A step-economic and one-pot access to chiral Cα-tetrasubstituted α-amino acid derivatives via a bicyclic imidazole-catalyzed direct enantioselective C-acylation†

Cα-Tetrasubstituted α-amino acids are ubiquitous and unique structural units in bioactive natural products and pharmaceutical compounds. The asymmetric synthesis of these molecules has attracted a lot of attention, but a more efficient method is still greatly desired. Here we describe the first sequential four-step acylation reaction for the efficient synthesis of chiral Cα-tetrasubstituted α-amino acid derivatives from simple N-acylated amino acids via an auto-tandem catalysis using a single nucleophilic catalyst. The synthetic efficiency is improved via a direct enantioselective C-acylation; the methodology affords the corresponding Cα-tetrasubstituted α-amino acid derivatives with excellent enantioselectivities (up to 99% ee). This step-economic, one-pot, and auto-tandem strategy provides facile access to important chiral building blocks, such as peptides, serines, and oxazolines, which are often used in medicinal and synthetic chemistry.


Introduction
C a -Tetrasubstituted a-amino acids are ubiquitous and unique structural units in many natural products and synthetic compounds, many of which exhibit signicant biological activities and physiological effects (Fig. 1). 1 Several feasible methodologies have been developed for the synthesis of C atetrasubstituted a-amino acids, ranging from the use of chiral auxiliaries to the presently reported catalytic models. 2 These include: (1) the enantioselective addition of alkyl, aryl, or even acyl precursors (Strecker reaction followed by hydrolysis) to ketimines, 3 and; (2) the asymmetric a-alkylation, arylation, or acylation of a-substituted amino acid derivatives. 4,5 However, due to the challenges imposed by C a -tetrasubstituted a-amino acids, efficient methods are still lacking, especially for the synthesis of a-acyl-substituted C a -tetrasubstituted a-amino acid derivatives. 5 The most commonly used approach relies on the Oacylation of azlactones followed by a Lewis base-catalyzed asymmetric O-to C-acyl transfer (well-known as Steglich rearrangement, red arrows in Scheme 1). Aer the initial work developed by Fu et al., 5a several improvements on the asymmetric Steglich rearrangement have been reported by many other research groups. 5b-k Our group has also developed an efficient bicyclic imidazole organocatalyst for this reaction, which can be easily synthesized from imidazole in only three steps. 5f However, the efficiency of this transformation is hindered by the prefabrication of unstable azlactones and relatively inert O-acylated azlactones. Such inefficiencies are present in many other transformations which lead to poor practical applicability. 5b, 6 In order to improve the synthetic efficiency, minimizing the number of steps of the synthetic sequence is important. This strategy, termed step-economy, has frequently been employed in advanced total syntheses by utilizing new reactions. 7 Herein, we would like to put forward a more general approach in which a two-step rearrangement process (AB + C / AB-C / C-AB) is replaced by a direct reaction (AB + C / C-AB) by altering chemoselectivity. As the rst example, we recently developed Fig. 1 Examples of C a -tetrasubstituted a-amino acid derivatives with biological activities and physiological effects. a bicyclic imidazole-catalyzed direct enantioselective C-acylation of benzofuran-2(3H)-ones and 2-oxindoles. 8 For the most part, the starting material is directly converted to the C-acylated product but not via the relatively inert O-acylated intermediate. As a result, the efficiency for the synthesis of the desired products is dramatically improved compared to the corresponding Black rearrangement. 9 Inspired by this, and in continuation of our interest in the construction of C a -tetrasubstituted a-amino acids, 10 we propose a direct enantioselective C-acylation of azlactones with the purpose of improving the synthetic efficiency for the preparation of 4-carboxyazlactones (pink arrow in Scheme 1). 11 To further improve the synthetic efficiency, simplifying the operations via a sequential process in one pot is an applicable method. This strategy, termed pot-economy, has been widely used for the construction of multiple bonds in one pot. 12 Generally, different catalysts are used for different reaction steps (tandem catalysis). Obviously, carrying out all the reaction steps with a single catalyst would be more benecial (auto-tandem catalysis). 13 This strategy, termed catalysteconomy, has oen been employed in two-step sequences but seldom in multi-step sequences. 13c During the initial research concerning the direct enantioselective C-acylation of azlactones, we found that the starting azlactones could be prepared via intramolecular O-acylation of in situ generated anhydrides (shown in the square brackets in Scheme 1); the produced 4carboxyazlactones could be readily derivatized by another acylation to produce various C a -tetrasubstituted a-amino acid derivatives. Therefore, by combining the aforementioned step-, pot-, and catalyst-economy strategies, we aimed to design a four-step acylation sequence for the efficient synthesis of chiral C a -tetrasubstituted a-amino acid derivatives and biologically active dipeptides from simple N-acylated amino acids. The reaction would proceed via auto-tandem catalysis using a single bicyclic imidazole nucleophilic catalyst (blue part in Scheme 1). To the best of our knowledge, this type of sequential reaction that achieves excellent enantioselectivity has not been reported previously.

