How PRKN helps break down METTL3 to promote cell aging

To investigate whether m6A methyltransferases drive senescence‐associated epigenetic remodeling, we compared proliferating and replicative senescent BJ fibroblasts and quantified global m6A alongside core methyltransferases. Cellular senescence was confirmed by SA‐β‐gal staining (Figure 1A,B) and increased expression of the senescence markers p16 (CDKN2A‐encoded protein) and p21 (CDKN1A‐encoded protein) (Figure 1C). Quantitative Slot Blot analysis revealed a significant reduction in global m6A levels in senescent cells compared to proliferating controls (Figure 1D,E), establishing senescence‐associated m6A hypomethylation.

Given this hypomethylation phenotype, we profiled core m6A methyltransferases. Western Blot analysis showed decreased protein abundance of METTL3 and METTL14 in replicative senescent cells (Figure 1F), whereas RT‐qPCR indicated unchanged transcript levels (Figure 1G). Notably, WTAP protein remained stable (Figure 1F), consistent with its unaltered mRNA expression (Figure 1G).

To determine whether this regulatory paradigm extends beyond replicative senescence, we established an irradiation‐induced senescence model (10 Gy X‐ray). Consistent with replicative senescence, irradiated cells exhibited comparable m6A reduction (Figure 1H–L) and METTL3 protein depletion without transcriptional alterations (Figure 1M,N). Critically, WTAP protein and mRNA levels remained unaltered in irradiated cells (Figure 1M,N), while METTL14 protein showed no significant reduction in this context (Figure 1M,N). These convergent dynamics—reduced METTL3 protein without transcriptional changes in both senescence models—implicate post‐translational mechanisms in its regulation, positioning METTL3 as a conserved senescence‐associated factor with post‐translational vulnerability.

Given the conserved discordance between METTL3 protein depletion and stable mRNA levels across senescence models (Figure 1F,G,M,N), we hypothesized accelerated protein turnover during senescence. To investigate this, cycloheximide (CHX) chase assays in replicative senescent cells revealed accelerated METTL3 degradation compared to proliferating controls (Figure 2A).

To further explore the specific post‐translational mechanisms responsible for METTL3 degradation, we assessed its stability in the presence of various inhibitors. Our data indicated that METTL3 degradation was inhibited by the proteasome inhibitor MG132, whereas no significant effects were observed with the lysosome inhibitor chloroquine (CQ) or the autophagy inhibitor 3‐MA (Figure 2B). Since ubiquitination is a key signal for proteasomal degradation, we next examined whether the ubiquitination of METTL3 was altered in senescent cells. The results revealed a significant increase in METTL3 ubiquitination in senescent cells (Figure 2C).

To identify the responsible E3 ligase, we performed UbiBrowser 2.0 analysis (Wang et al. 2022), which revealed seven candidates: BARD1, KDM1B, MID1, BMI1, PRKN (Parkin), SALL1 and SYVN1 (Figure 2D). To determine which candidate E3 ligases mediate METTL3 ubiquitination, qPCR analysis of these seven candidate E3 ligases showed that only PRKN and MID1 exhibited upregulated mRNA levels in senescent cells (Figure 2E). Subsequent Western blot analysis demonstrated exclusively PRKN protein elevation under senescence conditions (Figure 2F). Functional validation via siRNA‐mediated knockdown revealed that PRKN ablation increased METTL3 protein levels (Figure 2G), whereas MID1 silencing had no effect (Figure 2H). This differential regulation establishes PRKN as the principal E3 ligase governing METTL3 stability during cellular senescence.

To elucidate the molecular mechanisms by which PRKN mediates METTL3 degradation, we first investigated their interaction. Immunofluorescence (IF) analysis in BJ fibroblasts revealed evident PRKN‐METTL3 colocalization (Figure 3A). Co‐immunoprecipitation (co‐IP) in HEK293T cells further confirmed a robust physical interaction between PRKN and METTL3 (Figure 3B,C). To establish direct binding, bacterially expressed and purified His‐METTL3 and Myc‐PRKN proteins were purified and incubated in vitro. This reconstitution assay demonstrated that METTL3 directly binds PRKN (Figure 3D).

