A new approach to the asymmetric synthesis of pentadeca-1,3,5,7,9,11,13,15-octols and their derivatives is presented. It is based on the Sharpless asymmetric dihydroxylation (AD) of 4,4'-methylene[(1R,1'S,6R,6'S)-6-acetoxycyclohept-3-en-1-yl]bis(4-methoxybenzoate) (9), derived from a double [3+4] cycloaddition of the 1,1,3-trichloro-2-oxyallyl cation with 2,2'-methylenedifuran (1). The diol (-)-10, obtained in 98.4% ee from 9 with "AD-mix-beta(5x), was oxidised into (2R and 2S,4S,6R)-tetrahydro-2-hydroxy-6-((4S,6S)-(6-hydroxy-4-[(4-methoxybenzoyl)oxy]cyclohept-1-en-1-yl)-2-oxopropyl)-2H-pyran-4-yl 4-methoxybenzoates ((-)-18). By the combinations of Evans' anti and Nasaraka's syn reductions of aldol (-)-18 with the double Mitsunobu reaction, 16 diastereomeric pentadeca-1,3,5,7,9,11,13,15-octols and analogues can be obtained, in principle, with high enantio- and diastereoselectivities.

A great variety of natural products of biological interest includes polyketides (1,3-polyoxo, 1,3-polyols, aldols). 1 Several approaches for their synthesis have been proposed. 2,3 Inspired by the work of Lautens 4 and Hoffmann and co-workers, 5 who have converted 8-oxabicyclo[3.2.1]oct-6-en-3-one into 7-carbon-1,3-polyols and analogues, 6 and by that of Kaku et al.,7 who have transformed cyclohept-3-ene-1,6-diol into 1,3-polyols, we have proposed a new, noniterative asymmetric synthesis of longchain 1,3-polyols starting from the now readily available 2,2′-methylenedifuran (1). 8 The method imployed the double [3+4]-cycloaddition of the 1,1,3-trichloro-2-oxyallyl cation to 1. After reductive workup, a 45:55 mixture of meso-2 and (()-threo-2 was obtained in 55% yield and separated by fractional crystallizations. The meso compound was converted into meso-3, which was desymmetrized into diol (-)-4 by means of the Sharpless asymmetric dihydroxylation. 9 Further transformations imploying the combinations of Evans' anti 10 and Nasaraka's syn 11 aldol reductions with Mitsunobu reaction 12 allow to prepare, in principle, 16 diastereomeric pentadecane-1, 3,5,7,9,11,13,15-octols (e.g., (-)-5) and analogues (Scheme 1). If the syn relationship between the 4-methoxybenzoates at C-3 and C-13 (atom numbering of (-)-5) could be changed into an anti relative configuration, all possible stereomeric pentadecane-1,3,5,7,9,11,13,15-octols could be reached in both enantiomeric forms. We report here a solution to that problem.As already described, 8 diol (-)-4 was converted (Scheme 2) into tetrol (-)-6 in 75% overall yield. Treatment of (-)-6 with (MeO) 2 CMe 2 in acetone under acidic conditions (pyridinium paratoluenesulfonate) gave acetonide (-)-7 in 92% yield. Heating (-)-7 in acetonitrile in the presence of 12 equiv of DBU (1,8-diazabicyclo[5.4.0]undec-7-ene) induced the isomerization of (-)-7 into the primary paramethoxybenzoate (-)-8. Equilibrium ([(-)-8]/[(-)-7] ) 3.0) was reached after 4 h at 80 °C, and (-)-8 could be isolated in 70% yield (Scheme 2). This result can be interpreted in terms of acyl group migration from the C-3 to the C-1 position (atom numbering of (-)-5; IUPAC numbering: 2′′′,4′′′, see (-)-8), the primary ester (-)-8 being more stable than (-)-7 for steric reasons. Also for steric reasons, the intramolecular migration of the paramethoxybenzoyl group from the 1-oxy position to the 6-hydroxy group of the cycloheptenol moiety is forbidden. Attempts to displace the acyclic secondary alcohol at C-3 (atom numbering of (-)-5; IUPAC: 2′′′) using the Mitsunobu reaction ((EtOOC) 2 N 2 , PPh 3 , 4-NO 2 C 6 H 4 COOH) 12 were not met with success. They led to complex mixtures. Selective sulfonylation of the acyclic alcohol at C-3 (IUPAC: 2′′′) with CH 3 SO 2 Cl and p-TsCl (pyridine, Et 3 N) also led to intractable mixtures.Selective methanolysis of the acyclic paramethoxybenzoate at C-3 (IUPAC: 6′) was possible on treating (-)-7 in MeOH containing 7.2 equiv of Mg(OMe) 2 (40 °C, 8 h). This furnished triol (-)-9 in 68%...

The reactivity of various 2-oxyallyl cations toward 2,2'-methylenedifuran (1b), 2,2'-(hydroxymethyl)difuran (1c), 2,2'-(trimethylsilylmethylene)difuran (1d), and di(2-furyl)methanone (1e) has been explored. Difuryl derivatives 1c, 1d, and 1e refused to undergo formal double [4+3]-cycloadditions. Conditions have been found to convert 1b into meso-1,1'-methylenedi[(1R,1'S,5S,5'R)- (3) and (+/-)-1,1'-methylenedi[(1RS,1'SR,5SR,5'RS)-8-oxabicyclo[3.2.1]oct-6-en-3-one] (4) that do not require CF(3)CH(OH)CF(3) as solvent. High yields of meso-1,1'-methylenedi[(1R,1'S,2S,2'R,4R,4'S,5S,5'R)- (5) and (+/-)-1,1'-methylenedi[(1RS,1'RS,2SR,2'SR,4RS,4'RS,5SR,5'SR)-2,4-dimethyl-8-oxabicyclo[3.2.1]oct-6-en-3-one] (6) have been obtained when 1b was reacted with 2,4-dibromopentan-3-one (7h) and NaI/Cu.

Starting from inexpensive furan and furfuryl alcohol, a noniterative approach to the synthesis of pentadeca-1,3,5,7,9,11,13,15-octols and their derivatives has been developed. The method relies upon the double [4+3]-cycloaddition of 1,1,3-trichloro-2-oxylallyl cation with 2,2'-methylenedifuran and conversion of the adducts into meso and (±)-threo-1,1'-methylenebis (cis-and trans-4,6-dihydroxycyclohept-1-ene) derivatives. The latter undergo oxidative cleavage of their alkene moieties, generating 5-hydroxy-7-oxoaldehydes that are reduced diastereoselectively into either syn or anti-5,7-diols. Asymmetry is realized using either chiral desymmetrization with Sharpless asymmetric dihydroxylation or by kinetic resolution of polyols using lipase-catalyzed acetylations. All of the possible stereomeric pentadeca-1,3,5,7,9,11,13,15-octols and derivatives can be obtained with high stereoselectivity applying simple operations, thus demonstrating the high stereodiversity of this new, noniterative approach to the asymmetric synthesis of long-chain polyketides.

No abstract