Priming of Protective Anti-Listeria monocytogenes Memory CD8⁺ T Cells Requires a Functional SecA2 Secretion System

Abstract

The SecA2 auxiliary secretion system of Gram-positive bacteria promotes the export of virulence proteins essential for colonization of the host in the case of both Mycobacterium tuberculosis and Listeria monocytogenes, two intracellular bacteria causing diseases in humans. We and others have demonstrated that this secretion system is also linked to the onset of long-term CD8⁺ T cell-mediated protective immunity in mice. In the case of L. monocytogenes, expression of SecA2 inside the cytosol of infected cells correlates with the generation of CCL3-secreting memory CD8⁺ T cells that are required for protection against secondary challenge with wild-type (wt) L. monocytogenes. Since the SecA2 ATPase is well conserved among Gram-positive pathogenic bacteria, we hypothesized that SecA2 itself bears evolutionarily conserved motifs recognized by cytosolic pattern recognition receptors, leading to signaling events promoting the differentiation of CCL3⁺ memory CD8⁺ T cells. To test this possibility, we generated a stable L. monocytogenes chromosomal mutant that expressed a SecA2 ATPase bearing a mutated nucleotide binding site (NBS). Similarly to a SecA2 deletion mutant, the NBS mutant exhibited rough colonies, a bacterial chaining phenotype, an impaired protein secretion profile, and in vivo virulence in comparison to wt L. monocytogenes. Importantly, mice immunized with the SecA2 NBS mutant were not protected against secondary infection with wt L. monocytogenes and did not develop CCL3⁺ memory CD8⁺ T cells. NBS mutant and wt SecA2 proteins were expressed to comparable extents by bacteria, suggesting that SecA2 itself is unlikely to promote the induction of these cells. Rather, one or several of the SecA2 substrate proteins released inside the cytosol of infected cells may be involved.

INTRODUCTION

Induction of an efficient T and B cell-mediated adaptive immune response inside infected hosts requires several critical events that involve pathogen detection and the combination of adequate signals of activation. The amount of antigen and the duration of presentation (signal 1), the extent of costimulatory molecule upregulation (signal 2), and the inflammatory cytokines and chemokines that are secreted at the time of peptide presentation to T cells (signal 3) are critical parameters orchestrating the differentiation of T cells, e.g., their capacity to express distinct effector functions and to develop into memory T cells that confer protective immunity (13, 37, 43). Signals 2 and 3 depend mostly on danger signals which are given by highly conserved microbial molecules derived from pathogens that are sensed by innate immune sentinel cells such as dendritic cells (DCs) or macrophages from infected hosts. Detection of such pathogen-associated molecular patterns (PAMPs) is mediated by pattern recognition receptors (PRRs) that include cell surface- or intracellular compartment-expressed receptors such as the Toll-like receptors (TLRs) and intracytosolic receptors such as the Nod-like receptors (NLRs). Many reports have linked the set of PRRs expressed and triggered onto or inside antigen-presenting cells (APCs) scattered throughout the body to the shaping of primary T cell responses (29), and it is likely that these activation pathways may also affect the quality of the memory T cells generated.

Listeria monocytogenes is a Gram-positive intracellular bacterium that has been widely used to study the host innate and adaptive immune responses in mice. L. monocytogenes bears multiple TLR ligands that can trigger TLR2, which is involved in the recognition of bacterial peptidoglycan (PGN), lipoteichoic acid, and lipoproteins (39, 40), or TLR5, which recognizes bacterial flagellin (15). Other PRRs, such as the NLRs Nod1, Nod2, and NALP3, can also detect L. monocytogenes-derived motifs (19, 20, 27, 34, 42). Seminal studies from Portnoy's laboratory have shown that a TLR-independent cytosolic surveillance system involving NLR proteins (e.g., Nod2) is efficiently triggered upon bacterial entry into the cytoplasm and leads to a specific Myd88-independent transcriptional response that is composed largely of secretion of beta interferon (IFN-β) and IFN-responsive gene products (23, 28, 35). Following L. monocytogenes inoculation in mice, conventional DCs undergo maturation very early after inoculation and prime adaptive CD8⁺ T cells to further differentiate into protective memory CD8⁺ T cells that mediate protective immunity against a secondary challenge infection (18, 22).

