Identification of Salmonella SPI-2 secretion system components required for SpvB-mediated cytotoxicity in macrophages and virulence in mice (original) (raw)

Abstract

The Salmonella SpvB protein possesses ADP-ribosyl transferase activity. SpvB, acting as an intracellular toxin, covalently modifies monomeric actin, leading to loss of F-actin filaments in _Salmonella_-infected human macrophages. Using defined Salmonella mutants, different functional components of the SPI-2 type three secretion system (TTSS), ssaV, spiC, sseB, sseC, and sseD, were found to be required for SpvB-mediated actin depolymerization in human macrophages. Expression of SpvB protein in Salmonella was not affected by any of the SPI-2 mutants and the effects of these loci were not due to reduced numbers of intracellular bacteria. Interestingly, the major SPI-2 virulence effector, SifA, is not required for SpvB action. Further, caspase-3 activation is an additional marker of cytotoxicity in _Salmonella_-infected human macrophages. Caspase-3 activity depended on SpvB and SPI-2 TTSS function, but not on SifA. These human macrophage cell culture results were corroborated by virulence studies in mice. Using competitive infection of mice with mixed inocula of single and double mutants, spvBmut1 mutation did not have an effect independent of ssaJ mutation, essential for SPI-2 TTSS function. In contrast, competitive infection studies in mice confirmed that SpvB and SifA have independent virulence effects, as predicted by the macrophage studies.

Introduction

Nontyphoid Salmonella enterica can spread from the gastrointestinal tract to cause serious systemic infections, including bacteremia, meningitis, septic arthritis, and osteomyelitis. Disseminated, nontyphoidal salmonellosis, common in both the developed and developing world, produces considerable morbidity and significant mortality. This disease is particularly difficult to treat in patients with defects in T-cell function and is increasingly associated with multi-drug resistance (Lee et al, 2002).

The spv virulence locus is associated with severe, disseminated nontyphoid Salmonella infection in humans and has a major virulence effect in animal models of systemic disease (Fierer et al, 1992). The locus encodes the SpvB protein, which has an NAD-binding sequence conferring ADP-ribosyl transferase activity; the primary substrate of this enzyme has been identified as monomeric actin. The transfected enzyme prevents the polymerization of G-actin monomers to F-actin filaments in CHO cells and the spv locus markedly enhances Salmonella virulence in mice (Lesnick et al, 2001). Cytotoxicity, manifested by F-actin depolymerization and detachment of cultured human macrophages, is induced by spvB 10–24 ;h after intracellular infection (Libby et al, 2000; Browne et al, 2002). In human macrophages, actin depolymerization induced by SpvB enzyme activity is dependent on NAD binding, demonstrated by infection with spvBmut1, a mutant with point mutations at the NAD-binding site (Browne et al, 2002; Lesnick et al, 2001). Substantial characterization of SpvB structure has taken place showing that SpvB covalently modifies actin at Arg177, inhibiting its ATPase activity (Margarit et al, 2006). Thus, depolymerization of actin in human monocyte-derived macrophages provides a phenotypic assay of SpvB action in the human host cell cytoplasm (Browne et al, 2002).

Purified SpvB toxin is inactive when administered extracellularly, indicating that the SpvB protein cannot pass through the host cell plasma membrane. Salmonella enter and survive in host cells within a specialized vacuolar structure termed the _Salmonella_-containing vacuole (SCV). From the SCV, Salmonella bacteria transfer virulence effector proteins into the host cell cytoplasm using the type three secretion system (TTSS) encoded by the SPI-2 locus. Previous work showed that the SPI-2 genes ssaV and ssaJ, encoding core components of the TTSS, are required for SpvB-induced actin depolymerization in human macrophages (Browne et al, 2002). The TTSS is a complex, multi-protein apparatus, and additional genes encoding distinct functional components of the SPI-2 TTSS required for SpvB effects on actin have not been defined previously. During the intracellular phase of Salmonella infection, SCVs are processed along a unique vesicular trafficking pathway, governed by a number of effector proteins secreted by the SPI-2 TTSS (Holden, 2002; Knodler & Steele-Mortimer, 2003). SsaV and SsaJ are postulated to be essential for the assembly of the core TTSS structure spanning the inner and outer bacterial membranes (Freeman et al, 2002; Kuhle & Hensel, 2004). SpiC is required for the secretion of the SseB, SseC, and SseD proteins, which are assembled on the bacterial surface and form the translocon structure that inserts into the membrane of the SCV (Nikolaus et al, 2001; Freeman et al, 2002; Yu et al, 2002; Ruiz-Albert et al, 2003). However, SpiC also appears to affect SCV trafficking through interactions with host factors involved in the sorting pathway (Shotland et al, 2003); the relative contribution of this SpiC effector function to virulence remains unclear. The SifA effector protein has a substantial effect on virulence and alters SCV membrane dynamics through down-regulation of kinesin recruitment (Boucrot et al, 2005). Biochemical activities for several other SPI-2 TTSS effectors have been identified, but mutations in these loci have either modest or negligible effects on virulence in the mouse model (Knodler & Steele-Mortimer, 2003; Kuhle & Hensel, 2004).

