Post-translational Processing of RhoA

RhoA and related GTP-binding proteins are modified post-translationally at their carboxyl terminus to form a prenylcysteine methyl ester. The synthesis and post-translational modification of RhoA and Cdc42 were examined in the RAW264 macrophage cell line, and the effect of carboxyl methylation on protein turnover was determined. Cells were labeled with [35S]cysteine, and RhoA or Cdc42 was immunoprecipitated with specific antibodies. Both RhoA and Cdc42 were methylated rapidly in control cells, with little accumulation of unmethylated protein. Carboxyl methylation of RhoA was inhibited by incubation of cells with a carbocyclic adenosine analog, 3-deazaaristeromycin, resulting in the accumulation of unmethylated RhoA. Under these conditions, Cdc42 methylation was inhibited only partially. When methylation was inhibited, the RhoA half-life decreased from 31 to 12 h, and the Cdc42 half-life decreased from 15 to 11 h. The increased degradation of unmethylated RhoA demonstrates a novel function for carboxyl-terminal prenylcysteine carboxyl methylation in protecting RhoA and related proteins from degradation.

The Rho family of GTP-binding proteins is a family of important signal-transducing proteins which includes the Rho, Rac, and Cdc42 proteins (1,2). These proteins are integral components in the signaling pathways for many cellular functions, including the organization of cytoskeletal proteins required for cell motility, adhesion, and proliferation (2,3). The Rho proteins are also involved in signaling pathways for the activation of specific transcription factors (4). Different members of the Rho family appear to activate distinct signaling pathways. In fibroblasts, activation of Cdc42 is required for cytoskeletal changes involved in filipodia formation (5), whereas the activation of Rac is involved in lamellipodia formation and membrane ruffling (6), and Rho activation results in the formation of actin stress fibers and focal adhesions (7). These cytoskeletal changes are believed to be mediated through specific protein kinases that are activated by the GTPbound form of the Rho family proteins (8).
All members of the Rho protein family are modified posttranslationally at their carboxyl terminus. These proteins are synthesized initially with a carboxyl-terminal Cys-Axx-Axx-Xxx sequence (where Axx is usually an aliphatic amino acid, and Xxx is any amino acid), which is a signal for post-translational processing by the three sequential reactions of prenylation, truncation, and methylation (9). For RhoA and Cdc42, the carboxyl terminus is modified to form a geranylgeranylcysteine methyl ester (10,11). Lipid modification of the carboxyl terminus is believed to be important for the interaction of Rho with both guanine nucleotide exchange proteins and with downstream effector targets (2).
Carboxyl methylation of proteins containing a carboxyl-terminal prenylcysteine is catalyzed by a membrane-bound prenylcysteine methyltransferase (12). In yeast, the STE14 gene has been shown to encode a membrane-bound prenylcysteine methyltransferase that methylates both yeast RAS and the mating a-factor (13). Both the STE14 methyltransferase and the mammalian prenylcysteine methyltransferase activity can methylate peptide substrates that contain either a carboxylterminal farnesyl-or geranylgeranylcysteine (14,15). In mammalian cells, much of the methyltransferase activity is localized to microsomal membranes (16), but some activity has been reported in neutrophils to be associated with the plasma membrane (17). Because the mammalian methyltransferase has not been purified, it is not certain if multiple forms of the enzyme are present in the membranes.
Carboxyl methylation of some Rho family proteins can be stimulated in vitro by activation of the protein with GTP␥S 1 (18,19) or by activation of Rho-dependent signaling pathways in intact cells by the addition of chemoattractants (20). In their inactive form, the Rho proteins can form heterodimers with the Rho guanine nucleotide dissociation inhibitor protein (RhoGDI) (21). Unmethylated Cdc42 from brain (the brain form of Cdc42 is also known as G25K) purifies as a soluble complex with RhoGDI, and RhoGDI blocks methylation until Cdc42 is in the active GTP␥S-bound conformation (22). Upon activation, the conformational change brought about by guanine nucleotide binding and the subsequent change in interaction between the GTPase and RhoGDI result in translocation of the protein to the membrane, where methylation by the membrane-bound methyltransferase can occur (22).
