51.22 F
London
April 18, 2024
PI Global Investments
Silver

Silver-catalyzed direct conversion of epoxides into cyclopropanes using N-triftosylhydrazones


Optimization of the reaction conditions

At the outset, our investigations focused on the reaction of (2,3-epoxypropyl)-benzene (1a) and trifluoromethyl phenyl N-triftosylhydrazone (2a, 2.0 equiv.). A reaction system composed of catalytic TpBr3Ag(thf) and NaH (2.5 equiv.) at 60 °C in dichloromethane (DCM) provided the desired cyclopropane 3 in 58% yield and 10:1 diastereomeric ratio (d.r.) along with 90% deoxygenation product, trifluoroacetophenone 3’ (Table 1, entry 1), without the formation of competing C–O or C–H bond insertion products35,42,43,44,45,46,47,48. Control experiments revealed that the amounts of 2a and NaH were crucial to improving the efficiency, and the best result (92% isolated yield of 3 and quantitative trifluoroacetophenone 3’) was indeed obtained with 4.0 equiv. of 2a and 4.5 equiv. of NaH (entries 1, 2 vs 3). Other silver catalysts, such as AgOTf, AgOAc, Tp(CF3)2Ag(thf), and Tp*Ag, resulted in diminished yields and diastereoselectivity (entries 4–7). This unusual reactivity of TpBr3Ag(thf) is mainly due to the presence of a substantial steric hindrance scorpion ligand (TpBr3), which not only hinders the occurrence of a competitive carbene dimerization process but also enhances non-enantioselectivity. To our surprise, no conversion of 1a was observed when using TpBr3Cu(MeCN), Pd(OAc)2, and FeTPPCl as catalysts (entries 8–10). The use of Rh2(OAc)4 as catalysts results in modest efficiency in the deoxidation process, thus obtaining 3 in poor yield and selectivity (entry 11). The reaction yield and stereoselectivity could not be improved by decreasing (40 °C) or increasing (80 °C) the reaction temperature (entries 12 and 13). The substituents and positions on N-sulfonylhydrazones were also critical, as lower yields were observed with 4-methylphenyl- and 2,4,6-triisopropylphenyl N-sulfonylhydrazones (2b and 2c) (entries 14 and 15). Note that the easy decomposition of o-trifluoromethyl N-triftosylhydrazones compared to other functional groups substituted N-sulfonylhydrazones enabled the rapid deoxidation process35.

Table 1 Optimization of the reaction conditionsaView full size image

Substrate scope

With the optimized conditions in hand, we explored the scope and limitations of this transformation with a range of epoxides and N-triftosylhydrazone (Fig. 2). We first probed the reactivity of various epoxides for this oxygen-carbon exchange protocol using 4-chlorophenyl trifluoromethyl N-triftosylhydrazone as a carbon component (Fig. 2a). Mono-substituted epoxides with methyl, n-butyl, n-decyl, benzyl, and 2-phenethyl groups readily transformed into corresponding cyclopropane products 48 with good to excellent yields and good diastereomeric ratio (d.r.). Dialkyl ether and alkylaryl ether-tethered epoxides were tolerated well, giving rise to cyclopropane products 912 in good yields and diastereoselectivities (>20:1 d.r.), except for product 9 (7:1 d.r.). Pleasingly, functional groups such as ester, chloro, and tosylate groups at the terminus of alkyl epoxides underwent smooth reaction, affording cyclopropane products 1315 in good yields and diastereoselectivities. 3-Phenyl-2,2′-bisepoxide was transformed into biscyclopropane 16 with high diastereoselectivity albeit in lower yield under modified conditions for 24 h.

