the Rac1-bound ERK is activated, and the activated ERK phosphorylates the Rac1 T108. The T108-phosphorylated Rac1 translocates to the nucleus. The nuclear Rac1 interacts with various components in the nucleus to regulate cell functions such as cell division. On the other hand, a significant portion of Rac1 is geranylgeranylated and targeted to the plasma membrane, where it functions to regulate cell migration and cytoskeleton remodelling. The ratio between these two pools of Rac1 may vary depending on cell type and cell cycle. It is interesting that some early reports show that lipidated Rac1 could be targeted to the nucleus (38). Thus, it is possible that geranylgeranylated Rac1 could be further divided into distinct pools. Some of them will be associated in the membrane, and some of them will be targeted to the nucleus. As discussed above, various mechanisms have been proposed for targeting Rac1 to the nucleus.
While we do not know the percentage of Rac1 that is phosphorylated following EGF stimulation and the percentage of nuclear Rac1 that is phosphorylated, our data presented in Fig. 9 show that the total nuclear Rac1 following EGF stimulation is 16.6%. The percentage of phosphorylated Rac1 of the total nuclear Rac1 should be much lower than this because other mechanisms could also target Rac1 to the nucleus. We also showed that the membrane-associated Rac1 is 16% following EGF stimulation (Fig. 9). Thus, the pool of membrane-associated Rac1 is similar in size to that of the nuclear Rac1. As the membrane-associated Rac1 has been shown to carry out multiple functions in regulating cell morphology and migration, the similar amount of nuclear Rac1 could also carry out important cell functions. The functions of both nuclear Rac1 and the nonnuclear Rac1 are essential for the cell, and the alteration of their activities may contribute to cancer development.
We thank Mark R. Philips (New York University School of Medicine) for providing GFP-Rac1 and Gary Eitzen (University of Alberta) for providing GST-PAK.
This work was supported in part by grants from the Alberta Cancer Foundation, the Canadian Institutes of Health Research, and the Alberta Innovates-Health Solutions.
Published ahead of print 16 September 2013