Introduction

Cephalotane diterpenoids, characterized by unique and intricate carbon skeletons and notable antitumor activities, are distinctive compounds of the Cephalotaxus genus, comprising over 110 members1,2,3,4,5,6. Benzenoid cephalotane-type diterpenoids are a relatively novel subtype (Fig. 1a), as the isolation, named cephanolides A-D, was reported in 2017 by Yue and co-workers7. Up to now, eight other benzenoid cephalotane-type diterpenoids have been isolated from the seeds of Cephalotaxus fortunei var. Alpina by the same group2,6. Structurally, these molecules have a unique 6/6/6/5 tetracyclic core embedded with a bridged δ-lactone, containing5,6,7 contiguous stereogenic centers (Fig. 1b). In addition, the oxidation states of benzenoid cephalotane-type diterpenoids exhibit diversity at the C3, C7, C10, C13, and C20 positions.

Fig. 1: Background and reaction design.
figure 1

a Selected examples of benzenoid cephalotane-type diterpenoids. b Core skeleton of benzenoid cephalotane-type diterpenoids. c Representative cases of total syntheses of benzenoid cephalotane-type diterpenoids. d Our convergent synthetic strategy.

Full size image

Due to their intricate and diverse structures, as well as their potential biological activities, benzenoid cephalotane-type diterpenoids have garnered significant attention from synthetic chemists (Fig. 1c). Synthetic studies have been comprehensively reviewed by Yue and co-workers8. The total synthesis of benzenoid cephalotane-type diterpenoids was reported by Zhao and co-workers9. They utilized a palladium-catalyzed cascade cyclization reaction to forge the 6-5-6 cis-fused tricyclic core and successfully achieved the synthesis of (±)-cephanolide B and C (2-3) in 2018. The asymmetric total synthesis of cephanolide A (1) was achieved by Gao and co-workers through an intramolecular Prins cyclization, followed by a cation-mediated etherification/Friedel–Crafts cyclization10. Later, they also realized the asymmetric total synthesis of cephanolide B (2) using the same strategy11. Guided by chemical network analysis, Sarpong and co-workers realized the divergent total synthesis of (±)-cephanolides A-D (1-4) by taking advantage of an intramolecular Diels–Alder cycloaddition to forge the core skeleton12. One year later, they optimized the synthesis method to achieve the synthesis of (±)-cephanolides A-D (1-4), ceforalides C-D, and F-G13. The asymmetric total synthesis of cephanolides A-B (1-2) was completed by Cai and co-workers via an inverse-electron-demand Diels–Alder reaction14. Using a tandem intramolecular Pauson-Khand reaction, 6π-electrocyclization reaction, and oxidative aromatization strategy, ( ± )-cephanolide B (2) and cephanolides B-D (2-4) were synthesized by Yang’s group15 and Hu’s group16, respectively. The asymmetric total synthesis of cephanolides A-D (1-4) was reported by Zhai and co-workers via a palladium-catalyzed formal bimolecular [2 + 2 + 2] cycloaddition reaction17. In addition, other subgroups of cephalotaxus diterpenoids have also been extensively studied8,18,19,20,21.

A challenging synthetic strategy was envisioned wherein the core skeleton of benzenoid cephalotane-type diterpenoids was disconnected into two intermediates, such as aryl iodide and bridged δ-lactone, by breaking three C-C bonds (Fig. 1d). The strategy has its own unique challenges, which involves a cascade C(sp2) & C(sp3)–H activation and the precise construction of three C–C bonds and two chiral centers. The cascade C(sp2) & C(sp3)–H activation strategy, if effective, might also prove valuable for total syntheses of a variety of natural products featuring the 6/6/6/5 tetracyclic skeleton.

As widely acknowledged, C–H bond activation has been regarded as the Holy Grail of organic chemistry and the ideal strategy of organic synthesis22,23,24,25. The selective cleavage of C–H bonds enables the efficient formation of other functional groups or novel chemical bonds, thereby significantly enhancing the synthesis efficiency and atom economy for natural product syntheses26,27,28,29,30,31. The Catellani reaction, renowned as a classical C(sp2)–H bond activation reaction, was discovered by Catellani32 in 1997 (Fig. 2a). The reaction has been expanded by Catellani33, Lautens34,35,36, Dong37,38,39, and others40,41 as a very powerful tool used in the synthesis of complex molecules featuring polysubstituted arenes. To date, numerous synthetic chemists have employed the Catellani reaction to accomplish the sophisticated synthesis of natural products42,43,44,45,46,47,48,49,50,51,52,53,54.

