Synthesis of Secondary Amines via Self-Limiting Alkylation

Abstract

N-centered nucleophilicity increases upon alkylation and thus selective partial alkylation of ammonia and primary amines can be challenging: Poor selectivity and over alkylation are often observed. Here we introduce N-aminopyridinium salts as ammonia surrogates for the synthesis of secondary amines via self-limiting alkylation chemistry. Readily available N-aryl-N-aminopyridinium salts engage in N-alkylation and in situ depyridylation to afford secondary aryl alkyl amines without any overalkylation products. The method overcomes classical challenges in selective amine alkylation by accomplishing alkylation via transient, highly nucleophilic pyridinium ylide intermediates and can be applied in the context of complex molecular scaffolds. These findings establish N-aminopyridinium salts as ammonia synthons in synthetic chemistry and establish a strategy to control the extent of amine alkylation.

Graphical Abstract

N-alkylation chemistry,^1–3 which leverages the intrinsic N-centered nucleophilicity of trivalent nitrogen to forge new C–N bonds, is among the first reactions taught in introductory chemistry courses and comprises ~10% of all transformations carried out in pharmaceutical research and development.⁴ Despite the ubiquity of N-alkylation chemistry, significant challenges remain: N-centered nucleophilicity increases upon alkylation, which renders partial alkylation, for example to selectively access secondary amines, difficult (Figure 1a).⁵ In addition, substitution chemistry is often accompanied by competing elimination processes. Overalkylation and competitive elimination are particularly pronounced for application of ammonia alkylation in selective synthesis. To overcome some of these challenges, ammonia surrogates — phthalimides,⁶ sulfonamides,⁷ dioxazolones,⁸ benzotriazoles,⁹ and hydroxylamine derivatives¹⁰ — have been developed. While new amination reactions have been enabled by these reagents, the downstream chemistry is typically limited to deprotection to afford primary amines.

Figure 1.

(a) Amine alkylation with alkyl electrophiles is the most common route to secondary aryl-alkyl amines. (b) Reductive amination and C−N cross coupling are popular approaches for secondary aryl-alkyl amine synthesis. Alkyl electrophiles are readily available from natural sources. (C) Here, we demonstrate self-limiting alkylation of N-aminopyridinium salts enables selective synthesis of secondary amines.

In response to the intrinsic challenges of partial N-alkylation, an arsenal of methods, including reductive amination and metal-catalyzed C–N coupling, have been developed.^11–14 While these methods provide access to large families of amine products, N-alkylation, reductive amination, and C–N cross-coupling all require the use of amine-containing starting materials. These amine-containing starting materials are synthesized through several steps from naturally occurring feedstocks (e.g., alcohols or olefins), which can render them inconvenient starting materials. Thus, complementary methods, for example the enable selective N-functionalization of amine surrogates with alkyl electrophiles are of interest (Figure 1b).^15–20

Previously, we demonstrated the synthesis of secondary arylalkyl amines via C−H N-aminopyridylation followed by Ni-catalyzed cross-coupling with aryl boronic acids.²³ Although the scope for the C–N cross-coupling chemistry was broad, the initial C−H aminopyridylation exhibited significant limitations, notably working only with activated C−H bonds.

Motivated by the fundamental chemical challenges associated with selective partial ammonia alkylation and the ongoing need for new synthetic methods to access secondary amines, here, we demonstrate that N-aminopyridinium reagents are useful ammonia synthons in the selective construction of secondary aryl-alkyl amines. We describe the monoalkylation of N-aryl-N-aminopyridinium derivatives with readily available alkyl halides. We term this process as self-limiting alkylation because in contrast to classical N-alkylation chemistry, in which N-alkylation results in a more reactive nucleophile, our method achieves monoalkylation by accomplishing alkylation via a highly nucleophilic pyridinium ylide. Following alkylation, the obtained N-alkyl-N-pyridinium amine is a less reactive nucleophile. These results provide new modular disconnections to rapidly assembly secondary amines, extend the burgeoning chemistry of N-aminopyridinium salts as bifunctional amine synthons,^21–27 and introduce self-limiting alkylation as a conceptual approach for selective synthesis of secondary amines (Figure 1c).

We initially envisioned a two-step protocol for secondary amine synthesis based on sequential N-arylation and N-alkylation of N-aminopyridinium salts and set out to identify conditions for each of these transformations. Previous reports of N-arylation of N-aminopyridinium via either S_NAr or Pd-catalyzed coupling were limited to electron-deficient, heteroaryl halide coupling partners.^21,28 Work from our laboratory in C–H amination²³ and olefin aziridination²² efforts suggested that N-aminopyridinium salts can serve as plug-in replacements for sulfonamide reagents in many amination protocols. To access a family of N-aryl-N-aminopyridinium derivatives 3 needed to explore self-limiting alkylation chemistry, we extended the sulfonamide-to-aminopyridinium analogy to achieve highly efficient CuF₂-catalyzed Chan-Lam cross-coupling^{29, 30} of N-aminopyridinium salts and with aryl boronic acids (2) to affords a family of N-aryl-N-aminopyridinium salts (3) (Figure 2; see Supporting Information for optimization details). In no case were products of double arylation observed and the procedure could be readily translated to gram-scale synthesis: Compound 3a was prepared in 71% yield on an 8 mmol scale (2.2 g product). With access to a family of N-aryl-N-pyridinium amines, including those derived from electron-deficient, -neutral, and -rich substrates as well as pharmaceutically derived precursors, we turned our attention to developing alkylation chemistry that would provide access to secondary aryl alkyl amines.

