SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The ability of Salmonella enterica to invade and replicate within host cells depends on two type III secretion systems (TTSSs) encoded on pathogenicity islands 1 and 2 (SPI1 and SPI2). The current paradigm holds that these systems translocate two classes of effectors that operate sequentially and independently. In essence, the SPI1 TTSS mediates early events (i.e. invasion) whereas the SPI2 TTSS mediates post-invasion processes (i.e. replication, vacuole maturation). Contrary to this model, we have found in infected macrophages that a SPI1 effector, SopB/SigD, increased inducible nitric oxide synthase levels and nitric oxide production, host cell process previously known only to be a target of the SPI2 TTSS. Furthermore, SopB protein and message persist many hours after invasion. Our findings reveal an unanticipated potential for dialogue between the SPI1 and SPI2 TTSS and the host cell response.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Many Gram-negative pathogens use type III secretion systems (TTSSs) to translocate bacterial effector proteins into host cells where they modulate an impressive variety of host processes. Salmonella enterica serovars encode two  distinct  TTSSs  and  can  thus  translocate  different sets of effectors in a physically and temporally separate manner. The Salmonellapathogenicity island 1 (SPI1)-encoded TTSS, which is essential for invasion of epithelial cells, is induced extracellularly (Galan and Curtiss, 1989). Conversely, the SPI2 TTSS, which mediates biogenesis of the Salmonella-containing vacuole (SCV) and intracellular survival, is only induced intracellularly (Ochman et al., 1996; Cirillo et al., 1998; Hensel et al., 1998). According to the current model, SPI1 and SPI2 effectors modulate distinct processes with SPI1 effectors acting before SPI2 effectors. This model is dependent on the differential regulation of SPI1 and SPI2 expression and the degradation/inactivation of translocated SPI1 effectors (Galan, 2001; Knodler et al., 2002; Kubori and Galan, 2003). However, there are hints that SPI1 effectors may influence and complement post-invasion events previously attributed solely to the actions of SPI2 effectors (Garner et al., 2002; Steele-Mortimer et al., 2002).

Here we have specifically addressed the hypothesis that a SPI1 effector can play a significant role in post-invasion events. The inositol phosphatase SopB (SigD) was used as a model SPI1 effector because, unlike many SPI1 mutants, sopB deletion mutants are fully invasive. SopB activity is easy to follow by monitoring rapid activation of the serine/threonine kinase, Akt, an event that is solely SopB dependent in epithelial cells. Additionally a catalytically inactive SopB mutant (C462S) exists which does not activate Akt (Steele-Mortimer et al., 2000). Finally, SopB is required for intestinal fluid secretion in the calf model of disease (Galyov et al., 1997). Such marked phenotypes are important for defining specific functions for individual effectors, because considerable redundancy and overlapping functions exist for Salmonella effectors (Galan, 2001).

