Chemoselective Cu-Catalyzed Cross-Nucleophile Alkylarylation of Alkenes

Abstract

We report a general cross-nucleophile alkene alkylarylation which adds two different boronic acids, an alkyl and an aryl, across a vinylarene to afford 1,1-diarylalkanes. The excellent chemoselectivity derives from the distinct reactivities of the two boronic acids: the alkylboronic acid is selectively oxidized to an alkyl radical while the arylboronic acid favors transmetalation with the Cu(II)-catalyst. Mechanistic studies suggest that Lewis acid–Lewis base interactions between in-situ generated boroxines and added amine are critical for selectivity. A scope of 37 examples is presented with structurally and electronically diverse alkyl, vinyl, and aryl coupling partners. The synthetic utility is demonstrated in the preparation of Pimozide and MCF-7 analogues.

Graphical Abstract:

Oxidative cross-nucleophile coupling reactions require the two nucleophilic reagents to have distinct reactivity to achieve a selective transformation (Scheme 1a).¹ Alkyl- and arylboronic acids (R–B(OH)₂) are attractive coupling partners as they are easy-to-handle, air- and moisture-stable, and commercially available with over 33,000 currently purchasable.² Moreover, they exhibit versatile reactivities^2–5 depending on the nature of the R group; for example, mild oxidants^3c,6 are known to oxidize alkyl–B(OH)₂ while aryl–B(OH)₂ are superior transmetalating reagents (Scheme 1b).^2a,4b We hypothesized that, in the presence of a Cu-catalyst under oxidizing conditions, we could accomplish the alkylarylation of a vinylarene to chemoselectively add two distinct boronic acids across an alkene and generate a 1,1-diarylalkane, a prominent pharmacophore that is prevalent in pharmaceuticals and bioactive compounds (Scheme 1e), in a single step. This approach, shown in Scheme 1c, would involve a chemoselective functionalization with two distinct boronic acids under conditions wherein an alkyl–B(OH)₂ is selectively oxidized to the corresponding radical while an aryl–B(OH)₂ undergoes transmetalation.^7a,7c,7d,7e Notably, this proposed chemoselective single-electron oxidation of alkyl- over arylboronic acids is challenging to achieve as radical generation from both boronic acids is reported under photochemical^3d,8,9 and strongly oxidizing conditions^9,10.

Scheme 1.

Oxidative Cross-Nucleophile Alkene Alkylarylation Strategies and This Work

The importance of 1,1-diarylalkanes¹¹ have inspired considerable attention towards the development of multi-component synthetic strategies.^7,12 To date, many two-component strategies, such as hydro(amino)alkylation^13a, cross-couplings^13b-f, hydrogenation^13g, and conjugate addition^13h, have been developed. For example, our group disclosed the photochemical hydro(amino)alkylation of 1,1-diarylethenes with versatile alkyl radical precursors, e.g. α-TMS amines, to afford 1,1-diarylalkanes in good to excellent yields. While these strategies successfully furnish 1,1-diarylalkanes with high efficiency, three-component strategies are complementary as they allow for the flexible incorporation of both alkyl and aryl fragments simultaneously.

Developing a three-component alkene alkylarylation strategy faces significant chemoselectivity challenges (Scheme 1a) which require a careful reaction design to bypass undesired side reactions, e.g. two-component couplings, and, ultimately, achieve a chemoselective alkene alkylarylation. Recent advances have focused on transition metal-catalyzed alkene alkylarylation via radical mechanisms⁷ under redox-neutral and reductive conditions. Limited examples of oxidative alkylarylation have been reported and rely on the formation of a benzylic carbocation intermediate that can undergo electrophilic aromatic substitution with nucleophilic arenes (Scheme 1d) and, thus, electron-neutral and deficient styrene and arene derivatives, which cannot go through this pathway, do not participate.¹⁴ The regio- and chemoselective addition of two distinct boronic acids across a vinylarene is a conceptually different approach. It directly incorporates three easily accessible, bench stable reagents and can generate a wide array of 1,1-diarylalkanes some of which are not accessible by the aforementioned methods, i.e. those bearing two electron-deficient arenes (Scheme 1e).

Previously, we demonstrated the use of alkyl–B(OH)₂ as radical precursors in an oxidative three-component carboamination reaction. Mechanistic investigations revealed that single-electron oxidation of the alkyl–B(OH)₂ can be facilitated by a LA–LB interaction between the boron and an amine; notably, arylboronic acids do not participate under these reaction conditions.⁶ Encouraged by these findings, we envisioned that cross-nucleophile alkylarylation would proceed via the following mechanism: (1) the boronic acids and an aliphatic amine form LA–LB adducts I and II, (2) subsequent single-electron oxidation of I by Mn(IV)¹⁵ generates alkyl radical III which (3) adds across 2a generating benzylic radical IV, (4) transmetalation between II and [Cu(II)] affords [Cu(II)]–Ar V, (5) which reacts with IV to yield a formal Cu(III)-intermediate (VI), (6) subsequent reductive elimination affords the product (4a), and (7) oxidation of the resulting Cu(I) to Cu(II) closes the catalytic cycle (Figure 1a).

