Drosophila Rheb GTPase is required for cell cycle progression and cell growth

Regulation of cell size and number is required to establish a normal,correctly proportioned adult body size(Coelho and Leevers, 2000; Johnston and Gallant, 2002; Kozma and Thomas, 2002). Although a great deal is known about cell cycle regulation, little is known about the regulation of cell size. Recently, it has become clear that,although growth (increase in mass) usually occurs concomitantly with cell division, it is actually a distinct process(Edgar, 1999; Coelho and Leevers, 2000). Work in Drosophila has made it possible to characterize the separate effects of genes on these two processes during development of the whole organism (Edgar, 1999; Johnston and Gallant, 2002). The genes string (cdc25), cyclin E and E2F, for example, have been shown to be required for cell cycle progression but not for cellular growth(Neufeld et al., 1998).

Recently, the insulin and TOR (target of rapamycin) signaling pathways have been shown to regulate growth in Drosophila(Kozma and Thomas, 2002; Oldham and Hafen, 2003). Mutants in these pathways affect total organ and body size, as well as individual cell size. A downstream target of both pathways is dS6K, mutations in which affect only cell size, not cell number(Montagne et al., 1999). Most known components of the insulin and TOR signaling pathways affect both cell size and number; current understanding is that these effects on cell number are primarily via their effects on cell growth(Coelho and Leevers, 2000; Johnston and Gallant, 2002; Kozma and Thomas, 2002).

The Ras-like gene Rheb (Ras homolog enriched in brain) was recently shown to be required for growth of the yeast Schizosaccharomyces pombe. The rheb mutants (rhb1) arrest in G0/G1 as small, rounded cells (Mach et al.,2000; Yang et al.,2001), suggesting a role for Rheb in cell cycle progression and cell growth. Highly conserved Rheb genes have been described throughout the metazoa (Yamagata et al., 1994; Reuther and Der,2000; Urano et al.,2000; Urano et al.,2001; Im et al.,2002; Panepinto et al.,2002). To investigate the cellular role of Rheb, we have taken advantage of the unique suitability of Drosophila for both genetic and biochemical studies. Here, we show that Drosophila dRheb has both GTP binding and GTPase activities. Overexpression of dRheb results in tissue overgrowth and increased cell size in the whole organism, and transition into S phase and cell growth in culture. Conversely, reduction of dRheb activity results in reduced tissue growth and smaller cell size in the whole organism, as well as a G1 arrest and smaller cell size in culture. The results of treating S2 cells and flies with rapamycin, an inhibitor of TOR, suggest that the effects of dRheb are probably mediated by dTOR.

Flies were cultured at 25°C on standard food, in a yw or w genetic background. Overexpression of dRheb in the embryonic hindgut, in the third instar larval eye disc posterior to the morphogenetic furrow, in the dorsal compartment of the wing disc, in the posterior compartment of the wing disc or in clones of cells in the early eye imaginal disc was effected by combining the dRheb^AV4 allele or a UASdRheb construct with bynGAL4,UASGFP (Iwaki and Lengyel, 2002), GMR-GAL4 (III)(Ellis et al., 1993; Tapon et al., 2001), apGAL4 (Montagne et al.,1999), enGAL4(Neufeld et al., 1998) or eyFLP;act<y⁺<GAL4,UAS-GFP(Pignoni and Zipursky, 1997). The dRheb coding sequence was amplified from the GH15143 expressed sequence tag clone using primers 5′-GCTAAGATCTATGCCAACCAAGGAGCGCCACATA and 5′-GCGCCTCGAGTTACGATACAAGACAACC, and subcloned into XhoI/BglII-digested pUAST(Brand and Perrimon, 1993) to generate the UASdRheb construct. Germ-line transformation was by standard procedures (Ashburner,1989), as described previously(Green et al., 2002). Loss-of-function clones in the eye disc were produced by combining eyFLP with FRT(82B),w⁺ M(3)95A and FRT(82B)dRheb^AV4 [generated by recombination between dRheb^AV4 and FRT(82B),Sb(Xu and Rubin, 1993)]. Excision of the P-element in dRheb^AV4 was effected by introduction of transposase using Ki[δ2-3].

