(Received for publication, January 2, 1997, and in revised form, February 6, 1997)
From the Department of Pharmacology, University of
North Carolina, Chapel Hill, North Carolina 27599, ¶ Department
of Basic Medical Sciences, Purdue University,
West Lafayette, Indiana 47907-1246, the
Moffitt Cancer
Research Center, University of South Florida, Tampa, Florida 33612,
and the ** Department of Chemistry, University of Pittsburgh,
Pittsburgh, Pennsylvania 15260
Presently, nothing is known about the function of the Ras-related protein Rheb. Since Rheb shares significant sequence identity with the core effector domains of Ras and KRev-1/Rap1A, it may share functional similarities with these two structurally related, yet functionally distinct, small GTPases. Furthermore, since like Ras, Rheb terminates with a COOH terminus that is likely to signal for farnesylation, it may be a target for the farnesyltransferase inhibitors that block Ras processing and function. To compare Rheb function with those of Ras and KRev-1, we introduced mutations into Rheb that generate constitutively active or dominant negative forms of Ras and Ras-related proteins and were designated Rheb(64L) and Rheb(20N), respectively. Expression of wild type or mutant Rheb did not alter the morphology or growth properties of NIH 3T3 cells. Thus, aberrant Rheb function is distinct from that of Ras and fails to cause cellular transformation. Instead, similar to KRev-1, co-expression of Rheb antagonized oncogenic Ras transformation and signaling. In vitro and in vivo analyses showed that like Ras, Rheb proteins are farnesylated and are sensitive to farnesyltransferase inhibition. Thus, it is possible that Rheb function may be inhibited by farnesyltransferase inhibitors treatment and, consequently, may contribute to the ability of these inhibitors to impair Ras transformation.
Mutated forms of the three ras genes (H-, K-, and N-ras) are associated with 30% of all human cancers and encode potent transforming and oncogenic mutant proteins (1). Normal Ras proteins function as GDP/GTP-regulated molecular switches (2). Guanine nucleotide exchange factors (SOS and RasGRF/CDC25) promote formation of the active, GTP-bound state (2-4), whereas GTPase activating proteins (p120- and NF1-GTPase activating proteins) promote formation of inactive, GDP-bound Ras (5). Mutated Ras proteins contain single amino acid substitutions (at residues 12, 13, or 61) that render the proteins insensitive to GTPase activating protein stimulation and, consequently, persist as constitutively activated proteins. Ras proteins serve as key intermediate relay switches in diverse signaling pathways that control cell growth and differentiation (6-8). Consequently, mutated Ras proteins cause constitutive, ligand-independent activation of these pathways, thereby promoting to the aberrant growth of tumor cells.
Ras proteins are prototypes for a large superfamily of Ras-related proteins (>60 mammalian members) that function as GDP/GTP-regulated molecular switches (2, 6, 9). However, despite their strong amino acid sequence identity with Ras proteins (30-55%), the majority of these small GTPases lack the potent transforming potential of Ras proteins. Exceptions include TC21/R-Ras2 (10, 11), R-Ras (12, 13), RhoA (14-18), RhoB (19), and Rac1 (17, 20), where constitutively activated versions of these Ras-related proteins can cause tumorigenic transformation of NIH 3T3 cells. The transforming activities of TC21 and R-Ras reflect the fact that these two Ras-related proteins share complete identity with the core Ras effector domain (Ras residues 32-40) and, consequently, may activate downstream signaling pathways in common with those that mediate Ras transforming potential (21). However, although the Raf-1 serine/threonine kinase is clearly a critical effector important for Ras signaling and transformation (22-27), neither TC21 nor R-Ras cause activation of Raf-1 (28, 29). Thus, despite possessing complete identity with the core Ras effector domain, these two Ras-related proteins must cause transformation by activation of Raf-independent effector pathways. Therefore, while the core effector domain sequences of Ras, TC21, and R-Ras are clearly critical for effector interactions, sequences flanking this core region are likely to influence specific effector interactions. Consistent with this, recent mutagenesis studies have extended the boundaries of the Ras effector domain to include Ras residues 25-45 (30, 31).
