KEYWORDS: coenzyme F₄₂₀, F₄₂₀-dependent glucose-6-phosphate dehydrogenase, F₄₂₀H₂, folate, TB-drug resistant, gamma-glutamyl ligase, mycobacteria, polyglutamylation

F₄₂₀H₂ is the preferred substrate for the extension of glutamate chains of F₄₂₀ by FbiB.

We performed FbiB reaction with F₄₂₀-0, F₄₂₀-0H₂, F₄₂₀-2, or F₄₂₀-2H₂ as the substrates; the reduced forms were generated via reduction with sodium borohydride (10). Since F₄₂₀-2H₂ undergoes slow autoxidation to F₄₂₀-2 under air, we performed all reactions under strict anaerobic conditions. The reaction mixture was incubated for 48 h at 37°C. We observed that this incubation converted F₄₂₀-0 to F₄₂₀-2, and it did not transform F₄₂₀-2. In contrast, the reaction with F₄₂₀-0H₂ generated significant amounts of F₄₂₀-4 and F₄₂₀-5. These qualitative observations suggested that polyglutamylation of F₄₂₀ likely requires that the deazaflavin substrate is presented to the enzyme in a reduced form. However, this assay system proved unsuitable for more detailed analyses for the following reason. Here, the stoppered assay tubes containing anaerobic assay mixtures were incubated under air. Consequently, upon long incubation, a small amount of oxygen entered into the tubes, slowly oxidized F₄₂₀-0H₂ or F₄₂₀-2H₂, and the assay mixtures progressively turned yellow in color. It is known that deazaflavins are oxidized by oxygen, albeit very slowly (50). To avoid even a slight air-induced oxidation of F₄₂₀-0H₂ or F₄₂₀-2H₂, we opted for a system that allowed continuous regeneration of reduced deazaflavin.

Coenzyme F₄₂₀-dependent glucose-6-phosphate dehydrogenase (Fgd) is the major F₄₂₀ reduction system in mycobacteria (9,10,20,49).

Glucose-6-phosphate (G6P) + F₄₂₀ → 6-phospho-gluconate + F₄₂₀H₂.

Accordingly, we modified the FbiB reaction mixture by including Fgd and G6P to generate the reduced form of the deazaflavin (F₄₂₀-0H₂ or F₄₂₀-2H₂)in situ and maintain it in this form throughout the reaction period; in this design, the assay could be performed under air. At a desired time, the reaction was terminated by boiling the mixture. This treatment inactivated the enzymes and oxidized the substrate (F₄₂₀-0H₂ or F₄₂₀-2H₂) and products (F₄₂₀-nH₂); since Fgd was not active anymore, F₄₂₀ could not be maintained in a reduced state, and the products were detected as F₄₂₀-n. In this assay system with F₄₂₀-0H₂ as the substrate, a 20-h incubation at 37°C converted essentially all of the deazaflavin molecules into a form with seven glutamate residues (Fig. 3A); the product was detected as F₄₂₀-7 (Fig. 3A). In the respective control, which lacked Fgd and G6P, 91.54% of the product was F₄₂₀-2 and 6.54% was F₄₂₀-3 (Fig. 3B). With F₄₂₀-2 as the substrate and in the presence of Fgd and G6P (generating F₄₂₀-2H₂), F₄₂₀-7 was almost the sole (99.8%) product (Fig. 3C), and in the absence of the reduction and regeneration system only a minor portion (10.83%) of the substrate was transformed, and that too caused the addition of only one glutamate residue that generated F₄₂₀-3 (Fig. 3D).

The results presented above were further validated via mass spectrometry. Selected high-performance liquid chromatography (HPLC) fractions that were expected to contain F₄₂₀ species with 2, 5, 6, and 7 glutamate residues were analyzed via matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF) mass spectrometry. The resulting spectra exhibited the expected characteristics (observedm/z value [calculatedm/z value], F₄₂₀ species), as follows: 774.3 (773.598), F₄₂₀-2 (Fig. 4A); 1,161.7 (1,160.9), F₄₂₀-5 (Fig. 4B); 1,290.8 or 1,290.4 (1290.06), F₄₂₀-6 (Fig. 4B andC); and 1,419.4 (1,419.2), F₄₂₀-7 (Fig. 4C).

Consecutive glutamate chain extension by FbiB.

To elucidate the steps of the F₄₂₀ polyglutamylation process, the progress of the FbiB reactions with F₄₂₀-0H₂ and F₄₂₀-2H₂ as the substrates were followed. The data inFig. 5 show that FbiB catalyzed the addition of glutamate residues to F₄₂₀ molecules in a consecutive manner. With time, F₄₂₀ molecules with longer glutamyl chains appeared with concomitant disappearance of those with shorter glutamyl chains. Within 5 min from the initiation of the reaction, the enzyme transformed approximately 50% of the F₄₂₀-0H₂, generating the following products (in the indicated proportions) (Fig. 5A): F₄₂₀-1 (7.5%), F₄₂₀-2 (40%), and F₄₂₀-3 (1%). Within 15 min, F₄₂₀-0H₂ almost completely disappeared (remaining amount, 0.2% of the total F₄₂₀) and approximately equal amounts of F₄₂₀-2 and F₄₂₀-4 (50 and 45% of the total F₄₂₀, respectively), together with a very small amount of F₄₂₀-5 (4% of the F₄₂₀ population) were synthesized. After 30 min of incubation, F₄₂₀-5 and F₄₂₀-6 were the major products, while after 60 min these shifted to F₄₂₀-6 and F₄₂₀-7. F₄₂₀-7 was the sole reaction product after 120 min of incubation.

The enzyme (FbiB) also progressively extended the glutamyl side chain of F₄₂₀-2H₂ (Fig. 5B). Within 15 min of incubation, all of the substrate molecules were converted into forms with longer glutamyl chains, and after 2 h, only F₄₂₀-6 (60%) and F₄₂₀-7 (40%) were present in the reaction mixture.