Genomic DNA at the 3′ end of the P{y⁺,5XUAS}insert in dRheb^AV4 was isolated by inverse PCR(Sullivan et al., 2000). Amplified product was cloned into the pCR4-TOPO vector (Invitrogen) and sequenced. The insertion site was identified by FlyBLAST(http://www.fruitfly.org/blast/). For determination of lethal phase, eggs were collected from dRheb^AV4/TM6-GFP adults on yeasted apple juice agar plates for 2 hours. After aging for various times, larvae were washed onto fresh plates and scored under a fluorescence-equipped dissecting microscope for genotype, based on the presence or absence of green fluorescent protein (GFP). For effect of rapamycin on time of eclosion, 2-hour collections of 100 embryos each were placed in multiple vials containing standard food with and without 1μM rapamycin (Calbiochem).

In situ hybridization to whole-mount embryos was performed as described previously (Tautz and Pfeifle,1989). Digoxigenin RNA probes (Roche Molecular Biochemicals) were generated from cDNA templates of the dRheb open reading frame. BrdU incorporation into embryos for 1 hour was carried out as described(Smith and Orr-Weaver, 1991)and detected with anti-BrdU antibody (1:40, Becton Dickinson) using standard techniques (Ashburner, 1989). Scanning electron microscopy and fixation, embedding and sectioning of adult eyes were performed as described (Wolff and Ready, 1991; Sullivan et al., 2000). Light microscopy was performed with a Zeiss Axiophot and images were acquired with a Sony DKC-500 digital camera and processed with Adobe Photoshop. Cell size and number in the wing were determined as described by Montagne et al. (Montagne et al.,1999). Wing areas were measured using Image J software on 10-12 wings for each genotype; anterior was defined by wing vein L2 and anterior wing margin, posterior by vein L3 and posterior wing margin. Wing hair density was determined by counting wing hairs in a 0.0275 mm² area in the anterior and in a 0.0544 mm² area in the posterior of each wing. Sample standard error was ∼5% of the mean for all measurements.

S2 cells (1×10⁶) were seeded and maintained in Schneider's Drosophila medium (Gibco) supplemented with 10% fetal bovine serum(Gibco) at 25°C. Transfection was carried out using Cellfectin(Invitrogen) with 1 μg each of plasmids pPacFLAG-dRheb (expressing FLAG-dRheb using the actin 5C promoter) and pRmHa-GFP (transfection marker, expressing GFP using metallothionein promoter). For RNAi, dRheb template containing T7 promoter sequence was generated by PCR from the cDNA; double-stranded RNA (dsRNA) corresponding to the entire coding sequence (546 bp) was synthesized using T7 Megascript kit (Ambion). S2 cells were treated with 10 μg dRheb dsRNA with and without pPacFLAG-dRheb according to Clemens et al.(Clemens et al., 2000). For cell cycle analysis by fluorescence-activated cell sorting (FACS), transfected cells were fixed in 1% formalin/PBS and 70% ethanol, whereas RNAi-treated cells were fixed in 70% ethanol alone. After fixing, cells were stained with 25 μg ml^-1 propidium iodide and FACS was performed on a Becton Dickinson FACScan. For immunoblotting, total cell extracts were prepared in insect cell lysis buffer (BD Pharmingen) and electrophoresed on a 12% SDS polyacrylamide gel; after blotting, proteins were immunodetected with anti-FLAG antibody (Sigma).

