Genome-wide transcriptomic analysis reveals autophagy and HSP70-mediated mechanisms underlying thermal stress adaptation in Mytilus coruscus Authors: Chao Li, Siying Lv, Hua Jing, Heng Zhou, Xiao Liu Published: 2025 ABSTRACT Rising ocean temperatures driven by global warming pose a major threat to marine organisms. In this study, we examined physiological and molecular changes in the muscle, gill, foot, and mantle tissues of Mytilus coruscus under thermal stress, demonstrating that elevated temperatures significantly induce molecular heat shock responses and oxidative stress. Transcriptomic analysis further revealed conserved and divergent transcriptional responses across these tissues at the genome-wide level, implicating stress-related processes such as autophagy, heat response, and ROS regulation. Eight autophagy-related genes (ATGs) and five heat shock protein 70 genes (HSP70) were identified as significantly upregulated during thermal stress, with their heat resistance confirmed by functional analysis in Saccharomyces cerevisiae. Protein interaction analysis showed that ATG8c interacts with four HSP70 proteins, suggesting a cooperative role of autophagy and the heat shock response in regulating thermal stress. Collectively, these findings reveal organ-specific thermal stress responses in M. coruscus and highlight autophagy and HSP70 pathways as potential molecular targets for studying thermal adaptation in sessile marine species.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12864-025-12423-x. Background Global warming is driving a rise in atmospheric temperatures, which in turn contributes to increasing ocean temperatures. According to the IPCC, mean global sea surface temperatures (SST) are projected to rise by 0.2 °C to 4.4 °C by the end of the 21st century [1, 2]. In the Eastern China Seas, SST has increased by ~ 0.3 °C per decade from 1985 to 2022 [3]. In addition to gradual warming, extreme marine heatwaves are occurring with greater frequency, duration, and intensity worldwide and along China’s coasts [4–7]. Bivalves, a diverse group of mollusks distributed from intertidal to deep-sea habitats, are particularly sensitive to climate-driven thermal stress, which influences their growth, reproduction, feeding, behavior, and respiration [8, 9]. Thermal stress in bivalves may be acute, from sudden temperature spikes, or chronic, from prolonged exposure, both of which can cause sublethal impairments and mortality, with cascading consequences for populations and ecosystems [1]. Temperature is a key factor affecting growth and survival in bivalves, and tissue-specific differences in thermal responses have been widely reported at both physiological and molecular levels [10–14]. For example, heat stress reduces mussel foot protein abundance and impairs byssus production [12]. In addition, gill, mantle, and muscle tissues are particularly important because they are directly exposed to seawater and play critical roles in respiration, protection, and locomotion. At the cellular level, elevated temperatures disrupt homeostasis by altering enzyme activity, protein conformation, and membrane fluidity, leading to increased reactive oxygen species (ROS) and lipid peroxidation, which damage mitochondria and impair organismal health [15, 16]. With the increasing frequency of extreme temperature events, the capacity of mollusk species to tolerate thermal stress is a determinant of their long-term persistence [17, 18]. Autophagy, a conserved cellular process for degradation and recycling, maintains homeostasis under thermal stress [19]. It enhances thermotolerance by mitigating protein misfolding, reducing oxidative stress, and preserving organelle integrity [20]. In human cells, thermal stress increases autophagic activity, clearing damaged proteins and organelles [21], while also providing energy for essential functions during stress [22]. In bivalves, autophagy activity is often induced in response to both biotic and abiotic stresses [23–28]. In Crassostrea gigas, autophagy contributes to pathogen defense, including against Ostreid herpesvirus 1 (OsHV-1) and Vibrio species, supporting hemocyte function [29–31]. Under thermal stress, increased expression of Beclin 1 and LC3 (ATG8) was observed in Patinopecten yessoensis [24], while decreased SQSTM1/p62 levels were detected in Mytilus galloprovincialis [27, 28], together indicating the activation of autophagy under high-temperature conditions. Additionally, thermal stress inhibits mTOR, thereby activating autophagy [32]. The AMPK pathway promotes autophagy by inhibiting mTOR and phosphorylating ULK1 [33], with thermal stress activating AMPK in mollusks [34]. In C. gigas, mTOR inhibition with rapamycin enhances thermotolerance [35], and in Meretrix petechialis, a significant induction of the mTOR-mediated autophagy pathway was observed at 31 °C during Vibrio infection [26]. Autophagy responses also vary across thermal environments, with intertidal mollusks showing higher baseline levels than subtidal species [36–38]. Heat-induced reactive oxygen species (ROS) trigger autophagy [39], activating FoxO transcription factors to upregulate ATG genes [40]. These findings indicate that autophagy is integral to mollusk thermotolerance and adaptation to thermal variability. Heat shock proteins (HSPs), a conserved family of molecular chaperones, protect cells from thermal and other stressors [41–43]. Under thermal stress, HSPs are upregulated, supporting cellular homeostasis by facilitating protein folding, preventing aggregation, and promoting degradation of misfolded proteins [44]. Their role in enhancing thermotolerance is well-documented across animal taxa [45, 46]. In mollusks, HSPs contribute to a suite of physiological and molecular responses to thermal stress, including protein processing, antioxidant defense, immune activation, and anti-apoptotic mechanisms [47, 48]. In Crassostrea gigas, genes encoding HSP70, HSP90, and HSP100 have been identified and sequenced, with their expression induced by thermal stress [10]. Elevated HSP expression under thermal stress is observed across mollusk species, such as Haliotis discus hannai (Pacific abalone), mussels, and oysters [49, 50]. In bivalves, sublethal heat exposure transiently increases tolerance to lethal temperatures, a response associated with elevated HSP70 levels [51]. A similar correlation between HSP70 abundance and thermotolerance is reported in the arthropod Orchesella cincta following heat hardening [52]. HSPs also interact with autophagy pathways; HSP90 stabilizes Beclin-1 to promote autophagosome formation [53], while HSP70 targets damaged proteins for autophagic degradation [54], enhancing cellular resilience to thermal stress. The thick-shelled mussel (Mytilus coruscus), a bivalve native to the Northwest Pacific, is ecologically significant and economically valuable, particularly in Chinese aquaculture [55]. Rising seawater temperatures due to climate change necessitate an understanding of its physiological and molecular responses to thermal stress. As a marine ectotherm, M. coruscus is vulnerable to temperature fluctuations, which affect its