Figure 1.

Proposed Mechanism and Inherent Challenges in Chemoselectivities

This proposed cross-nucleophile alkylarylation faces several significant chemoselectivity challenges (Figure 1b): the rates of radical generation and transmetalation must be compatible to suppress side reactions, including homocouplings (5–6), cross-coupling (7), oxidation-elimination (8), protolytic-cleavage (9), and dimerization (10), as well as dialkyl- and diarylations of the alkene (11–12). Thus, a careful selection of the copper catalyst, oxidant, and a potential LB additive is critical to achieving the desired transformation by modulating the various relative rates.

Given these reaction hypotheses, we began our efforts towards the development of the Cu-catalyzed alkylarylation by choosing cyclohexylboronic acid (1a), styrene (2a), and phenylboronic acid (3a) as model substrates to afford alkylarylation product 4a (Table 1). Consistent with our hypothesis, the choice of LB significantly impacts the reaction outcome in both reactivity and chemoselectivity (entries 1–4; Table S1). Bulky, non-nucleophilic tertiary amines, e.g., triethylamine (TEA), are most efficient at promoting the desired alkylarylation affording 4a in up to 23% yield and successfully suppressing the formation of 5–8 and 11–12 (entries 1–2). More nucleophilic amines and strong inorganic bases are ineffective as they give poor yields of 4a along with 6 and 10 (entries 3,4). Next, we investigated the impact of ligands on the reaction (entries 5–8) and determined that rigid, sterically hindered L4 improves the yield to 56%. Despite the chirality imparted by L4, no enantiomeric excess is observed.¹⁶ Further optimizations identified the following reaction conditions to afford 4a in 75% GC yield (76% NMR yield): 1a (2.0 equiv), 2a (1.0 equiv), 3a (2.0 equiv), Cu(MeCN)₄PF₆ (10 mol %), L4 (10 mol %), MnO₂ (3.0 equiv), TEA (1.5 equiv), and powdered 4Å molecular sieves (MS) (40 mg) in THF (0.33 M) at 70 °C for 48 h (entry 10). Gratifyingly, air only moderately impacts the reaction efficiency (entry 11); 1.0 equiv of water has no deleterious effect, while 5.0 equiv of water reduces the yield to 37% (entries 12,13). Under the optimized conditions, model product 4a is isolated in 48% yield; this is largely due to the compound being difficult to visualize and its coelution with unreacted starting materials and side product 10. Alternative isolation conditions via simple filtration through silica afford 4a in 65% yield along with 7% of 10.

Table 1.

Select Optimization Studies^a

Entry	Deviation from the Condition Above	6	8	10	4a
1	None	-	-	22	23
2	DIPEA instead of TEA	-	-	14	21
3	DABCO instead of TEA	-	-	14	2
4	NaOEt instead of TEA	20	-	16	8
5	Bpy instead of L1	49	-	13	1
6	L2 instead of L1	-	-	17	19
7	L3 instead of L1	-	-	18	28
8	L4 instead of L1	-	-	8	56
9	with L4 as ligand, at 70 °C for 48 h	-	-	10	68
10	entry 9, 40 mg 4Å MS	-	-	10	75(65) c
11	entry 10, open to air for 10 min	-	-	17	45
12	entry 10, with 1.0 equiv water added	-	-	13	78
13	entry 10, with 5.0 equiv water added	-	-	39	37

^aIn-situ yields.

^b3a was used without further purification.

^cIsolated in 0.20 mmol scale.

We explored the scope of our alkylarylation reaction (Table 1). A wide range of electronically differentiated arylboronic acids,¹⁷ from electron-rich^17a to electron-deficient, participate in the alkylarylation to furnish the corresponding 1,1-diarylalkanes in moderate to good yields (13–30,44,47,50). Various functional groups, including halides (16,17,19,22,27–28,44,47), acid/base-sensitive moieties (18,24,26), and polyfluoroarenes (17,22,27), are tolerated under our conditions. Finally, sterically encumbered o-methylphenylboronic acid affords 20 without significant reduction in yield (45%).

