Asymmetric Synthesis of Diverse Glycolic Acid Scaffolds via Dynamic Kinetic Resolution of α-Keto Esters

Abstract

The dynamic kinetic resolution of α-keto esters via asymmetric transfer hydrogenation has been developed as a technique for the highly stereoselective construction of structurally diverse β-substituted-α-hydroxy carboxylic acid derivatives. Through the development of a privileged m-terphenylsulfonamide for (arene)RuCl(monosulfonamide) complexes with a high affinity for selective α-keto ester reduction, excellent levels of chemo-, diastereo-, and enantiocontrol can be realized in the reduction of β-aryl- and β-chloro-α-keto esters.

1. INTRODUCTION

The enantioselective construction of α-hydroxy carboxylic acids remains an active area of research due to their prevalence in biologically active molecules¹ and use in asymmetric synthesis.² Reliable methods to prepare these compounds include enantioselective glycolate aldol and alkylation reactions,^3,4 reduction or alkylation of α-keto esters,⁵ Passerini-type reactions,⁶ asymmetric cyanohydrin synthesis,⁷ and ester enolate oxygenations.⁸ Despite advances in these methodologies, the preparation of β-stereogenic glycolic acid derivatives remains much more challenging highlighting the importance of a generalizable strategy to access these substructures.

The reduction of α-keto esters to give β-stereogenic-α-hydroxy esters has largely been limited to the diastereoselective reduction of enantioenriched substrates.⁹ A more direct, efficient reaction manifold might arise from the asymmetric catalyst-controlled reduction of configurationally labile racemic β-substituted-α-keto esters. Such a reaction could in principle proceed with concomitant formation of two (or more) stereogenic centers in a single step and provide access to a number of functionalized glycolic acid derivatives (Scheme 1).¹⁰ This strategy pre-supposes the application of a dynamic kinetic resolution (DKR), a powerful tool for the conversion of racemic materials into enantiomerically enriched products.¹¹ In light of the prominence and utility of DKR reactions of α-stereogenic-β-keto esters, the absence of complementary isomeric variants from racemic α-keto esters was surprising. As part of our laboratory’s continued interest in glycolic acid synthesis,¹² we have recently developed a highly stereoselective dynamic kinetic resolution of β-stereogenic-α-keto esters via asymmetric transfer hydrogenation (DKR-ATH), yielding trisubstituted γ-butyrolactones (vide infra).¹³ It occurred to us that substantial product diversity might arise from a common mechanistic platform simply by varying the identities of the nonhydrogen substituents (X and Y) at the β-carbon. The successful creation of an attractive synthetic protocol would require: (1) simple routes to the needed racemic α-keto ester substrates; (2) reaction conditions that achieve rapid substrate racemization; (3) the identification of a reduction catalyst that is enantiomer-selective, provides strong facial bias during the diastereoselective reduction, and can be applied to functionally diverse substrates. The subject of this paper is the evaluation of this strategy and the presentation of a new (arene)Ru catalyst system for the asymmetric dynamic reduction of a range of racemic β-stereogenic-α-keto esters. The chemistry to be detailed was enabled by a seemingly trivial, yet ultimately crucial deviation from established art in asymmetric Ru-catalyzed transfer hydrogenation.

Scheme 1.

β-Stereogenic glycolic acid derivatives via reduction

2. RESULTS AND DISCUSSION

2.1. Ligand/Catalyst Design

Inspired by the efficacy of Noyori’s (arene)RuCl(monosulfonamide-DPEN)¹⁴ in both the asymmetric reduction of simple ketones¹⁵ and the dynamic reduction of α-substituted β-keto esters and amides,¹⁶ we took this complex as our point of departure in ligand/ catalyst design. Utilizing formic acid:triethylamine (5:2 mixture)¹⁷ as the organic reductant and 1a as a test substrate, a screen of ligands and precatalysts was undertaken (Table 1). Initial studies looked at the steric effects of the sulfonamide in Ru(II)-complexes possessing a (1S,2S)-diphenylethylenediamine (DPEN) backbone. Subjecting 1a to 2 mol % of the ruthenium dimer [RuCl₂(arene)]₂ and Noyori’s ligand L1 (Ru atom: L mole ratio 1:2) in DMF at 75 °C¹⁸ provided the desired γ-butyrolactone in high yield (90%) and diastereoselectivity (>20:1 dr), but with low levels of enantiocontrol (57:43 er, entry 1). DPEN-based ligands featuring bulkier sulfonamides (L2–L5) provided only modestly higher levels of selectivity (entries 2–5). Employing L2, a screen of (arene)Ru(II)-precatalysts was conducted to determine the role of the arene in the stereoselectivity of the reduction; however, no improvements were observed moving away from [RuCl₂(p-cymene)]₂ (entries 6 and 7). In addition to DPEN, 1,2-diaminocyclohexane and 1,2-aminoindanol were also investigated as chiral backbones (L6 and L7), but yielded comparable results (entries 8 and 9).

