Palladium and Lewis-Acid-Catalyzed Intramolecular Aminocyanation of Alkenes: Scope, Mechanism, and Stereoselective Alkene Difunctionalizations

Abstract

An expansion of methodologies aimed at the formation of versatile organonitriles, via the intramolecular aminocyanation of unactivated alkenes, is herein reported. Importantly, the need for a rigid tether in these reactions has been obviated. The ease-of-synthesis and viability of substrates bearing flexible backbones has permitted for diastereoselective variants as well. We demonstrated the utility of this methodology with the formation of pyrrolidones, piperidinones, isoindolinones, and sultams. Furthermore, subsequent transformation of these motifs into medicinally relevant molecules is also demonstrated. A double crossover ¹³C-labeling experiment is consistent with a fully intramolecular cyclization mechanism. Deuterium labeling experiments support a mechanism involving syn-addition across the alkene.

Graphical Abstract

1. INTRODUCTION

Nitriles are versatile building blocks in organic chemistry found in pharmaceuticals, cosmetics, and organic materials.¹ Recent advances in cyanation reactions using nonmetallic cyanating agents hold promise for minimizing the environmental concerns associated with metal cyanides.² The DuPont adiponitrile process represents a classic industrial application of the catalytic hydrocyanation reaction of alkenes.³ However, compared with the well-established cyanide addition reactions to polar functional groups,⁴ cyanide addition reactions to nonpolar C–C multiple bonds, such as alkenes and alkynes, are substantially more challenging.⁵

Recently, new catalytic approaches for accessing functionalized nitriles by cyanofunctionalization reactions of alkenes (or alkynes) have been discovered. The addition of a C–CN or heteroatom–CN σ-bond across a nonpolar C–C π-bond installs new C–CN bonds in conjunction with vicinal C–C,⁶ C–B,⁷ C–Si,⁸ C–Ge,⁹ C–Sn,¹⁰ C–Br,¹¹ C–S,¹² C–O,¹³ or C–N^14,15 bonds in a single operation. Nakao’s group reported palladium/Lewis-acid-catalyzed intramolecular oxycyanation^13a and aminocyanation^14a of alkenes through the activation of O–CN bonds of cyanates and N–CN bonds of cyanamides, respectively. These transformations, including a preliminary enantioselective variant, enable rapid construction of substituted dihydrobenzofuranes, indolines, and pyrrolidines. We independently discovered Rh(I)/BPh₃-catalyzed and B(C₆F₅)₃-catalyzed aminocyanation of alkenes to access indolines and tetrahydroquinolines (Scheme 1, part a).^14b For our metal free process, a recent computational study postulated a concerted asynchronous alkene addition mechanism.^14c Very recently, Shi and co-workers demonstrated an oxycyanation of methylenecyclopropanes triggered by N–CN bond cleavage, leading to the formation of benzo[d][1,3]oxazines.^13b

Scheme 1.

Synthesis of Lactams and Sultams by N–CN Bond Activation and Aminocyanation of Alkenes

Despite these advances, current aminocyanation methods are typically restricted to constructing heterocycles containing rigid aromatic backbones (e.g., indolines and tetrahydroquinolines) or requiring activated alkenes (e.g., methylenecyclopropanes). We envisioned that the cleavage of N–CN bonds of N-acyl and N-sulfonyl cyanamides (1), followed by subsequent addition across an unactivated alkene could take place without the templating effect of an aromatic ring. Further, the N-acyl or N-sulfonyl group, if placed in the forming ring, would enable the synthesis of substituted lactams or sultams bearing a nitrogen-substituted quaternary stereocenter (2, Scheme 1b).¹⁶ Unexplored in previous studies, these heterocycles are valuable synthetic targets found in medicinally relevant molecules, such as isoindolinone JM-1232 (a sedative drug)^17a and isoindolinone 8 (a renin inhibitor),^17b as well as sultam 9, a tumor necrosis factor-α converting enzyme (TACE) inhibitor (Scheme 1c).^17c Furthermore, a pre-existing stereocenter on the alkyl tether offers the potential for stereoselective cyanofunctionalization, which has been rarely examined.

