Keywords

10.1 Introduction

Fused polycyclic structures are found in many natural products, but although significant progress has been made in their synthesis, the development of efficient and practical strategies for constructing polycycles with adjustable substituents and functional groups remains challenging [1,2,3]. Aromatic compounds based on benzene, toluene, and xylenes, derived from petroleum, are readily available as basic precursors for synthesis. For example, aryl halides play a pivotal role in the synthesis of agrichemicals and pharmaceuticals via cross-coupling reactions [4, 5]. Nevertheless, although these readily available aromatic compounds are convenient starting materials for synthesizing biologically useful compounds with stereochemically rich scaffolds, their thermodynamic stability presents a challenge [6], and therefore, dearomatization reactions have been intensively explored [7, 8].

One of the most powerful dearomatization reactions is the oxidative dearomatization of phenols [9]. In particular, when phenols with attached nucleophiles are subjected to oxidants, such as hypervalent iodine reagent, dearomative oxidative cyclization is proceeded. This process is a promising method for assembling natural products that possess polycyclic framework quickly [10,11,12,13]. The advantages of this strategy include: (i) the widespread commercial availability of a variety of phenolic precursors, each featuring diverse substituents, (ii) the dearomatization reaction enables to the efficient construction of quaternary carbon centers, and (iii) the easy functionalization of the resulting dienones. Over the past few decades, there has been considerable research into the oxidative spirocyclization of phenol derivatives with nucleophiles [14, 15]. Reactions involving a range of nucleophiles such as carboxylic acids, amines, and aromatic rings have been documented. Furthermore, syntheses of various natural products based on this dearomative spirocyclization strategy have been reported, establishing this method as a promising strategy for the construction of complex frameworks (Scheme 10.1) [16, 17].

Scheme 10.1
Two chemical reactions. A, experiment by Kita's group. The reactant reacts with M e C H double bond C, O M e, O T M S and discorhabdin C is produced. The O H is converted into keto group. B, experiment by Gaunt's group. The reactant reacts with P I D A, T F E and produces morphine. The phenol moiety is converted.

Representative examples of dearomative spirocyclization in natural product synthesis

Full size image

Although oxidative phenolic coupling reactions of phenols with nucleophiles at the ortho-position to the para-position can efficiently construct fused ring structures, there are few reports describing this approach [18,19,20,21]. Moreover, phenol derivatives featuring the pendent nucleophiles in both ortho- and para-positions can be converted into complex fused polycyclic compounds by applying oxidative phenolic coupling reactions and subsequent intramolecular Michael reactions (Scheme 10.2). However, this strategy has not yet been applied to the synthesis of natural products. Therefore, we have studied the synthesis of alkaloids from phenols using an oxidative coupling reaction/aza-Michael reaction strategy. In this chapter, we summarize our recent progress with this strategy, focusing on its application to the synthesis of complex polycyclic alkaloids [22].

Scheme 10.2
A reaction pathway. Ortho para substituted phenol derivatives undergo an oxidative-phenolic coupling reaction. It gives fused skeletal dienes. The nucleophile is attached at para positions. Regioselective aza Michael's reaction produces two polycyclic compounds. These have chair conformations.

Overview of our strategy to construct complex polycyclic frameworks

Full size image

10.2 Total Synthesis of (+)-Gracilamine

10.2.1 (+)-Gracilamine

(+)-Gracilamine (1) is an alkaloid isolated from the plant Galanthus gracilis, a member of the Amaryllidaceae family, by Ünver and Kaya in 2005 (Fig. 10.1) [23]. (+)-Gracilamine (1) possesses a highly functionalized fused polycyclic structure consisting of five rings, A–E. Compare to other members of the Amaryllidaceae alkaloids (2–4), it possesses a complex structure with seven stereogenic centers, one of which is a quaternary carbon at the C3a position. This unique structure of 1 is of great interest to synthetic chemists, and since its first total synthesis by Ma and co-workers in 2012 [24], nine total syntheses [25,26,27,28,29,30,31], including ours [32], have been reported [33].

Fig. 10.1
Chemical structures of gracilamine, lycorine, galanthamine, and epimaritidine.

