Regulated degradation of the inner nuclear membrane protein SUN2 maintains nuclear envelope architecture and function

  1. Logesvaran Krshnan
  2. Wingyan Skyla Siu
  3. Michael Van de Weijer
  4. Daniel Hayward
  5. Elena Navarro Guerrero
  6. Ulrike Gruneberg
  7. Pedro Carvalho

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    The paper provides fundamental information through the identification of an E3 ligase and kinase/phosphatase regulatory machinery that regulates the inner nuclear membrane protein SUN2 using a GFP-based assay. The data reveal a model involving extraction of ubiquitylation of SUN2 from the membrane by p97, which is an important contribution to the field. Although the biochemical evidence is solid on the GFP-tagged SUN2 protein, one question is the extent to which this pathway works on endogenous SUN2 and the extent to which this is a quality control mechanism for turnover of unassembled SUN2 or whether it acts on the fully assembled complex.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Nuclear architecture and functions depend on dynamic interactions between nuclear components (such as chromatin) and inner nuclear membrane (INM) proteins. Mutations in INM proteins interfering with these interactions result in disease. However, mechanisms controlling the levels and turnover of INM proteins remain unknown. Here, we describe a mechanism of regulated degradation of the INM SUN domain-containing protein 2 (SUN2). We show that Casein Kinase 2 and the C-terminal domain Nuclear Envelope Phosphatase 1 (CTDNEP1) have opposing effects on SUN2 levels by regulating SUN2 binding to the ubiquitin ligase Skp/Cullin1/F-Box βTrCP (SCF βTrCP ). Upon binding to phosphorylated SUN2, SCF βTrCP promotes its ubiquitination. Ubiquitinated SUN2 is membrane extracted by the AAA ATPase p97 and delivered to the proteasome for degradation. Importantly, accumulation of non-degradable SUN2 results in aberrant nuclear architecture, vulnerability to DNA damage and increased lagging chromosomes in mitosis. These findings uncover a central role of proteolysis in INM protein homeostasis.

Article activity feed

  1. Version published to 10.7554/elife.81573 on eLife
  2. Version published to 10.7554/elife.80445 on eLife
  3. eLife

    Author Response

    Reviewer #2 (Public Review):

    This study has investigated the pathway for degradation of the inner nuclear membrane protein SUN2. Based on earlier studies that had searched for bTrCP substrates and interactors, the authors postulated that SUN2 might be a target of this ligase. They found two potential bTrCP recognition sites and showed that the second of these, Site2, is important for SUN2 turnover. A phospho-mimetic mutant is turned over faster, and a phospho-resistant mutant is turned over slower. The degradation is slowed by inhibitors of Neddylation (and therefore, of Cullin Ring Ligases), inhibitors of p97, and proteasome inhibitors. Using a genetic screen, they find bTrCP, components of the Cullin ring ligase, p97, the proteasome, and subunits of CK2. They use inhibitors to show that CK2 is needed for maximal SUN2 degradation, and a phosphatase called CTDNEP1 antagonises CK2-mediated SUN2 degradation. Using a non-degraded variant of SUN2, the authors show that its overexpression can influence nuclear morphology and various nuclear functions. In sum, the authors outline a pathway for regulated degradation of the inner nuclear membrane SUN2.

    The study is generally sound in its logic, well written, and appropriately interpreted for the most part. The data are of high quality. The findings are new and will provide a foundation to now examine how LINC complex abundance is regulated. I have a number of suggestions for improvement, listed in order of importance. Only the first two should require any experimental work, and the second item is potentially optional depending on the authors' response. The remaining items can be handled with adjustments to the manuscript.