Table 1.

Evaluation of chiral diamine ligands

entry	L	arene	er
1	L1	p-cymene	57:43
2	L2	p-cymene	70:30
3	L3	p-cymene	62:38
4	L4	p-cymene	70:30
5	L5	p-cymene	72:28
6	L2	C6Me6	62:38
7	L2	benzene	71:29
8	L6	p-cymene	73:27
9	L7	p-cymene	60:40

^aConditions: 1a (1.0 equiv), [RuCl₂(arene)]₂ (0.02 equiv), L (0.08 equiv), HCO₂H:NEt₃ (5.0 equiv), [1a]₀ = 0.1 M in DMF, 75 °C, 16 h.

Based on these preliminary findings that asymmetric transfer hydrogenation catalysts from this family present in the literature were found to provide inadequate levels of selectivity, it became clear that new chiral space would need to be explored in order achieve high levels of enantiocontrol. Utilizing the “mother diamine”/diaza-Cope approach to the synthesis of C₂-symmetric 1,2-diamines,¹⁹ screening of a number of chiral diamine backbones was conducted. The α-naphthyl/triisopropylbenzenesulfonamide ligand L8 considerably increased the selectivity (88:12 er, Table 2). To further optimize the ligand structure, perturbations of the sulfonamide were examined due to its apparent ability to directly impact the chiral environment (Table 1, L1 vs. L2). Since less sterically encumbering sulfonamides (L9–L12) resulted in erosions in selectivity, we sought to synthesize bulkier sulfonamides by exploring new chiral space at the 2,6-positions of the arylsulfonamide.²⁰ A number of diverse sulfonyl chlorides were synthesized through a one-pot double alkylation/sulfonylation of 1,3-dichlorobenzene.²¹ The simplest m-terphenyl sulfonamide variant L14, distinguished itself as being uniquely effective for providing high levels of enantioselectivity for the title reaction (95:5 er). The use of electron-withdrawing (L15) or releasing substituents (L16) provided no improvement in selectivity. The α-naphthyl backbone and m-terphenylsulfonamide operate synergistically; no improvement in enantiocontrol with the DPEN/m-terphenylsulfonamide ligand L5 was observed (Table 1, entry 5). The α-naphthyl ethylenediamine backbone has been used sporadically in asymmetric synthesis and the use of the m-terphenylsulfonamide for enantioselective catalysis is rarer still.²²

Table 2.

Evaluation of sulfonamides on α-naphthyl backbone

^aConditions: As described for Table 1.

2.2. Synthesis of Trisubstituted γ-Butyrolactones

This DKR-ATH was found to be applicable for a range of β-aryl α-keto esters.¹³ Electron-rich, electron-poor, and heteroaryl substituents were tolerated at the β-position providing γ-butyrolactone products in high yield and enantioselectivity. ²³ Additionally, the reduction of 1g was performed on a 10 g scale employing reduced catalyst loading (1 mol % Ru) yielding enantiopure lactone 2g in 72% yield following a single recrystallization. The absolute stereochemistry of the trisubstituted γ-butyrolactone products was determined by X-ray crystallographic analysis of 2b.²⁴

The obtained functionally rich γ-butyrolactones can be deployed in secondary transformations (Scheme 2). Diastereoselective alkylation of 2a employing allyl bromide and DBU provided tetrasubstituted lactone 3a bearing an all-carbon quaternary carbon center in high yield. Krapcho decarboxylation²⁵ gave access to α-unsubstituted lactone 3b (86% yield) that formally arises from cinnamic acid. When decarboxylation was preceded by alkylation with dibromomethane, dehalodecarboxylation²⁶ resulted and afforded α-alkylidene γ-butyrolactone 3c. This substructure is prominent in natural product chemistry and bioactivities within this subclass are well documented.²⁷

Scheme 2.

