Eur. J . Biochem. 174, 621 -627 (1988) 0 FEBS 1988

Conformational alterations detected by circular dichroism induced in the normal ras p21 protein by activating point mutations at position 12,59, or 61 Alfonso VALENCIA’, Luis SERRAN03, Rafael CABALLERO’, Paul S. ANDERSON4 and Juan Carlos LACAL4 ’ Instituto de Investigaciones Biomtdicas, Facultad de Medicina, Universidad Autonoma, Madrid ’ Departamento de Bioquimica, Facultad de Ciencias, Universiddd Complutense, Madrid

Centro de Biologia Molecular, Universidad Autonoma, Cantoblanco, Madrid Laboratory of Cellular and Molecular Biology, National Cancer Institute, Bethesda, Maryland

(Received November 25,1987) - EJB 87 1323

Activation of the oncogenic potential of ras oncogenes occurs by point mutations at codons 12, 13, 59, 61, and 63 of the sequences that codify for its product, a 21-kDa protein designated as p21. This activation has been postulated by computer models as modifiers of the structure of the protein, which may alter its biochemical and biological activities. We have expressed in bacteria the normal ras p21 and five mutated p21 proteins with mutations at positions 12, 59, 61, 12 plus 59, and 12 plus 61. Purification was carried out by solubilization from bacterial pellets in 7 M urea and chromatography through a Sephadex G-100 column to obtain > 95% purified proteins. Circular dichroic (CD) spectra showed that the normal protein and that activated by substitution of Alas’ to Thr5’ are very similar in their overall structure. By contrast, point mutations affecting either 12 or 61 residues substantially altered the structure of the proteins. When the parameters of Chen et al. [Biochemistry 11, 4120-4131 (1972)l were applied to the CD spectra, both normal and thr5’-mutated ras proteins showed a less organized structure than mutated proteins at position 12 or 61. Since the Thr5’ mutant has a more similar transforming activity than other activated proteins, but a GTPase activity similar to that of the normal protein, our results support the hypothesis that there is more than one mechanism of activation of the ras p21 protein. One of these mechanisms involves important structural alterations by point mutations at position 12 or 61 which reduce the GTPase activity of the protein. Another mechanism will be that induced by a substitution of Ala5’ to Thr59 which does not substantially alter the protein conformation. A putative alternative mechanism for the activation of this mutant is discussed.

The p21 product of the ras oncogene has been related to the production of tumors in man and animals [l -31. Point mutations within the coding sequence of the normal genes at positions 12,13,59,61, or 63 [4- 81 alter the biological activity of the protein, inducing a transforming 21-kDa product, p21. However, it is still unknown how such point mutations, which alter the p21 sequence, can affect the normal function of the protein.

The rczs p21 proteins have GDP/GTP-binding properties [9, 103, hydrolyze GTP 113-141 and when Thr is at position 59, they show a GTP-dependent, autophosphorylation ac- tivity [15]. Point mutations that affect the transforming ac- tivity of the protein at position 12, 59, or 61, do not substan- tially modify its ability to bind GDP or GTP or its location at the plasma membrane [16]. However, point mutations at positions 12 or 61, but not at 59, reduce the GTPase activity of the protein by 5- 10-fold 111 - 141.

Recently, it has been suggested that a reduction of the GTPase activity of the ras p21 protein is not a good estimate of its transforming potential. Proteins with substitution of Ala5’ to Thr5’ show a GTPase activity comparable to the normal counterpart but are efficient transforming proteins [14]. Moreover, a reduction of the GTPase activity is not

Correspondence to A. Valencia, Instituto de Investigaciones Bio- medicas, Facultad de Medicina de la Universidad Autonoma, Arzobispo Morcillo 4, E-28029 Madrid, Spain

enough to make a ras p21 protein transforming, since point mutations at position 61 which reduce the GTPase activity of the protein behave like normal proteins in vivo 1171.

Very recently, a model has been developed for the structure of the ras p21 molecule based on the similarities found be- tween ras proteins and other GTP-binding proteins, especially the EF-Tu [18]. This model defines the amino acid residues involved in the GTP-binding site as regions around positions 10-16,57-63 and 116-119.

