The SUMO activating enzyme subunit, SAE2, contributes SUMO protein bias for mitotic fidelity

  1. Alexandra K. Walker
  2. Alexander J. Lanz
  3. Mohammed Jamshad
  4. Alexander J. Garvin
  5. Peter Wotherspoon
  6. Benjamin F. Cooper
  7. Timothy J. Knowles
  8. Joanna R. Morris

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Mammalian cells possess three conjugatable SUMO variants: SUMO1 and the largely indistinguishable SUMO2 and SUMO3 (designated SUMO2/3). Some SUMOylated substrates are modified by both SUMO1 and SUMO2/3, while others show biased modifications towards SUMO1 or SUMO2/3. How preferential SUMO protein conjugation is coordinated is poorly understood.

Here, we examine a modification of the catalytic component of the human SUMO Activation Enzyme, SAE2. We observe that lysine 164 of SAE2 is deacetylated during mitosis in an HDAC6-dependent manner. We find that an acetyl-analogue mutant, SAE2-K164Q, biases the activation and conjugation of SUMO2>SUMO1 and discriminates SUMO1 and SUMO2/3 through their C-terminal tails.

Complementation of SAE2-depleted or inhibited cells with SAE2-K164Q restricts mitotic SUMO1-conjugates and increases multipolar spindle formation. We confirm the SUMO E1-dependent modification of the nuclear mitotic apparatus, NuMA, and find that the mitotic defects of both SAE2-K164Q complemented cells and HDAC6-inhibitor-treated cells are corrected by either over-expression of SUMO1 or by expression of a GFP-SUMO1-NuMA-K1766R fusion protein.

Our observations suggest a model in which the SAE1:SAE2 enzyme is deacetylated on early mitosis to encourage the conjugation of SUMO1 to support mitotic fidelity. These surprising data reveal that the SUMO-activating enzyme can bias SUMO variant conjugation.

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

    What follows is our revision Plan.

    Manuscript number: RC-2024-02794

    Corresponding author(s): Jo Morris

    [The "revision plan" should delineate the revisions that authors intend to carry out in response to the points raised by the referees. It also provides the authors with the opportunity to explain their view of the paper and of the referee reports.

    The document is important for the editors of affiliate journals when they make a first decision on the transferred manuscript. It will also be useful to readers of the reprint and help them to obtain a balanced view of the paper.

    If you wish to submit a full revision, please use our "Full Revision" template. It is important to use the appropriate template to clearly inform the editors of your intentions.]

    1. General Statements [optional]

    We feel the reviewers understood the paper well and made many reasonable points for improvement.

    In response to Reviewer three's concern about the reliance on SAE2 over-expression, in the 'Significance' section "One limitation is the strong reliance on the use of an actyl-mimicking mutant". We were minded not to rely on the mutant. Hence, the paper contains considerable data onthe HDCAC6 deacteylase, responsible for SEA2 deacetylation. We show that HDAC6 inhibition phenocopies SAE2-K164Q expression and, moreover, that the approaches which rescue the mitotic defects of SAE2-K164Q expression cells also rescue the defects of HDCA6 inhibited cells. These observations, we believe, overcome the concern.

    2. Description of the planned revisions

    Insert here a point-by-point reply that explains what revisions, additional experimentations and analyses are planned to address the points raised by the referees.

    Revisions.

    *R1: *As the authors state, SUMO1 conjugates decrease during mitosis and this is somewhat at odds with the proposed model regarding NuMA. The authors can detect a SUMOylated NuMA conjugate (fig. 4a). To test whether the proposed model is correct, the authors could check: a. Whether this form is indeed SUMO1-NuMA b. Whether it decreases upon expression of the SAE2K164Q variant.

    R2: Figure 4:The authors show a ML792 sensitive high molecular weight smear of NUMA in nocodazole treated cells. It would be very convincing if the authors could demonstrate whether endogenous NUMA is conjugated to SUMO1 or SUMO2 in mitosis by SUMO IPs and whether they can detect a change upon expression of SAE2 variants as in Figure 3a. By replicating this experiment, it would be important to demonstrate the presence of both free and conjugated SUMO paralogs in the input and paralog specific sumoylation in general (smear) and of NUMA in the IP.

    Response:These are important points. We intend to perform the suggested experiments to address which isoform NuMa is modified by, and what the impact of the variant is.

