Single prenyl-binding site on protein prenyl transferases

Abstract

Three distinct protein prenyl transferases, one protein farnesyl transferase (FTase) and two protein geranylgeranyl transferases (GGTase), catalyze prenylation of many cellular proteins. One group of protein substrates contains a C-terminal CAAX motif (C is Cys, A is aliphatic, and X is a variety of amino acids) in which the single cysteine residue is modified with either farnesyl or geranylgeranyl (GG) by FTase or GGTase type-I (GGTase-I), respectively. Rab proteins constitute a second group of substrates that contain a C-terminal double-cysteine motif (such as XXCC in Rab1a) in which both cysteines are geranylgeranylated by Rab GG transferase (RabGGTase). Previous characterization of CAAX prenyl transferases showed that the enzymes form stable complexes with their prenyl pyrophosphate substrates, acting as prenyl carriers. We developed a prenyl-binding assay and show that RabGGTase has a prenyl carrier function similar to the CAAX prenyl transferases. Stable RabGGTase:GG pyrophosphate (GGPP), FTase:GGPP, and GGTase-I:GGPP complexes show 1:1 (enzyme:GGPP) stoichiometry. Chromatographic analysis of prenylated products after single turnover reactions by using isolated RabGGTase:GGPP complex revealed that Rab is mono-geranylgeranylated. This study establishes that all three protein prenyl transferases contain a single prenyl-binding site and suggests that RabGGTase transfers two GG groups to Rabs in independent and consecutive reactions.

Protein prenylation is a lipid modification that affects between 0.5 and 2% of cellular proteins (1–4). Prenylation serves an essential function in the modified protein as a determinant of protein–protein as well as lipid–protein interactions. Prenylated proteins are modified covalently with either farnesyl or geranylgeranyl (GG) via thioether bonds to C-terminal cysteine residues.

Three distinct protein prenyl transferases, one protein farnesyl transferase (FTase) and two protein GG transferases (GGTase), have been isolated and characterized from a great variety of organisms, from yeast to plants to mammals (3). Protein prenyl transferases are named according to the lipid substrate and are classified in two groups according to the protein substrate: the CAAX prenyl transferases comprising farnesyl transferase and GGTase type-I (GGTase-I), and the Rab GG transferase (RabGGTase), also known as GGTase-II.

CAAX prenyl transferases recognize a conserved motif at the C termini of their substrates termed CAAX box (where C is cysteine, A is aliphatic, and X is a variety of amino acids present at the C terminus) (5). The cysteine residue is the site of prenylation, and the last amino acid in the CAAX sequence is the primary determinant for discrimination between the CAAX transferases. If the last amino acid in the CAAX sequence is a methionine or serine, the protein is a substrate of FTase whereas if it is a leucine, it becomes a substrate for GGTase-I (3). 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.

Previous biochemical characterization of FTase revealed that the enzyme possesses an farnesyl pyrophosphate (FPP)-carrier function (6). Stable FTase:FPP complexes were isolated, and the farnesyl moiety was transferred specifically to a Ras substrate. Comparable results were reported for GG pyrophosphate (GGPP) binding to GGTase-I (7). All three known protein prenyl transferases are evolutionarily conserved, ubiquitously expressed heterodimeric enzymes, composed of α and β subunits. The CAAX-recognizing enzymes share a common α subunit in combination with a distinct β subunit, which determines the different protein and lipid substrate specificity (3). Recently, the elucidation of three-dimensional structure of FTase revealed that the α subunit forms a crescent-shaped, double-layered, seven-helical hairpin domain that envelops part of the β subunit (8). The β subunit is also predominantly α helical possessing an α–α barrel, with six helices forming the inner barrel and six additional helices forming the outer barrel. The bottom end of the barrel is closed by a loop, and the top end is open to the solvent. A cleft formed by the interface of the α and β subunits is postulated to be the peptide (CAAX) binding site. The deep cleft in the β subunit is lined by conserved aromatic amino acids and is postulated to form a single FPP binding site.