The role of carboxyl methylation in Rho protein function has not been defined clearly. It has been reported that methylation may increase membrane attachment (23,24) and interactions with specific effector proteins (25). Whereas the prenylation and truncation steps are irreversible, carboxyl methylation can be reversed by hydrolysis of the methyl ester to form methanol. A number of esterase (26,27) and protease activities (27) in cell extracts have been reported to hydrolyze prenylcysteine methyl esters in vitro, but it has not been demonstrated which of these activities can act in vivo to hydrolyze methyl esters from protein of the Rho family. Several prenylcysteine analogs that inhibit carboxyl methylation also inhibit chemoattractantstimulated signal transduction, suggesting that methylation may have a direct role in these signal tranduction pathways (28). However, these compounds may disrupt chemoattractant signaling pathways by a mechanism that is independent of their inhibition of the methyltransferase (29 -31).
To investigate carboxyl methylation of the Rho protein family in vivo, the carboxyl-terminal processing of RhoA and Cdc42 was examined in the RAW264 macrophage cell line. The data reported here show that at least 90% of RhoA and Cdc42 becomes methylated in exponentially growing RAW264 cells. Inhibiting RhoA methylation decreases half-life of the protein significantly, indicating a novel role for carboxyl methylation in protecting RhoA from degradation.

EXPERIMENTAL PROCEDURES
Materials-A mouse monoclonal antibody, RhoA (26C4), directed against amino acid residues 120 -150 of human RhoA (Santa Cruz Biotechnology), was used for immunoprecipitation of RhoA. An anti-Cdc42 antibody was prepared from the serum of rabbits immunized with hexahistidine-tagged Cdc42 ((His) 6 -Cdc42). The cDNA for the human brain form of Cdc42Hs (also known as G25K) was obtained from Susan Munemitsu. The Cdc42Hs cDNA was subcloned into the pQE plasmid, and the protein was expressed in Escherichia coli and purified using the QIAexpress system (Qiagen). The (His) 6 -Cdc42 was purified further by chromatography using a Mono Q column (Pharmacia Biotech). The (His) 6 -Cdc42 was dialyzed and applied to a Mono Q column equilibrated with 20 mM Tris, pH 8, 1 mM EDTA, 1 mM dithiothreitol. The column was eluted at a flow rate of 1 ml/min with a linear gradient from 0 to 0.5 M NaCl in 25 min, and the peak of (His) 6 -Cdc42 was pooled. An affinity column was prepared by coupling (His) 6 -Cdc42 to cyanogen bromide-activated Sepharose 4B (Pharmacia). The anti-Cdc42 antibody was affinity purified by diluting the immune serum with 3 volumes of PBS and applying the diluted sample to the affinity column. The column was washed with PBS, and bound antibody was eluted with 0.05 M glycine, pH 2.3, and 0.15 M NaCl, into 0.25 volume of 0.5 M sodium phosphate, pH 7.7.
For immunoprecipitation, protein G-agarose (Santa Cruz Biotechnology) was used for binding the mouse monoclonal anti-RhoA antibody, and protein A-Sepharose (Pharmacia) was used to bind the rabbit anti-Cdc42 antibody. [ 35 S]Cysteine (stabilized, Ͼ1,000 Ci/mmol) was obtained from Amersham.
Growth and Labeling of RAW264 Cells-RAW264 cells were grown and subcultured as described previously (32). For labeling cells with [ 35 S]cysteine, confluent RAW264 cells were scraped into fresh MEM and 10% heat-inactivated fetal calf serum and plated in 75-cm 2 flasks at T/10 with MEM and 10% heat-inactivated fetal calf serum and grown for 48 h. The medium was then removed and replaced with serum-free MEM containing 20 M L-cysteine. The cells were incubated at 37°C for 1 h, the medium was then replaced with 3 ml of MEM containing 20 M [ 35 S]cysteine (0.25 mCi), and the cells were incubated for the indicated times. The medium was removed, and the cells were rinsed with 20 ml of warmed PBS. For chase experiments, the PBS rinse was removed from the flask and replaced with MEM and 10% heat-inactivated fetal calf serum (containing 0.2 mM L-cystine), and the incubation was continued at 37°C for the indicated times. Cells were harvested by scraping into 10 ml of ice-cold PBS and were collected by centrifugation at 250 ϫ g for 10 min. The cell pellets were frozen on a dry ice/methanol bath and stored at Ϫ70°C.
