|Home | About | Journals | Submit | Contact Us | Français|
Recognition of cytoplasmic bacterial flagellin by the Nod-like receptor ICE protease-activating factor (Ipaf) in macrophages leads to activation of caspase-1 and secretion of interleukin-1β (IL-1β). Salmonella possess two genes, fliC and fljB, that encode flagellin proteins. We examined the ability of purified FliC and FljB proteins to activate IL-1β secretion in the mouse macrophage-like J774 cell line and in mouse primary peritoneal cells. We found that purified FliC and FljB flagellins possessed a comparable ability to activate IL-1β secretion following introduction into the cytoplasm of J774 or primary cells. We also examined the ability of an attenuated Salmonella mutant strain (dam) deficient in DNA adenine methylase to activate IL-1β secretion. Compared to infection of primary cells with wild-type Salmonella, IL-1β secretion was reduced in cells infected with the dam mutant even though the two strains expressed similar levels of flagellin. As a control, cells infected with a flagellin-deficient (flhC) Salmonella strain did not show enhanced IL-1β secretion.
Salmonella are intracellular pathogens that establish infection by invasion and multiplication in the host's epithelial and monocytic cells (Chen and others 1996; Galen and Collmer 1999; Garcia-del Portillo 2001; Hueffer and Galan 2004). Salmonella infection of macrophages can lead to pyroptosis, a cytotoxic process that involves early cell death (Chen and others 1996; Hueffer and Galan 2004; Franchi and others 2006; Miao and others 2006). Bacterial flagellin is one Salmonella virulence factor (Miao and others 2007; Steiner 2007; Ramos and others 2004). Salmonella expresses two flagellin proteins, FliC (flagellin phase type 1/H1) and FljB (flagellin phase type 2/H2). A complex phase-variation process mediated by posttranscriptional control results in the production of only FliC or FljB at any given time (Aldridge and Hughes 2002; Bonifield and Hughes 2003). In addition to the role of flagellin in the formation of the filamentous flagella polymers and bacterial motility, flagellin also is recognized as a pathogen-associated molecular pattern (PAMP) by the host innate immune system (Kawai and Akira 2006).
Two kinds of cellular sensors of bacterial flagellin have been characterized, one extracellular and the other intracellular, that trigger innate immune responses (Miao and others 2007; Steiner 2007). During the course of infection, Salmonella sheds flagellin monomers (Garcia-Del Portillo and others 1999; Tallant and others 2004; Franchi and others 2006; Miao and others 2006; Mariathasan 2007). Extracellular flagellin monomers are recognized by the transmembrane Toll-like receptor 5 (TLR5) protein that signals through a MyD88-dependent pathway to activate NF-κB and induce proinflammatory cytokine and chemokine gene expression (Gewirtz and others 2001; Tallant and others 2004; Feuillet and others 2006; Miao and others 2007; Steiner 2007). The two Salmonella flagellins, FliC and FljB, display a comparable specific activity for activation of NF-κB (Simon and Samuel 2007a).
In addition to the extracellular recognition of flagellin by TLR5, flagellin subunits also are translocated into the host cell cytoplasm by a process dependent upon the Salmonella Pathogenicity Island I type III secretion system (SPI-1 TTSS) (Sun and others 2007). Intracellular flagellin is recognized by the Nod-like receptor ICE protease-activating factor (Ipaf), which causes activation of the inflammasome and caspase-1 dependant secretion of interleukin-1β (IL-1β; Franchi and others 2006; Miao and others 2006; Mariathasan 2007). Salmonella deficient in either flagellin or the SPI-1 TTSS fail to trigger significant caspase-1 activation, IL-1β secretion or host cell death (Franchi and others 2006; Miao and others 2006; Mariathasan 2007). However, the ability of Salmonella singularly deficient in either FliC or FljB flagellin to activate IL-1β secretion following infection is not completely clear. Phase-locked Salmonella strains that express only one of the two flagellin monomer proteins, either FliC or FljB, are comparably cytotoxic to macrophages, suggesting that FliC and FljB flagellins comparably induce caspase-1 activation and IL-1β production (Miao and others 2006). By contrast, studies with Salmonella mutants lacking either flagellin gene show significantly reduced IL-1β secretion by macrophages infected with FliC-deficient compared to FljB-deficient Salmonella, indicating that FliC is the main Salmonella flagellin required for IL-1β secretion and early death of macrophages (Franchi and others 2006). In addition to flagellin, bacterial DNA adenine methylase (Dam) also is a Salmonella virulence factor (Heithoff and others 1999; Heithoff and others 2001; Low and others 2001). Although not required for Salmonella viability, mutants of Salmonella lacking functional Dam methylase are attenuated for virulence in the mouse model (Heithoff and others 1999; Heithoff and others 2001; Shtrichman and others 2002; Simon and others 2007). The mechanistic basis of the reduced virulence associated with dam mutant strains of Salmonella is not precisely known.
