Novel Harziane Diterpenes from Deep-Sea Sediment Fungus Trichoderma sp. SCSIOW21 and Their Potential Anti-Inflammatory Effects

Five undescribed harziane-type diterpene derivatives, namely harzianol K (1), harzianol L (4), harzianol M (5), harzianol N (6), harzianol O (7), along with two known compounds, hazianol J (2) and harzianol A (3) were isolated from the deep-sea sediment-derived fungus Trichoderma sp. SCSIOW21. The relative configurations were determined by meticulous spectroscopic methods including 1D, 2D NMR spectroscopy, and HR-ESI-MS. The absolute configurations were established by the ECD curve calculations and the X-ray crystallographic analysis. These compounds (1, and 4–7) contributed to increasing the diversity of the caged harziane type diterpenes with highly congested skeleton characteristics. Harzianol J (2) exhibited a weak anti-inflammatory effect with 81.8% NO inhibition at 100 µM.


Introduction
The Trichoderma fungus, widely distributed in terrestrial and marine habitats, is a kind of important renewable natural resource with high economic value and application prospects. Among them, the species in the marine environment, together with Penicillium and Aspergillus, contributed to the discovery of more than half of the new terpenoids from marine fungi [1,2]. However, Trichoderma was rarely reported from deep marine ecosystems. During 2013 to 2019, a total of 151 novel compounds were reported from deep marine derived-fungi, of which 41.2% were from Penicillium, 28.1% were from Aspergillus, while only 1 Trichoderma was reported from the deep marine system [1].
During our ongoing investigations on inhibitors from deep-sea fungi [18][19][20][21][22][23] against nitric oxide (NO) production induced by lipopolysaccharide (LPS), Trichoderma sp. SC-SIOW21, which was isolated from sea sediment at a depth of over 1000 m, was found to be active. The subsequent cultivation of this strain resulted in the isolation of seven harziane diterpenes, including five new compounds. Herein, we report the isolation and identification procedures, as well as the anti-inflammatory, anti-fungal, and anti-bacterial activities of these compounds.

Results and Discussion
The fungus Trichoderma sp. SCSIOW21 was cultured at room temperature under static conditions. The BuOH extraction was fractioned and purified by silica gel, medium pressure ODS column chromatography, and semi-preparative HPLC to obtain seven harziane diterpenes ( Figure 1).
identification procedures, as well as the anti-inflammatory, anti-f activities of these compounds.