2. RESULTS AND DISCUSSION

2.1. Reaction Optimization.

The cyanamide substrates employed in previous studies typically required 5–6 steps to synthesize, including multiple purifications by column chromatography.^14a,b We prepared N-acyl cyanamide 1a from diphenylacetic acid, β-methallyl chloride, and p-toluidine in just three total steps.¹⁸ Our initial attempt at aminocyanation of 1a with a boron Lewis acid (BPh₃ or B(C₆F₅)₃) alone resulted in incomplete consumption of 1a and formation of byproduct 1a′, the product of reductive decyanation (Table 1, entries 1 and 2).¹⁹ Attempted aminocyanation of 1a with AlCl₃ led to decomposition (entry 3). Interestingly, aminocyanation reaction with CpPd(1-phenylallyl) (10 mol %), BPh₃ (40 mol %), and Xantphos (10 mol %) afforded γ-lactam 2a in 87% yield alongside a small amount of 1a′ (entry 4). This result showed that the palladium/Lewis acid strategy developed by Nakao’s group was also applicable to an alkyl-tethered cyanamide substrate.^14a Remarkably, a reaction using Pd(PPh₃)₄ and BPh₃ afforded a slightly lower yield of 2a than the Xantphos/CpPd(1-phenylallyl) system (81% versus 87%), offering a simple and affordable alternative catalytic system (entry 5). Other palladium sources, such as Pd(OAc)₂ and Pd₂dba₃ were less effective paired with Xantphos when compared with CpPd(1-phenylallyl) (entries 6 and 7).²⁰ Examination of bidentate phosphine ligands correlated the bite angle of ligands with their efficacy (entries 8–14).²¹ Thus, ligands with an increased bite angle generally afforded a higher yield of product, with Xantphos being optimal (entry 12). Nixantphos, which has a bite angle (114°) close to that of Xantphos (111°) performed similarly (entry 13), whereas further increasing the bite angle shut down the reaction (entry 14). A control reaction without BPh₃ gave a small amount of 2a along with other, unidentified, byproducts (entry 15). Under the optimized conditions of heating 1a in toluene at 80 °C in the presence of CpPd(1-phenylallyl) (10 mol %), Xantphos (10 mol %), and a boron Lewis acid (40 mol %), product 2a was isolated in 89% and 99% yield with BPh₃ and BEt₃, respectively (entries 16 and 17).

Table 1.

Optimization of Aminocyanation Conditions

entry	palladium	ligand	Lewis acid (equiv)	T (°C)	yield of 2a (%)a
1			BPh3 (0.5)	100	0b
2			B(C6F5)3 (0.5)	80	0b
3			AlCl3 (0.5)	80	0c
4	CpPd(1-phenylallyl)	Xantphos	BPh3 (0.5)	90	87
5	Pd(PPh3)4		BPh3 (0.5)	90	81
6	Pd(OAc)2	Xantphos	BPh3 (0.5)	90	67
7	Pd2dba3	Xantphos	BPh3 (0.5)	90	69
8	CpPd(1-phenylallyl)	dpped	BPh3 (0.4)	80	0b
9	CpPd(1-phenylallyl)	dppp	BPh3 (0.4)	80	23
10	CpPd(1-phenylallyl)	dppb	BPh3 (0.4)	80	49
11	CpPd(1-phenylallyl)	DPEphos	BPh3 (0.4)	80	72
12	CpPd(1-phenylallyl)	Xantphos	BPh3 (0.4)	80	93
13	CpPd(1-phenylallyl)	Nixantphos	BPh3 (0.4)	80	81
14	CpPd(1-phenylallyl)	DBFphos	BPh3 (0.4)	80	0b
15	CpPd(1-phenylallyl)	Xantphos		80	<20e
16g	CpPd(1-phenylallyl)	Xantphos	BPh3 (0.4)	80	89f
17g	CpPd(1-phenylallyl)	Xantphos	BEt3 (0.4)	80	99f

^aDetermined by ¹H NMR analysis using p-methoxyacetophenone as the internal standard.

