INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

In the absence of oxygen, bacteria degrade aromatic compounds derived from plant material and industrial sources to form benzoate, often in the form of benzoyl coenzyme A (benzoyl-CoA), as a common intermediate. The benzoate pathway is then the main conduit for the anaerobic processing of structurally diverse compounds, including aromatic hydrocarbons, phenols, halogenated aromatics, and phenylpropanoids, to central biosynthetic intermediates (9, 18, 20). Because of its importance in bioremediation and biomass recycling, the anaerobic benzoate degradation pathway has been given increased attention in recent years. It is now apparent that the pathway includes an initial critical ring reduction step, followed by a series of -oxidation-like reactions, culminating in ring cleavage by a hydrolytic, rather than a thiolytic, mechanism. Although there is agreement about the general metabolic strategy employed for anaerobic benzoate degradation, the structures of several of the pathway intermediates are not known with certainty, and two alternative degradation pathways have been proposed based on work with the phototrophic bacterium Rhodopseudomonas palustris and the denitrifier Thauera aromatica, that differ in the steps following benzoyl-CoA reduction (23, 24) (Fig. 1).

View larger version (20K): [in this window] [in a new window]   FIG. 1.   Alternative pathways for anaerobic benzoate degradation. Pathway A is similar to a sequence of reactions originally proposed by Dutton and Evans (13) and Guyer and Hegeman (22), with some modifications. Pathway B is a sequence of reactions proposed by Koch et al. (27) based on identification of cyclohex-1,5-diene-1-carboxyl-CoA and 6-hydroxycyclohex-1-ene-1-carboxyl-CoA as products of benzoyl-CoA reduction in cell extracts of R. palustris and T. aromatica. The product of benzoyl-CoA reduction by whole cells is uncertain but is probably one or more of three possible cyclohexadienecarboxyl-CoA isomers (21, 27), any of which could be subsequently reduced as shown. Enzymes that have been purified from R. palustris or T. aromatica are indicated in bold. Solid arrows indicate enzymatic activities that have been detected in benzoate-grown R. palustris or T. aromatica cells (1, 7, 19, 33). Dashed arrows indicate hypothetical enzymatic reactions.

An enzymatic activity that catalyzes the cleavage of 2-ketocyclohexanecarboxyl-CoA (2-ketochc-CoA), one of two proposed ring cleavage substrates of benzoate degradation (Fig. 1), has been detected in cell extracts of benzoate-grown cells of R. palustris (33). This activity was induced 10-fold by growth on benzoate, suggesting a role in anaerobic benzoate degradation (33). Recently a gene, termed badI, that encodes this ring cleavage activity was identified in a cluster of anaerobic benzoate degradation genes from R. palustris (15). When expressed in E. coli, the BadI protein had 2-ketochc-CoA hydrolase activity.

Here we report the purification and characterization of 2-ketochc-CoA hydrolase from R. palustris. A 2-ketochc-CoA hydrolase mutant was constructed and found to be unable to grow on benzoate, indicating that this enzyme is essential for anaerobic benzoate degradation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and culture conditions. R. palustris strains were grown anaerobically in light in defined mineral medium at 30°C as described previously (26). Carbon sources were added to a final concentration of 3 mM, except succinate, which was used at a final concentration of 10 mM. Growth was monitored spectrophotometrically at 660 nm. Large quantities of cells were grown in 20-liter glass carboys that were completely filled and stoppered. Cultures were stirred slowly and illuminated with 40- or 100-W incandescent light bulbs. Cells were harvested by centrifugation when they were in the mid- to late logarithmic phase of growth, washed once in 20 mM triethanolamine · hydrochloride, pH 7.5 (TEA buffer), and stored frozen at 70°C until used.