We present here a new approach to study the structural and biological implications of the activation by point mu- tations of the ras p21 proteins by means of circular dichroic (CD) analysis of bacterially expressed ras p21 proteins carrying different activating point mutations. Our CD spectra support the idea of more than one mechanism of activation of the ras p21 protein as previously proposed by us [14] and gives additional information for the understanding of the ras p21 protein conformation.

MATERIALS AND METHODS

Expression of ras p21 proteins in Escherichia coli

Expression of the normal ras p21 protein, p21H (Gly12- Ala5’); the viral p21 proteins codified by the BALB-MSV, p21H ( L y ~ ” - A l a ~ ~ ) ; and Harvey-MSV, p21H (Arg12 - Thr5’); as well as a chimeric protein with the normal Gly”

622

codon but mutated Thr59, p21H (GlyI2 -Thrs9), were performed in E. coli strains RRI or N4830, as previously described [14, 191. In addition, we have expressed two new proteins with a point mutation in position 61, p21H (GlyI2 - Ala59 - Led’), and a double-point mutated protein at po- sitions 12 and 61, p21H (Lysl - Ala59-Leuh l), utilizing a very similar strategy to that previously described [14, 191. Bacterial cells containing the ras gene expression vectors were grown in 1.5 ml NZY broth 1201 supplemented with 50 pg ampicillin/ ml at 30°C. When an As90 z 0.2 was reached, cells were transferred to 42°C and incubated at 250 rpm for 1 h.

Purification and characterization of bacterially expressed ras p21 proteins

Bacterial clones were treated as above for the ras p21 expression. Cells were collected by centrifugation at 2500 rpm for 10 min in a Sorvall RT 6000 centrifuge at 4”C, and pellets washed twice with 500 ml 30 mM Tris/HCl, 5 mM EDTA, 100 mM NaCl (pH 7.5) and sonicated for five 60-s periods in 50 ml of the same buffer. After centrifugation at 12000 rpm for 10min in a Sorvall SS34 rotor, ras p21 proteins were solubilized in 7 M urea, 20 mM Tris/HCl (pH 7.5) and centri- fuged at 30000 rpm for 30 min in a Beckman ultracentrifuge. A final step of purification was chromatography through a column (90 x 2.5 cm) of Sephadex G-100 in same buffer containing 7 M urea. Aliquots of 2ml were taken and analyzed by sodium dodecyl sulfate/polyacrylamide gel electrophoresis, as described previously (141. Fractions con- taining >95% pure p21 ras mutant proteins were pooled, dialized extensively against 20 mM Tris/HCl (pH 7.5) to re- move urea, and kept at -70°C until utilized. Biochemical activities of purified pas proteins were analyzed as previously described 1141.

Circular dichroic spectra

Circular dichroic (CD) spectra were recorded in a Mark I11 dicrograph (Jobin-Yvon). For spectra in the near-ultraviolet region (320 -250 nm) I-cm optical-path cells were utilized. The instrument was operated at a sensitivity of AA mm-’. Cells of 1 cm and of 0.1 cm optical paths were utilized for the spectra in the far-ultraviolet region (250- 190 nm) with the dicrograph sensitivity set at 5 x A A mm-’. Results are expressed as molar ellipticities in deg x cm2 x (dmol residue) -’ for the far-ultraviolet region or in deg x cm2 x (dmol protein)-’ in the near-ultraviolet region. The base line was measured each time for each spectrum and 10 spectra were recorded on newly prepared samples at least twice on different days. The different spectra for each protein were averaged and the standard deviation calculated. In order to analyze the differences between the spectra of the normal p21 and mutants, a t-test was performed at each wavelength. Estimation of the contribution of a-helix, p-form, p-turn or random coil to the secondary structure of the protein was carried out by a nonlinear regression method using the pa- rameters of Chen et al. 1211. Concentration of each protein was obtained by amino acid analysis performed in a Durrum D-5000 amino acid analyzer. Hydrolysis of proteins and peptides were carried out at 105°C in constant-boiling HCl, containing 0.1% phenol, in evacuated, sealed tubes for 24 h. Synthetic peptides were obtained from OCS Laboratories (Texas).

