Vanderbilt Institute of Chemical Biology



Discovery at the VICB







Exploring the Dynamics of Replication Fork Collapse


By: Carol A. Rouzer, VICB Communications
Published:  October 1, 2015



A comprehensive proteomics analysis of proteins at normal and stressed replication forks illuminates the role of ATR in preventing fork collapse.


DNA replication is an extremely complex process during which billions of base pairs must be accurately duplicated in a timely fashion. Primarily responsible for DNA synthesis is the replisome, a large assembly of polymerases, helicases, ligases, and nucleases, that quickly forms as a new origin of replication is converted into a replication fork (Figure 1). Equally important are proteins responsible for DNA packaging and chromatin formation, which also assemble at the fork. Considering the large number of proteins involved and the requirement for high fidelity and tight regulation, it is not surprising that cells are readily susceptible to replication stress from a wide variety of sources, including the presence of DNA damage, sequences that are difficult to replicate, or the activation of oncogenes. An important player in both normal replication and the response to stress is the ATR (ataxia telangiectasia mutated- and Rad-3-related) kinase. Known best for its role in the DNA damage response, ATR is activated in the presence of single-stranded DNA (ssDNA) at damage sites, leading to the phosphorylation of Chk1 (checkpoint kinase 1), which in turn, stops the cell cycle, allowing time for DNA repair. In addition, ATR regulates the firing of new origins of replication, stabilizes replication forks, and helps to restart damaged forks, although the exact mechanisms by which ATR carries out these functions remain unclear. Now, Vanderbilt Institute of Chemical Biology member David Cortez and his laboratory employ a combination of targeted click chemistry and proteomics to explore protein dynamics at the replication fork and the role of ATR in preventing fork collapse in response to replication stress [H. Dungrawala, et al., (2015) Mol. Cell, published online September 17, DOI: 10.1016/j.molcel.2015.07.030].


Figure 1. DNA replication is a complex process. The two strands of the double helix are unwound by the action of topoisomerase and helicase to form the replication fork. The single strands are stabilized by binding proteins, and the two strands are replicated in reverse directions, beginning with an RNA primer. This diagram shows the general structure of the replication fork as the process progresses down the DNA strand, but it only shows a few of the large number of proteins involved. Figure kindly provided by David Cortez (copyright 2015).


Key to their investigation of the mechanisms of DNA replication and replication fork structure was iPOND (isolation of proteins on nascent DNA), a technique previously developed in the Cortez lab. The iPOND approach uses 5-ethynyl-2′-deoxyuridine (EdU) to label newly synthesized DNA. Following the labeling period, treatment with formaldehyde cross-links the DNA to any associated proteins, and the DNA-protein complexes are then isolated. The researchers use click chemistry to attach a biotin tag to the alkyne functional group of EdU that has been incorporated into the DNA. This enables them to selectively isolate the newly synthesized DNA using streptavidin affinity purification. Reversal of the cross-links then frees the DNA-associated proteins for characterization (Figure 2). The timing of the EdU labeling determines the proteins captured by iPOND. A brief pulse followed by immediate isolation will lead to the detection of only the DNA synthesizing machinery. Longer pulses will lead to the capture of chromatin assembly proteins (Figure 2B). A brief pulse followed by a thymidine chase will result in the selective capture of proteins associated with mature chromatin (Figure 2C).


Figure 2. (A) Diagram illustrating the iPOND method. 1. Newly synthesized DNA incorporates EdU at the replication fork. 2. Formaldehyde is used to cross-link the DNA with any proteins associated at the fork region. 3. The cells are permeabilized, and reaction with the biotin azide reagent under click chemistry conditions attaches biotin to EdU in the DNA. 4. The cells are lysed and sonicated to release the DNA and break it into fragments. 5. DNA fragments are isolated by attachment to streptavidin-coated beads. 6. Following elution from the beads, the formaldehyde cross-links are reversed by high temperature incubation, releasing the proteins for analysis by western blot as shown in the figure or by mass spectrometry as carried out in the current study. (B) The region of DNA labeled with EdU (gold) starts at the replication fork and grows longer as the fork progresses along the DNA helix. With prolonged EdU incubation, the regions labeled initially undergo post-replication processes, such as chromatin assembly, and the associated proteins change as the process progresses. Newly synthesized DNA is associated with replication proteins such as PCNA and CAF-1. Over time, histones bind to the DNA in the process of chromatin assembly. (C) Incubation with a brief pulse of EdU followed by a chase of thymidine leads to EdU incorporation in a small segment of DNA that becomes more distant from the moving replication fork as the chase time increases. This allows the selective investigation of proteins at different stages of replication and post-replication processing. Figure kindly provided by Bianca Sirbu of the Cortez lab.