Results and Discussion
The fungus Trichoderma sp. SCSIOW21 was cultured at ro static conditions. The BuOH extraction was fractioned and purifi pressure ODS column chromatography, and semi-preparative H ziane diterpenes (Figure 1). Compound 1 was isolated as colorless crystal, with molecu using HRESIMS data. The IR spectrum showed strong absorption groups at 1734 and 1695 cm −1 , which was consistent with those re [3]. The 1 H NMR and 13 C NMR spectroscopy spectra along wit five methyls, four methylenes, four methines, and seven quaterna 1 and 2). The above NMR spectroscopy signal pattern was simil harziandione [3], except for 3 major differences: an additional hy absent methylene group, and an extra hydroxy group at δ 4.24 co one. The up-field shifts of H-8 to δ 4.24 and C-8 to δ 72.4 suggest to C-8 (Tables 1 and 2). 1 H-1 H COSY correlations between 8-OH a well as HMBC correlations from 8-OH to C-8 and C-7 also confirm ure 2). This conclusion was further secured by careful analysis o copy data, and compound 1 was named as harzianol K, with th shown in Figure 2. Compound 1 was isolated as colorless crystal, with molecular formula as C 20 H 28 O 3 using HRESIMS data. The IR spectrum showed strong absorption bands for two carbonyl groups at 1734 and 1695 cm −1 , which was consistent with those reported for harziandione [3]. The 1 H NMR and 13 C NMR spectroscopy spectra along with HSQC data suggested five methyls, four methylenes, four methines, and seven quaternary carbon atoms (Tables 1 and 2). The above NMR spectroscopy signal pattern was similar to the prior report for harziandione [3], except for 3 major differences: an additional hydroxy group at δ 5.31, an absent methylene group, and an extra hydroxy group at δ 4.24 compared with harziandione. The up-field shifts of H-8 to δ 4.24 and C-8 to δ 72.4 suggested this group connected to C-8 (Tables 1 and 2). 1 H-1 H COSY correlations between 8-OH and H-8, H-8 and H-7, as well as HMBC correlations from 8-OH to C-8 and C-7 also confirmed the elucidation ( Figure 2). This conclusion was further secured by careful analysis of 1D, 2D NMR spectroscopy data, and compound 1 was named as harzianol K, with the molecular framework shown in Figure 2.        The experimental CD spectrum of 1 was in accordance with the theoretically calculated ECD curve of the 2S, 5R, 6R, 8S, 13S, and 14S configuration. A total of 3 cotton effects were  (Figure 4a). Eventually, the stereocenters of 1 were determined as 2S, 5R, 6R, 8S, 13S, and 14S unambiguously through analysis of X-ray single-crystallography ( Figure 5). Compounds 2 and 3 were confirmed as known compounds, namely harzianol J [8] and harzianol A [13], by comparing their NMR spectroscopy data with those reported in the literature (Tables S2 and S3) [8]. Nevertheless, the absolute configuration of 2 was not determined previously. Herein we report it as 2S, 5R, 6R, 13S, and 14S by X-ray diffraction ( Figure 5).
Compounds 4-7 were all purified as colorless gum or amorphous solids. The molecular formulas of 4-7 were established as C20H30O3, C20H30O4, C20H30O3, and C20H30O3 based on HRESIMS data, respectively.
The IR spectrum of 4 showed strong absorption band for carbonyl group at 1716 cm −1 . The 1 H and 13 C NMR spectra of 4 (Tables 1 and 2) were similar to those of harzianol A (3) [13] except for two major differences: the lack of a methyl group and the presence of an extra hydroxy methylene group. The δH signals at 3.41, 3.28, 4.39 (OH) and the δC signal at 63.9 suggested that one methyl group was hydroxylated. The 1 H-1 H COSY cross-peaks between the hydroxy proton and methylene proton, methylene proton and H-5 (δH 2.13), Compounds 2 and 3 were confirmed as known compounds, namely harzianol J [8] and harzianol A [13], by comparing their NMR spectroscopy data with those reported in the literature (Tables S2 and S3) [8]. Nevertheless, the absolute configuration of 2 was not determined previously. Herein we report it as 2S, 5R, 6R, 13S, and 14S by X-ray diffraction ( Figure 5).
Compounds 4-7 were all purified as colorless gum or amorphous solids. The molecular formulas of 4-7 were established as C 20  The IR spectrum of 4 showed strong absorption band for carbonyl group at 1716 cm −1 . The 1 H and 13 C NMR spectra of 4 (Tables 1 and 2) were similar to those of harzianol A (3) [13] except for two major differences: the lack of a methyl group and the presence of an extra hydroxy methylene group. The δ H signals at 3.41, 3.28, 4.39 (OH) and the δ C signal at 63.9 suggested that one methyl group was hydroxylated. The 1 H-1 H COSY cross-peaks between the hydroxy proton and methylene proton, methylene proton and H-5 (δ H 2.13), along with the HMBC correlations from the hydroxy proton to C-5 (δ C 40.1) and C-18 (δ C 63.9), proved the hydroxy group connected to C-18 unambiguously. The molecular framework of 4 was consequently elucidated as harzianol L (Figures 1 and 2). The relative configuration of 4 was determined by ROESY spectra which showed the same correlation patterns as those of 1 (Figure 3). The absolute configuration of 4 was determined as 2R, 5S, 6R, 13S, and 14S by comparison of experimental CD spectrum with its calculated ECD data (Figure 4b).
The IR spectrum of 5 showed strong absorption band for carbonyl group at 1732 cm −1 . The NMR spectroscopy data of 5 was almost consistent with those of 4, except that a methylene group was missing, whereas an extra oxygenated methine group (δ H 4.21 and δ C 73.5) was detected. The signals suggested that one methylene group was oxygenated (Tables 1 and 2). 1 H-1 H COSY correlations between the hydroxy proton and H-8, between H-8 and H-7, confirmed the connection of the hydroxy group to C-8. The structure was then determined as harzianol M by a detailed analysis of 2D NMR data (Figures 1 and 2). In the ROESY spectra, H-8 showed correlations with Me-19, indicating the β configuration of the 8-hydroxy group (Figure 3). The absolute configurations of 5 were established as 2R, 5S, 6R, 8S, 13S, and 14S based on ECD calculation (Figure 4c).
The IR spectrum of 6 showed a strong absorption band for carbonyl group at 1734 cm −1 . The NMR spectroscopy spectra of 6 matched well with those of 5, with just 1 more extra methine group (δ H 2.26 and δ C 51.8) and 1 less oxygenated quaternary carbon signal (Tables 1 and 2). 1

