Amdigenol D, a long carbon-chain polyol, isolated from the marine dinoflagellate Amphidinium sp.

https://doi.org/10.1016/j.tetlet.2020.152376Get rights and content

Highlights

  • A novel long carbon-chain polyol compound, amdigenol D, has been isolated.

  • Structure elucidation was performed by NMR spectroscopy and MS/MS analyses.

  • Amdigenol D includes two common core structures of amphidinol analogs.

  • Amdigenol D has a C101-linear carbon-chain and is a longest amdigenol analog to date.

Abstract

A novel polyol compound, amdigenol D, produced by the dinoflagellate Amphidinium sp. has been isolated. Its structure was elucidated by NMR spectroscopy and MS/MS analyses. Amdigenol D consists of a C101-linear carbon backbone including two core structures of amphidinol analogs. Amdigenol D is the longest amdigenol analog identified to date.

Introduction

Amphidinol analogs [1] are widely occurring secondary metabolites produced by the dinoflagellates. Amphidinol analogs are composed of a common core structure including two tetrahydropyrans linked by a highly oxidized carbon chain, and two terminal carbon chains which have a variety of length and functional groups. While identification of many typical amphidinol analogs has been reported [2], we previously reported the isolation and structural elucidation of an unusual amphidinol analog, amdigenol A, from the Okinawan dinoflagellate Amphidinium sp., which appeared to have two core structures of amphidinol analogs [3]. We also previously reported other amdigenol analogs with the same carbon chain terminals as those of amdigenol A, from the same dinoflagellate [4]. Amdigenols A, E and G have the inhibitory activity of N-type Ca2+ channel-opening [4], but no information about the relationship between their structural similarity and the difference of their carbon-chain length, and their function or role for the dinoflagellate has been clarified. Because the total synthesis of amphidinol 3 has recently been achieved [5], studies on the structure–function relationship of typical amphidinol analogs are expected to progress in the future. However, it is difficult to supply amdigenols which are larger and more structurally complex synthetically. Isolation and structural determination of additional amdigenol analogs from the same dinoflagellate should give more information about the mechanism on the biological and physiological activities of amdigenols, or the function of the core structure and the two terminal chains of amphidinol analogs. In this manuscript, we report the identification of a novel amdigenol analog, amdigenol D.

The dinoflagellate Amphidinium sp. gathered at Ishigaki island, Okinawa, Japan was cultured in a sea water medium for two months, then the sea water medium (200 L) was filtered to remove the dinoflagellate. Separation of sea water medium with column chromatography with TSK G-3000S polystyrene gel, DEAE-Sephadex, and Sephadex LH-20 gave a fraction containing some compounds whose molecular weight were about 2000. This fraction was purified with continuous reversed-phase HPLC and negative ion exchange HPLC, to afford amdigenol D (1) (3.7 mg) as a colorless amorphous solid.

From the HR-ESI-MS spectra, the molecular formula of amdigenol D (1) was found to be C107H179NaO44S (m/z 2224.1477 for [M + H]+, m/z 2200.1499 for [M−Na]). 1H NMR of amdigenol D seemed to be similar to that of amdigenol A. The obvious differences in the 1H NMR signals between amdigenols D and A was one oxymethine proton signal (δH 4.42 ppm) and signals of internal protons in a conjugated olefin (δH 6.10 and 6.30 ppm).

Analyses of COSY, TOCSY and HMQC spectra led to the seven partial structures, C1–C13, C14–C16 (including C15–Me), C17–C21, C23–C33, C35–C36, C37–C39, and C41–C48 (Fig. 1). Analysis of the HMBC correlations identified some linkages of the partial structures, C13–C14, C16–C17, C21–C23 (including C22–Me), C33–C35 (including C34–CH2), and C36–C37. An HMBC correlation of H26/C30 indicated the presence of a six-membered ring ether, C26–C30. The chemical shifts of H37/C37 and H38/C38 (Table 1) were different to those of amdigenol A (δH37 4.16 ppm, δC37 69.1 ppm, δH38 1.86 and 1.49 ppm, δC38 38.1 ppm), but similar to those of H78/C78 and H79/C79 of amdigenol A (δH78 4.05 ppm, δC78 70.9 ppm, δH79 2.08 ppm, δC79 31.3 ppm). An HMBC correlation of H37/C41 was also observed. From these NMR spectral data, it was estimated that a tetrahydropyran, C37 to C41, was present. The correlation to H40 or C40 could not be observed in COSY or HMBC spectra, but the existence of a tetrahydropyran, not a tetrahydrofuran, was estimated by the similarily of chemical shifts on H78, H79, C78 and C79. The geometries of the olefins, C6–C7, C10–C11, and C22–C23 were all E due to the NOESY correlations. The coupling constants showed the geometries of the conjugated olefin, C44–C45 (15.0 Hz) and C46–C47 (15.3 Hz) were E and E, respectively. As a result of these analyses, partial structure C1–C48 was elucidated (Fig. 1).

