Manzamine A Synthesis Essay

2.1. Chemogenomic Profiling Suggests That Manzamine A Interferes with V-ATPases

We first evaluated whether Manzamine A was active against the yeast S. cerevisiae as the availability of genome-wide deletion collections for this eukaryotic model system enable powerful genetic assays to elucidate the mechanism of action for bioactive molecules [18]. At 10 μM, manzamine A inhibited growth by 30%, and was tested at this concentration by chemogenomic profiling using the heterozygous and homozygous deletion collections. The basic concept is that the haploinsufficiency profiling assay (HIP) identifies genes where one functional copy, as compared to two, confers hypersensitivity to the compound. This indicates pathways directly affected by the compound. The homozygous profiling assay (HOP), where both gene copies are deleted, identifies pathways that are compensating (synthetically lethal) for those directly affected by the compound [19]. Manzamine A was tested in two independent experiments and the HIP HOP profiles were generated (Figure 1A,B). Both HIP HOP experiments showed a good correlation when plotted against each other (Figure 1C) suggesting the manzamine A response was robust and reproducible. However, in the HIP experiment only sensitive strains of low relevance were detected, as defined by calculation of a z-score across the HIP HOP profiles of 2500 diverse compounds [20]. A z-score within a range of +/−7 indicates a strain that responds to many diverse chemotypes and is thus, not relevent for interpretation. A flat HIP profile may indicate the primary target is not a protein, or is not conserved in yeast. In contrast, the HOP profiles revealed reproducible sensitive hits with a z-score < −7 suggesting they are relevant and specific for the manzamine A mechanism of action. However, due to the synthetic lethality concept of the HOP experiments it can be challenging to interpret individual strain sensitivities. Thus, we used a comparative approach and correlated an individual manzamine A HOP profile to our database of over 2500 HOP profiles for diverse chemical substances. The best correlation was to the HOP profile for the second manzamine A experiment, followed by a different manzamine derivative, underlining the validity of the applied approach. Interestingly, the next best correlating compounds were metacycloprodigiosin and a set of cleistanthin derivatives. Analysis of the HIP HOP profiles and direct alignment against the manzamine A HOP profile revealed a conserved set of sensitive HOP strains (Figure 1E–H). Prodigiosins have been established as uncouplers of H+ ATPases including the v-ATPase [21,22,23], uncoupling ATP hydrolysis from proton transport.The identified cleistanthins have been experimentally validated to affect vacuolar morphology and acidification of vacuoles in S. cerevisiae similar to bafilomycin A1, an established v-ATPase inhibitor [20]. In addition, cleistanthin derivatives share a structural core also found in diphyllin, a natural product reported to potently inhibit v-ATPase [24]. Thus, the HIP HOP profiling provides two key observations: (1) the absence of reproducible, relevant HIP hits suggests a non-protein target effect, and (2) the reproducible HOP profile correlates with established v-ATPase uncouplers. Taken together, we conclude that in yeast, manzamine A likely acts as an uncoupler of v-ATPase activity.

Figure 1. Manzamine A HIP HOP Profiling. (A,B) HIP HOP profiling of manzamine A in two individual experiments. Both independent HOP experiments identified common hits that are relevant based on z-score (z-score < −7); (C) z-score alignment of the two individual manzamine A experiments; (D) For reasons explained in the results section individual gene functions of hits were not analyzed in detail but the manzamine HOP profile was used in a correlation analysis against a database of HOP profiles of 2500 diverse chemical structures. This revealed correlation with other manzamine A experiments, metacycloprodigiosin and a cluster of cleistanthin substances. (E) HIP HOP profiles of metacycloprodigiosin. (F) HOP profile alignment of manzamine A and metacyloprodigiosin. (G) HIP HOP profile of the best correlating cleistanthin compound. (H) HOP profile alignment of manzamine A and cleistanthin. HIP HOP profiles are plotted by sensitivity and relevance (z-score). Grey boxes represent strains with deletions in essential genes, black dots strains with deletions in non-essential genes.

Figure 1. Manzamine A HIP HOP Profiling. (A,B) HIP HOP profiling of manzamine A in two individual experiments. Both independent HOP experiments identified common hits that are relevant based on z-score (z-score < −7); (C) z-score alignment of the two individual manzamine A experiments; (D) For reasons explained in the results section individual gene functions of hits were not analyzed in detail but the manzamine HOP profile was used in a correlation analysis against a database of HOP profiles of 2500 diverse chemical structures. This revealed correlation with other manzamine A experiments, metacycloprodigiosin and a cluster of cleistanthin substances. (E) HIP HOP profiles of metacycloprodigiosin. (F) HOP profile alignment of manzamine A and metacyloprodigiosin. (G) HIP HOP profile of the best correlating cleistanthin compound. (H) HOP profile alignment of manzamine A and cleistanthin. HIP HOP profiles are plotted by sensitivity and relevance (z-score). Grey boxes represent strains with deletions in essential genes, black dots strains with deletions in non-essential genes.

