Selective Inhibition of Ras Interaction with Its Particular Effector by Synthetic Peptides Corresponding to the Ras Effector Region*

Ras proteins possess multiple downstream effectors of distinct structures. We and others demonstrated that Ha-Ras carrying certain effector region mutations could interact differentially with its effectors, implying that significant differences exist in their Ras recognition mechanisms. Here, by employing the fluorescence polarization method, we measured the activity of effector region synthetic peptides bearing various amino acid substitutions to inhibit association of Ras with the effectors human Raf-1 and Schizosaccharomyces pombe Byr2. The effect of these peptides on association with another effectorSaccharomyces cerevisiae adenylyl cyclase was also examined by measuring inhibition of the Ras-dependent adenylyl cyclase activity. The peptide corresponding to the residues 17–44 competitively inhibited Ras association with all the three effectors at the K i values of 1∼10 μm, and the inhibition was considerably attenuated by the D38A mutation. The peptide with the D38N mutation inhibited association of Ha-Ras with Byr2 but not with the others, whereas that with the P34G mutation inhibited association of Ha-Ras with Raf-1 and Byr2 but not with adenylyl cyclase. Thus, the specificity observed with the whole Ras protein was retained in the effector region peptide. These results suggest that the effector region residues constitute a major determinant for differential recognition of the effector molecules, raising a possibility for selective inhibition of a particular Ras function.

The ras genes are widely conserved from yeasts to mammals. Their protein products belong to a family of small guanine nucleotide-binding proteins that operate in key processes of intracellular signal transduction systems regulating cell growth and differentiation. In higher eukaryotes including vertebrates, Ras proteins bind directly and activate Raf (Raf-1, A-Raf, and B-Raf) proteins, which results in activation of the mitogen-activated protein kinase phosphorylation cascade (for reviews, see Refs. 1 and 2). In addition, other mammalian proteins have been shown to associate directly with Ras in a GTP-dependent manner. These include Ral-guanine nucleotide dissociation stimulator (3)(4)(5), phosphatidylinositol 3-kinase (6), the isoform of protein kinase C (7), AF-6 (8), and Rin1 (9).
However, the significance of their interaction with Ras is presently unclear. In the fission yeast Schizosaccharomyces pombe, its single Ras homologue, Ras1, is involved in signal transduction from mating pheromone receptors, and protein kinase Byr2 is its direct downstream effector (10,11). On the other hand, in the budding yeast Saccharomyces cerevisiae, adenylyl cyclase is an immediate downstream effector of a pair of Ras proteins, Ras1 and Ras2 (for a review, see Ref. 12).
Mutational studies of the effector molecules identified discrete Ras binding regions. The Ras binding site of mammalian Raf-1 was mapped to an 81-amino acid segment in its Nterminal regulatory domain (13)(14)(15). Recently another Ras binding site was identified in Raf-1, corresponding to the cysteine-rich region (16,17). In yeast adenylyl cyclase, the leucinerich-repeat domain, which is composed of tandemly repeated 23-amino acid leucine-rich motifs (18), was shown to bind directly to Ras (19). Byr2 contains a Ras binding site within its N-terminal 206 amino acid residues (11). Also, discrete Ras binding domains were identified and mapped in other Ras effectors (3)(4)(5)(6)(7)(8)(9). Comparison of the primary structures of these Ras binding domains revealed very little conservation among them except that those of Ral-guanine nucleotide dissociation stimulator, AF-6, and Rin1 share a weakly homologous region called RA domain (20). However, recent studies on the Ras binding regions of Raf-1 and Ral-guanine nucleotide dissociation stimulator revealed that they exhibit similar three-dimensional structures (21)(22)(23).
