Bacterial strains and plasmids used in this study are listed in Table 1. Spn was grown in Brain-Heart Infusion (Difco) broth supplemented with 10% heat-inactivated horse serum (Gibco) and 10% catalase (3 mg/ml) to an OD₆₀₀ of 0.8, correlating to approximately 1 × 10⁸ CFU/mL. Broth cultures were mixed 1:1 with a 50% glycerol solution and frozen at −80°C for future use. B. pertussis strains were maintained on Bordet–Gengou agar (BG) supplemented with 10% defibrinated sheep blood. Liquid cultures were grown in Stainer–Scholte broth with heptakis (2,6-di-O-methyl-β-cyclodextrin) (29). Escherichia coli strains were grown in Luria–Bertani medium. As necessary, the growth media were supplemented with appropriate antibiotics, chloramphenicol (Cm, 10 or 50 µg ml⁻¹), kanamycin (Km, 25 µg ml⁻¹), and streptomycin (Sm, 100 µg ml⁻¹).

Broth cultures were inoculated with Spn, Bps, or E. coli under conditions described above and cultured overnight. 1% w/v choline chloride (Sigma) was added to Spn cultures to limit bacterial autolysis. Following the culture period, bacteria were pelleted by centrifugation and washed 3–5× with ice cold PBS. Bacteria were disrupted by passaging 5× through a C3 Avestin emulsifier. Crude lysates were centrifuged (12,000 g) to pellet insoluble material. Protein was quantified with a BCA protein kit (Pierce). Lysates were aliquoted and stored at −80°C for future use.

Virus stocks were grown and tittered [median egg infectious dose (EID₅₀)] in 10-day-old fertilized chicken eggs (obtained from a local farm) eggs as described previously (35). Stocks were diluted in PBS, flash frozen, and stored at −80°C.

The 8–10–week-old female BALB/c mice were purchased from The Jackson Laboratories. Mice were housed in a biosafety level 2 facility with ad libitum access to food and water.

Mice were anesthetized with Avertin (2,2,2-tribromoethanol) by intraperitoneal (i.p.) injection. Virus (10³ EID₅₀) was administered via the intranasal (i.n.) route in 50 µl of PBS. Mediastinal lymph nodes (MLNs) were harvested 8 days following IAV infection. Cells were isolated and co-cultured in enriched medium in the presence of irradiated splenocytes previously pulsed with 10⁻⁷ M NP_147–155 peptide at a ratio of 1:10. 10% T-stim (Corning) was added as a source of IL-2. Cultures were restimulated weekly.

On day 6 or 7 postweekly stimulation, cells were cocultured in vitro with 10⁻⁶ M NP_147–155 peptide or PMA + IONO in the presence or absence of the indicated amount of lysate or PepN. Monensin and brefeldin A (BD Biosciences) were added to inhibit secretion of cytokines. For cells stimulated with peptide, BV420 conjugated NP_147–155/K^d tetramer (graciously supplied by the NIH tetramer facility) was included during the stimulation. This allowed tetramer labeling that otherwise may have been hampered as a result of TCR downregulation that occurs with stimulation. Anti-CD107a antibody (Biolegend) was also included during the stimulation period to capture granule release. Following the 5 h stimulation period, samples were stained with antimouse CD8α (Biolegend) and BV420 conjugated NP_147–155/K^d tetramer (necessary to identify antigen specific cells in the non-stimulated samples). Cells were then fixed and permeabilized (Cytofix/Cytoperm kit, BD Biosciences) followed by incubation with antibodies specific for IFNγ (Biolegend) and in some cases TNFα (Biolegend). Where indicated, data were normalized to % inhibition by the following: [%IFNγ⁺/%IFNγ⁺(No Lysate/PepN)] × 100 = %Functional; %Inhibition = 100 − %Functional. To determine cell viability, cells were incubated with 7-AAD (Biolegend) following antibody staining. Cells were then washed extensively prior to fixation and permeabilization.