    R2:Figures 2 C/Supplementary Figure 3c: The enzyme concentrations used in these reactions are much too high. To discriminate between thioester- and isopeptide-linked SUMO, the same samples should be analyzed in the absence (detection of thioester and isopeptide linkages) and presence of high concentrations of DTT (detection of isopeptide-linked SUMO only). The presented assay is problematic as it shows dimeric SUMO and RanGAP1:SUMO bands in the absence of ATP and no UBC9 but SAE2 thioester/isopetide formation in the absence of RanGAP1 (preferentially UBC9 should form a thioester/isopetide bond in this condition as higher molarities of UBC9 over E1 are used). Dimeric SUMO should not be detected unless disulfide bridges are formed between cysteines - this happens when DTT is not present in the reaction - under such conditions, SAE2 and UBC9 can also form disulfide bridges via their catalytic cysteines, impairing their enzymatic activity. In order to interpret the results correctly, it is important to add low concentrations of DTT (~0.1 mM) even in thioester reactions and to distinguish between thioester and isopeptide linkages.

    R2: Figure 2F/ Supplementary Figure 3d: Again, the enzyme concentrations are much too high and need to be reduced to a concentration where mainly RanGAP1 monosumoylation with SUMO1 is detected. As RanGAP1 is the most efficient SUMO substrate known, the enzyme concentrations and reaction time can be greatly reduced to limit the auto-modification of the enzymes and SUMO chain formation. Due to the efficient chain-forming activity of SUMO2, this is more difficult with SUMO2, but can be reduced by limiting the concentration of UBC9 in particular or by using a SUMO2 KallR mutant. In the reaction shown, the authors used only twice the molarity of SUMO compared to the substrate, too low taking into account SUMO2 chain formation, enzyme and substrate modification (The reaction should be limited by enzyme activity not by SUMO2). How can the authors be sure that the band they report as RanGAP1 high MW SUMO2 is indeed RanGAP1 modified and not SAE2 (in comparison to Suppl Figure 3b)?

    Response: We intend to repeat these assays, as suggested by the reviewer, reducing the enzyme concentrations and using low-concentration DTT. With the relevant controls and blots to show the identity of the RanGAP-SUMO2 product. Further, we will show control experiments with and without DTT that demonstrate the sensitivity of the SAE2~SUMO band to the reducing agent.

    R2: Figure 3 nicely shows that ML792-resistant SAE2 variants conjugate SUMO2 equally well, whereas SAE2 K163R is reduced and SAE2 K163Q appears to be abolished in SUMO1 conjugation. However, only high molecular weight SUMO conjugates are shown. What are the levels of free SUMO after overexpression of SAE2 variants and the indicated treatments?

    Response: We will attempt to show free SUMO levels in mitotic cells.

    R2: According to the work of Zhang et al from the Matunis lab (cited as reference 39 in the proposed study), SUMO conjugation is greatly reduced in nocodazole-arrested cells, but is restored after release in G1. Furthermore, SUMO1 and SUMO2 localize to different subcellular regions during mitosis. Have the authors tested whether SAE2 variants differ in their intracellular localization or alter the subcellular localization of SUMO1 and SUMO2 in interphase and mitotic cells?

    Response: We will examine the localisation of the SAE2 variants (see section below for the SUMO proteins).

    R3: It would be helpful if the authors could more clearly separate the two steps catalyzed by the E1. This would be needed to determine whether the accumulation of the SUMO1-AMP intermediate by the K164Q mutant is due to a faster rate of formation or a reduced rate of conversion to the thioester. They could test the AMP formation step in isolation in a straightforward manner by using the double mutant K164Q C173G and measuring a time course of SUMO1-AMP versus SUMO2-AMP build-up. Alternatively, they could try to isolate the second step by adding SUMO1-AMP versus SUMO2-AMP to the E1 de novo - although isolation of the intermediates may be more involved.

    Response: We intend to perform the first approach suggested, making and examining the double mutant's activity as suggested.

    R3: The reason for the isoform selectivity in the context of NuMA SUMOylation remains unresolved. The study would be significantly strengthened if the authors could address the question of whether the mitotic defects come from a lack of NuMA SUMOylation or the wrong type of SUMOylation. In other words, does it matter which isoform of SUMO is attached to NuMA? This could be addressed by also creating a SUMO2 fusion construct and testing whether that suppresses some of the phenotypes observed with the K164Q mutant and upon HDAC6 inhibition.