Study of bacteria lacking the SecA2 auxiliary secretion system demonstrated that this pathway directs the secretion of a specific subset of proteins that is involved in the pathogenicity of L. monocytogenes as well as that of Mycobacterium tuberculosis, although through a priori distinct mechanisms (7, 24) that both modulate the host immune inflammatory and effector responses. In fact, secretion systems of bacteria are critical protein structures that are responsible for the transport of proteins across hydrophobic membranes to their site of action (36). Most bacteria possess the highly conserved general secretion Sec pathway, which is necessary for viability and mediates the translocation across the bacterial membrane of proteins containing N-terminal signal recognition sequences. In the case of M. tuberculosis, SecA2 deficiency affects the ability of this bacterium to secrete superoxide dismutase A (SodA), a critical enzyme involved in M. tuberculosis escape and survival inside infected host cells by eluding the oxidative attack of the macrophages (21). In the case of L. monocytogenes, at least 19 proteins are secreted through this auxiliary secretory pathway (2, 10, 24, 25); among them, the L. monocytogenes murein hydrolases p60 and NamA are the two major SecA2 substrate proteins and account for the highly attenuated virulence in vivo of secA2 deletion mutant L. monocytogenes compared to the wild-type (wt) bacteria (∼500-fold) (24–26, 31). Both hydrolases can degrade L. monocytogenes PGN during growth, and the observed attenuated virulence most likely results from the impaired release of NLR ligands inside the cytosol of infected host cells, which subverts host pattern recognition.

Recently, using L. monocytogenes deficient in SecA2, we further found that mice immunized with this mutant formed memory T cell populations equivalent to those of mice that received wt bacteria (31, 32). However, these cells were suboptimally primed and did not differentiate into protective (CCL3-secreting) memory cells, which are required to confer protection against a secondary infection with wt L. monocytogenes (31, 32). Interestingly, data from Porcelli's laboratory have also linked the presence of SecA2 from the intracellular bacterium M. tuberculosis to the onset of CD8⁺ T cell-mediated immunity (16, 32). However, in that setting, the suggested molecular mechanisms involved in the priming of protective T cell-mediated responses were related to increased apoptosis of infected APCs, which favored more efficient cross-presentation and priming of M. tuberculosis-specific CD8⁺ T cells.

While only a few bacteria encode multiple SecA ATPases in their genome, all bacteria that possess such systems are Gram positive and exhibit pathogenicity toward their host. Since the SecA2 ATPase is well conserved among the Gram-positive pathogenic bacteria, we hypothesized that SecA2 itself could bear evolutionarily conserved PAMPs that are recognized by PRRs (NLRs) inside the cytosol of infected cells, leading to signaling events adequate to promote CCL3⁺ memory CD8⁺ T cell development. Alternatively, one or several of the SecA2-secreted proteins may be presenting or exhibiting an enzymatic activity that generates such PAMPs from the bacteria to trigger cytosolic PRRs. To discriminate between these hypotheses, we have generated stable, nonfunctional L. monocytogenes strains with chromosomal mutations of the SecA2 ATPase that have lost the capacity to promote the secretion of the SecA2 substrate proteins.

MATERIALS AND METHODS

Mice and bacteria.

Six- to 8-week-old BALB/cByJ female mice (Janvier Laboratories) were used in all experiments. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Committee of Animal Care and Use of the Regional Cote d'Azur. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Institut de Pharmacologie Moléculaire et Cellulaire (permit number B-06-152-5, delivered by the Veterinary Services of the Alpes-Maritimes Prefecture) and by the animal use committees at the Albert Einstein College of Medicine. All efforts were made to minimize suffering and provide humane treatment to the animals included in the study.

The L. monocytogenes 10403s strain was used in all experiments. The wt strain, a mutant expressing a SecA2 ATPase bearing a mutated nucleotide binding site (NBS-SecA2 mutant), and a SecA2 deletion mutant (Del-SecA2 mutant), expressing or not expressing the FLAG epitope, were on the 10403s L. monocytogenes strain background. The wt bacteria and the Del- and NBS-SecA2 mutants exhibit 50% lethal doses (LD₅₀s) of 3 × 10⁴, 10⁷, and 10⁷, respectively, in BALB/c mice (24, 31). All wt bacteria were prepared from clones grown from organs of infected mice. Stocks of bacteria were kept frozen at −80°C.

Site-directed mutagenesis of SecA2 nucleotide binding sites (NBS).

The wt secA2 gene was amplified from wt L. monocytogenes genomic DNA by PCR using the HF-Advantage kit (Clontech) and the primers 5′-AAACTGCAGTACCATCGCGTCCAGATGCATGACC-3′ and 3′-CGGGGTACCCAGTGCTTGCGGATGCGATTGTTGG-5′, which include the PstI/KpnI restriction sites for cloning into the pKSV7 transfer vector. The mutagenesis was performed using the QuikChange II XL site-directed mutagenesis kit (Stratagene) according to the manufacturer's guidelines and with the following primers: for NBS-I, 5′-GTGGCAGAAATGAAAACCGTCGAAGGTATAACATTGATGTCG and 3′-CGACATCAATGTTATACCTTCGACGGTTTTCATTTCTGCCAC, and for NBS-II, 5′-CTGGAATGACAGTAACAGCTATAACTGAGGAAGAAGAATTCCGTC and 3′-GACGGAATTCTTCTTCCTCAGTTATAGCTGTTACTGTCATTCCAG. Using this method, four mutations were introduced in the secA2 gene: (i) in NBS-I of SecA2, Gly98 and Lys101 were changed to valine (G98V) and isoleucine (K101I), respectively, and (ii) in NBS-II of SecA2, Gly366 and Lys369 were changed to valine (G366V) and isoleucine (K369I), respectively. The introduced mutations were verified by sequencing (Applied Biosystems 3100).

