Enantioselective synthesis of 2-oxazolidinones by ruthenium(ii)–NHC-catalysed asymmetric hydrogenation of 2-oxazolones† †Electronic supplementary information (ESI) available. CCDC 1584965. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c8sc01869c

A highly enantioselective synthesis of optically active 4-substituted 2-oxazolidinones was described via ruthenium(ii)–NHC-catalysed asymmetric hydrogenation of 2-oxazolones.


Introduction
Chiral 2-oxazolidinones, widely used as Evans' chiral auxiliaries (Scheme 1a, le), play a prominent role in modern organic synthesis. 1 Based on chiral 2-oxazolidinone auxiliaries, a wide range of asymmetric transformations has been developed to construct new chiral building blocks, which are frequently used in both natural product synthesis and drug discovery. 2 Furthermore, the chiral 2-oxazolidinone motif itself is common in pharmaceutically relevant molecules (Scheme 1a,right). 3 Thus, the synthesis of chiral 2-oxazolidinones has already attracted considerable attention. Conventionally, enantioenriched 2-oxazolidinones are synthesized by cyclization of the corresponding optically pure b-amino alcohols with C1building blocks like phosgene and its derivatives (Scheme 1b). These methods oen require toxic reagents and the preconstruction of the key stereocenter of the b-amino alcohols arises synthetic problems if they cannot be formed from natural enantiopure amino acids or related precursors. 2 Considering this, the exploration of orthogonal, efficient and divergent catalytic strategies for the construction of diverse 2-oxazolidinone derivatives is highly important for organic synthesis and drug discovery.
In the last decade, the asymmetric hydrogenation of unsaturated heterocycles has emerged as a conceptually powerful method to produce optically active cyclic compounds and has received signicant attention. 4 In this regard, a synthetic method utilizing the enantioselective hydrogenation of 2-oxazolones for the late-stage construction of the key stereocenter can be envisioned as a powerful alternative to prepare diverse optically active 2-oxazolidinones (Scheme 1c). Moreover, this strategy would also provide a general way to produce optically active b-amino alcohols since the transformation from 2-oxazolidinones to free b-amino alcohols is very convenient. 2 Recently, Zhang and co-workers reported the rhodium-catalysed asymmetric hydrogenation of 2-oxazolones, which afforded 4aryl substituted 2-oxazolidinones with moderate enantioselectivities. 5 To the best of our knowledge, this is the only precedent of an enantioselective synthesis of chiral 2-oxazolidinones by asymmetric hydrogenation of unsaturated heterocycles. As a continuous effort in the eld of arene and heterocycle hydrogenation, 6 we herein describe a highly enantioselective and practical hydrogenation of a broad scope of 2-oxazolones to access diverse enantioenriched 2-oxazolidinones catalysed by a ruthenium(II)-N-heterocyclic carbene (NHC) complex.

