Corynebacterium pseudotuberculosis strains were routinely maintained in Brain Heart Infusion broth or in BHI 1.5% bacteriological agar plates, at 37°C. Tween-80 was added to the broth cultures at 0.05% to prevent cell clumping. When necessary, the antibiotic kanamycin was used at 25 μg/ml. Escherichia coli strains were maintained in Luria–Bertani medium and employed in DNA manipulation experiments, according to standard protocols. Ampicillin (100 μg/ml) and kanamycin (50 μg/ml) were used where appropriate.

A chemically defined medium (CDM), previously optimized for growth of C. pseudotuberculosis (Moura-Costa et al., 2002), was used in all the experiments that included stress-generating agents. The composition of the CDM was as follows: autoclaved phosphate buffer pH = 7.4 [Na₂HPO₄·7H₂O (12.93 g/l), KH₂PO₄ (2.55 g/l), NH₄Cl (1 g/l), MgSO₄·7H₂O (0.20 g/l), CaCl₂ (0.02 g/l), and 0.05% (v/v) Tween-80]; 4% (v/v) MEM Vitamins Solution 100× (Invitrogen); 1% (v/v) MEM Amino Acids Solution 50× (Invitrogen); 1% (v/v) MEM Non-Essential Amino Acids Solution 100× (Invitrogen); and 1.2% (w/v) filter-sterilized glucose.

A conserved region of the sigE gene of C. pseudotuberculosis (GenBank: ADL20681.1) was initially isolated by a strategy employing degenerate primers. Basically, the amino acid sequences of the σ^E factors of the bacteria M. tuberculosis (Rv1221), C. diphtheria (DIP0994), C. glutamicum (NCgl1075), and C. efficiens (CE1177) were retrieved from public databases and aligned using the ClustalW tool. Highly conserved regions were identified in the σ^E factors of these bacteria and degenerate primers were designed according to the coding sequences of these factors, such that it would be possible to amplify a partial region of the sigE gene in all genomes analyzed. The primer pair sigE#1: GGMACCGCAGCDTTCGACGC and sigE#2: CGTCCRCGGTGRATWCGGGA was used for amplification of an expected 490 bp internal fragment of the sigE gene of C. pseudotuberculosis. PCR products around the expected size were purified from agarose gels, ligated into the vector pCR2.1 TOPO (Invitrogen), and sequenced according to standard protocols.

A plasmid carrying a fragment of the sigE gene was introduced into the wild-type strain 1002 of C. pseudotuberculosis by electroporation. This plasmid functions as a suicide vector for this bacterium, as it does not carry a functional corynebacterial replication origin. C. pseudotuberculosis clones that underwent recombination were selected by kanamycin resistance (25 μg/ml). Confirmation of recombination events in the C. pseudotuberculosis genome was obtained by Southern blot analysis using a [α³²-P]dCTP radiolabeled probe for the sigE gene, or through PCR reactions employing different combinations of primers that align in the start and stop codons of sigE with primers that align within the inserted plasmid: sigE#3-ATGACATCGAACAGTGGTTC/sigE#4-TTAGTGGGACATCGGTAGG; Kan#1-ATGATTGAACAAGATGGATTG/Kan#2-TTAATAATTCAGAAGAACTC; M13F/M13R (Invitrogen; data not shown).

Corynebacterium pseudotuberculosis strains were grown in CDM to early exponential phase (OD_540nm = ∼0.1), cultures were split into several aliquots and then incubated separately with different concentrations of various stress-generating agents, as follows: osmotic stress (NaCl, 0.5 and 1 M); acidic stress (HCl, to pH 4.5 and 5.5); detergent stress (SDS, 0.01% w/v and 0.05%); lysozyme, 500 and 750 μg/ml; alcoholic stress (ethanol, 2.5 and 5%); oxidative stress (H₂O₂, 1, 10, and 50 mM); nitric oxide stress (DETA/NO – diethylenetriamine NON-Oate, 0.1 and 1 mM). For starvation stress, CDM was prepared with low concentrations of glucose: 0.3 and 0.15%. For heat and cold shocks, cultures were incubated at 50 or 4°C for 30 min, and then returned to normal growth at 37°C.

Growth of the control and stressed cultures of C. pseudotuberculosis, of both wild-type (1002) and mutant (ΔsigE) strains, was monitored for 24 h in a LabSystems iEMS Absorbance Plate Reader (Thermo Fisher), at OD_540nm. Results were plotted using the GraphPad Prism software (GraphPad Software, Inc); the integrals of the areas under the curves (AUC) were determined, and a growth index (GI) was then calculated, as follows: GI (%) = [(AUC_{Treated culture}/AUC_{Control culture}) × 100]. Experiments were performed at least in triplicate.

