Identification of a Hydratase and a Class II Aldolase Involved in Biodegradation of the Organic Solvent Tetralin

Abstract

Two new genes whose products are involved in biodegradation of the organic solvent tetralin were identified. These genes, designated thnE and thnF, are located downstream of the previously identified thnD gene and code for a hydratase and an aldolase, respectively. A sequence comparison of enzymes similar to ThnE showed the significant similarity of hydratases involved in biodegradation pathways to 4-oxalocrotonate decarboxylases and established four separate groups of related enzymes. Consistent with the sequence information, characterization of the reaction catalyzed by ThnE showed that it hydrated a 10-carbon dicarboxylic acid. The only reaction product detected was the enol tautomer, 2,4-dihydroxydec-2-ene-1,10-dioic acid. The aldolase ThnF showed significant similarity to aldolases involved in different catabolic pathways whose substrates are dihydroxylated dicarboxylic acids and which yield pyruvate and a semialdehyde. The reaction products of the aldol cleavage reaction catalyzed by ThnF were identified as pyruvate and the seven-carbon acid pimelic semialdehyde. ThnF and similar aldolases showed conservation of the active site residues identified by the crystal structure of 2-dehydro-3-deoxy-galactarate aldolase, a class II aldolase with a novel reaction mechanism, suggesting that these similar enzymes are class II aldolases. In contrast, ThnF did not show similarity to 4-hydroxy-2-oxovalerate aldolases of other biodegradation pathways, which are significantly larger and apparently are class I aldolases.

The organic solvent tetralin (1,2,3,4-tetrahydronaphthalene) is produced for industrial purposes from naphthalene by catalytic hydrogenation or from anthracene by cracking, and it is widely used as a degreasing agent and solvent for fats, resins, and waxes, as a substitute for turpentine in paints, lacquers, and shoe polishes, and also in the petrochemical industry in connection with coal liquefaction (6). A concentration of tetralin higher than 100 μM inhibits bacterial growth (24). The toxicity of this compound is due in part to its lipophilic character, which results in accumulation in the cell membrane, leading to changes in the structure and function of the membrane (26, 27). In addition, tetralin also forms toxic hydroperoxides in the cell (5).

Tetralin is a bicyclic molecule composed of an aromatic moiety and an alicyclic moiety, which share two carbon atoms. In principle, initial transformation of tetralin may involve metabolism of either the aromatic ring or the alicyclic ring. A few bacterial strains that are able to grow on tetralin as a sole carbon and energy source have been described (24). By identifying accumulated intermediates, several authors have suggested that some bacteria, such as Pseudomonas stutzeri AS39 (23), initially hydroxylate and further oxidize the alicyclic ring, while other bacteria, such as Corynebacterium sp. strain C125 (25), initially dioxygenate the aromatic ring, which is cleaved in the extradiol position (meta-cleavage pathway); thus, the data indicate that tetralin may be metabolized in different ways. Strain TFA, which has been placed in the species recently renamed Sphingopyxis macrogoltabida (formerly a member of the genus Sphingomonas) (11, 30), is able to grow on tetralin as a sole source of carbon and energy. Tetralin is metabolized by strain TFA through a meta-cleavage pathway, which involves extradiol cleavage of the catechol derivative 1,2-dihydroxytetralin by a 1,2-dihydroxynaphthalene dioxygenase, followed by hydrolytic cleavage of the ring fission product (2, 10). Unlike other hydrolases involved in meta-cleavage pathways, the hydrolase ThnD cleaves a C—C bond which is part of the alicyclic ring, which results in a single long linear dicarboxylic reaction product (10), instead of cleaving the intermediate into a vinylpyruvate and a carboxylate.

In other biodegradation pathways, the vinylpyruvate product is further metabolized by a hydratase and an aldolase, yielding pyruvate and an aldehyde. The best-characterized reactions are those involving metabolism of 2-hydroxypenta-2,4-dienoate, the common intermediate of meta-cleavage pathways for many monoaromatic compounds and biphenyls. The substrate of the hydratase is the dienol tautomer of 2-hydroxypentadienoate, which is apparently tautomerized in the active site to 2-oxopent-3-enoic acid and subsequently hydrated, yielding 4-hydroxy-2-oxo-pentanoic acid (9, 15). This intermediate is further metabolized by an aldolase, which produces pyruvate and acetaldehyde (16). More distantly related hydratases and aldolases involved in biodegradation of homoprotocatechuate metabolize a larger dicarboxylic intermediate, 2-oxohepta-3-enedioic acid. In this case, the substrate of the hydratase appears to be the keto tautomer (19), which is hydrated to 2,4-dihydroxyhepta-2-enedioic acid, the substrate of the aldolase, yielding pyruvate and succinic semialdehyde (17, 21, 29).

