The Yersinia pseudotuberculosis Cytotoxic Necrotizing Factor (CNF Y ) Selectively Activates RhoA*

The cytotoxic necrotizing factors (CNF)1 and CNF2 from pathogenic Escherichia coli strains activate RhoA, Rac1, and Cdc42 by deamidation of Gln 63 (RhoA) or Gln 61 (Rac and Cdc42). Recently, a novel cytotoxic necrotizing factor termed CNF Y was identified in Yersinia pseudotuberculosis strains (Lockman, H. A., Gillespie, R. A., Baker, B. D., and Shakhnovich, E. (2002) Infect. Immun. 70, 2708–2714). We amplified the cnfy gene from genomic DNA of Y. pseudotuberculosis , cloned and expressed the recombinant protein, and studied its activity. Recombinant GST-CNF Y induced morphological changes in HeLa cells and caused an upward shift of RhoA in SDS-PAGE, as is known for GST-CNF1 and GST-CNF2. Mass spectrometric analysis of GST-CNF Y -treated RhoA confirmed deamidation at Glu 63 . Treatment of RhoA, Rac1, and Cdc42 with GST-CNF Y decreased their GTPase activities, indicating that all of these Rho proteins could serve as substrates for GST-CNF Y in vitro . In contrast, RhoA, but not Rac or Cdc42, was the substrate of GST-CNF Pull-down Experiments— The Rho-binding region, encoding the N- terminal 90 amino acids of rhotekin (rhotekin pull-down), the CRIB-domain (amino acids 56–272) of p21-activated kinase (PAK pull-down) or the Rac-binding domain of POSH (amino acids 282–371, POSH pull-down) were expressed as GST-fusion proteins in E. coli BL21. Overnight cultures diluted 1:10 (for rhotekin 1:100) and grown for Then, 0.1 m M isopropyl- (cid:1) - D -thiogalactopyranoside (final was added. 2 h after induction, cells were and lysed by sonication in rhotekin lysis buffer (20% m M m M m M MgCl and purified by affinity chromatography with glutathione-Sepharose (Am- ersham Pharmacia Biotech). Loaded beads were washed three times with rhotekin lysis buffer and once with buffer A. Toxin-treated cells were lysed in buffer A. For analysis of the total amount of the respective GTPase in the lysates, 1/10 of the volume was taken for Western blot analysis. 9/10 of the volume was incubated with protein-loaded beads for 1 h at 4 °C by head-over-head rotation. After incubation, beads were washed once with buffer A. Then SDS-sample buffer was added and samples were boiled and separated on SDS- PAGE. RhoA, Rac1, and Cdc42 were analyzed by immunoblotting with their specific antibodies. GTPases incubated GST-CNF GST-CNF1 a molar GTPases:toxin 20:1

The cytotoxic necrotizing factors (CNF) 1 -1 and CNF2 from pathogenic Escherichia coli strains belong to the group of pro-tein toxins that constitutively activate small GTPases of the Rho family. CNF1 and CNF2 catalyze the deamidation of RhoA at Gln 63 (Gln 61 of Rac1 and Cdc42, respectively), thereby inhibiting their intrinsic and GAP-stimulated GTPase activities (1,2). In CNF1-treated HeLa cells, activation of Rac is most prominent, and Cdc42 is only slightly activated. Persistent activation of the Rho GTPases leads to typical morphological changes caused by rearrangements of the cellular actin cytoskeleton, e.g. RhoA induces the formation of stress fibers, Rac1 leads to membrane ruffling, and Cdc42 induces the formation of filopodia (for review, see Ref. 3). Besides these direct effects on actin structures, the deregulation of Rho GTPases by CNF results in the formation of enlarged multinucleated cells, most likely by inhibition of cytokinesis (for review, see Ref. 4).
Recently, we reported that the cellular content of Rac1 decreases after CNF treatment, a process that is blocked by proteasome inhibitors (5,6). Doye et al. (6) showed that the CNF-activated form of Rac1 is ubiquitylated and subsequently degraded by a proteasome-dependent mechanism. Proteasomal degradation of CNF-activated Rac1 increases the motility of target cells and may be involved in the destruction of the barrier function of epithelial host cells by CNF-producing bacteria (6).
