Gliomasolides A–E, unusual macrolides from a sponge-derived fungus Gliomastix sp. ZSDS1-F7-2

Abstract

Five unusual macrolides, gliomasolides A–E (1–5), were isolated from a sponge-derived fungus Gliomastix sp. ZSDS1-F7-2. Their structures were established by extensive spectroscopic analyses and single-crystal X-ray diffraction studies. Structurally in common, these compounds feature a very rare 14-membered macrocyclic backbone lacking of methyl group. Cytotoxic effect of these metabolites against the growth of HeLa cell lines was evaluated, revealing that 1 exhibited moderate inhibitory effect with an IC₅₀ value of 10.1 μM.

---

Natural products have been a rich source of drugs and drug leads for many years.¹ Macrolides, frequently discovered from various marine microorganisms, are of particular importance to this end because of their significant biological effects and unique structures,² which therefore make them prominent targets for both synthetic chemists and pharmacologists.³ As versatile members of the large macrolides family, 14-membered macrolides attract tremendous attention since they often demonstrate excellent pharmacokinetics and efficacy: interesting examples involving clarithromycin and erythromycin,⁴ both of which have been approved by US FDA for the treatment of bacterial infections.

As a result of our continuous investigation of bioactive metabolites from marine organisms,⁵ we have discovered five unusual 14-membered macrolides, gliomasolides A–E (1–5) (Fig. 1), produced by the fungus Gliomastix sp. ZSDS1-F7-2 (see ESI†) from the marine sponge Phakellia fusca Thiele, which was collected from the South China Sea in March 2012. Herein, we report the isolation, structural elucidation, and biological activity of these unusual metabolites.

Fig. 1 Structures of compounds 1–5.

First, the strain Gliomastix sp. ZSDS1-F7-2 was incubated in 100 mL Erlenmeyer flask containing 25 mL of malt medium (20 g L⁻¹ of malt extract, 1 g L⁻¹ of peptone, and 30 g L⁻¹ of sea salt, 20 × 100 mL flasks) on a rotary shaker (180 rpm) at 25 °C to give the seed culture, which was transferred to the 1000 mL Erlenmeyer flask containing solid rice medium (200 g of rice, 200 g water, 6 g sea salt, 20 × 1 L flasks). After static incubation at 25 °C under daylight for 2 months, the culture was broken up with a spatula and extracted three times with 95% ethanol (EtOH). Then, the combined 95% EtOH crude extract was suspended in water and subsequently extracted with petroleum ether, ethyl acetate (EtOAc), and n-butanol. From the EtOAc extract, five novel macrolides, gliomasolides A–E (1–5), were isolated by serial column chromatography on silica gel, octadecyl silane (ODS), and preparative reversed-phase high-performance liquid chromatography (HPLC).

Compound 1 possessed a molecular formula C₁₈H₃₂O₄ as determined by HRCIMS ion peaks at m/z 313.2383 [M + H]⁺ and 295.2268 [M − H₂O + H]⁺ (calcd for C₁₈H₃₃O₄ and C₁₈H₃₁O₃, m/z 313.2379 and 295.2273, respectively), indicative of three double-bond equivalents (DBEs). ¹³C NMR spectrum confirmed the presence of 18 carbon atoms, which by HSQC experiment were classified into one methyl, eleven methylenes, three oxygenated methines, two olefinic carbons, and one carbonyl carbon (Table S1 in ESI†). Corresponding to the ¹³C NMR spectrum, resonances of three oxymethines at δ_H 4.31 (1H, ddd, J = 12.0, 8.0, 5.0 Hz, H-4), 3.64 (1H, m, H-7), 4.99 (1H, m, H-13), and two olefinic protons at δ_H 5.96 (1H, dd, J = 16.0, 1.0 Hz, H-2), 6.81 (1H, dd, J = 16.0, 8.0 Hz, H-3) were clearly observed in the low field region of the ¹H NMR spectrum of 1 (Table S1 in ESI†). Consideration of two DBEs occupied by one olefin and one carbonyl group, compound 1 has to be cyclized to meet the three DBEs required by molecular formula. ¹H–¹H COSY spectrum showing sequential cross peaks of from H-4 to H-7, together with the correlations of from H-4 to C-1 (δ_C 168.0)/C-2 (δ_C 123.3)/C-5 (δ_C 32.4)/C-6 (δ_C 31.3), and H-7 to C-5/C-9 (δ_C 24.9) in the HMBC spectrum (Fig. S1 in ESI†), clearly arranged hydroxyl groups at both C-4 and C-7 of 1. Further analyses of ¹H–¹H COSY, HSQC, and HMBC spectra (Fig. S1 in ESI†) established the structure of compound 1, being a 14-membered macrolide (Fig. 1). Specifically, the carbonyl carbon at C-1, conjugated with an olefinic double bond (Δ^2,3), was linked to C-13 via an oxygen atom to form a macrocyclic skeleton by the diagnostic correlations of from all H-2/H-3/H-13 to C-1 in the HMBC spectrum of 1. The obvious upfield shift for oxymethine CH-13 (δ_H/C 4.99/77.3) compared to CH-4/CH-7 (δ_H/C 4.31/72.7, 3.64/70.9, respectively), supported as well the formation of lactone between C-1 and C-13. The geometry of the double bond in 1 was determined to be E by the large coupling constant of H-2/H-3 (J_2,3 = 16.0 Hz). Fortunately, the single crystal of 1, suitable for X-ray diffraction analysis, was successfully obtained, corroborating the assignment of chemical structure of 1 (Fig. S2 in ESI†). Given the excellent absolute structure parameter [0.0(2)], the absolute stereochemistry of 1 was unambiguously identified as shown in Fig. 1. Compound 1, therefore, was completely elucidated and named gliomasolide A.