Because Akt is a known regulator of inducible nitric oxide synthase (iNOS) (Ozes et al., 1999; Park et al., 2003), we hypothesized that SopB may affect nitric oxide (NO) response after Salmonella infection. While iNOS appears to play a role in controlling Salmonella proliferation (Mastroeni et al., 2000; Vazquez-Torres et al., 2000), increased iNOS expression and NO release from macrophages were not observed until ≈ 8 h post infection (p.i.) (Cherayil and Antos, 2001). Although SPI2 effectors do not regulate the level of NO released, they have been shown to protect Salmonella from reactive nitrogen species. In addition, as yet unidentified SPI2 effectors perturb intracellular iNOS localization (Chakravortty et al., 2002). Here we show that SopB regulates iNOS activity for extended periods post invasion. Consistent with this observation, both sopB mRNA and protein levels persist in infected macrophage-like cells for many hours p.i. Thus, in addition to the historically defined role of SPI1 effectors in early host–cell interactions, these same effectors can influence host cell responses long after invasion. This significantly expands the host cell activities that can be potentially modulated by a single bacterial effector.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Phosphorylation and activation of the eukaryotic pro-survival kinase Akt is absolutely dependent on the SPI1 effector SopB in epithelial cells (Steele-Mortimer et al., 2000). Here we examined Akt activation in a murine macrophage-like cell line, RAW264.7, because macrophages serve as an important locale for intracellular Salmonella during systemic infections (Richter-Dahlfors et al., 1997; Fields et al., 1986). Serum-starved (DMEM + 0.5% FCS) RAW264.7 cells were infected with SPI1-induced serovar Typhimurium (S. Typhimurium) at a low multiplicity of infection (moi) (moi = 1). At the indicated times, whole-cell extracts were analysed by immunoblotting with polyclonal antibodies that recognize either phosphorylated Akt (Ser 473) or total Akt. Wild-type (WT) Salmonella induced rapid (within 10 min) and sustained phosphorylation of Akt in macrophages (Fig. 1A) similar to that seen in epithelial cells (Steele-Mortimer et al., 2000). In contrast, an isogenic sopB deletion mutant (ΔsopB) induced very little Akt phosphorylation up to 1 h p.i. However, at later time points (8 and 15 h p.i.), Akt phosphorylation was comparable in wild-type (WT) and ΔsopB-infected cells. A similar pattern of phosphorylation at Thr 308 of Akt was observed (data not shown). Thus SopB, while not essential, is a major stimulus of rapid Akt phosphorylation in Salmonella-infected macrophage-like cells. To visualize Salmonella-dependent Akt activation in macrophages, we examined green fluorescent protein (GFP)-Akt transiently transfected RAW264.7 cells by confocal microscopy. In infected epithelial cells membrane localization of Akt, specifically to Salmonella-induced ruffles, is SopB independent (Steele-Mortimer et al., 2000). Similarly, we found here that transiently expressed GFP-Akt in RAW264.7 cells was concentrated in membrane ruffles induced by either WT or ΔsopB strains 20 min p.i. (Fig. 1B, arrows). In contrast, indirect immunofluorescence using phospho-specific antibodies (Ser 473) showed that phosphorylated Akt was only detected in ruffles induced by WT Salmonella (Fig. 1B, arrowheads), like what has been reported for infected epithelial cells (Steele-Mortimer et al., 2000).

image

Figure 1. Rapid Akt phosphorylation in Salmonella-infected RAW264.7 cells is SopB dependent. A. Cells were serum starved (0.5% FCS) for 3 h before infection with SPI1-induced WT Salmonella or an isogenic sopB deletion mutant (ΔsopB) (moi = 1). Total cell extracts were collected at the times indicated p.i. Proteins were analysed by 10% SDS-PAGE and immunoblotting using polyclonal antibodies that recognize only phosphorylated Akt (Ser 473) or total Akt. B. Cells were transiently transfected with GFP-Akt (green) 24 h before infection, fixed 20 min p.i. and processed for confocal immunofluorescence microscopy using monoclonal anti-LPS antibodies to detect Salmonella (red) and polyclonal antibodies that recognize only phosphorylated Akt (Ser 473) (blue). Ruffles containing GFP-Akt (arrows) or phosphorylated Akt (arrowheads) are indicated. Images are the projection of a Z-series of 20 confocal images and insets are a single plane from the Z-series.

Download figure to PowerPoint

Because it has been shown that infecting RAW264.7 cells with SPI1-induced Salmonella at a high moi (moi = 100) induces apoptosis, we wished to discount cytotoxic effects in our experiments using low moi (0.5 and 5). To address this question, we infected RAW264.7 cells grown on glass coverslips with WT or ΔsopB strains at moi of 0.5, 5 and 100 for 2 or 15 h before incubating with a LIVE/DEAD stain that fluorescently labels the nuclei of dead cells with ethidium homodimer-1. Coverslips were then examined by fluorescence microscopy, the number of cells with positively stained nuclei were counted and percentage of dead cells calculated. As a positive control for cytotoxicity, cells were incubated with 0.1% saponin for 10 min before staining. Infecting cells with any Salmonella strain used in our experiments at a moi = 100 killed 60% of the cells by 15 h p.i. In contrast, less than 10% of RAW264.7 cells of the entire culture were dead by 2 or 15 h p.i. when infected at a moi of 0.5 or 5 (Fig. 2). Similar results were obtained measuring the release of lactate dehydrogenase from cells (data not shown). Thus under the infection conditions used in this study, cell death is minimal. Therefore, we used these infection conditions to investigate the role of SopB post invasion.

