OUP user menu

Shigella spp. and enteroinvasive Escherichia coli pathogenicity factors

Claude Parsot
DOI: http://dx.doi.org/10.1016/j.femsle.2005.08.046 11-18 First published online: 1 November 2005


Bacteria of Shigella spp. (S. boydii, S. dysenteriae, S. flexneri and S. sonnei) and enteroinvasive Escherichia coli (EIEC) are responsible for shigellosis in humans, a disease characterized by the destruction of the colonic mucosa that is induced upon bacterial invasion. Shigella spp. and EIEC strains contain a virulence plasmid of ∼220 kb that encodes determinants for entry into epithelial cells and dissemination from cell to cell. This review presents the current model on mechanisms of invasion of the colonic epithelium by these bacteria and focuses on their pathogenicity factors, particularly the virulence plasmid-encoded type III secretion system.

  • Invasion
  • Secretion
  • Chaperone
  • Regulation
  • Pathogenesis
  • Virulence
  • Plasmid
  • Type III secretion

1 Introduction

Bacteria of Shigella spp. and enteroinvasive Escherichia coli (EIEC) are responsible for shigellosis in humans. The burden of this disease was estimated to 150 million cases and 1 million deaths per year in the developing world [1]. Shigellosis is characterized by the destruction of the colonic epithelium provoked by the inflammatory response that is induced upon invasion of the mucosa by bacteria. Shigella is divided in four groups (or species), S. boydii, S. dysenteriae, S. flexneri and S. sonnei. However, these bacteria are so closely related to each other and to E. coli strains that, in fact, they all belong to the species E. coli [2]. Most studies have been conducted on S. flexneri and, notwithstanding strain specificities, results probably apply to other Shigella spp. and EIEC. As compared to commensal and other pathogenic E. coli strains, characteristic features of Shigella spp. and EIEC strains are the presence of a virulence plasmid (VP) of ∼220 kb and their ability to induce their entry into epithelial cells and disseminate from cell to cell. This review briefly presents the model on mechanisms of invasion of the epithelium by bacteria of Shigella spp. and focuses on their pathogenicity factors, particularly the VP-encoded type III secretion (TTS) system. This system consists of a TTS apparatus (TTSA) that spans the bacterial envelope, translocators that transit through the TTSA and insert into the host cell membrane to form a pore (translocon), effectors that transit through the TTSA and the translocon and are injected within the eukaryotic cell, chaperones that bind to translocators and some effectors prior to their transit through the TTSA, and transcription activators. Due to space limitations, priority for references was given to recent publications in which readers will find references to older, but not less interesting, works.

2 Pathogenesis

The current model on mechanisms of pathogenesis induced by bacteria belonging to Shigella spp. are derived from in vitro and in vivo studies using various cell types (epithelial cells, macrophages, monocytes, fibroblasts and red blood cells) and animal models of infection, such as the cornea in guinea pigs, ligated ileal loops in rabbits and lungs in mice, in which bacteria induce an inflammatory response leading to destruction of the corresponding epithelium. In the colonic mucosa, bacteria are proposed to cross the epithelial layer by invading M cells overlaying lymphoid follicles, which allows them to reach the basolateral pole of epithelial cells where they induce their uptake (Fig. 1). Entry into epithelial cells involves rearrangements of the cell cytoskeleton that extend beyond the zone of contact between the bacterium and the cell membrane, leading to membrane ruffling and engulfment of the bacterium within a vacuole [3]. Once internalized by epithelial cells, bacteria rapidly lyse the membrane of the entry vacuole and gain access to the cell cytoplasm where they multiply with a generation time of ∼40 min. By inducing actin polymerization at one of their poles, intracellular bacteria move within the cytoplasm of infected cells. This movement generates the formation of protrusions that contain one bacterium at their tip and are engulfed by adjacent epithelial cells, thereby allowing bacteria to disseminate from cell to cell without being exposed to the external milieu. Peptidoglycan fragments released by intracellular bacteria are detected by the Nod1 pathway, leading to phosphorylation and degradation of IκB, translocation of NF-κB to the nucleus and activation of NF-κB regulated genes [ [, []. Analysis of the transcriptome of infected epithelial cells showed, in particular, increased expression of the gene encoding IL-8, a potent chemoattractant for neutrophils [6]. Thus, epithelial cells actively participate in the detection and signaling of invasive bacteria to host defences. Bacteria released from M cells (after their initial uptake) or epithelial cells (after intracellular multiplication) interact with macrophages, escape from the phagocytic vacuole and induce apoptosis of infected cells. Apoptotic macrophages release pro-inflammatory cytokines, including IL-1 and IL-18, which, together with IL-8 released from infected epithelial cells, leads to recruitment of polymorphonuclear cells (PMN) at the site of infection. Transmigration of PMN destabilises the epithelial barrier and facilitates further invasion by luminal bacteria. Interactions of bacteria with host cells, innate and adaptive immune responses induced upon infection and vaccine developments are presented in recent reviews [ [– [].

