Biphenyl Dioxygenases: Functional Versatilities and Directed Evolution

Biphenyl is a compound in which two benzene rings are connected to each other. Polychlorinated biphenyls (PCBs) can be produced by the direct chlorination of biphenyl, by which 209 different compounds containing 1 to 10 chlorines can be produced. Because PCBs have been widely used for a variety of industrial purposes, these recalcitrant compounds are recognized to be some of the most serious environmental pollutants worldwide. Biphenyl-utilizing bacteria cometabolize PCBs into chlorobenzoic acids by using biphenyl-catabolic enzymes via an oxidative route (Fig. ​(Fig.1).1). Several biphenyl- and PCB-degrading bacteria, including both gram-negative and gram-positive strains, have been isolated to date (1, 18, 19, 81). Using these bacteria, many workers have studied the biochemical and genetic bases of PCB degradation in detail.

Open in a separate window

FIG. 1.

Catabolic pathway for degradation of biphenyl and organization of the bph gene cluster in P. pseudoalcaligenes KF707. Compounds: I, biphenyl; II, 2,3-dihydroxy-4-phenylhexa-4,6-diene (dihydrodiol compound); III, 2,3-dihydroxybiphenyl; IV, HOPD (biphenyl meta-cleavage compound); V, benzoic acid; VI, 2-hydroxypenta-2,4-dienoic acid; VII, 4-hydroxy-2-oxovaleric acid; VIII, pyruvic acid; IX, acetaldehyde; X, acetyl-coenzyme A. The BphR1 protein, belonging to the GntR family, is a transcriptional regulator involved in the expression of bphR1 and bphX0X1X2X3D. The function of orf3 remains unclear (83, 84). ISP, iron-sulfur protein.

Biphenyl dioxygenase (BphA) is a Resike-type, three-component enzyme, composed of a terminal dioxygenase and an electron transfer chain (Fig. ​(Fig.1)1) (12, 49). The former consists of a large subunit and a small subunit, associating as an α₃β₃ heterohexamer (11, 46). The latter consists of ferredoxin and its reductase and is involved in electron transfer from NADH to reduce the terminal dioxygenase. The terminal dioxygenase activates molecular oxygen to introduce it into the biphenyl molecule at the 2,3 position to obtain a 2,3-dihydro-2,3-diol, which is then dehydrogenated to 2,3-dihydroxybiphenyl by dihydrodiol dehydrogenase (BphB). The second dioxygenase, 2,3-dihydroxybiphenyl dioxygenase (BphC), does not require any external reductant and cleaves the 2,3-dihydroxylated ring between carbon atoms 1 and 2 to produce 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid (HOPD, the ring meta-cleavage product), which is then hydrolyzed to benzoic acid and 2-hydroxypenta-2,4-dienoate by a hydrolase (BphD). These upper pathway enzymes in biphenyl metabolism are encoded by the bph gene clusters, in which bphA1 and bphA2 encode a large and a small subunit of the terminal dioxygenase, bphA3 encodes ferredoxin, and bphA4 encodes ferredoxin reductase (Fig. ​(Fig.1)1) (15, 20, 28, 36, 48, 79). The bphB, bphC, and bphD genes encode a dehydrogenase, a ring-cleavage dioxygenase, and a hydrolase, respectively. Among these, the large subunit of terminal dioxygenase is crucially involved in the substrate specificity of biphenyl dioxygenase (40, 42). Therefore, evolutionary molecular engineering has been applied to large-subunit genes of different origins, creating novel dioxygenases. Evolved biphenyl dioxygenases thus obtained show enhanced and expanded degradation for not only PCBs, but also other related compounds (7, 8, 40, 42, 75-77). The use of evolved enzymes is also effective for the synthesis of high-value organic molecules in the pharmaceutical industries (53, 72).

In this communication, we review recent advances in studies on the function, regulation, and engineering of bph genes, particularly focusing on the versatile characteristics of biphenyl dioxygenases.

