Luria-Bertani (Miller) broth (LB) was from EMD Millipore (Gibbstown, New Jersey) and purchased through Fisher Scientific (Pittsburgh, PA). DNase I and lysozyme were from Sigma Chemical Company (St. Louis, Missouri). Isopropyl β-D-1-thiogalactopyranoside (IPTG) was from UBP Bio, LLC (Aurora, CO). Protein dye-binding reagent, bovine serum albumin standard, Precision Plus Protein^TM Dual Color Standards, 4% - 15% native gels, and Bio-Safe^TM Coomassie Stain were from Bio-Rad (Hercules, CA). Dithionite, dithiothreitol (DTT), Invitrogen Novex NativeMARK™ Protein Standards, and all other chemicals were purchased from Fisher Scientific, unless otherwise specified.

Burkholderia xenovorans DmrB with an N-terminal histidine tag (His₆-DmrB) was produced in Escherichia coli BL21 (DE3) RIL cells (Stratagene, La Jolla, CA) containing pET41a:his₆-DmrB, and cells were harvested by centrifugation, as described previously by McNamara et al. [12]. The cell pellet (about 3 g) was thawed from −20˚C and suspended in lysis buffer (50 mM tris(hydroxymethyl)aminomethane (Tris), pH 8.0, 200 mM NaCl, 20 mM imidazole) at a ratio of 2 ml of buffer per g of cell pellet. Cells were lysed using a French Press (Thermo-Fisher Scientific) at 20,000 psi and centrifuged at 32,000 × g for 30 min. The supernatant (cell-free extract, CFE) was mixed with nickel nitrilotriacetic acid (Ni-NTA) agarose beads (Qiagen, Chatsworth, CA) (0.25 ml of beads per ml of CFE) and incubated on ice with gentle rocking for 2 h. The beads were then poured in a plastic 10-ml column (Bio-Rad, Hercules, CA), and the flow-through fraction was collected. The beads were washed three times with 5 ml of wash buffer (50 mM Tris, pH 8.0, 200 mM NaCl, 30 mM imidazole). Some His₆-DmrB eluted upon washing twice with 2 ml of 50 mM Tris, pH 8.0, 200 mM NaCl, 100 mM imidazole, but more highly purified protein (~95% pure) eluted with two additions of elution buffer (2 ml) containing 250 mM imidazole. The protein was stored at 4˚C until use.

A codon-optimized gene encoding an NAD(P)H-flavin reductase (Fre) from S. enterica (NP_462864.1) [14] with an N-terminal six-histidine tag (His₆-Fre) was synthesized by GenScript (Piscataway, NJ), subcloned into the NdeI-HinDIII sites of pET41a, and transformed into E. coli BL21(DE3)-RIL cells. Cells were grown and harvested in the same manner as for DmrB, and protein was purified aerobically using the same protocol as DmrB, except that His₆-Fre tended to elute in 100 mM imidazole. All centrifugation and incubation steps were carried out at 4˚C.

His₆-Fre could be stored in its elution buffer and could remain stable for at least one month at 4˚C in that buffer. However, His₆-Fre was found to be sensitive to temperature fluctuations when stored in a refrigerator that was opened and closed frequently. Therefore, the active fraction (Elution 1, 100 mM imidazole) was separated into 200-μl aliquots, sealed in 2-ml vials, and kept in a 4˚C cold room to minimize temperature fluctuations. For enzymatic studies, His₆-Fre was maintained at 4˚C, either on ice or on Lab Armor cold beads (Fisher Scientific).

Electron transfer assays were conducted anaerobically in a chamber containing 98% N₂ and 2% H₂ (Coy Laboratories, Grass Lake, MI) or in anaerobic cuvettes sealed with Teflon-lined silicon septa. Solutions were purged with a flow of N₂ gas and then brought into the anaerobic chamber. If proteins were to be stored in the absence of O₂ after purification, samples were placed inside an anaerobic chamber and 2-mercaptoethanol was added to a final concentration of 15 mM.

Dithionite (10 mM) and DTT (up to 50 mM) were tested as artificial electron donors for His₆-DmrB by monitoring a decrease in the flavin absorbance peaks at 380 and 460 nm with either a Beckman model DU-800 spectrophotometer (Beckman Coulter, Inc., Brea, CA) using sealed anaerobic quartz cuvettes or a Bio-Rad SmartSpec Plus inside an anaerobic chamber. For dithionite reduction, purified protein (~1.25 mg of His₆-DmrB) was diluted to 990 µl with 20 mM Tris, pH 8.0, 100 mM sodium chloride, 5% glycerol and then mixed with 10 µl of 1 M dithionite. For DTT reduction tests, 50 µl of 1 M DTT and 1.25 mg of protein in 950 µl of 20 mM Tris, pH 8.0, 100 mM sodium chloride, 5% glycerol were mixed.