2.2. Effects of Manzamine A on Yeast Vacuolar Morphology and Proton Gradient

Next, we tested if effects on yeast vacuoles could be validated by a different approach. Inhibition of S. cerevisiae v-ATPase and modulating the proton gradient causes an increase in vacuole size and loss of luminal acidification, which can be assessed by fluorescence microscopy with (i) a vacuolar limiting membrane marker, Vph1-GFP, and (ii) the loss of LysoSensor fluorescence respectively [25,26]. We incubated Vph1-GFP labeled cells for 60 min with 5 × the IC30 concentration of the established v-ATPase inhibitor and manzamine A and analyzed the results in comparison to DMSO treated cells. Whereas DMSO treated cells showed clusters of 2–6 small vacuoles, bafilomycin A1 treatment resulted in formation of primarily a single large vacuole, (Figure 2A). In cells treated with 5 × the IC30 concentration of manzamine A, one enlarged vacuole was observed in more than 80% of the cells (n = 100). Testing vacuolar acidification by LysoSensor Green staining produced labeled vacuolar membranes in DMSO treated cells similar to the observed actions of the Vph1-GFP label. In both bafilomycin A1 and manzamine A treated cells, LysoSensor Green staining of the vacuole was completely absent suggesting a disruption of the pH gradient (Figure 2B). In summary, microscopic analysis of two v-ATPase dependent vacuolar parameters suggested similar effects of manzamine A to the established v-ATPase inhibitor bafilomycin A1.

Figure 2. Manzamine A affects vacuolar morphology and acidification in yeast, similar to bafilomycin. (A) Vacuolar morphology analysis using Vph1-GFP (a v-ATPase V0 domain) as marker. DMSO treated cells showed clusters of small vacuoles. In bafilomycin A1 treated cells one large vacuole was detected in almost all cells. Manzamine A treated cells displayed a few enlarged vacuoles similar to the situation observed in bafilomycin A1 treated cells. (B) Vacuolar acidification analysis using LysoSensor Green as marker. DMSO treated cells show staining of the vacuolar membranes. Treatment of cells with bafilomycin A1 or manzamine A results in abolishment of detectable vacuolar staining. Size bar represents 5 µM.

Figure 2. Manzamine A affects vacuolar morphology and acidification in yeast, similar to bafilomycin. (A) Vacuolar morphology analysis using Vph1-GFP (a v-ATPase V0 domain) as marker. DMSO treated cells showed clusters of small vacuoles. In bafilomycin A1 treated cells one large vacuole was detected in almost all cells. Manzamine A treated cells displayed a few enlarged vacuoles similar to the situation observed in bafilomycin A1 treated cells. (B) Vacuolar acidification analysis using LysoSensor Green as marker. DMSO treated cells show staining of the vacuolar membranes. Treatment of cells with bafilomycin A1 or manzamine A results in abolishment of detectable vacuolar staining. Size bar represents 5 µM.

2.3. Expression of V-ATPases in Pancreatic Cancer Cell Lines

Cells were seeded in 96-well plates and treated with manzamine A for 24 h. V-ATPase was detected by immunocytochemistry. V-ATPases were expressed in pancreatic cancer cell lines AsPC-1 and PANC-1 and to a less extent in BxPC-3 and MIAPaCa-2 (Figure 3). In contrast, the non-transformed cell line Vero showed only weak staining for v-ATPases. Manzamine A treatment had no effect on expression of v-ATPases.

Figure 3. Manzamine A treatment had no effect on expression of v-ATPases in pancreatic cancer cells. AsPC-1, PANC-1, BxPC-3 and MIA PaCa-2 pancreatic cancer cells, as well as non-malignant Vero cells were treated with 10 µM manzamine A or methanol (vehicle control). 24 h later expression of v-ATPases was analyzed by immunocytochemistry with mouse monoclonal antibodies. Cells were analyzed at a magnification of 20× or 40× (inserts). One representative experiment out of three is shown.

Figure 3. Manzamine A treatment had no effect on expression of v-ATPases in pancreatic cancer cells. AsPC-1, PANC-1, BxPC-3 and MIA PaCa-2 pancreatic cancer cells, as well as non-malignant Vero cells were treated with 10 µM manzamine A or methanol (vehicle control). 24 h later expression of v-ATPases was analyzed by immunocytochemistry with mouse monoclonal antibodies. Cells were analyzed at a magnification of 20× or 40× (inserts). One representative experiment out of three is shown.