In this report, we show that purified FliC and FljB flagellin proteins display a comparable capacity to activate IL-1β secretion when introduced into the cytoplasm of either immortalized J774 macrophage-like cells or primary peritoneal exudate cells (PECs). These primary cells infected with mutant Salmonella deficient in the Dam methylase secrete less IL-1β compared to cells infected with wild-type (WT) Salmonella, whereas as reported, cells infected with mutant Salmonella deficient in flagellin do not induce secretion of IL-1β (Franchi and others 2006; Miao and others 2006).
The Salmonella enterica serovar typhimurium WT and damΔ232 mutant strains used in this study were derived from strain ATCC 14028 (CDC 6516-60) and were previously described (Shtrichman and others 2002; Simon and others 2007). The flhC mutant strain also was as previously described (Chilcott and Hughes 2000; Badie and others 2007; Simon and Samuel 2007b). All Salmonella strains used were generously provided by M. Mahan (University of California, Santa Barbara).
Peritoneal exudate cells were prepared essentially as described (Li and others 1997). The experimental protocol was approved by the Institutional Animal Care and Use Committee. In brief, 8–10 week old female BALB/c mice were injected intraperitoneally with 1 mL of 4% sodium thioglycollate and 4% dextrose in phosphate-buffered saline (PBS) at pH 7.4; after 4 days PECs were harvested by peritoneal wash using a total of 5 mL of ice cold 30% sucrose. Primary PEC and the macrophage-like J774 cell line were maintained in DMEM (Invitrogen) supplemented with 10% (v/v) fetal bovine serum (Hyclone), 1 mM sodium pyruvate, 300 μg/mL l-glutamine, 100 μg/mL penicillin and 100 units/mL streptomycin (Invitrogen) in a 5% CO2 humidified environment at 37°C.
Mammalian cells cultured in 24-well plates in antibiotic-free medium were pretreated for 3 h with 10 ng/mL lipopolysaccharide (LPS) isolated from Salmonella typhimurium (Sigma). Subsequent infection with Salmonella was initiated by applying the bacterial inoculum at the specified multiplicity of infection (MOI) to the cell monolayer followed by centrifugation 1,0003 g for 10 min. Salmonella inoculum was prepared by dilution of an overnight shaking culture 1:40 in Luria-Bertani broth followed by incubation for ~3 h at 37°C with shaking (300 rpm) to achieve mid to late log phase growth. The inoculum was then washed and suspended in PBS and the bacterial concentration determined spectrophotometrically.
Monomeric flagellin proteins 1 (FliC) and 2 (FljB) purified from the Salmonella mutant strains fliC+/fljB− (flagellin 1) and fliC−/fljB+ (flagellin 2) were as described (Simon and Samuel 2007a). For transfection analyses, murine PEC or J774 cells seeded in 12- or 24-well plates were first treated for 3 h with 10 ng/mL LPS, then washed twice with serum-free media. Transfection with FliC or FljB flagellin, or bovine serum albumin as a control, was performed using a ratio of 70 ng protein per 1 μL of Lipofectamine 2000 (Invitrogen).
Cell-free culture supernatant fractions were analyzed 3 h after transfection for released IL-1β using an enzyme-linked immunosorbent assay (ELISA) kit (MLB00B, R&D systems). Duplicate samples were analyzed in each independent experiment.
