Arthproliferins A–D, Four New Sesterterpenes from the Mangrove-Sediment-Derived Fungus Arthrinium sp. SCSIO41221

Four new sesterterpenes, arthproliferins A–D (1–4), together with four known derivatives, were isolated and characterized from the mangrove-sediment-derived fungus Arthrinium sp. SCSIO41221. Their structures were determined using detailed nuclear magnetic resonance (NMR) and mass spectroscopic (MS) analyses. Some of the isolated compounds were evaluated for their cytotoxicity in vitro. The results revealed that terpestacin (6) exhibited significant activity with an IC50 value of 20.3 μM, and compounds 2 and 5 were found to show weak inhibitory effects against U87MG-derived GSCs.


Introduction
The mangrove forests are a complex ecosystem, and they are characterized by periodic tides that result in highly variable salinity and nutrient availability [1].To adapt to the frequent and extreme environmental changes, an active microbial community, which constitutes the second-largest ecological group of the marine microbes, was formed under these special ecological conditions.Among the mangrove microbial community, mangrovesediment-derived microbes not only play an essential role in creating and maintaining this biosphere, but also represent a rich source of secondary metabolites with novel structures and significant pharmacological activities, thus attracting a great deal of interest from scientists [2].Several important reviews have shown that mangrove-associated fungi are the dominant producers of new natural products with pronounced biological activities [2,3].Among these natural products of marine origin, the isolation of terpenoids has been exhaustively reported [4,5].As a special subclass of terpenoids, marine-derived sesterterpenes with diverse skeletal types are of great attention due to their significant biological activities, such as antimicrobial, cytotoxic, anti-inflammatory, and protein tyrosine phosphatase B inhibitory activities [6][7][8].
Terpestacin, a bicyclic sesterterpene containing a transfused [3.0.13] bicyclic skeleton that includes a 15-membered macrocycle with three geometrically defined trisubstituted olefins, was first reported from the filamentous fungus Arthrinium sp.FA1744 (ATCC74132) in 1993 [9][10][11][12].Heretofore, no more than 30 similar sesterterpenes have been obtained in nature.Examples include fusaproliferin from Fusarium proliferatum [13], fusaprolifins A and B from Fusarium proliferatum MA-84 [14], and 11-Epiterpestacin from Bipolaris sorokiniana NSDR-011 [15].Although very few have been found, they have shown significant biological activities [16].Great interest in these sesterterpenes has come from the finding that terpestacin was shown to effectively inhibit the formation of syncytia, the angiogenesis, without affecting the endothelial cell viability and the extracellular signal-regulated kinase activity [17,18].Several total syntheses of terpestacin have been reported to date, and a series of active derivatives have also been designed and synthesized [17,19,20].These results demonstrate that terpestacin analogues for treatment as anticancer and anti-HIV agents might offer promising research outcomes, and they encourage us that more effort is worth spending in this prospective field.

Results
The rice solid culture of Arthrinium sp.SCSIO41221 was extracted using an EtOAc extract.Several chromatographic methods, including an MPLC with silica and an ODS gel column, as well as a semipreparative HPLC with a C-18 column, were used for the isolation of the eight terpestacin analogous sesterterpenes (Figure 1).(ATCC74132) in 1993 [9][10][11][12].Heretofore, no more than 30 similar sesterterpenes have been obtained in nature.Examples include fusaproliferin from Fusarium proliferatum [13], fusaprolifins A and B from Fusarium proliferatum MA-84 [14], and 11-Epiterpestacin from Bipolaris sorokiniana NSDR-011 [15].Although very few have been found, they have shown significant biological activities [16].Great interest in these sesterterpenes has come from the finding that terpestacin was shown to effectively inhibit the formation of syncytia, the angiogenesis, without affecting the endothelial cell viability and the extracellular signalregulated kinase activity [17,18].Several total syntheses of terpestacin have been reported to date, and a series of active derivatives have also been designed and synthesized [17,19,20].These results demonstrate that terpestacin analogues for treatment as anticancer and anti-HIV agents might offer promising research outcomes, and they encourage us that more effort is worth spending in this prospective field.