For FLAG-expressing L. monocytogenes, two PCRs (HF-Advantage II; Clontech) were performed on wt L. monocytogenes genomic DNA to insert the FLAG sequence in frame at the C terminus of the secA2 gene inside the pKSV7 transfer vector and in the KpnI/PstI cloning sites. The first PCR used primers 5′-agtcGGTACCGGATGTCGTGGTAATCCCTACC and 3′-CCTTACTTGTCATCGTCATCCTTGTAATCGCCTTGGATTAAGCCGTCTG, and the second PCR used 5′-ATACAACTTGCGTCTGATTTTTTCCGATTACAAGGATGACGATGACAAGTAACC and 3′-CTGCAGCCGAATAGATTTCCGGCTACC. The pKSV7/NBS-SecA2 and pKSV7/SecA2-FLAG constructs were then digested with KpnI/BamHI and BamHI/PstI, respectively, and concomitantly inserted inside the KpnI/PstI cloning sites of pKSV7 to generate a hybrid construct bearing the Walker site mutations and the C terminus FLAG. All constructs were verified by sequencing.

Cloning and selection of an L. monocytogenes chromosomal mutant that expresses a SecA2 ATPase bearing a mutated NBS.

The mutated NBS-secA2 fragments (expressing FLAG or not) were then cloned into the pKSV7 transfer plasmid, which carries resistance to ampicillin in Escherichia coli (100 μg/ml) and to chloramphenicol (Cam) in L. monocytogenes (10 μg/ml) and also possesses a temperature-sensitive replication origin with a restrictive temperature (41°C). The recombinant pKSV7-NBS-secA2 was transformed into L. monocytogenes by electroporation, and transformed bacteria were selected at 37°C on Cam-brain heart infusion (BHI) (Sigma) plates. Overnight (O/N) liquid culture at a permissive temperature (30°C) with Cam, followed by two O/N steps at 41°C with Cam, forced plasmid integration. A single colony that integrated the plasmid was subsequently cultured O/N at 30°C for all excision and recombination steps. After subcultures at 30°C without selection, growing L. monocytogenes clones that harbored the correct mutations and/or the FLAG epitope inside the chromosome-encoded SecA2 were verified by DNA sequencing.

Light microscopy and EM images.

For light microscopy images of bacterial colonies, the distinct mutants and wt L. monocytogenes were grown on BHI agar plates at 37°C O/N, and the colonies were analyzed at a magnification of ×10. For electron microscopy (EM) studies, bacteria (NBS- and Del-SecA2 mutants and wt) were grown in BHI to log phase (optical density at 600 nm [OD₆₀₀] of ∼0.2 to 0.3), pelleted by centrifugation (4,000 × g), and fixed with 1.6% glutaraldehyde in 100 mM phosphate buffer (pH 7.5). Pellets of bacteria were incubated at 4°C in 100 mM cacodylate buffer (pH 7.0) containing 1% osmium tetroxide, washed with distilled water, and incubated for 2 h with 0.5% uranyl acetate buffer at room temperature in the dark. The pellets were then washed in water, dehydrated in an increasing acetone series, and embedded in epoxy resin. Blocks were thin sectioned using standard procedures and contrasted for observation on a Philips CM12 electron microscope.

Western blotting. (i) L. monocytogenes secretome.

Proteins from O/N culture supernatants of NBS- and Del-SecA2 mutants and wt bacteria were precipitated by treatment with trichloroacetic acid (TCA) at a final concentration of 6% on ice for 30 min. Precipitates were washed with cold acetone, dried, and suspended in SDS-PAGE loading buffer (50 mM Tris-Cl [pH 6.8], 2% SDS, 10% glycerol, 0.1% bromophenol blue) containing 5% dimethyl sulfoxide (DMSO). Samples were boiled for 5 min, run on an SDS–10% polyacrylamide gel, transferred to nitrocellulose, and probed with Listeria rabbit antiserum (BD Difco). Goat anti-rabbit IgG-peroxidase (Beckman Coulter) was used as a secondary probe, and bound antiserum was revealed using the Super Signal West Femto kit (Pierce).

(ii) Expression of wt and NBS mutant SecA2.

