Ras ( CXXX ) and Rab ( CClCXC ) Prenylation Signal Sequences Are Unique and Functionally Distinct *

Rab proteins typically lack the consensus carboxylterminal CXXX motif that signals isoprenoid modification of Ras and other isoprenylated proteins and, instead, terminate in either CC or CXC sequences (C = cysteine, X = any amino acid). To compare the functional relationship between the Ras CXXX and the Rab CC/CXC motifs, we have generated chimeric Ras proteins terminating in Rab carboxyl-terminal CC or CXC sequences. These mutant Ras proteins were not isoprenylated in vitro or in vivo, demonstrating that the CC and CXC sequences alone are not sufficient to replace a CXXX sequence to signal Ras isoprenoid modification. Surprisingly, chimeric Ras/Rab proteins terminating in significant lengths of carboxyl-terminal sequences from Rablb (7-139 residues), Rab2 (5-151 residues), or Rab3a (12 residues) were also not isoprenylated. These results demonstrate that the sequence requirements for isoprenoid modification of Rab proteins are more complex than the simple tetrapeptide CXXX sequence for isoprenoid modification of Ras proteins and suggest that the Rab geranylgeranyl transferase(s) requires recognition of protein conformation to signal the addition of geranylgeranyl groups. Finally, competition studies demonstrate that a common geranylgeranyl transferase activity is responsible for the modification of Rab proteins terminating in CC or CXC motifs.

costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ylation of the farnesylated cysteine). These CXXX-signaled modifications are critical for Ras association with the plasma membrane and for Ras transforming activity (4-6).
The importance of the CXXX motif for isoprenoid modification of Ras proteins has been demonstrated by several lines of study. First, mutant Ras proteins lacking either the cysteine or the XXX residues of the CXXX motif are completely blocked in isoprenoid modification (7-9). Second, synthetic tetrapeptides corresponding to the CXXX sequences of the Ras proteins are both efficient competitors of, and substrates for, Ras farnesyl transferase activity (10-12). Third, the introduction of the H-Ras CXXX tetrapeptide sequence to heterologous proteins such as protein A or the a subunit of the heterotrimeric Gi protein results in the prenylation of these normally nonprenylated proteins (8,13). Therefore, the CXXX tetrapeptide sequence is not only required but is also sufficient to signal for isoprenoid modification of Ras proteins and other isoprenylated proteins. Finally, the last residue of the CXXX sequence dictates the specific modification of the CXXX-containing proteins by different prenyl transferases which specifically promote the addition of either farnesyl or geranylgeranyl moieties (11, [14][15][16]. Although the Rab/YPT family of Ras-related proteins represents a major group of isoprenylated proteins (1, 17-20), most lack carboxyl-terminal CXXX sequences, and instead, typically terminate in either CC or CXC motifs (21-23). Mutant Rab proteins lacking their carboxyl-terminal CC or CXC sequences are not modified by prenylation (11,18,(24)(25)(26), are cytosolic and lack biological activity. Thus the CC and CXC motifs are functionally analogous to the Ras CXXX motif and are critical for the posttranslational processing and subcellular localization of Rab proteins.
Biochemical and genetic evidence indicates that proteins terminating in CC/CXC sequences are modified by a geranylgeranyl transferase(s) that is distinct from the activities that modify proteins terminating in CXXX sequences (11, 27-29). Furthermore, since the terminal 3 residues of the CXXX motif are critical to signal isoprenoid modification of Ras proteins (11,30), it is also clear that the sequence requirements for Rab isoprenoid modification are distinctly different from the sequences recognized by the CXXX prenyl transferases. Finally, the CC and CXC motifs signal modifications distinct from those signaled by CXXX sequences. The CXC motif of Rab3a has been shown to be modified by two geranylgeranyl moieties (17) and by carboxyl methylation (17, 31). The CC motif is modified by at least one, or possibly two, geranylgeranyl isoprenoids (18, 20, 32) but is not carboxylmethylated (20,31). Thus the CC and CXC sequences contrast with the CXXX sequences, and with each other, signaling structurally distinct carboxyl-terminal modifications.
