Novel Aldoxime Dehydratase Involved in Carbon-Nitrogen Triple Bond Synthesis of Pseudomonas chlororaphis B23 SEQUENCING, GENE EXPRESSION, PURIFICATION, AND CHARACTERIZATION*

Analysis of the nitrile hydratase gene cluster involved in nitrile metabolism of Pseudomonas chlororaphis B23 revealed that it contains one open reading frame encoding aldoxime dehydratase upstream of the amidase gene. The amino acid sequence deduced from this open reading frame shows similarity (32% identity) with that of Bacillus phenylacetaldoxime dehydratase (Kato, Y., Nakamura, K., Sakiyama, H., Mayhew, S. G., and Asano, Y. (2000) Biochemistry 39, 800–809). The gene product expressed in Escherichia coli catalyzed the dehydration of aldoxime into nitrile. The Pseudomonas aldoxime dehydratase (OxdA) was purified from the E. coli transformant and characterized. OxdA shows an absorption spectrum with a Soret peak that is characteristic of heme, demonstrating that it is a hemoprotein. For its activity, this enzyme required a reducing reagent, Na 2 S 2 O 4 , but did not require FMN, which is crucial for the Bacillus enzyme. The enzymatic reaction was found to be catalyzed when the heme iron of the enzyme was in the ferrous state. Calcium as well as iron was included in the

We have extensively studied the biological metabolism of toxic compounds (that have a triple bond between carbon and nitrogen) such as nitriles (R-C'N) (1)(2)(3)(4)(5) and isonitriles (R-N'C) (6,7). The microbial degradation of nitriles proceeds through two enzymatic pathways (8 -10): (i) nitrilase catalyzes the direct cleavage of nitriles into acids (R-C(ϭO)-OH) and ammonia (11)(12)(13), and (ii) nitrile hydratase (NHase) 1 catalyzes the hydration of nitriles to amides (R-C(ϭO)-NH 2 ) (14 -17), which are subsequently hydrolyzed to acids and ammonia by amidase (18 -20). These enzymes have received much attention in applied fields (2,8,21) as well as academic ones (13,(22)(23)(24)(25)(26). One of the fruits of our application-oriented nitrile studies is the current industrial production of acrylamide and nicotinamide involving NHase of an actinomycete, Rhodococcus rhodochrous J1 (1,9). On the other hand, NHase of Pseudomonas chlororaphis B23 (27), which was previously used as a catalyst for acrylamide manufacture (9,21), is now used for the production of 5-cyanovaleramide, a herbicide intermediate, at the industrial level (28). Because the amount of NHase produced comprises more than 50% of all soluble protein in P. chlororaphis B23 when this strain is cultured in the presence of methacrylamide (29), there must be a very interesting regulation mechanism for the enzyme expression (8). We have already cloned the NHase gene from this strain, and discovered the existence of a gene cluster containing the NHase and amidase genes (30). Although we have clarified its nitrile degradation mechanism at the protein and gene levels, an enzyme and its gene involved in the nitrile-synthetic pathway (which might be present in this strain) have never been identified.
In the biosynthesis of nitriles, aldoxime (R-CHϭN-OH) is one of the candidate intermediates (31). Indoleacetaldoxime, which is derived from tryptophan, is dehydrated to indoleacetonitrile in some higher plants. Aldoximes are also converted to cyanogenic glycosides and glucosinolates (31-34) (e.g. phenylpropionaldoxime to phenylethyl glucosinolate in Nasturtium officinale) (35). Other than these aldoximes, some naturally occurring aldoximes have so far been found in microorganisms; e.g. 2-(4-hydroxyphenyl)-2-oxoacetaldehyde oxime (which was discovered as a phosphorylated cholinesterase reactivator) in a fungus, Penicillium olsonii (36). However, there have only been two reports on purified enzymes (involved in nitrile synthesis from aldoxime) whose genes have been cloned but which belong to different superfamilies: (i) CYP71E1, a cytochrome P450 (which catalyzes the conversion of p-hydroxyphenylacetaldoxime to p-hydroxyphenylacetonitrile) in dhurrin biosynthesis of sorghum (37) and (ii) phenylacetaldoxime dehydratase (which * This work was supported in part by the 21st Century COE (tara. tsukuba.ac.jp/ϳcoe21/) Program of the Ministry of Education, Culture, Sports, Science, and Technology; by a grant-in-aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan; by the Industrial Technology Research Grant Program in '02 of the New Energy and Industrial Technology Development Organization of Japan; by the National Project on Protein Structural and Functional Analyses; and by a Research Grant (A) of the University Research Projects. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB093544.
