Conversion of NO₂ to NO by reduced coenzyme F₄₂₀ protects mycobacteria from nitrosative damage

Abstract

In mycobacteria, F₄₂₀, a deazaflavin derivative, acts as a hydride transfer coenzyme for an F₄₂₀-specific glucose-6-phosphate dehydrogenase (Fgd). Physiologically relevant reactions in the mycobacteria that use Fgd-generated reduced F₄₂₀ (F₄₂₀H₂) are unknown. In this work, F₄₂₀H₂ was found to be oxidized by NO only in the presence of oxygen. Further analysis demonstrated that NO₂, produced from NO and O₂, was the oxidant. UV-visible spectroscopic and NO-sensor-based analyses proved that F₄₂₀H₂ reduced NO₂ to NO. This reaction could serve as a defense system for Mycobacterium tuberculosis, which is more sensitive to NO₂ than NO under aerobic conditions. Activated macrophages produce NO, which in acidified phagosomes is converted to NO₂. Hence, by converting NO₂ back to NO with F₄₂₀H₂, M. tuberculosis could decrease the effectiveness of antibacterial action of macrophages; such defense would correspond to active tuberculosis conditions where the bacterium grows aerobically. This hypothesis was consistent with the observation that a mutant strain of Mycobacterium smegmatis, a nonpathogenic relative of M. tuberculosis, which either did not produce or could not reduce F₄₂₀, was ≈4-fold more sensitive to NO₂ than the wild-type strain. The phenomenon is reminiscent of the anticancer activity of γ-tocopherol, which reduces NO₂ to NO and protects human cells from NO₂-induced carcinogenesis.

Coenzyme F₄₂₀, a 7,8-didemethyl-8-hydroxy-5-deazaflavin derivative (Fig. 1), is a 2-electron or hydride transfer restricted redox catalyst (E^o′ = −360 mV) similar to the nicotinamides (E^o′ = −320 mV) (1, 2). F₄₂₀ is found in all methanogenic and certain nonmethanogenic archaea, where it participates in energy metabolism, NADP reduction, oxygen detoxification, and sulfite reduction (3–6). In the bacterial domain, F₄₂₀ is found in certain members of the Actinobacteria phylum, such as Mycobacterium species (7). These organisms express an F₄₂₀-dependent glucose-6-phosphate dehydrogenase (Fgd, Reaction 1) (7, 8).

The physiological fate of F₄₂₀H₂ produced by Fgd is unknown. An insertional inactivation of fbiC, an essential gene for the synthesis of the deazaflavin chromophore or catalytic unit of F₄₂₀ (9), renders Mycobacterium tuberculosis hypersusceptible to acidified nitrite (10). This in vitro treatment simulates an environment inside the phagosomes of an infected-activated macrophage, which produces nitric oxide (NO) by the action of inducible nitric oxide synthase (iNOS or NOS2) (11). Upon acidification of a phagosome, nitrite, a major product of NO oxidation, is converted to nitrous acid (HNO₂; pK_a = 3.16) (12), which in turn dismutates to NO and NO₂ (10, 13); NO₂ arises also from a reaction of NO with O₂ (14). These observations suggest that the pathogenic mycobacteria could use F₄₂₀H₂ to combat an attack of reactive nitrogen intermediates generated by the activated macrophage. We have tested this hypothesis. The resulting data point toward a mechanism that M. tuberculosis may use to counter the bactericidal actions of macrophages.

Fig. 1.

UV-visible spectra of oxidized and reduced forms of coenzyme F₄₂₀. F₄₂₀ was reduced with sodium borohydride.

Results and Discussion

In a phosphate-buffered aqueous solution, F₄₂₀H₂ was rapidly oxidized by the NO-releasing compound diethylamine-NONOate (DEA-NONOate) and oxygen (Fig. 2).

