Post-translational modification of H-Ras is required for activation of, but not for association with, B-Raf.

B-Raf is regulated by Ras protein and acts as a mitogen-activated protein (MAP) kinase kinase kinase in PC12 cells and brain. Ras protein undergoes a series of post-translational modifications on its C-terminal CAAX motif, and the modifications are critical for its function. To elucidate the role of the post-translational modifications in interaction with, and activation of, B-Raf, we have analyzed a direct association between H-Ras and B-Raf, and constructed an in vitro system for B-Raf activation by H-Ras. By using methods based on inhibition of yeast adenylyl cyclase or RasGAP activity and by in vitro binding assays, we have shown that the segment of B-Raf corresponding to amino acid 1-326 binds directly to H-Ras with a dissociation constant (Kd) comparable to that of Raf-1 and that the binding is not significantly affected by the post-translational modifications. However, when the activity of B-Raf to stimulate MAP kinase was measured by using a cell-free system derived from rat brain cytosol, we observed that the unmodified form of H-Ras possesses an almost negligible activity to activate B-Raf in vitro compared to the fully modified form. H-RasSer-181,184 mutant, which was farnesylated but not palmitoylated, was equally active as the fully modified form. These results indicate that the post-translational modifications, especially farnesylation, are required for H-Ras to activate B-Raf even though they have no apparent effect on the binding properties of H-Ras to B-Raf.

Ras protein is a plasma membrane-associated guanine nucleotide-binding protein that cycles between a GTP-bound active form and a GDP-bound inactive form, and operates in key processes of intracellular signal transduction systems that are involved in regulation of cell growth and differentiation. In higher eukaryotes including Caenorhabditis elegans, Drosophila melanogaster, and vertebrates, Ras is a key regulator that mediates signal transduction from cell surface tyrosine kinase receptors to the nucleus via activation of the MAP 1 kinase cascade (for reviews, see Refs. 1 and 2). Recent studies demonstrated that Ras makes a direct association with a serine/ threonine kinase Raf-1, a product of the c-raf-1 proto-oncogene (3)(4)(5)(6)(7)(8) and that this association leads to stimulation of the activity of Raf-1 to phosphorylate MAP kinase kinase (MEK) (for reviews, see Refs. 1 and 2). However, the precise mechanism of the Raf activation by active form of Ras remains to be clarified.
B-raf gene was discovered as a transforming gene in NIH3T3 cell transfection assays with human Ewing sarcoma DNA (9), and its protein product consists of 765 amino acid residues that contain three distinct regions of conservation with Raf-1; CR1, CR2, and CR3 (2,10). In contrast to the ubiquitous distribution of Raf-1 in a variety of mammalian organs, expression of B-Raf is confined to brain and testis (11). Another member of the Raf family, A-Raf, is expressed most abundantly in ovary and epididymis (11). Recent studies have shown that B-Raf, instead of Raf-1, is responsible for Ras-dependent activation of the MAP kinase pathway in PC12 cells and mammalian (rat and bovine) brain (12)(13)(14)(15).
Ras proteins undergo a series of post-translational modifications on their unique C-terminal region called a CAAX motif (C, cysteine; A, aliphatic; and X, any amino acid) (for reviews, see Refs. 16 -18). The first stage of the processing consists of three successive modifications of the CAAX motif: (i) farnesylation of the cysteine residue, (ii) proteolytic cleavage of the amino acids AAX, and (iii) methyl esterification of the new C-terminal cysteine. This first stage of modification converts the primary translation product into an intermediate form. In the case of H-Ras, it is further modified by acylation with palmitic acid on cysteine residues (Cys-181 and Cys-184) immediately upstream of the CAAX motif, finally yielding the post-translationally fully modified form. These modifications are essential for anchoring Ras proteins to the plasma membrane (19,20) and for a number of biological activities of Ras: malignant transformation of NIH3T3 cells (19,20), induction of neuronal differentiation of PC12 cells (21), and induction of germinal vesicle breakdown in Xenopus laevis oocytes (22) by activated Ras. The activity of H-Ras to activate Raf-1 in vivo was also reported to be dependent on the modifications (23). However, these in vivo experiments entail an inherent problem in separating the effect of the modifications on the activity of Ras from that on its membrane anchoring. Recently an in vitro pure reconstituted system was used to show that the post-translational modifications, especially farnesylation, are critical for activation of Saccharomyces cerevisiae adenylyl cyclase which is an immediate downstream effector of Ras in this organism (24). This suggested that the post-translational modifications are required for activation of Ras effectors. Efficient activation of MAP kinases by Ras in crude cell-free extracts from X. laevis oocytes was also reported to depend on the modifications (25,26). However, requirement of the modifications of Ras has not been examined in vitro for the Raf-1 activation because a cell-free system for the Raf-1 activation has not been established.
