Newly synthesized Rho A, not Ras, is isoprenylated and translocated to membranes coincident with progression of the G1 to S phase of growth-stimulated rat FRTL-5 cells.

Ras and Rho are involved in the regulation of signal transduction events governing cell growth and cell proliferation. Protein prenylation is essential for the activation and/or the translocation of these small GTPases; however, protein geranylgeranylation rather than farnesylation is required for G1/S transition. We studied prenylation and translocation of Ras and Rho A during G1/S progression in growth-stimulated rat thyroid FRTL-5 cells. Immunoblot analysis revealed that both Ras and Rho A were detected in membrane fractions at G0. Rho A was eliminated from the membrane fraction during G1 and was not detected on the membrane at mid-G1. Translocation of Rho A from the cytoplasm back to the membranes was observed during late G1 phase. In contrast, Ras remains in the membrane fraction through the cell cycle progression from G1 to S phase. The immunoprecipitation of Rho A from the membrane fraction demonstrated that newly synthesized Rho A, labeled by pulsing cells with [35S]methionine and [35S]cysteine, was geranylgeranylated and associated with the membrane in late G1. These results indicate that Rho A, not Ras, was eliminated from membrane fraction during G1 progression and that newly synthesized Rho A is geranylgeranylated and translocated to membranes during G1/S progression in growth-stimulated FRTL-5 cells.

Over the past five years, a number of small GTPases have been implicated in the signal transduction events governing cell growth and proliferation. Among the most thoroughly characterized is a small GTPase, Ras; Ras proteins function as relay switches transducing biological information from extracellular signals to the nucleus (1)(2)(3). Recently, a number of other small G-proteins, in particular members of the Rho family of GTPases, were also shown to be essential for cell cycle progression (4,5).
The activity of Ras and Rho proteins requires their attachment to the inner leaflet of the plasma membrane (1,6,7). This process is initiated by the covalent attachment of a hydrophobic farnesyl or geranylgeranyl groups to a cysteine at the fourth position from the COOH terminus of the proteins (8). Farnesyl or geranylgeranyl groups are derived from two major intermediate metabolites of the mevalonate, farnesylpyrophosphate, or geranylgeranylpyrophosphate, respectively. The enzymes that catalyze the attachment of farnesyl or geranylgeranyl moieties to proteins are now well characterized (9,10). Farnesyltransferase transfers farnesyl from farnesylpyrophosphate to Ras (11), whereas geranylgeranyltransferase I attaches the lipid geranylgeranyl from geranylgeranylpyrophosphate to Rho proteins (12).
It is well known that activation of HMG-CoA 1 reductase is essential for G 1 /S progression in various growth-stimulated cells (13,14); thus, the contributions of farnesylated and/or geranylgeranylated proteins to cell cycle control have been intensively investigated. Recent experiments using specific inhibitors of geranylgeranyltransferase I and farnesyltransferase indicate that protein geranylgeranylation, not farnesylation, is required for the G 1 to S phase transition of the cell cycle in mouse fibroblasts (15). Cyclin-dependent kinase (Cdk) enzymes are activated for entry into the S phase of the cell cycle (16). Elimination of Cdk inhibitor protein p27 Kipl during the G 1 to S phase is required for the activation process (17). We demonstrated that geranylgeranylation of Rho proteins is essential for the degradation of p27 Kip1 and facilitates the progression from G 1 to S phase in growth-stimulated rat FRTL-5 cells (18). These data indicate a pivotal role of protein geranylgeranylation, not farnesylation, in G 1 /S transition, although the detailed mechanism is still unclear.
This investigation was performed to gain further insights into the requirement of geranylgeranylpyrophosphate for G 1 /S transition in growth-stimulated FRTL-5 cells. We studied the following questions: When does the isoprenylation and membrane association of Ras and Rho small GTPases occur during G 1 /S phase progression in FRTL-5 cells? What correlation exists between growth-stimulated activation of HMG-CoA reductase and membrane association of Ras and Rho small GTPases? 8305), a strain of rat thyroid cells in continuous culture, were grown in Coon's modified Ham's F-12 medium supplemented with 5% calf serum and a six-hormone mixture consisting of thyrotropin (1 ϫ 10 Ϫ10 M), insulin (10 g/ml), hydrocortisone (0.4 ng/ml), human transferrin (5 g/ml), glycyl-L-histidyl-L-lysine acetate (10 ng/ml), and somatostatin (10 ng/ml) (19). This medium is referred to as 6H medium. For all experiments, cells were initially cultivated in 6H medium for at least 3 days. As appropriate to individual experiments, cells were then shifted to medium containing no thyrotropin, no insulin, and only 0.2% calf serum, which is referred to as 4H medium, for at least 5 days before use in individual experiments. Fluorescence-activated cell sorter analysis revealed that the percentage of cells in G 0 /G 1 at that point is over 95% (18). These cells are referred to as quiescent cells; the quiescent cells rechallenged with 6H medium are referred to as growth-stimulated cells.
