En Route to Enantioenriched Quaternary Stereocenters via Lewis Base/Palladium Cooperative Catalysis

Abstract

The prevalence of quaternary stereogenic centers in bioactive molecules coupled with innate challenges associated with their enantioselective preparation continues to provide powerful impetus for the development of catalytic asymmetric methods capable of their construction. Herein, we describe a cooperative isothiourea Lewis base-palladium catalyst system that enables the enantioselective alkylation of α-substituted-α-cyano esters with allyl methanesulfonate. While the levels of enantioselection are modest, this study represents the first time we have successfully constructed quaternary-substituted stereogenic centers using this Lewis base-palladium cooperative catalysis scheme. Further, this strategy constitutes a departure from ligand-based enantiocontrol and suggests that, when using acidic pro-nucleophiles, the development of protocols where Lewis base catalysis can outcompete direct deprotonation might be within reach.

Introduction

The enantioselective construction of quaternary stereogenic centers remains formidable but relevant due to the prevalence of such units in bioactive molecules.^[1] However, their formation is challenging both from the perspective of catalysis and stereoselectivity because the necessary forging of C(sp³)− C-(sp³) bonds must occur in sterically crowded environments.^[2,3] Of the methods available, palladium catalyzed allylic alkylation has emerged as one of the most versatile.^[4,5] Occurring principally at the electro-philic carbon this can be attributed to well established control over reaction regioselectivity and the strong influence of several ancillary ligand classes on enantioselectivity.^[6] Our interest in combining π(allyl)palladium electrophiles with C1-ammonium enolates under a synergistic Lewis base/palladium regime to address the issue of stereocontrol at the nucleophilic carbon presented an opportunity to interrogate the construction of quaternary stereogenic centers.^[7–22]

We have extensively exploited the rapid reaction of nucleophilic isothiourea Lewis base catalysts with pentafluorophenyl (Pfp) esters as a means to access the corresponding acylammonium species and thence C1-ammonium enolates, and we have capitalized on the latter being excellent nucleophiles for cationic π(allyl)metal electrophiles. Furthermore, we have been able to exploit the modular nature of such isothiourea Lewis base-transition metal cooperative catalysis to realize a general mechanistic scheme for enantioselective alkylation reactions where reactivity challenges and control over pertinent aspects of reaction regio- and diastereoselectivity can be directed by the chosen transition metal center/ancillary ligand assembly.^[8–17] Most recently both we and Zi have further extended this scheme to enable substrate activation via hydrido-palladium intermediates, which extends the range of substrates that participate in this cooperative platform

to cumulated alkenes and conjugated dienes.^[8–10] However, so far, this cooperative catalysis system enables the formation of tertiary stereogenic centers, with fully substituted centers so far being inaccessible (Scheme 1a). This requires comment with regards the scope of C1-ammonium enolate that is accessible via acyl ammonium ion enolization: (1) a C(sp²) substituent (aryl or alkenyl) is normally required in the α-position,^1[23] (2) the generation of stabilized C1-ammonium enolates from acidic precursors (malonate-type, where direct deprotonation might compete) has not been established, and (3) α,α-disubstituted C1-ammonium enolates are not accessible due to the presumed steric recalcitrance of α-branched acyl ammonium ions toward enolization.^2[24] We have be-gun to investigate these and herein provide proof-of-concept that (i) C(sp³)- and C(sp)-hybridized substituents are, in combination, tolerated in the α-position, (ii) stabilized C1-ammonium enolates can be accessed and do indeed engage in enantioselective bond construction, and (iii) that α,α-disubstituted C1-ammonium enolates can be accessed and enable the enantioselective construction of quaternary stereogenic centers (Scheme 1b).

Scheme 1.

(a) General depiction of breadth of Lewis base/ transition metal cooperative catalysis to forge tertiary stereo-centers. (b) This work: direct construction of quaternary stereo-centers via enantioselective isothiourea/palladium cooperative catalysis.

