Although NLRC4/IPAF activation by flagellin has been extensively investigated, the downstream signaling pathways and the mechanisms responsible for infection clearance remain unclear. In this study, we used mice deficient for the inflammasome components in addition to wild-type (WT) Legionella pneumophila or bacteria deficient for flagellin (flaA) or motility (fliI) to assess the pathways responsible for NLRC4-dependent growth restriction in vivo and ex vivo. By comparing infections with WT L. pneumophila, fliI, and flaA, we found that flagellin and motility are important for the colonization of the protozoan host Acanthamoeba castellanii. However, in macrophages and mammalian lungs, flagellin expression abrogated bacterial replication. The flagellin-mediated growth restriction was dependent on NLRC4, and although it was recently demonstrated that NLRC4 is able to recognize bacteria independent of flagellin, we found that the NLRC4-dependent restriction of L. pneumophila multiplication was fully dependent on flagellin. By examining infected caspase-1−/− mice and macrophages with flaA, fliI, and WT L. pneumophila, we could detect greater replication of flaA, which suggests that caspase-1 only partially accounted for flagellin-dependent growth restriction. Conversely, WT L. pneumophila multiplied better in macrophages and mice deficient for NLRC4 compared with that in macrophages and mice deficient for caspase-1, supporting the existence of a novel caspase-1–independent response downstream of NLRC4. This response operated early after macrophage infection and accounted for the restriction of bacterial replication within bacteria-containing vacuoles. Collectively, our data indicate that flagellin is required for NLRC4-dependent responses to L. pneumophila and that NLRC4 triggers caspase-1–dependent and –independent responses for bacterial growth restriction in macrophages and in vivo.
Legionella pneumophila is a flagellated, Gram-negative, facultative intracellular pathogen that is the main cause of a severe form of pneumonia called Legionnaires disease (1). The bacteria naturally survive in freshwater reservoirs and multiply in a variety of unicellular protozoa (2–4). Human infection occurs when contaminated water droplets are inhaled and the bacteria gain access to mammalian lungs, where they can infect and multiply in alveolar macrophages. Once inside the macrophages, L. pneumophila trigger the expression of a set of genes evolutionarily selected to allow replication in protozoan hosts (3). Among those genes are the dot/icm members, which encode a type IVB secretion system called Dot/Icm that facilitates the secretion of more than 200 bacterial proteins, called effectors, into the host cell cytoplasm (5, 6). Whereas the function of most of the effector proteins are unknown, it has been shown that some of these proteins modulate the biogenesis of the L. pneumophila-containing vacuole and create an intracellular niche that supports intracellular replication (5, 6).
Whereas free-living amoeba, human and guinea pig macrophages support the robust growth of L. pneumophila, macrophages from most inbred strains of mice are restrictive for bacterial multiplication (reviewed in Ref. 7). Macrophages from the A/J mouse strain are a notable exception; they are highly susceptible because of polymorphisms in the naip5/birc1e gene (8–10). NAIP5 belongs to a family of intracellular pattern recognition proteins called nucleotide-binding oligomerization domain (Nod)-like receptors (NLRs), which comprise cytosolic proteins that are functionally and structurally related. NLRs are composed of three distinct domains: a centrally located nucleotide-binding oligomerization domain; a C-terminal region containing a leucine-rich repeat domain; and a variable N-terminal region responsible for triggering effector functions (11, 12). Perhaps the most widely studied NLR is NLRP3, which triggers the caspase-1 activation in a process dependent on the adapter protein ASC in response to diverse stimuli (12). Other well-studied members of the NLRs are Nod1 and Nod2, which signal via the Rip2 kinase to trigger transcriptional activation of inflammatory genes in response to bacterial and parasite infection (13–15). Although it has been shown that Nod1 and Nod2 recognize L. pneumophila, these proteins play a minor role in growth restriction in mouse models (16–19). In contrast, inflammasome-triggering NLRs, such as NAIP5 and NLRC4/IPAF, play a key role in the restriction of L. pneumophila multiplication in murine macrophages (20–23). Although the mechanisms by which NLRC4 and NAIP5 restrict L. pneumophila replication remain largely obscure, it is certain that these mechanisms are dependent on caspase-1 and are triggered in response to flagellated bacteria (20–25).
