The C. cicadae strain AH 10-4 with higher yield of HEA was used in this study, which was isolated and purified in our laboratory in the previous study (Huang et al., 2021). Cordyceps cicadae stain AH 10-4 was plated to PDA medium and incubated at 26°C for 7 days. Three 25 mm² PDA cultures were punched out by the inoculation spatula and then transferred to the seed medium. The seed culture was then grown in a 250-mL flask containing 100 mL of its medium (per liter: glucose 20 g, peptone 10 g, yeast extract 10 g, KH₂PO₄ 2 g, MgSO₄ 0.5 g, ZnSO₄ 0.5 g, and pH natural) at 26°C and 140 rpm for 2 days (Huang et al., 2021).

The C. cicadae stain AH 10-4 was cultivated in 250-mL flasks containing 100 mL medium using 5% inoculum (v/v) and then cultured at 26°C and 140 rpm for 10 days. The flask samples were harvested every day. The fermentation medium was referenced to the seed medium (increasing the concentration of glucose to 30 g/L).

N⁶-(2-hydroxyethyl) adenosine, as an important nucleoside analog of fungi, was closely linked to nitrogen metabolism. In order to investigate the effects of compound and simple nitrogen sources on the production of adenosine and HEA, compound nitrogen sources (peptone, yeast extract) and simple nitrogen sources (ammonium sulfate, ammonium citrate dibasic, diammonium oxalate monohydrate, and ammonium citrate tribasic) were added to basal medium at 10 g/L supplementation, respectively, and incubated at 26°C and 140 rpm for 5 days. The composition of the basal medium was as following (per liter): glucose 30 g, KH₂PO₄ 2 g, MgSO₄ 0.5 g, ZnSO₄ 0.5 g, and pH natural.

To investigate the impact of predictive amino acids on HEA production, 30 g/L glucose with 0.25 g/L amino acids were added to 67 g/L yeast nitrogen base w/o amino acids and then cultured at 26°C and 140 rpm for 7 days, the amount of each ingredient added in Table 1. The 15 amino acids are valine, leucine, isoleucine, proline, phenylalanine, tyrosine, tryptophan, threonine, cysteine, methionine, asparagine, glutamine, lysine, arginine, and histidine. The five forecast amino acids are alanine, serine, glutamic acid, aspartic acid and glycine.

The C. cicadae stain AH 10–4 was cultivated in 250-mL flasks containing 100 mL medium using 5% inoculum (v /v). It was cultured at 26°C for 17 days by liquid static culture. In order to investigate the effect of amino acids on the production of HEA, amino acids (alanine, serine, aspartic acid, glycine as well as glutamic acid), were added at an addition of 5 g/L to the base medium (per liter: glucose 30 g, peptone 5 g, KH₂PO₄ 2 g, MgSO₄ 0.5 g, ZnSO₄ 0.5 g, pH natural). The incubation was then incubated at 26°C for 17 days.

Submerged fermentation was cultured with a 5% seed-cultured solution at 26°C, shaking at 140 rpm in the first 3 days and then static incubation for the latter 6 days. It was collected after 9 days incubation.

The fermentation broth of C. cicadae strain AH 10-4 was centrifuged at 4,500 rpm for 30 min to obtain mycelium, which was then oven-dried at 60°C to a constant weight. For the extraction of adenosine and HEA, the mycelium was ground into a powder that was sieved through a 60-mesh sieve. Then the powder was weighed and extracted by ultrasonic extraction with 20 fold of ultrapure water for 90 min with mixing at 10 min intervals. The extracts were centrifuged at 4,500 rpm for 30 min, and the supernatant was filtered through a Minisart R filter with a pore size of 0.22 μm for later use. The HPLC detection conditions for adenosine and HEA were consistent with the previous studies (Huang et al., 2021). The adenosine and HEA content in mycelium was measured in mg/g, with “g” representing the dry weight of the mycelium.