Figure 1

Model of pathogenesis induced by Shigella spp. Bacteria cross the epithelium barrier by entering into M cells (1). They are delivered to resident macrophages, in which they induce apoptosis (2), and reach the basolateral pole of epithelial cells (3), in which they induce their entry (4). Movement of intracellular bacteria (5) leads to the formation of protrusions and dissemination of bacteria within the epithelium (6). Release of cytokines and chemokines, including IL-1 by apoptotic macrophages (A) and IL-8 by infected enterocytes (B), promotes recruitment of monocytes that migrate through the epithelial barrier (C), facilitating entry of luminal bacteria into epithelial cells (D) and increasing invasion of the epithelium (E).

3 Chromosomal pathogenicity islands and black holes

The genomic sequence of two S. flexneri 2a strains, CP301 and 2457T, was recently determined [ [0, [1]. Genomes of these strains are very similar in size (4 Mbp), sequence and organization. As compared to E. coli K-12 strain MG1655, major differences are (i) the number of insertion sequences (IS), with 400 complete or partial IS in CP301 vs. 40 in MG1655; (ii) the number of genes that are inactivated or missing, 900 genes of MG1655 being absent or inactivated by frameshift mutations or insertion of IS in 2457T; and (iii) the presence of 195 genes specific to S. flexneri. Most specific genes are carried by islands that either encode an integrase, are located at tRNA sites, or are associated with IS or prophage remnants. A number of these genes encode proteins predicted to be membrane-associated or secreted or exhibiting similarities to adhesins. Genes carried by islands involved in pathogenesis include sit genes encoding an iron uptake system [12], sigA encoding an extracellular protease [13], gtr genes encoding proteins involved in glucosylation of the O antigen [14], shiA encoding a membrane-associated protein of unknown function [15] and stx genes encoding the Shiga toxin (in S. dysenteriae strains). Many chromosomal genes that are not specific of Shigella spp. are required for virulence, including those encoding the transcription termination factor Rho, the DNA topoisomerase I, nucleoid-associated proteins H-NS and FIS [16], tRNA modifying enzymes [17], the superoxide dismutase SodB, the protease DegP, the outer membrane protein OmpC, the periplasmic disulfide bond catalyst DsbA and proteins involved in lipopolysaccharide biosynthesis or cell division.

