Flavorubredoxin, an inducible catalyst for nitric oxide reduction and detoxification in Escherichia coli.

Nitric oxide (NO) is a poison, and organisms employ diverse systems to protect against its harmful effects. In Escherichia coli, ygaA encodes a transcription regulator (b2709) controlling anaerobic NO reduction and detoxification. Adjacent to ygaA and oppositely transcribed are ygaK (encoding a flavorubredoxin (flavoRb) (b2710) with a NO-binding non-heme diiron center) and ygbD (encoding a NADH:(flavo)Rb oxidoreductase (b2711)), which function in NO reduction and detoxification. Mutation of either ygaA or ygaK eliminated inducible anaerobic NO metabolism, whereas ygbD disruption partly impaired the activity. NO-sensitive [4Fe-4S] (de)hydratases, including the Krebs cycle aconitase and the Entner-Doudoroff pathway 6-phosphogluconate dehydratase, were more susceptible to inactivation in ygaK or ygaA mutants than in the parental strain, and these metabolic poisonings were associated with conditional growth inhibitions. flavoRb (NO reductase) and flavohemoglobin (NO dioxygenase) maximally metabolized and detoxified NO in anaerobic and aerobic E. coli, respectively, whereas both enzymes scavenged NO under microaerobic conditions. We suggest designation of the ygaA-ygaK-ygbD gene cluster as the norRVW modulon for NO reduction and detoxification.

In the accompanying article (17), we provide evidence for an inducible and robust NO-metabolizing and -detoxifying activity in anaerobic Escherichia coli. Attempts to biochemically identify the NO reduction system have been complicated by its instability. Moreover, the E. coli genome lacks a NOR belonging to either the cytochrome bc complex or cytochrome P450 families (1). The list of proteins displaying a reductase activity for NO in vitro with potential for function in E. coli is long and includes flavohemoglobin (flavoHb) (27,28), cytochrome c or cЈ (29), multi-heme nitrite reductase (2,30), copper-nitrite reductase (31), bacterioferritin (32), ribonucleotide reductase (33), Cu,Zn-superoxide dismutase (34), and terminal respiratory oxidases (35). However, none of these candidate systems, including NO-inducible flavoHb (17), have a demonstrated NO reduction function in cells. Moreover, unlike the inducible NOR in E. coli (17), none of these "NO reductases" are sensitive to inactivation by O 2 .
Genomic data and bioinformatics tools (NCBI Protein Database and BLAST) provided a strategy for identifying the E. coli system. E. coli ygaA encodes a protein bearing ϳ42% identity to the NO-modulated Ralstonia eutrophus transcription regulators NorR1 and NorR2 (36). Intriguingly, NorR homologs are also located adjacent to the flavoHb gene (hmp) in both Vibrio cholera (37) and Pseudomonas aeruginosa (38), suggesting a common control for NO detoxification systems in various organisms. Adjacent to ygaA and transcribed in the opposite direction with specific promoters are the genes ygaK encoding a flavorubredoxin (flavoRb) with a NO-binding diiron center and ygbD encoding a flavoRb reductase (39,40) with potential for a NOR function (see Fig. 1).
We demonstrate here the role of the E. coli NorR homolog YgaA in controlling NOR expression. We also elucidate the role of flavoRb (YgaK) and its reductase partner (YgbD) in the NO-induced anaerobic NOR activity in E. coli. We further demonstrate that NOR and NOD protect NO-sensitive [4Fe-4S]-containing (de)hydratases in critical anabolic and catabolic pathways and thus explain conditional growth defects observed with NO poisoning in E. coli. A mechanism for flavoRb-catalyzed NO reduction is envisioned. We suggest renaming the ygaA-ygaK-ygbD gene cluster as norRVW and annotating similar genes for a possible NO detoxification function.