Whole cell lysates were prepared, Cdk2 activity was measured, and [ 3 H]thymidine incorporation into DNA was determined as previously reported (19). HMG-CoA reductase activity was determined as reported by Grieco et al. (14). The cell cycle profiles of samples were analyzed, and crude membrane-or cytosol-containing fractions were prepared as previously reported (18,19). Pravastatin an HMG-CoA reductase inhibitor (20) was kindly provided by Dr. S. Kurakata (Sankyo Pharmaceutical Co., Ltd. Tokyo, Japan).
Growth-stimulated FRTL-5 cells that had been treated with pravastatin and incubated with GGPP liposomes (10 M) were pulse labeled with [ 35 S]methionine and [ 35 S]cysteine for a 6-h period, starting 18 h after the addition of 6H medium and pravastatin and 2 h after adding the GGPP-containing liposomes after the 6-h labeling period. Cells were washed and the incubation continued in unlabeled 6H medium plus pravastatin and GGPP liposomes. The membrane fraction was recovered at 30 h and solubilized. Rho A was prepared by immunoprecipitation as described previously (18); 35 S incorporated into Rho A was measured by fluorography after gel electrophoresis. In a separate experiment, growth-stimulated cells treated with pravastatin after the addition of 6H medium were incubated with liposomes containing [ 3 H]GGPP (18.5 MBq/mol) starting at 16 h, and membranes were recovered and processed in the same way to detect labeled geranylgeranyl in the Rho A.
Prenyl group identification of Rho A was performed by the method of Farrel et al. (21). In brief, labeled Rho A in the membrane fraction was delipidated, and the prenyl groups were removed by methyl iodide cleavage. The labeled lipids were separated by C18 reverse-phase HPLC as reported by Farrel et al. (21).
Preparation of Human Rho A cDNA-Full-length human Rho A cDNA encoding nucleotides 1-613, the entire coding region as reported by Fagan et al., 2 was obtained using polymerase chain reaction, using the following oligonucleotides as primers, 5Ј-CGTTGCCTGAGCAATG-GCTGCCA-3Ј and 5Ј-GCAAGGTTTCACAAGACAAGGC-3Ј, and using cDNA prepared from human kidney poly(A) ϩ RNA using a cDNA synthesis kit. Northern analysis of Rho A was performed as described previously (24).

Time Course of HMG-CoA Reductase Activation and DNA Synthesis in Growth-stimulated FRTL-5 Cells-Challenging
quiescent FRTL-5 cells with thyrotropin, insulin, and 5% calf serum resulted in the increased of HMG-CoA reductase activity and mevalonate synthesis after about 10 h (Fig. 1A). Activity reached a maximum at or near the mid-G 1 phase of the cell cycle (22 h) and returned to prestimulation levels when these cells entered the S phase (after 30 h) (Fig. 1A). Thymidine incorporation into DNA increased after 20 h and was maximal at 36 h. Thus, growth stimulation-induced activation of HMG-CoA reductase preceded the onset of DNA synthesis.
Time Course Study for p27 Elimination and Cdk2 Activation in Growth-stimulated FRTL-5 Cells-Significant elimination of p27 begins 24 h after exposure to growth stimulation, proceeds at a relatively slow rate for the next 6 h, and exponentially decreases thereafter until disappearance after 36 h (Fig. 1B). Immunoblotting reveals a coincident increase of the rapidly migrating, phosphorylated form of Cdk2 as a single major 33-kDa band (Fig. 1B), when p27 decreases toward nonmeasurable levels. The kinase activity associated with Cdk2 was increased at 24 h and maximal at 36 h (Fig. 1B). Thus, growth stimulation-induced activation of HMG-CoA reductase coincides with the onset of p27 elimination and Cdk2 activation.
