Transcriptomic and metabolomic insights into the long-term survival strategies of Ophiocordyceps sinensis following host immune perturbation

Abstract

Background

Ophiocordyceps sinensis is a premier medicinal fungus with a unique slow-killing parasitic strategy, yet the molecular basis of its long-term persistence in host hemocoel remains elusive. This fungus coexists with Thitarodes xiaojinensis larvae for over six months, ultimately leading to host death at the last instar before pupation. To decode the fungal survival strategies masked by host resistance, this study integrated omics approach combining sublethal imidacloprid-induced transcriptomics with longitudinal metabolomic profiling. We investigated the transcriptomic response of O. sinensis within the larval hemocoel under exposure to a sublethal dose of imidacloprid, a stressor utilized here to suppress host immunity via neuro-immune crosstalk. This approach allowed us to infer the core molecular responses of O. sinensis across three critical infection stages (30, 90, and 150 days post-infection, dpi).

Results

Differential gene expression and functional enrichment analyses revealed that O. sinensis significantly responds to host immune disturbance induced by sublethal dose of imidacloprid. During the early infection stage (30 dpi), Imidacloprid exposure suppressed fungal cell proliferation genes while activating secondary metabolite pathways and immune defense genes. This suggests that during a conventional infection, O. sinensis secondary metabolism was suppressed under the host immunity to maintain a balanced blastospore proliferation. Beyond similar response in fungal secondary metabolites and host immunity, imidacloprid also suppressed protein biosynthesis during the mid-infection stage (90 dpi), indicating that the fungus undergoes metabolism suppression while continuing to maintain balance between blastospore proliferation and host immunogenicity for the conventional infection. During the late stage (150 dpi), genes associated with hydrolytic activity, transport and cell wall degradation were upregulated while fatty acid synthesis genes were downregulated by imidacloprid, suggested that the conventional infection of O. sinensis at this stage exhibits metabolic polarization to shift blastospore to hypha, tightly regulating hydrolysis to balance nutrient acquisition with host viability and increasing fatty acid biosynthesis. Metabolomic analysis showed that the accumulation of chitin precursors in the late stage, including D-glucosamine-6-phosphate, marks the critical blastospore-to-hyphae transition. Furthermore, the accumulation of compounds like chlorogenic acid, L-glutathione oxidized, and hesperidin indicates the dual responses of secreting antimicrobials and deploying antioxidants to protect and sanitize the nutrient base.

Conclusions

This study provides insights into the dynamic, stage-specific gene expression profiles of O. sinensis that enable its long-term survival in the host hemocoel. Through coordinated regulation of proliferation, metabolism, secondary metabolite production, and immune evasion, the fungus sustains persistent infection and successfully completes its life cycle. These findings advance our understanding of the strategies employed by slow-killing fungal parasites.

Background

Chinese cordyceps refers to a unique fungus-insect complex formed by the infection of Hepialidae moth larvae by the entomopathogenic fungus Ophiocordyceps sinensis [43, 44]. Widely used in traditional Chinese medicine as whole fungus-insect complex, it is valued for its pharmacological properties, including anti-inflammatory, immunomodulatory, and antioxidant effects, and is commonly used in the treatment of respiratory diseases [2, 7–9, 39, 40, 46]. Due to its immense medicinal value and scarcity, the fungus-insect complex formed by O. sinensis commands an exorbitant market price, earning it the nickname soft gold. This scarcity stems primarily from its restricted distribution on the Qinghai-Tibet Plateau and a complex infection cycle [24, 26]. Furthermore, natural populations have declined precipitously in recent decades driven by over-harvesting and habitat degradation caused by climate change. Consequently, O. sinensis has been classified as Vulnerable on the IUCN Red List of Threatened Species [41], highlighting the urgent need for artificial cultivation and a deeper understanding of its infection biology.

The life cycle of O. sinensis is marked by distinct developmental stages. Ascospores are released in late August and disseminated by wind or rain, adhering to the larval cuticle where they penetrate using specialized proteases [24, 26]. Infected larvae can survive for two to three years without displaying overt disease symptoms. Eventually, they die and mummify 2–4 cm below the soil surface, where fungal stromata emerge for sexual reproduction [21]. Unlike fast-killing entomopathogens such as Beauveria and Metarhizium species, O. sinensis exhibits a slow and prolonged infection process with multiple morphological transitions in the host hemocoel [24, 26]. The infection cycle within the Hepialidae host is characterized by a prolonged persistent period, often exceeding six months. Initially, fusiform yeast-like unicellular blastospores colonize the hemocoel. During this parasitic phase, the fungus progresses through a proliferative stage (BP), characterized by low fungal density and active cell budding. As the infection persists for over five months, it reaches a stationary stage (BS) upon surpassing a population threshold; in this stage, fungal density stabilizes, and the host’s phenoloxidase activity and encapsulation reactions remain significantly inhibited [24, 26]. These early stages (BP and BS) are primarily defined by fungal cell density and morphological observations of budding. When the host larvae enter the final instar (just before pupation), a critical morphological switch is triggered: blastospores differentiate into segmented, multinucleated pre-hyphae (PreHy). Unlike the earlier stages, PreHy elongates via fission without sister-cell separation, a morphotype transformation that facilitates rapid expansion throughout host tissues. These finally transition into reticulated hyphae (Hy), leading to host mummification and death [24, 26]. This prolonged intra-host residence makes O. sinensis an ideal model for studying immune homeostasis maintenance in slow-killing parasitic systems.