To analyze the molecular mechanism underlying the requirement of the post-translational modifications for the Raf activation, we have established a cell-free system derived from rat brain cytosol in which exogenously added H-Ras protein can activate MAP kinase/ERK2 through activation of B-Raf and MEK. We have also examined the effect of the modifications on direct association of H-Ras with B-Raf, and the result is compared with that obtained on the B-Raf activation.

EXPERIMENTAL PROCEDURES
Materials-The post-translationally modified and unmodified forms of human H-Ras protein were purified from the membrane and cytosol fractions of Sf9 cells, respectively, which had been infected with a recombinant baculovirus carrying the human H-ras cDNA, as described previously (24). Two H-ras mutants encoding the proteins defective in the post-translational modifications, H-ras Ser-181,184 and H-ras Ser-186 , were constructed by site-specific mutagenesis using appropriate mutagenic oligonucleotides as described elsewhere (24). The effector mutant H-ras Asn-38 cDNA was provided by Dr. S. Yokoyama (University of Tokyo, Tokyo, Japan). Recombinant baculoviruses carrying the mutant H-ras genes were prepared as described before (24). The farnesylated but not palmitoylated H-Ras Ser-181,184 and the fully modified H-Ras Asn-38 proteins were solubilized and purified from the membrane fractions of Sf9 cells infected with recombinant baculoviruses carrying the respective H-ras genes similarly as described before (24). The unprocessed H-Ras Ser-186 protein was purified from the cytosol fraction of Sf9 cells infected with the corresponding baculovirus. A plasmid carrying the full-length human B-raf cDNA, pSL-Braf, was provided by Dr. W. Kolch (Institute for Clinical Molecular Biology and Tumor Genetics, Germany). The HindIII-ApaI and HindIII-XhoI fragments of the B-raf cDNA corresponding to amino acid 1-326 and 1-445, respectively, were cloned into the matching cleavage sites of pMAL-cRI (New England Biolabs Inc.) to produce pMAL-B-Raf(1-326) and pMAL-B-Raf(1-445), respectively. The MBP-B-Raf fusion proteins were purified from Escherichia coli harboring the corresponding pMAL-B-Raf plasmids by affinity chromatography on amylose resin (27). A segment of the human c-raf-1 cDNA corresponding to amino acid 1-206 of Raf-1 was amplified by a polymerase chain reaction with suitable primers and cloned into pMAL-cRI for expression as an MBP-fusion protein in E. coli. Plasmids for expressing GST-fusion protein of MEK (GST-MEK) or GST-fusion protein of a kinase negative mutant of ERK2 (GST-KNERK) in E. coli were obtained from Dr. A. Kikuchi (Hiroshima University, Hiroshima, Japan). A rat cDNA encoding the full-length ERK2 was cloned from a rat brain cDNA library using a polymerase chain reaction with suitable primers and cloned into pGEX-2T (Pharmacia Biotech Inc.) for production of GST-fusion protein of ERK2 (GST-ERK2) in E. coli. GST-MEK, GST-ERK2, and GST-KNERK proteins were purified by glutathioneagarose chromatography as described elsewhere (28). Purified recombinant RasGAP p120 produced in E. coli was provided by Dr. S. Hattori (National Center of Neurology and Psychiatry, Tokyo, Japan). The anti-H-Ras monoclonal antibody F235 was purchased from Oncogene Science Inc. (Manhasset, NY). The polyclonal antibody raised against the C terminus of B-Raf (ASPKTPIQAGGYGAFPVH) and that raised against the C terminus of Raf-1 (CTLTTSPRLPVF) were purchased from Santa Cruz Biotech. Inc. (Santa Cruz, CA). Myelin basic protein was purchased from Sigma.
