The Stress-Induced Virulence Protein InlH Controls Interleukin-6 Production during Murine Listeriosis

The Listeria strains used in this study are listed in Table 1 and were routinely grown in brain heart infusion (BHI) medium (Difco) at 37°C. Erythromycin (5 μg/ml) or chloramphenicol (7 μg/ml) was added for growth of strains carrying plasmids. For experiments involving growth at different phases, bacteria were first cultured at 37°C overnight in BHI medium and then diluted to obtain an optical density at 600 nm (OD₆₀₀) of 0.01 and grown at 37°C to an OD₆₀₀ of 0.5, 0.8, 1.0, 2.5, or 3.0. For experiments involving growth under various stress conditions, overnight cultures were diluted to obtain an OD₆₀₀ of 0.01 and grown at 37°C in BHI medium to an OD₆₀₀ of 0.2, and then they were pelleted and resuspended either in BHI medium at pH 5.5 after addition of HCl (acid stress), in BHI medium supplemented with 10 mM menadione sodium bisulfite (Sigma) (oxidative stress), or in BHI medium at 37°C or 45°C (for heat shock) and incubated for 1 h. In all cases bacteria were grown with shaking. The bacterial growth conditions used in tiling array experiments have been described previously (61). The eukaryotic cell lines used were human JEG-3 (ATCC HTB-36), Caco-2 (ATCC HTB-37), LoVo (CCL-229), HepG2 (ATCC HB-8065), and THP-1 (ATCC TIB-202) cells and RAW 264.7 murine macrophages (ATCC TIB-71). Cell lines were grown according to ATCC instructions. Peritoneal exudate macrophages (PEM) were prepared as previously described (1). Briefly, BALB/c mice were injected with 10% peptone, and after 48 h peritoneal exudate cells were harvested by peritoneal lavage.

Polyclonal rabbit InlH antibody (R138) was generated and affinity purified as described previously (20). Rabbits were immunized against recombinant glutathione S-transferase (GST)-tagged InlH protein (residues 36 to 518; kindly provided by W. D. Schubert [58]). We also used antibodies against InlA (47), InlB (9), SvpA (6), and EF-Tu (4). InlH at the bacterial surface was detected by performing immunofluorescence experiments as described previously (6) with coverslips incubated with the InlH antibody (1:10,000 dilution) in 5% bovine serum albumin in phosphate-buffered saline (PBS) for 1 h. After several washes, the coverslips were incubated with Alexa Fluor 488-labeled goat anti-rabbit secondary antibodies (1:400 dilution; Molecular Probes) for 1 h and mounted with 10 μl of Fluoromount G (Interchim). Total bacterial cells and cell wall extracts were prepared as described previously (30). Proteins were boiled in Laemmli sample buffer, resolved by 8% or 10% SDS-PAGE, and transferred to Hybond-C Extra transfer membranes (GE Healthcare). Bound antibodies were revealed using horseradish peroxidase (HRP)-conjugated goat anti-mouse or anti-rabbit (AbCys, France) antibodies and the ECL Plus blot detection reagent (GE Healthcare).

Two ∼900-bp fragments flanking the inlH gene were PCR amplified from EGD-e chromosomal DNA with primers inside the inlGHE locus. The primers used for the inlH 5′ flanking fragment were h1 (5′-CGCGGATCCAAATATCACTTGGAACTTAG-3′) and h2 (5′-AAAAGGCCTCATAATCCCTCTCCTTTTAT-3′), and the primers used for the 3′ fragment were h3 (5′-AAAAGGCCTTAACAACAAAAAAAGCTGAG-3′) and h4 (5′-CCGGAATTCTGTGAGATTAGATAAACCAG-3′). After restriction of the amplified 5′ and 3′ fragments with BamHI and StuI or StuI and EcoRI, 5′ and 3′ fragments were coligated in the thermosensitive plasmid pMAD digested by BamHI and EcoRI, yielding the pMAD-ΔinlH plasmid, which was verified by sequencing. This plasmid was prepared in Escherichia coli (BUG 2585) and electroporated into L. monocytogenes strain EGD-e. Independent colonies were used for allelic exchange in L. monocytogenes wild-type strain EGD-e, which was performed as described previously (56), generating independent ΔinlH isogenic deletion mutants (Table 1). Deletion of the entire inlH gene was confirmed by PCR amplification and sequencing using primers h10 (5′-TGCTTATGGTGAAGTAGTACC-3′) and h11 (5′-GCGGACTCACATCACTAATT-3′). To generate L. innocua expressing inlH, EGD-e chromosomal DNA was used to PCR amplify inlH (from the ribosome binding site to the stop codon) using primers h12 (5′-ACGAGCTCTAAAAGGAGAGGGATTATGAAAAAACG-3′) and h13 (5′-TCCCCCGGGCAGCTTTTTTTGTTGTTATTTTAC-3′). The DNA fragment was digested with SacI and SmaI and inserted into the SacI and SphI sites in the shuttle plasmid pP1 (53) downstream from the P_prot promoter, yielding pP_prot-inlH. This plasmid was prepared in E. coli (BUG2845), verified by sequencing, and then electroporated into L. innocua.