Previous studies have shown that O. sinensis employs dynamic virulence strategies to evade host immune defenses. Transcriptomic analyses have revealed the persistent overexpression of genes encoding transporters, glycoside hydrolases (GH18 family), tissue-penetrating enzymes, and mycotoxins during the symbiotic phase, as well as the activation of MAPK signaling pathways during the transition from blastospore to hyphal stages [44]. Additionally, a notable increase in lipid droplets has been observed in the BS stage compared to the BP stage, with corresponding upregulation of fatty acid biosynthesis genes and metabolite accumulation [24, 26]. These findings highlight the role of lipid metabolism in fungal development and pathogenicity. However, the precise strategies employed by O. sinensis to achieve long-term immune modulation and survival in its host remain unclear.

In this study, we distinguish between natural infection in wild environments and conventional infection, which represents the standardized biological progression following laboratory inoculation. To investigate fungal strategies during conventional infection, we used sublethal dose of imidacloprid (SLDI; LC₃₀ = 30 mg/L), a neonicotinoid insecticide known to disrupt the insect neuro-immune axis via competitive binding to nicotinic acetylcholine receptors [3], as an immune perturbation agent. Thitarodes xiaojinensis larvae were exposed to SLDI for 72 h at three key infection stages: 30 days (BP stage), 90 days (BS stage), and 150 days (PreHy stage) post-infection. By integrating next-generation RNA sequencing with metabolomic data, we characterized the molecular responses of O. sinensis to these immune fluctuations to reveal the mechanisms underpinning its persistent infection.

Materials and methods

Determination of sublethal imidacloprid dose

Imidacloprid (95% purity; Shanghai Jinsui Bio-Technology Co., Ltd., Shanghai, China) was selected as the immunomodulatory agent based on its specificity, efficacy, and safety profile. Ascospores were collected from mature O. sinensis fruiting bodies and isolated as single-spore strains on PDA medium (14–20 °C). Once a large number of conidia were produced, a suspension was prepared in 0.5% Tween-80 with a concentration of 1.0 × 104–1.0 × 106 conidia/mL. Inoculation was performed using a custom-designed apparatus specifically developed for Thitarodes xiaojinensis larvae (length: 1.5–2 cm; diameter: ~ 3 mm). Day 0 post-infection (0 dpi) is defined as the day the inoculation was performed. Two concentrations (30 and 60 mg/L) were prepared by dissolving imidacloprid in distilled water. Standard insect feed was soaked in each solution for 30 min to ensure uniform absorption. Third-instar Thitarodes xiaojinensis larvae (n = 20 per group) of similar body size were randomly assigned to feed on either medicated diets or untreated control diets (0 mg/L). All Thitarodes xiaojinensis larvae were maintained under controlled conditions (16 °C, 75% relative humidity, complete darkness). Mortality was assessed 72 h post-exposure, as this duration was sufficient to achieve a sub-lethal mortality effect and induce significant immune suppression; 90 individuals unresponsive to mechanical stimulation were considered dead. The LC₃₀ (lethal concentration causing 30% mortality) was determined based on mortality counts, and 30 mg/L was selected for subsequent immune perturbation experiments. It should be noted that this LC₃₀ value represents an empirical estimate derived from iterative dose–response trials, rather than a model-based calculation.

Sample collection and RNA extraction

Thitarodes xiaojinensis larvae and Ophiocordyceps sinensis isolates were provided by Dongguan HEC Cordyceps R&D Co., Ltd. (Guangdong, China). These larvae originated from a standardized laboratory-reared colony whose ancestral founders were collected from Xiaojin County, Sichuan Province, China. Using an artificially bred population ensured the uniformity of the experimental materials and eliminated the need for wild collection permits. Infected Thitarodes xiaojinensis larvae were maintained at 16 °C and 75% relative humidity in complete darkness to mimic conventional environmental conditions. Three key infection stages were analyzed:

• (i)Early infection stage (30 dpi): One-month post-infection, characterized by rapid fungal proliferation and active suppression of host immunity.
• (ii)Mid infection stage (90 dpi): Three months post-infection, marked by a stable equilibrium between fungal and host activity. The fungus fine-tunes host immune responses and begins affecting host physiology in preparation for behavior manipulation and future reproduction.
• (iii)Final infection stage (150 dpi): Five months post-infection, when O. sinensis transitions from yeast-like growth to hyphal differentiation, leading to extensive tissue colonization and host motor impairment.

At each infection stage (30, 90, and 150 dpi), 120 confirmed infected larvae were utilized. These were divided into control and imidacloprid-treated groups (n = 60 per group), with each group further partitioned into three biological replicates of 20 larvae. Following the 72-h exposure period, 10 surviving individuals were randomly selected from each replicate for sampling. This pooled sampling approach was employed to mitigate individual biological variability and ensure adequate RNA yield for sequencing. To ensure the accuracy of the infection stages, the infection status of each larva was verified upon dissection. Hemolymph was collected and examined via light microscopy; successful infection was defined by the visible presence of characteristic O. sinensis blastospores. Only confirmed infected larvae were assigned to the 30, 90, and 150 dpi groups for further analysis. Three biological replicates were prepared for each stage. Total RNA was extracted using TRIzol™ Reagent (Invitrogen, USA), and quality was assessed by 1.2% agarose gel electrophoresis, NanoDrop 2000 spectrophotometry (A260/A280 > 1.8), and Agilent 2100 Bioanalyzer analysis (RIN > 6.5) [29]. Only high-quality RNA samples were used for transcriptomic library construction.

