Our data indicate that DDB1 is required for maintaining genome stability in human cells. Silencing of DDB1 expression results in an accumulation of DNA damage and activation of cell cycle checkpoints. DDB1 protects the integrity of the genome as part of an E3 ubiquitin ligase complex containing Cul4A. One target of this complex that must be degraded to preserve genome integrity is the replication licensing factor Cdt1. Multiple levels of Cdt1 regulation exist in human cells. These mechanisms possess unique and necessary roles in the regulation of Cdt1, to properly restrain this protein and prevent the adverse cellular consequences associated with its misregulation.
A role for DDB1 in maintaining genome stability has previously been noted in other organisms. Shimanouchi and colleagues found that depletion of DDB1 in Drosophila melanogaster
promoted the loss of heterozygosity in somatic cells (53
). Holmberg et al. also found that deletion of DDB1 in S. pombe
enhances the mutation rate more than 20-fold (25
). These data correlate well with our data from human cells. Depletion or deletion of DDB1 results in genome instability, as assessed by the generation of DNA double-strand breaks in human cells (this study) or an enhanced mutation rate in yeast, and in neither organism is this instability the result of defective NER. In S. pombe
, the genetic instability associated with deletion of DDB1 could be partially suppressed by removing Spd1, an inhibitor of RNR (25
). DDB1 and Pcu4 (the Cul4 homologue in yeast) promote the ubiquitin-mediated degradation of Spd1 to allow association of the RNR subunits and production of deoxynucleoside triphosphates for DNA synthesis and repair (8
). A human homologue of Spd1 has not been identified; therefore, it is unclear whether this function of DDB1 is conserved in higher eukaryotes. However, removal of Spd1 could not completely alleviate the genetic instability created by the loss of DDB1 in S. pombe
. This suggests that DDB1 is required for other aspects of mutation avoidance in addition to regulating DNA replication through the degradation of Spd1. Likewise, we suggest that human DDB1 mediates genome stability, in part by controlling DNA replication through the degradation of Cdt1. However, the inability of Cdt1 depletion to completely alleviate the phenotype observed following DDB1 depletion indicates that additional functions of DDB1 are critical for its role in the maintenance of genome integrity.
The DDB1-Cul4A ubiquitin ligase complex regulates DNA replication in multicellular eukaryotes by mediating the degradation of Cdt1. Exposure to exogenous DNA-damaging agents induces the destruction of Cdt1 specifically by DDB1-Cul4A (24
). Our data indicate that DDB1 also has an important role in the regulation of Cdt1 in the absence of exogenous DNA damage. In agreement with our proposal, recent reports have implicated DDB1-Cul4A in the replication-dependent degradation of Cdt1 and identified PCNA as a critical mediator of the DDB1-Cul4A-dependent Cdt1 ubiquitination (2
). The destruction of Cdt1 by the DDB1-Cul4A ubiquitin ligase complex is one of three mechanisms known in human cells to regulate Cdt1. Additionally, the replication-dependent degradation of this protein can be accomplished by an SCFSkp2
ubiquitin ligase complex, which is targeted to Cdt1 by a CDK-mediated phosphorylation event (34
). Cdt1 is also functionally inhibited by the binding of geminin (59
). These multiple levels of Cdt1 regulation suggest that the proper restraint of this protein activity is crucial.
Disruption of Cdt1 regulation is detrimental to genome stability and cell viability (25
). Highly elevated levels of Cdt1 expression lead to significant amounts of rereplication in human cells, Drosophila
, and Xenopus laevis
). Rereplication induces activation of DNA damage response pathways in humans, Xenopus
, and yeast (20
). The phenotype observed after DDB1 depletion is consistent with the phenotype observed after rereplication. We found a significant increase in DNA damage and activation of both ATR- and ATM-mediated damage response pathways (Fig. and ). Notably, we see this enhancement of genome instability with only moderate increases in Cdt1 levels (Fig. ), in contrast to the significantly greater levels of Cdt1 overexpression that were used in previous studies. Furthermore, we observe direct evidence that rereplication is occurring in DDB1-depleted cells with the incorporation of BrdU by cells with greater than 4n DNA content (Fig. ). The rereplication observed after DDB1 depletion is not extensive, perhaps due in part to the restraints imposed by the checkpoint response (33
). The elevated levels of Cdt1 protein, rereplication, and DNA damage suggest that misregulation of Cdt1 is contributing to the phenotype observed after DDB1 depletion. In support of this hypothesis, we found that there was indeed deregulation of Cdt1, as the degradation of Cdt1 proteins was significantly delayed following depletion of DDB1 (Fig. ). Additionally, reducing Cdt1 protein levels in DDB1-depleted cells could prevent rereplication and eliminate approximately half of the DNA damage and checkpoint activation (Fig. ). Importantly, the level of Cdt1 reduction achieved in these experiments did not interfere with normal DNA replication. Therefore, the reduction in DNA damage by codepleting Cdt1 with DDB1 is unlikely to be an indirect consequence of slowing the cell cycle. Our results clearly demonstrate that depletion of DDB1 results in the misregulation of Cdt1, which stabilizes and elevates the cellular levels of this protein. The consequence of this misregulation is rereplication, which contributes to the DNA damage and checkpoint activation observed after DDB1 depletion.