Fig. 2: Scope of single-atom transmutation of epoxides with fluoroalkyl N-triftosylhydrazones.
figure 2

a Scope of epoxides. b Scope of N-triftosylhydrazones. c Gram-scale synthesis. Reaction conditions: a1 (0.15 mmol, 1.0 equiv.), 2 (0.6 mmol, 4.0 equiv.), NaH (0.68 mmol, 4.5 equiv.) and TpBr3Ag(thf) (10 mol%) in DCM (4.0 mL) at 60 °C for 6–12 h. bExtend reaction time to 24 h. cUsing 6.0 equiv. of 2. d1 (0.15 mmol, 1.0 equiv.), 2 (0.38 mmol, 2.5 equiv.), NaH (0.38 mmol, 2.5 equiv.) and TpBr3Ag(thf) (2.5–5 mol%) in DCM (4.0 mL) at 60 °C for 0.5–6 h. e1 (0.15 mmol, 1.0 equiv.), 2 (0.45 mmol, 3.0 equiv.), NaH (0.45 mmol, 3.0 equiv.) and TpBr3Ag(thf) (5 mol%) in DCM (4.0 mL) at 60 °C for 10 h. f1b (1.2 g, 10 mmol), 2a (9.9 g, 25 mmol), NaH (1.0 g, 25 mmol) and TpBr3Ag(thf) (219 mg, 2 mol%) in DCM (80.0 mL) at 60 °C for 2 h. Yields of the isolated product are given. Diastereomeric ratio (d.r.) determination by relative integration of 19F NMR spectra. PCP, 4-chlorophenyl. PMP, 4-methoxyphenyl.

This catalytic protocol also operates effectively with 2-(phenylethynyl)oxirane to furnish cyclopropane 17 (54% yield and >20:1 d.r.), where the internal alkynyl remains intact for [2 + 1] cycloaddition with carbenes. Under standard conditions, 1,2,7,8-diepoxyoctane readily converted to n-hexene-substituted cyclopropane 18 (80% yield and >12:1 d.r.) but produced the expected 1,4-dicyclopropylbutane 19 (74% yield and 8:1 d.r.) with 6 equiv. of 4-chlorophenyl trifluoromethyl N-triftosylhydrazone. Next, an array of aryl-substituted oxiranes was subjected to the reaction system. Aryl epoxides with various substitution groups (e.g., methyl, phenyl, chloro, methoxy, and trifluoromethyl) and electronic properties were well tolerated with this transformation, affording cyclopropanes 2028 in high yields (62–93%) and excellent diastereoselectivities (>20:1 d.r.). Similarly, disubstituted phenyl-, naphthyl-, and 2-thienyl-substituted epoxides were converted into cyclopropanes 2931 in good yield with >20:1 d.r. The scope of various disubstituted oxiranes was then examined. Several 2,2-disubstituted epoxides, including 2,2-dimethyl, sterically hindered 2,2-diaryl, and 2-(cyclo)alkyl-2-aryl-substituted epoxides, as well as cyclohexyl ketal, pyran, N-Cbz protected piperidine fused epoxides proved compatible, leading to highly substituted cyclopropanes 3240 in high yields with d.r. as high as 16:1. Similarly, various 2,3-disubstituted epoxides (e.g., cyclopentane, 2,3-dihydro indene oxirane, 2,3-dialkyl-, 2-alkyl-3-aryl-, and 2,3-diaryl-substituted epoxides) afforded the corresponding cyclopropanes 4145 with good to excellent yields and moderate to good d.r.

Finally, we explored the scope of fluoroalkyl N-triftosylhydrazones with the representative styrene oxide or 4-methoxystyrene oxide (Fig. 2b). Under optimized conditions, both electron-rich and electron-poor aryl trifluoromethyl N-triftosylhydrazones were well tolerated with this transformation, furnishing cyclopropane products 4655 in high yields and excellent d.r. Similarly, disubstituted phenyl-, naphthyl-, and furyl-substituted trifluoromethyl N-triftosylhydrazones reacted smoothly to afford corresponding cyclopropanes 5658 in 53–78% yield and >20:1 d.r. N-Triftosylhydrazones derived from trifluoromethyl vinyl/1,3-dienyl ketones were also found to be suitable carbene precursors, affording the corresponding alkenyl/dienyl substituted cyclopropanes 5962 in high yields (75–89%) with excellent diastereoselectivity. We also attempted to use alkyl N-triftosylhydrazones as carbene precursors (For details, please see Supplementary Fig. 1b), but no cyclopropane product was observed owing to the in situ generated alkyl trifluoromethyl silver carbene more prone to undergo β-hydrogen elimination to yield trifluoromethyl substituted olefins. Even methyl trifluoromethyl N-triftosylhydrazone also failed to yield the cyclopropanization product, likely due to the low reactivity of alkyl trifluoromethyl silver carbene. Interestingly, N-triftosylhydrazone derived from pentafluoroethyl phenyl ketone was successfully converted into pentafluoroethyl cyclopropane 63 in 74% yield with a d.r. of >20:1, demonstrating the capability of the present work in divergent synthesis. This protocol also successfully extended to N-triftosylhydrazones derived from ethyl phenylglyoxylate with good yield and excellent selectivity. Aside from trifluoromethyl ketones derived N-triftosylhydrazones, those derived from aromatic aldehydes performed well to afford desired 1,2-diaryl cyclopropanes 6567 in moderate diastereoselectivities. To demonstrate the robustness and scalability of this protocol, we performed a 10.0 mmol scale transformation to prepare cyclopropane 46 without lowering the yield or diastereoselectivity of the reaction (Fig. 2c). The relative configuration of 46 was confirmed with X-ray crystallography.

