Enantioselective Michael/Hemiketalization Cascade Reactions between Hydroxymaleimides and 2-Hydroxynitrostyrenes for the Construction of Chiral Chroman-Fused Pyrrolidinediones

In this paper, the organocatalytic asymmetric Michael addition/hemiketalization cascade reactions between hydroxymaleimides and 2-hydroxynitrostyrenes were developed, which provided a new protocol for building a chiral ring-fused chroman skeleton. This squaramide-catalyzed cascade reaction provided chiral chroman-fused pyrrolidinediones with three contiguous stereocenters in good to high yields (up to 88%), with excellent diastereoselectivities (up to >20:1 dr) and enantioselectivities (up to 96% ee) at −16 °C. Moreover, a scale-up synthesis was also carried out, and a possible reaction mechanism was proposed.


Introduction
The development of pharmaceutical science is inseparable from the discovery of lead compounds, primarily derived from natural products and analogues with biological activity. Ring-fused chroman skeletons are widely present in many natural products and analogues with bioactivity [1-11] (Figure 1). For example, myrtucommulone E, isolated from the Mediterranean folk herb Myrtucommulone, has α-glucosidase inhibitory activity [1]; rhodomyrtone, isolated from the leaves of myrtle, a small Indonesian shrub, has antibacterial activity [2]; miroestrol, isolated from the root of kudzu, a Thai herb, has estrogenic activity [3]; rhododaurichromanic acid A, isolated from the shoots and leaves of azalea from northern China and eastern Siberia, has anti-HIV activity [4] and so on. In addition to specific actions in the field of biomedicine, the ring-fused chroman skeleton also shows particular potential in the area of modern pesticide development. For example, greveichromenol, isolated from geronniang, the traditional herbal medicine of the Dai ethnic group, showed antitobacco mosaic virus activity, providing a lead compound for the development of antitobacco mosaic virus pesticide [5]. Therefore, the construction of the chroman skeleton has received great attention from organic and medicinal chemists. In recent years, a large number of synthetic strategies of chroman derivatives have been consistently reported [12][13][14][15][16][17][18][19][20][21][22][23][24][25][26].
2-Hydroxynitrostyrene plays an essential role in these reactions in the construction of chiral chroman derivatives. For example, in 2013, Zhu's group reported a new asymmetric oxa-Michael/Michael cascade reaction for the construction of enantiomerically enriched indolinone spiro-fused chromans; this protocol offered excellent stereo control under mild conditions [27] (Scheme 1a). In 2018, Xu's group developed an asymmetric catalytic method for the synthesis of polysubstituted chromans through an oxa-Michael-nitro-Michael reaction, and the squaramide-catalyzed domino reaction of 2-hydroxynitrostyrenes with trans-β-nitroolefins produced chiral chromans with excellent enantioselectivities, diastereoselectivities and yields [28] (Scheme 1b). At the same time, because of the importance 2-Hydroxynitrostyrene plays an essential role in these reactions in the construction of chiral chroman derivatives. For example, in 2013, Zhu's group reported a new asymmetric oxa-Michael/Michael cascade reaction for the construction of enantiomerically enriched indolinone spiro-fused chromans; this protocol offered excellent stereo control under mild conditions [27] (Scheme 1a). In 2018, Xu's group developed an asymmetric catalytic method for the synthesis of polysubstituted chromans through an oxa-Michael-nitro-Michael reaction, and the squaramide-catalyzed domino reaction of 2-hydroxynitrostyrenes with trans-β-nitroolefins produced chiral chromans with excellent enantioselectivities, diastereoselectivities and yields [28] (Scheme 1b). At the same time, because of the importance of the 2,5-pyrrolidinedione framework in biomedicine [29,30], Wang's group synthesized chiral chroman-fused pyrrolidinediones for the first time in excellent yields with excellent stereoselectivities using organocatalytic enantioselective [4+2] cyclization reaction [31] (Scheme 1c). Inspired by their work and continuing with our project with organocatalyzed domino or cascade reactions for the synthesis of bioactive heterocycles, we intend to synthesize chiral chroman-fused pyrrolidinediones using 2-hydroxynitrostyrenes as substrates to consolidate and develop this research result (Scheme 1d).  2-Hydroxynitrostyrene plays an essential role in these reactions in the constru of chiral chroman derivatives. For example, in 2013, Zhu's group reported a new metric oxa-Michael/Michael cascade reaction for the construction of enantiomerical riched indolinone spiro-fused chromans; this protocol offered excellent stereo contr der mild conditions [27] (Scheme 1a). In 2018, Xu's group developed an asymmetric lytic method for the synthesis of polysubstituted chromans through an oxa-Michael-Michael reaction, and the squaramide-catalyzed domino reaction of 2-hydroxynit renes with trans-β-nitroolefins produced chiral chromans with excellent enantiosel ties, diastereoselectivities and yields [28] (Scheme 1b). At the same time, because importance of the 2,5-pyrrolidinedione framework in biomedicine [29,30], Wang's synthesized chiral chroman-fused pyrrolidinediones for the first time in excellent with excellent stereoselectivities using organocatalytic enantioselective [4+2] cycli reaction [31] (Scheme 1c). Inspired by their work and continuing with our projec organocatalyzed domino or cascade reactions for the synthesis of bioactive heteroc we intend to synthesize chiral chroman-fused pyrrolidinediones using 2-hydroxynit renes as substrates to consolidate and develop this research result (Scheme 1d). Scheme 1. Previous reports and our work. (a. oxa-Michael/Michael cascade reaction [27]; b Michael/nitro-Michael reaction [28]; c. [4+2] cyclization reactin, [31]; d. our research plan.). Scheme 1. Previous reports and our work. ((a) oxa-Michael/Michael cascade reaction [27]; (b) oxa-Michael/nitro-Michael reaction [28]; (c) [4+2] cyclization reaction, [31]; (d) our research plan.).

