Alternaria sp. MG1 (code: CCTCC M 2011348) and Phomopsis sp. XP-8 (code: CCTCC M 209291), currently preserved in the China Centre for Type Culture Collection (Wuhan, China), were used in the study. Phomopsis sp. XP-8 is an endophytic fungus isolated from the bark of Tu-chung (Eucommia ulmoides Oliv.) by our group that produces lignins, including (+)-pinoresinol, (+)-pinoresinol monoglucoside, and (+)-pinoresinol diglucoside. The strains were cultivated in potato-dextrose broth (PDB) at 28°C and 160rpm for 7days. PDB medium was prepared with 20g potato, 2g dextrose, and 100ml distilled water.

The spore suspension (1×10⁷ spores/ml) of strain MG1 and XP-8 was separately prepared (Zhang et al., 2013) and co-cultured in PDB medium at the inoculation ratio of 5% each. After cultivation for 7days, the cells were collected from the culture by vacuum filtration and used for analysis of key genes expression. The cell-free cultures were then used for measurement of resveratrol production and the analysis of metabolite profile. For the control group, only 5% of MG1 or XP-8 was inoculated, and 3 replicates were set for each treatment.

Abiotic elicitors included physical and chemical factors. For physical induction, the 4-day culture of MG1 was separately treated with ultrasound (40kHz, 10min) and UV light (350nm; 10, 20, 30, 40min). For chemical induction, sodium butyrate was filtered through a 0.22μm hydrophilic membrane and then added to the 4-day culture of MG1 at final concentration of 100μm. OPC was kindly provided by Prof. Dongyan Shao (School of Life Sciences, Northwestern Polytechnical University). It was dissolved in dimethyl sulfoxide (DMSO) and added to the 4-day culture of MG1 at different final concentrations of 50, 100, 150, and 200μm of 1ml. To make sure whether the effect of OPC on the resveratrol biosynthesis is related to the solvent in which it is dissolved, 1ml of DMSO solution was also added to the MG1 culture as a separate treatment group. All treated cultures of MG1 were continuously cultivated to the 7th day. Then, cells were collected and stored for analysis of key genes expression. The cell-free cultures were used for the measurement of resveratrol production. Each treatment was performed in triplicate, and samples without elicitor treatment were always run in parallel as control.

Quantitative analysis of resveratrol was performed using HPLC (Water 2,695 system, Alliance, MA, United States) equipped with a WondaSil C18-WR column (5μm, 4.6mm×250mm, GL Sciences Inc., Japan). The cell-free cultures were extracted with equal volume of ethyl acetate three times (10h each time). The ethyl acetate phase was collected and vacuum concentrated in a rotary evaporator (Changzheng Instrument Manufacturing Co., Ltd., Zhengzhou, China) to dryness at 40°C. Then, the dried residue was dissolved in 2ml chromatographic grade methanol and filtered through a 0.22μm membrane (Solarbio, Beijing, China). The detection conditions for resveratrol were the same as that described before (Lu et al., 2019). The column was operated at 35°C. The mobile phase consisted of A: acetonitrile, B: ddH₂O. A multi-step gradient was used at a flow rate of 1ml/min according to the following steps: 0–28min, A: 5%, B: 95%; 28–33min, A: 60%, B: 40%; 33–40min, A: 85%, B: 15%; 40–50min, A: 5%, B: 95%. The detection wavelength was 306nm. The injection volume was 20μl.

Samples were prepared from 7-day growing cells with or without elicitor treatments. Total RNA was extracted, and cDNA was synthesized using commercial kits (Sangon Biotech Co., Ltd., Shanghai, China; Transgene Biotech Co., Ltd., Beijing, China). Primers of target genes are shown in Table 1. The reaction volume was set to 20μl in accordance with the operation manual of the TransStart Tip Green qPCR SuperMix Kit (Transgene Biotech Co., Ltd., Beijing, China). All reactions were conducted in triplicate. The program and qRT-PCR analysis were performed as previously described (Livak and Schmittgen, 2002). For each gene, the expression value was normalized with respect to the reference genes EF1 and TUBA and was calculated via 2^-ΔΔCT method (Che et al., 2016).

