Invasion of epithelial cells by Shigella flexneri involves entry and intercellular dissemination. Entry of bacteria into non-phagocytic cells requires the IpaA–D proteins that are secreted by the Mxi–Spa type III secretion machinery. Type III secretion systems are found in several Gram-negative pathogens and serve to inject bacterial effector proteins directly into the cytoplasm of host cells. In this study, we have analysed the IpgD protein of S. flexneri, the gene of which is located on the virulence plasmid at the 5′ end of the mxi–spa locus. We have shown that IpgD (i) is stored in the bacterial cytoplasm in association with a specific chaperone, IpgE; (ii) is secreted by the Mxi–Spa type III secretion system in amounts similar to those of the IpaA–D proteins; (iii) is associated with IpaA in the extracellular medium; and (iv) is involved in the modulation of the host cell response after contact of the bacterium with epithelial cells. This suggests that IpgD is an effector that might be injected into host cells to manipulate cellular processes during infection.
Bacteria of Shigella spp., including Shigella flexneri, are facultative intracellular pathogens that are the causative agents of shigellosis, a severe diarrhoeal disease in humans. Transmission occurs via the faecal–oral route or by contaminated food and water. These bacteria are able to enter into non-professional phagocytic cells, such as epithelial cells, and disseminate from cell to cell. The invasive phenotype is conferred by a virulence plasmid of 200 kb that encodes the factors necessary for bacterial entry and intracellular motility (Sansonetti et al., 1982). A 30 kb region of this virulence plasmid is necessary and apparently sufficient to confer the invasive ability to an Escherichia coli K-12 strain (Maurelli et al., 1985). This entry region contains 31 genes clustered in two divergently transcribed loci. The phenotypic characterization of mutants in most of these genes indicated that one locus encodes a specialized type III system, the Mxi–Spa secretion apparatus, and the other locus encodes secreted virulence factors, the IpaA–D proteins, and a cytoplasmic chaperone, IpgC (for a review, see Parsot, 1994). Type III secretion systems are present in a variety of Gram-negative pathogens, in which they serve to translocate bacterial effector proteins into the cytoplasm of the host cell (for a review, see Hueck, 1998). Several components of the secretion apparatus share sequence similarities with components of flagellar basal bodies. Electron microscopic analysis of the Salmonella typhimurium and S. flexneri systems revealed similar macromolecular structures spanning the bacterial envelope, with a needle emanating from the bacterial surface (Kubori et al., 1998; Blocker et al., 1999).
The analysis of S. flexneri mutants indicated that, in addition to components of the secretion machinery, the secreted IpaB, IpaC and IpaD proteins and the cytoplasmic IpgC chaperone are essential for entry into HeLa cells and for virulence in vivo, as assayed by the induction of keratoconjunctivitis in guinea pigs (Ménard et al., 1993). These proteins are also required for escape from the phagocytic vacuole once bacteria are internalized by macrophages (High et al., 1992) and for expression of a contact haemolytic activity on erythrocytes (Clerc et al., 1986). Subcellular localization and co-immunoprecipitation studies revealed that, under non-secreting conditions, IpaB and IpaC accumulate in the bacterial cytoplasm and associate independently with IpgC (Ménard et al., 1994a). Once secreted, IpaB and IpaC form a soluble complex in the extracellular medium, and latex beads coated with this complex are internalized by epithelial cells (Ménard et al., 1994a; 1996). It has been shown recently that the haemolytic activity results from the formation of a pore of about 25 Å in the membrane of erythrocytes, and that IpaB, IpaC, IpaA and IpgD, another protein encoded by the entry region, are associated with the membrane of lysed erythrocytes (Blocker et al., 1999).
Previous analysis of a S. flexneri ipgD mutant suggested that IpgD was a secreted protein and not required for entry into epithelial cells (Allaoui et al., 1993a). In the present study, we have further analysed the IpgD protein and investigated its potential role as an effector of bacterial and host cell interactions. Like IpaA–D proteins, IpgD is stored in the bacterial cytoplasm before it is secreted in response to activation of the type III secretion system. Storage of IpgD in the cytoplasm requires its association with a cytoplasmic chaperone, IpgE, encoded by the gene located immediately downstream of the ipgD gene. After secretion, IpgD forms a complex with IpaA in the extracellular medium. When used to infect tissue culture cells, the ipgD and ipgE mutants elicit entry structures that have an altered morphology compared with those produced by the wild-type strain. These results indicate that IpgD is implicated in the process of invasion of eukaryotic cells by S. flexneri and suggest that it is a translocated effector that modulates cellular responses induced upon contact of bacteria with epithelial cells.
