Enzymatic reduction of benzoyl-CoA to alicyclic compounds, a key reaction in anaerobic aromatic metabolism

Eur. J. Biochem. 205, 195-202 (1992) $2 FEBS 1992

Enzymatic reduction of benzoyl-CoA to alicyclic compounds, a key reaction in anaerobic aromatic metabolism Jurgen KOCH and Georg FUCHS Angewandte Mikrobiologie, Universitat Ulm, Federal Rcpublic of Germany

(Received October 2Y/December 16, 1991) - EJB 91 1453

Different anaerobic bacteria can oxidize a variety of aromatic compounds completely to C 0 2 via one common aromatic intermediate, benzoyl-CoA. It has been postulated that anaerobically the aromatic nucleus of benzoyl-CoA becomes reduced. An oxygen-sensitive enzyme system is described catalyzing the reduction of benzoyl-CoA to trans-2-hydroxycyclohexanecarboxyl-CoA in a denitrifying Pseudomonas species grown anaerobically on benzoate plus nitrate. The assay mixture consists of cell extract, [U-'4C]benzoyl-CoA, a [U-'4C]benzoyl-CoA-generating system (consisting of [U-' 4C]benzoate, purified benzoate-CoA ligase, Mg2 +-ATP, coenzyme A), an ATP-regenerating system (consisting of phosphoenolpyruvate, pyruvate kinase, myokinase), and a low-potential reductant [titanium(III) citrate]. The optimal pH is about 7, the specific activity 10 nmol benzoyl-CoA reduced min-' x mg-' protein. The apparent K, for benzoyl-CoA is below 50 pM, Five major products were found. One product is cyclohex-I -enecarboxyl-CoA which must have been formed by a benzoyl-CoA reductase. The other product is probably trans-2-hydroxycyclohexanecarboxyl-CoA rather than the cis-stereoisomer ; this product must have been formed by a cyclohex-I -enecarboxyl- CoA hydratase. Two other products are likely to be intermediates of benzoyl-CoA reduction to cyclohex-I -enecarboxyl-CoA, suggesting that the reduction reaction is more complex. An early formed fifth product is more polar than cyclohexanecarboxyl- or cyclohex-1-enecarboxyl-CoA.

The enzyme system is under oxygen control since it was not found in cells grown aerobically on benzoate. It is induced by aromatic compounds since its activity is low in cells grown anaerobically on acetate. The actual inducer is probably benzoyl-CoA rather than benzoate. This conclusion is drawn from the fact that the system is also present in cells grown anaerobically on phenol, phenylacetate, 4-hydroxybenzoate, or 2-aminobenzoate; the anaerobic metabolism of these compounds has been shown in this organism to proceed directly via benzoyl-CoA rather than via free benzoate.

The degradation of aromatic compounds by aerobic microorganisms has been studied in some detail; in all cases, molecular oxygen is required for the oxygenase-catalyzed attack on the stable aromatic ring structure (for review see [l]). However, it is generally ignored that many aromatic compounds can be completely mineralized to CO, by different anaerobic bacteria in the absence of molecular oxygen (for reviews see [2, 31). Benzoate served as the best studied model

Corrqondence to G. Fuchs, Angewandte Mikrobiologie, Universitat Ulm, Postfach 4066, D-7900 Ulm, Federal Republic of Germany

Enzymes. Benzoyl-CoA reductase (aromatic ring reducing) (EC 1.3.99. -); benzoate-CoA ligase (AMP-forming) (EC 6.2.1.25); phloroglucinol reductase (EC 1.3.99. -); resorcinol reductase (EC 1.3.99. -); cyclohex-I-enecarboxyl-CoA hydratase (EC 4.2.1. -); DNase I (EC 3.1.21.1); 'phenol carboxylase' (EC 4.1.1.-); 4- hydroxybenzoyl-CoA reductase (dehydroxylating) (EC 1. - . - . -); 4-hydroxybenzoate-CoA ligase (AMP-forming) (EC 6.2.1. -); phenylacetate-CoA ligasc (AMP-forming) (EC 6.2.1.21); benzoylformate (phenylglyoxylate) : acceptor oxidoreductase (EC 1.2.99. -); toluene dehydrogenasc (methylhydroxylating) (EC 1 . 1 7.99. -); 4-cresol de- hydrogcnase (methylhydroxylating) (EC 1.17.99.1); bcnzyl alcohol dchydrogenase (EC 1.1.1.90); benzaldehyde dehydrogenasc (NADP') (EC 1.2.1.7): 2-aminobenzovl-CoA rcductase (deaminat-

, I

ing) (EC 1. -. -. -); 2-aminobenzoate-CoA ligase (AMP-forming) (EC 6.2.1.-).

compound. Evans and coworkers [4,53 had proposed already in 1968 that the anaerobic attack on the aromatic ring should be by reduction. This proposal was corroborated by circum- stantial evidence based on studies with resting cells or cultures of phototrophic or nitrate-reducing bacteria grown anaerobically on benzoate [2,6 - 141. It was shown that they convert- ed benzoate to cyclohexanecarboxylate or cyclohex-I -enecarboxylate. Later, it became evident that benzoyl-CoA must be the actual intermediate of anaerobic metabolism, not only of benzoate but also of many other aromatic compounds [I 5 - I91 except for those with two or more /I-hydroxyl groups [20 - 221. Despite many efforts in different laboratories, the crucial enzyme catalyzing the postuhted reduction of benzoyl-CoA could not be demonstrated in vitro. We report for the first time on the enzyme system which catalyzes the reduction of the aromatic ring of benzoyl-CoA to cyclohex-I -enecarboxyl- CoA. This was demonstrated in extracts of a denitrifying Pseudomonas strain which is able to grow without oxygen in the presence of nitrate on benzoate, 2-aminobenzoate, 4- hydroxybenzoate, toluene, phenol, phenylacetate, or p-cresol. We have shown in this organism that all the aromatic growth substrates mentioned are metabolized anaerobically to C 0 2 via one common aromatic intermediate, benzoyl-CoA, as il- lustrated in Fig. l .

