SopE Acts as an Rab5-specific Nucleotide Exchange Factor and Recruits Non-prenylated Rab5 on Salmonella-containing Phagosomes to Promote Fusion with Early Endosomes*

Rab-GTPase regulates the fusion between two specific vesicles. It is well documented that, for their biological function, Rab proteins need to be prenylated for attachment to the vesicle membrane. In contrast, we showed in the present investigation that SopE, a type III secretory protein of Salmonella, translocates onto Salmonella-containing phagosomes (LSP) and mediates the recruitment of non-prenylated Rab5 (Rab5:ΔC4) on LSP in GTP form. Simultaneously, SopE present in infected cell cytosol acts as an Rab5-specific exchange factor and converts the inactive Rab-GDP to the GTP form. The non-prenylated Rab5 subsequently promoted efficient fusion of LSP with early endosomes. This is the first demonstration that a prenylation-deficient Rab protein retains biological activity and can promote vesicle fusion, if it is recruited on the membrane by some other method.

murium (a clinical isolate from Lady Harding Medical College, New Delhi, India) was obtained from Dr. Vineeta Bal of the National Institute of Immunology, New Delhi, India. Salmonella dublin wild type strain (2229) and SopE knockout mutant S. dublin (SE1 CMR) were kindly provided by Dr. E. E. Galyov. Bacteria were grown overnight in Luria broth (LB) at 37°C with constant shaking (300 rpm), washed twice in phosphate-buffered saline (PBS), and used in phagosome preparation.
Determination of the Molecules from the Host Cell Cytoplasm Recognized by the Salmonella-To determine the molecules from the host cell cytoplasm recognized by the Salmonella, these bacteria were incubated in the presence of biotinylated cytosol prepared from the macrophages. Macrophage cytosol was biotinylated using NHS-biotin as described previously (17). Briefly, 10 mg of macrophage cytosol was incubated with 5.5 mg of NHS-biotin for 2 h at room temperature in a biotinylation buffer (0.1 M sodium carbonate buffer, pH 9.0). Unreacted biotin was quenched with glycine, and biotinylated cytosol was separated from unreacted biotin using 1 ml of a G-25 Sephadex spin column. Biotinylated cytosol (2.5 mg) was incubated with 1 ϫ 10 8 Salmonella at 4°C for 1 h in PBS (pH 7.2). Subsequently, cells were washed five times with PBS and lysed in SDS sample buffer. Proteins were separated by 12% SDS-PAGE and were transferred onto nitrocellulose membrane. The biotinylated macrophage proteins bound to Salmonella were detected by Western blot analysis using avidin-HRP. Finally, cytosolic proteins recognized by the Salmonella were identified by Western blot analysis using specific antibodies against endocytic Rabs and actin.
Detection of Rab5 Binding Protein from Salmonella-To detect the Rab5 binding protein from Salmonella, Salmonella were grown overnight in LB and metabolically labeled with [ 35 S]methionine (18). Briefly, cells were washed three times with PBS and grown in methionine-free RPMI 1640 medium containing 1 mCi of [ 35 S]methionine with constant shaking (300 rpm) for 9 h at 37°C. The cells were washed five times with PBS to remove unincorporated radioactivity. Bacteria were lysed by sonication followed by incubation in 1% Triton X in PBS at 4°C for 30 min. Lysate was centrifuged at 12,000 rpm to get rid of unbroken cells and other debris. GST-Rab5 (200 g) was immobilized with glutathione beads as described previously (15) and incubated in the presence of the Salmonella lysate for 1 h at 4°C. Beads were washed (10,000 ϫ g for 5 min) three times to remove unbound proteins. Subsequently, the proteins were separated by 12% SDS-PAGE. The gel was dried and autoradiographed at Ϫ70°C. A similar experiment was carried out with unlabeled Salmonella lysate, and the indicated protein was identified using antibodies against SopE by Western blot analysis. Proteins were visualized using an appropriate HRP-labeled second antibody by ECL.
Specificity of Rab5 Binding with Salmonella-To determine the specificity of Rab5 binding with Salmonella, first the GST-Rab5 was biotinylated using the procedure as described previously (17). Biotinylated GST-Rab5 (10 g/ml) was incubated in PBS (pH 7.2) containing 1 ϫ 10 8 Salmonella at 4°C for 1 h in the presence or absence of a 100-fold excess of non-biotinylated Rab5 or SopE. Subsequently, cells were washed five times with PBS and lysed in SDS sample buffer. Proteins were separated by 12% SDS-PAGE and transferred onto a nitrocellulose membrane. The biotinylated Rab5 bound to Salmonella bacteria in the presence or absence of competitors were detected by Western blot analysis using avidin-HRP.