4 Overview of the VP

Sequence analysis of pWR100 and pCP301 from S. flexneri 5 and 2a strains, respectively, indicated that the VP is composed of a mosaic of ∼100 genes and numerous IS, these latter representing one-third of the VP [ [0, [8, [9]. The 30-kb region (designated the entry region) essential for entry into epithelial cells carries genes encoding components of a TTSA, translocators and some effectors, chaperones and transcription activators (Fig. 2). Genes encoding other effectors are scattered on the VP. The TTS system comprises 50 genes exhibiting a similarly low G+C content (34% G+C), indicating that it was acquired by lateral transfer. Two gene clusters exhibit a G+C content of 42%, one encoding IcsA/VirG (involved in movement of intracellular bacteria), a truncated PapC, UshA (a periplasmic UDP-sugar hydrolase) and PhoN1 (a periplasmic acid phosphatase) and the other encoding Orf185 and Orf186 (of unknown function), VirK (involved in production or localization of IcsA) and MsbB2 (an acyl transferase modifying the lipid A). The organization and G+C content of these two regions suggest that they were acquired from the same source, different from the entry region, and that they initially encoded Pap pili. The sepA gene encoding a secreted protease has a G+C content of 49%, suggesting that it came from a different source. The VP contains several gene clusters involved in plasmid replication, partition and post segregation killing and an incomplete transfer region. The presence of two seemingly intact and one incomplete partition systems exhibiting G+C contents that are different from one another and different from that of the replication region suggests that the VP carries elements from four different plasmids [18]. pWR100 and pCP301 differ by the presence or absence 16 fragments, most of which correspond to IS and deletions involving IS or direct repeats. This suggests that the common ancestor of these two VP contained more intact genes and less IS. Genes detected on one VP are not present on all VP: impCAB genes of pCP301 are absent on pWR100 and sepA and several other genes of pCP301 and pWR100 were not detected on the VP of a S. flexneri 6 strain [20]. Many S. flexneri strains carry two small plasmids, one of which encodes the Cld protein controlling the number of O antigen repeats, and often more than one large plasmid. For example, a 165-kb plasmid exhibiting 99.7% identity with plasmid R27 from Salmonella enterica serovar Typhi was present in strain 2457T used for sequencing [11].

Figure 2

Genetic map of the entry region and electron microscopy analysis of the TTSA. Top panel: schematic genetic map of the 30-kb entry region (in two parts); genes encoding transcription activators are in purple, effectors in red, chaperones in blue, components of the needle complex in green and innermembrane proteins in yellow. Bottom panel: structural analysis of the TTSA; (A) negative staining of one TTS apparatus on osmotically shocked bacteria; (B) deduced projection density map of images as shown in A; (C) negative staining of one purified needle complex; (D) average image of the major class of needle complexes (as shown in C); (E) surface representation of the volume of the needle complex, assuming cylindrical symmetry (adapted from [ [2, [4]).

5 Expression of VP genes

The regulation of expression of VP genes has been extensively studied [21]. Transcription of genes of the entry region is regulated by temperature and under the control of two VP-encoded proteins, VirF, a member of the AraC family of transcription activators, and VirB, a member of the ParB family of partition proteins. Binding of H-NS to virF and virB promoters prevents transcription of these genes at 30 °C [22]. At 37 °C, changes in DNA conformation lead to increased transcription of virF and activation of the virB promoter by VirF and production of VirB [23]. Activation of VirB-dependent promoters, including those of the entry region, might involve oligomerization of VirB and displacement of H-NS [24]. In addition to genes of the entry region, VirB controls transcription of ∼15 VP genes [25], most of which are part of the TTS system. Inactivation of orf81 encoding a transcription activator of the AraC family did not affect expression of any VP gene [25]. Since orf81 is inactivated or absent in other S. flexneri strains, this gene might be a remnant of a regulatory circuit that is no longer used.

