Activation of bacterial virulence-associated type III secretion systems (T3SSs) requires direct contact between a bacterium and a eukaryotic cell. In Yersinia pestis, the cytosolic LcrG protein and a cytosolic YopN-TyeA complex function to block T3S in the presence of extracellular calcium and prior to contact with a eukaryotic cell. The mechanism by which the bacterium senses extracellular calcium and/or cell contact and transmits these signals to the cytosolic compartment is unknown. We report here that YscF, a small protein that polymerizes to form the external needle of the T3SS, is essential for the calcium-dependent regulation of T3S. Alanine-scanning mutagenesis was used to identify YscF mutants that secrete virulence proteins in the presence and absence of calcium and prior to contact with a eukaryotic cell. Interestingly, one of the YscF mutants that exhibited constitutive T3S was unable to translocate secreted proteins across the eukaryotic plasma membrane. These data indicate that the YscF needle is a multifunctional structure that participates in virulence protein secretion, in translocation of virulence proteins across eukaryotic membranes and in the cell contact- and calcium-dependent regulation of T3S.
Yersinia pestis, the causative agent of plague, employs a type III secretion system (T3SS) to inject effector proteins, termed Yops, into eukaryotic cells (Cornelis, 2002). Yop secretion is triggered by contact between a bacterium and the surface of a eukaryotic cell. Effector Yops are injected into the targeted cell and are not found in substantial amounts in the extracellular milieu, indicating that the T3S process is activated only at the point of contact between the two cells (Forsberg et al., 1994; Rosqvist et al., 1994). In vitro, Yop secretion is blocked by millimolar levels of extracellular calcium and triggered by the removal of calcium from the growth medium (Michiels et al., 1990). The block in Yop secretion in the presence of extracellular calcium and prior to contact with a eukaryotic cell is dependent upon the secreted YopN protein (Forsberg et al., 1991) and the cytosolic TyeA (Iriarte et al., 1998; Cheng and Schneewind, 2000), SycN (Iriarte and Cornelis, 1999), YscB (Day and Plano, 1998) and LcrG (DeBord et al., 2001; Matson and Nilles, 2001) proteins. Mutational inactivation of any one of these genes results in unregulated, constitutive Yop secretion. Recent evidence suggests that both LcrG and the YopN-TyeA complex block Yop secretion from a cytosolic, not an extracellular, location (Cheng et al., 2001; Matson and Nilles, 2001; Ferracci et al., 2005). These results suggest that the bacterium requires a mechanism to sense and transduce the extracellular signal of low calcium or contact with a eukaryotic cell to the cytosolic compartment.
Virulence-associated T3SSs consist of a base structure that spans the bacterial inner and outer membranes and of a needle structure that extends approximately 40–60 nm from the surface of the bacterial cell (Kubori et al., 1998; Blocker et al., 1999; Hoiczyk and Blobel, 2001). The external needle primarily consists of a single small protein, termed YscF in Yersinia spp., that is secreted through the T3SS and subsequently polymerizes to assemble into the surface-exposed needle structure (Hoiczyk and Blobel, 2001; Journet et al., 2003). Current evidence suggests that the needle is a hollow tube that forms a conduit for protein secretion. In the presence of calcium the T3SS secretes only needle-type substrates (YscF and YscP) until needle assembly is completed (Edqvist et al., 2003). The YscP protein functions as a molecular ruler that directly determines the length of the needle (Journet et al., 2003). Upon completion of needle assembly, a type III secretion substrate specificity switch (T3S4) domain in the C-terminus of YscP directs a switch in substrate specificity from needle-type substrates to Yop-type substrates (Edqvist et al., 2003; Agrain et al., 2005). At this point the YopN-TyeA complex is predicted to be targeted to the T3SS where it blocks effector Yop secretion until the appropriate triggering signals are encountered. Yersinia strains deficient in expression of YscF do not make needles and cannot secrete Yops (Allaoui et al., 1995; Journet et al., 2003). In contrast, strains deficient in expression of YscP, secrete YscF constitutively and produce very long needles (Journet et al., 2003).
The assembly of the flagellar rod, hook and filament also requires a T3SS (Macnab, 2004). Rod, hook and filament subunits are transported through the flagellum's hollow central channel before being assembled at its tip. Remarkably, the assembled flagellar filament and the assembled needle structure of the Shigella flexneri virulence-associated T3SS have been shown to share a common helical architecture (∼5.6 subunits/turn with a 24 Å helical pitch), suggesting that the assembly of these surface structures follows a similar pathway (Samatey et al., 2001; Blocker et al., 2003; Cordes et al., 2003; Yonekura et al., 2003). The flagellar filament is capable of adopting alternate helical forms depending on the direction of filament rotation. These alternate helical structures play an important role in bacterial chemotaxis. We hypothesize that bacterial T3SS needles might also employ alternate conformations and/or have alternate functions. In this study we investigate the role of the Y. pestis YscF needle subunit in the assembly of a functional cell contact- and calcium-regulated T3SS.
