Skip to main content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
mBio. 2011 May-Jun; 2(3): e00016-11.
Published online 2011 May 17. doi: 10.1128/mBio.00016-11
PMCID: PMC3101776
PMID: 21586648

Streptococcus pneumoniae DNA Initiates Type I Interferon Signaling in the Respiratory Tract

Dane Parker, a Francis J. Martin, b Grace Soong, a Bryan S. Harfenist, a Jorge L. Aguilar, a Adam J. Ratner, a,c Katherine A. Fitzgerald, d Christian Schindler, c and Alice Prince a,b
Author information Article notes Copyright and License information PMC Disclaimer

Associated Data

Supplementary Materials

ABSTRACT

The mucosal epithelium is the initial target for respiratory pathogens of all types. While type I interferon (IFN) signaling is traditionally associated with antiviral immunity, we demonstrate that the extracellular bacterial pathogen Streptococcus pneumoniae activates the type I IFN cascade in airway epithelial and dendritic cells. This response is dependent upon the pore-forming toxin pneumolysin. Pneumococcal DNA activates IFN-β expression through a DAI/STING/TBK1/IRF3 cascade. Tlr4−/−, Myd88−/−, Trif−/−, and Nod2−/− mutant mice had no impairment of type I IFN signaling. Induction of type I IFN signaling contributes to the eradication of pneumococcal carriage, as IFN-α/β receptor null mice had significantly increased nasal colonization with S. pneumoniae compared with that of wild-type mice. These studies suggest that the type I IFN cascade is a central component of the mucosal response to airway bacterial pathogens and is responsive to bacterial pathogen-associated molecular patterns that are capable of accessing intracellular receptors.

IMPORTANCE

The bacterium Streptococcus pneumoniae is a leading cause of bacterial pneumonia, leading to upwards of one million deaths a year worldwide and significant economic burden. Although it is known that antibody is critical for efficient phagocytosis, it is not known how this pathogen is sensed by the mucosal epithelium. We demonstrate that this extracellular pathogen activates mucosal signaling typically activated by viral pathogens via the pneumolysin pore to activate intracellular receptors and the type I interferon (IFN) cascade. Mice lacking the receptor to type I IFNs have a reduced ability to clear S. pneumoniae, suggesting that the type I IFN cascade is central to the mucosal clearance of this important pathogen.

Introduction

The airway epithelium is a primary site for the recognition of both viral and bacterial pathogens and initiates the host defense against respiratory pathogens. Epithelial cells produce chemokines and cytokines, such as interleukin-8 (IL-8), IL-6, IL-1β, granulocyte-macrophage colony-stimulating factor (GM-CSF), and G-CSF, that function to recruit and activate phagocytic cells to eradicate organisms and infected cells (1). These mucosal epithelial cells are also a major source of the type I interferons (alpha interferon [IFN-α] and IFN-β) that are critical for viral clearance and coordinating immune regulation (2). The role of type I IFNs in pulmonary defenses is exemplified by the increased susceptibility of mice lacking the IFN-α/β receptor to influenza virus infection and secondary bacterial pneumonia (3). Bacterial induction of type I IFNs has been examined primarily within the context of infection due to intracellular pathogens and limited to pathways expressed in macrophages and dendritic cells (DCs) (4). These studies indicate that the release of bacterial pathogen-associated molecular patterns (PAMPs) within the phagolysosome, from the phagolysosome into the cytosol (5), or from bacteria within the cytosol also induces type I IFN responses, a situation which closely mimics viral infection (6). However, there is little data to indicate how nonphagocytic cells could produce type I IFNs in response to extracellular bacteria, such as the organisms which most commonly cause pneumonia.

The upper airway is the initial site of colonization by S. pneumoniae. Aspiration of these organisms into the lower airways induces a florid proinflammatory response ascribed to the stimulation of Toll-like receptors (TLRs) on the surface of airway and immune cells and activation of NF-κB-dependent signaling. S. pneumoniae produces a cholesterol-dependent, pore-forming toxin, pneumolysin (Ply), which has inherent proinflammatory activity (7) and is critical in pathogenesis (8, 9). As autolysis and the release of cell fragments accompany pneumococcal replication, the pneumolysin pore provides a conduit for these PAMPs to gain access to the mucosal airway cells (10). The presence of bacterial components, especially DNA, within the cytosol is a potent stimulus for the type I IFN cascade (11) that can be recognized by a number of recently described receptors, including RIG-1-like helicases, DAI (DNA-dependent activator of IFN regulatory factors), and several others (6, 11–15). In the experiments described herein, we demonstrate that pneumolysin and DNA are involved in the activation of type I IFN signaling, a response which contributes to pneumococcal clearance from the airway.

RESULTS

S. pneumoniae induces expression of type I IFN genes in the murine lung.

To test the hypothesis that S. pneumoniae activates type I IFN signaling, C57BL/6J mice were infected intranasally with 2 × 107 CFU of S. pneumoniae D39 for 24 h and induction of Ifnb and other cytokine-encoding genes in lung tissue was analyzed using quantitative reverse transcription (qRT)-PCR (Fig. 1A). A 10-fold increase in Ifnb was observed in addition to induction of the IFN-β-inducible genes Mx1, PKR, and LIF (16, 17). The chemokine KC (as a positive control) and the cytokine IL-6 (also a type I gene product) were up-regulated by almost 1,000-fold. To document that these type I IFN-dependent genes were induced through the expected JAK/STAT signaling cascade, phosphorylation of the STAT1 (P-STAT1) and STAT3 (P-STAT3) transcription factors in the infected mouse lung tissues was demonstrated and compared to a lipopolysaccharide (LPS) control (Fig. 1B).

An external file that holds a picture, illustration, etc. Object name is mbo0021111180001.jpg

S. pneumoniae activates the type I IFN response in vivo. Mice were infected with 2 × 107 CFU of S. pneumoniae and analyzed 24 h later. (A) qRT-PCR of type I IFN and inflammatory cytokine genes from RNA extracted from infected mouse lungs. Graphs display means with standard deviations (n = 3). (B) Immunoblots of transcription factors P-STAT1 and P-STAT3 from mouse lungs. β-Actin was used as a loading control. Each lane represents an individual mouse. Data are representative of two experiments. LPS from E. coli (50 µg per mouse) was used as a positive control.

Pneumolysin contributes to induction of type I IFN signaling in vivo and in vitro.

