Identification of amino acid residues required for Ras p21 target activation

Analysis of correlated mutations in Ras G-domain

Ras GTPases are most prevalent proto-oncogenes in human cancer. Mutations in Ras remain untreatable more than three decades after the initial discovery. At the amino acid level, some residues under physical or functional constraints exhibit correlated mutations also known as coevolving/covariant residues. Revealing intra-molecular co-evolution between amino acid sites of proteins has become an emerging area of research as it enlightens the importance of variable regions. Here, I have identified and analyzed the coevolving residues in the Ras GTP binding domain (G-domain). The obtained covariant residue position data correlate well with the known experimental data on functionally important residues. Therefore, it is of interest to understand these residue co-variations for designing protein engineering experiments and target oncogenic Ras GTPases.

Determinants of RasC specificity during Dictyostelium aggregation

RasC is required for optimum activation of adenylyl cyclase A and for aggregate stream formation during the early differentiation of Dictyostelium discoideum. RasG is unable to substitute for this requirement despite its sequence similarity to RasC. A critical question is which amino acids in RasC are required for its specific function. Each of the amino acids within the switch 1 and 2 domains in the N-terminal portion of RasG was changed to the corresponding amino acid from RasC, and the ability of the mutated RasG protein to reverse the phenotype of rasC(-) cells was determined. Only the change from aspartate at position 30 of RasG to alanine (the equivalent position 31 in RasC) resulted in a significant increase in adenylyl cyclase A activation and a partial reversal of the aggregation-deficient phenotype of rasC(-) cells. All other single amino acid changes were without effect. Expression of a chimeric protein, RasG(1-77)-RasC(79-189), also resulted in a partial reversal of the rasC(-) cell phenotype, indicating the importance of the C-terminal portion of RasC. Furthermore, expression of the chimeric protein, with alanine changed to aspartate (RasG(1-77(D30A))-RasC(79-189)), resulted in a full rescue the rasC(-) aggregation-deficient phenotype. Finally, the expression of either a mutated RasC, with the aspartate 31 replaced by alanine, or the chimeric protein, RasC(1-78)-RasG(78-189), only generated a partial rescue. These results emphasize the importance of both the single amino acid at position 31 and the C-terminal sequence for the specific function of RasC during Dictyostelium aggregation.

Farnesylation or geranylgeranylation? Efficient assays for testing protein prenylation in vitro and in vivo

BACKGROUND

Available in vitro and in vivo methods for verifying protein substrates for posttranslational modifications via farnesylation or geranylgeranylation (for example, autoradiography with 3H-labeled anchor precursors) are time consuming (weeks/months), laborious and suffer from low sensitivity.

RESULTS

We describe a new technique for detecting prenyl anchors in N-terminally glutathione S-transferase (GST)-labeled constructs of target proteins expressed in vitro in rabbit reticulocyte lysate and incubated with 3H-labeled anchor precursors. Alternatively, hemagglutinin (HA)-labeled constructs expressed in vivo (in cell culture) can be used. For registration of the radioactive marker, we propose to use a thin layer chromatography (TLC) analyzer. As a control, the protein yield is tested by Western blotting with anti-GST- (or anti-HA-) antibodies on the same membrane that has been previously used for TLC-scanning. These protocols have been tested with Rap2A, v-Ki-Ras2 and RhoA (variant RhoA63L) including the necessary controls. We show directly that RasD2 is a farnesylation target.

CONCLUSION

Savings in time for experimentation and the higher sensitivity for detecting 3H-labeled lipid anchors recommend the TLC-scanning method with purified GST- (or HA-) tagged target proteins as the method of choice for analyzing their prenylation capabilities in vitro and in vivo and, possibly, also for studying the myristoyl and palmitoyl posttranslational modifications.