survival, growth, and reproduction. For example, elevated temperatures reduce byssal thread production and increase substrate detachment in this species [12]. Given the ecological and economic roles of mollusks and the increasing threat of thermal stress, elucidating their thermal responses has become a research priority. Advances in high-throughput sequencing have accelerated investigations into mollusk adaptation to thermal stress [56, 57]. In this study, M. coruscus was selected as a model species to examine thermal stress response mechanisms in mollusks. We employed an integrated approach, combining multi-omics analysis, yeast stress screening, and gene functional validation, to characterize the thermal stress response of M. coruscus and identify candidate genes for enhancing thermotolerance. These findings contribute to predicting climate change impacts on mollusk populations and informing conservation and management strategies. Materials and methods Animal collection and husbandry Adult mussels (Mytilus coruscus) were collected from Shengshan Island, situated at coordinates 30°43′0.68″ N, 122°48′53.95″ E, in Zhoushan, Zhejiang Province, China. Following collection, the mussels, characterized by an average shell length of 8.3 ± 0.4 cm, shell height of 4.2 ± 0.5 cm, and shell width of 2.6 ± 0.3 cm, were transported to the laboratory. Upon arrival, the specimens were maintained in 100-liter aerated tanks containing natural seawater for an acclimation period of 10 days, during which the average seawater temperature was gradually increased 21 °C. The mussels were provided with a diet consisting of Chaetoceros muelleri and Isochrysis zhanjiangensis. Seawater within the tanks was replaced daily to ensure optimal conditions throughout the acclimation period. Experimental design and sample preparation Our experiment included four treatment groups maintained at 21 °C, 26 °C, 31 °C, and 36 °C. For each treatment, 30 mussels were randomly assigned to three 10-liter aerated tanks (10 mussels per tank, n = 30 per treatment). After 12 h of exposure, nine mussels per tank were dissected, and tissues from foot, gill, mantle, and muscle were collected, flash-frozen, and stored at − 80 °C. The three tanks served as biological replicates (n = 3). Within each replicate, three mussels were used for SOD and CAT activity assays, three mussels for ROS and MDA content assays, and the remaining three mussels for qPCR and transcriptome analysis. The temperature in each treatment group was initially set at 21 °C and incrementally increased by 1 °C every eight hours until the target temperature was achieved. Once reached, the designated temperature was sustained for 12 h prior to sampling. At the 12-hour mark following high-temperature stress exposure, tissues from the foot, gill, mantle, and muscle were excised, immediately flash-frozen in liquid nitrogen, and stored at − 80 °C for subsequent gene expression analysis and physiological assays. Based on the outcomes of the heat stress response gene expression and physiological experiments, tissues from the 21 °C (control) and 31 °C treatment group were selected and collected for transcriptome analysis. Antioxidative enzyme activities For antioxidant activity assays, 2 g samples of foot, gill, mantle, and muscle tissues were obtained by pooling tissues from three randomly selected mussels per tank (n = 3 biological replicates per treatment) to ensure sufficient sample mass and minimize individual variation. Each sample was combined with 18 mL of physiological saline, finely minced, and homogenized in an ice-water bath. The homogenates were subsequently centrifuged at 4,000 rpm for 10 min at 4 °C, and the resulting supernatants were retained for analysis. The activities of superoxide dismutase (SOD) and catalase (CAT) were determined following the manufacturer’s protocols for commercial assay kits (Nanjing Jiancheng Bioengineering Institute, catalog numbers A007-1-1 and A001-3-1, respectively). 2 g samples from the foot, gill, mantle, and muscle tissues were collected, rinsed with pre-chilled phosphate-buffered saline (PBS), mixed with the provided working solutions for 2 min using a vortex mixer, and incubated at 37 °C for 40 min to assess enzyme activity. SOD activity was quantified by measuring absorbance at 550 nm, with one unit of SOD activity (U/g protein) defined as the amount required to achieve a 50% inhibition rate of SOD. For CAT activity, the reaction involving the hydrolysis of H₂O₂ by CAT was terminated by the addition of ammonium molybdate, forming a yellow molybdate-H₂O₂ complex. CAT activity was then assessed by monitoring the change in absorbance at 405 nm. Three independent tissue samples from three individual mussels per treatment were used as biological replicates. ROS and MDA content assay The ROS content was determined using 2′,7′-dichlorofluorescein diacetate (DCFH-DA) in accordance with the manufacturer’s protocol for a commercial assay kit (Nanjing Jiancheng Bioengineering Institute, catalog number E004-1-1). 2 g samples from the foot, gill, mantle, and muscle tissues were collected, rinsed with pre-chilled phosphate-buffered saline (PBS), and treated with an appropriate volume (3 mL of 10 µM DCFH-DA for muscle and mantle, 2 mL of 10 µM DCFH-DA for foot and gill) of enzyme digestion solution. The samples were incubated in a water bath at 37 °C for 30 min to facilitate digestion. Following filtration, the resulting cell suspensions were analyzed using fluorescence spectroscopy at an excitation wavelength of 488 nm and an emission wavelength of 525 nm. The MDA content was assessed using the thiobarbituric acid reactive substances (TBARS) method, following the manufacturer’s instructions for a commercial assay kit (Nanjing Jiancheng Bioengineering Institute, catalog number A003-1). 2 g samples from the foot, gill, mantle, and muscle tissues were homogenized in a pre-chilled reagent. The homogenates were incubated in a water bath at 95 °C for 40 min. Subsequently, 0.1 mL of chloroform was added, and the mixture was vortexed for approximately 60 s to facilitate extraction. The samples were then centrifuged at 4,000 rpm for 10 min. After centrifugation, 0.2 mL of the supernatant was collected from each sample, and the absorbance was measured at 532 nm using a spectrophotometer. Three independent tissue samples from three individual mussels per treatment were used as biological replicates. Transmission electron microscopy analysis Transmission electron microscopy (TEM) was performed to examine ultrastructural changes in cellular organelles, particularly to assess the involvement of autophagy in the thermal stress response of M. coruscus tissues. Tissues (foot, gill, mantle, and muscle) from mussels in the 21 °C (control) and 31 °C treatment group after 12 h of exposure were used for TEM analysis Immediately after collection, the samples were fixed in 3% glutaraldehyde prepared in phosphate-buffered saline (PBS) and subjected to vacuum treatment overnight to ensure thorough penetration of the fixative. Subsequently, the samples were rinsed extensively with PBS to remove residual glutaraldehyde and then