Dbf4-Cdc7 Phosphorylation of Mcm2 Is Required for Cell Growth*

The Dbf4-Cdc7 kinase (DDK) is required for the activation of the origins of replication, and DDK phosphorylates Mcm2 in vitro. We find that budding yeast Cdc7 alone exists in solution as a weakly active multimer. Dbf4 forms a likely heterodimer with Cdc7, and this species phosphorylates Mcm2 with substantially higher specific activity. Dbf4 alone binds tightly to Mcm2, whereas Cdc7 alone binds weakly to Mcm2, suggesting that Dbf4 recruits Cdc7 to phosphorylate Mcm2. DDK phosphorylates two serine residues of Mcm2 near the N terminus of the protein, Ser-164 and Ser-170. Expression of mcm2-S170A is lethal to yeast cells that lack endogenous MCM2 (mcm2Δ); however, this lethality is rescued in cells harboring the DDK bypass mutant mcm5-bob1. We conclude that DDK phosphorylation of Mcm2 is required for cell growth.

The Cdc7 protein kinase is required throughout the yeast S phase to activate origins (1,2). The S phase cyclin-dependent kinase also activates yeast origins of replication (3)(4)(5). It has been proposed that Dbf4 activates Cdc7 kinase in S phase, and that Dbf4 interaction with Cdc7 is essential for Cdc7 kinase activity (6). However, it is not known how Dbf4-Cdc7 (DDK) 2 acts during S phase to trigger the initiation of DNA replication. DDK has homologs in other eukaryotic species, and the role of Cdc7 in activation of replication origins during S phase may be conserved (7)(8)(9)(10).
The Mcm2-7 complex functions with Cdc45 and GINS to unwind DNA at a replication fork (11)(12)(13)(14)(15). A mutation of MCM5 (mcm5-bob1) bypasses the cellular requirements for DBF4 and CDC7 (16), suggesting a critical physiologic interaction between Dbf4-Cdc7 and Mcm proteins. DDK phosphorylates Mcm2 in vitro with proteins purified from budding yeast (17,18) or human cells (19). Furthermore, there are mutants of MCM2 that show synthetic lethality with DBF4 mutants (6,17), suggesting a biologically relevant interaction between DBF4 and MCM2. Nevertheless, the physiologic role of DDK phosphorylation of Mcm2 is a matter of dispute. In human cells, replacement of MCM2 DDK-phosphoacceptor residues with alanines inhibits DNA replication, suggesting that Dbf4-Cdc7 phosphorylation of Mcm2 in humans is important for DNA replication (20). In contrast, mutation of putative DDK phosphorylation sites at the N terminus of Schizosaccharomyces pombe Mcm2 results in viable cells, suggesting that phosphorylation of S. pombe Mcm2 by DDK is not critical for cell growth (10).
DDK phosphorylates the N termini of human Mcm2 (19,20,28), human Mcm4 (10), budding yeast Mcm4 (26), and fission yeast Mcm6 (10). Although the sequences of the Mcm N termini are poorly conserved, the DDK sites identified in each study have neighboring acidic residues. The residues of budding yeast Mcm2 that are phosphorylated by DDK have not yet been identified.
In this study, we find that budding yeast Cdc7 is weakly active as a multimer in phosphorylating Mcm2. However, a low molecular weight form of Dbf4-Cdc7, likely a heterodimer, has a higher specific activity for phosphorylation of Mcm2 Size Exclusion Chromatography-The Superose 6 size exclusion column was pre-equilibrated in gel filtration solution containing 25 mM Tris-HCl, pH 7.5, 10% glycerol, 1 mM DTT, 0.1 mM EDTA, and 100 mM NaCl. 200 g of Cdc7 alone or 200 g of Cdc7 alone with 50 g of Dbf4 were first incubated for 30 min at 15°C. The samples were then subjected to Superose 6 size exclusion chromatography in gel filtration solution. Each fraction was then subjected to SDS-PAGE analysis followed by Coomassie staining. 7 l of each column fraction was incubated with 500 ng of full-length Mcm2 and 10 Ci of [␥-32 P]ATP in a final volume of 10 l for 30 min at 30°C, and the reactions were analyzed by SDS-PAGE followed by phosphorimaging.
