Rapid, Homogenous, B-Alkyl Suzuki–Miyaura Cross-Coupling of Boronic Esters

Abstract

A rapid, anhydrous Suzuki–Miyaura cross-coupling of alkylboronic esters with aryl halides is described. Parallel experimentation revealed that the combination of AntPhos, an oxaphosphole ligand, neopentyldiol alkylboronic esters, and potassium trimethylsilanolate (TMSOK) enables successful cross-coupling. In general, reactions proceed in under 1 h with good yields and high linear/branched (l/b) selectivities. Crucially, two literature examples which previously took >20 h to reach completion were accomplished in a fraction of the time with the method described herein. Mechanistic studies revealed that the reaction proceeds through a stereoretentive pathway and identified the boronic ester skeleton as a predominant pathway for deleterious protodehalogenation.

Graphical Abstract

INTRODUCTION

The venerable Suzuki–Miyaura cross-coupling (SMCC) has revolutionized C–C bond formation for synthetic chemistry.¹ Whereas it has been predominantly used for C_sp2–C_sp2 cross-coupling, within the past two decades, significant advances have been made for the analogous C_sp2–C_sp3 variant using aryl halides and alkylboronates as the coupling partners.^2,3 This type of reaction is of particular interest for the pharmaceutical industry, as potential drug candidates with greater sp³ character are considered to be more drug-like.^4–6 To accomplish these challenging cross-couplings, many early reports disclose the use of alkylboranes or boronic acids.^2a,7 More recently, the use of potassium alkyltrifluoroborate salts,^{2c,3b–e,8,9c,10a,11a–d} alkylboronic esters,^{9a,b,10b,11e–g,12–15} and MIDA alkylboronates^3f,16 has arisen as a popular strategy to circumvent the oxidative liabilities associated with alkylboronic acids or boranes. Despite these developments, a recent benchmarking study of methods for C_sp2–C_sp3 cross-coupling employing aryl halides and alkyl nucleophiles as coupling partners revealed that modern B-alkyl SMCC conditions succeeded in obtaining the cross-coupled product 61% of the time for 1° alkylboronates and 20% of the time for 2° alkylboronates, with an overall success rate of only 37%.⁶

The challenges with this important cross-coupling stem from the slow transmetalation of alkylboronates, and the ability of alkylboronates to undergo β-hydride elimination prior to transmetalation or before reductive elimination. To address these problems, many reports leverage the use of bulky, electron-rich ligands^3,7f,17 as well as the use of directing groups at proximal and distal positions, such as benzyl groups,⁹ ethers,¹⁰ carbonyl groups,¹¹ geminal 1,1- and 1,2-diboronates,¹⁴ and even hydroxides,¹⁵ to affect a variety of B-alkyl cross-couplings. Often, aqueous conditions are employed by necessity, as both potassium trifluoroborates and MIDA boronates require hydrolysis prior to transmetalation,^16a,18,19 or for solubilization of an inorganic base. Alternatively, with alkylboronic acids and esters, anhydrous conditions can be used, though these conditions are less common.^{2b,3a,7c,d,f,12,17a} Regardless of the scenario, the often biphasic nature of the reaction mixtures introduces problems related to mass transfer phenomena, resulting in irreproducible reaction kinetics and yields.

Recently, our laboratory disclosed a mechanistic study on the anhydrous, stoichiometric transmetalation of arylboronic esters with an arylpalladium(II) hydroxy dimer, demonstrating that transmetalation of boronic esters can occur without prior hydrolysis to a boronic acid.²⁰ Notably, the structure of the boronic ester greatly affected the rate of transmetalation, with glycol and dimethyl boronic esters providing significant rate increases. To translate these stoichiometric mechanistic insights to a catalytic system, anhydrous conditions must be used, as hydrolysis of the boronic ester to the boronic acid would preclude any beneficial rate enhancements from occurring. Thus, our group discovered that a rapid, homogeneous, and anhydrous C_sp2–C_sp2 SMCC could be realized by using neopentyldiol boronic esters in tandem with potassium trimethylsilanolate (TMSOK), a soluble base.²¹ Notably, in the original report, both methyl and cyclopropylboronic esters were successful coupling partners, but they required modified conditions to afford the cross-coupled products. Given the beneficial rate increases observed under anhydrous conditions with TMSOK, we sought to identify a method that could enable a rapid, homogeneous B-alkyl SMCC. Such a method could be of importance for drug discovery, as synthesis tends to be the bottleneck for the design-build-test-learn (DBTL) cycle.^6,22

RESULTS

Initial Cross-Coupling Experiments.