Adenylyl Cyclase Assay-Yeast adenylyl cyclase, overproduced in yeast TK35-1 harboring the plasmid YEP24-ADC1-CYR1, was solubilized from the crude membrane fraction with buffer C (50 mM MES/ NaOH, pH 6.2, 0.1 mM MgCl 2 , 0.1 mM EGTA, 1 mM PMSF, and 1 mM 2-mercaptoethanol) containing 1.5 M NaCl as described previously (29). The supernatant (15 g of protein) after centrifugation at 100,000 ϫ g for 1 h was used for each adenylyl cyclase assay. Measurements of adenylyl cyclase activity dependent on the GTP␥S-bound form of H-Ras and its inhibition by the purified MBP-B-Raf polypeptides were carried out as described previously (29,30).
Assay for GTPase Activity-The post-translationally modified and unmodified forms of H-Ras (2 pmol) were first loaded with [␥-32 P]GTP by incubation in a mixture containing 20 mM Tris/HCl, pH 7.4, 5 mM MgCl 2 , 10 mM EDTA, 1 mM DTT, and 1.7 M [␥-32 P]GTP (3,000 cpm/ pmol) at 30°C for 25 min, followed by addition of MgCl 2 to 15 mM. The GTPase reactions were initiated by addition of a fixed amount of the [␥-32 P]GTP-bound H-Ras into a reaction mixture (100 l) containing 20 mM Tris/HCl, pH 7.4, 10 mM MgCl 2 , 5 mM EDTA, 1 mM DTT, and 1 mM GTP with or without 400 pM recombinant RasGAP p120. After incubation at 25°C for various periods, the reaction was stopped by rapid filtration through a nitrocellulose filter, and the GTP-hydrolyzing activity was examined by measuring the residual amount of [␥-32 P]GTP retained on the filter as described previously (30). For measuring the inhibition of GAP activity, similar incubations were carried out for 5 min in the presence of various amounts of MBP-B-Raf(1-326), and the amounts of [␥-32 P]GTP remaining bound to H-Ras were measured as described above and compared with those without the MBP-B-Raf polypeptide.
In Preparation and Fractionation of the Rat Brain Cytosol-All manipulations were carried out at 0 -4°C. A rat (Wister, 300 g of body weight, the Center of Japan Biological Chemistry Co. Ltd.) brain was homogenized in 7 ml of buffer A (20 mM Tris/HCl, pH 8.0, 2 mM EDTA, 1 mM DTT, 2.5 mM MgCl 2 , 0.3 M sucrose, 10 g/ml leupeptin, 20 g/ml aprotinin, and 1 mM PMSF) with a Potter-Elvehjem homogenizer. The homogenate was centrifuged at 100,000 ϫ g for 1 h, and the resulting supernatant (20 mg of protein) was applied to a Mono S HR5/5 column (0.5 ϫ 5 cm) (Pharmacia) equilibrated with buffer B (20 mM HEPES/ NaOH, pH 6.8, 2 mM EDTA, 1 mM DTT, 2.5 mM MgCl 2 , 10 g/ml leupeptin, and 10 M (p-amidinophenyl)methanesulfonyl fluoride). A linear gradient elution was performed between 15 ml each of 0 and 1 M NaCl in buffer B, and 1-ml fractions were collected.