Steady-state concentrations of NO generated by the NO-donor DETA/NO in the cell culture medium were determined amperometrically using an NO-specific electrode (ISO-NOP, World Precision Instruments, Inc.) attached to an NO meter (ISO-NO Mark II, WPI). Briefly, the electrode was immersed in CDM and the electrode current allowed to stabilize under conditions identical to those used in the cell culture experiments before addition of different concentrations of DETA/NO and recording of the electrode response. Stock solutions of DETA/NO (Sigma-Aldrich) were prepared fresh and kept on ice in the dark, as described (Feelisch, 1998). Currents in pA were converted into NO concentrations by comparison to a standard calibration curve generated from either copper(I)-mediated decomposition of S-nitroso-N-acetyl-d,l-penicillamine or from iodide-mediated nitrite reduction, under otherwise identical conditions (Davies and Zhang, 2008).

Bone-marrow derived macrophages (BMMs) were obtained from the femurs and tibias of C57BL/6 or iNOS knockout mice (iNOS^−/−), according to a standardized protocol (Carvalho et al., 2011). Gamma (γ)-interferon was added at 20 U/well for macrophage activation. A Neutral Red assay was used to evaluate macrophage viability following C. pseudotuberculosis infection (MOI 5:1), as described (McKean et al., 2007).

C57BL/6 or iNOS^−/− mice were infected intraperitoneally with 10⁶ colony forming units (CFU) of the wild-type (1002) or mutant (ΔsigE) strain of C. pseudotuberculosis. On days 1 and 3 post-infection, animals were sacrificed and bacterial loads in the spleens were enumerated.

The 1002 (wt) and ΔsigE strains of C. pseudotuberculosis were grown in CDM to mid-exponential phase, cultures were split into two aliquots and 100 μM of the NO-donor DETA/NO were added to one aliquot of each strain. NO-treated cultures along with control cultures were further incubated for 1 h at 37°C. At this point, extracellular proteins were extracted by the three-phase partitioning technique and identified/quantified by a high-throughput proteomic strategy, based on a recently introduced method of liquid chromatography–mass spectrometry acquisition (LC–MS^E), exactly as previously described (Pacheco et al., 2011). Biological function annotations for the identified proteins were retrieved from the gene ontology (GO) database, using the web tool AmiGO¹ (Carbon et al., 2009). Protein regulation data were obtained from the CoryneRegNet v6.0 Database² (Pauling et al., 2011).

In order to evaluate the role played by the ECF sigma factor σ^E in C. pseudotuberculosis resistance to stress conditions faced during intracellular infection, a sigE-null mutant strain of this bacterium (ΔsigE) was generated by homologous recombination in the parental strain 1002 (wt). Both strains were submitted to a series of stress-generating agents in vitro, which aimed at mimicking conditions found by bacterial pathogens in the intraphagosomal environment (Rohde et al., 2007; Ehrt and Schnappinger, 2009; Schaible, 2009).

Mutation of sigE in C. pseudotuberculosis did not alter cell growth and morphology under normal conditions (Figure 1). Growth of the ΔsigE strain was also comparable to the 1002 (wt) strain under conditions of nutrient starvation (limiting glucose), and under osmotic, thermal, alcoholic, and oxidative stresses (Figure 1). On the other hand, the mutant strain was more sensitive to an acidic pH that resembles that found within an activated macrophage (pH = 5.5), and to cell surface stressors, namely SDS and lysozyme treatments (Figure 1), corroborating previous studies on the role of σ^E in resistance to environmental-stresses in other corynebacteria and mycobacteria (Manganelli et al., 2001; Park et al., 2008). Interestingly, growth of the ΔsigE strain was also more affected than that of the wt strain following exposure to different concentrations of the NO-donor DETA/NO (Figures 1 and 2A). NO electrode measurements indicated that the concentrations of the NO-donor used in our experiments, 0.1 and 1 mM, resulted in steady-state NO concentrations of ∼500 nM and 5 μM, respectively. The effect of these concentrations on growth of the two C. pseudotuberculosis strains was apparently more due to a bacteriostatic rather than bactericidal action of NO (Figure 2C), as reported for other bacterial pathogens (Ogawa et al., 2001; Voskuil et al., 2011). Notably, the ΔsigE strain was much more sensitive to the combination of the NO-donor and H₂O₂ (Figures 1 and 2B), reinforcing the requirement of σ^E for C. pseudotuberculosis resistance to combined NO/peroxide stress.