Purification of Salmonella-containing Phagosomes-Phagosomes containing wild type S. typhimurium, S. dublin (2229), or SopE knockout mutant S. dublin were prepared using a procedure described previously (15). Briefly, J774E clone macrophages (1 ϫ 10 8 ) were incubated with 1 ϫ 10 9 bacteria at 4°C for 1 h to allow binding. Then Salmonella were internalized into macrophages for 5 min at 37°C. Finally, cells were washed (300 ϫ g for 6 min) and resuspended (2 ϫ 10 8 cells/ml) in homogenization buffer (HB; 250 mM sucrose, 0.5 mM EGTA, 20 mM HEPES-KOH, pH 7.2) and homogenized in a ball bearing homogenizer at 4°C. Homogenates were centrifuged at a low speed (400 ϫ g for 5 min) at 4°C to remove nuclei and unbroken cells. The post nuclear supernatant was diluted with HB (1:3) and centrifuged at 12,000 ϫ g for 1 min in a microcentrifuge at 4°C. Subsequently, enriched phagosomal fractions were resuspended in 100 l of HB containing protease inhibitors and loaded onto a 1-ml 12% sucrose cushion. Samples were centrifuged at 1700 rpm for 45 min at 4°C, and the purified phagosomes were recovered from the bottom of the tube. The viability of the bacteria in the phagosomes was determined by lysing the respective phagosomes with solubilization buffer (PBS containing 0.5% Triton X-100) and plated them on an LB-agar plate. We found a similar number of colonies on the LB-agar plate, indicating the presence of a comparable number of viable bacteria in respective phagosomes. Biochemical characteriza-tion of these phagosomes demonstrated that they were free of endosome, lysosome, Golgi, and endoplasmic reticulum contamination as observed previously (15).
Electron Microscopic Observation of the Purified Phagosomes-To check the purity of the phagosomes, purified phagosomes were washed five times with ice-cold PBS and fixed in 1% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.2, washed, and postfixed with 1% OsO 4 in the same buffer. The phagosomes were rinsed and dehydrated in ethanol and embedded in araldite (16). Thin sections were doublestained with uranyl acetate and lead citrate and examined in a transmission electron microscope (JEOL 1200 EX 11).
Immunolabeling of Rab5 and SopE on Salmonella-containing Phagosomes-The SopE present on purified LSP and DSP was detected by immunogold labeling using a negative staining technique as described previously (15). Briefly, respective phagosomes were purified and washed five times with ice-cold homogenization buffer (250 mM sucrose, 0.5 mM EGTA, and 20 mM Hepes-KOH, pH 7.2) and sedimented by centrifugation. First, the purified phagosomes were absorbed with glow-discharged formver and carbon-coated nickel grids, and the specimens were quickly rinsed twice with HB and incubated for 30 min in blocking buffer (HB containing 3% skim milk and 0.1% gelatin). The samples were then incubated for 2 h with mouse anti-SopE antibody (monoclonal) diluted 1:40 in blocking buffer. Subsequently, the specimens were rinsed three times (5 min each) with blocking buffer and incubated for 1 h with goat anti-mouse conjugated with 12-nm colloidal gold at a 1:20 dilution. The samples were washed twice with HB and fixed in 1% glutaraldehyde in HB for 10 min. Finally, samples were sequentially washed with HB and distilled water, stained with 0.5% aqueous uranyl acetate for 1 min, blotted onto filter paper, and airdried. Similarly, Rab5 was immunolocalized on phagosomes containing either wild type S. dublin or the SopE knockout mutant of S. dublin using polyclonal rabbit anti-Rab5 antibody (1:200 dilution) followed by treatment with goat anti-rabbit IgG conjugated with 18-nm colloidal gold. Moreover, the presence of SopE was also determined on phagosomes containing either wild type S. dublin or the SopE knockout mutant of S. dublin using anti-SopE antibody.
Removing the Rab Protein from Phagosomes by GDI Treatment-To strip off the Rab protein from each phagosome, the phagosomes were treated with Rab-GDP dissociation inhibitor (GDI) as described previously (15). Briefly, respective phagosomes (150 g each) were preincubated with fusion buffer containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 20 g/ml leupeptin, and 20 g/ml aprotinin) for 20 min at room temperature in the presence of 1 mM GDP. Subsequently, 6 g of the purified GDI was added to one set of phagosomes in the reaction mixture, and incubation was carried out for another 10 min at room temperature. Phagosomes were sedimented by centrifugation (10,000 ϫ g for 5 min), and these Rab-stripped phagosomes were washed with PBS and used for indicated experiments.
Recruitment of Rab5 and Its Mutant Proteins on Salmonella-containing Phagosomes-To determine the recruitment of different forms of Rab5 proteins, purified phagosomes containing respective Salmonella were treated with Rab-GDI to remove the endogenous Rabs present on the phagosomes (15). Respective Rab fusion proteins were preincubated in the presence of J774E cytosol (3.5 mg/ml) for 30 min at 37°C in fusion buffer (250 mM sucrose, 0.5 mM EGTA, 20 mM Hepes-KOH, pH 7.2, 1 mM dithiothreitol, 1.5 mM MgCl 2 , 100 mM KCl, including an ATP-regenerating system, 1 mM ATP, 8 mM creatine phosphate, 31 units/ml creatine phosphokinase) for in vitro prenylation (19). Subsequently, recruitment of the different forms of Rab5 by the respective phagosomes was carried out by incubating the phagosomes at 37°C for 10 min in fusion buffer containing 1 mg/ml cytosol supplemented with 30 ng of the indicated purified GST-Rab5 proteins. The presence of Rab5 (50 kDa of GST-Rab5) on the phagosomes was determined by Western blot analysis using specific antibodies against Rab5. Similar studies were carried out to determine the recruitment of the Rab7 by respective phagosomes.
