Enantioselective 5-exo-Fluorocyclization of Ene-Oximes

The enantioselective 5-exo-fluorocyclization of ene-oxime compounds was demonstrated under phase-transfer catalysis. Although deprotonative fluorinations competed, the chemical yields and the ee values of the desired isoxazoline products were generally moderate to good. The absolute stereochemistry of the major isomer was determined to be S by comparison with the literature after transformation of the product to the corresponding iodinated isoxazoline.

Fluorine chemistry has contributed to pharmaceutical and agrochemical sciences, because an introduction of fluorine atom(s) at an appropriate position often improves the property of the parent compounds in terms of metabolic stability, lipophilicity, and so on [22][23][24][25][26][27]. Thus, a tremendous amount of fluorination reactions, including the asymmetric versions have been investigated [28][29][30]. Reflecting that alkene is a useful feedstock in organic chemistry and easy to prepare, asymmetric fluoro-functionalizations of alkenes have attracted increasing attention [31][32][33][34][35][36][37]. However, there is no report of the asymmetric fluorocyclization of ene-oximes, while its racemic version was recently reported in 2017 [38].
Our recent research interests include the asymmetric fluorofunctionalization of alkenes by phase-transfer catalysis. In 2015, we reported the first successful example of asymmetric fluorolactonization of ene-carboxylic acids using a hydroxymethyl carboxylate phase-transfer catalyst [39]. Based on this study, we recently developed a linked-binaphthyl dicarboxylic acid precatalyst 1, which was proven to be highly effective for the asymmetric fluorocyclization and the deprotonative fluorination of allylic amides (Scheme 1a) [40,41]. In these reactions, hydrogen bonding between the catalyst and the substrate was considered to be crucial for high asymmetric induction.
Considering pKa values of amide and oxime, we anticipated that oxime could interact with our anionic phase-transfer catalyst through hydrogen bonds, which would define the conformation of the substrate and/or a fluorinated carbocation intermediate. Our previous study suggested that the fluorocyclization Scheme 1. Fluorofunctionalizations of alkenes with linked-binaphthyl dicarboxylic acid 1 and Selectfluor: (a) Asymmetric 6-endo-type fluorocyclization of allylic amides; (b) Asymmetric 5-exo-fluorocyclization of ene-oximes.

Results and Discussion
Ene-oxime 2a [21] was chosen as a test substrate to optimize the reaction conditions (Table 1). At first, the reaction was carried out with 1 under the previous cyclization conditions [40] and the desired product, 3a, was obtained with 58% ee (entry 1). In this reaction, undesired byproducts were simultaneously observed by 1 H NMR analysis of the crude mixture. Although these byproducts could not be purified at this point, 1 H and 19 F NMR analyses suggested that the byproducts were deprotonative fluorination products 4a-6a. The enantioselectivity observed in chlorobenzene and benzene was almost the same with that in toluene, but the reaction rate became somewhat slower (entries 1-3). Use of CH2Cl2 and THF resulted in low chemical yield (entries 4 and 5). While Na2SO4 had a positive effect in improving the chemical yield in our previous case [40,41], it did not affect the reaction efficiency in the present reaction (entries 1 and 6). Among bases tested, Na3PO4 was found to be the base of choice in terms of the chemical yield of the desired product 3a (entries 6-11). The enantioselectivity was almost similar (59%-61%), irrespective of the basicity and counter cation. Proton sponge provided a low yield and a low enantioselectivity under the described conditions (entry 11). As seen in entry 12, the reaction rate became slower at 15 °C, but a better enantioselectivity was observed (69% ee). In all cases, the yields based on the recovered starting material were around 65%. Interestingly, the ee of 6a was as high as 77% under the optimized conditions (entry 12), albeit in only 4% yield. It should be noted that phosphoric acid 7, which is a commonly used precursor of anionic phase transfer catalysts [33], did not promote the present reaction (entry 14), confirming the better performance of our dicarboxylate catalyst.

