The WT strain of C. chrysosperma (CFCC 89981) used in this study was preserved in the forest pathology laboratory of Beijing Forestry University (Fan et al., 2020). Strains used in this study were regularly cultured on potato dextrose agar medium (PDA; 20% potato extract, 2% glucose, and 1.5% agar) at 25°C. For DNA isolation and RNA isolation, mycelia were cultured in liquid potato dextrose broth medium (PDB; 20% potato extract and 2% glucose) for 2 days at 150 rpm, 25°C. Populus bark broth medium (PBB; 30% 1-year-old polar branch extract and 1% glucose) was used for the extraction of crude secondary metabolites.

The sequence of terpene type gene cluster (GME3317_g to GME3324_g) were acquired from the draft genome sequence of C. chrysosperma, which had been sequenced by our laboratory (NCBI GenBank accession number JAEQMF000000000). The genome sequences of the terpene type gene cluster and the corresponding CDS sequences were committed to the Gene Structure Display Server (GSDS)¹ to analyze the number and alignment of introns and exons. TBtools (V0.66836) was used to visualize the relative expression level (Chen et al., 2020). The homologs of this terpene type gene cluster were searched in the genome of other microorganisms in the JGI database. The domain structures of terpene type gene cluster were annotated using the InterProScan tool.² In addition, phylogenetic analysis was conducted with MEGA 10.0 software using the full-length protein sequences and neighbor-joining method with 1,000 bootstrap replications.

The split marker method was used to construct the CcPtc1 gene deletion (ΔCcPtc1) mutants as previously described (Catlett et al., 2003; Goswami, 2012). According to this method, the upstream (∼1.3 kb) and downstream (∼1.2 kb) flanking sequences of CcPtc1 were amplified by primer pairs CcPtc1-5Ffor/CcPtc1-5Frev and CcPtc1-3Ffor/CcPtc1-3Frev, respectively (Supplementary Table 1). The hygromycin B resistance cassette (HPH) was amplified by specific primer pairs hygromycinfor and hygromycinrev, which included approximately 20 bp overlapped 5′ and 3′flanking sequences, respectively. Then, the resulting upstream and downstream fragments were fused with two-thirds of the hygromycin B resistance cassette by overlap PCR using primer pairs CcPtc1-5Ffor/HY-R and YG-F/CcPtc1-3Frev, respectively. The two overlapping fragments were directly transformed into protoplasts of the WT strain by using the PEG-mediated transformation, and the transformants were selected on TB₃ agar medium supplemented with 20 μg/ml hygromycin B. All transformants were identified by PCR assays with the primer pairs External-CcPtc1for/External-CcPtc1rev and Internal-CcPtc1for/Internal-CcPtc1rev to screening the successful replacement transformants. In addition, to analyze homologous recombination events in the transformants, southern blotting analysis was performed with the DIG High Prime DNA Labeling and Detection Starter Kit I, following the manufacturer’s protocol (Roche, Germany). BglI was used to digest the genomic DNA extracted from the WT strain and the transformants. The probes were amplified by the primers ProbeHPHfor and ProbeHPHrev from HPH and the primers ProbeCcPtc1for and ProbeCcPtc1rev for CcPtc1.

To generation the CcPtc1 gene complementation construct, a fragment containing the entire length of the CcPtc1 coding region along with native promoter sequence and terminator sequence was cloned from gDNA using the primer pair CcPtc1-Compfor/CcPtc1-Comprev. The resulting PCR products were co-transformed into protoplasts of the ΔCcPtc1-11 strains with a geneticin-resistant cassette. After that, we selected the transformants in TB₃ medium supplemented with 40 μg/ml geneticin. Successful complementation was confirmed by PCR with the primer pair Internal-CcPtc1for/Internal-CcPtc1rev. The complementation strain was named ΔCcPtc1/PTC1 in this study. All primers used in gene deletion and complementation were listed in Supplementary Table 1.

To analyze the differences in vegetative growth and conidiation among the WT, deletion mutants, and complemented strains, each strain was inoculated on the PDA plates at 25°C in the dark. The growth and conidial formation were observed at 24, 48, and 60 h and 30 days post inoculation. Each experiment was replicated at least three times.