Proteins were fractionated by SDS-polyacrylamide gel electrophoresis (10% acrylamide) gels, transferred to nitrocellulose membranes, and probed with an anti-flagellin 15D8 monoclonal antibody (Bioveris #15D8) at a dilution of 1:500. Detection and visualization of antibody–antigen complex formation was accomplished with the LICOR Odyssey infrared imager (Li-Cor Biosciences).
Exposure of mouse macrophage cells to LPS results in rapid transcriptional activation of the IL-1β gene (Fenton and others 1987; Dinarello 1996). The IL-1β functional precursor protein, pro-IL-1β, is sequestered within the cytoplasm and undergoes rapid proteolytic maturation and secretion following activation by a second signal, for example extracellular ATP through ASC (Perregaux and Gabel 1998; Mariathasan 2007; Mariathasan and others 2004) or intracellular flagellin through Ipaf (Franchi and others 2006; Miao and others 2006). While the ability of cytoplasmic flagellin to activate caspase-1 and stimulate the secretion of IL-1β from macrophages is well established (Franchi and others 2006; Miao and others 2006; Sun and others 2007), the relative ability of the two Salmonella flagellin proteins to elicit IL-1β processing and secretion is unclear. Studies using bacteria singularly deficient for one or the other flagellin proteins, FliC or FljB, gave inconclusive results (Franchi and others 2006; Miao and others 2006). As an independent approach to directly determine the ability of FliC flagellin and FljB flagellin to cause IL-1β release, we introduced purified FliC or FljB proteins (Simon and Samuel 2007a) into the J774 macrophage cell line by transfection. Cells were first stimulated with LPS to induce pro-IL-1β expression, which in the absence of infection or flagellin transfection, does not activate pro-IL-1β processing by caspase-1 and subsequent secretion (Franchi and others 2006; Miao and others 2006; Miao and others 2007; Steiner 2007).
As shown in Figure 1, transfected FliC and FljB flagellins stimulated a comparable release of IL-1β from J774 cells. The release of IL-1β following transfection of 30 ng of flagellin was not saturating and was quantitatively similar for FliC and FljB, whereas 100 ng protein gave a saturating IL-1β release with both flagellins (Fig. 1). By contrast, IL-1β production remained low when J774 cells were transfected with up to 300 ng of BSA and did not differ significantly from the low IL-1β production seen when either 300 ng of FliC flagellin protein was added without the transfection reagent or when the Lipofection 2000 transfection reagent was used alone without protein (Fig. 1).
The ability of purified flagellin monomer proteins to cause IL-1β release was also examined in primary cells. Thioglycollate-elicited PECs were transfected with increasing amounts of either FliC or FljB flagellin. For a protein transfection control, BSA again was used. As we observed for the immortalized J774 macrophage cell line, the FliC and FljB flagellins likewise showed a similar ability to stimulate IL-1β secretion in transfected PECs (Fig. 2). However, BSA failed to cause significant IL-1β release when transfected into PECs (Fig. 2). Interestingly, the amount of IL-1β released was nearly 10-fold higher from primary PECs (Fig. 2) as compared to J774 cells (Fig. 1) following transfection with purified Salmonella flagellins.
The characterization of flagellar mutants has firmly established that caspase-1 activation and subsequent IL-1β secretion from Salmonella-infected macrophages requires flagellin (Franchi and others 2006; Miao and others 2006; Sun and others 2007), a known virulence factor (Ramos and others 2004; Miao and others 2007; Steiner 2007). DNA adenine methylase is also a known Salmonella virulence factor. Mutants deficient in Dam are attenuated (Heithoff and others 1999; Low and others 2001); however, nothing is known regarding the ability of dam mutants to activate IL-1β secretion compared to WT Salmonella. We therefore measured the secretion of IL-1β by primary PECs following infection with WT, dam mutant or flagellin-deficient (flhC) Salmonella. As shown in Figure 3A, infection of PECs with WT Salmonella greatly enhanced the secretion of IL-1β in an MOI dependent manner, whereas infection with the dam mutant showed a reduced IL-1β production compared to WT that was less pronounced at high MOI. As anticipated, the flhC mutant failed to cause much production of IL-1β even at an MOI of 50 (Fig. 3A). Western blot analysis for bacterial flagellin using an antibody that recognizes the FliC and FljB flagellins with comparable avidity (Simon and Samuel 2007a) confirmed the absence of flagellin expression in the flhC mutant strain, and showed similar flagellin expression between the WT and dam mutant Salmonella strains (Fig. 3B). Finally, IL-1β secretion from J774 cells infected with WT Salmonella was modest even at high MOI (data not shown) compared to the robust release seen with WT-infected PEC cells (Fig. 3A).