More recently, a novel cytotoxic necrotizing factor, CNF Y , has been identified in Yersinia pseudotuberculosis (7). The cnfy gene is encoded by an open reading frame of 3,045 bp and is located on the chromosomal DNA and not on the virulence plasmid, which codes for the Yop proteins (8). The predicted amino acid sequence of CNF Y is about 61% identical to that of CNF1 from E. coli. The amino acid residues essential for catalytic activity in CNF1, Cys 866 and His 881 , are also conserved in CNF Y with the same spacing (7,9).
Lockman et al. (7) reported that broth supernatants and cell-free lysates of Y. pseudotuberculosis strains containing the complete cnfy gene induced the development of Hep-2 cells into giant, multinucleated cells. The morphology of these cells was similar to the CNF1-induced phenotype. Interestingly, the biological activity was independent of the presence of the 70-kDa virulence plasmid of Yersinia, which encodes the components of a type-III secretion system (8). Strains which produced CNF Y but were cured of their virulence plasmid still had the ability to induce the multinucleation of cells. Therefore, the activity of CNF Y seems to be independent of the type III secretion system of Yersinia. Moreover, CNF Y is released from the bacteria, probably by a different secretion mechanism, because the bacterial culture supernatant contains the biological activity. In contrast, it is believed that CNF1 is not released from E. coli but stays cell-associated, based on the fact that the biological activity is only found in the presence of the bacteria (7, 10). Despite the significant homology between CNF Y and CNF1, a CNF1 polyclonal antibody was not able to inhibit the CNF Y cytopathic effect (7).
Here, we cloned and expressed CNF Y from Y. pseudotuberculosis as a recombinant GST-fusion protein (GST-CNF Y ϭ 146 kDa; GST ϭ 26 kDa) and investigated its effects on the Rho GTPases RhoA, Rac1, and Cdc42, both in vitro and in intact cells. From our studies, we conclude that in intact cells, GST-CNF Y is a selective activator of RhoA.

EXPERIMENTAL PROCEDURES
Cell Culture and Preparation of Lysates-HeLa cells were grown in Dulbecco's modified Eagle's medium (12 mM L-glutamine) supplemented with 10% fetal calf serum, penicillin (4 mM), and streptomycin (4 mM) in a humidified atmosphere of 5% CO 2 at 37°C.
For intoxication with GST-CNF1 or GST-CNF Y , cells were treated with 400 ng of GST-CNF per ml of medium 1 day after seeding. For preparation of cell lysates, cells were washed twice with ice-cold phosphate buffered saline, lysed in buffer A (10% glycerol, 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1% Igepal, 2 mM MgCl 2 , and 0.5 mM phenylmethylsulfonyl fluoride) for 5 min at 4°C and harvested with a rubber policeman. Lysates were cleared by centrifugation (20 min, 20,800 ϫ g, 4°C), and supernatants were subjected to pull-down assays.
Primary cultures of neurons were prepared from hippocampi of newborn Wistar rats, as described previously (11). The dissociated neural cells were seeded on poly-D-lysine/laminin-coated coverslips. Incubation medium consisted of neurobasal medium supplemented with B27. Fresh medium was mixed 1:1 with medium, which had been conditioned for 3 or 4 days with hippocampal astroglial cultures. Cultures were incubated for 3 days at 37°C in a humidified atmosphere containing 8.6% CO 2 . Neurobasal medium and B27 were obtained from Invitrogen.
Actin-staining-Formaldehyde-fixed cells were washed three times with phosphate-buffered saline. The cells were then incubated with rhodamine-conjugated phalloidin (1 unit per coverslip) at room temperature for 1 h, washed again, and applied for fluorescence microscopy (bleaching preservative was Kaiser's glycerol gelatin from Merck). Micrographs were taken with an Axiocam camera (Zeiss).
Immunocytochemistry-Neurons were fixed with 4% paraformaldehyde at room temperature for 20 min. They were washed with phosphate-buffered saline and permeabilized with 0.1% (v/v) Triton X-100. Normal goat serum was used to block unspecific reactions. Thereafter, the neurons were incubated with the monoclonal mouse anti-␤-tubulin III antibody (Sigma). The resulting immune complexes were visualized with a Cy ™ 3-conjugated F(abЈ) 2 fragment of goat anti-mouse IgG (Dianova, Hamburg, Germany).