Total RNA was extracted from mycelium in shaker and stationary cultures (ammonium sulfate as nitrogen source) by the TRIzol method. The poly-T Oligo-attached magnet was used to enrich mRNA from the total RNA, which was subsequently fragmented in fragmentation buffer. And then a total amount of 1.5 μg mRNA was used to construct cDNA library for each sample by using Illumina’s NEBNext® Ultra™ RNA Library Preparation Kit. 200–250 bp cDNA fragments were screened and purified using the AMPure XP system (Beckman Coulter, Beverly, United States). The quality of the libraries was then assessed on an Agilent Bioanalyzer 2100 system and sequenced (150 bp paired-end) on an Illumina Novaseq 6000 platform. A total of six libraries were constructed with three biological replicates per sample.

The clean reads were obtained by removing reads with adapter, ploy-N, and low quality reads from the raw paired-end reads, which were then assembled using Trinity (Grabherr et al., 2011). To further explore the annotation function of the assembled and spliced transcripts, the assembled unigene was annotated against Nt, Nr, Swiss-Prot, GO, KOG, KEGG, and Pfam databases to obtain the annotation of the unigene.

The stationary and shaker fermentation mycelium was analyzed for differential expression using the DESeq R software package (Anders and Huber, 2010). The genes with a p value of less than 0.05 on an adjusted basis found by DESeq were then classified as differentially expressed genes. In addition, the mycelium from stationary and shaker fermentation cultures were analyzed for functional enrichment that was performed for GO and KEGG by using the GOsep R software package and KOBAS, respectively (Mao et al., 2005). The screening criteria for significantly enriched GO and KEGG by DEGS were Bonferroni corrected p values of less than or equal to 0.05.

The data were reported as mean ± SD and analyzed using one-way ANOVA with Duncan’s multiple rang test in SPSS software version 23.0 (IBM Corp., Armonk, NY, United States) to determine the significance of group differences. A p value less than 0.05 was considered statistically significant.

A time course analysis was conducted to study the kinetics of HEA production in strain AH 10–4 grown in shaker culture for 10 days. Firstly, in terms of biomass, it was seen that the biomass of mycelium increased with incubation time from Day 1 to Day 8, reaching the highest level of 2.403 g/100 mL on Day 8, and then slightly decreased (Supplementary Table S1). As was shown in Figure 3, after the 2 days culture, the content of adenosine reached to the highest level of 5.836 ± 0.486 mg/g, and then decreased gradually until Day 5. Distinct from adenosine, the content of HEA gradually increased with the incubation time, reaching its highest level on the 5th day (2.257 mg/g), and then slightly decreased since the sixth day. Interestingly, the content of HEA obviously increased from Day 2 to Day 5, in which time the content of adenosine decreased gradually, showing a correlation between the increasing trend in HEA production and the decreasing trend in adenosine production. Therefore, we speculated that there was a connection between the synthesis pathways of HEA and adenosine.

Due to the reason that adenosine and HEA belong to nucleosides produced by fungus, and nucleosides are known to be the derivatives of nitrogen metabolism. Given the advantages of the shaker culture, such as shorter fermentation time, it was chosen as the main method for obtaining mycelium. Thus, in order to investigate the effects of various nitrogen sources on the content of adenosine and HEA in mycelial, we measured the production of adenosine and HEA after 5 days of shaker culture, a time point that showed optimal of HEA production in Figure 3. As shown in Figure 4A, the HEA content was 0.047, 0.730, 0.284, and 0.547 mg/g mycelia dry weight by using ammonium sulfate, diammonium oxalate monohydrate, ammonium citrate dibasic, and ammonium citrate tribasic as nitrogen source, respectively. Among the ammonium salts, the use of diammonium oxalate monohydrate and ammonium citrate tribasic possessed more superior HEA content in mycelium, while the biomass of ammonium citrate tribasic was remarkably higher than that of diammonium oxalate monohydrate (p < 0.05) (Supplementary Table S2). Therefore, ammonium citrate tribasic was determined to be the optimum nitrogen source among the ammonium salts. Distinct from ammonium salts, the yields of HEA were 1.872 and 0.892 mg/g from yeast extract and peptone medium respectively, which were significantly superior to those from ammonium salts (p < 0.05). By comparison with ammonium salts, complex nitrogen sources (peptone, yeast extract) were more favorable for the accumulation of HEA production in mycelium during shaker fermentation, which might due to their enrichment in amino acids.