The NLRC4 inflammasome was shown to operate in response to several Gram-negative intracellular bacteria, including Shigella flexneri, Pseudomonas aeruginosa, and Salmonella enterica (26–30). Although flagellin expression by these bacteria was shown to trigger NLRC4 activation, flagellin-independent activation of NLRC4 was also reported (27, 29, 30). It was demonstrated that PrgJ, the basal body rod component of the Salmonella type III secretion system (T3SS) apparatus, triggers NLRC4 activation independently of flagellin (31). Furthermore, recognition of basal body rod proteins by NLRC4 was also reported for PrgJ orthologs present in Burkholderia pseudomallei, Escherichia coli, Shigella flexneri, and Pseudomonas aeruginosa (31). These studies indicated that NLRC4 is able to sense T3SS-expressing bacteria independent of flagellin. These bacterial models provided convincing demonstration of the broad role of NLRC4 in bacterial recognition. However, little is known regarding the mechanisms by which NLRC4 triggers the restriction of bacterial multiplication, possibly because these bacterial species are highly toxic for professional phagocytes, such as macrophages, dendritic cells, and neutrophils. Although much is known regarding the signals that trigger inflammasome activation in response to different bacterial pathogens, little is known about the effector responses that trigger the restriction of bacterial multiplication, particularly during in vivo infections. In this study, we used mice deficient for the inflammasome and L. pneumophila deficient for motility and flagellin to assess the NLRC4-dependent growth restriction and the signaling pathways that operate downstream of NLRC4 to restrict L. pneumophila multiplication in vivo and ex vivo. We found that NLRC4-dependent growth restriction was solely observed in response to flagellin-positive bacteria. Furthermore, NLRC4 signaling triggered caspase-1–dependent and –independent responses to restrict L. pneumophila multiplication in vivo and ex vivo.
Materials and Methods
The wild-type (WT) L. pneumophila JR32 strain was used (32). For competition experiments, the streptomycin-sensitive L. pneumophila F2111 strain was used (33). Non-pneumophila Legionella used included L. micdadei (ATCC 33218), L. bozemanii (ATCC 33217), L. gratiana (ATCC 49413), and L. rubrilucens (ATCC 35304). All bacteria grew in ACES or MOPS buffered charcoal–yeast extract (BCYE) agar [1% yeast extract, 1% 3-(N-morpholino)propanesulfonic acid (MOPS) or N-(2 acetoamido)-2-aminoethaneosulfonic (ACES), pH 6.9, 3.3 mM l-cysteine, 0.33 mM Fe(NO3)3, 1.5% Bacto agar, and 0.2% activated charcoal] at 37°C (34).
Construction of mutants
For the construction of mutant strains with fliI (lpg1757) clean deletion, PCR fragments were amplified using L. pneumophila genomic DNA as well as primer pairs fliI1/fliI2 (forward 5′-GGAGCTCGGTCATGTTTTAAACTACTATACC-3′; reverse 5′- AATGACCAATGAGTGACAGATTAGATCG-3′) and fliI3/fliI4 (forward 5′- TCTGTCACTCATTGGTCATTCATCACTATTCCTGAGAAACGC-3′; reverse 5′- GCTCTAGAGCGGATTTAATTCAAAAGCCCG-3′). The PCR products were used as a template for recombinant PCR using primer pair fliI1/fliI4. The resulting DNA fragment was digested with SacI and XbaI and cloned into pSR47S (pSR47S-fliI). This plasmid was used to construct fliI mutant L. pneumophila by allelic exchange (35). Bacterial mutants for flaA and dotA on the JR32 background were constructed using vectors pSR47S-flaA (36) and pSR47S-dotA (35), respectively.
