Organocatalytic Asymmetric [2 + 4] Cycloadditions of 3-Vinylindoles with ortho-Quinone Methides

Catalytic asymmetric [2 + 4] cycloadditions of 3-vinylindoles with ortho-quinone methides and their precursors were carried out in the presence of chiral phosphoric acid to afford a series of indole-containing chroman derivatives with structural diversity in overall high yields (up to 98%), good diastereoselectivities (up to 93:7 dr) and moderate to excellent enantioselectivities (up to 98% ee). This approach not only enriches the chemistry of catalytic asymmetric cycloadditions involving 3-vinylindoles but is also useful for synthesizing chiral chroman derivatives.


Introduction
Chiral indole derivatives are ubiquitous in biologically important natural products, pharmaceuticals and materials [1][2][3][4][5]. In recent years, vinylindoles have been recognized as versatile reactants for the synthesis of enantioenriched indole derivatives [6,7]. The 3-vinylindoles belong to a class of vinylindoles with multiple reactive sites and are widely applied in organocatalytic asymmetric cycloadditions and substitutions .

Introduction
Chiral indole derivatives are ubiquitous in biologically important natural products, pharmaceuticals and materials [1][2][3][4][5]. In recent years, vinylindoles have been recognized as versatile reactants for the synthesis of enantioenriched indole derivatives [6,7]. The 3-vinylindoles belong to a class of vinylindoles with multiple reactive sites and are widely applied in organocatalytic asymmetric cycloadditions and substitutions .

