Optical rotations were measured at 25°C using a JASCO P-1020 digital polarimeter (JASCO Ltd., Tokyo, Japan). The UV spectra were determined on a Techcomp UV2310II spectrophotometer (Techcomp, Ltd., Shanghai, China). Electronic circular dichroism (ECD) spectra were acquired with a JASCO J-815 CD spectrometer (JASCO Ltd., Tokyo, Japan) at 25°C. The NMR spectra were recorded on a DD2 NMR spectrometer (Agilent Technologies, Santa Clara, CA, USA; 500 MHz for ¹H and 125 MHz for ¹³C) and a JNM-ECP NMR spectrometer (JEOL, Japan; 600 MHz for ¹H and 150 MHz for ¹³C) using tetramethylsilane (TMS) as an internal standard, and CDCl₃ as solvent. The Electrospray ionization mass spectrometry (ESIMS) was taken on a Micromass Q-TOF spectrometer (Waters, Ltd., Milford, Massachusetts, USA), and high-resolution ESIMS spectra were collected on a Thermo Scientific LTQ Orbitrap XL spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Semipreparative HPLC was conducted on a Waters e2695 separation system (Milford, MA, USA), equipped with a Waters 2998 photodiode array detector and a Waters X-Bridge C₁₈ (5 μm, 10 × 250 mm) preparative column, and the flow rate was 2 mL/min. Column chromatography (CC) was performed using Silica gel (100–200, 200–300 mesh; Qingdao Marine Chemical Inc., Qingdao, China), Sephadex LH-20 (GE Healthcare, Pittsburgh, PA, USA), and octadecylsilyl silica gel (ODS) (40–63 μm, merck, MA, USA). Precoated silica gel plates (Yantai Zi fu Chemical Group Co., Yan Tai, China; GF254) were used for thin layer chromatography (TLC) analysis. Spots were detected by UV light (254 nm) and colored by spraying heated silica gel plates with 12% H₂SO₄ in H₂O containing saturated vanillin.

The studied strain was isolated from an unidentified mangrove sample collected from coastal marine habitats of the South China Sea, Haikou, China, in May 2017. This strain was identified as P. sclerotiorum HY5 by amplifying and analyzing its internal transcribed spacer (ITS) sequence, as well as morphological features. The ITS sequence of this fungus was deposited at the GenBank database (accession number MG827186), and the isolated strain was preserved in the Marine Agriculture Research Center, Tobacco Research Institute of Chinese Academy of Agricultural Sciences, Qingdao, China.

