Shigella invasion of macrophage requires the insertion of IpaC into the host plasma membrane. Functional analysis of IpaC.

Shigella infects residential macrophages via the M cell entry, after which the pathogen induces macrophage cell death. The bacterial strategy of macrophage infection, however, remains largely speculative. Wild type Shigella flexneri (YSH6000) invaded macrophages more efficiently than the noninvasive mutants, where YSH6000 induced large scale lamellipodial extension including ruffle formation around the bacteria. When macrophages were infected with the noninvasive ipaC mutant, the invasiveness and induction of membrane extension were dramatically reduced as compared with that of YSH6000. J774 macrophages infected with YSH6000 showed tyrosine phosphorylation of several proteins including paxillin and c-Cbl, and this pattern was distinctive from those stimulated by Salmonella typhimurium or phorbol ester. Upon addition of IpaC into the external medium of macrophages, membrane extensions were rapidly induced, and this promoted uptake of Escherichia coli. The exogenously added IpaC was found to be integrated into the host cell membrane as detected by immunostaining. The IpaC domain required for the induction of membrane extension from J774 was narrowed down within the region of residues 117-169, which contains a putative membrane-spanning sequence. Our data indicate that Shigella directs its own entry into macrophages, and the IpaC domain which is required for the association with its host membrane is crucial.

can multiply and move by inducing actin polymerization at one pole of the bacterium, which allows intracellular spreading of the bacterium within the cytoplasm as well as into the adjacent epithelial cells (6,7). Thus, the predominant feature of the Shigella pathogenicity is the ability to infect macrophage and epithelial cells, resulting in dissemination to adjacent epithelial cells.
A genetic study revealed that Shigella invasion of epithelial cells requires the numerous genes encoded by the 31-kilobase pair pathogenicity island on the large 230-kilobase pair plasmid (8). The pathogenicity island of Shigella flexneri contains 28 invasion-associated genes arranged in several transcribed regions (9). One of these contains the ipa genes whose products such as IpaA, IpaB, IpaC, and IpaD are required for bacteriacell interactions including induction of macropinocytosis from epithelial cells, whereas the mxi and spa regions are mostly involved in the formation of the type III secretion machinery (10 -12).
Studies have indicated that Shigella-induced macropinocytosis in epithelial cells occurs through a complicated process requiring the interaction of the type III-secreted effector proteins and host factors. In this process, the Ipa proteins (IpaA-D) are thought to play the most important roles. However, the involvement of Ipa proteins in the Shigella entry into epithelial cells is complicated and still unclear. Some of the Ipa proteins act both as regulators of the type III secretion system as well as effector proteins within the host cells. For example, IpaB and IpaD act as the molecular plug for regulating the type III secretion system (13)(14)(15). In addition, IpaB and IpaC, as well as Salmonella SipB and SipC (16) and Yersinia YopB and YopD (17), serve as a membrane pore in part of the type III machinery of the host cell plasma membrane, allowing the translocation of secreted effector proteins into the host cells (18). Upon contact to epithelial cells, a massive amount of IpaB and IpaC proteins is also secreted into the external medium from Shigella and promotes bacterial invasion by interacting with putative host receptors such as ␣ 5 ␤ 1 integrin and CD44 (19,20). However, these interactions would still be insufficient to elicit the macropinocytic events in epithelial cells required for bacterial entry. Further large scale rearrangement of actin dynamism is assumed to require a second signaling event within the host cytoplasm through the delivery of IpaA, IpaC, and IpgD via the type III secretion machinery (21,22). IpaA binds to vinculin, and the IpaA-vinculin complex together with F-actin promotes depolymerization of actin filaments required for modification of Shigella-induced membrane protrusions (23,24). IpaC somehow modulates actin dynamism, since formation of filopodia and lamellipodia can be induced when purified IpaC protein is added to semipermeabilized Swiss 3T3 cells or a ipaC clone is transfected into HeLa cells. The IpaC-induced membrane protrusions have been implicated in the activation of Cdc42 which in turn activates Rac1 (25). Although the mechanisms of IpaC-induced membrane protrusions are unknown, these studies have suggested that IpaC can somehow act as the effector for promoting Shigella invasion of epithelial cells.
Many invasive pathogenic bacteria have a specific strategy to resist the bactericidal activities of macrophages. For example, upon contact to macrophages, Yersinia activates the type III secretion system and delivers a set of Yop proteins into the cell (26 -28). The injected Yop proteins such as YopE, YopH, YopT, YopO, and YopP act in various ways to prevent phagocytosis and kill the macrophages, such as inducing destruction of actin filaments, interfering with cell signal transduction, or provoking apoptosis (29). Salmonella induces a macropinocytic event from epithelial cells and macrophages and delivers a subset of effector proteins from the type III secretion system encoded by the Salmonella pathogenicity island I, and then the pathogen can enter both types of cells. In the infection of macrophages, however, the pathogen eventually induces apoptosis to kill the cells (30,31), for which SipB encoded by Salmonella pathogenicity island I is delivered into macrophages (32,33). Salmonella typhimurium possesses an additional type III secretion system encoded by the Salmonella pathogenicity island II, which is activated inside the infected macrophage and delivers proteins such as SpiC and SifA that are required for bacterial proliferation in phagosomes and their maintenance (34,35). Shigella infects macrophages and disrupts the phagocytic vacuoles to escape into the cytoplasm, from which the bacterium later induces the cell death (36). A previous study (4) indicated that internalized Shigella induced macrophage apoptosis after 1-2 h postinfection, in which IpaB protein secreted via the type III secretion system within the macrophage cytoplasm played a crucial role. However, some other studies indicated that macrophage killing by apoptosis could only be induced if the cells were activated such as by stimulating with interferon-␥ (37,38). Whenever Shigella elicits different killing systems from the infected macrophages, the cell death results in release of massive amounts of interleukin-1␤, thus triggering a strong inflammatory response (39), and leads to an increase in the permeability of the epithelial barrier to Shigella entry and the migration of PMN 1 (40,41).