Suppression of integrin activation by activated Ras or Raf does not correlate with bulk activation of ERK MAP kinase

The rapid modulation of ligand-binding affinity ("activation") is a central property of the integrin family of cell adhesion receptors. The Ras family of small GTP-binding proteins and their downstream effectors are key players in regulating integrin activation. H-Ras can suppress integrin activation in fibroblasts via its downstream effector kinase, Raf-1. In contrast, to H-Ras, a closely related small GTP-binding protein R-Ras has the opposite activity, and promotes integrin activation. To gain insight into the regulation of integrin activation by Ras GTPases, we created a series of H-Ras/R-Ras chimeras. We found that a 35-amino acid stretch of H-Ras was required for full suppressive activity. Furthermore, the suppressive chimeras were weak activators of the ERK1/2 MAP kinase pathway, suggesting that the suppression of integrin activation may be independent of the activation of the bulk of ERK MAP kinase. Additional data demonstrating that the ability of H-Ras or Raf-1 to suppress integrin activation was unaffected by inhibition of bulk ERK1/2 MAP kinase activation supported this hypothesis. Thus, the suppression of integrin activation is a Raf kinase induced regulatory event that can be mediated independently of bulk activation of the ERK MAP-kinase pathway.

The GTP-binding protein Rho1p is required for cell cycle progression and polarization of the yeast cell

Previous work showed that the GTP-binding protein Rho1p is required in the yeast, Saccharomyces cerevisiae, for activation of protein kinase C (Pkc1p) and for activity and regulation of beta(1-->3)glucan synthase. Here we demonstrate a hitherto unknown function of Rho1p required for cell cycle progression and cell polarization. Cells of mutant rho1(E45I) in the G1 stage of the cell cycle did not bud at 37 degrees C. In those cells actin reorganization and recruitment to the presumptive budding site did not take place at the nonpermissive temperature. Two mutants in adjacent amino acids, rho1(V43T) and rho1(F44Y), showed a similar behavior, although some budding and actin polarization occurred at the nonpermissive temperature. This was also the case for rho1(E45I) when placed in a different genetic background. Cdc42p and Spa2p, two proteins that normally also move to the bud site in a process independent from actin organization, failed to localize properly in rho1(E45I). Nuclear division did not occur in the mutant at 37 degrees C, although replication of DNA proceeded slowly. The rho1 mutants were also defective in the formation of mating projections and in congregation of actin at the projections in the presence of mating pheromone. The in vitro activity of beta(1-->3)glucan synthase in rho1 (E45I), although diminished at 37 degrees C, appeared sufficient for normal in vivo function and the budding defect was not suppressed by expression of a constitutively active allele of PKC1. Reciprocally, when Pkc1p function was eliminated by the use of a temperature-sensitive mutation and beta(1-->3)glucan synthesis abolished by an echinocandin-like inhibitor, a strain carrying a wild-type RHO1 allele was able to produce incipient buds. Taken together, these results reveal a novel function of Rho1p that must be executed in order for the yeast cell to polarize.

Interaction of activated Ras with Raf-1 alone may be sufficient for transformation of rat2 cells

v-H-ras effector mutants have been assessed for transforming activity and for the ability of the encoded proteins to interact with Raf-1-, B-Raf-, byr2-, ralGDS-, and CDC25-encoded proteins in the yeast two-hybrid system. Transformation was assessed in rat2 cells as well as in a mutant cell line, rv68BUR, that affords a more sensitive transformation assay. Selected mutant Ras proteins were also examined for their ability to interact with an amino-terminal fragment of Raf-1 in vitro. Finally, possible cooperation between different v-H-ras effector mutants and between effector mutants and overexpressed Raf-1 was assessed. Ras transforming activity was shown to correlate best with the ability of the encoded protein to interact with Raf-1. No evidence for cooperation between v-H-ras effector mutants was found. Signaling through the Raf1-MEK-mitogen-activated protein kinase cascade may be the only effector pathway contributing to RAS transformation in these cells.