Time Course Study for Membrane Association of Ras and Rho A during G 1 Progression-Immunoblot analysis of membrane fractions for Ras and Rho A reveals that Rho A begins to disappear from the membrane 12 h after exposure to the Quiescent FRTL-5 cells were incubated with 6H medium for the indicated times. In A, HMG-CoA reductase activity assays were performed as described under "Experimental Procedures," and thymidine incorporation into DNA was measured during the last 2 h of each noted incubation period. Data in both cases are the means Ϯ S.D. of three independent experiments. To measure Cdk2 activation and the elimination of p27, cells were harvested by trypsinization. Cell lysates (30 g) were analyzed by immunoblotting with antibodies against Cdk2 and p27 as described under "Experimental Procedures." In C, membrane-containing fractions were prepared at the indicated times ("Experimental Procedures"), and 30 g of each sample was analyzed by immunoblotting with antibodies against Ras and Rho A. growth stimulus (Fig. 1C). Rho A in membranes is maximally diminished near the mid-G 1 phase of the cell cycle (18 h) but returns to the membranes in late G 1 18 -24 h after the growth stimulus started and is at near prestimulation levels by 36 h after stimulation (Fig. 1C). In contrast, the level of Ras in membrane fraction does not change throughout cell cycle progression (Fig. 1C). Thus, growth stimulation-induced elimination of Rho A from membranes is coincident with and precedes the onset of HMG-CoA reductase activation; reappearance of Rho A in membrane fractions is coincident with the elimination of p27 and Cdk2 activation Requirement of GGPP, not FPP, for DNA Synthesis and G 1 /S Transition-Pravastatin (1200 M) inhibited growth-stimulated DNA synthesis and induced G 1 arrest in FRTL-5 cells ( Table I). All of the effects of pravastatin on DNA synthesis and cell cycle progression were wholly reversed by the addition of liposomes containing 10 M GGPP, not FPP, 16 h after growth stimulation (Table I). The presence of GGPP during the first 16 h of culture does not reverse the inhibitory effect of pravastatin (Table I). GGPP is biosynthetically derived from the single condensation of FPP and isopentenylpyrophosphate. Because isopentenylpyrophosphate could not be synthesized in pravastatin-treated cells, FPP could not be converted to GGPP. These data indicate that GGPP can rescue the pravastatininduced G 1 arrest in the absence of upstream intermediates of cholesterol biosynthesis and that GGPP, not FPP, is required for G 1 /S transition in FRTL-5 cells.
Effect of Pravastatin and Isoprenoids on Growth Stimulation-induced Elimination and Reappearance of Rho A in Membrane Fraction-Pravastatin had no effect on the growth stimulation-induced elimination of Rho A from membranes; in contrast, the reappearance of Rho A in membranes was completely blocked by the presence of pravastatin (Fig. 2B, row 6HP versus row 6H); Rho A was detected only in the cytosolic fraction (Fig. 2C, second row). The inhibitory effect of pravastatin on the reappearance of Rho A in membranes was wholly reversed by the addition of 10 M GGPP, not FPP, when GGPP was added 16 h after stimulation (Fig. 2B, rows 6HPG and 6HPF, respectively). As was the case for growth (Table I), the presence of GGPP during the first 16 h of culture did not reverse the inhibitory effect of pravastatin on Rho A reappearance in the membrane (data not shown). The presence of Ras in membrane fraction did not change through cell cycle progression in the presence of pravastatin, nor did GGPP and FPP affect the level of Ras in membrane fraction (Fig. 3A). These results indicate that GGPP is required for the translocation of Rho A from the cytoplasm to membranes during the progression from G 1 to S phase in FRTL-5 cells but has no effect on Ras.

Time Course Study for the Expression of Rho A mRNA in
Growth-stimulated FRTL-5 Cells-Northern analysis showed that Rho A mRNA was present as a doublet as reported previously (25) and that Rho A mRNA levels did not change during G 1 progression (Fig. 3A). This indicates that the growth stimulation of FRTL-5 cells had no effect on Rho A gene expression.

Pulse Labeling of Rho A with [ 35 S]Methionine and [ 3 H] GGPP-Immunoprecipitation and fluorography of Rho A in the membrane fraction of cells pulsed with [ 35 S]methionine and [ 35 S]cysteine between 18 and 24 h after growth stimulation in
the presence of pravastatin and 10 M GGPP revealed that the Rho A present in the membrane fraction at 30 h was radiolabeled, i.e. was newly synthesized Rho A (Fig. 3B). In a separate experiment wherein the protein was labeled by supplementation with [ 3 H]GGPP, Rho A was radiolabeled when recovered from the membrane at 30 h (Fig. 3B). Rho A was not radiolabeled when 10 M cold GGPP was added to [ 3 H]GGPP (data not shown). HPLC analysis of the lipids associated with the protein showed that the protein on the membrane in this experiment was geranylgeranylated (Fig. 3C). This indicates that newly synthesized Rho A was geranylgeranylated and translocated to membranes in the presence of GGPP. DISCUSSION Ras and Rho small GTPases have each been shown to play a pivotal role in the regulation of cell cycle progression (1, 3).  2. Effect of pravastatin and isoprenoids on growth stimulation-induced elimination and reappearance of Rho A or Ras in the membrane and cytosolic fractions. Quiescent FRTL-5 cells were incubated in 6H medium for 36 h in the absence (6H) or the presence (6HP) of 1200 M pravastatin. 16 h after the shift to 6H growth medium, liposome containing FPP (6HPF) or GGPP (6HPG) were added to the medium of cells treated with pravastatin. Crude membranecontaining fractions were prepared at the times noted, as described under "Experimental Procedures." Each sample (30 g) was analyzed by immunoblotting with antibodies against Ras and Rho A.
