On account of the variation for the sugar chain, glycosphingolipids (GSLs) are classified into cerebroside, sulfatide, ceramide oligohexoside, globoside, and ganglioside. Gangliosides, sialic acid-containing GSLs, are especially enriched in brains and nervous tissues, and are involved in the regulation of many cellular events. Recently, a number of GSLs have been obtained from the marine invertebrates such as echinoderms, poriferans, and mollusks. We have also been researching the biologically active GSLs from the echinoderms for elucidating the structure-function relationships of GSLs, and for developing the novel medicinal resources. In this time, we conducted the isolation and characterization of biologically active ganglioside from the sea cucumber Stichopus japonicus and could obtain two ganglioside molecular species, SJG-1 and SJG-2. The structures of these ganglioside were characterized on the basis of spectral (^1H-NMR, ^<13>C-NMR, ^1H-^1H COSY, HSQC, and negative ion FAB mass spectra) and chemical (methanolysis and methylation analysis) evidences. Namely, SJG-1 and SJG-2 are ganglioside molecular species possessing phytosphingosine-type long chain base and nonhydroxy fatty acid. The sugar moiety of SJG-1 is unique in respect of that a sialic acid directly binds to glucose of cerebroside, while that of SJG-2 is quite unique differing from those of gangliosides obtained from not only other sea cucumbers, but vertebrate animals. The gangliosides SJG-1 and SJG-2 shows neuritogenic activities toward the rat pheochromocytoma cell line PC12 cell at 10μM in vitro, especially SJG-2 reveals the most potent activities in the ganglioside which have been obtained from echinoderms in our laboratory.
In the tropical zones of the Indian and Pacific Oceans, the crown-of-thorns starfish Acanthaster planci destroys coral by eating it, but there is no effective extermination of the starfish. Okinawa prefecture have conducted extermination of the starfish at a great cost. We sought to establish the method for clean, effective and ecological extermination of the starfish Acanthaster planci. We recently found that the sea urchin Toxopneustes pileolus had a feeding stimulant for this starfish. We established a simple bioassay system in the aquarium to study this stimulant, which was found to be composed of two unsaturated fatty acids, arachidonic acid (1) and α-linolenic acid (2). Since commercially available 1 and 2 also exhibited attracting activitiy for the starfish, 1 and 2 were found to be active principles from the sea urchin Toxopneustes pileolus. Other unsaturated fatty acids such as eicosapentaenoic acid (3), docosahexaenoic acid (4), γ-linolenic acid (5), and linoleic acid (7) showed no activity. We have set traps with α-linolenic acid for catching the starfish Acanthaster planci in the Okinawan sea at the depth of 1.5-2.5 meters. Size of the trap is 1.5 meters square. We could catch several starfishes in the sea. Studies on a receptor of the starfish for arachidonic acid and α-linolenic acid and pheromonal function of arachidonic acid and α-linolenic acid are currently in progress.
Amphidinolides G (1) and H (2), potent cytotoxic 27- and 26-membered macrolides, respectively isolated from marine dinoflagellates Amphidinium sp. (strain number Y-25). Recently we found new strains (Y-72 and Y-42) of the genus Amphidinium producing relatively large amount of 1 and 2. Here we describe the determination of relative and absolute stereochemistry of these compounds on X-ray diffraction analysis, synthesis of degradation product (3) of amphidinolide H (2), and interconversion between 1 and 2. We also isolated six new macrolides 3〜8 related to amphidinolides G (1) and H (2) from the Y-42 strain of dinoflagellates Amphidinium sp. The structures of compounds 3,4,5, and 6 were elucidated as 22-epi, 18-epi, 6,7-dihydro, and 6,7-dihydro-l8-epi forms of amphidinolide H (2), respectively, on the basis of detailed analyses of spectroscopic data including ^3J_<H,H>,^2J_<C,H>, and ^3J_<C.H> values and distance geometry calculation. On the other hand, the structures of compounds 7 and 8 were assigned as the 22-epi and 6,7-dihydro forms of amphidinolide G (1). Compounds 3 and 7 were as potent cytotoxic against L1210 and KB cells as amphidinolides G (1) and H (2), while cytotoxicity of compounds 4,5,6, and 8 were extremely weaker (1/100〜1/1000) than those of 1 and 2. These results indicated that allyl epoxide and stereochemistry of C-18 were important for its activity. The biosynthetic origins of amphidinolides G (1) and H (2) were investigated on the basis of 2D NMR data of ^<13>C-enriched samples obtained by feeding experiments with [1-^<13>C], [2-^<13>C], and [1,2-^<13>C_2] sodium acetates in cultures of a marine dinoflagellate Amphidinium sp. These incorporation patterns suggested that 1 and 2 were generated from three unusual C_2 units derived only from C-2 of acetates in addition to three successive polyketide chains. Furthermore it is noted that six oxygenated carbons of C-1, C-18, C-20, C-21, C-22, and C-26 in 1 and 2 were not derived from the C-1 carbonyl but from the C-2 methyl of acetates.
Recent progresses in glycobiology have shed light on significant roles of glycosidases in various biological functions, including immune response, oncogenesis, metastasis of tumors, viral and bacterial infections, and differentiation of neural cells. Specific inhibitors of glycosidases have potentials for treatment of a variety of diseases. α-Glucosidases are involved in glycoprotein processing and glycogenolysis and their inhibitors can be applied for treatment of diabetes, obesity, viral infections, and cancer. In our screening for α-glucosidase inhibitors from Japanese marine invertebrates, we encountered a marine sponge Penares sp. collected off Hachijo-jima Island whose hydrophilic extract was highly active. Bioassay-guided fractionation led to the isolation of two active compounds, penarolide sulfates A_1 and A_2. The molecular formula of penarolide sulfate A_1 (1) was determined to be C_<35>H_<62>NO_<15>S_3Na_3 on the basis of HR-FABMS and NMR data. Interpretation of 2D NMR spectra indicated that the structure of 1 was composed of partial structures a and b which were connected through long alkyl chains. Positions of three sulfate groups and length of alkyl chain were unambiguously determined by FAB-MS/MS experiments. The absolute stereochemistry of 1 was determined by combination of Marfey analysis and modified Mosher's method. Penarolide sulfate A_2 (2) had a molecular formula identical to that of 1, as determined by the HR-FABMS and NMR data. The structure of penarolide sulfate A_2 (2) was determined following the same procedure as in the case of 1. Penarolide sulfates A_1 (1) and A_2 (2) inhibited α-glucosidase with IC_<50> values of 1.2 and 1.5μg/mL, respectively; 30-40 times more potent than a known inhibitor 1-deoxynojirimycin. Interestingly, they were only marginally active against β-glucosidase at a concentration of 100μg/mL; castanospermine and 1-deoxynojirimycin showed IC_<50> values of 15 and 96μg/mL, respectively. All four compounds did not inhibit β-galactosidase at 100μg/mL. Because penarolide sulfates inhibited thrombin as did other sulfated compounds, it is of interest to examine whether the sulfate groups in these compounds are indispensable for their enzyme inhibitory activities.
Marine sponges are known to be treasure houses of bioactive metabolites, many being cytotoxic, to natural product chemists. This arises a question why sponges are all right with strong cytotoxins. The answer could be either reduced affinity between the toxin molecule and its target cellular component, and/or existence of another binding molecule for the toxin to lessen its availability as the free form. In order to investigate such speculations, proteinaceous components in the aqueous extract of the sponge Halichondria okadai were examined in terms of binding affinity to a polyethereal cytotoxin, okadaic acid (1), because this sponge is known to contain an abundant amount of 1. After preliminary fractionation by ammonium sulfate precipitation, considerable activity, which was further enhanced by removal of 1 with acetone precipitation, was perceived using the conventional binding assay using a tritiated 1 (2) and separation between the bound and the unbound by quick gel filtration. Active fractions were further separated through ion-exchange and other chromatographic techniques as guided by the same binding assay, eventually yielding two entities with molecular sizes of ca. 37kDa (OABP1) and 25kDa (OABP2) as indicated on SDS-PAGE. Conventional sequential analyses of their endoproteinase digests and the following PCR protocol indicated OABP1 to be the catalytic subunit of protein phosphatase 2A (PP2A), the known target enzyme of 1, but with much reduced affinity than mammalian ones, while OABP2 failed to show any homology. Based on the ample amount of the latter and on its low Kd of 0.9nM, this protein is most likely involved in sequestration of 1 from its action to PP2A. A complete amino acid sequence of OABP2 was obtained by 5'- and 3'-RACE techniques. Combined with its ESI mass spectrum showing multi-charged ions, deletion of amino-terminal methionine is the only post-translational modification to result in a protein of 189 amino acid residues.
