Interaction between human MCM7 and Rad17 proteins is required for replication checkpoint signaling

Genomic integrity in eukaryotic cells is continuously challenged by DNA‐damaging agents generated as by‐products of normal cellular metabolism or derived from environmental sources. To maintain the stability of their genomes, cells have evolved a battery of DNA damage response pathways, termed cell‐cycle checkpoints, which are differentially activated in response to distinct types of genotoxic stress (e.g., DNA double‐strand breaks (DSBs), ultraviolet (UV)‐induced photoproducts, or abnormal replication forks). Defects in checkpoint signaling lead to genetic instability, heightened sensitivity to DNA‐damaging agents, and increased susceptibility to cancer development in mammals (Elledge, 1996; Storchova and Pellman, 2004).

The process of DNA replication is particularly sensitive to genotoxic stress, and requires continuous surveillance in order to ensure that duplication of the genome is executed with high fidelity during S phase. Loss or malfunction of proteins that monitor DNA replication fidelity, including ATR, hRad17, and hChk1, leads to genomic instability and cell death, even in the absence of extrinsic genotoxic agents (Brown and Baltimore, 2000; Wang et al, 2003). Recently, the mechanisms whereby S‐phase cells detect and respond to damaged or aberrantly structured DNA have come under intensive scrutiny. Studies in yeast and Xenopus demonstrated that interfering with replication fork progression activates the S‐phase checkpoint, and that mutations in the replication machinery led to defective S‐phase checkpoint signaling (Lupardus et al, 2002; Stokes et al, 2002; Osborn and Elledge, 2003; Tercero et al, 2003). Although the underlying mechanisms are not well defined, mammalian cells also mount strong checkpoint responses to lesions that impede replication fork progression, such as UV‐induced cyclobutane–pyrimidine dimers (Heffernan et al, 2002). Compelling evidence implicates two phosphatidylinositol 3‐kinase‐related kinases, ATR and ATM, as proximal transducers in S‐phase checkpoint signaling pathways in mammalian cells (Abraham, 2001; Shiloh, 2003; Shechter et al, 2004). ATM is activated in response to DNA DSBs, which trigger specific alterations in histone structure (Bakkenist and Kastan, 2003), or the loading of the MRE11–Rad50–NBS1 complex at the site of strand breakage (Carson et al, 2003; Uziel et al, 2003; Lee and Paull, 2004). In contrast, ATR, together with its putative regulatory subunit, ATRIP (Cortez et al, 2002), is attracted to single‐stranded DNA (ssDNA) coated with replication protein A (RPA) (Zou and Elledge, 2003), which is a common intermediate produced during the processing of damaged DNA. Accumulating evidence suggests that checkpoint signaling through ATR is intimately linked to the process of DNA replication (Hekmat‐Nejad et al, 2000; Michael et al, 2000; Casper et al, 2002; Lupardus et al, 2002; Tercero et al, 2003; Marheineke and Hyrien, 2004; Shechter et al, 2004; Ward et al, 2004a, 2004b).

During G₁ phase of the cell cycle, replication origins in DNA are licensed by the assembly of a pre‐replication complex (pre‐RC) comprising the origin recognition complex (ORC), Cdc6, and MCM (minichromosome maintenance) complex (Waga and Stillman, 1998; Tye, 1999). Upon entry of the cell into S phase, pre‐RCs are converted into active replication forks by phosphorylation events carried out by the cyclin E–cdk2 and Dbf4–Cdc7 kinases. During fork progression through duplex DNA, unwinding and denaturation of the DNA is required to create a template for the replicative polymerases. The hexameric MCM complex, consisting of the hMCM2–7 proteins, is believed to function as a replicative helicase that mediates unwinding of DNA at origins of replication, as well as DNA in the path of active replication forks (You et al, 1999, 2002; Labib et al, 2000; Lei and Tye, 2001).

Previous studies demonstrated physical and functional interactions between ATR and the key checkpoint protein hRad17 during the cellular response to damaged DNA or replication stress (Bao et al, 2001; Stokes et al, 2002; Zou et al, 2002; Wang et al, 2003). DNA damage incurred during G₁, S, or G₂ phase of the cell cycle is sensed by hRad17, which, acting in a complex with the replication factor C (RFC) 2–5 subunits, loads three PCNA‐related proteins, hRad9, hHus1, and hRad1 (termed the ‘9‐1‐1 complex’), near the damaged site (Burtelow et al, 2000; Bermudez et al, 2003). Although the ATRIP‐ATR and 9‐1‐1 complexes are recruited independently to DNA damage sites (Kondo et al, 2001; Melo et al, 2001; Zou et al, 2002), the phosphorylation of many ATR substrates, including hChk1 and hRad17 itself, is highly dependent on the 9‐1‐1 complex (Zou et al, 2002; Jiang et al, 2003).