STRUCTURAL VERSATILITIES OF BIPHENYL CATABOLIC bph GENES

Genes for catabolic functions are considered to have adaptively evolved in nature by various genetic events, including mutation, recombination, gene transfer, and assembly, resulting in a family of diverse but highly related sequences. As a consequence, the bph genes are present on bacterial chromosomes (2, 3, 11, 20, 29, 35, 39, 54, 78), plasmids (30, 65, 85), and transposons (45, 51, 59, 73). The chromosomal 90-kb element (termed the bph-sal element) containing a bph gene cluster in Pseudomonas putida KF715 can be transferred to other P. putida strains at a high frequency (59). This conjugative element is then inserted into the chromosome of a new host. Tn4371, a 55-kb transposable element, displays a modular structure including a phage-like integrase gene (int), a Pseudomonas-like bph gene cluster, and RP4- and Ti-plasmid-like transfer genes (trb) (52).

The typical bph gene cluster, composed of bphR1-bphA1A2(orf3)bphA3A4BCX0X1X2X3D, is seen in Pseudomonas pseudoalcaligenes KF707 (20, 79, 83, 84) and Burkholderia sp. strain LB400 (14, 28, 54, 68). The bph gene clusters identified to date demonstrate that some are very similar but some are very different in terms of gene organization and the structure of each gene (Fig. ​(Fig.2).2). Thus, it is obvious that certain bph gene clusters can move among soil bacteria and have evolved from a common ancestor. Some bph genes are significantly rearranged. For example, the bph genes in Pseudomonas sp. strain KKS102 are shuffled, in that the bphA4 gene is located downstream of bphD (36, 37). The bph cluster in P. putida KF715 lacks the 3.5-kb bphX region, the genes of which are involved in the lower pathway of biphenyl catabolism (26, 59). The organization of the bph operon of the gram-positive strain Rhodococcus globerulus P6 (62) is similar to that in KF707; however, the genetic uniqueness of this strain was first demonstrated by the presence of multiple bphC genes (5, 38). In Rhodococcus erythropolis TA421, three of the seven bphC genes are located on a linear plasmid (41). More detailed features of the bph genes of rhodococci were reported for Rhodococcus sp. strain RHA1 (17, 47, 48, 85). RHA1 harbors huge linear plasmids, including pRHL1 (1,100 kb), pRHL2 (450 kb), and pRHL3 (330 kb). The major bph gene cluster, consisting of bphA1A2A3A4-bphC-bphB, is located on pRHL1. The bphDEF genes are located on pRHL2 (47). In pRHL2, bphB2, bphDEF, bphC2, and bphC4 are also located in three separate regions (71). A total of seven (or possibly six) bphC-like genes are found in strain RHA1, of which four (or possibly three) are located on plasmids and three are on a chromosome (Fig. ​(Fig.2)2) (63). The 2-hydroxypenta-2,4-dienoate metabolic pathway genes (lower pathway genes) and the 2-hydroxypenta-2,4-dienoate hydratase (bphE1), 4-hydroxy-2-oxovalerate aldolase (bphF1), and acetaldehyde dehydrogenase (acylating) (bphG) genes are located on the chromosome, in contrast to most catabolic genes for the upper biphenyl pathway, which are located on linear plasmids. These bphGF1E1 genes are indicated to be indispensably responsible for biphenyl metabolism (64).

Open in a separate window

FIG. 2.

Organization of bph gene clusters in various strains. KF707-bph, P. pseudoalcaligenes KF707 bph gene cluster (20, 79, 83, 84); LB400-bph, Burkholderia sp. strain LB400 (15, 28, 54, 68); KF715-bph, P. putida KF715 (26, 59); KKS102-bph, Pseudomonas sp. strain KKS102 (36, 37, 39); Tn4371-bph, Ralstonia eutropha A5 (57); B-356-bph, C. testosteroni B-356 (10, 78); P6-bph, R. globerulus P6 (5, 6, 61); M5-bpd, Rhodococcus sp. strain M5 (44, 45, 82); RHA1-bph, Rhodococcus sp. strain RHA1 (48, 71, 85). Homologous genes with the same function are depicted in the same colors.

