Introduction

In drug discovery, the incorporation of a main group element functionality, such as silicon, into drug candidates is a promising chemical strategy that can improve their biological activity and reduce toxicity1. Silicon, which belongs to group IV in the periodic table, shares a comparable bonding structure with carbon. However, silicon has a larger covalent radius compared to carbon, leading to enhanced lipophilicity and distinct bonding preferences2. The C/Si replacement is a widely used structural modification strategy that enables the incorporation of silicon-containing functionalities into organic compounds, thus branching into new chemical space3,4. This approach significantly alters the molecular physicochemical properties and in vivo activity of drug candidates. For example, C/Si exchange in the molecule of 4-fluorophenylpropylpiperidine has shown different types of bioactivity5 (Fig. 1a). The incorporation of sila-proline in a Pro-rich peptide (PPII) has been shown to increase cellular uptake by 20-fold and resistance to proteolytic degradation6 (Fig. 1a). This approach has led to the development of a series of unnatural sila-amino acids that have been applied in the synthesis of therapeutically relevant compounds7. Moreover, the incorporation of the silyl amine group into indomethacin has significantly improved its bioactivity8,9.

Fig. 1: Catalytic regioselective silylalkylation of alkenes.
figure 1

a C/Si-Isosteric replacement; b Incorporation of silicon functionalities; c The state of the art of carbosilylation of alkenes; d Bisfunctionalization of terminal alkenes; e This work: Ni-catalyzed regioselective silylalkylation of simple alkenes.

Along with the shift from flat targets to three-dimensional structures in medicinal discovery10, the development of modular methods for the synthesis of sp3-rich organosilicon compounds has stimulated growing interest, leading to significant advances in the hydrosilylation of alkenes11,12,13,14 and cross-coupling of alkyl coupling partners with silicon reagents15,16,17,18,19,20,21,22,23,24. Multicomponent reactions have shown advantages in terms of the abundance of starting materials and efficiency towards complex targets25,26,27. One such approach involves trapping the intermediates of silyl-metal28,29 or silyl radical species30,31,32 with alkenes to introduce another group (Fig. 1c). This approach requires the use of alkenes with adjacent π-bonds, benefiting the formation of π-benzyl or π-allyl stabilized metal intermediates to ensure reactivity. Recently non-activated olefins have also been explored, and the organoboron-catalyzed carbosilylation reactions involving silylium ions are particularly noteworthy33,34. Transition metal-catalyzed reactions were also developed, but the installation of a strong coordinating auxiliary group into alkenes plays a crucial role in the success of these protocols35,36. However, certain functional groups and certain carbon lengths are required, leading to limitations in the alkene scope37,38. In addition, a regioselective arylsilylation of terminal alkenes was achieved by nickel chain-walking catalysis, but this reaction scope is limited to alkenyl arenes39. This reflects the ongoing challenge of anchoring a metal catalyst at the carbon chain without the assistance of strong directing groups, as the carbosilylation of simple non-activated alkenes is rare40.

On the other hand, alkene addition reactions through both 1,2-addition and 1,1-addition fashions have been widely recognized41,42, while the addition reaction of simple aliphatic olefins in a highly selective head-tail addition is still underdeveloped (Fig. 1d). Motivated by our success in the borylative functionalization of alkenes by nickel chain-walking catalysis43,44,45, we moved our interest to the silylative functionalization of alkenes. We hypothesized that if a highly regioselective silylative functionalization of simple unactivated alkenes was achieved, it would provide an efficient avenue for assembling organosilicon compounds from readily accessible materials. Herein, we present our successful development of a nickel-catalyzed regioselective silylcarbofunctionalization of simple alkenes, wherein a silyl group is introduced into terminal olefins concurrent with the construction of a C(sp3)-C(sp3) bond (Fig. 1e). The combination of a stable Ni(II) salt and the inexpensive trans−1,2-diaminocyclohexane enables different regioselectivities from different types of alkenes. This work not only provides an approach to enable access to diverse bifunctional building blocks from abundant materials but also demonstrates a significant advancement in the chain-walking transformation of simple olefins.