    1. It is surprising that nowhere in the paper is an experiment directly and rigorously establishing that bTrCP is required for SUN2 degradation. I realise this is quite plausible from the shown experiments, but it seems to be a rather glaring oversight (apologies if I have missed it somewhere). At present, the current evidence for its role is the similarity of Site2 to a bTrCP recognition motif, the physical interaction of SUN2 with bTrCP, and the modulation of this interaction by mutants intended to mimic or eliminate phosphorylation. The inhibitor experiment is not strong evidence because it inhibits all CRLs. I would therefore recommend, at the least, to present an experiment knocking down or out bTrCP2 (i.e., FBXW11, which nicely showed up in the genetic screen). This simple experiment could be included in the validation experiments in Fig. S4b. It would be worth also including FBXW1A for comparison, and if needed, the double-knockdown. This seems essential to complete the study.

    We thank the reviewer for suggestion. These experiments have now been included in the manuscript. For further information, see response to Editor’s point 1.

    1. The experiments with TBCA are not complemented with knockdown experiments of CK2 subunits. I realise CK2 is essential, but cells can evidently tolerate acute knockdown sufficiently well to do experiments given that this came up in the CRISPR screen. I would think such knockdown experiments would strengthen the argument and mitigate any concern about the off-target effects of TBCA. Kinase inhibitors are often only partially specific, so arguments about the involvement of any kinase are stronger if inhibitor studies are complemented with genetic perturbations.

    We thank the reviewer for suggestion. We have now included knockdown experiments with independent sgRNAs that validate our conclusion on the role of CK2 in SUN2 degradation (Figure Supplement 4C). In addition, we would like to point out that besides the essential CK2 regulatory subunit CSNK2B, our genome widescreen also identified the catalytic subunit CSNK2A2 (non-essential as it is redundant with CSNK2A1) (see Figure S3A). Considering that our library contains 4 sgRNAs per gene, this makes a total of 8 sgRNAs targeting subunits of CK2. Importantly no other kinase was identified in our screen. Moreover, TBCA is well established as a specific CK2 inhibitor. Altogether, these various observations make us quite confident that CK2 is the prime kinase controlling SUN2 stability.

    1. Lines 173-183: MLN4924 is used interchangeably with inhibition of SCFbTrCP. But MLN4924 is an NAE inhibitor that indirectly inhibits all CRLs. It seems premature to invoke SCFbTrCP as being involved because the experiments have not yet established a role for this specific CRL (see point 1 above). Instead, the conclusion should be that the data indicate a role for one or more CRLs. At this point in the narrative, the only evidence that bTrCP is involved is the sequence similarity of site1 and site2 to canonical bTrCP recognition sites. However, this is not enough evidence as no experiments knocking down or knocking out bTrCP, or experiments showing a physical interaction, have been presented yet. That comes in the subsequent section.

    We thank the reviewer for pointing this out. The text was modified and additional data on the depletion of βTrCP has been included in the revised manuscript to support our conclusions.

    1. Line 195 - At this point in the narrative, there is no evidence that SUN2 is ubiquitinated by SCFbTrCP. This needs to be rephrased. I would think one can conclude at this point that SUN2 is degraded by a pathway that relies on a CRL, p97, and the proteasome. The degradation is controlled by Site2, potentially by phosphorylation (again, this has not really been established at this point in the story, even if it seems plausible based on the mutagenesis).

    The sentence has been modified.

    1. I think the discussion needs to include some thoughts on what the authors believe happens to the rest of the SUN2 trimer or more broadly, the LINC complexes. In other words, what is the consequence of degrading a single protein of a much larger complex? In this vein, the model shows monomeric SUN2. Is it worth showing that it is part of a trimer and part of the LINC complexes? Regardless of how the authors depict the model, discussing this issue seems worthwhile.

    We thank the reviewer for the suggestion. We observe that the turnover rates of endogenous SUN2 is affected by the exogenous expression of SUN2 and primarily its derivatives Site 2A and Site 2D (Figure 1E). The effects are likely due to the assembly of trimers containing both endogenous and exogenous SUN2. This observation also suggests that degradation of one of the subunits in the trimer leads to the degradation of the other two. However, in the current manuscript we do not directly test or analyse these models or look at SUN2 complexes.