For the detection of intracellular L. pneumophila flagellin, bacteria were grown in BCYE for the indicated time, and total cell lysates were used for Western blotting using the anti-flagellin Ab as previously described (25).For extraction of cell-associated extracellular proteins, bacteria were grown in BCYE medium for the indicated time at 37°C and suspended in 2 ml PBS followed by vigorously vortexing for 2 min to shear off the flagella. The cells were kept in an ice bath, and the procedure was repeated three times. The suspension was centrifuged at 20,000 × g at 4°C for 10 min to remove the bacterial cells. The flagella were precipitated with 3 ml acetone for 30 min at 20°C followed by centrifugation at 20,000 × g for 15 min at 4°C. The protein concentration was determined by Bradford assay (Sigma, St. Louis, MO). Proteins were resuspended in SDS-containing loading buffer, separated on a 15% SDS-PAGE gel, and transferred (Semidry Transfer Cell; Bio-Rad Laboratories, Hercules, CA) at 20 V for 40 min onto a nitrocellulose membrane (GE Healthcare, Piscataway, NJ) in transfer buffer (50 mM Tris, 40 mM glycine, and 10% methanol). Membranes were blocked overnight at 4°C in TBS with 0.1% Tween 20 and 5% nonfat dry milk and stained with primary Abs for 1 h at 25°C. Membranes were washed in TBS with 0.1% Tween 20 and incubated for 1 h at 25°C with the appropriate HRP-conjugated secondary Ab (1:3000 dilution; KPL, Gaithersburg, MD). Western blotting luminol reagent (Santa Cruz Biotechnology, Santa Cruz, CA) was used for Ab detection.
Growth of L. pneumophila in Acanthamoeba castellanii and macrophages
Bone marrow-derived macrophages (BMDMs) were prepared from the indicated mice and differentiated in L929 cells conditioned medium (LCCM) as previously described (37). BMDMs were replated in RPMI 1640 medium supplemented with 10% FBS and 5% L929 cells conditioned medium. Growth curves of L. pneumophila were performed as described (38). Briefly, Acanthamoeba castellanii (ATCC 30234) were cultured routinely at room temperature in ATCC medium 712. One hour before and after infection, A. castellanii cultures were maintained at 37°C in a 5% CO2 atmosphere in ATCC medium 712 without glucose, proteose peptone, and yeast extract. To measure intracellular multiplication, BMDMs or A. castellanii were plated at 2 × 105 cells/well in 24-well tissue culture dishes. Cultures were infected at multiplicity of infection (MOI) 10 followed by centrifugation for 5 min at 200 × g at room temperature and 1-h incubation at 37°C in a 5% CO2 atmosphere to allow for bacterial internalization. After 1 h, cultures were washed three times with PBS, and 1 ml medium was added to each well. For CFU determinations, the cultures were lysed in sterile water, and cell lysates were combined with cell culture supernatant from the respective well. Lysates plus supernatants from each well were diluted in water, plated on BCYE agar plates, and incubated for 4 d at 37°C for CFU determination. For competition assays, confluent A. castellanii were seeded in 175 cm2 tissue culture flasks and infected with 100 μl of a mix containing 3 × 103 of each bacterial strain. The flasks were incubated at room temperature without agitation. After 5 d of incubation, cells were lysed, and the total bacteria contained in the flasks was collected, diluted, and plated on BCYE plates with and without streptomycin for determination of the background of the bacterial colonies.
Immunofluorescence and the number of bacteria per vacuole
BMDMs were seeded at 2 × 105 cells/well on 13-mm glass coverslips contained in 24-well tissue culture dishes. The cells were infected with L. pneumophila with an MOI of 10 followed by centrifugation at 200 × g for 5 min and incubation for 1 h to allow phagocytosis. The cells were further washed three times with PBS at room temperature and incubated for an additional 9 h. The cultures were fixed and permeabilized for 10 min in ice-cold methanol. Coverslips were washed with PBS and blocked in PBS containing 5% BSA for 1 h. Bacteria were stained with a rabbit anti-L. pneumophila polyclonal Ab (1:2000) (25), followed by Alexa 488-conjugated goat anti-rabbit secondary Ab (1:300 dilution; Invitrogen). The coverslips were mounted onto Prolong Gold antifade reagent (Invitrogen) on glass slides for fluorescence microscopy. Images were acquired with a Leica TCS SP2 SE inverted microscope using a ×40, 1.25 numerical aperture oil objective (Leica HCX PL APO) and analyzed using Leica Advanced Fluorescence software. At least 100 infected cells were scored in each coverslip.