The published sequences of different CDPK proteins were chosen from UniProt database for multiple sequence alignment using the DNAMAN software package (version 7.0.2, LynnonBiosoft Company, United States). A phylogenetic tree was constructed using MEGA (Molecular Evolutionary Genetics Analysis) software version 7.0 with 1,000 bootstrap replicates. Sequences of the identified CDPK genes in MG1 are shown in Supplementary File 1.

All chemical data were presented as means ± standard deviation (SD) from triplicate determinations and processed with Origin 8.5. Multiple comparison analysis was performed by Duncan test using IBM SPSS Statistics 22.0 (SPSS Inc., Chicago, IL, United States). The significance level was set at p<0.05.

The production of resveratrol was detected at the seventh day under the controlled conditions using PDB medium at 28°C. From Figure 1, a product having the same retention time (25.582) as that of standard resveratrol (25.549), and its content was calculated to be around 160μg/l. To improve the production of resveratrol in MG1, different elicitors were selected and investigated. The results showed that ultrasound treatment did not significantly increase the production of resveratrol, but sodium butyrate had an inhibitory effect on its production. Co-culture with Phomopsis sp. XP-8 exhibited an obvious effect on the biosynthesis of resveratrol, which increased its production by approximately 40% (Figure 2A). In addition, the resveratrol production increased first and then decreased rapidly with the extension of UV irradiation time. It reached the highest value of 240.57μg/l at 20min, which increased by 45.8% compared with the control. The effect of different elicitors on the expression of key genes was also analyzed. In our previous study, a total of 5 PAL, 1 C4H, 6 4CL, and 1 CHS genes involved in the biosynthesis pathway of resveratrol were annotated in MG1 by genomic sequencing (Lu et al., 2019). Interestingly, no STS gene was annotated. Among these genes, CHS (Accession number: KX197069.1) was identified to synthesize both naringenin and resveratrol, while one of 4CL genes (Accession number: KX179648.1) synthesized p-coumaroyl CoA (Lu et al., 2021). Here, representative genes that have been functionally validated and those that are being functionally validated were selected for study. The results showed that under the same treatment condition, its effect on key genes expression was consistent with the trend of its effect on the production of resveratrol (Figure 2B). Among them, UV irradiation for 20min or co-culture with XP-8 significantly upregulated the expression of key genes of the resveratrol pathway, which was about 3–4 times that of the control.

Ultrasound is well-known reported to influence the biosynthesis efficiency of metabolites, growth rates, and enzymatic activity by improvement of cell permeability as well as mass transfer and nutrient uptake rates through cell membranes (Behzadnia et al., 2020). Valente et al. found that the antioxidant activity of exocellular metabolites of Botryosphaeria dothidea after ultrasound treatment was higher than the control, reaching a maximum value of 96% (Valente et al., 2020). In addition, mild intensity ultrasound (low energy and short exposure time) could significantly promote anticoagulant accumulation by marine Bacillus subtilis ZHX (Cao et al., 2020) and mycelial growth and exopolysaccharides (EPS) biosynthesis from Agaricus bitorquis (Quél.) Sacc. Chaidam (Lu et al., 2020). In the current study, ultrasound exposure (40kHz) for 10min caused no significant increase of resveratrol. This might be because resveratrol biosynthesis of MG1 is not sensitive to ultrasound disturbance. Similar result could be also found in the case of production of bioactive valepotriates in Valeriana glechomifolia (Russowski et al., 2013). Sodium butyrate is widely reported to show eliciting activity as a histone deacetylases inhibitor, which enables silent or cryptic gene clusters to express. The yield of enzymes and secondary metabolites from fungi could be increased, and new metabolites like phenolic compounds are generated after sodium butyrate treatment (El-Hawary et al., 2018; Mohammadipanah et al., 2020). However, 100μm of sodium butyrate had an adverse influence on the biosynthesis of resveratrol in MG1, which may be related to the inhibition of fungal growth by sodium butyrate. The biomass of some fungi decreased significantly after being treated with sodium butyrate, such as Phomopsis and Trichosporon spp. (Zhu et al., 2018; Cordeiro et al., 2019). Therefore, it is speculated that 100μm of sodium butyrate used in this study inhibited the biomass of MG1, which lead to insufficient supply of some important precursors of resveratrol, such as phenylalanine and tyrosine. It is noteworthy that the elicitor potential of sodium butyrate could be concentration-dependent (Fatemeh et al., 2020). Therefore, the failure to elicit the production of resveratrol at a single concentration of 100μm did not mean that its production under other concentrations or the production of other metabolites was not affected. A further dose–response experiment is required regarding the potency of this elicitor. UV radiation is an important mutagen for high-yielding strains. For example, strain Cladosporium phlei M0035 resulted in a total of 592mg/l phleichrome after UV-mutagenesis, more than seven-fold over wild type (Yi et al., 2011). Similarly, the highest amount of monacolin K produced by the mutated strain of Monascus spp. was 3 times greater than the control after the irradiation of UV light in the presence of 1.0‰ LiCl in the medium (Sun et al., 2011). In addition, UV radiation is also commonly used to increase the production rate and the catalytic activity of enzymes of fungi. The production of resveratrol synthesized by MG1 gradually increased with the increase of UV treatment time, which may be a self-adaptive protection to adverse environments. When the treatment time increased to 30min or even longer, it may damage the cell structure and enzyme activity, resulting in the inhibition of resveratrol biosynthesis.

