University of Groningen Functional analysis of the copper-dependent quercetin 2,3-dioxygenase. 2. X-ray absorption studies of native enzyme and anaerobic complexes with the substrates quercetin and myricetin

Quercetin 2,3-dioxygenase (2,3QD) is a mononuclear copper-dependent dioxygenase which catalyzes the cleavage of the heterocyclic ring of the flavonol quercetin (5,7,3′,4′-tetrahydroxy flavonol) to produce 2-protocatechuoyl-phloroglucinol carboxylic acid and carbon monoxide. In this study, X-ray absorption spectroscopy has been used to characterize the local structural environment of the Cu2+ center of Aspergillus japonicus 2,3QD. Analysis of the EXAFS region of native 2,3QD at functionally relevant pH (pH 6.0) indicates an active site equally well-described by either four or five ligands (3N(His) + 1-2O) at an average distance of 2.00 Å. Bond valence sum analysis confirms that the best model is somewhere between the two. When, however, 2,3QD is anaerobically complexed with its natural substrate quercetin, the copper environment undergoes a transition to a five-coordinated cage, which is also best modeled by a single shell of N/O scatterers at the average distance of 2.00 Å. This coordination is independently confirmed by the anaerobic complex with myricetin (5′-hydroxy quercetin). XANES analysis confirms that substrate binding does not reduce the Cu2+ ion. The present study gives the first direct insights into the coordination chemistry of the enzyme complexed with its substrates. It suggests that activation for O2 attack is achieved by monodentate substrate complexation to the copper ion through the 3-hydroxyl group. In addition, monodentate carboxylate ligation by the Glu73 side chain is likely to play a role in the fine-tuning of the equilibrium leading to the formation of the activated E‚S complex. Dioxygenases are ubiquitous enzymes which catalyze the transfer of both atoms of molecular oxygen into a substrate. They play an important role in the biosynthesis and catabolism of various metabolites and in several detoxification mechanisms (1). In particular, dioxygenases are essential mediators in the difficult degradation process of aromatic compounds (1, 2). As a consequence of the triplet ground state of molecular oxygen, dioxygenases often rely on the presence of a metallic cofactor to catalyze the transformation of singlet ground-state organic substrates (3). The most commonly employed metal is non-heme iron (4). However, enzymes containing other metallocenters (Cu, Mn, Mg) have also been reported (5-7). Quercetin 2,3-dioxygenase (2,3QD)1 is the only dioxygenase unambiguously known to depend on copper (7-9). The enzyme is produced in Pullularia (10), Fusarium (11), and Aspergillus (12) species when grown on rutin (quercetin 3-rhamnoglucoside) as a carbon source. In the degradative pathway of rutin the action of 2,3QD follows that of the esterase rutinase which releases quercetin from the glycoside (13). 2,3QD catalyzes the cleavage of the O-heteroaromatic ring of the flavonol quercetin (5,7,3′,4′-tetrahydroxy flavonol, QUE) to 2-protocatechuoyl-phloroglucinol carboxylic acid and carbon monoxide (Figure 1). The specificity of 2,3QD is not limited to QUE. In their study of Aspergillus flaVus 2,3QD, Oka et al. (14) discovered that the enzyme is able to catalyze the disruption of several polyphenolic flavonols. The rate at which they are degraded has been found to be influenced by the OH topology at the Aand B-rings. In general, the presence of 4′or 7-hydroxy groups increases the rate of cleavage and decreases the Michaelis constant (14). As a result, quercetin is processed about 2000 times faster than flavonol (3-hydroxy flavone). Myricetin (5,7,3′,4′,5′pentahydroxy flavonol, MYR), which possesses an additional meta 5′-OH group at the B-ring, is the substrate closest to QUE. It is processed only marginally slower (1.2 times). 2,3QD is the only non-iron dioxygenase for which a crystal structure is known. (9). The 1.6 Å resolution X-ray structure of the enzyme from Aspergillus japonicus solved at pH 5.2 shows that the enzyme is a glycosylated homodimer of about 100 kDa containing one copper ion per monomer (350 amino acids). The mononuclear copper center displays a double † The research described here was supported by The Netherlands Foundation for Chemical Research (CW) with financial aid from The Netherlands Foundation for Scientific Research (NWO). The EU supported the work at the EMBL outstation at DESY, Hamburg, through the HCMP Access to Large Installations Program. * Corresponding author. Phone (Bauke W. Dijkstra): +31-503634381. Fax: +31-50-3634800. E-mail: [email protected]. ‡ University of Groningen. § European Molecular Biology Laboratory. 1 Abbreviations: 2,3QD, quercetin 2,3-dioxygenase; E‚S, enzyme‚ flavonol complex; EXAFS, extended X-ray absorption fine structure; MYR, myricetin (5,7,3′,4′,5′-pentahydroxy flavonol); QUE, quercetin (5,7,3′,4′-tetrahydroxy flavonol); XANES, X-ray absorption near edge structure; XAS, X-ray absorption spectroscopy. 