Positive and negative modulation of H-ras transforming potential by mutations of phenylalanine-28

Conserved amino-acids of H-ras from residues 25 to 34 were mutated in human H-ras cDNA with a pre-existing valine-12 activating mutation ([V12]p21), and built into SV40-driven expression vectors. The influence of the introduced mutations was initially screened by transfection of Rat-1 cells to score foci of transformed cells. Non-conservative mutations of amino-acids 25 (tryptophan for glutamine), 27 (asparagine for histidine) and 34 (alanine for proline) did not abrogate the transforming potential of [V12]p21. The conservative mutation of phenylalanine-28 to tryptophan ([V12W28]p21) was also still transforming. Significantly, in the absence of the valine-12 activating mutation, tryptophan-28-ras ([W28]p21) was weakly transforming while, in contrast, [V12D28]p21 was unable to transform Rat-1 cells and retarded cell growth. Analysis of the binding and dissociation of GTP and GDP to normal and mutated p21 expressed in Escherichia coli showed that [V12D28]p21 and [D28]p21 do not bind GTP. The dissociation rate of both GTP and GDP bound to [W28]p21 is increased, suggesting a mechanism for its transforming potential in Rat-1 cells. These studies illustrate the importance of phenylalanine-28 in guanine nucleotide binding by p21H-ras. The mutations described could be valuable tools in investigations of cellular signal transduction involving small GTP-binding proteins.

Critical binding and regulatory interactions between Ras and Raf occur through a small, stable N-terminal domain of Raf and specific Ras effector residues

Genetic and biochemical evidence suggests that the Ras protooncogene product regulates the activation of the Raf kinase pathway, leading to the proposal that Raf is a direct mitogenic effector of activated Ras. Here we report the use of a novel competition assay to measure in vitro the relative affinity of the c-Raf-1 regulatory region for Ras-GTP, Ras-GDP, and 10 oncogenic and effector mutant Ras proteins. c-Raf-1 associates with normal Ras and the oncogenic V12 and L61 forms of Ras with equal affinity. The moderately transforming mutant Ras[E30K31] also bound to the c-Raf-1 regulatory region with normal affinity. Transformation-defective Ras effector mutants Ras[N33], Ras[S35], and Ras[N38] bound poorly. In contrast, the transformation defective Ras[G26I27] and Ras[E45] mutants bound to the c-Raf-1 regulatory region with nearly wild-type affinity. A stable, high-affinity Ras-binding region of c-Raf-1 was mapped to a 99-amino-acid subfragment of the first 257 residues. The smallest Ras-binding region identified consisted of N-terminal residues 51 to 131, although stable expression of the domain and high-affinity binding were improved by the presence of residues 132 to 149. Deletion of the Raf zinc finger region did not reduce Ras-binding affinity, while removal of the first 50 amino acids greatly increased affinity. Phosphorylation of Raf[1-149] by protein kinase A on serine 43 resulted in significant inhibiton of Ras binding. demonstrating that the mechanism of cyclic AMP downregulation results through structural changes occurring exclusively in this small Ras-binding domain.

Critical binding and regulatory interactions between Ras and Raf occur through a small, stable N-terminal domain of Raf and specific Ras effector residues

