H. gigas samples were collected from the Mariana Trench during a research expedition in September 2021 aboard the research vessel “Explore I”, using a macrobiological pressure-retaining sampler. The sampling sites were located at 11°19.6098′ N, 142°11.283′ E (10,895 m depth) and 11°22.584′ N, 142°35.3102′ E (10,910 m depth), where the ambient temperature typically ranges between 1-4°C (Yasuhara and Danovaro, 2016). All samples (n=6) were collected from the above two sites. Immediately after collection, H. gigas specimens were stored at −80°C. To ensure biological representativeness, gut contents from three selected individuals per site were processed independently. Before dissection, the samples were individually thawed at 4°C and rinsed with distilled water. Subsequently, the amphipods were dissected using a sterile scalpel to isolate intestinal tissues, which were then transferred into sterile centrifuge tubes for fungal isolation experiments. Fungi were isolated following the method described in our previous study (Cui et al., 2024). Briefly, 200 µL of diluted gut contents were spread onto potato dextrose agar (PDA), tryptone soy agar (TSA), and malt extract agar (MEA) plates and incubated at 20°C for 2–3 days. Visible colonies were transferred to fresh medium for purification, and single colonies were obtained through three rounds of subculturing.

Fungal identification was conducted based on a combination of morphological characteristics and conserved sequence analysis. Pure fungal cultures were grown on PDA at 28°C for 3 days. Fungal hyphae were collected from petri dishes and strain DNA was extracted using the TIAN combi DNA Lyse & Det PCR kit [Tiangen Biotech (Beijing) Co. Ltd.], following the manufacturer’s instructions. PCR amplification and sequencing were performed using the following primers: ITS1 and ITS4 for the internal transcribed spacer (ITS) region, and RPB2AM-1bf and RPB2AM-7R (Miller and Huhndorf, 2005) for the second largest subunit of the DNA-directed RNA polymerase II (RPB2) gene region (all primers used in this study were listed in Supplementary Table S1 ).

PCR amplification was performed under the following conditions: initial denaturation at 95°C for 3 min; 35 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min; final extension at 72°C for 5 min; and storage at 4°C (Deng et al., 2023). The amplified ITS and RPB2 sequences were subjected to a sequence similarity search using BLASTn against the NCBI nucleotide database. Highly similar sequences were aligned and comparatively analyzed using MEGA X. All sequences were uniformly trimmed at both 5’ and 3’ ends to ensure positional homology in the alignment. Phylogenetic trees were performed using the neighbor-joining algorithm in MEGA X, with nodal support evaluated through 1,000 bootstrap replications. Microscopic characterization of the target fungus was performed using an OLYMPUS BX53 microscope (Olympus Corporation, Tokyo, Japan) to observe the ascomycete fruiting bodies.

Screening for cellulase-producing fungi was conducted using a plate assay method. C. madrasense strains were initially screened for cellulase activity by culturing on agar medium supplemented with 1% carboxymethyl cellulose (CMC) as the sole carbon source. The plates were incubated at 28°C for 5 days. After incubation, the plates were stained with 1% Congo red staining solution for 15 min, followed by destaining with 1 M NaCl buffer solution for 15 min (n=3 biological replicates). The formation of a clear zone around the fungal colony was considered indicative of cellulase activity. Cellulolytic fungi were identified based on their ability to hydrolyze cellulose, as evidenced by the presence of a clearance zone surrounding the colony (Khokhar et al., 2012).

Fresh mycelial plugs (2 mm²) obtained from the edges of fungal colonies were inoculated onto PDA (0 M NaCl, 0.5 M NaCl) and 0 M NaCl microcrystalline cellulose agar (MCC) ( Supplementary Table S2 ), with 3 biological replicates set up for each condition. All plates were incubated at 28°C for 3 days, and colony morphology was observed. Additionally, mycelial fragments (2 mm²) were excised from the edges of actively growing colonies of each isolate and inoculated into the center of 0 M NaCl MCC and the plates were incubated at 15°C, 20°C, and 28°C (n=3 replicates per temperature). Colony diameters were measured daily for 3 days using a measuring scale, with three perpendicular measurements per colony to minimize observer bias.

