Proteomic signatures of sperm and extracellular vesicles associated with sperm freezability in Holstein bulls

Abstract

Semen cryopreservation and artificial insemination have crucial and beneficial effects on cattle breeding. The freezability of sperm, as primarily reflected by post-thaw sperm motility (PTM), is essential for evaluating semen quality. Some studies have shown notable differences in sperm freezability among various bulls. Here, we compared protein profiles of sperm cells and seminal plasma extracellular vesicles (SPEVs) in the high sperm freezability group and the low sperm freezability group of Holstein bulls to identify the important proteins and their mechanisms that affect sperm freezability. As a result, 432 and 394 differentially expressed proteins (DEPs) were identified in sperm and SPEVs between high and low freezability groups. The results of weighted correlation network analysis (WGCNA) showed that the blue module was significantly (r = 0.89, P = 9 × 10^− 6) associated with sperm freezability. In addition, the pathway analysis revealed that “Metabolic pathways” and “Oxidative phosphorylation” were the predominant biological processes represented. Furthermore, 17 DEPs were found located in the previously identified QTLs related to post-thaw sperm motility, indicating possible variation in their genes. Interestingly, the expression of 142 protein pairs in sperm were significantly (|r| >0.9, P < 0.05) correlated with their expression in SPEVs. Finally, we detected genetic variations in six important candidate genes (STK38, HSPA1A, HSP90B1, LPO, DNASE2 and CUTA), and found that a missense mutation (Chr23g. 23:27522566 A > G) in HSPA1A may affect sperm freezability by decreasing the expression of HSPA1A. Our study highlighted the different protein characteristics of sperm and SPEVs in samples with distinct sperm freezability. These proteins, together with relevant SNPs might be useful markers for selecting bulls with high sperm freezability.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37628-2.

Introduction

Since the first report of semen cryopreservation in the 18th century¹, semen cryopreservation and Artificial Insemination (AI) technologies have been widely used after years of development. Storing semen through cryopreservation and subsequent use through artificial insemination has the benefit of speeding up the dissemination of genetic diversity and facilitating the global distribution of genetically superior animals. For example, sperm from best breeding bulls can be used to fertilize numerous cows globally through the use of frozen semen. Thus, frozen semen and AI are important technologies in cattle breeding, which can accelerate genetic progress and selection. However, the rate of pregnancy in cows fertilized with frozen semen is lower compared to those inseminated with fresh semen². Part of the reason is that freezing sperm may damage sperm cells. Sperm cryopreservation is a continuous process that includes lowering temperature, dehydrating cells, freezing, storing and thawing. Due to physiological damage to sperm during this process, such as the changes in membrane structure and sperm metabolism, the motility and fertility of sperm are significantly reduced^3,4.

Currently, the requirement of PTM for frozen semen in most countries has reached 50%^3,5. Hitit et al.. (2020) carried out a statistical study on the sperm motility of Alta Genetics bulls after thawing of frozen semen from 2008 to 2016, including 100,448 frozen sperm phenotype records from 860 Holstein bulls⁶. The findings indicated that the average thawed sperm motility of bulls with high sperm freezability was 63.83%, while the average post-thaw sperm motility of bulls with low sperm freezability was 52.10%. Previous study has shown that some males with almost identical total sperm motility measured in fresh semen exhibited different sperm motility after cryopreservation⁷. In addition, some bulls with high reproductive performance in natural mating have poor semen freezability⁸. These studies indicate that there are significant differences in the survival ability of sperm under cryopreservation pressure among individuals. Moreover, it has been reported that the heritability of sperm motility after thawing in bulls is approximately 0.21⁹, indicating that genetic factors play a notable role in the freezability of bull sperm.

Semen is composed of sperm cells and seminal plasma, the latter of which contains a significant quantity of extracellular vesicles (EVs). EVs are double-layered phospholipid membrane vesicles with an average size of ~ 100 nm, originating from various types of cells¹⁰. As a transporter of bioactive compounds and crucial intercellular signaling molecules, it contains DNA, RNA, proteins, metabolites and lipids¹¹. Seminal plasma EVs (SPEVs) originate from a mixture of fluids produced by the testis, epididymis, and accessory glands. Many research studies have shown that SPEVs play a significant role in sperm function, particularly in sperm motility¹², sperm capacitation and acrosome reaction¹³. For example, SPEVs can interact with sperm, thereby extending the viable lifespan of sperm and enhancing the structural integrity of the sperm cell’s plasma membrane¹⁴. In our previous study, we analyzed proteins from SPEVs of boars with different sperm motility, and identified 76 differentially expressed proteins between the two groups¹⁵. However, little is known about the changes in protein levels of sperm and SPEVs with distinct freezability.

Several proteomic studies have been carried out to identify important protein markers for predicting sperm freezability in different animal species^7,16,17. Several important proteins related to sperm freezability have been reported, such as HSP90, BSPs and AQPs^18–20. It has been reported that the high expression of heat shock protein 90 (HSP90) in bull sperm was associated with high sperm freezability²¹. Additionally, Ardon observed that the levels of three BSP proteins (BSP1, BSP3, BSP5) were elevated in frozen and thawed bovine sperm samples compared to fresh sperm samples¹. Subsequently, Magalhaes et al. (2016) found that the BSP1 protein was lower expressed in the semen of bulls with good freezing resistance, but higher expressed in the semen of bulls with poor freezing resistance²². Furthermore, it was found that AQP3 and AQP7 were associated with freezability of pig and bovine sperm²⁰.

Moreover, some studies identified genes and SNPs related to sperm freezing resistance based on genetic association analysis. For example, Abril-Parreno et al. (2023) identified several SNPs and important genes (such as FHDC1 and ARFIP1) related to sperm freezing resistance based on genome-wide association analysis²³. Dementieva et al. (2024) identified important candidate genes (such as POU6F2, LPCAT4 and SLC39A12) and specific SNPs associated with various cryopreservation sperm abnormalities in Holstein cattle through genome-wide association analysis²⁴. Wang et al. (2025) identified key genes and variants associated with boar sperm freezability using whole genome resequencing and genetic association analysis²⁵. As a result, nine variants were found to be significantly correlated with sperm freezability.