S. pneumoniae pneumolysin is reported to interact with TLR4 (7, 18), which interacts with TRIF (TIR domain-containing adapter inducing IFN-β) to participate in the induction of type I IFN signaling. Recent data (19) have suggested that the immune response to pneumolysin does not involve TLR4. Pneumolysin also functions as a pore that allows bacterial PAMPs access to the epithelial cytosol and pathogen recognition receptors (10). We assessed Ifnb production in mice 4 h following intranasal infection with 107 CFU of wild-type (WT) S. pneumoniae D39 or a ply null mutant. There was 40-fold induction of Ifnb by the WT S. pneumoniae strain under these conditions but only a 3-fold induction in mice infected with the ply mutant (P = 0.029) (Fig. 2A). At this 4-h time point, LIF and Mx1 were expressed and this expression was pneumolysin dependent. Significantly reduced levels of KC and IL-6 were also associated with exposure to the ply mutant (P < 0.001).

An external file that holds a picture, illustration, etc. Object name is mbo0021111180002.jpg

Pneumolysin is involved in the induction of type I IFN. C57BL/6J mice were infected with 2 × 107 CFU of S. pneumoniae (S. p.) intranasally and studied at 4 h (A to D). (A) WT-infected lung data are shown in black, and ply mutant-infected lung data are shaded grey. *, P < 0.05 (Student’s t test; n = 8). Lung homogenates were analyzed for bacterial counts (B), neutrophils (percentage of Ly6G Ly6C+ CD45+ cells) (C), and CD11c+ cells (D) by FACS. Lines represent median values. *, P < 0.05 (WT versus ply mutant; Mann-Whitney test; n = 8, except for the PBS controls). (E) Murine nasal epithelial cells in primary culture were stimulated for the times indicated with WT or ply mutant S. pneumoniae D39 in the presence or absence of cytochalasin D (CytoD) (F). DMSO, dimethyl sulfoxide. *, P < 0.05 (Student’s t test; n = 3). (G) Mice were intranasally administered S. pneumoniae strains (P1121 background) expressing WT pneumolysin, no pneumolysin, or a nonhemolytic toxoid variant and examined 4 h later. Lung homogenates were immunoblotted for P-STAT1. Each lane represents an individual mouse. (H) AF488-labeled pneumolysin is visualized in murine nasal epithelial cells at 1 h poststimulation. Murine epithelial cells in primary culture following stimulation with S. pneumoniae D39 or purified pneumolysin (Ply) (I) or with ply lysates with and without purified pneumolysin or PdB (J). *, P < 0.05 compared to unstimulated cells (Student’s t test; n = 3). RNA was extracted and analyzed for levels of IFN-β by qRT-PCR. All graphs display means with standard deviations.

Despite equal numbers of WT and mutant bacteria in the lung at this 4-h time point (Fig. 2B), there were significant differences in the immune cell populations. Pneumolysin expression was required for the recruitment of neutrophils, as the ply mutant and the phosphate-buffered saline (PBS) control attracted similar proportions of neutrophils into the lung (Fig. 2C). Infection with the ply null mutant resulted in the retention of a significantly (P = 0.0002) larger percentage of CD11c+ cells (analogous to PBS mice), representing macrophages and DCs, than did infection with WT S. pneumoniae (Fig. 2D), which is consistent with the idea that Ply is a major immunostimulant (7, 20, 21).

To assess epithelial cell involvement in the signaling of the type I IFN response, we tested the induction of Ifnb in murine airway epithelial cells in primary culture. Murine airway epithelial cells stimulated with S. pneumoniae had 10-fold induction of Ifnb at 4 h compared with the 2-fold induction stimulated by the ply null mutant (Fig. 2E). Ifnb induction in epithelial cells was not inhibited by cytochalasin D, indicating that endocytosis was unlikely to be involved in this signaling (Fig. 2F).

Pneumolysin-dependent pore formation is involved in induction of type I IFN signaling.

The role of pore formation in the stimulation of Ifnb was examined using a Ply W433F mutant, a toxoid (PdB)-expressing strain in the S. pneumoniae P1121 background that has 0.1% of the WT hemolytic activity but is still capable of binding and oligomerizing into cellular membranes (22, 23). STAT1 phosphorylation in murine lungs exposed to the S. pneumoniae toxoid-expressing strain was substantially less than that induced by the P1121 parent strain (Fig. 2G). We used fluorescent (Alexa Fluor 488 [AF488]-labeled) Ply to follow the distribution of the toxin in polarized airway monolayers, which was limited to the apical surfaces, consistent with a role as a pore (Fig. 2H). Activation of Ifnb was not detected when purified recombinant Ply was applied to the cells (Fig. 2I), as a control for signaling that p38 phosphorylation was still detected (see Fig. S1 in the supplemental material). However, when Ply was applied in the presence of a ply pneumococcal cell lysate, a significant increase (P = 0.045) in Ifnb induction was observed, but not with a toxoid version of the protein (Fig. 2J and Fig. S2 in the supplemental material shows that the toxoid has significantly reduce hemolytic activity). These results suggest that pneumolysin pore activity is involved in the delivery of a ligand that activates type I IFN signaling.

Induction of Ifnb by S. pneumoniae is TLR4, MyD88, trif, and Nod2 independent.

We next examined which signaling pathways are involved in mediating pneumococcal activation of IFN-β. As it has been suggested that Ply functions as a TLR4 agonist (7, 18, 24), we tested the participation of TLR4 and its adaptor TRIF in mediating pneumococcal type I IFN signaling (Fig. 3). C57BL/6J WT, Tlr4−/−, and Trif−/− mice were intranasally inoculated with either PBS or 107 CFU of S. pneumoniae, and induction of Ifnb and STAT1 phosphorylation was examined 24 h later. Not only did all of the mice express Ifnb, but the Tlr4−/− mice had increased levels of P-STAT1 (Fig. 3A and B) and Ifnb production (3-fold higher than those of the WT; Fig. 3B), which were not due to increased bacterial numbers in the lung (data not shown), as has been reported by others (25). Trif−/− mice were also able to phosphorylate STAT1, indicating that endosomal TLR3 and TLR4, which use Trif/Tram adaptors, are not involved in the response to S. pneumoniae.

