Marine organisms contain various elements that are important for their survival and for their relationships with other living bodies. The nature, action and role of many of these elements have not yet been elucidated. Many marine organic compounds have been shown to be also useful for humans; as drugs, agricultural medicines, cosmetics and health foods [1–5]. In particular, a number of marine drugs have been developed in recent years. Marine natural products offer challenging targets to synthetic chemists due to their complicated structures that exhibit remarkable biological activities [6–11]. However, the chemical synthesis of marine drugs is of upmost importance as a synthetic supply is needed when it is difficult to maintain the amount of the natural source.

In the past several years, many new pyrroloiminoquinone alkaloids have been isolated from marine sources. The pyrroloiminoquinone alkaloid family consists of discorhabdins [12–16], prianosins [17], makaluvamines [18–21], batzellins [22–23], wakayins [24] and damirones [25–26], and have potent antitumor activities derived from their unique fused ring structures [27]. Among them, the discorhabdin alkaloids were isolated from marine sponges such as New Zealand sponges of the genus Latrunculia, the Okinawan sponge Prianos melanos, the Fijian sponge Zyzzya cf. Marsailis, etc. [12–16]. All of them have unique structures, with azacarbocyclic spirocyclohexanone and pyrroloiminoquinone structures. Figure 1 shows some representative discorhabdin alkaloids. Among the various isolated discorhabdins (A–X), discorhabdins A (1) [28,29], B (4) [28,29], D (3) [30], H [15], I [14], J [12], L [14], M [12], N [12], Q (5) [31], R [32], X (6) [33], and prianosins B (2), D [34] have a sulfur-containing fused ring structure. Discorhabdins S (11), T (12) and U (13) [13] have a methyl sulfide moiety. Discorhabdin W (14) [35] is a dimeric structure with a disulfide bond, while the others have no sulfur atom. Furthermore, discorhabdins F (9) [36], Q, S, T and prianosin B (2) contain a 16,17-dehydropyrroloiminoquinone moiety. Because of their prominent potent antitumor activity and unusual ring structure, pyrroloiminoquinone alkaloid synthesis has attracted the interest of many organic chemists, and over the last decade, the total synthesis of a few of them as well as synthetic approaches has appeared. Our group accomplished the first total syntheses of discorhabdin C (7) [37], makaluvamine F [38,39] and discorhabdin A (1) [40–43]. In 2009, we also accomplished the first total synthesis of prianosin B (2) [44]. In this review, we describe the total syntheses of the discorhabdins by various chemists to-date.

In 1986, discorhabdin C (7) was isolated from the sponge Latrunculia in New Zealand by Munro et al. and was found to exhibit a strong cytotoxicity against tumor cells (ED₅₀ = 30 ng/mL against P388 and <100 ng/mL against L1210) [45]. Discorhabdin C was the first isolated of the discorhabdins and much attention has been paid to the total synthesis of this attractive target.

Among the makaluvamines, makaluvamine F (48) has the most potent biological activities (e.g., cytotoxicity towards the human colon tumor cell line HCT-116 (IC₅₀ = 0.17 μM) and inhibition of topoisomerase II). Makaluvamine F (48) has an α-aminodihydrobenzothiophene skeleton, which is a labile N,S-acetal structure present in all sulfur-containing discorhabdins. Synthetic studies of the makaluvamines and discorhabdins have been carried out by several groups. However, in most cases, these efforts have been devoted only to the diverse preparations of the pyrroloiminoquinone and spirodienone units. To the best of our knowledge, the total syntheses of the sulfur-containing discorhabdins and makaluvamine F had not been reported until our synthesis, despite their potent cytotoxicity and their unique structure. This is probably due to the difficulty in construction of the labile and highly strained N,S-acetal skeletons. In our efforts to synthesize the discorhabdin alkaloids, we succeeded in the total synthesis of discorhabdin C (7) using the PIFA-mediated spirocyclization reaction. Thus, the remaining challenge for the total syntheses of these targets (the sulfur-containing family) was the construction of the labile and highly strained N,S-acetal skeletons. Our synthetic strategy for the total synthesis of makaluvamine F (48) is shown in Scheme 8.

In 1987, prianosin A [28] was isolated from the Okinawan sponge Prianos melanos by Kobayashi et al., and in 1988, discorhabdin A [29] was isolated from New Zealand sponges of the genus Latrunculia by Munro et al. Some years later, it was found that prianosin A and discorhabdin A were the same compound. Discorhabdin A (prianosin A) is antimicrobial and has a strong cytotoxicity (IC₅₀ values of 37 ng/mL, 14 ng/mL, 40 ng/mL and 13 ng/mL against L1210, L5178Y, P388 and xrs-6 in vitro, respectively, and ED₅₀ values of 50 ng/mL against P388). Our next synthetic target was discorhabdin A (1), which has a unique sulfur-containing fused ring system incorporating azacarbocyclic spirocyclohexanone and pyrroloiminoquinone systems, and shows the most powerful cytotoxic activity among the isolated discorhabdins.