FUNCTIONAL VERSATILITIES OF BIPHENYL DIOXYGENASES

Aromatic ring-hydroxylating dioxygenases involved in initial oxygenation are of particular importance because this reaction destabilizes the aromatic ring and initiates the degradation of aromatic compounds. These enzymes generally consist of a terminal dioxygenase and the reductase chain (12, 49). The terminal dioxygenase activates molecular oxygen and introduces it to the substrate. Some terminal dioxygenases are homomultimers, while others are heteromultimers that comprise a large (α) and a small (β) subunit. The reductase chain transfers electrons from NADH to the terminal dioxygenase (11). The biphenyl dioxygenase of P. pseudoalcaligenes KF707 is a class IIB-type three-component enzyme consisting of four subunits, including a large subunit (BphA1) and a small subunit (BphA2) of terminal dioxygenase, a ferredoxin (BphA3), and a ferredoxin reductase (BphA4). BphA1 is an iron-sulfur protein containing the motif Cys-X-His-X17-Cys-X2-His that forms a Rieske-type [2Fe-2S] cluster involved in electron transfer from ferredoxin. BphA1 and BphA2 are associated as an α₃β₃ heterohexamer and require Fe(II) for their activities (46). Oxygen activation is supposed to occur at the mononuclear iron center of BphA1.

The biphenyl dioxygenases of P. pseudoalcaligenes KF707 and Burkholderia sp. strain LB400 have been extensively studied with respect to the degradation of PCBs. These two enzymes show distinct differences in the ranges of PCBs used as substrates. The range of PCB congeners oxidized by the LB400 enzyme is much wider than that oxidized by the KF707 enzyme (16, 22). However, KF707 biphenyl dioxygenase has a higher activity for several di-para-substituted PCBs. The purified LB400 biphenyl dioxygenase has the remarkable ability to oxidize PCB congeners that contain up to four chlorines by introducing two hydroxyl groups at either the 2,3 or 3,4 positions. The specificity of the LB400 biphenyl dioxygenase for PCBs was correlated with the relative positions of the chlorine substituents on the aromatic rings rather than with the number of chlorine substituents on the rings (4). The attack by the biphenyl dioxygenase of Burkholderia sp. strain LB400 on several symmetrical ortho-substituted biphenyls or quasi ortho-substituted biphenyl analogues was investigated. 2,2′-Difluoro-, 2,2′-dibromo-, 2,2′-dinitro-, and 2,2′-dihydroxybiphenyl were accepted as substrates. Dioxygenation of all of these compounds shows a strong preference for the semisubstituted pair of vicinal ortho and meta carbons, leading to the formation of 2′-substituted 2,3-dihydroxybiphenyls by the subsequent elimination of HX (X = F, Br, NO₂, or OH) (67). The absence of 3,4-dioxygenase activity in KF707 is another significant difference between the two dioxygenases.

The biphenyl dioxygenase from Comamonas testosteroni B-356 transforms dichlorobiphenyls in the following order of apparent specificities: 3,3′-CB > 2,2′-CB > 4,4′-CB. PCB congeners such as 2,2′-CB exact a high energetic cost, produce a cytotoxic compound (H₂O₂), and inhibit the degradation of other congeners (33). The biphenyl dioxygenase from R. globerulus P6 exhibits the following ring substitution preference for six mono- and dichlorinated PCB congeners: no substitution > meta > para > ortho substitution. This enzyme shows a strict specificity for attacking at nonhalogenated ortho or meta vicinal carbons, as in the case of KF707 biphenyl dioxygenase (50).

