Although the innate immune response is the first line of defense against B. anthracis, its spores survive, germinate, and proliferate in macrophages, eventually bursting them to produce a lethal titer of infectious particles. The mechanism by which B. anthracis evades immune attack is not fully understood. Most studies have been focused on major virulence factors found on two plasmids (pXO1 and pXO2) that are responsible for exotoxins, capsule formation, and spore germination (2). These plasmid-borne virulence factors have been the prime candidates for anti-anthrax drug design. However, the ability of pathogens such as B. anthracis to survive in phagocytes also depends critically on the state of their oxidative stress defense system. Reactive oxygen species (ROS) play essential roles in innate immunity against many types of microorganisms (3–5). The antibacterial effects of ROS have been largely attributed to DNA and protein damage mediated by the Fenton reaction (6). This process generates hydroxyl radicals that react with DNA bases, sugar moieties, and amino acid side chains, causing various types of lesions (7). We showed that nonpathogenic B. subtilis utilizes its own nitric oxide (NO) to gain rapid protection against sudden oxidative damage (8). The mechanism of protection does not rely on transcriptional gene induction but rather on rapid suppression of DNA damage by preventing the Fenton reaction and direct activation of catalase (8). Here, we describe the key role of NO-synthase (bNOS)-derived NO in protecting germinating B. anthracis spores from macrophage oxidative attack, establishing the principal role of bacterial NO in pathogenicity.

Endogenous NO Protects B. anthracis from Oxidative Stress. To investigate the role of endogenous NO in defending B. anthracis from oxidative stress, we inactivated the bNOS-encoding gene with a Km cassette. Because the nos gene is not transcribed as part of any operon, and because inactivation was done by integration of the Km cassette in the middle of the gene, the mutation should not have any polar effects. The bNOS deletion did not significantly influence the growth rate in rich and sporulation media [supporting information (SI) Fig. 6]. To prove that the nos deletion abolished NO production, we took advantage of the highly specific copper-fluorescein-based NO fluorescent probe (CuFL) (9) (Fig. 1A). WT, but not Δnos, cells exhibited bright fluorescence upon treatment with CuFL, indicating that bNOS-deficient cells lost their ability to produce NO (Fig. 1A). Very low residual fluorescence is probably nonspecific, because no other NO-producing enzymes were found in B. anthracis. This result demonstrates that B. anthracis NOS generates NO in vivo under physiological growth conditions.

Fig. 1. NO protects B. anthracis against oxidative stress. (A) bNOS-dependent NO production in vivo. Representative fluorescent image of bacteria treated with Cu(II)-based NO-detecting probe (CuFL). B. anthracis Sterne (WT) and Δnos cells were grown in (more ...)

We previously established the mechanism of endogenous and exogenous NO protection against oxidative stress in B. subtilis (8). Because B. anthracis and B. subtilis are close relatives, NO is likely to protect B. anthracis from oxidative stress by the same mechanism. The susceptibility of NOS-deficient B. anthracis to peroxide was monitored qualitatively by a modified antimicrobial disk technique. Plates seeded with corresponding bacterial strain were incubated overnight with a filter paper disk saturated with 0.45M H₂O₂ placed on top of brain heart infusion (BHI) agar. Δnos cells formed a clear 5- to 7-mm zone around the H₂O₂ disk (Fig. 1B), whereas WT cells grew a complete lawn around the disk, thus demonstrating strong NOS-dependent resistance to hydrogen peroxide. A similar inhibition zone was also observed if the specific NOS inhibitors [N-methyl-l-arginine (NMLA) or N-nitro-l-arginine methyl ester (LNAME)] were added to the WT strain (Fig. 1B). Consistently, we found that the rate of H₂O₂ degradation in crude extracts of WT cells was ≈2-fold greater than in Δnos cells (Fig. 1C), suggesting that the antioxidant effect of endogenous NO was at least in part due to stimulation of catalase activity. Indeed, vegetative catalase is the major enzyme that detoxifies hydrogen peroxide in Bacilli (10). It is an iron-heme protein and therefore a natural target for NO. Our previous results indicate that NO activates B. subtilis catalase directly via an S-nitrosylation mechanism (8).

