Palladium-catalyzed enantioselective β-hydride elimination for the construction of remote stereocenters

Abstract

The β-H elimination is a crucial elementary step in transition-metal catalysis, but controlling the stereochemistry of this process has been underdeveloped. The limited works reported so far have only focused on creating axial chirality in allenes, and no report has been able to build central chirality using asymmetric β-H elimination. In this study, we report a Trost ligand-enabled enantioselective desymmetric β-H elimination reaction from π-allyl-Pd. This transformation provides rapid access to cyclohexenes bearing a C4-remoted stereocenter, and total synthesis of (-)-oleuropeic acid and (-)-7-hydroxyterpineol is demonstrated. Computational studies have shown that the β-H elimination is the rate-determining step, and the non-covalent interactions between the amide moiety of the Trost ligand and the benzene and cyclohexane moieties of the substrate play a key role in stereocontrol during the β-H elimination.

Beta-hydride elimination is a crucial elementary step in transition-metal catalysis, but controlling the stereochemistry of this process has been underdeveloped. Here, the authors report a Trost-ligand-enabled enantioselective desymmetric beta-hydride elimination reaction from π-allyl-palladium species, providing rapid access to cyclohexenes bearing a remote C4 stereocenter.

Introduction

The β-hydride elimination is a fundamental chemical process that appears in many transition-metal catalyzed transformations¹, such as the Heck reaction^2–5, Saegusa oxidation^6,7, and migratory cross-coupling reactions^8–11. Chemists have carried out several experimental and theoretical studies to investigate the chemoselectivity and regioselectivity of β-hydride elimination^12–24. However, there have been very few attempts to study the stereoselectivity because this elimination step transforms two sp³ carbons into an alkene without generating a common stereocenter. The development of methods to achieve enantioselective β-hydride elimination is an attractive research topic in asymmetric transition-metal catalysis^25–28. Frantz and coworkers reported the first successful asymmetric β-hydride elimination from vinyl Pd(II)-complexes derived from enol triflates. This reaction features an asymmetric desymmetrization strategy, producing 1,3-disubstituted allenes with high enantioselectivities that bear axial chirality²⁵. Zhang et al. disclosed a Heck reaction-triggered β-hydride elimination, and their Pd-Xuphos system could furnish 1,1,3-trisubstituted chiral allenes^26,27. More recently, Tan et al. reported an elegant asymmetric Heck reaction of anthracenylidenes, which utilized the asymmetric β-hydride elimination to establish an axially chiral anthracenylidene skeletons²⁸. Despite these achievements, the construction of central chiralities by asymmetric β-H elimination has never been addressed (Fig. 1a).

Fig. 1. Palladium-catalyzed enantioselective β-H elimination.

a Previous works on palladium-catalyzed enantioselective β-H elimination to construct axial chirality. b This work on palladium-catalyzed enantioselective β-H elimination to construct central chirality.

The β-H elimination from the π-allyl-palladium species in the Tsuji-Trost reaction can lead to the formation of undesired product 1,3-dienes^29–35. However, this reaction has not been fully explored for its synthetic potential, particularly in enantioselective variants. Our group has a long-standing interest in Pd-H involving asymmetric hydrofunctionalization of 1,3-dienes^36–43, which prompts us to consider the possibility of implying the reversed Pd-H elimination reaction to generate 1,3-dienes with a chiral center. Taking inspiration from the desymmetrization strategy^44–47 to prepare remote stereocenters^48–53, we proposed a desymmetric enantioselective β-H elimination from the π-allyl-Pd might yield 1,3-dienes with a remote central chirality (Fig. 1b). Specifically, 1-vinylcyclohexyl acetate, derived from 4-substituted cyclohexanone, was designed to react with LPd(0) to give π-allyl-Pd. The latter species was then isomerized to the C1-σ-Pd complex to occur β-H elimination. We believed that a proper chiral ligand could help differentiate the two adjacent cis-hydrogen atoms, making it possible to forge cyclohexenes bearing a C4-remoted stereocenter. Notably, more than three hundred natural products and bioactive molecules, including (-)-perillaldehyde, (-)-oleuropeic acid, β-bisabolenal and epitodomatuic acid, contain such structural units (statistics are based on the Reaxys database). However, only rare catalytic methods are available to access these chiral motifs. These elegant works include Liu’s cobalt-catalyzed olefin isomerization reaction^54,55 and Jacobsen’s asymmetric Wittig reaction⁵⁶.

