- Top of page
- Experimental procedures
- Supporting Information
Type III secretion systems (T3SSs) are protein injection devices essential for the interaction of many Gram-negative bacteria with eukaryotic cells. While Shigella assembles its T3SS when the environmental conditions are appropriate for invasion, secretion is only activated after physical contact with a host cell. First, the translocators are secreted to form a pore in the host cell membrane, followed by effectors which manipulate the host cell. Secretion activation is tightly controlled by conserved T3SS components: the needle tip proteins IpaD and IpaB, the needle itself and the intracellular gatekeeper protein MxiC. To further characterize the role of IpaD during activation, we combined random mutagenesis with a genetic screen to identify ipaD mutant strains unable to respond to host cell contact. Class II mutants have an overall defect in secretion induction. They map to IpaD's C-terminal helix and likely affect activation signal generation or transmission. The Class I mutant secretes translocators prematurely and is specifically defective in IpaD secretion upon activation. A phenotypically equivalent mutant was found in mxiC. We show that IpaD and MxiC act in the same intracellular pathway. In summary, we demonstrate that IpaD has a dual role and acts at two distinct locations during secretion activation.
- Top of page
- Experimental procedures
- Supporting Information
Type III secretion systems (T3SSs) are common virulence factors among Gram-negative bacteria. They are protein transport devices used for injecting effector proteins into host cells to modulate their responses in favour of the bacterium (Cornelis, 2006). To study conserved aspects of type III secretion, we use Shigella flexneri as a model system. It is the causative agent of human bacillary dysentery, a form of inflammatory diarrhoea, and uses the T3SS to promote its invasion into gut epithelial cells. The T3SS consists of a basal body spanning both bacterial membranes and a hollow extracellular needle serving as secretion channel. The distal end of the needle is topped by the tip complex, which is required for pore formation and protein translocation into host cells. In Shigella, the apparatus is assembled when the environmental conditions are appropriate for invasion, but secretion is blocked until physical contact with a host cell generates an activation signal (reviewed in Schroeder and Hilbi, 2008).
Protein translocation into host cells requires three proteins known as translocators (IpaB, IpaC and IpaD in Shigella). Of these, IpaD is hydrophilic and constitutively present atop mature needles (Espina et al., 2006; Sani et al., 2007; Veenendaal et al., 2007). It probably serves as a scaffold for the mostly hydrophobic proteins IpaB and IpaC, which later form the pore in the host cell membrane. IpaB is also constitutively present at needle tips, while the other hydrophobic translocator, IpaC, is only recruited to the needle tip upon activation (Veenendaal et al., 2007; Roehrich et al., 2010; Shen et al., 2010). In Yersinia, the needle tip is formed by the hydrophilic translocator LcrV: in this species no hydrophobic translocator has been found constitutively present atop needles (Mueller et al., 2005; 2008; Blocker et al., 2008).
Type III secretion is a tightly regulated process initiated only after host cell contact. As type III proteins are expressed when the environmental conditions are appropriate for invasion, secretion has to be prevented prior to cell contact. Recently, we proposed a model featuring gating mechanisms located in the cytoplasm that prevent secretion in the absence of activation signals (Martinez-Argudo and Blocker, 2010). We hypothesized that under non-activated conditions, a mechanism that represses translocator secretion, directly or indirectly involving IpaD and/or IpaB, is put in place. In addition, the gatekeeper protein MxiC prevents premature effector secretion (Botteaux et al., 2009; Martinez-Argudo and Blocker, 2010) and it requires a secretion signal to do so (Botteaux et al., 2009). After reception of the activation signal, such repression mechanisms would be counteracted and secretion happens in an ordered way: translocator proteins are secreted first, then MxiC (Kenjale et al., 2005; Martinez-Argudo and Blocker, 2010) followed by early and late effectors (Mavris et al., 2002; Parsot, 2009). As a ΔmxiC mutant shows weak and delayed secretion of translocators in response to activation when compared with the wild-type strain, we suggested that, in Shigella, MxiC also plays a role in activating translocator secretion (Martinez-Argudo and Blocker, 2010).