Recent evidence implicates mammalian Rab proteins as critical regulatory elements in intracellular trafficking (26,

24363
This is an Open Access article under the CC BY license. 33). The localization of different Rab proteins at distinct intracellular membrane compartments of the endocytic and exocytic pathways (3,34) is consistent with this function. For example, the Rablb protein has been localized to the endoplasmic reticulum and Golgi (33), while Rab3a protein is associated with secretory vesicles (35,36). Thus, current models suggest that each Rab protein interacts with two distinct membrane subcompartments and reversibly cycles between cytosolic and membrane fractions (34,37). While the functional distinction between the CXXX, CC, and CXC motifs is presently not known, it has been demonstrated that the CC/CXC carboxyl-terminal sequences are critical for Rab function in regulating vesicular trafficking in endocytosis and exocytosis (24-26), and it is possible that this function may not be replaced by the carboxyl-terminal CXXX-signaled modifications.
Since proteins terminating in CXXX uersus CC/CXC sequences undergo different modifications that are catalyzed by different enzymes, we have initiated studies to determine the Rab sequences required to signal isoprenoid modification and t o determine whether CXXX, CC, and CXC motifs have different functional roles. Our results demonstrate that the CC/CXC sequences cannot functionally replace the Ras CXXX sequence to promote Ras processing and membrane association, that additional upstream sequences as well as protein conformation seem to play a critical role in prenylation of the Rab proteins, and finally, that proteins terminating in CC or CXC sequences are modified by a common geranylgeranyl transferase activity.

EXPERIMENTAL PROCEDURES
Generation of Ras and Rab Carboxyl-terminal Mutants and Ras/ Rab Chimeric Proteins-Oncogenic H-Ras(61-Leu) or Rablb mutant sequences that encode carboxyl-terminal mutants (Table I) were generated by oligonucleotide-directed mutagenesis using Taq polymerase chain reaction DNA amplification as described previously (18). H-raslrablb, H-raslrab2, and H-raslrab3a chimeric genes were constructed using polymerase chain reaction amplification to generate specific DNA fragments from H-ras, rablb, rab2, and rab3a cDNA sequences using pairs of oligonucleotides to introduce the appropriate restriction sites at the 5'-and 3'-ends of the resulting DNA fragments (see Fig. 3). The resulting fragments were introduced into either PAT-RasH (Ras and Ras/Rab chimeras) or pET3a (Rab) bacterial expression vectors (38,39). All sequences were verified by dideoxy sequencing (40). Ras, Rab, or Ras/Rab proteins were expressed and isolated from Escherichia coli strains PR13-Q or BL-21(DE3) as described previously (18). The mutant and chimeric genes were also introduced into the pZIP-NeoSV(X)l retrovirus vector for expression in mammalian cells (41).
I n Vitro Protein Prenylation Assays-Specific prenylation of bacterially expressed Rablb carboxyl-terminal mutants, H-Ras carboxylterminal mutants, and H-Ras/Rablb chimeras were determined using a n in vitro rabbit reticulocyte lysate prenylation system as described previously (18). Purified proteins (1.2 p~) were incubated in a final volume of 50 p1 with nuclease-treated reticulocyte lysate (Promega)    (24 Ci/mmol). Incubations were conducted a t 30 "C for 1 h, and the reactions were stopped by addition of an equal volume of 2 X sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and analyzed by SDS-PAGE and fluorography. The Rablb carboxyl-terminal mutants were also incubated with partially purified Y P T l geranylgeranyl transferase activity (designated GGTase 11; provided by J. Gibbs, Merck Sharp and Dohme Laboratories, West Point, PA) (11) in the presence of [3H]GGPP, and [3H]GG incorporation was visualized by SDS-PAGE and fluorography.
Guanine Nucleotide Binding Assays-Two approaches were used to establish the ability of each Ras/Rab chimeric protein to bind GTP efficiently. First, 10 p1 of purified bacterially expressed proteins (50 pg/ml) were incubated in a 50-p1 reaction with 25 mM HEPES (pH 8.0) and 1 mM DTT, in the presence of 0.1 p~ [y-35S]GTP (800 Ci/mmol; Du Pont-New England Nuclear) for 10 min at 30 "C. Each reaction was then stopped by adding 1 ml of 25 mM Tris (pH 8.0), 100 mM NaCl, 30 mM MgC12,2 mM dithiothreitol, and 1 mg/ml bovine serum albumin. The amount of bound [Y-~'SS]GTP was measured by vacuum filtration and liquid scintillation counting (42). Second, bacterially expressed proteins were separated by SDS-PAGE and transferred to nitrocellulose filters, then incubated with 1 p~ [y3'S]GTP in 50 mM Tris, 0.3% Tween, 0.5 mM EDTA, 0.5 mM MgCl', and 10 pM of GTP (43). The bound ["SIGTP was then visualized by fluorography.
