Structure and synthesis of a novel homoisoflavanone from Scilla natalensis and synthesis of selected procyanidins through the C-4 functionalization of flavan-3-ols

[摘要] English: Scilla natalensis planch (Hyacinthaceae), commonly known as Wild squill, Blue squill, Blue hyacinth, Blouberglelie, Blouslangkop, Inguduza, is one of the plants that are widely used in traditional medicines and it grows naturally over large parts of Southern Africa. While the plant is widely used in traditional medicine by indigenous African people, phytochemical investigations have revealed this plant to contain a variety of biologically active compounds that show anti-inflammatory, antibacterial, antischistosomic, anthelmintic and cytotoxicity activity. In order to determine whether its traditional use is supported by actual pharmacological effects, it was decided to re-investigate the chemical composition of Scilla natalensis. Repeatetive column - and preparative thin layer chromatography together with acetylation of the methanol extract of the bulbs of the plant led to the isolation of five known compounds, 3',4'-Di-O-acetylchavicol, 4',4''-Di-O-acetyl-3-methoxynyasol, 5,7-Diacetoxy-3-(3'-acetoxy-4'-methoxybenzyl)chroman-4- one, 4'-O-acetyl-5,7-di-O-methyl-naringenin, and 2,3,4,5,6-Penta-O-acetyl-4'-O-methyl- apigenin-7-O-β-D-glucopyranoside as well as the novel homoisoflavanone, 5,6,7-triacetoxy-3- (3',4'-dimethoxybenzyl)-chroman-4-one. While the isolated metabolites were all identified and characterised by spectroscopic means involving 1- and 2D-NMR experiments, all of the known compounds were isolated from Scilla natalensis for the first time.Since the homoisoflavonoids have been found to possess widespread physiological activity and to give final proof of the structure of the isolated novel homoisoflavanone, in particular the position of the third OH on the A-ring, the synthesis of this compound was attempted. While several synthetic routes towards homoisoflavanones have been reported in literature, it was decided to follow the dihydrochalcone approach for the synthesis of this new homoisoflavanone. In this methodology the dihydrochalcone is subjected to α-alkylation with a C-1 fragment containing another leaving group that can be displaced in the final cyclization process for formation of the heterocyclic C-ring. The desired dihydrochalcone would become available by reduction of the chalcone which can be formed by aldol condensation of the appropriate acetophenone and benzaldehyde. In this instance, however, this synthetic approach was hampered by the unavailability of the required 2,3,4,6-hydroxyacetophenone. It was therefore decided to test the synthesis on a model compound, 2-hydroxyacetophenone, and to investigate the appropriate C-1 fragment to use, before attempting the challenging synthesis of the required acetophenone.Thus standard Claisen-Schmidt aldol condensation between 2-hydroxyacetophenone and 3,4-dimethoxybenzaldehyde afforded the required chalcone (68 % yield), which was subjected to hydrogenation over 5 % Pd/C to give the dihydro equivalent in quantitative yield. To introduce the C-1 fragment it was decided to utilise a modified Baker-Venkataraman rearrangement strategy followed by reduction of the ester functionality and subsequent Mitsunobu cyclization. While ester formation between ethylchloroformate and 2'-hydroxy-3,4-dimethoxydihydrochalcone proceeded well, the rearrangement part of the reaction led to the unexpected formation of 3-(3',4'-dimethoxybenzyl)-4-hydroxycoumarin. Although this product could be transformed into the desired homoisoflavanone, it would take three more steps and it was therefore decided to evaluate a Vilsmeier- Haack type α-formylation for introducing the additional carbon atom into the dihydrochalcone moiety. While treatment of the 2'-hydroxydihydrochalcone with N,N-dimethylformamide (DMF), PCl5 and BF3etherate afforded only the 3-(3',4'-dimethoxybenzyl)-isoflavone in 40 % yield, subsequent hydrogenation over 10 % Pd/C led to the isolation of three products, i.e. 3-(3',4'-dimethoxybenzyl)chromane (43%), 3-(3',4'-dimethoxybenzyl)chromanone (the desired homoisoflavanone) (3%), and the 3,4-cis- and trans-3-(3',4'-dimethoxybenzyl)chroman-4ols (3% each). Although the desired product was obtained in only 3% yield due to over-hydrogenation, the reaction was not repeated on larger scale as it was already established that the homoisoflavanone could indeed be formed in this way.In the second part of this dissertation the issue of determining the absolute configuration at the different chiral centres of flavonoids was to be addressed. Although this has up to now been done by circular dichroism (CD) measurement, this method has led to ambiguities and is plagued with a host of empirical rules that has to be applied. It was therefore decided to investigate theapplication of vibrational circular dichroism (VCD) to the determination of the absoluteconformation of flavonoids. In order to generate a data base and eventually apply the techniqueof VCD to the stereochemistry of proanthocyanidins, it was decided that the investigation shouldbe started form a flavonoid with only one chiral centre and systematically increase the number ofstereo centres until the level of oligomeric compounds is reached. Since (+)-catechin[(2R,3S)-(+)-3,3',4',5,7-penta-hydroxyflavan, and (-)-epicatechin [(2R,3R)-(-)-3,3',4',5,7-pentahydroxy-flavan] are freely available in optical active form and can be transformed into theirrespective enantiomers, the whole synthetic endeavour was based on these compounds. In thisdissertation the aim therefore was to functionalize (+)-catechin and (-)-epicatechin in the4-position followed by the synthesis of 4-arylflavan-3-ols and ultimately proanthocyanidins B1 toB4.Thus DDQ (2,3-dichloro-5,6-dicyano-1,4-benzoquinone) oxidation of tetra-O-methyl-(+)-catechin,tetra-O-benzyl-(+)-catechin, and tetra-O-benzyl-(-)-epicatechin in the presence of ethylene glycol,gave the 4-hydroxyethoxy derivatives in 46, 60, and 50 % yields respectively. Treatment of thelatter two compounds with perbenzylated fluoroglucinol under TiCl4 catalysis led to only thecorresponding 2,4-cis-4-arylflavan-3-ols in 65 and 55 % yields. The formation ofproanthocyanidins B1, B2, B3, and B4 were successfully achieved through similar coupling ofperbenzyled catechin and - epicatechin with their respective 4-hydroxyethoxy analogues. It has,however, to be mentioned that the characterization of the perbenzylated B1 to B4 products byNMR were virtually impossible, since the spectra of these compounds were very complicatedbecause of severe duplication of signals due to restricted rotation. In order to have all possibleisomers available in free phenolic form for VCD studies, debenzylation of the synthesised4-arylflavan-3-ols and procyanidins B1 to B4 as well as the synthesis of B5 to B8 will be attendedto during the candidate's PhD studies.

[发布机构] University of the Free State