Optimization of Reaction Conditions
Initially, we started our study with hydroxymaleimide 1a (3-ethyl-4-hydroxy-1-phenyl-1H-pyrrole-2,5-dione) and 2-hydroxynitrostyrene 2a as model substrates. We first tested the feasibility of the model reaction in the presence of 10 mol% cinchona-derived squaramide bifunctional catalyst C1 ( Figure 2) in dichloromethane (DCM) at room temperature. Under these conditions, the desired product 3aa was obtained in high yield (82%) with excellent diastereoselectivity (>20:1 dr), although with moderate enantioselectivity (53% ee). Encouraged by this important result and inspired by our previous work [32,33], we tried Temperature is a vital factor to affect the stereoselectivity in asymmetric organic reaction, so we evaluated reaction temperature at first. Unexpectedly, a lower temperature can lead to higher yield and enantioselectivity. To avoid the contingency of the case, we took another catalyst C2 to confirm this. We can easily find that a lower temperature is better for reaction from entries 1-4. Given the temperature conditions, we evaluated several catalysts (Table 1, entries 2, 4-12) next; the cinchona-derived bifunctional thiourea catalyst C9 had the best yield, while the enantioselectivity was ordinary and the diastereoselectivity was not good enough (entry 11, 90% yield, −69% ee, 16:1 dr). The cinchonaderived squaramide catalyst C6 has almost the best yield, as well as diastereoselectivity, but enantioselectivity is so low that we do not consider it (entry 8, 89% yield, 59% ee, >20:1 dr). In terms of enantioselectivity, cinchona-derived squaramide catalysts are better, compared to the diaminocyclohexane-derived squaramide catalyst C10 (entry 12, 81% yield, 59% ee, 9:1 dr). When taking into account entry 2 (80% yield, 87% ee, >20:1 dr) and entry 4 (83% yield, 73% ee, >20:1 dr), we easily find that the cinchona-derived squaramide catalyst C1 is superior to the cinchona-derived thiourea catalyst C2. Taking into account entry 2 (80% yield, 87% ee, >20:1 dr) and entry 5 (88% yield, 96% ee, >20:1 dr), we can easily find Temperature is a vital factor to affect the stereoselectivity in asymmetric organic reaction, so we evaluated reaction temperature at first. Unexpectedly, a lower temperature can lead to higher yield and enantioselectivity. To avoid the contingency of the case, we took another catalyst C2 to confirm this. We can easily find that a lower temperature is better for reaction from entries 1-4. Given the temperature conditions, we evaluated several catalysts (Table 1, entries 2, 4-12) next; the cinchona-derived bifunctional thiourea catalyst C9 had the best yield, while the enantioselectivity was ordinary and the diastereoselectivity was not good enough (entry 11, 90% yield, −69% ee, 16:1 dr). The cinchona-derived squaramide catalyst C6 has almost the best yield, as well as diastereoselectivity, but enantioselectivity is so low that we do not consider it (entry 8, 89% yield, 59% ee, >20:1 dr). In terms of enantioselectivity, cinchona-derived squaramide catalysts are better, compared to the diaminocyclohexane-derived squaramide catalyst C10 (entry 12, 81% yield, 59% ee, 9:1 dr).