Additionally, we analyzed the effect of PRKN on METTL3 ubiquitination and found that PRKN overexpression significantly enhanced the poly‐ubiquitination of METTL3 (Figure 3E). To assess whether PRKN influences the stability of METTL3, a cycloheximide (CHX) chase assay was performed. The results revealed that the depletion of PRKN delayed the degradation of METTL3 (Figure 3F). We further identified the specific type of ubiquitin chain added by PRKN using wild‐type (WT) ubiquitin and various mutants (K6‐, K11‐, K27‐, K29‐, K33‐, K48‐, and K63‐only) (Figure 3G). The K48‐only mutant, like WT ubiquitin, significantly increased PRKN‐induced METTL3 ubiquitination (Figure 3H). Moreover, overexpression of PRKN enhanced the K48‐linked ubiquitination of METTL3 in HEK293T cells (Figure 3I). These results demonstrate that PRKN catalyzes K48‐linked polyubiquitination of METTL3, targeting it for proteasomal degradation.

To identify the lysine residues on METTL3 targeted for K48‐linked ubiquitination, we purified Flag‐tagged METTL3 and performed multidimensional liquid chromatography coupled with tandem mass spectrometry (2D‐LC–MS/MS) analysis (Figure 4A). Combined with predictive modeling, this approach identified two candidate ubiquitination sites: K164 and K459 (Figure 4B). To validate these sites, we generated Flag‐METTL3 mutants containing single lysine‐to‐arginine substitutions (K164R and K459R) and assessed their ubiquitination status. Ubiquitination assays revealed that the K164R mutant exhibited significantly reduced K48‐linked ubiquitination, whereas the K459R mutation showed no effect (Figure 4C).

Critically, PRKN failed to mediate K48‐linked ubiquitination of the METTL3 K164R mutant (Figure 4D), establishing K164 as essential for PRKN‐dependent ubiquitination. To determine the functional consequence of this modification, we compared protein stability between wild‐type (WT) METTL3 and the K164R mutant. Unlike WT METTL3, the K164R mutant resisted PRKN‐mediated degradation (Figure 4E). Consistently, cycloheximide (CHX) chase assays demonstrated delayed degradation kinetics of the K164R mutant compared to WT METTL3 (Figure 4F). Comparative analysis of the primary amino acid sequences of METTL3 homologs across species revealed a high conservation of K164 among mammals (Figure 4G). These data establish that PRKN targets K164 for K48‐linked ubiquitination to drive proteasomal degradation of METTL3.

To evaluate the functional impact of METTL3 deficiency on cellular senescence, we first performed Western Blot analysis to detect the protein levels of the senescence markers p16 and p21 following METTL3 depletion. The results demonstrated that loss of METTL3 led to a significant upregulation of both p16 and p21 (Figure 5A). Subsequently, SA‐β‐gal staining assays were performed on both replicative and stress‐induced senescent cells, revealing that METTL3 deficiency resulted in a marked increase in the number of SA‐β‐gal positive cells (Figure 5B–E).

Given that telomere attrition triggers senescence and prior evidence implicating METTL3‐mediated m6A modification in telomere homeostasis (Chen et al. 2022), we hypothesized that METTL3 may regulate cellular senescence through its effects on telomere homeostasis. Immunofluorescence‐Fluorescence in situ hybridization (IF‐FISH) analysis demonstrated that METTL3 knockdown led to a significant increase in 53BP1 foci, with most of these foci colocalizing with telomeres in fibroblasts (Figure 5F–H). Furthermore, we observed that METTL3 deletion led to telomere uncapping but had no effect on telomere length (Figure 5I–K). These findings suggest that METTL3 plays a critical role in maintaining telomere structure and stability but is not directly involved in telomere lengthening.

Given the established association between shelterin complex dysfunction and telomere‐driven genomic instability (Raghuram and Mishra 2014), we performed systematic analysis of its core constituents: (i) TRF1 and TRF2, the double‐stranded DNA‐binding proteins that organize telomere loop architecture; (ii) POT1, the single‐stranded telomeric DNA‐binding protein critical for G‐overhang protection. RT‐qPCR and Western Blot analysis revealed a selective depletion of TRF2 and POT1 at the protein level in METTL3‐deficient cells, while their mRNA levels remained unchanged (Figure 5L,M). This observation implies that METTL3 likely regulates the translational efficiency or post‐translational stability of these shelterin components, rather than influencing their transcriptional output.