Growth stimulation induces expression of the HMG-CoA reductase activity in various cells (13,14), thereby allowing Ras and Rho A to be activated after isoprenylation with FPP or GGPP, which are derived from mevalonate (1,8,7). To see which of these small GTPases were playing a critical role in cell cycle progression during the growth of rat FRTL-5 cells, we compared time-dependent changes of Ras and Rho A in the membrane fraction with time dependent changes in HMG-CoA reductase activity, DNA synthesis, and the cell cycle regulating molecules, p27 and Cdk2, which control G 1 /S transition in growth-stimulated FRTL-5 cells.
In this study, both Ras and Rho A were detected in membrane fraction at G 0 . Growth stimulation of FRTL-5 cells in-duced the elimination of Rho A from membranes at mid-G 1 , but the level of Rho A in membranes returned to prestimulation levels at late G 1 . In contrast, Ras remained in membrane fraction throughout the cell cycle progression from G 1 to S phase. The elimination of Rho A from membranes precedes the activation of HMG-CoA reductase and is already significant at 12 h when the reductase activation is just starting. Thus, we demonstrate that growth stimulation of FRTL-5 cells causes an elimination of Rho A from membranes, not Ras, in early G 1 and that Rho A re-associates with the membrane in association with the activation of HMG-CoA reductase at mid-G 1 . Ras does not show such changes.
Pravastatin inhibition of HMG-CoA reductase activity in growth-stimulated FRTL-5 cells blocks DNA synthesis and induces G 1 arrest. The mevalonate metabolite, FPP, does not reverse this inhibitory effect of pravastatin, despite the fact that FPP is incorporated into cells as well as GGPP, as measured in cells exposed to liposomes containing either [ 3 H]FPP or [ 3 H]GGPP (18). This demonstrates that FPP is not required for G 1 /S transition of FRTL-5 cells. Vogt et al. also reported that protein farnesylation is not required for the G 1 to S phase transition of the cell cycle in mouse fibroblasts (15).
In this study, Ras was detected in the membrane fraction at G 0 , and the level of Ras in the membrane fraction did not change during G 1 /S progression. Furthermore, Ras in membranes was not affected by the treatment with pravastatin and supplementation with FPP. These results suggest that the farnesylation of Ras and its translocation into membranes does not occur during G 1 /S progression in FRTL-5 cells. These experiments may therefore explain why FPP is not required for G 1 /S transition in FRTL-5 cells; however, it remains unclear when Ras is farnesylated and translocated to the membrane during cell cycle progression.
In contrast to Ras, Rho A diminishes in the membrane fraction at the near mid-G 1 phase of the cell cycle (18 h). It is not clear whether Rho A lost from membranes during this period is further processed or not; the degradation pathway of Rho A has not yet been determined in detail. We have, however, examined the effects of several protease inhibitors on Rho A elimination in growth-stimulated FRTL-5 cells. MG-115, an inhibitor of the ubiquitin-proteasome pathway, had no effect on growth-stimulated elimination of Rho A from membranes (data not shown). It therefore seems unlikely that the ubiquitin-proteasome pathway is involved in growth-stimulated elimination of Rho A from membranes in FRTL-5 cells.
