One-dimensional and two-dimensional (2D) nuclear magnetic resonance (NMR) spectra were measured on a Bruker Avance 700 MHz NMR spectrometer (Fällanden, Switzerland) with Tetramethylsilane as an internal standard. High-resolution electrospray ionization mass spectrometry (HRESIMS) data were recorded on a maXis Q-TOF mass spectrometer in positive ion mode (Bruker, Fällanden, Switzerland). Electronic circular dichroism (ECD) and ultraviolet (UV) spectra were measured with a Chirascan circular dichroism spectrometer (Applied Photophysics). Optical rotations were measured using an MCP-500 polarimeter (Anton, Austria). High-performance liquid chromatography (HPLC) was performed on Hitachi Primaide with YMC ODS SERIES column (YMC-Pack ODS-A, YMC Co. Ltd., Kyoto, 250 × 10 mm I.D., S-5 μm, 12 nm). Column chromatography was carried out on silica gel (200–300 mesh, Jiangyou Silica Gel Development Co., Yantai, China), YMC Gel ODS-A (12 nm, S-50 μm YMC, MA, United States), and Sephadex LH-20 (40–70 μm, Amersham Pharmacia Biotech AB, Uppsala, Sweden). Spots were detected under UV light by heating after spraying with the mixed solvent of saturated vanillin and 5% sulfuric acid in water. The thin layer chromatography plates with silica gel GF254 (0.4–0.5 mm, Qingdao Marine Chemical Factory, Qingdao, China) were used for analysis and preparation.

The strain SCSIO06868 was isolated from the deep-sea sediment collected from the Indian Ocean (94°37.377′E; 2°59.853′S; depth 4,762 m). The internal transcribed spacer sequences of SCSIO06868 (494 base pairs, GenBank accession no. MZ277624) have 99% sequence identity to that of Penicillium citrinum DUCC5728 (GenBank accession no. 582768). Then, it was designated as a member of Penicillium sp. and named as Penicillium sp. SCSIO06868. The strain SCSIO06868 was stored on methylene blue agar (malt extract 15 g, agar 16 g, sea salt 10 g, water 1 L, pH 7.4–7.8) slants at 4°C and deposited at Key Laboratory of Tropical Marine Bio-resources and Ecology, Chinese Academy of Sciences.

The mass fermentation of this fungus was carried out in 1-L Erlenmeyer flasks. The fungus was inoculated in a liquid medium (2% maltose, 2% mannitol, 1% monosodium glutamate, 1% glucose, 0.3% yeast extract, 0.05% monopotassium phosphate, 0.03% MgSO₄⋅7H₂O, and 300-ml tap water/flask, 93 flasks, 28 L total) at 25°C under static condition for 35 days. After 35 days, the fermentation was soaked in ethyl acetate (500 ml/flask), and the mycelia were cut into small pieces and sonicated for 20 min. The ethyl acetate solution was concentrated under reduced pressure to gain a brown crude extract (59.0 g).

