6-Hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase and 6-oxocyclohex-1-ene-1-carbonyl-CoA hydrolase, enzymes of the benzoyl-CoA pathway of anaerobic aromatic metabolism in the denitrifying bacterium Thauera aromatica

Abstract

Benzoyl-CoA is a common intermediate in the anaerobic bacterial metabolism of many aromatic substrates. Two enzymes and ferredoxin of the central benzoyl-CoA pathway in Thauera aromatica have been purified so far. Benzoyl-CoA reductase reduces the aromatic ring with reduced ferredoxin yielding cyclohexa-1,5-diene-1-carbonyl-CoA [Boll, M. & Fuchs, G. (1995) Eur. J. Biochem.234, 921–933]. Dienoyl-CoA hydratase subsequently adds one molecule of water and thereby produces 6-hydroxycyclohex-1-ene-1-carbonyl-CoA [Laempe, D., Eisenreich, W., Bacher, A., & Fuchs, G. (1998) Eur. J. Biochem.255, 618–627]. Here two new enzymes, which convert this intermediate to the noncyclic product 3-hydroxypimelyl-CoA, were purified from T. aromatica and studied. 6-Hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase is an NAD⁺-specific β-hydroxyacyl-CoA dehydrogenase that catalyzes 6-hydroxycyclohex-1-ene-1-carbonyl-CoA + NAD⁺ → 6-oxocyclohex-1-ene-1-carbonyl-CoA + NADH + H⁺. 6-Oxocyclohex-1-ene-1-carbonyl-CoA hydrolase acts on the β-oxoacyl-CoA compound and catalyzes the addition of one molecule of water to the double bound and the hydrolytic C–C cleavage of the alicyclic ring, 6-oxocyclohex-1-ene-1-carbonyl-CoA + 2 H₂O → 3-hydroxypimelyl-CoA. The genes for both enzymes, had and oah, were cloned, had was overexpressed in Escherichia coli and the recombinant protein was purified. Hence, presumably all enzymes of the central benzoyl-CoA pathway of anaerobic aromatic metabolism from this organism have now been purified and studied and the corresponding genes have been cloned and sequenced.

Abbreviations

Bcr

benzoyl-CoA reductase

CV

column volume

Dch

cyclohexa-1,5-diene-1-carbonyl-CoA hydratase, dienoyl-CoA hydratase

dienoyl-CoA

cyclohexa-1,5-diene-1-carbonyl-CoA

enoyl-CoA hydratase

cyclohex-1-ene-1-carbonyl-CoA hydratase

enoyl-CoA

cyclohex-1-ene-1-carbonyl-CoA

Fdx

ferredoxin

GST

glutathione S-transferase

Had

6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase, β-hydroxyacyl-CoA dehydrogenase

IPTG

isopropyl thio-β-d-galactoside

Oah

6-oxocyclohex-1-ene-1-carbonyl-CoA hydrolase, β-oxoacyl-CoA hydrolase.

Aromatic compounds serve as substrates for microbial growth, under both aerobic and anoxic conditions. Aerobic pathways require molecular oxygen to introduce hydroxyl groups and to cleave the aromatic ring. The most common central intermediates of the aerobic pathways which are cleaved by dioxygenases are catechol (1,2-dihydroxybenzene), protocatechuate (3,4-dihydroxybenzoate) and gentisate (2,5-dihydroxybenzoate).

Anoxic metabolism uses different central intermediates, and in most of the cases studied the aromatic ring becomes reduced. Ring cleavage occurs hydrolytically after the aromatic substrate is reduced to an alicyclic compound with an oxo-group in the β position to a second carbonyl-group or to a thioester carbonyl-group. The most common central intermediate of the anoxic pathways is benzoyl-CoA. Two variants of the central benzoyl-CoA pathway have been found in the phototrophic bacterium Rhodopseudomonas palustris and in the denitrifying bacterium Thauera aromatica[1,2](Fig. 1). In both organisms benzoyl-CoA is reduced to a cyclic diene intermediate by an ATP-driven reaction catalyzed by benzoyl-CoA reductase (Bcr). The electron donor is a ferredoxin (Fdx) with low redox potential [5]. The main difference lies in the fact that in R. palustris the cyclic diene is reduced further, whereas in T. aromatica cyclohexa-1,5-diene-1-carbonyl-CoA (dienoyl-CoA) is hydrated to yield 6-hydroxycyclohex-1-ene-1-carbonyl-CoA. This reaction is catalyzed by dienoyl-CoA hydratase (Dch) [3].

Figure 1. Central benzoyl-CoA pathway of anaerobic aromatic metabolism as proposed for the denitrifying bacterium Thauera aromatica (β-subgroup of Proteobacteria) andthe phototrophic bacterium Rhodopseudomonas palustris (α-subgroup of Proteobacteria) [24]. (BcrCBAD) Benzoyl-CoA reductase, (redFdx) reduced ferredoxin, (Dch) dienoyl-CoA hydratase, (Had) 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (β-hydroxyacyl-CoA dehydrogenase), (Oah) 6-oxocyclohex-1-ene-1-carbonyl-CoA hydrolase (β-oxoacyl-CoA hydrolase), (BadDEFG) Benzoyl-CoA reductase, (redBadB) reduced ferredoxin, (BadK) enoyl-CoA hydratase, (BadH) 2-hydroxycyclohexane-1-carbonyl-CoA dehydrogenase, (BadI) 2-oxocyclohexane-1-carbonyl-CoA hydrolase.

Cell extract from T. aromatica converts benzoyl-CoA in the presence of ATP and a strong reductant to a noncyclic product, 3-hydroxypimelyl-CoA [1]. The conversion of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA to this final noncyclic product of the central benzoyl-CoA pathway in this microorganism requires at least two enzymes, a β-hydroxyacyl-CoA dehydrogenase and a hydrolytic ring cleavage enzyme. A gene cluster in T. aromatica that probably forms an operon and codes for the enzymes and electron donor of the central benzoyl-CoA pathway has been cloned and sequenced. It comprises the genes for the Fdx, the four subunits of benzoyl-CoA reductase, dienoyl-CoA hydratase and for two unknown proteins [4], that have yet to be studied. The deduced amino acid sequence of one of the unknown gene product, Had, has similarity with Zn-dependent alcohol dehydrogenases. This had gene has been proposed to code for β-hydroxyacyl-CoA dehydrogenase (Had). The other gene, oah, seems to code for the ring hydrolyzing enzyme, a β-oxoacyl-CoA hydrolase (Oah).

This study is aimed at purifying and studying the missing enzymes of the complete benzoyl-CoA pathway of anaerobic aromatic metabolism in T. aromatica.