Therefore, elucidation of the bacterial strategy for infection of macrophages is an important issue for understanding the pathogenicity of Shigella and for the development of novel attenuated vaccines that are still invasive but do not induce macrophage cell death. Nevertheless, the molecular basis for infection of macrophages by Shigella still remains poorly understood. Hence, we decided to investigate whether the pathogen had the ability to direct its own internalization into macrophages and induce the activation of macrophage functions. Our data strongly indicate that Shigella directs its own phagocytic events in macrophages by exploiting the ability of IpaC to be integrated into the host membrane. In this study, we provide direct evidence for the first time indicating that the membranespanning IpaC domain is critical for the inducing macropinocytic event in macrophages.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Cell Culture, Plasmids, and Media-S. flexneri 2a YSH6000 is the wild type strain, and S325 is a mxiA::Tn5 derivative of YSH6000 used as the negative control deficient in the type III secretion activity (42). S. typhimurium SB300 was obtained from J. E. Galan (Yale School of Medicine, New Haven, CT). Escherichia coli MC1061 was used as the host for constructing various plasmids. pBluescript SKϩ (Stratagene), pGEX-2T (Amersham Pharmacia Biotech), pCACTUS-Tp r (43), and pMW119Tp (44) were used for genetic engineering experiments. E. coli JM109 was used as the host for constructing pQE-30 (Qiagen)-borne plasmids. All primers used for construction of the various plasmids are listed in Table I. Strains containing pMW119Tp-or pCACTUS-Tp r -based plasmids were grown in Muller-Hinton broth (Difco) when selection for trimethoprim resistance was necessary. For all macrophage infection experiments, overnight cultures of the bacterial strains were diluted 50-fold in brain-heart infusion broth (Difco) and incubated at 37°C for 2 h. J774 cells (ATCC TIB-67) and THP-1 cells (ATCC TIB-202) were maintained in RPMI 1640 (Sigma), and RAW264.7 cells (ATCC TIB-71) were maintained in Dulbecco's modified Eagle's medium (Sigma). For the preparation of HMDM, monocytes were isolated from the peripheral blood of healthy donors using Ficoll-Paque (Amersham Pharmacia Biotech) following the manufacturer's protocol. The separated mononuclear cells were plated on coverslips, and nonadherent cells were removed after 1 h of incubation. The medium was then replaced by fresh RPMI 1640 containing 10% FBS and 50 ng/ml human recombinant granulocytemacrophage colony-stimulating factor (PeproTech). Monocytes were cultivated for 5-7 days, and the adherent population was shown to be Ͼ95% macrophages as determined morphologically by Giemsa's stain.
Preparation of Antibodies-The antibodies specific for IpaB, IpaC, and IpaD used for the immunoblots have been described previously (45). A fusion protein of IpaC tagged with six histidine residues at the N terminus was constructed using the QIAexpress system with a pQE-30 plasmid. A DNA fragment including the ipaC gene was amplified by PCR using an ipaC-1 primer containing a BamHI site and an ipaC-2 primer containing a SalI site together with pMYSH6000, the large virulence plasmid in YSH6000 as the template. The amplified BamHI-SalI fragment was cloned into pQE-30, resulting in the plasmid pQE-ipaC-fl. The fusion IpaC protein purified by nickel-nitrilotriacetic acid chromatography (Qiagen) was used to immunize rabbits. The anti-IpaC whole molecule antiserum was incubated with nitrocellulose membrane containing transferred fusion IpaC protein, and the Ig fraction specific for the whole molecule of IpaC (anti-IpaC-wm) was eluted. IpaC peptides named IpaC-n, encompassing amino acid residues 35-55 (ISTKQTQSSSETQKSQNYQQI), IpaC-m, encompassing amino acid residues 140 -169 (RTAETKLGSQLSLIAFDATKSAAENIVRQG), and IpaC-c, encompassing amino acid residues 228 -247 (KQIDTNITSPQT-NSSTKFLG), were synthesized. The anti-IpaC whole molecule antiserum was incubated with IpaC-n-, IpaC-m-or IpaC-c-conjugated, epoxyactivated Sepharose 6B (Amersham Pharmacia Biotech) to obtain an Ig fraction specific for each peptide; the antibodies obtained were named anti-IpaC-n, anti-IpaC-m, and anti-IpaC-c, respectively.
