The Broad Substrate Chlorobenzene Dioxygenase and cis-Chlorobenzene Dihydrodiol Dehydrogenase of Pseudomonas sp . Strain P 51 Are Linked Evolutionarily to the Enzymes for Benzene and Toluene Degradation *

The chlorobenzene degradation pathway of Pseudomonas sp. strain P51 is an evolutionary novelty. The first enzymes of the pathway, the chlorobenzene dioxygenase and the cis-chlorobenzene dihydrodiol dehydrogenase, are encoded on a plasmid-located transposon Tn5280. Chlorobenzene dioxygenase is a four-protein complex, formed by the gene products of tcbAa for the large subunit of the terminal oxygenase, tcbAb for the small subunit, tcbAc for the ferredoxin, and tcbAd for the NADH reductase. Directly downstream of tcbAd is the gene for the cis-chlorobenzene dihydrodiol dehydrogenase, tcbB. Homology comparisons indicated that these genes and gene products are most closely related to those for toluene (todC1C2BAD) and benzene degradation (bedC1C2BA and bnzABCD) and distantly to those for biphenyl, naphthalene, and benzoate degradation. Similar to the tod-encoded enzymes, chlorobenzene dioxygenase and cis-chlorobenzene dihydrodiol dehydrogenase were capable of oxidizing 1,2-dichlorobenzene, toluene, naphthalene, and biphenyl, but not benzoate, to the corresponding dihydrodiol and dihydroxy intermediates. These data strongly suggest that the chlorobenzene dioxygenase and dehydrogenase originated from a toluene or benzene degradation pathway, probably by horizontal gene transfer. This evolutionary event left its traces as short gene fragments directly outside the tcbAB coding regions.

Bacteria that are able to use mono-or dicyclic aromatic compounds as their sole source of carbon and energy under aerobic growth conditions are present ubiquitously in the environment (1,2). Considering the potential use and importance of such bacteria to help to remove many man-made polluting compounds, it is necessary to study the genetic and biochemical variations found among these types of bacteria and to investigate their evolutionary development (3). Only then can the limitations of existing metabolic pathways be understood and concepts be developed to select or engineer novel pathways (4,5). Of special interest will be to obtain degradation of highly recalcitrant chlorosubstituted aromatic compounds, such as chlorinated benzenes and biphenyls (6).
A very important enzyme complex in the aerobic degradation of many aromatic compounds is the multicomponent aromatic ring dioxygenase (7,8). Aromatic ring dioxygenases, such as benzoate dioxygenase (9), toluate dioxygenase (10), naphthalene (11), biphenyl, toluene (12), or benzene dioxygenases (13), are enzyme complexes with three-or four protein subunits. This complex catalyzes a redox reaction in which molecular oxygen is incorporated in the aromatic ring at the expense of the oxidation of NADH (7,8,12). The resulting intermediate is a cis-dihydrodiol derivative of the aromatic ring structure. A dihydrodiol dehydrogenase then catalyzes the (formal) oxidation of the dihydrodiol to a dihydroxy derivative and regenerates the reduced NADH. The components of the aromatic ring dioxygenase consist of different electron transport proteins (a ferredoxin and a reductase, or a combined ferredoxin-NADH reductase) and the terminal oxygenase (also called hydroxylase component or iron sulfur protein), which is thought to determine the substrate specificity of the enzyme and to carry out the substrate activation (7,8). Despite their structural similarities, remarkable differences in substrate spectrum are found among the different aromatic ring dioxygenases (12,14,15). We have focused on bacteria-degrading chlorinated benzenes, in particular Pseudomonas sp. strain P51. Upon growth on chlorobenzenes, strain P51 induces enzyme activities which catalyze the conversion of chlorobenzenes to chlorocatechols (16). The genes for these enzymes, the tcbAB genes, were cloned previously from a catabolic plasmid present in strain P51 and were proposed to encode an aromatic ring dioxygenase and a dihydrodiol dehydrogenase, similar to the enzymes of other aromatic pathways (16). Strain P51 also contains the genes for a so-called chlorocatechol oxidative pathway, which in strain P51 consists of an operon of four genes, tcbCDEF (17), and a regulatory gene tcbR (18). Interestingly, both gene clusters in strain P51 are located on a transmissible plasmid pP51. Furthermore, the tcbAB genes itself are part of a transposable element, Tn5280 1 (19). We therefore strongly believe that the chlorobenzene pathway is an evolutionary novelty in bacteria, formed by a novel combination of two existing gene clusters, perhaps through horizontal gene transfer.
