Contribution of Cysteine Desulfurase (NifS Protein) to the Biotin Synthase Reaction of Escherichia coli

Abstract

The contribution of cysteine desulfurase, the NifS protein of Klebsiella pneumoniae and the IscS protein of Escherichia coli, to the biotin synthase reaction was investigated in in vitro and in vivo reaction systems with E. coli. When the nifS and nifU genes of K. pneumoniae were coexpressed in E. coli, NifS and NifU proteins in complex (NifU/S complex) and NifU monomer forms were observed. Both the NifU/S complex and the NifU monomer stimulated the biotin synthase reaction in the presence of l-cysteine in an in vitro reaction system. The NifU/S complex enhanced the production of biotin from dethiobiotin by the cells growing in an in vivo reaction system. Moreover, the IscS protein of E. coli stimulated the biotin synthase reaction in the presence of l-cysteine in the cell-free system. These results strongly suggest that cysteine desulfurase participates in the biotin synthase reaction, probably by supplying sulfur to the iron-sulfur cluster of biotin synthase.

The last step of biotin biosynthesis, namely the conversion of dethiobiotin (DTB) to biotin, involves the insertion of a sulfur atom between the inactive methyl and methylene carbon atoms adjacent to the imidazolinone ring of DTB. Biotin synthase (BioB protein) is involved in this reaction. The respective BioB proteins of both Escherichia coli and Bacillus sphaericus are homodimers, which contain one [2Fe-2S] cluster per monomer (12, 18). This conversion reaction requires S-adenosyl-l- methionine (AdoMet) and a physiological reduction system consisting of flavodoxin, ferredoxin-NADP⁺ reductase, and NADPH (14, 19). The direct sulfur donor for the reaction has remained unclear; however, there is a report that the sulfur atom of l-cysteine was incorporated into biotin with low efficiency in the cell-free systems of B. sphaericus (8). In addition, Tes Sum Bui et al. recently proposed that DTB was converted to biotin in the absence of the possible sulfur donor and that the sulfur of biotin was derived from the iron-sulfur cluster of BioB protein (21). From these results, we thought that the sulfur, which would be derived from l-cysteine by an unknown metabolism, is incorporated into the iron-sulfur cluster of BioB protein and then incorporated into biotin. On the other hand, an assembly mechanism for the iron-sulfur cluster in nitrogenase of the nitrogen-fixing bacterium, Azotobacter vinelandii, has been well studied (9, 23, 24). Cysteine desulfurase (NifS protein) of A. vinelandii catalyzes the formation of sulfur and l-alanine from l-cysteine and is involved in the mobilization of sulfur necessary for the nitrogenase metallocluster core formation. Although the specific role of the nifU gene product (NifU protein) containing a [2Fe-2S] cluster is not yet known, it might function either to deliver the iron and sulfur necessary for the cluster formation or to provide an intermediate site for the cluster assembly. In addition, the nif gene cluster containing the nifS and nifU genes was cloned from Klebsiella pneumoniae, which is a nitrogen-fixing bacterium, and characterized (1, 2).

In the work presented in this paper, we investigated the contribution of cysteine desulfurase to the biotin synthase reaction of E. coli by using the NifS and NifU proteins of K. pneumoniae. Furthermore, we also examined the participation of the E. coli IscS and IscU proteins, which were homologous to the NifS and NifU proteins, respectively (6, 7, 22), in the biotin synthase reaction.

MATERIALS AND METHODS

Chemicals.

5′-Deazariboflavin was kindly supplied by W. Simon (F. Hoffmann-La Roche, Ltd., Basel, Switzerland). Other chemicals were purchased from either Sigma Chemical Co., Aldrich Chemical Co., Pharmacia Biotech Co., Bio-Rad Laboratories, Wako Chemical Co., or Takara Shuzo Co. d-DTB was purchased from Sigma Chemical Co. and purified to remove any biotin. The purification was performed by reverse-phase high-pressure liquid chromatography on CAPCELL PAK C₁₈ (20 mm [internal diameter] by 25 cm; Shiseido Co.) with the mobile-phase solvent containing 12% (vol/vol) acetonitrile and 0.1% (vol/vol) trifluoroacetic acid.

Bacterial strains and plasmids.

An E. coli bioB-deficient mutant, R875 (4), and K. pneumoniae M5a1 (5) were kindly supplied by A. Campbell (Stanford University) and T. Uozumi (Tokyo University), respectively. Plasmid pTrc99A and pBluescript II SK(+) were purchased from Pharmacia Biotech Co. and Toyobo Co., respectively. pUC18 and pUC19 were purchased from Takara Shuzo Co.

Assay for protein concentration and SDS-PAGE.

Protein concentrations were measured by bicinchoninic acid protein assay kit (Pierce) with bovine serum albumin as a standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out as described by Laemmli (11), and proteins were stained with Coomassie Brilliant Blue R-250. A broad-range or low-range molecular weight standard (Bio-Rad Laboratories) was used as the molecular weight marker. The broad-range molecular weight standard contained myosin (200 kDa), β-galactosidase (116 kDa), phosphorylase B (97.4 kDa), serum albumin (66.2 kDa), ovalbumin (45.0 kDa), carbonic anhydrase (31.0 kDa), trypsin inhibitor (21.5 kDa), lysozyme (14.4 kDa), and aprotinin (6.5 kDa). The low-range molecular weight standard contained phosphorylase B, serum albumin, ovalbumin, carbonic anhydrase, trypsin inhibitor, and lysozyme.

Assay for the amount of formed biotin.

The amount of formed biotin was measured by a microbiological method with Lactobacillus plantarum (20).

Construction of an expression plasmid for E. coli bioB gene.

