{"paper_id":"0d8ef8d7-b371-4b48-8fc9-3cd82b1ab4c0","body_text":"Lyu /1 \nASAP-ID: Proximity labelling with small tags \nRuohua Lyu1, Kiersten M. Ruff2, Catherine S. Palmer1, Alejandra E. Ramirez3, Angelique R. \nOrmsby1, Daniel J. Scott1,4, Rohit V . Pappu2, Diana Stojanovski1, David A. Stroud1,5,6, Danny \nM. Hatters1 \n \n1Department of Biochemistry and Pharmacology and Bio21 Molecular Science and \nBiotechnology Institute, 30 Flemington Road. The University of Melbourne, Victoria 3010 \nAustralia. \n2Department of Biomedical Engineering, Center for Biomolecular Condensates, Washington \nUniversity in St. Louis, St. Louis, MO 63130, USA. \n3Materials Characterisation & Fabrication Platform. The University of Melbourne, Parkville, \nVIC 3052, Australia \n4 Florey Institute of Neuroscience and Mental Health, The University of Melbourne, \nParkville, VIC 3052, Australia \n5Murdoch Children’s Research Institute, Melbourne, VIC 3052, Australia \n6Victorian Clinical Genetics Services, Royal Children’s Hospital, Melbourne, VIC, 3052, \nAustralia \nAbstract \nBiotinylation-based proximity labelling methods are valuable for discovering protein-protein \ninteractions within cellular systems. However, one limitation of these approaches is that most \nrequire fusing the target protein with the enzyme that biotinylates nearby proteins (i.e., \nTurboID or APEX2), which risks sterically disrupting the protein's native function. Here, we \npresent a method designed to reduce the steric impact of these fusions and offer greater \nflexibility in labelling modalities. The method, Antibody and Small-tag Assembly on Proteins \nfor Interaction Detection (ASAP-ID), involves a bipartite system. Target proteins are fused to \na small peptide antigen that recruits TurboID or APEX2 fused to an antibody directed to the \nantigen. Using two different antigen/antibody systems (SunTag and MoonTag), we show that \nASAP-ID can specifically label human Lamin A in cells. The method works when the target \nprotein and nanobody are co-expressed together in cis (ASAP-IDIC). We also demonstrate \nthat the approach works when the antibody fusion is added in trans to fixed cells post-\nexpression (ASAP-IDIT). ASAP-IDIT identified more than 448 known and previously \nundescribed interactors of lamin. We further used ASAP-IDIT to study how ALS-mutant \nprofilin 1 affected its interactome. The method identified proteins involved in protein quality \ncontrol that correlated with aggregation propensity. Moreover, the different mutants showed \nvariation in the cellular location where aggregates formed. ASAP-IDIT revealed preferences \nfor mitochondrial proteins for the two profilin mutants that tend to aggregate in the \ncytoplasm, C71G and M114T, and nuclear proteins for a mutant more prone to nuclear \naggregation. These findings position ASAP-ID as a powerful addition to the proximity \nlabelling toolkit, capable of probing subtle differences in interactomes in a less invasive \nmanner.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /2 \nIntroduction \nDefining the interacting partners of proteins in cells and tissues is one of the essential goals in \nbiomolecular research. A leading approach to doing this involves proximity labelling, which \ninvolves closely localised or transiently interacting proteins being biotinylated for subsequent \ncapture and detection. Most of the proximity labelling methods require the fusion of a \ncatalytically promiscuous biotin ligase or peroxidase to the protein of interest (1). When \nsubstrate is added, biotinylation of proteins occurs within several nanometres of proximity in \nlive cells, which can be captured by streptavidin affinity for identification by mass \nspectrometry (2). One of the first-developed proximity ligation approaches, BioID, used a \nvariant of Escherichia coli biotin ligase BirA* (3). BioID has a relatively slow catalytic \nactivity, often requiring hours in the presence of substrate for labelling to be sufficient for \nidentification of interactors. However, fast-acting variants of BirA*, for example TurboID, \nhave improved the reaction time to as low as 10 minutes (2). An alternative approach for \nfaster labelling involves APEX2, an engineered soybean ascorbic acid peroxidase derivative \nthat can label proteins within 30 seconds in the presence of hydrogen peroxide and substrate \n(4). \nBoth BirA* and APEX2 derivatives are fused to target proteins, which raises the potential for \nsteric effects that hinder the target protein's function and interactions. BirA* variants used in \nBioID and TurboID are about 38 kDa (2, 3). Efforts to reduce their size have brought them \ndown to about 20 kDa, but this comes at the expense of decreased activity or protein stability \n(5). APEX2 is smaller than BirA* but still relatively large at 27 kDa (6). \nOne notable challenge we aimed to address was the requirement for an enzyme to be directly \nfused to a protein of interest. Here, we describe a proximity labelling strategy that utilises an \nepitope-antibody system to recruit enzymes to a target protein fused to a small (15-19 amino \nacid) epitope tag. We validate this approach using Lamin as a well-characterised benchmark \nfor proximity labelling, and then apply it to determine how disease-relevant mutations in \nprofilin 1, which lead to protein aggregation, alter the set of proteins in proximity to profilin. \nOf particular interest was the method's ability to perform proximity labelling of target \nproteins on fixed cells by adding the proximity labelling enzyme in trans, offering unique \nmethodological advantages for research applications.  \nExperimental Procedures \nCell culture \nThe human embryonic kidney (HEK) 293T and HeLa cell lines were obtained from the \nAmerican Type Culture Collection (ATCC, Manassas, Virginia). HEK293T cells were \ncultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 2 mM L-\nglutamine and 10% v/v fetal bovine serum (complete DMEM). HeLa cells were cultured in \nDMEM supplemented with 10% v/v fetal bovine serum and 1% v/v penicillin-streptomycin \n(Thermo Fisher Scientific). To detach cells for passage, 0.05% w/v Trypsin and 0.02% w/v \nEDTA in PBS were used. Cells were maintained in a humidified incubator at 37 °C with 5% \nv/v atmospheric CO2. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /3 \n \nPlasmids and Cloning \nThe plasmid sequences developed for the project are listed in Table S1. Human lamin A, \nAPEX2, MoonTag, MoonTag nanobody, wild type and mutant profilin1 cDNA were \nsynthesised commercially (Twist Bioscience). The cDNA of the anti-SunTag scFv, HA tag, \nand GB1 were amplified by PCR from the plasmid pHR-scFv-GCN4-HaloTag-GB1-NLS-\ndWPRE (Addgene, #106303 (7)). The cDNA of TurboID was amplified from pcDNA V5-\nTurboID-NES (Addgene, #107169 (2)) by PCR. The scFv-HA-TurboID/APEX2-GB1 \nplasmids were generated from the pHR vector using restriction enzyme-mediated cloning. \nSunTag cDNA was amplified from pcDNA4TO-5xGCN4_v4-kif18b-24xPP7 by PCR \n(Addgene, #74927 (8)). The mCherry-SunTag-lamin and mCherry-Lamin expression \nplasmids were created by replacing GFP with mCherry in the pEGFP-C2 plasmid (Clontech \nLaboratories), and inserting the SunTag and lamin cDNA using restriction enzyme cloning. \nHiFi DNA assembly (New England Biolabs) was used to replace SunTag with MoonTag. The \nmCherry was swapped with APEX2 to generate the APEX2-lamin plasmid via restriction \nenzyme cloning. For the SunTag-tagged profilin1 (PFN1) expression plasmids, EGFP was \nreplaced with SunTag or SunTag-HA sequences in the Gateway cloning destination vector \n(Addgene, #122844 (9)) using HiFi DNA assembly. The wildtype and mutant PFN1 \nsequences were commercially synthesised in a Gateway Entry vector (Twist Bioscience). \nPFN1-SunTag and PFN1-SunTag-HA were transferred into Gateway expression vectors using \nthe Gateway LR Clonase Enzyme mix (Thermo Fisher Scientific). \nTransfection \nCells were seeded either 4.5 × 104 cells into 8-well µ-slides (Ibidi) for immunofluorescence \nstudies, 5 × 105 cells into 6-well plates (Corning) for Western Blot, 1.1 × 106 cells into each \nT25 flask (Thermo Fisher Scientific) for proteomic experiments or 4 × 106 cells into T75 \nflasks (Thermo Fisher Scientific) for antibody purification. Cells were cultured in complete \nDMEM for 18 hours before transfection. Cells were transfected with plasmids using \nLipofectamine 3000 (Thermo Fisher Scientific), following the manufacturer’s guidelines. \nFive hours after transfection, the media was replaced with complete DMEM. \nMethanol Fixation \nTwenty-four hours after transfection, cells were washed with phosphate-buffered saline (PBS) \nand then fixed with methanol (pre-chilled at –20 °C overnight) at –20 °C for 10 minutes. The \nmethanol was removed, and the cells were rinsed three times with PBS. \nParaformaldehyde Fixation \nTwenty-four hours after transfection, cells were rinsed with PBS and then fixed with 4% w/v \nparaformaldehyde at room temperature for 10 minutes. The paraformaldehyde was removed \nand the cells were incubated with 50 mM ammonium chloride at room temperature for 10 \nminutes and then washed three times with PBS. For immunofluorescence, the cells were \npermeabilised with 0.1% v/v Triton X-100/PBS at room temperature for 10 minutes and \nwashed three times with PBS. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /4 \nScFv/Nanobody-APEX2 Purification \nCells transfected with scFv-HA-APEX2-GB1 or Nanobody-HA-APEX2-GB1 plasmids were \nlysed using a Triton X-100-based lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 10% v/v \nGlycerol, 2 mM EDTA, 1% v/v Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and \nprotease inhibitor cocktail (Sigma, #11836170001), pH 8.0) on ice for 10 minutes. Cell \nlysates were extruded through a 30 G syringe ten times, followed by centrifugation at 21,000 \ng for 10 minutes at 4 °C. The supernatant was incubated with pre-washed anti-HA magnetic \nbeads (Thermofisher) at room temperature for 3 hours while rotating. The prewash involved \nrinsing the beads three times with TBS-T (Tris-buffered saline (TBS) 20 mM Tris, 150 mM \nNaCl containing 0.05% v/v Tween-20, pH 8.0). The anti-HA magnetic beads were washed \nthree times with TBS-T, and the APEX2-antibody was eluted using 2 mg/mL HA peptide (in \nTBS, MedChemExpress) by incubating at 37 °C for 20 minutes while rotating. The \nsupernatant was snap-frozen in liquid nitrogen and stored at –80 °C.  \nIn-trans APEX2 Biotinylation  \nThe purified scFv-APEX2 or nanobody-APEX2 was applied to fixed cells and incubated \novernight at 4 °C with shaking. The cells were rinsed three times with PBS and then \nincubated in complete DMEM containing 100 µM biotin-phenol (Sigma) for 10 minutes at \nroom temperature. The biotinylation reaction was initiated by adding H2O2 to a final \nconcentration of 1 mM for 1 minute, followed by quenching with an in-trans quenching \nbuffer (10 mM sodium ascorbate, 5 mM Trolox in PBS). Cells were subsequently washed \ntwice with the in-trans quenching buffer. \nIn-cis APEX2 Biotinylation  \nTwenty-four hours post-transfection, cells were rinsed once with PBS and incubated in \ncomplete DMEM containing 500 µM biotin-phenol (Sigma) for 30 minutes in a humidified \nincubator at 37 °C with 5% v/v atmospheric CO2. The biotinylation reaction was initiated by \nadding H2O2 to a final concentration of 1 mM for 1 minute and quenched by replacing the \nmedia with an in-cis quenching buffer (10 mM sodium azide, 10 mM sodium ascorbate, 5 \nmM Trolox in PBS). Cells were then washed with in-cis quenching buffer twice. \nIn-cis TurboID Biotinylation \nTwenty-four hours after transfection, cells were rinsed once with PBS and incubated in \ncomplete DMEM containing 500 µM biotin (Sigma) for 2 hours in a humidified incubator at \n37 °C with 5% v/v atmospheric CO2. Cells were then washed with cold PBS (chilled on ice \nfor 1 hour) three times. \nImmunofluorescence staining and imaging \nFor the biotinylation experiments, cells were blocked with Triton blocking buffer (0.1% v/v \nTriton X-100 and 5% w/v bovine serum albumin in PBS) for 1 hour at room temperature. \nStreptavidin-conjugated Alexa Fluor 488 (Thermo Fisher Scientific, #S11223) was added for \n1 hour at room temperature (1:500 dilution in the Triton blocking solution). After two washes \nwith PBS, cells were stained with Hoechst 33342 (1:1000 dilution in PBS, Thermo Fisher \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /5 \nScientific, #H3570) for 10 minutes at room temperature, followed by three more PBS washes. \nImmunofluorescence images were acquired on a LSM880 confocal microscope (Zeiss) with a \n60× objective. \nFor experiments not involving biotinylation, cells were blocked with blocking buffer (3% w/v \nbovine serum albumin in PBS) for 1 hour at room temperature and incubated with anti-SDHA \nantibody (1:500 dilution in the blocking buffer, Abcam, #AB14715) at room temperature for \n1 hour. After two washes with PBS, cells were probed with an anti-HA antibody (1:500 \ndilution in blocking buffer, Cell Signalling, #3724S). After another two washes with PBS, \ncells were stained with Anti-rabbit Alexa Fluor 568 (1:1000 dilution in PBS, Thermo Fisher \nScientific, #A-11011) and Anti-mouse Alexa Fluor 488 (1:1000 dilution in PBS, Thermo \nFisher Scientific, #A-11001) respectively at room temperature for 1 hour, with two washes of \nPBS after each incubation. Immunofluorescence images were acquired on a Zeiss Elyra \nLSM880 confocal microscope (Zeiss) with a 63×/1.4 NA Plan-Apochromat oil immersion \nobjective. Super-resolution structured illumination microscopy (SIM) was performed on a \nZeiss Elyra 7 Lattice SIM (Zeiss) with a 63×/1.4 NA Plan-Apochromat oil immersion \nobjective and a pco.edge sCMOS 4.2 CL HS camera. SIM reconstructions were performed \nwith Zeiss ZEN Black 3.1 SR software with the Lattice SIM2 processing function.  \nEnrichment with streptavidin and sample preparation for mass-spectrometry \nHEK293T cells were plated on T25 flasks and transfected with the lamin or PFN1 constructs. \nCells were then processed as described above for the methanol or paraformaldehyde fixation \nand biotinylation steps. For methanol fixation, cells were resuspended in 0.5 mL of buffer \ncontaining 8 M urea, 2% w/v SDS, 0.1 M DTT and incubated for 10 minutes at room \ntemperature, followed by adding 1 µL Universal Nuclease (Thermo Fisher Scientific) and \nanother 10 minutes incubation at room temperature. The lysate was transferred into LoBind \ntubes (Eppendorf) and centrifuged at 21,000 g for 10 minutes at room temperature. The \nsupernatant was transferred into new LoBind tubes and underwent the PD-10 desalting step. \nFor the paraformaldehyde fixation, cells were lysed in 0.5 mL of the buffer (10 mM Tris-HCl, \n140 mM NaCl, 1 mM EDTA, 0.5% w/v sodium deoxycholate, 1% v/v Triton X-100, 2% w/v \nSDS, 1 mM phenylmethylsulfonyl fluoride, and a protease inhibitor cocktail (Sigma), pH 8.0) \nfor 10 minutes on ice. The lysate was transferred into LoBind tubes and boiled at 100 °C for \n20 minutes, followed by incubation at 60°C for 2 h. The lysate was then centrifuged at 21,000 \ng for 10 minutes at room temperature. The supernatant was transferred into new LoBind \ntubes and desalted with PD-10 columns (Cytiva). Protein concentrations were determined by \noptical density (AU280 nm) on a Nanodrop (Thermo Fisher Scientific) using Bovine Serum \nAlbumin for the standard curve. Magnetic streptavidin beads (Thermo Fisher Scientific) were \nwashed with PBS twice. 600 µg of protein was incubated overnight with 100 µL of pre-\nwashed magnetic streptavidin beads (Thermo Fisher Scientific) at 4 °C with rotation. The \nunbound protein fraction was collected. Magnetic beads were washed twice with 1 mL of \nurea buffer (6 M urea, 100 mM Tris-HCl, pH 8.0) and once with 1 mL of PBS containing \n0.5% w/v SDS. The beads were then incubated with 200 µL of PBS containing 0.5% w/v \nSDS and 100 mM DTT at room temperature with shaking at 700 rpm. The beads were \nwashed twice with 1 mL of urea buffer and incubated with 200 µL of urea buffer containing \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /6 \n50 mM chloroacetamide for 2 h at room temperature in the dark. Beads were washed once \nwith 1 mL of urea buffer, once with 1 mL of PBS, and once with 1 mL of 50 mM ammonium \nbicarbonate (ABC), followed by a wash with 0.5 mL of 50 mM ABC. After washing, each \nsample was incubated with 100 µL of 50 mM ABC containing 1 µg of trypsin (Thermo \nFisher Scientific) at 37 °C overnight. The digested fractions were collected and acidified to \n1% v/v trifluoroacetic acid (TFA) by 10% v/v TFA. Stage tips were pre-packed with two \nlayers of Empore SPE Disks (Merck, #66884-U) membrane and activated with 50 µL of \n100% v/v acetonitrile (ACN) by centrifuging at 1,500 g, 22 °C until all liquid went through. \nTips were washed with 50 µL of 0.1% v/v TFA, 2% v/v ACN. Samples were loaded onto the \ntips and centrifuged as above. Next, tips were washed with 100 µL of 0.1% v/v TFA and 2% \nv/v ACN again and transferred into new LoBind tubes into which peptides were eluted with \n100 µL of 0.1% v/v TFA, 80% v/v ACN, dried by SpeedVac, and stored at –80°C.  \nMass spectrometry data acquisition \nLC-MS/MS was performed on an Orbitrap Ascend mass spectrometer (Thermo Scientific) \nequipped with a nanoflow reversed-phase-HPLC (Ultimate 3000 RSLC, Dionex) fitted with \nan Acclaim Pepmap nano-trap column (Dionex—C18, 100 Å, 75 µm× 2 cm) and an Acclaim \nPepmap RSLC analytical column (Dionex—C18, 100 Å, 75 µm× 50 cm). The tryptic \npeptides were injected into the enrichment column at an isocratic flow of 5 µL/min of 2% v/v \nACN containing 0.1% v/v formic acid for 5 minutes, applied before the enrichment column \nwas switched in line with the analytical column. The eluents were 5% DMSO in 0.1% v/v \nformic acid (solvent A) and 5% DMSO in 100% v/v ACN and 0.1% v/v formic acid (solvent \nB). The flow gradient was (i) 0-6min at 3% B, (ii) 6-7min, 3-4% (ii) 7-82 min, 4-25% B (iii) \n82-86min 25-40% B (iv) 86-87min, 40-80% B (v) 87-90min, 80-80% B (vi) 90-91min, 80-\n3% and equilibrated at 3% B for 10 minutes before the next sample injection. The mass \nspectrometer was operated in data-dependent acquisition mode, whereby complete MS1 \nspectra were acquired in a positive mode at 120,000 resolution. The ‘top speed’ acquisition \nmode (3 s cycle time) on the most intense precursor ion was used (i.e., charge states of 2 to \n7). MS/MS analyses were performed by 1.6 m/z isolation with the quadrupole, fragmented by \nHCD with a normalised collision energy of 30%. MS2 fragmented ion spectra were acquired \nat 15,000 resolution. Dynamic exclusion was activated for 30 s, and the AGC target was set \nto standard with auto maximum injection mode. \nMass spectrometry data analysis \nThe raw files were searched using MaxQuant (v. 2.4.3.0) with the UniProt human proteome \nas the reference (downloaded in January 2023). The enzyme specificity was set as trypsin, \nand the maximum number of missed cleavage sites permitted was two. Variable \nmodifications were used for all experiments: oxidation (M), acetylation (Protein N-term). A \nfixed modification used for all experiments was carbamidomethyl (C). Label-free \nquantification was applied without normalisation. 