By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
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.
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.
For the Shigella T3SS, ‘induction’ describes the burst of Ipa protein secretion upon host cell sensing (Menard et al., 1994) or addition of CR, a small amphipathic dye molecule which acts as an artificial inducer of its T3SS (Bahrani et al., 1997). Induction is a very fast event as around 50% of the Ipa proteins are secreted at 37°C in the 15 min following CR addition (Bahrani et al., 1997; Magdalena et al., 2002; Parsot et al., 2005). Prior to induction there is a slow, low level of Ipa protein secretion known as ‘leakage’, where around 5% of Ipa proteins are secreted (Magdalena et al., 2002; Parsot et al., 2005). ‘Constitutive secretion’ of Ipa proteins represents an unregulated and higher-level secretion than leakage and involves not only the Ipa proteins but also the ‘late effectors’, which are involved in later stages of infection. In the ΔipaB and ΔipaD mutant strains, constitutive secretion is very fast and detectable in minutes. It is thus named ‘fast constitutive secretion’ (Veenendaal et al., 2007). We recently showed that constitutive secretion has a different cause in a ΔipaB versus a ipaBΔ3 mutant (Martinez-Argudo and Blocker, 2010). As the latter mutant is in a constitutively activated state, we now call its phenotype ‘premature secretion’.
Isolation of ipaD mutants affected in secretion
To test our hypothesis that IpaD is involved in the reception or transduction of the activation signal, we sought to identify ipaD mutant strains that are unable to secrete after induction by CR. For this, we used a screening method based on the colour of Shigella colonies in plates containing CR. Wild-type Shigella colonies stain light red on CR-containing plates, while mutants unable to secrete effectors are white (Maurelli et al., 1984). Constitutive secretor mutant colonies, such as those containing bacteria that have lost their needle tip, are bright red on CR plates, probably because of the secretion of late effectors (Parsot et al., 1995). Upon random PCR mutagenesis of the ipaD gene, the mutant DNA libraries were cloned and amplified as described in Experimental procedures. After transformation of the mutant libraries into the Shigella ΔipaD mutant, we screened for white colonies in plates containing CR (Fig. S1). Around 500 000 transformants were screened and as expected a substantial proportion of them were bright red (10–20%, indicating expression of non-functional ipaD alleles). One hundred and ten ‘white’ colonies were selected and their plasmids isolated and retransformed into the ΔipaD mutant to ensure the white colour was not due to loss-of-function mutations elsewhere within the T3SS-encoding operons.
Twenty-two plasmids were confirmed as giving rise to white colonies and their ipaD gene was sequenced (Table 1). Mutant ipaD(L99P, N255D) was not white but slightly more red than wild-type when retransformed into the ΔipaD mutant; however, it showed an interesting phenotype (see below) and was further investigated. In cases where more than one mutation was found in the ipaD gene, individual mutations were cloned independently in order to identify which was responsible for the mutant phenotype and only the mutation generating the mutant phenotype was further investigated.
Table 1. ipaD mutants isolated in this study and their phenotypes
Times the mutant was found
Needle tip composition
The mutation responsible for the observed phenotype is indicated in bold letters. ‘WT’ indicates wild-type-like properties, + or ++ an increase and – or – – a decrease. ‘NA’ means not applicable.
aΔipaD is not inducible by CR, but nevertheless proteins are detectable as it is a fast constitutive secreter.
bMutant L99P only has a defect in IpaD secretion after CR induction.
cMutants K291I/T/N and T296I were found independently in two and three separate libraries respectively.