    1. Lines 225-226 - again, MLN4924 is not an inhibitor of SCFbTrCP, but rather a CRL inhibitor. The evidence for bTrCP being the key ligase is still missing at this point in the narrative.

    We now present evidence in an earlier figure that βTrCP is the F-Box involved in SUN2 degradation. In this context, the sentence appears correct.

    1. Fig. 5G is not especially convincing - to my eye, the effect on endogenous SUN2 is very similar to the effect on the transgene SUN2-site2A mutant, but simply a fainter exposure. Can the authors provide some numbers to allay this concern? It might well be that there is little difference between the behaviour of the endogenous and exogenous SUN2 in this experiment because they engage in heterotrimeric complexes. Also, why is the transgenic SUN2 not detected on the SUN2 blot? Would it not be evident at ~100 kD?

    We have consistently seen that SUN2 Site 2A is refractory to CTDNEP1 regulation. The blot has been replaced to better convey this result.

    The transgenic SUN2 is not detected in this blot because while the same cell lines were used for this experiment, to visualise the endogenous SUN2, doxycycline were not added to these cells. Thus, two sets of lysates were collected, one for cells that were treated with Doxycycline (transgene) and one without Doxycycline (endogenous). This is explained in the figure legend.

    1. In panel 1E, the heterologously expressed SUN2 protein has two bands, with the upper band being more readily degraded than the lower band in some cases. Is the upper band the phosphorylated product? Might be worth a comment if anything is known about what the two bands represent.

    We believe that the two bands do not correspond to different phosphorylated SUN2 forms. This is based on the analysis of SUN2 by SDS-PAGE in presence of Phostag reagent and the fact that two bands are seen both also for non-phosphorylatable and phospho-mimetic SUN2 derivatives. The appearance of two bands has been observed for other ERAD substrates characterized in our lab (for example Weijer et al. 2020) and appears to depend on the lysis conditions (see for example Figure 2 and 3).

    1. Worth mentioning in the main text that FBXW11 is bTRCP2. Also, it is worth noting whether bTRCP1 (FBXW1A) was a hit on the screen or not.

    Thanks for the suggestion. We have now included this information.

    Reviewer #3 (Public Review):

    The manuscript by Krshnan et al. reports a cellular mechanism akin to the endoplasmic reticulum-associated degradation (ERAD) that degrades SUN2, a nuclear inner membrane protein. The authors previously identified the Asi ubiquitin ligase complex that mediates the degradation of inner nuclear membrane proteins in budding yeast. In this manuscript, they identified the SCF β TrCP, and SCF as another ligase that regulates the ubiquitination and degradation of SUN2 in mammalian cells. The key findings include the identification of a substrate recognition motif that appears to undergo casein kinase (CK) dependent phosphorylation. Mutagenesis studies show that mutants defective in phosphorylation are stabilized while a phosphor-mimetic mutant is more unstable. They further show that the degradation of SUN2 requires the AAA ATPase p97, which allows them to draw the analogy between SUN2 degradation and Vpu-induced degradation of CD4, which occurs on the ER membrane via the ERAD pathway. Lastly, they show that the stability of endogenous SUN2 is regulated by a phosphatase and that over-expression of a non-degradable SUN2 variant disrupts nuclear envelope morphology, cell cycle kinetics, and DNA repair efficiency. Overall, the study dissects another example of inner nuclear envelope protein turnover and the involvement of a pair of kinase and phosphatase in this regulation. The data are of extremely high quality and the manuscript is clearly written. That being said, the following questions should be addressed to improve the robustness of the conclusions and to avoid potential misinterpretation of the data.

    1. Since SUN2 is normally incorporated into a SUN2-SYNE2-KASH2 LINC heterohexamer complex, the authors should be cautious with the use of over-expressed SUN2 in this study. Over-expressed SUN2 is expected to stay mostly as unassembled molecules and thus is likely degraded by a protein quality control mechanism that targets unassembled proteins. Consistent with this possibility, CK2 has been implicated in the regulated turnover of aggregation-prone proteins (Watabe, M. et al., JCS 2011). This mechanism would be potentially distinct from the one proposed for endogenous SUN2 degradation.