Insertion of Mu Transposons in the ygaA, ygaK, and ygbD Genes-Disruptions in the ygaA, ygaK, and ygbD genes were created by Mu transposon insertion. MudIIPR13 (cam R ), a generous gift of Dr. Danièlle Touati, was randomly transposed to pAlter9G10 by the method of Castilho et al. (43) as modified by Carlioz and Touati (41). Briefly, strain QC720 carrying a Mucts prophage and MudIIPR13 was transformed at 30°C with pAlter9G10. Mu transposition, phage growth, and cell lysis were induced at 44°C in LB medium prepared with 10 g of Tryptone, 5 g of yeast extract, and 10 g of NaCl prepared in 1 liter of deionized water and supplemented with 20 mM glucose. Phage lysates were used to transduce strain QC719 to ampicillin and chloramphenicol resistance, thus selecting for plasmids carrying a Mudlac transposon. Ampicillin-and chloramphenicol-resistant clones were further scored for ␤-galactosidase fusion phenotype on tetrazolium redlactose agar (44). White ␤-galactosidase-positive colonies were selected and screened for Mudlac insertions in ygaA, ygaK, and ygbD by restriction analysis. To enrich for NO-inducible, anaerobically expressed ␤-galactosidase gene fusions, ampicillin-and chloramphenicol-resistant colonies were replica-plated onto tetrazolium red-lactose agar with antibiotics and grown overnight under an atmosphere containing 180 ppm NO balanced with N 2 . Following an additional 24 h of growth in air, colonies that were originally red, but turned pink upon exposure to NO, were chosen for restriction analysis. Insertion sites were determined by sequencing from the 5Ј-end of the Mudlac insertions using primer 5Ј-AATACATCTGTTTCATTTG-3Ј. Plasmids with insertions at amino acids 311, 232, and 11 in the respective YgaA, YgaK, and YgbD open reading frames were isolated.
Construction of ygaA, ygaK, and ygbD Chromosomal Mutations-Plasmids were linearized with XbaI, and the DNA was used to transform strain JC7623 to chloramphenicol resistance (45). Several chloramphenicol-resistant, ampicillin-sensitive colonies were selected for analysis. Site-specific insertion in the ygaA, ygaK, and ygbD genes was confirmed by linkage to gshA using strain JTG10 as the recipient in P1 transduction analyses. Mutations were subsequently transduced to strain AB1157.
Bacterial Growth and Exposures-Cultures were grown in phosphate-buffered LB medium (8) unless otherwise indicated. Minimal salts medium (pH 7.0) contained 60 mM K 2 HPO 4 , 33 mM KH 2 PO 4 , 7.6 mM (NH 4 ) 2 SO 4 , 1.7 mM sodium citrate, 1 mM MgSO 4 , 10 M MnCl 2 , and 10 g/ml thiamin HCl and was supplemented with either 20 mM glucose or 2% potassium gluconate. Minimal medium was supplemented with 40 g/ml L-arginine, L-histidine, L-leucine, L-proline, and L-threonine or with 0.25% casamino acids as indicated. Overnight 5-ml aerobic cultures were grown at 37°C in 15-ml culture tubes with vigorous shaking. Overnight 10-ml anaerobic cultures were grown static at 37°C in 15-ml culture tubes. For gas exposures, cultures were grown at 37°C in rubber stopper-sealed 50-ml Erlenmeyer flasks continuously flushed with gas mixtures at 30 ml/min at a culture/flask volume ratio of at most 1:5 with vigorous shaking at 275 rpm. To minimize the disturbance of gases, culture aliquots were removed from flasks using a 1-ml tuberculin syringe connected via narrow tubing to the culture medium. Cell density was determined from culture absorbance at 550 nm and by plating and counting. An absorbance of 1.0 was taken to equal 3 ϫ 10 8 and 7 ϫ 10 8 bacteria/ml for cells grown in phosphate-buffered LB medium and minimal salts medium, respectively.