Total RNA from bacterial pellets was extracted using TRIzol reagent (Invitrogen) as described previously (61), and contaminating DNA was degraded using Turbo DNase according to the manufacturer's instructions (Ambion Inc., Austin, TX). RNA concentrations and quality were determined using an Experion automated electrophoresis station (Bio-Rad). cDNA synthesis was performed with an iScript cDNA synthesis kit (Bio-Rad) using 1 μg total RNA. For quantitative real-time PCR (qRT-PCR), the following primers were selected using the Integrated DNA Technologies Designer software and used: for inlH, 5′-TATACGTTCACAGGTTGGTATGA-3′ and 5′-TTCCTGTTGGGGTATTGGTAGTT-3′; for inlG, 5′-TCATGGGATGAAAGCTCAGGCAGA-3′ and 5′-TTCCAAGTTCCGTCTTCACCTCGT-3′; for inlE, 5′-TGTGAGCGCGCTTGCTGGATTAAA-3′ and 5′-TTGGAAAGTCCTGCAAGTGGTGTC-3′; for inlA, 5′-ACACGAGTAACGGGACAAATGCTC-3′ and 5′-ATCCGCCTGAAGCGTTGTAACTTG-3′; and for gyrA, 5′-GCGATGAGTGTAATTGTTG-3′ and 5′-ATCAGAAGTCATACCTAAGTC-3′. qRT-PCRs were performed using 25-μl mixtures containing 5 μl of a 1/10 dilution of cDNA, 0.75 μl of gene-specific primers (10 μM each), and 12.5 μl of iQ SYBR green supermix (Bio-Rad, Hercules, CA) with the MyiQ single-color RT iCycler PCR detection system and the MyiQ optical system software (Bio-Rad, Hercules, CA). The critical cycle threshold (C_T) was defined for each sample. The levels of expression of the genes tested were normalized using the gyrA genes of L. monocytogenes as an internal standard. The fold change was calculated using “the gene expression C_T difference” formula (57). Each assay was performed in triplicate and repeated with three independent RNA samples. Design of the tiling microarray, hybridization, and analysis have been described previously (61).

The relative rate of invasion of L. monocytogenes or L. innocua strains in mammalian cells was determined by a gentamicin assay. The Listeria strains were grown in BHI medium to an OD₆₀₀ of 2.5, washed in PBS, and diluted in Dulbecco modified Eagle medium (DMEM) so that the multiplicity of infection (MOI) was about 50 bacteria per cell. Bacterial suspensions were added to cells for 1 h, the cells were washed, and then incubation with 20 μg/ml gentamicin for 1 h was used to kill the extracellular bacteria. After washing, cells were lysed in 0.2% Triton X-100, and the number of viable bacteria released from cells was assessed by plating samples on BHI agar plates. For assays of survival in macrophages, isolated PEM or THP-1 monocytes were activated or not activated with phorbol myristate acetate (PMA) (0.5 μM) 24 h prior to infection. Macrophages were infected for 1 h with Listeria strains at an MOI of ∼10 bacteria per cell, 20 μg/ml gentamicin was added, and the number of viable bacteria was determined 12 h or 24 h later as described above. The level of entry or survival was determined by comparison with the number of bacteria present in the inoculum. Experiments were done in triplicate.

