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

The lipopolysaccharide (LPS) molecules of Shigella flexneri 2a have O antigen (Oag) polysaccharides with two modal chain length distributions. The chromosomal wzz(SF) gene results in short (S) type Oag chains [11-17 Oag repeat units (RUs)], and the pHS-2 plasmid-located wzz(pHS2) gene results in very long (VL) type Oag chains (>90 Oag RUs). S. flexneri wzz(SF) 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 wzz(SF) wzz(pHS2) 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. Wzz(O139) from Vibrio cholerae O139 and Wzz(ST) 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.