NO Consumption Assays-Aerobic and anaerobic NO consumption activities were measured amperometrically using a 2-mm ISO-NOP NO electrode (World Precision Instruments, Sarasota, FL) as previously described (16,17). Aerobic and anaerobic activities were measured at 37°C and at 1.0 and 1.5 M NO, respectively, unless otherwise indicated.
Statistical Analysis-Significance of differences between data (p Ͻ 0.05) were determined using Student's t test.

RESULTS
Inducible Anaerobic NO Consumption Activity Is Dependent upon ygaA, ygaK, and ygbD-To test the individual roles of the ygaA, ygaK, and ygbD gene products in the anaerobic NO consumption activity, we constructed strains carrying insertion mutations using random Mudlac transposition (Fig. 1). The anaerobic and aerobic NO consumption activities of strains AG200, AG300, and AG400, with Mudlac insertions in ygaA, ygaK, and ygbD, respectively, were measured and compared with those of the parental strain, AB1157. Strains bearing mutations in ygaA or ygaK showed no anaerobic NO consumption activity following exposure to 960 ppm NO under anaerobic growth conditions, whereas the ygbD mutant produced a rate ϳ40% lower than that of its parental strain ( Fig. 2A). None of the strains showed significant anaerobic NO consumption activity in the absence of NO exposure. We also measured the aerobic NO consumption activity of anaerobically induced cells. Each strain showed the normal basal level of aerobic NO consumption activity (Fig. 2B, compare white bars), and this activity was induced to high levels by NO (black bars). The absence of ygaA, ygaK, or ygbD resulted in small increases in the induction of the aerobic NO consumption activity (compare black bars). This aerobic NO consumption activity is fully attributable to flavoHb (20).
These results demonstrate that YgaA (NorR) and YgaK (NorV, flavoRb) are essential for anaerobic NO consumption. Thus, YgaA and YgaK constitute a novel modulon for NO reduction and detoxification in E. coli, with YgbD (NorW, fla-voRb reductase) acting as an accessory for NO reduction. YgaA, YgaK, and YgbD may decrease aerobic NO consumption activities by decreasing steady-state NO levels and flavoHb/NOD expression in anaerobic cells.
Dependence of the Anaerobic NO Consumption Activity upon NO-We examined the efficiency of the anaerobic NO consumption activity for NO scavenging. NOR showed an apparent K m(NO) value of 400 nM (Fig. 3, q) and was CO-resistant (E) and cyanide-resistant (17). The K m(NO) value is identical to the values estimated for the flavoHb-type NOD in E. coli (21) and for the cytochrome bc-type NOR in Pseudomonas perfectomarina (49). The results demonstrate the efficiency of the anaerobic system for NO scavenging. The results also indicate low affinities of the NO scavenger flavoRb (YgaK) for CO and cyanide.
Role of the ygaA, ygaK, and ygbD Gene Products in Protecting the NO-sensitive Aconitase-To define the NO detoxification function of ygaA, ygaK, and ygbD in cells, it is necessary to understand the protective role(s) of the system for critical target(s) of NO poisoning under physiological conditions. Thus, we tested the role of ygaA, ygaK, and ygbD in the inducible anaerobic protection of the NO-sensitive Krebs cycle enzyme aconitase (16,17). E. coli (AB1157) exposed to 480 ppm NO gas (Ͻ1 M) lost ϳ45% of the aconitase activity after 30 min (Fig. 4). Aconitase inactivation increased significantly (p Ͻ 0.05) in the presence of chloramphenicol, thus demonstrating the protective role for newly synthesized protein(s). Moreover, loss of aconitase activity in the ygaA (AG200) and ygaK (AG300) mutants was greater than in the parental strain, thus demonstrating critical roles for ygaA and ygaK in the adaptive protection.
In contrast, ygbD (AG400) was not essential for aconitase protection under these conditions.