Structure of (+)-gracilamine (1)

Full size image

10.2.2 Our Synthetic Plan of (+)-Gracilamine

The synthetic challenge in the case of 1 is to efficiently construct contiguous stereogenic centers while assembling a highly fused polycyclic ring system. As a starting point, our focus was on the stereochemistry of the C9a position. In our group, enantioselective 1,2-type aza-Friedel–Crafts (aza-FC) reaction of aldimine 6 and sesamol (7) by using our guanidine–bisthiourea catalyst 5 has been developed to provide optically active amine 8 (Scheme 10.3) [34, 35]. We envisaged that this enantioselective 1,2-type aza-FC reaction would provide a crucial breakthrough in addressing the synthetic challenges of 1, namely to enable the stereoselective connection of the A-ring and E-ring, in addition to constructing the desired stereochemistry at the C9a position.

Scheme 10.3
A reaction pathway. Aldimine reacts with sesamol in the presence of E t 2 O and produces an amine up to 86% e e. Guanidine bis thiourea catalyst is used in this reaction. The aldimine and sesamol moieties are fused, and N B o c is converted into N H B o c.

Enantioselective 1,2-type aza-Friedel–Crafts reaction catalyzed by guanidine–bisthiourea catalyst

Full size image

We planned to synthesize (+)-gracilamine (1) by employing the aza-FC reaction, as depicted in Scheme 10.4. We envisioned that 1 could be obtained by constructing the D-ring from enone 9. The enone 9 would be obtained by constructing the C-ring through a regioselective intramolecular aza-Michael reaction with the tricyclic dienone 10. We envisioned synthesizing the B-ring of 10, including the quaternary carbon at C3a, via a diastereoselective oxidative phenol coupling reaction of diarylmethylamine 11. The key issue in the oxidative phenol coupling reaction of 11 is whether the stereochemistry of C9a can be engaged for stereoselective construction of the C3a position. Diarylmethylamine 11, serving as the substrate for the oxidative phenol coupling reaction, can be obtained through the aforementioned enantioselective 1,2-type aza-FC reaction of aldimine 13 and sesamol (4).

Scheme 10.4
A reaction pathway. Gracilamine is converted into an enone. This enone gives tricyclic dienone by regioselective aza Michael reaction. The diaryl methylamine is formed by diastereoselective oxidative phenolic coupling. By enantioselective aza Friedel crafts reaction aldimine and sesamol are produced.

Our synthetic plan for (+)-gracilamine (1)

Full size image

10.2.3 Preparation of Optically Active Coupling Precursor 19 Based on Enantioselective 1,2-Type Aza-Friedel–Crafts Reaction

First, the aza-FC reaction was investigated using imine 14 (Scheme 10.5). The reaction of sesamol (4) with imine 3a at 20 °C in ether in the presence of catalyst 5 gave the aza-FC product 15 with 74% ee under the same conditions as described in our previous report. However, the reactivity of 14 was low due to the ortho-substituent, and the yield of 15 was only 33%. Therefore, we increased the reaction temperature to 40 °C and found that the desired aza-FC adduct 15 was obtained in 94% yield with 91% ee. In this organocatalytic reaction, the enantioselectivity-determining step is governed by entropy, and the enantioselectivity was improved by increasing the reaction temperature. The phenolic hydroxy group of the resulting 15 was converted to a trifluoromethanesulfonate group (16). The optical purity of 16 was improved to 99% ee by recrystallization from heated hexane. The C9a asymmetric carbon of (+)-1 was thus successfully constructed. Next, the phenolic coupling precursor of 19 was synthesized. Ozonolysis of the allyl group of 16 and reduction of the resulting aldehyde with sodium borohydride provided alcohol 17 in 87% yield. The hydroxy group of 17 was mesylated with methanesulfonyl chloride in DMF followed by sodium azide to afford azide 18 in 80% yield. In this process, when the mesylation was carried out in dichloromethane, an elimination reaction took place to give the vinyl compound as the major product. By carrying out the mesylation in DMF, the elimination reaction of the mesyl group was suppressed and the subsequent azidation could be achieved in a one-pot process. Then, the azide 18 was subjected to hydrogenolysis to remove the benzyl and triflate groups and reduce the azide group to an amine, and the resulting product was reacted with p-toluenesulfonyl chloride to give the precursor 19 for the oxidative coupling reaction.

Scheme 10.5
A reaction pathway. The reactant is a substituted benzene. C 1 is bonded to C H N B o c, C 2 is to C 3 H 5, C 5 is to B n O. It is converted into F C product. Then O H is converted into O T I. Ozonolysis of allyl group produces alcohol. This converts into azide. B o C H N converts into N H B o c.

