HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 46, 43659-43666, November 15, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/46/43659    most recent M207828200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Marvaud, J.-C. Articles by Popoff, M. R. Search for Related Content PubMed PubMed Citation Articles by Marvaud, J.-C. Articles by Popoff, M. R. Social Bookmarking                      What's this?

Clostridium perfringens Iota Toxin

MAPPING OF THE Ia DOMAIN INVOLVED IN DOCKING WITH Ib AND CELLULAR INTERNALIZATION^*

Jean-Christophe Marvaud, Bradley G. Stiles^§, Alexandre Chenal^¶, Daniel Gillet^¶, Maryse Gibert, Leonard A. Smith^§, and Michel R. Popoff

From the  CNR Anaérobies, Institut Pasteur, 75724 Paris Cedex 15, France, the ^§ Toxinology Division, United States Army Medical Research Institute of Infectious Diseases, Fort Detrick, Maryland 21702-5011, and the ^¶ Département d'Ingéniérie et d'Etudes des Protéines, CEA-Saclay, Gif sur yvette 91191, France

Received for publication, August 1, 2002, and in revised form, August 30, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Clostridium perfringens iota toxin consists of two unlinked proteins. The binding component (Ib) is required to internalize into cells an enzymatic component (Ia) that ADP-ribosylates G-actin. To characterize the Ia domain that interacts with Ib, fusion proteins were constructed between the C. botulinum C3 enzyme, which ADP-ribosylates Rho, and various truncated versions of Ia. These chimeric molecules retained the wild type ADP-ribosyltransferase activity specific for Rho and were recognized by antibodies against C3 enzyme and Ia. Internalization of each chimera into Vero cells was assessed by measuring the disorganization of the actin cytoskeleton and intracellular ADP-ribosylation of Rho. Fusion proteins containing C3 linked to the C terminus of Ia were transported most efficiently into cells like wild type Ia in an Ib-dependent manner that was blocked by bafilomycin A1. The minimal Ia fragment that promoted translocation of Ia-C3 chimeras into cells consisted of 128 central residues (129-257). These findings revealed that iota toxin is a suitable system for mediating the entry of heterologous proteins such as C3 into cells.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Iota toxin is one of the major lethal and dermonecrotic toxins of Clostridium perfringens type E and a member of the clostridial binary toxin group, which includes Clostridium botulinum C2 toxin, Clostridium difficile ADP-ribosyltransferase, and Clostridium spiroforme toxin (1-6). These toxins, which are also structurally related to the vegetative insecticidal proteins (VIP)¹ produced by Bacillus cereus (6, 7), are organized according to the classic A-B model (8). One domain (A) possesses enzymatic activity, and the other (B) domain is involved in binding to a membrane receptor and translocation of the toxin into a cell. Binary toxins are unique in that the functional domains are localized on separate proteins that are not linked in solution. Iota toxin consists of A (Ia, 47.5 kDa) and B (Ib, 94 kDa) components that undergo limited proteolysis for functional activity (3, 4, 9, 52). After binding to cell-associated Ib and gaining entry into the cytosol, Ia ADP-ribosylates globular skeletal muscle and non-muscle actin at Arg-177 which subsequently disrupts the actin cytoskeleton (10, 11).

Ib shares significant sequence identity (33.9%) with the protective antigen (PA) of anthrax toxins, suggesting that the two proteins have similar modes of action (3, 52). Anthrax toxin is also binary, and PA represents the common receptor-binding component that interacts with the two effecter proteins, edema factor and lethal factor (LF), subsequently mediating their entry into target cells (12).

The clostridial binary and anthrax toxins share a related mechanism of entry into cells via receptor-mediated endocytosis, which requires an acidification step for translocation into the cytosol (4, 13-16). The activated binding components of anthrax, iota, and C2 toxins (PA, Ib, and C2-II, respectively) recognize via their C terminus a cell surface carbohydrate or glycoprotein (C2-II), or they bind to unique surface proteins (PA or Ib) and then oligomerize (14, 15, 17, 18). The whole toxin (binding plus enzymatic components) is transported to the endosome where acidification promotes cytosolic entry of specific enzymatic components (edema factor and/or LF with PA; C2-I with C2-II; Ia with Ib) possibly through channels formed by oligomerized binding components (19). However, unlike anthrax and C2 toxins, iota toxicity is not blocked by alkalinizing agents such as chloroquine and monensin (16, 19). In addition, Ib can transcytose CaCo-2 cells and then promote Ia entry from the opposite cell surface (20).

The ability of bacterial toxins to translocate across cell membranes naturally avails them to be targeted protein delivery systems. Diphtheria toxin and Pseudomonas exotoxin A have been widely used to construct protein chimeras as tools for cell biology and therapeutic applications (21). Binary toxins represent an attractive model that warrants further development as a transport system because the functional domains are independent and localized on different protein chains. Thus, the N-terminal 300 residues of edema factor and LF share sequence similarities and reportedly contain the domain required for binding to PA (22). Fusion proteins between the 254 N-terminal residues of LF and the enzymatic domain of exotoxin A, diphtheria toxin, or tetanus toxin light chain are effectively translocated into cells in a PA-dependent manner (23-25).

