Structures of the Largest Amphidinol Homologues from the

Nov 9, 2017 - ... 2017 The American Chemical Society and American Society of Pharmacognosy. *Phone (M. Satake): +81-3-5841-4357. E-mail: msatake@chem...
0 downloads 6 Views 2MB Size
Article Cite This: J. Nat. Prod. 2017, 80, 2883-2888

pubs.acs.org/jnp

Structures of the Largest Amphidinol Homologues from the Dinoflagellate Amphidinium carterae and Structure−Activity Relationships Masayuki Satake,*,† Kimberly Cornelio,‡,§ Shinya Hanashima,‡ Raymond Malabed,‡ Michio Murata,*,‡,§ Nobuaki Matsumori,⊥ Huiping Zhang,∥ Fumiaki Hayashi,∥ Shoko Mori,▽ Jong Souk Kim,¶ Chang-Hoon Kim,¶ and Jong-Soo Lee*,# †

Department of Chemistry, School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan Department of Chemistry, Graduate School of Science, Osaka University, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043, Japan § JST ERATO Lipid Active Structure Project, 1-1 Machikaneyama, Toyonaka, Osaka 560-0043, Japan ⊥ Department of Chemistry, Graduate School of Sciences, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan ∥ RIKEN Center for Life Science Technology, 1-7-22, Suehiro-cho, Tsurumi-ku, Yokohama 230-0045, Japan ▽ Bioorganic Research Institute, Suntory Foundation for Life Sciences, Sikadai, Seika-cho, Soraku-gun, Kyoto 619-0284 Japan ¶ Department of Marine Bio-materials & Aquaculture, Pukyong National University, Busan 608-737, Korea # Department of Seafood and Aquaculture Science, College of Marine Science, Gyeongsang National University, Tongyeong, Kyungnam 650-160, Korea ‡

S Supporting Information *

ABSTRACT: Amphidinols are polyketide metabolites produced by marine dinoflagellates and are chiefly composed of a long linear chain with polyol groups and polyolefins. Two new homologues, amphidinols 20 (AM20, 1) and 21 (AM21, 2), were isolated from Amphidinium carterae collected in Korea. Their structures were elucidated by detailed NMR analyses as amphidinol 6-type compounds with remarkably long polyol chains. Amphidinol 21 (2) has the longest linear structure among the amphidinol homologues reported so far. The congeners, particularly amphidinol 21 (2), showed weaker activity in hemolysis and antifungal assays compared to known amphidinols.

A

proposed AM mode of action21−24 involves binding to lipids in the bilayer membrane, chiefly with the polyene part, which increases membrane permeability. The size of the pore/lesion is dependent on the features of the terminal polyol region. A hairpin-shaped conformation of the central region is stabilized by hydrogen bonds in amphipathic environments. In a previous study,24 we proposed a barrel-stave pore assembled as an amphidinol 3 (AM3, 3) complex based on solid-state NMR, where the polyol chain of AM3 forms the lining of the channel. However, previous structure−activity relationship (SAR) studies have revealed that the polyol portion has a relatively small influence on the membrane-disrupting activity as compared with the polyene chain.25 In this study, we report the structures of 1 and 2 and a structure−activity relationship study of the length of the polyol chain of AMs with respect to membrane-disrupting activity.

mphidinols (AMs) are polyketide metabolites produced by Amphidinium spp. dinoflagellates. Since the first member of this family1 was reported as a potent antifungal agent in 1991, 19 amphidinols2−10 and closely related analogues, luteophanols,11 lingshuiols,12,13 karatungiols,14 symbiopolyols,15 karlotoxins,16−18 carteraols,19 and amdigenols20 have been reported, forming an important group of natural bioactive marine products. Amphidinols and the related compounds have been isolated from dinoflagellates collected in Japan, New Zealand, Taiwan, Spain, and the United States. In the present study, two new amphidinol analogues, amphidinol 20 (AM20, 1) and amphidinol 21 (AM21, 2), were isolated from Amphidinium carterae collected in Korea (Chart 1). The general structure of AMs is best characterized by a very long carbon chain encompassing multiple hydroxy and olefinic functionalities. All AMs and congeners have a common central core structure that consists of two tetrahydropyrans linked by a C6 alkyl chain spacer. Therefore, structural variations among AMs reside in the terminal polyol and olefinic constituents. In addition to the antifungal activity, the structurally related karlotoxins16−18 and carteraols19 show lethal fish toxicity. The © 2017 American Chemical Society and American Society of Pharmacognosy

Received: April 20, 2017 Published: November 9, 2017 2883

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888

Journal of Natural Products

Article

Chart 1

Figure 1. Proposed cleavage sites of 1 in the negative ion MS/MS experiments.



