Combination Strategy of Genetic Dereplication and Manipulation of Epigenetic Regulators Reveals a Novel Compound from Plant Endophytic Fungus

The strategies of genetic dereplication and manipulation of epigenetic regulators to activate the cryptic gene clusters are effective to discover natural products with novel structure in filamentous fungi. In this study, a combination of genetic dereplication (deletion of pesthetic acid biosynthetic gene, PfptaA) and manipulation of epigenetic regulators (deletion of histone methyltransferase gene PfcclA and histone deacetylase gene PfhdaA) was developed in plant endophytic fungus Pestalotiopsis fici. The deletion of PfptaA with PfcclA and/or PfhdaA led to isolation of 1 novel compound, pestaloficiol X (1), as well as another 11 known compounds with obvious yield changes. The proposed biosynthesis pathway of pestaloficiol X was speculated using comparative analysis of homologous biosynthetic gene clusters. Moreover, phenotypic effects on the conidial development and response to oxidative stressors in the mutants were explored. Our results revealed that the new strain with deletion of PfcclA or PfhdaA in ΔPfptaA background host can neutralise the hyperformation of conidia in the PfptaA mutant, and that the ΔPfptaA ΔPfhdaA mutant was generally not sensitive to oxidative stressors as much as the ΔPfptaA ΔcclA mutant in comparison with the single mutant ΔPfptaA or the parental strains. This combinatorial approach can be applied to discover new natural products in filamentous fungi.


Introduction
Filamentous fungi produce clinically important secondary metabolites (SMs), including many natural products developed into pharmaceutical drugs. The biosynthesis genes for these SMs are usually clustered in a chromosome, and are called biosynthetic gene clusters (BGCs) [1]. Under laboratory cultural conditions, most of BGCs in silence are unexpressed or little expressed. In recent years, many strategies have been developed to effectively activate the silent expressed BGCs for the discovery of new natural products in filamentous fungi, including heterologous expression, promoter engineering, genetic dereplication, modulations of transcription factor, global regulator, epigenetic regulator, and combinational strategies [2,3].
Genetic dereplication is a powerful approach to discover novel compounds from unknown biosynthesis pathways. Eliminating major SMs is particularly effective to increase the odds of detecting minor SMs as well as being useful for the heterologous expression of biosynthetic genes from other fungal species [4]. Deletion of eight of the most highly

Construction of PfptaA Deletion Mutants
In previous studies, the pta gene cluster for pestheic acid biosynthesis has been identified in P. fici [28], and both ∆PfcclA and ∆PfhdaA mutants has been obtained, respectively [29]. Here, we constructed the single knockout mutant of PfptaA (PFICI_10824), double knockout mutants of PfptaA and PfcclA (PFICI_05127), and PfptaA and PfhdaA (PFICI_08988) using the previously described transformation method, individually [29] ( Figure 1 and Table S1). For deletion of PfptaA, we constructed a plasmid containing the upstream and downstream homologous arms of PfptaA with the resistance gene of G418 antibiotic. The fragments of deletion cassette of PfptaA were amplified by PCR, and the plasmid construct was transformed into the P. fici wild type (WT) strain. Then PfptaA deletion mutants were verified by diagnostic PCR analysis using designated primers (Figure 1a,b and Table S2). Subsequently, the PfptaA deletion plasmid construct was transformed into strains of TYXW7 (∆PfcclA) and TYXW8 (∆PfhdaA). The genomic DNAs of transformants were extracted and correct mutants were verified by diagnostic PCR using designated primers (Figure 1c,d; Figure S1 and Table S2).