To prove that NO indeed protected B. anthracis from oxidative stress, we examined the effect of a nontoxic single dose of NO (30 μM aqueous solution) on the survival of bacteria exposed to a lethal dose of hydrogen peroxide (10 mM). Within 5 sec of NO treatment, B. anthracis cells became ≈100 times more resistant to H₂O₂ challenge than nontreated bacteria (Fig. 1D, bar 2). The addition of NO simultaneously with or after H₂O₂ had no protective effect (Fig. 1D, bars 5 and 6), apparently because of rapid NO scavenging by the peroxide-derived radicals (10). Also, no cytoprotection was observed when oxidized NO or nitrite was added instead of NO (Fig. 1D, bars 7 and 8). Moreover, pretreatment with the same low concentration of H₂O₂ (30 μM) did not protect bacteria from a subsequent lethal dose (10 mM) of H₂O₂ (Fig. 1D, bar 9), indicating that the protective effect of NO is highly specific. These controls and the large molar excess (>300 times) of H₂O₂ rule out the possibility that the protective effect of NO is due to direct reaction with H₂O₂ or its products. A comparable level of protection from H₂O₂ was also achieved with the NO-donors methylamine hexamethylene NONOate (MAHMA NONOate) (bar 4), S-nitroso-N-acetylpenicillamine (SNAP), and propylamine propylamine NONOate (PAPA NONOate) (data not shown).

Although NO can activate various oxidative stress genes in bacteria to protect cells from reactive oxygen and nitrogen species (11–16), in our experiments, full protection was established within seconds of NO challenge (Fig. 1E), suggesting that gene induction is not required. Consistently, we found that inhibition of protein synthesis by the antibiotic chloramphenicol did not compromise NO-mediated cytoprotection (Fig. 1E).

Because in all of the experiments the NO-mediated effects were either identical or more potent compared with those observed in B. subtilis, we argue that NO protects B. anthracis by the same dual mechanism: suppression of the Fenton reaction and activation of catalase (8), with the former being predominant.

NO from B. anthracis Is an Essential Virulence Factor. To evaluate the contribution of bNOS to the pathogenicity of B. anthracis, the parental Sterne and Δnos mutant strains were compared for virulence after s.c. inoculation of their spores into A/J susceptible mice (17). Δnos spores exhibited a dramatic loss of virulence, with an LD50 increase of approximately three orders of magnitude compared with spores of the parental control strain (Fig. 2A). The nos gene is monocistronic, which excludes the possibility that the nos mutation exerted any polar effect leading to the virulent phenotype in vivo. Moreover, in comparison with WT, Δnos mutant cells were much more susceptible to killing by the macrophages (Fig. 2B). J774A.1 macrophages were challenged with B. anthracis spores at a 1:50 ratio for 40 min, followed by removal of noninternalized bacteria. Infected macrophages were lysed at different time intervals, and the surviving bacteria were plated on Luria–Bertani (LB) agar for colony-forming unit enumeration. To determine whether spores or vegetative cells were responsible for the marked difference in colony-forming units between WT and Δnos strains, each macrophage sample was spread on plates directly or applied to the plates after heat-mediated killing of vegetative cells. The results, expressed as a percentage of surviving cells, demonstrate that at 2 h after infection almost all of the spores germinated in macrophages (Fig. 2B Right). A control experiment showed that macrophages ingest Δnos and WT spores with the same efficiency. We conclude that germinating bacteria, not spores, account for the increased susceptibility to macrophage-mediated killing of Δnos B. anthracis.

Fig. 2. bNOS is essential for B. anthracis virulence in mice and survival in macrophages. (A) LD50 of anthrax spores as a function of bNOS activity. The indicated amounts of B. anthracis spores (Sterne or Δnos) were inoculated s.c. into 6- to 7-week-old (more ...)

Anthrax NO Production Occurs at the Early Stage of Infection. Considering that macrophages produce a large amount of NO to control infection (18–20) our seemingly paradoxical results can be rationalized if NO production by bacteria and macrophages is separated in time, leading to opposite outcomes. Indeed, it is well established that superoxide generation begins immediately after phagocytosis of bacteria. However, inducible NOS is activated only 8–12 h after infection, when the oxidative burst is largely over (4, 19). In contrast, B. anthracis germinating in macrophages would be expected to begin producing its own NO immediately from bNOS to antagonize the antimicrobial action of ROS. To test this hypothesis, we examined the level and source of NO production in infected macrophages at various times (Fig. 3). We first measured NO end products—nitrate and nitrite (NN). NO diffuses freely through cell walls and membranes, and, after escaping from cellular confinement, it is rapidly oxidized in the medium under aerobic conditions to form NN. Therefore, the level of NN in the media reflects NOS activity. We detected significant NN production in macrophages infected by WT, but not Δnos, spores within 2 h after infection (Fig. 3A). Consistently, bNOS-specific NO production in infected macrophages was demonstrated by using CuFL (Fig. 3B). At 2 h after infection, germinating WT (Sterne) spores caused a bright fluorescence signal indicative of a high level of NO. However, the NO level remained at a basal level after infection by Δnos spores or treatment with cytokines. Notably, at 18 h after infection, macrophages challenged with either WT or Δnos spores or cytokines generated the large amount of their own NO (Fig. 3B). Taken together, these results show that the only major source of NO at an early stage of infection (≈2 h) is that produced by ingested bacteria, not by macrophages.

