S-nitroso proteome of Mycobacterium tuberculosis: Enzymes of intermediary metabolism and antioxidant defense
- *Division of International Medicine and Infectious Diseases, Department of Medicine, and §Department of Microbiology and Immunology, Weill Medical College of Cornell University, New York, NY 10021; †Protein Center, Sloan–Kettering Institute, New York, NY 10021; and Programs in ‡Molecular Biology and ¶Immunology, Weill Graduate School of Biomedical Sciences of Cornell University, New York, NY 10021
-
Edited by Louis J. Ignarro, University of California School of Medicine, Los Angeles, CA (received for review August 19, 2004)
Abstract
The immune response to Mycobacterium tuberculosis (Mtb) includes expression of nitric oxide (NO) synthase (NOS)2, whose products can kill Mtb in vitro with a molar potency greater than that of many conventional antitubercular agents. However, the targets of reactive nitrogen intermediates (RNIs) in Mtb are unknown. One major action of RNIs is protein S-nitrosylation. Here, we describe, to our knowledge, the first proteomic analysis of S-nitrosylation in a whole organism after treating Mtb with bactericidal concentrations of RNIs. The 29 S-nitroso proteins identified are all enzymes, mostly serving intermediary metabolism, lipid metabolism, and/or antioxidant defense. Many are essential or implicated in virulence, including defense against RNIs. For each of two target enzymes tested, lipoamide dehydrogenase and mycobacterial proteasome ATPase, S-nitrosylation caused enzyme inhibition. Moreover, endogenously biotinylated proteins were driven into mixed disulfide complexes. Targeting of metabolic enzymes and antioxidant defenses by means of protein S-nitrosylation and mixed disulfide bonding may contribute to the antimycobacterial actions of RNIs.
Nitric oxide (NO) synthase (NOS)2 is critical for control of acute and chronic infection by Mycobacterium tuberculosis (Mtb) in mice and is expressed in macrophages of tuberculous lesions in humans (1–5). The reactive nitrogen intermediate (RNI) products of NOS2 exert potent time- and concentration-dependent mycobactericidal activity in vitro (5). However, immunity to Mtb is rarely sterilizing and may be undermined by the ability of Mtb to detoxify RNIs. If we knew which enzymes are targeted by RNI to suppress or kill Mtb, we might be able to inhibit these same enzymes with chemicals impervious to Mtb's anti-RNI defenses.
Two major reactions of RNIs with cell constituents are oxidation and S-nitrosylation (6–8). S-nitrosylation of specific cysteine residues in microbial proteases and Zn-dependent DNA-binding proteins accompanies the antimicrobial activities of RNI against coxsackievirus, HIV, Leishmania, Plasmodium, Trypanosoma, and Salmonella (7, 9–14). Despite the widespread role of RNI as antimicrobial and regulatory molecules (15), however, no systematic identification of S-nitroso proteins in any organism has been described.
We adapted a biotin-switch method that selectively replaces the NO moieties of target protein cysteine nitrosothiols with a disulfide linkage to a biotin derivative by means of a 1,6-diaminohexane spacer N-[6-(biotinamido)hexyl]-3′-(2′-pyridyldithio)propionamide (biotin-HPDP) (16). This method permits the detection of S-nitroso proteins by anti-biotin immunoblotting and their purification by streptavidin affinity chromatography. The work herein describes the identification of the Mtb S-nitroso proteome in an effort to define one set of specific targets of RNIs that may help explain the potent antimycobacterial properties of RNIs and identify potential targets for the development of novel antitubercular drugs.
Methods
Mycobacterial Growth Conditions and Lysate Preparation. Mtb H37Rv was grown in 7H9 broth with 10% oleic acid, albumin, dextrose, and catalase (OADC) media supplement (Difco). Cells were washed twice in PBS (pH 7.2) containing 0.05% Tween 80 and resuspended in PBS (pH 5.5) supplemented with 10% 7H9 broth containing 10% OADC to an OD at 580 nm of 0.3 absorbance units. Bacteria remained fully viable as assessed by colony-forming units and evenly dispersed over the course of 3 days under these conditions (data not shown) (17, 18). Bacteria were exposed to a final concentration of 30 mM sodium nitrite for 12 h at 37°C.