Construction of PfptaA Deletion Mutants
In previous studies, the pta gene cluster for pestheic acid biosynthesis has been identified in P. fici [28], and both ΔPfcclA and ΔPfhdaA mutants has been obtained, respectively [29]. Here, we constructed the single knockout mutant of PfptaA (PFICI_10824), double knockout mutants of PfptaA and PfcclA (PFICI_05127), and PfptaA and PfhdaA (PFICI_08988) using the previously described transformation method, individually [29] ( Figure 1 and Table S1). For deletion of PfptaA, we constructed a plasmid containing the upstream and downstream homologous arms of PfptaA with the resistance gene of G418 antibiotic. The fragments of deletion cassette of PfptaA were amplified by PCR, and the plasmid construct was transformed into the P. fici wild type (WT) strain. Then PfptaA deletion mutants were verified by diagnostic PCR analysis using designated primers ( Figure  1a,b and Table S2). Subsequently, the PfptaA deletion plasmid construct was transformed into strains of TYXW7 (ΔPfcclA) and TYXW8 (ΔPfhdaA). The genomic DNAs of transformants were extracted and correct mutants were verified by diagnostic PCR using designated primers (Figure 1c,d; Figure S1 and Table S2).

Assessment of Secondary Metabolites
To evaluate the modulation of secondary metabolites production via the deletion of the target genes, the strains were grown on a rice-based medium and the culture extracts were analyzed by HPLC and LC-MS. The results were changes of SM production profile in ΔPfpta ΔPfcclA and ΔPfpta ΔPfhdaA mutants, compared with WT and ΔPfpta strains (Figure 2a). A new peak was obviously detected along with 11 known SMs. Subsequently, the new peak was further isolated and purified to obtain pure compound 1, produced in ΔPfpta ΔPfhdaA mutant (Figure 2b). Compounds 7 and 12 produced in WT were disappeared in both ΔPfpta ΔPfcclA and ΔPfpta ΔPfhdaA mutants. Compound 10 produced in WT was obviously decreased in both of the double-deletion strains. Compounds 2, 5, 8, and 9 were increased in the deletion strains, in comparison with WT. Compounds 3 and 4 were novel peaks in both of the double deletion strains compared with WT, and compound 4 also was produced in ΔPfpta mutant. The known SMs identified in P. fici wild type, ΔPfcsnE (the fifth subunit of COP9 signalosome, CsnE), and ΔPfhdaA strains are pestaloficiol M (2), pestaloficin D (3), ficiolide J (4), asperpentyn (5), ficiolide C (6), isosulochrin (7), chloropupukeananin (8), ficiolide K (9), pestaloficiol J (10), hydroxyisoseiridin (11), and pestheic acid (12) [29,[33][34][35][36][37][38] (Figure 2c). All of these known structures were elucidated by comprehensive analysis with spectroscopy and HR-ESI-MS (Figures S8-S18).

Identification of Compound 1
The new compound 1 was only produced in ΔPfpta ΔPfhdaA mutant (Figures 2a and  S7). To elucidate the structure of 1, we made a scale-up fermentation and isolated it by a combination of UV-guided fractionation and retention time. Compound 1 was isolated as yellowish oil with the molecular formula C17H20O7, which was deduced by the protonated molecule HR-ESI-MS ion peak with [M + Na] + at m/z 359 ( Figure S7). The careful inspection of 1 H and 13 C NMR of 1 with previously identified compounds in fungal strains indicated that 1 was a 1, 3-enynes-based cyclohexanoid terpenoids. Further analysis of 1D NMR data provided the evidence that backbone of 1 is identical to siccayne which was first isolated as an antibiotic compound from the deuteromycete Helminthosporium siccans in 1981 [39]. The remaining NMR signals and unsaturation revealed that the hexose moiety is proposed. H-1′′/C-5′′ and H-5′′/C-1′′ HMBC correlations established that the hexose moiety underwent ring closure. The key HMBC correlation of H-1′′ with C-2 allowed us to assemble the intact planar structure by ether bond between the siccayne and hexose moiety ( Figure 2b; Table 1 and Figures S2-S6). Unfortunately, due to the lower content of 1, the configuration of hexose moiety remained unclear. Finally, compound 1 was determined to be a new structure and named pestaloficiol X. Glycosylation of small molecular plays a significant role in drug discovery and development [40]. As the precursor of 1, siccayne exhibited moderate antibiotic activity of Gram-positive bacteria and some fungal strains [39], and cytotoxic activity against multiple human cancer cell lines [41]. The glycosylated siccayne such as 1 was speculated to improve the solubility in water and druggability.

