SPSB3-mediated K48- and K63- linked ubiquitination and degradation of TUFM promote apoptosis induced by myocardial ischemia/reperfusion injury

Abstract

Hypoxia, caused by various factors, significantly threatens human health by damaging the heart and cardiomyocytes. Revascularization strategies can sometimes precipitate adverse clinical outcomes due to myocardial ischemia/reperfusion (I/R) injury. The primary objective of this investigation was to explore the mechanism underlying myocardial I/R injury to identify novel alleviating strategies. CRISPR/Cas9 technology was used for high-throughput screening of hypoxia/reoxygenation (H/R)-tolerant genes; afterward, the genes were intersected with differentially expressed genes in the GSE61592 dataset. Subsequent validation identified SPRY domain-containing SOCS box protein 3 (SPSB3) as a critical gene that confers resistance to H/R injury. Increased apoptosis and worse heart function were observed in mouse and cardiomyocyte models of myocardial I/R injury. These pathological alterations coincided with a marked elevation in SPSB3 expression. Furthermore, SPSB3 inhibition substantially attenuated the myocardial I/R injury-induced increase in cardiomyocyte apoptosis, decrease in cardiac function, and mitochondrial dysfunction. To identify SPSB3 substrates, we combined the results from mass spectrometry, ubiquitin-modified proteomics, and the MitoCarta3.0 database (cardiac mitochondrial proteins), yielding six candidate molecules. Co-immunoprecipitation and western blot analyses indicated that SPSB3 may bind to Tu translation elongation factor, mitochondrial (TUFM). In SPSB3-knockdown cardiomyocytes, additional TUFM knockdown partially reversed the protective effect of SPSB3 knockdown alone. Additionally, MG132 and Cycloheximide treatment effectively inhibited TUFM degradation. Further amino acid site mutagenesis and other analyses revealed that SPSB3 inhibition prevents K48- and K63-linked ubiquitination at the K259 residue of TUFM and TUFM degradation, providing a promising therapeutic avenue for mitigating myocardial I/R injury.

Graphical Abstract

1. Myocardial ischemia/reperfusion (I/R) injury upregulates SPSB3 expression, increases cell damage and apoptosis, and reduces cardiac function.

2. Knockdown of SPSB3 alleviates myocardial I/R injury-induced cell damage and apoptosis, restoring cardiac function.

3. In myocardial I/R injury, SPSB3 interacts with TUFM, promoting K48 and K63-linked ubiquitination at TUFM K259, which enhances apoptosis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10565-026-10199-8.

Introduction

Acute myocardial infarction (AMI) constitutes a severe, life-endangering cardiovascular disorder stemming chiefly from ischemic and hypoxic insults to the coronary vascular (Heusch and Gersh 2017). Although established therapeutic strategies, including thrombolysis and percutaneous coronary intervention, demonstrate considerable efficacy in limiting infarct expansion and maintaining global cardiac function, the adverse consequences of myocardial ischemia/reperfusion (I/R) injury remain a formidable clinical obstacle. Such persistent damage drives cardiomyocyte injury and death, thereby accelerating the progression toward several adverse clinical outcomes (Heusch 2024; Reddy et al. 2024). Alteration of the expression levels of specific genes or proteins can alleviate myocardial I/R injury; however, clinical applications are lacking. Presently, there are no available treatments for the effective prevention of myocardial cell death after myocardial I/R injury (Caccioppo et al. 2019). Thus, identifying the key pathogenic genes involved in this process could provide novel therapeutic targets for clinical intervention.

CRISPR/Cas9 library screening is a high-throughput method used to identify genes associated with cell death factors (Cheng et al. 2023; Shalem et al. 2014). This approach leverages lentiviral vectors encoding single-guide RNAs (sgRNAs) in addition to the Cas9 nuclease to generate comprehensive genome-wide knockout libraries (Joung et al. 2017). To identify effective therapeutic targets for myocardial I/R injury, we constructed a gene knockout library using sgRNA vector viruses and subjected the cells to hypoxia/reoxygenation (H/R) treatment to screen for H/R-resistant genes. In the present study, the SPSB3 gene was successfully identified. Elevated SPSB3 expression was closely related to enhanced apoptosis in myocardial I/R injury. Furthermore, SPSB3 knockdown reduced cardiomyocyte apoptosis and improved cardiac function following myocardial I/R injury, indicating that SPSB3 could serve as a novel therapeutic target.

Although no studies have directly linked SPSB3 to myocardial I/R injury, accumulating evidence suggests that SPSB3 exerts a pivotal influence on the pathogenesis of diverse pathological conditions, notably encompassing malaria and colorectal malignancy (Huo et al. 2017; Li et al. 2021). Protein ubiquitination represents a prevalent post-translational mechanism, modulates biological activity, and directs proteolytic degradation. Furthermore, it is indispensable to several mechanisms, encompassing, proliferation, apoptosis and damage repair (Kerscher et al. 2006; Popovic et al. 2014; Yadav et al. 2022). SPSB3, as a constituent of the E3 protein ligase complex, orchestrates in the regulation of protein degradation (Liu et al. 2018; Xu et al. 2024). To explore the mechanisms of SPSB3 in this context, we used mass spectrometry and post-translational modification omics results of ubiquitination to identify the potential substrate proteins. Our results revealed that Tu translation elongation factor, mitochondrial (TUFM) may serve as a substrate for SPSB3 and may be involved in the regulation of H/R-induced cardiomyocyte apoptosis.

Our findings are the first to indicate that SPSB3 interacts with TUFM, promotes its K48- and K63-linked ubiquitination at K259, and mediates TUFM degradation via the proteasome pathway, thereby affecting apoptosis and myocardial I/R injury. This study elucidated the function of SPSB3 within the context of myocardial I/R injury while evaluating its viability as a therapeutic target, offering potential strategies avenues for clinical intervention.

Results

Screening of genes related to AC16 cardiomyocyte death caused by H/R injury via the CRISPR/Cas9 library

We used whole-genome CRISPR/Cas9 library technology to screen AC16 cardiomyocytes under H/R conditions. This method involved the construction of a whole-genome sgRNA lentiviral library, inoculation of the cells, and the use of H/R lethal factors to obtain H/R-tolerant cells. The sgRNA levels in these cells were compared using sequencing. Eventually, a series of genes that could tolerate H/R conditions after knockout were identified. Thus, these genes represent possible therapeutic candidates for H/R injury. After 12 h of hypoxia and 6 h of reoxygenation (Wang et al. 2012; Yan et al. 2017), most wild-type AC16 cells had died, thus meeting the selection criteria (Supplementary Fig. 1). Human GeCKO Library A and lentiCRISPR v2 vector (developed by the Zhang Feng laboratory) were used for sgRNA transduction, followed by puromycin selection and high-throughput sequencing (Supplementary Fig. 2); the specific process is illustrated in Fig. 1A. A total of 68 sgRNA hits were identified (Supplementary File 2), which are presented in a bar chart (Fig. 1B) showing the data for the top 20 hits.

Fig. 1.

