Download PDF
Synlett 2024; 35(05): 521-534
DOI: 10.1055/s-0042-1751498
account
Biomimetic Synthesis

Total Synthesis of Mavacuran Alkaloids via Bioinspired and Non-Bioinspired Strategies

Cyrille Kouklovsky 
a   Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO), Université Paris-Saclay, CNRS, Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay, France
,
Erwan Poupon  ∗
b   BioCIS, Université Paris-Saclay, CNRS, Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay, France
,
Laurent Evanno  ∗
b   BioCIS, Université Paris-Saclay, CNRS, Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay, France
,
Guillaume Vincent  ∗
a   Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO), Université Paris-Saclay, CNRS, Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay, France
› Author Affiliations

Financial support for this work was provided by the Agence Nationale de la Recherche (ANR) (ANR-15-CE29-0001, ‘Mount Indole’), the Ministère de l’Enseignement Supérieur et de la Recherche (MESRI) (ED 2MIB, Université Paris-Saclay), the Fondation pour le développement de la chimie des substances naturelles et ses applications; sous l’égide de l’Académie des Sciences, Université Paris-Saclay and the Centre National de la Recherche Scientifique (CNRS).
› Further Information
 
 PDF Download Permissions and Reprints All articles of this category


Abstract

In this account, we report our endeavors towards the total synthesis of the mavacuran alkaloids and some of their highly natural complex bis-indoles. Our studies started with the hemisynthesis of voacalgine A and bipleiophylline, made an excursion to a related family of monoterpene indole alkaloids (total synthesis of 17-nor-excelsinidine) and ended with the total syntheses of several mavacuran alkaloids (16-epi-pleiocarpamine, 16-hydroxymethylpleiocarpamine, taberdivarine H, normavacurine, C-mavacurine, C-profluorocurine, and C-fluorocurine) via a combination of bioinspired and non-bioinspired synthetic routes.

1 Introduction

2 Bioinspired Hemisynthesis of Voacalgine A and Bipleiophylline

3 Total Synthesis of the Mavacuran Alkaloids

4 Bioinspired Oxidative Cyclization of a Geissoschizine Ammonium Derivative to Form the N1–C16 Bond and the E Ring

5 Non-Bioinspired Michael Addition to Form the C15–C20 Bond and the E Ring

6 Conclusion

7 Epilogue


#

Key words

monoterpene indole alkaloids - bipleiophylline - pleiocarpamine - C-mavacurine - taberdivarine H - C-fluorocurine - biomimetic - oxidative coupling

Biographical Sketches

Zoom Image

Cyrille Kouklovsky completed his Ph.D. at Université Paris-Sud (France) in 1989 under the supervision of Prof. Yves Langlois before undertaking postdoctoral studies with Prof. Steven V. Ley at the University of Cambridge (UK). He is currently a full professor of organic chemistry at Université Paris-Saclay (France). His research interests include asymmetric dipolar cycloaddition reactions and their synthetic applications. He was the president of the Organic Chemistry Division of the French Chemical Society between 2015 and 2019.

Zoom Image

Erwan Poupon is a full professor of natural product chemistry and drugs of natural origin at Paris-Saclay University (France). He is particularly interested in understanding the intimate mechanisms involved in the biosynthetic pathways of natural products that can explain the ‘emergence’ of molecular complexity. His other interests include the anticipation and discovery of new natural products from plants, marine invertebrates, and micro-organisms, as well as natural-product-based drug design.

Zoom Image

Laurent Evanno received his Ph.D. in 2007 from Université Pierre et Marie Curie, Paris (France), working on total synthesis under the supervision of Dr. Bastien Nay at the ‘Muséum National d’Histoire Naturelle’. He then undertook postdoctoral research with Professor Petri Pihko at the Helsinki University of Technology – TKK (Finland) in 2008 and with Professor Janine Cossy at ESPCI – Paris Tech (France) in 2009. Since 2010, he has been an assistant professor at Paris-Saclay University (France) and his research interests encompass biomimetic synthesis.

Zoom Image

Guillaume Vincent graduated in 2002 from CPE Lyon (Ecole Supérieure de Chimie, Physique, Electronique de Lyon, France) with a one-year internship at Dupont Pharma (Wilmington, DE, USA) with Dr. Patrick Y. S. Lam. He obtained his master’s degree (2002) and Ph.D. (2005) from Université Claude Bernard Lyon-1 (France) with Prof. Marco A. Ciufolini. After two periods of postdoctoral studies with Prof. Robert M. Williams at Colorado State University (USA) and with Prof. Louis Fensterbank and Prof. Max Malacria at Université Pierre et Marie Curie Paris-6 (France), he was recruited as a CNRS researcher in 2007 (‘Chargé de Recherche’ and then ‘Directeur de Recherche’ since 2019) at the Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO) of Université Paris-Sud, which is now Université Paris-Saclay (France). In 2018, he was awarded the Jean-Marie Lehn Prize (Advanced Researcher Prize) from the Organic Chemistry Division of the French Chemical Society.