Confocal Image Analysis-Neurons were imaged by using an MRC 1024 confocal system (version 3.2, Bio-Rad) with a krypton-argon laser and an Axiovert 135TV microscope (Zeiss, Oberkochen, Germany). Fluorescence of Cy ™ 3 was measured by using an excitation wavelength of 554 nm and an emission filter set at 576 nm. A Z-stock of 95 confocal images were obtained, and a vertical projection was produced by using Metamorph software from Vistron, Mü nchen, Germany.
Mutagenesis-The cnfy gene was amplified by PCR from genomic DNA of Y. pseudotuberculosis strain YPIII using 5Ј-GGATCCAT-GAAAAATCAATGGCAACATCAATATTT-3Ј as sense primer and 5Ј-CCCGGGGATATCTTTTCATTTCCCCCTGCC-3Ј as antisense primer, thereby introducing restriction sites for BamHI and SmaI. The resulting PCR product was cloned into the expression vector pGEX-2TGL and transformed into E. coli BL-21 for protein expression. Mutagenesis was performed by the QuikChange site-directed mutagenesis kit (Stratagene) based on the pGEX-2TGL-CNF Y construct using the following sense primers (and corresponding antisense primers): CNF Y (C866S), 5Ј-TTAACAGGGTCTACAGTTGTT-3Ј, CNF Y (H881A), 5Ј-TATGCTG-TAGCTACAGGAAATTCT-3Ј. All constructs were checked by DNA sequencing using a Dye Terminator Cycle sequencing kit with AmpliTaq DNA polymerase (Applied Biosystems).
Protein Preparation-For toxin purification, BL21 E. coli strains, carrying pGEX-CNF1, pGEX-CNF Y , and the mutated plasmids, respectively, were grown in LB medium. At A 0.6 , cells were collected by centrifugation and lysed by sonication in lysis buffer (20 mM Tris-HCl, pH 7.4, 10 mM NaCl, 1% Triton, 1 mM phenylmethylsulfonylfluoride, and 5 mM dithiothreitol) and purified by affinity chromatography with glutathione-Sepharose (Amersham Pharmacia Biotech, Freiburg, Germany). Loaded beads were washed five times with lysis buffer (without phenylmethylsulfonyl fluoride) at 4°C. The GST-CNF fusion proteins were eluted from the beads by glutathione (10 mM glutathione and 50 mM Tris-HCl, pH 7.4) twice for 10 min at room temperature.
Toxin-treated cells were lysed in buffer A. For analysis of the total amount of the respective GTPase in the lysates, 1/10 of the volume was taken for Western blot analysis. 9/10 of the volume was incubated with protein-loaded beads for 1 h at 4°C by head-over-head rotation. After incubation, beads were washed once with buffer A. Then SDS-sample buffer was added and samples were boiled and separated on SDS-PAGE. RhoA, Rac1, and Cdc42 were analyzed by immunoblotting with their specific antibodies.
Treatment of Recombinant GTPases with GST-CNF Y and GST-CNF1-Rho GTPases were incubated with GST-CNF Y or GST-CNF1 at a molar ratio of GTPases:toxin of 20:1 for 3 h at 37°C in a buffer containing 50 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 1 mM dithiothreitol, and 1 mM EDTA.
GTPase Assay-Recombinant Rho proteins were modified by GST-CNF. After incubation with the respective toxin, the proteins were loaded with [␥-32 P]GTP for 5 min at 37°C in loading buffer (50 mM Tris-HCl, pH 7.5, 10 mM EDTA, 2 mM dithiothreitol). MgCl 2 (12 mM, final concentration) and unlabeled GTP (2 mM, final concentration) were added. Loaded GTPases were incubated at 37°C for the indicated times with or without p50RhoGAP, and the GTPase activity was analyzed by a filter-binding assay.
Proteolytic Digestion in the Gel Matrix for Mass Spectrometric Analysis-The excised gel plugs of RhoA were destained for 1 h at 50°C in 40% acetonitrile/60% hydrogen carbonate (50 mM, pH 7.8) to remove Coomassie Blue, gel buffer, SDS, and salts. The plug was subsequently dried in a vacuum centrifuge for 15 min. Thereafter, 30 l of digestion buffer with trypsin was added, and digestion was carried out for 12 h at 37°C.