Chemical synthesis of CoA thioesters. 2-Ketocyclohexanecarboxylic acid (2-ketochc) was prepared from ethyl-2-cyclohexanonecarboxylate by alkaline hydrolysis as described by Dieckmann (11). The CoA thioesters of 2-ketochc and pimelate were synthesized from mixed anhydrides as described by Merkel et al. (28), except that the final alkali treatment step was omitted. The procedure for pimelyl-CoA synthesis yielded mainly pimelyl-CoA and only very small amounts of pimelyl-diCoA. CoA thioesters were purified with C₁₈ reverse-phase Sep-Pack cartridges (Millipore Corp., Milford, Mass.) as described previously (33), except 20 mM ammonium acetate buffer (pH 6.0) was used in place of 20 mM potassium phosphate buffer. Subsequent high-pressure liquid chromatography (HPLC) analysis was used to confirm that the purified CoA thioester substrates were free of contaminating CoASH. The product of the 2-ketochc-CoA synthesis reaction was analyzed by electrospray ionization mass spectrometry (ES MS) as described in the Results section. Since both CoASH and acyl-CoA thioesters have essentially the same extinction coefficients at A₂₅₄, the CoA thioester substrates were quantitated spectrophotometrically at A₂₅₄ according to a standard curve of known quantities of CoASH.

2-Ketochc-CoA hydrolase assays. Hydrolase activity was measured spectrophotometrically at 28°C as the decrease in A₃₁₄ of a Mg²⁺ substrate complex, as described previously (33). The standard reaction mixtures contained 50 mM Tris-HCl buffer (pH 8.5), 100 mM MgCl₂, and 1 mM CoA thioester substrate. This assay was used to test protein fractions for activity during protein purification. However, as described in the Results section, the spectrophotometric assay was suitable only for determining relative enzymatic activities, due to impurities in the substrate. Specific activities were determined by measuring product formation, as determined by HPLC analysis. Reactions were stopped by acidification with 100 mM HClO₄ (final concentration) at various time points over the period when the reaction rate was linear. Precipitated protein was removed by centrifugation (14,000 × g for 20 min), and the supernatant was neutralized with 100 mM KHCO₃ (final concentration). The reaction product, pimelyl-CoA, was separated from other components of the reaction mixture by HPLC using an Ultrasphere octyldecyl silane-C₁₈ reverse-phase (4.6 mm by 25 cm) column (Beckman Instruments, Fullerton, Calif.). The solvent system used was 20 mM ammonium acetate (pH 6.0) and methanol. The column was equilibrated with 20% methanol, and elution was achieved by a linear gradient of 20 to 80% methanol in 30 min. The absorbance of the effluent was monitored by scanning the region from 210 to 260 nm using a model 996 photodiode array detector (Waters Associates, Milford, Mass.). The quantity of product formed was determined by relating its peak area at 254 nm to a standard curve constructed with known quantities of pimelyl-CoA. Specific activities were expressed as micromoles of pimelyl-CoA formed per minute per milligram of protein. In some cases, HPLC fractions were collected, lyophilized, and resuspended in deionized water for molecular mass determination by ES MS at the University of Iowa-High Resolution Mass Spectrometry Facility or the University of Illinois Mass Spectrometry Laboratory.

Preparation of cell extracts. Approximately 33 g (wet weight) of cell paste was suspended in 60 ml of TEA buffer containing 1 mM MgCl₂, 0.1 mM EDTA, 0.1 mM dithiothreitol, DNase (5 µg/ml), RNase (5 µg/ml), 0.5 mM phenylmethylsulfonyl fluoride, and 0.5 mM -aminocaproate. Cells were lysed by several passages through a French pressure cell at 85 MPa, and cell debris was removed by centrifugation at 10,000 × g for 30 min at 4°C. The resulting supernatant was then subjected to centrifugation at high speed (100,000 × g) for 1 h at 4°C. The supernatant from the second centrifugation was termed crude cell extract.

Purification of 2-ketochc-CoA hydrolase. Crude cell extract was heated to 60°C for 5 min and then centrifuged at 10,000 × g for 20 min at 4°C to remove denatured protein. The supernatant was used for further purification. All subsequent steps were carried out at 4°C.

(i) Q-Sepharose chromatography. The clear, pale orange supernatant from the heat treatment was loaded onto a Q-Sepharose column (2.6 by 10 cm) (Pharmacia Biotech Inc., Piscataway, N.J.) equilibrated with TEA buffer. The column was washed extensively with TEA buffer and then developed with a linear gradient of 0 to 300 mM KCl in TEA buffer over 200 min at a flow rate of 5 ml/min. Fractions (10 ml) were collected and assayed spectrophotometrically for ring-cleavage activity. The enzyme activity eluted at approximately 180 mM KCl.