1

Introduction

In the framework of this cluster on biomimetic synthesis,[1] the objective of the present account towards the synthesis of mavacuran alkaloids[2] is to tell the story of the joint efforts of our two research teams at the Faculty of Pharmacy (BioCIS: Biomolécules, Conception, Isolement, Synthèse) and the Faculty of Science (ICMMO: Institut de Chimie Moléculaire et des Matériaux d’Orsay) of Université Paris-Saclay (formerly Université Paris-Sud). This project started with the objective of accomplishing the total synthesis of bipleiophylline (1),[3] which is a complex bis-indole alkaloid containing two units of pleiocarpamine (2),[4] a representative mavacuran alkaloid, connected via an aromatic spacer derived from 2,3-bishydroxybenzoic acid (3) (Figure [1]).

Zoom Image
Figure 1 Bipleiophylline and its biosynthetic precursor pleiocarpamine as our synthetic targets

The association of our two teams in reaching this goal was advantageous and highly complementary due to our respective expertise in bioinspired syntheses of indole alkaloids[5] and in the development of synthetic methodologies towards the benzofuroindoline unit[6] present in bipleiophylline. We were thrilled to be funded by the French Research National Agency (ANR) to tackle this ambitious target with a collaborative grant that we named ‘Mount Indole’, in reference to the front cover of an authoritative book by Hanessian and co-workers entitled ‘Design and Strategy in Organic Synthesis’ that depicted bipleiophylline at the top of a mountain.[7]

The mavacuran alkaloid sub-family[2] belongs to the monoterpene indole alkaloids which originate from strictosidine (6), the product of an enzymatic Pictet–Spengler reaction between tryptamine (4), the indolic part, and secologanin (5), the monoterpene part (Scheme [1]).[8] Among various biosynthetic pathways, the deglucosylation of strictosidine (6) and the subsequent reductive amination is a very important biosynthetic transformation that leads to the tetracyclic corynanthe skeleton of geissoschizine (7). This is a pivotal intermediate for several other sub-families of monoterpene indole alkaloids, for instance via divergent oxidative cyclization involving the C16-formyl ester leading to diverse pentacyclic structures. On the one hand, the coupling of the C16-formyl ester with the piperidinic C ring delivers either the sarpagan alkaloids 8 (C5–C16 coupling)[9] or the excelsinidine alkaloids 9 and 10 (N4–C16 coupling).[10] On the other hand, oxidative coupling of C16 at different positions of the indole moiety is also observed.[11] The strychnan alkaloids are the formal products of C2–C16 coupling and migration of the C3 carbon from C2 to the C7 position.[12] Alternatively, C7–C16 coupling yields the akuammilan alkaloids 12 and 13,[11] which have been postulated to be the biosynthetic precursors of the strychnan alkaloids via concomitant migrations of the C3 carbon from C2 to C7 and of the C16 carbon from C7 to C2.[11] Finally, the indolic nitrogen (N1) can also be involved in an oxidative cyclization yielding the mavacuran alkaloids 2 and 14 (N1–C16 coupling) which are of interest to us.[11] [13]

Zoom Image
Scheme 1 Biosynthesis of the mavacuran and other monoterpene indole alkaloids via divergent oxidative cyclizations of geissoschizine

Overall the N1–C16 bond is characteristic of the pentacyclic mavacuran skeleton, which encompass more than 20 natural products, some of which are presented in Figure [2]. It is interesting to note that this family of natural products was named after the isolation of one of its congeners, C-mavacurine (17), from a curare called mavacure used in Venezuela for the preparation of arrow poisons.[4a] [14]

The points of diversity in the structures are found mainly in the stereochemistry and nature of the oxygenated functional group borne by the C16 carbon [ester/carboxylate (2, 15, 18–23), primary alcohol (16, 17, 24), or both (14)], the N4-nitrogen [tertiary amine (2, 14–16, 20–23), N-methylammonium (17, 18, 24) or N-oxide (19)] or the substitution of the C2–C7 indolic bond [indoles (2, 14–19), reduced indoline (20) or oxidized indoline (21–24)] (Figure [2]).[2] [4a] [15]

Zoom Image
Figure 2 The mavacuran alkaloids

Fluorocarpamine (28) and C-fluorocurine (29) display a rearranged skeleton with a pseudo-indoxyl motif as the A/B rings and ring contraction of the C ring with migration of the C6 carbon from C7 to C2 via a pinacol rearrangement of the 2,7-dihydroxyindoline of respectively 2,7-dihydroxypleiocarpamine (23) and C-profluorocurine (24).[2] [4a] Some tetracyclic compounds are related to the mavacuran alkaloids due to the presence of the characteristic N1–C16 bond and E ring, but are missing one of the other rings.[2] Vinoxine (25) lacks the C ring via cleavage of the C6–C7 bond. Talbotine (26) and desformyl-talbotinic acid methyl ester (27) are missing the N4–C21 bond and thus the D ring.[2] [13a]