Direct Interaction of Different Forms of Rab5 with SopE-To determine the direct interaction of different forms of Rab5 mutant proteins with SopE, we developed an in vitro assay to study the protein-protein interaction. First, the recombinant SopE-(78 -240) (10 g/ml) was incubated in an ELISA plate for 3 h at 37°C in coating buffer (0.1 N sodium carbonate buffer, pH 9.5) to coat the protein, and subsequently, wells were washed thrice and incubated for 1 h at 37°C in blocking buffer (PBS containing 0.01% of Tween 20 and 2% of bovine serum albumin). Wells were washed three times with PBS-Tween, and 0.2 mg/ml of different constructs of GST-Rab5:WT or mutant proteins was incubated in PBS for 1 h at 37°C to allow binding. To determine the binding of Rab5 with SopE, wells were incubated with Rab5-specific polyclonal antibody (1:2000 dilution) in PBS-Tween for 1 h at 37°C. Excess antibodies were removed by washing the wells three times with PBS-Tween followed by three washes with PBS. Subsequently, wells were incubated with secondary antibodies labeled with HRP (1:10,000 dilution) for 1 h at 37°C, washed five times, and finally the HRP activity present in each well was measured as described previously (15) to measure the binding of different preparation of Rab5 with SopE.
Binding of SopE with Rab5-To determine the specificity of Rab5 binding with SopE, similar experiments were carried out using biotinylated Rab5 subsequently probed with avidin-HRP. First, the recombinant SopE-(78 -240) (10 g/ml) was coated in an ELISA plate as described in the previous section, washed, and incubated in the presence of different concentrations of biotinylated GST-Rab5:WT in PBS for 1 h at 37°C to allow binding. To determine the binding of biotinylated Rab5 with SopE, wells were washed and incubated with avidin-HRP (1 g/ml) for 30 min at 37°C. Unbound avidin-HRP were removed by washing the wells three times with PBS-Tween followed by three washes with PBS. Subsequently, HRP activity present in each well was measured as described previously (15) to measure the binding of Rab5 with SopE. Binding of biotinylated Rab5 (10 g/ml) with immobilized SopE (10 g/ml) was carried out in the presence of different concentrations of non-biotinylated Rab5 or SopE to determine the specificity.
Removal of Rab5 from the Cytosol-Immunodepletion of Rab5 from the cytosol was carried out using the procedure described previously (15). Briefly, 100 l of protein A/G plus-agarose (Santa Cruz Biotechnology, Santa Cruz, CA) was incubated with 10 l of anti-Rab5 antibody in PBS overnight at 4°C. The antibody-protein A/G agarose complex was washed and centrifuged at 10,000 ϫ g for 5 min at 4°C. Subsequently, 100 l of J774E cytosol (600 g) was added to the protein A/G agarose-antiRab5 complex and incubated for 2 h at 4°C to deplete the Rab5 from the cytosol. Subsequently, Rab5-depleted cytosol was separated from the agarose beads by centrifugation. Immunodepletion of Rab5 from the cytosol was confirmed by Western blot analysis using an anti-Rab5 antibody. Rab5-depleted cytosol was used for the in vitro fusion assay.
SopE-mediated Nucleotide Exchange of Rab5-The SopE-mediated nucleotide exchange of Rab5 was determined using an assay previously described (20), with some modifications. First, the respective Rab 5 proteins were loaded with GDP and subsequently, the exchange of GDP to [ 32 P]GTP was determined in the presence or absence of GST-SopE-(78 -240). Briefly, different mutants of Rab5 (60 pmol of each protein) was incubated at room temperature for 30 min in loading buffer (20 mM Tris-HCl (pH 7.5), 5 M GDP, 50 mM NaCl, 3 mM MgCl 2 , 0.1 mM dithiothreitol, 0.1 mM EDTA). Exchange reaction was carried out in 200 l of exchange buffer (20 mM Tris-HCl (pH 7.5), 5 M [ 32 P]GTP, 100 mM NaCl, 10 mM MgCl 2 , 0.5 mM dithiothreitol, 0.5 mg/ml bovine serum albumin) containing 15 pmol of GDP-bound Rab proteins with/without an equimolar concentration of GST-SopE-(78 -240) for 30 min at room temperature. Aliquots from the reaction were blotted onto a nitrocellulose membrane, and the membranes were extensively washed in icecold solution containing 20 mM Tris-HCl (pH 8), 100 mM NaCl, and 10 mM MgCl 2 to remove free radioactivity. The membranes were then transferred to scintillation vials, and counts were determined in the presence of scintillation fluid. A similar exchange reaction of Rab5 was carried out in the presence of different concentrations of SopE as indicated.
