Native Amides as Enabling Vehicles for Forging sp³–sp³ Architectures via Interrupted Deaminative Ni-Catalyzed Chain-Walking

Abstract

Herein, we disclose an interrupted deaminative Ni-catalyzed chain-walking strategy that forges sp³–sp³ architectures at remote, yet previously unfunctionalized, methylene sp³ C–H sites enabled by the presence of native amides. This protocol is characterized by its mild conditions and wide scope, including challenging substrate combinations. Site-selectivity can be dictated by a judicious choice of the ligand, thus offering an opportunity to enable sp³–sp³ bond formations that are otherwise inaccessible in conventional chain-walking events.

The ability to selectively functionalize sp³ C–H bonds has received considerable attention in medicinal chemistry programs.^1,2 Their popularity arises from the observation that drugs possessing a higher content of sp³-hybridized carbons are oftentimes more selective and deliver higher success in clinical trials.³ Recently, Ni-catalyzed chain-walking of unactivated olefins has gained momentum as a powerful, yet practical, alternative to commonly adopted sp³ C–H functionalization strategies,¹ allowing to promote innovative bond disconnections for forging sp³ architectures from simple, yet abundant, precursors.⁴ Despite the advances realized, site-selectivity in Ni-catalyzed chain-walking reactions is predominantly dictated by a subtle interplay between electronic and steric effects. Indeed, bond formation typically takes place adjacent to a stabilizing group on thermodynamic grounds,⁵ whereas functionalization at distal, primary sp³ C–H bonds is kinetically preferred (Scheme 1).⁶

Scheme 1. Canonical Ni-Catalyzed Chain-Walking Events.

Recent elegant disclosures have illustrated the viability for targeting other sp³ C–H sites within the alkyl side chain with strongly chelating bidentate 8-aminoquinoline,⁷ thioethers,⁸ protected amines,⁹ or transient directing groups via ketimine formation¹⁰ by utilizing alkyl/aryl halides or electrophilic nitrogen sources (Scheme 2, top). Driven by the prevalence of alkyl amines in a myriad of biologically relevant molecules and preclinical candidates,¹¹ we wondered whether we could establish “a la carte” site-selective interrupted Ni-catalyzed deaminative¹² chain-walking for forging sp³–sp³ linkages at the always elusive methylene sp³ C–H bonds with predictable and tunable selectivity by using weakly coordinating native amides (Scheme 2, bottom).¹³ The choice of the latter is not arbitrary, as these ubiquitous motifs possess an electron-rich carbonyl fragment suitable for metal chelation without the need for decorating the amide backbone with strongly chelating quinoline or pyridine backbones, hence setting the basis for applying these techniques to advanced synthetic intermediates.¹⁴ As part of our interest in the field,¹⁵ we report herein the successful realization of this goal. The protocol is distinguished by its mild conditions, broad scope, and exquisite site-selectivity, thus offering new opportunities in chain-walking and an unrecognized opportunity to enable formation of unactivated sp³–sp³ bonds in deaminative cross-couplings at methylene sp³ C–H sites.¹⁶ The protocol is inherently modular, allowing to establish regiodivergent scenarios for incorporating sp³–sp³ architectures at different methylene sp³ C–H sites by judicious choice of the ligand backbone.

Scheme 2. Interrupted Ni-Catalyzed Chain-Walking Events.

We began our studies by evaluating the Ni-catalyzed interrupted deaminative chain-walking reaction of 1a with 2a, readily accessible in large amounts from the corresponding tert-butyl 4-aminopiperidine-1-carboxylate in a single step (Table 1). After judicious screening of the reaction parameters,¹⁷ we found that a combination of NiI₂ (5 mol %), L1 (10 mol %), (EtO)₃SiH, and Na₂HPO₄ in DMA at 40 °C afforded the best results, giving rise to 3a in 81% isolated yield with exquisite β-selectivity (>150:1). As anticipated, the nature of the ligand was critical for success, with pyrox ligands providing better results than 2,2′-bipyridine congeners (entry 1 vs entry 6). Low β-selectivity was obtained with L3 and L4 lacking substituents adjacent to the nitrogen atom at either the pyridine or oxazoline motif (entries 3 and 4), whereas a selectivity switch was observed when utilizing nonsubstituted analogues, thus showing the subtleties of our reaction.¹⁸ In addition, nickel sources, silanes, or solvents other than NiI₂, (EtO)₃SiH, or DMA led to lower yields of 3a (entries 7–9). Similarly, conducting the reaction at room temperature led to significantly lower reactivity (entry 10). As expected, control experiments indicated that all of the reaction parameters were critical for success (entry 11).

