|Home | About | Journals | Submit | Contact Us | Français|
A variety of non-conjugated cis-olefins has been enantioselectively epoxidized with chiral ketones 2, and up to 92% ee has been obtained. The two prochiral faces of an olefin are likely stereodifferentiated by the relative hydrophobicity of the olefin substituents and/or allylic oxygen functionality.
Chiral ketones have been shown to be effective catalysts for asymmetric epoxidation of olefins.1 In our own studies, fructose-derived ketone 1 (Figure 1) has been shown to be a very effective catalyst for the epoxidation of trans- and trisubstituted olefins.2 It has also been found that oxazolidinone-bearing ketones 2 can give high ee’s for olefins such as conjugated aromatic cis-olefins,3a,c,d,e,g conjugated cis-dienes,3k enynes,3a,c,l styrenes,3b,c,d,f certain trisubstituted,3h,j and tetrasubstituted olefins.3i,j,4 For epoxidation of cis-olefins with ketones 2, the reacting C-C double bond has usually been conjugated with an aromatic group, an alkene, or an alkyne which directs the enantioselective epoxidation via an apparent attraction with the spiro oxazolidinone of the ketone catalyst (Figure 2). For asymmetric epoxidation of non-conjugated cis-olefins, 67% ee was obtained for cis-1-cyclohexyl-1-propene, and 94–97% ee was obtained for 3,3-ethylenedioxycycloalkenes with ketone 2a (Scheme 1).3a,c,5,6 To expand the substrate scope of chiral ketone-catalyzed epoxidation of non-conjugated cis-olefins and gain better understanding of stereodifferentiation factors, we undertook further investigations on asymmetric epoxidation of this class of olefins with glucose-derived ketones 2. Herein we wish to report our studies on this subject.
During our studies on asymmetric epoxidation of conjugated cis-dienes and enynes with ketones 2, it became apparent that the relative hydrophobicity of the olefin substituents had a significant effect on enantioselectivity.3k,l To further probe this hydrophobic effect on stereodifferentiation, cis-2-nonene was epoxidized with ketones bearing different N-substituents, giving 44%, 56%, 58%, 64% (Table 1, entry 1), and 54% ee respectively for ketones 2a–e.7 The ee initially increased with increasing length of the p-alkyl chain on the aryl ring of catalysts 2b–d (from Me to n-Bu) but decreased with further increasing length of the alkyl chain (n-C10H21) (2e). The epoxidations are run in aqueous solvent mixtures, and although the absolute configuration of the epoxide8a could not be unambiguously determined yet, the results suggest that the enantioselectivity is likely derived from hydrophobic interactions between the substrate and the catalyst with the hydrophobic n-hexyl group of the olefin aligned adjacent to the N-aryl group of the catalyst in the transition state (spiro C favored over spiro D) (Figure 3, X = H, n = 6).
The ee increased when the other substituent on the olefin became more hydrophilic (Table 1, entries 2–5) likely due to spiro C being further favored over spiro D (Figure 3).9 Up to 91% ee was obtained for cis-dec-4-enoic acid (entry 4). For entries 4 and 5 the products were obtained as five and six-membered lactones respectively.10 The lactone in entry 4 is the enantiomer of a natural product isolated from Streptomyces griseus and has been the subject of several synthetic investigations.11 The high enantioselectivity observed for entries 4 and 5 (Table 1) is likely due to the extreme difference in hydrophilicity between the two olefin substituents. The carboxylic acids are presumably deprotonated under the basic reaction conditions to give the corresponding carboxylates which are charged polar groups.
Good ee’s can also be obtained for certain allylic ethers (Table 1, entries 6–8). An allylic acetal was a very effective substrate (Table 1, entry 10), but an acyclic allylic ketal (entry 11) was not, in contrast to the cyclic ketals previously studied (Scheme 1). All-carbon analogs of an aromatic ether and a cyclic ketal gave much lower ee’s than their oxygen-containing counterparts (Table 1, entries 9 and 12) indicating the oxygen atoms are important for stereodifferentiation. The absolute configurations of the epoxides from 3,3-ethylenedioxycyclohexene (Scheme 1) and the allylic ethers of entries 6 and 8 (Table 1) indicate that ketal and ether substituents prefer to be proximal to the oxazolidinone of 2a during the transition states (Figure 4) (for methods used to determine absolute configuration, see Supporting Information).
