Article pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Total Synthesis of the Caged Indole Alkaloid Arboridinine Enabled by aza-Prins and Metal-Mediated Cyclizations Pei Gan, Jennifer Pitzen, Pei Qu, and Scott A. Snyder* Department of Chemistry, University of Chicago, 5735 South Ellis Avenue, Chicago, Illinois 60637, United States S Supporting Information *
ABSTRACT: Although alkaloid natural products possess incredible diversity when considered broadly, certain domains are sometimes shared by several members, even from different sub-collections. Such homology can point to potential synthetic strategies. Herein, we highlight how such an analysis of the natural product arboridinine pinpointed two key elements of structural similarity that suggested the value of a metal-mediated 6-endo-dig cyclization to fashion its tetracyclic indolenine core, as well as the need to develop what could be considered a reversed polarity aza-Prins cyclization to deliver its tertiary allylic alcohol and final cage structure. The power of the latter design element is highlighted by several failures in achieving similar functional group patterning through more traditional aza-Prins and Mannich cyclization strategies. Overall, these operations fueled an inaugural 13-step racemic synthesis of the target; exploration of varied solutions for the enantioselective preparation of a key 7-membered indole-containing piece afforded a 16-step formal asymmetric solution.
■
INTRODUCTION In 2015, Kam and co-workers reported the isolation and characterization of a unique pentacyclic indole alkaloid from a plant of the Kopsia genus found in Malaysia that they named arboridinine (1, Scheme 1).1 This relatively scarce natural product (1.5 mg/kg plant material) possesses an unprecedented cage skeleton that includes multiple rings of varying individual size (6- and 7-membered, as well as an indolenine heterocycle). It also includes two all-carbon quaternary centers as well as a tertiary alcohol. When certain elements of this molecule are considered in isolation, however, arboridinine does share some structural features with other alkaloid targets. For instance, if one excludes certain linking bonds within its core bicyclic ring fusions, it has a tetracyclic indolenine structure also found in other Kopsia-derived alkaloids (such as 2),2 including members of the akuammiline family of alkaloids, typified here by rhazinoline (3),3 strictamine (4),4 and scholarisine A (5).5 Similarly, if one focuses instead on the patterning of the alkene, tertiary alcohol, and neighboring amine, similar connectivities can be found in members of the coccinellid (6),6 daphniphyllum (7),7 and aconitine (8)8 families, among many others (such as 9).9 Herein, we delineate how these shared domains provided the inspiration to enable a rapid synthesis of the challenging fused ring systems of arboridinine (1) in 13-steps racemically and 16-steps asymmetrically.
that shared connectivity suggested a potential construction via an aza-Prins reaction.10 In a formal sense considering native functionality, that merger in an ionic manner would be of an umpolung type,11 since it would require the carbonyl carbon of an α,β-unsaturated ketone to serve as a nucleophile in attacking an intermediate iminium ion; that event would likely be difficult to achieve. As such, we wondered whether an appropriately functionalized γ-oxygenated allylic silane could be a competent alternative, wherein the goal of group X on the oxygen atom would be to prevent the enol carbon from serving as a competing nucleophile for Mannich chemistry. To the best of our knowledge, such an aza-Prins reaction in this format has not been demonstrated, noting that cycloadditions and Sakurai allylations using such γ-oxygenated silanes are known,12 and reports of an umpolung (or reverse polarity) aza-Prins reaction using allylic stannanes and boranes with varied forms of iminium species generation exist.13 Indeed, in recent syntheses of other alkaloids with such domains, such as the Gin synthesis of neofinaconitine,14 alternate sequences, such as Baeyer− Villiger oxidations of ketones and subsequent ester hydrolysis, have installed similar tertiary alcohols. The deployment of standard allylic silanes as nucleophiles for aza-Prins chemistry in alkaloid synthesis is, of course, legion, with many elegant cascade-based constructions reported.15 As shown retrosynthetically in Scheme 2, application of this designed aza-Prins transform to arboridinine (1) led back to intermediate 10. This material we hoped could be readily prepared from 11 by a Michael addition of a methyl-based
■
RESULTS AND DISCUSSION 1. Overarching Synthetic Strategy and Retrosynthetic Analysis. As denoted by the coloration of the amino/alcohol/ alkene motif for the molecules in the lower half of Scheme 1, © XXXX American Chemical Society
Received: July 24, 2017
A
DOI: 10.1021/jacs.7b07724 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Article
Journal of the American Chemical Society
