Shigella protein IpaH(9.8) is secreted from bacteria within mammalian cells and transported to the nucleus.

Various pathogenic bacteria such as Shigella deliver effector proteins into mammalian cells via the type III secretion system. The delivered Shigella effectors have been shown to variously affect host functions required for efficient bacterial internalization into the cells. In the present study, we investigated the IpaH proteins for their ability to be secreted via the type III secretion system and their fate in mammalian cells. Upon incubation in a medium containing Congo red, the bacteria secrete IpaH into the medium, but secretion of IpaH occurs later than that of IpaBCD. Immunofluorescence microscopy indicated that IpaH(9.8) is secreted from intracellular bacteria and transported into the nucleus. On microinjection of the protein, intracellular IpaH(9.8) is accumulated at one place around the nucleus and transported into the nucleus. This movement seems to be dependent on the microtubule network, since nuclear accumulation of IpaH(9.8) is inhibited in cells treated with microtubule-destabilizing agents. In nuclear import assay, IpaH(9.8) was efficiently transported into the nucleus, which was completely blocked by treatment with wheat germ agglutinin. The nuclear transport of IpaH(9.8) does not depend on host cytosolic factors but is partially dependent on ATP/GTP, suggesting that, like beta-catenin, IpaH(9.8) secreted from intracellular Shigella can be transported into the nucleus.

Various Gram-negative pathogenic bacteria possess a type III secretion system through which they deliver a set of proteins into the environment as well as into the target host cells required for infection (1,2). Although the mechanisms underlying the secretion of these proteins remain to be elucidated, the secreted proteins seem to be variously utilized for the infection process by affecting the target host functions. For example, Salmonella species deliver an array of effectors such as SipA, SipB, SipC, SopB/SigD, SopE, SptP, and SspH1 into the host cells during infection (3)(4)(5). These molecules modulate the host cellular signaling pathways and rearrange the cytoskeleton required for invasion of host cells. Enteropathogenic Escherichia coli and enterohemorrhagic E. coli also secrete effectors such as EspA, EspB, EspD, and Tir required for intimate attachment to the host cell surface (6 -9). Yersinia spp. delivers many effectors such as YopB, YopD, YopE, YopH, YopJ/P, YopM, YopN, YopT, or YpkA, some of which are used to evade killing by macrophages or for modulation of host immune responses (10 -12).
Upon contact of Shigella flexneri with epithelial cells, the bacteria quickly release proteins such as IpaA, IpaB, IpaC, IpgD, and VirA into the medium, and they seem to be injected into the host cells via the type III secretion machinery during invasion (13)(14)(15)(16)(17)(18)(19)(20). For example, IpaA injected into the host cytoplasm interacts with vinculin and somehow promotes depolymerization of F-actin accumulated around the bacterial entry site, which is required for efficient entry into the host cells (17,19). IpaC is mostly associated with the host plasma membrane as part of the type III secretion machinery, by which IpaC somehow stimulates Cdc42 activity, thus promoting protrusion of lamellipodia and filopodia around the site of bacterial entry (18). Furthermore, S. flexneri defective in ipaBCD expression, which causes deregulation of the type III secretion system, was shown to secrete an additional set of proteins including IpaH 9.8 , IpaH 7.8 , IpaH 4.5 , IpgB1, MxiC, MxiL, OspC1, OspB, OspD1, OspG and OspE1, and Spa32 (21). Although the question of whether each of these is also delivered into the host cells during Shigella infection of epithelial cells is still not clear, they suggested that Shigella delivers some of these proteins into host cells (21). Thus, the putative delivered proteins might perform important roles in bacterial infection or provide some benefit for the pathogen in the infected host cells.
Previous studies have indicated that five copies of the ipaH genes exist on the large plasmid (pWR100) of S. flexneri 5 (M90T), and these ipaH genes have seen cloned and sequenced (22,24,25). These were designated by the size of the HindIII fragment of pWR100 (e.g. IpaH 9.8 is encoded on a 9.8-kilobase fragment). All five copies have almost identical C-terminal halves. Although the N-terminal portions of ipaH differ in each gene, they all included a common leucine-rich repeat (LRR) 1 motif. Interestingly, LRR-containing proteins have also been found in a diverse group of bacteria and eukaryotes, although the biological significance of the LRR motif in each protein still largely remains speculative (26). Interestingly, the production of the IpaH proteins occurred later than that of IpaBCD as determined by monitoring the gene expression using the lacZ transcriptional fusion system (22). Furthermore, it has recently been reported that when S. flexneri 2a (2457T) infected mouse macrophages (J774) or human monocyte-derived macrophages, IpaH 7.8 facilitated escape from the vacuole (23).