Fig. 3. NO production in B. anthracis infected macrophages. (A) NN accumulation in clarified supernatants of J774A.1 macrophages 2 and 10 h after infection with Sterne or Δnos spores or noninfected control macrophages (MF). Experimental conditions are (more ...)

Model for the NO-Mediated Defense System in B. anthracis. To determine whether bacterial NO indeed protects B. anthracis from macrophage-inflicted oxidative damage, we monitored the state of bacterial genomic DNA in macrophages infected with either WT or Δnos spores. Formation of double-strand breaks in DNA after peroxide exposure is the major cause of bacterial death (21, 22). As mentioned earlier, this global damage to bacterial DNA is a result of the Fenton reaction (7), which can be suppressed by NO (8). The integrity of DNA can be analyzed by quantitative PCR (23). Accordingly, treatment of exponentially growing bacteria by H₂O₂ caused a significant decrease in the yield of the PCR product (Fig. 4, lanes 5 and 6). We next isolated bacterial DNA at 2 h after infection and analyzed it by the same quantitative PCR method. DNA obtained from nos mutant cells was damaged to a much greater extent that that from the parental strain (Fig. 4). The yield of full-length PCR fragment decreased ≈4-fold. No DNA damage was detected in either mutant or WT vegetative cells in the absence of macrophages (Fig. 4, lanes 3 and 4). These results demonstrate that bNOS-derived NO protects the B. anthracis chromosome from the macrophage-induced Fenton reaction at an early stage of infection.

Fig. 4. B. anthracis NO protects bacteria from the macrophage-inflicted oxidative damage. bNOS-dependent bacterial DNA protection during macrophage infection. Chromosomal damage was monitored by qPCR. A representative agarose gel shows a 3.6-kb PCR fragment amplified (more ...)

A large body of evidence implicates ROS in microorganism killing by stimulated phagocytes (3, 6, 18, 24). Our in vitro studies indicate that NO directly and rapidly activates a latent oxidative stress defense system in B. anthracis, which does not require additional protein synthesis (Fig. 1E) and thus must be available immediately upon spore germination. Taking together, the results presented in this report lead us to propose a novel antioxidant defense mechanism used by NOS-containing pathogenic bacteria such as anthrax (Fig. 5). In this mechanism, endogenous NO rapidly protects pathogens against immunological oxidative stress by suppressing the harmful Fenton reaction and directly activating catalase (8). Although the superoxide anion generated in the phagosome can react with NO to form toxic peroxynitrite (ONOO⁻), the inability of O₂⁻ to pass through the bacteria cell wall and membrane makes this event highly unlikely. Instead, H₂O₂, a product of spontaneous or enzymatic O₂⁻ dismutation, readily enters bacterial cells, where it is the major cytotoxic species because it fuels the damaging Fenton reaction inside bacteria. Therefore, the rapid NO preconditioning that occurs via a dual mechanism—the inhibition of the Fenton reaction and the activation of catalase (8)—renders bacteria resistant to immune oxidative attack (Fig. 5). The primary sequence and high-resolution structure of bNOS reveal several unique features that distinguish it from mammalian NOS counterparts (25, 26), suggesting that bNOS-specific small molecular inhibitors can be designed. Indeed, some potent bNOS inhibitors are described in ref. 27. Because bNOS represents an important virulence factor, its selective inhibition offers a previously uncharacterized and specific approach to the development of antimicrobial therapy.

Fig. 5. The proposed mechanism of the NO-mediated defense system in B. anthracis. Upon germination, bNOS, which has been accumulated in the spore during the sporulation phase (32, 33), generates NO that instantly protects the pathogen from H2O2 toxicity by a (more ...)