To further understand the checkpoint signaling functions of hRad17, we performed a yeast two‐hybrid screen aimed toward the identification of novel hRad17‐interacting proteins. In this study, we show that the MCM protein hMCM7 is a functionally important binding partner for hRad17 in human cells. The present findings indicate that, in addition to its contributions to normal DNA replication, hMCM7 is centrally involved in the activation of the ATR‐dependent S‐phase checkpoint by agents that induce DNA replication stress.

Using full‐length hRad17 as the bait, we performed a yeast two‐hybrid screen with a human fetal brain cDNA library as the prey. Of the 33 positive clones that emerged from this screen, one encoded the carboxyl‐terminal region of hMCM7, a component of the preinitiation complex assembled at origins of DNA replication (Waga and Stillman, 1998). To verify the yeast two‐hybrid results, we cotransfected HeLa cells with HA‐tagged hMCM7 and FLAG‐tagged hRad17, and observed a significant amount of FLAG‐hRad17 in the α‐HA immunoprecipitates from these cells (Figure 1A, upper panel). The converse co‐immunoprecipitation experiment revealed a modest but readily detectable amount of α‐HA immunoreactivity in FLAG‐Rad17 immunoprecipitates from the doubly transfected cells.

In subsequent experiments, we demonstrated that α‐hRad17 antibodies co‐precipitated endogenous hMCM7 from nontransfected HeLa cells (Figure 1B, left panel). Interestingly, exposure of cells to aphidicolin (Aph) or UV light increased the amount of hMCM‐7 that co‐immunoprecipitated with the endogenous hRad17. This experiment was repeated with three additional human cell lines (U2OS, A549, and human embryonic kidney 293T), and yielded similar results (not shown). Based on the prediction that ATR and its partner protein, ATRIP, reside in close proximity to the replication fork, we tested whether α‐ATRIP immunoprecipitates also contained hMCM7 and/or hRad17. Both proteins were indeed present in the α‐ATRIP immunoprecipitates, and, once again, the associations were increased by Aph or UV light exposure (Figure 1B, lower panel). Collectively, these results suggest that hMCM7 is closely associated with both ATRIP‐ATR and hRad17 in both the absence and presence of agents that induce replication stress.

Treatment of cells with inhibitors of DNA replication or UV light activates the ATR‐dependent S‐phase checkpoint (Abraham, 2001; Shiloh, 2003). To determine whether the interaction between hRad17 and hMCM7 was involved in genotoxic stress‐induced ATR activation, we silenced hRad17 or hMCM7 gene expression in HeLa cells by transfection with specific siRNAs. These reagents reduced the expression of their respective target proteins by at least 80%, relative to the control cells that received luciferase (Luc)‐specific siRNA. Both hRad17‐ and hMCM7‐depleted cells displayed significant defects in hChk1 phosphorylation at Ser‐345 after UV light exposure (Figure 2A). Furthermore, reduced expression of hMCM7 impaired UV‐induced phosphorylation of the Ser‐635 site in hRad17. These modifications of hChk1 and hRad17 are mediated by ATR in UV‐damaged cells (Liu et al, 2000; Bao et al, 2001; Bartek and Lukas, 2003). Loss of hMCM7 did not lead to global defects in checkpoint signaling, as the phosphorylation of hChk2 provoked by ionizing radiation (IR) was not perturbed in hMCM7 siRNA‐treated cells, whereas this response was clearly blunted in hRad17‐depleted cells. The latter results are consistent with the model that ATM, rather than ATR, serves as the primary upstream activator of hChk2 in response to IR‐induced DSBs (Bartek and Lukas, 2003; Shiloh, 2003).

