Purification and characterization of ferredoxin-NADP reductase encoded by Bacillus subtilis yumC gene Authors:

From Bacillus subtilis cell extracts, ferredoxin-NADP reductase (FNR) was purified to homogeneity and found to be the yumC gene product by N-terminal amino acid sequencing. YumC is a ~94 kDa homodimeric protein with one molecule of non-covalently bound FAD per subunit. In a diaphorase assay with 2,6-dichlorophenol-indophenol as an electron acceptor, the affinity to NADPH was much higher than to NADH, with Km values of 0.57 vs. >200 μM. Kcat values of YumC with NADPH were 22.7 and 35.4 s in diaphorase and in a ferredoxin-dependent NADPH-cytochrome c reduction assay, respectively. The cell extracts contained another diaphorase-active enzyme, the yfkO gene product, but its affinity for ferredoxin was very low. The deduced YumC amino acid sequence has high identity to that of the recently identified Chlorobium tepidum FNR. A genomic database search indicated that there are more than 20 genes encoding proteins that share a high level of amino acid sequence identity with YumC and annotated variously as NADH oxidase, thioredoxin reductase, thioredoxin reductase-like protein, etc. These genes are found notably in Gram-positive bacteria except for Clostridia, and less frequently in archaea and proteobacteria. We propose that YumC and C. tepidum FNR constitute a new group of FNR which should be added to the already established plant type, bacteria type, and mitochondria type FNR groups. Introduction Ferredoxin (Fd) is a small acidic electron transport protein that contains iron-sulfur cluster(s). The iron-sulfur clusters bound in Fds are [4Fe-4S], [3Fe-4S], and [2Fe-2S] (Beinert et al. 1997). In some cyanobacteria and heterotrophic bacteria, flavodoxins are able to functionally replace Fd. Fds and flavodoxins participate in a variety of redox reaction pathways, including photosynthesis, nitrogen fixation, etc., and some in the reduction of certain cytochrome P450s (Ifuku et al. 1994; Birch et al. 1995; Grinberg et al. 2000). In vertebrate adrenocortical mitochondria, adrenodoxin, a specific Fd, is an electron donor to cytochrome P450. The enzymes that mediate the redox reactions between NAD(P)/NAD(P)H and Fd are Fd-NAD(P) reductase (FNR) (EC 1.18.1.2 and EC 1.18.1.3 with NADP and NAD as the substrate, respectively) (Ceccarelli et al. 2004). The catalytic electron transport function between NAD(P)H and an artificial electron acceptor such as methyl viologen and 2,6-dichlorophenol-indophenol (DPIP) is called diaphorase activity and is shown by FNRs. FNRs contain FAD as a prosthetic group and are present in chloroplasts, mitochondria, and some bacteria. FNRs are often divided into several groups. The first includes the plant-type FNRs that are monomeric enzymes also found in eukaryotic algae as well as in cyanobacteria, and have been intensively reviewed (Arakaki et al. 1997). The second group is sometimes called the bacteria type and is composed of monomeric FNRs from some bacteria such as Escherichia coli (Bianchi et al. 1993) and Azotobacter vinelandii (Isas et al. 1995). There are significant similarities between these two groups of FNRs in 3-D structure (Prasad et al. 1998) as well as in amino acid sequence (Arakaki et al. 1997, Ceccarelli et al. 2004). These two FNR groups can be united to a plant type in a broader sense. The third group is sometimes called the adrenodoxin reductase or mitochondria type FNR and includes enzymes from mammalian mitochondria (Lin et al. 1990), yeast mitochondria (Lacour et al. 1998), the bacterium Mycobacterium tuberculosis (Fischer et al. 2002), etc. This type of FNR catalyzes essentially the same reaction as do the plant and bacteria types and there are certain amino acid sequence similarities among them. However, X-ray crystallographic studies indicate that