Excess Met1-ubiquitination leads to solid aggregate formation

  1. Stephanie Kaypee
  2. Minori Miyasaka
  3. Takuto Nakajima
  4. Hiromi Nishimura
  5. Jun-ichi Sakamaki
  6. Masaaki Komatsu
  7. Fumiyo Ikeda

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

The ubiquitin ligase HOIL-1 has a unique role in controlling quantity of Met1-linked/ linear ubiquitin chains in cells by coordinating action with the ubiquitin ligase HOIP. Both ligases are components of the Linear UBiquitin chain Assembly Complex (LUBAC), the only known ligase complex that is able to generate Met1-linked ubiquitin chains. Although importance of Met1-linked ubiquitin chains in inflammation and immunity is well established, physiological relevance of quantity of these chains remain unknown. Here, we demonstrate that cells expressing catalytically inactive HOIL-1 exhibited significantly higher numbers of α-Synuclein, tau, and amyloid beta aggregates. This phenotype is associated with a disruption in late-stage autophagic flux, wherein p62-positive aggregates fail to colocalize with lysosomal markers, leading to impaired clearance. Additionally, a biophysical transition in aggregate properties was observed in vitro , with mutant cells forming more rigid solid-like inclusions, shifting from dynamic and liquid-like condensates. Elevated Met1-linked ubiquitin chains, either through HOIL-1 catalytic inactivation or knockdown of the Met1-linked chain-specific deubiquitinase OTULIN, phenocopied the defects in aggregate clearance. These findings reveal a critical role of HOIL-1 catalytic activity in modulating aggregate clearance through autophagy and maintaining the quantity of Met1-ubiquitin chains, highlighting HOIL-1 as a key factor in proteostasis in neurodegenerative diseases.

Highlights

  • HOIL-1 catalytic activity prevents neurodegenerative protein aggregate accumulation

  • Inactive HOIL-1 impairs late-stage autophagic clearance of protein aggregates

  • Loss of HOIL-1 shifts aggregates to rigid, solid-like states

  • Excess Met1-ubiquitin chains drive aggregate solidification

Article activity feed

  1. Review Commons

    Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    Summary:

    The study 'Excess Met1-ubiquitination leads to solid aggregate formation' by Kaypee et. al suggests a previously unrecognised role for the E3 ligase HOIL-1 in clearing protein aggregates via autophagy (e.g. aggrephagy). In their model, toxic protein aggregates in cells are modified with ubiquitin chains, including M1-linked Ub-chains catalysed by LUBAC (of which HOIL-1 is a component). The HOIL-1 ubiquitin signal is posited to induce trafficking of aggregates to lysosomes for subsequent clearance. However, when HOIL-1 is inactive (catalytic C460A mutation), the pathway is interrupted. As a result, protein aggregates fail to clear, they increase in size and shift their biophysical properties from liquid-like to more rigid, insoluble aggregates. The authors explain their observations by an increasing amount of M1-linked chains on protein aggregates, which occur as a result of 'unrestrained HOIP activity' due to HOIL-1 inactivity (based on previous work). Increasing amounts of M1-chains are posited to promote aggregate formation, aggregate growth, and prevent clearance.

    The major claims made in this manuscript are the following:

    1. Following the induction of protein aggregate formation (e.g. alpha-synuclein, tau, beta-amyloid, p62 bodies), cells that express catalytically inactive HOIL-1 fail to clear protein aggregates and end up with more, larger and more rigid protein aggregates compared to cells that express WT HOIL-1.
    2. The observations made in 1. are due to disruptions in late-stage autophagic flux. While aggregates in cells that express WT HOIL-1 co-localise with autophagy and lysosomal markers, aggregates in cells that express mutant HOIL-1 show autophagic, but not lysosomal markers.
    3. An increase in M1-chains (either due to HOIL-1 inactivity or due to OTULIN knockout) is believed to be the cause for claims 1. and 2.