Healthy annual poplar twigs collected from the nursery garden in the Beijing forestry university were used for pathogenicity assay. Hyphal plugs of the WT, gene deletion mutants, and complemented strains were inoculated on the 20-cm-long twigs which scorched with a flat iron (5 mm in diameter). The inoculated twigs were incubated at 25°C for 5∼8 days under moist conditions and photographs were taken after 5 days. The experiments were repeated at least three times.

For analysis of metabolite production, the plugs of WT and ΔCcPtc1 were cultivated into PBB medium at 25°C and 150 rpm for 10 days. Culture filtrates of WT, ΔCcPtc1, and PBB medium were extracted with equal volume of ethyl acetate for three times, and the organic phase was saved. Organic phases were combined to obtain crude extracts of secondary metabolites. The organic phases were distilled using a rotary evaporator under 45°C and re-dissolved in 20 ml Dimethyl sulfoxide (DMSO). The phytotoxic activities of the crude extracts were tested on polar leaves using the leaf puncture method. PBB crude extract–treated and DMSO-treated samples were kept as control. Inoculated leaves were incubated in a petri dish under 25°C. The pictures were taken at 24 h after treatment. The experiment was performed with three biological replicates at least.

To analyze the expression levels of other genes among the terpene type gene cluster in the CcPtc1 deletion mutant, the WT, ΔCcPtc1-11 and ΔCcPtc1-14 strains were cultivated in PDB supplemented with sterilized poplar twigs at 25°C and 150 rpm for 48 h. Mycelium was harvested by Miracloth (Calbiochem). Then flash-frozen in liquid nitrogen and ground to powder. Total RNA was extracted from powder using RNA Easy Fast Plant Tissue Kit (TIANGEN, China) according to the manufacturer’s protocol. First-strand cDNA was prepared using ABScript II cDNA Fist-Strand Synthesis Kit (ABclonal, China).

Quantitative real-time PCR was performed with SuperReal PreMix Plus (ABclonal, China) on the Applied Biosystems 7500 Real-Time PCR system (Applied Biosystems). The CcActin gene of C. chrysosperma was used as an internal reference for all RT-qPCR experiments. Relative expression was calculated using the 2^{–Δ Δ Ct} method. The experiment was performed in triplicate with three independent technical replicates each. All primers used in the present study were listed in Supplementary Table 1.

To prepare the metabolome samples, the WT and ΔCcPtc1 strains was simultaneously cultured in PBB medium with the same conditions as crude extraction assay. Each strain contained six repeats and generated twelve datasets.

All samples were placed in the EP tubes and resuspended with prechilled 80% methanol by well vortex. After whirled (melted on ice for 30 s), sonification (6 min), and centrifuged (5,000 rpm, 4°C for 1 min), the supernatant of all samples was freeze-dried and dissolved with 10% methanol. Finally, the solution was injected into the LC-MS/MS system for further analysis. UHPLC-MS/MS was used to analyze the composition of the WT and ΔCcPtc1 strains using a Vanquish UHPLC system (Thermo Fisher, Germany) coupled with an Orbitrap Q ExactiveTM HF mass spectrometer (Thermo Fisher, Germany) in Novogene Co., Ltd. (Beijing, China). Samples were injected onto a Hypesil Gold column (100 × 2.1 mm, 1.9 μm) with a 12-min linear gradient at a flow rate of 0.2 mL/min. Q ExactiveTM HF mass spectrometer was operated in positive/negative polarity mode with spray voltage of 3.5 kV, capillary temperature of 320°C, sheath gas flow rate of 35 psi and aux gas flow rate of 10 L/min, S-lens RF level of 60, Aux gas heater temperature of 350°C.

The raw data generate by UHPLC-MS/MS were processed using the Compound Discoverer 3.1 (CD3.1, Thermo Fisher) to perform peak alignment, peak picking, and quantitation for each metabolite. The KEGG database,³ Human Metabolome Database (HMDB) database,⁴ and LIPIDMaps database⁵ were used to annotate these metabolites. Principal components analysis (PCA) and Partial least squares discriminant analysis (PLS-DA) were performed at metaX. We applied univariate analysis (t-test) to calculate the statistical significance (p-value). Metabolites with Variable Importance in Projection (VIP) > 1, fold change (FC) > 1.2 or FC < 0.833, and p-value < 0.05 were considered to be DAMs. The functions of these metabolites and metabolic pathways were annotated using the KEGG database. Enrichment metabolic pathways of differential metabolites were selected with x/n > y/N and p < 0.05.