post-fixed with 1% (w/v) osmium tetroxide at 4 °C for 12 h. Following postfixation, the samples underwent dehydration through a graded series of ethanol and acetone solutions. The dehydrated samples were then infiltrated and embedded in LR White resin. Transverse ultrathin sections were prepared using a Leica Ultracut R microtome, stained with uranyl acetate and lead citrate to enhance contrast, and examined using a JEOL 1230 transmission electron microscope. Three independent tissue samples from three individual mussels per treatment were used as biological replicates. RNA extraction and RT–qPCR analysis Total RNA was isolated from four tissues (foot, gill, mantle, and muscle) of mussels in each treatment group using TRIzol reagent (Thermo Fisher Scientific), and cDNA was synthesized using the Evo M-MLV Reverse Transcriptase Kit (Accurate Biology). Relative expression analysis was performed on these tissues. Quantitative real-time polymerase chain reaction (RT-qPCR) was conducted on a QuantStudio™ 3 Real-Time PCR System (Applied Biosystems) with SYBR Green Pro Taq HS (Accurate Biology) as the fluorescent dye. Relative expression levels of autophagy-related genes (ATGs) and heat shock protein 70 genes (HSP70 quantified using the 2−ΔΔCt method [58], as described by Livak and Schmittgen. To ensure robust normalization, we evaluated the expression stability of eight candidate reference genes (β-actin, UBFD1, EEF1A, TUBA, YBX1, RP-S15Ae, EEF1D, TEFA) under heat treatment using NormFinder and geNorm software tools [59, 60]. Based on these analyses, we selected CAC5418781.1 (encoding β-actin) as the most stable reference gene for normalizing gene expression in Mytilus coruscus (Fig. S1, Table S1). The specific genes and primers analyzed are listed in Table S1. Experiments were performed with three independent biological replicates to ensure reproducibility. Primer sequences utilized for the RT-qPCR assays are listed in Table S2. Three independent tissue samples from three individual mussels per treatment were used as biological replicates. RNA sequencing and transcriptome data analysis During sample collection, three mussels were randomly selected from each treatment group, with four tissues (foot, gill, mantle, and muscle) dissected from each individual. The samples utilized for RNA sequencing and transcriptome data analysis were derived from four tissues obtained from the 21°C and 31°C experimental groups, total 24 samples (3 mussels × 4 tissues × 2 treatment groups). Three independent tissue samples from three individual mussels per treatment were used as biological replicates.Complementary DNA (cDNA) was synthesized from the isolated RNA and used to construct adaptor-ligated libraries, which were sequenced on an Illumina HiSeq™ platform by Sangon Biotech (Shanghai) Co., Ltd. Raw sequencing reads were generated, and quality control was performed by trimming adapters and filtering out low-quality sequences using fastx_toolkit_0.0.14, SeqPrep, and Sickle. Low quality reads were removed by four standard criteria: (1) Trim low-quality bases (Q < 20) from the 3’ to 5’ direction of reads; (2) Trim low-quality bases (Q < 20) from the 5’ to 3’ direction of reads; (3) Use a sliding window approach to remove bases with quality scores below 20 at the read ends (window size: 5 bp). (4) Discard reads shorter than 35 nt, along with their paired reads. The resulting clean reads were aligned to the Mytilus coruscus reference genome (GenBank: GCA_011752425.2, available at https://www.ncbi.nlm.nih.gov/datasets/genome/GCA_011752425.2/) using HISAT2(version 2.1.0). Mapped reads were assembled into transcripts using Cufflinks. Gene expression levels were quantified as transcripts per million (TPM) using RSEM (version 1.2.31). Differentially expressed genes (DEGs) were identified with DESeq2 (version 1.22.1), applying a false discovery rate (FDR) threshold of < 0.01 and an absolute log2 fold change (|log2FC|) ≥ 1. Principal component analysis (PCA) of the DEGs was conducted in R (version 4.0.2), and Venn diagram analysis was performed online using the Venny tool (http://bioinfogp.cnb.csic.es/tools/venny/index.html). The k-means clustering algorithm was implemented in R to generate heatmaps and cluster DEGs based on their log2FC expression profiles, with correlation as the similarity metric and the number of clusters (k) set to 6. This analysis categorized the DEGs into six distinct clusters (C1–C6) according to their expression patterns. Gene Ontology Biological Process (GOBP) enrichment analysis of selected DEG clusters was performed using TBtools-II (version v2.121). Yeast transformation Full-length cDNAs of ATGs and HSP70s were amplified by PCR and subcloned into the pYES2 vector. The recombinant plasmids were transformed into Saccharomyces cerevisiae INVSC1 using the lithium acetate method, and transformants were selected on SG-URA medium (Coolaber, catalog number PM22725). For stress assays, yeast cells were cultured either on SG-URA medium supplemented with H₂O₂ (5 µM) at 28 °C for 3–4 days, or on SG-URA medium without H₂O₂ at 39 °C for 3–4 days. Yeast two-hybrid (Y2H) assay To construct the pGBKT7-BD plasmids, full-length cDNAs of autophagy-related genes were amplified via PCR and subcloned into the pGBKT7-BD vector (Clontech). Similarly, pGADT7-AD plasmids were generated by amplifying full-length cDNAs of the HSP70 gene and inserting them into the pGADT7 vector (Clontech). The yeast strain Saccharomyces cerevisiae Y2HGold (Clontech) was co-transformed with the recombinant pGBKT7-BD and pGADT7-AD plasmids using a standard lithium acetate method. Transformed yeast cells were plated on synthetic dropout (SD) media under two selective conditions: (i) SD medium lacking leucine and tryptophan (SD/-Leu/-Trp) to confirm successful co-transformation, and (ii) SD medium lacking tryptophan, leucine, histidine, and adenine, supplemented with 4 mg·ml⁻¹ X-α-Gal (SD/-Trp/-Leu/-His/-Ade + X-α-Gal) to assess protein-protein interactions. Yeast growth was monitored over 3–5 days at 30 °C, with blue coloration on X-α-Gal-containing plates indicating positive interactions. Three independent sample sets were used as biological replicates. Protein sequence and phylogenetic analysis Autophagy-related proteins were identified using BLASTP (version 2.5.0+) with default parameters [61]. The ATG protein sequence of Crassostrea gigas and query sequences from four model organisms—Caenorhabditis elegans, Dictyostelium discoideum, Saccharomyces cerevisiae, and Arabidopsis thaliana—were used [62, 63]. For phylogenetic analysis, 26 species were selected, comprising 24 mollusks and four outgroup species corresponding to the aforementioned model organisms. For HSP70 homologs in Mytilus coruscus, hmmscan from HMMER (version 3.1b2) was employed with default settings to search sequence databases [64]. To exclude false positives, sequences lacking conserved domains, as determined by NCBI CD-Search, were discarded. Redundancy was minimized by segregating FASTA files containing conserved domains by species ID and processing them with CD-Hit (version 4.8.1) using default parameters. Phylogenetic reconstruction