    Main methodologies used:

    The authors use two cellular systems. The first one is SH-SY5Y cells in which either WT or mutant HOIL-1 are transiently overexpressed (via the pcDNA3.1 plasmid), and physiologically important aggregates (Tau, Abeta, asyn) are induced. The second cellular system is MEF cells in which either WT or mutant HOIL-1 are endogenously expressed; in these cells aggregates are formed crudely through disruption of ribosomal translation. It is questionable if both systems can be compared. Aggregate formation is mainly monitored and quantified via fluorescent microscopy in both fixed and live cells, or via sucrose gradient fractionation to separate soluble and insoluble (=aggregate) fractions. The rigidity of protein aggregates is analysed in cells via FRAP, size and circularity measurements and 1,6-HD treatment, or in vitro after aggregate formation assays via size and circularity measurements. The observations are on the whole interesting, though the authors fail to discuss their data in light of previously published work. For example, HOIL-1 KO and KI animals were shown to feature polyglucososan bodies in brain, which is not mentioned. Also, McCrory et al on HOIL-1 chain types is not cited but seems relevant (Figure S4). Yet, the manuscript reports a number of interesting findings, more or less coherently, most useful for scientists embedded in current ubiquitin, autophagy, and LLPS fields. These reviewers believe that this manuscript will make a lot of sense in due course, and be well cited for a first description of the role of HOIL-1 in cellular quality control processes. A number of improvements seem required to consolidate the findings, and improve readability and impact.

    Major:

    1. Figure 1A-C: The authors transiently overexpress either WT HOIL-1L or catalytically inactive (C460A) HOIL-1 in SH-SY5Y cells, then induce and compare the formation of protein aggregates (alpha-synuclein, tau, amyloid-beta) in those cells over 72 h. More cells with aggregates were found in cells that overexpressed mutant HOIL-1L. While these findings are interesting, the cellular system used is artificial due to the transient overexpression of HOIL-1 (in presence of endogenous HOIL-1). Crucial controls are missing:

    a. Adding a condition in which no protein is overexpressed, for example via an empty pcDNA3.1 or GFP only vector. This would help ruling out secondary effects due to the transient overexpression. It would also allow to monitor whether the same amount of aggregates form in the empty ctr compared to when WT HOIL-1 is overexpressed.

    b. Figure legends and raw data points (?) in graphs do not match. The graphs show dubious statistics from 2-3 grey dots, while the figure legend refers to n=100 cells etc. This needs to be fixed.

    c. Showing Western Blots of HOIL-1, to better understand the levels of endogenous HOIL-1 vs overexpressed HOIL-1 in these cells, and to compare overexpression levels between WT and mutant HOIL-1.

    d. The study would also improve by western blotting and IF staining for other LUBAC components such as HOIP and SHARPIN. Do alpha-synuclein aggregates in both WT and mutant conditions co-localise with the other LUBAC components, and are there any differences between WT and mutants. This would further help strengthening the claims made in Figure S1A: '...suggesting that LUBAC is recruited to or retained within α-Synuclein aggregates.' And in the discussion: 'we found that LUBAC components were sequestered in aggregates, as evidenced by microscopy and gradient fractionation of soluble and insoluble proteins, confirming the direct involvement of LUBAC in aggregate processing.'

    1. Figure 2A-F: The authors change to a genetic-derived system (comparing endogenously expressed HOIL-1 WT with mutant HOIL-1 based on MEF cells from their mouse models). However, they use puromycin to produce aggregates from random protein homeostasis defects, which yes leads to aggregates, but is not as nice as the induced generation of neiurodegeneration-relevant aggregates. It was observed that after 2 h of puromycin treatment, cells accumulate p62-positive protein aggregates, and in during recovery (2 h washout), the aggregates in the HOIL1 mutant cells outgrow the aggregates in the WT HOIL1.

    a. However, the authors claim that: 'While Hoil-1+/+ MEFs efficiently cleared puromycin-induced p62 bodies,...', which is not supported by the data shown here. When comparing WT in panel C with WT in panel E, it becomes evident that the average number of p62 puncta before and after recovery is the same (around 5 puncta/cell in both pre and post washout conditions). A similar observation can be made for the mutant (around 12 puncta/cell in both pre and post washout conditions). Can the authors please amend their claims, or comment and perform a direct statistical comparison between the pre and post recovery conditions to test for clearance of p62 puncta in the WT after puromycin washout.

    b. The authors state that: 'These findings indicate that although HOIL-1 catalytic activity is dispensable for the initial formation of puromycin-induced aggregates, it is essential for their subsequent clearance.'

