Base-Mediated Claisen Rearrangement of CF3-Containing Bisallyl Ethers

We have previously clarified that the strongly electron-withdrawing CF3 group nicely affected the base-mediated proton shift of CF3-containing propargylic or allylic alcohols to afford the corresponding α,β-unsaturated or saturated ketones, respectively, which was applied this time to the Claisen rearrangement after O-allylation of the allylic alcohols with a CF3 group, followed by isomerization to the corresponding allyl vinyl ethers via the proton shift, enabling the desired rearrangement in a tandem fashion, or in a stepwise manner, the latter of which was proved to have attained an excellent diastereoselectivity with the aid of a palladium catalyst.


Introduction
It is widely understood that strategic entry of fluorine atoms or fluorinated groups to adequate molecules gave strong impact to the original character in many instances, and thus the development of novel methods for the construction of a variety of such compounds has attracted significant attention of researchers working in the field of synthetic organic chemistry, material science, and biologically active compounds [1][2][3][4][5]. For this reason, we have been studying to realize facile preparation of such molecules, and recently reported an interesting proton transfer starting from both propargylic [6] and allylic alcohols [7], enabling to form diverse α,β-unsaturated and saturated ketones, respectively, just by their treatment with very convenient as well as easy-to-handle tertiary amines. The representative example for the latter was described in Scheme 1. Presence of the electronwithdrawing CF 3 group was considered to play a crucial role in the increase of the acidity of a proton H a at the allylic position of 1. Actually its abstraction was realized by the action of such a weak base as diazabicyclo [5.4.0]undec-7-ene (DBU) under the toluene refluxing condition which resulted in the simple isomerization to the intermediates Int-1, followed by the conversion to their keto form 2 to complete this interesting sequence. This reaction mechanism was proved by our own computation for the transition state [7] as well as by experimental employment of the deuterated substrate by the other group [8]. On the basis of this successful as well as convenient proton shift process starting from the CF 3 -containing allylic alcohols 1, we envisaged its intriguing extension from the synthetic point of view: thus, our idea was that the bisallylic ethers 3 possibly synthesized in a facile manner by way of the O-allylation of 1 were recognized as the potential substrates for the Claisen rearrangement as long as the proton shift of 3 to 4 was possible. It is worthwhile to note that, except for our previous report [9], there are no such examples to prepare the compounds 5 by way of [3,3]-sigmatropic rearrangement irrespective of the substituents R 2 , while there are some precedented work on the alkylation route to alternatively get access to compounds like 5 [10][11][12][13]. We thus started our research for the novel utilization of the CF 3 -containing bisallylic ethers 3 as the potent substrates for the Claisen rearrangement via the facile isomerization to the corresponding allyl vinyl ethers 4.
CF3-containing bisallylic ethers 3 as the potent substrates for the Claisen rearrangement via the facile isomerization to the corresponding allyl vinyl ethers 4. Scheme 1. Isomerization of 1 to 2 by the proton shift and concept of the present study.

