Is ATP the Only Nucleoside Triphosphate among ATP, CTP, GTP, and UTP to Have a Role in Kinase Catalysis of Heme-Regulated Inhibitor toward eIF2α during Lung Cancer Development?
, Artur Sergunin, AlžbÄ›ta Farná, Tomáš Ovad, Toru Shimizu and Markéta MartÃnková *
Toni Kühl
Round 1
Reviewer 1 Report
Vávra et al. show in their manuscript entitled ‘Is ATP the only nucleoside triphosphate among ATP, CTP, GTP, and UTP to have a role in kinase catalysis of heme-regulated inhibitor toward eIF2α during lung cancer development?’ a comparative analysis applying kinetics studies to recombinant HRI to figure out how different nucleoside triphosphates may act as substrates for phosphorylation of eIF2α.
Their study is well-performed and little doubts arise from their kinetics analysis. The overall message is well presented and the manuscript is properly written. However, there are few aspects that need further clarification:
1.) Authors should refer to kcat instead of Vmax in their abstract and throughout their manuscript to avoid misunderstanding as Vmax is enzyme-concentration-dependent. It would also reduce some redundancy such as sentence ‘Moreover, the catalytic constants (turnover numbers) and catalytic efficiency values of the HRI reaction for the various nucleoside triphosphates were calculated from the apparent kinetic constants (Table 2), confirming the conclusion.’ on page 4. This sentence is redundant as the shown data are clearly explained before using only Vmax at always the same enzyme concentration which finally is the same as if the authors would have used kcat. Also, a combined Table from Table 1 and 2 would be recommended to reduce redundancy.
2.) With the errors given for their KM values I disagree calling the increase in affinity for UTP for the G202S mutant a ‘significant’ increase. If at all, it ‘slightly’ increases but in my opinion it remains more or less similar. Please adapt or explain why this is a significant increase!
Did the authors perform competition studies to detect which nucleotide might be more frequently utilized when available at the same time by, e.g., applying radiolabeled nucleotide derivatives?
How do the authors explain the sigmoidal curve shape for CTP?
3.) This reviewer has doubts on the heme-induced inhibition studies shown in the manuscript for several aspects:
- The authors did not cleave the His-tag from their protein. Heme-binding to His-tag of proteins is well described and cannot be excluded as a bias of the obtained results.
- It is not mentioned whether interactions of heme with single compounds, e.g. NTPs, have been tested in advance to avoid side reactions/unwanted interactions?
It would be also good to see an example for the IC50 plots applying heme to inhibit the activated complexes. Could the authors, please, show a corresponding plot as an example?
If the aforementioned aspects regarding heme binding cannot be clarified I would recommend to not include these data in the authors’ manuscript as the heme effect cannot unequivocally be determined with their current model compound.
4.) ‘However, the situation with the pyrimidine nucleoside triphosphates is dramatically different.’ Considering the 10 fold increased kcat/KM value I disagree calling the situation ‘dramatically’ changed.
5.) How did the authors assure appropriate folding of the eukaryotic HRI protein and the G202S mutant in E. coli? Are there structural data available?
Author Response
Reviewer: 1
Vávra et al. show in their manuscript entitled ‘Is ATP the only nucleoside triphosphate among ATP, CTP, GTP, and UTP to have a role in kinase catalysis of heme-regulated inhibitor toward eIF2α during lung cancer development?’ a comparative analysis applying kinetics studies to recombinant HRI to figure out how different nucleoside triphosphates may act as substrates for phosphorylation of eIF2α.
Their study is well-performed and little doubts arise from their kinetics analysis. The overall message is well presented and the manuscript is properly written.
Response: We thank reviewer # 1 very much for their kind words.
However, there are few aspects that need further clarification:
1.) Authors should refer to kcat instead of Vmax in their abstract and throughout their manuscript to avoid misunderstanding as Vmax is enzyme-concentration-dependent. It would also reduce some redundancy such as sentence ‘Moreover, the catalytic constants (turnover numbers) and catalytic efficiency values of the HRI reaction for the various nucleoside triphosphates were calculated from the apparent kinetic constants (Table 2), confirming the conclusion.’ on page 4. This sentence is redundant as the shown data are clearly explained before using only Vmax at always the same enzyme concentration which finally is the same as if the authors would have used kcat. Also, a combined Table from Table 1 and 2 would be recommended to reduce redundancy.
