Nickel-catalysed enantioselective alkene dicarbofunctionalization enabled by photochemical aliphatic C–H bond activation

The development of novel strategies to rapidly construct complex chiral molecules from readily available feedstocks is a long-term pursuit in the chemistry community. Radical-mediated alkene difunctionalizations represent an excellent platform towards this goal. However, asymmetric versions remain highly challenging, and more importantly, examples featuring simple hydrocarbons as reaction partners are elusive. Here we report an asymmetric three-component alkene dicarbofunctionalization capitalizing on the direct activation of C(sp3)–H bonds through the combination of photocatalysed hydrogen atom transfer and nickel catalysis. This protocol provides an efficient platform for installing two vicinal carbon–carbon bonds across alkenes in an atom-economic fashion, providing a wide array of high-value chiral α-aryl/alkenyl carbonyls and phosphonates, as well as 1,1-diarylalkanes from ubiquitous alkane, ether and alcohol feedstocks. This method exhibits operational simplicity, broad substrate scope and excellent regioselectivity, chemoselectivity and enantioselectivity. The compatibility with bioactive motifs and expedient synthesis of pharmaceutically relevant molecules highlight the synthetic potential of this protocol.

The efficient assembly of high-added-value chiral molecules from readily available, inexpensive hydrocarbons with minimal waste generation represents a long-standing challenge for synthetic chemists 1,2 .Two major challenges have hampered the broad implementation of these processes: on the one hand, C(sp 3 )-H bonds present a much lower reactivity compared to most functional groups.Furthermore, the presence of multiple non-equivalent C(sp 3 )-H bonds requires exquisite control over the chemoselectivity, regioselectivity and stereoselectivity to attain the desired outcome 3,4 .Recently, hydrogen atom transfer (HAT) photocatalysis has emerged as a mild strategy for direct C(sp 3 )-H functionalizations 5 .Upon irradiation with light, the excited state of a photocatalyst can abstract hydrogen atoms from strong and inert C(sp 3 )-H bonds to produce nucleophilic carbon-centred radicals.Easy-to-synthesize tetra-n-butylammonium decatungstate (TBADT) and diaryl ketones have attracted particular interest as HAT photocatalysts owing to their ability to cleave electronically and sterically accessible C-H bonds with high regioselectivity 6-9 .As a result, synthetically useful C(sp 3 )-H functionalizations promoting the formation of new C-C (refs.[10][11][12][13][14][15][16], C-O (refs.17,18), C-N (refs.19,20), C-F (refs.21,22) and C-S (ref.23) bonds have been realized capitalizing on this strategy.