An external file that holds a picture, illustration, etc. Object name is mbo0021111180003.jpg

Induction of IFN-β production by S. pneumoniae in mice with mutations in TLR4, Nod2, and innate adapters. WT and Tlr4−/−, Trif−/−, MyD88−/−, and Nod2−/− mutant C57BL/6J mice were infected intranasally with 107 CFU of S. pneumoniae and examined 24 h later. (A) Lung homogenates were assayed for STAT1 phosphorylation. Each lane represents an individual mouse. Shown are representative experiments. (B) Lung tissue was analyzed for induction of IFN-β by qRT-PCR (n = 4, except for Nod2−/− mutant mice [n = 6]).

Type I IFN signaling through the other endosomal TLRs (TLR7, TLR8, and TLR9) that utilize the adapter MyD88 (7) was also excluded. In response to S. pneumoniae, MyD88−/− null mice had increased activation of STAT1 and Ifnb induction compared to that of WT mice (Fig. 3A and B), consistent with increased levels of infection (26).

We next screened for involvement of the cytosolic receptor Nod2 (nucleotide-binding oligomerization domain 2), which senses peptidoglycan (27) and can be transported via the pneumolysin pore (10, 28). Although a link between Nod2 and type I IFN production has been observed (29–31), Nod2−/− mice still responded to S. pneumoniae (Fig. 3A and B).

Pneumococcal DNA activates IFN-β expression.

During active growth, pneumococci spontaneously lyse, releasing cellular components, including DNA. S. pneumoniae produces an autolysin (LytA, an amidase) required for autolysis and release of cell wall components (9). Incubation of epithelial cells with a lytA strain of S. pneumoniae did not activate type I IFN signaling (Fig. 4B), indicating that the inducing ligand is liberated from lysed cells. DNA is a potent activator of type I IFN signaling, and Ply could permit bacterial DNA to enter the cytosol, where it would interact with one or more of the several different types of DNA sensors that induce type I IFN signaling (11). Pneumococcal lysates incubated with DNase were found to stimulate significantly less (P = 0.046) Ifnb than control lysates (Fig. 4A), further implicating pneumococcal DNA as the ligand. To further implicate pneumolysin in enabling DNA entry into the epithelial cells, we demonstrated increased amounts of pneumococcal DNA in murine epithelial cells incubated with WT S. pneumoniae compared to the ply mutant (Fig. 4C). However, reconstitution of the system in vitro, i.e., adding purified Ply and S. pneumoniae DNA to murine airway epithelial cells in primary culture, did not result in Ifnb induction.

An external file that holds a picture, illustration, etc. Object name is mbo0021111180004.jpg

S. pneumoniae DNA activates type I IFN production. (A) Murine nasal epithelial cells were stimulated for 4 h with pneumococcal lysate that had been left untreated or treated with DNase. (B) Murine nasal epithelial cells were stimulated for 4 h with the WT or lytA mutant strain (n = 3). IFN-β induction was assessed by qRT-PCR and is displayed as fold induction over that in PBS-only controls. (C) Murine nasal epithelial cells were stimulated for 4 h with the WT and ply mutant strains (n = 3). DNA was extracted from the epithelial cells, and levels of S. pneumoniae 16S rRNA were measured by qRT-PCR. Un, untreated lysate. *, P < 0.05 compared to WT or untreated samples (Student’s t test). All results shown are representative of at least two independent experiments.

DAI participates in sensing of pneumococcal DNA.

There are numerous receptors that respond to foreign DNA but relatively few that do not involve TLRs. DNA receptors that stimulate IFN-β expression are available within both endosomes and the cytosol (11). DAI is a cytosolic double-stranded DNA sensor that responds to DNA from several sources (32), although it has been shown to be dispensable for DNA detection in cells isolated from DAI null mice (33). In response to S. pneumoniae in the lung, there was a significant (15-fold, P = 0.002) increase in DAI transcription, consistent with its known induction by IFN-β (32) and suggestive of involvement in pneumococcal signaling (Fig. 5A).

An external file that holds a picture, illustration, etc. Object name is mbo0021111180005.jpg

Pneumococcal DNA is sensed by the cytosolic DAI sensor. (A) Mice were infected with 107 CFU of S. pneumoniae D39 (S.p), and lung tissue was analyzed for induction of DAI 24 h later (n = 4). (B) Bone marrow-derived cells were stimulated with enzymatically treated S. pneumoniae lysates and then analyzed for Ifnb production by qRT-PCR (n = 3). (C) Bone marrow-derived cells from DAI−/− (n = 6) and STING−/− (n = 6) mice were transfected with pneumococcal DNA, and Ifnb induction was analyzed by qRT-PCR. Ifnb induction of pneumococcal DNA-stimulated knockout bone marrow-derived cells and their respective controls downstream of DAI (D and E) and other type I signaling components (F to H) (n = 3). Bone marrow-derived DCs and macrophages from DAI−/− (I) and STING−/− (J) mice (n = 6) and TBK1−/− (K) and IRF3−/− (L) mice (n = 3) were stimulated with S. pneumoniae for 2 h, and Ifnb levels were quantitated by qRT-PCR. Graphs display means plus standard deviations and are representative of two experiments (excluding STING data), and data are expressed as fold induction compared to that of the control cell line stimulated with PBS alone. *, P < 0.05 compared to unstimulated or WT samples (Student’s t test).

We took advantage of macrophage cell lines and dendritic cells in primary culture to evaluate the participation of the expected components of the DAI cascade in response to pneumococcal DNA. To confirm that pneumococcal DNA was a ligand in immune cells, we incubated pneumococcal lysate treated with DNase or RNase before stimulating dendritic cells (Fig. 5B). We observed a significant (90%, P = 0.007) decrease in Ifnb induction in the DNase-treated lysate. No loss of induction was observed in the RNase-treated sample. The requirement for DAI in recognizing pneumococcal DNA was confirmed in DAI-deficient macrophages (Fig. 5C). A significant (50%, P = 0.048) decrease in Ifnb induction was observed in DAI-deficient cells compared to that in WT controls. This was not specific to just pneumococcal genomic DNA, as DNA from Pseudomonas aeruginosa and Staphylococcus aureus also stimulated reduced IFN-β signaling in the DAI null background (bone marrow-derived DCs) (see Fig. S3 in the supplemental material; Fig. S4 with controls). We also investigated the role of the recently identified factor STING (stimulator of IFN genes) (14), which is essential for the recognition of non-CpG intracellular DNA species (Fig. 5C). Transfection of bone marrow-derived macrophages with pneumococcal DNA resulted in a complete absence of Ifnb induction in the STING−/− cells (P < 0.0001).