Response: We thank reviewer # 1 very much for their helpful input. The abstract and the manuscript were adjusted as suggested. The above-mentioned sentence was removed. Also, the new Table 1 was constructed according to the reviewer comments.
2.) With the errors given for their KM values I disagree calling the increase in affinity for UTP for the G202S mutant a ‘significant’ increase. If at all, it ‘slightly’ increases but in my opinion it remains more or less similar. Please adapt or explain why this is a significant increase!
Response: We thank reviewer # 1 and according to their suggestion we change the sentences as follows (a) on page 4, line 133 (Results) - “The apparent Michaelis constant decreased for the HRI G202S mutant compared to the HRI WT enzyme with regard to UTP (KMUTP), suggesting slightly higher affinity of the HRI G202S mutant to UTP than was observed for the HRI WT.” and (b) on page 8, line 263 (Discussion) – „However, in the case of UTP, the G202S mutation associated with cancer development slightly increased the affinity of the HRI to UTP compared to the HRI WT enzyme.“
Did the authors perform competition studies to detect which nucleotide might be more frequently utilized when available at the same time by, e.g., applying radiolabeled nucleotide derivatives?
Response: Thank you very much for the valuable suggestion. However, please note that our laboratory is not validated for handling radiolabeled nucleotides. We analyzed the kinetics of the HRI reaction by the phos-tag electrophoresis. The method can separate phosphorylated and non-phosphorylated forms of a protein (eIF2α in our case). Therefore, it was not necessary to utilize radiolabeled nucleotides. At the set conditions it is impossible to study a competition between nucleotides. However, in future studies we will try to develop a method to test the reviewer’s interesting idea.
How do the authors explain the sigmoidal curve shape for CTP?
Response: We thank reviewer # 1 for a very interesting question. It is evident that pyrimidine nucleotides are worse substrates than the purine ones. However, we were surprised that the pyrimidine nucleotides can serve as the phosphate donors for the HRI kinase reaction at all. Moreover, a specific mutation of HRI, namely G202S, is proposed to influence HRI affinity to ATP, given that the mutation position is located close to the supposed ATP binding site. Interestingly, we have observed the sigmoidal curve shape only for CTP only in the case of the HRI G202S. We think that the amino groups on the purine position 6 and pyrimidine position 4 in ATP and CTP, respectively could be the key to explain that sigmoidal observation. Because the important glycine 202 is mutated to serine in HRI G202S, the interaction with that pyrimidine nucleotide is changed causing the different behavior of the kinetics function. However, the possibility that the level of enzyme reaction is in this case so low, that the measurement is laden with great experimental error could also not be ruled out.
3.) This reviewer has doubts on the heme-induced inhibition studies shown in the manuscript for several aspects:
- The authors did not cleave the His-tag from their protein. Heme-binding to His-tag of proteins is well described and cannot be excluded as a bias of the obtained results.
Response: The reviewer’s opinion is reasonable. However, it is important to note that we compare the behavior of HRI free from His-tag (as was studied previously and the results are described in the reference number 12, namely Miksanova, M. et al. Biochemistry 2006, 45, 9894–9905, doi:10.1021/bi060556k.) and the current His-tag containing form. And because both forms of HRI behave in a similar way, we decided not to remove the His-tag from the currently studied proteins. Moreover, it is also possible to distinguish if the heme interacts with the specific His and Cys axial ligands or if it binds to His-tag by UV-VIS spectroscopy. And we checked that the His-tag does not interact with heme and therefore does not interfere with our study.
- It is not mentioned whether interactions of heme with single compounds, e.g. NTPs, have been tested in advance to avoid side reactions/unwanted interactions?
