Genetic modulation of Shigella flexneri 2a lipopolysaccharide O antigen modal chain length reveals that it has been optimized for virulence

Renato Morona, Craig Daniels{dagger} and Luisa Van Den Bosch

Department of Molecular Biosciences, University of Adelaide, Adelaide, South Australia, Australia, 5005

Correspondence
Renato Morona
renato.morona{at}adelaide.edu.au


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The lipopolysaccharide (LPS) molecules of Shigella flexneri 2a have O antigen (Oag) polysaccharides with two modal chain length distributions. The chromosomal wzzSF gene results in short (S) type Oag chains [11–17 Oag repeat units (RUs)], and the pHS-2 plasmid-located wzzpHS2 gene results in very long (VL) type Oag chains (>90 Oag RUs). S. flexneri wzzSF mutants are unable to form plaques on HeLa cell monolayers and F-actin comet tails, indicating that IcsA/VirG function in actin-based motility (ABM) is defective. An S. flexneri wzzSF wzzpHS2 double mutant had LPS with relatively short, random length Oag chains and, paradoxically, was able to form plaques and F-actin comet tails. The influence of Oag modal chain length distribution on virulence and related properties was investigated using complementation with different wzz genes. WzzO139 from Vibrio cholerae O139 and WzzST from Salmonella enterica serovar Typhimurium were fully functional in Shigella flexneri, resulting in LPS with either very short (VS) type Oag chains (2–7 Oag RUs) or long (L) type Oag chains (19–35 RUs), respectively. In the absence of VL-type Oag chains, the VS-, S- and L-type Oag chains were permissive for plaque and F-actin comet tail formation. However, in the presence of LPS with VL-type Oag chains, the VS- and S-type Oag chains but not the L-type Oag chains were permissive for plaque and F-actin comet tail formation. These data, and the results of a previous investigation, show that IcsA function in ABM requires LPS Oag chains with at least two but less than 18 RUs when VL-type Oag chains are co-expressed on the cell surface. However, in the absence of the VL-type Oag chains, LPS Oag chains with at least two but less than 90 RUs are able to support IcsA function in ABM. Indirect immunofluorescence staining of IcsA on the cell surface of the S. flexneri strains did not correlate with the observed effect of Oag chain length on plaque and F-actin comet tail formation. However, when intracellular bacteria lacking VL-type Oag chains were examined, an inverse correlation between Oag modal chain length and detection of IcsA was observed, i.e. staining decreased with increased modal length. It is hypothesized that Oag chains can mask IcsA and interfere with its function in ABM, and a model is presented to explain how LPS Oag and IcsA may interact. It is suggested that S. flexneri 2a has evolved to synthesize LPS with two Oag modal chain lengths, as S-type Oag chains allow IcsA to function in ABM in the presence of VL-type Oag chains that confer resistance to serum.


Abbreviations: ABM, actin-based motility; IF, immunofluorescence; L, long; LPS, lipopolysaccharide; Oag, O antigen; RU, repeat unit; S, short; VL, very long; VS, very short

{dagger}Present address: Program in Cell and Lung Biology, Hospital for Sick Children Research Institute, University of Toronto, McMaster Building Room 3026C, 555 University Ave, Toronto, Ontario M5G 1X8, Canada.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
The bacterium Shigella flexneri is able to cause bacillary dysentery in humans, and many of its virulence factors have been identified and extensively studied (Hale, 1991; Philpott et. al., 2000). This bacterium is able to invade intestinal cells, escape from the endocytic vacuole, replicate within the cytoplasm and initiate actin-based motility (ABM) that allows it to spread within and between cells. The infection results in cell killing and an acute pro-inflammatory response (Sansonetti, 2001). S. flexneri 2a, like many other Gram-negative enteric organisms, has O antigen (Oag) chains associated with its lipopolysaccharide (LPS) molecules. The Oag chains and their chain length distribution contribute to Shigella virulence, as mutations in Oag biosynthesis genes reduce virulence (Hong & Payne, 1997; Okada, et. al., 1991a, b; Okamura & Nakaya, 1977; Okamura et. al., 1983; Rajakumar et. al., 1994; Sandlin et al., 1995, 1996; Van Den Bosch et. al., 1997). In general, S. flexneri 2a strains produce LPS with two Oag modal chain lengths. The chromosomally encoded WzzSF protein results in LPS with short (S) type Oag chains with a modal length of 11–17 RUs (Morona et al., 1995, 2000). The S-type Oag chains seem to be required to promote efficient cell-to-cell spread, as S. flexneri 2a wzzSF : : KmR mutants are unable to form plaques on HeLa cell monolayers, have an F-actin comet tail formation defect and have reduced levels of cell surface IcsA (also called VirG) protein (Hong & Payne, 1997; Van Den Bosch et. al., 1997). IcsA is an outer-membrane protein that is essential for initiating ABM (Bernardini et al., 1989; Lett et al., 1989; Goldberg, 2001; Suzuki & Sasakawa, 2001). S. flexneri 2a additionally produces LPS with very long (VL) type Oag chains with a modal length of greater than 90 RUs; these are determined by the WzzpHS2 (CldpHS-2) protein, which is encoded by a small plasmid called pHS-2 (Stevenson et al., 1995). The VL-type Oag chains confer resistance to the bactericidal activity of serum and enhance the Sereny reaction in mice but are not needed for IcsA function in ABM (Hong & Payne, 1997).

In this study, we investigated the role of LPS Oag modal chain length distribution in S. flexneri virulence by assessing its impact on a number of in vivo and in vitro correlates of virulence. In addition, we investigated IcsA localization on the cell surface of intracellular S. flexneri bacteria. Our data suggest that Oag modal chain length distribution has been optimized for S. flexneri virulence, and we hypothesize that Oag chains can mask IcsA and affect its function in ABM.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Growth media and conditions.
Strains were routinely grown in Luria–Bertani (LB) broth and on LB agar as described previously (Van Den Bosch et al., 1997). Unless otherwise stated, S. flexneri was grown from a Congo-Red-positive colony (Van Den Bosch et al., 1997) in LB for 16 h with aeration, then diluted 1 in 50 into fresh LB and grown for 2 h with aeration. Under these growth conditions, the IcsA protein was in the intact 116 kDa form as determined by Western immunoblotting. Detection of the secreted 85–95 kDa {alpha}-domain of IcsA in the culture supernatant required 4–6 h growth (data not shown). Antibiotics were used at the following concentrations: 100 µg ampicillin ml-1; 25 µg chloramphenicol ml-1; 50 µg kanamycin ml-1; 10 µg tetracycline ml-1. Unless stated otherwise, strains were grown at 37 °C.

Bacterial strains and plasmids.
The bacterial strains and plasmids used in this study are described in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1. Bacterial strains and plasmids used in this study

 
Mini-Tn5-CmR mutagenesis.
Strains RMA961 and RMA696 were grown to mid-exponential phase, centrifuged (5000 r.p.m., 10 min, IEC Centra 4X centrifuge), resuspended in LB, mixed together at a ratio of 1 : 10 (0·3 and 3·0 ml), centrifuged as above, resuspended in 300 µl LB, plated onto a Millipore HA membrane filter overlaid on a pre-warmed (37 °C) LB agar plate and incubated for 6 h at 37 °C. The bacterial mating mix was harvested into LB, then plated onto selective medium (LB agar containing kanamycin and chloramphenicol). RMA696 colonies having random insertions of mini-Tn5-CmR were harvested, pooled and stored at -70 °C in an 80 % (v/v) glycerol/Bactopeptone (Difco) mix.

DNA methods and sequencing.
General DNA manipulation, plasmid DNA isolation, dye terminator DNA sequencing, PCR, DNA transformation and electroporation were performed as described previously (Baker et al., 1999; Daniels & Morona, 1999). The location of mini-Tn5-CmR in pRMCD72 was determined by digestion with SmaI, and cloning three SmaI fragments (~2·8, ~1·95 and ~0·4 kb) into the SmaI site of pBluescript-SK (Stratagene). DNA sequencing using dye-labelled M13 forward and reverse primers was then undertaken.