The bioB gene is located in 1.32 kb of the NcoI-HaeIII fragment on the chromosome in E. coli (16). We constructed a genomic library having a NcoI-HaeIII fragment of nearly 1.3 kb and selected a clone containing the bioB gene by doing a complementation test with an E. coli bioB-deficient mutant. The chromosomal DNA of E. coli HB101 was completely digested with NcoI and HaeIII, and the DNA fragments of 1.2 to 1.5 kb were isolated and inserted into NcoI and SmaI sites of the expression vector plasmid, pTrc99A. The resulting genomic library was transferred into the E. coli bioB-deficient mutant R875. Transformants were selected for biotin prototrophy to obtain a clone carrying the bioB gene. The hybrid plasmid having a 1.32-kb NcoI-HaeIII fragment containing the bioB gene was obtained and designated pTrcEB1 (see Fig. 5).

FIG. 5.

Structures of E. coli bioB and K. pneumoniae nifU and nifS gene expression plasmids pTrcEB1, pKNnif06, and pKNnif10.

Purification of BioB protein of E. coli.

E. coli JM109 having pTrcEB1 was cultivated in 2 liters of Terrific broth (17) containing 100 μg of ampicillin per ml at 30°C for 3 h. Isopropyl-β-d-thiogalactopyranoside (IPTG) was added at a 1 mM concentration to induce the expression of the bioB gene, and the cultivation was continued for another 3 h. Cells were collected by centrifugation at 8,000 × g for 20 min and washed with 20 mM Tris-HCl buffer (pH 8.1) containing 0.1 M NaCl. The cells were suspended in 60 ml of 20 mM Tris-HCl buffer (pH 8.1) containing 5 mM 2-mercaptoethanol (2-ME) and disrupted by French press in the presence of 0.25 mM phenylmethylsulfonyl fluoride (PMSF), 10 μg of deoxyribonuclease I per ml, and 10 μg of ribonuclease A per ml. Cell debris was removed by centrifugation at 15,000 × g for 30 min to obtain the cell-free extract. BioB protein was purified from the cell-free extract, with some modifications, as described by Sanyal et al. (18). During the purification steps, BioB protein was chased as a protein band (37 kDa) on SDS-PAGE gels. BioB protein purified to homogeneity was anaerobically incubated at room temperature for 1 h in 50 mM Tris-HCl buffer (pH 8.1) containing 1 mM dithiothreitol (DTT), 100 μM FeCl₃, and 50 μM Na₂S for holoformation of the iron-sulfur cluster of the enzyme. This mixture was passed through a gel filtration column (Sephadex G-25; Pharmacia Biotech Co.) to remove the excess iron and sulfide. After concentration by Centricon-30 (Amicon), the purified BioB protein was stored at −80°C until used. About 20 mg of BioB protein was obtained from 2 liters of cell broth.

Construction of an expression plasmid for K. pneumoniae nifU and nifS genes.

The K. pneumoniae nifU and nifS gene cluster was cloned by hybridization from the chromosomal DNA of K. pneumoniae M5a1. The chromosomal DNA was completely digested with BamHI, and 2.3- to 2.6-kb DNA fragments were obtained by agarose gel electrophoresis. The vector plasmid pUC19 was completely digested with BamHI and then treated with alkaline phosphatase to avoid self-ligation. The genomic DNA fragments prepared as described above were ligated with the cleaved pUC19, and the ligation mixture was transferred into E. coli JM109. The strains were selected for ampicillin resistance (100 μg/ml) on Luria-Bertani (LB) medium agar plate. Two thousand individual clones having the genomic DNA fragments were obtained as a genomic library. Selection of the clone having the nifU and nifS genes was carried out by colony hybridization. Two oligonucleotides having partial sequences of the nifU and nifS genes were synthesized based on the published sequences (2). The sequences were as follows: for the nifU probe, 5′-AGAGGAGCACGACGAGGGCAAGCTGATCTGCAAAT and for the nifS probe, 5′-CGTTGGTCAGCGTGATGTGGGCGAATAACGAAACC. A mixture of the labeled oligonucleotides was used as a probe for hybridization. Twenty-six candidates for the clone bearing the nifU and nifS genes were obtained. One of the obtained plasmids was analyzed using restriction enzymes, and 300 to 400 nucleotides of sequences at both ends of the inserted DNA fragments were confirmed. The determined sequences were identical to the nucleotide sequence of the K. pneumoniae nifU,S cluster published by Beynon et al. (2). The plasmid having the nifU,S cluster inserted in the opposite direction to the lac promoter was named pKNnif02.

The vector plasmid pBluescript II SK(+) was completely digested with HincII and BamHI. pKNnif02 was completely digested with VspI. The cleaved pKNnif02 was blunted and completely digested with BamHI. A 2.4-kb fragment containing the nifU,S cluster was isolated and inserted into the cleaved pBluescript II SK(+) to obtain pKNnif03. The vector plasmid pTrc99A was completely digested with KpnI and BamHI. A 2.4-kb KpnI-BamHI fragment containing the nifU,S cluster was isolated from pKNnif03 and ligated with the cleaved pTrc99A. The plasmid in which the nifU,S cluster was inserted at a site downstream of the trc promoter was finally obtained and named pKNnif04 (see Fig. 1A).

FIG. 1.

Expression of K. pneumoniae nifU and nifS genes in E. coli. (A) Structure of the expression plasmid of K. pneumoniae nifU and nifS genes. (B) SDS-PAGE analysis of the cell-free extracts. Lane 1, the cell-free extract of JM109 having pTrc99A (12 μg of protein); lane 2, the cell-free extract of JM109 having pKNnif04 (12 μg of protein); lane 3, molecular weight markers (low range).

Preparation of cell-free extract of E. coli with or without K. pneumoniae NifU and NifS proteins.