YscF is required to block Yop secretion in the presence of extracellular calcium
Alanine- and cysteine-scanning mutagenesis were used to identify YscF amino acid residues required to assemble a functional calcium- and cell contact-regulated T3SS. YscF residues were selected for mutagenesis based on the following criteria. First, residues highly conserved among all YscF homologues (I82, D77, L16 and S69) or specifically among YscF homologues from calcium-regulated T3SSs (D17, K24, N31, V34, D36, D46, N59, R73) were selected. Second, all amino acid residues with negatively charged side chains (D12, D15, D28, D29, D43, D53) potentially capable of high-affinity interactions with calcium were selected (Dudev and Lim, 2003). Lastly, several residues that were not conserved among YscF homologues (I13, K25, K32, S74) were selected. Oligonucleotide site-directed mutagenesis of plasmid pYscF was performed to replace the selected residues with alanine and/or cysteine (Table 1). The pYscF derivatives were electroporated into a yscF deletion strain and the expression and secretion phenotypes of the resultant strains were determined.
Table 1. Secretion phenotype of Y. pestis YscF point mutants.
Strain + plasmid
Strain + plasmid
RS, calcium-regulated secretion; CS, constitutive secretion in the presence and absence of calcium; NS, no secretion.
ΔyscF+ pYscF (K32A)
ΔyscF+ pYscF (K32C)
ΔyscF+ pYscF (V34A)
ΔyscF+ pYscF (D12A)
ΔyscF+ pYscF (D36A)
ΔyscF+ pYscF (D12C)
ΔyscF+ pYscF (D36C)
ΔyscF+ pYscF (I13A)
ΔyscF+ pYscF (D43A)
ΔyscF+ pYscF (I13C)
ΔyscF+ pYscF (D46A)
ΔyscF+ pYscF (D15C)
ΔyscF+ pYscF (D46C)
ΔyscF+ pYscF (L16A)
ΔyscF+ pYscF (D53A)
ΔyscF+ pYscF (D17A)
ΔyscF+ pYscF (N59A)
ΔyscF+ pYscF (D17C)
ΔyscF+ pYscF (N59C)
ΔyscF+ pYscF (K24A)
ΔyscF+ pYscF (S69A)
ΔyscF+ pYscF (K24C)
ΔyscF+ pYscF (S69C)
ΔyscF+ pYscF (K25A)
ΔyscF+ pYscF (R73A)
ΔyscF+ pYscF (K25C)
ΔyscF+ pYscF (R73C)
ΔyscF+ pYscF (D28A)
ΔyscF+ pYscF (S74A)
ΔyscF+ pYscF (D28C)
ΔyscF+ pYscF (S74C)
ΔyscF+ pYscF (D29A)
ΔyscF+ pYscF (D77A)
ΔyscF+ pYscF (D29C)
ΔyscF+ pYscF (D77C)
ΔyscF+ pYscF (N31A)
ΔyscF+ pYscF (I82A)
ΔyscF+ pYscF (N31C)
ΔyscF+ pYscF (I82C)
Thirty-nine mutants, with single alanine or cysteine substitutions in YscF, were analysed for YscF expression and Yop secretion following growth in the presence or absence of 2.5 mM calcium in TMH medium for 5 h at 37°C (Table 1). All of the YscF mutants expressed a stable YscF protein; however, the expression of YscF (D17C), YscF (L16A) and YscF (N59A) was significantly lower than that of the parent strain and of the other YscF mutants (data not shown). The parent strain Y. pestis KIM5-3001.P39, the yscF deletion strain complemented with pYscF, and 26 of the 39 YscF mutants exhibited normal calcium-regulated secretion (RS) of Yops. The yscF deletion strain and seven YscF point mutants (D17C, N31A, V34A, D77A, D77C, I82A, I82C) showed no secretion (NS) of Yops in the presence or absence of calcium. Interestingly, six YscF point mutants (I13C, D17A, D28A, D28C, D46A, D46C) exhibited constitutive secretion (CS) of Yops in the presence and absence of calcium. These results suggest that YscF, in addition to its established and essential function in Yop secretion, plays a role in the regulation of Yop secretion in response to extracellular calcium.
The Y. pestis T3SS is encoded on an approximately 70 kb virulence plasmid termed pCD1 (Perry et al., 1998). To confirm that the secretion phenotypes associated with select yscF mutations were due to specific amino acid replacements in YscF and not to other factors associated with multicopy expression, we regenerated each of these point mutations directly in yscF of plasmid pCD1. Y. pestis strains encoding YscF mutants with either single- (I13A, D17A, D28A, D46A, D77A) or double-alanine substitutions (D28A D46A and D28A D77A) were generated via oligonucleotide site-directed mutagenesis and allelic exchange. Double mutants were constructed to determine the phenotype of a Y. pestis strain that carries both a yscF mutation that results in a CS phenotype (D28A) and another yscF mutation that results in either a CS (D46A) or an NS (D77A) phenotype. The resultant YscF mutants, which carry defined mutations in yscF of pCD1, were used in all subsequent studies.