Results

Reaction development

In the study, the Ni-catalyzed silylalkylation of an alkene was chosen as a model to investigate the feasibility of the speculation mentioned above. Commercially available 1-octene (1a), the Suginome reagent (2a)46, and ethyl 4-bromobutyrate (3a) were used as model substrates. Ligand evaluation was conducted with NiBr2·DME as catalyst, CuI as cocatalyst, and LiOMe as an activator for the Si-B reagent47, in N-methylpyrrolidone (NMP) at 30 °C. It was found that the reaction chemoselectivity was greatly influenced by the nature of the used ligand. As shown in Fig. 2a, the ligands based on either 2,2’-bipyridine (L1), 1,10-phenanthroline (L2), bisoxazoline (L3), or 1,2-diphenylethylenediamine (L4) backbone failed to produce any three-component coupling products. In contrast, the ligands based on the trans−1,2-diaminocyclohexane backbone (L5-L9) were able to furnish a 1,8-regioselective silylalkylation product 4a, with L9 providing the optimal performance. The efficiency was further improved by the addition of KI, which was likely due to the coordination of the iodide anion inhibiting β-hydride elimination48. Further study found that the reactivity was highly sensitive to temperature: increasing the reaction temperature to 50 °C resulted in a very poor yield; while decreasing to 5 °C, the yield was improved to 82% with exclusive 1,8-regioselectivity, and this excellent terminal selectivity is kinetically controlled49. Notably, although the silyl group can also direct the formation of its α-metal complex50,51,52,53, no 1,1-regioselective isomeric product was detected with these conditions.

Fig. 2: Reaction development and substrate scope.
figure 2

a Reaction discovery; b Substrate scope. aReaction conditions: NiBr2·DME (5 mol%), L (7 mol%), CuI (15 mol%), 1a (0.4 mmol, 1.0 equiv), 2a (0.8 mmol, 2.0 equiv), 3a (0.6 mmol, 1.5 equiv), LiOMe (0.8 mmol, 2.0 equiv), NMP (2 mL), 30 °C, 36 h; b50 °C; c40 °C; d5 °C; e10 °C; fAddition of KI (0.5 equiv); gReaction conditions: NiBr2·DME (5 mol%), L9 (7 mol%), CuI (15 mol%), KI (0.5 equiv), 1 (0.4 mmol, 1.0 equiv), 2a (0.6 mmol, 1.5 equiv), 3 (0.8 mmol, 2.0 equiv), LiOMe (0.8 mmol, 2.0 equiv), NMP (2 mL), 5 °C, 36 h; hL8 instead of L9, KI (0.2 equiv); i30 °C, 36 h; jIsolated yield of the corresponding alcohol after oxidation.

Substrate scope

After the optimal reaction conditions were obtained, the generality of this Ni-catalyzed alkene silylalkylation reaction was explored. The scope of the alkyl electrophile was assessed first in combination with 1-octene (1a). As shown in Fig. 2b, an array of primary alkyl bromides bearing diverse substituents, including esters (4a-4c), ketyl (4d), ketones (4e and 4f), alkenylarene (4h), cyano (4i), alcohol (4j) and aryl chloride (4k), participated in this reaction to transfer the corresponding organosilicons in moderate to good yields with exquisite regioselectivity. Moreover, both cyclic (4l-4o) and linear (4p-4r) secondary bromides could also serve as successful coupling partners to deliver the corresponding products under identical reaction conditions. Both shortening the carbon length to five and extending the carbon length to eleven afforded the head-tail addition products with an identical level of reactivity (4q and 4r). Moreover, migratory 1,3-regioselective silylalkylation products, generated from thermodynamically favored alkyl intermediates, were isolated from allyl ether and allyl boronate54,55,56,57 (4s and 4w). When extending the carbon length of alkenyl boronate to four or six, the corresponding migratory 1,n-addition products were obtained with only a slight decrease in yields (4y and 4z). However, an increase in chain length led to deleterious reactivity for alkenyl ethers (4u and 4v), possibly due to the influence of the weak coordination of ethers.