Library preparation and RNA sequencing

Ribosomal RNA was depleted by poly-A selection to enrich for mRNA transcripts. Strand-specific cDNA libraries were prepared using the TruSeq Stranded mRNA Library Prep Kit (Illumina, USA) according to the manufacturer’s instructions. Briefly, mRNA was enzymatically fragmented, and first-strand cDNA was synthesized using random hexamer primers and reverse transcriptase. Second-strand synthesis incorporated dUTP to retain strand information. Adapters with unique molecular indices were ligated to both ends, followed by PCR amplification (12–15 cycles). Library quality was validated using Qubit fluorometry and Agilent 2100 Bioanalyzer. Paired-end sequencing (2 × 150 bp) was performed on the Illumina HiSeq 2500 platform, generating approximately 80 million reads per sample.

Differential gene expression and functional enrichment analysis

Raw sequencing reads were subjected to quality control using FastQC to evaluate base quality, guanine-cytosine (GC) content, and du plication levels. Reads containing adapter sequences or with Phred quality scores < 20 were removed. Clean reads were aligned to the O. sinensis reference genome (GWHGDWT00000000.1, Genome Sequence Archive) using HISAT2 with default parameters. Gene-level read counts were generated with featureCounts based on the corresponding annotation file.

Differential expression analysis was conducted using DESeq2 [38], with differentially expressed genes (DEGs) identified by an adjusted p-value (padj) < 0.05 and absolute log₂(fold change) > 1. Functional enrichment of DEGs was performed using Gene Ontology (GO: http://geneontology.org), Kyoto Encyclopedia of Genes and Genomes (KEGG: https://www.kegg.jp), and Gene Set Enrichment Analysis (GSEA v4.3.2). GO and KEGG terms were considered significantly enriched at a false discovery rate (FDR) < 0.05. GSEA results were filtered by normalized enrichment score (|NES|> 1.5) and q-value < 0.25.

Metabolite detection

Hemolymph was harvested from larvae across four experimental groups: Stage S (early, 30 dpi), Stage M (mid, 90 dpi), Stage L (late, 150 dpi), and an age-matched uninfected control group (corresponding to the 30-dpi stage). Prior to collection, the infection status of each larva in the S, M, and L groups was individually verified via microscopy to confirm the presence of O. sinensis. Each group consisted of 6 biological replicates, with each replicate comprising hemolymph pooled from 4 confirmed larvae to minimize individual biological variability. Metabolite profiling was conducted by Novogene Co., Ltd. (Beijing, China) using a Vanquish UHPLC system coupled to an Orbitrap Q Exactive™ HF-X mass spectrometer (Thermo Fisher Scientific, Germany). To achieve comprehensive metabolic coverage, two distinct chromatographic methods were employed. For the analysis of less polar compounds, samples were separated on a Hypersil Gold C18 column (100 × 2.1 mm, 1.9 µm) with a mobile phase consisting of 0.1% formic acid in water (A) and methanol (B). The gradient elution was programmed as follows: 2% B for 0–1.5 min; 2–85% B for 1.5–3 min; 85–100% B for 3–10 min; 100–2% B for 10–10.1 min; and 2% B for 10.1–12 min. For the analysis of more polar compounds, an ACQUITY UPLC BEH Amide column (100 × 2.1 mm, 1.7 µm) was used with a mobile phase of 5 mM ammonium acetate in 90% acetonitrile (A) and 5 mM ammonium acetate in 50% acetonitrile (B), under the following gradient: 2% B for 0–1.5 min; 2–100% B for 1.5–7 min; 100% B for 7–9 min; 100–2% B for 9–9.1 min; and 2% B for 9.1–12 min. The flow rate for both methods was maintained at 0.2 mL/min. The mass spectrometer was operated in both positive and negative ion modes with an electrospray ionization (ESI) source, using a spray voltage of 3.5 kV, a capillary temperature of 320 °C, and an auxiliary gas heater temperature of 350 °C. The raw UHPLC-MS/MS data were processed using the XCMS software for peak picking, alignment, and quantification. Metabolite identification was performed by matching the experimental secondary mass spectra against the self-built NovoMetDB database, with a mass deviation tolerance of 10 ppm. Following the removal of background ions based on blank samples, the quantitative results were normalized to obtain relative peak areas. To ensure data quality, features with a coefficient of variation (CV) greater than 30% in the quality control (QC) samples were excluded from further analysis.

Time-series analysis of metabolomic data

To identify temporal patterns in the metabolomic data, a soft clustering analysis was performed using the Mfuzz R package (version 2.62.0). Prior to clustering, the normalized metabolite abundance data was standardized using the standardise function within the package. The fuzzifier parameter m was estimated using the mestimate function. The optimal number of clusters c was determined by running the Dmin function with a range of cluster numbers (from 2 to 12) and selecting the value that maximized the minimum centroid distance. The dataset was then clustered using the mfuzz function, and metabolites with a membership value (α) greater than 0.7 were retained for further analysis.

