Glucuronidation of Methylated Quercetin Derivatives: Chemical and Biochemical Approaches

Abstract

Botanical supplements derived from grapes are functional in animal model systems for the amelioration of neurological conditions, including cognitive impairment. Rats fed with grape extracts accumulate 3′-O-methyl-quercetin-3-O-β-d-glucuronide (3) in their brains, suggesting 3 as a potential therapeutic agent. To develop methods for the synthesis of 3 and the related 4′-O-methyl-quercetin-7-O-β-d-glucuronide (4), 3-O-methyl-quercetin-3′-O-β-d-glucuronide (5), and 4′-O-methyl-quercetin-3′-O-β-d-glucuronide (6), which are not found in the brain, we have evaluated both enzymatic semisynthesis and full chemical synthetic approaches. Biocatalysis by mammalian UDP-glucuronosyltransferases generated multiple glucuronidated products from 4′-O-methylquercetin, and is not cost-effective. Chemical synthetic methods, on the other hand, provided good results; 3, 5, and 6 were obtained in six steps at 12, 18, and 30% overall yield, respectively, while 4 was synthesized in five steps at 34% overall yield. A mechanistic study on the unexpected regioselectivity observed in the quercetin glucuronide synthetic steps is also presented.

INTRODUCTION

Quercetin (3,3′,4′,5,7-pentahydroxyflavone, 1; Figure 1), a multipurpose natural product, is one of the most abundant flavonoids regularly consumed by humans.^1-4 Although most in vitro pharmacological effects of quercetin (1) have been evaluated by testing the aglycone,^5,6 in nature quercetin occurs more commonly as glycosides (with glucose as the usual attached sugar). The deglycosylation of quercetin can occur either pre or postabsorption, and further metabolism in vivo leads to conjugation of the hydroxyl groups with glucuronic acid and/or sulfate and the formation of methoxylated derivatives.^7-12 The position of these substitutions will consequently affect the biological activity of quercetin (1) in a positive or negative manner. For example, quercetin-4′-O-β-d-glucuronide was the most potent inhibitor (K_i = 0.25 μM) of xanthine oxidase among the four potential monoglucuronides of quercetin.¹³ Quercetin-4′-O-glucoside was also shown to be more potent than quercetin-3-O-glucoside in protection of rat intestinal mucosa homogenates against iron-induced lipid peroxidation.¹⁴ Because of the importance of the position of the glucuronide moiety for biological activity, much attention has been paid to the synthesis of quercetin glucuronides. However, quercetin (1) has five potential glucuronidation sites (the hydroxyl groups, Figure 1) and to synthesize any single monoglucuronide is a challenge using either enzymatic or synthetic approaches.^15-22

Figure 1.

Chemical structures of quercetin (1), and its glucuronidated derivatives described in this article (2–6). The potential sites for glucuronidation of 1 are highlighted in red and the glucuronide moiety in blue.

Previous studies by our group have shown that oral administration of a botanical supplement mixture from grapevine is effective in defending against neuropathology and cognitive impairment in aging.^23-27 These studies identified a series of biologically available phenolic metabolites in brain tissue extracts of Sprague-Dawley rats treated with the Cabernet Sauvignon polyphenol extract, including quercetin-3-O-β-d-glucuronide (2; Figure 1) and 3′-O-methyl-quercetin-3-O-β-d-glucuronide (also named isorhamnetin-3-O-glucuronide, 3; Figure 1).²⁷ Interestingly, neither 4′-O-methyl-quercetin-7-O-β-d-glucuronide (4; Figure 1) nor the other quercetin glucuronides (5 and 6; Figure 1) were identified in the brain tissue, suggesting that these molecules are not products of phase II metabolism or that only quercetin metabolites 2 and 3 can cross the blood/brain barrier.

To the best of our knowledge, although several synthetic approaches for obtaining quercetin-3-O-β-d-glucuronide (2) have been reported,^28-32 there is only one reported chemical synthesis for 3′-O-methyl-quercetin-3-O-β-d-glucuronide (3).³⁰ No synthetic approaches to obtain the methylated quercetin glucuronide derivatives 4, 5, and 6 have been described so far in the literature, with these compounds so far only having been identified as products of mammalian metabolism in in vitro models.^33-35 The preparation of 4, 5, and 6 should be valuable because these compounds could present, as previously determined for 3, anti-Alzheimer properties and/or they could be used as additional controls in mechanism of action studies of 3 in models of Alzheimer’s disease and other neurological disorders.^36-38

Here, we describe the synthesis and complete characterization of quercetin glucuronides 3–6, exploring both enzymatic and chemical synthetic approaches. In addition, we carried out theoretical calculations to explain the regioselectivity observed during the methylation and glucuronidation reactions described in the synthetic procedures.

MATERIALS AND METHODS

General Information.

The reactions were accomplished in dried glassware and round-bottomed flasks under an argon atmosphere using commercial reagents, distilled solvents, and anhydrous solvents unless otherwise noted. High-performance liquid chromatography (HPLC)-grade solvents were purchased from Fisher Scientific. Chemicals and solvents were of reagent grade and obtained from commercial sources without further purification. The reactions were monitored by thin-layer chromatography (TLC) on aluminum-backed precoated silica gel 60 F254 plates (Sigma, St. Louis, MO), and compounds were detected using a UV lamp (254 nm). Column chromatographic purification was performed using 230–400 mesh silica gel, unless otherwise noted. Microwave reactions were carried out in a Biotage microwave reaction kit (sealed vials) in an Initiator (Biotage). The wattage was automatically determined to maintain the desired temperature for the desired period time. Preparative HPLC was performed as described in de Fatima et al.³⁸ The column was eluted with an isocratic mixture of acetonitrile and water with formic acid (0.1%) (25:75, v/v), and the flow rate was set at 9 mL/min. The crude glucuronides were loaded in 60% (v/v) MeOH. Solubility of these compounds in 60% MeOH is poor and this limited the load amount, but this solvent gave the best compromise between solubility and solvent polarity.

Characterization Data.

Melting points, NMR analysis, and LC–tandem mass spectroscopy (MS/MS) data were recorded as in de Fatima et al.³⁸

The detection of newly synthesized metabolites was achieved using a hybrid triple quadrupole/ion trap mass spectrometer QTRAP 5500 from AB Sciex. Each compound was injected individually and directly into the mass spectrometer at a flow rate of 7 μL/min using electrospray ionization. Full and product ion scan modes were utilized to assess the precursor ion mass and MS/MS spectra, respectively.

The LC separation consisted of solvent A: H₂O/0.1% formic acid and solvent B: acetonitrile/0.1% formic acid with the following gradients of solvent B: 5 min 10%, 10 min 20%, 15 min 30%, 20 min 50%, 25 min 70%, and 30 min 95% with a flow rate of 5 μL/min for 30 min.

Synthesis of 3′-, 3-, and 4′-O-Methylquercetins and Their Glucuronide Derivatives.

Synthesis of 2-(3,4-Diacetoxyphenyl)-4-oxo-4H-chromene-3,5,7-triyl triacetate (7) by Classical Heating.³⁹

Quercetin (1) (3.00 g, 9.93 mmol) and sodium acetate (1.63 g, 19.85 mmol) were dissolved in acetic anhydride (20.0 mL, 0.20 mol). The reaction mixture was stirred vigorously under reflux for 18 h using an oil bath. The resulting mixture was cooled at room temperature and dichloromethane (90 mL) was added. Then, the mixture was washed with brine (2 × 50 mL) and the organic layer was dried over anhydrous sodium sulfate. The raw product obtained after evaporation was recrystallized with methanol to obtain 4.83 g (95% yield) of the white solid (7): mp 192–194 °C. ¹H NMR (500 MHz, CDCl₃): δ (ppm) 7.72 (1H, dd, ³J = 8.5 Hz, ⁴J = 2.2 Hz, H6′), 7.69 (1H, d, ⁴J = 2.2 Hz, H2′), 7.35 (1H, d, ³J = 8.5 Hz, H5′), 7.33 (1H, d, ⁴J = 2.4 Hz, H6), 0.6.87 (1H, d, ⁴J = 2.4 Hz, H8), 2.43 (3H, s, CH₃─OAc), 2.34 (6H, s, CH₃─OAc), 2.33 (6H, s, CH₃─OAc). ¹³C NMR (125 MHz, CDCl₃): δ (ppm) 170.2, 169.4, 168.0, 168.0, 168.0, 167.9, 157.0, 154.4, 153.9, 150.6, 144.5, 142.3, 134.2, 127.9, 126.6, 124.1, 124.0, 114.9, 114.0, 109.1, 21.3, 21.2, 20.8, 20.7.

Synthesis of 2-(3,4-Diacetoxyphenyl)-4-oxo-4H-chromene-3,5,7-triyl triacetate (7) by Microwave Heating.

The method used was that of Zhou et al.³⁹ with minor modifications. We obtained 4.98 g of 7 (98% yield): mp 192–194 °C.