In the course of our study of bioactive substances from marine organisms, we focused on a search for reversing substances of multidrug-resistance (MDR) in tumor cells and isolated agosterol A (1) from a marine sponge Spongia sp., Agosterol A having a polyhydroxylated sterol acetate structure reversed MDR caused by overexpression of membrane glycoproteins, P-glycoprotein (P-gp) or multidrug-resistance-associated protein (MRP1). We examined the action of agosterol A for both the P-gp-mediated MDR cells (KB-C2) and the MRP1-mediated MDR cells (KB-CV60). The accumulation of vincristine in both KB-C2 cells and KB-CV60 cells was recovered to the level equal to that in parental KB-3-1 cells by the treatment of agosterol A. Agosterol A also decreased the efflux of vincristine from both KB-C2 cells and KB-CV60 cells to the level equal to that from KB-3-1 cells. Agosterol A inhibited the [^3H]azidopine photolabeling to P-gp and also inhibited the uptake of [^3H]DNP-SG in the membrane vesicles prepared from KB-CV60 cells. From the above findings, it was indicated that agosterol A inhibited the drug transportation through P-gp and/or MRP1 by affecting those drug efflux pumps directly. In addition, the glutathione content in KB-CV60 cells was decreased by agosterol A. The syntheses of 4-deacetoxyagosterol A (7) and its 22-dehydroxy-analogue (22) were achieved by using ergosterol and 7,8-didehydrocholesterol as starting material, respectively. Furthermore, we synthesized the ^<125>I-labeled photoaffinity probe 23 of agosterol A and succeeded in photolabeling of MRP1. Interestingly, 23 showed the affinity to MRP1 only in the presence of glutathione. Vincrisitine and doxorubicin inhibited the binding of 23 to MRP1.
In our continuing studies on bioactive limonoids from Meliaceae plants, we isolated twelve new rings B,D-seco limonoids, which were four mexicanolide types compounds, named khayanone (4), kyayalactol (5) and seneganolides (6 and 7), seven rearranged phragmalin compounds, named khayanolides (8-14), and one glucoside, named khayanoside (15), together with three known compounds, methyl angolensate (1) and its 6-hydroxy (2) and 6-acetoxy (3) derivatives, as insect antifeedant from the stem bark of a mahogany Khaya senegalensis (Meliaceae) collected at Alexandria in Egypt. The relative structures of the new compounds were elucidated on the basis of spectroscopic means, containing ^1H-^1H COSY, HMQC, HMBC and NOESY NMR techniques, and the stereochemistry at C-6 in 4 and 8 was established as S by exciton chirality method in CD and X-ray crystal analysis, which is particularly of interest in contrast to R configulation in the compounds from a swietenia. In this study, compound 4 yielded acetal compounds at C-1 by the reaction with MTPA-chlorides or p-bromobenzoyl chloride. Conformational analysis and the application of dibenzoate chirality methord for the resulting 1,6-dibenzoate, elucidated the configuration of C-6 as S. Although many rings B,D-seco limonoids have been isolated, seneganolide (4) is the first report of C-19 oxygenated compound in mexicanolides and kyayanolide (15) is also the first occurrence of glycoside in rings B,D-seco limonoids. The most interesting question about rearranged phragmalin limonoids, khayanolides, is how they are formed from mexicanolides in the plant. Two possible pathways leading to the formation of khayanolides are proposed by modifying that proposed by Taylor for the biosynthesis of the phragmalin limonoid. Antifeedant activity of the isolated compounds was tested by a conventional leaf disk method against the larvae of Spodoptera littoralis (Boisduval). All of the new compounds showed moderate activities at 300-500ppm concentrations, which are comparable to those of azedarachins and trichilins from Melia species.
Extracts of medicinal plants were examined for the effects on the NGF-mediated neurite outgrowth from PC12D cells to evaluate their NGF-potentiating activities. In the methanol extracts, Gymmopteris rufa (LINN.) BERNH, Ruta graveolens LINN. and Picrorhiza scrophulariiflora PENNELL increased markedly the proportion of the neurite-bearing cells. In the case of ethyl acetate fractions. Equisetum giganteum LINN. produced the most powerful enhancement of the proportion of the neurite-bearing cells. In the water fractions. Imperata cylindrica, Ginseng Radix, Gymmopteris rufa (LINN.) BERNH, Gochnatia polymorpha (LESS) CAB and Picrorhiza scrophulariiflora PENNELL caused a weakly enhancement of the proportion of PC12D cells with neurites. We successfully isolated nardosinone, picrosides I and II. gersemiol and 9-hydroxysemperoside aglucone as enhancers of NGF-action. We studied the effects of nardosinone. picrosides. gersemiol and 9-hydroxysemperoside aglucone on NGF-induced neurite outgrowth and associated transmembrane signal transduction in PC12D cells. These compounds did not exhibit the neurotrophic activity but caused a concentration-dependent enhancement of the NGF-induced neurite outgrowth from PC12D cells. Nardosinone- and picrosides-induced enhancements of the NGF-action were abolished by GF109203X, a protein kinase C inhibitor. Furthermore, PD98059. a potent MAPK kinase inhibitor, completely blocked nardosinone- and picrosides-induced enhancements of the neurite outgrowth in the presence of NGF. These results suggest that nardosinone and picrosides activate PKC-MAPK-dependent signaling pathway. Interestingly, no increases in expression of phosphorylated MAPK were observed in nardosinone and picrosides-treated PC12D cells in the presence of NGF. On the basis of pharmacological data, it is suggested that nardosinone, picroside I or II enhances the NGF-induced neurite outgrowth from PC12D cells probably by amplifying a down-stream step of MAPK in the NGF receptor-mediated intracellular PKC-MAPK-dependent signaling pathway.
"Kwao Keur." which had been identified as Puraria mirifica (Leguminosae), has long been used in Thailand and Burma as a rejuvenating folk medicine and has a fascinating history. A potent estrogenic principle has been known to be an unusual phenol miroestrol (1). Although the possible presence of an alternative active component was suggested, there has been no isolation of any other powerful phytoestrogens. with further studies leading instead to the isolation of isoflavonoids. These situation made us re-investigate the estrogenic principles of P. mirifica. The bioassay-guided separation of the ethyl acetate extract of the tuberous roots of P. mirifica by chromatographic techniques resulted in the successful isolation of a new phytoestrogen. (+)-deoxymiroestrol (4). together with (+)-1 and (+)-isomiroestrol (7). The structure of (+)-1 had been determined by X-ray crystallographic analysis and its enantioselective total synthesis was recently reported. Thus. the structure of (+)-deoxymiroestrol (4) was established by comparison of its NMR data with those of (+)-1. The growth-promoting effect of them on MCF-7 human breast cancer cells showed the strongest activity with 4. Interestingly. 4 was easily converted into 1 and 7 by aerial oxidation. suggesting that 4 may be the actual phytoestrogen of P. mirifica. On the other hand daidzein (2), genistein (3), and coumestrol (6) belong to isoflavonoids were isolated as phytoestrogens with lower activity. In addition, it was found that kwakhurin (5). a characteristic isoflavonoid in this plant. also showed the same activity as 2.
Rose breeders have failed to make blue roses. This has been attributed to the lack of blue pigment in the petals. We revealed that mauve rose such as "M'me.Violet" and "Lavande" contain a small amount of blue pigment other than a red anthocyanin, cyanidin 3,5-diglucoside. The major blue pigment of the rose named Rosacyanin A and the minor red one was named Rosacyanin B. The structures of these pigments were elucidated. A high-resolution mass spectrometry showed that Rosacyanin B had the molecular weight of 419.0409 and molecular formula of C_<22>H_<11>O_9. The NMR data showed that Rosacyanin B had an extremely unique structure whose C-1 position of gallic acid is bound to the C-4 position of cyanidin by C-C bond formation. (Fig.2) Rosacyanin A has λ max 590nm (MeOH) of the ultraviolet and visible absorption spectrum, and a molecular formula of C_<56>H_<37>O_<31> which is calculated from the molecular weight of 1205.1319 obtained from high-resolution mass spectrometry. As a result of the observation of the isotope shift by the DH exchange of the solvent using a coaxial sample tube in ^<13>C NMR, it was found that the 3-position of flavylium of Rosacyanin B is bonded to the hexahydroxydiphenoyl part of Tellimagrandin II which is a kind of ellagitannin with ether linkage. (Fig.4) To our knowledge, this is the first report of the compound whose gallic acid binds to C-4 position of polyhydroxyflavylium. The only similar compound which binds gallic acid to catechin was obtained from Burkea africana and Peltophorum africanum. The dream of blue roses will come true if we can accumulate Rosacyanin A in rose petals.
Chitinase is an essential enzyme for the insect ecdysis. Inhibitors of chitinase would be expected to interrupt insect moulting and thus prevent maturation to adult reproductive stage. In the course of screening for chitinase inhibitors, we have found new compounds named argifin (1) and argadin (2) from the cultured broths of Gliocladium sp. FTD-0668 and Clonostachys sp. FO-7314, respectively. The structures of 1 and 2 were elucidated by amino acid analysis, NMR experiments including ^1H-^<13>C and ^1H-^<15>N HMBC, and other spectroscopic analyses. The structure of 1 was shown to be cyclo(N^ω-(N-methylcarbamoyl)-L-arginyl-N-methyl-L-phenylalany1-β-L-aspartyl-β-L-aspartyl-D-alanyl). The β-amide bonds of two aspartic residues were revealed by highfield shifts of free carboxylic carbons in acidic solution. The structure of 2 was deduced to be cyclo(N^ω-acetyl-L-arginyl-D-prolyl-homoseryl-histidyl-L-2-aminoadipyl) in which homoseryl γ-methylene bonded to histidyl α-amino residue to form a γ-lactam. The γ-lactam structure was confirmed by small shifts of homoseryl carbonyl carbon measured in D_2O alone and H_2O-D_2O, which suggested it bonded tertiary nitrogen. The IC_<50> values of 1 and 2 against Lucilia cuprina (blowfly) chitinase at 20℃ were 100 and 3.4nM, respectively. Their inhibitions were reduced about forty times at 37℃. They are the first chitinase inhibitors produced by fungi and showed inhibition in a submicromolar range. They arrested the moult of cockroach larvae upon injection into the ventral abdominal part. Therefore, cyclic pentapeptide containing N^ω-substituted-L-arginine could be an interesting lead for the development of novel insecticides.