Further understanding of functional domains and the translocation mechanisms employed by different toxins will help in designing a general system for delivering proteins into the cytosol of targeted eukaryotic cells. We have shown previously that the catalytic site of Ia is localized within the C terminus (26). In this report, we describe the minimal Ia domain that interacts with Ib by constructing varying sized fusions of Ia with C. botulinum C3 enzyme (23.5 kDa) and subsequently monitoring internalization into Vero cells via the Ib component. The C3 enzyme is an ADP-ribosyltransferase that selectively modifies Rho protein and leads to the loss of actin stress fibers (27, 28). This enzyme does not readily enter cells by itself, thus making it an ideal tool for measuring the biological activity of internalized fusion proteins. In this report, we show that iota toxin efficiently internalizes heterologous proteins, such as the C3 enzyme, into targeted cells.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Construction, Production, and Purification of the Fusion Proteins-- Standard procedures were used for genetic constructions (29). For the C3-Ia constructions, DNA encoding the C3 mature protein (C3^1-211) was amplified from the recombinant plasmid pMRP33 (30) with oligonucleotides generating a NcoI site at the 5'-end and EcoRI site at the 3'-end. The fragment was digested with NcoI and EcoRI and then inserted into pET28b (Novagen) yielding pMRP234. Seven fragments of the Ia gene were amplified by PCR from the recombinant plasmid pMRP67 (3, 52) using Vent polymerase (New England Biolabs) according to the manufacturer's recommendations. DNA fragments encoding the C terminus of Ia were amplified from pMRP252, which encodes for Ia mutated at the catalytic ADP-ribosylation site (E380A) (26). The oligonucleotides introduced EcoRI and SalI sites at the intended new 5'- and 3'-ends, respectively. Amplified fragments were ligated into the SmaI site of pUC18, and all constructs were verified by sequencing via the dideoxy chain termination method (29). Fragments cut with EcoRI-SalI were inserted into pMRP234. The different constructs depicted in Fig. 1 are designated according to the amino acid numbering for each encoded protein. The C3-Ia proteins were linked by four non-native amino acids (YPEF) as a result of the PCR manipulations. All constructs have a C-terminal extension of 9 residues resulting from the vector polylinker and a hexahistidine tag (Fig. 1).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the C3-Ia and Ia-C3 fusion proteins. Mature C3 was fused to the N or C terminus of various sized Ia fragments. The C3-Ia fusion proteins contain a C-terminal extension (VDKLAAALEHHHHHH), and the Ia-C3 molecules contained an N-terminal extension (MGSHHHHHHSSGLVPRGSHMASMTGGQQMGRGSEF) derived from the pET28a vector.

For the Ia-C3 constructions, DNA encoding the C3 mature protein (C3^1-211) was amplified as above with oligonucleotides generating a SalI site at the 5'-end and XhoI site at the 3'-end and was then inserted into pET28a (Novagen) yielding pMRP487. Eight Ia gene fragments were PCR amplified from pMRP252 with oligonucleotides adding an EcoRI site at the 5'-end and a SalI site at the 3'-end and were then cloned individually into pMRP487. The histidine tag and 26 subsequent N-terminal amino acids were derived from the pET28a vector. The Ia and C3 fragments were separated by two residues (Val-Asp) resulting from PCR and cloning manipulations, and the fusion proteins end at the natural stop codon from the C3 gene.

Production and Purification-- Recombinant plasmids were used to transform Escherichia coli BL21 (DE3) and were then grown at 37 °C in 40 µg/ml LB-kanamycin to an A_{600 nm} of 0.6-0.8. Recombinant protein production was then induced by the addition of 1 mM isopropyl -D-thiogalactosylpyranoside, and the cultures were incubated for an additional 4 h at 37 °C. The bacteria were disrupted by sonication and clarified by centrifugation. Recombinant protein was primarily recovered in the pellet following extraction with 50 mM NaH₂PO₄, 300 mM NaCl buffer (pH 8), containing 6 M guanidine HCl for 1 h at 4 °C. The mixture was then centrifuged and the supernatant filtered through a 0.45-µm membrane. Recombinant proteins were adsorbed onto a Co²⁺-charged column (Clontech), eluted with 250 mM imidazole, dialyzed against 50 mM NaH₂PO₄ buffer (pH 7.5) containing 300 mM NaCl plus 15% glycerol, and finally concentrated by dialysis against Aquacide II (Calbiochem).

