Author response:
Reviewer #1:
Summary:
García-Vázquez et al. identify GTSE1 as a novel target of the cyclin D1-CDK4/6 kinases. The authors show that GTSE1 is phosphorylated at four distinct serine residues and that this phosphorylation stabilizes GTSE1 protein levels to promote proliferation.
Strengths:
The authors support their Kindings with several previously published results, including databases. In addition, the authors perform a wide range of experiments to support their Kindings.
Weaknesses:
I feel that important controls and considerations in the context of the cell cycle are missing. Cyclin D1 overexpression, Palbociclib treatment and apparently also AMBRA1 depletion can lead to major changes in cell cycle distribution, which could strongly inKluence many of the observed effects on the cell cycle protein GTSE1. It is therefore important that the authors assess such changes and normalize their results accordingly.
We have approached the question of GTSE1 phosphorylation to account for potential cell cycle effects from multiple angles:
(i) We conducted in vitro experiments with puriIied, recombinant proteins and shown that GTSE1 is phosphorylated by cyclin D1-CDK4 in a cell-free system (Figure 2A-C). This experiment provides direct evidence of GTSE1 phosphorylation by cyclin D1-CDK4 without the inIluence of any other cell cycle effectors.
(ii) We present data using synchronized AMBRA1 KO cells (Figure 2G and Supplementary Figure 3B). As shown previously (Simoneschi et al., Nature 2021, PMC8875297), AMBRA1 KO cells progress faster in the cell cycle but they are still synchronized as shown, for example by the mitotic phosphorylation of Histone H3. Under these conditions we observed that while phosphorylation of GTSE1 in parental cells peaks at the G2/M transition, AMBRA1 KO cells exhibited sustained phosphorylation of GTSE1 across all cell cycle phases. This is evident when using Phos-tag gels as in the current top panel of Figure 2G. We now re-run one the biological triplicates of the synchronized cells using higher concentration of Zn+2-Phos-tag reagent and lower voltage to allow better separation. Under these conditions, GTSE1 phosphorylation is more apparent. In the new version of the paper, we will either show both blots or substitute the old panel with the new one. This experiment provides evidence that high levels of cyclin D1 in AMBRA1 KO cells affect GTSE1 independently of the speciIic points in the cell cycle.
(iii) The relative short half-life of GTSE1 (<4 hours) makes its levels sensitive to acute treatments such as Palbococlib or AMBRA1 depletion. The effects of these treatments on GTSE1 levels are measurable within a time frame too short to affect cell cycle progression in a meaningful way. For example, we used cells with fusion of endogenous AMBRA1 to a mini-Auxin Inducible Degron (mAID) at the N-terminus. This system allows for rapid and inducible degradation of AMBRA1 upon addition of auxin, thereby minimizing compensatory cellular rewiring. Again, we observed an increase in GTSE1 levels upon acute ablation of AMBRA1 (i.e., in 8 hours) (Figure 3B), when no signiIicant effects on cell cycle distribution are observed (please see Simoneschi et al., Nature 2021, PMC8875297 and Rona et al., Mol. Cell 2024, PMC10997477).
All together, these lines of evidence support our conclusion that GTSE1 is a target of cyclin D1-CDK4, independent of cell cycle effects. In conclusion, as stated in the Discussion section, GTSE1 has been established as a substrate of mitotic cyclins, but we observed that overexpression of cyclin D1-CDK4 induce GTSE1 phosphorylation at any point of the cell cycle. Thus, we propose that GTSE1 is phosphorylated by CDK4 and CDK6 particularly in pathological states, such as cancers displaying overexpression of D-type cyclins beyond the G1 phase. In turn, GTSE1 phosphorylation induces its stabilization, leading to increased levels that, as expected based on the existing literature, contribute to enhanced cell proliferation. So, the cyclin D1-CDK4/6 kinase-dependent phosphorylation of GTSE1 induces its stabilization independently of the cell cycle.
Reviewer #2:
Summary:
The manuscript by García-Vázquez et al identifies the G2 and S phases expressed protein
1(GTSE1) as a substrate of the CycD-CDK4/6 complex. CycD-CDK4/6 is a key regulator of the G1/S cell cycle restriction point, which commits cells to enter a new cell cycle. This kinase is also an important therapeutic cancer target by approved drugs including Palbocyclib. Identification of substrates of CycD-CDK4/6 can therefore provide insights into cell cycle regulation and the mechanism of action of cancer therapeutics. A previous study identified GTSE1 as a target of CycB-Cdk1 but this appears to be the first study to address the phosphorylation of the protein by Cdk4/6.
The authors identified GTSE1 by mining an existing proteomic dataset that is elevated in AMBRA1 knockout cells. The AMBRA1 complex normally targets D cyclins for degradation. From this list, they then identified proteins that contain a CDK4/6 consensus phosphorylation site and were responsive to treatment with Palbocyclib.
The authors show CycD-CDK4/6 overexpression induces a shift in GTSE1 on phostag gels that can be reversed by Palbocyclib. In vitro kinase assays also showed phosphorylation by CDK4. The phosphorylation sites were then identified by mutagenizing the predicted sites and phostag got to see which eliminated the shift.
The authors go on to show that phosphorylation of GTSE1 affects the steady state level of the protein. Moreover, they show that expression and phosphorylation of GTSE1 confer a growth advantage on tumor cells and correlate with poor prognosis in patients.
Strengths:
The biochemical and mutagenesis evidence presented convincingly show that the GTSE1 protein is indeed a target of the CycD-CDK4 kinase. The follow-up experiments begin to show that the phosphorylation state of the protein affects function and has an impact on patient outcomes.
