Catalytic Enantioselective Peroxidation of α, β-Unsaturated Aldehydes for the Asymmetric Synthesis of Biologically Important Chiral Endoperoxides

Abstract

We have developed an unprecedented highly enantioselective catalytic peroxidation of enals. Critical to this development is the discovery that varying the structure of the hydroperoxides has a significant impact on the enantioselectivity of the organocatalytic asymmetric peroxidation. This novel transformation enabled the development of the enantioselective route toward the core structure shared by all members of the stolonoxides family of anticancer natural products, a connected trans-3,6-disubstituted-1,2-dioxane and trans-2,5-disubstituted-tetrahydrofuran ring system. Our route also features an unprecedented cyclization cascade of a chiral bis(epoxy)hydroperoxide. The new methodology and synthetic strategy established from the current studies should be applicable to the enantioselective synthesis of a broad range of chiral 1,2-dioxolanes and 1, 2-dioxanes, thereby facilitating the biological and medicinal chemistry studies of peroxy natural products.

A large number of biologically interesting peroxy natural products contain either a 1,2-dioxolane (7) or 1,2-dioxane (1–6) bearing an acetylester substituent and two or more stereocenters (Scheme 1).¹ It is particular noteworthy that many of these perxoy natural products have been identified to be highly potent anticancer agents.^{1, 2} A general enantioselective approach toward these structural motifs would greatly facilitate the biological and medicinal chemistry studies of peroxy natural products. However, to our knowledge, such a synthetic approach has not yet been established.^{3, 4} The presence of non-adjacent stereocenters in these chiral 1, 2-dioxolanes and 1,2-dioxanes and the general lack of asymmetric peroxidations of broad substrate scope render the efficient stereoselective constructions of such motifs an outstanding challenge not only in synthetic strategy development but also in methodology development.^3c,5

Scheme 1.

Representative Peroxy Natural Products and Our General Strategy to Synthesize These Molecules

We envision that an asymmetric peroxidation of α, β-unsaturated aldehydes 12 followed by an oxidation of the aldehyde to the acid derivatives would provide an attractive route toward β-peroxy acid derivatives 10 (Scheme 1). Asymmetric functionalization of 10 at either carbon 4 or 5 followed by a cyclization with a nucleophilic attack by the hydroperoxide would generate 8 or 9.⁶ This strategy could provide a general approach toward chiral 1,2-dioxolanes and 1,2-dioxanes motifs, thereby facilitating the total synthesis of a broad range of peroxy natural products. Critical to the implementation of this strategy is the development of a general and highly enantioselective nucleophilic peroxidation of enal 12 that is highly general with respect to the β-alkyl substituent. In this manuscript, we wish to report the realization of such an unprecedented asymmetric peroxidation and the application of this new reaction to the development of a stereocontrolled, concise and flexible route for the construction of the connected trans-3, 6-disubstituted-1, 2-dioxane and trans-2, 5-disubstituted-tetrahydrofuran ring system, the core structure shared by all members of the stolonoxides family of marine natural products.⁷

Our previous investigations with the 9-amino cinchona alkaloid 18 led to the development of the first highly enantioselective catalytic peroxidation, namely the peroxidation of α, β-unsaturated ketones with hydroperoxides. We examined 18 for the promotion of a peroxidation of enal 13A with cumene peroxide. Although the reaction with 14a proceeded to completion to afford the desired peroxide 15 as the major product, the enantioselectivity afforded by 18 for this peroxidation was found to be very poor (entry 4, Table 1). Attempts to discover another active catalyst among cinchona alkaloid derivatives revealed that 18 is uniquely active (entries 1–4, Table 1). Although we could improve the enantioselectivity by 18 through further optimizations of other reaction parameters such as reaction temperature and solvent, we were still not able to obtain the β-peroxy aldehyde 15 in synthetically useful ee (entries 5–6, Table 1).

Table 1.

Peroxidation of α, β-Unsaturated Aldehyde 13a with Hydroperoxides 14

entry	peroxide	Cat.	solvent	temp (°C)	time (h)	conv.(%)a	ee (%) 15b
1	14a	DABCO	CH2Cl2	25	24	0	NA
2	14a	16	CH2Cl2	25	24	0	NA
3	14a	17	CH2Cl2	25	24	0	NA
4	14a	18	CH2Cl2	25	6	100	43
5	14a	18	CH2Cl2	0	24	60	67
6	14a	18	toluene	0	24	70	72
7	14b	18	toluene	0	24	56	78
8	14c	18	toluene	0	24	60	71
9	14d	18	toluene	0	24	80	91
10c	14d	18	toluene	0	48	100	91

^aDetermined by ¹H NMR.

^bDetermined by HPLC.

^c2.5 eq. of 14d was used.