When bacteria enter macrophages by opsonin-mediated phagocytosis, mouse-virulent Salmonella strains cause late cytopathology with characteristic findings of apoptosis (Lindgren et al, 1996; Libby et al, 2000; van der Velden et al, 2000; Santos et al, 2001; Kurita et al, 2003). In human intestinal epithelial cells infected with Salmonella, caspase-3 activation and DNA fragmentation are found as late markers of cytopathology, and this process was shown to require the SPI-2 TTSS and the virulence plasmid carrying the spv locus (Paesold et al, 2002). However, the role of SpvB-mediated actin depolymerization in caspase-3 activation has not been investigated.

In this paper, the different functional components of the SPI-2 TTSS required for SpvB to produce actin depolymerization during Salmonella infection of human macrophages were examined. The relationship between spvB and SPI-2 genes in the induction of cytotoxicity in human macrophages was also examined by investigating caspase-3 activation in vitro. The findings in macrophages predicted functional virulence relationships between the SPI-2 TTSS, SifA, and SpvB, and these hypotheses were confirmed by competitive infections in mice using mixtures of mutant strains.

Materials and methods

Bacterial strains

Single-colony isolates of the wild-type S. enterica serovar Dublin Lane (Libby et al, 2000) and S. enterica serovar Typhimurium 14028s or mutants derived from these strains were used. The mutations are as follows: spvBmut1 (Lesnick et al, 2001), _sifA_∷Tn_10d_Cm (Stein et al, 1996), _sifA_∷mTn5 (Beuzon et al, 2000), _spiC_∷Km (Uchiya et al, 1999), _ssaV_∷mTn5, ssaJ_∷mTn5 (Hensel et al, 1997), sseC∷_aphT, sseD_∷_aphT, sseB_∷_aphT (Nikolaus et al, 2001). The replacement of ssaJ with a cat gene cassette (ssaJ_∷_cat) was constructed using a modification of the method described by Datsenko & Wanner (2000). In double mutants, all mutant loci were transduced by P22 lysates (made using strain TH2423) into isogenic single-colony isolates of Typhimurium 14028s and Dublin Lane (Lesnick et al, 2001).

Human macrophage culture and measurement of cell detachment and intracellular F-actin

Blood was collected from normal donors, and the monocytes were purified by density gradient centrifugation and differential adherence (Libby et al, 2000). Monocytes were seeded into 24-well plates in RPMI with 15% autologous serum for 5–7 days to allow differentiation into macrophages. Single colonies of wild type or mutants, derived as described above, were grown to stationary phase in M9 minimal media (Fang et al, 1991). Macrophages were inoculated at a baseline multiplicity of infection (MOI) of 1–3 ;bacteria ;cell−1 or at five times (5 ×) the baseline MOI. Culture plates were incubated for 1 ;h at 37 ;°C, cells were washed, and fresh medium containing gentamicin (20µg ;mL−1) was added. After 18–24 ;h, wells were scraped. The contents were aspirated, fixed with 3.7% formalin in phosphate-buffered saline (PBS), and cyto-spun onto microscope slides. Intracellular actin was detected using phalloidin–rhodamine stain and counterstained for nucleic acid with 4′-diamidino-2-phenylindole (DAPI), at 1 ;µg ;mL−1 in PBS for 15 ;min. Slides were visualized by fluorescent microscopy, 300–600 cells for each infecting Salmonella strain were identified, and the proportions costaining for any detectable F-actin were determined. Only macrophages without any detectable actin staining were counted as negative (Browne et al, 2002).