Although previous studies have identified some proteins associated with sperm freezing resistance using proteomic approaches, these investigations were confined to analyzing differentially expressed proteins and lacked further exploration of functional genetic variations. In addition, there is currently no research that simultaneously investigated the proteomes of sperm and seminal plasma extracellular vesicle (SPEVs) to identify key proteins influencing sperm freezing resistance. The objective of this study was to characterize the proteomic signatures of bull sperm and SPEVs with distinct freezability, and further explore important genetic variations in candidate genes, aiming to identify potential biomarkers for predicting bull sperm freezability.

Materials and methods

Ethics statement

All protocols for the collection of semen samples were reviewed and approved by the Committee for Ethical Review of China Agricultural University. All procedures were carried out in accordance with the relevant guidelines and regulations for Experimental Animals of the Ministry of Science and Technology (Beijing, China), and complied with ARRIVE 2.0 guidelines.

Animals and sample collection

We obtained the semen phenotypes of 145 Holstein bulls from an artificial insemination center in China (Shandong OX Livestock Breeding Co.,Ltd, Shandong, China). All bulls were kept under the identical housing conditions and nutritional management throughout the process of semen collection. In this study, the fresh and post-thaw sperm motilities of these bulls were used to evaluate sperm freezability of each individual. At least ten records of frozen-thawed semen per bull were used for evaluation over a six-month period. Based on the proportion of unqualified frozen semen (defined as post-thaw sperm motility below 40%) in all tested semen samples of each bull, 15 Holstein bulls were selected for subsequent proteomic studies. The 9 individuals with the lowest proportion (< 30%) of unqualified frozen semen were divided into the high freezability (HF) group and the 6 bulls with the highest proportion (> 70%) of unqualified frozen semen were divided into the low freezability (LF) group. To validate the accuracy of grouping, semen samples were collected from each selected bull, and the sperm motility of fresh semen was immediately evaluated using the CASA system (IVOS 12.3, Hamilton Thorne Biosciences, USA). Briefly, after 5 min of incubation at 37◦C, 7 µL of sperm suspension was placed on a prewarmed glass slide. The glass slides were examined under an optical microscope using bright field at a total magnification of 200×. The percentage of motile sperm was estimated in five different regions for each sample. Meanwhile, each sample was used to produce frozen semen to evaluate post-thaw sperm motility. As a result, the average fresh sperm motility in both the HF and LF groups was approximately 70%. However, the average post-thaw sperm motility was higher than 40% in the HF group and less than 27% in the LF group. Finally, one ejaculate sample from each experimental bull was collected for proteomic analysis.

Isolation of sperm and SPEVs

Each fresh semen sample was centrifuged at 800× g and 17 °C for 20 min to separate sperm cells and supernatant. The supernatant was used for the isolation of SPEVs. SPEVs were isolated by ultracentrifugation following established protocols²⁶. Approximately 4–6 milliliters of seminal plasma from each bull were centrifuged at 10,000× g and 4 °C for 30 min to remove cellular debris. The supernatant was then transferred to a clean centrifugal tube and centrifuged at 12,000× g and 4 °C for 1 h (Centrifuge 5810 R, eppendorf, Germany) to further eliminate larger vesicles and contaminants. Subsequently, the supernatant was transferred to an ultracentrifugation tube and centrifuged at 120,000× g and 4 °C for 1.5 h (Optima XPN-100, Beckman, USA) to pellet small extracellular vesicles. The sediments were resuspended in DPBS (Gibco, USA), and the previous step was repeated to wash and purify the vesicles. The sediments were then resuspended in 2 mL of DPBS and filtered through 0.22-µm filters (Millipore, USA). The extracted SPEVs were observed and photographed using transmission electron microscopy (TEM) (HT770, Tokyo, Japan).

Nanoparticle tracking analysis (NTA)

The concentrations of SPEVs were diluted to 1 × 10⁶ ~ 1 × 10⁹ particles/mL with PBS. A ZetaView PMX 110 (Particle Metrix, Meerbusch, Germany) equipped with a 405-nm laser was used to detect the size of the isolated particles. A one-minute video recorded at a frame rate of 30 frames per second was used to analyze the motion of the particles using NTA software (ZetaView 8.02.28).

Western blot analysis

All the SPEV and sperm samples were lysed with RIPA buffer (Solarbio, Beijing, China) containing 1% protease inhibitor on ice for 30 min. Protein concentrations were measured using a BCA assay kit (Beyotime, Beijing, China). Twenty-five micrograms of protein were separated by SDS-PAGE and then transferred to PVDF membranes (Millipore, USA). The membranes were blocked with 5% (w/v) skim milk for 2 h, washed five times with 1× TBST, and then incubated with antibodies against CD9 (sc-13118, Santa Cruz, CA, USA), Tsg101 (sc-13611, Santa Cruz, CA, USA), Alix (sc-53540, Santa Cruz, CA, USA), Calnexin (10427-2-AP, Promega, Madison, WI, USA), HSPA1A (T55496F, Abmart, Shanghai, China) and α-Tubulin (11224-1-AP, Proteintech, IL, USA) for 12 h at 4 °C and then with secondary antibodies (SE131, Solarbio, Beijing, China) for 2 h at 37 °C, and detected with an ECL system.

Protein extraction

Sperm cells and SPEVs samples from 15 bulls were collected and stored at -80℃. After thawing, more than 400µL of lysate (7 M urea, 2 M thiourea, 0.1% PMSF protease inhibitor, 65mM DTT) was added to each sample, and followed by sonication (70 W, 5-s on and 10-s off, 3–5 times). The lysate was centrifuged at 13,572 g for 30 min, and the supernatant was collected as tissue extract. The protein concentration was determined using the Bradford assay (Beyotime, Shanghai, China).

Protein digestion

The protein digestion was carried out using in-gel digestion. The bands were cut into 1-2mm² and put them into EP tubes after SDS-PAGE electrophoresis (20 µg protein per sample) and Coomassie Brilliant Blue staining. Then, 500 µL of neat acetonitrile (ACN) was added, and the solution was incubated in tubes for 20 min. The waste liquid was removed, and destaining was repeated 1–2 times until the blue color faded. Next, 100 µL of 10 mmol/L DTT was added, and reduction was performed at 56 °C for 30 min. Then, 500 µL of ACN was added, the mixture was incubated for 5–10 min, and the liquid was removed. The sample was incubated in the dark for 30 min at room temperature after added 100 µL of the iodoacetamide (55mmol/L) solution. Subsequently, the gel was washed with 500 µL ACN, the liquid was removed, and the samples were freeze-dried for 20 min. Fifty microliters of trypsin solution (0.1 µg/µL) was added to cover the dried gel pieces, and the sample was incubated at 4 ℃ for 30 min. After the liquid was absorbed by the gel, the enzymatic hydrolysis buffer was refilled to completely soak the gel, and kept it at 37 ℃ for 15 h. Subsequently, 100 µL of extraction buffer I (5% TFA) was added to each tube, incubated the solutions for 1 h at 40 °C, and applied sonication at 30 min. Next, the extraction buffer was collected into a new tube, followed by freeze-drying of the solution. Afterward, we added 100 µL of buffer II [50% ACN, 2.5% TFA] to the gel pieces, incubated them for 1 h at 30 °C, and sonicated for 3 min at 30 min. Finally, we combined the extraction buffers, blow-dried the ACN with nitrogen, and freeze-dried the mixture.