Mice and in vivo infections
The mouse protocols were previously approved by the institutional ethics committee for animal care and research (comissão de ética em experimentação animal). Mice were bred and maintained at the Animal Facility of the University of São Paulo at Ribeirão Preto. C57BL/6 stock mice were from the institutional animal facility. Caspase-1−/− and NLRC4−/− mice were previously described (39, 40) and back-crossed to C57BL/6 mice for eight generations. For in vivo infections, male or female mice from 10 to 14 wk of age were anesthetized with 2.5% solution of 2,2,2-tribromoethanol (Sigma) by i.p. administration followed by intranasal inoculation with 40 μl PBS containing 2.5 × 106 bacteria/ml. Mice were euthanized at the indicated time points for the determination of CFU per lung. The lungs were harvested and homogenized in 5 ml sterile water for 30 s using a tissue homogenizer (Fisher Scientific). Lung homogenates were diluted in water, plated on BCYE agar plates, and incubated for 4 d at 37°C for CFU determination.
For cytokine determination, BMDMs were seeded at 2.0 × 105 cells/well in 24-well plates and infected with WT L. pneumophila or flaA mutants at an MOI of 10 for 12 h. For activation of the NLRP3 inflammasome, cultures were pretreated with 1 μg/ml LPS (Sigma) for 4 h and stimulated with 20 μM nigericin (Sigma) for 1 h. The cytokine in the supernatant was measured with a mouse IL-1β, IL-12p40, and TNF-α ELISA kit (OptEIA; BD Biosciences, Franklin Lakes, NJ) according to the manufacturer’s instructions.
Statistical analyses were performed using Prism 5.0 software (GraphPad, San Diego, CA). Unpaired Student t test was used to compare two groups. One-way ANOVA followed by multiple comparisons according to Tukey's procedure was used to compare three or more groups. Unless otherwise stated, data are expressed as the mean ± SD. Differences were considered statistically significant at a p value <0.05.
Construction of flaA and fliI mutants derived from L. pneumophila JR32 strain
Flagellin-deficient L. pneumophila (flaA mutant) is able freely to multiply in restrictive C57BL/6 BMDMs (20, 22, 36). However, it was previously demonstrated that compared with wild-type bacteria, the flaA mutants bind to macrophages less effectively due to the lack of motility (41). Furthermore, bacterial motility was required for the efficient evasion of phagosome/lysosome formation in macrophages (42). Thus, to address appropriately the role of flagellin for NLRC4-dependent restriction of L. pneumophila multiplication in macrophages and in vivo, we aimed to construct non-motile bacteria that express flagellin (fliI). The L. pneumophila fliI gene has previously been characterized and encodes an ATPase required for the secretion of flagellin to assemble the flagella (43). Therefore, although the fliI mutants were unable to export monomeric flagellin to assemble the flagellum, they could still activate macrophage innate immunity (Fig. 1A) (22, 36, 43). Thus, we constructed deletion mutants of fliI and flaA genes in the virulent L. pneumophila JR32 strain (WT L. pneumophila), which can be used for in vivo studies (32). By Western blotting, we confirmed that the flaA mutants lack flagellin expression. In contrast, when grown in solid BCYE plates, both WT L. pneumophila and fliI express intracellular flagellin from days 1 to 5 (Fig. 1B). To confirm further that fliI mutants fail to express extracellular flagellin, bacteria were grown for 24, 48, 72, 96, and 120 h in BCYE plates followed by extraction of cell-associated extracellular proteins as previously described (25). Proteins were separated by SDS-PAGE, blotted, and probed with anti-flagellin. We confirmed that whereas WT L. pneumophila express extracellular flagellin after the second day of culture in BCYE plates, the fliI and flaA mutants are deficient in expression of extracellular flagellin (Fig. 1C).