Fig. 2: C(sp2) & C(sp3)–H activation and retrosynthetic analysis of benzenoid cephalotane-type diterpenoids.
figure 2

a Catellani reported a C(sp2)–H bond activation reaction. b Catellani reported C(sp3)–H bond activation reaction. c Retrosynthetic analysis of benzenoid cephalotane-type diterpenoids. FGI functional group interconversion.

Full size image

Compared the C(sp2)–H bond activation in the Catellani reaction, the activation of C(sp3)–H bonds has been less explored, particularly in the formation of C(sp3)–C(sp3) bonds through the activation of a C(sp3)–H by transient alkylpalladium55,56. (Fig. 2b) The main challenges lie in the C(sp3)–H bond possessing high bond dissociation energy and a lack of stabilizing orbital interactions with the metal center57,58. Herein, we present the realization of the rare cascade C(sp2)–H and C(sp3)–H activation via Pd(0)/NBE-cocatalysis, resulting in the concise and convergent synthesis of cephanolides A-D (1-4) and ceforalide B (5).

The key cascade C(sp2)–H and C(sp3)–H activation reaction, featuring the construction of three C–C bonds and two chiral centers accurately, allows access to the core skeleton of benzenoid cephalotane-type diterpenoids, such as 10a and 10b, from iodobenzene derivatives (11a and 11b) and alkyl bromide 12 in one step, respectively (Fig. 2c). Thus, the asymmetric total synthesis of benzenoid cephalotane-type diterpenoids could be easily achieved from 10a and 10b via a late-stage modification strategy. The alkyl bromide 12 could be constructed from cyclohexenone 13 via reduction and acetalization. Ketone 13 could be prepared by Bamford-Stevens reaction and subsequent functional group interconversion (FGI) from compound 14, which could be derived from a known compound 1559.

Results

Our synthesis work commenced with the preparation of C10 unit acetal 12 from the known chiral compound 15, which was prepared from cyclohexenone in one step with 75% yield and 93% ee59 (Fig. 3).

Fig. 3: Synthesis of key intermediate 12.
figure 3

3, 5-DMP: 3,5-dimethylpyrazole.

Full size image

The alkylation of compound 15 with a 2-iodoethanol derivative, conditions previously reported by Zhu60, furnished compound 14 as a single diastereomer in up to 62% yield on decagram scale. Exposure of 14 to a one-pot olefin formation61/deprotection protocol yielded alcohol 16, which was then subsequently treated with CBr4 and PPh3, followed by CrO3 and 3,5-DMP to obtain cyclohexanone 13, along with a small amount of compound 17. Noteworthy, olefin 17 can convert to cyclohexanone 13 through the same oxidation. Cyclohexanone 13 was reduced by CeCl3·7H2O and NaBH4 to produce allyl alcohol 18, which was transformed to lactone 19 in the presence of NaH. Finally, to avoid potential stereoselectivity issues during the cascade reaction, lactone 19 was further converted to acetal 12, which has been proven effective in precisely constructing stereochemistry at the C1 and C10 positions by Zhao8, Yang14, and Hu15. Thus, lactone 19 was reduced by DIBALH to produce the corresponding aldehyde, which was then directly transformed into acetal 12 in the presence of chloroform and methanol to prevent the formation of the dimethyl acetal byproduct.

Cascade C(sp2) & C(sp3)–H activation

With acetal 12 in hand, our focus turned toward the pivotal palladium/NBE-cocatalyzed cascade C(sp2) & C(sp3)–H activation for the assembly of the tetracyclic skeleton. First, the reaction mechanism was proposed62,63,64 (Fig. 4). Iodobenzene derivative 11 and Pd(0) undergo oxidative addition to generate arylpalladium A, which is followed by insertion into NBE to produce intermediate B. Subsequently, C(sp2)–H activation generates aryl-norbornyl-palladacycle (ANP) C. Next, the ANP C undergoes oxidative addition with compound 12 to produce Pd(IV) intermediate D, followed by reductive elimination to obtain ortho-alkylated intermediate E. Then, the β-carbon elimination of intermediate E leads to the formation of intermediate F, which subsequently proceeds through an intramolecular migratory insertion to generate the crucial transient secondary alkylpalladium G. In the last stage, transient secondary alkylpalladium H undergoes a concerted metalation deprotonation (CMD) process to yield six-membered palladacycle I, and subsequent reductive elimination to provide the desired 10.