LC-MS/MS analysis

LC-MS/MS was performed on biological replicates from 15 Holstein bulls (HF n = 9; LF n = 6). The liquid chromatography tandem mass spectrometry (LC-MS/MS) detection system consisted of a nanoflow high-performance liquid chromatograph (HPLC) instrument (Easy nLC1200 System, Thermo Fisher) coupled to an Orbitrap Lumos mass spectrometer (Thermo Fisher) with a nanoelectrospray ion source (Thermo Fisher). Briefly, 0.5 µg of peptide mixture resolved in buffer A (0.1% formic acid (FA)) were loaded onto a 2-cm self-packed trap column (100-µm inner diameter, ReproSil-Pur C18-AQ, 5 μm; Dr Maisch HPLC GmbH, Ammerbuch, Germany) and separated on a 75-µm-inner-diameter column with a length of 50 cm (ReproSil-Pur C18-AQ, 2 μm; Dr Maisch) over a 78-min gradient (6–12% buffer B for 0–10 min, 12–30% buffer B for 10–95 min, 30–40% buffer B for 95–113 min, 40–95% buffer B for 133–134 min, 95% buffer B for 114–130 min, and then changed to 3% buffer B within two minute and equilibrated for 3 min). Mass spectra were acquired in a data-dependent manner, with an automatic switch between MS and MS/MS. MS spectra were acquired in the Orbitrap analyzer with a mass range of 400–1500 m/z at a resolution of 35,000 in the Orbitrap and a maximum injection time (MIT) of 50 ms. Collision-induced dissociation (CID) peptide fragments were acquired in the ion trap with a collision energy of 30, and 50 milliseconds (ms) activation time.

Protein identification and quantification

MS raw data files were analyzed using MaxQuant software (v. 1.5.2.8)²⁷ using default settings. MS/MS ion searches were performed against the UniProtKB Bos taurus FASTA database (Bos taurus, Proteome ID UP000009136). Cysteine carbamidomethylation was set as the fixed modification. Methionine oxidation and N-terminus acetylation were set as the variable modifications. Trypsin/P enzyme with 2 allowed missed cleavages and minimal peptide length of 7 amino acids were set. The false discovery rates (FDRs) of the peptide-spectrum matches (PSMs) and proteins were set to less than 1%.

Differential expression analysis

After quantile normalization of the proteomics data, the missing value was filled with the minimum value. Differentially expressed proteins (DEPs) analysis was performed using R software and consisted of two parts: (i) when a protein was expressed in both HF and LF groups, and had quantitative information in more than four individuals in the HF group and more than three individuals in the LF group, |log₂FC| >1 and P < 0.05 were set as the thresholds for significantly differentially expressed proteins between the HF and LF groups; (ii) proteins identified only in the HF or the LF group were considered as specifically expressed proteins. Both categories were deemed as differentially expressed proteins (DEPs). Two groups were compared by student’s t-test. The data are presented as the mean ± standard error. P ≤ 0.05 was considered statistically significant.

Weighted Co-expression network analysis

Normalized proteomic expression data from sperm and SPEVs were used to perform a correlation analysis with the psych package (version 2.4.3) in R. Two independent correlation matrices were constructed for the HF and LF groups, and Pearson correlation coefficients were calculated. A weighted correlation network analysis (WGCNA) was conducted using the WGCNA package (version 1.72.1) in R. The protein expression matrix and the sperm freezing resistance phenotypic matrix for all individuals were used as input. In the phenotypic matrix, nine individuals in the HF group were assigned a value of 1, and six individuals in the LF group were assigned a value of 0. The quantile normalized protein expression matrix contained the protein expression values of both sperm and SPEVs. The one-step method was applied to cluster proteins into modules based on pairwise Pearson correlations. The pickSoftThreshold function was used to determine the appropriate soft-thresholding power, followed by the blockwiseModules function to detect co-expressed modules. The labeledHeatmap function was then employed to visualize module–trait relationships. A soft-threshold power of 12 and a minimum module size of 60 were set for module identification. Each module was assigned a distinct color²⁸. Module–trait relationships were assessed as Pearson correlations between module eigengenes and traits. Hub proteins with strong trait correlations (gene significance, GS > 0.2, and module membership, MM > 0.8) within modules significantly associated with sperm freezability were selected for further analysis.

Functional enrichment analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis were performed using DAVID (https://david.ncifcrf.gov/). All genes expressed in bull sperm were subjected to gene set enrichment analysis (GSEA) using the R package clusterProfiler with KEGG pathway datasets. A P value < 0.05 was considered statistically significant.

SNPs detection and linkage disequilibrium analysis

To further explore the genetic variation sites that affect the freezing tolerance of bull sperm, six important candidate genes were selected for gene cloning and Sanger sequencing (ABI 3730XL, USA). We extracted DNA from the cryopreserved semen of bulls in the HF and LF groups. Sperm cells were separated by centrifugation, and then DNA was extraction from the sperm cells. Two mixed DNA pools were constructed by mixing equal amounts of DNA from the bulls in the HF and LF groups, and were used as templates for PCR to detect single-nucleotide polymorphisms (SNPs). 66 pairs of primers (Table S1) targeting the regions of genes were designed using Prime Premier 5.0 software based on the reference sequence (ARS-UCD2.0). These primers were synthesized by GENEWIZ Biotech Co. Ltd. (Suzhou, China). In a separate cohort of 99 Chinese Holstein bulls, a total of 18 SNPs were genotyped for linkage disequilibrium analysis. The linkage disequilibrium analysis of three SNPs within the HSPA1A gene was conducted using Haploview software.