Interestingly, when F₄₂₀-0H₂ was used as the substrate, a small amount of F₄₂₀-1 (7.5% of the total F₄₂₀) was detected after a short incubation time (5 min), and at this time point, 39.8% of the total deazaflavin pool was at the F₄₂₀-2 stage. The case for F₄₂₀-3 was similar, as it was detected in a very small amount in the reaction mixture only after 5 min of incubation (Fig. 5A) but not thereafter.Methanothermobacter marburgensis, which produces F₄₂₀ primarily in the F₄₂₀-2 form (>80% of the F₄₂₀ pool), contains traces of F₄₂₀ species with longer glutamyl side chains and F₄₂₀-1 (1).Methanosarcina barkeri, which produces F₄₂₀ with four and five glutamates as the major products, does not produce F₄₂₀-1 (1). WithM. tuberculosis FbiB, when F₄₂₀-2H₂ was the substrate, F₄₂₀-2 was detectable even after 30 min of reaction time, accounting for 0.13% of the total F₄₂₀ (Fig. 5B). These results suggest that when the oxidized form of the deazaflavin serves as the substrate, F₄₂₀-γ-glutamyl ligase produces F₄₂₀-1 only transiently and then rapidly converts it to F₄₂₀-2, and the latter is the first stable product of the reaction.

Requirement of F₄₂₀-dependent glucose-6-phosphate dehydrogenase (Fgd) for polyglutamylation of F₄₂₀in vivo.

In mycobacteria, Fgd is the predominant, if not exclusive, catalyst for the formation of F₄₂₀H₂ (9,10,20,49). For this reason, we hypothesized that Fgd is required for the polyglutamylation of F₄₂₀in vivo. To test this hypothesis, we determined the side chain characteristics of F₄₂₀ molecules extracted from both wild-type and Δfgd strains ofM. smegmatis mc² 155 (10,51) using an established protocol (1). The data presented inFig. 6A show that inM. smegmatis Δfgd, 50.5% of the deazaflavin pool was in the F₄₂₀-2 form and 31.33% was F₄₂₀-4, while in the wild-type strain, 67.7% of the pool was F₄₂₀-6. Moreover, the total F₄₂₀ content of the Δfgd strain was 36% of that seen with the wild-type strain; the values of the F₄₂₀ contents were normalized with respect to the protein content of the cell.

We also carried out similar experiments withM. bovis BCG and found that in the wild-type strain, 66.3% of the F₄₂₀ pool was in the F₄₂₀-5 and F₄₂₀-6 forms, and F₄₂₀-4, F₄₂₀-2, F₄₂₀-1, and F₄₂₀-0 represented 17.7, 8.2, 5.5, and 2.3% of the pool, respectively (Fig. 6B). In theM. bovis BCGfgd::Tn5367 strain which lacked a functionalfgd gene, F₄₂₀-4 and F₄₂₀-2 constituted 31.5 and 50% of the F₄₂₀ pool, respectively, F₄₂₀-0 was present in a substantial amount (18.4% of F₄₂₀ pool), and F₄₂₀-5 and F₄₂₀-6 were absent (Fig. 6B). These results clearly established the requirement of a reduced form of the coenzyme for the production of F₄₂₀ forms with more than two glutamate residues, especially F₄₂₀-5, F₄₂₀-6, and F₄₂₀-7, as physiologically relevant and also showed thatin vivo this need was fulfilled by F₄₂₀-dependent glucose-6-phosphate dehydrogenase (Fgd). The presence of a significant amount of F₄₂₀-4 inM. smegmatis andM. bovis BCG strains lacking Fgd activity could be rationalized by the fact that there are cell biosynthesis reactions, such that of hydroxymycolic acid dehydrogenase ofM. bovis BCG (11), that produce F₄₂₀H₂. However, such biosynthetic reactions are expected to be of much lower flux compared to that catalyzed by Fgd, which has catabolic-type activity, and supported with a high intracellular level of the respective electron-donating substrate, glucose-6-phosphate (7 to 30 μmol/gram protein) (8). Thesein vivo conditions explain whyM. smegmatis Δfgd andM. bovis BCGfgd::Tn5367 fail to produce F₄₂₀-5, F₄₂₀-6, and F₄₂₀-7, whereas the respective wild-type strains can (Fig. 6).

Our analysis of previously reported data validated the above-described model inM. tuberculosis. It also clearly linked the length of the side chain of F₄₂₀ to the sensitivity ofM. tuberculosis to delamanid. In an earlier study, severalM. tuberculosis strains that lacked a functionalfgd gene were found to be less sensitive to delamanid than were the wild-type, and the HPLC chromatograms of F₄₂₀ preparations obtained from the mutants were different from that of the wild-type strain (22). However, the observed HPLC peaks (Fig. 1 of reference22) were not characterized in terms of the structures of the respective F₄₂₀ species. Using our mass spectrometry-validated HPLC profiles (Fig. 3 and4) as a reference, we analyzed these previously published data and found that theM. tuberculosis strains with inactivatedfgd gene contained F₄₂₀-2, F₄₂₀-3, and F₄₂₀-4, whereas the F₄₂₀ species with longer polyglutamyl side chains (F₄₂₀-5, F₄₂₀-6, and/or F₄₂₀-7) were present in the wild-type strain (22). These results are similar to our observation withM. smegmatis andM. bovis BCG strains lacking a functionalfgd gene (Fig. 6). Taking all these results together, we concluded thatM. tuberculosis requires Fgd activity for the synthesis of longer polyglutamyl side chains of F₄₂₀, and by having F₄₂₀-5 through F₄₂₀-7, the pathogen is more sensitive to the bactericidal action of delamanid.