Our previous works found that two putative secondary metabolism related gene clusters (GME3437_g to GME3444_g and GME3317_g to GME3324_g) were regulated by the key pathogenicity factor CcPmk1 and CcSlt2. In this study, the backbone gene (GME3321_g) of gene cluster from GME3317_g to GME3324_g was functional characterized. GME3321_g contained a terpene cyclase-like 2 protein family domain (IPR 034686), therefore, we named it as CcPtc1 (Figure 1A). Additionally, GME3317_g, GME3322_g, and GME3323_g contains a conserved cytochrome P450 domain (IPR 001128), respectively (Figure 1A). The structure of the 8 genes among these terpene type gene cluster were characterized using the GSDS, which contained 0 to 7 introns (Supplementary Figure 1). In addition, we aimed to determine the putative homologs of this gene cluster in other fungal species. Therefore, the whole sequence of this gene cluster was used as queries to search through the genome sequences of Magnaporthe oryzae, V. dahliae, Botrytis cinerea, Cryphonectria parasitica, Saccharomyces cerevisiae, Sclerotinia sclerotiorum, and Fusarium species. Nevertheless, no significant hits were obtained in these fungal species.

Subsequently, we only identified the homologs of backbone gene CcPtc1 in other fungal species, and many homologs were found because of the conserved sequence of terpene cyclase-like 2 protein family domain. Homologos protein sequences of CcPtc1 were downloaded from four fungal species, including M. oryzae, Fusarium oxysporum, F. graminearum, and C. parasitica, to produce a phylogenetic tree (Figure 1B). The information of all these proteins were listed in Supplementary Table 2. Moreover, we analyzed the sequences and protein domain (IPR 034686) of terpene cyclase-like 2 proteins in these four fungi (Figure 1C). The overall amino acid identity of CcPtc1 to the Magor1| 2188, Fusoxalb1| 706066, and Fusgra3| 1251088 is with an identity of 44, 48, and 45%, respectively (Supplementary Table 2).

To investigate the expression of terpene type gene cluster during infection processes, we collected the expression data for these cluster genes from our previous transcriptome data of the initial infection process [0 and 1 days post-inoculation (dpi)] in poplar branches (Li et al., 2021). The expression levels of genes in this terpene type gene cluster were shown in Figure 2A, and half of the genes (4/8) were significantly upregulated at 1 dpi compared to 0 dpi. We then analyzed the expression levels of these cluster genes during the middle and late C. chrysosperma infection process (7 dpi and 15 dpi) of poplar branches. As shown in Supplementary Figure 2, most of the genes (7/8) were significantly upregulated at 7 dpi compared to 0 dpi, and 3 out of 8 genes were significantly upregulated at 15 dpi compared to 0 dpi. Notably, the expression of the backbone gene (CcPtc1) was significantly up-regulated during the whole infection processes.

In order to reveal the functions of CcPtc1 during the infection processes, we generated two CcPtc1 deletion mutants ΔCcPtc1-11 and ΔCcPtc1-14, which had been confirmed by the PCR and Southern blot analysis (Supplementary Figures 3A, B). For the complementation of ΔCcPtc1, a fragment containing the native promoter and the CcPtc1 coding sequence was isolated from genomic DNA and transformed into the ΔCcPtc1-11 strain and screened by PCR (Supplementary Figure 3C). One complemented strain (ΔCcPtc1/PTC1) and two mutant strains (ΔCcPtc1-11 and ΔCcPtc1-14) were used for subsequent phenotypic analyses.