dRheb coding sequence, amplified from a plasmid cDNA library(provided by F. Laski, University of California, Los Angeles) by PCR, was inserted into vector pET28a(+) (Novagen). Total cell lysates of bacteria expressing His-tagged dRheb, induced by IPTG, were prepared in lysis buffer(20 mM sodium phosphate, pH 7.4, 0.5 M NaCl, 0.5 mM MgCl₂, 10 μM GTP) using a French press. Protein was purified using Probond resin(Invitrogen) and stored in 50% glycerol at -20°C. GTP binding was determined by nitrocellulose filtration assay(Finlin et al., 2001). His-dRheb (2 μg) was incubated with [γ-³⁵S]GTP (1250 Ci mmol^-1, Amersham) in binding buffer (20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM DTT, 1 mM EDTA, 40 μg ml^-1 bovine serum albumin) in the presence of 1 mM MgCl₂ at 37°C. At various time points,samples were diluted into ice-cold wash buffer (20 mM Tris-HCl, pH 7.5, 50 mM NaCl, 10 mM MgCl₂, 1 mM DTT) and nitrocellulose-bound radioactivity was determined by scintillation counting. Specificity of nucleotide binding was assessed with [γ-³⁵S]GTP in the presence of a 20-fold excess of unlabeled competing nucleotide for 30 minutes, normalized to binding in the absence of competitor. GTPase activity(Tanaka et al., 1991) was assayed by measuring loss of ³²P from His-dRheb preloaded with[γ-³²P]GTP (3000 Ci mmol^-1) in binding buffer containing 0.1% Triton X-100, 4 μM GTP and 1 mM MgCl₂ at 37°C for 30 minutes. Unlabeled ATP (5 mM) was included to inhibit nonspecific phosphatase activity. Hydrolysis was initiated by adjusting the MgCl₂ concentration to 10 mM. Reactions were terminated at the indicated time points and bound radioactivity was assessed by nitrocellulose binding. All biochemical assays described here were carried out in triplicate.

Ectopic expression screens can be used to identify and characterize genes that, when expressed at a particular time or region, generate a phenotype of interest (Rorth et al., 1998). Lines containing random insertions of the Drosophila mobile element P(engineered to contain the 5XUAS binding site for the yeast transcription factor GAL4) are screened for phenotype in the presence of a particular GAL4 driver. Because P-elements usually insert near the transcription initiation site, this allows ectopic expression and/or overexpression of a particular gene (Rorth et al., 1998; Toba et al.,1999; Duchek and Rorth,2001).

We screened approximately 5000 P{y⁺,5XUAS} lines for lethality in combination with bynGAL4, which drives expression specifically in the developing posterior gut(Iwaki and Lengyel, 2002), to identify genes affecting hindgut morphogenesis. Among the roughly 7% of these lines that were lethal with bynGAL4, we identified one (designated AV4) that gave a dramatically enlarged (both in width and length) hindgut in the first instar larva (Fig. 1B,D). As determined by inverse PCR, this chromosome carries an insertion of the P-element in the 5′ UTR of dRheb; this allele is designated dRheb^AV4(Fig. 1A). In situ hybridization confirms that, as expected from the orientation of the inserted 5XUAS, combining dRheb^AV4 with bynGAL4 causes overexpression of dRheb mRNA in the hindgut(Fig. 1C). When the P-element insert of dRheb^AV4 is mobilized by transposase and precisely excised (confirmed by PCR), combination of the excision chromosome with bynGAL4 has no effect on hindgut size (data not shown), again confirming that the effects observed in combination with the bynGAL4 driver are due to the P{y⁺,5XUAS}insertion in dRheb.

The dRheb gene is predicted to encode a 182 amino acid protein that contains G-boxes characteristic of the Ras superfamily of G-proteins. Fig. 2A shows the predicted amino acid sequence of dRheb, together with those of the yeast and human Rhebs. Yeast and fly each have a single Rheb gene; the most similar human gene is hRheb1 (Gromov et al., 1995; Mizuki et al.,1996), which maps to chromosome 7(Mizuki et al., 1996). We have identified an additional, more divergent human Rheb gene, designated hRheb2, which maps to chromosome 12. The overall sequence identity between Drosophila and S. pombe Rheb is 51%, and that between Drosophila and human Rheb1 is 63%(Fig. 2A). Consistent with this high sequence conservation, dRheb, hRheb1 and hRheb2 can replace the function of S. pombe Rheb (data not shown).

Recombinant dRheb binds [γ-³⁵S]GTP (a non-hydrolysable GTP analog) in a manner that is both time and nucleotide-concentration dependent(Fig. 2B, data not shown). The binding of GTP to dRheb is enhanced by addition of Mg²⁺ (data not shown) and is specific for guanine nucleotides, with a preference for the binding of GTP over GDP (Fig. 2C). dRheb also exhibits intrinsic GTPase activity, as shown by the slow but significant loss of radioactivity from dRheb preloaded with[γ-³²P]GTP (Fig. 2D). Hydrolysis of GTP to GDP is confirmed by the appearance of ³²P in GDP (revealed by thin layer chromatography, data not shown)after incubation of dRheb with [α-³²P]GTP. Our detection of Rheb GTPase activity shows that the presence of arginine at the position corresponding to amino acid 12 of Ras [at which most Ras-family members have glycine (Bourne et al., 1991)]does not eliminate GTP-hydrolysing activity of the protein.