Neither DEA-NONOate nor O₂ alone was able to perform this reaction (Fig. 2). To ascertain that NO was responsible for oxidizing F₄₂₀H₂, we examined whether the oxidation rate of F₄₂₀H₂ can be controlled by the NO release rate. For this purpose, two other derivatives of diazeniumdiolates, spermine-NONOate and DETA-NONOate, which release NO slowly, were used. The half-lives of DEA-, spermine-, and DETA-NONOate to release NO are 2 min, 39 min, and 20 h, respectively (15). Indeed, with the last two compounds, the F₄₂₀H₂ oxidation was much slower than that with DEA-NONOate. With 80 μM dissolved O₂ and 25 μM F₄₂₀H₂, DEA-NONOate (0.5 mM), spermine-NONOate (0.5 mM), and DETA-NONOate (2.5 mM) oxidized F₄₂₀H₂ at the rates of 17.3, 0.94, and 0.22 μmol min⁻¹ L⁻¹, respectively. These data established that NO was one of the entities involved in the oxidation of F₄₂₀H₂.

Fig. 2.

Oxidation of F₄₂₀H₂ by nitric oxide in the presence of oxygen. Nitric oxide was generated in situ from NONOate derivatives. Into a 1-mL anaerobic solution of 50 mM potassium phosphate buffer (pH 7.0) in a stoppered cuvette with N₂ in the headspace, the following were added to the final concentrations and in the sequence indicated: (A) Reduced F₄₂₀ (20 μM), oxygen (80 μM), and DEA-NONOate (320 μM). (B) Reduced F₄₂₀ or F₄₂₀H₂ (20 μM), DEA-NONOate (320 μM), and oxygen (80 μM). The time of each addition is shown by an arrow. Oxidation of F₄₂₀H₂ was monitored spectrophotometrically at 420 nm. DEA-NONOate, 2-(N,N-diethylamino)-diazenolate-2-oxide sodium salt; DETA-NONOate, 2,2′-(hydroxynitrosohydrazino)bisethanamine; spermine-NONOate, N-[4-[1-(3-aminopropyl)-2-hydroxy-2-nitrosohydrazino]butyl]-1,3-propanediamine.

To understand the role of O₂ in the NO-dependent F₄₂₀H₂ oxidation, the effect of O₂ concentration on the reaction rate and amount of F₄₂₀ produced at fixed concentrations of DEA-NONOate and F₄₂₀H₂ was studied. The results presented in Fig. 3 show that when DEA-NONOate was present in excess, the rate of F₄₂₀H₂ oxidation and the amount of F₄₂₀ formed increased with the dissolved O₂ concentration.

Fig. 3.

Reaction rate and amount of F₄₂₀ produced at various dissolved oxygen concentrations and at fixed levels of DEA-NONOate and F₄₂₀H₂. The assay mixture contained 1 mL of anaerobic 100 mM potassium phosphate buffer (pH 7.0), with 20 μM F₄₂₀H₂ and varied volumes of O₂ added to the nitrogen headspace. Before the assay, the mixture was held at room temperature for 15 min to ensure equilibration with the added oxygen. The reaction was initiated by the addition of DEA-NONOate (final concentration, 320 μM) and was followed spectrophotometrically at 420 nm. The dissolved oxygen concentration was calculated by using a value of 43,800 atm/mol fraction for the Henry's constant for oxygen (40).

The above results led to the question of how a combination of NO and O₂ oxidized F₄₂₀H₂. In an aqueous solution, nitric oxide reacts rapidly and spontaneously with oxygen to produce mainly nitrite, via the intermediate formation of nitrogen dioxide (NO₂) and dinitrogen trioxide (N₂O₃) (Reactions 3–5) (14).

To test whether nitrite was the F₄₂₀H₂-oxidizing agent, in an experiment similar to that represented in Fig. 2, DEA-NONOate was incubated in an aerobic potassium phosphate buffer for 60 min at 25 °C (allowing dissipation of NO, NO₂, and N₂O₃ via Reactions 3–5), and then this solution was mixed with F₄₂₀H₂. This treatment did not result in F₄₂₀H₂ oxidation. Also, as determined from the UV-visible spectral characteristics, 1 mM sodium nitrite or 1 mM sodium nitrate did not oxidize F₄₂₀H₂. However, the subsequent addition of DEA-NONOate to these aerobic mixtures caused the oxidation of F₄₂₀H₂. Therefore, nitrite or nitrate was neither an F₄₂₀H₂-oxidizing agent nor an inhibitor of F₄₂₀H₂ oxidation. It is unlikely that N₂O₃ was responsible for F₄₂₀H₂ oxidation because this compound is not an oxidizing agent (16). Thus, it was hypothesized that NO₂, a fairly potent oxidant (16), was the compound that oxidized F₄₂₀H₂. This hypothesis was tested as described below.