Building on our finding that PRKN promotes proteasomal degradation of METTL3 (Figure 4), we investigated its role in cellular senescence by knocking down PRKN in pre‐senescent cells. Western Blot analysis revealed that PRKN deficiency restored METTL3 protein levels, subsequently rescuing expression of shelterin components TRF2 and POT1 while reducing senescent markers p16 and p21 expression (Figure 6A). Consistent with METTL3 functional recovery, global m6A levels reverted to those observed in proliferating cells (Figure 6B,C).

To determine whether PRKN knockdown alleviates METTL3 loss‐induced telomere dysfunction and senescence, we performed IF‐FISH and SA‐β‐gal staining assays in PRKN‐deficient cells. As expected, PRKN deficiency mitigated telomere dysfunction, evidenced by a decrease in telomeric 53BP1 foci formation (Figure 6D–F), and subsequently delayed the onset of cellular senescence, as demonstrated by a reduction in the proportion of SA‐β‐gal‐positive cells (Figure 6G,H). These results suggest that PRKN may regulate cellular senescence via the METTL3‐m6A‐telomere axis.

To further confirm that PRKN regulates cellular senescence by mediating proteasomal degradation of METTL3, we analyzed BJ cells overexpressing either wild‐type METTL3 (METTL3‐WT) or the ubiquitination‐resistant K164R mutant (METTL3‐K164R) following PRKN overexpression. Given that METTL3 mRNA levels remained unchanged (Figure 6I), PRKN overexpression specifically reduced METTL3‐WT but not METTL3‐K164R protein levels (Figure 6J). Concurrently, PRKN overexpression diminished TRF2/POT1 expression, while elevating p21 levels—effects rescued by METTL3‐K164R but not METTL3‐WT overexpression (Figure 6J). Consistent with this, PRKN overexpression promoted both total and telomeric 53BP1 foci formation, which were significantly attenuated in METTL3‐K164R‐ but not METTL3‐WT‐expressing cells (Figure 6K–M). Correspondingly, PRKN overexpression significantly increased the proportion of SA‐β‐gal‐positive cells in METTL3‐WT‐expressing cells, while METTL3‐K164R mutant cells exhibited resistance to this senescence acceleration (Figure 6N,O). Collectively, these findings demonstrate that PRKN‐mediated proteasomal degradation of METTL3 drives m6A‐dependent cellular senescence regulation.

Building on established evidence of accelerated cellular senescence in pulmonary systems—where senescent cells accumulate disproportionately during aging (Munoz‐Espin and Serrano 2014; Ovadya et al. 2018; Wang et al. 2009) and directly contribute to idiopathic pulmonary fibrosis (IPF) pathogenesis (Alvarez et al. 2017; Schafer et al. 2017; Zhang et al. 2024). To investigate PRKN's role in physiological aging and age‐related diseases, we quantified PRKN protein levels in lung tissues from young controls, aged mice, and mice with IPF. PRKN expression was significant upregulation in lung tissues from both aged mice (Figure S1A–D) and IPF models (Figure S1E–H). This convergent elevation of PRKN expression in two distinct but senescence‐exacerbated conditions reinforce its potential mechanistic role in pulmonary aging and fibrotic disease progression.

BJ, and HEK293T cells were obtained from American Type Culture Collection (ATCC) (Manassas, VA). Cells were cultured at 37°C in a humidified atmosphere containing 5% CO2. Cells were grown in Dulbecco's Modified Eagle Medium (DMEM) (Hyclone) with 10% fetal bovine serum (FBS) (Gibico) and 100 U/mL penicillin/streptomycin. All cells were confirmed to be free of for mycoplasma contamination.

The wild‐type METTL3 gene was amplified from HEK293T mRNA and cloned into pHAGE‐Flag vector. METTL3‐K164R and METTL3‐K459R were generated by site‐directed mutagenesis, introducing a lysine (K) to arginine (R) substitution at positions 164 and 459, respectively, based on the wild‐type METTL3 sequence. The wild‐type PRKN gene was synthesized from IGEbiology and cloned into pLVX‐AcGFP1‐N1 vector. For transfection, BJ cells were transfected using Lipofectamine 3000 (Invitrogen), and HEK293T cells were transfected using PEI (Yeasen) according to the manufacturer's recommended protocol. Chemically synthesized 21‐nucleotide siRNA duplexes were purchased from GenePharma and transfected using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer's instructions. The RNA oligonucleotides used in this study are as below: siMETTL3‐1: 5′‐CUGCAAGUAUGUUCACUAUGAdTdT‐3′; siMETTL3‐2: 5′‐GCACUUGGAUCUACGGAAUCCdTdT‐3′; siPRKN‐1: 5′‐GCUUAGACUGUUUCCACUUAUdTdT‐3′; siPRKN‐2: 5′‐GCCUUCUGCCGGGAAUGUAAAdTdT‐3′; siMIDI: 5′‐GUCGUUAGUCUGUGUAAUUdTdT‐3′.