The purified fungus was grown on PDA plates for 5 days at 28°C. The mycelial disc were inoculated into 500 mL Erlenmeyer flasks containing 200 mL of potato dextrose water (PDW) medium with 3% salinity and shaken on a rotary shaker (175 rpm) at 28°C for 3 days to obtain seed culture. Aliquots (5 mL) of the seed culture were transferred into 1,000 mL Erlenmeyer flasks containing 400 mL of production media, consisting of PDW medium with 3% salinity for 30 days at 28°C, and the fermentation scale was 80 L in total. After fermentation, the culture broth was filtered to separate the culture media and mycelia. The culture broth was extracted three times with equal volumes of EtOAc. The mycelia were mechanically broken, and then, extracted ultrasonically twice with a mixture (1:1, v/v) of dichloromethane (CH₂Cl₂) and methanol (MeOH), and concentrated in vacuo to obtain an aqueous solution, which was further extracted thrice with equivalent EtOAc. The culture broth and mycelia extracts were combined and evaporated under reduced pressure to yield EtOAc extract (74.61 g), which was then subjected to vacuum liquid chromatography (VLC) on silica gel using a step gradient elution of EtOAc–petroleum ether from 0 to 100% and 10 to 50% MeOH–EtOAc to afford six fractions (Fr.1–Fr.6) based on their TLC profiles. The Fr.2 was chromatographed repeatedly on silica gel CC eluting with mixtures of EtOAc–petroleum ether (10–50%, v/v) to give three subfractions (Fr.21–Fr.23). The Fr.23 was first fractionated via reverse silica gel CC eluting with a gradient MeOH–H₂O (from 50:50 to 100:0, v/v), and then, separated by Sephadex LH-20 CC eluting with CH₂Cl₂-MeOH (1:1, v/v) to obtain two subfractions (Fr.231–Fr.232). The Fr.231 was then chromatographed repeatedly with a gradient elution of EtOAc–petroleum ether (from 0 to 100%, v/v) and MeOH–EtOAc (from 10 to 50%, v/v) to give four subfractions (Fr.2311–Fr.2314). The Fr.2311 was applied to semipreparative HPLC separation using an isocratic elution of MeOH–H₂O with 0.1% trifluoroacetic acid (TFA) (75% MeOH in H₂O, v/v) to afford compound 7 (119.8 mg). The Fr.2313 was subjected to semipreparative HPLC using an isocratic elution of 70% MeCN in H₂O to yield compound 1 (26.8 mg). Following the same procedures, Fr.3 was also subjected to silica gel CC using a step gradient elution of EtOAc–petroleum ether (0 to 100%, v/v) and MeOH–EtOAc (10 to 50%, v/v) to give two subfractions (Fr.31–Fr.32). The Fr.31 was separated by octadecylsilyl (ODS) CC (MeOH–H₂O; from 50:50 to 100:0, v/v), and then, subjected to Sephadex LH-20 CC eluting with CH₂Cl₂-MeOH (1:1, v/v) to obtain two subfractions (Fr.311–Fr.312). The Fr.311 was further purified by using semipreparative HPLC with isocratic MeOH–H₂O (75:25, v/v) as mobile phase to yield compound 5 (166.3 mg). The Fr.312 was subjected to semipreparative HPLC using MeOH in H₂O (25%) to obtain compounds 2 (28.4 mg) and 3 (28.1 mg). The Fr.32 was eluted using a MeOH–H₂O gradient system (from 50:50 to 100:0, v/v) and sequentially subjected to Sephadex LH-20 CC (CH₂Cl₂-MeOH, 1:1, v/v) to give three subfractions (Fr.321–Fr.323). The Fr.322 was applied to semi-preparative HPLC (60% MeCN in H₂O) to afford compound 4 (164 mg). The Fr.323 was purified by semipreparative HPLC (75% MeOH in H₂O) to generate compound 6 (63 mg). In order to achieve the requirements of the NMR test, the purities of all the isolated compounds were >95% based on the peak area normalization methods.

Isochromophilone H (1): yellow amorphous powder; [α]D25.0 −15.9 (c, 0.46, MeOH); UV(MeOH) λ_max (log ε) 202 (3.38), 248 (3.23), and 387 (3.35) nm; ECD (c 1.08 mM, MeOH) λ_max (Δε) 223 (+0.34), 257 (−0.51), 312 (+1.18), 385 (−0.51) nm; ¹H, and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 463.1892 [M+H] ⁺(calculated for C₂₅H₃₁O₆Cl, 463.1882).

Sclerotiorin A (2): yellow amorphous powder; [α]D25.0 +3.1 (c, 0.37, MeOH); UV(MeOH) λ_max (log ε) 200 (3.46), 249 (3.34), and 390 (3.47) nm; ECD (c 1.23 mM, MeOH) λ_max (Δε) 238 (+0.84), 258 (−0.94), 312 (+3.82), and 387 (−1.13) nm; ¹H and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 405.1832 [M+H] ⁺(calculated for C₂₃H₂₉O₄Cl, 405.1827).

Sclerotiorin B (3): yellow amorphous powder; [α]D25.0 +28.8 (c, 0.34, MeOH); UV(MeOH) λ_max (log ε) 201 (3.72), 250 (3.67), and 392 (3.82) nm; ECD (c, 0.62 mM, MeOH) λ_max (Δε) 235 (+1.84), 259 (−1.97), 313 (+8.85), and 381 (−2.37) nm; ¹H and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 405.1835 [M+H] ⁺(calculated for C₂₃H₂₉O₄Cl, 405.1827).