was conducted using OrthoFinder (version 2.5.4) with default settings to identify single-copy orthologous genes [65]. Multiple sequence alignments were generated with MAFFT (version 7.505) under default parameters [66], followed by trimming of poorly conserved regions using trimAl (version 1.4.rev15) with default settings [67]. The phylogenetic tree was inferred using IQ-TREE (version 2.1.4-beta) with 1,000 rapid bootstrap replicates [68]. The best-fit model of protein evolution, LG + R10, was selected using ModelFinder, integrated within IQ-TREE [69]. Statistical analysis The statistical significance of differences was analyzed by the two-tailed Student’s t test or one-way analysis of variance (ANOVA). Significant differences were set at P < 0.05 and P < 0.01. Results Temperature-dependent thermal stress response in Mytilus coruscus To explore the thermal stress response in M. coruscus, we measured oxidative stress markers and analyzed the expression of thermal stress–related genes. Thermal stress disrupted oxidative homeostasis, with ROS levels increasing on average by ~ 3.7-fold at 31 °C and ~ 4.6-fold at 36 °C across four tissues (foot, gill, mantle, and muscle), most prominently in muscle and mantle (Fig. 1A). MDA levels also rose significantly from 31 °C onward, peaking at ~ 3.5-fold in mantle at 31 °C and ~ 3.1-fold in muscle at 36 °C (Fig. 1B). Antioxidant defenses were activated, with SOD activity peaking at 31 °C before declining at 36 °C; the highest induction was observed in muscle (~ 2.9-fold) and the lowest in gill (~ 2.3-fold). Similarly, CAT activity peaked at 31 °C and decreased at 36 °C, with the strongest induction in mantle (~ 6-fold) and the weakest in muscle (~ 3.2-fold) (Fig. 1C, D). Furthermore, we explored the heat response progression at the transcriptional level by examining the expression of thermal stress marker genes including HSF1, HSP20, HSP70, and HSP90A by RT-qPCR (Fig. 1E-H). Gene expression analysis revealed temperature-dependent upregulation of thermal stress markers. Expression was induced at 26 °C, reached near-maximal levels at 31 °C, and declined at 36 °C in most tissues. HSF1 and HSP20 were highly induced in all tissues, although HSP20 showed relatively lower induction in foot (Fig. 1E, F). HSP70 and HSP90A were strongly upregulated in mantle but showed limited induction in muscle (Fig. 1G, H). Together, these findings indicate that M. coruscus mounts a heterogeneous thermal stress response, with gills and mantle experiencing the strongest oxidative stress and antioxidant activation, which may in turn trigger autophagy to mitigate cellular damage. Fig. 1Dynamic analysis of physiological and molecular changes in Mytilus coruscus under thermal stress. A-D Physiological changes including ROS production (A), MDA content (B), SOD activity (C), and CAT activity (D) in the muscle, gill, foot, and mantle tissues of M. coruscus after thermal stress. E-H Relative expression levels of thermal-response-related genes (HSF1, HSP20, HSP70, and HSP90) in the same tissues post-stress. 21 °C served as the control group, while 26 °C, 31 °C, and 36 °C were designated as experimental treatment groups. Different letters above the bar charts indicate differences at the 0.05 significance level determined by one-way ANOVA; mean ± SE, n = 3 Dynamic analysis of physiological and molecular changes in Mytilus coruscus under thermal stress. A-D Physiological changes including ROS production (A), MDA content (B), SOD activity (C), and CAT activity (D) in the muscle, gill, foot, and mantle tissues of M. coruscus after thermal stress. E-H Relative expression levels of thermal-response-related genes (HSF1, HSP20, HSP70, and HSP90) in the same tissues post-stress. 21 °C served as the control group, while 26 °C, 31 °C, and 36 °C were designated as experimental treatment groups. Different letters above the bar charts indicate differences at the 0.05 significance level determined by one-way ANOVA; mean ± SE, n = 3 Transcriptomic analysis of thermal stress responses in Mytilus coruscus Given that physiological and molecular responses in M. coruscus peaked at 31 °C in prior analyses, we selected this temperature to investigate the molecular mechanisms underlying its thermal stress response. Transcriptomic profiling was performed on muscle, gill, foot, and mantle tissues after a 12-hour exposure to 31 °C, with 21 °C as the control. Principal component analysis (PCA) revealed both conserved and divergent transcriptional responses across these organs (Fig. 2A). In total, 9,089 differentially expressed genes (DEGs) were identified, including 1,495 in muscle, 2,038 in gill, 3,903 in foot, and 2,911 in mantle (Fig. 2B). The gill exhibited the largest number of DEGs, while the muscle showed the fewest. Notably, most DEGs were downregulated under thermal stress, indicating organ-specific transcriptional plasticity with an overall trend toward suppression. Fig. 2Transcriptomic analysis of thermal response in Mytilus coruscus. A Principal component analysis (PCA) was performed to assess the transcriptomic dimensionality of heat response in the muscle, foot, gill, and mantle tissues of Mytilus coruscus. B Numbers of upregulated (Up) and downregulated (Down) genes in each tissue compared to the control group. C Hierarchical clustering of differentially expressed genes (DEGs) in tissues exposed to 31 °C thermal stress. DEGs were grouped into six clusters (C1–C6) using a soft clustering method based on expression profiles. D GOBP enrichment analysis of DEGs in clusters 1, 3, and 5. E GOBP enrichment analysis of DEGs in clusters 2, 4, and 6 Transcriptomic analysis of thermal response in Mytilus coruscus. A Principal component analysis (PCA) was performed to assess the transcriptomic dimensionality of heat response in the muscle, foot, gill, and mantle tissues of Mytilus coruscus. B Numbers of upregulated (Up) and downregulated (Down) genes in each tissue compared to the control group. C Hierarchical clustering of differentially expressed genes (DEGs) in tissues exposed to 31 °C thermal stress. DEGs were grouped into six clusters (C1–C6) using a soft clustering method based on expression profiles. D GOBP enrichment analysis of DEGs in clusters 1, 3, and 5. E GOBP enrichment analysis of DEGs in clusters 2, 4, and 6 To characterize expression patterns, we applied Fuzzy C-Means Clustering method, which grouped DEGs into six clusters (C1–C6) (Fig. 2C). Cluster 1, 3, and 5 showed predominantly upregulated in gills, mantle, and muscle, respectively, while slightly downregulated in other organs (Fig. 2C). Gene Ontology Biological Process (GOBP) analysis of cluster 1 DEGs indicated enrichment in autophagy, TOR signaling, programmed cell death (PCD) regulation, hydrolase activity regulation, and immune response activation (Fig. 2D). Cluster 3 DEGs were associated with heat shock response, catalytic activity regulation, PCD regulation, Ras signaling, and hormone secretion (Fig. 2D). Cluster 5 DEGs were linked to negative regulation of ROS, cellular response to starvation, hexose response, organic acid catabolism, and cytokine response regulation (Fig. 2D). In