Synthesis and In Vitro Evaluation of Taxol oxetane ring D precursors

Abstract

A series of potential taxoid substrates was prepared in radiolabelled form to probe in vitro for the oxirane formation step and subsequent ring expansion step to the oxetane (ring D) presumably involved in the biosynthesis of the anticancer agent Taxol. None of the taxoid test substrates underwent transformation in cell-free systems from Taxus suggesting that these surrogates bore substitution patterns inappropriate for recognition or catalysis by the target enzymes, or that taxoid oxiranes and oxetanes arise by independent biosynthetic pathways.

Taxol is an established antineoplastic agent that represents a platform for the development of new, second generation drugs for the treatment of cancers and other diseases.^1,2 The supply of Taxol and its immediate precursors, by isolation from Taxus tissues³ or cell cultures⁴, is expected to be replaced by synthetic biology approaches^5,6 necessitating knowledge of the underlying pathway biochemistry. Whereas many of the biosynthetic steps en route to Taxol have been characterized⁷, the biochemistry of formation of the oxetane D-ring is still uncertain.

The complexity of the Taxol biosynthetic pathway has impeded determination of the timing of oxetane ring formation. Within the proposed 19 distinct enzymatic steps leading to Taxol, oxetane ring formation is anticipated to occur in mid-pathway, presumably following the formation of an acetylated taxadien-2,5,7,9,10,13-hexaol but before the formation of baccatin III, to which the C13-side chain is appended in three steps to complete the pathway to Taxol^7,8.

Theoretically feasible reaction mechanisms to account for the formation of the Taxol oxetane ring D have been proposed by several groups.^9–14 Based on an evaluation of known structures of naturally occurring taxoid derivatives, a simple and plausible mechanism leading to the oxetane D-ring of Taxol was first proposed by Potier and colleagues.¹² The co-occurrence of epoxy and oxetanyl ester taxoids in Taxus species led the authors to propose an enzyme-mediated acid catalyzed epoxyester/oxetaneester rearrangement mechanism involving protonation of a proposed β4,20-epoxide intermediate, backside attack of a C5-acetate moiety onto C4, and rearrangement to the expanded oxetane ring with the formal migration of the secondary 5α-acetoxy group to the tertiary C4α position of the taxane core (scheme 1, I).

Scheme 1.

Proposed Pathway to Taxol Oxetane Ring D.

This mechanism is thought to proceed via a reactive 1,3-dioxolan-2-ylium cation, formed by acetate-assisted opening of the protonated β4,20-oxirane ring¹⁵, but it could also be formulated as a concerted reaction⁷ (scheme 1, II).

From a biochemical perspective, the intermediate 4β,20-function (e.g. 4, scheme 2) could be formed from the corresponding double bond by a cytochrome P450 oxygenase or a flavin-dependent monooxygenase, since double bond epoxidations involving these enzyme types have been observed previously.^16,17 The subsequent ring expansion from epoxide to oxetane with acetate migration could involve a transferase-type enzyme or a mutase of unknown type; this transformation might also be mediated by a cytochrome P450 oxygenase, as somewhat related rearrangements to cyclic ethers catalyzed by P450-enzymes have been recently reported.^18,19

Scheme 2.

Synthesis of Radiolabeled Surrogates.

To explore the enzymology of the presumed epoxidation and oxetane formation reactions in vitro, a series of accessible substrate surrogates was synthesized in radiolabeled form (scheme 2). Although the precise nature (oxidation and acylation state) of the taxoid substrates involved in these reactions are yet to be discovered, we reasoned that allylic ester 1 could serve as a probe to study the initial oxirane ring formation, and the possibility that the resulting epoxide 2 might be processed further to oxetanyl taxoid 3 by subsequent ester-assisted ring expansion. In addition, the epoxy ester surrogates 6 and 7, displaying a more advance oxidation-acylation state, were thought to be good candidates to test the ester-assisted ring expansion reaction outlined in scheme 1, since oxetanyl taxoids with similar functionalities are known in nature.²⁰ Exhaustive literature searches to evaluate the relative abundances of the several hundred defined taxoids from Taxus species,^20–23 in combination with a wide range of biochemical studies,^7,9,11 have shown that none of the known 4β,20-epoxy taxoids bears a benzoate ester at the C2 position (C2-acetates are common), whereas over three-quarters of the characterized oxetane derivatives do bear a benzoate group at the C2 position.^20,21 Based on these structural observations, it is tempting to suggest that 4β,20-epoxy taxoids α-benzoylated at C2 (such as synthesized 7) might be transient intermediates of oxetane ring formation.

Labeled taxusin 1 was prepared by regioselective deacetylation at C13 of 1 using MeLi,²⁴ followed by re-acetylation with tritiated Ac₂O.²⁵ Taxusin-β4,20-epoxide 2 was obtained in low yield by peracetic acid epoxidation.²⁶ Commercially available 1-hydroxy baccatin I 4 was prepared, via diol 5b, in radiolabeled form (6) following the C13 deacetylation²⁴/reacetylation²⁵ procedure described for allylic alcohol 1. Benzoate 7 was synthesized by regioselective hydrolysis (K₂CO₃/MeOH)²⁷ at C2 of 4 affording diol 5a,²⁸ followed by benzoylation of the latter with Bz₂O.²⁹ After structural confirmation of the unlabeled compounds, the syntheses of 2 and 7 were carried out with tritiated Ac₂O and Bz₂O, respectively.

Surrogates (1, 2, 6 and 7) were evaluated as potential substrates for the proposed oxirane and oxetane ring formation reactions in crude, soluble and membranous enzyme preparations from Taxus cells under a broad range of redox-type^17,28 and transferase-type^29,30 reaction conditions. To rule out possible artefacts from non-enzymatic reactions, incubations of compounds 2, 6 and 7 were initially conducted without or with boiled Taxus cell preparations; both experiments revealed that these epoxy ester surrogates were stable under incubation conditions. Since the oxetane (ring D) formation step en route to Taxol is thought to occur at mid biosynthetic pathway,^7,8 the Taxus cell-free preparations were tested for taxoid hydroxylase³⁰ and taxoid acetyl transferase³¹ activities, confirming their catalytic competency for these early and intermediate pathway transformations. Radio-HPLC or HPLC-MS analysis of the reaction mixtures using authentic standards (e.g. baccatin IV and VI, 8) showed that none of the surrogates employed in the present study was converted to the expected epoxide or oxetane derivatives, hence demonstrating that compounds 1, 2, 6 and 7 did not act as functional substrates under the conditions used herein.

In contrast to previous Taxol biosynthetic studies, in which surrogate substrates were of value in defining the target reaction,^32,33 the present case did not yield useful information regarding precursors of taxoid oxiranyl or oxetanyl esters. These negative results suggest that the presumed epoxidation (1→2) and oxetane formation (2→3 and 6/7→8) reactions may require different conditions, or substrates with different substitution patterns to permit recognition and catalysis by the relevant enzymes. Alternatively, these results may indicate that 4β,20-epoxy ester and oxetanyl esters represent two distinct types of advanced taxoids formed by separate biosynthetic routes.