1% FDR was used. The match between \nruns setting was selected in the identification parameters. The mass tolerance for precursor \nions was 20 ppm, and the mass tolerance for fragment ions 20 ppm. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /7 \nWestern blot \nThe input and the unbound protein fractions collected during the enrichment using magnetic \nstreptavidin beads (Thermo Fisher Scientific) as described above were analysed by Western \nblot. Samples were mixed with SDS loading dye and dithiothreitol (DTT). The samples were \nboiled at 95°C for 10 minutes before electrophoresis. The boiled samples were loaded onto a \n12% fast-cast stain-free SDS-PAGE (Bio-Rad) and run at 130 V for 1 hour. After \nelectrophoresis, proteins were transferred to a 0.2 µm pore size PVDF membrane using the \niBlot transfer system (Thermo Fisher Scientific) for 7 minutes at 20 V . The membranes were \nblocked with a Western Blot blocking buffer (3% w/v BSA in PBS) at room temperature for 1 \nhour, then incubated with streptavidin-HRP (1:1000 dilution in the Western Blot blocking \nbuffer, Thermo Fisher Scientific, #S911) at room temperature for 30 minutes. Following \nincubation, the membranes were washed in PBS-T (PBS with 0.05% v/v Tween-20) with \nshaking at room temperature for 30 minutes. The enhanced chemiluminescence (Bio-Rad) \nreagent was prepared according to the manufacturer's instructions and applied to the \nmembrane for protein detection. Images were captured using the ChemiDoc imaging system \n(Bio-Rad). \nExperimental Design and Statistical Rationale  \nFor ASAP-ID and traditional proximity labelling constructs of APEX2-lamin fusion \nexperiments, controls were established differently. In the traditional fusion approach, free \nAPEX2 was expressed in cells as the background control and was compared to the APEX2-\ntagged protein of interest. In ASAP-ID, we compared the untagged protein against the \nepitope-tagged protein of interest. There were four biological replicates for each sample. \nFor the analysis of lamin proteomics data, the proteinGroups text file generated by MaxQuant \nwas loaded into Perseus (v 2.0.10.0), with the LFQ intensity of each sample as the main \nvariable. LFQ intensities were log2 transformed and filtered to remove rows where LFQ \nintensities were not reported in at least three SunTag-containing biological replicates in \nASAP-ID experiments or at least three APEX2-tagged protein of interest biological replicates \nin the direct fusion experiments. The data was then normalised to the pyruvate carboxylase \n(PC) protein by subtracting the PC LFQ intensities row from all other LFQ intensities in the \ncolumn. A default imputation was performed on non-SunTag-tagged samples in the ASAP-ID \nexperiments and free APEX2 samples in the direct fusion experiments. For ASAP-ID, a two-\ntailed two-sample t-test was conducted on log2 LFQ intensities comparing the SunTag-tagged \nand non-tagged protein of interest. The significance threshold for the methanol ASAP-ID was \nset as FDR < 0.05 and S0 = 1.7; for the paraformaldehyde ASAP-ID, it was FDR < 0.05 and \nS0 = 4.8. This threshold was set to exclude all proteins detected in the non-tagged lamin \nsamples on the assumption they represent background enrichment. For the traditional \nproximity APEX-lamin fusion method, a common threshold was applied: p-value < 0.05 and \nfold change > 2. Gene ontology terms were analysed using PANTHER GO-slim (v19.0) or \nSTRING (v12.0) with default settings. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /8 \nFor the ASAP-ID on PFN1, the non-tagged wildtype and mutant PFN1 served as controls and \nwere compared to the SunTag-tagged wildtype and mutant PFN1. Each sample had three \nbiological replicates. \nFor the analysis of PFN1 proteomics data, the proteinGroups text file generated by MaxQuant \nwas loaded into Perseus (v2.0.10.0), with the LFQ intensity of each sample as the primary \nvariable. Log2 transformed LFQ intensities were filtered to exclude those not identified in at \nleast two SunTag-containing biological replicates. The data was normalised to the PC protein \nby subtracting the PC protein intensity row from all rows. A default imputation was \nperformed on non-SunTag-tagged samples. LFQ intensities were then smoothed using \nPERCEPT (10), a tool that enhances focus on statistically meaningful changes in the data. \nThe scaled LFQ intensity log2 fold change after PERCEPT was loaded back into Perseus, \nfollowed by k-means clustering analysis. The optimal number of clusters was determined by \ncalculating the inertia with the sklearn machine learning library in Python. Actual clusters \nwere generated in Perseus using Euclidean distance. Gene ontology terms were analysed \nusing PANTHER GO-slim (v19.0) or STRING (v12.0) with default settings. \nFor the analysis of immunofluorescence images of PFN1-expressing HeLa cells, PFN1 was \nquantified in cells expressing the C71G, M114T, and G118V mutants using ImageJ FIJI \nsoftware. 25 to 38 cells from each variant were measured. The data were analysed statistically \nusing GraphPad Prism (v10). For two-sample comparisons, the data were analysed by an \nunpaired Student's t-test, and the data passed the Normality test. For multiple samples, we \nused one-way ANOV A with differences evaluated according to the two-stage linear step-up \nprocedure of Benjamini, Krieger and Yekutieli (11), which was used because it is sensitive \nand well-suited to detect the true positives even when data are positively dependent. We used \nthe C71G mutant as the reference sample for the test. \nSurfaces of PFN1 variants and SDHA were generated from super-resolution images using \nImaris software (v10.0). Imaris was also used to measure the distance between PFN1 and \nSDHA. \n \nResults \nThe logic of the Antibody and Small-tag Assembly Proximity-Interaction Detection (ASAP-\nID) method is outlined in Fig. 1A. It involves fusing a small epitope tag (e.g., SunTag or \nMoonTag) to the target protein, which binds to an antibody derivative (e.g. nanobody or \nsingle-chain variable fragment (scFv) chain) fused to APEX2 or TurboID. The SunTag is a \n19-amino-acid epitope for an scFv (12) and the MoonTag is a 15-amino-acid epitope for a \ncamelid variable heavy domain of heavy chain nanobody (13). The antibody derivative fusion \nalso contains the 6.2 kDa GB1 domain, derived from the immunoglobulin-binding domain of \nprotein G (14), which is designed to suppress aggregation (12). A further element is a HA tag, \nwhich is used for purification of the antibody derivative fusion to APEX2 or TurboID by \nimmunoaffinity chromatography. To apply ASAP-ID, cells are transfected to express the \nprotein of interest fused to the small epitope tag. Cells can be co-transfected with the \nantibody-biotinylating enzyme fusion for in-cis labelling (ASAP-IDIC) (Fig. 1B). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /9 \nAlternatively, cells can be transfected to express just the protein of interest-epitope tag fusion \nand fixed for in-trans labelling by application of the semi-purified antibody-APEX2 protein, \na variation of the method we call ASAP-IDIT (Fig. 1B). For ASAP-IDIT, the antibody-APEX2 \ncan be added as crude cell lysate from cells transfected with antibody-APEX2 or as a \npartially purified product using HA-immunoaffinity chromatography. ASAP-ID contrasts to \nthe traditional proximity labelling methods that involve fusing the biotinylating enzyme \ndirectly to the protein of interest (Fig. 1B). \nTo develop ASAP-ID, we first tested it on human lamin A (lamin), the target protein initially \nused to develop BioID (3). Lamin is a component of the nuclear lamina and is easy to \nvisualise by immunofluorescence confocal microscopy in transfected cells, as it is localised \nalmost exclusively at the nuclear membrane. When the scFv-APEX2 or scFv-TurboID was \nco-expressed with SunTag-tagged lamin (Fig. 1C & Fig. S1A), specific labelling was evident \nat the nuclear membrane, indicating that the ASAP-IDIC proximity labelling had worked. \nWithout the SunTag, biotinylation appeared to be non-specific and was visible throughout \ncytosol.   \nNext, we evaluated whether the in-trans labelling was effective for ASAP-IDIT. HEK293T \ncells were transfected with SunTag-tagged lamin, fixed with either methanol or \nparaformaldehyde, and then treated, in-trans, with scFv-APEX2 semi-purified via HA-\nimmunoaffinity chromatography or with crude cell lysate expressing scFv-APEX2. Both the \nsemi-purified and unpurified scFv-APEX2 caused specific biotinylation of SunTag-tagged \nlamin compared to lamin lacking the SunTag (Fig. 1D & Fig. S1B). The semi-purified scFv-\nAPEX2 in lysate maintained most of its activity after being stored as frozen aliquots for a \nmonth (Fig. S2). Specific labelling in trans was also observed with the MoonTag system \n(Fig. S1C), indicating that ASAP-IDIT can easily be adapted to different epitope tagging \nsystems. Moreover, sufficient semi-purified scFv-APEX2 can be obtained from ~20 ´ 106 \ncells and small-scale immunoprecipitation using anti-HA magnetic beads, which suggests this \nmethodology will be widely and easily accessible to researchers. \nNext, we assessed whether ASAP-IDIT would work in a proteomics format. For these tests, \nwe compared SunTag-lamin with untagged lamin. As a control, we performed a traditional \nproximity labelling experiment with lamin fused to APEX2 and compared it to APEX2 alone. \nAs expected, ASAP-IDIT resulted in a greater level of biotinylation when the SunTag epitope \nwas present (Fig. 2A & Fig. S3A). Unexpectedly, labelling of APEX2 alone had a greater \nlevel of biotinylation than when directly fused to lamin (Fig. S3B). This indicates that free \nAPEX2 in the cell exhibits a high level of non-specific reactivity to the proteome, but the in-\ntrans labelling strategy of ASAP-IDIT effectively avoids this because unbound APEX2 is \nwashed away. In principle, this supports the idea that ASAP-IDIT offers increased specificity \nfor the target protein.  \nQuantitative analysis of biotinylation-based proximity labelling data requires careful \nconsideration of the method used for abundance normalisation. In standard proteomics of cell \nlysates, protein abundances are typically normalised by median peptide counts, a method we \ncall “median normalisation.” This approach presumes that total protein levels are similar \nacross the datasets being compared, which is reasonable when comparing two lysates. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /10 \nHowever, since our experiments involve groups that do not necessarily have comparable total \nprotein levels, this assumption may not hold. Similar to immunoprecipitation experiments \nwith extensive washing of unbound material, normalising by median peptide counts can \noveremphasise background signals. Indeed, when we applied median peptide normalisation to \nthe ASAP-IDIT data, most known lamin interactors appeared “depleted” or not enriched, \nespecially when using a threshold (FDR < 0.05, S0 = 1.7) to exclude proteins enriched in the \ncontrol (Fig. 2B). \nTo find a more appropriate way to normalise the abundance of biotinylated proteins, we \nfollowed a previously described approach (15) that exploits the equal enrichment of four \nendogenously biotinylated proteins across experiments – these are proteins that are not \nexpected to be directly influenced by the APEX2 reaction, but which should otherwise be co-\npurified in proportion to the starting material (input). These proteins are Acetyl-CoA \nCarboxylase Alpha (ACACA), Propionyl-CoA Carboxylase subunit Alpha (PCCA), Pyruvate \nCarboxylase (PC) and Methylcrotonyl-CoA Carboxylase subunit 1 (MCCC1) (16). The \nbiotinylation of these four carboxylases is mediated by biotinoyl-5’-AMP (generated from \nbiotin and ATP) (16), which is distinct from the chemistry of APEX2-mediated biotinylation, \nwhich uses biotin-phenoxyl radicals (generated from biotin-phenol and H2O2). \nIn a traditional proximity labelling experiment (APEX2-lamin fusion versus APEX2 alone), \nthree of these enzymes appeared to be enriched at approximately the same level (Fig. S4), \nwhich is concordant with this hypothesis. However, ACACA appeared to have a higher level \nof enrichment in the APEX2-lamin eluates, suggesting that ACACA biotinylation is not \nindependent of APEX2-lamin localisation. None of these proteins are known interactors of \nlamin, but we decided to exclude ACACA for normalising the data. \nFor the ASAP-IDIT data and the traditional lamin-APEX2 proximity labeling data, \nnormalisation to PC, PCCA, and MCC1 proteins yields results that more accurately reflect \nthe levels of biotinylated proteins in total cell lysates via Western Blot and more effectively \nidentify known interactors with lamin (Fig. 2B & Fig. S5A-B). Interestingly, while ACACA \nappeared unsuitable for normalisation of the traditional lamin-APEX2 fusion data (Fig. S5C), \nit seemed to produce a similar result to normalisation using PC, PCCA, and MCC1 for \nASAP-IDIT (Fig. S5D), indicating that the confounding effects of ACACA are context-\ndependent. As PC, PCCA, and MCC1 provided similar results for normalisation, we focused \non further analysis using just the PC protein normalisation for simplicity, as it was the most \nabundant protein apart from ACACA of the four (Fig. S6). \nFollowing PC protein normalisation, ASAP-IDIT identified 19.8% of known lamin interactors \nas significantly enriched (3, 17) (45 out of 227) on the methanol-fixed cells, and 60.4% (137 \nout of 227) of them for the paraformaldehyde-fixed cells (Fig. 2B). This compared to the \ntraditional lamin-APEX2 fusion proximity labelling experiment, which identified 16.7 % of \nthe known lamin interactors (Fig. 2B). There was a strong correlation in enrichment of \nknown lamin interactors between labelling (in cis or in trans) and fixation methods (Fig. 2C), \nand each experiment yielded a significant GO enrichment for nuclear envelope proteins \n(Table S2), providing confidence in the validity of the methodology. The methanol fixation \ndid appear to have lower overall correlation to the other samples, which suggests some \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /11 \nproteins do not fix as evenly in methanol as with paraformaldehyde. We also noticed that the \nknown lamin interactors have higher protein intensities in the lamin-APEX2 fusion sample \nthan the ASAP-IDIT (Fig. 2C). This may reflect the more similar nature of the fusion \nexperiment to the original BioID approaches that gave rise to the list of lamin interactors used \nin this study. \nTo demonstrate the application of ASAP-IDIT, we focused on profilin 1 (PFN1) and its \nmutants, which are associated with amyotrophic lateral sclerosis (ALS) (18). PFN1, which is \n15 kDa, binds to monomeric actin through its actin-binding domain and has a role in \nregulating actin dynamics essential for various cellular processes, including cell motility, \ndivision and maintenance of cell shape (18). Beyond its interaction with actin, PFN1 binds to \npolyproline-rich sequences in proteins through its poly-L-proline (PLP) binding domain (19). \nPFN1 has other roles in gene regulation (20, 21), DNA damage response regulation (22), nuclear \nexport of actin (21) and membrane trafficking (23).  \nWe considered PFN1 (15 kDa) highly suitable for ASAP-IDIT because a large tag like APEX2 \n(27 kDa) or TurboID (38 kDa) would be much bigger than the target protein being studied \nand therefore carry a higher risk of steric interference. In fact, we found that GFP (27 kDa) \ntagging of PFN1, caused the wild-type PFN1 to form abnormal aggregates (Fig. S7A&B). \nPrior studies have found that four (untagged) ALS-related PFN1 mutants, C71G, M114T, \nE117G and G118V(18), formed distinct patterns of apparent aggregation compared to wild-\ntype PFN1 when transfected into mammalian cells (24). The addition of the SunTag did not \nappear to change the localisation of PFN1 in HEK293T or HeLa cells when we compared our \ndata (Fig. 3A) to published data (24). In essence, wild-type PFN1 formed a predominantly \ncytosolic pattern with no puncta or aggregates. E117G formed a similar pattern to wild-type \nPFN1, but with a higher proportion of PFN1 residing in the nucleus compared to the \ncytoplasm (24) (Fig. 3B). C71G, M114T and G118V all formed bright puncta characteristic \nof protein aggregates. The proportion of PFN1 in the nucleus compared to the cytoplasm \nincreased, progressively, in order for C71G, to M114T, followed by G118V (Fig. 3B). These \ndata suggest that subtle localisation differences and physical properties of the aggregates \nmight impact what proteins are colocalised with each mutant. \nIn total, 978 biotinylated proteins were enriched by proteomics upon application of ASAP-\nIDIT to the PFN1 variants (Table S3; Fig. 4A). PFN1 was consistently more abundant when it \nhad the SunTag, indicating that the proximity labelling had been effective. Of the proteins \nselectively enriched with WT PFN1 compared to the non-tagged PFN1, seven known \ninteractors of PFN1 were found (EIF5A, IPO9, CDK12, STUB1, CTTN, FLNC, and \nCYFIP1). This result provided confidence that the proximity labelling had successfully \ndetected the binding partners of PFN1. Five of these (EIF5A, IPO9, STUB1, CTTN, \nCYFIP1) were enriched with all PFN1 variants, indicating that the mutations do not disrupt \ntheir interaction with PFN1 (Fig. S8A). However, one of the ligand proteins, STUB1, was \nconsiderably more enriched in the C71G, M114T, and G118V variants suggesting a possible \ngain-of-function mechanism. Conversely, one protein, CTTN, showed a decrease in \nenrichment with the E117G, M114T, and C71G mutants, which may arise by a loss-of-\nfunction mechanism. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /12 \nWe postulated that proteins enriched with aggregated PFN1 would correlate with the mutants' \ndifferent localisation patterns. To examine this, we sorted the proteins into 13 clusters using \nk-means analysis (25) (Fig. S8B and Tables S3 & S4) and identified the patterns that \nemerged in the context of aggregation of the mutants based on imaging. Two of the clusters \nshowed patterns consistent with proteins enriching with the aggregates (clusters 1 and 9) \n(Fig. 4B). Specifically, the enrichment of these proteins increased with increasing \naggregation propensity of the PFN1 variants. Cluster 1 contained 12 proteins, including \nPFN1, FAF1, STUB1, DNAJC7, DNAJA1, NLN, POR, RBBP7, LARS2, CLPB, TOMM70 \nand IPO8. Four of these are heat shock binding proteins (DNAJA1, DNAJC7, STUB1 and \nFAF1) in GOMF: 0031072, enriched with an FDR of 0.0042 (Fig. 4B). Cluster 9 included 66 \nproteins, which like cluster 1, had functions linked to protein quality control, specifically the \nproteasome (PSMC1, PSMC2, PSMC4, PSMC6, PSMD2, PSMD7, PSMD11) (in KEGG \npathway hsa03050; enriched with an FDR of 1E-7) (Fig. 4B). Collectively, these data suggest \nthat aggregates recruit protein quality control machinery, which is consistent with what is \nknown about protein aggregation more generally (26). Moreover, another well-known ALS-\nassociated protein, TDP-43, was also in cluster 9, aligning with previous findings that mutant \nPFN1 can sequester endogenous TDP-43 (24). \nTwo other clusters stood out from the others. Cluster 3, with 69 proteins, showed selective \nenrichments for E117G and G118V mutants (Fig. 4C). In this cluster, 87% of these proteins \nare nuclear (GOCC:0005634, with enrichment FDR of 7.58E-14) (Fig. 4C). Wild-type PFN1 \nhas functions in the nucleus, including roles in gene expression regulation (21), DNA \nreplication (20), and DNA damage response and repair (22). In Cluster 3, 15 proteins \ninvolved in DNA repair (GOBP:0006281, with enrichment FDR of 1.87E-7), and 27 proteins \ninvolved in gene expression (GOBP:0010467, with enrichment FDR of 5.64E-7) were \nidentified. These findings suggest that the E117G and G118V mutants may interfere with the \nDNA repair and gene expression machinery and may do this through an aggregation-\nindependent mechanism, given that E117G doesn’t appear to aggregate. \nCluster 6 showed an enrichment of proteins for two mutants (M114T and C71G) over the \nothers (Fig. 4C). It is noteworthy that these two mutants had a greater aggregation in the \ncytosol than the other PFN1 variants. 75% of the proteinsin cluster 6 (44 of 55) are \nmitochondrial proteins (GOCC:0005739 with enrichment FDR of 1.46E-33) (Fig. 4C). The \npreferential cytoplasmic localisation of these variants may explain this co-enrichment pattern. \nFew studies have previously linked PFN1 to mitochondria. One reported that wildtype PFN1 \nregulates mitochondrial morphology, dynamics, and respiration and that M114T PFN1 forms \naggregates inside the mitochondria (27). Another reported a reduced mitochondrial content in \nM114T and C71G PFN1-transfected NSC-34 and HEK293T cells, but not in E117G and \nG118V PFN1-transfected cells(28). These findings agree with our data, suggesting that the \nM114T and C71G mutants localise or aggregate close to mitochondria where they influence \nvarious mitochondrial processes. \nTo investigate the localisation of PFN1, we conducted super-resolution fluorescence \nmicroscopy in HeLa cells and stained one of the mitochondrial proteins that we observed to \nbe enriched with M114T and C71G, SDHA, which is located in the inner matrix compartment \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /13 \nof the organelle where it is peripherally attached to the inner membrane, and co-stained the \ncells with PFN1 immunoreactivity. The surfaces of PFN1 aggregates and SDHA protein were \nreconstructed from z-stack images (Fig. S9). The shortest distance between each PFN1 \naggregate and SDHA was measured to evaluate their co-localisation. Strikingly, most of the \ncytosolic puncta of M114T and C71G are not within the mitochondria but are proximal to it; \nmore than half of the puncta were located within 0.5 µm of SDHA (Fig. 4D). Notably, SDHA \nwas in cluster 6, it is more enriched in the M114T and C71G mutants than the others. These \nresults suggest that ASAP-ID labelling is not only capable of detecting changes in the \nproximity of different PFN1 mutants but also that the labelling radius may extend to the \nmicrometre scale. \n \nDiscussion \nHere, we developed a new proximity labelling method, ASAP-ID, that removes the \nrequirement to fuse APEX2 or TurboID to the protein of interest genetically. This offers a \ngreat benefit in reducing the steric impact on the function of the protein of interest. ASAP-ID \nhas some similarities to other methods that couple biotinylation enzymes to antibodies \ntargeting the protein of interest or the post-translational modification (29, 30). The benefit of \nsuch approaches over ASAP-ID is that endogenous proteins can be directly targeted, which \nenables samples from patients or other organisms to be assayed. However, the approaches \ndepend on the quality of the antibody and its specificity, and it may be difficult to control for \nbackground. In principle, ASAP-IDIT could be developed to work on untagged endogenous \nproteins by fusing nanobodies or scFvs to the APEX2. Another use case is on tissue from \nmouse models harbouring whole body knock-in of a SunTag to a protein of interest. A similar \napproach using the similarly sized FLAG tag and classical immunoprecipitation to study \ntissue-specific differences in protein composition of the mitochondrial ribosome (31). \nOne of the questions that arose from this work is how far the proximity labelling reaches \nbeyond the target protein in ASAP-ID. This comes from the large number of interactors \ndetected for lamin (>300) and the fact that for PFN1 there was substantial labelling of \nmitochondrial proteins when it was clear by microscopy they were in physically distinct \nlocations. The labelling radius for APEX2 reaction products is determined by both the half-\nlife of the radical and the concentration of the quencher, glutathione, in the environment (32). \nThe predicted labelling radius of the APEX2 under cellular settings is 20 nm (4). 20 nm is \nlower than the resolution limit of the super-resolution imaging in our experiments. Hence, it \nis more likely that the limit is much larger than this, possibly more than 100 nm. The \ndiffusion of free radicals has been observed for the HRP-based TSA-seq method (33), which \nsuggested a limit of 1 µm, which is more consistent with our observations. Radii of up to 0.55 \nµm have also been reported in paraformaldehyde-fixed cells expressing APEX2 (34). \nAnother contributing factor to the distance of labelling may be the size of the protein \ncomplex in the ASAP-ID (epitope bound to antibody-APEX2 fusion). While the epitope is \nsmall and directs the scFv to bind to the first 10 amino acids of the SunTag (EELLSKNYHL) \n(12), the scFv creates a “linker” to the fused APEX2 that extends the labelling radius. The \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /14 \nproximity labelling range has been previously reported to increase through the addition of a \n25-nm linker consisting of 13 repeats of GGGGS in genetic fusions between BirA variant \nBioID2 and a target protein, Nup43 (35). In this example, the extended linker allowed a \ngreater number of proteins within the same complex to be captured, but also allowed a far \ngreater number of other proteins to be detected, which is consistent with our observations \nwith lamin and profilin (35). The scFv has a molecular weight of approximately 26.5 kDa, \nwhich corresponds to approximately 3 nm in width (36), which is far shorter than the \ndisordered 13´GGGGS repeat. A further factor that could influence labelling radii is the \nresidual level of glutathione. In cells, glutathione is present in millimolar concentrations and \nacts as an antioxidant (37). It can reduce free radicals, such as hydroxyl ions, and therefore \nquench the APEX2 biotinylation. However, upon fixation, most of the glutathione is expected \nto be oxidised (38) or washed away (39), which may attenuate the quenching effect, and in \nturn permit a more spatially promiscuous APEX2 reaction. \nFor applications of ASAP-ID there are several important considerations to note. For the in-cis \nco-expression approach of ASAP-IDIC, the constant expression and binding of the nanobody-\nAPEX2 to the epitope would be expected to confer steric interference, not dissimilar to the \ntraditional proximity methods involving the fusion of APEX2. However, it would offer the \nbenefit of capturing newly synthesised proteins more effectively than direct fusion because \nthe binding of the nanobody-APEX2 to the epitope would occur independently of the \nrequirement of APEX2 to fold.  