Table 1. GTP-binding activity of normal and muiated ras p21 proteins expressed in E. coli Purified ras p21 proteins were analyzed for GTP binding activity as previously described in 50 mM Tris/HC1, 10 mM MgC12, 5 mM dithiothreitol, 100 pg/ml BSA buffer [14]. Results are expressed as the molar ratio of [E-~’P]GTP to purified protein

~ ~~

Protein GTP binding

mol/mol

p21H (Gly’2-Alas9) 0.157 p21 H (Gly” - Thr5’) 0.460 p21H (Gly” - Ala5’ - Leu6’) 0.155 p21H (Ly~’~-AAla~’-Leu~’) 0.151 p21 H (Lys ’ - Ala5 ’) 0.147 p2 1 H (Arg ’ - Thr ”) 0.473

RESULTS

Purijiication and analysis of ras p21 proteins

Purification of the different ras p21 proteins was carried out as described under Materials and Methods. Yields of 10- 15% of the total p21 in bacterial crude 7 M urea extracts were obtained with a purity of >95%.

When purified proteins were analyzed for the in vitro known activities, similar levels of GTP binding were found among the normal protein, p21H (GlyI2 -Alas9) and those mutated at position 12, p21H (LysI2-Ala5’); 61, p21H (Gly12-Alas9-Leu6’); and 12 plus 61, p21H (LysI2- Ala59 - Leu6’). By contrast, the viral Harvey ras p21 protein, p21H (Arg”-Thrs9), and the chimeric protein, p21H (GlyI2 - Thr59), showed a threefold increased GTP-binding activity when compared to the normal p21 protein (Table 1). We have previously shown that the affinity constants of both types of proteins are similar, suggesting that the difference in their GTP-binding ability may not be related to intrinsic properties of the proteins [14].

We have also recently shown that the GTPase activity of the normal p21H (Gly” -Ala59) and the chimeric protein p21H ( G l ~ ’ ~ - T h r ~ ~ ) are similar [I41 and at least fivefold higher than that of the mutated proteins at position 12, 61, 12 plus 59, and 12 plus 61, These results suggested that point mutations that activate the transforming activity of the ras p21 protein may differentially alter its in vitro activities. These differences may be the result of a different structural alteration induced in the p21 protein by the specific activating point mutation or the result of the modification of the actual site where one of the activities resides. Thus, we sought to investi- gate the putative structural alterations of the p21 proteins induced by point mutations at position 12, 59, or 61, which are known to activate the oncogenic potential of the ras proto- oncogene product.

Circular dichroic spectra of ras p21 proteins

In Fig. 1, the far-ultraviolet CD spectra, together with the estimation of the standard deviations of the ras proto- oncogene p21H (Glyi2 -Ala59) and those of the transforming proteins p21H (Lys” -Ala59), p21H (GlyI2 p21H (ArgI2 -ThrS9), p21H (Gly’’ -Ala5’ - Leu6’), and p21H (Lys” -Ala59 -Leu6‘) are shown. It appears that the spectra of the normal protein and that of the Thr59 protein are similar, while the other four proteins showed important differences. In order to determine if these differences were statistically

623

- 3 220 210

D - 10 220 2LO

200 220 210 200 220 210

200 220 210 200 220 210 Wavelength Inm)

Fig. 1 . Circular dichroic spectra of rasp21 proteins in the fur ultraviolet. The actual concentration of purified proteins was estimated as described in Methods by amino acid analysis. Each protein was made 2.5 pM separately in 20 mM Tris/HCI, pH 7.5, and CD spectra recorded in a Mark 111 dicrograph. Ten different spectra were taken for each sample at two different path lengths (1 cm and 0.1 cm). The process was repeated once more on a different day and the data averaged. Standard deviations were calculated and represented at each wavelength as points. (A) Normal H-ras p21 protein; (B) p21 mutated at position 12; (C) p21 mutated at position 59; (D) p21 mutated at positions 12 and 59; (E) p21 mutated at positions 12 and 61 (see text for substitutions) . .

significant, we performed a t-test between the normal and the mutant H-ras p21 proteins. A graphic representation of such an analysis is shown in Fig. 2. These analyses indicate that the spectra for the normal p21 protein is significantly different from all the p21 mutants except that with a point mutation at position 59, p21 (Gly12 -ThrS9), especially around 220 nm. These analyses indicate as well that any of the other mutated proteins are more similar to each other than to the normal protein (data not shown). Table 2 summarizes the values obtained when the parameters of Chen et al. [21] were utilized to estimate the percentage of secondary structure. These data confirmed that the greater similarity to the normal ras p21 among all the proteins studied was the p21 protein activated by a point mutation at residue ThrS9, while the other p21 proteins showed different percentages of secondary structure.