Since a wide variety of tertiary hydroperoxides could be easily prepared in one step from readily available tertiary alcohols or 1,1’-disubstituted olefins,⁹ we turned our attention to investigate the effect of varying the structure of the hydroperoxides on the 18-catalyzed enantioselective peroxidation. We found the structure of the hydroperoxides has a significant impact on the enantioselectivity (entries 6–9, Table 1). By exploring this effect, we established a highly enantioselective peroxidation of enal 13A could be achieved with α-methoxydiphenyl hydroperoxide 14d (entry 10, Table 1). Importantly, we were able to extend this high enantioselectivity to the peroxidation of a considerable range of β-alkyl enals was quite general for all kinds of aliphatic enals (Table 2). The ability to tolerate steric variations of and the presence of olefins in the enals rendered this new reaction particularly valuable to facilitate the asymmetric synthesis of chiral 1,2-dioxolanes and 1,2-dioxanes as outlined in scheme 1.

Table 2.

Peroxidation of α, β-Unsaturated Aldehydes 13 with 14d^a

entry	enal		catalyst	yield (%) 15	ee (%) 15b
1		13A	18	60	91(R)c
2		13B	18	62	87(R)
3		13C	18	67	90(R)
4		13D	18	57	90(R)
5		13E	18	61	90(R)
6		13F	19	60	85(S)
7		13G	19	58	86(S)
8d		13H	19	54	87(S)
9		13I	19	55	87(S)
10		13J	19	58	85(S)

^aUnless noted, reactions were run with 0.20 mmol enal, 0.50 mmol 14d in 0.20 mL toluene at 0 °C for 2 days.

^bSee Supporting Information (SI).

^cAbsolute configuration of 15a was determined to be R (see SI).

^dReaction was carried out with 3.0 mmol 13H and 6.25 mmol 14d.

To demonstrate the synthetic consequence of this new asymmetric peroxidation, we next turned out attention to the enantioselective construction of the connected trans-3, 6-disubstituted-1, 2-dioxane and trans-2, 5-disubstituted-tetrahydrofuran ring system, the core structure shared by all members of the stolonoxides family of natural products. These peroxy natural products were isolated from the metabolites of the marine ascidian stolonica socialis.⁷ To date no total synthesis of any stolonoxide has been reported, although preliminary bioassay studies have shown that stolonoxides A-D (1–4) exhibited strong cytotoxicity against the tumor cell lines of mouse lymphoma P-388, human melanoma MEL-28, human prostate carcinoma DU-145, human lung carcinoma A-549, and human colon carcinoma HT-29 with IC₅₀ values ranging from 0.01 to 0.1 µg/mL.^7b In addition, stolonoxides C (1) and A (3) were also found to be inhibitors of the mitochondrial respiratory chain, affecting specifically the functionality of complex II (succinate:ubiquinone oxidoreductase) and complex III (ubiquinol:cytochrome c oxidoreductase) in mammalian cells.^7c Notably, these bioactivities have been attributed to the presence of common dioxane-tetrahydrofuran bicyclic core structure.^7b

As shown in our retrosynthetic analysis (Scheme 2), the stereochemically dense core 20 was to be constructed by an acid-promoted unprecedented hydroperoxide-initiated epoxide-opening cascade from the bis(epoxy)hydroperoxide 21. Intermediate 21 would be prepared from intermediate 22 via the newly developed asymmetric peroxidation of the epoxyenal 22 followed by a Shi epoxidation of the trans-olefin. The epoxyenal 22 could be prepared from known compounds epoxyiodide 25 and vinyliodide 24. Overall, the four stereocenters in 20 were to be created from the 19-catalyzed asymmetric peroxidation, a Shi olefin epoxidation, and D-aspartic acid.

Scheme 2.

Retrosynthesis for the Core Structre of (+)- Stolonoxides C and D

Our synthesis commenced from the coupling of epoxyiodide 25, available from D-aspartic acid via four simple operations in 58% overall yield and 99% ee,¹⁰ with trans-alkenyliodide 24 (Scheme 3).¹¹ 24 was first allowed to react with isopropylmagnesium chloride to generate the trans-alkenylmagnesium chloride,¹² which then coupled with epoxyiodide 25 to form 26 in THF. Through considerable optimizations, we found the use of a stoichiometric amount of CuI and 4.0 equivalents of HPMA was required for the reaction to proceed with high chemoselectivity in favor of the desired coupling reaction. Otherwise, the nucleophilic ring opening of the terminal epoxide became the dominant reaction.¹³ Deprotection of 26 was readily accomplished, and the resulting alcohol 27 was converted into the α, β-unsaturated ester 28 when subjected to a Dess-Martin oxidation followed by a Wittig olefination. Reduction of unsaturated ester 28 by DIBAL followed by an oxidation the allylic alcohol thus afforded the key enal intermediate 22. Catalytic asymmetric peroxidation of 22 with α-methoxydiphenyl hydroperoxide 14d applying the optimal condition described previously (Table 1) readily afforded the desired β-peroxyaldehyde, which was immediately subjected to sodium chlorite and then TMSCHN2 to form ester 30 in 57% yield over three steps. Importantly, the key peroxidation was found to proceed in high stereoselectivity to furnish chiral peroxide in 10:1 diastereomer ratio.