Protein analysis.
Analysis of IcsA proteins by SDS-PAGE and Western immunoblotting was performed as described previously (Morona et. al., 1995; Daniels & Morona, 1999). After transfer to a nitrocellulose membrane and incubation with a rabbit anti-IcsA antibody, bound antibody was detected using horseradish peroxidase-conjugated goat antibodies specific for rabbit immunoglobulins and enhanced chemiluminescence (Roche). Polyclonal rabbit anti-IcsA serum was used at 1 : 1000. The molecular mass markers used in this study were pre-stained (New England Biolabs; catalogue no. P7708S).

LPS analysis.
LPS was prepared by proteinase K treatment of whole-cell lysates, electrophoresed on SDS 15 or 20 % polyacrylamide gels and detected by silver-staining as described previously (Morona et al., 1991; Daniels & Morona, 1999).

Construction of pRMCD76.
Plasmid pRMCD77, a derivative of pET17b (Novagen) but with its BglII site removed by end-filling with dNTPs and Klenow, was the base vector used for cloning and expression of wzz genes (Daniels & Morona, 1999). pRMCD76, harbouring the wzzO139 gene (also called otnB) from Vibrio cholerae O139, was constructed as follows. A 1519 bp EcoRI–BglII fragment (corresponding to nucleotides 10654–12173 of the otn gene region, GenBank accession no. X90547) was isolated from pBSotn and ligated to EcoRI- and BamHI-digested pRMCD77. After transformation into Escherichia coli DH5{alpha}, plasmids from ampicillin-resistant colonies were screened by restriction enzyme digestion, resulting in the isolation of pRMCD76. DNA sequencing using T7 promoter and T7 terminator primers confirmed that the clone had the wzzO139 ORF.

Cloning, overexpression and purification of IcsA.
Plasmid pRMA920 has the icsA coding region under IPTG-inducible control and was used to overexpress IcsA. Oligonucleotides #2156 (5'-GCCG AAT TCA ACG GAA TCT TTT CAG GGG-3', EcoRI site) and #2170 (5'-GCG GAT CCC CAT GTG ATT TGC CTC G-3', BamHI site) were used in a PCR with pD10; the resulting product was purified (Qiagen), digested with EcoRI and BamHI, and ligated to likewise digested pTTQ181. After transformation into JM105, screening of ampicillin-resistant colonies resulted in pRMA920. Plasmid pRMA920 was electroporated into RMA763, resulting in RMA943. IcsA was overexpressed and purified as follows. A 1 l culture of RMA943 in LB with ampicillin was incubated with aeration at 37 °C; at an OD600 value of 1·0, IPTG was added to a final concentration of 0·4 mM, and the culture was incubated for a further 3 h. The cells were collected by centrifugation at 7000 r.p.m. (10 min, 4 °C, Beckman JA-10 rotor), resuspended in 40 ml of 20 % (w/v) sucrose/30 mM Tris/HCl (pH 8·1) and incubated on ice for 30 min. The cells were centrifuged as above, and the cell pellet was frozen in an ethanol/dry-ice bath for 30 min, thawed and dispersed vigorously in 3 ml of 3 mM EDTA (pH 7·3). Cells were lysed with a Branson ultra-sonifier (50 % cycle, intermittent), and by several cycles of freeze–thawing. Unlysed cells and inclusion bodies were removed by slow-speed centrifugation (7000 r.p.m., 10 min, 4 °C) and the supernatant containing whole-cell membranes and cytoplasm was centrifuged at 35 000 r.p.m. using a 80 Ti rotor for 90 min at 20 °C in a Beckman L8-80 ultracentrifuge. The pellet was resuspended in 1 ml of water, an equal volume of Triton solution [4 % (v/v) Triton X-100, 2 mM MgCl2, 50 mM Tris, pH 7·5] was added and the sample was vortex-mixed intermittently for 30 min at room temperature. The sample was then centrifuged at 35 000 r.p.m. for 90 min at 20 °C. The resulting pellet was mixed with 2 % (v/v) Triton X-100 in 10 mM Tris (pH 8·0)/8 M urea, and centrifuged at 18 000 r.p.m. for 20 min (Beckman centrifuge). A Rosenbusch-type extraction was then performed (Rosenbusch, 1974). The pellet was resuspended in SDS-PAGE sample buffer (Lugtenberg et al., 1975) without bromophenol blue and incubated at 37 °C for 40 min. The sample was centrifuged in a Beckman centrifuge at 15 000 r.p.m. for 60 min, at room temperature. The resulting supernatant was then applied to an S-200 Sephacryl column equilibrated with 10 mM Tris (pH 7·4)/2 % (w/v) SDS/5 mM EDTA/0·02 % (w/v) sodium azide, and the column was eluted with the same buffer. Fractions were collected and analysed on an SDS 12 % polyacrylamide gel. Under these conditions, IcsA was eluted close to the void volume. Fractions containing only the ~120 kDa IcsA protein were pooled; the final preparation was free of any other protein, as determined by Coomassie blue staining.

Preparation of rabbit anti-IcsA serum.
IcsA, isolated by column chromatography, was further purified by two rounds of SDS-PAGE. The gel slice containing IcsA was homogenized in Freund's complete adjuvant and used to immunize a rabbit by injection at multiple sites. The resulting antiserum was adsorbed four times with RMA763 bacterial cells. After the addition of an equal volume of sterile glycerol (100 %), antibodies were stored at -20 °C. Western blotting against S. flexneri 2457T whole cells, trichloracetic acid (8 %, w/v, final) precipitated culture supernatant and whole-cell membranes showed that the antiserum only detected the amino terminal {alpha}-domain of IcsA (data not shown).

Indirect immunofluorescence detection of IcsA on LB-grown bacteria.
The technique is a minor modification of a method by Klauser et al. (1990). Round glass coverslips were boiled for 1 min in 0·1 M HCl before storage in 95 % ethanol prior to air-drying for use. Poly-L-lysine (Sigma, catalogue no. P8920) [100 µl, 0·01 % (w/v) in H2O] was pipetted onto coverslips placed in a 24-well, flat-bottomed tissue culture tray (Costar) and incubated at room temperature for 5 min before being washed with PBS. Cells were formalin-killed by treating 10 ml of pelleted culture (5000 r.p.m., 10 min, IEC Centra 4X centrifuge) with 1 ml of 2 % (w/v) paraformaldehyde (Sigma) for 20 min. The cells were washed twice in PBS and resuspended in a final volume of 1 ml PBS. One-hundred microlitres of cells were added to each well, then centrifuged at 800 r.p.m. for 10 min (Hereaus Labofuge 400R). The wells were aspirated dry and washed three times with PBS. The primary antibody (1 : 100) in PBS with 10 % (w/v) fetal calf serum (FCS; IMVS, Adelaide) was added and incubated at 37 °C for 60 min. After three washes with PBS, the secondary antibody (1 : 80) (FITC-conjugated goat anti-rabbit antibody; Silenus, catalogue no. RDAF) in PBS with 10 % (w/v) FCS was added and left at 37 °C for 30 min, followed by a further two PBS washes. The coverslips were gently aspirated dry and mounted upside down on a clean glass microscope slide with 3 µl of Mowiol 4-88 (Calbiochem, catalogue no. 475904) containing 20 µg p-phenylenediamine ml-1 (Sigma, catalogue no. P6001) and sealed with acrylic nail polish. Bacteria were photographed with Kodak TMAX400 film using an Olympus B2 microscope equipped with phase-contrast and epifluorescence illumination, standard FITC filters and a 100xachromatic oil immersion lens.