Optimization of Reaction Conditions
Initially, we started our study with hydroxymaleimide 1a (3-ethyl-4-hydroxy nyl-1H-pyrrole-2,5-dione) and 2-hydroxynitrostyrene 2a as model substrates. W tested the feasibility of the model reaction in the presence of 10 mol% cinchona-d squaramide bifunctional catalyst C1 ( Figure 2) in dichloromethane (DCM) at roo perature. Under these conditions, the desired product 3aa was obtained in hig (82%) with excellent diastereoselectivity (>20:1 dr), although with moderate enant tivity (53% ee). Encouraged by this important result and inspired by our previou [32,33], we tried to reduce the reaction temperature to −16 °C to improve the enant tivity. Luckily, the results improved to 80% yield, >20:1 dr, and 87% ee. Furtherm tried to screen several catalysts, reaction solvent, and catalyst loading to further im the outcome and enantioselectivity. The results are outlined in Table 1. After a preliminary screening of the catalysts, a survey of the solvent effect using C3 as the organocatalyst concluded that dichloromethane (DCM) was still the best solvent among 1,2-dichloroethane (DCE), chloroform, toluene, acetonitrile, tetrahydrofuran (THF). Some solvents, such as 1,4-dioxane, will freeze at −16 • C; we do not consider these (Table 1, entries 13-17). It seems unexpected that when the reaction was carried out in THF, only trace products was detected by TLC. In general, THF will have an effect on stereoselectivity owing to compete hydrogen bond formation with catalyst, but the great influence on product yield may have other reason. We finally discovered that hydroxymaleimide 1a, one of the substrates, cannot dissolve well in THF, which could shed light on this phenomenon.
Afterwards, we further investigated the reaction with 5 and 2.5 mol% catalyst loading, respectively (Table 1, entries 18 and 19), and no improvements were obtained. Therefore, the optimal reaction conditions for this Michael/hemiketalization cascade reaction were to use a catalyst loading of 10 mol% of squaramide C3 in DCM at −16 • C for 24 h.

Substrate Scope
With the optimized conditions in hand, we then began to investigate the substrate scope and limitation of this reaction, and the results are summarized in Scheme 2.

Scaled-Up Synthesis
To prove the synthetic value of this cascade reaction, a scaled reaction of 1a and 2a with amplification ten times was carried out under standard conditions. As shown in Scheme 3, the desired product 3aa was obtained in slightly reduced yield (85%), with maintained diastereoselectivity and slightly lower enantioselectivity (>20:1 dr, 94% ee). This result shows that this asymmetric catalytic strategy has broad prospects for mass production. Firstly, we examined the tolerance of various hydroxymaleimides 1 under the optimized conditions. Various hydroxymaleimides with electron withdrawing and electron donating substituents at the 4-position on the benzene ring participated in the reaction easily, and the corresponding products 3ba-3ha could be generated in high yields (79-88%) with excellent stereoselectivities (up to >20:1 dr and up to >99% ee), except products 3da and 3fa, whose diastereoselectivity was 4:1 and 5:1, respectively. There is no clear rule of the influence of substituents on stereoselectivity. The effect of substituents on stereoselectivity cannot be explained by the electronic effect, because the enantioselectivity and diastereoselectivity are affected by many factors. Afterwards, the cascade process gave the desired products 3ia-3na with high stereoselectivity and good yield, even with the substituents at the 3-position or 3,5-position on the benzene ring. Unfortunately, substrate 1o substituted with phenyl did not work with 2a; no other addition was observed. The starting materials were recovered unaltered, probably because of the steric hindrance and the phenyl delocalization of the negative charge, so that the intramolecular Michael addition step cannot occur. Furthermore, we tested some substituents at different positions on 2-hydroxynitrostyrenes; most of them showed excellent results, including 3ab-3ae and 3ag (up to >20:1 dr, 92% ee). However, when the substituent is a nitro group, such as 2f, the reaction cannot work well, probably due to the strong electron-withdrawing effect of the nitro group, which lowers the nucleophilic reactivity of the corresponding phenoxy anion. The nitro group in 2f can block the catalyst by hydrogen bonding, and this may also hinder the reaction from proceeding.
We also tried to lower the temperature to increase the stereo control of the reaction. One can take 3ca as an example. When the reaction was carried out at −30 • C, the trace product was detected even after 72 h; when the reaction was carried out at −20 • C, the yield decreased to 78% and the diastereoselectivity remained at 10:1 dr.

