Different Phenotypes in Vivo Are Associated With ATPase Motif Mutations in Schizosaccharomyces pombe Minichromosome Maintenance Proteins

Corresponding author: Susan L. Forsburg, The Salk Institute, 10010 N. Torrey Pines Rd., La Jolla, CA., forsburg{at}salk.edu (E-mail)

Communicating editor: P. RUSSELL

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The six conserved MCM proteins are essential for normal DNA replication. They share a central core of homology that contains sequences related to DNA-dependent and AAA⁺ ATPases. It has been suggested that the MCMs form a replicative helicase because a hexameric subcomplex formed by MCM4, -6, and -7 proteins has in vitro DNA helicase activity. To test whether ATPase and helicase activities are required for MCM protein function in vivo, we mutated conserved residues in the Walker A and Walker B motifs of MCM4, -6, and -7 and determined that equivalent mutations in these three proteins have different in vivo effects in fission yeast. Some mutations reported to abolish the in vitro helicase activity of the mouse MCM4/6/7 subcomplex do not affect the in vivo function of fission yeast MCM complex. Mutations of consensus CDK sites in Mcm4p and Mcm7p also have no phenotypic consequences. Co-immunoprecipitation analyses and in situ chromatin-binding experiments were used to study the ability of the mutant Mcm4ps to associate with the other MCMs, localize to the nucleus, and bind to chromatin. We conclude that the role of ATP binding and hydrolysis is different for different MCM subunits.

THE minichromosome maintenance (MCM) proteins were originally identified in a genetic screen for budding yeast mutants that were defective for minichromosome maintenance (MAINE et al.. 1984 ) or cell cycle progression (MOIR and BOTSTEIN 1982 ; KEARSEY et al.. 1996 ). The MCM family has six members (Mcm2p through Mcm7p), which are highly conserved in all eukaryotes (reviewed in PASION and FORSBURG 2001 ). Each of the MCM proteins is essential for initiation and elongation of DNA replication (reviewed in LABIB et al.. 2000 ; PASION and FORSBURG 2001 ). These proteins share a central homology domain of 200 amino acids containing motifs characteristic of DNA-dependent and AAA⁺ ATPases, including both Walker A and B motifs (WALKER et al.. 1982 ; KOONIN 1993 ; NEUWALD et al.. 1999 ). Biochemical studies indicate that the MCM complex is formed by stoichiometric assembly of all six MCM proteins. However, the purification of subcomplexes by glycerol gradients and co-immunoprecipitation studies suggests that the complex contains a tightly associated core formed by Mcm4p, -6, and -7, which is loosely bound by Mcm2p and the strongly associated Mcm3p-Mcm5p dimer (reviewed in LABIB and DIFFLEY 2001 ; PASION and FORSBURG 2001 ).

Interestingly, the intact MCM complex does not have any detectable biochemical activity (ADACHI et al.. 1997 ; ISHIMI et al.. 1998 ; LEE and HURWITZ 2000 ; SATO et al.. 2000 ). However, a dimer formed by the trimer subcomplex Mcm4/6/7p possesses in vitro single-stranded DNA and ATP-binding activities as well as ssDNA-dependent ATPase and 3' to 5' DNA helicase activities (ISHIMI et al.. 1998 , 2000; LEE and HURWITZ 2000 , LEE and HURWITZ 2001 ). Mutational analysis within the Walker A and B motifs of mouse Mcm4p and Mcm6p has demonstrated the importance of these domains for the in vitro DNA helicase activity of the hexameric Mcm4/6/7p complex (YOU et al.. 1999 ).

The Walker A motif has the consensus sequence (G)xxxxGK[T/S], including the invariant lysine residue found in all ATP-binding proteins. The Walker A motif sequence in the MCMs contains a slight variation from the consensus, where the glycines in the GK(S/T) signature are substituted by alanine or serine (KOONIN 1993 ). On the basis of studies of known ATPases, it is suggested that the lysine in the A motif directly interacts with the phosphatyl group of bound ATP (WALKER et al.. 1982 ). The B motif typically has a hydrophobic structure D[D/E]. Apparently the common denominator for all variants of the B motif is at least one negatively charged residue preceded by a stretch of bulky hydrophobic residues predicted to form a ß-strand (KOONIN 1993 ). The acidic residues within this motif play an essential role in coordinating the magnesium ion and, probably, the water molecule that attacks the ß- bond of the ATP. Consistent with this hypothesis, it is postulated that motif B is involved in ATP hydrolysis rather than ATP binding (PAUSE and SONENBERG 1992 ; BROSH and MATSON 1995 ).

