Unexpected Desilylative-alkylation of 3-O-tert-Butyl-dimethylsilyl Galangin

Abstract

Biological activities of flavonoids have led to the creation of many therapeutic forms of plant flavonoids. Data on the chemical structures of a variety of flavonoids have been obtained, and the fundamental mechanisms of action of flavonoids as antioxidants, anti-inflammatories, cardiotonics, radioprotectors, antitumor agents and antiviral agents have been identified. However, little effort has been expended on synthetic flavonoid analogues presumably due to the difficulties in controlling the regiochemistry of the flavonoids. As a part of our ongoing efforts directed at the structureactivity relationship studies (SARs) of naturally occurring flavonoids, we have been interested in the regioselective alkylation of flavanones (naringenin, 1, Fig. 1) and flavones (apigenin, 2, Fig. 1). Herein, we report our recent attempts on the regioselective alkylation of a flavonol, galangin (3, Fig. 1). The flavonol has an additional hydroxyl group at the 3 position of the ring C (Fig. 1), which is known to be the primary site of alkylation. Thus, we envisaged that 3-Oprotected galangin would be selectively converted into the 7O-alkyl galangin under alkylating conditions, and set out to synthesize the key intermediate 4 starting from the commercially available chrysine 7 (Scheme 1). Treatment of chrysine 7 with Me2SO4 and K2CO3 in acetone provided the 5,7-di-O-methyl chrysine 8, which was subjected to the α-hydroxylation conditions to give the 5,7di-O-methyl galangin 9 in 60% yield. Protection of the 3-OH group with TBDMSCl and DMAP in anhydrous pyridine followed by Lewis acid-mediated demethylation provided the key intermediate 4, which was smoothly transformed into the alkylated product 5 upon treatment with substituted benzyl bromides (3-ClBnBr, 4-ClBnBr and 3-CNBnBr) and K2CO3 in acetone. However, under the alkylation conditions, the TBDMS protecting group was lost, and the NOESY analysis of 5 showed that there was no nOe correlation between the benzylic and A-ring protons (H6 and H8) (Fig. 2), which implied that the alkylation did not take place at the 7-O position. Instead, the benzylic protons of 5 showed strong nOe correlation with aromatic protons at the B-ring, which confirms that the alkylation site is 3-O rather than 7O. Protection of the 3-hydroxy group of 9 with TBDPSCl instead of TBDMSCl was attempted to provide more stable silyl ether but resulted in the same desilylative alkylation product 5 (data not shown). Based on this result, we presumed that the unexpected 3O-alkylated products were formed via desilylative-alkylation mechanism (Scheme 2). The flavonoid is deprotonated with K2CO3 to give an anion 11, which resonanced to the corresponding chromen-4-ol anion 12. The alkoxide ion then attacks the nearby TBDMS group to result in silyl migration (3-O to 4-O). The enolate anion at the 3-position 13, thus formed, attacks benzylic bromide to provide the 3-O-alkyl product 14, which resonances back to the stable aromatic form with concurrent loss of the TBDMS group upon aqueous work-up. In order to verify the desilylative-alkylation mechanism,