Catalytic Asymmetric Formal [3+2] Cycloaddition of Azoalkenes with 3-Vinylindoles: Synthesis of 2,3-Dihydropyrroles.

Chiral phosphoric acid-catalyzed highly enantioselective formal [3 + 2] cycloaddition reaction of azoalkenes with 3-vinylindoles has been established. Under mild conditions, the projected cycloaddition proceeded smoothly, affording a variety of 2,3-dihydropyrroles in high yields and excellent enantioselectivities, and also in a diastereospecific manner. As opposed to the common 4-atom synthons in the previous literature reports, azoalkenes served as 3-atom synthons. Besides, the observed selectivity was supported by primary theoretical calculation. The unique chemistry of azoalkenes disclosed herein will empower asymmetric synthesis of nitrogen-containing ring structural motifs in a broader context.

Introduction

1,3-Dipolar cycloadditions are well-established synthetic strategies in organic chemistry for the preparation of five-membered heterocyclic ring systems (Coldham and Hufton, 2005, Fang and Wang, 2018, Gothelf and Jørgensen, 1998, Hashimoto and Maruoka, 2015, Kissane and Maguire, 2010, Stanley and Sibi, 2008). In a typical normal-electron-demand 1,3-dipolar cycloaddition, nucleophilic 1,3-dipoles and electron-deficient dipolarophiles are utilized. Asymmetric versions of such cycloadditions often rely on synthetic strategy that lowers the lowest unoccupied molecular orbital (LUMO) of dipolarophiles (Cheng et al., 2019, Hashimoto et al., 2007, Kano et al., 2005, Liu et al., 2008, Pascual-Escudero et al., 2016, Sibi et al., 2004, Tong et al., 2013, Wang et al., 2015, Xu et al., 2018, Yang et al., 2017). In stark contrast, inverse-electron-demand 1,3-dipolar cycloadditions utilizing electrophilic 1,3-dipoles and nucleophilic dipolarophiles are much less common. Among the reported catalytic asymmetric inverse-electron-demand 1,3-dipolar cycloadditions, nitrones and vinyl ethers are commonly employed reaction partners (Figure 1) (Ashizawa et al., 2006, Bayón et al., 2000a, Bayón et al., 2000b, Hashimoto et al., 2011, Hori et al., 1998, Jensen et al., 1999, Jensen et al., 2000, Jiao et al., 2008, Mikami et al., 2001, Seerden et al., 1994, Seerden et al., 1995, Seerden et al., 1997, Simonsen et al., 1999a, Simonsen et al., 1999b, Suga et al., 2007, Suga et al., 2010, Yanagisawa et al., 2011) To date, there are only a handful of exceptions (Bartlett et al., 2017, Liu et al., 2016, Sohtome et al., 2017, Xu et al., 2015, Zhu et al., 2014). Sodeoka et al.’ employment of nitrile oxides as electrophilic 1,3-dipoles and Feng's utilization of enecarbamates as nucleophilic dipolarophiles are interesting examples, among others. To design inverse-electron-demand 1,3-dipolar cycloaddition processes, we recognized the importance of introducing alternative electrophilic 1,3-dipole surrogates, which ideally could be easily combined with various dipolarophiles, thus allowing for ready creation of useful molecular architectures.

Figure 1

Inverse-Electron-Demand 1,3-Dipolar Cycloadditions Utilizing Nitrones and Vinyl Ethers