196

p a GOOH COOH

Phenyl- acetic acid

Q OH

6 Phenol

OH p-Cresol

ATP AMP OH . PP

4-OH Benzoic . . .

,SCoA g NH*

aC120'& 2H

HC=O Benzoic acid

COOH CoA f H 2H Y

OH H20/2H @ N H c A F L

- non-aromatic products

/ Anthranilic acid

Toluene

Fig. 1. Proposed central position of benzoyl-CoA in the anaerobic degradation of different aromatic compounds by the denitrifying Pseudomonas sp. K 172. Note that this organism is able to grow not only in the absence of oxygen, but also in the presence of oxygen on many, though not all, of thc compounds menlioned. The anaerobic and the aerobic pathways are fundamentally different and both are strictly regulated. Enzymcs of phenol and 4-hydroxybenzoate metabolism: 'Phenol carboxylase' (EC 4.1.1. -), 4-hydroxybenzoate-CoA ligase (AMP-forming) (EC 6.2.1. -), 4-hydroxybenzoyl-CoA reductase (dehydroxylating) (EC 1. -. - . -). Enzymes of phenylacetate metabolism : phcnylacetate-CoA ligase (AMP-forming) (EC 6.2.1.21), an enzyme system oxidizing phenylacetate to benzoylformate, benzoylformatc (phenylglyoxy- 1ate):acceptor oxidorcductase (EC 1.2.99. -). Enzymes of 2-aminobenzoate metabolism: 2-aminobenzoate-CoA ligase (AMP-forming) (EC 6.2.1. - ). 2-aminobenzoyl-CoA reductasc (deaminating) (EC 1. - , - . -). Enzymes ofp-cresol mctabolisrn: 4-cresol dehydrogenase (methylhydroxylating) (EC 1.17.99. I ) , (4-hydroxy) benzaldehyde dehydrogenase (EC 1.2.1.7). Enzymes of toluene metabolism: toluene dehydrogenase (mcthylhydroxylating) (EC 1.17.99.-)), benzyl alcohol dehydrogenase (EC 1.1.1 .90), benzaldehyde dehydrogenase (NADP+ J (EC 1.2.1.7). En7ymes of benzoate metabolism : benzoate-CoA ligase (AMP-forming) (EC 6.2.1.25). Benzoyl-CoA reductase (aromatic ring reducing) (EC 1.3.99. -). the enzyme studied here, is suggested to be common to the degradation of all of these compounds (see Fig. 6).

MATERIALS AND METHODS

Materials

P.seudomonu.s strain K 172 was isolated in our laboratory [ 2 3 ] . Acetohacteriunz vcoodii DSM 1030" was obtained from Deutsclie Samrnlung vnn Mikroorganismen (Braunschweig). Chemicals were obtained from Fluka (Neu-Ulm, FRG), Aldrich-Chemie (Steinheim, FRG), Heraeus (Karlsruhe, FRG), or Merck (Darinstadt, FRG).

Biochemicals were from Boehringer (Mannheim, FRG) ; radioisotopes from ARC (American Radiolabeled Chemicals Inc./Biotrend Chemikalien GmbH, Koln, FRG); TLC plates and HPLC column from Merck (Darmstadt, FRG); scintillation cocktail rotiszint 2200 from Roth (Karlsruhe, FRG); gases from Linde (Hollriegelskreuth, FRG); viologen dyes from Serva (Heidelberg. FRG).

Growth of bacteria and preparation of cell extracts

Pseudomomis strain K 172 was grown at 30°C under anaerobic conditions in mineral salts medium, with 5 mM aromatic acid and 20 mM nitrate as sole sources of energy and cell carbon. The more toxic substrate phenol was applied at 3 mM concentration and supplied by repeated feeding. Growth de- termination, cell harvesting and storage, and preparation of cell extracts were as described [23, 241. A . woodii was grown anaerobically on fructose [25].