Although KRev-1/Rap1A also shares complete identity with the core Ras effector domain, it lacks any transforming potential and, instead, antagonizes the ability of Ras to transform cells (32-34). Since KRev-1 can interact with the Raf-1 serine/threonine kinase (35), as well as other candidate Ras effectors (e.g. RalGDS and related proteins) (36, 37), it is not clearly understood why KRev-1 is not transforming and how it antagonizes Ras function. One possibility is that KRev-1 interacts with, but fails to activate, Ras effectors important for Ras transformation. Thus, KRev-1 may form nonproductive complexes with Raf and other Ras effectors, thereby preventing their association with Ras. Consistent with this possibility KRev-1 fails to activate Raf or other downstream signaling activities that are stimulated by Ras.
Rheb (as homolog ighly nriched in rain) is a recently identified member of the Ras superfamily (38). The overall sequence of Rheb is distinct from other Ras-related proteins, but it shares strongest overall amino acid homology with human Rap2, yeast RAS1, and human H-Ras. Rheb also possesses sequence characteristics in common with Ras and Rap proteins. First, Rheb shows strong identity with the Ras and Rap core effector domains, suggesting that Rheb may share common effector interactions with Ras and KRev-1. The equivalent region of Rheb is quite similar to Ras with the first six amino acids being identical. Second, Rheb terminates with a CAAX motif (C, cysteine, A, aliphatic, and X, terminal amino acid) that is likely to signal for posttranslational modification by the C15 farnesyl isoprenoid (39-41). Except for the Ras proteins, Rap2B and two Rho family proteins (RhoB and RhoE) (42-44), all other prenylated Ras superfamily proteins are modified by the related C20 geranylgeranyl isoprenoid. Hence, Rheb function may be antagonized by FTIs that block Ras function (45).
In the present study, we introduced putative gain or loss of function mutations into Rheb and evaluated their biological properties in NIH 3T3 cells. We found that overexpression of wild type or mutant Rheb failed to cause morphologic or growth transformation of NIH 3T3 cells. Instead, like KRev-1, we observed that Rheb co-expression inhibited oncogenic Ras signaling and transformation and that Rheb complexed with Raf-1 in vitro. However, whereas KRev-1 is modified by the geranylgeranyl isoprenoid, like Ras, we determined that Rheb is modified by farnesylation, and this processing is blocked by FTI treatment. Finally, Rheb displays a subcellular location that is similar to that of Ras rather than the intracellular membrane location seen with KRev-1. Taken together, these results suggest that Rheb function is similar to that of KRev-1 rather than Ras and that FTI inhibition of Rheb function may be a component of FTI inhibition of Ras transformation.
We utilized
polymerase chain reaction-mediated DNA amplification to isolate
sequences corresponding to rheb from a randomly primed rat
fetal liver cDNA library (generous gift of T. Dawson). Five µl
(108 plaque-forming units/ml) of the library was subjected
to 30 cycles of amplification at an annealing temperature of 55 °C
using 5- and 3
-oligonucleotides that contain the coding sequences for the NH2- or COOH-terminal rat Rheb protein sequence (38),
together with flanking noncoding sequences that contain
BamHI restriction sites to facilitate subcloning into the
BamHI site of the Bluescript SKII plasmid vector. The
fidelity of the sequence was confirmed by dideoxy sequencing. The
rheb cDNA fragment was then subcloned into the
BamHI site of the pZIP-NeoSV(x)1 retrovirus for expression from the Moloney long terminal repeat promoter or the pCGN-hyg mammalian expression vector for expression of an
NH2-terminal hemagglutinin
(HA)1 epitope-tagged fusion protein
(HA-Rheb) from the cytomegalovirus promoter (46, 47). The
rheb cDNA sequence was also subcloned into the pGEX-2T
bacterial expression vector to encode a glutathione S-transferase (GST)-Rheb chimeric fusion protein.
