Synthesis of Non-Racemic Pyrazolines and Pyrazolidines by [3+2] Cycloadditions of Azomethine Imines

Abstract

Asymmetric [3+2] cycloadditions of azomethine imines comprise a useful synthetic tool for the construction of pyrazole derivatives with a variable degree of saturation and up to three stereogenic centers. As analogues of pyrrolidines and imidazolidines that are abundant among natural products, pyrazoline and pyrazolidine derivatives represent attractive synthetic targets due to their extensive applications in the chemical and medicinal industries. Following the increased understanding of the mechanistic aspect of metal-catalyzed and organocatalyzed [3+2] cycloadditions of 1,3-dipoles gained over recent years, significant strides have been taken to design and develop new protocols that proceed efficiently under mild synthetic conditions and duly benefit from superior functional group tolerance and selectivity. In this review, we represent the current state of the art in this field and detailed methods for the synthesis of non-racemic pyrazolines and pyrazolidines via [3+2] metal and organocatalyzed transformations reported since the seminal work of Kobayashi et al. and Fu et al. in 2002 and 2003 up to the end of year 2017.

1. Introduction

[3+2] Cycloadditions comprise a powerful tool for the construction of five-membered heterocycles [1,2,3,4]. Due to concertedness of the reaction mechanism [4,5,6], [3+2] cycloadditions are also the most straightforward way to obtain a (partially) saturated system with multiple stereogenic centers in a highly stereoselective manner [1,2,3,4,5,6]. The advent of asymmetric catalysis, which also took place in the field of [3+2] cycloaddition chemistry, nowadays enables the preparation of various types of non-racemic saturated heterocycles in a highly efficient manner. Within this context, the use of nitrile oxides, nitrones and azomethine ylides is well established [7,8], whereas asymmetric cycloadditions of azomethine imines remain much less explored [8,9,10,11]. To date, less than three dozen reports on the synthesis of non-racemic pyrazolines and pyrazolidines by [3+2] cycloaddition of azomethine imines can be found in the literature. However, the fact that all these examples have been published since 2003 points at increasing interest in this field.

Azomethine imines are 1,3-dipoles of the aza-allyl type represented with iminium amide and diazonium ylide mesomeric structures [9,10,11,12]. They are easily accessible from hydrazine derivatives, diazoalkanes and azo-compounds. Although many azomethine imines are stable compounds, also the unstable dipoles are easily generated in situ from the above precursors. [3+2] Cycloadditions of azomethine imines are of broad substrate scope as the reactions can be performed with all common types of dipolarophiles. Accordingly, reactions with acetylenes provide access to pyrazolines, whereas cycloadditions to olefins are employed to obtain the fully-saturated pyrazolidines. In terms of stereochemistry, [3+2] cycloadditions of azomethine imines furnish products with up to three newly-formed stereogenic centers [9,10,11] (Figure 1). In addition, the pyrazolidine derivatives obtained through these annulations can readily be transformed into differently-protected 1,3-diamine derivatives via N–N bond cleavage with SmI₂.

Shintani and Fu have published the first example of asymmetric [3+2] cycloadditions of azomethine imines to acetylenes in 2003. They utilized Cu(I)-based catalyst with chiral phosphaferrocene-oxazoline ligand to obtain the desired cycloadducts regioselectively and in excellent enantioselectivity [13]. Soon after, the same catalyst was also employed in kinetic resolution of azomethine imines by [3+2] cycloaddition [14]. Besides Cu(I)-, also Ag(I)- and Cu(II)-catalysts are amenable in cycloadditions to acetylenes [11,15]. In the reactions of olefins, however, asymmetric organocatalysts comprise the leading approach, although examples of metal-catalysis have also been reported [8,11]. Most probably, the combination of the proven viability of different catalytic systems, as well as the applicability of the pyrazoline scaffold in the development of drugs, agrochemicals, and materials (Figure 2) will continue to spur increased interest in these reactions.

This review covers the synthesis of non-racemic pyrazolines and pyrazolidines by [3+2] cycloadditions of azomethine imines until September 2017. The vast majority of examples are asymmetric catalyzed cycloadditions; however, the preparation of non-racemic cycloadducts by resolution of stereoisomers is also included.

2. Synthesis of Pyrazolines by Metal-Catalyzed Cycloadditions

Following the success of copper-catalyzed 1,3-dipolar cycloadditions of alkynes to azides [16,17], and to nitrones [18,19], in 2003, Fu and co-workers extended this useful mode of reactivity to N,N-cyclic azomethine imines as dipoles. They demonstrated that the reaction of 3-oxopyrazolidin-ium-2-ides 1 with terminal alkynes dramatically accelerated at room temperature in the presence of catalytic amount of CuI, presumably via the formation of copper acetylide as the more reactive dipolarophile, and gave the corresponding pyrazolo[1,2-a]pyrazole products 2 as single regioisomers (Scheme 1) [13]. To be able to control the stereoselectivity of the reaction, they utilized enantiomerically-pure oxazolines as privileged chiral ligands, since bidentate phosphines inhibited the catalysis by copper. While the bis(oxazoline) ligand L1 induced only modest stereoselection in the reaction of azomethine imine 1 (R = Ph) with ethyl propiolate, the level of enantioselectivity was considerably increased in the presence of P,N phosphaferrocene-oxazoline chiral ligand L2a bearing an isopropyl substituent on the oxazoline moiety (19% ee → 90% ee). Introduction of the sterically more demanding tert-butyl group (ligand L2b) or changing the planar chirality of the phosphaferrocene subunit as in ligand L3 resulted in a significant decrease of enantiomeric excess (90% ee → 58% ee and 80% ee, respectively). This was the first asymmetric Cu(I)-catalyzed [3+2] cycloaddition of azomethine imines with alkynes.

Their methodology was applicable to azomethine imines derived from aryl-, alkyl- and alkenyl-aldehydes, and thus, they prepared a series of the corresponding enantioenriched pyrazoline derivatives 2 with high chemical yields and ee values up to 98%. With regard to the dipolarophile coupling partner, the best yields and ee values were obtained with electron-poor alkynes having a carbonyl, an electron-deficient aromatic or heteroaromatic group. Simple aryl- or alkyl-acetylenes also reacted in the presence of the Cu(I)-phosphaferrocene-oxazoline L2a catalytic system, although they required gentle heating for a reasonable reaction rate, which consequently resulted in erosion of regioselectivity (~6:1). This protocol tolerates also substituents in the pyrazolidinone ring of an azomethine imine yielding sterically more congested bicyclic products 3 and 4 with excellent yields and enantioselectivities.

