Identification of ubiquitin Ser57 kinases regulating the oxidative stress response in yeast

Article activity feed

1. 

   eLife

   Jun 22, 2020

   ###This manuscript is in revision at eLife

   The decision letter after peer review, sent to the authors on May 19, 2020, follows.

   This manuscript sets out to identify a role for ubiquitin phosphorylation and to identify the kinases necessary for it. The same group has previously shown that serine 57 phosphorylation can be detected in yeast cells. Here they generate strains expressing only phosphomimetic or phosphonull mutants and asses their phenotype in terms of Ubiquitin linkage alone and effect on cell physiology. Among other phenotypes, they find that a strain expressing a non phosphorylable ubiquitin likely fails to mount a response to low doses of H2O2, leading to a slightly increased sensitivity to this chemical. They also find that treatment with H2O2 slightly increases the amount of phosphorylated Ubiquitin. They then go on to …

   ###This manuscript is in revision at eLife

   The decision letter after peer review, sent to the authors on May 19, 2020, follows.

   This manuscript sets out to identify a role for ubiquitin phosphorylation and to identify the kinases necessary for it. The same group has previously shown that serine 57 phosphorylation can be detected in yeast cells. Here they generate strains expressing only phosphomimetic or phosphonull mutants and asses their phenotype in terms of Ubiquitin linkage alone and effect on cell physiology. Among other phenotypes, they find that a strain expressing a non phosphorylable ubiquitin likely fails to mount a response to low doses of H2O2, leading to a slightly increased sensitivity to this chemical. They also find that treatment with H2O2 slightly increases the amount of phosphorylated Ubiquitin. They then go on to identify the kinases responsible for this phosphorylation using a screen in E. coli, which homes in two kinases, Vhs1 and Sks1.

   They delete both kinase and show that this, to a large extent phenocopies the expression of a non-phosphorylable Ubiquitin, and that expression of a phosphomimetic rescued some of the phenotypes of the kinase deleted strain. They also show that overexpressing one of the kinase increases the amount of phosphorylation on ubiquitin.

   Finally, they perform a similar screen using human kinases and human ubiquitin and identify a family of kinases that have the ability to phosphorylate ubiquitin in E. coli.

   All three reviewers found the work of interest. Yet, because pSer57-Ubiquitin is so rare, they expressed concerns that the observed phosphorylation of Ubiquitin could be an epiphenomenon of little incidence to cell function.

   First, the phenotypes of the alanine and aspartate mutants may be due to general effects on Ubiquitin rather than true phospho-Null and -mimetics effects. This concern is minimized by showing that the deletion and overexpression of the kinases phenocopy the ubiquitin mutants. Indeed analysis of the ubiquitin mutant is only valid in the light of this phenocopy. Yet, because of its importance, this point can and should be pushed further. For instance, while the asp mutant is sensitive to hydroxyurea, the ala mutant behaves as a wildtype. This is at odds with the fact that the KO of each kinase individually increase HU resistance. In this case at least, the effect of deleting the kinase does not appear to involve a decrease in the level of ser57 phosphorylation. How can this be reconciled? Also, while you show that expressing the 57Asp bypasses the need for the kinase in the H2O2 sensitivity assay, is it also the case for the HU and tunicamicin resistance bestowed by the deletion of the kinases? Please find in the specific points a list of experiments required to better pinpoint the phenocopy that is so essential for the relevance of this study. Also, overexpression Vhs1 causes a slight canavanine resistance, reminiscent of canavanine resistance confered by s57d expression. Vhs1 overexpression should therefore not confer canavanine resistance if expressed in a s57a background. This is important to strengthen the phenocopy.

   Second, while you show that both kinases can phosphorylate Ubiquitin in bacteria and in vitro, and that the overexpression of one of them increases the phosphorylation levels, you do not show how deletion of the kinase affect phosphorylation. This can and should be done, in particular to show if the observed increase in phosphorylation upon oxidative stress is mediated by these kinases.

   Third, given the low abundance of pS57 ubiquitin, it is hard to conceive that this modification has an important effect on global chain linkage, unless this rare modification is applied to an equally rare set of substrates (like for instance PINK-1 mediated phosphorylation of ubiquitin is limited to the pool of ubiquitin that is on mitochondrial proteins). This should be better emphasized throughout the manuscript so as not to mislead readers into believing that a substantial fraction of ubiquitin is subjected to phosphorylation.

   Fourth, in many cases, the experiments are not described in a sufficient amount of detail. For instance, vectors used herein are not described anywhere, nor is the way that all copies of ubiquitin have been replaced with mutant forms. The supplementary table 2 is absent, so is supplementary table 1. A much better methods section is required to ensure the reproducibility of the experiments. Better descriptions of numbers pertaining to quantitative analysis, statistical test employed an p-value threshold, description of error bars (Stdev, SEM...) are also needed in figures and legends.

   Here are other major points.

   1. In Figure 1A and Figure 3 supplement 1, the authors test the effect of ubiquitin phospho-mutants and absence of kinases, respectively, on the ability of yeast cells to recover from acute heat stress. Firstly, it is puzzling though the experimental conditions are the same (39ᵒC for 18 hours and shifted back to 26ᵒC for recovery) in both cases, the wild-type strain is as good as dead in Figure 1A while it grows fine in Figure 3 supplement 1. Importantly, to validate the resistance phenotype of the S57D mutant, the authors should rather over-express the kinases and see that cells grow better in this condition compared to the wild-type and much better compared to the S57A mutant.

   2. In Figure 1F, the authors employ anti-K48 ubiquitin and anti-K63 ubiquitin antibodies to show the specificity varies between S57A and the S57D mutant. The concern here is whether the serine mutants affect the binding of the antibodies. For instance, the epitope recognized by the anti-K63 ubiquitin antibody could involve the serine 57, however, when mutated to aspartate, the antibody can lose its ability to bind K63-linked ubiquitin. Is there a way to rule this out?

   3. The authors show that S57D increase K48 but decrease K63-linkages whereas S57A decrease K48 but increase K63-linkages upon H2O2 treatment (Figure 1F). What about overexpression or deletion of Vhs1 and Sks1? Does absence of the kinases impact the mutual abundance of ubiquitin K48 and K63 linkages in vivo? Gly-Gly peptides analysis of the data in the experiment from Figure 2G might answer this.

   4. Deletion of the kinases increase resistance to tunicamycin. However expression of S57A does not. To strengthen the case of the phenocopy, it is important to check if kinases have ubiquitin-independet effects and how much of the phenocopy is actually wrought by independent mechanisms.

   5. In general, the growth assays on tunicamycin, hydroxyurea or canavanin in F1S1, F3S2, F3S3 and F3S4 should rather be moved to the main figures.

   6. In Figure 4, human MARK kinases are found to trigger phosphorylation on UbS57 in vitro. It would be insightful to validate this finding in vivo and check whether phosphorylation of UbS57 also regulate the oxidative stress response in mammalian cells. I understand however that this might be take much longer to do than the timeframe which is allocated for revision. In this context, the authors may consider avoiding finishing the paper with these preliminary mammalian data and move them elsewhere in the manuscript. For instance, splitting data from Figure 2 in Figures 2 and 3 and moving figure 4C in the new Figure 2 (and figures 4A and B in supplement) would save some space to end the paper with the current Figure 3 and its supplements.

   Read the original source
2. 

   Version published to 10.1101/2020.06.20.162883 on bioRxiv

   Jun 20, 2020