Assay of Ras-dependent MAP Kinase Stimulation Activity-Ras-dependent MAP kinase stimulation activity was measured by phosphorylation of myelin basic protein in the presence of GTP␥S-bound or GDP-bound H-Ras. The sample (15 l) to be tested was preincubated for 10 min at 30°C in a final volume of 50 l containing 20 mM Tris/HCl, pH 8.0, 6 mM HEPES/NaOH, pH 6.8, 1.5 mM EGTA, 1 mM EDTA, 1 mM DTT, 5 mM MgCl 2 , 120 M ATP, 80 nM GST-MEK, and varying amounts of GTP␥Sor GDP-bound H-Ras, followed by addition of 10 l of 3 M recombinant GST-ERK2 and another incubation for 10 min. The final phosphorylation reaction was initiated by addition of 20 l of reaction mixture containing 20 mM Tris/HCl, pH 8.0, 1.5 mM EGTA, 1 mM EDTA, 1 mM DTT, 5 mM MgCl 2 , 50 M [␥-32 P]ATP (1,000 cpm/pmol), and 50 M myelin basic protein. After 20-min incubation at 30°C, a 30-l aliquot of the reaction mixture was spotted onto a piece of phosphocellulose paper (Whatman P81). After washing with 75 mM phosphoric acid, the radioactivity on the paper was measured by liquid scintillation spectrometry. To measure Ras-dependent phosphorylation of ERK2, the assay was performed similarly except that the reaction mixture contained purified GST-KNERK in place of GST-ERK2 and myelin basic protein and that GST-KNERK was absorbed onto glutathione-agarose resin after the final phosphorylation reaction. The resin was briefly washed with 20 mM Tris/HCl, pH 8.0, 1 mM EDTA, 1 mM DTT, and 5 mM MgCl 2 , and subjected to SDS-PAGE (10% gel) and autoradiography.
Immunodepletion Study-A 150-l aliquot of the active fraction (fraction 42) of the Mono S column chromatography was incubated with continuous mixing for 1 h at 4°C with 30 l of protein A-Sepharose resin alone or of the resin to which 1.5 g of either the anti-Raf-1 polyclonal antibody or the anti-B-Raf polyclonal antibody was attached. After a brief centrifugation, 15 l of the supernatant were assayed for the Ras-dependent MAP kinase stimulation activity as described above.

Measurement of Direct Association between B-Raf and H-Ras by Adenylyl Cyclase Inhibition
Assay-Although a previous study showed that B-Raf forms a complex with Ras and exists in an activated state in the complex (14), their direct association has not been rigorously demonstrated. We examined whether B-Raf could bind directly to H-Ras protein and thereby compete for it with yeast adenylyl cyclase in vitro as observed for Raf-1 and Schizosaccharomyces pombe Byr2 proteins (29,30). Because production of the full-length B-Raf protein in E. coli turned out to be impossible, we chose to express its Nterminal segments of 326 and 445 amino acid residues as MBP-fusion proteins. B-Raf(1-326) contained CR1 only, while B-Raf(1-445) encompassed both CR1 and CR2. Because the Ras-binding activity of Raf-1 had been shown to reside solely in the CR1 (6, 31), we expected that the N-terminal segments represent the Ras-binding activity of B-Raf. As shown in  (Fig. 1B). At each point of H-Ras concentration in the presence of the competitor, we obtained free H-Ras concentration available for adenylyl cyclase activation as that required for giving the same adenylyl cyclase activity in the absence of the competitor. A difference between the original and the free concentrations of H-Ras was regarded as that bound to the competitor, and a reciprocal of this value was plotted against a reciprocal of the free H-Ras concentration (Fig. 1C). This gave a series of straight lines for each value of the competitor, which converged on the horizontal axis. The data indicated that the competitor polypeptide bound directly to H-Ras protein and competitively sequestered it from adenylyl cyclase. The K d values of B-Raf(1-326) and B-Raf(1-445) for H-Ras were calculated from the points of intersection with the horizontal axis and determined to be about 7 nM (data not shown) and 5 nM (Fig. 1C), respectively. These values are comparable to those of Raf-1 for H-Ras, 3.5 nM, of yeast adenylyl cyclase for yeast Ras2, 7 nM, or of Byr2 for Ras2, 1 nM (see Ref. 29 for the data and a detailed description on the kinetic analysis).