Identification of Compound 1
The new compound 1 was only produced in ∆Pfpta ∆PfhdaA mutant (Figures 2a and S7). To elucidate the structure of 1, we made a scale-up fermentation and isolated it by a combination of UV-guided fractionation and retention time. Compound 1 was isolated as yellowish oil with the molecular formula C 17 H 20 O 7 , which was deduced by the protonated molecule HR-ESI-MS ion peak with [M + Na] + at m/z 359 ( Figure S7). The careful inspection of 1 H and 13 C NMR of 1 with previously identified compounds in fungal strains indicated that 1 was a 1, 3-enynes-based cyclohexanoid terpenoids. Further analysis of 1D NMR data provided the evidence that backbone of 1 is identical to siccayne which was first isolated as an antibiotic compound from the deuteromycete Helminthosporium siccans in 1981 [39]. The remaining NMR signals and unsaturation revealed that the hexose moiety is proposed. H-1 /C-5 and H-5 /C-1 HMBC correlations established that the hexose moiety underwent ring closure.
The key HMBC correlation of H-1 with C-2 allowed us to assemble the intact planar structure by ether bond between the siccayne and hexose moiety (Figures 2b and S2-S6 and Table 1 ). Unfortunately, due to the lower content of 1, the configuration of hexose moiety remained unclear. Finally, compound 1 was determined to be a new structure and named pestaloficiol X. Glycosylation of small molecular plays a significant role in drug discovery and development [40]. As the precursor of 1, siccayne exhibited moderate antibiotic activity of Grampositive bacteria and some fungal strains [39], and cytotoxic activity against multiple human cancer cell lines [41]. The glycosylated siccayne such as 1 was speculated to improve the solubility in water and druggability.

Proposed Biosynthesis Pathway of 1 in P. fici
Alkyne is a typical active group in many natural products with antitumor and anti-HIV activities. There have been several analogues of 1 with alkynyl pattern isolated in P. fici [29,33], Biscogniauxia sp. [42], Eutypa lata [43] and Aspergillus sp. [44] (Figure 3a). The biosynthesis gene cluster iac for iso-A82775C has been identified in P. fici. While the prenyltransferase IacE is responsible for the modification of isopentenyl, the mechanism of alkyne formation in the biosynthetic pathway was not elucidated in the work [45]. Inspired by this research, the bis gene cluster for biscognienyne B biosynthesis was identified in Biscogniauxia sp., in which cytochrome P450 enzyme (BisI) is confirmed to catalyze the alkynylation of the prenyl chain [46]. Surprisingly, there was no homologous gene of bisI in iac gene cluster in P. fici. In addition, the oxidoreductase gene iacJ in the iac cluster was unrelated to the formation of alkyne [45]. Simultaneously, another P450 monoxygenase (AtyI) was verified to catalyze dehydrogenation of the prenyl chain and to yield an alkene moiety in compound asperpentyn in Aspergillus sp. [47]. The cytochrome P450 gene Pfici_01577 located outside of iac gene cluster in P. fici was found by BlastP analysis to be the cytochrome P450 homologous protein of both BisI and AtyI with identity and similarity of both 89%/79% and 78%/87%, respectively (Figure 3b) [46,47]. So, there is an obvious difference in the biosynthetic gene clusters in different fungal species that produced siccayne and its analogues. In addition, glycosylation of natural products is catalyzed by glycosyltransferases (GTs). A phenolic GT MhGT1 identified in Mucor hiemalis exhibited broad substrate scope and regio-and stereospecificy [48]. The several O-GTs predicted in P. fici were not within the gene clusters and distributed throughout the genome [40]. The biosynthesis of 1 was proposed according to the above-mentioned studies. Compound 1 was synthesized via the biosynthesis pathway of iac gene cluster along with a putative cytochrome P450 enzyme (Pfici_01577), a putative hydroxylase (Pfici_01576), and an unknown glycosyltransferase in P. fici. The proposed biosynthetic pathways of pestaloficiol X (1) in P. fici was shown in Figure 3c. cytochrome P450 enzyme (Pfici_01577), a putative hydroxylase (Pfici_01576), and an unknown glycosyltransferase in P. fici. The proposed biosynthetic pathways of pestaloficiol X (1) in P. fici was shown in Figure 3c.