    As long as clearance of p62 bodies in the WT is not clearly shown, the second part of this sentence should be amended/removed.

    c. The experiments shown would improve by adding a t = 0 condition. How many p62 granules are present before puromycin treatment? Is there already a basal difference between WT and mutant HOIL-1L cells?

    1. Figure S1A: The authors claim that other LUBAC components co-localise to protein aggregates, based on sucrose gradient fractionation and the presence of the respective proteins in the insoluble fractions. Could the authors perform IF and stain for other LUBAC components (SHARPIN and/or HOIP) in their MEF cell system to directly validate this claim?
    2. Figure 3G-H: The authors created a GFP-mCherry-p62 reporter system in both their WT and mutant HOIL-1 MEF cells and performed live cell imaging following puromycin treatment, which allows monitoring of both aggregate formation and loss of GFP signal due to the acidic lysosomal localisation. Excitingly, the ratio of GFP/mCherry in the later timepoints is reduced in the WT compared to mutant HOIL-1, indicating that HOIL-1 activity is required to traffic p62 bodies to lysosomes.

    a. In panel G, a surprisingly large amount of p62 granules are present at t0, which (according to the relevant method section) is the time of puromycin treatment. This observation can be made for both WT and mutant cells. After 80 min of puromycin treatment in the WT, the majority of these puncta are cleared. Can the authors please comment on this high amount of p62 granules at t0 (before the effects of puromycin? And also on the observation that after 80 min there are now less granules than before puromycin? In case that t0 indicates the time of puromycin washout rather than puromycin addition, could this please be clarified in the methods or figure legend?

    b. Panel H would improve by adding the quantifications for t=0 (or ideally for all the time points).

    c. Fig S3C-D: Same comments as before but for GFP-mCherry-LC3.

    1. Fig. S4B and Fig 4A-B: The authors state that circular aggregates are more soluble and have more LLP characteristics, whereas non-circular aggregates are less soluble and have more aggregate-like characteristics. However, the aggregates shown in Fig. S4B are un-circular but easily dissolve in response to 1,6-HD treatment, which seems contradictory. On the other hand, the aggregates shown in Fig 4B in HOIL-1 mutant cells appear much rounder than the ones in S4B, but do not dissolve in response to 1,6-HD treatment. Can the authors please comment on these discrepancies?
    2. Fig. 4E-G. Here the authors suddenly switch to an in vitro aggregate-formation assay using mCherry-p62. In-vitro M1-chain reactions with either WT LUBAC or LUBAC with mutant HOIL-1L, together with the respective M1-chain reaction product, are added. This is not clear from the figure, and a schematic, as well as a gel (Coomassie) should be included to show component purity and indicate the biochemical in vitro nature of the experiment. It is good to have this breadth of methods, but does not help in the presentation if all figures look alike.

    a. The key difference to the cellular situation is p62 aggregates are not directly ubiquitinated here, and instead ubiquitin chains are (non-covalently) added to samples. Can the authors please make this important difference clearer in their text? Why not directly ubiquitinate mCherry-p62 via LUBAC (WT vs mutant HOIL-1) and then perform an aggregation assay on the reaction product?

    b. Can the authors please clarify whether the reaction was inactivated prior to addition to the aggregate-formation assay? If not, the enzymes might still be active at the point of aggregate formation, and the observed effects might be influenced by enzymatic activities and not only the presence of different M1-chain architectures.