Synthesis of 2-(3-Acetoxy-4-(benzyloxy)phenyl)-7-(benzyloxy)-4-oxo-4H-chromene-3,5-diyl diacetate (8).⁴⁰

A mixture of compound 7 (4.00 g, 7.81 mmol), benzyl chloride (0.9 mL, 7.81 mmol), potassium carbonate (4.32 g, 31.22 mmol), and potassium iodide (0.32 g, 1.95 mmol) in dry acetone (60 mL) was stirred at 40 °C using an oil bath for 24 h. Then, another batch of benzyl chloride (0.7 mL, 6.24 mmol) was added and the temperature was increased to 45 °C. After 24 h, benzyl chloride (0.5 mL, 4.68 mmol) was added again and the temperature was increased to 50 °C. After 72 h, the solvent was removed under vacuum, and the residue was dissolved in dichloromethane (60 mL), washed with brine (40 mL) and water (40 mL × 2), and dried over anhydrous sodium sulfate and concentrated. The raw product was purified by flash column chromatography [silica gel, hexane/EtOAc/CHCl₃ (8:4:1, 7:4:1, 6:4:1)] to give 3.33 g (70%) of a light yellow solid (8): mp 153–155 °C. ¹H NMR (500 MHz, CDCl₃): δ (ppm) 7.67 (1H, dd, ³J = 8.7 Hz, ⁴J = 2.2 Hz, H6′), 7.56 (1H, d, ⁴J = 2.2 Hz, H2′), 7.43–7.39 (10H, m, Ph), 7.08 (1H, d, ³J = 8.7 Hz, H5′), 6.90 (1H, d, ⁴J = 2.4 Hz, H6), 6.71 (1H, d, ⁴J = 2.4 Hz, H8), 5.17 (2H, s, CH₂), 5.14 (2H, s, CH₂), 2.43 (3H, s, CH₃─OAc), 2.33 (3H, s, CH₃─OAc), 2.32 (3H, s, CH₃─OAc). ¹³C{¹H} NMR (125 MHz, CDCl₃): δ (ppm) 170.2 (C═O─C4), 169.7 (C═O─OAc), 168.9 (C═O─OAc), 168.1 (C═O─OAc), 162.9 (C7), 158.1 (C5), 154.0 (C2), 152.7 (C4′), 150.8 (C9), 140.2 (C1′), 136.0 (Ph), 135.3 (Ph), 133.4 (C3), 128.9 (Ph), 128.8 (Ph), 128.7 (Ph), 128.4 (Ph), 127.7 (Ph), 127.3 (C6′), 127.2 (Ph), 123.0 (C2′), 122.5 (C3′), 113.7 (C5′), 111.3 (C10), 109.3 (C8), 99.8 (C6), 70.9 (CH₂), 70.8 (CH₂), 21.3 (CH₃), 20.8 (CH₃), 20.8 (CH₃). MS m/z: [M + H]⁺ calcd for C₃₅H₂₉O₁₀, 609.2; found, 609.1/[M + Na]⁺ calcd for C₃₅H₂₈O₁₀Na, 631.2; found, 631.1.

Synthesis of 7-(Benzyloxy)-2-(4-(benzyloxy)-3-hydroxyphenyl)-5-hydroxy-3-methoxy-4H-chromen-4-one (11).⁴¹

Compound 8 (1.5 g, 2.46 mmol) was dissolved in 60 mL of dry acetone. Then, sodium hydrogen carbonate (0.83 g, 9.86 mmol), benzyl triethyl ammonium chloride (0.22 g, 0.99 mmol), and anhydrous methanol (8 mL) were added. The reaction mixture was stirred vigorously at 40 °C for 14 h. After removing the acetyl groups, H₂O (30 mL), sodium hydrogen carbonate (0.41 g, 4.93 mmol), dimethyl sulfate (0.93 mL, 9.86 mmol), and acetone (60 mL) were added and the reaction mixture was stirred at 40 °C for 8 h. The mixture was concentrated under vacuum to give a residue that was acidified with HCl (10%) to pH 3–4. The mixture was extracted with dichloromethane (3 × 150 mL). The organic phase was concentrated under vacuum, and the residue was purified by flash column chromatography [silica gel, hexane/EtOAc/CHCl₃ (7:1.5:0.5, 7:2:0.5)] to afford 0.64 g (52%) of a yellow solid (11): mp 167–169 °C. ¹H NMR (500 MHz, CDCl₃): δ (ppm) 12.63 (1H, s, OH-5), 7.71 (1H, d, ⁴J = 1.9 Hz, H2′), 7.68 (1H, dd, ³J = 8.9 Hz, ⁴J = 1.9 Hz, H6′), 7.45–7.36 (10H, m, Ph), 7.01 (1H, d, ³J = 8.9 Hz, H5′), 6.51 (1H, d, ⁴J = 2.0 Hz, H8), 6.43 (1H, d, ⁴J = 2.0 Hz, H6), 5.79 (1H, s, OH-3), 5.20 (2H, s, CH₂), 5.13 (2H, s, CH₂), 3.87 (3H, s, OCH₃). ¹³C{¹H} NMR (25 MHz, CDCl₃): δ (ppm) 178.9 (C═O─C4), 164.6 (C7), 162.1 (C5), 158.8 (C9), 155.7 (C2), 148.1 (C4′), 145.8 (C3′), 139.3 (C3), 135.9 (Ph), 135.7 (Ph), 129.0 (Ph), 128.9 (Ph), 128.8 (Ph), 128.5 (Ph), 128.0 (Ph), 127.6 (Ph), 124.0 (C1′), 121.6 (C6′), 114.8 (C2′), 111.8 (C5′), 106.4 (C10), 98.8 (C6), 93.1 (C8), 71.3 (CH₂), 70.5 (CH₂), 60.3 (OCH₃). MS m/z: [M − H]⁻ calcd for C₃₀H₂₃O₇, 495.1; found, 495.1.

Synthesis of 7-(Benzyloxy)-2-(4-(benzyloxy)-3-methoxyphenyl)-3,5-dihydroxy-4H-chromen-4-one (12).^30,42

A mixture of compound 8 (1.5 g, 2.46 mmol), potassium carbonate (1.77 g, 12.82 mmol), dimethyl sulfate (0.28 mL, 2.96 mmol), and methanol (9 mL) in acetone (25 mL) was stirred under reflux for 2 h. Solvent was removed under air and the residue was acidified with HCl (10%) to pH 3–4. The mixture was extracted with dichloromethane (3 × 150 mL). The organic phase was concentrated under vacuum and the residue was purified by flash column chromatography [silica gel, hexane/EtOAc/CHCl₃ (8:2.5:1, 7:2:1)] to afford 0.48 g (39%) of a yellow solid (12): mp 173–175 °C. ¹H NMR (400 MHz, CDCl₃): δ (ppm) 12.22 (1H, s, OH-5), 7.47–7.34 (12H, m, H6′, H2′, Ph), 6.98 (1H, d, ³J = 8.4 Hz, H5′), 6.54 (1H, s, H8), 6.45 (1H, s, H6), 5.24 (2H, s, CH₂), 5.12 (2H, s, CH₂), 3.95 (3H, s, OCH₃). ¹³C{¹H} NMR (100 MHz, CDCl₃): δ (ppm) 175.9 (C═O─C4), 164.9 (C7), 162.1 (C5), 157.1 (C9), 156.6 (C2), 151.1 (C4′), 149.5 (C3′), 136.3 (Ph), 135.7 (Ph), 131.2 (C3), 128.9 (Ph), 128.8 (Ph), 128.5 (Ph), 128.3 (Ph), 127.6 (Ph), 127.3 (Ph), 122.3 (C1′), 122.1 (C6′), 113.2 (C5′), 111.5 (C2′), 105.7 (C10), 99.1 (C6), 93.6 (C8), 70.9 (CH₂), 70.6 (CH₂), 56.3 (OCH₃). MS m/z: [M − H]^- calcd for C₃₀H₂₃O₇, 495.1; found, 495.1.

Representative Procedure for Acid Glucuronidation: Synthesis of (2R,3R,4S,5S,6S)-2-(2-(Benzyloxy)-5-(7-(benzyloxy)-5-hydroxy-3-methoxy-4-oxo-4H-chromen-2-yl)phenoxy)-6-(methoxycarbonyl)-tetrahydro-2H-pyran-3,4,5-triyl triacetate (14).

Compound 11 (0.3 g, 0.60 mmol) and (2S,3S,4S,5R,6R)-2-(methoxycarbonyl)-6-(2,2,2-trichloro-1-iminoethoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate (0.72 g, 1.51 mmol) were dissolved in anhydrous dichloromethane (20 mL) in the presence of 4 Å molecular sieves. The reaction was stirred at room temperature for 30 min under argon and then cooled in an ice bath for 15 min. A solution of BF₃-etherate (0.1 mL, 0.91 mmol) in anhydrous dichloromethane (1 mL) was added dropwise over 10 min via a syringe. The reaction mixture was allowed to warm to room temperature and stirred for 16 h. Then, another batch of (2S,3S,4S,5R,6R)-2-(methoxycarbonyl)-6-(2,2,2-trichloro-1-iminoethoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate (0.72 g, 1.51 mmol) was added and the mixture was stirred for 30 min. The reaction was cooled in an ice bath for 15 min. A solution of BF₃-etherate (0.1 mL, 0.91 mmol) in anhydrous dichloromethane (1 mL) was then added drop by drop over 10 min via a syringe. The reaction mixture was allowed to reach room temperature and stirred for 16 h. After 32 h, the reaction was quenched by the addition of saturated sodium hydrogen carbonate solution (20 mL). The organic layer was separated, washed with brine (10 mL), dried over anhydrous sodium sulfate, and evaporated. The crude product (14) was directly subjected for subsequent debenzylation and hydrolysis of acetyl and methyl ester groups.

