Precise Mapping of the Proteasome Interaction Region (PIR) of p62/SQSTM1: Decoupling Condensate Formation from Proteasome Recruitment

Highlights

What are the main findings?

• This study uncovered a discrete and minimal proteasome-interacting region (PIR) within p62, precisely mapped to a six–amino acid stretch (residues 84–89; TMAMSY).
• The study demonstrates that loss of the PIR selectively disrupts the ability of p62 liquid–liquid phase–separated (LLPS) condensates to recruit 26S proteasome, resulting in impaired proteasome-dependent substrate degradation.

What are the implications of the main finding?

• These findings establish p62-mediated LLPS as a critical organizational hub for UPS activity, implying that perturbation of this mechanism may facilitate pathological processes, including neurodegeneration and malignant transformation.
• Precise identification of the PIR provides a molecular framework for dissecting the multifunctional roles of p62, enabling future studies to uncouple proteasome recruitment from other p62-dependent processes such as autophagy and signaling.

Abstract

p62/SQSTM1 is a multifunctional scaffold protein central to selective autophagy and, more recently, recognized as a regulator of ubiquitin–proteasome system-mediated degradation of intracellular proteins. Within phase-separated condensates, p62 has been shown to recruit and sequester the proteasome, yet the molecular basis for this interaction has remained largely unknown. Our previous study demonstrated that the ‘PB1’ domain (residues 1–123) of p62 is necessary for proteasome binding. However, this long stretch is also responsible for other functions of p62, such as condensate assembly and signal transduction. Thus, it was important to define more precisely the region responsible for interaction with the proteasome. In this study, we used systematic deletion variants of p62 and biochemical assays to delineate the minimal sequence within the PB1 domain responsible for proteasome binding. Our analyses revealed a small stretch of six amino acids (residues 84–89) that bind the proteasome and are distinct from the region responsible for condensate formation. Such a precise variant can serve as a useful tool to dissect how p62–proteasome interaction affects selective degradation and probably stress response, separating it from other p62 functions. Overall, this work advances our understanding of the structural determinants underlying p62’s dual role in autophagy and UPS regulation.

1. Introduction

Protein homeostasis, or proteostasis, is central to cellular health, ensuring that damaged or misfolded proteins are efficiently cleared, and that functional proteins are maintained at appropriate levels [1]. This balance is primarily maintained by two complementary degradation systems: the ubiquitin–proteasome system (UPS) and the autophagy–lysosome pathway [2,3]. Aberrations in either pathway result in damaged protein accumulation, aggregate formation, and disruption of signaling networks, which are hallmarks of aging and numerous diseases, including cancer and neurodegeneration [4]. Among the proteins coordinating these pathways, Sequestosome-1 (SQSTM1/p62) has emerged as a particularly versatile regulator, linking ubiquitination, autophagy, and proteasome function in a unique manner [5].

p62 is a 440–amino acid protein with a modular structure comprising several well-defined domains: an N-terminal Phox and Bem1p (PB1) domain, a zinc-finger ZZ domain, a ubiquitin-associated (UBA) domain, and an intrinsically disordered region (IDR) [6]. Through these regions, p62 interacts with a wide range of partners and modulates various signaling pathways. The PB1 domain enables oligomerization and the formation of higher-order assemblies, which is a prerequisite for the liquid–liquid phase separation (LLPS) [7] that drives p62 condensate formation [5]. These condensates, often referred to as “p62 bodies,” dynamically recruit a seed of ubiquitinated proteins and the UPS machinery, which are necessary for condensate assembly and efficient degradation [8]. Additionally, two short linear motifs in IDR are noteworthy: the LC3-interacting region (LIR), which mediates autophagosome formation and recruitment of ubiquitinated cargo, and the Keap1-interacting region (KIR), which links p62 to the oxidative stress response. The p62-mediated release of the transcription factor Nrf2 from the Keap1 protein during autophagy, and its translocation to the nucleus to initiate transcription, plays a key role in this process [8]. The C-terminal UBA domain binds polyubiquitin chains and thus contributes to condensate formation, and this activity is finely tuned by phosphorylation [8].

The ability of p62 to assemble into condensates has attracted significant interest. Unlike static protein assemblies such as aggregates, p62 condensates exhibit fluidity, recruit specific client proteins, and maintain the enzymatic activity of the sequestered components [5]. In mammalian cells, the proteasome and molecular chaperones localize within p62 condensates, indicating that these membraneless organelles function as specialized degradation compartments [9,10].

Earlier studies proposed that the N-terminal 210 amino acids of p62 are required for proteasome association [11]. A study from our laboratory later showed that a shorter region comprising residues 1–123, that contains the PB1 domain and a short extension that serves as a linker, was able to bind the proteasome [12]. However, deletion of this region abolished also the ability of p62 to form condensates [9,13]. Thus, it was clear that this segment harbors at least two functional and separate regions. Our aim was to generate a p62 variant that can still form condensates and function as a signaling hub, but is unable to recruit the proteasome. Such a variant has the potential to become a powerful tool with which one can test whether prevention of proteasome sequestration in p62 bodies is protective or detrimental in the contexts of stress, senescence, and neurodegenerative disease.