§ Both authors contributed equally to the results of this work. ʈ To whom correspondence should be addressed. Fax: 81-29-853-4605 (Institute). catalyzes the dehydration of phenylacetaldoxime to phenylacetonitrile) linked with nitrilase in nitrile metabolism of Bacillus sp. strain OxB-1 (38). In particular, regarding the phenylacetaldoxime dehydratase, only one paper (38) on analyses at the protein and gene levels has been published. Although the Bacillus enzyme has been found to bind loosely to heme and requires FMN for its enzymatic activity, the function of such cofactors remains unknown. Biochemical and genetic information on the enzyme remain limited. This enzyme is also expected to be applicable to the production of nitriles from the corresponding aldoximes, because the chemical dehydration of aldoximes usually requires harsh reaction conditions (39).
In the present study, we discovered a gene (oxdA) of a novel nitrile-synthesizing enzyme linked with NHase in P. chlororaphis B23. We constructed an Escherichia coli transformant overexpressing OxdA, and we purified OxdA and determined its enzymological, physicochemical, and spectroscopic properties. We also report that aldoxime dehydratase is a unique enzyme involved in the metabolism of aldoxime in vivo in P. chlororaphis B23.

EXPERIMENTAL PROCEDURES
Materials-All aldoximes were purchased from Tokyo Kasei Kogyo Co., Ltd. (Tokyo, Japan). DEAE-Sephacel, phenyl-Sepharose CL-4B, and a low molecular weight standard kit were obtained from Amersham Biosciences. TSK gel Butyl-Toyopearl 650M was purchased from Tosoh Co. (Tokyo, Japan). Cellulofine HAp was purchased from Seikagaku Kogyo Co. (Tokyo, Japan). Standard proteins for high performance gel filtration chromatography were obtained from Oriental Yeast (Tokyo, Japan). All other biochemicals were standard commercial preparations.
DNA Manipulations-Restriction endonucleases, DNA polymerase, and T4 DNA ligase were purchased from Toyobo Co., Ltd. (Osaka, Japan). Nucleotides were sequenced by the dideoxy chain-terminating method using an ABI Prism 310 genetic analyzer (Applied Biosystems, Foster City, CA). Unless otherwise stated, DNA manipulations were performed essentially as described by Maniatis et al. (41).
Expression and Purification of Recombinant Aldoxime Dehydratase-The coding sequence of the enzyme was amplified by PCR with pPCN4 (30) as a template. The following two oligonucleotide primers were used: sense primer (5Ј-CATATGGAATCTGCGATCGA-CACGC-3Ј) containing an NdeI recognition site (underlined) and 22 nucleotides of oxdA starting with the ATG start codon; and antisense primer (5Ј-ACGCGTCGACTCAGGTGGGCGCGACAACGGC-3Ј) containing an SalI recognition site (underlined) and 21 nucleotides that are complementary to the 3Ј-end sequence of oxdA ending with the TGA stop codon. The amplified DNA was subcloned into vector pUC18 and checked by DNA sequencing. The insert DNA was digested with NdeI and SalI and then inserted into the respective sites of pET-24a(ϩ). The resultant plasmid was designated as pET-oxdA; in this construction, oxdA was under the control of the T7 promoter. E. coli BL21-Codon-Plus(DE3)-RIL was transformed with pET-oxdA, and the recombinant cells were used for the overproduction and purification of aldoxime dehydratase.
The transformed cells were incubated with reciprocal shaking at 37°C in 240 ml of 2ϫ YT medium containing 50 g/ml kanamycin and 34 g/ml chloramphenicol. After overnight cultivation, the entire culture was inoculated into 48 liters of the same medium, followed by incubation with shaking at 37°C for 2 h. Isopropyl-1-thio-␤-D-galactopyranoside was then added to a final concentration of 0.1 mM to induce the T7 promoter, and further cultivation was carried out at 15°C for 7 days.