Materials and methods

Materials and bacterial strains

Chemicals were obtained from Fluka, Merck, Sigma, Roth or Pierce; biochemicals were from Boehringer or Gerbu. [phenyl¹⁴C]Benzoate was obtained from MSD Isotopes. Materials and equipment for solid-phase extraction and FPLC were from ICT, Pharmacia or BioRad. HPLC equipment was from Waters, Merck, Perseptive and Grom. Plasmid pGex-6P-1 for the production of the recombinant proteins Oah and Had was obtained from Pharmacia. Glutathione S-linked agarose was purchased from Sigma. T. aromatica strain K172 (DSM 6984) was isolated in our laboratory [6] and has been deposited in Deutsche Sammlung von Mikroorganismen (Braunschweig, FRG). Azoarcus evansii strain KB 740 (DSM 6898) was isolated by Braun & Gibson [7]. R. palustris (type strain) (DSM 123) was obtained from Deutsche Sammlung von Mikroorganismen.

Growth of bacterial cells and preparation of cell extracts

T. aromatica and A. evansii were grown anoxically at 28 °C in mineral salt medium. Benzoate or acetate and nitrate served as sole sources of cell carbon and energy. The genes for benzoyl-CoA reductase, dienoyl-CoA hydratase, 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase and ring hydrolase are expressed during growth on the aromatic substrate. The substrates were continuously fed in a molar ratio of 1 : 3.6 (benzoate/nitrate) or 1 : 1 (acetate/nitrate) from a concentrated stock solution, pH 7.4, e.g. containing 0.5 m benzoate and 1.8 m KNO₃. In addition, T. aromatica was grown at 28 °C in mineral salt medium under aerobic conditions with benzoate. Cultivation, cell harvesting, storage and preparation of cell extracts were performed as described previously [6,8].

R. palustris was grown photoheterotrophically in mineral salt medium in the light under anaerobic conditions with benzoate. Cultivation, cell harvesting, storage and preparation of cell extracts were carried out under conditions described previously [6,9].

Synthesis, purification and HPLC analysis of CoA-thioesters

Benzoyl-CoA. Benzoyl-CoA was synthesized from CoA and benzoic acid anhydride under anaerobic conditions according to the general method of Schachter and Taggart [10]. CoA (200 µmol) was dissolved in 20 mL of 0.1 m sodium bicarbonate, pH 8. After addition of 500 µmol benzoic acid anhydride, the reaction mixture was held for 4 h at room temperature. During this time the pH was controlled and portions of 100–500 µL of 0.1 m sodium bicarbonate, pH 8, were added to maintain pH 8. The formation of the thioester was tested with the nitroprusside assay for CoA [11]. After completion of the reaction the CoA-thioester solution was acidified to pH 2 by addition of ≈ 0.5 mL of 1 m HCl. Contaminant material was extracted three times with diethylether (3 × 50 mL). After freeze-drying the reaction mixture was dissolved in 5 mL of 20 mm ammonium formate buffer, pH 3.5, containing 2% (by volume) methanol (solvent 1). For further purification the sample was applied to a solid-phase extraction column (ICT; end-capped C18-material, 10 g; reservoir volume, 60 mL; flow rate, 0.5 mL·min⁻¹), which had been equilibrated with the same solvent (20 °C). After washing the column with 120 mL solvent 1, benzoyl-CoA was eluted with 80% (by volume) aqueous methanol. After evaporation of methanol under vacuum at 30 °C, the solution containing benzoyl-CoA was freeze-dried and stored at −20 °C. Purity and amount were analyzed by comparison of the UV spectrum with published data (benzoyl-CoA, ε₂₆₁: 21 100 m⁻¹·cm⁻¹) [12]. The yield of the CoA-thioester after synthesis and purification was 70–80%.

[Phenyl-¹⁴C]benzoyl-CoA. [¹⁴C]-labeled benzoyl-CoA was synthesized enzymatically from [phenyl-¹⁴C]benzoate (specific radioactivity: 4.5 GBq·mmol⁻¹) and CoA with enriched benzoate-CoA ligase from T. aromatica[13]. The assay contained 700 µL 100 mm Tris/HCl, pH 7.8, 5 mm MgCl₂, 0.3 mm[phenyl¹⁴C]benzoate, 2 mm ATP, 1 mm CoA, 0.5 mm NADH, 2 mm phosphoenolpyruvate, 8 nkat myokinase, 8 nkat pyruvate kinase and 20 nkat lactate dehydrogenase. The reaction was started by adding 2 nkat enriched benzoate-CoA ligase. The formation of [phenyl¹⁴C]benzoyl-CoA was followed by HPLC using a RP-C18 column (Grom; Grom-Sil 120 ODS-4 HE, 5 µm; 120 × 4 mm) with 11% (by volume) acetonitrile in 50 mm potassium phosphate, pH 6.7 (20 °C; solvent 2), as solvent at a flow rate of 1 mL·min⁻¹. The effluent was monitored using a flow-through scintillation counter with a solid scintillator cell. After 40 min of incubation at 37 °C the reaction was stopped by stirring the reaction mixture on ice.

Cyclohex-1-ene-1-carbonyl-CoA. Cyclohex-1-ene-1-carbonyl-CoA was synthesized according to the method of Gross and Zenk [14] via the esterification of the corresponding free acid (5 mmol) with N-hydroxysuccinimide (5 mmol) in 25 mL dried dioxane by adding dicyclohexylcarbodiimide (5 mmol, dissolved in 5 mL dried dioxane). The insoluble reaction product was removed by filtration and the filtrate containing the succinimidylester was evaporated at 35 °C under reduced pressure. The transfer of the cyclo-1-hexenyl moiety from the succinimide ester to CoASH was performed under the same conditions as described above for unlabeled benzoyl-CoA using 200 µmol of CoA and 400 µmol of ester. The yield of the CoA-thioester after synthesis and purification was ≈ 70%.

2-Hydroxycyclohexane-1-carbonyl-CoA. 2-Hydroxycyclohexane-1-carbonyl-CoA was synthesized enzymatically from cyclohex-1-ene-1-carbonyl-CoA with purified enoyl-CoA hydratase under aerobic conditions at 37 °C. The 0.5-mL assay mixture contained 3.0 mm cyclohex-1-ene-1-carbonyl-CoA in 100 mm potassium phosphate, pH 7.4, and 1 nkat enoyl-CoA hydratase (fraction after Blue Sepharose) from T. aromatica[3]. The obtained mixture of cyclohex-1-ene-1-carbonyl-CoA and 2-hydroxycyclohexane-1-carbonyl-CoA was used as the stock solution for spectrophotometric tests with different cell extracts of T. aromatica and R. palustris.