    Response. This is an excellent suggestion. We intend to make the constructs suggested and perform this experiment for our revision.

    R3. It would be helpful to show a time course of endogenous SAE2 acetylation over the cell cycle, using synchronized cultures.

    Response. We will attempt to gain a view of SAE2 acetylation over the cell cycle, which requires the precipitation of endogenous SAE following synchronisation.

    R3: Fig 2a: The figure would be easier to understand if the same colour scheme was used for S1 versus S2 to aid the comparison.

    Response: We will change this.

    R3: The title is not immediately understandable. "SUMO protein bias for mitotic stability" sounds a bit awkward. It would be clearer to be more explicit about isoforms.

    Response: We have considered: "HDAC6-Dependent Deacetylation of SAE2 enhances SUMO1 Conjugation for Mitotic Integrity", we have not changed it on the current manuscript so as not to confuse the reader - we will change it at the journal level.

    R3: Fig 2b: I don't understand the units of this graph. Why does normalization result in a value of zero, not 1? On this scale, what would a value of 1 signify? How can a value become negative? I would have expected values relative to the WT, with the WT being set to 1 or to 100%. The authors should also show the raw data for this plot.

    Response: The data will be normalised to the WT condition (1 instead of 0), and raw data shown.

    R3: Fig 2c: Please also show representative raw data.

    Response: Representative images will be shown.

    R3: Fig 2d,f: Again, the legend should explain what the plots were normalized to.

    Response: Inserted in the legend for Fig. 2d&f: 'The RanGAP1-SUMO1 products are normalised to the WT SAE1:SAE2:SUMO1-only condition (top) and the RanGAP1-SUMO2 products are normalised to the WT SAE1:SAE2:SUMO2-only condition (bottom).'

    R3 Fig S5b: The authors argue with the hydrogen bonding capacities of the different pairings. However, acetylation at K164 should not necessarily prevent a hydrogen bond to SUMO1-E93, considering that the "NH" group is likely still at a comparable distance to the carboxylate of E93 and could in principle undergo H-bonding unless prevented by the steric bulk introduced by the acetyl group. On the other hand, the K164-E93 interaction is the only electrostatic interaction among the 4 possible combinations. While a contribution is not easy to prove experimentally, I think the possibility of charge-charge interactions having an impact should be considered in the discussion.

    Response: Agreed. The figure will be redrawn, and the possibility will be discussed.

    R1 Fig. 2c: Why does C173G form a thioester with SUMO2 up to 40% of the WT?

    Response: We believe this arose in measuring background density in the blots in error. We will re-assess the method used.

    R3: Fig 4b: The images have very poor contrast. In addition, the merged image would be clearer if two different colours were used.

    Response: We will change one of the colours.

    3. Description of the revisions that have already been incorporated in the transferred manuscript

    Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. If no revisions have been carried out yet, please leave this section empty.

    R1:2. Please clarify the use of Dox addition in the text and legend earlier (is found currently in Supp. Fig 4).

    Response: Inserted before first result using doxycycline: 'Furthermore, we generated U2OS with a doxycycline-inducible (wild-type) WT FLAG-SAE2 or a FLAG-SAE2-K164R mutant.'

    R1.3. Fig. 4f: what is the difference between the first (invisible NUMA) bipolar and the second, NuMA visible bipolar spindle?

    Response: Fig. 4f now annotated with 'Untransfected' and 'GFP-NuMA transfected'.

    R1.4. ML972- should read ML792 on pg 8.

    Response: Corrected.

    R3: All the experiments showing acetylation are done with transfected FLAG-tagged constructs - are they overexpressed?

    Response: Supplemental Figure 4a illustrates that with the exception of the C173G mutant, the remainder WT, and K164-mutants are all expressed at near WT-levels and not over-expressed. The C-G-mutant is highly expressed.

    R3: On page 3, the authors could introduce a justification of why they tested IR treatment.

    Response: now justified.

    R3: The authors repeatedly use the word "codon" when they describe a site in the protein. Codon refers to mRNA, so the word "residue" would be more appropriate when talking about a protein.