Mycelial discs of uniform size were inoculated into an enrichment medium and incubated for 2 days. Subsequently, 1% of the enriched culture was transferred to MCC medium and further incubated at 180 rpm in a rotary shaker (n=3 biological replicates). Endoglucanase activity was quantified using the 3,5-dinitrosalicylic acid reagent (DNS) assay to measure the concentration of reducing sugars released by cellulase hydrolysis. The methods for measuring cellulase activity were performed in accordance with those described by Wei (2007). A 0.5 mL enzyme solution was mixed with 1 mL of 1% CMC-Na buffer solution and incubated at 50°C for 30 min. After terminating the reaction by adding 1.5 mL DNS and heating at 100°C for 10 min, the samples were immediately cooled on ice to minimize nonspecific color development. The absorbance at 540 nm was then measured (n=3 biological replicates). Substitute the absorbance into the glucose standard curve and calculate the glucose concentration ( Supplementary Figure S2 ). One unit of cellulase activity (U) was defined as the amount of enzyme required to catalyze the release of 1 μmol of glucose from 1% CMC-Na under specific reaction conditions (pH 4.8, 50°C, 30 min). The enzyme unit (symbol U), also known as the international unit (IU), is a measure of an enzyme’s catalytic activity (Helal et al., 2021).

To perform RNA sequencing and transcriptomics analysis, C. madrasense was incubated at 180 rpm at 15°C on MCC medium for 6 days, after which biomass was harvested. The mycelia were fully ground using liquid nitrogen and preserved in RNAiso Plus (Takara, Japan) at −80°C. The RNA extraction, transcriptomic sequencing, and bioinformatics analysis were accomplished by Novogene Bioinformatics Technology (Beijing, China). RNA integrity was assessed using an Agilent 2100 Bioanalyzer. Following standard protocols, high-quality RNA was used for mRNA enrichment, fragmentation, reverse transcription, end repair, amplification, and circularization to construct the sequencing library. After passing quality control, the library was loaded onto the Illumina sequencing platform for high-throughput sequencing. To minimize batch effects, all samples were harvested simultaneously, processed in the same RNA extraction batch, and sequenced in a single Illumina run. Gene expression levels were quantified as FPKM (fragments per kilobase of transcript per million mapped reads) using RSEM (v1.3.1). Analysis of differentially expressed genes (DEGs) was performed using DESeq2 (1.20.0). Q value ≤ 0.05 and |log2 Fold-Change| ≥ 1.0 were set as the threshold for significant differential expression.

To validate the reliability of the RNA-Seq results, selected differentially expressed genes were used to design primers for quantitative reverse transcription PCR (RT-qPCR) analysis. Total RNA was obtained from Novozymes and synthesized according to the instructions from the manufacturer of the PrimeScript RT Reagent Kit (Takara, Japan). Gene expression levels were quantified using the 2^-ΔΔCt method, and the fold changes obtained from RNA-Seq were compared with those from RT-qPCR. The correlation coefficient (R²) was calculated to assess the consistency between the two methods. β-actin was used as the internal reference gene, and the primer sequences are provided in Supplementary Table S1 .

All experiments were performed at least three times, with three biological replicates per sample. All data were analyzed using SPSS Statistics 27 for Windows (SPSS Inc.). After testing for normality using the Shapiro-Wilk test and Levene’s test for all data, one-way analysis of variance (ANOVA) and Tukey’s multiple comparison test were used to assess differences between groups. Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, while non-significant differences were denoted as n.s (P > 0.05). Results are presented as the mean ± standard error (SD) of three replicates. For multiple comparisons, we applied the Benjamini-Hochberg false discovery rate (FDR) correction with Q < 0.05 as the significance threshold. The line and bar graphs were drawn using Origin 2024b. The heatmap was drawn using TBtools software. Multi-group difference scatterplots were plotted using Omicshare tool.

An abyssal fungal strain was successfully isolated from the gut microbiota of amphipods inhabiting the Mariana Trench. Phylogenetic characterization based on ITS and RPB2 of conserved sequences confirmed the taxonomic classification of strain HM411 as C. madrasense ( Figure 1A ).