In this context, it is noteworthy that there is additional evidence supporting a link between the nature of the F₄₂₀ side chain and sensitivity to pretomanid and delamanid inM. tuberculosis. For example, anM. tuberculosis strain that lacks a functionalfbiB gene and consequently produces F₄₂₀-0, an F₄₂₀ species fully lacking the glutamyl side chain, is resistant to delamanid (22). Another study has found that a set ofM. tuberculosis strains that produce either F₄₂₀-0 or F₄₂₀-2 through F₄₂₀-4 were delamanid and pretomanid resistant (21); the statuses of thefbiB andfgd genes in these mutants are unknown.

The mechanistic basis for the requirement of a longer polyglutamate side chain for the effectiveness of F₄₂₀ in promoting an anti-M. tuberculosis activity of pretomanid and delamanid is unknown. We hypothesize that the enzymes that process delamanid and pretomanid to generate respective active drug forms require F₄₂₀ with longer side chains and F₄₂₀ with shorter chains are less effective in these activation reactions.

A hypothesis of the mechanistic basis for the requirement of reduced F₄₂₀ for polyglutamylation of F₄₂₀.

As mentioned above, the FbiB protein is composed of two distinct domains, one formed by the amino-terminal half (FbiB-N) and the other by the carboxy-terminal half of FbiB (FbiB-C), connected by a helical linker (44). Both domains are dimeric and bind F₄₂₀ with strong affinities;K_d (dissociation constant) values for a two-site model are ∼0.2 and ∼3 μM (44). It seemed that the FbiB-C domain is responsible for the extension of the glutamyl side chain beyond two glutamate residues. This is because, as detailed in the introduction, unlike full-length mycobacterial FbiB, standalone homologs of the amino-terminal half of FbiB (FbiB-N) that are found in evolutionarily deeply rooted methanogenic archaea, introduce only two glutamate residues into F₄₂₀ (44), and a recombinant form ofM. tuberculosis FbiB-N (MtbFbiB-N) exhibits similar activity, producing F₄₂₀-1 (44). However, heterologously producedMtbFbiB-C protein fails to ligate additional glutamate residues to F₄₂₀-1 (oxidized form) producedin vitro byMtbFbiB-N (44).

TheMtbFbiB-C domain is a homolog of the FMN-dependent family of nitroreductases (52), and an X-ray crystallographic study with a recombinant form of the protein expressed inE. coli has shown that it can bind F₄₂₀ and FMN but not both simultaneously, due to an irreconcilable steric clash (44). On the other hand, intactMtbFbiB purified from aMycobacterium smegmatis strain expressing this protein heterologously lacks FMN and contains F₄₂₀, whereas the same protein purified from a recombinantE. coli strain is faintly yellow, perhaps due to bound flavin (44). Also,in vitro binding data have demonstrated thatMtbFbiB-C produced inE. coli binds FMN poorly (K_d, ∼15 μM) and F₄₂₀ with 10-fold higher affinity (K_d, ∼1.5 μM) (44);MtbFbiB-N also binds F₄₂₀ tightly (K_d, ∼1.4 μM). TheMtbFbiB preparation used in our study lacked FMN or FAD, yet it was catalytically active. Thus, the observed binding of FMN toMtbFbiB-C is likely an artifact of expression of this domain inE. coli, and FMN may not have a role in the polyglutamylation of F₄₂₀. With this conclusion and the observed requirement of Fgd for the generation of F₄₂₀ with side chains carrying more than two glutamate residues inM. smegmatis andM. bovis BCG (Fig. 6), we hypothesize the following mechanisms for the polyglutamylation of F₄₂₀ (Fig. 2B). (i) Reduced F₄₂₀-0 or F₄₂₀-0H₂ is the physiological substrate for the polyglutamylation reaction. FbiB-N ligates the first two glutamates to F₄₂₀-0H₂, and the product, F₄₂₀-2H₂, is transferred to FbiB-C for further ligation reactions (green arrow inFig. 2B). Upon binding to FbiB-C, F₄₂₀-2H₂ induces structural changes to this domain, which in turn provide size and charge characteristics and hydrophobicity that are required for the side chain extension reactions. (ii) If F₄₂₀-0, an oxidized form, serves as the substrate, FbiB-N converts it to F₄₂₀-2, which instead of being transferred to FbiB-C is released from the enzyme, causing the polyglutamylation process to be aborted after the first round of ligation (red arrow inFig. 2B). Then Fgd reduces free F₄₂₀-2 to provide F₄₂₀-2H₂ to FbiB-C for further elongation (dotted arrow inFig. 2B). This entry of free F₄₂₀-2H₂ into the glutamylation process is less efficient. This hypothesis is consistent with the observation that in anin vitro system, the product with two glutamate residues in the side chain persists longer if the starting substrate is F₄₂₀-0 than what is seen with F₄₂₀H₂ (Fig. 5).

Conclusion and implications.

This report describes F₄₂₀H₂ as the true substrate of F₄₂₀-0-γ-glutamyl ligase for the attachment of more than two glutamate residues to the coenzyme. It also shows that in pathogenic and saprophytic mycobacteria the glutamylation process requires the participation of Fgd, F₄₂₀-dependent glucose-6-phosphate dehydrogenase (49), because in these organisms F₄₂₀H₂ is generated from F₄₂₀ by the action of Fgd (9,10,20,49). Interestingly, Fgd not only helps to generate F₄₂₀ with longer polyglutamyl side chains, but the catalytic properties of this enzyme are also influenced by the length of this side chain (29). These findings are relevant to the effort of developing new and effective TB drugs because, as mentioned above, the activation of two new TB drugs, delamanid and pretomanid, requires F₄₂₀ with longer polyglutamyl side chains. In addition, the observed link between the F₄₂₀ polyglutamylation reaction and Fgd brings further importance to this unusual glucose-6-phosphate dehydrogenase, which by itself is known to provide pathogenic resilience in TB-causingM. tuberculosis (9–11) and helps to convert a prodrug into an active TB drug (18,20–22).

