The Essential Role of Saccharomyces cerevisiae CDC6 Nucleotide-binding Site in Cell Growth, DNA Synthesis, and Orc1 Association*

Saccharomyces cerevisiaeCdc6 is a protein required for the initiation of DNA replication. The biochemical function of the protein is unknown, but the primary sequence contains motifs characteristic of nucleotide-binding sites. To study the requirement of the nucleotide-binding site for the essential function of Cdc6, we have changed the conserved Lys114 at the nucleotide-binding site to five other amino acid residues. We have used these mutants to investigate in vivo roles of the conserved lysine in the growth rate of transformant cells and the complementation of cdc6temperature-sensitive mutant cells. Our results suggest that replacement of Lys with Glu (K114E) and Pro (K114P) leads to loss-of-function in supporting cell growth, replacement of the Lys with Gln (K114Q) or Leu (K114L) yields partially functional proteins, and replacement with Arg yields a phenotype equivalent to wild-type, a silent mutation. To investigate what leads to the growth defects derived from the mutations at the nucleotide-binding site, we evaluated its gene functions in DNA replication by the assays of the plasmid stability and chromosomal DNA synthesis. Indeed, the K114P and K114E mutants showed the complete retraction of DNA synthesis. In order to test its effect on the G1/S transition of the cell cycle, we have carried out the temporal and spatial studies of yeast replication complex. To do this, yeast chromatin fractions from synchronized culture were prepared to detect the Mcm5 loading onto the chromatin in the presence of the wild-type Cdc6 or mutant cdc6(K114E) proteins. We found that cdc6(K114E) is defective in the association with chromatin and in the loading of Mcm5 onto chromatin origins. To further investigate the molecular mechanism of nucleotide-binding function, we have demonstrated that the Cdc6 protein associates with Orc1 in vitro and in vivo. Intriguingly, the interaction between Orc1 and Cdc6 is disrupted when the cdc6(K114E) protein is used. Our results suggest that a proper molecular interaction between Orc1 and Cdc6 depends on the functional ATP-binding of Cdc6, which may be a prerequisite step to assemble the operational replicative complex at the G1/S transition.

Saccharomyces cerevisiae Cdc6 is a protein required for the initiation of DNA replication. The biochemical function of the protein is unknown, but the primary sequence contains motifs characteristic of nucleotidebinding sites. To study the requirement of the nucleotide-binding site for the essential function of Cdc6, we have changed the conserved Lys 114 at the nucleotidebinding site to five other amino acid residues. We have used these mutants to investigate in vivo roles of the conserved lysine in the growth rate of transformant cells and the complementation of cdc6 temperature-sensitive mutant cells. Our results suggest that replacement of Lys with Glu (K114E) and Pro (K114P) leads to lossof-function in supporting cell growth, replacement of the Lys with Gln (K114Q) or Leu (K114L) yields partially functional proteins, and replacement with Arg yields a phenotype equivalent to wild-type, a silent mutation. To investigate what leads to the growth defects derived from the mutations at the nucleotide-binding site, we evaluated its gene functions in DNA replication by the assays of the plasmid stability and chromosomal DNA synthesis. Indeed, the K114P and K114E mutants showed the complete retraction of DNA synthesis. In order to test its effect on the G 1 /S transition of the cell cycle, we have carried out the temporal and spatial studies of yeast replication complex. To do this, yeast chromatin fractions from synchronized culture were prepared to detect the Mcm5 loading onto the chromatin in the presence of the wild-type Cdc6 or mutant cdc6(K114E) proteins. We found that cdc6(K114E) is defective in the association with chromatin and in the loading of Mcm5 onto chromatin origins. To further investigate the molecular mechanism of nucleotidebinding function, we have demonstrated that the Cdc6 protein associates with Orc1 in vitro and in vivo. Intriguingly, the interaction between Orc1 and Cdc6 is disrupted when the cdc6(K114E) protein is used. Our results suggest that a proper molecular interaction between Orc1 and Cdc6 depends on the functional ATPbinding of Cdc6, which may be a prerequisite step to assemble the operational replicative complex at the G 1 /S transition.
Cell cycle regulation is a complicated but highly coordinated process. It has a conserved mechanism among eukaryotes from yeast to human. The primary control of the eukaryotic cell cycle is provided by a family of cyclin-dependent kinases (CDKs) 1 and their associated cyclins, which regulate kinase activity. In unicellular yeast cells, a single prototype CDK gene, CDC28 in the budding yeast Saccharomyces cerevisiae or cdc2 ϩ in the fission yeast Schizosaccharomyces pombe functions at different cell cycle stages. Different cyclins activate the same kinase as different points in the cycle. It is now known that the central control of cell cycle progression by CDK complexes is regulated positively and negatively to monitor each step of the progression (1,2). This regulatory control, associated with checkpoints, orchestrates various types of cell cycle genes throughout the cell cycle.