Measurement of B-Raf Association with H-Ras by
Inhibition of RasGAP Activity-The affinity of B-Raf could not be determined for the post-translationally unmodified form of H-Ras by the adenylyl cyclase inhibition assay because the unmodified form of H-Ras possesses a negligible activity to stimulate yeast adenylyl cyclase (32). Therefore, we used a method based on inhibition of the GTPase-stimulating activity of RasGAP p120, whose interaction with H-Ras had been shown to be unaffected by the modifications (32). We purified the unmodified form of H-Ras from the cytosol fraction of Sf9 cells infected with the baculovirus carrying the H-ras cDNA. The purified protein migrated slower than the fully modified form upon SDS-PAGE (see Fig. 3A). Previous studies employing [ 3 H]mevalonate labeling confirmed that H-Ras purified from the cytosol fraction of Sf9 cells was not farnesylated (33). We also characterized the proteins by reversed phase chromatography, and found that H-Ras from the cytosol fraction eluted earlier than the modified form at the position expected for the unmodified form from the previous report (34) (data not shown). In the absence of Ras-GAP p120, [␥-32 P]GTP bound to Ras decreased at the same rate in the unmodified and modified forms of H-Ras. Also, no difference was observed in GTP-hydrolyzing activity stimulated by RasGAP p120 (Fig. 2A). H-Ras interacts directly with Ras-GAP p120 at its effector region (35), and the interaction is subject to competitive inhibition by the Ras effectors, Raf-1 and Byr2 (4,5,30). We examined whether B-Raf N-terminal polypeptides interfered with the activity of RasGAP to stimulate GTP-hydrolysis of H-Ras. As shown in Fig. 2B, we observed a dose-dependent inhibition of the RasGAP activity by MBP-B-Raf(1-326), and the extent of inhibition was not significantly affected by the post-translational modifications of H-Ras. The result suggested that the B-Raf N-terminal polypeptide has an almost equal activity to bind the unmodified and modified forms of H-Ras and thereby competitively sequester them from RasGAP p120.

In Vitro Binding of B-Raf N Terminus to the Modified and Unmodified Forms of H-Ras-
The post-translationally modified and unmodified forms of H-Ras were loaded with GTP␥S or GDP␤S, and examined for binding to MBP-B-Raf(1-326) and MBP-Raf-1(1-206), which had been immobilized on amylose resin as described under "Experimental Procedures." Both of the MBP-fusion proteins bound efficiently to the GTP␥S form but not to the GDP␤S form of both post-translationally modified and unmodified forms of H-Ras (Fig. 3A). The binding was abolished by a substitution of asparagine for aspartic acid at position 38 in the effector loop of H-Ras. This mutation had been shown to abolish the Raf-1 binding (4). We next examined whether the post-translational modifications affected the binding affinity of H-Ras to B-Raf (Fig. 3B). When increasing concentrations of the modified and unmodified forms of H-Ras were used for binding reactions with MBP-B-Raf(1-326), the amounts of bound H-Ras increased and finally reached a plateau level. No difference in the binding pattern was observed between the modified and unmodified forms (Fig. 3B). From the data, we estimated an apparent K d value of B-Raf to both forms of H-Ras as the concentration of H-Ras exhibiting a half maximal binding, and obtained a value of around 50 nM. This is a bit higher than that obtained by the adenylyl cyclase inhibition assay. These results indicated that H-Ras binds to the B-Raf N terminus in a GTP-dependent manner presumably at its effector domain and that the post-translational modifications of H-Ras have no effect on the binding.