    1. Fig. 5B-C: The M1-specific DUB OTULIN is knocked down (again, cells) to increase the overall amount of M1-linked Ub-chains present in cells. P62 aggregate formation is induced and the authors claim that the increase in M1-chains influences aggregate size. This claim would be strengthened if it was directly shown that M1-chains form on p62 aggregates in this assay, for example via IF using an M1-antibody (and potentially a total ubiquitin antibody). This would also enable to directly compare the abundance of M1-chains between conditions (ctr vs Otulin, WT vs HOIL-1L mutant).

    Minor:

    1. Figure 1 D: The authors state that 'Notably, HOIL-1 C460A was detected within these structures, as demonstrated by its colocalization with tau aggregates' and show a co-localisation comparison between WT and mutant HOIL-1L. This sentence implies that WT HOIL-1 was not detected in aggregates, however the chosen image of the WT cells does not show any obvious tau aggregate, even though aggregates were induced in this condition according to 1A-C. The better comparison would be to pick an image that includes a tau aggregate. Moreover, this experiment would benefit from quantification, calculating the percentage of total aggregates that co-localise with WT HOIL-1L vs with mutant HOIL-1L.
    2. Figure S1A: The authors state that 'Cells expressing HOIL-1 C460A displayed a pronounced accumulation of high-molecular-weight α-Synuclein species in the insoluble pellet fraction.' While the difference seen by Western Blot is apparent and seems to match Fig 1 A-C, it is not very strong. Moreover, the relevant comparison (WT vs mutant) is made between two different blots/membranes, and it is difficult to assess equal input solely based on the TCL lane. This experiment could be improved by normalising samples (for example via BCA) and by loading and imaging the two conditions (or at least the TCL and pellet samples) on the same membrane. The authors also state that: 'LUBAC subunits, HOIP, HOIL-1, and SHARPIN, were also enriched in this fraction, suggesting that LUBAC is recruited to or retained within α-Synuclein aggregates.' Both Sharpin and HOIL-1 seem to be present in similar levels in the pellet fractions of WT and mutant HOIL-1. Overall HOIP levels seem to be significantly increased in the mutant over the WT (see TCL lane), and to a similar level in the pellet fraction. It would be great if the authors could include these observations in their interpretation.
    3. Figure 2G: Similar to before, this experiment would improve if the authors could find a way to normalise samples between conditions prior to sucrose gradient fractionation or have the most relevant samples on the same blot. It is challenging to properly interpret the results while the bands in the total cell lysate (TCL) lane do not have similar intensities between samples. A blot in which only the TCL and the pellet samples of all conditions were loaded onto the same gel would solve this, allowing for a better comparison between conditions. Based on what is shown in panel G, the authors should amend their claim: 'Consistent with microscopic observations, denser fractions from Hoil-1C458A/C458A MEFs contained increased signals of p62 specifically during the recovery phase (Figure 2G).' It is not apparent that more p62 is present in the insoluble fractions of mutant HOIL-1 cells after puromycin treatment. The band intensities look very similar (This is different for the recovery condition, which shows a strong difference, as stated).
    4. Figure S2: Similar to before, the authors induce protein aggregate formation and compare cells endogenously expressing WT vs mutant HOIL-1L. The size of aggregates increases in mutant cells under proteotoxic stress. What happens to the number of aggregates per cell in these conditions? Does it also increase, or is it just the size?
    5. Fig. S4A: Here the authors analyse the circularity of p62 aggregates in HOIL-1L mutant cells after recovery from puromycin treatment. This experiment would improve if the same analysis could be performed for the WT cells and for the pre-recovery condition (under the condition that large enough granules are present), allowing to make a comparison between WT and mutant, as well as between pre- and post-recovery.
    6. Fig. 4H-J: The authors use a p62 mutant that is known for its enhanced ubiquitin affinity, repeat Fig. 4E-G and state: 'These aberrant condensates were similar to those observed in a reaction using wild-type p62.' Can the authors please comment on what they conclude from this similarity and why this experiment was performed? A quantitative comparison (condensate size and circularity) between WT p62 and mutant p62 may further be useful here.
    7. Fig. 5A: Here the authors pulldown M1-linked Ub-chains in WT vs mutant HOIL-1 cells, with or without puromycin-induced aggregate formation. More M1-chains are observed in mutant HOIL-1 cells under puromycin treatment, but the difference is very subtle. The conclusions drawn from this experiment could be strengthened by including alternative methods, for example (if available) Ub-AQUA to measure the abundance of M1-chains, or using the M1-antibody for IF analysis.
    8. Fig. S5B: The described differences between puromycin treatment and untreated conditions are extremely subtle on the anti-Ub blot, and absent in the anti-Met1 blot. I recommend that the authors remove this sentence, based on the shown data: 'we observed a modest increase in the signal of Met1-linked ubiquitin chains after puromycin treatment'.
    9. Fig. 5F-G: The authors went back to their more-artificial system from Fig 1, in which HOIL-1L was transiently overexpressed (WT or inactive mutant) in SH-SY5Y cells, alongside alpha-synuclein aggregate formation.