Results of CRISPR/Cas9 screening, bioinformatics analysis, and candidate molecule validation A Screening process of CRISPR/Cas9 library for H/R resistance genes. B A bar plot displaying the top 20 hits from the CRISPR/Cas9 library screen. C Venn diagram showing the intersection of the CRISPR/Cas9 screen and GSE61592 (I/R) upregulated genes, along with the corresponding sgRNA counts and ranks of the eight overlapping molecules. D Protein levels of MAU2 and SPSB3 in the sham and I/R groups. Quantitative data are presented in the accompanying graph (n = 3). E Volcano plot was used to show the differentially expressed genes between the con and I/R groups in the GSE61592 dataset. F GO enrichment analysis results of differentially expressed genes. G KEGG enrichment analysis results of differentially expressed genes. All data are presented as the mean ± SD. ns, not significant, ^***p < 0.001, determined via a two-tailed t-test

By intersecting our screening hits with genes showing elevated expression in the I/R group of the GSE61592 dataset, we identified eight candidate genes: MAU2, SPSB3, SMTNL2, KMT2C, LDHD, TSPAN32, PDZD2, and NDUFA9. Figure 1C shows the ranks of all molecules on the CRISPR/Cas9 screen. Notably, MAU2 and SPSB3 were among the top 20 hits out of the 68 genes and exhibited high abundance counts, suggesting a higher likelihood of these genes conferring resistance to H/R injury in cardiomyocytes. Consequently, MAU2 and SPSB3 were selected for subsequent validation. After 45 min of ischemia and 48 h of reperfusion (Ge et al. 2023; Jiang et al. 2022), we observed a marked upregulation of SPSB3 protein levels in the I/R group, whereas MAU2 protein expression was not significantly different (Fig. 1D). Further analysis of the GSE61592 microarray data revealed 855 differentially expressed genes (Fig. 1E). Gene Ontology (GO) enrichment analysis revealed substantial enrichment of cellular components (CCs) such as the mitochondrial inner membrane, protein complexes, and matrix (Fig. 1F). Additionally, the Kyoto Encyclopedia of Genes and Genomes pathway analysis found enrichment of the apoptotic pathway (Fig. 1G), suggesting that apoptosis mediated by mitochondrial pathways could be pivotal in myocardial I/R injury.

H/R injury significantly upregulated SPSB3 expression and induced cell apoptosis in AC16 cardiomyocytes

Figure 2A shows the schematic diagram of the H/R model, and relevant indicators were determined. We investigated the alterations in SPSB3 expression at the cellular level and the extent of cell damage and apoptosis. Lactate dehydrogenase (LDH) content and CCK8 activity were used to evaluate differences in cell damage between the two groups. H/R elevated LDH content in the cell supernatant, while reducing cardiomyocyte viability (Fig. 2B, C). In the H/R group, SPSB3 protein and mRNA levels were significantly elevated (Fig. 2D, E). Additionally, immunofluorescence results confirmed increased SPSB3 protein expression in the H/R group (Fig. 2F). Assessment of apoptosis-related proteins indicated that H/R decreased the Bcl2/Bax ratio while concurrently elevating the protein levels of cleaved Caspase-3/9 (Fig. 2G). TUNEL assay demonstrated a markedly elevated fraction of TUNEL-positive cells in the H/R group relative to the con group (Fig. 2H). Flow cytometry analysis using Annexin V-FITC/PI showed that H/R increased the degree of apoptosis (Fig. 2I).

Fig. 2.

H/R injury significantly upregulated SPSB3 expression and induced cell apoptosis in AC16 cardiomyocytes A Schematic diagram of cell hypoxia/reoxygenation modeling. B LDH levels in the culture supernatants of the con and H/R group cells (n = 3). C OD450 values were measured after the cells were incubated with CCK8 reagent for 1.5 h (n = 3). D Protein expression of SPSB3 between the two groups (n = 3). E mRNA expression of SPSB3 between the two groups (n = 3). F Immunofluorescence images in the control and H/R groups at 200 × magnification (n = 3). Scale bar = 50 μm. G Protein expression of Bax, Bcl2, cleaved Caspase-3, and cleaved Caspase-9 between the two groups (n = 3). H TUNEL staining outcomes for each group of cells at 200 × magnification (n = 3). Scale bar = 50 μm. I Proportions of apoptotic cells in each group, as measured via Annexin V/FITC flow cytometry (n = 3). All data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, determined via a two-tailed t-test

Myocardial I/R injury increased SPSB3 expression alongside heightened cardiomyocyte apoptotic rates and decreased cardiac function

Figure 3A presents the schematic diagram of the I/R model, with relevant indicators detected, and Fig. 3B shows electrocardiogram recordings from the mouse I/R model at various time points. Hematoxylin and eosin (H&E) staining revealed a disorganized myocardial structure in the I/R group, including partial fragmentation or dissolution of the nuclei and infiltration of inflammatory cells (Fig. 3C). I/R led to functional decline, manifesting as decreased ejection fraction (EF) and fractional shortening (FS) concomitant with increased left ventricular end-diastolic diameter (LVIDd) and left ventricular end-systolic diameter (LVIDs) [Fig. 3D]. The myocardial infarction area appeared white after the TTC staining. I/R exacerbated the extent of myocardial infarction (Fig. 3E). Serum LDH, CK-MB, and cTnI levels increased in the I/R group (Fig. 3F). Furthermore, I/R increased SPSB3 mRNA and protein levels (Fig. 3G, H). Immunohistochemical analysis revealed that I/R significantly upregulated SPSB3 expression (Fig. 3I). Furthermore, the assessment of apoptosis-related proteins indicated that I/R diminished the Bcl-2/Bax ratio while increasing the abundance of Caspase-3/9 (Fig. 3J). I/R treatment led to an increase in apoptotic cells, as detected by TUNEL staining (Fig. 3K).

Fig. 3.

Myocardial I/R injury increased SPSB3 expression alongside heightened cardiomyocyte apoptotic rates and decreased cardiac function A Schematic diagram of myocardial I/R injury modeling. B Electrocardiography at different times during surgery. C H&E staining in the sham and I/R groups, with red arrows marking the mitochondria at 200 × magnification (n = 6). Scale bar = 50 μm. D Representative ultrasound images of the sham and I/R groups and the EF, FS, LVIDd, and LVIDs results (n = 6). E TTC staining in each group (n = 6). Scale bar = 1 cm. F Serum levels of the myocardial enzymes CKMB, cTnI, and LDH (n = 6). G Protein expression of SPSB3 between the two groups (n = 6). H mRNA expression of Spsb3 between the two groups (n = 6). I Immunohistochemistry analysis revealed SPSB3 protein expression and localization in sham and I/R groups at 200 × magnification (n = 6). J Protein expression of Bax, Bcl2, cleaved Caspase-3, and cleaved Caspase-9 between the two groups (n = 6). K TUNEL staining outcomes of each group of cells at 200 × magnification (n = 6). Scale bar = 50 μm. All data are presented as mean ± SD. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, determined via a two-tailed t-test

