Cdc6 ATPase Activity Regulates ORC·Cdc6 Stability and the Selection of Specific DNA Sequences as Origins of DNA Replication*

DNA replication, as with all macromolecular synthesis steps, is controlled in part at the level of initiation. Although the origin recognition complex (ORC) binds to origins of DNA replication, it does not solely determine their location. To initiate DNA replication ORC requires Cdc6 to target initiation to specific DNA sequences in chromosomes and with Cdt1 loads the ring-shaped mini-chromosome maintenance (MCM) 2–7 DNA helicase component onto DNA. ORC and Cdc6 combine to form a ring-shaped complex that contains six AAA+ subunits. ORC and Cdc6 ATPase mutants are defective in MCM loading, and ORC ATPase mutants have reduced activity in ORC·Cdc6·DNA complex formation. Here we analyzed the role of the Cdc6 ATPase on ORC·Cdc6 complex stability in the presence or absence of specific DNA sequences. Cdc6 ATPase is activated by ORC, regulates ORC·Cdc6 complex stability, and is suppressed by origin DNA. Mutations in the conserved origin A element, and to a lesser extent mutations in the B1 and B2 elements, induce Cdc6 ATPase activity and prevent stable ORC·Cdc6 formation. By analyzing ORC·Cdc6 complex stability on various DNAs, we demonstrated that specific DNA sequences control the rate of Cdc6 ATPase, which in turn controls the rate of Cdc6 dissociation from the ORC·Cdc6·DNA complex. We propose a mechanism explaining how Cdc6 ATPase activity promotes origin DNA sequence specificity; on DNA that lacks origin activity, Cdc6 ATPase promotes dissociation of Cdc6, whereas origin DNA down-regulates Cdc6 ATPase resulting in a stable ORC·Cdc6·DNA complex, which can then promote MCM loading. This model has relevance for origin specificity in higher eukaryotes.

Jacob and Brenner (1) described the replicon model in 1963 that outlined in general terms how DNA replication might start at a specific location within chromosomes. This model formed the basis for discovery of initiator proteins that recognize origins of DNA replication. The bacterial initiator, DnaA (2), rec-ognizes the bacterial origin of DNA replication, oriC, by binding to multiple 9-mer DnaA boxes within the origin (3). However, DnaA also binds frequently to DnaA boxes outside of DNA replication origins, and these sites do not promote initiation of DNA replication (4); therefore, the origin is defined by a combination of specific protein-DNA interactions, including DnaA, histone-like protein HU, and integration host factor proteins and an arrangement of DNA sequences that form multiple binding sites in a specific conformation (5). It is only this particular conformation that allows the initiator to promote origin unwinding, to recruit other DNA replication proteins such as the DnaB helicase and helicase loading protein DnaC to the DNA, and consequently to initiate DNA synthesis (5).
The mechanism that specifies the location of origins of DNA replication in eukaryotic chromosomes is less clear particularly in mammalian cells, but it involves a combination of sequence specific protein-DNA interactions, epigenetic factors, and cell cycle timing (6,7). In Saccharomyces cerevisiae, specific DNA sequences are known to contribute the location of origins of DNA replication in chromosomes (8 -10). Knowledge of such sequences led to the discovery of the origin recognition complex (ORC) 3 (11), a six-subunit complex. Homologues of ORC exist in all eukaryotic species. Yeast origins contain several conserved genetic elements (12); the A element, which serves as a binding site for ORC (11), the B1 element, which is partially involved in ORC binding (13,14), and the B2 element as a potential binding site for loading the mini-chromosome maintenance (MCM) proteins (15,16). ORC in higher eukaryotes does not bind specifically to DNA but has a preference for ATrich (17,18) or supercoiled DNA (19).
Our understanding of how DNA sequences or locations within the chromosome are selected as sites of initiation of DNA replication is still poor since it is unclear what defines a start site for initiation of DNA replication. S. cerevisiae ORC binds to origins in a DNA sequence-specific fashion; however, this feature is not conserved in higher eukaryotes. But because the protein components required for initiation of DNA replication are conserved, it is expected that studies of how this process occurs in yeast will shed light on the process in all eukaryotes. In budding yeast it has been shown that the Orc1 subunit in ORC has to be in the ATP-bound form for ORC to bind DNA (11,20). Orc1-dependent ATPase, stimulated by an arginine residue in an adjacent Orc4 subunit, is also regulated in a DNA sequence-specific manner (20), but this ATPase activity has no influence on the selection of DNA sequences as origins of DNA replication (21). Although ATP-bound ORC can bind to origins of DNA replication specifically, its affinity cannot explain why some DNA sequences are selected as origins and others are not, e.g. ORC binding sites have been identified by chromatin immunoprecipitation analysis that do not act as replication sites (10).