ipaD mutant strains have differentially altered secretion profiles
To characterize the ipaD mutants, we analysed their secretion profile during exponential growth in non-inducing conditions (Fig. 1A) and following CR induction (Fig. 1B). We found two novel types of secretion profile. First, mutant ipaDL99P secreted a much higher level of T3S proteins in non-inducing conditions when compared to the ΔipaD/ipaD+ complemented strain, but still less than the ΔipaD mutant (Fig. 1A, c. 4 h incubation). Additionally, mutant ipaDL99P showed a defect in IpaD secretion (approximately 20% of the complemented strain) following CR induction, while secretion of IpaB, IpaC, MxiC and effectors was still induced to levels comparable with the complemented strain (Fig. 1B). In this 15 min assay, increased secretion under non-inducing conditions is only visible if it is ‘fast’ as for example in the ΔipaD mutant. In a second group of mutants (ipaDN186Y, ipaDK291 and ipaD(N273I, Q277L)), induction of secretion was impaired: in comparison to the ΔipaD/ipaD+ complemented strain they secreted only approximately 40%, 20%, 10% and 10% of IpaB, IpaC, IpaD and MxiC respectively. Mutants ipaDN292Y, ipaDT296I and ipaDQ299L showed a similar secretion pattern, although the defects in secretion were less pronounced (approximately 80%, 60%, 40% and 20% of IpaB, IpaC, IpaD and MxiC respectively; Fig. 1B). We named these Classes I, IIa and IIb respectively, starting the numbering with the most N-terminal mutation. When we analysed the expression of the IpaD mutant proteins, the level of the mutant proteins was similar to the complemented strain with the exception of the IpaDL99P, which was present at around 60% of the wild-type allele (Fig. 1C). As deletion of ipaD leads to constitutive secretion, we were concerned that the increased secretion in ipaDL99P under non-inducing conditions might be due to a decrease in IpaD expression. However, the ipaDL99P mutant secreted more IpaB and IpaC than a strain expressing much lower levels of the wild-type IpaD protein, indicating that constitutive secretion has a different cause in ipaDL99P (Fig. S2). Therefore, all IpaD mutants isolated are partially defective in secretion regulation.
All IpaD mutant proteins are secreted in a constitutive secretor background
As all the newly isolated ipaD mutants showed a defect in IpaD secretion, we wondered if the mutations made them ‘unfit’ for secretion, for example by affecting their conformation in the cytoplasm. To test this possibility we checked if the IpaD mutants were secreted in a constitutive secretor background. We and others have previously shown that the translocator proteins IpaC and IpaD are secreted constitutively in a ΔipaB background (Menard et al., 1994; Martinez-Argudo and Blocker, 2010). We generated a ΔipaB ΔipaD double KO strain and after transforming all the ipaD mutant plasmids individually into this strain we compared their secretion phenotype to that of the KO strain transformed with a wild-type copy of the ipaD gene. Analysis of the secretion profile by Western blot established that the IpaD mutant proteins were constitutively secreted by the ΔipaB ΔipaD/ipaD mutant strains, as efficiently as wild-type IpaD was secreted by the control strain ΔipaB ΔipaD/ipaD+ (Fig. S3). This indicates that our mutations are not impairing IpaD's ability to be secreted by the T3SS.
Class II ipaD mutants show defects in invasion of cultured cells
To test whether the defect in protein secretion after addition of CR in the ipaD mutants correlates with a defect in signal reception/transduction after host cell contact, we analysed whether the ipaD mutants are still able to invade HeLa cells. Gentamicin protection assays were performed as described in Experimental procedures. In this assay, the ΔipaD mutant is completely non-invasive (Menard et al., 1993). All Class II ipaD mutant strains, except for ipaDN292Y, showed a severe defect in invasion (c. 20% of the complemented strain, Fig. 2A). Mutant ipaDL99P was as invasive as the complemented strain. These results indicate that most of the mutations identified in ipaD that cause a defect in CR-induced secretion lead to a defect in invasion of cultured cells.
ipaD mutant strains have normal needle tip compositions
As all isolated ipaD mutants had a partial defect in secretion and as IpaD forms part of the tip complex, one possible explanation is that the mutants have a defect in the needle tip composition. Altered tip compositions are known to alter secretion regulation (Veenendaal et al., 2007). Therefore, we examined the tip complex composition of the ipaD mutants. Needles were purified from the different ipaD mutant strains as described in Experimental procedures and analysed by Western blot for the presence of the tip proteins IpaD and IpaB. All mutants had wild-type levels of both IpaD and IpaB within the levels of experimental variability (Fig. 2B). This was also the case in the ipaDL99P mutant, which was expressed at lower levels (Fig. 2C). This indicates that none of the mutants has a defect in needle tip complex composition.