ASAP-IDIT overcomes these steric effects, but the extra steps \nfor fixation and protein harvest may lead to a loss of protein identifications. For example, \nmethanol may not fix all proteins evenly and hence there could be leaching of some proteins, \nnotably hydrophobic proteins (40). Paraformaldehyde cross-linked molecules require heat-\nbased retrieval steps to reverse the cross-links. Unevenness in the efficiency of different \ncross-linked moieties may lead to biases in protein abundances that are recovered. Overall, \nour data showed that paraformaldehyde-fixation provides data that more closely correlates \nwith the traditional proximity approach (APEX2-lamin) than methanol-fixation, which \nsuggests that paraformaldehyde better preserves the representation of proximal proteins. \nOur application of ASAP-IDIT on PFN1 demonstrated the power of the approach, enabling us \nto highlight how mutations linked to ALS can lead to differences in endogenous protein \ncoaggregation. When considering the mutations on their pure tendency to aggregate, ASAP-\nIDIT showed a correlation between aggregation propensity and enrichment with protein \nquality control machinery, including DNAJA1, DNAJC7, STUB1, FAF1, the PSMC family, \nand the PSMD family. These proteins are also linked to protein aggregation in other studies \n(41-44). Both DNAJA1 and DNAJC7 are members of the J-domain protein (JDP) family, \nwhich functions as co-chaperones with Hsp70 proteins, aiding in protein folding and \npreventing aggregation (45). It has been shown that DNAJC7 preferentially binds and \nstabilises natively folded tau protein, which is a pathological protein in Alzheimer's disease, \nand prevents tau conversion into amyloids (41). DNAJA1 has been shown to have a similar \nbut lower effect on tau aggregation than DNAJC7 (41). Moreover, the loss of DNAJC7 is a \ngenetic risk factor for ALS (46). STUB1 encodes the C-terminus of Hsc70-Interacting Protein \n(CHIP), which functions as both a co-chaperone and an E3 ubiquitin ligase. CHIP interacts \nwith chaperones such as Hsp70 and Hsp90, facilitating the ubiquitination and subsequent \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /15 \nproteasomal degradation of misfolded proteins, thereby maintaining protein homeostasis (47). \nSTUB1 exhibits neuroprotective effects by regulating the degradation of mutant SOD1 (42). \nAdditionally, it is reported to interact with PFN1, ubiquitinating it for proteasome-mediated \ndegradation (43). Hence, its accumulation with aggregated PFN1 may reflect a protective role \nfor STUB1 in the clearance of misfolded protein aggregates. FAF1, also known as Fas-\nassociated factor 1, is a ubiquitin receptor that contains multiple ubiquitin-related domains. \nThe accumulation of FAF1 has been observed to play a key role in dopaminergic neuronal \ndegeneration, suggesting its involvement in the pathogenesis of Parkinson's disease through \nmechanisms related to protein aggregation (44). The PSMC family comprises ATPase \nsubunits in the 19S regulatory particle of the 26S proteasome, whereas the PSMD family \ncontains non-ATPase subunits in the 19S regulatory particle of the 26S proteasome (48). Both \nfamilies play important roles in the proteasome's function of degrading misfolded or damaged \nproteins (48, 49). Impairment of the ubiquitin-proteasome system has been demonstrated to \nresult in the accumulation of mutant PFN1 aggregates (18). Cells expressing PFN1 mutants \nC71G, M114T, and G118V exhibited numerous large protein inclusions following treatment \nwith the proteasome inhibitor MG132 (18). Cells expressing the E117G mutant displayed a \nmoderate level of aggregation, while those expressing wild-type PFN1 showed minimal \naggregate formation (18). These findings, in conjunction with our data, suggest that PFN1 \naggregates may sequester components of the protein quality control machinery, thereby \ncompromising proteostasis and contributing to the persistence and accumulation of misfolded \nprotein aggregates. \nThe other more intriguing result was the pattern of aggregate location in the cell. The C71G \nand M114T PFN1 mutants were more likely to be enriched with mitochondrial proteins, \nwhereas the E117G and G118V mutants were more enriched with nuclear proteins. This can \nbe explained by the preferred localisation of PFN1 aggregates in cells; the C71G and M114T \nhave a higher proportion of aggregates in the cytoplasm, whereas more aggregates of the \nG118V mutant appear in the nucleus.  \nThe selectivity of aggregate location may play a role disease pathomechanisms. The potential \nimpact of this is well illustrated by the aggregation behaviour of ALS-associated protein \nFused in Sarcoma (FUS) (50). FUS normally resides in the nucleus where it operates in \nfunctions related to transcription, mRNA splicing, and transport (51). In its role regulating \ntranscription, FUS binds DNA, recruits RNA polymerase II, and modulates transcriptional \nactivators (52). In its role in regulating mRNA splicing, FUS binds to heterogeneous nuclear \nribonucleoproteins (hnRNPs) and the U1 small nuclear ribonucleoprotein (snRNP) complex \n(53, 54).  In ALS patient-derived fibroblasts, FUS can form nuclear aggregates, which are \nassociated with disruptions in RNA polymerase II function (55). Mutations in the C-terminal \nregion of FUS that cause ALS, can instead, mediate mislocalisation and aggregation in the \ncytoplasm (56). In this context, the cytoplasmic FUS aggregates sequester RNA-binding \nproteins, leading to impaired RNA granule formation (57) and aberrant RNA splicing (56). \nCollectively, these findings illustrate the potential for nuclear or cytoplasmic aggregation to \ndrive different mechanisms of gain of toxic functions.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /16 \nOur data with PFN1 suggests that differential localisation of aggregation may selectively \nmodulate PFN1 dysfunction in DNA replication and DNA repair (20, 22), or mechanisms \ninvolving mitochondrial function (27).  \nAlso of relevance to ALS mechanisms was the identification of TDP-43 co-aggregating with \nPFN1 mutants in a manner correlated with PFN1 aggregation propensity (TDP-43 was in \ncluster 9). TDP-43 cytoplasmic mislocalisation from a normal nuclear localisation is a \nhallmark of ALS pathology, and the data here, as well as elsewhere, suggest PFN1 can \ninfluence this behaviour (24).   \nIn summary, ASAP-ID offers a flexible format that allows for the use of various biotinylation \nenzymes (TurboID and APEX2), epitope tags (SunTag and MoonTag), and labelling methods \n(methanol fixation or paraformaldehyde fixation for in trans labelling, as well as the in cis \nlabelling). All of these approaches have been demonstrated here to label the proximity protein \nof the model protein human lamin A in the immunostaining assay. These properties will make \nASAP-ID, particularly the adaptability and flexibility of ASAP-IDIT, a standout approach for \nversatile proximity labelling in cells and tissue samples. \nAcknowledgements \nImaging was conducted with support from the Biological Optical Microscopy Platform \n(BOMP) and the Materials Characterisation and Fabrication Platform (MCFP) at the \nUniversity of Melbourne. The proteomics data was collected by the Mass Spectrometry and \nProteomics Facility at Bio21 Institute, The University of Melbourne. The work was funded \nthrough grants to DMH (Australian Research Council: DP250100240 and DP230101050). \nDAS is supported by a National Health and Medical Research Council Investigator \nFellowship (GNT2009732). This work was supported in part by grants from the US Air Force \nOffice of Scientific Research (FA9550-20-1-0241) and the St. Jude Research Collaborative \non the Biology and Biophysics of RNP granules to RVP. \nData are available via ProteomeXchange with identifier PXD066625. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /17 \nThis article contains supplemental data and figures \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /18 \n \nFigure 1. ASAP-ID can label proximal proteins of the target protein with biotin. A. \nShown is a schematic of the ASAP-ID approach. The epitope can be either a SunTag or a \nMoonTag. The proximity ligase enzyme for biotinylating can be APEX2 or TurboID. The \nantibody, which is a nanobody or scFv, is fused to the biotinylating enzyme, allowing it to be \nrecruited to the tagged bait protein and proximal molecules labelled with biotin. B. \nIllustration of in-cis or in-trans versions of ASAP-ID (ASAP-IDIC and ASAP-IDIT \nrespectively) and the traditional fusion proximity labelling method. In ASAP-IDIC, the tagged \nbait protein and antibody-fused biotinylating enzyme are co-expressed in cells. In ASAP-\nIDIT, the tagged bait protein is first expressed in cells, and then the antibody-fused \nbiotinylating enzyme is introduced to the fixed cells for subsequent labelling. In the \ntraditional method, the bait protein is fused directly to the biotinylating enzyme. C. ASAP-\nIDIC with SunTag and APEX2. The SunTag-tagged or non-tagged Lamin was co-transfected \nwith scFv-APEX2 into HEK293T cells. Cells were then subjected to biotinylation and fixed \nwith methanol. D. ASAP-IDIT with methanol fixation. HEK293T cells were transfected with \nlamin constructs and then fixed with methanol, and the scFv-APEX2 fusion proteins were \nadded in-trans. Antibody-APEX2 was either purified using HA antibody affinity \nchromatography or expressed in cells, where cell lysates were added. \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /19 \n \nFigure 2. ASAP-ID identifies proximal proteins of lamin. A. Quantification of total \nbiotinylated proteins on Western Blot of an ASAP-IDIT conducted on HEK293 cells \ntransfected with the shown constructs (blots in Fig S3A; data shows densitometry data; \nmeans and SEM. T-test result shown: ***, P-value = 0.0004). B. V olcano plots using different \nnormalisation approaches. Labels above the graph depict the method of fixation for the \nASAP-ID