The ¹H, ¹³C NMR spectra of compound 2 were similar to those of 1 with the primary difference being the additional presence of signals attributable to both a hydroxyl group and a monosaccharide in 2 (Table S1 in ESI†), which was consistent with the HRCIMS showing ion peaks [M-C₆H₁₁O₅-H]⁻ at m/z 326.2092, (calcd for C₁₈H₃₀O₅, m/z 326.2093), and [M-C₆H₁₁O₅-H₂O-H]⁻ at m/z 308.2001, (calcd for C₁₈H₂₈O₄, m/z 308.1988), indicative of the molecular formula C₂₄H₄₂O₁₀ for 2. Detailed analyses of 2D NMR spectra (Fig. S1 in ESI†), confirmed the macrolide core of 2, being similar to that of 1, and suggested the monosaccharide was glucose with the α-anomeric structure determined by the coupling constant of H-1′/H-2′ (J_1′2′ = 4.0 Hz). Specifically, the additional presence of a hydroxyl group at C-5 of 2 was discerned by the ¹H–¹H COSY cross peak from H-4 (δ_H 5.05) to H-5 (δ_H 4.23) in conjunction with the HMBC correlations of from H-4 to both C-3 (δ_C 149.8) and C-5 (δ_C 72.5) (Fig. S1 in ESI†). Furthermore, the key HMBC correlation of from H-1′ (δ_H 5.72) to C-7 (δ_C 78.7) suggested the connection of glucose and macrolide moiety at C-7, establishing the planar structure of 2 (Fig. 1). The absolute configurations at C-4/C-7/C-13 of macrolide core in 2 were supposed to be the same as those of 1 based on the biogenetical viewpoint. This assumption, the glucose configuration, as well as the unambiguous structure elucidation of 2, were finally confirmed by the X-ray single crystal diffraction analysis (Fig. S3 in ESI†). Therefore, 2, being an unusual glycoside, was ascribed a trivial name gliomasolide B.