For analysis of TCR signaling, cells were stimulated with peptide or PMA + IONO as indicated for 15 min. For the induction of STAT1 phosphorylation, CTL were stimulated for 15 min with 5,000 Units of Universal Type 1 IFN (PBL). Following the stimulation period cells were fixed in 2% paraformaldehyde for 10 min at 37°C. Cells were then washed, stained with anti-CD8 antibody, and fixed/permeabilized with True-Nuclear Transcription Factor Staining kit (Biolegend) per the manufacturer’s instructions. The following antibodies were used: anti-ERK1/2 (Santa Cruz), antiphosphorylated ERK1/2 (ebioscience), antiphosphorylated p38 (ebioscience), antiphosphorylated ZAP-70 (BD Biosciences), antiphosphorylated JNK (BD Biosciences), antiphosphorylated STAT1 (BD Biosciences), and antiphosphorylated mammalian target of rapamycin (mTOR) (ebioscience).

Anion Exchange chromatography: Spn lysate was filtered and NaCl (Sigma) added to a final concentration of 10 mM. Lysate was then loaded onto a Q Sepharose HP column (GE Lifesciences). Proteins were eluted with an increasing gradient of NaCl (10 mM to 1 M) and fractions collected. Hydroxyapatite chromatography: Fractions from the previous step that contained inhibitory activity in our functional assay were pooled. Potassium phosphate (KPi, Sigma) was added to pooled fractions to a final concentration of 10 mM. These fractions were loaded onto a hydroxyapatite column (Macro-Prep ceramic hydroxyapatite resin, type 1, 40 µM, Bio-Rad). Proteins were eluted with a 10–500 mM KPi gradient; fractions were collected and tested for inhibition of T cell effector function.

To clone the pepN gene, the sequence of Spn strain EF3030 pepN was determined as follows: genomic DNA was subjected to PCR amplification using primers designed based on intergenic identity of strains D39, AP200, R6, and TIGR4 (Table 2). PCR products were purified with a Wizard PCR clean-up kit (Promega) and sequencing performed by Eton Bioscience Inc. (San Diego, CA, USA). DNA sequences were analyzed with ApE software (36) and the Basic Local Alignment Search Tool (http://www.ncbi.nlm.nih.gov/BLAST/).

Following confirmation of pepN gene sequence (submitted: GenBank KX522575), the gene was amplified from genomic DNA using primers to introduce a 5′ BamHI and a 3′ XhoI cut site [pepNF1-BamHI (cccggggatccatgcaagcagttgaacat), pepNR1-XhoI (cccgggctcgagttatgcatttccgtattg)]. PCR products were digested with BamHI and XhoI restriction enzymes (NEB) and purified. The pepN gene insert was then ligated into the pTHC-ΔP expression vector that enables cloning in frame with a His tag [for reference see Ref. (37)] using T4 DNA ligase (NEB). The resulting mixture was then transformed into XL-1 Blue E. coli (Stratagene), plated under selective conditions and colonies screened for pepN containing plasmid by double restriction digest with BamHI and XhoI.

Escherichia coli carrying Spn pepN were grown overnight under conditions to induce overexpression of PepN. The following day bacterial lysate was prepared as described above with the inclusion of 1 mM protease inhibitors (AEBSF, aprotinin, bestatin, E-64, leupeptin, and pepstatin A) (Pierce). Aliquots of lysate were run on a denaturing Bis-Tris SDS-PAGE resolving gel consisting of 10% acrylamide (BioRad). Proteins were visualized by staining with GelCode Blue Safe Protein Stain (Thermo). His-tagged PepN was purified by passage of lysate over 10 mL of Nickel Sepharose (GE Life Sciences). Bound protein was washed 3× with Tris buffer containing 50 mM sodium phosphate (Sigma), 10% glycerol (Fischer), 10 mM imidazole (Sigma), and 0.25% Tween-20 (Sigma). Tagged protein was eluted with buffer containing 150 mM imidazole. Following purification, PepN was dialyzed against PBS overnight at 4°C with PreScission protease (GE Life Sciences) to cleave the His-tag and remove any residual buffer components. The following day, PepN concentration was determined by BCA (Pierce) and then aliquoted and stored at −80°C for later use.