1 μg RNA was diluted to a concentration of 10 ng/μL in a buffer containing 6 × SSC and 7.4% methanal. The samples were heated at 65°C for 15 min to denature secondary structures, followed by immediate cooling on ice. Subsequently, RNA was loaded onto Hybond N+ membranes. After air drying, the membranes were crosslinked using a UV stratalinker at an energy setting of 1200 J/cm², applied six times. Membranes were then blocked for 1 h at room temperature in a solution of 5% non‐fat milk dissolved in PBST (0.1% Tween‐20). Following blocking, the membranes were incubated overnight at 4°C with anti‐6 mA antibody (Synaptic Systems, 1: 1000). On the second day, the membranes were washed five times, each for 5 min, with PBST. They were then incubated with HRP‐conjugated goat anti‐rabbit IgG secondary antibody (1:2000, 7074S, Cell Signaling Technology) for 1 h at room temperature. The washing step with PBST was repeated five times for 5 min each. Finally, the membranes were treated with enhanced chemiluminescence (ECL) reagent and developed for visualization of the antibody signals.

For immunoprecipitation, Cells were lysed with RIPA buffer (50 mM Tris pH 7.4, 0.25% sodium deoxycholic acid, 1% Triton, 150 mM NaCl, 1 mM EDTA), supplemented with protease inhibitor. whole‐cell lysates were incubated overnight with the appropriate antibodies, followed by incubation with Protein A/G beads (sc‐2003, SANTA CRUZ), or anti‐Flag beads (Sigma). The beads were then washed five times with RIPA buffer, and the bound immunoprecipitants were eluted with 2 × SDS loading buffer and resolved by SDS‐PAGE. The following antibodies were used for immunoblotting: anti‐METTL3 (1:2000, ab195352, Abcam), anti‐METTL14 (1:3000, 26158‐1‐AP, Proteintech), anti‐WTAP(1:5000, 10200‐1‐AP, Proteintech), anti‐PRKN (1:1000, 14060‐1AP), anti‐p16 (1:1000, 10883‐1‐AP, Proteintech), anti‐p21 (1:1000, 2947S, CST), anti‐TRF1 (1:500, T1948, Sigma‐Aldrich), anti‐TRF2 (1:1000 dilution, 66893‐1‐IG, Proteintech), anti‐POT1 (1:500 dilution, 10581‐1‐AP, Proteintech), HRP‐conjugated goat anti‐rabbit (1:2000, 7074S, Cell Signaling Technology) or horse anti‐mouse (1:2000, 7076S, Cell Signaling Technology).

Recombinant His‐METTL3 and Myc‐PRKN proteins (purchased from Targetmol) were incubated with Myc‐Beads in RIPA buffer (1% NP‐40, 0.25% sodium deoxycholate, 50 mmol/L Tris‐HCl (pH 7.4) and 150 mmol/L NaCl) at 4°C for 1 h. Myc peptides were used as a control. After incubation, beads were washed four times with RIPA buffer. The bound proteins were analyzed by Western Blot using anti‐METTL3 (1:2000, ab195352, Abcam) and anti‐PRKN (1:1000, 14,060‐1AP).

BJ cells grown on coverslips were transfected with constructs for METTL3/PRKN knockdown or overexpression. After 72 h, IF‐FISH was performed as previously described (Mao et al. 2016), with the following modifications: Cells were incubated with a primary antibody against 53BP1(1:100, NB100‐304, Novus Biologicals) for 1.5 h at room temperature; followed by incubation with Alexa Fluor 488/633 conjugated anti‐rabbit IgG sencondary antibody for 1h at room temperature. After antibody incubation, telomeres were denatured at 85°C and hybridized with Cy3‐labeled PNA telomeric G probes overnight at 37°C. Following hybridization, cells were washed counterstained with DAPI and mounted. Imaging was performed using a Zeiss microscope.