Assessment of Conidia Development in the Mutant Strains
Deletion of the cclA can observably decrease the amount of asexual spores and block the production of mature fruiting bodies and sexual development in A. nidulans [24]. To examine the effect on the morphology and conidia development, the strains of P. fici WT, ΔPfptaA, ΔPfcclA, ΔPfhdaA, ΔPfptaA ΔPfcclA, and ΔPfptaA ΔPfhdaA were cultivated on Potato Dextrose Agar (PDA) plate. The differences in morphology between the mutants and WT strain were shown, respectively (Figure 4a), and the differences in conidia number among all of the strains were analyzed (Figure 4b). In the ΔPfptaA mutant, the conidia number increased about 10-fold compared with P. fici WT. Both deletion of PfcclA and PfhdaA led to no obvious change in conidia number compared with P. fici WT. This suggested that PfptaA is involved in the conidia development and formation, and PfcclA and PfhdaA has little effect on conidia development and formation. Furthermore, deletion of PfcclA and PfhdaA in ΔPfptaA mutant had no differential effect on the conidia number compared with ΔPfcclA and ΔPfhdaA mutants, respectively, but both double mutant

Assessment of Conidia Development in the Mutant Strains
Deletion of the cclA can observably decrease the amount of asexual spores and block the production of mature fruiting bodies and sexual development in A. nidulans [24]. To examine the effect on the morphology and conidia development, the strains of P. fici WT, ∆PfptaA, ∆PfcclA, ∆PfhdaA, ∆PfptaA ∆PfcclA, and ∆PfptaA ∆PfhdaA were cultivated on Potato Dextrose Agar (PDA) plate. The differences in morphology between the mutants and WT strain were shown, respectively (Figure 4a), and the differences in conidia number among all of the strains were analyzed (Figure 4b). In the ∆PfptaA mutant, the conidia number increased about 10-fold compared with P. fici WT. Both deletion of PfcclA and PfhdaA led to no obvious change in conidia number compared with P. fici WT. This suggested that PfptaA is involved in the conidia development and formation, and PfcclA and PfhdaA has little effect on conidia development and formation. Furthermore, deletion of PfcclA and PfhdaA in ∆PfptaA mutant had no differential effect on the conidia number compared with ∆PfcclA and ∆PfhdaA mutants, respectively, but both double mutant strains had decreased numbers of conidia in comparison with ∆PfptaA mutant. The results indicated that deletion of PfcclA or PfhdaA neutralises the enhancement of conidia formation in ∆PfptaA host, suggesting interconnected regulatory network among these genes and products in conidia development and formation.