Results and Discussion
Next, we investigated the effect of the reaction temperature ( Table 2, entries 1-4) and found 0 • C to be the optimal reaction temperature (entry 1 vs. entry 3). Modulating the molar ratio of the reactants (entries 5-8) revealed that increasing the quantity of sesamolderived o-QM 2a improved the yield but decreased the enantioselectivity (entry 3 vs. entries 5-6), whereas increasing the quantity of 3-vinylindole 1a was detrimental to the reaction (entry 3 vs. entries 7-8). Therefore, the most suitable molar reagent ratio remained 1:1.2. Finally, some additives were screened (entries 9-13), and the optimal conditions for this [2 + 4] cycloaddition were set as shown in entry 12. occurred, as expected, to afford the chiral chroman derivative 3aa in a moderate yield with excellent diastereoselectivity, albeit with no enantio-control (43% yield, 96:4 dr, 0% ee). To control the enantioselectivity of this [2 + 4] cycloaddition, a series of CPAs 4 were screened (entries 2-7). The CPA (R)-4c bearing two bulky 3,3′-(1-naphthyl) groups facilitated the [2 + 4] cycloaddition with the highest enantioselectivity of 66% ee (for the major diastereomer) among the investigated catalysts (entry 3 vs. entries 1-2 and 4-7), which could be ascribed to the steric hindrance effect of the 3,3'-disubstituents of CPA in creating a chiral environment for controlling the enantioselectivity [64,65]. The subsequent evaluation of a series of representative solvents (entries [8][9][10][11][12] in the presence of CPA (R)-4c showed that toluene remained the most suitable solvent in terms of controlling the enantioselectivity (entry 3 vs. entries 8-12). Next, we investigated the effect of the reaction temperature ( Table 2, entries 1-4) and found 0 °C to be the optimal reaction temperature (entry 1 vs. entry 3). Modulating the molar ratio of the reactants (entries 5-8) revealed that increasing the quantity of sesamol-derived o-QM 2a improved the yield but decreased the enantioselectivity (entry 3 vs. entries 5-6), whereas increasing the quantity of 3-vinylindole 1a was detrimental to the reaction (entry 3 vs. entries 7-8). Therefore, the most suitable molar reagent ratio remained 1:1.2. Finally, some additives were screened (entries 9−13), and the optimal conditions for this [2 + 4] cycloaddition were set as shown in entry 12.   After establishing the optimal reaction conditions, we investigated the substrate scope of the 3-vinylindoles 1 for catalytic asymmetric [2 + 4] cycloadditions with sesamol-derived o-QM 2a. As shown in Table 3, a variety of 3-vinylindoles 1 bearing different R/R 1 groups underwent [2 + 4] cycloadditions to generate chiral indole-containing chroman derivatives 3 in overall good yields (54-98%) and moderate to excellent stereoselectivities (78:22 dr to 93:7 dr, 55-97% ee). In detail, C5-, C6-and C7-substituted 3-vinylindoles 1b-1f participated in the [2 + 4] cycloaddition with high yields and moderate enantioselectivities (entries 2-6). In addition, a series of ortho-, metaand para-substituted phenyl groups were utilized as R 1 groups for the 3-vinylindoles 1, and the corresponding substrates participated in [2 + 4] cycloaddition with moderate to good results (entries 7-13). Among these 3-vinylindoles, 1l-1m bearing para-substituted phenyl groups (R 1 ) delivered the corresponding products 3la-3ma with the best enantioselectivities (85% ee and 97% ee, entries [12][13]. Notably, these para-substituted substrates 1l-1m displayed a much higher capability in controlling the enantioselectivity than their orthoand meta-substituted counterparts (entries 12-13 vs. entries 7-8 and 10-11), which might be ascribed to the steric effect of the para-substituents. also be utilized as a reaction partner to yield the product 3ah with a high enantioselectivity of 87% ee (entry 8).  Next, the substrate scope of sesamol-derived o-QMs 2 was explored by catalytic asymmetric [2 + 4] cycloaddition with 3-vinylindole 1a (Table 4). This reaction was clearly amenable to participation by a series of sesamol-derived o-QMs 2a-2g bearing either electron-donating or electron-withdrawing groups at different positions of the phenyl ring (entries 1−7), producing chiral indole-containing chroman derivatives 3 in generally high yields (53-98%) and moderate to excellent diastereo-and enantioselectivities (75:25 dr to 89:11 dr, 60-98% ee). Among these o-QMs, 2f-2g bearing para-halogen-substituted phenyl groups delivered products 3af-3ag in the highest enantioselectivities of 97-98% ee (entries 4-7). Notably, o-QM 2h bearing a heteroaromatic 2-thiophenyl group could also be utilized as a reaction partner to yield the product 3ah with a high enantioselectivity of 87% ee (entry 8). a Unless otherwise indicated, the reaction was carried out at a 0.1 mmol scale in toluene (1.0 mL) with MgSO4 (50 mg) for 12 h using a 1:2a molar ratio of 1:1.2. b Isolated total yield of the diastereomeric mixtures. c The diastereomeric ratio (dr) was determined by 1 H NMR. d The ee value refers to that of the major diastereomer and was determined by HPLC. The asterisk * indicates chiral center. The structures of all products 3 were identified by their NMR, IR and HR MS data, and the ee value of all products 3 were calculated by their HPLC traces (see the Supplementary Materials). Although we tried to cultivate the single crystal from enantioenriched products 3, we failed to achieve this goal. So, the absolute configurations of chiral products 3 could not be determined. Nevertheless, when N-methyl-protected 3-vinylindole 1n was employed as a substrate in the reaction with sesamol-derived o-QM 2a under standard conditions (Figure 5a), the [2 + 4] cycloaddition occurred to generated product 3na in a moderate yield and diastereoselectivity (47% yield, 86:14 dr) albeit with an extremely low enantioselectivity (14% ee). Fortunately, we cultivated the single crystal of product 3na, whose relative configuration was determined to be (trans, trans) by X-ray diffraction analysis of the single crystal (CCDC 2100427, see the Supplementary Materials) (Figure 5b). The structures of all products 3 were identified by their NMR, IR and HR MS data, and the ee value of all products 3 were calculated by their HPLC traces (see the supplementary materials). Although we tried to cultivate the single crystal from enantioenriched products 3, we failed to achieve this goal. So, the absolute configurations of chiral products 3 could not be determined. Nevertheless, when N-methyl-protected 3-vinylindole 1n was employed as a substrate in the reaction with sesamol-derived o-QM 2a under standard conditions (Figure 5a), the [2 + 4] cycloaddition occurred to generated product 3na in a moderate yield and diastereoselectivity (47% yield, 86:14 dr) albeit with an extremely low enantioselectivity (14% ee). Fortunately, we cultivated the single crystal of product 3na, whose relative configuration was determined to be (trans, trans) by X-ray diffraction analysis of the single crystal (CCDC 2100427, see the supplementary materials) (Figure 5b).