SDS-PAGE and Immunoblotting-- Proteins were separated by 0.1% SDS and 10% polyacrylamide gel electrophoresis (SDS-PAGE) according to the method of Laemmli (31) and stained with Coomassie Blue. For immunoblotting analysis, proteins were transferred electrophoretically onto nitrocellulose (Hybond C, Amersham Biosciences) that was blocked by 5% (w/v) dry milk in phosphate-buffered saline (PBS) and then incubated overnight at room temperature with a 1:500 dilution of rabbit antibodies against C3 enzyme (28) or Ia (3, 52). Bound antibodies were detected with peroxidase-labeled protein A and chemiluminescence (Amersham Biosciences).

ADP-ribosylation Assays-- The RhoA-glutathione S-transferase (GST) fusion protein was produced from pGEX-2T RhoA in E. coli HB101 provided by A. Hall, and purified on a glutathione-Sepharose 4B column (Amersham Biosciences) by standard methods (32). MgCl₂ (5 mM) was added to all buffers to prevent loss of guanine nucleotides.

Recombinant C3 enzyme was produced in large amounts (about 10% of total protein) in E. coli strain Sure (Stratagene) from pMRP143 consisting of the DNA fragment coding for the C3 mature protein (30) and iota toxin gene promoter in vector pJIR750 (3, 52). Bacterial cell sonicate in 10 mM Tris-HCl buffer (pH 8.5) was clarified by centrifugation, treated with 2 mg/ml protamine sulfate (Merck) for 30 min at 4 °C, and centrifuged again. The supernatant was loaded onto a QAE-Separose A50 (Amersham Biosciences) column equilibrated in Tris buffer. The flow-through containing purified C3 enzyme was concentrated and consisted of one band (25 kDa) after SDS-PAGE and staining with Coomassie Blue.

The ADP-ribosylation solution was modified from Miyaoka et al. (33) and consisted of 50 mM triethanolamine HCl buffer (pH 7.5) containing 100 mM KCl, 5 mM MgCl₂, 10 mM dithiothreitol, 1 mM EDTA, 0.3 mM GDP, 10 mM thymidine, and 2.5 mM [³²P]NAD (PerkinElmer Life Sciences, 20,000 dpm/pmol). For in vitro ADP-ribosylation, recombinant C3, Ia, or one of the various chimeric constructs was added at a final concentration of 10⁸ M to 20 µl of the mixture. Bovine brain extract containing actin and Rho was used as the substrate for iota toxin and C3. The reaction was incubated for 1 h at 37 °C, stopped by adding 7 µl of the sample buffer, and separated by SDS-PAGE. The gel was dried for autoradiography.

The quantitative assays were performed as described above using 3 µg of purified RhoA-GST or 50 µg of total protein from a soluble Vero cell extract, and serial dilutions of C3 or fusion construct. After incubation at 37 °C for 30 min and subsequent SDS-PAGE, radioactive RhoA-GST bands were quantified by a PhosphorImager (Molecular Dynamics) and the IQMac program. For the assays of in vivo ADP-ribosylation of Rho, Vero cells were grown on six-well plates until confluent, and then various concentrations of C3 or fusion protein were incubated at 37 °C in the presence of Ib (molar ratio 1:1) for 4 h. The cells were washed twice with PBS and lysed with ADP-ribosylation buffer containing 1 µg/ml leupeptin, 1 µg/ml pepstatin, 1 mM phenylmethylsulfonyl fluoride, and 0.5% Triton X-100. Cell debris was removed by centrifugation, and the supernatant (50 µg of total protein) was ADP-ribosylated in vitro in a final volume of 20 µl of ADP-ribosylation buffer containing radiolabeled NAD and 10⁷ M C3.

Cytopathic Effects-- Vero cells were grown on coverslips at 37 °C in 5% CO₂ in Dulbecco's modified Eagle's medium supplemented with 5% fetal calf serum. Various concentrations of either a fusion construct or C3 enzyme with Ib (1:1 molar ratio) were added and incubated for the indicated times. Iota toxin (Ia and Ib at 1:1 molar ratio) was used as a control. The coverslips were washed twice in PBS, fixed with 3% paraformaldehyde in PBS for 20 min, permeabilized with 0.1% Triton X-100 in PBS for 5 min, and quenched with 50 mM NH₄Cl in PBS for 10 min. The cells were then incubated with 1 µg/ml FITC-phalloidin (Sigma) for 1 h at room temperature, washed, and mounted in Moviol as described previously (34).

Binding of Fusion Proteins to Cell-bound Ib-- Fluorescence-activated cytometry was done with Ib bound to Vero cells, or control cells without Ib, and fusion protein as described previously (18, 35). Cells were detached from culture flasks by 50 mM EDTA in Hanks' balanced salt solution (HBSS) lacking Ca²⁺ and Mg²⁺. Cells (4-8 × 10⁵/tube) were washed with ice-cold HBSS containing 0.2% bovine serum albumin (BSA) and incubated with 3 × 10⁸ M Ib in HBSS-BSA for 10 min at 37 °C. Cells were washed twice with ice-cold HBSS-BSA and incubated with a 3 × 10⁸ M fusion construct in the same buffer for 15 min on ice. After washing with ice-cold HBSS-BSA, cells were incubated with rabbit anti-Ia serum (1:400) for 1 h on ice, washed, and subsequently incubated for 1 h on ice with FITC-labeled goat anti-rabbit immunoglobulins (Sigma). After a final wash, cells were fixed with 0.5% paraformaldehyde and analyzed by FACsort flow cytometry (BD Biosciences).

