HETEROCYCLES

Short Paper | Regular issue | Vol. 87, No. 12, 2013, pp. 2615-2623
Received, 14th September, 2013, Accepted, 25th October, 2013, Published online, 5th November, 2013.
DOI: 10.3987/COM-13-12841

■ Iriomoteolides-4A and -5A, Hydrophilic Macrolides from Marine Dinoflagellate Amphidinium species

Keiko Kumagai,* Masashi Tsuda,* Eri Fukushi, and Jun Kawabata

Science Research Center, Kochi University, Oko Kohasu, Nankoku, Kochi 783-8506, Japan

Abstract

Two new macrolides, iriomoteolides-4a (1) and 5a (2), have been isolated from a marine benthic dinoflagellate Amphidinium sp. (strain HYA024), and the structures was assigned by detailed analyses of 2D NMR data. Iriomotelide-4a (1) is a 16-membered macrolide with five hydroxyl groups on the lactone ring and an isoprenoid-like side chain, while iriomoteloide-5a (2) is a 20-membered macrolide with five hydroxyl groups and three vicinally-located C1 branches. Iriomoteolides-4a (1) and 5a (2) exhibited moderate cytotoxic activity against antitumor cells.

Amphidinium dinoflagellates have been known as producers of polyketide-like metabolites with unique structural features.1,2 We have screened numerous Amphidinium strains by using genetic analyses,3 cytotoxic screening, and metabonomics analyses, and found an Amphidinium strain, named HYA024, which produced unknown cytotoxic macrolides. Three new cytotoxic 20-membered macrolides, iriomoteolides-1a, 1b, and 1c, and a 15-membered macrolide, iriomoteolide-3a, have been isolated from the strain.4-6 Further examination of the extract led to the isolation of two new 16- and 20-membered macrolides, iriomoteolides-4a (1) and -5a (2). Herein we describe the isolation and structure elucidation of 1 and 2.
The Amphidinium strain, HYA024 was monoclonally separated from sea sand collected off Iriomote Island, Japan. The cultured algal cells (15 g, dry weight) obtained from 400 L of the medium were extracted with the MeOH/toluene solvent system. The toluene soluble materials of the extract were subjected to SiO2 gel, C18, and NH2-SiO2 columns followed by C18 HPLC to afford iriomoteolides-4a (1, 0.001 %) and -5a (2, 0.001%). Known iriomoteolides were obtained from a less-polar fraction of the SiO2 gel column. The structures of 1 and 2 were shown in Figure 1.

Iriomoteolide-4a (1) showed the pseudomolecular ion peaks at m/z 531 (M+Na)+ in the positive-mode ESIMS spectra, respectively. The molecular formula, C29H48O7, of 1 was established by HRESIMS data [m/z 531.3336 (M+Na)+, ∆ +0.1 mmu]. 1H and 13C NMR data (Table 1) in CDCl3 assigned by using HMQC spectrum disclosed the presence of a total of 29 carbon signals due to an ester carbonyl, an sp2 quaternary carbon, seven sp2 methines, nine sp3 methines including six oxygenated ones, six sp3 methylenes, and five methyls.

The planar structure of 1 was elucidated on the basis of 2D NMR data measured in CDCl3. Analyses of 1H-1H COSY and TOCSY spectra revealed a continued spin network from H2-2 to H-23, H3-26, H3-27 and H3-28 (Figure 2). Three disubstituted E-double bonds at C-12, C-17 and C-21 were indicated by J(H-12/H-13) (15.6 Hz), J(H-17/H-18) (15.6 Hz) and J(H-21/H-22) values (15.1 Hz). Both of two singlet methyl signals (H3-25; δH 1.73, H3-29; δH 1.75) showed HMBC correlations for an sp2 quaternary and an sp2 methine carbons (C-23; δC 125.1, C-24; δC 133.2, respectively), suggesting the presence of an isobutene terminus. The HMBC correlation for H2-2 (δH 2.51, 2H)/C-1 (δC 174.5) implied an ester carbonyl was attached to C-2, and the relative low field resonance for H-15 (δH 4.79) suggested that C-15 was involved in an ester linkage with C-1. Thus, the planar structure of iriomoteolide-4a was revealed to be a 16-membered macrolide associated with five hydroxyl groups and four olefins.