Gentamicin Protection Assay-The invasion efficiency of bacteria was measured with a gentamicin protection assay (46). Briefly, a 24well tissue culture plate (Costar) was seeded with 2 ϫ 10 5 cells/well and incubated overnight at 37°C under 5% CO 2 . Each well was then in- 1 The abbreviations used are: PMN, polymorphonuclear leukocyte; HMDM, human monocyte-derived macrophage; m.o.i., multiplicity of infection; PBS, phosphate-buffered saline; mAb, monoclonal antibody; COZ, complement-opsonized zymosan; IOZ, IgG-opsonized zymosan; BOZ, bovine serum albumin-opsonized zymosan; PMA, phorbol myristate acetate; PY, phosphotyrosine; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis; LPS, lipopolysaccharide.  fected at an m.o.i. of ϳ500 for 30 min, before prewarmed tissue culture medium containing gentamicin was added at a final concentration of 200 g/ml. After a 15-min incubation, the medium was removed, and the macrophages were washed three times with PBS. PBS containing 0.5% Triton X-100 was then added to each well in order to lyse the cells.
After an appropriate dilution with saline, the lysed solution was plated onto L agar, and the number of viable bacteria was counted. Immunostaining Assay-Immunostaining was performed essentially as described previously (47). Mac-1 and Fc␥R were labeled with the anti-Mac-1 mAb M1/70 (BIOSOURCE International, Inc.) or the anti-Fc␥R mAb 2.4G2 (PharMingen), respectively. The anti-rat IgG antibody conjugated with fluorescein isothiocyanate (Sigma) was used as the secondary antibody. In order to measure the invasive capacity of the bacteria by immunostaining, the number of bacteria internalized in the macrophage cytoplasm was counted as follows: J774 cells seeded on coverslips were infected at an m.o.i. of ϳ30 and centrifuged for 10 min at 900 ϫ g. The infected J774 cells were then incubated for 15 min at 37°C under 5% CO 2 . Next the cells were washed with PBS and then fixed for 15 min with PBS containing 4% paraformaldehyde. The fixed cells were incubated in PBS containing anti-Shigella LPS antiserum for 1 h. To visualize extracellular bacteria, the cells were washed and then incubated with anti-rabbit IgG conjugated to fluorescein isothiocyanate (Sigma). To visualize intracellular bacteria, the cells were washed again and permeabilized by incubation with PBS containing 0.2% Triton X-100 for 20 min at room temperature. The cells were then treated once more with the anti-Shigella LPS antiserum and incubated with antirabbit IgG conjugated to Cy5 (Amersham Pharmacia Biotech). Immunostained samples were examined by confocal laser scanning microscopy (Radiance Plus, Bio-Rad).
Measurement of Spread in J774 Cells Infected with Shigella-J774 cells seeded on coverslips were infected at an m.o.i. of ϳ500 before being centrifuged for 10 min and incubated for 30 min at 37°C under 5% CO 2 . The cells were then washed with PBS and fixed in PBS containing 4% paraformaldehyde. The fixed cells were stained with Giemsa's solution. Photographs of the cells were taken with a CCD camera on the microscope, and the cell areas were measured using IPLab Spectrum software (Signal Analytics Corp.).
Scanning Electron Microscopy-Macrophages were grown on 12-mm round coverslips (Matsunami) before being infected by centrifugation as described above. For further fixation, the washed and fixed cells were treated with 2% OsO 4 and then processed for electron microscopy as described by Ginocchio et al. (48). Visualization of samples was carried out by an S800 scanning electron microscope (Hitachi).
Generation of a Nonpolar ipaC Mutant and pIpaC-For construction of a nonpolar mutant of ipaC, the aphA-3 (kanamycin resistance gene) cassette was used (49). Purified pMYSH6000 was used as the template for PCR. Nonpolar mutants of ipaC were constructed as follows: a DNA fragment encompassing nucleotides from position 1226 upstream of the 5Ј end of the ipaC gene to nucleotide 215 downstream from the 5Ј end was amplified by PCR using primers icm-1 containing an ApaI site and icm-2 containing an SmaI site. The ApaI-SmaI fragment was cloned into pBluescript, resulting in the plasmid pipaC-u. Another DNA fragment encompassing nucleotides from position 1076 downstream from the 5Ј end of the ipaC gene to nucleotide 2494 downstream from the 5Ј end was amplified using primers icm-3 containing an SmaI site and icm-4 containing a BamHI site. The SmaI-BamHI fragment was then cloned into pipaC-u, yielding the plasmid pipaC-ud. The aphA-3 cassette was cloned at the SmaI site of pipaC-ud in the correct orientation, resulting in the plasmid pBSipaC::Km r . The inactivated ipaC gene digested from pBSipaC::Km r at ApaI and NotI sites was subcloned into pCACTUS-Tp r , followed by introduction of the resultant plasmid named pipaC::Km r into YSH6000 by electroporation. Integration and selection of nonpolar ipaC mutants were achieved as described previously (43), and the obtained mutant strain, ⌬ipaC, was named TK001. For complementation of the ipaC gene defect in TK001, pIpaC was constructed by cloning the DNA fragment containing the ipaC gene into pMW119Tp. This fragment was amplified by PCR using primer icc-1 and icc-2 with pMYSH6000 as the template.
Analysis of Effector Protein Expression and Secretion-The effector protein solution released in the Shigella culture supernatants following stimulation with Congo red and the whole bacterial lysate were prepared as described previously (43). Each protein sample from the same number of bacteria was separated by 12% SDS-PAGE and immunoblotted with anti-IpaB, anti-IpaC, and anti-IpaD antibody.
