Tungsten-Catalyzed Regio-, and Enantioselective Aminolysis of trans-2,3-Epoxy Alcohols-New Entry to Virtually Enantiopure Amino Alcohols

Abstract

The first catalytic, enantioselective aminolysis of trans-2,3-epoxy alcohols has been accomplished. This stereospecific ring-opening process was efficiently promoted by our W-bishydroxamic acid catalytic system furnishing various anti-3-amino-1,2-diols as products in excellent regiocontrol and high enantioselectivities (up to 95 % ee). Moreover, virtually enantiopure 3-amino-1,2-diols could be obtained through combined asymmetric routes.

3-Amino-1,2-diol is a characteristic structural unit present in numerous biologically active compounds such as leukotriene-antagonist,^[1] human carbonic anhydrase inhibitor,^[2] and anti-inflammatory agents.^[3] Furthermore, 3-amino-1,2-diols are also important synthetic intermediates of cardiovascular,^[4] antibacterial,^[5] sedative agents,^[6] selective norepinephrine reuptake inhibitors^[7] and drug candidates to treat conditions ameliorated by monoamine reuptake^[8]. Most of these biologically important molecules are required to be virtually enantiopure (>99.8 % ee, <0.1% enantiomer) for the pharmaceutical applications. Thus it is highly desirable to develop a direct catalytic approach to prepare enantioenriched 3-amino-1,2-diols starting from readily available precursors. Although enantioselective ring-opening of epoxides using various nucleophiles, such as azide,^[9a,b] amines,^[9c–h] water,^[9i–m] alcohols,^[9d,h,l,m] phenols,^[9h,l–p] thiols,^[9q,r] halides,^[9s–v] and carbon nucleophiles^[9w-aa] has been intensively studied in the past decades and tremendous progress has been achieved, excellent results are usually obtained in the case of unfunctionalized terminal or meso epoxides, while kinetic resolution of 2,3-epoxy alcohols is still elusive.^[10,11] The challenge of this reaction remains not only in facial-selectivity, but also in regiocontrol. Recently, our group reported the first catalytic regioselective and stereospecific ring-opening of 2,3-epoxy alcohols promoted by achiral W-salts.^{[12, 13]} As a continuation of our research in this field, we report the first asymmetric aminolysis of 2,3-epoxy alcohols with various amines as nucleophiles using our W-bishydroxamic acid (W-BHA) catalytic system. Remarkably, this catalyst can promote both epoxidation and ring-opening reaction providing a new access to virtually enantiopure 3-amino-1,2-diols.

For optimization of the reaction conditions, we used aniline (1a) and racemic trans-2,3-epoxy cinnamyl alcohol (2a) as standard substrates. After careful screening of ligands, solvents and temperature we succeeded in establishing the optimum reaction conditions for the kinetic resolution of 2,3-epoxy alcohols through enantioselective aminolysis (Scheme 1).^[14]

Scheme 1.

Optimum reaction conditions for the asymmetric ring-opening of trans-2,3-epoxy cinnamyl alcohol with aniline as nucleophile.

The substrate spectrum of this reaction was then evaluated (Table 1). We first varied the structure of the amines: diverse substituted anilines 1a–k, hetereocyclic amines 1l and 1m, as well as secondary aromatic amines 1n–u were reacted with 2a. Generally, all the reactions proceeded smoothly at 55 °C under the catalysis of 2.5 or 5 mol% W-BHA providing the products 3a–u in 71–95 % yield and 84–93 % ee. Remarkably, all the reactions proceeded with complete regioselectivity in favor of the formation of C-3-regioisomers. Subsequently, we studied the substrate scope further by varying the structure of 2,3-epoxy alcohols. In the cases of substituted trans-3-phenylglycidol all the reactions provided the products 3v-bb in 78–95 % yield, complete regioselectivities and 90–95 % ee. Aliphatic trans-2,3-epoxy alcohols turned out to be less reactive and thus higher catalyst loading (10 mol%) and longer reaction time (7 or 24 h) were required to achieve good conversions. In these cases the reactions yielded the products 3cc-gg in 70–84 % yield, high regioselectivities (C3:C2> 95:5) with excellent asymmetric induction (92–94 % ee). Remarkably, all the C-3-regioisomers of 3cc-3gg could be easily separated from the corresponding minor C-2-regioisomers through column chromatography. One limitation of this method was observed in the cases of cis-, trisubstituted and terminal 2,3-epoxy alcohols which led to low yields (31–64 %) and enantioselectivities (40–74 % ee).