Role of Different Constructs of Rab5 in Fusion of Early Endosome with LSP in Vitro-In vitro fusion of phagosomes containing the biotinylated Salmonella with early endosomes containing avidin-HRP were carried out using a procedure similar to that described previously (15). Briefly, phagosomes were purified from macrophages, and endogenous Rab proteins were stripped off from the phagosomes by GDI-GDP treatment. To determine the role of different mutant proteins of Rab5, fusion of Rab-stripped LSP with early endosomes was carried out in the presence of fusion buffer (250 mM sucrose, 0.5 mM EGTA, 20 mM HEPES-KOH, pH 7.2, 1 mM dithiothreitol, 1.5 mM MgCl 2 , 100 mM KCl, including an ATP-regenerating system, 1 mM ATP, 8 mM creatine phosphate, 31 units/ml creatine phosphokinase, and 0.25 mg/ml avidin as scavenger) containing Rab5-immunodepleted, gel-filtered cytosol supplemented with different mutants of Rab5 (10 g of each GST-Rab5 preincubated in the presence of cytosol) as indicated. Fusion was carried out for 10 min at 37°C, and the reaction was stopped by chilling on ice. The HRP-avidin-biotin bacterial complex was recovered by centrifugation (10,000 ϫ g for 5 min) after solubilization of the membrane in solubilization buffer with 0.25 mg/ml avidin as scavenger. The enzymatic activity of avidin-HRP associated with the biotinylated bacteria was measured as fusion unit. The maximum fusion between endosomes and phagosomes (Control, LSP without GDI-GDP treatment) was ob-served at 0.5 mg/ml normal cytosol concentration, which was expressed as 1 unit of relative fusion. The HRP activity, corresponding to 1 unit, is mentioned in the legend to Fig. 6.

RESULTS
SopE Specifically Binds with Rab5-We have previously shown that live Salmonella-containing phagosomes specifically recruit Rab5 on the phagosomes and promote fusion with the early endosomes (15). We therefore investigated the possibility of bacterial proteins that might be involved in the recruitment of different factors from the host cell cytoplasm. The results To determine the molecules from the host cell cytoplasm recognized by the Salmonella, Salmonella were incubated in the presence of biotinylated cytosol prepared from the macrophages as described under "Experimental Procedures." Subsequently, biotinylated protein from the macrophage cytosol bound to Salmonella was detected by Western blot analysis using avidin-HRP (lane 1). To identify the macrophage protein bound to Salmonella, Western blot analysis was carried out using specific antibodies against Rab5 (lane 2), Rab7 (lane 3), and actin (lane 4). Proteins were visualized using HRPconjugated secondary antibody by ECL. Results from the Western blots are representative of three independent preparation. B, detection of Rab5 binding protein from Salmonella. To detect the Rab5 binding protein from the Salmonella, a GST pullout assay was carried out with Rab-GST from a metabolically labeled Salmonella lysate as described under "Experimental Procedures." The Salmonella proteins bound to the beads were analyzed by SDS-PAGE followed by autoradiography (lane 1). To identify the Salmonella protein bound to Rab5, Western blot analysis was carried out using specific antibodies against Salmonella secretory proteins like SopE (lane 2) and SipC (lane 3). Proteins were visualized using HRP-conjugated secondary antibody by ECL. Results from the Western blots are representative of three independent preparations. C, specificity of Rab5 binding with Salmonella. To determine the specificity of Rab5 binding with Salmonella, biotinylated Rab5 was incubated with Salmonella in the absence (Control) or the presence of excess non-biotinylated Rab5 (Rab5) or SopE (SopE) as described under "Experimental Procedures." The biotinylated Rab5 bound to Salmonella in the presence or absence of competitors was detected by Western blot analysis using avidin-HRP by ECL. Results from the Western blots are representative of three independent preparations. presented in Fig. 1A show that Salmonella specifically binds two proteins from the macrophage cytosol of apparent molecular mass of 25 and 42 kDa (lane 1). To identify these proteins, Western blot analyses were carried out with antibodies against different endocytic Rab proteins, which are of about 25 kDa and regulate vesicular trafficking. The anti-Rab5 antibody specifically detected the 25-kDa protein (lane 2) bound to Salmonella, while the anti-Rab7 antibody (lane 3) did not, demonstrating that Salmonella specifically recognized Rab5 from the host cells. Similarly, the 42-kDa protein bound to Salmonella was identified as actin (lane 4), which is consistent with the previous report that SipA, a protein from Salmonella, binds with actin to induce membrane ruffling, which facilitates the entry of the bacteria (21). To search for the bacterial surface protein that interacts with Rab5, GST-Rab5 was incubated with a Salmonella lysate obtained after growing the cells in the presence of [ 35 S]methionine. The data presented in Fig. 1B show that GST-Rab5 specifically picked up two proteins of apparent molecular masses of 30 and 50 kDa from the Salmonella lysate. In contrast, no proteins were detected when GST or GST-Rab7 was used under similar conditions. Western blot analysis with anti-SopE antibody revealed the 30-kDa protein as SopE. However, we have not yet identified the 50-kDa proteins, which also bind to Rab5. Similar results were also obtained using WT S. dublin (data not shown) indicating that SopE, a type III secretory protein of Salmonella, specifically binds with Rab5. To determine the specificity of SopE-mediated binding of Rab5, S. typhimurium was incubated with biotinylated GST-Rab5 in the presence and absence of non-biotinylated Rab5 or SopE. Western blot analysis presented in Fig. 1C shows that binding of biotinylated Rab5 with Salmonella is effectively competed by both Rab5 and SopE.
