1,3-Dipolar Cycloaddition of Azomethine Ylides to Alkenylboronates: Synthesis of Boronic Ester Substituted Pyrrolidine Derivatives and Asymmetric Approaches

Abstract

1,3-Dipolar cycloadditions have been known as one of the highly useful synthetic methodologies which provide a variety of biologically important heterocycles. Among those, the cycloaddition of an azomethine ylide to an alkene affords a pyrrolidine moiety which is another very useful class of compounds for natural product synthesis. For instance, 3pyrrolidinol derivatives have recently attract attention in novel routes to prolyl peptides and cyclopeptide alkaloides. Although 3-pyrrolidine derivatives are commercially available, the number of this type of derivatives is limited and valuable enough to explore further. There are several reports of the synthesis of 3-pyrrolidinol derivatives. Intrigued with those literature, we envisioned that the cycloaddition reaction of alkenylboronates with azomethine ylides might afford 3-boronic ester substituted pyrrolidine derivatives. The reason we chose alkenylboronate as a dipolarophile is that the boronic ester functionality can be converted into a variety of other useful functionalities such as alcohol, amine, aldehyde, caboxylic acid, and others. Here we report that 1,3-dipolar cycloaddition reaction of alkenylboronates with azomethine ylides and asymmetric cycloaddition using chiral auxilaries equipped to vinylboronates. First, we examined vinylboronate 1 and ester and toluenesulfonyl substituted vinylboronic esters, 2 and 3, which were prepared from hydroboration and transesterification of methyl propiolate and ethynyl p-toluenesulfone respectively. The cycloaddition of vinylboronate 1 to in situ generated Nbenzyl azomethine ylide was pretty much sluggish to afford cycloadduct only 20% yield under various conditions. However, when an ester or sulfone substituted alkenylboronate, 2 or 3, were employed, the reaction became much accelerated and smoothly proceeded to afford the product in good yields (85% and 75%). It may be noticed that electron deficient substituent in the alkenylboronate might activate this dipolarophile toward azomethine ylide. Similarly, N-methyl azomethine ylide also was equally good toward the activated dipolarophile 2 to give the product in good yield (scheme 1). Then we employed a different type azomethine ylide precursor, N-(methoxymethyl)-N-(trimethylsilylmethyl)benzylamine, which is triggered into an active ylide upon catalytic treatment of a strong acid. We were delighted to find that the reaction became much facilitated and afforded the products in good yields. The reaction of 1 with 10 under standard condition afforded the cycloadduct 6 in 75% yield which was markedly improved when compared to Scheme 1. It should be noticed that the reaction was well done at low temperature and the vinylboronate 1 was not sluggish toward an ylide any more in this case. It is assumed that in situ generated ylide in Scheme 2 may stay longer due to low temperature and have enough time to react with a dipolarophile rather than under Scheme 1. The compound 6 was chosen and subjected to oxidation/debenzylation to give 3pyrrolidinol (12) in overall 83% yield which is commercially available (Scheme 2). Then we turned our attention to asymmetric 1,3-dipolar cycloaddition by introducing appropriate chiral auxiliaries into a vinylboronic ester. There have been the number of examples of asymmetric 1,3-dipolar cycloadditions using chiral auxiliaries where most of them were attached onto the dipolarophiles. Some years ago, we reported the asymmetric cyloadditions of nitrile oxides to various alkenylboronates which were covalently bonded to α,α,α',α'-Tetraaryl-1,3dioxolan-4,5-dimethanol (TADDOL) chiral auxiliaries. In Scheme 1. Reaction conditions; i) CH2O, toluene, reflux.