Bacteria were grown in brain heart broth medium (Fluka) to exponential phase, and ∼10¹⁰ bacteria were pelleted, washed in phosphate-buffered saline (PBS)–1 mM phenylmethylsulfonyl fluoride (PMSF), and incubated in 40 mM Tris-HCl–150 mM NaCl containing 10 mg/ml chicken egg white lysozyme M (Sigma) at 37°C for 90 min in the presence of a cocktail of protease inhibitors (Thermo Scientific). The bacterial suspension was then sonicated (Misonix Ultrasonic Processors) with 1-min pulses for 20 min, pelleted, resuspended in PBS–1% SDS, and incubated for 30 min at 37°C. Approximately 3.3 × 10⁸ bacteria from the initial culture were loaded in SDS-PAGE sample buffer (62.5 mM Tris [pH 6.8], 2% SDS, 10% glycerol, 0.1 M dithiothreitol [DTT], 0.01% bromophenol blue) and separated on an 8% SDS-polyacrylamide gel. Proteins were then transferred to a nitrocellulose membrane, probed with a horseradish peroxidase (HRP)-coupled anti-FLAG monoclonal antibody (MAb) (clone M2; Sigma), and further revealed (Chemiglow West substrate) using the FluorChem FC2 imaging system (Alpha Innotech). Quantification was performed with AlphaView software.

Infection of mice and measurement of protective immunity.

For L. monocytogenes infections, bacteria were grown to logarithmic phase (OD₆₀₀ of ∼0.05 to 0.15) in BHI medium diluted in PBS and injected intravenously (i.v.) into lateral tail vein. In all experiments, mice were given a primary immunization with a 0.1× the LD₅₀ of bacteria. Secondary infections were carried out 1 month later with 10× the LD₅₀ of wt bacteria (3 × 10⁵). To measure bacterial titers in spleen and liver, organs were harvested and dissociated on metal screens in 10 ml of 0.1% X-100 Triton (Sigma). Serial dilutions were performed in the same buffer, and 100 μl was plated onto BHI medium plates.

Cell suspensions.

Organs were cut in very small pieces and incubated at 37°C for 20 min in Hanks balanced salt solution (HBSS) (Gibco) containing 4,000 U/ml of collagenase I (Gibco) and 0.1 mg/ml of DNase I (Roche). Red blood cells were lysed for 2 to 3 min in 170 mM NH₄Cl-17 mM Tris HCl (pH 7.4) and resuspended in fluorescence-activated cell sorter (FACS) buffer (PBS with 0.5% bovine serum albumin [BSA]) for further stainings. For intracellular CCL3 staining, all the steps were performed in the presence of 2 μg/ml of Golgi Plug (BD Pharmingen).

Antibodies and reagents.

The following MAbs were purchased from BD Pharmingen: anti-CD8α (53-6.7)-fluorescein isothiocyanate (FITC), -phycoerythrin (PE), -peridinin chlorophyll protein (PerCP), or -allophycocyanin; anti-CD3ε-FITC (145-2C11); and anti-CD62L-allophycocyanin (MEL-14). Donkey anti-rabbit-Alexa 647 was purchased from Molecular Probes. Goat polyclonal anti-CCL3 Ab was purchased from R&D systems. PE-conjugated LLO_91-99/H2-K^d tetramers were obtained from the NIH tetramer core facility (Emory).

Flow cytometry.

Cells were stained with the specified Abs in 100 μl of PBS containing 0.5% BSA (FACS buffer). For intracellular staining, splenocytes were incubated at 37°C with 5% CO₂ for 5 to 6 h in RPMI 1640 (Gibco) containing 5% FCS, 2 μg/ml Golgi Plug (BD Pharmingen), and 1 nM LLO_91-99 peptide (Mimotopes). Cells were then harvested, incubated for 20 min on ice with the indicated cell surface marker MAbs, fixed in 1% paraformaldehyde (PFA)-FACS buffer for 20 min on ice, permeabilized, and concomitantly stained for intracellular CCL3 (anti-CCL3 goat polyclonal Ab) for 30 min in 1× Perm/Wash (BD Pharmingen). CCL3 staining was revealed using donkey anti-goat-Alexa 647 MAb. In all cases, cells were washed in FACS buffer and analyzed on a FACSCalibur cytofluorimeter (Becton Dickinson).

Statistical analysis.

Statistical significance was calculated using an unpaired Mann-Whitney test and Instat software. All P values of 0.05 or less were considered significant and are referred to as such below.

RESULTS

Generation of an L. monocytogenes chromosomal mutant that expresses a SecA2 ATPase bearing mutated nucleotide binding sequences.

Previous reports have shown that both the human transporter associated with antigen processing (TAP) and M. tuberculosis SecA2 can be functionally inactivated by mutating at least one Walker motif, which prevents ATP binding and subsequent hydrolysis (17, 38). Therefore, to generate an L. monocytogenes strain that expresses a nonfunctional SecA2 ATPase, we performed targeted mutations inside the two Walker-type nucleotide binding sites (NBS-I and NBS-II) of the chromosomal gene encoding SecA2. We mutated the same two key positions putatively involved in ATP binding by replacing two glycines (mutations G98V and G366V) and two lysines (mutations K101I and K369I), respectively, in NBS-I and NBS-II of SecA2 (Fig. 1). Recombinant bacteria in which the wt NBS had been replaced by the mutated ones by homologous recombination were further selected by successive integration/excision/recombination cycles of the pKSV7 transfer vector bearing the mutagenized SecA2 cDNA. The presence of the mutations inside the bacterial chromosome was verified by sequencing.