In this study, we defined the minimal proteasome-binding region of p62, which we denote the Proteasome Interaction Region (PIR). Using deletion variants and biochemical assays, we identified the specific stretch of amino acids within the PB1 domain that is necessary for proteasome recruitment to p62 condensates. By combining microscopy and protein degradation assays, we evaluated the impact of the PIR on p62 condensates’ proteolytic function, thus providing mechanistic insights into how p62 modulates the UPS at the level of condensate biology.

2. Materials and Methods

2.1. Cell Culture

Cells cultured on 24-well glass-bottom plates (Cellvis, Cat. #P24-1.5H-N, Mountain View, CA, USA) were fixed in 4% paraformaldehyde (15 min, RT), permeabilized in 0.1% Triton X-100 (40 min), and blocked with 5% bovine serum albumin (BSA) in PBS (1 h). Primary antibodies: anti-p53 (1:1000, Santa Cruz Biotechnology, Cat. #sc-6243, Dallas, TX, USA) or anti-p62 (1:1000, Cell Signaling Technology, Cat. #88588S, Danvers, MA, USA) or anti-FLAG (1:1000, Sigma-Aldrich, Cat. #F1804, St. Louis, MO, USA) or anti-α6 (generated in mouse hybridoma cells [9]) or anti-LC3 (1:1000, Sigma-Aldrich, Cat. #L8918, St. Louis, MO, USA) or anti-NBR1 (1:1000, Sigma-Aldrich, Cat. #55474, St. Louis, MO, USA) were applied overnight at 4 °C. After PBS washes, Alexa Fluor-conjugated secondary antibodies (Thermo Fisher Scientific, Waltham, MA, USA) were added for 1 h at room temperature. Nuclei were counterstained with Hoechst 33342 (1 µg/mL, Sigma-Aldrich, St. Louis, MO, USA).

2.2. Confocal Microscopy and Immunofluorescence

Confocal images were acquired on an Olympus EVIDENT Fluoview™ FV4000 laser-scanning confocal microscope (EVIDENT, Tokyo, Japan) equipped with an oil immersion objective. Excitation and emission detection were configured according to the fluorophore’s spectra (405 nm: Hoechst, 488 nm: Alexa Fluor 488, 555 nm: Alexa Fluor 555, 633 nm: Alexa Fluor 647). Pinhole size was set to 1 Airy unit. Laser intensity and detector gain were kept constant across conditions. Images were exported as maximum intensity projections or single optical slices using FluoView FV4000 software (version 4.1.1, EVIDENT, Tokyo, Japan).

2.3. Image and Colocalization Analysis

Raw confocal images were processed by Olympus CellSens (version 4.3, Olympus/EVIDENT, Tokyo, Japan) and analyzed using Fiji/ImageJ (version 2025-08-08, National Institutes of Health, Bethesda, MD, USA) (NIH). Colocalization between p62 condensates and the proteasome was quantified by region-of-interest (ROI)–based analysis. p62 foci were segmented in the Alexa Fluor 488 channel, and the corresponding signal intensity from proteasome markers (e.g., Alexa Fluor 555 detecting the anti–α6) was measured within each ROI. Fluorescence intensities were normalized to the cytoplasmic background. Colocalization was quantified by calculating Manders’ correlation coefficients. At least 150 cells per condition were analyzed.

2.4. Cloning

Two types of plasmid vectors were used for expressing p62 constructs: pCS2+, and NSPI lentiviral plasmids [9]. All p62 deletion variants were first generated in pCS2+ for ease of cloning and testing. Each variant was designed to remove specific regions of the human SQSTM1/p62 coding sequence (Gene ID: 8878). An N-terminal FLAG epitope tag (DYKDDDDK) was added in-frame at the 5′ end of each insert in pCS2+ to facilitate detection of the expressed proteins. For the lentiviral NSPI constructs, an N-terminal RFP tag was fused to the p62 variants via a short linker (GGACGATGGGATCCACCGGTCGCCACC) to enable fluorescence tracking in live cells.

The p62ΔPB1 and p62WT were previously described [9]. They were subcloned into the NSPI expression vector containing an N-terminal RFP tag.

To generate the PB1 (1–123 aa) construct, the corresponding coding fragment was amplified by PCR and inserted into the NSPI–RFP vector by standard restriction–ligation cloning. The following primers were used for amplification:

PB1_F: CTACCGGAATTCATGGATTACAAGGATGAC

PB1_R: CTACCGGAATTCATGGATTACAAGGATGAC.

Site-directed deletion variants of p62 were created by site-directed mutagenesis using the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies, Cat. #210518, Santa Clara, CA, USA) according to the manufacturer’s instructions. Primers for generating point mutations, deletions, and insertions were designed using the QuikChange^® Primer Design Program (Agilent Technologies, Santa Clara, CA, USA) provided by Agilent, which calculates optimal primer sequences and melting temperatures based on the kit’s validated parameters. The primer sequences used to generate p62 deletion constructs are listed below:

77–90_F: CGGGGACTTGGTTGCCAAGGATGACATCTTCC

77–90_R: GGAAGATGTCATCCTTGGCAACCAAGTCCCCG

84–89_F: TCCAGTGACGAGGAATTGGTGAAGGATGACATCTTC

84–89_R: GAAGATGTCATCCTTCACCAATTCCTCGTCACTGGA

91–100_F: GGCCCATGTCCTACGTGGAGAAAAAAGAGTGCC

91–100_R: GGCACTCTTTTTTCTCCACGTAGGACATGGC

For deletions within the 20–100 amino acid region of p62, constructs were generated using individual forward primers specific to each deletion and a shared reverse primer. The primer sequences used are listed below:

