Aeromonas salmonicida Toxin AexT Has a Rho Family GTPase-Activating Protein Domain

Abstract

The N terminus of the Aeromonas salmonicida ADP-ribosylating toxin AexT displays in vitro GTPase-activating protein (GAP) activity for Rac1, CDC42, and RhoA. HeLa cells transfected with the AexT N terminus exhibit rounding and actin disordering. We propose that the Aeromonas salmonicida AexT toxin is a novel member of the growing family of bacterial RhoGAPs.

Aeromonas salmonicida is a fish pathogen, the causative agent of furunculosis, a fatal disease of salmonids (4). Exoenzyme T (AexT) is a 50-kDa toxin produced by A. salmonicida and translocated via the type III secretion pathway (2, 3). Recombinant AexT protein was reported to display ADP-ribosyltransferase enzymatic activity (2).

We have found that the N terminus of the AexT gene product is significantly similar to the Rho GTPase-activating protein (RhoGAP) domains of Salmonella enterica serovar Typhimurium SptP, Yersinia pestis YopE, and Pseudomonas aeruginosa ExoS and ExoT, showing 30%, 44%, 45%, and 36% amino acid identity with these gene products, respectively (11).

We have expressed and purified the N-terminal domain of the AexT gene product and tested its in vitro GAP activity for recombinant Rac1, RhoA, CDC42, and Ras. The GAP domain of AexT (AexTGAP), amino acids 93 to 255, was amplified using plasmid pJFFASAexT (2) as a template and primers ACATACATGCATGCCTGGTTCCGCGTGGCAGCGCGGTTTCCCCCGAGGATCTTCAGCG and GAGAGCCAAGCTTGTCGGCAGGCTCAACGTTAACTTC in a standard PCR (TaKaRa). The product was ligated into plasmid pQE-80L (QIAGEN). The final nucleotide sequence was verified using DNA sequencing analysis, as were all plasmids used in this work. Escherichia coli strain BL21(DE3) transformed with pQE-80L-AexTGAP was grown at 37°C in a Luria-Bertani (LB) medium containing 100 μg/ml ampicillin. When the optical density at 600 nm had reached 0.6, 50 μM IPTG (isopropyl-β-d-1-thiogalactopyranoside; Sigma I-5502) was added. Cells were harvested 2 h after IPTG induction. The bacterial pellet was resuspended in lysis buffer (25 mM Tris, pH 7.9, 6 mM beta-mercaptoethanol, 7.5 mM imidazole, 0.25 M NaCl, 1 mM phenylmethylsulfonyl fluoride [PMSF], 0.1 mg/ml DNase I, 0.2 mg/ml lysozyme, 0.02% NaN₃), passed through a Microfluidizer Processor M-110EHI (Microfluidics Corporation), and centrifuged. The supernatant was applied to a Ni-nitrilotriacetic acid agarose column (QIAGEN). The column was washed with wash buffer (25 mM Tris, pH 7.9, 6 mM beta-mercaptoethanol, 0.5 M NaCl, 20 mM imidazole, 0.02% NaN₃). AexTGAP protein was eluted with wash buffer supplemented with 80 mM imidazole. Eluent was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Highly purified AexTGAP was collected and dialyzed against dialysis buffer (25 mM Tris, pH 7.5, 1 mM dithiothreitol).

Human Rac1, RhoA, and CDC42 genes (templates pEGFP-Rac1, pEGFP-RhoA, and pEGFP-CDC42, respectively) and the Salmonella SptP gene (template pGEX-SptP) [12], amino acids 167 to 255 [SptPGAP]) were amplified and introduced into the pQE-80L vector as described for AexTGAP. The primers used were CTGAACGCGGATCCCTGGTTCCGCGTGGCAGCCAGGCCATCAAGTGTGTGGTGGTGG and CTTATCCCAAGCTTTTACAACAGCAGGCATTTTCTCTTCC for Rac1, CTGAACGCGGATCCCTGGTTCCGCGTGGCAGCGCTGCCATCCGGAAGAAACTGGTG and CTGAACGCAAGCTTTCACAAGACAAGGCACCCAGATTTTTTC for RhoA, CTGAACGCGGATCCCTGGTTCCGCGTGGCAGCGCTGCCATCCGGAAGAAACTGGTG and CTGAACGCAAGCTTTCACAAGACAAGGCACCCAGATTTTTTC for CDC42, and ACATACATGCATGCCTGGTTCCGCGTGGCAGCAGTAAGCAACCTTTACTCGATATCGCG and GAGAGCCAAGCTTTATTTTCTCAAGTTCAGCTGTAACATTTTTC for SptPGAP.