Given that protodehalogenation was previously observed with neopentyldiol cyclopropylboronic ester,^21a initial experiments began with THF-3,4-diol n-butylboronic ester (2a) (Scheme 1). Whereas the room temperature cross-coupling of 2a with 4-bromotoluene (1a) and t-Bu₃P–Pd-G3 failed to produce any cross-coupled product 4aa, heating the reaction mixture to reflux afforded 55% yield of 4aa. Unfortunately, reaction stalling was observed, and modulation of the aryl halide, solvent, or ester structure resulted in lower yields or significant protodehalogenation. To obtain satisfactory cross-coupling, we hypothesized that a different catalyst may be able to enable successful transmetalation of the slower alkylboronic ester while outcompeting deleterious protodehalogenation.

Scheme 1.

Initial Investigation using n-Butylboronic Ester 2a^a

^aReaction run with 0.20 mmol of aryl halide and 1.20 equiv of boronic ester. Yields determined by GC-FID analysis using biphenyl as internal standard.

Optimization of the Reaction by High-Throughput Experimentation.

Thus, a survey of 24, third-generation Buchwald precatalysts and two boronic ester scaffolds was commenced by running two 24 well plates using Radleys Greenhouse Plus Parallel synthesizer (see Supporting Information for details). The ligand set represents a broad selection of structural diversity for B-alkyl SMCC, including Buchwald-type ligands,^8,23 electron-rich phosphines, such as (Ad)₃P,²⁴ QPhos,²⁵ AntPhos,^17a,26 and BI-DIME,^17h,27 and bisphosphines.^7d,e The well plates were heated for 16 h at 100 °C combining 1a as the aryl halide and either 2a (Plate 1) or neopentyldiol n-butylboronic ester 3a (Plate 2) as the coupling partner with TMSOK in 1,4-dioxane (Figure 1A). Notably, the use of TMSOK as a base enabled preparation of stock solutions for liquid handling, decreasing the time for plate setup when compared to the laborious manual weighing of solid inorganic bases into each well.²⁸ Aliquots of the reaction mixtures were analyzed by gas chromatography and formation of 4aa, remaining 1a, and protodehalogenation of 1a were quantitated using biphenyl as an internal standard (see Supporting Information).

Figure 1.

HTE survey using four, 24 well plates. (A) First-generation screen focusing on ligand space. (B) Second-generation screen focusing on solvents and refined ligand space. Reactions run with 0.20 mmol of aryl halide and 1.20 equiv of boronic ester. Yields were determined by GC-FID analysis using biphenyl as an internal standard.

Four trends became apparent from the first-generation survey. First, whereas up to 22% of remaining 1a was found in all wells containing 2a, 1a was completely consumed in nearly every well containing 3a. Presumably, the lack of complete conversion of starting material that was observed with 2a as a coupling partner is due to the sequestration of TMSOK by a more Lewis acidic species that forms as the cross-coupling reaction proceeds. Second, as previously observed, reactions that used 3a had significantly higher amounts of protodehalogenation. Third, good l/b ratios (~20–25/1) were observed in nearly all cases with both 2a and 3a, with 2a providing significantly higher l/b ratios. Fourth, AntPhos-Pd-G3, BI-DIME-Pd-G3, and (o-Tol)₃P–Pd-G3 were among the top performing precatalysts in both plates, but, surprisingly, they gave higher yields with 3a. Additionally, Buchwald and bidentate ligands afforded similar yields between the two boronic esters; however, electron-rich ligands were uniquely effective for reactions that employed 2a.

In light of these observations, a second-generation survey with 3a, six solvents, and a refined set of seven, third-generation Buchwald precatalysts was performed using two 24 well plates (Figure 1B). In addition to the best performing ligands in the previous generation, new structural motifs that have been successful in B-alkyl SMCC were selected, such as Burke’s iPhos,^3f Tang’s bis-phosphine mono-oxide oxaphosphole (referred here as OxaPhosM),^17b and Capretta’s oxaphosphaadamantanes (CgMe-PPh(2,4-OMe)).²⁹ Whereas reactions employing DMF failed to provide 4aa in useful yields with any precatalyst, wells containing THF or 1,4-dioxane afforded 4aa (Plate 3) in comparable yields and selectivities to the first two plates, with AntPhos-Pd-G3, BI-DIME-Pd-G3, and iPhos-Pd-G3 representing the top performing systems. Curiously, the use of toluene or toluene/THF mixtures as the reaction solvent resulted in the formation of 4aa with higher yields than ethereal solvents for all precatalysts, including OxaPhosM-Pd-G3 and CgMePPh(2,4-OMe)-Pd-G3, two precatalysts that failed to significantly react in either THF or 1,4-dioxane (Plate 4). However, the heterogeneity of the reaction, induced by the poor solubility of TMSOK in toluene, precluded any further investigation of its use in the reaction. Nonetheless, both THF and 1,4-dioxane provided satisfactory results for their use in the cross-coupling.