contrast, clusters 2, 4, and 6 exhibited pronounced downregulation in gill but upregulation in the mantle (Fig. 2C). Cluster 2 DEGs were upregulated in muscle and mantle but downregulated in foot and gills, enriched in mitochondrial autophagy regulation, programmed cell death suppression, mitochondrial protein catabolism, hydrolase activity regulation, and DNA damage checkpoint signaling (Fig. 2E). Cluster 4 DEGs were downregulated in foot and gills but upregulated in muscle and mantle, associated with DNA repair, telomere maintenance, and endosomal lumen acidification (Fig. 2E). Cluster 6 DEGs were downregulated in gills but upregulated in muscle, foot, and mantle, linked to hyperoxia response, cytokine signaling, lipoxygenase pathway, and protein kinase A activation (Fig. 2E). Tissue-specific gene expression was further analyzed to reveal the varying responses to thermal stress across different tissues (Fig. S2A, B). In muscle, uniquely upregulated genes were enriched in protein polyglutamylation, ERBB signaling, and larval midgut cell death, while downregulated genes were associated with starvation response, organic anion transport, and carbohydrate derivative catabolism (Fig. S2C). In foot and gill, upregulated DEGs were linked to stress-related processes, including catalytic activity regulation, synaptic vesicle recycling, abiotic stimulus response, endoplasmic reticulum stress response, and mitochondrial autophagy, whereas downregulated genes were involved in growth-related functions such as cell motility, peptide hormone processing, muscle system processes, TOR signaling, cytokine response, and GTPase activity regulation (Fig. S2D, E). In mantle, uniquely upregulated genes were enriched in aminoglycan catabolism, insulin secretion regulation, and superoxide metabolism, while downregulated genes were associated with growth maintenance processes, including DNA repair, Golgi vesicle transport, and peptide biosynthesis (Fig. S2F). Together, these results demonstrate M. coruscus exhibits both conserved and transcriptional tissue-specific responses to thermal stress. Autophagy-mediated thermal stress response in Mytilus coruscus Transcriptomic and GOBP analyses revealed differential expression of ATGs across muscle, mantle, foot, and gill tissues under thermal stress at 31 °C, implicating a key role of autophagy in thermal stress regulation (Fig. 2D). In M. coruscus, we identified 26 autophagy-related (ATG) proteins, compared to 34 in C. gigas, 22 in Pecten maximus, 20 in Saccharomyces cerevisiae, and 36 in Homo sapiens [62, 63] (Fig. 3A). These ATG proteins span five functional complexes involved in autophagy: initiation, nucleation, elongation, completion, and autophagosome-lysosome fusion (Fig. 3B). Of the 26 ATGs, 15 were upregulated, with predominant expression in gill and foot tissues and minimal expression in muscle (Fig. 3B, Table S1). qPCR analysis confirmed that the expression of these 15 ATGs generally increased with temperature (21 °C, 26 °C, 31 °C, 36 °C), with organ- and gene-specific variation (Fig. 7A). Based on expression profiles, eight hub genes—ATG3, ATG4, ATG8c, ATG10, ATG13, ATG38, ATG101, and NBR1—were selected for further analysis. These genes showed significant upregulation across most organs, with very high expression levels in foot and gill (Fig. 7A). ATG13 and ATG101, part of the ATG1/ULK complex, were upregulated in all tissues, except that ATG13 was not upregulated in muscle and ATG101 was not upregulated in gills. Within the ATG8 conjugation system, ATG3 and ATG4 were upregulated in foot and mantle, while ATG8c was upregulated in muscle, foot, and mantle. ATG10 (ATG12-ATG5 conjugation system) was upregulated in all organs except gills and ATG38 ( PI3K complex I) in all except foot. NBR1, a selective autophagy receptor, was upregulated only in muscle. Transmission electron microscopy (TEM) analysis revealed putative autophagic vesicles near the nucleus in gill and foot cells, which is consistent with the high expression of ATG genes in these tissues and indicates enhanced autophagy activity during heat stress (Fig. 7B). These findings suggest that upregulated ATGs contribute to the thermal stress response via autophagy in M. coruscus. Overexpressed these eight ATGs in S. cerevisiae revealed that, compared with controls, ATG3, ATG4, ATG13, and ATG101 significantly enhanced yeast survival under both oxidative and thermal stress, whereas ATG8c and ATG10 improved survival exclusively under thermal stress (Fig. 7C-H). NBR1 increased sensitivity to heat but reduced sensitivity to H₂O₂, and ATG38 reduced sensitivity to H₂O₂ alone (Fig. 7E, H). These results indicate that ATGs differentially modulate stress tolerance, supporting their potential role in heat resistance in M. coruscus. Fig. 3Comparative analysis of autophagy pathway components in Mytilus coruscus. A Distribution of autophagy pathway components across 26 eukaryotic species. B Transcriptional analysis of the gene homologs of autophagy pathway components in My. coruscus under thermal stress. Autophagosome formation requires 26 core ATG proteins divided into five functional groups: (1) Members of the ATG1/ULK complex, (2) ATG9 (a transmembrane vesicle protein), (3) Members of the PI3K complex I, (4) Members of the ATG2–PROPPINs complex, (5) Members of the ATG12–ATG5 and ATG8 conjugation systems. NBR1, an archetypal selective autophagy receptor. Expression data for autophagy pathway homologs were derived from transcriptomic analyses Comparative analysis of autophagy pathway components in Mytilus coruscus. A Distribution of autophagy pathway components across 26 eukaryotic species. B Transcriptional analysis of the gene homologs of autophagy pathway components in My. coruscus under thermal stress. Autophagosome formation requires 26 core ATG proteins divided into five functional groups: (1) Members of the ATG1/ULK complex, (2) ATG9 (a transmembrane vesicle protein), (3) Members of the PI3K complex I, (4) Members of the ATG2–PROPPINs complex, (5) Members of the ATG12–ATG5 and ATG8 conjugation systems. NBR1, an archetypal selective autophagy receptor. Expression data for autophagy pathway homologs were derived from transcriptomic analyses Fig. 4Functional analysis of candidate ATG genes in Mytilus coruscus. A qPCR analysis of ATG genes in muscle, foot, gill and mantle under thermal stress (control: 21 °C; treatments: 26 °C, 31 °C, 36 °C). B Ultrastructure analysis of muscle, foot, gill and mantle under thermal stress (control: 21 °C; treatment: 31 °C). C-E Validation of ATG genes via yeast overexpression (control: 30 °C; heat: 39 °C; H₂O₂: 5 µM). F-H Survival count of yeast clones at OD₆₀₀ = 0.1, 0.01, 0.001, 0.0001, with pYES2 as negative control; heat plates at 39 °C for 4 days, H₂O₂ plates for 3 days. Red arrows indicate autophagic vesicles. * indicates significance level determined by two-tailed Student’s t test. Significance vs. control: *P < 0.05, **P < 0.01; mean ± SE, n = 3 Functional analysis of candidate ATG genes in Mytilus coruscus. A qPCR analysis of ATG genes in muscle, foot, gill and mantle under thermal stress (control: 21 °C; treatments: 26 °C, 31 °C, 