Given the importance of trifluoromethylcyclopropane motifs in bioactive molecules, epoxides derived from natural products and drugs were synthesized to showcase the versatility of the protocol (Fig. 3). Under standard conditions, epoxides derived from cocal, fenbufen, zingerone, nabumetone, and fenofibrate readily transformed into corresponding trifluoromethylcyclopropane products 6872 in moderate to high yield and diastereoselectivities. Similarly, the trifluoromethylcyclopropane functionality could be successfully installed onto the sulfonamide anti-gout drug probenecid 73 (63% yield and 17:1 d.r.). We attempted to use allose-derived epoxides as substrates for the reaction because the steric hindrance of the trisubstituted olefin intermediates prevented further conversion of cyclopropane. The derivatives of scopolamine, due to their alkalinity, result in catalyst deactivation and failure to obtain the corresponding cyclopropane product (For details, please see Supplementary Fig. 1b). The anti-gout drug probenecid was used to treat severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) infection49,50. Moreover, N-triftosylhydrazones derived from bioactive natural products (e.g., citronellol, cholesterol, and DL-menthol) and nonsteroidal anti-inflammatory drugs (e.g., indometacin and ibuprofen) effectively underwent this oxygen-carbon exchange reaction to afford cyclopropane products 7478 in high yields with excellent stereocontrol.

Fig. 3: Late-stage skeletal editing to access trifluoromethylcyclopropane containing quaternary carbon centers.
figure 3

a, b Scope of epoxides derived from natural products and drugs. c Scope of N-triftosylhydrazones derived from natural products and drugs. Reaction conditions: a1 (0.15 mmol, 1.0 equiv.), 2d (0.6 mmol, 4.0 equiv.), NaH (0.68 mmol, 4.5 equiv.) and TpBr3Ag(thf) (10 mol%) in DCM (4.0 mL) at 60 °C for 2–10 h. b1b (0.15 mmol, 1.0 equiv.), 2 (0.38 mmol, 2.5 equiv.), NaH (0.38 mmol, 2.5 equiv.) and TpBr3Ag(thf) (5 mol%) in DCM (4.0 mL) at 60 °C for 2 h. Yields of the isolated product are given. Diastereomeric ratio (d.r.) determination by relative integration of 19F NMR spectra.

Mechanistic investigations

We next conducted control experiments to gain some insights into the reaction mechanism (Fig. 4a). First, subjecting (2,3-epoxypropyl)-benzene or styrene oxide to standard conditions without N-triftosylhydrazones leads to no reaction and recovered epoxides in quantitative yields (Fig. 4a, eq. i), indicating the high stability of epoxides in basic conditions and N-triftosylhydrazone is necessary for deoxygenation process. However, the model reaction in the presence of N-triftosylhydrazones, both (2,3-epoxypropyl)-benzene and styrene oxide disappeared within 5 mins and produced olefin intermediate and cyclopropane, suggesting that the in situ-generated silver carbene may induce this sequential deoxygenation and cycloaddition process. When the reaction lasted for 30 mins, (2,3-epoxypropyl)-benzene produced allylbenzene (30%) along with the expected cyclopropane product 3 (68%), whereas styrene oxide gave the cyclopropane product 46 in 78% yield (Fig. 4a, eq. ii and iii), owing to the higher reactivity of aryl olefins toward cyclopropanation than alkyl olefins. After 6 hours of reaction, allylbenzene was completely consumed and converted to cyclopropane product 3 (92%, d.r. = 10:1) (Fig. 4a, eq. ii). Based on the control experiments, we proposed a reasonable deoxygenation and cycloaddition mechanism in Fig. 4b.