Results
The rice solid culture of Arthrinium sp.SCSIO41221 was extracted using an EtOAc extract.Several chromatographic methods, including an MPLC with silica and an ODS gel column, as well as a semipreparative HPLC with a C-18 column, were used for the isolation of the eight terpestacin analogous sesterterpenes (Figure 1).   1 and 2) of 1, with the aid of DEPT and HSQC (Figure S4 values of 15.9, 27.0, 13.6, 16.9, and 13.4), six methylenes (including an oxymethylene), three methines (including an oxymethine), three quaternary carbons (including a ketone carbon and an oxygenated quaternary carbon).All of these 1 H and 13 C NMR resonances accounted for five degrees of unsaturation.The remaining three degrees of unsaturation indicated that 1 is a tricyclic compound.These data showed great similarities to those of saponaroxin A [27], except for an oxygenated quaternary carbon and a disubstituted double bond instead of a trisubstituted double bond and a methylene in the 15-membered ring.This assumption was supported by the correlation of Me-21 to C-7, C-8, and C-9, of H-7 to C-5, C-8, and Me-21, and of H-6 to C-5 and C-8 in the HMBC spectrum (Figure 2).Thus, the gross structure of 1 was established.
Molecules 2023, 28, x FOR PEER REVIEW 3 of 11 formula was established as C25H36O4 and indicated eight degrees of unsaturation.The NMR data (Tables 1 and 2) of 1, with the aid of DEPT and HSQC (Figure S4 ), six methylenes (including an oxymethylene), three methines (including an oxymethine), three quaternary carbons (including a ketone carbon and an oxygenated quaternary carbon).All of these 1 H and 13 C NMR resonances accounted for five degrees of unsaturation.The remaining three degrees of unsaturation indicated that 1 is a tricyclic compound.These data showed great similarities to those of saponaroxin A [27], except for an oxygenated quaternary carbon and a disubstituted double bond instead of a trisubstituted double bond and a methylene in the 15-membered ring.This assumption was supported by the correlation of Me-21 to C-7, C-8, and C-9, of H-7 to C-5, C-8, and Me-21, and of H-6 to C-5 and C-8 in the HMBC spectrum (Figure 2).Thus, the gross structure of 1 was established.2.63, q (7.0, 14.0) 2.63, q (7.0, 16.0) 2.60, q (7.0, 14.0) 2.73, q (7.0, 14.0) 24 3.86, dd (7.0, 10.5) 3.86, dd (7.0, 10.5) 3.63, dd (7.0, 10.5) 3.87, dd (7.0, 10.5)  2.63, q (7.0, 16.0) 2.60, q (7.0, 14.0) 2.73, q (7.0, 14.0) 3.86, dd (7.0, 10.Compound 2 was isolated as a pale yellow oil.Based on the HRESIMS ion peak at an m/z of 441.2623 [M + Na] + (calcd for 441.2611), the molecular formula was established as C 25 H 38 O 5 and indicated seven degrees of unsaturation.Both the 1 H and 13 C NMR spectra of 2 showed a close similarity to those of 1 (Tables 1 and 2).However, a close comparison of the 13 C NMR spectroscopic data of 2 and 1 revealed some differences: one bisubstituted double bond in 1 was changed to an oxymethine and a methylene in 2. This assumption was supported by the correlation of Me-21 to C-7, C-8, and C-9, as well as of H-7 to C-5, C-6, C-8, and Me-21 in the HMBC spectrum (Figure 2).Thus, the planar structure of compound 2 was determined.
Compound 3 was isolated as a pale yellow amorphous solid and assigned the molecular formula of C 25 H 36 O 4 according to the following positive HRESIMS m/z value: 423.2511 [M + Na] + (calcd for 423.2506).Analysis of the 1 H and 13 C NMR data revealed that 3 (Tables 1 and 2) was found to contain an additional ketone carbon when compared with those of terpestacin [9][10][11].The key HMBC correlations of Me-22 to C-11, C-12, and C-13, as well as of H-9 to C-11 and H-10 to C-11, indicated the presence of an additional ketone carbonyl in C-11 (Figure 2); this was also indicated through comparison of the data with those that were in agreement with the data of terpestacin.
Compound 4 was isolated as a pale yellow amorphous solid.The HRESIMS of 4 established its molecular formula as C 25 H 38 O 5 , thus indicating seven degrees of unsaturation.The 1 H and 13 C NMR spectra of 4 (Tables 1 and 2) showed a close similarity to those of terpestacin, [9][10][11] with the exception that one trisubstituted double bond of the latter was replaced by an epoxy three-membered ring (δ C of 63.8 and 62.0, respectively), which was confirmed by the HRESIMS data.This assumption was supported by the correlation of Me-22 to C-11, C-12, and C-13, as well as of H-13 to C-12, C-14, and C-15 in the HMBC spectrum (Figure 2).Accordingly, the planar structure of 4 was constructed, as shown in Figure 1.
The similar relative configurations of compounds 1-4 were determined through analyses of the coupling constants and through NOESY experiments, as well as through a comparison with the data of terpestacin (6), which were originally determined using X-ray crystallography.A strong NOE was seen between Me-19 and H-14α, and the latter was not coupled with H-15, which are consistent with the literature concerning the trans relationship of 5-and 15-membered rings (Figure 3) [10,14,28].In the NOESY spectrum, the correlations of H-15 to H-3, H-13, and H-23, and of H-11 to H-13 revealed the β-configurations of these H atoms.The E-geometry of ∆3, ∆7, and ∆12 were deduced from the δ C values of Me-20, Me-21, and Me-22 (<20 ppm; Table 2) [10].The configuration of ∆6 in 1 was deduced as E as a result of using a large coupling constant of J 6,7 = 15.4Hz.The NOESY correlations of H-7 to H-13, H-21 to H-11 and H-7, and H-13 to H-15 in 2 deduced from H-7, H-21, H-11, and H-15 were on the same plane (Figure 3).In the NOESY spectrum of 4, the correlations of H-13 to H-15 and H-13 to H-11 deduced from H-11, H-13, and H-15 were on the same plane (Figure 3).The same relative configurations, similar to the CD spectrum and similar specific options of the terpestacin derivatives, together with the same biogenetic origin, indicated that they also shared an identical absolute configuration (Figure 4).tablished its molecular formula as C25H38O5, thus indicating seven degrees of unsaturation.The 1 H and 13 C NMR spectra of 4 (Tables 1 and 2) showed a close similarity to those of terpestacin, [9][10][11] with the exception that one trisubstituted double bond of the latter was replaced by an epoxy three-membered ring (δC of 63.8 and 62.0, respectively), which was confirmed by the HRESIMS data.This assumption was supported by the correlation of Me-22 to C-11, C-12, and C-13, as well as of H-13 to C-12, C-14, and C-15 in the HMBC spectrum (Figure 2).Accordingly, the planar structure of 4 was constructed, as shown in Figure 1.