RabGGTase acts on Rab proteins, a distinct family of the Ras GTPase superfamily (3, 9). Over 40 distinct genes in the Rab family have been identified, and most contain two cysteine residues near or at the C terminus in different motifs such as XXCC, XCXC, or CCXX. Structural analysis of prenylated Rab3a (ending in XCXC) purified from bovine brain revealed that both cysteines are modified with GG groups (10). The same finding was observed after analysis of recombinant Rab3a after in vitro prenylation by recombinant RabGGTase (11). Other members of the Rab family, including Rab1a (ending in XXCC) and Rab5a (ending in CCXX), also were shown to be digeranylgeranylated in vitro by RabGGTase (11). An important mechanistic difference between RabGGTase and the CAAX prenyl transferases is the recognition of the protein substrate. RabGGTase has a very low affinity for the Rab protein substrates, and Rabs are recognized only when present in a stable complex with Rab escort protein (REP) (12). REP:Rab appears to be the substrate for RabGGTase.

The main focus of this work was to study prenyl-binding to RabGGTase and compare it to the CAAX prenyl transferases. We show stable association between GGPP and RabGGTase and define the stoichiometry of the interaction to be 1:1, an important mechanistic issue because RabGGTase can transfer two GG groups to the same protein substrate. We also demonstrate that the CAAX prenyl transferases exhibit the same stoichiometry of GGPP-binding, suggesting that all protein prenyl transferases contain a single prenyl binding site.

EXPERIMENTAL PROCEDURES

Materials.

All-trans-[1-³H]-GGPP (20 Ci/mmol) was purchased from DuPont/NEN. Unlabeled all-trans-GGPP, all-trans-FPP, hydroxyalkoxypropyl Dextran type VI (also known as Lipidex 1000) and trypsin were obtained from Sigma. The chemical and radiometric purity of the prenyl pyrophosphates were assayed by thin-layer chromatography by using an n-propanol/ammonium hydroxide/H₂O (6:3:1) solvent system as described (13). The purity of the prenyl pyrophosphate groups was higher than 95%. Acetonitrile, water, trifluoroacetic acid and methanol used for HPLC were of HPLC grade and were purchased from Fisher. All other chemicals used were of analytical grade and were obtained from commercial suppliers.

Purification of Recombinant Rab, REP, and Protein Prenyl Transferases.

All procedures were conducted at 4°C. Recombinant histidine-tagged REP2, FTase, and GGTase-I were produced in insect cells by using the baculovirus system and were purified to homogeneity by using the methods described with minor modifications (14–17). Recombinant histidine-tagged Rab1a-CC and Rab1a-CS were overexpressed in bacteria and were purified as described (16).

In each case, after elution from Ni²⁺-Sepharose, the appropriate fractions were pooled, were concentrated by ultrafiltration, and were applied onto a Superdex 200 16/60 column (Pharmacia) equilibrated with 20 mM sodium Hepes (pH 7.2), 100 mM sodium chloride, and 1 mM DTT. For the chromatography of Rab1a, 3 mM MgCl₂ and 0.1 mM GDP were supplemented in the buffers whereas 10 μM ZnCl₂ was added for the chromatography of FTase and GGTase-I. The fractions corresponding to the appropriate molecular weight were pooled, were concentrated, were divided into multiple aliquots, and were stored at −70°C until used. We observed that, after long term storage, FTase and GGTase-I formed oligomers. In some experiments, FTase and GGTase-I were rechromatographed on Superdex 200 3.2/30 on a SMART system (Amersham Pharmacia BioTech) under identical buffer conditions to isolate the monomeric active form.