Class IIa and IIb ipaD mutants are affected in signalling
Induction of secretion was impaired in mutants from Classes IIa and IIb. We have hypothesized that, as part of the tip complex, IpaD is involved in conformational changes essential for the transduction of the activation signal (Veenendaal et al., 2007; Martinez-Argudo and Blocker, 2010). Therefore, we should be able to isolate mutants unable to transmit the signal. However, no mutant was completely uninducible. To test whether the combination of mutations would produce a stronger phenotype, we constructed two new mutant ipaD plasmids combining Class IIa mutations N186Y and K291I or Class IIb mutations N292Y, T296I and Q299L. Mutant ipaD(N186Y, K291I) considerably reduced the secretion levels following CR induction compared to single N186Y and K291I mutations, while the triple mutant ipaD(N292Y, T296I, Q299L) showed a phenotype similar to Class IIa single mutants (Fig. 3). Just like the original single mutants, these combinations showed normal leakage (not shown).
Our previous results indicate that the gatekeeper protein MxiC facilitates secretion of the translocators IpaB, IpaC and IpaD, as translocator secretion is only very weakly inducible in a ΔmxiC mutant (Martinez-Argudo and Blocker, 2010). We also found that translocator secretion is MxiC-independent in a ΔipaB or ΔipaD constitutive secretor background, but that it is still MxiC-dependent in a constitutively ‘on’ ipaBΔ3 strain, as in the wild-type strain (Martinez-Argudo and Blocker, 2010; Roehrich et al., 2010). In the ipaBΔ3 mutant IpaB is expressed lacking its last three C-terminal amino acids. This mutant is locked in a secretion ‘on’ conformation (i.e. similar to the secretion activated wild-type), as it is a constitutive secretor, invasive and recruits some IpaC to the needle tip (Roehrich et al., 2010). To rule out the possibility that the decreased secretion of Class II IpaD mutant proteins was due to lack of regulation by MxiC, we investigated if they were secreted in an ipaBΔ3 background. We made a plasmid expressing both ipaBΔ3 and ipaD into which all the mutations were subsequently cloned. All plasmids were transformed into the strain ΔipaB ΔipaD and their secretion profile analysed. For mutants belonging to Class II, secretion of IpaD became constitutive in all the mutants (not shown). This was also the case when non-inducible mutant ipaD(N186Y, K291I) was expressed in this background. Thus, impaired secretion of these IpaD mutant proteins is not due to a lack of regulation by MxiC. Moreover, the constitutive secretion pattern of all the strains was similar to that of the ipaBΔ3 ΔipaD/ipaD+ strain (not shown), indicating that the ipaBΔ3 mutation is epistatic over the ipaD Class II mutations.
These results, together with the fact that the composition of the needle tip is wild-type-like for those mutants (Fig. 2B), indicate that Class IIa and IIb mutants are affected in signalling from the needle tip and not in a downstream activation step.
Lack of secretion of the Class I IpaDL99P mutant protein is due to lack of regulation by MxiC
In the ipaDL99P mutant strain, only IpaD secretion was impaired after activation (Fig. 1B). As mentioned above, secretion of the translocator proteins from the cytoplasm is promoted by MxiC in wild-type and ipaBΔ3 backgrounds (Martinez-Argudo and Blocker, 2010). In order to examine if secretion of the IpaDL99P mutant protein was still dependent on MxiC, we analysed its secretion in the ipaBΔ3 background. When compared with ipaBΔ3 ΔipaD/ipaD+, strain ipaBΔ3 ipaDL99P showed a different pattern of secretion: IpaDL99P secretion was impaired in the ipaBΔ3 background, but constitutive in the ΔipaB background (Fig. 4). Therefore, the impairment in IpaDL99P secretion is probably due to a lack of intracytoplasmic regulation by MxiC.
IpaD has a dual role in T3SS regulation
The ipaDL99P strain secretes higher levels of T3S proteins in non-inducing conditions compared to the complemented strain (Fig. 1A). To further investigate this, we analysed the supernatant of exponential cultures under non-induced conditions by Western blot and found that the mutant secreted IpaB, IpaC, MxiC and effector proteins (Fig. 5A and not shown). This suggested that in the ipaDL99P mutant a regulatory function of IpaD is altered and that, as a result, the mechanism that prevents secretion before activation signal reception is impaired. On the other hand, Class II mutants did not show a defect in the prevention of secretion under non-induced conditions (Fig. 1A). Therefore, Class I and Class II mutants could affect two different aspects of IpaD function. To test this hypothesis, we made a double mutant combining mutation ipaDK291I (Class IIa) and ipaDL99P (Class I). If different functions were affected in these classes, one would expect to see both phenotypes in the combined mutant. However, if both classes affected the same function, dominance of one of the phenotypes would be more likely. The mutant ipaD(L99P, K291I) showed both phenotypes: it secreted IpaB, IpaC and MxiC under non-inducing conditions and was impaired in CR induction (Fig. 5). Together with the fact that secretion in the different mutant classes is differentially affected by an additional ipaBΔ3 mutation, this result suggests that IpaD has a dual role in regulation of T3SS secretion.