The study also brought a comparative model for studies on a long-standing mystery in the field of folate biochemistry. For last 55 years, it has been known that folate or vitamin B₉ must be reduced to the H₄folate stage before folylpolyglutamate synthetases can introduce more than one glutamate residues in the side chain of this coenzyme (53). However, the mechanistic basis for this requirement remains to be elucidated (33,34,36,37,53,54). This is an important knowledge gap in biochemistry, as the variation in the length of the polyglutamylate side chain of H₄folate impacts the metabolism of a broad range of organisms and has implications in the development of major human diseases such as cancer and resistance to anticancer drugs (33,34,37–41,53,54). Now, our study shows that F₄₂₀-γ-glutamyl ligase is more efficient in adding more than two glutamate residues to F₄₂₀ only if this deazaflavin coenzyme is presented to the enzyme in a reduced form. Consequently, an advancement in studies of the mechanism guiding the polyglutamylation in one system would help similar investigation on the other.

Reagents and chemicals.

Coenzyme F₄₂₀-2 was purified fromMethanothermobacter thermautotrophicus cells, as described previously (49). All other reagents and chemicals were purchased from standard suppliers.

Preparation of F₄₂₀-0 and F₄₂₀H₂.

F₄₂₀-0 was prepared by hydrolyzing F₄₂₀-2 with carboxypeptidase G (Sigma-Aldrich, St. Louis, MO) according to an available protocol (1) and purified further by use of HPLC as detailed below. The pool of F₄₂₀-0 fractions from HPLC was evaporated to dryness under a N₂ stream at room temperature, and the dried product was dissolved in water. A two-electron reduced form of F₄₂₀ (F₄₂₀H₂) was either prepared via reduction of F₄₂₀ with sodium borohydride (10) or generatedin situ by including purified recombinant F₄₂₀-dependent glucose-6-phosphate dehydrogenase (Fgd) and glucose-6-phosphate (G6P) in the assay at concentrations of 0.15 mg/ml and 5 mM, respectively.

Generation of recombinant proteins.

The method for generation of recombinant proteins has been described previously (55,56). In short, the DNA for the coding region ofM. tuberculosisfbiB (rv3262) was PCR amplified and cloned into the NdeI and BamHI sites of pTEV5 (55), generating the expression vector pUL3262; pTEV5 allows the expression of a cloned gene under the control of alacO/LacI-controlled T7 promoter, producing a protein linked to an NH₂-terminal His₆-tag via a tobacco etch virus (TEV) protease cleavage site (55,57). Similarly, an expression vector forM. smegmatis Fgd (Msmeg_0777), pEP0777, was developed using pET19b (MilliporeSigma), and it was designed to express the protein with a NH₂-terminal His₆-tag and enterokinase cleavage site. These plasmids were introduced intoE. coli C41(DE3) (58), and the resulting strains were cultured in LB medium with 100 μg/ml ampicillin at 37°C with shaking. When the cultures reached an optical density at 600 nm of 0.4 to 0.6, as measured by using a Beckman Coulter DU800 spectrophotometer (Brea, CA), protein expression was induced by adding isopropyl-β-d-1-thiogalactopyranoside (IPTG) to a final concentration of 0.4 mM, and the cultivation was continued for four more hours. The heterologously expressed proteins were purified using Ni-nitrilotriacetic acid (NTA) columns, as described previously (59); the proteins were eluted with imidazole at the following concentrations: FbiB, 150 to 200 mM; and Fgd, 250 mM. In each case, the fractions judged homogenous via SDS-PAGE analysis were pooled, and each protein pool was made free of imidazole and concentrated via 3 rounds of filtration through a 10-kDa cutoff membrane of a Macrosep Advance centrifugal device (Pall Corporation, Port Washington, NY), where at every step the retentate was mixed with 50 mM potassium phosphate buffer (pH 7.0) and centrifugation was performed at 5,000 ×g. To this concentrate, glycerol was added to a final concentration of 50% (wt/vol), and the protein solution was stored at −20°C until further use. Storage without glycerol inactivatedM. tuberculosis FbiB.

FbiB assays.

F₄₂₀-γ-glutamyl ligase activity of FbiB was assayed as described previously (44) at 37°C using F₄₂₀-0, F₄₂₀-0H₂, F₄₂₀-2, or F₄₂₀-2H₂ as the substrate, with a reaction mixture of the following composition: HEPES-NaOH buffer (50 mM; pH 8.5), NaCl (100 mM), MnCl₂ (5 mM),l-glutamate (10 mM), GTP (5 mM), F₄₂₀-0 or F_420-2 (40 μM), and FbiB (0.37 mg/ml). At the early stage of this work, where F₄₂₀-0H₂ and F₄₂₀-2H₂ generated via reduction with sodium borohydride (10) were used as the substrates, the reactions were performed under strict anaerobic conditions with nitrogen (14-kPa gauge pressure) as the headspace gas, following an established protocol (60); an assay mixture without FbiB was made anaerobic and FbiB from an aerobic stock was added to initiate the reaction. For the reasons stated in Results and Discussion, for part of the study, the reduced substrates were generatedin situ by including Fgd (0.15 mg/ml) and glucose-6-phosphate (5 mM) in the reaction mixture, and here, the reactions could be performed under air. In both cases, the reaction was stopped by placing the mixture in a boiling water bath for 10 min. Then, the respective supernatant was recovered via centrifugation at 20,800 ×g for 15 min, filtered through a 0.45-μm Acrodisc syringe filter (13 mm diameter; Pall Corporation, Port Washington, NY) for further clarification, and analyzed by HPLC.

HPLC analysis.