In nature, fungi lives in association with other organisms and this constant interaction affect the quality and quantity of secondary metabolites produced. Co-culture of different bacteria, bacteria and fungi as well as different fungi shows great advantages in improving the production of natural products with pharmaceutical properties, because they have similar pathways and could provide each other with complementary key genes for their biosynthesis. Co-cultivation of Paraconiothyrium SSM001 and Alternaria increased the production of taxol in SSM001 by 3 times (Soliman and Raizada, 2013). Similar studies could also be found in the enhancement of secondary metabolites subenniatins A and B, fusaristatin A, podophyllotoxins, and stemphyperylenol (Baldi et al., 2010; Chagas et al., 2013; Ola et al., 2013; Wang et al., 2013). Strains Alternaria sp. MG1 and Phomopsis sp. XP-8 are both plant endophytic fungi that isolated by our group in the previous study, so they were used to investigate the effect of co-culture system on the biosynthesis of resveratrol in MG1. The production of resveratrol in MG1 increased by 40% when co-cultured with XP-8, revealing microbial interaction may induce the expression of related silent gene clusters. Therefore, the metabolite composition of the culture broth from MG1 was analyzed by GC–MS to further clarify the effect of co-culture with XP-8 on the metabolite profile of MG1. As shown in Figure 3, the chromatographic peaks of different components exhibited good separation, and most of the peaks were well separated from the baseline. A total of 511, 1,458, and 1,535 distinct peaks were detected in the chromatograms for the culture broth of MG1, co-culture of MG1 and XP-8 (MG1 + XP-8), and XP-8, respectively. Then, after preprocessing of raw peaks with Chroma TOF software, strict quality control procedures, and peak identification, a total of 379, 832, and 875 metabolites were identified with a similarity above 200 in MG1, MG1 + XP-8, and XP-8, respectively.

As MG1 and XP-8 are both plant endophytic fungi and have similar metabolic pathways, we used the method of peak area normalization to only quantify the relative contents of unique secondary metabolites (from stilbene and flavonoid biosynthesis pathways) in MG1. As shown in Table 2, after co-culture with XP-8, the accumulation of resveratrol and piceatannol in MG1 increased, which was consistent with the above results, while naringenin, coniferyl alcohol, and sinapic acid were not detected compared with before co-culture. This indicated that co-culture promoted the synthesis of stilbene compounds in MG1 but completely inhibited the formation of naringenin and lignin compounds.

