P(III)/P(V)-Catalyzed Methylamination of Arylboronic Acids and Esters: Reductive C–N Coupling with Nitromethane as a Methylamine Surrogate

Abstract

The direct reductive N-arylation of nitromethane by organophosphorus-catalyzed reductive C–N coupling with arylboronic acid derivatives is reported. This method operates by the action of a small ring organophosphorus-based catalyst (1,2,2,3,4,4-hexamethylphosphetane P-oxide) together with a mild terminal reductant hydrosilane to drive the selective installation of the methylamino group to (hetero)aromatic boronic acids and esters. This method also provides for a unified synthetic approach to isotopically-labeled N-methylanilines from various stable isotopologues of nitromethane (i.e. CD₃NO₂, CH₃¹⁵NO₂ and ¹³CH₃NO₂), revealing this easy-to-handle compound as a versatile precursor for the direct installation of the methylamino group.

Graphical Abstract

Nitromethane (H₃C–NO₂) is an industrially important commodity chemical, primarily used as a solvent, stabilizer, and fuel additive.¹ Also, H₃C–NO₂ is a common and useful carbon pronucleophile in synthesis (Figure 1A, left),^2–5 but it is comparatively less developed as an amination reagent. Recently, Jiao and coworkers reported H₃C–NO₂ as a precursor to “H₂N–X” equivalents by reductive Nef-like decomposition,⁶ but an alternative reductive N-functionalization of H₃C–NO₂ that retains the “H₃C–N” substructure would trace a role for this inexpensive and easily-handled liquid as a potential surrogate for methylamine—a gaseous List 1 controlled precursor⁷—especially in catalytic coupling chemistry (Figure 1A, right).^8–11 At present, reductive N-functionalization of H₃C–NO₂ is limited to two isolated examples: Niggemann has reported reductive N-benzylation of nitromethane mediated by stoichiometric B₂pin₂ in the presence of excess BnZnBr,¹² and Suarez-Pantiga and Sanz have reported a triphenylphosphine-mediated reductive N-phenylation of nitromethane catalyzed by an oxomolybdenum(VI) compound under microwave irradiation.¹³ We report here a general catalytic method for the methylamination of arylboronic acids and esters with H₃C–NO₂ by reductive C–N coupling driven by P^III/P^V=O redox cycling (Figure 1B). These results expand the scope of reductive C–N coupling, generalize the use of H₃C–NO₂ as a methylamine surrogate in reductive cross coupling, and provide a unique pathway for introducing stable isotopes (¹⁵N, ¹³C, ²H) widely-used for metabolic tracing.¹⁴

Figure 1.

A) Uses of nitromethane in organic synthesis. B) Reaction sequence leading to formation of N-methylanilines via P^III/P^V=O-catalyzed reductive C–N coupling of nitromethane with boronic acids (esters); C) Electronic challenge presented by high frontier orbital gap of H₃C–NO₂. FMO=frontier molecular orbital.

Prior work has established P^III/P^V=O catalysis^15–17 as a viable approach to reductive N-functionalization of nitroarene (Ar–NO₂) substrates,^18–20 but the translation of this technique to reductive N-functionalization of H₃C–NO₂ requires that the desired C–N coupling sequence (Figure 1B) outcompete numerous unproductive but well-known decomposition pathways, both for H₃C–NO₂ (e.g. Nef reaction) and potential deoxygenation intermediates (e.g. tautomerization of H₃C–NO→H₂C=NOH and H₃C–N:→H₂C=NH, inter alia). In principle, the base-free conditions of the P^III/P^V=O catalyzed nitro functionalization, which depends on a rate-limiting (3+1) cheletropic addition of phosphine and nitro substrate, might suppress these processes. However, the nitro moiety of H₃C–NO₂ is inherently less reactive than Ar–NO₂ as cycloaddition partner due to a larger local frontier orbital energy gap (ΔΔE= 1.0 eV, Figure 1C),²¹ resulting in less favorable pairwise orbital interactions that therefore impose a stringent demand of the biphilic character of the phosphetane catalyst.

Investigation of the reductive N-functionalization of H₃C–NO₂ was initiated with these considerations in mind. Empirical observations (Table 1) indicate that a catalytic system comprising the organophosphorus O-atom transfer catalyst 1•[O]²² and a terminal hydrosilane reductant indeed successfully achieve a methylamination of arylboronic acids by P^III/P^V=O-catalyzed intermolecular C–N coupling. With 4-fluorophenylboronic acid (3) as a representative substrate, the reductive functionalization of 1.0 equiv of commercial reagent grade H₃C–NO₂ (2) by 10 mol% of 1•[O] and 2 equiv of PhSiH₃ gave a promising 35% yield of N-methyl-4-fluoroaniline (4) (Table 1, entry 1), but simply increasing equivalencies of inexpensive H₃C–NO₂ (3.0 equiv, entry 2) resulted in excellent yield of 4 (95% NMR yield, 90% isolated yield). A high yield was similarly obtained with phosphetane 1 as a precatalyst (entry 3)—consistent with P^III/P^V=O redox cycling—but omission of either the precatalyst 1•[O] or phenylsilane failed to provide methylamination product 4 (entries 4–5). Inexpensive and bench-stable polymethylhydrosiloxane (PMHS) is also a viable terminal reductant for the reaction, albeit with somewhat lower efficiency (85%, entry 6). In situ spectral monitoring of the reductive C−N coupling showed that isotopically enriched nitromethane-d₃ (δ 4.32 ppm) is cleanly converted to the product N-(methyl-d₃)-4-bromoaniline (δ 2.79 ppm), and no side products or long-lived intermediates were observed (Figure S2). The relatively high concentration of the P^III catalyst resting state 1 presumably drives the methylamination forward along the productive reaction pathway in preference to possible side pathways, including the favorable isomerization of nitrosomethane to formaldoxime.^23,24