To test this idea further, we wanted to characterize the genes for the chlorobenzene dioxygenase and for the cis-chlorobenzene dihydrodiol dehydrogenase of Pseudomonas sp. strain P51 in detail. We wanted to analyze further the capability of the enzymes to convert several different aromatic compounds. This would make it possible to compare the enzymes both genetically and biochemically with related enzymes from other aro-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EMBL Data Bank with accession number(s) U15298.
‡ To whom correspondence should be addressed. Tel.: 41-1-823-5438; Fax.: 41-1-823-5547; E-mail: vdmeer@eawag.ch. matic degradation pathways. The data in the paper indicate clearly that the chlorobenzene dioxygenase and the cis-chlorobenzene dihydrodiol dehydrogenase resemble most the toluene/ benzene-type enzymes. Furthermore, evidence is presented to show the relics of the horizontal gene transfer events which may have lead to the imprecise excision of a genetic element containing the genes for an aromatic ring dioxygenase and dihydrodiol dehydrogenase from a tod-like operon.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Growth Conditions-Pseudomonas sp. strain P51 has been described previously (16). Escherichia coli strains DH5␣ and TG1 were used routinely for plasmid cloning and single-stranded M13 phage preparation, respectively (20). E. coli BL21(DE3) was used for T7 RNA polymerase-directed expression of genes and gene fragments cloned in plasmid pET8c (21). We used the plasmids pUC18 and pUC19 (obtained from Boehringer Mannheim, Mannheim, Federal Republic of Germany) as cloning vectors for strain P51-derived DNA fragments. In general, E. coli strains were grown on LB medium at 37°C (20), supplemented with the appropriate antibiotics. For expression of active chlorobenzene dioxygenase and cis-chlorobenzene dihydrodiol dehydrogenase, however, the E. coli strains were cultivated at 25°C.
DNA Techniques and Sequence Analysis-All DNA techniques, such as plasmid DNA isolation, transformations, or DNA-enzyme digestions, were carried out according to established procedures described elsewhere (20). DNA sequence analysis was performed on both strands of the DNA by sequencing overlapping fragments cloned in M13mp18, as described elsewhere (17). The source of the DNA fragment containing the chlorobenzene dioxygenase and dihydrodiol dehydrogenase genes of strain P51 was plasmid pTCB60 (16). Restriction enzymes and other DNA-modifying enzymes were purchased from Life Technologies Europe (Paisley, UK), Appligene (Illkirch, France), or Boehringer Mannheim (Mannheim, FRG). Reagents for the polymerase chain reaction were obtained from Life Technologies Inc.