Approximately 20 ng of genomic DNA was used for genotyping by the Agena MassARRAY (Agena Bioscience, San Diego, USA). Detection primers and locus-specific PCR were projected using the MassArray Assay Designer 4.0 software (Agena Bioscience, San Diego, USA). The sample was first amplified by multiple PCR, and the PCR products were then used for locus-specific single-base extension reaction. The resulting products were desalted and transferred to a 384-element Spectro CHIP array. Genotype detection was performed using Matrix-Assisted Laser Desorption/IonizationTime of Flight Mass Spectrometry (MALDI-TOF-MS). The mass spectrograms were analyzed by the MassArray TYPER 4.0 software (Agena Bioscience, San Diego, USA). Then, we carried out the association analysis of eighteen SNPs separately using Chi-squared tests (SPSS 21.0). FDR P value < 0.05 was used as the significance level threshold. Hardy-Weinberg equilibrium (HWE) test was performed on each SNP. Pairwise linkage disequilibrium analysis for all SNPs was performed using Haploview 4.2 software²⁹ (Barrett et al., 2005).

Statistical analysis

Statistical tests were performed using the Statistical Package for the Social Sciences (SPSS) for Windows, release 21.0 (SPSS Inc., Chicago, IL, USA). Two groups were compared by student’s t-test. The data are presented as the mean ± standard error. P ≤ 0.05 was considered statistically significant.

Mediation analysis

To investigate the mediation linkages between proteins in SPEVs, proteins in sperm and the phenotypes of sperm freezability, we selected the DEPs in sperm and SPEVs overlapped with QTLs associated with post-thaw sperm motility, as well as DEPs in significant modules of WGCNA as candidate genes to analyze potential mediation effects. We selected 31 proteins in SPEVs and 32 proteins in sperm as candidate genes. The phenotype of individuals in the LF group is represented by 0, while the phenotype of individuals in the HF group is represented by 1. Mediation analysis was performed on the candidate genes using the SPSSAU (https://spssau.net/?100000001) using a linear regression–based model, with significance assessed by non-parametric bootstrap resampling (5,000 iterations, 95% CI).

Results

Phenotypic analysis and characterization of bull seminal plasma extracellular vesicles

We performed a proteomic study on the sperm cells and SPEVs of 15 Holstein bulls, including 9 bulls in the HF group and 6 bulls in the LF group. Phenotypic analysis showed that there were significant differences in post-thaw sperm motility between the HF and LF groups, while there was no significant difference in fresh sperm motility between the two groups (Figure. 1 A). Transmission electron microscopy (TEM) results showed that most SPEVs had the typical cup shape (Figure. 1B). The mean particle size was 101.9 nm, with the majority of vesicles ranging from 50 to 100 nm (Fig. 1C). The results of western blot showed that extracellular vesicle markers (Tsg101, Alix and CD9) were detected in both groups of SPEVs (Figure. 1D). In contrast, the negative marker of EVs (Calnexin) was not present in SPEVs from the bulls of two groups (Figure. 1D).

Fig. 1.

Phenotypic analysis and characterization of SPEVs isolated from bulls. A: Comparison of sperm motility before and after freezing between the high freezability (HF) group and low freezability (LF) group. Data are presented as mean ± SEM. ***: P < 0.001 (Student’s t-test), n = 9 for the HF group and n = 6 for the LF group. B: TEM image of bull SPEVs. Scale bars: 100 nm. C: NTA results showing that EVs derived from seminal plasma were approximately 50–100 nm in diameter. D: Western blotting results of EV marker Tsg101, Alix and CD9, and the negative marker Calnexin.

Proteome of bovine sperm and SPEVs

The protein profiles of sperm and SPEVs were investigated using the Label-free technique. The results show that the majority of ion deviations (i.e. delta values) are concentrated within the ± 5ppm range. A total of 2552 proteins were identified in sperm and SPEVs, among which 1943 proteins were identified in sperm and 2077 proteins were identified in SPEVs. In addition, 838 proteins were detected simultaneously in four groups (Figure. 2A). The top five proteins with highest expression in sperm were SFP1, TUBB4B, GAPDHS, TUBA3E and ODF1, while the five highest expressed proteins in SPEVs were NT5E, SFP1, DPP4, ENPEP and EHD4.

Gene set enrichment analysis (GSEA) of proteins expressed in SPEVs were significantly enriched in “Ether lipid metabolism”, “Fatty acid metabolism”, “Lysosome”, “Rap1 signaling pathway” and “Regulation of actin cytoskeleton” pathways (Figure. 2B). However, sperm proteins were enriched in “Cell cycle”, “Estrogen signaling pathway”, “Glutathione metabolism” and “MAPK signaling pathway” (Figure. 2C).

Fig. 2.

Proteomic features and differential expression analysis of sperm and SPEVs. A: Venn plot of proteins identified in four groups. B: Gene set enrichment analysis (GSEA) of SPEV proteome. C: Gene set enrichment analysis (GSEA) of sperm proteome. D: Volcano plot displaying significantly differentially expressed proteins (DEPs) between the HF and LF groups. The SPEVs group is on the left, and the sperm group is on the right. E: Heatmaps of DEPs in the HF and LF groups. The SPEVs group is on the left, and the sperm group is on the right. F: Venn plot of DEPs upregulated or downregulated in the SPEVs comparison group and sperm comparison group. Whether in the SPEVs group or the sperm group, the LF group was used as the control group.

Differentially expressed proteins in sperm and SPEVs between the two groups

A total of 123 significant differentially expressed proteins (DEPs) (|log2FC|>1 and P < 0.05) were identified in sperm between the HF and LF groups (Table S2), of which 73 were down-regulated and 50 were up-regulated in the HF group (Figure. 2D). By comparing the SPEVs protein profiles, 106 DEPs were identified between the two groups (Table S3), of which 63 were down-regulated and 43 were up-regulated in the HF group (Figure. 2D). The protein abundance of DEPs in sperm and SPEVs indicated that the clustering of the HF and LF groups was obviously separated (Figure. 2E). The up-regulated or down-regulated top ten proteins in the sperm and SPEVs comparison groups are shown in Table 1. In addition, 188 and 121 specific expressed proteins were only identified in the sperm of HF and LF groups, respectively, as well as 175 and 113 proteins were only identified in the SPEVs of HF and LF groups, respectively. Moreover, 27 DEPs showed consistent directions of gene expression changes in both the sperm and SPEVs comparison groups (such as ITGA6, EEFSEC, C2CD6, FN1, IPO5, OPA1 and GM2A), while 24 DEPs showed opposite expression patterns in the two comparison groups (such as ACTR2, MRAS, AP2S1, PROM1, SLC2A1, DNASE2 and ODF2) (Figure. 2F).