Later on, the same research group extended the above-mentioned asymmetric Cu(I)-catalyzed cycloaddition for the efficient preparation of enantiomerically-enriched C5-substituted azomethine imines 5 through kinetic resolution of their racemic mixtures (Scheme 2) [14]. The catalytic reaction of racemic azomethine imines 5 with ethyl propiolate in the presence of the previously used chiral phosphaferrocene-oxazoline ligand L2a was less efficient in kinetic resolution than the reaction performed with the catalytic system comprising CuI and ligand L4 bearing a bigger phenyl group on the oxazoline moiety. The employment of ligand L4 furnished a two-times higher selectivity factor s (s = rate of fast-reacting enantiomer/rate of slow-reacting enantiomer) than of ligand L2a (49 vs. 27) and allowed the use of even lower catalyst loading (1 mol % vs. 5 mol %). While C4-substituted (e.g., R = cyclohexyl) azomethine imine provided very low selectivity (s < 2) in kinetic resolution, a variety of C5-substituted dipoles can be effectively resolved by Cu/L4-catalyzed cycloadditions with ester- or amide-substituted alkyne. With this new catalyst system, they succeeded in preparing highly enanatio-enriched azomethine imines (ee’s up to 99%), which have a benzylidene, a heteroaryl methylidene, an alkenylidene or a cycloalkylidene substituent at the N1 of the dipole 5. Enantiomerically-pure azomethine imines 5 prepared in these kinetic resolutions served as precursors for the synthesis of useful pyrazolidinone derivatives 6, 7 and 8.

Arai et al. developed a chiral (S,S,S,S)-bis(imidazolidine)pyridine ligand L5 (Py-Bidine) and used it in the catalytic asymmetric cycloaddition of cyclic aryl-azomethine imines 9 with alkyl propiolates (Scheme 3) [20]. The best catalytic system, in terms of activity and enantioselectivity, turned out to be the combination of L5 and CuOAc as a source of Cu⁺ ions, which provided quantitative yield and 60% ee of the corresponding (R)-cycloadduct 10 (Ar = Ph, R = Et, R¹ = H) in the reaction of 2-benzylidene-5-oxopyrazolidin-2-ium-1-ide with ethyl propiolate at room temperature. Lowering the reaction temperature (r.t. → –20 °C) and the addition of base N,N-diisopropylethylamine (DIPEA) significantly improved enantioselectivity in CH₂Cl₂ (60% ee → 74% ee). Examination of different solvents, from nonpolar to polar protic and aprotic (PhMe, EtOH, THF, 1,4-dioxane, MeCN, CH₂Cl₂), revealed only a slight solvent effect on the reaction with the exception of MeCN, in which a low chemical yield of the product was obtained. Interestingly, while the use of Cu(I) salts is believed to be crucial for the formation of copper acetylide as a reactive dipolarophile, Cu(II) salts also exhibited significant activity in the presence of ligand L5. The authors also proposed a mode of action of the catalyst Cu(I)-L5 responsible for assuring an asymmetric [3+2] cycloaddition reaction. The in situ generated copper acetylide first coordinates to the (S,S,S,S)-Py-Bidine ligand to which (Z)-azomethine imine 9 approaches from an upper-side. Because the enantio-faces of the azomethine imine are differentiated on the molecular surface of the (S,S,S,S)-(Py-Bidine)–Cu complex, a suitable combination of catalyst and substrate is necessary for a high efficiency. Consequently, the azomethine imine 9 approaches in such a way as to minimize the non-bonded repulsion between the phenyl ring of the Py-Bidine and the aromatic ring of the azomethine imine (model A) resulting in bicyclic product 10 with R configuration. On the other hand, as presented in model B, steric repulsion between phenyl groups prevents the formation of the opposite (S)-enantiomer. Although the enantiomeric excesses of the bicyclic pyrazolo[1,2-a]pyrazolone products 10 were moderate (32–74%), this newly-designed ligand L5 has high potential for further development of chiral ligands to promote asymmetric [3+2] cycloaddition reactions.

To further explore the versatility of tridentate chiral pyridine ligands in asymmetric catalysis, the five-membered imidazolidine rings in L5 were replaced by differently substituted oxazolidines, which can easily be constructed in high optical purity from L-amino acid-derived alcohols 11 and symmetric pyridine(dicarboxaldehyde) [21]. The circular dichroism-high throughput screening approach was utilized to find the most efficient catalyst prepared from the in situ generated chiral polymer-supported bis(oxazolidine)pyridine ligands (Py-Bodine) and copper salts (Scheme 4).

Using this custom-made approach, the best candidate in the reaction of azomethine imine 12 (R¹ = Ph) with ethyl propiolate was established as Py-Bodine(Ala)–Cu(OAc)₂ furnishing the (R)-isomer of the cycloadduct 13 with high chemical yield (99%) and excellent enantioselectivity (94% ee). Interestingly, the Py-Bodine ligand prepared from proline-derived alcohol gave the (S)-isomer, but with significantly lower enantioselectivity (76% ee) for the same reaction. The complex of Py-Bodine(Ala) ligand L6 and Cu(OAc)₂ effectively catalyzed the reaction of various azomethine imines 12 bearing aryl or alkyl R¹ groups with alkyl propiolates providing the corresponding bicyclic pyrazoline products 13 in a highly enantioselective manner (ee’s up to 98%) and with nearly quantitative chemical yields in most cases (Scheme 5).

Multicomponent 1,3-dipolar cycloaddition is an attractive strategy for the construction of pyrazolines, particularly where less stable acyclic azomethine imines are used as dipoles. In this context, Marouka et al. reported on the one-pot three-component asymmetric [3+2] cycloaddition reaction of aldehydes 14, hydrazides 15 and terminal acetylenes 16 by employing the catalytic system composed of CuOAc, pyridine-bisoxazoline (R,R)-L7 and binaphthyl dicarboxylic acid (R)-L8 as chiral ligands (Scheme 6) [22]. While the ligand L7 is proposed to coordinate to the copper center, the chiral Brønsted acid L8 could interact with the azomethine imine via hydrogen bonding [23]. The labile acyclic azomethine imines generated in situ from aldehyde 14 and hydrazide 15 effectively coupled with the present acetylenes 16 in a highly regioselective and stereoselective manner. The corresponding 3,4-disubstituted pyrazolines 17 formed under mild conditions (r.t. or 40 °C), although for high yields, extended reaction times had to be applied (1–3 days). Notably, the use of axially-chiral dicarboxylic acid as a co-catalyst was not only beneficial to enantioselectivity, but also increased chemoselectivity by suppressing the formation of alkynylation by-product 18. The most effective acid was found to be binaphthyl dicarboxylic acid (R)-L8 bearing 3,3′-disilyl groups. The methodology was compatible with aliphatic cyclic and acyclic, aromatic and heteroaromatic aldehydes. Moreover, aromatic and aliphatic acetylenes can be used as coupling partners, and various substituents on the hydrazide moiety are tolerated. It was also found that chiral pair matching of ligands L7 and L8 is crucial for the efficient asymmetric induction since the combination of (S,S)-L7 and (R)-L8 led to a significant drop of the ee value (from 99% to 55%) of the pyrazoline product 17 prepared from N′-benzylbenzohydrazide, benzaldehyde and phenylacetylene.