Another protein involved in the sequential assembly of prereplicative complexes (pre-RCs) at origins of DNA replication is the Cdc6 protein (cdc18 in Schizosaccharomyces pombe) (22)(23)(24)(25)(26)(27). Cdc6 is a member of the AAAϩ family of proteins and has striking amino acid sequence similarity with Orc1, the largest subunit of ORC (28,29). Together with ORC and Cdt1, Cdc6 is required for loading the MCM proteins onto origin DNA (30 -36). Cdc6 binds to and changes the structure of ORC, and together they bind cooperatively to origin DNA with higher DNA sequence specificity than ORC binding alone (37,38). Specific nucleotides within the origin DNA are required for ORC⅐Cdc6⅐DNA complex formation, and Orc1 ATPase mutants are less efficient in forming the ORC⅐Cdc6⅐DNA complex (38). We were interested in understanding how Cdc6 contributes to the selection of certain DNA sequences in chromosomes of yeast to function as origins of DNA replication, a process with potential relevance to origin selection in higher eukaryotes. Because Cdc6 is also an AAAϩ protein, like Orc1-5 (38,39), and many initiation proteins use their ATPase activity in a regulatory fashion (40), we analyzed the contribution of Cdc6 ATPase activity to ORC⅐Cdc6 and ORC⅐Cdc6⅐DNA complex assembly and its impact on the selection of DNA sequences as origins of DNA replication.

EXPERIMENTAL PROCEDURES
Proteins-ORC expressed in insect cells from recombinant baculoviruses and Cdc6 expressed in Escherichia coli were purified to apparent homogeneity as described (11,20,38).
ATPase Assays-ORC and Cdc6 ATP hydrolysis was carried out as described (20), with minor modifications. 2.5 pmol of ORC and 2.5 pmol of Cdc6 were incubated for 30 min on ice in 12 l of ATPase buffer (25 mM HEPES, pH 7.6, 100 mM potassium glutamate, 5 mM magnesium acetate, 1 mM dithiothreitol, 1 mM EDTA, 1 mM EGTA, 0.1% (v/v) Triton X-100, 10% glycerol) containing 2.5 pmol of DNA (when indicated) and 100 M ATP (unless otherwise indicated). After the incubation 5 Ci of [␣-32 P]ATP (3000 Ci/mmol) was added and the reaction was started by moving the tubes from ice into a room temperature water bath. At 15,30,45, and 60 min 2-l aliquots were removed and stopped with 0.5 l of 2% SDS stop solution. 1 l of the samples were consequently spotted on TLC plates and developed.
Determination of Dissociation Constants-Dissociation constants were determined using a gel shift assay. Gel shift reactions were performed as described previously (38) in the absence of competitor DNA. After challenging the complex with 100-fold cold competitor DNA (ARS1 291 bp) aliquots were loaded onto a continuously running gel in the cold room at the indicated time points. k off , R values, and t1 ⁄ 2 rates were determined using KaleidaGraph 4.0 (Synergy software). All experiments were repeated at least twice with similar results.

Cdc6 ATPase Is Induced by ORC⅐Cdc6
Complex Formation-ORC and Cdc6 have been shown to form a stable complex in the presence of a slowly hydrolysable ATP analogue, ATP␥S, but not in the presence of ATP (38). This finding suggests that ATP hydrolysis regulates complex stability. To address this question in more detail we analyzed ATPase activities and complex formation of ORC and Cdc6 (Fig. 1). Cdc6 on its own had no ATPase activity as reported (31,41) (Fig. 1A). The ORC ATPase activity was similar to what has been reported (20). In contrast to a recent study (31) that used different salt conditions, we found very robust ATPase activity when Cdc6 was added to ORC, a 2.5-fold induction over the ATPase activity of ORC alone. This result suggests that ORC and Cdc6 form at least a transient complex in the absence of DNA that induces either ORC or Cdc6 ATPase activity. To identify the active ATPases in the ORC⅐Cdc6 complex we analyzed two Cdc6 ATPase mutants (Cdc6 E224G and Cdc6 N263A) and an ORC ATP hydrolysis mutant (ORC4R/5A) in ATPase assays. Cdc6 E224G has a mutation in the highly conserved Walker B motif (31,35) and is dominant lethal when overexpressed in yeast. Cdc6 N263A (42,43) has a mutation in the sensor 1 motif and is temperature-sensitive and lethal when overexpressed. Cdc6 belongs to the class of AAAϩ proteins, and mutations in Walker B or sensor 1 motifs frequently interfere with the ability of AAAϩ proteins to hydrolyze ATP (40). The two Cdc6 mutants individually displayed no ATPase activity, similar to wild type Cdc6 (Fig. 1A). The addition of Cdc6 N263A to ORC did not