Clumping of p62 Protein Controls Mitochondrial Grouping to Help Maintain Mitochondrial Quality

Endogenous and exogenous p62 have been reported to form droplet‐like bodies by phase separation in cultured cells (Bjørkøy et al. 2005; Sun et al. 2018). We observed the fast fusion and fission process between p62 bodies in cells and in vitro (Figure S1A,B) and fluorescence recovery after photobleaching (FRAP; Figure S1C,D), indicating the liquid‐like feature of p62 bodies in live HEK 293 cells, similar to the previous study reported (Sun et al. 2018).

Several ALS/FTD‐associated p62 missense mutations (absent in the control population) have been reported (Fecto et al. 2011; Chen et al. 2014; Rea et al. 2014), but it is still unclear whether these mutations influence phase separation of p62. To investigate this, we generated a series of p62 mutants, including R321C, L341V, P392L, and G425R (Figure 1A; Fecto et al. 2011; Teyssou et al. 2013), and analyzed their phase separation behavior. The results showed that p62 mutations exhibited slower FRAP and a more immobile fraction than wildtype (WT), indicating these p62 mutations impair the dynamics of p62 bodies (Figure 1B,C, Figure S1E). Moreover, we used a droplet fusion assay to determine the effect of these mutations on p62 condensate dynamics and liquidity (Figure 1D), and the statistical results about fusion timescale (as indicated by relaxation time τ; Ravindran et al. 2023; Gordon et al. 2025) showed that these mutations decreased droplet fusion dynamics (Figure 1E,F). Taken together, these findings suggest that ALS/FTD‐associated mutations accelerated the liquid‐to‐gel phase transition of p62 bodies.

According to previous studies, the PB1 domain (the N‐terminal dimerization domain) is necessary for p62 body formation (Ciuffa et al. 2015). To study the phase separation behavior of p62, we created a light‐inducible “OptoDroplet” model of p62 condensation to study “phase separation behavior,” since the previous paper showed that proteins with a self‐oligomerization domain for condensation are ideal for designing light‐inducible chimera constructs by replacing this self‐oligomerization domain with Cry2 (light‐inducible oligomerization element; Zhang et al. 2019). In detail, the PB1 domain (self‐oligomerization domain) of p62 was replaced by the blue (488 nm) light‐dependent self‐oligomerization domain (Cry2_PHR, hereafter referred to as Cry2). We named the resulting chimeric p62 protein (Cry2‐mCherry‐p62) “Opto‐p62,” which can temporally and spatially control p62 phase separation in live cells using blue light (Figure 2A). Unlike negative Opto‐Control (Cry2‐mCherry tag alone), Opto‐p62 dramatically underwent phase separation and formed p62 body‐like droplets (p62 condensates) after short‐term blue light stimulation, and p62 bodies could spontaneously disassemble after blue light withdrawal (Figure 2B,C, Video S1). These results indicate that Opto‐p62 is an effective and highly controllable tool for studying p62 phase separation.

As presented above, the liquid‐like feature of the p62 bodies is impaired by ALS/FTD‐related mutations, and we next explored the effect of ALS/FTD‐related p62 mutations on phase separation using Opto‐p62. To this end, we fused each p62 mutation protein (deletion of PB1 domain) to Cry2, and then intermittently photo‐activated Opto‐p62 using blue light to imitate chronic and intermittent assembly and disassembly of p62 condensates in live cells (Figure 2D, Figure S2A, Video S2). The results showed that the R321C mutation inhibited the formation of p62 condensates relative to WT and other mutations, indicating the R321C mutant displays a lower phase separation tendency. Moreover, we found that other p62 mutations weakened the ability of p62 condensates to disassemble during four stimulation cycles, indicating these light‐induced condensates containing p62 mutants display gel‐like behaviors, similar to “aging‐related” pathological protein (p62, TDP‐43, and FUS) aggregates in ALS/FTD (Figure 2D, Figure S2A, Video S2; Ling et al. 2013; Patel et al. 2015).