Theoretical Calculations of the Reaction Pathway and Key Transition States
To elucidate the reaction pathway and the interaction of CPA with the substrates, we carried out theoretical calculations on the reaction pathway of catalytic asymmetric [2 + 4] cycloaddition (see the supplementary materials) based on previous mechanistic studies [66,67]. As exemplified by the formation of product 3ma (Figure 6a), the key transition states (TSs) and the Gibbs free energy leading to the enantiomers of 3ma were determined, wherein TS-1 led to the major enantiomer (R,S,R)-3ma and TS-1′ led to the

Theoretical Calculations of the Reaction Pathway and Key Transition States
To elucidate the reaction pathway and the interaction of CPA with the substrates, we carried out theoretical calculations on the reaction pathway of catalytic asymmetric  [66,67]. As exemplified by the formation of product 3ma (Figure 6a), the key transition states (TSs) and the Gibbs free energy leading to the enantiomers of 3ma were determined, wherein TS-1 led to the major enantiomer (R,S,R)-3ma and TS-1 led to the minor enantiomer (S,R,S)-3ma. DFT calculations revealed that the two bulky 3,3 -(1-naphthyl) groups and the BINOL scaffold of CPA (R)-4c formed a pocket-like chiral environment to hold the two substrates of 1m and 2a in a confined orientation. Specifically, in TS-1, 3-vinylindole 1m was located above o-QM 2a in the chiral pocket of CPA (R)-4c, wherein the space above 2a was enough to make 1m have little steric effect on other groups. While in TS-1 , 1m was located below o-QM 2a, wherein the space below 2a was limited, thus making the phenyl group of 1m have some steric effect on the 1-naphthyl group of (R)-4c. This steric repulsion made TS-1 inferior to TS-1, which led to the formation of the major enantiomer (R,S,R)-3ma. It should be noted that the E-configuration of vinylindoles 1 has been retained as trans-configuration in products 3 due to a concerted [2 + 4] cycloaddition pathway as illustrated in TS-1. So, the diastereomeric ratio of product 3 reflects the stereoselectivity of the two adjacent chiral centers generated by the two individual substrates 1 and 2. In TS-1, CPA (R)-4c utilized its O-H group to form a strong hydrogen bond (b1 = 1.461 Å) with the C=O group of o-QM 2a, but there was no discernible hydrogen-bonding interaction between CPA (R)-4c and 3-vinylindole 1m. In addition, the calculations suggested that the [2 + 4] cycloaddition largely occurred via a concerted reaction pathway involving the formation of two new bonds (b2 = 2.526 Å, b3 = 1.970 Å). However, the longer bond length of b2 than b3 indicated that b3 (a C-C bond) formed slightly earlier than b2 (a C-O bond), which is in accordance with the reactivity of 3-vinylindole (based on the nucleophilicity of the vinyl group). In TS-1 , there were similar interactions between CPA (R)-4c and the substrates. However, the hydrogen bond (b1 = 1.520 Å) between (R)-4c and 2a in TS-1 was weaker than that in TS-1 (b1 = 1.461 Å), which resulted in a significantly higher Gibbs free energy barrier for the generation of TS-1 (24.9 kcal/mol) compared to that for TS-1 (19.7 kcal/mol). The calculated difference in the energy barriers for the two transition states of TS-1 and TS-1 of 5.2 kcal/mol explained the excellent experimentally obtained enantioselectivity of 3ma (97% ee).
Very interestingly, in the calculated transition states, there was no discernible hydrogenbonding interaction between CPA (R)-4c and 3-vinylindole 1m, which was seldom reported in CPA-catalyzed reactions involving 3-vinylindoles. To verify this issue, we performed a control experiment to investigate the role of the NH group in substrate 1m (Figure 6b). Namely, 3-vinylindole 1o, as N-methyl protected counterpart of 1m, was employed as a substrate in the [2 + 4] cycloaddition with o-QM 2a under standard conditions, which smoothly generated product 3oa in a moderate yield of 53% with a good diastereo-and enantioselectivity (91:9 dr, 83% ee). Compared to the results of product 3ma which was generated from N-unprotected 3-vinylindole 1m, the yield and the stereoselectivity of product 3oa were on a similar level, thus supporting the calculated activation mode that the NH group of 3-vinylindole 1m had no discernible hydrogen-bonding interaction with CPA (R)-4c.
It should be noted that the E-configuration of vinylindoles 1 has been retained as trans-configuration in products 3 due to a concerted [2 + 4] cycloaddition pathway as illustrated in TS-1. So, the diastereomeric ratio of product 3 reflects the stereoselectivity of the two adjacent chiral centers generated by the two individual substrates 1 and 2.

