Tumor Necrosis Factor α Stimulation of Rac1 Activity

We have previously demonstrated that both isoprenylcysteine carboxylmethyltransferase (ICMT) and one of its substrates, the RhoGTPase Rac1, are critical for the tumor necrosis factor α (TNFα) stimulation of vascular cell adhesion molecule-1 expression in endothelial cells (EC). Here, we have shown that ICMT regulates TNFα stimulation of Rac1 activity. TNFα stimulation of EC increased the membrane association of Rac1, an event that is essential for Rac1 activity. ICMT inhibitor N-acetyl-S-farnesyl-l-cysteine (AFC) blocked the accumulation of Rac1 into the membrane both in resting and TNFα-stimulated conditions. Similarly, the membrane-associated Rac1 was lower in Icmt-deficient versus wild-type mouse embryonic fibroblasts (MEFs). TNFα also increased the level of GTP-Rac1, the active form of Rac1, in EC. AFC completely suppressed the TNFα stimulation of increase in GTP-Rac1 levels. Confocal microscopy revealed resting EC Rac1 was present in the plasma membrane and also in the perinuclear region. AFC mislocalized Rac1, both from the plasma membrane and the perinuclear region. Mislocalization of Rac1 was also observed in Icmt-deficient versus wild-type MEFs. To determine the consequences of ICMT inhibition, we investigated the effect of AFC on p38 mitogen-activated protein (MAP) kinase phosphorylation, which is downstream of Rac1. AFC inhibited the TNFα stimulation of p38 MAP kinase phosphorylation in EC. TNFα stimulation of p38 MAP kinase phosphorylation was also significantly attenuated in Icmt-deficient versus wild-type MEFs. To understand the mechanism of inhibition of Rac1 activity, we examined the effect of ICMT inhibition on the interaction of Rac1 with its inhibitor, Rho guanine nucleotide dissociation inhibitor (RhoGDI). The association of Rac1 with its inhibitor RhoGDI was dramatically increased in the Icmt-deficient versus wild-type MEFs both in resting as well as in TNFα-stimulated conditions, suggesting that RhoGDI was involved in inhibiting Rac1 activity under the conditions of ICMT inhibition. These results suggest that ICMT regulates Rac1 activity by controlling the interaction of Rac1 with RhoGDI. We hypothesize that ICMT regulates the release of Rac1 from RhoGDI.

Some signaling proteins, such as small GTPases and the ␥ subunits of heterotrimeric G proteins, end with a CAAX (C for cysteine, A for aliphatic amino acids, and X for almost any other amino acid) sequence, which is post-translationally modified in three ways (1). Isoprenylation (farnesylation or geranylgeranylation) at cysteine is followed by cleavage of the last three amino acids (AAX) (1)(2)(3). Finally, the isoprenylated cysteine is carboxyl methylated by an endoplasmic reticulum resident enzyme (4,5) isoprenylcysteine carboxylmethyltransferase (ICMT) 1 (5). This carboxyl methylation provides additional hydrophobicity to the isoprenylated proteins. Carboxyl methylation is dependent on the first two modifications and is potentially reversible.
We demonstrated that ICMT and one of its substrates, the RhoGTPase Rac1, regulate in endothelial cells (EC) the inflammatory cytokine TNF␣ stimulation of redox-sensitive expression of vascular cell adhesion molecule-1 (VCAM-1) (6). These studies raised the possibility that ICMT may regulate VCAM-1 expression by regulating Rac1 activity.