IpgD is secreted by the Mxi–Spa secretion machinery in amounts similar to those of Ipa proteins
Previous analysis of an ipgD mutant (Fig. 1) indicated that a protein of the same size as IpgD (60 kDa) was found in supernatants of the wild-type strain, but not of the ipgD mutant, which suggested that IpgD was secreted (Allaoui et al., 1993a). To determine the localization of IpgD, we generated a monoclonal antibody (mAb) that specifically recognized IpgD. Cultures of the wild-type strain and ipgD and mxiD mutants in the exponential phase of growth were induced with Congo red to activate the Mxi–Spa secretion machinery (Bahrani et al., 1997), and the proteins present in supernatants were analysed by SDS–PAGE and Coomassie blue staining or immunoblotting using mAbs against IpaB, IpaC and IpgD (Fig. 2). IpgD was present in the supernatant of the wild-type strain but not in that of the mxiD mutant, which confirmed that it was a secreted protein and that its secretion required the Mxi–Spa machinery. Similar amounts of IpaB and IpaC were present in supernatants of the wild-type and ipgD strains, indicating that IpgD was not required for the secretion of Ipa proteins. Examination of the Coomassie blue-stained gel of proteins secreted by the wild-type strain suggested that the amount of IpgD was similar to that of the Ipa proteins. Western blot analysis of whole-cell extracts of the wild-type and mxiD mutant strains showed that IpgD was present in bacteria before exposure to Congo red, indicating that IpgD was expressed even when the secretion machinery was not active (data not shown). This was confirmed by measuring the β-galactosidase activity expressed from the ipgD–lacZ transcriptional fusion in strain SF134 (Allaoui et al., 1993a). As reported previously for transcription of ipaA–lacZ and mxiD–lacZ fusions (Bahrani et al., 1997), transcription of the ipgD–lacZ fusion increased during the exponential phase of growth and ceased in stationary phase (data not shown). These results indicated that IpgD, like the IpaA–D proteins, was stored in the cytoplasm of bacteria growing in laboratory media at 37°C and that its secretion was dependent on the Mxi–Spa secretion machinery.
IpgD is associated with IpgE in the bacterial cytoplasm
The observation that IpgD was stored in the cytoplasm before being secreted by the Mxi–Spa secretion machinery led us to investigate whether its storage, like that of IpaB and IpaC, involved an association with a specific chaperone. A likely candidate was the IpgE protein encoded by the gene located immediately downstream from ipgD on the virulence plasmid (Allaoui et al., 1993a). IpgE is predicted to have a small size (13.7 kDa) and an acidic pI (4.04), two features that are characteristic of chaperones for proteins secreted by type III secretion machineries (for a review, see Wattiau et al., 1996). To investigate the potential association of IpgD and IpgE, we constructed plasmids pKNE4 and pKNE6 that expressed a GST– or a His6–IpgD fusion protein, respectively, together with IpgE (Fig. 1). Each of these plasmids was introduced into the wild-type strain M90T, and cleared extracts of recombinant strains were passed over affinity resins. After extensive washing, proteins bound to the cobalt resin were eluted in the presence of imidazole and analysed by SDS–PAGE and Coomassie blue staining (Fig. 3A). Three proteins with apparent molecular weights of 38, 26 and 14 kDa, the last existing in greater amounts than the other two, were co-eluted with His6–IpgD. The N-terminal sequence of the 14 kDa protein (MEDLA) corresponded to the first five residues of IpgE (Allaoui et al., 1993a), and the sequences obtained for the 26 kDa protein (ATVS) and the 38 kDa protein (MKPVT) corresponded to the first residues of the ribosomal protein RpsB and the repressor LacI respectively. We, likewise, incubated cleared lysates of the recombinant strain harbouring pKNE4 (GST–IpgD and IpgE) with glutathione beads. After extensive washing, beads were treated with thrombin to separate IpgD from the GST moiety of the hybrid protein. SDS–PAGE and Coomassie blue staining of proteins eluted from the beads indicated that only one protein of 14 kDa co-purified with IpgD, and determination of the N-terminal sequence of this protein identified it as IpgE (data not shown).