The similar relative configurations of compounds 1−4 were determined through analyses of the coupling constants and through NOESY experiments, as well as through a comparison with the data of terpestacin (6), which were originally determined using Xray crystallography.A strong NOE was seen between Me-19 and H-14α, and the latter was not coupled with H-15, which are consistent with the literature concerning the trans relationship of 5-and 15-membered rings (Figure 3) [10,14,28].In the NOESY spectrum, the correlations of H-15 to H-3, H-13, and H-23, and of H-11 to H-13 revealed the β-configurations of these H atoms.The E-geometry of ∆3, ∆7, and ∆12 were deduced from the δC values of Me-20, Me-21, and Me-22 (<20 ppm; Table 2) [10].The configuration of ∆6 in 1 was deduced as E as a result of using a large coupling constant of J6,7 = 15.4Hz.The NOESY correlations of H-7 to H-13, H-21 to H-11 and H-7, and H-13 to H-15 in 2 deduced from H-7, H-21, H-11, and H-15 were on the same plane (Figure 3).In the NOESY spectrum of 4, the correlations of H-13 to H-15 and H-13 to H-11 deduced from H-11, H-13, and H-15 were on the same plane (Figure 3).The same relative configurations, similar to the CD spectrum and similar specific options of the terpestacin derivatives, together with the same biogenetic origin, indicated that they also shared an identical absolute configuration (Figure 4).Compound 4 was isolated as a pale yellow amorphous solid.The HRESIMS of 4 established its molecular formula as C25H38O5, thus indicating seven degrees of unsaturation.The 1 H and 13 C NMR spectra of 4 (Tables 1 and 2) showed a close similarity to those of terpestacin, [9][10][11] with the exception that one trisubstituted double bond of the latter was replaced by an epoxy three-membered ring (δC of 63.8 and 62.0, respectively), which was confirmed by the HRESIMS data.This assumption was supported by the correlation of Me-22 to C-11, C-12, and C-13, as well as of H-13 to C-12, C-14, and C-15 in the HMBC spectrum (Figure 2).Accordingly, the planar structure of 4 was constructed, as shown in Figure 1.
The similar relative configurations of compounds 1−4 were determined through analyses of the coupling constants and through NOESY experiments, as well as through a comparison with the data of terpestacin (6), which were originally determined using Xray crystallography.A strong NOE was seen between Me-19 and H-14α, and the latter was not coupled with H-15, which are consistent with the literature concerning the trans relationship of 5-and 15-membered rings (Figure 3) [10,14,28].In the NOESY spectrum, the correlations of H-15 to H-3, H-13, and H-23, and of H-11 to H-13 revealed the β-configurations of these H atoms.The E-geometry of ∆3, ∆7, and ∆12 were deduced from the δC values of Me-20, Me-21, and Me-22 (<20 ppm; Table 2) [10].The configuration of ∆6 in 1 was deduced as E as a result of using a large coupling constant of J6,7 = 15.4Hz.The NOESY correlations of H-7 to H-13, H-21 to H-11 and H-7, and H-13 to H-15 in 2 deduced from H-7, H-21, H-11, and H-15 were on the same plane (Figure 3).In the NOESY spectrum of 4, the correlations of H-13 to H-15 and H-13 to H-11 deduced from H-11, H-13, and H-15 were on the same plane (Figure 3).The same relative configurations, similar to the CD spectrum and similar specific options of the terpestacin derivatives, together with the same biogenetic origin, indicated that they also shared an identical absolute configuration (Figure 4).By comparing the 1 H, 13 C-NMR, and MS data with the literature values, the known compounds were identified as 21-hydroxyterpestacin (5) [28], terpestacin (6) [9][10][11], fusaproliferin (7) [13], and saponaroxin A (8) [27].Based on our isolated terpestacin, seventeen derivatives with various L-amino acid side chains were designed and synthesized.Their anticancer activities against U87MG-derived glioblastoma stem cells (GSCs) were evaluated, and two derivatives showed stronger activities related to their further development as anticancer agents for the treatment of GBM by targeting the GSCs [17].
All of the compounds were evaluated for their cytotoxic activities against the human cancer cells MDAMB-231, C4-2B, MGC803, MDA-MB-468, and A549 using the CCK-8 method, but they exhibited no inhibitory activities at 5 µM.The primary screening of compounds 2, 5, and 6 regarding the proliferation of U87MG-derived GSCs was carried out using an ATP-monitoring luminescence assay [17].The results revealed that terpestacin (6) exhibited significant activity, with an IC 50 value of 20.3 µM, compared to the positive control A1938, with an IC 50 value of 10.9 µM.Additionally, compounds 2 and 5 were found to show weak inhibitory effects.Furthermore, the levels of hypoxia-inducible Factor-1α (HIF-1α) and CD133 protein expression levels were evaluated using in vitro experiments after the treatment of the compounds 2, 5, and 6.Compared to terpestacin and the positive control 4-200, compound 5 showed stronger activity (Figure S26).Because of the diverse biological activities that have been reported regarding terpestacin [29,30], molecular docking studies of these isolated compounds were carried out in silico with HIF-1α (PDB: 3KCX) and UQCRB (PDB: 3BCC) active sites to further understand the binding mode between those derivatives with the protein.As a result, compounds 2, 5, and 6 appeared to interact with the HIF-1α protein with docking scores of −4.284, −5.875, and −4.565, respectively, as well as with the UQCRB protein with docking scores of −5.820, −8.023, and −7.425, respectively, according to the above experiment data (Figure 5).Although the in vitro tests of 7 were limited by the small amounts, the similar interactions in the 2D binding models of the HIF-1α and UQCRB-with better docking scores of −5.875 and −8.886, respectivelysuggested that 7 may act as a potential anticancer agent against U87MG-derived GSCs like its analogs.Compound 7 interacted with the HIF-1α active site mainly through hydrogen bonds between the Gln14 of the protein 3KCX, as well as with the ester groups and hydroxy.The OH-17 of 7 formed hydrogen bonds with the Ala296 of protein 3BCC.These results provide valuable information for the further development of anticancer agents.