To confirm that the expected distal components of the DNA-sensing cascade were involved, we transfected pneumococcal DNA into macrophage cell lines cells lacking TANK-binding kinase 1 (TBK1) (in a TNFR1 null background since TBK1 null mice are embryonic lethal [34, 35]) (Fig. 5D) and IFN regulatory factor 3 (IRF3) (Fig. 5E) transcription factors, both of which were required for Ifnb induction, results consistent with the involvement of DAI in the sensing of cytosolic DNA (15).

Participation of RIG-1, which stimulates helicase activity in a complex with MAVS, resulting in IFN-β expression through TBK1 and IRF3 (12, 13), was excluded (Fig. 5F), as was signaling through IRF5 (Fig. 5G) and IRF7 (Fig. 5H).

The significance of DNA sensing in the context of live organisms was also addressed. S. pneumoniae-induced Ifnb transcription in bone marrow-derived cells from DAI−/− mice was >50% decreased (P = 0.048) and that in cells isolated from STING−/− mice was >40% decreased (P = 0.0081) compared with that in controls (Fig. 5I and J). Macrophage lines deficient in TBK1 and IRF3 displayed an even greater reduction of Ifnb induction (Fig. 5K and L), confirming the participation of the expected components of DAI-mediated IFN-β signaling.

Type I IFN signaling contributes to S. pneumoniae clearance.

The physiological role of type I IFNs in the pathogenesis of pneumococcal infection was then explored by comparing the colonization of WT mice with that of Ifnar−/− mice. At 7 days postinoculation of 129 SvEv mice infected intranasally with 107 CFU of S. pneumoniae, Ifnar null mice had significantly higher bacterial counts (7-fold increase, P = 0.0102) in nasal lavage fluid than WT controls (Fig. 6A). There were increased numbers of monocytes in the lungs of infected Ifnar−/− mice (Fig. 6B; 14% versus 9%, P = 0.0051) and dendritic cell populations (Fig. 6C; 4% versus 1.6%, P = 0.0079) compared with the WT controls, consistent with the known effects of type I IFN signaling on the mobilization of immune cells to the lung (36). Neutrophil recruitment was equivalent between WT and Ifnar−/− mice (Fig. 6D). We did observe a difference in neutrophils in Fig. 2C, but this was in cells taken at an early time point from an intranasal infection, and pneumolysin can activate cytokines other than type I IFNs (37, 38). We did not observe a difference in bacterial numbers in the lung in this model (data not shown), consistent with other studies (39). Thus, despite the ability to recruit phagocytic cells, in the absence of type I IFN signaling, there was increased colonization with S. pneumoniae.

An external file that holds a picture, illustration, etc. Object name is mbo0021111180006.jpg

Type I IFN production contributes to pneumococcal clearance. WT and Ifnar−/− mutant mice (129/SvEV background for both) were infected with 107 CFU of S. pneumoniae intranasally and examined 7 days later. Shown are bacterial counts recovered from nasal lavage fluid samples (A) and percentages of monocytic cells (percentage of CD11b+ CD45+ cells) (B), DCs (percentage of CD11b+ CD11c+ CD45+ cells) (C), and neutrophils (percentage of Ly6G Ly6C+ CD45+ cells) (D) from lung homogenates analyzed by FACS. Each data point represents an individual mouse (n = 12). Data are combined from two independent experiments. Lines represent median values. *, P < 0.05 compared to WT mice (Mann-Whitney test).

DISCUSSION

In this report, we demonstrate the participation of airway epithelial cells in the initial recognition of an important bacterial pathogen, S. pneumoniae, in the respiratory tract. This immune response is mediated by activation of the epithelial type I IFN cascade by the intracellular accumulation of pneumococcal DNA. As this pathway is readily activated by viral nucleic acids that accumulate in airway epithelial cells, it is not unexpected that bacterial nucleic acids could have a similar effect. The upper respiratory tract is continually exposed to commensal flora, but the bacterial components, including DNA released from lysed organisms, do not appear to activate mucosal signaling. S. pneumoniae, a pathogen, is distinct from commensal flora by virtue of its expression of pneumolysin, the pore-forming toxin that enables ligands to gain access to the epithelial cytosol and initiate type I IFN signaling. As the induction of proinflammatory responses in the lung is so potentially deleterious to respiration, it seems logical that mucosal signaling is activated only by PAMPs that are sensed intracellularly, by organisms that have breached the barrier function of the airway mucosa. Our findings suggest that type I IFN signaling may be a general host response to perceived cellular invasion and does not discriminate among types of pathogens, whether viral or bacterial, but instead responds to the nature of the specific PAMPS within the cell.

In a model of nasopharyngeal colonization, the usual route of pneumococcal infection, the type I IFN response analogous to that induced by influenza virus contributed to bacterial clearance. Production of pneumolysin seems to be critical in triggering immune recognition of S. pneumoniae as a pathogen. The role of type I IFN signaling in the host response to pneumococcal infection has been previously reported by Weigent et al. (40), who demonstrated increased susceptibility to infection in mice treated with anti-IFN-α/β antibodies. Mancuso et al. (41) similarly investigated the importance of type I IFN signaling in S. pneumoniae sepsis and meningitis models. The significance of nasopharyngeal colonization in the initiation of host defenses against S. pneumoniae was recently highlighted in a microarray study that identified the induction of a number of type I IFN-related genes in murine nasal lymphoid tissue early in the course of a 6-week colonization (42). Less clear is how necessary type I IFN induction is in the context of active pneumococcal pneumonia, as models of pneumonia using direct tracheal inoculation of the organisms fail to demonstrate a notable phenotype (39). However, as upper airway colonization precedes the development of pneumonia, this is a critical stage in the pathogenesis of systemic pneumococcal infection.

It is increasingly apparent that the mucosal epithelium of the lung provides much more than a physical barrier to infection but actively participates in sensing and initiating immune signaling in response to inhaled pathogens, both viral and bacterial. This signaling includes not only the proinflammatory cascades set off by the apically exposed TLRs but also intracellular sensing systems. For at least two important pulmonary pathogens, S. aureus and S. pneumoniae, shed components that are either endocytosed (S. aureus protein A [36]) or taken up via pore formation (pneumolysin) trigger IFN-β production in both epithelial and immune cells. The ability of these PAMPs from Gram-positive extracellular bacteria to access intracellular receptors suggests a central role for the type I IFN cascade in mucosal defenses against respiratory pathogens of all types.