To determine the effects of CcPtc1 on the other genes among its cluster (GME3317_g to GME3324_g), we analyzed the expression levels of these genes in the WT and ΔCcPtc1 strains (Figure 2B) in mimetic infection process. The qRT-PCR assays show that, 5 out of 7 genes in this cluster exhibited significantly reduced expression levels in the CcPtc1 deletion mutants compared with those in the WT strain. These results suggest that this terpene type gene cluster, especially the backbone genes CcPtc1, may play important roles in the fungal pathogenicity of C. chrysosperma.

To determine the potential roles of CcPtc1 in C. chrysosperma vegetative growth, we cultivated the WT, CcPtc1 deletion mutants and complemented strains on PDA plates at 25°C for 60 h. Development phenotypic analysis showed that ΔCcPtc1 exhibited an apparent smaller colony diameter compared with the WT, and the growth defects of the knockout strains were restored in complementation strain (Figures 3A, B). Compared with the WT and ΔCcPtc1/PTC1 strains, the CcPtc1 deletion mutant showed approximately 17.13% reduction in hyphal growth on PDA plates.

To investigate whether CcPtc1 contributed to conidiation, we then calculated the number of pycnidium produced by the WT strain and CcPtc1 deletion mutants. The results suggest that ΔCcPtc1 produced significantly decreased amount of pycnidia compared to WT strain on PDA after 30 days (a reduction of approximately 34.5%) (Figures 3A, C). Taken together, these results suggest that CcPtc1 is required for conidiation and fungal growth.

As mentioned above, the expression level of CcPtc1 was significantly increased during the early infection stages. Thus, it prompted us to explore the roles of CcPtc1 in fungal virulence. We performed a pathogenicity test on detached poplar twigs with inoculating mycelial plugs of the WT, CcPtc1 deletion mutants, and complemented strains. As shown in Figure 4A, poplar twigs inoculated with WT strains exhibited severely typical symptoms at 4∼7 dpi, while the poplar twigs inoculated with ΔCcPtc1 deletion mutants showed significantly reduced lesion areas (over 70% reduction). In addition, the complemented strain ΔCcPtc1/PTC1 displayed a comparable lesion size as the WT strain (Figures 4A, B). Overall, these results demonstrate that the CcPtc1 contributes to fungal virulence during the infection process.

CcPtc1 was predicted as the backbone gene of the secondary metabolite gene cluster and it was required for fungal virulence, thus we speculated whether the impaired pathogenicity may result from the reduced production of secondary metabolites. To investigate the role of CcPtc1 in the production of secondary metabolites, toxicity tests were conducted. We collected culture filtrates of WT strain, ΔCcPtc1 strain, and uninoculated medium to obtain crude extracts of secondary metabolites. Toxicity of these crude secondary metabolites was tested on poplar leaves.

Infiltration the crude extracts of WT and ΔCcPtc1 into poplar leaves with leaf puncture method produced a brown ovoid region of necrosis which extend over the entire infiltration zone at 24 h after treatment (Figure 5A). On the contrary, the DMSO solvent and uninoculated medium control did not produce any necrotic symptom. Although necrotic symptoms were similar across WT and CcPtc1 deletion mutants, crude secondary metabolites from WT were more effective than those from ΔCcPtc1, the area of the necrotic lesions inoculated with the ΔCcPtc1 was reduced by approximately 50.3% compared with the WT strain (Figures 5A, B).

These findings verify that C. chrysosperma was likely to produce toxic secondary metabolites which cause damage on host plant tissue, and deletion of CcPtc1 may lead to reduction toxicity of secondary metabolites.

To clarify the differences in metabolites after the deletion of CcPtc1, six biological individuals for WT or ΔCcPtc1 strain were analyzed by mass spectrometry-based metabolomics, respectively. High Pearson correlation values between the quality control samples were obtained, indicating high data stability during LC-MS/MS analysis (Figure 6A). The markedly separation between WT group and ΔCcPtc1 group was validated by PCA with the first principal component (PC1) value of 29.52% and the second principal component (PC2) accounted for 11.64% of the variation (Figure 6B). The PCA analysis revealed that the metabolites of ΔCcPtc1 and WT strains were significantly different.