Live Salmonella Transport SopE onto Phagosomes and Recruit Rab5-To determine whether the SopE produced by Salmonella is transported onto the phagosomes, we purified phagosomes containing respective wild type and mutant Salmonella. The electron micrograph presented in Fig. 2A shows that the purified Salmonella-containing phagosomes are relatively pure without any contamination with other intracellular organelles. Moreover, biochemical characterization demonstrated that these phagosomes are free of endosome, lysosome, Golgi, and endoplasmic reticulum contamination (16). Subsequently, purified live Salmonella-containing phagosomes were probed with anti-SopE antibody followed by a second antibody conjugated with colloidal gold particles to determine the presence of SopE on the phagosomes. The immunolocalization studies presented in Fig. 2B demonstrate that SopE is present on LSP (panel a). In contrast, phagosomes containing dead Salmonella (DSP) do not show the presence of SopE (panel b). The results presented in Fig. 2B show that phagosomes containing live, SopE knockout mutant Salmonella (S. dublin, strain SE1) were unable to recruit Rab5 on LSP (panel d). However, phagosomes containing wild type S. dublin (strain 2229) recruited significant amounts of Rab5 (panel c) on LSP, comparable to that observed with phagosomes containing wild type S. typhimurium (15). The data presented in Fig. 2B (panel e) show that SopE is also localized on LSP harboring wild type S. dublin. It is pertinent to mention that mean numbers of gold particles are significantly higher in panels a, c, and e than in panels b, d, and f (Fig. 2b) as observed from 100 respective phagosomes, indicating that SopE is transported onto LSP and thereby possibly recruit Rab5 on Salmonella-containing phagosomes.
LSP Efficiently Recruits Prenylation-defective Rab5-To understand the mechanism of Rab5 recruitment by LSP, we studied the binding of LSP with various mutant forms of Rab5, namely Rab5:Q79L, a GTPase-defective mutant (22,23); Rab5-S34N, a dominant-negative mutant locked in GDP conformation (9, 24); Rab5:⌬C4, where the isoprenylation motif is deleted (24); and Rab5:WT. Respective Rab proteins were preincubated with macrophage cytosol in the presence of the ATP-regenerating system to allow in vitro prenylation (19). Phagosomes containing either S. typhimurium (WT), or S. dublin (WT) or SopE knockout mutant of S. dublin were treated with Rab-GDI to deplete the endogenous Rabs and incubated in the presence of indicated GST-Rab5 mutant protein in fusion buffer containing cytosol for 10 min at 37°C. Results presented in Fig. 3A show that phagosomes containing WT bacteria bind with Rab5:WT, Rab5:Q79L, as well as Rab5:⌬C4, which are in the GTP form. In contrast, Rab5:S34N, which is locked in the GDP form, did not bind with LSP indicating that only the Rab5   FIG. 3. A, recruitment of Rab5 and its mutant proteins on Salmonella-containing phagosomes. Phagosomes-containing respective Salmonella were stripped of endogenous Rab proteins by GDI-GDP treatment and were incubated in the presence of different mutants of GST-Rab5 fusion protein in the fusion buffer containing cytosol as described under "Experimental Procedures." To determine the binding of indicated GST-Rab5 (50 kDa) with phagosomes, respective phagosomal proteins (40 g of protein each per lane) were analyzed by 12% SDS-PAGE, transferred onto the nitrocellulose membrane, and subsequently probed with specific anti-Rab5 antibody. Proteins were visualized using appropriate HRP-labeled second antibody and ECL. Similarly, Rab7 was used as control. Western blots are representative of three independent preparations. B, interaction of different forms of Rab5 with SopE. To determine the direct interaction of different forms of Rab5 mutant proteins with SopE, the recombinant SopE-(78 -240) (10 g/ml) was coated in an ELISA plate, washed, and incubated with different forms of GST-Rab5:WT or mutant proteins (0.2 mg/ml). Finally, Rab5 binding with SopE was detected using Rab5-specific polyclonal antibody, subsequently probed with secondary antibodies labeled with HRP. The HRP activity present in each well was measured to quantitate binding, and the results are expressed as relative binding of three independent experiments Ϯ S.D.
in the GTP form is recognized by LSP. Significant binding of Rab5:⌬C4, a prenylation-defective Rab5 mutant, suggested that prenylation of Rab5 is not required for the binding of Rab5 with LSP. No significant binding of Rab7:WT with LSP was observed under similar conditions, indicating that LSP specifically binds with Rab5. It is pertinent to mention that all the forms of Rab5 excepting Rab5:⌬C4 were prenylated when preincubated in the presence of cytosol-containing labeled substrate (data not shown). Furthermore, phagosomes containing SopE knockout mutant Salmonella were unable to bind any form of Rab5 under similar conditions, demonstrating that SopE present on LSP is responsible for the recruitment of Rab5 (Fig. 3A).