image

Figure 2. moi-dependent cytotoxicity following SPI1-mediated invasion of RAW264.7 cells. Cells were infected with SPI1-induced WT Salmonella, ΔsopB or ΔsopB complemented with the empty plasmid pACYC184 (pACYC) at moi of 0.5, 5 and 100. At 2 h (A) or 15 h p.i. (B), cells were incubated with the LIVE/DEAD Viability/Cytotoxicity reagent for 30 min at 37°C, which stains the nuclei of dead cells. Cells were examined by fluorescence microscopy and the number of dead cells was counted. As a positive control for cytotoxicity, cells were incubated with 0.1% saponin (sap) for 10 min before staining. Results are the mean ± SD of three separate experiments (n = 100 cells).

Download figure to PowerPoint

Because iNOS is a downstream target of Akt (Park et al., 2003; Ozes et al., 1999), we next quantified iNOS activity by measuring nitrite accumulation in the culture supernatant using the Griess reagent. Cells infected with the ΔsopB (moi = 0.5) strain released 40% less nitrite than WT-infected cells (26.3 µM) 15 h p.i. (Fig. 3A). This reduction in nitrite release did not result from a defect in invasion (Fig. 4A) or replication (Fig. 4B) as measured by the gentamicin protection assay. To confirm the role of SopB, we also compared nitrite production in cells infected with the ΔsopB strain complemented with a plasmid expressing SopB (pDE), the catalytically inactive SopB (pDE C462S) (Steele-Mortimer et al., 2000) or the empty pACYC184 vector (pACYC). Although all plasmid-containing strains were slightly invasion and replication defective (Fig. 4A and B), complementation with pDE increased nitrite accumulation compared with pACYC (P < 0.005) or pDE C462S controls (P < 0.005) (Fig. 3A). Collectively these results show that catalytically active SopB is required for maximal iNOS induction by S. Typhimurium.

image

Figure 3. Nitrite production and iNOS induction are SopB dependent. A. RAW264.7 cells were infected with Salmonella strains (moi = 0.5) and culture supernatants collected 15 h p.i. to measure nitrite levels using the Griess reagent. Values are represented as percentage nitrite released by WT-infected cells. (WT nitrite concentration = 26.3 µM). B. Under the same infection conditions, whole-cell extracts were analysed by immunoblotting using polyclonal anti-iNOS antibodies and monoclonal anti-actin antibodies. iNOS levels were quantified after normalizing to actin levels and are expressed relative to the levels of iNOS protein in WT infected cells. All values represent the means ± SD from at least three separate experiments.

Download figure to PowerPoint

image

Figure 4. During SPI1-mediated infection of RAW264.7 cells, SopB does not affect invasion (A) or replication (B) as measured by the gentamicin protection assay. Cells were infected with Salmonella strains for 1 h or 15 h, solubilized and colony-forming units (cfu) enumerated. The Salmonella strains used were: WT; ΔsopB; ΔsopB complemented with the empty plasmid pACYC184 (pACYC), the plasmid containing the WT sopB (pDE) or the plasmid containing sopB with a point mutation (C462S) that encodes a catalytically inactive SopB (pDE C462S); a mutant completely lacking SPI2 (ΔSPI2) or the double mutant ΔSPI2/ΔsopB. Fold-replication represents the cfu recovered at 15 h divided by the cfu at 1 h. Results are the mean ± SD of at least three separate experiments each in triplicate.

