CuH-Catalyzed Regioselective Intramolecular Hydroamination for the Synthesis of Alkyl-Substituted Chiral Aziridines

Abstract

This report details a general and enantioselective means for the synthesis of alkyl-substituted aziridines. This protocol offers a direct route for the synthesis of alkyl-substituted chiral aziridines from achiral starting materials. Readily accessed allylic hydroxylamine esters undergo copper hydride-catalyzed intramolecular hydroamination with a high degree of regio- and enantio-control to afford the aziridine products in good to excellent yield in highly enantioenriched form. The utility of the products derived from this method is further demonstrated through the derivatization of chiral aziridine products to a diverse array of functionalized enantioenriched amines.

Graphical abstract

Optically active aziridines constitute a key functional group present in several classes of natural products and pharmaceutical agents.¹ Stereochemically well-defined aziridines also serve as useful building blocks or intermediates for the synthesis of a range of biologically active, nitrogen-containing compounds.² For instance, aziridines readily participate in nucleophilic ring opening and ring expansion reactions, as well as in cycloadditions with dipolarophiles.² Importantly, the stereochemistry embedded in the heterocycle can often be transferred to the final product to provide configurationally-pure amines. Consequently, a number of synthetic strategies for the stereoselective synthesis of aziridines have been developed.³ Prominent techniques include metal-catalyzed nitrene cycloadditions with olefins,⁴ carbene additions to imines,⁵ the aza-Darzens reaction,⁶ stereoselective azirine reductions⁷ and C–H activation-based approaches.⁸ Despite the utility of these methods, they are generally limited to the preparation of aziridines with electron- withdrawing N-substituents or require particular substitution patterns for productive reactivity. In general, alkyl-substituted aziridines, are especially challenging to access in a catalytic, enantioselective manner. One recent approach, disclosed by Lindsley, leverages the previously established enantioselective α-chlorination⁹ of aldehydes in a three-step one-pot protocol (Figure 1a).¹⁰ Nevertheless, no synthetic method allows for the direct preparation of chiral N-alkyl aziridines from achiral starting materials in a catalytic and asymmetric manner. We hypothesized that this synthetic challenge could be addressed through the intramolecular hydroamination of an allylic hydroxylamine ester 1, enabled by a catalytically generated chiral-phosphine ligated copper(I) hydride (CuH) species.

Figure 1.

Lindsley’s approach for the synthesis of chiral N-alkyl aziridines and the development of our strategy.

Recently, our group and others have adopted CuH catalysis as a general platform for the enantioselective hydrofunctionalization of olefins.¹¹ Among these transformations, the enantioselective hydroamination was successfully applied to symmetric and unactivated internal olefins (Figure 1b).^11e However, regioselectivity became problematic when unsymmetrical olefins were employed as substrates. Our own experience^11c,e–k and an example from the recent literature^11m suggested that a proximal polar substituent on alkenes could influence the regioselectivity of hydrocupration of the alkene by polarizing the C=C double bond. In the proposed aziridine synthesis (Figure 1c), we postulated that the presence of an allylic hydroxylamine ester would enforce the hydrocupration of the internal alkene to provide high regioselectivity for the desired regioisomer (favoring 2a over 2b). Following hydrocupration, the organocopper intermediate would undergo intramolecular amination to furnish chiral aziridine 3, in preference to the regioisomeric azetidine 4.