Representative Procedure for the Removal of benzyl and Acetyl Groups, and Methyl Ester Hydrolysis, to Afford the Final Glucuronides: Synthesis of (2S,3S,4S,5R,6R)-6-(5-(5,7-Dihydroxy-3-methoxy-4-oxo-4H-chromen-2-yl)-2-hydroxyphenoxy)-3,4,5-trihydroxy tetrahydro-2H-pyran-2-carboxylic acid (5).³³

To a suspension of crude (14) in a mixture of ethanol (120 mL) and cyclohexene (30 mL) was added 20% palladium hydroxide on charcoal (0.52 g). The reaction mixture was refluxed under argon for 60 min, cooled at room temperature, and filtered through a pad of Celite to remove the catalyst. The Celite pad was rinsed with methanol (3 × 20 mL). The combined filtrate was concentrated to dryness under air and the residue was dissolved in 300 mL of methanol/water (1:1 v/v). An aqueous solution of sodium carbonate (1 M, 9 mL) was added and the mixture was stirred at room temperature under argon for 4 h. Then, the mixture was cooled in an ice bath and Dowex 50 W resin (H⁺ form) was added with stirring until the pH reached 3–4. The mixture was filtered through a pad of Celite and the resin was washed with methanol (3 × 50 mL). The combined filtrate was evaporated under nitrogen. The residue was purified by preparative HPLC in eight injections (8 × 0.5 mL). The purity of the fractions was checked by liquid chromatography–mass spectrometry (LC─MS). Fractions with purity ≥98% were combined and partially evaporated under nitrogen. The aqueous remainder was lyophilized for 72 h to furnish compound 5 as a yellow solid (139.6 mg, 47% overall yield of three last steps). ¹H NMR (500 MHz, CD₃OD): δ (ppm) 8.01 (1H, d, ⁴J = 2.0 Hz, H2′), 7.77 (1H, dd, ³J = 8.6 Hz, ⁴J = 2.0 Hz, H6′), 6.96 (1H, d, ³J = 8.6 Hz, H5′), 6.42 (1H, d, ⁴J = 2.0 Hz, H8), 6.15 (1H, d, ⁴J = 2.0 Hz, H6), 4.90 (1H, inside of CD₃OD peak, H1″); 3.96 (1H, d, ³J = 8.7 Hz, H5″); 3.78 (3H, s, OCH3); 3.61–3.52 (3H, m, H2″, H3″, H4″). ¹³C{¹H} NMR (125 MHz, CD₃OD): δ (ppm) 185.5 (COOH), 180.2 (C═O─C4), 166.3 (C7), 163.3 (C5), 158.9 (C9), 157.3 (C2), 151.9 (C4′), 146.8 (C3′), 139.9 (C3), 126.2 (C6′), 123.4 (C1′), 119.4 (C2′), 117.3 (C5′), 106.2 (C10), 104.6 (C1″), 100.2 (C6), 95.1 (C8), 77.3 (C5″), 77.1, 74.8, 73.3 (C2″, C3″, C4″), 60.8 (OCH₃). ESI-MS m/z: [M − H]⁻ calcd for C₂₂H₁₉O₁₃, 491.1; found, 491.2.

Synthesis of (2R,3R,4S,5S,6S)-2-((7-(Benzyloxy)-2-(4-(benzyloxy)-3-methoxyphenyl)-5-hydroxy-4-oxo-4H-chromen-3-yl)oxy)-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (15).³⁰

Silver oxide (0.46 g, 2.00 mmol) and (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (0.24 g, 0.60 mmol) were dissolved in dry pyridine (4 mL) in the presence of powdered 3 Å molecular sieves. The mixture was stirred, in the dark under argon at 0 °C for 30 min. After this time, a solution of compound 12 (0.20 g, 0.40 mmol) in dry pyridine was added dropwise. Stirring was continued at 0 °C for 12 h. Then, another batch of (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (0.24 g, 0.60 mmol) was added and the mixture stirred for 12 h at 0 °C. After 24 h, a mixture of aqueous solutions of potassium chloride (10% w/v, 20 mL) and acetic acid (10% v/v, 100 mL) were added and the reaction stirred for 15 min and filtered through a pad of Celite. The solid residue was washed with water (2 × 100 mL) and the crude product was eluted with acetone (3 × 50 mL). The filtrate was evaporated under vacuum (no heat) and compound 15 was afforded as a light brown solid.

Synthesis of (2S,3S,4S,5R,6R)-6-((5,7-Dihydroxy-2-(4-hydroxy-3-methoxyphenyl)-4-oxo-4H-chromen-3-yl)oxy)-3,4,5-trihydroxyte-trahydro-2H-pyran-2-carboxylic acid (3).³⁰

Compound 15 was debenzylated without further purification, the acetyl and methyl ester groups were hydrolyzed, and the crude product was purified as described for compound 5, to give compound 3 (85.3 mg, 43% overall yield of three last steps) as a yellow solid. ¹H NMR (500 MHz, CD₃OD): δ (ppm) 7.99 (1H, d, ⁴J = 1.7 Hz, H2′), 7.52 (1H, dd, ³J = 8.4 Hz, ⁴J = 1.7 Hz, H6′), 6.87 (1H, d, ³J = 8.4 Hz, H5′), 6.36 (1H, d, ⁴J = 1.6 Hz, H8), 6.17 (1H, d, ⁴J = 1.6 Hz, H6), 5.46 (1H, d, ³J = 6.8 Hz, H1″), 3.93 (3H, s, OCH₃), 3.65 (1H, d, ³J = 8.9 Hz, H5″), 3.58–3.48 (3H, m, H2″, H3″, H4″). ¹³C{¹H} NMR (125 MHz, CD₃OD): δ (ppm) 179.3 (C═O─C4) 178.0 (COOH), 165.9 (C7), 163.0 (C5), 158.8 (C2), 158.4 (C9), 150.7 (C4′), 148.4 (C3′), 135.3 (C3), 123.6 (C6′), 123.0 (C1′), 115.9 (C5′), 114.5 (C2′), 105.7 (C10), 103.7 (C1″), 99.8 (C6), 94.7 (C8), 76.2 (C5″), 77.7, 75.7, 73.3 (C2″, C3″, C4″), 56.8 (OCH₃). ESI-MS m/z: [M − H]⁻ calcd for C₂₂H₁₉O₁₃, 491.1; found, 491.2.

Synthesis of 2-(3,4-Dihydroxyphenyl)-5,7-dihydroxy-3-methoxy-4H-chromen-4-one (16).⁴¹

To a suspension of compound 11 (0.3 g, 0.60 mmol) in a mixture of ethanol (120 mL) and cyclohexene (30 mL) was added 20% palladium hydroxide on charcoal (0.52 g). The reaction mixture was refluxed under argon for 60 min, cooled at room temperature, and filtered through a pad of Celite to remove the catalyst. The Celite pad was rinsed with methanol (2 × 20 mL). The combined filtrate was concentrated to dryness under air and the residue was purified by preparative HPLC in four injections (4 × 0.5 mL). The purity of the fractions was checked by LC─MS. Fractions with purity ≥98% were combined and partially evaporated under nitrogen. The aqueous remainder was lyophilized for 72 h to furnish compound 16 as a yellow solid (177.76 mg, 93%): mp 271–273 °C. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 12.72 (1H, s, OH-5), 7.56 (1H, d, ⁴J = 2.2 Hz, H2′), 7.46 (1H, dd, ³J = 8.5 Hz, ⁴J = 2.2 Hz, H6′), 6.91 (1H, d, ³J = 8.5 Hz, H5′), 6.41 (1H, d, ⁴J = 2.1 Hz, H8), 6.19 (1H, d, ⁴J = 2.1 Hz, H6), 3.79 (3H, s, OCH₃), 3.38 (1H, s, OH). ¹³C{¹H} NMR (125 MHz, DMSO-d₆): δ (ppm) 177.9 (C═O─C4), 164.1 (C7), 161.3 (C5), 156.3 (C9), 155.6 (C2), 148.7 (C4′), 145.2 (C3′), 137.7 (C3), 120.8 (C1′), 120.6 (C6′), 115.8 (C5′), 115.4 (C2′), 104.2 (C10), 98.5 (C6), 93.6 (C8), 59.7 (OCH₃). MS m/z: [M − H]⁻ calcd for C₁₆H₁₁O₇, 315.1; found, 315.0.

Synthesis of 3,5,7-Trihydroxy-2-(4-hydroxy-3-methoxyphenyl)-4H-chromen-4-one (17).³⁹

The benzyl groups of compound 12 (0.25 g, 0.50 mmol) were removed following the same procedure of obtaining and purifying compound 16 to afford 146.5 mg of 17 (92%): mp 305–307 °C. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 12.45 (1H, s, OH-5), 7.76 (1H, d, ⁴J = 2.0 Hz, H2′), 7.69 (1H, dd, ³J = 8.5 Hz, ⁴J = 2.0 Hz, H6′), 6.94 (1H, d, ³J = 8.5 Hz, H5′), 6.47 (1H, d, ⁴J = 1.9 Hz, H8), 6.20 (1H, s, H6), 3.84 (3H, s, OCH₃). ¹³C{¹H} NMR (125 MHz, DMSO-d₆): δ (ppm) 175.9 (C═O─C4), 164.0 (C7), 160.7 (C5), 156.2 (C9), 148.8 (C2), 147.4 (C3′), 146.5 (C4′), 135.8 (C3), 122.0 (C1′), 121.7 (C6′), 115.6 (C5′), 111.7 (C2′), 102.9 (C10), 98.3 (C6), 93.7 (C8), 55.8 (OCH₃). MS m/z: [M − H]⁻ calcd for C₁₆H₁₁O₇, 315.1; found, 315.0.