Large-Scale Synthesis of Product 3aa
Finally, the catalytic asymmetric [2 + 4] cycloaddition of 1a with 2a was carried out on a one mmol scale (Figure 7). The yield and stereoselectivity of this one-mmol-scale reaction were at the same level as those of the small-scale reaction ( Table 3, entry 1), which implied that the catalytic asymmetric [2 + 4] cycloaddition could be scaled up.

Organocatalytic Asymmetric [2 + 4] Cycloaddition of 3-Vinylindoles with o-Hydroxybenzyl Alcohols
To expand the substrate scope of this organocatalytic asymmetric [2 + 4] cycloaddition, we attempted to react 3-vinylindole 1a with o-hydroxybenzyl alcohol 5a as a precursor of o-QM (Table 5). In the presence of CPA (R)-4a (entry 1), the desired product 6aa was afforded in a moderate yield, albeit with a low stereoselectivity (51% yield, 67:33 dr, 42% ee). Then, a series of CPAs (R)-4 were screened. Among these CPAs, (R)-4e, bearing two 3,3 -(9-anthracenyl) groups, displayed the highest catalytic activity in delivering product 6aa with a better enantioselectivity than the other catalysts (entry 5 vs. entries 1-4 and 6-7), which could also be ascribed to the steric hindrance effect of the bulky 3,3 -disubstituents of CPA (R)-4e in controlling the enantioselectivity. Next, different solvents were evaluated in the presence of (R)-4e, revealing toluene to still be the most suitable solvent (entry 5 vs. entries [8][9][10][11][12]. Finally, the reaction temperature was modulated (entries [13][14][15], and the optimal reaction conditions were set as shown in entry 14. Table 5. Optimization of reaction conditions for [2 + 4] cycloaddition of 1a with 5a a . 4 and 6-7), which could also be ascribed to the steric hindrance effect of the bulky 3,3'-disubstituents of CPA (R)-4e in controlling the enantioselectivity. Next, different solvents were evaluated in the presence of (R)-4e, revealing toluene to still be the most suitable solvent (entry 5 vs. entries [8][9][10][11][12]. Finally, the reaction temperature was modulated (entries [13][14][15], and the optimal reaction conditions were set as shown in entry 14. With the optimal conditions in hand, we investigated the substrate scope of 3-vinylindole 1 in catalytic asymmetric [2 + 4] cycloaddition with the o-hydroxybenzyl alcohol 5a. As shown in Table 6, this [2 + 4] cycloaddition was amenable to participation by a wide range of 3-vinylindoles 1 bearing different R/R 1 groups. In detail, C5-, C6-and C7-substituted 3-vinylindoles participated in the [2 + 4] cycloaddition with the o-hydroxybenzyl alcohol 5a to generate the chiral indole-containing chroman derivatives 6 in moderate to good diastereo-and enantioselectivities (75:25 dr to 83:17 dr, 74-82% ee, entries 2-7). In addition, meta-and para-substituted phenyl groups were found to be suitable R 1 groups for the 3-vinylindoles 1, and the corresponding substrates participated in [2 + 4] cycloaddition with good results (entries 8-9). With the optimal conditions in hand, we investigated the substrate scope of 3-vinylindole 1 in catalytic asymmetric [2 + 4] cycloaddition with the o-hydroxybenzyl alcohol 5a. As shown in Table 6, this [2 + 4] cycloaddition was amenable to participation by a wide range of 3-vinylindoles 1 bearing different R/R 1 groups. In detail, C5-, C6-and C7-substituted 3-vinylindoles participated in the [2 + 4] cycloaddition with the o-hydroxybenzyl alcohol 5a to generate the chiral indole-containing chroman derivatives 6 in moderate to good diastereo-and enantioselectivities (75:25 dr to 83:17 dr, 74-82% ee, entries 2-7). In addition, metaand para-substituted phenyl groups were found to be suitable R 1 groups for the 3-vinylindoles 1, and the corresponding substrates participated in [2 + 4] cycloaddition with good results (entries 8-9). 