Some corrections of the review article by Willis and Rhind
Joel Huberman, Roswell Park Cancer Institute
18 August 2009
The article by Willis and Rhind attempts to provide a detailed overview of several aspects of the S-phase DNA damage checkpoint. One of the biggest strengths of this article is its detailed survey of historical literature concerning effects of DNA damage upon DNA replication (even before the term 'checkpoint' was coined). But that does not mean that this article is perfect. A weakness of this article is that, in their attempt to unify the literature with simple hypotheses, the authors overlooked some of the literature that does not support those hypotheses. The consequence is that the review article, in its current form, is somewhat confusing and at times misleading.
However, we believe that if the reader is made aware of the shortcomings of this article, then the information in this review will be more balanced and will have the potential to provide positive service to the research community.
Here we list some of the issues that we discovered when we read this article. The points raised are presented in the order of their occurrence in the text.
1. Page 3, left column, first paragraph, last sentence. The references [13,14] need to be placed at the position of the comma, just before the word "investigators". Otherwise readers may be led to believe that these references deal with investigations into how replication is affected by DNA damage, which they do not. The references deal exclusively with the process of DNA replication in undamaged cells.
2. Page 4, section on "Checkpoint Regulation of Origin Firing", third paragraph, first two sentences. In fact, strong evidence--obtained in our laboratory by two independent research techniques (flow cytometry and 2D gel analysis of replication intermediates)--shows that, at lower concentrations of MMS, Cdc25 is required for replication slowing in fission yeast ([45], and also Kumar and Huberman, On the slowing of S phase in response to DNA damage in fission yeast. J. Biol. Chem. 279:43574–43580, 2004). In the later publication ([45]), we provide a plausible explanation, based on levels of MMS employed, for why Kommajosyula and Rhind [46] failed to detect a role for Cdc25 in regulating the checkpoint in fission yeast. In that publication, we also show that the requirement for Cdc25 applies both to regulation of origin firing and to regulation of replication fork movement. It is interesting to note that one of the main messages of Willis' and Rhind's review is the extensive conservation of checkpoint responses between yeasts and vertebrates. If these authors had remembered our published reports, they could have strengthened this message.
3. Same paragraph, third (last) sentence. Although the final word is not yet in on how origins are regulated in yeast, the available evidence suggests that they are regulated by Cdc25 and Cdk in fission yeast (see previous paragraph) and by Cdc7-Dbf4 in budding yeast (e.g. Duncker et al., An N-terminal domain of Dbf4p mediates interaction with both origin recognition complex (ORC) and Rad53p and can deregulate late origin firing. PNAS 99:16087-16092, 2002). Thus it is not appropriate to describe the mechanism as "unknown" for either yeast.
4. Next paragraph, last sentence. Although the cited microarray results do indicate "that late-origin inhibition by the S-M checkpoint is a genome-wide phenomenon" in budding yeast, they do not do so for fission yeast. According to [52], only about 1/3 of fission yeast origins are regulated by the S-M checkpoint; in contrast, in budding yeast 2/3 are so regulated. According to [53], only 2% of fission yeast origins are so regulated. A slightly later study (Mickle et al., Checkpoint independence of most replication origins in fission yeast. BMC Mol. Biol. 8:112, 2007) concluded that only about 3% of fission yeast origins are significantly checkpoint-regulated, based on their own data plus a review of the data and conclusions of [52,53]. The checkpoint-regulated origins are primarily located near telomeres and thus are definitely not a "genome-wide" phenomenon.
5. Page 5, first column, third paragraph, lines 10-12. Here the authors mention the roles of ATR and Chk1 in slowing replication forks upon DNA damage, and they cite references [59,60]. However, [59] shows that Tim and Tipin are important for regulation of fork progression; it does not provide any data regarding the role of either ATR or Chk1 in regulating fork progression. The other reference, [60], investigates only the role of Chk1 in the camptothecin-induced S-phase checkpoint.
6. Same paragraph, lines 17-18. This sentence, "p53 is required for fork slowing in response to IR [63]" is correct, and its reference is accurate. The problem here is that Willis and Rhind failed to mention some important details contained in this reference. Shimura et al. [63] used DNA fiber analysis to show that low-dose (1 Gray) IR, which should introduce very few lesions, generates a global inhibition of replication fork movement, and this inhibition is dependent on the checkpoint protein, p53. The dependence of this effect on p53 would explain why many studies in mammalian cells failed to detect an effect of IR on fork movement. Many of those studies were carried out on transformed cell lines, which for one reason or another lack p53 function. An example is HeLa cells, for which loss of p53 function is a consequence of the binding of p53 by the HPV16 protein, E6. This evidence for global inhibition of replication fork movement indicates that one of Willis' and Rhind's favorite hypotheses--that checkpoints affect only the movement of replication forks that encounter damage--cannot be completely correct. We stress that this counter-evidence from [63] does not rule out the Willis and Rhind hypothesis; it only shows that reality is more complex than Willis and Rhind seem to think. In our view, it seems likely that--depending on the organism and the extent of DNA damage--checkpoints can regulate fork movement globally and/or only at forks stalled by damage. Another reason why Willis and Rhind should mention the details of [63] is that those details can help clarify something that is otherwise confusing in their review. As the review is currently written, the reader has no idea why some times mammalian cells retard replication fork movement in response to IR while at other times they don't. Shimura et al.'s [63] discovery that this response requires p53 provides a rational explanation for the two disparate sets of observations.