Figure 2.

Cu-catalyzed coupling of 1 with aryl boronic acids (2) to afford N-arylaminopyridinium salts (3). Conditions: 1 (1.0 equiv), 2 (2.0 equiv), CuF₂ (10 mol%), DMA (2 M), O₂, 70 °C; * Conditions: 1 (1.0 equiv), 2 (3.0 equiv), CuF₂ (10 mol%), N¹,N²-di([1,1’-biphenyl]-2-yl)benzene-1,2-diamine (20 mol%), DMA (2 M), O₂, 70 °C.

Using salt 3a, we rapidly identified conditions for a one-step N-alkyation, depyridylation cascade that delivered secondary amines selectively. Treatment of a MeCN solution of 3a with hexyliodide 4a and CsOAc at 70 °C resulted in N-alkylation to pyridinium amine 5a’ in 98% yield (Table 1). During optimization studies, we observed that while carboxylate and bicarbonate bases afforded alkylated product 5a’ (Table 1, Entries 1 and 2), tert-butoxide or carbonate bases afforded secondary amine 5a, the product of in situ depyridylation of 5a’, directly (Entries 3–5). Ultimately, Cs₂CO₃ was identified as the optimized base and promoted a one-pot alkylation/depyridylation sequence to furnish secondary amine 5a in 79% yield (see below for discussion of the depyridylation mechanism).

Table 1.

Optimization of the alkylation-depyridylation protocol. Conditions: 3a (1.0 equiv), 1-iodohexane (4a, 2.0 equiv), base (3.0 equiv), CH₃CN, 70 °C, 16 h; yields are measured using 1,3,5-trimethoxybenzene as the internal standard.

entry	base	yield of 5a’ (%)	yield of 5a (%)
1	CsOAc	98	0
2	NaHCO3	78	0
3	NatBuO	0	23
4	K2CO3	18	45
5	Cs 2 CO 3	0	79

The developed alkylation/depyridylation protocol provided access to a broad array of secondary aryl alkyl amines (Figure 3). Diverse functional groups, including long alkyl chains (5a and 5b), cyanides (5c), amides (5e), and protected alcohols (5f) are well-tolerated. The reaction could be accomplished from the corresponding alkyl bromide or triflate, as demonstrate for the synthesis of 5a, in 63% and 49% yield, respectively. While secondary allylic and benzylic iodides engage is efficient alkylation, as demonstrated by the examples 5g (81%) and 5h (62%), unactivated secondary alkyl iodides do not engage in alkylation (i.e., 5i). For highly electrophilic starting materials, e.g., 4j–4m, a lower reaction temperature was used to prevent overalkylation: methyl iodide engaged in selective monoalkylation to afford 5j in 96% yield; primary benzyl iodides 4k and 4l were converted to secondary amines 5k and 5l in 98% and 60% yield, respectively; and primary allyl iodide 4m afforded secondary amine 5m in 43% yield.

Figure 3.

Monoalkylation of 3. Conditions: 3 (1.0 equiv), 4 (2.0 equiv), Cs₂CO₃ (3.0 equiv), CH₃CN, t °C, 16 h; yields are isolated.

With respect to the N-arylaminopyridinium reaction partner (3), both electron-neutral and electron-donating (5n-5q) aminopyridinium derivatives were converted to their respective secondary aryl-alkyl amines efficiently. Moreover, thioflavin-T derived N-aminopyridnium derivative 3p was converted to methylated amine 5r, which is used in investigational studies of Alzheimer’s disease, in 74% yield.

The developed alkylation/depyridylation chemistry can be implemented in the context of natural products and drug molecules. Linoleyl iodide was transformed into 5s in 69% yield. Ibuprofen-derived alkyl iodide was converted to amine 5t in 58% yield. Isoxepac-derived and indomethacin-derived alkyl iodides were successfully coupled to give the corresponding products 5u and 5v in 42% and 45% yields, respectively. Biphenyl derived aminopyridinium 3c can also be coupled with indomethacin-derived alkyl iodide 4v to give the corresponding product 5w in 78% yield. The thioflavin T-derived N-aminopyridinum salt 3p can be coupled with ibuprofen-derived alkyl iodide 4t to yield the drug conjugate 5x in 70% yield. These examples highlight the compatibility of C–N bond construction with pharmaceutically relevant basic heterocycles, amides, carbamates, and basic amines.