Chemicals and Reagents. CuFL was prepared as described in ref. 28. Kanamycin (Km), erythromycin (Em), lysozyme, H₂O₂, MAHMA NONOate, PAPA NONOate, and SNAP were purchased from Sigma. Restriction enzymes and T4 ligase were purchased from New England BioLabs. PCR was carried out with Ex TaqDNA polymerase (TaKaRa). NO solution was prepared in an airtight device by bubbling NO gas (Aldrich) that had been purified from higher oxides by passing it through a 1 M solution of KOH into water until the concentration of dissolved NO reached ≈300 μM. Milli-Q-grade water was deaerated by boiling and then cooling under argon (Praxair). Immediately before reactions were performed, the NO concentration was measured with an iso-NO Mark II electrode (WPI Instruments).

Animals. A/J mice were purchased from The Jackson Laboratory. Six-week-old female mice, each weighing 18 to 22 g, were used in all experiments. This strain of mice is susceptible to spores of the B. anthracis Sterne strain (17).

Bacterial Strains, Plasmids, and Mammalian Cell Lines. B. anthracis Sterne 34F2 (pXO1⁺pXO2⁻) was used as a parent strain. Plasmids were constructed by using standard methods and amplified in E. coli TG1 [supE Δ (hsdM-mcrB)5 (r_k⁻m_kMcrB) thiΔ(lac-proAB) F′ (traΔ36 proAB+lacI^q lacZΔM15)]. Plasmids for B. anthracis transformation were first isolated from E. coli GM 2163 [F⁻ ara-14 leuB6 fhuA31 lacY1 tsx78 glnV44 galK2 galT22 mcrA dcm-6 hisG4 rfbD1 rpsL136 dam13::Tn9 xylA5 mtl-1 thi-1 mcrB1 hsdR2] in unmethylated form. All PCR fragments were amplified from B. anthracis Sterne 34F2 chromosomal DNA, using ExTaq DNA polymerase (TaKaRa). Oligonucleotide primers (IDT) used for PCR of the Sterne nos (JGI “Locus Tag” BAS5299) are shown in SI Materials.

To construct pKS1::baNOS, two ≈500-bp fragments upstream and downstream of B. anthracis nos were amplified by PCR and cloned into pKS1 (29). The resulting plasmid (pKS1::baNOS) carries the kanamycin (Km) resistance gene flanked with these fragments. Sterne strain cells were transformed with pKS1::baNOS, Km^R Em^S colonies were selected, and double cross-over recombination events were confirmed by PCR. Preparation of electroporation-competent B. anthracis cells was carried out as follows: A 50-ml culture of B. anthracis was grown in BHIG [BHI (Difco) and 0.5% glycerol] at 37°C to OD₆₀₀ ≈0.6. The cells were harvested by centrifugation at 4°C, and all subsequent steps were performed on ice. The cells were washed three times with an equal volume of 10% glycerol plus 1 mM Hepes (pH 7.0), resuspended in 2.5 ml of the same glycerol-Hepes solution, and kept on ice. Up to 5 μl of plasmid DNA (0.1–4 μg in water) was added to a 100-μl aliquot of cells, which was pulsed (2.5 kV, 25 μF, 200 Ω) in a 0.2-cm gap cuvette. The cells were resuspended in 1 ml of BHIG and incubated for 1.5 h at 30°C with aeration. Recovered cells were spread on LB medium or BHI agar plates, containing Km and Em. Colonies were visible after 20–24 h.

Bacteria were grown aerobically at 30 or 37°C in Luria–Bertani (LB) or brain heart infusion (BHI) media (Difco). The antibiotics Km and Em were used at final concentrations of 50 μg·ml⁻¹ and 300 μg·ml⁻¹ for E. coli or 100 μg·ml⁻¹ and 3 μg·ml⁻¹ for B. anthracis, respectively.

The macrophage cell line J774.A1 (TIB-67) was obtained from American Type Culture Collection and was maintained in DMEM containing 10% FBS (Gemini Bio-Products) and an antibiotic-antimycotic mixture (penicillin–streptomycin–fungizone) (Invitrogen) at 37°C with 5% CO₂.

