Enantiospecific, nickel-catalyzed cross-couplings of allylic pivalates and arylboroxines.

We have developed an enantiospecific, nickel-catalyzed cross-coupling of unsymmetric 1,3-disubstituted allylic pivalates with arylboroxines. The success of this reaction relies on the use of BnPPh2 as a supporting ligand for the nickel(0) catalyst and NaOMe as a base. This method shows excellent functional group tolerance and broad scope in both the allylic pivalate and arylboroxine, enabling the preparation of 1,3-diaryl allylic products in high yields with excellent levels of regioselectivity and stereochemical fidelity.

Transition-metal-catalyzed allylic substitution provides an important method of C–C bond formation.[1] In particular, enantioselective or -specific arylation of readily accessible 1,3-disubstituted secondary allylic electrophiles enables facile construction of enantioenriched products equipped with vinyl-substituted benzylic carbon stereocenters. These products are useful both as synthetic intermediates and for their potential biological activity.[2] An ideal cross-coupling method to deliver these valuable enantioenriched products would enable high levels of enantiospecificity (or enantioselectivity) and regioselectivity, wide functional group tolerance, and the use of an air-stable coupling partner (such as an arylboronic reagent).[3,4] Toward these goals, Rh- and Pd-catalyzed cross-couplings of allylic electrophiles and arylboronic acids have been developed to deliver highly enantioenriched products.[5−10] However, the additional goal of developing nonprecious metal catalysts is also important in lowering the cost, ensuring adequate supplies of catalysts for large-scale applications, and reducing the environmental impact of this chemistry.[11] In contrast to the methods developed with precious metal catalysts, the use of nonprecious metal catalysts for enantiospecific or -selective allylic arylations is less developed and often requires use of air-sensitive aryl nucleophiles that may have incompatibilities with some functional groups. For example, formation of enantioenriched products via Cu-catalyzed allylic arylations largely relies on the use of Grignard, organozinc, or organoaluminum reagents;[12−16] only a limited number of methods allow the use of arylboronates.[17−19]

Based on our studies of enantiospecific, Ni-catalyzed couplings of benzylic pivalates and arylboroxines,[20] we envisioned that Ni-based catalysts may also serve as efficient, nonprecious metal catalysts for highly enantiospecific and regioselective couplings of 1,3-disubstituted allylic pivalates with aryl boroxines. Prior art in this area has indeed demonstrated that nickel catalysts efficiently catalyze the cross-coupling of allylic electrophiles with aryl Grignard reagents, with both diastereo- and enantioselective variants reported.[21−23] Trost and Kobayashi have also shown that more functional group tolerant borate and boronic acid nucleophiles can be used in cross-couplings with allylic amines and carbonates, respectively.[24,25] With respect to the stereochemical outcome of these reactions, they have demonstrated that the arylation of cyclic allylic electrophiles occurs with inversion of configuration, and Kobayashi has exploited this in the arylation of cyclopentene diol derivatives to deliver enantioenriched products (Scheme 1A).[24a−24e,25] In addition, Uemura has reported a moderately enantioselective, Ni-catalyzed coupling of symmetric and terminal allylic acetates with arylboronic acids (Scheme 1B).[26] However, to our knowledge, there are no reports of enantiospecific, Ni-catalyzed cross-couplings of acyclic allylic electrophiles and arylboron reagents to deliver highly enantioenriched products.

Scheme 1

Enantioselective and Enantiospecific Nickel-Catalyzed Allylic Arylation with Arylboron Reagents

Herein we report the first enantiospecific, Ni-catalyzed cross-coupling of acyclic, 1,3-disubstituted allylic pivalates and arylboroxines (Scheme 1C). This method enables efficient transformation of readily available, highly enantioenriched allylic pivalates to valuable 1,3-diaryl allylic products in good yields with excellent regioselectivities and levels of stereochemical fidelity.