Overexpression of dRheb leads to enlargement of the larval hindgut, so we tested the generality of this effect by driving dRhebexpression in other tissues. In the larval salivary glands, dRheboverexpression also results in significant enlargement (data not shown). When dRheb is overexpressed in the developing eye imaginal disc, using the GMR-GAL4 driver, dramatically enlarged eyes are seen in the adult(Fig. 3A-F). Similarly,overexpression of dRheb in clones of cells in the eye and antennal discs, using the 'flip-out' method(Pignoni and Zipursky, 1997)and eyFLP, results in an enlarged eye, antenna and head(Fig. 3G-J). When compared with control eyes (Fig. 3K), the ommatidia of these eyes are much larger than normal, are not organized into the normal hexagonal array and are frequently flanked by extra bristles(Fig. 3L). The larger overall eye and ommatidial size, and the extra bristles further support the notion that dRheb overexpression promotes growth. The fact that bristles in the dRheb-overexpressing eyes are larger (thicker) than normal(Fig. 3K,L), together with the fact that each bristle is derived from a single cell(Wolff and Ready, 1993),suggests that dRheb might promote growth at the level of the individual cell.

Overexpression of dRheb also promotes growth in the wing, an organ consisting of two opposed, flat, single-layered epithelial sheets. Overexpression of dRheb in the dorsal epithelial compartment causes the wing to curve downwards (Fig. 4C,D), presumably owing to an increase in area of the dorsal compartment. When dRheb is overexpressed in the posterior compartment of the wing, the posterior area is increased by 36%(Fig. 4A,B), further supporting a role for dRheb in growth. To distinguish between effects on cell number versus cell size, we counted wing hairs (each of which is produced by a single cell) in normal and dRheb-overexpressing compartments(Fig. 4E,F). Because the cell density in the dRheb-overexpressing compartment is only 67% of that in the wild-type compartment, we conclude that dRheb-overexpressing cells are significantly larger (i.e. each cell occupies more area) than wild-type cells. Wing hair density, combined with compartment area measurements, shows that total cell number in the posterior compartment is the same whether or not dRheb is being overexpressed. Thus, in the wing, dRheb overexpression results in an increase in cell size but not cell number. This is similar to what has been described for overexpression of activated Ras1 and Myc; in Drosophila, these genes are believed to affect cell cycle progression via their primary function as regulators of cell growth (Johnston et al., 1999; Prober and Edgar,2000; Prober and Edgar,2001).

In the developing embryo, dRheb expression correlates with DNA replication. Immediately after fertilization, as rapid, syncytial nuclear divisions take place (Foe et al.,1993), dRheb mRNA is present at a high, uniform level. This level decreases as rates of nuclear division slow, begins increasing again during stage 11 and becomes significantly higher in tissues undergoing endocycles (i.e. midgut) and mitoses (i.e. central nervous system). Later in embryogenesis (stages 13-16), as revealed by in situ hybridization and BrdU incorporation, the same regions that show strong dRheb expression are also carrying out DNA synthesis (Fig. 5A, Flybase). Thus, embryonic mRNA expression patterns and levels are consistent with a role for dRheb in promoting S-phase.

To characterize the effect of dRheb further at the cellular level, Drosophila S2 cells were co-transfected with vectors expressing dRheb and GFP (a marker for transfection). Examination of the cell cycle profile of transfected (GFP plus dRheb-overexpressing) cells by FACS reveals that increased levels of dRheb result in an increase in the proportion of cells in S-phase (Fig. 5B). In addition, analysis of cell size by forward scatter analysis reveals that S2 cells overexpressing dRheb are slightly larger than control (GFP alone) cells (Fig. 5C).