Secretion of YopM and YopN by the YscF mutants was measured following growth of the bacteria in the presence and absence of 2.5 mM calcium in TMH medium for 5 h at 37°C (Fig. 1A). Analysis of Yop secretion patterns confirmed that the YscF mutants could be divided into three distinct classes. The parent strain and the I13A mutant exhibited normal RS of Yops. The yscF deletion mutant, the D77A mutant, and the D28A D77A double mutant showed NS of Yops in the presence or absence of calcium. Lastly, the yopN and tyeA deletion mutants, the D17A, D28A and D46A mutants, and the D28A D46A double mutant exhibited CS of Yops in the presence and absence of calcium. Complementation of each individual YscF mutant with plasmid pYscF restored normal RS of Yops (Fig. 2B), indicating that the defects associated with these strains were due to the amino acid change in YscF and not to polar effects or spontaneous mutations in other genes.
To confirm that the D28A, D46A and the D28A D46A double mutant constitutively secrete all highly expressed T3S substrates, select yscF point mutations were moved into plasmid pCD1 of Y. pestis KIM8.P39, a strain that does not express the plasminogen activator protease (Pla). The Pla protease is an outer membrane protein present in Y. pestis, but not in Yersinia enterocolitica or Yersinia pseudotuberculosis, that is responsible for the degradation of secreted Yops (Protsenko et al., 1983). Concentrated culture supernatant proteins from the parent, a yscF deletion mutant and the YscF point mutants were separated by SDS-PAGE and stained with Coomassie Blue R-250 (Fig. 1B). The parent strain and the I13A mutant showed normal RS of Yops. The yscF deletion mutant and the D77A mutant showed NS of Yops in the presence or absence of calcium. In contrast, a yopN deletion mutant, as well as the D28A, D46A and the D28A D46A double mutant secreted large amounts of Yops in both the presence and absence of calcium. These results confirm that YscF, like YopN, plays a critical role in establishing the calcium-dependent block of Yop secretion. Furthermore, we demonstrate that this novel YscF function can be genetically separated from YscF's function in Yop secretion. These results represent the first evidence that T3SS needles perform a role in the regulation of the T3S process.
Surface-exposed YscF is degraded by the Pla protease
SDS-PAGE and immunoblot analysis of YscF expression by both Pla+ and Pla–Y. pestis strains demonstrated that the approximately 9 kDa YscF protein is a substrate for the Pla protease (Fig. 2A). Expression and secretion of YscF by each of the YscF point mutants was determined by SDS-PAGE and immunoblot analysis (Fig. 2B). All of the YscF proteins were expressed, secreted and subsequently degraded by Pla with the exception of YscF (D77A) and YscF (D28A D77A). No export or degradation of YscF was detected in the D77A and the D28A D77A double mutant, suggesting that the mutant YscF proteins expressed by these strains were not surface-exposed. Essentially identical results were obtained using Pla- strains (data not shown), confirming that the absence of YscF (D77A) and YscF (D28A D77A) was likely due to a defect in the YscF-dependent T3S process and not due to rapid degradation of exported YscF by Pla.
Yersinia pestis YscF CS mutants assemble YscF needle polymers on their surface
To confirm that the I13A, D28A, D46A and the D28A D46A double mutant assemble T3S needles on their surface, we employed a protein cross-linking assay previously used to monitor needle assembly in Y. pestis (Ferracci et al., 2005). Bacterial cells grown at 37°C in the presence or absence of calcium were harvested, washed and treated with 5 mM BS3, a membrane impermeable, homobifunctional protein cross-linking agent. The presence of an anti-YscF antibody reactive ladder of bands that corresponded in size to various YscF-YscF multimers plus a variable number of BS3 molecules (mw = 572.43) was observed in samples obtained from the parent strain, I13A, D28A, D46A and the D28A D46A double mutant grown in both the presence and absence of calcium (Fig. 3). No YscF multimers were observed in the absence of cross-linker or in a secretion defective yscJ deletion mutant, the later observation confirming that BS3 does not enter the bacterial cell. These data indicate that an assembled YscF needle is present on the surface of the parent strain and the YscF CS mutants in both the presence and absence of calcium.
Yersinia pestis YscF CS mutants show altered responses to extracellular calcium
During an infection of a mammalian host, Y. pestis is primarily located extracellularly (Cornelis, 2000). The level of calcium present in this extracellular fluid or blood is approximately 1.0–1.2 mM. A similar level of extracellular calcium is required to block Yop secretion prior to contact with a eukaryotic cell. The YscF needle is one of a limited number of T3SS components that are exposed to the extracellular environment prior to the initiation of Yop secretion. This suggests that the YscF needle could play a direct role in sensing and/or responding to extracellular calcium. The ability of individual YscF mutants to sense and respond to extracellular calcium was investigated by measuring the secretion of YopM and YopN from bacteria grown in the presence of varying amounts of extracellular calcium (Fig. 4). No significant change in the amount of YscF associated with the bacterial cells was observed in response to the different levels of extracellular calcium (Fig. 4A), indicating that calcium does not dramatically alter the expression or stability of the different YscF proteins. As expected, secretion of YopM and YopN by the parent strain was blocked by extracellular calcium levels of 1.25 mM or higher (Fig. 4B). On the other hand, the I13A, D28A and D46A mutants required 2.5 mM, 5.0 mM and 7.5 mM calcium, respectively, in order to block Yop secretion. Furthermore, the D28A D46A double mutant, and the yopN and tyeA deletion mutants, secreted YopM and YopN regardless of the amount of extracellular calcium. Thus, the I13A, D28A, D46A and D28A D46A mutants exhibited increasing incremental defects in their ability to block Yop secretion in response to the level of extracellular calcium. These YscF mutants represent the first examples of T3SS mutants that require increased levels of calcium to block Yop secretion and suggest that YscF may play a role in sensing and/or responding to extracellular calcium. Interestingly, both of the residues that were mutated in the YscF CS mutants were aspartic acid residues, a common constituent of many calcium binding sites (Dudev and Lim, 2003).
