Dioxygenation of the Biphenyl Dioxygenation Product

Abstract

Two biphenyl dioxygenases (BphAs) were shown to catalyze dioxygenation of biphenyldienediol in the nonoxidized ring to form the respective symmetrical biphenyl-bis-dienediol. This novel metabolite served as a growth substrate for both BphA source strains. Its catabolism through the upper bph pathway of Burkholderia xenovorans LB400 was analyzed.

TEXT

Biphenyl dioxygenases (BphAs) catalyze the first step in the metabolism of biphenyl (BP) by aerobic bacteria. They dearomatize one of the biphenyl rings by converting it into a so-called biphenyldihydrodiol or biphenyldienediol (BP-DD), which is hydroxylated at positions 2 and 3 (for a structural formula of BP-DD and the carbon numbering used in this paper, see Fig. 2, left panel) (6). This is subsequently transformed into a catechol and further degraded by the well-characterized bph-encoded pathway (4). Several BphAs appear to possess a relaxed substrate specificity with respect to the nondioxygenated ring. Studies on polychlorobiphenyls (PCBs) in particular have shown that substituents are tolerated at several of its carbons (14). Furthermore, biphenyls with fluoro, bromo, nitro, and hydroxy substituents were accepted as substrates (15). Some BphAs even dioxygenate alkylated benzenes (16; H. Overwin and B. Hofer, unpublished results). Several lines of evidence were found indicating that dioxygenation products of BPs may also serve as substrates for a second dioxygenation, as previously described for polycyclic (hetero)arenes (2). On the grounds of gas chromatography-mass spectrometry (GC-MS) data, Haddock et al. (5) reported the likely 3,4,3′,4′-bis-dihydroxylation of a particular PCB, 2,5,2′,5′-tetrachlorobiphenyl. As also determined on the basis of GC-MS analysis, Kimura et al. (8) described the probable conversion of BP into a bis-catechol, 2,3,2′,3′-tetrahydroxybiphenyl (TeHBP), by the mutant strain Pseudomonas sp. strain KF712C1. Boyd et al. (3) reported 2′,3′-dioxygenation by Sphingomonas yanoikuyae B8/36 of the acetonide derivative of BP-DD. Shindo et al. (17) reported the transformation of BP into TeHBP by a recombinant strain, producing an artificial BphA variant and a BphB dehydrogenase. Those authors proposed a metabolic route involving the initial formation of BP-DD, followed by its dehydrogenation into 2,3-dihydroxybiphenyl (DHBP) and the subsequent dioxygenation of that intermediate at carbons 2′ and 3′, followed by a second dehydrogenation. In all those studies, however, the metabolite formed by direct dioxygenation of the aromatic ring of BP-DD, the biphenyl-bis-dienediol (BP-bis-DD), has never been described. Working with recombinant strains that efficiently produce BphAs, we reexamined the potential double dioxygenation of BP.

Fig 2.

Conversion of biphenyl into 2,3,2′,3′-tetrahydroxybiphenyl via two routes. Carbon numbering as used throughout this paper is indicated with compound 1. Compounds: 1, biphenyl; 2, biphenyldienediol [1-phenyl-cyclohexa-4,6-diene-cis-2,3-diol] (metabolite 1); 3a, 2,3-dihydroxybiphenyl; 3b, biphenyl-bis-dienediol [1-(1′-cyclohexyl-4′,6′-diene-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol] (metabolite 2); 4, 2,3-dihydroxy-biphenyl-4′,6′-diene-2′,3′-diol [1-(1′-phenyl-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol] (metabolite 3); 5, 2,3,2′,3′-tetrahydroxybiphenyl (metabolite 4). The stereochemistry at carbons 2′ and 3′ of compounds 3b and 4 is proposed; it is based on the likely assumption that, for the second dioxygenation, the two rings simply exchange their positions within the BphA active site. Enzymes: BphA, biphenyl-2,3 dioxygenase (EC 1.14.12.18); BphB, cis-2,3-dihydrobiphenyl-2,3-diol dehydrogenase (EC 1.3.1.56).

Detection, isolation, and characterization of BP-bis-DD.