Construction of the C9a stereogenic center and synthesis of the coupling precursor 19

Full size image

10.2.4 Synthesis of the Tetracyclic Core Structure of 1 by Oxidative Phenolic Coupling Reaction and Aza-Michael Reaction

Oxidative phenolic coupling reactions using hypervalent iodine reagents were investigated with the coupling precursor 19 (Scheme 10.6). As a result, we found that dienone 20 with the desired configuration at C3a was obtained as a single isomer in 80% yield by treatment with diacetoxyiodobenzene (PIDA) in HFIP [36]. In this reaction, the orientation of the NHBoc group at the C9a position gives rise to two possible transition states, denoted as TS-1 and TS-2. In TS-2, which yields diastereomer 21, the NHBoc group is directed in the pseudo-axial direction, leading to 1,3-diaxial repulsion with the side chain on the aromatic ring. Thus, it is expected that TS-1, which involves less steric hindrance, would be preferred, and indeed, the desired stereochemistry at C3a was predominantly formed.

Scheme 10.6
A reaction. The reactant has a phenol moiety. The C 3 and C 4 are substituted. It reacts with P I D A, H F I P and gives 2 transition states. T S 2 has steric repulsion due to the pseudo axial position of N H B o c. T S 1 gives single diastereomer compound. N H B o c is attached to the cyclopentane.

Oxidative phenolic coupling reaction of 19

Full size image

Having synthesized the tricyclic dienone 20, we subsequently investigated the regioselective aza-Michael reaction (Scheme 10.7). We found that this reaction proceeded at the sterically less hindered C7a position with p-toluenesulfonic acid in dichloromethane, and the tetracyclic enone 22 corresponding to rings A, B, C, and E in 1 was obtained in 70% yield. Then, the reduction of the double bond in 22 was investigated. However, the yield was low and the reproducibility was poor, probably due to the poor solubility of enone 22. Thus, we attempted to change the protecting group of 22 to increase the solubility.

Scheme 10.7
A reaction. A tricyclic dienone reacts with T s O H, H 2 O, C H 2 C l 2, and produces a tetracyclic enone. It has poor solubility. It reacts with H 2 and P d, C in M e O H solution, and the double bond is reduced in the final product.

Regioselective aza-Michael reaction of 20

Full size image

After the deprotection of the Boc group of 19 by using hydrogen chloride in methanol, the 2-(trimethylsilyl)ethoxycarbonyl (Teoc) group was introduced at the resulting amine by using N-[2-(trimethylsilyl)ethoxycarbonyloxy]succinimide (Teoc-OSu, Scheme 10.8). We then carried out the oxidative coupling reaction of 24, and dienone 25 with the desired configuration at C3a was obtained as a single isomer. Furthermore, the reaction of para-toluenesulfonic acid with 25 afforded the tetracyclic enone 26 in 82% yield via regioselective aza-Michael reaction at the C7a position. The resulting 26 was sufficiently soluble in various solvents, and subsequent reduction of the double bond of 26 by hydrogenation took place smoothly to give the ketone 27 quantitatively.

Scheme 10.8
A reaction pathway. The reactant is a substituted phenol. The para position is bonded to C H 2 C H 2 T s H N. When it is reacted with P I D A, B o c group is converted into N H T e o c. When it is reacted with T s O H H 2 O, an enone is formed. On hydrogenation it gives ketone.

Changing the protecting group from Boc to Teoc

Full size image

10.2.5 Total Synthesis of (+)-Gracilamine by Constructing D-Ring Based on Intramolecular Mannich Reaction

Next, the construction of the D-ring by means of the Mannich reaction was examined, aiming at the total synthesis of gracilamine (1) (Scheme 10.9). After investigation of various conditions, the intramolecular Mannich reaction of ketone 27 proceeded upon heating in the presence of α-keto ester 28 in a mixed solvent of cyclopentyl methyl ether (CPME) and trifluoroacetic acid (TFA) to afford 29 with the desired stereochemistry at the C8 position in 47% yield. The carbonyl group at C6 of 29 was reduced with sodium borohydride to give the desired 30 in 64% yield, together with 32, formed by further lactonization of the diastereomer 31, in 4% yield. A single crystal of 32 was obtained, and the absolute conformation was determined by X-ray crystallography. For the synthesis of 1, the conversion of the tosyl group of compound 30 to a methyl group was examined. However, removal of the tosyl group was troublesome, and decomposition of the substrate occurred when sodium naphthalenide was employed. Careful monitoring by TLC showed that the amine 33, which was generated by the deprotection of the tosyl group, was unstable under reducing conditions. Consequently, the slow addition of sodium naphthalenide competed with the deprotection of the nitrogen atom of 30 and the decomposition of 33. Therefore, all of the sodium naphthalenide was quickly added to the mixture, and the reaction was immediately quenched by adding ethanol and acetic acid, thereby suppressing the decomposition of 33. Then, the methyl group was introduced to the resulting amine in a one-pot fashion by reaction with formaldehyde in aqueous solution in the presence of NaBH3CN. This completed the total synthesis of (+)-gracilamine (1) [32].