First, on the basis of the hypothesis by Munro and co-workers [56], we examined the biosynthetically plausible route from makaluvamine F (48) using our previously developed oxidative spirocyclization reaction with PIFA. As a result, the oxidative cyclization of makaluvamine F (48) as well as the trimethylsilylated makaluvamine F using PIFA under various conditions yielded a complex mixture, probably due to the high reactivity of the iodine(III) reagent toward the sulfide group, the aminoiminoquinone skeleton, and the phenolic OH group in makaluvamine F (48) (Scheme 18).

Therefore, we altered the synthetic strategy. The new strategy included the pre-construction of the spirodienone system using the hypervalent iodine(III) reagent and the final introduction of the sulfur group to the cross-linked system. We first examined the shorter path via an amino acid derivative. The tritylation of the l-tyrosine methyl ester hydrochloride (73) with TrCl and Et₃N followed by mono-bromination with NBS yielded 74 (two steps, 65% yield). The coupling reaction of 74 with pyrroloiminoquinone 25, which was prepared by our previously developed PIFA-induced pyrroloiminoquinone formation, provided 75 as its CF₃CO₂H salt (75·TFA). We then examined the oxidative spirocyclization reaction of 75·TFA using PIFA. Although the various reaction conditions tested, this reaction did not give the corresponding spirodienone, but yielded a complex mixture; probably because generation of the free base of 75 was difficult. This might be because 75·TFA has an active methine proton and readily decomposed under basic conditions (even in the presence of Et₃N). Thus, we modified the synthetic strategy using an alternative route via the amino alcohol as follows: reduction of 74 with DIBAH followed by silylation of the resulting alcohol with TBSCl gave the bis-silylated compound 76, and a coupling reaction with pyrroloiminoquinone 25 yielded 77. The spirodienone formation of 77 using PIFA effectively proceeded in the presence of MK10 to give a diastereomeric mixture of 78 and 78′. Both diastereomers were readily separated by column chromatography on silica gel to give the less polar isomer 78 and the polar isomer 78′ (Scheme 19).

The major isomer 78, whose absolute stereochemistry (S configuration) of the spirocenter (C-6) is the same as that of natural discorhabdin A (1), was desilylated by BF₃·Et₂O, and then converted into the N,O-acetal intermediate 79 by an oxidative fragmentation reaction with C₆F₅I(OCOCF₃)₂ and MeOH. The p-methoxybenzylthio group was efficiently introduced in the presence of BF₃·Et₂O to give the unstable N,S-acetal A as a diastereomeric mixture. The debenzylation of A, a labile and highly functionalized compound, also required the mildest possible reaction conditions. Our initial strategy was to perform a mild debenzylation on the p-methoxybenzylsulfonium salt formed by the 1,4-addition of a sulfide group to the cyclohexadienone moiety. Accordingly, we treated a diastereomeric mixture of A with 30% HBr-AcOH followed by aqueous work up with MeNH₂ that produced the N-tosylated discorhabdin A (80). Ultimately, we developed an efficient one-pot transformation procedure yielding 80 from 79. The procedure using p-methoxybenzylthiol in 30% HBr-AcOH (36 h; −78~4 °C) followed by treatment with aqueous MeNH₂ mainly gave 80: the enantiomeric excess (ee) of 80 was confirmed to be >99% by HPLC analysis using a chiral column (DAICEL CHIRALCEL OD; n-Hex/i-PrOH = 65/35; 20 °C). Finally, removal of the tosyl group of 80 with NaOMe in THF gave discorhabdin A (1) in the optically pure form for the first time (Scheme 20. The synthetic product as its HCl salt was identical in all respects with the natural discorhabdin A including its optical rotation {synthetic discorhabdin A: [α]_D +388 (c = 0.06, MeOH); natural discorhabdin A: [α]_D +400 (c = 0.05, MeOH)} and ¹H NMR nOe data.

Furthermore, we also completed the total synthesis of the unnatural discorhabdin A, (−)-discorhabdin A, starting from the d-tyrosine methyl ester by the same route as described for (+)-discorhabdin A (1) {synthetic product: [α]_D -390 (c = 0.05, MeOH)}.

