Maintenance of constant genome ploidy is critical for eukaryotic organisms. If unchecked, disruption of the mechanisms that tightly couple DNA replication with the cell cycle may result in re-replication, aneuploidy and genomic instability. We have found that perturbation of Dup activity by geminin depletion results in the preferential re-replication of heterochromatic sequences in the Drosophila genome. Re-replication was limited to pericentromeric heterochromatic sequences which are marked by HP1 and H3K9 methylation. Euchromatin, including gene-poor late replicating sequences and polycomb repressed sequences, was resistant to re-replication. In the absence of geminin, a minimal complement of MCMs assembled on the chromatin was sufficent for re-replication. These findings suggest that the 2–3 fold increase in ploidy we observed was regulated by the specific re-assembly and activation of the heterochromatic pre-RCs.
The preferential re-replication of heterochromatic sequences in the absence of licensing controls was particularly striking given the established view that repressive chromatin environments are inhibitory to efficient origin activation 
. Classic experiments have clearly demonstrated that the heterochromatin in Drosophila
and other ogranisms is the last region of the genome to be duplicated during S phase 
. The complex nature of the heterochromatic sequences has hampered detailed analysis of the replication program in this part of the genome and it remains possible that the heterochromatin may be populated by a very limited number of ultra-efficient origins of replication. These may be required to ensure that the heterochromatin is duplicated in a timely manner at the end of S phase and that in the absence of licensing controls these origins are preferentially activated. Our genomic data did not identify any robust origins of replication in the heterochromatin that were consistently used across the cell population. Similarly, we found that the bulk of heterochromatin was re-replicated to similar ploidy levels suggesting that origin re-activation in the absence of geminin is a stochastic process.
Analysis of total DNA ploidy by FACS revealed that the majority of cells exhibited a DNA content greater than 8N following geminin depletion, suggesting geminin depleted cells had increased their total DNA content by at least 2-fold over that of G2 cells. The heterochromatin constitutes a minimum of 30% of the Drosophila
. Assuming that re-replication is specific to the heterochromatin, a 4.5 fold increase in heterochromatic DNA content would be sufficient to account for the increase in DNA ploidy we observe. However, we consistently observed, by multiple methods, only a 2–2.5 fold increase in heterochromatic DNA content (, Figure S2
). We speculated that the highly repetitive non-unique heterochromatic sequences might be preferentially re-replicated to higher ploidy levels. We tested this hypothesis by examining the genomic abundance of the 1.688 satellite DNA which accounts for 4% of the Drosophila
. Again, we only observed a 2–2.5 fold increase in the bulk levels of the 1.688 satellite DNA (Figure S6
). It is clear that the heterochromatin is preferentially re-replicated in the absence of geminin; however, we are unable to rule out the possibility that a limited amount of stochastic re-replication is also occurring in the euchromatin.
In higher eukaryotes, there are many more MCM complexes loaded onto the chromatin in G1 than are required to complete an unperturbed S phase 
. Although these excess MCMs are not required for completion of a normal S phase, they are critical for protecting the cell from genomic instability during replication stress 
. However, when we deplete geminin, we only observe a minimal complement of MCMs being re-loaded onto heterochromatic DNA. These results suggest that there is not a global re-assembly of the pre-RC throughout the genome as occurs in G1 and that limiting amounts of MCMs are sufficient for the greater than two-fold increase in ploidy we observe. The MCMs appear to be transiently associated with the heterochromatin as inhibition of DNA re-replication with aphidicolin results in an increase in detectable MCMs. We propose that in the absence of geminin, the MCMs are loaded onto heterochromatic sequences and that these pre-RCs are immediately activated for initiation of DNA replication.
, Dup activity is downregulated after origin firing through multiple mechanisms including Cul4-Ddb1 mediated proteolysis in S phase and inhibition by geminin during S, G2 and mitosis. Furthermore, we (Figure S1
) and others 
have reported that Dup/Cdt1 is degraded in the absence of geminin. Thus, geminin is only one factor that negatively regulates Dup/Cdt1 and its depletion may be insufficient to induce genome-wide re-replication. Therefore, the deregulation of geminin and Dup/Cdt1 may have distinct effects on replication control. This may, in part, explain the differences in sequences and chromatin environment which are preferentially targeted for re-replication in human and Drosophila
cell lines. In human cell lines, Cdt1 overexpression led to the preferential re-replication of early replicating sequences 
, while in Drosophila
cell lines, geminin depletion leads to the preferential activation of heterochromatic origins of replication. Future experiments will test whether Dup levels are critical for maintaining ploidy and selecting which origins are activated.
The observation that the re-replication at pericentromeric heterochromatin was not coupled with late replication timing suggests that origin selection during re-replication and the temporal control of DNA replication in S phase are regulated by distinct mechanisms. These data suggest that a key determinant of which sequences will re-initiate DNA replication is the local chromatin environment. Drosophila
pericentromeric heterochromatin is marked by H3K9 methylation which is maintained by Su(var)
3-9 and HP1
. ORC has been shown to interact with HP1 and localizes to heterochromatin by immunofluorescence in both interphase and mitotic nuclei 
. It is therefore possible that the increased density of ORC in the heterochromatin may stimulate the preferential re-assembly of the pre-RC at those sequences. However, recent studies using GFP tagged ORC and live imaging did not observe an increased density of ORC at the heterochromatic regions of the genome 
We found that cyclin A-CDK activity regulates the re-activitation of replication origins at two levels. First, cyclin A-CDK activity is required for the large increase in ploidy observed, consistent with the known role of CDK activity in activating the pre-RC for initiation 
. Second, cyclin A-CDK activity appears to differentially inhibit pre-RC re-assembly in the euchromatin and heterochromatin. The simultaneous depletion of both cyclin A and geminin results in the global re-assembly of the pre-RC in both euchromatin and heterochromatin (). In contrast, depletion of only geminin results in pre-RC re-assembly specific to the heterochromatin, suggesting that cyclin A-CDK activity specifically inhibits pre-RC assembly in the euchromatin. In humans, the N-terminal domain of ORC1 contains consensus CDK phosphorylation sites which can be phosphorylated in vitro
by cyclin A-CDK activity and may regulate the SCF/Skp2 mediated turnover of ORC1 during S phase 
. In Drosophila
, the ORC1 N-terminus also contains potential CDK phosphorylation sites, an additional O-box for APC mediated destruction 
, and is essential for the binding of ORC1 with HP1 
. We propose that the interaction between ORC1 and HP1 may protect ORC1 from inhibitory cyclin A-CDK signals or destruction by the APC, thereby differentially sensitizing heterochromatic and euchromatic origins of replication to un-licensed pre-RC assembly.