13

RESULTS AND DISCUSSION The dinoflagellate A. carterae was collected in South Korea, and 21.4 mg of amphidinol 20 (AM20, 1) and 4.7 mg of amphidinol 21 (AM21, 2) were isolated from 112 L of cultured cells using successive column chromatography. Amphidinol 20 was obtained as a pale yellow, amorphous solid. The UV absorptions at 260, 270, and 280 nm indicated that 1 had a conjugated triene chromophore in the molecule. The molecular formula of 1 was deduced to be C87H152O27 by the HRMS of sodium adduct ions observed at m/z 1652.0417. 1H and 13C NMR spectra measured in methanol-d4/pyridine-d5 (2:1) indicated that 1 possessed four doublet and two olefinic methyls. The NMR chemical shifts of five (C-82: δC 18.7; C-83: δC 15.5; C-84: δC 20.3; C-85: δC 12.7; C-86: δC 16.2) of six methyls identical to those in amphidinol 2 (AM2, 4) suggested that the structure of 1, including the central core, was similar to that of 4. Detailed analysis of the COSY, TOCSY, HSQC, and HSQC-TOCSY spectra led to elucidation of four partial structures (C-1 to C-31, C-33 to C-44, C-46 to C-56, and C-58 to C-80), which were interrupted by olefinic quaternary carbons. A polyol region from C-9 to C-20 was assigned by

C NMR and HSQC-TOCSY spectra due to severe overlapping of 1H NMR signals. Three methylenes around δC 22 and six methylenes around δC 37 corresponded to β and α methylenes of hydroxy-bearing carbons, respectively. The 13C chemical shifts and the number of those methylenes were identical to those of a polyol component in amphidinol 6 (AM6, 5). HSQC-TOCSY correlations from both β and α hydroxy methylenes to H-8, H-12, H-16, and H-20 were observed. These results indicated that the polyol structure from C-9 to C-20 consisted of repetition of a −CH2−CH2−CH2− CHOH− unit analogous with that of 5. Observed HMBC correlations H3-83/C-31, H3-83/C-32, H3-83/C-33, H3-86/C44, H3-86/C-45, H3-86/C-46, H-87/C-56, and H-87/C-58 enabled the assembly of the partial structures into the whole skeletal structure. The positions of the ether rings were also determined by HMBC correlations H-49/C-53 and H-64/C60. The configurations of the double bonds were deduced from proton coupling constants and 13C chemical shifts of Me-83 and Me-86. The large proton coupling constants of 14.9 Hz for H-2, 15.4 Hz for H-7, and 15.1 Hz for H-67 indicated that Δ2, Δ6, and Δ67 were in the E configuration. The 13C chemical 2884

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888

Journal of Natural Products

Article

shifts at δC 15.5 for Me-83 and δC 16.2 for Me-86 indicated an E configuration of Δ32 and Δ45. The NOE correlations H3-83/ H-34 and H3-86/H-47 also confirmed the E configuration. Positive and negative MS/MS experiments were conducted to confirm the structure (Figures S2, S3, and S4 in the Supporting Information). Unlike the MS/MS results of amphidinol 1, product ions generated from both termini were observed due to the absence of a charged site such as a sulfate ester on the terminal part of the molecule.1 Product ions (negative) at m/z 1542.0, 1500.0, and 1327.9 generated by α cleavage of hydroxy groups and the product ion at m/z 631.5 supported the polyol structure in 1 (Figure 1). Therefore, a planar structure of 1 was elucidated. The relative configurations of the ether rings were determined by proton/proton coupling constants and NOE correlations (Figure 2). The large proton coupling constants