Preparation of Cell Extracts and Immunoprecipitation-Frozen cell pellets were suspended in homogenization buffer (10 mM Tris, pH 7.5, 5 mM MgCl 2 ) containing the following protease and esterase inhibitors: 0.5 mM phenylmethylsulfonyl fluoride, 2.5 g/ml aprotinin, 10 g/ml E-64, and 20 g/ml ebelactone B (Calbiochem). Each cell pellet was suspended in 0.5 ml of ice-cold homogenization buffer with inhibitors and placed on ice for 10 min. The cells were then disrupted by 10 passages through a 21-gauge syringe needle, and the suspension was centrifuged at 4°C for 10 min at 500 ϫ g. The supernatant fluid was collected, and 50 l of 5% SDS and 10 mM dithiothreitol was added. The sample was heated for 5 min at 100°C, and the solution was diluted into 2 ml of PBS containing 1% Nonidet P-40 and 0.5% deoxycholate.
The solution was centrifuged at 18,000 ϫ g for 5 min at 4°C, and the supernatant fluid was collected. The antibodies (2-5 g) were added to 0.5-ml aliquots of the supernatant fluid, and the samples were rocked at 4°C overnight. A suspension of protein A-Sepharose (for the rabbit IgG anti-Cdc42 antibody) or protein G-agarose (for mouse IgG 1 anti-RhoA antibody) was added and the incubation continued for an additional 4 -5 h. The immunoprecipitates were collected by centrifugation at 2,000 ϫ g for 5 min, and the pellet was washed four times in 1 ml of PBS, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS. For each wash, the pellet was collected by centrifugation. The bound proteins were solubilized from the beads by the addition of SDS-PAGE sample buffer to the pellet and heating at 100°C for 5 min. The samples were then centrifuged for 3 min at 18,000 ϫ g, and the supernatant fluid was collected for SDS-PAGE. For two-dimensional PAGE, the bound protein was solubilized at room temperature in 30 l of sample lysis buffer containing 0.5% SDS (33). Each sample was centrifuged, and the supernatant fluid was applied to the isoelectric focusing tube gel and focused overnight at 500 V.
Gel Electrophoresis and Immunoblots-Proteins were separated by SDS-PAGE using 12% acrylamide Tris-glycine gels (Novex) or by twodimensional PAGE using a modification of the method of O'Farrell (33), using preblended pH 5.0 -8.0 carrier ampholytes (Ampholine, Pharmacia Biotech). After electrophoresis, the gels were stained with Coomassie Blue and treated with Enlightening (NEN Life Science Products). The gels were dried and exposed to X-Omat film (Kodak) at Ϫ70°C for 4 -9 days, and the developed films were scanned with a scanning densitometer (Molecular Dynamics) to quantitate the protein band or spot density.
For immunoblots, the proteins were transferred to polyvinylidene difluoride membranes. RhoA and Cdc42 were detected using anti-RhoA and anti-Cdc42 antibodies at a dilution of 1:500 in blocking buffer, and the bound antibody was localized by alkaline phosphatase-linked secondary antibody of either anti-rabbit IgG or anti-mouse IgG (1:2,000). Alkaline phosphatase activity was detected using nitro blue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate as substrates.