Virulence assays.
The ability of S. flexneri strains to form plaques on HeLa cell (laboratory stock) and BHK-26 cell (IMVS) monolayers was performed exactly as described by Oaks et al. (1985). The Sereny keratoconjunctivitis test was performed using guinea pigs, as directed by the Animal Ethics Committee of the University of Adelaide, and the severity of the reaction was scored according to Sereny reaction ratings (Sereny, 1957; Formal et al., 1958).

Invasion of HeLa cells and immunofluorescence (IF) staining.
HeLa cells were infected as follows. An 18 h-old culture of an S. flexneri strain was diluted 1 in 50 in LB (10 ml), grown for 1·5 h, centrifuged at 5000 r.p.m. for 10 min (IEC Centra 4X centrifuge), washed and resuspended at approximately 109 bacteria ml-1 in D-PBS [PBS with 0·1 % (w/v) CaCl2/0·1 % (w/v) MgCl2]. HeLa cells, seeded on glass coverslips in a 24-well tray and grown to give semi-confluence, were washed twice with D-PBS, overlaid with the 100 µl bacterial suspension and centrifuged at 2000 r.p.m. for 10 min (Hereaus Labofuge 400R, 8177 rotor). Bacterial entry was allowed for 60 min incubation at 37 °C in a humidified CO2 incubator (5 % CO2). Infected cells were washed with D-PBS three times and incubated with 0·5 ml MEM (Gibco) containing 50 µg gentamicin ml-1 (Gibco) for a further 2 h at 37 °C in a CO2 incubator, to allow intracellular spread. For IF staining, two different fixation methods were used. To detect F-actin comet tails, infected cells were fixed for 10 min in 3·7 % (w/v) paraformaldehyde in PBS and then permeabilized with 0·2 % (v/v) Triton X-100 in PBS for 1 min. To detect IcsA on the cell surface of intracellular bacteria, infected cells were fixed for 5 min at room temperature with 80 % acetone, incubated with PBS for 1 min, then permeabilized with 0·1 % (v/v) Triton X-100 in PBS for 1 min. After blocking in 1 % (w/v) FCS in PBS for 10 min, the infected cells were incubated at 37 °C for 30 min with either polyclonal anti-Shigella LPS (Denka Seiken, Japan, catalogue no. 310 111) (1 in 100) or rabbit anti-IcsA antiserum (1 in 100) as described above. After washing three times in PBS, coverslips were incubated for 15 min at 37 °C with either FITC-conjugated goat anti-rabbit or Texas-red-conjugated goat anti-rabbit (Amersham, catalogue no. N2034) secondary antibodies (1 in 100) as required. F-actin was visualized by staining with FITC–phalloidin (0·1 µg ml-1; Sigma, catalogue no. P5282), and propidium iodide (10 µg ml-1; Sigma, catalogue no. P4170) was used to counter-stain bacteria and cellular nuclei; FITC–phalloidin and propidium iodide were included with the secondary antibody incubation as required. The coverslips were mounted as described above. Preparations were examined with a Bio-Rad MRC-600 confocal laser-scanning microscope, using an Olympus IMT-2 inverted microscope and a 100xoil immersion, apochromatic, objective lens. FITC and Texas red images were collected simultaneously, and false-colour-merged using CONFOCAL ASSISTANT 4.02. Each image shown is from a single plane.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Isolation of plaque-forming, extrageneic suppressor mutants of RMA696
We have previously reported that a wzzSF : : KmR derivative of the virulent S. flexneri 2a strain 2457T (RMA696) is unable to form plaques on HeLa cells (Van Den Bosch et al., 1997). This strain has reduced levels of IcsA on its cell surface that correlate with its inability to form plaques on HeLa cells and to form F-actin comet tails. We reasoned that as a result of the wzzSF : : KmR mutation RMA696 had no absolute defect in plaque formation but that in some way it was limited in a factor(s) required for efficient plaque formation. To identify this factor, we isolated plaque-forming, extrageneic suppressor mutants of RMA696. RMA696 was randomly mutagenized with mini-Tn5-CmR. Several thousand transposon insertion mutants were pooled, then plated onto BHK-26 cell monolayers in a plaque assay. These cells were used because S. flexneri 2457T forms larger plaques on these cell monolayers compared to HeLa cell monolayers (Fig. 1). Plaques were observed at a frequency of 10-4 input bacteria. Bacteria were isolated from these plaques, purified and plated once more onto BHK-26 cell monolayers. The plating efficiency and plaque size were identical to that of the original parent strain 2457T (data not shown). The isolates were also able to plaque efficiently on HeLa cell monolayers (Fig. 1, and data not shown). One isolate (RMA982) was chosen for further characterization.



View larger version (47K):
[in this window]
[in a new window]
 
Fig. 1. Plaque formation by S. flexneri strains. The relative size and morphology of the plaques formed on HeLa and BHK-26 cell monolayers by the indicated S. flexneri strains are shown. Strains: 2457T (S. flexneri 2a wild-type); RMA696 (2457T wzzSF : : KmR); RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR); RMA3107 [RMA696(pRMCD77)]; RMA3108 [RMA696(pRMCD76)]; RMA3109 [RMA696(pRMCD78)]; RMA3110 [RMA696(pRMCD80)]. Strains RMA3232 [RMA982(pRMCD77)], RMA3231 [RMA982(pRMCD76)], RMA3233 [RMA982(pRMCD78)] and RMA3234 [RMA982(pRMCD80)] were identical to RMA982 with respect to plaque formation (data not shown). Arrowheads indicate plaques.

 
Characterization of RMA982
We analysed the LPS of strain RMA982 by SDS-PAGE and silver-staining (Fig. 2). The LPS of RMA982 has Oag chains of random length distribution, having no modal length regulation, with the majority of the Oag chains being less than 30 RUs in length. The Oag chains were shorter than those produced by the parent strain RMA696 (VL-type Oag) (Fig. 2). Significantly, RMA982 lacked LPS with VL-type Oag chains as produced by the parent strains; this suggested that in RMA982 mini-Tn5-CmR had inserted into the cld gene of pHS-2 (Stevenson et al., 1995). Plasmid DNA was prepared from RMA982 and transformed into E. coli DH5{alpha}; chloramphenicol-resistant colonies were obtained. Plasmid DNA (pRMCD72) from one isolate was subjected to restriction enzyme analysis that indicated that the 3·0 kb pHS-2 plasmid had an insertion of 5·3 kb corresponding to mini-Tn5-CmR (data not shown). Subcloning and DNA sequencing revealed that mini-Tn5-CmR was indeed inserted in pHS-2 between nucleotides 953 and 954 (GenBank accession no. M25995). This placed mini-Tn5-CmR inside the pHS-2-cld gene (herein termed wzzpHS2) ORF, disrupting translation after Thr-63. This result is consistent with the absence of LPS with VL-type Oag chains from RMA982. Furthermore, our observations are consistent with and confirm the results of Hong & Payne (1997), who found that S. flexneri 2a strain SA514rol with mutations in both wzzSF and pHS-2-cld was able to form plaques on HeLa cells. Consequently, we used strain RMA982 for further studies on the effects of LPS Oag modal length on virulence and related properties.



View larger version (34K):
[in this window]
[in a new window]
 
Fig. 2. LPS of a plaque-forming, extrageneic suppressor mutant of RMA696. LPS samples were prepared as described in Methods. Lanes: 1, 2457T; 2, RMA696 (2457T wzzSF : : KmR); 3, RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR). The positions of various types of LPS molecules are indicated on the left-hand side of the figure. R-LPS, rough LPS. Oag RU number is indicated on the right-hand side of the figure. Samples represent approximately 2x108 bacterial cells (except RMA982, 1x108 cells).