The pentacyclic mavacuran caged skeleton is highly strained which makes the C7 carbon of the indolic nucleus particularly reactive towards electrophiles. The resulting iminium at C2 could then be trapped by a nucleophile resulting in difunctionalization of the indole nucleus (Figure [3]). For this reason, pleiocarpamine (2) has a propensity to add to other monterpene indole alkaloids leading to bis-indole alkaloids such as villalstonine (30), hunterizeyline F (31), pycnanthinine (32), pleiocraline (33), plumocraline (34) and hunzylanine A (35).[2] Pleiocarpamine (2) can also react in nature with non-indolic electrophiles such as a pyrone-derived ortho-quinomethide to deliver pleiomaltinine (36) or an ortho-quinone leading to voacalgine A (37), and ultimately to bipleiophylline (1).[2]

Zoom Image
Figure 3 Pleiocarpamine-derived bis-indolic alkaloids and adducts

# 2

Bioinspired Hemisynthesis of Voacalgine A and Bipleiophylline

We tackled our endeavor to reach bipleiophylline, our synthetic summit, by making the final ascent from pleiocarpamine via voacalgine A. The reason being that the total synthesis of the strained mavacuran skeleton, and in particular pleiocarpamine, is a more difficult climb than it may seem. Fortunately, we were able to isolate almost 100 mg of pleiocarpamine as its formic acid salt from the bark of Alstonia balansae, with an isolation yield below 0.2%. It should be noted that the biomimetic hemisyntheses of the bis-indole natural products villastonine (30), macrocarpamine and plumocraline (34), and the pyrone-adduct pleiomaltinine (36) from natural pleiocarpamine (2) have already been reported.[16]

While we have previously developed several synthetic methods towards the synthesis of the benzofuroindoline moiety of bipleiophylline (1) and voacalgine A (37), none of them could be applied to the total synthesis of these natural products.[6] We therefore designed a new oxidative coupling between indoles and 2,3-bishydroxybenzoic acid (3) or its methyl ester 39 to yield respectively the isochromenoindoline 38 and benzofuroindoline 40 fragments of bipleiophylline (Scheme [2]).[17] [18] The application of this method to pleiocarpamine and 2,3-bishydroxybenzoic acid (3) led successively to the hemisyntheses of voacalgine A (37) and bipleiophylline (1).[17] This method proceeds via oxidation of the catechol of 3 with silver oxide to generate the corresponding ortho-quinone 41 onto which the nucleophilic indole of 2 (as its formate salt) could add to form a C–C bond, with subsequent interception of the iminium of 42 by the carboxylate yielding the isochromenoindoline of voacalgine A (37). Of note, before our synthesis of voacalgine, its structure was thought to contain a benzofuroindoline motif (‘the other half’ of bipleiophylline);[19] our work led us to revise its structure to 37 due to the chemoselectivity observed during the oxidative coupling and via 2D NMR. In the case of voacalgine A (37) and isochromenoindolines 38, HMBC correlations were observed between the same hydrogen of the aromatic platform and both the carbonyl and the C7 carbon. In contrast, these two carbons correlated with two different hydrogens of the aromatic platform for benzofuroindolines 40. In the second step of the synthesis, further oxidation with silver oxide in the presence of pleiocarpamine (2) (as its formate salt) led to addition of the latter to ortho-quinone 43. After the formation of a new C–C bond, the iminium 44 cyclized with the oxygen of the catechol to complete the formation of the benzofuroindoline of bipleiophylline. We thus reached the summit of our mountain starting from an already high altitude.

Zoom Image
Scheme 2 Bioinspired hemisyntheses of voacalgine A and bipleiophylline

# 3

Total Synthesis of the Mavacuran Alkaloids

In order to achieve a full ascent of our synthetic summit, it became necessary to climb the first part of the ascending route. In other words, to achieve the total synthesis of bipleiophylline, it was necessary to develop a total synthesis of pleiocarpamine, which had not been achieved when we started this project. The difficulty in constructing the mavacuran skeleton resides in its highly strained nature. The group of Sakai was the first to synthesize a mavacuran natural product through the hemisynthesis of 16-epi-pleiocarpamine (15) from geissoschizine.[20] Cleavage of the C3–N4 bond was necessary to give enough flexibility to the molecular backbone to construct the N1–C16 bond and the E ring via an intramolecular nucleophilic substitution on C16-chloroester 45. The C3–N4 bond was finally restored from 46 to complete the synthesis (Scheme [3]). A similar approach was adopted by Harley-Mason during the total synthesis of (±)-16-epi-pleiocarpamine (15), (±)-16-epi-pleiocarpamine N-oxide (19), (±)-normavacurine (16) and (±)-C-mavacurine (17).[21] [22] This first total synthesis of one natural member of the mavacurans was racemic and required more than 20 steps.