Bacteria deficient for flagellin or motility colonize A. castellanii less effectively
Next, we used bacteria deficient for flaA and fliI to investigate the requirement of flagellin and motility for L. pneumophila colonization and multiplication in the unicellular protozoa A. castellanii, which is one of the natural hosts for L. pneumophila multiplication in freshwater environments (2). Initially, we used WT L. pneumophila, fliI, and flaA to infect A. castellanii in 24-well plates and found that all bacteria effectively multiply intracellularly (Fig. 2A). To evaluate further the role of flagellum and motility for the colonization of protozoan cells, we performed competition assays in 175 cm2 tissue culture flasks containing A. castellanii, using a streptomycin sensitive WT strain of L. pneumophila (F2111) (33). We found that the two selected WT strains of L. pneumophila colonized the Acanthamoeba cells equally (48.6% F2111; 51.4% JR32) after 5 d of coculture (Fig. 2B). In contrast, flaA or fliI were unable to compete with flagellated bacteria, as we recovered a higher proportion of WT strain after 5 d of coculture (95.2% WT L. pneumophila, 4.8% flaA; 86.6% WT L. pneumophila, 13.4% fliI) (Fig. 2B). In this experiment, we also used bacteria deficient in dotA, which failed to compete with WT L. pneumophila as expected (Fig. 2B). These studies indicate that L. pneumophila flagellum and motility are important for the colonization of A. castellanii; therefore, they may contribute to the bacterial prevalence in natural freshwater environments.
Restriction of L. pneumophila multiplication in vivo is dependent on flagellin but not flagellum and motility
To assess the importance of flagellin and motility for the pathogenesis and innate immune restriction of L. pneumophila infection in mammalian hosts, we infected C57BL/6 mice with WT L. pneumophila, fliI, or flaA and evaluated bacterial CFU in the lungs. Whereas the WT L. pneumophila and the fliI mutants were readily restricted from the lungs of C57BL/6 mice, we found that the flaA mutants effectively multiplied and showed a maximum bacterial multiplication after 48 h of infection (Fig. 3). These data indicate that the immune restriction of L. pneumophila multiplication in mouse lungs is dependent on bacterial flagellin but not motility.
Flagellin-induced restriction of L. pneumophila multiplication is solely dependent on NLRC4
Next, we infected NLRC4−/− mice with WT L. pneumophila, fliI, or flaA bacteria to assess further whether the flagellin-dependent restriction of L. pneumophila replication is fully dependent on NLRC4 signaling. We found that WT L. pneumophila, fliI, and flaA were similarly able to multiply in the lungs of NLRC4−/− mice (Fig. 4A, 4B). These data support a major role for NLRC4 in the flagellin-dependent restriction of L. pneumophila multiplication and indicate that within the infection conditions used, L. pneumophila motility plays little part in the pathogenicity in a mouse model of Legionnaires disease. Notably, it was recently demonstrated that NLRC4 is able to recognize the Salmonella basal body rod component of the T3SS apparatus independent of flagellin (31). Thus, we simultaneously infected C57BL/6 and NLRC4-deficient mice with WT L. pneumophila and flaA to investigate if the NLRC4-dependent restriction of L. pneumophila multiplication required flagellin expression. As shown in Fig. 4C, regardless of the time postinfection, the replication of flaA was similar in the lungs of C57BL/6 and NLRC4−/− mice. These data indicate that within the assay’s sensitivity limits, the NLRC4-dependent restriction of L. pneumophila multiplication is fully dependent on flagellin. Moreover, in NLRC4−/− mice, WT L. pneumophila multiplied similarly to flaA mutants (Fig. 4C), thus confirming that TLR5 plays a minor role in restriction of L. pneumophila replication in murine models of infection (44).