experiment, or the traditional proximity labelling method by expression of the \nAPEX-lamin fusion (these cells are not fixed). Red dots represent previously established \nlamin interactors. C. Comparison of proteins identified by the different proximity labelling \nmethods. The lines represent the linear regression models with R2 values of 0.6402 (methanol \nversus paraformaldehyde), 0.2902 (Traditional proximity labelling versus methanol), and \n0.5739 (Traditional proximity labelling versus paraformaldehyde). \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /20 \n \nFigure 3. Localisation of PFN1 variants in cells. A. Immunostaining of the PFN1 variants \nfused to SunTag and HA in HEK293T cells and HeLa cells. The cells were then fixed with \nmethanol and stained with anti-HA antibody. B. Quantification of PFN1 distribution in the \nHeLa cells. The mean intensity of WT and E117G PFN1 in the nucleus and whole cell was \nmeasured in the PFN1 variant-transfected HeLa cells by using ImageJ FIJI software. The \nratio was calculated by dividing the mean PFN1 intensity in the nucleus by the mean intensity \nin the whole cell. T-test result shown: *, P<0.05. The number of puncta was measured in the \nC71G, M114T, and G118V PFN1-expressed HeLa cells using ImageJ FIJI software. The \npuncta ratio (nucleus/cell) was calculated by dividing the number of puncta measured in the \nnucleus divided by the number in the whole cell. Each dot represents one cell. Shown are the \nmultiple comparison FDR q values of the Two-stage linear step-up procedure of Benjamini, \nKrieger and Yekutieli (11) from a one-way ANOV A (P=0.0034). The linear trend P-value was \n0.008. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /21 \n \nFigure 4. ASAP-IDIT revealed interactomes of five PFN1 variants. A. Mass spectrometry \nintensity values of proteins detected across all SunTag-tagged and nontagged PFN1 variants. \nData are shown after the filtering and imputation analysis and normalised to PC. The dots \nrepresent different proteins. B. Cluster 1 and cluster 9 derived from k-means clustering. \nPathway enrichment analysis of cluster 1 and 9 was performed using STRING with default \nsettings. The enrichment score (signal) represents a weighted harmonic mean of the observed-\nto-expected ratio and –log (FDR). Full details of enrichment patterns are provided in Table \nS4. C. Same logic as panel B, except showing cluster 3 and cluster 6. D. Histogram of \nshortest distances between PFN1 puncta in the cytosol and SDHA surfaces. Surfaces of PFN1 \nmutants and SDHA were reconstructed in 3D using Imaris, and shortest distances were \nmeasured between these surfaces within the software. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /22 \nREFERENCES \n1. Qin, W., Cho, K. F., Cavanagh, P. E., and Ting, A. Y . (2021) Deciphering molecular \ninteractions by proximity labeling. Nature Methods 18, 133-143 \n2. Branon, T. C., Bosch, J. A., Sanchez, A. D., Udeshi, N. D., Svinkina, T., Carr, S. A., \nFeldman, J. L., Perrimon, N., and Ting, A. Y . (2018) Efficient proximity labeling in \nliving cells and organisms with TurboID. Nat Biotechnol 36, 880-887 \n3. Roux, K. J., Kim, D. I., Raida, M., and Burke, B. (2012) A promiscuous biotin ligase \nfusion protein identifies proximal and interacting proteins in mammalian cells. J Cell \nBiol 196, 801-810 \n4. Hung, V ., Zou, P., Rhee, H. W., Udeshi, N. D., Cracan, V ., Svinkina, T., Carr, S. A., \nMootha, V . K., and Ting, A. Y . (2014) Proteomic mapping of the human mitochondrial \nintermembrane space in live cells via ratiometric APEX tagging. Mol Cell 55, 332-341 \n5. Kubitz, L., Bitsch, S., Zhao, X., Schmitt, K., Deweid, L., Roehrig, A., Barazzone, E. C., \nValerius, O., Kolmar, H., and Béthune, J. (2022) Engineering of ultraID, a compact and \nhyperactive enzyme for proximity-dependent biotinylation in living cells. Commun Biol \n5, 657 \n6. Lam, S. S., Martell, J. D., Kamer, K. J., Deerinck, T. J., Ellisman, M. H., Mootha, V . K., \nand Ting, A. Y . (2015) Directed evolution of APEX2 for electron microscopy and \nproximity labeling. Nat Methods 12, 51-54 \n7. Liu, H., Dong, P., Ioannou, M. S., Li, L., Shea, J., Pasolli, H. A., Grimm, J. B., Rivlin, \nP. K., Lavis, L. D., Koyama, M., and Liu, Z. (2018) Visualizing long-term single-\nmolecule dynamics in vivo by stochastic protein labeling. Proc Natl Acad Sci U S A \n115, 343-348 \n8. Yan, X., Hoek, T. A., Vale, R. D., and Tanenbaum, M. E. (2016) Dynamics of \nTranslation of Single mRNA Molecules In Vivo. Cell 165, 976-989 \n9. Agrotis, A., Pengo, N., Burden, J. J., and Ketteler, R. (2019) Redundancy of human \nATG4 protease isoforms in autophagy and LC3/GABARAP processing revealed in \ncells. Autophagy 15, 976-997 \n10. Cox, D., and Hatters, D. M. (2024) PERCEPT: Replacing binary p-value thresholding \nwith scaling for more nuanced identification of sample differences. iScience 27, 109891 \n11. Benjamini, Y ., Krieger, A. M., and Yekutieli, D. (2006) Adaptive linear step-up \nprocedures that control the false discovery rate. Biometrika 93, 491-507 \n12. Tanenbaum, M. E., Gilbert, L. A., Qi, L. S., Weissman, J. S., and Vale, R. D. (2014) A \nprotein-tagging system for signal amplification in gene expression and fluorescence \nimaging. Cell 159, 635-646 \n13. Boersma, S., Khuperkar, D., Verhagen, B. M. P., Sonneveld, S., Grimm, J. B., Lavis, L. \nD., and Tanenbaum, M. E. (2019) Multi-Color Single-Molecule Imaging Uncovers \nExtensive Heterogeneity in mRNA Decoding. Cell 178, 458-472 e419 \n14. Gronenborn, A. M., Filpula, D. R., Essig, N. Z., Achari, A., Whitlow, M., Wingfield, P. \nT., and Clore, G. M. (1991) A novel, highly stable fold of the immunoglobulin binding \ndomain of streptococcal protein G. Science 253, 657-661 \n15. Frankenfield, A. M., Fernandopulle, M. S., Hasan, S., Ward, M. E., and Hao, L. (2020) \nDevelopment and Comparative Evaluation of Endolysosomal Proximity Labeling-\nBased Proteomic Methods in Human iPSC-Derived Neurons. Anal Chem 92, 15437-\n15444 \n16. Tong, L. (2013) Structure and function of biotin-dependent carboxylases. Cell Mol Life \nSci 70, 863-891 \n17. Go, C. D., Knight, J. D. R., Rajasekharan, A., Rathod, B., Hesketh, G. G., Abe, K. T., \nYoun, J. Y ., Samavarchi-Tehrani, P., Zhang, H., Zhu, L. Y ., Popiel, E., Lambert, J. P., \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /23 \nCoyaud, E., Cheung, S. W. T., Rajendran, D., Wong, C. J., Antonicka, H., Pelletier, L., \nPalazzo, A. F., Shoubridge, E. A., Raught, B., and Gingras, A. C. (2021) A proximity-\ndependent biotinylation map of a human cell. Nature 595, 120-124 \n18. Wu, C. H., Fallini, C., Ticozzi, N., Keagle, P. J., Sapp, P. C., Piotrowska, K., Lowe, P., \nKoppers, M., McKenna-Yasek, D., Baron, D. M., Kost, J. E., Gonzalez-Perez, P., Fox, \nA. D., Adams, J., Taroni, F., Tiloca, C., Leclerc, A. L., Chafe, S. C., Mangroo, D., \nMoore, M. J., Zitzewitz, J. A., Xu, Z. S., van den Berg, L. H., Glass, J. D., Siciliano, G., \nCirulli, E. T., Goldstein, D. B., Salachas, F., Meininger, V ., Rossoll, W., Ratti, A., \nGellera, C., Bosco, D. A., Bassell, G. J., Silani, V ., Drory, V . E., Brown, R. H., Jr., and \nLanders, J. E. (2012) Mutations in the profilin 1 gene cause familial amyotrophic lateral \nsclerosis. Nature 488, 499-503 \n19. Carlsson, L., Nystrom, L. E., Sundkvist, I., Markey, F., and Lindberg, U. (1977) Actin \npolymerizability is influenced by profilin, a low molecular weight protein in non-\nmuscle cells. J Mol Biol 115 \n20. Zhu, C., Iwase, M., Li, Z., Wang, F., Quinet, A., Vindigni, A., and Shao, J. (2022) \nProfilin-1 regulates DNA replication forks in a context-dependent fashion by \ninteracting with SNF2H and BOD1L. Nat Commun 13, 6531 \n21. Zhu, C., Kim, S. J., Mooradian, A., Wang, F., Li, Z., Holohan, S., Collins, P. L., Wang, \nK., Guo, Z., Hoog, J., Ma, C. X., Oltz, E. M., Held, J. M., and Shao, J. (2021) Cancer-\nassociated exportin-6 upregulation inhibits the transcriptionally repressive and \nanticancer effects of nuclear profilin-1. Cell Rep 34, 108749 \n22. Lee, C. J., Yoon, M. J., Kim, D. H., Kim, T. U., and Kang, Y . J. (2021) Profilin-1; a \nnovel regulator of DNA damage response and repair machinery in keratinocytes. Mol \nBiol Rep 48, 1439-1452 \n23. Dong, J., Radau, B., Otto, A., Müller, E., Lindschau, C., and Westermann, P. (2000) \nProfilin I attached to the Golgi is required for the formation of constitutive transport \nvesicles at the trans-Golgi network. Biochimica et Biophysica Acta 1497 \n24. Tanaka, Y ., Nonaka, T., Suzuki, G., Kametani, F., and Hasegawa, M. (2016) Gain-of-\nfunction profilin 1 mutations linked to familial amyotrophic lateral sclerosis cause seed-\ndependent intracellular TDP-43 aggregation. Hum Mol Genet 25, 1420-1433 \n25. MacQueen, J. (1967) Some methods for classification and analysis of multivariate \nobservations.  