Cu(I)-catalyzed 1,3-cycloaddition of azomethine imines with alkynes is believed to proceed via Cu-acetylide intermediate and is therefore restricted to terminal alkynes (Scheme 7, a). On the other hand, Lewis acid-catalyzed cycloaddition can be applied also to internal alkynes, which importantly extends the use of 1,3-cycloaddition methodology by allowing the synthesis of densely-substituted pyrazoline derivatives (Scheme 7, b).

In this context, the groups of Sakakura and Ishihara reported on the asymmetric [3+2] cycloaddition of N,N-cyclic azomethine imines 19 with propioloylpyrazoles 20 as dipolarophiles (Scheme 8) [24]. The reaction was catalyzed by the Cu(II)-complex of 3-(2-naphthyl)-L-alanine amide derivative L9 as chiral ligand, which enabled the preparation of (R)-cycloadduct products 22 in good-to-excellent isolated yields (70–98%) and high enantioselectivities (80–95% ee). Among Cu(II) salts tested, Cu(NTf₂)₂ exhibited the highest reactivity of the chiral catalytic system. This methodology tolerates meta- and para-substituted phenyl groups, heteroaryl and alkenyl groups on the azomethine imine coupling partner 19. Besides terminal pyrazole-containing alkyne 20 (R² = H), β-substituted propioloylpyrazoles as disubstituted acetylenes also effectively participated in the present cycloaddition to give fully-substituted pyrazoline derivatives 22. The activity and enantioselectivity of this novel π–cation catalyst can be finely tuned by changing the N-alkyl R group on the ligand L9. The synthesized cycloadducts 22 served as optically-active precursors for the synthesis of 1,3-diamines with three contiguous chiral centers, such as in 23.

On the basis of experimental observations and previous results [25], a transition state assembly for the asymmetric cycloaddition of 19 and 20 was proposed (Scheme 9). Since the carbonyl Re face of a coordinated alkyne is shielded by a naphthyl group through the intramolecular π–cation interaction, an azomethine imine approaches the Si face via more favored endo-TS by avoiding steric repulsion between the R group (e.g., cyclopentyl) of the amide ligand L9 and the R¹ group (e.g., Ph) of the azomethine imine 19 to give the corresponding (R)-product 22 as the major enantiomer. Less hindered N-cycloalkyl groups are successful in avoiding steric repulsion with the ethylene group of 19. Thus, the most efficient R groups are cyclopropyl and cyclobutyl giving the (R)-product 22 (R¹ = Ph, R² = H) with 93% ee and 95% ee, respectively, while the use of cyclohexyl and sterically similar isopropyl groups led to a significant decrease of enantioselectivity (54% ee and 60% ee). The N-methyl group, however, was too small for efficient enantio-differentiation. In the exo-TS, strong steric repulsion between the R¹ and cyclohexyl groups disfavors the addition of an azomethine imine to a coordinated alkyne, which would otherwise lead to the formation of (S)-enantiomeric product 22.

Copper-catalyzed 1,3-dipolar cycloadditions of N,N-cyclic azomethine imines to terminal alkynes, which proceed via Cu(I)-acetylide intermediate, generally afford 5,6-disubstituted bicyclic products of type 26. In 2012, Kobayashi et al. developed a silver amide-catalyzed cycloaddition of azomethine imines 24 with terminal alkynes to exclusively afford 5,7-disubstituted bicyclic products 25 (Scheme 10) [26]. This was the first protocol for the synthesis of 5,7-cycloadducts with reversed regioselectivity as in previously reported cycloadditions. It was found out that the strong basicity of AgHMDS was essential for this cyclization reaction. To efficiently suppress the formation of 1,2-adduct 27, the reaction has to be performed in the presence of molecular sieves using only a slight excess of alkyne. An asymmetric version of this highly regioselective cycloaddition was achieved by utilizing CuHMDS in combination with chiral 2,2′-bis(diphenylphosphino)-1,1′-binaphthalene (BINAP) ligand L10. The reaction tolerated electron-deficient and electron-rich aryl, alkenyl and alkyl R¹ groups on the azomethine imine 24 furnishing the products 25 in high yields and ee values in the range of 87–95%. With regard to the terminal alkyne, the reactions proceeded cleanly with electron-deficient and electron-rich aromatic, alkyl, silyl and protected alcohol R² groups, again with high enantioselectivities (ee’s up to 93%). Mechanistic studied revealed that, in contrast to the previously proposed concerted cyclization pathway [27], the cycloadduct products 25 are formed via a step-wise reaction mechanism. It involves a 1,2-addition of the metal acetylide 30 to the azomethine imine 24 followed by intramolecular cyclization of the Lewis acid-activated alkyne 31. The latter assumption was confirmed by successful cyclization of alkyne 28 to bicyclic product 29 with retained ee value. While CuHMDS combined with BINAP-type ligand enabled exclusive formation of products 25, a complete reversal of regioselectivity (formation of 26) was observed when bisoxazoline L11 or 2,2′-bipyridyl ligand was used.

3. Synthesis of Pyrazolidines

3.1. By Metal-Catalyzed Cycloadditions

Acylhydrazones can be visualized as imine equivalents, and their benzoyl derivatives have been shown to react with several nucleophiles under Lewis acid catalytic conditions or in the presence of Lewis base promoter. Acylhydrazones have also been explored as general 1,3-dipoles, and it has been found that the [3+2] cycloaddition reaction of acylhydrazones with olefins is possible under Lewis acid catalytic conditions [28]. Additionally, asymmetric intramolecular [3+2] cycloadditions of acylhydrazones have been described by Kobayashi using chiral zirconium catalyst leading to condensed pyrazolidine derivatives [29]. The chiral zirconium catalyst was prepared in situ from Zr(OPr)₄ (10 mol %) and selected (R)-BINOL ligand (12 mol %). The reaction was run in benzene or dichloromethane in the presence of 50 mol % of PrOH as an additive at room temperature yielding the corresponding fused pyrazolidine derivatives as major trans isomers (cis/trans ratio > 1/99) and with ee from 72 up to 97% (Scheme 11). Although that precise reaction pathway in these reactions is not clarified, it was clearly shown that cis/trans selectivity was controlled by the chirality of the chiral catalyst. For example, when (S)-L12 was used as a catalyst instead of (R)-L12 under the same reaction conditions, the corresponding cis isomer of cycloadduct of (S)-citronellal-derived 4-nitrobenzoylhydrazone was obtained with good selectivity.