SPSB3 knockdown reduced H/R-induced cardiomyocyte apoptosis

SPSB3 expression was modified by transfecting with a lentivirus, which knocked down SPSB3 expression. Based on the WB and RT-qPCR results, SPSB3-KD was selected for the follow-up experiments. Our findings demonstrated that the viral vector potently suppressed SPSB3 abundance, with significant downregulation observed at the transcript and protein levels (Fig. 4A, B); the same conclusion was obtained via immunofluorescence (Fig. 4C). Reduced SPSB3 expression alleviated H/R-induced damage to cardiomyocytes (Fig. 4D) and improved cell viability (Fig. 4E). The detection of apoptosis-related proteins revealed that the SPSB3-KD can partially attenuated the H/R-induced upregulation of cleaved Caspase-3/9 and numerically increased the decrease in the ratio of Bcl2/Bax, but there was no statistical difference (Fig. 4F). Annexin V-FITC/PI flow cytometry revealed that SPSB3-KD partially reduced the H/R-induced increase in apoptosis (Fig. 4G). JC-1 is a fluorescent lipid dye used to monitor mitochondrial membrane potential. When cells are damaged or undergo apoptosis, the mitochondrial membrane potential changes, with an increase in the proportion of green fluorescent cells. In the present study, JC-1 staining revealed that SPSB3-KD reduced the changes in mitochondrial membrane potential and partially reduced the increase in the proportion of H/R-induced green fluorescent cells (Fig. 4H). Furthermore, transmission electron microscopy (TEM) was used to characterize cellular ultrastructural morphology. To further delineate cellular ultrastructural architecture, we employed transmission electron microscopy (TEM). Through a systematic quantification of mitochondria across numerous cellular cross-sections, we classified organelles as damaged based on distinct morphological hallmarks, specifically cristae disruption and matrix edema. Subsequently, we derived the fractional ratio of damaged mitochondria relative to the total mitochondrial population. H/R treatment significantly elevated the fractional ratio of damaged mitochondria to total mitochondria. H/R-induced mitochondrial injury was effectively ameliorated by SPSB3 knockdown (Fig. 4I). Additionally, TUNEL staining revealed that SPSB3-KD partially attenuated the elevation in TUNEL-positive cells triggered by H/R (Fig. 4J).

Fig. 4.

SPSB3 knockdown reduced H/R-induced cardiomyocyte apoptosis A SPSB3 protein expression in each group (NC, SPSB3-KD, NC + H/R, SPSB3-KD + H/R) was detected via western blotting (n = 3). B mRNA expression of SPSB3 in each group (n = 3). C Immunofluorescence images in each group at 200 × magnification. Scale bar = 50 μm. D LDH levels in the culture supernatants of each group (n = 3). E OD450 values were measured after the cells in each group were incubated with CCK8 reagent for 1.5 h (n = 3). F Protein expression of Bax, Bcl2, cleaved Caspase-3, and cleaved Caspase-9 in each group and their statistical results (n = 3). G Proportions of apoptotic cells in each group as measured via Annexin V/FITC flow cytometry; the proportion of apoptotic cells is outlined. Quantitative data are presented in the accompanying graph (n = 3). H Fluorescence staining results of JC-1 probes in each group based on flow cytometry; the proportion of green cells is outlined. Quantitative data are presented in the accompanying graph (n = 3). I Representative results from transmission electron microscopy for each group are provided, with red arrows indicating the damaged mitochondria at 15,000 × magnification (n = 4). Scale bar = 1 μm. J TUNEL staining results for each group of cells at 200 × magnification (n = 3). Scale bar = 50 μm. All data are presented as mean ± SD. ns, not significant, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, determined via one-way ANOVA

SPSB3 knockdown reduced cardiac function impairment and myocardial apoptosis caused by I/R

We regulated SPSB3 expression in mice by injecting adeno-associated virus (AAV) into the tail vein. Four weeks after the virus injection, fluorescence microscopy revealed an effective cardiac infection that met the experimental requirements (Fig. 5A). Our findings demonstrated that the AAV9-shSPSB3 vector potently suppressed SPSB3 abundance, with significant downregulation observed at both transcript and protein levels (Fig. 5B, C), thereby attenuating the severity of I/R-induced myocardial infarction (Fig. 5D). Compared to AAV9-NC + I/R, AAV9-shSPSB3 significantly reduced inflammatory cell infiltration and myocardial disorganization (Fig. 5E). AAV9-shSPSB3 ameliorated cardiac dysfunction caused by I/R. Specifically, the AAV9-shSPSB3 + I/R group presented lower EF and LVFS values than did the AAV9-NC + I/R group, whereas LVIDd and LVIDs values were higher in the AAV9-shSPSB3 + I/R group (Fig. 5F). The I/R-induced elevation in myocardial enzymes (LDH and CK-MB) levels were attenuated by AAV9-shSPSB3 treatment (Fig. 5G). AAV9-shSPSB3 treatment counteracted the I/R-induced apoptotic signals, namely the increase in cleaved Caspase-3/9 and the decrease in the Bcl-2/Bax ratio (Fig. 5H). TUNEL staining revealed that AAV9-shSPSB3 partially reduced I/R-induced increase in apoptosis (Fig. 5I).

Fig. 5.

SPSB3 knockdown reduced cardiac function impairment and myocardial apoptosis caused by I/R A Results of spontaneous fluorescence in frozen myocardium after adeno-associated virus infection at 40 × magnification. Scale bar = 50 μm. B SPSB3 protein expression in each group (AAV9-NC + sham, AAV9-shSPSB3 + sham, AAV9-NC + I/R, and AAV9-shSPSB3 + I/R) was detected via western blotting (n = 6). C mRNA expression of Spsb3 in each group (n = 6). D TTC staining in each group (n = 6). Scale bar = 1 cm. E H&E staining in the sham and I/R groups, with red arrows indicating the mitochondria at 200 × magnification (n = 6). Scale bar = 50 μm. F EF, FS, LVIDd, and LVIDs results in each group (n = 6). G Serum levels of the myocardial enzymes CKMB and LDH (n = 6). H Bax, Bcl2, cleaved Caspase-3, and cleaved Caspase-9 protein expressions in each group (n = 6). I TUNEL staining outcomes of each group at 200 × magnification (n = 6). Scale bar = 50 μm. All data are presented as mean ± SD. ns, not significant, ^*p < 0.05, ^**p < 0.01, ^***p < 0.001, determined via one-way ANOVA

Selection of target proteins for SPSB3

We successfully constructed a lentivirus expressing Flag-SPSB3 and used it for transfection (Fig. 6A). AC16 cardiomyocytes were first transfected with Flag-SPSB3 and then subjected to either control (con) or H/R conditions. Immunoprecipitation (IP) experiments were performed using a Flag antibody after electrophoresis. After electrophoretic fractionation, Coomassie Brilliant Blue staining was applied to the gel to reveal the separated protein bands (Fig. 6B). Lanes 1 and 2 show the input lysates from control and H/R conditions, respectively, which served as positive controls. Lanes 3 and 4 correspond to the IP results using a control IgG antibody under the con and H/R conditions (labeled IgG1 and IgG2, respectively). Lanes 5 and 6 represent the IP results using an anti-FLAG antibody under the con and H/R conditions (labeled IP1 and IP2). Mass spectrometry was employed to determine proteins binding to the Flag-SPSB3 complex, and an ion chromatogram representing the SPSB3 protein was extracted and displayed, confirming the reliability of the mass spectrometry data (Fig. 6C). To identify the specific interacting proteins, we performed parallel control IP using non-specific IgG antibodies [con (IgG1): 44 samples, H/R (IgG2): 41 samples]. The proteins in the IP and IgG groups were considered background-binding proteins (con: 35 samples, H/R group: 26 samples). Specific interacting proteins (con: 132 samples, H/R group: 45 samples) were identified by removing background-binding proteins. To ensure a comprehensive profile, the union of specific interacting proteins from both conditions was performed, which resulted in 159 candidate proteins that could potentially bind to SPSB3 (Fig. 6D). SPSB3 acts as a part of E3 ligase, which is crucial for its biological function. Therefore, we focused on its role as an E3 ligase during myocardial I/R injury.