change the ATPase activity, in contrast to wild type Cdc6. Cdc6 E224G addition to ORC led to a reduction of overall ATPase activity. These results suggest that Cdc6 is a major ATPase within the ORC⅐Cdc6 complex. To analyze the contribution of ORC toward ATPase activity of the ORC⅐Cdc6 complex, we analyzed an ORC4R/5A double mutant (Fig. 1B). Mutations in the arginine finger of ORC4 (ORC4R) have been shown to abolish Orc1-dependent ATPase (21). ORC5A carries a mutation in the Walker A motif of Orc5 that blocks ATP binding to ORC5; however, this mutation has no significant phenotype in vitro and in vivo (20,44,45). We measured ATPase of the combined ORC4R/5A mutant in the absence and presence of Cdc6 (Fig.  1B). ORC4R/5A showed a very weak ATPase that was dependent on high but physiological relevant ATP concentrations (up to 1 mM). The addition of Cdc6 to the ORC4R/5A mutant lead to a significant increase in ATPase activity, and this was dependent on high ATP concentrations as well. High ATP concentrations, however, had no effect on the ORC⅐Cdc6-induced ATPase (data not shown). These results show that ORC4R/5A can stimulate Cdc6 ATPase. Taken together, the ORC and Cdc6 ATPase data suggest that both ORC and Cdc6 are active ATPases in the ORC⅐Cdc6 complex since ATPase mutants in both ORC and Cdc6 did not abolish ATPase activity completely.

Cdc6 ATPase Regulates ORC⅐Cdc6
Stability-Using glycerol gradient sedimentation, we have shown that in the presence of ATP, ORC and Cdc6 did not form a stable complex, whereas in the presence of the non-hydrolyzable ATP analogue, ATP␥S, a stable ORC⅐Cdc6 complex was observed (38). This suggests that ATPase activity controls the stability of the ORC⅐Cdc6 complex in the absence of DNA. The ORC mutants ORC-4R, an Orc1dependent ATP hydrolysis mutant (21), and ORC-1A, an Orc1 ATP binding mutant (20), did not form a stable ORC⅐Cdc6 complex in the presence of ATP (38). These results indicate that ORC ATPase in the ORC⅐Cdc6 complex does not regulate ORC⅐Cdc6 complex stability, suggesting that it is the Cdc6 ATPase that is responsible. This possibility was examined by analyzing Cdc6 ATPase mutants for their ability to form a stable complex with ORC ( Fig. 1, C and D). Cdc6 E224G and Cdc6 N263A were fractionated either individually or together with ORC in the presence of ADP, ATP, or ATP␥S by glycerol gradient sedimentation (Fig. 1C). Cdc6 E224G and Cdc6 N263A individually fractionated as monomers. The Cdc6 ATPase mutants co-fractionated with ORC in the presence of ATP and ATP␥S, as seen by the disappearance of monomeric Cdc6 and concomitant appearance of Cdc6 in the peak fraction containing ORC, which was verified by Western blot with anti-Cdc6 antibodies (Fig. 1D). In the absence of ORC, Cdc6 E224G and N263A were not found in the fraction corresponding to the ORC⅐Cdc6 (Fig. 1D). Previously we have shown that in the presence of ADP, ORC and Cdc6 do not form a complex (38). This is also true for Cdc6 E224G; however, Cdc6 N263A formed a weak complex with ORC in the presence of ADP, probably due to inefficient exchange of ATP for ADP.
The results so far show that Cdc6E224G and Cdc6N263A were capable of forming a stable complex with ORC in the presence of ATP, which was not true for wild type Cdc6 protein.
The results suggest that, in the absence of DNA, the Cdc6 ATPase negatively regulates ORC⅐Cdc6 complex stability. Because Orc1 and Orc5 ATP binding or ATPase activity have no influence on ORC⅐Cdc6 complex stability (38), it is the Cdc6 ATPase within the ORC⅐Cdc6 complex that regulates complex stability. DNA Sequence-dependent ATPase Induction in the ORC⅐Cdc6⅐DNA Complex-Because the ORC⅐Cdc6 complex is assembled on DNA, we were interested in determining how DNA modulates the ATPase activity in the complex. In performing these experiments, wild type and mutant versions of origin DNAs were used as well as nonspecific natural DNAs that do not function as an origin in yeast. Thus, the relationship between origin binding and ATPase activity could be assessed. The ATPase of ORC, ORC4R/5A, Cdc6, and the two Cdc6 ATP hydrolysis mutants (Cdc6 E224G and Cdc6 N263A) were analyzed in the absence and presence of GC-rich non-origin DNA, ARS1 origin DNA (containing the A, B1, and B2 elements), and mutant ARS1 A Ϫ or A Ϫ B1 Ϫ B2 Ϫ DNA (Fig. 2). Note that both the A and B2 elements are ORC binding sites; B2 is only used as an ORC binding site in the absence of a functional A element (11,38).
