Copper-Catalyzed Amidation of Primary and Secondary Alkyl Boronic Esters

Abstract

The oxidative, copper-catalyzed cross coupling of functionalized alkyl boronic esters with primary amides is reported. Through the identification of appropriate diketimine ligands, conditions for efficient coupling of both primary and secondary alkyl boronic esters with diverse primary amides, including acetamide, have been developed.

Graphical Abstract

The prevalence of nitrogen-containing molecules in the pharmaceutical, agrochemical, and materials industries continues to fuel the development of methods for the construction of C-N bonds. Over the past two decades, transition metal catalysis, including Buchwald-Hartwig,^[1] Lam-Chan,^[2] and Ullman-type^[3] couplings, has revolutionized the way that C_sp2-N bonds are prepared. However, due to the long-established propensity of alkylmetal intermediates to undergo β-hydride elimination, challenges with transmetallation of alkyl nucleophiles, and the relative reluctance of alkyl electrophiles to participate in oxidative additions,^[4] the analogous reactions at alkyl centers have remained largely undeveloped.

Motivated by recent calls for new technologies for amide synthesis,^[5] we recently reported the copper-catalyzed oxidative N-alkylation of primary amides using alkyl boronic acids (Figure 1).^[6] This transformation, an alkyl variant of the Lam-Chan reaction, is a rare example of a transition metal-catalyzed C-N cross coupling at alkyl centers capable of β-hydride elimination.^{[7],[8],[9],[10]} Importantly, not only does this method access biologically relevant secondary amides from non-canonical starting materials, it completely avoids over-alkylation to the corresponding tertiary amides, which are common byproducts using classical methods employing alkyl electrophiles.^[11] This process also provides a novel solution to the difficult conversion of alkyl boronic acids to alkyl amine derivatives.^[12] Despite these advances, the instability and difficulty of preparation of the alkyl boronic acids severely limited the functional group tolerance of the boron-containing coupling partner; only hydrocarbon, or simple ether-containing, primary boronic acids could be coupled. With regard to secondary systems, only cyclopentyl, cyclohexyl, and isopropyl boronic acids could be coupled, with the latter requiring forcing conditions. In all three cases, the yields were very modest.

Figure 1.

Second-Generation Approach to Alkyl Boronic Ester Amidation.

In contrast to alkyl boronic acids, alkyl boronic esters are both highly accessible and stable, and can be prepared with a vast range of functional groups.^[13] Unfortunately, our prior catalyst was incapable of coupling the latter reagents in synthetically useful yields.^{[6, 14]} Recognizing the practicality and synthetic accessibility of alkyl boronic esters, we endeavored to identify catalysts capable of coupling these reagents to primary amides.

We now report two new copper catalysts based on diketimine (nacnac) ligands that effectively couple a wide range of highly functionalized primary and secondary alkyl boronic esters to primary amides. These new catalysts allow direct preparation of secondary amides without over-alkylation or significant β-hydride elimination. These new catalysts significantly advance the ability to prepare highly functionalized amides via this alkyl Lam-Chan strategy.

We began our study investigating the coupling of pinacol boronic ester 2, which contains both an aryl ether and a basic heterocycle, with 4-fluorobenzamide (1, Table 1). Consistent with prior results, our previously optimized conditions (10 mol % CuBr, NaOSiMe₃, di-tert-butylperoxide (DTBP)) provided poor yield of the desired product 3 (entry 1). In contrast to the earlier reactions with alkyl boronic acids, decreasing the base loading resulted in slightly increased yield of 3 (entry 2). Other copper salts (e.g., CuI, CuOAc and Cu(OAc)₂) were found to catalyze the reaction; however the yields were largely comparable to the use of CuBr.^[15] The one exception was Cu(acac)₂, which provided a slightly higher yield (entry 3). An observed side-product from protodeborylation of 2 was suppressed by the addition of 4Å molecular sieves (4Å MS, entry 4).

Table 1.