The relative amounts of coenzyme F₄₂₀ derivatives with various lengths of glutamate side chain present in a mixture were determined by use of an HPLC unit (Prominence; Shimadzu Corporation, Kyoto, Japan) employing a Vydac analytical C₁₈ column (4.6 × 250 mm; particle size, 5 μm; catalog no. 218T54; Separation Group, Hesperia, CA). Solvent A was 2% acetonitrile in 27.5 mM sodium acetate-acetic acid buffer (pH 4.7), solvent B was acetonitrile (1), and the total flow rate was 0.6 ml/min. After the application of a sample under 100% solvent A, the following gradients of solvent B in solvent A were applied: 0 to 5 min, 0 to 5% B; 5 to 10 min, 5 to 15% B; 15 to 20 min, 25% B (isocratic); and 20 to 25 min, 25 to 0% B. Typical sample volume was 200 μl. A photodiode array detector (SPD M-20A; Shimadzu) was used for monitoring the elution at 400 nm and for obtaining the UV-visible spectra for the eluted compounds.

Mass spectrometric analysis of purified F₄₂₀ preparations.

In each case, an HPLC fraction (2 ml) containing a selected F₄₂₀ species was evaporated to dryness under a stream of nitrogen. The residue was dissolved in 200 μl of 50 mM ammonium formate-formic acid buffer (pH 3.6), and the resulting solution was mixed with 100 mg Biosil C18 silica powder (40 to 63 μm; Bio-Rad Laboratories, Inc., Hercules, CA) that was equilibrated with the above-mentioned buffer in a 1.5-ml polypropylene microcentrifuge tube. The mixture was centrifuged at 20,800 ×g for 5 min, and the supernatant was discarded. Then the silica powder was washed twice with two volumes of buffer (via resuspension, centrifugation, and removal of the supernatant), and coenzyme F₄₂₀ was eluted with methanol. The product was evaporated to dryness under a stream of nitrogen, dissolved in water, and analyzed via MALDI-TOF mass spectrometry at the School of Chemical Sciences Mass Spectrometry Laboratory at the University of Illinois at Urbana-Champaign. The Bruker peptide calibration mixture standard II (Bruker item no. 222570; mass range, 500 to 5,000 Da) was used for calibration, and the matrix was 2,5-dihydroxybenzoic acid. An UltrafleXtreme mass spectrometer (Bruker, Bremen, Germany) equipped with a smart beam II laser was used in the positive mode to acquire MALDI-TOF mass spectra. Samples were analyzed in the Reflectron mode.

Analysis of the side chain characteristics of F₄₂₀ inMycobacterium smegmatis andMycobacterium bovis BCG cells.

BothMycobacterium smegmatis mc² 155 (51) andMycobacterium bovis BCG (strain Pasteur 1173P2 (61) were grown in Middlebrook 7H9 liquid medium supplemented with 0.2% glycerol and 0.05% Tween 80; except in the latter case, the medium also contained 5% Middlebrook ADC (Becton, Dickinson and Company, Franklin Lakes, NJ). For the cultivation ofM. smegmatis Δfgd andM. bovis BCGfgd::Tn5367 strains, kanamycin (5 μg/ml) was included in the growth medium (9,10). The cells from the late-logarithmic phase of these cultures were analyzed for the side chain characteristics of F₄₂₀ by using a modified version of a previously described method (1). The cells were pelleted from a 150-ml culture via centrifugation at 10,000 ×g for 10 min at 4°C and resuspended in 2 ml of HPLC buffer A, as described above. The cell suspension was incubated in a boiling water bath for 15 min and centrifuged at 20,800 ×g for 15 min at room temperature. The resulting supernatant was stored at −20°C for 12 h and then centrifuged at 20,800 ×g for 15 min at room temperature. The supernatant from this stage was filtered through a syringe filter with 0.45-μm porosity, as described above, and the clarified supernatant was analyzed via HPLC.

Protein assay.

Protein concentration in solution was determined according to Bradford (62), using the Coomassie protein assay kit from Thermo Fisher Scientific (Waltham, MA). The protein content of a given amount ofMycobacterium smegmatis cells was determined by analyzing the respective cell lysate, which was prepared by use of a French pressure cell (49).