(ii) Phenyl-Sepharose chromatography. The active fractions from the Q-Sepharose column were pooled, and solid (NH₄)₂SO₄ was added slowly with stirring to give a final concentration of 1.7 M (NH₄)₂SO₄. The supernatant obtained after centrifugation at 10,000 × g for 20 min was loaded onto a phenyl-Sepharose HL column (2.6 by 10 cm) (Pharmacia) equilibrated with TEA buffer containing 1.7 M (NH₄)₂SO₄. The column was developed with a linear gradient of 1.7 to 0 M (NH₄)₂SO₄ in TEA buffer over 100 min at a flow rate of 5 ml/min. Fractions (10 ml) were collected, and those with activity were pooled and concentrated to 0.5 ml by ultrafiltration through a YM-30 membrane (Amicon, Beverly, Mass.). The enzyme activity eluted at approximately 0.2 M (NH₄)₂SO₄.

(iii) Gel filtration chromatography. The concentrated protein from the phenyl-Sepharose column was loaded onto a Superose 12 column (1.0 by 30 cm) (Pharmacia) equilibrated with TEA buffer. The column was run at a flow rate of 0.2 ml/min, and 2-ml fractions were collected. The enzyme eluted with 11 ml of buffer. Active fractions were pooled and concentrated and the buffer was changed to 1 mM potassium phosphate buffer (pH 6.8) with a Centriplus 10 concentrator (Amicon).

(iv) Hydroxyapatite chromatography. The sample was then loaded onto a hydroxyapatite column (1.6 cm by 55 cm) that had been equilibrated with 1 mM potassium phosphate (pH 6.8). The column was developed with a linear gradient of 0 to 400 mM potassium phosphate (pH 6.8) in 800 min at a flow rate of 1 ml/min. Fractions (2 ml) were collected, and those with activity were pooled. The hydrolase activity eluted at approximately 200 mM potassium phosphate.

Other analytical procedures. The native M_r was determined by gel filtration chromatography using a Superose 12 column (1.0 by 30 cm) (Pharmacia) equilibrated with TEA buffer. The following protein molecular weight standards were used to calibrate the column: ribonuclease A, 13,700; chymotrypsinogen A, 25,000; bovine serum albumin, 66,000; aldolase, 158,000; and ferritin, 440,000. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was carried out with 10% acrylamide gels by standard procedures (3). Separated proteins were visualized by staining with Coomassie blue R-250. Molecular weight standards were from Bio-Rad Laboratories (Richmond, Calif.). Purified 2-ketochc-CoA hydrolase (220 pmol) was subjected to SDS-10% PAGE and then electroblotted onto a ProBlott membrane (Perkin-Elmer Applied Biosystems, Foster City, Calif.) according to the manufacturer's instructions. The amino terminal sequence was determined with an automated sequence analyzer by the Genetic Engineering Facility at the University of Illinois Urbana-Champaign. The amount of protein was estimated by a dye-binding assay (Bio-Rad) with bovine serum albumin as a standard.

Immunoblotting. Polyclonal antiserum was prepared from a rabbit inoculated with purified enzyme at the Cornell University Center for Research Animal Resources (Ithaca, N.Y.). Standard protocols were used for analysis of protein expression (3). Cell extracts of R. palustris were typically separated on SDS-12% polyacrylamide gels and electroblotted onto an Immobilon polyvinylidene difluoride membrane (Millipore), and antigens were visualized by using alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin G (Bio-Rad).

Cloning, DNA manipulations, and mutant construction. Standard protocols were used for DNA cloning and transformation (3). Plasmids were purified on QIAprep spin columns (Qiagen Inc., Chatsworth, Calif.). R. palustris chromosomal DNA was isolated as described previously (14). Southern blots were performed with the Genius kit (Boehringer Mannheim Biochemicals, Indianapolis, Ind.). A badI::-Km mutant was constructed by inserting a 3.0-kb BamHI fragment from pHP45-Km, encoding the -Km cassette (17), into the unique BglII site on plasmid pDP105, which contains badI cloned into pUC18 (15). This construct was then cloned into the suicide vector pJQ200KS (34) to generate pDP201. The disrupted badI gene was then introduced into R. palustris on pDP201 by conjugation from E. coli S17-1 (36). Recombinants were selected as described previously (14, 31). The insertion of the -Km cassette into the open reading frame of badI was verified by Southern blot analysis (data not shown).