We present two new NMR techniques prepared by modification of HMBC, Constanttime HMBC (CT-HMBC) and J-resolved HMBC. CT-HMBC One problem of HMBC is that J-modulation due to H-H couplings cause line broadening of the ^<13>C-signals in the F_1 axis through coupling to the relevant protons during the t_1 period. This undesirable effect together with poor separation of carbon signals of complicated molecules sometimes makes strict assignments of cross peaks difficult. In order to solve this problem, we have developed a new technique, constant CT-HMBC by incorporation of the constant time method to HMBC. This modification suppresses the line broadening of cross peaks resulting in improvement of separation of carbon signals in the F_1 axis. Its application to complicated molecules proved the usefulness of this new technique. J-resolved HMBC In the conventional HMBC spectra, the cross peaks contain information on the long-range C-H J splittings and/or H-H J splittings. However, due to the poor resolution in the F_2 or F_1 dimension, it is difficult to measure directly the long-range C-H J couplings from HMBC cross peaks. In order to solve this problem, we have incorporated the J-scaling method into HMBC and developed a new technique, J-resolved-HMBC. This modification enables to determine easily long range C-H J couplings of complicated compounds and will be useful for conformational analysis of non-cyclic systems. Advantages of this method are simple operation, high sensitivity and general applicability.
Recently, we have found that 2-methoxy-2-(1-naphthyl)propionic acid (1, MαNP acid, Chart 1) was very useful for determination of absolute configurations of chiral alcohols including natural products. Racemic MαNP acid (1) was enantioresolved as its esters with various chiral alcohols. For example, a diastereomeric mixture of esters prepared from (±)-1 and (1R,3R,4S)-(-)-menthol (Scheme 2) was easily separated by HPLC on silica gel yielding esters (-)-7a and (-)-7b, the separation factor a being unusually large as 1.83 (Figure 1). The data of ^1H NMR chemical shift difference, Δδ=δ(R) ? δ(S), between diastereomers 7a and 7b, are much larger than those of conventional chiral auxiliaries, e.g., Mosher's MTPA and Trost's MPA acids (Figures 2-4). This acid 1 is therefore very powerful for determining the absolute configuration of chiral alcohols by the ^1H NMR anisotropy method (Figures 3 and 5). Solvolysis of the separated esters with NaOCH_3 in methanol and then with water, yielded enantiopure acid (S)-(+)-1, [α]D^<26> +67.4 (c 1.39, CHCl_3), and (R)-(-)-1. During those studies, we have also found that the acid 1 was very powerful for enantioresolution of various alcohols, especially aliphatic alcohols (Figure 6). The chiral acid 1 is a regular organic compound containing no hetero atoms, but its performance as a chiral auxiliary for enantioresolution of alcohols is great as shown below. The strong power of this acid 1 for diastereomers separation is demonstrated by the case of 2-butanol. Racemic 2-butanol was esterified with the chiral acid (S)-(+)-1; the obtained mixture of diastereomeric esters 8a and 8b was almost base-line separated by HPLC on silica gel (hexane/EtOAc 20: 1) as shown in Figure 6 (a): α=1.15. The acid 1 has thus a great ability to recognize the small difference between methyl and ethyl groups. As well known, chiral discrimination between methyl and ethyl groups is the most difficult in enantioresolution and also even in asymmetric reactions. Now we obtained a facile method to get enantiopure 2-butanol. We have applied this method to various 2-alkanols and 1-octyn-3-ol, which have been successfully resolved as shown in Figure 6. To determine the absolute configurations of the first-eluted diastereomeric esters by the ^1H NMR anisotropy method, the general scheme was proposed as shown in Figure 7. By applying the scheme, the absolute configurations of the first-eluted esters were determined as exemplified in Figure 8. To recover enantiopure alcohols, separated esters were reduced with LiAlH_4 or hydrolyzed with KOH/EtOH. When hydrolyzed with KOH/EtOH, the recovered acid 1 could be recycled for another resolution. The use of chiral MαNP acid 1 thus enables one to obtain enantiopure alcohols, and also to determine their absolute configurations simultaneously by the ^1H NMR anisotropy method.
Several eubacteria including Escherichia coli utilize an alternative mevalonate-independent pathway (nonmevalonate pathway) for the biosynthesis of isopentenyl diphosphate (IPP). The initial step of the nonmevalonate pathway is the formation of 1-deoxy-D-xylulose 5-phosphate (DXP) by condensation of pyruvate and glyceraldehyde 3-phosphate catalyzed by DXP synthase. In the second step, the intramolecular rearrangement and reduction of DXP occur simultaneously to yield 2-C-methyl-D-erythritol 4-phosphate (MEP). However, the following reactions leading to IPP from MEP remained unknown. To elucidate the remaining reactions, we employed a unique strategy for preparation of mutants of E. coli possessing a metabolic block(s) between MEP and IPP. Since such mutations would be lethal, we constructed an E. coli transformant expressing mevalonate kinase, phosphomevalonate kinase and diphosphomevalonate decarboxylase. Using this transformant as the parent strain, we prepared mutants with an obligatory requirement of mevalonate for growth. Thirty-three mutants from 60,000 colonies screened showed the expected phenotypes. With these mutants in hand, we were able to clone three genes, YACM, ychB and ygbB, that complemented the defect of each blocked mutant in synthesizing IPP from MEP. To identify the functions of these genes, plasmids for overexpression of these gene products were constructed. We found that the YACM gene product catalyzed a cytidylyl transfer to MEP to synthesize 4-(cytidine 5'-diphospho)-2-C-methyl-D-erythritol (CDP-ME). Next, the ychB gene product catalyzed phosphorylation of the tertiary hydroxy group in the CDP-ME molecule to produce 2-phospho-4-(cytidine 5'-diphospho)-2-C-methyl-D-erythritol (CDP-ME2P). Further, the ygbB gene product converted CDP-ME2P into 2-C-methyl-D-erythritol 2,4-cyclodiphosphate (MECDP).
Spicamycin (1), a novel nucleoside antibiotic isolated from Streptomyces shows potent antitumor activities. The structure of 1 is quite unique among the nucleoside antibiotics: while conventional adenine-nucleoside antibiotics bear a furanose at the N(9) position, 1 is glycosylated at the C(6)-amino group of adenine. Moreover, the sugar portion in 1 is a novel aminoheptose with the β-manno anomeric configuration. Here, we report the first total synthesis of one of congeners of spicamycin (SPM VIII, 3). The aminoheptose moiety (21) was synthesized stereoselectively from naturally abundant cyclitol, myo-inositol, and the key transformation, construction of the characteristic N-glycoside was accomplished by Pd-catalyzed coupling reaction of pyranosylamine (24) with 6-chloro-9-SEM-purine (7). The known racemic diol (±-11), prepared from myo-inositol in a one-step reaction, was converted into protected diol (±-12), which was optically resolved via corresponding (+)-O-acetylmadelates. The optically pure diol (12D) was transformed into azide-cyclitol derivative (16) in conventional 6 steps reactions. Removal of O-PMB and O-acetyl groups in 16 gave triol (17), whose glycol cleavage by periodate oxidation, followed by reduction with NaBH_4 provided acylclic triol (18). Acetonide formation and subsequent Moffatt oxidation generated unstable aldehyde (19), which was reacted with vinyllithium, followed by O-benzylation to afford Felkin-Anh product (20), stereoselectively. Ozonolysis of 20, followed by acid hydrolysis and subsequent acetylation afforded heptopyranosyl acetate (21). Treatment of 21 with TMSBr gave α-bromoheptose derivative, which was then reacted with TMSN_3 in the presence of TBAF to afford β-anomeric azide (22). Hydrogenation of diazide 22 in the presence of Lindlar catalyst provided β-heptopyranosylamine (23). Reaction of amine 23 with 7 in the presence of Pd_2(dba)_3 in toluene at 130℃ afforded the coupling product 24. Deacetylation of 24, followed by deprotection with BBr_3 provided spicamycin amino nucleoside (SAN, 2). Introduction of acyl side chain into 2 furnished the first total synthesis of SPM VIII (3).
Globomycin (1) was isolated as a 19-menbered cyclic depsipeptide antibiotic which consists of glycine, L-allo-isoleucine, L-allo-threonine. L-serine, N-Me-leucine, and 3-hydroxy-2-methylnonanoic acid, in 1978. 1 has been proven to be a specific inhibitor of Signal peptidase II, which processes the acylated precursor form of lipoproteins into apolipoprotein and signal peptide in Esherichia coli. 1 was used routinely to demonstrate the acylation of the newly identified lipoprotein, and since then it has been widely used to control the maturation of the lipopeptides. Although 1 has been an invaluable tool in studies of lipoprotein biosynthesis to date, its structure remains obscure for almost two decades. An initial structure elucidation of 1 only determined the absolute stereochemistry of L-allo-Thr-L-Ser-L-allo-Ile moiety. Relative stereochemistry of 3-hydroxy-2-methylnonanoic acid and absolute stereochemistry of N-Me-leucine moieties still remained ambiguous. The structure of 1 which was unequivocally determined by X-ray crystallography, revealed that the absolute stereochemistry of ambiguous moieties are (2R, 3R)-3-hydroxy-2-methylnonanoic acid and N-Me-L-leucine. The crystal structure of 1 suggested the carbonyl oxygen of the L-allo-Ile forms an intramolecular hydrogen bond to the NH of glycine. We also achieved an asymmetric total synthesis of 1 by convergent coupling of three fragments followed by macrocyclization. The physical and spectroscopic data of synthetic 1 were identical to those of natural globomycin.