Table 1.

Discovery and Control Experiments for Organophosphorus-Catalyzed N-methylaniline Synthesis.^a

Entry	silane	R3P=O	Yield (%)b
1c	PhSiH3	1•[O]	35
2	PhSiH3	1•[O]	95 (90)d
3	PhSiH3	1	94
4	PhSiH3	None	0
5	None	1•[O]	0
6e	PMHS	1•[O]	85

^aUnless otherwise noted, reactions were carried out with 1•[O] (10 mol %), phenylsilane (0.5 mmol, 2.0 equiv), 4-fluorophenylboronic acid (0.25 mmol) and nitromethane (0.75 mmol, 3.0 equiv) in CPME (0.5 mL) at 120 °C for 18 h.

^bYields were determined by ¹⁹F NMR integration with the aid of an internal standard.

^c1.0 equiv of 2 was used.

^dIsolated yield.

^e24 h reaction time.

As depicted in Figure 2A, the P^III/P^V=O-catalyzed methylamination method shows excellent functional group tolerance, several facets of which provide for chemoselectivities that complement transition-metal catalyzed C–N coupling methods. For example, halogen substituents are well-tolerated (5a, 8–10, 16) in the P^III/P^V=O-catalyzed methylamination and intermolecular C–N coupling occurs exclusively at the boronic acid position. Also, the orthogonal reactivity to the nitro group with respect to other nitrogen-containing functional groups permits selective aryl methylamination even in the presence of free −NH₂ groups (16) without explicit protection. Carbonyl functionalities such esters (13) and ketones (17, 19) that might be susceptible to acylation or condensation reaction with H₃C–NH₂ are retained when H₃C–NO₂ is used as a surrogate. Furthermore, substrates with other functional groups including benzyloxy (7), tetrahydropyranyl ether (14), thiomethyl (15), and trifluoromethyl (18) groups all yielded the corresponding N-methylaniline products in good yields. The method is also applicable to compounds with known bioactive core structures, such as derivatives of celecoxib (18), estrone (19), and Pittsburgh compound B²⁵ (20).

Figure 2.

Synthetic scope of P^III/P^V=O-catalyzed C–N coupling of nitromethane and arylboronic acids. See SI for full experimental details and conditions. Yields are reported for material obtained following purification and isolation.

The ability to use of H₃C–NO₂ as a simple synthetic precursor for the MeHN– group presents the opportunity for the preparation of diverse stable-isotope labeled products—useful tools in both laboratory²⁶ and pharmaceutical²⁷ research—through a unified synthetic strategy. As illustrated in Figure 2B, a suite of isotopically labeled N-methylaniline products 5b–5d are accessible with programmed labeling as dictated by the initial isotopic composition of the nitromethane isotopologue (viz. ²H₃C–NO₂ (2b), H₃¹³C–NO₂ (2c), H₃C–¹⁵NO₂ (2d)). The relative ease with which the isotopologues of nitromethane are accessed make this an attractive approach to potentially valuable stable isotope labeled compounds.

Initial attempts to directly transfer the foregoing P^III/P^V=O-catalyzed conditions to methylamination of 3-quinolylboronic acid resulted in a disappointing 19% yield of the target compound 21 (Figure 3A, yield in parenthesis). Studies showed that protodeboronation of the heteroarylboronic acid was a primary pathway limiting the productive C–N coupling in this case.²⁸ Correspondingly, P^III/P^V=O-catalyzed methylamination of 3-quinolylboronic acid 1,3-propanediolate ester showed a significant improvement in C–N coupling yield (93%), in line with the relative stability of boronic esters to protodeboronation as compared to their parent boronic acids.²⁷ This phenomenon is general; B-heteroaryl-1,3,2-dioxaborinanes including a variety of five- and six-membered heterocyclic derivatives, such as isoquinoline (22), indazole (23) pyridine (24, 27), thiophene (25) and pyrimidine (26), were successfully transformed to the corresponding N-methylamine derivatives with improved yields relative to their arylboronic acid congeners. B-Aryl-1,3,2-dioxaborinanes containing electron-donating (28) or electron-withdrawing group (29) transform smoothly to the desired products, showing the generality of utilizing boronate esters for methylamination reactions. Other common boronate ester residues such as a neopentyl glycol ester (e.g. Ar-Bnep 30, Figure 3B) and pinacol esters (e.g. Ar-Bpin 31–33, Figure 3C) gave similar efficiency in this methylamination process. Notably, subjection of benzene-1,4-diboronic acid bispinacol ester to the standard reaction conditions afforded single methylamination product 31 with excellent yield; evidently the second (pinacolato)boryl moiety becomes electronically deactivated for amination following an initial reductive C−N coupling.