Article
https://doi.org/10.1038/s41929-024-01153-0solvent and/or the participating reaction partners compromise the efficiency and selectivity of the overall process.
Here we show that the combination of a decatungstate or a diaryl ketone HAT photocatalyst with a chiral biimidazoline (BiIm) nickel catalyst enables the asymmetric dicarbofunctionalization of alkenes so that valuable α-aryl/alkenyl carbonyls and phosphonates as well as 1,1-diarylalkanes can be obtained in enantioenriched form through the regioselective activation of abundant hydrocarbons and their derivatives.
Catalytic three-component radical alkene difunctionalizations represent a straightforward and efficient approach for the rapid construction of molecular complexity, as two new functional groups can be simultaneously installed across the π system in a single operation 26-36 .However, controlling the enantioselectivity of the generated stereogenic centres remains a formidable challenge owing to the high reactivity and instability of the open shell radical intermediates participating in these transformations.Chiral copper 37-43 and nickel 44-52 complexes have opened a robust platform for asymmetric radical conjunctive cross-coupling of olefins.For example, impressive progress has been made in nickel-catalysed asymmetric radical relayed reductive coupling 46-49 , in which C(sp 3 )-hybridized electrophiles such as alkyl halides and perfluoroalkyl halides have been employed as radical precursors.In addition, processes merging photoredox catalysis and nickel catalysis have enabled the asymmetric redox-neutral dicarbofunctionalization of alkenes with alkyltrifluoroborates as competent synthons for these protocols 50,51 .All of the aforementioned transformations strongly rely on the use of pre-activated radical precursors, use that not only leads to the generation of additional waste but also results in poor step economy and atom economy.To improve the synthetic efficiency and expand the pool of radical precursors, the direct use of abundant hydrocarbons as coupling partners in the asymmetric dicarbofunctionalization of alkenes represents a more ideal but also highly challenging approach.
Although synergistic HAT processes and transition metal catalysis have been used in the enantioselective arylation, alkylation, alkenylation, acylation and cyanation of C(sp 3 )-H bonds (Fig. 1a) 53-64 , the direct activation of C(sp 3 )-H bonds towards the asymmetric difunctionalization of alkenes is certainly underdeveloped (Fig. 1b).Two main challenges need to be addressed: first, the identification of a chiral ligand that can overcome competitive two-component side reactions while efficiently imparting absolute stereocontrol in the formation of the new stereogenic centres in the three-component process; and second, the implementation of reaction conditions wherein no competitive activation of distinct, even more activated C-H bonds present in the Article https://doi.org/10.1038/s41929-024-01153-0nickel catalysts such as NiBr 2 •diglyme and NiBr 2 •3H 2 O delivered the product with slightly decreased yield and e.r. (Fig. 2, entries 10 and 11), while NiCl 2 •DME and NiCl 2 (Py) 4 led to a dramatically reduced yield and enantioselectivity (Fig. 2, entries 12 and 13).The reaction still proceeded smoothly and provided a comparable e.r.upon dilution (Fig. 2, entries 14 and 15).Control experiments confirmed that photocatalyst, nickel, ligand and light were all necessary for a successful outcome (Fig. 2, entry 16).The absolute configuration of product 1 was unambiguously confirmed by X-ray diffraction analysis.In parallel, optimal conditions for isopropanol as radical precursor were sought.Gratifyingly, the desired three-component coupling product 2 could be obtained in 85% yield and 96:4 e.r. in the presence of diaryl ketone PC I, NiBr 2 •DME, L5 and Na 2 CO 3 under 40 W, 390 nm LED irradiation at 5 °C.The influence of ligand, photocatalyst and solvent is shown in Fig. 2, entries 18-20.
The scope of C-H precursors was explored next (Fig. 4).As expected, cycloalkanes with various ring sizes such as cyclopentane (46), cycloheptane (47), cyclooctane (48) and cyclododecane (49) were well tolerated.In the last case, because of the lack of solubility of the reagent, an e.r. of 91:9 was obtained under standard reaction conditions, which was improved to 96:4 by increasing the amount of solvent and the use of L5 as ligand.Interestingly, 2,3-dimethylbutane (50) was selectively activated on the secondary C-H bonds owing to the high bond dissociation energies of the primary C-H bonds.Remarkably, hydrocarbon derivatives bearing different electron-withdrawing groups

Article
https://doi.org/10.1038/s41929-024-01153-0(-Cl, -Br, -CN, -COMe, -COOMe) were successfully functionalized with excellent control of the regioselectivities and enantioselectivities (51)(52)(53)(54)(55).This result can be attributed to the hydridic nature of tertiary C-H bonds and the electrophilic nature of excited decatungstate 65 , thus providing products that are selectively functionalized distal to the electron-withdrawing moieties.Complete regioselectivity was also observed in the case of alcohols containing tertiary C-H bonds, with 2,3-dimethylbutan-2-ol affording product 56 in 81% yield and 96:4 e.r.The compatibility of diverse functional groups further emphasizes the advantages of this photo-HAT/nickel dual-catalysed strategy, as these types of alkyl radicals were inaccessible with previously developed radical precursors.This protocol is not restricted to electron-neutral, unactivated C-H systems.Ethers were regioselectively functionalized at the α-oxy position (57 and 58), which is consistent with the selectivity of decatungstate for more electron-rich C-H bonds.However, methyl tert-butyl ether, toluene and Boc-protected pyrrolidine were not able to participate in this three-component reaction, producing two-component aryl-alkyl cross-coupling products instead (Supplementary Table 9).The aforementioned reaction conditions were not suitable for the direct functionalization of hydrogen bonds in the α-position to hydroxy functions.Inspired by literature precedents 25 , we were delighted to find that isopropanol (2) and pentan-3-ol (59) participated in this transformation when benzophenone derivative PC I was used as the photocatalyst under modified reaction conditions.Furthermore, extraordinary regiocontrol was achieved in the presence of both tertiary and primary C-H bonds adjacent to the oxygen atom, as only the target product 60 was formed when cyclopentyl methyl ether was used as substrate.This result highlights the ability of diaryl photocatalysts in site-selective activation.We next turned our attention to expanding the scope with respect to alkenes (Fig. 4).A series of acrylates with different substituents was investigated, generating highly enantioenriched α-aryl ester products (61)(62)(63)(64)(65)(66).Among them, a slight erosion in enantioselectivity was observed with less sterically hindered acrylate reactants, which was addressed by replacing ligand L1 with L5.Pleasingly, vinyl phosphonates were also suitable coupling partners under either decatungstate or diaryl ketone photocatalytic conditions.Chiral organophosphonates were obtained in good yields and excellent enantioselectivities (67 and 68).Other electron-deficient alkenes were also explored.The α,β-unsaturated ketones as well as acrylamides were viable electron