I n vitro assay of benzoyl-CoA reduction

The enzymatic reduction of benzoyl-CoA was studied in vitro under strictly anaerobic, reducing conditions at 30 C. The assay mixture consisted of cell-free extract (10000 x g supernatant) of Pseudomonas K 172; a [U-'4C]benzoyl-CoA- generating system consisting of [U-'4C]benzoate, MgZ +-AT€'. coenzyme A, and purified benzoate-CoA ligase from Pseudomonus KB 740 [15]; an ATP-regenerating system consisting of phosphoenolpyruvic acid, pyruvate kinase and myokinase; and a reducing system. The reducing system consisted either of hydrogen gas phase plus extract of A . wooclii (0.7 mg protein) containing 200 nkat methylviologen-reducing hydrogenase activity; or of different chemical reductants Ctitanium(II1) citrate [26], sodium dithionite, or sodium borohydride} at 3.5 mM concentration. The anaerobic assay mixture was prepared at 4"C, preincubated for 30 min at 30"C, and the reaction was started by adding cell-free extract of Psrudornonas K 172. After different incubation periods, samples were withdrawn. For the analysis of CoA-thioesters samples were acidified with H2S04 to pH 2 and centrifuged (20 min, 10000 xg, 4°C). For the analysis of thioester-bound acids samples were first treated with KOH (pH 12, 20 min, 8OVC), then acidified with H 2 S 0 4 to pH 2 , and centrifuged (lOOOOxg, 20 min, 4 T ) . The routinely used assay mixture (0.35 ml total volume) contained: 150 mM Mops/KOH pH 7 . 2 , 200 pM [U-14C]benzoate (130 kBq), 2 mM MgC12, 1.7 mM ATP, 1 mM dithioerythritol, 1.1 mM coenzyme A. 8

197

nkat benzoate-CoA ligase (3.5 pg protein), 3.5 mM titanium(II1) citrate, 3.3 mM phosphoenolpyruvic acid, 12 nkat myokinase (2 pg), 100 nkat pyruvate kinase (20 pg) and 50 pl cell-free extract (2 mg protein). In the early stages of this investigation a modified standard assay was used. It differed from the standard assay in that 150 mM sodium borate pH 8.5 was used instead of Mops buffer and that H2, A . woodii extract, and methyl viologen were used, in addition to Ti(II1).

The formation of [U-14C]benzoyl-CoA in the assays was controlled. At the end of the experiment, an aliquot of the assay mixture was acidified and centrifuged. The residual free beni-oate in the supernatant was analyzed by TLC and autoradiography and the amount of I4C in the benzoate area was determined. This amount of I4C was compared with I4C in the benzoate area of samples which were treated with KOH. The difference is due to ['4C]benzoyl-CoA.

Chromatographic separation, detection, preliminary identification, and quantitation of reaction products

[U-'4C]Benzoate and labelled products derived thereof were analyzed by TLC or HPLC using the following conditions. TLC was carried out on aluminum plates (20 cm x 20 cm) with 0.2mm silica gel (Kieselgel 60 FLS4; Merck, Darmstadt, FRG). The following solvent systems were used: (A) benzene/dioxan/acetic acid (8: 1 : 1, by vol.); (B) isopropyl ether/butanol (75 : 25, by vol.). TLC on reversed-phase RP- 18 F 2 5 4 ~ 0.25-mm glass plates (Merck) was performed with mcthanol/l50 mMpotassiumphosphatepH 3.5 (3:7, byvol.).

HPLC was carried out on a reversed-phase CIS- Lichrospher column (100 CH-18/2, 25 x 0.4 cm) from Merck (Darmstadt, FRG) with the following solvent systems: (A) 30% methanol, 70% 50 mM potassium phosphate pH 4; (B) 13% methanol, 87% 50 mM potassium phosphate pH 4. The flow rate was 1 ml x min- '. The separated organic acids were detected with a L-3000 photo diode array detector (Merck/ Hitachi, Darmstadt, FRG).

Syntheses The following compounds were not commercially avail-

able; they were synthesized and the products were charac- terized by ultraviolet spectroscopy, melting point (m. p.) and/ or boiling point (b. p.), and C, H analysis.

Cyclohex-1-enecarboxylic acid, which only recently became commercially available (Heraeus), was synthesized starting from 16.4 g cyclohexene by the method of Treibs and Orttmann [27]. Yield: 17% of theory; b.p. 138°C /1.87 kPa (lit. 136.5-137.5"C/1.87 kPa). m.p. 38°C. Later, the compound was prepared by alkaline hydrolysis of its ethyl ester.

cis-2-H ydroxycyclo hexanecarboxylic acid (cis-hexahydro- salicylic acid) was synthesized starting from 3 g 2-cyclohexanonecarboxylic acid analogously to the method for the synthesis of cis-4-hydroxycyclohexanecarboxylic acid [28]. Yield: 86% of theory; m.p. 78 "C (lit. 79 - 80 "C).

trans-2-H ydroxycyclohexanecarboxylic acid was synthesized starting from 3 g 2-cyclohexanonecarboxylic acid according to the method of Gardner et al. [29]. Yield: 46% of thcory; m.p. 11 1 "C (lit. I l l "C ).

2-0xocyc~ohexanecarboxy~ic acid was obtained starting from 5 g of its ethyl ester [30]. Yield: 93% of theory.

Purification of benzoate-CoA ligase Benzoate-CoA ligase was purified from denitrifying

Pseudomonas KB 740 cells which were anaerobically grown on

200 - I 150 Y. v

7 - 100 0 0

50 m u

0

A - B 'i zoo ,---l--l v

U 0 150

100 5 0

50 0 N

: o 0 2 4 6 8 1 0 0 5 10

Time ( rn in ) Protein (mg)

C D

6 7 8 9 0 2 4 6 8 1 0

PH [T i ( l l l ) ci t rate] (mM)