Download figure to PowerPoint

These results show that SopB alone does not account for the total NO release after Salmonella infection; therefore, it is possible that other effectors contribute to this response. Because SPI2 has previously been shown to affect iNOS localization (Chakravortty et al., 2002), we infected cells with a ΔSPI2 strain, lacking the entire SPI2 region. While this mutant has a replication defect (60% less than WT; Fig. 4B) comparable to other SPI2 TTSS mutants (Hensel et al., 1998; Ochman et al., 1996), it induced similar nitrite release compared with WT-infected cells (Fig. 3A). Cells infected with the double mutant ΔSPI2/ΔsopB released comparable amounts of nitrite to cells infected with the ΔsopB strain. These data may be interpreted in one of two ways. Either, SPI2 effectors are involved in reducing NO production so that, per bacterium, ΔSPI2 bacteria would induce more NO release than WT bacteria (Figs. 3A and 4B) or, alternatively, NO release is dependent on the number of bacteria initially internalized and SPI2 effectors play no role.

There are several possible explanations for the observed reduction in NO release because iNOS is subject to regulation at transcriptional, translational and post-translational levels (Alderton et al., 2001). We initially compared levels of iNOS transcript in WT and ΔsopB-infected cells by quantitative real-time polymerase chain reaction (PCR). These experiments revealed no difference in iNOS transcript levels over a 15 h time-course (data not shown) indicating that SopB-dependent iNOS regulation does not occur transcriptionally. We next measured iNOS protein levels in infected cells by immunoblotting. After 15 h, cells infected with WT Salmonella had increased iNOS levels compared with uninfected cells, while infection with the ΔsopB strain displayed an intermediate level of induction (Fig. 3B), a difference remarkably similar to that for nitrite accumulation (compare Fig 3A and B). Furthermore, complementation of ΔsopB with pDE resulted in increased iNOS levels compared with complementation with the empty vector pACYC (P < 0.05) or inactive SopB mutant pDE C462S (P < 0.05) (Fig. 3B), demonstrating that iNOS induction is SopB dependent. In agreement with our nitrite assay data, deletion of SPI2 did not significantly affect the ability of Salmonella to increase intracellular iNOS protein levels. Collectively, these results suggest that SopB controls NO release by regulating iNOS levels post-transcriptionally.

Because SopB influences NO and iNOS levels long after invasion, we next determined the length of time this effector persisted in infected cells. We carried out experiments similar to those previously used to demonstrate translocation of SopB to epithelial cell membranes (Marcus et al., 2002). Infected RAW264.7 cells were mechanically disrupted and centrifuged at low speed to separate cell lysates into soluble (host cell cytoplasm and membranes) and insoluble (intact bacteria, host cell nuclei, unbroken host cells and host cell cytoskeleton) fractions. Immunoblotting analysis of the fractions revealed SopB in intracellular WT bacteria up to 12 h p.i. (Fig. 5A, insoluble) although SopB translocated from WT Salmonella was below the limits of detection (data not shown). Therefore, to examine the persistence of translocated SopB we took advantage of the pDE plasmid, a low-copy-number plasmid that expresses SopB under the control of the native sopB promoter. Because of the multicopy nature of this plasmid (estimated at 10–15 copies; New England Biolabs), SopB levels in ΔsopB pDE bacteria are higher than the endogenous levels in WT bacteria. Like for WT infections, SopB persisted in intracellular ΔsopB pDE bacteria for up to 12 h p.i. (Fig. 5B, insoluble). In contrast to what would be predicted for a SPI1 effector, translocated SopB was detected in host cells for up to 12 h p.i. (Fig. 5B, soluble). Importantly, bacterial integrity was not compromised in these experiments as the cytosolic bacterial chaperone DnaK was not detected in the soluble fraction. As intracellular bacteria begin to replicate DnaK levels increased (Fig. 5A and B) in the insoluble fraction, results agreeing with those previously published (Buchmeier and Heffron, 1990). These data demonstrate that, after its translocation during the initial invasion event, SopB either persists or continues to be translocated into the host cell for many hours. The persistence of SopB may explain how a SPI1 effector can modulate cellular events, such as iNOS expression, that occur late in the infection process.

