A Novel Antioxidant Gene from Mycobacterium tuberculosis

Bacterial Strains and Growth Conditions.

The following mycobacterial strains were used in this study: M. smegmatis mc²155 (24), M. tuberculosis CB3.3 (22), M. tuberculosis H37Ra (American Type Culture Collection/[ATTC, Rockville, MD] No. 25177), M. tuberculosis H37Rv (ATTC No. 25618), M. bovis (ATTC No. 19210), M. africanum (ATTC No. 25420), M. gordonae (ATTC No. 14470), M. fortuitum (ATTC No. 6841), M. avium (ATTC No. 25291), M. intracellulare (ATTC No. 13950), M. kansassi (ATTC No. 12478). Mycobacterial strains were grown in Middlebrook 7H9 broth (Difco Labs. Inc., Detroit, MI) supplemented with 2% glycerol, 0.05% Tween 80, and ADC supplement (Difco Labs. Inc.) or plated on 7H11 agar (Difco Labs. Inc.). Luria-Bertani (LB) broth or LB agar were used for Escherichia coli strains XL1-Blue (Stratagene Inc., La Jolla, CA), HB101 (ATTC), DH5α (GIBCO/BRL, Gaithersburg, MD), M15 (QIAGEN Inc., Chatsworth, CA), GC4468 and DJ109 (a soxRS mutant, provided by T. Nunoshiba and B. Demple, Harvard School of Public Health, Boston, MA; reference 25), and JTG100 and its oxyR-deficient derivative, JTG101 (26), also the gift of B. Demple. Ampicillin and hygromycin B (Sigma Chemical Co., St. Louis, MO) were used at 100 μg/ml and 200 μg/ml, respectively, to grow E. coli. Hygromycin B was used at 50 μg/ml to grow mycobacteria.

Plasmids.

pBluescript (pBS; Stratagene Inc.) served as a cloning vector and pQE31 (QIAGEN Inc.) as an expression vector in E. coli. The shuttle vector pOLYG (a gift from Peadar O'Gaora and D.B. Young, Department of Medical Microbiology, St. Mary's Hospital Medical School, London, UK) is a derivative of p16R (27). pSMT3 is derived from pOLYG and contains the hsp60 promoter for overexpression of genes in mycobacteria (also a gift from P. O'Gaora).

Cloning of M. tuberculosis DNA Fragment Associated with Resistance against RNI.

Chromosomal DNA was isolated from M. tuberculosis CB3.3 as previously described (28) and digested with EcoRI and BamHI. A genomic library was constructed by ligation of the DNA fragments into the E. coli vector pBS. E. coli XL1-Blue was initially tested for growth in LB at various pH and NaNO₂ concentrations. At pH 6.0 and 10 mM NaNO₂, its growth was completely suppressed. Therefore, the genomic library was electroporated into strain XL1-Blue and the recombinants were screened for growth in acidified sodium nitrite (ASN; LB at pH 6.0 containing 10 mM NaNO₂).

Construction of pNO14.1 and its Open Reading Frame Mutant.

A HindIII–SmaI fragment of pNO14 was cloned into the HindIII and HincII sites of pBS. The resulting plasmid, pNO14.1, still contains the complete open reading frame (ORF)1. A point mutation was introduced at codon 12 of ORF1 (nucleotide 36 G → T), which created a stop codon (TGA) at that position. This mutation does not affect the amino acid sequence of the putative protein encoded by ORF2. The mutation was introduced by PCR mutagenesis. The final construct, pNO14.1-mut1, was confirmed by sequencing.

Analysis of the Resistant Phenotype.

The relative resistance to chemically generated RNI was tested by inoculating a 1:100 dilution of the overnight or log-phase culture of the strain to be tested into 3 ml of LB, pH 5.3, or 7H9, pH 5.3, containing NaNO₂ (Sigma Chemical Co.) at various concentrations. Resistance to paraquat (Sigma Chemical Co.), nalidixic acid (Sigma Chemical Co.), NaOCl (Aldrich Chemical Company, Inc., Milwaukee, WI), and ethanol was measured in LB, pH 7.0. Cultures were incubated at 37°C for the indicated periods. The chosen conditions resulted in at least a 10-fold reduction in CFU of the control E. coli HB101. The number of viable bacteria was determined by plating on LB agar. Alternatively, where indicated, a microplate assay was used that detects the ability of surviving bacteria to reduce a formulation of resazurin termed AlamarBlue^® (Sensititre/Alamar; AccuMed International Companies, Westlake, OH) to a fluorescent product (29).