Fig. 6.

Selection of target proteins for SPSB3 A The protein expression of SPSB3 in each group (Flag-NC, Flag-NC + H/R, Flag-SPSB3, and Flag-SPSB3 + H/R) were detected via western blotting (n = 3). B Coomassie Brilliant Blue staining of Flag-SPSB3 co-immunoprecipitates from AC16 cardiomyocytes after control or H/R treatment. Lanes 1&2: Input1 (con) and Input2 (H/R); 3 & 4: IgG control IPs, IgG1 (con) and IgG2 (H/R); 5 & 6: anti-Flag IPs, IP1 (con) and IP2 (H/R). C SPSB3 spectra extracted from the mass spectrometry detection results. D Venn analysis of specific SPSB3-interacting proteins. Specific interacting proteins were identified by subtracting background-binding proteins (both in IgG and the corresponding IP group) from anti-Flag IP results, yielding 132 con and 45 H/R proteins. The union of these specific interacting proteins resulted in 159 candidate proteins. E Distribution statistics of differentially expressed proteins and differentially modified sites in protein ubiquitination modification omics. F Volcano plot of differentially modified sites. G Subcellular localization classification rose plot of differentially modified proteins. H COG/KOG enrichment analysis of differentially modified proteins. I Venn diagram of the intersection of differential proteins in ubiquitination proteomics and mitochondrial genes and the combined specific interacting proteins from mass spectrometry results. All data are presented as mean ± SD. ^***p < 0.001, determined via one-way ANOVA

Consequently, we performed a comprehensive ubiquitination modification omics analysis of the con and H/R cell groups. Compared with the con group, 2827 proteins were upregulated, 3195 proteins downregulated, 9391 ubiquitination modification sites upregulated, and 8553 ubiquitination modification sites downregulated in the H/R group (Fig. 6E). The volcano plot presents the results of the obtained ubiquitination modification sites and shows information on the top five differential modification sites for the upregulated and downregulated sites (Fig. 6F). Subcellular localization profiling of the two distinct sets of differentially expressed proteins demonstrated that, in addition to the nucleus, cytoplasm, and plasma membrane, these proteins were significantly enriched in the mitochondria (Fig. 6G). The COG/KOG pathway enrichment analysis revealed enrichment of the “Post-translational modification” and “Translation” pathways (Fig. 6H). Six proteins, HSPD1, ATP5F1A, ATP5F1B, SLC25A3, TUFM, and VDAC1, were identified by intersecting the differentially expressed proteins from ubiquitination proteomics, mitochondrial genes located in the heart (from the MitoCarta 3.0 database), and the combined specific interacting proteins obtained via mass spectrometry (Fig. 6I).

SPSB3 could bind to TUFM and regulated H/R-induced apoptosis

Among the six candidate molecules, we precipitated ATP5F1A, ATP5F1B, TUFM, and VDAC1 with the Flag antibody (Fig. 7A). While protein levels of ATP5F1A, ATP5F1B, and VDAC1 remained statistically unchanged in the H/R model, we observed a significant downregulation of TUFM expression (Fig. 7B). We propose that TUFM could engage with SPSB3 as a substrate. The GST pull-down assay showed a combination of GST-SPSB3 and Flag-TUFM (Fig. 7C). Based on immunofluorescence results, SPSB3 and TUFM had co-localization in cells (Fig. 7D). Molecular docking prediction of SPSB3 and TUFM was performed. During the docking process of TUFM and SPSB3, the arginine 25 site of SPSB3 was bound to the glutamate 93 site of TUFM, with two covalent bonds between them. In addition, the arginine 19 site of SPSB3 was linked to the tyrosine 266 and histidine 112 sites of TUFM, with two covalent bonds with tyrosine and one covalent bond with histidine (Fig. 7E).

Fig. 7.

SPSB3 could bind to TUFM and regulated H/R-induced apoptosis A Co-IP results of Flag and candidate proteins, including HSPD1, ATP5F1A, ATP5FIB, SLC25A3, TUFM, and VDAC1. B ATP5F1A, ATP5FIB, TUFM, and VDAC1 expressions in the con and H/R groups (n = 3). C GST pull-down assays showed that GST-SPSB3 interacts with Flag-TUFM. D Immunofluorescence images of the con and H/R groups were observed under a confocal microscope, at 600 × magnification. Scale bar = 10 μm. E Molecular docking results of SPSB3 and TUFM. Purple lines indicate π‒π interactions or cation‒π interactions, and the sites belonging to SPSB3 are underlined. F–H The next experiments were divided into the following six groups: a, NC; b, NC + H/R; c, SPSB3-KD; d, SPSB3-KD + H/R; e, SPSB3-KD + TUFM_KD; and f, SPSB3-KD + TUFM_KD + H/R. F SPSB3 and TUFM expressions in the above-mentioned six groups; a line graph showing the changes (n = 3). G The expression of apoptosis-related proteins in each group (a-f), as detected via western blotting (n = 3–5). H Proportions of apoptotic cells in each group (a-f), as measured via Annexin V/FITC flow cytometry; the proportion of apoptotic cells is circled. (n = 3). Horizontal lines and accompanying labels show the statistical significance of intergroup comparisons. All data are presented as mean ± SD. ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, determined via one-way ANOVA

After SPSB3 knockdown, TUFM was knocked down to detect changes in SPSB3 and TUFM protein levels. When SPSB3 expression increased, TUFM expression decreased (Fig. 7F); in addition, SPSB3 regulated TUFM expression. H/R conditions significantly induced cellular apoptosis, as evidenced by a marked increase in cleaved Caspase-3/9 levels, coupled with decreased Bcl-2/Bax ratio. SPSB3 knockdown (SPSB3-KD) under H/R conditions attenuated these apoptotic markers, consistent with our previous observations. To investigate whether the anti-apoptotic effect of SPSB3 knockdown was mediated by TUFM, we generated a double-knockdown group (SPSB3-KD + TUFM_KD). Under H/R conditions, the protective effects conferred by SPSB3 knockdown were abolished by concomitant TUFM knockdown. Specifically, in the double knockdown group under H/R conditions, the Bcl-2/Bax ratio showed an intermediate value between the NC and SPSB3-KD groups; however, this difference was not statistically significant. Likewise, the abundance of cleaved Caspase-9 within the double-knockdown cohort was not significantly different from that in the NC or SPSB3-KD group. While cleaved Caspase-3 expression was substantially diminished in the double-knockdown cells relative to the NC, this reduction did not reach statistical significance when contrasted with the SPSB3-knockdown condition (Fig. 7G). Annexin V-FITC/PI flow cytometry under H/R conditions revealed that the double-knockdown group exhibited a markedly elevated apoptotic fraction relative to the SPSB3-KD group; however, no statistically significant divergence was observed when compared with the negative control (Fig. 7H).