However, in this methodology, the substrate scope was restricted, and the reactions were limited to only an intramolecular type of [3+2] cyclisation. Later on, the same group developed the efficient zirconium-catalyzed enantioselective [3+2] intermolecular cycloaddition of hydrazones to olefins leading to optically-active pyrazolidine, pyrazoline and 1,3-diamine derivatives [30]. In this case, benzoylhydrazones were found to be superior substrates over the 4-nitrobenzoyl analogues. The amount of olefin used also effected the reactivity, and the yields were significantly improved by using two equiv. of substituted olefins. Hydrazones derived from a β-branched aldehyde, a sterically-hindered aldehyde, enolizable aldehydes and functionalized aldehydes all readily reacted with ethene-1,1-diylbis(methylsulfane) without any side reactions, and the corresponding substituted pyrazolidines 35 were obtained in high yield and high levels of enantioselectivity (Scheme 12).

Furthermore, the variety of hydrazones derived from α-branched and β-branched aliphatic aldehydes reacted readily with different vinyl ethers as olefins furnishing the pyrazolidines 37 containing a N,O-acetal functionality in high yields and enantioselectivities (Scheme 13). In this particular case, p-nitrobenzoyl hydrazones were found to be more suitable derivatives than their benzoyl analogues due to their higher reactivity under the catalytic conditions. However, moderate diastereoselectivities were obtained in the majority of the cases (diastereomer ratio (dr) > 1/1), whereas both diastereomers showed excellent enantioselectivity in most instances (ee > 81%). Concerning the mechanistic aspect of the transformations, the authors suggested a concerted reaction pathway over the alternative stepwise pathway. The mechanism of the transformation was among the other facts elucidated on the observation that both diastereomers of the products have the same absolute configuration at the carbon atom bearing the R¹ substituent (Scheme 13).

The developed methodology was elegantly applied for the synthesis of enantioenriched compound MS-153 of biological importance. The [3+2] cycloaddition of hydrazone 38 to propyl vinyl ether proceeded with 76% yield and 84% ee for the major and 97% ee for the minor isomer. The p-nitrobenzoyl functionality was removed by treatment with LiAlH₄ at −78 °C in THF followed by nicotinoylation. The obtained intermediate 40 was a useful starting material for the synthesis of MS-153 derivatives. However, after the removal of the phenylthio group using Raney-Ni, ent MS-153 was obtained with high enantioselectivity (88% ee) (Scheme 14).

Phenyl and tert-butylsilane reagents (easily prepared from pseudoephedrine and phenyl or tert-butyltrichlorosilane as a mixture of diastereomers at silicon) introduced by Leigghton et al. were applied for acylhydrazone-enol ether cycloadditions [31]. Phenylsilane was found to facilitate the cycloaddition of N′-(3-phenylpropylidene)benzohydrazide with ethoxyethene to yield the corresponding pyrazolidine in 61% with 6:1 diastereoselectivity and 77% ee. The use of tert-butyl vinyl ether resulted in significant improvement in both diastereoselectivity (dr 24:1) and enantioselectivity (90% ee), and the transformation was found to perform the best in toluene at room temperature (Scheme 15). Variety of benzoylhydrazones derived from aliphatic (Ph(CH₂)₂CHO, BnOCH₂CHO, CyCHO, t-BuCHO) and (hetero)aromatic (benzaldehyde, 4-fluorobenzaldehide and furan-2-carbaldehyde) aldehydes reacted readily with tert-butyl vinyl ether (three equiv.) in the presence of phenylsilane (1.5 equiv.), yielding 3,5-disubstituted pyrazolidines 43 up to 85% yield, >19:1 dr and ee’s above 90% (Scheme 15). Unlike the Zr(II)-catalyzed cycloaddition of benzoylhydrazones 36 (Scheme 13), in this case, the β-substituted enol ethers reacted highly stereoselectively (>10:1 dr, 97% ee) with N′-(3-phenylpropylidene)benzohydrazides 42 providing enantio-enriched 3,4,5-trisubstituted pyrazolidine 43 (Scheme 15, Example A). Selectivity with the acetaldehyde-derived hydrazone (R² = Me, Scheme 15) is significantly reduced relative to that observed with larger R² groups when phenylsilane 41a is applied. Replacing the phenyl moiety in chiral silicon Lewis acid 41a with sterically more demanding tert-butyl group 41b resulted in significant improvement in stereoselectivity when more electrophilic N′-ethylidene-4-nitrobenzohydrazide 44 was reacted with tert-butyl vinyl ether to provide (5R)-3-(tert-butoxy)-5-methylpyrazolidine 45 (Scheme 15, Example B) in good yield of 85% and stereoselectivity (>15:1 dr, 98% ee). Two additional synthetic steps furnished the targeted MS-153 substrate in 70% overall yield and > 99% ee (Scheme 15, Example B) [32]. Furthermore, a concise synthesis of manzacidin C based on the second generation [33] chiral silicon Lewis acid, which promoted diastereo- and enantioselective acylhydrazone-alkene [3+2] cycloaddition, has also been reported by the same group [34]. The corresponding enantioenriched 3,5,5-trisubstituded pyrazolidine precursor was subsequently used for the synthesis of crucial 1,3-bisamide intermediate, which was obtained in 73% yield and high stereoselectivity (>20:1 dr and 94% ee).

In addition to the above-described silicon Lewis acid-promoted [3+2] cycloadditions of acylhydrazones, Tsogoeva et al. [35] developed a catalytic system comprised of chiral Brønsted acid (chiral BINOL phosphates) and silicon Lewis acid (R₂SiX₂). The combination of both acids acted cooperatively in the stereoselective intermolecular cycloaddition of acylhydrazones with cyclopentadiene providing a convenient and facile process for the synthesis of pyrazolidines. Assumed synergism between the BINOL phosphate (R)-L14 and Ph₂SiCl₂ was also tested. Lewis acid (dichlorodiphenylsilane) alone was inactive in catalysis, whereas BINOL phosphate itself showed low reactivity and moderate enantioselectivity (99:1 dr, 47% ee, example: R¹ = Et, R² = 4-(NO₂)-C₆H₄, Scheme 16). The combination of both components in ratio (R)-L14/ Ph₂SiCl₂ in 2/1 was established to be the most active, producing the corresponding pyrazolidines 48 as major cis isomers in good to excellent enantioselectivities (Scheme 16).