Northern Blot Analysis and Reverse Transcriptase PCR.

RNA was isolated from 20 ml logarithmically growing mycobacterial cultures according to the FastPrep FP120 bead beater apparatus (Bio-101, La Jolla, CA) protocol. The integrity of the RNA preparation was verified by the presence of two sharp rRNA bands, 1,500 and 3,100 nucleotides in length.

Two oligonucleotides, RNA-1 (5′-gacgcgctgatcgccgatctacgcgcgcatggtggtcgg-3′) and RNA-2 (5′-cggcaacgccggtgaacaacgcgcgggcatcctcgccc-3′) were labeled with dioxigenin using the oligonucleotide tailing kit (Boehringer Mannheim Biochemicals, Indianapolis, IN) to serve as probes for the transcripts of the two open reading frames, ORF1 and ORF2, respectively. Northern blots were carried out according the Genius^TM System User's Guide For Membrane Hybridization (Boehringer Mannheim Biochemicals). Reverse transcriptase (RT)–PCR was performed using the Perkin Elmer Gene Amp RNA PCR Kit (Perkin Elmer Corp., Norwalk, CT). In brief, 0.2–1 μg total RNA was transcribed by Moloney MuLV RT into cDNA with either random hexamer primers or a specific primer for the ORF1 RNA at 42°C for 15 min. cDNA specific primers were added (Ia: 5′-ctacccgcgcgcggagtgacgctgacc-3′; Ib: 5′-cggcaacgccggtgaacaacgcgcgggcatcctcgccc-3′; IIa: 5′-ggggatggcggtgggtgcggtgtcg-3′; IIb: 5′-gacgcgctgatcgccgatctacgcgcgcatggtggtcgg-3′), and the reaction was carried out with AmpliTaq (Perkin-Elmer, Norwalk, CT) DNA polymerase in a volume of 100 μl. The combined annealing and extending reaction was done at 60°C for 30 s.

Protein Expression.

noxR1 was cloned behind an inducible T5 promoter into the expression vector pQE-31 (QIAGEN Inc.). M15 (pREP4) pQE-31-ORF1 were grown in LB containing 100 mg/l ampicillin and 25 mg/l kanamycin to an OD₅₈₀ of 1.0 and induced with 1.5 mM IPTG (isopropyl-β-thiogalactopyranoside). After 4 h, bacteria were harvested and a sample of lysate was subjected to SDS-PAGE and Coomassie blue staining to check for overexpression of recombinant protein. Protein containing an NH₂-terminal histidine tag was purified on Ni-NTA resin columns (QIAGEN Inc.) and analyzed by SDS-PAGE. The NH₂-terminal sequence of the purified protein was established for 19 residues, sufficient to read beyond the tag and 7 residues into NoxR1 proper. The purified protein was injected in female New Zealand White rabbits (4 injections of 100 mg NoxR1, at 4 wk intervals). The resulting antiserum was used for immunoblot analysis of bacterial lysates or purified protein by standard procedures. Affinity-purified antibody was prepared as described in Results.

Studies in Macrophages.

An assay modified from that previously described (30) was used to determine the survival of M. smegmatis strains inside macrophages. Wild-type (iNOS^+/+) and iNOS-deficient mice (iNOS^−/−; C57BL/6x129/SvEv) (31) and wild-type and phox-91–deficient mice (C57BL6/J) (32) were injected intraperitoneally with 1.0 ml of sterile, freshly prepared 5 mM sodium periodate (Sigma Chemical Co.) in PBS 4 d before harvest. The mice were killed by cervical dislocation or CO₂ inhalation. Peritoneal cells were harvested by lavage with 10 ml ice-cold sterile RPMI (Sigma Chemical Co.), pH 7.2. Cells were collected by centrifugation for 10 min at 250 g at 4°C and resuspended in RPMI medium supplemented with 10% heat-inactivated fetal bovine serum (HyClone, Logan, UT), 1% glutamine (complete medium), and 10 μg/ml gentamicin. Viable cells were counted on a hemocytometer in the presence of trypan blue and the proportion of macrophages determined by differential count of Diff-Quik stained Cytospin (Shandon, Inc., Pittsburgh, PA) preparations. Peritoneal cells (4 × 10⁵, ∼50% macrophages) were plated in 96-well tissue-culture plates (Corning Medical and Scientific, Medfield, MA) at 100 μl per well. In some experiments, recombinant mouse IFN-γ (Genentech, South San Francisco, CA) was added at 50 U/ml. The plates were incubated at 37°C in 5% CO₂, and 12–24 h later the adherent monolayers were washed twice with sterile PBS to remove gentamicin containing medium; complete removal required that the plates be emptied with a hard flick in each wash. Fresh complete medium ± mIFN-γ was added, and 24–48 h after the initial plating the cells were washed again with sterile PBS, and reconstituted with complete medium before infection.