Sample Preparation for MALDI-TOF Mass Spectrometry-4-Hydroxy-␣-cyanocinnamic acid (Aldrich) was recrystallized from hot methanol and stored in the dark. Saturated matrix solution of 4-hydroxy-␣cyanocinnamic acid in a 1:1 solution of acetonitrile/aqueous 0.1% trifluoroacetic acid was prepared. 2 l of the proteolytic peptide mixture was mixed with 2 l of saturated matrix containing marker peptides (5 pmol of human ACTH (18 -39), clip (MW 2466, Sigma), and 5 pmol of human angiotensin II (MW 1047, Sigma)) for internal calibration. Using the dried-drop method of matrix crystallization, 1 l of the sample matrix solution was placed on the MALDI-TOF stainless-steel target and was allowed to air-dry several minutes at room temperature, resulting in a thin layer of fine granular matrix crystals.
Mass Spectrometry-MALDI-TOF mass spectrometry was performed on a Bruker Biflex mass spectrometer equipped with a nitrogen laser (l ϭ 337 nm) to desorb and ionize the samples. Mass spectra were recorded in the reflector positive mode in combination with delayed extraction.
Western Blot Analysis-For Western blotting, samples were subjected to SDS-PAGE and transferred onto polyvinylidene difluoridemembrane. Rac was detected with a specific antibody (anti-Rac, BD Biosciences). Rho was detected with a monoclonal antibody directed against the insert region (anti-RhoA, Santa Cruz Biotechnology, Santa Cruz, CA) and for Cdc42 detection, a monoclonal antibody (anti-Cdc42, Upstate Biotechnology, Lake Placid, NY) was used. Binding of the second horseradish peroxidase-coupled antibody was detected with enhanced chemiluminescent detection reagent (100 mM Tris-HCl, pH 8.0, 1 mM luminol (Fluka, St. Gallen, Switzerland), 0.2 mM p-coumaric acid, 3 mM H 2 O 2 ).

RESULTS
Recently, the nucleotide sequence of CNF Y from Y. pseudotuberculosis was described (7). It was shown that bacterial cell extracts of Yersinia induce multinucleation in Hep-2 cells as described for CNF1 produced by pathogenic E. coli strains.
Cloning and Expression of CNF Y -To analyze the activity of CNF Y , the genomic DNA of Y. pseudotuberculosis was purified. We amplified the cnfy gene by PCR and cloned it into the pGEX vector. The proper construct was checked by DNA sequencing. Sequencing revealed some differences in the N-terminal part of the protein (S193N, G305Q, and T395A), as compared with the sequence published by Lockman (7). Repeated PCR gave the same results, indicating that the differing sequence is not due to mistakes of the polymerase. CNF Y was then expressed as GST-fusion protein in BL21 cells and purified by affinity chromatography (Fig. 1A, lane 1).
Activity of Recombinant GST-CNF Y -The activity of the recombinant toxin was first studied by the change in electrophoretic mobility of modified RhoA. As shown in Fig. 1B, incubation of recombinant RhoA with GST-CNF1 (Fig. 1B, lane 2) or with GST-CNF Y (Fig. 1B, lane 3) but not with buffer (Fig.  1B, lane 1) caused an upwards shift of the GTPase in SDS-PAGE, indicating deamidation of RhoA by GST-CNF1 and GST-CNF Y . After incubation with GST-CNF Y for 3 h, only ϳ30% of the GTPase was shifted, whereas GST-CNF1 catalyzed modification of 100% of RhoA (Fig. 1B), indicating higher activity of the recombinant E. coli toxin in vitro. For CNF1, a catalytic dyad consisting of a cysteine and a histidine residue has been identified (9). These residues are identical in CNF Y . We mutated the conserved catalytic residues Cys 866 and His 881 in GST-CNF Y to serine or alanine, respectively, expressed the mutants (Fig. 1A, lanes 2 and 3), and studied their activities. Incubation of recombinant RhoA with these mutants does not change the electrophoretic mobility of the GTPase (Fig. 1B,  lanes 4 and 5). This indicates that CNF Y shares the same catalytic residues with CNF1. Deamidation of RhoA can be detected by MALDI-TOF mass spectrometry (1). We used this method to verify the GST-CNF Y -catalyzed deamidation of Glu 63 of RhoA. The recombinant GTPase was incubated with the toxin and subjected to SDS-PAGE. In-gel digestion with trypsin was performed, and the tryptic peptides were analyzed by MALDI-TOF mass spectrometry. As shown in Fig. 2, incubation of RhoA with GST-CNF Y (Fig. 2, A and B, control) led to an increase in mass of one Dalton in a peptide corresponding to amino acids 52-68 of the GTPase. This increase in mass corresponds to deamidation of Glu 63 . As shown for GST-CNF1catalyzed deamidation of Rho GTPases, this modification leads to the block of intrinsic and GAP-stimulated GTPase activity.