Role of the ygaA, ygaK, and ygbD Gene Products in Protecting E. coli from NO-mediated Growth Inhibition-To further define the functional importance of ygaA, ygaK, and ygbD for E. coli, we investigated the effects of specific insertion mutations in each of these genes on the anaerobic growth of E. coli exposed to a NO stress. Surprisingly, there was little growth inhibition with 240 ppm NO gas (Յ0.5 M) for cells growing on a rich phosphate-buffered LB medium (Fig. 5A). Higher NO exposure levels caused comparable growth inhibition of the mutants and the parental strain (data not shown). These results suggest a limited role for ygaA, ygaK, and ygbD and anaerobic NO metabolism in growth protection. Nevertheless, the exquisite sensitivity of aconitase to NO-mediated inactivation (Fig. 4) (8,16) strongly suggested that anaerobic growth protection may be better observed under conditions requiring aconitase function, the citric acid cycle, or other NO-sensitive metabolic pathways. Aconitase expression is relatively low under these conditions; and moreover, aconitase function is not expected to be limiting for E. coli growth with glucose supplied as the substrate for energy production (50).
The effect of NO on anaerobic growth of E. coli mutants was investigated under growth conditions demanding the function of putative NO-sensitive enzymes. The [4Fe-4S]-containing 6-phosphogluconate dehydratase of the gluconate-metabolizing Entner-Doudoroff pathway and the [4Fe-4S]-containing ␣,␤dihydroxyacid dehydratase of the branched-chain amino acid biosynthesis pathway are two enzymes that are predicted to be NO-sensitive (8,51). Anaerobic growth of strains AG200 and AG300 was significantly impaired by exposure to 240 ppm NO gas under growth conditions requiring gluconate metabolism (Fig. 5B) or amino acid biosynthesis (Fig. 5C). These results FIG. 2. NO metabolism in the ygaA, ygaK, and ygbD mutants. Anaerobic (A) and aerobic (B) NO consumption activities were measured for strains AB1157 (parental strain), AG200 (ygaA::lac), AG300 (ygaK::lac), and AG400 (ygbD::lac) grown under N 2 (white bars) or exposed to 960 ppm NO in N 2 (black bars).  4. Anaerobic protection of aconitase by the ygaA, ygaK,  and ygbD (norRVW) genes. The aconitase activities of AB1157 (parental strain), AG200 (ygaA::lac), AG300 (ygaK::lac), and AG400 (ygbD::lac) were measured in cells grown under N 2 and exposed to 480 ppm NO in N 2 for 30 min. Cells were harvested at t ϭ 0 and 30 min and assayed for aconitase activity as described under "Materials and Methods." Percent aconitase activity is expressed relative to aconitase activity measured in N 2 controls. 100% activity equals 11.5, 12.7, 14.0, 14.2, and 10.4 milliunits/mg of protein for AB1157, AB1157 ϩ chloramphenicol (CAM), AG200, AG300, and AG400, respectively. Chloramphenicol (200 g/ml) was added prior to NO exposures as indicated. Static cultures were grown overnight in 50 ml of minimal salts medium containing 10 mM glucose and 40 g/ml each histidine, arginine, proline, leucine, and threonine. Cells were washed and resuspended in 10 ml of minimal salts medium at a final A 550 of 0.6 and were incubated at 37°C for 30 min under N 2 prior to exposures (t ϭ 0). Minimal salts medium was chosen for these experiments to limit variations in aconitase activity that may be linked to rapid culture growth or to changes in substrate utilization (48). Error bars represent the S.E. determined from six independent experiments. Asterisks indicate a significance of p Ͻ 0.05 relative to the value for AB1157. demonstrate important, albeit conditional, roles for YgaA (NorR) and YgaK (NorV) in NO detoxification.