Boekelheide also started from a geissoschizine analog to access the mavacuran framework via formation of the N1–C16 bond through an oxidative rearrangement of the indole ring of (±)-47 into pseudoindoxyl (±)-48 followed by intramolecular opening of the epoxide to give the non-natural analog (±)-49 of C-fluorocurine.[23] Reduction of the C7-carbonyl and acidic treatment allowed C-ring expansion and restauration of the indole moiety. This racemic approach was rather straightforward with less than 10 steps to access (±)-19,20-dihydronormavacurine, which, however, is not a natural product (Scheme [3]).

Bosch described a notable approach in which the E ring and the N1–C16 and C2–C3 bonds were formed at an early stage via the addition of N-indolyl acetate 50 to pyridinium species 51.[24] (±)-2,7-Dihydropleiocarpamine (20) was finally synthesized via a photocyclization onto the C7 indolic position of α-chloroamide (±)-52 to form the C ring with a modest yield of 18% in this key step. (Scheme [3])

In this account, we will report our research enabling access to the mavacuran skeleton via two distinct strategies:

(1) The direct oxidative coupling of a C16 ester and the indolic N1-nitrogen and thus late stage E-ring formation of geissoschizine derivative 54 via quaternization of the N4-tertiary amine into an ammonium species to mask the reactivity of the latter and control the conformation of the geissochizine framework.[25] This diastereoselective approach, starting from tryptophan, allowed the total synthesis of enantiopure (+)-16-epi-pleiocarpamine (15), (+)-16-hydroxymethyl-pleiocarpamine (14) and (+)-taberdivarine H (18) (Scheme [3]). A related approach was reported at the same time by Takayama and co-workers via quaternization of N4 as an aminoborane and N1-H insertion with C16-diazoester (±)-53 to afford, in a racemic manner, (±)-16-epi-pleiocarpamine (15) and (±)-pleiocarpamine (2) as the minor product, as well as (±)-normavacurine (16) and (±)-C-mavacurine (17).[26]

(2) The stereoselective Michael addition of vinyl derivative 56 to N-indolyl acrylate (±)-55 and nucleophilic substitution to close the D ring at the end to synthesize (±)-taberdivarine H (18), (±)-16-epi-pleiocarpamine (15), (±)-normavacurine (16), (±)-C-mavacurine (17), (±)-C-profluorocurine (24) and (±)-C-fluorocurine (29) (Scheme [3]).[27]

These strategies have the advantages of being both competitive in terms of the numbers of steps and the efficiency for the key formation of the pentacyclic framework of this highly strained skeleton compared with the previous methods. Moreover, our initial approach is the first that allowed the total synthesis of any mavacuran alkaloid in enantiopure form.

Zoom Image
Scheme 3 Strategies towards the synthesis of the mavacuran alkaloids

# 4

Bioinspired Oxidative Cyclization of a Geissoschizine Ammonium Derivative to Form the N1–C16 Bond and the E Ring

The objective of our approach towards the mavacuran alkaloids relies on the postulated final step of the biosynthesis, i.e., the oxidative cyclization of geissoschizine to form the N1–C16 bond (Scheme [4]).

Zoom Image
Scheme 4 Desired N1–C16 bond and E-ring formation via oxidative cyclization of the geissoschizine template

To the best of our knowledge, oxidative cyclization of the full geissoschizine template has only been reported by Martin and co-workers via the diastereoselective formation of chloroindolenine 58 in 52% yield from desformylgeissoschizine (57) using tert-butyl hypochlorite in the presence of tin chloride, which is quite impressive considering the complexity of this skeletal reorganization (Scheme [5]).[28] Deprotonation of the ester of 58 induced cyclization to give the strychnan skeleton of 59 via formally the formation of the C2–C16 bond and migration of the C3 bond from C2 to C7 according to precedents from Kuehne then Massiot and then Rapoport.[29] The C7–C16 bond of the akuammilan could also be formed first before a skeletal reorganization leading to 59. This route is postulated among others as a biosynthetic pathway (see Scheme [1]).

Zoom Image
Scheme 5 Previous oxidative cyclization of the geissoschizine framework into the strychnan skeleton and skeletal interconversion between the akuammilan and strychnan frameworks

This hypothesis was supported by the fact that rearrangement of akuammiline (60) (akuammilan skeleton) into preakuammicine acetate (61) (strychnan skeleton) involving the interconversion of the C3 and C16 carbons from the C2 to C7 position[30] has been observed in the presence of boron trifluoride by one of our two research groups,[30a] and a similar transformation was also observed previously when using a strong base.[30`] [c] [d] Interestingly, the reverse bond migrations are also possible after saponification of preakuammicine acetate delivering rhazimol (62) (Scheme [5]). While this skeletal reorganization is possible in a laboratory, it does not mean that the biosynthesis of the strychnan alkaloids follows this pathway via the akuammilan template.