36 °C). B Ultrastructure analysis of muscle, foot, gill and mantle under thermal stress (control: 21 °C; treatment: 31 °C). C-E Validation of ATG genes via yeast overexpression (control: 30 °C; heat: 39 °C; H₂O₂: 5 µM). F-H Survival count of yeast clones at OD₆₀₀ = 0.1, 0.01, 0.001, 0.0001, with pYES2 as negative control; heat plates at 39 °C for 4 days, H₂O₂ plates for 3 days. Red arrows indicate autophagic vesicles. * indicates significance level determined by two-tailed Student’s t test. Significance vs. control: *P < 0.05, **P < 0.01; mean ± SE, n = 3 HSP70 regulation of thermal stress response in Mytilus coruscus Transcriptomic profiling of M. coruscus under 31 °C revealed heat response as a dominant process (Fig. 2E). GOBP enrichment highlighted the significant involvement of heat shock proteins (HSPs), particularly the HSP70 family, in thermal stress adaptation. Similar to autophagy-related proteins, HSP70 proteins mitigate various stresses, with a notable role in thermal stress. A total of 173 HSP70 proteins were identified and classified into six subfamilies (Classes I–VI) (Fig. S3). Among them, 9 HSP70 genes were high upregulated, predominantly from Class I, suggesting this subfamily’s key role in thermal stress response (Fig. 5A). Based on the phylogenetic clustering of HSP70 proteins, we selected five upregulated genes (HSP70-1, HSP70-2, HSP70-3, HSP70-4, and HSP70-5) for further study. qPCR analysis of these five HSP70 genes across muscle, gill, foot, and mantle confirmed temperature-dependent induction of these genes (Supplementary Table S1). Expression increased slightly at 26 °C and was significantly elevated at 31 °C and 36 °C, further indicating 31 °C as a critical threshold for thermal response in M. coruscus (Fig. 5B). Functional validation in Saccharomyces cerevisiae demonstrated that overexpression of HSP70-1, HSP70-3, HSP70-4, and HSP70-5 significantly enhanced yeast survival under heat stress (39 °C), whereas HSP70-2 showed no effect (Fig. 5D, G). Under oxidative stress (5 mM H₂O₂), overexpression of HSP70-2, HSP70-3, and HSP70-5 enhanced survival, while HSP70-1 and HSP70-4 did not (Fig. 5E, H). Collectively, These results suggest that specific HSP70 genes confer differential protection against thermal and oxidative stress, supporting their role in thermal stress resistance in M. coruscus. Fig. 5Functional analysis of Hsp70 family proteins in Mytilus coruscus. A Hierarchical clustering of HSP70 genes expression from transcriptomic data. B qPCR validation of HSP70 genes in muscle, foot, gill, mantle under thermal stress (control: 21 °C; treatments: 26 °C, 31 °C, 36 °C). C-E Validation of HSP70 genes via yeast overexpression (control: 30 °C; thermal stress: 39 °C; H₂O₂ stress: 5 µM). F-H Survival count of yeast clones at OD₆₀₀ = 0.1, 0.01, 0.001, 0.0001, with pYES2 as negative control; thermal plates at 39 °C for 4 days, H₂O₂ plates for 3 days. Different letters above the bar charts indicate differences at the 0.05 significance level determined by one-way ANOVA. * indicates significance level determined by two-tailed Student’s t test. Significance vs. control: *P < 0.05, **P < 0.01; mean ± SE, n = 3 Functional analysis of Hsp70 family proteins in Mytilus coruscus. A Hierarchical clustering of HSP70 genes expression from transcriptomic data. B qPCR validation of HSP70 genes in muscle, foot, gill, mantle under thermal stress (control: 21 °C; treatments: 26 °C, 31 °C, 36 °C). C-E Validation of HSP70 genes via yeast overexpression (control: 30 °C; thermal stress: 39 °C; H₂O₂ stress: 5 µM). F-H Survival count of yeast clones at OD₆₀₀ = 0.1, 0.01, 0.001, 0.0001, with pYES2 as negative control; thermal plates at 39 °C for 4 days, H₂O₂ plates for 3 days. Different letters above the bar charts indicate differences at the 0.05 significance level determined by one-way ANOVA. * indicates significance level determined by two-tailed Student’s t test. Significance vs. control: *P < 0.05, **P < 0.01; mean ± SE, n = 3 Interaction of autophagy and HSP70 in regulating thermal stress in Mytilus coruscus Transcriptomic analysis of M. coruscus under 31 °C revealed strong upregulation of numerous ATG and HSP70 genes. Functional validation in Saccharomyces cerevisiae further confirmed 8 ATGs and 4 HSP70 candidate genes as contributors to thermal stress resistance. Based on these results, we hypothesized that ATGs and HSP70 proteins may interact to regulate thermal stress in M. coruscus. To test this, we performed protein interaction predictions using the AlphaFold Server. Interaction confidence was evaluated using the predicted template modeling (pTM) and interface pTM (ipTM) scores, with a combined value > 0.75 indicating high interaction probability. This analysis identified 11 pairs of ATG and HSP70 proteins exceeding this threshold (Fig. 6A). Notably, ATG8c showed the strongest interaction potential, predicted to bind with five HSP70 proteins (Fig. 6B). For experimental validation, ATG8c was tested against the five HSP70s using the yeast two-hybrid system. Consistent with predictions, ATG8c interacted with all except HSP70-2 (Fig. 6C). These findings suggest that ATGs, particularly ATG8c, might regulate thermal stress through interactions with HSP70 proteins in M. coruscus, highlighting a cooperative mechanism underlying thermal stress response. Fig. 6Interaction analysis of autophagy proteins and heat shock proteins. A Protein interaction prediction using the AlphaFold Server, showing the total predicted template modeling (pTM) score and interface pTM (ipTM) score for autophagy-heat shock protein complexes. B Predicted 3D structural models of autophagy protein-heat shock protein interactions generated by the AlphaFold Server, highlighting key binding interfaces and conformational features. C Y2H analysis revealed that ATG8c interacts with HSP70-1, HSP70-2, HSP70-3, HSP70-4, and HSP70-5. SD− LT indicates a synthetic dropout medium lacking Leu and Trp; SD− LTHA indicates a synthetic dropout medium lacking Leu, Trp, His, and Ade with 4 mg mL− 1 X-α-Gal; AD indicates the activation domain of GAL4; BD indicates the DNA-binding domain of GAL4 Interaction analysis of autophagy proteins and heat shock proteins. A Protein interaction prediction using the AlphaFold Server, showing the total predicted template modeling (pTM) score and interface pTM (ipTM) score for autophagy-heat shock protein complexes. B Predicted 3D structural models of autophagy protein-heat shock protein interactions generated by the AlphaFold Server, highlighting key binding interfaces and conformational features. C Y2H analysis revealed that ATG8c interacts with HSP70-1, HSP70-2, HSP70-3, HSP70-4, and HSP70-5. SD− LT indicates a synthetic dropout medium lacking Leu and Trp; SD− LTHA indicates a synthetic dropout medium lacking Leu, Trp, His, and Ade with 4 mg mL− 1 X-α-Gal; AD indicates the activation domain of GAL4; BD indicates the DNA-binding domain of GAL4 Discussions Oxidative stress response to thermal stress in Mytilus coruscus Rising ocean temperatures caused by global warming threaten marine organisms, particularly intertidal species, resulting in ecological disruptions and