Scaled-Up Synthesis
To prove the synthetic value of this cascade reaction, a scaled reaction of 1a and 2a with amplification ten times was carried out under standard conditions. As shown in Scheme 3, the desired product 3aa was obtained in slightly reduced yield (85%), with maintained diastereoselectivity and slightly lower enantioselectivity (>20:1 dr, 94% ee). This result shows that this asymmetric catalytic strategy has broad prospects for mass production.

X-ray Diffraction Analysis
A single crystal of 3ca was obtained from the slow evaporation from the mixed solvents of methanol and dichloromethane. The absolute configuration of product 3ca was unambiguously determined by single-crystal X-ray diffraction analysis as (3aS,9S,9aR) (

X-ray Diffraction Analysis
A single crystal of 3ca was obtained from the slow evaporation from the mixed solvents of methanol and dichloromethane. The absolute configuration of product 3ca was unambiguously determined by single-crystal X-ray diffraction analysis as (3aS,9S,9aR) (

X-ray Diffraction Analysis
A single crystal of 3ca was obtained from the slow evaporation from the mixed solvents of methanol and dichloromethane. The absolute configuration of product 3ca was unambiguously determined by single-crystal X-ray diffraction analysis as (3aS,9S,9aR) ( Figure 3) [34] (See Supplementary Materials). The absolute configurations of the other products were assigned by analogy.

Controlled Reactions and Plausible Mechanism
For a better understanding of the mechanism of this cascade reaction, we designed three controlled reactions (Scheme 4). When hydroxymaleimide 1b and β-nitrostyrene 2h were used under the optimized conditions, the corresponding Michael addition product 4 was obtained with a yield of 20%. This result indicated that 1a is a suitable Michael donor to trigger the first Michael addition step of the Michael/hemiketalization cascade reaction. When the OH group in hydroxymaleimide 1a was protected with the acetyl group, as in substrate 1q, no reaction was observed when 1q reacted with 2-hydroxynitrostyrene 2a. Meanwhile, N-phenylmaleimide 1r also could not react with 2a. The last two control experiments indicated that the OH group in hydroxymaleimides is essential, and the hemiketalization reaction was the second step of the cascade reaction, instead of the oxa-Michael reaction as the first step.