The expression of pneumolysin is critical in the activation of type I IFN signaling by S. pneumoniae. Pneumolysin has been shown previously to facilitate the introduction of peptidoglycan into host cells, as well as induce host signaling from osmotic changes (28, 43). DNA from lysed commensal flora in the upper respiratory tract could potentially be immunostimulatory, if the organisms were able to deliver it into mucosal cells. The data suggest that pneumolysin facilitates entry of DNA via its pore formation. We observed the requirement for pneumococcal lysis and release of pneumococcal DNA as an autolysin null strain was unable to induce type I IFN signaling. We could obviate this requirement artificially by transfection of DNA or exposing naturally phagocytic DCs to cell lysates. This suggests that the ability to access cytosolic receptors may also contribute significantly to virulence. However, the addition of purified pneumolysin and DNA did not activate signaling (data not shown). The inability to artificially recapitulate the system in vitro indicates a missing component, alternative ligand, or stoichiometric effects necessary to activate signaling.

Several discrete receptors in both the cytosol and endosomes of the cells that make up the respiratory mucosa participate in type I IFN signaling. Our data suggest that DAI, whose role in immune signaling has been debated (15, 33), may be important in the response to pneumococcal DNA and other bacterial DNA. There was a significant induction of DAI expression by S. pneumoniae and a significant requirement for DAI to induce Ifnb in response to pneumococcal DNA. In addition, the contribution of other potential DNA sensors and signaling components (MAVS, RIG-I, and TLRs) was specifically excluded (although there was a trend to reduced induction in the MAVS null background, suggesting a limited role for RNA polymerase III [12, 13]). The expected downstream components of this pathway, including the recently identified STING (stimulator of IFN genes), TBK1, and IRF3 proteins, were also shown to be involved in the induction of Ifnb in the contexts of both DNA and the whole organism.

There is a growing literature detailing how bacterial pathogens induce the type I IFN cascade in macrophages (6, 30, 44–47). This report provides a second example of how an extracellular pathogen can activate type I IFN signaling, an intracellular response (36). We now show that epithelial cells similarly participate in the induction of type I IFN signaling. Exactly how the epithelial cells, which are not actively phagocytic, are stimulated to express IFN-β in response to other types of airway pathogens remains to be established. However, based upon our observations with the pneumococcus, it seems likely that a general induction of type I IFN signaling in the lung, including the airway mucosal cells, is an important consequence of both bacterial and viral infections of the lung.

MATERIALS AND METHODS

Bacterial strains.

S. pneumoniae strains D39 (48), D39 ply (49), 1725 (P1121, a derivative of P833, a human carriage isolate [50]), 1726 (P1121 ply [10]), 1727 (P1121 ply plyW433F+, a PdB toxoid-expressing strain [28]), and D39 lytA (51) were grown on Trypticase soy agar (supplemented with erythromycin for D39 ply) or broth supplemented with 200 U/ml catalase (Worthington). Plate cultures were grown at 37°C in the presence of carbon dioxide (5%). Escherichia coli strain BL21AI (Invitrogen), P. aeruginosa PAO1, and S. aureus USA300 were grown at 37°C on LB. All chemicals were purchased from Sigma unless otherwise stated.

RNA analysis.

RNA was isolated using the RNAqueous-4PCR kit (Ambion) with DNase treatment. cDNA was made using the iScript Synthesis Kit (Bio-Rad). qRT-PCR was performed using Power SYBR Green PCR Master Mix in a StepOne Plus thermal cycler (Applied Biosystems). Samples were normalized to β-actin or glyceraldehyde 3-phosphate dehydrogenase levels (sense, 5′ ACCACAGTCCATGCCATCAC 3′; antisense, 5′ TCCACCACCCTGTTGCTGTA 3′). 16S rRNA primers for S. pneumoniae were 5′ GCCTACATGAAGTCGGAATCG 3′ (sense) and 5′ TACAAGGCCCGGGAACGT 3′ (antisense). Primers for mouse actin, Ifnb, Mx1, PKR, LIF, KC, IL-6, and DAI have been described elsewhere (32, 36, 52).

Protein purification and labeling.

E. coli BL21AI carrying pET29a/pneumolysin or pneumolysin W433F (37, 53) was grown to mid-exponential phase and induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and 0.2% arabinose for 3 h at 37°C. Cells were lysed using Bugbuster (Novagen) and sonication in the presence of the HALT protease cocktail (Pierce). His-tagged protein was purified from clarified lysates with nickel-charged agarose resin using Poly-Prep gravity flow columns (Bio-Rad). Eluted protein was also passed through an endotoxin removal column (Detoxi-Gel columns; Thermo Scientific). When required, protein was concentrated to 2 mg/ml using Amicon centrifugation columns and labeled with AF488 according to the manufacturer’s instructions (AF448 labeling kit; Molecular Probes-Invitrogen). Hemolytic activity of proteins was confirmed (see Fig. S2 in the supplemental material).

Confocal microscopy.

AF488-labeled pneumolysin (25 µg/ml) was incubated with murine nasal epithelial cells in primary culture for 1 h. Cells were washed three times in PBS and fixed in 4% paraformaldehyde before being blocked in 5% normal goat serum with 0.3% Triton X-100 to permeabilize cells. Phalloidin conjugated to rhodamine was used to stain actin. Imaging was performed on a Zeiss LSM 510 META scanning confocal microscope and analyzed using LSM Image Browser software (version 4.2).

Cell culture.