In the first assay, NO₂ was generated in situ via a reaction between C-PTIO, a nitronyl nitroxide, and NO produced from DEA-NONOate in an anaerobic solution (17). The data in Fig. 4 show that the oxidation of F₄₂₀H₂ was observed only if both DEA-NONOate and C-PTIO were present in the reaction mixture, suggesting that F₄₂₀H₂ was oxidized by NO₂. Because of an overlap between the UV-visible spectra of C-PTIO and F₄₂₀H₂ (17), this system was unsuitable for spectroscopic studies. Therefore, further experiments were performed using pure NO₂ gas. The resulting data demonstrated that NO₂ did indeed oxidize F₄₂₀H₂ directly (Fig. 5). The UV-visible spectra of the reaction mixture (Fig. 5A) that was recorded immediately after each addition of NO₂ into an anaerobic solution of F₄₂₀H₂ in 100 mM potassium phosphate buffer (pH 7.0) showed the appearance of the oxidized form of F₄₂₀. The amount of F₄₂₀ produced was proportional to the amount of NO₂ added into the reaction mixture (Fig. 5B); because NO₂ would rapidly be converted to nitrite and nitrate in the phosphate-buffered solution we used (Reactions 6 and 7) (14), the ratio of NO₂ added and F₄₂₀ formed was not 1:1.

The UV-visible spectral changes shown in Fig. 5A clearly indicated that F₄₂₀H₂ was oxidized to F₄₂₀. This oxidation is known to be a 2-electron or hydride transfer process (2, 18). The observation leads to the hypothesis that F₄₂₀H₂ oxidation was coupled to 2-electron reduction of NO_{2 that} generated NO (Reaction 8).

This possibility was tested via measurements with an ISO-NOPF100H NO probe from World Precision Instruments. The addition of NO₂ in the nitrogen headspace of a sealed serum bottle containing anaerobic 100 mM potassium phosphate buffer (pH 7.0) created a high electrical current (Fig. 6) because the NO sensor detects NO₂, albeit with much less sensitivity than that for NO (Michael McIntosh and Xueji Zhang, World Precision Instruments, personal communication). However, the current dissipated quickly because of the conversion of NO₂ to nitrite and nitrate (Reactions 6 and 7). The overall result was a current spike (Fig. 6). When F₄₂₀H₂ was present in the reaction mixture, a similar current spike was also seen, but the residual current was higher than that recorded with NO₂ alone (Fig. 6). This residual current represented some of the NO generated from NO₂ by F₄₂₀H₂. NO is relatively stable in an anaerobic aqueous solution, but it can be dissipated via Reactions 4 and 5, if NO₂ is present. Because added NO₂ was lost quickly via Reactions 6 and 7, only a small amount of NO was generated. However, because of the lack of residual NO₂, this NO was not sequestered and provided a higher residual current. The amount of F₄₂₀ formed in this experiment was proportional to the amount of NO₂ added. Therefore, it was clear that F₄₂₀H₂ reduced NO₂ to NO. Based on the well-established ground-state chemistry of F₄₂₀ (2, 18), we hypothesize that this reduction occurs via a hydride transfer mechanism; this possibility is currently under investigation.

Fig. 4.

Oxidation of F₄₂₀H₂ under anaerobic condition by nitrogen dioxide generated from an in situ reaction between carboxy-PTIO and NO (from DEA-NONOate). Into 1 mL of anaerobic 50 mM potassium phosphate buffer (pH 7.0) in a stoppered cuvette with N₂ in the headspace, the following were added: 25 μM F₄₂₀H₂, 100 μM C-PTIO, and 250 μM DEA-NONOate. The point of each addition is shown by an arrow and the name of the compound. Oxidation of F₄₂₀H₂ was monitored spectrophotometrically at 420 nm. Carboxy-PTIO or C-PTIO, 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt. Other abbreviations are shown in the legend for Fig. 2.