NLRC4-dependent flagellin recognition triggers effector functions not exclusively dependent on caspase-1
Our data indicated that NLRC4-dependent restriction of L. pneumophila infection was fully dependent on flagellin. Because NLRC4 is known to trigger caspase-1 activation, which is critical for restricting bacterial multiplication, we investigated whether the NLRC4-mediated growth restriction was exclusively dependent on caspase-1. To test this, we initially infected caspase-1−/− mice with WT L. pneumophila, fliI, or flaA bacteria and measured bacterial growth in lungs of mice. We found that whereas all bacterial strains were able to multiply in the lungs of caspase-1−/− mice, the flaA bacteria multiplied slightly better than WT L. pneumophila or fliI (Fig. 5A, 5B). These results suggest that whereas NLRC4 is fully required for the flagellin-dependent restriction of L. pneumophila multiplication (Fig. 4), caspase-1 only partially accounted for the growth restriction in response to flagellin. Next, we performed infections in BMDMs to investigate if in isolated BMDMs, the flagellin-dependent growth restriction is only partially dependent on caspase-1 as observed in vivo. We found that flaA mutants multiplied significantly better than WT L. pneumophila in both C57BL/6 BMDMs (Fig. 6A) and BMDMs obtained from caspase-1−/− mice (Fig. 6B). A similar experiment performed with BMDMs generated from NLRC4−/− mouse indicates that the flagellin-dependent growth restriction in BMDMs is fully dependent on NLRC4 (Fig. 6C), which supports the in vivo results with NLRC4−/− mice (Fig. 4). Of note, the increased replication of flaA compared with WT L. pneumophila in caspase-1−/− BMDMs was not influenced by the infection dose because similar results were obtained with low (0.015) or high (10:1) MOI (Supplemental Fig. 1). To validate further the biological functions of the BMDMs used, we treated the cultures with LPS in combination with nigericin, which is known to trigger activation of the NLRP3 inflammasome (45). As expected, BMDMs from C57BL/6 and NLRC4−/− mice, but not BMDMs from caspase-1−/− mice, produced active IL-1β in response to LPS plus nigericin (Fig. 6D). As originally demonstrated (23), when BMDMs were infected with WT L. pneumophila, only C57BL/6 cells triggered a robust production of IL-1β; in contrast, IL-1β secretion was markedly reduced in NLRC4−/− and absent in caspase-1−/− BMDMs (Fig. 6D). Accordingly, as compared with WT L. pneumophila infections, the secretion of active IL-1β was significantly reduced when C57BL/6 BMDMs were infected with flaA and similar when NLRC4−/− BMDMs were infected with WT L. pneumophila or flaA (Fig. 6D). Nevertheless, C57BL/6, NLRC4−/−, and caspase-1−/− BMDMs were fully able to express IL-12p40 and TNF-α in response to WT L. pneumophila or flaA mutants (Fig. 6D). Collectively, the data presented in this study indicate that flagellin-dependent L. pneumophila restriction is fully dependent on NLRC4 but it is only partially dependent on caspase-1.
We reasoned that if NLRC4 is able to trigger caspase-1–independent responses for the restriction of L. pneumophila multiplication, then NLRC4−/− should support a higher bacterial replication compared with caspase-1−/−. Thus, we infected C57BL/6, caspase-1−/−, and NLRC4−/− mice with WT L. pneumophila and found a statistically significant increase in bacterial multiplication in the lungs of NLRC4−/− mice compared with that in caspase-1−/− mice after 72 h infection (Fig. 7A). We also found a tendency for increased replication in NLRC4−/− compared with that in caspase-1−/− at 48 and 96 h postinfection. However, the differences were not statistically significant, possibly due to experimental variability inherent to in vivo experiments. Nevertheless, these data support our findings that NLRC4 uses additional signaling pathways that are independent of caspase-1. To investigate further whether the NLRC4-dependent, caspase-1–independent mechanisms to restrict WT L. pneumophila infection operate in isolated BMDMs, we performed experiments using BMDMs obtained from these mice. We found that NLRC4-deficient BMDMs supported a higher bacterial multiplication than that of caspase-1−/− BMDMs (Fig. 7B). In contrast, when we used flaA mutants to infect BMDMs, we found that flaA mutants multiplied similarly in C57BL/6, caspase1−/−, and NLRC4−/− (Fig. 7C). These results support our findings that flagellin-positive L. pneumophila trigger multiple NLRC4 responses in the restriction of bacterial multiplication. Although caspase-1 clearly participates in NLRC4-dependent growth restriction, we also detected caspase-1–independent responses downstream of NLRC4.