Weighted co-expression network analysis (WGCNA) and hub metabolite identification

To further refine the clusters identified by Mfuzz and to identify key metabolites associated with the infection timeline, a Weighted Co-expression Network Analysis (WGCNA) was performed on the metabolites within each significant cluster using the WGCNA R package (1.73). First, an adjacency matrix was constructed by calculating the Pearson correlation between all pairs of metabolites, which was then transformed into a topological overlap matrix (TOM). The soft-thresholding power (β) was selected to ensure the network approximated a scale-free topology. Metabolite co-expression modules, representing groups of highly interconnected metabolites, were identified using hierarchical clustering of the TOM. The relationship between each module and the external trait (infection time) was assessed by calculating the Pearson correlation between the module eigengene (ME) and the time points. Modules showing a statistically significant correlation (p < 0.05) with infection time were selected for further investigation. Within these significant modules, hub metabolites were identified based on two criteria: high Module Membership (kME > 0.8) and high Metabolite Significance for the time trait (absolute correlation > 0.2). Metabolites satisfying both criteria were considered hub metabolites and were subjected to functional pathway and biological interpretation.

Results and discussion

Sequencing and reads mapping

To investigate how O. sinensis responds to sublethal dose of imidacloprid (SLDI, LC = 30 mg/L) during parasitism of T. xiaojinensis larvae, RNA sequencing was performed on samples collected at three key infection stages. RNA-seq libraries yielded an average of 87,521,146 raw reads per sample. After filtering out low-quality sequences (Q20 < 90%) and adapters, an average of 86,880,507 high-quality clean reads per sample was retained. All samples showed excellent sequence quality, with error rates below 0.03% and GC content ranging from 44.08% to 59.55%. Clean reads were mapped to the O. sinensis reference genome using HISAT2. Mapping rates varied by infection stage: 52.79%–71.00% at the early stage, 39.64%–81.80% at the mid stage, and 78.18%–91.15% at the final stage (Table 1). These differences likely reflect changing ratios of fungal to host RNA as infection progresses.

Table 1.

Summary of RNA-seq data quality and genome mapping statistics

Sample	Raw Reads	Clean Reads	Error Rate	Q20	Q30	GC Content	Mapping Rate
S_TR_1	78,331,932	77,880,564	0.02%	97.95%	94.38%	55.68%	65.41%
S_TR_2	76,713,844	76,223,490	0.02%	97.98%	94.45%	54.72%	62.50%
S_TR_3	78,029,656	77,576,182	0.02%	98.09%	94.64%	53.52%	52.79%
S_CK_1	85,618,318	85,087,932	0.03%	97.92%	94.31%	56.16%	69.22%
S_CK_2	82,664,536	82,164,882	0.03%	97.75%	93.90%	55.23%	62.69%
S_CK_3	83,364,116	82,826,002	0.03%	97.84%	94.14%	56.50%	71.00%
M_TR_1	101,847,910	101,175,668	0.02%	98.47%	95.46%	51.63%	39.64%
M_TR_2	107,920,506	106,813,968	0.02%	98.01%	94.62%	57.69%	81.80%
M_TR_3	93,487,498	92,581,934	0.02%	98.07%	94.69%	55.57%	64.09%
M_CK_1	100,388,444	99,495,040	0.02%	98.02%	94.51%	55.06%	65.19%
M_CK_2	101,264,494	100,478,204	0.02%	98.04%	94.63%	56.92%	76.11%
M_CK_3	95,923,306	95,192,844	0.03%	97.89%	94.18%	53.91%	64.09%
L_TR_1	88,178,360	87,536,618	0.02%	97.97%	94.57%	59.23%	88.92%
L_TR_2	82,932,726	82,313,354	0.03%	97.80%	94.12%	57.23%	78.18%
L_TR_3	78,512,850	77,997,720	0.02%	97.92%	94.40%	57.97%	81.46%
L_CK_1	81,225,538	80,629,920	0.02%	97.97%	94.47%	59.09%	88.29%
L_CK_2	81,211,612	80,668,130	0.03%	97.90%	94.30%	59.55%	91.15%
L_CK_3	77,764,976	77,206,672	0.02%	97.93%	94.38%	59.18%	88.91%

Transcriptional response of O. sinensis to imidacloprid during the early infection stage