Previous studies have shown that the Thrs9-mutated p21 is an activated protein with transforming activity comparable to those activated by point mutations at position 12 or 61 [8, 141. Thus, minor structural changes in the region around the Ala59 position, or long-distance interactions far from it, may account for the activation of the Thr” mutant. To test this possibility, we synthesized two peptides with the residue sequence 56-70 of the normal or the ThrS9 mutant and

analyzed their CD spectra in both the far and near ultraviolet. Both peptides showed similar spectra in the far ultraviolet (see Fig. 3) and in the near ultraviolet (not shown), suggesting that substitution of Ala59 by Thr59 is not by itself able to induce any important alteration in these peptides. Consequently it does not seem possible that the same sequence change will be able to induce a major change in the structure of this region in the whole protein; thus the conformational change related to the transforming activity is most probably due to long- range interactions or to a minor change (not detectable by CD) in the region around position 59 in ras p21 (Gly‘*- Thr5’).

The ras proteins are GDP/GTP-binding proteins which hydrolyze GTP [9 - 141. Both GDP/GTP-binding and GTPase activities are dependent on the concentration of Mg2+ or Mn2+ ions [22]. We next studied the effects of different concentrations of Mg2+, Mn2+, Ca2+, or GTP on the CD spectra in the far ultraviolet of both normal p21 and the transforming protein activated by a point mutation at position 12. The recorded CD spectra did not differ signifi- cantly from those obtained in the absence of GTP or metal ions (data not shown). These data suggest that these ligands do not induce any secondary structural changes detectable in

Table 2. Percentages of secondary structure in ras p21 proteins CD spectra were obtained as indicated in Methods and the percentage estimated with the parameters of Chen et al. [21]. RMS = root mean square

Protein a-Helix p-Structure Random coil RMS

p21 H (Gly - AlaS9) p21H (Gly” -Thrs9) p21H (Gly” - Ala59-Leu6’) p21H (Lys” -AlaS9-Leu6’) p21H (Lys1’-AAlas9) p22H (Arg”-ThrS9)

Yo

18.5 18.5 25.3 30.5 31.4 23.5

16.7 17.8 26.1 31.8 27.5 20.5

64.7 63.7 48.7 37.8 41.1 56.1

4.5 7.6

10.4 9.8 6.46 4.28

200 210 220 230 2LO 250 Wavelength (nm)

Fig. 2. Statistical significance between the different spectra oj’the nor- mal H-rasp21 andmutatedproteins. The spectra of the mutated H-ras p21 proteins from Fig. t were compared to the normal protein using a t-test analysis. The degree of significance of the test is represented at each wavelength. Normal p21 was compared to protein mutated at position 59 (U), 12 (A), positions 12 and 59 (A) , position 61 (0) and positions 12 and 61 (+)

the far ultraviolet. Thus, binding of these molecules to the ras p21 molecule does not induce any mayor alteration of the secondary structure of the protein. However, our results do not rule out the possibility of minor local changes induced in particular regions of the molecule in close proximity or even distant from that of the GTP- or Mg2+-binding sites.

Circular dichroic spectra of ras p21 proteins in the presence of SDS

In the previous experiments, we have demonstrated the existence of two groups of ras p21 proteins relative to the overall conformation of the proteins generated after point mutations at position 12, 59 or 61. These results indicate that point mutations at position 12 or 61 induce more drastic changes than substitution of AlaS9 to Thrs9. SDS is an anionic detergent able to alter protein conformation by changing the a-helix and fl-form content of many proteins. This results in an increase in the amount of ordered structure and a reduction in the random-coil content of proteins [23]. In order to analyze the degree of the differences in the CD spectra between the various ras p21 mutants, we studied the SDS-induced conformational changes on the various p21 proteins. CD spectra were obtained for each protein at different SDS con- centrations until a plateau in conformational changes was achieved. An SDS concentration of 2 mM showed maximal effect on the structure. Fig. 4 shows the changes induced at

