New Furan and Cyclopentenone Derivatives from the Sponge-Associated Fungus Hypocrea Koningii PF04

Two new furan derivatives, hypofurans A and B (1 and 2), and three new cyclopentenone derivatives, hypocrenones A–C (3–5), along with seven known compounds (6–12), were isolated from a marine fungus Hypocrea koningii PF04 associated with the sponge Phakellia fusca. Among them, compounds 10 and 11 were obtained for the first time as natural products. The planar structures of compounds 1–5 were elucidated by analysis of their spectroscopic data. Meanwhile, the absolute configuration of 1 was determined as 2R,3R by the comparison of the experimental and calculated electronic circular dichroism (ECD) spectra. All the isolates were evaluated for their antibacterial and antioxidant activity. Compounds 1, 10, and 12 all showed modest antibacterial activity against Staphylococcus aureus ATCC25923 (MIC, 32 μg/mL). In addition, compounds 1, 10 and 11 exhibited moderate DPPH radical scavenging capacity with IC50 values of 27.4, 16.8, and 61.7 µg/mL, respectively.


Introduction
Marine fungi harbor the potential to generate a multitude of structurally novel chemicals with diverse biological activities in part owing to harsh habitats [1,2]. More than one thousand new natural products have hitherto been harvested from marine-derived fungi. Excitingly, some of them with clinically relevant pharmacological activities will be probably developed into viable drug candidates exemplified by halimide, which entered into phase II clinical trials of cancer chemotherapy [3][4][5][6].
As an epitome of particular marine habitats, sponges often possess remarkable microbial biomass and diversity, and their microbial symbionts represent a precious wellspring of new scaffolds for drug discovery [7]. Hence, we focused on the sponge-derived fungus Hypocrea koningii PF04, which was associated with the marine sponge Phakellia fusca collected from Yongxing Island in the South China Sea [8].