7. Same paragraph, lines 18-20. Reference [64] describes experiments with cis-platin and UV, not CPT (camptothecin) and UV.
8. Page 5, paragraph overlapping columns 1 and 2 (and elsewhere). One of the confusing aspects of this review is its ambiguity regarding the role of the S-phase damage checkpoint in regulating fork movement. Does the checkpoint slow fork movement, or does the checkpoint speed fork movement? The fact is that the checkpoint can do both. It would have been helpful if Willis and Rhind had clarified the circumstances in which the checkpoint can slow forks and the circumstances in which it can accelerate them, and if they had provided a rationale for these different behaviors. In this paragraph, the observation that the checkpoint can speed forks in cells that aren't subjected to exogenous DNA damaging agents comes as a surprise, since previous paragraphs described how, in the presence of DNA damage, the checkpoint acts to slow forks down.
9. Page 5, section on "Stabilization of stalled forks, hints for fork slowing", line 8. Reference [69] isn't properly cited here. This reference describes experiments testing the effect of Brca2 (not a checkpoint kinase) on the stability of replication intermediates in cells treated with HU (not MMS).
10. Page 6, section on "Recombination, lesion bypass, and fork slowing", first paragraph, last two sentences. Indeed, as stated, vertebrates seem to require recombination to actively slow forks. However, yeast cells do not directly require recombination, as shown in references [70, 80]. Our concern is that, as this paragraph is written, readers are likely to conclude that yeast cells do require recombination. The relationship between recombination and fork slowing is more complex. It would have been helpful if that complexity had been made clear in this introductory paragraph, which would have prevented readers from jumping to erroneous conclusions.
11. Same section, next paragraph, line 7. Reference [83] does not show a relationship between recombination and fork movement. Instead, it implicates recombination in the initiation of DNA replication at origins.
12. Page 6, section on "Lesion density may differentiate between origin-based and fork-based checkpoint responses", first paragraph. We've already commented that the observation by Shimura et al. [63] of a p53-dependent checkpoint that globally regulates replication fork movement in response to low IR doses (very low lesion density) shows that there is not a simple relationship between lesion density and origin-based or fork-based checkpoint responses. Here we wish to point out an additional source of confusion in this paragraph. What do Willis and Rhind mean by "In these model systems"? The term "model systems" usually refers to yeast, but here the authors seem to be talking about mammalian cells. Then, the sentence, "These observations suggest that origin regulation is more important for slowing in vertebrates," is misleading--since (1) as already pointed out, normal (p53+/+) mammalian cells regulate forks even at low IR doses, and (2) in budding yeast origin regulation is a major component of checkpoint slowing, even at high lesion density [47, 66].
13. Page 8, left column, second paragraph. Please note that inhibition of origins can be graded, just like inhibition of replication forks. Whether origins, forks, or both are inhibited by a checkpoint, a more profound effect is expected in response to more severe damage, thresholds not withstanding. Also, the Merrick et al. [96] paper cited twice in this paragraph employed HeLa cells, which lack functional p53. If Merrick et al. had used normal cells, there is a high likelihood that their results and conclusions would have been different.
We hope that the reader will keep these pieces of information in mind while reading this otherwise interesting and thought-provoking review article.
Sanjay Kumar Division of Biology MC 147-75 210 Braun Labs California Institute of Technology 1200 E. California Blvd. Pasadena, CA 91125 sanjay@caltech.edu
Joel Huberman Dept. of Cancer Biology Roswell Park Cancer Institute Buffalo, NY 14263-0001 huberman@buffalo.edu
Competing interests
We (Sanjay Kumar and Joel Huberman) are co-authors of three papers on the intra-S-phase DNA damage checkpoint in fission yeast, two of which were not cited by Willis & Rhind, and the third of which was cited incorrectly. It is possible that our opinions regarding the Willis & Rhind review were influenced by our annoyance at their failure to cite our work. However, we have tried our best to be objective. We hope we have succeeded. Readers should judge for themselves.