For macrophage infections, 4 × 107 bone marrow-derived macrophages from wild-type or NOS2-deficient C57BL/6J mice were incubated in the absence and presence of 100 units/ml IFN-γ with 4 × 108 colony-forming units of M. tuberculosis H37Rv. After 4 h at 37°C, extracellular bacteria were removed by washing twice with prewarmed PBS followed by replacement of tissue culture media. Conditioned media sampled 16 h later were assayed for nitrite accumulation by using the Griess assay (19). Extracellular bacteria were removed again before lysis of macrophages by washing twice with prewarmed PBS. Intracellular bacteria were recovered by lysing the macrophages with 0.5% Triton X-100.
Bacterial lysates were maintained in the dark at or <4°C. Bacterial lysates were prepared by disrupting cells suspended in 50 mM Hepes, pH 6.5/1 mM EDTA/5 mM neocuproine, using Zirconia beads in a Mini Bead Beater (Stratech Scientific, Luton, U.K.) for 10 s at maximum setting.
Biotin-Switch Assay and Protein Identification. The biotin-switch assay (16) used anti-biotin mAb from Sigma. Biotin-labeled proteins were purified by preadsorbing the biotin-HPDP-labeled lysates onto Sepharose 4B, incubating protein lysates with agarose beads coupled to streptavidin and eluting biotin-HPDP-labeled proteins with the addition of 2-mercaptoethanol (2-ME) to a final concentration of 100 mM. Endogenously biotinylated proteins were purified from bacterial lysates by incubation using agarose beads coupled to monomeric avidin (Pierce) and eluted with free d-biotin at a final concentration of 2 mM. Purified proteins were separated by SDS/PAGE across an 8% polyacrylamide gel. Differentially detected species were excised, digested with trypsin, batch-fractionated on RP microtips, and the peptide mixtures were analyzed by MALDI-reTOF MS (UltraFlex TOF/TOF, Bruker Daltonics, Billerica, MA) (20, 21). Selected experimental masses (m/z) were taken to search a nonredundant protein database (NR; ≈1.7 × 106 entries; National Center for Biotechnology Information, Bethesda), by using the peptidesearch (M. Mann, Center for Eksperimentel Bioinformatik, University of Southern Denmark, Odense, Denmark) algorithm. MS sequencing of selected peptides was performed by MALDI-TOF/TOF (MS/MS) analysis of the same samples, using the UltraFlex instrument in “LIFT” mode. Fragment ion spectra were taken to search the NR by using the MASCOT MS/MS Ion Search program (Matrix Science, London). A molecular mass range of up to twice the predicted was covered, with a mass accuracy restriction >50 ppm.
Enzyme Assays. Mtb lipoamide dehydrogenase (Lpd) was overexpressed in Escherichia coli by using its natural coding sequence, purified, and assayed as described (22). Purified, recombinant Lpd was nitrosylated by incubation with 1 mM S-nitrosoacetylpenicillamine (SNAP) for 30 min at room temperature in the dark. SNAP and product N-acetylpenicillamine were removed by serial passage across a MicroBioSpin 6 gel filtration device (Bio-Rad).
Purified recombinant mycobacterial proteasome ATPase (Mpa) containing a C-terminal hexahistidine affinity tag was provided as a generous gift from G. Lin in our laboratory and assayed as described (23, 24). Purified, hexahistidine-tagged Mpa was nitrosylated by incubation with 1 mM S-nitrosoglutathione (GSNO) for 30 min at room temperature in the dark. Unincorporated GSNO and product glutathione were removed by serial passage across a MicroBioSpin 6 gel filtration device (Bio-Rad).