Reactions of lactone reduction product 2a

2.3. “Free” Glycolate Michael Adducts Using Chemoselective DKR-ATH of α,δ-Diketo Esters

2.3.1 DKR-ATH Approach to “Free” Glycolate Michael Adducts

Access to δ-oxygenated glycolic acid derivatives via asymmetric glycolate Michael reactions is limited. The most common approach to this class of compounds is the addition of stoichiometric chiral glycolate enolates to α,β-unsaturated ketones and esters,²⁸ and in each case the protected glycolate is obtained (Scheme 3a). Only recently has a catalytic enantioselective variant been disclosed; that method uses oxazolones as the α-hydroxy acid surrogate (Scheme 3b).²⁹ To the best of our knowledge, no catalytic enantioselective Michael addition of a free glycolate enolate has been reported.³⁰

Scheme 3.

Approaches to δ-oxygenated glycolic acid derivatives

Based on the success of the DKR-ATH of γ,γ-dicarboalkoxy-α-keto esters (1, vide supra), we postulated that β-substituted-δ-keto-α-hydroxy esters might be accessible via a chemoselective dynamic reduction of α,δ-diketo esters directly delivering the formal “free” glycolate Michael adducts. Implicit in this analysis is the need for complete site selectivity in the reduction. Methods for the selective reduction of an aldehyde in the presence of less reactive carbonyls, i.e. ketones and esters, are well-established. ³¹ While significant progress has been made in achieving the inverse process, the selective reduction of a ketone in the presence of an aldehyde,³² the discrimination between two ketones remains a challenging task. Examples of the latter include the chemoselective reduction of 2,4-diketo acid derivatives using cinchona-modified platinum catalysts³³ or baker’s yeast³⁴ and aluminium-mediated selective reductions of diaryl ketones.³⁵ Our tactic takes advantage of the heightened reactivity enjoyed by α-dicarbonyls and establishes a simple catalytic method for achieving the formal asymmetric glycolate Michael construction without recourse to auxiliary control or protection of the hydroxyl group (Scheme 3c).

2.3.2 Glyoxylate Stetter Addition with Enones

Indirect methods for the preparation of the requisite α,δ-diketo esters 5 have been reported by us^10b and others,³⁶ but the most direct and atom economical approach to these substrates would be a new Stetter reaction between commercial ethyl glyoxylate and α,β-enones.³⁷ This reaction was previously reported to be unsuccessful with thiazolium carbenes,^36c but we have found that effective catalysis can be realized by employing the Rovis triazolium carbene derived from salt 4.³⁸ As outlined in Table 4, this glyoxylate Stetter addition is highly efficient, tolerant of a number of ketonic substrates and substituents at the β-position and can be performed on a multigram scale (entry 1). Notably, with the 1,4-dien-3-one dibenzyli-deneacetone, exclusive monoaddition was observed (entry 5).

Table 4.

Scope of the glyoxylate Stetter addition with enones^a

entry	R1	R2	5	yield (%)b
1c	C6H5	C6H5	5a	97
2	C6H5	4-I-C6H4	5b	95
3	C6H5	4-MeO-C6H5	5c	86
4	C6H5	Me	5d	72
5	C6H5	(E)–CH=CHPh	5e	91
6	C6H5	piperonyl	5f	92
7	4-Cl-C6H4	C6H5	5g	94
8	4-MeO-C6H4	C6H5	5h	96
9	2-Me-C6H4	C6H5	5i	70
10	3-Me-C6H4	C6H5	5j	93
11	4-Me-C6H4	C6H5	5k	95
12	CO2Et	C6H5	5l	92
13	piperonyl	C6H5	5m	94
14	N-Boc-Indol-3-yl	C6H5	5n	74

^aConditions: Unless otherwise noted, all reactions were performed on a 2.0 mmol scale in PhCH₃ (4 mL) at ambient temperature.

^bIsolated yield.

^cReaction performed on a 20 mmol scale.