Response: Thank you very much for your valuable suggestions. However, please note that it is highly unlikely that heme could directly interact with nucleotides. We are studying a wide range of enzymes that convert/utilize various nucleotides (ATP – kinase, GTP – diguanylate cyclases, c-di-GMP- phosphodiesterase) and neither from our experience nor from literature there are any suggestions/clues that nucleotides can interact with heme. We must honestly answer the reviewer’s question, that we have not tested if nucleotides interact with heme, however, from our expertise, such a situation is highly improbable.
- It would be also good to see an example for the IC50 plots applying heme to inhibit the activated complexes. Could the authors, please, show a corresponding plot as an example?
Response: Thank you very much for your valuable suggestions. We added a new Figure 3 showing the IC50 plots HRI WT and HRI G202S in the presence of ATP.
If the aforementioned aspects regarding heme binding cannot be clarified I would recommend to not include these data in the authors’ manuscript as the heme effect cannot unequivocally be determined with their current model compound.
Response: The reviewer’s opinion is reasonable. However, it is important to note that the interaction with heme is the essence of the HRI function. It is a heme-regulated eukaryotic initiation factor 2α kinase. The interaction with heme was described in detail in several previous papers (for example number 10 and 12 from the manuscript list of references). We know, where and how heme interacts with HRI (one axial ligand, histidine, is in the N-terminal domain, while the other one, cysteine, is in the functional domain) and the sensing site is far away from the ATP and eIF2α binding sites. Anyway, we would prefer to include at least the results that both HRI enzymes (WT and G202S) are heme-regulated and that the heme inhibition of the kinase reaction is preserved without regard to the utilized substrate (in the case of all nucleotides). We hope that we have explained all the aspects above clearly. Anyway, if the reviewer and the editor should insist on the removal of this part, we will definitively follow their recommendation.
4.) ‘However, the situation with the pyrimidine nucleoside triphosphates is dramatically different.’ Considering the 10 fold increased kcat/KM value I disagree calling the situation ‘dramatically’ changed.
Response: We thank reviewer # 1 very much for their comments. The expressive word “dramatically” was removed, and the sentence is now as follows (page 8, line 257): „However, the situation with the pyrimidine nucleoside triphosphates is different.”
5.) How did the authors assure appropriate folding of the eukaryotic HRI protein and the G202S mutant in E. coli? Are there structural data available?
Response: Thank you very much for your valuable comments which are probably related to the fact, that the eukaryotic (even human) HRI was produced in prokaryotic (bacterial) cells. However, please note that we have a lot of experience with similar heterologous expressions. We have checked the UV-VIS spectra of the heme-bound forms of both HRI forms. It was obvious that the heme coordinates our proteins in the same way as described previously and that the coordination is specific. It could not be possible to observe that if the folding was impaired. Also, the specific interaction with its substrate, eIF2α is reserved only for the appropriately folded proteins. Unfortunately, the HRI structure is very flexible, as is typical for all heme sensing proteins and therefore there is no structure available. All discussions concerning the structural features are based on the indirect approaches.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The manuscript provides data that demonstrate HRI kinase nucleotide specificity. The finding that HRI kinase prefers ATP as substrate is highly predictable, as all kinases display such preference. The interesting part of the manuscript is the demonstration that HRI kinase, within the normal physiological range of nucleotides, can use GTP reasonably well and even can use CTP or UTP. The presented data are supportive of the conclusion that were drawn. Clearly, this manuscript will add some knowledge about the HRI kinase and should be reported.
However, there are some issues that the authors need to address. The major problem is that there is no explanation why the assays were done at 20oC, when the standard enzymatic reactions are usually performed at 37oC. At physiologic temperatures, the values for ATP, GTP, CTP and UTP may be different, and the observed affinity may also be different.
There are also several minor issues that need to be addressed. The presented data would be also more convincing, if some raw data containing the phosphorylated eIF2a are shown, even as an example. It would be very informative, if the data demonstrating that the initial rates are indeed calculated based on the linear section of the curve are presented in some form. Finally, it is not clear what were the concentrations of nucleotides in experiments described in Figure 1.