economic losses [70]. Reactive oxygen species (ROS) production are inevitable byproducts of aerobic metabolism; however, excessive ROS can disrupt cellular homeostasis by inducing lipid peroxidation and enzyme inhibition [71]. In bivalves, elevated temperatures induce oxidative stress and apoptosis via increased ROS, impairing antioxidant defenses [72–74]. In the present study, ROS levels increased significantly across all tissues (muscle, gill, foot, mantle) of M. coruscus following heat exposure at 31 °C (Fig. 1A), confirming thermal stress as a trigger for ROS accumulation. Concurrently, malondialdehyde (MDA), a marker of lipid peroxidation [75], was elevated under thermal stress (Fig. 1B), indicating enhanced membrane damage in response to heat exposure. To counteract these oxidative effects, antioxidant enzymes such as superoxide dismutase (SOD) and catalase (CAT) are mobilized to neutralize excessive ROS [76]. Consistent with this, our results revealed significant increases in SOD and CAT activities post-thermal stress (Fig. 1C, D), potentially reflecting enhanced posttranslational modifications or enzyme synthesis in response to elevated ROS [77]. Taken together, these findings indicate that rising temperature disrupt ROS and MDA homeostasis in M. coruscus, prompting increased SOD and CAT activities to mitigate oxidative stress. Transcriptional responses to thermal stress in Mytilus coruscus Marine organisms have evolved conserved mechanisms to enhance cellular thermal tolerance under thermal stress [78]. M. coruscus, an economically important bivalve extensively cultured in China, notably around Shengsi Island, exhibits sensitivity to elevated temperatures. The increasing frequency of marine heatwaves exacerbates this vulnerability, pushing seawater temperatures toward thermal tolerance limits [18, 79]. The ability of mollusks to cope with such environmental challenges depends largely on transcriptional regulation and physiological plasticity. Thus, characterizing these responses in M. coruscus provides critical insights into molecular mechanisms of heatwave adaptation and offers guidance for aquaculture management. In this study, we analyzed transcriptional and physiological responses to thermal stress in muscle, foot, gill, and mantle tissues continuously exposed to thermal stress (Figs. 1 and 2). Transcriptomic analysis revealed both conserved and divergent responses across these tissues (Fig. 2A). Most DEGs were downregulated, suggesting that thermal stress imposes broadly deleterious effects on cellular functions. These downregulated DEGs were enriched in developmental and growth-related processes, indicating impaired cellular homeostasis and growth under heat stress. Conversely, upregulated DEGs were associated with stress resistance pathways, reflecting adaptive mechanisms that counteract the damaging effects of elevated temperatures (Fig. 2D, E). Collectively, these findings provide a comprehensive overview of organ-specific transcriptional strategies underlying the thermal stress response in M. coruscus. Sessile mussels exhibit greater physiological and genetic plasticity than mobile organisms, compensating for their inability to behaviorally thermoregulate [80]. This plasticity involves cellular, transcriptomic, and metabolic adjustments, driven by gene expression and genetic variation [81, 82]. The genetic basis of heat response in mollusks is complex and polygenic, complicating identification of thermal stress-related genes via traditional transcriptomics, proteomics, phosphoproteomic, or metabolomics [13, 50, 81,]. In bivalves, reverse genetics has identified most of the currently known thermal stress genes [83]. In contrast, only a few forward genetics approaches, such as whole-genome sequencing and genome-wide association studies [82, 84], have been applied, while other methods like QTL mapping and mutant screens remain underutilized [85]. To overcome these limitations, we integrated transcriptomic data with a yeast-based screening system in M. coruscus. Saccharomyces cerevisiae, a well-established eukaryotic model, is widely used for functional gene analysis under stresses like heat, cadmium, and salt. This approach provided an effective means to validate candidate genes and assess their potential roles in thermal tolerance. Autophagy regulation of thermal stress response in Mytilus coruscus Autophagy, a conserved eukaryotic degradation pathway, mitigates cellular stress. In S. cerevisiae, over 30 autophagy-related (ATG) genes have been identified, with their molecular mechanisms well-characterized [86]. In bivalves, autophagy is similarly activated under environmental stress. For instance, in C. gigas, autophagy is activated during Ostreid herpesvirus 1 (OsHV-1) infection at both cellular and molecular levels [30], while in Mytilus edulis, autophagic activity in the digestive gland responds to diverse environmental challenges [87]. Thermal stress also induces autophagy as an adaptive mechanism, enhancing cellular protection in organisms like Caenorhabditis elegans [88], Sus scrofa [89], and Patinopecten yessoensis, where PyBeclin 1 and PyLC3 activate autophagy under acute thermal stress [24]. Despite these insights, little is known about the molecular components and regulatory roles of autophagy in M. coruscus during thermal stress. In this study, we identified 26 ATG genes in the M. coruscus genome, man of which exhibit significant transcriptional changes analysis at 31 °C, indicating their role in thermal stress adaption (Fig. 3B). Notably, several ATGs were upregulated, with transmission electron microscopy confirming autophagic structures in gill and foot tissues, where expression was markedly elevated compared to muscle, which showed minimal variation (Fig. 7A, B). These findings suggest that gill and foot tissues exhibit heightened autophagic sensitivity, representing key sites of thermal stress adaptation. Functional validation in S. cerevisiae showed that most upregulated ATGs enhanced yeast survival at 39 °C, except NBR1 (Fig. 7C-H). NBR1, a selective autophagy receptor, targets peroxisomes for degradation [90] and, in plants, degrades HSP90.1 post-thermopriming to modulate HSFA2 activity [91]. Interestingly, in our study, NBR1 overexpression reduced yeast survival under thermal stress but enhanced resistance oxidative stress (5 mM H₂O₂). This effect may result from excessive autophagy or HSPs degradation, which limitis heat shock factor (HSF) activity and compromise heat tolerance. Taken together, our results demonstrate that autophagy in M. coruscus is activated by thermal stress, with upregulated ATGs coordinating autophagosome formation and maturation to enhance thermal resistance. HSP70 gene family and thermal stress response in Mytilus coruscus Bivalves in marine ecosystems face multiple stressors, including elevated temperatures, low osmotic pressure, and reduced dissolved oxygen [92]. HSPs are critical for maintaining cellular homeostasis under such conditions, facilitating stress protein synthesis, protein folding, stabilization of denatured proteins, and their degradation [93]. Previous studies have