 
Construction of S. flexneri strains with different LPS Oag modal chain lengths
The data presented above, and those presented previously (Hong & Payne, 1997; Van Den Bosch et al., 1997), suggest that if S. flexneri only produces LPS with VL-type Oag chains then its ability to form plaques, and hence spread from cell to cell, is abrogated. Conversely, and paradoxically, strains with double mutations in genes known to regulate Oag modal chain length, resulting in LPS with Oag chains whose lengths are shorter than those observed in the parent strains, were able to plaque on HeLa and BHK-26 cells. We investigated the effect of varying the modal chain length on the ability of S. flexneri to form plaques on HeLa and BHK-26 cells. This was accomplished by complementation using the Salmonella enterica serovar Typhimurium wzzST gene which confers an LPS Oag modal chain length (L-type, 19–35 RUs) that is longer than that conferred by Shigella flexneri wzzSF (Daniels & Morona, 1999), and the otnB gene (herein termed wzzO139) from V. cholerae O139 whose LPS Oag has only one or a few RUs (Bik et. al., 1995, 1996).

Plasmids harbouring wzzO139 (pRMCD76), wzzSF (pRMCD78), wzzST (pRMCD80) and the control vector (pRMCD77) were electroporated into RMA696 (2457T wzzSF : : KmR) and RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR). The LPS molecules produced by the resulting strains were analysed by SDS-PAGE and silver-staining (Fig. 3), and the Oag modal chain lengths observed are summarized in Table 2. As expected, pRMCD78 (wzzSF) complemented the wzzSF : : KmR mutation in RMA982 and RMA696, resulting in LPS with either one (S-type) or two (S- and VL-type) Oag modal chain lengths, respectively, with the latter profile being almost identical to that of the parental strain 2457T (Fig. 3a, lanes 1 and 5). Strains RMA982 and RMA696 harbouring pRMCD80 (wzzST) produced LPS with either one (L-type) or two (L- and VL-type) Oag modal lengths, respectively, as expected (Daniels & Morona, 1999). RMA982 and RMA696 harbouring pRMCD76 (wzzO139) produced LPS with either one [very short (VS) type, 2–7 RUs] or two (VS- and VL-type) Oag modal lengths, respectively (Fig. 3, Table 2). One notable difference between the strains was that RMA3108 [RMA696(pRMCD76), wzzO139] (Fig. 3a, lane 3) and RMA3231 [RMA982(pRMCD76), wzzO139] (Fig. 3b, lane 3) had reduced levels of LPS with a single Oag RU (semi-rough LPS), and of LPS with VL-type Oag chains (RMA3108 only) compared with wzzSF- and wzzST-complemented strains.



View larger version (97K):
[in this window]
[in a new window]
 
Fig. 3. Effect of different Wzz proteins on LPS profiles. LPS samples were prepared as described in Methods. (a) Lanes: 1, 2457T; 2, RMA696 (2457T wzzSF : : KmR); 3, RMA3108 [RMA696(pRMCD76)]; 4, RMA3107 [RMA696(pRMCD77)]; 5, RMA3109 [RMA696(pRMCD78)]; 6, RMA3110 [RMA696(pRMCD80)]. (b) Lanes: 1, 2457T; 2, RMA696 (2457T wzzSF : : KmR); 3, RMA3231 [RMA982(pRMCD76)]; 4, RMA3232 [RMA982(pRMCD77)]; 5, RMA3233 [RMA982(pRMCD78)]; 6, RMA3234 [RMA982(pRMCD80)]. The positions of the various types of LPS are indicated on the left-hand side of (a) and (b). R-LPS, rough LPS; SR-LPS, semi-rough LPS. Oag RU number is indicated on the right-hand side of (a) and (b). Samples represent approximately 2x108 bacterial cells.

 

View this table:
[in this window]
[in a new window]
 
Table 2. Summary of LPS phenotypes, ability of strains to plaque on HeLa cells and Sereny assay results

 
Effect of LPS Oag modal chain length variation on HeLa plaque formation
Strains 2457T, RMA696 and RMA982, and derivatives carrying either control vector or plasmids with different wzz genes, were used to infect HeLa and BHK-26 cell monolayers to investigate the impact of different LPS Oag modal length combinations on plaque-forming ability. The results are shown in Fig. 1, and are summarized in Table 2. As expected, RMA3107 [RMA696(pRMCD77), control] did not form plaques, and RMA3109 [RMA696(pRMCD78), wzzSF] was able to form plaques; both the plaque size and efficiency of plaque formation were the same as that of 2457T (data not shown). Strain RMA3108 [RMA696(pRMCD76), wzzO139] formed plaques on HeLa and BHK-26 cells monolayers, showing that the VS-type Oag chains produced by this strain were permissive for plaque formation. Strain RMA3110 [RMA696(pRMCD80), wzzST] was unable to form plaques on HeLa and BHK-26 cells monolayers; no micro-plaques were observed and no effect on the cell monolayers was observed even at high multiplicities of infection. This showed that the L-type Oag chains produced by RMA3110 were, in some way, not permissive for plaque formation. All RMA982-based strains (lacking the VL-type Oag chains) were able to form plaques on HeLa and BHK-26 cell monolayers (Fig. 1, Table 2). Therefore, in the absence of the WzzpHS2-determined VL-type Oag chains, LPS Oag modal lengths ranging from VS to L were permissive for plaque formation by S. flexneri. Furthermore, a defined Oag modal length is not actually required, as the random length Oag chains of RMA982 LPS were permissive for plaque formation. These results indicate that the inability of strain RMA696 (and RMA3110) to form plaques is due to the presence of LPS with VL-type Oag chains.

Effect of LPS Oag modal chain length variation on Sereny reaction
To confirm the impact of LPS Oag modal length on S. flexneri virulence, we compared the abilities of strains 2457T, RMA696 (2457T wzzSF : : KmR), RMA3107 [RMA696(pRMCD77), control], RMA3108 [RMA696(pRMCD76), wzzO139], RMA3109 [RMA696(pRMCD78), wzzSF], RMA3110 [RMA696(pRMCD80), wzzST] and RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR) to provoke keratoconjunctivitis in the guinea pig eye (Sereny assay) (Table 2). Strains RMA696, RMA3107 and RMA3110 gave a reduced Sereny positive reaction, correlating with their inability to from plaques on HeLa cell monolayers, while strains RMA3108 and RMA3109 were fully virulent. RMA982 was also highly virulent in this assay and was identical to 2457T. Overall, the severity of the Sereny reaction correlated with the strains' plaque-forming ability. These results highlight the importance of LPS Oag modal length distribution for S. flexneri virulence.

Infection of HeLa cells and F-actin comet tail formation
We investigated if the LPS Oag modal length distribution had any affect on S. flexneri intracellular spread within HeLa cells, by comparing the abilities of RMA696 (2457T wzzSF : : KmR), RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR) and derivatives carrying either the control vector or plasmids encoding different wzz genes to form F-actin comet tails. The latter is indicative of the ability of IcsA to function in ABM. In particular, we wanted to determine if the reduced virulence and ability to plaque on HeLa cell monolayers shown by RMA3110 [RMA696(pRMCD80), wzzST] correlated with a defect in IcsA function.

HeLa cells were infected with 2457T, RMA696 (2457T wzzSF : : KmR), RMA3107 [RMA696(pRMCD77), control], RMA3108 [RMA696(pRMCD76), wzzO139], RMA3109 [RMA696(pRMCD78), wzzSF], RMA3110 [RMA696(pRMCD80), wzzST], RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR), RMA3232 [RMA982(pRMCD77), control], RMA3231 [RMA982(pRMCD76), wzzO139], RMA3233 [RMA982(pRMCD78), wzzSF], RMA3234 [RMA982(pRMCD80), wzzST] and RMA2041 (2457T {Delta}icsA : : TcR), which was included as a negative control. All strains tested were able to invade HeLa cells with similar efficiency (data not shown). 2457T formed F-actin comet tails (approx. 20 % of bacteria) (Fig. 4a) and as expected, RMA2041 (Fig. 4b) did not form F-actin comet tails nor did it show any intracellular spread. RMA696, RMA3107 and RMA3110 (Fig. 4c, e, h) rarely formed F-actin comet tails (<1 % of bacteria) but did exhibit some intracellular spread when compared to the IcsA-null mutant RMA2041, which formed a micro-colony clustered near the nucleus (Fig. 4b). RMA3108, RMA3109, RMA982 and all RMA982 derivatives were able to form F-actin comet tails (approx. 20 % of bacteria) (Fig. 4f, g, i, j, k, l). These data suggest that the inability of RMA3110 (like RMA696) to form plaques was a consequence of the IcsA in this strain being unable to efficiently nucleate actin polymerization and thus initiate ABM.