^bOnly 1a and 1a′ detected by NMR spectroscopy.

^cDecomposition of 1a.

^dBite-angle values obtained from ref 21: dppe 85°; dppp 91°; dppb 98°; DPEphos 102°; Xantphos 111°; Nixantphos 114°; DBFphos 131°.

^eComplex reaction mixture.

^fYield after column chromatography.

^gConditions applied to further substrate scope study. dppe = 1,2-bis(diphenylphosphino)ethane; dppp = 1,3-bis(diphenylphosphino)propane; dppb = 1,4-bis(diphenylphosphino)butane; DPEphos = bis[(2-diphenylphosphino)phenyl] ether; Xantphos = 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene; Nixantphos = 4,6-bis(diphenylphosphino)-10H-phenoxazine; DBFphos = 4,6-Bis(diphenylphosphino)dibenzofuran.

2.2. Scope of Aminocyanation of N-Acyl Cyanamides 1.

We examined the substrate scope of aminocyanation reactions, starting with cyanamides bearing various aryl substituents that we prepared from readily available anilines. (Table 2). N-Phenyl cyanamide 1b and substrates bearing tert-butyl, fluoro, chloro, methoxy, and acetoxy groups para to the cyanamide moiety cyclized into the corresponding pyrrolidones in good to excellent yields (2b–2g). Substrates containing an electron-withdrawing aryl substituent (R = Ac, Cl, CF₃, CO₂Me; 1h–1k) were more prone to reductive decyanation. By increasing the amount of Lewis acid and lowering reaction temperature, we obtained satisfactory yields for 1h–1k (67–92%). Substrates bearing one or two m-OMe substituents provide products 2l and 2m in good yield, whereas an o-methyl substitution resulted in a sluggish reaction and modest yield of 2n, possibly due to steric hindrance. A pendant olefin (1o) or a pyridine ring (1p) was compatible to the reaction conditions, demonstrating the compatibility of these typically coordinating functional groups with aminocyanation.

Table 2.

Substrate Scope of Aminocyanation Reactions of N-Acyl Cyanamides 1^a

^aReaction conditions: CpPd(1-phenylallyl) (10 mol %), Xantphos (10 mol %), BPh₃ or BEt₃, PhMe (1.0 mL). Reaction time: 30 h (2h, 2m); 36 h (2u, 2y, 2aa); 48 h (2n, 2s, 2ab, 2ah); 24 h (all other entries). All yields are isolated yields. The structures of 1b–1ak are shown in the Supporting Information.

^bRun with 15 mol % Pd/Xantphos.

^cRun in m-xylene.

^dRun with 20 mol % Pd/Xantphos.

^eUnconsumed staring material. Tol = 4-methylphenyl, PMP = 4-methoxyphenyl.

We then turned our attention to groups α to the acyl cyanamide carbonyl. We used aminocyanation to prepare pyrrolidones containing geminal methyl (2q and 2r) and benzyl groups (2s) α to the carbonyl, as well as spirocyclic products (2t–2v) in good yield (Table 2). In contrast, complete removal of the α substituents resulted in a sluggish reaction and poor yield of product 2w (33%), indicating that the Thorpe-Ingold effect contributes to chemoselectivity in aminocyanation.

Next, we directed our focus to alkene substitution. Substrates with alkyl substituents other than methyl performed with comparable yields (2x and 2y). The allylic benzyloxymethyl group in 1z proved challenging; aminocyanation provided 2z in much lower yield, even after raising the catalyst loading to 20 mol % of Pd. The cyclization of a styrenyl double bond was slower, but efficient with slightly higher catalyst loading (2aa, 85% yield, 15 mol % Pd used). We were concerned that aminocyanation across a monosubstituted alkene would be challenging due to competitive β-hydride elimination. Despite our concern, the aminocyanation of such an alkene (1ab) afforded the corresponding pyrrolidone 2ab in 47% yield.