An Escherichia coli strain producing the BphA of Burkholderia xenovorans LB400 (10) was used for the dioxygenation of BP. Resting cells were prepared basically as previously described (12), but with the modifications that cells were grown to an optical density at 600 nm (OD₆₀₀) of 0.9 to 1.0 before induction and that, after addition of IPTG (isopropyl-β-d-thiogalactopyranoside), cultures were further incubated at 30°C for 14 h. Cells resuspended in sodium phosphate buffer were supplemented with glucose to a concentration of 0.2% and were adjusted to 2.0 OD₆₀₀ units. A solution of BP was added to an Erlenmeyer flask to give a final nominal concentration of 62.5 μM. The solvent was evaporated, cells were added, and the transformation mixtures were incubated on a gyratory shaker at 30°C. At different time points, aliquots were withdrawn, cells were separated by centifugation, and the supernatants were analyzed by high-pressure liquid chromatography (HPLC), using an SC Lichrosphere 100 RP8 5-μm-pore-size column (4.6 by 125 mm) (Bischoff, Leonberg, Germany) connected to a diode array detector. Elution was done isocratically with 50% aqueous methanol. These experiments showed formation as well as depletion of the well-known BP-DD (metabolite 1) and the concomitant formation of metabolite 2 (Fig. 1). Its HPLC retention time indicated that the latter was substantially less hydrophobic than BP-DD, and its electronic absorption maximum showed a bathochromic shift relative to BP-DD (Table 1).

Fig 1.

Time course of the formation of the two products observed upon incubation of BP with recombinant cells producing BphA-LB400. Circles, metabolite 1; squares, metabolite 2. HPLC area units were converted to picomoles after purification and identification of the compounds (below).

Table 1.

HPLC retention times (t_r) and electronic absorption maxima (λ_max) of various metabolites

Compound	tr (min)a	λmax (nm)b	Identified as:
Metabolite 1	5.3	304	1-Phenyl-cyclohexa-4,6-diene-2,3-diol (Fig. 2; compound 2)
Metabolite 2	1.8	339	1-(1′-Cyclohexyl-4′,6′-diene-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol (Fig. 2; compound 3b)
2,3-Dihydroxy-biphenyl	10.4	247	NAc
Metabolite 3	2.8	293	1-(1′-Phenyl-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol (Fig. 2; compound 4)
Metabolite 4	3.9	247, 282	2,3,2′,3′-Tetrahydroxybiphenyl (Fig. 2; compound 5)
2,3-Dihydroxy-benzoic acid	3.1	246, 317	NA

^aRP8 column with 50% aqueous methanol (not acidified for metabolites 1 and 2) as eluent. For further details, see the text.

^bDetermined by diode array detector.

^cNA, not applicable.

We further observed that recombinant cells encoding a hybrid dioxygenase whose catalytic center is that of a BphA from Pseudomonas sp. strain B4-Magdeburg formed metabolite 2 in substantially larger amounts. As previously described (7), this enzyme was obtained by replacing a DNA segment of about 700 bp, which encodes the active site of the LB400 dioxygenase, by the respective PCR-amplified segment of Pseudomonas sp. strain B4-Magdeburg. Metabolite 2 was thus isolated from supernatants of cells producing this enzyme. After evaporation to dryness, it was extracted with an organic solvent such as acetone and was further purified by chromatography on a Phenomenex Gemini C₁₈ 10-μm-pore-size column (10 by 250 mm) (Phenomenex, Aschaffenburg, Germany) under isocratic conditions with 25% aqueous methanol.

The purified product was characterized by electrospray ionization-MS (ESI-MS) analysis, using an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific, Bremen, Germany) equipped with a nanospray ion source. A sample was dried, dissolved in methanol/water (2:1) to a concentration of about 10 pmol/μl, and transferred to gold-coated nanospray needles. Spectra were recorded in the positive ion mode at a resolution of 100,000 and compared with a simulated spectrum of BP-bis-DD, generated by Xcalibur software (Thermo Scientific). The determined mass (m/z = 245.0783) showed excellent agreement (deviation of 0.4 ppm) with the theoretical m/z value of 245.0784 for the sodium ion of a C₁₂H₁₄O₄ compound.

One-dimensional (1D) ¹H and ¹³C and 2D (COSY, HMBC) NMR spectra were recorded at 300 K on Bruker DPX-300 and DMX-600 NMR spectrometers. They clearly demonstrated the absence of aromatic protons and the highly symmetrical character of the metabolite. Signal assignments (Table 2) conclusively identified the compound as 1-(1′-cyclohexyl-4′,6′-diene-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol, the BP-bis-DD shown as compound 3b in Fig. 2.

Table 2.

Nuclear magnetic resonance data for 1-(1′-cyclohexyl-4′,6′-diene-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol (Fig. 2; compound 3b)^a

Atom (no.)	1H chemical shift (ppm)	Coupling constant (Hz)	13C chemical shift (ppm)
1/1′-C			138.1
2/2′-C	4.24	J2-3 = 5.7	67.8
		J2-4 = 1.3
3/3′-C	4.42	J3-4 = 2.4	71.8
		J3-5 = 2.5
4/4′-C	5.78	J4-5 = 9.7	133.6
5/5′-C	5.98	J5-6 = 5.8	124.3
6/6′-C	6.31		121.1
2/2′-O	3.47	J2-2OH = 5.9
3/3′-O	3.91	J3-3OH = 7.9

^aSolvent, acetone-d₆. Chemical shifts are given relative to the residual signals of the solvent (¹H, 2.05 ppm; ¹³C, 29.8 ppm).