Rac1, the most extensively studied Rac, is a ubiquitously expressed CAAX protein that plays a crucial role in endothelial cell biology by regulating gene transcription, actin remodeling, cell shape, motility, secretion, and proliferation (7)(8)(9)(10)(11)(12). Rac1 is an important mediator of the TNF␣ signaling cascade (13)(14)(15)(16). It participates in inflammatory signaling pathways to an important extent by activating the superoxide generating NADPH oxidase (17)(18)(19)(20)(21)(22) and MAP kinase pathways (23). The activation of Rac1 involves multiple steps. In resting cells, GDP-Rac1 is predominantly present in the cytosol as a complex with the Rho guanine nucleotide dissociation inhibitor, RhoGDI. After stimulation RhoGDI releases, by an unestablished mechanism, Rac1 (24), which then translocates to the membrane where GTP is exchanged for GDP by guanine nucleotide exchange factors. GTP-Rac1 then interacts with its effector molecules, e.g. p21-activated kinase (PAK), among others, to mediate its effects. GTP-Rac1 is inactivated by conversion of the GTP to GDP by GTPase activating protein. GDP-Rac1 is then extracted from the membrane by RhoGDI back into the cytosol, thereby completing the activation cycle. Thus, mechanisms regulating the accumulation of Rac1 into the membrane and GTP loading of Rac1 could be important in regulating its activity.
The mechanism(s) regulating the translocation of Rac1 from cytosol to the membrane is not completely understood. Protein kinase C activators and N-formyl-methionyl-leucyl-phenylalanine stimulate the translocation of Rac to the membrane (18,25). The tyrosine kinase inhibitor genistein blocked translocation of Rac in response to chemoattractant stimulation of human neutrophils (26) or to depolarization in EC (27). TNF␣ stimulation of neutrophils activates Rac translocation through a protein tyrosine kinase, but not through the protein kinase C pathway (25). Mutation analysis demonstrated Rac requires a polybasic region (KKRKRK) near the C terminus for its association with the membrane (28). Inhibitors of isoprenylation blocked Rac1 membrane translocation (29). However, inhibition of isoprenylation also prevents post-isoprenylation carboxyl methylation. Thus, the effect of inhibitors of isoprenylation on Rac translocation could also be due to the inhibition of carboxyl methylation. The role of carboxyl methylation in the translocation of Rac1 is not understood. We investigated the role of ICMT in the TNF␣ stimulation of Rac1 activity by assessing membrane-bound Rac1 and cellular GTP-Rac1 levels. We found that ICMT inhibition suppressed the accumulation of Rac1 into the membrane in both resting and TNF␣-stimulated EC. Furthermore, levels of cellular GTP-Rac1, the active form of Rac1, were also suppressed, and Rac1 was found to be mislocalized both from the plasma membrane and the perinuclear region. These effects of ICMT inhibition were found to be mediated, at least in part, through an increase in the interaction of Rac1 with RhoGDI. Thus, post-isoprenylation carboxyl methylation is important in the regulation of Rac1 activity.

EXPERIMENTAL PROCEDURES
Cell Culture and Reagents-Human aortic endothelial cells were purchased from Clonetics and were grown and maintained in growth medium EBM-2 supplemented with 2% fetal bovine serum and EGM-2 SingleQuots (hydrocortisone, human fibroblast growth factor, vascular endothelial growth factor, R3-insulin-like growth factor-1, ascorbic acid, human epidermal growth factor, GA-1000, and heparin). These cells were used between passages 3 and 7. Human dermal microvascular endothelial cells immortalized with simian virus 40 product large T antigen (HMEC-1) were grown and maintained in MCDB131 medium as described elsewhere (30). Icmt-deficient and wild-type control mouse embryonic fibroblast (MEFs) cell lines were isolated and cultured as described (31). N-acetyl-S-farnesyl-L-cysteine (AFC) was purchased from Biomol, and the stock solution was made at 100 mM in dimethyl sulfoxide (Me 2 SO). Antibodies against Rac1 were from Santa Cruz Biotechnology and Upstate Biotechnology. The anti-ICMT antibodies were raised against the 19-amino acid peptide (YKKRVPTGLP-FIKGVKVDL) present at the C terminus of human ICMT and were purified by affinity column.