On the other hand, the N–CN bond of N-alkyl cyanamide 1ac was much more inert toward activation and only 29% yield of 2ac was obtained. The aminocyanation was smoothly extended to the construction of isoindolinones from the cyclization of styrenyl alkenes (2ad–2af, 59–99% yield). The formation of six-membered lactams 2ag (94% yield) and 2ah (49% yield) were achieved with a higher catalyst loading. In addition, we identified several incompatible substrates under the aminocyanation conditions. Substrates bearing an electron-deficient alkene (1ai) or an aryl C–Br bond (1aj) failed to produce any detectable amount of products (2ai or 2aj). An N-acetyl substituent on the aryl ring (1ak) rendered the N–CN bond unreactive with our optimized catalyst, even at 120 °C.

2.3. Aminocyanation of N-Sulfonyl Cyanamides 3.

Following our previous success with metal-free aminocyanation of N-tosyl cyanamides, we prepared N-sulfonyl cyanamide 3a and examined the reaction conditions to access sultam 4a (Table 3). We targeted 4a since sultams are biologically active. Initially, heating 3a with 1.0 equiv of B(C₆F₅)₃ alone in toluene failed to produce any detectable amount of 4a (Table 3, entry 1). Our attempts with BF₃·OEt₂ or BPh₃ as catalysts were also unfruitful (entries 2 and 3). In contrast, aminocyanation of 3a proceeded under palladium/Lewis acid catalysis, affording 4a in 69% yield, along with a small amount of reductive decyanation byproduct 3a’ (entry 4). A brief screening of palladium sources (entries 5–7) and reaction temperatures (entries 8–10) revealed that the catalytic system employing readily available Pd₂dba₃ (5 mol %), Xantphos (10 mol %), and BPh₃ (40 mol %) afforded 4a in near quantitative yield (entry 9). A control experiment performed without BPh₃ returned only unconsumed starting material (entry 11).

Table 3.

Aminocyanation of N-Sulfonyl Cyanamide 3a

entry	palladium	Lewis acid	T (°C)	yield (%)a
1b		B(C6F5)3c	100	0d
2b		BF3·OEt2c	100	0d
3b		BPh3	100	0d
4	Pd(OAc)2	BPh3	100	69
5	Pd(TFA)2	BPh3	100	63
6	CpPd(1-phenylallyl)	BPh3	100	81
7	Pd2dba3	BPh3	100	85
8	Pd2dba3	BPh3	90	95
9	Pd2dba3	BPh3	80	99e
10	Pd2dba3	BPh3	70	82
11	Pd2dba3		100	0d

^aDetermined by ¹H NMR analysis using DMSO-d₆ as the solvent and p-methoxyacetophenone as the internal standard.

^bWithout Xantphos.

^c1.0 equiv.

^dUnconsumed starting material.

^eYield after column chromatography.

N-Sulfonyl cyanamides bearing various aryl subsitutents, including −tBu, −F, −Cl, −OMe, −CF₃, and −CO₂Me groups para or meta to the nitrogen atom, all proceeded smoothly to the corresponding sultams in consistently high yields (Table 4, 4ib–4i, 89–95% yield). Aminocyanation reactions that form a six-membered sultam (4j, 57% yield), aminocyanation with a OMe substituent para to the sulfonyl group (4k, 76% yield), or an acetal on N-aryl group (4m, 89% yield) were also successful. However, 2-methoxypyridine-containing substrate 3l only afforded a trace amount of product (4l) with our optimized conditions for sultam formation. The mass balance consisted mainly of 3l. This result contrasts with our successful preparation of 2p above, also containing a 2-methoxypyridine. Taken together, aminocyanation when basic heteroatoms are present is possible, but likely requires additional optimization on a case-by-case basis. Our results show that a variety of functional groups are compatible with sultam-forming aminocyanation reactions.

Table 4.