the 3-D structure of adrenodoxin reductase is more similar to those of glutathione reductase and disulfide reductase in a broader sense than to the plant-type FNRs (Ziegler et al. 1999). A bovine adrenodoxin reductase and its homolog from M. tuberculosis (FprA) exist as monomeric proteins (Fischer et al. 2002), while glutathione reductases exist as homodimer. With the entire genome sequences of many bacterial species now determined, several examples have been found, including Gram-positive bacteria and some archaea, whose genomes contain no genes coding for proteins with a high level of sequence identity to known FNRs. Recently, the occurrence of a novel type of FNR in the green sulfur bacterium Chlorobium tepidum was reported (Seo and Sakurai 2002). The enzyme shares much higher level of amino acid sequence identity with NAD(P)H-thioredoxin reductases than with plant-type, bacteria-type, or mitochondria-type FNRs. The C. tepidum enzyme is homodimeric and contains non-covalently bound FAD as do thioredoxin reductase, but has no di-cysteine motif (CXXC) essential for the catalytic function of the latter enzyme (Ronchi and Williams 1972). C. tepidum has another gene (CT0842, Eisen et al. 2002) that encodes a protein containing the di-cysteine motif which shares high levels of amino acid sequence identity with thioredoxin reductases from various organisms, and which appears to be a genuine thioredoxin reductase gene. BLAST homology searches (Altschul et al. 1997) with the deduced amino acid sequence of C. tepidum FNR identified a number of genes from Gram-positive bacteria and archaea that encode proteins which share a high level of amino acid sequence identity with that of C. tepidum. These genes are variously annotated as encoding thioredoxin reductases, thioredoxin reductase-related proteins, NADH oxidase, etc. The di-cysteine motif essential for thioredoxin reductase is also absent in all of these proteins. Bacillus subtilis is a gram-positive bacterium whose whole genome sequence has been elucidated (Kunst et al. 1997). Although the Fd was purified from E. coli cells overexpressing B. subtilis Fd gene (fer) and shown to support reduction of cytochrome P450 (BioI), the enzymes responsible for reduction of Fd have not yet been identified (Green et al. 2003). In this work, we show that the yumC gene of B. subtilis annotated as coding for a thioredoxin reductase-related protein (Kobayashi et al. 2003), is actually an FNR gene. To the best of our knowledge, this is the first report of an FNR from Gram-positive bacteria. Materials and methods Purification of ferredoxin-NADPH reductase from B. subtilis B. subtilis 168 CA (106309 Pasteur Culture Collection) stock culture was kindly provided by Drs. K. Kobayashi and N. Ogasawara of Nara Institute of Science and Technology. Cells were grown in Luria-Bertani medium for 18 h at 37C, collected by centrifugation at 10,000 × g for 20 min and stored at -80C until use. The cells were suspended in 50 mM 2-amino-2-hydroxymethyl-1,3-propanediol (Tris)-HCl buffer (pH 7.8) containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM n-caproic acid, and 1mM p-aminobenzamidine and disrupted at 0C using 6 1-min sonication cycles, each followed by a 1-min rest interval (VP-30s, TAITEC, Japan). The sonicate was centrifuged at 20,000 × g for 20 min at 4C, and the supernatant retained. The fraction obtained from the supernatant by precipitating between 35 % and 80 % saturation of ammonium sulfate at 4C was collected by centrifugation at 20,000 × g for 20 min, dissolved in 50 ml of 20 mM 2-morpholinoethanesulfonic acid (MES)-NaOH buffer (pH 6.5), and dialyzed three times against 5 l of the same buffer for 2 h each at 4°C. After removing undissolved proteins by centrifugation at 200,000 × g for 60 min, the supernatant was diluted with two volumes of the same buffer and