72:28 79 a Unless otherwise indicated, the reaction was carried out at a 0.1 mmol scale in a solvent (0.1 mL) at 25 °C for 6 h using a 1a:5a molar ratio of 1:1.2. b Isolated total yield of the diastereomeric mixtures. c The diastereomeric ratio (dr) was determined by 1 H NMR and HPLC. d The ee value refers to that of the major diastereomer and was determined by HPLC. e At 50 °C. f At 0 °C. g At −10 °C. The asterisk * indicates chiral center. Then, the substrate scope of o-hydroxybenzyl alcohols 5 was investigated for [2 + 4] cycloaddition with 3-vinylindole 1a under standard reaction conditions. As shown in Table 7, the o-hydroxybenzyl alcohols 5b-5c bearing a methyl group or a halogen group at the C5 position successfully participated in [2 + 4] cycloaddition with 3-vinylindole 1a, providing products 6ab-6ac in moderate to good diastereo-and enantioselectivities (68:32 dr to 81:19 dr, 73-76% ee, entries 2-3). In addition, aromatic R 1 groups with ortho, meta and para-substituents were successfully employed in the reaction, affording products 6ad-6af in overall good enantioselectivities (76-81% ee, entries 4-6). Then, the substrate scope of o-hydroxybenzyl alcohols 5 was investigated for [2 + 4] cycloaddition with 3-vinylindole 1a under standard reaction conditions. As shown in Table 7, the o-hydroxybenzyl alcohols 5b-5c bearing a methyl group or a halogen group at the C5 position successfully participated in [2 + 4] cycloaddition with 3-vinylindole 1a, providing products 6ab-6ac in moderate to good diastereo-and enantioselectivities (68:32 dr to 81:19 dr, 73-76% ee, entries 2-3). In addition, aromatic R 1 groups with ortho, meta and para-substituents were successfully employed in the reaction, affording products 6ad-6af in overall good enantioselectivities (76-81% ee, entries 4-6).  The structures of all products 6 were identified by their NMR, IR and HR MS data, and the ee value of all products 6 were calculated by their HPLC traces (see the supplementary materials). The relative configuration of product 6ma was determined to be (trans, cis) by a NOE experiment (see the supplementary materials) (Figure 8) and comparing the 1 H NMR spectra with that of a similar compound [18]. The structures of all products 6 were identified by their NMR, IR and HR MS data, and the ee value of all products 6 were calculated by their HPLC traces (see the Supplementary Materials). The relative configuration of product 6ma was determined to be (trans, cis) by a NOE experiment (see the Supplementary Materials) (Figure 8) and comparing the 1 H NMR spectra with that of a similar compound [18]. diastereomeric ratio (dr) was determined by 1 H NMR. d The ee value refers to that of the major d astereomer and was determined by HPLC. e At 0 °C for 6 h and then 25 °C for 2 h. f The ee value o the minor diastereoisomer. g At 25 °C for 2 h. The asterisk * indicates chiral center.
The structures of all products 6 were identified by their NMR, IR and HR MS data and the ee value of all products 6 were calculated by their HPLC traces (see the supple mentary materials). The relative configuration of product 6ma was determined to b (trans, cis) by a NOE experiment (see the supplementary materials) (Figure 8) and com paring the 1 H NMR spectra with that of a similar compound [18].