Identification of Reaction Conditions with Primary Boronic Esters.

entry	[Cu]	NaOSiMe3	ligand	additive	% Yield 3[a]
1	CuBr	2.2 equiv	-	-	33
2	CuBr	1.1 equiv	-	-	48
3	Cu(acac)2	1.1 equiv	-	-	61
4	Cu(acac)2	1.1 equiv	-	4Å MS	75
5	Cu(OAc)2	1.1 equiv	L1	4Å MS	90
6	Cu(OAc)2	1.1 equiv	L2	4Å MS	95
7	Cu(OAc)2	1.1 equiv	L3	4Å MS	94

^[a]Yields determined using ¹H NMR against an internal standard.

To further optimize the reaction, exogenous ligands were examined. Based on the superiority of Cu(acac)₂, we investigated the use of 1,3-diketimine (nacnac) ligands in combination with Cu(OAc)₂. Several electron-rich ligands from this class resulted in highly efficient catalysis. For example, the commonly employed 2,6-dimethylaniline-derived nacnac L1 provides the desired product in 90% yield (Table 1, entry 5). Nacnac ligands with ortho-methoxy groups (L2 and L3 entries 6 and 7) proved even more efficient. Notably, other ligands (such as 2,6-bipyridine, 1,10-phenanthroline, and 2,2',2"-tripyridine) that are commonly employed in aromatic Lam-Chan-type reactions provided much lower yields.^[2e],[15] Despite the similar performance of the nacnac ligands in Table 1, further investigation over a range of amides and primary boronic esters revealed that L3 was superior in nearly all cases.^[15][16]

The scope of primary boronic esters and primary amides is outlined in Scheme 1. Numerous heteroaromatic compounds are compatible, including basic and non-basic indoles (5, 6), pyridines (3, 8, 23, 26), aminopyrimidines (7), benzimidazoles (9), benzothiazoles (10), and thiophenes (4, 6, 11). Other tolerated groups include protic amine derivatives (17), aryl fluorides (e.g., 3–5, etc.), aryl and heteroaryl chlorides (6, 8, 9, 15), alkenes (25), aryl ethers (19–22), and carbamates (5, 17, 18). Nitriles can be used (8, 26), but aliphatic nitriles lead to some attenuation of the yield (16). Interestingly, for primary nitrile 16, ligand L2 provided higher yield than L3. This is counter to all other examples we examined and may be related to the donor ability of the nitrile.

Scheme 1.

Scope of the Coupling Reaction with Primary Pinacol Boronic Esters.

Notably, in the previously reported reaction of alkyl boronic acids, esters were not compatible. However, esters are well tolerated in the current transformation (4, 12, 15, 20). Free alcohols also can be used, albeit with somewhat suppressed yield (19). However, a range of protected alcohols, such as silyl ethers (21–24), pivalates (20) and THP groups (18) can also be used. Substrates bearing acetal- protected ketones can be coupled without incident (12). Significantly, thioethers are tolerated (14) without evidence (LC/MS) of oxidation of the sulfur atom. Aromatic (3–5, etc.) heteroaromatic (6, 23), alkyl (11, 13, and 26), and heteroalkyl (18) amides all participate. Electronic (14, 19–21, 23) and steric modulation (26) of the amide has little effect. A boronic ester bearing a β-stereogenic center in >99% ee coupled efficiently without erosion of ee (24), suggesting that β-hydride elimination and reinsertion does not occur. In some cases (noted in Scheme 1), LiOt-Bu proved more effective than NaOSiMe₃. In all cases, only minimal β-hydride elimination of the alkyl coupling partner occurs; at most 8% corresponding alkene was observed.^[17] Given the complexity and heteroatom content of the products outlined in Scheme 1, this reaction has remarkably broad scope for a transition metal-catalyzed transformation.

Previously, neither acetamide nor trifluoroacetamide were suitable coupling partners.^[6] The increased catalytic efficiency of the present conditions, however, now allows alkylation of these low molecular weight amides with a range of functionalized primary boronic esters (Scheme 2). The ability to cross-couple these small amides is of particular importance because of the relative ease of hydrolysis of the products. In this way, our method provides mild conditions for the formal conversion of boronic esters to primary amines that are competitive with current technologies for this challenging transformation.^{[12a, 12e, 18]}

Scheme 2.