    We thank reviewers for the suggestion to provide further genetic evidence of the involvement of βTrCP1 and 2 F-box proteins in the degradation of SUN2. We now show that maximum stabilization of endogenous (Figure Supplement S4D) and transgenic (Figure S2) SUN2 is observed upon simultaneous depletion of βTrCP1 and βTrCP2 indicating that these F-Box proteins are redundant. Depletion of βTrCP1 alone did not impact SUN2 levels while depletion of βTrCP2 increased SUN2 steady state levels, with the effect being more pronounced for overexpressed SUN2. Depletion of other F-Box proteins did not affect SUN2 levels indicating that the effect observed for βTrCP1 is specific (Figure S2B). These results are in line with the results of our genome wide screen (Figure 4 and S3) and the literature. The differences in the effects of βTrCP1 and βTrCP2 depletion likely result from the relative abundance of the two F-Box proteins in the HEK cells used in this study.

    1. Certain conclusions appear to be an overstatement. This is particularly the case for the title, which implies that SUN2 is a protein that undergoes regulated turnover (under certain physiological conditions). Given that CK2 is a constitutive kinase and that the authors have not identified the conditions under which the activity of CTDNEP1 is regulated, it is premature to make such a conclusion.

    We disagree with the reviewer in this point. We present clear evidence that the turnover rate of SUN2 (both overexpressed and endogenous) is regulated by opposing kinase/phosphatase activities. This per se implies a mode of regulation. Similar kinase/phosphatase balances regulate a plethora of physiologic processes (from cell cycle progression to DNA repair) and the term “regulation” is commonly used in these contexts. We agree with the reviewer that upstream events controlling SUN2 remain elusive however, we do present evidence the balance of CK2 and CTDNEP1 activities regulate SUN2 degradation.

    1. Likewise, the demonstration of the impact of SUN2 accumulation on different cellular pathways mainly relies on the over-expression of a non-degradable SUN2 mutant. Whether similar defects could be seen when the degradation of endogenous SUN2 is blocked remains an open question.

    It would be great to gene edit the SUN2 locus to introduce the desired mutations. But as pointed out this is not trivial, in particular considering that the desired mutations would need to be introduced in both chromosomal copies.

  4. eLife

    Evaluation Summary:

    The paper provides fundamental information through the identification of an E3 ligase and kinase/phosphatase regulatory machinery that regulates the inner nuclear membrane protein SUN2 using a GFP-based assay. The data reveal a model involving extraction of ubiquitylation of SUN2 from the membrane by p97, which is an important contribution to the field. Although the biochemical evidence is solid on the GFP-tagged SUN2 protein, one question is the extent to which this pathway works on endogenous SUN2 and the extent to which this is a quality control mechanism for turnover of unassembled SUN2 or whether it acts on the fully assembled complex.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #3 agreed to share their name with the authors.)

  5. eLife

    Reviewer #1 (Public Review):

    This paper describes an "ERAD-like" pathway for the turnover of the SUN2 protein. In this pathway, ubiquitylation of SUN2 in the nucleoplasm by the SCFbTRCP ubiquitin ligase leads to extraction of the membrane protein by p97 for delivery to the proteasome. This process involves phosphorylation of non-canonical degrons on SUN2 by CK2, which was identified using a genome-wide crispr screening approach. The CTDNEP1 phosphatase acts to reverse phosphorylation and stabilize SUN2. Non-degradable forms of SUN2 promote altered nuclear architecture and a delay in double-strand break repair. The conclusions are based on strong biochemical and cell biological data. The paper sets the stage for further analysis of how defects in SUN2 degradation alter additional nuclear processes.