Preparation of Bisallyl Ethers 3
Preparation of the CF3-containing allylic alcohols 1 was carried out in a stereoselective fashion following our own developed method like (1) reactions of adequate Grignard reagents with CF3CO2Et to construct the ketones CF3-C(O)-R 1 , (2) their Horner-Wadsworth-Emmons reactions with (EtO)2P(O)CH2C(O)R 2 [14], and (3) NaBH4 reduction of the resultant α,β-unsaturated ketones. Important to note is the fact that the sequences (1) and (2) could be performed without isolation of the intermediary trifluorinated ketones which allowed the possible formation of 1 even with a "small" R 1 (for example, 1d with a Et group as R 1 ) whose isolation is usually difficult due to their low boiling points and high volatility [7,15].
Optimization of the reaction conditions for the O-allylation was initially performed using the allylic alcohol 1a as the representative model whose results are summarized in Table 1. First of all, investigation of a base clarified that an excess amount of a 6 M NaOH aqueous solution was the best among tested for the construction of the desired 3a (Entries 1 to 5). Because BuLi unexpectedly recorded complete recovery (Entry 4), we further changed the reaction temperature or modified the reactivity by the addition of hexamethylphosphoric triamide (HMPA), but both did not give any fruitful results (Entries 15 or 16, respectively). Formation of the saturated ketone 2a as the byproduct was interpreted as a result of the abstraction of a proton at C 1 in 1a, followed by re-protonation at C 3 , and the stronger bases showed a clear tendency to prefer this isomerization except for BuLi. After fixing the base as 6 M NaOH aq., further brief check of a phase transfer catalyst (PTC) pointed out that Bu4NI was the reagent of choice (Entries 5 to 7) combined with dichloromethane (DCM) as a solvent (Entries 5, 8, and 9). Entries 10 to 14 were carried out for determination of the best amount of allyl bromide and it was concluded that 2.0 equiv would suffice for our purpose. Final examination on the concentration (Entries 12, 17, and 18), the equiv (Entries 12, 19, and 20) of NaOH, and the reaction period (Entries 12, 22, and 23) led to the final conclusion that the conditions shown in Entry 23 was the best of all for the O-allylation of 1a. Scheme 1. Isomerization of 1 to 2 by the proton shift and concept of the present study.

Preparation of Bisallyl Ethers 3
Preparation of the CF 3 -containing allylic alcohols 1 was carried out in a stereoselective fashion following our own developed method like (1) reactions of adequate Grignard reagents with CF 3 CO 2 Et to construct the ketones CF 3 -C(O)-R 1 , (2) their Horner-Wadsworth-Emmons reactions with (EtO) 2 P(O)CH 2 C(O)R 2 [14], and (3) NaBH 4 reduction of the resultant α,β-unsaturated ketones. Important to note is the fact that the sequences (1) and (2) could be performed without isolation of the intermediary trifluorinated ketones which allowed the possible formation of 1 even with a "small" R 1 (for example, 1d with a Et group as R 1 ) whose isolation is usually difficult due to their low boiling points and high volatility [7,15].