Cdc6 had no ATPase in the absence or presence of DNA ( Fig.  2A, lanes 1, 4, 7, 10, and 13). In the absence of DNA, ORC is an ATPase (lane 2), and the addition of Cdc6 to ORC led to a 2.5-fold increase in ATPase compared with ORC alone (Fig. 2A,  lane 3). GC-rich DNA and ARS1 A Ϫ B1 Ϫ B2 Ϫ , which do not support initiation of DNA replication, ORC⅐DNA, and ORC⅐Cdc6⅐DNA complex formation (38), did not suppress ORC ATPase activity ( Fig. 2A, compare lane 2 with lanes 5 and  14). Rather, they induced ORC⅐Cdc6 ATPase activity 5-fold stronger than the ATPase in the presence of ORC alone ( Fig.  2A, compare lane 2 with lanes 6 and 15). This result suggests that transient binding of ORC and Cdc6 to GC-rich DNA acti-vates the ATPase in either ORC or Cdc6. ARS1 origin DNA suppressed ORC ATPase by 8 -10-fold ( Fig. 2A, lanes 2 and 8) as reported (20). ARS1 DNA also promoted stable complex formation of the ORC⅐Cdc6⅐ARS1 complex (38). The addition of ARS1 DNA to reactions containing ORC and Cdc6 suppressed the ATPase activity of ORC⅐Cdc6 about 5-fold compared with the ATPase activity in the absence of DNA ( Fig. 2A, compare  lanes 3 and 9). An ARS1 A Ϫ mutant, which supports weak binding of ORC but does not support ORC⅐Cdc6 complex formation on DNA (38), suppressed ORC ATPase activity likely due to the weaker, non-perfect ORC binding site within the B2 element ( Fig. 2A, compare lane 2 with lane 11). Interestingly, the ARS1 A Ϫ mutant did not suppress ORC⅐Cdc6 ATPase but resulted in a similar ATPase as the ORC⅐Cdc6 complex in the absence of DNA ( Fig. 2A, compare lane 3 with lane 12). This result suggests that ORC can bind to the ARS1 A Ϫ via the weak ORC binding site within the B2 element but that addition of Cdc6 triggers ATPase activity that disrupts the ORC⅐Cdc6 binding to the ARS1 A Ϫ DNA. These data indicate that ORC⅐Cdc6 complex formation on DNA is regulated by ATPase activity and that the ORC and the ORC⅐Cdc6 complexes differ in the way they respond to non-origin DNA sequences, as exemplified by the striking differences between ARS1 and ARS1 A Ϫ DNA.
Cdc6 Is the Active ATPase in the ORC⅐Cdc6⅐DNA Complex-Next we were interested in identifying the DNA-regulated ATPase within the ORC⅐Cdc6⅐DNA complex. We, therefore, analyzed mutant versions of these proteins, including ORC4R/5A and Cdc6 E224G and N263A for their ability to regulate ORC⅐Cdc6 ATPase activity. ORC4R/5A, a very weak ATPase, was reproducibly suppressed by ARS1 DNA but not by GC-rich DNA or ARS1 A Ϫ B1 Ϫ B2 Ϫ (Fig. 2B, lanes 2, 5, 8, 11,  and 14). The addition of Cdc6 to ORC4A/5A resulted in stronger ATPase in the presence and absence of DNA, similar to the case of ORC and Cdc6, indicating that ORC ATPase is not the major ATPase within the ORC⅐Cdc6⅐DNA complex (Fig. 2B,  lanes 3, 6, 9, 12, and 15). Interestingly, based on the Cdc6⅐activated ATPase, this mutant complex was still able to distinguish between DNAs that acted as an origin and DNAs that did not.