During DNA replication, hMCM7 interacts with DNA as one member of the hexameric hMCM2–7 complex (Tye, 1999; Labib and Diffley, 2001). To determine whether the S‐phase checkpoint functions of hMCM7 was uniformly dependent on the MCM2–7 proteins, we silenced hMCM2 gene expression in HeLa cells by transfection with specific siRNA. Exposure to either siRNA reduced expression of the cognate MCM protein by at least 80%, relative to the Luc siRNA‐treated control cells. We noted that depletion of hMCM2 caused a moderate reduction of hMCM7 expression and vice versa, suggesting that loss of one MCM protein might negatively impact the stability of other members of the hMCM2–7 complex (Figure 2B). Nonetheless, reduced expression of hMCM2 had no effect on genotoxic stress‐induced hChk1 phosphorylation, whereas this response was profoundly suppressed in the hMCM7 siRNA‐treated cells (Figure 2B and C). Signal relay from ATR to hChk1 was also not perturbed in hMCM4 siRNA‐treated cells (not shown). Thus, these observations indicate that the efficiency of ATR signaling to hChk1 is particularly sensitive to reduced expression of the hMCM7 subunit.

We previously reported that UV light‐induced DNA damage triggers the appearance of ATR‐containing nuclear foci, which represent surrogate markers for ATR activation at sites of DNA damage (Tibbetts et al, 2000). In the present study, we determined whether the formation of these ATR nuclear foci was impaired in hMCM7‐ or hRad17‐depleted cells. As shown previously (Tibbetts et al, 2000), the majority of nuclei in nonirradiated cells exhibited a diffuse nuclear staining pattern for ATR, which shifted to a punctate staining pattern at 6 h post‐UV irradiation (Figure 3A). The formation of ATR nuclear foci was not impaired in hRad17‐depleted cells (Figure 3A); indeed, these cells displayed a clear increase in the number of nuclear foci‐positive cells in the absence of UV exposure (Figure 3B). The latter results suggest that knockdown of hRad17 expression triggers the accumulation of spontaneous, unrepaired DNA damage in cycling HeLa cells. The consequences of hMCM7 protein depletion were strikingly different, as the appearance of ATR nuclear foci was virtually abolished in both nonirradiated and UV‐irradiated cells (Figure 3A and B). Thus, hMCM7 expression appears to be critical for the recruitment of ATR into nuclear foci following cellular exposure to certain types of genotoxic stress.

Under normal conditions, recognition of DNA damage during S phase slows replication fork progression and inhibits the firing of late origins of replication (Lopes et al, 2001; Tercero and Diffley, 2001). We predicted that impairment of the UV‐induced DNA damage response might yield a UV‐resistant DNA synthesis (UVDS) phenotype, which is analogous to the radio‐resistant DNA synthesis (RDS) defect associated with loss of ATM and other proteins involved in the checkpoint response to DNA DSBs incurred during S phase (Painter and Young, 1980). We found that hMCM7 siRNA‐treated A549 cells displayed elevated levels of DNA synthesis at 30–120 min after UV light exposure (Figure 4A and B), consistent with the induction of UVDS. In contrast, these cells did not exhibit the RDS defect after IR exposure, indicating that hMCM7 expression was specifically required for S‐phase checkpoint activation in response to UV‐induced DNA damage. As expected, given the more global role of hRad17 in DNA damage signaling, transfection of the cells with hRad17 siRNA evoked both UVDS and RDS (Figure 4B and C).

Based on the above‐noted defects in ATR function in hMCM7‐depleted cells, we predicted that these cells might fail to delay S‐phase progression in response to UV‐induced DNA damage. However, testing of this prediction was complicated by the possibility that reduced hMCM7 might lead to an intrinsic defect in S‐phase progression. Consequently, we examined S‐ to M‐phase progression in hMCM7 siRNA‐treated HeLa cells with a nocodazole capture assay. Flow cytometric analyses of propidium iodide (PI)‐stained cells revealed that both hRad17‐ and hMCM7‐depleted cells accumulated with 4N DNA content in the presence of nocodazole, indicating that these cell populations had successfully replicated their DNA (Supplementary Figure S1A). To monitor more accurately cell‐cycle progression in hMCM7 siRNA‐treated cells, we pulse‐labeled these cells with BrdU and monitored the cell‐cycle status of the labeled cells by PI staining over the following 24 h. Relative to the Luc siRNA‐treated control cells, BrdU‐positive cells depleted of hMCM2, hMCM4 (not shown), or hMCM7 displayed little or no defect in S‐phase progression (Supplementary Figure S1B and C). In subsequent experiments, we transfected HeLa cells with siRNAs targeted against green fluorescent protein (GFP), hMCM7, or hRad17, and treated the cells with 200 J/m² UV and examined their cell‐cycle distributions at 24 h postirradiation. In GFP siRNA‐transfected cells, UV irradiation provoked an increase in the percentage of S‐phase cells after 24 h, consistent with a checkpoint‐mediated delay of cell‐cycle progression through S phase (Figure 5A). In contrast, hMCM7 siRNA‐treated cells arrested predominantly in G₂/M phase after UV light exposure. These results indicate that impaired S‐phase checkpoint activation in the hMCM7‐depleted cells favors the progression of cells bearing damaged DNA into G₂ phase, where they are captured by the G₂ DNA damage checkpoint.