The TTSA assembled by bacteria growing in broth at 37 °C is only weakly active and is activated upon contact of bacteria with epithelial cells (and deregulated in some mutants). TTSA activation (or deregulation) leads to increased transcription of 12 VP genes encoding TTSA substrates [ [5– [8]. Increased transcription of these genes in conditions of secretion is controlled by MxiE, a transcription activator of the AraC family produced by transcriptional frameshifting from two overlapping ORF [29]. The cis-acting element required for increased transcription of MxiE-controlled promoters is the 17-bp MxiE box (GTATCGTTTTTTTANAG) located between positions −49 and −33 with respect to the transcription start site [30]. There are 9 MxiE boxes on the VP and MxiE boxes are present in the 5′ region of chromosomal ipaH genes encoding TTSA substrates [25]. To be active, MxiE requires IpgC, the chaperone of the translocators IpaB and IpaC, acting as a co-activator [27]. In conditions of non-secretion, IpaB and IpaC are associated independently with IpgC, which titrates IpgC. In addition, MxiE is associated with the TTSA substrate OspD1 and its chaperone Spa15, which prevents MxiE from being activated by free IpgC [31]. Upon TTSA activation, transit of IpaB and IpaC liberates IpgC, but this is not sufficient to activate MxiE as long as OspD1 is present in the cytoplasm. Following transit of OspD1 through the TTSA, MxiE is liberated and an interaction between MxiE and IpgC would allow MxiE to bind to and activate transcription at MxiE-controlled promoters. On the basis of their expression profiles, TTSA substrates can be classified into three categories (Fig. 3): (i) those that are expressed independently of the TTSA activity; (ii) those that are expressed in conditions of non-secretion and induced in conditions of secretion; and (iii) those that are expressed only in conditions of secretion [25].

Figure 3

Repertoires of TTS effector genes controlled by VirB and/or MxiE. Transcription of genes controlled by VirB is independent upon the TTSA activity, whereas transcription of genes controlled by MxiE is activated (or increased) in condition of secretion (adapted from [25]).

6 The TTS apparatus

The TTSA encoded by mxi and spa genes of the entry region has been visualized by transmission electron microscopy on osmotically shocked bacteria [ [2, [3] and scanning electron microscopy on intact bacteria [14]. The core of the TTSA, designated the needle complex, consists in a needle, composed of MxiH and MxiI, and a platform, composed of MxiG, MxiJ, MxiD and presumably MxiM [ [3, [4] (Fig. 2). The platform can be assembled in the absence of MxiH and MxiI, but is not functional, i.e., unable to secrete TTSA substrates. In the needle, MxiH subunits are arranged in a helical fashion to form a 500-Å cylindrical structure with an outer diameter of 70 Å and a central channel of 20 Å [35]. Inner membrane proteins MxiA, Spa9, Spa24, Spa29 and Spa40 are not present in the needle complex, the purification of which involves extraction with detergents. Incorporation of MxiH and MxiI into the needle completes the TTSA and requires Spa47, the ATPase providing energy for transit of substrates through the TTSA. Steady-state amounts of MxiH and MxiI are reduced in secretion defective mutants, indicating that, in contrast to other TTSA substrates stored when the TTSA is not active, needle components are degraded when they are not delivered to the TTSA. MxiK and MxiN, both of which interact with Spa47, are required for transit of needle components, but not Ipa proteins (when the TTSA is artificially assembled by overexpression of MxiH or MxiI in mxiK or mxiN mutants) [36]. Inactivation of spa32 led to assembly of needles that were two times longer than those of the wild-type strain and inability to secrete Ipa proteins upon TTSA activation [ [7, [8], suggesting that Spa32 is required to control the needle length and switch the TTSA substrate specificity from needle components to other substrates. Spa32, Spa33 and MxiC transit through the TTSA. Analysis of a spa33 mutant indicated that Spa33 is required for IpaB and IpaC secretion, but the involvement of Spa33 in completion of the needle complex is not known [39]. Sequence similarities suggest that, like YopN of pathogenic Yersinia spp., MxiC might be involved in control of TTSA activity.

The TTSA is activated upon contact of bacteria with epithelial cells or exposure to the dye Congo red and is deregulated, i.e., constitutively active, by inactivation of ipaB or ipaD. The mechanism by which IpaB and IpaD control TTSA activity is not known, one possibility being that these proteins are required to form a complex that plugs the TTSA. It was recently shown that cholesterol-containing membranes trigger TTSA activity [40]. Genes whose expression is regulated by TTSA activity are induced upon entry into epithelial cells and no longer transcribed in bacteria growing in the intracellular compartment, suggesting that TTSA activity is turned on during entry and off in the cell cytoplasm [26]. The involvement of the TTS system in lysis of the cell membrane that surrounds bacteria in protrusions during cell-to-cell spread suggests that the TTSA is re-activated in protrusions [ [1, [2].