Computer analysis of DNA sequences. The BLAST network services at the National Center for Biotechnology Information (Bethesda, Md.) were used to search protein databases for similar sequences (2). Amino acid sequence similarities were calculated by using the GAP program from the University of Wisconsin Genetics Computer Group software package (version 9.0) (10). The multiple sequence alignment was constructed by using the CLUSTAL W multiple sequence alignment program at the Baylor College of Medicine Human Genome Center (37). The program BOXSHADE (version 3.21) was used to shade aligned sequences.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Substrate synthesis and product analysis. 2-Ketochc-CoA hydrolase activity was assayed by monitoring the decrease in A₃₁₄ of a 1 mM solution of substrate prepared as described in Materials and Methods. The assay depends on the loss of absorbance of a Mg²⁺-enolate complex that occurs when the alicyclic ring of 2-ketochc-CoA is cleaved (33). Although this assay was adequate for monitoring activity during purification, anomalous behavior was observed when attempts were made to determine the affinity of the purified hydrolase for its substrate. Only a small portion of the total substrate added was converted to the product, pimelyl-CoA, as determined by C₁₈ reverse-phase HPLC analysis. Modification of the HPLC parameters revealed that a small shoulder, representing no more than 5% of the total substrate peak area at A₂₅₄ and eluting at about 60% methanol, decreased in a manner that corresponded with the formation of pimelyl-CoA, which eluted at about 45% methanol. It subsequently became apparent that the predominant product formed by the method used for preparing the substrate from 2-ketochc and CoASH was pimelyl-CoA semialdehyde with an m/z of the (M-1) of 892, as determined by ES MS analysis. The desired enzyme substrate, 2-ketochc-CoA, made up only about 5% of the whole, as determined by the decrease in HPLC peak area at A₂₅₄ following reaction with the hydrolase. The reaction product comigrated with pimelyl-CoA on a C₁₈ reverse-phase HPLC column and had an m/z of the (M-1) of 908 when analyzed by ES MS. This is consistent with the incorporation of H₂O into 2-ketochc-CoA during ring cleavage by the hydrolase to yield pimelyl-CoA.

Purification of 2-ketochc-CoA hydrolase. A typical purification of 2-ketochc-CoA hydrolase from a culture of R. palustris grown anaerobically on benzoate is shown in Table 1. All steps were carried out in air, since preliminary studies showed that activity was not oxygen sensitive. The enzyme, which was present in the soluble fraction of cell extracts, was purified approximately 100-fold to apparent homogeneity in four chromatographic steps (Table 1). The purified protein was visualized as a single band on SDS-polyacrylamide gels stained with Coomassie blue (Fig. 2). About 8% of the activity present in crude cell extracts was recovered to yield 600 µg of pure enzyme from 725 mg of R. palustris extract. Improved yields could be obtained by eliminating the Superose 12 column step, but this resulted in enzyme that was only approximately 90% pure. The first 25 N-terminal amino acids of the purified hydrolase were determined to be the following: Met-Gln-Phe-Glu-Asp-Leu-Ile-Tyr-Glu-Ile-Arg-Asn-Gly-Val-Ala-Trp-Ile- Ile-Ile-Asn-Arg-Pro-Asp-Asp-Met. This amino acid sequence matched exactly that predicted from the nucleotide sequence of the badI gene (15).

View larger version (77K): [in this window] [in a new window]   FIG. 2.   SDS-PAGE analysis of active protein fractions obtained during purification of 2-ketochc-CoA hydrolase. Lanes: 1, crude cell extract (20 µg); 2, Q-Sepharose pooled fractions (10 µg); 3, phenyl-Sepharose pooled fractions (5 µg); 4, hydroxyapatite pooled fractions (2 µg). Numbers to the right of the gel are molecular masses (in kilodaltons).

Enzyme properties. The molecular mass of the purified 2-ketochc-CoA hydrolase, as determined by SDS-PAGE, was 34,700 Da. The molecular weight as determined by gel filtration, was 134,000, indicating that native 2-ketochc-CoA hydrolase is a homotetramer.