Compound 3, with the molecular formula C₁₈H₃₂O₆, was revealed by the HRCIMS showing ion peaks at m/z 326.2089 [M − H₂O]⁻, (calcd for C₁₈H₃₀O₅ at m/z 326.2093). The ¹H, ¹³C NMR spectra of 3 suggested structural similarity between 3 and 1, i.e., the α,β-conjugated carbonyl C-1 (δ_C 168.4), E configuration double bond Δ^2,3 (δ_H/C 6.18/123.2, δ_H/C 6.94/147.5, for C-2, C-3, respectively, and J_2,3 = 16.0 Hz) (Table S1 in ESI†), assuming the macrolide structure for 3, like 1. Further detailed analyses of 2D NMR spectra confirmed this assumption and established the planar structure of 3 (Fig. S1 in ESI†). In detail, the ¹H–¹H COSY spectrum displayed continuously sequential correlations of from H-2 to H-8, together with the correlations of H-3 with both C-4 (δ_C 73.6) and C-5 (δ_C 77.3), H-4 (δ_H 4.59) with C-2/C-3/C-5/C-6 (δ_C 71.1), and H-5 (δ_H 3.99) with C-7 (δ_C 71.2) in the HMBC spectrum, clearly indicated the arrangement of four continuous hydroxyl groups at C-4/C-5/C-6/C-7, respectively. Pleasingly, after trying a variety of solvents, we successfully obtained the single crystal of 3 from pyridine. The result of the X-ray analysis, consistent with that revealed in the NMR analyses, unequivocally assigned the 14-membered macrocyclic structure to 3 with the absolute configuration of all stereogenic centers determined as shown in Fig. 1 based on the absolute structure parameter being 0.09(19). Compound 3, therefore, was elucidated and given the trivial name gliomasolide C. Of particularly note, the molecular crystal of 3 exhibited a disorder in the part from C-1 to C-4 (Fig. 2), resulted from the rotation of the single bonds of both C1–C2 and C3–C4, indicative of the presence of two stable conformers in the solid state of 3, which revealed the high conformational flexibility of this unique macrolide. This observation agreed well with the NOESY spectrum by which the determination of 3 adopting at least two stable conformations in the CD₃OD solution could be deduced, evident from the presence of cross peaks from both H-2 and H-3 to all protons H-4/H-5/H-6/H-7. Given the coplanar, while trans relationship between the two olefinic protons H-2 and H-3 in 3, the observation of simultaneous correlations from both of them with several other same protons within one conformer might be impossible.

Fig. 2 Perspective drawings of 3 from X-ray crystallographic data.

To our surprise, similar to 3, the NOESY spectrum of each other macrolide reported here displayed as well the correlations from both olefinic protons H-2 and H-3 to several other same protons within the molecule (Fig. S9/15/21/27/33 in ESI†), suggesting more than one stable conformers coexisted in the solution of each of them, indicative of this type of unusual macrolide naturally has great conformational flexibility which might give it unique features in binding to the biological targets.

Compound 4, a white amorphous powder, was suggested to be an isomer of 3 in view of their closely similar ¹H, ¹³C NMR spectra (Tables S1 and S2 in ESI†) and the same molecular formula C₁₈H₃₂O₆ determined by the HRCIMS displaying ion peak at m/z 343.2103 [M − H]⁻ and 326.2089 [M − H₂O]⁻, (calcd for C₁₈H₃₁O₆ and C₁₈H₃₀O₅ at m/z 343.2121, 326.2093, respectively). Compared with 3, a key difference readily observed was the peak pattern arising from H₃-18 in the ¹H NMR spectrum, being doublet in 4versus triplet in 3, suggesting one hydroxyl group in the macrolide core of 3 was shifted to C-17 to furnish 4. The arrangement of the hydroxyl group at C-6 in 3 shifted, was secured by the well-resolved consecutive cross peaks from H-2 to H-7 in the ¹H–¹H COSY spectrum, together with the diagnostic HMBC correlations of from H-6 (δ_H 1.65, 1.83) to C-4 (δ_C 75.9)/C-5 (δ_C 72.7)/C-7 (δ_C 69.3) (Fig. S1 in ESI†). Furthermore, the assignment of a hydroxyl group at C-17 was clearly identified by the ¹H–¹H COSY correlation of from H-18 (δ_H 1.14) to H-17 (δ_H 3.70), along with the correlations from H-18 to C-17 (δ_C 68.4), and in turn, in the HMBC spectrum of 4. Detailed analyses of 2D NMR spectra involving ¹H–¹H COSY, HSQC, and HMBC (Fig. S1 in ESI†), unambiguously corroborated the planar structure of 4, being an unusual fourteen-membered macrolide (Fig. 1). The stereochemistry at C-4/C-5/C-7/C-13 in 4 was putatively assigned to be identical with that of 3 by the biogenetical viewpoint, which was supported as well by the comparison of the NOESY spectra of between 3 and 4, showing closely similar cross peaks (Fig. S21/27 in ESI†). The E olefinic double bond was assigned by the coupling constant of H-2/H-3 (J_2,3 = 16.0 Hz), while the absolute configuration at C-17 remained unresolved due partly to the limited amount obtained, which seriously challenged the successful application of Mosher's method. Compound 4, therefore, was elucidated and given the trivial name gliomasolide D.