DDK Analytical Kinase Assays-Kinase reactions were in a volume of 10 l and contained 5 mM Tris-HCl, pH 8.5, 10 mM MgCl 2 , 1 mM DTT, 50 M cold ATP, 10 Ci of [␥-32 P]ATP, and 1.5 g of DDK, as indicated for each reaction. The amount of Mcm2 in each reaction is described in the figure legends. Reactions were incubated at 30°C for 1 h. Reactions were stopped by addition of 5 l of 5ϫ SDS sample buffer (25% glycerol, 200 mM Tris base, 125 mM DTT, 2.5% SDS, 0.1% bromphenol blue), and the products were resolved on either 10 or 15% SDS-PAGE.
GST Pulldown Assay-Mcm proteins were first radiolabeled in a reaction volume of 100 l that contained 20 M full-length or fragment Mcm2 protein in kinase reaction buffer (5 mM Tris-HCl, pH 8.5, 10 mM MgCl 2 , 1 mM DTT, 500 M ATP, 100 Ci of [␥-32 P]ATP) containing 5 g of PKA. Reactions were incubated at 30°C for 1 h.
The 100-l GST-pulldown reaction contained 50 pmol of GSTtagged protein in GST-binding buffer (40 mM Tris-HCl, pH 7.5, 100 mM NaCl, 0.1 mM EDTA, 10% glycerol, 0.1% Triton X-100, 1 mM DTT, 0.7 g/ml pepstatin, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mg/ml bovine serum albumin) and varying amounts of radiolabeled protein as described in each figure. Reactions were incubated at room temperature for 1 h. Following incubation, reactions were added to 40 l of prepared glutathione-Sepharose and gently mixed. Binding of GSTtagged protein to the beads was performed for 20 min, with gentle mixing every few minutes. Once the binding was complete, the reaction mixture was aspirated, and the beads were washed two times with 0.5 ml of GST binding buffer. After the last wash, 30 l of 5ϫ SDS sample buffer were added to each reaction, and the samples were boiled for 10 min. Samples (20 l) were then analyzed by SDS-PAGE.
Services-DNA sequencing and mass spectroscopy were performed by the Vanderbilt University Facilities.

Low Molecular Weight Form of Dbf4-Cdc7
Exhibits High Specific Activity for Mcm2-The molecular architecture of the active Dbf4-Cdc7 kinase is currently not known. To determine the functionally active state of Cdc7, the protein was analyzed by size exclusion chromatography (Fig. 1A). Cdc7 eluted in a peak at a Stokes radius that corresponded to that of a 232-kDa standard (Fig. 1A, top gel). The predicted molecular mass of Cdc7 is 58 kDa, suggesting that the majority of Cdc7 is in a tetrameric state. The elution of Cdc7 in neighboring fractions suggests that Cdc7 may be a mixture of tetramer and other forms. These results are consistent with in vivo data that Cdc7 alone exists as a multimer in the cell (27).
To test if Cdc7 alone is active as a kinase at Mcm2, each fraction from the size exclusion column was incubated with To determine the functionally active state of Dbf4-Cdc7, Dbf4 was incubated with excess Cdc7, and the protein mixture was then subjected to size exclusion chromatography. Once again, a peak of Cdc7 alone was present at a Stokes radius that corresponded to a 232-kDa protein. However, an additional low abundance peak of Cdc7 and Dbf4 is present in a fraction that peaks at a Stokes radius that is between 68 and 158 kDa (Fig. 1B, top gel). The molecular mass of one subunit of Dbf4 plus one subunit of Cdc7 is 139 kDa, suggesting that this complex is a heterodimer of Dbf4-Cdc7. Dbf4 was not observed in the higher molecular weight fractions, suggesting that Dbf4 does not bind to multimeric Cdc7. The elution fractions were then analyzed for phosphorylation of Mcm2, and peak phosphorylation levels were observed at a molecular mass between 68 and 158 kDa ( Fig. 1, B, bottom gel, and C and D, filled triangles). Low levels of Cdc7 are present in the peak activity fractions, suggesting that a heterodimer of Dbf4-Cdc7 has a substantially higher specific activity for phosphorylating Mcm2 compared with the Cdc7 multimer (100-fold difference in specific activity, see Fig. 1D). Taken together, the data of Fig. 1 suggest that Cdc7 forms a multimer that is weakly active in phosphorylating Mcm2. Moreover, Dbf4 can bind to and activate a monomer of Cdc7, and the Dbf4-Cdc7 heterodimer has substantially higher specific activity than Cdc7 alone (100-fold higher specific activity).
