Effects of phosphorylation on Drp1 activation by its receptors, actin, and cardiolipin

  1. Ao Liu
  2. Anna L. Hatch
  3. Henry N. Higgs

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Drp1 is a dynamin family GTPase required for mitochondrial and peroxisomal division. Oligomerization increases Drp1 GTPase activity through interactions between neighboring GTPase domains. In cells, Drp1 is regulated by several factors including Drp1 receptors, actin filaments, cardiolipin, and phosphorylation at two sites: S579 and S600. Commonly, phosphorylation of S579 is considered activating, while S600 phosphorylation is considered inhibiting. However, direct effects of phosphorylation on Drp1 GTPase activity have not been investigated in detail. Here, we compare effects of S579 and S600 phosphorylation on purified Drp1, using phosphomimetic mutants and in vitro phosphorylation. Both phosphomimetic mutants are shifted toward smaller oligomers. Both phosphomimetic mutations maintain basal GTPase activity, but eliminate GTPase stimulation by actin and decrease GTPase stimulation by cardiolipin, Mff, and MiD49. Phosphorylation of S579 by Erk2 produces similar effects. When mixed with wildtype Drp1, both S579D and S600D phosphomimetic mutants reduce the actin-stimulated GTPase activity of Drp1-WT. Conversely, a Drp1 mutant (K38A) lacking GTPase activity stimulates Drp1-WT GTPase activity under both basal and actin-stimulated conditions. These results suggest that the effect of S579 phosphorylation is not to activate Drp1 directly. In addition, our results suggest that nearest neighbor interactions within the Drp1 oligomer affect catalytic activity.

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

    We have thoroughly revised the manuscript, taking into account all comments from all four reviewers. We have added new data (Supplemental Figure 2 and Supplemental Figure 4) in response to these comments.

    Reviewer 1

    The assessment of data reproducibility is currently uncertain due to the absence of replication and statistical analysis in the dataset. It is essential to provide explicit information regarding sample sizes or replicates for all data and figures, data should be presented as mean +/- SD/SEM, and the interpretation of results should be grounded in rigorous statistical analysis. The lack of experimental replicates and statistical analysis in most of the figures presented raises major concerns regarding the validity of the result.

    We have now added error bars for the graphs in Figure 3D, E, F, G, H; Figure 4 D, F, G, H, I, J; Figure 5 B, C, D, E, F, G; and Figure 6B, C, D. All GTPase assays have repeated three times. The mean ± S.D. (n = 3) is plotted for each condition. For high-speed pelleting assays, all assays have been conducted three times, and a representative assay is shown.

    Why was only one of the MiD proteins, specifically MiD49, studied, while MiD51 was not includedin the investigation?

    This is an excellent point. In our previous work (doi:10.1101/2023.07.31.551267), we found that MiD49 and MiD51 were strikingly similar in their abilities to activate Drp1 after their own activation with fatty acyl-CoA. We feel that the demonstration here with MiD49 suggests that a similar effect would occur with MiD51. Due to time constraints for the lead author, preparing more MiD51 protein was out of the scope of what could be done. We now add a line in the Discussion that results for MiD51 may be different.

    The author suggestion of Drp1 phosphorylation, based on the mobility of protein observed in SDS-PAGE gel (fig 4A, 5A, 6A), is not a sufficiently valid assessment. While western blot analysis is a valid method to assess Drp1 phosphorylation, it is essential to include replicates for semi-quantitation and demonstrate the reproducibility of the results. Moreover, it is recommended to incorporate Western blot analyses to provide additional support for the findings presented in Figures 5 and 6.