After its elimination from membranes at the near mid-G 1 phase of the cell cycle, the level of Rho A in the membranes returns to prestimulation levels at late G 1 . Pravastatin completely blocked the reappearance of Rho A in the membrane fraction at late G 1 , and Rho A was detected only in the cytosolic fraction. This inhibitory effect of pravastatin on the reappearance of Rho A in the membrane was completely reversed by GGPP, when GGPP was added 16 h after stimulation (at the onset of HMG-CoA reductase activation). On the other hand, the presence of GGPP during the first 16 h of culture did not reverse the inhibitory effect of pravastatin; the requirement of GGPP for G 1 /S transition and DNA synthesis had the same time course dependence. Thus, the inhibitory effect of pravastatin on G 1 /S transition and DNA synthesis was not reversed by GGPP during the first 16 h of culture but rather by its addition and continued presence beyond 16 h. These data indicate that growth-stimulated activation of HMG-CoA reductase is closely linked to the translocation of Rho A to membranes, not Ras, and to G 1 /S transition in FRTL-5 cells.
The membrane association of prenylated proteins has been reported to be cyclical (26). Such cycling of prenylated proteins FIG. 3. Analysis of Rho A mRNA levels and metabolism in growth-stimulated FRTL-5 cells. In A, quiescent FRTL-5 cells were incubated with 6H medium for the indicated times. Northern analysis was performed as described under "Experimental Procedures." In B, the growth-stimulated FRTL-5 cells treated with pravastatin and incubated with GGPP liposomes were pulse labeled with [ 35 S]methionine and [ 35 S]cysteine. In a separate experiment, cells treated with pravastatin were incubated with liposomes containing [ 3 H]GGPP. Rho A was recovered from solubilized membrane fraction by immunoprecipitation. Fluorography of radiolabeled Rho A was performed as described under "Experimental Procedures." In C, labeled Rho A, which was recovered from membrane fraction by immunoprecipitation, was delipidated. The prenyl groups were removed by methyl iodide cleavage, and the labeled lipids were separated by C18 reverse-phase HPLC as described under "Experimental Procedures." The arrows indicate elution times of geraniol (GO), farnesol (FO), and geranylgeraniol (GGO), respectively. could be caused by proteins such as the guanine nucleotide dissociation inhibitor (GDI) that appears to extract prenylated small GTPases out of the membrane and into the cytoplasm (27,28). Rho GDI makes a complex with the GDP-binding form of the Rho proteins and thereby inhibits the dissociation of GDP from and the subsequent binding of GTP to the Rho gene products (28). In addition to this regulatory function of Rho GDI, Rho GDI regulates the binding of Rho A in the membranes (29). It remains unclear whether GDI might be involved in growth-stimulated elimination of Rho A from membranes in FRTL-5 cells.
In this report, we performed pulse label studies using [ 35 S]methionine and [ 35 S]cysteine in the presence of pravastatin and in cells supplemented with 10 M GGPP. Pulse label experiments revealed that newly synthesized Rho A was translocated to membranes in the presence of GGPP. A separate labeling experiment using [ 3 H]GGPP-containing liposomes showed that the protein was also labeled with [ 3 H]GGPP; and HPLC analysis of the lipids associated with the protein showed that the protein was geranylgeranylated (data not shown). These results indicate that GGPP is required for geranylgeranylation of newly synthesized Rho A and its translocation to membranes. When pulse labeling experiments were performed, wherein the labeling preceded the initiation of the growth cycle, we could not detect significant mature Rho A (i.e. old geranylgeranylated Rho A) recycling to the membrane. Nevertheless, it cannot be concluded that some mature Rho A does recycle and that membrane association of such recycled mature Rho A is regulated by a process that requires GGPP through another mechanism. We can conclude that newly synthesized Rho A appears on the membrane but not that only newly synthesized Rho A appears on the membrane.
The possibility of re-geranylgeranylation of Rho A extracted from membranes seems to be unlikely. Because geranylgeranylation by geranylgeranyltransferase I has been demonstrated to require the carboxyl-terminal CAAX box of Rho proteins, (C, cysteine, A, aliphatic amino acid; X, leucine) (6). After prenylation by geranylgeranyltransferase I, the three COOH-terminal amino acids (the AAX residues) are removed from Rho A (30,31). Subsequent to proteolysis, the prenylated cysteine residue is carboxymethylated (23).
This study shows that growth stimulation eliminates Rho A from membranes just before the activation of HMG-CoA reductase in mid-G 1 phase and that the reappearance of the protein in membranes in late G 1 phase is associated with p27 elimination and Cdk2 activation. GGPP promotes the geranylgeranylation and the translocation of newly synthesized Rho A to membranes. In contrast, Ras in membrane fraction dose not change through cell cycle progression from G 1 to S phase. The difference in processing of these proteins during G 1 /S progression might explain the reason why GGPP, not FPP, is required for G 1 /S transition in FRTL-5 cells. The mechanisms underly-ing Rho A processing during G 1 /S progression, particularly the elimination of Rho A from membranes, require further investigation.