On the other hand, extensive mutational studies on Ras identified amino acid residues whose substitution abolished the ability to transform NIH3T3 cells, to induce neurite outgrowth in PC12 cells, to induce germinal vesicle breakdown in Xenopus laevis oocytes, or to associate with Raf-1 without affecting the guanine nucleotide binding properties (24 -31). These residues were concentrated in the positions 32-40 of mammalian Ras, so that this region was designated the effector region. This region almost matches with "switch I," one of the two regions that take significantly different conformations between the GTP-and GDP-bound forms of Ras (32,33). The x-ray crystallographic study on the complex between a Ras homologue Rap1A and Raf-1 Ras binding region has provided evidence for their association at the atomic level (21). In addition, the residues 26 -31 and 42-53 flanking the effector region were shown to be critical for effector activation even though they do not change their conformations upon GDP/GTP exchange (30,31,34). These residues have been proposed to constitute the "activator region" (31) or "constitutive effector region" (34), and this region was recently implicated in interaction with the cysteinerich region of Raf-1 (17).
Recently, we and others found that certain Ras mutants carrying substitutions in the effector region could discriminate the effector molecules (35)(36)(37)(38)(39). Specifically, more than 50 Ha-Ras mutants were examined by us for interaction with Raf-1, Byr2, and yeast adenylyl cyclase, the three effectors whose functions are firmly established by both genetic and biochemical evidences (36). We found that the D38N mutant lost the ability to associate with both Raf-1 and adenylyl cyclase while retaining the activity to associate with and activate Byr2. Also, the P34G mutation selectively abolished the ability of Ha-Ras to bind and activate adenylyl cyclase without affecting the interaction with Raf-1 and Byr2, suggesting that significant differences exist in the recognition mechanisms by which distinct effector molecules associate with Ras (36). These observations opened up a possibility for selective inhibition of the function of a particular effector by exploiting the differences in the Ras recognition mechanisms. Here, we demonstrate that short synthetic peptides corresponding to the effector region can be employed for this purpose by showing that they retain the ability to differentially recognize the effectors.

EXPERIMENTAL PROCEDURES
Materials-Synthetic peptides used in this experiment are shown in Table I. They were synthesized with a solid-phase method and purified by C18 reversed-phase high performance liquid chromatography (purchased from Peptide Institute Inc., Osaka, Japan). The peptides were dissolved in buffer A (20 mM Tris/HCl (pH 7.4), 50 mM NaCl, 5 mM MgCl 2 , 1 mM EDTA) or in dimethylformamide and centrifuged at 100,000 ϫ g for 30 min to remove insoluble materials. Their concentrations were determined spectroscopically by measurement of absorbance at 293 nm at pH 11.6. An MBP 1 fusion protein of the Raf-1 N-terminal 206 residues, MBP-Raf-1-(1-206), and a GST fusion protein of the Byr2 N-terminal 236 residues, GST-Byr2-(1-236), were described before (11,17,36). Yeast strain FS3-1 (MAT␣ his3 leu2 trp1 ura3 ade8 cyr1-2 ras2::URA3) carrying plasmids pAD4-GST-CYR1-(606 -2026) and YEP-HIS3-ADC1-CAP, which overexpressed a GST fusion protein of adenylyl cyclase and the cyclase-associated protein CAP, respectively, under the control of the ADC1 promoter, was described before (40). The posttranslationally fully modified and unmodified forms of human Ha-Ras were extracted from Sf9 cells infected with the baculovirus carrying Ha-ras and purified as described previously (41).
Measurement of Fluorescence Polarization-The purified unmodified form of Ha-Ras was labeled with fluorescein (*F) at its N-terminal ␣-amino group by reaction with fluorescein succinimidylester (Pan Vera Co., Madison, Wisconsin) at pH 7.0. The fluorescein-labeled Ha-Ras (*F-Ha-Ras) was separated from the unincorporated dye by gel filtration chromatography on Sephadex G-25. One microliter of *F-Ha-Ras (about 10 5 fluorescence units or 8 fmol/reaction) loaded with GTP␥S or GDP␤S was mixed with 1 ml of buffer A containing various concentrations of the purified MBP-Raf-1-(1-206) or GST-Byr2-(1-236) in a siliconized glass tube. The reaction mixture was incubated for 15 min at 30°C in the darkness followed by measurement of the fluorescence polarization value of each tube with BEACON (Pan Vera Co.) at an excitation wavelength of 490 nm and an emission wavelength of 520 nm. For assay of inhibition by the synthetic peptides, similar measurements were performed with 100 nM MBP-Raf-1-(1-206) or GST-Byr2-  in the presence of varying concentrations of the synthetic peptides.