Fig. 2. In vitro reduction of benzoyl-CoA to alicyclic products. (A) Time dependence of benzoyl-CoA consumption at different concentrations of titanium(ll1) citrate. The initial benzoate concentration was 200 pM. Titanium(II1) citrate concentration: ( 0 ) 0.1 mM; (V) 0.5 mM; (B) 1 mM; ( A ) 3 mM; (+) 5 mM; (0) 10 mM. Protein, 6 mg x ml- assay. (B) Dependence of the initial rate of benzoyl-CoA consumption on the amount ofprotein added/ml assay. The consumption ofbenzoate after 5 min of incubation is given. (C) Dependence of the initial rate of benzoyl-CoA consumption on the pH. The following buffers were used: Mops/KOH (pH 6.6-7.2), Tris/Cl (pH 7.0- 8.5). diethanolamine/Cl (pH 8.7- 9.3); pH 5.8 unbuffered. Bcnzoate consumed after incubation for 5 rnin ( 0 ) and 10 (A) is given at a protein concentration of 6 mg x ml-' assay. (D) Dependence of thc initial specific benzoate consumption rate on thc Ti(II1) concentration. The standard assay described in Matcrials and Methods was used except for A in which exogenous benzoate-CoA ligase was omit- ted.

benzoate and nitrate [15]. The purified enzyme had a specific activity of 2.5 pkat x mg-' protein.

Analytical methods

Protein was determined by the Bradford method [31]. 14C was determined by liquid scintillation counting using external standardization. Radioactive areas on TLC plates were located by autoradiography using X-ray film (Kodak X-Omat, Xar-352, Sigma Chemical Co., St. Louis, USA); radioactive spots were scratched off, and the TLC material was directly extracted with 4 ml scintillation cocktail. Organic standard acids were detected by 254-nm light, or by spraying [32] (a) with 1% (mass/vol.) vanillin in conc. H2S04, (b) with 5% (mass/vol.) potassium dichromate in conc. H,SO, or (c) with 0.1 % (massivol.) bromocresol green in ethanol.

RESULTS

Enzymatic reduction of ('4C]benzoyl-CoA in vitro

Extracts from cells of Pseudomonas K 172, which were grown anaerobically with benzoate and nitrate as sole carbon and energy sources, catalyzed the reductive conversion of [U-

198

Fig. 3. Analysis by two different TLC systems and autoradiography of the 14C-labelled products of (14Clbenzoyl-CoA conversion. (A, C) Separation by TLC system A. (B) Separation by TLC system B. In experiments shown in (A) and (B) the modified standard assay was used. (C) In this experiment the standard assay was used. Numbers 6 and 7 refer to products that were not found in the modified standard assay. (A, B) Track 1. schematic representation of the Rf values of reference compounds: (a) benzoate, (b) cyclohexanecarboxylate, (c) cyclohex-I-enecarboxylate, (d) 2-oxocyclohexanecarboxylate, (e) cis-2-hydroxycyclohexanecarboxylate, (0 rruns-2-hydroxycyclohexanecarboxylate, (g) pimelate, (h) benzoyl-CoA. Track 2, [U-14C]benzoate. Track 3, assay mixture after 15 min of incubation, without alkaline hydrolysis. Tracks 4- 11, assay mixture after 15 s, 30 s, 60 s, 2 min, 5 min, 10 min, 15 rnin and 30 min of incubation, after alkaline hydrolysis. Numbers 1-5 refer to the five major products found with this method; B refers to benzoate. For TLC systems and assay conditions see Materials and Methods.

14C] benzoyl-CoA to five major radioactive products which were probably all non-aromatic and accounted for 80-90% of the radioactivity added. The reaction was analyzed after alkaline hydrolysis of the coenzyme A thioesters [33, 341 in the assay mixture. The radioactive acids were separated by TLC in two solvent systems, detected by autoradiography, and the substrate and the products were quantitated from the amount of 14C in the radioactive spots. In the standard assay benzoyl-CoA was consumed at a specific rate of approxi- mdtely 10 nmol x min-' x mg-' protein; the reaction pro- ceeded linearly with time in the range of 0 - 5 rnin (Fig. 2A) and the rate was linearly dependent on the amount of protein added in the range of 0 - 6 mg protein per ml assay (Fig. 2 B). The pH optimum of benzoyl-CoA reduction determined in Mops/KOH, Tris/HCl and diethanolamine/HCl buffer was around 7 (Fig. 2C). In the standard 0.35-ml assay (0.2 mM (U-'4C]benzoyl-CoA, 2 mg protein, 3.5 mM Ti(II1) as reductant) the aromatic substrate was almost completely metabolized within 5 min. With respect to Ti(II1) concentration, the standard assay was suboptimal but was used for practical reasons in order to allow convenient sampling. The dependence of the reaction on Ti(II1) was strict, and the opti-

mal concentration was 10 mM or higher (Fig. 2A, D). The labelled products were still coenzyme A thioesters, the free acids were released only after alkaline hydrolysis of the thioester bond (Fig. 3A, B). The reduction of benzoyl-CoA to benzaldehyde and benzyl alcohol was therefore excluded. Five major 14C-labelled products, designated spots 1 - 5 in Fig. 3, were detected by different TLC and HPLC systems. One of the reaction products, spot 1, was preliminarily identi- fied as cyclohex-1-enecarboxylic acid (see below). This indi- cated that the aromatic nucleus of benzoyl-CoA was enzy- matically reduced to cyclohex-1-enecarboxyl-CoA. Whether this four-electron reduction was due to one four-electron- transferring or two two-electron-transferring enzymes is not known.