In this study we identified and sequenced the genes coding for the hydratase and the aldolase involved in tetralin biodegradation. Here we show that these enzymes are very similar to the homologous enzymes involved in homoprotocatechuate degradation. The product of the hydrolysis catalyzed by ThnD is a long dicarboxylic acid containing 10 carbon atoms, which is further metabolized by the hydratase. The action of the aldolase on the hydrated product yields pyruvate, which enters the central metabolism, and pimelic semialdehyde, which could be metabolized further.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Escherichia coli DH5α (8) was used for cloning and isolation of DNA for sequencing. E. coli NCM631/pIZ227 (7), a strain that produces the T7 RNA polymerase, was used to overproduce ThnD (hydrolase), ThnE (hydratase), and ThnF (aldolase). E. coli strains were routinely grown in Luria-Bertani medium. Strain TFA and its mutant derivative K7 (10) were grown in mineral medium (4) with tetralin in the vapor phase and β-hydroxybutyrate (8 mM) as the carbon and energy sources.

Sequencing was performed directly by using plasmid pIZ608 as the template (11). A 2.9-kb StuI-EcoRV fragment from plasmid pIZ608 was cloned into the SmaI site at the multiple cloning sequence of pBluescript II KS(+) (Stratagene), which yielded pIZ674. The cloned fragment harbored the thnDEF genes. pIZ675 was constructed by deleting a 658-bp SalI fragment from pIZ674, which eliminated the 3′ end of thnF.

Protein overexpression, crude extracts, and electrophoretic conditions.

For overexpression of thnDEF or thnDE, E. coli NCM631/pIZ227 was transformed with pIZ674 or pIZ675. pBluescript II KS was used as a control. The resulting transformants were grown in Luria-Bertani liquid medium at 26°C until the optical density at 600 nm was 0.7. Then they were induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) overnight (10 to 12 h). Cells were harvested by centrifugation, frozen in liquid nitrogen, broken with aluminum oxide 90 (Merck), and suspended in 0.01 to 0.02 volume of 20 mM Tris-HCl (pH 8.0)-100 mM NaCl. The lysate was centrifuged at 18,000 × g for 30 min at 4°C. The supernatant was frozen and kept at −80°C. Sample preparation and sodium dodecyl sulfate-polyacrylamide gel electrophoresis were performed essentially as described previously (13). Gels were stained with GELCODE Blue stain reagent (Pierce).

Determination of pyruvate.

In vitro reactions were carried out in 20 mM Na/K phosphate buffer (pH 7.2) by using crude extracts of wild-type strain TFA, mutant K7, or E. coli NCM631 harboring pIZ674 or pIZ675 and 2-hydroxy-4-(2-oxocyclohexyl)-2,4-butadienoic acid (OCHBDA) as the substrate. The reaction mixtures were incubated at 30°C for 5 h (in the case of TFA or K7 mutant extracts) or 2 h (for E. coli NCM631 extracts). The concentration of pyruvate produced was measured after removal of the protein by centrifugation in a Centricon-10 tube (Millipore). The filtered reaction mixture (900 μl) was incubated with 200 μM NADH and 5.5 U of lactate dehydrogenase. The amount of pyruvate was estimated by determining the amount of NADH consumed, which was monitored by measuring the decrease in absorbance at 340 nm.

Chemicals.

OCHBDA was produced biologically from 1,2-dihydroxytetralin by whole cells of an E. coli strain overproducing the extradiol dioxygenase ThnC (2). Chemically synthesized pimelic semialdehyde was obtained from the laboratory chemical collection of W. Reineke.

Identification of intermediates.