Analysis of SpvB protein expression in Salmonella

Strains used [wild-type _S. enterica_ serovar Dublin Lane (Libby et al, 2000) and mutant strains, spvBmut1 (Lesnick et al, 2001), spiC_∷Km (Uchiya et al, 1999), 12023 ssaJ∷Tn5 from Gunter Kan clone 1, sseC∷_aphT, sseD_∷_aphT, sseB_∷_aphT (Nikolaus et al, 2001)] were grown overnight to stationary phase in Luria–Bertani (LB) media without antibiotics. Overnight samples (1 ;mL) were pelleted, resuspended, and routinely analyzed by immunoblotting. Bacterial proteins were initially reacted with anti-SpvB rabbit polyclonal antibody. The secondary antibody used was horseradish peroxidase (HRP) goat α rabbit (KPL Gaithersburg, MD). Detection by fluorescence was per Pierce Super Signal (Pierce, Rockford IL), according to the manufacturer's instructions.

Measurement of caspase-3 activity

Human macrophages were prepared, cultured, infected, aspirated, and washed following the protocol described above, with single colonies of the wild-type S. enterica serovar Dublin Lane (Libby et al, 2000) or mutants derived from these strains, spvBmut1 (Lesnick et al., 2001), _sifA_mTn5 (Beuzon et al, 2000), and _ssaV_∷mTn5 (Hensel et al, 1997). After 24 ;h of infection, each well was carefully scraped using a Costar cell scraper 3010 (Corning, NY). Caspase-3 activity was determined using a commercial assay kit from Sigma-Aldrich according to the manufacturer's instructions.

Mouse infections and determination of the competitive index (CI)

BALB/c mice were infected by the intraperitoneal route as described previously (Lesnick et al, 2001). Single- and double-mutant strains were grown separately overnight, diluted, and mixed in an equal ratio, as determined by the optical densities of the cultures, to produce an inoculum of 3–5 × 104 ;bacteria. The actual inoculum was diluted and plated. At least 100 colonies were picked onto plates with kanamycin to determine the ratio of strains in the inoculum. Mice were sacrificed three days after infection; the spleens were homogenized and plated on LB agar containing chloramphenicol, and then picked onto kanamycin plates. CI was determined as the fraction of double mutants recovered from the spleen divided by the fraction of double mutants in the inoculum.

Results

SpiC, SseB, SseC, SseD as well as SsaV are essential for SpvB-induced actin depolymerization

Prior studies in human macrophages indicated that actin depolymerization induced by the Salmonella SpvB protein required a SPI-2 TTSS core structure with intact ssaV/ssaJ genes (Browne et al, 2002). SsaV and SsaJ are essential for the assembly of the core TTSS structure spanning the inner and outer bacterial membranes (Freeman et al, 2002; Kuhle & Hensel, 2004). In order to investigate the role of other SPI-2 TTSS functional components in the SpvB-mediated effect, the loss of filamentous actin was assayed in human macrophages 22 ;h after inoculation with wild-type S. enterica serovar Dublin Lane, spiC, sseB, sseC, sseD, or ssaV mutant strains.

SpiC is required for the secretion of translocon components through the core TTSS. Figure 1 shows that only 0.7% of macrophages infected with spiC mutant Salmonella, at a standard MOI of 1–3, demonstrated F-actin loss, a figure similar to that of _ssaV_-infected macrophages (1.7% loss). However, 47% of human macrophages infected with wild-type S. enterica serovar Dublin Lane lost their filamentous actin, at the same MOI (see Fig. 1). The standard inoculum, MOI of 1–3, was compared to an inoculum five times standard (5 × MOI). No increase in actin loss was observed when macrophages were infected with the spiC or ssaV mutants at 5 × MOI. In contrast, wild-type Salmonella infection at 5 × MOI showed an increase in actin loss to 69%. Experiments with these mutants were repeated more than three times with similar results. Analysis of viable bacterial numbers from each infection condition showed no significant difference between mutants and wild-type, verifying that the results seen were not simply due to lower numbers of intracellular bacteria associated with reduced replication fitness of the spiC or ssaV mutant (see Table 1). Human macrophages are bacteriostatic for Salmonella strains under the assay conditions, and survival/growth defects of Salmonella mutants are not seen (Libby et al, 2000; Browne et al, 2002). At these high MOI levels, the numbers of mutant bacteria were much higher than the numbers of wild-type bacteria previously shown to produce dramatic actin depolymerization (Browne et al, 2002). Thus, the results demonstrate that the spiC gene is essential for SpvB-induced F-actin depolymerization, in addition to the ssaV gene. Further, this requirement is absolute with no evidence of a dose response effect at higher MOIs.