Synthesis of 3,7-Bis(benzyloxy)-2-(3,4-dihydroxyphenyl)-5-hydroxy-4H-chromen-4-one (18).⁴³

A mixture of quercetin (1) (0.70 g, 2.32 mmol), borax (2.67 g, 6.95 mmol), benzyl triethyl ammonium chloride (0.08 g, 0.35 mmol), and anhydrous dimethylformamide (20 mL) was stirred at room temperature for 1 h to obtain the boron complex of quercetin. Benzyl chloride (1.1 mL, 9.26 mmol) and potassium carbonate (0.96 g, 6.95 mmol) were then added and the mixture was stirred for 24 h. A second batch of benzyl chloride (0.7 mL, 6.24 mmol) was added and the stirring continued for 24 h. Water (80 mL) was added to quench the reaction and the mixture was extracted with ethyl acetate (6 × 150 mL). The organic extracts were combined and washed with water (1 × 200 mL), dried over anhydrous magnesium sulfate, and concentrated under vacuum. The residue was purified by flash column chromatography [silica gel, hexane/EtOAc/MeOH (4:1:0.1, 3:1:0.1, 2:1:0.1)] to give 0.94 g (84%) of a yellow solid: mp 202–204 °C. ¹H NMR 500 MHz, DMSO-d₆): δ (ppm) 12.73 (1H, s, OH-5); 7.55 (1H, d, ⁴J = 2.2 Hz, H2′); 7.47–7.30 (11H, m, Ph, H6′); 6.87 (1H, d, ³J = 8.5 Hz, H5′); 6.76 (1H, d, ⁴J = 2.1 Hz, H8) 6.44 (1H, d, ⁴J = 2.1 Hz, H6); 5.22 (2H, s, CH₂); 5.01 (2H, s, CH₂). ¹³C{¹H} NMR (125 MHz, DMSO-d₆): δ (ppm) 178.2 (C═O─C4), 164.3 (C7), 161.2 (C5), 157.0 (C2), 156.4 (C9), 149.0 (C4′), 145.4 (C3′), 136.8 (C3), 136.6 (Ph), 136.3 (Ph), 128.7 (Ph), 128.6 (Ph), 128.4 (Ph), 128.3 (Ph), 128.3 (Ph), 128.0 (Ph), 121.1 (C6′), 120.9 (C1′), 115.9 (C2′), 115.7 (C5′), 105.5 (C10), 98.6 (C8), 93.2 (C6), 73.5 (CH₂), 70.2 (CH₂). MS m/z: [M − H]⁻ calcd for C₂₉H₂₁O₇, 481.1; found, 481.0.

Synthesis of 3,7-Bis(benzyloxy)-5-hydroxy-2-(3-hydroxy-4-methoxyphenyl)-4H-chromen-4-one (19).³⁹

Dimethyl sulfate (0.4 mL, 4.66 mmol), benzyl triethyl ammonium chloride (0.21 g, 0.93 mmol), sodium hydrogen carbonate (0.63 g, 7.46 mmol), and water (8 mL) were added to a solution of compound 18 (0.90 g, 1.86 mmol) in acetone (60 mL). The mixture was stirred at 45 °C for 6 h, cooled to room temperature, and water (50 mL) was added. The resulting mixture was extracted with ethyl acetate (3 × 50 mL) and the organic layers were dried over anhydrous sodium sulfate and evaporated under vacuum. The crude product was purified by column chromatograph [silica gel, hexane/EtOAc/CHCl₃ (9:2:0.5, 8:2:0.5, 7:2:0.5)] to give 0.66 g (71%) of a light yellow solid (19): mp 160–162 °C. ¹H NMR (400 MHz, CDCl₃): δ (ppm) 12.68 (1H, s, OH-5), 7.63 (1H, d, ³J = 7.8, Hz ²J = 2.4 Hz, H6′), 7.58 (1H, d, ²J = 2.4 Hz, H2′), 7.40–7.25 (10H, m, Ph), 6.85 (1H, d, ³J = 7.8 Hz, H5′), 6.47 (1H, d, ²J = 2.2 Hz, H8), 6.41 (1H, d, ²J = 2.2 Hz, H6), 5.74 (1H, s, OH-3′), 5.10 (2H, s, CH₂), 5.04 (2H, s, CH₂), 3.93 (3H, s, OCH₃). ¹³C{¹H} NMR (100 MHz, CDCl₃): δ (ppm) 178.9 (C═O─C4), 164.5 (C7), 162.1 (C5), 156.8 (C9), 156.5 (C2), 148.8 (C4′), 145.4 (C3′), 137.6 (C3), 136.5 (Ph), 135.9 (Ph), 128.9 (Ph), 128.8 (Ph), 128.4 (Ph), 128.4 (Ph), 128.3 (Ph), 127.6 (Ph), 123.7 (C1′), 122.0 (C6′), 114.7 (C2′), 110.2 (C5′), 106.3 (C10), 98.7 (C6), 93.1 (C8), 74.3 (CH₂), 70.5 (CH₂), 56.1 (OCH₃). MS m/z: [M − H]⁻ calcd for C₃₀H₂₃O₇, 495.1; found, 495.1.

Synthesis of (2R,3R,4S,5S,6S)-2-(5-(3,7-Bis(benzyloxy)-5-hydroxy-4-oxo-4H-chromen-2-yl)-2-methoxyphenoxy)-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (20).

Acid glucuronidation of compound 19 (0.20 g, 0.40 mmol) was carried out following the same procedure as for compound 14 to furnish a light orange solid (20), which was directly subjected to debenzylation.

Synthesis of (2R,3R,4S,5S,6S)-2-(2-Methoxy-5-(3,5,7-trihydroxy-4-oxo-4H-chromen-2-yl)phenoxy)-6-(methoxycarbonyl)-tetrahydro-2H-pyran-3,4,5-triyl triacetate (21).

The removal of the benzyl groups of the above quercetin moiety (20) was accomplished following the procedure for generation of compound 16 to obtain an orange solid (21), which directly underwent deacetylation.

Synthesis of (2S,3S,4S,5R,6R)-3,4,5-Trihydroxy-6-(2-methoxy-5-(3,5,7-trihydroxy-4-oxo-4H-chromen-2-yl)phenoxy)tetrahydro-2H-pyran-2-carboxylic acid (6)..^34,35

Compound 21 was dissolved in 10 mL of dry methanol, zinc acetate (0.37 g, 2.01 mmol) was added, and the solution was stirred at 75 °C for 24 h. Then, the solvent was removed under air, and the crude residue was dissolved in dimethyl sulfoxide (2.5 mL) and 35 mL of 20 mM of phosphate buffer (NaH₂PO₄/Na₂HPO₄—pH 7.2) was added. The mixture was heated in a water bath at 37 °C for 5 min, and then 140 mg of pig-liver esterase (PLE) was added. The solution was stirred at 37 °C for 8 h. Solvent was removed under nitrogen and the residue was purified by preparative HPLC in 10 injections (10 × 0.5 mL). The purity of the fractions was checked by LC─MS. Fractions with purity ≥98% were combined and partially evaporated under nitrogen. The aqueous remainder was lyophilized for 72 h to obtain compound 6 as a yellow solid (90 mg, 45% overall yield of four last steps). ¹H NMR (600 MHz, DMSO-d₆): δ (ppm) 7.94/7.98 (1H, dd, ³J = 8.7/8.7 Hz, ⁴J = 1.9/2.1 Hz, H6′), 7.81/7.91 (1H, d, ⁴J = 1.9/2.1 Hz, H2′), 7.18/7.18 (1H, d, ³J = 8.7/8.7 Hz, H5′), 6.52/6.52 (1H, d, ⁴J = 2.0/2.0 Hz, H8), 6.20/6.20 (1H, d, ⁴J = 2.0/2.0 Hz, H6), 5.13/5.16 (1H, d, ³J = 7.3/7.7 Hz, H1″), 3.86/3.87 (3H, s, OCH₃), 3.83/3.78 (1H, ds, ³J = 9.5/10.2 Hz, H5″); 3.33/3.31, 3.34/3.39, 3.40/3.43 (3H, m, H2″, H3″, H4″). ¹³C{¹H} NMR 150 MHz, DMSO-d₆): δ (ppm) 176.0/176.0 (C═O─C4), 170.3/167.6 (COOH), 164.1/164.0 (C7), 160.6/160.6 (C5), 156.2/156.2 (C9), 150.7/150.7 (C4′), 145.9/146.0 (C2), 145.7/146.2 (C3′), 136.3/136.3 (C3), 123.3/123.3 (C6′), 123.2/123.3 (C1′), 114.4/114.6 (C2′), 112.3/112.2 (C5′), 103.1/103.1 (C10), 100.1/100.6 (C1″), 98.2/98.3 (C6), 93.8/93.8 (C8), 76.1/75.8 (C3″), 75.4/75.6 (C5″), 72.9/73.1 (C2″), 71.5/71.1 (C4″), 55.8/55.8 (OCH₃). ESI-MS m/z: [M − H]⁻ calcd for C₂₂H₁₉O₁₃, 491.1; found, 491.2.

Synthesis of 3,5,7-Trihydroxy-2-(3-hydroxy-4-methoxyphenyl)-4H-chromen-4-one (22).³⁹

The benzyl groups of compound 19 (0.35 g, 0.70 mmol) were removed following the same procedure for obtaining and purifying compound 16 to afford 198.5 mg of compound 22 (89%): mp 253–255 °C. ¹H NMR (500 MHz, DMSO-d₆): δ (ppm) 12.45 (1H, s, OH-5); 7.68 (1H, d, ⁴J = 2.2 Hz, H2′), 7.64 (1H, dd, ³J = 8.5 Hz, ⁴J = 2.2 Hz, H6′), 7.06 (1H, d, ³J = 8.5 Hz, H5′), 6.42 (1H, d, ⁴J = 2.0 Hz, H8), 6.19 (1H, d, ⁴J = 2.0 Hz, H6), 3.84 (3H, s, OCH₃). ¹³C{¹H} NMR (125 MHz, DMSO-d₆): δ (ppm) 176.0 (C═O–C4), 164.3 (C7), 160.8 (C5), 156.3 (C9), 149.3 (C4′), 146.2 (C2), 146.2 (C3′), 136.2 (C3), 123.5 (C1′), 119.8 (C6′), 114.6 (C2′), 111.8 (C5′), 103.0 (C10), 98.3 (C6), 93.5 (C8), 55.6 (OCH₃). MS m/z: [M – H]⁻ calcd for C₁₆H₁₁O₇, 315.1; found, 315.0.