Yersinia pestis YscF mutants with altered responses to calcium exhibit defects in Yop translocation into HeLa cells
Upon contact with the surface of a eukaryotic cell, the T3SS translocates six effector Yops directly into the cytosol of the eukaryotic cell (Cornelis, 2002). The cytosolic LcrG- and YopN-TyeA-dependent block in Yop secretion ensures that Y. pestis does not secrete the Yops prior to contact with a host cell (Cheng et al., 2001; Day et al., 2003). To assess the role of YscF in the cell contact-dependent regulation of Yop secretion and translocation, we measured the ability of the various YscF mutants to translocate YopE into HeLa cells using a protein tag-based reporter system (Day et al., 2003). The parent strain, the yopB, tyeA, yopN and yscF deletion strains, and the YscF mutants, all of which also carry a deletion in yopE, were transformed with plasmid pYopE129-Elk. This plasmid codes for a truncated YopE protein with a C-terminal Elk-tag that is phosphorylated upon translocation into a eukaryotic cell. Following a 3 h infection of HeLa cells, the culture supernatants were removed, and the infected HeLa cell monolayers and the culture supernatants were analysed by SDS-PAGE and immunoblot analysis (Fig. 5). Total YopE129-Elk present in both fractions was detected with anti-Elk antipeptide antibodies. YopE129-Elk translocated into the HeLa cells was phosphorylated and detected with phospho-specific anti-Elk antipeptide antibodies.
All of the Y. pestis strains carrying pYopE129-Elk expressed approximately equal amounts of the YopE129-Elk protein. The parent strain translocated YopE129-Elk into the HeLa cells and released only small amounts of the tagged protein into the culture supernatant. As expected, the yscF deletion mutant (NS phenotype) and the D77A mutant (NS phenotype) did not secrete the YopE129-Elk protein into the culture supernatant and did not translocate YopE129-Elk into the HeLa cells. The I13A, D28A, D46A and D28A D46A mutants (CS phenotype), which required progressively greater amounts of calcium to block secretion (Fig. 4), showed an analogous progressive defect in their ability to translocate the YopE129-Elk protein into HeLa cells. The progressive defect in the ability of these strains to direct the YopE129-Elk protein into HeLa cells (D28A D46A < D46A < D28A < I13A) correlated with a progressive increase in the amount of YopE129-Elk secreted into the culture supernatant (D28A D46A > D46A > D28A > I13A). The CS phenotype of the D28A and D46A mutants was similar to that of the yopN and tyeA deletion strains, in that each of these strains translocated Yops inefficiently, but secreted increased amounts of Yops into the culture supernatant. Interestingly, the D28A D46A double mutant, which also secreted Yops constitutively, was unable to translocate the YopE129-Elk protein into HeLa cells. These results indicate that YscF is required to translocate secreted Yops across the eukaryotic membrane. Indeed, in the HeLa cell infection assay, the phenotype of the D28A D46A mutant was similar to that of the translocation defective yopB deletion mutant, which cannot form translocation-related pores in the eukaryotic membrane (Sory and Cornelis, 1994; Hakansson et al., 1996; Nordfelth and Wolf-Watz, 2001), although the YscF mutant secreted more YopE129-Elk into the culture supernatant. The inability to translocate the YopE129-Elk protein into HeLa cells was not solely due to the CS phenotype of the D28A D46A mutant, because all of the other CS mutants translocated significant amounts of this protein.
The Y. pestis YscF D28A D46A CS mutant is not cytotoxic to HeLa cells
Translocation of effector Yops into HeLa cells disrupts the cell's cytoskeleton and leads to an altered cell morphology that can be easily visualized using standard light microscopy (Rosqvist et al., 1991). To further evaluate the role of YscF in Yop translocation, we transformed the YscF mutants and control strains with plasmid pYopE (Day et al., 2003), which restores expression of the full-length YopE protein. The cytotoxicity of Y. pestis KIM8.P39 (parent), isogenic yopN, yopB and yscF deletion mutants and the isogenic YscF point mutants, all carrying plasmid pYopE, were evaluated using cultured HeLa cells as previously described (Rosenzweig et al., 2005). HeLa cells were infected at a multiplicity of infection (moi) of 50 for 2 h at 37°C. The morphology of the HeLa cells infected with the parent strain, the yopN deletion strain, the I13A mutant, and the D28A mutant displayed a dramatic cytotoxic effect, indicating that these strains injected sufficient quantities of effector Yops to disrupt the cell's cytoskeleton (Fig. 6). In contrast, HeLa cells infected with the translocation-defective yopB mutant, the secretion-defective yscF mutant and the D28A D46A mutant displayed a normal HeLa cell morphology, confirming that these strains were defective in Yop secretion and/or translocation. Interestingly, HeLa cells infected with the D46A mutant displayed an intermediate phenotype, suggesting that Yop injection was reduced in this strain compared with the parent strain and to the yopN deletion mutant. Complementation of the D46A and the D28A D46A mutants with plasmid pYscF restored the ability of these strains to elicit cytotoxicity on HeLa cells to parental levels. These results confirm that the D28A D46A mutant is unable to translocate Yops into a eukaryotic cell, a phenotype similar to that of a yopB or yopD deletion mutant. These results indicate that T3SS needles play a role in the translocation process, independent of their essential role in T3S.