To confirm the interaction between IpgD and IpgE, we constructed the plasmid pKNE8 that expressed a His6–IpgE fusion protein. Cleared lysates of a derivative of M90T harbouring plasmids pKNE8 and pMM100 were passed over a cobalt affinity resin, and proteins present in the flowthrough, bound to the column and eluted with imidazole, were analysed by SDS–PAGE and immunoblotting using mAbs against IpgD, IpaC and the His6-tag (Fig. 3B). Both His6–IpgE and IpgD were bound to the beads and eluted with imidazole, whereas IpaC, used as a control, was present only in the flowthrough. These results indicated that IpgD and IpgE were associated in the cytoplasm of S. flexneri and that only IpgE was specifically bound to IpgD in this compartment.
IpgE is required for the production and stabilization of IpgD
The association of IpgD with IpgE in the bacterial cytoplasm suggested that IpgE might act as a specific chaperone to stabilize IpgD. To test this hypothesis, a non-polar ipgE mutant was generated by allelic replacement to construct strain SB21 (Fig. 1). Cultures of M90T (wild type), SF701 (ipgD) and SB21 (ipgE) in the exponential phase of growth were induced with Congo red, and bacterial pellets and supernatants were analysed by SDS–PAGE and Coomassie blue staining or immunoblotting with mAbs against IpaB, IpaC and IpgD (Fig. 4A). The amount of IpgD that was present in the whole-cell extract of the ipgE mutant was greatly reduced compared with that present in the wild type, and no IpgD was detected in the supernatant. In contrast, the production and secretion of IpaB and IpaC were not affected by inactivation of ipgE. The ipgE mutant was transformed with plasmids pKNE8 (His6-IpgE) and pMM100 (LacI), and induction of expression of His6–IpgE by IPTG restored both the production and the secretion of IpgD (Fig. 4B), confirming that the decreased amounts of IpgD in the ipgE mutant were indeed the result of the absence of IpgE and not a cis-acting effect of the ipgE mutation.
To compare the stability of Ipa proteins and IpgD in wild-type and ipgE strains, we added chloramphenicol to bacterial cultures to inhibit de novo protein synthesis and analysed bacterial lysates for the presence of IpaB, IpaC and IpgD at different time points. In wild-type bacteria, no significant decrease in the amounts of IpgD, IpaB and IpaC was observed after 2 h of chloramphenicol treatment (Fig. 4C). In the ipgE mutant, the amounts of IpaB and IpaC were stable, whereas no IpgD was detected after 30 min of chloramphenicol treatment. To determine whether IpgE was sufficient to stabilize IpgD in the cytoplasm, the GST–IpgD fusion protein was expressed either alone or in combination with IpgE in E. coli. When GST–IpgD was expressed alone, there were small amounts of the full length 85 kDa protein, and the majority of the protein bound to the beads and present in the eluate fractions consisted of the GST moiety alone (Fig. 4D). In contrast, when GST–IpgD was expressed in the presence of IpgE, the full-length fusion protein was eluted as a stable protein, and only minor degradation occurred. We concluded from these results that IpgE was necessary and sufficient for the stability of IpgD in the cytoplasm and that it acted as a chaperone specific for IpgD.
IpgD is associated with IpaA in the extracellular medium
Association of the chaperone IpgC to IpaB and IpaC in the cytoplasm has been proposed to prevent the premature association of these proteins, which form a complex in the extracellular milieu (Ménard et al., 1994a). To investigate whether IpgD was associated with another protein in the extracellular medium, we took advantage of the fact that His6-IpgD was efficiently secreted by the Mxi–Spa type III secretion machinery. After exposure of the wild-type M90T derivative expressing His6–IpgD to Congo red, the recombinant protein was detected in the extracellular medium (Fig. 5). The secreted protein still carried the His6 tag, as indicated by its reactivity with a His6-specific mAb and its ability to bind to the cobalt resin. This result indicated that the presence of the His6 tag at the N-terminus of the recombinant His6–IpgD protein did not affect its secretion signal. Derivatives of the wild-type strain harbouring pKNE6 (His6–IpgD and IpgE) and pKNE8 (His6–IpgE) were induced with Congo red, and Triton X-100 was added to the supernatants before they were centrifuged at high speed to remove aggregated proteins. Soluble secreted proteins were incubated with the cobalt affinity resin for 1 h, and proteins bound to the resin were analysed by SDS–PAGE and Coomassie staining (Fig. 6). As His6–IpgE was not secreted (Fig. 5), the extract of M90T harbouring pKNE8 was used as a control to examine any non-specific interactions of secreted proteins with the cobalt beads. After incubation of the proteins secreted by this strain with the cobalt resin, neither IpgD nor any of the IpaA–D proteins were bound to the beads, which indicated that IpgD and the Ipa proteins had no intrinsic ability to bind to the resin. In the sample prepared from the strain expressing His6–IpgD, a protein of 60 kDa corresponding to His6–IpgD was bound to the resin, in amounts detectable by Coomassie staining. Examination of the Coomassie-stained gel indicated that a protein of 72 kDa, the size of IpaA, had also been retained on the resin. Western blot analysis using polyclonal antibodies against IpaA confirmed that this protein was indeed IpaA (Fig. 6). Furthermore, the use of a human serum convalescent from a shigellosis, which strongly recognized the secreted IpaA–D proteins, indicated that no Ipa proteins other than IpaA were bound to the resin (Fig. 6). These results suggested that IpgD and IpaA were associated in the extracellular medium.