Given the fact that the multivalent interactions between p62 and Ub drive p62 phase separation, which in turn helps to capture ubiquitinated cargos for the recognition by ATG8 family protein and subsequent autophagy (Sun et al. 2018; Zaffagnini et al. 2018; Isogai et al. 2011), and our observation that the co‐condensation of p62 and Ub in cultured neurons (Figure 3A), we wondered whether optical control of Opto‐p62 can affect condensation of ubiquitinated proteins, and vice versa. Our results showed that photo‐activation of Opto‐p62 could induce condensate formation of ubiquitinated proteins (Figure 3B,C). In addition, we created a light‐inducible model of Ub condensation by fusing mCherry‐Ub to Cry2, named Opto‐Ub (Cry2‐mCherry‐Ub), as shown in the previous paper to study the association between Ub and p62 (Hsieh and Yang 2019). We found that p62, unlike the negative control, could co‐condense with Opto‐Ub upon photo‐activation, and decrease the condensation threshold of Opto‐Ub (Figure 3D–F). Taken together, these results indicate that p62 and ubiquitinated proteins can positively cooperate with each other to form p62‐Ub condensates, which may display potential cellular functions in the regulation of autophagy.

To explore the potential contribution of p62‐Ub condensation to mitophagy, we used the combination of Antimycin A and Oligomycin A (A/O) treatments, which can abolish mitochondrial respiration to induce PINK1/Parkin‐mediated mitophagy in cultured cells (Yang et al. 2016). We found that the Parkin C431F mutation, a mutation in the RING2 domain that abolishes Parkin E3 ligase activity and Parkin‐mediated ubiquitination of mitochondria during mitophagy (Sun et al. 2022), strikingly impaired the translocation of p62 onto A/O‐treated damaged mitochondria (Figure S2B). Importantly, we also found that p62 failed to translocate onto mitochondria in cultured neurons with healthy mitochondria (Control group) or low levels of Parkin‐mediated ubiquitination (A/O group), compared with Parkin‐expressing neurons (Figure 3G,H). These data demonstrate that the remarkable translocation of p62 onto mitochondria depends on Parkin‐mediated ubiquitination, suggesting a potential role of p62‐Ub condensation in Parkin‐mediated mitochondrial quality control.

To decipher the special role of different autophagy receptors in the quality control of mitochondria, we carefully examined the contribution of p62 and OPTN to PINK1/Parkin‐mediated mitophagy in cells. Notably, we observed that p62 tended to associate with “grape‐like” mitochondrial clustering (Hsieh and Yang 2019), whereas OPTN colocalized with LC3 on dispersed mitochondria in p62 knockout cells (Figure 4A, Figure S2C), indicating that p62 depletion is associated with dispersed distribution of mitochondria in PINK1/Parkin‐mediated mitophagy. To further explore the special mechanism of mitochondrial clustering, we first tested the effect of the microtubule network (regulator of mitochondrial positioning) and p62 on mitochondrial clustering. In contrast to perinuclear “grape‐like” clusters after 4 h of A/O treatment, we found that disassembly of microtubule scattered grape‐like mitochondrial clusters to smaller mitochondrial clusters, whereas p62 depletion drove complete dispersion of mitochondria in the cytosol (Figure 4B, Figure S2D,E). These results indicate that p62 condenses damaged mitochondria to form mitochondrial clusters with small size, and the microtubule network stretches mitochondria clusters together to form integrated mitochondrial clusters (the clustering of depolarized mitochondria within a juxtanuclear region, compared with simple p62 phase separation‐mediated mitochondrial clusters). To assess the functional significance of p62 in mitophagy, we examined whether damaged mitochondria could still be degraded in p62‐depleted cells. The results showed that knockdown of p62 increased the process of mitochondrial clearance after A/O treatment (Figure 4C,D).

To further explore the relationship between membrane‐free p62 condensation and mitochondrial clustering, we evaluated the p62‐associated role with different expression levels in mitochondrial clustering during mitophagy. Surprisingly, in contrast to dispersive mitochondria in p62‐depleted cells (Figure 4E), we found that endogenous p62 condensates (puncta structures not autophagosomes) stuck to p62‐labeled mitochondria (Figure 4F), where p62 condensates acted as a “bridge” for linking individual mitochondria at the incipient stage of mitophagy in cells (Figure 4G).