Electron microscopy studies showed that the hexameric Mcm4/6/7p complex can actually form toroidal structures with a central cavity, comparable in size and shape to known hexameric DNA helicases (SATO et al.. 2000 ). The MCMs are thus unique among hexameric helicases, which generally are homomeric rather than heteromeric (reviewed in PATEL and PICHIA 2000 ). Interestingly, Archea Methanobacterium thermoautotrophicum contains one MCM-related protein that forms a double hexamer and possesses in vitro 3' to 5' DNA helicase activity (KELMAN et al.. 1999 ; CHONG et al.. 2000 ; SHECHTER et al.. 2000 ). These findings and the fact that the MCMs are necessary not only for the initiation of DNA replication but also for the elongation of replication forks suggest that the MCM complex or the core subcomplex might actually be the eukaryotic replicative fork helicase similar to bacterial DnaB protein or simian virus 40 (SV40) large T antigen (T Ag; reviewed in BAKER and BELL 1998 ; TYE and SAWYER 2000 ; LABIB and DIFFLEY 2001 ).

To better understand the in vivo role of ATP-binding and hydrolysis motifs and to establish the importance of cyclin-dependent kinase (CDK) phosphorylation in the function of the MCM complex, we constructed mutations in the conserved Walker A and B motifs and the putative CDK phosphorylation sites of core MCMs. We show that equivalent mutations in the A and B motifs of Mcm4, -6, and -7 proteins have different in vivo effects, suggesting that the role of ATP binding and hydrolysis in the function of each core MCM protein is different. Both motifs in Mcm4p are essential, while in Mcm6p they are dispensable. Mcm7p shows an intermediate phenotype: Only a conservative substitution of the crucial lysine in the A motif resulted in a functional protein. We found that not all equivalent mutations that abolished the in vitro helicase activity of the mouse hexameric Mcm4/6/7p subcomplex (YOU et al.. 1999 ) affect the in vivo function of the fission yeast MCM complex. Analysis of the Mcm4 mutant proteins showed different defects in MCM complex assembly, nuclear localization, or chromatin binding. Finally, mutations of the two putative CDK phosphorylation sites in Mcm4p and the single consensus site in Mcm7p had no effect upon the normal growth of cells, suggesting that these sites are not crucial for the regulation of MCM function.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains and manipulation:
Strains used in this study are listed in Table 1. Fission yeast strains were grown in yeast extract plus supplements (YES) or in Edinburgh minimal medium (EMM) with appropriate supplements as described (MORENO et al.. 1991 ). In this study "wild type" refers to strain FY254. Transformations were carried out by electroporation (KELLY et al.. 1993 ). Unless indicated, yeast plasmids used in this work contain the ura4⁺ marker. The full-strength nmt1⁺ promoter was used (BASI et al.. 1993 ). When studying complementation using constructs with the nmt promoter, transformants were selected on plates containing 15 µM thiamine (fully repressed). Different levels of induction in liquid culture were obtained by washing the cells twice with 10 ml of EMM followed by the addition of the indicated thiamine concentrations (JAVERZAT et al.. 1996 ). Overexpression experiments in liquid culture were carried out for a maximum of 23 hr following the removal of thiamine from the media. Overexpression on plates was done by streaking individual colonies on media lacking thiamine and incubating for 3–5 days at 32°. Vectors without inserts were used as negative controls.

Complementation analyses of temperature-sensitive strains were carried out by streaking at least four independent colonies of each transformation on EMM plates with appropriate supplements and incubating the cells at 36° for 3–5 days. When testing complementation using the nmt promoter, colonies were streaked on media with three different concentrations of thiamine: 15 µM thiamine (+ thiamine, maximum repression), 0.05 µM thiamine (low thiamine, medium induction), and no thiamine (- thiamine, maximum induction).