Fig. 1.

Schematic representation of the mutated positions in the nucleotide binding sites (NBS) of SecA2. The main secA2 gene coding regions are represented by NBS-I and -II (which bind ATP), PPB (which defines a putative binding region of SecA2-dependent proteins), and CD (which is a conserved domain of SecA proteins). The nucleotides modified by site-directed mutagenesis and the corresponding amino acids changes are indicated on the detailed NBS-I and -II sequences.

The deletion and NBS SecA2 mutants exhibit similar phenotypes.

We next analyzed the phenotype of the L. monocytogenes strain bearing the NBS mutations in the secA2 open reading frame (NBS-SecA2 mutant). NBS-SecA2 bacteria grown overnight on BHI agar plates formed rough colonies with jagged edges, similar to those of the SecA2 deletion mutant (Del-SecA2) and in striking contrast to the smooth edges of wt bacterial colonies (Fig. 2A). Electron microscopy (EM) pictures showed that NBS-SecA2 L. monocytogenes formed long filaments consisting of three to five individual bacteria exhibiting joined cell walls, a phenotype also found in Del-SecA2 bacteria but not in wt bacteria, which were mostly single or double units during division (Fig. 2B). This rough colony bacterial chaining phenotype resulted from the impaired secretion of the two main SecA2 substrate proteins, the L. monocytogenes autolysins p60 and NamA (24–26). Since the secretion of at least 17 other proteins was also affected in this mutant, we then compared the secretomes of in vitro-cultured NBS-SecA2, Del-SecA2, and wt bacteria (Fig. 2C). While some L. monocytogenes-derived proteins exhibited equal intensities in the supernatants from all bacteria, many other proteins detected in the culture supernatant of wt bacteria were reduced in NBS- and in Del-SecA2 bacteria to comparable extents and as previously described (24). Next, we tested whether the NBS-SecA2 mutant of L. monocytogenes was also impaired in its virulence in vivo, as described for the Del-SecA2 mutant (24, 31). We found that the mutants exhibit comparable LD₅₀s in mice (10⁷ bacteria) (data not shown), a result in agreement with previously published reports on Del-SecA2 L. monocytogenes by us and others. Collectively our data strongly suggest that the point mutations inside both NBSs of SecA2 give a phenotype comparable to that caused by deletion of the secA2 gene. Therefore, mutating these sequences has led to a nonfunctional auxiliary SecA2 secretion system.

Fig. 2.

Phenotypic and functional characterization of the NBS-SecA2 mutant of L. monocytogenes. (A) Morphology of wt, Del-SecA2, and NBS-SecA2 L. monocytogenes colonies grown on BHI agar plates at 37°C O/N as seen by light microscopy (magnification, ×10). wt bacteria exhibit smooth edges, whereas both SecA2 mutants (Del-SecA2 and NBS-SecA2) have craggy/rough edges. (B) Electron micrographs of wt, Del-SecA2, and NBS-SecA2 L. monocytogenes cultured at log phase. Left and right panels show low- and high-magnification views, respectively, of mutants and wt bacteria. Both mutants exhibit long filaments formed by multiple bacterial units, in contrast to individualized bacteria formed by wt L. monocytogenes. (C) SDS-PAGE analysis of TCA-concentrated stationary-phase culture supernatants from O/N-grown NBS-SecA2, Del-SecA2, and wt L. monocytogenes. Proteins were visualized by immunoblotting using a Listeria-specific antiserum.

Expression levels of wt and NBS mutant SecA2 proteins inside bacteria are comparable.

To rule out that the phenotype of NBS-SecA2 L. monocytogenes did not result from poor expression of the mutated SecA2 protein, we produced new recombinant wt and NBS-SecA2 L. monocytogenes strains that expressed a chromosome-encoded FLAG epitope at the C terminus of the translated SecA2 proteins, which allows for SecA2 detection (Fig. 3A). For this, we generated a construct including both mutated NBS-I and -II and an in-frame C terminus FLAG epitope, which was used to transfect wt L. monocytogenes and select for recombinant bacteria. With this approach, recombination could occur and generate both wt and NBS SecA2 FLAG-expressing bacteria. In fact, three wt and three NBS clones bearing a “flagged” SecA2 were obtained as the result of independent recombination events, which was verified by sequencing as well as smooth and rough colony formation. As documented earlier (Fig. 2A) (25), while all wt L. monocytogenes exhibited wt NBS and formed smooth colonies, the NBS mutants had either one or two mutated NBS and formed rough colonies. Next, we analyzed the protein expression levels of wt or NBS mutant SecA2 inside flagged bacteria. For this, recombinant or wt (control) bacteria were exponentially grown, and the whole bacterial pellet was extracted, Western blotted, and revealed using an anti-FLAG specific MAb (Fig. 3B). The results revealed that wt, NBS-I-mutated, and NBS-I- and -II-mutated SecA2 proteins were detected at the expected molecular mass (∼85 kDa) (not shown) and expressed at equivalent levels in all independently selected recombinant bacterial clones. Collectively these experiments therefore suggested that mutating the NBS inside the SecA2 protein affects primarily its ATPase function but not its expression levels inside the bacteria. Of note, mutating the NBS-I Walker site seems to be sufficient for the observed phenotype.