The changes of metabolites in MG1 culture broth after co-cultivation were further analyzed. A total of 65 new compounds were generated (Figure 4A), among which the more important secondary metabolites included: (−)-dihydrocarveol, salicylic acid, acetylsalicylic acid, 4-hydroxymandelic acid, caffeic acid, squalene, (+)-catechin, (+/−)-taxifolin, naringin, tricetin, and methyl jasmonate (Figure 4B). These chemicals exhibit a variety of biological activities and are widely used as fragrances, antioxidants, and drugs, etc. It is worth noting that (+/−)-taxifolin, naringin, and (+)-catechin are all belong to flavonoids. However, as mentioned above, their common precursor naringenin in MG1 was not detected after co-cultured with XP-8, indicating that the synthesis of naringenin was not actually inhibited, but was completely converted to different flavonoid compounds after co-culture. The results showed that co-culture with XP-8 promoted the further transformation of naringenin in MG1. Moreover, a total of 75 compounds were not detected in MG1 after co-cultivation with XP-8. This phenomenon, the complete bioconversion of a certain metabolite induced by co-culture, is commonly found in other studies. Eudistomin I is a B-carboline alkaloid, and its synthesis begins with tryptamine as a starting material. Rafidaf et al. (2020) found that the tryptamine that was detected in the control group was used up by the co-culture fungi and involved in the biosynthetic pathway to produce eudistomin I, which explains why tryptamine was not detected in the treatment group. In addition, three O-glycosylated isomers of naringenin dropped, with two of these reaching non detectable levels in okara fermented with co-culture strains of lactic acid bacteria (Jasmine et al., 2021). However, the conclusion needs to be supported by investigating whether the expression of genes involved in the synthesis of downstream flavonoids in MG1 is upregulated.

Proanthocyanidin is an important nutrient component in grape fruit and wine, and its existing form is mainly OPC. Endophytic fungi parasitizing on grapes (like skin, cob, and stem) inevitably enter the fermentation process with the crushing and pressing of grape raw materials, and their growth and physiological metabolism are affected by the main components of the wine. Therefore, in order to explore the influence of OPC on the biosynthesis of resveratrol in MG1, different concentrations of OPC were used to treat the MG1 cells. As shown in Figure 5A, the response of resveratrol biosynthesis to OPC treatment was dose-dependent. That was, its production increased with the increase of OPC concentration and reached the highest value (255.01μg/l) at 100μm of OPC, which was an improvement of 60.63% compared to the control, while its production decreased gradually when the OPC concentration was higher than 100μm and was significantly inhibited at 200μm of OPC. The effect of OPC treatment on the expression of key genes in the resveratrol pathway was basically consistent with the trend of its effect on the resveratrol yield (Figure 5B). However, at 200μm of OPC, the expression levels of PAL, C4H, and 4CL were all significantly upregulated, about 2–2.5 times that of the control, while the expression level of CHS was obviously downregulated. This indicated that the reaction catalyzed by the enzyme encoded by CHS gene was the rate-limiting step of the resveratrol pathway, leading to no significant increase in its yield. This conclusion was also valid with the corresponding relationship between the expression of CHS gene and resveratrol production under other OPC concentration conditions.

Many plant-derived secondary metabolites have been reported to react with protein kinases involved in the signal transduction of eukaryotic cell, thereby causing a series of biological responses (Molan et al., 2000). Ca²⁺ is an important second messenger involved in plant cell signal transduction. When plants are exposed to external stresses, the Ca²⁺ signals in the cell are transmitted to the downstream components through Ca²⁺ sensors to elicit the expression level of related genes, so as to respond to various environmental changes (Romeis and Herde, 2014). Calcium-dependent protein kinase (CDPK) is one of the important proteins involved in the regulation of Ca²⁺ and was found to cause significant accumulation of resveratrol when overexpressed in transgenic plant cells (Aleynova-Shumakova et al., 2014). CDPK proteins (CDPKs) mediate various Ca²⁺ signalings through CDPK cascades in response to various extracellular stimuli (Isamu et al., 2007). Therefore, as an important plant-derived secondary metabolite, whether OPC can activate Ca²⁺ channels by interacting with Ca²⁺ signal-related proteins, such as CDPK proteins, thereby activating the upregulation of downstream component (resveratrol) and related genes to resist stress environment caused by OPC, are not clear. Moreover, it is unclear whether there are CDPK proteins in MG1. To uncover the possible regulatory mechanism between OPC, CDPK protein and resveratrol biosynthesis in MG1, the expression level of CDPK protein under different OPC concentrations was investigated. Through mining the genomic data of MG1, two CDPK genes were found, named CDPK1 and CDPK2 (Supplementary File 1). The CDPK1 sequence consisted of 1,344bp encoding 447 amino acid resides and CDPK2 sequence consisted of 1,200bp encoding 399 amino acid resides. According to the results of multiple sequence alignment, two CDPK proteins from MG1 had high identity with other reported CDPKs from different species, among which CDPK1 and CDPK2 showed the identity of 81.65 and 65.69%, respectively. In addition, the results also revealed the presence of activation loop motif, ATP binding site, and calmodulin-binding domain of two CDPK proteins (Figures 6A,B), which are considered signature motifs of CDPK gene. MEGA7 phylogenetic analysis based on protein sequences showed that CDPK1 and CDPK2 were most closely related to proteins from Alternaria gaisen (KAB2111357.1) and Exserohilum turcicum (B2M1T6), respectively (Figures 6C,D). Interestingly, the representative motif of CDPK protein varies significantly among different species. For example, the identified calmodulin-binding domain (VKKNFNARRTLHAAIDTIRAINQLRA) of CDPK2 in MG1 was consistent with that in Arthrobotrys dactyloides, but completely different from those in Sporothrix schenckii (KARLRRGIELVKLTNRIEAL) and Coprinopsis cinerea (KDAVAVVRSILS, WKSAIASARALNR) (Tsai et al., 2002; Valle-Aviles et al., 2007; Masashi et al., 2018). However, no calmodulin-binding domain was identified of CDPK1 protein in MG1. In plants, the typical active domain is called EF-hand calcium-binding motifs, which generally has four. They are predominant calcium sensor (Wang and Shao, 2013).

