Putative ACP Phosphodiesterase Gene ( acpD ) Encodes an Azoreductase*

An FMN-dependent NADH-azoreductase of Esche-richia coli was purified and analyzed for identification of the gene responsible for azo reduction by microorganisms. The N-terminal sequence of the azoreductase conformed to that of the acpD gene product, acyl carrier protein phosphodiesterase. Overexpression of the acpD gene provided the E. coli with a large amount of the 23-kDa protein and more than 800 times higher azoreductase activity. The purified gene product exhibited activity corresponding to that of the native azoreductase. The reaction followed a ping-pong mechanism re-quiring 2 mol of NADH to reduce 1 mol of methyl red (4 (cid:1) -dimethylaminoazobenzene-2-carboxylic acid) into 2-aminobenzoic acid and N,N (cid:1) -dimethyl- p -phenylenedi-amine. On the other hand, the gene product could not convert holo-acyl carrier protein into the apo form under either in vitro or in vivo conditions. These data indicate that the acpD gene product is not acyl carrier protein phosphodiesterase but an azoreductase.

. The gene product (20 kDa) catalyzed the reduction of azo dyes (Roccellin, Sumifix Black, and Solar Orange) in the presence of NADPH.
In this paper, we report the 4500-fold purification and Nterminal sequencing of a native azoreductase of E. coli. A search of the translated data bases allowed identification of the acpD gene as the gene for the azoreductase. Overexpression of the acpD gene allowed the preparation of azoreductase, in 190-mg quantities, having no ACP phosphodiesterase activity. The biochemical properties of the azoreductase have also been revealed.

Assaying of Azoreductase
The standard assay system for azoreductase comprised 25 mM Tris-HCl (pH 7.4), 25 M methyl red, 0.1 mM NADH, 20 M FMN, and * 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.
‡ To whom correspondence should be addressed: Laboratory of Molecular Biochemistry, Dept. of Biomolecular Science, Faculty of Engineering, Gifu University, 1-1 Yanagido, Gifu 501-1193, Japan. Tel.: 8158-293-2640; Fax: 8158-230-1893; E-mail: kitade@biomol.gifu-u. ac.jp. enzyme, in a final volume of 2 ml. The reaction was initiated by addition of the enzyme. The initial reaction rate was determined by monitoring the decrease in absorbance at 430 nm in the first 1.5 min in a glass cuvette of 1.0-cm light path at 30°C. The enzyme activity was a linear function of the incubation time. One unit of methyl red reductase activity was defined as the amount catalyzing the degradation of 1 mol of methyl red/min at 30°C using a molar absorption coefficient of 23,360 M Ϫ1 cm Ϫ1 . When methyl red was replaced with other azo dyes, the following molar absorption coefficients and wavelengths were used: 21,560 M Ϫ1 cm Ϫ1 (ethyl red at 450 nm) and 18,680 M Ϫ1 cm Ϫ1 (Ponceau SX at 500 nm). We confirmed that the absorption of NADH did not affect the absorption of the azo dyes. One unit of menadione reductase activity was defined as the amount catalyzing the reduction of 1 mol NAD(P)H/min at 30°C using a molar absorption coefficient of 6220 M Ϫ1 cm Ϫ1 .

