Introduction

Functional groups (FGs) have a profound influence on the properties and functionalities of organic molecules, and manipulations of these groups serve as foundational reactions in synthetic chemistry. While the field has made significant progress in translocating functional groups (FGs) to remote positions1,2,3,4,5,6,7,8,9 a distinct category of FG manipulation remains comparatively underdeveloped: the direct migration of a FG to unactivated C–H bonds, achieved without introducing additional modifications to the molecular structure (Fig. 1a). In contrast to processes that necessitate the introduction or removal of functional groups, this approach maintains the same chemical composition while achieving a unique arrangement of atoms and bonds, embodying the principles of molecular editing10 and presenting new avenues for the design and synthesis of complex organic molecules.

Fig. 1: Background and reaction design.
figure 1

a General reaction profile for FG translocation to unactivated C–H sites without altering molecule’s overall structure. b Prior art for radical difunctionalization of unactivated alkenes through FG translocation. c Our two-step strategy for diverse difunctionalization of unactivated alkenes via FG translocation and straightforward subsequent transformations. d This work: cyano group translocation to alkenyl C(sp2)–H site by quinuclidine-based radical cation catalysis. FG, functional group.

Full size image

Alkenes occupy a central role not only as key scaffolds in drug-like molecules but also as vital starting materials in the realm of synthetic chemistry. Recent decades have witnessed a resurgence in radical chemistry, fueled by its exceptional reactivity and excellent functional group compatibility11,12,13. Particularly noteworthy is the substantial progress achieved in alkene 1,2-difunctionalization through radical-mediated processes, which enable the concurrent introduction of two distinct FGs at adjacent carbon centers14. Notably, the emerging paradigm of functional group translocation via radical intermediates has enabled a series of difunctionalization protocols (Fig. 1b)1,2,3,4,5,6. Mechanistically, these reactions commences with intermolecular electrophilic radical addition to the alkene. Subsequent intramolecular radical addition to the functional group leads to the formation of a cyclic radical intermediate. The transient radical intermediate is inclined to undergo an elimination step to achieve thermodynamic stabilization via functional group migration. This difunctionalization strategy employing functional group translocation typically necessitates the employment of an electronic radical precursor, consequently limiting the scope of functional groups (FGs) that can be incorporated. Moreover, each of these reactions generally demands rigorous optimization of reaction conditions, encompassing the choice of an appropriate catalyst system. To address these limitations, a practical approach might involve selectively directing the FG translocation to C(sp2)–H sites, process has the potential to transform an unactivated alkene into an activated one, thereby facilitating subsequent chemical modifications and enabling the straightforward and reliable difunctionalization of a diverse range of alkenes through dependable chemical reactions (Fig. 1c). To date, the direct translocation of FG to C−H sites without altering the molecule’s overall architecture remains in its nascent stages. The cyanide (CN) group, known for its versatility in organic chemistry, has been widely explored in group translocation reaction15,16,17,18,19,20,21,22,23. Recently, Xu et al. reported a novel process where cyanide groups migrate to unactivated C(sp3)–H sites via reversible hydrogen atom transfer (HAT), termed ‘C–H sampling’ process (Fig. 1a)24. With a sustained fascination for exploring radicals as catalysts25,26, we are captivated by the potential of radical cation species to function as catalysts27,28,29,30,31,32,33,34,35,36,37,38,39 for the CN translocation to unactivated C(sp2)–H sites (Fig. 1d). The catalytic reaction pathway entails a radical addition step, where the radical cation adds to an alkene, thereby shifting the spin center in preparation for the subsequent CN translocation. Following this, an ionic elimination reaction takes place, yielding the desired product while concurrently regenerating the radical, thus completing the catalytic cycle. Here, we show the direct cyano group translocation reaction to an alkenyl C(sp2)–H site, facilitated by quinuclidine-based radical cation catalysis. We envisioned these catalytic reactions proceeding through a mechanism involving the radical addition of a double bond by an electrophilic radical, specifically the quinuclidinium radical cation, which generates a β-quinuclidinium radical species40,41. This is then followed by a specific intramolecular CN 1,4-translocation, occurring via a 5-exo-dig cyclization and radical β-scission. Ultimately, this sequence leads to the formation of the alkenyl cyanide product, with the quinuclidine catalyst being recovered after the Hofmann elimination step. Given the versatility of alkenyl cyanide moiety in Michael or Giese reaction, and broad availability of various nucleophiles, we expect that such a type of direct CN translocation would contrasts and complements existing state-of-the-art reaction types. As such, it promises to be an intriguing and highly versatile platform for the diverse difunctionalization of unactivated alkenes.

