If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Pew Scholar in the Biomedical Sciences. To whom correspondence should be addressed: Dept. of Molecular Genetics, University of Texas Southwestern Medical Center, Dallas, TX 75235. Tel.: 214-648-8643; Fax: 214-648-8804;
* This work was supported by the Foundation Fighting Blindness, by National Institutes of Health Grant HL20948, and by the Perot Family Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Geranylgeranylation of Rab GTPases is an essential post-translational modification that enables Rabs to associate with intracellular membranes where they regulate exocytic and endocytic pathways. Geranylgeranylation is initiated by formation of a stable complex between newly synthesized Rab proteins and Rab escort protein (REP). The complex is recognized by Rab geranylgeranyl (GG) transferase, which transfers two GG groups to Rabs. The geranylgeranylated Rabs regulate vesicular movement by oscillating between an inactive GDP-bound form and an active GTP-bound form. In this study, I show that the kinetics of geranylgeranylation is influenced by the nucleotide status of nascent Rab. GDP-bound Rab is geranylgeranylated with 10–50-fold higher affinity than GTP-bound Rab (or GTP analog-bound Rab), as indicated by the apparent Km of the reaction. In vitro REP·Rab binding assays demonstrate that REP forms a stable complex only with the GDP-bound form of Rab but not the GTP-bound form, suggesting that the apparent Km effect in the prenylation reaction is due to a discrimination between the two different nucleotide-bound forms of Rab by REP. Inasmuch as Rabs are likely GTP-bound after synthesis and REP does not possess GTPase-activating protein activity, these results raise the possibility that a Rab GTPase-activating protein enhances the REP·Rab interaction prior to prenylation.
Post-translational modification of proteins with lipids is a frequent and important cellular regulatory mechanism that has only recently been appreciated (reviewed in
). The lipid moieties, including acyl groups, prenyl groups, or both, are essential for the function of the modified protein, because they serve as determinants for protein-membrane or protein-protein interactions. Two types of prenyl groups have been identified to date: farnesyl or geranylgeranyl (GG).
Both are attached to carboxyl-terminal cysteine residues via thioether bonds.
Protein prenylation is catalyzed by three cytosolic enzymes that can be classified in two groups: the “CAAX” prenyl transferases, which include farnesyl transferase and GG transferase-I, and the Rab GG transferase or GG transferase-II (
). CAAX prenyl transferases recognize a distinct motif at the carboxyl termini of their intracellular substrates. The motif is designated CAAX, where C is cysteine, A is aliphatic, and X is any amino acid. The CAAX sequence is the primary determinant for binding to the protein substrate. If the last amino acid in the CAAX sequence is a methionine or serine, the protein is a substrate of farnesyl transferase, whereas if it is a leucine it becomes a substrate for GGTase-I (
). CAAX-containing proteins include members of the Ras and Rho/Rac family of low molecular weight GTPases, γ-subunits of G-proteins, nuclear lamins, G-protein-coupled receptor kinases, and retinal cyclic GMP phosphodiesterase.
All proteins belonging to a large subfamily of low molecular weight GTPases, termed Rab in mammals and YPT/SEC4 in yeast, undergo geranylgeranylation even though they lack a classical CAAX motif at the carboxyl terminus. Instead, these proteins typically possess a double-cysteine motif, such as CC or CXC, of which both cysteines are modified by GG groups (
). This enzyme, called Rab GGTase or GGTase-II (previously designated Component B of Rab GGTase), is a heterodimer composed of an α-subunit and a β-subunit, both homologous to the corresponding α- and β-subunits of the CAAX prenyl transferases (
). However, in contrast with the CAAX prenyl transferases, Rab GGTase does not recognize Rab substrates unless they are complexed with another cytosolic protein, designated Rab Escort Protein (REP), previously designated Component A of Rab GGTase. REP is also known as the choroideremia gene product (see below) (
Previous studies suggested that newly synthesized Rabs bind REP, and the complex is then recognized by Rab GGTase. Two consecutive GG additions occur, possibly starting with the cysteine furthest from the carboxyl terminus. The product of the reaction, diGG-Rab·REP complex, likely dissociates after translocation of the diGG-Rab into a specific intracellular membrane (
). REP is absolutely required for in vitro prenylation of Rab proteins, and studies in yeast and humans suggest that it is also required in vivo. In Saccharomyces cerevisiae, the REP homolog MRS6/MSI4 is an essential gene (