In contrast to shaker fermentation, it has been demonstrated that stationary culture has the advantages of low energy consumption, low requirements for production conditions and easy scale-up (Liao et al., 2021). Therefore, we probed the ability of various nitrogen sources to yield HEA and adenosine after 17 days of stationary culture. In terms of biomass, complex nitrogen sources (yeast extract, peptone) promoted the growth of C. cicadae strain AH 10-4 mycelium better than ammonium salt (p < 0.05) (Supplementary Table S3), including ammonium citrate tribasic, ammonium citrate dibasic, and diammonium oxalate monohydrate (Figure 4B). Interestingly, ammonium citrate tribasic, ammonium citrate dibasic, and diammonium oxalate monohydrate were better than ammonium sulfate in stimulating mycelial growth of C. cicadae (p < 0.05). Complex organic nitrogen sources (yeast extract, peptone) were found to favor adenosine accumulation in the mycelium. As for HEA production, yeast extract and ammonium citrate tribasic were the most effective nitrogen sources, with HEA yields of 2.599 and 2.535 mg/g, respectively. These were followed by diammonium oxalate monohydrate and peptone (p < 0.05). Notably, stationary fermentation caused an increase in the HEA content within mycelium, while the opposite was found for the biomass. Therefore, diammonium oxalate monohydrate and ammonium citrate tribasic can be considered viable alternatives to complex nitrogen source to obtain high HEA yield industrially through stationary fermentation.

Submerged fermentation, a novel approach combining shaker and stationary fermentation, was investigated in order to identify an optimal fermentation method and nitrogen source for industrial application. We also explored the ability of different nitrogen sources to produce HEA of C. cicadae strain AH 10-4 under submerged fermentation. The result of mycelial biomass was shown in Figure 4C. Yeast extract was most effective for the growth of C. cicadae mycelium, which was followed at more than 1 g/100 mL for ammonium citrate, diammonium oxalate monohydrate and peptone as nitrogen source (Supplementary Table S4). Yeast extract as a nitrogen source yielded the highest levels of adenosine (3.416 mg/g) and HEA (4.691 mg/g) within C. cicadae strain AH 10–4 mycelium. The HEA concentration was 1.928, 0.065, 2.593, 0.450, and 2.578 mg/g mycelia dry weight by using peptone, ammonium sulfate, diammonium oxalate monohydrate, ammonium citrate dibasic, and ammonium citrate tribasic. The result indicated that diammonium oxalate monohydrate and ammonium citrate tribasic were comparable to peptone in enhancing HEA accumulation within the mycelium. Additionally, ammonium citrate tribasic and diammonium oxalate monohydrate were inexpensive and could be used as an alternative source of organic nitrogen. Notably, ammonium citrate tribasic had superior production of adenosine and HEA compared to diammonium oxalate monohydrate. Consequently, ammonium citrate tribasic was regarded as the optimal nitrogen source to produce higher HEA concentrations in C. cicadae strain AH 10-4 mycelium. Submerged fermentation increased the biomass and reduced incubation time than stationary fermentation. Importantly, this can be easily scaled-up in the fermentation industry.