We envisioned a cooperative isothiourea Lewis base-Pd catalyzed process that unites α-cyano esters with allyl electrophiles and expected this would proceed via stereoselective reaction between nitrile-substituted (E)−O−C1-ammonium enolates and a cationic π(allyl)Pd electrophile to deliver versatile enantioenriched quaternary stereocenter-containing products (Scheme 1b). If successful, such a scenario would achieve, for the first time, the construction of quaternary stereogenic centers via this cooperative mechanistic scheme. The direct enantioselective allylation of α-cyano esters and amides via Rh/Pd cooperative catalysis has been described by Sawamura, Sudoh, and Ito in what is undoubtedly one of the seminal examples of enantioselective cooperative catalysis.^[25] Accordingly, our attempts to achieve the same via Lewis base/palladium catalysis represents a most stringent test. Clearly, the enhanced acidity of the racemic α-cyano esters leads to the reasonable probability of a competing achiral alkylation via the corresponding stabilized anion. The crux of this problem can be summarized thus: Can Lewis base catalysis compete with direct enolization?

Results and Discussion

To interrogate the enantioselective construction of quaternary stereogenic centers and the potential for Lewis base catalysis to compete with direct deprotonation, we prepared racemic pentafluorophenyl methylcyanoacetate (1) (see Supporting Information). The enantioselective alkylation of 1 with allyl methanesulfonate catalyzed by Pd(PTh₃)₃ in cooperation with various isothiourea Lewis base catalysts (LB) was then explored (Scheme 2).^[17,26] Beginning with the prototypical isothiourea, (R)-benzotetramisole 3, the expected allylation product 2 was obtained in 26 % yield and with only very minor levels of enantioenrichment (44 : 56 er, 12 %ee). Assessment of other common isothiourea catalysts (4–7) revealed Bn-(4) and Me-substituted (5) isothioureas as those capable of delivering meaningful levels of enantioenrichment (ca 23 –32 %ee). In the case of 5 increasing the steric profile of the substituent from Me (5) to iPr (6) and tBu (7) completely eroded enantioselection. Finally, an attempt to increase the levels of enantioselectivity by extending the Bn substituent in 4 to the corresponding naphthyl 8 and therefore more effectively shielding one face of the putative nitrile-conjugated C1-ammonium enolate was not successful. The same was true of lengthening the aromaticity of 4 to the 1-naphthyl congener 9.

Scheme 2.

Identification of optimal isothiourea Lewis base catalyst.

In line with our original hypothesis, we interpreted this (loosely) as indication that the modest levels of enantioselection obtained using 4 and 5 were most likely due to a competing achiral background alkylation reaction emanating from direct deprotonation of substrate 1. This was confirmed by control experiments where efficient alkylation giving racemic 2 occurs (i) in the presence of both Lewis base (4) and iPr₂NEt (1.0 equivalent), and (ii) in the presence of only iPr₂NEt (see Supporting Information for details). Despite the modest levels of enantioenrichment observed in 2, we were nonetheless impressed by the ability of the isothiourea Lewis base to form the putative C1-ammonium enolate and undergo enantioselective alkylation, thus competing with direct deprotonation/racemic alkylation of 1.

Due to having greater quantities available we elected to use (R)-Bn-BTM (Lewis base 4) for the remainder of the study. With optimized conditions identified and having established that Lewis base catalysis could compete somewhat with direct deprotonation and deliver enantioenriched alkylation products, we next moved to assess the scope of the nucleophile to ascertain if the levels of enantioselectivity in alkylation might be general across this substrate class (Scheme 3). Gratifyingly, racemic α-cyano Pfp-esters bearing methyl, ethyl, benzyl, and elaborated benzyl substituents (2, 10 – 14), each gave quaternary stereocenter-containing alkylation products with similar efficiencies and conserved levels of enantioselection. The reaction was also effective for preparing thiophene-conaining product 15.

Scheme 3.

Substrate scope.

The synthetic utility of the process is further enhanced by the exceptional electrophilicity of the Pfp ester.^[7–22] Alkylation product 14 could be readily modified by chemoselective transformation to diverse ester products (16 –20) in the presence of the nitrile function (Scheme 4).

Scheme 4.

Product elaboration.

Having assessed the scope and utility of this enantioselective reaction we next sought to establish the absolute stereochemistry of the major enantiomer. Recrystallization of a scalemic mixture of 15 (69 : 31 er) from a saturated solution of 1 %Et₂O in pentane by slow evaporation gave colorless crystals of the major enantiomer in 99 : 1 er (confirmed by HPLC comparison, see Scheme 5). Subsequent single-crystal X-ray analysis established the absolute stereochemistry as R when derived from the (R)-Bn-BTM Lewis base catalyst. In addition to resolving the remaining structural questions regarding the absolute stereochemistry of the quaternary stereogenic centers in the major enantiomers of 2, 10–12, this also enables us to propose a tentative model for stereoinduction via correlation with our previous investigations.^[20,22]

Scheme 5.