Rac1, CDC42, and SptPGAP were expressed in E. coli strain BL21(DE3) as described for AexTGAP. RhoA was expressed in E. coli strain Rosetta(DE3). Purifications of Rac1, RhoA, CDC42, and SptP were performed as described for AexTGAP. All buffers (lysis, wash, and dialysis buffers) used for purification of Rac1, RhoA, and CDC42 were supplemented with 5 mM MgCl₂.

A C′-truncated human Ras gene product (amino acids 1 to 166) was cloned into pDEST14 vector (Invitrogen) using the Gateway cloning technology (Invitrogen), as indicated by the manufacturer. E. coli strain BL21 SI transformed with the pDEST14-Ras vector was grown at 30°C in LB medium lacking NaCl with 100 μg/ml ampicillin. When the optical density at 600 nm had reached 0.6, NaCl was added to give an 0.3 M concentration. Cells were harvested after an overnight incubation, resuspended in Ras lysis buffer (25 mM Tris, pH 7.9, 0.5 mM 1,4-dithioerythritol, 10 MgCl₂, 0.02% NaN₃, 1 mM PMSF, 0.1 mg/ml DNase I, 0.2 mg/ml lysozyme), passed through a Microfluidizer Processor M-110EHI (Microfluidics Corporation), and centrifuged. Supernatant was applied to a Q-Sepharose FF (Pharmacia) column. The column was washed with an increasing gradient of 0.05 to 0.2 M NaCl in buffer A (25 mM Tris, pH 7.9, 0.5 mM 1,4-dithioerythritol, 10 mM MgCl₂, 1 mM PMSF, 0.02% NaN₃). The eluent was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Purified Ras protein was collected, supplemented with 1 μM GDP, and applied to a Sephacryl S100 (Pharmacia) column. Highly purified Ras was eluted with buffer B (25 mM Tris, pH 7.9, 0.1 M NaCl, 5 mM MgCl₂, 1 μM GDP, 1 mM PMSF, 0.02% NaN₃).

A single-turnover GTPase assay was employed to determine the rate of GTP hydrolysis in the absence or presence of AexTGAP. AexTGAP has significantly accelerated the rate of GTP hydrolysis by Rac1, RhoA, and CDC42 GTPases but not of that by Ras (Fig. 1). The YopE, ExoS, and ExoT proteins have previously been reported to show GAP activity for Rho, Rac, and CDC42 (8, 9, 13). Surprisingly, the Salmonella protein SptP was reported to stimulate GTPase activity of Rac and CDC42 but not of Rho (5). We tested the recombinant SptPGAP RhoGAP activity using the same assay that we have used for AexTGAP. Indeed, we confirmed that SptPGAP activity is specific for Rac1 and CDC42 (data not shown).

FIG. 1.

In vitro GAP activity of AexTGAP. The rates of [γ-³²P]GTP hydrolysis by Rac1 (A), CDC42 (B), RhoA (C), and Ras (D) were determined using a single-turnover GTPase assay, in the presence or absence of AexTGAP or AexTGAP R143A. The GTPase (3 μM) was incubated with a 10-fold excess of [γ-³²P]GTP (NEN) in 25 mM Tris, pH 7.5, 1 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, and 5 mM EDTA for 5 (Rac1 and CDC42) or 10 (RhoA and Ras) minutes at 25°C, to enable the binding of radiolabeled GTP. Hydrolysis was initiated upon addition of 10 mM MgCl₂ and 0.3 mM unlabeled GTP at 25°C. AexTGAP was added to a concentration of 10 nM 1 min after initiation of hydrolysis. At each indicated time point, samples were transferred to 5% acid-washed charcoal in 20 mM phosphoric acid at 4°C. The charcoal was centrifuged, and inorganic phosphate in the supernatant was counted in a Beckman LS 2800 Liquid Scintillation Counter.