MS/MS analysis of amdigenol D was performed to confirm the result of NMR spectra analysis for the partial structure C1–C48. Observed fragment ion peaks of amdigenol D, m/z 827.4 and below, were the same as those of amdigenol A. Therefore, the partial structure C1 to C35 was the same as that of amdigenol A, including a sulfate and some hydroxyl groups. Three fragment ion peaks, m/z 929.5, 959.5, and 1001.5, confirmed the partial structure C36 to C42, including a tetrahydropyran moiety (Fig. 2). If this portion had been a tetrahydrofuran, observation of these fragment ion peaks would not have matched the proposed structure.

Analyses of 2D NMR spectra also revealed the partial structure C1′–C17′ as shown in Fig. 3. COSY, TOCSY and HMQC spectra showed the linkage of C1′–C17′. From NOESY correlation, three olefin geometries of C6′–C7′, C8′–C9′ and C13′–C14′ were all E. The chemical shifts of H1′–H17′ and C1′–C17′ could be assigned (Table 1), and these were similar to the H82–H98 and C82–C98 of amdigenol A. As a result, the partial structure C1′–C17′ of amdigenol D was analyzed as the opposite terminal of the linear carbon backbone, and aligned with amdigenol A. Observation of the fragment ion peak, m/z 1948.1, on the MS/MS measurement supported the existence of the terminal partial structure C1′–C15′.

Although amdigenol D has 107 carbon atoms, 2 correlations from methyl protons to corresponding carbons, 20 correlations from methylene or methyne protons to corresponding carbons, 26 correlations from oxymethylene or oxymethyne protons to oxidized carbons, and 21 correlations from olefin protons to olefin carbons were observed in the HMQC spectrum of amdigenol D. In the 1H NMR spectrum, some signals were observed as two protons, even though they were analyzed as one proton in the structural analysis of the partial structure C1–C48, for example H24 (δH 4.54 ppm), H34–CH2H 5.08 and 4.99 ppm). Additionally, amdigenol A, an analog of amdigenol D, has the same partial structures [3]. Therefore, it was thought that amdigenol D had the continuous portions in its structure, and some signals of the protons and carbons at these portions were overlapping.

Fragment ion peaks of amdigenol D, m/z 1139.6 and above, observed in the MS/MS experiment of amdigenol D were also analyzed (Fig. 4). Eight ion peaks, m/z 1167.6, 1223.6, 1281.6, 1369.7, 1413.7, 1483.8, 1687.9, and 1948.1, were 54 mass unit larger than corresponding ion peaks observed in the MS/MS measurement of amdigenol A [3]. The molecular weight of amdigenol D is 54 mass unit larger than that of amdigenol A. Therefore, these eight fragment ion peaks of amdigenol D would be observed by the result of the fragmentation at the same positions as amdigenol A. This means that amdigenol D is likely to have the same partial structure, C1′–C50′, as the C49–C98 part of amdigenol A.

From the molecular formula, it was inferred that C3H4 atoms existed between two partial structures C1–C48 and C1′–C50′. Because no cyclopropane, methyl or vinyl methyl group can be identified by NMR spectra other than those exist in two partial structures, two partial structures were linked by one olefin and one methylene. Of the two possible connections, the C49–C51 portion illustrated as (b) in Fig. 5 was not likely, as the chemical shift of H48 was not indicative of that of a proton suitable between two olefins. Therefore, it was determined that the C49–C51 portion illustrated as (a) in Fig. 5 was more likely.

As a result of these analyses, the planar structure of amdigenol D was determined as shown in Fig. 6. Detailed NMR spectra analysis assigned 1H and 13C NMR signals as shown in Table 1. The 1H and 13C chemical shifts of H8/C8–H41/C41 were assigned to be the same as those of H52/C52–H85/C85. An olefin geometry of C50-C51 was determined as E by the NOESY correlation of H50-H52 (Fig. 5). The chemical shifts of H49, C49 and C90 could not be determined exactly because of the heavily overlapping signals in the COSY, TOCSY, HMQC and HMBC spectra.