Purification of Azoreductase
All procedures were carried out at 4°C. Preparation of Cell Extracts-E. coli (JM109) cells that had been grown at 37°C as 10-liter cultures in LB medium (1.0% tryptone, 0.5% yeast extract, 1.0% NaCl, pH 7.0) were resuspended in 160 ml of buffer A (20 mM Tris-HCl (pH 8.0), 1 mM EDTA). Lysozyme was added to give a final concentration of 0.1 mg/ml, and then the solution was incubated at 30°C for 20 min. The viscous solution was chilled and subjected to sonication disruption (180-watt output, 30 s three times). Cell debris was removed by centrifugation at 12,500 ϫ g for 15 min.
Ion Exchange Chromatography on Q-Sepharose FF-The cell extract was applied at 0.6 ml/min to a Q-Sepharose FF column (16 ϫ 2.5 cm) that had been pre-equilibrated with buffer A. After washing of the column with 800 ml of buffer A at 0.6 ml/min, azoreductase was eluted with 400 ml of buffer A containing a linear gradient of NaCl, from 0 to 0.6 M. Each fraction (fraction size, 6 ml) was subjected to activity measurement; the seven fractions containing the highest levels of activity were pooled.
Affinity Chromatography on Blue-Sepharose-The pooled fraction was applied at 0.5 ml/min to a Blue-Sepharose column (6.5 ϫ 2.0 cm) that had been pre-equilibrated with buffer A. After washing of the column with 200 ml of buffer A, azoreductase was eluted with 70 ml of buffer A containing 0.4 mM NADH.
Gel Filtration Chromatography on Sephacryl S200 -The eluate was concentrated to 6 ml on a YM-10 membrane and then applied at 0.3 ml/min to a Sephacryl S200 column (90 ϫ 1.6 cm) that had been pre-equilibrated with buffer B (20 mM Tris-HCl (pH 7.5), 0.2 M NaCl, 2 mM 2-mercaptoethanol). The two fractions (fraction size, 3 ml) containing the highest activity were pooled.
Hydroxyapatite Chromatography on GIGAPITE-The pooled fraction was applied at 0.3 ml/min to a GIGAPITE column (3 ϫ 1.4 cm) that had been pre-equilibrated with buffer B. After washing of the column with buffer B at 0.3 ml/min, azoreductase was eluted with 60 ml of buffer B containing a linear gradient of potassium phosphate (pH 7.5), from 0 to 50 mM. The 10 fractions (fraction size, 3 ml) containing activity were pooled, and the buffer was exchanged with buffer A.
Ion Exchange Chromatography on Q-Sepharose HP-The enzyme solution was applied at 0.3 ml/min to a Q-Sepharose HP column (1.5 ϫ 1.4 cm) that had been pre-equilibrated with buffer A. After washing of the column with buffer A, azoreductase was eluted with 20 ml of buffer A containing a linear gradient of NaCl, from 0 to 0.6 M. The 10 fractions (fraction size, 2 ml) containing activity were pooled.
Hydrophobic Interaction Chromatography on Phenyl-Sepharose-Ammonium sulfate was dissolved in the pooled fraction to obtain 30% saturation (164 g/liter). The solution was applied at 0.3 ml/min to a phenyl-Sepharose column (1 by 0.8 cm) that had been pre-equilibrated with buffer A containing ammonium sulfate (30% saturation). The column was washed with the buffer, and then 40 ml of a linear gradient of ammonium sulfate, 30 -0% saturation, in buffer A was applied. The fraction containing activity was concentrated and stored at Ϫ20°C after adding glycerol to give 20% (v/v) concentration.

N-terminal Amino Acid Sequencing
One microgram of the purified enzyme was run on a 12.5% SDSpolyacrylamide Tricine gel (1 mm thick) (17), and then blotted onto a ProBlot ® poly(vinylidene difluoride) membrane (Applied Biosystems). The protein band was examined with an Applied Biosystems model 477A protein sequencer, fitted with an on-line model 120A analyzer for the detection of phenylthiohydantoin-amino acids.

Expression and Purification of the acpD Gene Product
The acpD gene was obtained by PCR using genomic DNA of E. coli strain JM109 as the template. Pfu turbo DNA polymerase (Stratagene) and oligonucleotide primers (sense, 5Ј-cggccatatgagcaaggtattagtt-3Ј containing an NdeI site; antisense, 5Ј-gcgctcgagttatgcagaaacaat-3Ј containing a XhoI site) were used for PCR. The resulting DNA was subcloned into the corresponding restriction site of pET22b to obtain a plasmid designated as pETacpD. The nucleotide sequence of the cloned DNA was confirmed by dideoxy sequencing. E. coli JM109 (DE3) harboring pETacpD was grown at 37°C in LB medium. Expression was induced with 0.5 mM isopropyl-1-thio-␤-D-galactopyranoside (IPTG) for 3 h with an optical density at 600 nm of 0.6. The recombinant AcpD was purified by Q-Sepharose, Blue-Sepharose, Sephacryl S200, and GIGA-PITE column chromatographies.