Establishment of a Cell-free System for Ras-dependent B-Raf Activation-It was shown that activation of the MAP kinase cascade by Ras protein is mediated by B-Raf, not by Raf-1, in a crude extract from rat brain (14). We established here a cellfree system derived from rat brain cytosol in which exogenously added H-Ras could activate protein kinases MEK and ERK2 in vitro. The cytosolic extract (20 mg of protein) was applied to a Mono S HR5/5 column equilibrated with buffer B, and washed with 20 ml of the same buffer. Subsequently a linear gradient elution was performed between 0 and 1 M NaCl in a total volume of 30 ml. One-ml fractions were collected and assayed for stimulation of the activity of MAP kinase to phosphorylate myelin basic protein in the presence and absence of the GTP␥Sbound H-Ras as described under "Experimental Procedures." The MAP kinase activity stimulated by H-Ras was eluted as a single peak around fractions 41-44 (Fig. 4A). The peak fraction 42 was used for subsequent studies. The Ras-dependent activation was not observed by either the GDP-bound H-Ras or the GTP␥S-bound H-Ras Asn-38 (Fig. 4B). The Ras-dependent phosphorylation of myelin basic protein also depended on the inclusion of recombinant GST-ERK2 in the reaction mixture (Fig.  4B), indicating that H-Ras induced the enhanced phosphorylation of myelin basic protein through the activation of ERK2. When GST-fusion protein of a kinase negative mutant of ERK2, GST-KNERK, was used as a substrate, the same fraction exhibited H-Ras-dependent stimulation of its phosphorylation, and this activity was dependent on the presence of recombinant MEK (Fig. 4C). Both the Ras-dependent phosphorylations of myelin basic protein and GST-KNERK were efficiently inhibited by the addition of MBP-B-Raf(1-326) (Fig. 4, B and C). This may be caused by competitive sequestration of Ras by the B-Raf N-terminal polypeptide from the endogenous Ras effector molecule. B-Raf, the major Raf family protein in brain, was detected in the peak fraction of the Rasdependent MAP kinase stimulation activity when examined by immunoblotting with the anti-B-Raf antibody (Fig. 4D). Next, we performed an immunodepletion study to examine whether B-Raf is responsible for the Ras-dependent stimulation of MAP kinase activity. The Ras-dependent activity was almost completely depleted by incubation with the anti-B-Raf antibody, but not with the anti-Raf-1 antibody (Fig. 4E). Thus, we concluded that H-Ras stimulated the MEK-MAP kinase pathway To examine which step in the process of modifications of H-Ras is critical for activation of B-Raf, we constructed and purified H-Ras Ser-181,184 , which was farnesylated but lacked the two cysteine residues to be palmitoylated, and H-Ras Ser-186 , which lacked the cysteine residue to be farnesylated and, therefore, was not modified at all. The activities of these mutants to stimulate B-Raf were examined similarly by using the in vitro system. As shown in Fig. 6A, H-Ras Ser-181,184 activated phosphorylation of myelin basic protein as efficiently as the fully modified form of H-Ras at the concentration of 5 nM. In con-trast, H-Ras Ser-186 had an almost negligible activity at the same concentration (Fig. 1A) or even at 20 nM (data not shown). Essentially similar result was obtained when the activities of H-Ras Ser-181,184 and H-Ras Ser-186 to stimulate S. cerevisiae adenylyl cyclase were examined (Fig. 6B). This result is consistent with our previous observation that farnesylation, not palmitoylation, of yeast Ras2 is essential for its ability to activate adenylyl cyclase in vitro (24). These results indicated that the post-translational modifications of H-Ras, especially the farnesylation step, are critical for the activation of B-Raf as well as of yeast adenylyl cyclase. DISCUSSION We have shown here that post-translational modifications of H-Ras are not required for association with one of its effector molecule, B-Raf, but are essential for activation of B-Raf. The farnesylation step of the modifications, not the palmitoylation, is shown to be responsible for this effect. B-Raf is a serine/ threonine kinase which is expressed specifically in neuronal tissues and testis (11), while Raf-1 is ubiquitously expressed in all cell types. Although the association of Raf-1 with Ras has been a subject of extensive investigation, B-Raf is not well analyzed for its interaction with Ras protein. In this paper we have shown for the first time that the N-terminal segment of B-Raf makes a direct association with H-Ras in a GTP-dependent manner. No association is observed with the effector mutant H-Ras Asn-38 , suggesting that B-Raf binds to the effector  1, 2, 6, and 7) and H-Ras Asn-38 ( lanes 5 and 10), and the unmodified form of H- Ras (lanes 3, 4, 8, and 9)  domain of Ras. Further, we have determined the K d value of B-Raf N terminus for the post-translationally modified H-Ras by the yeast adenylyl cyclase inhibition assay. The value is comparable to those of other Ras-effector molecules for their homologous Ras proteins. There exists little difference in the estimated K d values of B-Raf segments containing CR1 only and both CR1 and CR2, suggesting that CR1 contains a major Ras-binding site(s) as observed for Raf-1 (6,31).
