|Home | About | Journals | Submit | Contact Us | Français|
The catalytic enantioselective synthesis of densely functionalised organic molecules containing all-carbon quaternary stereocentres is a challenge to modern chemical methodology research. The catalytically controlled asymmetric α-alkylation of ketones represents another difficult task and has been of major interest to our and other research groups in the past. We now report a palladium-catalyzed enantioselective process that addresses both problems at once and allows the installation of vicinal all-carbon quaternary and tertiary stereocentres at the α-carbon of a ketone in a single step. This multiple bond forming process is carried out on readily available β-ketoester starting materials and proceeds via conjugate addition of an in situ-generated palladium enolate to activated Michael acceptors. In other words, the CO2-moiety of the substrate is displaced by a C-C fragment in an asymmetric cut-and-paste reaction. The products are obtained in high yield, diastereomeric ratio, and enantiomeric excess.
The enantioselective generation of carbon stereocentres having four other carbon substituents (also called quaternary or all-carbon quaternary stereocentres) is a task synthetic chemists are confronted with regularly1. Several methods have been developed to address this problem but catalytic approaches remain challenging2,3. Furthermore, when additional carbon stereocentres adjacent to the quaternary carbon are required, the available choices to install both within one reaction are extremely limited. While the combination of quaternary and secondary stereocentres can be introduced by aldol-, cycloaddition-, Michael- or even cross-coupling reactions, only a handful of catalytic methods exist that forge adjacent quaternary and tertiary carbon stereocentres. This motif was successfully introduced into organic molecules by α-functionalisation of carbonyl groups as demonstrated in recent reports. However, these approaches mostly rely on organic catalysts based on quinine or proline4,5,6,7 and only two enantioselective transition metal catalyzed examples have been published to date8,9. Interestingly, in both cases strongly coordinating substrates such as 1,3-diketones, β-ketoesters or α-cyanoesters are required to achieve good selectivity. Other examples, like the elegant work of Overman, demonstrated the feasibility of palladium catalysis to construct adjacent quaternary and tertiary or even vicinal quaternary stereocentres in a single double Heck-reaction10,11. These diastereoselective reactions employ an achiral catalyst and rely on substrate-directed stereocontrol. Palladium-catalyzed tandem reactions like this have found application in intramolecular diastereoselective coupling-reactions12. Intermolecular examples especially involving π-allyl-palladium species are less common, but some recent examples achieving high diastereoselectivities have been published13,14,15. Therefore, an approach to the formation of vicinal quaternary and tertiary stereocentres that employs a chiral palladium catalyst is appealing and one could imagine that an asymmetric reaction involving the formation of multiple bonds could achieve this goal.
The enantioselective α-functionalisation of ketones has been realised by means of palladium catalysis and over the past decade significant progress has been made in α-allylation16, vinylation17, and arylation18 reactions. A transition metal catalyzed α-alkylation of ketones has been reported by Jacobsen employing chromium-salen complexes as catalysts to yield quaternary stereocentres19,20. The requirement of tin-enolates as starting materials, which have to be synthesised in advance and are presumably toxic, is a strong limitation in this case.
Over the past few years, our group has been interested in novel methodologies for the generation of all-carbon quaternary stereocentres, which led to the development of an enantioselective palladium catalysed allylic α-alkylation of ketones. This process allows the construction of such quaternary carbon stereocentres starting from silylenolethers, enolcarbonates21 or readily available racemic β-ketoesters 1 (path a, Figure 1)22. We also demonstrated that the intermediate of this reaction, presumably a palladium enolate (A)23, can be intercepted by a strong electrophile like H+ to generate enantioenriched α-protonated products (path b)24,25.
To expand this enolate-trapping approach toward the broader goal of establishing vicinal stereocentres and to provide a general approach for α-functionalisation of ketones, we contemplated the use of other prochiral electrophiles (path c). Therefore, an electrophilic additive had to be identified meeting the following crucial requirements: (1) no interference with the oxidative addition of the palladium(0) catalyst into the allylester functionality should occur; (2) the reaction of the palladium-enolate with the electrophilic additive must be preferred to the intramolecular allylic α-allylation; (3) a scavenger for the allyl fragment must be provided. In our first experiments carbon electrophiles like acyclic or cyclic enones were probed for their capability to intercept the intermediate enolate but in these cases only the regular allylic alkylation product was observed. Toward that end, we were drawn to earlier reports by Yamamoto of intercepted enolate alkylation by highly electrophilic conjugate acceptors such as benzylidenemalononitrile 2a (Table 1)15,26,27. Herein, we describe the development of an enantioselective palladium-catalysed decarboxylative alkylation of ketones that forges all-carbon quaternary stereocentres with an adjacent tertiary stereocentre in good yield and stereoselectivity. This reaction represents a rare example for an asymmetric palladium catalyzed tandem process.
