#2由 sufang 在 六, 05/09/2015 – 20:02 發表。
Reviewers Comments
Reviewers Comments From PLoS Pathogens (06/20–0/7/31, 6 wks)
Reviewer #1: This is an interesting manuscript in which the model is advanced that EBV and KSHV Rta by binding to their respective genomes, induce local methylation which promulgates lytic replication and dissociation of CTCF, which is inhibitory to virus replication. While the model is intriguing, and some supportive data is presented, an underlying issue with the argument is that while Rta expression is temporally associated with such effects, it is not proven that the effects are causally related. In other words, induced Rta expression leads to a panoply of effects, culminating in viral replication and transcription of many genes, as well as well-described effects on cellular physiology. In the absence of more direct evidence, such as mutational analyses that abrogate Rta binding that lead to lack of CTCF binding, for example, these data are overinterpreted to support the model being proposed.
Specific points regarding the results include the following:
1. While there seem to be general trends toward increased methylation upon induction of Rta in EREV and ERKV cells, the effects are not robust and seem to decline in some cases over time, for example JUN AciI digestion from 24 to 48 hours, whereas the same effect is not seen with HpaII, casting some doubt on the quantitative aspects of these assays.
2. I also have some concern about the quantitative aspects of the EtBr-based measurements of ChIP data. For example, one would expect the input values to remain consistent across different experiments if equal amounts of input cells were used in the different IPs. However, the amount clearly varies over time in some samples but not others in Figure 5B. CTCF occupancy actually increases in some cases before declining. The binding of CTCF also changes differentially among the regions with EBV induction. While this may represent a physiologically specific dissociation from some sites versus others, this is not considered and the semiquantitative nature of the assay, combined with the moderate level of the effects, makes one question whether these are true differences.
3. The DNMT associations are very weak in the IPs from infected cells as opposed to the overexpression experiments and it would be more convincing if the percentage of input that was immunoprecipitable were demonstrated and if more of the gel were shown.
Reviewer #2: This paper addresses an interesting question of how Rta can both activate viral and repress host gene expression.
The authors address whether there is a link between Rta binding, CTCF binding, DNA methylation and gene expression in a series of experiments. First they compare the binding site sequences of Rta and CTCF sites. There appears to be both data and explanations omitted here, as the comparison method is not shown and the accepted method of showing a PWM is omitted. They conclude that the two types of binding site are similar but then display promoter maps where binding occurs in different places. They then ask whether Rta binding affects CTCF binding at the CCND1 promoter, using DAPA. This in vitro assay shows no obvious decrease in binding, but it is not a quantative assay, nor is there a control for a CTCF site without an Rta site adjacent to it. The conclusion on line 154 is not justified by this data.
The authors then ask whether CTCF and Rta binding at 4 host promoters changes following induction of Rta expression in cells. All RNAs are down regulated within 12 hours in the cells. The authors show that CTCF decreases at all 4 loci in the EBV cells but only at one in the KSHV cells. This analysis uses end-point PCR which is not quantitative and the presence of bands in the IgG and Rta minus Dox is disturbing. Q-PCR would provide robust quantitative data. The decrease of RNA occurs within 12 hours but it is 48 hours before a decrease in CTCF binding is evident. I would like to see CTCF associated genes which are not also regulated by Rta in this analysis.
They then assess DNA methylation status using restriction enzyme digestion and PCR analysis. This analysis uses end-point PCR, which is not quantitative. Several lines of data have decreased signal in the minus dox lanes which is unexpected and so a concern. This is particularly apparent in the 24hour data where the large changes in DNA methylation are scored, which raises the question of whether appropriate conclusions have been drawn. There is also no explanation offered for why DNA methylation would decrease back to starting levels at 48hours.
The authors then ask whether changing CTCF expression levels impacts on lytic replication. A control that is omitted here is to measure the impact on cell viability. The data show very modest overall changes. It would be appropriate to ask whether the % of lytic cells in the population increase which may show a more profound change.
The authors then assign BKFR4 as the lytic control region (RCR) of EBV. Adequate experimental evidence for this is not provided. ChIP analysis (with the same problems discussed earlier) revealed strong CTCF interactions at the RCR of EBV at 24 hours but a clear increase in gene expression a this time point. The conclusions shown in line 250 do not reflect this. The authors ask about DNA methylation and again see little change, except on the KSHV genome – again not reflected in the conclusions. The authors propose looping to play a role but do not address this experimentally. An association between DNMT3A and B with Rta is shown.
In the discussion, the authors suggest that Rta bound to DNA causes increased DNA methylation in local areas through its association with DNMT3s and that this prevents binding of CTCF leading to a blockage of cell cycle progression. Determining the order of events is essential but not possible from the data as presented as non-quantitative methods have been used and small effects are seen for each step which could mislead.
Minor:
Copy editing – for use of “a” and “the” is required.
Scales are required for figure 1B.
EREV8 and ERKV require referencing and full descriptions.
The word significant is used on line 189, which is not justified.
In the discussion, line 286 they claim to show here that Rta is a transcriptional repressor but have already published this (ref 13/14).