Unlabeled and [¹⁴C]-labeled cyclohexa-1,5-diene-1-carbonyl-CoA and 6-hydroxycyclohex-1-ene-1-carbonyl-CoA. These compounds were synthesized enzymatically from benzoyl-CoA or [phenyl¹⁴C]benzoyl-CoA with partially purified benzoyl-CoA reductase from T. aromatica, which contained traces of dienoyl-CoA hydratase [3]. The assay was performed under strictly anaerobic conditions. The 1 mL assay mixture contained 150 mm Mops/KOH, pH 7.3, 10 mm MgCl₂, 0.5 mm unlabeled or [phenyl¹⁴C]benzoyl-CoA (specific radioactivity: 4.5 GBq mmol⁻¹), 0.75 mm ATP, 1 mm phosphoenolpyruvate, 8 nkat pyruvate kinase, and 8 mm titanium (III)-citrate. The reaction was started by the addition of 1–2 nkat of enriched benzoyl-CoA reductase. After 20 min incubation at 37 °C the reaction was stopped by adding 100 µL of 1 m HClO_4. After centrifugation the supernatant was freeze-dried and stored at −20 °C. For large-scale preparation the assay mixture (100 mL) contained 150 mm Mops/KOH, pH 7.3, 10 mm MgCl₂, 2 mm unlabeled benzoyl-CoA, 11 MBq of [phenyl¹⁴C]benzoyl-CoA, 3 mm ATP, 4 mm phosphoenolpyruvate, 40 nkat of pyruvate kinase and 8 mm titanium (III)-citrate. The reaction was started by the addition of 20 nkat enriched benzoyl-CoA reductase and was stopped with ≈ 8 mL of 2 m formic acid. After removal of precipitated protein by centrifugation the product mixture was applied to a solid-phase extraction column (ICT; end-capped C18-material, 50 g; flow rate, 2 mL·min⁻¹) which had been conditioned with 140 mL methanol and equilibrated with 280 mL of solvent 1 (20 °C). After washing the column with 350 mL of solvent 1, benzoyl-CoA was eluted with 80% (by volume) aqueous methanol. After evaporation of methanol under vacuum at 35 °C, the solution containing cyclohexa-1,5-diene-1-carbonyl-CoA and 6-hydroxycyclohex-1-ene-1-carbonyl-CoA was freeze-dried and stored at −20 °C. This CoA-thioester mixture was used for determination of β-hydroxyacyl-CoA dehydrogenase activity during purification of the enzyme. For separation of these CoA-thioesters the dry powder was dissolved in 6 mL 50 mm potassium phosphate, pH 6.7, and applied in three runs to a preparative HPLC column (Grom; Grom-Sil 120 ODS-4 HE, 11 µm; 250 × 20 mm) with solvent 2 at a flow rate of 8 mL·min⁻¹ (20 °C). Two product fractions were collected and applied for desalting to an Isolute-RP-C18 extraction column (ICT; end-capped C18-material, 1 g; reservoir volume, 6 mL; flow rate, 0.5 mL·min⁻¹). The column had been conditioned with 6 mL of methanol and equilibrated with 12 mL of solvent 1 (20 °C). After washing the column with 12 mL of solvent 1, the CoA-thioester was eluted with 80% (by volume) methanol in water. The samples were freeze-dried. The purified and desalted fraction containing 6-hydroxycyclohex-1-ene-1-carbonyl-CoA was used for determination of specific activity and apparent K_m value of alcohol dehydrogenase. The purity and amount of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA was analyzed by comparison of the UV spectra with published data of benzoyl-CoA [12] under the assumption that the nonaromatic CoA-thioesters had the same absorption coefficient at 260 nm as benzoyl-CoA. The amount of [¹⁴C]-labeled CoA-thioester was determined using liquid scintillation counting. For more rigorous control of purity, an analytic HPLC column (Grom; Grom-Sil 120 ODS-4 HE, 5 µm; 120 × 4 mm) was used with solvent 2 (20 °C), or with a linear 2–15% acetonitrile gradient formed from acetonitrile and 50 mm potassium phosphate, pH 6.7, as solvent at a flow rate of 1 mL·min⁻¹. The effluent was monitored using a radioactivity monitoring analyzer with a solid scintillator cell, and by a photodiode array detector. The yield of the CoA-thioester after synthesis and purification was ≈ 30–40%.

Determination of enzyme activities

Benzoate-CoA ligase activity. The benzoate-dependent, MgATP-dependent and CoA-dependent formation of AMP (which parallels benzoyl-CoA formation) was measured in a coupled spectrophotometric assay at 30 °C as described previously [13].

Benzoyl-CoA reductase activity. Benzoyl-CoA reductase activity was determined as described previously [8]. The continuous spectrophotometric assay followed the benzoyl-CoA-and MgATP-dependent oxidation of reduced methyl viologen. It was performed under strict anaerobic conditions in stoppered glass cuvettes at 37 °C.

Enoyl-CoA hydratase activity. Enoyl-CoA hydratase activity was determined as described previously [3]. The continuous spectrophotometric assay followed routinely the hydration of cyclohexa-1,5-diene-1-carbonyl-CoA to 6-hydroxycyclohex-1-ene-1-carbonyl-CoA or the hydration of crotonyl-CoA to 3-hydroxybutyryl-CoA; the reactions were associated with a decrease of absorption at 320 and 260 nm respectively.

6-Hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (b-hydroxyacyl-CoA dehydrogenase, Had) activity. Had activity was determined in a continuous spectrophotometric assay at 37 °C and followed the 6-hydroxycyclohex-1-ene-1-carbonyl-CoA-dependent reduction of NAD⁺. The assay contained 100 mm potassium phosphate, pH 7.4, 1 mm NAD⁺ and 0.3 mm 6-hydroxycyclohex-1-ene-1-carbonyl-CoA or 0.5 mm CoA-thioester mixture with cyclohexa-1,5-diene-1-carbonyl-CoA and 6-hydroxycyclohex-1-ene-1-carbonyl-CoA in equal amounts [3] (see synthesis, purification and HPLC analysis of CoA-thioesters). The enzyme activity was recorded at 365 mm (ε for NADH = 3400 m⁻¹·cm⁻¹).

6-Oxocyclohex-1-ene-1-carbonyl-CoA hydrolase (β-oxoacyl-CoA hydrolase, Oah) activity. Oah activity was determined by discontinuous HPLC analysis of time-dependent substrate consumption and product formation. A 200-µL assay mixture contained 100 mm potassium phosphate, pH 7.4, 0.5 mm unlabeled or [phenyl¹⁴C]-labeled 6-hydroxycyclohex-1-ene-1-carbonyl-CoA (specific radioactivity: 4.5 GBq·mmol⁻¹), 1.0 mm NAD⁺, 1.5 mm sodium pyruvate, 10 nkat lactate dehydrogenase and 0.1 nkat recombinant Had. The reaction mixture was preincubated for 20 min at 37 °C, and the hydrolase reaction was then started by the addition of 10 µL of concentrated flow-through fractions from Zn⁺-chelating Sepharose chromatography containing Oah. For HPLC analysis samples of 30 µL were retrieved after 2, 5, 10, 20 and 40 min of incubation at 37 °C and adjusted to pH 4.0 by adding 3 µL 10% HCOOH (by volume). After centrifugation, the supernatants were applied to an analytical HPLC column (Grom; Grom-Sil 120 ODS-4 HE, 5 µm; 120 × 4 mm) which was equilibrated with 2% acetonitrile in 50 mm potassium phosphate, pH 6.7 (20 °C). HPLC separation of the reaction mixture was performed with a linear 2–15% acetonitrile gradient formed from acetonitrile and 50 mm potassium phosphate, pH 6.7. [Phenyl¹⁴C]-labeled compounds were quantified from the values obtained with a flow-through radioactivity monitor.