The morphological characteristics of C. madrasense HM411 were systematically investigated. When cultured on 0.5 M NaCl PDA, the strain exhibited white aerial hyphae ( Figure 1B ) and produced distinctive yellow exudates in the central region during prolonged incubation ( Figure 1C ). The colony demonstrated robust growth, reaching 6.50 ± 0.12 cm (n=3 biological replicates) in diameter following 7 days of incubation at 28°C. Microscopic examination revealed the presence of typical ascomata and ascospores ( Figures 1D, E ), with epidermal, spherical structures delineated by the filaments. The mature ascospores displayed an olivaceous-brown pigmentation and exhibited ellipsoidal to subglobose morphology.

As members of the genus Chaetomium are widely recognized for their cellulolytic capabilities in both cellulosic waste degradation and single cell protein (SCP) production (Darwish and Abdel-Azeem, 2020), we evaluated the cellulolytic potential of C. madrasense HM411. The strain demonstrated sustained growth on CMC agar and formed distinct cellulose-hydrolyzed hyaline rings upon Congo red staining, thereby confirming its cellulose-degrading capacity. These findings substantiate the cellulolytic potential of this abyssal fungal strain ( Figures 1F, G ).

To evaluate the cellulose-degrading specificity of hadal-derived C. madrasense, a comparative terrestrial strain (C. madrasense CGMCC 3.17112), originally isolated from cotton, was obtained from the China General Microbiological Culture Collection Center (CGMCC) and designated as C. madrasense HM412 ( Supplementary Figure S1 ). In the same medium, both strains exhibited similar colony morphology. Visually, the growth diameter of C. madrasense in 0 M NaCl MCC appeared larger than that in (0 M, 0.5 M NaCl) PDA under the same incubation conditions ( Figure 2A ), suggesting that cellulose may be a more favorable carbon source for C. madrasense than PDA components (n=3 biological replicates).

To assess the impact of temperature on the growth of the strains, we inoculated fungus discs of the same size on 0 M NaCl MCC, and measured the growth diameters of the two strains at different temperatures, with three biological replicates set up for each strain. The mean colony diameters of C. madrasense from marine and terrestrial sources, calculated from three independent replicates, were 3.66 ± 0.03 cm and 3.96 ± 0.03 cm, respectively, after 72 hours of incubation at 28°C. Whereas, the diameter of C. madrasense colonies of marine and terrestrial origin was 1.13 ± 0.02 cm and 0.98 ± 0.02 cm (n=6 measurements from 3 biological replicates). after 72 hours of incubation at 15°C, respectively ( Figure 2B ). The colony diameter of C. madrasense HM412 was larger than that of C. madrasense HM411 (p < 0.0001) under the incubation condition at 28°C, whereas at 15°C, the larger colony diameter was that of C. madrasense HM411 (p < 0.001), which might indicate that C. madrasense HM411 showed greater low-temperature tolerance than terrestrial source strains.

Temperature represents a critical environmental factor influencing the viability and enzymatic productivity of cellulase-producing microbial strains (Sohail et al., 2009; Rosyida et al., 2015). To evaluate the thermal adaptability of C. madrasense HM411, endoglucanase production was assessed across a temperature gradient (10, 15, 20, 25, and 28°C). The experimental results demonstrated that the optimal temperature range for endoglucanase production by C. madrasense HM411 was 25-28°C, with enzyme synthesis reaching a stabilization phase after 5 days of incubation. Notably, the strain exhibited remarkable psychrotolerant characteristics, maintaining metabolic activity and enzymatic production at 15°C, albeit with an extended stabilization phase observed at 16 days. These findings highlight the strain’s adaptive capacity to suboptimal thermal conditions while maintaining cellulolytic functionality. While, until the end of the experiment (i.e., 18 days), no production of endoglucanase was detected under 10°C culture conditions ( Figure 3A ).

To investigate the cellulolytic potential of two C. madrasense strains from different origins under low-temperature conditions, we evaluated their endoglucanase production capabilities at 15°C. Comparative analysis revealed that C. madrasense HM411 showed significantly higher enzymatic productivity than C. madrasense HM412 under identical conditions (p < 0.01). Specifically, C. madrasense HM412 reached the enzyme production stabilization phase at 22 days, demonstrating endoglucanase activity levels of 71.3 ± 3.4 U/mL (n=3 biological replicates), while C. madrasense HM411 achieved significantly higher activity levels of 87.5 ± 1.8 U/mL (n=3 biological replicates) during the same period ( Figure 3B ).