Results and Discussion
Ene-oxime 2a [21] was chosen as a test substrate to optimize the reaction conditions (Table 1). At first, the reaction was carried out with 1 under the previous cyclization conditions [40] and the desired product, 3a, was obtained with 58% ee (entry 1). In this reaction, undesired byproducts were simultaneously observed by 1 H NMR analysis of the crude mixture. Although these byproducts could not be purified at this point, 1 H and 19 F NMR analyses suggested that the byproducts were deprotonative fluorination products 4a-6a. The enantioselectivity observed in chlorobenzene and benzene was almost the same with that in toluene, but the reaction rate became somewhat slower (entries 1-3). Use of CH 2 Cl 2 and THF resulted in low chemical yield (entries 4 and 5). While Na 2 SO 4 had a positive effect in improving the chemical yield in our previous case [40,41], it did not affect the reaction efficiency in the present reaction (entries 1 and 6). Among bases tested, Na 3 PO 4 was found to be the base of choice in terms of the chemical yield of the desired product 3a (entries 6-11). The enantioselectivity was almost similar (59%-61%), irrespective of the basicity and counter cation. Proton sponge provided a low yield and a low enantioselectivity under the described conditions (entry 11). As seen in entry 12, the reaction rate became slower at 15 • C, but a better enantioselectivity was observed (69% ee). In all cases, the yields based on the recovered starting material were around 65%. Interestingly, the ee of 6a was as high as 77% under the optimized conditions (entry 12), albeit in only 4% yield. It should be noted that phosphoric acid 7, which is a commonly used precursor of anionic phase transfer catalysts [33], did not promote the present reaction (entry 14), confirming the better performance of our dicarboxylate catalyst.  1 toluene proton sponge 17 7 12 5 1 toluene Na3PO4 63 (58) 6 69 13 5 1 toluene Na2CO3 62 (58) 6 65 14 5 7 toluene Na3PO4 2 - 1 The reactions were carried out with 2a (0.1 mmol), precatalyst (10 mol %), Selectfluor (1.5 equiv), and base (1.5 equiv) at 25 °C, unless otherwise mentioned. 2 The yields were determined by 1 H NMR analysis using 1,1,2,2-tetrabromoethane as an internal standard. 3 The ee values were determined by HPLC analysis using a chiral stationary column. 4 Run with Na2SO4. 5 Run at 15 °C for 72 h. 6 Isolated yield.
To determine the absolute stereochemistry of the major isomer, the fluorinated isoxazoline 3a was transformed to the corresponding iodinated isoxazoline 8a with MgI2 using a sealed tube at 80 °C [42] (Scheme 2). Although the conversion was modest, 8a could be obtained without erosion of the ee value. The stereochemistry of the major isomer was determined to be S by comparing the retention time of HPLC analysis with that reported in the literature [21]. This result indicates that the major isomer of the present fluorocyclization is S. To determine the absolute stereochemistry of the major isomer, the fluorinated isoxazoline 3a was transformed to the corresponding iodinated isoxazoline 8a with MgI 2 using a sealed tube at 80 • C [42] (Scheme 2). Although the conversion was modest, 8a could be obtained without erosion of the ee value. The stereochemistry of the major isomer was determined to be S by comparing the retention time of HPLC analysis with that reported in the literature [21]. This result indicates that the major isomer of the present fluorocyclization is S. Having optimized the reaction conditions, other ene-oximes were transformed to the corresponding fluorinated isoxazolines (Figure 1). Reactions of para-and meta-methylated substrates proceeded to give the corresponding isoxazoline products with good ee values (3b, 3c). However, ortho-substituent retarded the reaction completely, probably due to the steric repulsion (3d). meta-Chlorinated and para-fluorinated isoxazolines were formed with 77% ee and 73% ee, respectively (3e and 3f). An ene-oxime bearing a cyclohexyl group was less reactive for the fluorocyclization and 3g was obtained in only 12% yield, even at room temperature. The chemical structure on the oxime side did not have a significant impact on the enantioselectivity (3h and 3i). To our delight, a substrate having a thiophen group provided 3j in 69% with 84% ee. We additionally performed the SDE (self-disproportionation of enantiomer) test by achiral column chromatography [43]. The difference between the first and the last fractions was less than 1% ee, suggesting that SDE did not occur. As usual, all fractions were collected after the chromatographic purification. Having optimized the reaction conditions, other ene-oximes were transformed to the corresponding fluorinated isoxazolines (Figure 1). Reactions of paraand meta-methylated substrates proceeded to give the corresponding isoxazoline products with good ee values (3b, 3c). However, ortho-substituent retarded the reaction completely, probably due to the steric repulsion (3d). meta-Chlorinated and para-fluorinated isoxazolines were formed with 77% ee and 73% ee, respectively (3e and 3f). An ene-oxime bearing a cyclohexyl group was less reactive for the fluorocyclization and 3g was obtained in only 12% yield, even at room temperature. The chemical structure on the oxime side did not have a significant impact on the enantioselectivity (3h and 3i). To our delight, a substrate having a thiophen group provided 3j in 69% with 84% ee. We additionally performed the SDE (self-disproportionation of enantiomer) test by achiral column chromatography [43]. The difference between the first and the last fractions was less than 1% ee, suggesting that SDE did not occur. As usual, all fractions were collected after the chromatographic purification.  To confirm the importance of hydrogen bond interaction between the oxime and the dianionic catalyst 1, the following control experiment was carried out. Thus, when O-methylated compound 9 was subjected to the described reaction conditions, no reaction occurred at all and the starting material 9 was just recovered (Scheme 3). Since even deprotonated products were not formed, it is likely that the hydrogen bond interaction has an important role in accelerating the fluorination step, in addition to the enantioselectivity control. To confirm the importance of hydrogen bond interaction between the oxime and the dianionic catalyst 1, the following control experiment was carried out. Thus, when O-methylated compound 9 was subjected to the described reaction conditions, no reaction occurred at all and the starting material 9 was just recovered (Scheme 3). Since even deprotonated products were not formed, it is likely that the hydrogen bond interaction has an important role in accelerating the fluorination step, in addition to the enantioselectivity control.