    Response: Agreed. Done.

    R3: Page 8: "confirmation" should be "conformation".

    Response: Done.

    R3:Page 8: "While we find a little role for..." - delete "a"

    Response: Done.

    *R2: *Supplementary Figure 2: Please indicate the size of the marker bands, the fraction numbers and which fractions were pooled for further analysis. Is there any explanation why SAE1:SAE2K164R eluates in two peaks, suggesting two complexes? How different are they in size?

    Response: Ladder markers added to each gel image. Fraction numbers added. Black box indicates fractions pooled. Figure updated with relevant recombinant protein preps generated for updated in vitro experiments. The additional SAE1:SAE2-K164R peak which appeared in the previous manuscript Supp. Fig. 2a eluted in the void volume and so we think it comprised aggregated SAE1:SAE2 protein, more recent preparations do not show it.

    R3: The authors should include a more detailed discussion of the importance of the absolute and relative concentrations of free SUMO1 versus SUMO2/3 as a possible mechanism to impose isoform bias. Specifically, they should consider the different KM values of the E1 for the isoforms. The literature says that the E1 has a lower KM (higher affinity) for SUMO1 than SUMO2/3 but also a lower kcat (considering both steps of its reaction together), resulting in an approximately equal Kcat/KM. This would mean that at low overall SUMO concentrations, SUMO1 would have an advantage, whereas with rising SUMO concentrations SUMO2/3 would be favoured (which might be particularly important during stress conditions). What part of the curve does the cellular environment reflect?

    Response: Yes, good point. Now included:

    R3: Fig 3g: Could the authors comment on the detrimental effects of both SUMO1 and SUMO2 in the WT background?

    Response: Comment included.

    R3: Fig 3h: typo ("Trasfect")

    Response: Done.

    R3: Fig 4f: The DAPI signal is hardly visible - better contrast would help.

    Response: Improved.

    R3: Fig S2: It would be appropriate to indicate which fractions were actually collected or combined during the purification.

    Response: Ladder markers added to each gel image. Fraction numbers added. Black box indicates the fractions pooled.

    4. Description of analyses that authors prefer not to carry out

    Please include a point-by-point response explaining why some of the requested data or additional analyses might not be necessary or cannot be provided within the scope of a revision. This can be due to time or resource limitations or in case of disagreement about the necessity of such additional data given the scope of the study. Please leave empty if not applicable.

    R2: According to the work of Zhang et al from the Matunis lab (cited as reference 39 in the proposed study), SUMO conjugation is greatly reduced in nocodazole-arrested cells, but is restored after release in G1. Furthermore, SUMO1 and SUMO2 localize to different subcellular regions during mitosis. Have the authors tested whether SAE2 variants differ in their intracellular localization or alter the subcellular localization of SUMO1 and SUMO2 in interphase and mitotic cells?

    Response: We have investigated SUMO isoform location. However, in our hands, using a range of SUMO antibodies, we do not see the previously reported localisations in mitotic wild-type cells, and thus, we are not able to assess the impact of the SAE variants. As our phenotypes are restricted to mitosis, we do not consider it worthwhile to look at interphase.

  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:

    In this manuscript, the authors report on an interesting regulatory mechanism that influences the balance between conjugation of the different SUMO isoforms, SUMO1 versus SUMO2/3. The authors describe that acetylation of a specific residue, K164, in the SUMO activating enzyme (E1) subunit, SAE2, biases the E1's preference towards SUMO2/3. Specifically, they use an acetylation-mimicking K164Q mutation to show that the acetylation state of SAE2 likely affects the affinity of the E1 to SUMO and the rate of thioester formation. With an antibody, they demonstrate the acetylation of SAE2 in cells. Mechanistically, they locate the cause of the isoform bias to a residue in the C-terminus of SUMO in proximity to K164 or SAE2, where SUMO1 carries glutamate, while SUMO2/3 has glutamine. Switching these residues between the SUMO isoforms reverses the isoform preference of the E1. Phenotypically, the SAE2 K164Q mutant induces mitotic problems that the authors attribute to the SUMOylation of the NuMA complex. They assign the deacetylation of SAE1 to HDAC6 and report that deacetylation occurs during mitosis. These results are consistent with a model that SUMO1 modification of the NuMA complex in mitosis is important for mitotic fidelity and that the cell cycle-dependent changes in the acetylation status of SAE2 promote this. Accordingly, fusion of SUMO1 to a NuMA subunit partially overcomes the problems induced by the K164Q mutant or the inhibition of HDAC6.