To characterize the transcriptional response of O. sinensis during early infection, we performed differential gene expression analysis under SLDI exposure using DESeq2 (padj < 0.05, |log₂FC|> 1). A total of 565 genes were significantly upregulated, whereas 227 genes were downregulated in the imidacloprid-treated group compared with controls. Upregulated genes were primarily linked to secondary metabolism, including trichodiene synthase (IPR010458), hydrolases for aromatic compound degradation (IPR052358), and α-ketoglutarate-dependent sulfate ester dioxygenase (IPR051323) (Fig. 1A). Collectively, these enzymes facilitate the conversion of farnesyl pyrophosphate (FPP) into tricyclic terpenoids [5, 31] and the oxidative breakdown of host-derived aromatic compounds into tricarboxylic acid (TCA) cycle intermediates such as succinate. Notably, carbohydrate metabolism genes such as sorbitol dehydrogenase (IPR045306) and glucosamine-6-phosphate isomerase (IPR004547) were also induced (Fig. 1A), suggesting activation of the sorbitol–fructosamine shunt pathway [14, 27]. This metabolic bypass may help circumvent host hexokinase inhibition [32], alleviate osmotic stress, and reinforce cell wall integrity through increased N-acetylglucosamine biosynthesis. Gene set enrichment analysis (GSEA) further confirmed activation of secondary metabolic processes (GO:0019748, NES = 1.7719, FDR = 0.05611) and carbohydrate homeostasis (GO:0033500, NES = 1.58, FDR = 0.100) (Fig. 1E, H). Collectively, these results suggest that in conventional infection (without SLDI treatment), O. sinensis may maintain their secondary metabolites at a low level to facilitate the establishment and maintenance of a stable, long-term symbiotic relationship.

Fig. 1.

Transcriptional reprogramming of O. sinensis during early stage infection (30 dpi) under sublethal dose of imidacloprid exposure. A, B Heatmaps showing the expression profiles of key upregulated genes (A) and key downregulated genes involved in cell proliferation (B) in the early-stage treatment (S_TR) versus control (CK) groups. Gene expression levels are represented by log₂(read counts + 1) values, with red indicating higher expression and blue indicating lower expression. C GO enrichment analysis of the DEGs. The bar chart displays the number of genes associated with the most significant GO terms, categorized into Biological Process (green), Cellular Component (blue), and Molecular Function (orange). D KEGG pathway enrichment analysis of the DEGs. The bubble plot shows the most significantly enriched pathways. The Gene Ratio (x-axis) represents the proportion of DEGs in a given pathway, the size of each bubble corresponds to the number of enriched genes (Counts), and the color indicates the statistical significance (p-value). E Gene ranking distribution for the top 5 activated (red) and top 5 suppressed (blue) pathways identified by Gene Set Enrichment Analysis (GSEA). F GSEA plot illustrating the significant suppression of the mitotic nuclear division pathway (GO:0140014; NES = −2.57044, FDR = 0.0566). G GSEA plot illustrating the significant suppression of the cytosolic ribosome pathway (GO:0022626; NES = −2.9008, FDR = 0.0566). (H) GSEA plot illustrating the significant activation of the secondary metabolic process pathway (GO:0019748; NES = 1.7719, FDR = 0.05611)

In addition, several upregulated genes were associated with immune defense. For instance, L-tryptophan decarboxylase (PsiD-like, IPR022237) was induced (Fig. 1A), suggesting enhanced synthesis of indole-derived secondary metabolites that may protect against host immune attack [15]. Members of glycoside hydrolase family 38 (GH38, IPR000602), which hydrolyze α-mannosidic linkages in N-glycans, were also upregulated (Fig. 1A), potentially altering fungal cell wall glycoproteins to evade immune recognition [19]. Furthermore, guanyl-specific ribonucleases (N1/T1) were induced (Fig. 1A), which may degrade extracellular RNAs, including host-derived antifungal RNAi effectors [23]. These findings suggest that SLDI disrupts host immunity, leading O. sinensis to enhance metabolic pathways that counteract immune defenses. From this, we can infer that under conventional infection, O. sinensis appears to reduce its virulence for establishment a balance withless-antagonistic interaction with host in the early colonization phase, which may be crucial for maintaining long-term symbiosis.

Conversely, downregulated genes under SLDI treatment were mainly associated with cell proliferation. Expression profiles indicated significant suppression of cell cycle progression. Key regulators of mitosis, including spindle pole body protein (Cut12/SUN family, IPR021589), DNA replication initiation factor Sld2 (IPR040203), and topoisomerase IIA (IPR001241), were strongly downregulated (Fig. 1B). Additionally, condensin complex subunits, topoisomerase II, and cytokinesis regulators such as Anillin/CDC14 were suppressed (Fig. 1B). GO enrichment analysis also revealed the significant enrichment of numerous terms related to cell division, including mitotic nuclear division (GO:0140014, p = 2.34E-08) and mitotic spindle organization (GO:0007052, p = 3.11E-09) (Fig. 1C). GSEA confirmed significant inhibition of mitotic nuclear division (GO:0140014, NES = –2.57044, FDR = 0.0566) and ribosome biogenesis (GO:0022626, NES = –2.9008, FDR = 0.0566) (Fig. 1E-G). These findings indicate that SLDI treatment suppressed genes critical for DNA replication, mitotic progression, and cytokinesis. Based on the transcriptomic shifts induced by imidacloprid, we infer that the early phase of conventional infection is defined by active blastospore proliferation, corroborating previous morphological reports [24, 26]. This high blastospore proliferation rate is corroborated by metabolic evidence, which reveals high concentrations of DNA synthesis precursors (dAMP, cluster3 yellow module, Fig. 4A, C, I, O). Such blastospore budding is essential for establishing fungal density within the larval hemocoel, creating a foundation for long-term persistent infection. These results suggest that O. sinensis creates a foundation for long-term symbiosis in the early infection stage by activating cell division, maintaining secondary metabolism and immunity at a low level.