Stoichiometry of Azo Reduction
Using the standard assay system (2 ml) with 7.8 ng of the purified acpD gene product (designated as AcpD), the methyl red degradation rate and NADH consumption rate in the first 1.5 min were determined by monitoring the absorbance at 430 and 340 nm, respectively. Because the absorbance of methyl red at 340 nm was a linear function of the absorbance at 430 nm under the given conditions, the absorbance at 340 nm in the standard assay system was corrected by calculation. The molar absorption coefficients were as described above.

Identification of Reaction Products
Methyl red was incubated in 40 ml of the standard assay system for 10 min at 30°C with the purified AcpD (20 g). For extraction of ABA, an aliquot (20 ml) of the reaction mixture was extracted with an equal volume of ethyl acetate three times, subsequent to adjusting the pH to 3 with 1 M HCl. The extracts were pooled and evaporated in a rotary evaporator. For DMPD, another aliquot (20 ml) was extracted with an equal volume of n-hexane three times, subsequent to adjusting the pH to 10 with 1 M NaOH. The extracts were pooled and evaporated in a rotary evaporator. Each residue was dissolved in 0.5 ml of acetonitrile. Each sample was diluted 200-fold with acetonitrile and then analyzed (20 l) with a Shimadzu HPLC system equipped with a model SPD-6AV variable wavelength detector (detection wavelength: 336 nm for ABA; 250 nm for DMPD) and a Jasco CrestPak ODS column (4.6 mm ϫ 150 mm). The mobile phase was composed of 25 mM phosphate and acetonitrile, 5:95 (25 mM phosphate:acetonitrile). The flow rate was 0.8 ml/min.

Overexpression and Purification of E. coli ACP
The gene for ACP was amplified using genomic DNA of E. coli strain JM109 as the template for PCR. The forward primer included an EcoRI site before the start codon: 5Ј-ctggaattcatgagcactatcga-3Ј. The reverse primer included a HindIII restriction site after the stop codon: 5Јctgaagcttttacgcctggtggc-3Ј. The PCR product was subcloned into the EcoRI/HindIII site of the pKK223-3 plasmid, and the resulting plasmid was designated as pKKacpP. E. coli JM109 was transformed with pKKacpP and grown at 37°C with LB medium. After the cultures had been grown to an optical density at 600 nm of 0.6, expression was induced with 0.5 mM IPTG for 3 h. The purification procedure followed the method of Therisod et al. (18).

In Vitro Assay for ACP Phosphodiesterase Activity
The assay system comprised 50 mM Tris-HCl (pH 8.5), 0.02 mM MnCl 2 , 25 mM MgCl 2 , 1 mM dithiothreitol, 20 g of holo/apoACP, and 3.7 g of purified AcpD, in a final volume of 0.1 ml. After incubation at 35°C for 12 h, a sample was analyzed by native PAGE (19). For the holoACP standard, 20 g of holo/apoACP was converted to holoACP under the conditions with 50 mM Tris-HCl (pH 8.8), 0.1 mM CoA, 25 mM MgCl 2 , 1 mM DTT, and 4.2 g of ACP synthase. ACP synthase was expressed and purified by a procedure based on that of Lambalot and Walsh (20). For the apoACP standard, Ser 36 -substituted ACP was used. PCR-based site-directed mutagenesis was carried out with substitution from Ser to Cys. The expression and purification procedure were the same as those for the wild type.

In Vivo Assaying of ACP Phosphodiesterase Activity
ACP was coexpressed with AcpD in E. coli and then analyzed by native PAGE. To obtain a plasmid (designated as pACYCacpP) that was compatible with pETacpD and produced ACP, the BamHI-HincII 0.6-kb DNA fragment of pACYC177 (21) was replaced by the 1.4-kb BamHI-PvuI fragment (the cohesive end of the PvuI site was blunt-ended) from pKKacpP. The 1.4-kb fragment contained a tac promoter, the ACP gene, and rrnB ribosomal RNA transcription terminators. E. coli JM109 (DE3) was cotransformed with pACYCacpP and pETacpD or pET21a (for a control experiment). Expression was induced by the addition of IPTG as described above.