Freshly electroporated M. smegmatis were grown to mid-log phase. Bacteria were opsonized in 10% fresh mouse serum for 30 min at 37°C, and 10 μl of the opsonized bacteria (∼2 × 10⁵) were added to each well. The plates were centrifuged for 5 min at 250 g to synchronize the infection, and were then incubated at 37°C for 30 min to allow phagocytosis. The wells were washed three times with sterile PBS to remove free bacteria. Complete medium (100 μl) containing 10 μg/ml gentamicin was added to each well, and the plates were incubated at 37°C. Gentamicin was added to prevent extracellular replication of mycobacteria that may not have been internalized or may have escaped from dying macrophages. Samples were taken at the indicated time points from individual wells in triplicate, as follows. The medium was removed and 50-μl aliquots were saved for nitrite determination. The cell monolayer was washed twice with PBS and lysed with 100 μl 0.1% sodium deoxycholate in PBS. Appropriate dilutions of the lysates were plated onto LB plates containing 50 μg/ml hygromycin B for CFU determinations.

To monitor macrophage production of NO, we measured nitrite in the culture supernatants as an accumulating oxidation product. The Griess reaction was performed as previously described (33). M. smegmatis itself produced no detectable nitrite under the conditions of these experiments, as evidenced in cultures with iNOS^−/− macrophages (see Fig. 7 B, inset).

Macrophage production of hydrogen peroxide was assessed by the horseradish peroxidase–catalyzed oxidation of fluorescent scopoletin to a nonfluorescent product, using a microplate format (34).

Decreased production of NO₂ ⁻ or hydrogen peroxide and diminished bactericidal activity could not be attributed to differential loss of macrophages from the monolayers, as monitored by measurements of adherent cell protein in the same cultures (34).

Cloning of an M. tuberculosis Gene Associated with Resistance to RNI.

E. coli XL1-Blue was electroporated with a genomic library of M. tuberculosis CB3.3 and exposed for 24 h to ASN (6 mM NaNO₂, pH 6.0). Protonation of NaNO₂ generates HNO₂, whose dismutation provides NO and nitrate, and, through reaction of NO with oxygen, other RNI (35, 36). The main products of these reactions are probably dinitrogen tri- and tetra-oxides (N₂O₃ and N₂O₄) as well as S-nitrosothiols, which have profound bacteriostatic effects (37, 38, 39). ASN plays a physiologic role in the microbicidal system of the stomach (40) and the combination of RNI and low pH also mimics aspects of the intraphagolysosomal milieu of the macrophage (5, 41). The chosen conditions killed E. coli XL1-Blue efficiently (>7 log₁₀ in 24 h). Surviving transformants were detected with a frequency of <10⁻⁶. One recombinant plasmid, called pNO14, consistently conferred upon XL1-Blue and four other E. coli hosts (DH5α, HB101, GC4468, and JTG100) as well as Salmonella typhimurium LT2 and 14028 (data not shown) an enhanced ability to resist ASN. By 12 h, transformation of HB101 with pNO14 afforded a 50-fold increase in survival above transformation of the same E. coli host with the vector alone (Fig. 1 A).

Means are not yet available for the facile genetic manipulation of M. tuberculosis. As a substitute mycobacterial species in which to analyze the function of pNO14, we chose M. smegmatis; the organism is fast-growing and easily transformable, and, as shown below, lacks a chromosomal copy of the gene carried by pNO14. The 680-bp mycobacterial DNA fragment of pNO14 was cloned into the shuttle plasmid pOLYG (27) and electroporated into M. smegmatis mc²155 (24). Neither the pOLYG–NO14– nor the vector-transformed strain was killed by sodium nitrate at pH 5.3, nor by sodium nitrite at pH 7.4 (Fig. 1 B). These findings excluded a mycobactericidal effect of acid plus nitrate (the nonreactive dismutation product of HNO₂), and of nitrite under conditions that disfavor the formation of NO. In contrast, when recombinant and control strains were exposed to nitrite at pH 5.3, both strains succumbed. Compared to its vector-transformed counterpart, M. smegmatis transformed with pOLYG–NO14 was increasingly less susceptible after 6 h, attaining a 100-fold relative advantage by 15 h (Fig. 1 C). Logarithmically growing M. smegmatis (Fig. 1 D) were more susceptible to ASN than those at stationary growth phase (Fig. 1 C). However, the protective effect of pOLYG–NO14 was manifest in both phases. Thus, a gene product encoded by NO14 conferred relative resistance to RNI upon both E. coli and M. smegmatis.