Table 1. Antifungal and Hemolytic Activities of AM21 (2) and AM20 (1) compounda

antifungal (MECb in μg/disk; Aspergillus niger) hemolysis (EC50c in μM; human erythrocytes)

AM625 (5)

AM225 (4)

DsAM725 (6)

AM21 (2)

AM20 (1)

>15

>15

6

6

8

>10d

1−3e

2.9

1.7

1.2

a

Compounds are arranged by descending molecular weight. bMEC = minimum effective concentration. Maximum amount used was 60 μg/ disk. cMaximum concentration used was 80 μM. dAt around 10 μM, AM21 did not induce 50% hemolysis, but blood cells were partially lysed. eHemolysis of AM20 showed low concentration dependence. Blood cells were partly lysed below 1 μM, but 100% hemolysis did not occur even at 20 μM.

liposomes consisting of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) or POPC/cholesterol (Chol) (Figure 3).26 Results indicated that both AM20 and AM21 have

Figure 2. NOE correlations of tetrahydropyrans in 1.

(11 Hz) of H-51/H-52 and H-52/H-53 indicated axial orientations of H-51 and H-53 in the ether ring. In particular, NOESY correlations observed between H-48/H-51, H-48/H53, and H-51/H-53 confirmed a chair conformation of the ether ring in a 1,3-diaxial orientation. An NOE correlation between H-49/H-50 indicated equatorial orientations for H-49 and H-50. Analogously, the NOE correlation observed for H60/H-65 and H-60/H-62 suggested a chair conformation of the ether ring. Protons H-60 and H-62 were in axial orientations, and the NOE correlation observed for H-63/H-64 indicated equatorial positions for H-63 and H-64. The partial structure from C-28 to C-80 in 1 agreed with that from C-13 to C-65 in 4. Additionally, the partial structure from C-2 to C-20 in 1 was identical to that from C-4 to C-22 in 5. HRMS of 2 suggested a molecular formula of C94H166O30. It was 146 amu higher than that of 1, and an additional CH, CH3, two CH2, and three CH(OH) signals were observed in the HSQC spectra. The connectivity −CH(OH)−CH2−CH(CH3)−CH(OH)−CH(OH)−CH2− was assigned and positioned from C-25 to C30 by the COSY and TOCSY experiments. The sequence was also confirmed by an HSQC-TOCSY experiment. Other signals agreed with those in 1. MS/MS experiments were also conducted to confirm the structure of 2 (Supporting Information). In the positive mode (Figure S11), prominent product ions were observed at m/z 1025.8, 753.6, and 751.6. In the negative mode (Figure S12) product ions at m/z 1646.1, 1474.0, 977.7, and 891.7 were observed. These product ions supported the structure of the polyol part (C-4 to C-39) in 2 that was deduced from the NMR data. Thus, the structure of amphidinol 21 was determined as shown in 2 (Chart 1). AM20 and AM21 showed hemolytic activity against human erythrocytes with EC50 (Table 1). The dose dependency of hemolysis of AM20 or AM21 was very weak as compared with other AM homologues. Additionally, AM20 and AM21 did not inhibit the growth of Aspergillus niger at 15 μg/disk in an antifungal activity assay (Table 1). Membrane-disrupting activities of AM20 and AM21 were examined by using artificial

Figure 3. Membrane permeability activity of AM20 (1) and AM21 (2) measured with pure POPC and POPC/Chol (9:1 mol ratio) large unilamellar vesicles (LUVs) containing calcein.

significantly weaker activities compared with other AM analogues with a maximum of 20% dye leakage even at the highest AM concentration used (80 μM). The potent analogue AM3 was reported to cause 100% calcein leakage at about 10 μM.25 In the present study, we successfully elucidated the planar structures of two new amphidinols with the largest molecular weight among the reported homologues. In addition to the common structural portion (gray boxes in Figure 4), there are an additional 12 and 17 stereogenic centers in AM20 and AM21, respectively, which greatly hamper their stereochemical assignment. Their complete structural elucidation could be 2885