    a. The claims made from this experiment would be stronger in the other cell system. Could OTULIN be transiently overexpressed in the MEF cells, to monitor the effect of aggregate formation and clearance. Again, staining aggregates with the M1-antibody would improve this experiment.

    b. The authors claim that HOIL-1 activity fine-tunes the function of HOIP within LUBAC (from discussion: 'This regulatory mechanism ensures that in the presence of functional HOIL-1, the overall quantity and potentially the architecture of Met1-linked ubiquitin chains are tightly controlled'). What is the quantity and architecture of M1-chains catalysed by LUBAC when HOIL-1 is very highly abundant, as it would be the case in this cellular overexpression system?

    1. Sentence structure: 'A comprehensive understanding of how Met1-linked ubiquitination, particularly through intricate regulation by Linear Ubiquitin Chain Assembly Complex (LUBAC) components, such as HOIL-1, influences aggregate dynamics and clearance; therefore, it is crucial to develop targeted therapeutic strategies against neurodegenerative proteinopathies'.

    Significance

    The conclusions drawn from this study are very intriguing and give LUBAC (and HOIL-1) a so far unrecognised role in the clearance of protein aggregates, which are a hallmark of several neurodegenerative diseases for which there are currently no cures. Some of the findings described in this manuscript have the potential to be of very high impact and interest to a broader community, in particular researchers interested in protein homeostasis, autophagy, ubiquitin biology and neurodegeneration. In fact, those findings might even expand to autophagic pathways that target other cargo than protein aggregates. Both the novelty aspect and the potential for translational/therapeutic applications comprise the major strength of this manuscript. However, multiple of the presented experiments are currently lacking crucial controls, show weak effect sizes or were performed in artificial settings that likely do not represent relevant in vivo conditions, overall weakening or not fully supporting the claims made. Consequently, further experiments, data re-analyses and validations were recommended to fully support all the claims made here.

    This review was written from the perspective of a researcher in the ubiquitin field.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary:

    In this manuscript, the authors show that the branching activity of the E3 ligase HOIL-1, a component of the LUBAC complex, contributes to the autophagic clearance of p62 bodies and protein aggregates. This activity is attributed to enhanced linear, unbranched ubiquitin chain formation by the second E3 ligase of LUBAC, called HOIP. The model systems employed are cell lines including MEFs expressing a catalytically dead version of HOIL-1. In addition, the authors perform in vitro reconstitution experiments with purified ubiquitin chains, the LUBAC complex and p62. The main message is that solid p62 bodies are poor substrates for autophagy and that linear, non-branched ubiquitin chains promote solidification. The mechanism remains unclear and some of the effects sizes are rather modest.

    Major comments:

    The key observations mentioned above are convincingly shown. Since the authors don't claim any detailed molecular mechanisms, the number of conclusions in this study are limited.