Role of YscF in the calcium-dependent regulation of Yop secretion
The dramatic effect of calcium on the growth of Y. pestis in culture was documented over 40 years ago (Higuchi et al., 1959). Likewise, the first demonstration that calcium functions to regulate Yop secretion was presented in 1984 (Heesemann et al., 1984). In the time since these dramatic discoveries, the CR growth properties and the CR T3SSs of all three human pathogenic yersiniae have been studied extensively; however, no protein or other molecule that both directly interacts with calcium and is involved in the T3S process has been identified. Thus, the mechanism by which the yersiniae sense and respond to extracellular calcium has remained a mystery. In this report we present data which demonstrate that the surface-localized YscF needle of the T3SS plays a role in the calcium-dependent regulation of Yop secretion in Y. pestis.
Previous studies have demonstrated that cytosolic LcrG and a cytosolic YopN-SycN-YscB-TyeA complex are required to block Yop secretion in the presence of calcium and prior to contact with a eukaryotic cell (Cheng et al., 2001; Day et al., 2003). The current model (see Fig. 7) suggests that the SycN-YscB chaperone is required to target the YopN-TyeA complex to the cytosolic face of the T3S complex (Day and Plano, 1998). In the presence of extracellular calcium, YopN initiates, but does not complete its secretion. Cytosolic TyeA, and possibly LcrG, are required to arrest the export of YopN (Cheng et al., 2001; Day et al., 2003). The arrested YopN-TyeA complex is hypothesized to disrupt the T3S process or directly block the entrance to the T3S pathway. Upon contact with a eukaryotic cell (or upon the removal of calcium), an unidentified signalling event activates the T3S process, triggering the secretion of YopN, immediately followed by secretion of the effector Yops. The mechanism by which the bacterium senses cell contact and removes the calcium-dependent block in Yop secretion is not understood.
The results presented here suggest that the YscF needle may be the extracellular T3SS component that senses extracellular calcium and/or participates in transmitting the calcium signal to the cytoplasmic compartment where the block in secretion is initiated. Several lines of evidence support this contention. First, YscF is present on the bacterial cell surface in the presence of calcium and prior to the initiation of Yop secretion. Second, calcium-dependent regulation of the T3S process is not imposed until assembly of the YscF needle is completed (Hoiczyk and Blobel, 2001). Third, we have identified YscF point mutants that secrete Yops constitutively in the presence and absence of calcium and prior to contact with a eukaryotic cell, indicating that YscF, like YopN and TyeA, is required to block Yop secretion in the presence of calcium. Fourth, a number of the YscF CS mutants required increased amounts of calcium to block Yop secretion (for example, D46A required a sixfold increase in extracellular calcium levels compared with the parent strain), suggesting that these mutations affected the ability of the T3SS to sense and/or respond to calcium. Fifth, Hoiczyk and Blobel (2001) demonstrated that the addition of calcium to purified Y. enterocolitica YscF needles caused the needles to aggregate, suggesting that calcium may affect the surface properties of the YscF needle. Lastly, three YscF mutants (D17A, D28A, D46A) that exhibited a CS phenotype were substitutions of alanine for aspartic acid, a common constituent of calcium binding sites (Dudev and Lim, 2003).
Electron micrographs of the Salmonella enterica SPI1 T3SS needle complex indicates that the PrgJ/PrgI needle structure originates from a location close to the inner membrane (Kimbrough and Miller, 2000; Marlovits et al., 2004). Consequently, changes in the conformation or helicity of the needle structure in response to changes in the extracellular environment could theoretically be propagated along the length of the needle to the inner membrane secretion complex where the T3S block is predicted to occur (Fig. 7). Alternatively, changes in the conformation of the needle structure could disrupt specific protein–protein interactions (For example YscF-YscP or YscF-YopN) that may be required to maintain the block in T3S.