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888

Journal of Natural Products

Article

Figure 4. Membrane-disrupting (hemolytic) activity29 dependence on chain length of amphidinol homologues. aUnlike AMs 4, 5, and 6, which showed clear hemolysis, AM21 even at 10 μM did not induce 50% hemolysis. bHemolysis of AM20 showed low concentration dependence; blood cells were lysed to a certain extent below 1 μM, but 100% hemolysis did not occur even at 20 μM. Thus, we could not determine the EC50 values of AM20 and AM21.

caused by their longer polyol chains, as the only structural difference between AM20/AM21 and AM6 resides in the polyol moiety (Figure 4). We speculate that the longer chain of AM20/AM21 may prevent their penetration into the membrane interior, which is considered necessary for forming the barrel-stave pore as illustrated in Figure 5. Another interesting feature of their activity is sterol dependency; the potency of AM20/AM21 was not dependent on the level of Chol in the membrane (Figure 3). This finding supports our previous hypothesis24 that the polyol chain of AM3 with potent activity can penetrate membranes in the presence of sterol, and its terminal portion reaches the opposite side of the bilayer as shown in Figure 5A. Their weak activity, on the other hand, can be explained by the carpet-model-type mechanism,30,31 in which the shallow interaction of the polyol portion somehow disrupts the integrity of lipid packing in bilayers (Figure 5B).

achieved through synthetic studies as previously shown for AM3.27,28 The structural difference of AM21 from AM20 is the presence of a C7 unit corresponding to the C-27−C-32 part. This unit in AM21 can also be found in the C-21−C-26 part of AM20 and the corresponding portion of AM2 (Figure 4), but not in AM3 or its homologues AM4 and AM6. These differences in structure can be attributed to the insertion of gene fragments encoding the corresponding modules in the PKS of dinoflagellates. As seen in Figure 3, AM20 and AM21 showed weak activity in membrane-disrupting assays; AM20 and AM21 did not reach 50% leakage activity at the highest concentration of 20 μM, while the EC50 values of AM2 and AM3 were 3.0 and 2.5 μM, respectively, under the same conditions.26 The hemolysis activity of AM21 was weak compared with shorter chain homologues29 (Table 1 and Figure 4). This difference is likely 2886

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888

Journal of Natural Products

Article

Figure 5. Schematic illustration of mode of action.24 (A) Barrel-stave channel formed by short-chain AMs, such as dsAM7. (B) A longer polyol chain hampers penetration of an AM molecule into the membrane interior. Instead, disruption of membrane packing may occur by the binding of longer AMs, such as AM21, which somewhat mimics the “carpet model” proposed for membrane-active peptides. filtration on Toyopearl HW-40 (Tosoh) with MeOH/H2O (1:1) and then on Sephadex LH-20 (Pharmacia) with MeOH. Fractions containing AM (154.7 mg) were chromatographed repeatedly on a reversed-phase column (RPAQUOUS, Develosil) with 80% MeOH. From 112 L of cultures, 21.4 mg of 1 and 4.7 mg of 2 were isolated. Amphidinol 20 (AM20, 1): pale yellow, amorphous solid; UV (MeOH) λmax (log ε) 260 (3.98), 270 (4.01), 280 (3.93) nm; ECD (50 μM, MeOH), λmax (Δε) 210 (22.95) and 258 (−12.61) nm; NMR data, Table S1 in the Supporting Information; HRESIMS m/z 1652.0417 [M + Na]+ (calcd for C87H152O27Na, 1652.0413). Amphidinol 21 (AM21, 2): pale yellow, amorphous solid; UV (MeOH) λmax (log ε) 260 (3.98), 270 (4.01), 280 (3.93) nm; ECD (50 μM, MeOH), λmax (Δε) 209 (23.55) and 247 (−7.25) nm; NMR data, Table S2; HRESIMS m/z 1776.1548 [M + H]+ (calcd for C94H166O30 + H, 1776.1536). Biological Assays. Dried specimens of Aspergillus niger (NBRC No. 31032) obtained from the NITE Biological Resource Center (Chiba, Japan) were cultured at 25 °C in a glucose/peptone liquid medium (2% glucose, 0.2% yeast extract, 0.5% peptone, 0.05% MgSO4, and 0.1% KH2PO4) for 2 days. An aliquot of the broth was spread onto an agar plate made of the same medium with 1.5% agar. The samples were dissolved in MeOH, spotted onto 8 mm paper disks, and placed on the agar plate containing A. niger mycelia. After incubation at 25 °C for 2 days, the zone of inhibition for each paper disk was measured. For the hemolytic assay, human red blood cells (RBCs) were separated from plasma by centrifuging blood cells suspended in phosphatebuffered saline (PBS) buffer (100 mM NaCl, 2.68 mM KCl, 8.10 mM Na2HPO4, and 1.47 mM KH2PO4; pH 7.4) at 2000 rpm for 5 min. The pellet obtained after centrifugation was dissolved in PBS buffer to make a 10 mL volume, which was further diluted 10 times to obtain a 1% RBC suspension (∼1.7 × 107 erythrocytes/mL).25 Samples for the hemolysis assay were prepared in 600 μL Eppendorf tubes by mixing 190 μL of 1% RBC suspension with 10 μL of AM20 or AM21 (in various concentrations) or PBS buffer as a negative control. For 100% hemolysis, 180 μL of Milli-Q H2O suspension was mixed with 20 μL of 10% RBC. After incubation at 37 °C for 18 h, the samples were centrifuged at 2000 rpm for 5 min, and the supernatant from each tube was subjected to colorimetric measurements at 450 nm on a microplate reader (Corona Electric multimode reader). The % hemolysis and the EC50 (concentration that caused 50% hemolysis) were determined from the dose−response curves generated. Calcein Leakage Assay. Pure POPC or POPC/Chol (9:1 mol ratio) large unilamellar vesicles (LUVs) were prepared as previously described.26 POPC lipids with and without cholesterol were dissolved in 2 mL of CHCl3/MeOH (1:1). Solvent was evaporated at 30 °C under vacuum for 2 h and subjected to overnight drying in vacuo. The resulting lipid film was suspended in 1 mL of 10 mM Tris-HCl (pH