In the present study, we attempted to determine whether the IpaH proteins can be secreted from S. flexneri 2a (YSH6000) via the type III secretion system, and we further investigated the fate of the secreted IpaH protein, IpaH 9.8 , in the host cells. Our results indicated that IpaH 9.8 , IpaH 7.8 , and IpaH 4.5 can be secreted through the type III secretion system but that this secretion occurred at a later stage. Importantly, the secretion from S. flexneri seemed to be stimulated primarily within the host cell cytoplasm, where the secreted IpaH 9.8 can be efficiently transported into the host cell nucleus.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Eukaryotic Cell Lines, and Growth Conditions-Bacterial strains and plasmids used are listed in Table I. All S. flexneriderived strains were grown routinely in brain heart infusion broth (Difco) at 37°C. HeLa cells were cultured in minimal essential medium (Sigma) with 10% fetal calf serum (Nichirei) at 37°C in the presence of 5% CO 2 . COS-7 cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal calf serum at 37°C in the presence of 5% CO 2 .
Plasmid Construction-Plasmid pIpaH 9.8 was constructed by ligating the EcoRI-BamHI fragment of the ipaH 9.8 gene into the corresponding sites of pTB101. Plasmids pGST-IpaH x (x: 9.8, 7.8, or 4.5) were constructed by ligating the EcoRI-BamHI fragment of the ipaH x gene into pGEX-2T. Plasmid pGST-␤-catenin was constructed by ligating the SalI-BamHI fragment of the ␤-catenin gene into pGEX-6P-1. Plasmid pFLAG-IpaH 9.8 was constructed by cloning the FLAG-IpaH 9.8 gene, which has a FLAG tag and linker (MDYKDDDDKVDGIDKLDIEF) fused to the N terminus, into the pMEsf-neo vector. Plasmids pHis-IpaH x (x: 9.8, 7.8, or 4.5) were constructed as follows. The ipaH x genes were cloned into pQE30 (Qiagen). Resultant plasmids were digested with EcoRI and HindIII, followed by ligation of the EcoRI-HindIII fragment containing the His-tagged ipaH x gene into pTB101.
Expression and Purification of Recombinant Proteins-Recombinant IpaH proteins were purified as follows. E. coli cells carrying pGST-IpaH x (x: 9.8, 7.8, or 4.5) were cultivated in L broth supplemented with ampicillin (50 g ml Ϫ1 ) for 3 h at 37°C. Expression was induced by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside and incubation for 2 h at 37°C. Bacteria were disrupted by sonication using an ultrasonic disruptor UD-200 (TOMY) for 1 min, 4 times with incubation on ice. Purification of the GST fusion proteins with glutathione-Sepharose 4B (Amersham Pharmacia Biotech) and cleavage of the GST proteins with thrombin were performed according to the manufacturer's protocol.
Recombinant ␤-catenin was purified as follows. E. coli cells carrying pGST-␤-catenin were cultivated in L broth supplemented with ampicillin (50 g ml Ϫ1 ) for 3 h at 37°C. Expression was induced by the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside and incubation for 2 h at 37°C. Bacteria were disrupted by sonication. Purification of the GST fusion protein with glutathione-Sepharose 4B and cleavage of the GST protein with PreScission Protease (Amersham Pharmacia Biotech) were performed according to the manufacturer's protocol.
Detection of Ipa Proteins in the Whole-cell Lysates and Culture Supernatants-Shigella cells cultivated in 5 ml of brain heart infusion broth at 37°C for 4 h were washed in ice-cold phosphate-buffered saline (PBS) and resuspended in 2 ml of PBS. After incubation at 37°C for 5 min, 6 l of 1% Congo red (CR) was added to the bacterial suspension followed by incubation for 10 min at 37°C. After centrifugation, the supernatant was passed though a 0.45-m pore size filter, and the proteins in the resultant supernatant and bacterial pellet were precipitated with trichloroacetic acid. Samples were separated by SDS-PAGE and immunoblotted with appropriate antibodies.
Collection of Ipa proteins at each time point was performed as follows. Shigella cells were cultivated in 2 ml of tryptic soy broth at 37°C for 3 h. After the addition of 10 l of 1% CR to the bacterial culture, the bacterial cells were incubated at 37°C for 0, 30, 60, 120, 180, or 240 min. Each sample was centrifuged, and the proteins in the resultant supernatant and bacterial pellet were precipitated with trichloroacetic acid. Each sample was separated by SDS-PAGE and immunoblotted with appropriate antibodies.