All purification procedures were performed at 0 -4°C. Potassium phosphate buffer (pH 7.0) containing 5 mM 2-mercaptoethanol was used throughout the purification. Centrifugation was carried out for 30 min at 15,000 ϫ g. The cells were harvested by centrifugation, washed twice with 100 mM buffer containing 1 mM dithiothreitol, and then disrupted by sonication (Insonator model 201M; Kubota, Tokyo, Japan) to prepare a cell-free extract. Cell debris was removed by centrifugation. The resulting supernatant was fractionated with ammonium sulfate (40 -60% saturation), followed by dialysis against 10 mM buffer. The dialyzed solution was applied to a DEAE-Sephacel column (5 ϫ 25 cm) equilibrated with 10 mM buffer. Protein was eluted from the column with 1.2 liters of 10 mM buffer, the concentration of KCl being increased linearly from 0.1 to 0.4 M. The active fractions were collected, and then ammonium sulfate was added to give 70% saturation. After centrifugation of the suspension, the precipitate was dissolved in 0.1 M buffer and brought to 20% ammonium sulfate saturation. The enzyme solution was placed on a TSK gel Butyl-Toyopearl 650M column (5 ϫ 22 cm) equilibrated with 10 mM buffer 20% saturated with ammonium sulfate. The enzyme was eluted by lowering the concentration of ammonium sulfate (from 20 to 5%) in 1.2 liters of the same buffer. The active fractions were combined and precipitated with ammonium sulfate at 70% saturation. The precipitate was collected by centrifugation, dissolved in 0.1 M buffer, and then dialyzed against 10 mM buffer containing 0.5% ammonium sulfate. The enzyme solution was put on a phenyl-Sepharose column (5 ϫ 22 cm) that had been equilibrated with the same buffer. After the column had been washed fully with 10 mM buffer containing 0.5% ammonium sulfate, the enzyme was eluted with 10 mM buffer. The active fractions were combined and pooled. Solid ammonium sulfate was added to the enzyme solution to give 70% saturation. After centrifugation of the suspension, the precipitate was dissolved in 0.1 M buffer, followed by dialysis against three changes of 2 liters of 1 mM buffer (pH 6.8). After centrifugation, the enzyme solution was loaded on a Cellulofine HAp column (5 ϫ 22 cm) equilibrated with 1 mM buffer (pH 6.8). The column was eluted with a linear gradient, 1-100 mM, of the buffer (pH 6.8). The enzyme solution was refractionated with ammonium sulfate (50 -60% saturation). The precipitate was collected by centrifugation and dissolved in 0.1 M buffer. The resultant solution was dialyzed against 10 mM buffer and then centrifuged. The homogeneity of the purified protein was confirmed by SDS-PAGE.
Enzyme Assays-All of the reactions were performed under linear conditions as to protein (ϳ2.5 g/ml) and time (ϳ5 min). The standard assay A mixture comprised 100 mM potassium phosphate buffer (pH 7.0), 5 mM butyraldoxime, and an appropriate amount of enzyme, in a total volume of 200 l. The reaction was started by the addition of butyraldoxime and carried out for 5 min at 30°C under aerobic conditions. The reaction was stopped by the addition of 100 l of 1 M KOH, and a supernatant was obtained by centrifugation (15,000 ϫ g, 5 min). The reaction product was determined with a gas chromatograph (GC-14BPF; Shimadzu, Kyoto, Japan) equipped with a flame ionization detector and a glass column (3.2 mm ϫ 2.1 m) packed with Gaskuropack 56 (80/100% mesh; GL-Science, Tokyo, Japan). Standard assay A was performed throughout the purification procedures. After the purification, the enzyme assaying in all of the following experiments was performed by means of standard assay B (anaerobically under reduced conditions), unless otherwise stated. In the case of standard assay B, Na 2 S 2 O 4 (final, 5 mM) was added to the buffer. One unit of aldoxime dehydratase activity was defined as the amount of enzyme that catalyzed the formation of 1 mol of butyronitrile/min from butyraldoxime under the standard assay A and B conditions, respectively. Specific activity is expressed as units/mg of protein.
Determination of the Prosthetic Group-Heme staining was carried out as described by Klatt et al. (42). Heme was identified, and its stoichiometry was estimated by conversion to the pyridine hemochrome (43). This procedure consisted of adding pyridine (final, 20%) and then KOH (final, 50 mM) and a small amount of Na 2 S 2 O 4 to the enzyme. After 15 min, the spectrum of the pyridine hemochrome was recorded. The hemochrome was determined using the published molar absorption coefficient (⑀) for the hemochrome of protoheme IX, 34.4 mM Ϫ1 cm Ϫ1 , at 557 nm (43). Acid-acetone heme extraction was carried out (44). The enzyme was added with rapid stirring to 9 volumes of 15 mM HCl in acetone at 0°C, followed by incubation for 10 min at 0°C in the dark. After a 10-min centrifugation at 17,000 ϫ g, the supernatant was pooled, and the same operation was repeated to prepare a pyridine hemochrome.
Molecular Mass Determination-The purified enzyme sample was applied to a Superose 12 HR10/30 column (Amersham Biosciences), which was attached to an Ä KTA purifier (Amersham Biosciences), and then eluted with 100 mM buffer containing 0.15 M KCl at the flow rate of 0.5 ml/min. For the enzyme in the ferrous state after reduction with Na 2 S 2 O 4 , on the other hand, 100 mM buffer containing 0.15 M KCl and 10 mM Na 2 S 2 O 4 was used. The absorbance of the effluent was recorded at 280 and 422 nm. The molecular mass of the enzyme was calculated from the mobilities of the standard proteins, glutamate dehydrogenase (290 kDa), lactate dehydrogenase (142 kDa), enolase (67 kDa), adenylate kinase (32 kDa), and cytochrome c (12.4 kDa).