Genetic and biochemical evidence suggests that the Ras protooncogene product regulates the activation of the Raf kinase pathway, leading to the proposal that Raf is a direct mitogenic effector of activated Ras. Here we report the use of a novel competition assay to measure in vitro the relative affinity of the c-Raf-1 regulatory region for Ras-GTP, Ras-GDP, and 10 oncogenic and effector mutant Ras proteins. c-Raf-1 associates with normal Ras and the oncogenic V12 and L61 forms of Ras with equal affinity. The moderately transforming mutant Ras[E30K31] also bound to the c-Raf-1 regulatory region with normal affinity. Transformation-defective Ras effector mutants Ras[N33], Ras[S35], and Ras[N38] bound poorly. In contrast, the transformation defective Ras[G26I27] and Ras[E45] mutants bound to the c-Raf-1 regulatory region with nearly wild-type affinity. A stable, high-affinity Ras-binding region of c-Raf-1 was mapped to a 99-amino-acid subfragment of the first 257 residues. The smallest Ras-binding region identified consisted of N-terminal residues 51 to 131, although stable expression of the domain and high-affinity binding were improved by the presence of residues 132 to 149. Deletion of the Raf zinc finger region did not reduce Ras-binding affinity, while removal of the first 50 amino acids greatly increased affinity. Phosphorylation of Raf[1-149] by protein kinase A on serine 43 resulted in significant inhibiton of Ras binding. demonstrating that the mechanism of cyclic AMP downregulation results through structural changes occurring exclusively in this small Ras-binding domain.

Effector domain mutations dissociate p21ras effector function and GTPase-activating protein interaction

The GTPase activity of p21ras is stimulated by GTPase-activating proteins (GAPs) such as p120GAP and the product of the neurofibromatosis 1 gene, which may negatively regulate p21 function. GAPs are also proposed effectors of ras. We have sought activating substitutions in c-H-ras in the region encoding the effector domain, on the rationale that such mutations would dissociate effector function from negative regulation by GAP. One such activating mutation, Pro-34-->Arg, encodes protein that is substantially bound to GTP in vivo. In vitro, this protein is not stimulated by GAPs, and its binding to p120GAP is grossly impaired. The results support the idea that the p21 structural requirements for effector function and GAP interaction are quite different and suggest that some molecule(s) other than p120GAP serves as the ras effector. In contrast to the results obtained with p120GAP, the Pro-34-->Arg p21 species is effectively coupled to the raf-1 product, as judged from electrophoretic mobility shifts of the Raf-1 phosphoprotein.

Effector domain mutations dissociate p21ras effector function and GTPase-activating protein interaction

The GTPase activity of p21ras is stimulated by GTPase-activating proteins (GAPs) such as p120GAP and the product of the neurofibromatosis 1 gene, which may negatively regulate p21 function. GAPs are also proposed effectors of ras. We have sought activating substitutions in c-H-ras in the region encoding the effector domain, on the rationale that such mutations would dissociate effector function from negative regulation by GAP. One such activating mutation, Pro-34-->Arg, encodes protein that is substantially bound to GTP in vivo. In vitro, this protein is not stimulated by GAPs, and its binding to p120GAP is grossly impaired. The results support the idea that the p21 structural requirements for effector function and GAP interaction are quite different and suggest that some molecule(s) other than p120GAP serves as the ras effector. In contrast to the results obtained with p120GAP, the Pro-34-->Arg p21 species is effectively coupled to the raf-1 product, as judged from electrophoretic mobility shifts of the Raf-1 phosphoprotein.

Requirement for posttranslational processing of Rac GTP-binding proteins for activation of human neutrophil NADPH oxidase

Rac1 and Rac2 are closely related, low molecular weight GTP-binding proteins that have both been implicated in regulation of phagocyte NADPH oxidase. This enzyme system is composed of multiple membrane-bound and cytosolic subunits and when activated catalyzes the one-electron reduction of oxygen to superoxide. Superoxide and its highly reactive derivatives are essential for killing microorganisms. Rac proteins undergo posttranslational processing, primarily the addition of an isoprenyl group to a carboxyl-terminal cysteine residue. We directly compared recombinant Rac1 and Rac2 in a human neutrophil cell-free NADPH oxidase system in which cytosol was replaced by purified recombinant cytosolic components (p47-phox and p67-phox). Processed Rac1 and Rac2 were both highly active in this system and supported comparable rates of superoxide production. Under different cell-free conditions, however, in which suboptimal amounts of cytosol were present in the assay mixture, processed Rac2 worked much better than Rac1 at all but the lowest concentrations. This suggests that a factor in the cytosol may suppress the activity of Rac1 but not of Rac2. Unprocessed Rac proteins were only weakly able to support superoxide generation in either system, but preloading of Rac1 or Rac2 with guanosine 5'-O-(3-thio-triphosphate) (GTP gamma S) restored activity. These results indicate that processing is required for nucleotide exchange but not for interaction with oxidase components.