Synthetic applications
The obtained α-aryl/alkenyl carbonyls and phosphonates as well as 1,1-diarylalkanes are often found as prevalent structural motifs in pharmacologically and biologically active molecules 66-69 .To further demonstrate the utility of this asymmetric multicomponent reaction, transformations of the products (Fig. 5a) and synthesis of pharmaceutically relevant molecules were performed.Specifically, through acidic hydrolysis followed by esterification or amidation processes, l-tyrosine and l-tryptophan derivatives (76 and 77) were obtained with high levels of diastereocontrol, which indicated the potential of this method for rapidly accessing complex amino acid derivatives.In addition, the reduction of α-aryl ester 3 was successfully accomplished by treatment with LiAlH 4 , delivering alcohol 78 with excellent stereofidelity (97:3 e.r.).Because alcohols are versatile intermediates in organic synthesis, several derivatizations of 78 were also carried out.Nucleophilic substitution reactions converted 78 into the corresponding β-aryl bromide (79), β-aryl amide (80) and β-aryl thiol (81) in high yields and enantioselectivities.To illustrate its practical value further, the method was applied to a concise synthesis of two glucokinase activators.The synthetic route towards piragliatin lead compound is shown in Fig. 5b (top).The combination of commercially available cyclopentane, tert-butyl acrylate and 4-bromo-2-chloro-1-(methylsulfonyl)benzene produced the corresponding three-component cross-coupling product 83 in 65% yield and 97:3 e.r.Hydrolysis of 82 with trifluoroacetic acid followed by amidation with 2-aminopyrazine offered piragliatin lead compound 83 in 72% yield and 97:3 e.r.Compared to previous reports, this method simplifies the operation steps by using cyclopentane instead of alkyltrifluoroborate as the synthetic precursors 50 .Similarly, enantioenriched glucokinase activator RO28-1675 (85, 97:3 e.r.) was successfully synthesized by using the abovementioned three-step synthetic procedure (Fig. 5b, bottom).