SPSB3 mediated K48- and K63-linked ubiquitination and degradation of TUFM

We investigated the mechanism underlying SPSB3-mediated TUFM expression. H/R treatment did not affect TUFM transcription but significantly reduced its translational abundance (Fig. 8A), suggesting a possible post-translational modification of TUFM protein. TUFM can be modified by ubiquitination in AC16 cardiomyocytes, as confirmed via a co-IP assay using a TUFM antibody (Fig. 8B). Cardiomyocytes were treated with cycloheximide (CHX) and MG132. In the case of inhibition of intracellular protein synthesis, TUFM protein expression gradually decreased in cells treated with CHX over time; the addition of MG132 inhibited this gradual decrease (Fig. 8C). Furthermore, co-transfection experiments using HEK293T cells with HA-ubiquitin and His-TUFM demonstrated that the introduction of Flag-SPSB3 significantly increased TUFM ubiquitination (Fig. 8D).

Fig. 8.

SPSB3 mediated K48- and K63-linked ubiquitination and degradation of TUFM A mRNA and protein expression of TUFM in the con and H/R groups (n = 3). B Using TUFM for IP and IB was performed on UB, SPSB3, and TUFM in AC16s. C Administration pattern at the same time point is illustrated: black arrow signifies CHX administration alone, whereas red arrow signifies CHX and MG132. The protein expression of TUFM under CHX or CHX + MG132 treatment (n = 3). D IP was performed using His-tag under different conditions in HEK293T cells, and IB was performed on His, Flag, and HA; and it also demonstrated the degree of change in ubiquitination. E After separately expressing different sites of ubiquitin in HEK293T cells, IP was performed using His-tag and IB was performed on His, Flag, HA, and β-tubulin. Values > 5 are marked; blue indicates a decrease, whereas red indicates an increase. F Heat map showing TUFM expression at different ubiquitination sites in the omics results. G After mutation at K259 lysine site to arginine (K259R) of TUFM in HEK293T cells, IP was performed using His-tag and IB was performed on His, Flag, HA, and β-tubulin

Furthermore, we constructed various ubiquitin mutants, each retaining one lysine residue at K6, K11, K27, K29, K33, K48, or K63 and cloned each mutant into an HA-tagged vector. HEK293T cells were co-transfected with His-TUFM, Flag-SPSB3, or HA-tagged plasmid vectors. The result showed that SPSB3-overexpressing cells exhibit higher levels of TUFM K6-, K48-, and K63-linked polyubiquitination. The ubiquitination level at the K48 site changed the most, followed by those at the K6 and K63 sites. Ubiquitination levels at other sites did not change significantly (Fig. 8E). The levels of TUFM protein decreased significantly at the K48 and K63 sites, whereas no significant change was observed at the K6 site. We observed a marked reduction in protein levels associated with both K48 and K63 ubiquitination sites, while K6-linked TUFM modification remained statistically unaltered. This indicates that the K48 and K63 sites may mediate SPSB3-induced TUFM degradation.

In the H/R group, SPSB3 expression and TUFM ubiquitination increased. Ubiquitination modification omics analysis revealed that the K259 site of TUFM increased in the H/R group, whereas the expression levels of other sites (K58, K91, K94, K237, K289, K314, K364 and K398) decreased (Fig. 8F). This suggests that K259 may be the amino acid site of TUFM that binds to ubiquitin. The subsequent data demonstrated that the ubiquitination function of SPSB3 in TUFM was lost when lysine 259 (K259) of TUFM was substituted with arginine (Fig. 8G).

Discussion

Fueled by population growth and aging, the prevalence of cardiovascular diseases is steadily increasing (Wang et al. 2025). Myocardial I/R injury remains a challenge for cardiovascular doctors (Welt et al. 2024). A large-scale CRISPR/Cas9 lentiviral library was used to transfect AC16 cells. Following H/R to simulate myocardial I/R injury, key genes linked to cardiomyocyte death were identified. The upregulation of SPSB3 in both H/R and I/R models suggests its potential critical role in the pathogenesis of myocardial I/R injury. Inhibition of SPSB3 expression attenuates cardiomyocyte apoptosis and ameliorates myocardial I/R injury. To dissect the molecular mechanisms involved, we employed various techniques, including 4D-fast DIA ubiquitination proteomics, mass spectrometry, and co-IP. To our current understanding, this investigation represents the first discovery and validation of the interaction between SPSB3 and TUFM proteins in cardiomyocytes. When SPSB3 binds to TUFM, it promotes K48- and K63-linked ubiquitination at K259, thereby regulating TUFM degradation via the proteasomal pathway. This interaction leads to increased apoptosis following myocardial I/R injury.

SPSB3 is highly homologous to other suppressor of cytokine signaling (SOCS) family members, such as SPSB1, SPSB2, and SPSB4, and has similar domains, including a SOCS box and an SPRY domain (Kile et al. 2002). SPSB3 may exert its biological effects by regulating various processes such as inflammatory responses and immune activation (Algoet et al. 2023; Davidson et al. 2019). At present, there is no conclusive evidence linking SPSB3 to myocardial I/R injury or similar I/R injuries in other organs. Our study demonstrates that inhibiting SPSB3 alleviates apoptosis triggered by I/R.

Protein–protein interactions can alter a protein's conformation, activity, or localization, thereby affecting its biological functions. As a core mechanism, they regulate cellular signaling, metabolic pathways, and gene expression (Jung-Klawitter et al. 2023; Yu et al. 2025). Protein–protein interactions can trigger changes in post-translational modifications (Zhao et al. 2025). Ubiquitination is a common post-translational mechanism that targets proteins for degradation or regulates signaling pathways (Chen et al. 2025). Cullin-RING E3 ubiquitin ligases (CRLs) constitute fundamental elements of the ubiquitin–proteasome machinery, exerting pivotal control over diverse cellular processes (Lu et al. 2021; Rolemberg Santana Travaglini Berti de Correia et al. 2024). CRLs specifically recognize substrates and promote their ubiquitination (Gong et al. 2021; Lu et al. 2019). They are composed of several modules, including the scaffold protein Cullin, RING protein RBX, substrate receptor, and adaptor protein ELOB/C (Horn-Ghetko et al. 2024; Petroski and Deshaies 2005). For instance, the CRL5 complex regulates protein degradation. It includes Cullin5, which uses SOCS box proteins to select substrates. These proteins link Elongin B/C to Cullin5, forming the functional CRL5 complex (Lumpkin et al. 2020; Yu et al. 2004). SPSB3 including a SOCS box domain. Regarding the underlying mechanism, the SPSB3-Cullin5 complex promotes the degradation of SNAIL (Liu et al. 2018), similarly, the CRL5-SPSB3 complex promotes the degradation of GAS (Xu et al. 2024). These studies suggest that SPSB3 likely operates as a constituent of the E3 ubiquitin ligase complex, and the present confirmed its role in myocardial I/R injury.