We have been studying on a bioluminescence system of an oceanic luminous squid. Symplectoteuthis oualaniensis L. In 1981, Tsuji and Leisman reported that a homogenate of the luminous organ of this squid gave light in the presence of monocations such as Na^+, K^+ and molecular oxygen at pH 7.8. In 1993, we succeeded in the extraction of a photoprotein. named symplectin, responsible for the bioluminescence in a 0.6M KCl solution from the same squid S. oualaniensis collected in Okinawa, Japan. We reported that this photoprotein includes dehydrocoelenterazine(2) as the chromophore which covalently binds with symplectin through a thioether linkage. The major evidence for this chromophore is that mixing of dehydrocoelenterazine(2) into apo-symplectin significantly enhanced the luminescence intensity. The reddish solution of 2 mixed to apo-symplectin solution instantaneously changed into yellowish color. Recently this phenomenon was successfully mimicked between tri-peptide GSH(glutathione) in place of symplectin and a 100%-^<13>C-labeled methoxy-dehydrocoelenterazine(17). The complex structure (18, 19) were determined by NMR (C-H COSY, HMBC and HOHAHA) and Q-TOF-MS/MS measurement. We found that fluorinated dehydrocoelenterazine(21) gave the most stable chromophore complex(22) with cysteine of symplectin analog; GSH(20). The photoprotein (symplectin) can only become soluble in high salt buffer containing KCl higher than 0.6M. The homogenate of the photogenic organ containing symplectin was first washed with 0.4M KCl solution (pH 6.0), and then extracted with 0.6M KCl buffer to give a crude symplectin solution. Addition of aliquot of the photoprotein solution into pH 8.0 buffer gives light, so it is possible to check the luminescence activity of symplectin. A selective hydrolysis occured with trypsin to convert the original 60kDa (symplectin) into 2 products, namely 40kDa and 15kDa. Both 60kDa and 40kDa proteins showed luminescence activity. The 15kDa was not luminescent, but it possessed the same N-terminal amino acid sequence as the 60kDa symplectin. This means the active center of symplectin exists in the 40kDa site. We have attempted to obtain the cDNA encoding the squid 60kDa symplectin by the cloning procedure based on the polymerase chain reaction (PCR). Recently, we have succeeded in cloning the cDNA and obtaining the amino acid sequence of the 60kDa symplectin. At the same time, some part of the sequence of symplectin has been deduced by means of enzymatic hydrolysis followed by nano-LC-Q-TOF-MS/MS analysis. The complete amino acid sequence of symplectin was obtained by using MS/MS analysis and cDNA. Symplectin has 11 cysteine residues in the total sequence and 8 cysteine residues in the active 40kDa site. One of the cysteine in the 40kDa site must be the active center of bioluminescence activity. The detection of the active site of the symplectin and the investigate molecular mechanisms on the interaction of the chromophore with the apo-symplectin are now in progress in our laboratory.
trans-Fused polycyclic ethers are the common frameworks of various marine polyether compounds, such as brevetoxins, ciguatoxins, and maitotoxin that have attracted many chemists from their characteristic structures, biosyntheses, and biological activities. The total syntheses of hemibrevetoxin B, brevetoxin B, and brevetoxin A had already been achieved. Nakanishi and Shimizu proposed the biogeneses in which these polyethers might be produced from the requisite polyepoxy precursors (Scheme 1). Their hypothetical routes seem to be plausible, though there has been not any chemical evidence that suggests the reaction mechanism. Many of the chemical syntheses of assembled polycyclic ethers from polyepoxy compounds has been performed through successive exo-cyclizations, while few successful endo-selective cyclizations in the polyepoxide system have been reported. We describe the first chemical realization of the biomimetic construction of the six-membered trans-fused tricyclic ether. In order to control the cyclization in an endo-mode, we have planned to activate an epoxide regioselectively by the coordination of Lewis acid between the oxygen atoms of a methoxymethyl group and an epoxide (Scheme 2). After many attempts, it was found that the use of La(OTf)_3 afforded 6- and 7-endo cyclization predominantly in the presence of La_2O_3 and H_2O. 8-endo-Selective cyclization was also achieved in the absence of H_2O (Scheme 3). We have extended the above results to the biomimetic reaction. anti-Diepoxy (11) and anti, anti-triepoxy alcohols (13) were prepared by Shi's asymmetric epoxidation of the respective (Z,Z)-diene 6 and (Z,Z,Z)-triene 12, which have been constructed by the procedure involving the syn-seiective aidol coupling followed by the syn-dehydration. On treatment with a 4:1 mixture of La(OTf)_3 and La_2O_3, in the presence of H_2O, racemic 9 corresponding to 11 provided a trans-fused bicyclic ether 14 in 52% yield, along with a trans-fused perhydropyranofuran 15 as well as monocyclic compounds 16 and 17 (Scheme 7). Under the same conditions, 13 afforded a trans-fused tricyclic ether 19 in 9.3% yield (Scheme 9). Thus we have succeeded in the first biomimetic construction of trans-fused bi- and tricyclic ethers from the di- and triepoxy alcohols in one pots, respectively.
In 1996 stellettadine A (1) was isolated from a marine sponge Stelletta sp. as a metamorphosis-inducing compound for ascidian larvae. Its absolute configuration was proposed as S. The first synthesis of (S)-1 was achieved by starting from (S)-citronellal (4). Although the naturally occurring 1 has been reported to be levorotatory, our synthetic (S)-isomer is dextrorotatory. It suggests that the absolute configuration of natural stellettadine A is not S but R. Naurol A is a cytotoxic metabolite isolated from Pacific sponges, and 2 was proposed as its structure. A mixture of (±)-2 and meso-2' was synthesized from 4-tert-butyldimethylsilyloxy-3-methyl-2-cyclohexen-1-one (10) employing the Stile coupling (12+13→2+2') as the key step. Although our synthetic sample (2+2') was a diastereomeric mixture at C-11, its spectral data (IR, UV, ^1H and ^<13>C NMR and MS etc.) were significantly different from those reported for natural naurol A. It was therefore concluded that the structure 2 proposed for naurol A was in error. In 1997 testudinariol A (3) was isolated from the marine mollusc Pleurobrancus testudinarius. This compound is thought to be defensive allomone of P. testudinarius. We studied the stereoselective synthesis of testudinariol A (3), and we synthesized the key intermediates 28 and 29 by employing intramolecular oxy-Michael addition and ene reaction as the key steps. The conversion of 28 and/or 29 to testudinariol A (3) is now under investigation.
Lipopolysaccharide (LPS) is a glycoconjugate located at the so-called outer membrane of Gram-negative bacterial cell surface and exhibits multiple and potent biological activities including both toxic and beneficial ones. LPS is generally composed of a polysaccharide portion and an acylated glucosamine disaccharide termed lipid A which is responsible for the biological activities of LPS. Re lipopolysaccharide (Re LPS) from E. coli Re mutant is one of the structurally most simple LPS found on living bacterial cell surface, being composed of only two moles of 3-deoxy-D-manno-2-octurosonic acid (Kdo) and lipid A. For precise investigation of effects of the additional Kdo residue(s) on the bioactivities and conformation, Re LPS has been successfully synthesized for the first time. The synthesis started from D-mannose and D-glucosamine which were converted into appropriately protected Kdo and lipid A components, respectively. The desired tetrasaccharide backbone consisting of Kdo-disaccharide and lipid A was constructed by stepwise condensations of these saccharide building blocks under elaborated efficient conditions and functional group manipulations in 0.9% total yield for 23 steps followed by efficient purification by liquid-liquid partition column chromatography. The synthetic preparation was identical in all respects with a natural counterpart. By employing the same synthetic strategy, several partial structures of Re LPS were also synthesized for precise functional studies. Biological activities as well as conformational studies by NMR will also be discussed in comparison with lipid A.
Oligosaccharides play a lot of important roles in biological processes and in pathology. While the automated synthesis of peptides and oligonucleotides was established decades ago, there have not been a general method for the synthesis of complex oligosaccharides. Recently the development of a broadly applicable synthetic strategy has been desired in this field. Here we wish to report the two efficient and practical strategies based on a one-pot sequential glycosidation and selective deprotection followed by glycosylation toward the automated synthesis of a oligosaccharide library. In the synthesis of heptasaccharide 3 having elicitor activity in soybean, we achieved a onepot 6-step glycosidation using a combination of different leaving groups, i.e., bromide (-Br), ethylthio group (-SEt), phenylthio group (-SPh), and fluoride (-F), with selective activators, AgOTf, MeOTf, DMTST, and HfCp_2(OTf)_2, respectively and differences of the reactivity of hydroxy groups of glycosyl acceptors. This method was applied to synthesize a library of tetra-, penta-, heptasaccharides using a manual synthesizer. For a study of the structure-activity relationship of elicitor active oligosaccharide in rice cell we demonstrated the synthesis of the oligosaccharide library having β(1→6) linkage at the various positions on β(1→3) tetrasaccharide backborn. Selective deprotection of three protecting groups (AOC, Lev, and Fmoc) of tetrasaccharide 4 was sequentially achieved to afford respective three mono-ols, three diols, and one triol, independently. Glycosylation of all glycosyl acceptors provided the corresponding oligosaccharides, respectively. These procedures were achieved on an automated synthesizer and by parallel column chromatography.