Fig. 3.

The wt and NBS-deficient SecA2 proteins are expressed to equivalent levels in L. monocytogenes. (A) Summary of independently obtained recombinant L. monocytogenes clones that express chromosome-encoded wt (+) or NBS-mutated (−) SecA2-FLAG proteins. While wt SecA2-FLAG bacteria formed smooth (S) colonies, NBS-SecA2-FLAG bacteria formed rough (R) colonies on BHI plates. (B) Western blot analysis of SecA2 expression in three distinct wt or NBS-mutated bacterial clones encoding or not encoding (wt control) a C-terminal SecA2-FLAG protein. The cell pellet fraction of ∼3.3 × 10⁸ log-phase bacteria was separated by SDS-PAGE, blotted, and revealed using a FLAG-specific MAb. The results shown are representative of two independent experiments that gave comparable results.

Mice immunized with the nonfunctional NBS SecA2 mutant of L. monocytogenes are not protected against a secondary infection with wt L. monocytogenes.

To assess if the SecA2 protein by itself or one (or several) of its substrate proteins is involved in inducing long-term protective immunity, we immunized mice with 0.1× the LD₅₀ of NBS-SecA2, NBS-SecA2-FLAG, Del-SecA2, or wt bacteria or with PBS. One month later, animals were challenged with 10× the LD₅₀ of wt bacteria and monitored for bacterial titers in the spleen and the liver (Fig. 4). As expected, animals that received PBS during primary injection (unprotected group) exhibited 130,961- and 9,012-fold more bacteria in the spleen and the liver, respectively, than mice primarily immunized with wt bacteria (protected group). As previously shown (31), mice immunized with the Del-SecA2 mutant failed to control the secondary infection and exhibited 6,643-fold (spleen) and 579-fold (liver) more bacteria than the protected mice. Likewise, animals that received either NBS-SecA2 or NBS-SecA2-FLAG L. monocytogenes during primary immunization exhibited 10,827- and 2,374-fold (spleen) and 1,921- and 40-fold (liver) more bacteria than protected mice and did not survive the secondary infection (data not shown). Taken together with the SecA2 protein expression data (Fig. 3), our results support the hypothesis that it is not the SecA2 ATPase itself but at least one of the SecA2 substrate proteins that is involved in the induction of protective memory CD8⁺ T cell-mediated immunity.

Fig. 4.

Mice immunized with NBS-SecA2 L. monocytogenes are not protected against secondary infection with wt bacteria. Mice (four or five per group) were immunized with 0.1× the LD₅₀ of the indicated bacteria (wt, 2 × 10³; Del-SecA2, NBS-SecA2, and NBS-SecA2-FLAG, 10⁶) or injected with PBS. One month later, mice were given a secondary challenge with 3 × 10⁵ wt bacteria and killed 48 h after the challenge infection. Data show the number of bacteria in the spleen and the liver for each individual mouse and represent pooled results from two experiments. P values were calculated between groups of mice immunized with wt versus NBS-SecA2 and NBS-SecA2-FLAG L. monocytogenes and PBS versus Del-SecA2 and NBS-SecA2 L. monocytogenes in the experiment shown. P values close to 0.05 have low statistical significance.

Induction of CCL3-secreting memory CD8⁺ T cells is impaired in mice immunized with NBS-SecA2 L. monocytogenes.

In a previous study, we showed that in contrast to mice receiving primary immunization with wt L. monocytogenes, animals immunized with Del-SecA2 L. monocytogenes lacked protective CCL3-secreting memory CD8⁺ T cells early after secondary infection (32). To assess whether the lack of immunological protection in mice primarily immunized with NBS-SecA2 L. monocytogenes correlates with the absence of the CCL3-secreting memory CD8⁺ T cells, we monitored the early secondary CD8⁺ T cell response in mice primarily immunized with NBS-SecA2, Del-SecA2, or wt L. monocytogenes or injected with PBS. Spleen cells from these groups of mice were harvested at 6 h after secondary infection and analyzed by FACS after staining with LLO_91-99/H2-K^d tetramers and anti-CD62L, a cell surface marker that defines effector memory cells (Fig. 5). As expected, while PBS-injected animals had only background levels of L. monocytogenes LLO_91-99/H2-K^d-specific CD8⁺ T cells (∼0.03 to 0.05%), mice immunized with wt and Del-SecA2 L. monocytogenes exhibited equivalent frequencies of these cells (0.57 and 0.51%, respectively), which secreted IFN-γ and tumor necrosis factor alpha (TNF-α) and expressed granzyme B (data not shown), a result in agreement with our previous reports (31, 32). We found similar frequencies of LLO_91-99/H2-K^d-specific CD8⁺ T cells in animals that received NBS-SecA2 bacteria (∼0.47%). Likewise 59.7 to 67.4% of these cells had lost CD62L expression in all groups of mice.