RabGGTase was purified by using the method published (18, 19) with minor modifications. In brief, Sf9 cells were co-infected with recombinant baculovirus encoding the α and the β subunits and were grown and lysed. The 100,000 × g supernatant was applied on a Q-Sepharose column (1.5 × 10 cm; Pharmacia) equilibrated with 50 mM Tris⋅HCl (pH 7.9) and 1 mM DTT and was run at a flow rate of 1 ml/min on an FPLC system. The adsorbed proteins were eluted with a 0.15–0.6 M linear gradient of NaCl in the same buffer, and fractions of 5 ml were collected and analyzed by SDS/PAGE for the presence of the enzyme. The appropriate fractions were pooled, concentrated, and applied onto a Superdex 200 16/60 column equilibrated in 20 mM sodium Hepes (pH 7.2), 150 mM NaCl, and 1 mM DTT. The flow rate was 1.5 ml/min, and fractions of 1 ml were collected. The fractions containing the enzyme were pooled and concentrated, and the concentration of ammonium sulfate was adjusted to 0.6 M. The material was loaded on a phenyl-Superose 10/10 column (Pharmacia) equilibrated with 50 mM sodium Hepes (pH 7.2), 0.6 M ammonium sulfate, and 1 mM DTT. The flow rate was 1 ml/min, and the adsorbed proteins were eluted by applying a linear gradient of ammonium sulfate from 0.6 to 0.24 M in the same buffer. The appropriate fractions were pooled, were dialyzed against 20 mM sodium Hepes (pH 7.2), 100 mM NaCl, and 1 mM DTT, were concentrated, were divided in multiple aliquots, and were stored at −70°C. The purified FTase, GGTase-I, and RabGGTase were delipidated according to a published method (20) modified as described (21).

[³H]GGPP Binding Assay.

A modification of the method described by Glatz and Veerkamp (20) was used. All polypropylene microcentrifuge tubes and micropipette tips used in the assays were siliconized. Purified and delipidated proteins (10 pmol) were incubated with various concentrations of [³H]GGPP in 10 mM potassium phosphate buffer (pH 7.4) at room temperature. The total concentration of prenyl group in the set of assay tubes ranged from 6.25 to 500 nM. The final volume was 200 μl, and the total concentration of alcohol did not exceed 2%. After 10 min, the tubes were chilled on ice for 10 min. Two-hundred microliters of a continuously stirred 50% (vol/vol) hydroxyalkoxypropyl Dextran suspension in 10 mM potassium phosphate buffer (pH 7.4) were added followed by vigorous stirring. The tubes were incubated for 30 min on ice and were vortexed every 10 min. The samples were centrifuged at 15,000 × g for 4 min, and the radioactivity in a 100-μl aliquot of the supernatant was determined by scintillation counting by using a Beckman LS 6000 IC scintillation counter. Each value was determined in duplicate and was corrected against a blank (no added protein) performed in parallel. The blanks were never >7% of the total amount of ligand added.

Preparation of RabGGTase:[³H]GGPP Complex.

One-hundred picomoles RabGGTase were incubated with 500 pmol [³H]GGPP (20 × 10⁶ dpm) in 100 μl of 10 mM potassium phosphate (pH 7.4) for 10 min at room temperature. The tube then was chilled on ice for 10 min, and 100 μl of a continuously stirred 50% (vol/vol) suspension of hydroxyalkoxypropyl Dextran in 10 mM potassium phosphate (pH 7.4) was added. The mixture was kept on ice for 30 min and was vortexed vigorously every 10 min. The sample was centrifuged at 15,000 × g for 4 min, and the supernatant was collected and used immediately.

RabGGTase Assays.

In single turnover assays, 5 pmol RabGGTase saturated with [³H]GGPP in a volume of 10 μl was added to a 15-μl prenylation mixture containing 50 mM sodium Hepes (pH 7.2), 5 mM MgCl₂, 1 mM DDT, 25 pmol Rab1a-CC or Rab1a-CS, and 50 pmol REP2 and was incubated for 5 min at 37°C. For quantitation purposes, the reaction was stopped by adding 250 μl of 4 mg/ml casein and 2 ml of 10% (vol/vol) HCl in ethanol, and the radioactivity was measured by a filter-binding assay (22, 23). Multiple turnover assays were performed as single turnover assays, except that the reaction mixture was supplemented with 4 μM GGPP and was incubated for 3 h at 37°C.