Next, we investigated the alkylboronic acid scope. Secondary alkylboronic acids display excellent compatibility with the conditions and incorporate both cyclic (13,21,22,30) and acyclic (23,24,43) alkyls into corresponding alkylarylation products. The medicinally relevant piperidine is tolerated as 30 is formed in a modest yield. The addition of aniline, as an essential LB activator, overcomes the thermodynamic challenges associated with primary alkyl radical formation and allows primary alkylboronic acids to participate (25–29). Further, alkylarylation products bearing useful synthetic handles, such as terminal alkenes (27), esters (28,47), and phthalimides (29) are readily obtained.

The vinylarene scope was explored; remarkably, electronically diverse styrene derivatives, from electron-rich to electron-deficient (13,19,31–45,47), are well tolerated. A wide array of functional groups can be incorporated on the vinylarene: esters (19), halides (35–37,39,43), polyfluoroarenes (39), as well as naphthyl (13), heterocycles (40–42), and a bioactive carbocyclic motif (44). 1-(p-Chlorophenyl)-1,3-butadiene reacts to furnish vinylated 1,1-diarylalkane 43 with high trans-selectivity (>20:1 dr) and 1:1 rr. 1,1-Diaryl acetic acid derivative 45, with a new quaternary carbon, is accessed via alkylarylation albeit in 13% yield. Finally, 1,1-diphenylethylene does not afford 46; rather, elimination product, analogous to 8, is observed.

Lastly, we sought to demonstrate the synthetic utility of this transformation (Table 2c). Notably, our reaction is scalable; on 3.0 mmol of 2, 610 mg of 19 is isolated without substantial reduction in yields (51%). Furthermore, the rapid syntheses of analogues of bioactive 1,1-diarylalkanes (Scheme 1e) 49¹⁸ and 50¹⁹ were demonstrated. 49 was prepared in good yield in three steps, including alkylarylation, ester reduction, and reductive amination.

Table 2.

Substrate Scope^a,b

^a0.2 mmol scale, in-situ NMR/GC yields in parenthesis.

^bSee SI for reaction conditions.

^cFormed as a 1:1 mixture of diastereoisomers.

With the substrate scope established, we next conducted mechanistic investigations into the origins of the excellent chemoselectivity. A series of control experiments were performed accordingly and identified that both 4Å MS and TEA are essential (Figure 2a). In the absence of molecular sieves, dimerization of the radical adduct is observed as the dominant pathway: 10 is formed in 46% yield along with 36% of 4a. TEA is essential for chemoselectivity, without it a low yield of 4a (22%) is observed along with 10% of elimination product 8, 9% of diarylation product 12, and trace 10 (5%).

Figure 2.

Mechanistic Investigation

Boronic acids and boroxine anhydrides are known to be in equilibrium and both are present in commercially available boronic acids.^20,21 Molecular sieves are known to facilitate the activation of MnO₂,²² however, as both 4 and 10 are observed in their absence we speculated that under the optimized conditions the MS serve a second role - promoting the dehydrative trimerization²¹ of the boronic acids (1a,3a) into the corresponding boroxines (1b,3b). Indeed, we observe both 1b and 3b in the presence of MS by ¹H-NMR and HRMS.²³ Further, we subjected freshly recrystallized boronic acids 1a and 3a to the alkylarylation conditions with varying amounts of MS (0–120 mg) (Figure 2b). Strikingly, under the optimized conditions, with 40 mg of MS, we observe only 34% yield of 4a vs. 75% with our standard reagents (Table 1, entry 10) along with a significant amount of the dimer 10 (30% yield). To obtain a comparable ratio of 4a:10 with the free boronic acids, an excessive amount—120 mg—of MS is required. The direct correlation between increased MS loading and yield of 4a is consistent with dehydrative trimerization occurring to afford boroxines 1b and/or 3b and thereby facilitating the reaction. Notably, when pure boroxines (1b and 3b) are used only 6% of 4a is observed, the addition of 1.0 equiv of H₂O restores reactivity, suggesting that some water is essential under our reaction conditions (See SI for more details).