IpaD and MxiC are both involved in the regulation of the mechanism that controls secretion
Intracytoplasmic MxiC is required for secretion of translocator proteins after activation (Martinez-Argudo and Blocker, 2010). Intriguingly, we generated a mutant in mxiC with a very similar phenotype to the ipaDL99P mutant: mxiC(E201K, E276K, E293K) is a triple mutant in a conserved negatively charged patch on the surface of MxiC that has been suggested to be involved in protein interactions (Deane et al., 2008). This mutant also secretes IpaB, IpaC, MxiC and effector proteins in non-inducing conditions. Furthermore, it is also impaired in IpaD secretion after induction, but not affected in induction of any of the other proteins (Fig. 6 and not shown). Our results suggest that both the ipaDL99P and mxiC(E201K, E276K, E293K) mutations cause premature activation of secretion.
When IpaDL99P and MxiC(E201K, E276K, E293K) were expressed together, we observed an increase in secretion under non-induced conditions in comparison to the single mutants (Fig. 7). These results strongly suggest that as MxiC, IpaD is involved in the mechanism that controls secretion.
Premature secretion in the ipaDL99P mutant is MxiC-dependent
We thus wondered whether IpaD is involved in the same activation route as MxiC or whether it has an independent role in controlling secretion. If the former was true, one would expect the ipaDL99P mutant to require MxiC for premature secretion. However, if IpaD was directly affecting secretion of the translocators IpaB and IpaC, ipaDL99P should lead to MxiC-independent constitutive secretion. To distinguish between these possibilities, we analysed whether IpaB and IpaC were secreted from a ΔipaD ΔmxiC/ipaDL99P strain under non-inducing conditions (Fig. 8). We found premature secretion in the ΔipaD ΔmxiC/ipaDL99P strain was reduced in comparison to ipaDL99P alone. Thus, premature secretion of IpaB and IpaC in ipaDL99P is dependent on the presence of MxiC.
Type III secretion systems are activated by host cell contact which we proposed leads to a conformational change in the needle tip complex that triggers secretion from the cytoplasm (Veenendaal et al., 2007; Blocker et al., 2008; Martinez-Argudo and Blocker, 2010). To investigate the role of the major component of the tip complex IpaD in transduction of the activation signal into the cytoplasm, we used a genetic screen to identify ipaD mutants unable to secrete after induction by CR. All previous mutations described in IpaD are silent or lead to at least partial loss-of-function and hence constitutive secretion (Picking et al., 2005; Barta et al., 2012b; Epler et al., 2012). Here we describe several mutations in ipaD that lead to separate defects in secretion activation (Class I and Class II) and use them to conclude that IpaD has two roles in this process.
IpaD is involved in signal transduction from the needle tip
All our results suggest that Class II ipaD mutants are affected in signalling from the tip complex. Their very identification supports the existence of at least two functional states of the needle tip. All Class II mutants have wild-type-like needle tip complex composition and show a decrease in protein secretion after induction by CR. Furthermore, the defect in secretion of IpaD Class II mutants is not due to their inability to be secreted by the T3SS or to a lack of regulation by MxiC. Finally, combining two ‘strong’ Class IIa mutations generates a non-inducible mutant ipaD(N186Y, K291I).
The ability of the mutants to be induced by CR correlates with their ability to invade HeLa cells: all Class II mutants except for ipaDN292Y showed significant decreases in invasion. Therefore, overall our genetic screen yields physiologically meaningful results and supports the notion that CR and host cells are sensed similarly, although the mechanism of sensing remains unknown. It is unclear why the decrease in CR inducibility and the ability to invade host cells do not correlate in mutant ipaDN292Y. There are a number of exceptions to the correlation between CR sensing and the ability to invade: three point mutants in mxiH are CR inducible but only poorly invasive (Kenjale et al., 2005). A reverse correlation was observed in ipaBΔ3, which is barely inducible by CR, does not perform contact-haemolysis (an assay for the ability to insert the translocon into host membranes) but is still invasive (Roehrich et al., 2010). Thus, even though these phenomena are clearly related, their detailed molecular mechanisms are not exactly the same. To understand the cause of these differences, we will need to understand the activation mechanism in molecular detail.
