Catalytic Asymmetric Synthesis of Secondary Nitriles via Stereoconvergent Negishi Arylations and Alkenylations of Racemic α-Bromonitriles

Abstract

The first method for the stereoconvergent cross-coupling of racemic α-halonitriles is described, specifically, nickel-catalyzed Negishi arylations and alkenylations that furnish an array of enantioenriched α-arylnitriles and allylic nitriles, respectively. Noteworthy features of this investigation include: the highly enantioselective synthesis of α-alkyl-α-aryl nitriles that bear secondary α-alkyl substituents; the first examples of the use of alkenylzinc reagents in stereoconvergent Negishi reactions of alkyl electrophiles; demonstration of the utility of a new family of ligands for asymmetric Negishi cross-couplings (a bidentate bis(oxazoline), rather than a tridentate pybox); in the case of arylzinc reagents, carbon–carbon bond formation at a remarkably low temperature (−78 °C), the lowest reported to date for an enantioselective cross-coupling of an alkyl electrophile; a mechanistic dichotomy between Negishi reactions of an unactivated versus an activated secondary alkyl bromide.

Many bioactive compounds bear a cyano group attached to a stereogenic carbon.¹ Furthermore, nitriles serve as versatile precursors to a diverse array of molecules, including heterocycles, aldehydes, ketones, amides, carboxylic acids, and amines.² Enantioenriched α-alkyl-α-arylnitriles are particularly noteworthy targets, since they can be transformed into α-aryl carboxylic acids such as α-arylpropionic acids, which are widely used as non-steroidal anti-inflammatory drugs (e.g., naproxen).³

A variety of methods have been described for the catalytic asymmetric synthesis of α-alkyl-α-arylnitriles.^4,5,6 For example, nickel-catalyzed hydrocyanation of vinylarenes can generate α-methyl-α-arylnitriles with excellent ee; however, there have been no reports of corresponding highly enantioselective hydrocyanations of substituted styrenes to produce nitriles that bear α-alkyl substituents larger than methyl. The conjugate addition of H–CN to β-aryl-α,β-unsaturated carbonyl compounds provides an alternate route to α-alkyl-α-arylnitriles, although current methods generally only furnish products with α-alkyl groups that are primary.

We have been interested in exploring the catalytic asymmetric synthesis of secondary α-arylnitriles through the reaction of an aryl nucleophile with a nitrile that bears a leaving group geminal to the cyano substituent; to the best of our knowledge, there is no precedent for this transformation using a chiral catalyst.⁷ However, while our study was underway, Falck reported an important investigation of Suzuki arylations of α-(sulfonyl)nitriles in the presence of an achiral palladium catalyst.^7c In the case of two enantioenriched electrophiles (one unhindered and one doubly activated), he established that carbon–carbon bond formation proceeds with inversion of stereochemistry with little or no erosion in ee; on the other hand, Falck did not describe the stereochemical outcome for Suzuki couplings of electrophiles that bear a secondary alkyl substituent. In this report, we provide a complementary approach wherein a chiral catalyst achieves stereoconvergent Negishi cross-couplings of racemic electrophiles with aryl, as well as with alkenyl, nucleophiles (eq 1).

(1)

Negishi reactions are an especially attractive family of cross-coupling processes, due to the compatibility of organozinc reagents with a wide range of functional groups and the typical lack of a need for a stoichiometric quantity of a Brønsted-basic additive.⁸ In previous studies, we determined that several classes of activated electrophiles are suitable coupling partners in asymmetric Negishi alkylations or arylations; in every instance, nickel with a pybox ligand proved to be the catalyst of choice.^9,10 Unfortunately, however, when we applied any of the arylation methods to the phenylation of an α-bromonitrile, we obtained disappointing results (<40% ee).

We decided to look beyond tridentate pybox ligands, and after substantial effort we determined that the Negishi phenylation of an α-bromonitrile can be achieved in good ee and yield in the presence of NiCl₂•glyme and a new, readily available bis(oxazoline) ligand¹¹ (94% ee, 98% yield; Table 1, entry 1).¹² The reaction proceeds at −78 °C, which is, to the best of our knowledge, the lowest temperature that has been employed for a cross-coupling of an alkyl electrophile.¹³

Table 1.

Stereoconvergent Negishi Phenylation of a Racemic α-Bromonitrile: Effect of Reaction Parameters^a

entry	variation from the “standard” conditions	ee (%)	yield (%)b
1	none	94	98
2	no NiCl2 • glyme	–	<2
3	no L	<2	3
4	PhZnX, instead of Ph2Zn	90	6
5	r.t., instead of −78 °C	4	16
6	1, instead of L	84	89
7	2, instead of L	8	88
8	3, instead of L	8	32
9	4, instead of L	32	25
10	0.6 equiv Ph2Zn	93	48
11	5% NiCl2 • glyme, 6.5% (S,S )-L	93	68

^aAll data are the average of two experiments.

^bYield determined by GC analysis versus a calibrated internal standard.

As illustrated in Table 1, essentially no cross-coupling of the α-bromonitrile is observed in the absence of NiCl₂•glyme (entry 2) or of bis(oxazoline) L (entry 3), as well as when an arylzinc halide is employed as the nucleophilic reaction partner (entry 4). The reaction proceeds in poor ee and yield when conducted at room temperature (entry 5), and ligands other than L provide modestly (bis(oxazoline) 1; entry 6) or substantially (bis(oxazoline) 2, pybox 3, and 1,2-diamine 4 (entries 7–9)) inferior results. Use of less Ph₂Zn (entry 10) or less catalyst (entry 11) leads to a decrease in yield, but no loss in enantioselectivity.