Cell‐cycle checkpoint malfunctions are frequently associated with heightened sensitivity to killing by DNA‐damaging agents. Treatment of hMCM7‐ or hRad17‐depleted HeLa cells with UV light revealed that both of these cell populations were sensitized to the antiproliferative and/or cell‐killing activities of UV light, relative to that observed in the Luc siRNA‐treated control cells (Figure 5B). Finally, we examined the clonogenic survival of hMCM7‐ or hRad17‐depleted cells after transient DNA replication stress imposed by Aph. Once again, recovery of the cells after release from Aph‐induced replication arrest was impaired in the hMCM7 and hRad17 siRNA‐treated cell populations (Figure 5C).

Mapping studies of the hRad17–hMCM7 interaction with the yeast two‐hybrid system revealed that a 43‐amino‐acid peptide fragment (designated MCM7‐5) from hMCM7 was sufficient to bind detectably to full‐length hRad17 (Supplementary Figure S2). Conversely, the minimal hRad17‐derived peptide capable of binding to an amino‐terminally truncated hMCM7 bait protein (fragment designated R27 in the hMCM7 deletion series) encompassed 249 amino acids from the carboxyl terminus of hRad17 (indicated as Rad17‐5 in Supplementary Figure S2). We constructed GFP fusion proteins containing the MCM7‐5 or Rad17‐5 fragments with the idea in mind that these fusion proteins, when ectopically expressed in human cells, might interfere with the physical and functional interactions between endogenous hMCM7 and hRad17.

In preliminary studies, we found that transient expression of a GFP‐tagged fusion protein containing the MCM7‐5 fragment completely blocked S‐phase entry and progression in otherwise unperturbed HeLa cells (results not shown). In contrast, HeLa cells transfected with the GFP‐Rad17‐5 construct continued to cycle normally, a finding that prompted us to focus on this fusion protein as a potential inhibitor of signal relay between hMCM7 and hRad17 in UV light‐ or Aph‐stressed cells. Co‐immunoprecipitation experiments revealed that endogenous hMCM7 associated with the ectopically expressed GFP‐Rad17‐5 fusion protein in transfected HeLa cells (Figure 6A). In contrast, α‐hMCM7 immunoprecipitates from GFP‐transfected cells contained no detectable GFP, in spite of the fact that intact GFP was expressed at much higher levels than the GFP‐Rad17‐5 fusion protein. Furthermore, GFP‐Rad17‐5 expression reduced the level of endogenous hRad17 present in α‐hMCM7 immunoprecipitates (Figure 6A), consistent with the idea that the Rad17‐5 fragment competitively interferes with the hMCM7–hRad17 interaction in intact cells. In subsequent studies, we found that expression of GFP‐Rad17‐5 strongly suppressed the phosphorylation of hChk1 in UV‐ and Aph‐treated cells (Figure 6B). Furthermore, like their hMCM7‐depleted counterparts, the GFP‐Rad17‐5‐expressing HeLa cells displayed the UVDS phenotype (Figure 6C).

To rule out the possibility that the GFP‐Rad17‐5 construct was simply acting as a dominant suppressor of hRad17 function, we determined whether the GFP‐Rad17‐5‐transfected cells retained the ability to phosphorylate hChk2 after IR exposure. This ATM‐dependent response also hinges on the function of hRad17 (see Figure 2A). Indeed, the GFP‐Rad17‐5‐transfected cells displayed a strong increase in basal hChk2 phosphorylation, which was modestly increased by IR or UV exposure, while hChk1 activation was virtually abolished (Figure 6D). The basal increase in hChk2 phosphorylation in GFP‐Rad17‐5‐expressing cells is consistent with the observation that these cells suffer spontaneous DNA damage (see Figure 7A), possibly due to disruption of the replication checkpoint. Regardless, we conclude from these results that the GFP‐Rad17‐5 construct is a selective, rather than global, inhibitor of hRad17 functions in radiation‐damaged cells.