Fig. 4: Proposed mechanism of cascade C(sp2) & C(sp3)–H activation.
figure 4

Arylpalladium (A), intermediate (B), aryl-norbornyl-palladacycle (ANP) (C), Pd(IV) intermediate (D), ortho-alkylated intermediate (E), intermediate (F), transient secondary alkylpalladium (G), transient secondary alkylpalladium (H), six-membered palladacycle (I). OA oxidative addition, RE reductive elimination.

Full size image

As shown in Table 1, an initial attempt with tri(2-furyl)phospine as the ligand and Cs2CO3 as the base was conducted at 110 °C. The C(sp2)–H activation product 20a was obtained in 43% yield along with the nucleophilic substitution byproduct 21, but the cascade C(sp2) & C(sp3)–H activation product was not observed (entry 1). When the reaction temperature was raised to 120 °C, the desired compound 10a was obtained, albeit accompanied by a significant amount of compounds 20a and 21 (entry 2). Further increasing the reaction temperature did not achieve a better result (entry 3). The yield of byproduct 21 indicates that nucleophilic substitution by acetate anions is fast, rendering them unable to participate in the CMD process of crucial C(sp3)–H activation. Therefore, to inhibit nucleophilic substitution and promote the CMD process, KOPiv, which has a larger steric effect, was added. However, the yield of 10a/20a (1/4.6) decreased to 34%, while the yield of 21/22 increased to 44% (entry 4). This result may indicate that the nucleophilic substitution of acetal 12 with carboxylate is much faster than the oxidative addition of acetal 12 with aryl-norbornyl-palladacycle (ANP) C (Fig. 4). To mitigate the tendency of nucleophilic substitution of acetal 12, PdCl2 was used instead of Pd(OAc)2 as a catalyst (entry 5). Although the nucleophilic substitution of acetal 12 is inevitable, the yield of nucleophilic substitution product 22 is slightly reduced. Subsequently, the dosage of KOPiv and delayed addition time were screened to promote the oxidative addition of 12 with ANP C (entries 6-10). It was found that the reaction proceeded more efficiently with 1.8 equivalents KOPiv when a 60 s delay was implemented for the addition (entry 7). Next, a few other ligands were tested; however, no improved results were obtained (entries 11-14). In the screening of solvents (entries 15-19), the optimal result was achieved using a mixed solvent of DMF and PhMe in a 1:1 ratio (entry 18). Notably, on a hundred-milligram scale, compound 10a was obtained in 42% yield, and the by-product 20a was not observed under these conditions (entry 20). Compound 10b was obtained when 11b reacted with acetal 12, with a delay in the time of adding KOPiv (entry 21). It is worth mentioning that the nucleophilic substitution compounds 21 and 22 could be reconverted to compound 12 by reduction and bromination (see Supplementary Information).

Table 1 Investigation of the palladium-mediated cascade C(sp2) & C(sp3)–H activationa
Full size table

