Plants of genus Tabernaemontana, widely distributed species of Apocynaceae family, are often used as traditional folk medicine due to their abundant bioactivities, such as antitumor, hypotensive, and analgesic activities [1-5]. The broad bioactivities are related to the structurally diverse monoterpenoid indole alkaloids (MIAs) that are common secondary metabolites. These MIAs have emerged as attractive targets for synthesis owing to the diverse biological properties as well as unique and structurally complex molecular architectures [6-9]. There has been tremendous interest in the synthesis of monoterpene indole alkaloid natural products isolated from Tabernaemontana species. Recently, two MIA-type natural products have been isolated from plants of genus Tabernaemontana. Taberdicatine B (1) [10], a bridged molecule isolated from Tabernaemontana divaricata, consists 3a-hydroxyhexahydropyrrolo[2,3-b]indole (3-OH—HPI) [11-13] skeleton and a highly rigid 6/5/5/6/6-fused pentacyclic framework, bearing 4 stereogenic centers with 3 quanternary chiral centers. 3-OH—HPIs are privileged scaffolds in some representative MIAs such as Alsmaphorazine D [14], Leuconodine E [15], and Hunteracine [16] illustrated in Fig. 1. Tabernabovine B (2) [17], isolated from Tabernaemontana bovina, is a type of rare rigid cage-shaped MIA natural product with similar 3-OH—HPI moiety, complicated by multi-substituted hexacyclic ring system featuring 6 stereogenic centers, 3 of which are quanternary chiral centers, all of which posed challenges to the total synthesis. We anticipated that the characteristic structural features of 1 and 2 present significant synthetic challenges, including rigid and cage-like framework as well as multiple quaternary stereogenic centers. The intriguing structures of 1 and 2 have rendered them appealing synthetic molecules. Since these two natural products have the same HPI skeletons and similar diazabicyclic core structures, we sought to develop a unified synthetic strategy involving first construction of 3-OH—HPI skeleton, and assembly of diazabicycles through different cyclization methods from different side chains. Herein, we reported the first total synthesis of (+)-taberdicatine B (1) and (+)-tabernabovine B (2).

We embarked on expeditious total synthesis of (+)-taberdicatine B (1) and (+)-tabernabovine B (2), and speculated that 1 and 2 might be accessed from common advanced intermediates by varying the substituent types. With this design, the retrosynthetic analysis of 1 and 2 is shown in Scheme 1. We anticipated that the synthesis of 1 could be realized from reductive cyanation of 3, tracking back to 4 through deprotection/reductive amination. We rationalized that 2 might be synthesized from 6 through reductive cyclization to furnish the rigid cage-like framework. In turn, 6 was assumed to be prepared from Dieckmann condensation of 7. Specifically, the retrosynthetic analysis revealed that tetracyclic compound 5 could be a versatile and advanced precursor. Further disconnection of 4 and 7 led to intermediate 5 with slight difference of substituents. The access to enantiopure 5 could be achieved through asymmetric dearomatization of 8 [18,19], and vicinal quaternary stereogenic centers were constructed in one-pot manner. 8 could be accessible from C—H activation of tryptamine 9 [20,21]. When the R¹ is hydrogen and R² is Cbz group in 5, acylation/alkylation of intermediate 5 gave compound 4. When the R¹ is methoxy group and R² is Boc group in 5, acylation/alkylation of intermediate 5 would give compound 7. As such, two parallel synthetic pathways leading to similar intermediates can be envisaged.

As depicted in Scheme 2, our synthesis of (+)-taberdicatine B (1) commenced with the construction of D ring through the 2-alkylation of tryptamine with 4-bromobutanoate to provide functionalized 2,3-disubstituted indole. A palladium-catalyzed norbornene-mediated cascade C—H activation was successfully applied to 9a, delivering 2-substituted tryptamine 11a in 82% yield [20,21]. Treatment of compound 11a with K₂CO₃ afforded lactam 8a in 84% yield [22,23]. Chiral phosphoric acid S-TRIP catalyzed asymmetric dearomatization of tryptamine derivative 8a was conducted to construct HPI skeleton through asymmetric bromocyclization by using readily available DABCO-derived bromine salt, and the one-pot direct hydrolysis of the bromide group occurred smoothly, affording tertiary alcohol 12a in 87% yield with 92% ee [18,19]. With enantioriched 12a in hand, a subsequent TES protection of the hydroxy group in 12a worked well, delivering 5a in 97% yield. It is noteworthy that the above-mentioned routes allowed multigram-scale reactions, which is beneficial for rapid accumulation of advanced intermediate 5a.