The two Cdc6 mutants, Cdc6 E224G (31) and N263A, did not increase the ATPase when compared with ORC alone in the absence or presence of any DNA (Fig. 2, C and D). Interestingly, the dominant negative Cdc6 E224G mutant resulted in reducing the ATPase of ORC in solution or in complex with non-origin GC-rich or ARS1 A Ϫ and A Ϫ B1 Ϫ B2 Ϫ DNA (Fig. 2D). Combined, these results suggest that Cdc6 is the major ATPase within the ORC⅐Cdc6 complex, independent of whether DNA is present or not. Furthermore, origin DNA, but not non-origin DNAs, suppressed the ORC⅐Cdc6 ATPase activity, suggesting that the ORC⅐Cdc6 complex and the Cdc6 ATPase within it could discriminate between functional and non-functional origin sequences. For example, even though the non-origin ARS1 A Ϫ DNA has an ORC binding site within it and this sequence is capable of suppressing ORC ATPase activity ( Fig. 2A, lanes 2 and 10), this DNA was incapable of suppressing the ORC⅐Cdc6 ATPase activity.
B1 and B2 Origin Mutations Induce ATPase within the ORC⅐ Cdc6⅐DNA Complex-The results presented in Fig. 2, combined with those in a previous report (38), suggest that multiple DNA sequences within an origin regulate the ATPase activity of the ORC⅐Cdc6 complex. We were, therefore, interested in quantifying the contribution of the genetically conserved B1 and B2 origin elements to the ATPase activity (Fig. 3A). A linker scanning mutant in B1, a point mutant in B1 (A838G), and a B2 linker scanning mutant were analyzed. The linker mutation of the B1 element has been shown to reduce the affinity of ORC; however, a point mutant in the B1 element, ARS1 A838G, and a linker scanning mutation of the B2 element have no influence on the ORC affinity for DNA (13). The B1 linker scanning mutation abolished and the point mutant reduced ORC⅐Cdc6 complex formation on DNA (38). The mutation of the B2 element had only a very weak influence on ORC⅐Cdc6 complex formation on DNA. 4 In ATPase assays with ARS1 B1 Ϫ , B1 A838G, and B2 Ϫ mutants, the addition of Cdc6 to ORC resulted in a 3-, 2.3-, and 1.8-fold activation of ATPase respectively, compared with the activity of ORC alone (Fig. 3A). ORC ATPase activity on its own was at the same time also slightly increased in these mutants compared with the activity on ARS1 DNA (Fig. 3A). In contrast, the A Ϫ mutant resulted in 8.5-fold activation of ORC⅐Cdc6 ATPase activity, indicating that mutations in the B elements have a 4 C. Speck and B. Stillman, unpublished data. weaker but significant impact on activation of ATPase. Thus, Cdc6 ATPase activity in the ORC⅐Cdc6 complex was most influenced by ORC binding sites that contribute most to origin activity, but additional sequences that do not affect ORC binding also alter the regulation of Cdc6ATPase activity.
Origins of Replication Suppress ORC⅐Cdc6 ATPase-We tested whether mutations in ARS A elements at other origins have a similar impact. The ARS607 and an ARS607 A Ϫ mutant DNAs were analyzed. ARS607 suppressed ORC ATPase efficiently, and the addition of Cdc6 resulted in increased ATPase activity, similar to ARS1 (Fig. 3B). Interestingly a mutation in the A element resulted in significant ORC ATPase, comparable with ORC ATPase in the absence of DNA, and the addition of Cdc6 induced very strong ATPase. This situation is similar to an ARS1 A Ϫ B1 Ϫ B2 Ϫ mutant, suggesting that ARS607 has no secondary ORC binding site within its B2 element. The ARS607 data suggest that Cdc6-induced ATPase is down-regulated at origins of replication due to ARS A elements.
Cdc6 ATPase Mutants Promote Binding of the ORC⅐Cdc6 Complex to Nonspecific DNA-The above-described experiments established that Cdc6 ATPase in the ORC⅐Cdc6 complex was regulated in an ORC and DNA sequence-specific manner. Furthermore, we have shown that Cdc6 influences the DNA binding specificity of the ORC⅐Cdc6 complex (37,38). We were, therefore, interested in determining if Cdc6 ATPase regulates the DNA sequence specificity of ORC⅐Cdc6 interactions with DNA.