Characterization of Isoprenylation of Ras/Rab Chimeric Proteins in Vivo-NIH 3T3 mouse fibroblasts were transfected by the calcium phosphate precipitation technique (44), and transfected cells were isolated by selection in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and 400 pg/ml G418 (Geneticin, GIBCO). NIH 3T3 cells expressing H-Ras/Rab chimeric proteins were established by cotransfection with an oncogenic c-raf cDNA plasmid construct. Details of the establishment of these cells will be described elsewhere.' NIH 3T3 cells expressing each mutant H-Ras and H-Ras/Rab protein were metabolically labeled for 16-18 h a t 37 "C in Dulbecco's modified Eagle's medium supplemented with 10% dialyzed calf serum and ['HIMVA at 100 pCi/ml or [35SJmethionine/ cysteine (Met/Cys; Translabel, ICN Pharmaceuticals) a t 100 pCi/ml. Fractionation into crude membrane (P100) and cystosolic (S100) fractions and partitioning into Triton X-114 were done as described previously (18). H-Ras and chimeric Ras/Rab proteins were then immunoprecipitated with the Y13-259 (45) and Rasll (46) anti-Ras monoclonal antibodies and analyzed by SDS-PAGE and fluorography.
Selective Inhibition of Protein Prenylation by Proteins with Different Carboxyl-terminal Motifs-To determine if proteins with distinct carboxyl-terminal motifs are modified by shared prenyl transferase activities, 0.75-1.0 p~ of bacterially expressed H-Ras, G25K, Rablb, and Rab3a were each incubated with nuclease-treated rabbit reticulocyte lysate and [3H]FPP (for H-Ras) or ['HH]GGPP (for G25K, Rablb, and Rab3a) alone, or in the presence of a >lO-fold molar excess (10 p~) of the competing protein or peptide substrate. Reactions were done for 1 h at 30 "C, and the proteins were then resolved on SDS-PAGE, and fluorographed after treatment with Amplify (Amersham). The gel slices containing the labeled proteins were also treated with H'O, to release the labeled material which was quantified by liquid scintillation as described previously (10,18). The G25K bacterial expression construct was provided by R. Cerione (Cornel1 University, Ithaca, NY) (47).

RESULTS
Rab CC or CXC Sequences Are Necessary but Not Sufficient for Rab Isoprenylation-In order to determine the carboxylterminal sequence requirements for Rab isoprenylation, we generated Ras or Rablb mutant proteins with altered carboxyl-terminal sequences ( Table I) locyte lysate or partially purified geranylgeranyl transferase (designated GGTase 11) and [3H]MVA to determine prenylation efficiency in vitro. In contrast to the wild type H-Ras protein, mutant H-Ras proteins that terminate in either CC or CSC sequences were not prenylated in vitro (Table I).
Furthermore, analysis of NIH 3T3 cells expressing the mutant proteins also failed to demonstrate incorporation of [3H]MVA in vivo (data not shown). Consistent with the absence of isoprenylation, fractionation analysis demonstrated that, like the nonprenylated H-Ras(AVLS) mutant, the CC and CSC mutant proteins were located exclusively in the cytosolic (SlOO) fraction (Fig. 1). Thus the Rab CC or CXC sequences alone cannot replace the Ras CXXX sequence to promote Ras isoprenylation and membrane association in vivo.
While a Rablb mutant protein lacking both carboxyl-terminal cysteine residues (Rablb(ACC)) was not a substrate for isoprenylation in vitro, mutant Rablb proteins retaining either cysteine residue (Rablb(CS) and Rablb(SC)) were radiolabeled with both [3H]MVA and [3H]GGPP (Fig. 2). These results suggested that both cysteines may be modified by isoprenoid addition and that the CC motif, like the CXC motif, may also signal modification by two geranylgeranyl moieties. In addition, the reduced incorporation of [3H] into these two mutant proteins, when compared to the wild type protein, demonstrates that both cysteine residues are required for optimal prenylation of Rablb. A Rablb mutant protein that retains the double cysteines, but lacks 7 amino acids upstream from the cysteine residues (Rablb(A7)CC), was not isoprenylated in vitro (Table I). Thus, Rablb isoprenylation is dependent on both the CC motif and sequences upstream of these cysteine residues. Finally, Rablb mutant protein terminating in either CXC (CKC or CSC) (Table I and Fig.   2, lanes m-o) or CXXX (CVLS) ( Table I) sequences were also efficiently isoprenylated in uitro, indicating that Rablb carboxyl-terminal sequences upstream of the CC motif are compatible with these two different prenylation signal sequences.