In the constitutively ‘on’ background ipaBΔ3 (Roehrich et al., 2010), Class II ipaD mutants behave as constitutive secretors indicating that signal transduction is unaffected by these mutations if the tip is already activated by a conformational change in IpaB. This supports the proposal that both IpaD and IpaB are involved in initiation of signal transduction (Veenendaal et al., 2007).
Mutations that affect signalling from the needle tip are located in or near the IpaD C-terminal helix
Class II mutations lie in the top of the C-terminal helix (Fig. 9) or on that face of the molecule (N186Y). As they all clustered in one part of the molecule, it is unlikely that we have overlooked another major patch of mutants in our screen. In all cases polar or charged amino acids are mutated into hydrophobic or aromatic residues. Intriguingly, the mutations that cause a stronger phenotype (N186Y, N273I + Q277L, K291I) line one side and the ones causing a weaker phenotype (N292Y, T296I, Q299L) line another side of the helix. The C-terminal part of the needle-associated proteins is required for assembly and needle tip binding: mutants containing short C-terminal deletions in the needle protein MxiH fail to produce needles (Kenjale et al., 2005; Fujii et al., 2012) and similar deletions in IpaB or IpaD reduce or abolish binding to the needle tip (Espina et al., 2006; Veenendaal et al., 2007; Roehrich et al., 2010). Indeed, no mutations were found further down the C-terminal helix of IpaD (residues 300 to 327; 328 to 332 were contained in our PCR primer). Mutations there might also prevent needle binding, leading to constitutive secretion. Such loss-of-function ipaD mutants are bright red on CR plates and were excluded in our screen. Deletion of the C-terminal three amino acids of IpaB also leads to a mutant that is constitutively ‘on’ (Roehrich et al., 2010). Therefore, the activation signal may be transmitted from the needle tip through the C-terminal helices of IpaD and IpaB. However, structural evidence for the importance of their C-terminal helices in signalling is presently unclear (Chatterjee et al., 2011; Lunelli et al., 2011; Rathinavelan et al., 2011; Barta et al., 2012b; Epler et al., 2012). Therefore, any conformational change that might or might not be occurring in Class II mutants within the tip complex remains to be established.
IpaD secretion upon induction may be only regulatory
Class I mutant ipaDL99P was specifically affected in induction of IpaD secretion in response to CR induction, but not in secretion of the hydrophobic translocators IpaB and IpaC, MxiC or effectors. It is thus unlikely that this mutant is affected in generation or transduction of the activation signal.
Mutant mxiC(E201K, E276K, E293K) shows the same phenotype as ipaDL99P. To our knowledge, these are the first mutants that show a different phenotype for secretion of the hydrophilic and hydrophobic translocators. These mutants thus allowed analysis of IpaD's hypothetical extracellular role after induction (Johnson et al., 2007). Neither mutant shows a defect in invasion, nor is mxiC(E201K, E276K, E293K) affected in survival within (not shown), HeLa cells. Thus, either IpaD secretion is not as reduced in vivo as it is in vitro or IpaD does not have a specific role in host cell invasion besides being the scaffold for translocon formation at the needle tip. Its secretion upon host cell contact might thus serve a regulatory function only: to remove IpaD from the cytoplasm.
IpaD has an intracellular role in the same pathway as MxiC
The Class I mutant ipaDL99P showed increased secretion of IpaB, IpaC and MxiC under non-inducing conditions, while its expression was reduced in comparison to the complemented strain. The IpaD N-terminal domain was proposed to act as an intramolecular chaperone (Johnson et al., 2007; blue domain in Fig. 9). Leucine 99 lies in the helix that connects this domain to the central coiled coil of IpaD. Thus, its mutation to proline might affect the conformation of this domain, leading to reduced protein stability. Even though decreased IpaD levels lead to constitutive secretion, premature secretion in ipaDL99P is not due to decreased expression, but to a specific effect of the mutation. That we found the same secretion phenotype in a mxiC(E201K, E276K, E293K) mutant, which has normal levels of IpaD, further supports this. Interestingly, deletions in the IpaD N-terminal domain (ipaDΔ41–80 and ipaDΔ81–120; Picking et al., 2005), which partially bind the needle tip (Veenendaal et al., 2007), secrete MxiC, IpaB and IpaC prematurely (not shown) like ipaDL99P. In Shigella and Salmonella deletion of IpaD or its homologue leads to constitutive secretion (Menard et al., 1994; Kaniga et al., 1995). In Yersinia and Pseudomonas the chaperone is a separate polypeptide (LcrG/PcrG; Lee et al., 2010; Sato and Frank, 2011) and its deletion has the same effect (DeBord et al., 2001; Lee et al., 2010). Therefore, a regulatory role of IpaD is probably connected to this domain. As T3SS chaperones are rarely secreted, this further suggests that this regulatory function operates intrabacterially. We cannot confirm the presence of an intrabacterial role of IpaD in type III secretion regulation directly by analysing non-secretable mutants as these are unable to form an external tip complex and are hence deregulated for secretion (Veenendaal et al., 2007).