Scope of Aminocyanation of N-Sulfonyl Cyanamides 3^a

^aReaction conditions: Pd₂dba₃ (5 mol %), Xantphos (10 mol %), BPh₃ (40 mol %), PhMe, 80 °C, 16 h. All yields are isolated yields.

^bRun for 24 h.

^cA trace amount of 4l is assigned to minor peaks in the ¹H NMR spectrum of the crude reaction mixture.

2.4. Synthetic Applications.

The aminocyanation reaction could be readily performed on a gram scale (Scheme 2). For instance, aminocyanation of cyanamides 1d and 1f gave the corresponding pyrrolidones 2d and 2f in excellent yields. Cyclizations of 1ad and 1af proceeded smoothly to the corresponding isoindolinones 2ad and 2af with reduced catalyst loading (5 mol % and 8 mol % of Pd/Xantphos, respectively).

Scheme 2.

Gram-Scale Aminocyanation Reactions^a

^aReaction conditions: (a) CpPd(1-phenylallyl) (10 mol %), Xantphos (10 mol %), BEt₃ (40 mol %), PhMe, 80 °C, 24 h; (b) CpPd(1-phenylallyl) (5 mol %), Xantphos (5 mol %), BEt₃ (60 mol %), PhMe, 80 °C, 24 h; (c) CpPd(1-phenylallyl) (8 mol %), Xantphos (8 mol %), BEt₃ (60 mol %), PhMe, 70 °C, 24 h. All yields are isolated yields.

To demonstrate the versatility of aminocyanation products in synthesis, we subjected 2ad to various chemical transformations (Table 5). Isoindolinone 2ad was hydrolyzed under basic conditions to either carboxylic acid 5a or primary amide 5b in good yield depending on reaction time. N-tert-Butyl amide 5c was obtained from 2ad in 88% yield via a Ritter reaction. Heating 2ad with n-butanol in the presence of p-TsOH gave ester 5d in 93% yield. A nickel-catalyzed methylation of the nitrile using AlMe₃ provided methyl ketone 5e in 77% yield.²² A palladium-catalyzed arylation reaction with p-tolylboronic acid quantitatively produced aryl ketone 5f.²³ Chemoselective reduction of the nitrile employing a NaBH₄–CoCl₂·6H₂O system generated amine 5g or, when combined with (Boc)₂O, the corresponding N-Boc carbamate 5h in good yield. Dibutyltin oxide-mediated cycloaddition of nitrile 2ad with trimethylsilyl azide gave tetrazole 5i in 76% yield.²⁴ Chemoselective deprotection of the N-PMP group with cerium(IV) ammonium nitrate afforded free isoindolinone 5j in 71% yield. Thioamide 5k was obtained in excellent yield by treating 2ad with Lawesson’s reagent. Chemoselective reduction of the amide moiety was also achieved with Charette’s procedure, furnishing isoindoline 5l in 76% yield.²⁵

Table 5.

Transformations of Isoindolinone 2ad^a

^aAll yields are of isolated material. Reaction conditions: (a) KOH, EtOH/H₂O, 90 °C, 40 h; (b) KOH, MeOH/H₂O, 90 °C, 2 h; (c) H₂SO₄ (1.2 equiv), tBuOH, 80 °C, 12 h; (d) TsOH·H₂O (2 equiv), nBuOH, 120 °C, 20 h; (e) Ni(acac)₂ (10 mol %), AlMe₃ (3 equiv), benzene, 50 °C, 5 h; (f) Pd(OAc)₂ (10 mol %), p-tolylboronic acid (2 equiv), 2,2′-bipyridyl (20 mol %), CF₃CO₂H (10 equiv), THF/H₂O, 90 °C, 28 h; (g) NaBH₄ (4 equiv), CoCl₂·6H₂O (1.5 equiv), MeOH, 0 °C, 2 h; (h) (Boc)₂O (2.5 equiv), NaBH₄ (4 equiv), CoCl₂·6H₂O (1.5 equiv), MeOH, 0 °C, 3 h; (i) TMSN₃ (3 equiv), nBu₂Sn(O) (0.5 equiv), PhMe, 100 °C, 48 h; (j) cerium(IV) ammonium nitrate (6 equiv), CH₃CN/H₂O, rt, 2 h; (k) Lawesson’s reagent (1 equiv), PhMe, 100 °C, 2.5 h; (l) Tf₂O (1.2 equiv), CH₂Cl₂, 0 °C to rt, 30 min, then Hantzsch ester (3 equiv), rt, 4 h. acac = acetylacetonyl; PMP = 4-methoxyphenyl.