Synthesis and Carboxyl Methylation of RhoA and Cdc42-
The synthesis and post-translational processing of RhoA and Cdc42 were examined by labeling RAW264 cells with [ 35 S]cysteine followed by immunoprecipitation of RhoA and Cdc42 from 35 S-labeled proteins in the cell extracts. Methylation of RhoA and Cdc42 occurs only after the carboxyl terminus is first modified by prenylation and cleavage of the carboxyl-terminal tripeptide. Because carboxyl methylation neutralizes the negative charge of the carboxyl-terminal ␣-carboxyl group, the methylated protein has a more basic isoelectric point (pI) than the unmethylated protein, and the two forms may be resolved by isoelectric focusing.
When immunoprecipitates of RhoA from [ 35 S]cysteine-labeled cell extracts were separated by two-dimensional PAGE, two labeled proteins with pI values of approximately 5.9 and 6.35 were observed migrating with the 23 kDa of RhoA (Fig. 1). Two pI forms of Cdc42 (pI values of approximately 6.4 and 6.9) were also observed in immunoprecipitates of Cdc42, after separation by two-dimensional PAGE. These two forms of RhoA and Cdc42 were also observed with immunoblots of cell extracts (Fig. 2). As shown below, when methylation was inhibited both RhoA and Cdc42 migrated with the more acidic pI form of the protein. Identification of the methylated proteins on the two-dimensional gel was confirmed by the presence of radiolabeled methyl esters in the proteins when cells were grown with [methyl-3 H]methionine. 2 These results indicated that for both RhoA and Cdc42, the more basic pI protein was carboxyl methylated, and the more acidic form was unmethylated.
The kinetics of RhoA synthesis and methylation were determined from immunoprecipitates of RhoA after cells were labeled with [ 35 S]cysteine for increasing times. After 1 h of labeling, 76% of the newly synthesized RhoA was already methylated; with increasing labeling times, the radioactivity in unmethylated RhoA reached a plateau, while radioactivity in the methylated RhoA continued to increase (Fig. 1). The kinetics indicate that most of RhoA becomes methylated a short time after the protein is synthesized. When further incorporation of radioactivity was blocked by a chase with unlabeled cysteine for 18 h, the unmethylated RhoA decreased significantly, but a detectable amount still remained (Fig. 1). In contrast, the methylated RhoA decreased only slightly during this time. As a result, after a chase of 18 h with unlabeled amino acid, unmethylated RhoA decreased from 16 to 6% of the total labeled RhoA.
In addition to the two spots identified as methylated and unmethylated RhoA, two minor protein spots were also observed in the RhoA immunoprecipitates. The first spot migrated directly above the methylated RhoA, and the second spot migrated between the methylated and unmethylated RhoA spots (Figs. 1B and 3). The identities of these proteins were not determined, and they may represent either further modifications of RhoA or related members of the Rho protein family which were recognized by the anti-RhoA antibody. The protein directly above the methylated RhoA also appears to be methylated because its migration shifts to the unmethylated pI when methylation was inhibited with 3-deazaaristeromycin (Fig. 3). The unprenylated RhoA precursor was not identified in these samples, but it would be expected to migrate just above the prenylated, unmethylated form of RhoA.
Similar kinetics for synthesis and methylation were observed with Cdc42 (data not shown). After 1 h of labeling, only 11% of the radioactivity was in the unmethylated Cdc42; at longer times the label in unmethylated Cdc42 reached a plateau, whereas the radioactivity in methylated Cdc42 continued to increase. After an 18-h chase, the radioactivity in both unmethylated and methylated Cdc42 decreased, such that 9% of the remaining Cdc42 was unmethylated. The total decrease in labeled Cdc42 after the chase indicates that the turnover of Cdc42 was more rapid than RhoA turnover. These results indicate that RhoA and Cdc42 methylation occurs rapidly after protein synthesis, resulting in less than 10% of either protein remaining unmethylated after 18 h.