Copper(II)-catalyzed diastereo- and enantioselective cycloadditions of azomethine imines to 2-acryloyl-3-pyrazolidinones were developed by Sibi et al. deriving the corresponding exo cycloadducts with high diastereoselectivities (up to > 96:4 cis/trans) and enantioselectivities for cis isomer up to 98% ee [36]. The Cu(OTf)₂/L15-catalytic system was found to be the most efficient in producing the corresponding cycloadducts 51 as a major syn diastereomers and with excellent enantioselectivities (Scheme 17). Chiral Lewis acid prepared from Cu(OTf)₂ and additional bis(oxazoline) ligands such as t-Bu-BOX, Ph-BOX and Bn-BOX were also studied as catalysts; however, when the Cu(OTF)₂/t-Bu-BOX system was used as a catalyst, syn products were formed in good yields and with lower diastereoselectivity. On the other hand, Cu(OTF)₂/Ph-BOX and Cu(OTF)₂/Bn-BOX catalytic systems behaved similarly yielding exo/endo product in ~5/1 and 82% ee of the major isomer. When applying the Cu(OTf)₂/L15-catalytic system, generally excellent yields and good diastereo- and high enantioselectivities were possible from cycloadditions of a variety of C5 and N′-arylmethylidene-substituted azomethine imines 50 and different α,β-unsaturated pyrazolidinone imides 49 (Scheme 17). However, when the methodology was extended to β-substituted α,β-unsaturated pyrazolidinone imides such as pyrazolidinone chrotonate, no reaction was observed or the cycloadducts were formed in moderate yields and in low enantioselectivities.

An additional example of copper(II)-catalyzed [3+2] cycloadditions of N,N′-cyclic azomethine imines is Cu(OAc)₂-catalyzed cycloaddition of azomethine imines to methyleneindolinones described by Sun and coworkers [37]. Among tested Lewis acids (Cu(OAc)₂, AgOTf, Sc(OTf)₃, Zn(OTf)₂, CuI, Cu(acac)₂ and Cu[MeCN]₄PF₆), Cu(OAc)₂ was found to be the most efficient in acetonitrile as a solvent of choice. Generally, when Cu(OAc)₂ was used as a Lewis acid catalyst, the reactions proceeded smoothly to afford the desired spiro adducts 55 in moderate to high yields (70–82%) and good stereoselectivities (9:1 up to 15:1) forming trans diastereomer as the major isomer (Scheme 18). Notably, the heteroaryl (2-thienyl) and alkyl substituted azomethine imines were tolerated in the reaction and derived the corresponding products 55p and 55q in 77%, 14:1 dr and 81%, 8:1 dr, respectively (Scheme 18). An enantioselective variant of this reactions was also carried out screening a series of chiral ligands (i-Pr- and t-Bu-substituted oxazoline ligands and i-Pr-Phosferrox ligand), affording the corresponding product 55a in moderate yield and low enantioselectivity (up to 48% ee).

Recently, Feng et al. developed highly diastereo- and enantioselective [3+2] cycloaddition of methyleneindolinones with N,N′-cyclic azomethine imines, which was successfully catalyzed by a N,N′-dioxide-ligated-Mg(OTf)₂ complex [38]. It should be stressed that unlike in the above-described Cu(OAc)₂-catalyzed examples [37], in this case, the major cis diastereomer prevailed. Three different alkaline metal salts (Ca(OTf)₂, Ba(ClO₄)₂, and Mg(OTf)₂ were tested in combination with several chiral N,N′-dioxide ligands derived from (S)-proline, (S)-Ramipril and (S)-pipecolic acid. However, the complex of Mg(OTf)₂ could offer only one isomer in nearly quantitative yield in the test reaction. The optimized reaction conditions constituting of 5 mol % of catalyst (L16-Mg(OTf)₂, 1:1) in dichloromethane at 30 °C in 48 h yielded the corresponding spiro adducts in excellent yields and enantioselectivities (Scheme 19). Methyleneindolinones bearing substituents at different positions of the phenyl ring had little influence on the selectivity (dr 19:1 in most cases with ee’s 90–99%). Furthermore, good results were also obtained when changing the electronic and steric nature at meta- or para-position on the aryl moiety of the N,N′-cyclic azomethine imines. However, the azomethine imines with a substituent at the ortho position in the phenyl ring in general gave lower yields of cycloadducts with decreased enantioselectivity (e.g., R² = 2-F-C₆H₄, 64% yield, 50% ee, Scheme 19). The relationship between the enantiomeric excess of the cycloadduct and L16 showed a weak negative nonlinear effect, suggesting that both monomeric and oligomeric catalytic species exist in the solution, but the monomeric complex might act as the more active one since its existence was clearly confirmed by observing the molecular mass of [L16 + Mg²⁺ + OTf⁻] in ESI-MS spectra.

An attractive and unique 1,3-dipolar cycloaddition reaction of azomethine imines 60 to allylic [39,40] and homoallylic [41] alcohols 59, based on a magnesium-mediated multinucleating chiral reaction system utilizing diisopropyl (R,R)-tartrate [(R,R)-DIPT] as a chiral reagent was developed by Ukaji and Inomata et al. The transformation was shown to be applicable to both aryl- and alkyl-substituted azomethine imines. The corresponding optically-enriched trans-pyrazolidines 61 were obtained with excellent regio-, diastereo- and enantioselectivity, with results up to 98% ee (Scheme 20). As an example, a mixture of allyl alcohol and (R,R)-DIPT was reacted with alkylmagnesium halide as a magnesium source followed by the addition of azomethine imine. The reactions were performed at 80 °C and alkanonitriles (MeCN or EtCN) were found to be the most suitable solvents. The use of a catalytic amount of (R,R)-DIPT was also effective when accompanied by the addition of one equiv. of MgBr₂. The proposed transition state model consists of a carbonyl oxygen of azomethine imine, rather than imine nitrogen, coordinated to the magnesium salt of (R,R)-DIPT. In this model, the pro-R allyloxy moiety approaches the azomethine imine group, giving the R,R,R-configuration (confirmed by the X-ray analysis of cycloadduct derivatives) of the cycloadducts.

The diastereodivergent process leading to enantiopure pyrazolidine derivatives 64 and 65 from 3-acryloyl-2-oxazolidinone 62 and N,N′-cyclic azomethine imines 63 employing Ni(II)-catalyzed asymmetric [3+2] cycloadditions was developed [42]. Among the tested catalytic systems, the in situ formed binaphthyl-quinolinediimine-based Ni(II) complex, (R)-L17–Ni(II), exhibited favorable activity yielding the corresponding pyrazolidine derivatives 64 as a major trans diastereomer and with high enantioselectivity. Under the catalysis of (R)-L17–Ni(ClO₄)₂·6H₂O in chloroform at 40–50 °C and in the presence of molecular sieves (MS) 4 Å, the reaction proceeded well and was independent of the electronic character of the benzene ring substituents (Scheme 21). It is of note here that the enantio- and diastereoselectivity improved in some cases upon increasing the reaction temperature from room temperature up to 50 °C, which could be attributed to enhanced ligand exchange rate of the transition state Ni(II)-complex. Although the reaction of C-cyclohexyl substituted azomethine imine resulted in moderate yield and 74% ee, the cycloaddition of naphthyl and heteroaryl analogues afforded excellent enantioselectivities (Scheme 21).