Cell Extract Preparation and Western Analysis-After treating with various reagents, cells in 100-mm dishes were lysed in 0.7 ml of ice-cold lysis buffer (50 mM HEPES, pH 7.4, 1% Triton X-100, 100 mM NaCl, 5 mM MgCl 2 , 10 g/ml each of leupeptin, aprotinin, and antipain and 1 mM each of phenylmethylsulfonyl fluoride, sodium fluoride, and sodium orthovanadate). The cells were vortexed and put on a rotor for 1 h at 4°C. Samples were spun for 10 min at 8,000 ϫ g to pellet debris, and the supernatant was removed and placed in new tubes. Equal amounts of total protein were size fractionated by SDS-PAGE using 10% gels and transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat milk (Bio-Rad). To assess the levels of various proteins, the membranes were probed with their respective antibodies. Membranes were then incubated with donkey anti-rabbit IgG horseradish conjugate and developed using enhanced chemiluminescence substrates.
Membrane Preparation in Hypotonic Buffer-EC were washed in ice-cold phosphate-buffered saline three times, drained, and incubated with 1 ml of ice-cold hypotonic lysis buffer (10 mM Tris-HCl at pH 7.4, 1.5 mM MgCl 2 , 5 mM potassium chloride, 1 mM dithiothreitol, 1.0 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin) for 5 min. Cells were scraped and homogenized with 15 strokes of a Dounce homogenizer. Homogenates were centrifuged at 700 ϫ g for 3 min to pellet nuclei and intact cells. The supernatants were spun at 100,000 ϫ g for 30 min at 4°C to sediment particulates before being removed. The crude membrane was gently washed with a hypotonic lysis buffer. Membrane and cytosol fractions were then assayed for total protein. Equal amounts were analyzed by Western blotting and quantified by scanning densitometry.
Preparation of Detergent-insoluble Membrane-EC were lysed in an ice-cold lysis buffer (phosphate-buffered saline, pH 7.0, 1% Triton X-100, 1 mM EDTA, 10 g/ml leupeptin and pepstatin, 1 mM phenylmethylsulfonyl fluoride, 2 mM sodium orthovanadate, and 10 mM NaF) prior to being centrifuged (500 ϫ g for 5 min) to separate unbroken cells and nuclei. After centrifugation of the supernatant (100,000 ϫ g for 30 min, 4°C), the pellet was dissolved in 2ϫ SDS-PAGE buffer and boiled for 5 min. Rac1 was immunodetected by Western analysis using an antibody specific to Rac1.
Rac1 Methylation Assays-The methylation of CAAX proteins by ICMT is base-labile. To determine the level of methylated Rac1, human aortic endothelial cells were grown to confluence in 100-mm dishes and labeled with L-[methyl-3 H]methionine (50 Ci/ml) in 4 ml of methionine-free medium (Clonetics) at 37°C overnight. Cell extracts were prepared after treating the cells with either AFC or TNF␣. Crude membrane was prepared as described earlier. The membrane was dissolved in 50 l of 1ϫ SDS-PAGE buffer, diluted to 1 ml with 1ϫ radioimmune precipitation assay buffer, and incubated with agaroseconjugated protein A/G (Santa Cruz Biotechnology) for 2 h at 4°C. After centrifugation at 5000 ϫ g for 5 min, the supernatant was collected and incubated with polyclonal antibodies specific to Rac1 (Santa Cruz Biotechnology) overnight at 4°C. Immune complexes were captured using agarose-conjugated protein A/G and separated by centrifugation. Pellets were washed three times with 1ϫ lysis buffer, and Rac1 was immunodetected by Western analysis. Bands corresponding to Rac1 were cut out from Western blots and carefully placed in 1.5-ml Eppendorf centrifuge tubes (without caps) containing 200 -500 l of 1N NaOH. Tubes were placed in 20-ml scintillation vials containing 5 ml of scintillation fluid. The vials were then capped and left at 37°C overnight to release the [ 3 H]methanol produced by hydrolysis of methyl esters. Radioactivity was determined by scintillation spectrometry.