Author Response
Reviewer: 2
The manuscript provides data that demonstrate HRI kinase nucleotide specificity. The finding that HRI kinase prefers ATP as substrate is highly predictable, as all kinases display such preference. The interesting part of the manuscript is the demonstration that HRI kinase, within the normal physiological range of nucleotides, can use GTP reasonably well and even can use CTP or UTP. The presented data are supportive of the conclusion that were drawn. Clearly, this manuscript will add some knowledge about the HRI kinase and should be reported.
Response: We thank reviewer # 2 very much for their kind comments and helpful input.
However, there are some issues that the authors need to address. The major problem is that there is no explanation why the assays were done at 20oC, when the standard enzymatic reactions are usually performed at 37oC. At physiologic temperatures, the values for ATP, GTP, CTP and UTP may be different, and the observed affinity may also be different.
Response: We thank reviewer # 2 for raising this very important point. We absolutely agree that human HRI works at 37℃ and our kinetics experiments were performed at room temperature (specifically at 20℃). The reason for that discrepancy is dictated by the recombinant HRI stability. We were thinking about that discrepancy also at the beginning of that study when the experiments were designed. We decided to apply the same reaction conditions as previously published in studies either from our laboratory or others, to be able to compare our recent results with those previous achievements (references 10, 12 and 27):
- Igarashi, J.; Murase, M.; Iizuka, A.; Pichierri, F.; Martinkova, M.; Shimizu, T. Elucidation of the Heme Binding Site of Heme-Regulated Eukaryotic Initiation Factor 2alpha Kinase and the Role of the Regulatory Motif in Heme Sensing by Spectroscopic and Catalytic Studies of Mutant Proteins. J. Biol. Chem. 2008, 283, 18782–18791, doi:10.1074/jbc.M801400200.
- Miksanova, M.; Igarashi, J.; Minami, M.; Sagami, I.; Yamauchi, S.; Kurokawa, H.; Shimizu, T. Characterization of Heme-Regulated EIF2alpha Kinase: Roles of the N-Terminal Domain in the Oligomeric State, Heme Binding, Catalysis, and Inhibition. Biochemistry 2006, 45, 9894–9905, doi:10.1021/bi060556k.
- Ricketts, M.D.; Emptage, R.P.; Blobel, G.A.; Marmorstein, R. The Heme-Regulated Inhibitor Kinase Requires Dimerization for Heme-Sensing Activity. J Biol Chem 2022, 298, 102451, doi:10.1016/j.jbc.2022.102451.
In our previous studies, the reaction temperature was even 15℃ (the first two papers above), because the new results concerning HRI were obtained at a slightly higher temperature (the third paper quoted above), we tested that also 20℃ is an appropriate condition for the assessment of HRI enzyme activity under various other conditions. Anyway, we believe that even if the specific values for ATP, CTP, GTP and UTP and their affinities may be different between observed values at room temperature (specifically at 20℃) and the values at the physiological temperature (specifically at 37℃), their mutual relationship and indicated behavior could be similar at 20 and 37℃.
Since it is definitely a very important point, we added the following sentences into the Discussion (page 7, line 235): „It should be noted that native human HRI works at 37℃ and our kinetics experiments were performed at 20℃, because of the enzyme stability. Anyway, we believe that even if the specific values for ATP, CTP, GTP and UTP and their affinities to HRI may differ at various temperatures, their mutual relationship and indicated behavior are similar at 20 and 37℃.“
There are also several minor issues that need to be addressed. The presented data would be also more convincing, if some raw data containing the phosphorylated eIF2a are shown, even as an example.
Response: We agree with reviewer # 2 and an example of raw data containing the non-phosphorylated and phosphorylated form of the HRI substrate, i.e. eIF2α are shown in new (revised) Figure 1, above each time course of eIF2α phosphorylation catalyzed by either HRI WT or HRI G202S mutant. As was added to the Figure legend, each raw data example represents one experiment with ATP as a substrate.
It would be very informative, if the data demonstrating that the initial rates are indeed calculated based on the linear section of the curve are presented in some form.