It was shown that B-Raf, not Raf-1, mediates the nerve growth factor-induced activation of the MAP kinase cascade through interaction with Ras in PC12 cell (13). In rat brain the association of MEK1 with Ras is dependent on B-Raf, not on Raf-1 (14). These data prompted us to establish a cell-free system for Ras-dependent stimulation of MAP kinase activity through B-Raf activation using rat brain cytosol. While this work was in progress, Yamamori et al. (15) reported establishment of such a system from bovine brain cytosol and purification of a Ras-dependent MEK kinase, which turned out to be a complex of B-Raf and 14 -3-3 proteins, although the possibility of presence of other minor components was not excluded completely. Here we also have constructed a similar in vitro system and employed it to quantitatively examine the effect of the post-translational modifications of Ras on its activity.
A number of studies showed that the post-translational mod- ifications are essential for the biological activities of Ras observed in vivo. However, in these systems, the effect of the modifications on the activity of Ras could not be examined separately from the effect on its membrane anchoring. A recent study using a baculovirus coinfection assay (23), which demonstrated that the modifications of Ras are required for Raf-1 activation, also cannot get rid of this fundamental problem.
Here we have employed the in vitro cell-free system for Rasdependent B-Raf activation to show that the post-translational modifications are critical for the ability of H-Ras to activate B-Raf. To determine which step of modifications is responsible for the acquisition of the ability to activate B-Raf, we also examined the B-Raf stimulation activity of H-Ras mutants, H-Ras Ser-181,184 which is farnesylated but not palmitoylated, and H-Ras Ser-186 which is neither farnesylated nor palmitoylated. The result clearly indicated that farnesylation but not palmitoylation is the critical step for the B-Raf activation by H-Ras. The effect of the modifications on the ability of H-Ras to bind to B-Raf was examined by both the RasGAP inhibition assay and the in vitro binding assay, and the result clearly indicated that the modifications are not required for efficient association between H-Ras and B-Raf N terminus.
The mechanism by which Raf is activated by Ras is still unclear. It was proposed that a major consequence of the direct association with Ras is to recruit Raf-1 to the plasma membrane, where Raf-1 is subject to activation by an unknown mechanism (36,37). This was based on the finding that attachment of the C-terminal peptide of K-Ras containing the farnesylation signal and the polybasic region to the Raf-1 C terminus abrogated the requirement of Ras for the Raf-1 activity (36,37). Obviously, the establishment of the cell-free system from rat brain cytosol strongly suggests that this membrane translocation model cannot be simply applied to B-Raf. The observed requirement of the post-translational modifications of H-Ras (especially farnesylation) for the activation process of B-Raf implies that the modifications are required for B-Raf to undergo a further conformational change for assuming its active conformation than that induced by the association with Ras. Alternatively, if an additional factor is required for the activation, the modifications may facilitate the association of Ras or B-Raf with it. This may reflect a difference in the activation mechanisms of Raf-1 and of B-Raf. Recently, it was reported that the H-Ras Ser-181,184 mutant has the ability to activate the MAP kinase cascade in vivo possibly through activation of Raf-1 as efficiently as the wild-type even though it is localized in the cytosol but not on the plasma membrane (38). The unmodified mutant H-Ras Ser-186 , also located in the cytosol, failed to activate the MAP kinase cascade. This is similar to our observations about B-Raf and adenylyl cyclase in this study and in a previous report (24). Another report (39) has appeared showing that the Ras2 Ser-318 mutant, which is farnesylated but not palmitoylated, is located in the cytosol and still maintains its full biological activity, whereas the unmodified mutant Ras2 Ser-319 completely lost the activity. This also suggests that the farnesylation of Ras protein, but not its membrane targeting, is required for yeast Ras to fulfill its function in the yeast cells. These results plus our observations suggest that the post-translational modifications, especially farnesylation, of Ras protein is generally essential for its ability to activate its effectors. The results were shown similarly as described in Fig. 5. B, adenylyl cyclase activity was measured in the presence of 10 pmol of the various forms of H-Ras as described in A.