Immunofluorescence Microscopy-The cells were grown on glass coverslips in 6-well plastic dishes, fixed with 4% paraformaldehyde-phosphate-buffered saline for 30 min, permeabilized in 0.2% Triton X-100 in phosphate-buffered saline for 5 min, and blocked with 3% bovine serum albumin. The cells were incubated with an antibody specific to Rac1 at 1:50 dilution overnight at 4°C. Specific antibody binding was detected by fluorescein isothiocyanate-labeled (green fluorescence) or TRITClabeled (red fluorescence) secondary antibody. Normal rabbit or mouse IgG (5 g/l) was used, instead of a primary antibody, as a negative control in each case. Images were acquired using a Bio-Rad 1024 MRC laser scanning confocal imaging system and represent basic z-sections of the cells.
Rac Pulldown Assays-Rac GTPase assays were performed essentially as described using a kit from Upstate Biotechnology. Briefly, GTP-Rac in cell lysates was pulled down using the Rac-binding domain of PAK linked to glutathione-agarose beads for 30 min at 4°C, washed, and then eluted with SDS-sample buffer. GTP-Rac was analyzed by Western blotting with a monoclonal anti-Rac antibody (Upstate Biotechnology).
Two-dimensional Gel Electrophoresis-Two-dimensional gel electrophoresis was performed as described previously (32) using a pH gradient of 3.5-10.

TNF␣ Stimulation of EC Increases the Membrane-associated
Rac1-The Rho GTPase Rac1 is a major target of the inflammatory signals stimulated by TNF␣ and is important in the TNF␣ stimulation of expression of vascular cell adhesion molecule-1 in EC (6). Rac1 association with membranes reflects its biological activity. To determine whether TNF␣ stimulates Rac1 activity, EC were treated with TNF␣ for 5-80 min, and a crude detergent-insoluble membrane fraction was prepared. Membrane-associated Rac1 was assessed by Western analysis. TNF␣ stimulation of EC for 5-80 min increased the level of membrane-associated Rac1 (Fig. 1). Total Rac1 level remained unchanged. Similar results were obtained when the membranes were prepared in buffer without detergent (data not shown). Thus, TNF␣ stimulation increases the level of membrane-associated Rac1.
TNF␣ Stimulation of EC Increases the Level of Membraneassociated Methylated Rac1-Isoprenylation is necessary for the translocation of Rac1 to the membrane (29). However, the role of the post-isoprenylation carboxyl methylation in the translocation of Rac1 is unknown. To determine whether translocated Rac1 is carboxyl methylated, Rac1 was immunoprecipitated from the membrane fraction, separated by SDS-PAGE, and analyzed by Western analysis. The band corresponding to Rac1 was cut out, and the level of carboxyl-methylated Rac1 was determined by vapor phase equilibration assays. EC membranes contained carboxyl-methylated Rac1 basally. TNF␣ stimulated a further increase in membrane-associated carboxyl-methylated Rac1 at 1 and 5 min (Fig. 2). Thus, TNF␣ stimulation of EC increases the level of membrane-associated carboxyl-methylated Rac1.
ICMT Inhibition Suppresses the Level of Membrane-associated Rac1-To determine whether ICMT modulates the membrane association of Rac1, we measured the effects of the ICMT inhibitor AFC on membrane translocation after TNF stimulation. As shown in Fig. 3A, TNF␣ stimulation increased the level of membrane-associated Rac1 and AFC pretreatment suppressed Rac1 to basal levels. AFC inhibited the level of membrane-associated Rac1 in resting cells. The total Rac1 level remained the same under these conditions. Thus, ICMT is involved in maintaining the membrane levels of Rac1 in both resting and TNF␣-stimulated EC.
To determine whether ICMT also affects Rac1 translocation in other cell types, we evaluated responses in Icmt-deficient and wild-type MEFs. As Fig. 3B shows, the membrane-associated Rac1 in resting wild-type MEFs was below the limit of detection. Stimulating the cells with TNF␣ for 1-60 min increased the level of membrane-associated Rac1. Under the same conditions the level of membrane-associated Rac1 in Icmt-deficient MEFs was very low, both in resting as well as in TNF␣-stimulated cells, although the total level of Rac1 expressed in these cells was higher than that in the wild-type cells. Thus, Icmt also appears to be important for the accumulation of Rac1 in the membrane of MEFs.