In S. cerevisiae, over 70 temperature-sensitive cell division cycle (cdc) mutations have been isolated that control events throughout the cell cycle (3,4). One of them, CDC6, is required in the late G1 and S phases of the cell cycle. At the nonpermissive temperature, the cdc6 mutant cells show a DNA synthesis defect (5). The cdc6 mutants undergo increased chromosomal loss and hyper-recombination, suggesting a fairly direct role in DNA replication (6). Tandem copies of different ARSs added to the mini-chromosome can suppress mitotic loss in cdc6, but not in cdc7, cdc9, cdc16, or cdc17 mutants (7), suggesting a role at replication origins. The CDC6 mRNA fluctuates periodically throughout the cell cycle (8 -10), and its nuclear entry is cell cycle-dependent (11). The 5Ј-untranslated region of the CDC6 gene is very similar to a group of cell cycle genes that are either precursor enzymes for DNA synthesis or are directly involved in DNA replication (12). The involvement of Cdc6 in DNA replication has also been supported by direct analysis of origin function in cdc6 -1 mutants using two-dimensional gel method (13). The studies suggest that the Cdc6 and Orc5 protein interact and determine the frequency of initiation of DNA replication in yeast. Genetic interaction between CDC6 and ORC6 has also been reported (14). The S. pombe cdc18 ϩ is homologous to S. cerevisiae CDC6 (15,16). It has been proposed that the cdc18 ϩ gene plays two roles in the cell cycle, mediating the initiation of DNA replication and preventing inappropriate mitosis. In S. cerevisiae, over-expression of CDC6 induces G 2 delay of the cell cycle (17). Both results suggest a possible second function of CDC6 at the G 2 /M phase boundary. Homologs of yeast Cdc6 have also been found in human (18) and Xenopus (19). The conservation of Cdc6 among species implies an essential role for this protein.
Genomic footprinting has defined two cell cycle stages regarding the formation of replication complexes (RC): pre-RC in G 1 phase and post-RC after S phase to M/G 1 border (20 -22). The post-RC footprint closely resembles that produced in vitro with purified ORC and Abf1, which protect A/B1 and B3 elements, respectively. The pre-RC is defined by its further protecting the B2 element of ARS1 in addition to the A/B1 and B3 elements. The data suggest that the initiation of DNA replication is not controlled by the binding of ORC and Abf1 to the origins; instead, modification of origins has to occur with the involvement of some additional factors. The Cdc6 protein is required for establishment and maintenance of pre-RC, because the arrested cdc6 -1 mutant produces a post-replication footprint (23)(24)(25). The Cdc6 protein is also required for the loading of MCM proteins (26 -30). In addition, the Cdc6 protein interacts with Cdc28 protein complex after p40 sic1 is degraded at late G 1 phase (31). We have recently found that the Cdc6 can stimulate Abf1 binding to the B3 domain of ARS1 DNA fragment (32). Currently, it is believed that interactions of Cdc6 with a variety of other protein complexes, such as the ORC, the MCM complex, and Cdc28 kinase are important in coordinating chromosome replication and cell cycle control (33). However, the detailed molecular mechanism of the CDC6 gene action is still largely unknown.
There is a conserved nucleotide-binding site in Cdc6 (34). It has been shown that the yeast Cdc6 is able to bind ATP and GTP (10). At present, the physiological roles of the nucleotidebinding site are still obscure. One effective way to study the function of the nucleotide-binding site is to carry out sitespecific mutagenesis experiments to alter the indispensable Lys residue, and examine its phenotypic consequence. In these studies, we have followed a series of steps to unravel the functions of the nucleotide-binding site. First, we changed the conserved Lys 114 at the nucleotide-binding site to five other amino acid residues. We show that replacement of Lys with Glu (K114E) and Pro (K114P) leads to loss of function in supporting cell growth, replacement of the Lys with Gln (K114Q) or Leu (K114L) yields partially functional proteins, and replacement with Arg yields a phenotype equivalent to wild-type. Second, based on the plasmid stability assay and chromosomal DNA synthesis assay, the roles of the conserved Lys is most likely required for DNA replication. This guides us to investigate the formation of a functional replication complex at G 1 /S transition. Third, we have prepared synchronous chromatin fractions of both mutant and wild-type proteins to examine the effect of the cdc6(K114E) mutation. We found that the loading of the cdc6(K114E) and Mcm5 proteins is drastically altered in the chromatin. Finally, at the molecular level, we demonstrated that Cdc6 interacts with Orc1 in vitro and in vivo. Interestingly, the interaction between Orc1 and Cdc6 is impaired when the cdc6(K114E) protein is used. Our results suggest that a proper molecular interaction between Orc1 and Cdc6 depends on the functional ATP binding of Cdc6. This may lead to assemble the operational replicative complex at the G 1 /S transition. The biological significance of the ATP binding in this process is discussed in this report.