IpaDL99P is also leaked and secreted after CR induction at lower levels than wild-type IpaD. This is not an intrinsic defect of the mutant, but a regulatory defect as the mutant protein is constitutively secreted in a ΔipaB mutant. While we cannot exclude that the decreased IpaD secretion itself is the cause of premature secretion in this mutant, it is more likely to be caused by the regulatory defect as the mxiC(E201K, E276K, E293K) mutant has the same general phenotype, but leaks normal levels of IpaD.
Indeed, our results suggest that secretion is prematurely activated in both the ipaDL99P and mxiC(E201K, E276K, E293K) mutants. In a combined mutant, premature secretion is further increased. We also found that presence of MxiC is necessary for premature secretion in the ipaDL99P mutant. This indicates that in these mutants secretion is activated from within the bacterial cytoplasm. Altogether, our results indicate that MxiC and IpaD act in the same pathway to regulate secretion in response to the absence or presence of the activation signal. Therefore, our findings also suggest a link between LcrG/PcrG and the MxiC homologues YopN/PopN in secretion activation.
We previously suggested that in Shigella, translocator secretion is prevented by a ‘repressor mechanism’ (R), that is put in place when the tip is finished and which is absent in ΔipaD and ΔipaB mutants. We have also shown that cytoplasmic MxiC is required for induction of translocator secretion (Martinez-Argudo and Blocker, 2010). We thus now propose that the mxiC(E201K, E276K, E293K) mutant has lost the ability to promote IpaD secretion but gained the ability to activate secretion of hydrophobic translocators and itself. In addition, the isolation and analysis of the ipaDL99P mutant now suggests that IpaD is directly involved in regulating secretion from the bacterial cytoplasm. It is plausible that the cytoplasmic function of IpaD is to prevent premature secretion before reception of the activation signal, i.e. that IpaD is a repressor of translocator secretion. In this case ipaDL99P would be a loss-of-function mutant as it is unable to fulfil its repression function. Such secreted repressors coupling induction to hierarchical secretion and expression are common in T3SSs (Brutinel and Yahr, 2008; Chevance and Hughes, 2008). Alternatively, IpaD could act as a co-activator of translocator secretion alongside MxiC. In that case, IpaDL99P would fulfil this function even though it should only do so after activation, thus it would be a gain-of-function mutant for translocator secretion. Either way, it is now clear that MxiC and IpaD are both involved in regulation of an early step in secretion activation within the bacterial cytoplasm.
We demonstrate here that IpaD controls secretion both by regulating the functional state of a transmembrane macromolecular machine and as part of a signal transduction pathway that is dependent on cytoplasmic proteins. By genetically separating these activities (in Class I vs. Class II mutants), we demonstrate that one protein pool acts indirectly on another pool of the same protein to orchestrate T3SS activation. To our knowledge, this has rarely been reported so far, possibly because such functions are difficult to separate. It may thus represent a novel regulatory paradigm.
Bacterial strains, plasmids and primers
Table 2 lists the strains and plasmids used in this study. S. flexneri strains were grown in Trypticase Soy Broth (Becton Dickinson) at 37°C with the appropriate antibiotics at the following final concentrations: ampicillin 100 μg ml−1, kanamycin 50 μg ml−1, tetracycline 5 μg ml−1, chloramphenicol 10 μg ml−1. For the inducers IPTG and arabinose we used the final concentrations indicated in the figure legends. Table S1 lists the primers used in this study.