In this study, we did transcriptomic analysis to compare mycelium under high-HEA-producing and low-HEA-producing conditions with the aim to find the different metabolic pathways related to HEA biosynthesis. Comparing HEA content in stationary (Figure 4B) and shaker cultured mycelium (Figure 4A), ammonium sulfate has a simple composition and a significant difference of HEA content, which was chosen a suitable nitrogen source for transcriptomics. Therefore, samples of mycelium cultured with ammonium sulfate as a nitrogen source in stationary (A1-3) and shaker (B1-3) cultures, including three biological replicates, were selected for transcriptomics. For the purpose of distinguishing transcriptomic features during HEA biosynthesis, RNA-seq libraries were performed by extracting total RNA from stationary and shaker fermentation mycelium, respectively. Then, the RNA-seq libraries were sequenced using Illumina Novaseq 6000. The raw reads obtained for the stationary culture and shaker fermentation samples ranged from 50,645,198 to 59,883,168, and raw reads were filtered to obtain clean reads ranging from 50,041,026 to 59,299,218 (Table 2). Among all six samples, the proportion of bases with quality value ≥20 was above 97.58% in Q20 and the proportion of bases with quality value ≥30 was above 93.55% in Q30 in all groups, which indicated that the sequencing results were satisfactory and could be used for subsequent analysis.

The clean reads were assembled into 9,713 unigenes and 23,356 transcripts using Trinity software (Trinity Technologies, CA, United States), with an average assembled single gene length of 2,721 bp (N50 = 4,483) (Supplementary Table S5). The 23,356 transcripts comprised 2,024 (8.67%) transcripts within the length range of 300–500 bp, 2,694 (11.53%) with 500–1,000 bp, 3,680 (15.76%) with 1,000–2,000 bp, and 14,958 (64.04%) with >2,000 bp. The 9,713 unigenes consist of 1,478 (15.22%) unigenes with length range of 300–500 bp, 1,627 (16.75%) with 500–1,000 bp, 1,872 (19.27%) with 1,000–2,000 bp, and 4,736(48.76%) with the length of more than 2,000 bp (Figure 5).

To annotate the gene functions of the stationary and shaker culture samples, the Nr, Nt, Pfam, KOG, Swiss-Prot, KEGG, and GO databases were used, which matched to 6,734, 6,393, 5,966, 2,171, 4,595, 1,443, and 4,636 unigenes, respectively, (Supplementary Table S6). The annotations of the unigenes sequences in the seven database species were shown in the Figure 6A, which indicated that the unigenes sequences were quite similar to the known proteins in the public databases. The 46.36% of the genes displayed an E-value of 0, indicating the reliability of these matches (Figure 6B). Regarding the identity distribution of the predicted proteins, the majority of hits (78.46%) in the NR database exhibited 80–100% identity with other fungi proteins, while only 0.03% of the sequences had 18–40% identity (Figure 6C). Furthermore, the distribution of species was also analyzed (Figure 6D), with the annotation results based on the NR library indicating that the genes of C. cicadae displayed the highest number of matches with the genes of Cordyceps fumosorosea.

Approximately 87–89% of the total reads were successfully mapped to the C. cicadae transcriptome (Supplementary Table S7), indicating a high mapping efficiency. We then examined the number of DEGs, with p value of less than or equal to 0.05 (Supplementary Figure S3A). A total of 4,121 DEGs were detected, with 2,038 upregulated and 2,083 downregulated genes (Supplementary Figure S3B). The Venn diagram analysis revealed that 6,499 genes were co-expressed genes (Supplementary Figure S4). These data indicated that the gene expression patterns remained consistent in mycelium cultured with ammonium sulfate as nitrogen source in stationary and shaker cultures.

To explore the functions of the identified unigenes in the C. cicadae transcriptome analysis, GO enrichment analysis was performed. The successfully annotated unigenes were grouped according to GO categories, including biological processes (BP), cellular components (CC) as well as molecular functions (MF). It was found that 9,713 unigenes were classified into 49 functional groups. The most common of these groups were metabolic process (2,516 unigenes), cell (1,984 unigenes), and catalytic activity (2,342 unigenes) (Figure 7). This analysis provided insights into the diverse functions and activities of the identified unigenes in C. cicadae.