Measurements of 6-phosphogluconate dehydratase activity following a 60-min exposure of parental or flavoRb-deficient cells to NO gas (240 ppm) demonstrated the protection that flavoRb (YgaK) and NO reduction afforded to 6-phosphogluconate dehydratase (Fig. 6, compare white and black bars). The results establish the sensitivity of this [4Fe-4S] enzyme to NO and further demonstrate the protection NO metabolism affords against metabolic NO poisoning.
Both flavoRb and flavoHb Detoxify NO under Microaerobic Growth Conditions-The anaerobic NO consumption activity in E. coli is O 2 -sensitive, decaying with a half-life of ϳ5 min in air (17), suggesting that flavoRb may function poorly, if at all, in the presence of O 2 . On the other hand, the aerobic NOD activity of flavoHb shows a rather high K m value for O 2 (60 -100 M) and is potently inhibited by NO under hypoxia (21,22), suggesting that flavoHb would be a poor catalyst for NO decomposition under O 2 -limiting conditions. Using mutants deficient in flavoRb and/or flavoHb, we compared the ability of flavoRb and flavoHb to function individually or together in NO detoxification under an atmosphere containing 0.5% O 2 (Յ5 M) and 240 ppm NO (Յ0.5 M). These solution concentrations are likely to be encountered by E. coli or other pathogens in tissues and organs. Under these conditions, either flavoRb (AG103) or flavoHb (AG300) alone was sufficient to protect E. coli from significant growth inhibition (Fig. 7). However, in the absence of both flavoRb and flavoHb defenses (AG301), NO caused severe growth arrest. These results support the hypothesis that flavoRb (NOR) and flavoHb (NOD) act together to provide an adequate capacity for NO detoxification in E. coli under varying conditions of O 2 and NO exposure. DISCUSSION Investigations of NO metabolism and toxicity continue to reveal a biological complexity of NO akin to that of O 2 . Indeed, O 2 transport/storage hemoglobins and myoglobins, the O 2 -utilizing cytochrome P450 monooxygenases, the respiratory cytochrome bc complex, and the O 2 -reducing cytochrome aa 3 all appear to have evolved from more ancient NO-or N 2 O-metabolizing enzymes in microbes (1,20,22,52,53).
During our investigations of NO toxicity and defenses in E. coli, we have observed a novel O 2 -sensitive and NO-inducible anaerobic pathway for reductive NO metabolism and detoxification (17). We have now demonstrated that the E. coli NorR homolog (b2709) encoded by ygaA (36) and the adjacent gene ygaK (encoding a flavoRb (b2710)) ( Fig. 1) are each required for expression of the efficient and inducible anaerobic NO metabolic activity (Figs. 2 and 3), the protection of NO-sensitive [4Fe-4S]-containing (de)hydratases (Figs. 4 and 6), and the resistance of E. coli to NO-mediated growth inhibition under various metabolic conditions (Fig. 5). Proximal ygbD (encoding FIG. 5. Anaerobic growth protection by the ygaA, ygaK, and ygbD (norRVW) genes. Strains AB1157 (parental strain), AG200 (ygaA::lac), AG300 (ygaK::lac), and AG400 (ygbD::lac) were grown continuously under N 2 (q) or were exposed to an atmosphere containing 240 ppm NO in N 2 (E) at the times indicated by the arrows. Cultures were grown in phosphate-buffered LB medium containing 20 mM glucose (A); in minimal salts medium containing 2% gluconate and 0.25% casamino acids (B); or in minimal salts medium containing 20 mM glucose and the required amino acids histidine, arginine, proline, threonine, and leucine (each at 40 g/ml) (C). Cultures were initiated with the respective 4% (A), 5% (B), and 10% (C) inocula from overnight static cultures grown in the same medium. In C, cells were centrifuged and washed with fresh minimal salts medium prior to inoculation. Approximate generation times are indicated for each condition. Data are representative of two or more experiments .   FIG. 6. Anaerobic protection of 6-phosphogluconate dehydratase by ygaK (norV). AB1157 (parental strain) (white bars) and AG300 (ygaK::lac) (black bars) were grown as described in the legend to Fig. 5. After 60 min of exposure to N 2 (Control) or 240 ppm NO in N 2 (NO), cells were harvested, and dehydratase activity was assayed as described under "Materials and Methods." Error bars represent the S.E. of four independent trials. The asterisk indicates a significance of p Ͻ 0.05 relative to the corresponding control value.