Some corrections of the review article by Willis and Rhind
18 August 2009
The article by Willis and Rhind attempts to provide a detailed overview of several aspects of the S-phase DNA damage checkpoint. One of the biggest strengths of this article is its detailed survey of historical literature concerning effects of DNA damage upon DNA replication (even before the term 'checkpoint' was coined). But that does not mean that this article is perfect. A weakness of this article is that, in their attempt to unify the literature with simple hypotheses, the authors overlooked some of the literature that does not support those hypotheses. The consequence is that the review article, in its current form, is somewhat confusing and at times misleading.
However, we believe that if the reader is made aware of the shortcomings of this article, then the information in this review will be more balanced and will have the potential to provide positive service to the research community.
Here we list some of the issues that we discovered when we read this article. The points raised are presented in the order of their occurrence in the text.
1. Page 3, left column, first paragraph, last sentence. The references [13,14] need to be placed at the position of the comma, just before the word "investigators". Otherwise readers may be led to believe that these references deal with investigations into how replication is affected by DNA damage, which they do not. The references deal exclusively with the process of DNA replication in undamaged cells.
2. Page 4, section on "Checkpoint Regulation of Origin Firing", third paragraph, first two sentences. In fact, strong evidence--obtained in our laboratory by two independent research techniques (flow cytometry and 2D gel analysis of replication intermediates)--shows that, at lower concentrations of MMS, Cdc25 is required for replication slowing in fission yeast ([45], and also Kumar and Huberman, On the slowing of S phase in response to DNA damage in fission yeast. J. Biol. Chem. 279:43574–43580, 2004). In the later publication ([45]), we provide a plausible explanation, based on levels of MMS employed, for why Kommajosyula and Rhind [46] failed to detect a role for Cdc25 in regulating the checkpoint in fission yeast. In that publication, we also show that the requirement for Cdc25 applies both to regulation of origin firing and to regulation of replication fork movement. It is interesting to note that one of the main messages of Willis' and Rhind's review is the extensive conservation of checkpoint responses between yeasts and vertebrates. If these authors had remembered our published reports, they could have strengthened this message.
3. Same paragraph, third (last) sentence. Although the final word is not yet in on how origins are regulated in yeast, the available evidence suggests that they are regulated by Cdc25 and Cdk in fission yeast (see previous paragraph) and by Cdc7-Dbf4 in budding yeast (e.g. Duncker et al., An N-terminal domain of Dbf4p mediates interaction with both origin recognition complex (ORC) and Rad53p and can deregulate late origin firing. PNAS 99:16087-16092, 2002). Thus it is not appropriate to describe the mechanism as "unknown" for either yeast.
4. Next paragraph, last sentence. Although the cited microarray results do indicate "that late-origin inhibition by the S-M checkpoint is a genome-wide phenomenon" in budding yeast, they do not do so for fission yeast. According to [52], only about 1/3 of fission yeast origins are regulated by the S-M checkpoint; in contrast, in budding yeast 2/3 are so regulated. According to [53], only 2% of fission yeast origins are so regulated. A slightly later study (Mickle et al., Checkpoint independence of most replication origins in fission yeast. BMC Mol. Biol. 8:112, 2007) concluded that only about 3% of fission yeast origins are significantly checkpoint-regulated, based on their own data plus a review of the data and conclusions of [52,53]. The checkpoint-regulated origins are primarily located near telomeres and thus are definitely not a "genome-wide" phenomenon.
5. Page 5, first column, third paragraph, lines 10-12. Here the authors mention the roles of ATR and Chk1 in slowing replication forks upon DNA damage, and they cite references [59,60]. However, [59] shows that Tim and Tipin are important for regulation of fork progression; it does not provide any data regarding the role of either ATR or Chk1 in regulating fork progression. The other reference, [60], investigates only the role of Chk1 in the camptothecin-induced S-phase checkpoint.
6. Same paragraph, lines 17-18. This sentence, "p53 is required for fork slowing in response to IR [63]" is correct, and its reference is accurate. The problem here is that Willis and Rhind failed to mention some important details contained in this reference. Shimura et al. [63] used DNA fiber analysis to show that low-dose (1 Gray) IR, which should introduce very few lesions, generates a global inhibition of replication fork movement, and this inhibition is dependent on the checkpoint protein, p53. The dependence of this effect on p53 would explain why many studies in mammalian cells failed to detect an effect of IR on fork movement. Many of those studies were carried out on transformed cell lines, which for one reason or another lack p53 function. An example is HeLa cells, for which loss of p53 function is a consequence of the binding of p53 by the HPV16 protein, E6. This evidence for global inhibition of replication fork movement indicates that one of Willis' and Rhind's favorite hypotheses--that checkpoints affect only the movement of replication forks that encounter damage--cannot be completely correct. We stress that this counter-evidence from [63] does not rule out the Willis and Rhind hypothesis; it only shows that reality is more complex than Willis and Rhind seem to think. In our view, it seems likely that--depending on the organism and the extent of DNA damage--checkpoints can regulate fork movement globally and/or only at forks stalled by damage. Another reason why Willis and Rhind should mention the details of [63] is that those details can help clarify something that is otherwise confusing in their review. As the review is currently written, the reader has no idea why some times mammalian cells retard replication fork movement in response to IR while at other times they don't. Shimura et al.'s [63] discovery that this response requires p53 provides a rational explanation for the two disparate sets of observations.