Infection of Cultured Cells and Immunofluorescence Microscopy-Cells were grown on coverslips to ϳ70% confluency in antibiotic-free medium. Cells were infected with Shigella at a multiplicity of infection of about 300 per cell. The plates were centrifuged at 900 ϫ g for 10 min after adding bacteria. After incubation for 15 min at 37°C in the presence of 5% CO 2 , the plates were washed three times with Hanks' balanced salt solution, into which fresh medium supplemented with 100 g ml Ϫ1 gentamicin, 60 g ml Ϫ1 kanamycin, and 100 M isopropyl-1thio-␤-D-galactopyranoside was added. The infected cells were incubated for an additional 4 h at 37°C in the presence of 5% CO 2 . Colchicine (1 g ml Ϫ1 ; Sigma) pretreatment of cells was performed for 30 min at 37°C in 5% CO 2 before infection. Nocodazole (10 g ml Ϫ1 ; Sigma) pretreatment was performed on ice for 1 h followed by incubation for 30 min at 37°C in 5% CO 2 before infection. Cytochalasin D (Sigma) was added to the fresh medium to a final concentration of 1 M after the medium change. After incubation, the cells were washed twice with Hanks' balanced salt solution and fixed with 4% paraformaldehyde in PBS for 20 min. For immunofluorescence studies, the coverslips with HeLa cell monolayers were incubated in 50 mM NH 4 Cl in PBS for 10 min, and the permeabilization of cells was carried out in 0.2% Triton X-100 in PBS for 20 min. After blocking for 30 min in 2% bovine serum albumin, the coverslips were incubated with primary antibodies in Tris-buffered saline for 1 h at 37°C. After washing, the samples were incubated with FITC-conjugated secondary antibodies for 1 h at 37°C. After washing, the samples were incubated with TO-PRO3 iodide (Molecular Probes) used to visualize host cell nuclei and infected bacteria for 30 min at 37°C. The coverslips were mounted in VECTASHIELD (Vector Laboratories) and observed with a confocal laser-scanning microscope (MicroRadiance Plus; Bio-Rad). Preparation of Cy3-labeled Proteins and Nuclear Localization Signal (NLS)-conjugated Allophycocyanin-Cy3-labeled IpaH 9.8 and ␤-catenin were prepared according to the manufacturer's protocol (FluoroLink-Ab Cy3 labeling kit; Amersham Pharmacia Biotech). Allophycocyanin (APC; Calbiochem) was conjugated with synthetic peptide (CYGGP-KKKRKVEDP) containing the SV40 large T antigen NLS as described previously (32).
Microinjection-Cy3-labeled IpaH 9.8 protein (about 1.5 mg ml Ϫ1 ) was injected through a glass capillary using a micromanipulator (model 5171, Eppendorf) into the cytoplasm of host cells grown on coverslips. The microscope stage was maintained at 37°C and 5% CO 2 during the time of injection and data acquisition. Nocodazole pretreatment was performed as described above.
In Vitro Nuclear Import Assay-HeLa cells were grown on coverslips to ϳ70% confluency in the absence of antibiotics. Cells were permeabilized in ice-cold transport buffer (20 mM HEPES, pH 7.3, 110 mM potassium acetate, 2 mM magnesium acetate, 5 mM sodium acetate, 0.5 mM EGTA, 2 mM dithiothreitol, 1 g ml Ϫ1 aprotinin, leupeptin, and pepstatin) containing 40 g ml Ϫ1 digitonin (Research Biochemicals Incorporated) for 5 min on ice. After digitonin permeabilization, ATPdepleted samples were pre-treated with transport buffer containing 0.1 units ml Ϫ1 apyrase (Sigma) and 2% bovine serum albumin for 5 min at 30°C. For wheat germ agglutinin (WGA; Sigma) treatment, permeabilized cells were incubated with 0.5 mg ml Ϫ1 WGA for 5 min on ice. After washing the cells with cold transport buffer, the coverslips were blotted and inverted over 10 l of test solution on a Parafilm sheet in a humidified box. The compositions of each test solution are indicated in the respective figure legends. The import reaction was performed for 30 min at 30°C or 4°C. For ϩATP conditions, 1 mM ATP, 0.5 mM GTP, 5 mM phosphocreatine (Sigma), 20 units ml Ϫ1 creatine phosphokinase (Sigma) were present during incubation. For ϩcytosol conditions, 20 mg ml Ϫ1 rabbit reticulocyte lysate (Promega) was present during the incubation. After rinsing the cells with ice-cold transport buffer, the cells were fixed with 4% paraformaldehyde in PBS.