    • We agree with the reviewer that additional information on the phosphorylation state of these proteins should be provided. We now include phospho-proteomic analysis for Erk2 phosphorylation of WT Drp1 and Drp1-S600D (Supplemental Table 1), showing that S579 is by far the predominant phosphorylation site. For WT Drp1, three lines of evidence now suggest efficient Erk2 phosphorylation of S579:
    1. Western blot using anti-phosphoS579
    2. Phosphoproteomic analysis
    3. Gel shift

    For the Drp1-S600D phosphorylation, we have phosphoproteomic and gel shift analysis. For isoform 6, we regrettably only have gel shift. However, given the fact that the effect of Erk2 treatment on actin-stimulated GTPase activity mimics what we found for WT-Drp1 and for Drp1-phosphoS579/S600D, we think it is highly likely that the equivalent phosphorylation (S629 in this case) has been affected.

    Data on phosphorylated peptides with replicates experiments should be presented.

    We now present these data, which have been significantly expanded since the initial submission (new Supplemental Table 1). While non-phophorylated S579 is still detected in both the WT and S600D phosphorylation reactions, the phosphorylated peptide is 2.2 and 2.3-fold more abundant, respectively. Our conclusion is that Erk2 efficiently phosphorylates S579, although stoichiometric phosphorylation was not obtained here. We have added statements in the relevant sections of the Result, and in the Methods. We have also added Supplemental Table 1 to show the spectral counts obtained from phospho-proteomic analysis, and have deposited the raw data files with the PRIDE consortium (access information in the Methods).

    Please provide additional context or specific details about the GFP-tagged Drp1 protein, such as the protein site where GFP was attached, as well as whether this tag could potentially impact the Drp1 GTPase activity and oligomerization. Figure 7C and D appear to suggest an increase in the GTPase activity of the GFP-Drp1 protein.

    We have now added these details to the Methods section, and have also added the complete amino acid sequence for the final purified construct in Supplemental Figure 4. We have also added that a previous study (PMID: 32901052) found that inclusion of GFP strongly inhibited Drp1 GTPase activity. We do not observe this effect here or in a previous study (PMID: 27559132), and provide possible reasons for this difference in the Methods. The reviewer points out that the activity of GFP-Drp1 appears higher than that of un-tagged Drp1 (comparing 7C with 7D). We find that the GTPase activity of Drp1 alone varies between 1 and 2 uM/min/uM protein depending on the preparation. This variation occurs for both untagged and GFP-tagged Drp1. This difference in basal activity from prep-to-prep might relate to differences between protein preparations, or exact amount of time required to freeze the aliquots of purified protein (we freeze small aliquots ( An optional experiment that would significantly enhance the biological relevance of the findings presented in the current study is to assess the morphology of mitochondria in cells expressing the phospho-mimetic mutant Drp1 proteins. This experiment would provide valuable insights into the functional consequences of Drp1 S579 and S600 phosphorylation on mitochondrial structure and dynamics.

    We fully agree that these would be valuable experiments. The issue is that a large number of experiments using phospho-mimetic mutants in cells have already been conducted, with varying results (Taguchi et al., 2007; Qi et al., 2011; Yu et al., 2011; Strack et al., 2013; Kashatus et al., 2015; Serasinghe et al., 2015; Xu et al., 2016; Brand et al., 2018; Han et al., 2020, Chang and Blackstone, 2007; Cribbs and Strack, 2007; Cereghetti et al., 2008; Wikstrom et al., 2013, Han et al., 2008，Wang et al., 2012 Jhun BS, Sheu, 2018, J Physiol). To conduct more targeted tests examining specific forms of Drp1 activation in cells (for example, through Mff, MiD proteins, actin, or cardiolipin) will require extensive work that is outside the scope here. Our feeling is that S579 phosphorylation is likely to recruit another molecule (probably a protein) that has an activating effect. We tried to test one possibility (NME3, mentioned in the Discussion) but failed to produce useable NME3 protein for these tests and, given time constraints for the lead author, could not address this further.

    Provide reference for method on actin polymerization.

    We have now added a reference in the ‘Actin preparation for biochemical assays’ section of the Methods (PMID 16472659).

    Rectify the error in referencing figure 3 panels within the figure legends of Supplemental Fig S1.