Optimization of the reaction conditions for the O-allylation was initially performed using the allylic alcohol 1a as the representative model whose results are summarized in Table 1. First of all, investigation of a base clarified that an excess amount of a 6 M NaOH aqueous solution was the best among tested for the construction of the desired 3a (Entries 1 to 5). Because BuLi unexpectedly recorded complete recovery (Entry 4), we further changed the reaction temperature or modified the reactivity by the addition of hexamethylphosphoric triamide (HMPA), but both did not give any fruitful results (Entries 15 or 16, respectively). Formation of the saturated ketone 2a as the byproduct was interpreted as a result of the abstraction of a proton at C 1 in 1a, followed by re-protonation at C 3 , and the stronger bases showed a clear tendency to prefer this isomerization except for BuLi. After fixing the base as 6 M NaOH aq., further brief check of a phase transfer catalyst (PTC) pointed out that Bu 4 NI was the reagent of choice (Entries 5 to 7) combined with dichloromethane (DCM) as a solvent (Entries 5, 8, and 9). Entries 10 to 14 were carried out for determination of the best amount of allyl bromide and it was concluded that 2.0 equiv would suffice for our purpose. Final examination on the concentration (Entries 12, 17, and 18), the equiv (Entries 12, 19, and 20) of NaOH, and the reaction period (Entries 12, 22, and 23) led to the final conclusion that the conditions shown in Entry 23 was the best of all for the O-allylation of 1a.
The optimized reaction conditions determined as above were employed for the synthesis of a variety of the CF 3 -containing bisallylic ethers 3 whose results are collected in Table 2. In spite of formation of the side product, ketones 2, in small amounts, good to excellent isolated yields were attained for the construction of the desired compounds 3 in many instances. However, this was not the case for the substrates 1 possessing alkylsubstituents as R 2 whose electron-donating effect seemed to lower the acidity of an OH group to slow down the reaction rate and consequently led to recovery of the substrates to some extent (Entries 8 and 9). In our previous report on the 1,3-proton shift of the compounds 1 to the ketones 2 [7], the clear substituent effect for R 2 was experimentally as well as computationally manifested, and aromatic groups were required for the smooth promotion of this isomerization by effective increase of the acidity of the proton at C 1 . It is important to mention that the clear contrast was pointed out for the substituents R 1 whose effect was, different from the instance of R 2 , only limited.   5 Alkoxide was prepared by the addition of BuLi at 0 °C for 10min, followed by changing the temperature as depicted. 6 Reaction was performed at reflux temperature. 7 3.0 equiv of HMPA was added.
The optimized reaction conditions determined as above were employed for the synthesis of a variety of the CF3-containing bisallylic ethers 3 whose results are collected in Table 2. In spite of formation of the side product, ketones 2, in small amounts, good to excellent isolated yields were attained for the construction of the desired compounds 3 in many instances. However, this was not the case for the substrates 1 possessing alkylsubstituents as R 2 whose electron-donating effect seemed to lower the acidity of an OH group to slow down the reaction rate and consequently led to recovery of the substrates to some extent (Entries 8 and 9). In our previous report on the 1,3-proton shift of the compounds 1 to the ketones 2 [7], the clear substituent effect for R 2 was experimentally as well as computationally manifested, and aromatic groups were required for the smooth promotion of this isomerization by effective increase of the acidity of the proton at C 1 . It is important to mention that the clear contrast was pointed out for the substituents R 1 whose effect was, different from the instance of R 2 , only limited.  5 Alkoxide was prepared by the addition of BuLi at 0 • C for 10min, followed by changing the temperature as depicted. 6 Reaction was performed at reflux temperature. 7 3.0 equiv of HMPA was added.