Structures of the ipaH Gene Family on the 230-kilobase
Plasmid of S. flexneri YSH6000 -The presence of three ipaH genes, ipaH 4.5 , ipaH 7.8 , ipaH 9.8 , on the large plasmid (pMYSH6000) of S. flexneri 2a YSH6000 was investigated by polymerase chain reaction cloning. The nucleotide sequences of the polymerase chain reaction-amplified ipaH genes indicated that all except ipaH 7.8 were almost identical to those of pWR100 (22,24,25). Our sequence data of ipaH 7.8 from pMYSH6000 contained one additional nucleotide at position 26 from the 5Ј end of ipaH 7.8 of pWR100, resulting in a 98-nucleotide extension from the 5Ј end of the ipaH 7.8 open reading frame (Fig. 1A). Consequently, the predicted N-terminal IpaH 7.8 sequence encoded by the ipaH 7.8 of pMYSH6000 possessed 33 additional amino acids at the N terminus of IpaH 7.8 on pWR100 (Fig. 1A). Accordingly, the deduced IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 sequences of pMYSH6000 were composed of four distinctive domains: (i) the N-terminal 60ϳ70-amino acid stretch; (ii) the following 200ϳ355 amino acid region containing 6 -9 LRRs; (iii) the intervening sequence bracketed by LRRs and the C-terminal conserved region; and (iv) the C-terminal conserved region (CTR) (Fig. 1B). Although the length of CTRs varied slightly between members of the IpaH family, they showed significant similarity with other cognates such as in the C-terminal regions of SlrP of Salmonella and y4fR of Rhizobium (33,34).
Type III-dependent Secretion of IpaH-Upon incubation of Shigella in PBS containing 0.003% CR, the bacteria rapidly deliver IpaA, IpaB, IpaC, IpaD, VirA, and IpgD into the medium via the type III secretion system (35). Shigella infectivity has been shown to be dependent on the type III secretion activity (27,36,37). To test whether S. flexneri secretes IpaH proteins into the medium via the secretion system, the bacteria were incubated in PBS with or without CR. Although the secretion of IpaB, IpaC, and IpaD could be readily detected on incubation in PBS with CR for 10 min (Fig. 2A, lane 1), IpaH secretion was not detected (Fig. 2A, lane 1). However, after long exposure of the X-ray film, a trace amount of IpaH secretion was detected as a 60-kDa band (data not shown). To confirm the potential ability of IpaH to be secreted via the type III secretion system, we introduced a cloned ipaH 9.8 plasmid (pIpaH 9.8 ) into YSH6000 or S325 (a type III-deficient mutant of YSH6000), and IpaH secretion into the medium containing CR was investigated by immunoblotting. As shown in Fig. 2A, secretion of IpaH 9.8 from YSH6000 but not from S325 was detected as a band of ϳ60 kDa.
To further investigate the ability of Shigella to secrete IpaH into the conditional medium via the type III secretion system, each IpaH protein was tagged with 6 histidines (His) (see "Experimental Procedures"). YSH6000 overexpressing each of the His-tagged IpaH proteins were examined for their ability to be secreted into the medium when incubated in PBS with or without CR. IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 were secreted from YSH6000 when incubated in PBS with CR (Fig. 2B), suggesting that at least these IpaH proteins can be secreted through the type III secretion system.
Delayed IpaH Secretion from Shigella in the Presence of Congo Red during Growth-It has been suggested that induction of ipaH expression in Shigella during growth in medium containing CR is markedly increased as determined by ipaH-lacZ fusion, whereas the expression of ipaBCDA under the same conditions occurred constitutively (22). Therefore, we investigated the kinetics of production within or secretion from YSH6000 of IpaH together with IpaBCD during growth for 4 h in tryptic soy broth in the presence or absence of CR. As shown in Fig. 3A, IpaBCD secretion from YSH6000 was detected as early as 30 min in medium with CR, whereas IpaH secretion was detected after a 2-h growth in medium with CR. Although IpaBCD production in YSH6000 could be detected readily during growth even in the absence of CR, IpaH production in bacteria was hardly detected during growth in the presence or absence of CR. In this experiment, none of the IpaH proteins including IpaBCD was secreted from S325 (a type III-deficient mutant of YSH6000), and S325 did not show production of IpaH proteins even with incubation in medium containing CR for 4 h (data not shown). On immunoblotting, two major protein bands were detected with the anti-IpaH antibody. To identify the secreted IpaH proteins corresponding to the 60-and 62-kDa protein bands, each IpaH protein was purified and compared with the sizes of the IpaH proteins secreted from YSH6000. As shown in Fig. 3B, the upper band corresponded to IpaH 4.5 , whereas the lower band corresponded to IpaH 7.8 and IpaH 9.8 . Also, the identity of the bands was confirmed by peptide mass fingerprinting using matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy. The upper band included IpaH 4.5 , whereas the lower band included IpaH 9.8 (data not shown). Due to unknown technical problems, fingerprinting for IpaH 7.8 was not detected in the lower band. These observations suggested that IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 can be secreted via the type III secretion system, albeit at a late stage, in medium containing CR.