General Experimental Procedures
Optical rotation values were achieved using an Anton Paar MCP-500 polarimeter

General Experimental Procedures
Optical rotation values were achieved using an Anton Paar MCP-500 polarimeter (Anton, Austria).The one-dimensional and two-dimensional (2D) nuclear magnetic resonance (NMR) spectra were obtained using a Bruker Avance spectrometer (AC 500 or AVANCE III HD 700 NMR) with tetramethylsilane (TMS) as an internal standard.High-resolution electrospray ionization mass spectrometry (HR-ESI-MS) data were measured using a Bruker microTOF-QII mass spectrometer (Bruker, Fällanden, Switzerland).CD spectra were measured using a Chirascan circular dichroism spectrometer (Applied Photophysics).YMC gel (ODS-A, 12 nm, S-50 µm) and silica gel (200-300 mesh) (Qingdao Marine Chemical Factory) were used for column chromatography.The silica gel GF254 (Qingdao Marine Chemical Factory, Qingdao, China) was used for TLC.A semipreparative HPLC was performed using a Hitachi L-2400 (diode array detector, Hitachi L-2455, Tokyo, Japan) using a YMC ODS column (YMC-pack ODS-A, 10 × 250 mm, 5 mm, YMC Co. Ltd., Kyoto, Japan).All solvents used were of analytical grade (Tianjin Fuyu Chemical and Industry Factory).Spots were detected on TLC under UV light or by heating after spraying with 5% H 2 SO 4 in EtOH (v/v).

