Facile access to chiral γ-butyrolactones via rhodium-catalysed asymmetric hydrogenation of γ-butenolides and γ-hydroxybutenolides

Abstract

The highly efficient Rh/ZhaoPhos-catalysed asymmetric hydrogenation of γ-butenolides and γ-hydroxybutenolides was successfully developed. This protocol provides an efficient and practical approach to the synthesis of various chiral γ-butyrolactones, which are synthetically valuable building blocks of diverse natural products and therapeutic substances, with excellent results (up to >99% conversion and 99% ee). Further follow-up transformations have been revealed to accomplish creative and efficient synthetic routes for several enantiomerically enriched drugs via this catalytic methodology.

A general and efficient rhodium(i)-catalyzed asymmetric hydrogenation of γ-butenolides and γ-hydroxybutenolides was disclosed, providing a wide range of chiral γ-butyrolactones in high yields and enantioselectivities with Rh/ZhaoPhos as a catalyst.

γ-Butyrolactone is an essential framework existing in large numbers of natural products¹ and pharmaceutical compounds² and shows unique and excellent biological activities (Fig. 1). Furthermore, enantiomerically pure γ-butyrolactones could serve as building blocks for the construction of a wide range of complex molecules,³ and diverse transformations have been developed,⁴ especially in the highly versatile furanone structure, to synthesize physiologically and therapeutically important reagents, for example brivaracetam,⁵ arctigenin,⁶ pilocarpine,⁷etc.

Fig. 1. Natural products containing the γ-butyrolactone core, and pharmaceutical compounds derived from γ-butyrolactones.

As commercially available or easily accessible compounds, γ-butenolides⁸ are important synthetic precursors towards γ-butyrolactones, and an impressive range of transformations have been developed for the enantioselective construction of chiral γ-butyrolactones.⁹ With the advancement in the field of asymmetric metal catalysis, enantioselective 1,4-addition and reduction of γ-butenolides have been considered representative methodologies of preparing optically active γ-butyrolactones. In contrast to the well-developed asymmetric 1,4-additions¹⁰ (Scheme 1a), with transition metal-catalysed 1,4-reduction of γ-butenolides as an alternative it would be difficult to achieve satisfactory catalytic conversion and stereoselectivity as well as broad substrate scope because of the lack of efficient metal catalysts.¹¹ The Buchwald group reported a CuH-catalysed enantioselective conjugate reduction of γ-butenolides;¹² however, the requirement of using high catalyst loading and a strong base additive could limit the application of this method in organic synthesis and the pharmaceutical industry. Recently, our group developed a bisphosphine-thiourea chiral ligand, ZhaoPhos,¹³ which exhibited extraordinary potential in rhodium¹⁴ and iridium-catalyzed¹⁵ asymmetric hydrogenation due to its powerful hydrogen-bonding and anion-binding ability resulting from the thiourea moiety. Inspired by this novel strategy, we herein demonstrated a Rh/ZhaoPhos catalysed asymmetric hydrogenation of easily accessible γ-butenolides and γ-hydroxybutenolides, which could be a straightforward method of approaching chiral γ-butyrolactones with excellent catalytic results (Scheme 1b).

Scheme 1. Current methods of preparing chiral γ-butyrolactones from γ-butenolides, and proposed Rh/ZhaoPhos-catalyzed hydrogenation.

We initiated our investigation with evaluation of various combinations of the rhodium precatalyst and different chiral phosphine ligands for the asymmetric hydrogenation of 4-phenylfuran-2(5H)-one 1a. After examining a series of commonly used chiral phosphine ligands and using THF as the solvent, very poor enantioselectivity was achieved with ligands such as DuanPhos, SegPhos, BINAP and so on (Table 1, entries 1–7). Employing ZhaoPhos as a ligand could facilitate the hydrogenation with a significantly improved enantioselective manner, and the desired product 2a was obtained in 79% conversion and 97% ee under 50 atm of H₂ at room temperature within 24 hours (entry 8). Furthermore, the solvent was crucial to the control of the reaction rate, and by screening several protic and aprotic solvents (Table 2, entries 1–9), we found that by employing DCM as solvent complete hydrogenation could be achieved under standard conditions, which afforded 2a with >99% conversion and in 98% ee (entry 8).

Ligand screening for the asymmetric hydrogenation of 1a.

Entrya	Ligand	Conv.b (%)	eec (%)
1	(RC,SP)-DuanPhos	14	1
2	(S)-SegPhos	100	2
3	(S)-BINAP	78	0
4	(R,S)-JosiPhos	100	2
5	(R)-MeO-BIPHEP	13	10
6	WalPhos	31	16
7	(R)-C3*-TunePhos	8	13
8	ZhaoPhos	79	97

^aUnless otherwise noted, all hydrogenations were carried out with a [Rh(NBD)₂BF₄]/ligand/1a (0.1 mmol) ratio of 1 : 1.1 : 100 in 1.0 mL of THF under H₂ (50 atm) at room temperature for 24 h.

^bDetermined by ¹H NMR.

^cDetermined by HPLC analysis, and the absolute configuration of 2a was determined as S by comparing the optical rotation data with the literature.

Solvent screening for the asymmetric hydrogenation of 1a.