General Methods. B. anthracis overnight cultures, grown in liquid brain heart infusion (BHI) media, were diluted 1:100 in fresh BHI and grown at 37°C with aeration to OD₆₀₀ ≈1.0, unless indicated otherwise. To determine H₂O₂ resistance, B. anthracis cells were swabbed on BHI agar plate, and 3MM disks were applied on the plate surface. Ten milliliters of 450 mM H₂O₂ was added on each disk, and plates were incubated at 37°C overnight. To prepare bacterial cell extracts, B. anthracis cells were harvested, dissolved in lysis buffer [20 mM Tris·HCl (pH 7.9) and 150 mM NaCl] containing 125 μg/ml lysozyme (Sigma), incubated for 5 min at 37°C, sonicated, and clarified by centrifugation. The protein concentration was determined by using a BioRad protein assay kit. Nitrite was measured in clarified infected macrophage cell culture supernatants, using the fluorometric nitrite assay kit (Cayman). Macrophages were lysed, and bacteria colony-forming units were counted on LB agar plates. One-half milliliter of a 5 × 10⁵ macrophage cell suspension was loaded in each well of a 24-well plate and incubated at 37°C with 5% CO₂ for 24 h. Spores were added to the macrophages (50:1) and incubated for 40 min Gentamicine (2 mg/ml) was then added to eliminate noningested spores and vegetative bacteria. After 20 min, the antibiotic media was replaced by fresh media and incubation was continued. At each subsequent hour, macrophages were lysed with 0.05% sodium desoxyholate and plated either directly or after heat-treatment on LB agar to determine the colony-forming units (vegetative cells or non-germinated spore). DNA manipulations and plasmid DNA isolation were performed by using standard procedures (30). B. anthracis spores were prepared in Difco sporulation medium as described in ref. 31.

Detection and Measurement of Bacterial NO. To quantify nitrate/nitrite (NN), cell culture supernatants were clarified by centrifugation and then filtered through YM-3 microcon (Millipore). NN were measured in the flow-through fraction with a fluorometric and calorimetric nitrate/nitrite assay kit (Cayman). NO production in vivo was detected by the NO-specific intracellular fluorescent CuFL probe, as described in ref. 9. CuFL was prepared immediately before use by mixing FL and CuCl₂ in a 1:1 ratio and then added to the growing bacterial cultures or macrophages to a final concentration of 10 μM. One hour later, fluorescent and visible images of bacteria treated with the NO-detecting probe were taken with a digital camera attached to an Axio microscope (Zeiss MicroImaging). The percentage of fluorescent bacteria was calculated by IPLab Scientific image processing software.

Catalase Activity Assay. Degradation of H₂O₂ was monitored in real time as a decrease in absorbance at 240 nm (10). Aliquots of extracts to be monitored or of pure catalase were mixed with 50 mM phosphate buffer (pH 7.0) and placed into a 1-ml quartz cuvette. H₂O₂ solution (40 mM) was added, and the kinetics of its degradation were recorded. Total H₂O₂ degrading activity was measured as the decrease of H₂O₂ concentration per milligram of total protein per second. OD₂₄₀ was converted to the concentration of H₂O₂ according to the calibration curve (10 mM H₂O₂ = 0.36 OD₂₄₀).

Measurement of DNA Damage. Macrophages infected with WT or Δnos spores were lysed and centrifuged to collect bacteria. Total genomic DNA was isolated from bacteria pellets according the GenElute bacterial genomic DNA kit protocol for Gram-positive bacteria (Sigma). DNA was extracted with phenol/chloroform and quantified by using a PicoGreen dsDNA quantitation reagent (Molecular Probes) and lambda phage DNA as a standard. An ≈3.6-kb fragment of AE017225 contig (JGI) was used for qPCR. Primer sequences were as follows: 5′-CTCAGCTGGTTAGAGCGCACGCCTG-3′ (forward) and 5′-CACCCCTTCTCCCGAAGTTACGGGGTC-3′ (reverse). PCR was performed by using Phusion DNA polymerase (Finnenzymes). The 50-μl PCR mixture contained 0.05–0.5 ng of genomic DNA as a template, 1.5 μM primers, 200 μM dNTPs (Fermentas), 5× Phusion GC PCR buffer, and 0.5 μl of DNA polymerase. DNA was subjected to 30 cycles of PCR (98°C for 30 s, 58°C for 30 s, 72°C for 9 min). PCR products were separated by electrophoresis in an 0.8% agarose gel, stained with ethidium bromide, scanned, and quantified with an AlphaImager (Imgen Technologies).

This work was supported by the National Institutes of Health Director's Pioneer Award (to E.N.) and National Science Foundation Grant NSF CHE-0611944 (to S.J.L.).