1–20_F: CTACCGGAATTCATGGATTACAAGGATGACGACGATAAGCGCCGCTTCAGCTTCTGC

1–39_F: CTACCGGAATTCATGGATTACAAGGATGACGACGATAAGGGTCCGGGACCCTGC

1–60_F: CTACCGGAATTCATGGATTACAAGGATGACGACGATAAGGGCGGCTTCCAGGCG

1–80_F: CTACCGGAATTCATGGATTACAAGGATGACGACGATAAGGAGGAATTGACAATGGCC

1–100_F: CTACCGGAATTCATGGATTACAAGGATGACGACGATAAGGAGAAAAAAGAGTGCCGG

P62_R: GGCTCTGGACACCATCCAGTATTCA

Ligation reactions were transformed into E. coli DH5α, and colonies were screened by PCR and restriction digest. Plasmid DNA was prepared using the Qiagen Miniprep kit (Qiagen, Cat. #27104, Hilden, Germany) and sequenced to confirm the precise deletion or mutation.

2.5. 1,6-Hexanediol Treatment

1,6-hexanediol (1,6-HD) was used to distinguish liquid-like (affected) from solid-like (not affected) states of p62 condensates. To assess condensate dissolution, cells were treated with 3.5% 1,6-HD for 2.5 min and subsequently fixed with 4% paraformaldehyde (PFA). A 3.5% (w/v) 1,6-HD solution was prepared by dissolving 35 mg of 1,6-HD powder in a total volume of 1 mL consisting of 250 µL complete DMEM and 750 µL sterile double-distilled water [9]. Fixed cells were then blocked in 10% goat serum and incubated with an anti-p62 antibody as described under Confocal Microscopy and Immunofluorescence.

2.6. Fluorescence Recovery After Photobleaching (FRAP)

FRAP experiments were performed in HeLa p62KO cells expressing RFP-p62WT, RFP-p62Δ77–90, or RFP-p62Δ84–89. Individual p62 condensates were photobleached using a circular ROI, and time-lapse images were acquired immediately before and after bleaching on Olympus EVIDENT Fluoview™ FV4000 laser-scanning confocal microscope (EVIDENT, Tokyo, Japan). Fluorescence intensity of the bleached ROI was measured in one pre-bleach frame and four post-bleach frames (1–4 s after bleaching) in 10 replicates per sample.

2.7. Co-Immunoprecipitation and Immunoblotting

Cells were lysed in ice-cold NP-40 lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% NP-40) supplemented with protease inhibitors, and clarified whole-cell lysates were obtained by centrifugation at 4 °C. Equal amounts of protein were incubated with anti-α6 antibody (generated in mouse hybridoma cells [9]) overnight at 4 °C with gentle rotation. Immune complexes were captured using Protein G agarose beads (Sigma-Aldrich, Cat. #11719416001, St. Louis, MO, USA) for 2–3 h at 4 °C. Beads were washed extensively with NP-40 lysis buffer, and bound proteins were eluted by boiling in SDS sample buffer.

Eluted proteins were resolved by SDS–PAGE and transferred to membranes. Membranes were blocked in 5% bovine serum albumin (BSA) and probed with anti-FLAG antibody (1:1000, Sigma-Aldrich, Cat. #F1804, St. Louis, MO, USA) to detect FLAG-tagged p62 and its mutants, or with anti-RPN1 antibody to verify co-precipitation of the fully assembled 26S proteasome. Whole-cell lysates (input controls) were analyzed in parallel and immunoblotted for FLAG-p62 and actin as a loading control. Primary antibodies were detected using appropriate HRP-conjugated secondary antibodies, and signals were visualized using enhanced chemiluminescence.

2.8. Proteasome Activity Reporter Assay (NES-GFP-CL1)

Proteasome activity was measured using the fluorescent degron reporter NES-GFP-CL1 [14]. The CL1 degron encodes an amphipathic α-helix that targets GFP for fast ubiquitin-dependent proteasomal degradation. GFP-CL1 was subcloned into an NSPI lentiviral vector with an N-terminal nuclear export signal (NES) to localize the reporter to the cytoplasm. HeLa p62KO cells reconstituted with RFP-tagged p62WT and the p62 deletion variants were stably transduced with NES-GFP-CL1 lentivirus, followed by puromycin selection (1 µg/mL, 7 days). Fluorescent protein–only controls for the CL1 reporter system were previously validated and shown not to contribute to proteasome-dependent degradation or condensate formation [9].