Herein, we disclose a Trost ligand-enabled enantioselective desymmetric β-H elimination process from π-allyl-Pd. This transformation provides rapid access to cyclohexenes bearing a C4-remoted stereocenter from 1-vinylcyclohexyl acetate. Application of this method to the total synthesis of (-)-oleuropeic acid and (-)-7-hydroxyterpineol is demonstrated.

Results and discussion

Reaction development

We commenced our study by using 1a as the model substrate, Pd₂(dba)₃·CHCl₃ (5 mol%) as the catalyst precursor, and THF as the reaction solvent. Various chiral ligands (12 mol%) were first investigated (Table 1). The desired reaction was very sensitive to the ligand’s structure (entries 1-11). Although (R,Sp)-Josiphos (L3), (R,R)-Ph-BPE (L4), and (S)-DTBM-MeO-Biphep (L9) showed very poor activity for this reaction, (S)-Phosphinooxazoline (L1), (R,Rp)-Ferrophox (L2), (S,S,S,S)-BIBOP (L5), (S,S)-QuinoxP (L6), Xuphos (L7), and (R)-MeO-Biphep (L8) were competent ligands and produced the product 2a in 37–72% yields. However, the enantioselectivity was quite low (less than 41% ee). When (S,S)-DACH-naphthyl Trost ligand (L10) was employed, we were pleased to find 2a was obtained in a 73% yield with 75% ee (entry 10). By switching to (R,R)-DACH-phenyl Trost ligand (L11), both the yield and ee could be significantly improved (96% yield, 89% ee, entry 11). We found that the catalyst loading could be reduced to 0.75 mol% Pd₂(dba)₃·CHCl₃ and 1.8 mol% L11 (entry 12). While the yield was maintained, the ee was slightly decreased. Then, using L11 as the optimal ligand, different solvents were screened. Changing the solvent to toluene, dichloromethane, or tert-butyl methyl ether resulted in dropped ee values (entries 13–15). However, when 1,4-dioxane was used, the desired product 2a was obtained quantitatively (90% isolated yield) with 90% ee (entry 16). Although one molecular AcOH was produced during the reaction, adding extra bases, such as K₂CO₃, ⁱPr₂NEt, and DBU, proved to be detrimental to either the ee value or the yield (comparing entry 16 with entries 17–19).

Table 1.

Optimization of reaction conditions.^a

Entry	L	Solvent	Base	Yield (%)b	ee (%)c
1	L1	THF	w/o	72	-32
2	L2	THF	w/o	46	41
3	L3	THF	w/o	<5	nd
4	L4	THF	w/o	8	-27
5	L5	THF	w/o	59	-6
6	L6	THF	w/o	37	-14
7	L7	THF	w/o	66	8
8	L8	THF	w/o	46	35
9	L9	THF	w/o	<5	nd
10	L10	THF	w/o	73	-75
11	L11	THF	w/o	96	89
12d	L11	THF	w/o	95	82
13d	L11	PhCH3	w/o	97	66
14d	L11	CH2Cl2	w/o	76	80
15d	L11	MTBE	w/o	98	78
16d	L11	1,4-dioxane	w/o	99 (90e)	90
17d	L11	1,4-dioxane	K2CO3	85	91
18d	L11	1,4-dioxane	iPr2NEt	82	89
19d	L11	1,4-dioxane	DBU	97	80

THF tetrahydrofuran, MTBE methyl tert-butyl ether, w/o without, DBU 1,8-diazabicyclo [5.4.0]undec-7-ene, nd not determined.