C. madrasense HM411 demonstrated superior endoglucanase activity and a significantly reduced fermentation period at 15°C compared to C. madrasense HM412. To elucidate the molecular mechanisms underlying the enhanced cellulase production and enzymatic activity of C. madrasense HM411 under low-temperature conditions, we conducted a comparative transcriptomic analysis of the two C. madrasense strains from distinct origins.

The reliability of the RNA-Seq data was supported by the raw read information ( Supplementary Table S4 ) and the high correlation and consistency observed among the three biological replicates ( Supplementary Figure S3A ). The RT-qPCR analysis results reflected that the trend of transcript level changes of these genes was consistent with those obtained from the comparative transcriptome analysis ( Supplementary Figure S4 ).

Comparative transcriptome analysis showed that there were 1,788 DEGs between C. madrasense HM411 and C. madrasense HM412 at 6 days of cellulose induction in 15°C (|log2FC| ≥ 1 and adjusted padj ≤ 0.05), of which 1,089 genes were up-regulated, and 699 genes were down-regulated ( Supplementary Figure S3B ).

We performed Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) enrichment analysis on DEGs and selected the top 10 items for further enrichment. These DEGs are mainly involved in protein synthesis, processing and degradation, primary metabolism, small molecule metabolism such as galactose, sucrose, fructose, aminosugar, and nucleotide sugar metabolism. In addition, the enrichment of differentially expressed genes in amino sugar and nucleotide sugar metabolism, starch and sucrose metabolism, and fructose and mannose metabolism may reflect enhanced metabolic flexibility in C. madrasense HM411, enabling better adaptation to low-temperature conditions through coordinated regulation of multiple metabolic pathways. This regulation of sugar metabolism may also enhance its metabolic efficiency and adaptive capacity ( Figure 4A ). GO enrichment analysis showed that genes involved in the regulation of intracellular metabolic activities, degradation of macromolecular substances were significantly enriched, in addition to genes related to transcription, regulation, protein synthesis, metabolism and transport using DNA as a template ( Figure 4B ). Energy metabolism is important for protein synthesis (Huang et al., 2017), and we hypothesize that up-regulation of related genes may benefit C. madrasense HM411 to resist low-temperature stress and produce cellulase. Next, TBtools (v1.120) was used to draw a heatmap containing the main marker genes associated with strains that resist low temperatures and produce enzymes. These include marker genes encoding enzymes associated with strain secretion. Notably, genes encoding secreted enzymes (e.g., endoglucanase CBH1, chitinase E3.2.1.14) were upregulated in C. madrasnese HM411. Additionally, genes associated with glycerol and choline synthesis, such as alcohol dehydrogenase (adh) and acetylcholinesterase (ACHE), were identified. These genes have been attributed to extracellular protein secretion and intracellular metabolism, including cell wall components, and cell membranes ( Figure 4C ), which may be related to cellulase production by C. madrasense HM411 at low temperatures.

CAZymes are key enzymes for nutrient conversion and utilization in fungi, mainly including 6 enzyme families: auxiliary activities (AAs), glycosyl hydrolases (GHs), carbohydrate binding modules (CBMs), carbohydrate esterases (CEs), glycosyl transferases (GTs), and polysaccharide lyases (PLs) (Hao et al., 2019; Li et al., 2020). We identified 378 genes as CAZyme genes by comparison with the CAZy database, of which 268 were up-regulated and 110 were down-regulated in expression ( Figure 5A ), including 37 secreted proteins. In C. madrasense HM411, the differential genes were mainly concentrated in the AA and GH families, in which the expression of glycoside hydrolase genes, GH5, GH6, GH7, GH11, GH16, GH18, and GH43 families, was significantly up-regulated (log₂FC > 1, padj < 0.05) ( Supplementary Table S5 ), and the major cellulase genes were distributed in the GH5, GH6, GH7, GH9 families.