Assessment of Oxidative Stress Response of the Mutant Strains
Epigenetic regulators also influence the fungal growth, development, infection, and their adaptation to environment. For instance, the loss of hdaA did not affect the growth rate of A. nidulans [49], but A. fumigatus ΔhdaA strain showed a statistically significant reduction of growth compared with the wild type [9]. HdaA was involved in sclerotia formation in A. flavus [50], and the deletion of hdaA reduced the oxidative stress tolerance of A. nidulans [49]. On the other hand, the deletion of cclA strongly reduced mycelial growth, asexual sporulation and spore germination, but did not impair the morphogenesis of specialized infection structures in C. higginsianum [21]. To assess any impact of target genes on oxidative stress response, P. fici WT strain and all of mutants were subjected to three oxidative reagents, including diamide, tert-butylhydroperoxide (tBOOH), and menadionesodium bisulfite (MSB). Separate deletion of PfptaA, PfcclA, and PfhdaA, partly lowered the growth rate on PDA medium, and the inhibition was more obvious in the ΔPfcclA mutant than the ΔPfptaA and ΔPfhdaA mutants. There were obvious differences of sensitivity to tBOOH, diamide, and MSB in different mutants in the third day and fifth day as determined via the measurement of colony diameter. The smaller colony size might indicate the more sensitivity. Almost all of the mutants were more sensitive to tBOOH, diamide, and MSB than P. fici WT, and displayed slower growth in the presence of oxidative stressors besides the ΔPfptaA ΔPfhdaA mutants treated with MSB (Figure 5a-d). Moreover, the MSB effects on colony size on the 3rd day and the 5th day were opposite between the P. fici WT and ΔPfptaA mutant. For the double genes-deficient mutants, the ΔPfptaA ΔPfcclA mutant was more sensitive to all of the stressors than the ΔPfptaA mutant, and only more sensitive to tBOOH than ΔPfptaA. The ΔPfptaA ΔPfhdaA mutant was not sensitive to MSB as well as ΔPfptaA mutant, and it was apparently not as sensitive as the ΔPfhdaA mutant to MSB. The effects on strain sensitivity to the oxidative stress agents were generaly different by deletion of epigenetic regulators in various fungal species. The A. nidulans ΔhdaA mutant increased susceptibility to oxidative stress compared with the wild type [49], but the growth of the A. fumigatus ΔhdaA mutant was not affected under

Assessment of Oxidative Stress Response of the Mutant Strains
Epigenetic regulators also influence the fungal growth, development, infection, and their adaptation to environment. For instance, the loss of hdaA did not affect the growth rate of A. nidulans [49], but A. fumigatus ∆hdaA strain showed a statistically significant reduction of growth compared with the wild type [9]. HdaA was involved in sclerotia formation in A. flavus [50], and the deletion of hdaA reduced the oxidative stress tolerance of A. nidulans [49]. On the other hand, the deletion of cclA strongly reduced mycelial growth, asexual sporulation and spore germination, but did not impair the morphogenesis of specialized infection structures in C. higginsianum [21]. To assess any impact of target genes on oxidative stress response, P. fici WT strain and all of mutants were subjected to three oxidative reagents, including diamide, tert-butylhydroperoxide (tBOOH), and menadionesodium bisulfite (MSB). Separate deletion of PfptaA, PfcclA, and PfhdaA, partly lowered the growth rate on PDA medium, and the inhibition was more obvious in the ∆PfcclA mutant than the ∆PfptaA and ∆PfhdaA mutants. There were obvious differences of sensitivity to tBOOH, diamide, and MSB in different mutants in the third day and fifth day as determined via the measurement of colony diameter. The smaller colony size might indicate the more sensitivity. Almost all of the mutants were more sensitive to tBOOH, diamide, and MSB than P. fici WT, and displayed slower growth in the presence of oxidative stressors besides the ∆PfptaA ∆PfhdaA mutants treated with MSB (Figure 5a-d). Moreover, the MSB effects on colony size on the 3rd day and the 5th day were opposite between the P. fici WT and ∆PfptaA mutant. For the double genes-deficient mutants, the ∆PfptaA ∆PfcclA mutant was more sensitive to all of the stressors than the ∆PfptaA mutant, and only more sensitive to tBOOH than ∆PfptaA. The ∆PfptaA ∆PfhdaA mutant was not sensitive to MSB as well as ∆PfptaA mutant, and it was apparently not as sensitive as the ∆PfhdaA mutant to MSB. The effects on strain sensitivity to the oxidative stress agents were generaly different by deletion of epigenetic regulators in various fungal species. The A. nidulans ∆hdaA mutant increased susceptibility to oxidative stress compared with the wild type [49], but the growth of the A. fumigatus ∆hdaA mutant was not affected under oxidative stress conditions compared with the wild type [9]. Moreover, conidial production was indistinguishable between A. fumigatus ∆hdaA and wild-type strains [9]. The A. fumigatus ∆cclA mutant was more sensitive to chemical 6-azauracil (6AU) compared with the wild type [19]. The cause leading to these different effects may be involved in the changes of SMs, and the regulatory mechanism should be explored in the future. oxidative stress conditions compared with the wild type [9]. Moreover, conidial production was indistinguishable between A. fumigatus ΔhdaA and wild-type strains [9]. The A. fumigatus ΔcclA mutant was more sensitive to chemical 6-azauracil (6AU) compared with the wild type [19]. The cause leading to these different effects may be involved in the changes of SMs, and the regulatory mechanism should be explored in the future.  Dextrose Agar (PDA) medium with appropriate antibiotics as required (Table S1). All of

