Synthesis and stereochemical determination of an antiparasitic pseudo-aminal type monoterpene indole alkaloid

5-Nor stemmadenine alkaloids, isolated from the genus Tabernaemontana, display a range of bioactivity. 16-Hydroxy-16,22-dihydroapparicine, the active component of an extract from the Tabernaemontana sp. (dichotoma, elegans, and divaricate), exhibited potent antimalarial activity, representing the first such report of the antimalarial property of 5-nor stemmadenine alkaloids. We, therefore, decided to attempt the total synthesis of the compound to explore its antimalarial activity and investigate structure and bioactivity relationships. As a result, we completed the first total synthesis of 16-hydroxy-16,22-dihydroapparicine, by combining a phosphine-mediated cascade reaction, diastereoselective nucleophilic addition of 2-acylindole or methylketone via a Felkin–Anh transition state, and chirality transferring intramolecular Michael addition. We also clarified the absolute stereochemistries of the compound. Furthermore, we evaluated the activity of the synthetic compound, as well as that of some intermediates, all of which showed weak activity against chloroquine-resistant Plasmodium falciparum (K1 strain) malaria parasites.


Introduction
Naturally occurring chemicals represent a treasure trove of compounds which hold promise as the seeds of discovery for drugs and medicines and which may facilitate the elucidation of structure and function investigations of bioactivity [1]. Ō mura's research group at the Kitasato Institute is a global pioneer in the search for bioactive agents that may be of use in developing drugs and medicines to fight to infection and combat tropical diseases (such as the filariases, malaria, trypanosomiasis, etc.), all originating from microbial metabolites. At present, 483 new compounds have been discovered, 26 of which have become useful, widely used agents in human and animal health, including the ground-breaking avermectins [2].
Malaria is one of the world's worst health and socioeconomic problems, causing widespread death, disease, disability, and economic loss. Infection arises when a protozoal parasite of the Plasmodium genus is transmitted to humans via the bites of blood-feeding mosquitoes. Plasmodium falciparum parasites cause the most deadly form of the disease, which can cause death in a few days, especially if cerebral malaria develops. Generally, most deaths occur in children under 5 years old, although deaths have been reduced markedly by recent global initiatives to tackle the disease [3][4][5]. Commonly used drugs to combat malaria include quinine, chloroquine, mefloquine, halofantrine, and sulfadoxine/pyrimethamine (Fig. 1). However, drug resistance in parasites has usually developed quickly, rendering many of these drugs useless, preventing effective treatment and hindering disease elimination efforts. In 1972, Professor Tu Youyou discovered artemisinin to be the active ingredient in the plant Artemisia annua, which was commonly used in China to treat fever. Artemisinin derivatives became the most effective therapeutic drugs against malaria [6]. The World Health Organization (WHO) recommends artemisinin-based combination therapies (ACTs) for malaria treatment [7], a multidrug approach requiring the use of artemisinin together with other drugs to help offset the pace of drug resistance to artemisinin developing and spreading. ACTs are already compromised because the safety of artemisinin with regard to use during first trimester pregnancy is yet to be established and, worse, resistance to artemisinin derivatives developed almost immediately in locations along the Thai-Cambodian border [8][9][10][11]. Therefore, inexpensive and potent antimalarial drugs, especially those that have different modes of action, are urgently required on a probably continuing basis due to the ability of the malaria parasites to quickly develop drug resistance.
Many of the therapies currently in development use known antimalarial pharmacophores (e.g., aminoquinolines and/or peroxides), which have been chemically modified to overcome the failures of their predecessors [12]. Although these compounds have been important in the treatment of malaria, it would be highly advantageous to discover chemotypes with novel action mechanisms [13]. However, despite important advances in our understanding of the Plasmodium genome, the identification and validation of new drug targets have been challenging [14][15][16].
16-Hydroxy-16,22-dihydroapparicine (1), a known 5-nor stemmadenine alkaloid, was identified at the Kitasato Institute as a main component of a leaf's MeOH extract from the plant Tabernaemontana dichotoma, which displayed antimalarial properties. The potent antimalarial activity of the complex leaf extract against chloroquineresistant Plasmodium falciparum (K1 strain) parasites in vitro, and its moderate selectivity (against MRC-5 strain human cells) are summarized in Table 1. Natural compound 1 was originally isolated from a leaf of Tabernaemontana dichotoma in 1984 by the Verpoorte group [17] (Fig. 2). The relative structural determination of 1 was based on detailed NMR study, yet the absolute stereochemistry was not determined. As 1 has the potential to contain antimalarial activity, we decided to attempt the total synthesis of 1 to confirm its stereochemistry and investigate its antimalarial effect.

