Field of the Invention

This invention relates to epoxides, their synthesis by the asymmetric epoxidation of enones, and their use. Background of the Invention

The enantioselective epoxidation of prochiral alkenes is a valuable methodology, which enables two stereogenic centres to be created in a single synthetic operation. Established methods tend to be limited to specific classes of substrate. The best known is the titanium tartrate- catalysed epoxidation of allylic alcohols, which was first reported by Sharpless as a stoichiometric method, in Katsuki et al , J. Am. Chem. Soc. (1980) 102:5974, and later adapted into a catalytic variant; see Gao et al, J. Am. Chem. Soc (1987) 109:5765.

More recently, epoxidations employing chiral

(salen)Mn(III) catalysts have been applied to a variety of alkene substrates, both unfunctionalised and functionalised; see Jacobsen, chapter 4.2 in Catalytic Asymmetric Synthesis, ed. I. Ojima (1993) VCH, New York.

Although both these known processes are proven as generic methodologies for laboratory-scale synthesis, reliance on metal-based catalysts and reagents means that operation on a large scale can be disadvantageous in terms of cost, work-up procedure and effluent disposal.

A third and potentially more economical methodology is the use of metal-free synthetic polypeptides such as poly- L-leucine as catalysts for the asymmetric epoxidation of prochiral α,5-unsaturated ketones of the general formula

R'R ² C=CR ³ -CO-X (ID

to give the corresponding optically-enriched epoxides

O ι 2C'-C>R_-CO-X (I)

This process was first reported by Julia et al , Angew. Chem. Int. Ed. Engl. (1980) 19:929. However, it is reported that high enantioselectivities are confined to traj-s-chalcone derivatives; see Julia et al , J. Chem. Soc, Perkin Trans. 1 (1982) 1317; Colonna et al , Tetrahedron (1983) 39:1635; Banfi et al , Tetrahedron (1984) 40:5297; Baures et al , Tetrahedron Lett. (1990) 31:6501; and Itsuno et al , J. Org. Chem. (1990) 55:5047. Thus, this reaction has been considered to be of restricted scope in organic synthesis.

Optically-enriched epoxides are especially suited to nucleophilic ring-opening reactions to give, in stereocontrolled fashion, products bearing heteroato functionality on adjacent chiral centres. In this respect, ( 2R,35)-syn-3-phenylisoserine synthons are reported by Boa et al , Contemporary Organic Synthesis (1994) 1:47, and references therein. Several methods proceed via trans- or cis-phenylglycidate intermediates, prepared by enantioselective oxidation (epoxidation and dihydroxylation) of styrene derivatives; see Greene, J.

Org. Chem. (1990) 55:1957; Jacobsen, J. Org. Chem. (1992) 57:4320; and Sharpless, J. Org. Chem. (1994) 59:5105. Although this is an effective overall strategy, provision of enantiopure phenylglycidates relies on the metal-based epoxidation methodologies described above, and aspects of the downstream chemistry are not well suited to operation on a large scale.

Compounds of formula I are known in racemic form. For example, compounds wherein R 1 i.s phenyl, R2 and R3 are each H, and X is t-butyl or cyclopropyl, are disclosed in EP-A- 0336841 and WO-A-0113066, and by Matano, J. Chem. Soc. Perkin Trans. I (1994) 2703, Meth-Cohn, iJb. 1517, and Treves, JACS (1967) 89:6257. The nature of the functional groups makes such compounds difficult to separate into constituent enantiomers, by conventional resolution techniques.

An optically-enriched epoxide of formula I (R = CF-, R ² = R ³ = H, X - t-butyl) is reported by Lin et al , J. Fluorine Chem. (1989) ∑H3-120. Its synthesis is from optically-enriched 1,1, 1-trifluoro-2-hydroxy-5, 5- dimethylhexan-4-one, using lithium diisopropylamide. This is not a commercial process.

Corey et al , Tetrahedron Lett. (1991) 32:2857, report the t-butyl glycidate 5 (see Scheme 1) as the product of a chiral Darzens reaction between t-butyl bromoacetate and benzaldehyde.

Summary of the Invention

It has surprisingly been discovered that asymmetric epoxidation of the type reported by Julia et al can tolerate a greater range of substituents than is indicated by the prior art. More particularly, the present invention enables the preparation of novel optically-enriched epoxides of formula I wherein R , R and R are each independently selected from H, R, R-CO- and R-O-CO-, each R independently being substantially a hydrocarbon group of up to 20 carbon atoms, and X is an alkyl or cycloalkyl group of up to 10 carbon atoms, provided that -CO-X is not enolisable. Novel epoxides of formula I constitute a further aspect of the invention. Description of the Invention The nature of each of R , R and R is not critical, providing that it does not interfere with the asymmetric epoxidation reaction. For example, R is essentially a spectator to the reaction, for which the non-enolisable nature of -CO-X is important. R may be, for example, a group of up to 10 carbon atoms. R and R , and optionally also R , may be H. R and R may be linked, e.g. together are -(CH ₂ ) ₄ -. Any group other than H may comprise C and H atoms only, or may comprise one or more heteroatoms and/or substituents. One preference for R is aryl or heteroaryl, optionally linked via a conjugating group to CR , e.g. phenyl (or substituted phenyl) .

Most preferably, X is t-butyl (for example when R ¹ is phenyl and R ² and R are each H) , since the substrates for epoxidation are readily available, or simply obtainable from inexpensive readily-available starting materials such as pinacolone. Another simple non-enolisable group is provided when X is cyclopropyl.