Furthermore, we found high expression level of p62 could form a giant condensate which contained multiple mitochondria inside the body during mitophagy (Figure 4H), and quantification data suggest that the clustering of mitochondria (as indicated by compaction index [CI]; Narendra et al. 2010) was enhanced by the expression level of p62 (Figure 4I). In addition, we showed that p62 depletion impaired the formation of perinuclear “grape‐like” clusters in mouse primary neuronal culture (Figure 4J, Figure S2D) and human neuronal SH‐SY5Y cells (Figure S3A). Our data indicate that p62 condensation is both necessary and sufficient for mitochondrial clustering during mitophagy.

To confirm the direct role of p62 phase separation on mitochondrial clustering, we analyzed the relationship between light‐induced p62 phase separation and mitochondrial distribution in p62‐depleted cells expressing Opto‐p62 (Figure 5A), which could dramatically phase separate into dense phase on damaged mitochondria from dilute phase (cytosol) upon stimulation (Figure 5B). As shown in 3D reconstruction, upon blue light‐induced phase separation, condensed p62 strongly colocalized with mitochondria and promoted mitochondrial clustering (Figure 5A). Furthermore, the mitochondrial track displacement program indicates that separated mitochondria tend to be compacted together into smaller regions along with activating blue light (Figure 5C). To explore the dynamic process of p62 condensation‐mediated mitochondrial clustering, we used time series modeling of live cell imaging and found that blue light‐induced p62 condensates dramatically condensed onto mitochondria and promoted mitochondrial clustering (Figure 5D). Quantification data further suggest that the CI of mitochondria (Narendra et al. 2010) was enhanced by light‐induced p62 condensation (Figure 5E). Furthermore, we established various SH‐SY5Y cell lines stably expressing Opto‐p62, and found a positive correlation between Opto‐p62 expression level and induction of Opto‐p62 condensates (Figure 5F,G). Importantly, we found that light‐induced p62 condensation significantly led to the clustering of mitochondria, even at relatively low Opto‐p62 expression level (Figure 5G). Along with the evidence that p62 condensates entrap mitochondria together in a Parkin‐mediated ubiquitination‐dependent manner (Figures 3G,H and 4F,G), these data suggest that p62‐Ub multivalent interaction‐mediated condensation drives clustering of ubiquitinated mitochondria during PINK1/Parkin‐mediated mitophagy. Taken together, our data demonstrate that unlike OPTN, which mediates autophagosome recruitment onto damaged mitochondria during mitophagy, p62 plays a different role in mitochondrial quality control by triggering damaged mitochondrial clustering through phase separation.

Given the fact that ALS/FTD‐related p62 mutations disrupt p62 phase separation which regulates mitochondrial clustering and clearance in PINK1/Parkin‐mediated mitophagy as mentioned above, we wondered whether ALS/FTD‐related mutations disturb mitochondrial clustering and clearance. Importantly, G425R mutation delayed mitochondrial clustering relative to WT and other mutations after 1.5 h A/O treatment (Figure S3B). Other mutations, but not WT, also perturbed mitochondrial clustering after A/O treatment in HEK 293 cells (Figure 6A), SH‐SY5Y cells (Figure S3C), mouse primary neuron (Figure 6D) and human induced pluripotent stem cell (iPSC)‐derived neurons (iNeurons; Figure 6E,F). Moreover, the CI analysis (Narendra et al. 2010) showed the same results (Figure 6B,C), indicating these p62 mutations, especially p62 G425R mutant, inhibit mitochondrial clustering. To further validate our results using mouse models, we tried to generate AAV‐based p62 (WT and ALS/FTD mutant) mouse model (Figure S3D), and then purified p62 bodies from mouse brain tissue and combined them with purified mitochondria from p62 knockout U2OS cells (Figure S3E), an assay of which is used to assess the ability of central nervous system‐derived p62 condensates on mitochondrial clustering. The results showed that p62 WT, but not P392L mutant condensates, promoted mitochondrial clustering (Figure S3F–H). To further study the effect of p62 condensation on mitochondrial clearance, we performed a long‐term A/O treatment assay. The results showed that, compared with WT p62, p62 mutations speeded up the clearance of damaged mitochondria (Figure S4A–C). To explore the contributions of p62 pathogenic mutations to neurotoxicity upon mitochondrial damage, we used chronic and mild Antimycin A treatment to imitate chronic mitophagy and disease environment and used MAP2 (a neuron‐specific marker) to examine neurite outgrowth. The results showed that ALS/FTD‐associated p62 pathogenic mutations, which impair mitochondrial clustering, reduce neurite outgrowth in primary cultured neurons upon mitochondrial damage (Figure 6G,H), indicating p62‐mediated mitochondrial clustering display protective role upon mitochondrial damage‐triggered mitophagy. Together, these results suggest that ALS/FTD‐related p62 mutations inhibit mitochondrial clustering and then speed up mitochondrial clearance, resulting in dysfunctional mitochondrial quality control.