With our optimized method, asymmetric Negishi cross-couplings of Ph₂Zn with a range of racemic α-bromonitriles proceed efficiently, furnishing α-phenylnitriles in good ee and yield (Table 2).¹⁴ A wide variety of functional groups are compatible with the reaction conditions, including an ether (entry 5), a carbamate (entry 6), an amide and a heterocycle (entry 7), a sulfonamide (entries 8 and 9), an unactivated alkyl chloride (entry 9), and an alkene (entries 11 and 12). In view of the limitations of other methods for generating highly enantioenriched α-alkyl-α-arylnitriles in the case of branched alkyl groups (e.g., via asymmetric hydrocyanation; vide supra), our high ee’s for Negishi cross-couplings of hindered electrophiles (entries 1–9) are particularly noteworthy.

Table 2.

Stereoconvergent Negishi Phenylations of Racemic α-Bromonitriles^a

entry	R1	ee (%)	yield (%) b
1	i-Pr	92	77
2		92	98
3		92	92
4c		92	92
5		92	94
6		90	96
7		85	95
8		91	94
9		90	94
10	Me	82	67 (83)d
11		78	88
12		76	94

^aAll data are the average of two experiments.

^bYield of purified product.

^cReaction temperature: −60 °C.

^dYield determined by GC analysis versus a calibrated internal standard.

This process for the Negishi arylation of α-bromonitriles is not limited to phenylation reactions. Thus, we can cross-couple with a range of nucleophiles in uniformly high ee and yield (Table 3), regardless of whether the aryl group is electron-rich (entries 1–3) or electron-poor (entry 4).¹⁵

Table 3.

Stereoconvergent Negishi Arylations of Racemic Bromonitriles^a

entry	Ar	ee (%)	yield (%) b
1		94	94
2		94	92
3		94	81
4c		93	99

^aAll data are the average of two experiments.

^bYield of purified product.

^cReaction temperature: −60 °C.

Although we have previously described examples of stereoconvergent Negishi reactions of activated electrophiles with alkyl- and arylzinc reagents, we have never been able to achieve corresponding cross-couplings with alkenylzinc reagents.⁹ We were therefore pleased to determine that we can apply the conditions (except reaction temperature¹⁶) that we developed for enantioselective Negishi arylations (Table 3) of α-bromonitriles to alkenylations (Table 4).¹⁷ Thus, as depicted in Table 4, we can accomplish an array of asymmetric alkenylations with good ee, thereby providing ready access to enantioenriched allylic nitriles.¹⁸ No racemization or isomerization (to the conjugated nitrile) of the allylic nitrile is observed under these mild conditions.

Table 4.

Stereoconvergent Negishi Alkenylations of Racemic α-Bromonitriles^a

entry	R	ee (%)	yield (%) b
1	(CH2)3Ph	80	64
2		86	59
3		89	78
4	Ph	91	94
5		92	92

^aAll data are the average of two experiments.

^bYield of purified product.

The enantioenriched allylic nitriles are suitable substrates for stereoselective functionalizations. For example, osmium-catalyzed dihydroxylation proceeds with good diastereoselectivity (eq 2).

(2)

We have previously proposed that a radical intermediate may be involved in the oxidative-addition step of certain nickel-catalyzed cross-couplings of unactivated secondary alkyl halides.¹⁹ One observation consistent with this hypothesis is the stereochemistry of cyclization/cross-couplings of electrophiles that bear pendant olefins. We have now determined that, as for Hiyama and Suzuki reactions,²⁰ Negishi reactions of secondary bromides 5 and 6 proceed with the same diastereoselectivities as observed for reductive cyclizations of these substrates by Bu₃SnH, which is consistent with a common intermediate (I) in the cyclization step for all of these processes (eq 3).

(3)

In this regard, it is interesting to note that the substrates depicted in entries 11 and 12 of Table 2, which have the potential to cyclize/cross-couple to furnish five-membered rings, afford only acyclic coupling products. To gain insight into whether cyclization of a radical intermediate might be occurring, but reversibly, in enantioselective cross-couplings of α-bromonitriles that bear pendant olefins, we examined the Negishi reaction of the electrophile illustrated in eq 4. If radical cyclization is occurring, but the acyclic radical preferentially proceeds to the cross-coupling product (Curtin–Hammett scenario), then the Z stereochemistry of the starting material would not be preserved in the acyclic product. However, we observe only the Z isomer, which suggests that radical cyclization is not occurring in this Negishi reaction of an α-bromonitrile, in contrast to our previous studies of cross-couplings of unactivated secondary alkyl bromides.²¹

(4)

In summary, we have developed the first method for the stereoconvergent cross-coupling of racemic α-halonitriles, specifically, nickel-catalyzed Negishi arylations and alkenylations that generate a variety of useful enantioenriched secondary α-arylnitriles and allylic nitriles, respectively. With regard to accessing secondary α-alkyl-α-arylnitriles, this work complements other prominent approaches, which are most effective for the synthesis of products in which the alkyl group is methyl (hydrocyanation of olefins) or a primary alkyl substituent (conjugate addition of HCN). For the first time for studies of asymmetric Negishi cross-couplings, an alkenylzinc reagent can serve as an effective nucleophilic partner, and a ligand other than a tridentate pybox is the ligand of choice; this latter observation is significant for future explorations of enantioselective Negishi reactions. Couplings with arylzinc reagents occur at a remarkably low temperature (−78 °C), the lowest temperature reported to date for such asymmetric carbon–carbon bond-forming processes. Finally, a mechanistic study provides a contrast between nickel-catalyzed Negishi reactions of unactivated vs. activated secondary alkyl bromides. An array of additional investigations of enantioselective cross-couplings of alkyl electrophiles are underway.