|Home | About | Journals | Submit | Contact Us | Français|
An enantioselective gram-scale synthesis of a key dihydroindolizine intermediate for the preparation of myrmicarin alkaloids is described. Key transformations in this convergent approach include a stereospecific palladium–catalyzed N-vinylation of a pyrrole with a vinyl triflate, a copper–catalyzed enantioselective conjugate reduction of a β-pyrrolyl enoate, and a regioselective Friedel-Crafts reaction. The synthesis of optically active and isomerically pure samples of (4aR)-myrmicarins 215A, 215B, and 217 in addition to their respective C4a-epimers is presented.
The myrmicarins are a family of structurally fascinating alkaloids isolated from the poison gland secretions of the African ant species Myrmicaria opaciventris (Figure 1).1 Despite significant isolation and purification challenges due to air and temperature sensitivity, elegant spectroscopic studies have revealed their molecular structures.2 The pyrroloindolizine core of myrmicarins 215A (1), 215B (2), and 217 (3) is a common structural motif within many of these alkaloids. Within the family, only the absolute stereochemistry of myrmicarin 237A (4) has been secured through an enantioselective synthesis.1a,3 Interestingly, the conversion of an unsaturated derivative of 4 to myrmicarin 217 (3) suggests the possible biogenesis of other myrmicarins from simpler indolizine derivatives.1b,4 The synthesis of (R)-myrmicarin 217 (3) and (R)-myrmicarin 215 as a mixture of olefin isomers has been reported starting with D-glutamic acid.5 The intriguing molecular structures of these poisonous alkaloids combined with challenges associated with their sensitivity provide an exciting arena to test and discover new methodologies for organic synthesis. Herein we describe a convergent synthesis of all naturally occurring tricyclic myrmicarin alkaloids employing an efficient approach to a pivotal optically active dihydroindolizine intermediate. The first preparation of isomerically pure samples of myrmicarins 215A and 215B is discussed. Key steps of the synthesis include an efficient palladium-catalyzed fragment coupling reaction, a copper catalyzed asymmetric conjugate reduction and a regioselective Friedel-Crafts reaction.
We envisioned utilization of the optically active dihydroindolizine 7 as a key intermediate for the preparation of myrmicarin alkaloids (Scheme 1). A regioselective Friedel-Crafts reaction of the pyrrole ring (C7a-alkylation) upon Brönsted–acid activation of the dimethoxyacetal 8 and elimination of methanol was expected to afford the bicyclic vinyl pyrrole 7. We planned to use a metal-catalyzed enantioselective conjugate reduction of the β-pyrrolyl enoate 9 to introduce the C4a-stereochemistry.6,7 A convergent synthesis of the pyrrolylenoate 9 was envisioned via a metal–catalyzed union of pyrrole 11 and readily available Z-vinyl triflate 10 (Scheme 1).
The synthesis of the required β-pyrrolylenoate 9 began with the Claisen condensation of the lithium enolate 12 and methyl 4-(dimethoxy)-butyrate (13)8 to give the β-ketoester 14 (Scheme 2). The lithium tert-butoxide additive was crucial to ensure rapid deprotonation of product 14, thus preventing the addition of a second equivalent of the lithium enolate 12.9 The trapping of the sodium enolate of β-ketoester 14 with Comins reagent (2-[N,N-Bis(trifluoromethylsulfonyl)amino]-5-chloropyridine, 15, Scheme 2) gave the vinyl triflate 10 in 82% yield with high Z-alkene selectivity (Z:E, >20:1).10 Early in our studies, we relied on the use of a copper–catalyzed N-vinylation of pyrroles for the synthesis of the requisite β-pyrrolyl enoates.7,11 However, due to difficulties associated with the synthesis of the necessary Z-β-iodoenoates12 containing acid sensitive functional groups, we turned our attention to the use of configurationally defined vinyl triflates.13 While copper-based catalyst systems were not effective, the use of palladium dibenzylideneacetone (Pd2(dba)3) and 2-dicyclohexyl-phosphino-2′,4′,6′-triisopropyl-1,1′-biphenyl (XPhos)14 provided a highly active catalyst system for the desired coupling reaction.15 Under optimal conditions, stereospecific coupling of pyrrole 11 and Z-vinyl triflate 10 provided the desired Z-β-pyrrolyl enoate 9 in 95% yield on multi–gram scale (Scheme 2). This represents the first example of a palladium–catalyzed N-vinylation of an azaheterocycle by a vinyl triflate. The convergent assembly of enoate 9 provides all of the carbons necessary for preparation of the tricyclic myrmicarins.