6 7 - 3 1 . , . I . I . , . , 9 200 210 220 230 2LO 250

Wavelength Inml Fig. 3. Far-ultraviolet CD spectra of two synthetic peptides of H-ras p21 protein. The CD spectra for (0) the normal p21 peptide Leus6- Asp-Thr-AlaS9-Gly-Gln-Glu-Glu-Tyr-Ser-Ala-Met-Arg-Asp-Gln- Cys” and (+) its ThrS9 analogue were obtained as described in Methods. Concentration 0.4 mM in 20 mM TrislHCl pH 7.5

222 nm in the ellipticities by increasing SDS concentrations of the normal and mutant (Gly”-ThrS9) p21 ras proteins; the changes observed are very similar, supporting their conformational similarities.

The inset to Fig. 4 shows the changes induced by SDS in the ellipticities at 222 nm for the p21 proteins with mutations at positions 12, and at 12 and 61. The differences observed between the normal and mutated proteins at position 59, and those with mutations at position 12, and 12 and 59, are maintained even in the presence of SDS. However, the folding induced by increasing concentrations of SDS was similar be- tween all the proteins tested. Thus, these results indicate that the general response to SDS treatment of the ras p21 proteins is similar although there are consistent differences due to the specific activating point mutations.

Circular dichroic spectra or ras p21 proteins in the presence of dithiothreitol

The CD spectra obtained under the plateau of con- formational changes obtained by addition of 2 mM SDS indi- cated that the increase in ellipticity at 220nm in ras p21 proteins is lower than that reported for other proteins [24, 25, 261. A possible explanation that could account for this difference is the existence of intradisulphide bridges within the structure of the protein as previously suggested 127, 281. These bridges would make certain regions of the protein less flexible, impeding its restructuring into an a-helix. This possi- bility is supported by the existence of four cysteine residues along the ras p21 protein which are conserved among all the

625

-6000 1

-8000 x 0 c .-

A

0

C

._ ._ -,,-loo00 - c -12000

-14000 0 0.5 1.0

ISDSI ImM)

1 I I 1 I l / / - l++

0.01 0.1 0.3 0.5 0.7 0.9 1.0 2.0 DTT lSDSl ImM)

Fig. 4. Variation of ellipticity of rasp21 proteins by treatment with SDS and dithiothreitol. Samples were prepared as described in Fig. 2 legend; CD spectra of each protein were recorded at different SDS concentrations (0.01 - 1 mM). Ellipticities at 222 nm were measured for each protein: p21H (Gly”-AlaS9) (m); p21H (Gly’2-Thr59) (0). At the plateau (2 mM SDS), 3 mM dithiothreitol (DTT) was added to each sample and the ellipticity at 222 nrn represented. Inset shows the comparison between the changes induced by SDS on the normal (A) and mutants at positions 12 (C), and 12 and 59 (B), together with the curves that fit best, produced by a third-degree polynomic

mammalian rmproteins at positions 51,80, 118, and 186 [29]. It has been demonstrated by genetic manipulation that Cys’ 86

is required for the biological activity of the rus p21 protein [30] and that this residue is the receptor of a molecule of palmitic acid which probably anchor the p21 protein to the cellular membrane [31]. However, the meaning of the other three highly conserved cysteines remains unknown.

We analyzed the possibility of the existence of disulphide bridges in the p21 molecule by obtaining the CD spectra of the ras p21 molecule in the presence of 2 mM SDS with the addition of dithiothreitol (1 mM); this induced an increase in the ellipticity around 220 nm (Fig. 4), suggesting the existence of at least an intradisulphide bridge that restrains the induc- tion of folding by SDS. Similar results were obtained when the other p21 proteins were analyzed under the same con- ditions (data not shown). Treatment of the bacterially expressed Harvey-MSV ras p21 protein with 1 mM dithiothreitol is sufficient to allow the titration of the six sulfhydryl groups of the protein [32]. Thus, these results strongly support our conclusion that there is at least one disulphide bridge involved in the proper conformation of the ras p21 protein.