3.1. One-pot asymmetric synthesis of substituted piperidines

Piperidine containing alkaloids are common in marine environments [20–22]. Although there have been several synthetic methods [23–25] reported for construction of this moiety, stereoselectivity remains a challenging task especially when three or more stereogenic centers or quaternary substituted carbons are present. The Shi group developed a novel cascade approach for the stereoselective synthesis of the piperidine moiety, based on their development of an intermolecular cross-double-Michael addition between α, β-unsaturated carbonyl compounds and nitroalkenes, facilitated by amines as Lewis bases (LB)[26]. Allylic nitro products are generated in the process via the β-elimination of the LB (Figure 2). The new cascade reaction encouraged Shi’s group to extend the cascade by involving an activated electrophilic intermediate for the construction of nitrogen containing heterocycles with two or more stereogenic centers.

Based on their previous results, which showed that the addition of the amine to nitroalkene is fast, they proposed a one-pot cascade approach for the asymmetric synthesis of the piperidine moiety [27]. They postulated that adduct A (Figure 3), the addition product of the amine with nitroalkene, was suitable for Michael addition type reaction with an activated carbonyl to generate adduct B. Ring closure then provides the substituted piperidine motif in a one-pot approach with the generation of three stereogenic centers. To validate this proposed one-pot cascade sequence, nitrostyrene was allowed to react with methyl vinyl ketone (MVK) in the presence of a primary amine (Scheme 5). Being a complex mixture with many possible side reactions (i.e., Baylis-Hillman reaction), it was surprising that substituted piperidines, 13a and 13b, were obtained in excellent yields and good diastereoselectivities with no side products. Solvent optimization of these one-pot conditions showed that THF gave the best yields with good diastereoselectivities (97%, dr = 7:1 for 13a; 85%, dr = 15:1 for 13b). The formation of 13 confirmed the formation of adduct B (Figure 3), which was trapped by a sequential Henry-aldol cyclization. Although three stereogenic centers were generated in piperidine 13 only two C-4 diastereoisomers were obtained, of which the cis isomer of C-3 nitro and C-4 hydroxy groups was the major product.

This one-pot reaction was compatible with a variety of nitroalkenes, amines and activated carbonyls. Different aryl and alkyl substituted nitroalkenes generated variation at the C-2 position. Moreover, both alkyl and aryl ketones were suitable for this cascade one-pot process giving different choices of substitutes on the C-4 position. Substituents on C-5 were possible through α-substituted enones, while β-substituted enones delivered variation at the C-6 position only when ammonia was used. It is interesting that only C-4 isomers were obtained in all cases.

The aza-Michael adduct B (Figure 3) was not observed in reaction NMR studies; moreover, the piperidine products were stable under strong basic conditions. These results strongly suggested that the irreversible Henry-aldol cyclization was the rate determining step, which accounts for the piperidine diastereoselectivity through the chair transition state. With the formation of the C-2 stereogenic center during the amine conjugate addition, it is possible that the stereochemistry of the final piperidine product could be selectively induced by chiral amines through its involvement in the spatial arrangement of the chair transition state. In this case, a new stereogenic center on the exocyclic C-7 will be introduced on the piperidine product, which may result in the formation of four diastereoisomers. To investigate this proposed chirality induction, arylethanamines 14ac were applied in the one-pot piperidine synthesis and as expected, piperidines 15ac were obtained in good yields (Scheme 6). The results showed excellent diastereoselectivity of the C-4 position in 15a and 15b with dr > 10:1. In addition, modest chirality induction by the C-7 position was observed with dr = 4:1.

X-ray analysis of the piperidine crystals revealed that the exocyclic C-7 adopted a staggered conformation with respect to the ring. Moreover, the chemical shifts of the C-8 and C-9 carbons were significantly shifted upfield in an anti-parallel position relative to the nitrogen lone pair electrons. This data suggest that the N1-C7 σ-bond rotation was restricted. Based on these results, a Henry-aldol cyclization chair transition state was proposed in Figure 4. The N-1 nitrogen lone pair electrons were placed in the axial position and the preferred staggered N1-C7 conformation could be achieved when the small group on C-7 was placed anti (axial orientation) to the N-1 nitrogen lone pair. Three syn-pentane interactions were expected in this staggered conformation (as shown in Figure 4); however, placement of the small group on C-7 anti to the lone pair will minimize these interactions. This proposed transition state was consistent with experimental observations. Also, the stability of this transition state could be influenced by the n-σ* electronic interaction between the nitrogen lone pair and the exocyclic axial small group.