We further characterized the interaction of Rab5 with SopE in an in vitro assay in which the recombinant SopE immobilized on ELISA plates was incubated with different GST-Rab5 mutant proteins. Binding of Rab5 with SopE was detected using a Rab5-specific primary antibody and HRP-labeled second antibody. Results presented in Fig. 3B show that Rab5: S34N, which is locked in the GDP form, does not bind with SopE. In contrast, SopE specifically binds with Rab5:WT, Rab5:Q79L, and Rab5:⌬C4 demonstrating that SopE can specifically bind only with the GTP form of Rab5. The fact that Rab5:⌬C4, where the prenylation site was deleted, still bound SopE further indicated that prenylation is not required for the binding of Rab5 with SopE.
Binding of SopE with Rab5-The data presented in Fig. 4A show that biotinylated Rab5 binds with immobilized SopE with saturation kinetics. Half-maximal binding of biotinylated Rab5 with SopE occurred at a concentration of about 0.5 g/ml, and maximum binding was observed at 5 g/ml biotinylated Rab5. The binding of biotinylated Rab5 to immobilized SopE was effectively inhibited by both unlabeled Rab5 and SopE (Fig. 4B) with 50% inhibition achieved at about 50 g/ml Rab5 or SopE, indicating the specificity of SopE binding with Rab5.
SopE Acts as a GDP/GTP Nucleotide Exchange Factor of Rab5-Our results (Fig. 3) demonstrated that Salmonella-containing phagosomes specifically bind Rab5 in its GTP form, the active form of the protein, which promotes endosome-endosome fusion. Furthermore, SopE can induce the GDP to GTP exchange of Rho-GTPases (20), which prompted us to investigate the role of SopE in the nucleotide exchange of Rab5. The results presented in Fig. 5A show that incubation of Rab5:WT and Rab5:Q79L in buffer alone significantly induces the nucleotide exchange of GDP to the GTP form over that obtained with Rab5:S34N, a mutant that is unable to exchange GDP to GTP. To determine the role of SopE, we measured the incorporation of [ 32 P]GTP molecules into GDP-loaded Rab5 in the presence of GST-SopE-(78 -240). Our results demonstrated that Rab5:WT and Rab5:Q79L incorporated more than 2-fold of GTP in the presence of GST-SopE-(78 -240) than did Rab5 alone, indicating that SopE enhances the nucleotide exchange of Rab5 (Fig.  5A). Furthermore, SopE was unable to induce the nucleotide exchange of Rab5:S34N, which is locked in the GDP form. Moreover, when Rab5 was incubated with increasing concentrations of SopE in the similar assay, the nucleotide exchange activity of Rab5 is proportional to the concentration of SopE present in the reaction (Fig. 5B).
Prenylation-defective Rab5 Mutant Is Functionally Active-To determine whether non-prenylated Rab5, e.g. Rab5: ⌬C4, recruited on the phagosomes is functionally active, we used an in vitro fusion assay. LSP (containing S. typhimurium WT) were incubated for 10 min with endosomes loaded with avidin-HRP at 37°C in the presence of cytosol and an ATPregenerating system. Results presented in Fig. 6 show that LSP efficiently fuse with early endosomes in 10 min (Control).
To establish the role of different forms of Rab5 in this fusion event, the endogenous Rab5 from the phagosomes were stripped off by Rab-GDI treatment, and fusion was carried out in Rab5-immunodepleted cytosol in the presence of indicated Rab5 mutant proteins. Data presented in Fig. 6 show that fusion of phagosomes with endosomes is inhibited in the Rab5depleted condition. Addition of Rab5:WT and Rab5:Q79L restored the fusion of the phagosomes with endosomes by more than 90%, whereas Rab5:S34N, which is locked in the GDP form did not stimulate the fusion (Fig. 6). Interestingly, Rab5: ⌬C4 stimulated the fusion of phagosomes with endosomes by more than 70% (Fig. 6). Our finding that Rab5:⌬C4, which is identical with Rab5:WT excepting for the deletion of the C terminus cysteine motif that is essential for prenylation, promotes fusion suggests that non-prenylated Rab5 is functionally active. However, when Rab-stripped LSP were pretreated with To determine the binding of Rab5 with SopE, the recombinant SopE-(78 -240) (10 g/ml) was coated on an ELISA plate, washed, and incubated in the presence of the indicated concentrations of biotinylated GST-Rab5:WT as described under "Experimental Procedures." Finally, binding of biotinylated Rab5 with SopE was detected using avidin-HRP. The HRP activity present in each well was measured to quantitate binding, and the results are expressed as the relative binding of three independent experiments Ϯ S.D. B, competition of binding of biotinylated Rab5 with SopE by non-biotinylated Rab5 or SopE. Binding of biotinylated Rab5 (10 g/ml) with immobilized SopE (10 g/ml) was carried out in the presence of indicated concentrations of non-biotinylated Rab5 or SopE as described under "Experimental Procedures." The HRP activity present in the absence of any competitor was taken as 100%, and the results are expressed as a percentage of biotinylated Rab5 bound of three independent experiments Ϯ S.D.
anti-SopE antibody, no fusion of LSP with endosomes could be detected (Fig. 6). We also found that phagosomes containing the SopE knockout mutant Salmonella did not support fusion with the endosomes in Rab-depleted condition (data not shown), indicating that SopE-mediated recruitment of Rab5 by phagosomes is responsible for promoting fusion with early endosomes.