DISCUSSION Investigations aimed at elucidating both normal and

altered biochemical functions of the ras gene family have shown that their products are GTP-binding proteins which exhibit GTPase activity [9- 141. Human oncogenes in the ras family are homologous to the normal proto-oncogene and become oncogenic by a series of unique point mutations which alter the protein sequence at position 12, 13, 59, 61 or 63 [4-81. Point mutations at position 32 or 61 have a drastic effect of impairing the GTPase activity of the protein, showing a 5 - 10-fold decreased activity. However, we have recently

demonstrated that a substitution of residue Ala59 to Thr59 originates a p21 protein which is able to hydrolyze GTP with similar efficiency to its normal counterpart [14]. This mutant shows a 3 -9-fold increase in the off-rate of bound guanine nucleotides when compared to the normal protein or other activated p21s [33]. Moreover, some point mutations at posi- tion 61, which reduce GTPase activity, are not able to activate their oncogenic potential [17]. Finally, a point mutation at position 119, which substituted a Glu119 to Ala119 residue, showed high transforming activity with a 20 - 40-fold de- creased affinity for either GDP or GTP [34]. These data strongly support the hypothesis of different structural alter- ations in the ras p21 molecules by different activating mu- tations.

Several computer models have suggested that the change from Gly12 to Lysi2 would result in a shortening of the loop between amino acids 10 - 16, producing a conformational mobility restriction [35, 361. These results are supported by a different mobility on PAGE of the mutated proteins [27]. Our CD analysis of the proto-oncogenic protein and several transforming proteins with point mutations at positions 12, 59, and 61 indicated that the normal product, p21H (GlyI2 - Ala59), and the transforming protein with a substitution Ala5’ to Thr59, p21H (Glyl’ -Thr59), are very similar in their gen- eral structure and are different from the other four p21 H-rus proteins. These differences are maintained even in the pres- ence of SDS and dithiothreitol, although the induction of folding by SDS is similar for all of them (see inset of Fig. 4). These results indicate that the differences between these two groups of proteins cannot be due only to local changes in the vicinity of the altered residue, but must be a reflection of a greater conformational change due to long-range alterations induced by mutations at positions 12 and 61.

Recently, McCormick et al. [18] have proposed a model for the tertiary structure of the rus p21 protein based on its homology to other GTP-binding proteins, especially EF-Tu,

626

where 42% of the amino acid sequences are identical or conservatively changed [37]. This model postulates that amino acids around position 12 are located in a loop related to the interaction of the phosphate groups of the GTP molecule, between a hydrophobic P-structure (amino acids 5 - 10) and an amphiphilic a-helix (amino acids 16-29). Amino acids 59 and 61 would be located in a P-turn which comprises residues 58-61, between a fl-structure (amino acids 50-57) and an or-helix (amino acids 61 -74), possibly implicated in the inter- action with Mgz+ ions [18].

We have proved by CD analysis of ras p21 proteins that addition of a reducing agent like dithiothreitol to the proteins, in the presence of SDS, induced further changes in the p21 proteins. This result strongly suggests the existence of an intrdchain disulphide bridge between two of the three highly conserved cysteines at positions 51, 80 or 118, in agreement with previous reports [32]. It has been previously reported that reducing agents such as dithiothreitol or 2-mercaptoethanol increase the ability of ras p21 proteins to bind GDP/GTP 1381. This effect could be explained if the disulphide bridge is made out of cysteines closely related or associated with the proper conformation of the GDP/GTP-binding site. If the proposed model for p21 structure by McCormick et al. 1181 is correct, a disulphide bridge between CysSo and Cysl" could explain these results satisfactorily, since residues 11 6 - 11 9 seem to be an important part of the binding site. However, if the orientation of the /3-form at residues 1 - 10 is the opposite of that which has been proposed by McCormick et al. [18], a disulphide bridge between residues Cyssl and CysSo will be more likely to occur.