Controlled Reactions and Plausible Mechanism
For a better understanding of the mechanism of this cascade reaction, we designed three controlled reactions (Scheme 4). When hydroxymaleimide 1b and β-nitrostyrene 2h were used under the optimized conditions, the corresponding Michael addition product 4 was obtained with a yield of 20%. This result indicated that 1a is a suitable Michael donor to trigger the first Michael addition step of the Michael/hemiketalization cascade reaction. When the OH group in hydroxymaleimide 1a was protected with the acetyl group, as in substrate 1q, no reaction was observed when 1q reacted with 2-hydroxynitrostyrene 2a. Meanwhile, N-phenylmaleimide 1r also could not react with 2a. The last two control experiments indicated that the OH group in hydroxymaleimides is essential, and the hemiketalization reaction was the second step of the cascade reaction, instead of the oxa-Michael reaction as the first step. Based on these experimental results and previous work [33], we proposed a plausible mechanism based on the absolute configuration of 3aa (Scheme 5). In the first step of Michael addition, the squaramide catalyst C3 initially promotes the formation of transition state A, and catalyst C3 works in a double activation model. 2-Hydroxynitrostyrene 2a is oriented and activated by the squaramide moiety through double hydrogen bonding and the OH group in 1a is deprotoned by the tertiary amine unit to form enolate, which is oriented by another hydrogen bond. 2-Hydroxynitrostyrene 2a is attacked by the enolate of 1a from the Si-face. In the second step of the hemiketalization reaction, the OH group in 2-hydroxynitrostyrene 2a is deprotoned by the tertiary amine unit in squaramide, and the newly formed carbonyl group in 1a is attacked by the deprotonated phenolic hydroxyl of 2a from the Si-face via transition state B, leading to the formation of (3aS,9S,9aR)-con- Based on these experimental results and previous work [33], we proposed a plausible mechanism based on the absolute configuration of 3aa (Scheme 5). In the first step of Michael addition, the squaramide catalyst C3 initially promotes the formation of transition state A, and catalyst C3 works in a double activation model. 2-Hydroxynitrostyrene 2a is oriented and activated by the squaramide moiety through double hydrogen bonding and the OH group in 1a is deprotoned by the tertiary amine unit to form enolate, which is oriented by another hydrogen bond. 2-Hydroxynitrostyrene 2a is attacked by the enolate of 1a from the Si-face. In the second step of the hemiketalization reaction, the OH group in 2-hydroxynitrostyrene 2a is deprotoned by the tertiary amine unit in squaramide, and the newly formed carbonyl group in 1a is attacked by the deprotonated phenolic hydroxyl of 2a from the Si-face via transition state B, leading to the formation of (3aS,9S,9aR)-configured 3aa and regenerates the bifunctional catalyst C3 after a protonation process.

Conclusions
In conclusion, we have successfully developed novel Michael addition/hemiketalization cascade reactions between hydroxymaleimides and 2-hydroxynitrostyrenes to synthesize chiral ring-fused chromans. Under mild conditions, a range of structurally diverse chiral chroman-fused pyrrolidinediones, containing hemiketals, were obtained in good to high yields with excellent stereoselectivities. Additionally, the potential utility of this methodology has been demonstrated by scaling-up synthesis. This cascade synthetic strategy is bound to be a powerful tool for medicinal chemistry studies.

Conclusions
In conclusion, we have successfully developed novel Michael addition/hemiketalization cascade reactions between hydroxymaleimides and 2-hydroxynitrostyrenes to synthesize chiral ring-fused chromans. Under mild conditions, a range of structurally diverse chiral chroman-fused pyrrolidinediones, containing hemiketals, were obtained in good to high yields with excellent stereoselectivities. Additionally, the potential utility of this methodology has been demonstrated by scaling-up synthesis. This cascade synthetic strategy is bound to be a powerful tool for medicinal chemistry studies.

General Information
Commercially available compounds were used without further purification. Solvents were dried according to standard procedures. Column chromatography was performed with silica gel (200-300 mesh). The melting points were determined with an XT-4 melting point apparatus and were not corrected. 1 H NMR spectra were measured with a Bruker Ascend 400 MHz spectrometer (Karlsurhe, Germany) and the chemical shifts were reported in δ (ppm) relative to tetramethylsilane (TMS) as the internal standard. 13 C NMR spectra were measured at 100 MHz with a 400 MHz spectrometer, and the chemical shifts were reported in ppm relative to tetramethylsilane and referenced to the solvent peak (CDCl 3 , δC = 77.00 ppm; CD 3 OD, δC = 49.05 ppm; acetone-d6, δC = 30.83 ppm). High-resolution mass spectra were measured with an Agilent 6520 Accurate-Mass Q-TOF MS system (Beijing, China), equipped with an electrospray ionization (ESI) source. Optical rotations were measured with a Krüss P8000 polarimeter (Beijing, China) at the indicated concentration with units of g/100 mL. Enantiomeric excesses were determined by chiral HPLC analysis, using an Agilent 1200 LC instrument (Beijing, China) with a Daicel Chiralpak IA, IB, IC, or AD-H column.

Procedure for the Asymmetric Synthesis of Compounds 3
In a small dried bottle, 1 (0.10 mmol), 2 (0.12 mmol), chiral organocatalyst C3 (6.0 mg, 0.01 mmol, 10 mol%) and DCM (1.0 mL) were added. The mixture was stirred at −16 • C for 24 h. After completion of the reaction, the residue was purified by flash column chromatography on silica gel to obtain the pure products 3 as solids. Racemates were prepared following a similar procedure with Et 3 N (20 mol%).