Preparation of Allyl Vinyl Ethers 4 and One-Pot Isomerization-Claisen Rearrangement from 3
The requisite bisallylic ethers 3 in hand, we have at first undertaken the base-promoted isomerization of 3 to 4 ( Table 3). A 0.5 equiv of DBU was employed as a base because of its high potency for our previous system to achieve the proton migration of 1 to 2 [7]. As a result, it was observed that the desired isomerization of the bisallyl ether 3a proceeded in a very smooth fashion at room temperature to furnish the corresponding allyl vinyl ether 4a as a single stereoisomer (Entry 1). Stereochemistry of 4a was presumed to be Z on the basis of the fact that the exclusive formation of the (Z)-enol silyl ether by the action of lithium diisopropylamide (LDA) to the structurally similar propiophenone was due to the unfavorable sterically repulsive interaction between the methyl and phenyl groups in the corresponding (E)-isomer [16]. Because of the attachment of an aromatic group as the R 2 was essential for the realization of the ready isomerization of 1 to 2, all compounds 3 employed here possessed benzene-based R 2 and thus, all of the products 4 were considered to have the same (Z)-stereochemistry. Under the room temperature stirring for 4 days, [3,3]-sigmatropic rearrangement of (Z)-4a simultaneously proceeded and the desired product 5a was constructed in 16% yield with the 88% syn preference (please refer to the Section 2.4. for the explanation of the syn selectivity).
Ph Et i 54 2 (0) 1 Yields determined by 19 F NMR were shown in parentheses. 2 In these cases, about 40% of the substrates were recovered.