In our initial experiments, we found that exposure of β-ketoester (±)-1a28 to catalytic amounts of [Pd2(dba)3] (dba = dibenzylideneacetone) and (S)-t-BuPHOX (L1, PHOX = phosphinooxazoline)29 in the presence of electrophile 2a indeed followed the desired reaction pathway, efficiently producing diastereomers 3a and 4a (82% conversion), albeit with moderate diastereomeric ratio and enantiomeric excess (1:3.1 dr, 42% and 63% ee; see supporting information for details). Following careful optimisation, the combination of an electronically modified PHOX ligand (L2, 12.5 mol%)30 and [Pd2(dba)3] (5 mol%) in 1,4-dioxane at 40 °C was found to produce diastereomers in high enantioselectivity. In the presence of only one equivalent of electrophile 2a, the diastereomeric products 3a and 4a were isolated in 99% yield, 1:6.1 dr, 77% ee and 87% ee, respectively (Table 1, entry 1).
We then applied these conditions to a series of substrates with different substitution at the α-carbon, the allyl functionality or the backbone of the cyclic ketone. The enantiomeric excess improved significantly when the α-substituent was changed to an ethyl group and the desired product was isolated with 99% ee (entry 2). Other α-substituents such as benzyl groups, esters or protected alcohols were tolerated as well (entries 4–7) but in some cases, allylic alkylation of the enolate was observed as a side reaction, leading to decreased yields (entries 4, 6, and 7). This side reaction was suppressed by lowering the reaction temperature to 23 °C (entry 5). Moreover, the reaction tolerates substitution on the allyl fragment as well (entry 8). In addition to cyclohexanones, 1,4-piperidinone derived substrates also furnished the corresponding products in high yield and good enantioselectivity with remarkably high diastereomeric ratio (entries 9 and 10).
Next, the scope of the electrophile was explored. The reaction performed well with arylidenemalononitriles derived from m- and p-anisaldehyde, allowing the introduction of substituted aryl groups at the stereocentre β to the ketone (Table 2, entries 1 and 2). Methyl or dimethylamino substituents were introduced at the 4-position, as well (entry 3 and 4). Although the diastereo-and enantioselectivity were impressive in the latter case, the yield was again low due to significant allylic enolate alkylation. Here, a change of the palladium precursor to Pd(dmdba)2 (dmdba = 3,3',5,5'-dimethoxybenzylideneacetone) slightly increased the reaction outcome toward the conjugate addition product (entry 5). Importantly, the versatile 2-furanyl substituent was installed with high yield and enantioselectivity as well (entries 6 and 7). Finally, a benzodioxole and a styrene group were successfully introduced at the tertiary centre in good yield and selectivity (entries 8 and 9) with exclusive regioselectivity in the last case.
As an initial probe to study the mechanism, we determined the absolute stereochemistry of diastereomeric products 3b and 4b to be (R,R) and (R,S) by single crystal X-ray analysis (Figure 2a, see supporting information for details). Interestingly, the absolute configuration at the α-quaternary centre is inverted, compared to the products from our standard asymmetric allylic alkylation (cf. 3b and 4b with 5)21, but matches the cyclohexanone products obtained from our protonation procedure (e.g., 6)24.
Strong electron-withdrawing groups on the Michael acceptor are necessary to achieve the desired conjugate addition. Although it was possible to employ reagents that differ from malononitriles like α-cyanoacetate and Meldrum's acid derivatives, the yields and stereoselectivities dropped significantly (Figure 2b). While products 3,4q were isolated only in 15% yield, it should be stated that in this case a series of two quaternary and one tertiary neighboring stereocentres were formed by a single palladium-catalyzed transformation, albeit with 64% ee. Notably, the conversion was significantly increased with the achiral, but more electronrich dppe ligand (96%) and the same two predominant diastereomers were observed as with ligand L2. Alternatively, inital attempts to employ alkyl-substituted acceptors were not successful, and instead, transfer of the allyl group to the alkylidenemalononitrile proceeded to yield adduct 7 (Figure 2c).