All animals were treated according to the Institut Pasteur guidelines for laboratory animal husbandry. Four animals were used for each experiment. Statistical analyses were performed using the Student t test. P values of <0.05 were considered statistically significant. For intravenous inoculation, 8-week-old BALB/c or C57BL6 female mice (Charles River) were injected intravenously with a sublethal bacterial inoculum containing 5 × 10⁴ CFU, and 48 h or 72 h after infection, livers and spleens were dissected and the numbers of CFU were determined by plating serial dilutions of organ homogenates on BHI agar. Heparin (250 U/ml) was added to blood samples, and 100-μl portions were plated on BHI agar. For oral inoculation, BALB/c or iFABP-human E-cadherin (hEcad) transgenic mice were inoculated intragastrically, as previously described (40), with a bacterial inoculum containing 10⁹ CFU. Infected intestinal tissue was dissected 24 h or 48 h postinoculation. Intestinal segments were rinsed in DMEM and incubated for 2 h in DMEM containing 100 μg/ml gentamicin to kill extracellular bacteria before homogenization. Numbers of CFU were determined after serial dilutions of organ homogenates were plated on BHI agar plates and incubated at 37°C for 48 h.

BALB/c mice were inoculated intravenously, as described above, with either the EGD1 or ΔinlH1 strain (4 mice per group) or the EGD-e or ΔinlH strain (6 mice per group). Tissues were homogenized gently using a MACS dissociator (Miltenyi Biotec) at 48 h postinfection. Samples were centrifuged at 1,000 × g for 15 min, and cytokine production was determined using supernatants, a mouse cytokine 20-plex kit (Biosource), and a Luminex-100 (Luminex), as recommended by the manufacturers. This assay allowed determination of the titers of several cytokines, including gamma interferon (IFN-γ), IL-1α, IL-1β, IL-6, IL-10, IL-12, and tumor necrosis factor alpha (TNF-α). Production of IL-6, IL-10, or TNF-α was also determined by an enzyme-linked immunosorbent assay (ELISA) by following the manufacturer's instructions (DuoSet kits; R&D Systems). Macrophages were infected with Listeria strains as described above. RAW 264.7 macrophages were infected for 16 h in the presence or absence of lipopolysaccharide (LPS) (1 μg/ml). PEM or THP-1 cells were activated with PMA (0.5 μM) for 24 h prior to infection and infected with Listeria for 12 h or 24 h. Secreted IL-6 was quantified by an ELISA. Data were analyzed using the Student t test. A P value of <0.05 was used as the threshold for significance.

Based on results obtained with green fluorescent protein (GFP) transcriptional fusions, it was previously proposed that the inlGHE cluster is organized as an operon with a polycistronic transcript expressed from the inlG promoter (53). However, both inlH and inlE are framed by a σ^B promoter sequence and a rho-independent terminator. This suggests that they are transcribed independently. Recent studies in our laboratory using tiling arrays made it possible to determine the transcriptional regulation of the entire L. monocytogenes EGD-e genome under diverse growth conditions (61). Our tiling arrays are powerful tools for mapping transcription start sites, as they cover both strands of the genome with 25-mer probes every 16 bp. As shown in Fig. 1A, transcription of inlG, inlH, and inlE in EGD-e and an isogenic ΔsigB strain grown to exponential phase in BHI medium (reference conditions) was compared to transcription of these genes under three other growth conditions, growth to stationary phase in BHI medium, 60 min of growth in human blood ex vivo, and 24 h of growth in the gut of mice. The inlH promoter led to a monocistronic transcript whose levels were substantially decreased in the ΔsigB mutant and significantly increased in bacteria grown to stationary phase in BHI medium. Importantly, inlH was upregulated in a σ^B-dependent manner in bacteria in the intestinal lumen of orally infected mice. In contrast, the levels of inlG and inlE transcripts remained very low under all of these conditions. In addition, the amount of the inlH transcript in a prfA mutant (data not shown) or in bacteria grown in human blood was not different from that in bacteria grown in BHI medium (Fig. 1A). By performing qRT-PCR (Fig. 1B) and immunoblotting of bacterial extracts using a specific InlH antibody (Fig. 1C), we confirmed that there was specific σ^B-dependent induction of inlH expression in stationary phase, while inlG and inlE expression remained very low (Fig. 1A and 1B). Together, these data indicate that the inlGHE gene cluster is not a polycistronic operon and that only one of these three inl genes, inlH, is substantially induced by stress stimuli in vitro and in vivo.