Oligonucleotide-directed mutagenesis (Stratagene Chameleon Mutagenesis
kit) was used to generate mutant rheb or Krev-1
cDNA sequences. All mutated sequences were sequenced to confirm
fidelity.
NIH 3T3 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and plasmid DNA transfections were performed by calcium phosphate precipitation using procedures that we have described previously (48). Focus formation inhibition assays were performed by co-transfecting NIH 3T3 cells in triplicate with oncogenic H-Ras(12V) driven by its genomic promoter (pUC-rasH(12V)) and a 10-fold molar excess of each rheb expression construct. The number of transformed foci was determined after 16 days. To determine the consequence of exogenously introduced wild type or mutant Rheb protein expression on the growth of NIH 3T3 cells, cells stably transfected with expression vectors encoding wild type or mutant Rheb proteins were isolated by selection of transfected cells in growth medium supplemented with 500 µg/ml G418 (Geneticin; Life Technologies, Inc.). Multiple, G418-resistant colonies (>100) were pooled to establish cell lines stably transfected with each rheb expression vector and were then used for in vitro growth assays (on plastic or in soft agar) using procedures described previously (48). Growth curves were performed by plating cells at 4 × 104 cells per 60-mm dish.
Transcriptional Activation of Ras-responsive Promoter ElementsTransient transfection transcription assays were done to determine if Rheb can activate transcription or block oncogenic Ras stimulation of transcription, from the ets/AP-1 Ras-responsive promoter element using procedures described previously (49). Briefly, cultures of NIH 3T3 cells were transfected with the pBL4X-CAT reporter plasmid, where chloramphenicol acetyltransferase (CAT) gene expression is regulated by an ets/AP-1 containing promoter and the indicated Rheb and/or Ras expression plasmids. After 48 h the cells were lysed and assayed for CAT activity as described previously (49).
Generation of Rheb-specific Polyclonal AntiseraTo generate Rheb-specific antiserum, we utilized Rheb COOH-terminal sequences that are distinct from the sequences of corresponding COOH-terminal sequences of Ras, Rap, and other Ras-related proteins (38). Polymerase chain reaction-mediated DNA amplification was used to introduce sequences corresponding to the Rheb COOH-terminal 33 amino acid residues (designated Rheb/C) into the pGEX-2T bacterial expression vector for expression of the GST·Rheb/C fusion protein. Purified GST·Rheb/C protein was injected intramuscularly into rabbits, followed by a second injection after 6 weeks. Antiserum with reactivity against Rheb was used to isolate affinity purified anti-Rheb antibodies using GST·Rheb/C, and the resulting affinity purified anti-Rheb antibodies were used for Western blot analysis of Rheb protein expression.
Raf Binding AnalysisTo determine if Rheb could bind to activated human Raf-1, Sf9 insect cells were co-infected with recombinant baculovirus expressing full-length Raf-1 and activated Ras (generous gift of J. Strom), then lysed in Nonidet P-40-containing detergent lysis buffer, and clarified by centrifugation at 103 rpm. GST-fusion proteins containing Ras, KRev-1, and Rheb sequences were first preloaded with nonhydrolyzable GTP (GMPPCP) and then incubated with 10 µl of the Raf-1 containing insect cell lysate in phosphate-buffered saline, 10 mM dithiothreitol for 1 h at 4 °C. The GST-fusion proteins were then isolated and resolved on SDS-polyacrylamide gel electrophoresis. The separated proteins were transferred to Immobilon membrane, probed with an anti-Raf antibody (C-12, Santa Cruz Biotech), and visualized by enhanced chemiluminescence.
Xenopus Oocyte Lysate MAPK Activation AssayOocyte lysate and activated Ras protein were prepared as described previously (50). GST-Rheb and GST-KRev-1 fusion proteins were first preloaded with GMPPCP (10 µg), then added to the oocyte lysate together with 100 ng of GTP charged H-Ras, an ATP regeneration system, and incubated at 20 °C. Samples were taken at time 0 and at 2 h and then assayed for their ability to phosphorylate myelin basic protein as described previously (51).