All these results suggest that there are two main mechanisms of stereodifferentiation operating for these olefins. For substrates in entries 1–5 (Table 1), the relative hydrophobicity of the olefin substituents is the dominant mechanism of stereodifferentiation, and ketone 2d is most effective. The second mechanism differentiates substituents that contain an allylic oxygen functionality from those that do not, and ketone 2a is most effective when this mechanism dominates. In the case of an allylic alcohol (Table 1, entry 2) both mechanisms might be operating in competition with each other. In this case the free hydroxyl group is apparently hydrophilic enough to override the second mechanism to give the observed epoxide configuration. Based purely on hydrophobic considerations, cis-3-nonen-1-ol (Table 1, entry 3) would be expected to give lower ee than cis-2-nonen-1-ol (entry 2) since the alkyl substituent is one carbon shorter, and the alcohol substituent is one carbon longer. However, higher ee is observed (82% vs 79% ee). This result indicates that the apparent attraction between oxygen-containing functionality and the oxazolidinone of the catalyst is significantly weakened when the oxygen-containing functionality is not in the allylic position.
Entry 6 is another case where both mechanisms of stereodifferentiation appear to be operating in competition with each other. However, in this case the methyl ether substituent is not hydrophilic enough to override the apparent ether-oxazolidinone attraction with ketone 2a, so the (2R,3S) enantiomer predominates. With ketone 2c hydrophobic properties again dominate, but only slightly, giving 14% ee of the (2S,3R) enantiomer.16 The high enantioselectivity observed in the epoxidation of the allylic acetal of entry 10 (Table 1) along with the previous observations of high ee with spirocyclic ketals (Scheme 1) indicate that two allylic oxygens on the same side of the olefin create an even stronger apparent attraction to the oxazolidinone of the catalyst than one. The allylic ketal of entry 11 (Table 1) was an ineffective substrate possibly because the olefin is too hindered or because it cannot adopt the optimal conformation for the interaction between the ketal of the olefin and the oxazolidinone of the catalyst due to A1,3 strain.
The origin of enantioselectivity for substrates which rely on differences in hydrophobicity of the olefin substituents is fairly straightforward (Figure 3). However, the origin of the apparent attraction for allylic oxygen functionality to the oxazolidinone of the catalyst is not clear, and there are several possible rationales. One possibility is that there is an attraction between the electron lone pairs on the oxygen atoms and a partial or transient positive charge on the oxazolidinone (Figure 5) which would favor spiro G (Figure 4 and Figure 5). Another possibility is that repulsion exists between the electron lone pairs of the oxygen atoms of the substrate and the fused ketal of the catalyst in spiro H, thus disfavoring this transition state. A better understanding awaits further studies.
In summary, the scope of the ketone-catalyzed asymmetric epoxidation reaction has been expanded to include several types of non-conjugated cis-olefins, and good to high ee’s have been obtained for a number of substrates. If the two substituents of non-conjugated cis-olefins have substantially different hydrophobic or electronic properties, this system could provide a good opportunity for asymmetric induction. This study opens up a new avenue for this system that will be valuable for further studies and designing new ketone catalysts in the future.
We are grateful for the generous financial support from the General Medical Sciences of the National Institutes of Health (GM59705-08). We thank Drs. Xuegong She, Hongqi Tian, and David Goeddel (Colorado State University) for some earlier studies.
Supporting Information Available Synthesis and characterization of olefins and epoxides, as well as the data for the determination of the enantiomeric excess and the NMR spectra of the epoxides (54 pages). This material is available free of charge via the Internet at http://pubs.acs.org.