The overexpression studies described above show that dRheb can promote two processes: increase in cell size and cell cycle progression. The correlation between dRheb expression and DNA synthesis in the embryo,and the effect of dRheb overexpression in cultured cells both suggest that dRheb promotes progression into S-phase. The increased size of cultured cells, eye bristles and wing cells overexpressing dRhebsuggests that dRheb promotes increase in mass of individual cells(cell growth). These two distinct effects of dRheb raise the question of whether dRheb affects the cell cycle and cell growth independently or via a pathway that coordinates both of these processes.

Although the results of overexpression studies described above suggest a role for dRheb in regulation of cell growth and cell cycle progression, removal of gene activity is necessary unequivocally to establish required function. Of great utility is the fact that the dRheb^AV4 allele, in addition to allowing ectopic expression when combined with a GAL4 driver, is also homozygous lethal. Precise excision of the P-element in the dRheb^AV4chromosome results in a reversion to viability, confirming that the dRheb^AV4 lethal phenotype is due to disruption of dRheb activity resulting from the P-element insertion. To determine the lethal phase of dRheb^AV4 homozygotes, we collected eggs for 2 hours from heterozygous parents and determined the number of surviving homozygous mutant larvae relative to heterozygous sibs as a function of time (Fig. 6A). Embryos homozygous for dRheb^AV4 hatch into first instar larvae that grow very slowly, move lethargically, and die by 72 hours without molting into second instar larvae(Fig. 6B,C). dRheb is thus required for growth of the whole organism.

To assess the requirement for dRheb in individual cells, we made clones of cells lacking dRheb in the eye. For this purpose, we recombined FRT(82B) onto the dRheb^AV4chromosome; combination of this chromosome with eyFLP should, in principle, allow generation of w;dRheb^-/- clones in the background of a w⁺;dRheb^+/- eye. In this background, however, dRheb^-/- clones were not detected,similar to what has been reported for clones lacking Ras1 function(Prober and Edgar, 2000). In both cases, the absence of detectable w clones is probably due to cell competition, a process in which faster-growing cells out-compete slowly growing cells, which are then eliminated by apoptosis(Morata and Ripoll, 1975). To give dRheb^-/- clones a growth advantage, we generated them in a Minute (M^+/-) background (Minutegenes regulate protein synthesis; thus M^+/+ cells grow more rapidly than surrounding M^+/- cells)(Andersson et al., 1994). Heads and eyes containing multiple dRheb^-/-M^+/+ clones are dramatically smaller than those containing dRheb^+/+clones (Fig. 7H,I). Sections of these eyes show that the dRheb^-/- ommatidia are smaller overall because they are composed of dramatically smaller cells(Fig. 7J cf. C).

Two important inferences can be drawn from analysis of these dRhebloss-of-function clones. First, because eyes bearing multiple dRheb^-/- clones are smaller than wild-type, dRhebmust play a required role in tissue growth. Second, the smaller size of individual dRheb^-/-M^+/+ cells(Fig. 7J) indicates that dRheb is required in a cell autonomous manner for cell growth(increase in mass).

To investigate whether these inferred roles of dRheb can be demonstrated at the cellular level, we used dsRNA(Clemens et al., 2000) to inhibit Rheb function in Drosophila S2 cells. Addition of dsRNA corresponding to the entire coding sequence of dRheb to the culture medium almost completely inhibits expression of FLAG-tagged dRheb(Fig. 8A). Characterization of cell cycle profiles by FACS showed a dramatic increase in the proportion of cells in G1-phase by day four to five in dRheb dsRNA-treated cells;this effect persisted to at least day 8(Fig. 8B; notice the increased proportion of cells in G2-phase in controls). In addition to having an effect on the cell cycle, inhibition of dRheb also has a significant effect on cell growth. Forward scatter analysis reveals a dramatic reduction in cell size after the addition of dRheb dsRNA(Fig. 8C). Both the diminution in cell size and the accumulation of cells in G1-phase after dRhebinhibition follow roughly the same time course (i.e. both are maximal by day five) (Fig. 8B,C).

From the loss-of-function studies in the whole organism and in cultured cells described above, we conclude that dRheb affects both cell growth (mass increase) and cell cycle progression promoting transition of cells from G1 to S phase.

