An Asymmetric Synthesis of 1,2,4-Trioxane Anticancer Agents via Desymmetrization of Peroxyquinols through a Brønsted Acid Catalysis Cascade

Abstract

The desymmetrization of p-peroxyquinols using a Brønsted acid catalyzed acetalization/oxa-Michael cascade was achieved in high yields and selectivities for a variety of aliphatic and aryl aldehydes. Mechanistic studies suggest that the reaction proceeds through a dynamic kinetic resolution of the peroxy hemiacetal intermediate. The resulting 1,2,4-trioxane products were derivatized and show potent cancer cytotoxicity.

Trioxanes are important scaffolds, which appear in molecules that exhibit antimalarial, anticancer and antibacterial activities.¹ In particular, artemisinin, administered as a part of a combination therapy for the frontline treatment of malaria, contains a 1,2,4-trioxane as the key pharmacophore. The recent emergence of an artemisinin resistant malaria strain² combined with the fact that artemisinin’s mode of action remains under debate³ increases the difficulty of treating malaria and makes the pursuit of novel therapeutic agents more urgent. One potential solution has been the development of new synthetic endoperoxides.⁴ Enantiomers of a few synthetic trioxanes have shown similar anti-malarial activities⁵ but stereochemistry has a demonstrated impact on anti-cancer activity.⁶ Current methods for the enantioselective synthesis of trioxanes are lengthy and use chiral starting materials or reagents.^5,6,7,8

We envisioned that enantioenriched trioxanes could be accessed quickly and enantioselectively through a desymmetrization of p-peroxyquinols via an acetalization/oxa-Michael cascade first reported by Jefford.^9,10 Cascade catalysis^11,12,13 and desymmetrizations¹⁴ are powerful methods utilized by our group and others to construct complex molecules containing multiple stereocenters in a rapid and efficient manner. Both enantioselective acetalization¹⁵ and oxa-Michael¹⁶ reactions are relatively unsolved problems. We were cognizant of the potential difficulties in this approach due to the inherent reversibility of both transformations particularly under acidic conditions. Nevertheless, we were emboldened by recent successes in this area.¹⁷

We began our investigation by studying the desymmetrization of p-peroxyquinol 2a, trivially accessed from cresol, using chiral Brønsted acid catalyst 5 (TRIP) which afforded the desired trioxane in good yield as a single diastereomer in 86% ee (entry 5, Table 1). Switching to the bis(2,4,6-triisopropylphenyl)spirobiindane phosphoric acid 6a developed by List^15e,18 improved the enantioselectivity to 96%. Other biindane Brønsted acids were screened but the parent acid 6a gave the best results. Lowering the catalyst loading from 10 mol % gave decreased reactivity, which could be restored through the use of thiourea 7 as a co-catalyst. Catalyst loadings as low as 2 mol% may be used at the expense of a longer reaction time (entry 7, Table 1). The use of thiourea 7 alone leads to no product.

Table 1.

Reaction optimization.

entry	variation from “standard” conditions	yield (%)b	ee (%)c
1	none	92	96
2	no 4Å MS	90	88
3	no 7	46	95
4	no 6a	<5	-
5	5, instead of 6a	93	86
6	no 7, 6a (10 mol%)	93	96
7	6a (2 mol%), 72h	88	98
8	6b, instead of 6a	65	95
9	23 °C, 48h, instead of 50 °C, 24	85	96

^aReactions were performed on 0.1 mmol scale (0.25 M solution).

^bIsolated yields of analytically pure material.

^cEnantiomeric excess determined by HPLC using a chiral stationary phase.

With our optimized reaction conditions in hand, we explored the aldehyde scope of the reaction (Figure 1). Paraformaldehyde and a variety of sterically hindered aliphatic aldehydes work well. Aldehydes containing alkyl halides, protected alcohols and protected amines are tolerated affording trioxanes in excellent selectivities. Aromatic aldehydes also participate in the reaction with high enantioselectivity but slightly decreased yields (Figure 1).

Figure 1.

Aldehyde substrate scope^a

^aConditions: 2a (0.3 mmol), 3 (1.25 equiv). All products formed as single diastereomers (>20:1). Enantiomeric excess determined by HPLC using a chiral stationary phase. ^bAbsolute configuration established by X-ray analysis. The rest were assigned by analogy. ^cent-6a used as catalyst.

The reaction also proved tolerant of substitution on the p-peroxyquinol. Products with esters, ethers, and multiple tetrasubstituted stereocenters are all isolated in good yields and selectivities (Figure 2).

Figure 2.

Peroxyquinol substrate scope^a

^aSee footnote a, Figure 1.

The enantiodetermining step is likely the oxa-Michael event based on the high enantioselectivity of the product formed from paraformaldehyde.¹⁹ We propose that the reaction proceeds via a dynamic kinetic resolution of peroxy hemiacetal (+/−)8a (Equation 1). To further test this hypothesis, racemic peroxy-hemiacetal 8ac was formed by heating p-peroxyquinol 2a with isobutyraldehyde. After excess aldehyde was removed, unpurified (+/−) 8ac was subjected to chiral acid 6a. To our delight, trioxane 4ac was formed in good yield as a single diastereomer in 94% ee. This suggests that peroxyhemiacetal 8ac is resolved through a dynamic kinetic resolution (Equation 1). Additionally, monitoring the reaction under standard conditions with catalyst 6a by HPLC, we note that the peroxyhemiacetal remains as a racemate throughout the course of the reaction. A crossover experiment subjecting 4ac to n-butyraldehyde showed that the oxa-Michael is not reversible under the reaction conditions (Equation 2).²⁰

(1)

(2)

The 1,2,4 trioxane products of the desymmetrization have a variety of synthetic handles for subsequent derivatization. A Luche reduction of 4ac forms the allylic alcohol in 4:1 dr and subsequent directed epoxidation delivers highly oxygenated cyclohexane 9 (Figure 3). Bromination of 4ac followed by elimination forms vinyl bromide 10 which allows for the incorporation of a variety of functional groups though cross coupling. Chemoselective reduction of the olefin in the presence of the peroxide may be achieved under the aegis of Rh/Al₂O₃ and Adams’ catalyst. This product can be further reduced under acidic conditions to reveal previously unreported diol 12.

Figure 3.

Trioxane derivatizations.

In addition to serving as the frontline antimalarial agent, artemisinin is cytotoxic toward cancer cells and the 1,2,4-trioxane is believed to play an important role.¹ Our products and their derivatives were screened for cytotoxicity against a variety of cancer cell lines. Compounds 4ac, 4an, and 11 show promising activity toward bone and lung cancer cell lines with in vitro IC₅₀’s from 3-25 μM (Figure 4). Importantly, the significant cytotoxicity of the semi-reduced trioxane 12 demonstrates that their activity is not due solely to the presence of the Michael acceptor.

Figure 4.

Anticancer activity of 1,2,4-trioxane products.

In conclusion we report the first catalytic enantioselective synthesis of trioxanes using a desymmetrization of p-peroxyquinols via an acetalization/oxa-Michael cascade. We propose that the reaction proceeds via a dynamic kinetic resolution of a peroxy-hemiacetal intermediate. The 1,2,4-trioxane products are easily derivatized and show promising cancer cytotoxicity. Further development of this reaction, antimalarial testing of these trioxanes and investigation of the cytotoxicity mode of action are currently underway.