In Vitro Substrate Specificity of Recombinant GST-CNF Y -Measuring of GTPase activity can be used to study substrate specificity of the toxin. Therefore, recombinant RhoA, Rac1, and Cdc42 were incubated with GST-CNF Y , GST-CNF1, or buffer, respectively, as indicated and loaded with [␥-32 P]GTP. The intrinsic or GAP-stimulated GTPase activity of the Rho proteins was then measured in a filter-binding assay. As shown in Fig. 3A, incubation of RhoA, Rac1, or Cdc42 with one of the two CNFs studied leads to the block of intrinsic GTP hydrolysis of the respective GTPase. The same effect was found for the GAP-stimulated GTPase activity (Fig. 3B). GST-CNF1 caused a stronger inhibition of the GTPase activity than GST-CNF Y under the conditions used. The finding that after this time period, still-unmodified GTPase was detected in the sample (Fig. 1, B and C) corresponds with the lower extent of inhibition of GTP hydrolysis by GST-CNF Y as compared with GST-CNF1. Nevertheless, the data show that in vitro, all three GTPases (RhoA, Rac, and Cdc42) can serve as substrates for CNF1 as well as for CNF Y .
GST-CNF Y Induces RhoA Morphology in HeLa Cells-To study the activity of the toxin in the living cell, HeLa cells were incubated with GST-CNF1, GST-CNF Y , and the inactive mutant GST-CNF Y (C866S) respectively, and morphological changes induced by the toxins were monitored by phase contrast microscopy. Moreover, the actin cytoskeleton of fixed cells was stained with rhodamine-phalloidin and analyzed by fluorescence microscopy. As expected, GST-CNF Y induced the rearrangement of the actin cytoskeleton of HeLa cells and led to polynucleation. Notably, the morphology of cells and toxininduced actin rearrangements observed with GST-CNF1 and GST-CNF Y , respectively, were different. As shown in Fig. 4, incubation of HeLa cells with GST-CNF1 for 2 h led to cell spreading (Fig. 4C). Also, the formation of membrane ruffles and, to a lesser extent, the formation of actin stress fibers and filopodia were evident already after 2 h of incubation with GST-CNF1, indicating strong activation of Rac and activation of RhoA and Cdc42. In contrast, incubation with GST-CNF Y for the same time period led to the formation of actin stress fibers, but no membrane ruffling, filopodia formation, or cell spreading were detected (Fig. 4D), indicating activation of RhoA but not of Rac and Cdc42 by CNF Y . However, after 24 h of incubation with GST-CNF Y , cells were flattened and showed mem- brane ruffling (Fig. 4E). As expected, the catalytically inactive mutant GST-CNF Y (C866S) had no effect on the morphology of HeLa cells (Fig. 4A). From our in vitro data presented above, the slower flattening could be explained by the minor catalytic activity of GST-CNF Y as compared with GST-CNF1. However, an even stronger stress fiber formation was observed after 2 h of GST-CNF Y treatment than was observed after GST-CNF1 treatment. This finding indicates that GST-CNF Y has even higher activity toward RhoA than GST-CNF1 in the living cell.
GST-CNF Y Selectively Activates RhoA in HeLa Cells-To clarify the in vivo substrate specificity of GST-CNF Y , we performed pull-down experiments with bead-coupled GTPasebinding domains of rhotekin, PAK, or POSH. These effector domains exclusively bind to the GTP-bound form of the GT-Pases. HeLa cells were treated with the two toxins in a time course and then lysed. The pull-down experiments were performed with increasing incubation times to follow activation of the respective GTPase. In addition, toxin-treated HeLa cells were fixed and stained with rhodamine-phalloidin after the same time periods to allow comparison of GTPase activation and morphology. As shown in Fig. 5, incubation of the cells for 2 h with GST-CNF1 led to strong activation of Rac and activation of RhoA and Cdc42 with the corresponding formation of membrane ruffles, stress fibers, and filopodia in HeLa cells (Fig. 5A, lane 1). With GST-CNF Y , a strong activation of RhoA was evident already after 1 h of incubation and was still present after 24 h of toxin treatment. In contrast, no activation of Rac and Cdc42 could be detected during the first 6 h of treatment with GST-CNF Y . In line with this, GST-CNF Y -treated HeLa cells showed a strong formation of stress fibers but no membrane ruffling or filopodia formation (Fig. 5, A and B, lanes  3-6). After 8 -24 h of treatment with GST-CNF Y , a slight activation of Rac occurred, which was accompanied by the formation of membrane ruffles (Fig. 5, A and B, lanes 7 and 8). Thus, activation of the GTPases detected in the pull-down experiments clearly depicts the morphological changes induced by GST-CNF Y in HeLa cells. All together, the data indicate that GST-CNF1 and GST-CNF Y have different substrate specificity in vivo, with GST-CNF Y specifically modifying RhoA.