To validate the results obtained from the screens, the top-yielding precatalysts, AntPhos-Pd-G3, iPhos-Pd-G3, and BI-DIME-Pd-G3, were subjected to further investigation in THF (Scheme 2A). Whereas the reaction mixtures containing iPhos and BI-DIME had significantly unconsumed 1a at 30 min, the reaction containing AntPhos had completely consumed 1a, affording 4aa in 91% yield. With conditions in hand, a control experiment was performed using n-butylboronic acid (5a) as the coupling partner (Scheme 2B). Critically, only a trace amount of product was observed, demonstrating the rate enhancements obtained with the boronic ester scaffold under anhydrous conditions. Finally, brief optimization of the base loading (1.40 equiv) and concentration (0.30 M) further increased the rate of the reaction, providing 92% yield of 4aa in 15 min (Scheme 2C).

Scheme 2.

Post-HTE Experiments^a

^aReaction run with 0.20 mmol of aryl halide and 1.20 equiv of boronic ester. Yields determined by GC-FID analysis using biphenyl as internal standard.

Evaluation of Reaction Scope.

Using these optimized conditions, the scope of the aryl halide structure was explored with boronate 3a (Table 1A). Most reactions were complete within 15–20 min and the desired products were obtained with good to excellent yields and high l/b selectivities (38 to 96%, l/b, 13:1 to >99:1). Whereas products formed from electron-neutral and electron-poor arenes (4aa–4ga) were obtained rapidly (<25 min) in generally good yields, products from electron-rich substrates were formed slowly, often requiring extended reaction times or elevated temperatures (4ha–4ja). Notably, products containing an acidic N–H, such as aniline 4ka, could be successfully cross-coupled in good yield using an additional equivalent of TMSOK. Products 4la and 4ma derived from (E)-2-bromostyrene and mesityl bromide, respectively, were formed in good yields and excellent l/b selectivities. The preparation of 1-pentylbenzene (4na), arising from the corresponding benzyl chloride, required elevated temperature and catalyst loading, but nevertheless proceeded in 38% and 13:1 l/b selectivity. Finally, products 4oa and 4ap containing simple heterocycles could be obtained in moderate yields, though higher catalyst loadings for 4oa and temperatures were required.

Table 1.

Reaction Scope of Aryl Halides and Neopentyldiol Boronic Esters^a

^aReactions run with 1.00 mmol of aryl halide and 1.20 equiv of boronic ester. The products are labeled accordingly with the numbering 4xy, where x = the letter of the corresponding halide, and y = the letter of the corresponding boronic ester. l/b = linear/branched ratio. Yields of isolated product after column chromatography.

^b1,4-Dioxane used at 100 °C.

^c2.4 equiv of TMSOK used.

^d4 mol % catalyst used.

^eBenzyl chloride used as the coupling partner.

^fProduct was further purified to analytical purity; see the Supporting Information.

Next, the scope of alkylboronic esters that could undergo cross-coupling with 2-bromonaphthalene (1c) was explored (Table 1B). In all examples, the coupling products were obtained rapidly and in good yields (50–94%) with perfect l/b selectivities. Products derived from methylboronic- and cyclopropylboronic esters (4cb and 4cc) were formed smoothly. Notably, the use of THF-3,4-diol boronic ester was not required for successful formation of 4cc.^21a Interestingly, 4cd was produced in good yield despite the increased steric hindrance introduced from the β-branched methylcyclohexylboronic ester. Products obtained from activated substrates such as benzylboronic ester 4ce could be successfully formed at elevated temperatures. In addition, products containing functional groups such as alkyl chlorides, silyl ethers and acetals were obtained in good yields (4cg, 4ch, 4cj). Notably, attempts to form 4ci containing a tert-butyl ester failed at 70 °C, presumably owing to formation of a palladium enolate that coordinates to the neopentyldiol boronic ester. However, running the reaction at 100 °C in 1,4-dioxane successfully enabled the production of 4ci. Finally, to demonstrate the method can produce more nitrogen-containing targets, 4qd and 4re⁶ were obtained in good yields from coupling with 3e and 3d, respectively.