We also exemplified the transformations of aminocyanation product 2af by preparing synthetic intermediates and analogues of biologically active isoindolinones (Scheme 3). Chemoselective reduction of the nitrile, followed by base-mediated formation of the piperazine ring gave isoindolinone 6 in 73% yield, which served as an advanced intermediate to PD-172938, a potent dopamine D₄ ligand.²⁶ In addition, a two-step hydrolysis/amide coupling sequence afforded amide 7 in 82% yield as a structural analogue to the anxiolytic drug pazinaclone.²⁷

Scheme 3.

Synthetic Applications of Isoindolinone 2af

2.5. Mechanistic Considerations.

We propose a catalytic cycle consistent with the Nakao group’s study.^14a Lewis-acid-promoted oxidative addition of the N–CN bond to the Pd(0) center gave rise to amidopalladium(II) cyanide complex I (Scheme 4, part a). Intramolecular addition of the N–Pd bond across the alkene double bond provides alkylpalladium(II) cyanide complex II, which, upon C–CN bond-forming reductive elimination, released the product and regenerated the active catalyst. The substrate scope study revealed a pronounced Thorpe-Ingold effect.

Scheme 4.

Mechanistic Considerations

We conducted a double crossover experiment to probe intramolecularity in Pd/Lewis-acid-catalyzed aminocyanation (Scheme 4, part b. A mixture of 1d and ¹³C-lableled cyanamide 1r ([¹³C]1r) were subjected to the aminocyanation conditions, affording the corresponding noncrossover products 2d and [¹³C]2r. The crossover products were not observed. This result indicated that BEt₃ did not facilitate the dissociation of Pd–CN bond in the proposed catalytic cycle. These results are consistent with the proposed intermediates I and II in Scheme 4, part a.²⁸

Substrates bearing an existing stereogenic center α to the carbonyl were prepared by routine alkylation reactions of enolates, offering a convenient way to examine the diastereoselectivity of the alkene addition step (Scheme 4c). For instance, cyclization of α-methyl-substituted cyanamide 1al gave lactams 2al in 59% yield and 9:1 d.r. The syn-relationship of the two methyl groups was determined by nOe experiments. Similarly, substrate substituted with a phenyl group α to the carbonyl (1am) yielded the corresponding lactam 2am in moderate yield (48%), but with excellent selectivity (>20:1 d.r.). Encouraged by these results, we further extended the scope of diastereoselective aminocyanation reactions to access functionalized lactams bearing an all-carbon stereocenter (2an), an N-phthalimide group (2ao), or a benzyloxy group (2ap) α to the carbonyl, resulting in generally good yield. The d.r. was excellent when one of the α-substituents was large, as in 2an (19:1 d.r.) and 2ao (20:1 d.r.), but lower when both were small (2ap, 3.2:1). Furthermore, cyclization of an alkene with an allylic stereogenic center proceeded cleanly to provide densely substituted lactam 2aq in 92% yield and 4.4:1 d.r.

To gain the further insight into the diastereoselectivity of the aminocyanation reactions, we studied the stereochemistry of the alkene addition step. We designed a terminal monodeuterated substrate (trans-1ad-d₁) to examine how the N–Pd bond inserts into the alkene. The possible outcomes are syn-addition, anti-addition, or a mixture of the two. A possible mode for anti-addition is through ionization of N–Pd bond to form an alkene complex, followed by attack of nitrogen on the alkene. This would be consistent with our null observation on cyano group crossover, if it occurred by the formation of a tight ion pair. The structures of the possible products (2ad-d₁-I and 2ad-d₁-II) are shown in Scheme 5.