Following a 5-h peptide stimulation in the presence of PepN, lymphocyte RNA was isolated using an RNeasy RNA isolation kit (Qiagen). cDNA was synthesized from mRNA by reverse transcription using Superscript III RT kit (Invitrogen) and random primers (Promega). For IFNγ and GAPDH mRNA analysis, commercially available Taqman primer-probe sets specific for the gene targets were used. RT-PCR (qRT-PCR) was performed using the Applied Biosystems 7500 real-time PCR system. Raw data values were normalized to GAPDH mRNA levels.

The zinc-binding domain of PepN was rendered incapable of binding zinc by introducing the following mutations: His292Ala, Glu293Ala, and His296Ala. Site directed mutagenesis was performed by Genscript as a fee for service.

To quantify the peptidase activity of PepN and PepN^Met−, 1 µg of each enzyme was incubated with 200 µM of lys-AMC (BACHEM) at 37°C for 2 h. Following incubation, samples were excited in the 355 nm range and signal was detected in the 460 nm range using a 96-well PolarStar plate reader (Omega, BMG LABTECH). 1 µg of trypsin (Gibco) was used as a positive control.

Data analysis was performed using GraphPad Prism (GraphPad Software). Significance was determined using either a two-tailed Student’s t-test or a repeated measure ANOVA with a Tukey’s posttest as appropriate.

To assess the potential for Spn to directly modulate an effector T cell response, we employed an in vitro model that allowed for rapid and high throughput assessment on CTL function. We produced a lysate from the 19F serotype strain (EF3030) used in our previous in vivo studies (28) by overnight growth and subsequent mechanical disruption. To test the ability to regulate T cell function, increasing amounts of lysate were added at the time of NP-peptide restimulation of a CD8⁺ NP-specific CTL line. In the absence of lysate over 95% of the cells produced IFNγ and TNFα and were cytolytic as evidenced by CD107a positivity (Figure 1A). However, as the amount of Spn lysate present during peptide restimulation increased, we observed a dose-dependent inhibition of IFNγ and TNFα production as well as cytolytic granule release (Figure 1A representative flow plots at 3 µg, Figure 1B averaged data across the dose response). These data demonstrate the potential for bacterial components from Spn to directly inhibit CD8⁺ T cell effector function in a dose-dependent manner.

To extend our results to effectors that had not been subjected to multiple cycles of in vitro stimulation, BALB/c mice were infected with the H1N1 influenza virus A/Puerto Rico/8/1934 and cells from the lung draining MLNs were restimulated ex vivo with peptide in the presence of NP_147–155/K^d tetramer (this approach negated any issues with adequate tetramer labeling that could result from peptide-mediated TCR internalization). Following restimulation, cells were stained with anti-CD8, anti-IFNγ, and tetramer (to identify antigen-specific cells in the non-stimulated sample). As shown in Figure 2A, we detected a robust IFNγ response with peptide restimulation that was severely diminished in the presence of lysate. On average approximately 47.9% of CD8⁺Tet⁺ cells from the MLN of infected mice produced IFNγ compared to 4.7% of cells treated with Spn lysate (Figure 2B), representing a 10.2-fold reduction in IFNγ-producing cells. One explanation for this effect was that IFNγ production was impaired in cells treated with lysate as a result of killing by PLY, the cytotoxin present in Spn (38, 39). To assess death mediated by the lysate, effectors from the MLN of influenza infected mice were stimulated in the presence of lysate and cell viability determined by exclusion of 7-AAD. As shown in Figure 2C, cells stimulated with 3 µg of Spn lysate, a dose that has strong inhibitory function for cytokine production, had similar levels of 7-AAD positivity as cells stimulated in the absence of lysate. These data establish that a dose that was maximally inhibitory is not toxic to the CD8⁺ effector cells. To ensure that cells treated with lysate were not impaired because they were in the early stages of death, cells were stimulated for 18 h in the presence or absence of Spn lysate. Similar to what we observed with the 5-h exposure, there was no difference in death in treated versus non-treated cells (Figure 2D). These data rule out the induction of death as a mechanism to explain the lack of effector function in cells treated with lysate.

