2017, Vol.46, No.9

A two-step transformation of α,β-unsaturated carboxylic acids to alkenylboronic esters has been achieved via thioesterification of carboxylic acids followed by rhodium-catalyzed decarbonylative borylation of thioesters. The latter reaction proceeds under mild conditions with broad functional-group tolerance, rendering a wide range of alkenylboronic esters easily accessible.

Alkenylboronic esters are useful synthetic building blocks that can be applied to a wide range of fields, including medicinal chemistry and materials science.1 Transition-metal-catalyzed hydroboration of alkynes is one of the most common methods for regio- and stereoselective preparation of alkenylborons (Figure 1A).1,2 Transition-metal-catalyzed borylation of (pseudo)haloalkenes, generally called Miyaura borylation, is also an efficient method for this purpose.3 However, available alkenylborons by conventional methods are limited and an alternative method with a broad scope of substrates is eagerly awaited. In this context, borylative transformation of a ubiquitous functional group bound to alkenes is a preferable approach. Herein, we report a novel synthetic method for alkenylboronic esters via formal decarboxylative borylation of α,β-unsaturated carboxylic acids, achieved via thioesterification of carboxylic acids followed by rhodium-catalyzed decarbonylative borylation of thioesters (Figure 1B).4

During the course of our recent studies on borylative cleavage of stable chemical bonds,5 we developed a two-step procedure for formal decarboxylative borylation of aromatic carboxylic acids.5c,6 This transformation includes thioesterification of carboxylic acids followed by rhodium-catalyzed decarbonylative borylation of thioesters feasible under mild conditions. Notably, the latter reaction proceeded at 80 °C and showed excellent tolerance to functional groups compared with other reported reactions of decarbonylative borylation of carboxylic acid derivatives such as amides and esters, which were performed at elevated temperatures (typically >150 °C).6a6c We anticipated that application of this method to α,β-unsaturated thioesters would allow for facile synthesis of diverse alkenylborons for several reasons. First, the starting materials, α,β-unsaturated carboxylic acids, are readily available; many are commercial and frequently found in pharmaceuticals and natural products.7 Moreover, recent development of carboxylation reactions has greatly increased the availability of α,β-unsaturated carboxylic acids.8 However, since our initial attempt of using (E)-thiocinnamic acid S-ethyl ester (2a) under the optimized conditions for aromatic thioesters5c afforded the desired product 3a in only a moderate yield (eq 1), we explored better conditions for the borylation of 2a.


Extensive screening of the reaction conditions revealed that borylation of 2a proceeded under considerably milder conditions than the conditions required for the borylation of aromatic substrates; heating a mixture of 2a, bis(pinacolato)diboron ((Bpin)2, 2 equiv), [Rh(OH)(cod)]2 (5 mol %), and PEt3 (30 mol %) in diethyl ether (Et2O) in a capped vial at 50 °C for 6 h afforded 3a in a high yield (Table 1, Entry 1). Of note, the reaction temperature was lowered to 50 °C and the use of base was avoided. Similar to the aromatic system,5c using a sterically unhindered trialkylphosphine ligand such as PEt3 was crucial for achieving an efficient transformation (Entries 1–3); other ligands including bulky trialkylphosphines (Entry 4), triarylphosphines (Entry 5), or bidentate phosphines gave poor results (Table S1). Reactions using 30 mol % of PEt3 (Rh:P = 1:3) showed the best result (Table S2). The use of rhodium methoxide or acetate complexes instead of [Rh(OH)(cod)]2 afforded similar results, whereas the use of [RhCl(cod)]2 was ineffective (Table 1, Entries 6–8), suggesting that the presence of a catalytic amount of base is essential to promote this reaction (Table S3). A variety of nonpolar aprotic solvents were available; ethereal solvents such as Et2O, cyclopentyl methyl ether (CPME), and tetrahydrofuran (THF) were particularly preferable (Table S4). The amount of [Rh(OH)(cod)]2 could be reduced to 1 mol %, although prolongation of the reaction time from 6 to 24 h was needed to obtain 3a in a reasonable yield (Table 1, Entry 9 and Table S5). The reaction was scalable without further optimization as demonstrated in a 4 mmol-scale experiment (Table 1, Entry 10). The amount of (Bpin)2 could be reduced to 1.2 equiv to obtain 3a in an acceptable yield, whereas the reaction using a substoichiometric amount of (Bpin)2 (0.7 equiv) afforded 3a in 47% yield, indicating that only one of the boron atoms in (Bpin)2 is utilized in the reaction (Table S6). The borylation of 2a with bis(neopentyl glycolato)diboron instead of (Bpin)2 under optimal conditions (Table 1, Entry 1) did not proceed and 2a was recovered almost quantitatively. Although the use of phenyl thioester 2b resulted in low conversion, the borylation of n-dodecyl thioester 2c gave a result comparable with that of ethyl thioester 2a, rendering the generation of a low-molecular-weight thiol compound with an unpleasant odor avoidable (eq 2).9


