In this study, Sarocladium terricola was isolated from soil in Anhui Province, China, and deposited in the Anhui Agricultural University with the accession number RCEF 6201. The morphological trait is mainly as followed: Mycelia hyaline, 1.5–1.9 µm wide; Phialides subulate, solitary, 12–24 µm long, 2–3 µm wide at the base; Conidia fusiform, hyaline, sharply, 7–9 × 3–4 µm; Chlamydospores not observed.

Strains were incubated with potato dextrose agar media (PDA: potato 200 g/L, glucose 20 g/L, agar 20 g/L, and 1 L distilled water) at 25°C for 1 week to produce spores and then collected using sterile water. The spore suspension was adjusted to a concentration of 1 × 10⁶/mL, then 1 mL spore suspension was inoculated into a 500 mL flask containing either 200 mL potato dextrose media (potato 200 g/L, glucose 20 g/L, and 1 L distilled water) or bean media (bean 30 g/L, glucose 20 g/L, and 1 L distilled water). Finally, the two types of media including HCBe (cluturing in bean media) and HCPo (cluturing in potato dextrose media) with difference in nitrogen source were incubated at 25°C with 150 rpm shaking for three days, and ergosterol was then detected using high performance liquid chromatography (HPLC). Three biological replicates were performed for each type of media.

We referred and improved the previous study (Nout et al., 1987): after three days, fresh mycelium was collected using a circulating water multi-purpose vacuum pump and placed in an oven at 45°C and allowed to equilibrate. To extract ergosterols, 0.5 g mycelium was mixed with 50 mL methanol in a 100 mL flask, disrupted with ultrasonic waves at 60°C for 30 min, and then centrifuged at 4000 g for 5 min to obtain supernatant. As a standard, 0.01 g ergosterol was placed in a 100 mL volumetric flask with methanol solvent added until volume reached 100 mL.

The ergosterol profiles were assayed using HPLC with a C18 column (250 mm × 4.6 mm × 5 µm). The analytical conditions were as follows: solvent system, anhydrous methanol; temperature, 35°C; injection volume, 10 µL; flow rate, 1 mL/min; and UV wavelength, 268 nm. Three technical replicates were assayed. The amount of ergosterol in the sample was determined by mass fraction (ω, unit is expressed in mg/kg), and the formula used for calculation was as follows:

A: peak area of ergosterol in sample.

As: peak area of ergosterol in standard working solution.

ρS  : mass concentration of ergosterol in standard working solution (unit: μg/mL).

V: final constant volume of sample solution (unit: mL).

m: mass of the test portion (unit: g).

f: sample dilution ratio.

S. terricola was incubated in potato dextrose agar media (PDA: potato 200 g/L, glucose 20 g/L, agar 20 g/L, and 1 L distilled water) and cultured for two weeks. Total cell DNA was extracted from mycelia using methods as described in a previous study (Zhao et al., 2022b) and detected using the DNA/Protein Analyzer and 1% agarose gel electrophoresis. Two approaches were used for sequencing: paired-end short reads (300 bp) using the Illumina NovaSeq 6000 platform and long reads (more than 20 kb) using the PacBio Sequel II platform. Low-quality reads were removed (end-polished, A-tailed, and ligated with the full-length adaptor for Illumina sequencing, and short reads less than 1000 bp for PacBio sequencing), leaving only high-quality reads. The high-quality reads were de novo assembled using SMRT (version 5.1.0) downloads from https://www.pacb.com/support/software-downloads with default parameters (Chin et al., 2013) and assessed using BUSCO (version 5.2.2) with Ascomycota gene set downloaded from https://busco-data.ezlab.org/v4/data/lineages/ascomycota_odb10.2020-09-10.tar.gz (Simão et al., 2015).