Response: We thank reviewer # 2 very much for their helpful input. In the Materials and Methods section is stated (page 10, line 387) that the initial velocity conditions for ATP and GTP are 1 min and for CTP and UTP 2 min. That was estimated from the time course experiments performed for both enzymes (HRI WT and HRI G202S) and for all substrates at their highest concentrations later utilized in the kinetics experiments. Therefore, the amount of reaction product (phosphorylated eIF2α, P-eIF2α) formed per min during the first 1 min (ATP and GTP) or 2 min (CTP and UTP) was equivalent to the initial velocity of the kinase reaction. In order to follow the reviewer recommendation, we have added additional inboxes below the Figure 1 plots zooming in the initial times of the presented functions, i.e. 1 min, respectively 2 min, to demonstrate the linear section of the curves.
Finally, it is not clear what were the concentrations of nucleotides in experiments described in Figure 1.
Response: We thank reviewer # 2 and we have immediately corrected our omission. Both, in the Figure 1 legend (page 3) and in the Materials and Methods section (page 10, line 376), the information was added. The time courses of the kinase reactions catalyzed by the HRI WT enzyme and the HRI G202S mutant were performed with the highest nucleotides concentrations utilized in the study i.e., 900 µM ATP, CTP, GTP or UTP (in the Materials and Methods section was initially only the information about ATP). We are sorry for the typo, and we are happy that the reviewer alerted us about that issue so we could immediately fix it.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
I would like to thank the authors for their insightful revision. Most of the points raised earlier have been addressed appropriately. However, three aspects still should be addressed to clearly avoid misunderstandings or misinterpretation of the data in the authors’ manuscript:
1.) The sigmoidal curve shape for CTP interaction should be briefly discussed in their manuscript as stated in the authors’ revision.
2.) I accept the authors’ discussion on the effect of heme-binding to a His-tagged protein assuming that most likely the affinity to the protein is sufficiently high to ignore binding of heme to the tag (or at least reducing it to a very minor amount). However, I would still insist on checking binding of heme to the nucleotides, especially because the concentrations of heme (1-200 µM) and nucleotides (900 µM) were chosen so high compared to the HRI concentration (0.35 µM) that the interaction between nucleotides and heme should not be competitive to HRI interaction. This could be done by, e.g., checking the shift of the Soret band in titration experiments monitored by UV/Vis spectroscopy or other techniques. This experiment is in my opinion mandatory especially because of the selection of such high nucleotide concentrations which in case of physiological interactions and concentrations might be important to know to evaluate the bias that has to be considered for the author’s results regarding their nucleotides’ effects on the activation of HRI. Especially in the light of availability of heteroatoms in nucleotides heme binding is an option that cannot be excluded.
Alternatively, I would also accept if the authors mention these concerns and that it might impact the precision of their data and, thus, only provides a rough idea what the ratios of IC50 values potentially are like. However, it should be mentioned somewhere that heme may bind more strongly to purine nucleotides (which may lower their available concentration) and therefore IC50 values may be higher.
3.) Regarding the discussion of the protein structures I was wondering if CD spectroscopy had been performed to compare HRI and its mutant among each other and at best to native material if available? This could help to get at least an idea on the availability of secondary structural elements and to understand if the G202S mutation impaired protein structure.
As a minor comment I would like to thank the authors for adding the IC50 plot example in figure 3. But, unfortunately, linkage to Figure 3 in the text is missing. Please add!
Author Response
Reviewer: 1
I would like to thank the authors for their insightful revision. Most of the points raised earlier have been addressed appropriately.
Response: We thank reviewer # 1 very much for their kind words and valuable suggestions, which have helped us to improve the manuscript significantly.
However, three aspects still should be addressed to clearly avoid misunderstandings or misinterpretation of the data in the authors’ manuscript:
1.) The sigmoidal curve shape for CTP interaction should be briefly discussed in their manuscript as stated in the authors’ revision.