1

1

1

1

Properties of 2-ketochc-CoA hydrolase as deduced from the badI sequence and similarities to other enzymes. The badI gene is predicted to encode a protein of 260 amino acids with a calculated molecular mass of 28.6 kDa. The predicted pI of BadI is 7.6. The BadI protein is most similar in amino acid sequence (~45% identity, 62% similarity) to MenB from a variety of bacteria. This enzyme, 1,4-dihydroxy-2-naphthoate synthase, catalyzes a ring closure reaction in the pathway of menaquinone biosynthesis (35). BadI is also homologous to enoyl-CoA hydratases (33% identity, 51% similarity) and 4-chlorobenzoyl-CoA dehalogenases (26% identity, 42% similarity) and thus appears to belong to the recently described crotonase superfamily of enzymes (4). Members of this enzyme superfamily have somewhat low overall amino acid sequence identity but share structural features proposed to be involved in catalysis (4, 12). An amino acid alignment between BadI and selected homologous enzymes is shown in Fig. 3.

View larger version (68K): [in this window] [in a new window]   FIG. 3.   Amino acid sequence alignment of 2-ketochc-CoA hydrolase with members of the crotonase superfamily of enzymes. 2KCH, 2-ketochc-CoA hydrolase of R. palustris, 260 amino acids; DHNS, 1,4-dihydroxy-2-naphthoate synthase of Escherichia coli, 284 amino acids (35); ECH, enoyl-CoA hydratase of rat mitochondria, 290 amino acids with leader sequence (29); CBD, 4-chlorobenzoyl-CoA dehalogenase of Pseudomonas sp. strain CBS-3, 269 amino acids (5). Amino acids identical to those in the 2KCH sequence are on a black background, and those similar to 2KCH are on a grey background. The boxed region indicates the crotonase superfamily signature sequence (30).

Immunoblot analysis of 2-ketochc-CoA hydrolase expression. Antiserum prepared against purified 2-ketochc-CoA hydrolase reacted with a single band of 35 kDa on immunoblots of SDS-PAGE-separated extracts of benzoate-grown R. palustris cells (Fig. 4). Barely detectable amounts of the hydrolase were synthesized by cells grown on succinate or pimelate. Cells grown anaerobically on compounds that are proposed to be metabolized via the benzoate pathway, including cyclohex-1-enecarboxylate, cyclohexanecarboxylate, and 4-hydroxybenzoate, contained high levels of 2-ketochc-CoA hydrolase. Cells grown on succinate in the presence of aromatic compounds that are not metabolized by R. palustris CGA009, including 3-chlorobenzoate, 2-aminobenzoate (anthranilate), protocatechuate, phenylacetate, and vanillate, did not synthesize significant amounts of the enzyme (data not shown).

View larger version (38K): [in this window] [in a new window]   FIG. 4.   Immunoblot analysis of 2-ketochc-CoA hydrolase expression using antiserum to purified enzyme. Lane BadI contains 1 µg of purified 2-ketochc-CoA hydrolase; lanes BEN, SUC, POB, CHC, 1CH, and PIM contain 20 µg each of crude cell extract of R. palustris grown on benzoate, succinate, 4-hydroxybenzoate, cyclohexanecarboxylate, cyclohex-1-enecarboxylate, or pimelate, respectively. Numbers to the left of the gel are molecular masses (in kilodaltons).

Construction and characterization of a badI mutant. Insertional inactivation of the badI gene with a -kanamycin resistance (-Km) cassette (Fig. 5) generated an R. palustris chromosomal mutant (CGA700) that was unable to grow on benzoate, 4-hydroxybenzoate, cyclohex-1-enecarboxylate, or cyclohexanecarboxylate under anaerobic conditions. However, CGA700 grew normally on succinate and pimelate. The benzoate growth defect of the badI mutant was not due to a polar effect of the insertion mutation on downstream genes, because a mutant with an insertion in badJ, the gene immediately downstream of badI, grew on benzoate at wild-type rates (32). Cell extracts of CGA700 grown on succinate plus benzoate contained no detectable 2-ketochc-CoA hydrolase activity.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Restriction maps of the region of the R. palustris CGA009 chromosome that encodes badI (benzoic acid degradation gene cluster) (top) and of subclones for the construction of a badI chromosomal insertion mutant. -Km is the omega kanamycin resistance cassette from pHP45::Km that contains transcriptional and translational terminators (17). WT, wild type.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Several of the proposed reactions of anaerobic benzoate degradation (Fig. 1) have been demonstrated in extracts of benzoate-grown cells (1, 19, 27, 33). However, with the exceptions of the initial thioesterification and first reductive reactions (14, 15), none of these activities has been shown to be required for benzoate metabolism. The results reported here show that BadI, the enzyme proposed to catalyze the final alicyclic ring-cleavage step, is essential for growth on benzoate in the absence of oxygen. 2-Ketochc-CoA hydrolase is also required for growth on the alicyclic compounds cyclohexanecarboxylate and cyclohex-1-enecarboxylate. It has been suggested that cyclohexanecarboxylate, cyclohex-1-enecarboxylate, and benzoate are degraded via a common set of enzymes (23). The growth phenotype of the badI mutant and the finding that the BadI protein is active with 2-ketochc-CoA as a substrate are consistent with this proposal. We were not able to determine whether BadI catalyzes the cleavage of 2-keto-6-hydroxycyclohexanecarboxyl-CoA, the alternate ring cleavage substrate that has been proposed to be generated during benzoate degradation (Fig. 1), since this substrate has yet to be prepared. Thus, the possibility that benzoate degradation in R. palustris may occur in part or even entirely via pathway B shown in Fig. 1 cannot be excluded.