Weaknesses:
It is not clear at which stage in the cell cycle GTSE1 is being phosphorylated and how this is affecting the cell cycle. Considering that the protein is also phosphorylated during mitosis by CycB-Cdk1, it is unclear which phosphorylation events may be regulating the protein.
In cells that do not overexpress cyclin D1, GTSE1 is phosphorylated at the G2/M transition, consistent with the known cyclin B1-CDK1-mediated phosphorylation of this protein. However, AMBRA1 KO cells exhibited high levels of cyclin D1 throughout the cell cycle and sustained phosphorylation of GTSE1 across all cell cycle points (Figure 2G and Supplementary Figure 3B). Please see also answer to Reviewer #1. Moreover, we show that, compared to the amino acids phosphorylated by cyclin D1-CDK4, cyclin B1-CDK1 phosphorylates GTSE1 on either additional residues or different sites (Figure 2H). Finally, we show that expression of a phospho-mimicking GTSE1 mutant leads to accelerated growth and an increase in the cell proliferative index (Figure 4C). However, we have not evaluated how phosphorylation affects the cell cycle distribution. We will perform FACS analyses and include them in the new version.
Reviewer #3:
Summary:
This paper identifies GTSE1 as a potential substrate of cyclin D1-CDK4/6 and shows that GTSE1 correlates with cancer prognosis, probably through an effect on cell proliferation. The main problem is that the phosphorylation analysis relies on the over-expression of cyclin D1. It is unclear if the endogenous cyclin D1 is responsible for any phosphorylation of GTSE1 in vivo, and what, if anything, this moderate amount of GTSE1 phosphorylation does to drive proliferation.
Strengths:
There are few bonafide cyclin D1-Cdk4/6 substrates identified to be important in vivo so GTSE1 represents a potentially important finding for the field. Currently, the only cyclin D1 substrates involved in proliferation are the Rb family proteins.
Weaknesses:
The main weakness is that it is unclear if the endogenous cyclin D1 is responsible for phosphorylating GTSE1 in the G1 phase. For example, in Figure 2G there doesn't seem to be a higher band in the phos-tag gel in the early time points for the parental cells. This experiment could be redone with the addition of palbociclib to the parental to see if there is a reduction in GTSE1 phosphorylation and an increase in the amount in the G1 phase as predicted by the authors' model. The experiments involving palbociclib do not disentangle cell cycle effects. Adding Cdk4 inhibitors will progressively arrest more and more cells in the G1 phase and so there will be a reduction not just in Cdk4 activity but also in Cdk2 and Cdk1 activity. More experiments, like the serum starvation/release in Figure 2G, with synchronized populations of cells would be needed to disentangle the cell cycle effects of palbociclib treatment.
In normal cells, GTSE1 is phosphorylated at the G2/M transition in a cyclin B1-CDK1dependent manner. During G1, when the levels of cyclin D1 peak, GTSE1 is not phosphorylated. This could be due to a higher affinity between GTSE1 and mitotic cyclins as compared to G1 cyclins or to a higher concentration of mitotic cyclins compared to G1 cyclins. We show that higher levels of cyclin D1 induce GTSE1 phosphorylation during interphase, but we do not rely only on the overexpression of exogenous cyclin D1. In fact, we observe similar effect when we deplete endogenous AMBRA1, resulting in the stabilization of endogenous cyclin D1. As mentioned in the Discussion section, we propose that GTSE1 is phosphorylated by CDK4 and CDK6 particularly in pathological states, such as cancers displaying overexpression of D-type cyclins (i.e., the overexpression appears to overcome the lower afIinity of the cyclin D1-GTSE1 complex). In sum, our study suggests that overexpression of cyclin D1, which is often observed in cancers cells beyond the G1 phase, induces phosphorylation of GTSE1 at all points in the cell cycle displaying high levels of cyclin D1. Please see also response to Reviewer #1. Concerning the experiments involving palbociclib, we limited confounding effects on the cell cycle by treating cells with palbociclib for only 4-6 hours. Under these conditions, there is simply not enough time for the cells to arrest in G1.
It is unclear if GTSE1 drives the G1/S transition. Presumably, this is part of the authors' model and should be tested.
We are not claiming that GTSE1 drives the G1/S transition. GTSE1 is known to promote cell proliferation, but how it performs this task is not well understood. Our experiments indicate that, when overexpressed, cyclin D1 promotes GTSE1 phosphorylation and its consequent stabilization. In agreement with the literature, we show that higher levels of GTSE1 promote cell proliferation. To measure cell cycle distribution upon expressing various forms of GTSE1, we will now perform FACS analyses and include them in the new version.
The proliferation assays need to be more quantitative. Figure 4B should be plotted on a log scale so that the slope can be used to infer the proliferation rate of an exponentially increasing population of cells. Figure 4c should be done with more replicates and error analysis since the effects shown in the lower right-hand panel are modest.
In Figure 4B, we plotted data in a linear scale as done in the past (Donato et al. Nature Cell Biol. 2017, PMC5376241) to better represent the changes in total cell number overtime. The experiments in Figure 4C were performed in triplicate. Error analysis was not included for simplicity, given the complexity of the data. We will include the other two sets of experiments in the revised version. While the effects shown in the lower right-hand panel of Figure 4C are modest, they demonstrate the same trend as those observed in the AMBRA KO cells (Figure 4C and Simoneschi et al., Nature 2021, PMC8875297). It's important to note that this effect is achieved through the stable expression of a single phosphomimicking protein, whereas AMBRA KO cells exhibit changes in numerous cell cycle regulators.
We appreciate the constructive comments and suggestions made by the reviewers, and we believe that the resulting additions and changes will improve the clarity and message of our study.