EGFP‐Vector, EGFP‐p62, mCherry‐LC3, BFP‐mito, EGFP‐OPTN, HA‐Parkin, FLAG‐Parkin, mCherry‐Parkin, GFP‐Parkin, HA‐Ub, pET‐15b, and mt‐Keima plasmids were described previously (Sun et al. 2022; Zhou et al. 2013). Lentiviral packaging plasmids were kindly provided by Qian Li (Shanghai Jiao Tong University). mCherry‐C1‐p62 was created by subcloning PCR product with BglII/EcoRI sites from EGFP‐N3‐p62 into mCherry‐C1 vector. Cry2‐mCherry vector was created by subcloning PCR from the original plasmid containing Cry2‐mCherry. Cry2‐mCherry‐Ub was created by subcloning PCR product with XhoI/BamHI sites from HA‐Ub into the Cry2‐mCherry vector. His‐mCherry vector was created by subcloning PCR product from the mCherry vector to pET‐15b. His‐mCherry‐p62 was created by subcloning PCR product with NotI/XhoI sites from EGFP‐p62 to His‐mCherry vector. GST‐GFP‐4xUb was produced by Genscript, and GST‐GFP‐4xUb‐His was generated by homologous recombination using ClonExpress II One Step Cloning Kit (Vazyme) with the following primers: AGCGGCCATCATCATCATCATCACTGAGCGGCCGCATCGTGA and TGATGATGATGGCCGCTGCTGCTGCCGCGCGGCACCAGTCCTCCTCGTAGTCGCAGGA. Lysine mutants (R321C, L341V, P392L, G425R) of Cry2‐mCherry‐tagged, mCherry‐C1‐tagged, and EGFP‐C1‐tagged p62 were generated by homologous recombination using ClonExpress II One Step Cloning Kit (Vazyme) with the following primers: 5′‐TGCCCTGAGGAACAGATGGAGTCGGATAACTG‐3′ and 5′‐ATCTGTTCCTCAGGGCACCCCTCGGACTCCAAGGC‐3′ for R321C mutant; 5′‐GACCCATGTGTCTTCAAAAGAAGTGGACCCGT‐3′ and 5′‐TTGAAGACACATGGGTCCAGTCATCATCTCCTC‐3′ for L341V mutant; 5′‐CTGACCTGCGGCTGATTGAGTCCCTCTCCCAG‐3′ and 5′‐AATCAGCCGCAGGTCAGCCTCTGGCGGGAGAT‐3′ for P392L; 5′‐ATGACATCAGAGCGGCTCTGGACACCATCCAG‐3′ and 5′‐AGCCGCTCTGATGTCATAGTTCTTGGTCTGCAGG‐3′ for G425R mutant. Cry2‐mCherry‐p62 (Δ1–122) was created by homologous recombination from EGFP‐N3‐p62 to Cry2‐mCherry vector with the following primers: 5′‐CTGTACAAGCACCCCAATGTGATCTGCGA‐3′; 5′‐ATTGGGGTGCTTGTACAGCTCGTCCATGCC‐3′; 5′‐TTGTGATCCGGACTCAGATCTCGAGAAG‐3′; 5′‐TCTGAGTCCGGATCACAACGGCGGGGGATG‐3′. Lentivirus‐Cry2‐mCherry‐p62 (Δ1–122) was created by homologous recombination from Cry2‐mCherry‐p62 (Δ1–122) to pHAGE‐MCS‐3 × FLAG vector with BamHI/XhoI sites with the following primers: 5′‐TCGGGTTTAAACGGATCCATGAAGATGGACAAAAAGACC‐3′; 5′‐GGGCCCTCTAGACTCGAGTCACAACGGCGGGGGATG‐3′. Lentivirus‐GFP‐p62‐WT, Lentivirus‐GFP‐p62‐L341V and Lentivirus‐GFP‐p62‐P392L were created by homologous recombination from GFP‐p62‐WT, GFP‐p62‐L341V, and GFP‐p62‐P392L to pHAGE‐MCS‐3 × FLAG vector with BamHI/XhoI sites with the following primers: 5′‐TCGGGTTTAAACGGATCCATGGTGAGCAAG‐3′; 5′‐GGGCCCTCTAGACTCGAGTCACAACGGCGG‐3′.