Response: We thank reviewer # 1 and according to their suggestion we added our explanation concerning the sigmoidal curve shape for CTP interaction into the discussion section (page 7, from the line 253): „Interestingly, we have observed the sigmoidal curve shape only for CTP only in the case of the HRI G202S. We think that the amino groups on the purine position 6 and pyrimidine position 4 in ATP and CTP, respectively could be the key to explain that sigmoidal observation. Because the important glycine 202 is mutated to serine in HRI G202S, the interaction with that pyrimidine nucleotide is changed causing the different behavior of the kinetics function. However, the possibility that the level of enzyme reaction is in this case so low, that the measurement is laden with great experimental error could also not be ruled out.”
2.) I accept the authors’ discussion on the effect of heme-binding to a His-tagged protein assuming that most likely the affinity to the protein is sufficiently high to ignore binding of heme to the tag (or at least reducing it to a very minor amount). However, I would still insist on checking binding of heme to the nucleotides, especially because the concentrations of heme (1-200 µM) and nucleotides (900 µM) were chosen so high compared to the HRI concentration (0.35 µM) that the interaction between nucleotides and heme should not be competitive to HRI interaction. This could be done by, e.g., checking the shift of the Soret band in titration experiments monitored by UV/Vis spectroscopy or other techniques. This experiment is in my opinion mandatory especially because of the selection of such high nucleotide concentrations which in case of physiological interactions and concentrations might be important to know to evaluate the bias that has to be considered for the author’s results regarding their nucleotides’ effects on the activation of HRI. Especially in the light of availability of heteroatoms in nucleotides heme binding is an option that cannot be excluded.
Alternatively, I would also accept if the authors mention these concerns and that it might impact the precision of their data and, thus, only provides a rough idea what the ratios of IC50 values potentially are like. However, it should be mentioned somewhere that heme may bind more strongly to purine nucleotides (which may lower their available concentration) and therefore IC50 values may be higher.
Response: Thank you very much for the valuable discussion and sharing with us an interesting point of view. As recommended by the reviewer, we incorporated following sentences into the discussion section of the manuscript (page 6, from the line 226 bellow):”However, because the concentrations of heme and nucleotides utilized in the experimental setting were three orders of magnitudes higher compared to the HRI concentration, an interaction between nucleotides and heme cannot be excluded and if so, such an interaction is not competitive to HRI-heme and HRI-nucleotide interactions. Such situation could hypothetically impact the precision of data. If heme interacts with nucleotides, it lowers its available concentration. Presented data only provides a rough idea what the ratios of IC50 values potentially are like.“
3.) Regarding the discussion of the protein structures I was wondering if CD spectroscopy had been performed to compare HRI and its mutant among each other and at best to native material if available? This could help to get at least an idea on the availability of secondary structural elements and to understand if the G202S mutation impaired protein structure.
Response: We thank reviewer # 1 very much for their comments. Unfortunately, we have not recorded CD spectra neither for HRI WT nor for its mutant not for a native material. Currently, we do not have an excess to that technique. However, we will search for an available machine, and we will employ that spectroscopy for future experiments. Now, we must rely on our experience, on the UV-VIS spectra of the heme-bound forms of both HRI forms and on the specific interaction between HRI and its substrate, eIF2α.
As a minor comment I would like to thank the authors for adding the IC50 plot example in figure 3. But, unfortunately, linkage to Figure 3 in the text is missing. Please add!
Response: We thank reviewer # 1 very much for pointing out this discrepancy. The linkage to Figure 3 was added into the main text (page 5, line 169).
Author Response File: Author Response.pdf
Reviewer 2 Report
In the revised version of the manuscript, the authors addressed the voiced critics. The authors provided the missing information, expanded the Figure 2 to include some examples of phosphorylation assay and provided a reasonable explanation why the assays were done at 20oC.
Author Response
Reviewer: 2
In the revised version of the manuscript, the authors addressed the voiced critics. The authors provided the missing information, expanded the Figure 2 to include some examples of phosphorylation assay and provided a reasonable explanation why the assays were done at 20oC.
Response: We thank reviewer # 2 very much for their kind comments and suggestions, which have helped us to improve the manuscript significantly.
Author Response File: Author Response.pdf
Round 3
Reviewer 1 Report
The authors have addressed all points adequately. I recommend publication in the current state.