Scheme 10.9
A reaction pathway. A ketone is converted into alpha keto ester. It reacts with T F A. The carbonyl group of the product converts into O H. Lactonization gives a ball and stick model. When the previous product is treated with N a, naphthalene, it gives an intermediate that produces gracilamine.

Total synthesis of (+)-gracilamine (1)

Full size image

10.3 Total Syntheses of Hasubanan Alkaloids

10.3.1 Hasubanan Alkaloid

Hasubanan alkaloids represent a group of alkaloids extracted mainly from plants of the genus Stephania [37]. These alkaloids have been known for a very long time. For example, Goto and Suzuki in Kitasato Institute reported the isolation of acutumine (43) for the first time in 1929 [38]. These alkaloids have a characteristic tetracyclic skeleton with a quaternary carbon at the C13 position, the so-called hasubanan scaffold (hasubanan skeleton 34). Hasubanonine (35) and metaphanine (36) are representative examples. Various analogs with different oxidation states on this scaffold have been reported, totaling more than 40 congeners. In addition to those compounds with the hasubanan skeleton, a number of compounds without this scaffold, such as stephadiamine (37), cepharatines 38–41, sinoracutine (42), and acutumine-type alkaloids 43 and 44, have been isolated to date [37].

Hasubanane alkaloids display a spectrum of biological activities, encompassing antibacterial properties for cepharatines [39], anti-amnesic activity and selective T-cell cytotoxicity for acutumine (43) [40], as well as opioid receptor affinity for enantiomers of the hasubanan skeleton [41]. Consequently, these alkaloids continue to be of great synthetic interest. Since the first total synthesis by Ide and Kitano from Kyoto University in 1966 [42], more than 20 total syntheses have been achieved using various methodologies [43,44,45,46]. Organic chemists have also explored synthetic methods targeting diverse scaffolds within this family of alkaloids, stimulated by the pioneering work of Herzon’s and Reisman’s groups [47,48,49,50].

10.3.2 Our Synthetic Approach for Hasubanan Alkaloid

We examined the synthesis of three different types of hasubanan alkaloids, metaphanine (36), stephadiamine (37), and cepharatines 38–41, based on a common strategy. The challenges in the synthesis were the stereoselective construction of the common quaternary carbon at C13 and the construction of the C- and D-rings in the late stage. Our synthetic approaches are depicted in Scheme 10.10. The characteristic five-membered C-ring of stephadiamine 48 would be constructed by ring contraction from the six-membered C-ring of the hasubanan skeleton 47. This skeleton 47 would be synthesized by C14-selective intramolecular aza-Michael reaction of the tricyclic dienone 46. On the other hand, the characteristic azabicyclo [3.3.1]nonane motif constituting the C- and D-rings of the cepharatine skeleton 50 would be constructed by a C5-selective intramolecular aza-Michael reaction of the dienone 46, followed by a 1,2-migration of the nitrogen atom from C5 to C6 in 49. We also considered that the dienone 46 would be synthesized by oxidative phenol coupling reaction of diarylethane derivative 45. In this case, we expected that the stereochemistry of the newly constructed C13 quaternary carbon would be controlled by the stereochemistry at C10 in 45.

Scheme 10.10
A reaction pathway. The reactant has a substituted benzene ring. Oxidative phenolic coupling gives a compound. The O H group of ring C is converted into keto group. It gives either hasubanab skeleton and stephadiamine skeleton by ring contraction or a cepharatine skeleton by 1, 2 migration.