Either wild type Cdc6 or the two Cdc6 ATPase mutants were tested for DNA binding in the absence or presence of ORC using a gel shift assay (Fig. 4, A and B) and a DNase I footprint assay (Fig. 4C). Cdc6 and its mutants had very weak and nonspecific binding to DNA (Fig. 4, A  and B, lanes 2, 11, and 21). In the absence of competitor DNA the addition of Cdc6 E224G and N263A formed an ORC⅐Cdc6⅐ARS1 complex like wild type Cdc6 (Fig. 4A, compare lanes 4 -9, 13-18, and 22-27). However, in the presence of competitor DNA, binding of the two mutant Cdc6 proteins were significantly impaired (Fig. 4B, compare lanes 4 -9, 13-18, and 22-27; note that in the presence of competitor DNA 5-fold more ORC was used to detect ORC binding). This result indicates that Cdc6 ATPase mutants reduced DNA sequencespecific complex formation. The DNase I footprint assay showed that Cdc6 ATPase mutants did not influence ORC binding at the A and B1 elements of the ARS1 origin; however, Cdc6-induced protection adjacent to the B1 element was lost (Fig. 4C, dotted line). More significant, the two Cdc6 mutants, but not the wild type Cdc6, induced significant nonspecific DNA binding in multiple regions outside the genetically defined origin (Fig. 4C, arrows). Because the Cdc6 mutants in the absence of ORC had no DNase I footprint, we suggest that the ORC⅐Cdc6 ATPase mutant stabilizes the ORC-mutant Cdc6 complex in regions outside the origin and, therefore, has reduced sequence specificity. Similar results have been seen when the ORC⅐Cdc6 complex was analyzed in the presence of ATP␥S (38).
The Mechanism of Cdc6-induced Sequence Specificity-The DNA binding experiments established that the Cdc6 ATPase activity contributes to sequence specific DNA binding to origin DNA; however, the mechanism of how this activity contributes to origin selection remained unclear. Theoretically, Cdc6 ATPase can regulate DNA sequence specific binding in several ways, by 1) stabilizing ORC binding to the origin, 2) destabilizing ORC on non-origin DNA (37) or by regulating Cdc6 interaction with ORC that is not bound to origin DNA, 3) binding preferentially to ORC that is origin bound, or 4) release of Cdc6 from non-origin DNA. Because many biological regulatory interactions are controlled at the level of dissociation and the footprint experiments suggested a dissociation defect of the Cdc6 ATPase mutants when in a complex with ORC, we measured the dissociation rates of ORC, Cdc6, and Cdc6 N263A in the presence of 32 P-labeled ARS1 or ARS1 A Ϫ DNAs. ORC, Cdc6, and origin DNA were incubated in the absence of competitor DNA and then challenged by the addition of 100ϫ wild type ARS1 DNA. The reactions were loaded at time intervals onto a running native polyacrylamide gel, and the dissociation rates were determined from the autoradiogram. The t1 ⁄2 of ORC on origin DNA was determined to be 64.9 min., indicating a very slow turnover (Fig. 5A, Table 1). The binding of ORC to ARS1 A Ϫ displayed increased turnover, with a 17.5ϫ decreased t1 ⁄2 of 3.7 min (Fig.  5B, Table 1). Wild type Cdc6 showed a biphasic dissociation rate from origin DNA, one fraction of Cdc6 coming off DNA immediately upon the addition of competitor and another fraction with a t1 ⁄2 of 62 min, similar to ORC (Fig. 5C, Table 1). Interestingly, 15 s after competitor DNA addition, all the Cdc6 was dissociated from an ORC⅐ARS1 A Ϫ DNA complex; the t1 ⁄2 was consequently below our detection limit (Fig. 5D, Table 1). The Cdc6 N263A mutant, incubated with ARS1 and ORC, showed in contrast to wild type Cdc6, a dissociation rate with a t1 ⁄2 of 29.7 min (Fig. 5E, Table 1). This lower t1 ⁄2 for ORC⅐Cdc6 N263A compared with the wild type ORC⅐Cdc6 probably reflects the measurement of the combined dissociation of two separate complexes, 1) Cdc6 N263A that is nonspecifically bound to the DNA and 2) the Cdc6 N263A that is within the ORC⅐Cdc6 N263A⅐DNA complex. In contrast, for wild type Cdc6, the nonspecifically bound Cdc6 immediately dissociated from the origin DNA upon the addition of competitor DNA, leaving only the ORC⅐Cdc6 complex to dissociate.
In the case of ARS1 A Ϫ the dissociation rate of Cdc6 N263A was measurable at t1 ⁄ 2 of 9.9 min (Fig. 5F, Table 1), which is in stark contrast to Cdc6 that had a dissociation rate 100ϫ faster. Interestingly Cdc6 N263A led to a 2.7-fold stabilization of ORC on ARS1 A Ϫ (compare Fig.  5, D and F), indicating that the ORC⅐Cdc6 N263A⅐ARS1 A Ϫ complex is overall more stable than the ORC⅐ARS1 A Ϫ complex, indicating that Cdc6 promotes moderately higher ORC⅐Cdc6⅐DNA complex stability in the absence of Cdc6 ATPase. This result also fits with the increased nonspecific binding of the ORC⅐Cdc6 N263A⅐DNA complex seen in the DNase I footprint assay in regions outside the primary ORC binding site (Fig. 4C). Overall, these results indicate that Cdc6 ATPase regulates the disassembly of the ORC⅐Cdc6 complex on non-origin DNA. The major mechanism is the release of Cdc6 from ORC bound to non-origin DNA due to Cdc6 ATPase. In the first lane ORC and ARS1 A Ϫ (no competitor DNA) were loaded as a reference. Note that ORC binds to A Ϫ DNA non-specifically in the absence of competitor DNA. F, gel shift assay of 32 P-labeled ARS1 A Ϫ DNA with ORC and Cdc6 N263A. In the first lane ORC and ARS1 A Ϫ (no competitor) were loaded as a reference. A ϩ B were used to determine ORC dissociation rates from DNA, and C-F were used to determine Cdc6 and Cdc6 N263A dissociation rates from ORC⅐DNA complexes.