Fig. 4.

Time-course metabolomic analysis identifies key metabolic modules and compounds. A, B Mfuzz clustering of metabolites detected in positive (A) and negative (B) ion modes, grouped into six distinct temporal profiles based on Z-scores. Blue lines represent individual metabolites, and red lines show the average profile for each cluster. C–H Heatmaps of the correlation between module eigengenes (MEs) from WGCNA and the infection time for Cluster 1 (C), Cluster 2 (D), Cluster 3 (E), Cluster 4 (F), Cluster 5 (G), and Cluster 6 (H). Each cell displays the Pearson correlation coefficient (top) and p-value (in parentheses). Red and blue indicate positive and negative correlations, respectively. I–N Scatter plots of Module Membership (MM) vs. Metabolite Significance (MS) for time in selected modules, with their source clusters indicated: the yellow module from Cluster 3 (I), the brown module from Cluster 1 (J), the green module from Cluster 1 (K), the blue module from Cluster 6 (L), the brown module from Cluster 2 (M), and the turquoise module from Cluster 6 (N). Metabolites highlighted in red are identified as hub metabolites within each module due to their high MM and MS values. O Heatmap displaying the relative abundance (Z-score) of representative hub metabolites across the infection time points (0, 30, 90, and 150 dpi). Red indicates high abundance, and blue indicates low abundance

Metabolic and functional reprogramming of O. sinensis during the mid-stage of infection (90 dpi)

At 90 days post-infection (dpi), O. sinensis reached a quasi-steady state within its host, characterized by stabilized fungal biomass and reduced proliferation. Under SLDI treatment, 506 genes were significantly upregulated and 386 were downregulated compared with untreated controls. Of the upregulated genes, 339 were successfully annotated. Many were linked to secondary metabolism, including terpene cyclase (IPR034686) and trichodiene synthase (IPR024652), which are essential for terpenoid biosynthesis and mycotoxin production [5, 13, 25, 28, 31]. Genes involved in glycolipid metabolism (e.g., sorbitol dehydrogenase, glucosamine-6-phosphate isomerase) and cell wall remodeling (e.g., glycoside hydrolase family 64 [IPR032477], α-mannosidase [IPR051323]) were also significantly induced, suggesting continued redirection of carbon flux toward N-acetylglucosamine biosynthesis [12]. This pathway reinforces the fungal cell wall, maintains osmotic balance, and mitigates host-imposed metabolic stress (Fig. 2A and 2B). The upregulation of glycoside hydrolase family 64 and α-mannosidase by SLDI suggests active remodeling of N-glycan structures on fungal cell wall proteins, which can reduce detection by host immune receptors [19]. In addition, numerous DEGs were associated with ribosome function, including amide biosynthesis (GO:0043604, p = 5.95E-08), ribosome biogenesis (GO:0042254, p = 8.06E-09), peptide metabolism (GO:0043043, p = 6.83E-08), ribosome (GO:0005840, p = 1.36E-13) (Fig. 2C), KEGG analysis also confirmed enrichment in ribosome assembly (Fig. 2D). GSEA revealed that SLDI treatment significantly activated preribosome assembly (GO:0030684, NES = 2.23, FDR = 0.046) and 5.8S rRNA maturation (GO:0000460, NES = 2.18, FDR = 0.046), while suppressing vesicle-mediated ER-to-Golgi transport (GO:0006888, NES = –2.01, FDR = 0.054) (Fig. 2E–H). These transcriptomic shifts under SLDI treatment imply that during the mid-stage of conventional infection, O. sinensis maintains defensive patterns similar to those in the early stage while further restricting protein biosynthesis to sustain the biotrophic equilibrium.

Fig. 2.

Transcriptional reprogramming of O. sinensis during mid stage infection (90 dpi) under sublethal dose of imidacloprid exposure. A, B Heatmaps showing the expression profiles of key upregulated (A) and downregulated (B) genes in the mid-stage treatment (M_TR) versus control (CK) groups. Gene expression levels are represented by log₂(read counts + 1) values, with red indicating higher expression and blue indicating lower expression. (C) GO enrichment analysis of the DEGs. The bar chart displays the number of genes associated with the most significant GO terms, categorized into Biological Process (green), Cellular Component (blue), and Molecular Function (orange). D KEGG pathway enrichment analysis of the DEGs. The bubble plot shows the most significantly enriched pathways. The Gene Ratio (x-axis) represents the proportion of DEGs in a given pathway, the size of each bubble corresponds to the number of enriched genes (Counts), and the color indicates the statistical significance (p-value). E Gene ranking distribution for the top 5 activated (red) and top 5 suppressed (blue) pathways identified by Gene Set Enrichment Analysis (GSEA). (F) GSEA plot illustrating the significant activation of the preribosome assembly pathway (GO:0030684; NES = 2.23, FDR = 0.046). G GSEA plot illustrating the activation of the 5.8S rRNA maturation pathway (GO:0000460; NES = 2.18, FDR = 0.046), confirming a boost in ribosomal biogenesis. H GSEA plot illustrating the suppression of the endoplasmic reticulum to Golgi vesicle-mediated transport pathway (GO:0006888; NES = −2.0076, FDR = 0.05383), indicating a shutdown of the secretory pathway