). Because Rab27 is expressed in the two retinal cell types that degenerate primarily in choroideremia, the retinal pigment epithelium and the choroid, we proposed that the pathogenesis of choroideremia may be related to Rab27 dysfunction (
Once prenylated, Rabs associate with the cytoplasmic leaflet of the membrane surrounding a nascent transport vesicle. There, Rabs are proposed to regulate the targeting and fusion of transport vesicles to the appropriate acceptor compartments in exocytic and endocytic pathways (
). Like other GTP-binding proteins, Rabs cycle between inactive GDP-bound and active GTP-bound forms. Rabs associate with membranes in the GDP-bound form and switch to a GTP-bound conformation in a reaction catalyzed by a GDP/GTP exchange factor (
). After fusion has occurred, the GTP is hydrolyzed in a reaction that is stimulated by a Rab-specific GTPase-activating protein (Rab GAP). This reaction forces Rabs into the GDP-bound form, which leads to their removal from the membrane by Rab GDI. In response to subsequent signals, Rab GDI mediates the reassociation of GDP-Rab with the donor membrane, where the cycle resumes. Rab GDI and REP share structural and functional properties, and it has been proposed that REP acts in the delivery of newly synthesized Rabs, whereas Rab GDI acts in their recycling (
The nucleotide status of Rabs is critical to their function, and it is therefore important to study the nucleotide dependence of Rab geranylgeranylation. This issue has been the subject of previous studies that yielded conflicting results. A study of Rab5 prenylation using a cell-free translation and prenylation assay in rabbit reticulocyte lysates suggested that there was a preference for prenylation of GDP-bound Rab (
In this study, I analyze the prenylation kinetics of GDP versus GTP-bound forms of Rabs using recombinant purified Rab GGTase and REP and show that there is a clear preference for the GDP-bound form. Furthermore, this effect is shown to be a direct consequence of a higher affinity of REP toward GDP-bound Rab. Inasmuch as Rabs are believed to be GTP-bound after synthesis, these results suggest that a Rab GAP may be involved in presenting newly synthesized Rabs to REP prior to geranylgeranylation.
All-trans-[1-3H]GGPP (15 Ci/mmol) and unlabeled all-trans-GGPP were purchased from American Radiolabeled Chemicals (St. Louis, MO). Guanine nucleotides were purchased from Boehringer Mannheim. [α-32P]GTP (3,000 Ci/mmol) was obtained from DuPont NEN.
Assay for Rab GG Transferase Activity
Rab GGTase activity was determined by measuring the amount of [3H]geranylgeranyl transferred from [3H]geranylgeranyl pyrophosphate to Rab proteins (
). Unless otherwise indicated, the standard reaction mixture contained the following concentrations of components in a final volume of 50 µl: 50 mM sodium Hepes (pH 7.2), 5 mM MgCl2, 1 mM Nonidet P-40, 1 mM dithiothreitol, 5.5 µM [3H]GGPP (3,000 dpm/pmol), and the indicated amounts of Rab proteins, REP-1, and Rab GGTase. After incubation for the indicated time at 37°C, the amount of ethanol/HCl-precipitable radioactivity was measured by filtration on a glass fiber filter (
). Briefly, transformed BL21 (DE3) Escherichia coli cells containing pET14b-Rab (canine Rab1a-CC, canine Rab1a-SS, human Rab2 (generously provided by Channing Der, University of North Carolina), human Rab3a, or human Rab5) (
) were grown and lysed, and the supernatant from a 105× g spin (1 h at 4°C) was subjected to nickel-Sepharose affinity chromatography (Pharmacia Biotech Inc.) under the conditions recommended by the manufacturer (Novagen). The eluted His-tagged Rab proteins (>90% pure as judged by SDS-polyacrylamide gel electrophoresis) were each dialyzed against buffer containing 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 3 mM MgCl2, 1 mM EDTA, 5 mM dithiothreitol, and 0.1 mM GDP) and stored in multiple aliquots at −70°C. Recombinant His-tagged REP-1 was prepared in Sf9 insect cells as described previously (
). Briefly, Sf9 cells were infected with recombinant baculovirus encoding REP-1, grown, and lysed, and a 105× g supernatant was subjected to nickel-Sepharose affinity chromatography as described above for the His-tagged Rab proteins. The His-tagged REP-1 (>90% pure as judged by SDS-polyacrylamide gel electrophoresis) was dialyzed against two changes of buffer containing 20 mM sodium Hepes (pH 7.2), 10 mM NaCl, 0.1 mM Nonidet P-40, and 1 mM dithiothreitol) and stored in multiple aliquots at −70°C. Recombinant RabGGTase was prepared as described previously (
). Briefly, Sf9 cells were co-infected with recombinant baculovirus encoding the α- and the β-subunits, grown, and lysed and a 105× g supernatant was chromatographed on Q-Sepharose and Superdex 200 columns, and the active fractions stored in multiple aliquots at −70°C.