Despite the major difference in the PCB degradation capacities between P. pseudoalcaligenes KF707 and Burkholderia sp. strain LB400, it is particularly interesting that the biphenyl catabolic bph genes of these two strains are nearly identical in gene organization and nucleotide sequence. The identities of these components between KF707 and LB400 are as follows: BphA1 (KF707) and BphA (LB400), 95.6%; BphA2 (KF707) and BphE (LB400), 99.5%; BphA3 (KF707) and BphF (LB400), 100%; and BphA4 (KF707) and BphG (LB400), 100% (15, 79). The major discrepancy is seen in the large subunit, in which 20 amino acids (including 1 amino acid that is lacking in KF707 BphA1) are different among 460 total amino acids (15, 79). Several lines of evidence, as follows, revealed that the large subunit of terminal dioxygenase is responsible for the recognition and binding of substrates and thereby for substrate specificity (21, 27, 40), although there are reports that the small subunit is also involved in substrate recognition (13, 32). (i) A hybrid dioxygenase composed of TodC1 (F1) and BphA2A3A4 (KF707), which was constructed by the replacement of KF707 bphA1 with todC1 (encoding an iron-sulfur protein of toluene dioxygenase from P. putida F1), exhibits a substrate specificity similar to that of the original toluene dioxygenase (21, 27). (ii) A hybrid biphenyl dioxygenase composed of BphA1 (LB400) and BphA2A3A4 (KF707) exhibits a wide-ranging PCB degradation capability similar to that of the original LB400 biphenyl dioxygenase (40). Thus, only a 20-amino-acid difference in the large subunits leads to a major difference in the PCB degradation capabilities of these two biphenyl dioxygenases. The KF707 enzyme primarily recognizes the 4′-chlorinated ring structure (97%) of 2,5,4′-CB and introduces a molecular oxygen at the 2′,3′ position. On the other hand, the LB400 enzyme primarily binds (recognizes) the 2,5-dichlorinated ring structure (95%) of the same compound and introduces O₂ at the 3,4 position. Kimura et al. constructed a variety of chimeric large-subunit genes by exchanging four common restriction fragments between the KF707 bphA1 and LB400 bphA1 genes (40). Upon expression in Escherichia coli cells, various chimeric biphenyl dioxygenases revealed that a relatively small number of amino acids in the carboxy-terminal half (among 20 different amino acids in total) are involved in the recognition of the chlorinated ring and the sites of dioxygenation. Further study revealed that the site-directed mutagenesis of Thr-376 (KF707) to Asn-376 (LB400) in the KF707 biphenyl dioxygenase resulted in the expansion of the range of biodegradable PCB congeners (40). Mondello et al. investigated the large-subunit proteins in more detail (55). A comparison of large-subunit protein sequences of KF707-type and LB400-type strains identified four regions (designated I, II, III, and IV) in which specific sequences were consistently associated with either a broad or narrow PCB substrate specificity. A combination of mutations between KF707-type BphA1 and LB400-type BphA in regions III and IV resulted in dramatic differences in the substrate specificity. Altering the regions in the LB400 BphA in order to correspond to those in the KF707 bphA1 sequence produced a narrow substrate specificity that was very similar to that of KF707. A stretch of seven amino acids, termed region III, is of particular interest. Some individual mutations within region III alone improved the PCB degradative activity, especially for di-para-substituted congeners. However, the highest improvements in activity were obtained from multiple amino acid modifications in region III, suggesting that the effects of these mutations are cooperative. Barriault et al. also constructed a biphenyl dioxygenase by using common restriction sites to exchange DNA fragments between Burkholderia sp. strain LB400 bphA and C. testosteroni B-356 bphA1, showing that modifications of the C-terminal portion of the LB400 α subunit can change the catalytic properties of the enzyme (8).

Oxygenase components from C. testosteroni B-356 and Rhodococcus sp. strain RHA1 were crystallized, and X-ray diffraction was measured (33, 57). The crystal structures of a Rieske ferredoxin of Burkholderia sp. strain LB400 (14) and an NADH-dependent ferredoxin reductase of Pseudomonas sp. strain KKS102 (69) were solved. This structural information may provide more detailed insight into the substrate specificity and mode of oxygenation of various biphenyl dioxygenases.