Purification of benzoyl-CoA reductase

Purification was performed at 4 °C under strictly anaerobic conditions in a glovebox with N₂/H₂ (95:5, by volume) as gas phase. All buffers contained 0.25 mm dithionite and 1 mm dithioerythritol as reducing agents. Preparation of benzoyl-CoA reductase started with extracts from 200 g cells of T. aromatica (wet mass). The purification procedure was performed in four steps including anion-exchange chromatography on DEAE–Sepharose, Mono-Q, chromatography on hydroxyapatite and gel filtration [8]. The enzyme was enriched 40-fold with a yield of 25% and had a specific activity of 0.48 µmol benzoyl CoA·min⁻¹·(mg protein)⁻¹.

Copurification of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (Had) and ring hydrolyzing β-oxoacyl-CoA hydrolase (Oah)

As both enzyme activities were oxygen stable, purification was performed at 4 °C under aerobic conditions. Preparation started with extract from 20 g cells of T. aromatica (wet mass).

DEAE–Sepharose chromatography. Extract (36 mL of 100 000 g supernatant) was applied to a DEAE–Sepharose column (Pharmacia, fast flow; diameter, 3.0 cm; volume, 85 mL) that had been equilibrated with 20 mm triethanolamine hydrochloride/KOH, pH 7.8, 10% (by volume) glycerol (referred to as buffer A) at a flow rate of 3 mL·min⁻¹. The column was washed with 1 bed volume of buffer A and with 3 bed volumes of 40 mm KCl in buffer A. Had activity was eluted with 80 mm KCl in buffer A in an volume of 160 mL.

Hydroxyapatite chromatography. The combined fractions from DEAE–Sepharose chromatography were applied in four runs to a FPLC column (diameter, 16 mm; volume, 20 mL) of Macro-Prep (BioRad; ceramic hydroxyapatite with diameter 40 µm) that had been equilibrated with buffer A at a flow rate of 2 mL·min⁻¹. The column was washed with 1 bed volume of buffer A and with 2 bed volumes of 600 mm KCl in buffer A. After reequilibration with buffer A (2 bed volumes) Had was eluted with a linear 0–80 mm potassium phosphate gradient (4 bed volumes) formed from buffer A and 100 mm potassium phosphate, pH 7.8, containing 10% glycerol. The Had eluted at 30–40 mm potassium phosphate in a 32-mL volume.

Blue Sepharose chromatography. The combined fractions containing Had activity were applied in 1-mL portions to a Blue Sepharose CL-6B affinity column (Pharmacia; volume, 1.5 mL; diameter, 5 mm). The column was equilibrated with 20 mm triethanolamine hydrochloride/KOH, pH 7.0, 10% (by volume) glycerol with a flow rate of 0.25 mL·min⁻¹. The column was washed with 1.5 mL of equilibration buffer and 4.5 mL of 120 mm KCl in equilibration buffer. Had activity was eluted with 0.5 mm NAD⁺ in 100 mm potassium phosphate, pH 7.0 (4.5 mL) in a volume of 3.5 mL. This fraction contained both Had and β-oxoacyl-CoA hydrolase.

Separation of Had and Oah

Separation of these two enzymes was accomplished by a Chelating Sepharose column (Pharmacia, fast flow; volume, 1.0 mL; diameter, 5 mm) which had been equilibrated with 0.44 mL 50 mm ZnCl₂, pH 3 and washed with 2 mL H₂O and 4 mL 100 mm KCl in 20 mm triethanolamine hydrochloride/KOH, pH 7.8, 4 mm MgCl_2, 10% (by volume) glycerol (referred to as buffer B). The 3.5-mL fractions from Blue Sepharose were applied at a flow rate of 0.25 mL·min⁻¹. After washing with 3 mL 500 mm KCl in buffer B Had activity was eluted with 100 mm glycine in 20 mm triethanolamine hydrochloride/KOH, pH 7.0, 4 mm MgCl₂ and 10% (by volume) glycerol in a 3.0-mL volume. Oah was not bound by Zn⁺-chelating Sepharose, the enzyme was found in the flow-through fractions.

Determination of native molecular mass

The native molecular mass of the enzymes was estimated using a FPLC Superdex 200 HR 10/30 gel filtration column (Pharmacia; diameter, 10 mm; volume, 24 mL), which had been equilibrated with 100 mm KCl containing 20 mm triethanolamine hydrochloride/KOH, pH 7.0 and 10% (by volume) glycerol; 100 µL of concentrated protein solution (50 µm) was applied at a flow rate of 0.2 mL·min⁻¹.

HPLC-analysis and UV spectra of substrates and products

A special 100-µL assay mixture was used for this purpose containing 100 mm potassium phosphate, pH 7.4, 0.2 mm unlabeled or [phenyl¹⁴C]-labeled 6-hydroxycyclohex-1-ene-1-carbonyl-CoA (specific radioactivity: 4.5 GBq·mmol⁻¹) and 0.2–0.4 mm NAD⁺. The reaction was started by the addition of 0.1 nkat of purified Had (fraction after Chelating Sepharose chromatography) or 0.3 nkat of Had after Blue Sepharose (contained also Oah activity). After 10 min of incubation at 37 °C the reaction was stopped by stirring the reaction mixture on ice. The assay was also performed with a NAD⁺ regenerating system consisting of 0.2 mm NAD⁺, 0.4 mm sodium pyruvate and 10 nkat lactate dehydrogenase. The conditions for HPLC analysis of substrates and products of Had and Oah reactions were as follows: samples of 40 µL were retrieved and directly applied to an analytical HPLC column (Grom; Grom-Sil 120 ODS-4 HE, 5 µm; 120 × 4 mm). HPLC separation of the reaction mixture was performed with a linear 2–15% acetonitrile gradient formed from acetonitrile and 50 mm potassium phosphate, pH 6.7. Substrate and products were monitored using a radioactivity monitoring analyzer with a solid scintillator cell as well as by a photodiode array detector. Retention times: 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, 20.5 min; Had product, 18.5 min; Oah product, 10.0 min.