Circular Dichroism (CD) Spectropolarimetry-- CD experiments were performed on a CD6 spectrodichrograph (Jobin-Yvon Instruments, Longjumeau, France), as described elsewhere (36). All spectra were recorded at 37 °C in a thermostated cell holder flushed constantly by N₂. Measurements were made in 10 mM NaH₂PO₄ buffer (pH 7.5). CD spectra were measured in the far-ultraviolet range from 190 to 250 nm in a quartz cell of 1-mm path length. The scans were recorded using a bandwidth of 2 nm and an integration time of 1 s at a scan rate of 0.5 nm·s¹. Each far-ultraviolet spectrum represents the average of 15 scans. The spectra were blank corrected, and a smoothing algorithm was applied with the minimum filter using CD6 software (CDMax, filter 5). The mean residue ellipticity (MRE) in far-UV was calculated using []_MRE = (100 × _m)/(C × l × N), where _m is the measured ellipticity in degrees, C is the molar concentration of the protein, l is the path length of the cell in centimeters, and N represents the number of residues. The values of []_MRE are expressed in degrees/cm²/dmol of residue.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Construction of Fusion Proteins-- Fusions between the mature C3 protein and various fragments of Ia were constructed recombinantly to identify the Ia domain that interacts with Ib (Fig. 1). A first set of Ia fragments was fused via the N terminus to C3 (designated as C3-Ia), and contained a C-terminal extension of 9 residues plus a hexahistidine tag encoded by the pET28b vector.

A second series of chimeras (designated as Ia-C3) consisted of C3 fused to the C terminus of various sized Ia fragments (Fig. 1). The Ia fragments contained an N-terminal hexahistidine tag followed by a stretch of 26 amino acids encoded by pET28a, the last of which is Phe. These chimeric proteins ended at the stop codon from the wild type C3 gene. Because the enzymatic site of Ia is located within the C terminus, constructs a, d, f, g, h, i, and j were derived from a mutated Ia gene (E380A) coding for an enzymatically inactive form of Ia (26).

The various fusion proteins were produced and then purified on a Co²⁺ column, yielding homogeneous fusion proteins that migrated according to their expected molecular masses in SDS-PAGE (Fig. 2). Western blots confirmed that these molecules reacted with anti-C3 and anti-Ia sera (data not shown).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   SDS-PAGE and Coomassie Blue staining of the C3-Ia and Ia-C3 fusion proteins (4 µg of total protein for each). The lettering corresponds to each fusion protein as shown in Fig. 1, and standards (kDa) are shown in the left margin.

In Vitro Enzymatic Activity of the Fusion Proteins-- Enzymatic specificity of each fusion protein was checked using bovine brain extract that contains large amounts of actin and other cellular proteins. All protein chimeras, like the C3 control, ADP-ribosylated a 21-kDa protein, whereas only wild type Ia ADP-ribosylated a 45-kDa protein corresponding to cellular actin (Fig. 3). This confirmed that the enzymatic activity of the fusion constructs was caused by C3 and not Ia, and there was seemingly no difference in ADP-ribosyltransferase activity between the chimeras and C3 (Fig. 3). However, further analysis by a quantitative ADP-ribosylation assay with purified RhoA-GST indeed revealed that the C3 fusion proteins retained 60-100% of the wild type C3 enzymatic activity as determined from initial reaction rates, except construct h (Ia^42-454-C3), which possessed only 40% of the native activity (Fig. 4 and Table I). Overall, these data indicated that Ia fragments fused via the C or N terminus to C3 did not adversely affect the enzymatic activity of C3.

View larger version (8K): [in this window] [in a new window]   Fig. 3.   In vitro ADP-ribosylation of bovine brain extract (40 µg of total protein) by 108 M Ia, C3, or fusion protein. The lettering corresponds to each fusion protein as depicted in Fig. 1. ADP-ribosylation was analyzed by SDS-PAGE and autoradiography. Standards (kDa) are indicated in the left margin.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Quantitative in vitro ADP-ribosylation of Rho by C3, C3-Ia (A), or Ia-C3 (B) constructs. ADP-ribosylation was assayed as described under "Experimental Procedures" using 3 µg of RhoA-GST and various concentrations of fusion proteins. 32P incorporation into RhoA-GST was measured, and a representative experiment from three is shown. The lettering corresponds to each fusion protein as shown in Fig. 1.