Proposed biosynthesis
The special architecture involved, embodying a 1-azabicyclo[4.2.2]decane, is probably the result of the C-5 tryptamine atom being excised from the alkaloid stemmadenine by a retro-Mannich reaction. Some in vitro transformations of stemmadenine-type to 5-nor stemmadenine-type alkaloids have provided further support for this biogenetic model, which the following summarizes.
To complete the total synthesis of (15S*,16S*)-16-hydroxy-16,22-dihydroapparicine (1), we designed a novel phosphineimine-mediated cascade reaction, without any isolated unstable intermediate (Scheme 4). The cascade reaction sequence was: (1) Staudinger reaction of an azide 21 with triphenylphosphine to generate phosphineimine intermediate 20 [62]; (2) intramolecular N-allylation of phosphineimine transformed into aminophosphinium 19 [63][64][65]; (3) aza-Wittig reaction of 19 with formaldehyde; and (4) intramolecular Mannich reaction; nucleophilic attack might be performed from the indole 3-position to iminium cation 18. We needed to solve two challenging issues. Firstly, the N-allylation of the phosphineimine group; phosphineimine has relatively high nucleophilicity, while the leaving group involves sufficient electrophilicity. Secondly, the formation of iminium cation using the aminophosphonium salt; there was no reported generation of iminium cation using the aminophosphonium salt and aldehyde via the aza-Wittig reaction. We found a solitary instance of the aminophosphonium salt with excess DMF to generating formamidinium salt [66]. However, the potential reactivity of the aminophosphonium salt has never been investigated. If we could overcome these challenges, an aminophosphonium salt (such as 19) could become a useful reactant for the aza-Wittig reaction. The key precursor 21 could be prepared from diastereoselective methylation of 2-acylindole 22 with completion of the To construct the C-15 stereocenter, we envisaged a remote stereocontrolled Michael reaction [72] of the a,bunsaturated carboxamide 25 with the crotonic acid derivative.