). Recombinant Rabs (1–3 mg) in a volume of 500 µl were mixed with 500 µl of exchange buffer (20 mM Tris-Cl (pH 7.5), 1 mM EDTA, 250 mM (NH4)2SO4, 1 mM dithiothreitol, and 0.05 mM Nonidet P-40), and the solution was adjusted to 5 mM EDTA and 250 mM (NH4)2SO4 and incubated in the presence of 33 units of alkaline phosphatase-agarose (Sigma) for 15 min at 22°C. Then, 250 µM (final) GMPPNP and 33 units of alkaline phosphatase-agarose were added, and the mixture was incubated for 30 min at 22°C. The samples were spun in a microfuge for 3 min, and the supernatant was collected and passed through a 5-µm filter, pre-washed with exchange buffer. The sample was collected and pooled with 2 ml of subsequent washes of the filter with exchange buffer. The sample was divided in three aliquots, each adjusted to 2 ml with exchange buffer and 500 µM desired nucleotide and subjected to Centricon 10 (Amicon) ultrafiltration for 45–60 min at 4°C. Each sample was then adjusted to 2 ml using loading buffer (20 mM Tris-Cl (pH 7.5), 5 mM MgCl2, 1 mM EDTA, 50 mM NaCl, 1 mM dithiothreitol, and 0.05 mM Nonidet P-40) and 200 µM desired nucleotide and spun again for 45–60 min at 4°C in Centricon 10. Finally, the sample was adjusted again to 2 ml with loading buffer, spun for 45–60 min at 4°C in Centricon 10, and stored in aliquots at −70°C. The efficiency of loading was determined by analysis of the bound nucleotide using MonoQ chromatography in a Smart System (Pharmacia Biotech Inc.) as described in the legend to Fig. 1.
Purification of Commercial Preparations of Guanine Nucleotides
Commercial preparations of GTP contain significant amounts of contaminating GDP. In order to remove contaminating nucleotides, 10–30 mg of the desired nucleotide was dissolved in water and loaded onto a 6-ml Q-Sepharose HP (Pharmacia Biotech Inc.) equilibrated in 10 mM Tris-Cl (pH 7.5). The column was washed at a flow rate of 2 ml/min at 4°C with 5 ml of the same buffer. Four consecutive gradients in LiCl2 were then used to elute the bound nucleotides: 1 ml of gradient between 0 and 5 mM, 9 ml of gradient between 5 and 10 mM, 50 ml of gradient between 10 and 18 mM, and 5 ml of gradient between 18 and 25 mM LiCl2. The peak fractions were analyzed by MonoQ chromatography as described in the legend to Fig. 1, pooled, and concentrated by ethanol precipitation. Four volumes of ice-cold ethanol were added to each pool, incubated at −20°C for 30–180 min and centrifuged at 3,500 × g, and the pellet was dried for 3 h at room temperature and resuspended in 50 mM Tris-Cl (pH 7.5). The concentration of each sample was determined by absorbance at 254 nm.