In the present context, we believed the value of pursuing this particular cyclization strategy would be three-fold: (1) it would allow rapid access to the core tetracycle 11 from relatively simple azepinoindole 12; (2) it would probe the power of this emerging cyclization reaction in a new format; and (3) it would require the synthesis of one of the 7-membered rings early in the sequence, hopefully in an asymmetric manner. One solution for this latter challenge, among several that we envisioned, included finding the means to differentiate the two esters within a precursor such as 13, a compound which we anticipated could arise from readily available Boc-protected tryptamine (14). 2. Development of Racemic and Enantioselective Syntheses of the Tricyclic Core and Its 7-Membered Ring. Our efforts to execute this plan began as shown in Scheme 3, starting with endeavors to effect both racemic and
Scheme 1. Structure of Arboridinine (1), Its Shared Patterning with Several Different Classes of Indole Alkaloids, and the Genesis of an Approach To Effect an azaPrins Reaction To Complete Its Core
Scheme 3. Syntheses of Aldehyde 16 in Racemic and Enantioenriched Formatsa
a Reagents and conditions: (a) bromomalonate 15 (1.0 equiv), Ir(ppy)3 (0.01 equiv), 2,6-lutidine (1.0 equiv), MeCN (1.0 M), blue LED, 23 °C, 24 h, 54% (87% brsm); (b) CH2Cl2/TFA (2/1), 23 °C, 30 min, then concentrate, K2CO3 (9.0 equiv), EtOH/MeOH (2/1, 0.05 M), 23 °C, 3 h, 83%; (c) LiAlH4 (10.0 equiv), THF (0.04 M), 0 °C, 0.5 h, then 45 °C, 4 h; (d) Boc2O (1.1 equiv), CH2Cl2 (0.04 M), 23 °C, 1 h, then DMSO (0.18 M), IBX (3.0 equiv), 23 °C, 18 h, 70% over two steps; (e) DIBAL-H (10 equiv), LiCl (10 equiv), THF, 0−23 °C, 14 h, 72%; (f) CuCl2 (0.1 equiv), 18 (0.1 equiv), BzCl (1.5 equiv), Et3N (1.1 equiv), CH2Cl2 (0.1 M), −78 °C, 14 h, 72%; (g) IBX (1.5 equiv), DMSO (0.15 M), 23 °C, 5 h; (h) Et3SiH (2.5 equiv), TFA (0.85 equiv), MeCN (0.1 M), 0−23 °C, 18 h, 68% over two steps, 96% ee; (i) K2CO3 (5.0 equiv), MeOH (0.1 M), CH2Cl2 (0.1 M), 23 °C, 0.75 h; (j) IBX (3.0 equiv), DMSO (0.1 M), 23 °C, 18 h, 71% over two steps. DMSO = dimethyl sulfoxide, IBX = 2-iodoxybenzoic acid, TFA = trifluoroacetic acid.
Scheme 2. Retrosynthetic Analysis of Arboridinine (1) Based on Two Key Cyclizations
enantiospecific preparations of one of the 7-membered rings of the target as part of the tricyclic unit defined in the retrosynthetic analysis delineated in the preceding section. Following precedent by Stephenson showing the ability to add malonyl radicals onto electron-rich aromatics,20 we were pleased to observe that photoirradiation of 15 in the presence of Boc-protected tryptamine (14) smoothly afforded the C-2 functionalized indole 13 in an isolated yield ranging from 46 to 54% depending on reaction scale; that yield range was 79−87% considering recovered 14. Three further steps involving deprotection/lactamization, reduction with LiAlH4, and then a one-pot reprotection/oxidation using Boc2O and IBX afforded a synthesis of the 7-membered ring within aldehyde 16 in 48 to 58% overall yield, delivering the final material as a racemate. This route proved quite scalable, and provided the material supplies for the remainder of the synthetic investigations. Identifying the means to effect an enantiospecific synthesis of 16 proved much more challenging. Initial investigations using
nucleophile and a subsequent series of functionalizations of its α,β-unsaturated ketone followed by Boc cleavage and iminium formation. This new precursor resembled the shared tetracyclic indolenine core of a number of indole alkaloids for which a range of creative approaches have been developed,16 including some by our group.16c,i Of those potential options, we anticipated that a metal-mediated 6-endo-dig cyclization of ynone 12 might be of particular value. Both we16i and a team led by Fujii and Ohno16g used distinct versions of such an alkyne-based cyclization to fashion the tetracyclic core of strictamine (4), while Taylor and co-workers developed a 5endo-dig variant to prepare spirocyclic indolenines17 and Wang identified an N-silyl indole variant for strictamine-type cores.18,19 B
DOI: 10.1021/jacs.7b07724 J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Article
Journal of the American Chemical Society Table 1. Screening of Conditions To Achieve the Mono-acetylation of Diol 17 with Enantiocontrol
entry
conditions
conversion (%)a,b
1 2 3 4 5 6 7 8 9 10 11 12
Candida cylindrica lipase, EtOAc, 40 °C, 48 h Pseudomonas cepacia lipase, EtOAc, 40 °C, 48 h Pseudomonas f luorescens lipase, EtOAc, 40 °C, 48 h pig liver esterase, EtOAc, 40 °C, 48 h porcine pancreatin, EtOAc, 40 °C, 48 h Candida antarctica lipase, EtOAc, 23 °C, 24 h Candida antarctica lipase, EtOAc, 40 °C, 216 h Candida antarctica lipase, EtOAc, 40 °C, 48 h Candida antarctica lipase, EtOAc, 80 °C, 24 h Candida antarctica lipase, THF, 40−60 °C, 48 h Candida antarctica lipase, CH2Cl2, 40−60 °C, 48 h Candida antarctica lipase, pH 7.4 phosphate buffer, 40−60 °C, 48 h
NR NR NR NR NR NR 39 10−20 50