To investigate the scope of the alkene component, a library of alkenes containing various functional groups was studied (Fig. 3). Primary (4aa-4ae), secondary (4af and 4ag), and tertiary (4ah-4aj) alkenyl alcohols reacted smoothly under the Ni-catalyzed system, producing the corresponding sila-alcohols in moderate to excellent yields (53%–95%) and with excellent regioselectivity. Even when the carbon chain length extended to seven or eleven (4ac and 4ad), the formation of carbon-carbon bonds at the γ-position of the alcohols was uniformly observed in these reactions. Notably, this observation is inconsistent with a recent alcohol-directed alkene difunctionalization reaction58, wherein 1,2-regioselectivity was uniformly obtained. Regretfully, allyl alcohols did not provide the corresponding silylation products under the current conditions (1ak and 1al). Comparing with the reactions with alkenyl ethers, we attribute this distinction to the stronger coordination ability of hydroxyl groups with metals compared to alkoxy groups, as demonstrated in a recent work by the Yu group59. In addition, the success could also extend to alkenyl esters, and the corresponding sila-esters (4am-4ap) were obtained in good yields with carbon-carbon formation exclusively at the β-position to the ester groups60. These results indicate that the involvement of five-membered nickelacycle intermediates in the reactions with alkenyl alcohols or alkenyl esters. Replacing the carboxylic acid group of amino acids with a silyl group can alter their biological activity61, making the synthesis of amino acid mimics highly valuable in medicinal chemistry. A series of amino acid mimics (4aq-4av) were efficiently prepared from vinyl or alkenyl amines. The transformation was also tolerant of various protecting groups, such as tert-butyloxy carbonyl (Boc), acyl (Ac), and benzyl (Bz). It is noteworthy that both the carbonyl and the nitrogen atom of amide groups can serve as directing groups62,63 in these mild reaction conditions (comparing 4as with 4au and 4av). Moreover, 3-butenamides afforded exclusive 1,2-addition products (4aw-4ay), but a poor site-selectivity was obtained when extending the chain length (4az)64,65. Comparing the reactions with alkenyl alcohols or alkenyl esters, these results indicate that the relatively stronger coordination ability of amide groups enables carbon-carbon formation from both five- and six-membered nickelacycle intermediates in this catalytic system66,67. Based on the above findings, regioselectivity-switchable reactions were achieved from homoallyl alcohol (Fig. 4a). In addition, endocyclic alkenes were able to deliver the corresponding cyclic sila-ethers (4bc and 4bd) in good yields with excellent stereochemistry, and a cyclic sila-ether (6) was efficiently synthesized by treating the product 4be with HCl gas (Fig. 4b).

Fig. 3: Scope of functionalized alkenes.
figure 3

Reaction conditions: NiBr2·DME (5 mol%), L9 (7 mol%), CuI (15 mol%), KI (0.5 equiv), 1 (0.4 mmol, 1.0 equiv), 2 (0.6 mmol, 1.5 equiv), 3 (0.8 mmol, 2.0 equiv), LiOMe (0.8 mmol, 2.0 equiv), NMP (2 mL), 5 °C, 36 h; a30 °C, 20 h.

Fig. 4: Extended scope and synthetic applications.
figure 4

a Switchable regioselectivity (1,2-addition vs. 1,4-addition); b Synthesis of cyclic sila-ethers; c Synthesis of silyl-substituted saturated aza-heterocycles; d Concise synthesis of alpha-lipoic acid.

Synthetic applications

The synthetic potential of this protocol was further demonstrated with downstream transformations of representative products. The first set of transformations involved the preparation of several silyl-substituted saturated aza-heterocycles, which are important pharmacophores ubiquitous in biologically relevant molecules68 (Fig. 4c). For instance, 2-substituted pyrrolidine (7) was prepared from 4bf, 3-substituted piperidine (8) from 4bg, and the Coniine69 analog 2-substituted piperidine (9) was obtained from 4bh. Finally, the synthetic value of this methodology was highlighted by simplifying the synthesis of a drug molecule (Fig. 4d). Alpha-lipoic acid (11) is a drug with numerous therapeutic functions, including the treatment of rheumatoid arthritis and Alzheimer’s disease70, and its previous synthetic routes required at least nine steps71,72. With this chemistry, a concise route toward alpha-lipoic acid was developed, and the new protocol only requires five steps from inexpensive commodity chemicals.