Phosphotyrosine-J774 cells were seeded on 6-well tissue culture plates (Costar) at a concentration of 4 ϫ 10 5 cells/well and incubated overnight at 37°C under 5% CO 2 . Shigella and Salmonella strains were added at an m.o.i. of ϳ500 and centrifuged for 10 min. COZ, IOZ, or BOZ was prepared as described previously (50,51). Opsonized zymosan particles were centrifuged onto the macrophage (particle:cell ratio ϭ 10:1) for 2 min. Treatment of macrophage with PMA (Sigma) was carried out at a final concentration of 100 ng/ml. After various incubation times, the cells were washed twice with ice-cold PBS and then lysed in 50 l of lysis buffer (10 mM Tris-HCl, pH 7.6, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM NaF, 1% Triton X-100, 100 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml pepstatin A, 10 mg/ml leupeptin) per well. The cell lysates were sonicated for 10 s and clarified by centrifugation at 15,000 ϫ g for 15 min. The supernatants were then assayed for protein concentration with SDS-PAGE. The supernatants were incubated with anti-PY mAb 4G10 (Upstate Biotechnology Inc.), anti-paxillin mAb 349 (Transduction Laboratories), or anti-c-Cbl antibody C-15 (Santa Cruz Biotechnology). Immunoprecipitations were performed overnight at 4°C. Protein G-Sepharose (Sigma) was used for precipitation of the immunocomplexes. The anti-PY antibody RC20 (Transduction Laboratories) was used for detection of phosphotyrosine proteins by immunoblot analysis. In the case of zymosan phagocytosis measured at 0, 15, 30, and 45 min, protein tyrosine phosphorylation was maximum at 15 min after addition (data not shown).
Sequential Observation of Macrophage Morphology-J774 or HMDM were seeded on 35-mm dishes. The cells were washed twice with prewarmed RPMI medium without FBS, and then medium containing 1/10th volume of recombinant protein solution prepared at 5 M was added. Macrophage shape was observed with an Axiovert 135 microscope (Zeiss) equipped with a SenSys 1400 CCD camera (Roper Scientific) in chamber maintained at 37°C with a 5% CO 2 atmosphere.
Phagocytosis Assay-The bacterial pellet of E. coli HB101 grown to middle log phase was suspended in the FBS-free RPMI medium containing recombinant protein at a final concentration of 0.5 M. Preparation of J774 cells and the phagocytosis assay were carried out under almost the same conditions as described for the gentamicin protection assay mentioned above. The cell medium was replaced with 0.5 ml/well RPMI without FBS prior to addition of the bacterial suspension. The bacterial pellet was added at an m.o.i. of ϳ1000 for 15 min. After gentamicin treatment for 30 min, the internalized bacteria were recovered and spread onto L-agar plates.
Construcion of GST-VirA, GST-IpaD, and IpaC Deletion Plasmids-The DNA fragments containing virA or ipaD gene amplified by PCR using primers virA-1 and virA-2 or ipaD-1 and ipaD-2, respectively, were cloned into pGEX-2T. After the purification under native conditions and digestion with thrombin according to the manufacturer's instructions, the GST fusion protein was dialyzed against Tris-buffered saline. pQE-ipaC-fl was digested with SacI and HindIII or StuI and HindIII in order to obtain the plasmids designated pQE-ipaC-⌬C1 (containing the coding region of IpaC amino acid residues 1-284) or pQE-ipaC-⌬C4 (containing the coding region of IpaC amino acid residues 1-169), respectively. After digestion, the DNA fragments containing the pQE-30 sequence were recovered from agarose gel and filled in using T4 DNA polymerase (Toyobo) followed by ligation. To obtain the plasmids named pQE-ipaC-⌬NC (containing the coding region of IpaC amino acid residues 117-284) and pQE-ipaC-⌬CTM (containing the ⌬C1 deleted with coding region of IpaC amino acid residues 117-169), the DNA fragments were amplified by PCR using primers NC-1 and NC-2 or CTM-1 and CTM-2, respectively, together with pQE-ipaC-⌬C1 as a template. These DNA fragments were digested with BamHI (for construction of pQE-ipaC-⌬NC) or SalI (for construction of pQE-ipaC-⌬CTM) and then ligated. To obtain the plasmid named pQE-ipaC-⌬C8 (containing the coding region of IpaC amino acid residues 1-112), the DNA fragment was amplified by PCR using pQE-ipaC-fl as a template and primers C8 -1 and C8 -2. This DNA fragment was then digested with SalI, and the amplified fragment was ligated. These pQE-30-based plasmids were introduced into the E. coli JM109 strain, and the strains harboring pQE-ipaC-⌬C1, pQE-ipaC-⌬C4, pQE-ipaC-⌬NC, pQE-ipaC-⌬CTM, and pQE-ipaC-⌬C8 expressing the truncated IpaC fusion proteins tagged with six histidine were designated ⌬C1, ⌬C4, ⌬NC, ⌬CTM, and ⌬C8, respectively. All histidine-tagged fusion protein was dialyzed against Tris-buffered saline after the purification under native conditions according to the manufacturer's protocol.