As shown in Fig. 6, IpgD was not recognized by the human antiserum. We then screened sera from 20 human patients by Western blotting for their reaction with IpgD. None of the sera recognized IpgD in bacterial supernatants or the purified His6–IpgD protein (data not shown), suggesting that, in contrast to IpaA–D proteins, IpgD is weakly immunogenic in humans.
Actin rearrangements induced upon contact of bacteria with epithelial cells are altered in the ipgD mutant
Previous analysis of the ipgD mutant indicated that IpgD is not absolutely required for the entry of S. flexneri into epithelial cells (Allaoui et al., 1993a). The demonstration that IpgD and IpaA were associated in the extracellular medium led us to investigate whether IpgD, like IpaA, was involved in the modulation of actin rearrangements triggered by bacteria upon contact with epithelial cells. HeLa cells were infected by the wild-type and ipgD strains for 15 min, and bacteria and F-actin were analysed by immunofluorescence and confocal scanning microscopy (Fig. 7). In contrast to the wild-type strain (Fig. 7A–E), which caused prominent cell surface and actin rearrangements around its attachment sites, the ipgD mutant (Fig. 7A′–E′) induced dense, cup-like actin rearrangements beneath adherent bacteria, which were also enriched for the cytoskeletal protein vinculin (Fig. 7B′ and C′). Z-projections of the confocal images (Fig. 7E and E′) and scanning electron microscopic analysis (Fig. 7F and F′) showed that the mutant provoked fewer actin rearrangements and less membrane ruffling on the cell surface. The same phenotype was also observed for the ipgE mutant (data not shown). This suggested that IpgD was involved in the formation of the entry structures that are triggered by bacteria on epithelial cells.
In this study, we have analysed the IpgD protein of S. flexneri, the gene of which is located on the virulence plasmid at the 5′ end of the mxi–spa locus. We have shown that IpgD (i) is stored in the bacterial cytoplasm in association with a specific chaperone, IpgE; (ii) is secreted by the Mxi–Spa type III secretion system in amounts similar to those of the IpaA–D proteins; (iii) is associated with IpaA in the extracellular medium; and (iv) is involved in the modulation of the host cell response after contact of bacteria with epithelial cells. These data and previous results on proteins secreted by the type III secretion machinery of other Gram-negative bacteria suggest that IpgD is an effector protein that might be injected into host cells to manipulate cellular processes during Shigella infection.
Proteins secreted by type III machineries are often stored in the bacterial cytoplasm in association with a specific chaperone (Wattiau et al., 1994; 1996). For example, IpaB and IpaC are independently associated with IpgC, which stabilizes IpaB and prevents the premature association of IpaB and IpaC before secretion (Ménard et al., 1994a). We used His-tagged recombinant proteins to investigate whether IpgD was associated with another protein in the cytoplasm of S. flexneri. IpgE, which is encoded by the gene located downstream from ipgD on the virulence plasmid (Allaoui et al., 1993a), was co-purified with a recombinant His6–IpgD protein, and IpgD was co-purified with a recombinant His6–IpgE protein expressed in S. flexneri. Inactivation of ipgE markedly decreased the amounts and the half-life of IpgD present in the cytoplasm, and expression of His6–IpgE restored production of IpgD in the ipgE mutant. Moreover, expression of IpgE together with a GST–IpgD recombinant protein in E. coli increased both the production and the stability of the recombinant protein. This observation suggests that stabilization of IpgD by IpgE not only functions when the protein is stored in the cytoplasm of S. flexneri before being released, but also when it accumulates in E. coli. The observation that His6-IpgE protein was not secreted and that IpgE was not bound to His6–IpgD in the extracellular milieu, in addition to its role in stabilizing IpgD in the cytoplasm of both S. flexneri and E. coli, suggested that the localization of wild-type IpgE was solely cytoplasmic.