View larger version (146K):
[in this window]
[in a new window]
 
Fig. 4. Detection of F-actin comet tails produced by S. flexneri growing within HeLa cells. HeLa cells infected with S. flexneri strains with various LPS Oag modal lengths and stained to detect F-actin comet tails are shown. S. flexneri cells were detected by indirect IF staining with a rabbit anti-Oag and a Texas-red-conjugated secondary antibody (red); F-actin comet tails were detected by staining with FITC–phalloidin (green) (see Methods). (a) 2457T; (b) RMA2041 (2457T {Delta}icsA : : TcR); (c) RMA696 (2457T wzzSF : : KmR); (d) RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR); (e) RMA3107 [RMA696(pRMCD77)]; (f) RMA3108 [RMA696(pRMCD76)]; (g) RMA3109 [RMA696(pRMCD78)]; (h) RMA3110 [RMA696(pRMCD80)]; (i) RMA3232 [RMA982(pRMCD77)]; (j) RMA3231 [RMA982(pRMCD76)]; (k) RMA3233 [RMA982(pRMCD78)]; (l) RMA3234 [RMA982(pRMCD80)]. Arrowheads indicate the locations of F-actin comet tails.

 
Effect of Oag modal chain length variation on IcsA production by LB-grown bacteria
We have previously reported that detection of IcsA on the S. flexneri cell surface was affected by wzzSF mutation, although no effect on IcsA production was observed (Van Den Bosch et al., 1997). We used Western immunoblotting to compare IcsA production by 2457T, RMA696 (2457T wzzSF : : KmR) and RMA696 derivatives carrying either the cloning vector or plasmids containing different wzz genes. The results show that all strains had similar levels of the 116 kDa IcsA protein (Fig. 5). Similar results were obtained for RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR) and RMA982 derivatives carrying either the control vector or plasmids containing different wzz genes (data not shown). Hence, the inability of RMA696, RMA3107 [RMA696(pRMCD77), control] and RMA3110 [RMA696(pRMCD80), wzzST] to form F-actin comet tails is unlikely to be due to an effect on IcsA production.



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 5. Detection by Western immunoblotting of IcsA production by S. flexneri strains. The S. flexneri strains indicated were grown in LB to early exponential phase, whole-cell lysates were prepared, electrophoresed on an SDS 12 % polyacrylamide gel and, after transfer to nitrocellulose, proteins were detected with an anti-IcsA antibody. Lanes: 1, 2457T; 2, RMA696 (2457T wzzSF : : KmR); 3, RMA3107 [RMA696(pRMCD77)]; 4, RMA3108 [RMA696(pRMCD76)]; 5, RMA3109 [RMA696(pRMCD78)]; 6, RMA3110 [RMA696(pRMCD80)]. The position of the 116 kDa IcsA protein is shown. Lower molecular mass bands are IcsA-derived products arising from either degradation or premature translation termination, and have been reported by others (Fukuda et al., 1995). Lanes contained approximately 2x108 bacterial cells.

 
We then investigated the cell surface expression of IcsA by S. flexneri strains having LPS with different Oag modal chain length distributions (grown to early-exponential phase in LB) by indirect IF detection with an anti-IcsA antibody. Polarly localized IcsA could be detected on the cell surfaces of 2457T, RMA696, RMA3107, RMA3108 [RMA696(pRMCD76), wzzO139], RMA3109 [RMA696(pRMCD78), wzzSF] and RMA3110 bacteria (approx. 40–60 % of bacteria) (Fig. 6a, c, e, g, i, k), although for RMA696 and RMA3107 relatively fewer bacteria in the cultures (approx. 20 % of bacteria) had detectable IcsA, a result similar to that reported for RMA696 (Van Den Bosch et al., 1997). Strains RMA982, RMA3232 [RMA982(pRMCD77), control], RMA3231 [RMA982(pRMCD76), wzzO139], RMA3233 [RMA982(pRMCD78), wzzSF] and RMA3234 [RMA982(pRMCD80), wzzST] (Fig. 6d, f, h, j, l) had polarly localized IcsA (approx. 50 % of bacteria) and were labelled more intensely than the corresponding wzzpHS2-positive strains. Thus, under the growth, fixation and labelling conditions used, the cellular localization of IcsA was apparently unaffected by Oag modal length distribution; however, strains unable to produce LPS with VL-type Oag chains were stained more intensely.



View larger version (84K):
[in this window]
[in a new window]
 
Fig. 6. (on facing page) Detection of IcsA on the cell surface of LB-grown S. flexneri strains. The indicated S. flexneri strains were grown to early exponential phase in LB, formalin-fixed and immunostained to detect IcsA on the cell surface with rabbit anti-IcsA antibodies and an FITC-conjugated secondary antibody (see Methods). Strains: (a) 2457T; (b) RMA2041 (2457T {Delta}icsA : : TcR); (c) RMA696 (2457T wzzSF : : KmR); (d) RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR); (e) RMA3107 [RMA696(pRMCD77)]; (f) RMA3232 [RMA982(pRMCD77)]; (g) RMA3108 [RMA696(pRMCD76)]; (h) RMA3231 [RMA982(pRMCD76)]; (i) RMA3109 [RMA696(pRMCD78)]; (j) RMA3233 [RMA982(pRMCD78)]; (k) RMA3110 [RMA696(pRMCD80)]; (l) RMA3234 [RMA982(pRMCD80)]. The specificity of the anti-IcsA antibody used was shown by the lack of labelling of the icsA mutant strain RMA2041 (b). In (a–l), the phase-contrast image is on the left-hand side and the IF image is on the right-hand side.

 
Detection of IcsA on the surface of intracellular bacteria
The detection of IcsA on the cell surface of LB-grown S. flexneri strains did not correlate with their Oag modal length distribution, their ability to form plaques and F-actin comet tails or their virulence. We investigated if there was a better correlation between the LPS Oag modal length distribution and cell surface IcsA by examining IcsA production by S. flexneri bacteria localized within HeLa cells. We modified our standard fixation procedure to detect IcsA production by intracellular S. flexneri bacteria, since the method used for LB-grown bacteria was unsuccessful. HeLa cells were infected with 2457T, RMA696 (2457T wzzSF : : KmR), RMA3107 [RMA696(pRMCD77), control], RMA3108 [RMA696(pRMCD76), wzzO139], RMA3109 [RMA696(pRMCD78), wzzSF], [RMA696(pRMCD80), wzzST], RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR), RMA3232 [RMA982(pRMCD77), control], RMA3231 [RMA982(pRMCD76), wzzO139], RMA3233 [RMA982(pRMCD78), wzzSF] and RMA3234 [RMA982(pRMCD80), wzzST], and the negative control RMA2041 (2457T {Delta}icsA : : TcR). IcsA was detected, albeit weakly, at the cell pole on the cell surface of strains 2457T, RMA696, RMA3107, RMA3108, RMA3109 and RMA3110 (approx. 50–80 % of bacteria) (Fig. 7a, c, e, f, g), although, as for LB-grown bacteria, RMA696 and RMA3107 were stained slightly less intensely than the other strains. This result does not correlate with ability of these strains to form plaques and F-actin comet tails. In contrast, IcsA was readily detected on the cell surface of RMA982, RMA3232, RMA3231, RMA3233 and RMA3234 (approx. 90–100 % of bacteria), and with the exception of RMA3234 (Fig. 7l), labelling of these strains was more intense than for 2457T (Fig. 7, compare a with d, i, j, k). Strains RMA982, RMA3232 and RMA3231 had IcsA located on lateral and polar regions of the cell surface (Fig. 7d, i, j), while RMA3233 and RMA3234 had IcsA at the cell pole, although it could also be weakly detected on lateral regions of RMA3233 (Fig. 7k, l). For intracellular bacterial strains unable to produce LPS with VL-type Oag chains (RMA982 background), there was an inverse correlation between detection of IcsA and LPS Oag modal length. These results contrast with the results obtained with LB-grown bacteria (Fig. 6).