Results and Discussion
We began by recovering intact Mtb from infected bone marrow-derived primary mouse macrophages to evaluate whether protein S-nitrosylation occurs in vivo. IFN-γ-activated wild-type macrophages produced NO (part of which underwent spontaneous oxidation to yield nitrite), and Mtb recovered from these cells contained at least 10 S-nitrosylated proteins. IFN-γ-treated NOS2-/- macrophages, used as a control, produced no nitrite, and Mtb recovered from these cells lacked detectable S-nitrosylated proteins (Fig. 1a). Thus, Mtb proteins are physiologically S-nitrosylated in the phagosome of immunologically activated macrophages.
S-nitrosylation of Mtb proteins in IFN-γ-treated macrophages and nitrite-treated bacterial cultures. Shown are anti-biotin immunoblots of Mtb lysates processed by means of the biotin-switch method. Numbers to left are M r markers. (a) Intracellular Mtb was recovered 16 h postinfection of resting (-) and IFN-γ-stimulated (+) (100 units/ml) bone marrow-derived macrophages from wild-type or NOS2-deficient C57BL/6J mice. Arrows denote species whose detection depended on both IFNγ and NOS2. Numbers beneath the lanes indicate concentrations of nitrite detected in the conditioned media. (b) S-nitrosylated proteins formed in intact Mtb exposed to nitrite under conditions that reduced the viable count by a factor of 104 without lysing the bacteria. Arrows denote biotin-HPDP-specific signals indicative of S-nitrosylated proteins. Stars mark endogenously biotinylated proteins whose signals increased (filled stars) or decreased (open stars) after exposure to nitrite. (Right) Confirmation that these signals arose from endogenous rather than exogenous biotin is shown. Two asterisks, on the far left, indicate background signals arising from incomplete alkylation (blocking) of cysteine sulfhydryls. (c) Dose-dependent and RNI-specific origin of anti-biotin signals. Symbols are as in b.
To identify S-nitroso proteins in Mtb systematically, it was necessary to scale-up, making it impractical to use macrophages as RNI generators.
Choice of an in vitro RNI-generating system was based on the following considerations. NOS2-derived RNIs are slowly but potently bactericidal against
Mtb in vitro (5). NOS2-deficient mice fail to control replication of Mtb, succumbing to infection more rapidly than their wild-type counterparts
in acute infections (1, 3). During the chronic phase of infection, Mtb-infected wild-type mice fed a highly specific NOS2 inhibitor succumb quickly
with a high bacillary burden, whereas counterparts fed an inactive enantiomer continue to control the infection (1). The composition of the milieu in the Mtb-containing phagosome is unknown. However, the Mtb phagosome of activated macrophages
is acidic (25) and is expected to contain nitrite, a freely diffusible product of NO autoxidation. Transcriptional adaptation of Mtb within
the phagosome of the activated macrophage depends on macrophage expression of NOS2, and a large component of that response
can be mimicked by exposing cultures of Mtb to mildly acidified nitrite (26). In contrast, the phagocyte oxidase is not required for mice to control Mtb infection (25), in part because of the protection afforded by Mtb's katG-encoded KatG peroxidase (27). Finally, the brief activity of phagocyte oxidase after uptake of Mtb (28) precedes the induction of NOS2, limiting the amount of peroxynitrite that might arise in the phagosome from the mutual reaction
of 
and NO, the short-lived products of those two enzymes. Peroxynitrite often leads to protein tyrosine nitration, but we could
detect no such modification of Mtb proteins in intraphagosomal Mtb (data not shown). In short, mildly acidified nitrite is
likely to reproduce a major component of the nitrosative and oxidative environment that Mtb encounters over a prolonged period
in the phagosome of activated macrophages.