Invasive Shigella Enters the Macrophage More Efficiently
Than the Noninvasive Mutant-To investigate whether the invasive S. flexneri would be able to direct internalization into macrophages, macrophage cell lines such as J774, RAW264.7, and THP-1 were infected with YSH6000 (wild type S. flexneri) or S325 (a noninvasive mutant deficient in the type III secre-tion system). At 30 min after the addition of bacteria, fresh medium containing gentamicin was added to the cultures and incubated for a further 15 min. The internalized bacteria were counted by plating the cell lysates on agar revealing that the numbers of YSH6000 internalized into J774, THP-1, and RAW264.7 were 7-, 5-, and 2 times higher than that of S325, respectively (Fig. 1A). Although the number of internalized bacteria varied among cell lines, the bacterial number associated with each cell line after cytochalasin D treatment showed no substantial differences between YSH6000 and S325 (data not shown). To ensure further the difference in bacterial invasive capacity, J774 cells infected with YSH6000 or S325 were fixed, and the internalized bacteria were immunostained with anti-Shigella LPS antiserum and anti-rabbit IgG conjugated with Cy5 (see "Experimental Procedures"). As shown in Fig.  1B, the number of internalized YSH6000 was significantly higher (ϳ10 times) than that of S325. These results suggest that wild type S. flexneri directs its own internalization into macrophages.
Shigella Invasion of Macrophages Elicits Cell Spreading-J774 infected with YSH6000 or S325 were stained by Giemsa's solution, and the cell morphology was investigated using microscopy. As shown in Fig. 2A, S325 was observed associated with macrophages; however, some of the macrophages infected with YSH6000 showed remarkable cellular spreading. Typically, the average area of cells infected with YSH6000 (157.0 Ϯ 4.9 m 2 , n ϭ 301) was significantly higher than that with S325 (71.0 Ϯ 1.3 m 2 , n ϭ 303) or that of cells without infection (72.9 Ϯ 1.1 m 2 , n ϭ 305). Scanning electron microscopy confirmed that J774 infected with YSH6000 extended a large scale lamellipodium (Fig. 2B). Also in some cases, infection of HMDM by YSH6000 (Fig. 2B) or J774 by YSH6000 (data not shown), membrane ruffles protruded and surrounded the bacteria. However, no J774 or HMDM cells infected with S325 showed such cellular responses, suggesting that S. flexneri has the ability to induce a large scale membrane extension including formation of lamellipodia and membrane ruffles in macrophages.
IpaC Is Essential for Efficient Uptake of Shigella by Macrophages-So far, studies have indicated that IpaC plays the most central role in inducing macropinocytic events in invasion of epithelial cells by Shigella (25). Hence, we constructed a nonpolar ipaC mutant (TK001) based on YSH6000 and introduced the wild type ipaC clone (pIpaC) into TK001. The ability of YSH6000, TK001, TK001/pIpaC, or S325 to secrete IpaB, IpaC, and IpaD into medium was investigated using a conditional medium (PBS containing 30 g/ml of Congo red) to stimulate the type III secretion activity. As shown in Fig. 3A, although the effectors such as IpaA and IpgD could still be secreted from TK001 at similar levels to those from YSH6000 or TK001/pIpaC, no IpaC was secreted from TK001 at all. An immunoblot with anti-IpaB, IpaC, and IpaD antibodies showed that IpaB and IpaD but not IpaC were present in the whole cell lysate from TK001, confirming that TK001 is deficient in IpaC production. We thus investigated TK001 together with TK001/ pIpaC, YSH6000, and S325 for their invasive capacity into J774 using the gentamicin-protection assay. The ability of TK001 to enter macrophages was reduced to less than 20% that of the wild type level (Fig. 3C). The reduced invasive capacity of TK001 was restored by the introduction of pIpaC (Fig. 3C). The appearance of cell spreading as well as the scale of lamellipodium by TK001 was also remarkably decreased as compared with that of YSH6000 (Fig. 3C). Fig. 3D shows a typical example of spreading by J774 cells infected with YSH6000, in which the cell dramatically spread as extending cell periphery with membrane ruffles and evoked local F-actin condensation around bacteria. J774 infected with YSH6000 or TK001 at 15 min after infection was subsequently investigated for the accumulation of Fc␥R and Mac-1 (CR3, ␣ M ␤ 2 integrin), since they are the major phagocytic receptors involved in uptake of particles or bacteria and depend on the rearrangement of the actin cytoskeleton (52)(53)(54). As shown in Fig. 3E, Fc␥R accumulated around the area of entry with YSH6000, but this was not observed with TK001. Mac-1 also accumulated but locally at the sites where YSH6000 was associated with J774, and thus Mac-1 foci were also present at the adhesion points along the spreading cell periphery (Fig. 3F). The results of these experiments strongly suggest that secreted IpaC is critical for inducing the phagocytic event that promotes Shigella invasion of macrophages. Tyrosine Phosphorylation of Macrophage Proteins Induced by Shigella-Fc␥R-mediated phagocytosis and Mac-1-mediated cell adhesions have been shown to stimulate host protein tyrosine phosphorylation (55)(56)(57). To assess whether the phagocytic events including macrophage spreading induced by Shigella would also involve protein tyrosine phosphorylation, J774 infected with YSH6000 or TK001 at various times were examined for protein tyrosine phosphorylation by