Requirements of the system

The enzymatic transformation of benzoyl-CoA required strictly anaerobic conditions and a low potential reducing agent; routinely titanium(II1) citrate was used. Another convenient reducing system consisted of hydrogenase plus H2 and methyl viologen; hydrogenase reduces methyl viologen and is

199

Table 1. Requirements of the enzymatic conversion of benzoate via activation to benzoyl-CoA and reduction to alicyclic compounds. Depen- dence of benzoate consumed after 5 min and 10 min of incubation in the standard assay on the individual assay components. The amount of [14C]benzoate, which was not activated as coenzyme A thioester, was controlled at the end of the experiment.

Assay Benzoate consumed Benzoate pre- after incubation for

5min 10min iment

sent at the end of the exper-

Complete - ATP -ATP regenerating

system, but plus ATP - exogenous benzoate- CoA ligase from Pseudomonas KB70 -coenzyme A -Ti(IlI) -extract - dithioerythritol

Complete, but aerobic

152 114 < 1 <1

158 166

188 194 15 38 19 31 4 11

148 I81 6 11

1 3 39

2

1 134

5 2 3 4

present at very high specific activity in extracts of A . woodii. Extract from A . woodii was required in catalytic amounts only and did not itself catalyze any benzoyl-CoA transformation. Other reducing agents such as sodium dithionite and sodium borohydride could also be used as reductants but they proved to be less effective. The dependence of the reaction on the individual test components is shown in Table 1. The assay contained a [U-'4C]benzoyl-CoA-generating system consisting of [U-'4C]benzoate, Mg2+-ATP, coenzyme A, and purified benzoate-CoA ligase (AMP-forming), and an ATP- regenerating system. The requirement for ATP and CoA may be indirect since they were necessary for benzoyl-CoA formation. Although ATP alone was sufficient, an ATP-regenerating system was used when the assay conditions were modified in order to avoid ATP limitation due to ATPases. For similar rcasons dithioerythritol and purified benzoate-CoA ligase were routinely included although they did not stimulate the standard assay: dithioerythritol stabilizes the oxygen-sensitive enzyme; the endogenous benzoate-CoA ligase in the extract may become limiting, for example when extracts from cells were tested which were grown on substrates other than benzoate

Kinetics of benzoyl-CoA consumption and of formation of five major products

The time course of benzoate consumption and formation of the five major products is shown in Fig. 4. According to the time course of benzoate consumption, the concentration at which half-maximal rate was observed was approximately 50 pM benzoyl-CoA, suggesting that the apparent K , for benzoyl-CoA is in the same range. The formation of products was analyzed in sodium borate pH 8.5 using a reducing system consisting ofboth 3.5 mM Ti(ll1) and H2/hydrogenase (Figs 3 and 4). Since 80-90% of the radioactivity was recovered and the product pattern did not vary within one sample, the hydrolysis products were reasonably stable. Only product 2 was found to be slightly volatile.

500

- 400 m

v

m 0 3

-c

=J 300 a e C

2.

.- c 200 .- c 0 0 0 71

of

.- 0 100

i 500

[L

0 0 10 20 30 60 120

Time (min)

Fig. 4. Time course of ['4C]benzoyl-CoA consumption and formation of the five major products which are also CoA thioesters. The radioactive spots on TLC plates were scratched off and 14C was determined by liquid scintillation counting. (+) Benzoate, corresponding to compound (a) in Fig. 3; (0 ) spot 1 = cyclohex-1-enecarboxylate, corrc- sponding to compound (c) in Fig. 3; (V) spot 2, (0) spot 3, (A) spot 4, (0) spot 5 = tuans-2-hydroxycyclohexanecarboxylate, corresponding to compound (0 in Fig. 3. The modified standard assay, as described in Materials and Methods, was used.

Products 2,3, and 4 appeared to be early products whereas products 1 and 5 appeared to be formed later. Some minor bands were seen on the TLC plates; most of them were located between the substrate spot and product spot 4. Since they appeared late (10 - 15 min) and contained little radioactivity they were neglected in the kinetics in Fig. 4. The main product pattern was reproducible, independent of the pH of the assay mixture and the source of extract (for cells grown anaerobically on different aromatic compounds see below). However, when borate buffer was replaced by Mops buffer, two additional radioactive products, spots 6 and 7, appeared (Fig. 3 C) which were more polar than trans-2-hydroxycyclohexanecarboxylic acid. It cannot be decided whether borate prevented the formation of these products or altered their chromatographic behaviour. Borate is known to react with vicinal hydroxyl groups.

Although the product pattern was the same in all tests the distribution of radioactivity in the individual spots varied depending on the test conditions (buffer, reducing agent, pH) and on the batch of cells.

Preliminary identification of the reaction products

Since most of the putative products of benzoyl-CoA reduction and further metabolism were not commercially available, some of them were synthesized and used for the preliminary identification of the products based on cochroma- tography. One of the reaction products after alkaline hydrolysis, spot 1, cochromatographed on TLC with authentic cyclohex-1 -enecarboxylic acid. When I4C was eluted f rom spot 1 and chromatographed by HPLC together with authentic cyclohex-1-enecarboxylic acid, radioactivity and standard coeluted. An HPLC separation of the radioactive substrate (benzoate) and products is shown in Fig. 5 A.

