Stretchable chiral pockets for palladium-catalyzed highly chemo- and enantioselective allenylation

Pyrazolones are a vital class of heterocycles possessing various biological properties and much attention is paid to the diversified synthesis of enantiopure pyrazolone derivatives. We describe here the development of diphenylphosphinoalkanoic acid based chiral bisphosphine ligands, which are successfully applied to the palladium-catalyzed asymmetric allenylation of racemic pyrazol-5-ones. The reaction affords C-allenylation products, optically active pyrazol-5-ones bearing an allene unit, in high chemo- and enantioselectivity, with DACH-ZYC-Phos-C1 as the best ligand. The synthetic potential of the C-allenylation products is demonstrated. Furthermore, the enantioselectivity observed with DACH-ZYC-Phos-C1 is rationalized by density functional theory studies.

In this work, inspired by conformationally flexible alkyl chain in the Feng's ligands, we develop the enantioselective allenylation of pyrazolones by fine tuning of the Trost ligands, which leads to the development of a class of stretchable chiral pocket (Fig. 1c).
Obviously, a higher enantioselectivity was desired and the success of conformationally flexible alkyl chain in the Feng's ligands [38][39][40][41] has caught our attention (Fig. 3a). It was reasoned that the rigid aryl linker in the Trost ligands may be replaced with a flexible alkyl linker for a class of stretchable chiral pockets seeking higher enantioselectivity. Thus, the ZYC-Phos ligands were designed and readily synthesized by amidation of diphenylphosphino alkanoic acids, which were prepared by nucleophilic substitution of ethyl chloroalkanoate or chloroacetic acid (Fig. 3b) for further optimization [42][43][44][45] .
A possible mechanism for the allenylation of pyrazolone is shown in Fig. 5a [48][49][50] . To better understand the superior enatioselectivity and the advantage of DACH-ZYC-Phos-C1 in the allenylation of pyrazolone, we have conducted X-ray single crystal diffraction studies: firstly, Pd(II)-DACH-ZYC-Phos-C1 complex (III) was prepared by the reaction of PdCl 2 and DACH-ZYC-Phos-C1 with excess amount of bases in toluene (Fig. 5b). Pd(II)-DACH-Phenyl Trost ligand complex (IV) was obtained from the reaction of Pd(OAc) 2 with DACH-Phenyl Trost ligand with excess amount of bases in THF 51 . Both complexes were then recrystalized from CHCl 3 /n-hexane to afford single crystals suitable for the X-ray diffraction study. It was obvious that the angle (P1-Pd1-P2) and the distance between P1 and P2 in III are greater than those in IV (113.21°vs 102.19°, 3.773 Å vs 3.489 Å). However, this complex III failed to catalyze the enantioselective allenylation under the standard conditions or with AcOH as the protic additive (Fig. 5d), indicating that the complex III is not really a catalytically active species but provides a stable coordination mode for the palladium catalysis. The X-ray crystal structures of III and IV were then taken as the starting geometries for all the following calculations involving these complexes by restoring the ligand's amide N-H.
DFT calculations were performed on the enantioselectivity determining C−C bond formation step of the reaction of benzyl buta-2,3-dienyl carbonates 1b with pyrazol-5-one 2a catalyzed by the DACH-ZYC-Phos-C1-ligated palladium catalyst (see computational methods in the Supplementary Information for details). The reported energies are Gibbs energies that incorporate the effect of the toluene solvent. Figure 6 shows the optimized structures and relative free energies of the competing transition states basing on the endomethylene-π-allyl palladium complexes 52 (see the Supplementary Information for the other less favorable transition structures with exo-methylene-π-allyl palladium complexes). These transition structures are denoted as TS_left_Si_a/b, TS_left_Re_a/b, TS_right_Si, and TS_right_Re, separately (left/right indicates that the side of methylene moiety referring to the π-allyl Pd unit). Among them, TS_left_Si_a with the Si-face attack of 2a anion, is found to be the most favorable one, which is consistent with the dominant formation of S-products observed in the experimental studies.
The ligand's right-side amide N-H is available for hydrogen bonding to either the carbonyl oxygen or the hydrazine nitrogen of the approaching 2a anion. The structures of TS_left_Si_a and TS_left_Si_b, leading to the formation of the allenylation product with S configuration, are stabilized by the hydrogen bonding forming between the ligand's N-H with the carbonyl oxygen and the hydrazine nitrogen, separately. In the structure of TS_left_-Si_a, the H···O distance is 1.81 Å, and the bond angle of N − H···O is 175.0°, which are consistent with the common hydrogen bonding parameters 53 . Based on the electron density at the bond critical point, the hydrogen binding energies (BE) were calculated to analyze the strength of the hydrogen bonding 54 . The bond energy of the hydrogen bonding in TS_left_Si_a was estimated to be about 8.1 kcal/mol. Although the influence of the π/π interactions between the aromatic rings of the ligand and N-Ph Table 1 The effect of temperature, concentration, ligands, and solvents on the asymmetric allenylation of pyrazolone 2a with allene 1a a . pyrazolones on the enantioselectivity could be excluded based on the results of (S)-3bu and (S)-3bv, we have also tried to analyze the non-covalent interactions (NCI) in TS_left_Si_a using the Multiwfn 55,56 and VMD programs 57 and no obvious π-π interactions are found so far from the analysis (see supplementary information for the NCI plot of TS_left_Si_a).