Intracellular Localization of Secreted IpaH-The present results and previous another study (22) led us to speculate that IpaH secretion would take place after bacterial entry into the host cells. To clarify the fate of IpaH secreted from Shigella during infection of epithelial cells, we performed indirect immunofluorescence laser-scanning confocal microscopy of HeLa cells infected with YSH6000 carrying pIpaH 9.8 . IpaC served as a reference type III-secreted protein. When the bacteria were allowed to come into contact with HeLa cells pretreated with cytochalasin D, thus preventing bacterial internalization, IpaH secreted from bacteria into the epithelial cells was not detected by 4 h of incubation (data not shown). In contrast, after infection of HeLa cells with YSH6000 carrying pIpaH 9.8 , internalized bacteria were detected by phase-contrast microscopy and multiplied within the cytoplasm (Fig. 4A; a and e). The internalized bacteria in the cytoplasm as well as the host cellular nucleus were both visualized by staining with TO-PRO3 (c and g). Although IpaH 9.8 visualized by the FITC signal was still visible in the cytoplasm, IpaH 9.8 was predominantly concentrated within the nucleus (b). Indeed, the FITC signal was mostly colocalized with the TO-PRO3 signal (d). In contrast, no IpaC signal was detected within the nucleus, but it was distributed evenly in the cytoplasm (Fig. 4A, lower panels). The

FIG. 3. Delayed IpaH secretion from S. flexneri and identification of the secreted IpaH proteins.
A, production and secretion of Ipa proteins after the addition of CR. After incubation for 3 h in tryptic soy broth, S. flexneri wild-type strain (YSH6000) was continuously grown in the presence (CRϩ) or absence (CRϪ) of CR. Samples were taken at the indicated time points, and similar amounts of bacterial lysates (ppt.) and supernatants (sup.) were subjected to SDS-PAGE followed by immunoblotting with antibodies specific for Ipa proteins. B, identification of IpaH secreted from YSH6000. Recombinant IpaH proteins were purified as described under "Experimental Procedures." WT, wild type. nuclear localization of IpaH was also reproducible when Caco-2 or COS-7 cells were infected with YSH6000 carrying pIpaH 9.8 (data not shown).
To further confirm the fate of intracellular IpaH 9.8 , COS-7 cells expressing FLAG-tagged IpaH 9.8 were constructed, and the intracellular distribution of IpaH 9.8 was investigated by immunofluorescence confocal microscopy with anti-FLAG-M5 antibody. As shown in Fig. 4B, FLAG-IpaH 9.8 was localized in the nucleus. Furthermore, Cy3-labeled IpaH 9.8 together with FITC-labeled IgG were injected into COS-7 cells, and the cells were examined after a 30-min incubation. As shown in Fig. 4C, IpaH 9.8 was distributed within the whole cell including the nucleus (left panel), whereas IgG was located only in the cytoplasm (right panel). The distribution was also confirmed by immunohistological method using unlabeled IpaH 9.8 protein (data not shown). These results strongly suggested that IpaH secretion occurred after internalization of Shigella into host cells, in which the secreted IpaH proteins were transported into the nucleus.
