Total synthesis of (−)-dihydroprotolichesterenic acid via diastereoselective conjugate addition to chiral fumarates

Abstract

A diastereoselective conjugate addition of a variety of monoorganocuprates, Li[RCuI], to chiral fumarates to provide funtionalized succinates has been developed. The utility of this reaction is demonstrated in a concise total synthesis of (−)-dihydroprotolichesterenic acid that required only four steps and proceeded in an overall 31% yield.

Functionalized succinic acid derivatives and compounds that may be derived therefrom are structural motifs that are commonly found in biologically active molecules of natural and non-natural origin. Some notable examples include the matrix metalloproteinase inhibitor BB-1101 (1),¹ the antifungal and antibacterial agent dihydroprotolichesterenic acid (2),² the glaucoma drug pilocarpine (3),³ and the anti-inflammatory and antiviral agent antrodin D (4) (Figure 1).⁴

Figure 1.

Biologically active molecules accessible through succinate derivatives.

Because of the obvious importance of the functionalized, four-carbon building blocks found in these and other compounds, there has been considerable interest in developing efficient methods for enantioselective and stereoselective synthesis of substituted succinates.⁵ Toward this end, three general diastereoselective approaches using chiral auxiliaries have been reported. These include alkylations of enolates,⁶ oxidative cross couplings of enolates,⁷ and radical mediated conjugate additions to unsaturated carbonyl compounds (Scheme 1).⁸ However, these methods typically suffer from low stereoselectivity or production of toxic stoichiometric byproducts. Furthermore, none of the current strategies to access the succinate core enable access to any desired diastereomer from a single starting material.

Scheme 1.

Diastereoselective Routes to Access Succinates

Diastereoselective conjugate additions that are directed by chiral oxazolidinones have been extensively explored since their initial report in 1993.⁹ In a useful advance in the area, Bergdahl reported that the diastereofacial selectivity of copper-mediated conjugate additions to chiral N-crotonyl oxazolidinones can be reversed by varying Lewis acid additives and solvents, thus enabling a single starting material 5 to be converted into the two diastereomeric products 6 and 7. (Scheme 2).¹⁰ Prior to this discovery, the only tactic that could be used to direct an organometallic reagent to the other face of the carbon-carbon double bond of a crotonate derivative was to employ an enantiomorphic chiral auxiliary on the starting material.

Scheme 2.

Bergdahl’s Switchable Conjugate Addition.¹⁰

In the context of work directed toward the stemofoline alkaloids, we had occasion to expand the utility of the Bergdahl chemistry to γ-alkoxy crotonates,¹¹ and we then queried whether it might also be extended to fumarate derivatives such as 10. If the regio- and stereoselective additions of organometallic reagents to fumarates could be controlled, it would be possible to access a variety of useful succinate derivatives. However, we could not be certain of the regiochemical outcome of cuprate additions to 10 because two carbon atoms are β to an activating carbonyl group. Curran and Sibi had each shown that radical additions proceeded β to the imide moiety of such compounds,⁸ and Evans had found that Mukaiyama-Michael reactions with chiral fumarates also occurred at the carbon atom β to the imide moiety.¹² However, there was no guiding precedent for the reactions of such substrates with organocuprate-derived reagents.

In order to probe the feasibility of inducing addition to chiral fumarates at the carbon atom distal to the chiral auxiliary, we prepared the chiral fumarate 10 in 78% yield from commercially available monomethyl fumarate 8 and oxazolidinone 9 in the presence of pivaloyl chloride, lithium chloride, and triethylamine (Scheme 3).¹³

Scheme 3.

Preparation of Chiral Fumarate 10

We quickly found that implementing Bergdahl’s conditions for the TMSI-mediated conjugate addition of monoorganocuprate reagents, Li[RCuI], to 10 required little optimization. Indeed, reaction of 10 with the monomethylcuprate Li[MeCuI] derived from methyllithium and (CuI)₄(DMS)₃ in the presence of TMSI proceeded with a high degree of regio- and diastereoselectivity to furnish the mono substituted succinate 11a (R = Me) in 89% yield and 19:1 dr (Entry 1, Table 1). None of the regioisomeric addition product was detected in the NMR spectrum of the reaction mixture. The absolute stereochemistry of 11a was established by its subsequent conversion into 2 (vide infra).

Table 1.

Diastereoselective Conjugate Addition

Entry	Alkyllithium	Yieldb	drc
a	MeLi	89%	19:1
b	EtLi	72%	19:1
c	n-BuLi	83%	19:1
d	PhLi	82%	19:1

^a13C NMR spectrum chemical shifts (ppm) for 10.