Biology of the Rap proteins, members of the ras superfamily of GTP-binding proteins

Images

Functional interaction between p21rap1A and components of the budding pathway in Saccharomyces cerevisiae

The rap1A gene encodes a 21-kDa, ras-related GTP-binding protein (p21rap1A) of unknown function. A close structural homolog of p21rap1A (65% identity in the amino-terminal two-thirds) is the RSR1 gene product (Rsr1p) of Saccharomyces cerevisiae. Although Rsr1p is not essential for growth, its presence is required for nonrandom selection of bud sites. To assess the similarity of these proteins at the functional level, wild-type and mutant forms of p21rap1A were tested for complementation of activities known to be fulfilled by Rsr1p. Expression of p21rap1A, like multicopy expression of RSR1, suppressed the conditional lethality of a temperature-sensitive cdc24 mutation. Point mutations predicted to affect the localization of p21rap1A or its ability to cycle between GDP and GTP-bound states disrupted suppression of cdc24ts, while other mutations in the 61-65 loop region improved suppression. Expression of p21rap1A could not, however, suppress the random budding phenotype of rsr1 cells. p21rap1A also apparently interfered with the normal activity of Rsrlp, causing random budding in diploid wild-type cells, suggesting an inability of p21rap1A to interact appropriately with Rsr1p regulatory proteins. Consistent with this hypothesis, we found an Rsr1p-specific GTPase-activating protein (GAP) activity in yeast membranes which was not active toward p21rap1A, indicating that p21rap1A may be predominantly GTP bound in yeast cells. Coexpression of human Rap1-specific GAP suppressed the random budding due to expression of p21rap1A or its derivatives, including Rap1AVal-12. Although Rap1-specific GAP stimulated the GTPase of Rsr1p in vitro, it did not dominantly interfere with Rsr1p function in vivo. A chimera consisting of Rap1A1-165::Rsr1p166-272 did not exhibit normal Rsr1p function in the budding pathway. These results indicated that p21rap1A and Rsr1p share at least partial functional homology, which may have implications for p21rap1A function in mammalian cells.

Functional interaction between p21rap1A and components of the budding pathway in Saccharomyces cerevisiae

The rap1A gene encodes a 21-kDa, ras-related GTP-binding protein (p21rap1A) of unknown function. A close structural homolog of p21rap1A (65% identity in the amino-terminal two-thirds) is the RSR1 gene product (Rsr1p) of Saccharomyces cerevisiae. Although Rsr1p is not essential for growth, its presence is required for nonrandom selection of bud sites. To assess the similarity of these proteins at the functional level, wild-type and mutant forms of p21rap1A were tested for complementation of activities known to be fulfilled by Rsr1p. Expression of p21rap1A, like multicopy expression of RSR1, suppressed the conditional lethality of a temperature-sensitive cdc24 mutation. Point mutations predicted to affect the localization of p21rap1A or its ability to cycle between GDP and GTP-bound states disrupted suppression of cdc24ts, while other mutations in the 61-65 loop region improved suppression. Expression of p21rap1A could not, however, suppress the random budding phenotype of rsr1 cells. p21rap1A also apparently interfered with the normal activity of Rsrlp, causing random budding in diploid wild-type cells, suggesting an inability of p21rap1A to interact appropriately with Rsr1p regulatory proteins. Consistent with this hypothesis, we found an Rsr1p-specific GTPase-activating protein (GAP) activity in yeast membranes which was not active toward p21rap1A, indicating that p21rap1A may be predominantly GTP bound in yeast cells. Coexpression of human Rap1-specific GAP suppressed the random budding due to expression of p21rap1A or its derivatives, including Rap1AVal-12. Although Rap1-specific GAP stimulated the GTPase of Rsr1p in vitro, it did not dominantly interfere with Rsr1p function in vivo. A chimera consisting of Rap1A1-165::Rsr1p166-272 did not exhibit normal Rsr1p function in the budding pathway. These results indicated that p21rap1A and Rsr1p share at least partial functional homology, which may have implications for p21rap1A function in mammalian cells.