To demonstrate the advantages of using anhydrous conditions with boronic esters, two comparisons to literature procedures were executed. First, using anhydrous conditions with TMSOK, 4re was formed from neopentyldiol boronic ester 3e in 53% yield after 5 min. In contrast, 4re had been previously prepared in 53% yield after 72 h using a SMCC with potassium benzyltrifluoroborate and aqueous cesium carbonate.⁶ Similarly, using potassium phenethyltrifluoroborate and aqueous potassium carbonate, 4sk had been previously synthesized using an SMCC in 74% yield after 16–20 h.³⁰ Using the method described herein, 4sk was obtained in 70% yield after only 20 min. Thus, preparation of both 4re and 4sk highlights the rate enhancements obtained with a soluble base and a boronic ester under anhydrous conditions.

DISCUSSION

Stereospecificity of the Cross-Coupling.

Finally, two sets of mechanistic experiments were performed to investigate the stereospecificity of the cross-coupling with a 2° boronic ester and the origin of protodehalogenation. First, two stereodefined cyclopropylboronic esters, trans-3l³¹ and cis-3l,³² were prepared as a single diastereomer by cyclopropanation of the corresponding (Z)- and (E)-styrenes. The diastereospecificity (ds)³³ of the reaction was determined by using the optimized coupling conditions with 1c. Respectively, products trans-4cl and cis-4cl were obtained in 90% and 94% ds, suggesting that the reaction proceeds through a highly stereoretentive process (Scheme 3).³⁴

Scheme 3.

Stereospecificity of the Coupling Reaction

^aReactions run with 1.00 mmol of aryl halide and 1.20 equiv of boronic ester. Yields of isolated product after column chromatography. ^bProduct was further purified to analytical purity; see the Supporting Information.

Origin of Protodehalogenation.

Second, the origin of protodehalogenation was investigated by deuterium labeling studies. In analogy to other palladium-catalyzed cross-couplings,³⁵ this side reaction was hypothesized to occur by reductive elimination of an arylpalladium(II) hydride (8) (Figure 2A). Consequently, three potential pathways were identified in which 8 could be generated from an arylpalladium(II) halide (7) (Figure 2B). In path A, β-hydride elimination could occur from the backbone of the boronic ester skeleton, such as coordination of 7 to an alkoxide, such as 9, resulting from fragmentation of the neopentyldiol boronic ester skeleton upon binding of TMSOK. In path B, β-hydride elimination could proceed from the solvent, as in THF adduct 10 or 1,4-dioxane adduct 11.³⁶ In path C, β-hydride elimination could occur from the alkyl chain, such as alkylpalladium(II) complex 12. Initially, investigation of path A and B was chosen (Figure 2C). Thus, neopentyldiol n-butylboronic ester 3a-d₄³⁷ was subjected to the standard reaction conditions with 2,4,6-triisopropylbromobenzene (1s), a substrate for which protodehalogenation greatly outcompeted cross-coupling. After heating in 1.4-dioxane-h₈ at 100 °C for 25 min, 13-h₁/d₁ was obtained with 77% deuterium incorporation. Executing the same experiment with 3a-h₄ in 1,4-dioxane-d₈ resulted in formation of 13-h₁/d₁ with only 2% deuterium incorporation. Together, these results suggest that reduction of the arene predominantly occurs by Path A.

Figure 2.

(A). Plausible mechanism for the origin of protodehalogenation. (B). Hypothetical pathways for the source of the hydride. (C). Deuterium labeling experiments. Reactions run with 0.25 mmol of aryl halide and 1.20 equiv of boronic ester. Yields of isolated product after column chromatography. Deuterium incorporation was determined from mass spectrometry; see the Supporting Information for details.

CONCLUSION

In conclusion, a rapid, homogeneous, B-alkyl SMCC has been described, providing a variety of products in 1 h or less. Notably, parallel experimentation involving the use of a soluble base, TMSOK, enabled discovery of a precatalyst and boronic ester that could successfully perform cross-coupling while preventing reaction stalling and outcompeting deleterious protodehalogenation. Mechanistic studies revealed that transmetalation proceeds through a stereoretentive pathway, and the boronic ester skeleton was revealed to be predominantly responsible for the origin of protodehalogenation. Owing to the homogeneity of the reaction, this method could be exceptionally useful for applications in library synthesis and high-throughput experimentation. Additionally, we expect that the rate enhancements for transmetalation provided by alkylboronic esters under anhydrous conditions may enable new reactivity for other metal-catalyzed transformations.

Data Availability Statement

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