The hydrogen bonding in TS_left_Si_b features a longer H···N distance of 2.18 Å and a smaller N − H···O bond angle of 163.1°, suggesting a weaker interaction, which estimated to be 4.8 kcal/ mol. Similarly, hydrogen-bonding interactions also exist in the structure of TS_left_Re_a and TS_left_Re_b, which would provide the allenylation product with R configuration. TS_left_Re_a, is stabilized by the hydrogen bonding formed between the ligand's N-H and the carbonyl oxygen of 2a anion, which was estimated to be about 6.3 kcal/mol. Moreover, TS_left_Re_a suffers the unfavorable interaction between 2a anion and the methylene-π-allyl moiety, which have been clearly shown in the Newman projection along the forming C-C bond in Fig. 6. Both factors contribute to the lower stability of TS_left_Re_a as compared to TS_left_Si_a by 4.5 kcal/mol. The bond energy of the hydrogen bonding in TS_left_Re_b was estimated to be about 5.4 kcal/mol, which is around 2.7 kcal/mol weaker than the one in TS_left_Si_a (about 8.1 kcal/mol), thus, accounting for the preference of TS_left_Si_a over TS_left_Re_b by 2.3 kcal/mol. Furthermore, TS_right_Si is also stabilized by the hydrogen bonding formed between the ligand's N-H and the carbonyl oxygen of 2a anion. No corresponding TS is able to be located with hydrogen bonding formed between the ligand's N-H and the hydrazine nitrogen. The bond energy of the hydrogen bonding in TS_right_Si was estimated to be about 5.8 kcal/mol. What's more, the steric repulsions between 2a anion and the methyleneπ-allyl moiety destabilize TS_right_Si. Both factors lead to the less stability of TS_right_Si than TS_left_Si_a by 3.3 kcal/mol. In addition, there is no hydrogen bonding existing in TS_right_Re. And unfavorable steric interactions between the two phenyl groups of 2a anion and the DACH-ZYC-Phos-C1 ligand further destabilize TS_right_Re, making it the least stable transition state (10.0 kcal/mol higher in energy than TS_left_Si_a). The calculated enantioselectivity, basing on the energy difference between TS_left_Si_a and TS_left_Re_b of 2.3 kcal/mol, is 94%, which is a Reaction conditions: 1 (0.5 mmol), 2 (1.2 equiv), Pd 2 (dba) 3 •CHCl 3 (2.5 mol%), and DACH-ZYC-Phos-C1 (5.0 mol%) in toluene (25 mL) at 60°C unless otherwise noted. The ee of 3 was determined by chiral HPLC. b The reaction time was 2 h. c 2l (1.4 equiv) was used and the reaction time was 13 h. d The reaction time was 11 h. e ent-DACH-ZYC-Phos-C1 was used and the reaction time was 11 h. f Pd 2 (dba) 3 •CHCl 3 (3.0 mol%), and DACH-ZYC-Phos-C1 (6.0 mol%) at 80°C. g Recovery of 1a was 9% as determined by the 1 H NMR analysis of the crude product using mesitylene as the internal standard. Cbz benzyloxycarbonyl, TMS trimethylsilyl.
in perfect agreement with the experimental ee value of 93% (Table 2, entry 2). Similar calculations were then conducted to compare with DACH-Phenyl-Trost ligand. The optimized structures and relative free energies of the competing transition states of the enantioselectivity-determining C−C bond formation step, associated with the endo-methylene-π-allyl palladium complexes, are illustrated in Fig. 7a, which are denoted as TS'_left_Si_a/b, TS'_left_Re_a/b and TS'_right_Si, separately. It should be noted that the structure of TS'_right_Re has not been located successfully, despite all the efforts, probably due to the severe steric repulsions. Different from the DACH-ZYC-Phos-C1 participated reaction, all the transition structures with methylene moiety lying on the left side to the π-allyl Pd unit suffer the steric interactions causing by 2a anion with the phenyl group in phenyl-Trost ligand, which reduce their stability. Nevertheless, via avoiding the unfavorable hindrance caused by the nucleophile 2a anion with the phenyl group in phenyl-Trost ligand, TS'_right_Si becomes the most favorable transition state. The energy difference between the diastereomeric TS'_right_Si and TS'_left_Re_b is found to be 1.2 kcal/mol. The calculated enantioselectivity of 71% in favor of also the S enantiomer is in agreement with the experimental ee value of 71% (Fig. 7b).
In summary, we have developed a DPPAA-based ligand for the highly chemo-and enantioselective allenylation of pyrazol-5-ones with benzyl 2,3-dienyl carbonates (up to 97% ee). Many   synthetically useful functionalities were tolerated under the catalysis of Pd/DACH-ZYC-Phos-C1. A rationale has been provided based on the X-ray diffraction studies of Pd(II)-DACH-ZYC-Phos-C1 complex and DFT calculations. In addition, these types of stretchable chiral pockets may also provide flexiblity and show great potential applications in catalytic enantioselective allylation, allenylation, and other reactions. Such studies are being actively pursued in our laboratory with very promising results and will be reported in due courses.

Data availability
All data that support the findings of this study are available in the online version of this paper in the accompanying Supplementary Information (including experimental procedures, compound characterization data, and spectra). The X-ray crystallographic coordinates for structures of (S)-3ak, Pd (