FIG. 7. Microaerobic growth protection by flavoRb (YgaK) and flavoHb (HMP). AB1157 (parental strain), AG103 (hmp::Tn5), AG300 (ygaK::lac), and AG301 (hmp::Tn5 ygaK::lac) cultures were grown in minimal salts medium plus 2% gluconate and 0.25% casamino acids under a hypoxic atmosphere containing 0.5% O 2 in N 2 . Cultures were maintained under hypoxia (q) or were exposed to 240 ppm NO under hypoxia (E) at the times indicated by the arrows. Cultures were initiated with 5% inocula from overnight 10-ml static cultures grown in the same medium. Generation times were estimated for each condition and are given in minutes. Data are representative of two or more experiments. a NADH:(flavo)Rb oxidoreductase (b2711)) appears to play an ancillary role in anaerobic NO metabolism and detoxification, suggesting the existence of other flavoRb reductases. We propose the designation of the E. coli ygaA-ygaK-ygbD gene cluster (Fig. 1) as the nitric oxide reduction modulon norRVW. Similar modulons are also found in the Salmonella typhimurium and Klebsiella pneumoniae genomes.
We can infer in part from the work of others (36,54) that NorR (YgaA) acts as 1) the receiver of phosphate from a separate unidentified NO-sensing histidine:aspartate phosphotransferase (kinase) and 2) the regulator of ygaK-ygbD (norVW) transcription in a two-component sensor-receiver regulatory system. Future investigations will be aimed at understanding these regulatory systems.
E. coli flavoRb (NorV) belongs to a superfamily of proteins encoded in the genomes of anaerobic archaea and facultative eubacteria, including Methanococcus, Desulfovibrio, Pyrococcus, Dehalococcus, Treponoma, Clostridium, Salmonella, Klebsiella, and the photosynthetic cyanobacterium Synechocystis (39,55). E. coli flavoRb shares 34% amino acid identity with the Desulfovibrio gigas rubredoxin:O 2 oxidoreductase (ROO) (55) and shares key amino acids in the diiron center of ROO (Fig. 8). ROO differs most notably from flavoRb in that the rubredoxin domain exists as a separate protein (55). Non-heme iron and FMN stoichiometries and electromagnetic properties of E. coli flavoRb support the presence of a diiron center, a tightly bound FMN, and mononuclear iron in the rubredoxin domain (40). ROO has been proposed to function in O 2 reduction and detoxification or oxidative stress protection in anaerobes (56,57). However, a role for ROO and flavoRb in anaerobic NO reduction and detoxification now appears more likely because these organisms often lack an identifiable NOR (1), but clearly express O 2 -reducing respiratory chains (58).