Ib-mediated Binding of Ia and Chimeric Toxins to Cells-- Interaction of fusion proteins with Ib on the cell surface was assessed by flow cytometry (Table I). All chimeras containing the 129-257 Ia domain possessed 48-90% of the binding activity to cell-bound Ib, except construct c (C3-Ia^42-291; 34% binding activity), relative to whole molecule Ia fused to C3. Other constructs (d, g, j, and l) lacking this domain, or containing a shortened version, possessed less than 10% of the wild type Ia binding activity. These data strongly suggested that the central region of Ia, consisting of residues 129-257, was intimately involved in interactions with cell-bound Ib.

Cytopathic Effects of the Chimeric Toxins to Vero Cells-- To evaluate the biological activity of each fusion protein compared with iota toxin, we investigated their ability to enter Vero cells in the presence of Ib and subsequently disrupt the actin cytoskeleton. Vero cells were grown on coverslips and then treated with various concentrations of Ia, C3, or fusion protein in the midst of Ib (molar ratio 1:1) for 4 h. The cells were then stained for F-actin and analyzed for normal shape and stress fibers. Characteristic actin depolymerization induced by C3 fusion proteins or iota toxin is shown in Fig. 5. Iota toxin, which ADP-ribosylates actin monomers, induced a complete depolymerization of the actin cytoskeleton, whereas C3 and iota-C3 fusion proteins preferentially disrupted stress fibers and induced wrinkled cortical actin as described previously (37).

View larger version (148K): [in this window] [in a new window]   Fig. 5.   Actin cytoskeleton changes because of iota toxin or chimeric proteins. Vero cells were treated with 107 M wild type Ia (iota), C3-Ia (e and f), or Ia-C3 (h, i, k, l, m, n, and o) construct in the presence of Ib. After 4 h at 37 °C, the cells were fixed and stained with FITC-phalloidin. Control cells (ctrl) showed numerous stress fibers, whereas the actin cytoskeleton was completely depolymerized by wild type iota toxin. Ia-C3 constructs containing the Ia 129-257 domain induced loss of stress fibers and wrinkled cortical actin. The C3-Ia constructs e and f caused weak actin cytoskeletal changes, and no significant modification was observed with the other C3-Ia constructs (not shown).

To correlate fusion protein effects upon the actin cytoskeleton after enzymatic modification of Rho, we measured the level of Rho ADP-ribosylation in Vero cells treated with each chimeric toxin. Vero cells were incubated with various concentrations of C3 or iota-C3 fusion proteins and Ib (mol ratio 1:1) for 4 h, the cells were lysed, and 50 µg of soluble protein extract was ADP-ribosylated in vitro with radiolabeled NAD and C3 as described under "Experimental Procedures." Results are expressed as the percentage of ADP-ribosylated Rho in vivo among untreated cells (Fig. 6) and the half-maximal inhibition concentration (Table I). There was a good correlation between toxin concentrations required to modify the actin cytoskeleton and in vivo ADP-ribosylation of Rho.

View larger version (19K): [in this window] [in a new window]   Fig. 6.   In vivo ADP-ribosylation of Rho in Vero cells by C3 fusion proteins and Ib. Cells were treated with various concentrations of a C3-Ia or Ia-C3 chimera and an equal molar concentration of Ib for 4 h at 37 °C. The lettering corresponds to each fusion protein as shown in Fig. 1. Cells were lysed, and 50 µg of total soluble protein was used for in vitro ADP-ribosylation with [32P]NAD and C3 (107 M). The results are expressed as a percentage of the value obtained from control cells not incubated previously with toxin (100% ADP-ribosylation). A representative experiment from three is shown.

Results expressed as the percentage of cells with a normal aspect and ADP-ribosylation of Rho in vivo revealed that the fusion proteins containing C3 linked to the N-terminal part of Ia were poorly active (Figs. 5 and 6 and Table I). The most biologically active C3-Ia fusions were constructs e (C3-Ia^129-330) and f (C3-Ia^129-454). Interestingly, constructs a (C3-Ia^42-454), b (C3-Ia^42-387), and c (C3-Ia^42-291) were inactive although they overlap the Ia domain from the active e and f proteins. Constructs a, b, and c retained the propeptide and proteolytic sites from unprocessed Ia, and previous studies have found that unprocessed Ia is partially activated by a cellular protease(s) (9). Cleavage between C3 and Ia domains from constructs a, b, and c by a cell protease(s) was possibly responsible for this lack of activity. However, the C3-Ia fusion proteins devoid of the Ia propeptide and proteolytic sites were also not very active, unless the -chymotrypsin site generated by DNA cloning between C3 and Ia fragments supported cleavage by a cell protease(s) (9). Another possibility is that C3 enzyme linked to the N terminus of an Ia fragment was poorly translocated into cells through Ib or was impaired in the recognition and modification of Rho. Because the in vitro enzymatic activity of C3 fusion proteins toward Rho was not altered significantly, the ability of Ib to transport these chimeric molecules conformationally intact into the cytosol represents a critical step. According to the overall activity of constructs e and f, the central region of Ia seemed intimately involved in translocation.