Fig. 4: Mechanism investigation.
figure 4

a Control experiments. b Proposed mechanism. c Computed fragmentation/[2 + 1] cycloaddition mechanism. Energies are given in kcal mol−1 and most H atoms in 3D structures are omitted for clarity.

To gain a better understanding of the reaction mechanism and the specific interactions responsible for the experimentally observed stereoselectivity, we performed systematic density functional theory (DFT) calculations for the reaction between trifluoromethyl phenyl diazomethane 2a’ and (2,3-epoxypropyl)-benzene at the B3LYP/6-31 G(d,p)-GD3(BJ)-SDD (Ag,Br) level using PCM (CH2Cl2) solvent model. The calculations indicated that the most favorable pathway was net oxygen-carbon exchange via cycloaddition of the in situ generated olefin from deoxygenation of epoxides with excess silver carbene. Figure 4c depicts the calculated free energy profiles for the catalytic cycle. First, the dinitrogen extrusion from 2a’ occurs in the presence of an Ag catalyst via the transition state TS1’ (∆G≠ = 8.6 kcal mol−1) to form an electrophilic silver carbene Int1 (For details, please see Supplementary Fig. 2a). Then the Int1 either binds to (2,3-epoxypropyl)-benzene 1a to form Int2 (TS1, ∆G≠ = 1.3 kcal mol−1) or reacts with excess 2a’ to provide the Pr’ via TS4 (∆G≠ = 4.7 kcal mol−1) (For details, please see Supplementary Fig. 2b), which is consistent with the experimental results. However, the calculations suggest that the formation of Int2 is preferable due to the low energy barrier required. As a result, Int2 readily undergoes intramolecular ring-opening isomerization via TS2 (∆G≠ = 2.9 kcal mol−1) to form trifluoromethyl acetophenone 3a and allylbenzene Int3. Finally, the concerted [2 + 1] cycloaddition of Int3 with the excess silver carbene Int1 yields the cyclopropane product 3-syn (∆G≠ = 4.7 kcal mol−1) and 3-anti (∆G≠ = 1.0 kcal mol−1). Note that competitive allylic C–H bond insertion might occur to some extent due to the presence of allylic C–H bond51. Our calculations demonstrated that the C–H bond insertion reaction at the allylic position occurs via the three-membered ring transition state formed by Int1 and Int3 (TS5, ∆G≠ = 3.7 kcal mol−1). When compared to the C–H insertion process, the [2 + 1] cycloaddition is both thermodynamically and kinetically favored and thus produces cyclopropane products 3-anti and 3-syn.

The cyclopropane (3anti and 3syn) formation transition states TS3 and TS3’ are shown in Fig. 5. According to our calculations, the number of C–H—F hydrogen bonding interactions in TS3’ is more than that of TS3, resulting in a sterically more hindered configuration of TS3’ (Fig. 5). The color-filled reduced density gradient (RDG) analysis52,53 for the cyclopropanation process demonstrated that TS3 has a stronger non-covalent interaction of mutual attraction, while a stronger steric hindrance exists in TS3’. Thus, the formation of 3-anti is highly favorable over 3syn.

Fig. 5: The number of C–H—F hydrogen bonding interactions (left).
figure 5

color-filled reduced density gradient (RDG) isosurface for TS3 and TS3’ (isovalue set to 0.5): (blue) areas of attraction (covalent bonding); (green) vdW interaction; (red) areas of repulsion (steric and ring effects) (right).



Source link

Related posts

Outcrop Silver Announces Marketing Service Agreement – Company Announcement

D.William

Newrange Gold Transforms into Pinnacle Silver and Gold

D.William

East of England Co-Op in Silver End raided by armed crook demanding staff to open till

D.William

Leave a Comment

* By using this form you agree with the storage and handling of your data by this website.