A ping-pong mechanism for DmrB has been proposed in which an enzyme (Fre) reduces a loosely bound FMN cofactor (FMN-2) of DmrB [12]. To begin to test this hypothesis, an assay using Fre from S. enterica [14] was developed by monitoring a decrease in the flavin absorbance peaks at 380 and 460 nm in the presence of NADH and Fre with a six-histidine tag (His₆-Fre). To conserve protein resources, a small-scale (200 μl) assay was also developed using a Bio-Rad SmartSpec Plus and plastic trUView cuvettes (Bio-Rad) inside the anaerobic chamber. Stock solutions of [5 mM FMN in 20 mM Tris pH 8.0] and [10 mM NADH in 50 mM 3-(N-morpholino)propanesulfonic acid (MOPS), pH 6.8] were used. The standard assay (200 μl) consisted of 194 μl of 50 mM MOPS, pH 7.0, 2 μl of 5 mM FMN (50 μM final concentration) and 2 μl of 10 mM NADH (100 μM final concentration). Although an E. coli version of Fre is reported to use both NADH and NADPH as co-substrates [15] [16], under the conditions studied in the current work, the S. enterica His₆-Fre could use only NADH efficiently. The reaction was initiated by addition of 2 μl of His₆-Fre enzyme (stock enzyme concentration was 2 to 2.5 mg protein/ml).

For experiments to estimate the reaction rate, the protein amounts used above were too high, and thus the enzyme solution was diluted 50-fold with assay buffer. His₆-Fre was kept on 4˚C beads during these assays since it would lose activity at higher temperatures throughout the day, as described above. For studies in the absence of exogenous FMN, the standard assay (200 μl) included 196 μl of enzyme and 2 μl of 10 mM NADH (100 μM final). The reaction was initiated by the addition of His₆-Fre. The solution buffering the enzyme was used as assay buffer.

All assays were conducted anaerobically. His₆-DmrB (2.5 mg) was diluted with 983 µl of assay buffer (20 mM Tris, pH 8.0, 100 mM sodium chloride, 5% glycerol), and then 10 µl of 10 mM NADH (100 μM final concentration) and 7 µl of 3.33 μg/µl His₆-Fre (20 μg) were added (1 ml final volume). No exogenous FMN was added. NADPH (100 µM final concentration) was also tested as an artificial electron donor in the presence of His₆-Fre.

Methanogen Methanosarcina thermophila cells were previously grown on acetate by Joseph Scott (University of Florida, Gainesville), stored in liquid N₂, and lysed as described previously [17]. Cells (2.5 g) were mixed with breakage buffer (5 ml of 50 mM TES, pH 7.0, 10 mM MgCl₂, 5% glycerol, and 10 mM 2-mercaptoethanol) and 2 μl DNase and broken anaerobically with one passage using a cold (4˚C) French Press (ThermoScientific). After anaerobic centrifugation at 32,000 × g (4˚C) for 45 min, the supernatant (cell-free extract) was flash-frozen and stored as pellets in liquid N₂ until use.

It has previously been shown that CO and extracts of the sulfate-reducing archaeon Archaeoglobus fulgidus could be used to reduce archaeal flavoprotein A (AfpA), a DmrB homolog from A. fulgidus [11]. Electrons from CO presumably flow to a large carbon monoxide dehydrogenase (CODH) complex in the extracts, where they are transferred to a ferredoxin or a ferredoxin-like protein, and then to an archaeal flavoprotein AfpA [11] that is a homolog of DmrB. Using similar reasoning, we tested the ability of CO and extracts from the methanogen M. thermophila to reduce purified His₆-DmrB.

His₆-DmrB was purified as described above and stored at 4˚C until use. Enzyme solutions were centrifuged before use and then desalted using prepacked 8.3-ml columns of Sephadex G-25 (GE Healthcare, Marlborough, MA, purchased from Fisher Scientific) to separate any FMN that might have become unbound during storage.