The crude extract was subjected to silica gel column chromatography, which was eluted with dichloromethane and methanol (MeOH) mixed solvent in a step gradient (100:0–5:1, v/v) and separated into seven fractions (Fr-1–Fr-7). Fr-1 (3.2 g) was applied to Sephadex LH-20 column eluted with MeOH, reversed-phase C₁₈ medium pressure liquid chromatography (MPLC) eluted with MeOH/water (H₂O) (10:90–100:0, v/v), and semipreparative HPLC [72% CH₃OH/H₂O with 0.3‰ trifluoroacetic acid (TFA), 2 ml/min] to afford compounds 7 (2.2 mg, t_R = 20.2 min) and 8 (18.4 mg, t_R = 22.2 min). Fr-2 (6.3 g) was subjected to Sephadex LH-20 column eluted with MeOH, reversed-phase C₁₈ MPLC eluted with MeOH/H₂O (10:90–100:0, v/v), and semipreparative HPLC (2 ml/min) to gain compounds 3 (4.0 mg, 54% CH₃CN/H₂O, t_R = 31.0 min), 5 (4.5 mg, 35% CH₃CN/H₂O with 0.3‰ TFA, t_R = 33.6 min), 6 (7.6 mg, 35% CH₃CN/H₂O with 0.3‰ TFA, t_R = 35.0 min), and 10 (6.4 mg, 33% CH₃CN/H₂O, t_R = 22.0 min). Fr-3 (4.6 g) was purified with Sephadex LH-20 column eluted with MeOH, reversed-phase C₁₈ MPLC eluted with MeOH/H₂O (10:90–100:0, v/v), and semipreparative HPLC (2 ml/min) to obtain compounds 4 (2.9 mg, 72% CH₃OH/H₂O with 0.3‰ TFA, t_R = 8.2 min) and 9 (42.2 mg, 70% CH₃OH/H₂O with 0.3‰ TFA, t_R = 14.6 min). Fr-4 (3.7 g) was purified with Sephadex LH-20 column eluted with MeOH, reversed-phase C₁₈ MPLC eluted with MeOH/H₂O (10:90–100:0, v/v), and semipreparative HPLC (75% CH₃CN/H₂O with 0.3‰ TFA, 2 ml/min) to yield compound 1 (30.5 mg, t_R = 12.2 min). Fr-5 (1.3 g) was applied to Sephadex LH-20 column eluted with MeOH, reversed-phase C-18 MPLC eluted with MeOH/H₂O (10:90–100:0, v/v), and semipreparative HPLC (2 ml/min) to get compounds 11 (76.5 mg, 40% CH₃CN/H₂O with 0.3‰ TFA, t_R = 11.4 min) and 12 (60% CH₃CN/H₂O, t_R = 15.2 min, 6.0 mg). Fr-6 (1.6 g) was subjected to Sephadex LH-20 column eluted with MeOH, reversed-phase C₁₈ MPLC eluted with MeOH/H₂O (10:90–100:0, v/v), and semipreparative HPLC (75% CH₃CN/H₂O with 0.3‰ TFA, 2 ml/min) to obtain compound 2 (2.5 mg, t_R = 12.8 min).

Penicillenol G1 (1): Pale white solid; [α]D25 –156.6 (c 0.10, MeOH); UV (MeOH) λ_max (log ε) 219 (3.21), and 278 (2.79) nm; ECD (1.06 mM, MeOH) λ_max (Δε) 209 (+ 5.26), 227 (–5.27), and 281 (−4.38) nm; ¹H and ¹³C NMR data (Table 1); HRESIMS m/z 284.1866 [M + H]⁺ (calcd for C₁₅H₂₆NO₄, 284.1856).

Penicillenol G2 (2): Yellowish oil; [α]D25 + 46.0 (c 0.10, MeOH); UV (MeOH) λ_max (log ε) 242 (2.87) and 279 (3.12) nm; ECD (1.06 mM, MeOH) λ_max (Δε) 213 (–3.08), 238 (+ 2.50), and 271 (2.89) nm; ¹H and ¹³C NMR data (Table 1); HRESIMS m/z 284.1862 [M + H]⁺ (calcd for C₁₅H₂₆NO₄, 284.1856).

Penicillenol H (3): Yellow oil; [α]D25 –21.0 (c 0.10, MeOH); UV (MeOH) λ_max (log ε) 227 (2.81) and 285 (3.04) nm; ECD (0.71 mM, MeOH) λ_max (Δε) 213 (2.87), 231 (–2.76), and 289 (–1.27) nm; ¹H and ¹³C NMR data (Table 1); HRESIMS m/z 284.1858 [M + H]⁺ (calcd for C₁₅H₂₆NO₄, 284.1856).

Coniochaetone N (4): Yellow solid powder; [α]D25 + 46.4 (c 0.10, MeOH); UV (MeOH) λ_max (log ε) 203 (4.11), 226 (4.15), 243 (4.22), and 342 (3.56) nm; ¹H and ¹³C NMR data (Table 2); HRESIMS m/z 285.0373 [M + Na]⁺ (calcd for C₁₃H₁₀NaO₆, 285.0370).

X-Ray crystallographic analysis of penicillenol G1 (1): Moiety formula: C₁₅H₂₅NO₄ (M = 283.36 g/mol), colorless needle, crystal size = 0.6 × 0.03 × 0.03 mm³, trigonal, space group C2; unit cell dimensions: a = 20.4710(5) Å, b = 4.85100(10) Å, c = 33.3706(10) Å, V = 3154.94(15) ų, Z = 8, ρ_calcd = 1.193 g cm^–3, T = 101(2) K, μ(Cu Kα) = 0.698 mm^–1. A total of 31,816 reflections were measured with 6,225 independent reflections (R_int = 0.0496, R_sigma = 0.0341). Final R indices [I > 2σ (I)]: R₁ = 0.0348, wR₂ = 0.0872. Final R indexes [all date]: R₁ = 0.0407, wR₂ = 0.0896, Flack parameter = 0.07(8). Largest diff. peak and hole = 0.18 and –0.20 eÅ^–3.

