Terpenoids From the Coral-Derived Fungus Trichoderma harzianum (XS-20090075) Induced by Chemical Epigenetic Manipulation

The soft coral-derived fungus Trichoderma harzianum (XS-20090075) was found to be a potential strain to produce substantial new compounds in our previous study. In order to explore its potential to produce more metabolites, chemical epigenetic manipulation was used on this fungus to wake its sleeping genes, leading to the significant changes of its secondary metabolites by using a histone deacetylase (HDAC) inhibitor. The most obvious difference was the original main products harziane diterpenoids were changed into cyclonerane sesquiterpenoids. Three new terpenoids were isolated from the fungal culture treated with 10 μM sodium butyrate, including cleistanthane diterpenoid, harzianolic acid A (1), harziane diterpenoid, harzianone E (2), and cyclonerane sesquiterpenoid, 3,7,11-trihydroxy-cycloneran (3), together with 11 known sesquiterpenoids (4–14). The absolute configurations of 1–3 were determined by single-crystal X-ray diffraction, ECD and OR calculations, and biogenetic considerations. This was the first time to obtain cleistanthane diterpenoid and africane sesquiterpenoid from genus Trichoderma, and this was the first chlorinated cleistanthane diterpenoid. These results demonstrated that the chemical epigenetic manipulation should be an efficient technique for the discovery of new secondary metabolites from marine-derived fungi.


INTRODUCTION
Marine fungi have been proved to possess the potential ability to produce structurally unique and biologically active secondary metabolites (Blunt et al., 2018). However, it has become a crucial issue to discover microbial natural products due to repeating isolation of known compounds at the traditional methods involving bulk culture of the organism and subsequent fractionation and bioassay to determine if specific fractions hold any bioactive metabolites (Spraker and Keller, 2014). It has been revealed that fungi possess far more gene clusters encoding secondary metabolites than GRAPHICAL ABSTRACT | Terpenoids from T. harzianum induced by chemical epigenetic manipulation. their characterized compounds (Khaldi et al., 2010). In order to solve this challenge, a number of manipulations have been used to regulate the production of secondary metabolites from fungi, such as One strain many compounds (OSMAC) (Pan et al., 2019), co-culture , interspecies crosstalk , and heterologous expression (Huo et al., 2019;Zhang et al., 2019). Among these methods, chemical epigenetic manipulation has been demonstrated to be a promising strategy to wake the silent biosynthetic gene clusters to obtain novel compounds and has been applied to the marine fungi (Asai et al., 2013). For instance, histone deacetylase (HDAC) inhibitor, suberoylanilide hydroxamic acid (SAHA), was applied to an algicolous strain of Aspergillus wentii, giving rise to three new norditerpenoids with potent bioactivities (Li et al., 2017). Similarly, treating deepsea-derived Eutypella sp. fungus with a combination of HDAC inhibitor (SAHA) and DNA methyltransferase (DNMT) inhibitor (5-azacytidine) resulted in the discovery of three new eremophilane-type sesquiterpenoids with nitric oxide inhibitory activities (Niu et al., 2018). These cases might demonstrate that chemical epigenetic manipulation could efficiently excavate novel secondary metabolites from marine-derived fungi. However, the successful examples of chemical epigenetic manipulation applied to marine-derived fungi are not abundant enough to confirm the conclusion.
Trichoderma species are widespread, highly competitive soil-borne fungi. They display a successful antagonism against a variety of other fungi (Hanson, 2005). Fungus Trichoderma harzianum is known to be a biocontrol agent against phytopathogenic fungi extensively applied in agriculture (Hassan et al., 2019;Dufresne et al., 2020). This fungus has some applications in other aspects of agriculture, such as improving drought tolerance in rice genotypes (Pandey et al., 2016) and increasing plant productivity (Poveda et al., 2019). One of the mechanism of these bioactivities of T. harzianum is considered to be its ability to produce metabolites with various activities, such as antifungal harzianopyridone (Ahluwalia et al., 2015) and 6-pentyl-α-pyrone (Chen et al., 2012), plant growth promoter harzianic acid (Vinale et al., 2013), and plant growth regulator harzianolide (Cai et al., 2013). Marine-derived T. harzianum can also produce substantial active secondary metabolites (Liang et al., 2019;Yamada et al., 2019;Zhao et al., 2019). More than 60 compounds have been isolated from marine-derived T. harzianum that further demonstrated the potential ability of this fungus to produce natural products with diverse structures. Up to now, there has no research to study the secondary metabolites of T. harzianum through epigenetic modification.
During our ongoing investigation to discover bioactive marine natural products, we have also obtained new metabolites from marine-derived fungi by using chemical epigenetic modification Chen et al., 2016;Wu et al., 2019Wu et al., , 2020. In our previous work, a series of harziane diterpenoids and hydroxyanthraquinones have been discovered from the fungus T. harzianum (XS-20090075) isolated from a soft coral collected from the South China Sea (Shi et al., 2018;Zhao et al., 2019). In order to obtain more new bioactive compounds, chemical epigenetic manipulation was employed on this fungal strain to mine its potential ability to produce metabolic products. Screening chemical epigenetic modifying agents resulted in the significant changes of its metabolic profile by using HDAC inhibitors. Subsequently, besides harziane diterpenoid, new metabolic products were discovered from this strain including cleistanthane diterpenoids and cyclonerane sesquiterpenoids. Herein, we report the epigenetic modification on this fungus, and the isolation, structural characterization and bioactivity evaluation of these metabolites.

