Nickel‐Catalyzed Monofluoromethylation of Aryl Boronic Acids

Abstract

Aryl boronic acids can be monofluoromethylated under nickel catalysis. The utility of this method is demonstrated by the monofluoromethylation of a borylated and acyl-protected derivative of the statin drug ezetimibe. Mechanistic investigations indicate that a fluoromethyl radical is involved in the NiI/NiIII catalytic cycle. The incorporation of fluorine atoms into small organic molecules can often drastically enhance the metabolic stability, lipophilicity, and bioavailability, and can also increase the receptor-binding affinity and the selectivity relative to the parent molecule.1 As a result, fluorine-containing compounds have been widely used in pharmaceutical and agrochemical products.2 Recently, the selective introduction of monofluoromethyl groups into small organic molecules has emerged as a new strategy in drug design. Notably, many biologically active molecules, including the widely prescribed afloqualone, fluticasone propionate, and the anaesthetic sevoflurane, contain a CH2F group as an essential motif.3 Although a variety of methods for the transition-metal-catalyzed trifluoromethylation4 and difluoromethylation3b, 5 of arenes have been developed in the past several decades, the incorporation of methyl groups containing a single fluorine (CH2F) into arenes has been studied to a lesser extent and remains a challenge.6 The only example of a palladium-mediated direct monofluoromethylation of pinacol phenylboronate was reported by the Suzuki group in 2009.7 Therein, a stoichiometric amount of palladium and a large excess of the boronic ester (40 equiv) were required, and the yield was modest (57 %, Scheme 1). More recently, using fluoromethyl 2-pyridyl sulfone reagents, Hu and co-workers reported a copper-mediated monofluoromethylsulfonylation of aryl iodides.8a,8b Herein, the requirement of the coordinating 2-pyridylsulfone group prohibits access to other stabilized monofluoromethylating groups. As in the report from Suzuki and co-workers, this reaction is also hampered by the high loading of a transition metal (0.3–2.0 equiv of copper), and also suffers from operational inconvenience imparted by the requisite stepwise addition of reagents. Herein, we report the first example of a nickel-catalyzed monofluoromethylation of aryl boronic acids, wherein the fluoromethylating reagents bear either a phenylsulfone or an ethoxycarbonyl moiety.9–11 Transition-metal-promoted monofluoromethylation. We commenced our study with phenylboronic acid (1 a) as the pilot substrate and PhSO2CFHI8 (2) as the coupling partner in the presence of a catalytic amount of [Ni(acac)2] (5 mol %) in dichloromethane at 100 °C. To our delight, the desired fluoro(phenylsulfonyl)methylated product 3 a was obtained in 45 % yield when XantPhos (10 mol %) was used as the ligand (Table 1, entry 2). Furthermore, a careful survey of bases (entries 3–6) and solvents (see Table S2 in the Supporting Information) was then performed, which showed the combination of K2CO3 and dicholoromethane to be optimal. To improve the yield further, a variety of phosphine and diamine ligands were examined next. In contrast to previous reports on nickel-catalyzed reactions, in which diamine ligands are the most effective, phosphine ligands, including diphosphines (BINAP, dppe and dppf) and monophosphines (PPh3), afforded the desired fluoromethylated product in up to 87 % yield of isolated product (entry 20). Meanwhile, no improvement was found when the reaction temperature was changed (entry 21). Additionally, the investigation of aryl boron reagents showed that triphenylboroxine, usually in equilibrium with 1 a, gave the isolated products in slightly lower yield (75 %, entry 23), while phenylboronic ester and trifluoroborate afforded no desired product (see Table S4 in the Supporting Information). Last, the control experiment indicated that none of the fluoromethylated product was obtained without nickel catalyst (entry 24). Entry Ligand Base Yield [%][b] 1 XantPhos No 0 2 XantPhos Cs2CO3 45 3 XantPhos K2CO3 48 4 XantPhos KOtBu 37 5 XantPhos KOAc 0 6 XantPhos DABCO 0 7[c] XantPhos K2CO3 7 8[d] XantPhos K2CO3 0 9[e] XantPhos K2CO3 28 10 