Total Synthesis of Decahydroquinoline Poison Frog Alkaloids ent-cis-195A and cis-211A

The total synthesis of two decahydroquinoline poison frog alkaloids ent-cis-195A and cis-211A were achieved in 16 steps (38% overall yield) and 19 steps (31% overall yield), respectively, starting from known compound 1. Both alkaloids were synthesized from the common key intermediate 11 in a divergent fashion, and the absolute stereochemistry of natural cis-211A was determined to be 2R, 4aR, 5R, 6S, and 8aS. Interestingly, the absolute configuration of the parent decahydroquinoline nuclei of cis-211A was the mirror image of that of cis-195A, although both alkaloids were isolated from the same poison frog species, Oophaga (Dendrobates) pumilio, from Panama.


Introduction
The skin extracts of Neotropical poison frogs contain a variety of lipophilic alkaloids, and over 800 alkaloids have been isolated or detected to date [1]. Many of these alkaloids show remarkable biological activities on the nervous system such as nicotinic acetylcholine receptors [2,3]. Methods for chemical synthesis of poison frog alkaloids are needed to investigate the biological activities of poison frog alkaloids, as only minute amounts of natural alkaloids can be obtained from skin extracts [1]. Decahydroquinolines are a relatively large subgroup of poison frog alkaloids, and over 50 types have been detected. Among them, the alkaloid cis-195A is the parent member of this class that was originally isolated from a Panamanian population of Oophaga (Dendrobates) pumilio in 1969 [4]. The structure and isolated from a Panamanian population of Oophaga (Dendrobates) pum structure and absolute configuration of cis-195A were determined b raphy, and several total syntheses have also been reported for this com alkaloid cis-211A was isolated from skin extracts of the same species o [23]. However, no total synthesis of this alkaloid has been reported, a figuration remains unknown to date ( Figure 1). As part of a program d the synthesis of poison frog alkaloids [24][25][26][27][28][29][30][31][32][33][34][35][36][37][38], herein, we report the to cis-195A and cis-211A. Both syntheses proceed via the common and k The synthesis of cis-211A also enabled the determination of its absolu

Results and Discussion
Hydrogenation of known allyl derivative 1 [39] provided the este verted to enaminoester 4 via thiophenyl derivative 3. The Michael-type reaction to 4 [25] gave adduct 5 as a single isomer in excellent yield. T was elongated by the Arndt-Eistert reaction sequence to afford the h which was transformed into the methyl ketone 8 via the correspondin Lemieux-Johnson oxidation of 8 provided the aldehyde 9, which was tion by treatment with DBU in refluxing benzene to yield the cis-fu without generating the trans-fused enone 10t. Selective formation of 1 the preferential formation of conformer A in the starting material 9 ow Thus, epimerization at the 3-position of 9 occurred first, and then cy to provide the enone 10c as shown in Figure 2. With enone 10c in han for the divergent synthesis of ent-cis-195A and cis-211A. The conjuga to 10c with Me2CuLi followed by treatment of the resulting enolate wit agent [40] afforded the common and key intermediate 11 (Scheme 1).