Figure 1

Percentage of the total amount of human monocyte-derived macrophages in the culture well with complete actin depolymerization, as determined via F-actin staining with phalloidin–rhodamine 20 ;h after infection. Values represent the means of results for three or four wells. Macrophages infected with wild-type Salmonella enterica serovar Dublin are compared with those infected with ssaV and spiC mutant strains at either the standard inoculum MOI of 1–3, or five times the standard inocula (5 × MOI). Experiments testing these mutants were repeated three or more times with similar results.

Table 1

Viable bacteria recovered from the total population of human monocyte derived macrophages after infection with wt Salmonella enterica serovar Dublin, spiC, or ssaV mutants at 5 × MOI

Table 1

Viable bacteria recovered from the total population of human monocyte derived macrophages after infection with wt Salmonella enterica serovar Dublin, spiC, or ssaV mutants at 5 × MOI

Next, to further investigate the components of the SPI-2 TTSS needed for SpvB to induce host cell actin depolymerization, human monocyte-derived macrophages were infected with wild-type S. enterica serovar Dublin Lane, sseB, sseC, sseD, or spiC mutants. SseB, SseC, and SseD form the putative translocon on the bacterial surface that interacts with the phagosome membrane and transfers Salmonella effector proteins into the host cell. In order to eliminate the possibility that these mutants would lead to decreased numbers of intracellular bacteria, these experiments were performed using 5 × the standard inocula of 1–3, termed 5 × MOI. Following 20 ;h of infection with Salmonella enteriditis serovar Dublin mutant sseC or sseB, 3–6% of human macrophages in vitro had lost F-actin staining, comparable to infection with the spiC mutant where 5% of cells showed F-actin loss (Fig. 2). The sseD mutant infection also gave negligible F-actin loss (data not shown). In contrast, after infection with the parent wild-type S. enterica serovar Dublin Lane, c. 60% of human macrophages show F-actin depolymerization. Intracellular bacterial CFU showed no significant differences between mutant, or wild-type S. enterica serovar Dublin Lane infection (data not shown). These experiments were repeated with similar results. These studies indicate that the translocon components SseB, SseC, or SseD are required in addition to SpiC for SpvB to produce actin loss in the cytoplasm. Use of a high MOI with the SPI-2 mutants should unmask the presence of any secondary gene products that could compensate for the absence of these genes. As none were found, it was concluded that these genes are an absolute requirement for SpvB-induced actin loss in the host macrophage.

Figure 2

Percentage of the total amount of human monocyte-derived macrophages in the culture well with complete actin depolymerization, as determined via F-actin staining with phalloidin–rhodamine 20 ;h after infection. Values represent the means of results for three or four wells. Macrophages infected with wild-type Salmonella enterica serovar Dublin at five times standard MOI (MOI × 5) are compared with those infected with sseB, sseC, or spiC mutant strains at five times standard MOI (5 × MOI). Experiments testing these mutants were repeated three or more times with similar results.

Bacterial SpvB expression is not affected by any of the SPI-2 mutations

We then looked to see whether these genes, which appeared essential to SpvB action, had any effect on intrabacterial SpvB expression, or whether their involvement was more likely attributable to their function in the SPI-2 TTSS. SPI-2 genes have not thus far been implicated in the regulation of SpvB expression in Salmonella; nevertheless, it was possible that SpvB expression was deficient in the SPI-2 mutant strains used. To evaluate SpvB production levels across the bacterial strains employed, the levels of SpvB protein in the wild-type bacteria and in each of the SPI-2 mutants were measured directly, using an SpvB-specific antibody combined with immunoblotting. Figure 3 shows that SpvB is present in wild-type Salmonella serovar Dublin and in all mutants used in this study. Moreover, the levels of SpvB expression observed in wt Samonella serovar Dublin are comparable to those obtained from ssaV, spiC, sseB, sseD, sseC mutant Samonella serovar Dublin. These data provide direct confirmation that these genes are not involved in SpvB production and that their involvement is more likely attributable to their function in the SPI-2 TTSS.