Accordingly, suspensions of Mtb were treated with nitrite at pH 5.5. As the acidity of the medium approaches a pH of ≈4.5, similar to that of an Mtb-containing phagosome in an IFN-γ-activated macrophage (25), nitrite is partially protonated to nitrous acid (pKa ≈3.8), which dismutates to generate NO and •NO2 (15). At pH 5.5, 0.5 mM nitrite in 0.5 ml generates about as much NO as 3 × 105 activated macrophages over 24 h (29). Further, the concentration of nitrite required to kill Mtb increases with bacterial culture density and decreases with time of exposure (30). Thus, to recover sufficient biomass while keeping Mtb intact, we used millimolar concentrations of nitrite and short treatment times, then removed nitrite and any adsorbed proteins with extensive detergent washes and concentrated bacteria by centrifugation. Exposure of Mtb at 1.5 × 108 colony forming units/ml to 30 mM sodium nitrite (pH 5.5) for 12 h at 37°C produced a 4 log10 reduction in colony-forming units without loss of sedimentable organisms. Treatment with an equimolar concentration of nitrate (pH 5.5), which does not generate RNIs, did not affect bacterial viability and served as a control (30, 31).
Mtb treated under these conditions was then physically disrupted and subjected to the biotin-switch assay (Fig. 1b). Omission of the exogenous biotinylated reagent revealed three endogenously biotinylated species. Peptide mass fingerprinting confirmed two of these species to contain the biotinylated subunit of acetyl CoA carboxylase (AccA3) (encoded by the accA3 gene) and one to contain pyruvate carboxylase (Pca) (encoded by the pca gene). These biotinylated proteins are conserved in all organisms (32, 33). With the inclusion of the biotin label, multiple S-nitrosylated proteins were observed, but only in lysates prepared from Mtb exposed to nitrite. These effects were concentration-dependent and no S-nitrosylation was observed in the presence of an equimolar concentration of nitrate (Fig. 1c). S-nitrosylated proteins detected in nitrite-treated cultures of Mtb included species migrating at the same apparent M r as Mtb proteins S-nitrosylated in the phagosome of IFN-γ-activated macrophages. Unexpectedly, signals from the endogenously biotinylated proteins of Mtb were also affected by nitrite in a dose- and RNI-dependent manner, as will be discussed subsequently.
S-nitrosylated proteins were enriched by streptavidin-based affinity chromatography, selectively released from their sulfhydryl-derivatized biotin labels with 2-ME and identified by a combination of peptide mass fingerprinting using MALDI-TOF MS, and MS sequencing using MALDI-TOF/TOF MS/MS. In this way, 29 proteins were unambiguously identified (Table 1). For simplicity, these proteins will be referred to as the S-nitroso proteome, recognizing limitations inherent to proteomic techniques, such as a bias against trace and insoluble proteins.
The S-nitroso proteome of Mtb proved to be highly enriched for proteins encoded by genes predicted to be essential or required for optimal growth based on transposon-mediated insertional mutagenesis (62% of the S-nitroso proteome vs. 15% of the Mtb genome) (34). Moreover, the Mtb S-nitroso proteome is comprised predominantly of enzymes involved in intermediary and lipid metabolism. This functional composition is disproportionate to the distribution of Mtb genes encoding such functions. For example, 62% of the enzymes in the S-nitroso proteome but only 30% of ORFs in the Mtb genome are involved in intermediary metabolism. Likewise, 14% of the enzymes of the S-nitroso proteome but only 6% of the ORFs in the Mtb genome are annotated to function in lipid metabolism. The composition of the S-nitroso proteome also does not correspond to the most abundant proteins of the input lysates (seen on 1D gels) or the most abundant proteins of Mtb seen in published 2D gels (55) (data not shown).