Assay for Complex Formation between Rab1a and REP-1 by Gel Filtration Chromatography
REP·Rab complex formation was determined by gel filtration chromatography essentially as described before (
). Reaction mixtures in loading buffer in a final volume of 50 µl, containing 2 µM Rab1a and 2 µM REP-1 were incubated for 10 min at 37°C and loaded onto a Superdex 200 3.2/30 using a SMART system (Pharmacia Biotech Inc.). The column was equilibrated in buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM MgCl2, and 1 mM dithiothreitol at a flow rate of 50 µl/min, the samples were injected, and the material eluting between 0.75 and 1.95 ml was collected in 100-µl fractions. An aliquot (25 µl) of fractions 4–12 was subjected to SDS gel electrophoresis and transferred to nitrocellulose filters, and the proteins were identified by immunoblot analysis as described below. The Superdex 200 column was calibrated with the following gel filtration molecular mass standards (Bio-Rad): thyroglobulin (670 kDa), aldolase (160 kDa), ovalbumin (45 kDa), and myoglobin (17 kDa), which eluted at 1.07 (fraction 4), 1.36 (fraction 7), 1.60 (fractions 9 and 10), and 1.77 ml (fraction 11), respectively.
Assay for Complex Formation between Rab1a and REP-1 by Affinity Chromatography and Guanine Nucleotide Analysis
) was loaded with [α-32P]GTP by incubation for 30 min at 30°C in 100 µl of buffer containing 50 mM Tris-Cl (pH 7.5), 1 mM EDTA, 2 mMβ-mercaptoethanol, 1 mg/ml bovine serum albumin, and 200 µM [α-32P]GTP (2, 200 dpm/pmol). After incubation, the sample was adjusted to 5 mM MgCl2, and the unincorporated nucleotide was removed by rapid centrifugation on G-25 quick spin columns (Boehringer Mannheim) according to the instructions of the manufacturer. Reaction mixtures containing 50 mM sodium Hepes (pH 7.2), 5 mM MgCl2, 1 mM EDTA, 1 mMβ-mercaptoethanol, 5 µM unlabeled GGPP, 2 µM [α-32P]GTP-Rab1a, in the presence or the absence of 2 µM His-REP-1 and 1 µM RabGGTase, in a final volume of 25 µl, were incubated for 15 min at 37°C and subjected to nickel-Sepharose affinity precipitation essentially as described before (
). Briefly, 10 µl of nickel-Sepharose beads equilibrated in 25 µl of wash buffer (20 mM sodium Hepes (pH 7.2), 10 mM NaCl, 0.1 mM Nonidet P-40, and 1 mMβ-mercaptoethanol) were added to the reaction mixtures and incubated for 20 min at room temperature with gentle shaking, and the unadsorbed proteins were removed by centrifugation. The nickel-Sepharose beads were washed twice with 1 ml of wash buffer, vortexed in 10 µl of stop solution (0.2% SDS, 5 mM EDTA, and 2 mM dithiothreitol), heated at 70°C for 2 min, and spotted on polyethyleneimide-cellulose thin-layer chromatography plates to analyze the guanine nucleotides present in the sample. The plates were developed in 0.75 M potassium phosphate buffer (pH 3.5), dried, and autoradiographed using a Fuji phosphorimager.
Antibodies and Immunoblotting
D576 is a polyclonal antibody directed against recombinant canine Rab1a produced as described below, and IgG was purified on a protein A-Sepharose column (Pharmacia Biotech Inc.) as described (
). For immunoblot analysis, samples were subjected to SDS gel electrophoresis (Laemmli, 1970), and polypeptides were transferred to polyvinylidene difluoride membranes and stained with 0.2% Ponceau S in 5% acetic acid or Coomassie Blue. The membranes were incubated in 10 ml of phosphate-buffered saline with 0.2% polyoxyethylenesorbitan monolaureate (Tween 20) throughout the immunoblot protocol. The filters were incubated with primary antibody with gentle agitation for 1 h at room temperature followed by two washes for 10 min. The filters were then incubated for 20 min with secondary antibody, donkey anti-rabbit IgG conjugated to horseradish peroxidase (Amersham). The blots were finally washed twice for 10 min, and bound IgG was visualized by chemiluminescence using either the ECL system (Amersham Corp.) or SuperSignal system (Pierce). All blots were calibrated with prestained molecular weight markers (Bio-Rad).