Complementation of null strains was carried out using diploid strains containing one allele of the mcm4⁺, mcm6⁺, or mcm7⁺ genes disrupted with his3⁺ or ura4⁺. These strains were transformed with ura4⁺ or leu1⁺ plasmids. Transformants were selected on EMM plates lacking uracil and histidine. At least two independent transformed diploids for each transformation were induced to undergo meiosis, which was followed by a random spore analysis. The spore preparations were plated to identify Ura⁺, His⁺ colonies, which were analyzed by FACS to determine DNA content. Plasmids were rescued from haploid strains and analyzed by restriction mapping or sequencing to confirm the presence of the mutation.

For construction of strains FY1602, FY1603, FY1604, and FY1605 (Table 1), plasmids pEBG56 (pJK148 nmt-mcm4HA leu1⁺), pEBG57 (pJK148 nmt-mcm4D AHA leu1⁺), pEBG58 (pJK148 nmt-mcm4K AHA leu1⁺), and pEBG59 (pJK148 nmt-mcm4K RHA leu1⁺) were linearized with NruI within the leu1⁺ sequence and integrated into the haploid strain FY254. Leu⁺ transformants were selected and the integration event was confirmed by streaking and replica plating the cells on YES media. Expression of wild-type and mutant Mcm4-HA proteins was verified by immunoblot.

Protein extracts, immunoblotting, and immunoprecipitations:
Cell lysates were prepared by glass bead lysis according to MORENO et al. (1991) in lysis buffer (50 mM HEPES, pH 7.0, 50 mM KCl, 5 mM MgAc, 100 mM sorbitol, 0.1% Triton X-100) with the addition of 1 mM dithiothreitol (DTT), 1 mM ATP, and protease inhibitors. Lysates were cleared by spinning at 16,000 x g for 20 min at 4°. Total protein concentrations were determined by BCA protein assay (Pierce Chemical, Rockford, IL). For immunoblot analyses, samples were fractionated by 7% SDS-PAGE and transferred to Immobilon-P (Millipore, Bedford, MA). Antibodies to Mcm6p (serum 5899), Mcm5p (serum 5897), and Mcm4p (serum 5898) were previously described in SHERMAN et al. (1998); to Mcm2p (serum 5616) in FORSBURG et al. (1997); to Mcm3p (serum 6178) in SHERMAN and FORSBURG 1998 ; and to Mcm7p (serum 6184) in LIANG and FORSBURG 2001 . Monoclonal anti-hemagglutinin (HA) 12CA5 antibody was a kind gift of Tony Hunter, and monoclonal anti--tubulin antibody was purchased from Sigma (St. Louis; T5168). Detection was carried out using anti-rabbit or anti-mouse HRP-conjugated secondary antibodies (Sigma or Jackson ImmunoResearch Laboratories, West Grove, PA) and enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Immunoprecipitations were performed with 600 µg of total protein at a concentration of 1 µg/µl and 1 µl of 12CA5 antibody, for 4 hr at 4°. Fifty microliters of Protein A-Sepharose CL-4B (Sigma; 1:1 in lysis buffer) were added and incubated for 1–2 hr at 4°. Pellets were washed four times with 1 ml of cold lysis buffer and boiled after the addition of 80 µl of 2x sample buffer. Ten microliters per lane of each immunoprecipitation were loaded on a SDS-PAGE.