To identify the products of the reactions catalyzed by the hydratase and the aldolase in the tetralin pathway, crude extracts from induced cells of NCM631/pIZ227 harboring pIZ675 or pIZ674 were used. The in vitro reactions were performed as described above. To stop the reactions, the reaction mixtures were acidified to pH 2, and the products were extracted from the supernatants with Bakerbond spe extraction columns (Mallinckrodt Baker B. V.), as previously described (10). The products were methylated with diazomethane (3) or trimethylsilylated with hexamethyldisilane and dimethylchlorosilane (1:1 [vol/vol]). The reaction mixtures were incubated for 15 min at room temperature. After this, the products were analyzed by gas chromatography-mass spectrometry (GC-MS) by using a Fission mass selective detector (MD800; VG Analytical, Manchester, United Kingdom) with a DB-5 MS fused silica column (length, 30 m; inside diameter, 0.25 mm; film thickness, 0.25 μm; J & W Scientific, Folsom, Calif.). The following column temperature programs were used: 150°C (5 min isothermal) to 240°C (final hold for 15 min), increasing at a rate of 4°C/min for trimethylsilylated products; and 60°C (5 min isothermal) to 230°C (final hold for 15 min), increasing at a rate of 4°C/min for methylated products. The injector temperature was 250°C, and the transfer line temperature was also 250°C. The carrier gas (helium) flow rate was 1 ml min⁻¹. The end of the column was inserted directly into the ion source block. The mass spectra were generated in electron ionization-positive mode at 70 eV.

Sequence analysis comparison.

The resulting 1,680-bp sequence in all reading frames was initially compared to the sequences in the databases by using the BLASTp and tBLASTn programs (1). Sequences which showed high levels of similarity to the sequence of strain TFA were aligned by using the CLUSTAL X program (31) with default parameters. A distance matrix and a phylogenetic tree were constructed by using the neighbor-joining method (22) and were visualized with the TreeView program.

Nucleotide sequence accession number.

The nucleotide sequence reported here has been deposited in the DDBJ, EMBL, and GenBank nucleotide sequence databases under accession no. AF498315.

RESULTS

Sequencing of thnE and thnF and sequence analysis of their gene products.

Complete sequencing of the downstream region of thnD and comparison of the translated sequence in all reading frames to the sequences in the databases revealed a new potential open reading frame, designated thnE, that showed similarity to hydratases involved in biodegradation pathways. The start codon of thnE appears to be GUG. Downstream of thnE, a new open reading frame was identified, whose predicted amino acid sequence was similar to the sequences of aldolases; this open reading frame was designated thnF (Fig. 1). The start codon of thnF is just 5 nucleotides downstream of the stop codon of thnE, and the putative Shine-Dalgarno sequence of thnF is within the coding region of thnE, resulting in probable translational coupling between the two genes.

FIG. 1.

(A) Genomic region of strain TFA involved in tetralin biodegradation, showing the divergent thnC and thnD genes and the newly identified thnE and thnF genes. The arrows represent divergent operons. The triangles represent locations of KIXX insertions in mutants unable to grow on tetralin as the only carbon source. aa, amino acids. (B) Diagram of the thn genes cloned in the vector pBluescript II KS, which are transcribed from the φ10 promoter of phage T7 (large arrow).

A dendrogram resulting from a comparison of amino acid sequences exhibiting significant similarity to the ThnE amino acid sequence is shown in Fig. 2. The enzymes can be divided into four groups according to their similarity relationships. In general, members of one group show 30 to 40% identity with members of the other groups. Groups I and II comprise hydratases involved in degradation of monoaromatic or polyaromatic compounds. However, the biochemically characterized enzymes of these two groups catalyze the same reaction, which consists of hydration of 2-hydroxypenta-2,4-dienoate (9, 15). Enzymes belonging to group III are oxalocrotonate decarboxylases. These enzymes produce 2-hydroxypenta-2,4-dienoate by decarboxylation of oxalocrotonate in the dehydrogenative route of meta-cleavage pathways of aromatic compounds and may be physically associated with the hydratases that catalyze the next reaction (9, 28). Group IV, which includes ThnE, comprises hydratases involved in degradation of homoprotocatechuate and 4-hydroxyphenylacetate, which hydrate 2-oxohept-3-ene-1,7-dioic acid (17, 21), and hydratases potentially involved in degradation of ethylbenzene and biphenyls, which have not been biochemically characterized (14, 18). ThnE showed more than 50% identity with all other proteins belonging to the same group, although the level of identity was highest (62.4% over the whole length) with EtbE/BphE from Rhodococcus sp. strain RHA1 (14).

FIG. 2.