DISCUSSION
Intracellular trafficking of phagosomes depends on vesicular membrane composition as well as intravesicular content and involves dynamic modulations of the phagosomal membrane (6,13) brought about by fusion with other endocytic vesicles and recruitment of various proteins from the cytosol. Recent studies have shown that small GTP binding proteins of the Rab family regulate intercompartmental transport (1,2). Intracellular pathogens modulate the recruitment of these proteins on phagosomes for their survival by avoiding or inducing specific interactions of phagosomes with other vacuolar compartments (12,25). Recently, we have shown that live Salmonella-containing phagosomes (LSP) recruit the early-acting Rab5, and the fusion factors, NSF and ␣-SNAP, to promote fusion with early endosomes (15) thus avoiding transport to the lysosomes so that live Salmonella could persist in a low acidity compartment lacking active lysosomal enzymes (16). In the present study, we sought to delineate how Salmonella-containing phagosomes specifically recruit Rab5 to modulate the maturation of the phagosomes.
Salmonella have evolved a complex protein secretion system termed type III to deliver bacterial effector proteins into host cells, which serve to modulate host cellular function (21,26). Hardt et al. (20) showed that SopE, also a type III secretory protein of Salmonella, stimulates the GDP to GTP nucleotide exchange of several Rho-GTPases, which modulate the cytoskeletal architecture to facilitate entry of Salmonella into epithelial cells. However, uptake of Salmonella, even the noninvasive mutant organism, in macrophages is mediated through lectinophagocytosis (27). We have previously shown that uptake of metabolically labeled live or dead Salmonella by macrophages was essentially the same (16), supporting the previous observation that Salmonella enters into macrophages through a host cell-mediated mechanism such as phagocytosis as opposed to pathogen-induced membrane ruffling. Our previous studies (e.g. Ref. 15) have shown that live Salmonellacontaining phagosomes specifically recruit Rab5 on the phagosomes and promote fusion with the early endosomes. We therefore investigated the possibility of bacterial proteins that might be involved in the recruitment of different factors from the host cell cytoplasm, and our results demonstrated that SopE specifically binds with Rab5 (Fig. 1). In light of this finding, we inferred that SopE from the bacteria should be transported onto the phagosomes where they participate in the recruitment of Rab5 on the phagosomes to promote fusion with the early endosomal compartments. Thus, we looked for the presence of SopE on the surface of the phagosomes containing wild type and mutant Salmonella. The immunoelectron micrograph presented in Fig. 2B demonstrates that SopE is indeed In vitro fusion of Rab-stripped (GDI-GDPtreated) phagosomes containing the biotinylated Salmonella with early endosomes containing avidin-HRP were carried out in the presence of an ATP-regenerating system containing Rab5-immunodepleted cytosol (0.5 mg/ml) supplemented with different mutants of Rab5 (10 g of each GST-Rab 5) as described under "Experimental Procedures." The maximum fusion between endosomes and phagosomes (Control, LSP without GDI-GDP treatment) was observed at 0.5 mg/ml normal cytosol concentration, which was chosen as 1 unit; the results are expressed as the relative fusion of three independent experiments Ϯ S.D. One unit corresponds to 13.7 ng of HRP activity/mg of protein in control fusion. present on LSP containing wild type S. typhimurium or S. dublin (panels a and e), and this facilitates the binding of Rab5 on LSP (panel c). However, live SopE knockout mutant Salmonella-containing phagosomes (S. dublin, strain SE1) lacked SopE on the phagosomes (panel f) and were unable to recruit Rab5 on LSP (Panel d). Taken together, these results demonstrate that SopE is transported onto LSP and thereby helps recruit Rab5 on Salmonella-containing phagosomes.