Ochlephilone (4): orange amorphous powder; [α]D25.0 +195.3 (c, 0.54, MeOH); UV (MeOH) λ_max (log ε) 204 (3.89), 250 (4.04), 395 (4.25) nm; ECD (c 1.30 mM, MeOH) λ_max (Δε) 239 (+2.84), 267 (−0.56), 310 (+6.11), and 340 (+7.14) nm; ¹H and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 383.1856 [M+H] ⁺(calculated for C₂₃H₂₆O₅, 383.1853).

Isochromophilone IV (5): yellow amorphous powder; [α]D25.0 −69.1 (c 0.44, MeOH); UV(MeOH) λ_max (log ε) 200 (3.33), 265 (2.98), 388 (3.83) nm; ECD (c 0.63 mM, MeOH) λ_max (Δε) 257 (−5.03), 285 (+1.04), 325 (+0.32), 383 (−7.82) nm; ¹H and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 395.1627 [M+H] ⁺(calculated for C₂₁H₂₇O₅Cl, 395.1620).

Isochromophilone J (6): yellow amorphous powder; [α]D25.0 +113.3 (c, 0.15, MeOH); UV(MeOH) λ_max (log ε) 208 (3.38), 265 (3.15), and 390 (3.94) nm; ECD (c 1.42 mM, MeOH) λ_max (Δε) 214 (−1.87), 255 (+3.47), 321(−0.34), and 388 (+1.88) nm; ¹H and ¹³C NMR data (Tables 1, 2); HRESIMS m/z 353.1524 [M+H] ⁺(calculated for C₁₉H₂₅O₄Cl, 353.1514).

Phytotoxicity was evaluated by seed germination methods. The bioassay experiments were performed on representative weeds in farmlands, including two types of grass [wild oat (Avena fatua L.), ryegrass (Lolium perenne L.)] and two broadleaf species [redroot amaranth (A. retroflexus L.), and velvetleaf (Abutilon theophrasti Medikus)], based on previously reported assay methods with some modifications (Travaini et al., 2016; Adetunji et al., 2018).

The tested seeds were pre-incubated in 9-cm diameter Petri dishes with 5 mL of distilled water for about 5 h at 25°C. After that period, the seeds of the target weeds were disinfected with 5% sodium hypochlorite for 10 min and rinsed with distilled water. One layer of sterile filter paper was placed at the base of each Petri dish (for A. retroflexus L., 12-well plates were used). Then, 3 mL (for A. retroflexus L., 330 μL) of the methanolic solution containing compounds were dropped on the filter paper. The final concentrations of tested compounds were 500, 250, 125, 62.5, and 31.25 μg/mL, respectively. The equivalent sterilized water was added to each well and Petri dish after the escape of methanol. Twenty-five (for A. retroflexus L., 10) viable seeds of weeds were placed on a filtrate paper. The herbicide glufosinate-ammonium was used as a positive control, and the methanol solution was then used as solvent control. Lids were sealed with Parafilm and incubated at 28°C, with 12 h supplemental light provided by 400 W Philips lamps and 26°C with 12 h darkness per day. The radicle and plumule lengths were measured and inhibition rates were calculated after 4 days. All treatments were carried out in triplicate. The experimental results are expressed as the mean ± SD, and the EC₅₀ values were calculated from the regression equations.