VERSATILE REGULATION OF bph GENES

Despite detailed biochemical and genetic analyses of the bph genes of various bacteria, our knowledge concerning regulation has remained unclear for a long time. Recently, the regulatory mechanisms of the bph genes of several strains have been studied. The regulation of the bph gene cluster, bphR1-bphA1A2(orf3)bphA3A4BCX0X1X2X3D, in P. pseudoalcaligenes KF707 was recently reported in some detail by Watanabe et al. (83, 84). In this system, two regulatory genes, bphR1 and bphR2, were identified. The bphR1 gene is located just upstream of bphA1, but bphR2 is separated from the other bph genes. The BphR1 protein belongs to the GntR family and the BphR2 protein belongs to the LysR family, showing a high similarity (81%) to NahR (the naphthalene and salicylate catabolic regulator) (66). Both regulatory proteins act as activators, and at least six transcriptional start sites are mapped in this gene cluster. Thus, there are two regulatory systems as follows: (i) bphR1-dependent transcription for bphR1 itself, bphX0X1X2X3, and bphD and (ii) bphR2-dependent transcription for bphA1A2(orf3)A3A4BC. In this regulatory system, it is believed that the BphR2 protein first activates the transcription of bphA1A2(orf3)A3A4BC to convert biphenyl to the meta-cleavage compound (HOPD), which binds to BphR1 to activate this protein. The activated BphR1 protein binds to the promoter-operator regions of bphR1 itself and to bphX0, bphX1, and bphD to promote the transcription of these genes (Fig. ​(Fig.3).3). The transcription of the bph locus of Burkholderia sp. strain LB400, whose bph genes are very similar to those of KF707, was investigated (9). In this system, the ORF0 protein (corresponding to KF707 BphR1) mediates the activation of the bphA1 promoter. The four major 5′ ends were mapped between 25 and 70 bp upstream of the start codon of the bphA1 gene. Sequence elements between approximately positions 710 and 1080 upstream were required in cis for full functioning of the respective promoter(s). It should be noted that the regulatory mechanisms of the bph genes are totally different between P. pseudoalcaligenes KF707 and Burkholderia sp. strain LB400, despite the fact that the bph genes of these two strains are nearly identical. The expression of the bph genes of Pseudomonas sp. strain KKS102 is also induced by the ring meta-cleavage product (HOPD), as in the case of P. pseudoalcaligenes KF707 (61). The bph genes [bphEGF(orf4)A1A2A3BCD(orf1)A4R] of strain KKS102 constitute an operon whose expression is strongly dependent on the pE promoter located upstream of the bphE gene. A bphS gene, whose deduced amino acid sequence shows homology with the GntR family of transcriptional repressors, was identified in the upstream region of the bphE gene. Disruption of the bphS gene resulted in constitutive expression of the bph genes, suggesting that BphS negatively regulates the pE promoter. Gel retardation and DNase footprinting analyses demonstrated specific binding of BphS to the pE promoter region and identified four BphS binding sites. The binding of BphS is abolished in the presence of HOPD (60). Thus, the BphS protein acts as a repressor in strain KKS102, unlike BphR1 of strain KF707, which acts as an activator.

Open in a separate window

FIG. 3.

Proposed transcriptional regulation of bph genes in P. pseudoalcaligenes KF707. Two regulatory systems are involved in this regulation. The BphR2 protein positively regulates the bphA1A2A3A4BC genes and allow biphenyl to convert to HOPD. The BphR1 protein binds with HOPD and activates transcription of the bphR1 gene itself. The BphR1 protein is also involved in the expression of the bphX region and bphD. See the text for a detailed explanation. The data for the figure are from reference 83.

The bphEGForf4A1A2A3BCD operon in Tn4371 in Ralstonia eutropha A5 is transcribed from a σ⁷⁰ promoter, and the bphS gene product (GntR-like regulator) negatively regulates the transcription of the bph gene cluster as a repressor (56). The bph operon in gram-positive Rhodococcus sp. strain M5, bpdC1C2BADEF, is suggested to be regulated by the two-component signal transduction system of bpdS and bpdT (43). In this system, BpdS and BpdT seem to function as a sensor histidine kinase and a response regulator, respectively. Recently, Takeda et al. (80) reported the regulatory system of Rhodococcus sp. strain RHA1 in more detail. The transcription of the bphA1A2A3A4C1B operon, located on the linear plasmid pRH1, is positively regulated by a set of two-component regulatory genes (termed bphS and bphT). The bphS and bphT genes promote transcriptional induction by various aromatic compounds, such as biphenyl, benzene, and substituted benzenes. The possible induction mechanism by bphST is presented as follows. In the absence of biphenyl, bphST genes are constitutively transcribed from the adjacent bphSp promoter at the basal level. In the presence of biphenyl, biphenyl activates the bphS product (BphS), which then activates the bphT product (BphT) by phosphorylation. The activated BphT protein promotes transcription from bphA1p and induces expression of the bphA1A2A3A4C1B and bphST genes. Thus, it is interesting that GntR-like regulators appear to be common in bph clusters from gram-negative bacteria, whereas gram-positive bacteria have two-component regulatory systems to control bph expression. It is also true that the regulation of the bph genes is very versatile from strain to strain. These versatilities reflect the fact that certain bph genes are foreign genes derived from other strains and are regulated in different fashions in the new host strains.