Proteasome-dependent NES-GFP-CL1 degradation was quantified via cycloheximide (CHX) chase assays [15]. Cycloheximide (Sigma-Aldrich, Cat. #C4859, St. Louis, MO, USA, 100 µg/mL) was added to block translation, and cells were harvested at defined time points. Cell lysates were prepared in RIPA buffer containing protease inhibitors, resolved by SDS-PAGE, and transferred onto nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). Membranes were blocked in 5% non-fat milk in Tris-buffered saline with 0.1% Tween-20 (TBST) for 1 h at room temperature, followed by overnight incubation with primary antibodies: anti-GFP (1:1000, Cell Signaling Technology, Cat. #2955S, Danvers, MA, USA) or anti-FLAG (1:1000, Sigma-Aldrich, Cat. #F1804, St. Louis, MO, USA) or anti-p53 (1:1000, Santa Cruz Biotechnology, Cat. #sc-6243, Dallas, TX, USA). After washing with TBST, membranes were incubated with HRP-conjugated secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA, 1:5000) for 1 h at room temperature. Protein signals were detected using chemiluminescence (ECL; GE Healthcare, Chicago, IL, USA). Band intensities were quantified using GelAnalyser (version 19.1, István Lázár, Budapest, Hungary) [16] and normalized to time 0 values. Degradation kinetics were expressed as a relative NES-GFP-CL1 signal over time. Linear regression analysis was used to calculate degradation slopes, and results were compared across cell lines relative to wild-type controls.

2.9. Structural Modeling of the p62Δ84-89 Variant

To assess the structural impact of deleting residues 84–89 within the PB1 domain of p62, comparative homology modeling was performed using the wild-type PB1 structure (PDB: 2KKC) as a template. A mutant model (PB1Δ84-89) was generated using the SWISS-MODEL server (https://swissmodel.expasy.org accessed on 30 September 2025) via loop reconstitution with the sequence as input [17,18,19,20,21].

For additional validation, an AlphaFold2 model of the Δ84–89 sequence was generated using the ColabFold (version v1.5.5) [22] implementation. Resulting models produced by SWISS-MODEL and AlphaFold2 were aligned to full-length PB1 structure using PyMOL [23] (version 2.5, Schrödinger, LLC, New York, NY, USA). Root-mean-square deviations (RMSD) were calculated for all atoms across aligned residues.

Model quality and stereochemical parameters were evaluated using MolProbity (version 4.5, Duke University, Durham, NC, USA) [24]. Ramachandran plot statistics, rotamer distributions, Cβ deviations, and bond/angle RMS values were recorded for both the wild-type and mutant models. Structural overlays and pLDDT were visualized in PyMOL.

2.10. Modelling of Protein–Protein Docking of p62 and RPN10

To explore potential interactions between p62 and the 19S proteasomal subunit RPN10 (S5a/PSMD4), protein–protein docking analysis was performed. Based on domain organization and known binding functions, the C-terminal region of RPN10 (residues 190–377) which contains two ubiquitin-interacting motifs responsible for binding polyubiquitin chains and ubiquitin-like receptors—was selected for docking modelling. The crystal structure of this region (PDB ID: 1YX4) was retrieved from the Protein Data Bank, with all non-protein atoms and the co-crystallized ubiquitin molecule removed prior to docking. The structure of p62 was obtained from the AlphaFold2 database (entry AF-Q13501-F1).

Docking was performed using the HADDOCK2.4 [25] web server (https://wenmr.science.uu.nl/haddock2.4/, accessed on 1 July 2025), with residues 80–93 of p62 manually defined as active, based on prior functional evidence implicating this region in proteasome engagement. All residues of RPN10 were treated as active to allow unbiased sampling of potential interaction sites. The standard HADDOCK protocol was applied, consisting of: (1) rigid-body energy minimization, (2) semi-flexible interface refinement, and (3) final refinement in explicit solvent. Output models were clustered automatically based on interface RMSD using the default cutoff of 7.5 Å, and clusters were ranked by HADDOCK score.

Top-scoring clusters were identified based on the HADDOCK score. For the representative model from the top-ranked cluster, molecular interactions were analyzed using PyMOL [23] (Schrödinger LLC). Hydrogen bonds and salt bridges were defined by interatomic distances of ≤3.5 Å between donor and acceptor atoms. Van der Waals contacts were identified between non-polar side chains located within ≤4.0 Å. Buried solvent-accessible surface area (SASA) and surface complementarity were also computed to quantify the interface properties. Structural figures and distance measurements were generated using PyMOL’s visualization and measurement tools.

2.11. Evolutionary Conservation Analysis of the p62 PIR

To assess the evolutionary conservation of residues within the proteasome-interacting region (PIR) of p62, a comparative sequence analysis was performed across vertebrate orthologs. The full-length human p62 (SQSTM1) protein sequence was used as a query for (version 2.14.1, National Center for Biotechnology Information, Bethesda, MD, USA) searches against the NCBI non-redundant protein database, restricted to Vertebrata, with an E-value cutoff of 1 × 10⁻²⁰. Retrieved sequences were filtered to retain entries with ≥80% query coverage and ≥50% sequence identity.

Redundant sequences were removed using CD-HIT clustering at a 90% sequence identity threshold, yielding a non-redundant set of 78 p62 orthologs. The resulting sequences were aligned using MUSCLE with default parameters.

Residue-level conservation was quantified relative to the human sequence by calculating, for each alignment position, the percentage of sequences containing the same amino acid as the human residue (excluding gaps). Residues were classified as highly conserved (≥80% identity), moderately conserved (50–79%), or weakly conserved (<50%). Conservation analysis focused on residues 80–95 encompassing the PIR region, and interface residues identified in the docking model were annotated accordingly.

2.12. Generative Artificial Intelligence Tools

Generative artificial intelligence tools were used in the preparation of this manuscript to assist with phrasing and structural organization of the abstract. Scientific content, data generation, analysis, and interpretation were performed and written by the authors.