The molecular docking was conducted by AutoDockTools (Version 1.5.6) (Morris et al., 2008). The crystal structure of SARS-CoV-2 main protease (PDB ID: 6LU7) was retrieved from the Protein DataBank¹ (Jin et al., 2020). The structures were generated in ChemBio3D Ultra 14.0 (ChemBioOffice version 14.0), followed by an MM2 calculation to minimize the conformation energy. The original ligand and crystal water were removed before the docking calculation. The hydrogens were added to the structure of 6LU7, and Kollman united partial charges were assigned. A Lamarckian genetic algorithm was applied as a default search algorithm and set the grid box within the size of 46 × 44 × 46 Å, with the spacing of 0.375 Å. During the docking, the default parameters were used if it was not mentioned. The docking pose that had the lowest binding energy was represented as the most favorable binding conformation.

All compounds (1–12) were tested for antibacterial activities against five pathogenic bacteria using the method of agar filter paper diffusion. Compounds that had inhibition zone were evaluated in 96-well plates using a modification of the broth microdilution method (Pang et al., 2018). Ampicillin and gentamicin were used as a positive control for Gram-positive and Gram-negative bacteria, respectively.

The antiviral activities of all compounds (1–12) against SARS-CoV-2 M^pro were evaluated through the method mentioned in the previous report (Li et al., 2020). Hydroxychloroquine showed potent inhibitory activity against SARS-CoV-2 M^pro with Ki = 0.36 μm and was used as a positive control.

Penicillenol G1 (1) possessed the elemental composition of C₁₅H₂₅NO₄ with 4 degrees of unsaturation as established by its ¹³C NMR data and a protonated molecule at m/z 284.1866 in the HRESIMS spectrum. Its 1D NMR data (Table 1) displayed three methyls [δ_C/H 20.4/1.29 (d, J = 7.0 Hz, CH₃-7), 17.3/1.16 (d, J = 6.3 Hz, CH₃-16), and 14.4/0.89 (t, J = 7.7 Hz, CH₃-15)], three sp³ methines [(δ_C/H 69.0/3.73 (brs, CH-5), 68.0/4.10 (dq, J = 6.3, 2.8 Hz, CH-6), and 37.5/3.58–3.67 (m, CH-9)], four sp³ methenes (δ_C/H 28.3–34.8/1.23–1.74), and four sp² non-protonated carbons (δ_C 178.0 C-2, 102.3 C-3, 196.2 C-4, and 193.8 C-8). The ¹H-¹H COSY correlations (Figure 2) of H₃-16/H-9/H₂-10/H₂-11 and H₃-15/H₂-14, along with four overlapping sp³ methenes in the ¹H NMR, indicated the presence of a 2-isooctyl group. The ¹H-¹H COSY correlations of H₃-7/H-6/H-5 verified that there was a 1-hydroxyethyl group directly connected to C-5. Comparison of the NMR data of 1 with those of penicillenol A₁ (5) (Lin et al., 2008; Yoda et al., 2010) showed that they only differed by an absence of the singlet methyl in 1. The planar structure of 1 was further confirmed by its heteronuclear multiple bond correlation (HMBC) correlations (Figure 2) of H-6 to C-4, H₃-7 to C-5, H₂-10 to C-8 and C-12, and H₃-16 to C-8 and C-10. The configuration of a double bond at C-3 was determined as Z based on that the chemical shift of acylamino (δ_C 178.0, C-2) was in a lower field than those of normal ones, which caused by the hydrogen bond between the oxygen atom at C-2 and hydroxy at C-8 (Aoki et al., 2000). Thus, the ECD of 1 (Figure 3) displayed a positive Cotton effect at 209 nm (Δε = + 18.39), a negative Cotton effect at 227 nm (Δε = –15.33), and a negative Cotton effect at 281 nm (Δε = –4.38), and the trend of which was consistent with that of 5. Thus, the absolute configuration of C-5 in 1 was determined as S. Mosher’s method was tried to confirm the absolute configuration of C-6 but failed. Fortunately, the single crystal of 1 was obtained, and the absolute configuration of 1 was established as 5S, 6R, 9S by analyzing the X-ray crystallographic data (Figure 4). Compound 1 was named as penicillenol G1.