Dbf4 , and this sequence was added solely for the purpose of accurate quantitation. We mixed PKA-labeled fulllength Mcm2 with GST-DDK, GST-Dbf4, or GST-Cdc7 and then isolated complexes bearing GST with glutathione beads (GST pulldown, Fig. 2A). When 6.6 pmol (66 nM), 20 pmol (200 nM), or 66 pmol (660 nM) of Mcm2 was added to the 100-l binding reactions, more than half of the Mcm2 bound to GST-Dbf4-Cdc7 (GST-DDK) or GST-Dbf4 proteins (Fig. 2B). Thus, GST-DDK or GST-Dbf4 binds tightly to full-length Mcm2. In contrast, GST-Cdc7 binds weakly to Mcm2, because Mcm2 binding levels were only slightly greater than those achieved with GST alone (Fig. 2, A and B). The weak binding of Cdc7 to Mcm2 is consistent with the size exclusion data, which demonstrate that high concentrations of Cdc7 alone phosphorylate Mcm2 to very low levels (Fig. 1A).
We next determined whether the Mcm2-7 complex also binds to DDK, Dbf4, or Cdc7, because the Mcm2-7 complex provides the motor function for the replication fork helicase (29 -31). We reconstituted Mcm2-7 complexes as described previously (31), except the Mcm3 subunit in this complex bears a PKA site. Using the same GST-pulldown approach, we found that the PKA-labeled Mcm2-7 complex binds tightly to Dbf4 or DDK but weakly to Cdc7 (Fig. 2, A and C). DDK phosphorylates The data in Fig. 2 suggest that Dbf4 does not activate Cdc7 that is pre-bound to Mcm2. It is more likely that Dbf4 binds to Mcm2 and then recruits Cdc7 to activate phosphorylation of Mcm2. It is also possible the Dbf4 first binds to Cdc7, and the DDK complex then binds to and phosphorylates Mcm2.  (Fig. 4, A and B). tion is plotted as a function of Mcm2 fragment concentration, the importance of amino acids 161-182 is evident (Fig. 4B). The plot also reveals that at low concentrations of Mcm2, region 204 -278 is important for maximal DDK phosphorylation of Mcm2.
We next tested the ability of DDK to bind to Mcm2 fragments to identify the region for DDK docking on Mcm2. GST-DDK pulldown of PKA-radiolabeled Mcm2 fragments was accomplished, and the quantity of Mcm2 fragment bound as a function of input amount was plotted (Fig. 4C). A slight increase in binding is detected as Mcm2 fragment length is increased from 1-160 to 1-182, suggesting that residues 161-182 bind weakly to DDK. A substantial increase in binding is observed as the Mcm2 fragment length is increased from 1-203 to 1-278. These data suggest that region 204 -278 binds tightly to DDK.  4D). In contrast, DDK was completely inactive in phosphorylating Mcm2-(183-278) (Fig. 4D). However, DDK was roughly equally efficient in binding to each of these Mcm2 fragments (Fig. 4E). Therefore, the substantial difference in phosphoryla- To determine whether Ser-164 and Ser-170 are phosphorylated by DDK, these residues were mutated to alanine, and DDK phosphorylation and binding were examined. S614A exhibited a slight defect in DDK phosphorylation compared with wild-type Mcm2-(1-278), whereas S170A was modestly defective in DDK phosphorylation (Fig. 5, B and C). The double mutant S164A/ S170A was not phosphorylated by DDK at all (Fig. 5, B and C).
Sequence alignment between budding yeast Mcm2 and budding yeast Mcm4 reveals that the DDK phosphorylation sites on Mcm2 are in the same region as the DDK phosphorylation sites on Mcm4 (Fig. 5E). In budding yeast Mcm4, seven serine and threonine residues between residues 161 and 178 are likely phosphoacceptor sites for DDK (26). In budding yeast Mcm2, the two serine residues between residues 161 and 178 are phosphorylated by DDK. Thus, although the number of residues    (Fig. 5F). Although there is some disagreement as to the particular residues of human Mcm2 that are phosphorylated by DDK (10,19,20), in every published study the human DDK phosphorylation sites do not align with those found here for budding yeast Mcm2.