With robust conditions for self-limiting alkylation in hand, we turned our attention to understanding 1) the origins of the observed partial alkylation and 2) the mechanistic basis for in situ depyridylation. The following mechanistic studies were carried out with N-pyridinium salt 3a, but importantly, electron withdrawing groups are not needed for efficient depyridylation (i.e., 5n, 5o, 5p, and 5x).

To investigate the origin of the observed partial alkylation selectivity, we treated an independently prepared sample of ylide 3a’ with alkyl iodide 4a and observed the rapid formation of alkylation pyridinium salt 5a’ (Figure 4a, see Section C.1 of Supporting Information). In contrast, neither p-trifluoromethyl aniline 6 nor amine 5a undergo deprotonation in the presence of Cs₂CO₃ nor undergo alkylation to an appreciable extent upon subsequent exposure to 4a (Figure 4b, see Section C.3 of Supporting Information). These data indicate that the presence of an electron-withdrawing pyridinium substituent in 3a, which lowers the N–H pKa and enables access to ylide 3a’, is critical to efficient alkylation: Ylide 3a’ is more nucleophilic than amine 6 and thus undergoes alkylation under conditions that 6 is unreactive. In contrast to typical amine alkylation reactivity trends, in situ depyridylation renders the products of alkylation (i.e., 5a) less nucleophilic than the starting material (i.e., ylide 3a’).

Figure 4.

(a) Reaction of ylide 3a’ with 4a yields intermediate 5a’. (b) Aniline 6 does not undergo alkylation. (c) Depyridylation of 5a’ proceeds with Na₂CO₃ and 15-crown-5. (d) Proposed mechanism for the alkylation/depyridylation sequence.

To investigate the mechanism of the unanticipated in situ depyridylation, we treated an independently synthesized sample of compound 5a’ with each of the reagents present during the alkylation reaction (i.e., 3a, 3a’, 4a, and Cs₂CO₃; see Section C.4 of Supporting Information). Of these reactions, treatment of 5a’ with excess Cs₂CO₃ uniquely resulted in depyridylation to afford 5a in 79% yield along with pyridine (26%) and pyridine-derived products (see Section C.5 of Supporting Information). In addition, while the one-pot alkylation/depyridylation protocol described above utilizes 3 equivalents of Cs₂CO₃, alkylation of 3a in the presence of 1 equivalent of Cs₂CO₃ resulted in alkylated pyridinium amine 5a’ (78%), not secondary amine 5a (see Section C.6 of Supporting Information). Based on these observations, we hypothesize that carbonate serves two roles: As a base to generate nucleophilic pyridinium ylides and as a reductant to promote in situ depyridylation.³¹

Consistent with the carbonate-as-reductant hypothesis, while Na₂CO₃ does not promote efficient depyridylation of 5a’, addition of 15-cr-5 to sequester Na⁺ and generate a more reducing carbonate source, promotes efficient depyridylation (60% yield of 5a, Figure 4c). Analysis of the reaction headspace revealed the formation of CO₂ during successful deprotection reactions whereas unproductive conditions (i.e., Na₂CO₃ without 15-cr-5) did not evolve CO₂ (see Section C.7 of Supporting Information). Finally, exposure of pyridinium salt 5a’ to Cs₂CO₃ results in the formation of a low-energy absorbance centered at 450 nm (see Section C.7 of Supporting Information). This observation is consistent with the formation of a carbonate-to-pyridinium electron donor-acceptor (EDA) complex (5”).^32–35 Attempts to trap the putative radical intermediate with common EPR spin traps, such as N-tert-butyl-α-phenylnitrone (PBN), were unsuccessful, which is consistent with the expected short lifetime of this reactive intermediate. Together, these data indicate that Cs₂CO₃ promotes depyridylation of alkylated pyridinium amine 5a’ via thermally promoted electron transfer within a discrete EDA complex.

In conclusion, we introduce self-limiting alkylation chemistry as a platform for partial amine alkylation. We describe a one-pot synthesis of secondary amines via self-limiting alkylation of N-aryl-N-aminopyridinium salts. Chan-Lam coupling of aryl boronic acids with N-aminopyridinium triflate provides access to a diverse set of N-aryl-N-pyridinium amines. Deprotonation of these salts affords highly nucleophilic pyridinium ylides that engage is facile substitution chemistry with alkyl halides. The resulting pyridinium salts are much less nucleophilic than the ylide precursor, which enforces selective monoalkylation. In situ reductive cleavage of the N–N bond (i.e., removal of pyridinium moiety) affords secondary amines. Mechanistic experiments suggest that this unique depyridylation is triggered by electron transfer from Cs₂CO₃. Notably, Cs₂CO₃ serves a dual function, acting both as a base and a reductant in this transformation. The N-arylation, N-alkylation sequence can be applied in the context of complex, pharmaceutically relevant molecules. Unlike amination methods that necessitate amine-based starting materials, the presented method harnesses readily available alkyl electrophiles in partial alkylation chemistry. The resulting method represents a new approach to secondary amines and validate N-aminopyridinium compounds as ammonia surrogates in synthetic chemistry.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.³⁶