Construction of tcb Expression Clones-To obtain overexpression of the individual components of the dioxygenase system and the dehydrogenase, we constructed a number of translational fusions with the start codon present on pET8c ( Fig. 1) (21). Hereto, artifical NcoI sites were created on the DNA to be cloned by applying PCR amplification in the presence of a mutagenic primer. For clones starting with the tcbAa gene, we amplified a 200-bp DNA fragment ranging from the start of tcbAa until the first SphI site downstream. The PCR fragment was then cleaved with NcoI and SphI and ligated with a 2-kb SphI-ScaI fragment of plasmid pTCB71 (16) and, with pET8c, cut with NcoI and EcoRV. After transformation, this resulted in plasmid pTCB113. The sequence of the PCR-amplified part of this plasmid was determined and found to be identical with that of the wild-type tcbAa gene. From plasmid pTCB113 we then constructed pTCB115 by exchanging the 1.2-kb MluI-ScaI fragment from pTCB113 with the 2.0-kb MluI-PstI fragment, which contains tcbAb, tcbAc, and part of tcbAd. In plasmid pTCB114 the frame of the tcbAb gene was interrupted by cutting the DNA with AatII, filling the ends by using Klenow enzyme, and religating. Plasmid pTCB147 was constructed from pTCB115 by cutting with SacII, removing the 1.0-kb fragment, creating blunt ends with T4 DNA polymerase treatment, and religating. Clones starting with the tcbAc gene were made by introducing a NcoI site at the start codon of tcbAc similarly as described above. A small region of the DNA was amplified by PCR between the start codon of tcbAc and the first downstream SalI site, then digested with NcoI and SalI, and ligated with a 1.3-kb SalI-EcoRI fragment containing the region of tcbAc and tcbAd and with pET8c, cut with NcoI and EcoRI. This resulted in plasmid pTCB117. A DNA fragment containing the complete tcbAd and tcbB sequence was then introduced in plasmid pTCB117 to form pTCB120. A deletion derivative was created which lacked the complete tcbB sequence by cutting this plasmid with BamHI and SacI, removing protruding ends, and religating (pTCB116). Finally, a clone containing only tcbB was made by cloning a 1.2-kb EcoRI-MroI fragment of pTCB119 in pUC19 cut with EcoRI and SmaI. This plasmid was named pTCB149.
The pET8c-derived plasmids did not, unfortunately, express active Tcb enzymes in E. coli (see below). Therefore, we constructed a clone containing the complete tcbAaAbAcAd gene sequence in pUC19. Plasmid pTCB115 was hereto cut with MluI and BamHI, and a 4.7-kb fragment was isolated and ligated with a 3.0-kb MluI-BamHI fragment of pTCB60. After transformation, this resulted in plasmid pTCB130. We then removed the complete 4.2-kb XbaI-BamHI insert of plasmid pTCB130 and ligated this fragment into pUC19, cut with XbaI and BamHI. This resulted in plasmid pTCB144. Colonies of E. coli containing this plasmid turned blue green when growing on LB plates at 30°C, which indicates synthesis of active dioxygenase in the cells (11).
Overproduction of Individual Components of the Dioxygenase and Dehydrogenase-To test if the observed ORFs could be translated to proteins of the predicted size, the pET8c-derived plasmids containing the various parts of the tcb genes were tested for expression in E. coli BL21(DE3). These cultures were grown on LB medium at 37°C to an A 550 of 0.5. We then induced T7-mediated gene expression by adding isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM to the cell culture. Cultures were then allowed to grow for another 2 h, after which the cells were harvested from a 1-ml sample. The obtained cell pellet was resuspended in 50 l of protein sample buffer according to Laemmli (50), and 5-10 l were analyzed on 12.5% SDS-PAGE.
Analysis of Chlorobenzene Dioxygenase and Chlorobenzene Dihydrodiol Dehydrogenase Activity-Chlorobenzene dioxygenase activity was tested by analyzing the formation of dihydrodiol intermediates from aromatic substrates incubated with washed cell suspensions of E. coli DH5␣ (pTCB144). The cultures were inoculated from a single colony in 200 ml of LB medium with 50 g/ml ampicillin and grown at 25°C for about 36 h. The optical density of the cultures had then reached an A 550 of 3.5. The cell culture was then centrifuged at 4000 rpm for 10 min at 20°C, and cells were resuspended in 50 ml of M9 mineral salts medium (20). This was repeated once more, and, after a final centrifugation step, the cells were resuspended in 10 ml of M9 medium and stored briefly on ice until use. A series of glass-stoppered tubes with a volume of 15 ml were filled with the following reagents: 4.5 ml of M9 containing 1 mM glucose, 50 l of a methanol solution containing the aromatic substrates, and 0.5 ml of the cell suspension. The final A 550 of the cells in the assay was between 0.7 and 1.0. The tubes were incubated on a rotary shaking platform at a temperature of 30°C. We tested the following aromatic substrates at a final concentration in the assay of 0.5 mM: 1,2-dichlorobenzene, toluene, biphenyl, naphthalene (all previously dissolved in methanol), and benzoate. For each time point of the assay, two tubes were incubated. Samples of 1.0 ml were then taken from the tubes and centrifuged for 1 min at 13,000 rpm to remove the cells. The supernatant was transferred to a fresh tube and analyzed for the presence of dihydrodiol intermediates by HPLC (see below).