Proteolytic Digestion of Prenylated Rabs and Prenyl-Peptide Analysis.

Single turnover prenylation reaction mixtures (50 μl) were denatured by adding 24 mg of urea and 5 μl of 100 mM DTT followed by incubation at 50°C. After 15 min, the tubes were cooled on ice, and 322 μl of H₂O, 5.5 μl of 500 mM CaCl₂, and 6.25 μl of 4 mg/ml trypsin in 100 mM ammonium bicarbonate (pH 8.0) were added. The digests were incubated at 37°C overnight (≈18 h). The generated peptides were analyzed by HPLC on a 250 × 4.5 mm Vydac C₁₈ column (The Separations Group). The concentrations of acetonitrile and trifluoroacetic acid in the samples were adjusted to 50% and 0.06%, respectively. The chromatograms were developed in a binary system consisting of solvent A [0.06% trifluoroacetic acid in Milli-Q water (Millipore)] and solvent B (0.052% trifluoroacetic acid in acetonitrile). After injection onto the column, peptides were separated by a 25-min linear gradient from 50 to 90% of solvent B at 0.5 ml/min, and fractions of 0.5 ml were collected. The radioactivity in each fraction was measured by scintillation counting. The column was reconditioned with five volumes of chloroform between each analysis. Multiple turnover assays using Rab1a-CS and Rab1a-CC, respectively, were subjected to identical proteolysis and HPLC analysis to generate mono- and digeranylgeranylated peptides used as controls.

Protein Determination.

The extinction coefficient of the purified proteins was determined by using the method of Edelhoch (24). The concentrations of the pure recombinant proteins then were measured by using their respective extinction coefficient at 280 nm. The extinction coefficients used were RabGGTase, 137,300 M⁻¹⋅cm⁻¹; Ftase, 164,715 M⁻¹⋅cm⁻¹; GGTase-I, 128,756 M⁻¹⋅cm⁻¹_; REP2, 53,680 M⁻¹⋅cm⁻¹; and Rab1a, 32,380 M⁻¹⋅cm⁻¹.

RESULTS

GGPP Binding to RabGGTase.

Lipid binding assay using hydroxyalkoxypropyl dextran (HAPD) was used previously for determination of the stoichiometry of binding of hydrophobic ligands such as acyl-CoA and fatty acids to fatty acid-binding proteins (20). We adapted this procedure to study association of GGPP with protein prenyl transferases. The assay is based on the ability of HAPD to bind to free lipids at 4°C. GGPP binding to RabGGTase was determined by estimation of the radiolabeled GGPP complexed with the enzyme in solution once the free ligand was sedimented with HAPD by gentle centrifugation. A typical result of the binding assay is presented in Fig. 1. GGPP exhibited saturable binding to RabGGTase. In an assay in which 10 pmol of enzyme were added per tube, the binding curve reaches a plateau (B_max) at 9.2 ± 0.55 pmol of GGPP bound, indicating a stoichiometry of 0.92:1 (moles of GGPP:moles of enzyme). No significant loss of protein caused by nonspecific adsorption to the dextran or to the surface of the tube was detected. The addition of 100 μM Triton-X100 to the incubation buffer, as recommended by Vork et al. (25) to limit the nonspecific surface adsorption of proteins or lipids, did not change the binding properties of RabGGTase (data not shown). We routinely delipidated the purified RabGGTase before the binding assays to prevent possible interference by prebound GGPP to the purified enzyme. However, we found no significant difference between the delipidated enzyme versus the nondelipidated one (data not shown). REP acts as an essential cofactor of RabGGTase and binds prenylated Rab, raising the possibility that it may bind GGPP and/or influence GGPP binding to RabGGTase (12). However, we detected no significant binding of GGPP to REP (Fig. 1) or to the REP:Rab complex (data not shown). Also, the presence of an excess of REP over RabGGTase in the binding assay did not change significantly the affinity of the enzyme for GGPP or its stoichiometry of binding (Fig. 1).