Role of YscF in the cell-contact-dependent regulation of Yop secretion
We hypothesize that contact with a eukaryotic cell could generate a low-calcium signal that directly triggers the T3S process. Penetration of the eukaryotic membrane by the YscF needle may allow the bacterium to directly sense the low-calcium environment of eukaryotic cytosol, generating a signalling event that could be propagated to the base of the T3SS. Alternatively, the T3SS needle could sense the low calcium environment of eukaryotic cytosol by directly docking with the pore-forming LcrV–YopB–YopD translocation complex (Neyt and Cornelis, 1999; Marenne et al., 2003). If correct, either of these hypotheses would provide a plausible explanation for why only the needle complexes that are in direct contact with a eukaryotic cell are activated. Although, the hypothesis that T3SS needles directly penetrate eukaryotic membranes is controversial, several lines of evidence suggest that this may be possible. First, Hoiczyk and Blobel (2001) have demonstrated that isolated Y. enterocolitica YscF needles are hydrophobic in nature. Furthermore, Hoiczyk and Blobel (2001), Matsumoto et al. (1976), and Bavoil and Hsia (1998) have shown electron micrographs of T3SS needles that appear to penetrate eukaryotic membranes. On the other hand, Y. enterocolitica strains that express YscF needles, but no LcrV, YopB or YopD are unable to form translocation pores or inject Yops into eukaryotic cells (Marenne et al., 2003). Furthermore, the calcium-regulated T3SS of enteropathogenic Escherichia coli (EPEC) uses a filament-like structure composed of the EspA protein to connect the EscF needle structure to the translocon components embedded in the eukaryotic membrane (Sekiya et al., 2001; Wilson et al., 2001). These studies suggest that direct contact between the needle and the eukaryotic plasma membrane is not sufficient, or even required, to trigger the injection of virulence proteins. Further studies will be required to determine the role of YscF in the cell contact-dependent events that trigger Yop secretion in response to cell contact.
Role of YscF in the translocation of Yops across the eukaryotic membrane
In addition to YscF's role in secretion and in the regulation of secretion, the results presented suggest that the YscF needle also plays a role in translocating effector Yops across the eukaryotic plasma membrane. The D28A D46A double mutant secreted wild-type levels of Yops in the presence and absence of calcium in vitro, but was completely defective for Yop translocation. Furthermore, the defect in translocation associated with this mutant could be complemented by providing a wild-type copy of yscF in trans, confirming that the translocation defect was due to the alanine substitutions in YscF. The expression and assembly of YscF (D28A D46A) needles could disrupt the translocation process through a number of different mechanisms. One possibility is that the YscF (D28A D46A) needle has a structure, or is locked in a conformation, that is unable to interact productively with the eukaryotic plasma membrane or is unable to dock with the LcrV–YopB–YopD translocation complex. Alternatively, the D28A and D46A substitutions may disrupt the YscP-dependent needle length control mechanism. Mota et al. (2005) have recently demonstrated that needle length needs to be coordinated with the length of specific adhesins in order to ensure efficient Yop translocation. BS3 cross-linking studies indicate that the I13A, D28A, D46A, as well as the D28A D46A double mutant assemble needle structures on their surfaces; however, transmission electron microscopy (TEM) failed to visualize the wild-type YscF, YscF (I13A), YscF (D28A), YscF (D46A) or YscF (D28A D46A) needles in Y. pestis. Interestingly, deletion of the yscP gene from the parent, I13A, D46A and the D28A D46A double mutant resulted in the expression of a small number of extremely long YscF needles (200–1700 nm in length; data not shown) that could be readily visualized by TEM. The needles were similar in size and dimension to the needles observed in a Y. enterocolitica yscP deletion mutant (Journet et al., 2003). These observations confirm that the YscF CS mutants can assemble needle structures. Although the length of the needles in the YscP + background could not be determined, no extremely long needles were visualized, indicating that YscP-dependent needle length control was likely operating in both the wild-type and YscF CS mutants.
Calcium overlay experiments failed to demonstrate direct binding of calcium to a maltose-binding protein (MBP)–YscF fusion protein (data not shown). These results were not unexpected given that millimolar levels of extracellular calcium are required to block Yop secretion. These types of experiments are typically used to identify eukaryotic calcium binding proteins with Kd values in the nM or low µM range. Moreover, calcium binding to YscF may require assembly of the YscF needle. Future studies will be aimed at purifying the assembled wild-type and mutant YscF needles. Purified needles will facilitate calcium binding studies, as well as studies aimed at identifying any helical or conformational changes that may occur in the presence or absence of calcium.
Bacterial strains and growth conditions
Yersinia pestis strains were routinely grown in heart infusion broth (HIB) or on tryptose blood agar (TBA) plates (Difco) at 28°C. For routine growth and secretion assays, Y. pestis strains were grown in TMH medium in the presence or absence of 2.5 mM CaCl2 as described (Day et al., 2003). The levels of CaCl2 in TMH medium were varied from 0 to 7.5 mM. Escherichia coli DH5α was used for routine cloning experiments. Antibiotics were routinely used at the following concentrations: ampicillin 50 µg ml−1, streptomycin 50 µg ml−1, kanamycin 25 µg ml−1 and chloramphenicol 20 µg ml−1 (Table 2).
Construction of Y. pestis yscF, yopN, tyeA, yopB and yopD deletion strains
In-frame deletions in yscF and yopD that removed DNA sequences encoding residues 6–74 of YscF and 28–283 of YopD were constructed by the polymerase chain reaction (PCR)–ligation–PCR technique (Ali and Steinkasserer, 1995). Amplified DNA fragments containing a deletion in yscF or yopD flanked by approximately 700 bp of upstream and downstream DNA were inserted into SmaI-digested plasmid pRE112 (Edwards et al., 1998), generating suicide vectors pRE112-YscF1 and pRE112-YopD1 respectively.