Electrospray mass spectroscopy of substrates and products

A 100-µL enzyme assay was used that contained 100 mm potassium phosphate, pH 7.4, 1.8 mm unlabeled 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, 1 mm NAD⁺, 2 mm pyruvate and 20 nkat lactate dehydrogenase. The reaction was started by the addition of 0.2 nkat of purified Had (fraction after Chelating Sepharose chromatography) or 0.6 nkat of Had after Blue Sepharose (contained also Oah activity). After 35 min of incubation at 37 °C the reaction was stopped by stirring the reaction mixture on ice. For desalting of the obtained CoA-thioesters the samples were applied to an analytical polystyrene column (Perseptive; Porus R2, 20 µm; 100 × 4.7 mm) with a linear 0–15% acetonitrile gradient formed from acetonitrile and 0.1% trifluoroacetic acid (by volume), pH 2.0, as solvent at a flow rate of 5 mL·min⁻¹. The effluent was monitored using a photodiode array detector. The collected fractions were freeze-dried. Electrospray mass spectroscopy analysis of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and of products of Had and Oah was performed on a Finnigan TSQ700 mass spectrometer with electrospray interface. The dry powder containing the products was dissolved in 20 µL 30% aqueous methanol (by volume) and 0.1% trifluoroacetic acid (by volume), pH 2.0; 5-µL aliquots of this sample were applied directly to the mass spectrometer. In case of the Oah product the mass was detected in the product peak obtained after HPLC separation using a RP-C18 column (Vydac; 5 µm; 150 × 8 mm) directly coupled to the mass spectrometer. The Oah product was eluted by a linear gradient of 0–25% acetonitrile in 0.1% trifluoroacetic acid (by volume) at a flow rate of 20 µL·min⁻¹ provided by a dual-piston pump 140B of Applied Biosystems.

Construction of gene-expressing vectors for T. aromatica Had and Oah genes

The complete had gene was amplified from the T. aromatica chromosome via PCR using the primers hadBamf (5′-GTT CAG TCG GAT CCT GGA GGA GTG-3′) and hadEcorev (5′-GCT TAA AAG GAA TTC CGT TCA GGG-3′), which contained BamHI and EcoRI restriction sites, respectively. The 1065-bp PCR product was cut with both enzymes and inserted into BamHI/EcoRI-cut pGEX-6P-1 creating pGexhad. It contained the had gene as a fusion to the glutathione S-transferase (GST) gene from Schistosoma japonicum under the control of a tac promoter. Directly upstream of the 5′-end of the had gene a recognition sequence for PreScission protease which itself is a fusion of GST with human Rhinovirus 3C protease was located. By the same token, the complete ring hydrolase gene was amplified from the T. aromatica chromosome via PCR using the primers oahBamf (5′-CAC AGG AGT TCG GAT CCA TGA ATC CGA C-3′) and oahXhorev (5′-GGGTGC TGG CTC GAG CTC ATC ACT TGG C-3′), which contained BamHI and XhoI sites. The 1131 bp PCR product was cut with both enzymes and ligated into BamHI/XhoI-cleaved pGex-6P-1 creating pGexoah. This constructed plasmid contained the same properties as described above. Both constructed plasmids were checked by complete DNA sequence determination of the inserts.

Had and Oah expression and purification

The GST-fusion vectors containing the genes for Had and Oah were transformed into E. coli BL21(DE3) [15]. Further treatment continued analogously. Bacteria were grown at 30 °C to an D_578nm of 1.0 in 500 mL batches of Luria–Bertani media containing ampicillin (50 µg·mL⁻¹) and induced by the addition of 100 µm isopropyl thio-β-d-galactoside (IPTG). Cultures were incubated for an additional 3 h and bacteria were subsequently harvested by centrifugation. The bacterial pellet was resuspended in 50 mL of NaCl/P_i (150 mm/10 mm) containing 1 mm PhCH₂SO₄F and disrupted with a French Press (3 × 137 MPa). Cell debris was removed by centrifugation for 45 min at 23 000 g. The resulting supernatant was directly loaded onto a GST-linked agarose column (1 × 10 cm) which was previously equilibrated with 5 column volumes (CV) of NaCl/P_i. Unbound protein was removed by washing the column with 5 CV of NaCl/P_i. GST-fusion protein was eluted with elution buffer containing 50 mm Tris/HCl, pH 8.0, 10 mm reduced glutathione. GST-fusion protein-containing fractions were pooled and dialyzed overnight at 4 °C against PreScission protease cleavage buffer (50 mm Tris/HCl, pH 7.0, 150 mm NaCl, 1 mm EDTA and 1 mm dithiothreitol). Digestion of the fusion protein was performed by adding 2 units of PreScission protease for each 100 µg of fusion protein and incubating for 5 h at 4 °C. Removal of the cleaved GST-tag and traces of uncleaved fusion protein was achieved by applying a GST-linked agarose column. As expected, all recombinant native T. aromatica Had and Oah was detected in the flow-through while the protease, the removed GST-tag and uncleaved GST-fusion remained on the column. The activity of the recombinant protein was tested as described above. Purity of the recombinant proteins was analyzed on 12% SDS/PAGE.

SDS/PAGE

SDS/PAGE (12.0 or 12.5% polyacrylamide) was performed as described by Laemmli [16]. Proteins were visualized using Coomassie blue staining [17].

Other methods

Protein was routinely determined by the BCA Protein Assay Reagent (Pierce). In addition, the method of Bradford [18] using BSA as standard and the Lowry method [19] were used. N-terminal protein sequences were determined as described recently [20].

Results

β-Hydroxyacyl-CoA dehydrogenase activities acting on 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and 2-hydroxycyclohexane-1-carbonyl-CoA in different bacteria

Extracts of T. aromatica cells grown under different conditions were analyzed for the capacity to oxidize 6-hydroxycyclohex-1-ene-1-carbonyl-CoA. Activity was found in the soluble cell fraction after anaerobic growth with benzoate (Table 1). Enzyme activity was specific for NAD⁺; no reaction was observed with NADP⁺. When cells were grown aerobically with benzoate the activity was 20 to 25-fold lower. This indicates that expression of the enzyme activity is under negative control by oxygen. Cells grown anaerobically with the nonaromatic substrate acetate contained one-seventh of the enzyme activity suggesting that gene expression is positively controlled by an intermediate of the central benzoyl-CoA pathway. The enzyme activity in the closely related denitrifying bacterium A. evansii when grown with benzoate and nitrate was similar to T. aromatica. R. palustris also contained substantial amounts of this enzyme activity. However, NADH formation ceased soon after starting the reaction by adding substrate. Therefore the specific activity in Table 1 represents only a minimal value. The rapid leveling-off of the reaction measured in the direction of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA oxidation with NAD⁺ may be due to several reasons. Extracts may contain high thioesterase activity which hydrolyzes the CoA thioester substrate. The equilibrium of the reaction is far on the substrate side. Also, the ring hydrolyzing enzyme activity may not be high enough under the experimental conditions to remove the product 6-oxocyclohex-1-ene-1-carbonyl-CoA rendering the overall reaction irreversible. It is to be expected that the equilibrium constant of the NAD⁺-dependent oxidation of the β-hydroxyacyl-CoA compound is far on the substrate side. As NAD⁺ was virtually stoichiometrically reduced in extracts of the denitrifying bacteria, it is concluded that the cell extract contained an enzyme that further metabolizes the product of the reaction (see below).