Dendrogram showing the best tree obtained by the neighbor-joining method from the alignment of 27 sequences showing significant similarity to the hydratase ThnE. Scale bar = 0.1% divergence. The ThnE sequence is enclosed in a box. The GenBank accession numbers for other hydratases are as follows: EtbE/BphE Rsp RHA1, D78322; BphE Nar pNL1, XylI Nar pNL1, and XylJ Nar pNL1, AF079317; HpaH Sdu 2229, AF144422; HpaH Eco W, Z37980; HpcG Eco C, X81446; TbhG Bce AA1, AF001356; AtdE Asp YAA and AtdH Asp YAA, AB008831; DmpE Ppu CF600 and DmpH Ppu CF600, X60835; NahK Pst AN10 and NahL Pst AN10, AF039534; XylI Ppu PWW0, M94186; PhnK Psp DJ77, AF073359; CumE Pfl IP101, D63377; IpbE Ppu RE204, AF006691; BphE Psp LB400, X76500; BphX1 Pps KF707, D85853; XylJ Ppu PWW0, M64747; TodG Ppu F1, U09250; BphE Psp KKS102, D16407; PhnH Psp DJ77, U97697; Cmt Ppu F1, U24215; MhpD Eco K12, D86239; and MhpD Cte TA441, AB024335.

Comparison of ThnF to other aldolases in the databases showed similarity to aldolases involved in degradation of homoprotocatechuate and 4-hydroxyphenylacetate, which use 2,4-dihydroxyhept-2-ene-1,7-dioic acid as the substrate (17, 29); aldolases potentially involved in degradation of ethylbenzene and biphenyls (14, 18); and other known or presumed aldolases from different bacteria. Interestingly, BLAST analysis did not reveal significant homology of ThnF to other aldolases involved in the degradation of aromatic compounds, which use 4-hydroxy-2-oxovalerate as the substrate. These enzymes are 80 to 90 residues larger than ThnF. Forced alignment of ThnF and 4-hydroxy-2-oxovalerate aldolases revealed several large gaps spread along the alignment, which account for the different sizes of the proteins, and the global alignment score was always negative (data not shown).

Characterization of a thnF insertion mutant.

A collection of nonpolar mutants formed by insertion of the KIXX cassette at different locations of the genomic region containing the thn genes was constructed previously (11) (Fig. 1). The cassette in mutant K7 was inserted into the SalI site of thnF, resulting in truncation of the gene beyond codon 169. Mutant K7 was unable to use tetralin as a sole carbon source, indicating that the aldolase encoded by thnF was essential for tetralin biodegradation. Wild-type and mutant K7 strains were grown in mineral medium containing 8 mM β-hydroxybutyrate plus tetralin provided in the gas phase. These conditions allowed growth of mutants unable to metabolize tetralin, as well as expression of thn genes. Crude extracts from the two strains were used to transform OCHBDA, the ring fission product of the meta-cleavage pathway of tetralin, and the substrate of the hydrolase ThnD. Hydrolysis of OCHBDA results in production of 2-hydroxydec-2,4-diene-1,10-dioic acid (10). Further transformation by a hydratase and an aldolase should produce pyruvate and a semialdehyde. Transformation of 36 μM OCHBDA by crude extracts of the wild-type strain resulted in accumulation of 14 μM pyruvate, 39% of the total pyruvate which could result from complete transformation of OCHBDA. However, transformation of OCHBDA by crude extracts of mutant K7 did not result in accumulation of detectable pyruvate. These results suggested that pyruvate is produced in the degradation pathway of tetralin and that ThnF is required for this production.

Heterologous production of ThnE and ThnF and characterization of their functions.

In order to confirm the production of pyruvate by ThnF and to characterize in more detail the substrates and products of the reactions catalyzed by ThnE and ThnF, these proteins, together with the hydrolase ThnD, were produced in E. coli, and the reactions catalyzed by them were characterized. To do this, plasmids pIZ674 and pIZ675 were constructed (Fig. 1) and introduced into the overproducing strain NCM631/pIZ227 (7).

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the total proteins obtained from the overproducing strains after induction revealed a single prominent band, which corresponded to the hydrolase ThnD. ThnE and ThnF were not produced in sufficient amounts to be visually detected in gels (data not shown). This was probably due to low rates of translation, since the start codon of thnE is GUG and its Shine-Dalgarno sequence is very poor. A similarly poor Shine-Dalgarno sequence precedes the start codon of thnF. In addition, thnF may be translationally coupled to thnE. Nevertheless, both ThnE and ThnF were produced in sufficient amounts to detect their activities (see below).