cis-Chlorobenzene dihydrodiol dehydrogenase activity was tested in E. coli DH5␣ (pTCB149). Cultures were grown on 50 ml of LB medium at 37°C to an A 550 of 1.0, after which the cells were harvested and a cell extract was prepared as described previously (16). The reaction mixture for dihydrodiol dehydrogenase activity contained 0.65 ml of 20 mM sodium phosphate buffer, pH 7.5, 25 lof20mMNAD ϩ solution, 50 l of the cell extract, and 50 l of dihydrodiol substrate. As substrates we used the supernatants of the whole cell incubations after 2 h (see above). The assay mixture was incubated at 37°C, and the change in A 340 was measured on a spectrophotometer. When no more changes in A 340 were observed, the assay mixture was analyzed on HPLC to check for the disappearance of the dihydrodiol and the presence of the dihydroxy compound.
Dihydrodiol intermediates were extracted from supernatants of the whole cell incubations after 2 h with an equal volume of ethyl acetate and dried with sodium sulfate. Samples were derivatized with BSTFA and subjected to GC-MS analysis as described elsewhere (22).

RESULTS
Sequence Determination of the tcbAB Genes-We determined the nucleotide sequence of the region containing the tcbAB genes of Pseudomonas sp. strain P51 on both strands of the DNA. The tcbAB genes are located on a stretch of 5,402 base pairs which lay between IS1066 and IS1067 (19) (Figs. 1 and 2). The region showed the presence of five large unidirectional ORFs, encoding the different subunits of the chlorobenzene dioxygenase and the dihydrodiol dehydrogenase. Sequence homologies with other known dioxygenases allowed the assignment of putative protein functions to each of the ORFs (Table  I). We propose to designate the genes as follows: tcbAa, coding for the large subunit of the terminal oxygenase; tcbAb, encoding the small terminal oxygenase subunit; tcbAc, the ferredoxin; tcbAd, the NADH reductase; and tcbB, the dihydrodiol dehydrogenase. Except for a small 109-bp gap between tcbAa and tcbAb and 8 bp between tcbAb and tcbAc, the ORFs were contiguous on the DNA. Downstream of tcbB, another ORF was found, which showed homology to catechol 2,3-dioxygenases, such as todE (23). This ORF, however, appeared to be interrupted by IS1067, causing a premature ending (Fig. 2).
Homologies with Other Bacterial Aromatic Ring Dioxygenases and Dihydrodiol Dehydrogenases-The amino acid sequences predicted for the TcbAa, -Ab, -Ac, -Ad, and TcbB proteins were compared with those from other bacterial aromatic ring dioxygenases by using the GCG programs FASTA, DIS-TANCES, and PILEUP (24,25). The alignments and distance calculations showed for almost every individual component of the dioxygenases and for the dihydrodiol dehydrogenase a clustering in four different families (Fig. 3 and Table II). One family is formed by the dioxygenases which are composed of two components (and three protein subunits), i.e. benzoate dioxygenase of Acinetobacter calcoaceticus (9) and toluate dioxygenase of Pseudomonas putida (10). A second family contains the three-component dioxygenases of naphthalene metabolism, such as those encoded by the nah (26), ndo (27), pah (28), or dox (29) genes. The third family is composed of the dioxygenases for biphenyl and chlorobiphenyl conversion in Gramnegative bacteria (30 -32), and a fourth one of the benzene (13,33), toluene (23), and chlorobenzene dioxygenases. Two recently published sequences for a biphenyl dioxygenase from two Gram-positive microorganisms aligned more closely with the toluene/benzene family than with the biphenyl family itself (34,35) (Fig. 3). The only exception in the alignments was the clustering of the reductase components. In this case, the positions of the reductases from P. paucimobilis KKS102, Rhodococcus sp. RHA1, and R. globerulus P6 appeared to be intermediate (Table II). In general, the reductases seem to have diverged substantially more than the other components of the dioxygenases and the dihydrodiol dehydrogenase.