Five proteins (19%) of the S-nitroso proteome (phosphoenolpyruvate carboxykinase, Mpa, glutamine synthetase, mycocerosic acid synthase, and acetohydroxyacid synthase) have been individually validated as important for virulence and/or persistence in animal models of tuberculosis through targeted gene disruption or enzymatic inhibitor studies (24, 31, 35–38). Ten proteins (≈50% of the S-nitroso proteins involved in intermediary or lipid metabolism) are involved in pathways important for the virulence and/or persistence of Mtb: mycolic acid synthesis (mycocerosic acid synthase, polyketide synthase 13, and fatty acyl-AMP ligase), gluconeogenesis (phosphoenolpyruvate carboxykinase and malate synthase), branched chain amino acid synthesis (acetohydroxyacid synthase), nitrogen assimilation (asparagine synthase, glutamine synthetase, and glutamate synthase), and iron metabolism (mycobacterial ortholog of bacterioferritin) (35–41).
Two enzymes, Lpd and Mpa, were specifically implicated in the defense of Mtb against peroxynitrite, peroxides, or nitrite (24, 31, 42). Also identified was catalase (KatG), an enzyme crucial to Mtb's defense against reactive oxygen intermediates as revealed when NOS2 is absent (27). Thus, in addition to enzymes of intermediary and lipid metabolism, the Mtb S-nitroso proteome includes enzymes that participate in Mtb's defenses against oxidative or nitrosative injury.
Three (DnaK, GroEL2, and Tuf) of the 29 proteins (10%) are devoid of cysteine residues and might be considered false-positives of the assay (33). However, these proteins were not detected in lysates prepared from nonnitrosylated bacteria and they specifically eluted from streptavidin columns with the addition of 2-ME. Thus, these proteins appear to have been captured by complexing with S-nitrosylated proteins through their chaperonin activities (43–45). Remarkably, in contrast to their homologs in E. coli, 11 of the 16 (69%) predicted heat shock/chaperonin proteins of Mtb, including GroEL, DnaK, and Tuf, are devoid of cysteine residues (33) and this fact may imply a relative resistance of the Mtb chaperone machinery to oxidative stress.
As noted earlier, treatment of Mtb with nitrite also induced dose-dependent increases in the anti-biotin immunoreactivity of one endogenously biotinylated species and dose-dependent decreases of two others (Figs. 1c and 2). These effects were not seen in lysates prepared from Mtb treated with equimolar concentrations of nitrate. With the addition of the thiol-specific reductant, 2-ME, in the absence or presence of additional nitrite, the biotinylated species of the highest apparent M r disappeared with an accompanying increase in the signal intensities of the two smaller biotinylated species. Thus, the largest biotinylated species is a disulfide-bonded complex containing a biotinylated protein. In the presence of nitrite, formation of this complex was enhanced. These biotinylated species were purified by monomeric avidin chromatography and identified by peptide mass fingerprinting. The two smaller species were identified as the biotin-containing subunit of acetyl CoA carboxylase (encoded by accA3), which catalyzes the conversion of acetyl CoA to malonyl CoA, the first committed step in fatty acid biosynthesis, and pyruvate carboxylase (encoded by pca), which catalyzes the conversion of pyruvate to oxaloacetate and serves an essential anaplerotic role during periods of metabolic shunting. The largest biotinylated species contained both AccA3 and catalase (encoded by katG), a protein involved in antioxidant defense, also identified as a target for S-nitrosylation. AccA3 and Pca are structurally homologous, biotin-dependent, multidomain complexes that contain highly conserved cysteine residues, and whose assembly is associated with the self-interaction of two biotin-containing domains (encoded by accA3 and pca, respectively) (46, 47). In solution, KatG similarly exists as a homodimer and contains a highly conserved cysteine (48). Thus, it is possible that, within the cell, RNIs can oxidize cysteine thiols to generate mixed intermolecular disulfide bonds (49). Such mixed disulfide formation may interfere with the proper assembly and function of oligomeric protein complexes, such as AccA3 and Pca (46, 47), or alter the stability of proteins such as KatG, in which isosteric mutation of a highly conserved cysteine to serine resulted in increased proteolytic degradation (50).