2.3.3 Chemoselective DKR-ATH of α,δ-Diketo Esters

Our investigation into the chemoselective DKR-ATH began with an examination of the reduction of α,δ-diketo ester 5a. As shown in Table 5, our initial studies confirmed the feasibility of selectively reducing the α-keto ester in the presence of an aryl ketone, as we observed exclusive formation of δ-keto-α-hydroxy ester 6a as a single diastereomer. Subjecting 5a to our previously optimized reaction conditions, 2 mol % of [RuCl₂(p-cymene)]₂ and diamine ligand L14 (Ru atom:L mole ratio 1:2) in DMF at 70 °C using HCO₂H/Et₃N as the reductant, provided 6a in 96% yield and a 91:9 er (entry 1). Lower levels of selectivity were observed when the reduction was conducted at room temperature (87:13 er, entry 2). Further optimization revealed that high levels of enantioselectivity (97:3 er) could be obtained by performing the reaction in DMSO at room temperature (entry 6). The chemoselectivity observed is remarkable in light of the extensive use of (arene)RuCl(sulfonamide) complexes for asymmetric transfer hydrogenation of aryl ketones.^14,15,17

Table 5.

Chemoselective DKR-ATH: reaction optimization ^a

entry	solvent	T (°C)	yield (%)a	erc
1	DMF	70	94	91:9
2	DMF	rt	93	87:13
3	2-MeTHF	70	90	65:35
4	DCE	70	91	78:22
5	DMSO	70	96	97:3
6	DMSO	rt	98	97:3

^aConditions: 5a (1.0 equiv), [RuCl₂(p-cymene)]₂ (0.02 equiv), L14 (0.08 equiv), HCO₂H:NEt₃ (5.0 equiv), [5a]₀ = 0.1 M, 16 h.

^bIsolated yield.

^cEnantiomeric ratio determined by SFC analysis.

With optimal reaction conditions in hand, we next explored other substrates in the chemoselective dynamic reduction (Table 6). For all substrates examined, exclusive reduction of the α-keto ester was observed, irrespective of the electronic characteristics of the δ-ketone. High yields and enantioselectivities were obtained for substrates incorporating electron-rich, electron-poor, and heteroaryl substituents at the β-position. The level of selectivity for 6l (entry 12, 91:9 er) is noteworthy as this demonstrates that the scope is not limited to β-aryl substrates (vide infra). ³⁹ Additionally, other reducible functional groups remained intact: the retention of the α,β-enone in 6e under the reaction conditions further highlights the catalyst’s strong preference for the α-keto ester (entry 5). Aryl halides were also tolerated (entries 2,7). To evaluate the catalytic efficiency of this system, the reduction of 5a was performed using 0.05 mol % of [RuCl₂(p-cymene)]₂; no loss in reaction efficiency was observed as 6a was obtained in 98:2 er (entry 1).

Table 6.

Chemoselective dynamic reduction scope^a

entry	R1	R2	6	yield (%)b	erc
1d	C6H5	C6H5	6a	97	98:2
2	C6H5	4-I-C6H4	6b	95	96:4
3	C6H5	4-MeO-C6H5	6c	86	96:4
4	C6H5	Me	6d	72	93:7
5	C6H5	(E)–CH=CHPh	6e	91	99:1
6	C6H5	piperonyl	6f	92	99:1
7	4-Cl-C6H4	C6H5	6g	94	98:2
8	4-MeO-C6H4	C6H5	6h	96	96:4
9	2-Me-C6H4	C6H5	6i	70	83:17
10	3-Me-C6H4	C6H5	6j	93	98:2
11	4-Me-C6H4	C6H5	6k	95	98:2
12	CO2Et	C6H5	6l	92	91:9
13	piperonyl	C6H5	6m	94	97:3
14	N-Boc-Indol-3-yl	C6H5	6n	95	98:2

^aConditions: as described for Table 5.

^bIsolated yield.

^cEnantiomeric ratio determined by SFC analysis.

^dReaction performed using 0.05 mol % [RuCl₂(p-cymene)]₂.