Other Techniques
Protein concentrations were determined using protein assay reagent (Bio-Rad), with bovine serum albumin as the standard. SDS-PAGE was carried out using 12.5% gels as described by Laemmli (22). The proteins on gels were stained using a Silver Stain Kit II (Wako Pure Chemicals, Osaka, Japan). The native molecular weight of the protein was determined by gel filtration on a Superdex 200 HR column (30 ϫ 1.0 cm) that had been equilibrated with Tris-buffered saline buffer (20 mM Tris-HCl (pH 7.5) and 0.15 M NaCl). Calibration of the column was carried out with the following proteins: alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), carbonic anhydrase (29 kDa), and cytochrome c (12.4 kDa), all from Sigma.

RESULTS
Purification of Azoreductase-The azoreductase activity in the 12,500 ϫ g supernatant of the E. coli homogenate was 0.014 units/mg. The enzyme was purified from the supernatant ( Fig.  1) with an overall activity recovery of 0.7% and about 4500-fold purification by a combination of anion exchange, dye affinity, gel filtration, hydroxyapatite, and hydrophobic interaction chromatographies (Table I). The azoreductase exhibited molecular masses of 23 and 42 kDa on SDS-PAGE and gel filtration, respectively.
Amino Acid Sequence Analysis-The N terminus of the azoreductase was not blocked and revealed to be NH 2 -Ser-Lys-Val-Leu-Val-Leu-Lys-. The sequence conformed to the acpD gene product (AcpD), known as ACP phosphodiesterase (6).
Identification of AcpD as the Azoreductase-To confirm the azoreductase activity of AcpD, the acpD gene was overexpressed in E. coli. In all purification procedures ( Fig. 2 and Table II), AcpD behaved the same as the native azoreductase. Furthermore, AcpD exhibited identical migration with the azoreductase in an SDS-PAGE gel (Fig. 1, lane 8). The enzyme activity was also the same as that of the azoreductase (Table  III). These results indicate that the azoreductase is identical with AcpD.
Spectroscopic Properties of AcpD-The absorption spectrum of AcpD (0.47 mg/ml) exhibited a single peak at 278 nm, ranging in wavelength from 250 nm to 600 nm. The single peak indicated the absence of a flavin cofactor. The fluorescence emission of AcpD (0.22 mg/ml, 11 M) showed a maximum at around 310 nm, ranging from 280 to 380 nm (Fig. 3). When the enzyme solution was titrated with FMN or FAD, the fluorescence at 310 nm was gradually quenched. The degree of quenching by 12 M (slightly more than the stoichiometric amount) FMN and FAD was 70 and 30%, respectively (Fig. 3). The greater quenching by FMN implied the FMN specificity of AcpD.
Steady State Kinetic Analysis of AcpD-Double-reciprocal plots of initial velocity versus NADH or methyl red concentration resulted in parallel patterns (Fig. 4, A and B). The patterns are consistent with a ping-pong mechanism. Secondary plots of the intercepts against the reciprocals of variable substrate concentrations gave linear correlations. The K m values for NADH and methyl red were calculated to be 31.6 and 17.9 M, respectively.
Identification of Reaction Products-The products of the  a Because of the lability of this enzyme at lower protein concentration, the specific activity and yield were reduced.  standard reaction of AcpD were extracted under acidic (pH 3) and basic (pH 10) conditions. HPLC analyses of the extracts gave peaks of ABA (retention time, 2.23-2.25 min) and DMPD (retention time, 4.16 -4.23 min). The retention times of ABA and DMPD completely coincided with those of each authentic compound. In the absence of AcpD, NADH, or FMN, no peak other than that of methyl red (retention time, 5.2 min) was observed. Thus, AcpD catalyzed the reductive cleavage of methyl red into ABA and DMPD.
Spectrophotometrically, 1 mol of methyl red was reduced with 2 mol of NADH. This coincided with the expected stoichiometry for reductive cleavage of the azo bond of methyl red.
ACP Phosphodiesterase Activity in Vitro-acpD was origi-nally reported as the gene coding ACP phosphodiesterase. The phosphodiesterase activity was examined here by a method based on native PAGE analysis. For the assay, overexpressed and purified ACP was used as the substrate. The purified ACP was separated into two bands on native PAGE (Fig. 5, lane 1). N-terminal sequence analyses of the materials confirmed its identity as ACP. The lower band was confirmed to be that of holoACP because of the convergence of the upper band and the lower band with the ACP synthase reaction in the presence of CoA (Fig. 5, lane 3). ACP synthase transfers the 4Ј-phosphopantetheine moiety from CoA to Ser 36 of apoACP to produce holoACP. On the other hand, the apo type was confirmed to be the upper band by using a mutant ACP in which the Ser 36 residue was replaced by Cys. The mutant ACP only gave the upper band (Fig. 5, lane 4). The purified AcpD was subjected to the ACP phosphodiesterase assay as described (6). However, the reaction did not affect the ratio of apo-and holoACP, even in the presence of an excess amount of AcpD (Fig. 5, lane 2).