Preparation of Allyl Vinyl Ethers 4 and One-Pot Isomerization-Claisen Rearrangement from 3
The requisite bisallylic ethers 3 in hand, we have at first undertaken the base-promoted isomerization of 3 to 4 (Table 3). A 0.5 equiv of DBU was employed as a base because of its high potency for our previous system to achieve the proton migration of 1 to 2 [7]. As a result, it was observed that the desired isomerization of the bisallyl ether 3a proceeded in a very smooth fashion at room temperature to furnish the corresponding allyl vinyl ether 4a as a single stereoisomer (Entry 1). Stereochemistry of 4a was presumed to be Z on the basis of the fact that the exclusive formation of the (Z)-enol silyl ether by the action of lithium diisopropylamide (LDA) to the structurally similar propiophenone was due to the unfavorable sterically repulsive interaction between the methyl and phenyl groups in the corresponding (E)-isomer [16]. Because of the attachment of an aromatic group as the R 2 was essential for the realization of the ready isomerization of 1 to 2, all compounds 3 employed here possessed benzene-based R 2 and thus, all of the products 4 were considered to have the same (Z)-stereochemistry. Under the room temperature stirring for 4 days, [3,3]-sigmatropic rearrangement of (Z)-4a simultaneously proceeded and the desired product 5a was constructed in 16% yield with the 88% syn preference (please refer to the Section 2.4. for the explanation of the syn selectivity).   19 F NMR were shown in parentheses. 2 Diastereomer ratios determined by 19 F NMR was shown in brackets. 3 Reaction was conducted at room temperature. 4 Reaction was conducted at 40 °C. 5 Toluene was used as a solvent.

Isolated Yield 1 (%)
Because of the relatively slow reaction rate as described in Entry 1, attempt to raise the reaction temperature was conducted to find out that reflux in THF gave strong influence for the conversion of 3a which was significantly accelerated with recording almost the conversion of 3a which was significantly accelerated with recording almost quantitative combined yields of 4a and 5a (Entry 2). Solvent change to toluene provided 45 • C difference for the reflux temperature which seemed to suffice for the sequential isomerization-Claisen rearrangement to produce 91% of the desired product 5a as a 68:32 diastereomer mixture only in 3 h (Entries 3-4). Necessity of 15 h for 3b to complete the reaction was interpreted as the destabilization of the transition state by the electron-donating MeO group at the p-position of a phenyl moiety in R 1 (Entry 8). This is in sharp contrast to the case of the substrate 3c with a fluorine atom at the same position and 3 h reflux in toluene was enough to afford 91% yield of the product 5c (Entry 11). In both instances of 3b and 3c, 40 • C and room temperature reactions alternatively led to the formation of the allyl vinyl ethers 4b and 4c in moderate to good yields, respectively (Entries 6 and 10).
The similar substituent effect of R 1 was clearly understood by comparison of the results in Entries 12-23 with the others, and retardation of the present process was observed by incorporation of alkyl groups for R 1 but in a less effective manner than the case of R 2 which completely inhibited the elimination of a proton from C 1 and thus, no transformation of 1 to 2 was observed [7]. In the case of 3d with R 1 = Et, 48 h reflux was necessary for the direct transformation to 5d which was attained in 87% isolated yield but with the decreased diastereoselectivity to 55:45. Almost similar outcomes were recorded for the substrates 3e to 3g with a Ph(CH 2 ) 2 moiety as R 1 in terms of their chemical yields as well as the stereoselectivities irrespective of variation of the reaction periods.
For acquiring the relationship between the effect of the reaction temperature and time towards the diastereoselectivity, the isolated product 5a (syn:anti = 74:26) was submitted to the standard rearrangement conditions using toluene as a solvent (Scheme 2). Thus, reflux for 3 h demonstrated a slight decrease of the isomeric ratio to 68:32 which was further lowered to 66:34 after the prolonged reaction time to 24 h. On the other hand, the initial proportion was completely retained when 5a was stirred at room temperature, or under reflux without the base, DBU. This brief study proved that a weak base DBU was found to have an ability for epimerization of 5a at least in part, which was nicely explained by the experimental facts shown in Entries 1, 2, and 4 as well as 7 and 8 in Table 3: the lower the reaction temperature became, the better the diastereomer ratios were obtained. For the instances of the substrates 3 with an alkyl substituent as R 1 , the lower selectivity obtained was similarly understood as a consequence of the requirement of longer reaction times which would offer the higher chance of epimerization at the carbonyl α position in 5. for the direct transformation to 5d which was attained in 87% isolated yield but with the decreased diastereoselectivity to 55:45. Almost similar outcomes were recorded for the substrates 3e to 3g with a Ph(CH2)2 moiety as R 1 in terms of their chemical yields as well as the stereoselectivities irrespective of variation of the reaction periods.
For acquiring the relationship between the effect of the reaction temperature and time towards the diastereoselectivity, the isolated product 5a (syn:anti = 74:26) was submitted to the standard rearrangement conditions using toluene as a solvent (Scheme 2). Thus, reflux for 3 h demonstrated a slight decrease of the isomeric ratio to 68:32 which was further lowered to 66:34 after the prolonged reaction time to 24 h. On the other hand, the initial proportion was completely retained when 5a was stirred at room temperature, or under reflux without the base, DBU. This brief study proved that a weak base DBU was found to have an ability for epimerization of 5a at least in part, which was nicely explained by the experimental facts shown in Entries 1, 2, and 4 as well as 7 and 8 in Table 3: the lower the reaction temperature became, the better the diastereomer ratios were obtained. For the instances of the substrates 3 with an alkyl substituent as R 1 , the lower selectivity obtained was similarly understood as a consequence of the requirement of longer reaction times which would offer the higher chance of epimerization at the carbonyl α position in 5. Scheme 2. Effect of DBU for the epimerization of the rearranged product 5a.