Gene prediction was performed using four tools, Augustus (version 3.2.1) (Stanke et al., 2008), Genemark-ES (version 4.21) (Ter-Hovhannisyan et al., 2008), Genewise (version 2.20) (Birney et al., 2004), and SNAP (version 2010-07-28) (Johnson et al., 2008), and EvidenceModeler (version 1.1.1) (Haas et al., 2008) was then used to make integrated gene models. A total of ten databases were used to functionally annotate genes, namely Gene Ontology (GO) (Ashburner et al., 2000), Kyoto Encyclopedia of Genes and Genomes (KEGG) (Kanehisa et al., 2016), Cluster of Orthologous Groups of proteins (COG) (Galperin et al., 2015), Swiss-Prot (Consortium, 2015), nr (https://www.ncbi.nlm.nih.gov/protein/), Pathogen Host Interactions (PHI) (Torto-Alalibo et al., 2009), Fungal Cytochrome P450 (Fischer et al., 2007), Carbohydrate-Active Enzymes (CAZy) (Levasseur et al., 2013), Virulence Factor Database (VFDB) (Chen et al., 2016), and Type III secretion system Effector protein (T3SS) (Vargas et al., 2012) using Diamond with an e-value less than 1 × 10⁻⁵. In addition, antiSMASH (fungal version) (Medema et al., 2011), Extensive de novo TE Annotator (EDTA) pipeline (version 1.9.5) (Ou et al., 2019), RNAmmer (version 1.2) (Lagesen et al., 2007), and tRNAscan-SE (version 2.0.5) (Lowe and Eddy, 1997) were used to predict gene clusters of secondary metabolites, repetitive elements, rRNAs, and tRNAs, respectively.

In this study, a total of 26 species of ascomycetes ( Table S1 ), including 25 genomes downloaded from genomic databases on the NCBI website and the strain of S. terricola (RCEF 6201) sequenced in this study, were used for phylogenomic analysis. This was performed using single copy ortholog genes that were identified using OrthoFinder (version 2.5.4) (Emms and Kelly, 2019) and aligned using MAFFT (version 7) with default parameters (Katoh and Standley, 2013). A Maximum Likelihood (ML) tree was reconstructed with RaxML (version 8.2.12) (Stamatakis, 2014) using the best optimal model of PROTGAMMAILGF with 100 bootstrap replicates.

Six samples (three biological replicates from both types of growth media) of S. terricola were collected, and total RNA was extracted and purified using methods as described in a previous study (Ren et al., 2021). The transcriptome libraries were constructed and sequenced using the Illumina NovaSeq 6000 platform. After quality control and cleaning of raw reads, the high-quality reads were aligned and assembled using HISAT2 (version 0.1.6-beta) (Kim et al., 2015) and Cufflinks (version 2.2.1) (Trapnell et al., 2012), respectively.

Bowtie2 (version 2.2.5) (Langmead and Salzberg, 2012) was used to count the proportion of high-quality reads of each sample gene compared to the genome of S. terricola, and RSEM (version 1.2.12) (Li and Dewey, 2011) was then used to convert to standardized gene expression levels, viz. reads per kilobase of exon model per million mapped reads (FPKM values). DESeq2 (Wang et al., 2009) was used to identify differentially expressed genes (DEGs), using the criteria of fold change (FC) ≥ 2 and false discovery rate (FDR)< 0.05. Functional enrichments of DEGs were performed using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway databases.

The present study measured the ergosterol content in Sarocladium terricola grown in two different growth media: potato dextrose media and bean media. The study found that ergosterol synthesis was more efficient in potato dextrose media, which resulted in a significantly higher content of ergosterol compared to bean media. Specifically, the ergosterol content in potato dextrose media was 4509 mg/kg, while in bean media, it was only 2382 mg/kg. This difference in ergosterol content between the two growth media was statistically significant with a p-value less than 0.01, as shown in Figure 1 .