Rheb Prenylation AnalysesFor in vitro prenylation analysis (52), 5 µg of recombinant H-Ras, GST-KRev-1, and GST-Rheb proteins were added to 45 µl of rabbit reticulocyte lysate (Promega), together with [3H]mevalonate, [3H]-farnesylpyrophosphate (FPP), or [3H]geranylgeranylpyrophosphate (GGPP) (DuPont NEN) (25, 2.5, and 2.5 µCi, respectively), either with or without the indicated concentrations of the FTI-277 FTI. We previously demonstrated that FTI-277 can inhibit potently farnesyltransferase in vitro (IC50 = 500 pM), is highly selective for farnesyltransferase over geranylgeranyltransferase I (IC50 = 50 mM), and can selectively inhibit prenylation of H-Ras, but not geranylgeranylated KRev-1, in whole cells (53). For prenylation analysis in NIH 3T3 cells, cells stably expressing wild type Ras or Rheb were incubated with varying concentrations of FTI-277 for 72 h. The cells were then fractionated into crude membrane (P100) and supernatant fractions (S100) as described previously (54), then resolved by SDS-polyacrylamide gel electrophoresis, and exposed for autoradiography to visualize incorporated label.
Immunofluorescence MicroscopyCells expressing HA epitope-tagged H-Ras, Rheb, or KRev-1 protein were grown on coverslips for 18 h at 37 °C and then fixed with 3.7% formaldehyde in phosphate-buffered saline and permeabilized with 0.5% Triton X-100 in Tris-buffered saline (150 mM NaCl, 50 mM Tris, pH 7.6, 0.1% NaN3). To determine the localization of the tagged proteins, anti-HA monoclonal antibody (Babco) was diluted 1:50 and used to stain the fixed cells for 1 h at 25 °C. The cells were then washed in phosphate-buffered saline and incubated with rhodamine-conjugated goat anti-mouse antibody (Cappel, Durham, NC) for 1 h at 25 °C. Coverslips were viewed on a Zeiss Axiophot microscope, and fluorescence micrographs were taken on T-max 400 film (Eastman Kodak Co.).
Rheb shows strong sequence homology with the Ras effector
domain (Fig. 1). To determine whether Rheb function
could regulate cell growth, we generated mutant Rheb proteins with
mutations analogous to those that convert Ras and other Ras-related
proteins into constitutively active or inactive proteins. Whereas
Rheb(64L) contains a mutation that is analogous to the Q61L mutation
that renders Ras proteins constitutively active and transforming (1), Rheb(20N) contains a mutation that is analogous to the S17N mutation that results in dominant negative mutants of Ras and Ras-related proteins (28, 55-57). To evaluate the consequences of Rheb protein expression on the growth properties of NIH 3T3 cells in
vitro, we isolated NIH 3T3 cells stably transfected with
expression vectors that encoded wild type or mutant Rheb proteins. To
avoid clonal variation differences that may complicate our comparisons,
we utilized early passages of pooled populations of multiple
G418-resistant colonies (>100) for our analyses.
We first determined if these G418-resistant cell populations expressed
the exogenously introduced Rheb sequences. Western blot analysis using
a rabbit anti-Rheb polyclonal antiserum that recognized the unique
COOH-terminal amino acid sequence of Rheb was done. Consistent with the
divergence of this portion of Rheb with other Ras-related proteins
(58), this antiserum did not recognize the closely related H-Ras,
KRev-1, or TC21/R-Ras2 proteins (data not shown). A 21-kDa protein was
detected in the control, empty vector-transfected cells, indicating
that Rheb is expressed in NIH 3T3 cells (Fig.
2A). The Rheb(WT)-transfected cells expressed a 2- to 3-fold elevation of protein at this position, whereas a slower
migrating band was detected in the Rheb(64L)-transfected cells. This
altered mobility is seen when the equivalent mutation is introduced
into Ras and Ras-related proteins. Cells transfected with Rheb(20N) did
not show elevated nor altered forms of protein. This is similar to
observations with 17N mutant versions of Ras and Ras-related proteins,
where expression is typically lower than that seen with the wild type
protein (29, 55). This may be due to a reduced stability of the protein
or, alternatively, to the fact that high level expression of this
mutant is not tolerated by NIH 3T3 cells.