Recrystallization of a scalemic mixture of 15 from 1 % Et₂O in pentane gave 15 99 : 1 er. Single crystal x-ray analysis revealed the absolute stereochemistry of the major enantiomer as (R) – when derived from the (R)-Bn-BTM Lewis base catalyst 4.

The sense of enantioselectivity is consistent with our prior disclosures describing enantioselective allylation via Lewis base/Pd-cooperative catalysis.^[8,14–22] Accordingly, we propose a tentative stereoselectivity model whereby allylation occurs via the (E) O C1-ammonium enolate 21 (Scheme 6). Here, the oxyanion and nitrile moieties are anti-co-planar, and the ensemble is rigidified by a n_O→σ*_C−S interaction resulting in shielding of the Si face by the benzyl (Bn) substituent of the catalyst which directs alkylation to the Re face giving 22. Based on our control experiments we expect competing racemic allylation occurs via the putative malonate-type stabilized enolate 23.

Scheme 6.

Tentative stereoinduction model.

We further propose that this is operative within the plausible mechanistic scenario depicted in Scheme 7. This is based on our prior investigations where the racemic α-cyano Pfp ester reacts with the isothiourea Lewis base catalyst (NR₃) giving the corresponding (E)−O−C1-ammonium enolate via enolization of a presumed ammonium ion intermediate (where PfpO^— or NR₃ would act as an appropriate base). The C1-ammonium enolate then undergoes stereoselective alkylation via reaction with cationic π(allyl)Pd(II) (see Scheme 6 for stereoselectivity model). Turn over of the Pd catalyst would occur by alkene exchange in the usual way, whereas the Lewis base (NR₃) turnover would occur via phenolate (or phenol, not shown) rebound.^[22,27]

Scheme 7.

Plausible mechanism.

Conclusions

We have developed an enantioselective alkylation of nitrile-substituted α-branched esters with allyl mesylate that is catalyzed by a cooperative isothiourea and palladium system. This enables the preparation of quaternary stereogenic centers, expands the scope of C1-ammonium enolate that is accessible, and provides proof-of-concept that Lewis base catalysis can compete with direct enolization in the preparation of stabilized C1-ammonium enolates from acidic malonate-type substrates. The further utility of products is shown by the chemoselective modification of the Pfp esters in the presence of the nitrile moiety. We expect the results described herein will lead to an expansion of C1-ammonium enolates that serve as effective ester enolate equivalents for enantioselective carbon-carbon bond formation, and further contribute to the catalysis arsenal being developed for quaternary stereogenic center construction.

Experimental Section

General Information

Commercial reagents were purified prior to use. Unless otherwise noted, all reactions were carried out with distilled and degassed solvents under an atmosphere of dry N₂ in oven-dried glassware with standard vacuum-line techniques. All reactions were carried out in Teflon screw cap reaction vials with magnetic stirring unless otherwise indicated. Dichloromethane and tetrahydrofuran were purified under a positive pressure of dry argon by passage through two columns of activated alumina. Ethanol and 1,4-dioxane were purchased from Sigma-Aldrich and was dried over 3Å and 4Å activated molecular sieves respectively. All workup and purification procedures were carried out with reagent grade solvents (purchased from Sigma-Aldrich) in air. Standard column chromatography techniques using ZEOprep 60/40–63 μm silica gel were used for purification. Liquids and solutions were transferred via syringe or cannula.