Initial absorbance values at 340 nm and 460 nm of a small sample (100 μl) of as-purified His₆-DmrB were determined using plastic trUView cuvettes (Bio-Rad) and a Bio-Rad Smart-Spec Plus spectrophotometer to establish an amount within the detection limit of the instrument [1.5 absorbance units (AU) or less]. If on-scale, then a quartz cuvette was sealed anaerobically and flushed with CO (or N₂ for controls) for 3 min to clear away any O₂ that would interfere with reduction studies. Using anaerobic technique, 1 ml of His₆-DmrB protein sample was injected into the CO-flushed quartz cuvette and an initial reading at 460 nm was taken. A Beckman model DU-640 spectrophotometer (Beckman Coulter) was used for this and continuing steps. Using an anaerobic syringe, 5 μl of M. thermophila cell extract containing CODH was injected into the cuvette containing CO and His₆-DmrB. FMN reduction was followed at 460 nm (ε_{460 nm} = 12,500 M⁻¹∙cm⁻¹) [18].

To produce an expression plasmid for heterologous protein production, a DNA sequence encoding MM1853 (accession number AAM31549.1) with an N-terminal histidine tag (his₆-MM1853) was synthesized by GenScript with codon optimization for E. coli. The modified gene was cloned first into pUC57 and then subcloned into the NdeI and BamHI sites of the pET41a expression vector (Novagen). Chemically competent BL21(DE3)-RIL cells were transformed and selected on plates containing 50 µg kanamycin/ml, creating the cell line MP01.

For His₆-MM1853 production, 1 L of LB broth with kanamycin (50 μg/ml) was inoculated with 25 ml of an overnight culture of MP01 cells and incubated with shaking (180 rpm) at 37˚C until an OD₆₀₀ of 0.6 was reached. Gene expression was induced with IPTG to 1 mM. After 6 h of incubation at 30˚C, the cells were harvested using a Sorvall RC-5C Plus centrifuge (Fisher Scientific) with an F9 rotor at 5000 rpm (4700 × g) at 4˚C for 15 min. The cell pellet was washed with 30 ml of 50 mM Tris, pH 8.0, and the cells were centrifuged in an F13 rotor at 5000 rpm (4300 × g). The pellet was stored at −80˚C until use.

During the aerobic purification of His₆-MM1853, all manipulations were carried out on ice. Frozen E. coli (MP01) cells (~3 g wet weight from 1 L of medium) were suspended in 6 ml of lysis buffer (50 mM Tris, pH 8.0, 200 mM sodium chloride, 20 mM imidazole). The cells were broken using a French press and centrifuged as described for DmrB. The supernatant (cell-free extract, CFE) was mixed with 1.5 ml of Ni-NTA agarose and incubated with shaking on ice for 2 h. The column was developed as described above for the purification of His₆-DmrB.

Protein concentrations were measured using the Bradford assay [19] with bovine serum albumin (Bio-Rad) as the standard. SDS-PAGE [20] with Bio-Safe^TM Coomassie Stain (Bio-Rad) was used to analyze protein solubility and purity.

DTT (10 mM) was tested as an artificial electron donor for His₆-MM1853 using ~1.25 mg of purified protein in a final volume of 1 ml with the same protocol described for His₆-DmrB. In addition, the NADH/Fre system was tested for the ability to reduce His₆-MM1853 (0.5 mg of protein in a final volume of 1 ml) as described for His₆-DmrB. The ability of NADPH to serve as an electron donor in the presence of His₆-Fre was also examined using an NADPH concentration of 100 µM.

The first step in developing an activity assay for DmrB is to identify an electron donor that reduces the FMN prosthetic group of the enzyme without directly reducing the substrate (dihydromethanopterin) through a nonenzymatic process. The UV-visible absorbance spectrum of oxidized FMN has peaks near 278, 380, and 460 nm; therefore, the ability of different chemical compounds to reduce the tightly bound FMN of His₆-DmrB can be conveniently measured as a decrease in the absorbance at these wavelengths. Initially, several common chemical electron donors were selected for analysis. These included dithionite, with a standard redox potential at pH 7 of −0.66 V, compared to FMN (−0.207 V) (21), dithiothreitol (DTT, −0.33 V) [21], β-mercaptoethanol (−0.26 V) [22], and NADH (−0.32 V).