The molecular formula of penicillenol G2 (2), which was the same as 1, was established as C₁₅H₂₅NO₄ by its NMR data (Table 1) and a protonated molecule at m/z 284.1862 in the HRESIMS. Its NMR data were nearly the same as those of 1, and only the chemical shifts of C-5 (δ_C/H 68.1/3.94, brs), C-6 [δ_C/H 68.7/4.08 (qd, J = 6.3, 3.5 Hz)], and C-7 [δ_C/H 17.0/1.10 (d, J = 6.3 Hz)] in 2 have some small differences. Compound 2 had the same planar structure with 1, which was confirmed by the ¹H-¹H COSY and HMBC spectra (Figure 2). Thereby, the differences mentioned earlier might be caused by their different configurations. The ECD spectrum of 2 showed a negative Cotton effect at 213 nm (Δε = –3.08), a positive Cotton effect at 238 nm (Δε = + 2.5), and a positive Cotton effect at 271 nm (Δε = + 2.89), which exhibited a reverse trend with 1 and 5, but was consistent with penicillenol A2 (6) (Lin et al., 2008; Sengoku et al., 2012). Therefore, the absolute configuration of C-5 in 2 was determined as R. Some synthetic chemists demonstrated that the natural penicillenol A1 (5) (Yoda et al., 2010), penicillenol A2 (6) (Sengoku et al., 2012), and penicillenol C1 (7) (Kempf et al., 2013) were all have the same absolute configurations of 6R and 9S, and the absolute configuration of compound 1 was also determined as 6R and 9S by the X-ray diffraction study. Thus, considering the same biosynthetic pathway (Yin et al., 2019) and comparison of the spectroscopic data with 1, the absolute configurations of C-6 and C-9 in 2 were deduced as 6R, 9S.

Penicillenol H (3) was obtained as a yellow oil. The ¹³C NMR data and a protonated molecule at m/z 284.1858 in the HRESIMS of 3 suggested that its molecular formula was C₁₅H₂₅NO₄ with 4 degrees of unsaturation. The NMR data (Table 1) of 3 were similar to those of 1, except that the 1-hydroxyethyl group in 1 was displaced by oxygenated methylene (δ_C/H 59.3/3.94, qd, J = 12.6, 2.8 Hz, CH₂-6) and an N-methyl group (δ_C/H 26.8/3.03, s, CH₃-17) was added in 3. The extinction was determined by its ¹H-¹H COSY cross-peak of H₂-6/H-5 and HMBC correlations of H₂-6 to C-4 and H₃-17 to C-2 and C-5. The planar structure of 3 was further established by its 2D NMR (Figure 2). The configuration of a double bond at C-3 was designated as Z by the chemical shift of C-2 (δ_C 178.0) (Aoki et al., 2000). The absolute configuration of C-5 and C-9 in 3 were determined as S and R same as compound 1 by their similar ECD spectrum (Figure 3), of which 3 showed a positive Cotton effect at 213 nm (Δε = + 2.87), two negative Cotton effects at 231 nm (Δε = – 2.76), and 289 nm (Δε = – 1.27), respectively. Compound 3 was named penicillenol H.