Cellulose degradation is carried out by three classes of enzymes, mainly endoglucanase (EG) (EC 3.2.1.4), cellobiose hydrolase (CBH) (EC 3.2.1.91), and β-glucosidase (BG) (EC 3.2.1.21) (Henriksson et al., 2000; Yuan et al., 2001; Amer et al., 2017). In addition, oxidoreductases such as cellobiose dehydrogenase (CDH) (EC 1.1.99.18) may act synergistically with classical GH to accelerate the enzymatic conversion of polysaccharides (Cameron and Aust, 2001). Significant up-regulation of CELB (log₂FC = 5.69, padj < 0.05), CBH1 (log₂FC = 3.11, padj < 0.05), and bglX (log₂FC = 2.13, padj < 0.05) involved in the cellulose metabolism pathway was detected in RNA-Seq, corresponding to EG, CBH, and BG, respectively ( Figure 5B ), in addition to CDH (log₂FC = 3.46, padj < 0.05), a cellobiose dehydrogenase that accelerates the enzymatic conversion of polysaccharides ( Supplementary Table S6 ). The degradation product of cellulose, glucose, is specifically recognized and bound by glucose transport proteins and enters the cell by changing its own conformation to participate in glycolysis and ethanol fermentation pathways. The up-regulation of HK (log₂FC = 1.08, padj < 0.05) and galM (log₂FC = 4.76, padj < 0.05) genes within the starch and sucrose metabolic pathways, along with the enrichment of GAPDH, ALDH, and AKR1A1 genes in the glycolytic pathway, was observed. These findings suggest that C. madrasense HM411 cells may possess enhanced capabilities for cellulose degradation and ethanol synthesis, thereby providing efficient energy metabolism for cellular functions.

Cellulase secretion begins with the transcription of DNA in the nucleus, and in the RNA-Seq we detected up-regulation of the TFIIH2 (log₂FC = 1.22, padj < 0.05) and XPB (log₂FC = 1.34, padj < 0.05) genes. TFIIH2 assists RNA polymerase II to bind to the promoter region of the gene, promotes the formation of the transcription initiation complex, and stabilizes the structure of the transcription initiation complex (Schilbach et al., 2017). Up-regulation of RP-L30 (log₂FC = 1.18, padj < 0.05) and RP-L24 (log₂FC = 1.09, padj < 0.05) genes involved in encoding the 60S large subunit of the ribosome was detected, and the L30 protein helps to maintain the normal function and stability of the ribosome and ensures that the translation process proceeds efficiently (White et al., 2004). The L24 protein helps maintain ribosome activity and accuracy, ensuring proper decoding of mRNA and accurate amino acid incorporation (Dresios et al., 2000). The up-regulation of TFIIH2, RP-L30, and RP-L24 genes may indicate that C. madrasense HM411 has a better DNA transcription capacity than C. madrasense HM412 under the same low temperature conditions.

Peptides entering the ER lumen need to be folded with the assistance of the binding protein BiP1, a homolog of Kar2 in Streptococcus that plays a role in the folding of ER proteins (Seppä and Makarow, 2005). Thus, up-regulation of the BIP (log₂FC = 1.28, padj < 0.05) gene may help C. madrasense HM411 maintain intracellular protein homeostasis. Unfortunately, endoplasmic reticulum stress is induced when misfolded proteins accumulate in the endoplasmic reticulum. IRE1 is able to sense this stress signal and, through its own nucleic acid endonuclease activity, splices the X-box-binding protein 1 (XBP1) mRNA to convert it into the active XBP1 protein, a potent transcription factor that enters the nucleus to activate a series of factors involved in ER protein folding, mass control, lipid synthesis and transport, among other processes (Xu et al., 2016). Therefore, we hypothesized that C. madrasense HM411 attempted to restore protein folding homeostasis by enhancing the molecular chaperone-assisted folding function through the up-regulated BIP gene. And there were no significant changes in PDI, ERO1 and IRE1 genes, indicating that oxidative protein folding and IRE1-mediated UPR signaling pathway in C. madrasense HM411 cells are relatively stable at the current stage and have not yet been significantly activated, and that the cells still have a certain ability to cope with protein folding under the existing mechanism. Folded proteins undergo complex assembly and modification in preparation for vesicle transport, which begins with vesicle growth.