RNI-induced intermolecular disulfide formation involving AccA3 and Pca. Anti-biotin immunoblot of endogenously biotinylated Mtb proteins demonstrates nitrite-induced, thiol-sensitive complex formation of endogenously biotinylated Mtb proteins. The anti-biotin signal α was identified by peptide mass fingerprinting as the biotin carboxylase-biotin carboxyl carrier protein subunit of AccA3 and β as Pca. The species labeled γ consisted of both AccA3 and KatG.
The techniques used to define the S-nitroso proteome do not reveal the proportion of each target protein that is S-nitrosylated. However, because the total quantity of AccA3 and Pca is represented by their anti-biotin signals, the impact of RNIs on the endogenously biotinylated proteins of Mtb is revealed as complete. The nitrite-dependent loss of the anti-biotin signal from Pca may reflect the formation of a mixed disulfide with Pca that masks its biotin epitope. Thus, in addition to S-nitrosylation and methionine oxidation (6), nitrite can give rise to species that oxidize cysteine thiols, leading to the formation of mixed, intermolecular, disulfide-bound complexes.
To test the functional impact of S-nitrosylation on protein function, we selected two proteins for study in vitro, Lpd (encoded by Rv0462) and Mpa (encoded by Rv2115c). Lpd is the E3 component of the pyruvate dehydrogenase complex and also forms an integral component of Mtb's peroxynitrite reductase/peroxidase (42). Transposon-mediated insertional mutagenesis identified lpd as a gene whose function in Mtb is either essential or required for normal growth (34). The catalytic activity of Lpd depends on two highly conserved cysteine residues (22). mpa encodes a mycobacterial proteasome-associated ATPase recently identified as a component of Mtb's RNI defense system (24, 31). Mpa contains three cysteine residues, all of which are conserved in Mycobacterium leprae, and one that lies close to the ATP-binding site (33).
Purified, recombinant Lpd or Mpa was incubated with a molar excess of SNAP or GSNO, respectively, and S-nitrosylated, as confirmed by the biotin-switch method (Fig. 3 a and c). S-nitrosylation was associated with inhibition of both substrate-dependent, Lpd-mediated consumption of NADH (Fig. 3b) and Mpa-mediated hydrolysis of ATP (Fig. 3c). S-nitrosylation and enzymatic inhibition of both Lpd (Fig. 3 a and b) and Mpa (data not shown) were concentration-dependent and reversed by treatment with the thiol-specific reductant, 2-ME (data not shown). There was no inhibition of enzyme activity after incubation of Lpd or Mpa with the nitrosylation-incompetent derivatives (N-acetylpenicillamine or glutathione, respectively) or vehicle alone (Fig. 3 b and c). Thus, S-nitrosylation inhibits the enzymatic activities of both Lpd and Mpa.
S-nitrosylation is associated with inhibition of dihydrolipoamide dehydrogenase (Lpd; Rv0462) and mycobacterial proteasome-associated ATPase (Mpa; Rv2115c). (a Left Upper) Anti-biotin immunoblotting of purified, recombinant Mtb Lpd incubated with SNAP, N-acetylpenicllamine (AP), or vehicle (H2O) and assayed by using the biotin-switch method. (Right Upper) Anti-biotin immunoblotting of purified, recombinant Lpd incubated with vehicle, 2 mM SNAP, or 2 mM SNAP, followed by treatment with the thiol-specific reductant, 2-ME, and assayed using the biotin-switch method. (Images presented were taken from a single gel with intervening lanes removed.) (Lower) Ponceau S staining of nitrocellulose membrane before anti-biotin immunoblotting. (b) Effects of S-nitrosylation on Lpd-catalyzed lipoamide-dependent NADH consumption in the presence of a saturating concentration of substrate under initial rate conditions. Data are means ± SD for four independent experiments. (c Left) Anti-biotin immunoblotting of purified, recombinant Mtb Mpa incubated with GSNO, glutathione (GSH), or vehicle (H2O) and assayed by using the biotin-switch method. (Right) Effects of S-nitrosylation on Mpa-catalyzed ATP hydrolysis in the presence of a saturating concentration of substrate under initial rate conditions. Means ± SD for three independent experiments are shown.