Proteins homologous to IpgD and IpgE, designated SigD and SigE, respectively, have been described in S. typhimurium. Their genes were identified in a screen for invasion-defective mutants, and transposon insertion in sigE, which is located downstream from sigD, abolished the secretion of a protein that had the size of SigD (Hong and Miller, 1998). The phenotype of the ipgE mutant suggests that SigE might also be required for the stability of SigD in the cytoplasm of S. typhimurium.
The recombinant His6–IpgD, which contained 10 additional residues at its N-terminus, was still secreted by the Mxi–Spa type III secretion system. Studies on the Yop virulon of Yersinia spp., which represents the paradigm of type III secretion, suggested that the secretion signal is located in the N-terminus of the Yop proteins, as fusion proteins containing the first 15 and 17 residues of YopE and YopH, respectively, were secreted efficiently (Sory et al., 1995). Replacement of the first six residues of YopH by the first 12 residues of β-galactosidase resulted in a hybrid protein that was not secreted (Michiels and Cornelis, 1991). Anderson and Schneewind (1997; 1999) suggested that, in Yersinia spp., mRNA translation and type III secretion are coupled, and that the signal for secretion of Yop proteins was contained in the mRNA. This is probably not the case for the IpaA–D and IpgD proteins of S. flexneri, as these proteins are stored in the cytoplasm and can be secreted upon activation of the secretion machinery (Ménard et al., 1994b; Watarai et al., 1995). Further studies are required to identify the secretion signal of the IpaA–D and IpgD proteins.
In type III secretion systems, specific cytoplasmic chaperones not only stabilize proteins in the bacterial cytoplasm, but also prevent premature association of proteins before secretion (Ménard et al., 1994a). Binding of IpgD to IpgE in the cytoplasm suggested that, once secreted, IpgD might be associated with another protein. Indeed, after induction of secretion, IpaA was co-purified with His6–IpgD. The interaction between IpaA and IpgD was specific, as none of the other Ipa proteins was retained on the resin, and IpaA was not bound to the resin in the absence of His6–IpgD. Association of IpgD and IpaA in the extracellular medium raised the possibility that both proteins might share the same chaperone in the cytoplasm, i.e. that IpaA, like IpgD, might be associated with IpgE. However, similar amounts of IpaA were detected in the extracts of the wild-type and ipgE strains, and IpaA was not co-purified with His6–IpgE. This indicated that IpgE is neither required for stability nor associated with IpaA. The latter result also indicated that IpgD is the only protein chaperoned by IpgE and suggested that the association of IpgD with IpgE prevents the association of IpgD with IpaA in the cytoplasm. As neither IpaA nor IpaD are associated with IpgC (Page et al., 1999) and genes encoding a protein with features characteristic of a chaperone are not present in the vicinity of the ipaA and ipaD genes, storage and secretion of IpaA and IpaD might not require a specific chaperone.
Previous analysis of the ipgD mutant indicated that IpgD is not absolutely required for the entry of S. flexneri into HeLa cells (Allaoui et al., 1993a). The demonstration that IpgD and IpaA form a complex after being secreted prompted us to investigate whether IpgD, like IpaA, was involved in the modulation of actin rearrangements induced during the formation of an entry focus. After contact with epithelial cells, bacteria induce the formation of membrane extensions that rise around them and fuse into ruffle-like structures (Adam et al., 1995). An ipaA mutant is still able to enter into epithelial cells, although less efficiently than the wild type, and triggers entry structures that differ from those induced by the wild type: they are more diffuse and do not contain the focal adhesion-like structure beneath attached bacteria (Tran Van Nhieu et al., 1997). Molecular analysis revealed that IpaA interacts with the cytoskeletal protein vinculin and that this complex leads to the depolymerization of actin filaments (Tran Van Nhieu et al., 1997; Bourdet-Sicard et al., 1999). Immunofluorescence and scanning electron microscopy (SEM) analysis of the entry structures induced by the ipgD and ipgE mutants allowed us to detect a phenotype resulting from the absence of IpgD. Whereas the wild-type strain triggered massive actin-rich, membranous cell surface rearrangements around its attachment sites, the ipgD and ipgE mutants induced actin rearrangements that were mostly focused around adherent bacteria, where dense actin cups were observed. These cups contained other components found in the entry cups of the wild type, such as vinculin, α-actinin, paxillin and talin (Watarai et al., 1996; Tran Van Nhieu et al., 1997; K. Niebuhr, unpublished results). SEM analysis revealed that slight membrane ruffling was sometimes observed when several bacteria attached to the cells in close vicinity. These results suggest that IpgD is involved in the modulation of the cell response upon bacterial entry, even though it is not absolutely required for entry of bacteria into cultured epithelial cells.