    Major comments:

    The experiments are largely performed in a well-controlled manner, and overall, the study is very convincing. I would like to suggest a few experiments that would strengthen the authors' conclusions, and there are a few minor issues with some of the figures.

    1. It would be helpful if the authors could more clearly separate the two steps catalyzed by the E1. This would be needed to determine whether the accumulation of the SUMO1-AMP intermediate by the K164Q mutant is due to a faster rate of formation or a reduced rate of conversion to the thioester. They could test the AMP formation step in isolation in a straightforward manner by using the double mutant K164Q C173G and measuring a time course of SUMO1-AMP versus SUMO2-AMP build-up. Alternatively, they could try to isolate the second step by adding SUMO1-AMP versus SUMO2-AMP to the E1 de novo - although isolation of the intermediates may be more involved.
    2. The reason for the isoform selectivity in the context of NuMA SUMOylation remains unresolved. The study would be significantly strengthened if the authors could address the question of whether the mitotic defects come from a lack of NuMA SUMOylation or the wrong type of SUMOylation. In other words, does it matter which isoform of SUMO is attached to NuMA? This could be addressed by also creating a SUMO2 fusion construct and testing whether that suppresses some of the phenotypes observed with the K164Q mutant and upon HDAC6 inhibition.
    3. The authors should include a more detailed discussion of the importance of the absolute and relative concentrations of free SUMO1 versus SUMO2/3 as a possible mechanism to impose isoform bias. Specifically, they should consider the different KM values of the E1 for the isoforms. The literature says that the E1 has a lower KM (higher affinity) for SUMO1 than SUMO2/3 but also a lower kcat (considering both steps of its reaction together), resulting in an approximately equal Kcat/KM. This would mean that at low overall SUMO concentrations, SUMO1 would have an advantage, whereas with rising SUMO concentrations SUMO2/3 would be favoured (which might be particularly important during stress conditions). What part of the curve does the cellular environment reflect?
    4. It would be helpful to show a time course of endogenous SAE2 acetylation over the cell cycle, using synchronized cultures. All the experiments showing acetylation are done with transfected FLAG-tagged constructs - are they overexpressed? Is is not possible to work with endogenous SAE2?

    Minor comments:

    • The title is not immediately understandable. "SUMO protein bias for mitotic stability" sounds a bit awkward. It would be clearer to be more explicit about isoforms.
    • On page 3, the authors could introduce a justification of why they tested IR treatment.
    • The authors repeatedly use the word "codon" when they describe a site in the protein. Codon refers to mRNA, so the word "residue" would be more appropriate when talking about a protein.
    • Page 8: "confirmation" should be "conformation".
    • Page 8: "While we find a little role for..." - delete "a"
    • Fig 2a: The figure would be easier to understand if the same colour scheme was used for S1 versus S2 to aid the comparison.
    • Fig 2b: I don't understand the units of this graph. Why does normalization result in a value of zero, not 1? On this scale, what would a value of 1 signify? How can a value become negative? I would have expected values relative to the WT, with the WT being set to 1 or to 100%. The authors should also show the raw data for this plot.
    • Fig 2c: Please also show representative raw data.
    • Fig 2d,f: Again, the legend should explain what the plots were normalized to.
    • Fig 3g: Could the authors comment on the detrimental effects of both SUMO1 and SUMO2 in the WT background?
    • Fig 3h: typo ("Trasfect")
    • Fig 4b: The images have very poor contrast. In addition, the merged image would be clearer if two different colours were used.
    • Fig 4f: The DAPI signal is hardly visible - better contrast would help.
    • Fig S2: It would be appropriate to indicate which fractions were actually collected or combined during the purification.
    • Fig S5b: The authors argue with the hydrogen bonding capacities of the different pairings. However, acetylation at K164 should not necessarily prevent a hydrogen bond to SUMO1-E93, considering that the "NH" group is likely still at a comparable distance to the carboxylate of E93 and could in principle undergo H-bonding unless prevented by the steric bulk introduced by the acetyl group. On the other hand, the K164-E93 interaction is the only electrostatic interaction among the 4 possible combinations. While a contribution is not easy to prove experimentally, I think the possibility of charge-charge interactions having an impact should be considered in the discussion.