immunoprecipitation and immunoblotting with anti-PY antibodies. At 15 min post-infection of J774 with YSH6000 but not with TK001, proteins corresponding to the molecular masses of 44, 45, 60, 66, 68, 70, 80, 100, and 116 kDa were phosphorylated, and the levels of phosphorylation were maximum at 30 -45 min postinfection (Fig. 4A). After a 15-min infection of J774 with YSH6000 but not with TK001, tyrosine phosphorylation of paxillin was detected by immunoprecipitation with anti-paxillin monoclonal antibody (Fig. 4B). When the tyrosine-phosphorylated proteins in J774 infected with YSH6000 were precipitated by anti-PY, proteins corresponding to 66, 68, and 70 kDa were recognized with the anti-paxillin mAb (Fig. 4C). Macrophage spreading is thought to involve the tyrosine phosphorylation of some proteins including c-Cbl (58). Tyrosine phosphorylation of c-Cbl was also detected at 30 min post-infection with YSH6000, at which time the level of tyrosine-phosphorylated c-Cbl was 1.5 times higher than that with TK001 (Fig. 4D). When tyrosine-phosphorylated proteins in J774 infected with YSH6000 were precipitated with anti-PY, proteins corresponding to 100 and 116 kDa were recognized with the anti-c-CbI antibody (Fig. 4E). To see whether the protein tyrosine phosphorylation in macrophages infected with YSH6000 would be similar to that induced by other known stimulators that induce membrane extensions from macrophages (31,59), the profiles of tyrosine-phosphorylated proteins induced in J774 were compared with that by Salmonella infection or PMA stimulation. As shown in Fig. 4, F and G, the pattern of protein tyrosine phosphorylation induced by Salmonella and PMA were different from that by Shigella. Although the 66-kDa protein, which was identified as paxillin, was strongly phosphorylated by Shigella, it was not so prominent as compared with the cells stimulated with Salmonella or PMA (Fig. 4, F and G). Furthermore, the patterns of tyrosine-phosphorylated proteins in J774 induced by YSH6000 was different from that of phagocytosis of zymosan opsonized with C3bi (COZ), IgG (IOZ), or bovine serum albumin (BOZ) (Fig. 4H). These results suggest that Shigella entry into macrophages stimulates protein tyrosine phosphorylation, including paxillin and c-Cbl, but that pattern is rather unique to Shigella invasion.
Exogenously Added IpaC Can Stimulate Cell Spreading and Uptake of Bacteria by Macrophage-To assess the role of IpaC in stimulating macrophage uptake of Shigella, the ability of the IpaC protein to induce macrophage spreading was investigated using phase contrast microscopy with a time-lapse imaging system. Upon addition of purified IpaC into the external medium at final concentration of 0.5 M (20 g/ml), 20 -30% of J774 or HMDM cells responded by extending lamellipodia as early as 2 min after addition. The scale of lamellipodia was maximum at 5 min, at which the periphery of extended lamellipodia also showed ruffling (Fig. 5A, arrowheads), although this was mostly over by 30 min. Fig. 5B shows a typical spreading of a J774 cell stained with rhodamine-phalloidin, where F-actin condensation appeared along the leading edge of the lamellipodium (Fig. 5B). When an E. coli strain such as HB101 was added together with the purified IpaC, uptake by J774 was increased up to ϳ1.6-fold that by HB101 alone (Fig. 5C). Under the same conditions, addition of other purified effector proteins such as IpaD or VirA had no appreciable effect on bacterial uptake. In contrast to J774 cells, the addition of the same amount of IpaC as above into the external medium of HeLa or Caco-2 cells did not stimulate any changes in cell shape or uptake of HB101 (data not shown). Thus, these results suggest that IpaC has the ability to induce membrane extensions in macrophages and promote uptake of bacteria by the cell.
Functional Domain of IpaC Required for Inducing Macrophage Spreading-To determine the functional domain of IpaC involved in induction of the membrane extensions in macrophages, various truncated IpaC versions were constructed as the histidine-tagged proteins (Fig. 6A). Purified truncated IpaC proteins were checked by SDS-PAGE (Fig. 6B), and equal amounts of each of protein were added into the external medium of J774, and the cells were observed using the phase contrast microscopy or immunostaining with the anti-IpaC-wm antibody. Fig. 6A summarizes the ability of each IpaC-truncated protein to induce membrane extensions or associate with the macrophage membrane. Fig. 6C (top) shows a typical cellular response to the addition of a truncated IpaC protein such as ⌬C1, which results in membrane extensions. Addition of ⌬C4 and ⌬NC could also induce membrane extensions in J774 as large as that caused the fulllength IpaC (data not shown). The fluorescence signals for the full-length IpaC and the truncated versions (⌬C1, ⌬C4, and ⌬NC) were found associated with the cell membrane (Fig. 6C, top). Importantly, when the membrane-associated fluorescence signals were analyzed at the vertical axis, the signals for ⌬C1 were found to be integrated into host plasma membrane detected for membrane-spanning signals (Fig. 6C), whereas the other IpaC truncated versions such as ⌬CTM or ⌬C8 were not associated with the cells at all (data not shown). Taking these results together it strongly suggest that the residues encompassing positions 117-169 (see Fig. 4A) are sufficient for inducing the membrane extensions and the association with the macrophage membrane.