    Thank you, we have changed this.

    The inclusion of full length isoform 6 is commendable. However, there is no mentioned of isoform6 in the method section.

    Thank you for pointing this out. We have added description of the construct and referenced our previous paper that used it.

    Since papers deposited in bioRxiv have not undergone peer review, reference #7 should not becited as references in scholarly work.

    Reference 7 has so far been reviewed by a peer-review journal. e are addressing reviewers’ concerns and will re-submit soon. We do not know how to rectify the issue of referencing this work, because it describes an extensive amount of groundwork for the MiD proteins. Our hope is that this work will be in press by the time the work reviewed here is ready for publication.

    Please provide details about the calculation of GTPase activity and the distinctions between the specific GTPase activity and total GTPase activity shown in figure 8D-F.

    We now describe these calculations in the “GTPase assay” section of the Methods.

    Reviewer 2

    Overall, the experiments described here are carried out with rigor and the conclusions drawn are of significance to understanding how phosphorylation regulates Drp1 functions.

    Thank you for these kind comments!

    Phosphorylation of both the serine residues appears to elicit a common effect in that they inhibitDrp1's stimulated GTPase activity. This would suggest that phosphorylation affects Drp1's self-assembly as tightly packed helical scaffolds. Instead of sedimentation analysis, an EM analysis of helical scaffolds on cardiolipin-containing membrane nanotubes or in the presence of soluble adaptors causing Drp1 to form filaments would provide a direct readout for defects in self-assembly.

    This is an excellent point, and we would love to conduct this work. Given our current EM infrastructure and expertise, these experiments would take extensive time for us to do. We do have a collaborator who could carry these out, but feel that the time it would take even for them to do this correctly is beyond that which we have (the lead author is transitioning to their next career phase). We have added the point that further EM studies of this type are necessary to test the effect on Drp1 assembly more directly.

    I am not sure of the rationale for experiments reported in Fig. 7 and 8. If the idea was to test if hetero oligomerization with WT Drp1 rescues defects associated with phosphorylated Drp1 then this could be stated explicitly in the manuscript. GFP-Drp1 is used as a WT mimic but a previous report (PMID: 30531964) indicates that this construct is severely defective in stimulated GTPase assays, much like the K38A mutant. But the rationale of using these constructs is not quite apparent. Is the intention to test if defects seen in the phospho-mimetic mutants of Drp1 can be rescued by the presence of a 'seed' of WT Drp1. If so, then this could be stated explicitly in the manuscript. But regardless, I am not quite sure what this data set achieves in terms of addressing mechanism.

    We apologize for not being clearer in our explanation of these experiments. Our goal was to test the effects of partial Drp1 phosphorylation on overall Drp1 activity, which likely mimics more accurately the cellular situation (wherein only a portion of the Drp1 population is likely to be phosphorylated even upon kinase activation). We now discuss these experiments in a clearer manner. For the GFP-Drp1, we do not observe the effect on GTPase activity shown in that previous manuscript by another laboratory, either here or in previous studies (eg, PMID: 27559132). In the Methods, we now provide a discussion of these differences and possible reasons for them, as well as providing the complete amino acid sequence of our GFP-fusion construct in Supplemental Figure 4.

    Finally, it would have been nice to see if the phospho-mimetic mutants of Drp1 produce the same effects on mitochondrial structure as those reported earlier. Reanalyzing their effects in a cellular assay becomes important because it would consolidate this work for the readers to evaluate the'true' effects of phosphorylation on Drp1 functions. If the phospho-mimetic mutants fare in a manner like those previously reported, then it signifies that stimulation in GTPase activity is not a readout that directly correlates with Drp1 functions. If not, then the results presented here would establish a comprehensive analysis of in vitro biochemical activities and in vivo functions of the phospho-mimetic mutants.

