Synthesis of Optically Pure 1,4-Dioxanes from Alkenes Promoted by Organoselenium Reagents

M. Tiecco, L. Testaferri, F. Marini, S. Sternativo, C. Santi, L. Bagnoli and A. Temperini

Dipartimento di Chimica e Tecnologia del Farmaco, Sezione di Chimica Organica, Università di Perugia, Italy. E-mail: tiecco@unipg.it

Received: 12 June 2001 / Uploaded: 7 August 2001

Introduction

An alkene can be regio and stereospecifically converted into a selenoether, a selenoalcohol or a selenoamide through an anti addition reaction mediated by an electrophilic organoselenium reagent in the presence of the appropriate nucleophile. [1] These molecules can then be further functionalized by taking advantage of the versatile chemistry of the selenium compounds. Thus, for example, the selenium moiety can be easily displaced either directly by reductive deselenenylation (c) or, after oxidation to the corresponding selenoxide or selenone by elimination (a) or by substitution (b), respectively. Several other conversions of the organoselenium compounds are also possible. [1]

The alkoxyselenenylation reaction of alkenes occurs easily also in the cases in which diols are used as nucleophiles. [2] We now report that the alkoxyselenenylation of alkenes using optically pure diols can be employed as the starting and key step to effect an efficient and stereocontrolled synthesis of several isomeric tetrasubstituted 1,4-dioxanes in an enantiomerically pure form. Very few methods are reported in the literature for the synthesis of these heterocycles in optically active form. [3] Moreover, these dioxanes can be cleaved to afford optically active diols and this represents a further useful application of the presently described procedure.

Results and discussion

The results reported in the following scheme and in the table refer to the reactions carried out with methyl styrylacetate 1 and the (R,R)-diols: (2R,3R)-2,3-butandiol 2a, (1R,2R)-1,2-diphenyl-1,2-ethanediol 2b and (1R,2R)-1,2-cyclohexandiol 2c.

The synthetic procedure involves the following three steps:

= The regio and stereospecific anti addition of diol 2a-c on the alkene 1 promoted by N-(phenylseleno)phthalimide which generates the alkoxyselenides 3a-c and 7a-c. The two enantiomerically pure diastereoisomers are separated by column chromatography and then submitted to the next step.

= The elimination reaction of the selenoxides obtained by oxidation with hydrogen peroxide to afford the Michael acceptors 4a-c and 8a-c.

=The 6-exo-trig cyclization reactions promoted by NaH. The attack of the nucleophilic oxygen on both the diastereotopic faces should generate the two enantiomerically pure diastereoisomers 5, 6 from 4 and 9, 10 from 8, respectively.

As a matter of fact not all the possible isomers were formed probably because the high steric requirements of some of these dioxanes discourage their formation. In fact, as indicated in the following table, which summarizes the results obtained, compound 8a gives rise to the single isomer 9a, whereas 8b and 8c do not give any cyclization product.
 

Starting from the commercially available (S,S)-diols the enantiomers of the dioxanes described above can be obtained. This was demonstrated in the case of the (2S,3S)-2,3-butandiol by the optical rotations of the intermediates and of the final dioxanes as well as by the proton nmr spectra recorded in the presence of chiral shift reagents.

Interesting results were obtained when the reaction was effected starting from the nitrile 11 in the presence of (1R,2R)-1,2-diphenyl-1,2-ethanediol 2b. In this case in fact, as indicated in the following scheme, all the four possible dioxanes were separated.
 
 

For all the 1,4-dioxanes obtained the stereochemical assigments, which are indicated in the following table, could be easily effected on the basis of proton and carbon NMR spectra and of some NOESY experiments.

The four ring protons could be unambiguously assigned on the basis of their chemical shifts, multiplicity and coupling constants. In compounds 5a-d, 6a-d, 9d and 10d the values of the vicinal coupling constants JAB and JCD clearly suggest [3e][4] that the molecular geometries are those indicated in the table. Further support came from the results of the NOESY experiments which demonstrated that a strong dipolar interaction occurs between the protons reported in the third column.

Structural attribution of compound 9a was not so straightforward because of the observed values of the two vicinal coupling constants. However, the NOESY spectrum indicated that strong dipolar interactions exist between MeA and HC and between MeB and HD. This suggests that both the methyl groups occupy an axial position. Furthermore, in analogy with what it is observed in the case of the phenyl substituted cyclohexanes, [5] it seems reasonable to assume that the phenyl group occupies an equatorial position. In order to explain the observed values of the vicinal coupling constants it can be suggested that the molecule do not assume a perfect chair conformation but it must be partially distorted.

It is interesting to note that compounds 9b,c and 10a-c, which in both the possible chair conformations must have the phenyl groups or the fused ring in axial positions, were not formed.
 

References