Shigella pathogenicity is a multigenic phenomenon involving the participation of genes on the unstable large virulence plasmid and chromosomal PAIs [12–14, 17, 28, 31–34]. Mobile genes encode key factors that help Shigella invade tissue and maintain its intracellular viability [13, 17, 35–38]. The pathogenicity of the strain decreases markedly once the mobile genes are deleted [4, 32, 33].
Several studies have been conducted to detect virulence genes in Shigella by mPCR, targeting ipaH, ial, and rfc or stx1 for serotype identification [3, 5, 7, 39]. In 2005, Thong  first described a new mPCR system to detect S. flexneri 2a by targeting four virulence genes (ipaH, ial, set1A and set1B). This mPCR system was able to determine, in a single reaction, whether genes related to pathogenesis of a particular Shigella strain are associated with the chromosome or plasmid, and whether the serotype of the particular strain can be grouped under S. flexneri 2a [4, 5]. In our present study, Thong’s mPCR system was modified to identify S. flexneri 2a strains and their virulence using only three virulent genes (ipaH, ial, and set1B). We omitted set1A from the mPCR system, as both set1B and set1A genes have been shown to exist in tandem on PAI-1 of the bacterial chromosome, and they share the same promoter [5, 21]. The low prevalence of ial (45/86, 52.3%) verifies that the cell-entry region on the large virulence plasmid of S. flexneri is prone to loss or deletion. The high prevalence of the set1B gene (69/86, 80.2%) verifies that in the rural regions of Zhengding, the isolated epidemic strain of Shigella was S. flexneri 2a. All of our mPCR results were confirmed by serological tests. We confirmed that comparable decreases in virulence occur following the deletion of essential elements in the large virulence plasmid (ipaH and set1B for SF68; and ipaH for SF36) [35–38]. A clinical SF51 isolate was found to retain ial but had lost set1B, and demonstrated an obvious decrease in HeLa cell invasion. This indicated to us that the chromosome locus around set1B may influence virulence.
The location of set1B is known to be in Shigella PAI-1 [7, 20], which exists exclusively in S. flexneri 2a. At least four major virulence genes are present in PAI-1 (pic, set1A, set1B, and sigA). The autotransporter SigA exhibits cytopathic effects on HEp-2 cells , and the autotransporter Pic exhibits hemagglutination and mucinolytic activities in vitro[20–23, 41–43]. Upstream from pic are two IS elements, IS911 and IS629, followed by pic itself, and then a perD IS element . This implies that pic can be spontaneously deleted.
The upstream element int, downstream element orf30, cytopathic factor gene sigA, and the hemagglutinin gene pic on PAI-1 of SF51 were sequenced to verify whether SF51 lost the whole PAI-1 or only part of the genetic locus around set1B. Our results revealed that the entire pic gene on PAI-1 was deleted in this case, whereas other genes (sigA, int, and orf30) were unaffected (Figure 1). This result also suggests that a decrease in virulence of SF51 is not related to sigA, but may be associated with pic deletion.
To confirm that the decreased virulence phenotype in SF51 was associated with deletion of pic, we knocked out pic from the SF301 strain to produce SF301-∆ pic. Additionally, complementation strains SF301-∆ pic/pPic and SF51pic/pPic were constructed to demonstrate that the decreased virulence of SF51 was associated with the deletion of pic. Using gentamicin protection assays, we showed that the Hela cell invasion potential of the pic knockout strains, SF51 and SF301-∆ pic, was decreased compared with the wild-type SF301 strain. This decreased virulence was partially recovered by introducing pSC-pic.
Previous studies have demonstrated that purified recombinant protein Pic (prepared from E.coli HB101 (pPic1)) is not involved in cytotoxic effects on HT29-C1 and HEp-2 cells [24, 25]. However, the findings from our current study show that both the clinical and constructed pic-deleted mutants possessed a decreased tendency for cell invasion compared with SF301. Virulence was partially recovered through the insertion of a complementary pic gene into these deletion mutants. Because Pic did not elicit cytopathic effects on epithelial cells, it may be associated with a less efficient interaction process with host cells, lacking any assistance from bacterial effectors. This phenomenon has also been observed by Vidal et al. , who examined the EPEC autotransporter EspC. Purified EspC requires a higher concentration (300 μg/ml vs. 50 μg/ml for other autotransporter cytotoxins) and a longer incubation time (8 h vs. 1 h for EPEC host cells) to produce the same cytotoxic effects as other EPEC isolates. Further studies have confirmed that EspC translocation into epithelial cells results in cytopathic effects in HeLa cells, but require participation of types III and V secretion systems. The mechanism by which Pic is interacted with epithelial cells remains unknown and warrants further study. Further, differences in results observed with Pic regarding decreases in cytopathic effects are likely also associated with other cell lines. Differences in invasion efficiency between Hela cells and HEp-2 cells have been observed for Streptococcus pyrogenes, Campylobacter jejuni and Salmonella typhimurium[45–47]; however, the reasons for these differences remain unclear, and further study is required to clarify this.
The mouse Sereny test is commonly used to the test the invasiveness of Shigella. In our work, the virulence of SF51 and SF301-∆ pic was obviously decreased. This was partially recovered by the introduction of pSC-pic into deletion mutants. Our findings support the conclusion that pic is associated with the invasion potential of S. flexneri 2a.
Harrington et al.  used a mouse model treated with streptomycin to show that Pic promotes intestinal colonization by comparing intestinal colonization abilities of wild-type E. coli 042 and pic mutants (E. coli 042 pic::aph3 and E. coli 042PicS258A). They demonstrated that the constructed mutants (E. coli 042 pic::aph3 and E. coli 042PicS258A) contained significant defects that adversely affected colonization of mice gastrointestinal tracts compared with E. coli 042. Further work by Harrington et al. suggested that a possible mechanism of promoting intestinal colonization depended on the mucinase activity of Pic. They also showed that this effect is associated with the serine protease catalytic residue in Pic. The research of Harrington et al. supports our findings that Pic is involved in bacterial invasion ability. Whether a decrease in virulence is associated with the mucinase activity of Pic, or other biological activities, should be investigated further.