Materials and Methods
The detailed procedures for the synthesis and characterization of the products ar given in Appendix A section.

Conclusions
In summary, we performed catalytic asymmetric [2 + 4] cycloaddition o 3-vinylindoles with ortho-quinone methides and their precursors in the presence of chira phosphoric acid. This approach was used to synthesize a series of indole-containin chroman derivatives with structural diversity in overall high yields (up to 98%), goo diastereoselectivities (up to 93:7 dr) and moderate to excellent enantioselectivities (up t 98% ee). This approach not only enriches the chemistry of 3-vinylindole-inolved catalyt asymmetric cycloadditions but is also useful for the enantioselective synthesis of chira chroman derivatives.

Supplementary Materials:
The following are available online. NMR and HPLC spectra of produc 3 and 6, NOE spectrum of product 6ma, X-ray single-crystal data for product 3na, and theoretic calculations of the reaction pathway.

Materials and Methods
The detailed procedures for the synthesis and characterization of the products are given in Appendix A section.

Conclusions
In summary, we performed catalytic asymmetric [2 + 4] cycloaddition of 3-vinylindoles with ortho-quinone methides and their precursors in the presence of chiral phosphoric acid. This approach was used to synthesize a series of indole-containing chroman derivatives with structural diversity in overall high yields (up to 98%), good diastereoselectivities (up to 93:7 dr) and moderate to excellent enantioselectivities (up to 98% ee). This approach not only enriches the chemistry of 3-vinylindole-inolved catalytic asymmetric cycloadditions but is also useful for the enantioselective synthesis of chiral chroman derivatives.
Supplementary Materials: The following are available online. NMR and HPLC spectra of products 3 and 6, NOE spectrum of product 6ma, X-ray single-crystal data for product 3na, and theoretical calculations of the reaction pathway.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available in this article.

Conflicts of Interest:
The authors declare no conflict of interest.
Sample Availability: Samples of the compounds are not available from the authors.

Appendix A Experimental Section
1 H and 13 C NMR spectra were measured at 400 and 100 MHz, respectively. The solvents used for NMR spectroscopy were acetone-d 6 and CDCl 3 , using tetramethylsilane as the internal reference. HR MS (ESI) was determined by an HR MS/MS instrument. The X-ray source used for the single crystal X-ray diffraction analysis of compound 3na was MoKα (λ = 0.71073), and the thermal ellipsoid was drawn at the 30% probability level. Analytical grade solvents for the column chromatography were used after distillation, and commercially available reagents were used as received. Substrates 1 were synthesized according to the literature method [17]. Substrates 2 and 5 were synthesized according to the literature method [52,68].

General procedure for the synthesis of products 6
To the mixture of 3-vinylindoles 1 (0.1 mmol), ortho-hydroxybenzyl alcohols 5 (0.12 mmol), catalyst 4e (0.01 mmol) were added toluene (1 mL). Then, the reaction mixture was stirred at 0 • C for 6 h. After completion of the reaction, which was indicated by TLC, the reaction mixture was directly purified through flash column chromatography to afford products 6.