For example, membrane-active peptides that permeabilize lipid bilayers by the carpet model mechanism are generally less potent than those by the barrel-stave mechanism; the carpet peptide has to accumulate on the surface of the membrane to a certain extent to disrupt bilayer integrity, while the barrel peptide can self-assemble to form a pore even in low membrane concentrations.31 Another possible cause of the reduced biological activities of AM20 and AM21 may be their higher solubility due to a longer polyol chain and more hydroxy groups. As seen in AM homologues such as AM4 and AM9, higher water solubility due to a sulfate substitution always results in reduced biological activity.6 Heavier hydration also prevents the polyol chain from approaching the membrane surface, which is a necessary step to disrupt the membrane by either the carpet or barrel-stave mechanism.



EXPERIMENTAL SECTION

General Experimental Procedures. UV spectra were recorded on a V-630BIO spectrophotometer (JASCO). NMR spectra were recorded on a Bruker AVANCE III HD 600 MHz instrument in CD3OD/C5D5N (2:1). 1H and 13C chemical shifts (δ) are reported in ppm and referenced to the CD3OD residual solvent signal (δH 3.33 and δC 47.8). High-resolution MS and MS/MS spectra were measured with an Orbitrap Elite FT mass spectrometer (Thermo Fisher Scientific, Inc.). HPLC was performed on a Younglin SP 930 D pump (Younglin Instruments) equipped with a Younglin 730 D UV/ vis spectrophotometer at 272 nm. Chemicals and solvents were purchased from Wako Pure Chemicals and Kanto Kagaku and used without further purification. Culture of Dinoflagellates. Amphidinium carterae was isolated in Goseong Province, Korea. The isolated species was taxonomically identified by comparison of DNA sequences (GenBank accession numbers AY460579 and AY460584) with A. carterae strain nos. 550 and 551 from the Korea Marine Microalgae Culture Collection. A unialgal culture was cultured in seawater with f/2 supplement at 25 °C and harvested by centrifugation when the cell density reached 1.87 × 103 cells/mL. Isolation of 1 and 2. Harvested dinoflagellate cells (110 g wet) were extracted with MeOH three times. The extract was partitioned between CHCl3 and 40% MeOH. After elimination of MeOH, the aqueous layer was extracted with 1-butanol. The extract was fractionated on an ODS column by stepwise elution each with a 2fold column volume of 60%, 85%, and 100% MeOH in this order. The fraction eluting with 85% MeOH was further fractionated by gel 2887