We next addressed whether the InlH protein was efficiently produced, targeted, and anchored to the bacterial surface in response to specific stresses known to be controlled by σ^B, such as the stationary phase, heat, acid, or oxidative stress. As expected, InlH was more abundant on the surface of bacteria in stationary phase than in exponential phase, while it was absent from the surface of ΔinlH or ΔsigB isogenic strains, as well as the surface of ΔsrtA bacteria, from which it was released into the supernatant (Fig. 1D). An increased amount of InlH at the bacterial surface was also detected when EGD-e bacteria were grown at pH 5.5, shifted from 37°C to 45°C (Fig. 1D), or grown in 7.5% NaCl (data not shown). Interestingly, the strongest induction of InlH production was observed when bacteria were grown in the presence of menadione, a quinone that produces superoxide radicals and thus subjects the cells to oxidative stress (Fig. 1D). Together, these results demonstrate that InlH is a stress-induced surface protein anchored by sortase A.

We confirmed by performing an immunoblot analysis of different bacterial cell fractions that InlH, like InlA, is present mainly in the cell wall compartment (Fig. 2A). InlB, which is present in protoplasts, and EF-Tu, which is an abundant cytosolic protein that is also found in the cell wall fraction (4, 52), were used as controls. Strikingly, the amounts of InlA in the cell wall and supernatant of the ΔinlH mutant grown until stationary phase were larger than the amounts in the parental strain. However, increased InlA production did not result from increased transcription, as the amounts of the inlA transcript were similar in wild-type and ΔinlH bacteria, as shown by qRT-PCR (Fig. 2B). Disruption of inlH in three other ΔinlH mutants obtained in independent mutagenesis experiments also led to increases in the amounts of InlA in the cell wall (Fig. 2C) and supernatant (data not shown).

Since L. monocytogenes strains carry either inlH or its homolog inlC2, we examined InlA production in other wild-type strains and isogenic ΔinlH or ΔinlC2 strains. Cell wall fractions were extracted from EGD-e and two other serotype 1/2a strains grown to stationary phase, BUG 2204 (referred here to as “EGD-1” [Table 1]), in which the inlGHE cluster was originally identified and which was used to study the role of inlH in virulence (53, 58), and BUG 600 (referred here to as “EGD-2” [Table 1]), which contains inlC2 instead of inlH (19). Loads were controlled by the amounts of SvpA, another peptidoglycan-binding protein that is abundant in Listeria cell walls (6, 52) (Fig. 2D), or with EF-Tu (data not shown). InlC2 was efficiently detected in the cell wall of strain EGD-2 and was absent in the ΔinlC2 mutant, confirming that the InlH antibody also recognizes InlC2. InlA was more abundant in the cell wall fraction of the ΔinlC2 strain, as well as in the EGD-e ΔinlH strain (Fig. 2D). In contrast, the basal levels of InlH and InlA were very low in the cell wall of EGD-1 bacteria, and the InlA level was not increased in the cell wall of the isogenic ΔinlH-1 bacteria.

Together, these results suggest that in strains in which inlA is expressed well, inlH induction under stress conditions may limit inlA expression at the posttranscriptional level.

To search for a function of InlH in the infectious process, we first examined the ability of ΔinlH bacteria to enter mammalian cells. We used cell lines in which Listeria entry occurs through InlA (Caco-2 cells), InlB (HepG2 cells), or both InlA and InlB (JEG-3 and Lovo cells) pathways or independent of InlA and InlB (RAW 264.7 macrophages). The entry of ΔinlH bacteria after 2 h of infection was similar to that of the wild-type strain for all cell lines (Fig. 3A) except Caco-2 cells, for which entry was 2-fold greater (Fig. 3B). The increased amount of InlA in ΔinlH bacteria described above probably accounts for the more efficient invasion of Caco-2 cells, in which entry occurs mainly by the InlA/E-cadherin pathway. Indeed, heterologous expression of InlH in the nonadherent and noninvasive organism Listeria innocua did not increase adherence or entry, confirming that InlH plays no role in these processes (Fig. 3B and data not shown). In addition, intracellular replication, actin-based motility, and cell-to-cell spread were not affected in cultured ΔinlH mutant cells (data not shown).