Expression Analyses.

The cloned 680-bp fragment hybridized to genomes of the members of the M. tuberculosis complex. None of the other mycobacterial strains tested hybridized with this probe (Fig. 2 A). Thus, among mycobacteria, the cloned DNA appears to be specific for the members of the M. tuberculosis complex associated with human and rodent disease.

Plasmid pNO14's 680-bp EcoRI/BamHI fragment from M. tuberculosis included two overlapping ORFs oriented in opposite directions (Fig. 2 B). ORF1 encodes a putative protein of 152 amino acids (Fig. 2 C) in an unusual pattern of alternating hydrophilic and hydrophobic segments that are shorter than typical membrane-spanning domains (Fig. 2 D). The putative protein was predicted to have an M_r of 15,514 daltons and a pI of 10.6, the basicity chiefly reflecting the relative abundance of Arg residues (9.2 mole percent). Despite the presence of four cysteines, no structural motifs were recognized, nor were there homologous sequences in the database. ORF2 did not contain a stop codon. Therefore, we isolated by colony hybridization a plasmid termed pNO16, which contained the full length ORF2, consisting of 1074 bases, from a library of M. tuberculosis CB3.3 (Fig. 2 B). The putative protein encoded in ORF2 is 40% identical to XP55, a Streptomyces lividans secretory product of unknown function (42).

To find out which ORF was responsible for the observed phenotype, we determined which was transcribed in recombinant M. smegmatis. In Northern blots with an oligonucleotide specific for ORF1, an ∼400-bp transcript was detected in RNA purified from logarithmically growing M. smegmatis pOLYG–NO14 but not from M. smegmatis pOLYG (Fig. 3 A). In contrast, neither RNA preparation hybridized with an oligonucleotide specific for ORF2. Exposure of M. smegmatis pOLYG–NO14 and M. smegmatis pOLYG to ASN before extraction of their RNA did not alter the Northern blot results (data not shown), suggesting that ORF1 but not ORF2 was expressed during manifestation of the phenotype conferred by transformation with pOLYG–NO14. The gene corresponding to ORF1 was designated noxR1.

Northern blots were negative when the noxR1 probe was applied to total RNA isolated from M. tuberculosis CB3.3. However, RT-PCR produced amplified products of the expected size (325 bp and 240 bp) with two different sets of primers specific for noxR1 mRNA (Fig. 3 B). Primers specific for ORF2 did not amplify any product (data not shown). No amplification occurred without RT or after adding RNAse A (Fig. 3 B), excluding the possibility that the products were amplified from genomic sequences. Thus, noxR1 is transcribed in M. tuberculosis.

To analyze the expression of NoxR1 protein, we prepared affinity-purified antibody against a recombinant fusion protein. noxR1 was cloned behind the inducible T5 promoter (pQE30–noxR1) and overexpressed in E. coli M15 pREP4 in fusion with a hexahistidine-containing tag. Attempts to force high-level expression in E. coli M15 via pQE30–noxR1 immediately led to inhibition of growth, and only a small amount of IPTG-induced product was recognizable by Coomassie blue staining of bacterial lysates separated by SDS-PAGE. Nevertheless, we were able to purify a single polypeptide by nickel column chromatography with the expected M_r (∼17 kD, based on the 15.5-kD deduced M_r of NoxR1 plus 1.1 kD for the fused tag). The purified protein was identified as NoxR1 by NH₂-terminal amino acid sequencing, and no contaminating sequences were detected. The ostensibly pure NoxR1 was used to raise a rabbit antiserum. Antibody was affinity purified by subjecting chromatographically purified NoxR1 to SDS-PAGE and blotting NoxR1-containing gel slices to a nitrocellulose membrane, to which specific antibody was bound and eluted. The affinity-purified antibody did not detect any protein in E. coli HB101 dependent on transformation with pNO14. The techniques used for the immunoblot analysis may have been insufficiently sensitive to detect NoxR1 when it is expressed at low levels. Immunoblots were completely negative with M. smegmatis pOLYG-NO14 and M. tuberculosis CB3.3 (data not shown). We then tried to overexpress NoxR1 in M. smegmatis using the hsp60 promoter. As in E. coli, overexpression of the hsp60–noxR1 translational fusion impaired the growth of M. smegmatis (data not shown). Next, E. coli HB101 were transformed either with wild-type noxR1 or with a mutant in which a single bp change introduced a premature stop at codon 12 in ORF1, without affecting ORF2 (see Materials and Methods). Wild-type noxR1 encoded by pNO14.1, but not its ORF-1 mutant pNO14.1-mut1, protected the bacteria from ASN (Fig. 3 C). Thus, we have been able to detect, purify, and sequence NoxR1 protein only when a NoxR1 fusion was overexpressed in association with toxicity; we have not been able to detect NoxR1 protein under conditions where lower levels of expression were presumed and a phenotype was conferred. Nonetheless, translation of the noxR1 transcript appears to be required to confer resistance to ASN.