Results and Discussion
Hydrogenation of known allyl derivative 1 [39] provided the ester 2, which was converted to enaminoester 4 via thiophenyl derivative 3. The Michael-type conjugate addition reaction to 4 [25] gave adduct 5 as a single isomer in excellent yield. The ester moiety of 5 was elongated by the Arndt-Eistert reaction sequence to afford the homologated ester 6, which was transformed into the methyl ketone 8 via the corresponding Weinreb amide 7. Lemieux-Johnson oxidation of 8 provided the aldehyde 9, which was subjected to cyclization by treatment with DBU in refluxing benzene to yield the cis-fused enone 10c [25] without generating the trans-fused enone 10t. Selective formation of 10c was explained by the preferential formation of conformer A in the starting material 9 owing to the A 1,3 strain. Thus, epimerization at the 3-position of 9 occurred first, and then cyclization proceeded to provide the enone 10c as shown in Figure 2. With enone 10c in hand, the stage was set for the divergent synthesis of ent-cis-195A and cis-211A. The conjugate addition reaction to 10c with Me 2 CuLi followed by treatment of the resulting enolate with Comins' triflating agent [40] afforded the common and key intermediate 11 (Scheme 1).  Finally, global hydrogenation of 11 and deprotection of the methyl carbamate moiety in 12 using TMSI in CHCl3 at 50 °C provided ent-cis-195A, as shown in Scheme 2. The 1 H and 13 C-NMR spectra of synthetic ent-cis-195A were in good agreement with those reported in the literature [22].  Finally, global hydrogenation of 11 and deprotection of the methyl carbamate moiety in 12 using TMSI in CHCl3 at 50 °C provided ent-cis-195A, as shown in Scheme 2. The 1 H and 13 C-NMR spectra of synthetic ent-cis-195A were in good agreement with those reported in the literature [22]. Finally, global hydrogenation of 11 and deprotection of the methyl carbamate moiety in 12 using TMSI in CHCl 3 at 50 • C provided ent-cis-195A, as shown in Scheme 2. The 1 H and 13 C-NMR spectra of synthetic ent-cis-195A were in good agreement with those reported in the literature [22].  Finally, global hydrogenation of 11 and deprotection of the methyl carbamate moiety in 12 using TMSI in CHCl3 at 50 °C provided ent-cis-195A, as shown in Scheme 2. The 1 H and 13 C-NMR spectra of synthetic ent-cis-195A were in good agreement with those reported in the literature [22]. The enol triflate 11 was converted to olefin 13 by palladium-catalyzed reduction. Epoxidation of 13 by mCPBA proceeded smoothly to give the epoxide 14, unfortunately, as a 1:1 mixture of epoxide 14. Hydroboration of 13 using BH 3 -SMe 2 in toluene provided the alcohol 15 and the mixture of alcohols 16 and 17. The structure of 15 was determined by NOESY. The NOESY experiments of 15 revealed a syn relationship between the methyl group at C-5 and H-6 based on the NOESY correlations from H 3 -5 to H-6. However, the separation of 16 and 17 was difficult at this stage. For completion of the synthesis of cis-211A, inversion of the hydroxyl group in 15 was necessary. For this purpose, we subjected 15 to the Mitsunobu reaction; however, all attempts failed and resulted in the recovery of 15. Next, we examined hydroxyl inversion of 15 via ketone 19. Any oxidations of 15 using Swern, a SO 3 -pyridine complex, PCC, PDC, DMP, or TPAP were not successful, and the starting material was recovered. Only oxidation using AZADOL ® [41] proceeded smoothly to yield the desired ketone 19 in good yield. In addition, the AZADOL ® oxidation of the mixture of 16 and 17 afforded the ketones 19 and 20 in 34% and 66% yield, respectively, which could be easily separated by column chromatography. Thus, we succeeded to obtain the ketone 19 from 13 in 64% (49% + 15%) overall yield, as shown in Scheme 3. The conformation of ketone 19 is depicted in 19-A. The reduction of 19 from the concave face was needed to obtain the desired alcohol 16. We expected that the reduction of 19 would proceed from the concave face, as shown in 19-A, because of the steric hindrance of the α-axial methyl group. However, the use of a small reducing agent like NaBH 4 or LiAlH 4 reduced 19 from the convex face to afford 15 as the sole product. To secure the reduction from the concave face, we tried the large reducing agent L-Selectride ® ; however, the reduction did not proceed, and only ketone 19 was recovered. Fortunately, Super-Hydride ® , a moderately sized reducing agent, was the best match for this substrate, and the reduction proceeded from the concave face to provide the desired alcohol 16 as the major product (16:15 = 9:1). The final deprotection of the urethane moiety in 16 was also troublesome. First, we applied the same reaction conditions used for ent-cis-195A (TMSI in refluxing CHCl 3 ) to cleave the methyl carbamate; however, the reaction did not proceed. Then, other reaction conditions, such as the use of n-PrSLi/HMPA or KOH/i-PrOH in a sealed tube at 130 • C, resulted in the recovery of the starting material. Finally, we used TMSI in refluxing MeCN, and the reaction proceeded cleanly to yield cis-211A, as shown in Scheme 3. The 1 H-and 13 C-NMR spectra of synthetic cis-211A were in good agreement with the reported values [23]. The absolute stereochemistry of natural cis-211A was determined unambiguously by the present synthesis to be 2R, 4aR, 5R, 6S, and 8aS by comparison of the optical rotation of synthetic cis-211A ([α]D 25 −11.5 (c 0.2, CHCl3)) with the reported value ([α]D −11.7 (c 1.0, CHCl3)) [23]. Interestingly, cis-195A and cis-211A were both isolated from the same poison frog, Oophaga (Dendrobates) pumilio (Dendrobatidae) from Panama; however, the absolute stereochemistry of the parent decahydroquinoline nuclei of The 1 H-and 13 C-NMR spectra of synthetic cis-211A were in good agreement with the reported values [23]. The absolute stereochemistry of natural cis-211A was determined unambiguously by the present synthesis to be 2R, 4aR, 5R, 6S, and 8aS by comparison of the optical rotation of synthetic cis-211A ([α] D 25 −11.5 (c 0.2, CHCl 3 )) with the reported value ([α] D −11.7 (c 1.0, CHCl 3 )) [23]. Interestingly, cis-195A and cis-211A were both isolated from the same poison frog, Oophaga (Dendrobates) pumilio (Dendrobatidae) from Panama; however, the absolute stereochemistry of the parent decahydroquinoline nuclei of cis-195A is opposite to that of cis-211A. The NMR spectra ( 1 H-NMR, 13 C-NMR) of all synthesized compounds are listed in Supplementary Materials.