The following siRNAs (GenePharma) were used: Negative control: A non‐targeting oligonucleotide; Human p62: 5′‐CAUGUCCUACGUGAAGGAUGATT‐3′; Mouse p62: 5′‐GGAACUCGCUAUAAGUGCATT‐3′.

His‐mCherry‐p62 and GST‐GFP‐4xUb‐His were expressed and purified from BL21 Escherichia coli cells, respectively. After being induced by IPTG at 16°C overnight, cells were then centrifuged and resuspended in lysis buffer (20 mM Tris–HCl, 500 mM NaCl, 20 mM imidazole, 5% Glycerol, 0.5 mM PMSF, 2 mM βME) and then the resuspension solution was sonicated and pelleted at 10,000 g at 4°C for 40 min. The supernatant was added onto Ni Sepharose FF (Vazyme), and then washed column with lysis buffer. After eluted with imidazole (250 mM), the proteins were cleaved by thrombin (Solarbio) overnight at 4°C to cleave His tag and further purified and filtered by Amicon Ultra‐0.5 (Merck). Finally, the proteins were stored in the stock buffer (500 mM NaCl, 50 mM Tris–HCl, 2 mM DTT) at −80°C and analyzed by SDS‐PAGE.

Human embryonic kidney 293 (HEK 293) cells, HEK 293T, SH‐SY5Y, Neuro‐2a, and U2OS cells were cultured in 90% Dulbecco modified essential medium (DMEM; Gibco, 11995500), 10% fetal bovine serum (FBS; Gibco, 10099141C), 100 μg/mL streptomycin, and 100 U/mL penicillin (Gibco, 15140122). The p62 knockout U2OS cell lines mediated by CRISPR‐Cas9 were kindly provided by Yanfen Liu (ShanghaiTech University).

The plasmids were wrapped by Lipofectamine 2000 reagent (Invitrogen, 11668019) in Opti‐MEM (Gibco, 31985070) without serum medium for 15 min, and then transfected into the HEK 293 cells according to the manufacturer's instructions. The Opti‐MEM containing plasmids was added to cultured cells after discarding the culture medium, and the culture medium was added to these cells after 4 h. After transfection for 24 h, the cells were treated with drug or observed. The siRNA was wrapped by RNAiMAX transfection reagent (Invitrogen) in Opti‐MEM (Gibco, 31985070) for 15 min, and then transfected into the cells upon splitting according to the manufacturer's instructions.

The preparation of primary mouse cortical neurons was previously described (Xia et al. 2016). Briefly, cortical neurons were obtained from the cortex of C57BL/6 mouse embryos at embryonic Day 17. The dissociated cortical neurons were cultured in a neurobasal medium (Gibco, 21103049) with 10% FBS for 8 h, and then the medium was replaced with a supplemented neurobasal medium without serum. After being cultured for 7–9 days, the neurons were used in experiment.