Suicide vectors pRE112-YscF1, pUK4134.6 (Day and Plano, 1998), pUK4134.P2 (Day and Plano, 1998) and pRE112-ΔyopB (Day et al., 2003) were used to move in-frame deletions in yscF, yopN, tyeA and yopB into Y. pestis KIM5-3001.P39 by allelic exchange, generating strains KIM5-3001.P61 (ΔyscF), KIM5-3001.P62 (ΔyopN), KIM5-3001.P63 (ΔtyeA) and KIM5-3001.P64 (ΔyopB) respectively. Suicide vectors pRE112-YscF1 and pRE112-YopD1 were used to move in-frame deletions in yscF or yopD into Y. pestis KIM8-3002, generating strains KIM5-3002.P61 (ΔyscF) and KIM5-3002.P65 (ΔyopD).
Generation of point mutations in yscF of plasmid pYscF
Plasmid pYscF expresses the full-length YscF protein and was constructed by PCR amplification of yscF using oligonucleotide primers that anneal upstream of the yscF ribosome-binding site (YscF-UP1; 5′-TTTGAATTCAGCAGTATCAGG TTTGGCAGA-3′) and downstream of the yscF stop codon (YscF-DN4; 5′-TTTAAGCTTATGCAATTCGCTTCTTCGTG G-3′). The 446 bp amplification product was digested with EcoRI and HindIII (underlined) and inserted downstream of the Plac promoter in EcoRI- and HinDIII-digested plasmid pBCSK–.
Alanine- and cysteine-scanning mutagenesis of yscF of plasmid pYscF was accomplished using the PCR–ligation–PCR technique as previously described (Ferracci et al., 2004). The DNA sequence of the entire yscF gene present in each derivative of plasmid pYscF was confirmed by DNA sequence analysis.
Generation of point mutations in yscF of plasmid pCD1
Point mutations in the yscF gene of plasmid pCD1 of Y. pestis KIM5-3001.P39 and KIM8-3002 were constructed by site-directed mutagenesis and allelic exchange. Oligonucleotide primers YscF-pCD1-1 (5′-TGGGATACCATGGAATTGAGG TTT-3′) and YscF-pCD1-4 (5′-TAATCCAACCTGGCTCTCAT TGGG-3′) were used to amplify a 1.77-kb fragment of plasmid pCD1 containing yscF with approximately 750 bp of upstream and downstream flanking DNA. The amplified DNA fragment was inserted into the EcoRV site of plasmid pST-Blue1 (Novagen), generating plasmid pST-YscF. Site-directed mutagenesis of yscF of plasmid pST-YscF was performed using the QuikChange XL Site-Directed Mutagenesis Kit (Strategene) according to the manufacturer's instructions. Complementary oligonucleotides were designed to change either one or two yscF codons and to add or remove a restriction endonuclease site. Mutagenized plasmids were screened for the desired mutation by restriction endonuclease digestion and confirmed by DNA sequence analysis. To facilitate allelic exchange the mutated yscF insert was removed from pST-YscF by digestion with SacI and KpnI and inserted into SacI- and KpnI-digested suicide vector pRE112. The resultant pRE112-derivatives were used to move the yscF mutations into plasmid pCD1 of Y. pestis KIM5-3001.P39 and Y. pestis KIM8-3002 by allelic exchange. Y. pestis clones containing the individual yscF mutations were identified by PCR amplification and subsequent restriction endonuclease digestion.
Elk-tag translocation assay
HeLa cells were routinely grown at 37°C in 5% CO2 in RPMI-1640 medium (Sigma) supplemented with 10% (v/v) fetal calf serum (FCS) and 100 µg ml−1 penicillin and streptomycin. HeLa cells were seeded into six-well tissue culture plates containing 2.5 ml of RPMI per well at a density of 5 × 105 cells per well. Cells were allowed to adhere for 24 h. Before infection with Y. pestis, the cell monolayers were washed twice with RPMI without FCS and without antibiotics and incubated with 900 µl of the same medium for 30 min. Y. pestis strains were grown overnight at 27°C in HIB and diluted (400 µl in 1600 µl of HIB) the next day prior to reinitiating growth. After 3 h of growth at 27°C, bacterial cells were harvested, washed once with RPMI and resuspended in the same medium. HeLa cell monolayers were infected with Y. pestis at an moi of 30 for 3 h at 37°C in 5% CO2. After 3 h, the culture supernatants were removed, and the infected monolayers were lysed by the addition of 100 µl of 1X SDS-PAGE lysis buffer containing mammalian cell protease- (P-8340) and phosphatase-inhibitor (P-2850) cocktails (Sigma). Release of YopE129-Elk into the tissue culture medium was analysed by separating the non-adherent bacteria from the media supernatant by centrifugation at 14 000 g for 2 min. Proteins secreted into the media supernatants were precipitated by addition of TCA to 10% (v/v) final concentration. The TCA pellets from 500 µl of supernatant were each resuspended in 50 µl of 1X SDS-PAGE lysis buffer and both supernatant and infected HeLa cell monolayers were analysed by SDS-PAGE and immunoblot analysis with Elk-1 (♯9182) or phosphospecific Elk-1 (♯9181) antibody preparations (Cell Signaling Technology).