Intracellular Trafficking of IpaH-We postulated that the movement of IpaH toward the nucleus may be mediated through interaction with the intracellular trafficking system. Thus, we injected Cy3-labeled IpaH 9.8 into COS-7 cells, and the cells were observed at 5 and 30 min after injection with a microscope equipped with a cooled CCD camera. Although IpaH 9.8 -associated fluorescence varied among cells, the IpaH signal, which was initially dispersed within the cytoplasm at 0 min (data not shown), was concentrated around the periphery of the nucleus. In some case, the IpaH-associated fluorescence was accumulated at one place around the nucleus (Fig. 5A). This was not due to an artificial interaction caused by Cy3 modification, since the addition of an excess mount of unlabeled IpaH 9.8 competed with the accumulation of Cy3-labeled signal at one place on the nuclear membrane (without unlabeled protein, 95.9%; with 10-fold excess mount of unlabeled protein, 38.2%; with 100-fold mount of unlabeled protein, 20.0%). When COS-7 cells were pretreated with nocodazole, this phenomenon was not observed (Fig. 5A). Importantly IpaH 9.8 signal coincided with regions of high microtubule density containing ␥-tubulin, which indicates a microtubule-organizing center (Fig.  5B). About 80% of the cells showed a similar distribution. Furthermore, HeLa cells treated with various drugs known to affect cytoskeletal dynamics were infected with YSH6000 carrying pIpaH 9.8 . At 4 h post-infection, HeLa cells treated with or without cytochalasin D showed an accumulation of IpaH 9.8 within the nucleus as determined by immunofluorescence microscopy (Fig. 6A), suggesting that microfilaments are not in- volved in the movement. In contrast, in HeLa cells treated with nocodazole or colchicine IpaH 9.8 were distributed in the cytoplasm as well as the nucleus. We examined the structure of microtubules using a mixture of anti-␣-and -␤-tubulin antibodies. These drugs caused a loss of the normal microtubule structure, and the tubulin that was visible appeared to be aggregated (data not shown). Quantitative assay for the nuclear accumulation of IpaH 9.8 in HeLa cells treated with the drugs revealed that nuclear accumulation of IpaH was significantly decreased in cells treated with nocodazole and colchicine (Fig.  6B). These results strongly suggested that the microtubule network is involved in the intracellular trafficking of IpaH 9.8 toward the periphery of the nucleus and transport into nucleus facilitated.
Nuclear Transport of IpaH-Small proteins of less than 40ϳ60 kDa can enter the nucleus passively, albeit in a concentrationdependent manner, whereas some macromolecules larger than 40ϳ60 kDa are actively transported across the nuclear pore complex (38). The active nuclear import of proteins has been shown to be mediated by specific amino acid sequences, which are referred to as NLSs. NLS-mediated nuclear import requires soluble factors such as the importin family (39 -42). Although it seemed to have no NLS, IpaH 9.8 protein was efficiently transported into the nucleus, suggesting that IpaH acts as nuclear transported protein. Therefore, we used a digitonin-permeabilized HeLa cell transport assay and investigated the intracellular fate of exogenously added IpaH using immunofluorescence microscopy, since digitonin permeabilizes the cytoplasmic membrane but not the nuclear membrane (42). In this experiment, we investigated the fate of ␤-catenin and allophycocyanin conjugated to SV40 T-antigen NLS peptide (NLS-APC) as a control. Exogenously added ␤-catenin was rapidly imported into the nucleus in the absence of cell lysate without the addition of ATP/ GTP, whereas under these conditions NLS-APC cannot be imported (43). Our results showed that IpaH 9.8 migrated rapidly into the nucleus in the absence of exogenously added cytosol and ATP/GTP at 30°C (Fig. 7, panel a). The nuclear transport of IpaH 9.8 was inhibited as the concentration of cytosol added to the medium was increased (panel m). To investigate the requirement of ATP/GTP for nuclear import, permeabilized cells were pretreated with apyrase, since apyrase decreases the ATP level (43,  44). Upon the addition of apyrase, the nuclear transport of IpaH 9.8 was inhibited (panel j), although the extent of inhibition was less compared with that for NLS-APC, and the import of IpaH 9.8 into the nucleus was sensitive to temperature (panel d) and was inhibited by WGA (panel g), a specific inhibitor of nuclear transport across the nuclear pore complex (45). These results indicated that the nuclear transport of IpaH 9.8 is independent of cytosolic factors but dependent on temperature and partly on ATP/GTP. It is thus likely that the nuclear transport of intracellular behavior of IpaH 9.8 is similar to that of ␤-catenin.