Results

Reaction development

Within the realm of covalent radical catalysis, the dual characteristics of relative stability and the remarkable ability to reversibly participate in radical addition reactions serve as the crucial factors that establish radicals as catalysts. After extensive exploration, we were delighted to discover that the quinuclidine stood out as an exceptionally potent precatalyst for the proposed CN translocation reaction. Initially, we selected 2-(but-3-en-1-yl)-2-phenylmalononitrile 1 as a model substrate with which to test the proposal. Following a series of systematic endeavors, the optimized conditions were established (Table 1; see Supporting Information for detail). By utilizing 1,2,3,5-Tetrakis(carbazol-9-yl)-4,6-dicyanobenzene (4CzIPN, 1 mol%) as the photocatalyst, ethyl quinuclidine-4-carboxylate (N3, 15 mol%) as the nitrogen precatalyst, and bulky 2,4,6-triisopropylbenzene disulfide ((TripS)2, 20 mol%) as the cocatalyst, the desired 1,4-CN translocation product 2 could be generated in 91% NMR yield under 455 nm blue LEDs irradiation in acetonitrile at room temperature for 12 h (Table 1, entry 1). In the absence of light, photocatalyst, or quinuclidine precatalyst, no product was detected (Table 1, entry 2–4). During the investigation, we found CN translocation product 2 slowly polymerized under the photocatalytic condition42, and subsequently discovered that the thiol catalyst played a role in reducing the rate of product polymerization. Removal (TripS)2 or substitution of 2,4,6-triisopropylbenzenethiol (TripSH) for (TripS)2 resulted in decreased reaction yields (Table 1, entry 5 and 6).

Table 1 Optimization of reaction conditionsa
Full size table

The influence of the quinuclidine catalyst scaffold on the reaction efficiency was investigated (Table 1, entry 7–12). The position and characteristics of substituents on quinuclidine significantly influence the reactivity of the reaction. Quinuclidine (N1), 4-CN (N2), and 4-CPh2OH (N4) substituted quinuclidine gave lower yield while ethyl quinuclidine-4-carboxylate N3 turned out to be the best precatalyst (Table 1, entry 7–9). N5, which possesses an OAc group at the 3-position, displayed comparable efficiency in catalyzing the reaction (Table 1, entries 10). Specifically, catalysts N6 where the free NH2 group is substituted at the 3-position, exhibited minimal capacity to catalyze the reaction (Table 1, entry 11). In addition, those substituted at the 2-position, exemplified by natural products of Cinchona alkaloid family such as quinine N7 or hydroquinidine N8 (Table 1, entry 12), were ineffective to catalyze the reaction. Tertiary amines such as DABCO and Et3N were also ineffective (Table 1, entry 13), further highlighting the unique nature of quinuclidine as a covalent radical catalyst. Other heteroatom catalysts, such as triphenylphosphine and tetramethylthiourea, which have been reported to undergo SET to give the corresponding radical cation, did not work, with all starting materials recovered.