Mouse nasal epithelial cells grown polarized from septa were isolated from the indicated strains of adult mice and grown as described elsewhere (54). Cells were stimulated with 5 × 108 CFU/ml of bacteria. Purified pneumolysin was applied to cells at a final concentration of 2 µg/ml for 4 h. Experiments were performed with cytochalasin D at 20 µM and included a 30-min preincubation of inhibitor before the organism was applied. Pneumococcal lysate experiments were performed using 5 mg/ml of protein. S. pneumoniae lysates (1 mg/ml) were treated with 100 U/ml DNase (Ambion) or 25 µg/ml RNase (digestion verified by gel electrophoresis) for 2 h at 37°C before heat inactivation at 70°C for 10 min. DCs and macrophages were generated from bone marrow isolated from femurs and tibias (from DAI and STING null mice). DCs were cultured for 7 days in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), streptomycin, penicillin, and 20 ng/ml of GM-CSF (PeproTech). Macrophages were grown in Dulbecco’s modified Eagle’s medium with 10% FBS, ciprofloxacin, and 20% L929 supernatant. Immortalized macrophages (from TBK1, IRF3, IRF5, IRF7, and MAVS null mice) were prepared as described previously (55) and stimulated with 4 × 107 CFU/ml S. pneumoniae. Cell transfections were carried out using bacterial genomic DNA at a final concentration of 2 µg/ml. DNA was isolated using the DNeasy kit (Qiagen), and transfections were performed for 6 h using FuGENE according to the manufacturer’s instructions (Roche).

Mouse studies.

Lung infections with S. pneumoniae were performed using 6-week-old C57BL/6J mice. Nod2−/− mice were obtained from Jackson Laboratories (originally from Regeneron Pharmaceuticals). Mice were anesthetized with 100 mg/kg ketamine and 5 mg/kg xylazine and intranasally inoculated with 2 × 107 to 5 × 107 CFU of the organism. LPS from E. coli (50 µg per mouse) was used as a control. Colonization studies were performed with Ifnar−/− and WT 129/SvEV mice. Six- to 7-week-old mice were inoculated with S. pneumoniae D39 without anesthesia. For fluorescence-activated cell sorter (FACS) analysis, red blood cells were lysed from lung homogenates. Cells were suspended in PBS with 2% FBS and stained with combinations of phycoerythrin-labeled anti-CD45 (Caltag Laboratories), fluorescein isothiocyanate-labeled anti-Ly6G/Ly6C (clone RB6-8C5; BD Biosciences-Pharmingen), PerCP-Cy5.5-labeled anti-CD11b (M1/70; eBioscience), and allophycocyanin-labeled anti-CD11c (N418; eBioscience) antibodies in the presence of 10% normal mouse serum and Fc block (2.4G2; BD Biosciences). Negative controls were stained with relevant isotype-matched antibodies. Cells were fixed with 1% paraformaldehyde and analyzed on a FACScalibur (Becton Dickinson) using CellQuest software (version 3.3; BD). Cells were gated on their side and forward scatter and CD45 expression. Data were analyzed using WinMDI (version 2.8; Joseph Trotter). All mouse infections were performed under the guidelines of the Institutional Animal Care and Use Committee of Columbia University.

Western blotting.

Phosphorylation of STATs was detected using P-STAT1 (Y701; Abcam) and P-STAT3 (Y705; Cell Signaling Technology) and normalized to β-actin (Sigma), while phosphorylation of p38 was detected using P-p38 (T180/Y182; Cell Signaling Technology) and normalized to total p38 (Cell Signaling Technology). All immunoblot assays were followed by secondary antibodies conjugated to horseradish peroxidase (Santa Cruz Biotechnology Inc.). Samples from at least two individual mice were typically analyzed on two separate occasions. Protein separation, transfer, and immunoblotting were performed as previously described (36).

Statistics.

Significance of data that followed a normal distribution was determined using a two-tailed Student t test, and for data that did not follow a normal distribution, a nonparametric Mann-Whitney test was used. Statistics were performed with GraphPad Prism software.

SUPPLEMENTAL MATERIAL

Figure S1

Activation of p38 mitogen-activated protein kinase activity by purified pneumolysin. Murine epithelial cells in primary culture were incubated with purified pneumolysin for the specified times, and phosphorylation of p38 was detected by Western blotting. Total p38 was detected as a loading control. Download Figure S1, TIF file, 0.073 MB.

Figure S2

Hemolytic activity of purified pneumolysin proteins. Pneumolysin and the toxoid PdB were incubated at various concentrations with sheep red blood cells (1%) in Hanks balanced salt solution for 30 min at 37°C in CO2 before centrifugation of red blood cells and measurement of supernatant at 415 nm. One hundred percent lysis was determined by incubating red blood cells with 1% Triton X-100. Download Figure S2, TIF file, 0.042 MB.

Figure S3

DAI senses bacterial DNA. Bone marrow-derived DCs from WT and DAI−/− mutant mice (n = 6) were transfected with genomic DNA from S. pneumoniae (S. p), P. aeruginosa (P. a), and S. aureus (S. a), and IFNb induction was analyzed by qRT-PCR. Un, unstimulated. *, P < 0.05 (Student’s t test). Download Figure S3, TIF file, 0.046 MB.

Figure S4

Ability of major knockout cell lines to signal. Bone marrow-derived DCs from DAI (A) and STING (B) null mice were transfected with poly(dA-dT), and IFNb levels were quantitated. TBK1 (C) and IRF3 (D) macrophages stimulated with S. pneumoniae were measured for KC induction. A trend to reduced signaling is seen with the TBK1 and IRF3 null cells, as this pathway can also lead to NF-κB signaling. pAT, poly(dA-dT); un, unstimulated; n.s., not significant. Download Figure S4, TIF file, 0.092 MB.

ACKNOWLEDGMENTS

We thank Connie Woo, Tracie Seimon, Ira Tabas, Neil Paragas, and Jonathan Barasch for Tlr4−/− and Trif−/− mice; Jeffery Weiser for strains; and Osamu Takeuchi, Shizuo Akira, Glen Barber, and Charlie Kim for DAI and STING bone marrow.

D.P. was the recipient of an NHMRC Overseas Biomedical Fellowship. This work was supported by NIH grant 1R21AI083491.

We have no competing financial interests.

D.P. performed the experiments, F.J.M. assisted with mice and FACS experiments, G.S. performed confocal microscopy, B.S.H. and J.L.A. provided technical support, A.J.R., K.A.F., and C.S. provided reagents, A.J.R. and A.P. supervised the research, and D.P. and A.P. conceived the experiments and wrote the manuscript.

Footnotes

Citation Parker D, et al. 2011. Streptococcus pneumoniae DNA initiates type I interferon signaling in the respiratory tract. mBio 2(3):e00016-11. doi:10.1128/mBio.00016-11.