^aReaction conditions, unless otherwise noted: 1a (0.10 mmol), Pd₂(dba)₃·CHCl₃ (5 mol%), ligand (12 mol%), THF (0.5 mL), 50 °C, 12 h.

^bYields were determined by ¹H NMR analysis of crude reaction mixtures.

^cEnantiomeric excess (ee) values were determined by HPLC analysis on a chiral stationary phase.

^d1a (0.20 mmol), Pd₂(dba)₃·CHCl₃ (0.75 mol%), L11 (1.8 mol%), solvent (0.5 mL), 50 °C, 12 h.

^eIsolated yield was given in the parentheses.

Substrate scope

Having the optimal conditions in hand, we began to explore the substrate generality of this reaction (Fig. 2). The 4-aryl substituted substrates were first examined. Phenyl groups with electron-donating (Me, MeO) or electron-withdrawing (Cl, F, CN, CF₃) substituents were well tolerated, and the desired products (2a–2g) were obtained in 81-94% yields with 88–93% ee. The yield of the electron-withdrawing substituted phenyl group is slightly higher than that of electron-donating. We observed no significant differences in reactivity and enantioselectivity when comparing para-, meta-, and ortho-substitutions (2b, 2h vs. 2j and 2i vs. 2g). In addition to phenyl groups, substrates with other aromatic rings such as naphthalene and quinolone were also competent for this elimination reaction (2k–2m). Next, we turned our attention to the 4-alkyl substituted substrates. Aliphatic groups such as isopropyl (2n), tert-butyl (2o), and polar groups such as dimethyl malonate (2p, 2q) and Meldrum’s acid moiety (2r) all showed good tolerance to the reaction. When the R group was 2,3-dihydrobenzofuran, 2s was produced in 96% yield and 1:1 dr ratio, and 84% ee and 91% ee were obtained respectively for the two diastereomers.

Fig. 2. Scope of the reaction with respect to the vinylarene.

Reaction conditions: 1 (0.20 mmol), Pd₂(dba)₃·CHCl₃ (0.75 mol%), L11 (1.8 mol%), 1,4-dioxane (0.5 mL), 50 °C, 12 h. Enantiomeric excess (ee) values were determined by HPLC on a chiral stationary phase. ^aWhen substrates with cis-1-vinyl-4-aryl relative configuration were used in this investigation, moderate enantioselectivities were given. For example, cis-1-vinyl-4-phenyl substrate 1a affords 2a in 62% yield with −55% ee under the standard conditions. ^bYields were determined by ¹H NMR analysis of crude reaction mixtures. ^cThe X-ray crystallographic data have been deposited at the Cambridge Crystallographic Data Center (CCDC) under deposition number CCDC 2306164 (2d) can be obtained free of charge from www.ccdc.cam.ac.uk/data_request/cif.

We next explored the possibility of constructing a C4-heteroatom substituted chiral center. Substrates with 4-oxygen or nitrogen substitution were therefore evaluated. Products with 4-O-benzyl (2t) and 4-O-benzoyl (2u) had slightly reduced ee values (72 and 82%, respectively), while products with C4-N-phthalimide (2v), C4-N-benzoxazolinone (2w), and C4-N-benzotriazole (2x) had good to excellent enantioselectivities. The l-aspartic acid condensed amine substrate was also compatible, yielding 2y in 98% yield and >20:1 dr. This elimination reaction was also reliable in establishing quaternary stereocenters. Both heterosubstituted (3a, 3b) and all-carbon quaternary stereocenters (3c, 3d) were efficiently constructed with 86–94% ee. Furthermore, C4-spirobicyclic products were also achievable, as demonstrated in the formation of 3e and 3f.