DIRECTED EVOLUTION OF BIPENYL DIOXYGENASES

Enzymes that comprise different catabolic pathways exhibit a peculiar substrate specificity for different aromatic compounds or catabolic intermediates. Sequence analyses of aromatic ring-degradative enzymes revealed that they can be grouped into families that are similar in size and amino acid sequence (12, 23, 58). Enzymes belonging to the same family have evolved from a common ancestor to acquire a new catabolic function through various genetic events, such as gene transfer, recombination, duplication, multiple point mutation, deletion, and integration (24, 25, 31). Thus, we could learn how new degradation abilities appeared through a long historical period. Gene manipulation techniques have opened up a way to alter the function of aromatic ring dioxygenases. Thus, mutant enzymes with an enhanced degradation ability for biphenyl and its related compounds and also with a novel capability to transform the heterocyclic aromatic compounds can be generated.

DNA shuffling is a method for random recombination of selected genes in vitro by fragmentation and PCR reassembly (74). This technique was applied to the bphA1 gene of P. pseudoalcaligenes KF707 and the bphA gene of Burkholderia sp. strain LB400 (42) because the large subunits of the biphenyl dioxygenases of these two strains are crucially responsible for substrate specificity in a different manner. E. coli cells expressing shuffled (evolved) bph genes were incubated with biphenyl, 4-chlorobiphenyl (4-CB), 2,2′-dichlorobiphenyl (2,2′-CB), 4,4′-dichlorobiphenyl (4,4′-CB), 2,5,4′-trichlorobiphenyl (2,5,4′-CB), 4-methylbiphenyl (4-MB), diphenylmethane (DM), and dibenzofuran (DF). E. coli cells expressing the original KF707 BphA1 enzyme and E. coli cells expressing the original LB400 BphA enzyme exhibited major differences in the formation of the ring meta-cleavage yellow products for many biphenyl compounds. Large amounts of yellow compounds were produced from 4,4′-CB, 2,5,4′-CB, and DM by the KF707 enzyme, but not by the LB400 enzyme. In contrast, large amounts of yellow compounds were produced from 2,2′-CB and DF by the LB400 enzyme, but not by the KF707 enzyme. Thus, major differences can be seen in these two parental enzymes that are used for shuffling evolution. E. coli cells expressing some evolved BphA1 proteins exhibited interesting features in the production of ring meta-cleavage yellow compounds. One such E. coli clone carrying pSHF1045 exhibited an enhanced production of yellow compounds from biphenyl, 4-CB, 4-MB, and 4,4′-CB relative to E. coli expressing the original KF707 enzyme. The same clone produced yellow compounds from DF and 2,2′-CB but no yellow compound from 2,5,4′-CB, from which a 3,4-dihydrodiol compound is produced as a dead-end product, as did E. coli expressing the LB400 enzyme. Another E. coli clone carrying pSHF1072 gained a novel degradation activity for toluene and benzene and produced indigo from indole. The same clone exhibited a much higher activity toward monocyclic aromatic compounds such as ethylbenzene, butylbenzene, and isopropylbenzene than did E. coli expressing the KF707 enzyme (76). The deduced amino acid sequences of such evolved large subunits showed only a few amino acid changes from the original enzymes. Barriault et al. also did family shuffling of a targeted region of the large-subunit genes from Burkholderia sp. strain LB400, C. testosteroni B-356, and R. globerulus P6. Some variants showed a high activity toward 2,2′-CB, 3,3′-CB, 4,4′-CB, and 2,6-CB (7).