The inhibition rate (expressed as a percentage) was calculated as follows:

Compound 1 was isolated as a yellow, amorphous powder, and its molecular formula was deduced as C₂₅H₃₁O₆Cl by HRESIMS (Supplementary Figure S7), corresponding to 10 degrees of unsaturation. The chlorine atom was confirmed by an isotopic peak for [M+H]⁺:[M+H+2]⁺ with an intensive ratio of 3:1 in the molecule. It existed as inseparable mixtures of two isomers according to HPLC analysis on either ODS or chiral column, due to spontaneous isomerization. The ¹H and ¹³C NMR spectra (Tables 1, 2) of 1 showed two sets of resonances with a ratio of 1:4 for the 1a and 1b isomers. The ¹H NMR spectroscopic data and heteronuclear singular quantum correlation (HSQC) correlations (Supplementary Figure S3) of 1a revealed seven methyl groups, including five singlets (δ_H/δ_C 1.45/24.6, 1.58/21.5, 1.89/20.1, 3.32/49.2, and 3.74/52.2), one doublet (δ_H/δ_C 1.01/20.9), one triplet (δ_H/δ_C 0.85/12), one methylene (δ_H/δ_C 1.33/30.2, 1.44/30.2), three aliphatic methines (δ_H/δ_C 2.64/34, 3.04/58.2, and 3.80/44.6), and five olefinic protons (δ_H/δ_C 5.49/145.1, 6.14/118.8, 6.53/105.8, 7.41/133.4, and 7.56/146). Additionally, the ¹³C NMR spectra of 1a revealed the presence of 25 carbons, including one conjugated ketone carbonyl at δ_C 188.4, one ester carbonyl at δ_C 169.3, one oxygenated quaternary olefinic carbon at δ_C 157.7, two sp³ oxygenated quaternary carbons signal at δ_C 83.6/105.6, and four quaternary olefinic carbons signal at δ_C 109.3/114.2/129.8/140.3. These NMR spectroscopic data indicated that 1a belonged to the family of azaphilones and the planar structure was the same as that of isochromophilone C (Luo et al., 2018). Further examination found that the key nuclear overhauser effect spectroscopy (NOESY) correlations (Figure 2, Supplementary Figure S6) had obvious differences between 1a and isochromophilone C, indicating they were diastereoisomers.

The relative configuration of 1a was established by NOESY correlations (Figure 2, Supplementary Figure S6) and corresponding proton coupling constants (Luo et al., 2018; Qian et al., 2019). The large coupling constant (J = 12.6 Hz) between H-2' and H-8 suggested these two protons were on the opposite orientation. The NOESY correlations observed for H₃-18, H-8, and 1'-OCH₃ indicated that these protons were located on the same face. Thus, the stereochemistry of the azaphilone skeleton of 1a was determined. Furthermore, the coupling constant between H-9 and H-10 (J = 15.6 Hz) in addition to the NOESY correlations between H-9/H-12 and 17-CH₃, and between H-10 and 16-CH₃ illustrated that the double bond at C-9 and C-10 was E configuration, and C-11/C-12 was Z configuration. Therefore, the relative configuration of 1a, differing from that of isochromophilone C, was assigned as rel-(7R,8R,9E,11Z,1′R,2′S), and given the name as isochromophilone H.

The ¹H and ¹³C NMR data of 1b were similar to those of 1a. The differences between them were the NMR data of C-9 to C-13, and C-17, indicating 1a/1b were a pair of E/Z isomers, which was confirmed by the NOESY correlations of H-9/17-CH₃, and H-10/H-12. Hence, 1b was defined as rel-(7R,8R,9E,11E,1′R,2′S)-isochromophilone H.

Compounds 2–7 were also isolated as six pairs of C-11 E/Z isomers. Among them, 2a, 3a, 4a, 5a, and 6a were new compounds with 11-(Z) configuration, which were confirmed by NOESY correlations of H-9/H-12 and 17-CH₃, and H-10/16-CH₃. The relative configuration of 5b was also determined for the first time by the NOESY correlations (Supplementary Figure S34), combined with the proton coupling constants. The coupling constants calculated for H-8 (J = 10.2 Hz) and H-8a (J = 4.8, 10.2, and 13.2 Hz), indicated the ax/ax relationship of these two protons. The NOESY correlations observed for H-1α/H₃-18 and H-8 suggested that these protons were cofacial. Accordingly, the relative configuration of 5b was established as rel-(7R,8R,8aR).