Figure 3.

Synthetic scope of P^III/P^V=O-catalyzed C–N coupling of nitromethane and (hetero)arylboronic esters. See SI for full experimental details and conditions. Yields are reported for material obtained following purification and isolation. ^a Yields for reactions with boronic acids were determined by ¹H NMR with dibromomethane as the internal standard.

Competition experiments evidence a differential reactivity of H₃C–NO₂ as compared to Ar–NO₂ substrates in P^III/P^V=O-catalyzed reductive C–N coupling. As illustrated in Figure 4A, an equimolar mixture of four componenets—namely H₃C–NO₂ (2), 4-NC-C₆H₄–NO₂ (34), 3,5-(F₃C)₂-C₆H₃–Bpin (35) and 4-H₃CO-C₆H₄–B(OH)₂ (36)—was subjected to standard reductive C–N coupling conditions catalyzed by 1•[O] in a one pot manner. In the event, only two of the possible four products—specifically methylaniline 29 and diarylamine 37—were observed in greater than 5% yield. Consistent with the FMO rationale in Figure 1,¹⁵ⁱ this observation implies that nitroarene 34 reacts more quickly than nitromethane 2 via (3+1) cheletropic addition with 1, and goes on to couple selectively with the more reactive boronic acid partner 36. In a subsequent event, H₃C–NO₂ then reacts to give 29 by methylamination of the more inert boronic ester 35, proceeding without diminished yield for these reaction partners.²⁹ This kinetically-controlled chemoselectivity was further demonstrated by the modular synthesis of N,N’-difunctionalized phenylenediamine (Figure 4B, 42). The three-component coupling of H₃C–NO₂, Ph–B(OH)₂ (40), and 4-nitrophenylboronic ester (41) produced 42 via sequential amination reactions with almost complete suppression of undesired methylaminated product (6).

Figure 4.

A) Reaction of a four component mixture illustrating differential reactivity of H₃CNO₂ vs ArNO₂ and ArB(OH)₂ vs ArB(OR)₂. (B) Selective three-component coupling via sequential C–N coupling reactions.

A synthesis of the anxiolytic triflubazam³⁰ (45, Figure 5) contextualizes the reductive C–N coupling of nitromethane within the growing portfolio of P^III/P^V-catalyzed methods promoted by phosphetane catalyst 1•[O]. Specifically, P^III/P^V-catalyzed reductive C–N coupling of phenylboronic acid and 4,4,5,5-tetramethyl-2-(2-nitro-5-(trifluoromethyl)phenyl)-1,3,2-dioxaborolane (43), followed in situ by P^III/P^V-catalyzed amidation^15h,31 with monoethyl malonate gives diarylamide 44 in a one-pot sequence. Subsequent P^III/P^V-catalyzed methylamination by C–N coupling with H₃C–NO₂ installs the MeHN– moiety and induces intramolecular cyclization to close the diazepinedione ring, furnishing the medicinal target 45 in an overall 51% yield for the two-pot, three-step, all-P^III/P^V-catalyzed sequence.

Figure 5.

Two-pot, three step synthesis of triflubazam by an all-P^III/P^V-catalyzed sequence. Reaction conditions: (a) 43 (1.0 equiv), PhB(OH)₂ (1.1 equiv), Ph₂SiH₂ (4.0 equiv), 1•[O] (30 mol%), CPME, 120 °C ; then add monoethyl malonate (1.2 equiv), diethylbromomalonate (1.5 equiv), 40 °C; (b) 44 (1.0 equiv), H₃C–NO₂ (3.0 equiv), Ph₂SiH₂ (3.0 equiv), 1•[O] (15 mol%), CPME, 120 °C.

In summary, we have demonstrated that nitromethane is an inexpensive and easy-to-handle synthetic equivalent for installation of the MeHN– fragment via P^III/P^V=O catalysis. Readily-available boronic acids and esters are selectively methylaminated in the presence of various functional groups and heteroaromatics. The method serves as a robust complementary tactic to transition metal catalyzed C–N coupling techniques relying on the use of MeNH₂ or related surrogates. With respect to our efforts to develop P^III/P^V=O-catalyzed reductive C–N coupling reactions as modular approach to complex amine synthesis, this study represents our first success with nitroalkane substrates and portends future developments in the deoxygenative N-functionalization of higher nitroalkane homologues to access more elaborate N-alkyl products