Nevertheless, an oxidative cyclization method has been developed by the group of Ma in the context of the total synthesis of the akuammilan alkaloids on simplified geissoschizine structures that are missing two rings, such as in 63 and 64 that led respectively to 65 and 66.[31] The reaction proceeds via deprotonation of both the NH indole (N1) and the malonate (C16), followed by the addition of iodine as an oxidant which led to the formation of the C7–C16 bond (Scheme [6]). It turned out that this reaction could be substrate dependent since the group of Zhu reported that tricyclic β-tetrahydrocarboline (±)-67 led in a chemodivergent manner to N1–C16 bond formation to give (±)-68.[32] In certain cases, this chemoselectivity could be controlled by the choice of the oxidant as we demonstrated with β-tetrahydrocarboline 69, which possesses an additional substituent in comparison to 67.[33] After formation of the bis-anion with LiHMDS, the addition of ferrocenium hexafluorophosphate led to formation of the C7–C16 bond to give 70, while employing phenyliodine(III) bis(trifluoroacetate) delivered the N1–C16 product 71, which could then be converted with TBAF into talbotine analog 72 as a mixture of diastereoisomers.

Zoom Image
Scheme 6 Indole-malonate oxidative cyclizations related to the akuammilan and mavacuran alkaloids

In this context, our ambition was to apply Ma’s oxidative conditions to geissoschizine itself in order to promote the N1–C16 bond formation and to obtain pleiocarpamine (2). In order to ensure a scalable and robust access to geissoschizine we identified that the combination of two known total syntheses of geissoschizine would be optimal.[34] According to Martin and co-workers, the diastereoselective Mannich reaction between silyl dienol ether 74 and chiral tryptophan-derived imine 73 yielded β-tetrahydrocarboline 75.[35] The only undescribed step of our synthesis is the allylation of 75 with iodoalkenyl-containing bromide 76 to generate intermediate 77 from Cook’s synthesis.[36] An intramolecular reductive Heck reaction delivered the tetracyclic geissoschizine framework with a 2:1 diastereoselectivity, from which removal of the benzyl ester yielded 16-desformylgeissoschizine (57) and then (+)-geissoschizine (7) itself via a formylation reaction (Scheme [7]).

Zoom Image
Scheme 7 Oxidative cyclizations of geissoschizine derivatives for the total synthesis of (–)-17-nor-excelsinidine, (+)-16-hydroxymethylpleiocarpamine, (+)-16-epi-pleiocarpamine and (+)-taberdivarine H

The stage was now set to test the key oxidative cyclization. After a few experiments, we found that upon double deprotonation with KHMDS and addition of I2 as the oxidant, an oxidative cyclization occurred. However, it did not lead to pleocarpamine (2) or even strictamine (13) as we had hoped via coupling of the C16-formyl ester and one of the positions of the indole nucleus. Instead, the C16-formyl ester reacted with the N4-tertiary amine with spontaneous desformylation leading to the excelsinidine skeleton 78. Compound 78 was saponified to give the zwitterionic natural product (–)-17-nor-excelsinidine (10), which was synthesized for the first time (Scheme [7]).[34]

At this stage, we reasoned that formation of the N1–C16 bond of pleiocarpamine could be facilitated by bringing in close proximity the indolic nitrogen N1 and the C16 carbon. Therefore, a ring contraction of chlorolactam 79 derived from 16-desformylgeissoschizine (57) was sought. However, (–)-17-nor-excselsinidine (10) was obtained once again via concomitant nucleophilic substitution of the C16-chlorolactam by the N4-tertiary amine and opening of the lactam upon treatment of 79 under basic conditions (Scheme [7]).[34]

We realized that the lone pair of the N4-tertiary amine was more reactive than the indole nucleus. Moreover, based on earlier conformational studies by Sakai and co-workers, geissoschizine (7) adopts a trans-conformation in which the N4-tertiary amine and the C16-formyl ester are in close proximity, which probably facilitates their oxidative coupling.[37] Taking all these considerations together, we acknowledged that in order to access the mavacuran skeleton, we needed to sequester the highly reactive lone pair of the N4-tertiary amine and to push the conformation of geissoschizine into the cis-conformation in which the C16-carbon and the indole would be spatially able to interact. It appeared to us that transforming the N4-tertiary amine into an ammonium species would fulfil both objectives since the team of Gaich proved that the N-(4-bromo)benzylammonium derivative of geissoschizine adopts a cis-conformation according to X-ray analysis.[38]

However, we were concerned that performing a double deprotonation on an already positively charged intermediate may result in decomposition. Nevertheless, we decided to pursue this strategy.