The results of the Ni(II)-catalyzed reaction are comparable to those of analogous Cu(II)-catalyzed reaction [36], where exo-cycloadduct (yielding the cis-diastereomer) was the major product, complementarily since in these cases, the endo-cycloaddition occurred. The difference in diastereoselectivity between the Cu(II)- and the Ni(II)-catalyzed cycloaddition of structurally-similar substrates can be explained by the alternative approach of the azomethine imine dipole depending on the metal geometry of catalytic chiral Lewis acid complex. The Cu(II)-catalyzed cycloaddition most probably proceeds via a distorted square-planar complex, which favors the exo-approach of a dipole. However, the Ni(II)-catalyzed [3+2] cycloaddition likely proceeds through an octahedral complex forcing the endo-approach of a dipole.

Togni et al. developed Ni(II)-Pigiphos enantioselective 1,3-dipolar cycloaddition of various C,N-cyclic azomethine imines 66 to α,β-unsaturated nitriles [43]. When N-benzoylamino-3,4-dihydroisoquinolinium betaine was reacted with acrylonitrile in the presence of 5 mol % of [Ni(PPP)(MeCN)](BF₄)₂ in dichloromethane, the completion of the reaction was observed after 0.5 h yielding 3,4-regioisomer, with not only good diastereomeric access of the trans-isomer, but also excellent enantioselectivity (96% ee). Under the optimized reaction conditions, the substrates functionalized with different electron-donating or electron-withdrawing groups yielded moderate enantiomeric excesses, form 60% up to 88% ee (Scheme 22). The exo-selectivities were good to moderate, decreasing from 13.3:1 to 5.6:1 in decreasing the electron-donating ability of the substituent (Me, H, Br, F) on the position para to the 1,3-dipole. The relative trans-stereoselectivity of the major cycloadduct 67 was established in each case by ¹H and 2D NMR spectroscopy. Furthermore, the absolute configurations (3R,4R) could be established by X-ray crystallography of enantiopure single crystals of the major diastereomer. The potential influence of a more functionalized cyanoolefins on the exo-selectivity of the transformation was also investigated using crotononitrile, methacrylonitrile, trans-cinnamonitrile and cis-2-pentenonitrile. The reactivity of this substituted acrylonitriles was much more inferior compared to the acrylonitrile, as no cycloaddition occurred with trans-cinnamonitrile and cis-2-pentenenitrile under the same reaction conditions. Furthermore, methacrylonitrile yielded 52% of the corresponding regioisomeric mixture of cycloadducts (3,4-cycloadduct/3,5-cycloadduct, 3.7/1) after 48 h at 40 °C. However, addition of crotononitrile (four equiv., cis/trans mixture) to N-benzoylamino-3,4-dihydroisoquinolinium betaine at r.t. over two days resulted largely in the formation of the trans-3,4-cycloadduct (dr = 7.3/1) in 66% yield and 62% ee of the trans-isomer (Scheme 22, e.g., 67g). On the basis of the above-mentioned results, it can be concluded that the steric size of the dipolarophile is a crucial factor in these 1,3-dipolar cycloadditions.

Feng et al. represented the first example of asymmetric reaction of azomethine imines 70 with alkylidene malonates 69 using N,N′-dioxide-Ni(II) complexes [44]. Different metal sources such as Sc(OTf)₃, Mg(ClO₄)₂ and Co(ClO₄)₂·6H₂O were evaluated; however, the N,N′-dioxide-Ni(ClO₄)₂·6H₂O promoted the reaction giving rise to the corresponding endo-cycloadduct 71 as a single isomer (dr > 99:1). Different nickel sources were also examined (Ni(BF₄)₂·6H₂O and NiBr₂), but no improvement in the cycloaddition was observed. The counter anion has a crucial effect on both the yield and the selectivity in the azomethine imine-alkylidene malonate cycloaddition. Upon the optimization of the reaction conditions, the ratio of 1:1.2 between Ni(ClO₄)₂·6H₂O and N,N′-dioxide ligand L18 was chosen as optimal, and the reaction performed the best in dichloromethane. It is important to note that the solvent effects the reaction greatly since very low conversions were obtained in solvents such as toluene, THF, MeCN and EtOAc. The catalytic composition was also investigated, and a slightly positive nonlinear effect was obtained between the enantiomeric excess of the ligand and the product, suggesting that the minor oligomeric aggregation of the N,N′-dioxide-Ni(ClO₄)₂·6H₂O might exist in the catalytic system. However, the ESI-MS studies of Ni(ClO₄)₂·6H₂O and chiral ligand in a ratio of 1:1.2 in dichloromethane confirmed the existence of the monomeric nickel species [L18 + Ni⁽²⁺⁾ + ClO₄⁻] in solution. Using the optimized catalytic system, a variety of dipolarophiles, (cyclo)alkyl- and (hetero)aryl-substituted alkylidene malonates, reacted with N,N′-cyclic azomethine imines, deriving the corresponding products as sole trans-diastereomers with excellent enantioselectivities (from 83% up to 97% ee) (Scheme 23). To demonstrate the scalability, a gram-scale reaction was performed between diethyl 2-benzylidenemalonate and (Z)-1-benzylidene-1λ⁴-pyrazolidin-3-one under the optimized catalytic system to give the corresponding cycloadduct in 85% yield and 92% ee.

3.2. By Organocatalyzed Cycloadditions

Rueping et al. efficiently employed acidic N-triflylphosphoramide Brønsted acid C1 as an organocatalyst in highly enantioselective cycloadditions between various alkenes and azomethine imines derived from N-benzoylhydrazone precursors. Thus, under optimized reaction conditions, [3+2] cycloaddition of N-benzoyl protected hydrazones 72 (derived from aliphatic, aromatic and heteroaromatic aldehydes or ethyl glyoxylate) to cyclopentadiene (73) afforded pyrazolidine derivatives 74 in 51–99% yields with excellent stereoselectivity (dr ≥ 96:4, 87–98% ee). Next, α-methylstyrene 75 and its (hetero)aryl analogues were used as dipolarophiles in [3+2] cycloaddition with N-benzoyl-protected hydrazones 72, as azomethine imine precursors, furnishing valuable pyrazolidine derivatives 76, bearing a quaternary and a trisubstituted stereocenter at the 3- and 5-positions, in 52–95% yields and 80–96% ee as single diastereomers (Scheme 24) [45].

Suga and co-workers reported cycloadditions of N,N-cyclic azomethine imines 77 to acrolein (78) catalyzed by L-proline C2 and (S)-indoline-2-carboxylic acid C3. The initial cycloadducts 8 were, after reduction with NaBH₄, isolated as the corresponding alcohols. Under the optimized reaction conditions, the L-proline C2-catalyzed reactions gave the cycloadducts 79 in 54–89% yields as the major endo-diastereomers (endo/exo = 83:27 to 99:1) with modest to good enantioselectivity (31–83% ee). Azomethine imines bearing o-chlorophenyl- and alkyl-substituents gave the corresponding products 79 in merely low to modest enantioselectivity (31–50% ee). Interestingly, reactions catalyzed by (S)-indoline-2-carboxylic acid C3 furnished the corresponding cycloadducts 79 with high exo-selectivity (exo/endo = 91:9 to 99:1) and good to excellent enantioselectivity (75–98% ee), in 29–95% yields. Different substituents on the azomethine imines were well tolerated. The authors have shown that exo-cycloadduct can be isomerized into endo-cycloadduct in the presence of L-proline C2 (Scheme 25) [46].