Fig. 5.

Effector memory but not CCL3-secreting CD8⁺ T cells are present during secondary infection in mice given a primary immunization with NBS-SecA2 L. monocytogenes. BALB/c mice (10 to 12 mice per group) were immunized with PBS or 0.1× the LD₅₀ of the indicated bacteria and challenged 30 days later with 10× the LD₅₀ of wt bacteria. At 6 h after secondary challenge, spleen cells from individual mice were stained for expression of the cell surface markers CD8, CD3, CD62L, and H2-K^d/LLO_91-99 tetramers and/or restimulated in vitro with LLO_91-99 peptide for intracellular CCL3 staining and further processed for FACS analysis. Dot plots are representative of the staining obtained for CD8⁺ T cells expressing H2-K^d/LLO_91-99 tetramers. The right upper panel shows the mean frequency (± standard error of the mean [SEM]) of H2-K^d/LLO_91-99 tetramer-positive cells among the total CD8⁺ T cells. The left lower panel shows the mean frequency of H2-K^d/LLO_91-99 tetramer-specific CD62L^low (effector memory) CD8⁺ T cells among the total tetramer⁺ CD8⁺ T cells. The right lower panel represents the mean frequency (± SEM) of CCL3-secreting cells among the total CD3⁺ CD8⁺ T cells. Data are pooled results from three independent experiments with n = 15 mice. P values were calculated between groups of mice immunized with wt versus Del-SecA2 and NBS-SecA2 L. monocytogenes or injected with PBS. NS, not significant.

As expected, while L. monocytogenes-specific memory CD8⁺ T cells from mice immunized with wt bacteria secreted CCL3 upon reactivation (∼0.21%), memory cells from animals primarily infected with Del-SecA2 exhibited lower frequencies of these cells (∼0.13%) (32). Importantly, mice that received NBS-SecA2 also had similar frequencies of CCL3-secreting memory CD8⁺ T cells (∼0.13%), equivalent to that of mice immunized with the Del-SecA2 mutant. Since spleens from NBS-SecA2-, Del-SecA2-, and wt-infected mice exhibited comparable numbers of cells following the challenge infection (not shown), the absolute numbers of bacterium-specific CD8⁺ T cells that secreted CCL3 cells were close to background levels in mice immunized with the mutated bacteria. Collectively, these results therefore suggest that like in mice primed with Del-SecA2 (32), the lack of CCL3⁺ memory CD8⁺ T cells in mice receiving primary immunization with NBS-SecA2 L. monocytogenes correlates with their inability to control a secondary infection.

DISCUSSION

In a previous study, we demonstrated that the SecA2 ATPase of the intracellular bacterium L. monocytogenes, a well conserved ATPase among Gram-positive bacteria, is required for the induction of CD8⁺ T cell-mediated protection against secondary L. monocytogenes infection (31). We showed that SecA2 needs to be expressed inside the cytosol of infected cells. In the present work, we investigated whether SecA2 by itself or via one or several of its substrate proteins triggers the adequate signaling events at the time of Listeria-specific CD8⁺ T cell priming. For this purpose, we constructed and characterized an L. monocytogenes mutant which carries two point mutations in the secA2 gene that abolish the ATPase function of the protein yet do not affect its expression by the bacteria. Our results suggest that one SecA2-dependent protein (or more) is essential for the onset of protective CCL3-secreting memory CD8⁺ T cells. One important question raised by these data is to elucidate which SecA2-dependent protein(s) induces these cells and by which mechanism.

L. monocytogenes growth inside the cytosol of infected cells triggers specific cytosolic signals (23, 28), and SecA2 secretes one or several proteins involved in this pathway that contribute to activating APCs for CCL3⁺ memory CD8⁺ T cell induction. Along these lines, in mice immunized with heat-killed L. monocytogenes, which does not access the cytoplasm of cells and with which suboptimal activation of DCs is observed (30), CD8⁺ T cells are primed but do not develop into memory cells capable of protecting against secondary listeriosis (41). These cells never differentiate into IFN-γ/TNF-α-secreting cytolytic T lymphocytes (22). Likewise, mice immunized with L. monocytogenes lacking listeriolysin O (LLO), a pore-forming toxin of the bacteria that is essential for their escape from the primary vacuole of phagocytosis to the cytosol of infected cells (an essential step for inducing long-term immunity) (4, 5), are unable to resist a secondary infection with wt bacteria. In that case too, L. monocytogenes-specific CD8⁺ T cells are primed and differentiate into memory cells. However, in contrast to heat-killed L. monocytogenes-primed CD8⁺ T cells, these cells undergo differentiation into IFN-γ/cytolytic effectors, suggesting that these effector activities of memory CD8⁺ T cells are not sufficient to confer protective immunity (3, 12). Thus, induction of cytolytic and IFN-γ/TNF-α-secreting CD8 T cell effectors does not require access of live bacteria to the cytosol of infected cells but rather requires adequate qualitative signals (13). Similarly, chemically inactivated or irradiated replication-deficient L. monocytogenes, on which the immunogenic structures of the bacteria are still preserved, can prime protective IFN-γ⁺ memory CD8⁺ T cells, but protection becomes dependent on memory CD4⁺ T cells (8).