Azoalkenes, also known as 1,2-diaza-1,3-dienes, have proven to be versatile synthetic building blocks in organic chemistry (Attanasi et al., 2002a, Attanasi et al., 2002, Attanasi et al., 2009, Attanasi and Filippone, 1997, Lopes et al., 2018; Wei et al., 2019). Their characteristic 1,3-conjugate systems have been utilized synthetically; azoalkenes were shown to be a valuable acceptor in 1,4-conjugate additions, displaying excellent reactivity toward a wide variety of nucleophiles (Attanasi et al., 2011a, Attanasi et al., 2011b, Attanasi et al., 2012, Attanasi et al., 2013a, Attanasi et al., 2013b, Ciccolini et al., 2019, Mantenuto et al., 2015, Miles et al., 2015, Preti et al., 2010). Another attractive synthetic application of azoalkenes is the cycloaddition reaction, which serves as a powerful strategy for the construction of nitrogen-containing heterocycles. In the currently available mode of cycloaddition, Wang and others employed azoalkenes as 4-atom (A4) synthons (Int-I), making use of C4 electrophilicity and N1 nucleophilicity of azoalkenes for various asymmetric formal [4 + n] cycloaddition processes (Chen et al., 2012, Gao et al., 2013, Huang et al., 2016, Tong et al., 2014, Wei et al., 2017, Wei et al., 2018, Wei and Wang, 2015, Zhang et al., 2018, Zhang and Song, 2018). Very recently, our group discovered an azoalkene-enabled enantioselective dearomatization of indoles (Mei et al., 2020). Given the ubiquitous existence of nitrogen-containing cyclic structural motifs, we questioned whether it might be possible to utilize azoalkenes as a carbon-carbon-nitrogen (CCN) 1,3-diplole surrogate, a 3-atom (A3) synthon (Int-II) in asymmetric formal [3 + 2] cycloaddition reactions, and thereby to access a broad range of nitrogen-containing ring systems (Attanasi et al., 2002a, Attanasi et al., 2002, Attanasi et al., 2005, Attanasi et al., 2013a, Attanasi et al., 2013b, Clarke et al., 1983, Karapetyan et al., 2008, Mari et al., 2017, Ran et al., 2017, Sommer, 1979). We reasoned the hydrazine-enamine tautomerization could play a key role, and fine-tuning of the system and judicious selection of potential reaction partners are of crucial importance to the successful implementation of synthetic plans (Figure 2). In particular, we believe that the current under-developed status of inverse-electron-demand 1,3-dipolar cycloadditions, in combination of rich chemistry of azoalkenes and anticipated broad applicability of the methodology, make the proposed strategy highly attractive and worthwhile investigating.

Figure 2

Employment of Azoalkenes As a Reaction Partner in Enantioselective Formal Cycloaddition Reactions

2,3-Dihydropyrroles are common structural motifs that are widely present in biologically significant molecules, and they are also valuable intermediates in organic synthesis (Augeri et al., 2005, Cantín et al., 1999, Hertel and Xu, 2002, Herzon and Myers, 2005, Kawase et al., 1999, Marti and Carreira, 2005, Petersen and Nielsen, 2013). Although approaches to access racemic 2,3-dihydropyrroles are well documented (El-Sepelgy et al., 2018, Jiang et al., 2017, Liang et al., 2017, Liang et al., 2018, Ma et al., 2018, Zhu et al., 2009, Zhu et al., 2011), reports on catalytic asymmetric synthesis of 2,3-dihydropyrroles are scarce. In an early example, Gong et al. documented a catalytic asymmetric formal [3 + 2] cycloaddition reaction between isocyanoesters and nitroolefins for the synthesis of optically enriched 2,3-dihydropyrroles (Guo et al., 2008). More recently, Miura and Murakami, as well as the Fokin group, reported enantioselective preparation of 2,3-dihydropyrroles via RhII-catalyzed asymmetric annulations of triazoles with alkenes (Kwok et al., 2014, Miura et al., 2013). As part of our continuous interests in developing enantioselective cycloaddition reactions for the preparation various heterocyclic ring systems (Chan et al., 2019, Han et al., 2014, Han et al., 2016, Li et al., 2019, Ni et al., 2017, Yao et al., 2016, Wu et al., 2019), we questioned the feasibility of establishing an effective asymmetric synthesis of 2,3-dihydropyrroles via a formal [3 + 2] cycloaddition reaction, by utilizing azoalkenes as an electrophilic CCN 1,3-dipole surrogate and employing simple 3-vinylindoles (Gioia et al., 2008, Li et al., 2018, Sun et al., 2016, Tan et al., 2011, Yang et al., 2019, Zhang et al., 2018, Zheng et al., 2015) as a C2 reaction partner (Figure 3). In this report, we document a formal [3 + 2] cycloaddition reaction for enantioselective creation of 2,3-dihydropyrroles under the catalysis of chiral phosphoric acid (CPA) (Akiyama, 2007, Terada, 2008, Terada, 2010, Wu et al., 2015, Yu et al., 2011). The projected progress could be identified as a formal inverse-electron-demand 1,3-dipolar cycloaddition reaction, wherein azoalkene served as a CCN 1,3-dipole surrogate, a 3-atom synthon.