Here, 725 metabolites (data have been de-duplicated) were detected with varied abundance in both strains including 550 metabolites in positive polarity mode (Figure 6C) and 175 metabolites in negative polarity mode (data not shown). Among them, the abundance of 193 metabolites were significantly changed including 90 metabolites significantly decreased their abundance and 103 metabolites significantly increased their abundance (Figure 6D). According to KEGG, HMDB, or LIPID MAPS databases, these 193 DAMs were categorized to the following major classes: biosynthesis of secondary metabolites (14 annotations), nucleosides, nucleotides, and analogs (4 annotations), organic acids and derivatives (14 annotations, 1 annotation was included in biosynthesis of secondary metabolites), carbohydrates and carbohydrate conjugates (10 annotations, 1 annotation was included in biosynthesis of secondary metabolites), and lipids and lipid-like molecules (38 annotations, 4 annotations were included in biosynthesis of secondary metabolites) (Figure 7A and Supplementary Table 3). Importantly, lipid and organic acids were the category that enriched by most of DAMs in response to CcPtc1 deletion. These observations suggest that different abundant metabolites are resulted from CcPtc1 deletion and might be the agents for CcPtc1 to modulate fungal phenotypes.

Functionally annotation of the 193 DAMs revealed that four metabolic pathways of the C. chrysosperma were identified including starch and sucrose metabolism, linoleic acid metabolism, pantothenate and coenzyme A (CoA) biosynthesis, and biosynthesis of unsaturated fatty acids (Figure 7B).

The schematic map of significant enrichment metabolic pathways and changes level of key metabolites identified in the mycelia culture filtrates was shown in Figure 8. The key metabolites included trehalose 6-phosphate, D-Fructose 6-phosphate, arachidonic acid, 9-KODE, palmitic acid, pantothenate, and pantothenol were significantly influenced by CcPtc1 (Figure 8). As shown in Figure 8, deletion of CcPtc1 resulted in a decrease in the content of D-Fructose 6-phosphate, pantothenate and palmitic acid, and an increase in the content of trehalose 6-phosphate, 9-KODE, arachidonic acid, and pantothenol.

Moreover, to explore the changes of terpenoids in ΔCcPtc1 mutants, we further analyzed the metabolomic data and identified 17 terpenoid related metabolites (Supplementary Table 4), which could be categorized into four major categories: diterpenoids, monoterpenoids, sesquiterpenoids, and terpene glycosides. Among the 17 terpene metabolites detected, the number of monoterpenoids and sesquiterpenoids was the largest, of which 9 metabolites belonged to monoterpenoids and 5 metabolites belonged to sesquiterpenoids (Figure 9A and Supplementary Table 4). Further analysis revealed that 6 terpenoid related metabolites were significantly alterations in the ΔCcPtc1 compared to the WT, including down-regulated metabolites, (+)-ar-turmerone, pulegone, ethyl chrysanthemumate and genipin, and up-regulated metabolites cuminaldehyde and (±)-abscisic acid (Figure 9A).

We focused on the biosynthesis of pantothenate and CoA which are significantly enriched in positive polarity mode. Remarkably, we discovered that early intermediates of the pantothenate and CoA biosynthesis pathway (pantothenate and pantothenol) were significant changed in ΔCcPtc1: the amounts of pantothenate was significant down-regulated, and in contrast, the amounts of pantothenol was remarkably up-regulated compared to the WT strain. Furthermore, the ΔCcPtc1 strain showed significant perturbations of pantothenate. Previous studies have showed that pantothenate plays dominant role in regulation of carbohydrate, lipid, and nucleic acid metabolism. As shown in Figure 6 and Supplementary Table 3, the lipid metabolites accounted for the largest number (27 out of 79) (Figure 6 and Supplementary Table 3). Besides, a certain amount of carbohydrates metabolites was also enriched (7/79) (Figure 6 and Supplementary Table 3). Importantly, the lipid (16/27) and carbohydrate (4/7) metabolites were generally reduced in CcPtc1 mutant. It is suggested that deletion of CcPtc1 could reduce the accumulation of pantothenate, which resulted in a significant changed in the content of lipid and carbohydrates metabolites (Figure 9B).

In this study, we found that CcPtc1 markedly affected fungal development, pathogenicity, and toxic secondary metabolites in C. chrysosperma. Taken together, these results shows that CcPtc1 acts as a crucial virulence factor.