Scaled-up reaction and application

The scalability of this reaction was demonstrated in a gram-scaled preparation of 2a with a decreased catalyst loading (0.12 mol% Pd₂(dba)₃·CHCl₃ and 0.29 mol% L11). While the yield dropped slightly, the enantioselectivity was well maintained (Fig. 3a). The diene moiety in the resulting elimination products could undergo diversified synthetic transformations. Using 2a as an example (Fig. 3b), the terminal alkene occurred the dihydroxylation with AD-mix-α, giving 4a in 75% yield with >20:1 dr. Further cleaving the diol by oxidation with NaIO₄ gave aldehyde 4b in 72% yield with the constant ee. The diene was extended to unsaturated ester 4c by reaction with ethyl acrylate under Grubbs II catalyst. Moreover, regioselective hydroboration of 2a with HBpin was achieved by [Ir(cod)Cl]₂ catalyst, affording 4d in 90% yield⁵⁷. Considering the abundant conversion chemistry of boranes, 4d might serve as a useful synthon in many other transformations. Finally, the Diels-Alder reaction of 2a with N-benzylmaleimide was achieved by heating at 80 °C in toluene, forging diastereomers 4e in nearly quantitative yield with 1.3:1 dr. Furthermore, the application of this method to total synthesis of (-)-oleuropeic acid and (-)-7-hydroxyterpineol was demonstrated (Fig. 4). Starting from substrate 5, the enantioselective elimination reaction was conducted with 0.75 mol% Pd₂(dba)₃·CHCl₃ and 1.8 mol% ent-L11, affording (S)-6 in 75% yield with 90% ee. The terminal olefin moiety of (S)-6 was dihydroxylated with AD-mix-α reagent to give 7. Then removal of the acetate and cleavage of the dihydroxyl groups, giving aldehyde 9 in 67% combined yield. Finally, aldehyde 9 underwent the Pinnick oxidation or Luche reduction, furnishing (-)-oleuropeic acid and (-)-7-hydroxyterpineol, respectively.

Fig. 3. Scaled-up reaction and synthetic transformation of 2a.

a Scaled-up reaction under reduced catalyst loading. b Synthetic transformation of the product 2a.

Fig. 4. Total synthesis of (-)-oleuropeic acid and (-)-7-hydroxyterpineol.

Starting material 5 was converted to 6 in 75% yield with 90% ee under the standard conditions. The latter 6 was then transformed into a common intermediate 9 with three steps. Selective oxidation or reduction of the aldehyde moiety of 9 produced (-)-oleuropeic acid 10 and (-)-7-hydroxyterpineol 11, respectively.

Mechanistic Studies

The kinetic isotope effect (KIE) experiments were conducted to get insights into the reaction mechanism (Fig. 5). A comparison of the initial reaction rates of 1a and the deuterated d-1a revealed a KIE value of 3.5, indicating the potential involvement of β-hydride elimination in the rate-limiting step⁵⁸.

Fig. 5. Kinetic isotope effect experiments.

The reaction of 1a or d-1a with 0.375 mol% Pd₂(dba)₃·CHCl₃ and 0.9 mol% L11 in d8-THF at 32 °C was performed, and the initial reaction rate constants k_H and k_D were determined. The KIE values was calculated as k_H/k_D.

Subsequently, DFT calculations were performed to elucidate the detailed mechanism (Fig. 6). The reaction commences with the allylic substitution of the acetate moiety by L11-Pd(0), forming the η³-π-allyl-Pd intermediate Int C. The transition state for this step, TS B, was found to have an energy barrier of 14.3 kcal/mol. Int C then isomerizes to the η¹-allyl-Pd Int D, where the adjacent H atom acts as a σ-donor ligand to the Pd center. Subsequently, β-hydride elimination occurs via the transition state TS E, giving rise to the observed product (R)-2a and forming the palladium-hydride species Int F. Finally, Int F undergoes a rapid reductive elimination through a five-membered transition state TS G, regenerating the Pd(0) catalyst Int A along with one molecule of HOAc. The overall energy profile confirms that the rate-limiting step is the β-hydride elimination, consistent with the findings from the KIE experiments. We also calculated the pathway to generate the (S)-2a (indicated in red in Fig. 6). The corresponding transition state of β-hydride elimination for (S)-2a was located as TS E′, which is 5.2 kcal/mol disfavored comparing with TS E. These results suggest that (R)-2a would be the dominant product, consistent with the experimental results.