Control of cell growth and cell cycle progression can, in principle, be regulated by parallel independent pathways or through a signal that coordinates both (Coelho and Leevers,2000). In both yeast and Drosophila, mutations in cell-cycle-specific genes (such as cyclin E) result in cell cycle arrest with an associated increase in cell size owing to continued cell growth(Johnston et al., 1977; Neufeld et al., 1998),although overexpression of these genes results in smaller cells. Because loss of dRheb function in both cultured cells and in the whole organism results in reduced cell size, it is likely that dRheb coordinates cell cycle and cell growth. We therefore considered the possibility that dRheb might impinge on the insulin and TOR signaling pathways, which are major contributors to the regulation of cell growth in both Drosophila and mammalian cells(Kozma and Thomas, 2002; Oldham and Hafen, 2003). Because dRheb larvae exhibit a growth arrest similar to dTORmutants and larvae starved for amino acids, we used rapamycin treatment to investigate whether dRheb interacts, either in the whole organism or in cultured cells, with dTOR. Under normal growth conditions, flies with either one or two copies of dRheb eclose (emerge from the pupal case) at the same time; in the presence of rapamycin, however, larvae with only one wild-type copy of dRheb grow more slowly, eclosing two days later than their wild-type sibs (Fig. 9A). Reduced dRheb thus sensitizes the organism to the growth-inhibiting effect of rapamycin. We also examined possible involvement of dTOR in dRheb function in S2 cells, in which, as expected, treatment with rapamycin causes cells to decrease in size(Fig. 9B). Significantly, the cell-growth-promoting effect of overexpressing dRheb is blocked by rapamycin (Fig. 9C). This latter result indicates that the effect of dRheb overexpression depends on the functional activity of dTOR; in other words, dTOR is epistatic to (downstream of) dRheb.

We show here that, both in vivo and in culture, dRheb is required for cell growth. A role for dRheb in cell cycle progression in S2 cells in culture is indicated by observations that dRheboverexpression drives cells into S-phase, whereas inhibition of dRhebexpression leads to G1 arrest. Consistent with this, dRheb expression in the embryo correlates with domains carrying out DNA replication. A role for dRheb in promoting cell growth is shown in vivo by the enlarged cell size (but not cell number) in wings overexpressing dRheb, and the smaller cell size in dRheb^-/- clones in the eye. In culture, dRheb overexpression leads to a small increase in cell size,whereas inhibition of dRheb expression causes a dramatic reduction in cell size. The effect of dRheb on growth is probably mediated by dTOR. Mutants in dTOR exhibit phenotypes similar to those of dRheb loss-of-function mutants, namely small cell size and growth arrest similar to what is seen during amino acid starvation(Britton and Edgar, 1998; Oldham et al., 2000; Zhang et al., 2000). Our experiments using rapamycin show that flies with reduced levels of dRheb are hypersensitive to the drug and that the effect of overexpressing dRheb in S2 cells is blocked by rapamycin. These results, taken together, suggest that Rheb might function upstream of TOR.

An important question is whether Rheb affects downstream effectors of TOR,most particularly S6K, which modulates growth by controlling protein synthesis. Recent studies have demonstrated a link between dRheb and the dTOR/dS6K pathway (Saucedo et al.,2003; Stocker et al.,2003). Transient expression of dRheb in Drosophila S2 cells was found to lead to the activation of p70S6K,and inhibition of dRheb expression by small interfering RNA blocked amino-acid-induced activation of p70S6K(Saucedo et al., 2003). In addition, in Drosophila dRheb mutants, p70S6K activity was reported to be decreased (Stocker et al.,2003). We have recently observed that transient expression of human Rheb in mammalian cells (HEK293) induces activation (phosphorylation) of S6K (C.L.G. and F.T., unpublished). Although these results might be taken to suggest that Rheb acts solely through TOR and S6K, it is important to realize that dS6K^-/- flies develop into adults with reduced body size (Montagne et al., 1999),whereas, as shown here, dRheb^-/- flies do not survive. Thus, dS6K might not be the only target of dRheb.