View larger version (118K):
[in this window]
[in a new window]
 
Fig. 7. Detection of IcsA on the cell surface of S. flexneri strains growing within HeLa cells. HeLa cells were infected with S. flexneri strains with various LPS Oag modal lengths and then fixed with acetone. IcsA was detected by indirect IF staining with rabbit anti-IcsA and an FITC-conjugated secondary antibody (green); S. flexneri cells were detected by counter-staining with propidium iodide (red) (see Methods). Strains: (a) 2457T; (b) RMA2041 (2457T {Delta}icsA : : TcR); (c) RMA696 (2457T wzzSF : : KmR); (d) RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR); (e) RMA3107 [RMA696(pRMCD77)]; (f) RMA3108 [RMA696(pRMCD76)]; (g) RMA3109 [RMA696(pRMCD78)]; (h) RMA3110 [RMA696(pRMCD80)]; (i) RMA3232 [RMA982(pRMCD77)]; (j) RMA3231 [RMA982(pRMCD76)]; (k) RMA3233 [RMA982(pRMCD78)]; (l) RMA3234 [RMA982(pRMCD80)]. The specificity of the anti-IcsA antibody used was shown by the lack of labelling of the icsA mutant strain RMA2041 (b). Propidium iodide was used as a counter-stain to detect both the bacteria and the HeLa cell nuclei (seen here as large, oval-shaped, red-staining bodies). This avoided any potential problems with anti-Oag antibodies interfering with detection of IcsA by anti-IcsA antibodies. Within (a–l), an enlargement of a typical bacterium is shown.

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
We isolated and characterized strain RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR), a plaque-forming, virulent derivative of RMA696 (2457T wzzSF : : KmR), and found it to be similar to S. flexneri strain SA514rol, a wzzSF : : KmR wzzpHS2 : : CmR double mutant characterized by Hong & Payne (1997). While RMA982 was fully virulent in a guinea-pig-based Sereny assay, SA514rol gave a weak reaction in the mouse-based Sereny test (Hong & Payne, 1997). This difference may be due to the S. flexneri strain used and/or test animal differences. Paradoxically, while S-type modal length Oag chains seem to be needed for intracellular spreading and virulence (Hong & Payne, 1997; Van Den Bosch et al., 1997), the LPS molecules of RMA982 and SA514rol had random length Oag chains (Hong & Payne, 1997; this study). Further experiments resolved this paradox, and suggested that S. flexneri 2a Oag modal chain length has been optimized during evolution of the bacterium into a human pathogen.

We used plasmids carrying wzzO139, wzzSF and wzzST to complement the wzz mutations in RMA696 and RMA982. The ability of the V. cholerae wzzO139 gene (carried on pRMCD76) to fully complement the S. flexneri wzzSF : : KmR mutation, resulting in a novel LPS with a VS-type Oag modal chain length, was surprising since the WzzO139 protein has a low level of sequence identity with WzzSF and functions in V. cholerae O139 capsule polysaccharide (CPS) biosynthesis (Bik et. al., 1995, 1996; Morona et al., 2000; Attridge et al., 2001). Confirming this function, pRMCD76 was recently shown to complement a V. cholerae O139 otnB (wzzO139) mutant and restored production of O139 CPS (Attridge et al., 2001). While the ability of wzzO139 to complement the S. flexneri wzzSF : : KmR defect and regulate LPS Oag chain length clearly shows that WzzO139 can act as a Wzz protein, its role in V. cholerae O139 LPS/capsule biosynthesis may be altered such that it is involved in O139 CPS biosynthesis rather than LPS Oag synthesis.

Like RMA982, the strains that were unable to produce the VL-type Oag chains formed plaques and F-actin comet tails regardless of their LPS Oag modal length status (Table 2). Different results were obtained when the complemented strains were able to produce the VL-type Oag chains. In this situation, VS- and S-type Oag chains but not L-type Oag chains were permissive for IcsA function in ABM. Our data, combined with those of Sandlin et al. (1996), who found that semi-rough LPS with a single Oag RU does not support plaque formation, show that IcsA function in ABM is compatible with and requires LPS with at least two Oag RUs but less than 90 RUs, a relatively broad range. It should be noted that we did not test Oag modal lengths between 35 and 90 Oag RUs. In the presence of LPS with VL-type Oag chains this requirement becomes more stringent, and the LPS must have Oag chains with two to less than 18 RUs.

While unknown, different types of LPS molecules may be randomly distributed on the S. flexneri cell surface. Alternatively, rough LPS molecules (LPS lacking Oag chains) and LPS molecules with particular Oag chain lengths may be preferentially localized to certain regions of the cell (e.g. polar versus lateral regions), and there might be a preferred association between LPS molecules of different size and particular outer-membrane proteins. Based on our observations, LPS with VL-type Oag chains in the absence of LPS with S-type Oag chains (as in RMA696 and RMA3107) place IcsA molecules in an environment where they cannot function, perhaps due to the VL-type Oag chains masking and/or interfering with the interaction of IcsA with host-cell proteins (see below). While the WzzST-determined L-type Oag chains on their own are permissive for IcsA function, unlike the VS- and S-type Oag chains, they cannot counteract the inhibitory effect of VL-type Oag chains. Hence, RMA3110 LPS, which has L-type and VL-type Oag chains, behaves like the RMA696 LPS, which has VL-type Oag chains, and affects IcsA function in ABM.

Overall, we found that detection of IcsA produced by various S. flexneri strains when examined by indirect IF microscopy did not correlate with the observed impact of LPS Oag chain length on the ability of the strains to form plaque and F-actin comet tails (Figs 6 and 7). Presumably, the epitopes recognized by IcsA antibodies are more accessible than the sequences that interact with host-cell proteins involved in initiating ABM. With the exceptions of RMA696 and RMA3107 (which have only VL-type Oag chains), detection of IcsA on LB-grown S. flexneri strains was unaffected by Oag modal chain length; this was also the case for S. flexneri strains producing LPS with VL-type Oag chains resident within HeLa cells. However, when S. flexneri strains unable to produce VL-type Oag chains were growing inside HeLa cells, we found that detection of IcsA was influenced by LPS Oag chain length. In this situation, a reduction in IcsA detection was observed which correlated with an increase in LPS Oag chain length. These results suggest that VL-type Oag chains can mask and interfere with the binding of IcsA antibodies. One possible explanation for the differences observed between LB-grown and intracellular bacteria is that different fixation conditions were used to prepare the samples for IF staining, and these may have affected the detection of IcsA epitopes which can be masked by the VL-type Oag chains.

Interestingly, strains RMA982, RMA3232 [RMA982(pRMCD77), control] and RMA3231 [RMA982(pRMCD76), wzzO139], when resident within HeLa cells, had IcsA at their cell poles and on lateral regions of their cell surfaces (Fig. 7), and these strains were also able to form plaques and F-actin comet tails. The cell surface localization of IcsA in these strains and their phenotype is inconsistent with the hypothesis that strict polar IcsA localization is needed for ABM within cells (Goldberg, 2001; Suzuki & Sasakawa, 2001).