Recently we have reported that a point mutation that alters AlaS9 to Thr59 activates the transforming activity of the p21 protein without affecting its GTPase activity [14]. Similarly, several point mutations at position 61 reduce 5 - 10-fold the hydrolytic activity of the protein with no effect on their biological activity [17]. The evidence presented here is in agreement with these data, since the normal p21 and the protein activated by a point mutation at position 59 (Ala59 to Thr59) have a similar structure. Both proteins showed a similar GTPase activity in vitro which is 5 - 10-fold greater than that of p21 proteins mutated at positions 12,61,12 plus 59, or 12 plus 61. Moreover, we have not detected by CD analysis any alteration of the structure in peptides carrying Alas9 or Thr59. Thus, the region around residues 55-81 is likely to be the target of a very minor conformational alter- ation around this region, not detectable by CD analysis or an alteration of the structure in a region distant from this se- quence, which affect an unknown function of the p21 protein.

Recently we have reported that p21 proteins carrying a Thrs9 residue have a 3 -9-fold increased GDP and GTP inter- change than normal proteins or proteins mutated at residues 12 or 61 [33]. Thus, the region around amino acid 59 might be involved in the GDP/GTP interchange of the protein. This hypothesis is substantiated by the ability of monoclonal antibody Y13-259 to block GDP/GTP interchange of ras proteins in vitro [33], since this antibody recognizes an epitope with involvement of amino acid residues 63-73 [28, 331.

The work presented here provides additional evidence for the existence of at least two alternative mechanisms of acti- vation of the transforming potential of ras p21 proteins. One of these mechanisms involves a drastic structural alteration, like that induced by point mutations at residues 12 or 61, which are reflected in a decreased GTPase activity of the protein. This mechanism implies that the p21 . GTP complex is the actual active form of p21, as is the case for other CTP-

binding proteins 1381. A second mechanism of activation of p21 proteins must involve a rather minor alteration of the p21 structure not detectable by CD analysis, which leaves the GTPase activity intact.

The CD spectra were performed in the Departamento de Bio- quimica, Facultad de Ciencias, Universidad Complutense, Madrid, Spain. We are indebted to F. Montero for use of the facilities. We appreciate the important contribution of J. Avila, A. Cano, S. Aaronson and S. Tronick in the interpretation of results and in the revision of the manuscript; and M. Ruta for the generation of p21H (Lys" - Alas' - Led1) by in vitro mutagenesis. A. V. is supported by the Comisibn Asesorapara la Investigacibn Cientifica y Ticnica, R. C. by a grant from the Plan Formacibn Personal lnvestigador, and L. S. by a grant from Fondo de Investigaciones Sanitarias.

REFERENCES 1. Cooper, G. (1982) Science (Wash. DC) 217, 801 -806. 2. Weinberg, R. A. (1982) Adv. Cancer Res. 36, 149-163. 3. Bishop, J. M. (1983) Annu. Rev. Biochem. 52, 301 -354. 4. Tabin, C. J., Bradley, S. M., Bargmann, C. I., Weinberg, R. A.,

Papageorge, A. G., Scolnick, E. M., Dhar, R., Lowy, D. R. & Chang, E. H. (1982) Nature (Lond.) 300, 143-149.

5. Reddy, E. P., Reynolds, R. K., Santos, E. & Barbacid, M. (1982) Nature (Lond.) 300, 149 - 152.

6. Shimizu, V., Birnbaum, D., Ruley, M. A., Fasano, O., Suard, Y., Edlund, L., Taparowsky, E., Goldfarb, M. & Wigler, M. (1983) Nature (Lond.) 304,497 - 500.

7. Yuasa, Y., Srivastava, S. K., Dunn, C. Y., Rhim, J. S., Reddy, E. P. & Aaronson, S. A. (1983) Nature (Lond.) 303,715-779.

8. Fasano, O., Aldrich, T., Tamanoi, F., Taparowsky, E., Furth, M. & Wigler, M. (1984) Proc. Nut1 Acad. Sci. USA 81,4008- 401 2.

9. Scolnick, E. M., Papageorge, A. G. & Shih, T. Y. (1979) Proc. Natl Acad. Sci. USA 76, 5355 - 5359.

10. Shih, T. Y., Papageorge, A . G., Stokes, P. E., Weeks, M. 0. & Scolnick, E. M. (1980) Nature (Lond.) 287,686-691.

11. McGrath, J. P., Capon, D. J., Goeddcl, D. V. & Levinson, A. D. (1984) Nature (Lond.) 310, 644-649.

12. Gibbs, J. B., Sigal, I. S., Poe, M. & Scolnick, E. M. (1984) Proc. Natl Acad. Sci. USA 81, 5104-5108.

13. Sweet, R. W., Yokoyama, S., Kamata, R., Feramisco, J. R., Rosenberg, M. & Gross, M. (1984) Nature (Lond.) 311, 273- 275.