Substrate scope evaluation

With the optimal conditions in hand, we scaled the model reaction. Notably, the desired CN translocation product 2 was efficiently formed at a scale of 5 mmol (Fig. 2). The structure of 2 was further confirmed by X-ray crystallography. Additionally, a diverse range of phenyl groups with various aromatic substituents and substitution patterns (3−8) were successfully accommodated. The versatility of the reaction was further demonstrated with malononitriles featuring an elongated benzyl group (9 and 10). Furthermore, a wide array of alkyl chains incorporating diverse heterocycles, such as thiophene (11), pyridine (12), and indole (13), were observed to be compatible. Naphthyl (14), along with smaller rings like cyclopropyl (15), cyclobutane (16), and oxetane (17) motifs, were also found to tolerate the reaction conditions. Encouraged by these findings, we expanded our exploration to substrates with various common functional groups. Acetals (18), esters (19), ethers (20), halides (21), amines (22), and even free hydroxyl groups (23) were well-tolerated under the reaction conditions. Unsaturated bonds, including enynes (24) and alkenes (25), were also found to be compatible with the optimal conditions. Moreover, tertiary alkyl groups (26−29) were suitable for the current CN translocation reaction. Beyond malononitrile derivatives, we further investigated substrates containing only one CN functional group. A range of ‘mono’-nitrile substrates with radical-stabilizing α-substituents like α-cyano esters (30 and 31), ketones (32), and heteroarenes (33) were found compatible under the reaction condition. α-Cyanated amino esters were also tested, leading to novel derivatives (34) with moderate efficiency. In addition to achieving the translocation of the CN group through the cleavage of a C–C bond, we have also conducted experiments to test the CN group translocation via the cleavage of a C–N bond, using cyanamides as the substrates. The cyanamide derived from aniline exhibited low reactivity (35). However, the cyanamide derived from p-toluenesulfonamide performed well and produced the cyanamide product in a moderate yield (36). When malononitrile derivative 37 was used as a substrate, the desired CN translocation product 38 was only generated in trace amounts. Furthermore, upon 39 was used as the substrate, no product 41 was formed, affirming the challenge of translocating a CN functional group through a three- or six-memberedring transition state43. Interestingly, substrate 40, which differs from 39 only by the introduction of a phenyl group at the terminal site of the alkene, could undergo CN translocation, yielding product 42 with exclusive 1,4-translocation selectivity. With internal alkene 43 as a substrate, both 1,4-CN translocation products 44 and 45 were observed, which confirms the existence of a quaternary ammonium as catalytic intermediate. Furthermore, we successfully applied this CN translocation process to more complex substrates derived from bioactive molecules or drugs, such as Indomethacin (46), Telmisartan (47), and Progesterone (48). These findings highlight the broad applicability and versatility of our CN translocation reaction.

Fig. 2: Substrate scope for cyano group translocation to alkenyl C(sp2)–H site.
figure 2

aThe reaction conditions were as follows: conducted with substrate (0.30 mmol) using 4CzIPN (1 or 5 mol%), N3 (15 mol%), and (TripS)2 (20 mol%) in CH3CN (1.5 mL) at room temperature for 12–48 h; Isolated yields. b5 mmol scale reaction. cThe diastereomeric ratios (d.r.) of the products are approximately 1:1 determined by 1H NMR. dYields and ratios were determined by 1H NMR.

Full size image

Synthetic application

Next, the implementation of this CN translocation strategy for the difunctionalization of compounds was showcased through a series of classic Michael addition and Giese radical addition reactions (Fig. 3). The results presented in Fig. 3a demonstrate that the CN translocation product can function as a versatile linchpin electrophile, enabling quick access to a variety of difunctionalized products through reactions with a broad spectrum of nucleophiles, including those based on carbon (49 and 50), nitrogen (51), oxygen (52), thiol (53), halide (54), and hydrogen (55). In addition to the ionic strategy, Giese-type radical addition offers an alternative pathway for the functionalization of molecules, as exemplified by the photo catalytic alkylation44 (56), silylation45 (57), and phosphonation46 (58) reactions (Fig. 3b). It is nontrivial to directly introduce nucleophilic radicals using previous translocation strategies for the difunctionalization of alkenes due to the inherent polarity mismatch effect47,48. This limitation can be overcome by our CN translocation to C(sp2)–H strategy, which employs a wide range of nucleophilic radical precursors. Furthermore, we illustrate that the CN group can be diversely transformed into industrially important adipic acid and hexamethylenediamine derivatives (59−61).

Fig. 3: Diverse difunctionalizations of alkenes through sequential cyano translocation and nucleophilic addition.
figure 3

a General two-step strategy for diverse difunctionalizations of alkenes. b Diverse difunctionalizations of alkenes through sequential cyano translocation and Michael addition. c Diverse difunctionalizations of alkenes through sequential cyano translocation and Giese radical addition.