Strains, Media and Culture Conditions
P. fici CGMCC3.15140 and its correct transformants were grown at 25 • C on Potato Dextrose Agar (PDA) medium with appropriate antibiotics as required (Table S1). All of the strains were grown on the rice medium with water at 25 • C for 20 days for extraction and analysis of secondary metabolites. Escherichia coli DH5α and Agrobacterium tumefaciens AGL-1 were propagated at 37 • C in LB medium with appropriate antibiotics for plasmid DNA amplification and transformation, respectively.

Plasmids for Deletion of PfptaA
The plasmids and primers are listed in Table S1 and Table S2, respectively. PCR amplifications were executed in the T100TM Thermal cycler (Bio-Rad, Hercules, CA, USA). TransStart-FastPfu DNA polymerase as a High-Fidelity DNA polymerase (TransGene Biotech, Beijing, China) was used to amplify the gene fragments. PCR screenings for transformants were performed by using 2×Taq Mix kit (Tiangen Biotech, Beijing, China). PCR reaction and thermal profiles were referred to the manufacturer s instructions. The restriction enzymes used in this work were obtained in New England Biolabs (New England Biolabs Inc. (NEB), Ipswich, MA, USA). To generate the deletion cassette, we used Fusion PCR strategy as described previously [51]. Briefly, G418 was amplified from the pAG1-H3-G418, and around 1.1 kb of fragments upstream and downstream of the gene PfptaA were amplified from P. fici genomic DNA using the designed primers. The three PCR fragments were ligated into the T-vector p-Blunt, and then were amplified for transformation in P. fici strains.

Transformation in P. fici
For creation of deletion of PfptaA (PFICI_10824) mutants in P. fici wild type, ∆PfcclA, and ∆PfhdaA mutants, the deletion cassette was amplified with the template of pYYJ1.1 using primers 10824-5f-FL and 10824-3f-RL. The DNA cassette fragments were transformed into P. fici WT, TYXW7.1 and TYXW8.1 as described previously [29]. Then candidate colonies were singled out after culturing on PDA with G418 resistant at 28 • C for 5 days. The disruption mutants were verified using diagnostic PCR with primers inside and outside of the gene PfptaA (Figure 1a and Table S2).

Oxidative Stress Sensitivity Assays
Different oxidative reagents were used to estimate the stress sensitivity of the mutants according to the method described previously [35]. Moreover, the colony diameters are the indicator supplementing with the following stress-generating agents: 0.5 mM diamide, 0.5 mM menadionesodium bisulfite (MSB), and 1.8 mM tert-butylhydroperoxide (tBOOH), respectively. The strains were incubated at 25 • C for 5 days, and colony diameters were measured on the 3rd day and 5th day. Three replicates were performed for each experiment.