EXPERIMENTAL PROCEDURES
Strains, Plasmids, and Media-The yeast cdc6 -1 strain, 611, has been described (31). Strain K4055 was CDC6::hisGURA3hisG,trp1-1::TRP1 MET-CDC6 with URA3 looped out (9). Strain BJ2168 (a prc1-407 prb1-1122 pep4 -3 leu2 trpl ura3-52) is a protease-deficient strain. Plasmid carrying alleles mutated at lysine 114 were generated by cloning a 1.8-kb HindIII-EcoRI fragment encompassing the CDC6 ORF into M13mp19, mutagenizing as described below, and cloning the resulting alleles into YEp352 (35). Lysine 114 has been changed to Glu (K114E), Gln (K114Q), Arg (K114R), Pro (K114P), and Leu (K114L) (see Fig. 1). The YEp352 subclones are designated: YEp352-K114E, YEp352-K114L, YEp352-K114R, YEp352-K114P, and YEp352-K114Q (Table I). Plasmid YGp12 for the expression of untagged proteins were prepared by cloning the wild-type and K114E allele under the control of the GAL1,10 promoter in a 2-m yeast vector with URA3 as selectable marker. Plasmid YGp102 is the same construct except that LEU2 is the selection marker (Fig. 2). YCp5N is derived from a single-copy shuttle vector YCplac111 and contains a 0.7-kb 5Ј upstream sequence of the CDC6 gene (34). A T7 tag was inserted into 3Ј end of the CDC6 open reading frame in which the TAG codon of CDC6 gene had been removed, and the fusion was subcloned into YCp5N. YEp-HA-Mcm5 is generated by inserting a MCM5 fragment into a high copy vector, YEp352, in which Mcm5 is HA-tagged at its C-terminal 751 amino acids and the expression of the MCM5 gene is driven by its own 444-bp promoter sequence (36). A full-length of ORC1 gene was subcloned into bacterial expression vector pET28b, pET28a-ORC1, for the expression of 6xHis-Orc1 protein YGp181-His-T7-ORC1 contains 6xHis and T7 tag at the N terminus of the ORC1 gene driven by the Gal promoter with Leu2 selection marker (Fig. 7C). YGp123 is the vector with the insertion of 0.8-kb Gal1-Gal10 promoter on the vector YEplac112 with Trp1 as the selection marker. The wild-type CDC6-T7 tag, cdc6(K114E)-T7 tag and temperature-sensitive cdc6 -1-T7 tag genes were subcloned into YGp123, individually, for co-expression experiments (Fig. 7C). Yeast media YPD and CSM medium SD-ura were purchased from Bio101, Inc. All chemicals were ordered from Sigma; and restriction enzymes were purchased from Life Technologies, Inc. or New England Biolabs.
In Vitro Mutagenesis-Muta-Gene in vitro mutagenesis kit (Bio-Rad) was used to carry out the experiments as described below. The mutagenic (GGT CCG CCT GGC ACT GGC (G/C)(G/C/T)G ACT) and universal primers were synthesized (12,37); T7 DNA polymerase instead of the Klenow fragment of DNA polymerase I was used in the reaction (38). The mutagenized DNA is shown in Fig. 1.
Plasmid Stability Assay-Yeast cdc6 -1 shows elevated chromosomal loss. The wild-type and mutant cdc6 genes were transformed into strain cdc6 -1 to examine their plasmid stability. A colony grown on selective medium was resuspended in 0.2 ml of water. A 0.1-ml sample was used to inoculate 5 ml of nonselective media (either YPD or SD plus uracil), and cultures were grown at 30°C with aeration for 5-10 generations. Dilutions of the initial suspension were plated on YPD plates, and colonies were counted to determine the initial concentration of cells. These plates were then replica-plated to SD-ura to determine the percentage of plasmid-bearing cells.
PFGE Labeling-The PFGE labeling method was used to investigate yeast chromosomal DNA synthesis (39). In this method, yeast cells are first labeled with 32 P i in vivo and chromosomal DNA is then resolved by pulsed field gel electrophoresis. Briefly, 20 ml of yeast cell culture was grown at room temperature to the early log phase (ϳ10 7 cells/ml) and arrested by ␣-factor for 90 min. The cells were then washed with 20 ml of low phosphate medium (LPM) three times, and resuspended into the same volume of LPM. After heat treatment (37°C), 50 Ci of radioactive 32 P i was added. The culture was grown for another 1 h, and uptake was quenched with 50 mM cold phosphate. The labeled cells were harvested and then washed with phosphate-buffered saline buffer, and molecules were separated using the Bio-Rad CHEF-DR II Megabase DNA pulse field electrophoresis system. The resulting gel was dried and autoradiographed.