    Overall, the authors are quite careful regarding their conclusion, and therefore the ones that are made in this manuscript are generally well supported. The data regarding the clearance of the p62 bodies presented in Figure 3 should be backed up with additional data. The authors could add a macroautophagy inhibitors such as VPS34 IN1 and/or perform the clearance experiments in a ATG KO/KD cell line to corroborate the contribution of macroautophagy to the clearance. In addition, a proteasome inhibitor should be used for comparison.

    The expertise and resources for the experiments mentioned above are expected to be well within the authors' capacity and should be doable within a few weeks.

    Some of the effects sizes (e.g. Fig. 5 and S5) are very small and it is possible that some of them are below statistical significance if the number of replicates are increased.

    Minor comments:

    Figure 1D should be quantified, for example using PCC, Pearson correlation coefficient. Figure S1 should be quantified. Figure S3: It should be explained how the region for the profiles are shown were selected.

    It is suggested to include a scheme of the LUBAC complex and its E3 ligase activities in Figure 1A. This will make it more accessible for readers, who are not so familiar with this complex, in particular as HOIP and HOIL can be easily confused. The authors may also want to clarify this in the abstract.

    Significance

    As mentioned in the summary. The authors report the observation that an excess of linear ubiquitin chains produced by HOIP in the absence of HOIL-1 activity results in the solidification of p62 bodies and reduced clearance by autophagy. This observation is novel and will be interesting for the proteostasis field.

    This reviewer is expert in autophagy and protein degradation, but less so in the LUBAC complex.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Background

    GWAS analyses carried out several years ago identified over 40 genetic loci associated with increased risk of developing Alzheimers and other types of dementia. They included loci encoding the E3 ubiquitin ligase HOIL-1 and the protein SHARPIN, two of the three components of the Linear Ubiquitin Assembly Complex (LUBAC) (Bellenguez et al, 2022, cited in the paper). The third component of LUBAC, HOIP, is the only E3 ubiquitin ligase known to catalyse the formation of Met1-linked ubiquitin (also known as linear ubiquitin). HOIL-1 is one of the few E3 ubiquitin ligases that attaches ubiquitin to serine and threonine residues in proteins forming ester bonds (Kelsall et al. 2019) and has been reported to restrict the HOIP-catalysed formation of Met1-linked ubiquitin (Kelsall et al. 2019; Fuseya et al. 2020; Rodriguez Carvajal et al. 2021).

    Summary of Paper's findings:

    In this study the authors report that HOIL-1 catalytic activity prevents neurodegenerative protein aggregation (synuclein, tau, A) in the human SH-SY5Y neuroblastoma cell line or in mouse embryonic fibroblasts (MEFs) expressing a catalytically inactive mutant of HOIL-1. They argue that this is achieved by maintaining the dynamic, liquid-like properties of protein condensates through regulation of Met1-linked ubiquitin chain levels, thereby facilitating efficient clearance via the aggrephagy pathway. They report that loss of HOIL-1 activity leads to excess Met1-ubiquitylation that drives the transition to rigid, solid-like aggregates resistant to autophagic degradation. In support of this conclusion, they also report that the siRNA knock-down of Otulin, a deubiquitylase that hydrolyses Met1-linked ubiquitin specifically, produces the same effect . The reframing of HOIL-1 as a key factor for fine-tuning ubiquitylation to maintain cellular protein homeostasis is an interesting development and the paper is generally well-written, focused and concise. Further work is required however, to fully convince these reviewers that the effects observed are entirely attributable to excess Met1-linked ubiquitylation, as claimed.