Allylic hydroxylamine ester 1aa (Table 1) was chosen as a model substrate to begin our study. Subjecting 1aa to a solution of CuH catalyst (generated from Cu(OAc)₂, (S)-DTBM-SEGPHOS (L1)) and a stoichiometric amount of dimethoxymethylsilane as the hydride source, resulted in the formation of the desired aziridine product 3a¹² in 74% yield and 88% ee. Competing N–O bond reduction by the CuH catalyst accounted for the major side product, amine 5a. In order to suppress this deleterious reduction pathway, the reaction of substrates containing different hydroxylamine esters were examined. While use of electron-deficient p-trifluoromethyl benzoate 1ab exhibited a higher tendency towards reduction, sterically hindered 2,4,6-trimethylbenzoate 1ad provided 3a in comparable yield but lower enantioselectivity (entries 2 and 4). The incorporation of electron-rich p-N,N-dimethylamino benzoate 1ac gave 3a in higher yield and improved enantioselectivity (80% yield, 94% ee, entry 3). By switching to pivalate 1ae, the enantioselectivity was further improved to 96% ee (entry 5). These last two results are in line with what we have seen in olefin hydroamination processes.¹³ In all cases, only a trace amount of the analogous azetidine was seen.¹⁴ Use of other commercially available chiral bisphosphine ligands (L2–L4) did not result in any enhancement over our initial choice of ligand, DTBM-SEGPHOS (L1) (entries 5–8). The bench-stable precomplexed copper catalyst (S)-CuCatMix¹³, consisting of a mixture of Cu(OAc)₂/L1/PPh₃ in a 1:1.1:1.1 ratio, offered a slightly improved yield and also simplified the reaction protocol (entry 9). Lastly, further optimization of reaction temperature (4 °C) and concentration (1 M) allowed the desired product to be obtained in 89% yield by GC analysis (81% isolated yield) with an excellent level of enantioselectivity (98% ee, entry 10).

Table 1.

Optimization of CuH-catalyzed chiral aziridine synthesis^a

entry	R	Ligand	Temperature	yielda	ee
1	1aa	L1	40 °C	74%	88%
2	1ab	L1	40 °C	57%	66%
3	1ac	L1	40 °C	80%	94%
4	1ad	L1	40 °C	77%	20%
5	1ae	L1	40 °C	78%	96%
6	1ae	L2	40 °C	9%	–44%
7	1ae	L3	40 °C	4%	–94%
8	1ae	L4	40 °C	68%	−94%
9	1ae	(S)-CuCatMixb	40 °C	82%	96%
10c	1ae	(S)-CuCatMixb	4 °C	89% (81%d)	98%

^aYields determined by GC analysis with 1-dodecane as internal standard, 0.1 mmol scale.

^bPrepared from Cu(OAc)₂/L1/PPh₃ (1:1.1:1.1).

^c2 mol % catalyst loading, 1 M concentration, 18 h.

^dIsolated yield on a 0.5 mmol scale.

Having identified optimized conditions, the substrate scope of the N-substituent was explored (Table 2). In most cases, substituents on the nitrogen atom of varying electronic and steric properties were well-tolerated and their transformation proceeded with high yields and excellent enantioselectivities. In addition, an aziridine with a pendant aryl chloride (3e) was successfully prepared, allowing cross-coupling technology to be considered for downstream derivatization. A substrate containing a methyl ester and a free phenol (3i) was transformed to product with excellent efficiency and stereoselectivity. Moreover, structurally diverse N-alkyl groups including a silyloxyethyl (3j), a cyclohexyl (3k) and a simple methyl group (3l) could all be employed using our procedure. The presence of an ortho-isopropyl group on the N-benzyl substituent (3h) resulted in diminished enantioselectivity and represents a limitation of the current procedure, although a smaller ortho-methyl group (3g) was readily accommodated.

Table 2.

Substrate scope of CuH-catalyzed chiral aziridine synthesis – variation of substitution on the nitrogen atom^a

^aAll yields represent average isolated yields of two runs conducted on a 0.5 mmol scale.

^b40 °C reaction temperature.

^c5 mol % catalyst used.

^d3 equivalents of (MeO)₂MeSiH were used.