3. Results

3.1. Identification of the Proteasome-Interacting Region (PIR) in p62

Previous studies demonstrated that deletion of the ‘PB1’ domain (either residues 1–123 [9] or 1–200) disrupts p62’s interaction with the proteasome. However, p62 is known to interact with the proteasome via at least two domains: (i) the PB1 domain primarily mediates proteasome recruitment for ubiquitin-mediated degradation of ubiquitinated protein targets in LLPS condensates [8]; and (ii) the UBA domain that facilitates targeting of ubiquitinated proteasome for autophagy during stress [2]. To determine whether the PB1 domain alone is sufficient for proteasome recruitment independent of other regions, we generated a FLAG-tagged PB1 construct comprising residues 1–115 (Figure 1 and Figure S1 (left panel)), which includes the ‘canonical’ PB1 domain (residues 1–102) and an adjacent linker that stabilizes it by modulating oligomerization and structural conformation. HeLa p62KO cells were transfected with constructs encoding PB1 (1–115), full-length wild-type (WT) p62 (p62WT), or a PB1-deleted p62 (residues 1–123; p62ΔPB1), followed by immunofluorescence analysis to evaluate proteasome recruitment and condensate formation. As shown in Figure 1A, PB1 alone was capable of forming cytosolic clusters that colocalized with the proteasome. As a control, p62WT formed condensates that were colocalized with the proteasome, whereas p62ΔPB1 did not. These results indicate that the PB1 fragment can cluster and probably serve as a scaffold for proteasome recruitment.

To determine whether the PB1 domain itself is capable of driving condensate assembly, we examined its behavior under conditions that disrupt weak hydrophobic interactions. HeLa p62KO cells expressing either FLAG-PB1 or FLAG-p62WT were treated with 1,6-hexanediol and processed for immunostaining. Whereas full-length p62 formed cytoplasmic condensates that rapidly dissolved upon 1,6-HD treatment, the FLAG-PB1 fragment remained in cytoplasmic clusters (Figure S2). Thus, although PB1 is sufficient for proteasome recruitment, it is not capable of forming liquid-like p62 condensates.

Having established the importance of the PB1 domain in proteasome recruitment, we aimed to identify a specific shorter region in p62 that is responsible for the binding of the proteasome. To map this region, we designed a series of p62 deletion variants focusing on the N-terminal portion (Figure S1 middle panel). Progressive truncations were generated in ~20 amino acid steps (Δ1–20, Δ1–39, Δ1–60, Δ1–80, and Δ1–100), and the truncated variants were expressed in HeLa p62KO cells for immunofluorescence analysis.

As shown in Figure 1B, p62 retained the ability to form clusters and recruit the proteasome with deletions up to ~60 amino acids from the N-terminus. However, deletion beyond 60 residues abolished clustering. These variants displayed a diffuse cytoplasmic pattern. In the three truncated variants that displayed association of p62 with the proteasome, the 1,6-hexnediol “test” demonstrated that the association represents clustering and not condensates (Figure S3). The reason is most probably the lack of residue Lys7 in all these variants, which is necessary for endowing the membraneless organelles with their liquid characteristics [9]. The complete disappearance of the association in the variant which lacks the N-terminal 80 amino acids may reflect the lack of Asp69, which is known to interact with Lys7 in condensate formation (which is lacking, anyway, in all variants) [26], but in this case may reflect its role in clustering. The importance of Lys7 and Asp69 in various forms of self-association of p62 necessitated preserving them, while still searching for the PIR in a region that is upstream of residue 60.

Based on these observations, we modified our deletion strategy to generate internal deletions, attempting to better preserve p62’s structural and functional capacity. Accordingly, we focused on the region spanning residues 76–101 and generated three constructs—p62Δ77–90, p62Δ84–89, and p62Δ91–100—which were fused to RFP to enable quantitative assessment of proteasome colocalization (Figure S1 right panel, Figure 1C). Consistent with our initial mapping, both p62Δ77–90 and p62Δ84–89 showed markedly reduced colocalization with the proteasome, whereas p62Δ91–100 retained proteasome association (Figure 1C(ii)).

1,6-HD “test” confirmed that all three variants were capable of forming punctate structures and remained hexanediol-sensitive, indicating that they still behaved as condensates (Figure 2A). To determine whether impaired proteasome recruitment reflected altered properties of the condensates, we next performed FRAP analysis on p62WT, p62Δ77–90, and p62Δ84–89 (Figure 2B). p62WT condensates exhibited rapid fluorescence recovery, characteristic of a highly dynamic liquid-like state. p62Δ77–90 recovered to a similar extent as WT. The p62Δ84–89 variant showed overall recovery behavior comparable to WT, although a modest reduction in normalized intensity was observed at the 2 s time point. This isolated delay can suggest a subtle perturbation of early-phase molecular exchange rather than a global defect in condensate dynamics (Figure 2B(ii)). Taken together, these results indicate that neither Δ77–90 nor Δ84–89 dramatically affects the physical characteristics of p62 condensates, and that loss of proteasome colocalization in these mutants occurs without major impairment of condensate fluidity.