Results and discussion
We initially focused on screening the reaction conditions for the rst three steps in the sequence in one pot. (4-Methoxybenzoyl) alanine 1a was selected as the starting material and benzylamine was used as the nucleophilic reagent to derivatize the produced 4-carboxyazlactone shown in the square brackets in Table 1. Over the past decade, we have developed a series of bicyclic imidazole nucleophilic catalysts which have been successfully applied in a number of different reactions by our group 5f, 8,9,14 and other groups. 15 We envisaged that the bicyclic imidazole would be able to catalyze all the steps in the acylation sequence. Several bicyclic imidazole catalysts were screened for the sequential acylation of 1a with benzyl chloroformate as the acylating agent and di(isopropyl)ethylamine (DIPEA) as a base to afford product 2a (Table 1, entries 1-4). Catalysts bearing an alkoxy group gave better ee than those bearing an acyloxy group, and catalyst OBn-DPI was found to give the best result. Replacing the acylating agent, benzyl chloroformate with allyl chloroformate or phenyl chloroformate, afforded the desired products in lower yield and ee (entries 5 and 6, respectively). Use of ethyl chloroformate only afforded a trace amount of the product (entry 7). The base triethylamine (TEA) was also applied in this reaction but no product was observed (entry 8). The effect Scheme 1 Our new approach to enantiopure C a -tetrasubstituted aamino acid derivatives by a four-step acylation sequence using autotandem catalysis. of solvents was studied and toluene provided the product with the highest ee and satisfactory conversion (entries 4,[9][10][11][12][13][14]. Aer preliminary screening, the catalyst OBn-DPI, acylating agent benzyl chloroformate, base DIPEA and solvent toluene were chosen as the optimal conditions for further study, affording the desired product in 78% yield and with 91% ee (entry 4). Next, the effect of reaction temperature was researched ( Table 2). At a lower reaction temperature of 0 C for 36 h, product 2a was obtained with same yield and better ee compared to that of 20 C (entries 1 and 2). When the reaction temperature was reduced to À20 C for 36 h, the yield of 2a was reduced to 68% while the enantioselectivity increased to 95% (entry 3). Extending the reaction time to 48 hours gave the product in an increased yield (76%, entry 4). So the reaction was conducted for a longer time (72 h) to give better yield when reaction temperature was further decreased (entries 5-7). Finally, À55 C was found to be the optimized temperature with 78% yield and 99% ee (entry 6). Then the equivalents of ClCOOBn were investigated and it was found that 3.0 equivalents of ClCOOBn afforded the best results (entries 6, 8 and 9). Therefore, the sequential reaction of substrate 1a with 3.0 equivalents of benzyl chloroformate, 4.0 equivalents of DIPEA and 20 mol% OBn-DPI at À55 C over 72 h was optimal for both reactivity and enantioselectivity, leading to the product 2a in 78% yield and 99% ee (entry 8).
Having established the optimal reaction conditions, we investigated the nucleophile in the last step of the sequential process, employing (4-methoxybenzoyl)alanine 1a as the initial substrate (Scheme 2). Firstly a number of different alkyl amines and a phenyl amine were tested as nucleophiles (2a-d). Amines with less steric hindrance showed higher yields (2a, b vs. 2c, d). Cholamine, bearing both amino and hydroxyl groups, was also used as the nucleophile. Due to the wide difference in the activity of the alcohols and amines for this reaction, only the product 2e was obtained, via attack of the amino group, with good yield and excellent enantioselectivity (98% ee). Secondly, excess methanol was employed as a nucleophile in this reaction to give the corresponding product 2f with similar results. In addition, a variety of amino acid esters gave their corresponding enantiomerically pure dipeptides, which will be of particular use in the elds of biology and medicine (2g-r). When using an enantiomerically pure amino acid ester which contained a chiral center, the corresponding dipeptide products bearing two stereocenters were obtained with high dr (>99 : 1, 2l-r).