We and others (48) have shown previously that constitutively activated mutants of Ras cause both morphologic and growth transformation of NIH 3T3 cells. In contrast, we found that cells transfected with wild type or mutant rheb sequences showed morphologies that were indistinguishable from that of untransformed NIH 3T3 cells (data not shown). Additionally, in contrast to observations with pZIP-ras(17N)-transfected NIH 3T3 cells (55), no significant reduction in the appearance of G418-resistant colonies was seen for pZIP-rheb(20N)-transfected cells (data not shown). We then determined if the growth rates of NIH 3T3 cells were altered by expression of exogenously introduced rheb sequences. As shown in Fig. 2B, the growth rates on plastic of wild type or 64L rheb-transfected cell populations were indistinguishable from those displayed by the empty vector-transfected cells. Furthermore, whereas Ras-transformed NIH 3T3 cells will proliferate in growth medium supplemented with low serum (1-5%) (data not shown) (48), all three Rheb-transfected populations showed a requirement for growth medium supplemented with 10% calf serum (Fig. 2C). Finally, we also tested their ability to form colonies in soft agar and found them to be negative (data not shown). Thus, in contrast to Ras, aberrant wild type or mutant Rheb function did not cause growth transformation, or inhibition, of NIH 3T3 cells.
Rheb Antagonizes Ras Signaling and TransformationSince Rheb
lacked the growth promoting activity seen with constitutively activated
mutants of Ras, TC21, or R-Ras, we next determined if Rheb was more
similar to KRev-1/Rap1A and, instead, could antagonize Ras function
(32, 33). For these analyses, we used the constitutively activated
KRev-1(63E) mutant as a positive control for these co-transfection
focus formation inhibition assays (33). Like the consequences of
co-transfection of pZIP-Krev-1(63E), co-transfection of pZIP
expression vectors encoding Rheb(WT) or Rheb(64L) also caused a greater
than 50% reduction in oncogenic H-Ras(12V) focus-forming activity
(Fig. 3A). In contrast, co-transfection with
Rheb(20N) did not cause a significant reduction in Ras focus-forming activity. This transformation inhibition was specific, since neither co-transfection with pZIP-rheb WT or 64L caused any
reduction in v-Mos focus-forming activity (Fig. 3B).
We next determined the ability of Rheb to modulate the ability of oncogenic Ras to stimulate transcription from a Ras-responsive promoter element. For this analysis, we used a reporter plasmid where CAT expression is controlled by a promoter containing the ets/AP-1 Ras-responsive promoter element. Whereas activated Ras caused enhanced transcription from this reporter plasmid (5-10-fold), no enhancement was seen with cultures when transfected alone with expression vectors encoding either wild type or mutant Rheb. As described previously, co-transfection of KRev-1(63E) caused a drastic impairment of oncogenic Ras(12V)-stimulated activity. Similarly, co-transfection of Rheb(WT) or Rheb(64L) caused a greater than 75% reduction in oncogenic H-Ras(12V)-induced transcription activation (Fig. 3C).
We also determined if Rheb could antagonize oncogenic Ras activation of
MAPKs. Xenopus frog oocyte lysates have been shown to
contain all the necessary cellular components to mediate oncogenic Ras
activation of p42/p44 MAPKs in vitro (59). Therefore, we determined the ability of Rheb to modulate this activity. As described previously, the introduction of recombinant Ras protein complexed with
nonhydrolyzable GTP caused the activation of MAPKs, as measured by the
ability to phosphorylate myelin basic protein (51). Addition of
recombinant Rheb(64L) or KRev-1(63E) protein, preloaded with a
nonhydrolyzable analog of GTP (GMPPCP), caused a near-complete reduction in Ras activation of MAPKs (Fig. 4). These
results, when taken together with the ability of Rheb to block Ras
transformation and transcription activation, suggest that Rheb shares
the same Ras inhibitory activity as KRev-1.