Total synthesis of madindoline A was achieved. Stereoselective construction of quaternary carbon was realized by alkylation of enolate derived from α,β-unsaturated imide possessing chiral auxiliary. The γ-proton of the α,β-unsaturated imide 9 was abstracted with NaHMDS to give enolate 12 stereoselectively which facilitated alkylation at α-position from bottom side to afford β,γ-unsaturated imide 10 in high stereoselectivity (Table 1, Scheme 2). This method was applied to the cyclopentenedione moiety of madindoline A (Scheme 3). The α,β-unsaturated imide 14 was alkylated to obtain imide 15 in 61% yield with 8.4:1 ds. Imide 15 was converted into ethylketone 17 which was oxidized to give triketone 19. Treatment of triketone with DBU in benzene at room temperature facilitated regioselective cyclization to afford cyclopentenedione 20 in high yield. This procedure was applied to the total synthesis of madindoline A. In the reductive amination step for the coupling of sterically hindered aldehyde 16 and acid-sensitive amine 23, we found that Sn(OTf)_2-NaBH(OAc)_3 method was effective. After the reductive coupling, side chain was transformed to triketone 28 which was subjected to regioselective intramolecular condensation to construct a madindoline skeleton. Removal of TBS group from 29 with TBAF completed the total synthesis of madindoline A.
The study on the reactivity of 1-azatrienes toward 6π-azaelectrocyclization is limited and the application of this reaction to the natural product synthesis is also very limited in the literature. This is presumably because the azaelectrocyclization process has been believed to require high temperature, and in addition, the method for preparation of 3-cis-1-azatrienes as precursors of the electrocyclization was not well established. During the course of our recent studies on the syntheses of the enzyme inhibitors and then on the elucidation of their inhibitory mechanism, we found that the (E)-3-carbonyl-2,4,6-trienal compounds inhibited the hydrolytic ability of phospholipase A_2 (PLA_2) by the formation of the dihydropyridine derivatives resulting from the reaction with the particular lysine residues of PLA_2 via 6π-electrocyclization of the intermediary Schiff base. Moreover, we found that the presence of both the C4 ester group and the double bond attached at C6 in 1-azatrienes significantly contributed to the acceleration of the azaelectrocyclization. In this conference, we disclose in detailed (1) the new and efficient method for the synthesis of the 3-cis-1-azatriene derivatives and the results of their 6π-azaelectrocyclization, and then we evidently concluded that the combination of the C4-carbonyl and the C6-olefin groups in 1-azatrienes actually contributed to the remarkable acceleration of the azaelectrocyclization based on the orbital interaction between HOMO of the olefin part and LUMO of the azadiene part under the reverse electron demanded mode. (2) This conclusion was supported by the molecular orbital calculation of the azatriene. Moreover, (3) the results obtained here were applied to the synthesis of the ocular age pigment A2-E, by focusing on two types of the new and efficient one-pot synthesis of substituted pyridines, which involve the sequence of (a) smooth 6π-azaelectrocyclization and dehydration, and (b) azaelectrocyclization and oxidation. The later sequence of the reactions is compatible with Nakanishi's hypothetical metabolic pathway of A2-E.
Coelenterazine is well known as a chromogenic compound of aequorin and as luciferin of various bioluminescent marine organisms including the sea pansy Renilla reniformis and the deepsea shrimp Oplophorus gracilorostris. The luminescent reaction was basically accepted as an oxidation reaction in which several key peroxy intermediates and/or transition-states seem to be involved. We have developed a new method to synthesize imidazopyrazinones by using Pd-mediated cross couplings. To investigate the process of bioluminescence reaction and the chiral environment of its transiton-state, we designed a number of new stable coelenterazine analogs and transition-state analogs with an indanone core structure. These analogs were synthesized in 11 steps. Having examined the inhibitory activities for bioluminescence reaction, we found proteinous components in bioluminescence showed different specificity for the chiral transition-state analogs. It suggests that the transition-state analogs might be useful to investigate the chiral environment of bioluminescent reactions.
Strychnine (1) is a representative member of the Strychnos family of indole alkaloids and has been a target of several synthetic investigations. In recent years, three research groups have culminated their research in the asymmetric total synthesis of (-)-1. To date, however, no catalytic asymmetric synthesis has been accomplished except for enzymatic method. We describe herein the development toward a novel stable, storable and reusable asymmetric catalyst (R,R)-La-linked-BINOL 10 and our synthetic studies of (-)-strychnine (1) employing catalytic asymmetric Michael reaction catalyzed by multifunctional asymmetric catalysts as the key step. We previously developed a highly practical catalytic asymmetric Michael reaction using AlLibis(binaphtoxide) (ALB 5). At high concentration (31M) the reaction was highly accelerated and the use of only 0.3mol% of (R)-ALB 5 promoted the Michael reaction of 6 with 7 to completion within 10h, giving Michael adduct 8 in 95% yield and 99% ee. Furthermore, we developed novel highly stable, storable and reusable asymmetric catalyst (R,R)-La-linked-BINOL 10, which also catalyzed this reaction efficiently to give Michael adduct 8 in 94% yield and >99% ee. The catalyst 10 was easily prepared from La(O-i-Pr)_3 and 1.0 equivalent of linked-BINOL 9, which were mixed in THF followed by removal of solvent under reduced pressure to afford 10 as a pale-yellow powder. This air-stable catalyst 10 is storable under air at ambient temperature. After 4 weeks storage, no change in catalytic activity, in terms of both chemical yield and ee, was observed using stock catalyst 10. In addition, the catalyst 10 can be reused several times. After the examination of the scope and limitations of different substrates, we found that the catalyst 10 promoted the Michael reaction of a variety of cyclic enones (n=0-4) with various malonates to afford Michael adducts with good to excellent ee's. The complex 10 was also effective for the Michael reaction of acyclic enones. To the best of our knowledge no efficient catalytic Michael reaction of 8- and 9-membered ring enones with malonates has been reported to date and this is the first example of a Michael reaction where the catalyst shows such broad generality. Having obtained nearly optically pure 8 in large quantities, we next planned to synthesize (-)-strychnine (1) using the Michael adduct 8 as a chiral building block. From the many disconnections, we chose the strategy based on the Stille coupling reaction, intramolecular 1,4-addition reaction of amine to enone, and intramolecular alkylation. This approach requires the precursor α-iodoenone 38, which could be prepared by regioselective enone forming reaction followed by iodination of the corresponding enone. Because the (E)-side chain of 40 would evolve to F and G rings of (-)-strychnine (1), stereoselective introduction of this unit would solve the stereochemical problem posed by the allylic ether double bond at C(20). The synthesis of the (E)-α,β-unsaturated ester 40 was achieved by using a modified Overmann's procedure. In this reaction, we found that the E/Z-selectivity and yield of this reaction are highly dependent on the reaction temperature. At ?55℃ the best result, 70% (2 steps) and E:Z=>15:1, was obtained. In the oxidation step of the corresponding allylic alcohol using MnO_2 the kinetic resolution was observed. (E)-allylic alcohol was oxidized much faster than (Z)-allylic alcohol to afford almost pure (E)-α,β-unsaturated aldehyde 45. After the introduction of C-ring moiety by reductive amination, we next examined the regioselective enone forming reaction. As we expected, more bulky base gave higher selectivity (48:47=up to 6:1). Using these inseparable mixtures, DMAP promoted the iodination of 48 to afford the α-iodoenone 38 with unreacted 47. Using the α-iodoenone 38, we next examined the Stille coupling reaction. At first we could get only
Lepadin A (1) isolated from the tunicate Clavelina lepadiformis represents the first example of a decahydroquinoline alkaloid from a marine natural source. Subsequently, the closely related compounds, lepadins B (2) and C (3), along with lepadin A have been found in the flatworm Prostheceraeus villatus and its tunicate prey C. lepadiformis. Both lepadins A (1) and B (2) have been shown to exhibit significant in vitro cytotoxicity against human cancer cell lines. We report here the first synthesis of (-)-lepadins A and C, as well as a new synthesis of (-)-lepadin B. The key feature involves the use of a stereoselective intramolecular hetero-Diels-Alder reaction of an N-acylnitroso compound and a Suzuki-type cross-coupling reaction for elaboration of the octadienyl side chain. Upon treatment of the chiral hydroxamic acid 15, prepared from enantiopure 2,4-O-benzylidene-2,4-dihydroxybutanal (8) as a C_3 chiral synthon, with Pr_4NIO_4 under the aqueous conditions, the in situ generated acylnitroso compound 16 was subjected to intramolecular Diels-Alder reaction to yield the trans-oxazinolactam 17 in a stereoselective manner. Subsequent hydroxylation at C-7 by Davis methodology and introduction of the methyl group were achieved by means of a tandem Grignard reaction-NaBH_3CN reduction in a highly stereocontrolled manner. The oxazine 20 thus obtained was converted to the common key intemediate 35 via transformations involving intramolecular aldol condensation of the keto aldehyde 26, epimerization at the C-5 stereocenter, hydroxyl-directed hydrogenation of the octahydroquinoline 32, and Takai olefination of the aldehyde 34. Subsequent elaboration of the octandienyl side chain was achieved by palladium-catalyzed Suzuki-type cross-coupling reaction between 35 and the alkenyldihydroxyborane to afford the coupling products 36 or 39, which was converted into (-)-lepadin A (1) and (-)-lepadin B (2), or (-)-lepadin C (3), respectively.