  6. eLife

    Reviewer #2 (Public Review):

    This study has investigated the pathway for degradation of the inner nuclear membrane protein SUN2. Based on earlier studies that had searched for bTrCP substrates and interactors, the authors postulated that SUN2 might be a target of this ligase. They found two potential bTrCP recognition sites and showed that the second of these, Site2, is important for SUN2 turnover. A phospho-mimetic mutant is turned over faster, and a phospho-resistant mutant is turned over slower. The degradation is slowed by inhibitors of Neddylation (and therefore, of Cullin Ring Ligases), inhibitors of p97, and proteasome inhibitors. Using a genetic screen, they find bTrCP, components of the Cullin ring ligase, p97, the proteasome, and subunits of CK2. They use inhibitors to show that CK2 is needed for maximal SUN2 degradation, and a phosphatase called CTDNEP1 antagonises CK2-mediated SUN2 degradation. Using a non-degraded variant of SUN2, the authors show that its overexpression can influence nuclear morphology and various nuclear functions. In sum, the authors outline a pathway for regulated degradation of the inner nuclear membrane SUN2.

    The study is generally sound in its logic, well written, and appropriately interpreted for the most part. The data are of high quality. The findings are new and will provide a foundation to now examine how LINC complex abundance is regulated. I have a number of suggestions for improvement, listed in order of importance. Only the first two should require any experimental work, and the second item is potentially optional depending on the authors' response. The remaining items can be handled with adjustments to the manuscript.

    1. It is surprising that nowhere in the paper is an experiment directly and rigorously establishing that bTrCP is required for SUN2 degradation. I realise this is quite plausible from the shown experiments, but it seems to be a rather glaring oversight (apologies if I have missed it somewhere). At present, the current evidence for its role is the similarity of Site2 to a bTrCP recognition motif, the physical interaction of SUN2 with bTrCP, and the modulation of this interaction by mutants intended to mimic or eliminate phosphorylation. The inhibitor experiment is not strong evidence because it inhibits all CRLs. I would therefore recommend, at the least, to present an experiment knocking down or out bTrCP2 (i.e., FBXW11, which nicely showed up in the genetic screen). This simple experiment could be included in the validation experiments in Fig. S4b. It would be worth also including FBXW1A for comparison, and if needed, the double-knockdown. This seems essential to complete the study.

    2. The experiments with TBCA are not complemented with knockdown experiments of CK2 subunits. I realise CK2 is essential, but cells can evidently tolerate acute knockdown sufficiently well to do experiments given that this came up in the CRISPR screen. I would think such knockdown experiments would strengthen the argument and mitigate any concern about the off-target effects of TBCA. Kinase inhibitors are often only partially specific, so arguments about the involvement of any kinase are stronger if inhibitor studies are complemented with genetic perturbations.

    3. Lines 173-183: MLN4924 is used interchangeably with inhibition of SCFbTrCP. But MLN4924 is an NAE inhibitor that indirectly inhibits all CRLs. It seems premature to invoke SCFbTrCP as being involved because the experiments have not yet established a role for this specific CRL (see point 1 above). Instead, the conclusion should be that the data indicate a role for one or more CRLs. At this point in the narrative, the only evidence that bTrCP is involved is the sequence similarity of site1 and site2 to canonical bTrCP recognition sites. However, this is not enough evidence as no experiments knocking down or knocking out bTrCP, or experiments showing a physical interaction, have been presented yet. That comes in the subsequent section.

    4. Line 195 - At this point in the narrative, there is no evidence that SUN2 is ubiquitinated by SCFbTrCP. This needs to be rephrased. I would think one can conclude at this point that SUN2 is degraded by a pathway that relies on a CRL, p97, and the proteasome. The degradation is controlled by Site2, potentially by phosphorylation (again, this has not really been established at this point in the story, even if it seems plausible based on the mutagenesis).

    5. I think the discussion needs to include some thoughts on what the authors believe happens to the rest of the SUN2 trimer or more broadly, the LINC complexes. In other words, what is the consequence of degrading a single protein of a much larger complex? In this vein, the model shows monomeric SUN2. Is it worth showing that it is part of a trimer and part of the LINC complexes? Regardless of how the authors depict the model, discussing this issue seems worthwhile.