XPhos K2CO3 20 11 SPhos K2CO3 0 12 BINAP K2CO3 81 13 dppf K2CO3 79 14 dppe K2CO3 89 15 P(o-tol)3 K2CO3 11 16 P(2-furyl)3 K2CO3 31 17 bpy K2CO3 0 18 TMEDA K2CO3 27 19 DMEDA K2CO3 41 20 PPh3 K2CO3 97 (87[f]) 21[g] PPh3 K2CO3 70 22[h] PPh3 K2CO3 53 23[i] PPh3 K2CO3 75[f] 24 PPh3 K2CO3 0[j] With the optimized conditions in hand, we next investigated the scope of the procedure by testing various aryl boronic acids in reaction with 2. As shown in Scheme 2, para-, meta-, and even ortho-substituted aryl boronic acids are all effective cross-coupling partners, affording the corresponding fluoromethylated arenes in good yields. Aryl boronic acids substituted with electron-donating groups are fluoromethylated smoothly to give the desired product with good yields (3 b–3 i). A range of halogenated boronic acids are also well-tolerated by this method (3 j–3 m). The presence of halo substituents in the products offers the potential for further synthetic elaboration through metal-catalyzed coupling reactions. In particular, aryl boronic acids containing electron-withdrawing groups, including trifluoromethyl (3 n–3 p), cyano (3 q), ester (3 r), and nitro groups (3 s), were also fluoroalkylated with satisfactory yields in our catalytic system. Although a higher catalyst and ligand loading is typically required, these arenes are usually ineffective partners in nickel-catalyzed cross-coupling reactions.10a, 11 Additionally, boronic acids derived from polycyclic and heterocyclic arenes, such as p-biphenyl (3 t), 2-naphthyl (3 u), thiophene (3 v–3 w), dibenzothiophene (3 x), and dibenzofuran (3 y), are also compatible with the reaction conditions. Finally, even a tyrosine derivative and a flavanone derivative may be monofluoromethylated, leading to 3 z and 3 za, respectively, both in satisfactory yields. Scope of substituted phenylboronic acid substrates. Reaction conditions (unless otherwise noted): PhSO2CFHI (0.2 mmol, 1.0 equiv), 1 (2.0 equiv), [Ni(acac)2] (5 mol %), PPh3 (10 mol %), K2CO3 (1.5 equiv), CH2Cl2, 100 °C, 24 h. Yields of isolated products are reported. [a] [Ni(acac)2] (20 mol %), PPh3 (40 mol %). [b] [Ni(acac)2] (10 mol %), PPh3 (20 mol %). [c] [Ni(acac)2] (20 mol %), PPh3 (40 mol %), and 1 (2 equiv) were added in two batches (50 % each time). [d] [Ni(acac)2] (10 mol %), PPh3 (20 mol %), and 1 (2 equiv) were added in two batches (50 % each time). [e] No racemization. To test whether this method could be extended to fluoromethylating reagents other than sulfone 2, we examined ethyl 2-bromo-2-fluoroacetate (4) as a fluoromethylator. To our excitement, phenylboronic acid was fluoroacetated in excellent yield (94 %, Scheme 3, 5 a). The optimized reaction conditions were applied to a variety of boronic acids, furnishing 2-fluoro-2-aryl acetates in yields on par with those observed in the fluoro(phenylsulfonyl)methylations (Scheme 3). Both electron-donating groups, such as methyl, methoxy, and thiomethyl (5 b–5 g), and electron-withdrawing groups, such as fluoro, chloro, bromo, trifluoromethyl, cyano, ethoxycarbonyl, acyl, and even an aldehyde (5 h–5 r), are well-tolerated using this protocol. A dibenzothiophene-derived boronic acid is also a suitable coupling partner of this fluoroacetation, leading 5 u in a good yield. Again, biologically relevant tyrosine- and flavanone-derived boronic acids were monofluoromethylated smoothly under the reaction conditions to give the fluoroalkyl products 5 v and 5 w, respectively, with acceptable yields. Notably, a vinyl boronic acid was also compatible with this method, giving the product in an acceptable yield (63 %, 5 x). While electron-deficient aryl- and heteroaryl boronic acids were not suitable coupling partners in the palladium-catalyzed method developed by Qing and co-workers,9d our nickel-catalyzed system enables these reagents to be useful coupling partners. Scope of substituted phenylboronic acid substrates. Reaction conditions (unless otherwise noted): EtO2CCFHBr (0.2 mmol, 1.0 equiv), 1 (2.0 equiv), [Ni(acac)2] (5 mol %), PPh3 (10 mol %), K2CO3 (1.5 equiv), CH2Cl2, 100 °C, 24 h. Yields of isolated products are reported. [a] [Ni(acac)2] (10 mol %), PPh3 (20 mol %). [b] [Ni(acac)2] (10 mol %), PPh3 (20 mol %) and 1 (2 