Studies on the reactivity of cis-4-benzyloxy-1,2-epoxycyclohexane

The reaction course of the selenation-oxidation-elimination sequence carried out from cis-4-benzyloxy-1,2-epoxycyclohexane ( cis-1 ) is studied. Contrary to literature precedents, this transformation leads to an unexpected diastereomeric cyclohexenol, whose formation can be interpreted by stereoelectronic grounds. In addition, the base induced rearrangement of cis-1 with lithium amide bases is also discussed. In the absence of external additives, a mixture of cyclohexenols arising from the competition of syn and anti elimination processes is observed. However, in the presence of Li salts, the corresponding cyclohexenol arising from an apparent anti elimination pathway predominates. A mechanistic rationale is proposed to account for these observations.


Introduction
In the course of our recent research, the use of cis-4-benzyloxy-1,2-epoxycyclohexane (cis-1) has disclosed interesting reactivity features that deserve some attention.Thus, a literature report 1,2 describes the use of cis-1 as starting material for the synthesis of cis-5-benzyloxy-2-cyclohexenol (cis-2) by phenyl selenation followed by oxidation to the corresponding selenoxide and in situ elimination (Scheme 1).

Results and Discussion
Attempts to reproduce the above sequence required the preparation of epoxide cis-1, which was obtained uneventfully from 4-benzyloxycyclohexene following literature protocols. 3Reaction of cis-1 with sodium phenylselenide (obtained in situ by reaction of diphenyldiselenide with NaBH 4 in EtOH) was carried out as described in the literature. 1,2Mechanistic considerations concerning the putative reaction pathway involved in the transformation of cis-1 into the expected allylic alcohol cis-2, would require phenylselenide attack to afford phenylselenyl derivative 3 through a chelated reactive conformation cis-1A(Na), in agreement with the trans-diaxial attack imposed by the Fürst-Platner rule. 4The above reaction course would be imperative for the subsequent syn elimination of selenoxide 4 5 required to give alcohol cis-2 (see Scheme 2).However, taking into account the relatively low chelating ability of the Na ion to force the above reactive conformation cis-1A(Na), 6 a thorough examination of the reaction outcome was undertaken.Thus, operation of a non-chelating reactive conformation cis-1B on reaction of cis-1 with phenylselenide would afford phenylselenide 5 (Scheme 3), whose oxidation to selenoxide 6, followed by syn-elimination, 5 would lead to the isomeric allylic alcohol cis-7, as depicted in Scheme 3. As expected from the above hypothesis, and contrary to literature precedents, 1,2 alcohol cis-7 was formed in this process instead of cis-2.Formation of cis-7 was confirmed by comparison of its spectroscopical data with those described in the literature for this alcohol 7 and isomeric cis-2. 8n light of these results, we explored the potential of epoxide cis-1 as starting material for the synthesis of isomeric alcohols cis-2 and cis-7.In this context, synthesis of cis-7 is described in the literature from base-induced rearrangement of epoxide cis-1. 7However, despite the well recognised synthetic usefulness of this transformation, 9,10 the reaction outcome can be dramatically affected by the nature of the base, the solvent, and the reaction temperature, among others. 11,12Thus, computational and experimental studies carried out on cyclohexane oxides have shown a switch from syn-β elimination in non polar solvents to a more energetically favourable anti-β elimination in polar solvents, such as HMPA. 13Our experiments carried out from epoxide cis-1 under different reaction conditions (base, solvent, temperature, and LiClO 4 as chelating agent) are shown in Table 1.An initial experiment in LDA/Et 2 O at rt (entry 1) showed the formation of a roughly 1:1 mixture of allylic alcohols cis-2 and cis-7, which can be interpreted as a result of the operation of competing anti and syn elimination processes, respectively, from the most stable conformation cis-1B, (Scheme 4). 14Based on our previous results, 15 addition of LiClO 4 (5 equiv/mol) is known to drive the reaction mixture towards a chelated reactive conformation cis-1A(Li) (Scheme 5).This conformation was expected to favour the operation of a syn elimination leading ultimately to cis-2.However, contrary to our assumption, alcohol cis-7 was the major one under the above conditions (entry 2).Similar results were obtained in the presence of THF as a solvent (entry 3), although unreacted starting epoxide cis-1 was the major or exclusive one at lower temperatures (entries 5, 6).The use of a larger excess LDA (3 equiv/mol, entry 4) was unsuccessful, since noticeable amounts of alcohol 8 (R=iPr) was also observed (see Scheme 5).This side reaction was even more important under reflux conditions, since 8 (R=iPr) was the major compound (entry 7).Moving from LDA to other lithium amide bases led to similar results, irrespective of their steric demand or reaction conditions (entries 8-10).In all cases, only alcohol cis-7 was formed as a result of a base-induced rearrangement process.Finally, contrary to literature precedents, 7 the use of HMPA did not improve the reactivity of the lithium amides, since poor conversions were observed in all cases.The above results, in the presence of LiClO 4 as chelating agent, can be interpreted by considering Scheme 5. Thus, due to the well recognized ability of Li ions to promote a reactive chelating conformation in epoxy cyclohexanes, 6,15 conformation cis-1A(Li) can account for the reactivity of epoxide cis-1 in the presence of LiClO 4 (5 equiv/mol).The commonly accepted syn elimination pathway for this kind of LDA promoted rearrangements would require a previous Li-ligand exchange to accommodate the amide base in a proper orientation for a subsequent abstraction of the vicinal pseudoaxial proton 10 (conformation cis-1C, Scheme 5).This exchange process might be slower than an alternative anti elimination pathway leading ultimately to alcohol cis-7.This reactive conformation would also explain formation of amino alcohols 8 as a result of nucleophilic attack of the amide base following a trans-diaxial pathway.In summary, the above results represent an additional proof of the generality of the Fürst-Platner rule on the reactivity of epoxy cyclohexane derivatives, as evidenced by the results obtained from epoxide cis-1 on reaction with phenyl selenide anion.On the other hand, studies on the base-induced rearrangement of epoxide cis-1 in the presence of LiClO 4 show the ability of this additive to facilitate an anti elimination process leading ultimately to allylic alcohol cis-7 as the major reaction product.