Next, our attention was turned to the construction of final E ring. To this end, the introduction of the crucial quaternary stereocenter next to the carbonyl of lactam of 5a was first executed. One-pot installation of an ester group followed by alkylation of the resulting 1,3-dicarbonyl derivative with ethyl iodide, gave 13a as single diastereoisomer in 88% yield. The diastereoselectivity during the formation of the quanternary stereogenic center could be attributed to the steric hindrance. Since the Re face was blocked by the N-protecting group, ethyl iodide attack from Si face was favored, therefore furnishing 13a as single diastereoisomer. Selective and partial reduction of the ester group with DIBAL at −78 ℃ led to the corresponding aldehyde 4 in 84% yield [24]. 4 was then subjected to NCbz deprotection/reductive amination sequence under the treatment of Pd/C, affording 3 in 98% yield [25,26]. To our delight, the final six-membered E ring was successfully formed and the main skeleton of (+)-taberdicatine B (1) was constructed. Then, LAH reduction of the lactam of 3 gave the corresponding labile hemiaminal intermediate, which was employed immediately in the next step without purification. Subsequent cyanation of hemiaminal with TMSCN and boron trifluoride diethyl etherate provided nitrile 14 as single diastereoisomer in 76% overall yield from 3 [27-29]. Finally, the total synthesis of 1 was completed by treating 14 with K₂CO₃ and H₂O₂ [30,31], followed by one-pot desilylation of TES group, and 1 was obtained in 84% yield. Synthetic 1 exhibited identical ¹H and ¹³C NMR data to those reported for natural product (+)-taberdicatine B.

As summarized in Scheme 3, total synthesis of (+)-tabernabovine B (2) commenced with the elaboration of 5b, which could be prepared on a multigram scale from tryptamine 9b. The synthetic route began with 2-alkylation of the indole ring of tryptamine, delivering 11b in 84% yield. Due to the steric hindrance of 7‑methoxy substitution on the indole ring in comparison to 11a, the reactivity of 11b is weaker than that of 11a, therefore the synthetic route for D ring in 2 was different from the previous route explored in 1. Thus, we converted the ester group in 11b to carboxyl group first, which is more reactive in the presence of condenser to proceed lactamisation. As we expected, the acid intermediate obtained from the hydrolysis of the ester group was smoothly converted to 8b in the presence of the condenser EDCI. After conversion of 11b into 8b [32], S-TRIP catalyzed asymmetric dearomatization, bromide hydrolysis, and TES protection proceeded well to give 5b with excellent enantioselectivity and efficiency. Next, the same one-pot acylation/alkylation conditions were used in 5b to get 13b, however poor diastereoselectivity was obtained. Then, various bases were screened, and it was found that the alkylation step needs the use of extra base. When Cs₂CO₃ was used as base in the alkylation step, the quaternary stereocenter was formed with exclusive diastereoselectivity in 76% yields over two steps. It is reasonable to assume that due to the steric hindrance of methoxy group in comparison to 5a, the reactivity of 5b is different from that of 5a. Removal of Boc group of 13b by TFA, followed by N-alkylation of the resulting free amine with ethyl chloroacetate afforded 7 in 90% yield for two steps. Dieckmann condensation of 7 with LiHMDS gave β-keto ester 15 in 81% yield [33,34], and the relative configuration of 15 was determined by X-ray analysis. Thus, the assembly of seven-membered E ring was accomplished.