Incubation of the purified BP-DD with BphA-producing cells ultimately confirmed its direct transformation into the bis-DD.

BP-bis-DD as metabolite.

Formation of BP-bis-DD demonstrated that dioxygenation of the second ring does not necessarily require conversion of the BP-DD into the catechol, as proposed by Shindo et al. (17). Further experiments showed, however, that the Shindo pathway is also possible. When DHBP (Sigma-Aldrich, Munich, Germany) was incubated with either of the BphA cells, HPLC analysis (RP8 column, 50% aqueous methanol, acidified with 0.085% phosphoric acid) showed disappearance of this catechol and concomitant formation of the more hydrophilic metabolite 3 (Table 1), presumed to be the dienediol of the educt. This was examined by LC-MS analysis, which was carried out using ESI-time of flight-MS (ESI-TOF-MS) (Maxis; Bruker) and an Agilent 1200 series DAD-UV HPLC system (C₁₈ Acquity UPLC BEH column [2.1 by 50 mm; 1.7 μm pore size] [Waters]; solvent A, 0.1% formic acid in water; solvent B, 0.1% formic acid in acetonitrile; gradient, 5% B for 0.5 min, increasing to 100% B in 19.5 min, maintained at 100% B for 5 min; flow rate = 0.6 ml/min; UV detection at 200 to 500 nm). The determined m/z value of 243.0619 was consistent with the sodium ion of the proposed metabolite (calculated m/z = 243.0628; deviation of 3.7 ppm). The structure was further confirmed by subsequent incubation of the supernatant of this transformation with a recombinant E. coli strain expressing gene bphB of strain LB400 (13). Under these conditions, metabolite 3 was converted into metabolite 4 (Table 1), which showed the UV spectral characteristics of TeHBP (9). The corresponding molecular mass (calculated m/z = 219.0652 for the protonated ion) was verified by LC-MS (found m/z = 219.0646; deviation of 2.7 ppm). These results prompted us to investigate if BP-bis-DD was a substrate for BphB. If conversion would occur, formation of the same metabolites as described above, namely, 1-(1′-phenyl-2′,3′-diol)-cyclohexa-4,6-diene-2,3-diol (metabolite 3) as the intermediate and TeHBP (metabolite 4) as the end product, would be expected. HPLC analysis (performed under the conditions described above for metabolite 3) clearly showed turnover of BP-bis-DD and formation of TeHBP. Only traces were observed of an intermediate with the HPLC retention time and spectral characteristics of metabolite 3. These results indicate that the bph-encoded pathway is able to convert BP into TeHBP via two routes: either by double dioxygenation, followed by double dehydrogenation, or by two subsequent cycles of single dioxygenation and single dehydrogenation (Fig. 2).

We also examined if TeHBP can be further catabolized through the upper bph pathway of strain LB400. Thus, it was separately incubated with recombinant E. coli strains producing either only the ring-cleaving dioxygenase BphC (14) or BphC and the hydrolase BphD (12), respectively. In the first experiment, the transient appearance of yellow coloration was observed, indicating formation of a short-lived ring-cleavage product. This finding is in accordance with spontaneous intramolecular cyclization of the latter via nucleophilic attack of the 2′-hydroxy group of the nonoxidized ring, as described by Kohler et al. (9). In the second experiment, a fraction of the ring-cleavage product might escape this reaction, if it is a substrate for BphD. In this case, it should be converted into 2,3-dihydroxybenzoate, in analogy to the transformation of the “normal” pathway metabolite DHBP into benzoate. Using an authentic reference (Sigma-Aldrich), the formation of 2,3-dihydroxybenzoate was indeed confirmed by HPLC (under the conditions described above for metabolite 3). A growth experiment carried out in M9 medium (11) supplemented with trace elements (1) additionally demonstrated that strain LB400 (as well as the other BphA source organism, Pseudomonas sp. strain B4-Magdeburg [7]) was able to further metabolize 2,3-dihydroxybenzoate and to utilize it as the sole carbon source.

Finally, BP-bis-DD (3 mM) was assayed as a growth substrate with both BphA source strains. When incubated in the minimal medium described above, growth was observed with both bacteria, in accordance with the described metabolism of the compound through the bph-encoded pathway. These results clearly show that the more hydrophilic degradation products originating from the double dioxygenation of biphenyl are not dead-end metabolites but can be exploited by the microbes as carbon and energy sources. It does not appear unlikely that the findings described here also analogously apply to the aerobic bacterial metabolism of other bicyclic aromatic substrates.