Upon treatment with para-methoxybenzyl bromide, 16-desformylgeissoschizine (57) delivered ammonium derivative 80, while the corresponding C16-malonate 54a could also be obtained from 57 via temporary protection of the indole nitrogen (Scheme [7]).[25]

To our delight and despite our initial concerns, double deprotonation with KHMDS followed by treatment with I2 as an oxidant allowed selective formation of the N1–C16 bond and the pentacyclic framework of the mavacuran alkaloids for the first time in our hands. A modest yield of 30% of cyclization product 81 was obtained from monoester 80 along with recovery of 52% of the starting material. Removal of the PMB group from 81 with BBr3 restored the tertiary amine and delivered (+)-16-epi-pleiocarpamine (15) in 25% yield.[25b] Oxidative coupling from the C16-malonate 54a was more efficient since a 76% yield of pentacyclic ammonium derivative 82a was obtained over the benzylation and oxidative coupling steps.[25a] After removal of the PMB group with BBr3, a Krapcho decarboxylation was performed which also delivered (+)-16-epi-pleiocarpamine (15) (Scheme [7]). The pseudo-symmetry of the malonate could be broken since the two esters were in different spatial environments induced by the shape of the mavacuran framework. Indeed, the ester lying on the convex face was more easily accessible than that protected by the concave face. Therefore, selective reduction of the ester on the convex face by DIBAL-H gave 16-formyl-pleiocarpamine (84), which is likely to be the biosynthetic precursor of 16-hydroxymethyl-pleiocarpamine (14) and pleiocarpamine (2). Further reduction of the aldehyde of 84 with NaBH4 generated (+)-16-hydroxymethyl-pleiocarpamine (14). Next, deprotonation of the alcohol of 14 with NaH and heating resulted in a desformylation reaction and enabled the isolation of (+)-16-epi-pleiocarpamine (15). Unfortunately, we were never able to isolate pleiocarpamine (2) itself. Under each set of conditions, its more stable epimer 16-epi-pleiocarpamine (15) was isolated. All attempts to epimerize 16-epi-pleiocarpamine into the postulated kinetic product pleiocarpamine failed.

Having noticed that several mavacuran alkaloids display both an N-methylammonium moiety and the stereochemistry of 16-epi-pleiocarpamine (compounds 17, 18, 24 and 29) (see Figure [2]), the quaternization was performed with methyl iodide to form N4-methylammonium derivative 54b. The oxidative cyclization smoothly delivered pentacyclic N-methylammonium derivative 82b and subsequent double saponification and decarboxylation produced zwitterionic (+)-taberdivarine H (18) (Scheme [7]).[25]


# 5

Non-Bioinspired Michael Addition to Form the C15–C20 Bond and the E Ring

While the bioinspired oxidative cyclization of geissoschizine to generate the N1–C16 bond was the first approach that allowed us to access the mavacuran framework, it was not the first strategy that we investigated. Actually, our first thought was to perform an intramolecular 1,4-addition of a vinyl halide to an N-indolyl acrylate 85, which unfortunately failed (Scheme [8]). Nevertheless, after the success of the biomimetic coupling, we went back to the Michael addition approach, but applied the intermolecular version with the addition of 56 to 55.[27]

Zoom Image
Scheme 8 Approaches for the 1,4-addition of a vinyl halide to an N-indolyl acrylate to form the C15–C20 bond and the D ring

The N-indolyl acrylate functions of 85 and 55 were installed via Pictet–Spengler reactions of N-substituted tryptamines 86–88 with aldehydes 89a,b containing respectively a keto-ester (89a) or a protected keto-ester (89b). The reaction of N4-allyltryptamine 86 and aldehyde 89a yielded hemiaminal (±)-90 after reaction of the indolic nitrogen and the carbonyl of the keto-ester. Dehydration with trifluoroacetic anhydride delivered the desired N-indolyl acrylate (±)-85. The corresponding N4-Me and N4-PMB Michael acceptors 55a,b were obtained via hydrolysis of the keto-ester acetals of Pictet–Spengler products (±)-91a,b (Scheme [9]). These two compounds and the synthetic route employed[39] are closely related to vinconate (OC-340, the N4-ethyl analogue of 55) that was in clinical trial for the treatment of cognitive disorders.[40]

Zoom Image
Scheme 9 Michael addition approaches for the total syntheses of (±)-taberdivarine H, (±)-16-epi-pleiocarpamine, (±)-normavacurine, (±)-C-mavacurine, (±)-C-profluorocurine and (±)-C-fluorocurine

As already mentioned, the intramolecular addition of the vinyl iodide of (±)-85 to the N-indolyl acrylate moiety failed, despite having precedents in the total synthesis of other families of monoterpene indole alkaloids.[36] [41] [42] This failure is probably due to a disfavored conformation leading most of the time to the dehalogenated product (±)-92, such as under reductive Heck or radical conditions. The generation of a vinyllithium intermediate via iodide/lithium exchange with n-butyllithium was also not productive for inducing the intramolecular Michael addition. Nevertheless, it was noted that in THF, an excess of n-butyllithium added in a 1,2-fashion to the ester, thus affording (±)-93, which is indeed not surprising knowing the textbook reactivity of organolithium reagents. More interestingly, when we switched THF to toluene as the solvent, we were very surprised to observe the 1,4-addition product (±)-94 of n-butyllithium to the N-indolyl acrylate moiety with a very high diastereoselectivity (Scheme [9]).[43] We postulated with the help of DFT studies that this selectivity in favor of the 1,4-addition in toluene is controlled by the formation of a complex A involving the N4-tertiary amine and an aggregate of the organolithium reagent.[44] After a few years focused on the successful bioinspired oxidative cyclization presented previously, we desired to publish this intriguing observation on the intermolecular 1,4-addition of n-butyllithium on our tetracyclic Michael acceptor. During the publication process, we realized that we could take advantage of this unexpected reactivity to design a revised retrosynthesis via the corresponding intermolecular 1,4-addition strategy (cf. Scheme [8]).