Table 1.

The top 10 deps identified in sperm and SPEVs between the HF and LF groups, respectively.

Protein ID	Gene name	Log2FC Sperm	Log2FC SPEVs	P value	Regulated
U3GXK6	COX3	7.05		3.30E-02	up
A0A6F8Z1 × 1	GLYCAM1	6.29		1.29E-03	up
E1BI28	GM2A		5.46	8.66E-04	up
Q2KJA1	SH3GL1		5.19	2.67E-03	up
E1BNQ4	NAXD		5.07	7.11E-04	up
Q1RMP3	CUTA		4.86	1.71E-02	up
P68509	YWHAH		4.66	1.01E-02	up
F1MG08	SPPL2A		4.05	1.84E-02	up
Q32PA1	CD59	-7.05		1.03E-02	down
Q32PB3	SPACA4	-6.77		1.66E-02	down
F1N0T3	LOC617406	-6.41		4.74E-03	down
Q32L78	TRAPPC6B	-6.40		5.86E-03	down
A1L5A6	ADRM1	-5.71		1.83E-02	down
F1MLK1	RSF1	-5.67		3.07E-03	down
A0A3Q1M5N0	UBE2D3	-5.35		1.60E-02	down
V6F7 × 8	REXO2	-5.00		5.60E-03	down
A0A3Q1MBX8	CCDC50		-5.75	4.01E-02	down
G3MZM7	LOC104993921		-4.37	1.75E-02	down
F1N339	PTPN11		-4.25	1.21E-02	down
A2VDZ0	PPP2R5A		-4.16	3.34E-02	down

To assess the biological significance of these DEPs, bioinformatics analyses were conducted. Enrichment analysis revealed a total of 190 significant GO terms for DEPs in sperm and SPEVs between the two groups. The top ten significant terms for each category were shown in Figure. 3A and Figure. 3B, respectively. In addition, the DEPs identified in sperm were significantly enriched in 25 KEGG pathways (Table S4), and the DEPs identified in SPEVs were significantly enriched in 29 KEGG pathways (Table S5). We found that many metabolic pathways were the predominant biological processes represented, such as carbon metabolism, glutathione metabolism, citrate cycle and biosynthesis of amino acids. Interestingly, four significant pathways (metabolic pathways, lysosome, endocytosis and phagosome) were enriched simultaneously in the sperm and SPEV groups. Meanwhile, 11 DEPs were identified both in sperm and SPEVs (Figure. 3C). Among these 11 DEPs, six proteins (ARPC3, AASS, GALC, ITGA6, ACTB and GM2A) showed consistent directions of gene expression changes in the comparison groups of sperm and SPEVs, while five proteins (AP2S1, ACTR2, HACD3, SEC22B and DANSE2) showed opposite expression patterns in the two comparison groups.

Fig. 3.

Functional enrichment analysis of DEPs. A: The top 10 biological processes (BP), cellular component (CC) and molecular function (MF) enriched in DEPs identified by SPEVs comparison group (P value < 0.05). B: The top 10 biological processes (BP), cellular component (CC) and molecular function (MF) enriched in DEPs identified by sperm comparison group (P value < 0.05).C: The top 10 important pathways enriched in DEPs in SPEVs and sperm comparison groups, respectively. Four pathways shared in the two comparison groups.

Protein Co-expression modules associated with sperm freezability

A weighted correlation network analysis (WGCNA) of all proteins from sperm and SPEVs identified two significant co-expressed modules (Figure. 4 A). We found that the blue(418) module, which included 92 proteins in sperm and 326 proteins in SPEVs, was significantly positively (r = 0.89, P = 9 × 10 − 6) associated with high sperm freezability (Figure. 4 A). On the contrary, the salmon(74) module (including 7 proteins in sperm and 67 proteins in SPEVs) showed a significantly negatively (r= -0.76, P = 0.001) associated with high freezability (Table S6). These data indicated that the proteins in these two modules have different roles with completely opposite effects on sperm freezability. It is worth noting that there were 120 unique DEPs in the blue(418) module, among which 37 were detected in SPEVs and 84 were detected in sperm, and one protein was detected in both SPEVs and sperm. In addition, 32 DEPs were identified in the salmon(74) module, including 8 DEPs in SPEVs and 24 DEPs in sperm (Figure. 4B).

To gain further understanding of the roles played by these proteins, GO and KEGG pathway enrichment analysis were conducted using proteins in the blue(418) and salmon(74) module, respectively. As a result, we found that the proteins in the salmon module enriched in 13 important GO terms (Figure. 4 C, Table S7), such as acrosomal vesicle, sperm flagellum and binding of sperm to zona pellucida. In contrast, proteins in the blue module that showed a positive correlation with sperm freezability were significantly enriched in 132 GO terms (Figure. 4D, Table S8), such as mitochondrion, ATP-dependent microtubule motor activity, mitochondrial inner membrane, sperm midpiece, ATP binding, microtubule-based movement, hydrogen ion transmembrane transport and protein binding. Furthermore, proteins in the blue module were enriched significantly in metabolic pathways, oxidative phosphorylation, carbon metabolism, thermogenesis, proteasome, glycolysis/gluconeogenesis, biosynthesis of amino acids, fatty acid degradation, pyruvate metabolism, cGMP-PKG signaling pathway and fatty acid metabolism were the predominant biological processed represented (Figure. 4E, Table S9). Notably, LPO (GS = 0.97) and VAMP8 (GS = 0.88) in the blue module exhibited the most significant positive correlation with high sperm freezability. Conversely, ODF2 (GS = 0.74) and LYZL6 (GS = 0.70) involved in sperm flagellum showed the most significant negative correlation with high sperm freezability. However, the proteins in the salmon module were not significantly enriched in any pathway.

Fig. 4.

Weighted gene correlation network analysis (WGCNA) of proteins in sperm and SPEVs. A: Two modules (blue(418) and salmon(74) modules) were significantly (P < 0.001) associated with sperm freezability. B: The venn diagram of DEPs identified in sperm and SPEVs, as well as proteins in significant modules. C: GO enrichment results of proteins in the salmon module. D: GO enrichment results of proteins in the blue module. E: KEGG enrichment results of proteins in the two significant modules. Copyright permission for KEGG pathway image obtained from Kanehisa Laboratories^30,31.