The relative stereochemistry of nine chiral centers in the macrocyclic ring was assignable from analyses of sp3-sp3 bond rotations based on 1H-1H coupling constants and NOESY data (Figure 3). Magnitudes of 1H-1H coupling constants were estimated by selective population transfer (SPT) experiments and intensities of correlations observed for 1H-1H COSY spectrum, when signals were overlapped with other signals or had multiple couplings. Nevertheless, the nJCH were not obtained due to the small amount of samples. Compound 1 is difficult to get by re-cultivation of this dinoflagellate from the reason why the macrolide producing ability may decrease significantly or be lost.
The 1,3-syn relation for 4-OH and 6-OH was inferred by relatively large 1H-1H couplings for H-4/H-5a and H-5a/H-6, and rather small ones for H-4/H-5b and H-5b/H-6 and NOESY correlations for H-4/H-6 and H-5b/H-7 (Figures 3a, 3b, and 3i). Bond rotation analyses as shown in Figure 3c was suggested to be erythro for C-6−C-7. The 1,3-anti relation for the 7-methyl (C-26) and the 9-hydroxyl groups was deduced from coupling magnitudes of H-7/H2-8 and H2-8/H-9 and NOESY correlations as shown in Figures 3d and 3e. Both erythro configurations for C-9−C-10 and C-10−C-11 were assigned by 1H-1H coupling constants for H-9/H-10, H-10/H-11, and H-11/H-12 (4.5, 2.8 and 6.3 Hz, respectively) and NOESY correlations for H-8a/H-12 and H-8b/H-10 (Figure 3f, 3g, and 3j). Thus the relative configuration was concluded to 4S*, 6S*, 7R*, 9R*, 10R*, and 11S*. On the other hand, the C-15−C-16 bond (Figure 3h) was elucidated to be erythro by NOESY correlations for H-14a/H3-27, H-14b/H-15, H-15/H-17, and H-15/H3-27 as well as 1H-1H coupling constants for H-15/H-16 (8.0 Hz). Nevertheless, the relation between C-11 and C-15 through four carbon−carbon bonds is not

assignable unambiguously, because transannular NOE was not observed.
HRESIMS data [m/z 531.3336 (M+Na)+, ∆ +0.1 mmu] of iriomoteolide-5a (2) established the molecular formula C31H52O7. 1H and 13C NMR data (Table 2) in CDCl3 assigned by using HMQC spectrum disclosed the presence of a total of 31 carbon signals due to an ester carbonyl, two sp2 quaternary carbons, four sp2 methines, two sp2 methylenes, ten sp3 methines including six oxygenated ones, seven sp3 methylenes, and five methyls.

Detailed analyses of 1H-1H COSY and TOCSY spectra revealed three spin systems from H-2 to H-13, H3-26, H3-27 and H3-28, from H2-15 to H2-21, and from H-23 to H3-25 and H3-31 (Figure 4). J(H-4/H-5) and J(H-17/H-18) values (15.4 and 15.8 Hz, respectively) were suggestive of both E-geometries for two disubstituted double bonds at C-4 and C-17. HMBC correlations for H3-28 (δH 1.15)/C-14 (δC 148.5), H2-29 (δH 5.07 and 5.01)/C-14, and H2-29/C-15 (δC 43.6) indicated that C-13 was attached to C-15 through an exomethylene unit (C-14−C-29). Connection of C-21 and C-23 via another exomethylene unit (C-22−C-30) was deduced from HMBC correlations for H2-30 (δH 4.93, 2H)/C-21 (δC 29.4), H2-30/C-23 (δC 48.9) and H3-31 (δH 0.99)/C-22 (δC 151.1). The methyl proton on C-2 (H3-26, δH 1.17) showed an HMBC correlation to the carbonyl carbon (C-1, δC 175.7), indicating the attachment of the ester carbonyl to C-2. The relatively low-field resonance of H-19 (δH 5.30) suggested the existence of 20-membered macrolactone ring with an ester linkage between C-1 and C-19. Thus, the planar structure of iriomoteolide-5a was concluded to be 2.