E. coli JM109 having pKNnif04 or pTrc99A was aerobically cultivated in Terrific broth containing 100 μg of ampicillin per ml at 30°C for 3 h. IPTG (1 mM) was added to induce gene expression, and the cultivation was continued for another 3 h. Cells were harvested by centrifugation at 8,000 × g for 20 min, washed once with 20 mM Tris-HCl buffer (pH 8.1) containing 0.1 M NaCl and 1 mM EDTA, and washed twice with the same buffer without 1 mM EDTA. The cells were suspended in about 2 volumes of 0.1 M Tris-HCl buffer (pH 7.5) containing 0.2 mM 2-ME against packed cell volume. The cell suspensions were degassed and purged by argon gas, and the cells were disrupted by sonication in a sealed tube with argon gas. After sonication, insoluble materials were removed by centrifugation at 100,000 × g for 30 min. The resulting supernatant was used as a cell-free extract. To confirm the expression of the nifU and nifS genes, the cell-free extracts were subjected to SDS-PAGE (Fig. 1B). The cell-free extracts were stored at −80°C in sealed tubes with argon gas until used.

Purification of K. pneumoniae NifU and NifS proteins.

E. coli JM109 having pKNnif04 was aerobically cultivated in 1 liter of Terrific broth containing 100 μg of ampicillin per ml at 26°C for 3 h. IPTG (1 mM) was added to induce expression of the nifU and nifS genes, and the cultivation was continued for another 3 h. Cells (5.4 g [wet weight]) were harvested by centrifugation at 8,000 × g for 20 min, washed once with 20 mM Tris-HCl buffer (pH 7.4) containing 0.1 M NaCl and 1 mM EDTA, and washed twice with the same buffer without 1 mM EDTA. A 5.4-g (wet weight) quantity of cells was obtained and stocked at −80°C until used.

All the column operations were anaerobically performed at room temperature, and other operations were anaerobically performed at 4 to 10°C unless otherwise stated. The NifU and NifS proteins were chased as protein bands on SDS-PAGE gels. The cells were thawed with about 40 ml of 20 mM Tris-HCl buffer (pH 7.4) containing 5 mM DTT and disrupted by French press in the presence of 0.5 mM PMSF, 10 μg of deoxyribonuclease I per ml, 10 μg of ribonuclease A per ml, and 5 mM pyridoxal phosphate. The cell debris was removed by centrifugation at 7,700 × g for 30 min, and the insoluble fraction was removed by centrifugation at 48,000 × g for 30 min. The soluble fraction was filled up to 50 ml with the same buffer, and streptomycin sulfate was added at a final concentration of 1% (wt/vol). The insoluble residue was removed by centrifugation at 48,000 × g for 20 min, and solid ammonium sulfate was added to the supernatant at 30% saturation. After gentle stirring at room temperature for 10 min, the precipitate containing NifU and NifS proteins was obtained by centrifugation at 48,000 × g for 10 min. The precipitate was resuspended in 20 mM Tris-HCl buffer (pH 7.4) containing 5 mM DTT and centrifuged at 48,000 × g for 30 min. The supernatant was loaded on RESOURCE Q (6 ml; Pharmacia Biotech Co.) equilibrated with 20 mM Tris-HCl buffer (pH 7.4) containing 5 mM DTT. After washing with the same buffer, elution was done with 150 ml of 0 to 0.5 M NaCl linear gradient. The NifU and NifS proteins, which coeluted in 20 ml of fraction around 0.3 M NaCl, were collected and concentrated to 3.5 ml by PM-30 (Amicon). The concentrated protein solution was passed through HiPrep Sephacryl S-200HR 26/60 (Pharmacia Biotech) with 20 mM Tris-HCl buffer (pH 7.4) containing 5 mM DTT and 0.25 M NaCl. All NifS protein was recovered as a protein complex with NifU protein, and a portion of NifU protein was recovered as a monomer. The NifU/S complex was eluted at a molecular mass position of about 140 kDa, and the NifU monomer was eluted at a molecular mass position of about 35 kDa. The fraction (18 ml) containing the NifU/S complex (approximately 6 mg of protein) and the fraction (24 ml) containing the NifU monomer (approximately 4.5 mg of protein) were concentrated to 3 ml by CentriPlus-30 (Amicon) and stored at −80°C.

Effects of K. pneumoniae NifS and NifU proteins on biotin synthase reaction in DAF photoreduction system.

Biotin synthase activity was assayed by the amount of biotin formed from DTB in the 5′-deazariboflavin (DAF) photoreduction system. The basal reaction mixture contained 100 μM DTB, 50 μM DAF, 1 mM AdoMet, 200 μM l-cysteine, and 16 μM BioB protein in 100 mM Tris-HCl buffer (pH 7.5). Mixtures of the cell-free extracts prepared from E. coli JM109 having pTrc99A and E. coli JM109 having pKNnif04 at various ratios were added to the basal reaction mixture at the final concentration of 20 mg of protein/ml. In another experiment, the purified NifU/S complex and/or NifU monomer was added to the basal reaction mixture at a final concentration of 13 and/or 30 μM, respectively, in the presence of 10 mM DTT. The reaction mixtures were degassed and purged with argon to make anaerobic conditions for stabilization of the photoreduced DAF. The reactions were performed at 30°C for 80 min with irradiation with a 300-W halogen bulb and stopped by boiling for 5 min. After the reaction mixture was centrifuged, the supernatant was assayed for formed biotin. The amount of biotin produced by the enzymatic reaction was determined by the differential between results from the reacted and nonreacted mixtures.

Construction of expression plasmids for the E. coli fldA and fpr genes for flavodoxin and ferredoxin-NADP⁺ reductase, respectively.

The nucleotide sequences of the fldA and fpr genes have been reported by Osborne et al. (15) and Bianchi et al. (3), respectively. We isolated DNA fragments having the fldA and fpr genes by using PCR. Two pairs of primers, fldA-1 and fldA-2 for the fldA gene and fpr-1 and fpr-2 for the fpr gene, were synthesized based on the published sequences. The sequences of the primers were as follows: fldA-1, 5′-GGCACCATGGCTATCACTGGCATC; fldA-2, 5′-CCGGCTGCAGTGAGTCTACGCCGC; fpr-1, 5′-GGCCACCATGGCTGATTGGGTAAC; and fpr-2, 5′-AGCTGGATCCCGTGCCGTTTATCG. PCRs were carried out by using E. coli HB101 chromosomal DNA as a template. Amplified DNA fragments were cloned into pUC18, and the nucleotide sequences of the inserted DNA fragments were confirmed. Then a 0.56-kb NcoI-PstI fragment containing the fldA gene was cloned into the NcoI and PstI sites of the expression vector pTrc99A, and a 0.78-kb NcoI-BamHI fragment containing the fpr gene was cloned into the NcoI and BamHI sites of pTrc99A. The constructed plasmids were named pTrcfldA and pTrcfpr, respectively.