ICMT Inhibition Suppresses the GTP-Rac1 Levels in TNF␣stimulated EC-After accumulation into the membrane, Rac1 is loaded with GTP by guanine nucleotide exchange factors. GTP-Rac1, the active form of Rac1, interacts with its effector molecules, e.g. PAK, to perform its functions. To validate the conclusions about Rac1 activity derived from the results obtained with membrane association studies, we performed Rac1 pulldown assays to measure directly GTP-Rac1 levels in EC. As shown in Fig. 4A, TNF␣ stimulation of the cells for 1-35 min increased the level of GTP-Rac1 in HMEC-1 cells. Pretreatment of the cells with AFC (25-50 M) inhibited the increase in GTP-Rac1 levels in TNF␣-stimulated EC (Fig. 4B). Thus, ICMT regulates the TNF␣ stimulation of GTP-Rac1 levels in EC.
Mislocalization of Rac1 in AFC-treated EC-To elucidate the role of ICMT in the localization of Rac1, we performed confocal microscopy. As Fig. 5A demonstrates, Rac1 is present in the plasma membrane and also in the perinuclear region of resting cells. Treatment of EC with AFC mislocalized Rac1 from both the membrane and the perinuclear region to the nucleus.
To determine whether AFC affects the methylation status of Rac1, we performed two-dimensional gel electrophoresis. In resting EC, Rac1 had two major bands, "M" and "U" (Fig. 5B, lower panel). AFC treatment of the cells suppressed the level of band M and concomitantly increased the intensity of the U band of Rac1 (Fig. 5B, upper panel). Thus, the U band appears to represent unmethylated Rac1 and the M band likely represents methylated Rac1. AFC seems to suppress the level of methylated Rac1. These results suggest that ICMT is necessary for the localization of Rac1 to the plasma membrane and the perinuclear region in EC and that the effect is manifested by methylated Rac1.
To determine whether ICMT also plays a role in Rac1 localization in other cells types, we performed studies on Icmtdeficient and wild-type MEFs. As shown in Fig. 5C, Rac1 is localized in the perinuclear region of resting wild-type MEFs but is diffusely located in Icmt-deficient MEFs. Thus, Icmt also plays a role in the targeting and intracellular localization of Rac1 in MEFs.
ICMT Inhibition Suppresses the TNF␣ Phosphorylation of p38 MAP Kinase-p38 MAP kinase requires Rac1 for its activation by TNF␣ and has also been implicated in the regulation of expression of vascular cell adhesion molecule-1 in EC (33). To determine whether ICMT plays a role in regulating the signaling pathways downstream of Rac1, we investigated the effect of AFC on p38 MAP kinase activation by Western analysis. As shown in Fig. 6A, lane 2, TNF␣ stimulated the phosphorylation of p38 MAP kinase in EC severalfold above base-  and 4), whereas the solvent Me 2 SO had no effect (lane 5). Under the same conditions the total amount of p38 MAP kinase remained unchanged. Thus, ICMT also regulates the TNF␣ stimulation of phosphorylation of p38 MAP kinase that is downstream of Rac1.
To determine whether ICMT also plays a role in the TNF␣ stimulation of p38 MAP kinase phosphorylation in other cell types, we performed studies on MEFs. As shown in Fig. 6B, TNF␣ stimulation of wild-type MEFs for 15 min robustly induced the phosphorylation of p38 MAP kinase. Under the same conditions, the TNF␣ stimulation of p38 MAP kinase phosphorylation was significantly reduced in Icmt-deficient MEFs. The expression of p38 MAP kinase was also lower in Icmt-deficient versus wild-type MEFs. Taking into account the reduction in p38 MAP kinase expression, the TNF␣ stimulation of p38 MAP kinase phosphorylation in Icmt-deficient MEFs remained significantly lower compared with that in wild-type MEFs. Thus, Icmt plays a role in the TNF␣ stimulation of p38 MAP kinase phosphorylation in MEFs.