Since surface levels of InlH are significantly increased by acid stress and oxidative stress, two conditions occurring in activated macrophages, we also investigated whether InlH contributes to L. monocytogenes survival in murine peritoneal exudate macrophages (PEM). Compared to nonactivated (NA) PEM, PMA-activated cells were, as expected, more phagocytic following infection with wild-type or inlH-deficient Listeria (Fig. 3C). However, InlH did not influence either the survival or the proliferation of bacteria in resting or PMA-activated PEM. The same results were obtained with human THP-1 macrophages (data not shown). Thus, InlH is neither an adhesin nor an invasin and does not promote bacterial intracellular survival in macrophages.

A role for InlH in L. monocytogenes murine listeriosis was demonstrated in a previous study using the EGD-1 and ΔinlH-1 strains (Table 1) (58). However, because we found that the level of inlH expression in strain EGD-e was higher than that in strain EGD-1 in vitro (Fig. 2D), we examined whether inactivation of inlH in this genetic background could have a more pronounced effect on bacterial colonization of organs. As expected, the numbers of bacteria recovered from the blood, livers, and spleens of BALB/c mice 72 h following intravenous (i.v.) inoculation of the EGD-e ΔinlH strain were consistently lower than the numbers of bacteria recovered when the parental EGD-e strain was used. In contrast, there was greater variability in the number of surviving EGD-1 and ΔinlH-1 bacteria than in the number of surviving EGD-e and ΔinlH bacteria (Fig. 4A).

We next examined the virulence of the EGD-e ΔinlH mutant in C57BL/6 mice, a mouse background known to be more resistant to L. monocytogenes infection. In particular, these mice are altered in the IFN-β response to infection, which plays an important role in host defense against L. monocytogenes (22). Inactivation of inlH also resulted in a significant L. monocytogenes multiplication defect in the blood, livers, and spleens of C57BL/6 mice (Fig. 4B). Together, these results confirm that InlH is a factor that is involved in systemic L. monocytogenes infection.

A previous report showed that there were significant reductions in the counts of viable bacteria in livers and spleens for a inlGHE mutant compared to the wild-type strain at 24 h postinfection following oral inoculation of mice (53). Since expression of inlH is specifically regulated by σ^B and is induced in the intestinal lumen, we looked for a possible role for InlH in bacterial invasion of the gastrointestinal tissue, which is the early stage of listeriosis. The first experiments were performed with BALB/c mice, in which the InlA/E-cadherin pathway is not functional and in which Listeria preferentially uses the M-cells of the Peyer's patches as an entry portal (15). There was not a significant decrease in the number of ΔinlH bacteria in the small intestine or the colon 24 h after oral inoculation compared to the number of wild-type bacteria (Fig. 4C). Next we used transgenic mice expressing human E-cadherin in enterocytes (40), in which translocation of L. monocytogenes across the gut epithelium occurs through InlA-dependent invasion of enterocytes. There was not a significant difference between the numbers of ΔinlH and parental EGD-e bacteria in the small intestine 24 h (data not shown) or 48 h (Fig. 4C) following oral inoculation, whereas invasion of this tissue by the inlA mutant was, as expected, impaired. Together, these results strongly suggest that InlH is not involved in L. monocytogenes penetration of the intestinal epithelium.

L. monocytogenes infection is known to induce innate immune responses and cytokine production, which are essential for host survival (48, 51, 64, 67). Because InlH does not play a role in internalization or survival in cells in vitro but does play a role in vivo, we examined its effect on the modulation of innate immune responses to L. monocytogenes in vivo. Cytokine production in the liver and bacterial loads were quantified after 48 h of infection following i.v. inoculation. This time point was chosen because it corresponds to peaks in various cytokine titers (49) and to upregulation of inlH expression in the spleen during mouse listeriosis (11).

It has been reported that cytokine levels in tissues increase with the L. monocytogenes bacterial load (35). To test the effect of inlH inactivation on cytokine production independent of the bacterial load, we first performed experiments with the EGD-1 and ΔinlH-1 strains, for which there was not a statistically significant difference in the number of bacteria surviving in the liver at 48 h postinfection (Fig. 5A). Strikingly, inactivation of inlH led to a significant 2-fold increase in IL-6 production, while the amounts of several cytokines, such as IL-12, IFN-γ, IL-10, IL-1α, and IL-1β, remained unchanged (Fig. 5A). These results were confirmed with the EGD-e and isogenic ΔinlH strains. IL-6 production was 4-fold greater in the livers and spleens infected with the ΔinlH bacteria than in the livers and spleens infected with wild-type bacteria (Fig. 5B). In contrast, the amounts of TNF-α and IL-10 decreased in livers infected with the ΔinlH bacteria, possibly as a consequence of the ∼2- to 3-fold decrease in the bacterial load. Independent repeated experiments showed that IL-6 production still increased after 72 h of infection in the absence of InlH (data not shown). Together, these results strongly suggest that InlH plays a role in a process involved in IL-6 synthesis or secretion during L. monocytogenes infection.