Figure 3

SpvB protein identified via Western blotting using SpvB-specific antibody. The signal corresponds to a known 65.6 ;kDa size of SpvB protein. The signal observed from bacterial cultures is as follows: lane 1, wt Salmonella enterica serovar Dublin; lane 2, ssaV_∷_Tn5 mutant; lane 3, ssaJ_∷_Tn5 mutant; lane 4, sseB_∷_aphT mutant; lane 5, sseC_∷_aphT mutant; lane 6, sseD_∷_aphT mutant; lane 7, spiC mutant Salmonella enterica serovar Dublin.

The sifA gene is not required for F-actin loss in infected cells

The only SPI-2 effector that approaches the virulence effect of SpvB is SifA. SifA alters SCV positioning and membrane dynamics and is the factor necessary for the formation of tubular membranous structures, termed Sifs, for _Salmonella_-induced filaments. Therefore, whether SpvB-mediated actin depolymerization required the intracellular action of SifA was determined. The effects of infection of human macrophages with sifA knockout mutants were compared with infection with spvBmut1 or wild-type S. enterica serovar Dublin Lane at MOI values between 3 and 6. After 22 ;h of infection, c. 70% of macrophages infected with wild-type S. enterica serovar Dublin Lane have lost their F-actin (Fig. 4). Similarly, 70–80% of cells infected with two independent sifA mutants show actin depolymerization. In contrast, only 8% of _spvBmut1_-infected macrophages showed filamentous actin loss. Analysis of bacterial CFUs in each infection condition showed no significant differences (data not shown); verifying that the results seen were not simply due to the reduced replication fitness of the sifA or spvBmut1 mutant strains. These experiments were repeated with similar results (data not shown). This analysis of SifA shows that this effector does not appear to have a role in SpvB-induced actin depolymerization.

Figure 4

Percentage of the total amount of human monocyte-derived macrophages with complete actin depolymerization, as determined via F-actin staining with phalloidin–rhodamine 22 ;h after infection. Values represent the mean of the results for triplicate wells. Salmonella enterica serovar Dublin wild-type infected cells are compared with those infected with spvB or sifA mutant strains. Experiments with these mutants were tested three or more times with similar results.

Wild-type Salmonella serovar Dublin infection of human macrophages activates caspase-3 through a mechanism dependent on SpvB and the SPI-2 TTSS but independent of SifA

_Salmonella_-induced cytotoxicity of human macrophages mediated by the spv locus produces DNA fragmentation and nuclear changes consistent with apoptosis. It was predicted that caspase-3, a principal executioner of apoptosis, would be activated in human macrophages by the cytotoxic mechanism dependent on SpvB-induced actin depolymerization, but independent of SifA-mediated SCV membrane dynamics. Primary human macrophages were inoculated with the standard MOI of 1–3, and caspase-3 activity was assayed 20 ;h postinfection, at a time when c. 50% of the macrophages infected with wild-type bacteria show complete loss of F-actin. Caspase activation in human monocyte-derived macrophages infected with wild-type serovar Dublin was >10-fold higher than in uninfected cells (Fig. 5). Caspase-3 activity was markedly reduced in cells infected with the spvBmut1 strain (P<0.00005 compared with wild-type), and was even less when the ssaV mutant was used (P<0.00005). In contrast, the sifA mutation had no effect on caspase-3 activation. Bacterial CFU in each infection condition showed no significant differences (data not shown), thus the results obtained cannot be due to differences in proliferation between mutant and wild-type bacteria. These results suggest that the SPI-2 TTSS and SpvB are important cofactors in the induction of the classical pathway of apoptosis in late cytotoxcity produced by Salmonella infection in human macrophages, while SifA-mediated effects on the SCV are not required for the apoptosis process.