Various N-protected a-amino acid substrates 1 were tested in the sequential process using methyl 3-aminopropanoate as the nucleophile (Scheme 3). Firstly, N-substituted amino acid substrates bearing a methoxy group at the para-, meta-, and ortho-positions of the phenyl ring were tested in the domino reaction (2j, 2s and 2t). All these products were obtained in good yields and with excellent enantioselectivities. The effect of different substituents at the para-position of the phenyl ring was studied (2u-x). Substrates bearing electron-donating groups such as Me or tBu, and electron-withdrawing groups such as F or Cl, all afforded the corresponding products in good yields and with excellent enantioselectivities. The presence of an electron-withdrawing group enhances the acidity of the substrate and its corresponding intermediates in the two Oacylation steps and the C-acylation step (see proposed mechanism below), thus reducing the reaction time of these substrates to 8 h, much shorter than the reaction time for substrates bearing electron-donating groups (72 h). A substrate without a substituent group on the phenyl ring also gave the corresponding product 2y in 80% yield and with 98% ee within 10 h. We then employed simple N-benzoyl amino acids to study the inuence of the R group on the stereocenter (2y-ae). The substrate bearing a phenyl substituent gave the corresponding product 2ae with only 57% ee, and the substrate bearing an isopropyl group only gave the O-acylated byproduct. All other tested substrates gave the corresponding products in good yields and with excellent enantioselectivities. It's worth mentioning that products 2ab and 2ac were both obtained with 99% ee.
Having explored the substrate scope of both nucleophiles and amino acids, we focused on elucidating the effect of the catalyst OBn-DPI in each step of the acylation sequence (Scheme 4). Firstly, benzoylalanine 1h was reacted with 1.0 equivalent of benzyl chloroformate in the presence or absence of catalyst OBn-DPI to test the effect of the catalyst in the rst two O-acylation steps (step I and step II). In the presence of the catalyst, 1h afforded the cyclized product 3 in 63% yield at À55 C in 10 min; however in the absence of catalyst, 1h only gave the desired product 3 in 35% yield. We next subjected 3 to reaction with benzyl chloroformate in the presence or absence of catalyst OBn-DPI to test effect of the catalyst in the third C-acylation step (step III). In the presence of the catalyst, 3 afforded C-acylated product 4 in 93% yield in 10 h, whereas in the absence of catalyst, 3 was unable to afford C-acylated product 4 giving O-acylated compound 5 in 26% yield. Compound 5 was then employed as a reactant at À55 C in the presence of the catalyst OBn-DPI; no rearrangement product 4 was observed aer 24 h. These experimental results strongly support that 3 undergoes a direct Cacylation to afford 4. Finally, 4 was reacted with an amino acid ester in order to test the effect of the catalyst in the last step (step IV). In the presence of the catalyst, 4 afforded nal product 2y in 85% yield at 20 C in 18 h, whereas 2y was only obtained in 46% yield in the absence of catalyst.
Based on these results, a mechanism for the four-step acylation sequence has been proposed (Scheme 5). First, the catalyst OBn-DPI and benzyl chloroformate generate active species A. In the presence of DIPEA, N-substituted amino acid substrates 1, such as 1h, attack A to generate the intermediate mixed anhydride B. Reaction of B and OBn-DPI gives intermediate C which generates azlactone 3 via an intramolecular cyclization. In step III, azlactone 3 attacks the active species A to form C-acylated azlactone 4 via an enantioselective direct C-acylation. This chiral compound 4 and catalyst OBn-DPI form active intermediate D, which is attacked by the amino acid ester to further generate valuable dipeptide 2y via N-acylation. Experimental studies suggest that OBn-DPI acts as a nucleophilic catalyst in all the steps of the acylation sequence with compound 4 being formed from direct C-acylation.
Small peptides play a signicant role in many biological, pharmaceutical, and environmental applications. 16 Therefore the stereoselective synthesis of such biomolecules is of great importance. Here, we describe the concise synthesis of small Scheme 3 Expanding the racemic amino acid substrates. All the yields are isolated yields. Reaction time for 2s-v is 72 h, reaction time for 2w-x is 8 h, reaction time for 2y-ad is 10 h.