7. Same paragraph, lines 18-20. Reference [64] describes experiments with cis-platin and UV, not CPT (camptothecin) and UV.
8. Page 5, paragraph overlapping columns 1 and 2 (and elsewhere). One of the confusing aspects of this review is its ambiguity regarding the role of the S-phase damage checkpoint in regulating fork movement. Does the checkpoint slow fork movement, or does the checkpoint speed fork movement? The fact is that the checkpoint can do both. It would have been helpful if Willis and Rhind had clarified the circumstances in which the checkpoint can slow forks and the circumstances in which it can accelerate them, and if they had provided a rationale for these different behaviors. In this paragraph, the observation that the checkpoint can speed forks in cells that aren't subjected to exogenous DNA damaging agents comes as a surprise, since previous paragraphs described how, in the presence of DNA damage, the checkpoint acts to slow forks down.
9. Page 5, section on "Stabilization of stalled forks, hints for fork slowing", line 8. Reference [69] isn't properly cited here. This reference describes experiments testing the effect of Brca2 (not a checkpoint kinase) on the stability of replication intermediates in cells treated with HU (not MMS).
10. Page 6, section on "Recombination, lesion bypass, and fork slowing", first paragraph, last two sentences. Indeed, as stated, vertebrates seem to require recombination to actively slow forks. However, yeast cells do not directly require recombination, as shown in references [70, 80]. Our concern is that, as this paragraph is written, readers are likely to conclude that yeast cells do require recombination. The relationship between recombination and fork slowing is more complex. It would have been helpful if that complexity had been made clear in this introductory paragraph, which would have prevented readers from jumping to erroneous conclusions.
11. Same section, next paragraph, line 7. Reference [83] does not show a relationship between recombination and fork movement. Instead, it implicates recombination in the initiation of DNA replication at origins.
12. Page 6, section on "Lesion density may differentiate between origin-based and fork-based checkpoint responses", first paragraph. We've already commented that the observation by Shimura et al. [63] of a p53-dependent checkpoint that globally regulates replication fork movement in response to low IR doses (very low lesion density) shows that there is not a simple relationship between lesion density and origin-based or fork-based checkpoint responses. Here we wish to point out an additional source of confusion in this paragraph. What do Willis and Rhind mean by "In these model systems"? The term "model systems" usually refers to yeast, but here the authors seem to be talking about mammalian cells. Then, the sentence, "These observations suggest that origin regulation is more important for slowing in vertebrates," is misleading--since (1) as already pointed out, normal (p53+/+) mammalian cells regulate forks even at low IR doses, and (2) in budding yeast origin regulation is a major component of checkpoint slowing, even at high lesion density [47, 66].
13. Page 8, left column, second paragraph. Please note that inhibition of origins can be graded, just like inhibition of replication forks. Whether origins, forks, or both are inhibited by a checkpoint, a more profound effect is expected in response to more severe damage, thresholds not withstanding. Also, the Merrick et al. [96] paper cited twice in this paragraph employed HeLa cells, which lack functional p53. If Merrick et al. had used normal cells, there is a high likelihood that their results and conclusions would have been different.
We hope that the reader will keep these pieces of information in mind while reading this otherwise interesting and thought-provoking review article.
Sanjay Kumar
Division of Biology MC 147-75
210 Braun Labs
California Institute of Technology
1200 E. California Blvd.
Pasadena, CA 91125
sanjay@caltech.edu
Joel Huberman
Dept. of Cancer Biology
Roswell Park Cancer Institute
Buffalo, NY 14263-0001
huberman@buffalo.edu
Competing interests
We (Sanjay Kumar and Joel Huberman) are co-authors of three papers on the intra-S-phase DNA damage checkpoint in fission yeast, two of which were not cited by Willis & Rhind, and the third of which was cited incorrectly. It is possible that our opinions regarding the Willis & Rhind review were influenced by our annoyance at their failure to cite our work. However, we have tried our best to be objective. We hope we have succeeded. Readers should judge for themselves.