• 1.Bair TB, Isabelle DW, Daniels L. 2001. Structures of coenzyme F₄₂₀ inMycobacterium species. Arch Microbiol176:37–43. doi: 10.1007/s002030100290. [DOI] [PubMed] [Google Scholar]
• 2.Eirich LD, Vogels GD, Wolfe RS. 1978. Proposed structure for coenzyme F₄₂₀ fromMethanobacterium. Biochemistry17:4583–4593. doi: 10.1021/bi00615a002. [DOI] [PubMed] [Google Scholar]
• 3.Eirich LD, Vogels GD, Wolfe RS. 1979. Distribution of coenzyme F₄₂₀ and properties of its hydrolytic fragments. J Bacteriol140:20–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 4.Peck MW.1989. Changes in concentrations of coenzyme F₄₂₀ analogs during batch growth ofMethanosarcina barkeri andMethanosarcina mazei. Appl Environ Microbiol55:940–945. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 5.Greening C, Jirapanjawat T, Afroze S, Ney B, Scott C, Pandey G, Lee BM, Russell RJ, Jackson CJ, Oakeshott JG, Taylor MC, Warden AC. 2017. Mycobacterial F₄₂₀H₂-dependent reductases promiscuously reduce diverse compounds through a common mechanism. Front Microbiol8:1000. doi: 10.3389/fmicb.2017.01000. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 6.DiMarco AA, Bobik TA, Wolfe RS. 1990. Unusual coenzymes of methanogenesis. Annu Rev Biochem59:355–394. doi: 10.1146/annurev.bi.59.070190.002035. [DOI] [PubMed] [Google Scholar]
• 7.Greening C, Ahmed FH, Mohamed AE, Lee BM, Pandey G, Warden AC, Scott C, Oakeshott JG, Taylor MC, Jackson CJ. 2016. Physiology, biochemistry, and applications of F₄₂₀- and F_o-dependent redox reactions. Microbiol Mol Biol Rev80:451–493. doi: 10.1128/MMBR.00070-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 8.Hasan MR, Rahman M, Jaques S, Purwantini E, Daniels L. 2010. Glucose 6-phosphate accumulation in mycobacteria: implications for a novel F₄₂₀-dependent anti-oxidant defense system. J Biol Chem285:19135–19144. doi: 10.1074/jbc.M109.074310. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 9.Purwantini E, Daniels L, Mukhopadhyay B. 2016. F₄₂₀H₂ is required for phthiocerol dimycocerosate synthesis in mycobacteria. J Bacteriol198:2020–2028. doi: 10.1128/JB.01035-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 10.Purwantini E, Mukhopadhyay B. 2009. Conversion of NO₂ to NO by reduced coenzyme F₄₂₀ protects mycobacteria from nitrosative damage. Proc Natl Acad Sci U S A106:6333–6338. doi: 10.1073/pnas.0812883106. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11.Purwantini E, Mukhopadhyay B. 2013. Rv0132c ofMycobacterium tuberculosis encodes a coenzyme F₄₂₀-dependent hydroxymycolic acid dehydrogenase. PLoS One8:e81985. doi: 10.1371/journal.pone.0081985. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 12.Lapalikar GV, Taylor MC, Warden AC, Scott C, Russell RJ, Oakeshott JG. 2012. F₄₂₀H₂-dependent degradation of aflatoxin and other furanocoumarins is widespread throughout the actinomycetales. PLoS One7:e30114. doi: 10.1371/journal.pone.0030114. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 13.Taylor MC, Jackson CJ, Tattersall DB, French N, Peat TS, Newman J, Briggs LJ, Lapalikar GV, Campbell PM, Scott C, Russell RJ, Oakeshott JG. 2010. Identification and characterization of two families of F₄₂₀ H₂-dependent reductases from Mycobacteria that catalyse aflatoxin degradation. Mol Microbiol78:561–575. doi: 10.1111/j.1365-2958.2010.07356.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 14.Gurumurthy M, Rao M, Mukherjee T, Rao SP, Boshoff HI, Dick T, Barry CE III, Manjunatha UH. 2013. A novel F₄₂₀ -dependent anti-oxidant mechanism protectsMycobacterium tuberculosis against oxidative stress and bactericidal agents. Mol Microbiol87:744–755. doi: 10.1111/mmi.12127. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 15.Ebert S, Fischer P, Knackmuss HJ. 2001. Converging catabolism of 2,4,6-trinitrophenol (picric acid) and 2,4-dinitrophenol byNocardioides simplex FJ2-1A. Biodegradation12:367–376. doi: 10.1023/A:1014447700775. [DOI] [PubMed] [Google Scholar]
• 16.Nakano T, Miyake K, Endo H, Dairi T, Mizukami T, Katsumata R. 2004. Identification and cloning of the gene involved in the final step of chlortetracycline biosynthesis inStreptomyces aureofaciens. Biosci Biotechnol Biochem68:1345–1352. doi: 10.1271/bbb.68.1345. [DOI] [PubMed] [Google Scholar]
• 17.Wang P, Bashiri G, Gao X, Sawaya MR, Tang Y. 2013. Uncovering the enzymes that catalyze the final steps in oxytetracycline biosynthesis. J Am Chem Soc135:7138–7141. doi: 10.1021/ja403516u. [DOI] [PubMed] [Google Scholar]
• 18.Singh R, Manjunatha U, Boshoff HI, Ha YH, Niyomrattanakit P, Ledwidge R, Dowd CS, Lee IY, Kim P, Zhang L, Kang S, Keller TH, Jiricek J, Barry CE 3rd. 2008. PA-824 kills nonreplicatingMycobacterium tuberculosis by intracellular NO release. Science322:1392–1395. doi: 10.1126/science.1164571. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 19.Cellitti SE, Shaffer J, Jones DH, Mukherjee T, Gurumurthy M, Bursulaya B, Boshoff HI, Choi I, Nayyar A, Lee YS, Cherian J, Niyomrattanakit P, Dick T, Manjunatha UH, Barry CE III, Spraggon G, Geierstanger BH. 2012. Structure of Ddn, the deazaflavin-dependent nitroreductase fromMycobacterium tuberculosis involved in bioreductive activation of PA-824. Structure20:101–112. doi: 10.1016/j.str.2011.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 20.Stover CK, Warrener P, VanDevanter DR, Sherman DR, Arain TM, Langhorne MH, Anderson SW, Towell JA, Yuan Y, McMurray DN, Kreiswirth BN, Barry CE, Baker WR. 2000. A small-molecule nitroimidazopyran drug candidate for the treatment of tuberculosis. Nature405:962–966. doi: 10.1038/35016103. [DOI] [PubMed] [Google Scholar]