DISCUSSION
In previous studies we have analyzed ORC⅐Cdc6 complex formation in the presence and absence of DNA and showed that Cdc6 confers increased sequence specificity to the ORC⅐Cdc6 complex (37,38). We also showed that ORC and Cdc6 form a ring-shaped complex with a surface of similar dimension and shape to the ring-shaped surface of the MCM-hexamer, which likely functions as a MCM loading machine before initiation of DNA replication (38). In this study we concentrated on analyzing the mechanism by which Cdc6 increases DNA sequence specificity or origin recognition and stability of the ORC⅐Cdc6⅐DNA complex, both of which are relevant to understanding why origins of DNA replication locate to specific regions within chromosomes for pre-RC formation and MCM loading (31).
Six of the seven proteins within the ORC⅐Cdc6 complex are predicted to belong to the class of AAAϩ proteins (38,39). ATP binding and hydrolysis frequently regulate the function of AAAϩ proteins (40). In our previous study we focused on the role of ORC ATP binding and ORC ATPase in regulating ORC⅐Cdc6 complex (38). In this study the contribution of Cdc6 ATPase on ORC⅐Cdc6 interaction and sequence specificity was investigated. Mutants interfering with Cdc6 ATPase have been characterized in vivo; Cdc6 E224G has a mutation in the highly conserved Walker B motif (31,35) and is dominant lethal when overexpressed in yeast. Cdc6 N263A (42,43) has a mutation in sensor 1 and is temperature-sensitive and lethal when overexpressed. In the absence of DNA we found that Cdc6 addition to ORC resulted in rapid Cdc6-induced ATP hydrolysis and dissociation from ORC. Cdc6⅐ADP, which is produced during ATP hydrolysis, is not capable of binding to ORC (38). Cdc6 mutants defective for ATPase did not show increased ATPase in the presence of ORC and resulted in stable complex formation. The Cdc6 E224G mutant actually inhibited ORC ATPase activity, perhaps explaining why this mutant is a dominant lethal. The biological function of ORC-induced Cdc6 ATPase in the absence of DNA is likely the forced disassembly of the complex to ensure that the ORC⅐Cdc6 complex only forms on origin DNA (38).
In the presence of origin DNA, a stable ORC⅐Cdc6 complex was formed (31,38), and Cdc6 ATPase activity was induced, but this low level ATPase could originate from partial binding of the ORC⅐CDC6 complex to regions outside the origin that are unable to suppress Cdc6 ATPase. In the presence of non-origin DNA, however, the Cdc6 ATPase activity was much higher, and the complex was not stable on DNA. Mutations in conserved genetic elements of ARS resulted in activation of ATPase activity, which was dependent on Cdc6 ATPase. The A element represents the primary binding site for ORC (11), the sequence of A elements are most conserved within different ARS, and point mutants between the A-element have drastic phenotypes on plasmid stability (11,12). Mutations in the A element have also been shown to reduce ORC binding (11). We suggest that Cdc6 activated ATPase, and consequently, disassembly of the ORC⅐Cdc6⅐DNA complex is a reason for the low plasmid stability of ARS A box mutants in addition to the reduced affinity of ORC for DNA. Similarly, mutations in the B elements are important for origin function; however, in this case only simultaneous deletion of at least two of the B elements results in nonfunctional origins (12). Mutations in the B1 element reduce slightly the affinity of ORC for origin DNA (13,38), and mutations in the B2 element can be compensated by Cdc6 overexpression (15). We found that mutations in the B elements resulted in increased Cdc6 ATPase activity, which promotes disassembly of the complex. Cdc6 overexpression may be sufficient to force ORC⅐Cdc6 complex formation on the defective origin and, therefore, could compensate for B2 mutations.