To this end, lithium/halogen exchange from vinyl iodide 56 with n-BuLi was conducted under carefully controlled conditions, and the resulting vinyllithium reagent, used in large excess, added to the Michael acceptors (Scheme [9]).[27] The temperature of this process is key to both ensuring efficient lithium/halogen exchange and to avoid decomposition of the thus generated vinyllithium intermediate. The tetracyclic compounds (±)-95a,b could be cyclized to give the pentacyclic framework of the mavacuran alkaloids as ammonium salts via formation of an allyl halide and heating to achieve the intramolecular nucleophilic substitution. In the case of the N4-methyl substrate (±)-95b, a final saponification delivered (±)-taberdivarine H (18).

As for the N4-PMB substrate (±)-95a, removal of the N-PMB group to restore the tertiary amine was effected with TFA and anisole to yield (±)-16-epi-pleiocarpamine (15). Following steps described by Harley-Mason[21a] and then Takayama,[26] the latter was converted into (±)-normavacurine (16) via reduction with LiAlH4, and then into (±)-C-mavacacurine (17) via treatment with methyl iodide. Note that we were not able to access this last natural product from the N-methylammonium salt arising from (±)-95b (Scheme [9]).

After this non-biomimetic approach to the mavacuran skeleton, the end-game to access C-fluorocurine (29) was bioinspired. We were thus able to perform dihydroxylation of the indole nucleus of (±)-C-mavacacurine (17) to give (±)-C-profluorocurine (24), which had never been obtained by total synthesis.[45] [46] Finally, biomimetic pinacol rearrangement mediated by hydrogen chloride in methanol via intermediate 96 delivered the pseudo-indoxyl moiety and the first total synthesis of C-fluorocurine (29) (Scheme [9]).[45] [46]


# 6

Conclusion

Our journey on the chemistry of the mavacurans has been accompanied by ups and downs. By performing the hemisynthesis of bipleiophylline, we reached the heights and our ‘Mount Indole’ before what we expected via the start of our ascent from an already advanced base camp, i.e., pleiocarpamine. However, climbing up to this base camp from the bottom was the most perilous part due to the highly strained nature of the mavacuran scaffold. Several strategies were envisioned and we finally discovered two routes towards the skeleton of the mavacuran alkaloids: (1) Biomimetic oxidative cyclization of a geissoschizine ammonium derivative to form the N1–C16 bond and the E ring, and (2) the non-bioinspired intermolecular and unorthodox 1,4-addition of a vinyllithium reagent to form the C15–C20 bond and then the D ring via an intramolecular nucleophilic substitution. The first approach allowed the first total synthesis in an enantioenriched form of any mavacuran alkaloid: (+)-16-epi-pleiocarpamine (15), (+)-16-hydroxymethyl-pleiocarpamine (14) and (+)-taberdivarine H (18) were obtained in respectively 15, 16 and 14, steps and 1.8%, 2.4% and 3.4% overall yields in their longest linear sequences. The length of the synthesis of 16-epi-pleiocarpamine was significantly shortened to 11 steps, albeit with a reduced overall yield of 1.1% using a monoester instead of a malonate for the key oxidative cyclization. The second approach represents the shortest total syntheses of (±)-16-epi-pleiocarpamine (15), (±)-normavacurine (16), (±)-C-mavacurine (17) and (±)-taberdivarine H (18), with respectively 9, 10, 11 and 10 steps and overall yields of 5.0%, 4.6%, 2.7% and 3.2% in the longest linear sequences. The first total syntheses of (±)-C-profluorocurine (24) and (±)-C-fluorocurine (29) were achieved in respectively 12 and 13 steps and 1.3% and 0.5% overall yields in the longest linear sequences.

Unfortunately, the epimerization at C16 is the last crevasse to cross to make the junction with the base camp: the total synthesis of pleiocarpamine. We are in the pursuit of epimerization conditions allowing (1) deprotonation of the hydrogen at C16 lying on the concave face which makes it difficult to reach, and (2) kinetic control during the reprotonation.


# 7

Epilogue

A few days after the submission of the first version of this personal account, the group of Ueda and Tokuyama published enantioselective total syntheses of pleiocarpamine, voacalgine A and bipleiophylline.[47] The authors overcame the challenge of making the pleiocarpamine skeleton with the required stereochemistry at C16, which was the final problem that we were unable to solve.