A further product of the enzymatic reduction of benzoyl- CoA (spot 5 ) was also formed in the first few minutes, but

A 1 3 50

I

0 : o 20 30 40

l i m e ( m i r )

1

0 10 2 0 30 40

TNme ( m l n )

B G 32

c sc 0 10 20 30 40 50

- i v e (nsn)

50 n

0 10 20 30 40 50

idme (mmn)

Fig. 5. Corhromatography uf "C-labelled products on IIPLC with nuth- rntic refcrence compounds. (A) Cochromaiography of' one of the luhcllcd pioduct.; corrcxponding to spot 1 with authentic cyclohcx-l- eiisc.irboxylatc (solvciit system A); 1. ben/oatc. I I . cyclohcx-l- cnccarho\!latc (€3) ('(,chromatography of one of the Iahelled product\ corresponding to spot 5 with /ron.\-2-hydroxycSclohcxane- carhoxyl;itc (solvent system B): 111. rran.s-2-hydroxyr~clohcxane- ciirhoxyldtc: 11'. c.;.\-~-hydrc~xyc!clohc~.anec.irboxylate. I:or solvent systems scc M;itcrials and Methods.

accumul;ited and finally accounted for halfofthe radioactivity in the products; in contrast. the amount of l4C. in all the other product.; levclcd off early or dccreased: this presumably secondary ''C-labelled product cochromatographed with ituthentic rrcr~r.c-2-h~drc~sycyclohexanecarboxylic acid rather than with the us-isomer on 'I'LC. Spot 5 was scratched off to identify the corresponding product in the HPLC' run. This

Tablr 2. k:ffect of different growth substrates mid ox!gcn un tho actkit! of the benxoyl-CoA redwtasc system. Extracts wcrc prepared from cell.; harbcstcd in the cxponcntial growth phase. I he st;indartl .tssa! was used.

ExtrJcts from cclls Bcn/oatc consumed Bcnzoatc prc- grown on aftcr incubation for sent at thcencl

01' the cxper- 2 . 5 min 5 min imcnt

knzoatc . anaerobically 54 100 1 BcnLoatc.. aerobicall! < I < I X I Acetate, anaerobically < I 2 5

anacrobicallv 1 0 74 I

anaerobically 28 64 2

anaerobically 62 120 1

Phenol. .inaerobically 5 0 117 17 4-H ydrox) bewoatc.

Phsn ylacctate.

2-Aininohcnzoate.

"l'he iiicreascd amount of benzoate in thc cxpcriment with aerobically grown cell5 miiy be due to Inhihition of bcnzoatc-Co:\ ligiisc or due to 1ncrc;lsed ihiocsteraw activity. Both the inhibitory effect of unknown compounds in the extract and diffcrcnt Ievcls or thiocstera\e wcrc observed in thi\ organism.

product also cochromatopraphed on HPLC with r r m v - 2-hydroxycyclohcxanecarboxylic acid rather than c i s - 2 - hydroxycarboxylic acid (Fig. 5 B). The kinetics of its formation indicates that i t originatcd from c)clohcx-l -cnecarboxyl- CoA via water addition irons to the doublc bond. catalyzed by a sepitratc cnzyme. Further oxidation to 2-oxocyclohcsanc- carhoxyl-CoA was not ohservcd under the reducingconditions o f the assay. All of the products shown (spots 1 - 5 ) could bc extracted aftcr acidification into dicthylcther with an avcragc yield ol'about 90°/b. Spots 2,3, and 4 did not cochroinatograph with any of the following standards: 2-oxocyclohexanc- carhoxylic acid. ( is-2-hydroxycyclohcsanecarbo~ylic acid. cyclohcxanecarboxylic acid. pimclic acid

Regulatory properties of benzoyl-CoA reductase (aromatic ring reducing)

The formation of tran.\.-'-hydroxScyclohex;iIiccarhox) I - CoA from bcnzoyl-CoA requires at least two cnzymes. ii benzoyl-CoA reductose (aromatic ring reducing) (EC' 1.3.09. - and a cyclohcx-1-enecarboxyl-CoA hydratasc (K 4.2.1. - ): the former enzyme system inay consist of more thiiii one enzyme. The synthesis of the postulated new enLyme system bcnzoyl-CoA reductasc (aromatic ring reducing). which wiis

measured as benmyl-CoA consumption. appears to be undcr tu ofold regulatory control ('l'ablc 2). Hcnzo! I-CoA reduction was not detected in cells grown under aerobic conditions o n benzoate and was almost not detcctable in cells growii under anaerobic conditions on acetate plus nitrate. This indicates thiit the enzcme H';IS undcr anaerobic control m d \ias induced by benzoate or bcnzoyl-CoA. Not only anaerobic benyoate- grown cells contaiticd this en7yme activity. but also cells grown anaerobically with nitrate on such diverse compounds as phenol. 4-hydrosybcnzoatc. phcnylacctatc. or 2-aminoben7oatc. Thcse compounds, however, have in common that thcy are metabolized iinacrobically via bcnyoyl-CoA rather than ben- zoiite (Fig. 1). Since benzoic acid is not an intermediate 111

their dcgradiition. thc actual inducer of henzovl-CoA rc-

20 1

a C

,SCoA ,SCoA c=o c-0

f d

SCoA SCoA

Fig. 6. Initial reactions of anacrobic benzoyl-CoA degradation via rc- duction of the aromatic ring by Pseudomonas sp. K 172. Note that cyclohexanecarboxyl-CoA appears to be a side product rather than an intermediate of aromatic ring reduction. (a) Benzoyl-CoA, (c) cyclohex-1-enecarboxyl-CoA, (b) cyclohexanecarboxyl-CoA, (0 ~runs-2-hydroxycyclohexanecarboxyl-CoA, (d) 2-oxocyclohexane- carboxyl-CoA; the indication a-f refers to the acid moiety of the CoA thiocsters which was used as standard in Fig. 3.

ductase hencc appears to be bcnzoyl-CoA rather than benzoic acid.