To further investigate the effect of the stereochemistry of the hydroxyl group at the 6-position on the inhibitory activity against nicotinic acetylcholine (ACh) receptors, we also synthesized 6-epi-211A by deprotection of the methoxycarbonyl group in 15, as shown in Scheme 4. The 1 H-and 13 C-NMR spectra of synthetic cis-211A were in good agreement with the reported values [23]. The absolute stereochemistry of natural cis-211A was determined unambiguously by the present synthesis to be 2R, 4aR, 5R, 6S, and 8aS by comparison of the optical rotation of synthetic cis-211A ([α]D 25 −11.5 (c 0.2, CHCl3)) with the reported value ([α]D −11.7 (c 1.0, CHCl3)) [23]. Interestingly, cis-195A and cis-211A were both isolated from the same poison frog, Oophaga (Dendrobates) pumilio (Dendrobatidae) from Panama; however, the absolute stereochemistry of the parent decahydroquinoline nuclei of cis-195A is opposite to that of cis-211A. The NMR spectra ( 1 H-NMR, 13 C-NMR) of all synthesized compounds are listed in Supplementary Materials.
To further investigate the effect of the stereochemistry of the hydroxyl group at the 6-position on the inhibitory activity against nicotinic acetylcholine (ACh) receptors, we also synthesized 6-epi-211A by deprotection of the methoxycarbonyl group in 15, as shown in Scheme 4. Nicotinic ACh receptors are ligand-gated cation channels [42,43]. Homomeric α7and heteromeric α4β2-pentamers are the major subtypes of nicotinic receptors found in the central nervous system [44]. It has been reported that (−)-cis-195A, a natural cisdecahydroquinoline alkaloid (formerly referred to as Pumiliotoxin C), blocks ganglionic nicotinic ACh receptors in pheochromocytoma PC12 cells [45] and that the synthetic analog (+)-cis-195A is more potent than (−)-cis-195A at inhibiting nicotinic receptor activity [3]. Here, we examined the effects of ent-cis-195A, cis-211A, and 6-epi-211A on α7and α4β2nicotinic ACh receptors ectopically expressed in Xenopus oocytes. If the criterion for partial inhibition by an alkaloid is defined as a ≥ 20% decrease in the peak amplitude of AChelicited currents, then ent-cis-195A (1-10 µM) showed no apparent inhibitory effects on α7and α4β2-receptor-mediated currents ( Figure 3A,B). cis-211A and 6-epi-211A at 10 µM partially inhibited α7-receptor-mediated currents by 38% and 31%, respectively, while both alkaloids at 10 µM showed negligible effects on α4β2-receptor-mediated currents ( Figure 3C-F). Analysis of their structure-activity relationship suggested that the 6-hydroxy moiety of cis-211A and 6-epi-211A might contribute to the partial blockade of α7-nicotinic ACh receptors. The ligand-binding assays showed that none of these alkaloids affected [ 3 H]nicotine and [ 3 H]methyllycaconitine binding to rat whole brain membranes (data not shown). Therefore, cis-211A and 6-epi-211A were believed to act as noncompetitive blockers of α7-nicotinic receptors, although they were less potent and not as highly selective.