image

Figure 5. Translocated SopB persists for at least 12 h p.i. in RAW264.7 cells. Translocated SopB (soluble) and SopB in internalized Salmonella (insoluble) were monitored in WT (A) and ΔsopB pDE (B) infected cells. At the times indicated (1–12 h p.i.), cells were fractionated by mechanical disruption followed by low-speed centrifugation to yield an insoluble fraction (intact bacteria, host cell nuclei, unbroken host cells and host cell cytoskeleton) and a soluble fraction (host cell cytoplasm and membranes). Fractions were analysed by immunoblotting using affinity-purified polyclonal anti-SopB antibodies (SopB is indicated by an asterisk), monoclonal anti-DnaK antibodies and polyclonal anti-calnexin antibodies. The cytosolic bacterial chaperone DnaK was monitored as a control for Salmonella lysis and calnexin was used as a loading control for host cell proteins. Molecular mass markers are shown in kDa on the left side.

Download figure to PowerPoint

To expand upon these results, we used quantitative real-time PCR to directly measure sopB mRNA levels in intracellular bacteria. As predicted from previous studies in epithelial cells (Boddicker and Jones, 2004), sopB transcript levels steadily declined after invasion by SPI1-induced WT Salmonella(Fig. 6A), whereas pipB (a SPI2 effector) mRNA levels increased with maximal expression occurring at 4–8 h p.i. (Fig. 6B). Nevertheless, we found that sopB mRNA levels were detectable for at least 8 h p.i. overlapping with the maximal induction of pipB (Fig. 6). In addition, transcripts of prgI, a structural gene of the SPI1 TTSS (Pegues et al., 1995; Kubori et al., 2000), and invF, a key regulator of SPI1 genes (Darwin and Miller, 1999; Eichelberg and Galan, 1999), were detectable 4 and 8 h p.i. (data not shown). After normalization to gene copy number, the reductions in sopB mRNA levels over 8 h were proportional for WT and ΔsopB pDE infections (Fig. 6A), confirming that sopB regulation and stability are comparable whether chromosomally or episomally encoded. These data represent the first characterization of transcript levels in Salmonella after SPI1-mediated invasion. They show that SPI1 gene expression can continue for several hours p.i. and significantly overlap with the induction of SPI2 genes.

image

Figure 6. sopB transcript levels decline after invasion, but persist for at least 8 h p.i. (A) sopB and (B) pipB mRNA levels from infected RAW264.7 cells (moi = 5) were measured by quantitative real-time PCR at the times indicated p.i. RNA and DNA were isolated concurrently using TRIzol. mRNA levels were normalized to genome number for each sample and each value expressed relative to that at 1 h p.i. *  = ratio of mRNA to gene copy number at 1 h p.i. (for sopB) or 4 h p.i. (for pipB). Representative data from one experiment are shown. Measurements are means from triplicate wells ± SD.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

It has recently been suggested that some SPI1 effectors persist in host cells for several hours after translocation (Marcus et al., 2002; Kubori and Galan, 2003). Here we have addressed the hypothesis that a SPI1 effector, SopB, may contribute significantly to post-invasion host cell interactions previously attributed to the SPI2 TTSS. The expression, translocation, persistence and function of SopB were analysed after SPI1-mediated invasion of RAW264.7 cells. Importantly, we found Salmonella-induced cytotoxicity to be very dependent on bacterial load, in agreement with others (Fig. 2; Weiss et al., 2004), so we used low moi to minimize cell death in our studies. Under such infection conditions, SopB-dependent Akt phosphorylation occurred rapidly and was localized to membrane ruffles as seen in epithelial cells (Steele-Mortimer et al., 2000). However, in macrophages we also observed SopB-independent Akt phosphorylation that was initially minor (up to 1 h p.i.) but later the level of phosphorylation was comparable to that seen in WT-infected cells. Induction of NO release and iNOS were SopB dependent. Again iNOS induction is not absolutely dependent on SopB and may result from SopB-independent Akt activation or the action of unidentified effectors. It is typical for SPI1 effectors to observe minor changes in host response because of redundancy of effector function (Zhou et al., 2001; Zhou and Galan, 2001). In epithelial cells and macrophages, SopB expression can be detected for at least 8–12 h p.i. (Marcus et al., 2002 and Fig. 5). The persistence of SopB protein was corroborated by real-time PCR, which showed that sopB transcript persists for at least 8 h p.i. Using SPI1-induced Salmonella, our data show that sopB expression decreased over time intracellularly, whereas a previous study found no change, a discrepancy almost certainly resulting from the fact that Eriksson et al. (2003) infected cells with bacteria grown to minimize SPI1 induction. In cultured macrophages it is unclear if SPI1-mediated invasion or a phagocytic method of internalization better represents the mechanism of Salmonella infection in vivo.