Although iriomoteolide-5a (2) possessed ten chiral centers, it was limiting for 1,2- and 1,3-relations to elucidate the stereochemistry from bond rotation analyses. The erythro relation for C-2−C-3 bond (Figure 5a) was suggested by NOESY correlations for H-4/H3-27 and H3-26/H3-27 and the small coupling magnitude for H-2/H-3. The J(H-10/H-11a), J(H-10/H-11b), J(H-11a/H-12), and J(H-11b/H-12) values (4.0, 6.1, 7.3, and 4.3 Hz, respectively) implied gauche for H-10−H-11a and H-11b−H-12 and anti for H-10−H-11b and H-11a−H-12 (Figure 5b and 5c). Both gauche relations for C-9−C-12 and C-10−C-13 were elucidated from NOESY correlations for H-9b/H-12, H-10/H-13, and H-10/H3-28, thus indicating the 1,3-anti relation for two hydroxyl groups at C-10 and C-12. The C-12−C-13 and C-23−C-24 bonds were concluded to be threo and erythro, respectively, from 1H-1H coupling data and NOESY correlations as shown in Figures 5d and 5e. Although relatively long-range NOE’s for H-5/H-7, H-13/H-16 and H-15a/H-18 were observed, the relative stereochemistries for these three units and three isolated chiral centers of C-7, C-16, and C-19 remained unknown.

Iriomoteolide-4a (1) is a novel 16-membered macrolide having a unique carbon skeleton associated with five hydroxyl groups and an isoprene-like side chain. 1 is the second example with a 16-membered macrocyclic ring in Amphidinium macrolides.7 The isoprene-like side chain may be generated from a polyketide chain with C1 branches derived from C-2 of acetate.8 Iriomoteolide-5a (2) is a novel 20-membered macrolide possessing 5-hydroxyl groups and three portions of vicinally locating C1 branches. Although two classes of 20-membered macrolides such as amphidinolides A and U had been isolated from the symbiotic dinoflagellate Amphidinium species,9,10 the carbon chain length and C1-branched and oxygen-substituted positions for 2 are quite different from those of these known 20-membered macrolides. Our preliminary in vitro screening on antitumor activity showed that iriomoteolide-4a and -5a exhibited moderate cytotoxicity against human B lymphocyte DG-75 (IC50: 0.8 and 1.0 µg/mL, respectively).