Fungal Material
The culture of Arthrinium sp.SCSIO41221 was isolated from a mangrove sediment sample collected in Sanya (18 • 13 50.2N, 109 • 37 15.8 E) in August 2010.The internal transcribed spacer (ITS) sequences of SCSIO41221 shared a similarity of 99% to that of Arthrinium sp.LH11 (GenBank accession no.HQ832842.1).On the basis of its molecular biological protocol and morphological analyses, strain SCSIO 41221 was identified as Arthrinium sp. and designated as Arthrinium sp.SCSIO 41221.The strain was stored on PDA slants at 4 • C and deposited in the RNAM Center.

Extraction and Isolation
The fungus was inoculated in PDA agar medium (infusion of 200 g of potato, 20 g of dextrose, 2.5 g of NaCl, and 1000 mL of distilled water) as the seed medium and incubated at 25 • C on a rotating shaker (180 rpm) for 2 days.Then, 10 mL seed solution were inoculated in rice medium (200 g of rice, 0.5 g of NaCl, and 200 mL of distilled water) in 1000 mL Erlenmeyer flasks.The mass fermentation of this fungus was carried out at 25 • C under static conditions for 50 days.After 50 days, the fermentation was cut into small pieces, sonicated for 20 min, and soaked in acetone three times.The acetone extract was concentrated under reduced pressure to afford an aqueous solution, and then the aqueous solution was extracted using EtOAc to gain 51 g of a crude gum.