Because both deletions (p62Δ77–90 and p62Δ84–89) impaired proteasome colocalization and LLPS properties, we next assessed whether these regions contribute to stable association with the 26S proteasome. Co-immunoprecipitation experiments (Figure S4) showed that neither p62Δ77–90 nor p62Δ84–89 was able to co-precipitate with the proteasome, indicating that this region is required for proteasome engagement. Given that both deletions produced comparable phenotypes, we focused subsequent analyses on the smaller 84–89 segment, as its selective removal is expected to minimally perturb the overall PB1 structural integrity

To determine whether any individual residue within this minimal region was solely responsible for proteasome binding, we generated a panel of single-amino acid substitutions spanning the 84–89 sequence (Figure S5). IF analysis of these mutated variants showed that none of the substitutions completely abolished proteasome colocalization; all variants retained detectable α6 recruitment to p62 condensates (Figure S5). These results indicate that no single amino acid is responsible for the interaction. Instead, the entire 84–89 stretch appears to act as the minimal PIR.

3.2. Impact of PIR Deletion on p62 Structure

3.2.1. Structural Modeling and Validation of the p62 PB1 Domain and the p62Δ84–89 Variant

To assess the structural impact of deleting residues 84–89 within the PB1 domain of p62, we generated two independent structural models using the solved PB1 structure (PDB ID: 2KKC [27], (WT PB1)) as a reference (Figure 3A(i,ii) and Figure S6). In the first approach, the p62Δ84–89 model was produced by removing the corresponding residues from the wild-type sequence and reconstructing the missing segment using SWISS-MODEL loop rebuilding (Figure 3A(iii)). In parallel, a second Δ84–89 model was generated de novo using AlphaFold2, without imposing any structural restraints from the wild-type (Figure 3A(iv)). Superposition of each Δ84–89 model onto the WT PB1 domain (Figure 3A(iii,iv)) revealed that both reconstructions retain the overall PB1 domain topology. The AlphaFold2-derived Δ84–89 model deviated from the WT PB1 structure with an RMSD of 1.140 Å, whereas the SWISS-MODEL reconstruction showed a smaller deviation of 0.117 Å. These values indicate that the predicted global PB1 fold is preserved in silico, although such superpositions do not assess protein stability or folding experimentally. To further contextualize these predictions, we examined per-residue AlphaFold confidence scores (pLDDT; encoded in the B-factor field) for the WT and Δ84–89 PB1 models (Table S1). This analysis revealed that confidence reduction in the Δ84–89 model is localized primarily to the deletion junction region, whereas regions outside the junction retain moderate-to-high predicted confidence. Thus, while the deletion introduces localized structural uncertainty, it does not result in a domain-wide loss of fold in silico.

Stereochemical validation using MolProbity [24] indicated that the SWISS-MODEL PB1 Δ84–89 structure exhibits acceptable geometry, with no major outliers and only minor deviations confined to the reconstructed loop region (Figure S7A), comparable to those observed in the WT PB1 template (Figure S7B). These metrics indicate that loop deletion and rebuilding did not introduce severe geometric artifacts, supporting structural interpretability of the model, while not constituting validation of folding or stability.

3.2.2. Docking Analysis of the p62 PIR Interaction with the UIM-Containing Arm of RPN10

At this stage, we sought to examine the structural basis of the interaction between the p62 proteasome-interacting region (PIR; residues 84–89) and the 19S proteasomal subunit RPN10, which has been reported to bind p62 directly [11]. To this end, we performed protein–protein docking using the HADDOCK platform [25,28] docking an AlphaFold2 model of full-length p62 (AF-Q13501-F1) against a C-terminal fragment of human RPN10 encompassing residues 196–306. This fragment corresponds to the UIM2-containing arm of full-length RPN10 and was derived from PDB ID 1YX4 (Figure 3B(i)).

This analysis was designed to test the hypothesis that the p62 region encompassing residues 80–95 can function as a direct docking interface for RPN10. The top-ranked docking solution revealed that this p62 segment associates along a surface groove of the UIM2-containing region of RPN10, forming an interface characterized by electrostatic/polar interactions and a central hydrophobic packing region (Figure 3B(i,ii); Tables S2 and S3). The p62–RPN10 interface involves residues Asp80, Glu81, Thr84, Met85, Met87, Ser88, Val90, Lys91, Asp92, and Asp93, encompassing and extending beyond the core PIR motif (TMAMSY; residues 84–89). Key contacts include hydrogen bonds formed by p62 Asp80 and Glu81 with RPN10 Gln297 and Ser295 (corresponding to residues 102 and 100 in the PDB fragment), as well as interactions involving p62 Lys91 and Asp93 with RPN10 Glu226 and Arg237 (corresponding to residues 31 and 42 in the PDB fragment), mediated by a combination of hydrogen bonds and a salt bridge (Figures S8A(i,ii) and S9A(i,ii)). In addition, p62 residues Thr84 through Val90 form close hydrophobic contacts with a surface patch on RPN10 comprising Leu276, Pro277, Leu279, and Tyr290 (full-length numbering), displaying high surface complementarity (~0.7–0.85) and substantial buried solvent-accessible surface area (>70% for several residues; Table S3; Figure S10). Notably, this segment corresponds to the loop region deleted in the p62Δ84–89 variant.