Improvement of the Diastereoselectivity of the Claisen Rearrangement Products 5
Because it was our interpretation that the low diastereoselectivity of the rearranged products 5 were attributed to the requirement of the harsh conditions like refluxing in toluene at 111 °C, modification of the present system was planned by the addition of appropriate activators for the purpose of lowering the reaction temperature. As described in Table 4, 20 mol% of typical Lewis acids like TiCl4, BF3·OEt2, and AlCl3 were independently added to a DCM solution containing 4a at 0 °C to prove that cleavage of the allyl ether part attached to 1 as shown in Table 2 occurred as the main pathway to yield the ketone 2a along with a minor quantity of the requisite rearranged product 5a (Entries 1 to 3, Table  4). Further decrease of the reaction temperature in the case of TiCl4 was carried out with the expectation to inhibit this unfavorable route, but 2a was the sole product obtained even at −80 °C with complete suppression of the construction of 5a (Entry 4).

Scheme 2.
Effect of DBU for the epimerization of the rearranged product 5a.

Improvement of the Diastereoselectivity of the Claisen Rearrangement Products 5
Because it was our interpretation that the low diastereoselectivity of the rearranged products 5 were attributed to the requirement of the harsh conditions like refluxing in toluene at 111 • C, modification of the present system was planned by the addition of appropriate activators for the purpose of lowering the reaction temperature. As described in Table 4, 20 mol% of typical Lewis acids like TiCl 4 , BF 3 ·OEt 2 , and AlCl 3 were independently added to a DCM solution containing 4a at 0 • C to prove that cleavage of the allyl ether part attached to 1 as shown in Table 2 occurred as the main pathway to yield the ketone 2a along with a minor quantity of the requisite rearranged product 5a (Entries 1 to 3, Table 4). Further decrease of the reaction temperature in the case of TiCl 4 was carried out with the expectation to inhibit this unfavorable route, but 2a was the sole product obtained even at −80 • C with complete suppression of the construction of 5a (Entry 4).   19 F NMR were shown in parentheses. 2 Diastereomer ratios determined by 19 F NMR was shown in brackets. 3 Inseparable isomers 6 were observed by 19 F NMR (8% in Entry 9, 6% in Entry 10, and 4% in Entry 11).
At the next stage, application of [PdCl2·(PhCN)2] was implemented on the basis of our previously successful experience on this catalyst for the Ireland-Claisen rearrangement [9,19,20]: as a result, 10 mol% of this catalyst was found to be quite effective for excellent conversion of 4a to 5a even at room temperature. After brief examination of the solvent, toluene recorded smooth transformation without forming 2a and the desired product 5a was obtained in 70% isolated yield (Entries 7 to 9). The additional benefit was the increase of the diastereoselectivity from 68:32 (Entry 4, Table 3) to 95:5 (Entry 9) where, in line with our expectation, the lowering of the reaction temperature from 111 °C (toluene reflux) to 25 °C would play a significant role at least in part. This pertinent condition was also able to produce the better consequence for both substrates 4b (R 1 : p-MeOC6H4) and 4c (R 1 : p-FC6H4), attaining the same level of excellent selectivity in 81% and 65% isolated yields, respectively (Entries 10 and 11). One drawback of this Pd-catalyzed process is the formation of the byproducts 6 in a range of 4 to 8% [21] (Entries 9 to 11) which was not possible to be separated completely by silica gel column chromatography. As described in Scheme 3, subjection of the thermally rearranged product 5a as a 70:30 diastereomer mixture to this Pd-catalyzed conditions led to the formation of 10% of 6a (determined by 19 F NMR), which unambiguously proved that, at least in part, 6a was obtained as the result of isomerization of 5a.
1 Yields determined by 19 F NMR were shown in parentheses. 2 Diastereomer ratios determined by 19 F NMR was shown in brackets. 3 Inseparable isomers 6 were observed by 19 F NMR (8% in Entry 9, 6% in Entry 10, and 4% in Entry 11).
At the next stage, application of [PdCl 2 ·(PhCN) 2 ] was implemented on the basis of our previously successful experience on this catalyst for the Ireland-Claisen rearrangement [9,19,20]: as a result, 10 mol% of this catalyst was found to be quite effective for excellent conversion of 4a to 5a even at room temperature. After brief examination of the solvent, toluene recorded smooth transformation without forming 2a and the desired product 5a was obtained in 70% isolated yield (Entries 7 to 9). The additional benefit was the increase of the diastereoselectivity from 68:32 (Entry 4, Table 3) to 95:5 (Entry 9) where, in line with our expectation, the lowering of the reaction temperature from 111 • C (toluene reflux) to 25 • C would play a significant role at least in part. This pertinent condition was also able to produce the better consequence for both substrates 4b (R 1 : p-MeOC 6 H 4 ) and 4c (R 1 : p-FC 6 H 4 ), attaining the same level of excellent selectivity in 81% and 65% isolated yields, respectively (Entries 10 and 11). One drawback of this Pd-catalyzed process is the formation of the byproducts 6 in a range of 4 to 8% [21] (Entries 9 to 11) which was not possible to be separated completely by silica gel column chromatography. As described in Scheme 3, subjection of the thermally rearranged product 5a as a 70:30 diastereomer mixture to this Pd-catalyzed conditions led to the formation of 10% of 6a (determined by 19 F NMR), which unambiguously proved that, at least in part, 6a was obtained as the result of isomerization of 5a. les 2021, 26, 4365 8 of 19 Scheme 3. Control experiment for the olefinic isomerization of 5a.