Table 2. Shigella flexneri strains and plasmids used in this study
PCR mutagenesis was carried out with Taq DNA polymerase (New England Biolabs) by using standard or error prone (3 mM Mg2+ and 4 mM Mg2+) reaction conditions using primers ipaD_NdeI and ipaD_BamHI and pUC18_ipaD as template. PCR fragments were purified, digested with NdeI and BamHI and cloned into plasmid pUC18_ipaD (Picking et al., 2005) digested with the same enzymes. Ligation mixtures were electroporated into DH5α. The transformation mixture was incubated for 16 h, and plasmid DNA was then extracted and kept at −20°C.
Screening of the ipaD mutant libraries
To identify non-inducible mutants, the DNA from each of the three mutant libraries was electroporated into S. flexneri ΔipaD mutant and screened for white colonies on TCSB agar plates containing 100 μg ml−1 CR (Serva). Putative ‘white’ colonies were selected and their plasmids isolated and retransformed into the ΔipaD mutant to ensure the white colour was not due to loss-of-function mutations elsewhere within the T3SS-encoding operons. Candidate plasmids were sequenced to identify the mutation(s) responsible for the mutant phenotype.
Construction of non-polar knockout mutant strains
The ΔipaB ΔipaD mutant was generated by using the Lambda Red system as described by Datsenko and Wanner (2000). Briefly, to replace the wild-type ipaD gene, a tetracycline resistance cassette with 50 bp flanking regions homologous to the ipaD gene was amplified from strain TH2788 (Frye et al., 2006) using the primers ipaD_KO_tetF and ipaD_KO_tetR (Table S1). The ipaD gene was exchanged for a tetracycline cassette in the single mutant strain SF620 (ΔipaB) following the procedure previously described (Martinez-Argudo and Blocker, 2010). To generate strain ΔipaD ΔmxiC, the same procedure was used to exchange the mxiC gene for a tetracycline cassette in the mutant strain SF622 (ΔipaD) using the primers mxiC_KO_tetF and mxiC_KO_tetR (Table S1). All gene replacements were confirmed by sequencing.
Construction of plasmids
To express the ipaD mutations in combination with ipaBΔ3, a plasmid encoding both ipaBΔ3 and ipaD was constructed. The ipaD or ipaDL99P genes were amplified by PCR using primers ipaD_PstI_F and ipaD_BamHI (Table S1) and pUC18_ipaD or pIMA230 (pUC18_ipaDL99P) as template. PCR products were purified, digested with PstI and BamHI and cloned into pDR2 (pUC19_ipaBΔ3) digested with the same enzymes, giving rise to plasmids pIMA241 and pIMA242 respectively.
To construct the double ipaD mutants, ipaD(L99P, K291Y) and ipaD(N186Y, K291Y), a HindIII-BamHI fragment from plasmid pIMA233 (ipaDK291I) was cloned into pIMA230 and pIMA231, digested with the same enzymes, giving rise to pIMA238 and pIMA237 respectively.
Plasmid pIMA239 (ipaD(N292Y, T296I, Q299L)) was made by two-step PCR. The first step consisted of two PCR reactions, one carried out with primer ipaD_NdeI and reverse primer ipaD_tripleR containing the three desired mutations and the other with forward primer ipaD_tripleF containing the mutations and reverse primer ipaD_BamHI (Table S1). The two PCR fragments were used as a template for the second PCR using primers ipaD_NdeI and ipaD_BamHI. The PCR product was purified, digested with HindIII and BamHI and cloned into pUC19_ipaD digested with the same enzymes.
To express ipaD and ipaDL99P from an inducible plasmid we chose plasmid pBAD/Myc_HisA (Invitrogen). The ipaD or ipaDL99P genes were amplified by PCR using primers ipaD_NcoI_F and ipaD_PstI_rev (Table S1) and pUC18_ipaD or pIMA230 (pUC18_ipaDL99P) as template. PCR products were purified, digested with NcoI and PstI and cloned into pBAD/Myc_HisA digested with the same enzymes, giving rise to plasmids pIMA243 and pIMA244 respectively.
To overexpress the needle protein MxiH, the mxiH gene was amplified by PCR using the Shigella virulence plasmid pWR100 (Buchrieser et al., 2000) as template and primers mxiH_RBS and mxiH_HindIII (Table S1). The PCR product was purified, digested with SacI and HindIII, and cloned into the IPTG inducible plasmid pACT3 (Dykxhoorn et al., 1996) giving rise to pIMA212.