The molecular formula of compound 4 was C₁₃H₁₀O₈, established by its ¹³C NMR data (Table 2) and a sodium adduct ion peak at m/z 285.0373 in the HRESIMS spectrum. Its ¹H NMR data were simple and showed two aromatic protons (δ_H 7.54, brs, CH-7; 7.32, brs, CH-9) at meta-position, an oxygenated methine (δ_H 5.30, d, J = 7.0 Hz, CH-1), and two sp³ methylenes (δ_H 2.45–2.53, m, 2.00, brt, J = 11.2 Hz, CH₂-2; 3.19, dt, J = 17.5, 7.7 Hz, 2.89, ddd, J = 18.2, 9.1, 2.8 Hz, CH₂-3). Besides the corresponding carbons (δ_C 109.7, CH-7; 113.1, CH-9; 71.5, CH-1; 34.4, CH₂-2; 30.7, CH₂-3), there were eight sp² non-protonated carbons in its ¹³C NMR, which indicated the presences of an α,β-unsaturated ketone (δ_C 182.2, C-12; 122.9, C-13; 176.1, C-4) and a carboxyl (δ_C, 167.9, C-14). Its NMR data were very similar to those of coniochaetone L (Guo et al., 2019), except that the methoxy group at C-1 in coniochaetone L was replaced by hydrogen in 4. The speculation was further confirmed by its HMBC and ¹H-¹H COSY spectra (Figure 2). The positive specific rotation of 4 ([α]D25 + 46.4, MeOH) that was consistent with that of coniochaetone L ([α]D25 + 25.3, MeOH) suggested that the configuration of C-1 was R. Thus, compound 4 was established as R-1,8-dihydroxy-9-oxo-1,2,3,9- tetrahydrocyclopenta[b]chromene-6-carboxylic acid and named as coniochaetone N.

In addition, the eight known compounds (5–12) (Figure 1) were identified as penicillenol A1 (5) (Lin et al., 2008; Yoda et al., 2010), penicillenol A2 (6) (Lin et al., 2008; Sengoku et al., 2012), penicillenol C1 (7) (Lin et al., 2008; Kempf et al., 2013), penicillenol C2 (8) Lin et al., 2008), scalusamide C (9) (Tsuda et al., 2005), (E)-7-(3-methyl-4-oxo-6,7,8,8a-tetrahydro-4H-pyrrolo[2,1-b][1,3]oxazin-2-yl)hept-2-enoic acid (10) (Lai et al., 2013), terretrione D (11) (Shaala and Youssef, 2015), and (2R)-2,3-dihydro-7-hydroxy-6,8-dimethyl-2-[(E)-prop-1-enyl] chromen-4-one (12) (Li et al., 2007; Zhang et al., 2019) by comparison of their physical and spectroscopic data with those in the literature.

All isolated compounds (1–12) were tested for their antiviral activities against SARS-CoV-2 M^pro in vitro through the method mentioned in the reported literature (Li et al., 2020). Compound 3 showed weak inhibitory activity against M^pro enzyme, which was responsible for the COVID-19 pandemic. When treated with 50 μm of 3, the relative enzyme activity of SARS-CoV-2 M^pro was 46.64%, and that of the positive control hydroxychloroquine was 5.28% with the same concentration. To better understand the interactions between compounds and SARS-CoV-2 M^pro, molecular docking research was performed to mimic the interactions between compound 3 and M^pro enzyme of SARS-CoV-2 (PDB ID: 6LU7) by utilizing the AutoDockTools. Molecular docking results demonstrated that compound 3 could interact with the SARS-CoV-2 M^pro enzyme at the entrance of the catalytic pocket, with the calculated binding affinities of –4.98 kcal/mol. The 2D binding model for 3 (Figure 5) showed two hydrogen bonds and two intermolecular hydrophobic interactions. Two hydrogen bonds were formed between the carbonyl group at C-4 and Thr-26, as well as between the hydroxyl group at C-6 and Thr-24. The lengths of the two hydrogen bonds were 2.0 and 1.8 Å, respectively. These results suggested that compound 3 could insert into the active site of the enzyme and bind tightly to the catalytic amino acid residues by different types of interactions to inhibit SARS-CoV-2 M^pro.

Compounds 1–12 were evaluated their antibacterial activities against five pathogenic bacteria Escherichia coli (ATCC 25922), Enterococcus faecalis (ATCC 29212), Klebsiella pneumonia (ATCC 13883), Staphyloccocus aureus (ATCC 29213), and methicillin-resistant S. aureus (MRSA). Compounds 1, 3, 5, and 6 with 50 μg/disc showed inhibition zones against S. aureus. Compounds 1, 3, and 5 with 50 μg/disc showed an inhibition zone against MRSA (Supplementary Figure 1). Furthermore, their minimum inhibitory concentrations (MICs) were tested, and the results are shown in Table 3. Compound 3 displayed potent inhibitory activities against S. aureus and MRSA with MIC values of both 2.5 μg/ml. Ampicillin was used as a positive control against S. aureus and MRSA with MIC values of 1.56 and 0.39 μg/ml, respectively.