Our data show a correlation between LPS Oag modal length distribution and the ability of S. flexneri strains to form plaques and F-actin comet tails, and the results are consistent with strains RMA696 (2457T wzzSF : : KmR), RMA3107 [RMA696(pRMCD77), control] and RMA3110 [RMA696(pRMCD80), wzzST] having an ABM defect. However, indirect IF detection of IcsA revealed no major difference between these three strains and the related ABM-competent strains RMA3108 [RMA696(pRMCD76), wzzO139] and RMA3109 [RMA696(pRMCD78), wzzSF]. To explain these observations, we hypothesize that plaque formation only occurs when a minimum number of IcsA proteins are simultaneously active to initiate F-actin comet tail formation. Below this threshold level, sustained ABM sufficient to form plaques is not achieved. This is consistent with our observations that RMA696, RMA3107 and RMA3110 can form F-actin comet tails, albeit very infrequently. Additionally, Magdalena & Goldberg (2002) have recently estimated that a threshold level of 4000 molecules of IcsA is needed to achieve efficient ABM. In the model to be described below, we suggest that Oag chains influence the availability of these IcsA molecules to act in ABM.

A model to explain our observations is shown in Fig. 8. In the wild-type strain 2457T, IcsA epitopes and ABM initiation domains are accessible to antibodies and host proteins due to the presence of LPS with S- and VL-type Oag chains. The presence of LPS with shorter Oag chains (e.g. VS- to S-type) in the outer membrane counteracts the effect of the VL-type Oag chains, thereby allowing IcsA to function in ABM, while LPS with VL-type Oag chains can still confer serum resistance (Hong & Payne, 1997). In strain RMA696 (2457T wzzSF mutant), the LPS only has VL-type Oag chains. These mask and interfere with the IcsA domain(s) that interact(s) with host-cell protein(s) (e.g. N-WASP) (Goldberg, 2001; Suzuki & Sasakawa, 2001), and also slightly reduce antibody binding, presumably due to a decrease in epitope accessibility. In this situation, initiation of ABM is inefficient. In strain RMA982 (2457T wzzSF wzzpHS2 double mutant), relatively short Oag chains are present on the cell surface. In this situation, IcsA can interact with host proteins and hence, ABM can be efficiently initiated. In strain RMA3110 [RMA696(pRMCD80), wzzST], the VL-type Oag chains still interfere with the ability of IcsA to initiate ABM as the WzzST-determined L-type Oag chains are too long to compete with and counteract the effect of the VL-type Oag chains. The L-type Oag chains by themselves do not interfere with the ability of IcsA to initiate ABM (strain RMA3234). Indirect support for this model comes from our observation that detection of IcsA on intracellular S. flexneri bacteria unable to co-express VL-type Oag chains is affected by Oag modal chain length. We have also directly demonstrated, using in situ enzymic hydrolysis, that Oag chains can mask and sterically hinder detection of IcsA by indirect IF (R. Morona & L. Van Den Bosch, unpublished data). This model does not take into account any conformational differences between Oag chains with different modal length distributions or any interactions that may occur between the Oag chains.



View larger version (37K):
[in this window]
[in a new window]
 
Fig. 8. Model for interaction between IcsA and LPS Oag chains. Possible interactions between LPS molecules with different Oag chain length and the IcsA protein in the outer membrane (OM) of S. flexneri. Strains: 2457T (wild-type); RMA3110 [RMA696(pRMCD80)]; RMA696 (2457T wzzSF : : KmR); RMA982 (2457T wzzSF : : KmR wzzpHS2 : : mini-Tn5-CmR); RMA3234 [RMA982(pRMCD80)]. The IcsA protein has sequences (solid regions) that interact with host proteins (e.g. N-WASP, not shown) to initiate ABM. Four types of LPS are shown. 2457T has LPS with both S- and VL-type Oag chains, and IcsA is fully functional. RMA696 has LPS with VL-type Oag chains, and these VL-type Oag chains interfere with IcsA function (solid region is obscured). RMA982 LPS has relatively short, random length Oag chains, and IcsA is fully functional. RMA3110 has LPS with both L- and VL-type Oag chains, and IcsA is not functional in this situation (solid region is obscured), whereas in RMA3234, which has LPS with L-type Oag chains, IcsA is fully functional. Presumably, in RMA696 and RMA3110, the Oag chains mask and/or interfere with the IcsA sequences (solid regions) required to interact with host proteins and the process of nucleation of actin polymerization is inefficient. Antibodies (not shown) are able to bind IcsA on the cell surface of these strains presumably because reactive epitopes (not shown) are still accessible to antibodies in the polyclonal IcsA antiserum used. Under certain conditions (e.g. fixation and immunostaining of intracellular bacteria), the absence of VL-type Oag chains results in enhanced immunostaining of IcsA.

 
The L-type Oag modal chain length conferred by WzzST is approximately double that conferred by WzzSF (S-type). All E. coli-derived wzz genes characterized thus far fall into three subclasses depending on the Oag modal length conferred: Short (S), Long (L) or Intermediate (I) (Franco et al., 1998). Superficially at least, there is no reason to assume that one modal length would be an advantage over another. Our results suggest that during its evolution into a successful human pathogen, S. flexneri 2a has selected the wzzSF subtype of wzz over other wzz genes, such as wzzST. This is likely to have occurred to counterbalance the effect of the VL-Oag chains which confer serum resistance (due to wzzpHS2) but which are not permissive for IcsA function. Presumably, S. flexneri 2a strains harbouring the invasion plasmid and icsA arose first. This allowed ABM and intra- and intercellular spreading, but only by those strains that had a permissive LPS Oag structure. These strains potentially either lacked a wzz gene or had wzz genes conferring the S-, I- or L-type Oag modal lengths. Certain S. flexneri 2a strains then obtained the pHS-2 (encoding wzzpHS2) plasmid, resulting in enhanced serum resistance, and those with wzzSF were the more successful pathogens. Presumably, S. flexneri and enteroinvasive E. coli strains exist in the environment that vary in their relative virulence levels as a result of the wzz genes they carry. Further studies are in progress to investigate the influence of LPS Oags on IcsA function in ABM.


   ACKNOWLEDGEMENTS
 
This work was supported by a project grant from the National Health and Medical Research Council of Australia. Dr Uwe Stroeher is thanked for supplying pBSotn. C. D. was the recipient of an Australian Postgraduate Research Award.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Attridge, S. R., Fazeli, A., Manning, P. A. & Stroeher, U. H. (2001). Isolation and characterisation of bacteriophage-resistant mutants of Vibrio cholerae O139. Microb Pathog 30, 237–246.[CrossRef][Medline]

Baker, S. J., Gunn, J. S. & Morona, R. (1999). The Salmonella typhi melittin resistance gene pqaB affects intracellular growth in PMA-differentiated U937 cells, polymyxin B resistance and lipopolysaccharide. Microbiology 145, 367–378.[Abstract]

Bernadini, M. L., Mounier, J., d'Hauteville, H., Coquis-Rondon, M. & Sansonetti, P. J. (1989). Identification of icsA, a plasmid locus of Shigella flexneri that governs bacterial intracellular spread through interaction with F actin. Proc Natl Acad Sci U S A 86, 3867–3871.[Abstract]

Bik, E. M., Bunschoten, A. E., Gouw, R. D. & Mooi, F. R. (1995). Genesis of the novel epidemic Vibrio cholerae O139 strain: evidence for horizontal transfer of genes involved in polysaccharide synthesis. EMBO J 14, 209–216.[Abstract]

Bik, E. M, Bunschoten, A. E., Willems, R. J., Chang, A. C. & Mooi, F. R. (1996). Genetic organization and functional analysis of the otn DNA essential for cell-wall polysaccharide synthesis in Vibrio cholerae O139. Mol Microbiol 20, 799–811.[Medline]

Daniels, C. & Morona, R. (1999). Analysis of Shigella flexneri Wzz (Rol) function by mutagenesis and cross-linking: Wzz is able to oligomerize. Mol Microbiol 34, 181–194.[CrossRef][Medline]

De Lorenzo, V., Herrero, M., Jakubzik, U. & Timmis, K. N. (1990). Mini-Tn5 transposon derivatives for insertion mutagenesis, promoter probing, and chromosomal insertion of cloned DNA in gram-negative eubacteria. J Bacteriol 172, 6568–6572.[Medline]