H-1 H COSY correlations between the methine proton and H-3, H-15
suggested the methine group was located at C-3. The molecular framework of 6 was consequently established as harzianol N through a detailed analysis of 2D NMR spectroscopy spectra (Figures 1 and 2). The absolute configurations of 6 were determined as 2S, 5S, 6R, 8S, 13S, and 14S through detailed analysis of ROESY spectra and ECD calculation (Figures 3 and 4d).
The IR spectrum of 7 showed strong absorption band for carbonyl group at 1718 cm −1 . The 1 H and 13 C NMR spectroscopy data of 7 were similar to those reported for harzianol A (3) (Table S3) [13], with an extra oxygenated methine group (δ H 3.65 and δ C 73.5) and a disappeared methylene group, indicating the oxygenation of the methylene group (Tables 1 and 2). The molecular framework was confirmed as harzianol O (Figures 1 and 2) through a detailed analysis of 2D NMR spectroscopy data, including the key COSY correlation between the methine proton and H-14 (δ H 2.07), which suggested the hydroxy group connected to C-15. The ROESY correlations between H-15 and Me-19 suggested the β configuration of the 15-hydroxy group (Figure 3). The absolute configurations of 7 were determined as 2S, 5R, 6R, 13S, 14S, and 15R by ECD calculation.
The anti-inflammatory activity of compounds 1-7 was measured by NO production inhibitory assay [20]. The cytotoxicity of these compounds was tested to avoid false-positive results due to cell death, and none of them showed cytotoxicity at the concentrations of 25-100 µM ( Figure 6). Hazianol J (2), harzianol A (3) and harzianol O (7) exhibited the strongest NO production inhibitory activity at 100 µM with inhibitory rates at 81.8%, 46.8%, and 50.5%, respectively. The IC 50 of Hazianol J (2) was 66.7 µM, while harzianol L (4) and harzianol K (1) only showed weak inhibition at the highest concentration of 100 µM (Figure 6). Compounds without "top" hydroxy groups at C-8 and C-18 (2,3, and 7) exhibited higher NO production inhibitory activities compared to the compounds with more hydroxy groups (1, 4, 5, and 6). These hydroxy groups may reduce the membrane permeability and reduced the activities.
strongest NO production inhibitory activity at 100 µM with inhibitory rates at 81.8%, 46.8%, and 50.5%, respectively. The IC50 of Hazianol J (2) was 66.7 µM, while harzianol L (4) and harzianol K (1) only showed weak inhibition at the highest concentration of 100 µM ( Figure 6). Compounds without "top" hydroxy groups at C-8 and C-18 (2,3, and 7) exhibited higher NO production inhibitory activities compared to the compounds with more hydroxy groups (1, 4, 5, and 6). These hydroxy groups may reduce the membrane permeability and reduced the activities. All of the compounds were examined for their activities against plant pathogenic fungi (Helminthosporium maydis, Gibberella sanbinetti, Botrytis cinerea Pers, Fusarium oxysporum f. sp. cucumerinum, Penicillium digitatum). None of the compounds exhibited obvious activities at the test concentration of 100 µg/mL. Since fungi from Trichoderma sp. are widely used as bio-control agents, many harziane diterpenes were investigated All of the compounds were examined for their activities against plant pathogenic fungi (Helminthosporium maydis, Gibberella sanbinetti, Botrytis cinerea Pers, Fusarium oxysporum f. sp. cucumerinum, Penicillium digitatum). None of the compounds exhibited obvious activities at the test concentration of 100 µg/mL. Since fungi from Trichoderma sp. are widely used as bio-control agents, many harziane diterpenes were investigated against plant pathogenic fungi [3,9,10,16,24]. However, the results were controversial. Although harziandione and isoharziandione, the structure of which was latterly revised as harziandione [10], were mentioned as antifungal agents, the activities of the pure compounds were not clarified in the original literature [3,24]. Harzianone was found to be inactive against Colletotrichum lagenarium and Fusarium oxysporum at 30 µg/disk using a disk diffusion assay [10]. Deoxytrichodermaerin and harzianol A were not active against Botrytis cinerea, Fusarium oxysporum, Glomerella cingulata, and Phomopsis asparagi at 40 µg/disk [16]. Harzianone E was not active against Candida albicans by traditional broth dilution assay [9]. According to the previous studies and our results, harziane diterpenes did not show anti-fungal activity.