Table 1. Optimization of the reaction conditions
Table 1. Optimization of the reaction conditions
Entry [Rh]2 Ligand Yield/%a
1 [Rh(OH)(cod)]2 PEt3 86 [73]b
2 [Rh(OH)(cod)]2 PMe3 80
3 [Rh(OH)(cod)]2 P(n-Bu)3 70
4 [Rh(OH)(cod)]2 PCy3 0
5 [Rh(OH)(cod)]2 PPh3 0
6 [Rh(OMe)(cod)]2 PEt3 79
7 [Rh(OAc)(cod)]2 PEt3 72
8 [RhCl(cod)]2 PEt3 1
9c [Rh(OH)(cod)]2 PEt3 64
10d [Rh(OH)(cod)]2 PEt3 [82]b

aYields determined by GC analysis, unless otherwise noted. bIsolated yields in brackets. cThe reaction was performed using 1 mol % of [Rh(OH)(cod)]2 for 24 h. d769 mg (4.00 mmol) of 2a was used.

The optimized conditions (Table 1, Entry 1) were applicable to decarbonylative borylation of a wide range of α,β-unsaturated thioesters 210 (Table 2). Derivatives of 2a with an electron-donating para-substituent on the aromatic ring, such as 2d2h, or those with an electron-withdrawing para-substituent, such as 2i and 2j, were smoothly transformed into the corresponding alkenylboronates 3d3j in moderate to excellent yields. Notably, substrates bearing a protic hydroxy or amino group, such as 2f2h and 2m, were applicable, demonstrating the excellent functional-group tolerance of this method. Heteroaromatic-containing substrates such as thiophene and indole derivatives 2k2m were also transformed to their corresponding alkenylboronates 3k3m in good yields. α,β-Unsaturated thioesters 2n2q bearing a group sensitive to low-valent transition metals, such as a halo or tosyloxy group, also participated in this reaction to furnish borylated (pseudo)halostyrenes 3n3q, respectively,11 from which a sequential transformation is feasible. The borylation of 2r with a methylsulfanyl group proceeded smoothly to obtain the borylated compound 3r. During this reaction, neither the possible rhodium-catalyzed borylation at the ortho-position to the sulfenyl group12 nor borylative cleavage of the C(aryl)–S bond5b was observed, showing the high chemoselectivity of this transformation. Excellent chemoselectivity was also observed in the competition reaction between α,β-unsaturated thioester 2a and aryl thioester 4 under the optimized conditions for α,β-unsaturated thioesters; 2a was completely consumed to afford alkenylboronate 3a, while 4 was left almost untouched (eq 3).


Table 2. Decarbonylative borylation of various α,β-unsaturated thioesters 2a
Table 2. Decarbonylative borylation of various α,β-unsaturated thioesters 2a

aIsolated yields are shown. bYields for the reactions conducted for 24 h. cThe reaction was conducted in THF at 80 °C. dYields for the reactions conducted under the optimized conditions for the borylation of aromatic thioesters (eq 1, ref 5c) are shown in parentheses. eThe reaction was conducted for 48 h. fThe reaction was conducted for 96 h. gThe reaction was conducted for 18 h.

Cinnamic thioesters bearing a substituent on the benzene ring at a position other than para were also applicable (Table 2). For example, borylation of meta- or ortho-methyl substrates took place to afford desired products 3s and 3t in high yields. Moreover, α,β-unsaturated thioesters 2u2w bearing a methyl or methoxy group at either of the olefinic carbons also participated in this reaction to afford trisubstituted alkenylboronates 3u3w. In these cases, the substituents retarded the reactions likely because of their steric hindrance. To obtain the products, the reaction had to be conducted at a higher temperature with prolonged reaction times. For decarbonylative borylation of these sterically hindered substrates 2u2w, the optimal conditions for aromatic thioesters5c gave better results, affording 3u3w in higher yields. These alkenylborons, particularly 3w that serves as an acylboron equivalent,13 are difficult to prepare by conventional methods because of the low availability of the corresponding starting materials, clearly demonstrating the utility of our method that uses readily available α,β-unsaturated carboxylic acids as starting materials. The reaction of (E)-β-n-hexylacrylic thioester 2x also proceeded by extending the reaction time; further heating did not improve the yield likely because of the low stability of 3x under the reaction conditions. Throughout the studies on decarbonylative borylations, we have not observed any byproducts formed via 1,4-addition of a boryl group14 or a thiolate.