Fig. 5.

Oxidation of F₄₂₀H₂ by nitrogen dioxide gas under anaerobic condition. (A) To 1 mL of an anaerobic 25 μM F₄₂₀H₂ solution in 50 mM potassium phosphate buffer (pH 7.0) in a 3-mL quartz cuvette, nitrogen dioxide gas was added by using a gas-tight syringe. The UV-visible spectrum of the solution was recorded before NO₂ addition and then after each addition of 250 μL of (10.22 nmol) NO₂. (B) Relationship between NO₂ added and F₄₂₀ generated. Note: Most of the NO₂ added was likely converted into NO₂⁻ and NO₃⁻.

Fig. 6.

Nitric oxide production from the reaction of nitrogen dioxide and F₄₂₀H₂. One milliliter of NO₂ (1 atm) was added to the nitrogen headspace of a stoppered, 10-mL serum bottle containing 7.5 mL of anaerobic 100 mM potassium phosphate buffer (pH 7.0) with 20 μM F₄₂₀H₂ (blue) or without F₄₂₀H₂ (pink). NO production was measured by using an NO sensor (ISO-NOPF100H) (World Precision Instruments).

Because F₄₂₀ catalyzes an NAD(P)H-type hydride transfer, the reactivity of the latter toward NO₂ was examined. NADH reacts with NO₂ in the presence of oxygen (19). In this process, NO₂ is reduced to nitrite, and an NAD· radical is produced as an intermediate, and the latter reacts with oxygen to generate superoxide (

). We found that in an anaerobic assay system of the type shown in Fig. 5, NO₂ does not oxidize NADH or NADPH; in each case the concentration of the reduced cofactor was 160 μM. Therefore, in its ability to reduce NO₂ to NO, F₄₂₀H₂ was distinct from NAD(P)H.

To examine whether nitrogen dioxide caused nitrosation of the phenol moiety of F₄₂₀, the products from a reaction similar to that shown in Fig. 2A, but with F₄₂₀H₂ and DEA-NONOate at 50 and 500 μM concentrations, respectively, were characterized. In a C₁₈-silica column-based reversed-phase HPLC analysis employing a diode array detector (20), the deazaflavin product from the reaction and authentic F₄₂₀ exhibited the same retention time and UV-visible spectrum [supporting information (SI) Fig. S1]. A nitrophenol generated from a reaction between a phenol and NO₂ binds more tightly to C₁₈-silica than the corresponding phenol (21, 22). These results ruled out the possibility of nitrosation of F₄₂₀H₂ with NO₂ or NO.