Bioassay
Cytotoxicity was assayed using the CCK-8 method, which was previously described [31].Five human cancer cell lines (human breast cancer cells MDAMB-231, human breast cancer cells C4-2B, human gastric cancer cells MGC803, human breast cancer cells MDA-MB-468, and human breast cancer cells A549) were used in the cytotoxic activity assay.These cell lines were obtained from the National Infrastructure of Cell Line Resource (NICR) and Shanghai Cell Bank through the Chinese Academy of Sciences.Taxol was used as a positive control.
A preliminary screening of compounds 1-8 on the proliferation of U87MG-derived GSCs was determined using an ATP-monitoring luminescence assay as previously described [17].A1938 was used as a positive control.The abilities of compounds 2, 5, and 6 to influence the expression of HIF-1α and CD133 were also investigated as previously described [17].U87MG was obtained from the Korean Cell Line Bank (Seoul, Korea).Statistical analyses were performed using SPSS 9.0 software.

Molecular Docking Analysis
The molecular docking was conducted using the Schrödinger 2017-1 suite as described previously [32].The crystal structures of human HIF-1α with Clioquinol inhibitor (PDB ID: 3KCX) and UQCRB (PDB ID: 3BCC) were collected from the protein data bank (http://www.pdb.org,accessed on 15 September 2023).The structures of these docking compounds were generated in ChemBio3D Ultra 20.0, followed by an MM2 calculation to minimize the conformation energy.The 3D structures of the binding models were generated using PyMol molecular graphics software (Maestro Version 11.1.011,MMshare Version 3.7.011,Release 2017-1, Platform Windows-x64) (Schrödinger 2017-1, Schrödinger Inc., New York, NY, USA).The XYZ coordinates of the protein 3KCK with the ligand were −20.95, 25.69, and 7.97, respectively.The XYZ coordinates of the protein 3BCC with the ligand were 8.78, 128.61, and 83.01, respectively.

Conclusions
In summary, the chemical investigation of the marine mangrove-sediment-derived fungus Arthrinium sp.SCSIO41221 led to the isolation of four new sesterterpenesarthproliferins A-D (1-4)-and four known derivatives-21-hydroxyterpestacin (5), terpestacin (6), fusaproliferin (7), and saponaroxin A (8). Their planar structures and absolute configurations were elucidated using detailed NMR, MS spectroscopic analyses, and com-parison of the coupling constants, as well as through NOESY experiments and determining the ECD spectra with terpestacin, which was determined using X-ray crystallography.Terpestacin has been proven to display attractive anti-HIV and anticancer activities [11,30,33].It can also inhibit angiogenesis both in vitro and in vivo via the downregulation of the ROS/HIF1α/VEGF pathway [17,30].All of these isolated terpestacin derivatives were evaluated with respect to their cytotoxic activities.Compounds 2 and 5 were found to show weak inhibitory effects against U87MG-derived GSCs compared to terpestacin, which exhibited significant activity with an IC 50 value of 20.3 µM.Further related enzymatic bioassays were evaluated, and compound 5 showed stronger abilities than terpestacin in HIF-1α and CD133 protein degradation.Molecular docking with the HIF-1α and UQCRB proteins was also performed to understand the inhibitory activity.

Figure 1 .
Figure 1.Structures of compounds 1-8.Compound 1 was isolated as a pale yellow amorphous solid.Based on the HRESIMS ion peak determined at an m/z of 423.2516 [M + Na] + (calcd for 423.2506), the molecular

Figure 1 .
Figure 1.Structures of compounds 1-8.Compound 1 was isolated as a pale yellow amorphous solid.Based on the HRESIMS ion peak determined at an m/z of 423.2516 [M + Na] + (calcd for 423.2506), the molecular formula was established as C 25 H 36 O 4 and indicated eight degrees of unsaturation.The NMR data (Tables1 and 2) of 1, with the aid of DEPT and HSQC (FigureS4) experiments, revealed resonances that were suggestive of a disubstituted double bond (δ H 5.69 for
a indicates recorded at 700 MHz in CD 3 OD.b indicates recorded at 500 MHz in CD 3 OD.c indicates recorded at 700 MHz in DMSO-d6.