Our comprehensive synthetic approach for hasubanan alkaloids

Full size image

10.3.3 Construction at C10 Stereogenic Center in Hasubanan Skeleton

We commenced with the synthesis of ketone 58 to investigate the construction of the stereochemistry at the C10 position (Schemes 10.11 and 10.12). First, the commercially available aromatic aldehyde 51 was subjected to bromination to give 52 in 62% yield. Then, TMS cyanohydrin 53, the precursor of the A-ring part, was obtained by reaction with trimethylsilyl cyanide in the presence of zinc iodide to give 52 quantitatively. The C-ring precursor of 57 was obtained from phenol 54 by protecting the phenolic hydroxy group with a methoxymethyl group, reduction of the aldehyde to an alcohol with sodium borohydride and regioselective bromination with N-bromosuccinimide to give 55 in 93% yield (three steps), followed by the introduction of an allyl group by means of the Stille coupling reaction with tributylallyltin in the presence of Pd catalyst to give 56, which was then treated with methanesulfonyl chloride.

Scheme 10.11
Two reactions. A, A ring part. 3, 4 dimethoxy benzaldehyde reacts with B r 2, M e C N. B r atom is added to the ring. When treated with T M S C N, C H O group is replaced by C H C N O T M S group. B, C ring part. Meta hydroxy benzaldehyde. The O H becomes O M O M, C H O is converted into C O M s.

Synthesis of A- and C-ring synthons 53 and 57

Full size image
Scheme 10.12
A reaction. The reactant has a benzene ring. The C 2 is bonded to B r, C 4 and C 5 are bonded to methoxy, C 1 is bonded to C H C N O T M S. This reacts with L i H M D S, T H F, n B u 4 N F and gives a ketone.

Coupling reaction of A- and C-ring synthons

Full size image

With the A-ring precursor 53 and C-ring precursor 57 in hand, these were coupled to obtain ketone 58 (Scheme 10.12). That is, deprotonation of the A-ring cyanohydrin 53 with lithium hexamethyldisilazide, mesylate 57 promoted the alkylation reaction, and then, TMS group was deprotected with tetrabutylammonium fluoride to give ketone 58 in 78% yield.

With the ketone 58 in hand, we then examined the enantioselective reduction of the carbonyl group at C10 in 58 (Table 10.1). Initially, we examined asymmetric reduction with CBS catalysts [51]. In the presence of catalysts 63, 64, and 65 bearing methyl, ortho-methylphenyl, and n-butyl groups, however, the enantioselectivities of the alcohol 59 were low or moderate (16–51% ee, entries 1–5). In contrast, with the substrate 60 lacking the bromo group on the aromatic ring, the corresponding alcohol 61 was obtained with 99% ee (entry 6). This result strongly suggested that the low enantioselectivity in this asymmetric reduction was due to steric hindrance around the carbonyl groups of 58.

Three chemical structures. It presents the structure of the tested catalysts. Different types of catalysts can be obtained by replacing R with M e, o ToI, n B u. It presents the structures of S B T M and 2 S, 3 R hyper B T M.
Table 10.1 Stereoselective construction at C10
Full size table

Next, we investigated the kinetic resolution of racemic alcohol rac-59 using chiral isothioureas [52]. In the presence (S)-BTM (66, 10 mol%) and isovaleric anhydride as the acylating agent, kinetic resolution proceeded and the optically active 59 was obtained in 60% ee (entry 7). When we examined the more reactive chiral isothiourea catalyst (2S,3R)-HyperBTM (67) [53], kinetic resolution proceeded more efficiently, and 59 was obtained in 96% ee (entry 8). By decreasing the reaction temperature to −60 °C, the enantioselectivity was improved to 99% ee (entry 9). This kinetic resolution strategy allowed us to construct the stereochemistry at C10 of the hasubanan skeleton. The ester 62 obtained by kinetic resolution was converted almost quantitatively to ketone 58 by hydrolysis of the ester group and subsequent oxidation with Dess–Martin periodinane.

10.3.4 Construction of Hasubanan Skeleton

With the optically active 59 in hand, we next synthesized the precursor 71 for the oxidative coupling reaction (Scheme 10.13). The hydroxy group at C10 in 59 was protected with a TIPS group, and the silyl ether 68 was subjected to ozonolysis/reduction to give the alcohol 69 in 75% yield. After mesylation of the resulting hydroxy group of 69, azide 70 was obtained by reacting with sodium azide in a one-pot process. After reduction of the azide group of 70 under the Staudinger conditions, the protection of the resulting amine by a Boc group and the removal of MOM group by using a catalytic amount of carbon tetrabromide in 2-propanol were carried out, respectively, to give phenol 71 [54].