ATR‐deficient cells rapidly accumulate damaged DNA in the absence of extrinsic genotoxic stress, likely due to a breakdown in the fidelity of DNA replication (Brown and Baltimore, 2000; de Klein et al, 2000). Given the evidence that GFP‐Rad17‐5 expression disrupts signal transmission through ATR to hChk1, we predicted that cells transfected with this construct might also display elevated levels of spontaneous DNA damage. Consequently, we stained the cells with a phosphospecific antibody that recognizes an ATM/ATR‐dependent phosphorylation site (Ser‐139) in histone H2AX. The phosphorylated form of H2AX (γH2AX) accumulates rapidly at sites of DNA damage in cells treated with IR, UV, or DNA replication inhibitors (Rogakou et al, 1998, 1999; Ward and Chen, 2001). Expression of GFP‐Rad17‐5 in HeLa cells provoked a greater than seven‐fold increase in γH2AX‐positive nuclei, compared with the GFP‐transfected control cells (Figure 7A and B). We repeated these experiments with hMCM7 siRNA‐treated cells, and observed that loss of hMCM7 also caused a clear increase in γ‐H2AX‐positive nuclei (Figure 7C). In contrast, no increase in γH2AX staining was seen in the hRad17‐depleted cells. The latter results suggest that loss of hRad17 does not lead to spontaneous DNA damage, although we have not ruled out the possibility that hRad17 expression is required for maximal phosphorylation of histone H2AX in UV‐damaged cells.

The hexameric MCM complex plays central roles in DNA replication initiation and elongation (Labib et al, 2000). In this study, we have defined an additional role for the hMCM7 subunit of the MCM complex in the activation of the S‐phase DNA damage checkpoint by UV light and Aph. Our results argue that hMCM7 participates in functionally important interactions with at least two key elements of the S‐phase checkpoint machinery—hRad17 and ATRIP‐ATR. Manipulations that interfere with hMCM7 function strongly inhibit the formation of ATR nuclear foci, as well as ATR‐dependent hChk1 activation in UV‐ or Aph‐stressed cells. Furthermore, disruption of hMCM7 function triggered a DNA damage response in the absence of extrinsic genotoxic stress, consistent with previous evidence that the ATR–Chk1 pathway plays a continuous role in genome maintenance during normal DNA replication (Casper et al, 2002; Cha and Kleckner, 2002; Marheineke and Hyrien, 2004). Taken together, these results strengthen the notion that components of the DNA replication fork function as both DNA damage sensors and effectors of checkpoint signaling in S‐phase cells (Tercero et al, 2003).

Previous studies identified at least three binding partners for hRad17: Ku‐interacting protein (Chang et al, 1999), hRad1 (Rauen et al, 2000), and protein phosphatase 5 (Ali et al, 2004). With the exception of the hRad1 interaction, the functional significance of these associations with hRad17 remains unclear. Our results indicate that binding of hRad17 to hMCM7 is critical for S‐phase checkpoint signaling in response to replicative stress. Studies in fission yeast have shown that Rad17 is constitutively bound to chromatin, and that the level of chromatin‐bound Rad17 is increased during replication stress (Kai et al, 2001). More recent results indicate that the hRad17–RFC2–5 complex binds most avidly to DNA duplexes containing single‐stranded gaps coated with RPA (Zou et al, 2003). These structures are abundant in the vicinity of active replication forks (Waga and Stillman, 1998). In light of the current findings, we propose that RPA recruits the hRad17–RFC2–5 complex to sites of DNA replication, where hRad17 encounters and binds to hMCM7 in the vicinity of the replication fork. We further speculate that replication fork stalling induced by agents such as UV or Aph triggers a local increase in the level of RPA‐bound ssDNA, an event that provokes the recruitment of sufficient hRad17–RFC2–5, together with ATRIP‐ATR, to elicit a full‐blown S‐phase damage checkpoint response. This model is consistent with the notion that DNA polymerase stalling induced by Aph causes DNA helicase activity to become uncoupled from DNA polymerases in the Xenopus egg extracts, an event that would lead to abnormally long stretches of ssDNA (Walter and Newport, 2000).