Synthesis of (2R,3R,4S,5S,6S)-2-((3,5-Dihydroxy-2-(3-hydroxy-4-methoxyphenyl)-4-oxo-4H-chromen-7-yl)oxy)-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (23).

Silver oxide (0.27 g, 1.18 mmol) and (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate (0.24 g, 0.59 mmol) were dissolved in dry pyridine (5 mL) in the presence of powdered 3 Å molecular sieves. The mixture was stirred, in the dark, under argon at 0 °C for 30 min. After this time, a solution of compound 22 (150.0 mg, 0.47 mmol) in dry pyridine was added dropwise. Stirring was continued at 0 °C for 18 h. After this time, a mixture of aqueous solutions of potassium chloride (10% w/v, 15 mL) and acetic acid (10% v/v, 75 mL) was added and the reaction mixture was stirred for 15 min and filtered through a pad of Celite. The solid residue was washed with water (2 × 90 mL) and the crude product was eluted with acetone (3 × 60 mL). The filtrate was evaporated under vacuum (no heat) and compound 23 was furnished as a light-yellow solid.

Synthesis of (2S,3S,4S,5R,6R)-6-((3,5-Dihydroxy-2-(3-hydroxy-4-methoxyphenyl)-4-oxo-4H-chromen-7-yl)oxy)-3,4,5-trihydroxytetrahydro-2H-pyran-2-carboxylic acid (4).^34,35

Compound 23 was not further purified. The acetyl groups were removed and the product was purified by preparative HPLC as described for compound 21 to yield compound 4 (142.4 mg, 61% overall yield of the two last steps) as a yellow solid. ¹H NMR (600 MHz, DMSO-d₆): δ (ppm) 7.71 (1H, d, ⁴J = 2.0 Hz, H2′), 7.69 (1H, dd, ³J = 8.7 Hz, ⁴J = 2.0 Hz, H6′), 7.09 (1H, d, ³J = 8.7 Hz, H5′), 6.76 (1H, d, ⁴J = 1.8 Hz, H8), 6.41 (1H, d, ⁴J = 1.8 Hz, H6), 5.03 (1H, d, ³J = 7.6 Hz, H1″), 3.84 (3H, s, OCH₃), 3.51 (1H, d, ³J = 9.9 Hz, H5″), 3.27 (1H, dd, ³J = 8.9; 8.7 Hz, H3”), 3.22 (1H, dd, ³J = 8.6; 7.7 Hz, H2″), 3.12 (1H, dd, ³J = 8.9; 9.7 Hz, H4″). ¹³C{¹H} NMR (150 MHz, DMSO-d₆): δ (ppm) 176.2 (C═O–C4), 171.3 (COOH), 162.9 (C7), 160.2 (C5), 155.7 (C9), 149.4 (C4′), 147.0 (C2), 146.2 (C3′), 136.6 (C3), 123.3 (C1′), 119.8 (C6′), 114.6 (C2′), 111.8 (C5′), 104.6 (C10), 99.7 (C1″), 98.9 (C6), 94.2 (C8), 76.6 (C3″), 73.5 (C5″), 73.0, (C2″),72.0 (C4″), 55.6 (OCH₃). ESI-MS m/z: [M – H]⁻ calcd for C₂₂H₁₉O₁₃, 491.2; found, 491.0.

Enzymatic Approaches to Quercetin Glucuronide Synthesis.

Enzymatic Assays.

Commercially available mammalian uridine diphosphate glucuronosyltransferase (UGT) enzymes (protein stock concentration 5 mg/mL) prepared from insect cell lines that had been infected with baculovirus containing the cDNAs of human UGTs (UGT1A-1, 3, 4, 6, 7, 8, 9, 10 and UGT2B-4, 15, and 17) and UDP-glucuronic acid (UDP-GlcA) were purchased from Corning Life Sciences NY.

Enzymatic assays were performed according to an established method.⁴⁴ Briefly, to 100 μL of the reaction mixture containing a final concentration of 0.25 mM acceptor substrate (compounds 22, 17, and 1), 1 mM UDP-GlcA, 50 mM Tris-HCl buffer (pH 7.5), were added 100 μg of UGT enzyme and the mixture was incubated at 37 °C for 1 h. The reaction was quenched by adding an equal volume of MeOH, centrifuged at 10,000g for 10 min at 4 °C, and the products were analyzed by C18 reverse-phase HPLC.

Enzymatic Synthesis of (2S,3S,4S,5R,6R)-6-((3,5-Dihydroxy-2-(3-hydroxy-4-methoxyphenyl)-4-oxo-4H-chromen-7-yl)oxy)-3,4,5-trihydroxytetrahydro-2H-pyran-2-carboxylic acid (4).

Mammalian UGTs 1A3, 1A7, and 1A9 were used to catalyze the glucuronidation of compound 22. The enzyme assay described above was scaled up as follows. The reactions were performed in 10 2 mL Eppendorf tubes, each containing 40 μL of 25 mM UDP-GlcA, 100 μL of 250 mM Tris-HCl (pH 7.5), 40 μL of 20 mM of 22 solution, and 120 μL of 5 mg/mL UGT enzyme. The reaction volume was made up to 480 μL with deionized water, and the reaction mixtures incubated at room temperature. After 6 h of incubation, an additional 40 μL of 25 mM UDP-GlcA, 40 μL of 250 mM Tris-HCl (pH 7.5), and 120 μL of 5mg/ml UGT enzyme were added. Finally, after a further 6 h of incubation, an additional 40 μL of 25 mM UDP-GlcA and 120 μL of 5 mg/mL UGT enzyme were added, and the reaction mixture was incubated for a further 12 h before quenching by the addition of 400 μL of methanol. The 10 reaction tubes were combined, the solvent was removed under nitrogen gas, and the aqueous fraction desalted and concentrated using Water Sep-Pak Plus C18 cartridges as described previously.³⁹ HPLC was then used to separate and collect compound 6, monitoring at 350 and 254 nm. The fraction obtained was dried under nitrogen gas, verified by HPLC and LC─MS in comparison with chemically synthesized authentic standard, and stored at −80 °C.

HPLC Analysis for Enzymatic Reactions.

Reverse-phase HPLC (Agilent HP1200 HPLC with a diode array detector) was used to analyze the enzymatic reaction products using a C18 Hypersil Gold column (250 × 4.6 mm dimensions and particle size 5 μm). The solvent used was 0.1% formic acid (solvent A) and 100% acetonitrile (solvent B) with solvent gradient 95% solvent A, 5% solvent B for 5 min; 85% solvent A, 15% solvent B for 5 min; 77% solvent A, 23% solvent B for 15 min; 67% solvent A, 33% solvent B for 5 min; 60% solvent A, 40% solvent B for 5 min; 0% solvent A, 100% solvent B for 5 min; and 95% solvent A, 5% solvent B for 5 min. Glucuronidated products were monitored at 350 nm and 254 nm.

Computational Methods.

Theoretical calculations were performed with General Atomic and Molecular Electronic Structure System (GAMESS)⁴⁵ and with Gaussian⁴⁶ on a supercomputing system. Theoretical calculations for the mechanistic study were performed at the PM3 level,^47-49 with solvation effects being incorporated through the polarizable continuum model (PCM).^47-49 Geometry optimization, to generate low-energy conformations, was achieved via a two-step approach. The large initial set of conformations generated from the Spartan’18 Parallel Suite was screened at the GAIO-M062X/6-311+G(d,p)-sp//PBE/Midi level of theory.^50-54 This was followed by a second optimization with the GAIO-M062X/6-311+G(d,p)-sp//B3LYP/6-31+G-(d,p)^50-54 step of conformers within 5.0 Kcal/mol of the lowest energy conformer from the previous step—this also included the keyword, int = ultrafine, as well as incorporation of the SMD (solvation model based on density) solvent model.^50-54

RESULTS AND DISCUSSION

Enzymatic Glucuronidation of 4′-O-Methylquercetin.

To test the potential for enzymatic glucuronidation of a quercetin derivative, we selected 4′-O-methylquercetin (22, Figure 2), a non-natural substrate for mammalian phase II metabolism. We first screened all available recombinant mammalian UGTs (UGT1A-1, 3, 4, 6, 7, 8, 9, and 10; UGT1A9 and 1A7; and UGT2B-4, 15, and 17) in in vitro assays comparing 4′-O-methylquercetin (22), 3-O-methylquercetin (16—Figure 3), and nonsubstituted quercetin (1) according to Blount.⁴⁴ After separation of the products from the substrate by HPLC, peak areas of all glucuronidated products were summed and overall activity was expressed relative to the activity of the most active enzyme (UGT1A9 for all three substrates) (Figure 4). The enzymes with highest activities with 22 were in order, UGT1A9, UGTs 1A1, 1A7, and 1A10, whereas UGTs 1A4, 1A6, and 2B4 had relatively little to no activity (Figure 4). The HPLC retention times were also compared with those of authentic standards from which we determined that the major products generated by mammalian UGTs from 22 were the 3′-O- and 7-O-glucuronides. Diglucuronides were also detected with UGT1A7 and UGT1A10 (SI1, SI2, and Table 1). No enzyme generated the 3′-O-glucuronide. In the case of UGT1A7, the diglucuronide was the major product.

Figure 2.