Mechanistic studies
Several experiments were designed to probe the reaction mechanism of this dual-catalytic transformation.The desired alkylarylation reaction was completely suppressed by adding 2,2,6,6-tetramethyl-1-piper idinyloxy (TEMPO; 3.0 equiv.).TEMPO adducts 86 and 87 were detected by high-resolution mass spectrometry (HR-MS) in the mixture (Fig. 6a), thus demonstrating the intermediacy of an alkyl radical along the reaction pathway.Next, radical clock experiments were conducted with cyclopropyl alkene 88 under the standard reaction conditions, affording product 89 in 20% yield through a sequential radical addition, ring opening and arylation process (Fig. 6b).Product deracemization or Giese-type addition of radicals to the double bond followed by enantioselective α-arylation were ruled out based on controlled experiments (Supplementary Discussion).Subsequently, the reactivity and catalytic efficiency of presynthesized 4-CF 3 C 6 H 4 Ni(II)L1Br complex 90 was further investigated.Only a trace amount of 7 was detected with stoichiometric complex 90, cyclohexane and tert-butyl acrylate.One possible reason is that complex 90 is unstable in the absence of aryl bromides, and it possesses strong visible-light absorption properties 54 .By contrast, using 10 mol% of complex 90 instead of NiBr 2 •DME/L1 as catalyst in the presence of aryl bromides, product 7 was obtained in 67% yield with an identical e.r.(95:5; Fig. 6c).Taken together, these results indicate that the putative aryl-Ni(II) complex might not be a productive intermediate in the main catalytic cycle.In addition, laser flash photolysis experiments were performed to study the quenching of the excited state of TBADT (TBADT*) in the presence of increasing concentrations of cyclohexane, tert-butyl acrylate and 4-bromobenzonitrile, respectively.A clear decay of TBADT* was observed in the presence of cyclohexane following a linear Stern-Volmer behaviour (bimolecular rate constant k = 3.14 × 10 7 M −1 s −1 ).By contrast, tert-butyl acrylate and 4-bromobenzonitrile were not able to quench TBADT* (Fig. 6d).These results confirm the activation of cyclohexane by the excited state of the photocatalyst to form carbon radicals, in line with the radical trapping and clock experiments summarized in Fig. 6a,b.
On the basis of the aforementioned mechanistic studies and precedents in the literature 24 , a plausible mechanism for this TBADT/ nickel dual-catalysed enantioselective alkene dicarbofunctionalization can be proposed (Fig. 6f).Excited-state tetrabutylammonium decatungstate I would be formed under photoexcitation conditions, enabling the abstraction of a hydrogen atom from C(sp 3 )-H nucleophiles.This process generates singly reduced decatungstate II and carbon-centred radical III.Disproportionation of II provides doubly reduced decatungstate IV and regenerates ground-state TBADT.In parallel, radical addition of III to an olefin would form adduct V, which can be captured by Ni(0) species VI to furnish alkyl-Ni(I) intermediate VII.Subsequent oxidative addition of alkenyl bromide to Ni(I) species VII generates alkyl-Ni(III)-aryl intermediate VIII.Reductive elimination produces the desired dicarbofunctionalization product as well as Ni(I) species IX.Both catalytic cycles converge to completion through final single-electron transfer between this Ni(I) species and doubly reduced TBADT IV to regenerate the active Ni(0) catalyst VI and singly reduced TBADT II.This process is supported by cyclic voltammetry studies (Fig. 6e), as the reductive potential of [W 10 O 32 ] 5-/ [W 10 O 32 ] 6-(E 1/2 = -1.85V versus Fc/Fc + in MeCN, reductive peak observed at -1.93 V) is more negative than that of Ni I /Ni 0 (E 1/2 = -1.74V versus Fc/Fc + in MeCN, reductive peak observed at -1.82 V).A Ni I /Ni II /Ni III catalytic sequence is also possible (Fig. 6f, grey).Alternatively, Ni(I) species VII could be obtained through the combination of Ni(I) species IX with radical adduct V, followed by the subsequent single-electron reduction of alkyl-Ni(II) intermediate X.

Conclusions
In summary, the combination of photoredox-mediated HAT with a nickel-catalysed radical relay has enabled a highly enantioselective three-component difunctionalization of alkenes.Unactivated hydrocarbons can be directly used as coupling agents without the requirement of pre-activated radical precursors, making this strategy more attractive for step-economical and atom-economical synthesis.Site-selective activation and the conversion of alkanes, ethers and alcohols were achieved by employing decatungstate or diaryl ketone photocatalysts.The method exhibits mild conditions, broad substrate scope and excellent regioselectivity, chemoselectivity and enantioselectivity, providing a versatile synthesis of value-added enantioenriched α-aryl/ alkenyl carbonyls/phosphonates and 1,1-diarylalkanes.The synthetic potential of this method has been demonstrated by derivatization of the products and the synthesis of medicinally relevant molecules.We believe that the lessons obtained here will inspire the development of more asymmetric multicomponent reactions enabled by C(sp 3 )-H activation in the future.

Fig. 3 |
Fig. 3 | Scope of (hetero)aryl/alkenyl bromides.The reaction and conditions are shown at the top.a Standard reaction conditions I are the same as in Fig. 2. b L5 was used as ligand, and acetone/PhCF 3 (1 ml:1 ml) was used as solvent.c NiBr 2 •3H 2 O was used instead of NiBr 2 •DME.d Two 40 W 390 nm Kessil lamps were used.d.r., diastereomeric ratio.

Fig. 4 |
Fig. 4 | Scope of C-H precursors and alkenes.The reaction and conditions are shown at the top.a Standard reaction conditions I are the same as in Fig. 2. b L5 was used as ligand, and acetone/PhCF 3 (1 ml:1 ml) was used as solvent.c Standard reaction conditions II are the same as in Fig. 2. d Two 40 W 390 nm Kessil lamps were used.