Table 1. Specific activities of β-hydroxyacyl-CoA dehydrogenase(s) acting on 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and 2-hydroxycyclohexane-1-carbonyl-CoA in different gram-negative bacteria and after growth with different substrates. The activity was determined in spectrophotometric assays at 37 °C, pH 7.4, as described in Materials and methods. The cell extract is the 100 000 g supernatant.

			Specific dehydrogenase activity (nmol·min^−1·mg^−1)
Cell extract	Growth substrates	Electron acceptor	6-hydroxycyclohex-1-ene-1-carbonyl-CoA	2-hydroxycyclohexane-1-carbonyl-CoA
a Of different preparations of cell extracts.
T.  aromatica	Benzoate, HNO3	NAD^+	650–750^a	21
		NADP^+	< 1	< 1
	Benzoate, O2	NAD^+	30	6
	Acetate, HNO3	NAD^+	105	100
A.  evansii	Benzoate, HNO3	NAD^+	820	29
R.  palustris	Benzoate, anaerobic (in the light)	NAD^+	≥ 120	≥ 40

Similarly, the capacity to oxidize 2-hydroxycyclohexane-1-carbonyl-CoA was assayed. The substrate was formed from cyclohex-1-ene-1-carbonyl-CoA by purified enoyl-CoA hydratase. β-Hydroxyacyl-CoA dehydrogenase activity acting on the 2-hydroxycyclohexane compound was present in both T. aromatica and A. evansii cells grown anaerobically with benzoate, but at 30-fold lower activity compared with dehydrogenase activity acting on the 6-hydroxycyclohex-1-ene compound. Interestingly, cells of T. aromatica grown anaerobically with acetate contained high activity (Table 1). The enzyme was present in R. palustris, but exact determination of the specific activity in cell extract was not possible for reasons discussed above. Perrotta and Harwood [20] reported a specific activity of 30 nmol· min⁻¹·mg⁻¹ protein, and of 280 nmol·min⁻¹·mg⁻¹ for the reverse reaction.

Copurification of β-hydroxyacyl-CoA dehydrogenase and β-oxoacyl-CoA hydrolase from T. aromatica

6-Hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (Had) and 6-oxocyclohex-1-ene-1-carbonyl-CoA hydrolase (Oah) were purified from cell grown anaerobically with benzoate and nitrate (Table 2). Since both enzyme activities were oxygen stable, purification was performed under aerobic conditions. After three chromatography steps which included a Blue Sepharose affinity column with NAD⁺-depended elution, a colorless protein fraction was obtained that showed two distinct bands of relative molecular mass (M_r) 38 000 and 40 000 on SDS/PAGE (Fig. 2). Amino-acid sequencing revealed that the N-terminus of the M_r 38 000 protein exactly corresponded to the predicted sequence of the β-hydroxyacyl-CoA dehydrogenase, Had, the N-terminus of the 40 kDa protein to the putative β-oxoacyl-CoA hydrolase, Oah. This fact explained that NADH formation was stoichiometric and the reaction went to completion when a limiting amount of the β-hydroxyacyl-CoA substrate was applied to the assay, no matter which enzyme fractions were used. Evidently the hydrolase removed the product of the dehydrogenase in an irreversible reaction.

Table 2. Purification of β-hydroxyacyl-CoA dehydrogenase (Had) and β-oxoacyl-CoA hydrolase (Oah) from extracts of T. aromatica grown anaerobically with benzoate and nitrate. The two enzymes were separated only in the last step, Chelating Sepharose chromatography. Since the substrate for β-oxoacyl-CoA hydrolase was not available in large amounts only the specific activity of the purified enzyme will be given. The values in the individual fractions refer to β-hydroxyacyl-CoA dehydrogenase activity (NAD⁺). ND, not determined.

Purification step	Protein (mg)	Activity (µmol·min^−1)	Specific activity (µmol·min^−1·mg protein^−1)	Purification (-fold)	Yield (%)
^a 100 000 g supernatant.
Cell extract^a	1476	1116	0.7	1	100
DEAE-Sepharose	431	833	1.7	2.7	75
Hydroxyapatite	113	379	3.4	4.8	34
Blue Sepharose	22	193	8.7	12.4	17.3
Chelating Sepharose
-Dehydrogenase	4.9	58	11.8	16.8	5.2
-Hydrolase	5.2	≥ 6.8	≥ 1.3	n. d.	ND

Figure 2. SDS-PAGE of selected fractions obtained during purification of Had and Oah. (1) Oah and Had after Blue Sepharose. (2) Oah after Chelating Sepharose. (3) Had after Chelating Sepharose. (4) Molecular mass standard proteins. For details of purification and assays see Materials and methods.

Purification and properties of β-hydroxyacyl-CoA dehydrogenase (Had)

Separation of the two M_r 38 000 and 40 000 enzymes was accomplished by Chelating Sepharose (Fig. 2). The purified Had was composed of one single subunit of M_r 38 000. The native molecular mass determined by gel filtration was 78 000 ± 10 000 suggesting an α₂ composition of the native enzyme. The UV–visible spectrum showed only the protein band at 280 nm. Hydroxyacyl-CoA dehydrogenase activity was insensitive to 10 mm EDTA, 1 mm 2,2′-bipyridyl, or 1 mm pyrazole. Activity was not increased after adding 2–20 mm of ZnCl_2.

The equilibrium constant of the reaction was determined by HPLC separation and UV detection at 260 nm plus ¹⁴C-radiodetection of substrates and products. When equimolar concentrations (200 µm) of [phenyl¹⁴C]6-hydroxycyclohex-1-ene-1-carbonyl-CoA and NAD⁺ were applied equilibrium concentrations of 186 µm NAD⁺ and 14 µm NADH and 186 µm[¹⁴C]-labeled 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and 16 µm[¹⁴C]-labeled product (see below) were obtained. Hence the equilibrium constant corrected for H⁺ concentration at pH 7.4, K′, was 6 × 10⁻³.

Because of the unfavorable equilibrium constant the catalytic properties of the enzyme were investigated with the Blue Sepharose fraction; this fraction also contained Oah, which irreversibly removes the product of Had. The enzyme was specific for NAD⁺, no reaction was observed with NADP⁺. Similarly, the enzyme was specific for the substrate, no activity was observed with 2-hydroxycyclohexane-1-carbonyl-CoA. Therefore, the dehydrogenase activity with 2-hydroxycyclohexane-1-carbonyl-CoA observed with cell extract must be due to another dehydrogenase. Optimal pH was 7.2–7.4, the apparent K_m value determined at 3.5 mm concentration of NAD⁺ was 60 ± 20 µm.