The steady-state levels of methylated and unmethylated RhoA and Cdc42 were determined from immunoblots of RAW264 cell extracts after separation by two-dimensional PAGE. After transfer to the membranes, Cdc42 or RhoA was  detected with either anti-Cdc42-or anti-RhoA-specific antibodies (Fig. 2). Similar to the results obtained with immunoprecipitates, each antibody detected two immunoreactive proteins that comigrated with the radiolabeled forms of Cdc42 and RhoA in the immunoprecipitates. Again, the major Cdc42 immunoreactive spot detected on the blots corresponded to the more basic, methylated, form of Cdc42, and the more acidic unmethylated form was estimated to be less than 10% of the total Cdc42 ( Fig. 2A). Similar results were obtained using the anti-RhoA antibody, where the unmethylated RhoA comprised approximately 10% of the total immunoreactive RhoA (Fig. 2B). The immunoblots and the immunoprecipitates both gave results consistent with Cdc42 and RhoA being mostly methylated in RAW264 cells under normal growth conditions.
Inhibition of RhoA and Cdc42 Methylation-To examine the properties of unmethylated RhoA and Cdc42, methylation was inhibited by incubating RAW264 cells with the carbocyclic adenosine analog, 3-deazaaristeromycin (DZAri), a potent inhibitor of S-adenosylhomocysteine hydrolyase (34). This compound has been shown previously to inhibit S-adenosylhomocysteine hydrolase in RAW264 cells, resulting in the accumulation of S-adenosylhomocysteine (AdoHcy) (35), which in turn inhibits of a variety of methyltransferases. Although AdoHcy hydrolase is inhibited by DZAri, the concentration used here is not toxic to RAW264 cells (36), and protein synthesis continues at a normal rate (37). The resulting increase in cellular AdoHcy inhibits the prenylcysteine methyltransferase, and methylated proteins are identified by a shift in migration to a more acidic pI when methylation is blocked.
RAW264 cells were pretreated for 1.5 h with 100 M DZAri to allow time for an inhibitory concentration of AdoHcy to accumulate followed by labeling with [ 35 S]cysteine for 3 h in the presence of the inhibitor. Under these conditions, the label accumulating in unmethylated RhoA increased, accompanied by a corresponding decrease in methylated RhoA (Fig. 3 and Table I). The data are consistent with the more basic pI form of the protein being carboxyl methylated RhoA and the more acidic protein being the unmethylated form of RhoA. A similar shift in labeling of Cdc42 was also observed after treatment with DZAri ( Fig. 4 and Table I). After 3 h of labeling in the presence of DZAri, 69% of the newly synthesized RhoA was unmethylated compared with only 10% unmethylated RhoA in control cells (Table I). After a chase for 18 h with unlabeled cysteine in the presence of DZAri, the radioactivity in unmethylated RhoA decreased, but the label in methylated RhoA remained fairly constant, resulting in a decrease in total RhoA radioactivity (Table I).
Cdc42 methylation was only partially inhibited in the presence of DZAri, resulting in 33% of the labeled Cdc42 being unmethylated after 3 h of labeling ( Fig. 4 and Table I) compared with 6% unmethylated Cdc42 in control cells. After a chase period in the presence of DZAri, unmethylated Cdc42 decreased to 9% of the total labeled Cdc42. Therefore, the increase in cellular AdoHcy inhibited RhoA methylation to a greater extent than the inhibition of Cdc42 methylation.
Effect of RhoA and Cdc42 Methylation on Protein Turnover-The effect of carboxyl methylation on protein turnover was examined for RhoA and Cdc42 by measuring their degradation under conditions where the proteins were either methylated or unmethylated. RAW264 cells were labeled with [ 35 S]cysteine followed by a chase period with unlabeled cysteine to block further labeling of newly synthesized protein, and the radioactivity in RhoA or Cdc42 was determined over time (Fig. 5). Under control conditions, approximately 90% of the labeled RhoA and Cdc42 was methylated at the start of the chase period (Table I). Because these proteins were more than 90% methylated, the observed protein turnover in control cells is approximately equal to the turnover for the methylated forms of RhoA and Cdc42. The apparent half-life of RhoA in control cells was approximately 31 h (Fig. 5A). Treatment of cells with DZAri inhibited methylation such that 69% of radiolabeled RhoA was unmethylated after 3 h of labeling (Table I). Inhibiting methylation increased RhoA turnover significantly, causing the apparent half-life to decrease from 31 to 12 h (Fig. 5A). The turnover of total cell protein was determined from precipitation of total proteins with cold 5% trichloroacetic acid from cell extracts of control and DZAri-treated cells, and no significant change in total protein turnover was observed with DZAri treatment (data not shown).