DISCUSSION
In the present study, we investigated the secretion of IpaH proteins (IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 ) encoded by the large plasmid (pMYSH6000) of S. flexneri 2a (YSH6000) via the type III secretion system. Our results indicated that IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 were secreted via the type III secretion system when bacteria were incubated in PBS containing CR. These findings agreed with those of other studies (21), indicating that IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 can be secreted from S. flexneri 5 (M90T) into the medium under conditions of inactivation of the ipaBCDA genes (21). We found that the secretion of IpaH proteins from YSH6000 during growth in the medium stimulated the type III secretion system at a late stage such as after 2 h, whereas the secretion of the IpaBCD under the stimulated conditions occurred as early as after 30 min (see Fig. 3A). The delayed secretion phenotype displayed by IpaH proteins from Shigella in activation of the type III secretion system may be reflected by the differences in the transcriptional control system from that required for ipaBCDA genes, as reported by Demers et al. (22). Indeed, they showed that the levels of transcription of ipaH genes, but not ipaBCDA genes, were markedly increased during growth in the presence of CR or in the ⌬ipaBCDA mutant, conditions that enhanced type III secretion (22). Although the mechanism underlying the delayed IpaH expression remains to be elucidated, the delayed IpaH secretion suggested that IpaH plays different roles from IpaBCD in Shigella infection.
The assignment of the open reading frame for the ipaH 7.8 gene of YSH6000 in the present study was different from that reported previously for the M90T strain (24), since an addi-tional nucleotide was found at position 26 from the 5Ј end of ipaH 7.8 of M90T (see Fig. 1A). The discrepancy might have been due to an error in sequencing of M90T ipaH 7.8 , since Buchrieser et al. (21) very recently reported that the N-terminal amino acid sequence of the secreted IpaH 7.8 from M90T was consistent with that of our predicted IpaH 7.8 sequence of YSH6000. Therefore, the extended N-terminal 33-amino acid residues of IpaH 7.8 also showed similarities with those of IpaH 4.5 and IpaH 9.8 (see Fig. 1C). Although the precise roles of the Nterminal sequences of IpaH 4.5 , IpaH 7.8 , and IpaH 9.8 are still to be investigated, the N-terminal sequence may contain amino acid sequences required for secretion via the type III secretion system.
Our observations suggested that the secretion of IpaH from Shigella occurs after invasion of epithelial cells. When YSH6000 carrying pIpaH 9.8 (a cloned ipaH 9.8 plasmid) was infected into HeLa cells pretreated with cytochalasin D, IpaH 9.8 was not detected in the cells by indirect immunofluorescence staining with anti-IpaH antibody. 2 In contrast, when HeLa cells were incubated with the bacteria for 4 h without treatment, Shigella were internalized into the HeLa cells, and secreted IpaH was detected within the nucleus, although a small amount of IpaH was present in the cytoplasm. Since the IpaC secreted from the intracellular Shigella at 4 h post-infection was observed only within the cytoplasm and not in the nucleus, we concluded that IpaH secretion would be stimulated after bacterial entry into the host cytoplasm and that IpaH 9.8 can be transported into the nucleus. Similarly, YopM of Yersinia pestis was previously reported to be transported into the nucleus (46). YopM also possesses 15 repeats of LRR, showing similarity with the N-terminal portion of IpaH, but lacks the CTR, which exists in all IpaH family proteins including other homologous proteins such as SlrP of Salmonella typhimurium or y4fR of Rhizobium (see Fig. 1A). Although whether SlrP and y4fR would also be translocated into the host nucleus remains to be elucidated, based on the primary structural features and nuclear transport behavior of IpaH 9.8 and YopM, the N-terminal portion of IpaH proteins containing the LRR is probably involved in nuclear transport.
Since IpaH 9.8 secreted from intracellular Shigella at 4 h post-infection appeared to be accumulated in the nucleus, we further investigated the fate of IpaH 9.8 including intracellular trafficking by different approaches. In COS-7 transfectants expressing FLAG-tagged IpaH 9.8 , the IpaH signal was mostly detected in the nucleus. Furthermore, the microinjection of Cy3-labeled IpaH 9.8 together with FITC-IgG into COS-7 cells revealed that although the FITC fluorescence signal was detected only in the cytoplasm at 30 min after injection, Cy3 fluorescence was detected in both the cytoplasm and nucleus, implying that IpaH 9.8 acts as a mammalian nuclear transport protein.
