The Degradation Pathway of COP9 Signalosome–Cullin-RING Ubiquitin Ligase Complexes via Autophagy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This work is well written, clear even for a non-specialist of this topic and useful for the readership overall. I don't have major issues, but only few minors:

• minor language editing is needed, especially when longer sentences are included

• Authors used many abbreviations, although these appear properly described at the first citation in the text, it is a bit difficult to follow the narrative overall. It may be good to introduce a figure describing the main players for instance in the introduction.

• Figure 5 should be different than the one reported in ref 20 (doi.org/10.1042/EBC20253018)

Author Response

Reviewer 1

Point-by-point response for manuscript biomolecules-4050964

We wish to thank the reviewer for his constructive critique of our work, which has been instrumental in improving the quality of the present manuscript! In the following, we provide a point-by-point response to the reviewers’ concerns and explain how we have addressed them.

• minor language editing is needed, especially when longer sentences are included

We edited languages especially in longer sentences:

Line 47, lines 48-50, lines 346-348 and more. Changes are labeled in red.

• Authors used many abbreviations, although these appear properly described at the first citation in the text, it is a bit difficult to follow the narrative overall. It may be good to introduce a figure describing the main players for instance in the introduction.

It is difficult to explain the abbreviations already in the introduction. In revised Figure 5 legend, we explain again as many abbreviations as possible. There is an explanation of abbreviation in lines 412 – 412.

• Figure 5 should be different than the one reported in ref 20 (doi.org/10.1042/EBC20253018)

We modified Figure 5 to make it different from that in ref 20 and to shape it more appropriately according to the text of the manuscript.

 

Reviewer 2 Report

Comments and Suggestions for Authors

In manuscript : 'The Degradation Pathway of COP9 Signalosome-Cullin-RING- 2
Ubiquitin Ligase Complexes via Autophagy' the authors provided a possible breakdown pathway of the said protein complex using comprehensive confocal microscopy based studies. Overall their approach looks rigorous and the studies are in line with the hypothesis. However, I have comments on figure 2E:

1. The signal for lamp2 is weak in wild type
2. Magnifications of control vs csn5i3 seem to differ
3. Lamp2 signals seem to bleed into the nucleus , which indicates non specificity
4. Csn7a association with lamp2 seems weak, both antibodies need better specificity, a Pearson correlation coefficient to indicate association is recommended.                                                      The authors should provide explanations on 1 to 4. Alongside, the paper can be strengthened by adding some structural details about the association and dissociation of the protein protein complexes using respective structures (if they are available in PDB or if predicted 'reliable' structures are available in AlphaFold database).

Author Response

Reviewer 2

Point-by-point response for manuscript biomolecules-4050964

We wish to thank the reviewer for his constructive critique of our work, which has been instrumental in improving the quality of the present manuscript! In the following, we provide a point-by-point response to the reviewers’ concerns and explain how we have addressed them.

1. The signal for lamp2 is weak in wild type

 

Reviewer 2 is right, anti-Lamp2 is a weak antibody. However, the original Figure 2E shows a clear signal. Statement of Figure 2E is the nuclear export of CSN-CRL complexes in presence of CSN5i-3 shown by CSN7A antibody in LiSa-2 cells. This is clearly confirmed in HeLa cells in Figure 3C.

2. Magnifications of control vs csn5i3 seem to differ

 

The magnification in control and in CSN5i-3 is the same. In control cells, two cells are fused which brings you to the assumption.

3. Lamp2 signals seem to bleed into the nucleus, which indicates non specificity

 

You are right, Lamp2 is a weak antibody but its bleeding into the nucleus is minimal in Figure 2E in presence of CSN5i-3.

4. Csn7a association with lamp2 seems weak, both antibodies need better specificity, a Pearson correlation coefficient to indicate association is recommended. The authors should provide explanations on 1 to 4.

 

In Figure 2E shows that nuclear CSN-CRL complexes are spelled into the cytosol. Lamp 2 antibody indicates the localization of lysosomes but not the association with lysosomes. Association of CSN-CRL complexes and lysosomes is only clearly shown in Figure 3C, D and Figure S1C in presence of CQ. In future manuscript we will take into consideration to calculate Pearson correlation coefficient.

 

5. Alongside, the paper can be strengthened by adding some structural details about the association and dissociation of the protein protein complexes using respective structures (if they are available in PDB or if predicted 'reliable' structures are available in AlphaFold database).

This is a very interesting idea to show structural details of protein-protein complexes. We will use it in future manuscripts.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript presents an interesting idea about how the CSN-CRL assemblies are cleared by autophagy. This happens either by bulk autophagy after serum starvation or by a more selective macroautophagic route upon CSN inhibition.