Figure 5

Caspase-3 activation expressed as a percentage of positive control, where positive control is equivalent to caspase-3 levels detected in 1 million apoptotic jurkat cells. Caspase-3 activation detected by colorimetric assay in uninfected monocyte-derived human macrophages is compared with that obtained from human macrophages after 20 ;h of infection with wild-type Salmonella enterica serovar Dublin or ssaV, spvB, or sifA mutant strains. The figure is composed of means for each mutant from three or more samples. The means obtained from three or more experiments were then combined in this figure. The SE bar represents the variance seen across experiments.

The virulence effect of spvB in mice is dependent on the SPI-2 TTSS but independent of sifA

Previous work has shown that the ADP-ribosylation of actin is essential for the virulence phenotype of SpvB. Thus, given the results described above, the virulence effect of SpvB in mice should also require the SPI2 TTSS function, but not the sifA gene. To investigate this prediction, competitive infection of mice with mixed inocula of single and double mutants was used. The hypothesis was that the spvB mutation would not further decrease the virulence of an SPI-2 knockout in the core ssaJ gene, because the SpvB protein should not have an effect independent of SPI-2 function. In contrast, the spvB mutation is predicted to decrease the virulence of a sifA mutant, because both effectors function independently in the macrophage assay.

The mice were infected with an equal ratio of the ssaJ_∷_cat single mutant and the ssaJ_∷_cat/spvBmut1 double mutant, and then determined the ratio of the strains in the spleens after 3 days was determined. The competitive index was calculated as the fraction of double mutants recovered from the spleens divided by the fraction of double mutants in the actual inoculum. A CI significantly <1 indicates that the double mutant was less virulent than the single mutant. As seen in Table 2, the CI for the experiment comparing the _ssaJ_ single mutant with the _ssaJ/spvB_ double mutant was 0.996, indicating that the introduction of the _spvB_ mutation into the _ssaJ_ mutant strain had no detectable effect on virulence (_P_>0.5), as predicted from the macrophage assays measuring actin depolymerization. These findings were not affected by the particular serovar used, because similar results were obtained for the corresponding single- and double-mutant strains of serovar Dublin (mean CI=0.89). In contrast, the CI for the experiment comparing the sifA single mutant with the sifA/spvB double mutant was 0.088. This result shows that recovery of the double mutant after 3 days of infection was about 10-fold reduced relative to its representation in the inoculum (P<0.00001), indicating that the sifA and spvB genes have independent contributions to virulence, as predicted from the macrophage infection assays. Similar independent virulence effects of spvB and sifA were found in competitive infections using serovar Dublin (data not shown).

Table 2

Competitive index (CI) values for mixed infections of BALB/c mice with Salmonella enterica serovar Typhimurium strains using equal numbers of single and double mutants in the inocula

Table 2

Competitive index (CI) values for mixed infections of BALB/c mice with Salmonella enterica serovar Typhimurium strains using equal numbers of single and double mutants in the inocula

Discussion

This work demonstrates that the SPI-2 genes ssaV, spiC, sseB, sseC, and sseD are essential for SpvB-induced actin depolymerization in the host cell cytoplasm. Following entry into host cells, Salmonella normally resides inside the SCV, and induction of SPI-2 expression is mediated by the environmental conditions in the vacuole. SsaV and SsaJ are essential for the assembly of the core TTSS structure spanning the inner and outer bacterial membranes. SpiC is required for the secretion of the SseB, SseC, and SseD proteins, which are assembled on the bacterial surface and form the translocon structure that inserts into the membrane of the SCV (Nikolaus et al, 2001; Freeman et al, 2002; Yu et al, 2002; Ruiz-Albert et al, 2003). As each component of the translocon, as well as SpiC, is essential, the translocon structure apparently needs to be complete for SpvB-induced actin depolymerization of host cells to take place. Because SpvB-mediated effects on F-actin require only the C-terminal enzyme domain without any bacterial cofactors, the assay for depolymerization is a sensitive test for the transport of SpvB into the host cell (Lesnick et al, 2001). Our results suggest that SpvB is secreted and translocated into the cell cytoplasm by the SPI-2 encoded TTSS and translocon structures. These results cannot be explained by replication defects in the mutant-infected strains as intramacrophage CFU were similar, with no significant or substantial differences between wild-type and the mutants employed. Further, use of five times the standard MOI should unmask the presence of any secondary gene products encoding alternative secretion mechanisms that could compensate for the absence of these genes. These results also verified that expression levels of SpvB protein were identical in wild-type Salmonella serovar Dublin and the bacterial mutants ssaV, ssaJ, spiC, sseB, sseC, and sseD. This provides direct observation that SpvB is produced normally in these mutants.