Proceedings of 5-th Berkeley Symposium on Mathematical Statistics and \nProbability/University of California Press \n26. Rajendran, A., and Castañeda, C. A. (2025) Protein quality control machinery: \nregulators of condensate architecture and functionality. Trends in Biochemical Sciences \n50, 106-120 \n27. Read, T. A., Cisterna, B. A., Skruber, K., Ahmadieh, S., Liu, T. M., Vitriol, J. A., Shi, \nY ., Black, J. B., Butler, M. T., Lindamood, H. L., Lefebvre, A. E., Cherezova, A., \nIlatovskaya, D. V ., Bear, J. E., Weintraub, N. L., and Vitriol, E. A. (2024) The actin \nbinding protein profilin 1 localizes inside mitochondria and is critical for their function. \nEMBO Rep 25, 3240-3262 \n28. Teyssou, E., Chartier, L., Roussel, D., Perera, N. D., Nemazanyy, I., Langui, D., Albert, \nM., Larmonier, T., Saker, S., Salachas, F., Pradat, P. F., Meininger, V ., Ravassard, P., \nCote, F., Lobsiger, C. S., Boillee, S., Turner, B. J., Seilhean, D., and Millecamps, S. \n(2022) The Amyotrophic Lateral Sclerosis M114T PFN1 Mutation Deregulates \nAlternative Autophagy Pathways and Mitochondrial Homeostasis. Int J Mol Sci 23 \n29. Santos-Barriopedro, I., van Mierlo, G., and Vermeulen, M. (2023) Off-the-shelf \nproximity biotinylation using ProtA-TurboID. Nat Protoc 18, 36-57 \n30. Li, X., Zhou, J., Zhao, W., Wen, Q., Wang, W., Peng, H., Gao, Y ., Bouchonville, K. J., \nOffer, S. M., Chan, K., Wang, Z., Li, N., and Gan, H. (2022) Defining Proximity \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /24 \nProteome of Histone Modifications by Antibody-mediated Protein A-APEX2 Labeling. \nGenomics Proteomics Bioinformatics 20, 87-100 \n31. Busch, J. D., Cipullo, M., Atanassov, I., Bratic, A., Silva Ramos, E., Schöndorf, T., Li, \nX., Pearce, S. F., Milenkovic, D., Rorbach, J., and Larsson, N. G. (2019) MitoRibo-Tag \nMice Provide a Tool for In Vivo Studies of Mitoribosome Composition. Cell Rep 29, \n1728-1738.e1729 \n32. Qin, W., Cho, K. F., Cavanagh, P. E., and Ting, A. Y . (2021) Deciphering molecular \ninteractions by proximity labeling. Nat Methods 18, 133-143 \n33. Chen, Y ., Zhang, Y ., Wang, Y ., Zhang, L., Brinkman, E. K., Adam, S. A., Goldman, R., \nvan Steensel, B., Ma, J., and Belmont, A. S. (2018) Mapping 3D genome organization \nrelative to nuclear compartments using TSA-Seq as a cytological ruler. J Cell Biol 217, \n4025-4048 \n34. Tran, J. R., Paulson, D. I., Moresco, J. J., Adam, S. A., Yates, J. R., Goldman, R. D., \nand Zheng, Y . (2021) An APEX2 proximity ligation method for mapping interactions \nwith the nuclear lamina. J Cell Biol 220 \n35. Kim, D. I., Jensen, S. C., Noble, K. A., Kc, B., Roux, K. H., Motamedchaboki, K., and \nRoux, K. J. (2016) An improved smaller biotin ligase for BioID proximity labeling. Mol \nBiol Cell 27, 1188-1196 \n36. Asaadi, Y ., Jouneghani, F. F., Janani, S., and Rahbarizadeh, F. (2021) A comprehensive \ncomparison between camelid nanobodies and single chain variable fragments. Biomark \nRes 9, 87 \n37. Singh, V ., Gera, R., Prasad Purohit, M., Patnaik, S., and Ghosh, D. (2017) Fluorometric \nEstimation of Glutathione in Cultured Microglial Cell Lysate. Bio Protoc 7, e2304 \n38. Downs, C. A., Kreiner, L., Zhao, X. M., Trac, P., Johnson, N. M., Hansen, J. M., \nBrown, L. A., and Helms, M. N. (2015) Oxidized glutathione (GSSG) inhibits epithelial \nsodium channel activity in primary alveolar epithelial cells. Am J Physiol Lung Cell \nMol Physiol 308, L943-952 \n39. Franco, R., and Cidlowski, J. A. (2012) Glutathione efflux and cell death. Antioxid \nRedox Signal 17, 1694-1713 \n40. Grosvenor, A. J., Deb-Choudhury, S., Middlewood, P. G., Thomas, A., Lee, E., Vernon, \nJ. A., Woods, J. L., Taylor, C., Bell, F. I., and Clerens, S. (2018) The physical and \nchemical disruption of human hair after bleaching - studies by transmission electron \nmicroscopy and redox proteomics. Int J Cosmet Sci 40, 536-548 \n41. Hou, Z., Wydorski, P. M., Perez, V . A., Mendoza-Oliva, A., Ryder, B. D., Mirbaha, H., \nKashmer, O., and Joachimiak, L. A. (2021) DnaJC7 binds natively folded structural \nelements in tau to inhibit amyloid formation. Nat Commun 12, 5338 \n42. Urushitani, M., Kurisu, J., Tateno, M., Hatakeyama, S., Nakayama, K., Kato, S., and \nTakahashi, R. (2004) CHIP promotes proteasomal degradation of familial ALS-linked \nmutant SOD1 by ubiquitinating Hsp/Hsc70. J Neurochem 90, 231-244 \n43. Choi, Y . N., Lee, S. K., Seo, T. W., Lee, J. S., and Yoo, S. J. (2014) C-Terminus of \nHsc70-interacting protein regulates profilin1 and breast cancer cell migration. Biochem \nBiophys Res Commun 446, 1060-1066 \n44. Sul, J. W., Park, M. Y ., Shin, J., Kim, Y . R., Yoo, S. E., Kong, Y . Y ., Kwon, K. S., Lee, \nY . H., and Kim, E. (2013) Accumulation of the parkin substrate, FAF1, plays a key role \nin the dopaminergic neurodegeneration. Hum Mol Genet 22, 1558-1573 \n45. Marszalek, J., De Los Rios, P., Cyr, D., Mayer, M. P., Adupa, V ., Andréasson, C., \nBlatch, G. L., Braun, J. E. A., Brodsky, J. L., Bukau, B., Chapple, J. P., Conz, C., \nDementin, S., Genevaux, P., Genest, O., Goloubinoff, P., Gestwicki, J., Hammond, C. \nM., Hines, J. K., Ishikawa, K., Joachimiak, L. A., Kirstein, J., Liberek, K., Mokranjac, \nD., Nillegoda, N., Ramos, C. H. I., Rebeaud, M., Ron, D., Rospert, S., Sahi, C., Shalgi, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /25 \nR., Tomiczek, B., Ushioda, R., Ustyantseva, E., Ye, Y ., Zylicz, M., and Kampinga, H. \nH. (2024) J-domain proteins: From molecular mechanisms to diseases. Cell Stress \nChaperones 29, 21-33 \n46. Farhan, S. M. K., Howrigan, D. P., Abbott, L. E., Klim, J. R., Topp, S. D., Byrnes, A. \nE., Churchhouse, C., Phatnani, H., Smith, B. N., Rampersaud, E., Wu, G., Wuu, J., \nShatunov, A., Iacoangeli, A., Al Khleifat, A., Mordes, D. A., Ghosh, S., Consortium, A., \nConsortium, F., Project Min, E. C., Consortium, C. R., Eggan, K., Rademakers, R., \nMcCauley, J. L., Schule, R., Zuchner, S., Benatar, M., Taylor, J. P., Nalls, M., Gotkine, \nM., Shaw, P. J., Morrison, K. E., Al-Chalabi, A., Traynor, B., Shaw, C. E., Goldstein, D. \nB., Harms, M. B., Daly, M. J., and Neale, B. M. (2019) Exome sequencing in \namyotrophic lateral sclerosis implicates a novel gene, DNAJC7, encoding a heat-shock \nprotein. Nat Neurosci 22, 1966-1974 \n47. Ballinger, C. A., Connell, P., Wu, Y ., Hu, Z., Thompson, L. J., Yin, L. Y ., and Patterson, \nC. (1999) Identification of CHIP, a novel tetratricopeptide repeat-containing protein \nthat interacts with heat shock proteins and negatively regulates chaperone functions. \nMol Cell Biol 19, 4535-4545 \n48. Haertle, L., Buenache, N., Cuesta Hernández, H. N., Simicek, M., Snaurova, R., \nRapado, I., Martinez, N., López-Muñoz, N., Sánchez-Pina, J. M., Munawar, U., Han, \nS., Ruiz-Heredia, Y ., Colmenares, R., Gallardo, M., Sanchez-Beato, M., Piris, M. A., \nSamur, M. K., Munshi, N. C., Ayala, R., Kortüm, K. M., Barrio, S., and Martínez-\nLópez, J. (2023) Genetic Alterations in Members of the Proteasome 26S Subunit, AAA-\nATPase (PSMC) Gene Family in the Light of Proteasome Inhibitor Resistance in \nMultiple Myeloma. Cancers (Basel) 15 \n49. Kröll-Hermi, A., Ebstein, F., Stoetzel, C., Geoffroy, V ., Schaefer, E., Scheidecker, S., \nBär, S., Takamiya, M., Kawakami, K., Zieba, B. A., Studer, F., Pelletier, V ., Eyermann, \nC., Speeg-Schatz, C., Laugel, V ., Lipsker, D., Sandron, F., McGinn, S., Boland, A., \nDeleuze, J. F., Kuhn, L., Chicher, J., Hammann, P., Friant, S., Etard, C., Krüger, E., \nMuller, J., Strähle, U., and Dollfus, H. (2020) Proteasome subunit PSMC3 variants \ncause neurosensory syndrome combining deafness and cataract due to proteotoxic \nstress. EMBO Mol Med 12, e11861 \n50. Taylor, J. P., Brown, R. H., Jr., and Cleveland, D. W. (2016) Decoding ALS: from genes \nto mechanism. Nature 539, 197-206 \n51. Fujii, R., Okabe, S., Urushido, T., Inoue, K., Yoshimura, A., Tachibana, T., Nishikawa, \nT., Hicks, G. G., and Takumi, T. (2005) The RNA binding protein TLS is translocated to \ndendritic spines by mGluR5 activation and regulates spine morphology. Curr Biol 15, \n587-593 \n52. Schwartz, J. C., Ebmeier, C. C., Podell, E. R., Heimiller, J., Taatjes, D. J., and Cech, T. \nR. (2012) FUS binds the CTD of RNA polymerase II and regulates its phosphorylation \nat Ser2. Genes Dev 26, 2690-2695 \n53. Lagier-Tourenne, C., Polymenidou, M., Hutt, K. R., Vu, A. Q., Baughn, M., Huelga, S. \nC., Clutario, K. M., Ling, S. C., Liang, T. Y ., Mazur, C., Wancewicz, E., Kim, A. S., \nWatt, A., Freier, S., Hicks, G. G., Donohue, J. P., Shiue, L., Bennett, C. F., Ravits, J., \nCleveland, D. W., and Yeo, G. W. (2012) Divergent roles of ALS-linked proteins \nFUS/TLS and TDP-43 intersect in processing long pre-mRNAs. Nat Neurosci 15, \n1488-1497 \n54. Calvio, C., Neubauer, G., Mann, M., and Lamond, A. I. (1995) Identification of hnRNP \nP2 as TLS/FUS using electrospray mass spectrometry. Rna 1, 724-733 \n55. Schwartz, J. C., Podell, E. R., Han, S. S., Berry, J. D., Eggan, K. C., and Cech, T. R. \n(2014) FUS is sequestered in nuclear aggregates in ALS patient fibroblasts. Mol Biol \nCell 25, 2571-2578 \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint \n\n Lyu /26 \n56. Rezvykh, A. P., Ustyugov, A. A., Chaprov, K. D., Teterina, E. V ., Nebogatikov, V . O., \nSpasskaya, D. S., Evgen'ev, M. B., Morozov, A. V ., and Funikov, S. Y . (2023) \nCytoplasmic aggregation of mutant FUS causes multistep RNA splicing perturbations \nin the course of motor neuron pathology. Nucleic Acids Res 51, 5810-5830 \n57. Takanashi, K., and Yamaguchi, A. (2014) Aggregation of ALS-linked FUS mutant \nsequesters RNA binding proteins and impairs RNA granules formation. Biochem \nBiophys Res Commun 452, 600-607 \n58. The PyMOL Molecular Graphics System, Version 3.0 Schrödinger, LLC.  \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted August 24, 2025. ; https://doi.org/10.1101/2025.08.20.671413doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}