In a very original manner, they constructed the indole and the B ring as the last event (Scheme [10]). N-Cbz glutamine-derived 97 was converted in 5 steps into the Michael-acceptor-containing vinyl iodide 98. Related to their recent total synthesis of vinoxine,[42] the authors performed a radical-mediated intramolecular Michael addition of a vinyl radical onto the N-phenylamino-acrylate moiety to construct the C16–C20 bond and the E ring in 99. The required relative cis stereochemistry at C15 and C16 was fortunately obtained, albeit with significant isomerization of the alkylidene. The C ring of compound 101 was obtained via reductive amination with aldehyde 100 to form the N4–C6 bond, C2-lactam reduction and intramolecular aza-Henry reaction to form the C2–C7 bond. The B ring, the indole and the C7–C8 bond of (+)-pleiocarpamine (2) were constructed on gram scale via a palladium-catalyzed reaction from the corresponding α-nitroalkane 101.

Zoom Image
Scheme 10 Total syntheses of (+)-pleiocarpamine, (+)-voacalgine A and (+)-bipleiophylline by the group of Ueda and Tokuyama

Related to our previous hemisyntheses, oxidative coupling of (+)-pleiocarpamine (2) with 2,3-bishydroxybenzoic acid (3) delivered (+)-voacalgine A (37) and then (+)-bipleiophylline (1) in impressive yields. Instead of our use of silver oxide (Ag2O), the group of Ueda and Tokuyama performed the two successive oxidative biomimetic couplings by deploying iron octacarboxyphthalocyanine complex (FePc(CO2H)8 as the catalyst in the presence of oxygen (Scheme [10]), a process closely related to the catalytic system (iron phthalocyanine [FePc] and t-BuOOH) described by Chen and co-workers for the synthesis of isochromanoindolines.[18]


#
#

Conflict of Interest

The authors declare no conflict of interest.

Acknowledgment

We are in debt to the Ph.D. students and postdoctoral researchers that performed the synthetic work over almost a decade: Natacha Denizot, Maxime Jarret, David Lachkar, Audrey Mauger, Aurélien Tap, Sarah Benayad and Victor Turpin. We thank Prof. Vincent Gandon and Dr. Guillaume Bernadat for DFT studies. We thank Dr. Georges Massiot (Université de Reims and CNRS, France) and Prof. Luc Angenot (University of Liège, Belgium) for insightful discussions concerning the isolation and chemistry of the mavacuran alkaloids and their NMR analysis.


Corresponding Authors

Erwan Poupon
BioCIS, Université Paris-Saclay, CNRS
Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay
France   

Laurent Evanno
BioCIS, Université Paris-Saclay, CNRS
Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay
France   

Guillaume Vincent
Institut de Chimie Moléculaire et des Matériaux d’Orsay (ICMMO), Université Paris-Saclay, CNRS
Bâtiment Henri Moissan, 17 Avenue des Sciences, 91400 Orsay
France   

Publication History

Received: 14 July 2023

Accepted after revision: 23 August 2023

Article published online:
02 November 2023

© 2023. Thieme. All rights reserved

Georg Thieme Verlag KG
Rüdigerstraße 14, 70469 Stuttgart, Germany


   Permissions and Reprints All articles of this category
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Zoom Image
Figure 1 Bipleiophylline and its biosynthetic precursor pleiocarpamine as our synthetic targets
Zoom Image
Scheme 1 Biosynthesis of the mavacuran and other monoterpene indole alkaloids via divergent oxidative cyclizations of geissoschizine
Zoom Image
Figure 2 The mavacuran alkaloids
Zoom Image
Figure 3 Pleiocarpamine-derived bis-indolic alkaloids and adducts
Zoom Image
Scheme 2 Bioinspired hemisyntheses of voacalgine A and bipleiophylline
Zoom Image
Scheme 3 Strategies towards the synthesis of the mavacuran alkaloids
Zoom Image
Scheme 4 Desired N1–C16 bond and E-ring formation via oxidative cyclization of the geissoschizine template
Zoom Image
Scheme 5 Previous oxidative cyclization of the geissoschizine framework into the strychnan skeleton and skeletal interconversion between the akuammilan and strychnan frameworks
Zoom Image
Scheme 6 Indole-malonate oxidative cyclizations related to the akuammilan and mavacuran alkaloids
Zoom Image
Scheme 7 Oxidative cyclizations of geissoschizine derivatives for the total synthesis of (–)-17-nor-excelsinidine, (+)-16-hydroxymethylpleiocarpamine, (+)-16-epi-pleiocarpamine and (+)-taberdivarine H
Zoom Image
Scheme 8 Approaches for the 1,4-addition of a vinyl halide to an N-indolyl acrylate to form the C15–C20 bond and the D ring
Zoom Image
Scheme 9 Michael addition approaches for the total syntheses of (±)-taberdivarine H, (±)-16-epi-pleiocarpamine, (±)-normavacurine, (±)-C-mavacurine, (±)-C-profluorocurine and (±)-C-fluorocurine
Zoom Image
Scheme 10 Total syntheses of (+)-pleiocarpamine, (+)-voacalgine A and (+)-bipleiophylline by the group of Ueda and Tokuyama