We planned to introduce the C4a-stereochemistry of the myrmicarin alkaloids through an enantioselective conjugate reduction of enoate 9 (Scheme 3). Recently, a variety of efficient catalytic methods have been reported for the synthesis of optically active β-amino acid derivatives.16 In particular, we were interested in the utilization of a recently reported methodology for the catalytic asymmetric reduction of β-azaheterocyclic enoates.7 Gratifyingly, the copper-catalyzed reduction of enoate 9 in the presence of (S)-2,2′-bis(diphenylphosphino)-1,1′-binaphthyl (BINAP) and the stoichiometric reductant polymethylhydrosiloxane (PMHS) proceeded efficiently to provide the optically active β-pyrrolyl ester 8. After minor optimization for the isolation of the substrate at hand, the reduction of enoate 9 proceeded to give the (R)-β-pyrrolyl ester 8 in 89% yield and 85% ee on a 2-gram scale (Scheme 3).17,18 It should be noted that upon completion of our total synthesis of myrmicarin 217 from the β-pyrrolyl ester 8 prepared using (R)-BINAP, we discovered that the conjugate reduction of enoate 9 had proceeded unexpectedly7 to give the (S)-β-pyrrolyl ester 8.19
With a multi-gram enantioselective synthesis of both enantiomers of β-pyrroly ester 8 at hand, we turned our attention to the synthesis of the dihydroindolizine 7, a key intermediate for the synthesis of the myrmicarin alkaloids (Scheme 1). Optimal conditions (acetone–acetic acid–water, 2:1:1, 40 °C) were identified for the quantitative conversion of the β-pyrrolyl ester 8 to the bicyclic vinyl pyrrole with good regioselectivity (>10:1) for the desired C7a-cyclization product 7 (Scheme 3).20 The formation of the desired dihydroindolizine 7 as the major regioisomer was confirmed by a 9.2% nOe between the C9-methylene of the propanoyl group and the pyrrole C2a-methine upon irradiation of the C2a-methine (Scheme 3). Compared to earlier studies employing N-alkylpyrrole derivatives, the use of the acyl pyrrole 11 in the preparation of 7 not only provided a more convergent synthesis, but the C8-carbonyl also afforded greater stability for isolation and storage of key intermediates towards myrmicarin alkaloids.21 Hydrogenation of the dihydroindolizine (R)-7 gave the tetrahydroindolizine (R)-16 in 96% yield. The corresponding tetrahydroindolizine (S)-16 was prepared with the same efficiency starting with (S)-8.
The introduction of the third ring of the pyrroloindolizine structure required the reduction of the C3-ester 16 to the corresponding C3-alcohol 17 (Scheme 4). Selective reduction of the C3-ester 16 was accomplished in a single operation by transiently protecting the C8-ketone as a silyl enol ether. In this sequence, conversion of the C8-ketone (R)-16 to the corresponding triisopropylsilyl enol ether derivative, reduction of the tbutyl ester by lithium aluminum hydride, and protodesilylation by addition of aqueous hydrochloric acid (pH 1) provided the desired alcohol (R)-17 in 91% yield. That this sequence could be executed in a single flask greatly simplified the preparation of alcohol 17.18 Treatment of the alcohol (R)-17 with a mixture of triphenylphosphine, iodine, and imidazole led to the conversion of the primary alcohol to the primary iodide (R)-18 in 91% yield. Based on prior literature precedent22 for cyclization of N-(ω-alkyl) radicals onto pyrroles containing an electron–withdrawing group, we first chose to explore free–radical cyclization strategies for the conversion of the iodide 18 to the tricyclic ketone 19. Accordingly, treatment of the iodide 18 with tri-nbutyltin hydride and 2,2′-azobisisobutyronitrile (AIBN) in toluene at reflux did provided tricycle 19, albeit in low yields and with poor mass recovery (ca. 30–40%). Further attempts using oxidative radical cyclization23 conditions did not lead to a significant improvement. Ultimately, the optimal conditions for the synthesis of tricyclic ketone 19 were found to involve an intramolecular C2-alkylation of the pyrrole nucleus by activation of the primary iodide 18.24 Under strictly anhydrous conditions, treatment of (R)-18 with silver tetrafluoroborate in a dichloromethane–benzene (3:1) solvent mixture at ambient temperature furnished the tricyclic ketone (R)-19 in 75% yield (Scheme 4). The identical sequence starting with ketoester (S)-16 furnished the corresponding alcohol (S)-17, primary iodide (S)-18, and the ketone (S)-19 in 87%, 86%, and 73% yield, respectively.