Structure determination
However, the spectral data of synthetic (±)-1 did not agree with that of naturally occurring 1 [17]. In particular, analysis of synthetic (±)-1, showed a ROESY relationship between H-18 or H-19 and 16-Me. Consequently, the relative stereochemistry of synthetic (±)-1 was determined to be a 15S*,16S*-configuration. Data of synthetic (±)-(15S*,16S*)-1 were then compared with naturally occurring compound (Table 2), with 1 H and 13 C NMR indicating differences of chemical shift (differences of all positions are shown in the experiment section). In 1 H NMR, 16-Me and H-6a,b signals were registered more than 0.20 ppm and, furthermore, the 13 C signals of the piperidine ring were greatly shifted from those seen in natural occurring 1. Therefore, we expected that the 16-Me group in naturally occurring 1 was on the opposite face for the tri-substituted exo-cyclic olefin. Accordingly, the relative stereochemistry was anticipated to be the 15S*,16R*-configuration.
To clarify the cascade reaction mechanism, we attempted the experiment outlined in Scheme 9. At first, to provide the corresponding primary amine, a Staudinger reaction of (±)-34 with PPh 3 was carried out under reflux condition to obtain the piperidine-indole (±)-37, without acidic activation of the 3-nitropyridinyl group. ESI massmonitoring of the first reaction allowed phosphineimine 35 to be easily generated from (±)-34 and PPh 3 without transformation into primary amine via solvolysis. In a timedependent change, phosphineimine smoothly converted to the aminophosphonium cation 36. Though the 3-nitropyridinyl group was a low electrophile, it was unnecessary for acidic activation. We inferred that the 1,3-allylic strain [92] was a key component, occurring via the tri-substituted    olefin. Therefore, the 3-nitropyridinyl group was located within close proximity of the phosphineimine group. Subsequent intramolecular Mannich reaction of piperidine-indole (±)-37 provided (±)-(15S*,16R*)-1 in 43 % yield, using formaldehyde and PPTS. We subsequently expected that the aza-Wittig reaction of 36 with formaldehyde could assist in generating the iminium cation precursor 38 in a cascade reaction.
In the next stage, we established the absolute stereochemistry of 1. In order to accomplish asymmetric total synthesis, we used the chiral methylketone 30 (Scheme 10), which could be supplied from azidobutyrolactone 39, including the appropriate functional groups. If 39 formed acetylbutyrolactone 40, its acetyl and ester moiety could be transformed into E-ethylidene and azido groups, respectively. Acetylbutyrolactone 40 was, therefore, our key intermediate, with synthetic manners for related compounds having already been reported by Smith's group and others [93][94][95][96]. We expected that 40 would involve a C-15 stereocenter being constructed by the intramolecular chirality transferring Michael reaction. We expected to perform via 5-exo-cyclization in the ketoester 41, which should be stereo-specifically constructed by the Baldwin rule [97] and Thorpe-Ingold effect [98,99]. Synthesis of the optically pure tri-substituted 40 began from commercially available (-)-(R)-methyl lactate 42, which, after with four steps of preparation, provided the ketoester (?)-41 in excellent yield (Scheme 11). With the optically pure (?)-41 in hand, we attempted the intramolecular chirality transferring Michael reaction [100][101][102][103][104]. Through extensive optimization, we found a suitable condition to provide (?)-40 in 91 % yield as a single diastereomer, and assignment of the relative stereochemistry was derived from the coupling constants and Scheme 11 Asymmetric synthesis of methylketone (?)-30 Scheme 12 End game of the total synthesis of (?)-(15S, 16R)-1 NOE correlation between a and c protons. The key factor of the intramolecular chirality transferring Michael reaction was the solvent's effect; polar solvent was stabilized to the anticipated transition state. The acetyl group of (?)-40 converted into the ethylidene moiety along with the separable Z-isomer. The tri-substituted olefin moiety was determined to be of E-configuration by NOE correlation. The configuration of the C-3 stereocenter of (-)-43 was determined after simple modification; hydrogenation of (-)-43 obtained a single diastereomer, and the stereochemistry was confirmed to be S-configuration by NOE and ROESY correlation. Compound (-)-43 was transformed into primary alcohol 44 by stepwise preparation; at first, selective hydrolysis of the ethyl ester group under basic condition generated carboxylic acid, followed by the corresponding acid anhydride. The furnished carboxylic anhydride was immediately reduced to the desired (-)-44 in 82 % yield over the two steps [105,106]. Subsequently, four steps functionalization provided the chiral methylketone (?)-30 in excellent yield without racemization. The optical purity of the (?)-30 (97 %ee) was confirmed by chiral HPLC analysis. The R-isomer of (-)-30 was prepared in the same asymmetric synthetic manner from (-)-(S)-methyl lactate.

Biological activity
Naturally occurring and synthetic compounds were tested for antimalarial activity against Plasmodium falciparum parasites (chloroquine-resistant K1 strain and chloroquinesusceptible FCR3 strain) and for cytotoxicity (against human MCR-5 cells) [107][108][109], in comparison with the first-line antimalarial artemisinin.

Conclusion
We achieved the first total synthesis of (?)-(15S,16R)-16hydroxy-16,22-dihydroapparicine (1) and the (-)-enantiomer and determined the absolute stereochemistry of naturally occurring 1. The synthesis involved a novel cascade reaction for efficient construction of the 1-azabicyclo[4.2.2]decane, including a pseudo-aminal moiety, via a Staudinger reaction, N-allylation, aza-Wittig reaction, and Mannich reaction. In addition, we developed a new method using diastereoselective 1,2-addition of methylketone, using N-TBSOM protecting the indole nucleophile and intramolecular chirality transferring Michael reaction with neighboring group participation. In particular, intramolecular chirality transferring Michael reaction proved to be a useful method for synthesis of the chiral tri-substituted butyrolactone. We established an effective enantioselective synthetic route for the production of pseudo-aminal alkaloids.
Synthetic (?)-(15S,16R)-1 exhibited moderate/weak antimalarial activity against chloroquine-resistant Plasmodium falciparum parasites and there is a possibility that the structurally unique compounds may be useful for the development of novel antimalarial drug candidates.