Our findings that SopB is expressed and persists in host cells for at least 8 h after invasion support our hypothesis that SPI1 effectors have the potential to influence host cell responses after the induction of the SPI2 regulon. Furthermore, SopB-dependent stimulation of iNOS demonstrates that SPI1 and SPI2 overlap in targeting host cell functions. Previously, the SPI1 effector, SopE2, has also been implicated in iNOS regulation. However, the mutants used were invasion compromised and the decrease in iNOS levels could not be restored by sopE2 complementation, indicating that the effect on iNOS regulation probably resulted from invasion deficiencies (Cherayil et al., 2000). It is possible that both effectors contribute to the regulation of iNOS as they have been shown to have some overlap in function (Zhou and Galan, 2001; Zhou et al., 2001). SPI2 has been shown to prevent delivery of iNOS to the SCV, thus spatially regulating iNOS function (Chakravortty et al., 2002), but we could not detect any significant difference in iNOS localization in our study (data not shown). Again, this discrepancy may result from different experimental conditions used to infect macrophages and will require further investigation.

Previous work has shown a role, although minor, for NO in controlling Salmonella infection in primary mouse macrophages from iNOS knockout mice (Vazquez-Torres et al., 2000). Our results suggest that the decrease in NO release (40%) from a cultured macrophage-like cell line is not significant enough to affect bacterial replication. In fact, it has been shown that reducing nitrite release by > 95% results in less than a twofold increase in replication in WT-infected RAW264.7 cells (Chakravortty et al., 2002). These in vitro findings are consistent with the mouse model of Salmonella pathogenesis demonstrating that iNOS plays a minor role during the early stage of infection (less than 1 week), whereas the oxidative burst from NADPH phagocyte oxidase is the major mechanism for controlling Salmonella proliferation early in the course of infection (Mastroeni et al., 2000).

SopB specifically stimulates iNOS production and early (up to 1 h p.i.) activation of Akt. It remains unclear if the SopB-dependent Akt activation observed within 1 h of invasion results in SopB-specific iNOS induction. We believe that SopB increases iNOS protein levels post-transcriptionally because real-time PCR did not reveal any differences in iNOS transcript (not shown). The factor(s) responsible for the SopB-independent iNOS induction remain unclear, but bacterial lipopolysaccharide is known to stimulate iNOS production (Lyons et al., 1992) and has recently been shown to be the major stimulus of TLR4 induction of iNOS transcription (Vazquez-Torres et al., 2004). However, this does not preclude the involvement of other, as yet unidentified, type III effectors in iNOS regulation.

Historically, there is a well-established requirement for SPI1 effectors in early events in Salmonella pathogenesis, but our results demonstrate that a SPI1 effector also contributes to later host cell interactions. Indeed, because SPI2 is not induced immediately after invasion, it is likely that other SPI1 effectors will also play significant roles in post-invasion events, as has been suggested for intracellular replication and cholesterol trafficking in non-phagocytic cells (Garner et al., 2002; Steele-Mortimer et al., 2002). Further research on the expression kinetics, longevity and function of SPI1 effectors in host cells will probably confirm the complementary nature of SPI1 and SPI2 effector actions in Salmonella pathogenesis.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Cell culture and bacterial strains

RAW264.7 mouse macrophage-like cells (ATCC ♯ TIB-71) were obtained from the American Type Culture Collection and maintained as directed in DMEM containing 10% heat-inactivated fetal calf serum. Low-passage-number (<11) cells were seeded in six-well dishes at 1.5 × 106 cells per well and grown overnight until 70–80% confluent. S. enterica serovar Typhimurium SL1344 WT, ΔsopB, ΔsopB pDE (WT SopB), ΔsopB pDE C462S (catalytically inactive SopB) and ΔSPI2::kan strains were as described (Steele-Mortimer et al., 2000; Knodler et al., 2003). The SL1344 ΔSPI2/ΔsopB strain was constructed by P22 transduction from SL1344 ΔSPI2::kan (Knodler et al., 2003) to the ΔsopB strain.