The ability to use F₄₂₀H₂ in reducing NO₂ to NO would be useful to M. tuberculosis in combating the host defense system. Although both NO and NO₂ exhibit antimycobacterial activities, NO₂ is more potent in killing M. tuberculosis (23). Therefore, because an activated macrophage produces NO and converts this compound in its acidified phagosomes to NO₂ either via the formation and dismutation of nitrous acid (13) or a reaction of NO with O₂ (14) for bringing about a more aggressive attack, M. tuberculosis could use Fgd-generated F₄₂₀H₂ to reduce NO₂ back to NO and lower the effectiveness of the antibacterial action of macrophages. This hypothesis is consistent with the observed hypersensitivity of M. tuberculosis fbiC strain to acidified nitrite (10). A mycobacterial cell contains a substrate level of glucose 6-phosphate (24). Therefore, F₄₂₀H₂ produced via the Fgd reaction (Reaction 1) is an effective vehicle for delivering substantial reducing power for NO₂ detoxification. To provide an experimental validation for this deduction, the effect of NO₂ on the viability of Mycobacterium smegmatis, a close relative of M. tuberculosis and a nonpathogen, was studied. In this experiment, actively growing cultures of wild-type, fgd::aph, and Δfbic::aph strains of M. smegmatis were exposed to NO₂, and then the numbers of the living cells in these cultures were determined (Fig. 7). NO₂ was generated in situ through a reaction between dissolved O₂ and NO (derived from spermine-NONOate; half-life, 39 min); an exposure time of 6 h was selected because it provided an observable yet mild effect on the viability of the wild-type cells. The survivability of a strain was expressed in terms of the ratio of the colony-forming units from an NO₂-exposed culture and that from a control culture (without NO₂ exposure). The values for these ratios for the wild-type, fgd::aph, and Δfbic::aph strains were 0.8, 0.25, and 0.2, respectively (Fig. 7). Therefore, an inability to synthesize F₄₂₀ or to produce F₄₂₀H₂ made the organism ≈4-fold more sensitive to NO₂. When the fgd::aph and Δfbic::aph were complemented with Fgd and Fbic expression vectors, respectively, they were able to tolerate NO₂ even better than the wild-type cells (Fig. 7); the increased tolerance was likely caused by above-normal level of F₄₂₀H₂ generated by the elevated levels of FbiC and Fgd proteins. These results established a requirement for F₄₂₀H₂ in the reduction of NO₂. However, it remains possible that the observed phenotypes of the mutant strains were not solely the result of a loss of the chemical reaction described in this report but also in part caused by the lack of the activity of a yet to be identified NO₂ detoxification enzyme that requires F₄₂₀H₂. The overall phenomenon is similar to the inhibition of neoplastic transformation of human cells by γ-tocopherol (25). It has been suggested that the nitrosation of primary amines of DNA bases by NO, which leads to cancer-causing mutations, requires oxidation of this oxide to NO₂ (25). Because γ-tocopherol efficiently reduces NO₂ back to NO, it protects cells from NO₂-induced carcinogenesis (25).

Fig. 7.

Requirement for reduced F₄₂₀ for NO₂ detoxification in M. smegmatis. The sensitivities of wild-type, fgd::aph, Δfbic::aph, fgd::aph (pEP-fgd), and Δfbic::aph (pEP-fbiC) strains to NO₂ were studied. In each case, the cells were exposed to NO₂ by incubating a corresponding diluted suspension in growth medium with spermine-NONOate under air; a control culture received water in place of spermine-NONOate solution. In each case, including for the control, the number of surviving cells per milliliter in a culture is presented as a percentage of the value recorded for a sample drawn before incubation with spermine-NONOate (or water); as a result, the value even for the control culture was <100%. Each error bar was calculated from measurements in triplicate. Gray bar, control culture; black bar, NO₂-exposed culture.

It has been known that PA-824, a bicyclic nitroimidazole and a promising candidate drug for the treatment of tuberculosis, acts as an antimycobacterial agent only when the organism produces F₄₂₀ and reduces it to F₄₂₀H₂ (9, 20, 26). A recent report shows that the basis for this activity is the F₄₂₀H₂-dependent production of NO from PA-824 (27); NO kills M. tuberculosis anaerobically. The imidazole ring of PA-824 is reduced with F₄₂₀H₂ by a nitroreductase called Ddn (27). This reaction is followed by the release of the nitro group of the compound as nitrous acid that breaks down to NO (27). Therefore, F₄₂₀-dependent reactions not only provide a protection to the mycobacteria from nitrosative stresses, but also can be exploited to impose such a stress on these organisms.

In summary, F₄₂₀H₂-dependent NO₂ reduction reaction is a defense tool for the mycobacteria against NO₂ stress. It should be noted that often NO has been cited as the agent that kills M. tuberculosis under both aerobic and low-oxygen or hypoxic conditions (28, 29), although, as mentioned above, NO₂ has been shown to be a more potent antimycobacterial agent (23), and it is produced under aerobic conditions (10, 13, 14). Therefore, the F₄₂₀H₂-dependent defense against nitrosative stress would be useful to M. tuberculosis when it grows aerobically and causes active tuberculosis.

Materials and Methods

Chemicals.