Our data indicate that rereplication occurs in DDB1-depleted cells despite the presence of other mechanisms that operate to suppress the refiring of replication origins. Previous data from human systems have suggested that the DDB1-Cul4 and SCFSkp2
mechanisms of Cdt1 destruction are redundant in the absence of exogenous DNA damage (45
). However, the disruption of a single mode of Cdt1 regulation in other organisms was shown to be sufficient to have adverse effects. In Xenopus
, disruption of the DDB1-Cul4 degradation pathway stabilizes Cdt1 and induces significant levels of rereplication (2
). Depletion of Cul4 from Caenorhabditis elegans
also stabilizes Cdt1 and results in massive rereplication, with cells exhibiting up to 100n DNA content (73
). The stabilization of Cdt1 and the presence of rereplication after DDB1 depletion argue against redundant roles for DDB1-Cul4A and SCFSkp2
in the destruction of Cdt1 in human cells. Additionally, this suggestion is supported by our observation that expression of a Cdt1 mutant insensitive to DDB1-Cul4A degradation results in greater DNA damage than expression of either wild-type Cdt1 or a mutant that is insensitive to the SCFSkp2
destruction pathway (Fig. ). We propose that the DDB1-dependent degradation of Cdt1 is particularly important because the loss of DDB1 creates a situation in which one other mechanism of Cdt1 regulation is also inactivated. Depletion of DDB1 generates DNA damage that activates cell cycle checkpoints, which in turn function to inactivate CDK complexes (50
). Since the SCFSkp2
-mediated destruction of Cdt1 requires a CDK-mediated phosphorylation event, this pathway is inhibited by the presence of active checkpoints. Therefore, disruption of DDB1 eliminates both ubiquitin-dependent mechanisms of Cdt1 regulation, which will likely result in further rereplication, greater DNA damage, and amplification of genome instability (Fig. ). It should be noted that cells generate DNA damage intrinsically as a consequence of respiration and DNA metabolism. Replication forks encounter DNA lesions and experience difficulty in replicating through specific regions of the genome during every round of DNA synthesis. The act of growing cells in culture can also increase cellular stresses. Thus, the distinction between DNA damage-dependent Cdt1 degradation and replication-dependent degradation is largely one of degree.
FIG. 9. Model highlighting the importance of DDB1 in Cdt1 regulation. (A) Geminin, DDB1-Cul4A, and SCFSkp2 cooperate to restrain the activity and levels of Cdt1, ensuring that origins of replication fire only once per S phase. (B) siRNA depletion of DDB1 disrupts (more ...)
DDB1-dependent regulation of Cdt1 is not sufficient to explain all of the genome instability that arises from DDB1 depletion. Codepletion of Cdt1 with DDB1 eliminates rereplication; however, it does not completely prevent DNA damage or checkpoint activation. Additional genome maintenance functions of DDB1 likely have cell cycle dependency, since the loss of DDB1 from G0-arrested cells does not cause DNA damage (Fig. ). Identifying other substrates of DDB1-Cul4A whose degradation is important in preventing genetic instability will be important. Another question that remains unanswered is the mechanism by which rereplication activates checkpoint pathways. Our analyses suggest that rereplication causes DNA double-strand breaks in a manner that is distinct from those arising at chromosome fragile sites during replication stress. It is unclear whether these breaks are at random locations or whether they may cluster near specific genomic regions that are prone to rereplication. The breaks do not appear to cluster near centromeres. This suggests that attachment of the mitotic spindle to a rereplicated centromere on a single chromatid, and the subsequent breakage of the chromosome during anaphase, is not a major mechanism contributing to these breaks. One possibility is that rereplication causes double-strand breaks as the second fork originating from a refired origin encounters the Okazaki fragments generated by the original replication fork. A second possibility is that the forks initiated at a refired origin are defective, stall, and eventually collapse, thus generating DNA damage.
Recent studies have highlighted a role for replication stress as an early event in the initiation of cancer (5
). Aberrant DNA replication in precancerous lesions produces DNA damage and activates checkpoint pathways. Inactivation of these damage response pathways is associated with tumor progression due to genetic instability. Disruption of DDB1-dependent functions also has the ability to create replication stress and genetic instability, perhaps providing an avenue through which cancer progression can be facilitated. Changes in the activity of the DDB1-Cul4A ubiquitin ligase complex, or in mechanisms regulating replication origin firing, may therefore play important roles in the process of tumorigenesis.