Membrane association and subsequent biological functions of Rab proteins have been attributed to the C-terminal isoprenylation, which is characteristic of these proteins (28 -30). Isoprenylation of Rab proteins occurs at the C-terminal motifs, which include CC (Rab1, Rab2, Rab9, and Rab10), CXC (Rab3, Rab4, Rab6, Rab7, Rab13, and Rab14), CCXXX (Rab11), CXXX (Rab8, Rab12), and CCXX (Rab5) (X can be any amino acid residue) (31)(32)(33)(34)(35)(36). Deletion of the C-terminal tetrapeptide motif (CCSN) of Rab5 abolishes post-translation isoprenylation, membrane association, and homotypic fusion between endosomes (24). Similar studies (8,9,19,22,24,(37)(38)(39) have shown that deletion of the C-terminal motif of other Ras/Rab proteins results in failure to attach with the target membrane and inhibition of the specific transport process. In contrast, we have demonstrated that Rab5:⌬C4 is recruited on LSP (Fig. 3A), which indicates that binding of Rab5 with LSP is independent of Rab prenylation. However, Rab5 locked in GDP conformation (Rab5:S34N) is unable to bind LSP, indicating that SopE present on LSP specifically recognize Rab5 only in the GTP form. These results are further supported by the fact that immobilized SopE specifically binds Rab5:WT, Rab5:Q79L, and Rab5:⌬C4 but not Rab5:S34N (Fig. 3B). Moreover, binding of biotinylated Rab5 with SopE is competed by unlabeled Rab5 (Fig. 4B). Essentially, results presented in Figs. 3 and 4 demonstrate that SopE acts as the Rab5-specific determinant and mediates the binding of Rab5 in GTP form on LSP. The major functional significance of our observations on SopE-mediated recruitment of non-prenylated Rab5 is that there could be a region of Rab5 outside the prenylation motif, which is specifically recognized by SopE.
Small molecular weight GTP binding proteins of the Rab family regulate vesicular transport. Rab proteins cycle between an active GTP-bound form and an inactive GDP-bound form, the latter being mainly present in the cytosol. Rab-specific guanine nucleotide exchange factor (40) (GEF) catalyzes the conversion of Rab-GDP to Rab-GTP and mediates the particular transport event through other accessory proteins. After the membrane fusion, GTPase-activating protein increases the GTPase rate of Rab and converts them into their GDP-bound state, which is, finally, retrieved by a cytosolic protein termed GDI. GDI delivers the GDP-bound Rab to the membrane and is subsequently reactivated by GEF. Recent studies have shown that SopE acts as a specific GEF on Rho-GTPase proteins such as Cdc42 and Rac to induce membrane ruffling to facilitate Salmonella invasion (20). The results presented in Fig. 5, A and B, clearly demonstrated that SopE also acts as a GEF for Rab5 but not for Rab7. We (15) and others (20,25,26) have shown that SopE produced by the bacteria is transported to the cytosol across the phagosomal membrane, and thus, SopE is transiently present on the phagosomal membrane. Because the live bacteria constitutively produces SopE, this transient phenomenon is extended to a continuous presence of SopE on the phagosomal membrane. Thus, it could be possible that SopE present in the infected cytosol first converts the inactive Rab5 into active GTP-bound conformation, and subsequently, Rab5 in the GTP-bound state is recognized by SopE present on the Salmonella-containing phagosomes. Therefore, the presence of Rab5 in the GTP form on LSP promotes their continuous fusion with early endosomes, thereby inhibiting transport of LSP to the downstream lysosomal compartment, as observed in our previous studies (16).
To determine whether prenylation-defective Rab5 mutant is functionally active, we used in vitro fusion of Rab-stripped LSP with early endosomes in the presence of Rab5-immunodepleted cytosol supplemented with different Rab5 mutant proteins. The results presented in Fig. 6 demonstrate that addition of Rab5:⌬C4 promotes significant fusion between LSP with early endosomes. Prenylation-defective mutants of other Rabs do not mediate the transport between respective compartments. This is due to the fact that prenylation-defective Rab proteins do not bind to the target membrane (8,9,19,22,24,39) and thus fail to trigger downstream events in vesicle fusion. Rab5:S34N, which is locked in the GDP form, does not promote the fusion between LSP and early endosomes, which is consistent with previous demonstrations that this mutant protein does not support homotypic fusion between early endosomes (9,23,24,28). Furthermore, treatment of Rab-stripped LSP with anti-SopE antibody inhibits the Rab5-mediated fusion between LSP and early endosomes (Fig. 6), indicating that SopE-mediated recruitment of Rab5 on LSP promotes the fusion. Therefore, our data indicate that prenylation-defective Rab5 protein is functionally active when it is recruited on LSP through SopE.
Thus, it appears that prenylation of Rab proteins in general is only required for their attachment with the membrane.
In conclusion, our results demonstrate that SopE acts as a nucleotide exchange factor for Rab5 and mediates the specific recruitment of Rab5 in the GTP form on LSP, irrespective of prenylation, thus promoting fusion of LSP with early endosomes. In contrast to the current concept of Rab function, that prenylation of Rab protein is required for membrane attachment and biological function, this is the first demonstration that a non-prenylated Rab protein can sustain its biological activity of promoting fusion when recruited on the target membrane. Thus, these results indicate that prenylation of Rab proteins is not essential for their biological function; it is simply required for membrane attachment. The physiological significance of this finding derives from the fact that SopE acts as an Rab5-specific exchange factor and thereby mediates the recruitment of Rab5 in the GTP form on phagosomes containing live Salmonella. This constitutes a salvage mechanism that ensures the sustained fusion of LSP with early endosomes, independent of Rab5 prenylation, thereby inhibiting targeting of live Salmonella to the lysosomes and their eventual destruction.