NOESY correlations also clarified the relative configuration of two tetrahydropyrans, A and B (Fig. 7). Correlations of H24–H27, H24–H28, H24–H30, H25–H28, H25–H30 and H28–H30 revealed the relative configuration of ring–A as shown in Fig. 7(a). Correlations of H37–H38a (δH 2.06 ppm), H37–H42, H37–H43, H38a–H39, H39–H42, H39–H43 and H40–H42 revealed the relative configuration of ring–B as shown in Fig. 7(b). From the similarity of chemical shifts and NOESY correlations at ring–C and ring–D, tetrahydropyrans, C and D seemed to have the same relative configuration as A and B, respectively. These relative configuration of tetrahydropyrans were the same as those of amdigenol analogs and general amphidinol analogs.

In order to confirm the structure of amdigenol D, a degradation reaction of amdigenol D was carried out to reveal the structure of the obtained smaller segments. During the structural analysis of amdigenol A [3], an ethenolysis with Grubbs catalyst was used for cleaving olefins of amdigenol A, but allyl vic-diol cleavage occurred as a side reaction [6]. Therefore, the diol cleavage reaction with Grubbs catalyst [7] was used to perform the degradation of amdigenol D (Scheme 1). As the degradation reaction produced many segments, the clean isolation of each segment was difficult. Therefore, the reaction mixture was separated by reverse-phase HPLC and a fraction of segments with a sulfate group was obtained. This fraction was then subjected to ESI-MS measurement and some ion peaks of segments, C1–C16 (2) and C1–C24 (3), with expected structures were observed [8]. The diol cleavage reaction, which had originally been seen as a side reaction in an ethenolysis with Grubbs catalyst, was shown to be a useful method for the decomposition of long carbon-backbone polyol compounds.

The structure of amdigenol D is characterized by including two core structures of the amphidinol analog. Amdigenol A was also thought to have two core structures of the amphidinol analog, but one of the two was the structure of a biosynthetic precursor [3]. Therefore, amdigenol D is expected to be biosynthesized by first constructing a long carbon-chain backbone and then being oxygen-functionalized. Identification of amdigenol D confirmed that amdigenol analogs whose molecular weight were over 2000 mass unit were huge amphidinol analogs with two core structures in their backbone. The impact of an increase in the number of cores on the shape of the compound and biological activity or function are issues for future research with a variety of amphidinol and amdigenol analogs. All of the identified amdigenol analogs, amdigenols A, D, E and G, have the same two side chain structures. Amphidinol analogs produced by other dinoflagellates have the same core structure as amdigenols, but different linear side-chain structures. The function and role of the side-chain in amphidinol analogs also needs to be established.

In conclusion, we performed the isolation and structural determination of a novel long carbon-backbone polyol compound, amdigenol D. Amdigenol D contains two core structures of amphidinol analogs and is the longest amdigenol analog isolated and structurally determined to date. Further studies on the relative and absolute configurational analyses and the biological and the physiological activities of amdigenol D are underway.

Section snippets

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

We thank Dr. T. Horiguchi (Hokkaido University) for identifying the dinoflagellate. This study was supported in part by a Grant-in-Aid for Scientific Research from JSPS (18K05333). We thank Alison McGonagle, PhD, from Edanz Group (https://en-author-services.edanzgroup.com/ac) for editing a draft of this manuscript.

References (8)

  • K. Washida et al.

    Tetrahedron lett.

    (2006)
    R.M. Van Wagoner et al.

    Tetrahedron Lett.

    (2008)
    J. Peng et al.

    J. Am. Chem. Soc.

    (2010)
    S.A. Rasmussen et al.

    J. Nat. Prod.

    (2017)
    S.-J. Huang et al.

    Tetrahedron Lett.

    (2009)
    X.-C. Huang et al.

    Tetrahedron Lett.

    (2004)
    Y. Doi et al.

    J. Org. Chem.

    (1997)
    T. Kubota et al.

    Tetrahedron

    (1998)
    T. Kubota et al.

    Mar. Drugs

    (2005)
    N. Hanif et al.

    J. Nat. Prod.

    (2010)
  • T. Oishi et al.

    J. Am. Chem. Soc.

    (2020)
  • M. Kita et al.

    Tetrahedron

    (2007)
  • C. Han et al.

    Tetrahedron lett.

    (2008)
    C. Han et al.

    Tetrahedron

    (2011)
There are more references available in the full text version of this article.

Cited by (1)

  • Marine natural products

    2022, Natural Product Reports
View full text