In our results, low-temperature stress induced changes in the transcript levels of DEGs associated with cell wall components ( Figure 6A ). We detected the activated expression of the genes related to chitinase synthesis, which attack the crystal structure of chitin by disrupting glycosidic glycoconjugates, which further leads to a decrease in chitin content within the cell wall. The β-mannosidase-encoding gene (manB) showed up-regulated expression (log₂FC = 1.12, padj < 0.05). Such an enzyme is indispensable for the complete degradation of mannans. Moreover, the decline in mannans will result in a reduction in mannoprotein molecules. Mannan is a macromolecular polysaccharide that is composed of mannose and glucose via β-(1-2) glycosidic linkages and β-(1-3) glycosidic linkages. The expression of genes related to mannose metabolizing enzymes will decrease the mannose content, thereby influencing the formation of mannan. In addition, key genes responsible for prolonging β-1,3-glucan were down-regulated. Taken together, we hypothesize that reduced cell wall stability may favor improved tolerance to low temperatures. Furthermore, the key genes accountable for elongating β-1,3-glucan were down-regulated. All things considered, we put forward the hypothesis that the decreased stability of the cell wall might contribute to enhanced tolerance to low temperatures.

Low temperature or heat stress is capable of disrupting the stability and fluidity of the membrane. Meanwhile, the stability and fluidity of the membrane dictate the cell’s tolerance towards stress conditions and directly influence a multitude of membrane associated physiological processes (Santos et al., 2019). KEGG analysis of DEGs showed enrichment of glycerolipid and glycerophospholipid metabolic pathways involved in the response to low temperature stress. Enrichment of genes associated with the glycerol ester and glycerophospholipid metabolic pathways was observed. The upregulation of ALDH (log₂FC = 1.17, padj < 0.05), adh (log₂FC = 2.82, padj < 0.05), and DAK (log₂FC = 4.86, padj < 0.05) expression in the glyceride metabolic pathway led to increased intracellular glycerophospholipid levels. Glycerophospholipid acyl ketone participates in pyruvate synthesis during glycolysis, which subsequently promotes the tricarboxylic acid cycle (TCA) to generate substantial ATP through substrate phosphorylation, thereby supplying energy for cellular activities in fungal cells. In the glycerophospholipid metabolic pathway, the upregulation of GDE1 enhances the hydrolysis of glycerophosphodiester and accelerates glycerophospholipid metabolism. This may result in alterations in intracellular glycerophospholipid levels, potentially influencing the composition and fluidity of cell membranes.

To verify the above analysis, we incubated C. madrasense HM411 and HM412 at 15°C and 28°C for 6 days ( Figure 6B ). Subsequently, RNA was extracted, and the expression of genes involved in the synthesis of E3.2.1.14, manB, adh, and ACHE were determined using β-actin as an internal reference. It should be noted that the relative expression of manB (7.6 ± 2.5-fold, ΔΔCt = -2.62 ± 0.33, n=3 biological replicates), adh (11.4 ± 4.6-fold, ΔΔCt = -2.48 ± 0.30, n=3 biological replicates), E3.2.1.14 (1.2 ± 0.2-fold, ΔΔCt = -4.77 ± 0.15, n=3 biological replicates) and ACHE (30.3 ± 9.8-fold, ΔΔCt = -6.41 ± 0.40, n=3 biological replicates) genes in HM411 at 15°C was higher than that at 28°C. The increased relative expression of manB and E3.2.1.14 may directly influence intracellular chitin content. This could lead to reduced cell wall stability in C. madrasense HM411, and decreased obstruction of extracellular substances. Consequently, the uptake of nutrients, such as glucose and amino acids, into the cell may be facilitated. Glucose, amino acids and other nutrients enter the cell smoothly, and enhance intracellular material metabolic activities. The increased expression of adh and ACHE genes may have enhanced cell membrane fluidity, facilitating the movement of various transporter proteins and signal receptors. This improved membrane dynamics enables cells to more rapidly and accurately perceive external low-temperature signals, which are subsequently transduced intracellularly to activate the expression of relevant genes and initiate the cellulase synthesis pathway.