The ARS1 A838G mutation is of particular interest because this mutant binds to ORC well but is partially defective in the Cdc6⅐dependent extended footprint on the ARS1 origin DNA (38) and has reduced ability to suppress the Cdc6-induced ATPase activity of the ORC⅐Cdc6 complex. We, therefore, suggest that specific nucleotide sequences modulate origin utilization by targeting the ORC⅐Cdc6 complex rather than ORC alone, again suggesting that Cdc6 contributes to origin selection.
It is known that origins with weak ORC binding sites can still function efficiently (46). We suggest that these DNA sequences bind the ORC⅐Cdc6 complex more efficiently to promote pre-RC formation. Furthermore, when Cdc6 is overexpressed in certain circumstances, re-replication of only a subset of origins of DNA replication occurs (47,48). Because Cdc6 contributes to origin selection and utilization, this may be one reason why not all origins are re-replicated when Cdc6 is present in excess.
Cdc6 ATPase within the ORC⅐Cdc6⅐DNA complex, first described by Randell et al. (31), regulates the stability of the ORC⅐Cdc6 complex. The function of the ORC⅐Cdc6 complex is to load in cooperation with Cdt1 the potential MCM helicase onto DNA to form a pre-RC complex (49). Based on our new findings that Cdc6 ATPase controls the stability of ORC⅐Cdc6 complex on DNA, we suggest that it is the stability of the ORC⅐Cdc6 complex on origin DNA, but not on non-origin DNA, that regulates directly the location in chromosomes and amount of MCM loading. Cdc6 ATPase contributes to the selection of DNA sequences that can promote MCM loading and, hence, formation of a pre-RC. The ORC⅐Cdc6 complex stability depends on several factors, many of which are hardwired into the DNA sequence at origins, namely the A and B elements, but we envision that chromatin structure might also affect complex stability.
It has been shown that Cdc6 availability is further regulated by association with cyclin (50), and Cdc6 destruction is regu-

of ORC-DNA and ORC⅐Cdc6⅐DNA complexes
Based on the data exemplified in Fig. 5, dissociation rates (k off ), R values, and t 1/2 were calculated using KaleidaGraph 4.0 (Synergy software). The rates were calculated for ORC or Cdc6 as indicated in bold. ORC⅐Cdc6⅐A Ϫ dissociation rates were too fast to be determined and, therefore, are labeled (ND). The t 1/2 was estimated to be lower than 0.125 min since this is the lower detection limit of this assay. lated by ubiquitin-mediated proteolysis (51)(52)(53). Cyclin-Cdc6 interaction and proteolysis of Cdc6 regulate the overall availability of Cdc6 during the cell division cycle, but when Cdc6 is available to bind to ORC before pre-RC assembly, DNA sequence information is a significant determinant in modulating Cdc6 activity and, hence, where on the DNA pre-RCs are formed. Because subtle changes in the DNA sequence influence Cdc6 ATPase activity, Cdc6⅐ORC complex stability can also be represented as having a specific probability at any given DNA sequence and, therefore, contributes to selection of functional origins. This fits well with the recent data in fission yeast in which origins of DNA replication are selected in a stochastic manner dependent on DNA sequence and do not function every cell cycle (54 -56). We suggest that Cdc6 (cdc18 in S. pombe) ATPase activity influences those DNA sequences that are selected as origins of DNA replication and even the stochastic firing of origins in each S phase. Such a scenario may exist for the selection of sites of DNA replication in the chromosomes of plant and animal cells where ORC binding is not known to be sequence specific. In fact, Cdc6 is a rate-limiting component for the initiation of DNA replication in vertebrate cells (36,57), and we suggest that it is the interaction between ORC and Cdc6 and the subsequent ATPase activities of these proteins that determines when and where in chromosomes initiation of DNA replication takes place. Cdc6 ATPase activity within the ORC⅐Cdc6 complex is suppressed by origin DNA when compared with the activity in the absence of DNA and activation of Cdc6 ATPase, due to nonorigin DNA, leads to disassembly of Cdc6 from DNA. It is possible that ORC bound to non-origin DNA has an altered conformation that stimulates Cdc6 ATPase upon complex formation. This results in ADP⅐Cdc6, which cannot interact with ORC and is released from the ORC⅐DNA complex. Once ADP⅐Cdc6 is in solution, it has to be recharged with ATP. This could be a passive reaction or actively modulated by a protein.

Complex
On the other hand a specific protein present during pre-RC assembly might modulate Cdc6 ATPase to promote several rounds of MCM loading. Alternatively, a protein may stimulate yeast Cdc6 ATPase activity on origin DNA to promote dissociation of Cdc6 from the origin before its destruction at the G 1 /S phase transition and, hence, hinder re-replication in a single cell cycle.