Cell-surface cross-linking of the YscF needle with BS 3
Cultures were grown at 37°C in 2 ml of TMH with or without 2.5 mM calcium chloride. After 3 h, 1 ml volumes were harvested by centrifugation at 8000 g for 5 min at 4°C. Bacterial pellets were gently resuspended in 1 ml of cold 20 mM HEPES, with or without 2.5 mM calcium chloride (pH 8). Bacterial surface proteins were cross-linked for 30 min at room temperature (RT) with the non-cleavable, membrane-impermeable, amine reactive cross-linker bis(sulfosuccinimidyl) suberate (BS3) (Pirece) at a final concentration of 5 mM. Unreacted BS3 molecules were quenched by the addition of 20 mM Tris-HCl, pH 8 for 15 min at RT. BS3-treated bacterial cells were collected by centrifugation at 12 200 g for 5 min at 4°C and treated as noted in SDS-PAGE and immunoblotting.
SDS-PAGE and immunoblotting
Whole-cell (pellets) and secreted proteins (supernatants) were prepared and analysed by SDS-PAGE followed by staining with Coomassie blue or immunoblotting, which was done essentially as described (Day et al., 2003). YopM and YopN were detected with polyclonal antibodies raised in rabbits against the whole molecules. YscF was detected with affinity purified antipeptide antibodies raised in rabbits against peptides corresponding to the N-terminal 20 residues (NH2-MSNFSGFTKGTDIADLDAVA-COOH) or C-terminal 20 residues (NH2-NSTIVRSMKDLMQGILQKFP-COOH) of YscF (ProteinTech Group).
HeLa cells were seeded in 6-well plates at a density of 1.5 × 105 cells per well in RPMI medium containing 10% FCS and 100 µg ml−1 penicillin and streptomycin. Cells were allowed to adhere for 24 h. Before infection with Y. pestis, the cells were washed twice with RPMI without FCS and without antibiotics, and incubated in 900 µl of the same medium for 1 h. Y. pestis strains were added to wells at an moi of 50 for 2 h at 37°C in 5% CO2.
Calcium overlay assay
The expression plasmid pMBP-TEV-YscF was constructed by the PCR–ligation–PCR technique (Ali and Steinkasserer, 1995). Oligonucleotide primers MBP-BglII (5′-CAAGCTGAT TGCTTACCCGATCGC-3′) and MBP-TEV (5′-ACCCTGGAA GTACAGGTTCTCCAC-3′) were used to amplify an 849 bp fragment of pFS1468 (Schubot and Waugh, 2004) encoding a C-terminal fragment of the MBP with a TEV protease recognition site. Oligonucleotide primers YscF-MBP1 (5′-ATGAGTAACTTCTCTGGATTTACG-3′) and YscF-MBP2 (5′-TTTAAGCTTTTATGGGAACTTCTGTAGGAT-3′) were used to amplify the entire 264 bp yscF open-reading frame. The resultant amplification products were combined, treated with T4 polynucleotide kinase and ligated using T4 DNA ligase. Oligonucleotide primers MBP-BglII and YscF-MBP2 were used to amplify the 1113 bp ligated fragment. The amplified product was digested with BglII and HindIII and inserted into BglII- and HinDIII-digested pMAL-C2X (New England Biolabs), generating plasmid pMBP-TEV-YscF. The pMBP-TEV-YscF expression vector and pMAL-C2X were electroporated into E. coli BL21. BL21 cultures (100 ml) were grown at 37°C in HIB to an OD600 = 0.4, induced with 1 mM IPTG and incubated for an additional 3 h at 37°C. Washed cells were lysed by passage through a French pressure cell at 4°C. Unlysed cells and large debris were removed by centrifugation at 10 000 g for 10 min at 4°C. The crude lysate (supernatant) was applied to an amylose column and purified according to the manufacturer's protocol (New England Biolabs).
45Ca overlay experiments were carried out essentially as described by Zorzato and Volpe (1988). Approximately 1 µg of partially purified MBP and MBP-YscF were separated by SDS-PAGE and electrotransferred to poly(vinylidene difluoride) membranes. The blot was extensively washed in overlay buffer (10 mM imidazole-HCl, pH 7.0, 70 mM KCl, 0.5 mM MgCl2) and incubated with overlay buffer containing 45Ca (0.4 mM CaCl2; 100 µCi ml−1) at RT for 1 h. The blot was washed three times with dH2O, air dried and exposed to film for 24 h.
Yersinia pestis strains were grown in TMH medium for 3 h at 37°C. Whole bacterial cells were stained for 30 s in a 4% solution of uranyl acetate on formvar-coated copper grids followed by two rinsing steps with dH2O. Grids were allowed to air dry and were viewed on Philips 300 TEM at 60 kV in the Edward Dower Electron Microscopy Laboratory on the University of Miami Coral Gables Campus.
We thank Ariel Blocker, Ken Fields and Franco Ferracci for critical reading of the manuscript and for useful comments. We also acknowledge Jeffrey R. Prince for helpful advice and for use of the Edward Dower Electron Microscopy Laboratory. This work was supported by Public Health Service Grants AI-39575 and AI-50552 from the National Institutes of Health.