Metal Analysis-All glassware was soaked in 2 M HCl overnight and then exhaustively rinsed with distilled water before use. Prior to analysis, the enzyme was dialyzed against 1 mM potassium phosphate buffer (pH 7.0). The dialysis had no effect on the enzyme activity. The enzyme sample containing 0.94 mg of protein/ml was analyzed with an inductively coupled radiofrequency plasma spectrophotometer, Shimadzu ICPS-8000 (27.120 MHz). The metal contents of the enzyme sample were determined from the calibration curves for standard solutions.
Immunoblot Analysis-A female New Zealand White rabbit was immunized four times at 2-week intervals. The first and second injections were performed intradermally with the purified OxdA (500 g) in Freund's complete adjuvant (Difco). The booster injection was of 500 g of antigen homogenized in an equal volume of incomplete adjuvant (Difco).
Western blotting analysis was carried out as described below. Proteins were separated on SDS-polyacrylamide gels under reducing conditions and then electroblotted onto a polyvinylidene difluoride membrane (Sequi-Blot TM ; Bio-Rad). After blocking with 1% skim milk, the blots were incubated with rabbit antisera directed against OxdA at room temperature for 2 h and then incubated with horseradish peroxidase-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for 1 h. The immunoreactive proteins were detected with an ECL Western blotting detection kit (Amersham Biosciences).
Analytical Methods-The absorption spectra were recorded with a Shimadzu UV-1700 PharmaSpec spectrophotometer. The protein con-centrations were determined with a Nakalai Tesque Co., Inc. (Kyoto, Japan) protein assay kit with bovine serum albumin as the standard. Native PAGE and SDS-PAGE were carried out as described by Davis (45) and Laemmli (46), respectively. For SDS-PAGE, the following marker proteins were used to determine the relative molecular mass of the enzyme: phosphorylase b (94 kDa), bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), soybean trypsin inhibitor (20.1 kDa), and ␣-lactalbumin (14.4 kDa).

Nucleotide Sequence of the Open Reading Frame and Identification and
Purification of the Gene Product-In a previous study (30), based upon the results of SDS-PAGE analysis and enzyme assaying, we demonstrated that the region corresponding to the presumed 38-kDa protein (which would be located upstream of the amidase gene in P. chlororaphis B23) is required for maximum expression of the NHase gene in the E. coli transformant. Here, therefore, we sequenced the upstream region of the amidase gene in pPCN4 and found one open reading frame (ORF) clustered in the same direction with the structural genes of amidase, subunits ␣ and ␤ of NHase, the presumed 47-kDa protein (P47K), and OrfE. The nucleotide sequence of the ORF consists of 1,056 nucleotides (Fig. 1A), encoding 352 amino acid residues with a calculated molecular mass of 40,127 Da. A possible ribosome-binding site (GAGGA) is located 8 nucleotides upstream from the ATG. A search with the BLAST program revealed that the deduced amino acid sequence of the ORF exhibits low similarity (32% identity) with that of the phenylacetaldoxime dehydratase gene (oxd) of Bacillus sp. OxB-1 (38) (Fig. 1B). These genes have been found to comprise a superfamily, our enzyme being the second member of this superfamily.
An expression plasmid (pET-oxdA) for the ORF was constructed and introduced into E. coli BL21-CodonPlus(DE3)-RIL. The recombinant protein was obtained as an insoluble form almost completely when the E. coli transformant was cultured at 37°C. To increase the amount of the soluble enzyme form, various culture conditions were examined. Under the optimum conditions for 7 days at 15°C, we succeeded in expressing the recombinant protein at a level corresponding to about 10% of the total amount of soluble protein (data not shown). Cell-free extracts prepared from the E. coli transformant carrying the ORF dehydrated butyraldoxime as a substrate, yielding butyronitrile. This finding demonstrated that this ORF encodes aldoxime dehydratase. Thus, the gene was designated as oxdA.
OxdA was purified through the purification steps described under "Experimental Procedures." The purity of OxdA was confirmed by migration of the protein as a single band corresponding to a molecular mass of ϳ38 kDa on SDS-PAGE (Fig.  2). The M r of the native enzyme was 76.4 kDa according to the results of gel filtration chromatography, indicating that OxdA consists of two identical subunits. The purified OxdA showed specific activity of 0.758 units/mg (under aerobic conditions) ( Table I).