The genome assembly of S. terricola spans 11 scaffolds and contains 30.27 Mb with a GC content of 54.07% ( Table 1 ; Figure 2 ). We predicted a total of 10,326 protein-coding gene (PCG) models using four gene prediction tools. Among these gene models, 5,926, 4,220, 1,277, 3,044, 9,055, 1,425, 124, 321, 42, and 3,661 were functionally annotated using GO, KEGG, COG, Swiss-Prot, nr, PHI, P450, CAZy, VFDB, and T3SS databases, respectively. In addition, we predicted 41 gene clusters of secondary metabolites, including 9 NRPS-like, 11 NRPS, 12 T1PKS, 2 T2PKS, 5 terpene, and a single cluster of both beta-lactone and phosphonate. Repetitive elements accounted for 0.81% of the whole genome. Moreover, we detected 282 non-coding RNAs (ncRNAs) in this genome.

A phylogenomic tree was reconstructed using single copy ortholog genes from 26 species, with Agaricus bisporus, Mortierella alpina, and Backusella circina serving as outgroups. A total of 268 single copy ortholog genes were identified through OrthoFinder (Emms and Kelly, 2019), and were used to create an alignment dataset with a length of 339,589 characters. In the Maximum Likelihood phylogenomic tree, S. terricola of the family Sarocladiaceae was found to be within a clade that included members of the family Bionectriaceae, such as Acremonium chrysogenum and A. citrinum (refer to Figure 3 ).

To investigate the impact of ergosterol biosynthesis in different media, we performed transcriptome analyses of S. terricola. We sequenced six transcriptome libraries, three from both bean and potato dextrose media, resulting in a total of 37,043,298 to 42,033,580 clean reads per sample. A total of 5.56 Gb to 6.31 Gb of clean reads were generated, with more than 95.93% of clean reads meeting the Q30 quality threshold ( Table S2 ). Furthermore, over 97.65% of the clean reads mapped to the genome of S. terricola. In each sample, a total of 9,176-9,738 genes (88.86-94.31% of the total number of genes) were found to be expressed.

The differential expression analysis indicated that the HCBe group had a slightly higher number of down-regulated genes compared to up-regulated genes (2221 vs. 1919) when compared to the HCPo group. Furthermore, 5880 genes did not exhibit any significant differences in their expression levels between the two groups, as depicted in Figure 4 . To gain insights into the biological functions of the differentially expressed genes, we performed GO and KEGG pathway enrichment analyses, which are shown in Figures 5 , S1 . Nineteen GO categories were significantly enriched with over 100 DEGs, including cellular processes, metabolic processes, membrane, membrane parts, catalytic activity, and binding, as illustrated in Figures 5A , S1A . Additionally, the KEGG pathway enrichment analysis identified several major pathways associated with the set of DEGs, including global and overview maps, carbohydrate metabolism, signal transduction, cell growth and death, and amino acid metabolism, as shown in Figures 5B , S1B .

In this study, we annotated eleven genes ( Figure 5 ) associated with ergosterol biosynthesis from the KEGG database, which were lanosterol synthase (ERG7; S2G02520), sterol 14alpha-demethylase (CYP51; S1G01142), delta7-sterol 5-desaturase (ERG24; S3G03213), methylsterol monooxygenase (ERG25; S5G06091), sterol-4alpha-carboxylate 3-dehydrogenase (ERG26; S2G02191), 3-keto steroid reductase (ERG27; S7G08361), sterol 24-C-methyltransferase (ERG6-1: S10G10199 and ERG6-2: S1G00302), Delta7-sterol 5-desaturase (ERG3; S7G08143), sterol 22-desaturase (ERG5; S5G06474), and delta24(24(1))-sterol reductase (ERG4; S1G00326). Out of these genes, five genes (S3G03213, S2G02191, S1G00302, S1G00326, and S7G08361) were significantly down-regulated in the HCPo group, whereas three genes (S5G06474, S5G06091, and S7G08143) were significantly up-regulated, as shown in Figure 6 .