In order to analyze the nucleotide dependence of Rab prenylation, an effective way to load the desired nucleotide after removing all the prebound GDP had to be devised. I adapted a procedure described by Wittinghofer and co-workers (
) for the loading of Rab proteins. The procedure involves the exchange of pre-bound GDP for GMPPNP in the presence of alkaline phosphatase. Alkaline phosphatase degrades GDP or GTP to guanosine but does not hydrolyze GMPPNP. Rabs will not bind guanosine, and therefore at the end of the exchange reaction, Rabs are efficiently depleted of GDP and loaded with GMPPNP. The latter can later be exchanged with the desired nucleotide owing to its lower affinity for Rabs when compared with GDP or GTP. This procedure was used to load Rabs with GDP, GTP, and the slowly hydrolyzable analog of GTP, GMPPNP. After loading, an aliquot was analyzed for the presence of guanine nucleotides by ion exchange chromatography. Under the conditions used, GDP, GTP, and GMPPNP elute differentially as demonstrated in Fig. 1A. Loading of Rab1a with GDP or GMPPNP resulted in homogenous preparations of GDP-Rab1a (Fig. 1B) and GMPPNP-Rab1a (Fig. 1D), respectively. The GTP-Rab1a preparation contained a small amount of GDP (8%), possibly due to intrinsic hydrolysis during the procedure and some GMPPNP (24%) due to incomplete exchange after the initial exchange of GMPPNP for pre-bound GDP (Fig. 1C).
To analyze the effect of the bound nucleotide on the kinetics of Rab prenylation, the loaded Rabs were used as substrates in the in vitro prenylation assay in the presence of recombinant RabGGTase and REP-1. When GDP-loaded Rab1a was used in the assay, an apparent Km of 2.1 µM was obtained (Fig. 2). This result is similar to previously reported Km values obtained with recombinant Rab1a (
). In contrast, there was a dramatic change in the apparent Km of the reaction when GTP-Rab1a or GMPPNP-Rab1a were used. Because of the low affinity, I was unable to obtain sufficiently concentrated preparations of GTP-Rab1a or GMPPNP-Rab1a to saturate the reaction. However, it is clear that the apparent Km is at least 10-fold higher for GTP-Rab1a and 50-fold for GMPPNP-Rab1a when compared with GDP-Rab1a. To test the generality of this effect, Rab2, Rab3a, and Rab5 were loaded with either GDP and GMPPNP and tested in the in vitro prenylation assay (Fig. 3). In every case, there was a dramatic difference in apparent Km comparable with what was observed in Fig. 2 for Rab1a. The least dramatic change was observed with Rab2, possibly because the GMPPNP-Rab2 preparation was contaminated with 8% of GDP (data not shown).
There are two likely explanations for the observed effect on the prenylation kinetics. One is that REP binds with higher affinity to GDP-Rab, and the second is that REP binds equally well to both forms of nucleotide loaded Rab but that only REP·GDP-Rab is recognized by RabGGTase. To begin to discriminate between these two alternative possibilities, a competition experiment was devised using a mutant Rab1a, designated Rab1aSS. Rab1aSS is unable to serve as a prenyl acceptor because the two cysteine residues that accept geranylgeranyl groups are mutated to serine residues. Rab1aSS retains the ability to bind to REP (
). Susceptibility to Rab1aSS inhibition can serve as an indirect measure of REP-Rab binding. Rab1aSS was loaded with either GDP or GMPPNP and tested as an inhibitor of the prenylation of GDP-Rab1a (Rab1aCC). When increasing amounts of GDP-Rab1aSS were added, we observed up to a 70% reduction in the amount of [3H]GG incorporated into GDP-Rab1a-CC when the mutant protein was present at 30-fold higher concentrations (Fig. 4). At the same high concentration in the assay, GMPPNP-Rab1aSS failed to inhibit the reaction. This result suggests that the GTP-bound Rabs are unable to bind REP with high affinity.
To further test the hypothesis that REP has low affinity for GTP-bound Rabs, two distinct direct REP·Rab binding assays were developed. In the first experiment, a gel filtration assay to ascertain REP·Rab complex formation was used (
) (Fig. 5A). When GDP-Rab1a was incubated with REP-1 and the mixture was subjected to gel filtration under the same conditions, a significant fraction of the GDP-Rab1a (approximately 50%) was shifted to fractions 7 and 8, co-eluting with REP-1 (Fig. 5B). In contrast, when GTP-Rab1a was used only a small amount shifted to fractions 7 and 8 (Fig. 5C), and when GMPPNP-Rab1a was used no significant amount of Rab1a was found eluting in fractions 7 and 8 even after long exposures (Fig. 5D). These results confirm the suggestion derived from the previous experiments that REP has higher affinity for GDP-bound Rabs.