Preparation of Synchronous Chromatin Fractions-Yeast chromatin was prepared according to Lue and Kornberg (40) with slight modifications. Briefly, yeast strain K4055 cells harboring plasmid fusions with co-transformation of YEp-Mcm5-HA and YCp5N-Cdc6-T7, or YCp5N-cdc6(K114E)-T7. The transformant cells were grown on the CSM medium in the presence of methionine (MET ϩ ) to shut off the endogenous CDC6. The cells were synchronized at G 1 with ␣-factor at final concentration of 10 g/ml at room temperature for 4 h. The cells were then released from ␣-factor block, grown on the fresh MET ϩ medium, and collected at the indicated time points. Aliquots of cells were harvested at different time intervals to monitor the degree of synchrony by measurement of percentage of budded cells and FACS analysis (38). Yeast cells were washed once with 50 mM ice-cold EDTA and incubated in a solution containing 20 mM EDTA/2% ␤-mercaptoethanol at 30°C for 30 min. After incubation, the cells were centrifuged and washed once with 1 M sorbitol. Spheroplasts were generated by digesting the cells with 100 g/ml yeast lytic enzyme (ICN) in a solution containing 1 M sorbitol, 5 mM ␤-mercaptoethanol at 37°C for 60 min. The spheroplasts were cooled on ice for 10 min, centrifuged at 3000 ϫ g for 10 min at 4°C, and washed once with the 1 M ice-cold sorbitol. The spheroplasts were collected by centrifugation and lysed in a buffer containing 20% (w/v) Ficoll 400 (Sigma), 20 mM HEPES-KOH, pH 7.5, 20 mM KCl, 5 mM MgCl 2 , 3 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM EDTA with a Teflon/glass homogenizer. Cell debris were removed by four spins at 3000 ϫ g for 5 min at 4°C. Uniform white supernatants were recovered and were spun at 36,000 rpm and 4°C for 30 min. Nuclear pellets were resuspended in 500 l of the spheroplast lysis buffer described above and subjected to Ficoll 400 gradient centrifugation (20 -45%) at 36,000 rpm and 4°C for 30 min. Chromatin pellets were resuspended in a buffer H/0.1 (50 mM HEPES-KOH, pH 7.5, 100 mM KCl, 5 mM magnesium acetate, 1 mM EDTA, 0.02% Nonidet P-40, 1 mM dithiothreitol, 1 mM PMSF, and 10% glycerol) and stored at Ϫ80°C.
Western Blots to Detect Mcm5 and Cdc6 Proteins-10 g of protein from each chromatin sample were mixed with 4ϫ SDS loading buffer and boiled in a water bath for 5 min. The samples were resolved by SDS-polyacrylamide gel (8 -10%) electrophoresis at 4°C for 4 h, and proteins were blotted onto an Immobilon-P polyvinylidene difluoride transfer membrane (Millipore). The protein blots were probed either with 12CA5 monoclonal antibody (Boehringer Mannheim) for detecting HA-tagged proteins, or with anti-T7 tag monoclonal antibody (Novagen) for detecting T7-tagged proteins, followed by alkaline phosphataseconjugated secondary antibody. The protein samples were detected with the Tropix chemiluminescent system.
In Vitro and in Vivo Interaction between Orc1 and Cdc6 -500 ml of bacterial culture containing expressed 6xHis-T7-Orc1 tagged protein was harvested and resuspended into 15 ml of Buffer I (20 mM Tris-HCl, pH 8, 500 mM NaCl, 20 mM imidazole, 0.1% Triton X-100) in the presence of 20 mM PMSF and 20 mM benzamidine. 10 mg of lysozyme and 2 mg of DNase were added and incubated at 4°C for 30 min. The mixture was centrifuged at 13,000 rpm for 20 min. The supernatant was collected (ϳ15 ml) and mixed with 0.2 ml of Ni-NTA (Qiagene) at 4°C for 60 min. The resulting slurry was packed onto a column and washed with 20 ml of Buffer II (20 mM Tris-HCl, pH 8, 150 mM NaCl, 50 mM imidazole, 0.1% Triton X-100). The Orc1-charged matrix was ready for Cdc6 interaction assay. GST-Cdc6 fusions were prepared from 250 ml of culture. Crude extracts were made in Buffer II with 20 mM PMSF and 20 mM benzamidine. Aliquot was used for SDS-polyacrylamide gel electrophoresis to analyze their expression (Fig. 7B, lanes 4 -6). The rest (ϳ15 ml) was passed through Orc1-charged matrix 10 times for the interaction studies.

RESULTS
Site-directed Mutagenesis of the Nucleotide-binding Site of the CDC6 Gene-In vitro site-specific mutagenesis is a powerful tool to probe the relationship between the structure and activity of proteins, because the amino acid residues responsible for a particular function can be identified directly (41). Previous analysis of the primary sequence (34) has shown that CDC6 contains a motif conserved among known ATPases, GXXGXGKT, ranging from residues 108 -115 in CDC6 (Fig. 1A). This element plays a role in binding the pyrophosphate moiety of nucleotides, and the Lys residue is essential for electrostatic interaction between the proteins and the nucleotides (42). In order to examine whether the conserved Lys residue represents an essential amino acid residue as in other nucleotide-binding proteins, we have mutated the Lys residue to five other residues. The mutagenic primers were similar to the wild-type sequence except for one or two altered nucleotides at the first two bases in the lysine codon (Fig. 1, A and B). The following substitutions were made at the conserved lysine: to glutamate to test the effect of introducing a negative charge, to proline to test the effect of a general disruption, to glutamine to test the effect of substituting an amide, to leucine to test the effect of introducing a neutral amino acid, and to arginine to test the effect of changing the configuration of the positive charge. Although both Lys and Arg are basic amino acids, none of the nucleotide-binding sites known so far contain Arg. Substitution of arginine for lysine results in a protein with partial function in some cases. For example, this substitution in yeast RAD3 was found to abolish its ATPase and DNA helicase activities, but not to change the ability to bind ATP (43). Mutagenesis was confirmed by DNA sequencing (Fig. 1C).