Cofactor of Aldoxime Dehydratase-OxdA was brownish red in solution. The existence of heme was expected in OxdA. Thus, heme staining was carried out after native PAGE for our aldoxime dehydratase. The gel was stained with 3,3Ј-dimethoxybenzidine/H 2 O 2 (42) and Coomassie Brilliant Blue for heme and protein, respectively. Both methods gave a single band, and the positions of the bands were the same (data not shown). Besides, the absorption spectrum of this enzyme had a specific peak at 415 nm (Fig. 3A). Thus, OxdA was identified as a hemoprotein. To examine the properties of the heme in the enzyme, pyridine hemochrome was prepared. Fig. 3C shows absorption peaks at 418.5 (Soret), 524 (␤-band), and 556 (␣-band) nm. Furthermore, the heme prosthetic group was extracted from the enzyme on HCl/acetone-treatment (44). The spectrum of the pyridine hemochrome in the extract was identical with that of the pyridine hemochrome of the enzyme (data not shown). The extractability of the chromophore with acid-acetone and the shape of its pyridine hemochrome spectrum confirmed that the aldoxime dehydratase carries protoheme IX as the prosthetic group. The heme content of the enzyme was calculated to be 0.69 mol/mol of subunit, as determined from the value for ⑀ at 556 nm in the spectrum of its pyridine hemochrome. Quantitative analysis of iron in OxdA with an inductively coupled radiofrequency plasma spectrophotometer revealed that OxdA (as a homodimer) contains 1.62 mol of iron/mol. On the basis of these results, it is concluded that OxdA contains iron and heme in a ratio of 1:1 and that all of the iron would be present in the heme molecule. The heme content of the preparation is not equimolar as to the protein, suggesting that a little of the heme in OxdA is lost during the purification procedure. It has been reported that the Bacillus phenylacetaldoxime dehydratase loses heme during its purification and that the purified enzyme contains only 0.35 mol of heme/mol of enzyme (38).
Effect of the Redox Environment on the Enzyme Activity-In order to confirm the state of the iron in the heme molecule, redox spectra of the purified OxdA were examined. Under aerobic conditions, OxdA showed a strong absorption peak at 415 nm and weak absorption in a broad region (i.e. 520 -580 nm) (Fig. 3A). Under the same conditions, the specific activity of the purified OxdA was not so high (0.758 units/mg; Table II), although it was not inactive. The addition of Na 2 S 2 O 4 to OxdA caused an increase and shift from 415 to 428 nm of the absorption peak under anaerobic conditions (Fig. 3A). The activity increased by more than 250 times (namely from 0.758 to 197 units/mg) in the anaerobic reduced environment (Table II). Upon the subsequent addition of CO to OxdA reduced by the addition of Na 2 S 2 O 4 , a further increase and shift of the absorption peak from 428 to 419 nm were observed (Fig. 3A), suggesting that this peak is due to an enzyme-CO complex. The addition of an exogenous ligand of the ferrous state of the heme, CO, caused 49% inhibition of the activity (namely from 197 to 101 units/mg) (Table II); but the substrate was dehydrated even in the presence of CO. These findings suggest that CO, which would have bound beforehand at the ferrous heme of the enzyme, was replaced by butyraldoxime during the enzyme reaction. On the other hand, the absorption spectrum of OxdA was unchanged with CO under aerobic conditions (data not shown). Upon the addition of K 3 [Fe(CN) 6 ] (which is an oxidizing agent), OxdA showed a shift of the absorption peak from 415 to 412 nm under aerobic conditions, whereas it showed a shift of the absorption peak from 415 to 419 nm upon the addition of 10 mM KCN under aerobic conditions (Fig. 3B). Under these conditions, enzyme activities have never been detected (Table II). These spectral features indicate that the heme iron of the purified OxdA is in the ferric state but that the enzyme functions in the ferrous state. In all of the following experiments described below, therefore, all procedures were performed after reduction by the addition of Na 2 S 2 O 4 ; standard assay B described under "Experimental Procedures" was used as the enzyme assay unless otherwise stated.
For determination of the molecular mass of the ferrous enzyme, gel filtration chromatography was carried out again un- der reduced conditions. The reduced OxdA sample prepared by the addition of Na 2 S 2 O 4 was applied to a Superose 12 HR10/30 column and then eluted with 100 mM buffer containing 0.15 M KCl and 10 mM Na 2 S 2 O 4 . The absorbance of the effluent was recorded at 422 nm. A single peak corresponding to 76.2 kDa was detected (data not shown). Also, the spectrum of the pooled fraction was confirmed to be identical to that of the ferrous state. These findings, together with the finding that OxdA is most active under reduced conditions, indicate that the native OxdA in the ferrous state is also a homodimer.