The human iPSCs (National Collection of Authenticated Cell Cultures) were cultured in mTeSR1 medium (StemCell Technologies, 85850) supplemented with 10 μM ROCK inhibitor Y‐27632 (MedChemExpress) on Matrigel‐coated dishes. For neuronal differentiation modified by STEMdiff Differentiation Kit, iPSCs were dissociated with Accutase (Corning) and centrifuged at 300 g for 5 min. The iPSCs were seeded into 96‐well ultra‐low attachment plates (1 × 10⁵ cells/well) and cultured in Differentiation Medium to form embryoid bodies. After 9 days, the embryoid bodies were dissociated and cultured in Differentiation Medium to form iNeurons. After being cultured for 14–21 days, the iNeurons were used in experiment. The establishment of Opto‐p62‐expressing monoclonal cell lines was described previously (Luo et al. 2023). Briefly, lentivirus was produced with HEK 293T cells transfected with psPAX2 (packaging plasmids), pMD2.G (envelope plasmid), and Virus‐Cry2‐mCherry‐p62 (Δ1–122; transfer plasmid) by using Lipofectamine 2000. After 24 and 48 h, the viral supernatant in HEK 293T cells was respectively collected and filtered by a 0.45‐μm filter unit (Millipore), and then the viral supernatant along with 8 μg/mL polybrene (HanBio Biotechnology) was added to cultured SH‐SY5Y cells for 24 h. These cells were diluted into 96‐well plates (one‐cell‐per‐well), observed under a microscope after 7 days, and selected and cultured to establish monoclonal cell lines.

The procedure of Lentivirus‐GFP‐p62‐WT, Lentivirus‐GFP‐p62‐L341V, or Lentivirus‐GFP‐p62‐P392L was similar to that mentioned for the production of Lentivirus‐Opto‐p62. Then the harvested lentiviruses were concentrated by ultracentrifugation (25,000 rpm for 2 h at 4°C) and resuspended using 1xPBS. Finally, the viruses were stored at −80°C. Lentivirus‐mCherry‐Parkin was from HanBio Biotechnology.

The following chemicals were used for treatment: nocodazole (5 μg/mL; MedChemExpress), Antimycin A (1 or 5 μg/mL; Sigma, A8674), and Oligomycin A (1 or 5 μg/mL; Selleck Chemicals, S1478).

Mitochondrial isolation was previously described (Sun et al. 2022). Briefly, the cells were harvested in mitochondrial separation reagent A with PMSF (Beyotime Biotechnology) and ground using a Dounce tissue grinder (KIMBLE). The homogenates were centrifuged at 600 g for 10 min at 4°C to get supernatant total proteins, and then the supernatant was centrifuged at 11,000 g for 10 min at 4°C to get supernatant cytosolic proteins. The isolated mitochondria in sediment were resuspended using stock buffer (500 mM NaCl, 50 mM Tris–HCl, 2 mM DTT). Finally, the samples were analyzed by SDS‐PAGE.

The neonatal pups (C57BL/6 mice, P0) were anesthetized under ice for 5 min and then were bilaterally injected with 2 μL of AAV particles using intra‐cerebroventricular injection (ICV) at the speed of 0.2 μL/min. After the injection, the needle remained in place for 1 min, and the mice were placed in an incubator and then returned to their home cage. The AAV‐PHP.eB‐CAG > Kozak‐p62 WT‐3 × Flag‐P2A‐EGFP‐Poly(A) (titer: 2.5 × 10¹³ GC/mL) and AAV‐PHP.eB‐CAG > Kozak‐p62 P392L‐3 × Flag‐P2A‐EGFP‐Poly(A) (titer: 3 × 10¹³ GC/mL) were purchased from Cyagen Biosciences (Guangzhou, China). The experimental protocols were evaluated and approved by the Ethics and Animal Care and Use Committee of Soochow University.

The biological sample collection of brain tissues was previously described (Sun et al. 2025). Briefly, the collected mouse brains were first fixed using 4% paraformaldehyde overnight and dehydrated using 30% sucrose, and then encapsulated in OTC at −80°C and used by frozen sectioning at −20°C. In the purification experiment of p62 body in brain tissues, the collected brain samples were first homogenized in cell lysis buffer (1% NP‐40, 25 mM Tris–HCl [pH 7.6], 150 mM NaCl, and 1% sodium deoxycholate with protease inhibitor cocktail [Roche, 4693132001]). The homogenates were centrifuged at 13,000 g for 20 min at 4°C, and the supernatant containing p62 bodies was used to perform mitochondrial clustering experiment.