We currently consider that decarbonylative borylation of α,β-unsaturated thioesters 2 proceeds via a mechanism similar to that of aromatic thioesters;5c the decarbonylation occurs via acylrhodium(I) intermediate III that is generated in situ by transmetalation of rhodium(I) hydroxide I with (Bpin)2 followed by oxidative addition of 2 (Scheme 1). After decarbonylation affording the alkenyl(boryl)rhodium(III) species IV, reductive elimination occurs to form alkenylboronate 3 with the regeneration of I. Liberation of carbon monoxide was confirmed by its detection using a gas detector tube (Figure S1), which supports this mechanism. Moreover, a competition reaction between 2e and 2i showed that α,β-unsaturated thioesters with an electron-deficient olefin moiety are preferred for borylation (Scheme S1), which is consistent with the borylation of aromatic substrates.5c In addition, borylation of 2a was not inhibited by the addition of a radical scavenger such as 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO), indicating that a radical species is not involved in this reaction (Table S7).

Interestingly, decarbonylative borylation of α,β-unsaturated thioesters 2 proceeded without the addition of a base, whereas the use of a catalytic amount of a base, such as potassium acetate, was needed to achieve efficient decarbonylative borylation of aromatic thioesters.5c Our previous work suggested that the acetate anion contributed to the regeneration of the active catalyst from the deactivated rhodium complex. Considering that α,β-unsaturated thioesters 2 have an electron-deficient alkene that can easily coordinate with an electron-rich rhodium center,15,16 we anticipate that 2 serves as a transient ligand to regenerate the catalytic activity, enabling decarbonylative borylation under base-free conditions.

Our hypothesis that α,β-unsaturated thioester 2 serves as a transient ligand was also supported by the promotion of decarbonylative borylation of aromatic thioester 4 in the presence of an electron-deficient alkene instead of potassium acetate (Scheme 2A). We previously reported that the use of polycarbonyl rhodium(I) complex ([Rh(SEt)(CO)2]2) alone did not catalyze decarbonylative borylation of 4, although the reaction proceeded smoothly in the presence of a catalytic amount of potassium acetate to quantitatively give borylarene 5.5c We found that the addition of an equimolar amount of 4-fluorostyrene partly recovered the catalytic activity of the rhodium(I) complex to afford 5.17 In contrast, decarbonylative borylation of 2a was partially promoted by [Rh(SEt)(CO)2]2 even under base-free conditions (Scheme 2B and Table S3, Entry 9), indicating that electron-deficient alkenes contributed to the recovery of the activity of the rhodium catalyst. Although a detailed mechanism for the regeneration of the catalytic activity is unclear, we assume that α,β-unsaturated thioester 2 serves to accelerate ligand exchange from carbon monoxide to a phosphine via reversible coordination to the rhodium center (Scheme 2C).

In summary, we have developed a synthetic method for alkenylboronic esters via thioesterification of α,β-unsaturated carboxylic acids and subsequent rhodium-catalyzed decarbonylative borylation that proceeds under mild conditions with a broad functional-group tolerance. Since a number of carboxylic acids are easily available, this method expedites the development of valuable molecules via well-established organoboron chemistry. Further investigations, including detailed mechanistic studies and application of the method to densely functionalized carboxylic acids, are currently underway in our group.

This work was supported by Platform for Supporting Drug Discovery and Life Science Research funded by Japan Agency for Medical Research and Development (AMED) (T.N.), JSPS KAKENHI Grant Numbers 15K05509 (C; T.N.), and Special Postdoctoral Researcher Program from RIKEN (H.O.).

Supporting Information is available on http://dx.doi.org/10.1246/cl.170549.

This paper is dedicated to Professor Teruaki Mukaiyama in celebration of his 90th birthday (Sotsuju).

T. Niwa

T. Hosoya

H. Ochiai

M. Isoda