It is interesting that to a large extent the gene organization within the clusters has been conserved as well, but clearly differs between them. For example, the benzene/toluene family has the gene order: large subunit of the terminal oxygenase, small subunit, ferredoxin, reductase, and, in three cases, dehydrogenase. The group of the Gram-positive biphenyl dioxygenases, which appeared to be the cluster closest to that of the benzene/toluene dioxygenases, has an identical organization of the genes encoding the core dioxygenase, but differs in genes located downstream. The family of the Gram-negative biphenyl dioxygenases has a genetic organization comparable to that of the benzene/toluene family, but in two cases contain an extra ORF between the genes for the terminal oxygenase and that for the ferredoxin. Also in these cases, the upstream regions lack homology between each other or with the corresponding regions of the benzene/toluene family. The biphenyl dioxygenase of strain KKS102 lacks the extra ORF, but also lacks the gene for the reductase at this position (Fig. 3). A stronger difference is found with the family of the naphthalene dioxygenases. Here the gene order is reductase, ferredoxin, large subunit of the terminal oxygenase, small subunit, dehydrogenase. In the doxencoded system, no (clustered) gene for a reductase was described (29), but this may similarly be located directly up-stream. The naphthalene reductases (i.e. nahAa and pahAa) are also substantially shorter than those of the others (329 amino acids versus approximately 410) and biochemically different (7). In the case of the benzoate and toluate dioxygenase, the largest difference in gene order with the others is found in the presence of a gene for the combined ferredoxin and reductase function (i.e. benC and xylZ) (8,9).
Expression of the tcb-encoded Gene Products-We cloned all open reading frames from the tcbAB cluster in the expression vector pET8c under transcriptional control of the T7 promoter in E. coli BL21(DE3), to test if the gene products would have the size as predicted from the amino acid sequence. Upon induction in E. coli, we could detect all predicted protein bands and deletion derivatives on SDS-PAGE (Fig. 4). Interestingly, in some cases, read-through from one ORF into the other occurred only sparsely. For example, the tcbAb gene could not be visibly expressed in clones containing tcbAa upstream of tcbAb. Only by using plasmids in which part of tcbAa was deleted, such as pTCB147, we found detectable expression of tcbAb.On the other hand, clones starting with tcbAc would also express downstream ORFs when present, such as in plasmid pTCB116 and pTCB120. Most protein bands observed on SDS-PAGE which were attributed to expression from a tcb gene were of the size expected from computer predictions ( Table I). The exception was TcbB, which migrated at a smaller apparent molecular mass than predicted (22 kDa instead of 33.1 kDa).
Chlorobenzene Dioxygenase and Dihydrodiol Dehydrogenase Activity in E. coli-The functionality of the tcbAB gene products was then tested by measurement of their enzymatic activity in E. coli. We cloned the complete DNA fragment with the tcbAaAbAcAd genes starting at the ATG codon of the tcbAa gene in pET8c. To our surprise, we could not detect any measurable activity of the chlorobenzene dioxygenase with this plasmid (not shown). We think that this may be caused by the unbalanced expression of the different ORFs when expressed from the T7 promoter (see above) and by the formation of inclusion bodies. The genes were then removed from pET8c and cloned in pUC19 under control of the lac promoter (pTCB144). E. coli (pTCB144) showed the typical formation of blue-green colonies when grown on LB agar, due to the formation of indigo. This color became more pronounced when the colonies were incubated at 25°C.
Whole cells of E. coli (pTCB144) were incubated with different aromatic substrates in minimal medium in order to produce the cis-dihydrodiols. E. coli (pTCB144) cells rapidly produced one single metabolite as detected by HPLC, when incubated with 1,2-dichlorobenzene, toluene, biphenyl, or naphthalene (Fig. 5). No conversion of benzoate was detected with these cells. The UV spectra of the intermediates of toluene, naphthalene, and biphenyl incubation on HPLC were in agreement with the max values published previously for the corresponding dihydrodiols (36 -38). For the product of 1,2-dichlorobenzene incubation, we observed a UV spectrum similar to that of toluene, although with a max at 272 nm. Further information on the identity of the four intermediates was obtained by GC-MS analysis of the BSTFA-derivatized form (Fig. 6). All   3 Si] ϩ ) dominate the four spectra. In the case of the product of dichlorobenzene, the ion at m/z 289 is formed by the loss of chlorine (mass of 35) from the molecular ion. Following the line of evidence for the formation of cis-dihydrodiols from aromatic compounds by toluene and naphthalene dioxygenase (36 -38), for example, all our results are in agreement with the proposed cis-dihydrodiol structures which would be formed during conversion of toluene, 1,2-dichlorobenzene, biphenyl, and naphthalene by the chlorobenzene dioxygenase.