Fudecalone was isolated by Omura et al. in 1995 from a culture broth of Penicillium sp. FO-2030 as an anticoccidial sesquiterpene. The structure was elucidated mainly by NMR experiments and the conformation was reported to be A. However, our synthetic Fudecalone, which has the same configuration as reported, showed the other conformation B and the NMR spectral data were not identical. According to MM3 calculation, B was more stable than A. So, we thought interesting that Fudecalone existed as an unstable conformation because of the extra-high energy barrier between A and B, and investigated conformational isomerization from B to A. The selenide 11, whose conformation was determined to be nonsteroid type (A) by NOE experiments, was prepared in seven steps from the intermediate 8 with B type conformation. But after oxidative elimination, the conformation returned to A type, so we doubted the proposed relative stereochemistry and attempted to synthesize transoctalone (C). We adopted the intermediate 8 again, isomerized its octalone skeleton from cis to trans and converted the trans-isomer (14) into 16, whose NMR spectral data were identical with those of the natural Fudecalone. In addition to that, NOE experiment revealed the hydroxy group of hemiacetal to be β-oriented. In conclusion, we synthesized Fudecalone as a racemate and determined its relative configuration as 1S^*, 3aS^*, 6aS^* and 10aS^*. The synthesis of optically active compound is now in progress.
Recently, cytotoxic triterpene polyethers, which are thought to be biogenetically squalene-derived natural products, have been isolated from both marine and terrestrial plants. Among them, our synthetic target in this symposium is glabrescol (1) isolated from the branches and wood of Spathelia glabrescens (Scheme 1). Our synthetic strategy for the proposed structure of glabrescol (1) is based on taking its intrinsic symmetry into consideration, and on the sequential hydroxy-directed anti oxidative cyclizations of acyclic bishomoallylic alcohols with vanadium catalyst and TBHP to stereoselectively construct such THE rings via epoxides (Scheme 2). Optimized conditions for the double cyclization of 9 (0.02 equiv. VO(acac)_2, 2.5 equiv. TBHP, 2 equiv. TFA, CH_2Cl_2, rt, 30min) provided the desired triTHF ether 14 as a major product (Scheme 3). The treatment of 14 under similar conditions gave the original meso structure 1 as the predominant product in 30% yield. Unfortunately, the NMR spectra of our synthetic 1 were not identical with those of the natural glabrescol kindly provided by Jacobs. The three remaining possible meso structures 21,24,25 were also synthesized; however, their NMR spectra were again inconsistent with those of the natural product (Scheme 4). Although Jacobs et al. proposed a meso structure for glabrescol based on the optical inactivity and the presence of fifteen signals in the ^<13>C NMR spectrum, the above results cannot support any meso structures for glabrescol. The other possibilities fulfilling the criteria are that glabrescol is C_2 symmetric and racemic or that glabrescol is C_2 symmetric and the value of the specific rotation is near zero. Thus, we embarked on the enantioselective total synthesis of the C_2 symmetric structure 4 possessing the same relative stereochemistry as that of longilene peroxide (3) (Scheme 5). Fortunately, the spectral characteristics including the CD spectrum of the synthetic 4 were identical to those of the natural glabrescol. Thus, the correct stereostructure of glabrescol must be revised from the Cs symmetric 1 to the C_2 symmetric 4 with the indicated absolute configuration.
In 1997 CP-263,114 (Phomoidride B, 1) was isolated from the culture broth of an unidentified fungus by a Pfizer group and shown to inhibit squalene synthase as well as ras farnesyl transferase. In addition, CP-263,114 has a unique, densely functionalized polycyclic skeleton that consists of a bridgehead double bond, a γ-lactone-acetal, and a maleic anhydride moiety. These structural features and interesting bioactivities have attracted much interest from synthetic chemists. Herein we report a stereoselective total synthesis of 1. As illustrated in our retrosynthetic analysis of CP-263,114, we planned to construct the basic skeleton A (bicyclo[4.3.1]decene) by means of intramolecular Diels-Alder reaction. The precursor of Diels-Alder reaction could be prepared from three units; an aldehyde unit B, an acryloyl derivative, and a 1,3-diene unit C. The 1,3-diene unit 7 was synthesized by 1,4-addition reaction of alkenyl copper reagent 5 to the reactive allenic ester 3, followed by an introduction of one more methoxycarbonyl group. On the other hand, the aldehyde unit 13 was prepared from (S)-epichlorohydrin in a 11-step sequence. After the assembly of 7 and 13, boron-mediated diastereoselective aldol reaction, and subsequent Parikh-Doering oxidation afforded the enone 15. An intramolecular Diels-Alder reaction was performed by treatment of 15 with ZnCl_2・OEt_2 in the presence of a small amount of pyridine. After the construction of the bicyclic system, a novel protocol for the formation of the thiomaleic anhydride 20 via thiobutenolide 19 was employed. Finally homologation of the ester and subsequent acid-promoted formation of γ-lactone-acetal 23 were achieved to complete the total sythesis of (-)-CP-263,114 (1). The synthetic 1 was identical in all respects with the natural CP-263,114 [^1H NMR, ^<13>C NMR, HRMS and [α]_D ?10°(c = 0.25, CH_2Cl_2) (lit. [α]_D ?11°(c = 0.5, CH_2Cl_2))]. Accordingly, the absolute configuration of CP-263,114 should be the one depicted in the last scheme.
Marine cyanobacteria are a rich source of a variety of structurally-unique and biologically-active natural products. Isolation and structure elucidation of kalkitoxin, a potent neurotoxin from the marine cyanobacterium Lyngbya majuscula, is presented. Isolation of this compound was aided by bioassay-guided fractionation using the brine shrimp and gold fish toxicity assays. The planar structure of this thiazoline-containing lipid was elucidated by standard 1D and 2D NMR data. The stereochemistry of C-3 was determined to be R by Marfey's method, while the relative stereochemistry within the aliphatic chain of kalkitoxin was suggested to be 7R^*, 8S^*, 10S^* by J-based configuration analysis using a new NMR pulse sequence, HSQMBC, and a cryoproba NMR technology. Five synthetic isomers of kalkitoxin were compared to the natural compound by ^1H and ^<13>C NMR and CD spectroscopy. ^<13>C NMR chemical shifts showed very small differences in the range of less than 0.2ppm to natural kalkitoxin. However, the 3R,7R,8S,10S,2'R isomer showed maximal ^<13>C NMR differences of 0.026ppm with an average difference of only 0.008ppm. In addition, its CD spectrum was identical with natural kalkitoxin. Natural kalkitoxin is strongly ichthyotoxic (LC_<50> 700nM) and potently brine shrimp toxic (LC_<50> 170nM). Synthesized kalkitoxin was similarly potent in the brine shrimp assay (LC_<50> 170nM). Interestingly the synthesized enantiomer of kalkitoxin was relatively inactive (LC_<50> 9300nM). Kalkitoxin also showed potent NMDA receptor-mediated neurotoxicity (LC_<50> 3.86±1.91nM).
Most people are familiar with the sight of a young seedling bending towards a window or the brightest source of light to which it is exposed. This directional growth response is known as phototropism, which is caused by a lateral gradient of growth-promoting auxin in the bending organ (Cholodny-Went theory). In contrast to the C.-W. theory, Bruinsma and Hasegawa recently found that the shaded half did not contain more auxin than the illuminated ones. Instead it was found that the even distribution of auxin was accompanied by a lateral gradient of growth inhibitory substances during phototropic curvature. We have isolated some photo-induced growth inhibitory substances related to phototropism, raphanusanins from radish (Raphanus sativus var. hortensis f. gigantissimus Makino) hypocotyls, benzoxazolinones from light-grown maize (Zea mays L.) shoots, indolylacetonitrile from light-grown shoots (Brassica oleacea L.), and 8-epixanthatin from sunflower hypocotyl, and uridine from oat coleoptile. And chemical analyses have shown phototropic stimulations to cause curvature by inducing a local unequal distribution of growth-inhibiting substances that antagonize auxin in its cell-elongating activity. Finally, the mechanism of phototropism in plants is described in Scheme 4.
During the course of screening for modulators of signal transduction of mammalian cells, we have discovered two novel inhibitors, indocarbazostatin (1) and indocarbazostatin B (2) from a culture broth of a Streptomyces sp. (Fig. 1). In our assay system, 1 inhibited NGF-induced neurite outgrowth from PC12 cells at 6nM, which is approximately 33 times higher than that of K252a. We would like to present structures, biological activities and some biosynthetic aspect of these natural products during this symposium. Among 4671 extracts, only one culture produced active compounds of interest. The producing strain identified as Streptomyces sp. TA-0403 was cultivated at 30℃ for 3 days in 500mL cylindrical flasks each containing 70mL of the medium. The fermentation broth (35L) was centrifuged and the mycelial cake was extracted with acetone. After the removal of acetone, pure compounds 1 and 2 were isolated as shown in Scheme 1. The HR FAB-MS, UV, ^1H and ^<13>C NMR data indicate the chromophore system of these compounds to be hetero-atom substituted indolo[2,3-a]-pyrrolo[3,4-c]carbazole-5,7(6H)-dione. The structure of 1 was deduced from ^1H-^1H COSY, PFG HMBC, PFG HMQC, and DIF NOE data. The NMR assignments for 1 are summarized in Table 1. The relative and absolute configurations of the sugar moiety of 1 were clarified by MM2 and MOPAC calculations (Fig. 2) and CD analyses. Also, the structure of 2 was obtained through a similar approach. Thus, the structures of 1 and 2 were determined as shown in Fig. 1. Indocarbazostatin (1) and indocarbazostatin B (2) are new members of indolocarbazole-type bioactive molecules. The compounds 1 and 2 inhibited NGF-induced neurite outgrowth from PC12 cells at 6nM and 24nM respectively, whereas K252a inhibited at 200nM under our assay conditions (Fig. 3). K252a showed cytotoxicity at a concentration three times higher than the minimal effective concentration. However, 1 did not show obvious cytotoxicity to PC12 cells at a concentration nine times higher than the minimal effective concentration. The IC_<50> values for PKC were: 1 2.0nM, 2 8.5nM and K252a 35nM. The enhancement of productivity of 1 and 2 by UV radiation rearing of the producing strain, and their biosynthetic studies will be discussed in the symposium also.