¹H and ¹³C NMR spectra were recorded at room temperature: Varian I400 (¹H-NMR at 400 MHz and ¹³C NMR at 100 MHz), Bruker B500 (¹H-NMR at 500 MHz and ¹³C NMR at 125 MHz), Varian I500 (¹H-NMR at 500 MHz and ¹³C NMR at 125 MHz), and Varian I600 (¹H-NMR at 600 MHz and ¹³C NMR at 150 MHz) using deuterium lock. Data for ¹H-NMR spectra are quoted relative to chloroform, dichloromethane, or methanol as an internal standard (7.26, 5.32, or 3.31 ppm respectively) and data for ¹³C NMR spectra are quoted relative to chloroform, dichloromethane, or methanol as an internal standard (77.23, 53.84, or 49.00 ppm respectively) and are reported in terms of chemical shift (δ ppm). Data are reported as follows: chemical shift, multiplicity (s = singlet, d =doublet, t = triplet, q= quartet, br=broad, m= multiplet), coupling constants (Hz), and integration. Infrared spectra (IR) were obtained on a PerkinElmer Spectrum Two FTIR Spectrometer and Bruker TENSOR II FTIR Spectrometer and recorded in wavenumbers (cm⁻¹). Melting points were obtained on a Thomas Hoover capillary melting point apparatus without correction. High Resolution Mass (HRMS) analysis was obtained using Electrospray Ionization (ESI), Electron Impact Ionization (EI), or Atmospheric Pressure Chemical Ionization (APCI) and reported as m/z (relative intensity) for the [M] ⁺, [M + H]⁺, or [M + Na]⁺ molecular ion. Chiral HPLC analyses were performed on an Agilent 1200 Series system with specified column.

General Alkylation Procedure

The specified pentafluorophenyl ester (0.20 mmol, 1.0 equiv.), (R)-benzyl-benzotetramisole (10.7 mg, 0.04 mmol, 20 mol%.), and Pd(PTh₃)₃ (18.9 mg, 0.02 mmol, 10 mol%) were added to an oven-dried 8 mL vial containing a magnetic stir bar with a Teflon septa insert screw cap. The vial was evacuated and backfilled with N₂ (3 ×). Anhydrous 1,4-dioxane (2.0 mL, 0.1 M) and allyl methanesulfonate (22.7 μL, 0.20 mmol, 1.0 equiv.) were added to the vial sequentially via syringe. The reaction was stirred at r.t. for 16 h. The reaction was diluted with petroleum ether and filtered through activated, acidic Al₂O₃, rinsing with Et₂O. The filtrate was concentrated by rotary evaporation and the crude residue was further purified by column chromatography [SiO₂, specified eluent].

Analytical details for 11 (details for all other compounds are described in the supplementary information): Alkylation product 11 was prepared according to General Alkylation Procedure (above). An internal standard (1,2,4,5-tetramethylbenzene) was added to the crude product and the yield (35 %) calculated by ¹H-NMR. Compound 11 was obtained (17.9 mg, 0.047 mmol, 24 %) as a colorless oil following purification by column chromatography [SiO₂, 1 – 3 % Et₂O in pentane]. The enantiomeric ratio (68 : 32) was determined by chiral HPLC in comparison with the racemate (see below).

[α]_D²⁰ −12.2 (c = 1.0, CHCl₃). IR (film): 3082, 2929, 2246, 1785, 1645, 1518, 1343, 1165, 1080, 1030, 995, 977, 934, 701 cm−¹. ¹H-NMR (500 MHz, CDCl₃) δ = 7.42–7.32 (m, 5H), 5.94 (ddt, J = 17.4, 10.5, 7.3 Hz, 1H), 5.40–5.33 (m, 2H), 3.38 (d, J = 13.7 Hz, 1H), 3.24 (d, J = 13.8 Hz, 1H), 2.88 (dd, J = 13.8, 7.6 Hz, 1H), 2.76 (dd, J = 13.9, 7.0 Hz, 1H) ppm. ¹³C-NMR (126 MHz, CDCl₃) δ = 165.2, 141.9 (m), 141.2 (m), 139.9 (m), 139.0 (m), 137.0 (m), 133.0, 130.1, 129.4, 129.0, 128.5, 124.4 (m), 122.5, 117.3, 51.9, 42.8, 42.1 ppm. ¹⁹F-NMR (376 MHz, CDCl₃) δ=−151.07 (d, J= 17.4 Hz), −156.12 (t, J = 21.6 Hz), −161.40 (dd, J= 21.6, 17.5 Hz) ppm. HR-MS (EI, 15.0 V): m/z calcd for [M] ⁺ C₁₉H₁₂F₅NO₂: 381.0788, found: 381.0792. HPLC analysis using a chiral column (Chiralpak IA 3 μ column, 22 °C, 0.5 mL/min, 1 % isopropanol/hexanes, 210 nm, t_major = 15.351 min, t_minor = 14.267 min).

Supplementary Material

Supporting information for this article is available on the WWW under https://doi.org/10.1002/hlca.202400089

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.