The absolute configurations of all the isolated compounds were determined by comparison of experimental ECD spectra, and biosynthetic considerations. Among these compounds, the stereogenic carbon at C-13 in the side chain moiety was established to be S due to the aliphatic branch of this kind of azaphilones having a shared biosynthetic pathway (Gao et al., 2013). The absolute configuration of C-7 in compounds 1–5 was assigned to be R based on positive Cotton effects at 312 (Δε + 1.18, 1), 312 (Δε + 3.82, 2), 313 (Δε + 8.85, 3), 310 (Δε + 6.11, 4), and 325 nm (Δε + 0.32, 5), respectively (Figures 3, 4; Qian et al., 2019). In addition, combined with the NOESY correlations, the absolute configurations of 1a and 1b were assigned as 7R,8R,13S,1′R,2′S (Luo et al., 2018). The absolute configurations of 2, 3, 4, and 5 were suggested to be (7R,8R,19S), (7R,8R,19R), (7R,8R,3′R), and (7R,8R,8aR) by comparisons of experimental ECD data to those reported in the literature (Figure 4; Arai et al., 1995; Matsuzaki et al., 1995; Qian et al., 2019). Contrary to ECD data of 5 and epi-isochromophilone III, the ECD spectrum (Figure 4) in MeOH of 6 exhibited Cotton effects at 388 (Δε +1.88), 321 (Δε −0.34), 255 (Δε +3.47), and 214 nm (Δε −1.87), indicating the absolute configuration at C-7 of 6 was S. Thus, 6 was identified as a new compound, and the absolute configurations of 6a and 6b were confirmed as 7S,8R,8aS,13S (Hemtasin et al., 2016). Compounds 2b–5b were identified as the isomers of 2a–5a with 11-(E) configuration reported in the literature (Arai et al., 1995; Matsuzaki et al., 1995; Pairet et al., 1995; Qian et al., 2019). Compounds 7a and 7b were identified as isochromophilone Ib and Ia by comparing their NMR data with those in the literature (Omura et al., 1993; Matsuzaki et al., 1995).

Numerous studies have reported that azaphilones have broad-spectrum biological activities. However, there were relatively few reports about their phytotoxic activity of them. Chaetomugilin A, D, S, and O, four chlorine-containing azaphilone derivatives, isolated from the endophytic Chaetomium globosum TY1, showed a higher response index and lower IC₅₀ values to eight species of herbaceous plant seeds than positive control glyphosate (Wang et al., 2017). Chaetomugilin D and J, isolated from the EtOAc extract of the fermentation medium of C. globosum, exhibited phytotoxicity to lettuce seeds, with IC₅₀ values for root inhibition of 24.2 and 22.6 ppm, respectively, while those for shoot inhibition were 27.8 and 21.9 ppm, respectively (Piyasena et al., 2015). Acetosellin, isolated from large-scale cultures of the fungus Cercosporella acetosella, inhibited the growth of the root of Lepidium sativum and Zea mais at 640 μM (Gianluca et al., 2002). In the present work, all the isolated compounds were evaluated for their phytotoxicity against four weeds species (A. fatua L., L. perenne L., A. retroflexus L., A. theophrasti Medikus) in farmland. The experimental results indicated that sclerotiorin B (3), ochlephilone (4), and isochromophilone I (7) exhibited potent phytotoxicity against the growth of radicle and plumule of A. retroflexus L., with EC₅₀ values ranging from 234.87 to 320.84 μM, compared to positive control glufosinate-ammonium (Table 3). Compounds 4 and 7 also showed inhibitory activities against the growth of velvetleaf (A. theophrasti Medikus; Table 4). These tested compounds had no significant inhibitory effects on the growth and germination of wild oat (A. fatua L.) and ryegrass (L. perenne L.).