14. Lacal, J. C., Srivastava, S. K., Anderson, P. S. & Aaronson, S. A. (1 986) Cell 44, 609 - 61 7.

15. Shih, T. Y., Stokes, P. E., Smythers, G. W., Dhar, R. & Oroszlan, S. (1982) J . Biol. Chem. 257, 11 767 - 11 773.

16. Finkel, T., Der, C. J. &Cooper, G. M. (1984) Cell 37, 151-158. 17. Der, C. J., Finkel, T. & Cooper, G. M. (1986) Cell 44, 167 - 176. 18. McCormick, F., Clark, B. F. C., LaCour, T. F. M., Kjeldgaard,

M., Norskov-Lauritsen, L. & Nyborg, J. (1985) Science (Wash.

19. Lacal, J. C., Santos, E., Notario, V., Barbacid, M., Yamazaki, S., Kung, H., Seamans, C., McAndrew, S. & Crowl, R. (1984) Proc. Natl Acad. Sci. USA 81, 5305 - 5309.

20. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular cloning: a laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.

21. Chen, Y. H., Yang, J. T. & Martinez, H. M. (1972) Biochemistry

22. Manne, V. & Kung, H. (1985) Biochem. Biophys. Res. Commun.

23. Lundahl, P., Creijer, E., Sandberg, M., Cardell, S. & Erickson,

24. Yang, J., Bewley, T. A., Chen, G. C. & Li, C. H. (1977) Proc.

25. Barbero, S. L., Montero, F. & Franco, L. (1984) J. Protein Chem.

DC) 230,78 -82.

11,4120-4131.

128,1440-1446.

K. (1986) Biochim. Biophys. Acta 873, 20-26.

Natl Acad. Sci. USA 74, 3225 - 3228.

3,455-463.

627

26. Valencia, A., Cano, A., Caballero, R. & Pestaiia, A. (1980) Bio-

27. Srivastava, S. K., Yuasa, Y., Reynolds, S. H. & Aaronson, S. A.

28. Sigal, I. S., Gibbs, J. B., D’Alonso, J. S. & Scolnick, E. M. (1986)

29. Shih, T. Y. &Weeks, M. 0. (1984) Cancer Invest. 2, 109-123. 30. Willurnsen, B. M., Christensen, A., Hubbert, N. L., Papageorge,

A. G. & Lowy, D. R. (1984) Nature (Lond.) 310,583-586. 31. Willumsen, B. M., Norris, K., Papageorge, A. G., Hubbert, N.

32. Hattori, S., Yamashita, T., Copeland, T. D., Oroszlan, S. & Shih,

chim. Biophys. Acta 911, 11 -38.

(1985) Proc. Natl Acad. Sci. USA 82, 38-42.

Proc. Natl Acad. Sci. USA 83,4725 -4129.

L. & LOWY, D. R. (1984) EMBO J. 3, 2581 -2585.

T. Y. (1986) J. Biol. Chem. 261, 14582-14586.

33. Lacal, J. C. & Aaronson, S. A. (1986) Mol. Cell Biol. 6, 4214- 4220.

34. Sigal, I. S., Gibbs, J. B., D’Alonzo, J. S., Temeles, G., Wolanski, B. S., Socher, S. H. & Scolnick, E. (1986) Proc. Nut/ Acad. Sci.

35. Pincus, M. & Brandt-Ranf, W. P. (1985) Proc. Natl Acad. Sci.

36. Murakami, M. (1985) J. Theor. Bid 114, 195-198. 37. Halliday, K. R. (1984) J. Cyclic Nucleotide Protein Phosphoryla-

38. Manne, V., Yamazaki, S. & Kung, H. (1984) Proc. NatZAcad.

39. Gilman, A. G. (1984) Cell 36, 571 - 579.

USA 83,Y 52 - 9 56.

USA 82, 3596 - 3600.

tion Res. 9, 435-448.

Sci. USA 81, 6953 - 6951.