Functional Complementation of cdc6 -1 Mutant Cells-To test the in vivo function of the nucleotide-binding site, we examined whether the mutant genes can complement the cdc6 -1 temperature-sensitive growth defect. We transformed the cdc6 -1 mutant strain with plasmids expressing the mutant genes. Transformants were selected at the nonpermissive temperature. As shown in Table I, the wild-type and K114L, K114Q, and K114R are able to complement the defective mutant cells at nonpermissive temperature. Mutant K114E, K114P, and vector alone, however, fail to transform. The few colonies observed for K114E and K114P are presumably due to the recombination between the episomal and chromosomal copies. These results imply that the Lys to Glu mutation may The mutagenic primer is identical to the wild-type CDC6 DNA sequence depicted above except that the nucleotide A is changed to G, C, and/or T, resulting in the Lys residue altering to other amino acid residues. B, alternations of the codon and the amino acid residue are summarized. C, the DNA sequences of mutagenized cdc6 genes are shown on the bottom.
perturb the interaction between Cdc6 and nucleotides; the Lys to Pro mutation may alter steric structure resulting in the loss of CDC6 gene function. Thus, the result supports that the Lys residue is essential for CDC6 gene function.
A trivial explanation of the above findings is that the mutant proteins are not efficiently present. To ensure that failure of the mutant alleles to complement cdc6 -1 was due to a defective Cdc6 protein, and not merely due to absence of protein, we compared the steady-state levels of wild-type Cdc6 and the cdc6(K114E) proteins. The cdc6(K114E) protein was chosen, because it had the most drastic defect in the complementation assay. Since Cdc6 is present at very low levels in yeast and is barely detectable with the polyclonal Cdc6 antibody, we cloned the wild-type and mutant alleles under the control of the GAL1,10 promoter and introduced this construct into yeast vectors marked with URA3 or LEU2 (YGp12 and YGp102, respectively). Protein blots of extracts of cells carrying the plasmids were probed with Cdc6 antibody. As shown in Fig. 2, both mutant and wild-type proteins were expressed at the same level, demonstrating that expression levels were not the cause of the lack of in vivo function of cdc6(K114E) protein.
Effects on the Growth Rate from cdc6 -1 Transformant Cells-To further evaluate the effect of mutant alleles in the cdc6 -1 background, cdc6 -1 was transformed with the mutant plasmids, the wild-type plasmid and vector alone, and transformants were selected at room temperature. Transformant cells were grown to log phase at the permissive temperature (25°C), then transferred to the nonpermissive temperature (37°C) and incubated for another 8 h, during which growth rates were determined (Fig. 3A). Cells harboring the wild-type plasmid grew at a rate similar to wild-type strains in this medium (t d ϭ 120 min). For mutants K114E, K114P, and vector alone, cells essentially ceased to divide after 2 h and totally arrested at 37°C. The remaining strains grew more slowly than wild-type, with doubling times of 125-130 min for K114R, and 150 -160 min for K114L and K114Q at 37°C. Thus, all of the mutants generated show at least some impairment in growth rate, indicating that an intact lysine residue is critical for optimal CDC6 function. Similar results can be observed from the plates incubated at 25°C (permissive temperature) and 37°C (nonpermissive temperature). No colonies were formed in the K114E and K114P transformants (Fig. 3B). As the growth curves of the K114E and K114P transformants are similar to cdc6 -1 cells transformed with vector only, it is clear that the K114E and K114P are loss-of-function proteins that cannot rescue the cdc6 -1 mutation.
Plasmid Stabilities in cdc6 -1 Cells Expressing Mutant cdc6 Genes-Temperature-sensitive mutation cdc6 -1 shows enhanced plasmid loss at the nonpermissive temperature, but plasmids may be stabilized by the inclusion of a functional copy of CDC6 on the plasmid. We tested the above mutated alleles of cdc6 for their ability to confer stability on plasmids. As shown in Fig. 4, cdc6 -1 cells bearing a plasmid with the wild-type CDC6 gene maintains ϳ85% of plasmid after 10 generations under nonselective conditions, a loss rate similar to wild-type backgrounds. The K114R plasmid confers slightly lower stability (ϳ70%), and K114L and K114Q plasmids result in 45-50% plasmid-containing cells after 10 generations. K114E and K114P transformants, however, show dramatic plasmid loss, similar to that of YEp352 transformants. The loss of plasmid stability is most likely due to the defect(s) of DNA synthesis. To explore what mechanism affects the growth rate and plasmid DNA stability, we investigated the chromosomal DNA synthesis.