ICMT Regulates the Interaction between Rac1 and RhoGDI-
In resting cells, RhoGDI solubilizes and sequesters Rac1 in the cytosol by masking its isoprenyl group. Stimuli that activate Rac1 are thought to release GDP-Rac1 from RhoGDI through mechanisms that are unclear. GDP-Rac1 translocates to the membrane, where it is loaded with GTP. After performing its function in the membrane, GTP-Rac1 is converted to GDP-Rac1 through the action of RhoGTPase activating proteins. The GDP-Rac1 is then extracted by RhoGDI back to the cytosol. We tested the hypothesis that a decrease in Rac1 level in the membrane after ICMT inhibition could be due to an increase in the interaction of unmethylated Rac1 and RhoGDI. AFC also has an isoprenyl group that can interfere with the interaction of Rac1 with RhoGDI by competing for the isoprenyl binding pocket present within the RhoGDI structure (34 -37). For this reason, we used Icmt-deficient MEFs. These cells have been characterized and contain no residual biochemical activity capable of methylating isoprenylated substrates, such as AFC or N-acetyl-S-geranylgeranyl-L-cysteine (38). Rac1 was immunoprecipitated from lysates of Icmt-deficient and wild-type MEFs. The level of RhoGDI in the immunoprecipitates was deter-

FIG. 3. ICMT inhibitor AFC suppresses the levels of Rac1 in the membrane of both resting and TNF␣-stimulated EC.
A, confluent HMEC-1 cells were treated in serum-free medium with AFC (50 M) or the solvent Me 2 SO before stimulation with TNF␣ for 15 min. The cells were then harvested and crude membrane preparations were prepared as described under "Experimental Procedures." Rac1 levels in the membrane preparations were determined by Western analysis, using antibodies specific to Rac1. The data presented are the mean Ϯ S.E. of three experiments. *, p Ͻ0.05). B, confluent Icmt-deficient and wild-type MEFs were treated with TNF␣ for the indicated time intervals, and crude membranes were prepared in the presence of Triton X-100. Rac1 levels in the membrane were determined by Western analysis using antibodies specific to Rac1. The data presented are representative of two experiments.

FIG. 4. Inhibition of carboxyl methylation reduces the amount of GTP-Rac1 in TNF␣-stimulated EC.
A, confluent HMEC-1 in serum-free medium were treated with TNF␣ at different time intervals, and the Rac1 activity was measured by GTP-Rac1 pulldown assays as described under "Experimental Procedures." B, confluent cells in serum-free medium were treated with either AFC or Me 2 SO before stimulation with TNF␣ for 15 min and Rac1 activity assays were performed. The data presented are the average of two experiments. mined by Western analysis. As shown in Fig. 7A, the level of Rac1-associated RhoGDI was severalfold higher in Icmt-deficient versus wild-type MEFs. Similar results were obtained when RhoGDI was immunoprecipitated from the lysates and the level of associated Rac1 was determined by Western analysis (Fig. 7B). The interaction between RhoGDI and Rac1 re- mained lower in wild-type than in Icmt-deficient MEFs even when proteins from wild-type cells were used at twice the amount of Icmt-deficient cells in immunoprecipitation studies. Thus, in the absence of ICMT, RhoGDI appears to increase its association with Rac1, which might inhibit Rac1 translocation to the membrane.