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888

Journal of Natural Products

Article

7.5) containing 60 mM calcein and agitated for 30 min. A freeze−thaw cycle was carried out five times to obtain multilamellar vesicles. To obtain LUVs, the lipid suspensions were passed through a polycarbonate membrane filter (200 nm pore size) 19 times using a Lipsofast extruder (Avestin, Inc.). The excess calcein was separated from the calcein-entrapped LUVs by gel filtration using a Sepharose 4B (Sigma-Aldrich) with 10 mM Tris-HCl buffer (pH 7.5) containing 1 mM EDTA and 150 mM NaCl. The lipid and Chol concentrations were measured using a phospholipid C-test and Chol E-test (Wako Pure Chemical Industries, Ltd.). For the calcein leakage measurements, 20 μL of the lipid suspension was added to 170 μL of PBS buffer in a 96-well microplate. Fluorescence measurements (490 nm excitation/517 nm emission) were subsequently acquired (Corona Electric multimode reader). In all samples, lipid concentration was set to 27 μM. After the readings, 10 μL of AM20 or AM21 was added to the wells, and fluorescence measurements were acquired again. Finally, 20 μL of 10% (v/v) Triton X-100 was added to each well to cause 100% leakage, and fluorescence measurements were acquired.



(7) Morsy, N.; Houdai, T.; Matsuoka, S.; Matsumori, N.; Adachi, S.; Oishi, T.; Murata, M.; Iwashita, T.; Fujita, T. Bioorg. Med. Chem. 2006, 14, 6548−6554. (8) Meng, Y.; Van Wagoner, R. M.; Misner, I.; Tomas, C.; Wright, J. L. C. J. Nat. Prod. 2010, 73, 409−415. (9) Nuzzo, G.; Cutignano, A.; Sardo, A.; Fontana, A. J. Nat. Prod. 2014, 77, 1524−1527. (10) Cutignano, A.; Nuzzo, G.; Sardo, A.; Fontana, A. Mar. Drugs 2017, 15, 157. (11) Kubota, T.; Takahashi, A.; Tsuda, M.; Kobayashi, J. Mar. Drugs 2005, 3, 113−118. (12) Huang, X.; Zhao, D.; Guo, Y.; Wu, H.; Lin, L.; Wang, Z.; Ding, J.; Lin, Y. Bioorg. Med. Chem. Lett. 2004, 14, 3117−3120. (13) Huang, X.; Zhao, D.; Guo, Y.; Wu, H.; Trivellone, E.; Cimino, G. Tetrahedron Lett. 2004, 45, 5501−5504. (14) Washida, K.; Koyama, T.; Yamada, K.; Kita, M.; Uemura, D. Tetrahedron Lett. 2006, 47, 2521−2525. (15) Hanif, N.; Ohno, O.; Kitamura, M.; Yamada, K.; Uemura, D. J. Nat. Prod. 2010, 73, 1318−1322. (16) Van Wagoner, R.; Deeds, J. R.; Satake, M.; Ribeiro, A. A.; Place, A. R.; Wright, J. L. C. Tetrahedron Lett. 2008, 49, 6457−6461. (17) Peng, J.; Place, A. R.; Yoshida, W.; Anklin, C.; Hamann, M. T. J. Am. Chem. Soc. 2010, 132, 3277−3279. (18) Waters, A. L.; Oh, J.; Place, A. R.; Hamann, M. T. Angew. Chem., Int. Ed. 2015, 54, 15705−15710. (19) Huang, S.; Kuo, C.; Lin, Y.; Chen, Y.; Lu, C. Tetrahedron Lett. 2009, 50, 2512−2515. (20) Inuzuka, T.; Yamada, K.; Uemura, D. Tetrahedron Lett. 2014, 55, 6319−6323. (21) Swasono, R. T.; Mouri, R.; Morsy, N.; Matsumori, N.; Oishi, T.; Murata, M. Bioorg. Med. Chem. Lett. 2010, 20, 2215−2218. (22) Oishi, T.; Kanemoto, M.; Swasono, R.; Matsumori, N.; Murata, M.; Houdai, T.; Matsuoka, S.; Matsumori, N.; Murata, M. Biochim. Biophys. Acta, Biomembr. 2004, 1667, 91−100. (23) Houdai, T.; Matsuoka, S.; Morsy, N.; Matsumori, N.; Satake, M.; Murata, M. Tetrahedron 2005, 61, 2795−2802. (24) Espiritu, R.; Matsumori, N.; Tsuda, M.; Murata, M. Biochemistry 2014, 53, 3287−3293. (25) Morsy, N.; Konoki, K.; Houdai, T.; Matsumori, N.; Oishi, T.; Murata, M.; Aimoto, S. Biochim. Biophys. Acta, Biomembr. 2008, 2008, 1453−1459. (26) Morsy, N.; Houdai, T.; Konoki, K.; Matsumori, N.; Oishi, T.; Murata, M. Bioorg. Med. Chem. 2008, 16, 3084−3090. (27) Tsuruda, T.; Ebine, M.; Umeda, A.; Oishi, T. J. Org. Chem. 2015, 80, 859−871. (28) Ebine, M.; Kanemoto, M.; Manabe, Y.; Konno, Y.; Sakai, K.; Matsumori, N.; Murata, M.; Oishi, T. Org. Lett. 2013, 15, 2846−2849. (29) Morsy, N.; Houdai, T.; Matsuoka, S.; Matsumori, N.; Adachi, S.; Oishi, T.; Murata, M.; Iwashita, T.; Fujita, T. Bioorg. Med. Chem. 2006, 14, 6548−6554. (30) Boland, M. P.; Separovic, F. Biochim. Biophys. Acta, Biomembr. 2006, 1758, 1178−1183. (31) Shai, Y. Biopolymers 2002, 66, 236−248.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jnatprod.7b00345. ESIMS, MS/MS, 1D 1H and 13C NMR spectra, 1H−1H COSY, HSQC, and HMBC, and table of 1H and 13C NMR chemical shifts in CD3OD/C5D5N (2:1) for 1 and 2 (PDF)