Our study showed that myocardial I/R injury induced increased SPSB3 expression, thereby promoting TUFM ubiquitination and degradation. Western blot and flow cytometry analyses indicated that double-knockdown group partially reversed the protective effect conferred by SPSB3-KD against H/R injury, although some inter-group comparisons showed numerical differences that did not reach statistical significance. These non-significant trends could stem from the constrained cohort size, and future studies with greater statistical power are required to confirm this potential effect. Nevertheless, these results offer fresh insights into the molecular mechanism of myocardial I/R injury, indicating that the SPSB3–TUFM–apoptosis signaling axis could be pivotal in the pathogenesis of this condition. TUFM encodes the mitochondrial translation elongation factor Tu. TUFM is multiubiquitinated by RNF185 (Chen et al. 2023) and deubiquitinated by USP5 (Xu et al. 2019). In another model of DOX-induced myocardial injury, TUFM expression decreased, and PANoptosis was inhibited (Bi et al. 2022). In addition, TUFM inhibits CASP8-mediated apoptosis in cancer (Choi et al. 2022). These findings matched our study's outcomes.

Our mutagenesis analysis of key ubiquitin linkage sites (K6, K11, K27, K29, K33, K48, and K63) revealed that K48 or K63 site mutation reduced TUFM protein levels. Concurrently, K6, K48, or K63 site mutation enhanced TUFM ubiquitination. Collectively, these findings suggest that SPSB3 promotes TUFM ubiquitination at lysine residues K6, K48, and K63. Furthermore, ubiquitination at the K48 and K63 sites appears to be involved in targeting TUFM for degradation, ultimately leading to decreased TUFM protein expression. Consistent with previous reports, K6-linked ubiquitin chains can mediate non-degradative functions, such as modulating chromatin recruitment and maintaining protein stability (Fonseca et al. 2024; Luo et al. 2024). In contrast, K48-linked ubiquitination is primarily associated with targeting substrates for degradation via the proteasome pathway (Kong et al. 2023), a pathway that aligns with the experimental findings of the present study. Beyond its roles in regulating signaling pathways (Lan et al. 2025), DNA damage repair (Xia et al. 2025), and immune responses (Wang et al. 2024), K63-linked ubiquitination can also contribute to the maintenance of protein stability (Yuan et al. 2025). Notably, emerging evidence indicates that this modification can mediate degradation via the proteasome pathway. For example, upregulation of HACE1 promotes K63-linked ubiquitination of ATG5, leading to its subsequent degradation (Gao et al. 2025). THAP11 promotes the ubiquitination of Nsp1β at both K48 and K63 sites, leading to its degradation (Chen et al. 2025). RNF186 targets HMGB1 for K48/K63-linked ubiquitination, thereby triggering its degradation via the proteasome pathway (Du et al. 2023). These studies demonstrate that K63-linked ubiquitination can operate independently or accompany with K48-linked ubiquitination to regulate the degradation of substrate protein. Based on the experimental data presented in this study, our results support the notion that ubiquitination at both the K48 and K63 residues participates in regulating the degradation of TUFM. However, further studies are needed to clarify whether K63-linked ubiquitination influences TUFM degradation via mitochondrial stress or involvement in other signaling pathways.

Our results provide novel insights into the molecular mechanisms underlying I/R injury, emphasizing the pivotal function of the SPSB3–TUFM–apoptosis signaling axis. By demonstrating that targeting this axis could alleviate apoptosis, we identified SPSB3 as a promising target for therapeutic intervention. While future development of SPSB3-TUFM blockers represents a rational therapeutic avenue, their clinical translation will depend on successful pharmacological optimization and stringent safety evaluation.

However, this study had several limitations. First, although the CRISPR/Cas9 screen systematically identified 68 candidate genes associated with H/R tolerance, we were unable to individually validate all hits due to constraints of time and resources. This limitation prevented us from obtaining a comprehensive functional confirmation of every potential target from the screen. Future studies may further validate the specific roles of the remaining candidate genes. Second, in vitro models may not reflect the in vivo cardiac environment, thereby limiting the generalizability of the findings. However, no studies have been conducted on human tissue samples or ventricular cardiac organoids derived from hiPSCs. We hope that future research will validate these results with more advanced evidence. Third, reasons for the increased expression of SPSB3 during myocardial I/R injury require further investigation. While specific time points were chosen for modeling in this investigation, the long-term consequences remained unevaluated. Extending the observation period may provide valuable clinical insights.

Materials and methods

Cell culture and establishment of H/R model

AC16 cells (SCC109, Millipore, USA) were cultured in DMEM/F-12 (PM150310, Pricella, China). The culture medium was fortified with 10% fetal bovine serum (164210; Pricella, China) and 1% penicillin–streptomycin solution (PB180120; Pricella, China). To develop the H/R model, cardiomyocytes were incubated in glucose-free, serum-free DMEM/F-12 (PM150322, Pricella, China) with 5% CO2 and 95% N2 for 12 h and transferred to fresh medium with glucose and serum and normoxic conditions for 6 h.

HEK293T cells (KGG3101-1, KeyGEN, China) were cultured in DMEM High Glucose (KGL1211-500, KeyGEN, China).

Lentiviral transfection

A cellular suspension was formulated at a density of 3–5 × 10⁴ cells/mL, from which 1.5 mL aliquots were distributed into individual wells of six-well plates and supplemented with 500 μL of growth medium before vigorous mixing. The SPSB3 knockdown viral stock (Syngentech, China) was diluted to 1 × 10⁸ TU/mL for transfection; 20 μL of the diluted viral mixture and polybrene (5 μg/mL; HB-PB-500, HANBIO, China) were added to each well. The medium was replaced after 12 h. Transfection efficiency was evaluated after 72 h, followed by selection with puromycin (5 μg/mL; HB-PU-500, HANBIO, China). For TUFM knockdown viral stock (HANBIO, China), the same procedure was followed, using blasticidin (4 μg/mL; HB-BSD-500, HANBIO, China) for selection. Supplementary File 3 presents basic information on the lentivirus vector.

Plasmid transfection

Using human cDNA as a template, the CDS sequences of SPSB3 and TUFM genes were identified through PCR‐based cloning and constructed into commercial mammalian overexpression vectors pcDNA3–Flag and pcDNA3–His via recombination. PcDNA3–HA–ubiquitin and TUFM gene mutant plasmids (pcDNA3–His–TUFM [WT], pcDNA3–His–TUFM [K293R]) were similarly designed and constructed. All plasmids were validated by sequencing. Supplementary File 3 shows the base sequences of the WT and K293R mutant plasmids.

CRISPR/Cas9 library screening of cells resistant to H/R injury

The "Human GeCKO v2 library" (1,000,000,048, Addgene, USA) was packaged into lentivirus and used to infect AC16 cardiomyocytes at a low MOI in order to establishing a cell pool in which each cell carries a single-gene knockout. Subsequently, the cells were screened using puromycin, and only AC16 cells that contained the sgRNA library (a mixture of nearly 30,000 single-gene knockout cells from the entire genome) remained. Using an in vitro H/R model as a lethal condition, we believe that surviving cardiomyocytes could tolerate H/R damage. Surviving single-gene knockout cell complexes were collected and sequenced. The screening results were obtained using this high-throughput strategy of knocking out individual genes.

Animals and myocardial ischemia/reperfusion model

Male C57BL/6 J mice (6–8 w; weight 25 ± 2 g) were obtained from Beijing HuafuKang Biological Technology Co., Ltd. All animals were pre-adapted for 4 weeks after arriving at the facility before the experiments were performed. Mice were anesthetized with isoflurane. In the I/R group, transient ligation of the left anterior descending coronary artery was performed for 45 min, subsequently followed by a 24 h reperfusion interval. Conversely, in the sham group, sutures were passed without ligation.