It can be seen from Figure 7 that different concentrations of OPC had no effect on the expression of CDPK2, while the expression of CDPK1 first increased and then decreased with the increase in OPC concentration and reached the maximum at 100μm OPC, which was about 4.5 times that of the control. In addition, the trend of CDPK1 expression was positively correlated with that of resveratrol production. Therefore, the influence of OPC on the resveratrol biosynthesis in MG1 may be related to the regulation of CDPK1, and the specific mechanism needs to be further analyzed. Although little is known about the regulated mechanism of CDPKs in fungi, many of the signal-induced responses have been well documented in controlling spore germination and differentiation, cell division, and protein sorting in the secretory pathway (Tsai et al., 2002).

To further improve the production of resveratrol, UV light combined with OPC was employed to treat MG1 cells. As shown in Figure 8, synergistic treatment increased the production of resveratrol by 70.37% and finally reached 276.31μg/l. In our previous study, ethanol was found to have a positive effect on resveratrol synthesis due to its role as a precursor and carbon source. It increased the production of resveratrol in a concentration-dependent manner, with the highest increase by 26.31% (Lu et al., 2019). In addition, salicylic acid (SA) and methyl jasmonate (MeJA) were reported to be related to phenylpropane pathway genes toward accumulation of stilbenes and flavonoids. Results showed that SA induction resulted in a 33.32% increase in resveratrol production, whereas MeJA did not affect the resveratrol production nor the expression of related genes. Compared with all single factor treatments, combination treatment of UV light and OPC showed great potential in promoting the biosynthesis of resveratrol in MG1. This strategy is also used to increase the production of resveratrol in other microorganisms. Villa-Ruano et al. (2021) found that citral and Glucanex (mixtures of hydrolytic enzymes) produced up to 225.1mg/l of resveratrol in Arcopilus aureus MaC7A, which was significantly higher than that from either citral (187.8mg/l) or Glucanex (198.3mg/l). And, a final production of 237.6mg/l was obtained by combination of tyrosine (200mg/l), citral (50mg/l), thymol (50mg/l), and Glucanex (100mg/l) in batch cultures. Therefore, synergistic effect of multiple elicitors is an effective way to increase the yield of secondary metabolites derived from plants or microbes and is widely employed in other studies. The combination of yeast extract (YE) and Ag+(300 micromol x L(−1)) or and Co2+ (100 micromol L(−1)) led to the highest tanshinone I and tanshinone IIA content, which was nearly 14-fold and 14.5-fold of the control, respectively (Yan et al., 2006). Serial treatment of methyl jasmonate (MJ), salicylic acid (SA), and YE at 24-h intervals enhanced the accumulation of dihydrosanguinarine (2.5 times) and sanguinarine (5.5 times) in Eschscholtzia californica suspension cultures (Cho et al., 2008). This strategy was more effective than single elicitors. In addition, the combined elicitors of MJ, jasmonic acid, and chitosan synergistically promoted greater valtrate production than each individual elicitor addition. It was 1.4-fold higher than the maximum increase obtained by individual elicitor addition (10.58mg/g with 100mg/lMJ).