Coenzyme F₄₂₀ was purified from Methanothermobacter thermautotrophicus as described (30, 31). 2-(N,N-diethylamino)-diazenolate-2-oxide sodium salt (DEA-NONOate), N-[4-[1-(3-aminopropyl)-2-hydroxy-2-nitrosohydrazino]butyl]-1,3-propanediamine (spermine-NONOate), 2,2′-(hydroxynitrosohydrazino)bisethanamine (DETA-NONOate), 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt (carboxy-PTIO or C-PTIO) were from Cayman Chemicals. Sodium borohydride was from Sigma–Aldrich. F₄₂₀H₂ was generated by reducing F₄₂₀ with sodium borohydride in an anaerobic water solution under nitrogen atmosphere in a rubber stopper-sealed glass tube followed by neutralization with hydrochloric acid. Only freshly made F₄₂₀H₂ was used in the experiments. Nitrogen dioxide gas was prepared in microscale (32). The anaerobic techniques used in this work have been described in ref. 30.

F₄₂₀H₂ Oxidation Assays.

The assays were carried out as described in ref. 30. A 1-mL anaerobic aqueous solution containing 50 mM potassium phosphate buffer (pH 7.0) and 20 or 25 μM F₄₂₀H₂ under a N₂ headspace in a 12-mm × 75-mm (8.5 mL) KIMAX borosilicate tube (Kimble/Kontes) sealed with a cutoff black butyl rubber stopper no. 00 was used for each assay. To this solution, DEA-NONOates, spermine-NONOate, and DETA-NONOate were added from respective 50 mM anaerobic stock solutions in 10 mM sodium hydroxide. For C-PTIO, a 50 mM stock was prepared in water. NaNO₂ or NaNO₃ was added from an anaerobic 100 mM stock solution in water. NO₂ was added by using a gas-tight syringe. Oxidation of F₄₂₀H₂ was followed spectrophotometrically at 420 nm (at 25 °C), and the rates were calculated by using a value of 41.4 mM⁻¹ cm⁻¹ for the extinction coefficient of oxidized F₄₂₀ at this wavelength (31).

NO Measurement Assay.

NO produced in an anaerobic F₄₂₀H₂ oxidation–NO₂ reduction assay was measured with an Apollo 1000 single-channel, free radical detector employing an amperometric-type NO probe (ISO-NOPF100H; World Precision Instruments). The data were collected and analyzed by using a WPI DataTrax. The measurement was performed under anaerobic conditions. For this purpose, a 10-mL serum bottle containing 7.5 mL of 100 mM potassium phosphate buffer (pH 7.0), with 25 μM F₄₂₀H₂ was sealed with a 20-mm septum stopper (Bellco Glass), and the content was made anaerobic via 3 cycles of evacuation and pressurization with N₂ (30); a control vial was prepared without F₄₂₀H₂. The final headspace gas was N₂ at 1.3 atm. Then a 1-inch 18-gauge hypodermic needle with a Luer hub (Precision Glide needle; Becton Dickinson), which was shortened to 0.5 inches by cutting part of its tip end, was inserted into the bottle through the rubber stopper. Through this needle the probe was lowered in a manner where 2 mm of the probe tip was inside the buffer; before use, the probe was polarized according to the manufacturer's instructions. After the current output stabilized, NO₂ gas was added to the headspace of the bottle through the rubber stopper by using a syringe, and the probe output was measured. To avoid contamination of air, the syringe was flushed with N₂ before use.

Determining the Sensitivities of M. smegmatis to NO₂.

A strain to be tested was grown in 7H9-glycerol medium with 0.05% Tween 80 at 37 °C until the early-logarithmic stage of growth. Then a sample of this culture was diluted with fresh medium to a 600-nm optical density of 0.5 [as measured with a Lambda 25 UV-visible spectrophotometer (PerkinElmer)]. One hundred microliters of this diluted culture was added to 900 μL of fresh medium, and this mixture was supplemented with 10 μL of 100 mM spermine-NONOate right after the latter was prepared in water; a control culture received water in place of the spermine-NONOate solution. The tubes with dilute cell suspensions (±spermine-NONOate) were incubated aerobically in 13-mm × 100-mm borosilicate culture tubes (catalog no. 73750-13100; Kimble/Kontes) for 6 h at 37 °C in a C25 incubator shaker (New Brunswick Scientific) operating at 240 rpm. Then the number of surviving cells in each such suspension was estimated via plating on 7H9-glycerol-agar medium and determining the number of colony-forming units.