Topological Analysis of IpaC on the Macrophage Membrane-For analysis with TMpred, a program for prediction of putative membrane-spanning regions (60) in IpaC, the residues encompassing 121-139 and 169 -191 were predicted to be putative transmembrane domains, whereas the remaining Nterminal and C-terminal regions would be present on the external side of the plasma membrane. To determine the membrane topology of IpaC, three IpaC antibodies (anti-IpaC-n, -m, and -c), which recognized three distinctive regions in IpaC (see Fig. 7A, top bar), were purified using synthetic peptides (see "Experimental Procedures"). The resulting anti-IpaC-n, anti-IpaC-m, and anti-IpaC-c antibodies were found to recognize the N-terminal IpaC portion, a hydrophobic region bracketed by the two putative transmembrane regions (TM1 and TM2) and a region following TM2, respectively, by using various truncated IpaC versions (Fig. 6A) in immunoblotting (data not shown). J774 cells on two coverslips for staining with each antibody were incubated with purified ⌬C1 for 5 min. In the experiment, ⌬C1 was used instead of the full-length IpaC due to the ease in its preparation as compared with the fulllength IpaC. After fixation, one of the coverslips was treated with 0.2% Triton X-100 to permeabilize the host cell membrane, and the other one was not. The cells on both coverslips were then immunostained with each anti-IpaC antibody. The fluorescent signal of IpaC stained with the anti-IpaC-m antibody was significantly decreased in nonpermeabilized cells (Fig. 7A, e) as compared with that in the permeabilized cell (Fig. 7A, b) or nonpermeabilized cell stained with the other antibodies (Fig. 7A, d and f). These data together with the results in Fig. 6 suggest that the residues 140 -168 of IpaC exist as the cytoplasmic loop, whereas the preceding N-terminal portion of TM1 and the following C-terminal portion of TM2 exist as the external membrane domains (Fig. 7B). DISCUSSION Shigella infection of the resident macrophages beneath M cells overlaying the solitary lymphoid nodules in the colon constitute an early critical infection event. The bacteria released from the macrophages then basolaterally invade the colonic epithelial cells, whereas the infected macrophages and epithelial cells produce inflammatory mediators, the major cause leading to bacillary dysentery (61). Nevertheless, the strategy for Shigella infection of macrophages has remained largely speculative. Therefore, in the present study, we investigated whether Shigella could direct its own internalization into macrophages, like the case for invasion of epithelial cells.
Our data indicate that Shigella can initiate its own internalization into macrophages, and the secreted IpaC protein plays a major role in triggering the phagocytic event. This conclusion is based on the following results. (i) The efficiency of invasion of macrophage by wild type S. flexneri was significantly higher than that by noninvasive mutants. (ii) Infection of macrophages by wild type S. flexneri but not the noninvasive mutants induced a large scale membrane extension, including lamellipodia and membrane ruffles. (iii) The addition of purified IpaC protein into the external medium of macrophages rapidly induced membrane extensions and mediated bacteria uptake.
Although the mechanisms underlying the IpaC-mediated cell spreading event are still to be elucidated, some surface receptors such as Mac-1 on the macrophage might be involved in the phagocytic event. Incubation of J774 cells with wild type Shigella but not with the ipaC deletion mutant generated Mac-1 foci at the site of bacterial attachment as well as along the periphery of the extended cell membrane, strongly indicating that an integrin-dependent adhesion event, which is a prominent feature of the Mac-1-mediated macrophage adherence, had been stimulated. This observed macrophage response is consistent with the previous report by Renesto  The tyrosine-phosphorylated proteins precipitated with 4G10 were separated on 7.5% SDS-PAGE gel and blotted with RC20. The lower panel shows the proteins around 45 kDa after a long exposure. Numbers to the right indicate the molecular weight standards. The arrows on the left side indicate the signals specifically increased in cells infected with YSH6000. B, paxillin in the lysate from cells infected with YSH6000 or TK001 was precipitated with an anti-paxillin mAb. The tyrosine-phosphorylated proteins and paxillin were detected by RC20 and anti-paxillin mAb, respectively. C, tyrosine-phosphorylated proteins or the paxillin in cells infected with YSH6000 for 30 min were precipitated with 4G10 or anti-paxillin mAb, respectively. The tyrosine-phosphorylated proteins and paxillin were detected by RC20. The arrows indicate the paxillin band. D, c-Cbl in cells infected with YSH6000 or TK001 was precipitated with an anti-c-Cbl antibody. The tyrosine-phosphorylated proteins and c-Cbl were detected by RC20 and anti-c-Cbl antibody, respectively. Signal intensity of the bands was measured by NIH image software. The lower band of c-Cbl is the truncated version (70). E, tyrosine-phosphorylated proteins or the c-Cbl in the lysate from the cells infected with YSH6000 for 30 min were precipitated with 4G10 or anti-c-Cbl antibody, respectively. RC20 was used in immunoblotting to detect the tyrosine-phosphorylated proteins. The arrows shown at the right side of photograph indicate the c-Cbl band. F and G, tyrosine-phosphorylated proteins in J774 cell infected with S. typhimurium (F) or simulated with PMA (G) were analyzed as described in A. H, tyrosine-phosphorylated proteins in J774 cell engulfing complement-opsonized zymosan (COZ), IgG-opsonized zymosan (IOZ), or bovine serum albumin-opsonized zymosan (BOZ) were analyzed as in A. The lysates were prepared from cells 15 min after the addition of the zymosans. The lysate from the cells infected with YSH6000 (wild type) or TK001 (ipaC mutant) for 30 min was used as the positive and negative controls, respectively. which PMN suspended in medium became adherent to culture plates coated with FBS when wild type S. flexneri but not the ipaC mutant was added into the external medium.