The biosynthesis of clinically important aminocyclitol antibiotics, particularly containing 2-deoxystreptamine, has been studied. The first-step enzyme, 2-deoxy-scyllo-inosose synthase (DOIS), was purified to homogeneity from butirosin-producing Bacillus circulans for the first time as heterodimer consisting of 23kDa and 41kDa subunits. Partial sequences of the polypeptides were deduced and reverse-genetics approach consisting of oligonucleotide probe and hybridization allowed identifying the responsible gene btrC for the larger subunit. The deduced amino acid sequence appeared to have significant homology with microbial dehydroquinate synthase (DHQS). The btrC gene was heterologously over-expressed in E. coli and the expressed BtrC was recovered by rather simple purification procedures. Mechanistic studies showed that the stereochemistry of Bacillus DOIS was identical with that of Streptomyces DOIS, but different from DHQS. The possible DOIS-substrate interaction has been postulated on the basis of the studies with substrate analogs and molecular modeling. The butirosin biosynthetic gene cluster was analyzed by gene-walking starting from btrC, and more than a dozen ORFs were identified. Gene disruption of the ORFs, btrB, btrC, btrD, and btrM, were carried out and clearly showed that every disruptants lost the ability of antibiotic production. Therefore, the above-mentioned gene cluster, which has not yet been fully deduced, appeared to be responsible for the biosynthesis of butirosin. The function of each ORF has further been discussed by homology search.
All fungal polyketide synthase (PKS) genes so far cloned code for iterative type I PKSs different from bacterial modular type I PKSs and type II PKSs. In order to clarify how fungal PKSs control their reactions to produce specific product compounds, we have expressed several fungal PKSs in heterologous hosts and identified their product compounds. In this study, we carried out functional analysis of two fungal PKSs involved in DHN-melanin biosynthesis. The PKS1 gene cloned from the phytopathogenic fungus Colletotrichum lagenarium was assumed to code for a 1,3,6,8-tetrahydroxynaphthalene (T4HN, 1) synthase. To identify its actual function, the PKS1 gene was expressed in a heterologous fungus Aspergillus oryzae. Production of T4HN (1) by the transformant confirmed the function of PKS1 PKS as T4HN synthase. Cell-free extract prepared from the transformant showed PKS enzyme activity to form T4HN (1). Furthermore, cell-free studies identified that malonyl-CoA serves as not only extender units but also as a starter unit of enzymatic T4HN (1) formation. Aspergillus fumigatus is an important opportunistic pathogen causing aspergillosis and its conidial melanization is a key factor in its virulence. Presence of a PKS gene (alb1) was identified in its melanin biosynthetic gene cluster. T4HN (1) was expected to be a product compound of Alb 1p PKS, but naphthopyrone YWA1 (4) was produced by the heterologous expression of alb1. Homologous complementation and heterologous co-expression studies revealed that Ayg1 protein coded in the gene cluster is involved in T4HN (1) formation with Alb1p PKS. Further cell-free studies identified that Ayg1 protein can convert YWA1 (4) to T4HN (1) possibly by hydrolytic C-C bond cleavage.
Multiple alignments of primary structures of many kinds of prenyltransferases which participate in the most fundamental prenyl-chain backbone synthesizing process in isoprenoid biosynthesis have revealed the presence ofseven conservedregions in the primary structures of (E)-prenyl diphosphate synthases. However, no information has been available about the structures of (Z)-prenyl diphosphate synthases until our recent isolation of the gene for the undecaprenyl diphosphate synthase of Micrococcus luteus B-P 26 by use of the method "colony autoradiography" The amino acid sequence of the (Z)-prenyl diphosphate synthase istotally different from those of (E)-prenyl chain elongating enzymes. Protein data base searches for the sequences similar to that of the undecaprenyl diphosphate synthase yielded many unknown proteins which have not yet been characterized. Two of the proteins have recently been identified as the undecaprenyl diphosphate synthase of Escherichia coli and the dehydrodolichyl diphosphate synthase of Saccharomyces cerevisiae, indicating the presence of five highly conserved regions typical of (Z)-prenyl chain elongating enzymes. The crystal structure of M. luteus B-P 26 UPP synthase was determined at 2.2 A resolution as the first three dimensional structure among cis-prenyl chain elongating enzymes (Fig. 5). This enzyme shows a novel protein fold, which iscompletely different from the structures for the enzymes relating to isoprenoid biosynthesis, which is believed to have a common structure, the so-called "isoprenoid synthase fold". The UPP synthase has a large cleft composed of highly conserved residues among cis-prenyl chain elongating enzymes. A P-loop motif, which is composed of several conserved residues was found in the UPP synthase, suggesting a binding site for the diphosphate moiety of the allylic substrate FPP.
Based upon the idea of the stepwise mechanism of biomimetic olefin cyclimtion and enzymatic cyclization of oxidosqualene via conformationally flexible cationic intermediates, significant attention has been focused on each step of sterol biosynthesis. Today, the biosyntheses of phytosterols such as dammaranoid, lupanoid, oleananoid, and tirucallanoid are explained by the cyclization of oxidosqualene via a bicyclic cation 1, tricyclic 6/6/5-cation (pre-C ring cation) 2, secondary 6/6/6-cation 3, and 6/6/6/5 cation 4. In the animal kingdom, steroids are also constructed through the corresponding boat-form B-ring intermediates. Transformation of 1 to 2 is selective cyclization to 6/5 trans fused system even though 6/5 cis is generally more stable. Transformation of 2 to 3 involves ring expansion of a tertiary cationic substrate into a secondary cation, namely, an anti-Markovnikov cation. We have investigated the possibilities of the transformations of 1 to 2 and 2 to 3 as a chemical reaction by preparing model cations. When bicyclic tert-alcohol 8 was treated with BF_3 ・Et_2O, 6/6/5-tricyclic compound 10 and 11 were formed and the structures were confirmed by X-ray analysis. Thus the trans selective 6/5 cyclization took place via cation 12. Now we will be able to state that 6/5 trans selective cyclization from 1 to 2 does not depend upon any special enzyme effect. This may be due to the steric repulsion between the angular methyl and C-12 and C-13 methylenes for 6/5 cis cyclization as illustrated in 13. Next problem to be solved is the Markovnikov' wall. We chose a diol 18 as the model compound and investigated the chemical behavior of the generated cation 19. Most of the Lewis acids such as BF_3・Et_2O, SnCl_4, Sc(OTf)_3, FeCl_3, or TiF_4, and CF_3SO_3H leads to a hydride shift to cation 20 leading spiro cyclic ether 21. This hydride shift corresponds to the initiation of the backbone rearrangement on the D-ring of steroid On the other hand, TiCl_4, selectively induces rearrangement to secondary cation 22 by the ring expansion that corresponds to the C-ring formation of sterol biosynthesis, and afforded cis-ether 23, along with chlorides 24 and 25. AlCl_3 and ZrCl_4 induce further rearrangement into six-membered ring tert-cation 29 and afforded only isomeric cis-ether 26. Thus, the transformation of 2 to 3 will also be able to achieve without enzyme.
Reveromycins A-D (1-4) are novel polyketide-type antibiotics isolated from the genus Streptomyces as inhibitors of mitogenic activity induced by the epidermal growth factor (EGF) in a mouse epidermal keratinocyte. The characteristic structural features of reveromycins include a 6,6- or a 5,6-spiroketal system bearing a hemisuccinate, two unsaturated side chains and two alkyl groups. Their strong biological activity as potential drugs and their synthetically challenging, unique structure have attracted the attention of synthetic organic chemists and the total synthesis of reveromycin B (2) has been independently accomplished by three groups. We now report the first asymmetric total synthesis of reveromycin A (1), which is the major and most bioactive compound of the reveromycins. Our retrosynthetic analysis of reveromycin A (1) is as follows. The unsaturated left and right side chains should be produced by the Horner-Wadsworth-Emmons reaction, Julia coupling, and Wittig reactions during the later stage due to their instabilities. The construction of the hemisuccinate of the C18 tert-hydroxyl group might be achieved by acylation under high pressure. The main problem should be the construction of the 6,6-spiroketal core 5 in which the C18 tert-hydroxyl group and C19 side chain are axially oriented. The inherent instability of the 6,6-spiroketal system in 5 may cause some difficulties during the synthetic studies, e.g., easy transketalization of 5 (X=H) into the stable 5,6-spiroketal 7, transformation of 5 into the undesired 6,6-spiroketals 6 in the unnatural form via an equilibration. The construction of 5 could be achieved by the intramolecular ketalization of the ketone 8, which would be synthesized via the coupling reaction of the Weinreb amide 9 and alkyne 10. The coupling reaction of 9 and the lithio derivative of 10 followed by hydrogenation furnished the saturated ketone 11, which was treated with CSA to give the 6,6-spiroketals 12 in the natural form and 13 in 54% and 27% yields, respectively. Deprotection of the MTM group in 12 with MeI-NaHCO_3 proceeded to afford the alcohol 16. Succinylation of 16 with mono-allyl succinate and DCC at 1.5GPa efficiently proceeded to give the succinate 17. Desilylation of 17 followed by the Dess-Martin oxidation gave the aldehyde, which was subjected to the Horner-Wadsworth-Emmons reaction with (EtO)_2P(O)CH_2C(Me)=CHCO_2Allyl to give the (E,E)-dienoic esters 18. Deprotection of the MPM group in 18 with DDQ afforded the alcohol 19. The 6,6-spiroketal 13 in the unnatural form was converted into the desired 19 via the treatment of the alcohol 21 with CSA. The treatment of alcohol 23 derived from 19 with 2-mercaptobenzothiazole under the modified Mitsunobu conditions with TMAD and n-Bu_3P gave the sulfide which was subjected to the Mo(VI)-mediated oxidation to give the sulfone 24. The one-pot Julia olefination of 24 and the aldehyde 28 stereoselectively produced the (6E,8E)-diene 25. Deprotection of TES group in 25 followed by Dess-Martin oxidation provided aldehyde which was subjected to the Wittig reaction with a neutral phosphorane to give the protected reveromycin A (27). Finally, successive deallylation with Pd(Ph_3P)_4-Ph_3P and desilylation with TBAF in DMF gave reveromycin A (1), which was identical with the natural product.