With 15 in hand, the subsequent construction of caged F ring was investigated. We first screened reduction conditions for the formation of β‑hydroxy ester (Table 1). In the presence of NaBH₄ (Table 1, entry 1), a pair of diastereomers 6a (desired, 24% yield) and 6b (57% yield) were mainly obtained with trace amount of 6c. The relative configurations of 6a-6c were elucidated separately through demethoy-6a (CCDC 2255376), demethoxy-6b derivative (CCDC 2255374) and demethoy-6c (CCDC 2255375) by X-ray crystallographic analysis. Since the stereochemistry of the carbon chiral center adjacent to ester functionality is essential for the following reductive cyclization to assemble the caged ring, more productive condition was needed. When NaBH₄ reduction with NH₄Cl as buffering reagent [35-37] was carried out, the diastereoselectivity did not increase (Table 1, entry 2). An attempted addition of metal salt such as MgCl₂ as chelating reagent [38] did not improve the selectivity (Table 1, entry 3). When Me₄NBH(OAc)₃ [39] was used as reductant, no reaction occurred and 15 was recovered (Table 1, entry 4). ⁿBu₄NBH₄ reduction [40] also did not give better result (Table 1, entry 5). When the reaction was perfomed in the presence of bulky reductants such as l-selectride [41-43] and LiAl(O^tBu)₃H [44], the reactions were complicated, and only undesired 6b was obtained (Table 1, entries 6 and 7). Ru-catalyzed transfer hydrogenation (Table 1, entry 8) [45-49] and Crabtree's catalyst (Table 1, entry 9) [50,51] failed to give the desired β‑hydroxy ester. After much experimentation on selective hydride reduction, an acceptable and reliable hydrogenation reaction condition was found. Hydrogenation of the β-keto ester 15 using H₂ over PtO₂ under mild condition allowed the generation of two diastereomers α‑hydroxy esters 6a in 46% yield and 6b in 42% yield (Table 1, entry 10) [52,53]. One of the resulting diastereomer 6b can be further transferred to 6a through hydroxy oxidation and ketone reduction, which implied that β-keto ester 15 underwent enolization when the hydrogenation occurred. We also conducted an extensive evaluation of reaction conditions to promote epimerization of 6b to the desired stereo-configuration by the enolization/kinetic protonation sequence. Despite screening for bases, reaction times, and temperatures, we were unable to achieve the epimerization of 6b. In almost all cases, 6b was recovered, and no other products were generated. By treating 6b with KHMDS and then quenching with D₂O, it was found by NMR that the α-H of the ester group was partially deuterated. We assumed that the enolization of the ester could occur under the reaction condition, but the proton was still subsequently attacked from the side with relatively small steric hindrance, affording configuration-preserving starting materials. The failure of enolization of 6b under strongly basic conditions by the kinetic protonation prompted us to evaluate thermodynamic conditions for the epimerization. A wide range of bases were examined, such as DBU, ^tBuOK and NaOEt, however all conditions lead to decomposition or unreacted starting material. Since 6b can be transferred to 6a, the condition using H₂ over PtO₂ (Table 1, entry 10) was applied to synthesize 6a. Reductive cyclization of 6a with LAH afforded cage-like skeleton 16 in 76% yield [54,55]. As a result, the crucial caged F ring was successfully formed via ester reduction/amide semireduction/cyclization sequence. We also investigated reduction of 15 with LAH to get 16 via ketone reduction/ester reduction/amide semireduction/cyclization sequence directly, however complex reaction mixture was obtained along with 10% yield of desired 16. Since the stereochemistry of the carbon chiral center adjacent to hydroxy functionality of 16 was opposite to the natural product, we planned to invert the stereochemistry by DMP oxidation and NaBH₄ reduction. The expected epimerization occurred and afforded 17 in 73% yield over two steps. The total synthesis was wrapped up with desilylation of TES group with TBAF, and thereby (+)-tabernabovine B (2) was obtained in 98% yield. The ¹H and ¹³C NMR data of synthetic 2 was in agreement with those reported for natural product (+)-tabernabovine B.

In conclusion, we have developed efficient approaches for the first total synthesis of (+)-taberdicatine B (10 steps, 26.9% overall yield) and (+)-tabernabovine B (15 steps, 7.3% overall yield) from tryptamine derivatives. The key steps involved one-pot asymmetric bromocyclization/hydrolysis for the assembly of HPI skeleton, Dieckmann condensation to form β-keto ester for the assembly of seven-membered ring, and an ester reduction/amide semireduction/cyclization sequence for the formation of the cage-like framework. The prominent feature of this synthetic strategy is that (+)-taberdicatine B and (+)-tabernabovine B were accessed from common advanced intermediates by varying the substituents. Thus, the strategy could be applied to the synthesis of natural product analogues for medicinal investigation to explore potential biological activities, and is expected to be shown in total syntheses of other MIAs containing 3-OH—HPI skeleton.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

We thank the Qingdao Marine Science and Technology Center (No. 2022QNLM030003-2), the Fundamental Research Funds for the Central Universities, Taishan Scholar Program of Shandong Province (No. tsqn202103152), National Natural Science Foundation of China (No. 22171251) for financial support.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109816.