Formal, S. B., Dammin, G. J., LaBrec, E. H. & Schneider, H. (1958). Experimental Shigella infections: characteristics of a fatal infection produced in guinea pigs. J Bacteriol 75, 604–610.[Medline]

Franco, A. V., Liu, D. & Reeves, P. R. (1998). The Wzz (Cld) protein in Escherichia coli: amino acid sequence variation determines O-antigen chain length specificity. J Bacteriol 180, 2670–2675.[Abstract/Free Full Text]

Fukuda, I., Suzuki, T., Munakata, H., Hayashi, N., Katayama, E., Yoshikawa, M. & Sasakawa, C. (1995). Cleavage of Shigella surface protein VirG occurs at a specific site, but the secretion is not essential for intracellular spreading. J Bacteriol 177, 1719–1726.[Abstract]

Goldberg, M. B. (2001). Actin-based motility of intracellular microbial pathogens. Microbiol Mol Biol Rev 65, 595–626.[Abstract/Free Full Text]

Hale, T. L. (1991). Genetic basis of virulence in Shigella species. Microbiol Rev 55, 206–224.[Medline]

Hong, M. & Payne, S. M. (1997). Effect of mutations in Shigella flexneri chromosomal and plasmid-encoded lipopolysaccharide genes on invasion and serum resistance. Mol Microbiol 24, 779–781.[CrossRef][Medline]

Klauser, T., Pohlner, J. & Meyer, T. F. (1990). Extracellular transport of cholera toxin B subunit using Neisseria IgA protease beta-domain: conformation-dependent outer membrane translocation. EMBO J 9, 1991–1999.[Abstract]

Lett, M.-C., Sasakawa, C., Okada, N., Sakai, T., Makino, S., Yamada, M., Komatsu, K. & Yoshikawa, M. (1989). virG, a plasmid-coded virulence gene of Shigella flexneri: identification of the virG protein and determination of the complete coding sequence. J Bacteriol 171, 353–359.[Medline]

Lugtenberg, B., Meijers, J., Peters, R., van der Hoek, P. & van Alphen, L. (1975). Electrophoretic resolution of the major outer membrane proteins of Escherichia coli K12 into four bands. FEBS Lett 58, 254–258.[CrossRef][Medline]

Magdalena, J. & Goldberg, M. B. (2002). Quantification of Shigella IcsA required for bacterial actin polymerization. Cell Motil Cytoskeleton 51, 187–196.[CrossRef][Medline]

Morona, R., Brown, M. H., Yeadon, J., Heuzenroeder, M. W. & Manning, P. A. (1991). Effect of lipopolysaccharide core synthesis mutations on the production of Vibrio cholerae O-antigen in Escherichia coli K-12. FEMS Microbiol Lett 82, 279–286.[CrossRef]

Morona, R., Mavris, M., Fallarino, A. & Manning, P. A. (1994). Characterization of the rfc region of Shigella flexneri. J Bacteriol 176, 733–747.[Abstract]

Morona, R., Van Den Bosch, L. & Manning, P. A. (1995). Molecular, genetic, and topological characterization of O-antigen chain length regulation in Shigella flexneri. J Bacteriol 177, 1059–1068.[Abstract]

Morona, R., Van Den Bosch, L. & Daniels, C. (2000). Evaluation of Wzz/MPA1/MPA2 proteins based on the presence of coiled-coil regions. Microbiology 146, 1–4.[Free Full Text]

Oaks, E. V., Wingfield, M. E. & Formal, S. B. (1985). Plaque formation by virulent Shigella flexneri. Infect Immun 48, 124–129.[Medline]

Okada, N., Sasakawa, C., Tobe, T., Yamada, M., Nagai, S., Talukder, K. A., Komatsu, K., Kanegasaki, S. & Yoshikawa, M. (1991a). Virulence-associated chromosomal loci of Shigella flexneri identified by random Tn5 insertion mutagenesis. Mol Microbiol 5, 187–195.[Medline]

Okada, N., Sasakawa, C., Tobe, T., Talukder, K. A., Komatsu, K. & Yoshikawa, M. (1991b). Construction of a physical map of the chromosome of Shigella flexneri 2a and the direct assignment of nine virulence-associated loci identified by Tn5 insertions. Mol Microbiol 5, 2171–2180.[Medline]

Okamura, N. & Nakaya, R. (1977). Rough mutant of Shigella flexneri 2a that penetrates tissue culture cells but does not evoke keratoconjunctivitis in guinea pigs. Infect Immun 17, 4–8.[Medline]

Okamura, N., Nagai, T., Nakaya, R., Kondo, S., Murakami, M. & Hisatsune, K. (1983). HeLa cell invasiveness and O antigen of Shigella flexneri as separate and prerequisite attributes of virulence to evoke keratoconjunctivitis in guinea pigs. Infect Immun 39, 505–513.[Medline]

Philpott, D. J., Edgeworth, J. D. & Sansonetti, P. J. (2000). The pathogenesis of Shigella flexneri infection: lessons from in vitro and in vivo studies. Philos Trans R Soc Lond B Biol Sci 355, 575–586.[CrossRef][Medline]

Rajakumar, K., Jost, B. H., Sasakawa, C., Okada, N., Yoshikawa, M. & Adler, B. (1994). Nucleotide sequence of the rhamnose biosynthetic operon of Shigella flexneri 2a and role of lipopolysaccharide in virulence. J Bacteriol 176, 2362–2373.[Abstract]

Rosenbusch, J. P. (1974). Characterization of the major envelope protein from Escherichia coli. Regular arrangement on the peptidoglycan and unusual dodecyl sulfate binding. J Biol Chem 249, 8019–8029.[Abstract/Free Full Text]

Sandlin, R. C., Lampel, K. A., Keasler, S. P., Goldberg, M. B., Stolzer, A. L. & Maurelli, A. T. (1995). Avirulence of rough mutants of Shigella flexneri: requirement of O antigen for correct unipolar localization of IcsA in the bacterial outer membrane. Infect Immun 63, 229–237.[Abstract]

Sandlin, R. C., Goldberg, M. B. & Maurelli, A. T. (1996). Effect of O side-chain length and composition on the virulence of Shigella flexneri 2a. Mol Microbiol 22, 63–73.[Medline]

Sansonetti, P. J. (2001). Rupture, invasion and inflammatory destruction of the intestinal barrier by Shigella, making sense of prokaryote–eukaryote cross-talks. FEMS Microbiol Rev 25, 3–14.[CrossRef][Medline]

Sereny, B. (1957). Experimental keratoconjunctivitis shigellosa. Acta Microbiol Acad Sci Hung 4, 367–376.[Medline]

Stark, M. J. (1987). Multicopy expression vectors carrying the lac repressor gene for regulated high-level expression of genes in Escherichia coli. Gene 51, 255–267.[CrossRef][Medline]

Stevenson, G., Kessler, A. & Reeves, P. R. (1995). A plasmid-borne O-antigen chain length determinant and its relationship to other chain length determinants. FEMS Microbiol Lett 125, 25–30.

Suzuki, T. & Sasakawa, C. (2001). Molecular basis of the intracellular spreading of Shigella. Infect Immun 69, 5959–5966.[Free Full Text]

Van Den Bosch, L., Manning, P. A. & Morona, R. (1997). Regulation of O-antigen chain length is required for Shigella flexneri virulence. Mol Microbiol 23, 765–775.[Medline]

Yanisch-Perron, C., Vieira, J. & Messing, J. (1985). Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33, 103–119.[CrossRef][Medline]

Received 21 November 2002; revised 14 January 2003; accepted 22 January 2003.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Morona, R.
Articles by Van Den Bosch, L.
Articles citing this Article
PubMed
PubMed Citation
Articles by Morona, R.
Articles by Van Den Bosch, L.
Agricola
Articles by Morona, R.
Articles by Van Den Bosch, L.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2003 Society for General Microbiology.