In their prior applications of iPOND, the Cortez lab had primarily relied on immunoblotting to identify the captured proteins. However, realizing that this approach limited them to studying proteins for which antibodies were available, they turned for the current study to mass spectrometry to enable a much more comprehensive identification of the replication fork-associated proteins. In addition, they applied SILAC (stable isotope labeling of amino acids in cell culture) to provide relative quantification of proteins between treated and control samples (Figure 3).





Figure 3. SILAC method for protein quantitation. (A) Cells are grown in the presence of lysine and arginine uniformly labeled with 12C and 14N or 13C and 15N to produce “light” and “heavy” cells, respectively. The cells are grown with the labeled amino acids until all cellular proteins contain those amino acids. In this example, the heavy cells are then exposed to the desired treatment while the light cells serve as controls. The cells are lysed and protein extracts are prepared and quantified. The heavy and light extracts are then combined in equal quantities of protein and subjected to proteomics analysis by mass spectrometry. (B) The mass spectrometer can detect the difference between the masses of lysine- or arginine-containing peptides derived from each heavy- and light-labeled protein. In this example, the peptide contains one lysine residue and no arginine residues. Thus, the mass of the heavy peptide is 8 Da greater than that of the light peptide. If the signals from the heavy and light peptides are the same (center), then the treatment had no effect on the levels of this protein. If the signal from the light protein is larger (left), then the treatment reduced the levels of this protein, while if the signal of the heavy protein is larger (right), the treatment increased the levels of the protein, relative to those in the control.



In their initial iPOND-SILAC experiments using HEK293 cells, the investigators labeled the heavy cells with EdU while leaving the light cells unlabeled. This experiment allowed them to distinguish proteins that were specifically isolated as a result of the presence of EdU in adjacent DNA versus those captured nonspecifically. The results identified 864 specifically captured proteins. In a second experiment, the investigators treated the light cells with EdU for 10 min and the heavy cells for 10 min followed by a 1 h chase. This experiment was designed to identify proteins that were truly associated with the replication fork (light cells) versus those present in bulk chromatin (heavy cells). The results identified 218 replication fork-associated proteins that included, as expected, helicases, polymerases, nucleases, ligases, chromatin and histone remodeling proteins, and replication stress proteins.


Having established the effectiveness of iPOND-SILAC to correctly identify proteins associated with the replication fork, the investigators next explored the effects of replication stress on that proteome. They induced stress in the cells by treatment with hydroxyurea, which blocks replication by inhibiting ribonucleotide reductase, thereby depleting the available pools of deoxyribonucleotide triphosphate precursors (Figure 4). After labeling cells for 10 min with EdU, the researchers added hydroxyurea for periods of 5 min through 24 h in the continued presence of EdU. The results identified 68 proteins that were increased in abundance at stalled replication forks over at least two time points. Most of these were replication stress response proteins, including BRCA1-BARD1, SMARCAL1, WRN, BLM, and ATR.




Figure 4. Following an EdU pulse label, hydroxyurea (yellow) is added to the cells to stall replication. Over time, DDR response proteins (red, orange, green) assemble at the replication fork to repair damage that results from prolonged replication stress. Figure kindly provided by Bianca Sirbu of the Cortez lab.



The investigators became interested in the mechanisms that specifically draw proteins to stalled replication forks. They noted that ATR and ATR-related proteins increase rapidly upon fork stalling. Thus proteins, such as MDC1, 53BP1, and RNF169 that also appear quickly may do so following the phosphorylation of histone H2AX by ATR. Other proteins are recruited to stalled forks following increases in the levels of ubiquitinated proteins in the region. The investigators noted that there was no increase in ubiquitinating enzymes at the stalled forks, but a decrease in the deubiqutinating protein, USP1-WDR48 could explain the accumulation of ubiquitin that they observed during replication stress. Similarly, some proteins are recruited to the stalled fork through an interaction with poly-ADP-ribosylated (PARylated) proteins, which also increased in the region of the stalled fork. Again, no increase in PARP1, the enzyme responsible for PARylation was observed, but a decrease in PARG, which removes the poly-ADP-ribosyl groups could explain the observed increase in PARylated proteins.


The researchers used the statistical approach of unsupervised hierarchical clustering to identify proteins that exhibited similar patterns of abundance changes in response to hydroxyurea treatment. Proteins known to be part of the same complex behaved similarly in this regard. This led the investigators to hypothesize that similar patterns of abundance changes might be used to identify new protein complex associations. This hypothesis was supported in the case of ZNF644, a protein that was highly enriched at normal replication forks and exhibited an abundance pattern similar to that of the G9a/GLP methyltransferase complex. Indeed, ZNF644 co-immunoprecipitated with G9a, consistent with an association between the two proteins. In addition, the sequence of ZNF644 shares substantial identities with that of WIZ, a protein known to interact with the G9a/GLP complex. RNAi-mediated knockdown of ZNF644 in HEK293 cells resulted in decreased cell proliferation and increased sensitivity to replication stress, confirming an important role for the protein in DNA replication.