Rheb Complexes with Raf-1 in Vitro
Although Rheb shares
sequence identity with the core Ras effector domain, the failure of
Rheb to activate transcription from the ets/AP-1
Ras-responsive element, or MAPKs in oocyte lysates, suggests that Rheb
fails to activate Ras effectors important for mediating Ras function.
Instead, the ability of Rheb to inhibit Ras activity may be a
consequence of Rheb formation of inactive complexes with Ras effectors
such as Raf-1. KRev-1 can interact with, but not activate, Raf-1 (35).
To address this possibility, we determined the ability of chimeric
GST-fusion proteins containing activated Rheb(64L) or KRev-1(63E) to
bind insect cell-expressed Raf-1 in vitro. GTP-complexed
GST-Rheb(64L), Ras(61L), and KRev-1(63E), but not GST alone, formed
stable complexes with Raf-1 in vitro (Fig.
5). Thus, like KRev-1, Rheb may antagonize Ras function by forming inactive complexes with key Ras effector proteins such as
Raf-1.
Rheb Is Modified by the Isoprenoid Farnesyl
Like Ras and
KRev-1, Rheb terminates with a COOH-terminal CAAX motif (38)
(Fig. 1). Since the nature of the terminal amino acid typically
dictates whether the CAAX sequence signals for modification
for farnesylation (where X is Ser, Met, or Gln) or for
geranylgeranylation (where X is Leu or Phe) (40), we
suspected that Rheb may be farnesylated. To determine the status of
Rheb processing, we utilized a rabbit reticulocyte lysate assay using recombinant Rheb protein (52). Recombinant H-Ras and KRev-1 proteins
were used as controls for farnesylation and geranylgeranylation, respectively. When incubated with tritiated mevalonate, the essential precursor for the farnesyl- and geranylgeranyl-pyrophosphate (FPP and
GGPP) that are the sources for the isoprenoid groups used for protein
prenylation, all three proteins showed incorporation of label (Fig.
6A). When radiolabeled FPP was used, only
H-Ras and Rheb showed significant incorporation of label. In contrast, when [3H]GGPP was used, only KRev-1 was labeled.
Furthermore, addition of FTI-277 (from 0.1 to 5 µM)
blocked both H-Ras and Rheb farnesylation in the reticulocyte lysate
assay (Fig. 6B). Thus, like Ras, Rheb is preferentially
modified by farnesylation in vitro.
To determine whether Rheb was also farnesylated in vivo, we determined if Rheb membrane association could be blocked in cells treated with FTI-277. In control, untreated cells, H-Ras was found to be present exclusively in the P100 fraction, whereas comparable levels of Rheb were present in both the P100 and S100 fractions. Treatment of cells with 10 µM FTI-277 caused both H-Ras and Rheb to shift to the S100 fraction. These results suggest that Rheb is farnesylated in vivo and that this modification promotes Rheb membrane association.
Rheb Subcellular LocalizationWhereas Ras proteins are
associated with the inner face of the plasma membrane (40), KRev-1 has
been localized to intracellular membrane compartments (Golgi and
endoplasmic reticulum) (60, 61). To compare the subcellular location of
Rheb with Ras and KRev-1, we established NIH 3T3 cells that stably
expressed HA epitope-tagged versions of wild type Rheb, H-Ras, and
KRev-1. Indirect immunofluorescence analyses were then done using the anti-HA epitope antibody (Fig. 7). In agreement with
previous studies, we found that KRev-1 showed an intracellular membrane localization. In contrast, both Ras and Rheb showed a plasma membrane localization, with an enrichment of expression in apparent membrane ruffles. Thus, the subcellular location of Rheb is more similar to that
of Ras.