Figure 3

Our Hypothesis: Construction of 2,3-Dihydropyrroles from Azoalkenes and Simple Alkenes

Results and Discussion

Reaction Development

Our investigation was initiated by identifying optimal conditions for the model reaction between azoalkene 1a and vinylindole 2a (Table 1). TRIP-CPA 4a effectively catalyzed the reaction, furnishing 2,3-dihydropyrrole 3a in excellent yield and moderate stereoselectivities (entry 1). The solvent screening was subsequently carried out, and chloroform was found to be the best solvent (entries 1–5). Next, the catalytic effects of different CPA catalysts (4b−f) were examined. Catalysts 4b and 4e had excellent controls on diastereoselectivities, but enantiomeric controls were less ideal (entries 6 and 9). Although 4c was less effective (entry 7), 4d was completely ineffective (entry 8). To our delight, the SiPh3-derived CPA 4f was found to be an excellent catalyst, leading to the formation of 2,3-dihydropyrrole 3a in excellent yield and excellent enantioselectivity and diastereoselectivity (entry 10). Lowering the reaction temperature or adding molecular sieves did not result in enhancement (entries 11 and 12). Under the optimized reaction conditions, the desired 2,3-dihydropyrrole 3a was obtained in 96% yield, and with 94% ee and >20:1 dr.

Table 1

Optimization of the Reaction Conditions

Entry	4	Solvent	Yield (%)a	ee (%)b	Drc
1	4a	CH2Cl2	95	70	6:1
2	4a	Toluene	90	53	8:1
3	4a	THF	<5	–	–
4	4a	DCE	85	70	7:1
5	4a	CHCl3	86	72	11:1
6	4b	CHCl3	80	28	>20:1
7	4c	CHCl3	92	62	7:1
8	4d	CHCl3	95	0	2:1
9	4e	CHCl3	88	54	>20:1
10	4f	CHCl3	96	94	>20:1
11d	4f	CHCl3	94	92	>20:1
12e	4f	CHCl3	92	91	>20:1

Reaction conditions: 1a (0.1 mmol), 2a (0.12 mmol), and the catalyst (0.001 mmol) in the solvent specified (1 mL) at room temperature for 0.5 h.

Yields refer to isolated yields.

The ee values were determined by HPLC analysis on a chiral stationary phase.

The dr values were determined by 1H NMR analysis of the crude mixture.

The reaction was carried out at 0°C.

Molecular sieves (4 Å) were added.

Scope

With the optimal reaction conditions in hand, the substrate scope with regard to azoalkenes was evaluated (Table 2). Azoalkenes bearing different R1 groups such as methyl (1a), ethyl (1b), and n-propyl (1c) were well tolerated. When azoalkenes containing different C=C double bond appended ester groups, e.g., CO2Et (1a), CO2Me (1d), CO2Bu (1e), CO2Bn (1f), and CO2Pr (1g) were utilized, consistently high chemical yields and enantio- and diastereoselectivities were attainable.

Table 2

Employing Different Azoalkenes

Entry	R1/R2(1)	3	Yield (%)a	ee (%)b	Drc
1	Me/Et(1a)	3a	96	94	>20:1
2	Et/Et(1b)	3b	90	95	>20:1
3	nPr/Et(1c)	3c	94	83	>20:1
4	Me/Me(1d)	3d	85	94	>20:1
5	Me/tBu(1e)	3e	86	92	>20:1
6	Me/Bn(1f)	3f	95	91	>20:1
7	Me/iPr(1g)	3g	92	91	>20:1

Reaction conditions: 1 (0.1 mmol), 2a (0.12 mmol), and 4f (0.001 mmol) in CHCl3 (1 mL) at room temperature for 0.5 h.

Yields refer to isolated yields.

The ee values were determined by HPLC analysis on a chiral stationary phase.

The dr values were determined by 1H NMR analysis of the crude mixture.

The reaction scope with regard to vinylindoles was subsequently investigated (Figure 4). Different substituted aryl groups could be installed at the terminal position of vinylindoles, regardless of electronic nature and substitution pattern (3h−3o). Moreover, vinylindoles bearing a dichloro-substituted phenyl ring (3p), a 2-naphthalenyl (3q), or a 2-thiophenyl substituent (3r) were found suitable for the reaction. In all the examples examined, the desired 2,3-dihydropyrrole products were obtained in high yields and with excellent ee values and all the reactions proceeded in a diastereospecific manner.

Figure 4

Reaction Scope of Vinylindoles

Reaction conditions: 1 (0.1 mmol), 2 (0.12 mmol), and 4f (0.001mmol) in CHCl3 (1 mL) at room temperature for 0.5 h. Yields refer to isolated yields; the ee values were determined by HPLC analysis on a chiral stationary phase.