7963 Biochemistry 2002, 41, 7963-7968 10.1021/bi015974y CCC: $22.00 © 2002 American Chemical Society Published on Web 06/01/2002 coordination (Figure 2). One coordination is pseudotetrahedral with three histidine residues (His66, His68, and His112) and a water molecule (Wattd) as ligands. The other coordination is mixed trigonal bipyramidal/square pyramidal and has the same histidine residues, a solvent molecule (Wattb), and the carboxylate side chain of Glu73 as ligands. The different positions of Wattd and Wattb and the double conformation of Glu73 determine the two different coordination geometries. Glu73 coordinates the metal only in its minor conformation. In its principal conformation, the carboxylate side chain points away from the metal center. The early biochemical studies on A. flaVus 2,3QD (14) and extensive biomimetic synthetic work (15-21) have led to the proposal that the formation of a 2,3QD‚flavonol complex constitutes the first catalytic step. However, no structural data for such a complex is as yet available. Instead, its existence has even been challenged by the finding that the anaerobic addition of QUE to A. niger 2,3QD did not result in any changes in the EPR spectrum (8). To investigate this in more detail, we have undertaken an X-ray absorption spectroscopic study on A. japonicus 2,3QD in solution at functionally relevant pH (6.0). The vantage point given by the knowledge of the crystal structure of this enzyme provided a good platform for the refinement of the native EXAFS data which, in turn, resulted in a properly calibrated initial ligand set for the characterization of the 2,3QD‚QUE and 2,3QD‚MYR complexes. MATERIALS AND METHODS Sample Preparation. 2,3QD was purified as previously described (9). Protein purity was ascertained by SDS-PAGE gel electrophoresis. For XAS measurements, a glycosylated form of the enzyme was used (∼25% w/w sugar content). EXAFS samples were concentrated in Centricon concentrators (Amicon Inc., Danvers, MA) to a protein concentration of about 3-5 mM in 50 mM MES buffer, pH 6.0. Atomic absorption measurements indicated full occupation of the metal site. The 2,3QD‚flavonol complexes were prepared by anaerobic addition of the flavonols to a final concentration of 30 mM. As previously noted by Oka et al. for A. flaVus 2,3QD (14), a bathochromic shift of the ∼367 nm flavonolic band (367 nm f 380 nm in the case of quercetin) was observed by UV-vis spectroscopy upon incubation of the substrates with the enzyme. The samples were loaded into 1 mm thick polypropylene holders with Kapton foil windows and directly flash frozen in liquid nitrogen. XAS Measurements. X-ray absorption spectra were recorded at 20 K at the D2 bending magnet beam line of the EMBL Outstation Hamburg at DESY. The synchrotron was operating at 4.5 GeV with ring currents ranging from 50 to 100 mA. A Si(111) double crystal monochromator and a focusing mirror with a cutoff energy of 21.5 keV were used throughout the study. The X-ray absorption spectra of the samples were recorded in fluorescence mode by means of a 13-element Ge solid-state detector (Canberra, Meriden, CT). EXAFS spectra were measured with ∼0.4 eV steps in the edge region and 0.04 Å-1 steps in the k range from 2 to 13 Å-1. Integration times varied from 1 s in pre-edge region to 5 s at k ) 13 Å-1 for a total integration time of approximately 60 min/scan. Total exposure time per sample was typically 24 h. Eo and ∆Eo were 8983 and -6.6 eV, respectively. Data Analysis. Data reduction based on standard methods (22) was performed with the local set of EXPROG programs (23). For each sample, the scans were inspected for edge consistency, normalized by the edge jump, and averaged. Only in the case of the native sample, a slight photoreduction of the metal was observed (see Supporting Information). However, this caused no significant change in the EXAFS region as determined by the analysis of individual scans. EXAFS spectra were extracted by subtracting the slowly varying atomic background fitted by three cubic splines. The raw EXAFS spectra were converted to k space and weighted by k3 to compensate for the smaller amplitude at high k owing to the decay of the photoelectron wave. The analysis of the EXAFS spectra utilizing rapid curved multiple scattering FIGURE 1: Scheme of 2,3QD-mediated dioxygenation of the substrate quercetin (5,7,3′,4′-tetrahydroxy flavonol). Atom nomenclature is indicated in the substrate. FIGURE 2: Copper coordination in 2,3QD as derived from the crystal structure (PDB code 1JUH) (9). Two geometries are present: a distorted tetrahedral (coordination ∼70% of the total) with His66, His68, His112, and Wattd as ligands and a trigonal bipyramidal coordination with a strong square pyramidal component (τ index (32) ) 0.63) formed by the same histidine residues, Wattb and Glu73. The apical ligands of the latter are His66 and Glu73. All histidine residues coordinate through their N 2 atoms while Glu73 is bound in a syn-monodentate fashion through its O 1 atom. Averaged metal-ligand distances (average calculated over the four molecules present in the crystallographic asymmetric unit) are given in Å. This figure was generated with the program MOLSCRIPT (33) and rendered with the program RASTER3D (34). 7964 Biochemistry, Vol. 41, No. 25, 2002 Steiner et al.