    We fully agree that these would be valuable experiments. The issue is that a large number of experiments using phospho-mimetic mutants in cells have already been conducted, with varying results (Taguchi et al., 2007; Qi et al., 2011; Yu et al., 2011; Strack et al., 2013; Kashatus et al., 2015; Serasinghe et al., 2015; Xu et al., 2016; Brand et al., 2018; Han et al., 2020, Chang and Blackstone, 2007; Cribbs and Strack, 2007; Cereghetti et al., 2008; Wikstrom et al., 2013, Han et al., 2008，Wang et al., 2012 Jhun BS, Sheu, 2018, J Physiol). To conduct more targeted tests examining specific forms of Drp1 activation in cells (for example, through Mff, MiD proteins, actin, or cardiolipin) will require extensive work that is outside the scope here. Our feeling is that S579 phosphorylation is likely to recruit another molecule (probably a protein) that has an activating effect. We tried to test one possibility (NME3, mentioned in the Discussion) but failed to produce useable NME3 protein for these tests and, given time constraints for the lead author, could not address this further.

    Previous work reports that the effect of actin on the GTPase activity of Drp1 is biphasic but the binding to actin is not. This is quite confounding, and the authors could perhaps explain why this is the case.

    The reviewer makes an excellent point, which we now explain further in the manuscript. We have also discussed this in doi:10.1101/2023.07.31.551267 (see Figure 2D in that work). Our interpretation is that it is the density of Drp1 bound to the actin that provides the activation, by positioning the GTPase domains in close proximity. As the amount of actin increases, the Drp1 becomes more dispersed on the filaments, and activation decreases. We observe the same effect for MiD49 and MiD51 oligomers (see the above-mentioned reference).

    The manuscript cites PMID: 23798729 for expression analysis of slice variants but PMID:29853636 provides a more compressive analysis. The authors could cite this work.

    Thank you for this reference. We were unaware of it, but are very glad to know of it now. We now include this reference. In particular, in the legend to Figure 1C (table of splice variants), we now state that this table is for human Drp1, and that additional splice variants have been identified for murine Drp1 (PMID 29853636).

    Reviewer 3

    The splendid results of the manuscript willbe interesting to the researchers in the related fields.

    Thank you for this nice comment!

    The manuscript provided well-organized biochemistry results for comparisons between phosphorylation of Drp1 S579 and S600. It is the reviewer's comments that the authors may include experiments that manipulate Drp1 phosphorylation at different amino acids in cells. Such experiments will provide strong support for this manuscript.

    • We fully agree that these would be valuable experiments. The issue is that a large number of experiments using phospho-mimetic mutants in cells have already been conducted (Taguchi et al., 2007; Qi et al., 2011; Yu et al., 2011; Strack et al., 2013; Kashatus et al., 2015; Serasinghe et al., 2015; Xu et al., 2016; Brand et al., 2018; Han et al., 2020, Chang and Blackstone, 2007; Cribbs and Strack, 2007; Cereghetti et al., 2008; Wikstrom et al., 2013, Han et al., 2008，Wang et al., 2012 Jhun BS, Sheu, 2018, J Physiol). To conduct more targeted tests examining specific forms of Drp1 activation in cells (for example, through Mff, MiD proteins, actin, or cardiolipin) will require extensive work that is outside the scope here. Our feeling is that S579 phosphorylation is likely to recruit another molecule (probably a protein) that has an activating effect. We tried to test one possibility (NME3, mentioned in the Discussion) but failed to produce useable NME3 protein for these tests and, given time constraints for the lead author, could not address this further.

    The authors discussed the known factors that involved in Drp1 activation, such as its receptors, actin and cardiolipin. Recent JCB paper (J. Cell Biol. 2023 Vol. 222 No. 10 e202303147) indicates that intermembrane space protein Mdi1/Atg44 may play a role in coordinating mitochondria fission with Dnm1 (Drp1 in yeast cells). It will be valuable if the manuscript could also discuss the potential factor.