We have previously demonstrated that some species of non-pneumophila Legionella express flagellin and trigger activation of the NLRC4 inflammasome (25). Moreover, activation of the NLRC4 inflammasome contributes to the pulmonary clearance of flagellated Legionella spp. such as L. micdadei, L. bozemanii, L. gratiana, and L. rubrilucens (46). To address further whether NLRC4 triggers caspase-1–independent responses for the clearance of these species, we infected C57BL/6, NLRC4−/−, and caspase-1−/− mice and measured the CFU in the lungs of mice infected for 4 and 72 h. We initially confirmed the NLRC4 inflammasome participates in the pulmonary clearance by these species as we detected higher bacterial loads in the lungs of NLRC4−/− and caspase-1−/− mice compared with those in C57BL/6 mice (Fig. 8). Importantly, bacterial loads detected in the lungs of NLRC4−/− mice were higher than those found in the lungs of caspase-1−/− mice (Fig. 8). These data indicate that whereas both NLRC4 and caspase-1 participate in the restriction of the infection by L. micdadei, L. bozemanii, L. gratiana, and L. rubrilucens, NLRC4 may use caspase-1–independent responses for restriction of bacterial replication.
The NLRC4-dependent response that leads to the restriction of L. pneumophila multiplication operates early after macrophage infection
To investigate further the mechanisms by which macrophages restrict the multiplication of flagellated bacteria, we measured bacterial replication early after bacterial infection. BMDMs were infected with WT L. pneumophila or flaA bacteria. The bacterial multiplication within L. pneumophila-containing vacuoles (LCVs) were evaluated after 10 h of macrophage infection. We found that in C57BL/6 BMDMs, most of the cells infected with WT L. pneumophila contained a few replicating bacteria per LCV (Fig. 9A). In contrast, a higher proportion of LCVs found in caspase-1−/− BMDMs contained 2–3 or 4–10 bacteria per vacuole (Fig. 9B). In NLRC4−/− BMDMs, we found that a higher proportion of LCVs contained more than 10 bacteria per vacuole (Fig. 9C). Representative images of the cultures infected with WT L. pneumophila are shown in Fig. 9D–9F. These data suggest that both NLRC4 and caspase-1 contribute to an early restriction of L. pneumophila multiplication in macrophages. However, macrophages lacking NLRC4 are more permissive for bacterial replication compared with those deficient in caspase-1. This result supports our finding that NLRC4 triggers the restriction of L. pneumophila replication by mechanisms dependent and independent of caspase-1. To evaluate further whether the early restriction of bacterial replication by macrophages required flagellin, we also used flaA bacteria in the experiments. As shown in Fig. 9G–9I, most of the LCVs harboring flaA mutants contained more than 10 bacteria, independent of the host macrophages used. To investigate if the flagellin-dependent growth restriction observed in caspase-1−/− BMDMs was influenced by the infection dose, we performed an experiment in which we infected caspase-1−/− BMDMs with WT L. pneumophila and flaA at different MOI and measured the bacterial replication within the LCVs. We found that regardless of the MOI used, the flaA mutants more effectively replicated in the recently established LCVs compared with WT L. pneumophila (Supplemental Fig. 2).
Collectively, our studies indicate that flagellin recognition by NLRC4 triggers caspase-1–dependent and –independent responses in the restriction of bacterial replication. Furthermore, these responses operate early after macrophage infection and comprise the macrophage autonomous mechanisms for bacterial growth restriction.
Engagement of the NLRC4 inflammasome is critical for the restriction of L. pneumophila replication in mouse macrophages. In this study, we demonstrated that NLRC4 triggers caspase-1–dependent and –independent responses in the restriction of L. pneumophila infection in vivo and in macrophages. The NLRC4-dependent and caspase-1–dependent response, known as NLRC4 inflammasome, has been extensively characterized and accounts for pore formation, pyroptosis, and modulation of phagosome maturation as well as restriction of L. pneumophila replication in macrophages (20, 21, 23–25, 47–50). Besides the NLRC4 inflammasome, we demonstrated that NLRC4 also uses caspase-1–independent signaling in the restriction of L. pneumophila replication. This previously unappreciated response operates early after macrophage infection and represents a cell-autonomous mechanism for the restriction of bacterial replication. Although the mechanism by which this novel pathway leads to infection control remains unclear, our data suggest that this pathway operates in macrophages and requires flagellin; thus, one can speculate about the participation of the NAIP5 protein in this pathway. Although it is clear that NAIP5 contributes to caspase-1 activation in response to L. pneumophila infection of macrophages (22–24, 50, 51), caspase-1–independent NAIP5 responses has also been suggested (48, 52, 53). Thus, it is possible that NAIP5 participates in this caspase-1–independent response downstream of NLRC4 reported in the current study. Regardless of the possible participation of NAIP5, we describe in this study a novel caspase-1–independent, NLRC4-dependent response in the restriction of the infection by flagellated species of Legionella such as L. pneumophila, L. micdadei, L. bozemanii, L. gratiana, and L. rubrilucens. This previously unappreciated pathway requires bacterial flagellin, operates early after macrophage infection, and has a direct influence on the pulmonary infection of mice in vivo.