Yeast Chromosomal DNA Synthesis in cdc6 -1 and Transformant Cells-The PFGE labeling technique is a novel method, which is highly specific to chromosomal DNA synthesis in yeast. In this method, yeast cells are first labeled by 32 P in vivo and chromosomal DNA molecules are then resolved by pulsed field gel electrophoresis. DNA synthesis can be measured by the 32 P incorporation into individual chromosomal DNA molecules. We have previously used the PFGE labeling technique to verify that cdc6 -1 cells fail to synthesize chromosomal DNA molecules (39). In Fig. 5, we examined DNA synthesis by the PFGE method in cdc6 -1 cells transformed with different cdc6 Lys 114 mutants. The cultures were synchronized with ␣-factor (at START), 32 P label was added, and cells were incubated at the non-permissive temperature. PFGE was used to resolve the labeled chromosomal DNA molecules. Cells transformed with wild-type CDC6 and vector YEp352 only, are used as the positive and negative controls, respectively. The DNA pattern (Fig.  5A, lanes 1 and 7) and 32 P incorporation (Fig. 5B, lanes 1 and  7) for the controls are as expected. In the K114E transformant (lane 6) and K114P transformant (lane 4), ethidium bromide staining revealed a similar chromosomal pattern, yet little radioisotope incorporation into chromosomal DNA was observed. However, in cells transformed with K114R, K114Q, and K114L plasmids, a nearly wild-type amount of 32 P incorporation was observed at the nonpermissive temperature (lanes 2, 3, and 5). Taken together (Figs. 3-5), we conclude that the Vector only 4.0 ϫ 10 3 0

FIG. 2. Expression of mutant and wild-type Cdc6 proteins in yeast.
On the left margin, the protein markers from Life Technologies, Inc. are indicated: myosin (205,000), phosphorylase b (97,400), bovine serum albumin (68,000), ovalbumin (43,000), carbonic anhydrase (29,600), and lysozyme (18,000). CDC6 was expressed under the control of the GAL1,10 promoter as described under "Experimental Procedures." We created strains carrying either one or two copies of the wild-type or mutant allele by transforming strain BJ2168 with YGp12 alone or YGp12 and YGp102 constructs together. Cells were grown in raffinose, and then 2% galactose was added for 3 h; cells were then harvested. Fifty g of each extract was electrophoresed and blotted, and the blot was probed with polyclonal Cdc6 antibody described previously. K114E and K114P are loss-of-function mutations, K114Q and K114L are partially functional proteins, and K114R functions equivalently to wild-type. Since the only fully functional mutant represents a conservative change in the conserved ATPbinding site, we argue that a functional ATP-binding site in CDC6 is required for DNA synthesis in vivo.
Effect of Mutant K114E on the Assembly of the Replicative Complex-It has been shown that the Cdc6 is essential for establishment and maintenance of pre-replicative complexes (pre-RCs) and the loading of MCM proteins at G 1 phase on chromatin is Cdc6-dependent. Since the initiation of DNA replication requires the prior assembly of pre-RCs, we further tested the hypothesis that the mutation of ATP-binding domain of Cdc6 would lead to the failure of the pre-RC assembly. We selected the mutant cdc6(K114E) for these studies, because it showed the most drastic effect on the CDC6 gene functions. To do this, we transformed wild-type CDC6 and mutant cdc6(K114E) genes with its native promoter bearing on the single copy vector into strain K4055, respectively. In this K4055 strain, the chromosomal CDC6 gene is under the control of the MET3 repressive promoter. In the absence of methionine in the medium, the chromosomal wild-type CDC6 gene is expressed to support the cell growth. In the presence of methionine, the chromosomal wild-type CDC6 gene is shut-off and endogenous Cdc6 protein is rapidly degraded, resulting in a null mutant phenotype. In the cdc6(K114E) transformant, the wild-type protein is basically replaced by the mutant one. The impaired function of the mutant cdc6 protein, if any, can be examined. We have investigated the temporal and spatial regulation of the Cdc6 nucleotide-binding function(s) as shown in Fig. 6.
We determined the association of the cdc6(K114E) mutant protein with chromatin and the loading of Mcm5 onto chromatin origin in chromatin-binding assay. Chromatin was prepared from the yeast cell released from G 1 -phase block at the indicated time points and probed either with anti-T7 tag (for detecting the association of Cdc6 with chromatin) or with monoclonal antibody 12CA5 (for detecting HA-tagged Mcm5 associated with chromatin). In the presence of methionine, the endogenous Cdc6 was shut off, so growth of the cell was totally dependent on Cdc6 produced from the plasmids harboring in the cell. The cultures were synchronized by the ␣-factor arrestand-release method. Based on the percent of budded cells and FACS analysis, the peak of the S phase is around ϳ60 min (Fig.  6A, upper panel) in the wild-type cells. Orc2, which was used as an internal control showed a constant level throughout the cell cycle (data not shown). The association of wild-type Cdc6 with chromatin was in a cell cycle-dependent manner, i.e. Cdc6 was detected when the cell was released from G 1 phase block. It remained on chromatin during G 1 phase, decreased at the start of S phase and disappeared from chromatin during G 2 and M phases, and reappeared after M phase (Fig. 6A, bottom panel). The association of Mcm5 with chromatin showed a quite similar pattern to that of the Cdc6 protein. i.e. Mcm5 was loaded onto chromatin during G 1 phase, replaced from chromatin during S phase and was not detected during G 2 and M phases. The association of cdc6(K114E) mutant proteins with chromatin, however, was dramatically different from that the wild-type Cdc6 (Fig. 6B). The cdc6(K114E) mutated proteins can associate with chromatin around 20 min, but then faded quickly from the chromatin. Under the same condition, the association of Mcm5 with chromatin was determined in the presence of the mutant cdc6(K114E) proteins. The loading of the Mcm5 protein onto chromatin was delayed for at least 60 min after the cell was released from the G 1 phase block (Fig. 6B, lower panel). There was no second time associating or re-loading for Cdc6 and Mcm5, suggesting that the cell cycle was arrested and the cell ceased to grow. Consistent with these observations, atypical FACS pattern in the mutant transformants was observed. Presumably, this is due to defective DNA synthesis and/or abnormal morphology. Thus, the data suggested the improper chromatin assembly and the perturbation of the cell cycle progression in the cdc6(K114E) background.