Purification of E. coli flavodoxin and ferredoxin-NADP⁺ reductase.

Flavodoxin and ferredoxin-NADP⁺ reductase were isolated from the cell-free extracts of E. coli JM109 having pTrcfldA and E. coli JM109 having pTrcfpr, respectively. E. coli JM109 having pTrcfldA or pTrcfpr was cultured at 37°C in 1 liter of LB medium containing 100 μg of ampicillin per ml. After 2.5 h of cultivation, 2 mM IPTG was added, and the cultivation was continued for another 4 h. The cells were harvested and washed twice with saline (0.85% NaCl). The cells were suspended in 0.1 M Tris-HCl buffer (pH 7.5) containing 2 mM DTT, 0.2 mM PMSF, 10 μg of deoxyribonuclease I per ml, and 10 μg of ribonuclease A per ml. Then 50 μM flavin mononucleotide or 50 μM flavin adenine dinucleotide was added to the cell suspensions of E. coli JM109 having pTrcfldA or E. coli JM109 having pTrcfpr, respectively, and the cells were disrupted by French press. Cell debris was removed by centrifugation at 15,000 × g for 30 min to obtain a cell-free extract. Purification of flavodoxin and ferredoxin-NADP⁺ reductase from each cell-free extract was carried out by modified methods as described by Sanyal et al. (19). Finally, about 50 mg of flavodoxin protein and about 5 mg of ferredoxin-NADP⁺ reductase protein were obtained. SDS-PAGE analysis showed that both the flavodoxin and ferredoxin-NADP⁺ reductase were more than 90% pure.

Effects of K. pneumoniae NifS and NifU proteins on biotin synthase reaction in physiological reduction system.

The basal reaction mixture contained 16 μM BioB protein, 3 μM ferredoxin-NADP⁺ reductase, 50 μM flavodoxin, 100 μM DTB, 1 mM NADPH, 1 mM AdoMet, and 10 mM DTT in 100 mM Tris-HCl buffer (pH 7.5). To the basal reaction mixture 200 μM FeCl₃, 100 μM Na₂S, 200 μM l-cysteine, 15 μM NifU/S complex, and/or 40 μM NifU monomer were added as additional factors. The reaction was anaerobically carried out at 30°C for 3 h and stopped by boiling for 5 min. After centrifugation of the reaction mixture, the supernatant was assayed for formed biotin. The amount of biotin produced by the enzymatic reaction was determined by the differential between the results from the reacted and nonreacted mixtures.

Effects of NifU and NifS proteins on the biotin synthase reaction in the growing cell system.

To coexpress the E. coli bioB gene and the K. pneumoniae nifU and nifS genes in E. coli, we constructed two hybrid plasmids, pKNnif06 and pKNnif10. The nifU,S cluster or the nifU gene was inserted at a site downstream of the bioB gene in pTrcEB1. First, a 2.4-kb VspI-BamHI fragment containing the nifU,S cluster was isolated from pKNnif02. The VspI site of the fragment was changed to the BamHI site with BamHI linker to obtain a BamHI cassette having the nifU,S cluster. The cassette was inserted into the cleaved pTrcEB1 with BamHI. Finally, pKNnif06 in which the bioB, nifU, and nifS genes were expressed under the control of the trc promoter was constructed (see Fig. 5). Next, a HpaI-BamHI fragment containing the nifS gene was deleted from pKNnif06 to construct pKNnif10 (see Fig. 5). E. coli JM109 was transformed with pTrc99A, pKNnif06, or pKNnif10.

E. coli JM109 having pTrc99A, pTrcEB1, pKNnif06, or pKNnif10 was cultivated with 50 ml of PC medium (2% glycerol, 5% proteose peptone, 2% casamino acid, 1% K₂HPO₄, 0.05% KCl, 0.05% MgSO₄ · 7H₂O, 0.001% MnSO₄ · 4-6H₂O, and 0.001% FeSO₄ · 7H₂O; pH 7.0) containing 200 μg of DTB per ml and 100 μg of ampicillin per ml in a 500-ml flask. The cultivation was carried out aerobically at 30°C for 3 h on a rotary shaker (at 100 rpm). IPTG was added to induce the expression of the bioB, nifU, and nifS genes at a concentration of 1 mM, and the cultivation was continued for another 27 h. After cultivation, 1.5 ml of the culture broth was centrifuged to remove cells, and the supernatant was assayed for formed biotin.

To confirm the expression of the bioB, nifU, and nifS genes, 1 ml of culture broth was sampled from each flask 3 h after the induction, and cells were collected by centrifugation. The cells were suspended in 0.1 M Tris-HCl buffer (pH 7.5) with 2 mM DTT and disrupted by sonication. The insoluble fractions were removed by centrifugation at 100,000 × g for 30 min, and the soluble fractions were subjected to SDS-PAGE (data not shown).

Construction of expression plasmids for the E. coli iscS and iscU genes.