To further study the ability of REP to form a stable complex with GTP-bound Rab, I used a different REP·Rab binding assay (
). This REP·Rab binding assay is based on the selective precipitation of his-tagged REP-1 using nickel-Sepharose. If a stable complex between his-tagged REP-1 and non-His-tagged Rab1a is formed, the nickel-Sepharose pellet will contain detectable amounts of guanine nucleotide associated with Rab in the REP·Rab complex. Non-His-tagged Rab1a was loaded with [α-32P]GTP and incubated with REP-1 in the presence or the absence of Rab GGTase. After precipitation of his-tagged REP-1 with nickel-Sepharose, the nucleotide present in the pellet and supernatant was analyzed by thin-layer chromatography. When His-tagged REP-1 was added to the incubation mixture, essentially all of the nucleotide in the nickel-Sepharose pellet was GDP (Fig. 6, lane 2). When His-tagged REP-1 was omitted from the incubation mixture, no nucleotide was present in the pellet (Fig. 6, lane 1). When Rab GGTase was present and prenylation occurred, the nucleotide in the nickel-Sepharose pellet was approximately 75% GDP and 25% GTP (Fig. 6, lane 3). When the supernatant of the reaction was analyzed, the majority of the nucleotide was GTP (80%), as expected because Rab1a was loaded with GTP (Fig. 6, lane 4). This experiment clearly indicates that even in the presence of an excess of GTP-Rab, REP forms a stable complex preferentially with GDP-Rab. A somewhat puzzling finding in this experiment is that the presence of Rab GGTase in the incubation mixture leads to the co-precipitation of some GTP-bound Rab with REP (Fig. 6, lane 3), whereas no GTP-Rab coprecipitates with REP in its absence (Fig. 6, lane 2). There is no obvious explanation for this finding. It is possible that the higher concentration of GTP-Rab in the assay leads to transient complex formation with REP. If the binding of Rab GGTase to the REP·Rab complex is not nucleotide-sensitive, then prenylation will occur and result in the formation of a prenylated GTP-Rab·REP complex. This complex may be sufficiently stable to allow detection in this assay.
In the experiment shown in Fig. 6, the REP·Rab complex contained [32P]GDP even though [32P]GTP was the only nucleotide added in the loading reaction. One source of the [32P]GDP was a small contamination of the [32P]GTP stock with [32P]GDP. It is also possible that some GDP-Rab was derived from GTP-Rab by GTP hydrolysis during the loading and incubation procedures. Rabs have slow intrinsic GTPase activity, and it is possible that REP could stimulate their intrinsic GTPase activity, i.e., REP could act as a GAP. To test this hypothesis, REP was incubated with GDP-, GTP-, or GMPPNP-Rab, in the presence of [3H]GGPP and Rab GGTase, and the initial rate of GG transfer was measured. The reaction rate was extremely fast when GDP-Rab was used (Fig. 7, open squares) and slow in the presence of GTP-Rab (Fig. 7, open squares) or GMPPNP-Rab (Fig. 7, open triangles). The difference in the rate of geranylgeranylation of GDP-Rab versus GTP-Rab provided the basis for an assay that measures the formation of GDP-Rab. If GDP-Rab is generated from GTP-Rab, the rate of the reaction will increase. When REP was preincubated with GTP-Rab for 30 min prior to the addition of GGPP and Rab GGTase, the rate of the reaction did not change when compared with the rate obtained in the absence of preincubation (Fig. 7, compare open and closed circles). It is concluded that the REP-Rab preincubation did not result in the formation of GDP-Rab, suggesting that REP has no GAP activity. We also attempted to measure the ability of REP to induce GTP hydrolysis using a standard GAP assay, where Rab1a was loaded with [α-32P]GTP and the hydrolysis of the bound nucleotide was determined by thin-layer chromatography (
). Under many different experimental conditions, no significant REP-dependent stimulation of GTP hydrolysis was observed (data not shown).
The current study suggests that REP has a strict preference for GDP-bound Rabs. The formation of a stable REP·GDP-Rab complex initiates digeranylgeranylation by Rab GGTase, followed by membrane translocation of digeranylgeranylated GDP-Rab.