Compound 5 was a white amorphous powder, its molecular formula C₁₈H₃₂O₅ was assigned by the HRCIMS displaying ion peaks at m/z 328.2248 [M]⁻ and 310.2136 [M − H₂O]⁻ (calcd for C₁₈H₃₂O₅ and C₁₈H₃₀O₄ at m/z 328.2250, 310.2144, respectively), indicative of containing one less hydroxyl group than 3. Albeit the similar ¹H, ¹³C NMR spectra, the obvious difference between 5 and 3 in the ¹H NMR spectrum, ascribing to the peak pattern of H-3 (dd in 3versus ddd in 5), was clearly observed, indicating the oxymethine CH-4 in 3, adjacent to the olefin functionality, was replaced by a methylene moiety in 5, which was supported as well by the chemical shifts of CH₂-4 (δ_H 2.44, 2.54/δ_C 42.1) (Table S2 in ESI†), and HMBC correlations of from H-4 to C-2 (δ_C 125.7)/C-3 (δ_C 146.9) in 5. Further extensive analyses of 2D NMR involving ¹H–¹H COSY, HSQC, and HMBC (Fig. S1 in ESI†), confirmed the fourteen-membered macrolide structure for 5, with all C-5, C-7, and C-9 being hydroxylated, similar to those of 3. Particularly, the assignment of one hydroxyl group linked at C-9, revealed by the consecutive cross peaks of H-7 (δ_H 4.02)/H-8b (δ_H 1.53)/H-9 (δ_H 3.77) in the ¹H–¹H COSY spectrum, was supported as well by the HMBC correlations of from H-9 to C-7 (δ_C 68.3)/C-11 (δ_C 21.8), H-7/H-10 (δ_H 1.40, 1.46)/H-11 (δ_H 1.20, 1.58) to C-9 (δ_C 69.2). The stereochemistry (2E/5R/7R/13R) of 5, being identical with that of 3, was assumed by the biogenetic viewpoint and the proton coupling constant of J_2,3 (16.0 Hz). Determination of stereochemistry at C-9 of 5 was not straightforward, the J-based configuration analysis was envisioned to resolve this problem, but ending in failure, due in part to this conformation-flexible macrolide adopting at least two stable conformers in the solution, just like 3, which seriously reduced the efficacy of this methodology. Currently, the stereochemistry at C-9 remains unresolved. Compound 5, therefore, was given the trivial name gliomasolide E.

The cytotoxicity of all the isolated compounds against HeLa (human epithelial carcinoma cell line) cells was evaluated using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) method, revealing that only 1 exhibited moderate inhibitory effect with an IC₅₀ value of 10.1 μM. Further biological evaluations of these metabolites using various screening assays are going on in our laboratory.

In conclusion, our investigation discovered five novel 14-membered macrolides from a sponge-derived fungus Gliomastix sp. ZSDS1-F7-2. Structurally, gliomasolides A–E, lacking methyl groups on the macrocyclic skeleton, are very rare with only two similar relatives being sch725674 and gloeosporone,⁶ both of which are popular synthetic targets.⁷ Of particularly note, in view of most macrolide drugs approved are glycosides: examples involving clarithromycin, erythromycin, and telithromycin etc.,^4,8 gliomasolide B (2), being the first glycoside of this type of 14-membered macrolide, might be very interesting as a template for the design and development of drugs. Gliomasolide C (3), featuring the presence of four consecutive cis hydroxyl groups in the macrolide backbone, is structurally unique, likely providing a target of interest for realizing a total synthesis. The discovery of five structurally novel macrocyclic lactones, gliomasolides A–E, is of great interest as well since it is the first report of chemistry from the marine-derived fungus of the genus Gliomastix. Not only does our investigation extend the structural varieties of known macrolides, it also demonstrates the tremendous potential of marine fungi as a prolific source of novel secondary metabolites.

Acknowledgements

The authors thank Dr Herman H. Y. Sung and Prof. I. D. Williams (Department of Chemistry, School of Science, Hong Kong University of Science and Technology) for measurements of X-ray crystallography and Mr Rui Feng (Division of Life Science, School of Science, Hong Kong University of Science and Technology) for performing NMR experiments. This research was financially supported by the General projects of the National Natural Science Foundation of China (Grant nos 21172094, 41376155, 21372100, 31201028, U1301131), a key program of the technology office in Guangzhou (Grant no. 2013000000163), and the National High Technology Research and Development Program (“863” Program) of China (Grant no. 2013AA092902). The authors also acknowledge the support from the Biotechnology Research Institute at the Hong Kong University of Science and Technology.