The absolute configuration and syn-stereochemical relationship of the α-hydroxy ester products were determined by converting 6c to lactone 3b via a Baeyer-Villiger oxidation followed by in situ lactonization; spectral data and optical rotation were in agreement with those previously obtained by us for 3b.¹³

To determine if the [RuCl₂(p-cymene)]₂/L14 catalyst system was uniquely effective for the reduction of α-keto esters, the chemoselectivity of transfer hydrogenation catalysts known to reduce simple ketones was evaluated with 5a (Scheme 5). The use of ligand L2 in the reduction, which is known to reduce acetophenone,⁴⁰ also afforded 6a as the sole product, albeit with lower levels of enantioselectivity (69:31 er). This result caused us to wonder if the δ-ketone was possibly undergoing in situ “protection” as the lactol 7 following reduction of the α-keto ester and that this intermediate masked the δ-ketone from further reduction. To test this hypothesis, the reduction was monitored by ¹H NMR spectroscopy in DMSO-d⁶, but 7 was not detected: only the diketo ester 5a and hydroxy ester 6a were observed (Figure 1).

Scheme 5.

Investigations into the observed chemoselectivity

Figure 1.

In situ monitoring of the reduction of (±)-5a by 1H NMR spectroscopy in DMSO-d⁶. Blue panel: α,δ-diketo ester 5a; Red panel: in situ monitoring of reduction of α,δ-diketo ester 5a to 6a (~50% conversion); Green panel: α-hydroxy-δ-keto ester 6a (with added HCO₂H/Et₃N).

We then examined the transfer hydrogenation of several other ketone substrates using the standard reaction conditions used in Table 6 (Scheme 6). Interestingly, acetophenone and ethyl 2-ethyl-3-oxobutanoate, which are prototypical test substrates for new transfer hydrogenation catalysts, are not reduced with this catalyst system. This lack of reactivity further highlights the unique preference for α-keto esters conferred by the newly developed terphenylsulfonamide/di-α-napthylethylenediamine ligand. Reducing tert-butyl 2-oxo-4-phenylbutanoate under the standard conditions proceeded efficiently, imparting good levels of enantioselection and highlighting the potential applicability of this catalyst for simple α-keto esters.

Scheme 6.

Other keto ester reductions

2.4. Chlorohydrin Synthesis

2.4.1 DKR-ATH Approach to Optically Active Chlorohydrins

The investigations described above provide simple access to useful enantiomerically enriched glycolate building blocks, but Scheme 1 implies a goal of diversification in product structures that had not yet been realized. In considering new substrates that might be useful for DKR-ATH reactions, the potential integration of β-halo substituents was appealing on several levels. Optically active halohydrins are fundamental building blocks in organic chemistry and these functional arrays can be converted to their derived enantioenriched epoxides or engage in nucleophilic substitution to provide a variety of functionalized product classes (Scheme 7). The emergence of halohydrin dehalogenase (HheC), an enzyme produced by Agrobacterium radiobacter AD1, as a biocatalyst for the kinetic resolution of racemic haloalcohols highlights the importance of methods for the preparation of optically pure terminal halohydrins.⁴¹ The catalytic asymmetric preparation of halohydrins has been limited principally to desymmetrization reactions of epoxides⁴² and alkenes⁴³ or kinetic resolution of terminal epoxides.^{42d,42g,42l,42m,44} Methodology designed to directly access internal halohydrins from unsymmetrical precursors is largely underdeveloped (vide infra). We sought to develop a stereoselective Ru-catalyzed dynamic kinetic resolution of β-chloro-α-keto esters that would provide an efficient route to β-chloro-α-hydroxy esters.

Scheme 7.

Secondary transformations of chlorohydrin products.

Access to chiral β-chloro glycolic acid derivatives is currently limited to enzymatic processes or stereospecific opening of glycidic esters with strong acids. While enzymatic reductions⁴⁵ and kinetic resolutions⁴⁶ have been shown to impart good levels of diastereo-and enantioselectivity, these processes are substrate limited and lack generality. Chloride addition to optically pure glycidic esters often necessitates harsh reaction conditions, suffers from non-ideal regio- and stereoselectivity, and lacks significant precedent for aliphatic substrates.⁴⁷ The dynamic kinetic resolution of α-chloro-β-keto esters has been developed for some time,⁴⁸ but the dynamic kinetic resolution of β-chloro-α-keto esters that would provide isomeric products is heretofore unknown. Thus, we sought to develop a simple, flexible synthesis of β-chloro glycolic acid derivatives employing a dynamic kinetic resolution asymmetric transfer hydrogenation (DKR-ATH) of β-chloro-α-keto esters (Figure 2).