Adeno-associated virus infection

Adenoviral vectors were administered via tail vein injection according to the following protocol. The tail was disinfected with 70% alcohol, and the virus was injected slowly into the tail vein (200 μL/mouse) and held for 5–10 s after completion to prevent virus reflux (AAV9—cTNT—miR30—m—Spsb3 – ZsGreen, target sequences: GCTCCATTATTGGTGTGCACTTGGA, titer of Virus: 2.8 × 10^11 vg/mL; Control for AAV9-cTNT-ZsGreen: titer of Virus: 2.5 × 10^11 vg/mL, HANBIO, China). I/R surgery was performed 4 weeks later.

Echocardiography

Echocardiography was performed using a VINNO 6VET/6LAB high-resolution ultrasound system. For echocardiography, mice were anesthetized using isoflurane and positioned supinely. The parasternal long-axis view was acquired by positioning the ultrasound transducer at the left sternal border between the third and fourth intercostal spaces with the probe oriented toward the 11 o'clock position (relative to the cranial direction). To evaluate contractile function of the left ventricle, we measured the EF, LVFS, LVIDd, and LVIDs.

TTC staining

The infarction area was estimated by staining with a 2% TTC solution (G3005; Solarbio, China). Upon anesthesia, murine hearts were excised and thoroughly washed with phosphate-buffered saline (PBS). Afterward, the heart was frozen at −20℃ for 30 min before being sliced into 1-mm Sects. (4–6 sections per heart). The tissue sections underwent incubation in TTC solution at 37 °C for a duration of 30 min under continuous agitation. Digital images of the stained tissue were acquired using a standard camera, followed by quantitative evaluation with Image-Pro Plus software.

H&E staining

Fixation in 4% PFA for 48 h, mouse heart tissues were dehydrated through a graded ethanol series, embedded in paraffin, and sectioned for hematoxylin–eosin (H&E) staining. Tissue sections were dewaxed; stained with hematoxylin and eosin (C0105, Beyotime, China), dehydrated with ethanol, immersed in xylene for transparency, sealed with neutral gum.

Measurement of CKMB, cTnI and LDH

Blood samples were collected 24 h after reperfusion. Thereafter, the samples were incubated at room temperature for 30 min and then centrifuged to separate the plasma. CK-MB (JL12422; Jianglai, China) and cTnI (JL11280; Jianglai, China) levels were measured using ELISA kits, whereas LDH levels were assessed using a lactate dehydrogenase activity detection kit (G0685; Solarbio, China).

Immunohistochemistry (IHC)

Mouse heart tissue was sectioned into 4 μm slices, dewaxed, and placed in citrate‒sodium citrate buffer (pH = 6.0). Antigen retrieval was conducted via microwave irradiation at maximum power for 7.5 min, followed by natural cooling. The experiment was performed using an immunohistochemistry kit (KIT9710; MXB, China) with an anti-SPSB3 antibody (PA5-31,502; Invitrogen, USA) diluted 1:150 overnight. A DAB chromogenic working solution (ZLI-9018, ZSGB-BIO, China) was prepared, and all sections were developed for the same amount of time. Following a 5-min hematoxylin counterstaining of the nuclei, the specimens underwent dehydration and were subsequently mounted using neutral resin.

Frozen slice preparation and photography

Cardiac specimens were immersed in 4% PFA solution and maintained at 4 °C for a duration of 48–72 h after 24 h of reperfusion, dehydrated in 30% sucrose solution until it sank, inverted, and dehydrated again. Thereafter, heart was frozen in optimal cutting temperature (OCT) [4583, SAKURA, USA] compound, sliced into 10 μm sections, incubated at 37 °C for 20 min, and examined under a fluorescence microscope for signs of viral infection.

TUNEL staining

Mouse heart tissue was cut into 4-μm thick paraffin sections, dewaxed, and subjected to antigen retrieval. Following the manufacturer's protocol, TUNEL assays were conducted using KeyGEN BioTECH reagents, wherein cells exhibiting TUNEL positivity were visualized by emitting either green (KGA1407, KeyGEN, China) or red fluorescence (KGA1405, KeyGEN, China). Cellular nuclei were labeled using DAPI (C0065, Solarbio, China). In vitro, cells were treated under different conditions for 12 h. Subsequently, cells were fixed using 4% PFA for 30 min at ambient temperature, followed by permeabilization with 1% Triton X-100 for 15 min. Then samples were treated with 0.3% H2O2-methanol for 15 min prior to undergoing TUNEL and DAPI staining. Final visualization was conducted via fluorescence microscopy.

Cell viability and LDH release analysis

The viability of AC16 cells was evaluated using Cell Counting Kit-8 (CCK-8) [KTA1020, Abbkine, China]. Cell damage was assessed using an LDH assay kit (KGA7403; KeyGEN, China).

Immunofluorescence (IF)

AC16 cardiomyocytes were plated at 50,000 cells per well. Different treatment conditions were applied to the cells after 12 h, fixed using 4% PFA, and permeabilized with 0.2% Triton X-100. To minimize non-specific interactions, samples were blocked with 10% goat serum at 37 °C for 30 min. For single staining, an anti-SPSB3 antibody (#PA5-31,502, Invitrogen, USA) was diluted at 1:200 and incubated overnight. In dual-staining experiments, samples were concurrently exposed to a cocktail of anti-SPSB3 and anti-TUFM antibodies (67,802–1-Ig, Proteintech, China), each diluted 1:150, followed by overnight incubation, followed by the addition of a fluorescent secondary antibody (KGC6215-0.1, Keygen, China; KGC6214-0.1, Keygen, China) incubated at 37 °C for 1 h. Prior to image acquisition, nuclear counterstaining was performed using DAPI.

Flow cytometry

Cellular apoptosis was quantified employing an Annexin V-FITC/PI assay kit (556,547, BD, USA), while mitochondrial membrane potential was assessed utilizing a JC-1 staining kit (C2006, Beyotime, China).

Transmission electron microscope

Cell specimens were immobilized via exposure to 2.5% glutaraldehyde followed by 1% osmium tetroxide. The materials then underwent rinsing with PBS and progressive dehydration through a graded series of ethanol and acetone solutions. After infiltration with acetone and embedding agent 812 (02660-AB, SPI, USA) at 60 °C, the samples were polymerized for 48 h. Ultrathin sections were then cut and subjected with a double-staining solution comprising 2% uranyl acetate in saturated ethanol and 2.6% lead citrate, dried overnight, and observed via transmission electron microscopy.

Western blot analysis

Total protein was isolated from the tissue and cells using RIPA lysis buffer (ZLI-9018, Seven, China) supplemented with PMSF (HY-B0496, MCE, USA). Protein lysates were resolved via SDS-PAGE and subsequently transferred to PVDF membranes (10,600,023; GE Amersham, USA). Following a 2 h blocking step with 5% non-fat milk, the membranes underwent washing with TBST. Incubation with specific primary antibodies was performed overnight at 4 °C, after which the membranes were exposed to corresponding secondary antibodies for 90 min at ambient temperature (dilution ratios and information for each antibody are provided in Supplementary File 4). Immunoreactivity was detected using an enhanced chemiluminescence reagent (BMU102, Abbkine, China) with an Amersham Imager 600/680.