However, there are still several important conclusions that need to be proven before we can accept them. In particular:

The manuscript uses 1 mM CSN5i-3 and 30 mM chloroquine (CQ) in important experiments. These concentrations are unusually high for cell-based signaling/trafficking studies and could lead to non-specific stress responses, membrane/lysosomal dysfunction, or a collapse of global proteostasis. Check that these are unit errors (e.g. µM vs. mM) and correct them consistently.

The main evidence for autophagy is that CQ stops the loss of CSN subunits/cullins and that CSN components are found in the same place as "ATG8"/RAB7 vesicles. This supports the idea of a connection between autophagy and lysosomes, but it does not yet prove that macroautophagic flux is the mechanism. Add measurements of how well the cells are working (LC3 processing, p62/SQSTM1 turnover) with and without lysosomal blockade, plus appropriate positive controls.

The Discussion says that the process of CUL3/CUL4 self-ubiquitylation is what triggers selective macroautophagy, and it references experiments using HA-ubiquitin transfection and HA-IP. As things stand, it is not yet clear that the ubiquitin signal comes from CUL3/CUL4A themselves, rather than other species that go with them, or adaptor proteins. First, perform denaturing IP of CUL3 and CUL4A. Then, use anti-ubiquitin (or the reciprocal strategy) to establish direct modification. Test how much the CRL catalytic machinery depends on "self-ubiquitylation", which is a key part of the model.

The text says that there is a link between these structures and the process of cell death. At the moment, these statements are only qualitative. Add colocalisation quantification (Pearson/Manders) and include per-cell statistics.

Provide exact n for each quantified panel, define “independent replicate,” and show individual data points where feasible.

Author Response

Reviewer 3

Point-by-point response for manuscript biomolecules-4050964

We wish to thank the reviewer for his constructive critique of our work, which has been instrumental in improving the quality of the present manuscript! In the following, we provide a point-by-point response to the reviewers’ concerns and explain how we have addressed them.

• The manuscript uses 1 mM CSN5i-3 and 30 mM chloroquine (CQ) in important experiments. These concentrations are unusually high for cell-based signaling/trafficking studies and could lead to non-specific stress responses, membrane/lysosomal dysfunction, or a collapse of global proteostasis. Check that these are unit errors (e.g. µM vs. mM) and correct them consistently.

We checked our manuscript and could not find 1 mM CSN5i-3 or 30 mM chloroquine. These might be conversion errors. We used 1 μM CSN5i-3 and 30 μM chloroquine.

• The main evidence for autophagy is that CQ stops the loss of CSN subunits/cullins and that CSN components are found in the same place as "ATG8"/RAB7 vesicles. This supports the idea of a connection between autophagy and lysosomes, but it does not yet prove that macroautophagic flux is the mechanism. Add measurements of how well the cells are working (LC3 processing, p62/SQSTM1 turnover) with and without lysosomal blockade, plus appropriate positive controls.

In Supplementary materials we confirmed in another cell line (HeLa cells) that CSN components colocalize with RAB7 and that the process is inhibited by chloroquine (CQ) (Figure S1C). We have already shown that LC3/ATG8 is accumulated in presence of CQ in western blot (Figure 3A) and in immunofluorescence (Figure 3C) indicating the cells are working well. We performed new experiments as you suggested and confirmed in Figure S2B with p62 and again with LC3/ATG8 antibody (accumulation of p62 and conjugated LC3) that cells are working well under this condition. There was no cell death in our experiments. We added new text on line 239 in red.

• The Discussion says that the process of CUL3/CUL4 self-ubiquitylation is what triggers selective macroautophagy, and it references experiments using HA-ubiquitin transfection and HA-IP. As things stand, it is not yet clear that the ubiquitin signal comes from CUL3/CUL4A themselves, rather than other species that go with them, or adaptor proteins. First, perform denaturing IP of CUL3 and CUL4A. Then, use anti-ubiquitin (or the reciprocal strategy) to establish direct modification. Test how much the CRL catalytic machinery depends on "self-ubiquitylation", which is a key part of the model.

In Figure 4C, HA-tagged ubiquitin immunoprecipitation shows clear specific bands of CUL3 and CUL4A only in presence of CSN5i-3, which run higher than normal CUL3 and CUL4. This bands seems to be ubiquitylated CUL3 and CUL4. In addition, if you inhibit the CRL catalytic machinery by MLN4924 (Figure 4E), the process is blocked. Moreover, we did reciprocal IP in Figure 2B. Immunoprecipitation of CUL3 shows a clear modified CUL3 band only in presence of CSN5i-3. As you recommended, we performed CUL4A immunoprecipitation under denaturing conditions and results are presented in Figure S2A. Data shows direct modification of CUL4A with Ub. Text was changed from line 292 and in Material and methods from line 97 in red.  