The immunoblot experiment was previously described (Zhou et al. 2013; Lv et al. 2017; Liu et al. 2017; Yang et al. 2018; Wang et al. 2018). In detail, the processed cells were lysed to obtain proteins using cell lysis buffer (1% NP‐40, 25 mM Tris–HCl [pH 7.6], 150 mM NaCl, and 1% sodium deoxycholate with protease inhibitor cocktail [Roche, 4693132001]). The proteins with different molecular weights in cell lysates were separated by SDS‐PAGE and then transferred onto the PVDF membrane (Merck Millipore, IPVH00010) for immunoblot analysis. The following antibodies were used: anti‐GAPDH antibody (Proteintech, 60004‐1‐Ig), anti‐TOM20 antibody (Proteintech, 11802‐1‐AP), anti‐COXIV antibody (Proteintech, 11242‐1‐AP), anti‐HSP60 antibody (Proteintech, 66041‐1‐Ig), and anti‐p62 antibody (Santa Cruz, sc‐28359).

The cells firstly were transfected with needed plasmids and then treated with indicated chemicals. Next, the cells were imaged using a live cell incubating system. Immunofluorescence assay was previously described (Wu et al. 2015; Yu et al. 2016; Chen et al. 2018; Ren et al. 2016; Fang et al. 2017). In brief, the cultured cells were fixed by paraformaldehyde (4%, 10 min) and then permeabilized by Triton X‐100 (0.1%, 10 min). Finally, the cells were stained by proper primary antibodies and matching fluorescent secondary antibodies. The following primary antibodies were used: anti‐TOM20 antibody (Proteintech, 11802‐1‐AP), anti‐p62 (Santa Cruz, sc‐28359), anti‐p62 (Cell signaling technology, D6M5X), anti‐p62 (Selleck, D16H4), anti‐HSP60 (Proteintech, 66041‐1‐Ig), anti‐Ub (Enzo Life Sciences, ENZ‐ABS840), anti‐HA antibody (Santa Cruz, sc‐7392), and anti‐MAP2 (Proteintech, 17490‐1‐AP). The cells were visualized using a Nikon fluorescent microscope (Ti2) with a pco.edge 4.2 bi sCMOS camera or a Nikon (A1R) confocal microscope with NSPARC super resolution imaging system. For in vitro imaging of p62 and mitochondria, the lysate was co‐incubated with primary and secondary antibodies (1:1000) at room temperature for 1 h.

The optogenetic tool was activated by blue light (0.5%, 1% or maximum power, 488 nm wavelength) for 1 s or continuous stimulation in cells, and then the cells were imaged by 561 nm wavelength for us to observe the condensate recovery process. For FRAP, the condensates were photobleached by 40% power of 488 nm light (for green fluorescence) or 561 nm light (for red fluorescence) on showed space and then were imaged to show the fluorescence recovery process. In quantification of FRAP results, the intensity of pre‐photobleaching droplets was normalized to 100% and the first photobleaching droplets were normalized to 0%. In Opto‐p62‐mediated mitochondrial clustering experiment, p62‐depleted cells expressing Opto‐p62 were first pre‐treated by 5 μg/mL A/O for 1.5 h, and then were activated by blue light (488 nm wavelength) and imaged by 561 nm wavelength for us to observe the mitochondrial clustering process.

Immunoblot densitometry was performed by Image processing and analysis in Java (ImageJ) software. Cellular localization of mitochondria, LC3, OPTN, p62, and Ub was determined by ImageJ software. Mitochondria track displacement, 3D reconstruction, morphometry, and area analyses were performed by Imaris 9.0.1. The charts about the obtained data were performed by Prism 8.0 (GraphPad) software. p‐values were displayed in figure legends. The formula of compaction index (CI; Narendra et al. 2010) indicates the circled perimeter with the same area as the interested object (P₀) divided by the actual perimeter of the interested object (P_real), and the following is the specific formula:Compaction IndexCI=P0Preal=2πArearealπPreal

The relaxation time (τ; Ravindran et al. 2023; Gordon et al. 2025) was obtained using NonLinearRegression function of aspect ratio (A.R.) over time (t), and the following is the specific equation (A.R.₀ indicates initial aspect ratio):A.R.=1+A.R.0−1·exp−tτ