documented HSP upregulation in Mytilus edulis and increased HSP production in Mytilus spp. under thermal stress [94, 95]. Among them, the HSP70 family is particularly noted for its role in thermal stress response across mollusks [96]. In this study, 173 candidate HSP70 genes containing the conserved HSP70 domain were identified from the M. coruscus genome, exceeding counts in other bivalves such as Crassostrea gigas (113), Patinopecten yessoensis (61), and Chlamys farreri (65) [49, 96, 97]. Such expansion in HSP70 gene copy number may reflect species-specific regulatory adaptations to environmental fluctuations, particularly temperature [83]. Transcriptomic analysis under thermal stress at 31 °C revealed that most HSP70 genes were downregulated, contrary to expectations. Only nine genes were significantly upregulated, exhibiting high elevated expression across all organs tested (Fig. 5A). Similar patterns have been observed in other bivalves. In Magallana gigas, only 31 of 113 HSP70s were upregulated after 12 h at 28 °C, with most downregulated or not expressed [49]. Likewise, in M. coruscus, only 24 HSP70s were differentially expressed after 3 h at 28 °C [50]. These findings suggest that widespread downregulation of HSP70s under heat stress is common. Likely, only a few HSP70s are induced for stress defense, while others involved in growth and development are repressed, reflecting a trade-off between stress response and normal functions. Additionally, our single-temperature (31 °C) and single-time-point (12 h) design may not fully capture dynamic HSP70s expression. Functional validation in Saccharomyces cerevisiae showed that four of these five upregulated HSP70 genes conferred thermal stress resistance at 39 °C, suggesting a key role in thermal stress response (Fig. 5C-H). Together, these findings provide a comprehensive characterization of the HSP70 gene family in M. coruscus and establish a foundation for further investigation into their functional roles in thermal stress adaptation in bivalves. Interaction between HSPs and ATGs in thermal stress regulation in Mytilus coruscus Both HSPs and ATGs contribute to stress responses, prompting investigation into their potential interactions, particularly under thermal stress [77]. In mammals, Ulk1 interacts with the Hsp90-Cdc37 complex to stabilize and activate Ulk1, facilitating Atg13 phosphorylation and its recruitment to damaged mitochondria [53]. Autophagy-inducing stress elevates acetylated HSP70 levels, which bind the Beclin-1–Vps34 complex to enhance autophagy [98]. Additionally, HSP90 stabilizes Beclin 1 via an evolutionarily conserved domain, supporting TLR-mediated autophagy during microbial infections [53]. While HSP70 and autophagy are known to regulate thermal stress in animals, their interactions in this context remain underexplored. In M. coruscus, we combined AlphaFold-based protein interaction predictions [99] with yeast two-hybrid validation to demonstrate HSP70-autophagy interactions (Fig. 6). These findings provide evidence that HSP70 and ATGs coordinately regulate thermal stress response in M. coruscus, extending prior observations in other systems. Based on our findings, we propose a model for how ATGs and HSP70s regulates thermotolerance (Fig. 7). Under thermal stress, excessive reactive oxygen species (ROS) and 3,4-Methylenedioxyamphetamine (MDA) accumulate in tissues, leading to oxidative stress. This stress activates the expression of autophagy-related genes (ATGs) and heat shock proteins (HSP70s). ATGs and HSP70s may interact to enhance protein homeostasis and cellular survival, thereby improving thermotolerance. Meanwhile, many HSP70s primarily involved in growth and development are transcriptionally repressed under heat stress, reflecting a trade-off between stress defense and normal cellular functions. Fig. 7Proposed model of the regulatory interplay between autophagy-related proteins (ATGs) and heat shock proteins (HSP70s) in the thermal stress response of Mytilus coruscus. Thermal stress induces excessive accumulation of reactive oxygen species (ROS) and 3,4-Methylenedioxyamphetamine (MDA) resulting in oxidative stress. In response, autophagy-related genes (ATGs) and heat shock protein genes (HSP70s) are activated, with potential crosstalk between these pathways contributing to protein quality control and enhanced thermotolerance. Concurrently, a subset of HSP70s primarily involved in growth and developmental processes is transcriptionally repressed, suggesting a trade-off between cellular stress defense and normal physiological functions Proposed model of the regulatory interplay between autophagy-related proteins (ATGs) and heat shock proteins (HSP70s) in the thermal stress response of Mytilus coruscus. Thermal stress induces excessive accumulation of reactive oxygen species (ROS) and 3,4-Methylenedioxyamphetamine (MDA) resulting in oxidative stress. In response, autophagy-related genes (ATGs) and heat shock protein genes (HSP70s) are activated, with potential crosstalk between these pathways contributing to protein quality control and enhanced thermotolerance. Concurrently, a subset of HSP70s primarily involved in growth and developmental processes is transcriptionally repressed, suggesting a trade-off between cellular stress defense and normal physiological functions Conclusion This study provides a comprehensive analysis of autophagy-related genes (ATGs) and heat shock proteins (HSP70s) in the thick-shelled mussel M. coruscus and their roles in thermal stress response. Oxidative stress indicators, enzyme activities, and gene expression analyses revealed tissue-specific patterns under elevated temperatures, suggesting heterogeneous responses across organs. Transcriptomic and functional assays demonstrated that only a subset of HSP70s are strongly induced, while many are downregulated, reflecting a trade-off between stress defense and normal cellular functions. Moreover, yeast functional assays and protein interaction analyses indicated that ATG8c and HSP70s may cooperate to enhance thermotolerance. Together, these findings deepen our understanding of the molecular mechanisms underlying thermal stress adaptation in bivalves, and provide valuable genetic resources for future studies on stress tolerance and aquaculture improvement of M. coruscus. Supplementary Information Supplementary Material 1: Fig. S1. Gene expression stability of the candidate reference genes calculated by NormFinder (A) and geNorm (B). Fig. S2. Transcriptomic analysis of tissue-dependent heat response in M. coruscus. Fig. S3. Phylogenetic analysis of HSP70 proteins in M. coruscus. Table S1. Candidate genes selected for RT-qPCR and gene function analysis.Table S2. Primers used for RT-qPCR analyses. Supplementary Material 1: Fig. S1. Gene expression stability of the candidate reference genes calculated by NormFinder (A) and geNorm (B). Fig. S2. Transcriptomic analysis of tissue-dependent heat response in M. coruscus. Fig. S3. Phylogenetic analysis of HSP70 proteins in M. coruscus. Table S1. Candidate genes selected for RT-qPCR and gene function analysis.Table S2. Primers used for RT-qPCR analyses.