Fungal Strain and Fermentation
The fungal strain, which was isolated from the South China deep-sea sediment sample (2134 m depth), was identified as Trichoderma sp. SCSIOW21 by ITS sequencing and morphology analysis. Its sequence data was deposited at GenBank (accession number: KC569351.1) and the strain was deposited at the Laboratory of Microbial Natural Products, Shenzhen University, China. The fungal strain was activated on potato dextrose agar dishes containing 3% sea salt at 28 • C for 3 days and cultured in modified rice broth (rice 50.0 g sprayed with 3% sea salt water 60.0 mL for each 500 mL flask) statically at room temperature for 30 days.

Extraction and Isolation
A total of 100 mL of water saturated BuOH were added in each of the Erlenmeyer flasks which contained fermentation broth. The BuOH extract was collected after 12 h and evaporated under vacuum. The extraction was repeated three times and the total yield was 12.9 g.
The BuOH extract was subjected to a silica gel chromatography with a gradient of  Tables 1 and 2

X-ray Crystal Analysis of Compounds 1 and 2
The crystals of compounds 1 and 2 were obtained from concentrated MeOH solutions and 1 suitable crystal for each compound was selected. The crystals were scanned using Cu Kα radiation (λ = 1.54184 Å) on the XtaLAB AFC12 (RINC) Kappa single diffraction instrument, the structures of which were solved by the Olex2 software, the SHELXT [25], and the SHELXL [26] package with the parameters corrected by the least-squares minimization method.
The single-crystal data has been submitted to the Cambridge Crystallographic Data Centre database, with CCDC 2093540 for 1 and CCDC 2093541 for 2. The data can be downloaded for free from the website http://www.ccdc.cam.ac.uk/ (accessed on 7 November 2021).

ECD Computational Methods
The conformations of compounds 1 and 4-7 were searched by Marvin Sketch software (optimization limit = normal, diversity limit = 0.1) ignoring the rotation of methyl and hydroxy groups. Geometric optimization of the molecules in MeOH (Figures S49-S53) was carried out at 6-31G (d, p) level using DFT/B3LYP through Gaussian 09 software [27], within the 3 kcal/mol energy threshold from the global minimum [28]. The ECD curve was simulated based on TD-DFT calculations and drawn with sigma = 0.3 by SpecDis software (version 1.71, Berlin, Germany). The calculated data was also produced by Boltzmann's weighting and magnetization based on experimental values.

MTT and NO Production Inhibitory Assay
The cytotoxicity and NO production inhibitory activity were examined using RAW 264.7 macrophages, and the detailed methods were reported previously [20].

Anti-Fungal Activities
The anti-fungal activities were tested on a 96-well plate by mycelial growth inhibitory assay [29], using actidione as the positive control. Five plant pathogenic fungal species (Helminthosporium maydis, Gibberella sanbinetti, Botrytis cinerea Pers, Fusarium Oxysporum f.