Scheme 4 Experimental studies on the effect of catalyst OBn-DPI and direct C-acylation. Yields were calculated from 1 H NMR spectra.
optically active peptides from the sequentially acylated product 2n (Scheme 6). Firstly, C a -tetrasubstituted a-amino acid 6 could be obtained from hydrogenation of 2n under 3 atm hydrogen pressure with Pd(OH) 2 /C in quantitative yield. Then, coupling of 6 with several different amino acids or dipeptides afforded a variety of valuable small chiral peptides in good yields (7a-g).
In particular, coupling of 6 with enantiomerically pure amino acids or dipeptides which contained a chiral center afforded products bearing three stereocenters with high dr (7b-d, 7f-g). Furthermore, the utility of this acylation sequence was demonstrated via the enantiodivergent synthesis of a-methyl serine from 2f by manipulating the different reactivities of the two ester groups (Scheme 7). Hydrogenation of the benzyl ester group of 2f with Pd(OH) 2 /C in MeOH at 20 C for 12 h provided 8 in quantitative yield. Reduction of 8 with LiBH 4 in THF at ambient temperature for 12 h afforded the (R)-isomer of N-protected a-methyl serine, (R)-10, in 83% yield. Hydrolysis of the methyl ester group of 2f with KOH in THF/H 2 O at 20 C provided 9 in 78% yield, which was reduced with LiBH 4 to give the (S)isomer of N-protected a-methyl serine, (S)-10, in 85% yield.
In addition, we realized the synthesis of 4,4 0 -disubstituted oxazoline from 2a, as shown in Scheme 8. Reduction of 2a with LiBH 4 in THF provided the corresponding alcohol 11. Mesylation of 11 with methanesulfonyl chloride in DCM at 0 C in the presence of triethylamine, followed by intramolecular cyclization at 20 C in one pot, nally afforded oxazoline 12 in 94% yield (74% overall from 2a). This method allows for the practical synthesis of chiral oxazolines, which can be used as chiral ligands for asymmetric catalysis. 17 Experimental Under a N 2 atmosphere, the substrate 1h (0.2 mmol, 44.7 mg), the catalyst OBn-DPI (0.04 mmol, 8.6 mg) and DIPEA (0.8 mmol, 132.2 mL) were dissolved in anhydrous toluene (2 mL) and cooled to À55 C in a dry two-necked ask. ClCOOallyl (0.6 mmol, 84.5 mL) was then added and the vial was sealed with a septum. The reaction mixture was stirred at À55 C for 10 h and then the temperature was gradually raised to 20 C. Methyl 3-aminopropionate hydrochloride (0.3 mmol, 41.9 mg) and DIPEA (0.3 mmol, 49.6 mL) were dissolved in anhydrous toluene (1 mL) in another dry ask and stirred for 10 min. Then the mixture was transferred into the former reaction ask and the reaction mixture was stirred at 20 C for 36 h. The reaction mixture was quenched with 0.2 M HCl (5 mL) and extracted with DCM (5 mL Â 3). The combined organic phases were dried over Na 2 SO 4 . Aer ltration, the residue was puried by column Scheme 5 Proposed mechanism for the four-step acylation sequence via auto-tandem catalysis (the anion is omitted in A, B, C, and D for clarity).
Scheme 6 Synthesis of small peptide. All the yields are isolated yields. 1 H NMR spectroscopy of the crude reaction products was used to assess diastereoisomeric ratios (dr).
Scheme 7 Enantiodivergent synthesis of a-methyl serine. All the yields are isolated yields.
Scheme 8 Synthesis of 4,4 0 -disubstituted oxazoline. All the yields are isolated yields. chromatography (petroleum ether/ethyl acetate) to give the corresponding product 2y. The ee value was determined by chiral HPLC analysis aer purication by column chromatography (petroleum ether/ethyl acetate).

Conclusions
In summary, the rst four-step sequential acylation reaction for the concise asymmetric synthesis of C a -tetrasubstituted aamino acid derivatives via auto-tandem catalysis has been successfully developed. This step-economic, one-pot, and autotandem strategy is promoted by a direct enantioselective Cacylation. Through four acylations catalyzed by a single chiral bicyclic imidazole, the simple N-acylated amino acids could be smoothly converted to the corresponding C a -tetrasubstituted aamino acid derivatives with excellent enantioselectivities (up to 99% ee). Signicantly, the obtained products, particularly dipeptides, are potentially biologically active compounds. These products can be further transformed to other biomolecules and important chiral building blocks such as small peptides, asubstituted serines and 4,4 0 -disubstituted oxazolines.

Conflicts of interest
The authors declare that they have no conict of interest.