A fluorescence polarization value is given by a difference between the vertical and horizontal fluorescence intensities divided by a sum of the vertical and horizontal fluorescence intensities and expressed as mP units (42,43). It reflects the molecular size of a fluorescently labeled molecule (42,44,45). There occurs a large increase in the molecular size of *F-Ha-Ras upon complex formation with the effector proteins, causing an increase in the fluorescence polarization value. Thus, the change in fluorescence polarization of *F-Ha-Ras is proportional to the amount of the complex formed with the effector molecule.
Adenylyl Cyclase Inhibition Assay-Adenylyl cyclase was solubilized from the crude membrane fraction of the yeast FS3-1 carrying pAD4-GST-CYR1-(606 -2026) and YEP-HIS3-ADC1-CAP with buffer C (50 mM Mes (pH 6.2), 0.1 mM MgCl 2 , 0.1 mM EGTA, 1 mM ␤-mercaptoethanol) containing 1% LubrolPX, 0.6 M NaCl, and 1 mM phenylmethylsulfonyl fluoride. After centrifugation at 100,000 ϫ g for 1 h at 4°C, the resulting supernatant was used for the measurement of adenylyl cyclase activity in the presence of varying concentrations of the GTP␥Sbound form of posttranslationally fully modified Ha-Ras as described before (19,40). For the measurement of inhibition of adenylyl cyclase activities by the synthetic peptides, varying amounts of the peptides were added to the reaction mixtures before commencing the reaction.

RESULTS
Binding of Fluorescently Labeled Ha-Ras to N-terminal Regions of Raf-1 and Byr2--We employed the fluorescence polarization method to quantitatively analyze the association of Ha-Ras with its various effectors. The change in fluorescence polarization value of *F-Ha-Ras is proportional to the amount of a complex formed with its effector molecule. In this experiment, *F-Ha-Ras was loaded with GTP␥S or GDP␤S and examined for association with various concentrations of MBP-Raf-1-(1-206) (Fig. 1A) or GST-Byr2-(1-236) (Fig. 1B). *F-Ha-Ras alone exhibited an mP value of approximately 100, and the complex formation of *F-Ha-Ras with the effector polypeptides was detectable at the effector concentration as low as 10 nM. The binding was almost saturated at 100 nM for MBP-Raf-1-(1-206) or at 500 nM for GST-Byr2-(1-236) with the increase of the mP value by approximately 30 units. The extent of increase was consistent with that predicted from the increase in molecular size upon the complex formation with the effector polypeptides based on the observation that the reciprocals of the mP values correlate linearly with the reciprocal values of the molecular weights of the fluorescently labeled proteins (44,45). This suggested that the binding reactions occurred stoichiometrically. The maximal bindings were reached within 10 min of incubation, and the values remained stable for the following 30 min (data not shown). The binding reactions were dependent on the GTP-bound configuration of *F-Ha-Ras, and almost no binding was observed for MBP or GST only (Fig. 1, A and B). Since the *F-Ha-Ras concentration used was very low at 8 pM, essentially all of MBP-Raf-1-(1-206) or GST-Byr2-(1-236) added to the reactions could be regarded as in a free form. When reciprocals of the fluorescence polarization increase values were plotted against reciprocals of the effector concentrations, a straight line could be drawn with an intersect with the horizontal axis giving K d values of 35 nM for MBP-Raf-1-(1-206) and 90 nM for GST-Byr2-(1-236) (Fig. 1C). These values are comparable although somewhat higher than those estimated from the inhibition profiles of Ras-dependent adenylyl cyclase activity by the same effector fragments (11,19).