To determine whether TNF␣ stimulation can influence Rac1/ RhoGDI interaction in cells, we measured the level of Rac1associated RhoGDI in Icmt-deficient MEFs. As shown in Fig.  7C, Rac1-associated RhoGDI level is higher in resting Icmtdeficient versus wild-type cells. TNF␣ stimulation for 1 h did not change the level of Rac1-associated RhoGDI in Icmt-deficient cells. Under the same conditions 1 h of TNF␣ stimulation suppressed the level of Rac1-associated RhoGDI in wild-type cells. Thus, ICMT also appears to play a role in the TNF␣stimulated dissociation of Rac1 from RhoGDI. DISCUSSION Multifunctional Rac proteins play a central role in signaling pathways regulating cellular proliferation, cytoskeletal organization, transcriptional activation, and superoxide generation (7-9, 11, 29, 39, 40). To perform its functions, Rac1 has to be activated, which involves multiple steps. In resting cells GDP-Rac1 is predominantly present as a complex with RhoGDI in the cytosol. After stimulation of the cells, RhoGDI releases Rac1 by an unestablished mechanism (24), which then translocates to the membrane where GTP is exchanged for GDP by guanine nucleotide exchange factors. GTP-Rac1 then interacts with its effector molecules, e.g. PAK, among others, to mediate its effects. After completing its function, GTP-Rac1 through the action of GTPase activating protein is converted to GDP-Rac1, which is then extracted from the membrane by RhoGDI back into the cytosol, thereby completing the Rac1 activation cycle. Previous studies have suggested that isoprenylation is critical for the accumulation of Rac1 into the membrane (29). Our present studies suggest that the inhibition of the post-isoprenylation carboxyl methylation is sufficient to inhibit membrane Rac1 and GTP-Rac1 levels in EC. Somewhat surprisingly, the interaction of Rac1 with RhoGDI was dramatically increased in Icmt-deficient cells, suggesting that ICMT may be involved in the release of Rac1 from RhoGDI.
The crude membranes prepared in detergent-free buffer from resting cells contained Rac1, the level of which was further increased after TNF␣ stimulation of the cells (data not shown). Similar results were obtained when the membranes were prepared in the presence of the detergent Triton X-100, indicating that Rac1 associates stably with the membrane. These results are consistent with previous studies that have suggested that Rac1 is present in detergent-insoluble fractions (41). Furthermore, we found evidence that the Rac1 stably associated with the membrane was carboxyl methylated. The inhibition of ICMT suppressed the level of membrane-associated Rac1, both in resting as well as in TNF␣-stimulated cells. This reduction in the level of Rac1 could be due to a decrease in the translocation of Rac1 from the cytosol to the membrane or a decrease in the retention of Rac1 in the membrane. Because interaction of Rac1 with its inhibitor RhoGDI is increased when ICMT is inhibited, it is possible that the inhibition of Rac1 accumulation in the membrane involves a translocation mechanism. However, we cannot exclude the possibility that carboxyl methylation also affects the retention of Rac1 in the membrane. Regardless, our studies suggest that ICMT plays a role in regulating Rac1 levels in the membrane.
Rac contains a C20 geranylgeranyl group at its C terminus, which anchors Rac to lipid membranes (42). In the cytosol, RhoGDI proteins solubilize Rho proteins by shielding the geranylgeranyl group from the solvent (36,(43)(44)(45). The RhoGDI/Rho interaction is based on two contacts. First, the geranylgeranyl group inserts into a deep pocket made by the immunoglobulinlike ␤ sandwich of RhoGDI; second, the switch I and II regions of Rho interact with the regulatory arm of RhoGDI, a small helixloop-helix motif (36). The second interaction of RhoGDI-Rho makes it partially dependent on the confirmation of Rho. RhoGDI interacts preferentially with the GDP-bound form of Rho proteins (46), although equal interactions with both GDP-and GTP-bound forms have been reported in some cases (47,48). In general, Rac-GDP is considered to be present in a complex with RhoGDI in the cytosol, whereas Rac-GTP is preferentially associated with membranes (43,49,50). Recent studies have suggested that dissociation of RhoGDI and membrane translocation are required for the activation of Rac by the nucleotide exchange factor Tiam (24). However, the mechanism by which Rac1 is dissociated from RhoGDI and translocated to the membrane is not well understood. Observations of the dramatic increase in the interaction of RhoGDI