Since macrophages are an important source of IL-6 during bacterial infections, we investigated the effect of InlH on IL-6 production in RAW 264.7 or THP-1 macrophages or in murine PEM. The level of IL-6 secreted by resting or LPS-activated RAW 264.7 macrophages infected for 16 h with EGD-e was similar to the level secreted by macrophages infected with ΔinlH bacteria (Fig. 5C). Likewise, InlH did not influence the production of IL-6 in PMA-activated PEM or THP-1 macrophages infected for 12 h or 24 h (Fig. 5C and data not shown). Thus, InlH does not play a role in a signaling pathway involved in production of IL-6 by macrophages.

We found that stress-responsive protein InlH of L. monocytogenes contributes to evasion of the host immune response by decreasing the IL-6 cytokine level. Innate immunity is essential for early control of L. monocytogenes infection, until T-cell-mediated immune responses eradicate intracellular bacteria (for reviews, see references 51, 64, and 67). Recent studies have highlighted different strategies by which L. monocytogenes evades innate defenses, including nonrecognition by innate receptors through N deacetylation of the peptidoglycan, manipulation of host cell autophagy, and listeriolysin O (LLO) toxin-mediated signaling pathways inducing T-cell apoptosis or repression of several immune genes (for a review, see reference 16). Since immunity against Listeria involves coordinated engagement of different immune cells and multiple cytokines at different stages and places in the infection process, it is likely that several different bacterial factors control these cells and/or extracellular molecules in each tissue.

To identify such factors, in vivo studies are necessary, as shown here for InlH, because in vitro experiments with cultured cells cannot reproduce the complex environment that pathogens have to cope with during natural infections. As an illustration, expression of the inlH gene is not induced in bacteria grown in cultured enterocytes or macrophages or in human blood ex vivo, but it is upregulated in the gut and in the spleen during infection in mice (Fig. 1) (11, 12, 31, 61). Likewise, InlC2, an InlH homolog present in a subset of L. monocytogenes strains, elicits a humoral response in vivo, which is consistent with its expression during infection (66). Here, using different L. monocytogenes and mouse genetic backgrounds, we show that inlH mutants have a reduced capacity to infect organs following i.v. inoculation of mice compared to the wild-type strains, whereas they do not have any defect in entry or persistence in cultured mammalian cells in vitro. Importantly, mice infected with inlH mutants have significantly higher levels of IL-6 than mice infected with wild-type bacteria, while the levels of other cytokines that also control primary L. monocytogenes infection, such as IFN-γ and TNF-α (48, 49, 64), elicited by inlH mutants are similar to the levels elicited by wild-type bacteria.

By what mechanism could L. monocytogenes alter the immune response through InlH-mediated downregulation of IL-6? Transcriptional regulation of the IL-6 gene is complex and involves at least five transcription factors, NF-κB, NF-IL-6, C/EBP, AP-1, and CREB, which are engaged differently depending on the cell type and on the stimulus (27, 37). Inhibition of the general NF-κB pathway by InlH is unlikely, as the production of other NF-κB-regulated cytokines, such as TNF-α and IL-1, is not increased by inactivation of inlH. In addition, InlH does not influence IL-6 production by activated macrophages, suggesting that it is not involved in a general signaling pathway controlling IL-6 expression. Nevertheless, InlH may control IL-6 expression or secretion in other immune cells, such as T cells. Alternatively, InlH could contribute to a pathway that modulates long-term survival of specific IL-6-secreting cells or the intercellular signals that are critical for activation of these cells. These hypotheses will be addressed in further studies.