    Major comments:

    1. The causal link between elevated Met1-linked chains and solid-like aggregates in cells is the central claim of the paper. Throughout the study the authors use inactive HOIL-1 to enhance aggregate formation, which they attribute to increased Met1-linked ubiquitylation, something observed by themselves and others previously (Kelsall et al. 2019; Fuseya et al. 2020; Rodriguez Carvajal et al. 2021). However, the immunoblot for Met1-linked ubiquitin (Fig 5A) is not very convincing. In addition, the authors have not excluded the possibility that the loss of HOIL-1 enzyme activity has other effects on ubiquitylation, such as a change in the architecture of the ubiquitin chains caused by the absence of HOIL-1 catalysed formation of oxyester linkages. Many/most ubiquitin chains formed in cells contain more than one ubiquitin linkage type. It is therefore important for the authors to perform immunoblots for other ubiquitin linkage types, such as Lys63-linked ubiquitin, and to include these results in Fig 5.
    2. The reviewers also think that the authors' claims that the transition of condensate property is linked to elevated Met1-linked ubiquitin chains would be strengthened by performing the biophysical assays (FRAP and 1,6-hexanediol resistance) after Otulin knockdown/knockout (and ideally also with Otulin rescue). This will provide direct biophysical evidence linking Met-1 linked chain elevation to condensate liquidity and 1,6-HD sensitivity.
    3. The authors have not shown any evidence that Met1-linked chains are more enriched at the sites of protein aggregation. Would the authors be able to demonstrate direct spatial colocalization of Met1-Ub with the analysed aggregates?
    4. Do the authors know if the effects that they are seeing are general effects on autophagy? For example, is starvation-induced autophagy similarly impaired in the cells studied? A simple flux-style experiment looking at LC3-II levels and p62 with starvations vs puromycin (-/+ bafilomycin) would be informative here.

    Minor comments:

    1. The authors show that loss of HOIL-1 catalytic activity causes p62 bodies to transition from dynamic liquid-like states to rigid solid-like states and claim this as a more general effect on protein aggregates. But the study does not directly demonstrate a liquid-to-solid transition for the disease-relevant α-synuclein, tau, or Aβ aggregates, limiting the generalisation of the claim beyond p62 bodies. Perhaps the authors should modify the text to better reflect this (or, even better, consider treating α-synuclein/tau/Aβ aggregates with 1,6-hexanediol to measure the response). [optional]
    2. Given that the blots presented in Fig S1A appear to come from different membranes, and high-molecular-weight species of α-synuclein seem to exist in the insoluble pellet fraction of both WT and C460A expressing cells, the reviewers would caution against concluding anything about differences, which can only be assessed if the samples are run side-by-side on the same gel.
    3. The Methods section says that two different total ubiquitin antibodies were purchased, but which one was used in Figure 5 and other figures are not stated. Please clarify.
    4. On page 10 ABIN1 is mentioned but it is not mentioned that it is the protein product of the TNIP1 gene that is mentioned in the Introduction. This will confuse to many readers.
    5. 1st paragraph of Discussion line 5 from bottom:- change "oof" to "of".

    Significance

    Prior research has established that the components of LUBAC are recruited to, and are components of, protein aggregates. A link between LUBAC and selective autophagy has also been established previously. The significance of this paper is that it identifies the catalytic function of HOIL-1 as a brake on the activity of LUBAC in proteostasis. The reviewer and co-reviewer are not experts in autophagy or aggregate formation in dementia but, if those reviewers who are find the data presented in these areas to be convincing, then this paper may be the first to suggest a molecular mechanism by which polymorphism/mutation of HOIL-1 leads to increased formation of the aggregates observed in Alzheimer's and other dementias. The results presented in the paper also suggested that initial autophagosome recruitment to aggregates is intact but subsequent late-stage autophagy is impaired. Hence, the study begins to identify the specific step that fails. However, as the authors themselves acknowledge, validation of these potentially exciting findings using in vivo models of neurodegeneration should be the aim of future studies. The paper combines the molecular dissection of ubiquitin and autophagy pathways to understand the causes of neurodegenerative disease. The paper will therefore be of interest to a broad audience, encompassing both the basic research and clinical research communities.

    Reviewers field of expertise: Biochemists and cell biologists with an interest in ubiquitin and cell signalling.

  5. Version published to 10.64898/2026.01.20.700516 on bioRxiv