Synthesis of 4′-O-methylquercetin-7-O-β-d-glucuronide (4) and 4′-O-methylquercetin-3′-O β-d-glucuronide (6). (a) BnCl, Borax, 48 h (18: 84%); (b) NaHCO₃, BnN(Et)₃Cl, MeOH:H₂O (1:4, v/v), acetone, Me₂SO₄, 50 °C*, 8 h (19: 71%); (c) (2S,3S,4S,5R,6R)-2-(methoxycarbonyl)-6-(2,2,2-trichloro-1-iminoethoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate, BF₃·Et₂O, DCM, 4 Å molecular sieves, 0 °C-to r.t, 48 h; (d) Pd(OH)₂/C (20%), cyclohexene/EtOH (1:4, v/v), reflux* 1 h (22: 89%); (e) Zn(OAc)₂ MeOH, reflux* 1 h; (f) DMSO, 20 mM of NaH₂PO4/Na₂HPO4—pH 7.2, PLE, 37 °C**, 8 h (6:45%, 3 steeps from 19); (g) (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate, Ag₂O, Py, 3 Å molecular sieves, 0 °C, 18 h; and (e,f) (4: 61%, 3 steps from 22). *An oil bath was used as a heat source. **A water bath was used as a heat source.

Figure 3.

Synthesis of 3′-O-methylquercetin-3-O-β-d-glucuronide (3) and 3-O-methylquercetin-3′-O-β-d-glucuronide (5). Reagents and conditions: (a) i: Ac₂O, NaOAc, reflux* under argon, 24 h (7: 95%) or ii: Ac₂O, NaAc, MW, 15 min (7: 98%); (b) BnCl, K₂CO₃, KI, acetone, 35–40 °C*, 72 h (8: 70%; 9: 25%; 10: <5%); (c) NaHCO₃ BnN(Et)₃Cl, MeOH, H₂O, acetone, Me₂SO₄, 40 °C*, 22 h (11: 52%); (d) Me₂SO₄, K₂CO₃, MeOH, acetone reflux*, 2 h (12: 39%: 13: 8%); (e) (2S,3S,4S,5R,6R)-2-(methoxycarbonyl)-6-(2,2,2-trichloro-1-iminoethoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate, BF₃.Et₂O, DCM, 4Å molecular sieves, r.t, 32 h; (f) Pd(OH)₂/C (20%), cyclohexene/EtOH (1:4, v/v), reflux* 1 h; (g) Na₂CO₃ (1.0 M), MeOH/H₂O (1:1, v/v) (5: 47%, 3 steps from 11); and (h) (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate, Ag₂O, Py, 3 Å molecular sieves, 0 °C, 24 h (3: 43%, three steps from 12). *An oil bath was used as a heat source.

Figure 4.

Relative activity of mammalian UGTs for glucuronidation of quercetin (blue), 3-O-methylquercetin (orange), and 4′-O-methylquercetin (grey). The relative activities for the acceptor substrates are shown with respect to UGT1A9 (100%), the most active enzyme toward all three substrates. Activity of UGT1A9 for quercetin (1) was 2.93 nmol/min/mg protein, for 3-O-methylquercetin (16) was 4.05 nmol/min/mg protein, and for 4′-O-methylquercetin (22) was 0.827 nmol/min/mg protein.

Table 1.

Glucuronidated Products Detected from Activity of Mammalian UGTs with 4′-O-Methyl-quercetin (22)^a

	glucuronide products
UGT enzyme	4′-O-Me-Q-3′-O-GlcA	4′-O-Me-Q-7-O-GlcA	4′-O-MeQ;di-O-GlcA
UGT1A1	+++	+++	−
UGT1A3	+++	+++	−
UGT1A4	+	+	−
UGT1A6	+	+	−
UGT1A7	+++	+++	+++
UGT1A8	+++	+++	−
UGT1A9	+++	+++	−
UGT1A10	+++	+++	+++
UGT2B4	+	+	−
UGT2B15	+++	+++	−

^aKey: (+++), glucuronide detected as a major product; (+) trace of glucuronide detected; and (−) no glucuronide detected. See SI1 and SI2 for HPLC and LC–MS analysis.

For enzymatic synthesis of a reference sample of 6, the assay procedure described in the Materials and Methods section was used, with scale-up to a final volume of 870 μL in each reaction vial, and incubation for a longer time period (24 h) with stepwise addition of the sugar donor UDP-GlcA and the UGT enzyme at 6 h intervals. The fraction obtained was dried under nitrogen gas, verified by HPLC and LC─MS in comparison with the chemically synthesized authentic standard (SI1 and SI2), and then stored at –80 °C. The final yield was 1.1 mg from 2.53 mg of 22 and 19.4 mg of UDP-glucuronic acid, obtained using approximately 18 mg of mammalian UGT enzyme, at a reagent cost of $1740 for the enzyme and sugar donor alone. This cost is prohibitive for use in large-scale experiments, and we, therefore, reassessed chemical synthetic approaches to the methyl-quercetin glucuronides.

Chemical Synthesis of 3′-O-Methylquercetin-3-O-β-d-glucuronide (3) and 3-O-Methylquercetin 3′-O-β-d-glucuronide (5).

Our first synthetic approach for obtaining 3′-O-methylquercetin 3-O-β-d-glucuronide (3) and 3-O-methylquercetin-3′-O-β-d-glucuronide (5) started with the complete acetylation of quercetin (1) following the methodology developed by He and coworkers.³⁹ The peracetylated quercetin (7) was obtained with 95% yield under conventional heating; however, 24 h under reflux was necessary to fully consume the starting material. The reaction time was reduced dramatically by using microwave irradiation (MW) as a source of energy for 15 min, with the same yield (Figure 3, step a). The benzylation of 7 was achieved by the direct substitution of the acetyl group by benzyl groups after treating 7 with 2.0 equivalents of benzyl chloride under basic conditions in acetone (Figure 3, step b).⁴⁰ Unfortunately, under the reaction conditions described by Cheng,⁴⁰ the major product was the monobenzylated derivative 9 (61%). We were able to improve the yield of 8 by using one more equivalent of benzyl chloride and by adding this reagent in batch (1.0, 0.8, and 0.7 equivalents with 24 h between additions), which gave 8 and 9 with 70 and 25% yield, respectively. A small amount of derivative 10 was also observed (less than 5%). Compound 8 was characterized in detail using 1D- and 2D-NMR. The long-range correlation between the methylene protons of the benzyl groups and C-4′ and C7 can be seen in the heteronuclear multiple bond correlation (HMBC) spectrum (SI9).

The selective monomethylation of 8 was carried out using two different approaches. First, the use of NaHCO₃ (2 equiv) as the base and dimethyl sulfate (4 equiv) as the methyl source³⁹ furnished compound 11 with 52% yield (Figure 3, step c). This procedure includes two steps; the removal of acetyl groups, which was done by monitoring with TLC and LC─MS (SI3) followed by the methylation reaction over 8 h. Under these conditions, the preferred position of methylation is OH-3 (see theoretical modeling below). However, Zhou³⁹ reported obtaining compound 12 using the same method.

Our second approach was done by using K₂CO₃ (5.2 equiv) as the base and dimethyl sulfate (1.2 equiv) as the methyl group source. The reaction was accomplished in only 2 h in a single step⁴² to achieve compound 12 with 39% yield. A small amount of dimethylated derivative (13, 8%) was observed (Figure 3, step d). Under these conditions, the preferred position of methylation is OH-3′ (see theoretical modeling below).

Compounds 11 and 12 were characterized by 1D- and 2D-NMR. Long-range correlation between the methoxy protons and C-3 of 11 (SI13) and C-3′ of 12 (SI17), respectively, is evident in the HMBC spectra.

Glucuronidation at position 3′ was carried out under acid conditions considering the charge over the oxygen at this position. (2S,3S,4S,5R,6R)-2-(methoxycarbonyl)-6-(2,2,2-trichloro-1-iminoethoxy)tetrahydro-2H-pyran-3,4,5-triyl triacetate was used as the glucuronyl donor and BF₃·Et₂O as Lewis acid (Figure 3, step e). Compound 14 was not purified to proceed to the next steps. The two last steps for the synthesis of 3-O-methylquercetin-3′-O-β-d-glucuronide (5) involved the hydrogenolysis of the benzyl groups and the hydrolysis of the methyl ester and acetyl groups (Figure 3, steps f and g). The glucuronide 5 was obtained with 47% yield (three steps; Figure 3) after purification by preparative HPLC.

Compound 5 was characterized using 1D- and 2D-NMR. In the HMBC spectrum, the long correlation between methoxy protons and C-3 and between the anomeric proton (H-1″) and C-3′ (SI21) is evident. Table 3 provides the NMR spectroscopic data for the glucuronide products.

Table 3.

NMR Spectroscopic Data for 3′-O-Methylquercetin-3-O-β-d-glucuronide (3), 3-O-Methylquercetin-3′-O-β-d-glucuronide (5), 4′-O-Methylquercetin-7-O-β-d-glucuronide (4), and 4′-O-Methylquercetin-3′-O-β-d-glucuronide (6)^a