Although the present study has focused exclusively on hMCM7, this protein is generally believed to function in the context of the hMCM2–7 complex during DNA replication. Remarkably, however, depletion of either hMCM4 or hMCM2 failed to disrupt the activation of hChk1 by UV or Aph, indicating that the checkpoint signaling function of hMCM7 may not be obligatorily linked to its role in DNA replication. A recent report provides strong support for the idea that MCM proteins are functionally intertwined with the S‐phase checkpoint machinery (Cortez et al, 2004). This study demonstrated that two MCM proteins, hMCM2 and hMCM3, are phosphorylated by ATR and ATM, respectively, in cells exposed to genotoxic stress. These findings complement an earlier study, which showed that hMCM4 also undergoes rapid phosphorylation in cells exposed to UV light or DNA replication inhibitors (Ishimi et al, 2003). Although the functional significance of these phosphorylation events is unknown, it seems clear that, as is the case for ATR and hRad17, the interplay between ATR and the MCM proteins will prove to be bidirectional. Cortez and Elledge also found that hMCM7 was a direct binding partner for the ATR‐associated protein ATRIP, and showed that partial depletion of hMCM7 interfered with UV‐induced hChk1 activation in U2OS cells. Taken together with our findings, these results provide compelling evidence to support the hypothesis that hMCM7 plays a specialized role in the relay of DNA damage signals from the replication fork to proximal elements of the S‐phase checkpoint pathway.

Costanzo et al (2003) have defined a distinct mechanism of S‐phase checkpoint initiation in Xenopus egg extracts (Costanzo et al, 2003). These investigators demonstrated that the Dbf4–Cdc7 kinase was a critical target of the ATR‐dependent S‐phase checkpoint induced by the topoisomerase II inhibitor etoposide. Inhibition of Dbf4–Cdc7 by this checkpoint mechanism interferes with the phosphorylation of the MCM complex by this protein kinase, thereby suppressing conversion of pre‐RCs to active replication forks. The inhibition of replication origin firing via this checkpoint mechanism complements the pathway outlined in the present study, which relies on hMCM7 function to slow the progress of and/or stabilize pre‐existing replication forks that encounter fork‐stalling lesions. The common element that leads to the recruitment of ATR in both of these S‐phase checkpoint pathways is the formation of RPA‐coated ssDNA (Costanzo et al, 2003; Zou and Elledge, 2003; Zou et al, 2003).

Our results indicate that hRad17 and hMCM7 interact in a pathway that includes ATR and mediates S‐phase checkpoint activation in response to genotoxic stress. However, it is clear that binding to hMCM7 is not the only mechanism by which hRad17 receives and transmits DNA damage signals in human cells. For example, we observed that UV‐ or IR‐induced hChk2 phosphorylation was not impaired in hMCM7‐deficient cells, whereas this event was strongly suppressed in cells depleted of hRad17. This outcome is consistent with previous evidence that hChk2 phosphorylation and activation are more tightly coupled to ATM than to ATR (Matsuoka et al, 2000; Bartek and Lukas, 2003). The checkpoint functions of hRad17 are not confined to S phase (Elledge, 1996; Caspari and Carr, 1999), suggesting that, at other points in the cell cycle, DNA damage signals are relayed to hRad17 through RPA and hRad17‐interacting proteins other than hMCM7.

A striking outcome of our studies was that decreased hMCM7 expression abolished the formation of ATR‐containing nuclear foci in UV‐treated cells. In contrast, depletion of hRad17 actually promoted the formation of such foci in the absence of genotoxic agents. These observations reinforce the idea that the ATRIP‐ATR and 9‐1‐1 complexes are recruited independently to damaged chromatin (Zou et al, 2002). While RPA may be required for efficient recruitment of ATRIP‐ATR to ssDNA (Zou and Elledge, 2003), the present findings, together with those of Cortez et al (2004), suggest that hMCM7 is centrally involved in the assembly and/or stabilization of ATRIP‐ATR‐containing foci in S‐phase cells. In line with this model, we found that α‐ATRIP immunoprecipitates from cycling cells contain both hMCM7 and hRad17, and that the amounts of co‐precipitating hMCM7 and hRad17 are increased after cellular exposure to UV or Aph (Figure 1B).

In summary, the present findings highlight the hMCM7 protein as a critical node for checkpoint signal transmission from stalled replication forks. A major unresolved issue is whether the checkpoint signaling functions of hMCM7 are strictly tied to its role in DNA replication as part of the MCM complex, or whether this facet of hMCM7 function is independent of its contributions to normal DNA replication. Regardless, the present studies add significant strength to the idea that the replication fork functions as both a sensor of DNA damage and an effector of checkpoint signaling in cells that incur DNA damage during S phase (Cimprich, 2003; Katou et al, 2003; Tercero et al, 2003; Oehlmann et al, 2004).

We thank the members of the Abraham laboratory for helpful discussions, and members of the Wei Jiang laboratory for advice and reagents. This research was supported by grants from the National Institutes of Health (CA097950), Department of Defense (DAMD17‐02‐1‐0730), and Johnson & Johnson.