Collectively these results suggest that access of L. monocytogenes to the cytosol of infected cells and growth, as well as the presence of a sufficient amount of bacterially derived immunogenic structures, are prerequisites for inducing fully functional (e.g., CCL3⁺, IFN-γ⁺, TNF-α⁺, and cytolytic) protective memory CD8⁺ T cells in vivo. In this report, our data obtained using both the Del-SecA2 (31) and the NBS-SecA2 mutants, which both escape and grow inside the cytosol of infected cells at rates equivalent to that of the actA protective mutant of L. monocytogenes (11), may indicate that one or more SecA2 substrate proteins are driving the induction of the protective CCL3⁺ memory CD8⁺ T cells. Since the expression levels of wt and NBS-mutated SecA2 proteins are comparable inside L. monocytogenes, this further supports our interpretation and makes less likely that the SecA2 protein itself or one of its derived motifs is mediating optimal T cell priming signals.

How might this SecA2-dependent protein act inside the cytosol of infected cells? One possibility is that the SecA2 substrate protein bears a motif that is recognized by an intracytosolic PRR yet to be determined. Many reports have linked the set of PRRs expressed and triggered on or inside APCs to the outcome of primary T cell responses (29). As for primary responses, such activation pathways are likely to affect the quality of the memory T cells generated. TLRs bind to and detect conserved PAMPs either at the cell surface or in the cytosol inside the lysosome/endosome internal membranes. For instance, TLR3, TLR7, and TLR9 are expressed almost exclusively inside intracellular compartments such as endosomes and recognize motifs released from degraded nucleic acids (RNA and DNA) of viruses or bacteria (1, 9, 33). While L. monocytogenes expresses several TLR ligands, TLRs have been found only at the cell surface or inside specific intracellular compartments and have not been detected inside the cytosol of infected cells. Rather, detection of cytosolic L. monocytogenes involves other PRRs such as NLRs. For instance, bacterial PGN degradation releases l-Ala-d-Glu-meso-diaminopimelic acid (DAP-PGN) that activates Nod1 as well as muramyl dipeptides (MDP) that are sensed by Nod2 and NALP3. However, L. monocytogenes encodes a PGN-N-deacetylase which modifies its PGN, therefore allowing the bacteria to escape Nod1- and TLR2-mediated detection by the innate immune system (6). Thus, it seems less likely that Nod1 participates in L. monocytogenes recognition inside the host cell cytosol but rather that Nod2, NALP3, or other cytoplasmic sensors of the cells do so.

Generation of DAP-PGN and MDP which trigger NLRs involves PGN degradation by bacterial enzymes inside the cytosol. Interestingly, the most abundant SecA2-secreted proteins are the two autolysins p60 and NamA/MurA, which are involved in L. monocytogenes PGN digestion (25, 26) and are predicted to release MDP and likely other products from PGN. Indeed, L. monocytogenes lacking either or both of these autolysins exhibits a rough and bacterial chaining phenotype comparable to that of the Del- and NBS-SecA2 mutants and is impaired in its ability to digest PGN. However, since MDP are also present in intact PGN (6, 14), they should be recognized via Nod2 or NALP3 as efficiently in wt and SecA2-deficient L. monocytogenes. Therefore, they might not account for the distinct outcome of CD8⁺ T cell priming in mice immunized with these bacteria.

From the studies from Portnoy's laboratory, however, it is clear that very specific genetic programs are triggered upon L. monocytogenes access to the cytosol of infected macrophages and that this leads to the expression of an IFN regulatory factor 3 (IRF-3)- and NF-κB-dependent, MyD88-independent inflammatory response (23). Both genomic DNA and MDP from the bacteria synergistically mediate the induction of the cytosolic response, which likely contributes to T cell responses. In the case of bacteria lacking a functional SecA2 ATPase, one hypothesis may be that the p60 and NamA autolysins trigger such a cytosolic response by releasing NLR ligands distinct from MDP inside the cytoplasm of infected cells, leading to the activation of signaling pathways adequate to promote the differentiation of CCL3⁺ memory CD8⁺ T cells. Identification of such NLR ligands from bacteria may be of broad interest for therapeutic vaccination approaches.