Rab Geranylgeranyl Transferases May Require Recognition
of Protein Conformation for Isoprenylation-To determine the carboxyl-terminal sequence requirements for isoprenoid modification of Rab proteins, we generated a series of Ras/Rab chimeric proteins that contain the amino-terminal sequences of H-Ras but in which varying lengths of Rablb (7-139 residues), Rab2 (5-151 residues), or Rab3a (12 residues) replaced the carboxyl-terminal region of H-Ras (Fig. 3). Surprisingly, none of these chimeric proteins was isoprenylated in vitro when incubated with reticulocyte lysate (representative examples are shown in Fig. 4, lanes d, f, and i) or partially purified GGTaseII (data not shown). Analysis of these chimeric Ras/Rab proteins in transfected NIH 3T3 cells demonstrated that these mutant proteins, in contrast to H-Ras(GlL), were not metabolically labeled with [3H]MVA, were present exclusively in the crude cytosolic (SlOO) fraction (Fig.  5A), and did not partition into the detergent phase after Triton X-114 extraction (Fig. 5B). Thus, despite possessing significant carboxyl-terminal sequences from Rab proteins, these chimeric proteins were still not substrates for isoprenylation i n vitro or in vivo.
The lack of isoprenoid modification of the Ras/Rab chimeric proteins suggested that Rab carboxyl-terminal sequences alone are not sufficient for Rab geranylgeranyl transferase modification. Alternatively, the improper folding of the combined amino-terminal Ras and carboxyl-terminal Rab sequences, or the inaccessibility of the carboxyl termini to the prenyl transferase, may account for the lack of efficient isoprenoid modification of these chimeric proteins. To address the first possibility, we analyzed these chimeric proteins for their ability to bind [35S]GTPyS. Bacterially expressed chimeric Ras/Rab proteins were analyzed by incubation with M [35S]GTPyS, followed by vacuum filtration through nitrocellulose filters, or by separation on SDS-PAGE and transfer to nitrocellulose followed by incubation with [3sS] GTPyS. Like authentic H-Ras, Rablb, and Rab2 proteins, all chimeric Ras/Rab proteins efficiently bound [3sS]GTPrS (data not shown), indicating that they were properly folded to produce a functional conformation. To address the possibility that the carboxyl-terminal sequences were not available for isoprenylation, we generated chimeric Ras/Rablb proteins which terminated in CVIM instead of in CC residues. In contrast to their non-CXXX containing counterparts, both the Ras/Rablb.l5(CVIM) and Ras/Rablb.33(CVIM) chimeric proteins were good substrates for isoprenylation (Fig.   4, lanes g and h), indicating that their carboxyl-terminal regions were accessible for isoprenoid addition. Furthermore, both Ras/Rablb.l5(CVIM) and Ras/Rablb.33(CVIM) displayed potent transforming activity when expressed in NIH 3T3 cells (data not shown), which provides further indication that these chimeric proteins are properly folded.