Table 1. Optimization of the Reaction Conditions^a.

^aConditions: 1a (0.12 mmol), 2a (0.10 mmol), NiI₂ (5 mol %), L1 (10 mol %), (EtO)₃SiH (0.20 mmol), Na₂HPO₄ (0.20 mmol), DMA (0.10 M) at 40 °C for 15 h.

^bGC yields using dodecane as internal standard.

^cIsolated yield.

As shown in Table 2, our interrupted chain-walking sp³–sp³ bond formation turned out to be widely applicable regardless of the substitution pattern at the amide backbone. Indeed, excellent yields and exclusive β-selectivities were observed for a series of amides derived from acyclic amines (3a–3c), anilines (3d), piperidine (3e), piperazine (3f), morpholine (3g), and pyrrolidine (3h). Even secondary amides exhibiting acidic N–H bonds delivered 3j and 3l in good yields and exquisite β-selectivity. Notably, excellent regiocontrol was obtained with Weinreb amides with just 1 mol % NiI₂ (3m), thus leaving ample room for further manipulation by using the amide backbone as a masked carbonyl fragment.^17,19 These results contribute to the perception that our protocol does not operate under substrate control and that L1 dictates the β-selectivity pattern. Note, however, that azetidine 3i and primary amides (3k) led to slightly lower regioselectivities, probably due to a poor orbital overlap in the former and an ineffective coordination mode in the latter.²⁰ Next, we turned our attention to study the influence of the olefin on both reactivity and selectivity. Importantly, our interrupted β-selective deaminative chain-walking event could be applied across a wide number of substrates with exclusive β-selectivity. It is worth noting that long-range chain-walking events are within reach with exclusive β-selectivity, albeit in slightly lower yields (3a vs 3n–q). More importantly was the observation that both E- and Z-internal olefins could participate equally well in our β-selective sp³–sp³ bond-forming reaction (3n–3y), even by utilizing advanced reaction intermediates (3z). Notably, 1,1-disubstituted olefins or even trisubstituted olefins could be employed as substrates with similar ease (3s, 3t). Equally interesting was the ability to extend this technique to cyclic amides, obtaining the targeted cis β-alkylated 3u and 3w. Interestingly, 3-cyclohexenamide only delivered γ-selective 3v—unambiguously characterized by X-ray diffraction—in 51% yield. At present, we tentatively ascribe these results with an enhanced chelation of the nickel catalyst to the amide backbone through syn-pentane interactions in the well-defined chair conformation of cyclohexane rings.^17,21 Although one might argue that the presence of a pending, coordinating ester or an arene on the alkyl side chain might compromise site-selectivity,^5a,6b this was not the case, and 3x and 3y were both obtained in good yields, thus highlighting the strong chelation exerted by the native amide backbone.¹³ The successful preparation of 3aa–3ak with pyridinium salts arising from aliphatic or cyclic secondary amines showcases the generality of our β-alkylation beyond 2a. While the utilization of pyridinium salts derived from benzyl amines led to the targeted products 3aj and 3ak, the coupling of the corresponding primary alkyl congeners was better suited with alkyl iodides, delivering 3al–3ar in good yields as exclusive β-isomers, even when utilizing capsaicin as substrate (3at vs 3z). As expected, 3u and 3as were both obtained with exquisite regioselectivity en route to the corresponding β- or γ-alkylated products, thus illustrating the intriguing site-selectivity observed depending on the size of the pre-existing ring.

Table 2. Scope of the Ni-Catalyzed Deaminative Interrupted Chain-Walking β-Alkylation of Native Amides^a.