SifA is an SPI-2 secreted effector shown to have an effect on virulence approaching that of SpvB (Stein et al, 1996; Beuzon et al, 2000). However, our studies in human monocyte-derived macrophages found that SpvB-induced actin depolymerization was independent of sifA. It has been reported in sifA mutants that the SCV membrane is unstable and bacteria are released free into the cytoplasm, but our results indicate that SifA-dependent membrane dynamics do not affect the ability of SpvB to be secreted and released into the host cell.

The findings that we report in human macrophages for genes involved in SpvB-induced effects were corroborated in the mouse model of systemic Salmonella infection. Using competitive infection of mice with mixed inocula of single and double mutants, it was demonstrated that spvB did not have a virulence effect independent of SPI-2 function. These findings clarify and extend the results of Shea et al. (1999), who reported that the presence of an spvA mutation appeared to further attenuate the virulence of an ssaV mutant. The authors also employed a competitive index method in mixed infections of mice. The present studies used the spvBmut1 point mutation, which exclusively alters the ADP-ribosylation enzyme activity specific to the SpvB protein and does not have polar effects on expression of the downstream genes. In contrast, the spvA mutation is likely to be polar on the downstream spvBCD genes (Krause et al, 1991). Although spvA does not appear to have a virulence effect in mice (Roudier et al, 1992), transcription of not only spvB but also spvC and spvD proteins would be affected. The SpvC and SpvD proteins could have an accessory virulence role in vivo, as suggested in an earlier mutational analysis (Roudier et al, 1992). However, our studies did show that the spvB mutation caused a highly significant decrease in virulence when combined with the sifA mutation. This finding indicates that spvB and sifA gene functions are entirely independent of each other, also predicted from the cell culture studies with human macrophages.

The ADP-ribosylating capacity of SpvB has been shown as required to induce apoptosis in eukaryotic cells (Kurita et al, 2003). In human epithelial cell lines, delayed cell death and caspase-3 activation has been shown to be dependent on the spv locus together with SPI-2 (Libby et al, 2000; Paesold et al, 2002). Our results suggest that the spvB active site mutation, spvBmut1, and SPI-2 are required for caspase-3 activation in _Salmonella_-infected human macrophages, whereas the SPI-2 effector SifA was not. Because caspase-3 activity showed dependence on the ADP-ribosylating activity of SpvB, SpvB may be one of the effectors of cell death in _Salmonella_-infected macrophages. The role of host cell death mechanisms in the pathogenesis of infection remains unclear. However, the major role of the SpvB protein in virulence suggests that both actin depolymerization and the induction of host cell apoptosis through caspase-3 are important mechanisms of Salmonella pathogenesis.

Acknowledgements

The authors wish to thank Dr Eulalia Valle for her excellent assistance with caspase-3 assays and Richard Castle for assistance with the sifA studies. In addition, the authors appreciate the generosity of Murry Stein, Sam Miller, Eduardo Groisman, David Holden, and Michael Hensel in supplying key mutants. This work was supported by NIH grants AI062758 to S.H.B., AI32178 and DK35108 to D.G.G., and AI47884 to J.F.

References

(

2000

)

Salmonella maintains the integrity of its intracellular vacuole through the action of SifA

.

EMBO J

19

:

3235

–

3249

.

(

2005

)

The intracellular fate of Salmonella depends on the recruitment of kinesin

.

Science

308

:

1174

–

1178

.

(

2002

)

Genetic requirements for _Salmonella_-induced cytopathology in human monocyte-derived macrophages

.

Infect Immun

70

:

7126

–

7135

.

(

2000

)

One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products

.

Proc Natl Acad Sci USA

97

:

6640

–

6645

.

(

1991

)

Growth regulation of a Salmonella plasmid gene essential for virulence

.

J Bacteriol

173

:

6783

–

6789

.

(

1992

)

Salmonella typhimurium bacteremia: association with the virulence plasmid

.

J Infect Dis

166

:

639

–

642

.