The dual recognition of a given PAMP by unique extracellular and intracellular sensors creates an opportunity to tune subsequent host immune responses to infection. Such a conceptual mechanism appears operative both for bacteria, in the case of flagellin as a protein PAMP sensed by cell surface membrane-associated TLR5 and cytosolic Nod-like receptors (Miao and others 2007; Steiner 2007), and for viruses in case of double-stranded RNA as a nucleic acid PAMP sensed by endosome-associated TLR3 and cytosolic RIG-like receptors and PKR (Samuel, 2001, 2007; Uematsu and Akira 2007; Yoneyama and Fujita 2007). In the case of Salmonella, by virtue of the dual recognition of flagellin by TLR5 when present in the extracellular milieu and Ipaf when present in the cytosol of the host, flagellin can both cause IL-1β induction through a TLR5-dependent process and subsequent IL-1β secretion through an Ipaf-dependent process. Thus, intracellular pathogens such as Salmonella have the potential to modulate the immune response in a manner that is distinct from that of extracellular pathogens.
The reason that Salmonella expresses two different flagellin proteins by a phase-variation mechanism, in the context of Salmonella pathogenesis, is not yet clear. TLR5 appears to recognize a conserved domain region of flagellin required for protofilament formation and bacterial motility (Smith and others 2003; Andersen-Nissen and others 2005). We earlier established that the two Salmonella flagellins, FliC and FljB, possess a comparable specific activity for activation of NF-κB-dependent gene expression (Simon and Samuel 2007a). Our results presented herein extend this finding to the intracellular Ipaf-dependent cytosolic sensing of flagellin, whereby FliC and FljB displayed a comparable specific capacity to activate IL-1β secretion. Interestingly, Salmonella do appear to preferentially express the FliC flagellin during the course of infection in the intact animal (Ikeda and others 2001). The reason for this preferred expression of FliC over FljB in vivo remains unclear.
Dam deficient Salmonella are attenuated for virulence in the mouse model, and provide subsequent immunity to WT challenge (Heithoff and others 1999; Heithoff and others 2001; Shtrichman and others 2002; Simon and others 2007). While much is known regarding the phenotype of dam mutant Salmonella in the mouse, comparatively little has been revealed regarding specific biochemical defects seen following infection. Dam mutations are pleiotropic, showing reduced proinflammatory cytokine responses, increased sensitivity to bile, decreased invasion of intestinal epithelial cells, reduced M-cell cytotoxicity, altered expression of known virulence factors and increased release of periplasmic proteins (Garcia-Del Portillo and others 1999; Heithoff and others 1999; Pucciarelli and others 2002; Shtrichman and others 2002; Balbontin and others 2006; Simon and others 2007; Badie and others 2007). Our data suggest that dam mutant Salmonella are further impaired in their ability to cause IL-1β release in macrophages. While the mechanism responsible for the reduced release is unclear, our results are consistent with the notion that translocation of Salmonella proteins may in some way be impaired in dam mutant Salmonella (Garcia-Del Portillo and others 1999). This impairment may conceivably be at the level of expression of a component of the SPI-1 TTSS apparatus. Alternatively, the impaired ability of dam mutant to cause IL-1β release in macrophages may be due to the loss of integrity of bacterial surface structures integral to SPI-1 TTSS and virulence protein translocation function (Miao and others 2006). The contribution of impaired IL-1β release to the attenuated phenotype observed in the mouse model is at present unclear, as the nonflagellated flhC strain is fully virulent in the mouse model (Schmitt and others 2001).
This work was supported in part by the National Institute of Allergy and Infectious Diseases, U.S. Public Health Service. We thank Drs. Sambuddho Mukherjee and Cyril George for helpful discussions.