Experimental Section
General Procedures.Solvents were distilled prior to use and dried by standard methods. 16elting points are uncorrected.FT-IR spectra are reported in cm -1 . 1 H and 13 C NMR spectra were obtained in CDCl 3 solutions at 300 MHz (for 1 H) and 75 MHz (for 13 C), respectively, unless otherwise indicated.Chemical shifts are reported in delta (δ units, parts per million (ppm) relative to the singlet at 7.24 ppm of CDCl 3 for 1 H and in ppm relative to the center line of a triplet at 77.0 ppm of CDCl 3 for 13 C. 7 Sodium borohydride (950 mg, 25 mmol) is added portionwise to a solution of diphenyldiselenide (3.7 g, 12 mmol) in EtOH (10 mL) under nitrogen.The reaction mixture is stirred at rt for 10 min.The mixture is then treated with a solution of epoxide cis-1 (1g, 4.89 mol) in EtOH (5 mL).After stirring for 45 min at rt, the reaction mixture is diluted with THF (7 mL), treated with 30% H 2 O 2 (5.2 mL added dropwise), and heated to reflux temperature with vigorous stirring.After 8h, the mixture is concentrated in vacuo, treated with H 2 O (10 mL) and extracted with Et 2 O (3 x 20 mL).The combined organic extracts are dried over MgSO 4 , filtered and evaporated to afford a crude residue which was flash chromatographed on hexanes/EtOAc (2/1) to afford alcohol cis-7 (350 mg, 35 % yield).General procedure for the reactions of epoxide cis-1 with lithium amide bases Lithium amide bases were typically prepared by treatment at -78ºC of a solution of the corresponding amine (4.5 mmol) in the required solvent (4 mL) with BuLi (2.5 mL of a 1.6 N solution in hexanes).This affords a solution containing 4 mmol LDA, approximately.The above mixture is allowed to warm to the required temperature and next treated with a solution of epoxide cis-1 (500 mg, 2.45 mmol for a base/substrate ratio of 1.5) containing LiClO 4 (1.3 g, 12.25 mmol) in the required solvent (6 mL).The reaction mixture was quenched by careful addition of H 2 O (1 mL).The organic phase was extracted, dried, and evaporated in vacuo to afford a residue which was purified by flash chromatography to afford the reaction products (see Table 1).

Scheme 2 .
Scheme 2. Proposed mechanism to account for the formation of cis-2 by syn elimination of phenyl selenoxide intermediate 4.

Scheme 3 .
Scheme 3. Proposed mechanism to account for the formation of cis-7 by syn elimination of phenyl selenoxide intermediate 6.

Scheme 4 .
Scheme 4. Proposed reaction mechanisms to account for the formation of alcohols cis-2 and cis-7 from a common non-chelated reactive conformation cis-1B.

LDAScheme 5 .
Scheme 5. Mechanistic interpretation to account for product distribution on reaction of epoxide cis-1 with LDA in the presence of LiClO 4.

Table 1 .
Reactivity of epoxide cis-1 under basic conditions in the presence of LiClO 4