Salmonella infection of macrophages

Salmonella were grown under conditions to induce SPI1 gene expression. Briefly, overnight shaking 37°C cultures were diluted 1:33 and grown shaking at 37°C for 3.5 h to late log phase (Steele-Mortimer et al., 2000). Bacteria were added to RAW264.7 cells at an moi = 0.5 (nitrite release and iNOS immunoblot experiments), moi = 5 (real-time PCR experiments) or moi = 15 (SopB immunoblots) for 10 min at 37°C. Extracellular bacteria were then removed and monolayers washed twice in growth media (GM). Infected cells were then incubated for 20 min in GM, 60 min in GM + 100 µg ml−1 gentamicin to kill extracellular bacteria and finally GM + 10 µg ml−1 gentamicin for the remainder of the experiment. For gentamicin protection assays, cells were solubilized at 1 h or 15 h p.i. and plated onto Luria–Bertani (LB) agar for bacterial enumeration. In experiments measuring nitrite accumulation, iNOS protein levels and Salmonella invasion and replication  RAW264.7  cells  were  incubated  in  GM containing low serum (0.5%) for 3 h before infection and throughout the experiment.

Transfection and immunofluorescence

RAW264.7 cells were transfected with GFP-Akt (Watton and Downward, 1999) using FuGENE 6 (Roche). Plasmid DNA was prepared using an EndoFree Plasmid Maxi kit (QIAGEN). Cells were infected with Salmonella 24 h after transfection. Cells were fixed 20 min p.i. in 2.5% paraformaldehyde for 10 min and permeabilized in PBS containing 10% goat serum and 0.2% saponin. Immunofluorescence microscopy was as described (Steele-Mortimer et al., 2000). Briefly, cells were incubated with rabbit polyclonal anti-phospho Akt (Ser 473) antibodies (1:100; Cell Signaling), which recognize only Akt phosphorylated at Ser-473, and mouse monoclonal anti-lipopolysaccharide (LPS) antibodies (1:2000; Biodesign International). Primary antibodies were detected using donkey anti-rabbit Cy5 (1:1000; Jackson Immuno Research) and goat anti-mouse Alexa Fluor 568 secondary antibodies (1:1000; Molecular Probes). Images were collected with a Perkin Elmer UltraView spinning-disc confocal system connected to a Nikon Eclipse TE2000-S microscope using a 60×, 1.4 NA oil immersion objective. Each micrograph is a projection of 20 confocal sections (0.3 µm thick) processed with ImageJ and deconvolved using Priism 4.1.2.

Cytotoxicity assays

RAW264.7 cells were grown overnight on glass coverslips before infection with Salmonella (moi = 0.5, 5 or 100) as described above. At 2 h or 15 h p.i., media were removed and cells were incubated with LIVE/DEAD Viability/Cytotoxicity reagent (Molecular Probes) for 30 min at 37°C. As a positive control for cytotoxicity, 0.1% saponin was added to cells for 10 min at 37°C. Coverslips were washed twice in PBS and mounted on slides with Mowiol (Calbiochem). Cells were examined by fluorescence microscopy and the number of cell nuclei stained with ethidium homodimer-1 was enumerated. The results are the mean ± SD of three separate experiments (n = 100 cells for each condition).

Nitric oxide measurements

Culture supernatant nitrite, the final oxidation product of NO, was measured by combining culture supernatants 1:1 with the Griess reagent (Green et al., 1982). Absorbance at 550 nm was measured using a Bio-Tek FL600 microplate reader (BIO-TEK Instruments) and nitrite concentration determined using a sodium nitrite standard curve.

Statistical analysis

Unless otherwise stated, all results are the average ± SD of at least three separate experiments. P-values were determined using the t-test and Fisher's method of combining P-values.