Functional Orc1 and Cdc6 Interaction Depends on the Normal ATP-binding Site of Cdc6 -During the process of Cdc6 purification, we consistently observed the co-fractionation of Orc1 with Cdc6. In the glycerol gradient centrifugation, both Orc1 and Cdc6 proteins co-sediment in a large 450 -660-kDa protein complex. 2 Due to their structural similarity, it has been proposed that the Cdc6 and Orc1 may be in a single gene family (44). This information led us to investigate the relationship between Cdc6 and Orc1 in vitro and in vivo.
In the in vitro interaction studies, bacterially expressed Orc1-His-tagged protein retained on the nickel-chelating matrix was used as a protein affinity column. The same vector without the Orc1 insert was used as negative control. Yeast extracts, as the Cdc6 protein source, were then passed through the Orc1-charged matrix. After extensive washing, the interacting components were analyzed by protein blots. The presence of Cdc6 protein can be detected by anti-Cdc6 polyclonal antibodies, indicating an interaction between Orc1 and Cdc6 proteins (Fig. 7A). To test whether this is a direct interaction between Cdc6 and Orc1, we used bacterially expressed wildtype GST-Cdc6 and mutant GST-cdc6 proteins for interaction studies. In the Escherichia coli background, one would assume that yeast adaptor protein(s), if any, can be avoided. In Fig. 7B  (lane 1), the wild-type GST-Cdc6 was able to interact with Orc1, suggesting a direct interaction. Intriguingly, there was no detectable GST-cdc6(K114E) band (Fig. 7B, lane 2). This result suggests that Orc1 could not interact with mutated cdc6(K114E) protein. To avoid artifact, we used another mutant protein GST-cdc6 -1 as control. The mutation point of the cdc6 -1 protein has been determined as Gly 260 to Glu muta-2 L. Feng and A. Y. Jong, unpublished data. tion. 2 This mutation leads to a temperature-sensitive mutation. However, there is no defect in Orc1/Cdc6 interaction (Fig.  7B, lane 3). The failure of Orc1/cdc6(K114E) interaction is not due to the absence of protein, as the same level of GST-Cdc6 (lane 4), cdc6(K114E) (lane 5), and cdc6 -1 (lane 6) were expressed and used in this experiment.
In in vivo interaction studies, 6xHis-T7-Orc1-tagged protein was expressed in yeast. This strain co-expressed wild-type Cdc6-T7 tag, mutant cdc6(K114E)-T7 tag, or cdc6 -1-T7 tag proteins, separately. The 6xHis-T7-Orc1-tagged protein was isolated by nickel-chelating matrix, then concentrated, and probed with anti-T7 monoclonal antibody. In Fig. 7C, there is an obvious co-precipitation between Orc1 and Cdc6 (lane 1), but the interaction is insignificant in our observation of Orc1 and cdc6(K114E) mutant protein (lane 2). Again, normal interaction between Orc1 and temperature-sensitive cdc6 -1 protein was observed (lane 3), suggesting that the temperature sensitivity of cdc6 -1 follows a defective mechanism different from the one in ATP-binding mutation. Taken together, the results suggest that ATP binding may induce or maintain a proper conformation of Cdc6, which is important for Orc1 interaction. The failure of the proper interaction may lead to the impairment of replicative complex assembly.