Plasmid pDR67 (mxiC(E201K, E276K, E293K)) was made by three subsequent PCR steps. In the first step four PCR fragments were generated using pWR100 as a template and primers mxiC_SacI and mxiC_E201K_rev, mxiC_E201K_for and mxiC_E276K_rev, mxiC_E276K_for and mxiC_E293K_rev and mxiC_E293K_for and mxiC_BamHI (Table S1). In the second step the first two and the last two fragments were combined using primers mxiC_SacI and mxiC_E276K_rev or mxiC_E276K_for and mxiC_BamHI respectively. The final PCR product was generated by combining these fragments using primers mxiC_SacI and mxiC_BamHI. It was purified and digested with SacI and BamHI. The digested fragment was cloned into pACT3 (Dykxhoorn et al., 1996) restricted with the same enzymes.
All plasmids were verified by sequencing.
Analysis of protein synthesis and secretion
Total level of protein expression
Shigella flexneri strains were grown at 37°C until mid-exponential growth (OD600 ≈ 1) was reached. Samples of the cultures, representing the total protein fraction, were denatured in Laemmli sample buffer. Samples from equivalent cell numbers were separated by SDS-PAGE and used for Western blot analysis.
Shigella flexneri strains were grown until OD600 ≈ 1 (mid-exponential phase). Cultures were centrifuged at 15 000 g for 10 min at 4°C and supernatants from equivalent cell numbers were denatured in Laemmli sample buffer, subjected to SDS gel electrophoresis and Silver-stained (Silver Xpress kit, Invitrogen) or used for Western blot analysis.
Congo red induction
Bacteria collected during mid-exponential growth (OD600 ≈ 1) were resuspended at OD600 = 5 in phosphate-buffered saline (PBS). CR (Serva) was added at a final concentration of 200 μg ml−1 to induce T3SS activity. After incubation at 37°C for 15 min, the samples were centrifuged at 15 000 g for 10 min at 4°C and the supernatants were denatured, separated by SDS-PAGE and Silver-stained (Silver Xpress kit, Invitrogen) or used for Western blot analysis.
Western blot analysis
For Western blot analysis, proteins separated by SDS-PAGE were transferred onto PVDF membrane (Immobilon FL, Millipore) and hybridized with the following primary antibodies: IpaB, IpaC, IpaD and MxiC (see Martinez-Argudo and Blocker, 2010 for details). Fluorescent secondary antibodies (goat anti-rabbit-Alexa680, Invitrogen; goat anti-rabbit-DyLight800 and goat anti-mouse-DyLight800, Pierce) were visualized and quantified using an Odyssey infrared imaging system (Li-Cor). For quantifications, standard curves obtained by serial dilution of a sample from the complemented strain were run on the same gel to control for linearity of the signal. Numbers are averages of two independent experiments.
To analyse tip complex compositions we isolate needles sheared from the surface of bacteria that overexpress the needle protein MxiH and hence make very long needles (as normal length needles are too short to be efficiently sheared off). Needles were purified from non-activated strains containing plasmid pIMA212 (pACT3_mxiH) as described in Veenendaal et al. (2007). Briefly, 250 ml of cultures were grown in TCSB overnight with 200 μM IPTG to induce overproduction of MxiH from plasmid pIMA212. Needles were sheared off, separated by differential centrifugation and precipitation and further purified by size exclusion chromatography and fractions containing needles were pooled. Samples were denatured in Laemmli sample buffer, separated by SDS-PAGE and Silver-stained (Silver Xpress kit, Invitrogen) to normalize the amount of MxiH prior to Western blotting to detect IpaB and IpaD.
Shigella flexneri invasion of HeLa cells was assessed with a gentamicin protection assay as described previously (Roehrich et al., 2010). Briefly, HeLa cells were seeded into 24-well plates and infected with bacteria grown to mid-exponential phase at a multiplicity of infection of 100. After centrifugation for 10 min at 900 g, the infected cells were incubated for 30 min at 37°C. Cells were then washed with PBS and gentamicin was added to kill extracellular bacteria. After further incubation for 1 h, cells were lysed in 0.1% Triton X-100; a dilution series was prepared and plated on LB agar plates. Colonies were counted the next day.
This work was supported by Wellcome Trust project grant 088231 to AJB and IMA and by a University of Bristol Centenary Postgraduate studentship to ADR. We thank Richard Sessions (Bristol) for helpful discussions.