7 TTS chaperones, translocators and effectors

Due to regulation of TTSA activity by external signals, TTSA substrates are stored in the bacterial cytoplasm prior to their transit through the TTSA. Dedicated chaperones stabilize translocators and some effectors and/or maintain them in a secretion-competent state while they are stored in the bacterial cytoplasm [43]. Four chaperones are encoded by the VP, (i) IpgA, the chaperone for IcsB [44]; (ii) IpgE, the chaperone for IpgD [45]; (iii) Spa15, the chaperone for IpaA, IpgB1, OspB, OspD1, OspC2 and OspC3 [ [6, [7]; (i) and IpgC, the chaperone for IpaB and IpaC [48]. IpgC and Spa15 are also involved in the regulatory mechanism by which TTSA activity controls transcription of a set of genes encoding effectors [ [7, [1].

Translocators are proposed to be IpaB and IpaC as the sequence of these proteins exhibits two hydrophobic regions and purified IpaB and IpaC interact with lipid membranes [ [9, [0] and both proteins are essential for contact hemolytic activity (which reflects the ability of bacteria to insert a pore of a defined size in the membrane of erythrocytes) and associate with the membrane of erythrocytes [32]. IpaB interacts with numerous partners, probably in a sequential order: it binds to IpgC in the bacterial cytoplasm and probably to IpaD within the TTSA, forms trimers [50] and associates with IpaC in the external milieu. It also interacts with CDC44 on the cell surface [51], cholesterol in the cell membrane [52] and caspase-1 in the cytoplasm of infected macrophages [53]. Although not part of the translocon, IpaD is probably required for insertion of IpaB and IpaC in cell membranes [54]. IpaB and IpaC are required for lysis of the membrane of the vacuole surrounding bacteria internalized by macrophages and epithelial cells and the membrane of protrusions during cell-to-cell spread [ [1, [2].

Effectors are defined as proteins that, via the TTSA and the translocon, are injected into the cell, where they affect cellular functions. Binding of IpaB to caspase-1 and a direct role of IpaC in promoting cellular signaling cascades that lead to internalization of bacteria in epithelial cells [55] suggest that translocators are also endowed with effector functions. The repertoire of effectors, identified by N-terminal sequencing of proteins secreted by a TTSA-deregulated strain and sequence analysis of the virulence plasmid and, more recently, the chromosome consists in ∼25 proteins encoded by the VP and 3–5 proteins encoded by the chromosome (depending on strains). These latter belong to the IpaH family and, so far, there is no evidence that the chromosome encodes other TTSA substrates. Members of the IpaH family are characterized by a variable N-terminal domain containing 7–10 repeats of a LRR motif and a constant C-terminal domain. In addition to IpaH proteins, several VP-encoded effectors belong to multigene families, with three ospC genes, three ospD genes, two ospE genes and two ipgB genes [18].

The roles of effectors are starting to be elucidated, but their exact function, i.e., enzymatic or binding activities, remains elusive in most cases. IpgD specifically dephosphorylates phosphatidyl-inositol 4,5-biphosphate (PtdIns(4,5)P2) into phosphatidyl-inositol 5-phosphate upon entry of bacteria within epithelial cells, which uncouples the plasma membrane from the cellular cytoskeleton [56]. IpaA binds to the focal adhesion protein vinculin, which promotes depolymerization of actin filaments [57]. IpgB1, directly or indirectly, stimulates activities of Rho GTPases Rac1 and Cdc42, contributing to the formation of membrane ruffles during entry [58]. VirA binds αβ-tubulin dimers and destabilizes microtubules, which also leads to Rac1 activation and formation of membrane ruffles [59]. IpgD, IpaA, IpgB1 and VirA facilitate the formation of entry structures, indicating that several effectors act synergistically to promote internalization of bacteria. IcsB was shown to interact with IcsA (VirG) and this interaction was proposed to prevent association of IcsA with the autophagy protein Atg5 [60].