The participation of σ^E in C. pseudotuberculosis resistance to the conditions found within the host was first evaluated by infection of C57BL/6 mice with the wt and mutant strains of this bacterium, followed by analysis of bacterial persistence in mouse spleens. After 3 days of infection, the ΔsigE strain of C. pseudotuberculosis was virtually undetectable in the spleens of C57BL/6 mice, whereas the 1002 (wt) strain still persisted (Figure 3A). We then evaluated whether nitrosative stress was a determining condition for the lowered persistence of the ΔsigE mutant in the host. For this, we infected C57BL/6 or iNOS knockout (iNOS^−/−) mice, which are unable to produce NO in the intraphagosomal environment, and compared bacterial loads in mouse spleens 3 days post-infection. Again, the C. pseudotuberculosis strain lacking σ^E was unable to persist in C57BL/6 wild-type mice (Figure 3B); nevertheless, the ΔsigE mutant displayed a surprisingly high ability to survive in the iNOS^−/− animals, following an initial 10⁶ CFU inoculum (Figure 3B).

Wild-type C. pseudotuberculosis has been shown to possess a profound capability to kill macrophages in culture (Stefańska et al., 2010), but this natural ability is also affected in the ΔsigE strain (Figure 3C). However, only ca. 30% of iNOS^−/− macrophages survived after 4 h infection with this mutant (Figure 3C). This corroborated the importance of σ^E in C. pseudotuberculosis resistance to NO generated intracellularly.

The extracellular proteomes of the 1002 (wt) and ΔsigE strains of C. pseudotuberculosis were profiled by a high-throughput proteomics method (LC–MS^E), before and after treatment by 100 μM of the NO-donor DETA/NO.

In total, 281 extracellular proteins were confidentially identified by this strategy, in the four groups studied: (i) 1002 (wt) untreated; (ii) 1002 (wt) + DETA/NO; (iii) ΔsigE untreated; (iv) ΔsigE + DETA/NO. This represented 104 different extracellular proteins of C. pseudotuberculosis (Figure 4A). Proteins exported after NO-treatment could be detected in both 1002 (wt) and ΔsigE strains (Figure 4A). The mutant strain exported a higher number of proteins in comparison to the wt strain, under normal growth conditions, and after NO-stress (Figure 4A). The majority (67%) of these differential proteins of the ΔsigE strain are predicted in silico to be truly exported proteins (data not shown). Moreover, the proteins commonly identified in the four groups presented very similar concentrations in all samples (Figure 4B) showing a high reproducibility of the methods used for extraction and identification of the exoproteins. This demonstrates that the qualitative differences observed between the exoproteomes are in fact due to biological variations of the extracellular proteomes.

Three proteins were commonly identified in the 1002 (wt) and ΔsigE strains exclusively after treatment by NO (Figure 4A). Only a few proteins were differentially exported in the 1002 (wt) strain in response to NO, compared to the mutant strain (Figure 4A). These included proteins predicted to be involved in metal ion transport and cell redox homeostasis. A putative dioxygenase was also specifically detected in the exoproteome of the wt strain (Table 1). On the other hand, 17 differential proteins could be detected in the extracellular proteome of the ΔsigE-null strain following NO-stress (Figure 4A). Metal ion transport and cell redox homeostasis were again highly represented predicted biological functions in this protein set (Table 1). Some of these proteins have already been seen to participate in C. pseudotuberculosis responses to other environmental stress conditions, according to the CoryneRegNet Database (Table 1).

A few proteins primarily considered to have a cytoplasmic location have been identified following NO-stress (Table 1). These included the chaperonin GroEL, which was found in the exoproteomes of both wt and ΔsigE strains of C. pseudotuberculosis, and the proteins glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and dnaK, differentially exported only in the mutant strain (Table 1). These cytoplasmic proteins have been consistently identified in recent exoproteome studies of various bacteria (Sengupta et al., 2010; Dreisbach et al., 2011). Besides, environmental cues such as osmotic stress and host contact have been shown to induce exportation of some of these proteins (Chitlaru et al., 2007; Mattinen et al., 2007; Pumirat et al., 2009). However, the molecular mechanisms that account for non-classical (leaderless) secretion of specific cytoplasmic proteins is still to be completely understood (Pasztor et al., 2010; Yang et al., 2011). Moonlighting roles have already been demonstrated for these proteins while in the extracellular environment, including adhesion to host cells and evasion of host’s immune mechanisms (Matta et al., 2010; Jin et al., 2011).

Noticeably, while most of the proteins commonly identified between the exoproteomes of the NO-treated strains presented very similar concentrations (Figure 4B), a single protein annotated as a “putative secreted protein” (ADL21925.1) was highly secreted only in the ΔsigE strain. This protein has also been shown to be differentially regulated during osmotic stress, according to CoryneRegNet.