As discussed in the introduction, the specificity of the nucleotide requirement has been studied previously with conflicting results. A study of Rab5 prenylation using a cell-free translation and prenylation assay in rabbit reticulocyte lysates suggested a preference for GDP-bound Rab (
). The authors compared the prenylation kinetics of wild type and mutant Rab5 proteins that alter the affinity for guanine nucleotides. Wild type Rab5 was prenylated faster than mutants believed to assume preferentially the GTP-bound state (Rab5Q79L) or the nucleotide-free state (Rab5N133I). More recently, an analysis of Rab6 revealed that Rab6Q72R, a GTPase-deficient mutant equivalent to Rab5Q79L was also less efficiently prenylated upon expression in insect cells (
). In contrast, in vitro assays using purified or partially purified components yielded opposite results. In the original study where the purification of Rab GGTase was reported, we found in preliminary experiments that Rab3a prenylation could occur in the presence of either GDP, GTP, or GTP analogs (
) on in vitro prenylation of Rab6 using partially purified enzyme.
How could these studies have generated conflicting results? The studies using purified or partially purified enzymes were performed under conditions that generated very low amounts of product, typically less than 2% of substrate added. Under such conditions, even a small amount of GDP-Rab present in the GTP preparation could have served as substrate and masked the nucleotide effect. The current results were only possible after subjecting Rabs to a procedure that completely strips the pre-bound GDP and replaces it with the desired nucleotide, and the use of recombinant proteins in the assay. As shown above, the results demonstrate dramatically different kinetics between GDP- and GTP-bound substrates.
The dramatic change in the apparent Km of Rab prenylation when GTP-bound substrate was used is likely due to the low affinity of the REP·GTP-Rab complex, suggesting that the Km of the reaction is largely determined by the affinity of the REP·Rab complex. An alteration of the apparent Km of Rab geranylgeranylation was reported in two previous studies. In one study, a chimeric Rab6/Ras protein was constructed composed of Rab6 with an eight-amino acid substitution for Ras sequences in the loop3/β3 region (residues 60–67). Prenylation assays using this chimeric protein exhibited a 30-fold higher apparent Km than ones containing wild type Rab6 (
). A direct demonstration that equilibrium binding constants of Rab27 with either REP-1 and REP-2 are altered is not yet available. Nevertheless, the possibility arises that the Km effect is a direct consequence of a lower affinity REP-2·Rab27 interaction.
The differential affinity of REP toward GTP- and GDP-bound Rabs is not entirely surprising because REP shows structural and functional homology to Rab GDI, a protein with strict preference for GDP-Rab (
). The present findings imply that at least one domain of Rab that interacts with REP lies within a region that is nucleotide-sensitive. There are at least four regions in Rabs proposed to be involved in binding REP: the amino-terminal region including a conserved lysine residue (
). Of these, the effector domain region is the one that is most altered upon nucleotide exchange and therefore is a good candidate for a major interacting domain. A more detailed analysis will be needed to test these hypotheses.
The evidence presented here suggests that REP is unable to act as a GAP for Rab proteins. This finding is in agreement with a previous suggestion that GTP hydrolysis is not necessary for prenylation of Rab proteins in reticulocyte lysates (
). Therefore, it remains to be explained how REP recognizes newly synthesized Rabs. It is unlikely that Rabs are prenylated passively because they slowly hydrolyze their bound GTP. Rab GTPase rates are extremely slow, and it may take hours to convert all GTP into GDP. Evidence suggests that prenylation is efficient and that at steady-state there is no significant pool of unprenylated Rabs (
). Alternatively, it is possible that REP acts as a true chaperone by binding to Rabs as the polypeptide is released from the translational machinery and forcing nascent Rabs to the GDP-bound conformation. The same role could be played by molecular chaperones, which would present GDP-Rab to REP. It is also impossible to rule out at present that the initial form of the newly synthesized protein is preferentially occupied with GDP. However, a more likely alternative is that a putative Rab GAP recognizes newly synthesized Rabs, stimulates GTP hydrolysis, and presents GDP-Rab to REP, so it can undergo prenylation. It is interesting to note that a cytosolic Rab GAP activity has been described (
), raising the possibility that such activities exist. Future studies detailing the molecular mechanisms that regulate REP·Rab interaction should provide further glimpses into the role of prenylation in Rab function and its importance to cells.
I thank the members of my lab for many insightful suggestions, Mike Brown and Joe Goldstein for support and critical reading of the manuscript, and Richard Gibson for preparing recombinant proteins.