Reviewer #3: General Description:
The manuscript by Chen et al addresses an interesting topic namely how Rta, a gamma herpesvirus lytic protein, modulates CTCF occupancy and DNA methylation of cellular and viral DNA regulatory regions. The work originates from a previous finding, by the same group, showing that Rta induces senescence by modulating expression of several cell cycle regulatory proteins – Rta enhanced expression of p21, p27 and cyclin E and repressed expression of CDK1, CDK4 and CDK6. In the current manuscript the authors demonstrate the Rta represses expression of four cellular genes (Myc, CCND1, Jun and CDK6) and activates lytic viral genes. The model proposed in the study suggests that Rta interacts with two DNA methyl transferases (DNMT3A and DNMT3B). Recruitment of DNMTs to cellular promoters suppresses their activation; on the contrary, methylation of viral promoters triggers their activity. Increase in cellular and viral DNA methylation was associated with dissociation of CTCF from both types of promoters.
General remarks:
While the manuscript poses an interesting model for the capacity of Rta to augment cellular and viral DNA methylation and to interfere with CTCF binding, the data presented in the manuscript does not fully support this model. In many occasions the changes were minimal. For example, in Fig 2B, expression of Rta has almost no effect on the level of CTCF associated with MYC, CCND1, or Jun promoters in KSHV cells. In Fig 5, the authors concluded that expression of EBV Rta reduced binding of CTCF to EBV and KSHV regulatory regions. Most of the changes ranged between 10 to 20%; in some instances CTCF binding increased rather than decreased. Furthermore, the effect of Rta on methylation of EBV and KSHV regulatory regions (Fig 6) was marginal except for one site (KSHV – LCR at 48h). Are these minute changes statistically significant and more importantly are they biologically meaningful?
The authors state that the experiments were repeated two or three times. A representative experiment is shown. With such small changes I would recommend that the authors provide the average of at least three experiments and to calculate the p-values.
Activation of viral DNA replication reduces DNA methylation [1]. When does viral DNA replication occur in EREV8 and ERKV cells? According to a previous report describing these cell lines, the gp350 and K8.1 proteins were detected after 48h [2].
Specific remarks:
Fig 1. The authors used biotinylated oligonucleotides that contain Rta and CTCF binding sites to demonstrate that Rta and CTCF can bind simultaneously in the absence of epigenetic DNA modifications, particularly DNA methylation. The experiment does not provide any evidence that the two proteins bind to the same DNA molecule, i.e. there could be two populations of protein-DNA complexes, an Rta-DNA complex and a CTCF-DNA complex.
The model could be tested experimentally by first methylate the DNA probes then perform the pull down assay. Also, would expression of DNMT3A and 3B block CTCF binding to the biotinylated probe? Can the authors pull down DNMTs in a manner dependent on the presence of Rta? What is the effect of a non-DNA binding Rta mutant on recruitment of DNMTs and on dissociation of CTCF from the probe?
Fig 2. What is the level of Rta and CTCF proteins in these experiments? In Fig 2B – 24h and 48h, in the absence of Dox, the Rta antibody immunoprecipitated the promoters of MYC, CCND1, Jun and CDK6, why? The authors did not exclude the possibility that less CTCF is associated with these promoters mainly due to reduction in the level of the CTCF protein.
Fig 3. The authors stated in line 246 that binding to CTCF started to decrease 24h after Dox treatment. However, the data shows that CTCF binding to EBV-LCR and EBV-RCR increased from 0.9 to 1 and from 1.1 to 1.2, respectively. At 48h, CTCF binding to EBV-oriLyt and KSHV-LCR also increased from 0.8 to 0.9. These results contradict the model provided by the authors.
Fig 7. The figure illustrates the capacity of Rta to interact with DNMT3A and DNMT3B. However, this result does not demonstrate that Rta recruits DNMT3A or DNMT3B to the studied cellular and viral promoters(as stated by the authors line 286). Association of Rta with DNMTs could be irrelevant to the proposed model. Chromatin immunoprecipitation experiments should be performed to demonstrate that expression of Rta recruits DNMTs to the studied promoters.
1. Fernandez AF, Rosales C, Lopez-Nieva P, Grana O, Ballestar E, et al. (2009) The dynamic DNA methylomes of double-stranded DNA viruses associated with human cancer. Genome research 19: 438-451.
2. Chen YJ, Tsai WH, Chen YL, Ko YC, Chou SP, et al. (2011) Epstein-Barr virus (EBV) Rta-mediated EBV and Kaposi’s sarcoma-associated herpesvirus lytic reactivations in 293 cells. PloS one 6: e17809.
#3由 sufang 在 三, 08/06/2014 – 09:00 發表。
好好認識一下methylation與demethylation
In the male pronucleus of zygote immediately after fertilization;
In mouse primordial germ cells (PGCs) between E8.5-11.5 day old embryos;[1]
Possibly in amphibia – during midblastula transition.
Electroconvulsive stimulation-induced demethylation of neurotrophic factor genes in dentate gyrus neurons in the mouse brain.
#4由 sufang 在 五, 08/01/2014 – 22:13 發表。
DNMT3A/DNMT3B 黏在nucleosome上 and SF’s comments
|
#4由 sufang 在 五, 08/01/2014 – 22:13 發表。 DNMT3A/DNMT3B 黏在nucleosome上 and SF’s comments是從這篇review p727, 左邊第二段中看到的:
Unlike DNMT1, DNMT3A and DNMT3B are firmly anchored to nucleosomes, perhaps allowing these enzymes to remain closely associated with the DNA methylation that they produce to facilitate epigenetic inheritance37. 所以，可不可能Rta bind to RRE後，block TET的功能，換句話說，附近有Rta bind的 CpG island, 5-hydroxymethylcytosine應該比control低。
–新的實驗、考慮只用24或48小時。放棄Kinetics. 因為細胞proliferation或是長滿時global methylation再control組實在變化很大。
–7/11放在LINE上面的幾點:
|