In contrast to the N-linked chimeras, C3 fused to the C terminus of Ia possessed cytotoxic levels equivalent to those of wild type iota toxin (Table I). Construct h (Ia^42-454-C3), in the presence of Ib, modified the actin cytoskeleton and ADP-ribosylated cellular Rho in a dose-dependent manner. The concentration that induced a 50% cytopathic effect after 4 h was about 10⁹ M, similar to the wild type iota toxin (Table I). The C3 fusions containing Ia^62-454, Ia^42-257, or Ia^129-257 and incubated with Ib plus Vero cells had identical cytopathic effects compared with the full-length Ia^42-454. These chimeric toxins were neutralized by preincubating them with a 1:100 dilution of anti-Ib or anti-C3 serum before adding to cells (data not shown). Incubation of any fusion protein, C3 alone, or C3 plus Ib (10⁷ M), with Vero cells did not cause any morphological changes even after 24 h (Fig. 7 and data not shown). Ia^129-257 (construct m) was the smallest fragment, when fused to C3, which induced significant cytopathic effects in the midst of Ib. Relative to construct m, deletion of just 27 amino acids from the N terminus (Ia^156-257) or 35 amino acids from the C terminus (Ia^42-222), respectively, decreased the in vivo ADP-ribosylation activity of Rho 30-fold and >1,000-fold, in the presence of Ib (Fig. 6 and Table I).

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Influence of -chymotrypsin and bafilomycin A1 (Baf) on the ADP-ribosyltransferase activity of construct k (Ia42-257-C3) in Vero cells and time course of intoxication. A, 107 M construct k was treated with 10 µg/ml -chymotrypsin (30 min at 37 °C) prior to incubation with Vero cells and equal molar Ib (k+alpha). Vero cells were also pretreated with 200 nM bafilomycin for 30 min at 37 °C prior to incubation with k and Ib (k + Baf) at the indicated concentrations. As a control, purified C3 and Ib were incubated at equal concentrations (C3 + Ib). Vero cells were incubated with toxins for 4 h at 37 °C in a CO2 incubator, lysed, and processed for Rho ADP-ribosylation as indicated in the Fig. 6 legend. B, Vero cells were treated with 107 M construct i (Ia62-454), m (Ia129-257), or wild type Ia in the presence of Ib. At the indicated times, cells were lysed and processed for Rho (constructs i and m) or actin ADP-ribosylation (wild type iota toxin, Ia). The same protocol for Rho ADP-ribosylation was used to measure actin ADP-ribosylation with the exception that C3 was replaced by 107 M Ia. Mean values of three experiments are shown.

Activation of the chimeras with -chymotrypsin, as reported for the native Ia component (9), was tested with construct k (Ia^42-257-C3). No significant increase in ADP-ribosylation of cellular Rho in vivo was observed (Fig. 7); however, it cannot be ruled out that activation may occur via endogenous cellular proteases.

CD Spectropolarimetry of Fusion Proteins-- We also evaluated the effect of structural deletions within Ia on the overall folding of fusion proteins by far-ultraviolet CD, which provides information on the secondary structure of proteins. Fig. 8 shows the spectra obtained for fusion proteins l (Ia^42-222-C3), m (Ia^129-257-C3), n (Ia^156-257-C3), and o (Ia^177-257-C3) as well as those for purified C3 and Ia. All spectra from the fusion proteins exhibited a positive signal below 200 nm, a minimum around 208 nm, and a shoulder between 220 and 230 nm. These signals are indicative of folded proteins containing secondary structures and exclude the possibility of large unfolded regions (38). Although the spectrum of fusion protein l displays a decreased MRE, its overall shape is similar to that of the three other chimeras. The spectra are characteristic of proteins containing both -helices and -sheets, revealing a contribution of signals from both C3 and Ia (Fig. 8, inset). Indeed, the crystal structures of C3 (39, 40) and VIP2 (41), an analog of Ia, produced by B. cereus, both reveal / content that was evident for fusion proteins h (Ia^42-454-C3) and a (C3-Ia^42-454), which contain full-length Ia linked via the N and C terminus of C3, respectively (not shown). Overall, the results revealed that a loss of binding activity to Ib and subsequent cell intoxication by these fusion proteins was not connected to a dramatic unfolding of the Ia fragments.

View larger version (23K): [in this window] [in a new window]   Fig. 8.   Far-ultraviolet CD spectra of the fusion proteins l (Ia42-222-C3), m (Ia129-257-C3), n (Ia156-257-C3), and o (Ia177-257-C3), as well as the isolated C3 and Ia proteins (inset).

Cell Entry of Chimeric Protein k (Ia^42-257-C3) and Time Course of Intoxication-- Both iota and C2 toxins enter cells by receptor-mediated endocytosis and require a low pH-dependent membrane translocation step inhibited by bafilomycin A1 (4, 14, 15). However, in contrast to C2 toxin, the cytopathic effects of iota toxin are not blocked by weak bases or ionophores such as chloroquine and monensin (16, 19). These pharmacological drugs were tested with construct k incubated with Ib and Vero cells. Bafilomycin A1 (200 ng/ml) prevented modification of the actin cytoskeleton either by Ia or construct k (10⁷ M) with an equimolar concentration of Ib (Fig. 7A), unlike 0.5 mM chloroquine and 10 µM monensin (data not shown). This suggests that construct k enters Vero cells via Ib by a mechanism similar to that of wild type iota toxin. Similar findings were also evident with constructs i (Ia^62-454-C3) and m (Ia^129-257-C3) (data not shown).