GST-CNF Y Induces RhoA Morphology in Rat Hippocampal Neurons-To further substantiate our observation that CNFy selectively activates RhoA, we studied the morphology of a different type of cell and tested the effect of GST-CNF Y on rat hippocampal neurons. In these cells, it is possible to distinguish morphological changes induced by active RhoA from changes induced by active Rac and Cdc42 activation. It has been shown that RhoA activation in hippocampal neurons leads to the retraction of dendrites, whereas constitutive activation of Rac or Cdc42 induces the formation of small fingerlike extensions along the dendrites (12). We treated rat hippocampal neurons with GST-CNF1 or GST-CNF Y and studied the morphology induced. As shown in Fig. 6, incubation of the neurons for 4 h with GST-CNF1 or with GST-CNF Y led to retraction of dendrites from the total neurite length of ϳ500 m to Ͻ100 m in toxin-treated cells, indicating RhoA activation.
Moreover, the induction of small finger-like extensions, indicating activation of Rac and/or Cdc42 from control levels with 3 spikes per neuron, increased to 12 spikes in mean only when the cells were treated with GST-CNF1. In contrast, incubation of the neurons with GST-CNF Y for 4 h led to a clear retraction of dendrites but not to the formation of small extensions, suggesting that RhoA was activated exclusively. Activation of RhoA in this cell system was verified by pull-down experiments with rhotekin. Moreover, no activation of Rac or Cdc42 by GST-CNF Y was found in PAK pull-down experiments (data not shown).
All together, our data strongly suggest that CNF Y can be used as a new tool for permanent and selective activation of RhoA in mammalian cells without concomitant activation of Rac and Cdc42. DISCUSSION We and others analyzed the molecular mechanism and the structure-function relationship of the cytotoxic necrotizing factor (CNF1) of E. coli (for review, see Ref. 4). Recently, the nucleotide sequence of the CNF Y from Y. pseudotuberculosis was described (7). CNF Y shows 61% identity in its amino acid sequence with CNF1 all over the protein. The activity of the toxin was analyzed using bacterial cell extracts. In these studies, GST-CNF Y induced multinucleation in Hep-2 cells in the same manner as described for GST-CNF1.
In the present study, we cloned and expressed the CNF Y of Y. pseudotuberculosis for the first time and investigated the activity of the recombinant toxin in vitro and in vivo. We used different methods established for the analysis of CNF1 to analyze the activity of CNF Y . First, we analyzed the GST-CNF Ycatalyzed modification of recombinant RhoA by determination of the different migration behavior of unmodified and deamidated RhoA in SDS-PAGE. Incubation of the GTPase with the recombinant toxin led to the same shift as induced by GST-CNF1, indicating the same modification. The same activity was found for the thrombin-cleaved toxins (not shown). Deamidation at Glu 63 of RhoA by GST-CNF Y was further verified by the increase in mass of one Dalton in a peptide corresponding to amino acids 52-68 of the GTPase detected by MALDI-TOF mass spectrometry. This increase in mass corresponds to deamidation of Glu 63 . Deamidation of RhoA leads to the block of intrinsic and GAP-stimulated GTPase activity, which can be analyzed in a filter-binding assay. We show that GST-CNF Y deamidates RhoA, Rac, and Cdc42 in vitro and leads to the block of the GTP-hydrolyzing activity of the respective GTPase. In CNF1, a catalytic dyad of Cys 866 and His 881 was identified (9). These residues are identical in CNF Y . Mutation of these amino acids in GST-CNF Y completely blocks the catalytic activity of the Yersinia toxin. As expected from the overall identity of 61%, including the catalytic amino acids, the data indicate that CNF1 and CNF Y catalyze the same modification, namely the deamidation of Gln 63 or Gln 61 of RhoA, Rac, and Cdc42, respectively.  3-8) for the same time periods as in A and lysed. For analysis of the total amount of the respective GTPase in the lysates, 1/10 of the volume was taken for Western blot analysis. 9/10 of the volume was incubated with bead-coupled GTPase-binding domains of rhotekin, PAK, or POSH. The amounts of activated RhoA, Rac1, and Cdc42 were analyzed by immunoblotting with specific antibodies. The total amount of the GTPases was analyzed by Western-blotting of the lysates.