Inhibition of Ras Effector Association by Synthetic Peptides-We next examined the ability of synthetic peptides encompassing the Ha-Ras effector region with various single amino acid substitutions (Table I) to interfere with the Ha-Ras-Raf-1 and Ha-Ras-Byr2 associations. The peptide corresponding to the residues 17-44 was chosen because this peptide had been used successfully in inhibition of the Ras binding to Raf-1 or to protein kinase C (7,46). In addition, similar experiments were carried out with synthetic peptides SII and AD, corresponding to the switch II region (residues 58 -76) of Ha-Ras and to a part of the activator region (residues 39 -51), respectively (Table I). Fig. 2 shows inhibition profiles of the association of *F-Ha-Ras with MBP-Raf-1-(1-206) ( Fig. 2A) or with GST-Byr2-(1-236) (Fig. 2B) by increasing concentrations of the various peptides. Wild-type effector region peptide effectively inhibited the Ras association with both of the effectors with 70% inhibition for Raf-1 and 55% inhibition for Byr2 at the concentration of 10 M, whereas that carrying the D38A mutation was much less effective. In addition, the specificities of inhibition by the various peptides were clearly different between the two effectors.
Although P34G substitution did not notably affect the inhibitory activity of the peptide against both Raf-1 and Byr2, D38N substitution selectively impaired the inhibitory activity against complex formation of *F-Ha-Ras with Raf-1-(1-206) but not with Byr2- (1-236). Both SII and AD peptides had no inhibitory effect up to the concentration of 10 M. Kinetic analyses of the inhibition patterns were carried out as follows to determine K d values of each peptide for the two effectors.
The fluorescence polarization value (Y) is proportional to the amount of *F-Ha-Ras bound to the effector and hence formulated as Y ϭ Y max ⅐ X/(K Ras ϩ X), where Y max , X, and K Ras represent the maximal polarization value, the effector concentration, and the K d value of *F-Ha-Ras for the effector, respectively. Assuming that the association with *F-Ha-Ras is competitively inhibited by a peptide whose concentration (P) is in a vast excess over X, the concentration of the effector sequestered by the peptide is X ⅐ P/(K p ϩ P), where K p represents the K d value of the peptide for the effector. Thus, the free effector concentration available for association with *F-Ha-Ras is given by X ⅐ (1 Ϫ P/(K p ϩ P)). The fluorescence polarization value (Y p ) in the presence of the peptide (P) can be formulated as Y p ϭ Y max ⅐ X ⅐ (1 Ϫ P/(K p ϩ P))/(K Ras ϩ X ⅐ (1 Ϫ P/(K p ϩ P))). From the two equations is derived Y/Y p ϭ (K Ras /K p ⅐ (K Ras ϩ X)) ⅐ P ϩ 1. When Y/Y p was plotted against P for inhibition by the various peptides of association with Raf-1 (Fig. 2C) or with Byr2 (Fig. 2D), a straight line can be drawn for either case that converged on the longitudinal axis at Y/Y p ϭ 1. The K p value can be calculated as ϪK Ras ⅐ P 0 /(K Ras ϩ X), where P 0 represents the P value at the intersect on the horizontal axis (Fig. 2, C and  D). Thus, the K p values for MBP-Raf-1-(1-206) of wild-type, P34G, D38N, and D38A peptides were calculated as 2.1, 4.4, 28, and 28 M, respectively, given that K Ras is 35 nM and X is 100 nM. Similarly, the K p values of the same set of peptides for GST-Byr2-(1-236) were calculated as 3.8, 2.0, 4.7, and 39 M, respectively.
Inhibition of Ha-Ras-dependent Adenylyl Cyclase Activity by Synthetic Peptides-The effects of the various synthetic peptides on association between Ha-Ras and yeast adenylyl cyclase were examined by measuring inhibition of Ha-Ras-dependent adenylyl cyclase activity (Fig. 3). Wild-type peptide inhibited adenylyl cyclase activity obtained by the 8 nM post-translationally modified form of Ha-Ras by a half at approximately 21 M (Fig. 3A). A notable difference from Raf-1 and Byr2 was observed in the effect of P34G substitution, which was found totally ineffective in inhibition of adenylyl cyclase activity. Assuming that the inhibitory associations of the peptides are competitive with that of Ha-Ras, the inhibition constants (K i ) for the peptides can be derived from a conventional equation where V, V max , X, K m , and I represent the adenylyl cyclase activity, the maximal adenylyl cyclase activity, the concentration of Ha-Ras, the K d value for Ha-Ras, and the inhibitor peptide concentration, respectively. Rearranging the equation according to Lineweaver and Burk Adenylyl cyclase activities dependent on various concentrations of Ha-Ras were subjected to inhibition by varying concentrations of the peptides, and a reciprocal of the activity was plotted against a reciprocal of the Ha-Ras concentration. This gave a series of straight lines that converged on the longitudinal axis for each concentration of the peptide, wild type (Fig.  3B), D38N (Fig. 3C), and D38A (Fig. 3D), as predicted from the    (Fig. 3E). The K i values could be estimated from intersects of the straight lines on the horizontal axis and was determined to be 10, 56, and 68 M for wild-type, D38N, and D38A peptides, respectively.