DISCUSSION

We have described for the first time in vitro the long- postulated key reaction of anaerobic aromatic metabolism, the enzymatic reduction of benzoyl-CoA. In this central intermediate the aromatic nucleus is not activated for the following reduction by two or more [j-hydroxyl functions. This work provides the basis for convenient assays and for purification of the new enzymcs. The products of in vitro transformation of benzoyl-CoA and the deduced reaction sequences are summarized in Fig. 6. Our findings support most of thc con- clusions derived from other kinds of experiments, notably reported by Evans and coworkers (for reviews see [2], for recent work [I I ] ) who demonstrated cyclohexanecarboxylate or cyclohcx-1-enecarboxylatc in growing cultures. Howcver, we could not find one of the postulated products, cyclohexanecarboxyl-CoA.

The elaborated assay will allow us to scale up the in vitro system and to isolate sufficient amounts of products to deter- mine by NMR techniques their structures and the sites and stereospecificity of hydrogen addition to the ring. This detailed knowledge will be required in order to discuss the novel reactions and the possible mechanism of ring reduction. Since one-electron-donating reducing compounds were active, the protons must be derived from water. The purification of the ring-reducing system proves to be difficult due to its oxygen sensitivity.

Benzoyl-CoA reductase (aromatic ring reducing) can be considered the central enzyme of the anaerobic metabolism of many, though not all, aromatic compounds. These include not only benzoate and analogues [ 1 5 , 171, but also phenolic compounds [24,35- 371, aniline 1381, toluene [39] and possibly other aromatic hydrocarbons, phenylacetic acid and analogues [16, 40,411, as well as aromatic amino acids and possibly other compounds formed in plant phenylpropane metabolism. Many of these compounds were shown in different bacteria to be degraded in the absence of molecular oxygen

via benzoyl-CoA (Fig. 1). There may be another ring-reducing system for 4-hydroxybenzoyl-CoA [19] but this has not been shown so far.

Benzoyl-CoA reductase catalyzes an intriguing reaction and it is to be expected that the electron-withdrawing coenzyme-A thioester bond would facilitate ring reduction. It has been postulated that benzoate may be covalently enzyme- bound [9]; this would also require the activation of benzoate. The reduction of benzene with hydrogen to cyclohexadiene is thermodynamically unfavorable whereas the four-electron reduction to cyclohexene is slightly exergonic. Interestingly, a specific chemical four-electron reduction of benzene to cyclohexene has been shown using hydrogen as reductant and pentaamineosmium(I1) as catalyst [42]. The analogy to the reduction of benzene suggests that the biological reduction of benzoyl-CoA may involve a four-electron rather than two- electron transfer. The nature of the physiological electron donor of the enzyme remains to be shown. Three further early labelled products were found. Two products seem to be rather unpolar (compared with cyclohex-1 -enecarboxylic acid); the others showed chromatographic behavior similar to truns-2- hydroxycyclohexane carboxylate. This shows that the reaction is more complex than anticipated, that subsequent two-electron steps cannot be excluded, and that the scheme in Fig. 6 needs modification.

The chemical reduction of benzene by Birch reduction is a radical reaction involving solvated electrons. It may well be that the biological ring reduction is also by a radical mechanism and that a metal is involved in the enzymatic electron transfer. Growth of two sulfate-reducing bacteria on benzoate was selenium-dependent and/or molybdenum-dependent, in contrast to growth on lactate 1431, suggesting that molybdenum may be a possible candidate. The specific in vitro rate of benzoyl-CoA reduction with 10 mM Ti(Il1) (10-20 nmol benzoyl-CoA transformed min- ' x mg- ' protein) is not too far from the estimated specific benzoate consumption rate; at a generation time of 6 h the estimated in vivo benzoate consumption rate is 64 nmol x min-' x mg-' protein. The reaction was not stimulated by boiled cell extract. These two facts suggest that no additional freely dissociable cocatalyst is required for the reaction.

Two other enzymes reducing an aromatic nucleus have been reported which, however, act on those phenolic compounds in whch the aromatic ring is activated by two or three m-hydroxyl functions. This class of aromatic compounds can be considered almost non-aromatic. Phloroglucinol (1,3,5- trihydroxybenzene) reductase (EC 1.3.99. -) is an NAD(P)H- dependent enzyme catalyzing the two-electron reduction of phloroglucinol [20]; this enzyme and its catalytic hydride transfer mechanism are probably totally unrelated to the enzyme discovered here. Resorcinol (1,3-dihydroxybenzene) reductase (EC 1.3.99. -) catalyzes the two-electron reduction of resorcinol to cyclohexane-I ,3-dione, was pyridine-nucleotide- independent and used reduced viologen dyes as artificial electron donors [21].

Thanks are due to our collaborators Dr. Werner Dangcl for help in the initial stages of this investigation, to Dr. Martina Gotz-Hcrm for syntheses of reference compounds, and to Christd Lochmeyer and Brigitte Oswaldfor a kind gift of benzoate-CoA ligase. This work was supported by the Deutschc Forschun~sgemeinschaft Schwerpimktpro- gramm 'Neuartige Renktionen und Kuralysemechunismen in nnaerohen Mikroorgunismen and by the Fonds deer Chemischen Industrie.