Correlation maps between sperm proteins and SPEVs proteins

To investigate the protein interactions between sperm and SPEVs, we conducted a global correlation analysis on the abundance of proteins between sperm and SPEVs of the HF group and LF group, respectively. The correlation coefficient matrices generated by the cross-correlation of all protein expression values were clustered into the global correlation maps of HF and LF samples, respectively (Figure. 5 A and 5B). We observed that the global correlation maps showed distinct patterns between the HF and LF groups. Moreover, there were 142 protein pairs with a correlation greater than 0.9 in sperm and SPEVs. These protein pairs contain a total of 83 SPEVs proteins and 105 sperm proteins, which may play an important role in signal transmission between sperm and SPEVs. 86 pairs of proteins were positively correlated in both the HF and LF groups, and 6 pairs of proteins were negatively correlated in both the HF and LF groups. Interestingly, we found a significant positive correlation between the expression levels of some proteins in sperm and other proteins in SPEVs in the HF group. However, in the LF group, the expression of these identical proteins showed a significant negative correlation between sperm and SPEVs (Figure. 5 C). The opposite situation has also been observed (Figure. 5D). A total of 10 pairs of proteins were negatively correlated in the HF group, while positively correlated in the LF group.

Among 142 protein pairs (|r| > 0.9), we found significant correlations between the expression levels of multiple proteins and the expression level of one protein. For example, a number of genes (RANBP17, DNAJA4, FLAD1, PAMD11, NUTF2M PSMB4 and HSP90AA1) are significantly (P < 0.05) positively correlated with PSMC5 gene in both the HF and LF groups (Figure. 5E). Moreover, this protein highly correlated network contains five important DEPs, including two DEPs (ODF2 and DBI) of SPEVs and 3 DEPs (ART3, DNASE2 and ABCA3) in sperm (Figure. 5E).

Fig. 5.

Global correlation analysis between the expression of proteins in SPEVs and proteins in sperm. A: The global correlation map of all proteins in SPEVs and sperm of the HF group. B: The global correlation maps of all proteins in SPEVs and sperm of the LF group. C: Heatmap of some proteins positively correlated in the HF group and negatively correlated in the LF group. The proteins at the bottom come from the SPEV, and the proteins on the right come from the sperm. The red gene names represent the upregulated proteins, while the blue gene names represent the downregulated proteins. * represents DEPs. D: Heatmap of some proteins negatively correlated in the HF group and positively correlated in LF group. E: The correlation network shows protein pairs with correlation coefficients greater than 0.9. The circles indicate that proteins were identified in SPEVs, while triangles indicate that proteins were identified in sperm. Red circles or triangles represent DEPs. The green lines indicate a positive correlation between the proteins in the HF and LF groups. The pink lines indicate a positive correlation between protein pairs in the HF group and a negative correlation between protein pairs in the LF group. The purple line indicates a positive correlation between protein pairs in the LF group and a negative correlation between protein pairs in the HF group.

Identification of key proteins related to sperm freezability

To further identify key proteins related to sperm freezability, we searched the Animal QTL database (https://www.animalgenome.org/cgi-bin/QTLdb/index) for QTLs related to sperm freezability. The results showed that 3 QTLs were associated with post-thaw motility, including 543 genes (Table S10). Among these genes, 17 (9 + 8) genes overlapped with DEPs in this study (Figure. 6 A). In addition, there are 24 (16 + 8) overlapping genes between hub proteins in important modules identified in sperm and DEPs identified in sperm (Figure. 6A). In total, 63 genes appeared in two or more analysis results simultaneously. Among them, 32 (16 + 8+8) proteins were present in sperm, 31 (11 + 9+8 + 3) proteins were present in SPEVs, and 8 proteins overlapped between sperm and SPEVs.

To evaluate whether certain proteins in SPEVs can mediate the impact of proteins in sperm on the sperm freezability, we performed mediation analysis using 63 proteins mentioned above. The results showed that 43 significant mediation linkages between these proteins (P < 0.05, Figure. 6B). Most of these linkages were related to CUTA, STX7, FISP2 and IPO5 in SPEVs and LPO, PTPRJ and CLPP in sperm. For example, PTPRJ is a protein specifically expressed in sperm of the HF group. However, the expression of PTPRJ was significantly downregulated in SPEVs of the HF group. Our mediation analysis suggested that PTPRJ in SPEVs may contribute to the sperm freezability by increasing expression level of PTPRJ in sperm (Figure. 6B).

Fig. 6.

Identification of key proteins and related to sperm freezability. A: Venn plot of proteins related to sperm freezability in the QTLs database, DEPs and hub proteins in significant modules of WGCNA. B: Mediation analysis of proteins in SPEVs, proteins in sperm, and phenotypes. C: Linkage disequilibrium (LD) pattern for the three genotyped SNPs within the HSPA1A gene. D: The protein expression levels of HSPA1A in sperm at the loci of SNPs (a) rs135145204, (b) 23:g.27520786G > T and (c) rs385826597 of different genotypes.

Identification of important genetic variation in candidate gene related to sperm freezability

To explore the sperm freezability markers of Holstein bulls, we detected the genetic variations of important candidate genes. Eighteen SNPs were identified within six candidate genes, including STK38, HSPA1A, HSP90B1, LPO, DNASE2 and CUTA. There were 7 mutations in exons, 3 intronic mutations, 2 promoter region mutations, 5 3’UTR mutations and 1 5’UTR mutation. Interestingly, all three SNPs were located within exon of HSPA1A. In addition, we found that the allele A > G of rs385826597 is a missense mutation, and cause the starting codon methionine (ATG) into a threonine (ACG) (Table 2). Furthermore, these three SNPs were found to be in linkage disequilibrium (Figure. 6C).

Table 2.

SNPs information of candidate genes associations with bull sperm freezability.