To determine the intoxication time course of wild type iota toxin and C3-iota chimeras, we analyzed the modification of Rho induced by constructs i and m over time compared with actin ADP-ribosylation by iota toxin. Construct i encompasses most of Ia, and m contains the minimal sequence of Ia mediating Ib-dependent internalization. Within 1 h, Ia-C3 fusion molecules i and m, in the presence of Ib, readily ADP-ribosylated Rho in vivo (Fig. 7B) with similar kinetics (Table I). Actin ADP-ribosylation induced by iota toxin was assayed like the in vivo ADP-ribosylation of Rho, except that 10⁷ M wild type Ia was used instead of C3 enzyme. Kinetics of the actin ADP-ribosylation with iota toxin were similar to that of Rho ADP-ribosylation with constructs i and m (Fig. 7B). This indicated that Ib efficiently mediated the internalization of chimeric molecules i and m just like wild type Ia.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Ib shares significant sequence homology and probably four structural and functional domains with PA (12, 42). The most highly conserved regions lie within the N-terminal domains. Domain 1' (residues 168-258) of mature PA forms a hydrophobic surface considered to be the binding site for edema factor and LF. Domain 2 (residues 259-487) contains a -barrel and large flexible loop implicated in translocating toxin across a lipid membrane. Domains 1 and 2 of PA, respectively, share 34 and 41% sequence identity with the corresponding regions of Ib, whereas domain 4 (residues 596-735) recognizes a cell surface receptor and shares less than 10% identity with Ib (12, 42), thus supporting previous results suggesting that PA and Ib interact with unique receptors (16). Like PA, the N-terminal domain (residues 1-106) of Ib is important for Ia docking, and the C-terminal domain (residues 465-665) is involved in binding to an unknown cell surface protein (35). The receptor binding domain has also been located within the C terminus of the closely related C. botulinum C2-II molecule (17). It is possible that the binding components of various bacterial binary toxins, such as PA, Ib, and C2-II, use a similar mechanism to translocate an enzymatic component into the targeted cell. The minimum region of LF which interacts with PA consists of 224 residues (23, 43). The main focus of our study was to determine the minimal Ia fragment that interacts with Ib and is subsequently translocated into a cell.

In addition to the inherent similarities of anthrax and C2 toxins with iota, a recent investigation has elucidated the three-dimensional structure of B. cereus VIP2, an ADP-ribosyltransferase that is structurally homologous to Ia and shares 30% amino acid identity (41). VIP2 consists of two structurally homologous but nonsymmetrical domains (N and C domains) that probably arose from gene duplication. The C domain is involved in NAD binding and catalysis (26, 41), whereas the N domain interacts with the cell-binding component. Our results show that the Ia central region, residues 129-257, is critical for interacting with cell-bound Ib and mediating Ib-dependent internalization of C3 chimeric molecules into Vero cells. The 129-257 segment of Ia corresponds to the C-terminal half of the N domain, which forms a compact core containing eight -strands and one -helix (1-5-2-8) (41). Deletion of the two C-terminal -strands (fusion protein l) drastically reduced fusion protein interactions with cell-bound Ib, and deletion of the -strand within the N terminus, which decreased the cytotoxic activity, probably had a similar effect. These data indicated that the whole core structure is required for effective Ia interactions with Ib. This structural organization of the Ia binding domain resembles that of LF. The N-terminal domain of LF involved in binding to PA was mapped to residues 40-263, which forms a compact structure of 11 packed helix bundles and 4 -strands. The first 27 residues of LF are unstructured, and residues 28-39 contain a helix projecting away from the molecule, both of which are not required for binding activity (43, 44). Like LF, it was initially discovered that the N terminus of C2-I (amino acids 1-225) permits C2-II-dependent transport of fusion proteins. However, the minimal region of C2-I involved in binding to C2-II and translocation of fusion proteins has more recently been localized to amino acids 1-87, which encompasses helices 1-4 of VIP2 (43-45). This contrasts with the minimal region of Ia corresponding to the -strand core of the N-terminal domain of VIP2 (41).

Iota and C2 toxins are part of two distinct families of binary toxins (6) that probably evolved from a common ancestor, yet the functional domains of these proteins are seemingly organized in a slightly different way. Therefore, in contrast to Ia, C2-I does not possess a signal peptide or propeptide sequence and thus does not require proteolytic activation (9). In addition, the enzymatic component of C2 toxin only recognizes cellular actin, whereas iota toxin modifies all actin isoforms (45). These toxins do not share common epitopes, nor do the heterologous components form a biologically functional hybrid toxin (6). Now, it appears that the enzymatic components of iota and C2 toxin also differ in domain interactions with their binding components.