Stoichiometry-The stoichiometry of aldoxime consumption and nitrile formation during the dehydration of aldoxime was examined in a reaction mixture consisting of 100 mM potassium phosphate buffer (pH 7.0), 5 mM butyraldoxime, 5 mM Na 2 S 2 O 4 , and 31.25 nM enzyme, in a final volume of 400 l. The reaction was carried out at 30°C under anaerobic conditions. After a 5-min incubation, the amounts of residual butyraldoxime and butyronitrile were determined. The butyronitrile formed and the butyraldoxime remaining amounted to 1.93 and 3.12 mM, respectively. No formation of other compounds was observed. The results indicated that butyronitrile was formed stoichiometrically with the consumption of butyraldoxime.
Substrate Specificity and Kinetic Properties-The ability of OxdA to catalyze the dehydration of various aldoximes was examined. Among the tested aldoximes, only three were active as substrates for the aldoxime dehydratase, whereas aromatic aldoximes such as benzaldoxime and pyridine-4-aldoxime were inert. The k cat value for pyridine-4-aldoxime (0.090 min Ϫ1 ) was much lower than those for butyraldoxime and acetaldoxime (5.4 and 5.6 min Ϫ1 , respectively), suggesting that OxdA prefers aliphatic aldoximes to aromatic ones as substrates (Table III). Because of their commercial unavailability, not only arylalkyland alkyl-aldoximes (e.g. phenylacetaldoxime), but also pure Eand Z-stereoisomers of aldoximes, which are not listed in Table  III, could not be examined. Butyraldoxime comprised a 60:40 mixture of E/Z-stereoisomers (commercial communication). The complete consumption of E/Z-butyraldoxime in the enzyme reaction suggests that OxdA accepts both the E-and Z-forms of butyraldoxime as substrates. The specific activity of OxdA for butyraldoxime (197 mol/min/mg; Table II) is much higher than that of the Bacillus phenylacetaldoxime dehydratase for this substrate (0.93 mol/min/mg), although the assay conditions were not the same. Not only butyronitrile but also all of the other tested nitriles did not act as substrates for OxdA, despite the addition of a large amount of the enzyme and a long incubation period.
Effects of Temperature and pH on the Activity and Stability of the Enzyme-The effects of pH and temperature on the enzyme activity were examined. OxdA was active over a broad pH range (pH 5.0 -10.0). Maximum activity was observed at pH 5.5; another high activity level was observed at pH 9.4 (Fig.  4A). The reason for such an interesting pH optimum pattern, which has never previously been observed for the Bacillus phenylacetaldoxime dehydratase (38), is unclear. The optimal temperature was 45°C (Fig. 4B).
The stability of OxdA was examined at various pH values. After the enzyme had been incubated at 30°C for 15 min in a 0.1 M concentration of the buffers citrate/sodium citrate (pH 4.4 -6.5), potassium phosphate buffer (pH 6.7-7.9), Tris/HCl buffer (pH 7.9 -8.8), and NH 4 Cl/NH 4 OH buffer (pH 8.9 -10.4), an aliquot of each enzyme solution was reduced by the addition of Na 2 S 2 O 4 , and then the enzyme activity was measured anaerobically. OxdA was most stable in the pH range of 6.0 -8.0, 40% of its initial activity being retained even at pH 10.0.
Spectral Interaction of Aldoximes with Aldoxime Dehydratase-The addition of butyraldoxime (final concentration 50 mM) to OxdA (final concentration 0.23 mg/ml) reduced by Na 2 S 2 O 4 under anaerobic conditions led to the rapid appearance of a peak at 416 nm; an intense difference spectrum exhibiting a peak at 412.5 nm and a trough around 430 nm has obtained (Fig. 5). Butyronitrile was unable to give the 416-nm peak when added to the Na 2 S 2 O 4 -reduced enzyme (data not shown). These findings suggest that the substrate (aldoxime) really approaches the ferrous heme of the enzyme. As the enzymatic reaction proceeded, the peak at 416 nm disappeared (Fig. 5). After a 10-min reaction, a spectrum close to that at the start of the reaction was observed. Under the same conditions with butyraldoxime as the substrate in the reaction mixture, the aldoxime dehydratase assay showed the complete consumption of butyraldoxime and the concomitant appearance of butyronitrile at 10 min; no substrate remained in the reaction mixture at 10 min. These findings demonstrated that the absorption peak at 416 nm is due to the intermediate in the butyronitrile synthetic reaction. On the other hand, when ac- etaldoxime was added as the substrate (final, 50 mM) to OxdA (final, 0.23 mg/ml) reduced by Na 2 S 2 O 4 , a difference spectrum similar to that in the case of butyraldoxime as the substrate was observed (data not shown), although there was a slight difference between them. Under these conditions, the addition of acetaldoxime gave a peak at 415 nm; a difference spectrum characterized by a peak at 411 nm and a trough around 430 nm was obtained. After a 30-min reaction, a spectrum close to that at the start of the reaction was observed, and the complete consumption of acetaldoxime and the concomitant appearance of acetonitrile were further confirmed. These results obtained on enzyme assaying and spectral analyses demonstrate that the substrate really approaches the ferrous heme of