Figure 2.

Dynamic kinetic resolution of β-chloro-α-keto esters.

2.4.2 Direct Catalytic β-Chlorination of α-Keto Esters

The relative dearth of direct, catalytic β-functionalizations of α-keto esters presented an obstacle to the implementation of this synthetic plan; in particular, methods for the direct β-halogenation of α-keto esters are scarce.^{9k,45b,45c,49} Only two examples of the direct β-chlorination of singly activated α-keto esters have been reported and those require either long reaction times or harsh reaction conditions.^45b,49a In order to synthesize the requisite β-chloro substrates for the anticipated DKR-ATH, the development of a mild chlorination reaction of α-keto esters was pursued. A screen of various Cu(II)-diamine and Ni(II)-diamine complexes led to the identification of Sodeoka’s Ni(OAc)₂-diamine complex 8⁹ⁱ as an effective catalyst for the β-chlorination of α-keto esters 9 using N-chlorosuccinimide (NCS) under mild reaction conditions.⁵⁰ As outlined in Table 7, the chlorination proceeds with good selectivity for the singly halogenated product, is tolerant of a variety of functionalized aliphatic substrates, and can be performed on a multigram scale (entry 1). The elevated acidity of β-aryl substrates favored dichlorination (entry 9); therefore, the requisite β-aryl substrates were prepared via Darzens reaction.⁵¹

Table 7.

Ni(II)-catalyzed β-chlorination of α-keto esters.^a

entry	R	10	mono:dib	yield (%)c
1d	–CH2Ph	10a	9:1	81
2	–CH2p-ClPh	10b	10:1	83
3	–CH2p-MeOPh	10c	12:1	84
4	–CH2CH2Ph	10d	8:1	78
5	–CH2CH=CH2	10e	13:1	86
6	–CH2C≡CTMS	10f	10:1	79
7	–(CH2)2CH3	10g	13:1	87
8	–(CH2)3OBn	10h	13:1	86
9	–Ph	10i	1:>20	trace

^aUnless otherwise noted, all reactions were performed on a 1.0 mmol scale.

^bDetermined by ¹H NMR analysis of the crude reaction mixture.

^cIsolated yield.

^dPerformed on 8.0 mmol scale.

2.4.3 DKR-ATH of β-Chloro-α-Keto Esters

The reduction of β-chloro-α-keto esters with NaBH₄ proceeds with high levels of diastereoselectivity to afford the syn-diastereomer via Felkin-Ahn control.^46a Initial investigations into the DKR-ATH of 10a revealed that ethylenediamine-derived 11a also afforded excellent levels of syn-selectivity (Figure 3); however, the diastereochemical outcome was powerfully influenced by ligand selection. Upon switching to Noyori’s parent Ru(II)-complex possessing a (1S,2S)-diphenylethylenediamine (DPEN) backbone (11b),¹⁴ a significant erosion in syn-diastereoselection was observed, albeit with promising levels of enantioselectivity. The bulkier triisopropyl-DPEN ligand (11c) gave modest anti-diastereoselection with improved enantiocontrol. Exploiting the unique properties associated with the terphenylsulfonamide (vide supra), the DPEN-derivative 11d led to appreciable ligand-controlled diastereoselection providing anti-12a with excellent levels of enantioselectivity. The α-naphthyl backbone (11e) employed in the reduction of β-aryl-α-keto esters was found to provide slightly higher levels of diastereoselection albeit with a small loss in enantioselectivity. The diastereoselectivity and enantioselectivity were further improved in DMF at 0 °C employing only 1 mol % of the Ru catalyst (in this case, the conveniently prepared and stored dehydrohalgenated variant of 11d). Considering diastereoselectivity only, the continuum expressed by Figure 3 (>20:1 syn:anti → 1:9 syn:anti at 298 K) represents approximately 2.5 kcal/mol modulation of diastereomeric transition states through simple substituent modifications on a common ligand framework.

Figure 3.