General Experimental Procedures
Optical rotations were measured on a JASCO P-1020 digital polarimeter. UV spectra were recorded on a Beckman DU 640 spectrophotometer. ECD spectra were obtained on a Jasco J-815-150S circular dichroism spectrometer. IR spectra were recorded on a Nicolet-Nexus-470 spectrometer using KBr pellets. NMR spectra were measured on an Agilent DD2 500 MHz NMR spectrometer (500 MHz for 1 H and 125 MHz for 13 C), using TMS as an internal standard. The ESIMS and HRESIMS spectra were obtained from a Micromass Q-TOF spectrometer and a Thermo Scientific LTQ Orbitrap XL spectrometer, respectively. The crystallographic data were collected on a Bruker APEX-II CCD diffractometer equipped with graphite monochromatized Cu Kα radiation. Semi-preparative HPLC was performed on a Waters 1525 system coupled with a Waters 2996 photodiode array detector. A Kromasil C 18 semi-preparative HPLC column (250 × 10 mm, 5 µm) was used. Silica gel (Qing Dao Hai Yang Chemical Group Co.; 200-300 mesh), Sephadex LH-20 (Amersham Biosciences) and octadecylsilyl silica gel (Unicorn; 45-60 µm) were used for column chromatography (CC). Precoated silica gel GF 254 plates (Yantai Zifu Chemical Group Co., Yantai, China) were used for thin-layer chromatography.

Fungal Material
The fungal strain T. harzianum

Antibacterial and Antifungal Activity Assay
The antibacterial and antifungal activity was evaluated by the conventional broth dilution assay (Appendino et al., 2008).

DNA Topo I Inhibition Bioassay
The Topo I inhibitory activity was measured by assessing the relaxation of supercoiled pBR322 plasmid DNA (Bogurcu et al., 2011). Camptothecin (CPT) was used as a positive control. The gel was stained with Gelred and visualized under UV illumination and then photographed with a Gel imaging system.

AChE Inhibition Bioassay
The AChE inhibition activity was measured based on the modified Ellman's method (Ellman et al., 1961). Huperzine A and galanthamine were used as positive drugs. The inhibition rates of AChE were calculated using Origin 8.0 software.