Next, we turned our attention to understanding the role of TEA. We hypothesize that TEA serves as a Lewis base in the generation of I and II and propose that these adducts are crucial for matching the rates of single-electron oxidation of the alkyl–[B] and transmetalation of the [B]–Ar with Cu(II). As seen in Table S1, decreasing the Lewis basicity to aniline slows transmetalation, as 4, 6, and 12 are not observed; rather, 8 and 10 along with the carboamination product are formed in 9%, 16%, and 12% yield, respectively. Likewise, employing a more basic additive, NaOEt, leads to transmetalation outcompeting oxidation, 4, 6, and 10 in 8%, 20%, and 16% yield, respectively. The optimized conditions employ TEA, which has an intermediate pK_a, thus matching the rate of single electron oxidation^3d with transmetalation.²⁴

To support the generation of I and II, we monitored the formation of the LA–LB adducts, between organoborons (1a*,1c*,3a,3b) and TEA, in various ratios by ¹¹B-NMR. Intriguingly, despite the addition of TEA in an excessive amount (10. equiv), we do not observe a significant change in the chemical shifts of free boronic acids (1a*,3a*) (see S41-42 and S44-45 in SI for more details). On the other hand, the more Lewis acidic boroxines (1c,3b) undergo an observable LA–LB interaction²⁵ in the presence of TEA as a significant up-field shift is observed (Figure 2c; S43 and 46-49).

We performed reaction profile studies with free boronic acids 1a and 3c, and styrene 2a to understand how molecular sieves affect relative rates of oxidation and transmetalation and their impact on the overall chemoselectivity (51, 10, and 52). With boronic acids acting as coupling partners in absence of MS, radical generation becomes faster than transmetalation, thus, radical dimer 10 forms rapidly, up to 46%, within 2 h. When boroxines are generated in-situ upon adding MS, on the other hand, the rates of the elementary steps are compatible with the rate of 51 formation; radical dimerization is suppressed while enhanced yields of 51 and 52 are observed.

Together, these mechanistic studies further support that the formation of boroxine intermediates is key to achieving chemoselective alkylarylation through kinetically matched radical generation and transmetalation.

We then performed mechanistic studies to confirm that selective single-electron oxidation of the alkylboron is occurring. Competition spin-trapping experiment, with TEMPO as a radical scavenger, leads to the exclusive formation of TEMPO-alkyl adduct 53 (Figure 3a). The lifetime of the alkyl radical was measured by a radical-clock experiment with 1d; both cyclized 55 and uncyclized 56 are observed in 6:1 ratio (Figure 3b). The formation of uncyclized product 56 indicates that the rate of radical addition is fast, and competitive to that of 5-exo-trig cyclization (1.0$\times$10⁵ s⁻¹ at 40 °C)²⁶. This rapid radical addition explains our high selectivity for three-component alkylarylation over the unobserved two-component cross-coupling (7) between 1 and 3. The formation of an analogous cross-coupling product 7 was observed only when an alkene partner is absent (see SI 41).

Figure 3.

Mechanistic Investigation

Lastly, we investigated mechanism of our C(sp³)–C(sp²) bond formation (Figure 3c-d). A control experiment confirms that the Cu-catalyst is essential for an arylation to occur (Figure 3c). Moreover, a modest, but significant, enantioinduction, 16% ee, with enantiomerically-enriched ligand, (S,S)-(4-F-Ph)PyBOX, is observed. This supports that the Cu catalyst mediates the C(sp³)–C(sp²) bond formation (Figure 3d).²⁷

Taking these mechanistic insights into consideration, we propose the following modified mechanism (Figure 3e) for this cross-nucleophile alkylarylation reaction: (1) an irreversible dehydrative trimerization of boronic acids occurs to generate both alkyl and arylboroxines in-situ, (2) LA–LB adducts (I’ and II’) form between boroxines and the added amines, (3) single-electron oxidation of II’ by Mn(IV) generates nucleophilic radicals III’, (4) which then add across alkenes to afford stable benzyl radicals IV’, (5) transmetalation between II’ and [Cu(II)] occurs through proposed LA-coordination²⁸ and subsequently generates boroxine byproducts and [Cu(II)]–Ar V’, (6) which reacts with IV’ to yield a formal Cu(III)-intermediate (VI’), (7) subsequent reductive elimination to afford alkylarylation product and [Cu(I)], and, finally, (8) oxidation of Cu(I) to Cu(II) occurs to close the catalytic cycle.

In conclusion, we establish the viability of a general oxidative cross nucleophile strategy for the alkylarylation of vinyl arenes employing two easy-to-handle boron nucleophiles. This efficient and modular method presents an operationally simple, versatile, and scalable approach for the synthesis of valuable 1,1-diarylalkanes. The demonstrated substrate scope includes 37 examples with an array of alkyl- and arylboronic acids and vinylarenes participating. The synthetic utility of the method is demonstrated by the preparation of Pimozide and antibreast cancer derivatives. Mechanistic studies support that key LA–LB interactions between both boroxines and the added amine is critical to matching the rates of oxidation and transmetalation and therefore achieving the desired chemoselective alkylarylation reaction.

Supplementary Material

ASSOCIATED CONTENT

Supporting Information

The Supporting Information is available free of charge on the ACS Publications website.

Experimental Procedures, Characterization, Spectral Information (PDF)