AUTHOR INFORMATION

Corresponding Authors

*Phone (M. Satake): +81-3-5841-4357. E-mail: msatake@ chem.s.u-tokyo.ac.jp. *Phone (M. Murata): +81-6850-5774. E-mail: murata@chem. sci.osaka-u.ac.jp. *Phone (J.-S. Lee): +82-55-772-9145. E-mail: [email protected]. ORCID

Masayuki Satake: 0000-0001-8508-3551 Nobuaki Matsumori: 0000-0003-0495-2044 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was partly supported by Scientific Research KAKENHIs (S), (A), and (C) with Grant Nos. 16H06315, 25242073, and15K01798, respectively, and also by Korea Institute of Marine Science & Technology Promotion (Project No. D11111112H390000140).



REFERENCES

(1) Satake, M.; Murata, M.; Yasumoto, T.; Fujita, T.; Naoki, H. J. Am. Chem. Soc. 1991, 113, 9859−9861. (2) Paul, G. K.; Matsumori, N.; Murata, M.; Tachibana, K. Tetrahedron Lett. 1995, 36, 6279−6282. (3) Paul, G. K.; Matsumori, N.; Konoki, K.; Sasaki, M.; Murata, M.; Tachibana, K. In Harmful and Toxic Algal Blooms; Yasumoto, T.; Oshima, Y.; Fukuyo, Y., Eds.; I.O.C. of UNESCO: Paris, 1996; pp 503−506. (4) Paul, G. K.; Matsumori, N.; Konoki, K.; Murata, M.; Tachibana, K. J. Mar. Biotechnol. 1997, 5, 124−128. (5) Morsy, N.; Matsuoka, S.; Houndai, T.; Matsumori, N.; Adachi, S.; Murata, M.; Iwashita, T.; Fujita, T. Tetrahedron 2005, 61, 8606−8610. (6) Echigoya, R.; Rhodes, L.; Oshima, Y.; Satake, M. Harmful Algae 2005, 4, 383−389. 2888

DOI: 10.1021/acs.jnatprod.7b00345 J. Nat. Prod. 2017, 80, 2883−2888