• The text says that there is a link between these structures and the process of cell death. At the moment, these statements are only qualitative. Add colocalization quantification (Pearson/Manders) and include per-cell statistics.

A quantification of autophagic degradation is given in Figure 1B, Figure 3B, and Figure 4F. The quantification of immunofluorescence microscopic images would go beyond the scope of this manuscript. However, it is a good indication for future projects.

• Provide exact n for each quantified panel, define “independent replicate,” and show individual data points where feasible.

The exact n for independent replicates is shown in figure legends.

 

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have presented results on autophagy-mediated degradation of CSN–CRL complexes, addressing following concerns can improve the paper:

- The mechanistic distinction between bulk autophagy (serum starvation) and selective macroautophagy (CSN5i-3 treatment) would benefit from clearer experimental separation, including quantitative flux assays ( LC3 turnover) to more rigorously demonstrate pathway specificity rather than reliance on inhibitor-based inference alone.

-The authors have presented results on cullin self-ubiquitylation as the trigger for selective macroautophagy, addressing following concerns can improve the paper: additional biochemical evidence clarifying the ubiquitin topology (mono- vs poly-ubiquitylation and linkage type) is needed to substantiate the proposed signaling role, as this distinction is central to autophagy receptor engagement and currently remains inferential.

-The authors have presented results on nuclear export and cytosolic degradation of intact CSN–CRL complexes, addressing following concerns can improve the paper: the manuscript would be strengthened by direct interrogation of the nuclear export mechanism (CRM1 dependence or mutational analysis of putative export signals) to move beyond descriptive fractionation and imaging data.

-The authors have presented results on autophagosome association via ATG8 and RAB7, addressing following concerns can improve the paper: identification or functional validation of the proposed “unknown autophagy receptors” is essential, as the absence of p62/TOLLIP interaction leaves a key mechanistic gap in how CSN–CRL complexes are selectively recognized by the autophagy machinery.

Author Response

Reviewer 4

Point-by-point response for manuscript biomolecules-4050964

We wish to thank the reviewer for his constructive critique of our work, which has been instrumental in improving the quality of the present manuscript! In the following, we provide a point-by-point response to the reviewers’ concerns and explain how we have addressed them.

-           The mechanistic distinction between bulk autophagy (serum starvation) and selective macroautophagy (CSN5i-3 treatment) would benefit from clearer experimental separation, including quantitative flux assays ( LC3 turnover) to more rigorously demonstrate pathway specificity rather than reliance on inhibitor-based inference alone.

Reviewer 4 is right; it would be a good idea to discriminate bulk autophagy and selective macroautophagy by quantitative measurement of LC3 turnover. However, in our case we see clear differences between autophagic pathways by the ubiquitylation of complexes in selective macroautophagy (Figure 1A, Figure 4C), similar to proteaphagy.

-           The authors have presented results on cullin self-ubiquitylation as the trigger for selective macroautophagy, addressing following concerns can improve the paper: additional biochemical evidence clarifying the ubiquitin topology (mono- vs poly-ubiquitylation and linkage type) is needed to substantiate the proposed signaling role, as this distinction is central to autophagy receptor engagement and currently remains inferential.

This is a very interesting point raised by the reviewer. After immunoprecipitation of HA-Ub in Figure 4C and in new experiment shown in Figure S2A, we detect clear mono-ubiquitylated cullin bands according to the molecular weight. In our additional experiments (data not shown) we could not see any influence of CSN associated deubiquitylating enzymes, USP15, USP48, CYLD, on mono-ubiquitylation of cullins. Further investigations into this topic are the subject of our future research.

-           The authors have presented results on nuclear export and cytosolic degradation of intact CSN–CRL complexes, addressing following concerns can improve the paper: the manuscript would be strengthened by direct interrogation of the nuclear export mechanism (CRM1 dependence or mutational analysis of putative export signals) to move beyond descriptive fractionation and imaging data.

We did preliminary experiments with leptomycin B without a significant influence on nuclear export of CSN5i-3 inhibited CSN-CRLs (data not shown). So far, nothing is known about CSN-CRL nuclear exports in Mammalia. This is one of our most important future projects.

-           The authors have presented results on autophagosome association via ATG8 and RAB7, addressing following concerns can improve the paper: identification or functional validation of the proposed “unknown autophagy receptors” is essential, as the absence of p62/TOLLIP interaction leaves a key mechanistic gap in how CSN–CRL complexes are selectively recognized by the autophagy machinery.

This mechanistic gap is an interesting part of our future projects.

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Response accepted 

Reviewer 3 Report

Comments and Suggestions for Authors

the authors replied to all my concerns

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have improved the paper and it is ready to be published.