Scope of the Coupling with Acetamide and Trifluoroacetamide.

In the previous work using boronic acids, only the simplest secondary substrates (cyclopropyl, cyclohexyl, isopropyl) could be coupled, and then only in very modest yield. Significantly, no functionality could be tolerated on the secondary alkyl reagents. We attributed this to instability of secondary boronic acids.^[13] In sharp contrast, functionalized secondary pinacol boronic esters are both easily accessible and stable.^[19] Thus, we sought conditions that would allow coupling of these reagents as a means of accessing branched amine derivatives.

Using the coupling of 4-fluorobenzamide and 2-octylboronic acid pinacol ester as a model, we found that the optimized conditions for the coupling of primary boronic esters using L3 provided very little cross-coupled product (Table 2, entry 1). Alarmingly, the small amount of product was observed as a 3:1 mixture of the desired product 33 along with the rearranged linear product 34.^[20] Reactions excluding nacnac ligand (entry 2), or using dimethylaniline-derived L1 (entry 3) provided very similar results. In stark contrast, however, we found that o-anisidine derived ligand L2 provides excellent yield of the product 33, with minimal amounts of rearranged product (entry 4).

Table 2.

Optimization of the Coupling with Secondary Boronic Esters.

entry	ligand	yielda	ratio 33 : 34
1	L3	8	74:26
2	-	10	80:20
3	L1	3	74:26
4	L2	82	97:3
5	L4	36	95:5
6	L5	36	>95:5

^aYield and regiospecific were determined by ¹H NMR analysis of crude reaction mixtures.

Intrigued by these exciting, yet unexpected results, we investigated two other nacnac ligands. First, we examined ligand L4, derived from unsubstituted aniline. Although this ligand did not provide as high yield as L2, it was highly regiospecific (Table 2, entry 5). This suggests that the rearrangement observed with ligand L3 and L4 is steric in nature. Next, we examined the use of L5, which bears p-methoxy substitution (entry 6). This ligand also provided high regiospecificity, but low yield of 33. This suggests that the importance of the o-methoxy groups in L2 on the yield of the reaction is not due to electronic factors but rather that the ether substituents might participate in metal chelation - possibly modulating or stabilizing the catalytic center. Further investigations into the coordination environment of this catalyst are underway.

With optimized conditions in hand for the coupling of amides with secondary boronic esters, the scope was examined (Scheme 3). Gratifyingly, this process provides convenient and selective access to highly functionalized α-branched amides. Secondary boronic esters bearing ethers (37), carbamates (38), heterocyclic amides (39), imides (40), alkenes (44), and acetals (45) underwent smooth coupling with various alkyl and aryl amides. Like with the primary substrates, this reaction shows the remarkable ability to tolerate heterocyclic scaffolds. In no case was isomerization to the linear amide observed.

Scheme 3.

Reaction Scope with Secondary Boronic Esters.

The present reaction appears to be stereoconvergent (Scheme 4). Boronic esters cis-46 and trans-46 were prepared and independently subjected to the cross coupling. Identical diastereomeric ratios were observed (determined by ¹⁹F NMR of crude mixture). This lack of stereospecificity suggests a mechanism that does not involve simply invertive or retentive transmetallation.^[21] Studies are underway to further investigate the mechanism of this transformation.

Scheme 4.

Stereoconvergence in the cross-coupling reaction

In conclusion, we have developed a mild, selective synthesis of secondary amides via the cross coupling of alkyl boronic esters and primary amides using non-canonical starting materials. This cross coupling is a rare example of a transition metal-catalyzed C-N bond construction at an alkyl center, and both primary and secondary alkyl boronic esters participate. Key to the high heterocycle and functional group tolerance was identification of diketimine ligands to support the copper center. Investigations to understand the role of ligand structure, as well as the expansion to other heteroatomic coupling partners is currently underway.