Specific isoprenoid modification is required for function of normal, but not oncogenic, Ras protein

While the Ras C-terminal CAAX sequence signals modification by a 15-carbon farnesyl isoprenoid, the majority of isoprenylated proteins in mammalian cells are modified instead by a 20-carbon geranylgeranyl moiety. To determine the structural and functional basis for modification of proteins by a specific isoprenoid group, we have generated chimeric Ras proteins containing C-terminal CAAX sequences (CVLL and CAIL) from geranylgeranyl-modified proteins and a chimeric Krev-1 protein containing the H-Ras C-terminal CAAX sequence (CVLS). Our results demonstrate that both oncogenic Ras transforming activity and Krev-1 antagonism of Ras transforming activity can be promoted by either farnesyl or geranylgeranyl modification. Similarly, geranylgeranyl-modified normal Ras [Ras(WT)CVLL], when overexpressed, exhibited the same level of transforming activity as the authentic farnesyl-modified normal Ras protein. Therefore, farnesyl and geranylgeranyl moieties are functionally interchangeable for these biological activities. In contrast, expression of moderate levels of geranylgeranyl-modified normal Ras inhibited the growth of untransformed NIH 3T3 cells. This growth inhibition was overcome by coexpression of the mutant protein with oncogenic Ras or Raf, but not with oncogenic Src or normal Ras. The similar growth-inhibiting activities of Ras(WT)CVLL and the previously described Ras(17N) dominant inhibitory mutant suggest that geranylgeranyl-modified normal Ras may exert its growth-inhibiting action by perturbing endogenous Ras function. These results suggest that normal Ras function may specifically require protein modification by a farnesyl, but not a geranylgeranyl, isoprenoid.

Specific isoprenoid modification is required for function of normal, but not oncogenic, Ras protein

While the Ras C-terminal CAAX sequence signals modification by a 15-carbon farnesyl isoprenoid, the majority of isoprenylated proteins in mammalian cells are modified instead by a 20-carbon geranylgeranyl moiety. To determine the structural and functional basis for modification of proteins by a specific isoprenoid group, we have generated chimeric Ras proteins containing C-terminal CAAX sequences (CVLL and CAIL) from geranylgeranyl-modified proteins and a chimeric Krev-1 protein containing the H-Ras C-terminal CAAX sequence (CVLS). Our results demonstrate that both oncogenic Ras transforming activity and Krev-1 antagonism of Ras transforming activity can be promoted by either farnesyl or geranylgeranyl modification. Similarly, geranylgeranyl-modified normal Ras [Ras(WT)CVLL], when overexpressed, exhibited the same level of transforming activity as the authentic farnesyl-modified normal Ras protein. Therefore, farnesyl and geranylgeranyl moieties are functionally interchangeable for these biological activities. In contrast, expression of moderate levels of geranylgeranyl-modified normal Ras inhibited the growth of untransformed NIH 3T3 cells. This growth inhibition was overcome by coexpression of the mutant protein with oncogenic Ras or Raf, but not with oncogenic Src or normal Ras. The similar growth-inhibiting activities of Ras(WT)CVLL and the previously described Ras(17N) dominant inhibitory mutant suggest that geranylgeranyl-modified normal Ras may exert its growth-inhibiting action by perturbing endogenous Ras function. These results suggest that normal Ras function may specifically require protein modification by a farnesyl, but not a geranylgeranyl, isoprenoid.