We next examined the scope of substituents that could be appended to the olefin (Table 3). The analogous 3-ethylaziridine 3m was successfully obtained from the corresponding crotyl hydroxylamine ester in 62% isolated yield and 93% ee. A benzyl ether (3n), an N-tosylate (3r), a silylether (3t), and a ketal (3v) were also compatible with the reaction conditions. Furthermore, high yields and enantioselectivities were observed in products derived from substrates that contained heterocyclic fragments. Among the substrates tested, a pyridine (3p), a quinoline (3q), a furan (3f, Table 2), a benzoxazole (3s), a benzothiazole (3t), and a tosyl-protected benzimidazoles (3r) were all well-tolerated under this CuH-catalyzed protocol. Substrates bearing large alkene substituents such as a diphenylmethyl (3l and 3o) and a cyclohexyl (3p) group also underwent the desired aziridination reaction without loss of efficiency or selectivity. Notably, other C=C double bonds in positions distal to the hydroxylamine ester (3u) were left intact, suggesting that the hydroxylamine ester group not only controls the regioselectivity but also activates the adjacent alkene for chemoselective hydrocupration. The presence of a homoallylic benzyl ether did not perturb the regioselectivity of the reaction (3n). Also, a hydroxylamine ester containing a chiral ketal group (1v) can be converted to aziridine 3v with greater than 20 : 1 diastereoselectivity. To demonstrate the scalability of this procedure, product 3o was synthesized on a 5.0 mmol scale. Full conversion of the substrate was achieved with a reduced catalyst loading of 0.5 mol % to efficiently provide aziridine 3o (1.78 g of 3o, 90% yield, 98% ee).

Table 3.

Substrate scope of alkene substituents and heterocyclic substituents^a

^aAll yields represent average isolated yields of two runs conducted on a 0.5 mmol scale.

^b5.0 mmol scale, 0.5 mol % catalyst used, 40 h reaction time.

^c5 mol % catalyst used.

^d48 h reaction time.

During the course of our studies, we found that the nature of olefin geometry had a significant influence on reactivity and enantioselectivity. Under the optimized aziridination conditions, a low conversion of (Z)-1ae (Table 2) was observed, with only trace amounts of aziridine product 3a detected. Although full conversion could be achieved by raising the reaction temperature to 40 °C, the aziridine product 3a was obtained in only 12% isolated yield and 71% ee. The majority of the mass balance was accounted for by reductive cleavage of the N–O bond 5a, a side reaction exacerbated by the low reactivity of the (Z)-configured alkene.^11g

Chiral aziridines are well known as versatile intermediates in organic synthesis.² To showcase the synthetic utility of N-alkyl aziridines, aziridine 3o was converted into a series of chiral amines (Scheme 1). Terminal aziridine 3o underwent regioselective ring opening with azide¹⁵ and acetic acid¹⁶ to provide chiral azido amine 6 and amino alcohol 8, respectively. Subjecting 3o to hydrogenation conditions provided chiral secondary amine 7. Further, LiI-catalyzed insertion of carbon dioxide into 3o afforded oxazolidones 9 as a 9.5 : 1 mixture of regioisomers.¹⁷ In a reaction for which we could find no precedent, treatment of 3o with PPTS afforded chiral pyridinium salt 10·OTs, which allowed the absolute stereochemistry of aziridine product 3o to be established by X-ray crystallography.¹⁸

Scheme 1.

N-alkyl aziridine derivatization^a and X-ray crystallographic structure of 10

^aReaction conditions: a) TMSN₃ (5 equiv), AcOH (5 equiv), MeCN, rt. b) Pd/C (10 mol %), H₂ (1 atm), MeOH, rt. c) AcOH (5 equiv), CH₂Cl₂, rt. d) LiI (1 equiv), CO₂ (1 atm), 40 °C. e) PPTS (1.5 equiv), CH₂Cl₂, rt.

In summary, a highly enantioselective aziridination reaction was achieved through CuH-catalyzed regioselective intramolecular hydroamination. The mild reaction conditions of this protocol allow for compatability of a wide range of functional groups and heterocyclic substitutents. Efforts to extend this intramolecular hydroamination strategy to encompass more types of nitrogen-containing heterocycles are currently under way.