Immunoblotting

Cell extracts were prepared as described (Steele-Mortimer et al., 2000). Briefly, infected RAW264.7 monolayers from six-well plates were solubilized by adding 0.5 ml of hot SDS sample buffer containing 50 mM dithiothreitol. Samples were boiled for 5 min and proteins separated by SDS-polyacrylamide gel electrophoresis (10%) and transferred to nitrocellulose membranes (Bio-Rad). Membranes were blocked in Tris-buffered saline (TBS) containing 0.1% (v/v) Tween 20 and 1% (w/v) bovine serum albumin Fraction V (Merck kgaA). Membranes were probed with rabbit polyclonal anti-phospho Akt Ser-473 (1:1000; Cell Signaling), rabbit polyclonal anti-Akt (1:1 000; Cell Signaling), mouse monoclonal anti-actin (1:10 000; Santa Cruz), rabbit polyclonal anti-calnexin (1:40 000; Stressgen), mouse monoclonal anti-DnaK (1:5 000; Stressgen) or rabbit polyclonal anti-iNOS antibodies (1:20 000; Calbiochem). Goat anti-mouse horseradish peroxidase (HRP) and goat anti-rabbit HRP secondary antibodies (Cell Signaling) were used at 1:10 000 dilution. Secondary antibodies were detected using SuperSignal West Femto (Pierce) according to the manufacturer's instructions. Membranes were scanned using a Kodak Image Station 1000 and iNOS protein amounts quantified after normalizing to actin levels.

Analysis of intracellular SopB levels

Fractionation of infected host cells was as described (Knodler et al., 2003). Monolayers from two 10 cm dishes were infected (moi = 15) for 10 min as described above, washed in cold PBS and gently scraped and resuspended in 0.3 ml of homogenization buffer [3 mM imidazole pH 7.4, 250 mM sucrose, 0.5 mM EDTA, Protease Inhibitor cocktail III (Calbiochem)]. Cells were mechanically disrupted by two to three passes through a 22-gauge needle. A low speed centrifugation (7000 g) was used to pellet insoluble material (intact bacteria, host cell nuclei, unbroken host cells and host cell cytoskeleton) from soluble material (host cell cytoplasm and membranes). The insoluble pellet was resuspended in 0.3 ml of hot SDS sample buffer and 60 µl of hot 6× SDS sample buffer was added to the soluble supernatant fraction. Samples were resolved on 10% SDS-PAGE gels and transferred to nitrocellulose. Immunoblotting with affinity-purified polyclonal anti-SopB antibodies was as described previously (Marcus et al., 2002).

Real-time PCR

RNA and DNA were isolated simultaneously using TRIzol (Invitrogen) according to the manufacturer's instructions, from infected RAW264.7 cells grown in six-well plates. cDNA was synthesized using 2 µg of RNA and the TaqMan Reverse Transcription Reagent Kit (Applied Biosystems) according to the manufacturer's instructions. Primers and probes (Table 1) were designed using primer express® 1.5 version (Applied Biosystems). Real-time PCR was performed according to the manufacturer's instructions in an ABI Prism 7700 Sequence Detector (Applied Biosystems). The amount of cDNA was normalized to sopB gene copy number.

Table 1.  Quantitative real-time PCR primers and probes.
sopB fwd primer5′-GGCGGCGAACCCTATAAAC-3′
sopB rev primer5′-GGGTACCGCGTCAATTTCAT-3′
sopB probe5′-(VIC™)CGCACAACGCCTTGCCATGTTAGC-3′(TAMRA)
pipB fwd primer5′-GCGCTAACATGTCCGGTGTA-3′
pipB rev primer5′-GCACCATTTAGTTTGGTGTCAGTT-3′
pipB probe5′-(VIC™)AACCGCTGCAATTCTATTCGGCTCAGA-3′(TAMRA)

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We are grateful to Debbie Crane, John Carlson, Jens Gieffers, Scott Grieshaber, Rey Carabeo, Ken Fields and Roberto Rebeil for expert assistance and advice.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References