In 2006, Copp et al. reported the semi-synthetic preparation of discorhabdins P (81) and U (13) from natural discorhabdins C (7) and B (4). In New Zealand waters, Latrunculia sp. sponges are a rich source of the discorhabdin alkaloids, with organisms capable of producing either discorhabdins A (1), B (4) or C (7) as their major secondary metabolites, typically in isolated yields of 25–35 mg/g dry sponge weight [58]. The aim of the Copp et al. study was to investigate the methylation of discorhabdin C (7) in an effort to prepare discorhabdin P (81) or related N-methyl analogs for further biological testing, and for the first time, to prepare similar analogs of discorhabdin B (4) in an effort to expand the structure-activity relationship understanding of this class of compounds. Discorhabdin C (7) was reacted with CH₃I in dry acetone to yield a single major product, the 13-N-methyl analog discorhabdin P (81) in 54% yield (Scheme 21). Discorhabdin B (4) was reacted with CH₃I in dry acetone to yield two products, discorhabdin U (13) and 82. The products of this reaction were dependent upon the molar equivalents of CH₃I used: with a large excess (10–25 equivalent) yielding predominantly discorhabdin U (13), while a two molar equivalent of CH₃I yielded the mono-methyl 82. The order of the methylation of discorhabdin B (4) preferentially favors the thio group and a large excess of CH₃I enables the methylation at the N-13 pyrrole position.

Although, some 20 years ago, Kobayashi et al. also reported one such family, prianosin B (2), which has the 16,17-dehydropyrroloiminoquinone structure, its synthesis has not yet been reported [17]. For the synthesis of prianosin B (2), the construction of the 16,17-dehydropyrroloiminoquinone structure is an important issue. We have succeeded in the first total synthesis of the sulfur-containing discorhabdin, discorhabdin A [40–43]. We postulated that the prianosins and discorhabdins with the 16,17-dehydropyrroloiminoquinone moiety can be synthesized by dehydrogenation of the pyrroloiminoquinone of the discorhabdin A (1) or discorhabdin A intermediate (Scheme 22).

We first examined the dehydrogenation reaction of the simple spirodienone (83) with various oxidants as the model reaction. However, reactions with oxidants, such as DDQ, CAN, MnO₂ and Pd/C (Yamamura’s conditions) [59], gave poor results (Scheme 23).

We next tried the use of nucleophiles, because White et al. reported that the dehydrogenation and detosylation of the pyrroloiminoquinone system using sodium azide as a nucleophile [60]. From a detailed study of nucleophiles, we found that the use of a catalytic amount of NaN₃ produced the best result. Since it was revealed that the detosylation and dehydrogenation reaction of the pyrroloiminoquinone unit with a catalytic amount of NaN₃ proceeded in good yield, the total synthesis of prianosin B (2) was next studied (Scheme 24). In our own total synthesis of discorhabdin A (1), the N,O-acetal intermediate (79) was prepared from the l-tyrosine methyl ester hydrochloride in eight steps. The sulfur cross linkage reaction of the intermediate was carried out using p-MeOBnSH and 30% HBr-AcOH followed by aqueous MeNH₂ to give the sulfur-linked pyrroloiminoquinone compound 80. The treatment of 80 with NaN₃ in DMF caused detosylation and dehydrogenation to produce prianosin B (2) in 48% yield. The spectral data were identical to that reported for the natural prianosin B (2) ([α]²³_D +362 (c 0.405, CHCl₃), lit. [α]³⁰_D +360 (c 0.1, CHCl₃) [17]).

For determining the reaction mechanism, two studies were carried out. One was the measurement of the mass spectrum of the reaction mixture. A suspension of spirodienone 83 and NaN₃ in DMF was stirred under nitrogen at 70 °C for 40 min. The mass spectrum of the reaction mixture revealed the ion peak of TsN₃. The other was the reaction of the N-H spirodienone 85 and TsN₃ under basic conditions. Thus, NaH and TsN₃ were added to a solution of 85 in DMF and the aromatic compound 86 was obtained in 41% yield (Scheme 25).

A plausible reaction mechanism for the present dehydrogenation reaction of pyrroloiminoquinone using NaN₃ is illustrated in Scheme 26. It would involve the nucleophilic attack of N₃⁻ on the tosyl residue followed by the addition of metalo enamine to TsN₃ to produce the intermediate I [61–63]. Dehydrogenation would proceed by the intramolecular elimination via a six-membered transition state to produce the intermediate III and isomerization to produce 86 [64–65]. Reproduction of the azide anion during the reaction would make possible the use of the catalytic amount of NaN₃. The reproduction of the azide anion was deduced from the presence of toluenesulfinic acid observed in the ¹H-NMR spectrum (Scheme 26).

In this review, we described the existing studies on the total synthesis of the discorhabdins. Because of their prominent potent antitumor activity and unusual ring structure, synthesis of the discorhabdins has attracted the interests of many organic chemists. Many synthetic approaches have appeared over the past decade, however, only few discorhabdins have been completely synthesized. We have accomplished the total syntheses of discorhabdins C, A and prianosin B. Discorhabdin A has the strongest activity in vitro, but no activity in vivo. Therefore, we are synthesizing a variety of discorhabdin analogs [66]. Although many synthetic studies of the discorhabdin alkaloids have already appeared, biological studies including structure activity relationship or the mode of action of the discorhabdins have not yet significantly occurred [67–71]. In the next stage, it is necessary to study the structure activity relationship and the mode of action of the discorhabdins.