the enzyme during the enzymatic reaction; namely the heme molecule in OxdA under reduced conditions plays an essential part in the catalytic reaction. Aldoxime Dehydratase Activity in P. chlororaphis B23-To confirm the function of OxdA in vivo, P. chlororaphis B23 was grown in several culture media. When P. chlororaphis B23 was grown in the medium containing 1% sucrose and 0.5% (NH 4 ) 2 SO 4 as carbon and nitrogen sources, respectively, the OxdA activity was not observed. On the other hand, P. chlororaphis B23 grew slowly when butyraldoxime (0.05%) was used as the sole carbon and nitrogen source, although the strain did  The reaction was carried out at 30°C in the standard assay B mixture except that various aldoximes were used as the substrate in place of butyraldoxime. The K m and k cat values were determined by the nonlinear least squares regression method using the program Kaleida Graph (Synergy Software, Reading, PA). NA, no activity could be detected. not grow when butyraldoxime was used at the concentration of 0.2%. These findings suggested that butyraldoxime at a high concentration would be toxic for the strain, but butyraldoxime at a low concentration can be used as carbon and nitrogen sources. Upon replacement of sucrose and (NH 4 ) 2 SO 4 with butyraldoxime, the OxdA activity was detected with 0.561 units/mg under standard assay B. These findings also indicated that OxdA was an inducible enzyme.
Moreover, the formation of OxdA in P. chlororaphis B23 cells was examined by means of immunoblot analysis. Cell-free extracts were prepared from the strain grown with butyraldoxime and then subjected to SDS-PAGE. A prominent band of OxdA, migrating to a position corresponding to a molecular weight of 38,000, was detected with antiserum raised against OxdA. No OxdA was detected when the strain was grown without aldoxime. Immunoblot analysis showed the strong coincidence of the OxdA formation with the OxdA activity in vivo (Fig. 6).
Sequence analysis of the upstream region of the amidase gene in the "industrial microorganism" P. chlororaphis B23 revealed the existence of the aldoxime dehydratase gene exhibiting weak sequence similarity to the Bacillus phenylacetaldoxime dehydratase gene, whose protein was recently characterized biochemically (38), and it was found to be completely different from all of the enzymes reported above not only in sequence but also in enzymological and physicochemical properties.
In order to characterize this aldoxime dehydratase enzyme well, we attempted to express the recombinant OxdA of P. chlororaphis B23 in E. coli. Although the recombinant phenylacetaldoxime dehydratase of Bacillus sp. OxB-1 has almost never been obtained in an active form, because of its production as inclusion bodies (Ͻ10 units/liter of culture), we here succeeded in the overproduction of OxdA, which allowed the characterization of this enzyme in various ways after purification.
OxdA of P. chlororaphis B23 has been found to be a hemoprotein comprising protoheme, as judged from the isolation of the pyridine hemochrome and the results of spectroscopic analyses. Although the absorption spectrum of the Bacillus phenyl- a Methanol was added to the reaction mixture to the final concentration of 5% (v/v) to enhance the solubitity of the tested compound. acetaldoxime dehydratase is not affected by the addition of various compounds (i.e. reducing reagents, oxidizing reagents, substrates, products, and common exogenous heme ligands), a variety of spectral shifts were observed on the addition of each of various compounds to OxdA. Spectroscopic analyses and enzyme assaying have revealed that the enzyme reaction proceeds when the heme iron of OxdA is in the ferrous state. In the case of the Bacillus phenylacetaldoxime dehydratase, FMN is necessarily required, although the precise nature of the FMN requirement is unclear (38). As for the cytochrome P450s that act on aldoximes, NADPH is required for reduction of the heme to the ferrous state (47,48). However, no enhancement of the enzyme activity has been observed on the addition of many kinds of cofactors, including FMN and NADPH, to OxdA, which would be in the ferric state under aerobic conditions. OxdA exhibits much higher activity with the addition of Na 2 S 2 O 4 , which is a strong reducing reagent under anaerobic conditions (Table II). These findings indicate that OxdA may not be easy to reduce. There may be a mechanism that causes the reduction of OxdA to the ferrous state. When the heme iron of OxdA is in the ferrous state under anaerobic conditions, the enzyme has been confirmed to be a dimer. This is the first report that an aldoxime-degrading enzyme acts as a dimer, although it has been reported that the Bacillus phenylacetaldoxime dehydratase is a monomer (38). Metal analysis of OxdA demonstrated that iron and calcium atoms are included in the enzyme. Both metal ions remain even during the purification procedure or dialysis, suggesting that the two metals are tightly associated with the enzyme. Whereas the iron should be derived from the heme, calcium may act as another cofactor. Further studies are required to determine the function of the calcium.