    • Thank you for this comment. We now include Mdi1/Atg44 as a possible factor that might be influenced by Drp1 phosphorylation. Two points we would like to make here are: there doesn’t seem to be an Mdi1 homologue in mammals, so the equivalent factor must be identified before testing; and Mdi1 is an inter-membrane space protein, so any effect of Drp1 phosphorylation on coordinated functioning with Mdi1 would either require an intermediary factor or exposure of the IMS in some way.

    Keywords cannot represent the manuscript. It is recommended that the authors use other words to for the current manuscript.

    We have removed K38A from this list. The other key words are not mentioned in the Abstract.

    Reviewer 4

    The authors showed that the binding of Drp1 to actin depends on salt concentrations (Fig. 2Band C). In the presence of 65 mM NaCl, the phosphomimetic mutants showed decreased binding to actin. The GTPase assay is performed with 65 mM KCl, in which actin did not stimulate GTP hydrolysis of the phosphomimetic mutants. In contrast, with 140 mM NaCl, the S579D Drp1 exhibits slightly enhanced actin binding compared to WT Drp1. Could the authors assess the actin-activated GTPase activity in the 140 mM salt condition to test if actin activates GTP hydrolysis ofS579D Drp1 more potently than WT?

    This is a good point by the reviewer. However, with limited time for the first author, we chose to focus on the reviewer’s other comments (see below).

    Both phosphomimetic mutants show reduced activation for GTP hydrolysis in the presence of cardiolipin, Mff, and MiD49. Is this because the mutants have a lower affinity for these interactors? Or do they bind with the same affinity but experience diminished activation? The data suggests the latter scenario, potentially resulting from decreased oligomerization properties. Can the authors provide more insights on this, for example, by measuring their interaction in the presence of GMP- PCP, which fully induces oligomerization in all three forms of Drp1?

    • These are interesting ideas, and we conducted experiments similar to what the reviewer described: co-sedimentation experiments with combinations of Drp1 and Mff under three nucleotide states: no nucleotide, GMP-PCP, and GTP. We used Mff for these experiments because we have this protein in abundance, and have previously characterized this construct as a trimer in PMID 34347505. We use a high concentration of Mff (50 mM) versus Drp1 (1.3 mM) because of the relatively low affinity between the two proteins (shown in PMID 34347505). We find the following:
    1. In the absence of nucleotide, Mff does not cause an increase in pelletable Drp1 for any of the Drp1 constructs.
    2. In the GTP state, the presence of Mff greatly increases the amount of Drp1 in the pellet, suggestive of increased Drp1 oligomerization. This effect occurs for all Drp1 constructs (WT, S579D and S600D mutants), but the amounts of both Drp1 and Mff in the pellets are about 50% less for both mutants than for the WT construct. This result suggests a decrease in oligomerization for the mutants, but not necessarily a decrease in Mff binding.

    I'm curious what happens to oligomerization if GTP is added instead of nonhydrolyzable GMP-PCP (Fig. 1D). Does this lead to higher oligomerization in the mutants compared to WT since the mutants seem to have lower GTPase activity? This might explain why phosphorylation increases mitochondrial localization of Drp1 in cells seen in some studies.

    This is another interesting thought, and we describe the new experiments we conducted in the response to the previous comment. Essentially, while GTP does cause a slight increase in pelletable Drp1, the increase is somewhat similar for all constructs. As described in the last comment, the addition of Mff causes a substantial increase in pelletable Drp1 for both WT and the mutants. This result suggests that, while the basal oligomeric state of Drp1 (in the absence of nucleotide) is reduced for the mutants (our original analytical ultracentrifugation data), the mutants appear to be capable of responding to GTP and Mff in a similar manner to WT. We acknowledge that the assay used here (pelleting) lacks the precision required to draw detailed conclusions on oligomerization or interaction with Mff, and we try to reflect this in our discussion of the data. We do feel, however, that these data are useful to report, in guiding future study.

    Please include the number of experimental repeats and error bars where applicable.