noxR1 also Confers Resistance to ROI and H⁺.

To more fully explore the phenotype afforded by expression of noxR1, we made use of a fluorescent, dye-reduction microplate assay whose results corresponded almost perfectly (correlation coefficients r² >0.96) to the results of the more laborious colony-forming agar-plate assay after exposing bacteria to RNI in vitro or to the intraphagosomal milieu of macrophages (29). Not only was E. coli HB101 rendered far more resistant to ASN (2.5 mM NaNO₂, pH 5) by expression of noxR1 (data not shown), but the bacteria also better resisted S-nitrosoglutathione (GSNO; 2 mM, pH 5.0), a physiologic and bacteriostatic (37, 43, 44) source of several RNI, including ammonia (45; Fig. 4, A and C). By 6 h, the survival advantage was close to 100-fold. GSNO was more bactericidal at pH 5.0 (Fig. 4, A and C) than at pH 7.0 (Fig. 4 E), but noxR1 conferred protection under both conditions.

Unexpectedly, noxR1 also conferred resistance to H₂O₂ (survival advantage, >100-fold) by 6 h, and in this case H⁺ also appeared to show a small synergistic effect on the killing (pH 5.0, Fig. 4, B and D; pH 7.0, Fig. 4 F). Furthermore, noxR1 protected E. coli from the synergistic cytotoxicity afforded by GSNO plus H₂O₂ at concentrations of each agent that were harmless singly (Fig. 4 G), and from the cytotoxicity afforded by the combination of three species likely to be simultaneously present in some phagosomes: GSNO, H₂O₂, and H⁺ (Fig. 4 H).

The results to this point did not allow us to distinguish whether noxR1 might protect relatively specifically against RNI and ROI, or might instead confer a general survival or repair function effective against virtually any threat to bacterial viability. The distinction can be hard to draw, since so many forms of insult, such as heat shock and ultraviolet irradiation, lead directly or indirectly to formation of ROI. Nevertheless, we subjected noxR1- and vector-transformed E. coli to several other stresses: elevated concentrations of H⁺, sodium chloride, detergent, ethanol, nalidixic acid, paraquat, hypochlorous acid, and heat. noxR1 did confer relative resistance to acid at pH 4.5, the lowest level reported in phagosomes (46) as well as at pH 4.0 (Fig. 5 A). At pH 5 and above, no growth inhibition was detectable (Fig. 1 B and data not shown). In addition, noxR1 protected E. coli against killing by hypochlorous acid (10 mM). In contrast, noxR1 afforded no protection against the growth-inhibiting effects of high salt, SDS, heat, ethanol, nalidixic acid, or paraquat (Fig. 5, B–E).

Independence of noxR1 Effects from the oxyR and soxRS Regulons.

oxyR and soxRS are multigenic regulons (25), each of which is activated by and confers resistance to both ROI and RNI in E. coli (47–49). In M. tuberculosis, oxyR is disrupted and soxRS is undescribed (9, 10, 50, 51), whereas in E. coli neither regulon contains any sequence homologous to noxR1. Nonetheless, we wished to test whether noxR1 from M. tuberculosis might function in recombinant E. coli through activation of oxyR or soxRS. The genes controlled by these factors number at least 19, including those encoding superoxide dismutase, NADPH:ferredoxin oxidoreductase, fumarase, DNA repair endonuclease IV, catalase, alkylhydroperoxide reductase, and glutathione reductase. Accordingly, we transformed E. coli deficient in either of these two regulons and their corresponding wild-type strains with pNO14 or pBS vector alone. In all four hosts, noxR1 conferred resistance to ASN (Fig. 6). The degree of protection depended on the host cell type, but not on its expression of oxyR or soxRS. Thus, genes dependent upon oxyR or soxRS are dispensable for the function of noxR1, unless noxR1 can substitute for oxyR or soxRS to induce their expression.