EXPERIMENTAL
General. Optical rotations were measured on a JASCO DIP-370 polarimeter. IR spectra were recorded on a JASCO FT/IR-5300 spectrophotometer. 1H, 13C, and 2D NMR spectra were measured on a Bruker AMX-500 spectrometer using 2.5 mm micro cells for CDCl3 (Shigemi Co., Ltd.). Chemical shifts in CDCl3 are reported in ppm with reference to the solvent residual proton and carbon signals (δH 7.26 and δC 77.0). 1H-1H Coupling constants were estimated by SPT analyses and magnitudes of 1H-1H COSY correlations. ESIMS spectra were obtained on a JEOL JMS 700-TZ spectrometer at -80 V as a focus voltage using a sample dissolved in MeOH with flow rate of 200 µL/min.
Isolation. Cultivation and extraction were described previously.4 The toluene-soluble fractions (2 g) obtained from the harvested HYA024 cells (15.3 g, from 400 L of culture) were subjected to SiO2 column chromatography (40 x 200 mm) using a stepwise elution of CHCl3 (200 mL) and CHCl3/MeOH (98:2, 200 mL and then 95:5, 200 mL). The fraction eluted with (CHCl3/MeOH, 95:5) was chromatographed successively by using a C18 (MeCN/H2O, 7:3) and then NH2-SiO2 columns (n-hexane/EtOAc, 2:1). A macrolide-containing fraction was separated by C18 HPLC [YMC-Pack Pro C18, 5 µm, YMC Co., Ltd., 10 x 250 mm; eluent, MeCN/H2O (60:40); flow rate, 2 mL/min; UV detection at 210 nm] to afford iriomoteolides-4a (1, 0.001%) and 5a (2, 0.001%).
Iriomoteolide-4a (1). Colorless amorphous; [α]D20 -20 (c 0.02, CHCl3); IR (neat) νmax 3300 (broad) and 1721 cm-1; 1H and 13C NMR data (Table 1); ESIMS (pos.) m/z 531 (M+Na)+; ESIMS (neg.) m/z 543 and 545 [ca. 3:1, (M+Cl)-]; HRESIMS m/z 531.3336 [calcd for C29H48O7Na, (M+Na)+: 531.3336].
Iriomoteolide-5a (2). Colorless amorphous; [α]D22 +65 (c 0.02, CHCl3); IR (neat) νmax 3420 (broad) and 1718 cm-1; 1H and 13C NMR data (Table 2); ESIMS (pos.) m/z 559 (M+Na)+; ESIMS (neg.) m/z 571 and 573 [ca. 3:1, (M+Cl)-]; HRESIMS m/z 559.3615 [calcd for C31H52O7Na, (M+Na)+: 559.3615].
Cytotoxic Assay. Human B lymphocyte DG-75 cells were seeded at a density of 5000 cells per well into 96-well plates in culture medium containing 10% FBS. After 72 h, the number of viable cells was counted using Cell Counting Kit 8 (Dojindo Co., Kumamoto, Japan) according to the manufacturer's instructions. The assay reagent is a tetrazolium compound (WST-8) that is reduced by live cells into a colored formazan product measured at 450 nm using a microplate reader (Bio-Rad, USA). The viability of the treated groups was estimated as a percentage of control groups. The cytotoxicity was shown as the concentration causing a 50% reduction of cell growth (IC50). Doxorubicin and 5-fluorouridine (IC50: 0.04 and 1.2 µg/mL, respectively) were used as authentic samples, and the experiments were repeated in triplicate wells.

ACKNOWLEDGEMENTS
We thank Y. Endo, Y. Nagakita, and Y. Fukuda, for help with dinoflagellate cultivation and S. Oka, Center for Instrumental Analysis, Hokkaido University, for ESIMS measurements.

References

1. J. Kobayashi and M. Tsuda, Nat. Prod. Rep., 2004, 21, 77. CrossRef
2. J. Kobayashi and T. Kubota, J. Nat. Prod., 2007, 70, 451. CrossRef
3. R. Iwamoto, J. Kobayashi, T. Horiguchi, and M. Tsuda, Phycologia, 2005, 44 Supplement, 104.
4. M. Tsuda, K. Oguchi, R. Iwamoto, Y. Okamoto, J. Kobayashi, E. Fukushi, J. Kawabata, T. Ozawa, A. Masuda, Y. Kitaya, and K. Omasa, J. Org. Chem., 2007, 72, 4469. CrossRef
5. M. Tsuda, K. Oguchi, R. Iwamoto, Y. Okamoto, E. Fukushi, J. Kawabata, T. Ozawa, and A. Masuda, J. Nat. Prod., 2007, 70, 1661. CrossRef
6. K. Oguchi, M. Tsuda, R. Iwamoto, Y. Okamoto, J. Kobayashi, E. Fukushi, J. Kawabata, T. Ozawa, A. Masuda, Y. Kitaya, and K. Omasa, J. Org. Chem., 2008, 73, 1567. CrossRef
7. M. Ishibashi, M. Takahashi, and J. Kobayashi, Tetrahedron, 1997, 53, 7827. CrossRef
8. J. Kobayashi, M. Takahashi, and M. Ishibashi, J. Chem. Soc., Chem. Commun., 1995, 1639. CrossRef
9. J. Kobayashi, M. Ishibashi, H. Nakamura, Y. Ohizumi, T. Yamasu, T. Sasaki, and Y. Hirata, Tetrahedron Lett., 1986, 27, 5755. CrossRef
10. M. Tsuda, T. Endo, and J. Kobayashi, Tetrahedron, 1999, 55, 14565. CrossRef