    6. Lines 225-226 - again, MLN4924 is not an inhibitor of SCFbTrCP, but rather a CRL inhibitor. The evidence for bTrCP being the key ligase is still missing at this point in the narrative.

    7. Fig. 5G is not especially convincing - to my eye, the effect on endogenous SUN2 is very similar to the effect on the transgene SUN2-site2A mutant, but simply a fainter exposure. Can the authors provide some numbers to allay this concern? It might well be that there is little difference between the behaviour of the endogenous and exogenous SUN2 in this experiment because they engage in heterotrimeric complexes. Also, why is the transgenic SUN2 not detected on the SUN2 blot? Would it not be evident at ~100 kD?

    8. In panel 1E, the heterologously expressed SUN2 protein has two bands, with the upper band being more readily degraded than the lower band in some cases. Is the upper band the phosphorylated product? Might be worth a comment if anything is known about what the two bands represent.

    9. Worth mentioning in the main text that FBXW11 is bTRCP2. Also, it is worth noting whether bTRCP1 (FBXW1A) was a hit on the screen or not.

  7. eLife

    Reviewer #3 (Public Review):

    The manuscript by Krshnan et al. reports a cellular mechanism akin to the endoplasmic reticulum-associated degradation (ERAD) that degrades SUN2, a nuclear inner membrane protein. The authors previously identified the Asi ubiquitin ligase complex that mediates the degradation of inner nuclear membrane proteins in budding yeast. In this manuscript, they identified the SCF β TrCP, and SCF as another ligase that regulates the ubiquitination and degradation of SUN2 in mammalian cells. The key findings include the identification of a substrate recognition motif that appears to undergo casein kinase (CK) dependent phosphorylation. Mutagenesis studies show that mutants defective in phosphorylation are stabilized while a phosphor-mimetic mutant is more unstable. They further show that the degradation of SUN2 requires the AAA ATPase p97, which allows them to draw the analogy between SUN2 degradation and Vpu-induced degradation of CD4, which occurs on the ER membrane via the ERAD pathway. Lastly, they show that the stability of endogenous SUN2 is regulated by a phosphatase and that over-expression of a non-degradable SUN2 variant disrupts nuclear envelope morphology, cell cycle kinetics, and DNA repair efficiency. Overall, the study dissects another example of inner nuclear envelope protein turnover and the involvement of a pair of kinase and phosphatase in this regulation. The data are of extremely high quality and the manuscript is clearly written. That being said, the following questions should be addressed to improve the robustness of the conclusions and to avoid potential misinterpretation of the data.

    1. Since SUN2 is normally incorporated into a SUN2-SYNE2-KASH2 LINC heterohexamer complex, the authors should be cautious with the use of over-expressed SUN2 in this study. Over-expressed SUN2 is expected to stay mostly as unassembled molecules and thus is likely degraded by a protein quality control mechanism that targets unassembled proteins. Consistent with this possibility, CK2 has been implicated in the regulated turnover of aggregation-prone proteins (Watabe, M. et al., JCS 2011). This mechanism would be potentially distinct from the one proposed for endogenous SUN2 degradation.
    2. Certain conclusions appear to be an overstatement. This is particularly the case for the title, which implies that SUN2 is a protein that undergoes regulated turnover (under certain physiological conditions). Given that CK2 is a constitutive kinase and that the authors have not identified the conditions under which the activity of CTDNEP1 is regulated, it is premature to make such a conclusion.
    3. Likewise, the demonstration of the impact of SUN2 accumulation on different cellular pathways mainly relies on the over-expression of a non-degradable SUN2 mutant. Whether similar defects could be seen when the degradation of endogenous SUN2 is blocked remains an open question.

  8. Version published to 10.1101/2022.07.15.500172v1 on bioRxiv