The dihydrodiols were then incubated with cell extracts of E. coli (pTCB149), which expresses the dihydrodiol dehydrogenase, and analyzed by HPLC. In the case of 1,2-dichlorobenzene, we found that the dihydrodiol was converted to one single product. This product cochromatographed with authentic 3,4dichlorocatechol, and the UV spectra of the two compounds were identical. Biphenyl dihydrodiol was enzymatically converted to a compound with an identical retention time and UV spectrum as 2,3-dihydroxybiphenyl and, similarly, toluene dihydrodiol to a compound with identity to 3-methylcatechol. In the case of naphthalene 1,2-dihydrodiol, the presumed product 1,2-dihydroxynaphthalene could not be detected, as it became autooxidized quickly as described earlier (39,40). DISCUSSION Chlorobenzene Dioxygenase Belongs to the Toluene and Benzene Dioxygenase Subclass-Pseudomonas sp. strain P51 has the ability to use chlorinated benzenes as sole carbon and energy source. The enzyme catalyzing the initial dioxygenation of the aromatic ring structure was presumed to be a threecomponent aromatic ring dioxygenase, like those found in other aerobic bacteria. Here we have shown that the genes for the chlorobenzene dioxygenase are contiguous on the DNA and indeed code for four protein subunits, two of which make up the terminal oxygenase, one the ferredoxin, and the last one the NADH reductase. Following the genes for the dioxygenase is a gene coding for a dihydrodiol dehydrogenase. All genes were shown to be functional by using expression studies and enzyme activity assays. Both biochemical and genetic evidence indicate that the chlorobenzene dioxygenase belongs to a subclass of aromatic ring dioxygenase enzymes to which the toluene and benzene dioxygenases also belong.
Our studies with the Tcb dioxygenase showed that it is not specific for catalyzing the conversion of 1,2-dichlorobenzene only, but capable of converting toluene, naphthalene, and biphenyl. Benzoate was not converted by the Tcb dioxygenase, which is like other characterized three-component aromatic ring dioxygenases (8,12). The outcome of the whole cell incu- bations suggests that naphthalene and biphenyl are converted even faster than 1,2-dichlorobenzene and toluene. However, we do not know in what way solubility of these compounds, different uptake, and excretion rates influence this outcome. We cannot exclude that other products, such as other stereoisomers or monohydroxylated rings, were formed in the incubations, since no attempts were made to determine a mass balance in the whole cell incubations. Furthermore, we did not determine the absolute stereoconfiguration of the products. HPLC analysis, however, suggested the formation of single intermediates. UV and mass spectrum of these compounds were in agreement with the structures of cis-dihydrodiols, as published previously (36 -38). In the subsequent enzyme incubations, we obtained good evidence that the TcbB dihydrodiol  (24). Circles show a graphic representation with a circle diameter propertional to the calculated distances. Circle filling indicates an observed clustering in PILEUP at a distance higher than 0.8. cross-hatched circle, benzoate/toluate family; checkered circle, naphthalene family; solid circle, biphenyl family; diagonally lined circle, toluene/ benzene family.