The E. coli iscS and iscU gene cluster was cloned by PCR from the chromosomal DNA of E. coli HB101. Two PCR primers, iscSU-1 and iscSU-2, were synthesized based on the E. coli genome sequence database (EMBL accession number AE000339). The sequences of the primers were as follows: for iscSU-1, 5′-ACGCGATCGACGTTAAGTTACG and for iscSU-2, 5′-ACCCTTTACCGCGGTTAGCCAG. Amplified DNA fragments were cloned into pUC18, and the nucleotide sequences of the inserted DNA fragments were confirmed. The plasmid having the iscS,U cluster inserted in the same direction as the lac promoter in the vector was named pECisc01. The pECisc01 was blunted after digestion with EcoRI, and a 1.8-kb fragment having the iscS,U cluster was obtained by digestion with BamHI. The vector plasmid, pTrc99A, was digested with NcoI and blunted. After digestion with BamHI, the cleaved pTrc99A was ligated with the fragment having the iscS,U cluster. The plasmid in which the iscS,U cluster was inserted at a locus downstream of the trc promoter was finally obtained and named pECisc05 (see Fig. 6A). A 1.2-kb fragment having the iscS gene was obtained from pECisc01 by PCR with the iscS-1 and iscS-2 primers. The sequences of the primers were as follows: for iscS-1, 5′-GATCTCTAGATGAGTGATGTACGGAG and for iscS-2, 5′-GATCCTGCAGTCCTGATTCCGATACC. An amplified DNA fragment was digested with XbaI and PstI and cloned into pUC18 to construct pECisc02. The pECisc02 was digested with BamHI and blunted. Then, a 1.2-kb fragment having the iscS gene was obtained by digestion with PstI. The pTrc99A blunted as described above was digested with PstI and ligated with the fragment having the iscS gene. Finally, pECisc06, in which the iscS gene was inserted downstream of the trc promoter, was obtained (see Fig. 6A). E. coli JM109 was transformed with pECisc05 or pECisc06.

FIG. 6.

Expression of the E. coli iscS and iscU genes in E. coli. (A) Structures of expression plasmids pECisc05 and pECisc06. (B) SDS-PAGE of the soluble fractions prepared from cells of the recombinant strains. Lanes 2 and 6: molecular weight markers (broad range). Lane 1, JM109 having pTrcEB1; lane 3, JM109 having pTrc99A; lane 4, JM109 having pECisc05; lane 5, JM109 having pECisc06. Protein (12 μg) of the soluble faction was subjected to SDS-PAGE.

Effects of E. coli IscS and IscU proteins on biotin synthase reaction in the cell-free system.

E. coli JM109 having pTrcEB1 was grown in FM37 medium (2% glycerol, 10% corn steep liquor, 0.2% proteose peptone, 0.2% yeast extract, 1% K₂HPO₄, 0.05% KCl, 0.05% MgSO₄ · 7H₂O, 0.001% MnSO₄ · 4-6H₂O, and 0.01% FeSO₄ · 7H₂O; pH 7.0) containing 100 μg of ampicillin per ml. E. coli JM109 having pTrc99A, pECisc05, or pECisc06 was grown in Terrific broth containing 100 μg of ampicillin per ml. The cultivation was aerobically carried out at 37°C, and IPTG was added at 1 mM for E. coli JM109 having pTrcEB1 or at 0.2 mM for other strains after 2 h of cultivation. The cultivation was continued for another 4 h. Cells were harvested by centrifugation at 8,000 × g for 20 min and washed with 20 mM Tris-HCl buffer (pH 7.5) containing 0.1 M NaCl three times. Cells were suspended in about 2 volumes of 0.1 M Tris-HCl buffer (pH 7.5) containing 2 mM DTT and 0.2 mM PMSF against the packed cell volume. The cell suspensions were degassed and purged by argon gas, and the cells were disrupted by sonication in a sealed tube with argon gas. Cell debris was removed by centrifugation at 15,000 × g for 30 min, and the supernatants were used as cell-free extracts. To confirm the expression of the genes, the cell-free extracts were centrifuged at 100,000 × g for 30 min, and the supernatants were subjected to SDS-PAGE (see Fig. 6B). The cell-free extracts were stored at −80°C in sealed tubes with argon gas until used.

The cell-free extract prepared from E. coli JM109 having pTrcEB1 was used for the conversion of DTB to biotin as BioB protein. The basal reaction mixture contained the cell-free extract (final concentration, 16.7 mg of protein/ml) of E. coli JM109 having pTrcEB1, 100 μM DTB, 1 mM NADPH, 1 mM AdoMet, 5 mM DTT, 2 mM l-cysteine, 200 μM FeCl₃, and 100 mM Tris-HCl buffer (pH 7.5). Mixtures of the cell-free extract of E. coli JM109 having pTrc99A and the cell-free extracts of E. coli JM109 having pECisc05 or pECisc06 at various ratios were added to the basal reaction mixture at a final concentration of 4.2 mg of protein/ml. The reaction was anaerobically performed at 37°C for 3 h and stopped by boiling for 5 min. After centrifugation of the reaction mixture, the supernatant was assayed for formed biotin. The amount of biotin produced by the enzymatic reaction was determined by the differential between the results from the reacted and nonreacted mixtures.

RESULTS

Expression of the K. pneumoniae nifU and nifS genes in E. coli.

We constructed an expression plasmid, pKNnif04 (Fig. 1A) to overproduce the K. pneumoniae NifU and NifS proteins in E. coli. E. coli JM109 having pKNnif04 was grown at 30 or 37°C, and expression of the nifU and nifS genes was induced by the addition of IPTG. To confirm production of the NifU and NifS proteins, whole-cell proteins and soluble fractions prepared from the collected cells were subjected to SDS-PAGE analysis. In cells grown at both 37 and 30°C, overproduction of the NifU (30-kDa) and NifS (43-kDa) proteins was observed on SDS-PAGE gels, but the amounts of NifU and NifS protein in the soluble fraction were small compared with those in whole-cell proteins (data not shown). This result indicated that the NifU and NifS proteins produced were insoluble in E. coli. Since more of the soluble NifU and NifS proteins were recovered from the E. coli cells grown at 30°C than at 37°C, we cultured cells at temperatures lower than 30°C for the expression of the nifU and nifS genes thereafter.

Effect of the cell-free extracts having K. pneumoniae NifU and NifS proteins on biotin synthase reaction in the DAF photoreduction system.

