Bismuth(III) salts provide a convenient, inexpensive, and environmentally benign source of the corresponding Brønsted acid.^1-3 For example, the hydrolysis of bismuth(III) bromide is known to afford two equivalents of hydrogen bromide and insoluble bismuth oxybromide, in which the former is responsible for the observed catalysis.^2a We recently demonstrated that this reagent promotes the intramolecular reductive and two-component etherification reactions for the construction of cis- and trans-2,6-disubstituted tetrahydropyrans, respectively.² A striking feature of this approach is the ability to initiate the cyclization based on the nature and proximity of the triorganosilyl ether.

In a program directed toward exploring the synthetic utility of bismuth(III) salts, we envisioned using a bismuth(III) salt to accomplish an intramolecular oxa-conjugate addition for the stereoselective synthesis of cyclic ethers.^4,5 Although the intramolecular oxa-conjugate addition with alkoxides has been extensively studied, there is a surprising paucity of Brønsted acid catalyzed versions of this transformation.^6,7 Herein, we now describe the intramolecular bismuth-mediated oxa-conjugate addition of tethered trialkylsilyloxy substituted α,β-unsaturated ketones 1 for the stereoselective construction of cis-2,6-disubstituted tetrahydropyrans 2 (Eq. 1).

Equation 1

Preliminary studies focused on the development of the intramolecular bismuth-mediated oxa-conjugate addition using the triethylsilyloxy substituted α,β-unsaturated ketone 1a (R = PhCH₂) as outlined in Table 1.⁸ Treatment of 1a with a catalytic amount of bismuth(III) bromide in reagent grade acetonitrile at room temperature, furnished the 2,6-disubstituted tetrahydropyrans 2a/3a in 16% yield, favoring the cis-diastereoisomer 2a (entry 1). Although the yield was not particularly encouraging, the high stereoselectivity and the fact that the unreacted α,β-unsaturated ketone 1a was completely recovered stimulated further optimization studies. The low conversion was initially attributed to an inefficient protodesilylation of the secondary triethylsilyl ether by the hydrogen bromide.^11,12 This prompted the examination of related bismuth(III) salts (entries 4–7), in which the halides gave analogous results (entries 4–5), whereas the pseudo halide (entry 6) afforded 2a with somewhat improved efficiency and diastereoselectivity. Gratifyingly, bismuth(III) nitrate, which is commercially available as the pentahydrate, gave the cis-2,6-disubstituted tetrahydropyran 2a in quantitative yield and with excellent diastereoselectivity (entry 7).^13,14 This dramatic improvement was attributed to the increased water content rather than the acidity of the requisite Brønsted acid. Consistent with this hypothesis, the addition of water to the bismuth(III) bromide mediated reaction, provided the cis-2,6-disubstituted tetrahydropyran 2a in an analogous manner to bismuth(III) nitrate pentahydrate (entry 2 vs 7). Moreover, the addition of the hindered base 2,6-di-tert-butyl-4-methylpyridine to neutralize the Brønsted acid, led as expected to no observable reaction (entries 3 and 8).

Table 1 Optimization for the intramolecular bismuth-mediated oxa-conjugate addition (Eq. 1; 1a, R = PhCH2)a

Table 2 summarizes the application of the optimized reaction conditions (Table 1, entry 7) to various tethered triethylsilyloxy substituted α,β-unsaturated ketones.¹⁵ The intramolecular oxa-conjugate addition is tolerant of both linear (entries 2 and 3) and branched substituents (entries 4 and 5), in which even the tert-butyl substituent (entry 6) affords the cyclic ether in 99% yield with excellent diastereoselectivity. The intramolecular oxa-conjugate addition was also applicable to substrates containing protected heteroatoms (entries 7–10), which is significant given that these products provide important synthons for synthetic applications.^6,7 Interestingly, the nitrogen containing substrate required an electron-withdrawing group, which presumably avoids the basic nitrogen sequestering the Brønsted acid (entry 7). Another interesting feature of this study was that the benzyl protected alcohol derivative afforded lower diastereoselectivity (entry 8), which was attributed to the formation of a hydrogen-bonded hemiacetal that alters the facial selectivity in the oxa-conjugate addition.¹⁶ This problem was easily circumvented by either replacing the benzyl ether with an electron-withdrawing protecting group (entry 9), or a bulky trialkylsilyl ether (entry 10), which result in higher diastereoselectivity. Moreover, the latter example is particularly impressive, since it demonstrates that certain primary trialkylsilyl ethers may be tolerated, without competitive desilylation.

Table 2 Scope of the intramolecular bismuth(III) nitrate pentahydrate mediated oxa-conjugate additiona

In an attempt to extend the scope of this reaction, tertiary triethylsilyl ethers were also examined in the context of the diastereoselective intramolecular oxaconjugate addition (Eq. 2). Treatment of the triethylsilyloxy substituted α,β-unsaturated ketone 4 under the optimized reaction conditions, albeit with 20 mol % bismuth(III) nitrate pentahydrate, furnished the tetrahydropyrans 5/6 in 94% yield, with ≥ 19:1 diastereoselectivity favoring 5.¹⁸ The increased loading is presumably due to the increased stability of the tertiary trialkylsilyl ethers toward protodesilylation.

Equation 2

In conclusion, we have developed a general and highly diastereoselective intramolecular Brønsted acid catalyzed oxa-conjugate addition for the construction of cis-2,6-disubstituted tetrahydropyrans using bismuth(III) nitrate pentahydrate. This work demonstrates the importance of water in facilitating the hydrolysis of the bismuth(III) salt, to afford the Brønsted acid necessary to promote protodesilylation. Furthermore, we have provided compelling evidence that it is the Brønsted acid rather than the bismuth(III) salt, which is responsible for the catalysis. Finally, we extended this methodology to a tertiary alcohol nucleophile, thereby making this a mild and efficient method that will be particularly attractive for applications to target directed synthesis.

We sincerely thank National Institute of Health (GM054623) for generous financial support. We also thank Johnson and Johnson for a Focused Giving Award, and Pfizer Pharmaceuticals for Creativity in Organic Chemistry Award (P.A.E.).