A number of extremely potent enediyne antitumor antibiotics have been isolated from culture filtrates of Streptomyces species. Although these antibiotics are believed to generate p-benzyne-type biradicals through cycloaromatization, no direct EPR measurements of carbon radical species generated from enediyne antibiotics have been reported. A chromoprotein antibiotic, C-1027, which is powder of a 1:1 complex composed of a carrier apoprotein and a DNA cleaving enediyne chromophore (1), shows a steady EPR spectrum at room temperature. Here, we characterize three paramagnetic species existing in C-1027 antibiotic by EPR spectroscopy: a phenyl radical (3), a radical pair between a peptide carbon radical and 3, and a peroxy radical. These radical species are supplied constantly by dynamic processes through hydrogen abstraction by p-benzyne type biradical (2) from apoprotein. The present results provide strong evidence for equilibration between 1 and 2, and a chemical mechanism of deterioration leading to self-inactivation of the C-1027 antibiotic.
Previously, we have synthesized various natural products enantioselectively by using baker's yeast reduction of acyclic α-hydroxyketones as the chilarity induction method. In this paper, we describe the enantioselective synthesis of two triterpene ethers possessing symmetrical structure, e.g. glabrescol with C_s symmetry and testudinariol A with C_2 symmetry. 1. What is the structure of glabrescol? Glabrescol, isolated from Spathelia glabrescens, is a triterpene having five continuously linked tetrahydrofuran (THF) rings, to which novel meso-type structure 5 has been assigned. Starting from (+)-9 obtained in high enantiomeric purity by yeast reduction, two 15-carbon segments 7 and 8 were synthesized. Coupling of them, followed by diastereoselective oxygenation and THF ring formation resulted in the synthesis of a compound having the structure 5 and one of its diastereomers 21. However, their NMR spectra were not identical to those of natural glabrescol. Thus, the structure of glabrescol remains to be clarified. 2. Synthesis of testudinariol A Testudinariol A has been isolated from the marine mollusc Pleurobrancus testudinarius and C_2 symmetric structure 6 has been assigned. Because of its C_2 symmetry, 6 could be synthesized by the homocoupling of the segment such as 22. Starting from the compound 23 which has been prepared during our synthesis of hippospongic acid A (4) and employing hydroboration and ene reaction, the bromide 40 was prepared and treated with Ni(cod)_2. The coupling product was a mixture of (Z,Z)-isomer and (Z,E)-isomer (1:1). Although the separation of them was unsuccessful, NMR spectra revealed the presence of testudinariol A (6) in the mixture. Further study for other method to combine the segments without formation of undesired isomer is under proceeding.
The gibberellins, plant growth regulators, are divided into two groups, the larger of which is C_<19> gibberellins [gibberellic acid: GA_3 (1) is a typical representative] and most of the remaining have 20 carbons. The latter possess the ent-gibberellane carbon skeleton. GA_<12> (2), GA_<111> (3), and GA_<112> (4) belong to C_<20> gibberellins and 2 is presumed to be a common intermediate in the biosynthesis of all gibberellins. A significant array of biological activities and structural complexity have made gibberellins popular targets for total synthesis. Although GA_3 (1) is produced commercially by the large-scale fermentation of the fungus Gibberella fujikuroi, most of the C_<20> gibberellin syntheses have been achieved through many functional group manipulations, in spite of starting with tricyclic compounds or natural gibberellins. An efficient synthetic route to C_<20> gibberellins would make it possible to confirm tentative new structures and to explore their biological activities. Herein we would like to present novel synthetic routes to C_<20> by a radical cyclization, and by a combination of palladium-catalyzed cycloalkenylation reaction and reverse electron demand intramolecular Diels-Alder reaction.
We have studied a gene cluster responsible for the biosynthesis of the diterpene compound, terpentecin, produced by a Streptomyces strain. We have prepared a mutant lacking the 3-hydroxy-3methylglutaryl CoA (HMG-CoA) reductase activity from this strain, which uses both the mevalonate and nonmevalonate pathways for the formation of isopentenyl diphosphate, by screening terpentecin non-producing mutants. Terpentecin is known to be synthesized via the mevalonate pathway. The gene encoding HMG-CoA reductase (hmgr) was cloned and identified by complementation of the mutant, using a self-cloning system developed in this study for strain MF730-N6. By searching a flanking regions of the HMG-CoA reductase gene, we were able to,identify two gene clusters, which encode the mevalonate pathway enzymes and the terpentecin biosynthetic enzymes. In the former cluster, seven genes encoding geranylgeranyl diphosphate synthase (GGPPS), mevalonate kinase (MK), mevalonate diphosphate decarboxylase (MDPD), phosphomevalonate kinase (PMK), an unknown protein, HMG-CoA reductase, and HMG-CoA synthase were identified. In the latter cluster, additional seven genes, which were suggested to encode the terpentecin biosynthetic genes, were found. Two genes found in the terpentecin biosynthetic gene cluster, which have similarities to eucaryotic diterpene cyclase genes and a eubacterial pentalenolactone synthase gene, respectively, were expressed in Streptomyces lividans together with the GGPPS gene under the control of the ermE^* promoter. The transformant produced a novel cyclic diterpenoid with the same basic structure as that of terpentecin. To the best of our knowledge, this is the first report about a eubacterial diterpene cyclase. A dxs gene, which encodes a 1-deoxy-D-xylulose 5-phosphate (DXP) synthase participating in the initial step of the nonmevalonate pathway was also cloned from strain MF730-N6. The dxs, hmgr and GGPPS cyclase genes were used as probes for Northern blot analysis to examine temporal expression of the nonmevalonate and mevalonate pathways. A transcript of the dxs gene was detected throughout the cultivation. On the other hand, messages of the hmgr gene and the GGPPS cyclase gene were not detected at early growth phase but appeared when terpentecin production started, suggesting that the nonmevalonate pathway was essential for growth and the mevalonate pathway was mainly used for the secondary metabolism.
Benzoisochromanequinone (BIQ) antibiotics are a class of aromatic polyketides produced by Streptonzyces spp. Actinorhodin (ACT) produced by Streptomyces coelicolor A3(2) has served as one of the pioneering model compounds for studying the genetic programming of aromatic polyketides whose basic carbon skeleton is formed by a type II polyketide synthase. Elucidation of the biosynthetic pathway of ACT would lead to comprehensive understanding of secondary metabolism because S. coelicolor is currently the subject of a genome sequencing project. The present study focuses on the BIQs, the granaticins (GRAs), whose nascent polyketide is identical to that of ACT. The post-PKS steps involve different pyran-ring stereochemistry and glycosylation. Comparison of the complete gene clusters for ACT and GRA is therefore of great interest. The entire GRA biosynthetic gene cluster was cloned from S. violaceoruber Tu22, and expressed in the heterologous host, S. coelicolor A3(2) strain CH999. Chemical analysis of the recombinants carrying the cosmid, pOJ446-22-24, showed comparable production of GRAs. Sequencing analysis of the cosmid identified 37 complete open reading frames (ORFs). There are 15 ORFs that resemble each other on the act and gra clusters. Although the fact that granaticin has the opposite stereochemistry from ACT led us to search for the gra homologues for actVI-ORF1, the stereochemical determinant at C-3 of ACT, the gra cluster lacks obvious candidates. There exist gra ORF products having a characteristic nucleotide binding motif: ORFs 5,6,17,22, and 26. The products of ORFs 17, 22, and 26 are most likely to be involved in the DOH pathway because of their high degree of homologies with other deduced gene products responsible for deoxysugar formation. An explored possibility is that the stereochemistry in the pyran ring of GRA biosynthesis would be controlled by ORF 5/6. A pathway-specific transcriptional activator gene (gra-ORF9) was identified. The gene product (named SARP: Streptomyces antibiotic regulatory protein) is presumed to bind the direct repeat (DR) sequences in the promoter regions of Streptomyces antibiotic biosynthetic genes. A search for such DR sequences in the gra cluster revealed an array of hexameric repeats separated by fixed numbers of nucleotides corresponding one or two turns of the DNA helices.