The effects of NifU and NifS proteins on the biotin synthase reaction were subjected to preliminary examination in the DAF photoreduction system by using BioB protein and cell-free extracts having NifU and NifS proteins as described in Materials and Methods. The cell-free extracts (Fig. 1B) prepared from E. coli JM109 having pKNnif04 and E. coli JM109 having pTrc99A (vector control) were mixed at various ratios and added to the reaction mixture. The addition of the cell-free extract having NifU and NifS proteins enhanced the activity, but the enhancing effect was saturated at about a 1.7-fold increase, as shown in Table 1.

TABLE 1.

Effect of K. pneumoniae NifS and NifU proteins on biotin synthase activity in the DAF photoreduction system

Cell-free extracts ratioa	Biotin formed (ng/ml)	Relative index (%)
100:0	532.8	100
84:16	688.3	129.2
68:32	829.8	155.7
36:64	927.3	174.0
0:100	811.3	152.3

^aThe basal reaction mixture was described in Materials and Methods. Values shown are the ratios of cell-free extract prepared from JM109 having pTrc99A (vector control) to cell-free extract prepared from JM109 having pKNnif04 (NifS and NifU proteins) that were added to the basal reaction mixture. The reactions were anaerobically carried out at 30°C for 80 min. 

Purification of K. pneumoniae NifU and NifS proteins.

We next purified NifU and NifS proteins from the cell-free extract of E. coli JM109 having pKNnif04. All of the NifU and NifS proteins behaved together through the ammonium sulfate fractionation and the ion-exchange column chromatography. In the gel filtration (the final step of the purification), a nearly purified NifU and NifS mixture was separated into two fractions as shown in Fig. 2. All NifS protein was coeluted with equivalent NifU protein at a molecular mass position of about 140 kDa (NifU/S complex), and NifU protein alone was eluted at a molecular mass position of about 35 kDa (NifU monomer). Both fractions were colored a typical iron-sulfur red. The results indicated that NifS protein possibly existed as a heterotetramer complex with NifU protein (might be NifU₂NifS₂ form) in E. coli. On the other hand, NifU protein seemed to exist in both monomeric form and complexed form with NifS protein in E. coli. Moreover, the purified NifU/S complex and the NifU monomer both showed the typical iron-sulfur-red color. The result indicated that the NifU protein produced in E. coli contains the iron-sulfur cluster, as does the A. vinelandii NifU protein (9).

FIG. 2.

Gel filtration of K. pneumoniae NifU and NifS proteins. (A) Gel filtration pattern. Molecular weight markers consisted of thyroglobulin (670 kDa), bovine gamma globulin (158 kDa), chicken ovalbumin (44 kDa), equine myoglobin (17 kDa), and vitamin B12 (1.4 kDa). (B) SDS-PAGE analysis of fractions of gel filtration. NifU/S∗ and NifU∗∗ indicate NifU/S complex and NifU monomer, respectively.

Effects of the NifU/S complex of K. pneumoniae on the biotin synthase reaction in the DAF photoreduction system.

We examined the effects of the purified NifU/S complex and the NifU monomer on the biotin synthase reaction in the DAF photoreduction system by using BioB protein as described in Materials and Methods. As shown in Table 2, the addition of 13 μM NifU/S complex or 30 μM NifU monomer resulted in about fourfold higher activity. In addition, an additive effect of NifU/S complex and NifU monomer was observed.

TABLE 2.

Effect of K. pneumoniae NifU/S complex and NifU monomer on biotin synthase activity in the DAF photoreduction system^a

Additional proteins		Biotin formed (ng/ml)	Relative index (%)
NifU/S complex	NifU monomer
−	−	54.15	100
+	−	202.75	374
−	+	203.51	376
+	+	453.83	831

^aThe basal reaction mixture was described in Materials and Methods. NifU/S complex (13 μM) and/or NifU monomer (30 μM) was added to the basal reaction mixture. The reactions were anaerobically carried out at 30°C for 80 min. 

Effects of several supplements, iron ion, sulfur ion, and l-cysteine, on the biotin synthase reaction were examined in the DAF photoreduction system by using BioB protein in the presence of NifU/S complex as shown in Fig. 3. Addition of iron and sulfur ions enhanced the activity up to about 4.5-fold. But a similar positive effect of iron and sulfur ions was also observed in the absence of the NifU/S complex (data not shown). On the other hand, the highest activity was shown in the presence of iron ion and l-cysteine, and the activity was enhanced to about 5.3-fold. This result suggested that the NifU/S complex completely replaced the positive effect of sulfur ion in the presence of l-cysteine and that the NifU/S complex contributed to the biotin synthase reaction as a supplier of sulfide from l-cysteine.

FIG. 3.

Effects of l-cysteine, iron, and sulfur ions on the stimulation by K. pneumoniae NifU/S complex in the DAF photoreduction system. The basal reaction mixture contained 100 μM DTB, 50 μM DAF, 1 mM AdoMet, 10 mM DTT, 16 μM BioB protein, and 13 μM NifU/S complex in 100 mM Tris-HCl buffer (pH 7.5). Either 200 μM FeCl₃ (Fe³⁺), 100 μM Na₂S (S²⁻), or 200 μM l-cysteine (Cys) or a combination of them was added to the basal reaction mixture as an additional factor. The reaction was carried out at 30°C for 80 min as described in Materials and Methods.

Effects of NifU/S complex and NifU monomer of K. pneumoniae on the biotin synthase reaction in the physiological reduction system.

We purified flavodoxin and ferredoxin-NADP⁺ reductase from E. coli as a physiological reduction system, as described in Materials and Methods, and examined effects of the NifU/S complex and NifU monomer on the biotin synthase reaction in the natural enzyme system consisting of BioB protein, flavodoxin, and ferredoxin-NADP⁺ reductase as shown in Fig. 4. First, we confirmed the effect of iron and sulfur ions on the biotin synthase reaction in the physiological reduction system. Iron and sulfur ions stimulated the reaction by about ninefold, and this response to the addition of iron and sulfur ions was quite similar to that in the DAF photoreduction system. The addition of NifU/S complex and l-cysteine stimulated the reaction more than did the combination of iron and sulfur ions. Iron and sulfur ions did not further enhance the activity in the presence of NifU/S complex and l-cysteine (data not shown). Furthermore, NifU monomer also enhanced the activity in the presence of l-cysteine, but an additive effect of NifU/S complex and NifU monomer was not observed.