with Rac1 in Icmt-deficient cells lead to the speculation that ICMT may promote the dissociation of Rac1 from RhoGDI by carboxyl methylating it. This is an issue that needs to be addressed in the future. In recent studies, calcium and PKC-dependent phosphorylation of RhoGDI have been proposed to promote the release of bound Rac and RhoA and subsequent translocation to the plasma membrane (8,(51)(52)(53)(54)(55)(56). It was shown recently (57) that integrin signals act on the Rac/RhoGDI interaction, inducing release of Rac to sites of cell adhesion. Phosphorylation of RhoGDI by Pak1 has also been shown to mediate dissociation of Rac (58). Recently, two other RhoGDI displacement factors, the neurotrophin receptor p75NTR and ezrin/radixin/moesin, have been shown to activate RhoA in 293T and Swis 3T3 cells, respectively (59, 60). Thus, reduced affinity of FIG. 7. ICMT regulates interaction between Rac1 and RhoGDI. Cell lysates were prepared from confluent wild-type and Icmt-deficient MEFs. A, Rac1 was immunoprecipitated from the lysates containing equal amounts of protein. The immunoprecipitated proteins were analyzed by Western analysis using antibodies specific to Rho-GDI. B, Rho-GDI was immunoprecipitated from the lysates containing equal amounts of protein. The immunoprecipitated proteins were analyzed by Western analysis using antibodies specific to Rac1. The data presented are representative of two experiments. C, wild-type and Icmt-deficient MEFs were stimulated with TNF␣, and Rac1 was immunoprecipitated from the cell lysates. The Rac1-associated RhoGDI was estimated by Western analysis using antibodies specific to RhoGDI. RhoGDI for Rac may be a common feature preceding its translocation. Our data suggest that ICMT regulates the association of RhoGDI with Rac1; however, whether ICMT can reduce the affinity of RhoGDI for Rac1 remains to be demonstrated.
Confocal microscopy revealed that Rac1 was present in the plasma membrane as well as in the perinuclear region of EC. It is possible Rac1 may have unique roles in these different locations as has been observed in the case of Ras proteins (61). The presence of Rac1 in the perinuclear region raises the possibility that it is involved in the production of intracellular superoxide in this location, because the other subunits of NADPH oxidase, e.g. gp91phox, p22phox, p47phox, and p67phox, have mainly a perinuclear distribution in EC (62). Plasma membrane GTP-Rac can preform a number of functions such as those related to agonist-induced cytoskeletal reorganization but also, more specifically, to formation of reactive oxygen species that are stimulated by TNF␣ and are involved in its signal transduction. The mislocalization of Rac1 from the perinuclear region and plasma membrane in cells after AFC treatment infers that ICMT plays a critical role in regulating multiple functions of Rac1 that are dependent upon its targeting to specific cellular compartments.
The activity of Rac1 has been proposed to be regulated by upstream GTPases such as H-Ras, K-Ras, and ␤␥ subunits of heterotrimeric G proteins (8, 54 -56). Because these CAAX proteins are also substrates of ICMT, it is possible that this enzyme regulates the activity of Rac1 by modulating their methylation.
Because AFC inhibits the TNF␣ stimulation of Rac1 activity, it might also inhibit the events downstream of Rac1. To test this hypothesis we investigated the effect of AFC on the TNF␣ stimulation of p38 MAP kinase, an event that has been shown to be dependent upon Rac1 in several cell types (63). Consistent with our hypothesis, AFC inhibited the TNF␣ stimulation of phosphorylation of p38 MAP kinase. Similarly, the TNF␣ stimulation of p38 phosphorylation was significantly lower in Icmtdeficient compared with wild-type MEFs. In the future, it will be interesting to investigate which other events downstream of multifunctional Rac1 are also regulated by ICMT.
In summary, we have demonstrated that ICMT regulates membrane accumulation and GTP loading of Rac1, events required for the biological activity of Rac1. The inhibition in the activity of Rac1 under the conditions of ICMT inhibition is due, at least in part, to an increase in the interaction of Rac1 with RhoGDI. These studies support the notion that ICMT could release Rac1 from RhoGDI. Because carboxyl methylation is a reversible event, ICMT may be an important regulator of the activity of Rac1. Further studies will shed more light on the mechanisms by which ICMT regulates the activity of Rac1 and will determine which of the events downstream of Rac1 are controlled by ICMT.