Crude extracts of the overproducing strain enriched for ThnD and ThnE were used to transform OCHBDA. After all OCHBDA had been transformed, the potential products were trimethylsilylated to distinguish between the enol and keto tautomers of the putative hydrated product. GC-MS analysis showed three major peaks, two of which were the trimethylsilylated enol and keto tautomers of the hydrolysis product, 2-hydroxydec-2,4-diene-1,10-dioic acid and 2-oxodec-4-ene-1,10-dioic acid, respectively (data not shown). The mass spectrum of the third compound detected showed a weak peak at m/z 520 that was thought to be the molecular ion, whose molecular mass matched that of the trimethylsilylated enol tautomer of the putative hydrated product, 2,4-dihydroxydec-2-ene-1,10-dioic acid (DHDDA). A more abundant ion at m/z 505 resulted from the loss of one methyl group, which is characteristic of trimethylsilylated compounds, and suggested that the molecular ion of the compound should be at m/z 520 (data not shown). Since the molecular mass of the trimethylsilylated keto tautomer of the hydrated product, 2-oxo-4-hydroxydec-1,10 dioic acid, is 448 Da, the presence of these peaks with significantly higher molecular masses indicated that the compound was the enol tautomer DHDDA. The base peak of the spectrum at m/z 333 corresponded to a four-carbon molecule with a double bond, one carboxyl group, and two hydroxyl groups, and all three groups were trimethylsilylated ([-CH(OTMSi)—CH=C(OTMSi)—CO(OTMSi)]⁺t), which resulted from cleavage of the molecular ion between C-4 and C-5. The ion at m/z 147 resulted from a typical rearrangement of two trimethylsilyl groups which were close together and corresponded to (CH₃)₃—Si=O⁺—Si(CH₃)₃. The mass spectrum is consistent with a dicarboxylic C₁₀ compound having two hydroxyl groups which were also trimethylsilylated and, therefore, indicates that DHDDA is the major product resulting from the hydration reaction.

To characterize the products of the reaction catalyzed by the aldolase ThnF, extracts of the overproducing strain enriched in ThnD, ThnE, and ThnF were also used to transform OCHBDA. GC-MS of the methylated reaction products showed that the keto tautomer of the hydrolysis product was still evident, but neither its enol form nor any of the tautomers of the hydrated product was detected (data not shown), showing that the aldolase ThnF had fully transformed its substrate. The most abundant reaction product had a mass spectrum with the largest peak at m/z 158, which matched the molecular mass of methylated pimelic semialdehyde (data not shown). However, the peak at m/z 157, which corresponded to M⁺-1, was clearly more abundant. The peak at m/z 128 corresponded to M⁺-1 which had lost an aldehyde group. The base peak at m/z 115 corresponded to the molecular ion which had lost CHO—CH₂. The ion at m/z 74 resulted from a MacLafferty rearrangement and provided evidence of a methylated carboxylic group. To confirm the identity of the most abundant reaction product, chemically synthesized pimelic semialdehyde was methylated and analyzed in a similar way. The mass spectrum of methylated pimelic semialdehyde (data not shown) confirmed that this was one of the products of the reaction catalyzed by ThnF.

The second product of the cleavage reaction, pyruvate, was not detected by GC-MS, probably because under the chromatographic conditions it eluted with the solvent front. However, transformation of 82 μM OCHBDA by extracts of E. coli enriched in ThnD, ThnE, and ThnF resulted in accumulation of 21 μM pyruvate, as estimated spectrophotometrically by using lactate dehydrogenase. In contrast, transformation of OCHBDA by extracts containing just ThnD and ThnE did not result in pyruvate accumulation, showing that ThnF is required for pyruvate production.

DISCUSSION

Two genes, designated thnE and thnF, which code for a hydratase and an aldolase, respectively, were identified downstream of the previously characterized thnD gene, which codes for a hydrolase required for tetralin biodegradation (10) (Fig. 1). Identification was initially based on an amino acid sequence comparison of the products and was subsequently confirmed by characterizing the substrates and products of the reactions catalyzed by the enzymes.

Sequence alignment of ThnE and similar hydratases showed that they can be classified in four groups according to similarity. Groups I and II comprise 2-hydroxypentadienoic acid hydratases involved in catabolic pathways for different aromatic compounds. The best-characterized enzymes are XylJ encoded by plasmid pWW0 (group I) and MhpD from E. coli (group II). The substrate of both enzymes is the enol form 2-hydroxy-2,4-pentadienoate, and ketonization to 2-oxo-4-pentenoate, previous to the hydration reaction, is catalyzed by the hydratase in a highly stereoselective way (9, 15, 28). It is interesting that in spite of the fact that these hydratases catalyze hydration of the same substrate in similar ways, the sequence divergence among proteins belonging to the two groups is very high, as high as the sequence divergence with enzymes which are not even hydratases (group III in Fig. 2).