Like RhoA, Cdc42 was methylated rapidly under control conditions (Table I), and the half-life of methylated Cdc42 was approximately 15 h under these conditions (Fig. 5B). The halflife of methylated Cdc42 was significantly shorter than the 31-h half-life of methylated RhoA. Treatment of cells with DZAri resulted in only 33% of the radiolabeled Cdc42 remaining unmethylated after 3 h (Table I), and only a small increase in Cdc42 turnover was observed, decreasing the apparent halflife from approximately 15 to 11 h (Fig. 5B). However, 67% of the labeled Cdc42 was still methylated under these conditions, so the observed Cdc42 turnover was equal to the sum of the rates for both methylated and unmethylated Cdc42. The results indicated that incubation of cells with DZAri increased the turnover of RhoA and Cdc42, and the increased turnover was correlated with the inhibition of methylation. S]cysteine, and Cdc42 was immunoprecipitated from the cell extracts followed by separation by two-dimensional PAGE. The labeling conditions and treatment with 3-deazaaristeromycin are described in the legend for Fig. 3.

DISCUSSION
This study examined the synthesis and methylation of RhoA and Cdc42 in RAW264 cells, and the results demonstrated an increased turnover of RhoA when methylation of the carboxylterminal prenylcysteine was blocked. Under the normal growth conditions, both RhoA and Cdc42 were approximately 90 -95% methylated (Fig. 2), and the kinetics of methylation indicated that both proteins are methylated a short time after their synthesis (Fig. 1). The lag between synthesis and methylation appeared to be a slightly longer for RhoA than for Cdc42, since at early labeling times unmethylated RhoA accumulated to a greater extent than the unmethylated Cdc42 (24% unmethylated RhoA versus 11% unmethylated Cdc42 after 1 h). However, after an 18-h chase period, approximately equal proportions of RhoA and Cdc42 were methylated.
Raising the cellular AdoHcy concentration inhibited methylation of RhoA and Cdc42, but the inhibition was greater for RhoA than for Cdc42 (Figs. 3 and 4). The reason for the difference in inhibition is not clear. A previous report on the mechanism of prenylcysteine methyltransferase in mammalian membranes reported that the enzyme has an ordered Bi Bi mechanism, where AdoHcy is a mixed type inhibitor with respect to the methylation peptide substrate (38), so AdoHcy may be a better inhibitor with RhoA as the methylation substrate. The preferential inhibition of RhoA methylation could also indicate the presence of two distinct methyltransferases in the membranes, with each one having a different susceptibility to inhibition by AdoHcy. Regardless of the detailed mechanism, the results indicate that raising AdoHcy levels in vivo can preferentially inhibit the methylation of specific protein substrates.
Although methylation of RhoA and Cdc42 occurred rapidly after synthesis, a small fraction of labeled RhoA and Cdc42 remained unmethylated even after an 18-h chase. The persistence of radioactivity in the unmethylated protein suggests the presence of a small pool of unmethylated RhoA and Cdc42 which is either methylated very slowly or is never methylated. This pool of unmethylated protein could be caused by localization of the protein to a separate compartment that is inaccessible to the methyltransferase. Alternatively, unmethylated protein may be generated by the slow removal of methyl esters from a previously methylated protein.
The turnover of Cdc42 was significantly faster than RhoA turnover. In control cells, the half-life of Cdc42 was approximately 15 h, whereas the half-life of RhoA was approximately 31 h. The half-life for mammalian proteins can range from minutes to hundreds of hours (39), so the half-lives of both RhoA and Cdc42 are within an intermediate range for mammalian proteins. Protein degradation is a complex process, and many structural features can influence protein turnover. Protein turnover can be regulated by covalent modifications, and several covalent modifications have been shown to increase degradation of specific proteins (40). For RhoA, ADP-ribosylation of the protein by the Clostridium difficile toxin A has been shown to lower the level of RhoA dramatically after an overnight incubation (41), causing the loss of polymerized actin fibers and rounding up of the cells.