The enantiomerically enriched tricyclic ketone 19 provided expedient access to isomerically pure samples of tricyclic myrmicarins 1–3 (Scheme 4). As noted during isolation studies,1b myrmicarins 1–3 were found to be sensitive to air oxidation, giving the corresponding C6-C7 alkene derivatives25 followed by rapid decomposition.18 It should be noted that the propensity toward air oxidation in these pyrroloindolizine derivatives is higher in the absence of the C8-carbonyl. According to literature precedent, heating a mixture of ketone (R)-19 and lithium aluminum hydride in 1,4-dioxane gave (+)-(R)-myrmicarin 217 ((+)-3, [α]20D = +72.1 (c 0.050, CH2Cl2), lit.: [α]20D = +88 (c 1, CH2Cl2), Scheme 4) in 85% yield.5a The enantiomeric (−)-(R)-myrmicarin 217 ((−)-3, [α]20D = −67.4 (c 0.074, CH2Cl2)) was prepared via the same procedure in 99% yield.18
The stereospecific synthesis of the sensitive myrmicarin 215A (1) was achieved via a two–step sequence from the ketone 19 (Scheme 4). Dehydration of the C8–ketone (R)-19 using 2-chloro-3-ethylbenzoxazolium tetrafluoroborate (20)26 gave the corresponding C8-C9–alkyne, which proved to be a particularly sensitive intermediate.27 Successful isolation of pure material was only possible by an immediate isolation and purification of the alkyne at less than 30% conversion (12–19%, 75–80% based on recovered starting material). Partial reduction with Lindlar catalyst under an atmosphere of dihydrogen provided exclusively (−)-(R)-myrmicarin 215A ((−)-1, [α]20D = −53.8 (c 0.045, CH2Cl2), Scheme 4) in 74% yield. Similarly, the ketone (S)-19 was converted to the (+)-(S)-myrmicarin 215A ((+)-1, [α]20D = +49.8 (c 0.090, CH2Cl2)).
Synthesis of (+)-(R)-myrmicarin 215B commenced with the reduction of the ketone (R)-19 with lithium aluminum hydride to the corresponding C8–alcohol as a mixture of diastereomers (~1:1.5). Acid catalyzed dehydration then afforded isomerically pure (+)-(R)-myrmicarin 215B ((+)-2, [α]20D = +60.4 (c 0.044, CH2Cl2), Scheme 4) in 61% yield.28 Starting with ketone (S)-19 the corresponding (−)-(S)-myrmicarin 215B ((−)-2, [α]20D = −58.5 (c 0.085, CH2Cl2)) was prepared in 74% yield. Interestingly, (4aR)-myrmicarins 215A (1) and 215B (2) rotate plane–polarized light in opposite directions. While myrmicarins 215A (1) and 215B (2) were isolated and characterized as a mixture, our analysis of the isomerically pure samples has confirmed the previously reported 1H and 13C NMR assignments of the mixture.18
A convergent gram-scale synthesis of the key dihydroindolizine 7 for the synthesis of myrmicarin alkaloids in optically active form is described. The enantioselective synthesis of myrmicarins (−)-215A (1), (+)-215B (2), and (+)-217 (3) in addition to their respective enantiomers is described. Central to the success of this approach was the use of an efficient and stereospecific palladium–catalyzed fragment coupling of a Z-vinyl triflate and a pyrrole, a copper–catalyzed enantioselective conjugate reduction and two Friedel-Crafts reactions to form the pyrroloindolizine core of myrmicarins. The chemistry described here sets the stage for the synthesis of more complex members of this family of alkaloids and our studies in this area will be reported in due time.
M.M. is a Dale F. and Betty Ann Frey Damon Runyon Scholar supported by the Damon Runyon Cancer Research Foundation (DRS-39-04). M.M. is a Firmenich Assistant Professor of Chemistry. A.E.O. acknowledges a Robert T. Haslam Presidential Graduate Fellowship. We acknowledge generous financial support by the donors of the American Chemical Society Petroleum Research Fund, MIT, Amgen Inc., and NIH-NIGMS (GM074825).