It is not so clear which subunit of the aromatic ring dioxygenases determines the substrate specificity of the enzyme and why the toluene/benzene subclass enzymes have such a wide substrate spectrum. The remarkable potential of the Tod enzyme to catalyze incorporation of oxygen into a wide range of aromatic substrates has been well studied and explored (12,41). For instance, the Tod dioxygenase oxidizes biphenyl, the main substrate of the group of bph-encoded enzymes. On the other hand, biphenyl dioxygenase from P. pseudoalcaligenes KF707 does not oxidize toluene (14,42). This limitation supposedly arose in the small subunit of Bph dioxygenase, because when hybrid enzymes between Tod and Bph were constructed, some were found (e.g. BphA1TodC2AB) that gained both the Bph and Tod substrate range. The Bph-dioxygenases were mostly studied for their capability to convert (poly-) chlorinated biphenyls. For example, differences in polychlorinated biphenyls-congener specificity were found between the Bph dioxygenase from Pseudomonas sp. strain LB400 and P. pseudoalcaligenes KF707, which in this case were attributed to changes in the large subunit of the terminal oxygenase (15). It will be interesting to study in more detail whether the Tcb dioxygenase has acquired any new substrate specificities which enable it to convert higher chlorinated aromatic compounds more efficiently.
Gene Rearrangements in the Evolutionary Divergence of Aromatic Ring Dioxygenases-The large pool of genetic data on aromatic ring dioxygenase systems from different aerobic bacteria makes it possible to speculate about the different events which have taken place in the course of the evolutionary development of these microorganisms (43). The accumulation of small events (e.g. mutations) has likely led to the divergence of the different individual genes and gene products, as shown in Fig. 3. However, more striking larger genetic changes have also occurred. Rearrangements on the DNA have caused differences in gene order of the aromatic ring dioxygenase. This becomes obvious when we compare the gene order of the toluene/benzene family and the biphenyl family on one side, and that of the naphthalene family on the other (Fig. 3). In the naphthalene family, the genes for the reductase and ferredoxin have inverted their position with respect to the genes encoding the terminal oxygenase. The gene for the reductase may also be at a different position, as in the bph gene cluster of P. paucimobilis strain KKS102 (31,44). Similarly, rearrangements have caused differences in the organization of genes located downstream of the aromatic ring dioxygenase, like the ones encoding the dihydrodiol dehydrogenase or the meta-cleavage enzymes (Fig. 3). These DNA rearrangements must have had their effects on gene expression and on enzyme synthesis, perhaps due FIG.6. Electron ionization mass spectra of the trimethylsilyl derivatives of the products of whole cell incubations with E. coli (pTCB144) and the following substrates. A, 1,2-dichlorobenzene; B, toluene; C, biphenyl; D, naphthalene. The structural formula of the proposed product has been drawn above the respective mass spectrum. The fragmentation pattern is discussed in the text. to improper signals on the DNA or changed stability and structure of the mRNA. For instance, how is it achieved that the right molar proportions of all components of the dioxygenase are synthesized? The benzene dioxygenase of P. putida ML2, which is transcribed from one single gene cluster with operonic organization, apparently has intracellular molar proportions of 1:0.45:0.8 of ISP-␣ subunit/ferredoxin/reductase (45). Regulation of the right molar amounts may become less obvious when the reductase gene is not directly transcriptionally coupled. This may reflect the idea that the aromatic ring dioxygenase is a kinetic enzyme complex with a rather loose association between oxygenase component, ferredoxin, and reductase (45). For the tcbAB genes it was interesting to notice that we could not see expression of tcbAb in E. coli under control of the T7 RNA polymerase from plasmid constructs on which tcbAa was still present, although expression was clearly forced upon the system in this case.
A major gene rearrangement has probably taken place in Pseudomonas sp. strain P51. In this microorganism the genes for the aromatic ring dioxygenase and the dihydrodiol dehydrogenase were most likely transposed from their original position by the activity of two insertion elements (19). The relics of this process can still be seen on the DNA (Fig. 7). Upstream of the gene tcbAa there are parts of a gene similar to todF (46), and downstream of tcbB there is a large part of a gene for a metacleavage enzyme similar to todE (23). It may have been a necessity for the organism to inactivate such a meta-cleavage enzyme since these are known in some microorganisms to interfere with the metabolically productive ortho-cleavage of chlorinated catechols, which arise as intermediates in chlorobenzene degradation (47). We believe that the strong genetic and biochemical similarity between the tcbAB and the tod system is good evidence to assume that the tcbAB genes originated in a microorganism which could use toluene or benzene as the sole carbon and energy source.