CO-IP and mass spectrometry

Cells were incubated on ice for 10 min in NP-40 lysis buffer (BL653A; Biosharp, China) supplemented with a protease inhibitor cocktail (HY-K0010; MCE, USA). Thereafter, they were centrifuged at 14,000 rpm for 10 min at 4 °C. IP was performed according to the manufacturer’s instructions (10007D, Invitrogen, USA). IP used 4 μg of primary antibody mixed with Ab Binding & Washing Buffer, incubated with protein A/G beads for 30 min. Following a wash step, the protein extract was thoroughly mixed with the beads for an additional 30 min. After subsequent washing, elution and loading buffers were added, and the product was denatured at 70 °C for 5 min. The proteins were subjected to electrophoresis, and the gel was rinsed and stained with Coomassie Brilliant Blue for 30 min, washed, and analyzed using mass spectrometry.

Antibody information

Detailed specifications for all antibodies used in this study, including the target antigen, catalog number, RRID (where applicable), manufacturer, and working dilution for each application, are provided in Supplementary File 4.

Glutathione S-transferase (GST) pulldown

His-TUFM (47-455aa)-Flag and GST-SPSB3-His fusion proteins were expressed in Escherichia coli BL21 (DE3) constructed as vectors pET-30a or PGEX-4 T-1. The washed GST-purified magnetic beads were mixed with the binding solution and washed twice to remove any impurities. Protein samples containing the GST tag were added to the treated magnetic beads, properly mixed, and incubated for 30 min under ambient conditions. Following two wash cycles with gentle pipetting, samples containing the target proteins were added to the protein-magnetic bead complexes described above, mixed well, and maintained at room temperature for an additional 30-min period. Subsequent to two further washing steps, the eluates were collected, and western blotting was performed to verify the results.

Real-time qPCR

MRNA was extracted from mouse left ventricular tissue and cardiomyocytes using RNAiso reagent (9108, TAKARA, Japan). Complementary DNA (cDNA) was synthesized by reverse transcription, followed by quantitative PCR (qPCR) (using kits RR047A and RR820A from TAKARA, Japan, respectively).

Quantitative proteomics of ubiquitination modification with 4D-Fast DIA

The con and H/R cell collection required more than 70 μL precipitation (2.0 × 10⁷ cells). Protein samples were enzymolyzed into a peptide mixture, the components of the enzymolyzed peptide mixture were separated USING liquid chromatography, the modified peptide segments were enriched with high-quality ubiquitination-modified antibodies and biomaterials, and the specimens were transferred to liquid chromatography-tandem mass spectrometry for analytical profiling and quantitative assessment.

Drug treatment of CHX and MG132

The cells were treated with 100 μg/mL of cycloheximide (HT-12320, MCE, USA) and 10 μM of MG132 (S2619, Selleck, USA) and then collected after coincubation for approximately 2, 4, and 6 h.

Molecular docking

The UniProt website was used to search for the UniProt ID of the proteins, followed by AlphaFold2 for structure prediction and file downloads. PyMOL was used to visualize and modify the PDB files, which were uploaded to the HDOCK website for molecular docking. The structure with the highest docking score was selected, the interaction forces were analyzed in detail, and the results were depicted in Adobe Illustrator.

Data source and processing

Public datasets and databases were used in this study. The sources and specific analytical procedures are described below.

1) GSE61592 dataset.

The dataset GSE61592 (www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE61592) was obtained from the public available Gene Expression Omnibus database. The raw data (CEL files) were downloaded for subsequent analysis. The initial quality assessment via UMAP demonstrated a clear separation between the experimental groups (Supplementary Fig. 3).

The GSE61592 was used to analyze differentially expressed genes (DEGs) via GEO2R with |fold change (FC)|≥ 2 and p < 0.05; the largest FC probe was retained for multiple probes, and results with multiple genes were excluded. A volcano plot was generated using the ggplot2 package in R Studio (version R-4.1.1) with FC = 2 and p = 0.05. Enrichment analysis utilized enrichGO from clusterProfiler with org.Mm.eg.db, applying p and q cut-offs of 0.05 and 0.2, respectively, and BH correction, showing the top 15 entries. EnrichKEGG was also performed using the BH method and visualized with a bar plot. Venn diagrams were created using R Studio to display the overlapping DEGs among the groups. Bioinformatics analysis of the results of the ubiquitination modification proteomics was performed using the PTM Bio-Cloud platform.

2) MitoCarta 3.0 Database.

The MitoCarta 3.0 database (www.broadinstitute.org/mitocarta) is a mitochondrial protein database constructed using integrated high-throughput proteomics and bioinformatics. It was established by systematically screening and validating mitochondrial proteins through large-scale protein identification across multiple cell lines and tissues, combined with the curation of existing mitochondrial gene databases. Proteins in the MitoCarta3.0 database are stringently curated through a multi‑evidence quality control pipeline that integrates data from mass spectrometry, GFP‑localization experiments, or established functional databases (Rath et al. 2021).

A catalog of human mitochondrial genes was obtained from the MitoCarta 3.0(file: Human.MitoCarta3.0. xls).This comprehensive resource provides detailed annotations, including submitochondrial localization and pathway associations, for 1,136 human mitochondrial genes, along with integrated mitochondrial localization scores for approximately 20,000 human protein-coding genes. From this dataset, genes specifically localized to the cardiac tissue were selected based on the column heart_total_peak_intensity_log10. This filtering yielded a subset of 539 heart-enriched mitochondrial genes that were subsequently used for downstream analyses.

Statistical analysis

Experimental data were analyzed using GraphPad Prism 10, with findings reported as mean values ± standard deviations (SD). Inter-group comparisons between two cohorts were performed via Student's t-test, while variations across multiple groups were evaluated using ANOVA. Statistical significance was defined as a p-value below 0.05; conversely, p-values of 0.05 or higher were considered non-significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ contributions

Y.Y., Y.P. and X.L. conceived and designed the study. Y. Y. contributed to most of the cellular and molecular biology experiments. Y.X. and Y.T. provided technical and material support. Y.Y. and X.W. collected the animal specimens. Y.Y. compiled public data. Y.Y., Y.P., and X.L. analyzed and interpreted the data. Y.Y. wrote the manuscript. Y.T. contributed to the performance of additional experiments and guided manuscript revision and responses to the reviewers. Y.P., X.L. and Y.T. supervised the study. All authors reviewed and approved the final version of the manuscript.

Funding

This study was funded by a grant from the Science and Technology Joint Project of Liaoning Province (grant no. 2025-MSLH-780).

Data availability

Data available in “Mendeley data” at https://data.mendeley.com/preview/gkhtb4s3nd?a=10a54ab0-77fb-48a6-8a55-6e21ae5c87d1 (https://data.mendeley.com/preview/t44vhjw4zf?a=5791baae-75b5-4b82-b9c3-8261764bea7b).

Declarations

Ethics approval

The animal study protocols were approved by the Ethics Committee of Shengjing Hospital affiliated with China Medical University (Approval No. 2022PS1153K) in accordance with the National Institutes of Health Guidelines for Laboratory Animal Care and Use.

Clinical trial number

Not applicable.

Consent for publication

All authors read and approved the manuscript.

Competing interests

The authors declare no competing interests.