SNP	Gene name	Gene region	Position	Alleles [A/B]	Mutation type	Changes in Amino acids
rs135164605	STK38	3’UTR	23:10417465	C/T
23:g.10432393T > G		Intron6	23:10432393	T/G
23:g.10421180G > C		Exon11	23:10421180	G/C	Missense	E332A
23:g.10,419,332 C > G		Intron13	23:10419332	C/G
rs136094277		3’UTR	23:10419122	C/A
23:g.10418664G> A		3’UTR	23:10418694	G/A
rs381580439		3’UTR	23:10418156	A/G
23:g.10417595G > A		3’UTR	23:10417595	G/A
23:g.10,460,292 C > G		promoter	23:10460292	C/G
rs385826597		Exon1	23:27522566	A/G	missense	M5T
rs135145204	HSPA1A	Exon1	23:27522454	C/T	synonymous
23:g.27520819G > T		Exon1	23:27520819	G/T	synonymous
5:g.67613978T > G	HSP90B1	Intron11	5:67613978	T/G
5:g.67,598,097 A > G		promoter	5:67598097	A/G
rs135860093	LPO	Exon8	19:9234370	G/T	Missense	A311S
19:g.9,242,904 C > T		Exon9	19:9242904	C/T	synonymous
rs42312321	DNASE2	Exon4	7:12677843	T/C	Missense	V152A
rs136744810	CUTA	5’UTR	23:7593210	A/C

As the possible effects of mutations on gene expression, we measured HSPA1A protein expression in sperm with different genotypes at the loci of three SNPs. The results indicated that individuals with the TGG genotype have almost no expression or low expression level of HSPA1A in their sperm (Figure. 6D). This WB result was consistent with the lower expression of HSPA1A in sperm from the HF group. These findings provided additional support for the association between these SNP markers and sperm freezability trait.

Discussion

In this study, we used bulls with almost identical fresh sperm motility, but the semen of these individuals exhibited distinct sperm freezability phenotypes (high or low freezability) after cryopreservation. We investigated the proteomic characteristics of sperm and SPEVs in the HF and LF groups and identified some important proteins related to sperm freezability, such as HSP90B1, HSPA1A, BSP3, OGDH, LPO, SLC26A8, TEX101, VAMP3 and LY6K (Figure. 7). In addition, our findings suggested that an important SNP within the HSPA1A may affect sperm freezability of bulls.

Fig. 7.

Important proteins in the model that affect sperm freezability. The image below shows a zoomed-in section of the sperm. Proteins in sperm are represented by rectangles, while proteins in SPEVs are represented by ovals. Pink color indicates proteins upregulated in the HF group, while blue color indicates proteins downregulated in the HF group. This schematic was created with BioRender (https://biorender.com/).

Through comprehensive analysis of the results of DEPs and WGCNA, a total of 167 important proteins related to sperm freezability were identified in this study. Many proteins in the HSP family were identified in our results. The HSP family mainly consists of HSP70, HSP40, HSP90, HSP100, HSP60 and small heat stress proteins³². Among them, the HSP90 family proteins have been extensively studied in sperm freezing resistance. HSP90 family proteins are a series of highly conserved family proteins produced by cells under sudden stress to maintain normal physiological functions³³. HSP90 has been found to be localized in the tail of sperm²¹ and participate in regulating sperm motility¹⁸. It has been reported that the expression of HSP90 was positively correlated with sperm quality³⁴. In addition, the expression level of HSP90 was significantly higher in fresh semen than that in frozen-thawed semen³⁴. In this study, it was observed that the level of HSP90B1 expression, a member of the HSP90 protein family, was notably elevated in the SPEVs of the HF group compared to the LF group. Besides, HSP90B1 was assigned to the blue module, which is significantly positively correlated with high sperm freezability, indicating that the high expression of HSP90B1 may be beneficial for the freezability of sperm. HSP90B1 was significantly enriched in the “protein processing in endoplasmic reticulum” pathway, indicating that the HSP90B1 may regulate sperm freezability by affecting protein folding and assembly.

In addition, a recent study revealed a positive association between the expression level of HSPA1A (HSP70) in sperm cells and the redox status of the sperm³⁵. Spinaci et al. found that the cellular distribution of HSPA1A protein changes when sperm are subjected to stress³⁶. Furthermore, HSPA1A plays a key role in the MAPK signal pathway³⁷, which is involved in sperm damage induced by heat³⁸. In this study, HSPA1A is specifically expressed in sperm of the LF group and enriched in important GO terms such as “ATP binding” and “Protein processing in endoplasmic reticulum”, indicating that HSPA1A have a significant impact on the processes involved in the freezing and thawing of sperm. In addition, our findings suggested that a SNP (g23:27522566, rs385826597) in HSPA1A cause the starting codon methionine (ATG) into a threonine (ACG). Furthermore, the WB results showed that HSPA1A protein was almost not expressed or significantly reduced in the sperm of bulls with the TGG genotype. Considering that the other two SNPs in HSPA1A are synonymous mutations, we speculate that the missense mutation rs385826597 may reduce the protein expression of HSPA1A or affect protein function. Moreover, the WB result was consistent with the proteomic sequencing data, which showed lower expression of HSPA1A in sperm from the HF group and higher expression in the LF. Although HSPs, including HSPA1A, are generally considered cytoprotective, elevated expression may also signify excessive stress rather than improved resilience. Clinical studies showed that infertile men have higher HSP70 levels in their semen, which was negatively correlated with sperm concentration and motility³⁹. Considering that fresh semen samples prior to freezing were used in this study, the lower HSPA1A expression observed in HF bulls may reflect a more stable oxidative environment, while the increased expression in LF bulls could indicate heightened stress sensitivity, thereby impairing freezing tolerance. Therefore, we propose that the missense mutation rs385826597, together with two synonymous SNPs in HSPA1A, represents promising molecular markers for predicting and selecting bulls with superior semen cryotolerance.

The BSP family proteins, including BSP1, BSP3, and BSP5, have been identified as potentially associated with sperm cryopreservation^1,19. Previous study reported that the BSP family proteins are secreted by the seminal vesicles, encompassing BSP-A1/BSP-A2 (BSP1), BSP-A3 (BSP3), and BSP-30kD (BSP5)⁴⁰, which account for 50–65% of the total proteins in seminal plasma⁴¹. BSP family proteins cover the surface of spermatozoa by partially inserting themselves into the spermatid plasma membrane¹. Therefore, excessive exposure of sperm to BSP may have negative effects on the integrity of the sperm plasma membrane. The present investigation revealed a significant upregulation of BSP3 protein in the LF group. Additionally, it was assigned to the salmon module, which significantly negatively correlated with sperm freezability, indicating that the high abundance of BSP3 may be associated with low freezability.