Conidia Counting
The conidia counting in P. fici wild type and its mutants were performed as described previously [52]. These strains were grown at 25 • C on PDA medium in 90 mm plate for 14 days. Three replicates were performed for each culture sample. Then, we used ddH 2 O with 0.1% (v/v) Tween to flush the plates repeatedly and remove the hyphae and impurities through Miracloth. The filtrate containing conidia was centrifuged and concentrated to 1 mL of volume. The number of conidia from different mutants and wild type was determined using a blood-cell-counting plate. Values are means of three replicates for each culture are presented. Data were analyzed using the GraphPad Prism 8.0 performing Tukey-Kramer multiple comparison test at p ≤ 0.05. Asterisks indicate statically significant differences in mean values.

Analytical Methods for HPLC and LC-MS
Analysis of secondary metabolites was performed on a Waters HPLC system (Waters 2998, Photodiode Array Detector) with an ODS column (C18, 250_4.6 mm, Waters Pak, 5 µm). Water (A) and acetonitrile (B), both with 0.1 % (v/v) formic acid, were used as solvents at a flow rate of 1 mL/min. The substances were eluted with a linear gradient from 5-100% B in 40 min, then washed with 100 % (v/v) solvent B for 5 min and equilibrated with 5 % (v/v) solvent B for 5 min. UV absorptions at 236 nm were illustrated. LC-MS analyses of secondary metabolites was determined using an Agilent 1200 Accurate-Mass QTOF LC/MS system (Agilent Technologies, Santa Clara, CA, USA) with Agilent ZORBAX Eclipse column (C18 Plus, 2.1 × 4.6 mm, 3 µm) and an electrospray ionization (ESI) source. Water (A) and acetonitrile (B), both with 0.1 % (v/v) formic acid, were used as solvents at a flow rate of 1 mL/min. The substances were eluted with a linear gradient from 5-100% B in 40 min, then washed with 100 % (v/v) solvent B for 5 min and equilibrated with 5 % (v/v) solvent B for 5 min.

Isolation and Identification of New Compound
The ∆PfptaA ∆PfhdaA mutant was cultivated in flasks. Each flask contained 80 g rice and 120 mL distilled water and was sterilized by autoclave. A total of 10 kilograms of rice media were made, and static cultured at 25 • C for 20 d. The rice culture was extracted three times with ethyl acetate. The organic phase was evaporated to dryness under reduced pressure to afford the residue (7.6 g). The crude residue was applied on a C-18 ODS column using a stepped gradient elution of MeOH-H 2 O yielding 10 subfractions (fractions 1-10).

Conclusions
In summary, we developed a combined approach to modulate the secondary metabolic profile in filamentous fungi. We deleted the epigenetic regulators PfcclA or PfhdaA in a host lacking major SM pestheic acid. This led to significant changes in secondary metabolic profiles. We discovered 1 novel SM named pestaloficiol X (1) as well as 11 other known compounds with obvious yield changes in ∆PfptaA ∆PfcclA or/and ∆PfptaA ∆PfhdaA mutants. Compound 1 was only produced in the ∆Pfpta ∆PfhdaA mutant strain, suggesting a successful activation by modulation of histone acetylation and Pfpta dereplication. Moreover, the deletion of PfptaA in ∆PfcclA or ∆PfhdaA background host did not overproduce conidia as seen in the single mutant ∆PfptaA. We also found that the ∆PfptaA ∆PfhdaA mutant is generally not sensitive to oxidative stressors in comparison with the single mutant ∆PfptaA or the parental strains, whereas the ∆PfptaA ∆PfcclA mutant was more sensitive. Our findings support that combination strategy of genetic dereplication and manipulation of epigenetic regulators is an efficient approach to discover novel SMs in plant endophytic fungi P. fici as well as is a valuable strategy to be applied for new natural product discovery in filamentous fungi.