Given the homologous structure of ROO (55) and the redox and NO-binding properties of E. coli flavoRb (40), we envision a catalytic NO reduction mechanism in which two NO molecules bind to the diferrous (Fe 2ϩ -O-Fe 2ϩ ) center (Fig. 8), and each is univalently reduced to form two nitroxyl anions and a diferric center. Two nitroxyl molecules would then combine to form N 2 O and water, as suggested for the cytochrome bc-type NOR (1,59). For turnover, NADH-dependent flavoRb reductase (NorW) or other reductases (Fig. 2) would supply two electrons to the diferric center via the rubredoxin domain and the proximal FMN in flavoRb. Alternatively, we can suppose a mechanism in which the diferrous center binds NO and reduces NO by two electrons to produce an Fe 3ϩ -O-Fe 3ϩ -NO 2Ϫ (H ϩ ) intermediate that reacts with the second NO molecule to form N 2 O. An analogous mechanism has been suggested for cytochrome P450nor (60). The cyanide and CO resistance and high NO affinity of the activity (Fig. 3) may be explained by this novel non-heme iron mechanism for NO reduction. Reductive activation of O 2 to peroxo (Fe 3ϩ -OOH) or oxenoid (Fe 4ϩ ϭO) intermediates by non-heme diiron centers such as those in deoxyhemerythrin, methane monooxygenase, and stearoyl-(acyl carrier protein) ⌬ 9 -desaturase is well documented (61,62). The formation of these reactive O 2 intermediates may account for the rapid and irreversible O 2 -mediated inactivation of the flavoRb-type NOR activity in E. coli (17). O 2 sensitivity may also explain the low (ϳ0.2 s Ϫ1 ) and progressively diminishing in vitro O 2 reductase activity reported for flavoRb (40). Future studies will be aimed at understanding the mechanism of NO reduction by flavoRb and the mechanism of O 2 inactivation. It is important to point out that reaction intermediates and mechanisms of the two-heme cytochrome bc-type NOR and cytochrome P450nor remain to be fully elucidated.
Labile [4Fe-4S] (de)hydratases have explained the conditional toxicity of superoxide to various organisms (41,47,48,51), and it is becoming increasingly apparent that these enzymes may also illuminate important mechanisms of NO toxicity and the physiological roles of NODs, NORs, and other NO defenses in various models (7,16,17). Our investigations demonstrate that anaerobic NO toxicity is due, at least in part, to the inactivation of [4Fe-4S]-containing (de)hydratases, including aconitase and 6-phosphogluconate dehydratase, and that inducible anaerobic NO metabolism via flavoRb is important for their protection . NO sensitivity of the branchedchain amino acid pathway ␣,␤-dihydroxyacid dehydratase may contribute to growth defects observed in minimal medium (Fig.  5C), and aconitase inactivation and poisoning of the amphibolic reactions of the citric acid cycle are expected to have complex and pleiotropic effects on cells.
NORs and NODs are likely to be important for the resistance of microbes to the immune system (3)(4)(5) and the NO present in various niches (1,3). It is hoped that a greater understanding of the complex roles of NODs and NORs and their regulators in NO detoxification and microbial pathogenesis may lead to the development of novel therapies. Thus, inhibitors of NODs, NORs, or their regulators may be useful for enhancing the antibiotic action of NO on pathogens or tumor cells (7). Furthermore, greater knowledge of NO metabolism may be particularly important for understanding the pathogenesis and persistence of certain organisms. For example, we have noted an internal 68-amino acid deletion in the flavoRb-coding sequence (norV) encompassing the critical FMN-binding flavodoxin domain in the genomes of enterohemorrhagic E. coli O157:H7 isolates from the Michigan (63) and Sakai (64) outbreaks. The norV mutation located among wild-type norR and norW genes suggests an important, yet puzzling role for NO in the pathogenesis and virulence of O157:H7. It seems unlikely that the deletion increases flavoRb function. Rather, the truncated flavoRb may, in combination with the NO FIG. 8. Diiron center structure and the NO reduction mechanism. Key diiron center ligands shown in the D. gigas ROO structure are conserved in E. coli flavoRb (40,55). Reduction of NO is proposed to occur in the diiron center because NO binds to the diferrous center (Fe 2ϩ -O-Fe 2ϩ ) to form an EPR-active ferric nitroxyl (40). A second NO molecule binding to the diferrous center and a second one-electron reduction are required for N 2 O formation as proposed for the two-heme diiron center of NorB (1). Alternatively, bound NO may be reduced by two electrons, as suggested for NO in cytochrome P450nor (60)  produced by intestinal epithelial cells and other immune cells, increase the synthesis of the bacterial toxins that make O157:H7 so devastating.