It has been reported that SPEVs influence the function of mammalian sperm, especially sperm motility^15,16,42,43. The absence of related enzymes in spermatozoa can affect sperm function and lead to various signaling and metabolic defects⁴⁴. OGDH plays a crucial role as a pivotal enzyme within the tricarboxylic acid cycle (TCA)⁴⁵. We found that OGDH was an up-regulated DEP in SPEVs of HF group. Besides, the expression of OGDH in the sperm of the HF group was higher than that in the sperm of the LF group, although it did not reach a significance level. Its high expression may ensure that sperm have sufficient energy during activity. Additionally, SLC26A8(also known as Testis Anion Transporter 1 (Tat1)) localizes to the annulus of mammalian sperm and is crucial for the terminal differentiation of mouse sperm⁴⁶. Notably, sperm from SLC26A8-null male mice displayed structural defects⁴⁷. In the present study, SLC26A8 was an upregulated DEP in the SPEVs of the LF group. However, the expression trend of SLC26A8 was downregulated in the sperm of the LF group, whereas it was upregulated in the sperm of the HF group. Therefore, we hypothesize that there may be underlying causes preventing the transfer of SLC26A8 from SPEVs to sperm in the LF group, thereby impairing sperm function.

A recent study suggested that TEX101 interactome (known as TEX101 and its interaction network) was associated with male fertility⁴⁸. TEX101 is a glycoprotein that plays a crucial role in the processes of spermatogenesis and sperm function, located on the surface of germ cells⁴⁹. In human studies, enzyme-linked immunosorbent assay (ELISA) was employed to assess TEX101 levels in seminal plasma across various groups of patients, such as fertile men, individuals from couples with unexplained infertility, and men with sperm concentrations falling below the established reference range⁵⁰. The results indicated that TEX101 levels in pre-vasectomy men who were healthy and fertile were significantly higher compared to those observed in individuals with unexplained infertility, oligospermia, and azoospermia. Likewise, our study showed that the expression level of TEX101 in SPEVs of the HF group bulls was higher than that of the LF group bulls. In addition, we found that VAMP3 was significantly upregulated in SPEVs of the HF group bulls. It has been reported that VAMP3 is used to transport TEX101 protein to the cellular membrane⁵¹. Therefore, we speculate that the interaction between VAMP3 and TEX101 may have an important impact in maintaining membrane integrity, indirectly affecting membrane damage after sperm freezing-thawing. Furthermore, we found that the Lymphocyte Antigen 6 Family Member K (LY6K) was highly expressed in the SPEVs of HF group bulls. The GPI anchored protein complex LY6K/TEX101 has been demonstrated to play a role in facilitating sperm migration to the oviduct and in mediating sperm binding to the zona pellucida of the ovum in mice⁵². Our results indicated that TEX101 and its interaction proteins may play a crucial role in the post freezing vitality of spermatozoa.

Moreover, our study revealed the important relationship between proteins in sperm and proteins in SPEVs. The results showed distinct protein expression correlation patterns of sperm and SPEVs between the HF and LF groups. There were 142 protein pairs with a correlation greater than 0.9 in sperm and SPEVs, including five DEPs (ART3, DNASE2, ODF2, ABCA3 and DBI). In addition, our mediation analysis suggested that some proteins in SPEVs may affect the sperm freezability by affecting the proteins in sperm cells. For example, PDCD6, PTPRJ, STX7, IPO5, GCLC, FLOT1 and CUTA in SPEVs might contribute to the freezability of sperm by increasing the expression of LPO in sperm. In this study, importin-5 (IPO5) was identified concurrently in the blue module, as well as among the DEPs of the sperm group and the SPEV group. In addition, the protein expression of IPO5 was significantly upregulated in both sperm and SPEVs of HF group. The first importin identified as exhibiting a testicular expression profile was IPO5. During rodent testis development, cellular localization analyses of IPO5 proteins revealed their distinct patterns of cell-specific production⁵³. Several studies demonstrated that the normal expression of IPO5 is crucial for the development and maintenance of sperm function^54–56. Additionally, it is widely recognized that the freezing thawing process causes physical and chemical stress on the sperm plasma membrane, leading to oxidative stress and the formation of reactive oxygen (ROS)^57,58. It has been found that LPO may help maintain appropriate levels of H₂O₂ in cells, thereby protecting sperm from damage and inflammation caused by H₂O₂^58,59. Further research is required to explore the molecular mechanisms underlying the interactions of these proteins and their impact on sperm motility following the process of cryopreservation.

However, it should be noted that the proteomic analysis in this study was based on a relatively small sample size. While the findings provide valuable associations between sperm and seminal plasma extracellular vesicle proteins and sperm freezability, the limited number of animals may reduce the statistical power and the ability to fully elucidate underlying mechanisms. Future studies with larger cohorts are warranted to validate these results and further explore the mechanistic pathways involved.

Conclusion

The present study described a comprehensive analysis of the proteome of sperm and SPEVs from Holstein bulls with different post-thawing sperm motility. We found a number of important proteins in sperm and SPEVs associated with sperm freezability. In addition, three important SNPs in HSPA1A can serve as sperm freezability potential biomarkers for distinguishing between the excellent and poor freezer breeding bulls. These findings provide new perspectives and better understanding for further study of sperm freezability.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

SPEV

Seminal plasma extracellular vesicles

DEPs

Differentially expressed proteins

WGCNA

Weighted correlation network analysis

QTLs

Quantitative trait loci

SNPs

Single nucleotide polymorphisms

AI

Artificial Insemination

EVs

Extracellular vesicles

HF

High freezability

LF

Low freezability

GS

Gene significance

MM

Module membership

GO

Gene Ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

GSEA

Gene set enrichment analysis

HWE

Hardy-Weinberg equilibrium

LD

Linkage disequilibrium

TEM

Transmission electron microscopy

NTA

Nanoparticle tracking analysis

ROS

Reactive oxygen species

Author contributions

LJ conceived and designed the study. JC, CZ and XZ performed the experiments. JC and BL analyzed the data. YW and YZ provided technical support. YZ and YL assisted in collecting experimental samples. LJ, JC, BL and CZ wrote the paper and LJ and BL revised the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported financially by the National Key Research and Development Program of China (2022YFD1302200), the earmarked fund for CARS36, and National Natural Science Foundation (3210200137).

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the dataset identifiers PXD053574 and PXD053624(http:/www.ebi.ac.uk/pride/).

Declarations

Competing interests

The authors declare no competing interests.