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-generatingsystem (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 -enecarboxylCoA 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.-); 4hydroxybenzoyl-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 dehydrogcnase (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 (deaminating) (EC 1. -. -. -); 2-aminobenzoate-CoA ligase (AMP-forming) (EC 6.2.1.-). , I
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 circumstantial 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 converted 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 -enecarboxylCoA. 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, 4hydroxybenzoate, 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 illustrated in Fig. l .
196 p
a
GOOH
COOH
-
Phenylacetic acid
Q OH
Phenol
..
OH p-Cresol
.
AMP PP
4-OH Benzoic aC120'&
6
ATP
OH
non-aromatic products
.
2H ,SCoA
HC=O f H 2H
Benzoic acid
Y OH
g
NH*
COOH CoA
@NHcAFL
H20/2H
/
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 (phenylglyoxy1ate):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
I n vitro assay of benzoyl-CoA reduction
Materials
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-CoAgenerating 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 H 2 S 0 4 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 m M MgC12, 1.7 mM ATP, 1 mM dithioerythritol, 1.1 mM coenzyme A. 8
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 determination, cell harvesting and storage, and preparation of cell extracts were as described [23, 241. A . woodii was grown anaerobically on fructose [25].
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.
-'i zoo ,--
A
I Y.
B
200
v
U
150
0
5
v
-
7 100
150 100
0
0 0
50
50 m
N 0
u
0
0
2
4
6
8
1
0
:
o
5
0
10
Time ( r n i n )
Protein (mg)
C
D
Chromatographicseparation, 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 RP18 F 2 5 40.25-mm ~ glass plates (Merck) was performed with mcthanol/l50 mMpotassiumphosphatepH 3.5 (3:7, byvol.). HPLC was carried out on a reversed-phase CISLichrospher 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
6
7
PH
8
9
0
2
4
6
8 1 0
[Ti(lll) c i t r a t e ] (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 7.08.5). 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 omitted.
The following compounds were not commercially available; they were synthesized and the products were charac- benzoate and nitrate [15]. The purified enzyme had a specific terized by ultraviolet spectroscopy, melting point (m.p.) and/ activity of 2.5 pkat x mg-' protein. or boiling point (b. p.), and C, H analysis. Cyclohex-1-enecarboxylic acid, which only recently became commercially available (Heraeus), was synthesized Analytical methods starting from 16.4 g cyclohexene by the method of Treibs and Protein was determined by the Bradford method [31]. 14C Orttmann [27]. Yield: 17% of theory; b.p. 138°C /1.87 kPa was determined by liquid scintillation counting using external (lit. 136.5-137.5"C/1.87 kPa). m.p. 38°C. Later, the com- standardization. Radioactive areas on TLC plates were lopound was prepared by alkaline hydrolysis of its ethyl ester. cated by autoradiography using X-ray film (Kodak X-Omat, cis-2-H ydroxycyclo hexanecarboxylic acid (cis-hexahydroXar-352, Sigma Chemical Co., St. Louis, USA); radioactive salicylic acid) was synthesized starting from 3 g 2-cyclo- spots were scratched off, and the TLC material was directly hexanonecarboxylic acid analogously to the method for the extracted with 4 ml scintillation cocktail. Organic standard synthesis of cis-4-hydroxycyclohexanecarboxylicacid [28]. acids were detected by 254-nm light, or by spraying [32] (a) Yield: 86% of theory; m.p. 78 "C (lit. 79 - 80 "C). with 1% (mass/vol.) vanillin in conc. H2S04, (b) with 5% trans-2-H ydroxycyclohexanecarboxylic acid was synthe- (mass/vol.) potassium dichromate in conc. H,SO, or (c) with sized starting from 3 g 2-cyclohexanonecarboxylicacid ac- 0.1 % (massivol.) bromocresol green in ethanol. cording to the method of Gardner et al. [29]. Yield: 46% of thcory; m.p. 111 "C (lit. I l l "C ). 2-0xocyc~ohexanecarboxy~ic acid was obtained starting RESULTS from 5 g of its ethyl ester [30]. Yield: 93% of theory. Enzymatic reduction of ('4C]benzoyl-CoA in vitro Purification of benzoate-CoA ligase Extracts from cells of Pseudomonas K 172, which were Benzoate-CoA ligase was purified from denitrifying grown anaerobically with benzoate and nitrate as sole carbon Pseudomonas KB 740 cells which were anaerobically grown on and energy sources, catalyzed the reductive conversion of [U-
198
Fig. 3. Analysis by two different TLC systems and autoradiographyof the 14C-labelledproductsof (14Clbenzoyl-CoAconversion. (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 Rfvalues 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 approximdtely 10 nmol x min-' x mg-' protein; the reaction proceeded 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 identified as cyclohex-1-enecarboxylicacid (see below). This indicated that the aromatic nucleus of benzoyl-CoA was enzymatically reduced to cyclohex-1-enecarboxyl-CoA.Whether this four-electron reduction was due to one four-electrontransferring 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. Dependence 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
=J
500
m
400
v
m
-c
0 3
Benzoate consumed after incubation for 5min
10min
Benzoate present at the end of the experiment
ea
300
i 500
.-C 2.
c
.c
200
0 0 0
.-
71
Complete - ATP -ATP regenerating system, but plus ATP - exogenous benzoateCoA ligase from Pseudomonas KB70 -coenzyme A -Ti(IlI) -extract - dithioerythritol Complete, but aerobic
152
$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-generatingsystem (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 -enecarboxylCoA 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.-); 4hydroxybenzoyl-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 dehydrogcnase (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 (deaminating) (EC 1. -. -. -); 2-aminobenzoate-CoA ligase (AMP-forming) (EC 6.2.1.-). , I
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 circumstantial 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 converted 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 -enecarboxylCoA. 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, 4hydroxybenzoate, 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 illustrated in Fig. l .