DISCUSSION
Cdc6 plays an essential role in the initiation of DNA replication. Little, however, is known about structure/function relationships in the CDC6 gene. All eukaryotic Cdc6 contains a conserved region known as Walker A-motif, GXXGXGKT (18). This element may play a role in binding the pyrophosphate moiety of nucleotides. The Lys residue is essential for electrostatic interaction between the proteins and the nucleotides. Secondary structural predictions show that the structure around residues 108 -115 in yeast Cdc6 are consistent with the model for an ATP-binding protein (data not shown). In order to examine whether the conserved Lys residue represents an essential amino acid residue as in other nucleotide-binding proteins, we have mutated the Lys residue to five other amino acids. In general, many proteins with the GXXGXGKT nucleotide-binding site consensus sequence possess ATPase activity, including the helicase, RAD3 (43). In other ATP-and GTP-binding proteins, this sequence has been suggested to mediate functional conformational change (45). It has recently been demonstrated that the bGST-Cdc6 can bind ATP and GTP in vitro. It has also been proposed that Cdc6 is active as an ATPase (10). The in vivo function(s) of nucleotide binding in Cdc6 is still unclear. Site-directed mutagenesis of the conserved Lys may be useful to delineate its biological function in vivo. Indeed, we generated five point-mutated cdc6 at Lys 114 of the nucleotide-binding site. We have used four approaches to explore the effect on the cell functions of various point-mutants of cdc6: (i) complementation of the temperature-sensitive strain, (ii) effect on the growth rate, (iii) determination of plasmid stability in the transformant mutants, and (iv) determination of DNA synthesis in these mutants. The PFGE labeling method to examine DNA synthesis in cdc6 -1 and transformant cells was particularly informative (Fig. 5). Since synthesis on individual chromosomes giving chromosomal size DNA products is observed directly, it may be that this technique creates the fewest experimental artifacts among all current available methods for yeast chromosomal DNA synthesis. Our data clearly demonstrated that a functional nucleotide binding activity is required for the CDC6 in DNA synthesis. Our results also demonstrate that the change of Lys to Glu (K114E) and to Pro (K114P) lead to loss-of-function in all of these respects; change of the Lys to Gln (K114Q) and to Leu (K114L) leads to partially functional proteins. Interestingly, Cdc6(K114R) shows a behavior equivalent to wild-type. No functional ATPase to date contains an Arg in place of Lys at the same position (46), implying that nucleotide hydrolysis may not be a major function of this domain in Cdc6. This leads to the model that nucleotide binding may be involved in conformational or allosteric changes in the protein, resulting in sequential events at chromatin in the G 1 /S transition of the cell cycle.
How do these observations reconcile with the in vivo function of initiation of DNA replication? The Cdc6 protein is required for establishment and maintenance of pre-RC, and the Cdc6 protein is also required for the loading of MCM proteins. In addition, the Cdc6 protein interacts with Cdc28 protein complex after p40 sic1 is degraded at late G 1 phase (31). The Cdc28 protein kinase may then activate the initiation of DNA repli- Trp1. 500 ml of culture was induced with 2% galactose for 4 h. The tagged Orc1 was purified by Ni-NTA matrix, concentrated 10-fold, and probed with anti-T7 antibody on the protein blots. The T7 monoclonal antibody can detect not only 6xHis-T7-Orc1 tag protein but also Cdc6-T7, if there is an interaction between Orc1 and Cdc6 proteins. The upper arrow indicates the predicted size of Orc1 and the lower arrow indicates the predicted size of Cdc6. cation (47). It is generally accepted that a stepwise assembly of the functional replication complex is required for the coordination of the DNA replication and cell cycle control (for review, see Ref. 48). It is well established that ATP is essential for prokaryotic DNA replication, but is much less understood in eukaryotic DNA replication. We have systematically investigated the temporal and spatial regulation of the wild-type CDC6 and mutated cdc6(K114E) gene function(s) to explore the possible involvement of ATP in this process (Fig. 6). This system is dynamic and useful for understanding the molecular events from late G 1 phase (a stable protein-origin complex) into the S phase (a mobile DNA machinery). Indeed, we demonstrated that the impaired function of cdc6(K114E) leads to an abnormal assembly of the replicative complex. At this stage, the preparation of replicative complexes was still quite crude, preventing us from drawing a conclusive picture. Nevertheless, two major findings allowed us to unravel the possible molecular mechanism action of ATP binding. First, we have demonstrated that there is a direct interaction between yeast Orc1 and Cdc6 proteins. The information is consistent with observations from many other laboratories regarding the structure and functions of the ORC-origin complex, and the interaction between Cdc6 and Orc1 is conserved in eukaryotes (50,51). Moreover, we demonstrated that the interaction between Orc1 and Cdc6 is defective in the cdc6(K114E) mutation. This observation is original. The simplest model is that the ATP-binding mutation fails to form Orc1/Cdc6 complex properly, which results in a deficiency in the formation of an operational prereplicative complex. Thus, the requirement of ATP for Cdc6/ Orc1 interaction is involved in a quite early stage of DNA replication. This possibility is supported by our chromatin studies (Fig. 6).
In summary, with the availability of these mutants, we have clearly demonstrated that the in vivo function of the conserved Lys 114 in the nucleotide-binding site of CDC6 is required for cell growth and DNA synthesis. We have also demonstrated that, at the molecular level, the mutant cdc6 at its nucleotide binding fails to interact with Orc1. The fact that the ATPbinding domain is highly conserved among Cdc6 and Orc1 protein families suggests their similar role in the initiation of DNA replication (18,19,47). Perhaps by sequentially contributing ATPase activity or ATP-governed conformational change to the ATP-dependent ORC/MCM functions result in the ordered events during initiation of replication (18,30,44). The highly conserved central domain of the MCM and RF-C family proteins also contains a predicted consensus motif for DNA-dependent ATPase (49). It is tempting to speculate that the MCM proteins of the pre-replicative complex, driven by the Cdc6 and Orc1 ATPase or ATP-governed conformational change, start to unwind the origin sequence during initiation of replication. ATP binding in the MCM and RF-C families may follow a similar mechanism. Further studies of the roles of ATP in Cdc6/Orc1, MCM, and RF-C families may reveal this complicated, yet extremely important regulation in DNA replication and cell cycle progression.