Discussion on the Reaction Mechanism
First of all, on the diastereoselectivity of the present Claisen rearrangement, production of the syn-isomer was anticipated on the basis of our previous report as shown in Scheme 4 [9]. Thus, Michael addition of butyroyloxazolidinone-based enolate to allyl 4,4,4-trifluorobut-2-enoate, followed by the capture of the resultant enolate by TMSCl furnished the intermediary ketene silyl acetal Int-2 [22,23]. This intermediate Int-2 then experienced the Ireland-Claisen rearrangement with the aid of a catalytic amount of [PdCl2·(PhCN)2] to afford the rearranged product 7 in a highly stereoselective manner along with the unrearranged Michael adduct 8 in 63% and 32% yields, respectively. Stereochemistry of 7 was crystallographically confirmed as 2,3-anti,3,4-syn [24], the latter of which was conveniently explained by the application of the Cieplak rule [25]. Because incipient transition states (TS) in general are electron-deficient, reactions favorably occur from the face where the better stabilization is accomplished by the electron donation from the adjacent C-C bond (Figure 1). Suppose that the rearrangement occurred from the siface (by way of TS-si) in our previous instance, TS σ* ≠ should be better stabilized by electron-donation from the electronically richer σC-R1 orbital rather than the participation of the electron-deficient σC-CF3. This rule consistently elucidated the preferential formation of the syn-isomer as determined by X-ray crystallographic analysis. In the present case, because of the similar [3,3]-sigmatropic rearrangement of the structurally similar substrates with the R-C(CF3)H-branched structure at the same position as well as activation by the same catalyst, we believe that the major isomers anticipated for the present case was the syn isomers.

Discussion on the Reaction Mechanism
First of all, on the diastereoselectivity of the present Claisen rearrangement, production of the syn-isomer was anticipated on the basis of our previous report as shown in Scheme 4 [9]. Thus, Michael addition of butyroyloxazolidinone-based enolate to allyl 4,4,4trifluorobut-2-enoate, followed by the capture of the resultant enolate by TMSCl furnished the intermediary ketene silyl acetal Int-2 [22,23]. This intermediate Int-2 then experienced the Ireland-Claisen rearrangement with the aid of a catalytic amount of [PdCl 2 ·(PhCN) 2 ] to afford the rearranged product 7 in a highly stereoselective manner along with the unrearranged Michael adduct 8 in 63% and 32% yields, respectively. Stereochemistry of 7 was crystallographically confirmed as 2,3-anti,3,4-syn [24], the latter of which was conveniently explained by the application of the Cieplak rule [25]. Because incipient transition states (TS) in general are electron-deficient, reactions favorably occur from the face where the better stabilization is accomplished by the electron donation from the adjacent C-C bond (Figure 1). Suppose that the rearrangement occurred from the si-face (by way of TS-si) in our previous instance, TS σ* = should be better stabilized by electron-donation from the electronically richer σ C-R1 orbital rather than the participation of the electron-deficient σ C-CF3 . This rule consistently elucidated the preferential formation of the syn-isomer as determined by X-ray crystallographic analysis. In the present case, because of the similar [3,3]-sigmatropic rearrangement of the structurally similar substrates with the R-C(CF 3 )H-branched structure at the same position as well as activation by the same catalyst, we believe that the major isomers anticipated for the present case was the syn isomers.