Construction of M. smegmatis ΔfbiC::aph and fgd::aph Strains.

A ΔfbiC strain was generated by replacing the corresponding coding sequence with a kanamycin resistance cassette (aph) via homologous recombination (33). For this purpose, a 926-bp upstream region (UP) and 920-bp downstream region (DN) of M. smegmatis fbiC coding sequence (Msmeg_5126) were PCR-amplified with the following primer pairs: for the UP element, ccgaggatcctcgagaaggcaagggtctcggtcggcgtggtgag (forward) and gctcgaattccaaagccactcccgaataactccacgag (reverse) (underlined sequences, BamHI and EcoRI sites, respectively); for the DN element, cggccgaattcaccacatccagaccagctggg (forward) and gcccggctcgaggccgtcgaacacgagattca (reverse) (underlined sequences, EcoRI and XhoI sites, respectively). The UP and DN amplicons were digested with the above listed restriction enzymes. These restricted DNA fragments and a res-Km-res element that was excised from pCG122 (34) as an EcoRI fragment were cloned into BamHI and XhoI sites of pBluescript to obtain pEPfbiCKO1. Then the fbiC knockout construct was excised from pEPfbiCKO1 with XhoI and was cloned into the XhoI site of pJQ200KS (35) to obtain the suicide plasmid pEPfbiCKO2; pJQ200KS does not replicate in mycobacteria, but in an integrated form it allows a sacB-mediated sucrose counterselection and confers resistance to gentamicin in mycobacteria. In pEPfbiCKO2, the cloned fragments formed the following arrangement: 5′UP3′…5′res-Km-res3′…5′DN3′. Transformation of M. smegmatis mc²155 (wild-type strain) with pEPfbiCKO2 via electroporation (36) and selection of the transformants on gentamicin and kanamycin provided a merodiploid strain. By counterselecting segregants derived from the merodiploid strain on sucrose in the presence of kanamycin, a M. smegmatis ΔfbiC::res-Km-res strain was obtained; we call this strain ΔfbiC::aph. A Southern blot with EcoRI-digested chromosomal DNA and PshAI-digested pEPfbiCKO2 as the probe confirmed this genotype and showed that the mutant did not carry a remnant of pJQ200KS. An fgd::aph strain was constructed by inserting a kanamycin resistance cassette (Amersham Pharmacia Biotech) as a BamHI fragment at the BglII site of the gene; the details of this construction will be presented elsewhere. A Southern blot analysis confirmed the fgd::aph genotype of this strain. Assays for F₄₂₀ and Fgd in cell lysates as described (24, 37) showed that the fgd::aph and ΔfbiC::aph strains lacked Fgd activity and F₄₂₀, respectively.

Construction of Complementing of Plasmids for M. smegmatis fbiC and fgd.

The constructs were based on pSMT3, a mycobacteria–Escherichia coli shuttle vector that allows selection for hygromycin resistance and expression of cloned genes constitutively from the hsp60 promoter in mycobacteria (38). To generate the fbiC complementing plasmid (pEP-fbiC), the corresponding coding sequence along with a 222-bp upstream region (bearing the fbiC promoter and ribosome-binding site) was PCR-amplified from M. smegmatis chromosomal DNA with primers GATATCCGCGGACAGAGACAG (forward) and AAGCTTGACCTCGCGGTGAAG (reverse) (underlined sequence: EcoRV and HindIII sites, respectively) and cloned into EcoRV and HindIII sites of pSMT3; the selection of the upstream element was based on an earlier report (39). Similarly, the Fgd expression plasmid (pEP-fgd) was constructed by amplifying the corresponding coding sequence along with the 264-bp upstream region with the primers GGTTTTCGGGATCCAGATGGCAGTGCCGAAG (forward, with a BamHI site) and GAGCACTGCAGCTCGCGCGAGCACCCGCACG (reverse, with a PstI site) and cloning the amplicon into BamHI- and PstI-digested pSMT3.