196 p
a
GOOH
COOH
-
Phenylacetic acid
Q OH
Phenol
..
OH p-Cresol
.
AMP PP
4-OH Benzoic aC120'&
6
ATP
OH
non-aromatic products
.
2H ,SCoA
HC=O f H 2H
Benzoic acid
Y OH
g
NH*
COOH CoA
@NHcAFL
H20/2H
/
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 (phenylglyoxy1ate):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
I n vitro assay of benzoyl-CoA reduction
Materials
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-CoAgenerating 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 H 2 S 0 4 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 m M MgC12, 1.7 mM ATP, 1 mM dithioerythritol, 1.1 mM coenzyme A. 8
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 determination, cell harvesting and storage, and preparation of cell extracts were as described [23, 241. A . woodii was grown anaerobically on fructose [25].
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.
-'i zoo ,--
A
I Y.
B
200
v
U
150
0
5
v
-
7 100
150 100
0
0 0
50
50 m
N 0
u
0
0
2
4
6
8
1
0
:
o
5
0
10
Time ( r n i n )
Protein (mg)
C
D
Chromatographicseparation, 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 RP18 F 2 5 40.25-mm ~ glass plates (Merck) was performed with mcthanol/l50 mMpotassiumphosphatepH 3.5 (3:7, byvol.). HPLC was carried out on a reversed-phase CISLichrospher 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
6
7
PH
8
9
0
2
4
6
8 1 0
[Ti(lll) c i t r a t e ] (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 7.08.5). 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 omitted.
The following compounds were not commercially available; they were synthesized and the products were charac- benzoate and nitrate [15]. The purified enzyme had a specific terized by ultraviolet spectroscopy, melting point (m.p.) and/ activity of 2.5 pkat x mg-' protein. or boiling point (b. p.), and C, H analysis. Cyclohex-1-enecarboxylic acid, which only recently became commercially available (Heraeus), was synthesized Analytical methods starting from 16.4 g cyclohexene by the method of Treibs and Protein was determined by the Bradford method [31]. 14C Orttmann [27]. Yield: 17% of theory; b.p. 138°C /1.87 kPa was determined by liquid scintillation counting using external (lit. 136.5-137.5"C/1.87 kPa). m.p. 38°C. Later, the com- standardization. Radioactive areas on TLC plates were lopound was prepared by alkaline hydrolysis of its ethyl ester. cated by autoradiography using X-ray film (Kodak X-Omat, cis-2-H ydroxycyclo hexanecarboxylic acid (cis-hexahydroXar-352, Sigma Chemical Co., St. Louis, USA); radioactive salicylic acid) was synthesized starting from 3 g 2-cyclo- spots were scratched off, and the TLC material was directly hexanonecarboxylic acid analogously to the method for the extracted with 4 ml scintillation cocktail. Organic standard synthesis of cis-4-hydroxycyclohexanecarboxylicacid [28]. acids were detected by 254-nm light, or by spraying [32] (a) Yield: 86% of theory; m.p. 78 "C (lit. 79 - 80 "C). with 1% (mass/vol.) vanillin in conc. H2S04, (b) with 5% trans-2-H ydroxycyclohexanecarboxylic acid was synthe- (mass/vol.) potassium dichromate in conc. H,SO, or (c) with sized starting from 3 g 2-cyclohexanonecarboxylicacid ac- 0.1 % (massivol.) bromocresol green in ethanol. cording to the method of Gardner et al. [29]. Yield: 46% of thcory; m.p. 111 "C (lit. I l l "C ). 2-0xocyc~ohexanecarboxy~ic acid was obtained starting RESULTS from 5 g of its ethyl ester [30]. Yield: 93% of theory. Enzymatic reduction of ('4C]benzoyl-CoA in vitro Purification of benzoate-CoA ligase Extracts from cells of Pseudomonas K 172, which were Benzoate-CoA ligase was purified from denitrifying grown anaerobically with benzoate and nitrate as sole carbon Pseudomonas KB 740 cells which were anaerobically grown on and energy sources, catalyzed the reductive conversion of [U-
198
Fig. 3. Analysis by two different TLC systems and autoradiographyof the 14C-labelledproductsof (14Clbenzoyl-CoAconversion. (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 Rfvalues 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 approximdtely 10 nmol x min-' x mg-' protein; the reaction proceeded 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 identified as cyclohex-1-enecarboxylicacid (see below). This indicated that the aromatic nucleus of benzoyl-CoA was enzymatically reduced to cyclohex-1-enecarboxyl-CoA.Whether this four-electron reduction was due to one four-electrontransferring 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. Dependence 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
=J
500
m
400
v
m
-c
0 3
Benzoate consumed after incubation for 5min
10min
Benzoate present at the end of the experiment
ea
300
i 500
.-C 2.
c
.c
200
0 0 0
.-
71
Complete - ATP -ATP regenerating system, but plus ATP - exogenous benzoateCoA ligase from Pseudomonas KB70 -coenzyme A -Ti(IlI) -extract - dithioerythritol Complete, but aerobic
152