Previous studies (57,(63)(64)(65) have indicated that clustering of Fc␥R and Mac-1 stimulate protein tyrosine phosphorylation and local rearrangement of the actin cytoskeleton, although the tyrosine-phosphorylated proteins vary depending on the cell signaling pathway and the nature of ligands including surface receptors engaged in phagocytic events (53). Therefore, we investigated the profile of induced protein tyrosine phosphorylation in J774 infected with wild type Shigella or ipaC deletion mutant. The results indicated that Shigella infection of macrophages induced tyrosine phosphorylation of a set of proteins including paxillin and c-Cbl but that it was rather distinctive from those induced by Salmonella infection or PMA stimulation (see Fig. 4, E and F). In macrophages and PMN, paxillin is thought to be tyrosine-phosphorylated in the cell spreading and adhesion mediated by the ␣ M ␤ 2 integrin (Mac-1) (56,59) or phagocytosis mediated by Fc␥R (55). Tyrosine phosphorylation of c-Cbl has been thought to be stimulated through macrophage spreading, for which the integrin bound to an extracellular matrix such as fibronectin is indicated to be involved in the interaction of c-Cbl with Src family kinases (58). Overexpression of c-Cbl in COS-7 cells or macrophages has been shown to stimulate Fc␥R-mediated phagocytosis (66). Although the increased level of c-Cbl tyrosine phosphorylation in J774 during Shigella infection was marginal, the increase was reproducible in three independent experiments, suggesting that phosphorylated c-Cbl might also contribute to Shigella invasion of macrophages. Although tyrosine phosphorylation of some other proteins such as Vav or Syk has also been found to be stimulated by Fc␥R-mediated phagocytosis or Mac-1-mediated cell adhesion (55, 65, 67), so far no proteins other than paxillin and c-Cbl have yet been shown to be convincingly tyrosine-phosphoryl-ated in J774 cells during Shigella infection using immunoprecipitation and immunoblotting assays. Nevertheless, these results lead us to speculate that the entry into macrophages involves the activation of some cellular signaling pathways that might be unique to the Shigella entry process.
In infection of the macrophage by Shigella, we focused on the role of IpaC, one of the type III secreted proteins, since previous studies (25,68) indicated that IpaC has the ability to induce actin reorganization from nonprofessional phagocytic cells such as epithelial or fibroblastic cells including the activation of Cdc42. Our results showed that when purified IpaC was added to the external medium of cells, we observed that only macrophage cells responded quickly to extend lamellipodium, a reaction that was never observed in nonprofessional phagocytic cells (25,68). Indeed, whereas the addition of the purified IpaC protein into the external medium of epithelial cells such as HeLa or Caco-2 cells showed no appreciable effect on cell shape, addition of the IpaC protein into J774 cells promoted uptake of bacteria such as E. coli HB101.
IpaC has been shown to be composed of three distinctive domains (25). The first domain is the N-terminal portion encompassing the first residue to residue 116. The second domain is the central hydrophobic portion encompassing residues 117-190. The third domain is the remaining C-terminal portion. Consistent with a recent report (69) on epithelial cells, we demonstrated that the central hydrophobic domain of IpaC is essential for binding to the host plasma membrane. The experiment to narrow down the functional domain of IpaC indicated that the central IpaC region encompassing residues 117-169, which contains one of the putative two membrane-spanning domains, is required for the association with the host membrane as well as for inducing macrophage spreading. Indeed, we could directly visualize the IpaC protein integrated into the macrophage plasma membrane using immunofluorescence staining. Furthermore, we also determined the membrane topology of IpaC by immunostaining with three distinctive anti-IpaC antibodies. Our data strongly indicate that IpaC is integrated via the two membrane-spanning regions with an ϳ30-amino acid cytoplasmic loop, whereas the N-and C-terminal portions exist on the external side of the plasma membrane (see Fig. 7B). Since some IpaC truncated versions capable of associating with the host membrane were able to stimulate macrophage cell spreading, although no direct evidence has yet been obtained, IpaC might have some association with or effect on some putative host receptor(s) mediating the induction of membrane extension via activation of some cellular signal transduction pathway such as activation of Cdc42 (25). If so, the macrophage system would be useful for investigating the downstream signal transduction pathways evoked upon integration of IpaC into the host membrane.
In summary, our data provide for the first time clear evidence supporting the idea that Shigella directs its own invasiveness into macrophages, and the IpaC involved in the invasion of epithelial cells also appears to be critical for triggering of the phagocytic event. The information on the distinctive functional organization of the IpaC domains based on their effect on macrophage spreading will also provide an important insight into understanding the precise role of IpaC in Shigella pathogenesis.