The indole moieties in the vinylindole structures could also be varied, and the results are summarized in Figure 5. A wide range of vinylindoles bearing various substituted indoles were employed, and the corresponding 2,3-dihydropyrrole products 3 (3s−3z, 3a’−3e′) were obtained in good to very good yields, and with consistently excellent enantioselectivities, as well as perfect diastereoselectivities. Notably, the electronic nature and position of the indole substituents did not appear to have much influence on the reaction, and this trend held true for 5,6-dichloro-substituted substrate (3e′). The absolute configurations of 2,3-dihydropyrrole products were assigned based on X-ray crystallographic analysis of 3y.

Figure 5

Further Reaction Scope of vinylindoles

Reaction conditions: 1a (0.1 mmol), 2 (0.12 mmol), and 4f (0.001 mmol) in CHCl3 (1 mL) at room temperature for 0.5 h. Yields refer to isolated yields; the ee values were determined by HPLC analysis on a chiral stationary phase. The absolute configurations of the annulation products were assigned based on X-ray crystallographic analysis of 3y (CCDC 1957145).

Mechanistic Investigations

We carried out a few further experiments to gain a better understanding of this reaction process. When methyl-substituted vinylindole 2b was employed, only a moderate ee value of 64% was obtained (Scheme 1 Equation 1), which suggested the importance of aryl moiety in vinylindole substrates for asymmetric induction. When 2-methyl-substituted vinylindole 2c was utilized, a dearomatization of indole occurred, furnishing the pyrroloindoline product 5 in good yield and excellent enantioselectivity (Scheme 1 Equation 2). It is intriguing to note that such subtle difference in substrate structure could result in totally different chemoselectivity. The presence of a 2-methyl group may render indole higher nucleophilicity at the C3-position, thus favoring the dearomative process (Mei et al., 2020). Furthermore, no reaction was observed when N-methyl vinylindole 2d was employed (Scheme 1 Equation 3), indicating the indispensability of hydrogen bonding interactions between CPA and the substrates, not only in asymmetric induction but also in reaction activation.

Scheme 1

Control Experiments

On the basis of our experimental results, we have also constructed the models with the aid of computation using 1a and 2a as substrates to obtain some insights into the reaction selectivity. A plausible reaction pathway was proposed (Figure 6) where the asymmetric 1,4-addition of vinylindole 2a to azoalkene 1a is initiated via a hydrogen-bonding activation mode and both substrates can be activated simultaneously by the CPA catalyst within its chiral pockets. Vinylindole 2a adopts the s-cis geometry in order to reach the electrophilic site in 1a. The resulting intermediate A has its conformation locked for the 5-exo attack [3 + 2] of the iminic N to the spatially adjacent C=C bond (N1-C 3.98 Å, path a) to afford the experimentally observed product after proton transfer and tautomerization steps. Alternatively, hydrazone-enamine tautomerization may occur, furnishing intermediate B, which undergoes cyclization to afford the observed [3 + 2] product 3a via path c. In comparison, the 6-exo [4 + 2] attack was deemed difficult to occur. In intermediate A, path b is unfavorable, likely due to the ring strain under such a rigid structure (N2-C 5.32 Å), whereas in intermediate B, pathway d is unlikely because of the reduced nucleophilicity of the amide nitrogen. Indeed, the [4 + 2] products formed via path b or d were never observed in this reaction.

Figure 6

A Plausible Reaction Mechanism Accounting for the Selectivity

Conclusions

In conclusion, we have established a formal [3 + 2] cycloaddition reaction, utilizing azoalkenes as an electrophilic reaction component and simple alkenes as a nucleophilic partner. In the presence of chiral phosphoric acid, the reaction proceeded smoothly, furnishing a wide range of functionalized 2,3-dihydropyrroles in good yields and in a highly enantioselective and diastereospecific manner. It is noteworthy that the projected progress could be identified as a formal inverse-electron-demand 1,3-dipolar cycloaddition reaction, wherein azoalkenes served as CCN 1,3-dipole surrogates, 3-atom synthons, as opposed to the common 4-atom synthons in the previous literature reports. With the current successful demonstration of chiral 2,3-dihydropyrrole synthesis and theoretical understanding of the observed chemoselectivity, we anticipate the unique chemistry of azoalkenes disclosed herein will empower asymmetric synthesis of nitrogen-containing ring structural motifs in a broader context. Our findings in this direction will be reported in due course.

Limitations of the Study

A brief examination showed that the present method is not compatible with N-methyl-substituted vinylindole and 2-methyl-substituted vinylindole for the construction of corresponding 2,3-dihydropyrroles.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.