Scheme 5.

Mechanistic Study of Alkene Addition Step

Aminocyanation of trans-1ad-d₁ yielded a single diastereomer, however, assigning the position of deuterium in 2ad-d₁ was challenging due to the exocyclic stereogenic center. The ¹H NMR spectrum of unlabeled 2ad showed the diastereotopic protons (atom 13 and 31 in Figure 1) as well-resolved signals (δ = 2.89, 2.70 ppm), showing that the two protons are in distinct electronic environments. Thus, we hypothesized that the two diastereotopic protons could be assigned by nOe experiments. As expected, the downfield diastereotopic proton in 2ad shows the through-space interactions with aromatic proton 24 on the isoindoline and 32/35 on the para-methoxyphenyl group. The upfield diastereotopic proton is only observed to enhance the signal of proton 32/35 on the para-methoxyphenyl group. We also ran an nOe experiment for the only proton alpha to the nitrile (δ = 2.88 ppm) in the ¹H NMR of 2ad-d₁, and we observed nOe for aromatic proton 24 and 32/35. These data support the assignment of our labeled product as 2ad-d₁-I, assuming a conformation where the methyl and nitrile groups are anti. This assumption seemed reasonable, but we sought additional data to confirm our assignment.

Figure 1.

Optimal geometry of 2ad and nOe results.³⁰

To bolster our assignment of 2ad-d₁, we utilized computational methods²⁹ to predict the chemical shifts of the diastereotopic protons in 2ad. Conformational minimization in silico confirmed our assumptions about the anti-orientation of the nitrile and methyl group in the dominant conformers. The diastereotopic protons (labeled as 31 and 13 in Figure 1) were respectively predicted at δ = 2.80 and 2.62 ppm (Δδ = 0.18 ppm), while the experimental chemical shifts are δ = 2.89, 2.70 ppm (Δδ = 0.19 ppm) in the ¹H NMR of 2ad. Compared with the spectrum of 2ad-d₁, the more upfield peak around 2.70 ppm is not observed, confirming that proton 13 is the one labeled as deuterium. Combined with nOe results, we assign 2ad-d₁-I as the diastereomer produced by of aminocyanation of trans-1ad-d₁. These results are consistent with the alkene addition proceeding via a syn-addition pathway.

Based on syn-aminopalladation, we proposed a model (shown in Scheme 6) to elucidate the origin of the diastereoselectivity. During the ring-forming step, the molecule may adopt a chairlike conformation. Meanwhile, the larger substituent adjacent to the carbonyl prefers to be oriented in a pseudoaxial position to minimize strain between the R-group and the carbonyl. Hence, the cyclization reactions afforded the diastereomers we observed (2al–2ap).

Scheme 6.

Proposed Model for the Diastereoselectivity

3. CONCLUSION

In summary, we developed an intramolecular aminocyanation of alkenes by N–CN bond activation of N-acyl and N-sulfonyl cyanamides, accessing a broad range of nitrogen-containing heterocycles, including pyrrolidones, piperidinones, isoindolinones, and sultams. The synthetic utility of the method was demonstrated in a variety of transformations of the resulting lactam heterocycles, including those leading to medicinally relevant intermediates. Additionally, diastereoselective aminocyanation reactions, previously underdeveloped, were successfully applied to the formation of densely substituted pyrrolidones. We have shown through isotope labeling experiments that the reaction proceeds via a syn-addition pathway and that the cyclization is intramolecular due to the lack of crossover in cyanide labeled experiments. These data add clarity to the mechanistic picture for the N–CN bond activation and functionalization promoted by palladium and Lewis acid catalysts.

Supplementary Material

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.8b01330.

Experimental details for the preparation of new compounds, tabulated characterization data (¹H NMR, ¹³C NMR, melting points, MS, and IR), and copies of NMR spectra (PDF)