DDK Phosphorylation of Ser-164 of Mcm2 Is Not Required for Budding Yeast Cell
Growth-To determine whether DDK phosphorylation of Ser-164 of Mcm2 is required for cell growth, we performed plasmid-shuffle assays in budding yeast cells lacking endogenous MCM2 (mcm2⌬) (Fig. 6). The deletion is complemented by wild-type MCM2 under expression control of its native promoter on a low copy centromere plasmid bearing a URA3 marker. The cells also contain a plasmid that expresses either wildtype of mutant MCM2 under control of the GAL1, galactose-inducible promoter. As expected, the cultures grow equivalently on media containing galactose (Fig. 6A, right). Growth on media containing 5-FOA and galactose selects for cells that have lost the URA3 complementing plasmid and are expressing only the MCM2 allele (or empty vector) indicated to the left of Fig. 6A. As seen in 10-fold serial dilutions, the mcm2 S164A allele complements nearly as well as wild type in this background, a result confirmed by measuring the growth rates of cells taken from the 5-FOA plates and grown in media containing galactose (Fig. 6B,  top graph). These results suggest that DDK phosphorylation of S164A is not required for yeast cell growth.

DDK Phosphorylation of Ser-170 of Mcm2 Is Required for Budding
Yeast Cell Growth-A similar experimental strategy was then used to determine whether DDK phosphorylation of Ser-170 of Mcm2 is required for yeast cell growth. When mcm2⌬ cells were grown in the presence of galactose and 5-FOA, cells expressing mcm2 S170A did not grow at all at any dilution, in marked contrast to the growth of cells expressing mcm2 S164A or wild-type MCM2 (Fig. 6A, top  left panel). These data suggest that Ser-170 of Mcm2 is required for yeast cell growth.
Serine 170 of Mcm2 is a target of DDK phosphorylation, but it may also be required for an additional unknown function. To determine whether the cell death caused by expression of mcm2 S170A is related to DDK function, the plasmid-shuffle assay was performed in cells harboring the genetic bypass for DDK function, mcm5-bob1. These cells are also deleted for CDC7 (cdc7⌬). When mcm2⌬,mcm5-bob1,cdc7⌬ cells were  (10,19,20). grown in the presence of galactose and 5-FOA, cells expressing mcm2 S170A grew on agar plates in serial 10-fold dilutions, whereas cells harboring the vector alone were dead at every dilution (Fig. 6A, bottom left panel). Thus, the mcm5-bob1 DDK bypass rescues the cellular defect caused by the expression of mcm2 S170A , suggesting that the lethality of mcm2 S170A is primarily due to the absence of DDK phosphorylation of Ser-170 of Mcm2. It is not clear why mcm5-bob1 cells expressing mcm2 S170A required higher concentrations to grow on agar plates compared with cells expressing wild-type MCM2, but it may be that in the mcm2⌬,mcm5-bob1,cdc7⌬ strain, MCM2-Ser-170 is involved in a Cdc7-independent function. This is the first study of an mcm mutant with a lethal phenotype that is rescued by mcm5-bob1. mcm2⌬,mcm5-bob1,cdc7⌬ yeast colonies grown in galactose and 5-FOA were then grown in galactose-containing liquid media, and the rate of growth was determined (Fig. 6B, bottom  graph). mcm5-bob1 cells expressing mcm2 S164A grew, although the rate of growth was slightly decreased compared with cells expressing wild-type MCM2 for unknown reasons. Cells expressing mcm2 S170A grew in liquid media, albeit slightly less fast than cells expressing wild-type MCM2, suggesting that the lethal phenotype of mcm2 S170A in ⌬mcm2 cells is related to DDK phosphorylation of Ser-170 (Fig. 6B, bottom graph). Taken together, the data strongly support that DDK phosphorylation of Ser-170 of Mcm2 is required for budding yeast cell growth.

DISCUSSION
In this study we find that Cdc7 exists in solution primarily as a multimer, and this species at high concentrations is weakly active in phosphorylating Mcm2. Dbf4 binds to a low molecular weight form of Cdc7, forming what might be a heterodimer of Dbf4-Cdc7, and this species phosphorylates Mcm2 with substantially higher specific activity. Dbf4 alone binds tightly to Mcm2, whereas Cdc7 alone binds weakly to Mcm2, suggesting that Dbf4 recruits Cdc7 to phosphorylate Mcm2. DDK phosphorylates Mcm2-(1-278) but not Mcm2-(279-C). DDK phosphorylates Ser-164 and Ser-170 of budding yeast Mcm2, and these two serine residues are positioned in a highly acidic region. Although expression of mcm2 S164A supports cell growth, expression of mcm2 S170A does not. However, the lethality of mcm2 S170A is rescued by the DDK-bypass mutation mcm5-bob1. We conclude that DDK phosphorylation of Ser-170 of Mcm2 is required for normal cell growth.