3.2.3. Evolutionary Conservation of the p62 PIR

To assess the evolutionary conservation of the p62 PIR, we analyzed a multiple sequence alignment of 78 p62/SQSTM1 orthologs spanning vertebrate species. Conservation was quantified as the percentage of non-gap residues matching the human amino acid at each alignment position. This analysis revealed that several residues within and adjacent to the PIR are highly conserved, including Asp80, Glu81, Met85, Met87, Tyr89, and Val90 (>90% identity), while other residues within the loop region exhibit moderate conservation (e.g., Thr84 and Ser88). These results support the functional relevance of this region and are consistent with its proposed role in mediating p62–proteasome interactions (Supplementary Table S4).

Together, these results suggest that residues 84–89 contribute to the predicted p62–RPN10 interface and constitute a proteasome-interacting region within p62.

3.2.4. Functional Characterization of PIR and Its Requirement for Proteasome-Mediated Degradation Within p62 Condensates

To assess whether deletion of residues 84–89 does not affect other p62 cellular functions, such as its interaction with autophagy-involved components, we examined its ability to interact with LC3B (a well-established binding partner which is involved in autophagosome assembly) and Neighbor of BRCA1 (NBR1), an autophagy receptor [13]. NBR1 contains its own PB1 domain, which interacts with the PB1 domain of p62 through PB1–PB1 domain interactions, thereby cooperating in the autophagosomal degradation of ubiquitinated cargo. p62KO cells that were reconstituted with either p62WT or p62Δ84–89 were treated with chloroquine, a lysosomotropic agent, and analyzed by IF staining. As shown in Figure 4A, both p62WT and p62Δ84–89 condensates colocalized with LC3B, indicating that this deletion does not compromise autophagic flux or the p62–LC3B interaction. Likewise, NBR1 staining (Figure 4B) also revealed colocalization with both p62WT and p62Δ84–89 condensates, confirming that deletion of the seven–amino acid PIR segment does not compromise scaffold integrity or autophagy-associated functions of p62.

To further elucidate the role of the PIR segment in the proteasomal degradation of ubiquitinated substrates within p62 condensates, we employed the NES-GFP-CL1 reporter protein [9]. In this construct, GFP is fused at its C-terminus to the CL1 degron—a 16-amino-acid amphipathic helix that sensitizes the protein for rapid degradation by the UPS, and at its N-terminus to a nuclear export signal (NES) that sequesters the reporter in the cytoplasm. HeLa p62KO cells were transduced with NES-GFP-CL1 and the RFP-tagged p62 variants, and subjected subsequently to cycloheximide (CHX) chase. As shown in Figure 5A, p62WT efficiently promoted degradation of the reporter protein, whereas both p62ΔPB1 and p62Δ84–89 variants failed to do so. Quantitative analysis of degradation kinetics for three biological repeats (Figure S11A) revealed that the decay slope of NES-GFP-CL1 signal in cells expressing p62Δ84–89 was reduced by ~71% compared to p62WT.

To further dissect this effect and assess recruitment of the reporter protein to p62 condensates, we performed degradation assays followed by IF staining. As shown in Figure 5B(i,iii) (upper panels), NES-GFP-CL1 localized within both p62WT and p62Δ84–89 condensates. However, degradation following CHX chase was observed only in cells expressing p62WT (Figure 5B(i–iii), middle panels). Treatment of cells with MG132, a proteasome inhibitor, along with CHX, abolished the degradation of the reporter protein in p62WT-expressing cells, with no effect observed in both p62Δ84–89 and p62ΔPB1-expressing cells (Figure 5B(i–iii), lower panels). No colocalization or degradation of NES-GFP-CL1 was observed in p62ΔPB1-expressing cells (Figure 5B(ii), all panels), consistent with the inability of this variant to form condensates and promote degradation.

Next, we wanted to further validate the role of the PIR in proteasome-mediated degradation, examining p53, a bona fide endogenous proteasome substrate, which has recently been shown to be degraded within cytosolic p62 condensates [29]. As shown in Figure 5C, during the CHX chase, p53 level declined rapidly in the presence of p62WT, but remained almost stable in cells expressing either p62Δ84–89 or p62ΔPB1. Quantitative analysis of the degradation kinetics for three biological repeats showed a ~74% reduction in slope steepness for p62Δ84–89 compared with p62WT (Figure S11B).

Microscopy analysis demonstrated that p53 colocalized with both p62WT and p62Δ84–89 condensates (Figure 5D(i,iii), upper panels), whereas no colocalization was observed in p62ΔPB1-expressing cells (Figure 5D(ii), upper panels). While p53 degradation occurred efficiently within p62WT condensates (Figure 5D(i), middle panels), it was markedly attenuated in p62Δ84–89 condensates or in the presence of p62ΔPB1 (Figure 5D(ii,iii), middle panels). Treatment of cells with MG132, along with CHX, abolished p53 degradation in p62WT–expressing cells. p53 in cells expressing p62Δ84–89 or p62ΔPB1 is stable, and therefore the effect of MG132 and CHX is minimal, if they have an effect at all (Figure 5D(i–iii), lower panels).

Taken together, these findings demonstrate that residues 84–89, within the PB1 domain of p62, mediate proteasome recruitment to p62 condensates. This enables the condensates to act as proteolytic hubs required for degradation of both the reporter substrate (NES-GFP-CL1) and endogenous substrates, such as p53.