A protein homologous to IpgD, SopB, has been identified in Salmonella dublin (Galyov et al., 1997). A sopB mutant was as invasive as the wild-type strain, and the protein seems to be implicated in the inflammatory response and induction of fluid secretion in bovine ligated ileal loops. SopB was shown to be translocated into eukaryotic cells (Galyov et al., 1997), and the purified protein exhibited phosphatidylinositol phosphatase activity (Norris et al., 1998). We have found a similar enzymatic activity for IpgD in vitro and in infected cells (unpublished data), and this suggests that IpgD is also translocated into epithelial cells and endowed with a phosphatase activity. Further studies will be needed to determine which host cell signalling pathways IpgD interferes with by dephosphorylating phosphatidylinositides, which play a critical role in many processes in the eukaryotic cell. The phenotype of the ipgD mutant, which showed dense actin cups around attached bacteria, was the opposite of that observed for the ipaA mutant, which only showed disorganized filopodial protrusions around bacterial attachment sites (Tran Van Nhieu et al., 1997). This suggests that IpaA and IpgD act in concert to co-ordinate and control the membrane and cytoskeletal rearrangements induced at an early stage of the interaction between bacteria and epithelial cells upon entry of S. flexneri.
Bacterial strains and growth media
S. flexneri strains are derivatives of the wild-type strain M90T (serotype 5) (Sansonetti et al., 1982). Strains M90T-Sm (SmR), SF701 (ipgD) and SF401 (mxiD) have been described previously (Allaoui et al., 1992; 1993a,b). E. coli strains are derivatives of E. coli K-12: DH5α was used for plasmids carrying an oriT origin of replication; DH5α-λpir was used for derivatives of the suicide vector pGP704 (Miller and Mekalanos, 1988); and SM10-λpir was used to transfer derivatives of pGP704 to S. flexneri. Bacteria were grown in Luria–Bertani (LB) medium or tryptic caseine soy broth (TCSB) at 37°C. Antibiotics were used at the following concentrations: ampicillin, 100 µg ml−1; kanamycin, 50 µg ml−1; streptomycin, 100 µg ml−1; and tetracycline, 10 µg ml−1.
Construction of recombinant plasmids expressing His6 and GST hybrid proteins
The ipgD gene was amplified by polymerase chain reaction (PCR) with oligonucleotides IpgD1 (5′-CCACGCGGATCCATGCACATAACTAATTTGGGA-3′) and IpgD2 (5′-CCACCGGAATTCTTATACAAATGACGAATACCC-3′). ipgE was amplified using primers IpgE1 (5′-CCACCGGTCGACTTAATACCCCTTCATTCTTCG-3′) and IpgE3 (5′-CCACGCGGATCCATGGAAGATTTAGCAGATGTT-3′). A fragment containing both ipgD and ipgE was amplified using primers IpgD1 and IpgE1. Forward and reverse primers contained BamHI and SalI or EcoRI sites, respectively, to allow cloning of PCR products between the BamHI and SalI or EcoRI sites of the vectors. The vector pQE30 (Qiagen) was used to construct plasmids pKNE6 (His6–IpgD and IpgE) and pKNE8 (His6–IpgE), and the vector pGEX4T-2 (Pharmacia) was used to construct pKNE10 (GST–IpgD) and pKNE4 (GST–IpgD and IpgE). Recombinant plasmids derived from pQE30 were obtained by transformation of a derivative of DH5α that harboured plasmid pMM 100. This latter plasmid carried the lacI gene, which allowed control of the production of the recombinant proteins that were expressed under the lac promoter of the vectors.