Scheme 10.13
A reaction pathway. Hydroxy group of the reactant is bonded with TIPS group when it is reacted with TIPS C l imidazole. On ozonolysis, alcohol is formed. Azide is formed when it reacts with sodium azide. The N 3 is converted into N H B o c and O M O M is converted into O H in the final product.

Synthesis of the oxidative phenolic coupling precursor 71

Full size image

The oxidative coupling reaction was examined with phenol 71 as depicted in Scheme 10.14. After investigating various reaction conditions, we found that the oxidative coupling product 72 was obtained as a single diastereomer in 34% yield by treatment with PIDA in HFIP in the presence of MeOH at 0 °C [55]. The high diastereoselectivity in this reaction can be explained as follows. In this reaction, two transition states, TS-3 and TS-4, are possible with respect to the orientation of the C10 substituent. The reaction proceeds predominantly through TS-3, which has less steric repulsion than TS-4, yielding 72 with the desired stereochemistry at C13.

Scheme 10.14
A reaction. The reactant has 2 substituted benzene rings bonded by C O TIPS C H 2. When it is treated with P I D A, M e O H, it gives T S 3 and T S 4. T S 4 has steric hindrance, so further product is not obtained. T S 3 gives a single diastereomer. It is a dienone.

Oxidative phenolic coupling reaction of 71

Full size image

Then, the regioselective aza-Michael reaction of the dienone 72 was examined (Scheme 10.15). Although we explored various acidic or basic conditions, unfortunately, the desired aza-Michael reaction adduct at C14 did not proceed, and only the C5-adduct of 74 was obtained. Therefore, we decided to reduce the less hindered olefin at C5–C6 of 72 by hydrogenation in the presence of Rh/C. The aza-Michael reaction of the resulting enone 75 took place at C14 upon treatment with hydrochloric acid. However, under these conditions, the resulting 76 proved to be unstable, undergoing elimination of the silyl ether at C10 to afford an oxonium cation 77 followed by intramolecular cyclization to give carbamate 78.

Scheme 10.15
A reaction. The reactant is a dienone. By aza Michael reaction at C 5, it gives a product. The N C bond is formed. Aza Michael's reaction at C 14 gives another product. The reduced product reacts with H C l and gives carbamate. The oxonium cation converts into carbamate by cyclization.

Initial attempts at intramolecular aza-Michael reaction

Full size image

Thus, we explored the intramolecular aza-Michael reaction with the dienone 79, which was obtained by removal of the bromo group of 72 with formic acid in the presence of a palladium catalyst, under a variety of conditions (Table 10.2). Firstly, acidic conditions were examined. In the case of TFA, the reaction of 79 predominantly provided the undesired C5-adduct 81 (80/81 = 1:4.4, entry 1). With hydrochloric acid, 80 and 81 were obtained in a 1:1 ratio (entry 2). We then examined basic conditions. When DBU was used, the reaction did not proceed, resulting in the recovery of substrate 79, presumably due to the weak basicity of DBU (entry 3). Strong bases such as NaH and KOt-Bu afforded mixtures of 80 and 81 with little or no selectivity (entries 4, 5). Interestingly, in the case of potassium tert-butoxide, we observed a slight selectivity for the C14-adduct 80 (80/81 = 1.9:1, entry 6). Subsequent optimization of the reaction conditions demonstrated the efficacy of using HMPA as a co-solvent, and the selectivity was significantly enhanced to give 80 predominantly (80/81 = 7.3:1, entry 7). The diastereomers were separable by silica gel column chromatography, affording the desired C14-adduct 80 in 51% isolated yield (entry 7). The tetracyclic hasubanan skeleton was thus constructed.

Table 10.2 Investigation of aza-Michael reaction
Full size table

10.3.5 Total Syntheses of (–)-Metaphanine and (+)-Stephadiamine

With the tetracyclic hasubanan skeleton 80 in hand, our attention moved to the total synthesis of (–)-metaphanine (36) (Scheme 10.16). First, oxidation at the C8 position in 80 was examined. Oxidation reaction of 80 with the electron-deficient oxaziridine rac-82 proceeded smoothly to give α-hydroxyketone 83 as a single diastereomer in 71% yield. The reason for the stereoselectivity is presumably that the TIPS group on C10 shields the β-face of 80, thereby influencing the preferred direction of approach of rac-82. The resulting hydroxy group of α-hydroxyketone 83 was oxidized with Dess–Martin periodinane to give a diketone, whose TIPS group was deprotected with HF followed by hydrogenolysis of the olefin to give hemiketal 85. Finally, asymmetric total synthesis of (–)-metaphanine (36) was achieved by deprotection of the Boc group of 85, followed by methylation of the resulting amine under reductive amination conditions.