α,α-Bis[3,5-di(trifluoromethyl)phenyl]prolinol C4 was applied as the organocatalyst of choice for the stereoselective cycloaddition of N,N-cyclic azomethine imines 80 to α,β-unsaturated aldehydes 81, furnishing the expected cycloadducts 82 in 50–95% yields with excellent stereoselectivities (exo/endo = 81:19 to 98:2; 82–97% ee). Both electron-withdrawing and electron-donating substituents in the para-position of benzene ring of azomethine imines gave the corresponding products with excellent stereoselectivities. Regarding the applied enals 81, the ones with an alkyl group worked very well, whereas reactions with cinnamaldehyde failed to give the expected products. Azomethine imine, derived from isobutyraldehyde, was successfully reacted with crotonaldehyde in the presence of MacMillan’s first generation organocatalyst C5 [47], giving the corresponding cycloadduct in 40% yield, 77% ee and exo/endo = 95:5 (Scheme 26) [48].

Chen et al. applied 9-amino-9-deoxyepiquinine C6 in the presence of 2,4,6-triisopropylbenzenesulfonic acid (TIPBA) for an efficient cycloaddition of various alkyl and (hetero)aryl substituted N,N-cyclic azomethine imines 84 to 2-cyclopenten-, 2-cyclohexen- and 2-cyclohepten-1-ones 83, giving the desired products 85 in excellent diastereoselectivity (dr > 99:1), excellent enantioselectivity (85–95% ee) and in 67–99% yields. When 9-amino-9-deoxyepiquinidine C7 was applied, enantiomeric cycloadducts were obtained, retaining the same level of stereoselectivity (Scheme 27). The 6-hydroxy group on the quinolone part of the organocatalysts C6 and C7, as well as the presence of the sterically bulky TIPBA additive were essential for the high stereoselectivity of these transformations [49].

A chiral bis-phosphoric acid C8, bearing triple axial chirality, was developed by Wang and co-workers and successfully utilized in the 1,3-dipolar cycloaddition of N,N-cyclic azomethine imines 86 to methyleneindolinones 87. The chiral spiro[pyrazolidin-3,3’-oxindole] products 88 have been formed in 68–94% yields, good to excellent diastereoselectivity (dr from 6:1–20:1) and excellent enantioselectivity (91–99% ee). Azomethine imines bearing different (hetero)aryl substituents, as well as different electronic nature, bulkiness or position of the substitution patterns on the methyleneindolinones had negligible effect on the efficiency and stereoselectivity of transformations under the optimized reaction conditions. DFT calculations have been performed to shed the light on the mechanism of the 1,3-dipolar cycloaddition (Scheme 28) [50].

Kerrigan and co-workers reported 1,3-dipolar cycloadditions of the in situ-generated ketenes (formed from the corresponding acyl chlorides 90 and DIPEA) to azomethine imines 89 in the presence of quinuclidine organocatalyst C9 as the catalyst of choice, thus furnishing (2R,3S)-bicyclic pyrazolidinones 91 in 52–99% yields, excellent enantioselectivity (≥96% ee) and moderate to high diastereoselectivity (dr 3:1–27:1). The method displays tolerance to aryl groups of azomethine imine containing both electron-donating and electron-withdrawing substituents, while with respect to ketene substituents, an acetoxy and different alkyl substituents were employed. With the acetoxy group, a markedly improved trans-diastereoselectivity was observed (dr 12:1 to 27:1). If pseudo enantiomeric catalyst C10 was applied, enantiomeric (2S,3R)-bicyclic cycloadducts 91 were obtained, retaining the same level of stereoselectivity (Scheme 29) [51].

Proline-derived thiourea organocatalyst C11 was used by Kesavan et al. for the formal [3+2] cycloaddition of azomethine imines 92 to malononitrile (93), yielding bicyclic pyrazolidinones 94 with embedded enaminonitrile functionality in 68–92% yield and 41–98% ee. Azomethine imines containing cyclohexyl, heteroaromatic and diversely substituted aromatic substituents were used. 2-methylmalononitrile also gave the expected product (78% yield, 70% ee), while other dipolarophiles such as ethyl 2-cyanoacetate, benzoylacetonitrile and (phenylsulfonyl)acetonitrile failed to react. Presumably, the azomethine imine gets activated via hydrogen bonding with the thiourea moiety, whereas the tertiary amine deprotonates/activates the malononitrile (Scheme 30) [52].

Chi and co-workers developed a new C–C bond activation of γ-mono-chlorine-substituted cyclobutenones 96 enabled by an isothiourea Lewis base organocatalyst C12. The 1,2-addition of Lewis base C12 to 96 furnishes intermediate 97, which undergoes electrocyclic ring opening to generate vinyl enolate intermediate 98. Selective α-addition of 98 to azomethine imine 95 followed by cyclization gives the final products 28 in excellent diastereo- and enantioselectivity (cis-selective, up to ≥ 98% ee). The developed reaction tolerates different aromatic substituents at the β-carbon of cyclobutenone 96, while the mono-chlorine in the 3-position is essential for the positive outcome of the reaction. With regard to azomethine imine 95, substituents in position 8 other than hydrogen result in lower yields and enantiomeric excess, presumably due to steric hindrance (Scheme 31) [53].

Using chiral phosphine organocatalyst C13, Shi et al. managed to employ δ-substituted allenic esters 101 in an asymmetric [3+2] cycloaddition to C,N-cyclic azomethine imines 100, yielding tetrahydroquinoline derivatives 102 in 57–93% yields, with high diastereoselectivity (> 20:1 dr) and good enantioselectivity (68–93% ee). The employed allenoates 101 react as C₂-synthons via their δ- and γ-positions, for which the authors proposed a plausible reaction mechanism. Both simple and sterically-demanding alkyl allenoates 101 have been successfully employed, containing benzyl and various aryl substituents in the δ-position, while the reaction with allenoate containing δ-methyl-substituent failed. In the reaction of δ-Ph substituted allenoates 101 with various azomethine imines 100, the electron-rich alkyl-substituents on the azomethine imine seem to have a beneficial influence on the enantioselectivity of the reaction. N,N-cyclic azomethine imine does not react with δ-substituted allenic esters under the applied reaction conditions (Scheme 32) [54].