Discussion on the Reaction Mechanism
First of all, on the diastereoselectivity of the present Claisen rearrangement, production of the syn-isomer was anticipated on the basis of our previous report as shown in Scheme 4 [9]. Thus, Michael addition of butyroyloxazolidinone-based enolate to allyl 4,4,4-trifluorobut-2-enoate, followed by the capture of the resultant enolate by TMSCl furnished the intermediary ketene silyl acetal Int-2 [22,23]. This intermediate Int-2 then experienced the Ireland-Claisen rearrangement with the aid of a catalytic amount of [PdCl2·(PhCN)2] to afford the rearranged product 7 in a highly stereoselective manner along with the unrearranged Michael adduct 8 in 63% and 32% yields, respectively. Stereochemistry of 7 was crystallographically confirmed as 2,3-anti,3,4-syn [24], the latter of which was conveniently explained by the application of the Cieplak rule [25]. Because incipient transition states (TS) in general are electron-deficient, reactions favorably occur from the face where the better stabilization is accomplished by the electron donation from the adjacent C-C bond (Figure 1). Suppose that the rearrangement occurred from the siface (by way of TS-si) in our previous instance, TS σ* ≠ should be better stabilized by electron-donation from the electronically richer σC-R1 orbital rather than the participation of the electron-deficient σC-CF3. This rule consistently elucidated the preferential formation of the syn-isomer as determined by X-ray crystallographic analysis. In the present case, because of the similar [3,3]-sigmatropic rearrangement of the structurally similar substrates with the R-C(CF3)H-branched structure at the same position as well as activation by the same catalyst, we believe that the major isomers anticipated for the present case was the syn isomers. In the case of the thermal rearrangement shown in Table 3, there are two distinct groups of 5a-c and 5d-g in terms of the diastereoselectivity obtained which was interpreted in the Section 2.2. as a result of the lower activating character by the electron-donating R 1 moieties in the latter group, thereby requiring a longer reaction time with possible higher chance of epimerization. Another explanation would be made from the standpoint of steric requirement: the smaller substituents for the latter group as R 1 (Et and PhCH2CH2 with the revised Taft Es values (Es') [26] of 0.08 and 0.35, respectively) than the case of the former (for example, 2.31 and 0.78 Es' values for Ph and CF3, respectively: the bigger numbers indicate the more steric bulkiness) would alleviate the unfavorable steric factors for the approach from the electronically less favorable re-face, thereby the diastereoselectivity was more or less lowered.

Conclusions
As depicted above, we have succeeded in the development of a new route of the Claisen rearrangement starting from the base-mediated isomerization of the bisallylic alcohols 3 to the allyl vinyl ethers 4, and the following formation of the rearranged products 5 were found to be realized in a one-pot manner when the reactions were conducted under toluene reflux conditions. An excellent alternative method was to treat 4 with such a catalyst as [PdCl2·(PhCN)2] in toluene, and in spite of contamination by a small amount of the isomerized products 6, the highest diastereomeric ratio of 95:5 was recorded which are well compared with the ones of 68:32 obtained by the direct thermal rearrangement from 3a. The weakly acidic HFIP was another choice for the present process which, in spite of requirement of 3 days for completion, the reaction nicely proceeded with recording good yield as well as stereoselectivity.

General Information
Unless otherwise noted, materials were obtained from commercial suppliers and were used without further purification. All manipulations involving air-sensitive materials were performed under argon. Anhydrous Et2O, THF and DCM were purchased and were used without further purification. In the case of the thermal rearrangement shown in Table 3, there are two distinct groups of 5a-c and 5d-g in terms of the diastereoselectivity obtained which was interpreted in the Section 2.2. as a result of the lower activating character by the electron-donating R 1 moieties in the latter group, thereby requiring a longer reaction time with possible higher chance of epimerization. Another explanation would be made from the standpoint of steric requirement: the smaller substituents for the latter group as R 1 (Et and PhCH 2 CH 2 with the revised Taft Es values (Es') [26] of 0.08 and 0.35, respectively) than the case of the former (for example, 2.31 and 0.78 Es' values for Ph and CF 3 , respectively: the bigger numbers indicate the more steric bulkiness) would alleviate the unfavorable steric factors for the approach from the electronically less favorable re-face, thereby the diastereoselectivity was more or less lowered.

Conclusions
As depicted above, we have succeeded in the development of a new route of the Claisen rearrangement starting from the base-mediated isomerization of the bisallylic alcohols 3 to the allyl vinyl ethers 4, and the following formation of the rearranged products 5 were found to be realized in a one-pot manner when the reactions were conducted under toluene reflux conditions. An excellent alternative method was to treat 4 with such a catalyst as [PdCl 2 ·(PhCN) 2 ] in toluene, and in spite of contamination by a small amount of the isomerized products 6, the highest diastereomeric ratio of 95:5 was recorded which are well compared with the ones of 68:32 obtained by the direct thermal rearrangement from 3a. The weakly acidic HFIP was another choice for the present process which, in spite of requirement of 3 days for completion, the reaction nicely proceeded with recording good yield as well as stereoselectivity.