equiv) were added in two batches (50 % each time). [c] [Ni(acac)2] (20 mol %), PPh3 (40 mol %), and 1 (2 equiv) were added in two batches (50 % each time). [d] [Ni(acac)2] (20 mol %), PPh3 (40 mol %). [e] No racemization. To gain insight into the mechanism of this transformation, a number of control experiments were carried out. First, when the reaction was performed in the presence of 1.0 equivalent of a radical scavenger, namely 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO), the monofluoromethylated product 3 a was not obtained [Eq. (1), Scheme 4]. This observation is consistent with reports that the oxidative addition step in nickel-catalyzed cross-couplings proceeds through a radical pathway.12 We were also able to trap the PhSO2CHF⋅ radical using β-pinene 6 under our standard conditions, affording the ring-opened diene 7 as a mixture of isomers [Eq. (2), Scheme 4), which further implicated the generation of the fluoro(phenylsulfonyl)methyl radical in the catalytic cycle. Furthermore, catalytic (5 mol %) and stoichiometric amounts of [Ni(PPh3)4] were subjected to the standard conditions, respectively, but gave none or only trace amounts of 3 a, thus indicating that the Ni0 species might not be involved in the catalytic cycle (see the Supporting Information for a detailed discussion). Radical trapping experiments. On the basis of the above results and previous reports,13 a plausible mechanism involving a NiI/NiIII catalytic cycle was proposed (Scheme 5). The transmetalation between aryl boronic acid 1 and the initial NiI species A, which may be produced through the comproportionation reaction of in situ generated Ni0 and the remaining NiII species,13a,13j could generate NiI–Ar species B, which subsequently activates 2 through a single-electron transfer to afford NiII intermediate C and the corresponding alkyl radical. Intermediate C is then further oxidized to NiIII species D through an oxidative radical addition, followed by reductive elimination to liberate the final monofluoromethylated product 3. Possible mechanism. The reductive desulfonylation of 3 mediated by Na(Hg) proceeded smoothly with both electron-donating and electron-withdrawing substituents on the phenyl ring in MeOH/THF at −20 °C,14 affording the fluoromethyl arenes 8 a–8 f in excellent yields (Scheme 6). In addition, the fluoroacetated arenes 5 can be hydrolyzed under very mild conditions to afford the corresponding α,α-fluoroarylethanoic acids in excellent yield (95 % for 5 s, see the Supporting Information for details).15 Reductive desulfonylation. Reaction conditions: 3 (0.1 mmol, 1.0 equiv), Na(Hg) (10.0 equiv), Na2HPO4, MeOH/THF, −20 °C. Yields of isolated products are reported. [a] 19F NMR yield. To demonstrate both the potential pharmaceutical relevance and the functional-group tolerance of this method, we next attempted to perform a fluoromethylation of ezetimibe, a drug known to lower plasma cholesterol levels.2c As shown in Scheme 7, the monofluoromethylation of boronic acid 9, derived from acetated ezetimibe, proceeded smoothly to afford the fluoromethylated ezetimibe 10 in 62 % overall yield after subsequent deacylation and desulfonylation. This type of late-stage modification could be highly useful for drug discovery and development as a straightforward method for the synthesis of fluorinated analogues. Fluoro(phenylsulfonyl)methylation of ezetimibe derivative 9 and elaboration to ezetimibe-CH2F 10. Reaction conditions: [a] PhSO2CFHI (1.0 equiv), 9 (2.0 equiv), [Ni(acac)2] (10 mol %), PPh3 (20 mol %), K2CO3 (1.5 equiv), CH2Cl2, 100 °C, 83 %. [b] K2CO3, MeOH/THF, 82 %. [c] Na(Hg), MeOH, 91 %. In summary, we have developed novel monofluoromethylation and monofluoroacetation reactions of aryl boronic acids using nickel catalysis. Both electron-rich and electron-deficient aryl boronic acids were compatible with this new transformation. Mechanistic investigations indicated that a monofluoromethyl radical is involved in the catalytic cycle. Further examination of the mechanistic details of this catalytic cycle, and the application of this technology to the modification of complex molecules, are ongoing in our laboratory. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.