Figure 1.

Binding of [³H]GGPP to RabGGTase and REP. Ten picomoles of RabGGTase (filled circles), REP2 (filled squares), or RabGGTase with 25 pmol of REP2 (open circles) were incubated in 200 μl of 10 mM potassium phosphate buffer (pH 7.4) with varying concentrations of [³H]GGPP (0–0.5 μM). Free lipid was sedimented by using HAPD, and the specific binding was determined as described in Experimental Procedures. All points represent mean of duplicate incubations.

A competition study using this binding assay was devised to analyze the binding specificity of the RabGGTase for different prenyl groups. The enzyme was incubated with a constant amount of [³H]GGPP and increasing concentrations of nonradiolabeled prenyl pyrophosphates. After the incubation, free lipids were sedimented by using HAPD, and the amount of enzyme complexed to [³H]GGPP remaining in solution was determined by scintillation counting. As shown in Fig. 2, GGPP inhibited competitively the binding of [³H]GGPP to the enzyme. However, neither geranyl pyrophosphate nor FPP at concentrations up to 1.6 μM were able to prevent [³H]GGPP from binding to the enzyme. When we studied FPP binding to all three prenyl transferases by direct binding and competition studies, we observed that FTase is the only protein prenyl transferase that binds FPP with high affinity (data not shown). Using this technique, we have not been able to study accurately whether there is low affinity FPP binding to RabGGTase as described for GGTase-I (26).

Figure 2.

Inhibition of [³H]GGPP binding to RabGGTase by prenyl pyrophosphates. A fixed concentration of [³H]GGPP (0.625 μM) and various concentrations (0–1.6 μM) of GGPP (filled circles), FPP (open circles), or geranyl pyrophosphate (GPP; filled triangles) were incubated with 10 pmol of RabGGTase in 200 μl of 10 mM potassium phosphate buffer (pH 7.4). The binding of [³H]GGPP was determined by using HAPD as described in Experimental Procedures. All points represent mean of duplicate incubations.

Stoichiometry of GGPP binding.

To compare the GGPP-binding stoichiometry with the CAAX prenyl transferases, we used this assay to analyze GGPP binding to all three enzymes under similar experimental conditions. As shown in Fig. 3, the hyperbolic response observed for all three protein prenyl transferases is very similar. The observed stoichiometry was 1.04:1 (moles of GGPP:moles of enzyme) for GGTase-I and 1.03:1 for FTase, values comparable to those obtained with RabGGTase. We conclude that the stoichiometry of binding for the three enzymes is almost identical, predicting a 1:1 molar ratio in the interaction.

Figure 3.

Binding of [³H]GGPP to three protein prenyl transferases. Ten picomoles of RabGGTase (filled circles), GGTase-I (open circles), or FTase (filled squares) were incubated in 200 μl of 10 mM potassium phosphate buffer (pH 7.4) with varying concentrations of [³H]GGPP (0–0.5 μM). The free lipid was sedimented by using HAPD, and the specific binding was determined as described in Experimental Procedures. All points represent mean of duplicate incubations.

Analysis of Prenylated Rab after Single Turnover.

We analyzed the transfer of prebound GGPP in a single enzymatic turnover. First, we incubated RabGGTase with a molar excess of [³H]GGPP and isolated the RabGGTase:[³H]GGPP complex after sedimentation of free [³H]GGPP using HAPD. Then, the RabGGTase:[³H]GGPP complex was incubated with an excess of Rab acceptor in the presence of REP under conditions that allowed transfer of GG to Rab. The covalent attachment of the GG moiety to Rab represents a single turnover reaction because the only GGPP available for the reaction was precomplexed with RabGGTase. After prenylation, we analyzed the reaction products with a filter-binding assay (not shown) and proceeded with the HPLC analysis of the Rab1a-derived prenylated peptides generated. Reaction products were digested extensively, and the resulting peptides were analyzed by HPLC. As shown in Fig. 4, the elution profile of the radiolabeled ([³H]GG-containing) peptides coincided with the position of monoprenylated products. This experiment suggests that only one GG group is transferred to Rab in a single turnover and confirms the 1:1 stoichiometry of the interaction between GGPP and RabGGTase. To test whether a second GG group is transferred in a consecutive step, we performed a single turnover reaction and then allowed it to proceed by adding more GGPP. We observed that the resulting product was digeranylgeranylated (Fig. 4).