REFERENCES

1. Gómez MI, Prince A. 2008. Airway epithelial cell signaling in response to bacterial pathogens. Pediatr. Pulmonol. 43:11–19 [PubMed] [Google Scholar]
2. Shornick LP, et al. 2008. Airway epithelial versus immune cell Stat1 function for innate defense against respiratory viral infection. J. Immunol. 180:3319–3328 [PubMed] [Google Scholar]
3. García-Sastre A, Biron CA. 2006. Type 1 interferons and the virus-host relationship: a lesson in détente. Science 312:879–882 [PubMed] [Google Scholar]
4. Decker T, Müller M, Stockinger S. 2005. The yin and yang of type I interferon activity in bacterial infection. Nat. Rev. Immunol. 5:675–687 [PubMed] [Google Scholar]
5. Stanley SA, Johndrow JE, Manzanillo P, Cox JS. 2007. The type I IFN response to infection with Mycobacterium tuberculosis requires ESX-1-mediated secretion and contributes to pathogenesis. J. Immunol. 178:3143–3152 [PubMed] [Google Scholar]
6. Uematsu S, Akira S. 2007. Toll-like receptors and type I interferons. J. Biol. Chem. 282:15319–15323 [PubMed] [Google Scholar]
7. Malley R, et al. 2003. Recognition of pneumolysin by Toll-like receptor 4 confers resistance to pneumococcal infection. Proc. Natl. Acad. Sci. U. S. A. 100:1966–1971 [PMC free article] [PubMed] [Google Scholar]
8. Berry AM, Paton JC. 2000. Additive attenuation of virulence of Streptococcus pneumoniae by mutation of the genes encoding pneumolysin and other putative pneumococcal virulence proteins. Infect. Immun. 68:133–140 [PMC free article] [PubMed] [Google Scholar]
9. Kadioglu A, Weiser JN, Paton JC, Andrew PW. 2008. The role of Streptococcus pneumoniae virulence factors in host respiratory colonization and disease. Nat. Rev. Microbiol. 6:288–301 [PubMed] [Google Scholar]
10. Ratner AJ, Lysenko ES, Paul MN, Weiser JN. 2005. Synergistic proinflammatory responses induced by polymicrobial colonization of epithelial surfaces. Proc. Natl. Acad. Sci. U. S. A. 102:3429–3434 [PMC free article] [PubMed] [Google Scholar]
11. Hornung V, Latz E. 2010. Intracellular DNA recognition. Nat. Rev. Immunol. 10:123–130 [PubMed] [Google Scholar]
12. Ablasser A, et al. 2009. RIG-I-dependent sensing of poly(dA:dT) through the induction of an RNA polymerase III-transcribed RNA intermediate. Nat. Immunol. 10:1065–1072 [PMC free article] [PubMed] [Google Scholar]
13. Chiu YH, Macmillan JB, Chen ZJ. 2009. RNA polymerase III detects cytosolic DNA and induces type I interferons through the RIG-I pathway. Cell 138:576–591 [PMC free article] [PubMed] [Google Scholar]
14. Ishikawa H, Ma Z, Barber GN. 2009. STING regulates intracellular DNA-mediated, type I interferon-dependent innate immunity. Nature 461:788–792 [PMC free article] [PubMed] [Google Scholar]
15. Takaoka A, et al. 2007. DAI (DLM-1/ZBP1) is a cytosolic DNA sensor and an activator of innate immune response. Nature 448:501–505 [PubMed] [Google Scholar]
16. Haller O, Staeheli P, Kochs G. 2007. Interferon-induced Mx proteins in antiviral host defense. Biochimie 89:812–818 [PubMed] [Google Scholar]
17. Quinton LJ, et al. 2008. Alveolar epithelial STAT3, IL-6 family cytokines, and host defense during Escherichia coli pneumonia. Am. J. Respir. Cell Mol. Biol. 38:699–706 [PMC free article] [PubMed] [Google Scholar]
18. Srivastava A, et al. 2005. The apoptotic response to pneumolysin is Toll-like receptor 4 dependent and protects against pneumococcal disease. Infect. Immun. 73:6479–6487 [PMC free article] [PubMed] [Google Scholar]
19. McNeela EA, et al. 2010. Pneumolysin activates the NLRP3 inflammasome and promotes proinflammatory cytokines independently of TLR4. PLoS Pathog. 6:e1001191 [PMC free article] [PubMed] [Google Scholar]
20. Kadioglu A, et al. 2000. Host cellular immune response to pneumococcal lung infection in mice. Infect. Immun. 68:492–501 [PMC free article] [PubMed] [Google Scholar]
21. Maus UA, et al. 2004. Pneumolysin-induced lung injury is independent of leukocyte trafficking into the alveolar space. J. Immunol. 173:1307–1312 [PubMed] [Google Scholar]
22. Alexander JE, et al. 1998. Amino acid changes affecting the activity of pneumolysin alter the behaviour of pneumococci in pneumonia. Microb. Pathog. 24:167–174 [PubMed] [Google Scholar]
23. Korchev YE, et al. 1998. A conserved tryptophan in pneumolysin is a determinant of the characteristics of channels formed by pneumolysin in cells and planar lipid bilayers. Biochem. J. 329(Pt. 3):571–577 [PMC free article] [PubMed] [Google Scholar]
24. Park JM, Ng VH, Maeda S, Rest RF, Karin M. 2004. Anthrolysin O and other gram-positive cytolysins are Toll-like receptor 4 agonists. J. Exp. Med. 200:1647–1655 [PMC free article] [PubMed] [Google Scholar]
25. Branger J, et al. 2004. Role of Toll-like receptor 4 in gram-positive and gram-negative pneumonia in mice. Infect. Immun. 72:788–794 [PMC free article] [PubMed] [Google Scholar]
26. Albiger B, et al. 2005. Myeloid differentiation factor 88-dependent signalling controls bacterial growth during colonization and systemic pneumococcal disease in mice. Cell. Microbiol. 7:1603–1615 [PubMed] [Google Scholar]
27. Girardin SE, et al. 2003. Nod2 is a general sensor of peptidoglycan through muramyl dipeptide (MDP) detection. J. Biol. Chem. 278:8869–8872 [PubMed] [Google Scholar]
28. Ratner AJ, Aguilar JL, Shchepetov M, Lysenko ES, Weiser JN. 2007. Nod1 mediates cytoplasmic sensing of combinations of extracellular bacteria. Cell. Microbiol. 9:1343–1351 [PMC free article] [PubMed] [Google Scholar]
29. Leber JH, et al. 2008. Distinct TLR-and NLR-mediated transcriptional responses to an intracellular pathogen. PLoS Pathog. 4:e6 [PMC free article] [PubMed] [Google Scholar]