Scheme 10.16
A reaction pathway. A tetracyclic hasubanan skeleton reacts in K H M D S. The product reacts with D M P. Alpha hydroxy ketone is oxidized. Hydrogenolysis produces a product with C O C bond. When it is treated with H C l, H C H O, it gives metaphanine. The ball and stick model is presented.

Total synthesis of (–)-metaphanine (36)

Full size image

We next examined the conversion of (–)-metaphanine (36) to (+)-stephadiamine (37) by contraction of the C-ring of the hasubanan skeleton (Scheme 10.17). In this transformation, it is necessary not only to contract the C-ring, but also to introduce the α-tertiary amine with construction of the lactone. After investigating various conditions, we found that the aza-benzilic acid type rearrangement of 36 occurred efficiently through the formation of the iminium intermediate 86. This rearrangement was achieved by simply treatment 36 with NH3 in MeOH at room temperature, resulting in a quantitative yield of compound 37. Although the biosynthetic pathway of (+)-stephadiamine (37) has not yet been established in detail [44, 56, 57], we consider that 36 is likely to be a biosynthetic precursor of 37 via a similar reaction, based on the structural similarity between 36 and 37.

Scheme 10.17
A reaction. Metaphanine reacts with N H 3, M e O H and gives an aza bezlllc type rearrangement. Then stephadiamine is formed as the product.

Total synthesis of (+)-stephadiamine (37)

Full size image

10.3.6 Investigation of Oxidative Phenolic Coupling for Cepharatines

Cepharatines are classified into A- and C-types (38, 40) and B- and D-types (39, 41), depending on the substitution pattern on the A-ring (Fig. 10.2). In order to synthesize these cepharatines, selective cyclization reaction at the ortho-position in addition to the para-position is necessary. After investigation of various reaction conditions and substrates, we found that dienone 88 (corresponding to the A-, C-types) and 89 (corresponding to the B-, D-types) were obtained in a 1:1 ratio in 43% yield from 87 bearing a hydroxy group at C4 (Scheme 10.18). NMR experiments confirmed that the absolute stereochemistry at the C13 position of 89 was opposite to that of 88.

Fig. 10.2
Eight chemical structures. These are hasubanan skeleton, hasubanonine, metaphanine, stephadiamine, cepharatine A, cepharatine C, cepharatine B, cepharatine D, sinoracutine, acutumine, and acutuminine.

Structures of representative hasubanane alkaloids

Full size image
Scheme 10.18
A reaction pathway. The reactant has two substituted benzene rings bonded with C O T I P S C H 2. It reacts with P I D A, M e O H, and gives two products. The phenolic O H is converted into keto group. C 13 is bonded to C H 2 C H 2 N H B o c.

Oxidative phenolic coupling reaction of 87

Full size image

The diastereoselectivity of the products 88 and 89 of the oxidative coupling reaction can be explained as shown in Fig. 10.3. Considering the orientation of the C10 substituent of 87, four transition states, TS-5 to TS-8, are possible in which the orientation of the substituent at C10 is equatorial. In the coupling reaction at the ortho-position, leading to 88, there is steric repulsion between the C10 silyl ether and the bromo group on the A-ring in TS-6. Therefore, TS-5 is preferred, leading to the coupling product 88 with the anti-configuration of the C10 and C13 positions. On the other hand, in the coupling reaction at the para-position, steric hindrance occurs between the ethylamino side chain on the C-ring and the bromo group in TS-7. Therefore, the reaction proceeds preferentially through TS-8 to give coupling product 89 with syn-configuration at the C10 and C13 positions.

Fig. 10.3
Four chemical structures of possible transition states. T S 5 is the favored T S for C 10 and C 13 anti. T S 6 is disfavored T S for C 10 and C 13 syn. T S 7 is disfavored T S for C 10 and C 13 anti. T S 8 is favored T S for C 10 and C 13 syn.