The procedure used to thioesterify 2-ketochc with CoASH yielded products consisting primarily of pimelyl-CoA semialdehyde and only small amounts of the desired compound, 2-ketochc-CoA. In fact, pimelyl-CoA semialdehyde predominated to such an extent that we were unable to effectively separate and purify significant amounts of 2-ketochc-CoA from synthesis mixtures. Because of the impurity of the substrate, we were unable to determine the affinity of the hydrolase for its substrate. Enzymatic synthesis of the substrate, as the enzymes involved in the conversion of cyclohex-1-enecarboxyl-CoA to 2-ketochc-CoA are purified, should make it possible to synthesize and purify quantities of 2-ketochc-CoA sufficient to conduct these sorts of studies. Cyclohex-1-enecarboxyl-CoA can be easily synthesized in pure form and in large amounts (33).

The product of the alicyclic ring cleavage reaction catalyzed by the BadI protein was determined by ES MS to be pimelyl-CoA, indicating that ring fission occurs by a hydrolytic mechanism. Hydrolytic enzymes that cleave carbon-carbon bonds are unusual, and the lack of obvious cofactors or prosthetic groups associated with BadI made it difficult to draw inferences about possible mechanistic features of the enzyme. In this regard, the deduced amino acid sequence of the protein was particularly helpful because it suggests that BadI is homologous to members of the crotonase superfamily of proteins (4). The deduced amino acid sequence of BadI aligns along its entire length with naphthoate synthases, enoyl-CoA hydratases, and 4-chlorobenzoyl-CoA dehalogenases. These enzymes have low overall sequence identity but contain short segments of high amino acid similarity (Fig. 3). Although these enzymes catalyze seemingly disparate reactions, all are proposed to form a thioester-enolate intermediate during catalysis (12). Recent structural determinations for 4-chlorobenzoyl-CoA dehalogenase and enoyl-CoA hydratase (6, 16) show that the active sites of these enzymes contain an "oxyanion hole" (4) required for polarization of the substrate and stabilization of enolate intermediates during catalysis. Ongoing studies of critical mechanistic features of 4-chlorobenzoyl-CoA dehalogenase and enoyl-CoA hydratase (6, 8, 16, 38) should be helpful in future work aimed at identifying amino acid residues of 2-ketochc-CoA hydrolase that are important for catalysis.

The amino acid identities shared by 2-ketochc-CoA hydrolase (BadI) and naphthoate synthases (MenB) are intriguing in that the former enzyme catalyzes a ring fission reaction upon addition of water whereas the latter catalyzes a ring closure reaction (25). We examined the possibility that BadI can operate in the reverse direction to catalyze the formation of 2-ketochc-CoA from pimelyl-CoA. However, this reaction did not occur under the conditions tested.

The fact that many members of the crotonase superfamily, including 2-ketochc-CoA hydrolase, participate in metabolic sequences that are part of, or related to, fatty acid degradation pathways is consistent with the conclusion that members of this superfamily have a common ancestry. The close relationship between naphthoate synthase, an enzyme involved in synthesis of an electron carrier, and 2-ketochc-CoA hydrolase, a dissimilatory enzyme, suggests that the badI gene may have been most recently recruited from a biosynthetic gene cluster and adapted for use in a catabolic pathway.