The precise mechanism of how clostridial binary toxins translocate proteins across lipid membranes remains an intriguing mystery. The enzymatic component is thought to cross the endosomal membrane via a channel formed by oligomerized binding components inserted into the lipid bilayer (19, 46, 48, 49). As reported for other bacterial proteins, such as diphtheria toxin and anthrax toxins, the enzymatic component is translocated through a channel in an unfolded form followed by refolding within the cytosol (47). It is very plausible that proteins recombinantly linked to Ia could impair unfolding and proper refolding, thus inhibiting efficient translocation into the cytosol. It is not known whether the N or C terminus of Ia is first translocated through the Ib-induced channel. However, it is possible that a specific sequence is not required at the extremity of Ia to drive it through an Ib-induced channel. Indeed, the Ia-C3 proteins that were transported most efficiently did not contain a conserved sequence within the N-terminal region of the various Ia fragments. All of these constructs were seemingly transported by the same mechanism used by wild type iota toxin, as evidenced from inhibition by bafilomycin A1 but not by chloroquine or monensin. It still remains an enigma as to what triggers internalization of Ia or the enzymatic components of other bacterial binary toxins like anthrax or C2, into cells.

Several protein toxins such as diphtheria toxin, Pseudomonas exotoxin A, C2, and anthrax naturally traverse the cell membrane and have been used successfully for delivering heterologous proteins into the cytosol of targeted cells (21, 45, 50). Toxins differ via recognition of unique cell surface receptors, entry mechanisms, and delivery of proteins into various cell compartments. Those specific aspects of each toxin might be of special interest for some drug applications. Additionally, bacterial binary toxins such as iota have functional domains localized on separate proteins that facilitate their use as biological tools. Heterologous proteins need only to be fused to one part of an enzymatic component leading to a shorter construct versus the use of a single chain holotoxin that may induce steric problems. In addition, short recombinant proteins are generally easier to produce. Iota toxin offers the distinct advantage of requiring a short Ia fragment to interact with Ib and undergo subsequent translocation. An Ia fragment of just 129 residues is sufficient to internalize C3 into cells via an Ib-dependent mechanism that is as efficient as the wild type toxin, whereas the minimum size for the enzymatic component (LF) of anthrax toxin which promotes translocation of fusion proteins through PA consists of 224 N-terminal residues (23, 43). Moreover, fusion to the enzymatic component does not alter folding and receptor binding activity of an independent binding component, unlike that reported for a single diphtheria toxin-C3 chimera (51). Finally, the results presented here suggest that iota toxin represents a potential biological tool that may be widely applicable for the introduction of heterologous proteins into cells.

	ACKNOWLEDGEMENTS

We thank P. Binder and C. Dane for supporting this project.

	FOOTNOTES

^* This work was supported by DGA Contract 0034056 and a Direction Générale de l'Armement fellowship (to J. C. M.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: CNR Anaérobies, Institut Pasteur, 28 rue du Dr. Roux, 75724 Paris Cedex 15, France. E-mail: mpopoff@pasteur.fr.

Published, JBC Papers in Press, September 6, 2002, DOI 10.1074/jbc.M207828200

	ABBREVIATIONS

The abbreviations used are: VIP, vegetative insecticidal protein(s); BSA, bovine serum albumin; C2I, enzymatic component of C2 toxin; C2II, binding component of C2 toxin; FITC, fluorescein isothiocyanate; GST, glutathione S-transferase; HBSS, Hanks' balanced salt solution; Ia, enzymatic component of iota toxin; Ib, binding component of iota toxin; LF, lethal factor from Bacillus anthracis; MRE, mean residue ellipticity; PA, protective antigen from B. anthracis; PBS, phosphate-buffered saline.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				D. Fehr, S. E. Burr, M. Gibert, J. d'Alayer, J. Frey, and M. R. Popoff Aeromonas Exoenzyme T of Aeromonas salmonicida Is a Bifunctional Protein That Targets the Host Cytoskeleton J. Biol. Chem., September 28, 2007; 282(39): 28843 - 28852. [Abstract] [Full Text] [PDF]

				H. Barth, K. Aktories, M. R. Popoff, and B. G. Stiles Binary Bacterial Toxins: Biochemistry, Biology, and Applications of Common Clostridium and Bacillus Proteins Microbiol. Mol. Biol. Rev., September 1, 2004; 68(3): 373 - 402. [Abstract] [Full Text] [PDF]

				M. Nagahama, A. Yamaguchi, T. Hagiyama, N. Ohkubo, K. Kobayashi, and J. Sakurai Binding and Internalization of Clostridium perfringens Iota-Toxin in Lipid Rafts Infect. Immun., June 1, 2004; 72(6): 3267 - 3275. [Abstract] [Full Text] [PDF]
</