On the other hand, hydratases belonging to group IV use a substrate that is significantly different, since it is longer and dicarboxylic. Based on the isomerase activity ascribed to the decarboxylase HpcE, which ketonizes 2-hydroxyhept-2,4-diene-1,7 dioic acid to 2-oxohept-4-ene-1,7-dioic acid, it has been proposed that the substrate of the hydratase HpcG is the keto tautomer (19, 20, 21). ThnE also hydrates a long, 10-carbon, dicarboxylic substrate. When the products of the sequential hydrolysis and hydration reactions were analyzed, both the dienol and the keto tautomers of the hydrolysis reaction were detected; these compounds could be substrates of ThnE. Interestingly, characterization of the activity of purified hydrolase ThnD showed that the dienol tautomer is the only reaction product detected (10). The presence of both tautomers in the coupled reactions with crude extracts may suggest that the hydratase could ketonize the dienol form to some extent prior to its hydration. This would be in contrast to what has been proposed for the similar hydratase HpcG. However, the evidence is not strong, and both tautomers are presented in the catabolic pathway shown in Fig. 3. On the other hand, analysis of the hydrated product identified it as the enol form, DHDDA, while its keto form was not detected. This is consistent with the enol form of the product of the hydration reaction catalyzed by HpcG (group IV) and in contrast to the keto form detected after hydration of 2-hydroxy-2,4-pentadienoate by more distantly related hydratases belonging to groups I and II.

FIG. 3.

Characterized reactions of the tetralin catabolic pathway. 1,2-DHT, 1,2-dihydroxytetralin; HDDDA, 2-hydroxydec-2,4-diene-1,10-dioic acid; ODDA, 2-oxodec-4-ene-1,10-dioic acid.

Sequence comparison of the aldolase ThnF revealed significant similarity to aldolases different from 4-hydroxy-2-oxovalerate aldolases involved in catabolic pathways of different aromatic compounds, which are partners of the hydratases belonging to groups I and II. The substrates of the aldolases similar to ThnF are 2,4-dihydroxy-2-ene-dicarboxylic acids of different lengths, which are transformed into pyruvate and a semialdehyde (12, 20). Consistent with this, characterization of the reaction catalyzed by ThnF indicated that the products obtained from DHDDA are pyruvate and the seven-carbon molecule pimelic semialdehyde.

The best-characterized enzyme among the aldolases similar to ThnF is DdgA, encoded by the locus garL, which catalyzes the reversible aldol cleavage of 2-dehydro-3-deoxygalactarate to pyrvate and tartronic semialdehyde. DdgA is one of the class II aldolases, which require a divalent cation for activity, although it has a novel reaction mechanism (20). Its crystal structure reveals the importance of a number of residues which are metal ligands (Asp179 and Glu153) and are in van der Waals contact with the three carbons of pyruvate (Gly176, Ser178, and Leu216) or with the aldehyde (His50 and Arg75) (20). These residues, including the metal ligands, are conserved in ThnF and in most other similar aldolases. The high level of conservation of the active site residues of DdgA, including the metal ligands, indicates that these enzymes are class II aldolases and may have the novel reaction mechanism determined for DdgA. This mechanism was studied in the reverse condensation reaction and showed that DdgA first catalyzes the enolization of pyruvate and in a second step catalyzes the condensation reaction. This clearly indicates that the substrate of the forward cleaving reaction must be in its enol form and that C-2 of pyruvate is ketonized after the substrate is cleaved. Consistent with this, the characterized substrates of the similar aldolases HpcH from E. coli (21) and ThnF are the enol forms.

In contrast, the substrate of 4-hydroxy-2-oxovalerate aldolases is the keto form. At least MhpE has been shown to be one of the class I aldolases (16), which do not require a metal cation and utilize an imine linkage between a lysine of the active site and the C-2 carbonyl of the substrate. Since the overall similarity of 4-hydroxy-2-oxovalerate aldolases is very high (about 80% identity), it is probable that all of them are class I aldolases. Therefore, aldolases in the catabolic pathways of aromatic compounds can be divided into two broad classes of aldolases, which do not exhibit significant sequence similarity and whose reaction mechanisms should be different.

The results presented in this paper extend the characterization of the tetralin biodegradation pathway from extradiol cleavage of the catechol derivative up to products which are general metabolic compounds (Fig. 3), although pimelic semialdehyde should be further metabolized by an aldehyde dehydrogenase which remains unidentified. Our results also show that genes having apparently different origins have been recruited to generate this catabolic pathway.