The data presented here demonstrate significantly faster degradation of RhoA when the protein is unmethylated, indicating a role for carboxyl methylation in protecting the protein from degradation. For Cdc42, protein turnover was increased only slightly by raising the level of AdoHcy; however, a large fraction of labeled Cdc42 was still methylated under these conditions ( Fig. 4 and Table I). The decrease in half-life was specific for proteins that were carboxyl methylated because total protein degradation did not change. In addition to the Rho proteins, a number of other signal-transducing GTP-binding proteins contain a carboxyl-terminal prenylcysteine methyl ester, including members of the Ras family and the G-protein and transducin ␥-subunits. Therefore, prenylcysteine carboxyl methylation may be an important mechanism to regulate the turnover of these proteins as well.
Methylation could decrease RhoA degradation by several mechanisms. Methylation could decrease protein degradation by inhibiting the activity of carboxypeptidases. Hrycyna and Clarke (42) previously reported a soluble yeast carboxypeptidase that can slowly degrade peptides ending in a carboxylterminal farnesylcysteine, and they proposed that carboxyl methylation may "cap" the carboxyl terminus of yeast RAS, making the protein less susceptible to degradation by these enzymes. Methylation may also cause the protein to bind more tightly to specific proteins or to membrane components, resulting in a conformation that is more resistant to proteolysis. Alternatively, the unmethylated carboxyl-terminal prenylcysteine residues may be a signal for degradation which is recognized by specific proteases, or it may direct transport of the protein to a site for degradation, such as the lysosomes or proteosomes.
Because much of the prenylcysteine methyltransferase activity is localized to the microsomal fraction (16), methylation may be required for normal transport and localization of the newly synthesized Rho proteins. A precedent for methylation FIG. 5. Effect of carboxyl methylation on RhoA and Cdc42 turnover. RAW 264 cells were labeled with [ 35 S]cysteine for 3 h followed by a chase with unlabeled cysteine for the indicated times, as described under "Experimental Procedures." RhoA or Cdc42 were then immunoprecipitated, separated by SDS-PAGE, and the radioactivity was detected by autofluorography. The labeled bands corresponding to RhoA or Cdc42 were quantitated by scanning densitometry. The radioactivity remaining in RhoA or Cdc42 is expressed as a percent of the radioactivity in each protein at the start of the chase. Each data point is the average of duplicate determinations. Panel A, radioactivity in RhoA immunoprecipitated from U, control cells; f, ϩ0.1 mM 3-deazaaristeromycin. Panel B, radioactivity in Cdc42 immunoprecipitated from E, control cells; Ⅺ, ϩ0.1 mM 3-deazaaristeromycin. altering peptide transport is the yeast a-factor, where methylation of the a-factor by the STE14 methyltransferase is required for efficient transport of the peptide out of the cell by the STE6 transporter protein (43).
The effect of methylation on RhoA turnover suggests that RhoA levels in vivo may be influenced by competition between methyltransferase and the methylesterase activities. Methionine depletion would decrease the level of AdoMet and result in decreased protein methylation, which would increase degradation of the Rho proteins. In this manner, the level of RhoA and related proteins could be regulated by the cellular methionine pool. Unmethylated RhoA could also be generated by removal of the methyl ester by an esterase from a previously methylated protein. Therefore, activation of a methylesterase could also increase the amount of unmethylated protein, leading to increased protein degradation.
Previous studies have proposed that carboxyl methylation of Rho family of GTP-binding proteins is important for their function in signal transduction. The results presented here provide evidence that carboxyl methylation has a novel role in increasing the half-life RhoA and related proteins. In this role, methylation may not be directly involved in the signal transduction pathway, but methylation may influence signaling by changing the level of RhoA. Studies of additional proteins will be required to determine if the carboxyl methylation of carboxylterminal prenylcysteine is a general mechanism for altering the degradation of these signal-transducing proteins.