FIG. 4.

Effects of the K. pneumoniae NifU/S complex and NifU monomer on the biotin synthase reaction in the physiological reduction system. The reaction was anaerobically carried out at 30°C for 3 h as described in Materials and Methods. □, basal reaction mixture; ■, mixture with FeCl₃ and Na₂S; ○, mixture with NifU/S complex and l-cysteine; ▵, mixture with NifU monomer and l-cysteine; ●, mixture with NifU/S complex, NifU monomer, and l-cysteine.

Effects of NifU and NifS proteins on the biotin synthase reaction in the growing cell system.

We evaluated the effect of NifU and NifS proteins on the biotin synthase reaction in the E. coli growing cell system. For coexpression of the E. coli bioB gene and K. pneumoniae nifU and nifS genes in E. coli, we constructed pKNnif06 having the bioB, nifU, and nifS genes and pKNnif10 having the bioB and nifU genes (Fig. 5) and introduced them each into E. coli JM109. Since NifS protein was completely insoluble in E. coli in the absence of NifU protein, we could not construct a coexpression system of the bioB and nifS genes for the evaluation in the growing cell system.

E. coli recombinant strains were cultivated in PC medium containing 200 μg of DTB per ml at 30°C for 30 h as described in Materials and Methods. To confirm the expression of the genes from plasmids, soluble fractions were prepared from the recombinant cells induced by the addition of IPTG and subjected to SDS-PAGE (data not shown). BioB, NifU, and NifS proteins were coproduced in E. coli JM109 having pKNnif06. On the other hand, only BioB and NifU proteins were coproduced in E. coli JM109 having pKNnif10. The expression level of BioB protein in these recombinant strains was nearly equal to that in E. coli JM109 having pTrcEB1.

The productivity of biotin from DTB of the recombinant strains is shown in Table 3. E. coli JM109 having pKNnif06 produced 2.3-fold biotin compared with the control strain, E. coli JM109 having pTrcEB1; however, the productivity of E. coli JM109 having pKNnif10 showed that only NifU protein was not effective on the biotin production from DTB. In conclusion, NifU and NifS proteins together remarkably stimulated the biotin production from DTB in the growing cell system.

TABLE 3.

Effect of K. pneumoniae NifS and NifU proteins on the conversion activity in the growing cell system^a

Recombinant strain	Protein(s) expressed from plasmid(s)	Biotin production (μg/ml)
JM109(pTrc99A)	None	0.02
JM109(pTrcEB1)	BioB	2.24
JM109(pKNnif06)	BioB, NifS, NifU	5.29
JM109(pKNnif10)	BioB, NifU	2.09

^aThe cultivation was carried out at 30°C for 30 h, as described in Materials and Methods. 

Effects of E. coli IscS and IscU proteins on the biotin synthase reaction in the cell-free system.

E. coli iscS and iscU genes coding for homologs of NifS and NifU proteins, respectively, were cloned from the chromosome, and the expression plasmids pECisc05 and pECisc06 (Fig. 6A) were constructed. We confirmed overproduction of the IscS and IscU proteins in E. coli JM109 having pECisc05 or pECisc06 by SDS-PAGE. As shown in Fig. 6B, IscS protein was overproduced in both recombinant strains, and the expression level was nearly equal for both strains. In addition, IscU protein was also overproduced in E. coli JM109 having pECisc05.

We used the cell-free extract prepared from E. coli JM109 having pTrcEB1 instead of BioB protein and examined the effects of the E. coli IscS and IscU proteins on the biotin synthase reaction in the cell-free system. The cell-free extract containing the Isc protein(s) was prepared from E. coli JM109 having pECisc05 or pECisc06 and added to the reaction mixture together with the cell-free extract of E. coli JM109 having pTrc99A (vector cell-free extract) at various ratios as described in Materials and Methods. As shown in Table 4, the biotin synthase activity was enhanced by the addition of the cell-free extracts with IscS protein to about 1.5-fold under anaerobic conditions. The enhancing effect of the IscS protein under aerobic conditions was lower than that under anaerobic conditions (data not shown). This result suggested that the E. coli IscS protein contributed to the biotin synthase reaction, probably by supplying sulfur to the BioB protein in the same way that the K. pneumoniae NifS protein did; however, there was only a little additional enhancement of the biotin formation by IscU protein.

TABLE 4.

Effect of E. coli IscS and IscU proteins on biotin synthase activity in the cell-free system

isc plasmid (isc gene)	Additional cell-free extract (mg/ml)		Formed biotin (μg/ml)
	Isca	Vectorb
Control	0	4.2	8.74
pECisc05 (iscS, iscU)	0.4	3.8	11.8
	2.1	2.1	12.4
	4.2	0	14.5
pECisc06 (iscS)	0.4	3.8	11.4
	2.1	2.1	11.6
	4.2	0	13.3

^aIsc cell-free extract was prepared from JM109 having pECisc05 or pECisc06. 

^bVector cell-free extract was prepared from JM109 having pTrc99A. The basal reaction mixture was described in Materials and Methods. The Isc cell-free extract and the vector cell-free extract were added to the basal reaction mixture at various ratios. The reactions were anaerobically carried out at 37°C for 3 h. 

DISCUSSION

In this study, we first examined the effect of the NifS and NifU proteins of K. pneumoniae on the biotin synthase reaction of E. coli by using the cell-free extract prepared from E. coli cells in which NifS and NifU proteins were overproduced. The biotin synthase reaction was remarkably stimulated by the addition of the cell-free extract in the presence of l-cysteine. The stimulation was correlated with the amount of the added cell-free extract, but a saturation of the stimulation was observed at about 1.7-fold stimulation.