applied to Matrex Blue A (Amicon), a dye ligand affinity gel column (2.8 × 10 cm) equilibrated with 20 mM MES-NaOH buffer (pH 6.5) at 4C. The column was washed with two column volumes of the equilibration medium and the bound proteins were eluted by a total of 432-ml linear gradient of 20 mM MES-NaOH (pH 6.5) buffer as a starting solution and 0.1 M glycine-NaOH (pH 11.5) buffer containing 1.5 M KCl as a feeding solution. Eluted fractions were assayed for NADPH-DPIP diaphorase activity. High activity fractions were pooled, concentrated by ultrafiltration (YM-10, Amicon) and applied to a gel-permeation column (Sephacryl S-200 HR 26/60, Pharmacia) equilibrated with 50 mM Tris-HCl buffer (pH 7.8) containing 150 mM NaCl at 4C. The proteins were eluted with the same buffer at a flow rate of 20 ml h. Fractions rich in the diaphorase activity were pooled, desalted by ultrafiltration (YM-10, Amicon) and applied to an anion exchange column (Mono Q 10/10, Pharmacia) equilibrated with 50 mM Tris-HCl (pH 7.8) at room temperature. After washing with two column volumes of the same buffer, the proteins were eluted from the column with a 70-ml linear gradient of NaCl concentration ranging from 0 to 350 mM in 50 mM Tris-HCl (pH 7.8) yielding two major peaks of diaphorase activity (Fig. 1). Peak fractions were pooled separately and purified as follows. The buffer was changed to 10 mM MES-NaOH (pH 6.5) by ultrafiltration (YM-10, Amicon), and the concentrate (2 ml) was applied to Matrex Red A (Amicon), a dye affinity column (1 × 10 cm) equilibrated with 20 mM MES-NaOH (pH 6.5). After washing with two column volumes of the same buffer, the activity was eluted as single-peak fractions, with a 24-ml linear gradient of 20 mM MES-NaOH (pH 6.5) buffer and 0.1 M glycine-NaOH (pH 11.5) buffer containing 1.5 M KCl. The buffer was changed to 50 mM Tris-HCl (pH 7.8) by ultrafiltration (Ultrafree 4, Millipore) and the concentrate was applied to an anion exchange column (Mono Q 5/5, Pharmacia) equilibrated with 50 mM Tris-HCl buffer (pH 7.8). The proteins were eluted with a 40-ml linear gradient of NaCl, from 0 to 400 mM in 50 mM Tris-HCl (pH 7.8) and stored at -80C until use. Preparation of ferredoxin from B. subtilis Fd was purified from the flow-through fraction from the Matrex Blue A column in FNR purification as described above. The purification procedures were essentially as described by Seo et al. (2001) for purification of Fds from C. tepidum. During the initial course of purification, Fd was detected by a cytochrome c reductase assay in the presence of spinach FNR that was prepared essentially according to Seo et al. (2001). Because crude preparations were contaminated with Fd-independent NAD(P)H-cytochrome c reductase activity, Fd-dependent cytochrome c reductase activity was estimated by subtracting the former activity obtained in the absence of spinach FNR. For purification, the flow-through fractions were applied to a diethylaminoethyl (DEAE)-cellulose column (2.6 × 15 cm, DE52, Whatman) equilibrated with 20 mM MES-NaOH buffer (pH 6.5), and after washing with two column volumes of the same buffer, Fd was eluted with a linear gradient of NaCl concentration from 0 to 1 M. Active fractions were concentrated by ultrafiltration (YM-3, Millipore) and applied to a gel-permeation column (Sephacryl S-200 26/60, Pharmacia). Active fractions were collected and a saturated ammonium sulfate solution in 50 mM Tris-HCl (pH 7.8) was added to 2 M, and after standing for 20 min centrifuged at 20,000 × g for 20 min. The supernatant was applied to a hydrophobic column (Phenyl Superose 5/5, Pharmacia) equilibrated with the same buffer containing 2 M ammonium sulfate, and Fd was eluted with 30 ml of a reverse linear gradient of ammonium sulfate, concentration from 2 M to 0.8 M. Fd-rich fractions were collected, desalted by ultrafiltration (YM-3, Amicon) and applied to a Mono Q 5/5 column equilibrated with 20 mM bis(2-hydroxyethyl)iminotris(hydroxymethyl)methane (BisTris)-HCl buffer (pH 6.5). Fd was eluted with a 22.5 ml linear gradient of NaCl, concentration from 0 to 0.6 M, in 20 mM BisTris-HCl buffer (pH 6.5), and the final preparation was stored at -80C until use. Fd concentration was estimated from ε390= 16.0 mM cm (Green et al. 2003). Enzyme activity assays In this report, the turnover rates are expressed in number of molecules of oxidized NAD(P)H or reduced NAD(P) by one molecule of native-form enzyme per second. NAD(P)H-DPIP diaphorase activity was assayed at 24C by monitoring the reduction of DPIP as the decrease of absorbance at 600 nm (ε600 = 21.8 mM cm) for 1 min. The reaction mixture (500 μl) contained 50 mM 2-[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES)-NaOH (pH 7.0), 0.1 mM DPIP, 10 mM glucose 6-phosphate (G7250, Sigma), 5 U of recombinant glucose-6-phosphate dehydrogenase (Leuconostoc mesenteroides, G8404, Sigma), NAD(P)H and diaphorase (YumC or YfkO, see Results) as indicated in figure legends and tables. The diaphorase activity was calculated by subtracting the blank rate obtained in the absence of YumC and YfkO from the observed rate. Cytochrome c reduction activity was assayed at 24C by monitoring the increase in absorbance at 550 nm for 1 or 3 min under aerobic conditions. The reaction mixture (500 μl) contained 50 mM HEPES-NaOH (pH 7.0), 0.1 mM cytochrome c (horse heart, ε550 = 27.8 mMcm, C2506, Sigma) and either 7.6 nM YumC or 40 nM YfkO with NADPH and Fd as indicated in figure legends and tables. NAD(P)H oxidase activity was assayed at 24C by monitoring the decrease of absorption at 340 nm (ε340 = 6.22 mMcm) for 3 min. The reaction mixture contained 100 mM potassium phosphate (pH 7.0), 0.15 mM NAD(P)H and either 111 pmol of YumC or 51 pmol of YfkO in the presence or absence of externally added flavin mononucleotide (FMN) or FAD as indicated in the tables. Photoreduction of Fd was assayed under strictly anoxic conditions by a heterologous assay system containing C. tepidum photochemical reaction center (3.84 μM bacteriochlorophyll a), 2 μM B. subtilis Fd, 95 nM YumC or 101 nM YfkO, 0.1 mM NADP, 20 mM Tris-HCl (pH 7.8), 20 mM NaCl, 5 mM sodium ascorbate, 0.1 mM DPIP, 0.1% Triton X-100, 5 mM D-glucose, 1.25 units glucose-oxidase, 5 × 10 -3 units catalase, and 0.25% (v/v) ethanol in a 1 ml mixture (Seo et al. 2001). Miscellaneous methods SDS-PAGE analysis was performed as described by Laemmli (1970), and protein bands in the gel were visualized by silver staining (Daiichi Kagaku, Tokyo). The native molecular mass was deduced by gel-permeation chromatography on Superdex 200 10/30 (Pharmacia) with a buffer containing 50 mM Tris-HCl (pH 7.8) and 200 mM NaCl at a flow rate of 0.5 ml min with Molecular weight marker kit MW-GF-200 (Sigma) for standards: cytochrome c (horse heart, 12.4 kDa), carbonic anhydrase (bovine erythrocytes, 29.0 kDa), albumin (bovine serum, 66.0 kDa), alcohol dehydrogenase (yeast, 150 kDa), beta-amylase (sweet potato, 200 kDa), and blue dextran (2,000 kDa). For determination of enzyme-bound flavins, the enzymes were treated with trichloroacetic acid (5%, w/v), and the flavins in the neutralized supernatant (pH 7.0) were determined by assuming ε450= 11.3 mM cm for FAD and ε445 = 12.5 mM cm for FMN. N-terminal amino acid sequence was analyzed by the Edman degradation method with Procise 491 or Procise cLC 494 sequencer (Applied Biosystems). Genome sequence data were obtained from the Swissprot (sp), trembl (tr), pir (pir), prf (prf) databases on the GenomeNet server (http://www.genome.ad.jp, Kyoto Univ). Sequence alignments were obtained using CLUSTAL W (Thompson et al. 1994) on the GenomeNet server (Kyoto Univ.). Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 2.1 (Kumar