Although fungal biomass remained stable at this stage, transcriptional profiling under SLDI exposure revealed suppression of key cell cycle regulators, including mitotic spindle assembly factors and DNA replication initiators. Thus, in the conventional persistent infection, O. sinensis does not entirely halt proliferation but instead sustains a basal level of blastospore production to maintain population stability within the host. This aligns with previous findings that blastospore numbers remain essentially unchanged at 90 dpi [24, 26]. The metabolic analysis also suggests that O. sinensis sustained a low- profile, energy-conserving standoff with the host, as the immunomodulator Albiflorin [30], plant hormone analogs (IBA, Kinetin) [4, 33], and ADP (cluster 1, green module, Fig. 4A, C, J, O) are rapidly cleared or their production is suppressed in this stage.

In summary, during the mid-stage of conventional infection, O. sinensis adopts similar responses with the early infection stage, but suppressed protein biosynthesis. Unlike rapid proliferation in the early stage, O. sinensis engages in low-level blastospore proliferation to maintain fungal population density within the host. These results reflect the fungus’s adaptation to a long-term, balanced coexistence with its host.

Transcriptional response of O. sinensis to imidacloprid during the terminal stage (150 dpi)

At 150 days post-infection (dpi), O. sinensis entered the final infection stage, marked by pronounced host deterioration and substantial lipid droplet accumulation within fungal cells, an essential energy reserve for imminent hyphal germination and stroma development. Transcriptomic profiling under SLDI exposure identified 188 DEGs, including 125 upregulated and 63 downregulated genes. Upregulated genes were related with ketone body involving in fatty acid degrading and aromatic compound catabolism, including acetoacetate decarboxylase (IPR023375) and prephenate dehydratase (IPR001086) [16, 18, 36]. Genes involved in secondary metabolism (e.g., terpene cyclase-like 2, IPR034686), nutrient transport (e.g., malic acid transporter, IPR030185,iron transporter Fet3-like, IPR044130), and cell wall degradation (e.g., glycoside hydrolase 64, IPR032477; β−1,3-glucanase, IPR037398) were also significantly induced. In contrast, fatty acid synthesis genes, including fatty acid synthase type I (IPR041550) and fungal FAS (IPR050830) were consistently downregulated (Fig. 3A and 3B). The GO term lipid biosynthesis (GO:0008610, p = 0.007, Fig. 3C) was significantly enriched among the DEGs. KEGG pathway analysis also showed enrichment in lipid metabolism (ko09103, p = 0.0185, Fig. 3D). These observations that SLDI exposure suppresses lipid synthesis leads us to infer that, during conventional infection, O. sinensis actively enhance fatty acid biosynthesis and inhibit its degradation. This is essential for accumulating the reserves needed to fuel morphogenesis from blastospore to pre-hyphae. This is directly validated by metabolomic data, which identifies the massive accumulation of triglycerides (specifically TG 12:0_12:0_14:0, Fig. 4 O) as the core metabolic event in this final phase. This result is consistent with previous findings showing that lipid droplet accumulation in the pre-hyphal (PreHy) stage is indispensable for energy storage required during mycelial germination [24, 26].

Fig. 3.

Transcriptional reprogramming of Ophiocordyceps sinensis during late stage infection (150 dpi) under sublethal dose of imidacloprid exposure. A, B Heatmaps showing the expression profiles of key upregulated genes (A) and downregulated lipogenesis-related genes (B). Gene expression levels are represented by log₂(read counts + 1) values, with red indicating higher expression and blue indicating lower expression. Genes are hierarchically clustered based on their expression patterns. C GO enrichment analysis of the DEGs. The bar chart displays the number of genes associated with the most significant GO terms, which are categorized into Biological Process (green) and Molecular Function (orange). D KEGG pathway enrichment analysis of the DEGs. The bubble plot shows the most significantly enriched pathways. The Gene Ratio (x-axis) represents the proportion of DEGs in a given pathway. The size of each bubble corresponds to the number of enriched genes (Counts), and the color indicates the statistical significance (p-value)

Under SLDI exposure, GO enrichment analyses revealed significant enrichment of hydrolase activity on glycosyl bonds (GO:0016798, p = 7.91E-05), transmembrane transporter activity (GO:0022857, p = 0.024), and general transporter activity (GO:0005215, p = 0.049). Notably, upregulation of malic acid transporters and Fet3-like multicopper oxidases [17, 35] reflected intensified host tissue degradation and efficient nutrient assimilation (Fig. 3C and 3D). This aggressive degradative activity under SLDI-induced stress suggests a stress response, which implies that during conventional infection, these hydrolytic processes might be tightly regulated to balance nutrient acquisition with host viability. No significantly enriched gene sets were detected by GSEA at this stage.

Our metabolomic analysis revealed a significant late-stage accumulation of D-glucosamine-6-phosphate and N-acetylglucosamine-1-phosphate (cluster 6, blue module; Fig. 4B, H, L, O), which are direct precursors for chitin synthesis [4, 6, 33, 45]. This finding points to the activation of cell wall biogenesis, a critical preparatory step for the morphological transition from blastospores to hyphae. The resulting chitin provides the structural material necessary for extensive polarized hyphal growth.