Ye and co-workers developed N-heterocyclic carbene-catalyzed [3+2] cycloaddition of α-chloroaldehydes 104 to C,N-cyclic azomethine imines 103. N-heterocyclic carbene was generated in situ from the corresponding triazolium salt C14 and DIPEA. The corresponding pyrazolidinone products 105 were obtained in good yields (65–93%), with moderate to good diastereoselectivity (3:1–8:1 dr) and excellent enantioselectivity (90–99% ee). It was found that both β-(hetero)aryl and alkyl α-chloroaldehydes 104, as well as azomethine imines 103 containing various N-arylcarbonyl groups worked well, affording the corresponding products 105 with high enantioselectivities. On the other hand, the reaction of α-chlorophenylacetaldehyde with azomethine imine, under optimized reaction conditions, furnished the desired product in decreased yield (30%) and enantioselectivity (55% ee), though significantly improved diastereoselectivity (dr > 20:1) (Scheme 33). A plausible catalytic cycle was postulated. It includes N-heterocyclic carbene addition to α-chloroacetaldehyde, thus generating Breslow intermediate [55], which decomposes to azolium enolate. Azolium enolate reacts with the azomethine imine in an addition-cyclization sequence, affording the final products 105 [56].

An organocatalyzed inverse electron demand 1,3-dipolar cycloaddition of C,N-cyclic azomethine imines 106 with azlactones 107 has been established by Su and co-workers employing noncovalent quinidine-derived bifunctional thiourea organocatalyst C15. The initial tetracyclic intermediates 108 are unstable and rearrange into the final products 109. Products 109 were obtained in high yields (85–99%) with excellent stereoselectivity (88–98% ee, dr ≥ 14:1). In general, azlactones prepared from different amino acids were very well tolerated, as well as azlactones containing diverse (hetero)aryl substituents in the C2-position. Azomethine imines with different substituents on the aromatic ring and various substituents at the 4-position of the phenyl ring of the arylsulfonyl protecting group were very well compatible with the developed protocol (Scheme 34) [57].

Zhu et al. developed an inverse electron demand cycloaddition between C,N-cyclic azomethine imines and electron-rich enecarbamates applying chiral phosphoric acids as organocatalysts. For the cycloaddition of benzyl N-vinyl carbamate (111) to azomethine imines 110 (containing both electron-donating and electron-withdrawing substituents, irrespective of their position on the aromatic ring), phosphoric acid C16 turned out to be the optimal catalyst, yielding the corresponding products 112 in good to high yields (68–83%) with excellent regio- and stereo-selectivity (dr ≥ 19:1, 88–98% ee). On the other hand, cycloadditions to dipolarophiles 113 possessing an electron-rich (Z)-internal bond needed re-optimization. Thus, the binol-based phosphoric acid C17 was found to furnish the desired cycloadducts 114 in high yields (89–95%), excellent diastereoselectivity (dr ≥ 19:1) and good to excellent enantioselectivity (80–98% ee). Again, different substitution patterns (substituents) on the aromatic ring of azomethine imines 110 were well tolerated, whereas increasing the size of the β-substituent of the (Z)-enecarbamate resulted in deceleration of the reaction, which at room temperature led to decreased enantioselectivity. Finally, reaction with benzyl [(E)-prop-1-en-1-yl]carbamate afforded the corresponding product with significantly reduced stereoselectivity (Scheme 35) [58].

Wang and co-workers developed an asymmetric synthesis of tetrahydroquinoline derivatives via a [3+2] cycloaddition controlled by dienamine catalysis. Thus, the reaction of α,β-unsaturated aldehydes 116, containing (hetero)aromatic substituents in the γ-position, with C,N-cyclic azomethine imines 115 (either unsubstituted or containing 5-, 6- and 7-Me/MeO/Br substituents) catalyzed by chiral prolinol silyl ether C18 [59] furnished the desired tricyclic systems 117 in high yield (82–92%) with excellent stereoselectivity (dr > 25:1, 90–99% ee). This reaction proceeds via 4,5-reactivity of dienamine reactive intermediates. In stark contrast, reactions performed with aliphatic α,β-unsaturated aldehydes 116 with azomethine imines 115, performed under identical reaction conditions, yielded different [3+2]-cycloaddition products 119. In the latter case, the reaction proceeds via 3,4-reactivity of the iminium ion reactive intermediates. Products 119 were formed in 79–83% yields, dr > 25:1 and 80–94% ee (Scheme 36) [60].

Maruoka and co-workers developed an organocatalyzed inverse electron demand cycloaddition of C,N-cyclic azomethine imines to electron-rich vinyl ether and vinylogous hydrazone catalyzed by axially chiral dicarboxylic acids. Thus, under optimized reaction conditions (organocatalyst C19), azomethine imines 120, bearing either electron-rich or electron-deficient substituents, reacted with tert-butyl vinyl ether (121) to furnish the expected cycloadducts 122 in excellent yields (90–99%) and stereoselectivities (92–97% ee, exo/endo > 95:5). Reactions of azomethine imines 120 with acrolein-derived vinylogous hydrazone 123 needed reoptimization of the reaction conditions (organocatalyst C20). The corresponding cycloadducts 124 were formed in excellent yields (94–99%) and high enantioselectivity (65–92% ee), albeit with modest diastereoselectivity (exo/endo = 2.8:1–6.7:1). The reaction with methacrolein-derived α-substituted vinylogous hydrazone generated preferentially the endo-diastereomer (exo/endo = 1.0:2.4) with moderate enantioselectivity (68% ee), while β-substituted hydrazone derived from crotonaldehyde failed to give the desired product (Scheme 37) [61].

4. Conclusions and Outlook

Since the seminal papers of Kobayashi [29] and Fu [13] over 30 asymmetric [3+2] cycloadditions of azomethine imines have been reported. Asymmetric cycloadditions to alkynes are limited to terminal acetylenes and the use of chiral Cu(I)-based catalysts [11]. The regioselectivity of cycloadditions is invertible by the use of Ag(I) catalysts. Although applicability of Cu(II)-based catalyst has been demonstrated [24], asymmetric cycloadditions to internal alkynes are still very scarce. Therefore, the development of asymmetric catalysts that would enable the enantioselective and regioselective preparation of cycloadducts from non-symmetrical acetylenes represents a synthetic challenge to be met. This is of particular importance, because such asymmetric reactions would allow for the preparation of analogues of Eli-Lilly’s γ-lactam antibiotics. On the other hand, the majority of catalytic asymmetric [3+2] cycloadditions of azomethine imines to alkenes have been performed using various organocatalysts, although examples of transition-metal-catalyzed reactions have also been reported [11,37,38,57]. The use of other dipolarophiles, such as cumulenes, imines, thiones, and nitriles, is another unexplored field of asymmetric cycloadditions of azomethine imines. So far, only reactions with allenoates [62], ketenes [51] and nitriles [52] have been reported.

In conclusion, the number of representative examples of asymmetric [3+2] cycloadditions of azometnine imines to various types of dipolarophiles is currently sufficient to support viability and broad scope of these reactions. However, as this research topic is now at the end if its infancy period it offers interesting and attractive challenges to synthetic chemists.