OxdA of P. chlororaphis B23 shows the most significant activity toward butyraldoxime, yielding butyronitrile through its reaction. For the NHase of this strain (27), butyronitrile is a better substrate and exhibits the highest affinity (K m ϭ 1.03 mM) among all of the tested nitriles. Other aliphatic nitriles such as propionitrile and acrylonitrile are remarkably active as substrates for the enzyme, although aromatic nitriles are barely hydrated by the NHase (27). The amidase linked with this NHase can also act on aliphatic amides including butyramide and isobutyramide, rather than aromatic amides, as substrates (55). In this study, the kinetic parameters of OxdA indicated that aliphatic aldoximes are more effective substrates than aromatic aldoximes. We observed a similar trend in substrate specificity among the NHase, amidase and OxdA, which co-exist in the same gene cluster in P. chlororaphis B23 (Fig. 7).
The aldoxime dehydratase reaction is surprising, particularly in the dehydration of the substrate (aldoxime) and the formation of the carbon-nitrogen triple bond, both of which have rarely been observed for enzymes known thus far. In this reaction scheme, H 2 O is apparently removed from the substrate, whereas the reaction mixture contains water. Such an efficient dehydration reaction is very unusual, and the catalytic mechanism in which the heme would be involved is interesting. Although both aldoxime dehydratase and CYP71E1 (cytochrome P450) act on the same substrate, "aldoxime," and are classified as a group of aldoxime-degrading enzymes, they exhibit no sequence similarity to each other; the two enzymes thus do not share a common ancestor. Further biochemical analyses, including three-dimensional analysis, of these enzymes are required to clarify the differences in their reaction mechanisms. Such analyses will also be significant from a biotechnological perspective for the production of nitriles from the corresponding aldoximes. OxdA is an interesting target with a unique catalytic mechanism, because the enzymatic reaction proceeding even in the presence of CO is rare.
Our report is the first demonstration of a gene cluster consisting of the structural genes of aldoxime dehydratase, amidase, and NHase, whereas aldoxime dehydratase activity is detected in some microorganisms with NHase activity (56 -58) (Fig. 7). This gene association strongly supports the theory that these three enzymes are closely related in the following metabolism: aldoxime 3 nitrile 3 amide 3 acid and ammonium. In Bacillus sp. strain OxB-1 (38,56), on the contrary, there is different aldoxime metabolism: aldoxime 3 nitrile 3 acid and ammonium, through the combination of phenylacetaldoxime dehydratase and nitrilase. In this strain, the phenylacetaldoxime dehydratase gene is coded by the region just downstream from the nitrilase gene, which is present just upstream of a possible regulatory gene for the expression of the nitrilase gene (38). We have also discovered sequence similarity (42% identity) between OxdA and an unknown ORF (designated as Psyr0978 under GenBank TM accession number NZ_AABH01000002) of Pseudomonas syringae pv. syringae FIG. 7. Organization of the gene cluster for aldoxime metabolism and the corresponding metabolic pathway. Microorganisms were as follows: P. chlororaphis B23 (B23), Bucillus sp. OxB-1 (OxB-1), and P. syringae pv. syringae B728a (Psyr). Regulatory protein, nitrilase, and aldoxime dehydratase in P. syringae, which we presumed, are registered as Psyr0980, Psyr0979, and Psyr0978, respectively, under GenBank TM accession number NZ_AABH01000002.
B728a, a phytopathogenic bacterium (data not shown). In P. syringae, the same adjacent gene organization is observed (Fig.  7), although the sequence of each of the constituent ORFs in this organization is only reported in this bacterial genome data base, and its function has never been identified. On the other hand, we demonstrated that OxdA was in fact utilized in P. chlororaphis B23. There were significant correlations between the addition of aldoxime in the medium and OxdA formation, indicating that OxdA is involved in the degradation of aldoxime in vivo (Fig. 6). Here, we have the following question. Do the bacteria with aldoxime-degrading activity also have aldoximesynthetic ability? In glucosinolate-producing plants (e.g. Arabidopsis), aldoximes have been shown to be formed from amino acids by cytochrome P450 belonging to the CYP79 family (33, 59 -61). Further sequence analyses of the upstream and downstream regions of the aldoxime-related gene cluster in the bacteria with aldoxime-degrading activity may provide information on such aldoxime biosynthesis.