Our initial investigations focused on the coupling of pivalate 1a, easily prepared in 96% ee via a CBS reduction/pivalation sequence, with phenylboroxine (Table 1). Under optimal conditions for the coupling of benzylic pivalates [Ni(cod)2, no additional phosphine ligand, PhMe], desired coupling product 2 was obtained in excellent yield. However, only 18% ee of product was observed (entry 1). Changing the solvent to MeCN resulted in higher enantiomeric excess (ee) of the product (57% ee) in similar yield (entry 2). In an effort to improve the stereochemical fidelity further, we investigated phosphine ligands to support the Ni catalyst. Although bidentate phosphine ligands led to poor ee’s of product (entries 3, 4), the use of monodentate phosphine ligands resulted in significant improvement (entries 5–7). Among the ligands examined, benzyl diphenyl phosphine (BnPPh2) proved to be the best (entry 7). Reducing the reaction temperature gave an increased yield (entry 8). Further increases in yield and ee were obtained when the catalyst loading was lowered to 2 mol % (entries 9, 10). The observation of higher levels of stereochemical fidelity at lower catalyst loadings may suggest a Ni-mediated epimerization pathway.[28] Under these optimized conditions, the use of phenylboronic acid in place of phenylboroxine led to a lower yield but with the same ee (entry 11). The reason for this difference is currently unclear, but does not seem to be due to the reaction’s sensitivity to water. Addition of water to the reaction resulted in only a slightly reduced yield (entry 12).

Table 1

Optimization of Reaction Conditionsa

entry	mol % [Ni]	ligand (mol %)	temp (°C)	yieldb (%)	eec (%)
1d	10	none	50	95	18
2	10	none	50	96	57
3	10	dppp (11)e	50	90	57
4	10	XantPhos (11)e	50	30	27
5	10	PCy3 (22)	50	80	84
6	10	CyPPh2 (22)	50	83	85
7	10	BnPPh2 (22)	50	85	87
8	10	BnPPh2 (22)	rt	90	87
9	5	BnPPh2 (11)	rt	90	92
10	2	BnPPh2 (5)	rt	92	94
11f	2	BnPPh2 (5)	rt	72	93
12g	2	BnPPh2 (5)	rt	86	94

Conditions: pivalate 1a (96% ee, 0.1 mmol, 1.0 equiv), (PhBO)3 (1.0 equiv), Ni(cod)2, ligand, NaOMe (2.0 equiv), CH3CN (0.25 mL, 0.4 M), 4 h, unless otherwise noted.

Determined by 1H NMR analysis using 1,3,5-trimethoxybenzene as internal standard.

Determined by HPLC analysis using a chiral stationary phase.

PhMe instead of CH3CN.

dppp = 1,3-bis(diphenylphosphino)propane. XantPhos = 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene.

PhB(OH)2 (3.0 equiv).

H2O (1.5 equiv) added.

Under our optimized conditions (Table 1, entry 10), we observed a broad scope with respect to the arylboroxine (Scheme 2).[27] In all cases in Schemes 2 and 3, excellent regioselectivity and E/Z selectivity were observed; only α-aryl products with E-olefins were formed. High yields and excellent levels of stereochemical fidelity were obtained with arylboroxines bearing either electron-donating or -withdrawing groups (2–13). A wide range of functional groups were tolerated including ether (3), thioether (4), vinyl (5), halide (6–8), trifluoromethyl (9), ester (10), ketone (11), nitrile (12), and acetal (13) groups. In particular, the vinyl, bromide, chloride, and thioether groups highlight the mildness of these reaction conditions; undesired Ni-catalyzed reactions of these groups do not occur under our conditions. Notably, the Ni(0) catalyst generated in situ from NiBr2, BnPPh2, and Zn was also effective in this arylation, providing product 2 in the same ee as with Ni(cod)2, albeit in somewhat reduced yield (eq 1). These conditions enable a benchtop setup of the arylation reaction.