BJ cells were transfected with METTL3 or control siRNA for 2 weeks, following treated with 0.5 μg/mL Nocodazole for 15 h to enrich for metaphase cells. These cells were collected and then hypotonic treatment using 75 mM KCl solution for 30 min at 37°C, followed by fixation with methanol: acetic acid (3:1) twice. Fixed cells were spread onto clean cold slides and digested with 0.5 mg/mL RNase A for 10 min at 37°C. Telomeres were denatured at 85°C and hybridized with FITC‐labeled PNA telomeric C probes. Chromosome were counterstained with DAPI. Imaging was perform using a Nikon microscope.

Total RNA was extracted from cells using the Trizol reagent (Invitrogen) according to the manufacturer's instructions. For RT‐qPCR analysis, cDNA was synthesized from the extracted RNA using HiScript III RT SuperMix for qPCR (+gDNA wiper) (R323‐01, Vazyme) following the manufacturer's protocol. The synthesized cDNA was then analyzed by qPCR using the 2 × RealStar Green Fast Mixture (A311‐10, GenStar). The PCR primer sequences used for the amplification are provided in Table 1.

BJ cells were cultured in 10 cm dishes, knocking down with siRNA of METTL3 or NC for 1 month. Total DNA extracted from the cells using the HiPure Blood DNA Mini Kit (D3111, Magen) according to the manufacturer's instructions. For the quantitative PCR analysis, 50 ng of the extracted DNA was used. then analyze was performed using the 2 × RealStar Green Fast Mixture (A311‐10, GenStar) with telomere‐specific primers. Albumin, a single‐copy gene, was used as the reference gene. The reaction concentration of DNA and primers were maintained at 2 ng/μL and 10 μM respectively. And the special qPCR procedures refer to published article (Hsieh et al. 2016). The specific PCR primer sequences were provided in the Table 2.

Cells were cultured in 6‐well plates and subjected to knocking down or overexpressing METTL3/PRKN for 1 week. For inducing stress‐induced premature senescence, cells were exposed to X‐ray irradiation at 10 Gy. Immediately after irradiation, the medium was changed, and cells were cultured for an additional week. Subsequently, SA‐β‐gal staining was performed using the senescence β‐Galactosidase Staining Kit (Beyotime) according to the manufacturer's instructions. Briefly, cultured cells were washed with PBS and fixed at room temperature for 15 min. Fixed cells were stained with SA‐β‐gal staining solution at 37°C overnight, Images were captured using a Zeiss microscope.

Lung tissues were fixed in 4% paraformaldehyde at room temperature overnight, then transferred to 70% ethanol and subsequently embedded in paraffin. Sections of 4 μm thickness were prepared on slides, dewaxed in xylene, and rehydrated (100%, 95% and 70%) followed by PBS buffer. Histopathological analysis was performed on these paraffin‐embedded lung sections stained using Masson's trichrome (MXB Biotechnologies) according to using standard procedures.

For immunohistochemistry, sections were blocked with 5% goat serum in PBS buffer and then incubated overnight at 4°C with primary antibodies. The primary antibodies used included anti‐p21 (1:1000, ab188224, Abcam) and anti‐PRKN (1:100, 14,060‐1AP). After incubation, sections were washed three times with PBS and then incubated with an HRP‐conjugated anti‐rabbit IgG secondary antibody (KPL Inc). Following secondary antibody application, sections were washed again three times with PBS. The detection was carried out using DAB (diaminobenzidine) for enzyme activity visualization and hematoxylin for DNA staining, following established protocols.

The Immunohistochemistry samples are acquired from Canfeng zhang et al. with permission from EMBO Journal (Zhang et al. 2024).

Data are presented as mean ± standard error of mean (SEM). All analyses were performed using GraphPad Prism (Version 8.0.1). Normality was assessed per experimental group via Shapiro–Wilk tests (α = 0.05), with variance homogeneity evaluated by Brown‐Forsythe tests (α = 0.05). For two‐group comparisons, two‐tailed unpaired Student's t‐tests were applied when both assumptions were satisfied; Welch's t‐tests were utilized when normality held but variances differed significantly; Mann–Whitney U tests were employed for non‐normal distributed data. In multi‐group comparisons, one‐way ANOVA with Holm‐Sidak's post hoc testing was conducted only when both normality and variance homogeneity assumptions were met. When either assumption was violated, non‐parametric analysis was performed using Kruskal‐Wallis tests followed by Dunn's post hoc correction with family‐wise error rate control.