and DNA methytransferase (DNMT) inhibitors with different concentrations (Supplementary
Harzianolic acid (1) was obtained as yellow, needle crystals. The molecular formula C 20 H 29 O 5 Cl of 1 was determined by HRESIMS spectrum, indicating six degrees of unsaturation. The  IR spectrum showed hydroxyl (3445 cm −1 ) and carboxyl (2921, 1700 cm −1 ) characterized absorption bands. The 1 H NMR, 13 C NMR (Table 1) and HSQC spectra of 1 exhibited two methyl groups, six methylene groups including one olefinicmethylene, seven methines including two olefinic-methines and two oxy-methines, and five non-protonated carbons including one carboxyl group and one olefinic carbon. These NMR signals combined with its degrees of unsaturation indicated that 1 should be a tricyclic diterpenoid belonging to the family of cleistanthane-type diterpenoid. The spectroscopic feature of 1 was similar to those of zythiostromic acid B which was isolated from the fungus Zythiostroma sp. derived from aspen Populus tremuloides Michx (Ayer and Khan, 1996). However, the hydroxyl substituent at C-5 in zythiostromic acid B was replaced by a hydrogen atom in 1, indicating by the methine signal of H-5 in 1 H NMR and the upfield of C-5 in 13 C NMR. This was also confirmed by the COSY cross peak of H-5/H-6 and HMBC correlations from H-18 and H-20 to C-5 (Figure 3). The COSY correlation of H-12/12-OH, HMBC correlation from 12-OH to C-11 and the downfield shift of C-12 in 13 C NMR indicated a hydroxyl group locating at C-12. The cross peaks from 14-OH to C-8, C-13, C-14, C-15 in HMBC, as well as the downfield shift of C-14 in 13 C NMR suggested a hydroxyl group anchoring at C-14. The ESIMS of 1 exhibited chlorine isotope peak ratio of 3:1 revealed that a chlorine substitute should occur in 1. The methylene at C-17 in zythiostromic acid B was replaced by the methine in 1, indicating the chlorine was located at C-17, which was confirmed by the downfield shifts of C-17 in 13 C NMR and H-17 in 1 H NMR.
The relative configuration of 1 was determined by NOESY data. The NOESY spectrum of 1 measured in DMSOd 6 was not enough to indicate its relative configuration (Supplementary Figure S8). So it was measured again in acetoned 6 (Supplementary Figure S12). The NOESY correlations of H-5/H-7b, H-5/H-9, and H-5/H-18, and the correlations of H-9/H-15 and H-18/3-OH indicated that H-5, H-7b, H-9, H-15, H-18, and 3-OH were in the same face (Figure 4). The correlations of H-12/H-7a and H-12/H-20 suggested H-7a, H-12, and H-20 were in another face (Figure 4). The NOESY correlation between H-17 and 14-OH indicated the E configuration of the double bond at C-13 and C-17 (Figure 4).
The absolute configuration of 1 was first attempted to solve by the modified Mosher's method. Unfortunately, this method was unsuccessful due to the limited quantity of 1 and the multiple hydroxyls in 1 resulting in the complex products. Fortunately, the single crystals of 1 suitable for X-ray diffraction analysis was obtained by slowly crystallization from the mixture solvent of CH 3 OH/CH 2 Cl 2 /H 2 O (20:20:1). The stereochemistry of 1 was undisputed confirmed to be 3R,4R,5R,8R,9S,10R,12S,14S,13E by Cu Kα X-ray diffraction with a Flack's parameter of 0.050(5) (Figure 5).
Harzianone E (2) was obtained as a colorless oil. The molecular formula of C 20 H 28 O 3 was determined by HRESIMS indicating seven degrees of unsaturation. The IR absorption bands at 3561 and 1685 cm −1 indicated the presence of hydroxyl and carbonyl groups. The analyzing of 1 H NMR and 13 C NMR spectra of 2 (Table 2) suggested that 2 belonged to the family of harziane diterpenoid and was very similar to harziandione which was first isolated from the biological control agent T. harzianum Rifai (Ghisalberti et al., 1992). The only difference between 2 and harziandione was the oxy-methylene (δ C 67.1, δ H 4.38) at C-20 in 2 instead of the methyl (δ C 22.5, δ H 2.12) at C-20 in harziandione, revealing a hydroxyl substituent at C-20. This was confirmed by the correlations from H-8b to C-20, and from H-20 to C-10 in HMBC of 2 (Figure 3). The relative configuration of  (Figure 4). The NOESY correlation between H-5 and H-19 suggested that H-5 and H-19 were located on another face of the molecule (Figure 4). Thus, the relative configuration of 2 was determined as 2S * ,5R * ,6R * ,13S * ,14S * .