Dbf4 May Recruit a Monomer of Cdc7 to Target Mcm2 for Phosphorylation-In this study we report that Cdc7 alone is a multimer that is weakly active in phosphorylating Mcm2. Cdc7 is present at high levels during G 1 and S phase, whereas Dbf4 levels peak in S phase (18). It is unlikely that Cdc7 activity during G 1 is physiologically relevant, given the low levels of Mcm2 phosphorylation observed in this study, and no activity was reported in other investigations (7,8). In contrast, we report that Dbf4-Cdc7 is stable as a low molecular weight form, likely a heterodimer, and this species phosphorylates Mcm2 to substantially higher levels than Cdc7 alone. Thus, during S phase, Dbf4 may bind to a monomer of Cdc7, thereby forming a highly active heterodimer. The high molecular weight form of Cdc7 may be less active than the monomeric form because the high molecular weight form does not efficiently bind to Dbf4 or Mcm2. If this model is correct, multimerization of Cdc7 may be an auto-inhibitory mechanism to prevent high level Cdc7phosphorylation of Mcm2 in the absence of Dbf4 (Fig. 7A).
We also find that Dbf4 and DDK bind tightly to Mcm2 and the Mcm2-7 complex but that Cdc7 alone does not bind tightly to these Mcm proteins. Thus, Dbf4 may recruit Cdc7 to Mcm2 and to the Mcm2-7 complex (Fig. 7B). Our in vitro data are consistent with a previously published in vivo report suggesting that Dbf4 recruits Cdc7 to origins of replication (23), and a recent study suggesting that Dbf4 recruits Cdc7 to Mcm2-7 complexes that are tightly bound to origins (24). or deleted for mcm2 and cdc7 and bearing the mcm5-bob1 mutation (mcm2⌬, mcm5-bob1, and cdc7⌬, bottom plates) were used. The cells harbor wild-type MCM2 on a plasmid with a URA3 selectable marker and mutant mcm2 on a LEU2 plasmid with expression controlled by a galactose-inducible promoter.
In the presence of galactose and 5-FOA, only mcm2 under the control of the galactose-inducible promoter is expressed (left plates). In the presence of galactose (right plates), wild-type MCM2 is expressed as well. Cells were plated in serial 10-fold dilutions. B, colonies from A were grown in liquid media, and the rate of growth was measured as function of time. The top graph is for mcm2⌬ cells, and the bottom graph is for mcm2⌬,mcm5-bob1,cdc7⌬ cells.  (26). Furthermore, the DDK phosphorylation site region is conserved between these two proteins ( Fig. 4)  Essential Requirement for DDK Phosphorylation of Mcm2 Is Likely to Be Conserved from Yeast to Humans-In this study, the lethal phenotype of Mcm2 S170A is rescued by the mcm5-bob1 mutation, providing very strong evidence that DDK phosphorylation of Mcm2 is required for budding yeast cell growth. In a previous study of human cells, mutation of DDK phosphorylation sites to alanine resulted in inhibition of DNA replication, whereas mutation of these sites to glutamate did not (20). Thus, data from yeast and human support the notion that DDK phosphorylation of Mcm2 is essential for DNA replication and cell growth. These results suggest that the function of DDK at Mcm2 is conserved from yeast to human. Interestingly, the DDK phosphorylation sites of human Mcm2 do not align with the kinase sites of yeast Mcm2. We speculate that although the function of DDK phosphorylation of Mcm2 is conserved from yeast to human, the mechanistic details may be slightly different between these species.  (10,26). Thus, DDK phosphorylation of Mcm proteins likely influences the binding of key helicase accessory proteins such as Cdc45. However, the relationship between DDK phosphorylation of Mcm4 and Cdc45 binding is complex, because overexpression of Cdc45 is lethal to strains expressing mutations of mcm4 deficient in DDK phosphorylation (26). Thus, phosphorylation of Mcms by DDK does not appear to directly stimulate the interaction between Mcms and Cdc45. In this study, we report that DDK phosphorylation of Mcm2 can be suppressed by a mutation in mcm5 (mcm5-bob1). Changes in the conformation of Mcm5 as a result of the mcm5-bob1 mutation have recently been reported (35). This observation suggests that phosphorylation of Ser 170 of Mcm2 may be partially mimicked by a conformational change in a different Mcm subunit (Mcm5) (Fig. 7D). Interestingly, Mcm5 is positioned adjacent to Mcm2 in the Mcm2-7 complex (31,36). We propose that DDK phosphorylation of Ser-170 of Mcm2 results in a conformational change in the Mcm2-7 complex that allows for the proper positioning of accessory proteins, such as Cdc45 and GINS, which ultimately leads to the activation of the replication fork helicase (Fig. 7E).