Scheme 2

Scope of Arylboroxine

Conditions: pivalate 1a (96% ee, 0.3 mmol, 1.0 equiv), (PhBO)3 (1.0 equiv), Ni(cod)2 (2 mol %), BnPPh2 (5 mol %), NaOMe (2.0 equiv), CH3CN (0.75 mL, 0.4 M), rt, unless otherwise noted. Average yields (±3%) and ee’s (±1%) of isolated products of duplicate reactions. Ee determined by HPLC analysis using a chiral stationary phase. es = (ee of product)/(ee of starting material). 1a (95% ee).

Scheme 3

Scope of Pivalates

Conditions: pivalates 1b–1i (0.3 mmol, 1.0 equiv), (PhBO)3 (1.0 equiv), Ni(cod)2 (2 mol %), BnPPh2 (5 mol %), NaOMe (2.0 equiv), CH3CN (0.75 mL, 0.4 M), rt, unless otherwise noted. Average yields (±3%) and ee’s (±1%) of isolated products of duplicate reactions. Ee determined by HPLC analysis using a chiral stationary phase. es = (ee of product)/(ee of starting material); corresponding starting materials and their ee’s in parentheses. Single experiment.

With respect to the allylic pivalate, pivalates with both electron-rich and -deficient aryl substituents efficiently underwent arylation (Scheme 3, 14–20). As had been true of the arylboroxine, aryl chlorides are also well tolerated on the allylic pivalate, again highlighting the orthogonality of this reaction to aryl chlorides (17–19). In addition to the functional groups already demonstrated on the arylboroxine, a tertiary amine was well tolerated (18). Allylic pivalates with naphthyl and heteroaryl substituents were also successfully coupled (21, 22). In terms of the alkyl substituent (R) on the allylic pivalate, increased steric hindrance did not diminish reactivity, but lower levels of stereochemical fidelity were observed with cyclohexyl-substituted 1i (23, 24). The X-ray crystallographic analysis of product 17 shows that its absolute configuration is S,[29] which indicates that the reaction proceeds with inversion of configuration. To confirm further the stereochemical outcome and demonstrate the utility of this arylation method, compound (R)-12, prepared in 95% ee via the arylation of (S)-1a, was converted to (S)-ketoprofen (26), an anti-inflammatory drug, via a two-step manipulation (Scheme 4). Comparison of the optical rotation of 26 to previously reported values confirms its absolute configuration is S.[30] The absolute configurations of other products were assigned by analogy.

Scheme 4

Synthesis of (S)-Ketoprofen

Consistent with previous Ni-catalyzed allylations, we propose that this reaction proceeds through a π-allylnickel intermediate.[21−26,31] To test this hypothesis, we subjected pivalate 27 (96% ee), a regioisomer of 1a, to the reaction conditions. Product (R)-2 was formed in 81% yield and 83% ee with inversion of the absolute stereochemistry (Scheme 5). The reason for the mild erosion of ee is unclear, but may be due to epimerization of a faster formed π-allylnickel intermediate. The observation that both regioisomers of starting material (1a and 27) lead to the same product (2) supports the intermediacy of an η3-allylnickel complex. Also consistent with an η3-allylnickel intermediate, the arylation of 1,3-diaryl-substituted 28 resulted in a 1:1.5 ratio of regioisomers 29 and 30 (Scheme 5). Despite the modest regioselectivity, it is notable that 30 was obtained with 91% es.

Scheme 5

Regioselectivity Studies

In conclusion, we have developed an enantiospecific, Ni-catalyzed cross-coupling of enantioenriched secondary allylic pivalates and arylboroxines to deliver 1,3-diaryl allyl products. The method features mild conditions that allow broad functional group tolerance on both the allylic pivalate and arylboroxine. In all cases, good yields and excellent levels of regioselectivity, E/Z selectivity, and stereochemical fidelity were observed. In addition, air-stable, less expensive NiBr2 can be used as an alternative to Ni(cod)2. Studies exploring other substrate classes as well as details of the mechanism are currently ongoing and will be reported in due course.