Crystal structure determination of the H-Ras protein demonstrated that the amino and carboxyl-terminal sequences reside on the same face of the protein (48,49). Therefore, we  Rab2 (light hatched bones), and Rab3a (black boxes) sequences are shown. The carboxyl-terminal residues of H-Ras were replaced by 7 to 139 carboxyl-terminal residues from Rablb (cg), 5 to 151 carboxyl-terminal residues from Rab2 (i-m), or 12 carboxyl-terminal residues from Rab3a ( p ) . The numbers after each chimera indicate the number of amino acids from carboxylterminal of Rab proteins that replace analogous amino acids from H-Ras. In addition, a RabS/Ras/Rab2 chimera ( n ) was generated which contains the first 60 amino-terminal residues of Rab2 followed by H-Ras residues 61-160 and terminates with the 65 carboxyl-terminal amino acids of Rab2.    lanes u and b), Ras/ Rablb.ll (lanes c and d )  Modification of Rab Proteins Terminating with CC or CXC Sequences-To further characterize the prenyl transferase activities responsible for the isoprenylation of proteins terminating in CXXX, CC, or CXC sequences, we performed competition experiments to measure the ability of heterologous peptides or proteins to block isoprenylation of bacterially expressed H-Ras (carboxyl-terminal CXXX motif for farnesylation), G25K (carboxyl-terminal CXXX motif for geranylgeranylation), Rablb (CC), or Rab3a (CXC) proteins using the reticulocyte lysate assay (Fig. 6). For each assay 1 pM of each substrate protein was incubated either alone, or in the presence of a 10-fold molar excess of the competing peptide or protein, and the incorporation of either [3H]FPP or [3H] GGPP into the substrate protein was determined. H-Ras isoprenylation was selectively inhibited by a synthetic peptide corresponding to the carboxyl terminus of the farnesylated yeast a-factor (CVIA) (50), but not by full length G25K (CVIF), Rablb (CC), or Rab3a (CSC) protein (Fig. 6A). Similarly, G25K isoprenylation was inhibited by a synthetic heptapeptide corresponding to the carboxyl terminus of the geranylgeranylated 7 -6 subunit of heterotrimeric G proteins (KFFCAIL) (16,51,52), but not by full length H-Ras, Rablb, or Rab3a protein (Fig. 6B). In contrast to the ability of CXXX-containing peptides to efficiently inhibit the isoprenylation of Ras or G25K proteins, Rablb and Rab3a peptides corresponding to the last 8 carboxyl-terminal amino acids of these proteins did not inhibit the isoprenylation of either Rablb or Rab3a when used at peptide concentrations of up to 50 p~ (data not shown). These data, together with the inability of the chimeric Ras/Rab proteins to serve as efficient substrates for isoprenylation, provide further evidence that Rab prenyl transferase(s) requires recognition of Rab sequences distant from the carboxyl termini.
To determine if Rablb and Rab3a are modified by distinct  or common geranylgeranyl transferase activities, the ability of excess full-length Rab proteins to block isoprenylation of the corresponding Rab protein was determined. While Rablb isoprenylation was not inhibited by G25K or H-Ras proteins, geranylgeranylation of Rablb was inhibited (by 84%) when assayed in the presence of a 10-fold molar excess of Rab3a ( Fig. 6C and Fig. 7, lanes c-e). Similarly, Rab3a geranylgeranylation was inhibited (by 93%) when assayed in the presence of a 10-fold molar excess of Rablb, but not H-Ras or G25K proteins (Fig. 6D and Fig. 7, lanes Liz). This level of inhibition correlated well with that predicted by the respective K , values for these substrates (K,,, for Rablb = 1.0 PM; K , for Rab3a = 2.3 PM) if their prenylation were catalyzed by the same enzyme(s). The predicted inhibition of Rablb prenylation by Rab3a under these conditions is 71% and that for Rab3a by Rablb is 88%. Thus these data suggest that the CC and CXC proteins are recognized by the same geranylgeranyl transferase(s).

DISCUSSION
While the carboxyl-terminal CC and CXC sequences of Rab proteins are functionally analogous to the Ras CXXX se-quence in that they signal isoprenoid modification and promote membrane association, three previous observations have suggested that the sequence requirements for prenyl transferase recognition and isoprenoid modification of Rab proteins may be different from those of Ras and other CXXX-containing proteins. First, in contrast to the ability of CXXX-containing peptides to inhibit Ras isoprenylation, synthetic peptides containing the Rab CC or CXC sequences were not effective inhibitors of, nor substrates for, isoprenylation (11,28,29). Second, it has been shown that the noncysteine XXX residues of the CXXX motif are the critical residues for efficient recognition and modification by the CXXX prenyl transferases (11,30,53). Third, a mutant yeast Rasl protein that terminated with the 6 carboxyl-terminal residues of yeast YPTl was not a substrate for GGTase I1 activity (11). Despite these indications, our observations that chimeric Ras/Rab proteins terminating in significant carboxyl-terminal sequences from Rablb, Rab2, or Rab3a were not isoprenylated were unexpected. These results clearly demonstrate that the sequence requirements for Rab isoprenylation are strikingly different from those for isoprenylated proteins that terminate in CXXX sequences, and suggest that Rab geranylgeranyl transferase(s) requires recognition of sequences outside the Rab carboxyl termini, and possibly protein conformation, for isoprenoid modification. Whereas CXXX carboxyl-terminal sequences are both necessary and sufficient to signal isoprenoid modification, our results indicate that multiple regions of the CC-and CXC-terminating Rab proteins combine to act as structural determinants of the isoprenylation signal.