ACP Phosphodiesterase Activity in Vivo-An in vivo assay was also carried out to examine the function of AcpD. Because proteins other than the ACP-derived species were observed to remain in the stacker gel (23), crude ACP preparations could be analyzed by native PAGE (Fig. 6). The cell extract of E. coli overproducing AcpD (Fig. 6, lane 3) was compared with that of the control E. coli (Fig. 6, lane 4) by native PAGE. ACP was observed to be exclusively in its functionally active holo form regardless of AcpD overexpression. DISCUSSION To gain insights into the gene and properties of azoreductase, we purified native azoreductase from the E. coli JM109 strain and identified acpD as the gene encoding the azoreductase. Overexpression of the acpD gene allowed the isolation of more than 190 mg of the gene product (AcpD). AcpD was catalytically similar to the native azoreductase, and exhibited the same molecular weights under denatured and nondenatured conditions as those of the native azoreductase as well as N-terminal sequence identity.
Identification of the Reaction Product-HPLC analysis indicated that the decolorization of methyl red by AcpD is followed by cleavage of the molecule into colorless compounds, ABA and DMPD. The results proved that the decolorization of methyl red by AcpD was associated with reductive cleavage of the azo bond.
Kinetic Properties of AcpD-The decolorization of methyl red by AcpD was analyzed using double-reciprocal plots of initial velocity versus NADH or methyl red concentration. The parallel lines obtained with the assays suggest that, as with methyl red, the catalysis mode of AcpD is a ping-pong Bi-Bi mecha-  nism (Fig. 4). This mechanism suggests that FMN mediates electron-transfer from NADH to methyl red. This catalysis mode is generally observed for oxidoreductases containing a flavin prosthetic group such as quinone oxidoreductase. In view of the quinone (menadione) reductase activity observed for AcpD (Table III), the proposed catalysis mode is feasible.
Apparently, AcpD required 2 mol of NADH for the decolorization of 1 mol of methyl red. This stoichiometry implies that two cycles of the ping-pong mechanism were required for the cleavage (Scheme 1). In any case, the expression system reported here will provide a sufficient amount of azoreductase and greatly facilitate mechanistic studies on the reductive cleavage of azo dyes.
Comparison of AcpD with Other Enzymes-Previously, two azoreductases were purified from E. coli K12 (24). These enzymes, 12.5 and 28 kDa on SDS-PAGE, utilized NAD(P)H for Ponceau SX degradation. For AcpD, the molecular mass was 23 kDa on SDS-PAGE (Fig. 1), and NADPH was virtually ineffective as a cofactor. Furthermore, AcpD could not degrade Pon-ceau SX (Table II). Therefore, AcpD is different from the azoreductases of 12.5 and 28 kDa.
Orange II azoreductase (3) was purified from Pseudomonas KF46. The enzyme exists as a monomer with a molecular mass of 30 kDa, and there is no enzyme-bound FAD. Although this enzyme requires NAD(P)H as an electron donor for its activity, a flavin cofactor is not required. Because AcpD requires FMN for its activity (Table III), AcpD must belong to a different protein family from Orange II azoreductase.
Azoreductase from Bacillus sp. was reported recently (5). The NADPH-dependent azoreductase comprises 178 amino acids and contains an NADH-binding motif (GXGXXG). Although this molecular size is similar to that of AcpD, we could not find this motif or significant homology in AcpD. NAD(P)H:quinone acceptor oxidoreductase (NQO1, EC 1.6.99.2) is a FAD-containing enzyme that catalyzes the reduction of azo dyes (25) as well as quinones, and thus sequence comparison between AcpD and NQO1 from rat was performed. NQO1 consists of two separate domains (26): a major, "catalytic domain" (residues 1-220), and a small, "C-terminal domain" (residues 221-273). The catalytic domain is involved in the binding of the FMN moiety of FAD and the nicotinamide ribose of NADP ϩ , whereas the C-terminal domain is responsible for the binding of adenine riboses from FAD and NADP ϩ (26). AcpD exhibited moderate homology (Fig. 7) to the "catalytic domain" and highly conserved the residues involved in the hydrogen bonds with the FMN moiety and nicotinamide ribose moiety. On the other hand, a part corresponding to the