In these experiments, we noted that in some cells the cytoplasmic Cy3 signal, which was associated with IpaH 9.8 , was concentrated around the microtubule-organizing center. However, the accumulation was not predominant in cells pretreated with nocodazole, a microtubule-destabilizing agent, suggesting that intracellular IpaH 9.8 is accumulated in the vicinity of the nuclear surface through association with the microtubule network(see Fig. 8). This was also suggested by investigation of the fate of IpaH 9.8 secreted from intracellular Shigella, in which accumulation of the secreted IpaH 9.8 within the nucleus was almost completely blocked in cells pretreated with nocodazole or colchicine. Although the precise mechanism underlying the concentration of IpaH 9.8 remain to be investi- gated, as indicated for YopM of Y. pestis (46), the intracellular trafficking system mediated by microtubule networks would be important for the condensation of IpaH around the microtubule-organizing center. Trafficking along the microtubules in mammalian cells is mediated by various motor proteins such as dyneins and kinesins, which are involved in transport of vesicles and macromolecules within the cytoplasm (47). The capsids of adenovirus and herpes simplex virus are transported to the periphery of the nucleus with the aid of dynein, thus facilitating subsequent transport of the genome into the nucleus (48 -51). In this regard, it is likely that the microtubule networks would also be engaged in the intracellular trafficking of IpaH (and YopM) toward the periphery of the nuclear surface, although the putative motor proteins have not yet been identified.
Nucleoplasmic protein trafficking from the cytoplasm into the nucleus has been shown to be mediated by various transport systems. Small proteins of less than 40ϳ60 kDa passively diffuse by following a concentration gradient, whereas proteins larger than 40ϳ60 kDa, if they possess an NLS, are actively transported across the nuclear pore complex (38). For example, the SV40 large T-antigen, a 94-kDa protein, possesses PKKKRKV, the first NLS identified, and is actively transported into the nucleus (52,53). Accordingly, IpaH 9.8 with an estimated molecular weight of around 60 kDa seems to have no such NLS. In this context, we investigated the fate of Cy3labeled IpaH 9.8 with regard to import into the nucleus together with that of allophycocyanin-conjugated SV40-T-antigen NLS peptide (NLS-APC) and ␤-catenin in digitonin-permeabilized HeLa cells. Nuclear transport of NLS-APC has been shown to be dependent on transport factors such as importin ␣/␤ (39 -42), whereas ␤-catenin has no such requirement (43,44). In digitonin-permeabilized cell transport assay, in the absence of cytosol and without a supply of ATP/GTP, IpaH 9.8 was still efficiently imported into the nucleus. Furthermore, nuclear transport was sensitive to temperature and was inhibited by WGA acting as a nuclear pore plug, suggesting that the nuclear transport of IpaH 9.8 would not occur in a concentration-dependent manner, but rather the import appears to occur in an active fashion. Upon the addition of apyrase, which causes a decrease in ATP level, the nuclear transport of IpaH 9.8 was partially inhibited. Although the nuclear transport of IpaH 9.8 partially requires ATP, this requirement for IpaH 9.8 was relatively less than that for classical nuclear transported proteins such as NLS of SV40 large T antigen. Based on these results, we concluded that IpaH 9.8 could be transported into the nucleus, and similarly, ␤-catenin would be internalized into the nucleus independently of host-soluble factors (see Fig. 8). Although no direct evidence has yet been obtained, we speculate that the other IpaH proteins may also be imported into the host cell nucleus, since they showed similarities with the LRR motifs of IpaH 9.8 and YopM (22,25).
How intracellular Shigella can sense the host cytoplasmic environment and trigger the secretion of IpaH is still unknown. However, since IpaH secretion would occur mainly after bacterial entry into the host cells, and activation of the type III secretion system would be stimulated by bacterial contact with the host cells or growth under conditions such as in media containing CR, we speculate that rapidly motile intracellular Shigella within the cytoplasm, allowing direct contact with the inner surface of the cytoplasmic membrane, might be triggered to activate the type III secretion system as proposed in Fig. 8.
Finally, it is worth contemplating the significance of nuclear transport of IpaH 9.8 including other IpaH proteins in Shigella infection. Interestingly, with mutation of the ipaH 7.8 gene of the large plasmid of S. flexneri 2457T or both the ipaH 7.8 and ipaH 4.5 genes, although the bacterial invasion of epithelial cells had no effect, the mutants induced an exaggerated Sereny response in guinea pig eyes (54), suggesting that ipaH 7.8 plays a role in modulating the inflammatory response elicited by infection (23). Although it is unclear whether the roles of each of the IpaH proteins in Shigella infection are similar to each other, the enhanced inflammatory reaction in the Sereny test suggests that the nuclear-transported IpaH proteins participate in modulating gene expression involved in the production of inflammatory mediators such as interleukin-8. If the nuclear transport of IpaH is important for Shigella infection of the human colon, it is necessary to identify the targeting genes or factors of IpaH 9.8 protein.