Entrya	Solvent	Conv.b (%)	eec (%)
1	MeOH	29	96
2	EtOH	19	97
3	CF3CH2OH	59	98
4	EtOAc	>99	97
5	Toluene	91	95
6	THF	79	97
7	1,4-Dioxane	30	16
8	DCM	>99	98
9	DCE	25	96

^aUnless otherwise noted, all hydrogenations were carried out with a [Rh(NBD)₂BF₄]/ligand/1a (0.1 mmol) ratio of 1 : 1.1 : 100 in 1.0 mL of solvent under H₂ (50 atm) at room temperature for 24 h.

^bDetermined by ¹H NMR.

^cDetermined by HPLC analysis, and the absolute configuration of 2a was determined as S by comparing the optical rotation data with the literature.

With optimized reaction conditions in hand, we then continued to investigate the substrate generality (Table 3). A considerable number of β-aryl substituents (2a–l) on γ-butenolides were investigated, and the hydrogenation proceeded smoothly to yield the desired products with excellent yields (95 ∼ 99%) and in high enantioselectivities (97 ∼ 98% ee), in which both electron-withdrawing (2b–2d) and electron-donating (2e–2j) substituents on the aryl ring at different positions were well tolerated. The hydrogenation of substrates with the bulky naphthyl moiety (1k) and heterocyclic unit (thienyl, 1l) was feasible, giving the corresponding product with excellent results (95 ∼ 99% yields and 98% ee). In addition, the alkyl substituents (2m–n) were explored as well, and it is noteworthy that α-methyl substituted lactone 2n could be produced with 84% or 82% ee respectively from the substrates featuring either endo- (1n) or exocyclic (1n′) unsaturation, in which this transformation was rarely achieved with good yield and in acceptable enantioselectivity according to previous studies.¹⁶ The reactions of α,β-disubstituted γ-butenolides 1o and 1p were also conducted; however, no desired products were obtained.

Substrate scope of the hydrogenation of γ-butenolides 1^a,b,c.

^aUnless otherwise noted, all hydrogenations were carried out with a [Rh(NBD)₂BF₄]/ligand/1a (0.1 mmol) ratio of 1 : 1.1 : 100 in 1.0 mL of DCM under H₂ (50 atm) at room temperature for 24 h.

^bIsolated yield.

^cDetermined by HPLC analysis, and the absolute configuration was determined by comparing the optical rotation data with the literature.

Apart from the hydrogenation of γ-butenolides, γ-hydroxybutenolides 3 were also successfully reduced under the same reaction conditions with CF₃COOH as an additive (see the ESI† for details), and consequently afforded γ-butyrolactones 2 directly in high yields (95 ∼ 99%) and with excellent enantioselectivities (up to 99% ee) (Table 4). We suggested that CF₃COOH could activate the hydroxy group of lactol, which could be protonated to form an oxonium salt, and then hydrogenated by the catalyst.^15e,f γ-Hydroxybutenolides could be naturally occurring as well as easily prepared, and on account of realizing these transformations, the scope of the starting material for preparing chiral γ-butyrolactones was extremely expanded, which further improved the synthetic value of the Rh/ZhaoPhos catalyst in the field of asymmetric hydrogenation.

Substrate scope of the hydrogenation of γ-hydroxybutenolides 3^a,b,c.

^aUnless otherwise noted, all hydrogenations were carried out with a [Rh(NBD)₂BF₄]/ligand/CF₃COOH/1a (0.1 mmol) ratio of 1 : 1.1 : 10 : 100 in 1.0 mL of DCM under H₂ (50 atm) at room temperature for 24 h.

^bIsolated yield.

^cDetermined by HPLC analysis, and the absolute configuration was determined comparing the optical rotation data with the literature.

To demonstrate the synthetic utilities of this method, a gram-scale hydrogenation of substrate 1m (Scheme 2a) was conducted with a low catalyst loading of 0.02 mol%. To our delight, the hydrogenation underwent smoothly under 70 atm of H₂ at 50 °C using CF₃COOH (10 mol%) as the additive (adding CF₃COOH could not only stabilize the cationic rhodium catalyst to increase the TON of the reaction, but also activate the carbonyl of α,β-unsaturated lactone to facilitate the hydrogenation), and the chiral γ-butyrolactone 2m was afforded in 98% yield and with 95% ee. We further transformed this chiral building block via a ring-opening/recyclization strategy to achieve the total synthesis of brivaracetam in high optical purity (96.9 : 3.1 dr and >99% ee) within only 3 steps (67% overall yield). Moreover, the 2-step synthesis of arctigenin analogue 7 was realized using γ-butenolide 1e as the starting material (79% overall yield, Scheme 2b), which presented a shorter synthetic route compared to previous studies.^6c

Scheme 2. Synthetic applications.

In summary, we have developed a highly enantioselective Rh/ZhaoPhos-catalyzed hydrogenation, providing a series of synthetically useful chiral γ-butyrolactones with readily available γ-butenolides and γ-hydroxybutenolides as starting materials. In addition, this strategy was applicable to the construction of various natural products and therapeutic substances, and scalable and concise syntheses of the pharmaceutical drugs brivaracetam and arctigenin were accomplished through this methodology, which demonstrated its practical utilities.

Data availability

All experimental procedures, characterization, and computational data for this study, can be found in the ESI.†

Author contributions

The manuscript was written through the contributions of all authors. All authors have approved the final version of the manuscript.

Conflicts of interest

There are no conflicts to declare.