Notably, we observed marked differences in cell morphology induced by GST-CNF1 compared with GST-CNF Y . HeLa cells treated with GST-CNF Y showed strong formation of stress fibers, but no flattening, membrane ruffling, or formation of filopodia within the first 6 h of toxin treatment, indicating activation of RhoA, but suggesting that Rac and Cdc42 are not activated. After 24 h of treatment with GST-CNF Y , however, the cells were spread out, and membrane ruffling was detectable. In all in vitro studies performed, we detected stronger activity of GST-CNF1 as compared with the recombinant Yersinia toxin.
In contrast, we found a stronger activation in living cells of RhoA by GST-CNF Y than after incubation with the same amount of GST-CNF1. For greater insight into the morphological differences observed, we performed pull-down experiments with bead-coupled GTPase-binding domains of rhotekin, PAK, or POSH, respectively, to exclusively precipitate activated Rho, Rac, or Cdc42 from lysates of toxin-treated cells. To allow comparison of GTPase activation and morphology, HeLa cells were treated with the toxins for the same time periods, fixed, and stained for F-actin. As known from previous studies, incubation of the cells for 2 h with GST-CNF1 leads to activation of RhoA, Rac, and Cdc42, with preference for Rac and with the corresponding formation of stress fibers, membrane ruffles, and filopodia in HeLa cells (13). In contrast, with GST-CNF Y , strong activation of RhoA was evident after 1 h of incubation. RhoA activation was still present after 24 h of toxin treatment, but no activation of Rac and Cdc42 was detected by pull-down experiments during the first 6 h of treatment with GST-CNF Y and, subsequently, no membrane ruffling or formation of filopodia was observed. Activation of the GTPases detected in the pull-down experiments clearly depicts the morphological changes induced in HeLa cells.
The finding of a selective activation of RhoA in intact cells is supported by the toxin-induced morphology of rat hippocampal neurons. In these cells, GST-CNF1 and GST-CNF Y induced retraction of dendrites, indicating activation of RhoA. In contrast, only GST-CNF1 but not GST-CNF Y induced the formation of small finger-like extensions in these cells, indicating that RhoA is activated exclusively by the Yersinia toxin.
Our data show that Rac and Cdc42 is not activated by GST-CNF Y in intact cells within the first 6 h of toxin treatment, whereas RhoA is activated strongly already after 1 h of incubation with GST-CNF Y . In line with this, we detected no degradation of Rac in GST-CNF Y -treated cells, whereas GST-CNF1-activated Rac is degraded by a proteasome-dependent pathway (Refs. 5 and 6 and data not shown). This finding again suggests that Rac is not activated by GST-CNF Y in intact cells. In contrast, we found a decrease of total RhoA in cells treated with the Yersinia toxin for 24 h (compare Fig. 5B, lane 8). Similar results were obtained with thrombin-cleaved CNF Y (not shown).
There are many examples of toxins with a broader substrate specificity in vitro as compared with their actions in the living cell (14 -17). Moreover, substrate specificity has been shown to be even cell-type dependent. For example, the Pseudomonas aeruginosa exotoxins ADP-ribosyltransferase modifies a different set of small GTPases in different cell lines (18). The reason for the tighter specificity of CNF Y in living cells might be due to a different uptake mechanism or different localization within the cell and remains to be analyzed.
All together, our data strongly indicate that CNF1 and CNF Y have different substrate specificity in intact cells. CNF Y is a specific and direct activator of RhoA in HeLa cells and primary hippocampal neurons and can be used as an important new tool for the permanent activation of RhoA without concomitant activation of other Rho family GTPases like Rac and Cdc42.