Structural basis for certain naturally occurring bioflavonoids to function as reducing co-substrates of cyclooxygenase I and II

Abstract

Background: Recent studies showed that some of the dietary bioflavonoids can strongly stimulate the catalytic activity of cyclooxygenase (COX) I and II in vitro and in vivo, presumably by facilitating enzyme re-activation. In this study, we sought to understand the structural basis of COX activation by these dietary compounds.

Methodology/principal findings: A combination of molecular modeling studies, biochemical analysis and site-directed mutagenesis assay was used as research tools. Three-dimensional quantitative structure-activity relationship analysis (QSAR/CoMFA) predicted that the ability of bioflavonoids to activate COX I and II depends heavily on their B-ring structure, a moiety known to be associated with strong antioxidant ability. Using the homology modeling and docking approaches, we identified the peroxidase active site of COX I and II as the binding site for bioflavonoids. Upon binding to this site, bioflavonoid can directly interact with hematin of the COX enzyme and facilitate the electron transfer from bioflavonoid to hematin. The docking results were verified by biochemical analysis, which reveals that when the cyclooxygenase activity of COXs is inhibited by covalent modification, myricetin can still stimulate the conversion of PGG(2) to PGE(2), a reaction selectively catalyzed by the peroxidase activity. Using the site-directed mutagenesis analysis, we confirmed that Q189 at the peroxidase site of COX II is essential for bioflavonoids to bind and re-activate its catalytic activity.

Conclusions/significance: These findings provide the structural basis for bioflavonoids to function as high-affinity reducing co-substrates of COXs through binding to the peroxidase active site, facilitating electron transfer and enzyme re-activation.

Figure 1. Chemical structures of the bioflavonoids used in this study.

The structure of flavone is enlarged to show the numbering of different carbon positions.

Figure 2. The 3-D QSAR/CoMFA analysis showing the correlation between the experimentally-determined COX-stimulating activity values and the predicted values.

Nine representative bioflavonoids plus flavone were used in the analysis. The experimental values were based on measuring PGE₂ production, which was determined in our previous study . Statistical parameters (q ², r ², PC, SEE, and F) for the CoMFA models of COX I and II are also listed in the figure.

Figure 3. The 3-D QSAR/CoMFA color contour maps for COX I (A) and COX II (B).

Note that quercetin was shown in the ball and stick format inside the field for demonstration. Oxygen, carbon, and hydrogen atoms are colored in red, gray, and blue, respectively. The contours of the steric maps are shown in yellow and green, and those of the electrostatic maps are shown in red and blue. Green contours indicate regions where a relatively bulkier substitution would increase the COX activity, whereas the yellow contours indicate areas where a bulkier substituent would decrease the COX activity. The red contours are regions where a negative-charged substitution likely would increase the COX activity whereas the blue contours showed areas where a negative-charged substitution would decrease the COX activity. Bioflavonoids with a higher ability to activate the COX enzymes are correlated with: (i) more bulkier substitute near green; (ii) less bulkier substitute near yellow; (iii) less negative charge near blue; and/or (iv) more negative charge near red. The figure in the lower panel shows the 90° rotation around the x-axis of the figure shown in the upper panel.

Figure 4. Identification of the binding sites for COX I and II by molecular docking approach.

A. The superimposed structures of COX I and II in complex with hematin and AA. The white labels indicate the two potential binding sites for quercetin as identified by the Active-Site-Search function in InsightII. B. Docking results for quercetin in Site-2 of COX I. C. Docking results for quercetin in Site-2 of COX II. D. Enlarged view of the interaction of quercetin with hematin and key amino acid residues in the peroxidase active sites of COX I. E. Enlarged view of the interaction of quercetin with hematin and key amino acid residues in the peroxidase active site of COX II. The protein structure was shown in the ribbon format in A, B and C. COX I was colored in pink and COX II in purple. AA was colored in light green for COX I and dark green for COX II. Quercetin was colored in light red for COX I and magenta for COX II. Carbon atoms in hematin were colored in yellow for COX I and orange for COX II whereas nitrogen atoms were colored in blue, oxygen atoms in red, and magnesium in silver. The green dashes represent the hydrogen bonds. Hematin, AA, key amino acid residues, and quercetin are shown in the ball and stick format. For amino acid residues, oxygen atoms are shown in red, carbon atoms in gray, and nitrogen atoms in blue. Hydrogens are omitted in these molecules.

Figure 5. Myricetin stimulates the catalytic activity of COX I and II (with or without aspirin pretreatment) when [¹⁴C]AA or PGG₂ is used as substrate.

The incubation mixtures consisted of 20 µM [¹⁴C]AA (0.2 µCi) or 10 µM PGG₂ as substrate, COX I or COX II as enzyme (0.5 or 0.97 µg/mL, respectively), 10 mM EDTA, 1 mM reduced glutathione, 1 µM hematin, and myricetin in 200 µL Tris-HCl buffer (100 mM, pH 7.4). The reaction was incubated at 37°C for 5 min and terminated by adding 15 µL of 0.5 N HCl to each test tube. Ethyl acetate (600 µL) was added immediately for extraction. The dried extracts were re-dissolved in acetonitrile or EIA buffer (Cayman Co. Michigan, USA), and the metabolites were analyzed using HPLC (with radioactivity detection) when [¹⁴C]AA was used as substrate or using an EIA kit when PGG₂ was used as substrate. Note that in this experiment, the COX I and II enzymes with or without aspirin pretreatment were both tested. For aspirin pretreatment, enzymes were pre-incubated with aspirin at 0.5 mM for COX I or 5 mM for COX II for 30 min at room temperature and then were immediately used as the enzyme source in the assay.

Figure 6. Schematic depiction of the catalysis and inactivation mechanism of COX enzymes and their interaction with bioflavonoids.

PPIX is for protoporphorin IX. Quercetin structure is shown as a representative bioactive bioflavonoid. Events in the peroxidase cycle are labeled with numbers to denote the sequence of occurrence.