Mechanistic studies

Control experiments were carried out to determine the role of the nickel salt, copper salt, and ligand in the three-component reaction (Fig. 5a). The silicon reagent 2a was recovered from the reaction without copper, indicating that 2a can not directly react with nickel salt and the copper salt plays an important role in the transmetallation of the silicon reagent to nickel73,74 (Fig. 5b). Without a ligand, only C–Si bond formation products were detected, revealing the ligand’s crucial role in the formation of C–C bond. To shed light on the C–C bond formation step, radical probe experiments were conducted (Fig. 5c). The reaction with 6-bromo-1-hexene (3f) produced a cyclization product 4bj and a ring-opening product 4bk’ was dictated in the reaction with (bromomethyl)cyclopropane (3g), suggesting that the generation of alkyl radicals from the alkyl bromides is likely involved in the C–C bond formation step. Based on the results obtained from the experiments and previous studies39, a catalytic cycle for this bimetallic cooperative catalyzed reaction is proposed. As shown in Fig. 5d, the reaction begins with the transmetallation of the Sugimone reagent (2) with a copper salt (II) to form a Cu-[Si] species (III), which further transmetallates with a nickel salt to generate the Ni-[Si] species (V). Olefin migratory insertion is followed to furnish an alkyl-Ni(II) species (VI). When an alkene with a neighboring chelating group is used, this species then captures an alkyl radical (X) to produce a Ni(III) species (VII). Finally, reductive elimination delivers a 1,2-silylalkylation product (4) and a Ni(I) species (VIII). The intermediate VIII reacts with the alkyl halide (3) to generate the alkyl radical (X) and the Ni(II) catalyst (IV) and to close the cycle. When an aliphatic olefin or an alkene with a remote chelating group is used, rapid chain-walking occurs on the intermediate VI, leading to generate a thermodynamically stable species (VI’). Then the species undergoes radical addition and reductive elimination (via VII’) to deliver the corresponding migratory product.

Fig. 5: Mechanistic investigations and proposed catalytic cycle.
figure 5

a Control experiments; b Transmetalation of the Suginome reagent; c Radical probe experiment; d Proposed mechanism.

Discussion

In conclusion, a highly regioselective silylalkylation of non-activated terminal alkenes has been developed, producing an efficient approach to sp3-rich organosilicon building blocks. The reaction regioselectivity is determined by the alkene and the choice of bisnitrogen-ligated nickel catalyst is crucial to the success of these transformations, leading to mild reaction conditions, exquisite site-selectivity, wide substrate scope, and exceptional functionality tolerance. Follow-up transformations of selected products demonstrate the synthetic potential of this chemistry, including the preparation of important pharmacophores and the streamlining of a synthetic route for a drug molecule. Overall, this study has the potential to significantly impact medicinal chemistry research.

Methods

Model procedure

Under N2 atmosphere in a glove box, an oven-dried 10 mL reaction tube equipped with a magnetic stir bar and sealed with a rubber stopper was used. Sequentially, NiBr2·DME (6.2 mg, 0.02 mmol, 5 mol%), L9 (10.0 mg, 0.028 mmol, 7 mol%), LiOMe (30.4 mg, 0.8 mmol, 2.0 equiv), CuI (11.4 mg, 0.06 mmol, 15 mol%), KI (33.0 mg, 0.20 mmol, 0.5 equiv) and PhMe2SiBpin (220.8 mg, 0.8 mmol, 2.0 equiv) were added. Subsequently, anhydrous NMP (1 mL), alkenes (0.4 mmol, 1.0 equiv), alkyl bromide (0.6 mmol, 1.5 equiv) and anhydrous NMP (1 mL) were added in sequence, and the mixture was stirred at room temperature. After stirring at 5 °C for 36 h (or stirring at 30 °C for 24 h), the resulting mixture was quenched with water (2 mL) and further diluted with ethyl acetate (3 mL). The mixture was then extracted with ethyl acetate, and the combined organic layers were dried over anhydrous Na2SO4, filtered, and concentrated under a vacuum. The crude material was separated on a silica gel column to afford the desired product.