Using hydroxyurea treatment and iPOND-SILAC analysis, the Cortez lab was now prepared to investigate the role of ATR in preventing replication fork collapse. Prior work had suggested that ATR prevents fork collapse by stabilizing the replisome. The investigators reasoned that if this were true, then inhibition of ATR should lead to replisome dissociation from the fork under stress conditions. To test this hypothesis, they treated cells with hydroxyurea in the presence or absence of VE821, an ATR inhibitor. The iPOND-SILAC results showed that, consistent with its general role in preventing replication fork collapse, blocking ATR activity led to the rapid accumulation of proteins associated with replication stress and double strand break responses, such as RPA, SMARCAL1, BLM, WRN, FNACJ, ATM, MRN, and DNA-PK. To their surprise, however, inhibition of ATR had very little effect on the level of replisome proteins at the replication fork. In fact, they observed a small increase in replisome proteins associated with the replication fork in VE821-treated cells. They obtained similar results when replication stress was induced by the DNA polymerase-α inhibitor aphidicolin in the presence and absence of VE821. Addition of an inhibitor of CDC7 (cell division cycle 7-related protein kinase) to prevent any initiation of new replication forks blocked the small increase in replisome protein association that occurred in cells treated with the ATR inhibitor. The investigators concluded that ATR has no effect on replisome stability and that the slight increase in replisome association with the replication fork that they observed upon ATR inhibition was due to a low level of formation of new replication forks.


As noted above, a primary function of ATR is to phosphorylate Chk1, leading to cell cycle arrest and a cessation of DNA synthesis. Thus, one mechanism by which ATR prevents fork collapse during replication stress is by preventing the firing of new origins of replication. This is particularly important at times of replication stress, because fork collapse leads to the exposure of ssDNA, which must be protected through the binding of replication protein A (RPA). RPA also plays an important role in normal DNA replication, so by preventing the firing of new origins of replication, RPA can be conserved for use in stabilizing and repairing stalled forks. Consistent with these considerations, prior reports had suggested that prevention of new origin firing could preserve RPA and decrease fork collapse in stressed cells treated with an ATR inhibitor. However, when the Cortez lab investigators tested this hypothesis using cells treated with hydroxyurea in the presence of VE821 to inhibit ATR with and without an inhibitor of CDC7 to prevent new origin firing, their iPOND-SILAC results indicated that CDC7 inhibition had no effect on the amount of RPA present at stalled replication forks. They noted, however, that some double-strand break sensing proteins and DNA repair proteins were affected by the presence of the CDC7 inhibitor while others were not. This led them to propose that two kinds of collapsed replication forks occur when ATR is inhibited. The first kind, generated at new origins of replication, recruits proteins that repair double-strand breaks at collapsed forks using nonhomologous end joining (NHEJ). In contrast, when preexisting forks collapse, they recruit proteins that repair double-strand breaks by homologous recombination (HR). The investigators note that preferential use of HR at the latter forks is advantageous to the cells, as HR leads to a more accurate repair than NHEJ.


When the Cortez lab investigators treated cells with hydroxyurea, leading to the formation of collapsed replication forks, they found that 80% of the forks could recover and resume DNA synthesis after they removed the hydroxyurea. However, when the cells were treated with hydroxyurea in the presence of the ATR inhibitor, only 20% of the forks recovered. If new origin firing was prevented, replication fork collapse in the presence of ATR inhibition was delayed but not prevented. Overexpression of RPA also delayed fork collapse in cells treated with hydroxyurea and an ATR inhibitor. In fact, RPA overexpression completely blocked hydroxyurea-induced fork collapse in cells treated with inhibitors of both ATR and CDC7. These findings confirm that ATR functions to stabilize stalled replication forks. However, the increased fork collapse observed upon ATR inhibition is not explained by failure to block new origin firing, because fork collapse still occurs, even when firing is inhibited by other mechanisms (CDC7 inhibition). The authors propose that ATR may, instead, stabilize replication forks through recruiting and regulating other enzymes involved in the replication stress response.


The results highlight the value of iPOND-SILAC in the investigation of replication under both normal and stressed conditions. This approach has provided important new insights into the role of ATR in replication fork stabilization, distinguishing two distinct forms of replication fork collapse in the absence of ATR, and ruling out stabilization of the replisome as one of ATR’s mechanisms of action. iPOND-SILAC also proved valuable in identifying new protein complex interactions in the case of ZNF644, and can readily be applied to the exploration of additional protein complex interactions in the future.



View Mol. Cell article: The Replication Checkpoint Prevents Two Types of Fork Collapse without Regulating Replisome Stability







The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center