Notes and references

1. J. W. Blunt, B. R. Copp, R. A. Keyzers, M. H. G. Munro and M. R. Prinsep, Nat. Prod. Rep., 2015, 32, 116–211 and the previous reviews in this series.
2. (a) T. Teruya, H. Sasaki, K. Kitamura, T. Nakayama and K. Suenaga, Org. Lett., 2009, 11, 2421–2424; (b) K. Kito, R. Ookura, S. Yoshida, M. Namikoshi, T. Ooi and T. Kusumi, Org. Lett., 2008, 10, 225–228; (c) P. A. Wender and K. E. Longcore, Org. Lett., 2009, 11, 5474–5477; (d) A. R. Pereira, Z. Cao, N. Engene, I. E. Soria-Mercado, T. F. Murray and W. H. Gerwick, Org. Lett., 2010, 12, 4490–4493; (e) Y. Inahashi, M. Iwatsuki, A. Ishiyama, A. Matsumoto, T. Hirose, J. Oshita, T. Sunazuka, W. Panbangred, Y. Takahashi, M. Kaiser, K. Otoguro and S. Ōmura, Org. Lett., 2015, 17, 864–867.
3. (a) M. Hangyou, H. Ishiyama, Y. Takahashi and J. Kobayashi, Org. Lett., 2009, 11, 5046–5049; (b) T. I. Lazarova, S. M. Binet, N. H. Vo, J. S. Chen, L. T. Phan and Y. S. Or, Org. Lett., 2003, 5, 443–445; (c) R. Tello-Aburto, E. M. Johnson, C. K. Valdez and W. A. Maio, Org. Lett., 2012, 14, 2150–2153; (d) I. Paterson, S. J. Fink, L. Y. Lee, S. J. Atkinson and S. B. Blakey, Org. Lett., 2013, 15, 3118–3121; (e) Y. Tanabe, E. Sato, N. Nakajima, A. Ohkubo, O. Ohno and K. Suenaga, Org. Lett., 2014, 16, 2858–2861.
4. (a) P. B. Fernandes, D. J. Hardy, D. McDaniel, C. W. Hanson and R. N. Swanson, Antimicrob. Agents Chemother., 1989, 33, 1531–1534; (b) M. A. Boivin, M. C. Carey and H. Levy, Pharmacotherapy, 2003, 23, 5–8.
5. (a) W. J. Xu, X. J. Liao, S. H. Xu, J. Z. Diao, B. Du, X. L. Zhou and S. S. Pan, Org. Lett., 2008, 10, 4569–4572; (b) J. Wang, X. Wei, X. Qin, X. Lin, X. Zhou, S. Liao, B. Yang, J. Liu, Z. Tu and Y. Liu, Org. Lett., 2015, 17, 656–659.
6. (a) S. W. Yang, T. M. Chan, J. Terracciano, D. Loebenberg, M. Patel and M. Chu, J. Antibiot., 2005, 58, 535–538; (b) W. L. Meyer, Helv. Chim. Acta, 1987, 70, 281–291.
7. (a) J. D. Moretti, X. Wang and D. P. Curran, J. Am. Chem. Soc., 2012, 134, 7963–7970; (b) A. K. Bali, S. K. Sunnam and K. R. Prasad, Org. Lett., 2014, 16, 4001–4003; (c) J. D. Trenkle and T. F. Jamison, Angew. Chem., Int. Ed., 2009, 48, 5366–5368; (d) N. R. Curtis, A. B. Holmes, M. G. Looney, N. D. Pearson and G. C. Slim, Tetrahedron Lett., 1991, 32, 537–540; (e) A. Sharma, S. Gamre and S. Chattopadhyay, Lett. Org. Chem., 2005, 2, 547–549; (f) S. V. Ley, E. Cleator, J. Harter and C. J. Hollowood, Org. Biomol. Chem., 2003, 1, 3263–3264.
8. N. Scheinfeld, J. Drugs Dermatol., 2004, 3, 409–413.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details, tables of NMR data, additional figures, crystallographic data files (CIF), as well as spectra of all novel compounds 1–5 (¹H NMR, ¹³C NMR, 2D NMR, and HRMS). CCDC 1057054-1057056. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c5ra08559d

---