Table 1.

Kinetic resolution by ring-opening reactions of trans-2,3-epoxy alcohols with various amines as nucleophiles.^[a–^d]

^[a]Unless otherwise specified, reactions were performed on a 0.25 mmol scale of amines 1 using 2.3 equiv. racemic trans-2,3-epoxy alcohols (2) and x mol% W(OEt)₆, 1.2x mol% BHA and 20 mol% H₂O₂ (30 %) at 55 °C in 2.5mL THF.

^[b]For catalyst loading and reaction time, see Supporting Information (SI): Page 4.

^[c]Yields of the isolated products.

^[d]All enantiomeric excesses were determined by HPLC-analysis on chiral stationary phase.

Since the ring-opening process is a kinetic resolution of racemic 2,3-epoxy alcohols, we selected two reactions to determine the s-factors. In both cases the recovered 2,3-epoxy alcohols were obtained in good enantioselectivities.^[15]

Recently, our group developed a W-catalyzed enantioselective epoxidation of allylic alcohols using the same BHA as ligand.^[16] Thus, we tested the possibility of combining the epoxidation and the ring-opening reaction in a one-pot procedure. Unfortunately, the results were not promising at all.

This observation suggested that the required catalysts for epoxidation and ring-opening may be enantiomers. Indeed, the enantioenriched 2ee provided by the W-catalyzed epoxidation employing (S, S)-BHA barely reacted with aniline under the catalysis of (S, S)-BHA-W, but reacted smoothly with (R, R)-BHA-W catalyst to furnish 3ee almost exclusively (>99.8 % ee, Scheme 2).

Scheme 2.

Investigation of the Ligand-effect on the Outcome of the Ring-opening Reaction.

These amazing results open a new entry to synthesize virtually enantiopure compounds through a combined asymmetric routes consisting of an initial asymmetric reaction followed by a kinetic resolution process of the resulting functional group.^[17] Remarkably, in general >99.8%ee would be rather difficult to achieve except enzymatic transformations. By implementation of our strategy a variety of virtually optically pure amino alcohols (up to > 99.8 % ee) have been successfully prepared and the results are summarized in Table 2. It is noteworthy that compound 3j is reported to be a human carbonic anhydrase IX (hCA IX) inhibitor.^[2]

Table 2.

Synthesis of virtually enantiopure compounds through combined asymmetric routes.^[a–^h]

^[a]For detailed reaction conditions, see SI page 16–17.

^[b]Yields of the isolated products.

^[c]All enantiomeric excesses were determined by HPLC-analysis on chiral stationary phase.

^[d](S, S)-BHA was employed.

^[e]5 mol% WO₂(acac)₂ and 5.5 mol% BHA were used.

^[f](R, R)-BHA was employed.

^[g]1.5 Equiv aniline was used.

^[h]1.5 Equiv epoxide was used.

As an illustration of the utility of this method, three biologically active compounds 4,^[18] 5^[8] and 6^[19] have been readily synthesized through derivatization of the ring-opening products ent-3o, ent-3r and ent-3k (Scheme 3).

Scheme 3.

Derivatization of the ring-opening products to biologically active compounds in optically pure form.

In summary, we developed the first catalytic asymmetric aminolysis of 2,3-epoxy alcohols, which was efficiently promoted by the W-BHA catalytic system. This kinetic resolution method demonstrated a broad substrate scope and was applicable to various amines and a variety of trans-2,3-epoxy alcohols. Furthermore, the sequential reaction of epoxidation/ring-opening (kinetic resolution) provides a new route to prepare virtually enantiopure products for pharmaceutical use.