Structure Elucidation
Hypofuran A (1) was afforded as a colorless oil. The molecular formula C10H14O4 was deduced from HRESIMS data (m/z 199.0965 for [M + H] + ), indicative of four degrees of unsaturation. The IR spectrum of 1 showed characteristic absorptions for a hydroxy group (3386 cm −1 ) and a furan ring (3123, 1584, 1522, 888 cm −1 ) [19]. The 1 H NMR spectrum (Table 1) showed a 2,5-disubstituted furan ring at δH 6.42 (d, H-5) and 6.24 (d, H-6), one acetal methine at δH 5.89 (s, H-1), an exchangeable proton at δH 5.23 (t, OH-8), one hydroxymethyl group at δH 4.37 (d, H-8), two oxygenated methines at δH 3.70 (m, H-3) and 3.67 (m, H-2), and two methyls at δH 1.27 (d, H-10) and 1.22 (d, H-9). The 13 C NMR and DEPT data ( Table 2) displayed 10 carbons, including two olefinic quaternary carbons, five methines (two olefinic ones), one oxymethylene, and two methyls. The hydroxymethyl group was located at C-7 (δC, 155.9) in the furan ring on the basis of HMBC correlations from H2-8 to C-6 and C-7. The hydroxymethylfuran ring accounted for three of four degrees of unsaturation, thus requiring an additional ring for 1. The remaining ring was assigned as a dioxolane ring from the HMBC correlations of H-1/C-2 and C-3 and COSY correlation of H-2/H-3. A closer examination of the COSY spectrum implied that two methyls were tethered to C-2 and C-3 of the dioxolane ring, respectively, which was confirmed by correlations of H-2/H3-10 and H-3/H3-9. Further HMBC correlations of H-1/C-4 The relative stereochemistry of 1 was disclosed by a NOESY experiment ( Figure 2). Two intense NOE interactions between H-1/H3-9 and H-1/H-2 suggested that H-1, H-2, and H3-9 were in the same orientation of the dioxolane ring while H-3 and H3-10 were cofacial. Thus, the relative configurations at C-2 and C-3 of 1 were assigned as 2S,3S or 2R,3R, respectively. In order to assign the absolute configuration of 1, we chose to compare its experimental electronic circular dichroism (ECD) spectrum with the correspondingly time-dependent density functional theory (TDDFT) calculated one, which has proved to be a powerful and reliable approach for determining the absolute configuration of natural products [20,21]. Six low energy conformers above 1% population were generated using conventional initial Merck Molecular Force Field (MMFF) and DFT geometry optimization method ( Figure 3 and Supplementary Information), for which ECD spectra were calculated with the TZVP basis set and four different functionals (B3LYP, BH&HLYP, PBE0, CAM-B3LYP). The experimental ECD spectrum (MeOH) of 1 showed a strong negative Cotton effect around 218 nm. With all four methods, the mirror image curve of the Boltzmann-averaged ECD spectrum of (2S,3S)-1 was in good accordance with the experimental one, thus suggesting a (2R,3R) absolute configuration ( Figure 4).   Hypofuran B (2) was furnished as a yellow powder. The HRESIMS data (m/z 267.0634 for [M + Na] + ) supported a molecular formula of C14H12O4. The IR absorption bands at 3404 cm −1 (a hydroxy group), together with 3020, 1614, 1516, 905 cm −1 (a furan ring), implying 2 possessed a hydroxymethylfuran ring [19]. The assumption was further supported by comparison of its corresponding NMR spectra of 1. The HMBC correlations of H-8 (δH, 7.40)/C-7 (δC, 137.7) and H-14 (δH, 9.63)/C-7 and C-8 (δC, 136.0) revealed that a formyl group (C-14, δC, 193.4) was linked to a vinyl group (C-7-C-8), constructing an acrylaldehyde moiety. Furthermore, the remaining characteristic signals at δH 9.61 (s, 4-OH), 6.99 (d, 8.5 Hz, 2H), and 6.82 (d, 8.5 Hz, 2H) were assignable to a para-hydroxyphenyl ring, established by the HMBC correlations of H-2 and H-6/C-4. The connectivity of the hydroxymethylfuran ring, acrylaldehyde moiety and para-hydroxyphenyl ring were accomplished by analysis of the HMBC spectrum. The crucial cross-peak of H-8/C-10 (δC, 117.0) clearly suggested that the hydroxymethylfuran ring was connected to the acrylaldehyde moiety, which was attached to the para-hydroxyphenyl ring based on the correlations of H-14/C-1 (δC, 123.5) as well as H-2 and H-6/C-7 in the HMBC spectrum. The geometry of the double bond (C-7-C-8) was determined as E by the NOESY correlation between H-14 and H-8 ( Figure 2). According to the aforementioned information, the structure of 2 was unambiguously assigned as depicted in Figure 1.  Table 2). The COSY cross-peak of H-4/H-5 in conjunction with HMBC correlations of H-2/C-1, C-3, C-4, and C-5, H-4/C-3 and C-8, and H2-5/C-1, C-3, and C-4 delineated a cyclopentenone moiety. Furthermore, the COSY cross-peaks of H2-6/H2-7 and H2-9/H2-10 suggested the presence of two spin systems, which were connected through an ester bond based on the HMBC correlations from H2-6, H2-7, and H2-9 to C-8 ( Figure 2). Finally, C-6 was adjacent to the cyclopentenone moiety, as evident by the HMBC cross peaks of H-2 and H-4/C-6 and H-6/C-3, while C-10 was linked to the benzene ring, as indicated by the HMBC correlations of H-9 and H-10/C-11 and H-10/C-16, thereby establishing the structure of 3.

Biological Activity
All the isolates (1-12) were evaluated for antibacterial activities against Gram-positive Staphylococcus aureus ATCC25923 and methicillin-resistant Staphylococcus aureus (MRSA) ATCC43300 as well as Gram-negative Escherichia coli ATCC25922. The MIC values for the compounds 1-12 were no less than 64 μg/mL against the tested pathogens, except for 1, 10 and 11. These three compounds selectively inhibited the growth of S. aureus with the same MIC value of 32 μg/mL. Meanwhile, Compounds 1-12 were also tested for antioxidant activities using DPPH radical scavenging assay. The results (Table 3) showed that compounds 1, 10 and 11 had moderate DPPH radical scavenging activities with IC50 values of 27.4 ± 7.4, 16.8 ± 4.3 and 61.7 ± 3.3 µg/mL, respectively, and the rest of the compounds did not show DPPH radical scavenging capacities (IC50 > 128 µg/mL).