MxiC belongs to a family of T3SS proteins that includes Yersinia YopN/TyeA, enteropathogenic Escherichia coli SepL and Salmonella InvE and SsaL (Pallen et al., 2005). Besides similar sequences, the proteins have conserved domain topologies and structures (Schubot et al., 2005; Deane et al., 2008; 2010) and all block effector secretion in the absence of an activation signal. However, mutants do not show the same phenotype with respect to the control of translocator secretion. In ΔinvE, ΔssaL, ΔsepL and ΔmxiC mutants, secretion of translocators is impaired (Kubori and Galan, 2002; Coombes et al., 2004; O'Connell et al., 2004; Deng et al., 2005), whereas it is constitutive in ΔyopN or ΔtyeA mutants (Iriarte et al., 1998; Ferracci et al., 2005). How the gatekeepers perform their tasks is unclear, but Salmonella InvE interacts with chaperone–translocator complexes (Kubori and Galan, 2002) and Shigella MxiC interacts with the ATPase in vitro (Botteaux et al., 2009).
Despite intense research we still do not understand how the activation signal generated after contact with the host cell is transduced to the cytoplasmic side of the T3SS to initiate secretion. The first part of the T3SS that establishes contact with the host cell is probably the needle tip. Therefore, it is likely that the tip complex senses the host cell. Experimental data from the Shigella and Yersinia systems suggest that the needle is directly involved in transmitting the activation signal. Particular single amino acid mutations in the needle protein lead to altered secretion phenotypes and tip complex composition (Kenjale et al., 2005; Torruellas et al., 2005; Veenendaal et al., 2007).
The Shigella needle protein MxiH is essentially a coiled coil hairpin that polymerizes into the helical needle using both of its termini (Cordes et al., 2003; Kenjale et al., 2005; Deane et al., 2006; Fujii et al., 2012; Loquet et al., 2012). Similar to the needle protein MxiH, IpaD contains a longer central coiled coil and requires its extreme C-terminus for binding to the needle. In addition, it has two globular domains, one at the N-terminus and one linking the two helices of the coiled coil (Johnson et al., 2007; Veenendaal et al., 2007). The C-terminal globular domain is a site of interaction with IpaB, while the N-terminal globular domain of IpaD has been proposed to act as a self-chaperone (Johnson et al., 2007). In the LcrV class of hydrophilic tip proteins, the chaperone is formed by a separate polypeptide called LcrG (Lawton et al., 2002).
The needle tip was first visualized in the Yersinia T3SS (Mueller et al., 2005). We proposed that the tip contains four molecules of IpaD and one molecule of IpaB (Johnson et al., 2007; Veenendaal et al., 2007; Blocker et al., 2008). The crystal structure of IpaD allowed construction of a pentamer based on dimer contacts in one of the crystal forms with a helical rise similar to that of the MxiH needle (Cordes et al., 2003; Johnson et al., 2007). In the proposed IpaD pentamer, the position between the fourth and first IpaD molecules is structurally different from all other positions. This final position might thus be filled by IpaB (Johnson et al., 2007; Veenendaal et al., 2007; Blocker et al., 2008). The model puts the proposed central coiled coil and the C-terminal globular domain of IpaB in an equivalent conformation to that of IpaD, positioning IpaB's domain homologous to pore-forming colicins and its putative transmembrane helices in the ideal place to interact with the host cell membrane (Johnson et al., 2007; Barta et al., 2012a). Together with mutagenesis data this suggests that IpaB could be the host cell sensor and that a conformational change in the tip complex might be signalling event triggering T3SS activation (Veenendaal et al., 2007; Roehrich et al., 2010; Shen et al., 2010).
Congo red (CR) is a small amphipathic dye molecule that was originally used to assess infectivity of Shigella strains: only colonies binding CR on agar plates were infective, while ‘white’ colonies were not (Payne and Finkelstein, 1977). It was later found that CR specifically activates type III secretion (Parsot et al., 1995) and that its target is likely in the tip complex (Veenendaal et al., 2007). All T3SS component-knockout (KO) strains that are not inducible by CR are also non-invasive, suggesting that CR and host cell contact induce T3SS activation using a related mechanism (Menard et al., 1993; 1994; Blocker et al., 2001).
To test our hypothesis that, as part of the tip complex, IpaD is involved in conformational changes crucial for signal transduction, we sought to identify ipaD point mutants unable to respond to host cell contact. For ease of analysis, we chose to set up a genetic screen for mutants insensitive to induction by CR. All mutants isolated showed a wild-type-like tip complex composition. Yet, six out of seven mutants were impaired in host cell sensing and induction of type III secretion. The position of these mutations highlights the importance of the C-terminal helix of IpaD in signal transduction. The remaining ipaD mutant was unable to prevent secretion under non-inducing conditions. Parallel analysis of a mxiC mutant with a similar phenotype further supports the regulatory roles of both IpaD and MxiC in translocator secretion. Taken together, these data demonstrate that IpaD has a dual role during secretion activation.