Bacterial pathogens produce a broad array of bacterial toxins (40). To determine whether the inhibitory effect observed following addition of Spn lysate could be attributed to a non-specific effect of adding bacterial lysate to T cells during stimulation, we generated lysates from the respiratory pathogen Bordetella pertussis (BP536) and a laboratory strain of E. coli (XL-1 Blue). Lysates from overnight cultures of these bacteria were prepared in the same manner as the Spn lysate. NP-specific CTL were then restimulated with peptide in the presence or absence of these bacterial lysates. While T cells stimulated with peptide in the presence of Spn lysate showed 80% inhibition in cytokine production (Figure 3A) stimulation with the B. pertussis or E. coli lysate showed minimal inhibition (Figure 3A). These data support the hypothesis that the T cell inhibitory activity was not a general property of bacterial lysates, but rather was mediated by a pneumococcal product(s).

We next tested the hypothesis that the ability to inhibit T cell effector function was a general property of Spn, i.e., that lysates generated from strains encompassing a variety of serotypes exhibited this capability. Spn currently has over 90 known serotypes, which induce a disease encompassing a range of severity (41). We included in our panel pathogenic strains such as TIGR4 or D39 (capsular serotypes 4 and 2, respectively) as well as an unencapsulated strain, MNZ1113 (Table 1). Lysates from all strains were generated in the same manner as EF3030 lysates. Influenza-specific CTL were restimulated with peptide in the presence or absence of each lysate. As shown in Figure 3B, we found that all of the lysates were capable of inhibiting IFNγ production. Interestingly, there was divergence with regard to the degree of inhibition conferred by lysates from the different strains. The pattern did not correlate with any particular capsular serotype. These data establish the ability of multiple strains of Spn to inhibit CD8⁺ T cell responses.

To identify the component responsible for the inhibitory activity, we subjected the lysate to heat (100°C) and proteinase K treatment and found both of these treatments completely ablated the ability of the Spn lysate to inhibit IFNγ production by CTL (data not shown). Based on these results supporting the proteinaceous nature of the component, we passed the Spn lysate over an anion exchange column under constant flow. Fractions were collected every minute. Based on the elution pattern of protein, fractions 13–50 were tested for inhibitory activity in our peptide restimulation assay. Maximal inhibitory activity was observed in fractions 31–33 (Figure 4A). These fractions were pooled and further purified by passage over a hydroxyapatite column. The protein-containing fractions (9–14) were tested for inhibitory activity. The majority of activity was found in fractions 11–13 with maximal activity in fraction 12 (Figure 4B).

Fractions 11–13 were run on a denaturing polyacrylamide gel and proteins visualized by Coomassie staining. Each fraction contained multiple proteins, with the majority of protein eluting in fraction 11, despite maximal inhibition occurring with fraction 12 (Figure 4C). Upon close inspection, the intensity of a band at approximately 100 kDa (Figure 4C, arrow) was observed to be most correlative with inhibitory activity. This band was excised, extracted, and subjected to liquid chromatography and mass spectrometry to identify protein candidates, one of which was the pneumococcal metalloenzyme PepN.

To determine whether Spn PepN was in fact responsible for the inhibition of CTL effector function, the protein was produced in E. coli. PepN has been described in other closely related bacterial species, e.g., Streptococcus thermophilus, but the pneumococcal PepN has not been characterized to date. We designed primers in the intergenic regions surrounding the PepN gene based on genetic alignment of strains of pneumococcus for which sequences are available. Following PCR amplification and sequencing to confirm identity, the pepn gene was cloned from EF3030 into the E. coli-based protein overexpression vector pTHC-ΔP. This vector is engineered to add a sequence encoding a 6× His tag to the N terminus of the cloned gene. Following transformation, E. coli positive for the pepn containing vector were selected and cultured under conditions that induced the production of PepN.