A method of random-priming recombination (70) is also a powerful tool for evolutionary molecular engineering of an enzyme. The bphA1 gene of P. pseudoalcaligenes KF707 was subjected to this mutagenesis. One of the resultant biphenyl dioxygenases thus obtained exhibited novel multifunctional oxygenase activities (75). This evolved enzyme attacked at the angular position adjacent to the hetero atom of heterocyclic aromatic compounds such as dibenzofuran and dibenzo-p-dioxin (angular dioxygenation). The same enzyme also introduced two atoms of molecular oxygen into the aromatic ring of dibenzofuran and dibenzo-p-dioxin (lateral dioxygenation). Furthermore, the enzyme exhibited sulfoxidation for dibenzothiophene and monooxygenation for fluorene. Based on the structural information developed from crystallographic analyses of naphthalene dioxygenase (34), Suenaga et al. constructed 12 site-directed BphA1 mutants with changes in the amino acids that coordinate the catalytic nonheme iron center (77). The Ile335Phe, Thr376Asn, and Phe377Leu biphenyl dioxygenase mutants exhibited altered regiospecificities for various PCBs compared with the wild-type biphenyl dioxygenase. In particular, the Ile335Phe mutant acquired the ability to degrade 2,5,2′,5′-CB by 3,4-dioxygenation and showed bifunctional 2,3-dioxygenase and 3,4-dioxygenase activities for 2,5,2′-CB and 2,5,4′-CB. Furthermore, two mutants, the Phe227Val and Phe377Ala mutants, introduced molecular oxygen at the 2,3 position, forming 3-chloro-2′,3′-dihydroxybiphenyl with concomitant dechlorination.

Another successful application by modified biphenyl dioxygenases is the bioconversion of a variety of heterocyclic aromatic compounds, such as flavone, flavanone, and ionized aromatics. A recombinant E. coli strain expressing pSHF1072, carrying biphenyl dioxygenase, converted 1-methoxynaphthalene, dibenzothiophene, xanthene, 1-phenylpyrazole, 2-phenylpyridine, and 4-phenylpyrimidine into their corresponding cis-dihydrodiols (53). Recombinant Streptomyces lividans expressing the same enzyme converted flavone, 6-hydroxyflavone, 7-hydroxyisoflavone, and trans-chalone to the corresponding mono- or di-hydroxylated compounds (72). The same evolved enzyme could also transform the molecular structure of a variety of aromatic compounds, including carboxylic acids or amines such as 1-naphthoic acid, 2-(1-naphthyl) acetic acid, diphenylamine, and 1-benzyl-4-piperidone. These ionized aromatics were converted to the corresponding 1,2-dihydrodiol or mono- or trihydroxy forms. According to the three-dimensional structure model constructed based on the naphthalene dioxygenase (34), diphenylamine can be well accommodated within the active site of the evolved BphA1 (Fig. ​(Fig.4).4). The hydroxylation site of diphenylamine is located adjacent to the catalytic iron in evolved BphA1, while the location of the diphenylamine in the active site of the wild-type BphA1 enzyme is far from the iron molecule. These products that were converted by evolved biphenyl dioxygenase are potentially useful as versatile starting materials for the chemical synthesis of pharmaceuticals and biologically active organic molecules.

Open in a separate window

FIG. 4.

Binding model of evolved biphenyl dioxygenase with diphenylamine. The three-dimensional structure of the biphenyl dioxygenase (pSHF1072) was constructed based on that of naphthalene dioxygenase (34). The binding model of dioxygenase with diphenylamine was constructed with MOE software (CGI Co. Ltd.). Phe-277 (depicted in magenta) in the original enzyme was changed to Tyr (in green) in the evolved enzyme along with three other amino acids. The configuration of the substrate (diphenylamine) is depicted in red in the original enzyme and in lime in the evolved enzyme. Note that the 2,3 position of diphenylamine faces the catalytic mononuclear iron center in the evolved enzyme (depicted as a large red ball).

CONCLUSIONS

Biphenyl-utilizing bacteria are ubiquitously distributed in nature. These bacteria are considered to be involved in the final stage of plant lignin degradation as well as with other aromatic degraders. Biochemical and genetic studies on PCB degradation provide us knowledge about how microorganisms acquire new and novel degradation capabilities for man-made xenobiotic compounds. Biphenyl dioxygenase is an interesting enzyme that provides a good model system for molecular evolutionary engineering. One major advantage of this technology is that only minimal prior information is required. It has been demonstrated that evolved biphenyl dioxygenases can be used for the degradation of PCBs and other environmental pollutants, including dioxins and chlorinated ethenes. Moreover, the use of evolved biphenyl dioxygenases is effective for the synthesis of high-value organic molecules, because many of the products generated by these enzymes are difficult to synthesize by existing methods of organic chemistry. It should be feasible to complement the methods of combinatorial chemistry with biotechnological methods.

Acknowledgments

REFERENCES