X as alkyl can readily be converted to alkoxy by the Baeyer-Villiger reaction. As indicated below, compounds in which X is t-butoxy are of particular interest. The present invention provides, for example, an asymmetric route to a ,/3-unsaturated esters or other carboxylates such as amides, in order to access, for instance, the phenylisoserine component of the anticancer natural product taxol or intermediates for the antihypertensive drug diltiazem.

Obviously a benefit of the methodology is that either enantiomer of the epoxide can be obtained with equal facility by using the appropriate catalyst, e.g. either the L- or the D-polyamino-acid. Other catalysts can be used, as may be found effective, by trial and error. The catalyst may be a material obtainable by nucleophile- promoted oligomerisation of an amino-acid carboxy anhydride. An alternative catalyst is the immobilised catalyst system described by Itsonu et al, J. Org. Chem. (1990)

The conditions reported by Julia et al , supra , for the asymmetric epoxidation comprise a three-phase system of poly-amino-acid catalyst, an organic solvent such as n- hexane, and an aqueous phase containing a large excess of both oxidant (hydrogen peroxide) and alkali (sodium hydroxide) . For the economic utilisation of the methodology for the manufacture of bulk single-enantiomer intermediates to, e.g. pharmaceuticals, it would be desirable to reduce the need for any excesses of reagents. It has been discovered that, by the use of solutions of perborates, the amount of alkali, e.g. hydroxide, required in the reaction can be cut down substantially. As

a result, apart from the saving in reagents, substrates may be used that are otherwise sensitive to the high alkali concentrations present. In the novel conditions, the oxidation system comprises the poly-a ino-acid catalyst, an organic solvent such as dichloromethane, and an aqueous phase containing (sodium) perborate and alkali, e.g. sodium hydroxide. In addition, some phase transfer catalyst such as Aliquat 336 is added. A discovery is that only one equivalent of the sodium hydroxide is required. These oxidation conditions may apply to other heterogeneous oxidations.

Scheme 1 below illustrates reactions according to this invention, shows useful embodiments l, 2 and 3, and an important, illustrative use of products of this invention. All these embodiments may be generalised to the scope of the invention.

The following Examples illustrate the invention. As shown in Table 1, a variety of epoxyketones (R = R - H) have been prepared in good to excellent yield and excellent optical purity (Example A is for comparison) . All these epoxidations were carried out at ambient temperature in a three-phase system with an organic solvent, a catalytic amount of poly-L-leucine synthesised according to Flisak et al , J. Org. Chem. (1993) 58:6247, or poly-D-leucine synthesised in the same way from D-leucine, and with a large excess of oxidant. Preactivation of the catalyst, by stirring the mixture for 6 hours before addition of the α,S-unsaturated ketone, resulted in a shorter reaction time, e.g. of 1 to 3 days. Preferred solvents for these reactions are hydrocarbons such as hexane or chlorinated solvents such as dichloromethane. Optical purities (as given in Table 1) were as determined by HPLC on a Chiralpac AD column, and absolute configurations assigned as [2R,3S] for those epoxides obtained from using poly-L-leucine. The catalyst could be recovered and reused.

Table 1

Conditions (i) = poly-L-leucine/H-0 ₂ /NaOH/CHCl ₂ (ii) = poly-D-leucine/H ₂ 0 ₂ /NaOH/CHCl ₂

The results summarised in Table 1 show that the epoxidation reaction has a broad substrate specificity and is therefore not restricted to chalcones. Satisfactory enantioselectivities were obtained, including a case where the substrate has a second conjugated double bond (Example 4) . Example A was a relatively prolonged reaction, by comparison with Examples 1 and 2, owing to the possibility of enolisation.

With reference to Scheme 1, as exemplification of the value of epoxyketones 1, a further aspect of the present invention is the use of (IS, 2R)-1,2-epoxy-4,4-dimethyl-1- phenyl-3-pentanone 2 (synthon 3 is an alternative) in processes for the preparation of taxol side-chain synthons such as (2R,3S)-N-benzoy1-3-phenylisoserine synthons 4, wherein the t-butyl ketone functionality serves as a masked carboxylate. As summarised in Scheme 1, conversion of 2 to

4 can be accomplished by either of the following sequences:

(a) Baeyer-Villiger oxidation to produce t-butyl 2,3- epoxy-3-phenylpropanoate 5, inversion of configuration at C-3 to afford cis-epoxide 6, nucleophilic ring opening at the benzylic position with either ammonia or azide anion (followed by reduction to the amine) ,

N-benzoylation and optional acid-catalysed deesterification. In contrast to similar prior art processes for the corresponding n-alkyl ester

(McChesney, Tetrahedron: Asymmetry (1994) 5:1683;

Jacobsen, J. Org. Chem. (1992) 57:4320), during

epoxide ring opening with ammonia the t-butyl ester provides effective protection against unwanted amidation at C-l, and allows final unmasking of the C- 1 carboxyl group to be carried out under mild, non- hydrolytic conditions.

(b) Similar to (a) , but with Baeyer-Villiger oxidation carried out at the penultimate stage. By this route the t-Bu ketone provides effective masking for the C-l carboxyl group through much of the synthesis. Scheme 2 shows another use for a compound of the invention, i.e. in the synthesis of an α-hydroxyester precursor to -2-naphthylalanine.

Scheme 1

(2) (5) (6)

C) (4)

Scheme 2

mCPBA

H_,P<_C

a TsCt. _%N, DMAP bNaN_DMF

CCπ ccn