Metabolomic evidence from the terminal phase reveals two sets of chemical changes related to environmental control and self-preservation. First, the detection of compounds with known antimicrobial properties, such as chlorogenic acid (cluster 6, blue module, Fig. 4B, H, L, O) and perillic acid (cluster 6, turquoise module, Fig. 4B, H, N, O) [34],Ch oi et al., 2010), suggests a mechanism to inhibit microbial competitors and secure the nutrient-rich host. Second, the metabolome indicated intense internal oxidative stress, marked by a significant increase in oxidized L-glutathione (cluster 2, brown module, Fig. 4A, D, M, O). Correspondingly, a notable accumulation of antioxidant and anti-inflammatory molecules was also detected, including hesperidin, artesunate and undecylenic acid (cluster2, brown module, Fig. 4A, D, M, O) [1, 10, 37], which likely serves to counteract this stress. The significant accumulation of trehalose-6-phosphate (cluster 6, blue module,Fig. 4B, H, L, O) indicates the synthesis of trehalose. This disaccharide acts as both a key energy reserve and a potent protectant against environmental stresses such as desiccation, freezing, and oxidative stress [20, 22]. Together, these metabolic features suggest a dual function: sanitizing the external environment while neutralizing internal physiological hazards, thereby preserving the host resources for exclusive use.

During the terminal stage of conventional infection, O. sinensis accumulates amounts of lipids (fats) as a primary energy resource and actively produces the building blocks for chitin, which are needed to remodeling cell walls for its transition from blastospores to infection hyphae. Furthermore, the presence of antimicrobial compounds suggests a mechanism to inhibit microbial competitors. Simultaneously, the accumulation of antioxidants and protective sugars, points to a response to the stressful chemical environment in preparation for the host's mummification.

Conclusion

This study reveals the dynamic and stage-specific responses of O. sinensis during its long-term interaction with its insect host. In the early stage of infection, O. sinensis establishes a foundation for this interaction by activating cell division while maintaining its secondary metabolism and immunity at a low level. This is followed by a mid-stage characterized by a stable state, where the fungus engages in low-level proliferation and suppresses protein biosynthesis, likely to reduce immunogenicity while maintaining its presence within the host. Finally, the interaction shifts dramatically in the terminal stage. This phase involves the massive accumulation of lipid reserves for energy and the production of chitin building blocks in preparation for its morphological change from blastospores to hyphae. Furthermore, the chemical environment is remodeled to facilitate this transition, characterized by the secretion of antimicrobial compounds to inhibit competitors and the accumulation of antioxidants and protective sugars to prepare for the host's mummification. This study provides insights into the molecular mechanism for a slow-killing fungus, providing a map of metabolic adaptation and immune evasion that serves as a valuable framework for studying other chronic pathogens. Nevertheless, several limitations should be noted. While our RNA-seq analysis provides a robust transcriptomic overview, independent experimental validation (e.g., via qRT-PCR) was not performed for the identified genes. Additionally, although imidacloprid was employed as an immunomodulatory agent based on its specific action on insect nicotinic acetylcholine receptors (nAChRs), potential direct physiological effects of this compound on O. sinensis cannot be entirely ruled out and should be considered when interpreting the fungal-host interaction dynamics.

Abbreviations

Dpi

Days post infection

O. sinensis

Ophiocordyceps sinensis

T. xiaojinensis

Thitarodes xiaojinensis

BP

Blastospores proliferation stage

BS

Blastospores stationary

PreHy

Pre-hyphae

Hy

Hyphae

GH

Glycoside hydrolases

SLDI

Sublethal dose of imidacloprid

S

Early infection stage

M

Mid infection stage

L

Late infection stage

GC content

Guanine-cytosine content

cDNA

Complementary Deoxyribonucleic Acid

dUTP

Deoxyuridine Triphosphate

dAMP

Deoxyadenosine monophosphate

PCR

Polymerase Chain Reaction

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

GSEA

Gene Set Entichment and Analysis

NES

Normalized enrichment score

RNA

Ribonucleic Acid

DEGs

Differentially expressed genes

FDR

False Discovery Rate

TCA cycle

Tricarboxylic Acid Cycle

RNAi

Ribonucleic Acid Interference

FPP

Farnesyl pyrophosphate

FAD

Flavin Adenine Dinucleotide

FAS

Fatty acid synthase

GSA

Genome Sequence Archive

Authors’ contributions

LTY and ZJY wrote the main manuscript text; ZL and LXJ analyed the data; LWS, LQP and QZM edited the manuscript; XMC. LWJ and LXZ reviewed and edited the manuscript. All authors reviewed the manuscript.

Funding

This work was supported by the technical development contract: Interaction mechanisms and sapplications of Ophiocordyceps sinensis with its host insects from Sunshine Lake Pharma Co., Ltd, Dongguan, China (Contract No. FW-ZY190711-01).

Data availability

The raw sequence data reported in this study have been deposited in the Genome Sequence Archive [42] at the National Genomics Data Center [11], China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences, under accession number CRA026124 (https://ngdc.cncb.ac.cn/gsa). Additionally, the metabolomic data have been deposited in the OMIX database (https://ngdc.cncb.ac.cn/omix) under accession number OMIX012455.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.