^bIsolated yields after purification by flash column chromotography.

^cBased on HPLC analysis using a Chiralcel OD-H column.

The observed regioselectivity in the reaction of 10 with a monomethylcuprate reagent is consistent with other conjugate additions that proceed via different mechanistic pathways.^8,12 The enhanced reactivity at the carbon atom β to the imide moiety of 10 can be attributed to its greater electropositive character that is reflected by the relative chemical shifts of the two olefinic carbon atoms in the ¹³C NMR spectrum of 10.¹⁴ Namely, the chemical shift of the carbon atom β to the imide moiety (δ = 133.4 ppm) is 1.5 ppm further downfield than the carbon atom α to the imide (δ = 131.9 ppm). The presence of TMSI, which is proposed to interact selectively with the oxygen atom of the imide carbonyl group,¹⁰ in the reaction mixture would be expected to further increase the electropositive character of this carbon atom.

Toward elucidating the scope of this process, the reactions of the chiral fumarate 10 with monoorganocuprates derived from several other alkyllithium reagents were examined (Table 1). In each case, the additions proceeded with high regio- and diastereoselectivity to give 11b–d. The absolute stereochemistry of 11b–d was assigned based on analogy with 11a. Although monoorganocuprates derived from organolithium species and (CuI)₄(DMS)₃ added smoothly to 10 in the presence of TMSI, efforts to switch the diastereoselectivity of this reaction according to the Bergdahl protocol have thus far been unsuccessful. We have examined a number of alternate reaction conditions. For example, use of alkylcuprates derived from Grignard reagents, MgBr[RCuI], as described by Bergdahl did not give detectable amounts of any conjugate addition product.¹⁰ Reactions in which various chelating Lewis acids such as MgBr₂, ZnCl₂, and MnCl₂ were used with monoorganocuprates, Li[RCuI], gave only small amounts of products arising from 1,4-addition. Hence, the ability to control the diastereofacial selectivity in cuprate additions to chiral fumarate derivatives by the simple expedient of changing conditions as was done for 5 remains an unsolved problem.

Having established the underlying feasibility of inducing highly regio- and diastereoselective additions to 10, we sought to exemplify the utility of the process by applying it to a synthetic problem. We thus identified (−)-dihydroprotolichesterinic acid (2), a member of the paraconic acid family of natural products,¹⁵ as a suitable target. (−)-Dihydroprotolichesterenic acid has been synthesized three times in racemic form^16a–c and twice in enantiomerically pure form.^15d–e However, we envisioned that a more concise enantioselective synthesis of 2 could be achieved using our methodology.

With the key intermediate 11a in hand, all that remained to complete a synthesis of 2 was a diastereoselective aldol reaction, which we envisioned would proceed with concomitant lactone formation, and removal of the chiral auxiliary. Although an aldol reaction of a substrate similar to 11a had been reported,¹⁷ use of those conditions for 11a gave 12 in only 24% yield. After some experimentation, we discovered that enolization of 11a at a higher concentration (0.9 M) with n-Bu₂BOTf (1.5 equiv) in the presence of Hünig’s base (1.7 equiv), followed by the addition of myristyl aldehyde (1.2 equiv) and reacting for 15 h at 0 °C furnished the lactone 12 in 53% yield (95% based upon recovered starting material) and >95:5 dr (based on ¹H NMR of the crude reaction mixture) (Scheme 4). This aldol reaction was surprisingly sluggish, and despite repeated efforts, we were unable to identify conditions that would provide complete conversion. Hydrolysis of the oxazolidinone ring in 12 under standard conditions delivered (−)-dihydroprotolichesterenic acid (2) in 85% yield, thereby completing a four step, enantioselective synthesis of 2 that proceeded in 31% overall yield (56% based upon recovered starting material). The structure of synthetic 2 was confirmed by x-ray crystallography,¹⁸ and it also gave gave ¹H and ¹³C NMR spectra and optical rotation data consistent with those published.^16a

Scheme 4.

Synthesis of (−)-Dihydroprotolichesterenic Acid

In summary, we report the first diastereoselective conjugate addition of alkyl and aryl monoorganocuprates, Li[RCuI], to chiral fumarates, thereby providing a rapid and stereoselective entry to substituted succinate derivatives. The practical utility of the method was demonstrated by its application to the concise, enantioselective synthesis of (−)-dihydroprotolichesterinic acid (2).