Ligand-controlled switch in diastereoselectivity.^a

^aReactions were performed on 0.155 mmol scale employing 5 equiv. HCO₂H:NEt₃ (5:2). ^bDetermined by ¹H NMR analysis of the crude reaction mixture. ^cDetermined by chiral SFC analysis. ^dPerformed with 1 mol % catalyst in DMF (0. 1 M) at 0 °C for 10 h. Complex 11da is the dehydrohalogenated variant of 11d; the structure is illustrated in Table 8.

With optimized reaction conditions in hand the relationship between α-keto ester structure and reaction stereoselectivity was assayed (Table 8). A variety of aliphatic substrates were found to be amenable to the reaction conditions providing β-chloro α-hydroxy esters with excellent levels of diastereo- and enantioselection. Alkene, alkyne, and benzyloxy functionality was tolerated under the reaction conditions offering value-added functional handles. The method was also scalable (entry 1). The efficiency of these aliphatic substrates under the DKR-ATH reaction conditions is a marked structural departure from the β-aryl and β-ester requirements in antecedent work from our group and the paradigm that aryl groups are necessary for high levels of enantiocontrol due to ligand/substrate π/C–H interactions.⁵²

Table 8.

β-aliphatic substrates in the DKR-ATH.^a

entry	R	12	yield (%)b	drc	erd
1e	–CH2Ph	12a	89	12:1	99:1
2	–CH2p-ClPh	12b	90	16:1	99.5:0.5
3	–CH2p-MeOPh	12c	93	20:1	98.5:1.5
4	–CH2CH2Ph	12d	93	16:1	98.5:1.5
5	–CH2CH=CH2	12e	95	>20:1	98:2f
6g	–CH2C≡CTMS	12f	91	>20:1	96.5:3.5f
7	–(CH2)2CH3	12g	94	>20:1	98.5:1.5f
8	–(CH2)3OBn	12h	91	18:1	98:2

^aUnless otherwise noted, all reactions were performed on 0.155 mmol scale employing 5 equiv. HCO₂H:NEt₃ (5:2).

^bIsolated yield of anti-diastereomer.

^cDetermined by ¹H NMR analysis of the crude reaction mixture.

^dDetermined by chiral SFC/HPLC analysis.

^ePerformed on 6.5 mmol scale.

^fDetermined following conversion to the benzoate (see Supporting Information).

^gPerformed at 23 °C for 10 h.

Compatibility with β-aryl substrates was also demonstrated under the optimized reaction conditions, providing adducts with excellent levels of enantioselectivity (Table 9). The electronic character of the aromatic ring was found to significantly impact the diastereoselectivity of the reaction. Electron-releasing groups engendered excellent diastereocontrol (entries 2, 3, and 10) whereas electron-withdrawing groups provided somewhat lower diastereoselection (entries 4, 7, and 8).

Table 9.

Scope of β-aryl substrates in the DKR-ATH.^a

entry	Ar	13	yield (%)b	drc	erd
1	Ph	13a	93	14:1	99.5:0.5
2	o-MeOPh	13b	95	>20:1	99:1
3	m-MeOPh	13c	94	19:1	99.5:0.5
4	m-NO2Ph	13d	74	5:1	97.5:2.5
5	p-ClPh	13e	85	10:1	98.5:1.5
6	p-CF3Ph	13f	80	8:1	98.5:1.5
7	p-CNPh	13g	82	6:1	98:2
8	p-NO2Ph	13h	73	4:1	99:1
9	p-MePh	13i	91	14:1	99.5:0.5
10	p-MeOPh	13j	93	>20:1	99:1

^aReactions were performed on 0.155 mmol scale employing 5 equiv. HCO₂H:NEt₃ (5:2).

^bIsolated yield of anti-diastereomer.

^cDetermined by ¹H NMR analysis of the crude reaction mixture.

^dDetermined by chiral SFC/HPLC analysis.