    We have now added number of experimental repeats and error bars for the graphs in Figure 3D, E, F, G, H; Figure 4 D, F, G, H, I, J; Figure 5 B, C, D, E, F, G; and Figure 6B, C, D.

  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 #4

    Evidence, reproducibility and clarity

    During mitochondrial division, a mechanochemical GTPase, Drp1, interacts with its receptor proteins, phospholipids, and the actin cytoskeleton. These interactions regulate the mitochondrial recruitment of Drp1 and its activities, including oligomerization and GTP hydrolysis. Drp1 undergoes serine phosphorylation at two primary sites (S579 and S600 in human isoform 3). It has been suggested that S579 phosphorylation activates Drp1, while S600 phosphorylation inhibits Drp1. However, the biochemical effects of these phosphorylations on Drp1's activity are mostly unexplored. The current study by Liu et al. addresses this crucial question in extensive biochemical assays using recombinant proteins. First, the authors showed that phosphomimetic Drp1 mutations (S579D or S600D) have a reduced ability to oligomerize. Second, both mutants exhibited decreases in their activation for GTP hydrolysis by actin, and concurrently, they demonstrated reduced binding to actin. Third, the Drp1 phosphomimetic mutants showed decreased activation for GTP hydrolysis by cardiolipin and two receptor proteins, Mff and MiD49. This reduction was also evident when Mff was combined with actin. The authors confirmed these results by phosphorylating WT Drp1 at S579 in vitro using the protein kinase Erk2. The effects of phosphorylation seem consistent across Drp1 isoforms with different alternative exons.

    Specific comments

    1. The authors showed that the binding of Drp1 to actin depends on salt concentrations (Fig. 2B and C). In the presence of 65 mM NaCl, the phosphomimetic mutants showed decreased binding to actin. The GTPase assay is performed with 65 mM KCl, in which actin did not stimulate GTP hydrolysis of the phosphomimetic mutants. In contrast, with 140 mM NaCl, the S579D Drp1 exhibits slightly enhanced actin binding compared to WT Drp1. Could the authors assess the actin-activated GTPase activity in the 140 mM salt condition to test if actin activates GTP hydrolysis of S579D Drp1 more potently than WT?
    2. Both phosphomimetic mutants show reduced activation for GTP hydrolysis in the presence of cardiolipin, Mff, and MiD49. Is this because the mutants have a lower affinity for these interactors? Or do they bind with the same affinity but experience diminished activation? The data suggests the latter scenario, potentially resulting from decreased oligomerization properties. Can the authors provide more insights on this, for example, by measuring their interaction in the presence of GMP-PCP, which fully induces oligomerization in all three forms of Drp1?
    3. I'm curious what happens to oligomerization if GTP is added instead of nonhydrolyzable GMP-PCP (Fig. 1D). Does this lead to higher oligomerization in the mutants compared to WT since the mutants seem to have lower GTPase activity? This might explain why phosphorylation increases mitochondrial localization of Drp1 in cells seen in some studies.
    4. Please include the number of experimental repeats and error bars where applicable.

    Significance

    Overall, the current study provides a comprehensive set of biochemical data for the role of Drp1 phosphorylation using cutting-edge in vitro assays. One might expect that S579 phosphorylation would enhance some of Drp1's action while S600 phosphorylation would show the opposite impact. Unexpectedly and interestingly, the authors found that both phosphorylations decrease Drp1's activities. Therefore, this work significantly advances our mechanistic understanding of mitochondrial division.

  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 #3

    Evidence, reproducibility and clarity

    Summary

    The manuscript aimed to clarify the effects of Drp1 phosphorylation on its activation under different factors, such as receptors, actin and cardiolipin. The manuscript performed experiments with purified proteins to examine the correlation of protein-protein interactions and enzymatic activity. The authors specifically focused on the phosphorylation of s S579 and S600 of Drp1 isoform 3, which is the most abundant in HeLa, HL60 and PC12 cells. The results demonstrated the difference of post-translational modification on S579 and S600. The manuscript went further to suggest that additional factors may exist for the Drp1 activation by S579 phosphorylation.