Treatment with the powerful reductant dithionite rapidly reduced the tightly bound FMN of His₆-DmrB (data not shown), as observed previously for the archaeal flavoprotein homologs AfpA and DmrX [10] [11]. Unfortunately, dithionite is too strong for use in a DmrB assay because it can also nonenzymatically reduce the substrate dihydromethanopterin. We then tested DTT at concentrations from 30 to 100 mM, but DTT was unable to reduce the FMN prosthetic group of His₆-DmrB over the period of one hour (Figure 2). This result was unexpected since DTT effectively reduces the archaeal homolog DmrX and has been used to develop a successful activity assay for that enzyme [10]. We hypothesize that the 4Fe-4S clusters of DmrX may facilitate electron transfer to FMN in the archaeal dihydromethanopterin reductases, while the lack of iron-sulfur clusters in His₆-DmrB may prevent the use of DTT as an electron donor. Neither NAD(P)H nor β-mercaptoethanol could influence the absorbance peaks of DmrB (data not shown), which is consistent with the published results for DmrX and AfpA [10] [11].

Having failed to identify a chemical reductant suitable for a DmrB assay among our selected chemical candidates, we next examined the hypothesis of McNamara et al. (12) that an electron transport system consisting of exogenous FMN, NADH, and NAD(P)H-dependent flavin reductase (Fre) may provide electrons to DmrB. Preliminary experiments in the laboratory provided ambiguous yet promising preliminary results that NADH combined with exogenous FMN and

recombinant S. enterica Fre could partially reduce the flavin peaks of His₆-DmrB (unpublished data). Moreover, in a control experiment omitting exogenous FMN, the remaining components (NADH and Fre) unexpectedly appeared sufficient to partially reduce the FMN prosthetic group of His₆-DmrB.

To investigate these results in a more systematic way, we first purified His₆-DmrB using meticulous anaerobic technique and recorded the UV-visible spectrum from 300 to 550 nm (Figure 3(a), line 1). Upon addition of NADH, the NADH peak at 340 nm interfered with the flavin peak of DmrB at 380 nm (Figure 3(a), line 2), and thus only the second flavin peak near 460 nm was used monitor electron flow between the electron donor system and His₆-DmrB. When the reduction assay was initiated by adding 20 µg of His₆-Fre, a slight increase (0.050 absorbance unit at 460 nm) was initially observed (data not shown), indicating that the injected solution might not have been fully anaerobic. However, after 10 min, a decrease of 0.163 absorbance unit was observed (Figure 3(a), line 2), and the absorbance continued to decrease in subsequent time points. In a

typical experiment, the reaction appeared to be complete within about 65 min (Figure 3(a), line 4), since the 460-nm peak did not decrease substantially at later time points. The decrease in absorbance at 460 nm is consistent with electron transfer from NADH to Fre to DmrB; however, contrary to the prediction of the ping-pong mechanism [12], exogenous FMN did not appear to be a requirement for Fre-mediated electron transfer to DmrB.

To quantify the rate of His₆-DmrB reduction in each time interval by the NADH/Fre system using the molar extinction coefficient for FMN at 460 nm (12,500 M⁻¹∙cm⁻¹) [18], the following equation was used:

[ H i s 6 − D m r B F M N r e d u c e d ] / m i n = Δ A b s 4 6 0 n m / 1 2 , 5 0 0 M − 1 ⋅ c m − 1 / a s s a y i n t e r v a l ( m i n ) (1)

To estimate an initial rate of His₆-DmrB reduction, the absorbance at 460 nm was monitored at 5 min intervals, and the data points at 0 min, 5 min, and 10 min were used for calculating the velocity. In this experiment, the initial absorbance of His₆-DmrB at 460 nm was 0.733 (Figure 3(b), blue line). After 5 min, a decrease of 0.081 absorbance unit was observed, corresponding to a reduction rate of 1.3 µM/min. In the interval from 5 min to 10 min, the absorbance decreased by 0.077 (1.2 µM/min). Over the next 10-min interval (from 10 to 20 min), the rate dropped to 0.522 µM/min. As a control, when was His₆-Fre inactivated by heat, no loss of absorbance at 460 nm was observed (Figure 4(b), red line).