position	compound 3		compound 5		compound 4*		compound 6**
	1H δ (ppm)CD3OD J (Hz)	13C{1H} δ (ppm) CD3OD	1H δ (ppm)CD3OD J (Hz)	13C{1H} δ(ppm) CD3OD	1H δ (ppm)DMSO-d6 J (Hz)	13C{1H} δ (ppm) DMSO-d6	1H δ (ppm) DMSO-d6 J (Hz)	13C{1H} δ (ppm) DMSO-d6
2		158.8		157.3		147.0		145.9/146.0
3		135.3		139.9		136.6		136.3/136.3
4 (C═O)		179.3		180.2		176.2		176.0/176.0
COOH		178.0		180.5		171.3		170.3/167.6
5		163.0		163.3		160.2		160.6/160.6
6	6.17 (d) 1H, J = 1.6	99.8	6.15 (d) 1H, J = 2.0	100.2	6.41 (d) 1H, J = 1.8	98.9	6.20/6.20 (d) 1H, J = 2.0/2.0	98.2/98.3
7		165.9		166.3		162.9		164.1/164.0
8	6.36 (d) 1H, J = 1.6	94.7	6.42 (d) 1H, J = 2.0	95.1	6.76 (d) 1H, J = 1.8	94.2	6.52/6.52 (d) 1H, J = 2.0/2.0	93.8/93.8
9		158.4		158.6		155.7		156.2/156.2
10		105.7		106.2		104.6		103.1/103.1
1′		123.0		123.4		123.3		123.2/123.3
2′	7.99 (d) 1H, J = 1.7	114.5	8.01 (d) 1H, J = 2.0	119.4	7.71 (d) 1H, J = 2.0	114.6	7.81/7.91 (d) 1H, J = 1.9/2.1	114.4/114.6
3′		148.4		146.8		146.2		145.7/146.2
4′		150.7		151.9		149.4		150.7/150.7
5′	6.87 (d) 1H, J = 8.4	115.9	6.96 (d) 1H, J = 8.6	117.7	7.09 (d) 1H, J = 8.7	111.8	7.18/7.18 (d) 1H, J = 8.7/8.7	112.3/112.2
6′	7.52 (dd) 1H, J = 8.4, 1 1.7	123.6	7.77 (dd) 1H, J = 8.6, 2.0	126.2	7.69 (dd) 1H, J = 2.0, 8.7	119.8	7.94//7.98 (dd) 1H, J = 1.9/2.1, 8.7/8.7	123.3/123.3
1″	5.46 (d) 1H, J = 6.8	103.6	4.90	104.6	5.03 (d) 1H, J = 7.6	99.7	5.13/5.16 (d) 1H, J = 7.3/7.7	100.1/100.6
2″, 3″, 4″	3.58–3.48 m (3H)	77.7, 75.7, 73.3	3.61–3.52 m (3H)	77.1, 74.8, 73.3	3.22 (dd), 1H* J = 8.6, 7.7; 3.27 (dd), 1H* J = 8.9, 8.7; 3.12 (dd), 1H* J = 8.9, 9.7	73.0, 76.6, 72.0	3.33/3.31, 3.34/3.39, 3.40/3.43**	72.9/73.1, 76.1/75.8, 71.5/71.1
5″	3.65 (d) 1H, J = 8.9	76.2	3.96 (d) 1H, J = 8.7	77.3	3.51 (d) 1H*, J = 9.9	73.5	3.83/3.78 (d) 1H, J = 9.5/10.2	75.4/75.6
OCH3	3.93 (s) 3H	56.8	3.78 (s) 3H	60.8	3.84 (s) 3H	55.6	3.86/3.87 (s) 3H	55.8/55.8

^a*The presence of residual water peak impacted peak integration of the glucuronic acid.

^**The presence of the residual water peak also impacted peak integration of the glucuronic acid. In addition, hydrogen bonding between the ─COOH of glucuronic acid and C3─OH resulted in the presence of two stable conformations for this quercetin glucuronide, which was evident in the ¹H NMR spectrum, hence NMR shifts and coupling constants are reported in pairs (see SI for details).

Many authors have reported that the glucuronidation of the hydroxyl group can be successfully accomplished under Koenigs–Knorr reaction conditions.^30,55 Thus, our first attempt for obtaining the glucuronide 15 was to use the conditions reported by Needs.³⁰ He described the preparation of 15 with 51% yield by treating 12 with silver oxide (Ag₂O) and (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)-tetrahydro-2H-pyran-3,4,5-triyl triacetate in pyridine at 0 °C. However, when we attempted these conditions, we obtained 15 in low yields (8%) (Table 2, entry 1). As we progressed by adding more equivalents of the glucoronosyl donor and silver oxide, the yield improved (Table 2, entries 2–4). Next, we applied an intermittent injection method with a higher concentration of the glucuronosyl donor (3.0 equiv). After 24 h of the total reaction time, 15 was obtained at a 53% yield (Table 2, entry 5).

Table 2.

Different Koenigs–Knorr Reaction Conditions for Preparation of 4′,7-Di-O-benzyl 3′-O-methylquercetin-3-O-methyl-2,3,4-tri-O-acetyl-β-d-glucuronate (15)

entry	sugar donor	Ag2O	time (h)	yield (%) of 15c,d
1a	1.25	2.5	18	8
2	1.5	3.0	18	19
3	2.0	4.0	18	27
4	2.5	5.0	24	32
5	3.0 (1.5 + 1.5)b	5.0	24	53

^aReaction conditions from Needs.³⁰

^bReagents were added in batch (periods of time of 12 h).

^cYield was calculated by HPLC without purification.

^dUnder these reaction conditions, only the β-anomer was formed.

The linkage formed between glucuronic acid and the aglycone may be either axial or equatorial (α or β with respect to the sugar moiety). In enzymatic glucuronidation, the β-isomer is generally formed. The stereochemical outcome of chemical glycosidation can be determined by the nature of the protecting group at position 2 of the sugar donor. Possession of an acetyl protecting group at this position allows for the formation of an acetoxonium ion intermediate that facilitates the formation of the β-glucuronide by blocking attack on the bottom face of the ring.^16-21

To remove the benzyl and acetyl groups and hydrolyze the methyl ester, we followed the procedure described above, and 3′-O-methylquercetin-3-O-β-d-glucuronide (3) was obtained with 43% yield (3 steps; Figure 3).

Compound 3 was characterized in detail using 1D- and 2D-NMR. In the HMBC spectrum, the long correlation between methoxy protons and C-3′ and between the anomeric proton (H-1″) and C-3 (SI25) is also evident. Table 3 provides the NMR spectroscopic data for the glucuronide products.

Chemical Synthesis of 4′-O-Methyl quercetin 3′-O-β-d-glucuronide (6) and 4′-O-Methyl quercetin 7-O-β-d-glucuronide (4).

Novel 4′-O-methylquercetin 7-O-β-d-glucuronide (4) and 4′-O-methylquercetin 3′-O-β-d-glucuronide (6) were successfully synthesized in six and five steps from quercetin (1), respectively (Figure 2).

The preparation of 18 was accomplished with 84% yield from the selective benzylation of 1 using borax as the chelating agent to protect positions 3′ and 4′ and benzyl chloride and K₂CO₃ in the presence of benzyltriethylammonium chloride as a phase-transfer catalyst (Figure 2, step a).³⁹ In order to improve the yield, benzyl chloride was added in two batches at intervals of 24 h each.

Next, the selective methylation of 18 furnished compound 19 with 71% yield (Figure 2, step b). In this case, the hydroxyl group at position 3′ remains unresponsive to methylation because of the differences in charge over oxygen atoms in the 3′ and 4′ positions.

Compounds 18 and 19 were characterized using 1D- and 2D-NMR. The long correlation between the methylene protons of benzyl groups and C-3 and C-7 carbons for compound 18 (SI37) and between methoxy protons and C-4′ for compound 19 (SI41) is seen in the HMBC spectra.

The glucuronidation of 19 was performed under acidic conditions using the same procedure described to synthesize compound 14 (Figure 2, step c). Without further purification, compound 20 was subjected to the next reaction to remove benzyl groups using Pd(OH)₂ over charcoal and a mixture of ethanol and cyclohexene (Figure 2, step d). However, when we tried to remove the acetyl groups and methyl ester from the glucuronosyl moiety, compound 6 was not obtained. Jurasekova²² reported the study of the chemical modifications undergone by flavonoids, especially by quercetin, under alkaline conditions by UV–visible absorption, Raman, and surface-enhanced Raman scattering spectroscopy. It was demonstrated that there exists a clear structure-instability correlation concerning mainly to the C3─OH group in the C-ring to pH ~ 12. The opening of the flavanone into its corresponding chalcone occurs in alkaline solution. Thus, it was necessary to use Zn(OAc)₂ under reflux to eliminate the acetyl groups followed by enzymatic hydrolysis using PLE to hydrolyze the methyl ester (Figure 2, steps e, f).

Compound 6 was obtained with 45% overall yield in the four last steps after purification by preparative HPLC. It was characterized using 1D- and 2D-NMR. The long correlation between methoxy protons and C-4′ and between the anomeric proton (H-1″) and C-3′ (SI47) is observed in the HMBC spectrum. Table 3 provides the NMR spectroscopic data for the glucuronide products.

The NMR spectra obtained for 6 in DMSO-d₆ gave more signals than expected, warranting more detailed characterization. A pair of signals was observed in the NMR spectra for each nuclei in 6 with a peak intensity ratio of approximately 1:0.67 in the ¹H NMR spectrum (Figure 5). Full chemical shift assignment of the pair of signals, using a suite of 2D NMR experiments (SI46 and SI47), showed that the pair of signals both corresponded to 6. A second ¹H NMR spectrum acquired at 50 °C resulted in a change in the relative peak integrals to approximately 1:0.85 (SI43). This, therefore, indicated some conformational equilibrium exhibited by 6. It is worth noting that this conformational equilibrium was not observed for the methanol solutions of 3 and 4 as well as the DMSO-d₆ solution of 5 presented in this work.

Figure 5.

¹H NMR spectrum of compound 6. The inset shows expanded region with H6 resonances from the pair of conformational isomers, present in a ratio of approximately 1:0.67. The large water peak indicated precluded integration of some glucuronic acid proton signals.