Full size image

Mechanistic considerations

To gain a more comprehensive understanding of the reaction mechanism, a series of mechanistic studies were conducted (Fig. 4). Initially, two radical scavengers, namely (2,2,6,6-tetramethylpiperidin-1-yl)oxyl (TEMPO) and 2,6-di-tert-butyl-4-methylphenol (BHT), were utilized in preventive experiments. In both cases, a significant loss in yield of 2 was observed, indicating the involvement of a radical-mediated mechanism (Fig. 4A). Then, the deuteration experiment was performed. Substrate 1 was subjected to standard reaction conditions in the presence of 1.0 equiv of D2O. The CN translocation product 2 was generated in 87% yield with 41% d-incorporation product into the benzylic position (Fig. 4B). This observation hints at the likely presence of a benzylic anion species as an intermediate, potentially formed through a photocatalytic radical polar crossover process. It is noteworthy that, despite a lower yield, a similar percentage of deuterium incorporation was noted in the absence of (TripS)2, suggesting that (TripS)2 may do not participate in the catalytic cycle of the CN translocation reaction (Fig. 4B). (TripS)2 was later proven to play a role in inhibiting, to a certain extent, the undesired polymerization of the product under the photocatalytic conditions (Fig. 4C). To elucidate the mechanism of covalent radical catalysis wherein the radical catalyst reversibly adds to the double bond, we designed substrate 62 specifically to capture the pivotal intermediate. When 62 was subjected to standard reaction conditions, in the presence of a stoichiometric amount of N3 and 1.0 equivalent of H2O, we successfully isolated the quinuclidine-bound quaternary ammonium hydroxide product 63. These results conclusively confirm the hypothesis of a covalent catalysis mechanism based on an aminium radical cation (Fig. 4D). The time-course experiments with light on/off cycling indicated that the yield of 2 showed no significant increase in the absence of light under standard conditions, aligning with a catalFytic radical-polar crossover pathway instead of a radical chain propagation pathway (Fig. 4E). The fluorescence Stern-Volmer quenching experiment involving 4CzIPN was conducted with substrate 1, quinuclidine N3, and (TripS)2, and indicated that the excited-state photocatalyst tends to be quenched by quinuclidine (Fig. 4F). Based on the information presented above, we conceived a catalytic cycle in which the excited-state photocatalyst 4CzIPN initially oxidizes quinuclidine to its corresponding aminium radical cation49, as depicted in Fig. 4G. This electrophilic N-centered radical cation subsequently undergoes intermolecular addition with the olefin acceptor 1, leading to the formation of a new C–N bond and an adjacent carbon-centered radical40,50. Drawing inspiration from prior research on radical CN translocation, this radical intermediate proceeds through a 5-exo-dig cyclization, resulting in the formation of an imido radical intermediate. This intermediate then undergoes β-scission to yield a more stable benzyl radical intermediate. This radical can then accept an electron from the reduced state of the photocatalyst, generating the corresponding anion species. Subsequently, a Hofmann elimination and proton transfer occur, liberating the quinuclidine catalyst while producing the CN translocation product 2. Alternatively, the generated carbanion intermediate may undergo deprotonation of the α-C–H bond in the migrated CN group via a five-membered cyclic transition state. Through E2 or E1cb mechanism, this facilitates proton transfer to the carbanion center with concomitant elimination of the quinuclidine moiety and double-bond formation.

Fig. 4: Mechanistic studies.
figure 4

a Radical prevention experiments. b Isotope exchange experiments. c Probing the role of disulfide. d Isolation of catalytic analogs of intermediate. e Light on/off experiment. f Stern−Volmer experiment. g Proposed mechanism.

Full size image

Conclusions

In summary, we have demonstrated a straightforward and practical protocol leveraging photocatalytic quinuclidine-based radical cation catalysis to selectively effectuate 1,4-cyano translocation to alkenyl C(sp2)–H sites. This reaction is noteworthy for its extensive substrate scope, exceptional group tolerance, 100% atom economy, and operation under redox-neutral conditions, alongside its suitability for late-stage functionalization. By employing N-centered radical cations as catalysts, the approach concurrently facilitates radical translocation and ionic Hofman elimination, thereby restoring the alkenyl double bond and enabling positional exchange between a CN group and the alkenyl C(sp2)–H bond. This ultimately yields alkenyl cyanide, a versatile intermediate that can be readily transformed into a diverse array of difunctionalized products through reactions with a broad spectrum of nucleophiles. The continual exploration of radical covalent catalysis in organic synthesis is ongoing in our laboratory.

Methods

General procedure for 1,4-cyano translocation to alkenyl C(sp2)–H sites by photocatalytic quinuclidine-based radical cation catalysis

In an oven-dried Schlenk tube, to the solution of corresponding substrate (0.3 mmol, 1.0 equiv.) in anhydrous acetonitrile (1.5 mL) was added ethyl quinuclidine-4-carboxylate (N3) (8.3 mg, 0.045 mmol, 0.15 equiv.), (Trip)2S (28.2 mg, 0.06 mmol, 0.2 equiv.) and 4CzIPN (1 mol%). The reaction solution was degassed by three cycles of freeze-pump-thaw, and the Schlenk tube was charged with nitrogen subsequently. The reaction tube was placed on the magnetic stirrer equipped with 455 nm LEDs. The reaction solution was stirred and monitored by TLC at room temperature until the reaction is completed. The reaction solution was concentrated under vacuum, then purified by flash column chromatography on silica gel to afford the corresponding product.