Total synthesis of cephanolides A-D and ceforalide B

After obtaining the key compounds 10a and 10b, benzenoid cephalotane-type diterpenoids were synthesized through late-stage modification. Consequently, 10a was oxidized using BF3·Et2O and m-CPBA to produce lactone 23, which was subsequently deprotected with BBr3 to yield cephanolide B (2)9,15,16 (Fig. 5a). For the synthesis of cephanolide A (1), compound 10a was hydrolyzed in the presence of 3 M HCl (aq.)65 resulting in compound 24. The structure of 24 was confirmed by X-ray crystallographic analysis (see Supplementary Information). Compound 24 underwent Swern oxidation and Pinnick oxidation to yield compound 25. The Davis oxaziridine has been demonstrated to be efficacious in introducing a hydroxyl group at the C3 position17, resulting in compound 26 with 71% yield. The carbonyl of compound 26 was reduced by sodium borohydride and then underwent lactonization to give lactone 27 in the presence of Mukaiyama reagent. Lactone 27 was subjected to Suárez oxidation12,14,17,66 to form the tetrahydrofuran ring of compound 28, followed by the removal of the protective group with BBr3 to yield cephanolide A (1). From compound 10b, three additional benzenoid cephalotane-type diterpenoids were synthesized (Fig. 5b). Compound 10b was oxidized to produce lactone 29, which was then converted to cephanolide C (3) using the NHPI /Co(OAc)2-catalyzed oxidation as reported by Hu16. To efficiently introduce a methyl ester group in the late stage, the selective bromination at the C15 position was attempted. The expected compound 31a was obtained in a modest yield using TBCA/TFA67, along with a small amount of nonseparable C13 brominated by-product 31b. Finally, the Pd-catalyzed methoxycarbonylation12,17 resulted in compounds 32a and 32b, which were further oxidized to cephanolide D (4) and ceforalide B (5) via NHPI/Co(OAc)2-catalyzed oxidation.

Fig. 5: Total Synthesis of Benzenoid Cephalotane-type Diterpenoids.
figure 5

a Total Synthesis of Cephanolides A-B (1-2). b Total Synthesis of Cephanolides C-D (3-4) and Ceforalide B (5). m-CPBA = m-chloroperbenzoic acid, 4 Å MS = 4 Å molecular sieves, DCM = dichloromethane, DMSO = dimethyl sulfoxide, THF = tetrahydrofuran, KHMDS = potassium bis(trimethylsilyl)amide, BRSM = based on recovered starting material, CMPI = 2-chloro-1-methylpyridinium iodide, PIDA = phenyliodine(III) diacetate, NHPI = N-hydroxyphthalimide, o-DCB = o-dichlorobenzene, TBCA = tetrabromocinnamic acid, dppf = 1,1’-bis(diphenylphosphino)ferrocene, DMF = N,N-dimethylformamide, TBAB = tetrabutylammonium bromide.

Full size image

Discussion

In summary, we have developed a cascade C(sp2)–H and C(sp3)–H activation strategy to efficiently access the characteristic and synthetically challenging 6/6/6/5 tetracyclic skeleton of the benzenoid cephalotane-type diterpenoids, which was used to complete a short total synthesis of cephanolide B (2). A late-stage modification and divergent strategy were employed to complete the syntheses of four additional benzenoid cephalotane-type diterpenoids. The key feature of the cascade C(sp2) & C(sp3)–H activation reaction is the construction of three C–C bonds (two C(sp2)–C(sp3) bonds and a C(sp3)–C(sp3) bond) and two chiral centers accurately in one step via palladium/NBE-cocatalysis. Analogous strategies may prove effective in the preparation of various natural products. This encourages us to further investigate the synthesis of natural products through cascade C(sp2) & C(sp3)–H activation.

Methods

Cascade C(sp2) & C(sp3)–H activation

In a 50 mL round-bottom flask, 11a (182.0 mg, 0.9 mmol), PdCl2 (32.0 mg, 20 mol%), tri(2-furyl)phospine (83.6 mg, 40 mol%), Cs2CO3 (880.0 mg, 2.7 mmol) were added and charged with argon more than three times (The flask was plugged with rubber plugs and sealed with a parafilm). Then, 5.5 mL DMF was injected into the flask via plastic syringes. The resulting light-yellow suspension was stirred vigorously at room temperature for 15 min. Norbornene (339.0 mg, 3.6 mmol) and compound 12 (353.0 mg, 1.35 mmol) were dissolved in 5.5 mL DMF, and the mixture was injected into the tube via plastic syringes. Subsequently, the reaction tube was placed in an oil bath at 120 oC, stirred for 2 min, and then a suspension of potassium pivalate (11.0 mL in toluene) was added over 50 s. After the reaction was completed (4 h), the flask was cooled to room temperature and filtered to remove the solid residue. The filtrate is concentrated under reduced pressure, and the residue is purified by silica gel column chromatography (petroleum ether/ethyl acetate =  50/1 to 10/1) to afford compound 10a (119.1 mg, 42%, white solid) and compound 22 (244.5 mg, 64%, colorless oil).