As shown here, dRheb exhibits intrinsic GTPase activity despite having arginine at the position corresponding to glycine-12 of Ras, a residue known to be crucial for GTPase activity. Our results are in agreement with the reported intrinsic GTPase activity of mammalian Rheb by Im et al.(Im et al., 2002), who also described that the ratio of GTP to GDP bound to Rheb in mammalian cells is higher than that bound to Ras. The possibility that dRheb might also have high GTP:GDP ratio could explain why overexpression of wild-type dRheb has significant growth promoting effects, as shown here. In our experiments, we did not use dRhebQ63L, a mutant analogous to the constitutively active Ras,because it has previously been shown that expression of S. pombeRhebQ64L or RhebS20N (analogous to a dominant negative Ras) in S. pombe cells has no detectable effect(Mach et al., 2000). Similar observations were made with mammalian Rheb(Clark et al., 1997).

Our finding of GTPase activity for dRheb raises the possibility that there is a dRheb GTPase-activating protein (GAP). A candidate for such an activity is the TSC1/TSC2 complex; TSC1 and TSC2 are tumor suppressors that act to regulate TOR negatively (Gao and Pan,2001; Potter et al.,2001; Tapon et al.,2001; Radimerski et al.,2002). Interestingly, the C-terminal portion of TSC2 contains a GAP domain that is reported to function as a GAP for Rap1 and Rab5(Wienecke et al., 1995; Xiao et al., 1997). Our results linking dRheb to TOR raise the possibility, currently under investigation, that TSC2 is a GAP for Rheb.

We have previously reported that, in S. pombe, Rheb inhibition leads to cell cycle arrest at G0/G1 phase(Yang et al., 2001). Because this growth arrest can be rescued by expression of dRheb, there is conservation of function between Rhebs in two evolutionarily very distant organisms. It is therefore significant that TOR and TSC homologues exist in S. pombe (Kawai et al.,2001; Weisman and Choder,2001; Matsumoto et al.,2002). Notably, disruption of tor2⁺ in S. pombe leads to growth arrest, whereas disruption of tsc2⁺ leads to amino acid uptake defect(Weisman and Choder, 2001; Matsumoto et al., 2002); both of these phenotypes are reminiscent of those seen for yeast rhebmutants (Urano et al., 2000; Yang et al., 2001). Further experiments are needed to examine possible genetic interactions between the tor, tsc and rhb1 genes in S. pombe. As for mammalian Rheb, it has been reported that Rheb inhibits activation of Raf kinase and that overexpression of Rheb in NIH3T3 cells antagonizes Ras transformation (Clark et al.,1997; Im et al.,2002). It will be important to examine whether mammalian Rheb also functions in the insulin/mTOR signaling pathway.

Our observations suggesting that Rheb is involved in the TOR pathway could have implications for cancer therapy, because TOR activity appears to be upregulated in several human cancers(Hidalgo and Rowinsky, 2000). Because of its interaction with TOR, Rheb might be an important player in tumorigenesis. As previously reported, farnesylation is required for Rheb activity as well as its cellular localization(Clark et al., 1997; Urano et al., 2000; Yang et al., 2000); farnesyltransferase inhibitors (FTIs), anticancer drugs currently being evaluated in clinical trials, might block Rheb function. It is interesting to note that FTI has been shown to inhibit p70S6K activation in mammalian cells (Law et al.,2000). TOR function is blocked by rapamycin, initially used as an immunosuppressant but currently under investigation as an anticancer drug(Crespo and Hall, 2002). Therefore, both Rheb and TOR can be targets for anticancer drug therapy. Further investigations are expected to provide insights that will be relevant for the design of future drug therapies.

We thank J. Clemens and L. Zipursky for advice on small interfering RNA experiments with S2 cell lines, J. Merriam for assistance with the ectopic expression screen, J. Canon and C. Evans for advice on histology, G. Jackson,P. Gupta, U. Banerjee, J. Merriam and the Bloomington Stock Center for providing fly lines, and F. Laski and M. Smith for providing the Drosophila plasmid cDNA library and the pPacFLAG, pRmHa-GFP vectors. This work was supported by NIH grants CA41996, CA32737 and HD09948 to FT and JAL, respectively, as well as by a grant from the UCLA Jonsson Comprehensive Cancer Center to JAL. Flow cytometry performed in the UCLA Flow Cytometry Core Facility was supported by NIH grants CA16042 and AI28697.