FIG. 7. Amino acid sequence alignment of AcpD and the catalytic domain (1-220 residues) of rat quinone reductase (NQO1).
Black and gray boxes indicate identical and similar amino acids, respectively. The similar amino acid groups are as follows: D and N; E and Q; S and T; K and R; F, Y, and W; L, I, V, and M. Arrowheads indicate the residues involved in the binding of the FMN moiety of FAD in rat NQO1 (26). Circles show the residues involved in the binding of nicotinamide ribose in rat NQO1 (26). SCHEME 1 C-terminal domain of NQO1 is lacking in AcpD. These observations corroborate the cofactor specificity of AcpD from a structural aspect and imply that AcpD belongs to the quinone oxidoreductase family.
ACP Phosphodiesterase Activity-ACP phosphodiesterase activity of AcpD was not detected in vivo or in vitro in this study. On the other hand, phosphodiesterase activity had been detected in vitro by monitoring the release of phospho-[ 3 H]pantetheine from labeled ACP in an earlier study (6). This discrepancy could be accounted for by misidentification of ACP phosphodiesterase. In the early study (6), the ACP phosphodiesterase was finally purified from a preparative SDS-PAGE gel by electroelution, because the phosphodiesterase remained active even after separation by SDS-PAGE. However, the purified fraction could be poor in authentic ACP phosphodiesterase but rich in AcpD because of the lower resolution on electroelution. Therefore, the protein band of ACP phosphodiesterase was misidentified as that of AcpD and the determined N-terminal sequence (NH 2 -Ser-Lys-Val-Leu-Lys-Ser-Xaa-Ile-Leu-Ala-Gly-Tyr-Ser-) was regarded as that of ACP phosphodiesterase. Moreover, the acpD gene was determined by another group (27) on the grounds of the correspondence of a translation sequence of the E. coli genome with the wrong N-terminal sequence of ACP phosphodiesterase. The function of the acpD gene has been demonstrated for the first time in this paper.
Despite the lack of ACP phosphodiesterase activity, genes homologous to acpD of E. coli have been found in many microorganisms, as described above. This wide distribution suggests the essential function of AcpD.
In conclusion, we have identified the gene responsible for azoreductase, constructed an efficient expression system, and biochemically characterized the enzyme. The results will contribute to the development of a biodegradation system for azo dyes via an understanding of the molecular mechanisms of azoreductase. To more accurately reflect its newly determined function, we propose the redesignation of the acpD gene as azoR.