Purification and properties of β-oxoacyl-CoA hydrolase (Oah)

As shown below, the product of Had was 6-oxocyclohex-1-ene-1-carbonyl-CoA. This compound was apparently the substrate of the ring hydrolyzing enzyme. Oah was tested by following the consumption of [¹⁴C]-labeled substrate and the formation of [¹⁴C]-labeled product from 6-oxocyclohex-1-ene-1-carbonyl-CoA by HPLC, followed by UV and radiodetection. The specific enzyme activity determined at pH 7.4, 37 °C and 0.5 mm[¹⁴C]-labeled substrate was 1.3 µmol·min⁻¹·mg⁻¹. However, this value seems to be underestimated because of high activity losses during Chelating Sepharose chromatography. This assumption was made due to the fact that assays using the enzyme fractions before this step never showed the accumulation of 6-oxocyclohex-1-ene-1-carbonyl-CoA, when 6-hydroxycyclohex-1-ene-1-carbonyl-CoA was added as a substrate of Had. Since these fractions contained both Had and Oah activities, the specific activity of Oah must have been at least as high as that of Had (compare with Table 2), i.e. >11 µmol·min⁻¹·mg⁻¹.

The enzyme was separated from the dehydrogenase by Chelating Sepharose chromatography (Fig. 2). The enzyme consisted of one M_r 40 000 subunit, the native molecular mass was 170 000 ± 10 000. This suggests that the native enzyme has an α₄ composition.

Substrates and products of Had and Oah

Neither the free acids nor the CoA thioesters of the alicyclic compounds, which served as the substrate of the two enzymes, nor labeled compounds were commercially available. Therefore all compounds were synthesized enzymatically starting from unlabeled or [¹⁴C]-labeled benzoyl-CoA and were purified by HPLC and solid-phase extraction. Their UV spectra are shown in Fig. 3. Figure 4 shows the [¹⁴C] elution profile of different assays to which purified [¹⁴C]6-hydroxycyclohex-1-ene-1-carbonyl-CoA (0.2 mm) was added. In Fig. 4A, the assay before addition of Had is shown. In Fig. 4B, the equilibrium concentrations of substrate and product after completion of the reaction are shown. This assay contained 0.2 mm NAD⁺ initially. In Fig. 4C, the Had reaction was assayed with a NAD⁺ regenerating system. Clearly the complete conversion of substrate (1) to product (2), 6-oxocyclohex-1-ene-1-carbonyl-CoA, can be seen. This is explained by pulling the reaction forward due to removal of the product NADH. In Fig. 4D, the only product (3) obtained after addition of Oah to assay (Fig. 4B) is shown. Obviously the dehydrogenase reaction in Fig. 4B is pulled in the unfavorable direction by presence of Oah, which removes the product in an irreversible reaction forming 3 hydroxypimelyl-CoA (Fig. 1).

Figure 3. Ultraviolet spectra of substrates and products of Had and Oah. (A) 6-hydroxycyclohex-1-ene-1-carbonyl-CoA. (B) 6-oxocyclohex-1-ene-1-carbonyl-CoA. (C) 3-hydroxypimelyl-CoA.

Figure 4. HPLC separation of substrate [phenyl¹⁴C]6-hydroxycyclohex-1-ene-1-carbonyl-CoA and of derived labeled products. (A) 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, 0.2 mm. (B) Equilibrium concentrations obtained when purified Had, 0.2 mm 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and 0.2 mM NAD⁺ were initially added. (C) Products obtained when NAD⁺ was continuously regenerated. (D) Products obtained when purified Had and purified Oah were added; 0.2 mm 6-hydroxycyclohex-1-ene-1-carbonyl-CoA and 0.4 mm NAD⁺ were initially added. The figures show the [¹⁴C]-elution profiles.

The identification of the compounds (1)–(3) in Fig. 4 was based on mass spectroscopy (data not shown). The substrate of Had, 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, has been characterized previously using NMR techniques [3]. It showed a mass peak of 892.4 (theoretical value: 892.2), the product of Had a mass of 890.2, the product of Oah a mass of 926.1. These latter two values are identical or very close to the expected molecular masses of 6-oxocyclohex-1-ene-1-carbonyl-CoA (890.2) and of 3-hydroxypimelyl-CoA (926.2), respectively. Therefore the reactions catalyzed by Had and Oah are 6-hydroxycylohex-1-ene-1-carbonyl-CoA + NAD⁺→ 6-oxocyclohex-1-ene-1-carbonyl-CoA + NADH/H⁺, and 6-oxocyclohex-1-ene-1-carbonyl-CoA + 2 H₂O → 3-hydroxypimelyl-CoA, respectively.

Purification of the recombinant Had and Oah from E. coli

The genes for Had and Oah have been cloned and sequenced previously [4]. Sequencing of the first 10 N-terminal amino acids confirmed the identity of the corresponding genes. The N-terminus of Oah was identical with the deduced sequence, in case of Had the N-terminal methionine was missing. The properties of Had and Oah from T. aromatica are shown in Tables 3 and 4, respectively.

Had and Oah were stable under aerobic assay conditions. Therefore, expression of recombinant GST-fusion proteins was performed in the presence of oxygen. No predominant overexpression of the two recombinant proteins could be seen, when the extracts were analyzed on a 12% SDS/PAGE (not shown). A single affinity chromatographic step for the purification of the soluble GST-tagged protein was employed using a reduced glutathione agarose chromatography. When the soluble fractions containing the recombinant protein were applied onto the column, all nonfusion proteins passed through. Two GST-fusion proteins of M_r 64 000 and 34 000, which eluted with reduced glutathione, were the GST-tagged Had and presumably Oah, respectively. After cleavage of the GST-tag (M_r 26 000) with precission protease, chromatography on GST-linked agarose was repeated, and the two recombinant proteins could be obtained almost pure in the flow-through. While the recombinant Had showed an expected band of M_r 38 000 on a 12% SDS/PAGE, recombinant Oah showed a band of M_r 8000; hence the GST–Oah fusion protein contained only an M_r 8000 fragment of Oah. Both enzyme activities were analyzed under the same conditions as the proteins isolated from extracts of T. aromatica. Already the M_r 64 000 recombinant fusion protein GST–Had was capable of converting 6-hydroxycyclohex-1-ene-1-carbonyl-CoA to 6-oxocyclohex-1-ene-1-carbonyl-CoA with a specific activity of 25.8 µmol·min⁻¹·mg⁻¹ fusion protein (15.6 µmol·min⁻¹· mg⁻¹ protein part of M_r 38 000 Had). When the GST-tag was cleaved with protease the Had-protein showed an activity of 6.3 µmol·min⁻¹·mg⁻¹ purified protein. The conversion of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA to 6-oxocyclohex-1-ene-1-carbonyl-CoA could be clearly attributed to the enzyme Had as analyzed by HPLC (Fig. 4). As expected, the recombinant M_r 8000 Oah-fragment was enzymatically inactive in the Oah assay and most likely was produced by action of an endogenous protease in E. coli from the native Oah-fusion protein.