What is the biophysical basis for the selectivity of the S-nitroso proteome? NO or nitrosonium ion (NO+) reacts preferentially with the thiolates of cysteine residues when the pKa of the sulhydryl is lowered by nearby cationic side chains on other amino acids (7). Moreover, micellar catalysis favors S-nitrosylation in hydrophobic environments where NO can accumulate (51). Thus, S-nitrosylation may be favored on cysteine thiolates in protein interiors. These conditions characterize cysteines that participate in active-site chemistry. Enzymes that use active-site cysteines often have redox activity. Together, these features may explain, in part, why so few proteins were found to be S-nitrosylated, why so many of the proteins are enzymes, and why so many of the enzymes are redox-active. Indeed, of the >100 S-nitroso proteins identified through in vitro studies of individual proteins and proteomic surveys of brain, mesangium, endothelium, and mitochondria, nearly 60% represent enzymes, among which are the mammalian orthologs of Lpd and aconitase (7, 52–54). The remainder of described S-nitroso proteins are largely proteins involved in channels, transport, structure, DNA binding, and storage, including the mammalian ortholog of bacterioferritin (7).
RNIs are broadly active as antimicrobial molecules, but candidate targets have not been identified systematically. The remarkable convergence of the Mtb S-nitroso proteome with known virulence and/or persistence factors important to the pathogenesis of tuberculosis supports a significant role for protein S-nitrosylation as an effector of the antimicrobial activities of RNIs. The relative contribution of S-nitrosylation to the antimicrobial activities of RNIs awaits detailed studies of individual target proteins to evaluate the specificity and stoichiometry of S-nitrosylation-dependent effects. That RNIs can target three proteins important in the antioxidant defense system of Mtb (Lpd, Mpa, and KatG) emphasizes the dynamic interplay between the pathogen's efforts to subvert host immune responses and the host's efforts to disable these same microbial defenses (27, 31, 42). The identification of evolutionarily selected targets inhibited in Mtb by the host immune system may point to useful targets for the development of antitubercular drugs.
Acknowledgments
We thank G. Lin for purified, recombinant Mpa; R. Bryk for expert advice; and W. D. Johnson, Jr., for enthusiastic support of this work. This work was supported by National Institutes of Health Grants P01AI56293 (to C.F.N.) and T32AI007613 and K08AI0061393 (to K.Y.R.). The Department of Microbiology and Immunology is supported by the William Randolph Hearst Foundation.
Footnotes
-
↵ ∥ To whom correspondence should be addressed at: Department of Microbiology and Immunology, Weill Medical College of Cornell University, 1300 York Avenue, Box 65, New York, NY 10021. E-mail: cnathan{at}med.cornell.edu.
-
Author contributions: K.Y.R. and C.F.N. designed research; K.Y.R. and H.E.-B. performed research; K.Y.R., H.E.-B., P.T., and C.F.N. analyzed data; and K.Y.R. and C.F.N. wrote the paper.
-
This paper was submitted directly (Track II) to the PNAS office.
-
Abbreviations: Mtb, Mycobacterium tuberculosis; NOS, NO synthase; RNI, reactive nitrogen intermediate; biotin-HPDP, N-[6-(biotinamido)hexyl]-3′-(2′-pyridyldithio)propionamide; 2-ME, 2-mercaptoethanol; Lpd, lipoamide dehydrogenase; SNAP, S-nitrosoacetylpenicillamine; Mpa, mycobacterial proteasome ATPase; GSNO, S-nitrosoglutathione; KatG, catalase; pca, pyruvate carboxylase; accA3, acetyl CoA carboxylase.
- Copyright © 2005, The National Academy of Sciences