Arginine 143 of AexT is equivalent to the arginine finger motif in other bacterial GAPs (Fig. 2) (5, 8, 13). We mutated this arginine to alanine, using the QuikChange site-directed mutagenesis kit (Stratagene) and primers GGCAATGGCCCCCTGGCCTCTCTCTGCACCGCG and CGCGGTGCAGAGAGAGGCCAGGGGGCCATTGCC, following the manufacturer's instructions. Expression and purification of the mutant AexTGAP R143A followed the same procedure as for AexTGAP. AexTGAP R143A was unable to stimulate GTPase activity of Rac1, RhoA, or CDC42 in the in vitro GTPase assay.

FIG. 2.

Arginine finger motif of RhoGAPs. The arginine finger motifs of mammalian p50RhoGAP, Salmonella enterica serovar Typhimurium SptP, Pseudomonas aeruginosa ExoS and ExoT, Yersinia pestis YopE, and Aeromonas salmonicida AexT are shown. A conserved glycine and the catalytic arginine residues are shaded.

The AexT gene has been shown to cause cell rounding and cytotoxicity in RTG-2 (rainbow trout gonad) cells infected with A. salmonicida (2). SptP, ExoS, ExoT, and YopE RhoGAPs have been reported to disrupt the eukaryotic actin cytoskeleton and cause cell rounding (6, 9, 10, 13). To investigate the function of the AexT protein in a cellular context, we constructed a mammalian expression vector containing a green fluorescent protein-fused AexTGAP gene. AexTGAP and AexTGAP R143A sequences were amplified using primers CTGAACGCAAGCTTGCCATGGCCGTTTCCCCCGAGGATCTTC and CTGAACGCGGATCCGCGTCGGCAGGCTCAACGTTAAC in a standard PCR (TaKaRa) and ligated into plasmid pEGFP-N1 (Clontech). Mammalian HeLa cells were transiently transfected with the pEGFP-N1-AexTGAP vector. GFP-AexTGAP-expressing cells appeared rounded, with a disordered actin structure (Fig. 3). HeLa cells expressing the pEGFP-N1 control vector appeared with a typical morphology, with a normal, well-defined actin cytoskeleton. HeLa cells that were transiently transfected with the mutant pEGFP-AexTGAP R143A did not round up. Our results suggest that arginine 143 of AexT is indeed a functional arginine finger and that AexTGAP acts by a mechanism similar to that of the other bacterial GAPs.

FIG. 3.

AexTGAP causes cell rounding and disorder of the actin cytoskeleton in transfected HeLa cells. Transient transfections of pEGFP-N1 (A and B), pEGFP-N1-AexTGAP (C and D), and pEGFP-N1-AexTGAP R143A (E and F) into HeLa cells using calcium phosphate were performed according to a standard protocol (11a). Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline 24 h posttransfection and permeabilized with 0.025% saponin in phosphate-buffered saline. Actin microfilaments were visualized with rhodamine-conjugated phalloidin (Sigma P-1951) staining (B, D, and F). Images were obtained with a confocal laser scanning microscope (MRC 1024; Bio-Rad). GFP, green fluorescent protein.

Bacterial GAPs are a remarkable family of proteins. Though sharing limited sequence similarity, they all exhibit strikingly similar effects on eukaryotic cells. They target Rho-family GTPases to subvert the eukaryotic cytoskeleton to their own advantage. All bacteria known to encode a GAP utilize the type III secretion system to translocate these toxins into the host cytosol (1, 7). We have shown here that the N-terminal domain of the A. salmonicida AexT toxin is a novel member of this family of proteins. Increasing genomic data may reveal additional prokaryotic GAPs and extend our knowledge of bacterial virulence.