It is well established that the NLRC4 inflammasome is activated in response to flagellin from different bacterial species (20, 24–26, 28, 50, 54). However, it was recently demonstrated that rod proteins from the basal body of the T3SS apparatus are able to engage NLRC4 independent of flagellin (31). We tested if L. pneumophila were able to activate NLRC4 independent of flagellin. By using isogenic mutants for fliI, which express flagellin but are non-motile, we demonstrated that NLRC4-dependent restriction of L. pneumophila multiplication is fully dependent on flagellin and not influenced by bacterial motility. Thus, it is possible that the flagellin-independent NLRC4 activation previously reported is exclusive to T3SS-expressing bacteria that express the rod proteins from the basal body of that secretion apparatus (31).
It was previously shown that L. pneumophila mutants for flaA bind to macrophages less effectively due to their lack of motility (41). Moreover, bacterial motility was required for the efficient evasion of phagosome/lysosome formation in macrophages (42). These studies indicated that the flagellum-mediated motility of L. pneumophila was important for pathogenesis, and they thus provide an explanation for why L. pneumophila express genes, such as flagellin, that account for the restriction of bacterial replication in mammalian macrophages. In this study, we provide an alternative, but not exclusive, explanation; by performing competition experiments, we were able to demonstrate that mutants defective for flagellum and motility were defective for colonization of A. castellanii. Because the natural habitats for L. pneumophila are freshwater environments, it is feasible that flagellum expression facilitates the colonization of the protozoan hosts in the freshwater environments. Thus, despite the detrimental role of flagellin during mammalian host infection, the natural environments may impose a selective pressure for expression of L. pneumophila flagellum genes. This may explain the fact that the majority of L. pneumophila clinical isolates are flagellin-positive (L.M. Massis and D.S. Zamboni, unpublished observations). Notably, the closely related, soil-living pathogen Legionella longbeachae lacks genes for flagellum biosynthesis and may thus avoid recognition by mammalian pattern recognition receptors such as NLRC4 (25, 51, 55). This feature may account for why L. longbeachae is highly and often lethal in mouse models of infection (46, 56).
Regardless of the selective pressure imposed for the maintenance of L. pneumophila flagellum during its natural history, our studies demonstrate that expression of flagellin allows the immune system to recognize L. pneumophila via intracellular pattern recognition receptors such as NLRC4. The engagement of NLRC4 triggers at least two distinct pathways for the restriction of bacterial replication: a well-described caspase-1–dependent response called NLRC4 inflammasome and a previously unappreciated response that is independent of caspase-1. Together, these responses effectively account for the bacterial growth restriction in macrophages and influence the outcome of pulmonary infection in a murine model of Legionnaires disease.
We are grateful to Gabriel G. Marques and Maira Nakamura for technical assistance.
This work was supported by PEW, Instituto Nacional de Ciência e Tecnologia de Vacinas/Conselho Nacional de Desenvolvimento Científico e Tecnológico (INCTV/CNPq), and Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Grants 06/52867-4 and 10/50959-4 (to D.S.Z.). C.V.H. is the recipient of a master fellowship from CNPq. M.S.F.P. (07/55852-0), G.F.M. (07/51133-0), L.M.M. (08/56725-5), and J.I.H. (07/53076-3) are recipients of fellowships from FAPESP. D.S.Z. is a research fellow from CNPq.
The online version of this article contains supplemental material.
Abbreviations used in this article:
buffered charcoal–yeast extract
bone marrow-derived macrophage
L. pneumophila-containing vacuole
multiplicity of infection
nucleotide-binding oligomerization domain
type III secretion system
The authors have no financial conflicts of interest.