Construction of the ipgE mutant SB21
The ipgE mutant was constructed by allelic exchange of the wild-type ipgE gene carried by the pWR100 virulence plasmid with a mutated ipgE gene carried by a non-replicative plasmid as described previously (Allaoui et al., 1992 ; 1993a,b). Plasmid pAB4 (Fig. 1) was constructed by cloning the 2300 bp HindIII DNA fragment of PHS5103 (Baudry et al., 1987) into the HindIII site of the vector pUC8. In pAB4, the ipgE gene is located in proper orientation dowstream of the lac promoter of the vector. Plasmid pMP1 was constructed by replacing the 200 bp ClaI–BglII fragment of pAB4, internal to ipgE, by the 850 bp EcoRI–HincII fragment of pUC18K, carrying an aphA-3 cassette that confers resistance to kanamycin (Kn) (Ménard et al., 1993). In this construct, the aphA-3 gene was in the same orientation as ipgE, and insertion of the cassette introduced a translational stop at codon 19 of ipgE and a translational start codon in frame with the last 36 codons of ipgE, thereby allowing a potential translational coupling with the downstream ipgF gene. The mutator plasmid pMPD4 was constructed by cloning the PvuII–SalI fragment of pMP1, encompassing the mutated ipgE gene and its flanking regions, between the EcoRV–SalI sites of the suicide vector pGP704 that confers resistance to ampicillin (Miller and Mekalanos, 1988). Plasmid pMPD4 was transferred to S. flexneri M90T-Sm by conjugal mating, with selection for transconjugants on plates containing Sm and Km. Clones in which a double recombinational event had exchanged the wild-type ipgE gene for the mutated copy carried by pMPD4 were screened for their sensitivity to Amp. The structure of the derivative of pWR100 carrying the ipgE mutation was confirmed by PCR and Southern blot analyses, and the resulting strain was designated SB21.
Preparation of mAbs against IpgD
Balb/C mice were immunized with a mixture of proteins secreted by the wild-type strain M90T after activation of the type III secretion machinery. Briefly, bacteria in the exponential phase of growth were harvested by centrifugation, resuspended in PBS and incubated in the presence of Congo red for 30 min at 37°C. Samples were centrifuged to eliminate bacteria, and the antigen mixture was prepared by passing the supernatant over a HR-200 column (Pharmacia) to remove the dye. The immunization and lymphocyte–myeloma fusion procedure for the generation of mAbs have been described in detail previously (Niebuhr et al., 1998). The resulting hybridoma supernatants were screened by enzyme-linked immunosorbent assay (ELISA) and Western blotting on the antigen mixture. Hybridoma cells were subcloned by limiting dilution, and the antigen recognized by the resulting mAbs were identified by testing each mAb on culture supernatants of different S. flexneri mutants and on lysates of derivatives of E. coli harbouring recombinant plasmids. This procedure led to the isolation of the IpgD-specific mAb 20D9.
Preparation of protein extracts, SDS–PAGE and immunoblotting
Bacteria in the exponential phase of growth were harvested by centrifugation, and lysates were prepared by resuspending pellets in SDS–sample buffer. Whole-cell lysates corresponding to 3 × 107 bacteria were analysed by SDS–PAGE. For the analysis of secreted proteins, bacterial pellets were resuspended in a 10-fold smaller volume of PBS, and Congo red was added to a final concentration of 1 mg ml−1. Bacterial suspensions were incubated for 30 min at 37°C to induce secretion and centrifuged to remove bacteria. Supernatant samples corresponding to proteins secreted by 108 bacteria were analysed by SDS–PAGE. Immunoblotting was carried out with the anti-IpgD mAb 20D9 (this work), the anti-IpaB mAb H16 (Barzu et al., 1993), the anti-IpaC mAb J22 (Phalipon et al., 1992), an anti-His tag mAb (Clontech) and polyclonal antibodies against IpaA (Tran Van Nhieu et al., 1997). ECL substrate (Amersham) was used for subsequent chemoluminescence detection. For N-terminal sequence determination, proteins were transferred onto a polyvinylidene difluoride (PVDF) membrane, stained with amido black and subjected to Edman degradation. For the examination of protein stability, 30 µg ml−1 chloramphenicol was added to bacterial cultures grown to the logarithmic phase. Samples were taken after certain time points, and cell lysates were analysed by Western blotting.