8 Other VP genes involved in pathogenicity

In addition of the TTS system, the VP carries several genes involved in pathogenicity. icsA (virG) encodes the outer membrane protein directly responsible for promoting the movement of intracellular bacteria. IcsA binds the Neural Wiskott-Aldrich syndrome protein (N-WASP), leading to the formation of a complex with Arp2/3 that induces actin polymerization [ [1, [2]. IcsA is asymmetrically distributed on the bacterial surface and enriched at the old pole of the bacterium [63]. A certain proportion of IcsA is released into the culture medium following cleavage by the VP-encoded outer membrane protease IcsP (SopA) [ [4, [5]. Inactivation of virK led to a decreased production of IcsA by an unknown mechanism [66]. The sepA gene encodes a secreted serine protease of the IgA protease family that, like IcsA, is transported across the outer membrane by an autotransporter C-terminal domain. Inactivation of sepA in S. flexneri 5 led to an attenuated phenotype upon infection of rabbit ileal loops [67]. The absence of sepA in other S. flexneri strains [20] suggests that a protein functionally equivalent to SepA might be encoded by the chromosome or the VP in some strains. The VP msbB2 gene and the chromosomal msbB1 gene encode acyl transferases that catalyze the acyl–oxyacyl linkage of a myristate on the hydroxy-myristate attached to the 3′ position of glucosamide disaccharide in lipid A. This modification contributes to lipid A toxicity. Both MsbB1 and MsbB2 are required for the modification of all lipid A molecules and mutants in either msb1 or msbB2 are attenuated [68].

9 Concluding remarks

Invasion of the colonic epithelium by bacteria involves various interactions between bacteria, both extra- and intracellular, and different cell types. Moreover, these interactions, their outcomes and their consequences are evolving during infection, as cell populations are changing, in nature, relative numbers and states of activation. Likewise, external signals influence repertoires of bacterial genes that are expressed by extra- and intracellular bacteria. Thus, infection is far more complex than what is studied using homogenous populations interacting in vitro. Nevertheless, in vitro models of infection, imperfect as they are, have been essential in deciphering some events that take place during infection. In many cases, in vitro results have been validated using animal models, mostly infection of ligated loops in rabbits and lungs in mice. However, we still have much to learn on the functions of bacterial virulence factors and the responses of infected cells.

Comparison of S. flexneri 2a and E. coli K-12 genomes indicates that many genes are inactivated or missing in S. flexneri strains. Inactivation of these genes suggests that their products were detrimental to invasive bacteria. Indeed, introduction of some E. coli K-12 genes, such as ompT and cadA, into S. flexneri interfered with some virulence traits [ [9, [0]. Acquisition, or assembly, of the VP allowed an ancestral E. coli strain to explore a new environment, but exploiting this “opportunity” required numerous adjustments of the chromosome. Understanding why these genes have been inactivated might shed light on metabolic constraints imposed by an intracellular lifestyle. On the other hand, we start to have ideas on functions of genes carried by pathogenicity islands and the VP. However, as exemplified by stx genes present only in S. dysenteriae 1 strains or by orf81 present on pWR100 and inactivated on pCP301, genes specific to particular strains or isolates might not be of general relevance to the invasive behaviour. Ongoing sequencing projects for S. dysenteriae and S. sonnei strains should lead to a better evaluation of genes that are conserved, or inactivated, between members of Shigella spp. and, likewise, comparison of several VP should help to understand the history of the VP. The TTS system is central to the pathogenicity of Shigella spp. and EIEC strains, being essential for entry of bacteria in epithelial cells and killing of infected macrophages and dendritic cells. The large repertoire of effectors suggests that the TTS system is used for additional functions, possibly to modulate cell responses induced by invading bacteria [71]. The differential regulation of genes encoding TTS effectors suggests that different effectors might be required at different times following contact of bacteria with host cells. Elucidating the function of these effectors and their role(s) at various stages of infection is a challenging prospect.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
View Abstract