We then investigated the probable reason for this conformational equilibrium. Density functional theory (DFT) calculations of compound 6 at GIAO-M06-2X/6-311+G(d,p)-sp⁵⁶//B3LYP/6-31+G(d,p),^57,58 following exhaustive conformational search in Spartan’18 Parallel Suite (Spartan’18 v. 1.4.4 (2019) Wavefunction, Inc., Irvine, CA), resulted in 23 conformers >1% Boltzmann population. The six lowest energy conformations and their respective Boltzmann populations are shown in Figure 6. Three conformations from these six lowest energy conformers (Figure 6a-c) exhibited hydrogen bonding between the glucuronic acid residue, via ─COOH, and the B ring of the quercetin subunit, via C3─OH. Typical hydrogen bond length is between 2.6 – 3.1 Å and can be observed up to ~3.5 Å. A fourth conformation with a ─COOH/C3─OH internuclear distance of 3.69 Å may have a very weak hydrogen bonding interaction, whereas the remaining two conformations do not show this hydrogen bonding (respective internuclear distances far greater than 3.5 Å, Figure 6d-f). The presence of hydrogen bond interactions potentially confers restricted rotation to the C2–C1′ bond, which in turn allows for slow rotational conformation averaging that can now be observed on the NMR time scale, as is the case with 6. In contrast, DFT modeling of 4 and 5 did not reveal any instance of this hydrogen bonding interaction (SI61, Tables SI1 and SI2).

Figure 6.

DFT-derived (GIAO-M06-2X/6-311+G(d,p)-sp//B3LYP/6-31+G(d,p)) lowest energy conformations of 6, showing their respective Boltzmann population distributions and the internuclear distances between glucuronic acid ─COOH and the C3─OH of the B ring of the quercetin moiety. The ─COOH/C3─OH internuclear distances are 2.56, 2.59, 2.68, 3.69, 6.80, and 7.87 Å, respectively, for (a–f). Potential hydrogen bonding can be formed between ─COOH and C3─OH at internuclear distances <~03.5 Å, resulting in possible restricted rotation around C2–C1′.

In order to obtain 4′-O-methylquercetin-7-O-β-d-glucuronide (4), it was necessary first to perform the debenzylation of 19 to generate 4′-O-methylquercetin (22), with 89% yield (Figure 2, step d). Next, the direct glucuronidation of 22 was performed under Koenigs–Knorr reaction conditions. By adding 1.25 equivalents of (2R,3R,4S,5S,6S)-2-bromo-6-(methoxycarbonyl)tetrahydro-2H-pyran-3,4,5-triyl triacetate and 2.5 equivalents of Ag₂O, the glucuronide 4 was obtained in 61% yield (3 steps) after the complete hydrolysis of the acetyl groups and methyl ester of the sugar moiety in a regioselective manner (Figure 2, steps g, e, and f).

In addition, we observed the formation of another regioisomer, 4′-O-methylquercetin-3-O-glucuronide, as noted from the LC─MS chromatogram of the reaction (Figure 7); however, it was not further characterized. We think that the other most susceptible position to link the glucuronic moiety is position 3 because the position 3′ needs acidic conditions to be glucuronidated.³⁰ The ratio of both regioisomers on the chromatogram is 1:0.05. To the best of our knowledge, regioselective glucuronidation of 4′-O-methylquercetin at the 7-position has not been previously reported.

Figure 7.

LC chromatogram and MS spectra in the negative mode for the direct basic glucuronidation of 22 (Figure 2, steps g, e, and f) to obtain 4 as a major product.

Compound 4 was fully characterized by 1D- and 2D-NMR. From the HMBC spectrum, it is possible to observe long correlation between methoxy protons and C-4′ and between the anomeric proton (H-1″) and C-7 (SI56). Table 3 provides the NMR spectroscopic data for the glucuronide products.

Mechanistic Theoretical Modeling.

The purpose of this subsection is to provide qualitative insight to some mechanistic details behind the studied reactions. It is by no means a detailed computational account on thermodynamic and kinetic modeling, which would require accurate energies computed with DFT coupled with a moderate-sized basis set. Given the size of the molecular systems studied in this research, such calculations would be extremely time consuming. This is far beyond the scope of the present work, where we will be using the computational data as a guide to improve the understanding of the underlying mechanistic pathways. To achieve this goal in a cost-effective fashion, we decided to perform the calculations at the semiempirical PM3 level,^59-61 with water solvation effects being incorporated through the PCM.^59-61 Semiempirical methods are widely known to be valuable in calculations of large molecular systems and have been successfully employed in a variety of situations.⁶² These calculations were performed with the GAMESS⁶³ using default parameters.

Figure 8 shows the synthetic steps to obtain 11 from 8. In the initial steps, the acetyl groups of 8 are first hydrolyzed (TLC and LC─MS data-SI3). The possibility of the subsequent loss of a proton from positions 3 and 3′ provides two active sites (negatively charged oxygen atoms) for the intermediate. To investigate the energy barrier for the loss of a proton from both sites, the saddle points corresponding to the hydrogen abstraction by the OH anion were calculated. This was followed by intrinsic reaction coordinate (IRC) calculations in the reactants’ direction. A subsequent minimization of the reactants gave a zero-point energy (ZPE) barrier height of less than 0.1 kcal mol⁻¹ for both positions. This result can be interpreted as both positions (3 and 3′) will readily lose a hydrogen atom. We next considered the reaction between the intermediate formed (but now with no hydrogen atoms at positions 3 and 3′) and Me₂SO₄, calculating the barrier heights involved in this process. The aforementioned steps of saddle-point optimizations, IRC calculations in the reactants’ direction, and the subsequent reagent minimization gave a barrier height for reaction at position 3 of ~43 kcal mol⁻¹, which is ~2 kcal mol⁻¹ higher than the same reaction at position 3′. The implication of this result is that the reaction at 3 may be somewhat slower than the reaction at 3′, allowing the product to be mostly methylated at 3′. However, equilibrium constant calculations for the reaction between the intermediate formed and dimethyl sulfate (Figure 8) provided a possible explanation for why the major product is as described in Figure 8. Reaction at position 3 gave an equilibrium constant that is 60 times larger than the reaction through position 3′. Thus, when equilibrium is attained, the products associated with the reaction through position 3 will exist in much larger quantity than the products associated with the reaction through position 3′, thus allowing the former to be the major product (Table SI5).

Figure 8.

Selective methylation of 4′,7-di-O-benzylquercetin triacetate (8) to obtain 4′,7-di-O-benzyl-3-O-methylquercetin (11).

Under different conditions (Figure 9), 12, and a lesser amount of 13, were obtained from 8. In this case, the hydrolysis of the acetyl groups and the methylation reaction occur at the same time; thus, the most important question was to know why the acetyl groups leave compound 8 from position 3′ before leaving from positions 3 and 5. Likewise, we, therefore, calculated the barrier heights corresponding to the hydrolysis of the acetyl groups. Here, we first looked at the barrier heights for the reaction represented by step 2 (Figure 10), which were lower than 0.6 kcal mol⁻¹ for the three positions. This means that at 55 °C, these three reactions, in practice, will occur very rapidly. However, in the first step of the process (Figure 10), it was observed that the barrier height for the reaction through position 3′ (~7 kcal mol⁻¹) is much lower than the barrier heights for positions 3 and 5 (12 and 19 kcal mol⁻¹, respectively). At this level of theory, this can, therefore, be considered the rate-determining step, indicating that position 3′ could be the first to lose the acetyl group (Figure 11). Therefore, the methylation is predicted to occur at this position (Table SI6).

Figure 9.

Methylation of 4′,7-di-O-benzylquercetin triacetate (8) to obtain 4′,7-di-O-benzyl-3′-O-methylquercetin (12) and 4′,7-di-O-benzyl-3,3′-di-O-methylquercetin (13) with 39% and 8% yield, respectively.

Figure 10.

Steps for hydrolysis of acetyl groups.

Figure 11.

IRC paths for the OH attack on three different acetyl groups (PM3/PCM level).

Finally, for the regioselective glucuronidation of 22, we wanted to explain why this reaction occurs through position 7 before occurring through positions 3, 3,′ and 5 (Figure 12).

Figure 12.

Regiospecific glucuronidation of compound 22.

The barrier heights for the S_N2 attack of the OH groups at the four available positions of 22 on the acetoxonium cation (Figure 2) were calculated. The respective saddle-point, IRC calculations, and minimizations led to a barrier height [excluding ZPE effects] for position 7 of 1.2 kcal mol⁻¹, more than 1 kcal mol⁻¹ lower than for the other three positions. Inclusion of ZPE effects makes all barrier heights below zero kcal mol⁻¹, position 7 having the lowest energy and, therefore, correspondingly being predicted as the fastest reaction (Figure 13, Table SI7) at the PM3/PCM level.

Figure 13.

IRC paths for the OH attack on four different OH groups (PM3/PCM level).

In summary, although enzymatic approaches offer the conceptually simplest route for regioselective glycosylation of flavonoids, these are currently of limited value for the synthesis of 4′-O-methyl quercetin 3-O-β-d-glucuronide because of the low product yields as well as the high cost of the cosubstrate UDP-glucuronic acid. Cell-based approaches using engineered E. coli to generate this cosubstrate and also harbor the necessary UGT enzyme have been reported,^52,53 but are problematic with membrane-bound mammalian enzymes such as UGT1A9. Plant UGTs are soluble enzymes that work better in bacteria. However, although plant genomes often contain several hundred different UGT genes,^54,55 few of their products have been shown to catalyze glucuronidation, let alone with the necessary regiospecificity of the acceptor. Structure-based protein design may in the future yield plant-derived enzyme catalysts for the reactions studied here.

Chemical synthetic approaches to obtain 3′-O-methylquercetin-3-O-β-d-glucuronide (3) and provide routes to the synthesis of 3-O-methylquercetin-3′-O-β-d-glucuronide (5), 4′-O-methylquercetin-7-O-β-d-glucuronide (4) and 4′-O-methylquercetin-3′-O-β-d-glucuronide (6) with acceptable yields of 12, 18, 34, and 30% respectively, have been developed and discussed. We believe that the synthesis of the latter four compounds has not been described previously. A number of analytical tools were applied for complete characterization of the final products as well as the intermediates in the synthetic routes, including MS and NMR. In addition, theoretical calculations supported the synthetic pathways that yielded the final products.