Neuronal nicotinic ACh receptors play a role in cerebral and retinal physiology. The blood-brain barrier (BBB) and inner blood-retinal barrier (BRB) directly segregate the brain and retina, respectively, from the circulating blood. It has been reported that putative nicotine-and verapamil-sensitive cationic drug transport systems at the BBB and inner BRB, respectively, are involved in the facilitative distribution of their substrates to the central nervous system [46,47]. To evaluate the recognition of ent-cis-195A and cis-211A as substrates for these cationic drug transport systems, we performed an inhibition study using conditionally immortalized rat BBB and inner BRB model cells, known as TR-BBB13 and TR-iBRB2 cells [48,49]. As shown in Table 1, ent-cis-195A exhibited an inhibitory effect on [ 3 H]nicotine transport into TR-BBB13 cells and [ 3 H]verapamil transport into TR-iBRB2 cells by more than 40%. In addition, the presence of cis-211A significantly attenuated [ 3 H]nicotine and [ 3 H]verapamil uptake by TR-BBB13 and TR-iBRB2 cells, respectively, by at least 29%. These results suggest that ent-cis-195A and cis-211A are recognized by the cationic drug transport systems at the BBB and inner BRB. It is possible that these derivatives reach the brain and retina via the cationic drug transport systems and show neuronal effects in the CNS.
α7-and α4β2-receptor-mediated currents ( Figure 3A,B). cis-211A and 6-epi-211A at 10 μM partially inhibited α7-receptor-mediated currents by 38% and 31%, respectively, while both alkaloids at 10 μM showed negligible effects on α4β2-receptor-mediated currents ( Figure 3C-F). Analysis of their structure-activity relationship suggested that the 6-hydroxy moiety of cis-211A and 6-epi-211A might contribute to the partial blockade of α7nicotinic ACh receptors. The ligand-binding assays showed that none of these alkaloids affected [ 3 H]nicotine and [ 3 H]methyllycaconitine binding to rat whole brain membranes (data not shown). Therefore, cis-211A and 6-epi-211A were believed to act as noncompetitive blockers of α7-nicotinic receptors, although they were less potent and not as highly selective. Neuronal nicotinic ACh receptors play a role in cerebral and retinal physiology. The blood-brain barrier (BBB) and inner blood-retinal barrier (BRB) directly segregate the brain and retina, respectively, from the circulating blood. It has been reported that putative nicotine-and verapamil-sensitive cationic drug transport systems at the BBB and inner BRB, respectively, are involved in the facilitative distribution of their substrates to the central nervous system [46][47]. To evaluate the recognition of ent-cis-195A and cis-211A as substrates for these cationic drug transport systems, we performed an inhibition study using conditionally immortalized rat BBB and inner BRB model cells, known as TR-BBB13 and TR-iBRB2 cells [48,49]. As shown in Table 1, ent-cis-195A exhibited an inhibitory effect on [ 3 H]nicotine transport into TR-BBB13 cells and [ 3 H]verapamil transport into TR-iBRB2 cells by more than 40%. In addition, the presence of cis-211A significantly attenuated [ 3 H]nicotine and [ 3 H]verapamil uptake by TR-BBB13 and TR-iBRB2 cells, respectively, by at least 29%. These results suggest that ent-cis-195A and cis-211A are recognized by the  In summary, we achieved the total syntheses of ent-cis-195A and cis-211A in a divergent process from the key and common intermediate 11. The absolute stereochemistry of natural cis-211A was determined to be 2R, 4aR, 5R, 6S, and 8aS by comparison with the data obtained from our total synthesis. The inhibitory effects of ent-cis-195A, cis-211A, and 6-epi-211A on nicotinic ACh receptors were also investigated. The results showed that cis-211A and 6-epi-211A had better inhibitory effects on the α7-receptor than that of ent-cis-195A, and none of the compounds showed inhibitory effects on the α4β2-receptor at the same concentration. These results suggested that cis-211A and 6-epi-211A could be applied as important tools for studying the brain and nervous system. More interestingly, the absolute configuration of the decahydroquinoline nuclei of cis-211A was a mirror image of that of cis-195A, even though both alkaloids were isolated from Oophaga (Dendrobates) pumilio from Panama.