To validate the importance of this n-σ* electronic interaction in the chirality enhancement, the methyl group was replaced in the chiral amine with a carboxylate group which is less bulky and would produce a stronger n-σ* interaction. As expected, the reaction of the amino esters 16a, b with enone 12a and nitroarene 10a yielded the corresponding piperidine products in moderate yields (Scheme 7). Interestingly, piperidines 17a, b were obtained in their diastereomerically pure form. Complete chirality induction was achieved through control via exocyclic asymmetry. This new one-pot method opens the door for the asymmetric synthesis of several natural products that contain a piperidine moiety.

3.2. One-pot organo-catalytic synthesis of quinolizidine derivatives

An elegant one-pot method was developed recently by Franzen and Fisher [28] for the asymmetric synthesis of substituted quinolizidines. This moiety is widely represented in several alkaloids isolated from plants, ants and marine organisms. Their retrosynthetic analysis for the indolo[2,3a]quinolizidine skeleton is highlighted in Figure 5. They postulated that the stereogenic center 11b could be generated through the asymmetric acyliminium ion cyclization of the imine (19). This imine could be obtained from the aldehyde precursor 20, in which the aldehyde is apparently the adduct of an enantioselective Michael addition.

This retro-analysis required the addition of an activated amide 22 to the unsaturated aldehyde 21, which has not yet been reported in the literature. To accomplish their synthetic plan, they optimized the enantioselective Michael addition of cinnamic aldehyde 23 and an activated indol substituted amide 22, using different proline derivatives as organocatalysts (Figure 6), exploring different solvents, temperatures, as well as various acids needed for the cyclization step of the acyliminium ion (Scheme 8). Dichloromethane (DCM) was the solvent of choice and yielded compounds 24a and 24b with full conversion at room temperature, but with low enantioselectivity (88% ee). The enantioselectivity was increased to 94% by lowering the temperature to 3 °C, while further cooling (−20 °C) resulted in only 5% conversion. The proline derivative (S)-A showed the best selectivity, B was less selective and less active, while C and D were inactive for this reaction. When TFA was used in the acid-catalyzed cyclization of the acyliminium ion, the indoloquinolidine products were obtained nonselectively as a 1:1 mixture. However, some selectivity of 24a over 24b was observed when HCl was used. It was also noted that cooling down the reaction mixture prior to the addition of HCl increased the selectivity up to 85:15. The optimized conditions are highlighted in Scheme 8.

This one-pot, two step method was applied to several aromatic α, β-unsaturated aldehydes with good to excellent enantioselectivity. The indolyl moiety of 22 was replaced with an electron-rich phenyl group, which should give direct access to the benzo[a]quinolizidine skeleton. The reaction between 23 and the activated amide 25, in the presence of the organocatalyst A, and subsequent addition of HCl, delivered the benzo[a]quinolizidines 26a and 26b with good to high enantioselectivity (Scheme 9). It was noted that the formation of the benzo[a]quinolizidine moiety required stronger acidic conditions (40 mol%) relative to the indolo[2,3a]quinolizidine (20 mol%). This is explained by the poorer nucleophilicity of the phenyl ring compared to the 3-indolyl moiety.

The absolute configuration of one of the benzo[a]quinolizidine derivatives was established by X-ray analysis as 2R,3S,11bS, which provided important mechanistic insight (Scheme 10). The aryl groups in the catalyst will shield the Re face of the iminium intermediate 27, which will help to establish the S configuration on C-2 through the unshielded Si face. Intramolecular imine formation with epimerization of the stereochemically labile stereogenic center at C-3 will establish the more thermodynamically stable 2R,3S-trans configuration of the intermediate imine 28. This acyliminium ion undergoes an electrophilic aromatic substitution with the aromatic moiety to give quinolizidine products.

The great diastereoselectivity observed in the acid catalyzed cyclization could be explained based on the reaction conditions. Considering the synthesis of indoloquinolizidines 24a and 24b (Scheme 11), the major isomer was 24a with the indolyl moiety in an axial orientation. Although the formation of 24a resembles a higher energy product, reaction via the transition state I (Scheme 11) is under kinetic control (−78 °C), owing to less steric hindrance from the equatorial α protons relative to the theromodynamic equatorial product.

As a possible application of this one-pot method in the total synthesis of biologically active marine alkaloids, the tricyclic core of schulzenine alkaloids could be easily constructed using Franzen and Fisher’s method (Figure 7). Schulzenine alkaloids were recently isolated from the marine sponge Penares schulzei [29]. Schulzenine A–C inhibits α-glucosidase with IC50 values of 48–170 nM. Figure 8 represents the proposed one-pot reaction that could be used for the synthesis of the tricyclic core of schulzenine alkaloids.

0 comments

Leave a Reply

Your email address will not be published. Required fields are marked *