(

2002

)

SpiC is required for translocation of Salmonella pathogenicity island 2 effectors and secretion of translocon proteins SseB and SseC

.

J Bacteriol

184

:

4971

–

4980

.

(

1997

)

Functional analysis of ssaJ and the ssaK/U operon, 13 genes encoding components of the type III secretion apparatus of Salmonella Pathogenicity Island 2

.

Mol Microbiol

24

:

155

–

167

.

(

2002

)

Trafficking of the Salmonella vacuole in macrophages

.

Traffic

3

:

161

–

169

.

(

2003

)

Taking possession: biogenesis of the Salmonella-containing vacuole

.

Traffic

4

:

587

–

599

.

(

1991

)

Molecular analysis of the virulence locus of the Salmonella dublin plasmid pSDL2

.

Mol Microbiol

5

:

307

–

316

.

(

2004

)

Cellular microbiology of intracellular Salmonella enterica: functions of the type III secretion system encoded by Salmonella pathogenicity island 2

.

Cell Mol Life Sci

61

:

2812

–

2826

.

(

2003

)

Intracellular expression of the Salmonella plasmid virulence protein, SpvB, causes apoptotic cell death in eukaryotic cells

.

Microb Pathog

35

:

43

–

48

.

(

2002

)

Multidrug-resistant non-typhoid Salmonella infections in a medical center

.

J Microbiol Immunol Infect

35

:

78

–

84

.

(

2001

)

The Salmonella spvB virulence gene encodes an enzyme that ADP-ribosylates actin and destabilizes the cytoskeleton of eukaryotic cells

.

Mol Microbiol

39

:

1464

–

1470

.

(

2000

)

The Salmonella virulence plasmid spv genes are required for cytopathology in human monocyte-derived macrophages

.

Cell Microbiol

2

:

49

–

58

.

(

1996

)

Macrophage killing is an essential virulence mechanism of Salmonella typhimurium

.

Proc Natl Acad Sci USA

93

:

4197

–

4201

.

(

2006

)

A steric antagonism of actin polymerization by a Salmonella virulence protein

.

Structure

14

:

1219

–

1229

.

(

2001

)

SseBCD proteins are secreted by the type III secretion system of Salmonella pathogenicity island 2 and function as a translocon

.

J Bacteriol

183

:

6036

–

6045

.

(

2002

)

Genes in the Salmonella pathogenicity island 2 and the Salmonella virulence plasmid are essential for _Salmonella_-induced apoptosis in intestinal epithelial cells

.

Cell Microbiol

4

:

771

–

781

.

(

1992

)

Characterization of translation termination mutations in the spv operon of the Salmonella virulence plasmid pSDL2

.

J Bacteriol

174

:

6418

–

6423

.

(

2003

)

SseA is a chaperone for the SseB and SseD translocon components of the Salmonella pathogenicity-island-2-encoded type III secretion system

.

Microbiology

149

:

1103

–

1111

.

(

2001

)

Salmonella enterica serovar typhimurium induces cell death in bovine monocyte-derived macrophages by early sipB-dependent and delayed sipB-independent mechanisms

.

Infect Immun

69

:

2293

–

2301

.

(

1999

)

Influence of the Salmonella typhimurium pathogenicity island 2 type III secretion system on bacterial growth in the mouse

.

Infect Immun

67

:

213

–

219

.

(

2003

)

The Salmonella SpiC protein targets the mammalian Hook3 protein function to alter cellular trafficking

.

Mol Microbiol

49

:

1565

–

1576

.

(

1996

)

Identification of a Salmonella virulence gene required for formation of filamentous structures containing lysosomal membrane glycoproteins within epithelial cells

.

Mol Microbiol

20

:

151

–

164

.

(

1999

)

A Salmonella virulence protein that inhibits cellular trafficking

.

EMBO J

18

:

3924

–

3933

.

(

2000

)

Salmonella pathogenicity island 1-independent induction of apoptosis in infected macrophages by Salmonella enterica serotype typhimurium

.

Infect Immun

68

:

5702

–

5709

.

(

2002

)

SpiC is required for secretion of Salmonella Pathogenicity Island 2 type III secretion system proteins

.

Cell Microbiol

4

:

531

–

540

.

Author notes

Editor: Hans Kusters

© 2007 Federation of European Microbiological Societies