Diastereoselectivity in the oxidative coupling reaction for 88 and 89

Full size image

10.3.7 Total Syntheses of (–)-Cepharatine A and C

Total synthesis of (–)-cepharatine A (38) and (–)-cepharatine C (40) based on the intramolecular aza-Michael reaction with dienone 88 was also examined (Scheme 10.19). Initially, the bromo group of 88 was removed with HCO2Na in the presence of Pd(PPh3)4 to afford the dienone 90. Previous study had revealed that the aza-Michael reaction at the C5-position proceeds preferentially under acidic conditions (Table 10.2). Thus, hydrochloric acid was applied to 90, and as expected, the aza-Michael reaction proceeded at the C5-position to give enone 91, followed by elimination of the C10 hydroxy group to afford conjugated dienone 92 in 61% yield. The phenolic hydroxy group of 92 was protected as methoxymethyl ether, and the resulting 93 was subjected to oxidation reaction at the C6 position. After several investigations, we found that oxaziridine rac-94 was effective in this case, and the α-hydroxy ketone 95 was obtained in 59% yield. After removal of the MOM and Boc groups with trifluoroacetic acid, the methyl group was introduced into the resulting amine by reacting with formaldehyde and NaBH3CN to furnish 96 in 41% yield. The synthesis of (–)-cepharatine A (38) was explored using 96, employing a cascade reaction that involved a retro aza-Michael reaction and subsequent hemiaminal formation (Scheme 10.20). Despite examining a various reaction conditions involving both acids and bases, the desired 38 was not produced. Instead, only the decomposition of compound 96 was observed. Interestingly, however, when 96 was left at room temperature for 48 h, the desired cascade reaction proceeded spontaneously, and 38 was obtained in 91% yield. Finally, (–)-cepharatine C (40) was synthesized from (–)-cepharatine A (38) by following Reisman’s protocol with sulfuric acid and trimethyl orthoformate [50].

Scheme 10.19
A reaction pathway. A dienone reacts with H C O 2 N A and the bromine atom is removed. It reacts with H C l and an intermediate is formed. It reacts with M O M C l, then with K H M D S, T F A, and gives the final products, cepharatine A and cepharatine C.

Total syntheses of (–)-cepharatine A (38) and C (40)

Full size image
Scheme 10.20
A reaction pathway. The reactant undergoes retro aza Michael reaction, keto enol isomerization, hemiaminal formation reactions to give the final product. The H of M e N H attacks the C of C double bond O to get the final product.

Proposed mechanism of 1,2-migration of the nitrogen atom at C5

Full size image

10.3.8 Total Syntheses of (+)-Cepharatine B and D

Total syntheses of (+)-cepharatine B (ent-39) and (+)-cepharatine D (ent-41) from enone 89 were also investigated (Scheme 10.21). Following the synthetic route shown in Scheme 10.19, α-hydroxyketone 99 was synthesized from dienone 89 in three steps. The Boc and MOM groups of 99 were then deprotected with trifluoroacetic acid. Interestingly, in the case of 99, a retro aza-Michael reaction/hemiaminal formation cascade reaction proceeded simultaneously with the deprotection step to give hemiaminal 100. Finally, methylation of the amino group of 100 gave (+)-cepharatine B (ent-39) in 37% yield. The phenolic hydroxy group of ent-39 was methylated by treatment with TMS diazomethane to afford (+)-cepharatine D (ent-41).

Scheme 10.21
A reaction pathway. The reactant contains TIP S O, O H, O M e, N H B o c groups. When it reacts with H C l, the N of N B o c forms a bind with C. This compound reacts with M O M C l. The O H is converted into O M O M. It is reacted with K H M D S, T F A, and H C H O. Cepharatine B and D are formed.

Total syntheses of (+)-cepharatine B (ent-39) and D (ent-40)

Full size image

10.4 Conclusion

In this chapter, we have described our efforts to synthesize fused polycyclic alkaloids based on the strategy of dearomatization and intramolecular aza-Michael reactions. Dearomative oxidative phenolic coupling proceeds with environmentally benign hypervalent iodine (e.g.,.PIDA) as an oxidant, giving tricyclic dienones with quaternary carbon centers. In the intramolecular aza-Michael reaction of the resulting dienones, we found that the reaction can be driven regioselectively toward either of two reaction sites with similar electronic states by choosing the appropriate reaction conditions. The present strategy showed that highly fused three-dimensional compounds can be efficiently synthesized from acyclic or aromatic compounds. This approach should also be applicable to the synthesis of natural products with various other complex frameworks.