^aAs for Table 1, entry 1 (0.30 mmol scale).

^bNiI₂ (1 mol %), L1 (2 mol %).

Taking into consideration the exclusive β-selectivity observed in Table 2, we wondered whether our interrupted chain-walking deamination proceeded via five-membered nickelacycles or acrylamides formed upon olefin isomerization prior to sp³–sp³ bond formation. To this end, the reaction of 1a and 2a was monitored by both ¹H NMR spectroscopy and GC analysis.¹⁷ Under the limits of detection, not even traces of acrylamide were detected in the crude mixtures. In addition, no deuterium incorporation was observed at the β-position when utilizing 1a-d₂ as substrate, thus arguing against acrylamides as reaction intermediates (Scheme 3). However, it was not particularly straightforward to rigorously distinguish whether sp³–sp³ bond formation occurred via reductive elimination of in situ generated nickelacycles or Giese-type addition of alkyl radical intermediates to α,β-unsaturated amides. High diastereoselectivity for the reaction of 1z was anticipated in the former; on the contrary, statistical mixtures of diastereoisomers in 3au–3av would indicate a radical-type pathway. As shown in Scheme 3, high yields and dr > 20:1 were obtained for both products, with an anti-stereochemistry unequivocally confirmed by X-ray diffraction of 3av. In line with this observation, the reaction of enantioenriched 1aa resulted in 3aw with an excellent selectivity profile and, more importantly, with preservation of the chiral integrity at the α-position. Taken together, these findings advocate the notion that the formation of Ni–I precedes sp³–sp³ bond formation.

Scheme 3. Preliminary Mechanistic Experiments.

Conditions for 1z and 1aa: amide (0.30 mmol), alkyl iodide (0.45 mmol), NiI₂ (10 mol %), L1 (20 mol %), (EtO)₃SiH (0.60 mmol), Na₂HPO₄ (0.60 mmol), DMA (0.10 M) at 50 °C for 15 h.

Encouraged by the results of Table 2, we wondered whether the nature of the ligand might dictate the site-selectivity of our interrupted Ni-catalyzed deaminative chain-walking event. As shown in Scheme 4, regiodivergency could be accomplished in the reaction of 1a and 2a by a judicious choice of the ligand. Specifically, a Ni/L1 regime delivered 3a with exclusive β-selectivity (>150:1) and high TON (68). Notably, the utilization of L7 or L8 gave rise to 4a and 5a with γ- and δ-selectivity, respectively.²² These results should be interpreted against the challenge that is addressed, offering a gateway to develop “a la carte” site-selective deaminative sp³–sp³ bond formations by controlling the motion at which Ni catalysts promote chain-walking reactions. Gratifyingly, our chain-walking scenario based on the Ni/L7 couple could be extended to internal olefins with either primary or secondary alkyl electrophiles (4b, 4c). In addition, the successful preparation of 4d from capsaicin stands as a testament to the impact that this technique might have in the context of late-stage diversification of advanced intermediates.

Scheme 4. Regiodivergent β/γ/δ sp³–sp³ Bond Formations.

Conditions for 4a–d: amide (0.36 mmol), pyridinium salt or alkyl iodide (0.30 mmol), NiBr₂·diglyme (5 mol %), L7 (10 mol %), (EtO)₃SiH (0.60 mmol), Na₂HPO₄ (0.72 mmol), DMA (0.10 M) at rt for 15 h.

In summary, we report the successful utilization of native amides for forging sp³–sp³ architectures via interrupted site-selective deaminative nickel chain-walking catalysis at methylene sp³ C–H sites. This method is characterized by its mild conditions, excellent site-selectivity, and wide scope, including challenging substrate combinations and advanced intermediates. A judicious choice of the ligand backbone allows the effective discrimination among different methylene sp³ C–H sites, offering an opportunity to promote regiodivergent strategies when forging sp³ architectures. Further studies into the mechanism and extension to related transformations are currently ongoing.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.2c12915.

• Experimental procedures and spectral and crystallographic data (PDF)

Author Contributions

J.R. and H.W. contributed equally to this work.

The authors declare no competing financial interest.