The produced NifS protein formed a complex with the NifU protein (NifU/S complex) in E. coli. NifU protein existed in a monomer form (NifU monomer) in addition to existing in the NifU/S complex. We purified NifU/S complex and NifU monomer from the cell-free extract of E. coli cells in which the nifS and nifU genes were overexpressed. The effects of NifU/S complex and NifU monomer on the biotin synthase reaction were examined in the pure enzyme system with the purified BioB protein. In the pure enzyme system by using the artificial reduction system with DAF, NifU/S complex stimulated the biotin synthase reaction to about 5.3-fold in the presence of iron ion and l-cysteine. Moreover, we examined the effects of NifU/S complex and NifU monomer on the biotin synthase reaction in the pure enzyme system, but with the physiological reduction system consisting of flavodoxin, ferredoxin-NADP⁺ reductase, and NADPH. The addition of NifU/S complex and l-cysteine stimulated the reaction more than did the addition of iron and sulfur ions. Although NifU monomer alone enhanced the activity, the additive effect of NifU/S complex and NifU monomer was not observed in the presence of l-cysteine. These results indicated that NifU/S complex is effective in the in vitro biotin synthase reaction. NifU monomer is also effective in the in vitro reaction, but the additive effect of NifU/S complex and NifU monomer seems to be dependent on the reducing potential. Although the biotin synthase reaction was stimulated by NifU/S complex with l-cysteine as well as by iron and sulfur ions, the reaction had nearly stopped after 1 h of incubation.

To evaluate the effects of the NifS and NifU proteins on biotin biosynthesis in the growing cell system, we constructed systems for the coexpression of the K. pneumoniae nifS and nifU genes with the E. coli bioB gene. NifU/S complex and NifU monomer remarkably enhanced the biotin productivity from DTB in the growing cell system. However, the positive effect of NifU monomer alone was not observed in the growing cell system. The result suggested that the enhancing effect of the NifU monomer observed in the in vitro reaction system might be an artifact and not a physiological effect. The reaction efficiency of BioB protein depended on the BioB concentration itself in the pure enzyme system using the BioB protein of Bacillus subtilis or E. coli (our unpublished result). Therefore, the iron-sulfur cluster of NifU monomer might affect the reaction as a stabilizer of the iron-sulfur cluster of BioB protein.

Experiments with the NifS and NifU proteins of K. pneumoniae indicated that NifU/S complex is effective in the presence of l-cysteine on the biotin synthase reaction of E. coli in both in vitro and in vivo reaction systems. We could not evaluate the effect of NifS protein alone, because NifS protein was not produced as a soluble protein in E. coli. However, the stimulation by NifU/S complex is estimated to be due to a cysteine desulfurase activity of NifS protein.

Recently, it has been reported that some non-nitrogen-fixing bacteria, including E. coli, have proteins homologous to the NifS protein (13, 22) and that the IscS protein, which is an NifS homolog of E. coli, seems to be involved in assembly of the iron-sulfur cluster of dihydroxy-acid dehydratase (6, 7). We examined the contribution of the IscS protein to the biotin synthase reaction in the cell-free system. The biotin synthase activity was enhanced by the addition of IscS protein in the presence of l-cysteine, and the enhancement was correlated with the amount of IscS protein added. This result indicates that the IscS protein contributes to the biotin synthase reaction as does the NifS protein. There was only a little additional enhancement by the IscU protein, which is a homolog of the NifU protein in E. coli.

In conclusion, cysteine desulfurases, such as the NifS and IscS proteins, stimulated the biotin synthase reaction in the presence of l-cysteine. These results strongly suggest the participation of cysteine desulfurase in the biotin synthase reaction as a sulfur supplier to BioB protein. Tes Sum Bui et al. (21) and Gibson et al. (10) have proposed that the sulfur of biotin is derived from the iron-sulfur cluster of BioB protein. In addition, the sulfur of l-cysteine was incorporated into biotin with low efficiency in the cell-free systems of B. sphaericus (8) and B. subtilis (our unpublished result). Therefore, the sulfur might be released from l-cysteine by cysteine desulfurase and transferred into the iron-sulfur cluster of biotin synthase. Then, the sulfur of the iron-sulfur cluster of biotin synthase is incorporated into biotin, followed by the regeneration of the iron-sulfur cluster of biotin synthase by cysteine desulfurase. However, in this study, the regeneration system of the iron-sulfur cluster stimulated the conversion activity, even though the turnover number was less than 1 per mol of biotin synthase monomer (Table 2 and Fig. 4). Sanyal et al. have reported that the deduced iron-sulfur cluster in BioB protein was unstable, and the cluster was destroyed (18). Moreover, the sample of BioB protein used in our study might contain apo-BioB protein (without [2Fe-2S] clusters). Therefore, apo-BioB protein must have existed in the reaction mixture even though BioB protein did not catalyze. The generation system must have stimulated the conversion activity via rebuilding the iron-sulfur cluster of apo-BioB protein or stabilizing the iron-sulfur cluster of BioB protein. Any contribution of NifU or IscU protein solely on the biotin formation from DTB is unclear.

In the pure enzyme system, the biotin synthase reaction nearly stopped after 1 h of incubation in the presence of the regeneration system of the iron-sulfur cluster consisting of cysteine desulfurase and l-cysteine. This result suggests a possibility that an unknown factor(s) might further be involved in the enzyme turnover at a catalytic level in the pure enzyme system. However, the activity of biotin synthase is still far from the catalytic level, even in the cell-free system in the presence of the regeneration system. In the meantime, growth of E. coli biotin-deficient mutants, such as R875 (lacking bioB), is supported by only a little biotin (less than 1 ng/ml). These findings would imply that the bioconversion reaction of DTB to biotin in microorganisms is noncatalytic by nature, because their biotin requirement is so minute.