In general, the diastereoselectivity is good for most products and some examples show remarkable results with only trace amounts of a second diastereomer formed. The amount of direct allylic alkylation, as observed in some cases, was lowered either by a temperature change or by employing a different palladium precursor and suggest a possible dependence on the coordination behaviour of the acceptor and the catalyst. Indeed, initial preliminary experimental results suggest a resting state that is different from the η1-allyl-palladium carboxylate observed for the enantioselective allylic alkylation of ketones and may involve coordination of the electrophile31. Detailed studies are currently underway and comment on the specific details of the process would be speculation at this stage. Although it is not possible to provide a detailed picture at this time, a plausible simplified catalytic cycle is outlined (Figure 2d). First, palladium(0) catalyst B reacts with substrate 1 under the loss of CO2 to form η1-allylpalladium enolate A. Next, conjugate addition to 2 occurs and results in stabilised intermediate C, consisting of a π-allyl-palladium cation and a deprotonated malononitrile. Reductive alkylation then yields the diastereomeric products 3,4 and regenerates catalyst B.
In conclusion, a highly enantio- and diastereoselective palladium-catalyzed α-alkylation process has been developed that proceeds via the trapping of an intermediary palladium-enolate species by conjugate addition to a prochiral activated Michael acceptor. The reaction allows the asymmetric construction of densely functionalised molecules possessing an all-carbon quaternary centre neighboring a tertiary centre from simple racemic β-ketoester starting materials. Currently, expansion of this methodology to include other substrates and applications in multistep synthesis are in progress. Finally, mechanistic studies are also underway and will be reported in due course.
A flame-dried 50 mL Schlenk-tube was charged with [Pd2(dba)3] (13.7 mg, 0.015 mmol, 5 mol%) and L2 (22.2 mg, 0.0375 mmol, 12.5 mol%) under argon atmosphere and 3 mL of freshly distilled 1,4-dioxane were added. After stirring for 30 min at 23 °C, 1 (0.3 mmol, 1.0 equiv) and 2 (0.3 mmol, 1.0 equiv) were added simultaneously. The resulting yellow-green solution was stirred at the reported temperature for the reported amount of time (see Table 1 and and2).2). The consumption of starting material was monitored by tlc (KMnO4-stain) or by analysis of a small nmr sample. The solvent was removed under reduced pressure and the diastereomeric ratio determined by crude 1H nmr. Isolation and separation of products 3 and 4 was achieved by flash chromatography in hexanes/ethyl acetate mixtures in the given combined yields (Table 1 and and2).2). The enantiomeric excess was determined by either hplc or sfc analysis of the purified product (see supporting information).
The authors wish to thank NIH-NIGMS (R01 GM 080269-01 and a postdoctoral fellowship to D.E.W.), the German Academic Exchange Service (DAAD, postdoctoral fellowship to J.S.), Abbott Laboratories, Amgen, Merck, Bristol-Myers Squibb, Boehringer Ingelheim, the Gordon and Betty Moore Foundation, and Caltech for financial support. We thank Sarah Reisman for helpful discussions. Lawrence Henling and Michael Day are gratefully acknowledged for X-ray crystallographic analysis. The Bruker KAPPA APEXII X-ray diffractometer was purchased via an NSF CRIF:MU award to the California Institute of Technology, CHE-0639094.
Author ContributionsJ.S. planned and carried out the described experimental work and wrote the manuscript. D.E.W. and S.C.V. took part in the initial reaction development and screening experiments. B.M.S. initiated and directed the project. All authors commented on the manuscript.
Supplementary information and chemical compound information accompany this paper at www.nature.com/naturechemistry. Reprints and permission information are available online at http://npg.nature.com/reprintsandpermissions/.
Jan Streuff, The Arnold and Mabel Beckman Laboratories of Chemical Synthesis, Division of Chemistry and Chemical Engineering, California Institute of Technology, MC 101-20, 1200 E. California Blvd., Pasadena, CA 91125 (USA), Fax: (1) 626-395-8436.
David E. White, The Arnold and Mabel Beckman Laboratories of Chemical Synthesis, Division of Chemistry and Chemical Engineering, California Institute of Technology, MC 101-20, 1200 E. California Blvd., Pasadena, CA 91125 (USA), Fax: (1) 626-395-8436.
Scott C. Virgil, Caltech Center for Catalysis and Chemical Synthesis, California Institute of Technology, 1200 E. California Blvd., Pasadena, CA 91125 (USA)
Brian M. Stoltz, The Arnold and Mabel Beckman Laboratories of Chemical Synthesis, Division of Chemistry and Chemical Engineering, California Institute of Technology, MC 101-20, 1200 E. California Blvd., Pasadena, CA 91125 (USA), Fax: (1) 626-395-8436.