    Major Concerns

    1. The manuscript provided well-organized biochemistry results for comparisons between phosphorylation of Drp1 S579 and S600. It is the reviewer's comments that the authors may include experiments that manipulate Drp1 phosphorylation at different amino acids in cells. Such experiments will provide strong support for this manuscript.
    2. The authors discussed the known factors that involved in Drp1 activation, such as its receptors, actin and cardiolipin. Recent JCB paper (J. Cell Biol. 2023 Vol. 222 No. 10 e202303147) indicates that intermembrane space protein Mdi1/Atg44 may play a role in coordinating mitochondria fission with Dnm1 (Drp1 in yeast cells). It will be valuable if the manuscript could also discuss the potential factor.

    Minor Concerns

    Keywords cannot represent the manuscript. It is recommended that the authors use other words to for the current manuscript.

    Significance

    The authors provided a model for Drp1 catalytic activity. The splendid results of the manuscript will be interesting to the researchers in the related fields.

  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 #2

    Evidence, reproducibility and clarity

    The results presented here indicate that phosphorylation of either serine residue negatively affects Drp1's tendency to oligomerize and substantially reduces stimulation of its GTPase activity in the presence of actin, adaptor proteins and cardiolipin-containing vesicles.

    Overall, the experiments described here are carried out with rigor and the conclusions drawn are of significance to understanding how phosphorylation regulates Drp1 functions.

    Major Comments:

    1. Phosphorylation of both the serine residues appears to elicit a common effect in that they inhibit Drp1's stimulated GTPase activity. This would suggest that phosphorylation affects Drp1's self-assembly as tightly packed helical scaffolds. Instead of sedimentation analysis, an EM analysis of helical scaffolds on cardiolipin-containing membrane nanotubes or in the presence of soluble adaptors causing Drp1 to form filaments would provide a direct readout for defects in self-assembly.
    2. I am not sure of the rationale for experiments reported in Fig. 7 and 8. If the idea was to test if hetero oligomerization with WT Drp1 rescues defects associated with phosphorylated Drp1 then this could be stated explicitly in the manuscript. GFP-Drp1 is used as a WT mimic but a previous report (PMID: 30531964) indicates that this construct is severely defective in stimulated GTPase assays, much like the K38A mutant. But the rationale of using these constructs is not quite apparent. Is the intention to test if defects seen in the phospho-mimetic mutants of Drp1 can be rescued by the presence of a 'seed' of WT Drp1. If so, then this could be stated explicitly in the manuscript. But regardless, I am not quite sure what this data set achieves in terms of addressing mechanism.
    3. Finally, it would have been nice to see if the phospho-mimetic mutants of Drp1 produce the same effects on mitochondrial structure as those reported earlier. Reanalyzing their effects in a cellular assay becomes important because it would consolidate this work for the readers to evaluate the 'true' effects of phosphorylation on Drp1 functions. If the phospho-mimetic mutants fare in a manner like those previously reported, then it signifies that stimulation in GTPase activity is not a readout that directly correlates with Drp1 functions. If not, then the results presented here would establish a comprehensive analysis of in vitro biochemical activities and in vivo functions of the phospho-mimetic mutants.

    Minor comments:

    1. Previous work reports that the effect of actin on the GTPase activity of Drp1 is biphasic but the binding to actin is not. This is quite confounding, and the authors could perhaps explain why this is the case.
    2. The manuscript cites PMID: 23798729 for expression analysis of slice variants but PMID: 29853636 provides a more compressive analysis. The authors could cite this work.