Figure 4.

HPLC analysis of [³H]GG-peptides after single turnover prenylation. Single turnover prenylation reactions using Rab1a-CC were carried out by using a preformed RabGGTase:GGPP complex as described in Experimental Procedures. The reactions were followed by the addition of excess unlabeled GGPP (filled circles) or buffer alone (open circles), were denatured, were digested with trypsin, and were chromatographed by reverse-phase HPLC on a C₁₈ column, and the radioactivity present in each elution fraction was measured by scintillation counting as described in Experimental Procedures. The arrows indicate the elution positions of mono- and di-geranylgeranylated peptides obtained from multiple turnover prenylation of Rab-1a-CS and Rab-1a-CC, respectively.

DISCUSSION

The major conclusion from the present study is that the three known protein prenyl transferases possess a single prenyl-binding site. We developed a prenyl-binding assay that allows a quantitative recovery of enzyme-bound prenyl groups and a determination of the stoichiometry of interaction. Using this assay, we show here that RabGGTase exhibits stable GGPP-binding, a property similar to that originally described for FPP and FTase (6). The enzyme has high affinity for GGPP and little, if any, for FPP. This data agrees with a previous report that showed that FPP was unable to inhibit RabGGTase activity (18) and from unpublished findings in which we failed to incorporate farnesyl into Rab substrates using RabGGTase.

We demonstrate that 1 mol of RabGGTase binds 1 mol of GGPP, suggesting that it possesses a single binding site for GGPP. We also show the same stoichiometry for the CAAX prenyl transferases. Despite a number of previous studies on lipid substrate binding to CAAX prenyl transferases, the stoichiometry of binding never was demonstrated conclusively (6, 7, 14, 26–30).

The finding that RabGGTase contains only one GGPP binding site has important mechanistic implications. It was shown that the recombinant enzyme can efficiently digeranylgeranylate three different substrates, Rab1a, Rab3a and Rab5a (11), with very little, if any, accumulation of monoGG-Rab. In a subsequent study, it was suggested that each GG addition is an independent reaction leading to two consecutive products, monoGG-Rab and diGG-Rab (16). The new evidence presented in this study further suggests that the two GG additions are independent because RabGGTase contains only one GGPP-binding site.

Taking the new findings in consideration, we propose that the mechanism for the Rab digeranylgeranylation reaction is as follows. RabGGTase loaded with GGPP binds REP:Rab and catalyzes the transfer of the first GG group to Rab (31). After this transfer has occurred, the first GG group covalently attached to Rab may dissociate from the RabGGTase active site and interact with REP, Rab, or the REP:Rab interface. The displacement of the transferred GG group from the RabGGTase binding site may be driven by the binding of another molecule of GGPP to RabGGTase. In FTase, product release depends on binding of a new substrate, either FPP or Ras (28). The second GGPP seems to be used preferentially on monoGG-Rab rather than on a new unprenylated Rab, judging from the almost exclusive generation of digeranylgeranylated Rabs in vitro and in vivo (11). RabGGTase processes preferentially monoGG-Rab substrates either because it exhibits higher catalytic efficiency toward REP:monoGG-Rab substrate as compared with REP:Rab or because the dissociation of REP:monoGG-Rab is slower than the second GG transfer reaction. Future studies should attempt to distinguish between a processive and a nonprocessive mechanism by determining whether RabGGTase dissociates from REP:monoGG-Rab after the first GG transfer and by measuring the affinity of RabGGTase toward its two REP:Rab substrates.