• 21.Hartkoorn RC, Ryabova OB, Chiarelli LR, Riccardi G, Makarov V, Cole ST. 2014. Mechanism of action of 5-nitrothiophenes againstMycobacterium tuberculosis. Antimicrob Agents Chemother58:2944–2947. doi: 10.1128/AAC.02693-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 22.Fujiwara M, Kawasaki M, Hariguchi N, Liu Y, Matsumoto M. 2018. Mechanisms of resistance to delamanid, a drug forMycobacterium tuberculosis. Tuberculosis108:186–194. doi: 10.1016/j.tube.2017.12.006. [DOI] [PubMed] [Google Scholar]
• 23.Mayerl F, Piret J, Kiener A, Walsh CT, Yasui A. 1990. Functional expression of 8-hydroxy-5-deazaflavin-dependent DNA photolyase fromAnacystis nidulans inStreptomyces coelicolor. J Bacteriol172:6061–6065. doi: 10.1128/jb.172.10.6061-6065.1990. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24.Glas AF, Maul MJ, Cryle M, Barends TR, Schneider S, Kaya E, Schlichting I, Carell T. 2009. The archaeal cofactor F0 is a light-harvesting antenna chromophore in eukaryotes. Proc Natl Acad Sci U S A106:11540–11545. doi: 10.1073/pnas.0812665106. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 25.Eker AP, Kooiman P, Hessels JK, Yasui A. 1990. DNA photoreactivating enzyme from the cyanobacteriumAnacystis nidulans. J Biol Chem265:8009–8015. [PubMed] [Google Scholar]
• 26.Zhong D.2015. Electron transfer mechanisms of DNA repair by photolyase. Annu Rev Phys Chem66:691–715. doi: 10.1146/annurev-physchem-040513-103631. [DOI] [PubMed] [Google Scholar]
• 27.Li H, Graupner M, Xu H, White RH. 2003. CofE catalyzes the addition of two glutamates to F₄₂₀-0 in F₄₂₀ coenzyme biosynthesis inMethanococcus jannaschii. Biochemistry42:9771–9778. doi: 10.1021/bi034779b. [DOI] [PubMed] [Google Scholar]
• 28.Li H, Xu H, Graham DE, White RH. 2003. Glutathione synthetase homologs encode alpha-l-glutamate ligases for methanogenic coenzyme F₄₂₀ and tetrahydrosarcinapterin biosyntheses. Proc Natl Acad Sci U S A100:9785–9790. doi: 10.1073/pnas.1733391100. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29.Ney B, Carere CR, Sparling R, Jirapanjawat T, Stott MB, Jackson CJ, Oakeshott JG, Warden AC, Greening C. 2017. Cofactor tail length modulates catalysis of bacterial F₄₂₀-dependent oxidoreductases. Front Microbiol8:1902. doi: 10.3389/fmicb.2017.01902. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 30.Gupta R, Wells CD, Hittel N, Hafkin J, Geiter LJ. 2016. Delamanid in the treatment of multidrug-resistant tuberculosis. Int J Tuberc Lung Dis20:33–37. doi: 10.5588/ijtld.16.0125. [DOI] [PubMed] [Google Scholar]
• 31.Silva DR, Dalcolmo M, Tiberi S, Arbex MA, Munoz-Torrico M, Duarte R, D'Ambrosio L, Visca D, Rendon A, Gaga M, Zumla A, Migliori GB. 2018. New and repurposed drugs to treat multidrug- and extensively drug-resistant tuberculosis. J Bras Pneumol44:153–160. doi: 10.1590/s1806-37562017000000436. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 32.Pstragowski M, Zbrzezna M, Bujalska-Zadrozny M. 2017. Advances in pharmacotherapy of tuberculosis. Acta Pol Pharm74:3–11. [PubMed] [Google Scholar]
• 33.Cossins EA, Chen L. 1997. Folates and one-carbon metabolism in plants and fungi. Phytochemistry45:437–452. doi: 10.1016/S0031-9422(96)00833-3. [DOI] [PubMed] [Google Scholar]
• 34.Shane B.2008. Folate and vitamin B₁₂ metabolism: overview and interaction with riboflavin, vitamin B6, and polymorphisms. Food Nutr Bull29:S5–S19. doi: 10.1177/15648265080292S103. [DOI] [PubMed] [Google Scholar]
• 35.Bailey LB, Stover PJ, McNulty H, Fenech MF, Gregory JF III, Mills JL, Pfeiffer CM, Fazili Z, Zhang M, Ueland PM, Molloy AM, Caudill MA, Shane B, Berry RJ, Bailey RL, Hausman DB, Raghavan R, Raiten DJ. 2015. Biomarkers of nutrition for development—folate review. J Nutr145:1636S-1680S. doi: 10.3945/jn.114.206599. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 36.Bertacine Dias MV, Santos JC, Libreros-Zuniga GA, Ribeiro JA, Chavez-Pacheco SM. 2018. Folate biosynthesis pathway: mechanisms and insights into drug design for infectious diseases. Future Med Chem10:935–959. doi: 10.4155/fmc-2017-0168. [DOI] [PubMed] [Google Scholar]
• 37.Green JM, Matthews RG. 2007. Folate biosynthesis, reduction, and polyglutamylation and the interconversion of folate derivatives. EcoSal Plus2007. doi: 10.1128/ecosalplus.3.6.3.6. [DOI] [PubMed] [Google Scholar]
• 38.Kim SE, Cole PD, Cho RC, Ly A, Ishiguro L, Sohn KJ, Croxford R, Kamen BA, Kim YI. 2013. Gamma-glutamyl hydrolase modulation and folate influence chemosensitivity of cancer cells to 5-fluorouracil and methotrexate. Br J Cancer109:2175–2188. doi: 10.1038/bjc.2013.579. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 39.Longley DB, Harkin DP, Johnston PG. 2003. 5-fluorouracil: mechanisms of action and clinical strategies. Nat Rev Cancer3:330–338. doi: 10.1038/nrc1074. [DOI] [PubMed] [Google Scholar]
• 40.Visentin M, Zhao R, Goldman ID. 2012. The antifolates. Hematol Oncol Clin North Am26:629–648, ix. doi: 10.1016/j.hoc.2012.02.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 41.Zhao R, Titus S, Gao F, Moran RG, Goldman ID. 2000. Molecular analysis of murine leukemia cell lines resistant to 5,10-dideazatetrahydrofolate identifies several amino acids critical to the function of folylpolyglutamate synthetase. J Biol Chem275:26599–26606. doi: 10.1074/jbc.M002580200. [DOI] [PubMed] [Google Scholar]