rheb was originally identified as a gene that was rapidly and transiently induced by seizures in the hippocampus, by N-methyl-D-aspartate-dependent synaptic activity in the long term potentiation paradigm, by serum stimulation of Balb/c 3T3 fibroblasts, and by growth factor (nerve, epidermal, and fibroblast growth factors) stimulation of PC12 pheochromocytoma cells (38). Since Rheb protein shows strongest amino acid identity with Rap and Ras proteins, it has been suggested that Rheb may play an important role in receptor-mediated signaling pathways in fibroblast and neuronal cells. To evaluate Rheb function, we generated mutants of Rheb that harbored mutations analogous to those that result in either constitutively activated or dominant negative inhibitory mutants of Ras and Ras-related proteins. In contrast to Ras, we found that aberrant Rheb expression did not cause morphologic or growth transformation of NIH 3T3 cells. Instead, we found that Rheb inhibited oncogenic Ras signaling and transformation. Thus, Rheb exhibited the anti-Ras activity similar to that seen with KRev-1/Rap1A rather than the growth promoting actions of Ras. However, like Ras, we found that Rheb was modified by the farnesyl isoprenoid and that its membrane association was inhibited by FTI treatment.
Rheb exhibits strongest overall sequence similarity with yeast RAS1 and human Rap2 (43 and 38% identity, respectively) (38). In particular, Rheb shows strong sequence identity with residues that define the core effector domain of Ras (Ras residues 32-40) and KRev-1. Since rheb mRNA (38) and protein (data not shown) expression was transiently induced by growth promoting extracellular stimuli, we speculated that Rheb may exhibit the growth promoting activity associated with Ras proteins. However, NIH 3T3 cells stably transfected with wild type Rheb or mutant Rheb did not undergo the morphologic or growth (no growth in low serum or in soft agar) transformation that is seen with cells expressing constitutively activated mutants of Ras. Furthermore, a putative dominant negative mutant of Rheb, analogous to the Ras(17N) dominant negative mutant, also did not show the potent growth inhibitory activity of NIH 3T3 cells that is caused by Ras(17N) (55). Thus, Rheb does not appear to influence the growth properties of fibroblasts. However, whether Rheb is a positive regulator of growth or differentiation of neural cells, such as PC12 cells, remains a possibility. For example, KRev-1 has been shown to be growth stimulatory when expressed in PC12, but not NIH 3T3, cells (62).
Instead, we found that both wild type and a putative constitutively activated mutant of Rheb (64L) showed the ability to antagonize oncogenic Ras stimulation of transcription from Ras-responsive promoter elements, activation of MAPKs in Xenopus oocyte lysate in vitro, and focus formation in NIH 3T3 transformation assays. Thus, like KRev-1, Rheb may function as a negative regulator of Ras function. The comparable ability of wild type and Rheb(64L) to antagonize Ras function is not entirely surprising, since Rheb possesses an arginine and serine at residues homologous to Ras residues 12 and 13 (Rheb residues 15 and 16). Ras and Rap proteins contain glycine residues at both positions and either substitution renders Ras proteins constitutively active and transforming (1). Thus, wild type Rheb may already exist as a constitutively active protein and the 64L mutation may not have caused any significant shift in Rheb GDP/GTP cycling. In contrast, the 17N mutant of Rheb, which is likely to render the protein constitutively GDP bound (55), lacked the ability to antagonize Ras. Thus, we speculate that the GTP-complexed form of Rheb is responsible for its anti-Ras activity.
The strong sequence identity seen between the core Ras effector domain and the homologous region of Rheb suggests one possible mechanism for Rheb antagonism of oncogenic Ras function. Since the Ras effector domain is critical for Ras binding of Raf and other candidate effectors (21), Rheb may form nonproductive complexes with Ras effectors and, consequently, prevent oncogenic Ras signaling and function. Support for this possibility is provided by our observation that, like KRev-1, Rheb can complex with Raf-1 in vitro and that Rheb inhibited oncogenic Ras activation of MAPKs, which is a Raf-dependent activity. Furthermore, Rheb inhibited Ras stimulation of transcription from the ets/AP-1 Ras-response promoter element, suggesting that Rheb can complex with Raf-1 in vivo. Whether Rheb can complex with other candidate Ras effectors, such as RalGDS (36, 63, 64) or phosphatidylinositol 3-phosphate kinase (65), will be interesting to determine. However, it remains possible that Rheb may interact with Rheb-specific effectors and mediate its own signaling activities that indirectly counter the transforming actions of oncogenic Ras and shift the balance of activity to reverse Ras transformation.