Scheme 3.

Synthesis of the Core Structure of (+)- Stolonoxides C and D

Reagents and Conditions: (a) 24 (1.67 equiv.), ⁱPrMgCl·LiCl, THF, −10 °C; then Cul, HMPA/THF, −30 °C, 66%. (b) TBAF, AcOH(1.2 equiv.), THF, 92%. (c) Dess-Martin periodinane, DCM, 88%. (d) Ph₃Ph=CHCO₂Et, DCM, 94%. (e) DIBAL-H, THF, −78°C, 89%. (f) MnO₂, DCM, rt, 87%. (g) (1) 19 (10mol%), TFA (0.2 equiv.), 14d (2.0 equiv.), toluene, 0 °C; (2) NaClO₂, NaH₂PO₄, t-BuOH/H₂O, rt; (3) TMSCHN₂, MeOH/benzene, rt, 57%, 82% d.e. (h) 31 (1.0 equiv.), Oxzone, k₂CO₃, BuN₄HSO₄, DMM/CH₃CN/Buffer, 68%. (i) TFA, CHCl₃, −30 °C, 75%. (j) TESCI, TEA, DCM, 0 °C, 85%. (k) same condition as (h), 66%. (l) TBAF, buffer (pH=7), THF, 69%. (m) CsOH, (CF₃)₂ CHOH/MeOH, rt, 52%.

Following our original plan we carried out a Shi epoxidation¹⁴ to convert 30 to 32 in 68% yield and attempted a cascade of deprotection of peroxyketal and epoxide-opening to reach dioxane-tetrahydrofuran bicyclic core structure 20 directly from 32. However, these attempts were not successful in spite of our extensive efforts, as subjections of 32 to various acidic conditions applied to the removal of α-methoxydiphenyl group led to decompositions of 32 resulted from the ring-opening of the bisepoxide motif. After considerable experimentations, we developed an alternative route to accomplish the synthesis of 20 from 30. Specifically, we first converted 30 to hydroperoxide 33 in 75% yield upon treatment of 30 with TFA at −30 °C. TES protection of 33 followed by Shi epoxidation transformed 34 into 35 in a highly diastereoselective fashion as determined by ¹³C NMR analysis. Deprotection of 35 with TBAF under neutral condition¹⁵ generated the bis(epoxy)hydroperoxide 21, the required precursor for the key ring opening cascade, in 69% yield.

The construction of the connected trans-3, 6-disubstituted-1, 2-dioxane and trans-2, 5-disubstituted-tetrahydrofuran ring system from 21 via an unprecedented cyclization cascade turned out to be very challenging. We initially attempted the cyclization cascade in the presence of various acids such as camphorsulfonic acid, trifluoroacetic acid, trichloroacetic acid, and triflic acid. Unfortunately, all of these attempts were unsuccessful due to the decomposition of 21 to form unidentifiable mixtures. A trace amount of the desired cyclization product 20 was detected when the cyclization cascade was promoted with a heterogeneous acid, Amberlyst 15.^6c Upon further screening and optimization, we found a catalytic amount of phosphomolybdic acid (PMA) supported on silca gel^6d could promote the desired cyclization cascade to afford 20 in 15% isolated yield along with the tetrahydrofuran-tetrahydrofuran linked bicyclic compound 36 in 36% yield.¹⁶ We next turned our attention to the promotion of the cyclization cascade under basic conditions. Treatments of 21 with common inorganic and organic bases such as KOH, LiOH, DAMP and Et₃N led to the formations of complex mixtures. After extensive experimentations, we found that cesium hydroxide could promote the desired cyclization cascade in a mixed solvent of hexafluoroisopropanol and methanol (7:3) to afford 20 in 52% yield (Scheme 3).^6e

In conclusion, we have developed the first highly enantioselective catalytic peroxidation of enals. It is noteworthy that we are able to achieve synthetically useful enantioselectivity only after the discovery that varying the structure of the hydroperoxides has a significant impact on the enantioselectivity of the organocatalytic asymmetric peroxidation. We have demonstrated the synthetic utility of this transformation in the development of an enantioselective route toward the core structure of stolonoxides C and D featuring an unprecedented cyclization cascade of a chiral bis(epoxy)hydroperoxide. Importantly, the new methodology and synthetic strategy established from the current studies should be applicable to the enantioselective synthesis of a broad range of chiral 1,2-dioxolanes and 1,2-dioxanes, thereby facilitating the biological and medicinal chemistry studies of peroxy natural products.