• 42.Sohn KJ, Smirnakis F, Moskovitz DN, Novakovic P, Yates Z, Lucock M, Croxford R, Kim YI. 2004. Effects of folylpolyglutamate synthetase modulation on chemosensitivity of colon cancer cells to 5-fluorouracil and methotrexate. Gut53:1825–1831. doi: 10.1136/gut.2004.042713. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 43.Choi KP, Bair TB, Bae YM, Daniels L. 2001. Use of transposon Tn5367 mutagenesis and a nitroimidazopyran-based selection system to demonstrate a requirement forfbiA andfbiB in coenzyme F₄₂₀ biosynthesis byMycobacterium bovis BCG. J Bacteriol183:7058–7066. doi: 10.1128/JB.183.24.7058-7066.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 44.Bashiri G, Rehan AM, Sreebhavan S, Baker HM, Baker EN, Squire CJ. 2016. Elongation of the poly-gamma-glutamate tail of F₄₂₀ requires both domains of the F₄₂₀:gamma-glutamyl ligase (FbiB) ofMycobacterium tuberculosis. J Biol Chem291:6882–6894. doi: 10.1074/jbc.M115.689026. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 45.Bermingham A, Derrick JP. 2002. The folic acid biosynthesis pathway in bacteria: evaluation of potential for antibacterial drug discovery. Bioessays24:637–648. doi: 10.1002/bies.10114. [DOI] [PubMed] [Google Scholar]
• 46.Bourne CR.2014. Utility of the biosynthetic folate pathway for targets in antimicrobial discovery. Antibiotics (Basel)3:1–28. doi: 10.3390/antibiotics3010001. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 47.Salcedo E, Cortese JF, Plowe CV, Sims PF, Hyde JE. 2001. A bifunctional dihydrofolate synthetase-folylpolyglutamate synthetase inPlasmodium falciparum identified by functional complementation in yeast and bacteria. Mol Biochem Parasitol112:239–252. doi: 10.1016/S0166-6851(00)00370-4. [DOI] [PubMed] [Google Scholar]
• 48.Purwantini E, Daniels L. 1998. Molecular analysis of the gene encoding F_420-dependent glucose-6-phosphate dehydrogenase fromMycobacterium smegmatis. J Bacteriol180:2212–2219. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 49.Purwantini E, Daniels L. 1996. Purification of a novel coenzyme F_420-dependent glucose-6-phosphate dehydrogenase fromMycobacterium smegmatis. J Bacteriol178:2861–2866. doi: 10.1128/jb.178.10.2861-2866.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 50.Spencer R, Fisher J, Walsh C. 1976. Preparation, characterization, and chemical properties of the flavin coenzyme analogues 5-deazariboflavin, 5-deazariboflavin 5′-phosphate, and 5-deazariboflavin 5′-diphosphate, 5′ → 5′-adenosine ester. Biochemistry15:1043–1053. doi: 10.1021/bi00650a015. [DOI] [PubMed] [Google Scholar]
• 51.Snapper SB, Melton RE, Mustafa S, Kieser T, Jacobs WR Jr.. 1990. Isolation and characterization of efficient plasmid transformation mutants ofMycobacterium smegmatis. Mol Microbiol4:1911–1919. doi: 10.1111/j.1365-2958.1990.tb02040.x. [DOI] [PubMed] [Google Scholar]
• 52.Akiva E, Copp JN, Tokuriki N, Babbitt PC. 2017. Evolutionary and molecular foundations of multiple contemporary functions of the nitroreductase superfamily. Proc Natl Acad Sci U S A114:E9549–E9558. doi: 10.1073/pnas.1706849114. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 53.Griffin MJ, Brown GM. 1964. The biosynthesis of folic acid. III. Enzymatic Formation of dihydrofolic acid from dihydropteroic acid and of tetrahydropteroylpolyglutamic acid compounds from tetrahydrofolic acid. J Biol Chem239:310–316. [PubMed] [Google Scholar]
• 54.Toy J, Bognar AL. 1990. Cloning and expression of the gene encodingLactobacillus casei folylpoly-gamma-glutamate synthetase inEscherichia coli and determination of its primary structure. J Biol Chem265:2492–2499. [PubMed] [Google Scholar]
• 55.Rocco CJ, Dennison KL, Klenchin VA, Rayment I, Escalante-Semerena JC. 2008. Construction and use of new cloning vectors for the rapid isolation of recombinant proteins fromEscherichia coli. Plasmid59:231–237. doi: 10.1016/j.plasmid.2008.01.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 56.Susanti D, Loganathan U, Compton A, Mukhopadhyay B. 2017. A reexamination of thioredoxin reductase fromThermoplasma acidophilum, a thermoacidophilic euryarchaeon, identifies it as an NADH-dependent enzyme. ACS Omega2:4180–4187. doi: 10.1021/acsomega.7b00640. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 57.Blommel PG, Fox BG. 2007. A combined approach to improving large-scale production of tobacco etch virus protease. Protein Expr Purif55:53–68. doi: 10.1016/j.pep.2007.04.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 58.Miroux B, Walker JE. 1996. Over-production of proteins inEscherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol260:289–298. doi: 10.1006/jmbi.1996.0399. [DOI] [PubMed] [Google Scholar]
• 59.Susanti D, Loganathan U, Mukhopadhyay B. 2016. A novel F₄₂₀-dependent thioredoxin reductase gated by low potential FAD: a tool for redox regulation in an anaerobe. J Biol Chem291:23084–23100. doi: 10.1074/jbc.M116.750208. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 60.Mukhopadhyay B, Daniels L. 1989. Aerobic purification ofN⁵,N¹⁰-methylenetetrahydromethanopterin dehydrogenase, separated fromN⁵,N¹⁰-methylenetetrahydromethanopterin cyclohydrolase, fromMethanobacterium thermoautotrophicum strain Marburg. Can J Microbiol35:499–507. doi: 10.1139/m89-077. [DOI] [PubMed] [Google Scholar]
• 61.Brosch R, Gordon SV, Garnier T, Eiglmeier K, Frigui W, Valenti P, Dos Santos S, Duthoy S, Lacroix C, Garcia-Pelayo C, Inwald JK, Golby P, Garcia JN, Hewinson RG, Behr MA, Quail MA, Churcher C, Barrell BG, Parkhill J, Cole ST. 2007. Genome plasticity of BCG and impact on vaccine efficacy. Proc Natl Acad Sci U S A104:5596–5601. doi: 10.1073/pnas.0700869104. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 62.Bradford MM.1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem72:248–254. doi: 10.1016/0003-2697(76)90527-3. [DOI] [PubMed] [Google Scholar]