Plausible Biogenetic Pathways Proposed for 1-14
The emerged main metabolic pathway of T. harzianum (XS20090075) was the biosynthesis of harziane diterpens at the traditional experimental condition in rice culture (Zhao et al., 2019). By chemical epigenetic manipulation, the biosynthesis of harziane diterpens was depressed. Only one harziane diterpen, harzianone E (2), was obtained in this study, which might be the hydroxylation and oxidation derivative of harziane ( Figure 7A) (Adelin et al., 2014). Absorbingly, the new produced cleistanthane diterpenoid, harzianolic acid A (1), might reveal the activation of cleistanthane diterpenoid biosynthesis pathway. Compound 1 may be generated from CPP to cleistanthadiene by cyclization and rearrangement (Bai et al., 2018), followed by the oxidation and halogenation of cleistanthadiene ( Figure 7A). In contrast, the biosynthetic pathway of cyclonerane sesquiterpenoids was activated, resulting in the discovery of a series of cyclonerane sesquiterpenoids, 3,7,11-trihydroxy-cycloneran (3) and 4-11. The biosynthesis pathway of these sesquiterpenoids may start from FPP or NPP to 9 through cyclization (Evans et al., 1976), followed by oxidation, hydroxylation, hydration or reduction procedures ( Figure 7B). Interestingly, the new produced ophioceric acid (12) might reveal the activation of the biosynthesis of african sesquiterpenoids due  Frontiers in Microbiology | www.frontiersin.org 9 April 2020 | Volume 11 | Article 572 to epigenetic manipulation. Compound 12 may be produced from FPP to african-2-ene through cyclization (Wawrzyn et al., 2012), followed by oxidation reaction (Figure 7B).

Bioactivities of Compounds 1-14
All of the isolated compounds (1-14) were evaluated for their antibacterial, antifungal, DNA topoisomerase I (Topo I) inhibitory and acetylcholinesterase (AChE) inhibitory activities.
The results indicated that only 2 and 4 exhibited weak antibacterial activity against P. angustum.

DISCUSSION
Our continuing research on the metabolic products from the fungal strain T. harzianum (XS-20090075) revealed the presence of the main products, harziane diterpenoids, in rice culture (Zhao et al., 2019). Attractively, the metabolic profiles of the fungal strain have been changed significantly after chemical epigenetic manipulation, resulting in the emergence of cyclonerane sesquiterpenoids as main metabolic products and the restraining of the production of harziane diterpenoids. Interestingly, a new type of diterpenoid, cleistanthane diterpenoid, arisen owing to epigenetic modification. It should be noted that the cleistanthane diterpenoid and african sesquiterpenoid were the first time to be discovered from Trichoderma species and this was the first time to discover cleistanthane diterpenoid containing chlorine. This study was the first time to regulate the secondary metabolites of fungus T. harzianum applying chemical epigenetic manipulation, leading to the significant changes of its metabolic profiles. Evidently, this attempt will offer a new approach to mine the secondary metabolites from T. harzianum. Furthermore, this research was also a powerful evidence to prove that chemical epigenetic manipulation could be an efficient way to wake the silence genes of marine-derived fungi to discover new natural compounds, and even to promote the research and development of marine natural products.

CONCLUSION
In summary, the chemical epigenetic manipulation of the soft coral-derived fungus T. harzianum (XS-20090075) led to the significant changes of its metabolic profiles. The epigenetic modification suppressed the biosynthesis pathway of harziane diterpenoids while excited the pathway of cyclonerane sesquiterpenoids, resulting in the alteration of main metabolic products from harziane diterpenoids to cyclonerane sesquiterpenoids (3-11). Intriguingly, two new biosynthesis pathway of cleistanthane diterpenoid and african sesquiterpenoid were found to be activated by epigenetic modification, illustrated by the appearance of cleistanthane diterpenoid harzianolic acid A (1) and african sesquiterpenoid ophioceric acid (12). Noticeably, the cleistanthane diterpenoids and african sesquiterpenoids were the first time to be discovered from Trichoderma and this was the first time to discover cleistanthane diterpenoid containing chlorine. These results demonstrated that the chemical epigenetic manipulation should be an efficient technique for the discovery of new secondary metabolites from marine-derived fungi.

DATA AVAILABILITY STATEMENT
The datasets generated for this study can be found in the GenBank(NCBI) accession number KU866299; The CCDC number of crystals of 1 was 1910127, 12 was 1910138.

AUTHOR CONTRIBUTIONS
TS contributed to the fermentation, extraction, isolation, and manuscript preparation. C-LS, YL, and D-LZ reviewed and amended the manuscript. FC contributed to the quantum chemistry calculation. X-MF, J-YY, J-SW, and Z-KZ contributed to the bioactivities test. C-YW was the project leader, organized and guided the experiments and manuscript writing.