Since Rheb antagonized oncogenic Ras transforming activity, yet showed no growth inhibitory activity in untransformed cells, it suggests that Rheb specifically antagonizes oncogenic, but not normal, Ras. This preferential antagonism is similar to that seen with KRev-1, which also did not show inhibitory activity of untransformed NIH 3T3 cells (32). Similarly, the cytosolic Ras(61L,186S) dominant protein has also been shown to selectively antagonize oncogenic versus normal Ras function (66). Thus, if the inhibitory actions of Rheb, KRev-1, and Ras(61L,186S) are due to competition for Ras effectors important for transformation, these effectors may not be critical for normal Ras function. Alternatively, their inhibitory actions may be incomplete and simply cause a reduction, but not loss, of oncogenic Ras function to levels that are no longer sufficient to cause transformation. Support for this second possibility is provided by observations that oncogenic Ras transformation of Rat-1 rodent fibroblast cells is dependent on a threshold level of expression (67).
How KRev-1 antagonizes Ras transformation is still not fully understood. Furthermore, it is still unclear whether antagonism of Ras function is the physiological role of KRev-1, since very high levels of KRev-1 expression were required to antagonize Ras transformation (32). At present, no clear function for KRev-1 in non-phagocytic cells has been defined (68). Therefore, while we have demonstrated that Rheb can antagonize Ras function, whether this is the true physiological role for this Ras-related protein or whether it plays a role distinct from Ras remains to be determined. An identification of Rheb-specific effector targets may provide important clues for Rheb function.
FTI inhibition of K-Ras transformation can occur at concentrations that are not sufficient to impair K-Ras prenylation (69, 70). Since there is evidence that K-Ras4B may undergo modification by a different isoprenoid lipid (geranylgeranol) when in the presence of FTIs (71, 72), it has been speculated that FTI inhibition of K-Ras4B transformation may occur as a consequence of FTI action on other farnesylated proteins. Two key features of such an FTI target are 1) that it undergoes modification by farnesylation which can be inhibited at FTI concentrations that inhibit K-Ras4B transforming activity and 2) that the function of this protein could modulate Ras transforming potential. Our observation that Rheb is also farnesylated, is sensitive to FTI inhibition, and can modulate Ras transforming activity implicates it as a possible target for FTI inhibition of Ras transformation. It was previously speculated that Rheb might be a brain-specific protein (38), in which case it would not be a good candidate for mediating responses to FTI compounds in tumors of non-neural origin. However, we have now examined a panel of human tissues by Western blot and found Rheb to be ubiquitously expressed in most non-muscle tissue, with particularly high levels being present in the testes and ovary (data not shown). Therefore, the expression of Rheb in vivo is compatible with a role in the response to FTI compounds of a broad range of tumors.
To further evaluate such a role for Rheb in FTI action, it will be interesting to determine if a geranylgeranylated version of Rheb, which should exhibit a farnesylation-independent membrane association, decreases the ability of FTIs to inhibit K-Ras4B transforming activity.
In summary, our studies show that Rheb function is distinct from Ras proteins and instead is more similar to KRev-1 when evaluated in NIH 3T3 cells. However, it remains possible that Rheb may exhibit growth or differentiation promoting activity in other cell types. Thus, while our studies establish that Rheb can function more like Rap than like Ras in fibroblasts, Rheb function in other cell types may provide additional clues to determine the physiological role of Rheb. Finally, since Rheb is farnesylated, whether the biological actions of FTIs involve alterations in Rheb function await further understanding of the role of Rheb in cell physiology.
We thank Adrienne Cox for critical comments, Carol Martin and Sarah Johnson for excellent technical assistance, and Jennifer Parrish for preparation of the figures and manuscript.