IL-6 is a key cytokine in the control of inflammation and has mainly proinflammatory effects, although anti-inflammatory activities have also been reported (5, 62). It is also involved in other physiological functions, such as hematopoiesis, bone metabolism, and cell differentiation. IL-6 is strongly induced during acute bacterial infections, and studies with IL-6-deficient mice demonstrated the critical role of IL-6 in clearance of L. monocytogenes infections (17, 38, 42, 48). It is noteworthy that IL-6 seems to play an important role in the neutrophil response to Listeria (17). It has been proposed that in IL-6-deficient mice neutrophils may be less effective in killing bacteria, as suggested by the finding that IL-6 primes neutrophils for an oxidative burst, and that neutrophils may die from apoptosis more quickly (48). Our data suggest that L. monocytogenes uses InlH to counterbalance IL-6 production by immune cells in response to bacterial molecules that activate the NF-κB pathway (16, 32, 44, 63). The contribution of InlH to downmodulation of IL-6, together with other L. monocytogenes properties involved in evading host innate immunity (16), provides a rational basis accounting for the absence of strong inflammation observed during listeriosis. Indeed, in contrast to Listeria, other Gram-positive pathogens, such as Staphylococcus aureus and Streptococcus pyogenes, are pyogenic and elicit exacerbated inflammatory responses during systemic infections.

The inlH transcript is specifically upregulated by σ^B in the mouse intestinal tract (Fig. 1) (61), where L. monocytogenes encounters salt, acidic, or osmotic stress stimuli (21, 59), which, as we show here, induce production of the InlH protein at the bacterial surface. In addition, previous studies have established that σ^B plays a key role in regulating genes important during invasion of and survival in the gastrointestinal tract (23, 36, 60). InlH may thus contribute to L. monocytogenes infection of the gastrointestinal tract. However, we could not detect any defect in inlH mutants in early invasion of mouse intestinal tissue, suggesting that, in contrast to InlA, InlH is not involved in crossing of the intestinal barrier. Yet, InlH contributes to bacterial survival in deeper tissues. Since the InlH protein is highly expressed at the bacterial surface in response to heat and oxidative stresses, we suggest that inlH is induced and contributes to survival of L. monocytogenes in host microenvironments where these stresses may impede bacterial proliferation. These stresses could include hypoxia and fever that develop upon inflammation. Oxidative stress that occurs due to accumulation of reactive oxygen species (ROS) in activated immune cells may also be responsible for the induction of inlH in specific tissues. It is noteworthy that ROS are involved in the enhanced release of IL-6 during hypoxia (2). Given the variety of immune cells known to be infected by Listeria, which include not only macrophages, which do not respond to InlH, but also dendritic cells, neutrophils, and T cells (3, 13), it will be a challenge to define the cells in which inlH is induced in vivo in order to understand the role of the product of this gene in counteracting IL-6 secretion.

The inlG and inlE genes surround inlH. Our transcriptome studies using tiling arrays indicated that inlG and inlE are transcribed at very low levels in strain EGD-e and are not induced in any of the growth conditions tested under which inlH is active. This is consistent with the previously reported absence of inlE expression in L. monocytogenes strain 10403S (45). Thus, inlH is likely to act independent of inlG and inlE. We observed that inactivation of inlH or of its homolog inlC2 correlates with an increase in production of the InlA protein but not with an increase in production of the inlA transcript in bacteria growing in stationary phase. This suggests that in wild-type bacteria inlH or inlC2 expression interferes with the expression of inlA at a posttranscriptional level. InlH/InlC2 and InlA are LPXTG proteins covalently anchored to the peptidoglycan by sortase A. Hence, bacteria may adjust the amount of InlA protein in order to facilitate InlH/InlC2 secretion or anchoring at the bacterial surface during the infection process. These results and the results of previous studies (46, 50) suggest that there is an intricate network that allows L. monocytogenes to regulate internalin expression in response to changing environmental cues.

There are several examples of bacterial effectors that specifically target cytokine pathways to enhance pathogenesis (14, 28, 55). Among them are two LRR proteins that, in contrast to InlH, stimulate IL-6 cytokine production. In L. monocytogenes, InlB, besides being an invasin, activates the NF-κB pathway and expression of the IL-6 gene (44). In Enterococcus faecalis, disruption of the elrA gene that encodes an LRR virulence protein correlates with decreased IL-6 levels in vivo (10). It is thus likely that like other bacteria, L. monocytogenes has a repertoire of proteins that either block or activate the immune response during the different stages of the infectious process. Following tissue invasion that stimulates inflammatory processes, InlH may act as an immunosuppressor to increase L. monocytogenes persistence in tissues. Understanding the mechanism that InlH uses to control IL-6 production and identifying its host cell partner(s) are challenges for future work.