AmericanChemicalSociety.com

With the support of this ACS-PRF grant we have made significant progress in the discovery of new dehydrative reactions of unsaturated alcohols.  Our work has been focused on developing allylic and propargylic alcohols as electrophiles in gold-catalyzed processes whereby a tethered nucleophile adds to the pi-bond with concomitant loss of water (Scheme 1).  This is a new mode of reactivity in the burgeoning area of homogenous gold catalysis, which enables the preparation of a variety of ubiquitous structural motifs and as such should find use in total synthesis.

Scheme 1.  Au-catalyzed dehydratve cyclization reactions.

At the outset, it was our intention to search for new catalytic heterofunctionalization reactions of alkenes that proceed under mild reaction conditions and would be suited for use in complex molecule synthesis.  Through a series of preliminary experiments, it was found that allylic alcohols are extremely reactive towards gold catalysts.  When substrates containing an allylic alcohol tethered to a nucleophile were treated with 1 mol % Ph_3­PAuCl and 1 mol% AgOTf in methylene chloride, a clean cyclization smoothly proceeds with elimination of water. Careful control experiments demonstrated that both the gold and silver complexes are necessary for the reaction. Further studies were designed to probe the effects of substitution on the allyl moiety and the results are shown in Figure 1.  A variety of substrates performed well in the reaction including cis olefins, aldol adducts, tertiary allylic alcohols, tri-substituted olefins (forming a quaternary stereogenic center), and other assorted functional groups.  The reaction proceeds in high yield in all cases with high cis selectivity when applicable.  In addition, the reactions could be performed at temperatures as low as -78 °C and the catalyst loadings could be reduced to 0.1 mol % without negative consequences.

Based on these results, the analogous propargylic diols would be predicted to undergo similar reactions, although the products produced would be allenol ethers.  We reasoned that these reactive intermediates could be utilized in an efficient process if an additional alcohol were present in the molecule, and would form unsaturated spiroketals such as those contained in natural products like spirastrellolide A and okadaic acid.  The reaction sequence would be predicted to proceed from the triol 15 to the spiroketal 17 via the allenol ether 16 (Scheme 2).  In principle, both processes could be catalyzed by a cationic metal.  This route was particularly attractive because the olefin would be precisely placed according to the general scheme leaving no ambiguity in the sizes of the rings formed, which is often a challenging issue.

      Gratifyingly, this reaction sequence works extraordinarily well, smoothly generating the desired products in high yield (Figure 2).  Interestingly, the reaction functions best when Au[P(t-Bu)­₂(o-biphenyl)]Cl is employed with AgOTf in THF at 0 °C.  The loading can be reduced to 2 mol% with reaction times ranging from minutes to just less than 1.5 h.  Control experiments suggest that Au[P(t-Bu)­₂(o-biphenyl)]OTf is the active catalyst and rule out the possibility that the reaction is catalyzed by TfOH. 

      A likely catalytic cycle for the reaction is presented in Scheme 3.  Experimental evidence using the diol generated by protection of the C1 or C9 hydroxyl group suggests that either the C1 or C9 alcohol may cyclize first but both events eventually lead to the same product.

      This proposal lead to the idea that the dual catalytic cycles could function independently and provide unique reaction pathways.  If cycle B were to be accessed with substrates devoid of a second nucleophilic alcohol, a different mechanism for further reactions from 31 would be necessary.  To address this, heteroatom substituted substrates such as 34 were prepared and treated under the same reaction conditions (Figure 3).  These systems rapidly aromatize to form 5-membered heteroaromatics in high yield.  Additionally, it was found that AuCl could be used as catalyst in an open vessel without taking any precautions to exclude air or moisture.  The reaction is very simple to perform (essentially no precautions are necessary), rapid, high-yielding, exceptionally clean (no additional product purification is generally necessary), and the catalyst loadings are exceedingly low (as low as 0.05 mol%). Efforts are currently underway to utilize catalytic cycle A (Scheme 3).

Since these aromatic heterocycles were easily prepared, we sought to explore the synthesis of additional heterocycles, namely 2-H-chromenes.  In the process, it was found that endo-cyclization reactions of allylic alcohols can also be catalyzed by Au-salts, but the conditions required are much more forcing (Figure 4).  Starting from substituted benzaldehydes, a wide variety of differentially substituted chromenes are readily available in two steps.

The results outlined above demonstrate that unsaturated alcohols undergo a variety of dehydrative transformations under extremely mild conditions.  These methods enable facile preparation of highly useful cyclic ethers, monounsaturated spiroketals, and five-membered heteroaromatics from readily prepared diols and triols.  The reactions are rapid and generally high yielding providing a concise synthesis of functional building blocks in short order.  Additional reactions of this ilk are being developed and will be reported in due course.