    Significance

    This manuscript reports an extensive and systematic evaluation of the effects of serine phosphorylation on Drp1's GTPase activity in the presence of actin, adaptor proteins and cardiolipin-containing vesicles. Drp1 is predominantly phosphorylated at two sites, S579 and S600 (numbering based on isoform 1). A large body of literature indicates that phosphorylation at S579 activates Drp1 functions while phosphorylation at S600 inhibits Drp1 functions. But as the authors point out, the effects of phosphorylation on the biochemical functions of Drp1 have not been reported, which is what the present manuscript does.

  5. 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

    The present study aims to delineate the effect of S579 and S600 phosphorylation on Drp1 oligomerisation and GTPase activity. Using phospho-mimetic mutant Drp1 proteins, in conjunction with GTPase activity and phosphorylation assays, as well as size exclusion chromatography, the authors conclude that phosphorylation of residue S579 does not activate Drp1 directly. Notably, the authors did not perform cell-based assays to assess mitochondrial fission. The abstract concludes by stating, "our results suggest that nearest neighbour interactions within the Drp1 oligomer affect catalytic activity". However, this assertion appears to lack clarity and direct support from the presented results. Further clarification or evidence linking the observed data to this conclusion would enhance the overall comprehensibility and validity of the study's findings.

    Major comments

    • The assessment of data reproducibility is currently uncertain due to the absence of replication and statistical analysis in the dataset. It is essential to provide explicit information regarding sample sizes or replicates for all data and figures, data should be presented as mean +/- SD/SEM, and the interpretation of results should be grounded in rigorous statistical analysis. The lack of experimental replicates and statistical analysis in most of the figures presented raises major concerns regarding the validity of the result.
    • Why was only one of the MiD proteins, specifically MiD49, studied, while MiD51 was not included in the investigation?
    • The author suggestion of Drp1 phosphorylation, based on the mobility of protein observed in SDS-PAGE gel (fig 4A, 5A, 6A), is not a sufficiently valid assessment. While western blot analysis is a valid method to assess Drp1 phosphorylation, it is essential to include replicates for semi-quantitation and demonstrate the reproducibility of the results. Moreover, it is recommended to incorporate Western blot analyses to provide additional support for the findings presented in Figures 5 and 6.
    • Data on phosphorylated peptides with replicates experiments should be presented.
    • Please provide additional context or specific details about the GFP-tagged Drp1 protein, such as the protein site where GFP was attached, as well as whether this tag could potentially impact the Drp1 GTPase activity and oligomerisation. Figure 7C and D appear to suggest an increase in the GTPase activity of the GFP-Drp1 protein.
    • An optional experiment that would significantly enhance the biological relevance of the findings presented in the current study is to assess the morphology of mitochondria in cells expressing the phospho-mimetic mutant Drp1 proteins. This experiment would provide valuable insights into the functional consequences of Drp1 S579 and S600 phosphorylation on mitochondrial structure and dynamics.

    Minor comments

    • Provide reference for method on actin polymerisation.
    • Rectify the error in referencing figure 3 panels within the figure legends of Supplemental Fig S1.
    • The inclusion of full length isoform 6 is commendable. However, there is no mentioned of isoform 6 in the method section.
    • Since papers deposited in bioRxiv have not undergone peer review, reference #7 should not be cited as references in scholarly work.
    • Please provide details about the calculation of GTPase activity and the distinctions between the specific GTPase activity and total GTPase activity shown in figure 8D-F.

    Significance

    The current investigation holds promise for advancing our understanding of the impact of post-translational modifications, specifically those occurring at the S579 and S600 sites, on Drp1 activity. Nevertheless, the absence of experimental replication and comprehensive statistical analysis introduces notable concerns regarding the credibility and replicability of the findings.

    Audience: Basic research that focus on mitochondrial morphology and Drp1 biology.

    I lack expertise in velocity analytical ultracentrifugation and, as a result, am unable to provide an assessment regarding the validity and accuracy of the assay.

  6. Version published to 10.1091/mbc.e23-11-0427
  7. Version published to 10.1101/2023.08.20.554022v1 on bioRxiv