Fig. 6. DFT calculations.

Calculation method: M06(SMD)/SDD(Pd)/6-311 + G(d,p)//B3LYP(D3)/SDD(Pd)/6-311 G(d).

The IGMH analysis⁵⁹ revealed that attractive van der Waals interactions occur between the amide moiety of the ligand and the benzene and cyclohexane moieties of the substrate in the transition state TS E (Fig. 7). To investigate these weak interactions between the ligand and substrate, energy decomposition analysis (EDA) was performed (Fig. 8). The EDA analysis showed that there was more steric repulsion (ΔΔE_exchange = 0.95 kcal/mol) and less dispersion effects (ΔΔE_dispersion = 0.98 kcal/mol) between the ligand and substrate in TS E as compared to TS E′. However, the electrostatics (ΔΔE_{electrostatics} = −3.03 kcal/mol) and induction (ΔΔE_induction = −1.55 kcal/mol) compensated for the steric repulsion and resulted in additional non-covalent stabilization energy in TS E (ΔΔE_total = −2.64 kcal/mol).

Fig. 7. Origin of the stereoselectivity.

a Calculated transition states TS E and TS E′ for β-H elimination step that form (R)-2a and (S)-2a, respectively. b Non-covalent interaction analysis with IGMH. The acetate counter anion in TS E and TS E′ was omitted for clarity.

Fig. 8. Rationalization of the stereochemistry.

a Studies of the non-covalent interactions between the ligand and substrate by energy decomposition analysis. The EDA was performed with density-fitted sSAPT0 using the SAPT module (with jun-cc-pVDZ basis set) in the PSI4 package. b The ΔΔE_dispersion, ΔΔE_{electrostatics}, ΔΔE_induction, and ΔΔE_exchange between TS E and TS E′.

In summary, we have developed a palladium-catalyzed enantioselective desymmetric β-hydride elimination from π-allyl-Pd for the construction of remote stereocenters. This transformation features low catalyst loading, base-free conditions, and wide functional group tolerance. A series of 1,3-dienes bearing a C4-chiral center were synthesized efficiently from 1-vinylcyclohexyl acetates. Synthetic application of this method to the total synthesis of (-)-oleuropeic acid and (-)-7-hydroxyterpineol was demonstrated. DFT calculations indicated that the β-hydride elimination is both a rate-limiting and enantioselectivity-determining step. The non-covalent interactions between the chiral ligand and substrate play a key role in stereocontrol.

Methods

General procedure for Pd catalyzed β-H elimination

In an Ar-filled glove box, (R,R)-DACH-phenyl Trost ligand L11 (2.5 mg, 0.0036 mmol, 1.8 mol%) and Pd₂dba₃ ∙ CHCl₃ (1.6 mg, 0.0015 mmol, 0.75 mol%) were dissolved in dry 1,4-dioxane (0.4 M, 0.5 mL) to stir at room temperature for 1 h. To the solution, substrate (0.2 mmol) was added. The resulting mixture was stirred at 50 °C for 12 h. The mixture was then purified by silica gel column chromatography to give the desired product.

Author contributions

W.Z. conceived the concept and directed the project, conducted the DFT calculations, and wrote the paper with feedback from all other authors. Shaozi Sun made the initial discovery, performed the optimization, and explored the substrate scope. Shaozi Sun and Shengnan Sun conducted the preparation of the starting materials.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Data availability

The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information Files (including experimental details, NMR data, HPLC and NMR spectrums) and Supplementary Data 1 (DFT calculated Cartesian coordinates), as well as from the corresponding author on request. The X-ray crystallographic coordinates for the structure reported in this study have been deposited at the Cambridge Crystallographic Data Centre (CCDC), under deposition numbers CCDC 2306164 for 2d. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre via www.ccdc. cam.ac.uk/data_request/cif.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-57437-x.