    Significance

    The results presented here are interesting and novel. Importantly, the authors provide a molecular model for a new mechanism of how the SUMO system achieves isoform specificity, which is a still very poorly understood phenomenon. The manuscript makes a significant advance by contributing an important new aspect of how the SUMO conjugation machinery chooses between isoforms. The manuscript is strong by providing very good evidence for its conclusions. One limitation is the strong reliance on the use of an actyl-mimicking mutant; this limitation could be overcome by placing a bit more emphasis on detecting endogenous SAE2 acetylation.

    Audience: The study should be relevant to a broad audience, given the impact of the SUMO system on cellular regulation; after all, the study addresses a very fundamental problem in the field. In addition, it should be of interest to researchers studying regulation of mitosis.

  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:

    Walker et al characterized lysine 164 acetylation of the catalytic SUMO activating enzyme subunit SAE2 and observed that this modification causes a bias towards SUMO2/3 over SUMO1 involving their C-terminal tails. While several enzymes appear to mediate SAE2 acetylation, HDAC6 is responsible for deacetylating SAE2 in mitosis, thereby promoting mitotic SUMO1 modification. The nuclear mitotic apparatus, NuMA, was identified as a putative mitotic SUMO1 substate upon SAE2 deacetylation. Replacement of endogenous SAE2 with an acetylation mimetic SAE2-K164Q mutant restricts SUMO1 conjugation of NuMA resulting in multipolar spindle formation that can be rescued either by overexpression of SUMO1 or by SUMO1-NuMA fusion.

    Major comments:

    • Figures 2 C/Supplementary Figure 3c: The enzyme concentrations used in these reactions are much too high. To discriminate between thioester- and isopeptide-linked SUMO, the same samples should be analyzed in the absence (detection of thioester and isopeptide linkages) and presence of high concentrations of DTT (detection of isopeptide-linked SUMO only). The presented assay is problematic as it shows dimeric SUMO and RanGAP1:SUMO bands in the absence of ATP and no UBC9 but SAE2 thioester/isopetide formation in the absence of RanGAP1 (preferentially UBC9 should form a thioester/isopetide bond in this condition as higher molarities of UBC9 over E1 are used). Dimeric SUMO should not be detected unless disulfide bridges are formed between cysteines - this happens when DTT is not present in the reaction - under such conditions, SAE2 and UBC9 can also form disulfide bridges via their catalytic cysteines, impairing their enzymatic activity. In order to interpret the results correctly, it is important to add low concentrations of DTT (~0.1 mM) even in thioester reactions and to distinguish between thioester and isopeptide linkages.
    • Figure 2F/ Supplementary Figure 3d: Again, the enzyme concentrations are much too high and need to be reduced to a concentration where mainly RanGAP1 monosumoylation with SUMO1 is detected. As RanGAP1 is the most efficient SUMO substrate known, the enzyme concentrations and reaction time can be greatly reduced to limit the auto-modification of the enzymes and SUMO chain formation. Due to the efficient chain-forming activity of SUMO2, this is more difficult with SUMO2, but can be reduced by limiting the concentration of UBC9 in particular or by using a SUMO2 KallR mutant. In the reaction shown, the authors used only twice the molarity of SUMO compared to the substrate, too low taking into account SUMO2 chain formation, enzyme and substrate modification (The reaction should be limited by enzyme activity not by SUMO2). How can the authors be sure that the band they report as RanGAP1 high MW SUMO2 is indeed RanGAP1 modified and not SAE2 (in comparison to Suppl Figure 3b)?
    • Figure 3 nicely shows that ML792-resistant SAE2 variants conjugate SUMO2 equally well, whereas SAE2 K163R is reduced and SAE2 K163Q appears to be abolished in SUMO1 conjugation. However, only high molecular weight SUMO conjugates are shown. What are the levels of free SUMO after overexpression of SAE2 variants and the indicated treatments? According to the work of Zhang et al from the Matunis lab (cited as reference 39 in the proposed study), SUMO conjugation is greatly reduced in nocodazole-arrested cells, but is restored after release in G1. Furthermore, SUMO1 and SUMO2 localize to different subcellular regions during mitosis. Have the authors tested whether SAE2 variants differ in their intracellular localization or alter the subcellular localization of SUMO1 and SUMO2 in interphase and mitotic cells? Can the authors comment on the proportion of SAE2 that is acetylated?
    • Figure 4:The authors show a ML792 sensitive high molecular weight smear of NUMA in nocodazole treated cells. It would be very convincing if the authors could demonstrate whether endogenous NUMA is conjugated to SUMO1 or SUMO2 in mitosis by SUMO IPs and whether they can detect a change upon expression of SAE2 variants as in Figure 3a. By replicating this experiment, it would be important to demonstrate the presence of both free and conjugated SUMO paralogs in the input and paralog specific sumoylation in general (smear) and of NUMA in the IP.