DISCUSSION
Synthetic peptides encompassing the Ras effector region, in particular the one corresponding to the residues 17-44, were used successfully to competitively inhibit association of Ras with its effector proteins including Raf-1 (46) and the isoform of protein kinase C (7). Contact epitope scanning using a series of smaller peptides suggested that many residues across the residues 17-51 may be involved in direct interaction with Raf-1 N-terminal fragment (47). In the present study, we have shown for the first time that introduction of certain single amino acid substitutions can confer the ability to discriminate Ras effectors, human Raf-1, fission yeast Byr2, and budding yeast adenylyl cyclase upon the effector region peptide. The association of the wild type and its mutant peptides with the three effectors was quantitatively examined by measuring competitive inhibition of the binding of the effectors to Ha-Ras, which was mon-itored by the fluorescence polarization method (42)(43)(44). The use of this method enabled us to measure the equilibrium binding constants of the effector molecules for Ras in solution and to determine the K d values for association with the peptides from kinetic analyses of the inhibition profile. This is certainly advantageous over other nonequilibrium procedures such as an enzyme-linked immunosorbent assay and an affinity precipitation of the complex formed, both of which require the physical separation of free and bound ligands considering that the binding nature of Ras to Raf-1, for example, is substoichiometric and the dissociation rate is very high (48).
The results clearly demonstrated selectivity in binding of the peptides to the distinct effectors. Specifically, D38N substitution attenuated the affinity of the effector region peptide for both Raf-1 and adenylyl cyclase by about 10-fold without affecting that for Byr2. The importance of Asp-38 of Ras for interaction with Raf-1 had been indicated from extensive mutational and structural studies on the whole Ras protein and its homologue Rap1A (21, 24 -33). In contrast, P34G substitution almost abolished the interaction with adenylyl cyclase but did not appreciably affect that with Raf-1 or Byr2. These specificities, summarized in Table II that residues 17-44 are sufficient to assume a specific conformation that interacts selectively with a particular set of the effectors depending on a single amino acid substitution. The results strongly suggest that the effector region residues by themselves constitute a major determinant for differential recognition of the effector molecules, although we cannot exclude the possibility that residues outside of the effector region may contribute to this recognition process. Since other Ras effector molecules such as Ral-guanine nucleotide dissociation stimulator and phosphatidylinositol 3-kinase were also known to possess distinct requirements for the effector region residues from Raf-1, Byr2, or adenylyl cyclase and from one another (37,38), it may be possible to find a synthetic peptide carrying certain amino acid substitutions that specifically recognize any member of the multiple Ras effector molecules.
Our results have also shown that it is possible to achieve selective inhibition of the association of Ras with a particular effector molecule without significantly affecting those with other effectors by employing the antagonist peptide with a relatively small molecular size around 3 kDa. This may open up  a possibility to design a compound that mimics a specific conformation of the effector region peptide bearing the particular substitution, which is in theory deducible from the x-ray crystallographic or nuclear magnetic resonance spectroscopic analysis. From a viewpoint of effectiveness in cancer drug therapy, it is possible that the effector-specific inhibition, which could potentially be achieved by administration of such a compound, is superior to the total inhibition of Ras functions by, for example, a compound that directly inactivates Ras. In addition, the effector region peptides carrying a series of substitutions might be found useful as effector-specific inhibitors for analysis of individual cellular functions of the distinct effector molecules.