4. Discussion

The main goal of this study was to identify the proteasome-interacting region within the PB1 domain of p62. The PB1 domain (residues 1–101) performs multiple functions, including scaffolding and self-oligomerization required for condensate formation, making it important to separate these activities. We identify a short sequence within the p62 PB1 domain (amino acids 84–89) as a PIR that is required for proteasome recruitment to p62 condensates and for efficient proteasome-mediated degradation of substrate proteins.

Several lines of evidence support this conclusion. First, PB1 alone is sufficient to colocalize with proteasome, and progressive N-terminal truncations localize the interaction regions to amino acids residues 77–90, with loss-of-function observed in deletion of residues 84–89 (Figure 1C and Figure S4). Functionally, removing residues 84–89 decouples condensate formation from proteasome recruitment: p62Δ84–89 retains PB1-dependent condensation, LC3 association, and NBR1 co-assembly (Figure 4), yet it is defective in driving the turnover of the model protein NES-GFP-CL1 and endogenous p53 (Figure 5). Together, these data distinguish two core activities of p62: (i) condensate formation/protein scaffolding, and (ii) proteasome interaction, pinpointing the minimal binding region to a six amino acid stretch within the PB1 domain.

The C-terminal 210 amino acids were previously implicated in proteasome association [11], and our earlier work showed that residues 1–123 are sufficient to slow proteasomal degradation [9]. By narrowing the interaction site to residues 84–89, we show that the PB1 domain contains distinct functional modules: regions that mediate self-association and filament formation, and a separate, discrete interface responsible for proteasome interaction.

Structural modeling provides a framework for the functional separation (Figure 3). Deletion of residues 84–89 does not disrupt the overall PB1 fold in silico, but instead causes a localized perturbation at the deletion site, reflected by reduced AlphaFold confidence (pLDDT) values (Table S1). Superposition of Δ84–89 and wild-type PB1 models yields low RMSD values, indicating that global PB1 topology is largely preserved, consistent with the retained ability of p62Δ84–89 to form puncta and recruit PB1-binding partners such as NBR1.

Docking analyses suggest that the 84–89 region contributes to the proteasome-interaction surface, such that its deletion could alter local surface geometry required for productive engagement. This is supported by evolutionary conservation of key residues within and adjacent to the PIR across vertebrate p62 orthologs, consistent with functional constraint on a protein–protein interaction interface (Table S4). Together, these observations support the existence of a discrete proteasome-docking element within the PB1 domain that is separable from its role in p62 self-assembly.

Functionally, the PIR appears to be a prerequisite for routing proteins to active proteasomes within p62 condensates. Two lines of experimental evidence strongly support this notion. First, NES-GFP-CL1 and p53 both accumulate within p62Δ84–89 puncta and fail to undergo efficient degradation (Figure 5). Second, CHX-chase kinetics show marked slope reductions in the degradation of the two proteins (~71 and ~74% for NES-GFP-CL1 and p53, respectively [Figure S11]) for p62Δ84–89 versus p62WT. These effects are difficult to attribute to generic defects in condensate formation or cargo capture; rather, they point to a loss of proteasome access or retention within the condensate microenvironment when the PIR is removed.

These findings may have broader implications for proteostasis during stress. p62 condensates are a hallmark of proteotoxic conditions [5,30,31,32], yet the functional consequences of proteasome engagement within these assemblies have remained unclear. Our data suggest that this balance may be governed by whether condensates are able to physically recruit the proteasome via the PIR.

For all these reasons, defining the precise proteasome-binding sequence in p62 has been essential: without knowing where and how p62 captures proteasomes, it has been impossible to dissect condensate-specific signaling roles from proteasome-dependent degradation functions. Separation-of-function variants such as p62Δ84–89 allow exactly this type of analysis. They preserve condensate assembly and signaling while selectively removing proteasome engagement, enabling direct tests of whether condensates contribute to protection or pathology in cells and in vivo. Six-residue PIR provides a tractable target for precise modulation. From a therapeutic perspective, PIR-blocking peptides or small molecules can prevent maladaptive proteasome sequestration in aggregates, for example, whereas PIR mimetics might enhance local proteasome recruitment in contexts where accelerated clearance is beneficial.

Several limitations of this study should be noted. Most experiments were performed in HeLa p62-knockout cells using ectopic expression, and future proximity-labeling approaches (e.g., BioID or APEX), as well as reconstitution with purified proteasomal components (e.g., 19S/20S), will be important to define direct binding partners and quantitative affinities. In addition, docking calculations were performed using a representative conformer from the NMR ensemble and were intended to provide a structural rationale for the experimentally identified interaction region, rather than to define a unique or quantitative binding mode. Because ubiquitin was not included in the docking calculations, the model cannot determine whether p62 binding to the UIM2 region of Rpn10 competes with ubiquitin engagement. Notably, the predicted p62 interface partially overlaps, but does not fully coincide with, the ubiquitin-binding surface defined by previous studies [33].

In summary, we delineate a minimal proteasome-interacting motif within the p62 PB1 domain (aa 84–89) that is essential for proteasome recruitment and selective degradation, yet dispensable for condensate formation and autophagic partnering. This separation-of-function explains prior domain-level observations, provides mechanistic insight into how p62 coordinates UPS and autophagy, and offers practical tools to interrogate, and potentially therapeutically modulate proteasome availability in stress, aging, and neurodegeneration.