Fungal Material
The fungus H. koningii PF04 was isolated from the tissue of the sponge P. fusca collected from Yongxing Island in the South China Sea. The fungus was identified by its rDNA amplification and sequence analysis of the ITS region (GenBank accession no. FJ941853). A voucher strain was deposited at the School of Life Sciences and Biotechnology, Shanghai Jiao Tong University, Shanghai, China.

Fermentation
The strain was initially grown on PDA medium in a Petri dish for 7 days. A single colony was inoculated into seed medium (potato 200 g, dextrose 20 g, sea water 1000 mL) in 250 mL Erlenmeyer flasks on a rotatory shaker (180 rpm) at 25 °C for 48 h. Subsequently, the large scale fermentation was carried out in 50 × 250 mL Erlenmeyer flasks (80 g of rice and 120 mL of sea water), each of which was inoculated with seed medium (10 mL). The fungus PF04 was cultured under static conditions at 25 °C for 40 days.

Extraction and Isolation
The fermented material was extracted with acetone (3 × 5 L). The organic solvent was concentrated under reduced pressure and partitioned with EtOAc (1.5 L) and H2O (1.5 L) to yield the EtOAc extract (24.5 g). The extract was subjected to vacuum liquid chromatography (VLC) on silica gel column

Antibacterial Activity Assay
The antibacterial activities were evaluated against three different bacteria (S. aureus ATCC25923, methicillin-resistant S. aureus (MRSA) ATCC43300 and E. coli ATCC25922) in 96-well microplates as described by Correa with some modifications [33]. The final concentrations of each compound in the wells were 256, 128, 64, 32, 16, 8, 4 and 2 μg/mL. Chloramphenicol was used as positive control. Each assay was carried out in triplicates.

DPPH Radical Scavenging Activity Assay
The DPPH free radical scavenging assay was performed by a modified method [34]. Briefly, the tested compounds were dissolved in methanol to serial concentrations. Each well in 96-well microplates containing 100 μL of sample solution and 100 μL of DPPH solution (methanol, 0.2 mM) was incubated at 37 °C for 30 min. Ascorbic acid was used as positive control. The blank control experiment was added methanol without any dissolved compound. The absorbance was measured at 517 nm on a microplate reader. DPPH scavenging rate (%) = (1 − absorbance of compound/absorbance of control) × 100. All experiments were taken in triplicates. The IC50 value (the concentration of a compound to scavenge 50% of DPPH radicals) was calculated from nonlinear regression analysis using the GraphPad Prism software 5.0 (GraphPad Software, San Diego, USA). Results are presented as the means ± SD.

Theoretical ECD Calculations
All calculations were performed with the Gaussian 09 program using various functionals (B3LYP, BH&HLYP, PBE0, CAM-B3LYP) and TZVP basis set. See Supplementary Information for more details of the DFT calculation.

Conclusions
Chemical investigation of the sponge-associated fungus Hypocrea koningii PF04 has resulted in the isolation and characterization of five new compounds, hypofurans A and B (1 and 2) and hypocrenones A-C (3-5), along with seven known secondary metabolites (6-12) encompassing two new natural products (10 and 11), representing a paradigm of chemical diversity. Structurally, hypofurans A was a furan derivative featuring a dioxolane ring where the absolute configuration was ascertained by comparing its experimental and calculated ECD spectra. With respect to the biogenetic relationships of all the new compounds, the plausible avenues furnishing them were postulated. In a small panel of antibacterial assays, compounds 1, 10 and 12 displayed modest inhibitory activities against S. aureus ATCC25923. In addition, compounds 1, 10 and 11 exhibited moderate DPPH radical scavenging activity. Collectively, this article showcased marine fungi served as new bioactive secondary metabolite producers.