The absolute stereochemistry of the products was determined by comparison to the known epoxide (2R,3S)-14,⁵³ which was synthesized from chlor0hydrin (2S,3R)-13a upon exposure to KO^tBu (Scheme 7). To highlight the synthetic utility of the enantioenriched chlorohydrins as synthetic building blocks, illustrative secondary transformations were pursued. Treatment of chlorohydrin 13a with NaN₃ afforded the azido alcohol 15 representing a formal synthesis of the paclitaxel C13 side-chain.^47b Notably, the syn-product 15 is stereocomplementary to the anti diastereomer obtained from the glycidic esters that one might derive from Darzens or Weitz-Scheffer reactions. Following triflate formation of chloroalcohol 12a, chemoselective displacement with NaN₃ affords α-azido-β-chloro ester 16 providing access to β-chloro-α-amino acid derivatives.

The DKR-ATH is also amenable to the reduction of β-fluoro-α-keto esters. Preliminary investigations have revealed that ketone 17 can be reduced under the optimized reaction conditions to afford the derived fluorohydrin 18 in excellent yield as a mixture of diastereomers (Scheme 8). Despite the lack of diastereocontrol in the reaction,⁵⁴ excellent levels of enantioinduction were observed for both diastereomers. This initial finding is quite encouraging in light of the responsiveness of diastereocontrol to ligand structure in this reaction family (vide supra).

Scheme 8.

DKR-ATH of β-fluoro-α-keto ester 17.

3. CONCLUSIONS

In summary, we have designed new (arene)Ru(monosulfonamide) asymmetric transfer hydrogenation catalysts that have led to the successful development of a highly modular dynamic kinetic resolution of β-substituted-α-keto esters. The productive merger of a common mechanistic framework and a new m-terphenylsulfonamide-based catalyst system allow for rapid, atom-economical construction of highly functionalized glycolic acid derivatives with excellent levels of chemo-, diastereo-, and enantioselectivity.

A formal asymmetric glycolate Michael reaction has been established via a catalytic asymmetric chemoselective dynamic reduction of α,δ-diketo esters. The latter are prepared via a new atom economical carbene-catalyzed Stetter reaction between commercial ethyl glyoxylate and α,β-enones. The enantioselective reduction proceeds with high enantio- and diastereoselectivity for a number of substrates. Initial investigations into the origin of the observed selectivity suggests that the [RuCl₂(p-cymene)]₂/L14 catalyst system is uniquely effective for the reduction of α-keto esters, as other ketone substrates (even acetophenone) are unreactive under the standard reaction conditions.

Additionally, a highly stereoselective synthesis of β-chloro glycolic acid derivatives via asymmetric transfer hydrogenation was developed. A Ni(II)-catalyzed chlorination of aliphatic α-keto esters was developed to provide the requisite β-chloro-α-keto esters. In the reduction of these ketones, careful catalyst tuning allows for a remarkable ligand-controlled inversion of the preference for syn-selectivity to provide access to anti-chlorohydrins. The DKR-ATH proceeds with high levels of diastereo- and enantioselectivity for a number of aliphatic and aromatic substrates. The obtained chlorohydrins are versatile chemical building blocks for valuable secondary transformations.

These studies collectively provide diverse glycolate-based building blocks for synthesis. This study has highlighted some of the preparative and practical aspects of these reactions, but open questions with respect to mechanism clearly remain (Scheme 9). Additional studies to understand the catalyst-substrate interactions that account for the high levels of selectivity observed are ongoing, with the goal of utilizing of this information in extensions to the dynamic kinetic resolution of other useful frameworks.

Scheme 9.

Open mechanistic questions.

Scheme 4.

Stereochemical analysis of δ-keto-α-hydroxy esters

Table 3.

DKR-ATH substrate scope^a

entry	R2	2	yield (%)b	erc
1	4-Cl-C6H4	2b	94	96:4
2	4-Me-C6H4	2c	84	95.5:4.5
3	4-MeO-C6H5	2d	90	95:5
4	4-CN-C6H5	2e	88	95:5
5	2-Me-C6H5	2f	82	89:11
6d	piperonyl	2g	72	95:5
7	2-furyl	2h	91	95:5
8	N-Ts-indol-3-yl	2i	91	96.5:3.5
9	N-Boc-indol-3-yl	2j	88	96:4

^aConditions: As described for Table 1.

^bIsolated yield.

^cEnantiomeric ratio determined by HPLC/SFC analysis.

^d[RuCl₂(p-cymene)]₂ (0.5 mol %), L14 (2 mol %), 10 g scale, >99.5:0.5 recrystallized.