    Minor comments:

    • Supplementary Figure 2: Please indicate the size of the marker bands, the fraction numbers and which fractions were pooled for further analysis. Is there any explanation why SAE1:SAE2K164R eluates in two peaks, suggesting two complexes? How different are they in size?

    Significance

    The finding that E1 acetylation regulates SUMO paralog specificity is very exciting, particularly because of its link to key regulatory mitotic functions. Overall, the findings are intriguing and supported in part by various biological and biochemical methods. However, some concerns remain unsatisfactorily addressed, as outlined above.

    The findings provide a novel basic concept of how E1 enzyme regulation contributes to the specification of modifier selectivity, demonstrates cross-talk with other PTMs and reveals a biological function. Therefore, the study is of interest to a broad audience.

  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

    Summary:

    In their manuscript, Walker et al. investigate the physiological role of deacetylation of the SAE2 subunit of the SUMO E1 enzyme. They find that SAE1:SAE2-acK164 is deacetylated in an HDAC6-dependend manner and use a series of biochemical assays to show that deacetylation of the SAE2 subunit shifts the bias of the SUMO E1 towards SUMO1 conjugation in vitro, proposing a mechanism that is similar to the one that the NEDD8 E1 employs to discriminate between NEDD8 and ubiquitin.

    The authors continue to examine the role of different SAE2 variants in different cellular stresses and show that the acetyl-mimicking SAE2K164Q variant displays reduced levels of high molecular weight SUMO1 conjugates in mitotic cells. This variant cannot support proper mitotic spindle formation leading to the appearance of multipolar spindles and centromere-containing micronuclei. Finally, they go on to identify the mechanism underlying these phenotypes and examine NuMA SUMOylation. They test SUMOylation-refractive NuMA variants as well as an already published SUMO1-NuMA fusion that mimics the SUMOylated protein form. They propose a model in which deacetylation of SAE2 changes the bias of the SUMO E1 to increase SUMO1-NuMA conjugation during mitosis, promoting bipolar spindle formation.

    Major point:

    As the authors state, SUMO1 conjugates decrease during mitosis and this is somewhat at odds with the proposed model regarding NuMA. The authors can detect a SUMOylated NuMA conjugate (fig. 4a). To test whether the proposed model is correct, the authors could check:

    a. Whether this form is indeed SUMO1-NuMA

    b. Whether it decreases upon expression of the SAE2K164Q variant.

    Minor points:

    1. Fig. 2c: Why does C173G form a thioester with SUMO2 up to 40% of the WT?
    2. Please clarify the use of Dox addition in the text and legend earlier (is found currently in Supp. Fig 4).
    3. Fig. 4f: what is the difference between the first (invisible NUMA) bipolar and the second, NuMA visible bipolar spindle?
    4. ML972- should read ML792 on pg 8.

    Significance

    General assessment:

    This is a thorough study with complex but well controlled experiments and contains a large amount of valuable information. A point could be further clarified in order to provide further support the proposed model.

    Advance:

    The document brings understanding on the regulation of the SUMO conjugation system a step forward and links it to a physiological context.

    Audience:

    basic science: the Ubiquitin family field and also the mitosis-cytoskeleton field. Applied science concerning the use of SUMO inhibitors in cancer.

    Expertise: SUMO regulation of the cytoskeleton during mitosis (yeast system)

  5. Version published to 10.1101/2022.03.18.484840v2 on bioRxiv
  6. Version published to 10.1101/2022.03.18.484840v1 on bioRxiv