Effect of noxR1 on Survival of M. smegmatis within Macrophages.

The observation that noxR1 protects M. smegmatis from the antibacterial effects of RNI and ROI in vitro prompted us to explore the effect of this gene on the survival of M. smegmatis inside activated macrophages, where the full complement of antibacterial mechanisms is undefined.

Macrophages were collected from the peritoneal cavity of mice 4 d after intraperitoneal injection of sodium periodate (52). Periodiate, a lymphocyte mitogen, stimulates cytokine production (53). Macrophages from periodate-injected mice have a respiratory burst capacity typical of macrophages activated by infection of the host with mycobacteria (54) or macrophages treated in vitro with cytokines (55). Periodate-elicited macrophages respond to further inductive signals, such as in vitro incubation with IFN-γ, by producing NO (33). Both properties were confirmed in wild-type macrophages in this experiment; in addition, M. smegmatis alone was sufficient to trigger NO production in periodate-elicited macrophages without exposure to exogenous cytokines in vitro (Fig. 7, insets).

To vary the genotype of the macrophages along with that of the mycobacteria, wild-type mice of C57BL/6x129/ SvEv background were matched with NOS2-deficient mice on the same background (31). Similarly, wild-type C57BL/6 mice were compared with Phox91-deficient (respiratory burst, oxidase-null) mice backcrossed to C57BL/6. As expected, H₂O₂ production was preserved and NO production was abolished in macrophages from NOS2-deficient mice, while the reverse was the case in macrophages from Phox91-deficient mice; neither class of knock-out macrophages overproduced the opposite product (Shiloh, M.U., unpublished data). In all settings described below, results were similar with or without addition of exogenous IFN-γ to the already activated macrophages (data not shown), and have been pooled.

On average, activated macrophages from wild-type C57BL/ 6x129/SvEv mice killed 80% of ingested pOLYG-transformed (control) M. smegmatis by 5.6 h and 97% by 24 h (Fig. 7 A; interpolation from data in Table 1). When M. smegmatis was transformed with noxR1, it took the wild-type macrophages twofold longer to kill 80% of them. Moreover, at each time point tested (6, 12, and 24 h), approximately twofold more bacteria survived (Fig. 7 A; Table 1).

Macrophages that were genetically incapable of expressing NOS2 (31) were no less efficient at killing M. smegmatis than wild-type macrophages of the same genetic background. 80% of the bacteria were killed by 6 h and 96% were killed by 24 h. (Fig. 7 B; Table 1). This indicated that RNI do not play a major role in the control of M. smegmatis in vitro, in contrast to the situation with M. tuberculosis (see references 16 and 17, and the references cited therein) and M. leprae (56). Nonetheless, expression of noxR1 protected the bacteria, delaying the time of 80% killing by a factor of 2.7-fold and resulting in 1.6–1.7-fold more surviving organisms at each time point tested (Table 1). This indicated that the protective action of noxR1 is not directed exclusively against RNI, consistent with the evidence in vitro showing that noxR1 also protects against ROI.

Wild-type C57BL/6 macrophages killed wild-type (pOLYG-transformed) M. smegmatis more slowly (80% killing by 12.5 h) than did wild-type C57BL/6x129/SvEv macrophages (80% killing by 5.6 h). Nonetheless, expression of noxR1 in M. smegmatis delayed the 80% killing time by a factor of 1.75-fold and resulted in 1.6–2.5-fold more surviving organisms at each time point tested (Fig. 7 C; Table 1).

C57BL/6 macrophages deficient in Phox91 were strikingly impaired in killing wild-type M. smegmatis; within the 24 h period of observation, 80% killing was not often attained (Fig. 7 D; Table 1). This indicated that ROI play a prominent, albeit not an exclusive, role in killing M. smegmatis, in contrast to the situation with M. tuberculosis, where a minimal role for ROI was evident (16). In ROI-deficient macrophages, M. smegmatis that expressed noxR1 survived 1.4–1.9-fold better than vector-transformed M. smegmatis at each time point tested (Fig. 7 D; Table 1). These findings suggested that ROI and another product represent redundant killing mechanisms for M. smegmatis, the former more effectively than the latter; in the absence of ROI, less potent killing by RNI, H⁺, or another product is manifest, against which noxR1 affords protection.