Synthetic peptides corresponding to the CXXX sequences of Ras proteins have been very useful reagents for the identification and analysis of CXXX farnesyl and geranylgeranyl transferases (10,14,16) and have allowed a rapid characterization of the structural requirements for prenylation of CXXX-terminating proteins (11,53,16). Ras CXXX tetrapeptides are sufficient to both inhibit Ras isoprenylation and to serve as substrates for Ras farnesyl transferase activity (1 1, 12, 53). Furthermore, CXXX tetrapeptides have also greatly facilitated the affinity purification of Ras prenyl transferases and have provided a basis for the design of potential inhibitors of oncogenic Ras activity in human tumors (10). By contrast, since prenylation of Rab proteins requires recognition of primary or higher order structure besides the carboxyl-terminal prenylation domain, such carboxyl-terminal peptides will be less useful for the characterization of Rab prenyl transferases. In particular, assays of Rab prenylation and determination of specific sequence requirements for reactivity of the Rab substrates will require intact Rab proteins. Recent studies have identified geranylgeranyl transferase activities (designated GGTase I1 or Rab GGTase) for Rab proteins (1 1, 28, 29) which are distinct from the activity that geranylgeranylates proteins that terminate with CXXX sequences (GGTase I). The results from competition experiments described in this study demonstrated that Rablb (CC motif) and Rab3a (CXC motif) were competitors of the isoprenylation of the reciprocal protein in a reticulocyte lysate assay. These results provide evidence that geranylgeranyl additions to CC or CXC terminating proteins are catalyzed by a common geranylgeranyl transferase activity(ies). Additionally, the successful isoprenylation of mutant Rablb proteins in which CXC sequences replaces the authentic CC sequence suggests that the CC and CXC motifs require common upstream sequences, or protein conformation, for signaling geranylgeranyl addition. Finally, the recent demonstration that a partially purified Rab3a geranylgeranyl transferase activity also efficiently modified Rabla is also consistent with a common prenyl transferase for these two motifs (29). Our observation that Rablb mutants terminating in either SC or CS are still geranylgeranylated are in agreement with results described by Kinsella and Maltese (32) and suggests that the prenylation of the 2 cysteines are independent, nonsequential reactions. Whether these two geranylgeranyl modifications are catalyzed by a single enzymatic activity remains to be determined.
While it is clear that the Ras C X X X and Rab CC or CXC motifs serve analogous roles in triggering isoprenoid modification and membrane association (6), the distinct processing signaled by each of these motifs suggests the possibility of distinct roles in regulating protein function. Proteins terminating in C X X X sequences are modified by either single farnesyl or geranylgeranyl moieties, and the isoprenylated cysteine also undergoes carboxyl methylation. In contrast, Rab proteins terminating in CC or CXC sequences apparently undergo modification by geranylgeranyl groups on both cysteines, but only the CXC proteins are further modified by carboxyl methylation. One aim of our studies has been to determine whether Ras function can be promoted by the modifications signaled by the CC or CXC motifs. The nonprenylated and cytosolic nature of mutant Ras proteins that terminate with CC or CXC sequences clearly indicate that these motifs alone do not functionally replace the C X X X motif for promoting Ras function. However, since these mutant proteins are not processed, how the function of Ras proteins in regulating signal transduction pathways in the cell is facilitated by Rab carboxyl-terminal modifications remains t o be answered.
At present, there is limited information available addressing whether there is a functional basis for the distinct prenylation motifs. Our recent observation that normal Ras proteins that terminate in mutant C X X X sequences signaling for geranylgeranyl, rather than farnesyl, addition are potent inhibitors of cell proliferation suggests a distinct role for distinct prenylation motifs (54). In contrast, the observation that yeast YPT1, which is essential for cell viability (43) and facilitates endoplasmic reticulum (ER) to Golgi transport (55), could still support yeast viability when the CC motif is substituted with the H-Ras C X X X sequence ( C V L S ) might argue against a unique role for the different motifs in protein function. However, it was not determined whether this carboxyl-terminal mutant was correctly processed or whether it retained wild type function in transport. Therefore, at present, it is not known whether the C X X X , C C , and CXC motifs possess distinct roles in promoting protein function. Clearly, further analysis of this question will be required to establish the functional significance of different carboxyl-terminal sequences that signal protein prenylation. We have recently demonstrated the contribution of Rablb protein in ER to Golgi transport (33), and we are presently determining whether mutant Rablb proteins that terminate in either CXC or C X X X sequences also function in ER to Golgi transport.