As a first step, we determined whether the lysate prepared from E. coli expressing PepN exhibited inhibitory activity. NP-specific CTL were stimulated with peptide in the presence of lysates from Spn, empty vector transformed E. coli, or E. coli expressing PepN-His. As expected, the lysate derived from E. coli transformed with the empty vector induced minimal inhibition of IFNγ (Figure 5A). In agreement with our previous results, significant inhibition of IFNγ production was observed following treatment with Spn lysate (Figure 5A, black bar). Addition of lysate prepared from E. coli expressing PepN also resulted in robust inhibition of cytokine production (Figure 5A, hatched bar), supporting the hypothesis that PepN mediates inhibition of CD8⁺ CTL effector function.

Given the evidence of PepN inhibitory activity in the lysate of E. coli transformed with the PepN vector, PepN was purified by passage over a nickel column. The His tag was subsequently cleaved. The purified proteins were assessed by gel electrophoresis (Figure 5B). The difference in mobility between the His-tagged protein (PepN-His) and the cleaved protein (PepN) indicated successful removal of the tag.

In order to establish that PepN from Spn was sufficient for mediating the inhibition of CTL effector function, we added titrated amounts of purified PepN versus Spn lysate during stimulation. To control for non-specific effects of adding purified E. coli-expressed protein, we used Bacillus anthracis coenzyme A-disulfide reductase (BACoADR) and l-α-glycerophosphate oxidase (GlpO), which were purified in a similar manner to PepN. As shown in Figure 6A, increasing amounts of Spn lysate resulted in a corresponding increase in inhibition of IFNγ production following peptide stimulation. The Spn lysate was capable of inhibiting approximately 50% of the response with 0.75 µg. In contrast, the purified PepN exhibited 50% inhibition of IFNγ production at 0.025 µg of PepN, which corresponds to an approximately 33-fold increase in the activity compared to the lysate. Inhibition was not a non-specific effect as the addition of 3 µg of BACoADR or GlpO did not alter the production of IFNγ (Figure 6A). The loss of IFNγ production was associated with a failure to produce IFNγ mRNA (Figure 6B). While peptide stimulation resulted in a 349-fold increase, message levels increased by only 9.6-fold when PepN was present, supporting regulation at the level of transcription.

To determine whether PepN treatment had cytotoxic effects in our T cell function assay, cells were stained with 7-AAD as previously described. As shown in Figure 6C, PepN did not induce a significant increase in 7-AAD positivity at any of the concentrations tested. These data show PepN-mediated inhibition occurs in the absence of cell death.

We hypothesized the ability of PepN to inhibit IFNγ was dependent on its known enzymatic activity. Like other metalloproteases, PepN activity relies on metal ion interaction with the binding pocket in the enzymatic domain of the protein (42). In order to test the dependence on peptidase activity, we generated a mutant PepN that was incapable of binding zinc as a result of mutation of His292Ala, Glu293Ala, and His296Ala (PepN^Met−) (43). Following generation of the PepN mutant, we tested enzymatic activity by measuring the cleavage of lysine from an aminomethylcoumarin (AMC) substrate. Trypsin served as a positive control. As shown in Figure 7A, trypsin treatment resulted in robust hydrolysis of lysine-AMC. Purified PepN also hydrolyzed the lysine-AMC substrate. In contrast, PepN^Met− failed to hydrolyze the substrate (Figure 7A). When normalized to the relative activity of wild-type PepN, the metal negative mutant exhibited only 0.3% activity (Figure 7B) confirming that the mutations in the zinc-binding domain rendered PepN enzymatically inactive.

To test the inhibitory activity of the PepN mutant, CTL were stimulated in the presence or absence of PepN^Met− and IFNγ production was measured. As expected, PepN inhibited nearly all IFNγ production induced by peptide stimulation (Figure 7C). Surprisingly the enzymatically inactive PepN exhibited similar inhibitory activity. Thus, the ability of PepN to inhibit CTL effector responses is not dependent on its aminopeptidase activity, demonstrating a novel function of this protein.