The DEGs identified between stationary and shaker fermentation were further used for GO enrichment analysis. It was found that the DEGs were mainly related to “RNA processing” (116), “ncRNA processing” (76), “nucleolus” (66), “ribosome biogenesis” (69), “preribosome” (40), “peptidase activity” (113), “peptidase activity, and acting on L-amino acid peptides” (93) (Figure 8). These up- and downregulated genes were crucial for the notable difference in HEA production between stationary and shaker-cultured mycelium. Thus, these DEGs provided potential targets for future functional studies aiming to elucidate the underlying mechanisms involved in HEA production.

To seek metabolic pathways related to HEA synthesis, we annotated all differentially expressed genes against the KEGG database. The DEGs of top 20 pathways were categorized into four branches: cellular processes, environmental information processing, information processing and metabolism. Among the top 20 pathways, the highest number of differential genes are shown in Figure 9. The largest group of the top 20 enriched pathways was responsible for metabolism, including “Alanine, aspartate and glutamate metabolism” (20), “Glyoxylate and dicarboxylate metabolism” (20), “Galactose metabolism” (14), “Nitrogen metabolism” (10), “Starch and sucrose metabolism” (21), “Amino sugar and nucleotide sugar metabolism” (28), ‘beta-Alanine metabolism” (13), “Ether lipid metabolism” (10), “Biosynthesis of secondary metabolites” (130), “Glycosphingolipid biosynthesis—globo and isoglobo series” (4), “Valine, leucine and isoleucine degradation” (15), “Arginine biosynthesis” (10), “Steroid biosynthesis” (12), and “Metabolic pathways” (342). We speculated that HEA synthesis might be closely linked to pathways such as “alanine, aspartate and glutamate metabolism,” “nitrogen metabolism” and “β-alanine metabolism.” It was well known that amino acids provided the backbone of the purine metabolic pathway. However, the relationship of alanine, aspartate and glutamate metabolism to the synthesis of HEA needs to be further investigated. Furthermore, HEA is an adenosine derivative, so transcriptional studies of amino acid synthesis pathways and purine metabolic pathways of metabolism in combination with analyses of biosynthesis genes have the potential to dissect biosynthesis pathways of HEA.

As shown in Figure 4A, the complex nitrogen source promoted the accumulation of HEA in mycelium during shaker fermentation. The complex nitrogen sources had a superior production of HEA since they were rich in amino acids. In addition, alanine, aspartate, and glutamate as well as arginine metabolic pathways were significantly different, which is consistent with our speculation that amino acids can be used as substrates for HEA synthesis (Figure 9). Due to HEA as a purine derivative, glutamic acid, glycine, serine, and aspartic acid are involved in purine metabolism. In addition, the former research has also indicated that alanine promotes HEA accumulation in mycelium (Lei et al., 2014). Therefore, we used yeast nitrogen base w/o amino acids supplemented with 15 amino acids as the basal medium, and then the five presumed amino acids were added, respectively, to explore their relationship with the HEA content. The result was as shown in Table 3, there was no statistically significant difference in the content of HEA between the group with the addition of the 15 amino acids and the control group, so we inferred that the 15 amino acids had no relevance to the biosynthesis of HEA (Table 3). However, the group supplemented with glutamate exhibited significantly higher HEA content compared to the other groups.

In addition, in order to further validate, we also added the presumed 5 amino acids (Gly, Glu, Ser, Asp, and Ala) to the peptone medium and then HEA content was tested, respectively, after 17 days of stationary fermentation, respectively. Addition of amino acids (glutamate, glycine, serine, aspartate, and alanine) resulted in lower biomass compared to the control (Supplementary Table S8). As shown in Figure 10. When peptone was used as nitrogen source, the adenosine content of the strain AH 10-4 fermenting mycelium was 2.332 mg/g and HEA content was 1.422 mg/g. Interestingly, the addition of amino acids to based medium led to a decrease in HEA content. Hence, it was speculated that excess nitrogen source inhibited HEA synthesis after 17 days of fermentation. Among the peptone medium added with five amino acids, glutamate showed a significant effect on the HEA content in the fermenting mycelia of the strain AH 10-4, which is consistent with the results in Table 3.