Development of methodology for the synthesis of 4-arylflavan-3-ol lactones

[摘要] English: The phenolic content of the heartwood of Peltophorum africanum (African wattle) and Burkea africanum (Red syringa) proved to contain a variety of compounds including the novel 4-arylflavan-3-ol lactones, 6-(3,4-dihydroxyphenyl)-6,6a,8,12b-tetrahydroisochromeno[ 3,4-c]chromene-3,8,10,11,12-pentaol and 6-(3,4,5-tri-hydroxylphenyl)- 6,6a,8,12b-tetrahydroisochromeno[3,4-c]chromene-3,8,10,11,12-pentaol respectively. Several attempts at the synthesis of these compounds to give final proof of the structure including the stereochemistry at the three chiral centres, failed due to the fact that the nucleophilicity of the pyrogallol ring of the gallic acid analogue (both protected and fee phenolic) is reduced to such an extent by the presence of the carbonyl group that it cannot effectively react with a C-4 electrophile generated on the flavan-3-ol starting material. In addition it became evident that the starting materials and products are sensitive to more drastic acid and basic reaction conditions to the extent that no desired products could be isolated from the attempted coupling of gallic acid analogues to C-4 functionalised flavan-3-ols. In order to alleviate these problems, it was envisaged to utilize a nucleophile without the carbonyl attached to it, thus changing the substrate to a pyrogallol entity containing a 1- hydroxypropyl - or allyl substituent. After coupling at C-4 of the flavan-3-ol the required carbonyl could then be introduced by consecutive water elimination and isomerization (1-hydroxylpropyl substituent) or isomerization (allyl substituent) followed by ozonolysis (with non-reductive work-up) and subsequent esterification. Since the starting materials for the synthetic strategy were not available commercially, both the flavan-3,4-diol and 1-(3',4',5'-trimethoxyphenyl)propan-1-ol had to be synthesized. While the phenylpropan-1-ol analogue was obtained in 64 % yield through utilisation of a Grignard reaction between 3,4,5-trimethoxybenzaldehyde and ethylmagnesium bromide, the required 3',4',7-trimethoxyflavan-3,4-diol became available through synthesis and subsequent manipulation of the appropriate chalcone. Thus trans-2'- ethoxymethoxy-3,4,4'-trimethoxychalcone, obtained by standard Claisen-Schmidt condensation of 2-ethoxymethoxy-4-methoxyacetophenone and 3,4- dimethoxybenzaldehyde, was epoxidized with dimethyldioxirane to give the chalcone epoxide in 98 % yield. Deprotection and cyclization to the dihydroflavonol were accomplished in 67.5 % overall yield via treatment of the chalcone epoxide with benzylmercaptan and tin(iv)chloride, followed by reaction of the subsequent-hydroxy--benzylmercaptodihydrochalcone with silver tetrafluoroborate. Altough the syn- and anti-isomers of the mercaptodihydrochalcone were observed, both of these isomers led to only the 2,3-trans-dihydroflavonol. Finally, NaBH4 reduction of the dihydroflavonol gave the flavan-3,4-diol 70 % yield. Since the starting materials for - and products from the coupling reaction are known to be acid/base sensitive, it was decided to functionalise the flavan-3,4-diol through a mercaptan leaving unit that could be activated by thiophilic Lewis acid in order to induce coupling with the pyrogallol moiety. Silver tetrafluoroborate catalysed model reactions between 4-benzylmercaptochroman and the aromatic nucleophiles, resorcinol and methylated pyrogallol, gave the 4-arylchromans in 62 and 72% yield respectively. When the nucleophile was changed to 1-(3',4',5'-trimethoxyphenyl)propan-1-ol or the dehydrated version, 1-(3',4',5'-trimethoxy-phenyl)-1-propene, however, no couple could be detected. Since no apparent change in nucleophilicity could be identified as the cause of the reaction not giving any product, the failure can probably be ascribed to steric congestion brought about by the methoxy- and allyl - or propyl groups adjacent to the required point of reaction. As it is known from literature that radicals play an important role in the in vivo synthesis of many natural products, a biomimetic approach towards the synthesis of the 4- arylflavan-3-ol lactones was considered as next alternative. It was therefore envisaged that generation of a phenolic radical at C-2 of the pyrogallol ring of a flav-3-ene 3-gallate ester moiety could result in the formation of the desired lactone. Thus tetra-O-methyl catechin was converted into the 3-keto compound by mild IBX (2-iodoxybenzoic acid) oxidation in 61 % yield. Treatment of the catechin derivative with LDA followed by quenching of the enolate with t-butyldiphenylchlorosilane led to 3-tertbutyldiphneylsilyloxy- 3',4',5,7-tetramethoxyflav-3-ene in 84 % yield; thus proving that the double bond was indeed in the right position (between C-3 and C-4 and not C-2 and C-3) for the lactone to be formed. With the position of the double bond established, the enolization reaction was repeated with 2-bromobenzoyl chloride as model electrophile and the enol ester obtained in 22 % yield. In this instance, however, the desired enol ester was accompanied by the 2,3-unsaturated isomer (43 %). With all the uncertainties round the oxidation and enol formation sorted out, the final step in this strategy for the synthesis of the target lactones can now attempted with confidence. AIBN initiated radical cyclization of 3-O-(3,4,5-trimethoxybenzoyl)-3',4',7-trimethoxyflav-3-ene and 3-O-(3,4,5-trimethoxybenzoyl)-3',4',5',7-tetramethoxyflav-3-ene will, however, receive attention in a subsequent PhD study.

[发布机构] University of the Free State