General Information
1 H and 19 F NMR spectra were measured on a JEOL ECX-500 spectrometer at 500 and 470 MHz, respectively. 13 C NMR spectra were recorded on a JEOL JNM-ECX-500 spectrometer at 125 MHz. Chemical shifts were reported in parts per million (ppm) downfield from TMS (δ = 0) for 1 H NMR. For 13 C NMR, chemical shifts were reported in the scale relative to CDCl3. For 19 F NMR, chemical shifts were reported in a scale relative to CFCl3 external standard (δ = 0 ppm). Column chromatography was performed with silica gel N-60 (40-100 μm) purchased from Kanto Chemical Co., Inc. TLC analysis was performed on Silica gel 60 F254-coated glass plates (Merck). Visualization was accomplished by means of ultraviolet (UV) irradiation at 254 nm or by spraying an ethanol solution of 12-molybdo(VI)phosphoric acid as a developing agent.
Dehydrated dichloromethane (CH2Cl2) and toluene were purchased from Wako Pure Chemical Industries, Ltd. Dehydrated tetrahydrofuran (THF), diethyl ether (Et2O), and benzene were purchased from Kanto Chemical Co., Inc. Selectfluor was purchased from Aldrich. Other reagents were purified by usual methods.
The substrates were synthesized according to the literature [21]. The catalysts were synthesized according to the literature [40]. The 1 H, 13 C and 19 F NMR spectra and HPLC data of compounds 3 are available in the Supplementary Material.

General Information
1 H and 19 F NMR spectra were measured on a JEOL ECX-500 spectrometer at 500 and 470 MHz, respectively. 13 C NMR spectra were recorded on a JEOL JNM-ECX-500 spectrometer at 125 MHz. Chemical shifts were reported in parts per million (ppm) downfield from TMS (δ = 0) for 1 H NMR. For 13 C NMR, chemical shifts were reported in the scale relative to CDCl 3 . For 19 F NMR, chemical shifts were reported in a scale relative to CFCl 3 external standard (δ = 0 ppm). Column chromatography was performed with silica gel N-60 (40-100 µm) purchased from Kanto Chemical Co., Inc. TLC analysis was performed on Silica gel 60 F 254 -coated glass plates (Merck). Visualization was accomplished by means of ultraviolet (UV) irradiation at 254 nm or by spraying an ethanol solution of 12-molybdo(VI)phosphoric acid as a developing agent.
Dehydrated dichloromethane (CH 2 Cl 2 ) and toluene were purchased from Wako Pure Chemical Industries, Ltd. Dehydrated tetrahydrofuran (THF), diethyl ether (Et 2 O), and benzene were purchased from Kanto Chemical Co., Inc. Selectfluor was purchased from Aldrich. Other reagents were purified by usual methods.
The substrates were synthesized according to the literature [21]. The catalysts were synthesized according to the literature [40]. The 1 H, 13 C and 19 F NMR spectra and HPLC data of compounds 3 are available in the Supplementary Material.

Conclusions
In this paper, we have demonstrated the enantioselective 5-exo-fluorocyclization of ene-oximes using the linked-binaphthyl dicarboxylic acid precatalyst 1. The corresponding fluorinated isoxazolines were obtained with up to 84% ee, and the stereochemistry of the major isomers were determined to be S after transformation to known isoxazoline 8a with an iodomethyl unit. A control experiment revealed that hydrogen bond interaction of the oxime group is extremely important for the reaction acceleration and the enantioselectivity control. Further applications of the present fluorinating system are underway in our laboratory.