Discussion

We have purified and studied two new enzymes of the benzoyl-CoA pathway of anaerobic aromatic metabolism in T. aromatica,β-hydroxyacyl-CoA dehydrogenase (Had) and β-oxoacyl-CoA hydrolase (Oah), respectively. These enzymes were identified on the basis of their N-terminal amino acid sequence and the genes were found next to the genes coding for benzoyl-CoA reductase. A gene cluster in T. aromatica that probably forms an operon and codes for the enzymes and electron donor of the central benzoyl-CoA pathway has been cloned and sequenced [4]. Our results confirm that there are two variants of the central benzoyl-CoA pathway in T. aromatica and R. palustris (Fig. 1).

Had is the third enzyme in the anaerobic benzoyl-CoA pathway and catalyzes the NAD⁺-dependent oxidation of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, the product of dienoyl-CoA hydratase [3], to 6-oxocyclohex-1-ene-1-carbonyl-CoA. The properties of this purified enzyme are summarized in Table 3. The amino acid sequence of Had showed similarities with Zn-dependent long-chain alcohol dehydrogenases, e.g. benzyl alcohol dehydrogenase II of Pseudomonas putida (XylW, D63341). Because of the insensitivity of the β-hydroxyacyl-CoA dehydrogenase activity for EDTA, bipyridyl and pyrazole, it may not be a metal-binding protein. The residue proposed to act as a proton donor in the NAD⁺-dependent oxidation of β-hydroxyacyl-CoA was His47 as shown for benzyl alcohol dehydrogenase [22]. R. palustris contains a NAD⁺-dependent dehydrogenase, the badH gene product. BadH is a short-chain alcohol dehydrogenase, which has been shown to oxidize the secondary alcohol 2-hydroxycyclohexane-1-carbonyl-CoA to 2-oxocyclohexane-1-carbonyl-CoA, the putative ring-cleavage substrate R. palustris[23,24]. The low enzyme activity with 2-hydroxycyclohexane-1-carbonyl-CoA given in Table 1 does not represent the actual value, but a minimal specific activity. The enzyme assay with cell extract was severely impaired. In accordance with the differences of the metabolic pathway of benzoyl-CoA, the two β-hydroxyacyl-CoA dehydrogenases in the two organisms (Had and BadH) show no significant amino acid sequence similarity. The catalyzed reactions of Had in T. aromatica and BadH in R. palustris are shown in Fig. 1.

Table 3. Properties of 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (Had) from T. aromatica.

Property	Dehydrogenase
Substrates	6-hydroxycyclohex-1-ene-1-carbonyl-CoA
	NAD^+
Products	6-oxocyclohex-1-ene-1-carbonyl-CoA
	NADH
Specific activity	with 6-hydroxycyclohex-1-ene-1-carbonyl-CoA, 11.8 µmol·min^−1·mg^−1
Apparent Km	60 ± 20 µm, 6-hydroxycyclohex-1-ene-1-carbonyl-CoA
pH optimum	7.2–7.4
Native molecular mass	78 000 ± 10 000
Subunit Mr	38 000
Suggested composition	α2
Catalytic number per	7 s^−1
subunit

Oah is the ring-cleaving enzyme in the central benzoyl-CoA pathway in T. aromatica. Purified Oah catalyzed the conversion of the alicyclic 6-oxocyclohex-1-ene-1-carbonyl-CoA to the aliphatic 3-hydroxypimelyl-CoA. This implies the existence of two partial reactions, enoyl-CoA hydratase and carbon–carbon bond cleaving hydrolase activity, both catalyzed by Oah. A proposed reaction mechanism is shown in Fig. 5. The Oah gene, oah, follows upstream of the dienoyl-CoA hydratase gene and showed 34% identity and 48% similarity to the corresponding badI gene in R. palustris. The function of badI has been deduced from the N-terminal amino acid sequence of the purified protein that catalyzed hydrolytic carbon ring opening of 2-oxocyclohexane-1-carbonyl-CoA producing pimelyl-CoA [2]. Differences in ring cleavage substrates are reflected in differences in the genes. The T. aromatica oah gene is 45% longer than the R. palustris badI gene, with an additional 35 bp domain at the N-terminus, an additional central 51 bp domain and an additional 27 bp domain at the C-terminus. This may reflect the need of the T. aromatica enzyme to act as an enoyl-CoA hydratase, in addition to a ring hydrolyzing enzyme. The deduced amino-acid sequences of the common part of the sequence share only ≈ 34% identity. Both oah in T. aromatica and badI in R. palustris showed sequence similarities with the dihydroxynaphthoic acid synthase from various organisms [25]. This synthase catalyzes the formal reaction of the hydrolase, i.e. the formation of the aromatic ring in naphthoate synthesis from the precursor o-succinylbenzoyl-CoA [26].

Figure 5. Proposed reaction mechanism of theconversion of 6-oxocyclohex-1-ene-1-carbonyl-CoA to 3-hydroxypimelyl-CoA, probably catalyzed by Oah.

We did not succeed in overexpressing oah in E. coli to prove that the enzyme catalyzes in fact two additions of water. If a contaminant enoyl-CoA hydratase would have caused the addition of one molecule of water to the double bond, one would have expected the formation of the corresponding β-hydroxy compound, which was not observed. However, this possibility cannot be ruled out completely.

Presumably all enzymes of the central benzoyl-CoA pathway in T. aromatica have been studied. Downstream of the T. aromatica gene cluster coding for proteins of the anaerobic benzoyl-CoA pathway two ORFs were found whose function is unknown [4]. They show no strong similarity to known proteins and a homologous version is found in R. palustris, as well as in A. evansii. This coincidence may suggest that these gene products play a role in anaerobic benzoate metabolism, possibly in regulation [4,24].

Acknowledgements

This work was supported by the Deutsche Forschungsgemeinschaft, the Fonds der Chemischen Industrie and the Landesschwerpunktprogramm ‘Biokatalysatoren’. Thanks are also due to P. Hörth and Prof. W. Haehnel, Freiburg, for mass spectroscopical analysis and to Prof. H. Schägger, Frankfurt, for amino acid sequencing. Thanks are due to Prof. W. Buckel, Marburg, for helpful discussions and suggestions concerning the ring cleavage enzyme.

Footnotes

1. Enzymes: benzoyl-CoA reductase (dearomatizing) (EC 1.3.99.15); 6-hydroxycyclohex-1-ene-1-carbonyl-CoA dehydrogenase (EC 1.1.1.-); 6-oxocyclohex-1-ene-1-carbonyl-CoA hydrolase (EC 3.7.1.-).