Given the failure to produce IFNγ, we determined whether PepN treatment impaired signaling through the TCR. One of the earliest events to occur following TCR engagement is phosphorylation of the kinase ZAP-70 [for review see Ref. (44)]. To assess activation of ZAP-70, CTL were stimulated with peptide-pulsed splenocytes for 15 min in the presence or absence of PepN at which point cells were fixed. Peptide stimulation induced a robust increase in the MFI of phospho-ZAP-70 which was inhibited in the presence of PepN (Figures 8A,B), indicating PepN treatment prevents initiation of TCR signaling.

Given the early nature of this signaling defect, we hypothesized that we could restore function through direct stimulation of downstream mediators of T cell activation using PMA and IONO. PMA and IONO directly activate protein kinase C-θ and induce Ca⁺ flux, respectively. Figure 8C shows that PMA + IONO stimulation induced a robust IFNγ response in effector cells that was blocked by the presence of PepN. The inability of PMA + IONO treatment to promote IFNγ production in the presence of PepN suggested this immunoregulatory protein also inhibited distal molecules in the TCR signaling pathway.

Given the failure of PMA + IONO to overcome the negative effects of PepN, we evaluated the signaling mediators distal to the targets of PMA and IONO action, i.e., ERK1/2, JNK, and the p38 mitogen-activated protein kinase (44). Phosphorylation of these molecules mediates the activation and nuclear translocation of transcription factors where they promote production of mRNA. We hypothesized that PepN treatment impaired activation of these molecules. To test this, CTL were stimulated with PMA and IONO in the presence or absence of PepN, followed by staining with antiphospho-ERK1/2, antiphospho-JNK, and antiphospho-p38 antibodies. We observed increases in the level of p-ERK1/2, p-p38, and p-JNK following 15 min of stimulation with PMA and IONO (representative data shown in Figure 9A and averaged data in Figures 9B–D). In contrast, CTL treated with PepN showed significantly reduced levels of these phosphorylated molecules. Interestingly, the amount of phospho-ERK1/2 in PepN-treated cells was significantly below that in non-stimulated cells (approximately threefold) (Figure 9E), suggesting the ability of PepN to modulate basal levels of p-ERK1/2. To determine whether the reduced level of phosphorylated protein was the result of protein degradation, we evaluated total ERK1/2 as a representative signaling molecule. No decrease was observed following PepN treatment (Figure 9E). Together these data show that PepN regulates T cell effector function by inhibiting phosphorylation of the TCR signaling cascade at multiple steps. This raised the interesting possibility that signaling was generally inhibited by PepN treatment.

The Akt pathway is activated downstream of signaling through TCR and is responsible for regulation of cellular metabolism. This pathway operates via the global regulator mTOR. When CD8⁺ T cells are activated, they are placed under metabolically intensive conditions. Glycolysis, amino acid uptake, and protein synthesis must all increase in order to facilitate appropriate effector function (45). Thus, regulation of this pathway has a profound impact on the cell. To assess mTOR activation, NP-specific CTL were stimulated for 15 min with PMA and IONO in the presence or absence of PepN as above. Following stimulation, a rapid increase in the levels of phospho-mTOR was observed (Figures 9F,G). In stark contrast to the previous signaling molecules analyzed, there was no effect of PepN treatment on mTOR phosphorylation (Figures 9F,G). These data argue that the effect on TCR signaling by PepN is not global in nature, but instead impacts selected molecules/arms of the pathway.

We were interested in determining whether the negative effects on signaling extended beyond the TCR signal transduction pathway. To test this hypothesis, we examined signaling through the type I IFN receptor by assessing STAT1 phosphorylation [for review see Ref. (46)]. Treatment with type 1 IFN resulted in an increase in the level of phospho-STAT1 that was significantly inhibited by PepN (Figures 9H,I). Thus, regulation of effector cells by PepN is not limited to the TCR signaling pathway.