202

REFERENCES 1.

2.

3.

4. 5. 6.

7. 8. 9.

10.

11.

12.

13. 14. 15.

16.

17.

18.

19.

20.

21.

22.

Young, L. Y. (1984) in Microbial degradation of organic com-

Evans, W. C . & Fuchs, G. (1988) Annu. Rev. Microbiol. 42,289-

Berry, D. F., Francis, A. J. & Bollag, J.-M. (1987) Microbiol. Rev.

Evans, W. C. (1977) Nature (London) 270, 17-22. Dutton, P. L. & Evans, W. C. (1968) Biochem. J . 109, 5P. Blake, C. K. & Hegeman, G. D. (1987) J . Bacteriol. 169,4878-

Dutton, P. L. & Evans, W. C. (1969) Biochem. J . 113, 525-536. Guyer, M. & Hegeman, G. (1 969) J . Bacteriol. 99,906 - 907. Hegeman, G. (1988) in Microbialmetabolism and the carbon cycle,

Nozawa, T. & Maruyama, Y. (1988) J . Bacteriol. 170, 2501 -

Nozawa, T. & Maruyama, Y. (1988) J . Bacteriol. 170, 5778-

Taylor, B. F., Campbell, W. L. & Chinoy, I. (1970) J . Bacteriol.

Taylor, B. F. & Heeb, M. J. (1972) Arch. Microbiol. 83,165 - 171. Williams, R. J. & Evans, W. C. (1975) Biochem. J . 148, 1 - 10. Altenschmidf U., Oswald, B. & Fuchs, G. (1991) J . Bacteriol.

Dangel, W., Brackmann, R., Lack, A., Magdy, M., Koch, J., Oswald, B. & Fuchs, G. (1991) Arch. Microhiol. 155,256-262.

Geissler, J. F., Harwood, C. S. & Gibson, J. (1988) J . Bucteriol.

Hutber, G. N . & Ribbons, D. W. (1983) J . Gen. Microbiol. 129,

Merkel, S. M., Eberhard, A. E., Gibson, J. & Harwood, C . S.

Haddock, J. D. & Ferry, J . G. (1989) J . Biol. Chem. 264,4423-

Kluge, C., Tschech, A. & Fuchs, G. (1990) Arch. Microbiol. 155,

Krumholz, L. R., Crawford, R. L., Hemling, M. E. & Bryant, M.

pounds, pp. 487, Marcel Dekker, New York.

317.

51.43-59.

4883.

Harwood Academic Publishers, Chur.

2505.

5784.

102,430-437.

173, 5494-5501.

170, 1709-1714.

241 3 - 2420.

(1989) J . Bacteriol. 171, 1-7.

4427.

68 - 74.

P. (1987) J . Bacteriol. 169, 1886-1890.

23.

24.

25. 26.

27. 28.

29.

30. 31. 32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

Tschech, A. & Fuchs, G. (1987) Arch. Microhiol. 148. 213-

Tschech, A. & Fuchs, G. (1989) Arch. Microbiol. 152. 594-

Eden, G. & Fuchs, G. (1982) Arch. Microbiol. 133, 66-74. Zehnder, A. J. B. & Wuhrmann, K. (1976) Science 194, 1165-

Treibs, W. & Orttmann, H. (1960) Chem. Ber. 93, 545. Hardegger, E., Plattner, P. A. & Blank, F. (1944) Helv. Chim.

Gardner, H. D., Perkinjun., W. H. & Watson, H. (1910)J. Chem.

Dieckmann, W. (1901) Liebigs Ann. Chern. 317, 98-109. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. Krebs, K. G., Heusser, D. & Wimmer, H. (1967) in DCnnschicht-

chromafogruphie (Stahl, E., ed), pp. 81 3 - 861, Springer-Vcrlag, Berlin.

Decker, K. (1959) Die akfivierte Essigsuure, pp. 60-1 10, Ferdinand Enke Verlag, Stuttgart.

Schachter, D. & Taggart, J . V. (1976) J . Biol. Chem. 203, 925- 933.

Gliickler, R., Tschech, A. & Fuchs, G. (1989) FEBS Left . 251, 237 - 240.

Knoll, G. & Winter, J. (1989) Appl. Microhiol. Biotechnol. 30.

Lack, A,, Tommasi, I., Aresta, M. & Fuchs, G. (1991) Eur. J .

Schnell, S. & Schink, B. (1991) Arch. Microhiol. 155, 183-

Altenschmidt, U. & Fuchs, G. (1991) Arch. Microbiol. 156,152-

Sembiring, T. &Winter, J. (1989) Appl. Microhiol. Biotechnol. 31,

Seyfried, B., Tschech, A. & Fuchs, G. (1991) Arch. Microhiol.

Harman, W. D. & Taube, H. (1988) J . Am. Chem. SOC. 110,

Widdel, F. (1980) Thesis, Universitut, Gottingen.

217.

599.

1166.

Acta 27, 793 - 800.

SOC. 97,1756 - 1772.

318 - 324.

Biochem. 197,473-479.

190.

158.

89-92.

155,249-255.

7906 - 7907.

Documents

Enzymatic reduction of benzoyl-CoA to alicyclic compounds, a key reaction in anaerobic aromatic metabolism