Results and discussion
Initially, the hydrogenation of 4-phenyloxazol-2(3H)-one (1a) was attempted under 50 bar H 2 in n-hexane at room temperature in the presence of our previously-developed ruthenium(II)-NHC catalyst, which is prepared in situ from [Ru(2methylallyl) 2 (COD)], the NHC precursor (R,R)-SINpEt$HBF 4 , and NaOt-Bu. 7,8 However, the desired product was not observed, presumably due to catalyst deactivation by coordination of the metal with free N-H (Table 1, entry 1). To circumvent catalyst deactivation, a variety of protecting groups were investigated, of which the 4-methoxybenzyl (PMB) protected carbamate 1d was found to be suitable for hydrogenation, providing the desired 4phenyloxazolidin-2-one with 85% ee and in 95% isolated yield (entry 2-4). To further improve the reaction conditions, a solvent screen was conducted (entries [4][5][6][7][8]. No reaction occurred in dichloromethane presumably due to catalyst decomposition (entry 5), and cyclohexane (entry 8) was found to be slightly superior to other solvents (n-hexane, toluene and THF) for enantioselectivity. Decreasing the reaction temperature to 0 C further improved enantiocontrol, providing the desired product with 95% ee and in 93% yield (entry 9). Finally, a solvent mixture of cyclohexane/THF ¼ 20/1 was used to improve the solubility of the substrate, to afford the chiral 2oxazolidinone 2d in 99% yield and 95% ee (entry 10).
With the optimised reaction conditions in hand (Table 1, entry 10), the substrate scope of the reaction was explored (Schemes 2   % ee values were determined by HPLC analysis using a chiral stationary phase. a Using a solvent mixture of cyclohexane/THF (1/1). b Using THF (2.0 mL). c At À10 C. and 3). First, the variation of the protecting group from N-PMB to N-methyl afforded the products with similar results (Scheme 2, 2d and 2e). Next, the positional inuence of substituents on the phenyl ring was investigated. Methyl groups in the 2-, 3-and 4positions were well-tolerated, providing the corresponding 2oxazolidinones with excellent enantioselectivities and in high yields (2f-h). The electronic effect of the substituents was also examined. Both, electron-rich and electron-decient substrates (1i and 1j respectively) underwent hydrogenation to smoothly afford products 2i and 2j. Halogenated substrates 1k-m were also used to provide the desired products 2k-m with excellent enantioselectivities, in very high yields and without the formation of dehalogenated byproducts. Notably, the catalytic system showcased a robust reactivity, tolerating various functional groups and useful motifs (such as SMe, 1,3-benzodioxole, morpholine, CO 2 Me and SO 2 Me) to provide the corresponding products 2o-s with 91-96% ee and in 76-99% yield. These functional groups and motifs (2i-s) provide an excellent opportunity for further applications of the 2-oxazolidinone products. In addition, the absolute conguration of 2s was determined to be (R) by X-ray crystallographic analysis. 9 The absolute conguration of all other products was assigned by analogy. Additionally, substrates with condensed-ring and heteroaromatic moieties were also tested. Both 1-and 2-naphthyl-substituted substrates were tolerated under the standard conditions (2t and 2u). Remarkably, thiophene and pyridine containing substrates did not poison the Ru-NHC catalyst, producing the corresponding products (2v and 2w) with 93% ee and 94% ee respectively. We further explored the substrate scope with 4-alkyl substituents (Scheme 3). Alkyl substituted substrates with different steric demand were systematically tested. Simple 4methyloxazolidin-2-one 2x was obtained with moderate enantioselectivity. Better control of the stereoselectivity was observed when introducing n-butyl substituent (2y). Isopropyl, cyclopropyl, and cyclohexyl substituents were successfully employed, affording the corresponding products with 91-94% ee and in 92-95% yield (2z-ab). Finally, tert-butyl substituted 2-oxazolone 1ac was tolerated to give the product 2ac with 90% ee and in 83% yield. These results indicate that bulky alkyl groups are benecial for the enantioinduction. In addition, bicyclic 2oxazolidinone 2ad was also obtained by the developed method, albeit with moderate ee.
We then proceeded to demonstrate applications of this methodology (Scheme 4). Scale-up of the hydrogenation of 4-(1naphthyl) substituted substrate 1t to gram-scale provided the corresponding 2-oxazolidinone 2t with 95% ee and in 99% yield (Scheme 4a). Remarkably, the catalyst loading was successfully reduced to 0.2 mol%, which greatly increases the synthetic viability of this protocol. The deprotection of the PMB group was conveniently completed to afford 3 in 99% yield and without loss of enantiomeric excess. 6d Hydrolysis of 2t using NaOH liberated the N-PMB b-amino alcohol 4 in 92% yield. The enantiopurity of product 3 was readily increased to >99% ee aer recrystallization from ethyl acetate. Optically pure b-amino alcohol 5 was then prepared in quantitative yield by cleavage of the 2-oxazolidinone 3 using diethylenetriamine (Scheme 4b). 10 The absolute conguration of b-amino alcohol 5 was reaffirmed to be (R) by comparing the optical rotation to the literature data. 11 The b-amino alcohol 5 was further transformed into 1naphthyl-substituted bisoxazoline ligand 6 in 74% yield by reaction with dimethylmalononitrile and Zn(OTf) 2 . 11a,12 Furthermore, aryl iodide and different N-alkyl substituted substrate 1ae was well tolerated under the established reaction conditions (Scheme 4c). Hydrogenation of 1ae using (S,S)-Scheme 3 Substrate scope of 4-alkyl substituted 2-oxazolones. For detailed conditions, see ESI. † Yields of isolated products after column chromatography are reported. % ee values were determined by HPLC analysis using a chiral stationary phase. a Using 5 mol% of the catalyst. SINpEt$HBF 4 as the carbene ligand precursor, furnished the oxazolidinone 2ae (92% ee and 96% yield), a key synthetic intermediate employed in the synthesis of the alkaloid (À)-aurantioclavine. 13

Conclusions
In summary, we have developed a protocol for the catalytic enantioselective hydrogenation of 2-oxazolones to obtain optically active 2-oxazolidinone derivatives. The ruthenium(II)-NHC catalyst system enabled a broad range of substrates to be successfully hydrogenated with excellent enantioselectivities (up to 96% ee) and in yields (up to 99%), thus to make the approach practical for the rst time. Various functional groups and synthetically useful motifs were well-tolerated. The synthetic utility of this protocol was further demonstrated by performing a reaction on a gramscale with a reduced catalyst loading; the obtained enantioenriched product was readily converted into an optically pure bamino alcohol and subsequently a bisoxazoline ligand. The formal synthesis of (À)-aurantioclavine was enabled by the functional group tolerance towards an iodine substituent at the aryl ring and a varied N-alkyl chain.

Conflicts of interest
There are no conicts to declare.