Purification of His6 and GST hybrid proteins
Production and purification of hybrid proteins containing a His6 tag or GST at their N-terminal extremity were performed according to the instructions given by Qiagen (His6) and Pharmacia (GST). Briefly, bacteria were grown in 50 ml of TCSB at 37°C until the optical density (OD) of the culture reached values of 0.8–1.0, and IPTG was added to a final concentration of 1 mM. After 2 h of induction at 37°C, bacteria were harvested by centrifugation at 1500 g for 10 min. Pellets were resuspended in 5 ml of buffer A (50 mM NaH2PO4, 10 mM Tris, 100 mM NaCl, pH 8.0) supplemented with 1% Triton X-100 and protease inhibitors (Boehringer), sonicated three times for 20 s, and cell debris was removed by centrifugation at 1500 g for 10 min. To prepare cleared lysates, supernatants were subsequently centrifuged at 400 000 g for 30 min. For the preparation of secreted proteins, bacteria were resuspended in PBS containing 1 mg ml−1 Congo red and incubated at 37°C for 30 min. Bacteria were eliminated by centrifugation, and supernatants were supplemented by the addition of 1% Triton X-100 and protease inhibitors and centrifuged at 15 000 g for 20 min to remove aggregated proteins. Cleared lysates and supernatants were mixed with 500 µl of Talon cobalt–sepharose resin (Clontech) or glutathione–sepharose (Pharmacia) and incubated for 1–2 h at 4°C with gentle agitation. Beads were washed three times with 10 ml of PBS containing 0.5% Triton X-100 and, in the case of His6 fusion proteins, 10 mM imidazole. Proteins bound to the resins were either resuspended in Laemmli sample buffer or eluted with PBS containing 0.5% Triton X-100 and 100 mM imidazole (His6) or with 100 mM Tris, pH 8.0, containing 50 mM glutathione and 100 mM NaCl (GST).
HeLa cell infection and microscopy analysis
HeLa cells were grown in minimum essential medium (MEM; Gibco BRL) supplemented with 10% fetal calf serum, glutamine and non-essential amino acids in the absence of antibiotics. For infection experiments, cells were grown on 12 mm coverslips in six-well tissue culture dishes and placed in serum-free medium 2 h before infection. Bacteria in the exponential growth phase were harvested by centrifugation and resuspended in an equal volume of serum-free MEM, and 2 ml of bacterial suspensions, corresponding to approximately 3 × 108 bacteria, were added to each well. After centrifugation at 2000 g for 10 min at 20°C, dishes were incubated at 37°C for 15 min. Coverslips were rinsed three times with PBS and, for immunofluorescence labelling, cells were fixed in 4% paraformaldehyde in PBS and permeabilized with 0.2% Triton X-100 in PBS. Bacteria were labelled with an anti-LPS antiserum at a dilution of 1:200 and an anti-rabbit secondary antibody coupled to Cy5; vinculin was labelled with mAb hVin1 (Sigma) diluted 1:50 and an anti-mouse secondary antibody coupled to Cy3; and F-actin was visualized by fluorescein isothiocyanate (FITC)-coupled phalloidin. Samples were analysed by confocal microscopy with a Zeiss LSM 510. For SEM, samples were fixed with 1.6% glutaraldehyde, 1% tannic acid in 0.1 M phosphate buffer, pH 7.4, rinsed twice for 5 min in the same buffer complemented with 3.7% sucrose and post-fixed in 1% osmium in 0.1 M cacodylate buffer. After one rinse in the same buffer, one rinse in distilled water and one 5 min wash in 50% ethanol, cells were stained for 30 min in 50% (v/v) uranyl acetate, dehydrated in a graded series of ethanol and treated for 10 min with hexamethyldisilane (Bray et al., 1993). After drying, coverslips were mounted on pins, coated with gold using a Polaron E5100 SEM coating unit and examined on a JEOL JSM-35CF scanning electron microscope.
The authors wish to thank Drs Maria Mavris and Frank Ebel for helpful comments and critical reading of the manuscript, Armelle Phalipon for the gift of serum samples from human shigellosis patients, and the Hybridolab of the Institute Pasteur for providing the facilities to generate monoclonal antibodies. The confocal microscope was purchased with a donation from Marcel and Liliane Pollack. This work was supported in part by grants from the GIP-HMR (Hoechst-Marion-Roussel), the Belgian FRSM (Fonds National de la Recherche Scientifique Médicale, convention 3.4611.99) and the Actions de Recherche Concertées (ARC; convention 98/03-224) de la Direction Générale de La Recherche Scientifique-Communité Française de Belgique. K.N. was supported by an EMBO long-term postdoctoral fellowship, and N.J. is a recipient of a fellowship from the ARC.