30. Pandey AK, et al. 2009. NOD2, RIP2 and IRF5 play a critical role in the type I interferon response to Mycobacterium tuberculosis. PLoS Pathog. 5:e1000500 [PMC free article] [PubMed] [Google Scholar]
31. Stockinger S, et al. 2004. IFN regulatory factor 3-dependent induction of type I IFNs by intracellular bacteria is mediated by a TLR- and Nod2-independent mechanism. J. Immunol. 173:7416–7425 [PubMed] [Google Scholar]
32. Takaoka A, Yanai H. 2006. Interferon signalling network in innate defence. Cell. Microbiol. 8:907–922 [PubMed] [Google Scholar]
33. Ishii KJ, et al. 2008. TANK-binding kinase-1 delineates innate and adaptive immune responses to DNA vaccines. Nature 451:725–729 [PubMed] [Google Scholar]
34. Bonnard M, et al. 2000. Deficiency of T2K leads to apoptotic liver degeneration and impaired NF-kappaB-dependent gene transcription. EMBO J. 19:4976–4985 [PMC free article] [PubMed] [Google Scholar]
35. Perry AK, Chow EK, Goodnough JB, Yeh WC, Cheng G. 2004. Differential requirement for TANK-binding kinase-1 in type I interferon responses to Toll-like receptor activation and viral infection. J. Exp. Med. 199:1651–1658 [PMC free article] [PubMed] [Google Scholar]
36. Martin FJ, et al. 2009. Staphylococcus aureus activates type I IFN signaling in mice and humans through the Xr repeated sequences of protein A. J. Clin. Invest. 119:1931–1939 [PMC free article] [PubMed] [Google Scholar]
37. Aguilar JL, et al. 2009. Phosphatase-dependent regulation of epithelial mitogen-activated protein kinase responses to toxin-induced membrane pores. PLoS One 4:e8076 [PMC free article] [PubMed] [Google Scholar]
38. Bernatoniene J, et al. 2008. Induction of CC and CXC chemokines in human antigen-presenting dendritic cells by the pneumococcal proteins pneumolysin and CbpA, and the role played by Toll-like receptor 4, NF-kappaB, and mitogen-activated protein kinases. J. Infect. Dis. 198:1823–1833 [PubMed] [Google Scholar]
39. Shahangian A, et al. 2009. Type I IFNs mediate development of postinfluenza bacterial pneumonia in mice. J. Clin. Invest. 119:1910–1920 [PMC free article] [PubMed] [Google Scholar]
40. Weigent DA, Huff TL, Peterson JW, Stanton GJ, Baron S. 1986. Role of interferon in streptococcal infection in the mouse. Microb. Pathog. 1(4):399–407 [PubMed] [Google Scholar]
41. Mancuso G, et al. 2007. Type I IFN signaling is crucial for host resistance against different species of pathogenic bacteria. J. Immunol. 178:3126–3133 [PubMed] [Google Scholar]
42. Joyce EA, Popper SJ, Falkow S. 2009. Streptococcus pneumoniae nasopharyngeal colonization induces type I interferons and interferon-induced gene expression. BMC Genomics 10:404 [PMC free article] [PubMed] [Google Scholar]
43. Ratner AJ, et al. 2006. Epithelial cells are sensitive detectors of bacterial pore-forming toxins. J. Biol. Chem. 281:12994–12998 [PMC free article] [PubMed] [Google Scholar]
44. Bogdan C, Mattner J, Schleicher U. 2004. The role of type I interferons in non-viral infections. Immunol. Rev. 202:33–48 [PubMed] [Google Scholar]
45. Gratz N, et al. 2008. Group A streptococcus activates type I interferon production and MyD88-dependent signaling without involvement of TLR2, TLR4, and TLR9. J. Biol. Chem. 283:19879–19887 [PMC free article] [PubMed] [Google Scholar]
46. Henry T, Brotcke A, Weiss DS, Thompson LJ, Monack DM. 2007. Type I interferon signaling is required for activation of the inflammasome during Francisella infection. J. Exp. Med. 204:987–994 [PMC free article] [PubMed] [Google Scholar]
47. O’Connell RM, et al. 2004. Type I interferon production enhances susceptibility to Listeria monocytogenes infection. J. Exp. Med. 200:437–445 [PMC free article] [PubMed] [Google Scholar]
48. Avery OT, MacLeod CM, McCarty M. 1944. Studies on the chemical nature of the substance inducing transformation of pneumococcal types: inductions of transformation by a desoxyribonucleic acid fraction isolated from pneumococcus type III. J. Exp. Med. 79:137–158 [PMC free article] [PubMed] [Google Scholar]
49. King SJ, et al. 2004. Phase variable desialylation of host proteins that bind to Streptococcus pneumoniae in vivo and protect the airway. Mol. Microbiol. 54:159–171 [PubMed] [Google Scholar]
50. McCool TL, Cate TR, Moy G, Weiser JN. 2002. The immune response to pneumococcal proteins during experimental human carriage. J. Exp. Med. 195:359–365 [PMC free article] [PubMed] [Google Scholar]
51. Berry AM, Lock RA, Hansman D, Paton JC. 1989. Contribution of autolysin to virulence of Streptococcus pneumoniae. Infect. Immun. 57:2324–2330 [PMC free article] [PubMed] [Google Scholar]
52. Soong G, et al. 2006. Bacterial neuraminidase facilitates mucosal infection by participating in biofilm production. J. Clin. Invest. 116:2297–2305 [PMC free article] [PubMed] [Google Scholar]
53. Gelber SE, Aguilar JL, Lewis KL, Ratner AJ. 2008. Functional and phylogenetic characterization of vaginolysin, the human-specific cytolysin from Gardnerella vaginalis. J. Bacteriol. 190:3896–3903 [PMC free article] [PubMed] [Google Scholar]
54. Antunes MB, et al. 2007. Murine nasal septa for respiratory epithelial air-liquid interface cultures. Biotechniques 43:195–204 [PubMed] [Google Scholar]
55. Hornung V, et al. 2008. Silica crystals and aluminum salts activate the NALP3 inflammasome through phagosomal destabilization. Nat. Immunol. 9:847–856 [PMC free article] [PubMed] [Google Scholar]
Articles from mBio are provided here courtesy of American Society for Microbiology (ASM)
-