Ding and Ferry [11] showed previously that archaeal flavoprotein (AfpA, an archaeal homolog of DmrB) could be reduced by CO in the presence of CODH and ferredoxin. Both AfpA and DmrB lack iron-sulfur clusters, raising the possibility that the archaeal CO/CODH/ferredoxin system may act as an artificial electron donor system for His₆-DmrB in vitro. When 2 µM His₆-DmrB in 1 ml of anaerobic buffer was incubated with an atmosphere of CO, the FMN cofactor remained oxidized (Figure 4, line 1). However, when the reduction assay was initiated by adding 10 µl of partially purified CODH-ferredoxin, a sudden decrease of 0.105 absorbance unit was immediately observed (Figure 4, line 2). After 3 min, a further decrease of 0.05 absorbance unit occurred (Figure 4, line 3). The reaction appeared to be complete within 6 min since the 460-nm peak did not decrease substantially at later time points. His₆-DmrB remained reduced, and re-oxidation of the flavin was not observed within two hours of adding CODH/ferredoxin (data not shown). The decrease in absorbance at 460 nm is consistent with rapid electron transfer from CODH through ferredoxin to His₆-DmrB. Thus, the archaeal CODH-ferredoxin system from M. thermophila can be used successfully as an artificial electron donor for His₆-DmrB from the bacterium B. xenovorans.

When the assay was initiated by adding CODH and ferredoxin, the rate could not be measured accurately because the first decrease to 0.105 absorbance units occurred within 15 seconds of adding the catalyst. In the time it took to mix and record the first absorbance reading, the reduction was nearly two-thirds complete. A general approximation of a rate might be made by dividing the change in absorbance by the extinction coefficient and 0.25 min, which gives an estimate of about 34 µM FMN reduced/min. After 3 min, an additional decrease of 0.046 absorbance unit was observed with a corresponding rate of 1.22 µM/min.

A second flavoprotein, MM1853 from the methanogen Methanosarcina mazei, has homology to the Orf19 protein of M. extorquens. These proteins have been hypothesized to play a role in H₄MPT biosynthesis based on mutagenesis studies in M. extorquens showing that deletion of the orf19 gene abolishes the ability to synthesize H₄MPT [13]. Because the crystal structure of MM1853 (pdb 2NR4) contains one FMN per subunit, we predict that the enzyme functions as an oxidoreductase required for H₄MPT biosynthesis. However, this protein has not yet been characterized biochemically. In this work, we tested the ability of MM1853 to be reduced by the same electron donor candidates as for His₆-DmrB. The gene for MM1853 was overexpressed in E. coli cells, and after cell lysis and centrifugation, about 90% of the His₆-MM1853 appeared to be soluble, as shown by the distribution of the 23-kDa protein between the cell-free extract and insoluble pellet in SDS-PAGE (Figure 5, lanes 3 and 4). The molecular mass of the overproduced protein is similar to the predicted molecular mass of His₆-MM1853 based on the amino acid sequence (approximately 21.6 kDa) plus the mass of the six-histidine tag. The target protein His₆-MM1853 was purified by Ni-NTA chromatography. Although His₆-MM1853 was abundant in the fractions eluted with 100 mM imidazole, this fraction was contaminated with three additional proteins (Figure 5, lane 7). When His₆-MM1853 was eluted with 250 mM imidazole, a single 23-kDa band was observed on the SDS-PAGE gel (Figure 5, lane 8). The purified protein had a yellow color, consistent with the presence of FMN.

DTT (10 mM) reduced the FMN of His₆-MM1853 by about 20% within the first 5 min of adding reductant, indicated by a decrease in the absorbance peaks at 380 nm and 460 nm (Figure 6, trace 2). By 30 min, the FMN appeared completely reduced (Figure 6, trace 3). Therefore, in contrast to the result with His₆-DmrB (Figure 3), His₆-MM1853 was able to receive electrons from DTT. NADH and NADPH were also examined for the ability to donate electrons directly to the cofactor of His₆-MM1853. However, neither NADH nor NADPH

could directly transfer electrons to His₆-MM1853 without His₆-Fre added (Figure 7, line 1).

Because NADH alone was unable to reduce the FMN of His₆-MM1853, we tested whether His₆-Fre could facilitate electron transfer from NADH to His₆-MM1853. When the assay was initiated by adding His₆-Fre, a decrease of 0.135 absorbance unit was observed compared to the original absorbance at 380 nm, while the 460-nm peak was decreased completely to baseline level (Figure 7, line 2). For the 460-nm peak, the reaction appeared to be complete within 5 min since the absorbance at this wavelength did not decrease substantially at later time points (Figure 7, lines 3 and line 4). After 15 min, a further slight decrease of 0.007 absorbance unit was observed only at 380 nm (Figure 7, line 3), which might

correspond to NADH oxidation. The decrease in peaks corresponding to oxidized FMN is consistent with electron transfer from NADH through Fre to MM1853. Thus, the NADH/Fre system can be used as a convenient and effective electron donor system for MM1853.