In situ chromatin-binding assay and immunofluorescence analysis:
An in situ chromatin-binding assay was adapted from the protocol described in KEARSEY et al. (2000) as follows. After the Zymolyase treatment, cells were washed with cold STOP buffer as indicated KEARSEY et al. (2000) and once with 10 ml of PEMS buffer (100 mM PIPES, pH 6.9, 1 mM EGTA, 1 mM MgSO₄, 1.2 M sorbitol). Cells were resuspended in 4 ml of PEMS + protease inhibitors and split in two. One-tenth volume of PEMS + 10% Triton X-100 was added to one of the tubes and 1/10 volume of PEMS + 0.25% Triton X-100 was added to the other. After an incubation for 10 min at room temperature, the cells were spun down and resuspended in 20 ml of FIX solution (0.1 M K-phosphate, pH 6.5, 10% methanol, 3.7% formaldehyde) for 20 min at room temperature. After pelleting cells and washing once in 10 ml of PEMS, the cells were blocked in 1 ml PEMBAL buffer (1x PEM, 1% BSA, 0.1% NaN₃, 100 mM L-lysine monohydrochloride) for 30 min at room temperature. Cells were incubated with monoclonal anti-HA 16B12 antibody (BabCO) overnight at room temperature at a 1:5000 dilution. After three 1-ml washes with PEMBAL the cells were incubated with Cy3-conjugated donkey anti-mouse Ig-G antibody (no. 715-165-150; Jackson ImmunoResearch Laboratories) for 1 hr at room temperature in the dark at 1:500 dilution. Following washes with PEMBAL, cells were stained with 4',6-diamidino-2-phenylindole (DAPI) and heat fixed onto microscope slides. Microscopy was performed with a Leitz Laborlux S microscope or a Leica DMR microscope. Images were captured with a SPOT2 charge-coupled device digital camera or a Hamamatsu digital camera directly into Adobe Photoshop or Improvision Openlab, respectively.

Flow cytometry:
Cells were fixed in ice-cold 70% ethanol and stained for flow cytometry as described previously (SAZER and NURSE 1994 ) except that the cells were stained in a final concentration of 1 µM Sytox Green (Molecular Probes, Eugene, OR). Flow cytometry was performed on a FACScan (Becton Dickinson, San Jose, CA), and data analysis was carried out using Cell Quest software for Macintosh.

Plasmid and mutant construction:
Plasmid features are described in Table 2. Nucleotide changes were introduced using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) following the manufacturer's instructions. The constructs were sequenced to confirm the mutations and that PCR mutagenesis introduced no extra changes. Details of plasmid construction are available upon request.

Kinase assays:
To purify substrate protein, 500-ml cultures of BL21(DE3)pLysS Escherichia coli (Invitrogen, San Diego) transformed with pMGC28 or pMGC29 were induced to express protein, and then his-tagged N terminus Mcm4 was purified under native conditions using Ni-NTA agarose (Qiagen, Valencia, CA) according to manufacturer's instructions.

For kinase assays, wild-type yeast (FY254) transformed with pSLF172-cdc13-HA (BROWN et al.. 1997 ) was grown at 32° in EMM + supplements + thiamine and then washed and grown in EMM + low thiamine (0.05 µM thiamine) for 18 hr at 32°. Cell lysates were prepared by glass bead lysis in HB buffer (25 mM MOPS, pH 7.2, 60 mM ß-glycerophosphate, 15 mM p-nitrophenylphosphate, 15 mM MgCl₂, 15 mM EGTA, 1 mM DTT, 1% Triton X-100, 1 mM PMSF, and protease inhibitors) as described above. To precipitate Cdc2/Cdc13-HA kinase, 0.5 µl 12CA5 antibody was added to 300 µg extract and rotated at 4° overnight. Fifty microliters Protein A-Sepharose CL-4B (Sigma; 1:1 in lysis buffer) was then added and rotated at 4° for 1 hr. Pellets were washed four times in HB buffer, and then all HB was removed and 20 µl kinase reaction buffer (5 µCi [-³²P]ATP and 25 µM cold ATP in HB buffer) was added. After addition of 2 µg substrate protein [histone H1 (Boehringer Mannheim, Indianapolis), N terminus Mcm4 wild type, or N terminus Mcm4 T15A T112A], reactions were incubated at 32° for 30 min and boiled after addition of 4 µl 5x sample buffer. Proteins were fractionated by 10% SDS-PAGE. The Coomassie-stained gel was scanned into Canvas 5 for Macintosh and ³²P incorporation was determined using a Phosphoimager (Molecular Dynamics, Sunnyvale, CA).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Rationale:
Mutations of the crucial lysine in the Walker A motif of several ATPases disrupt ATP hydrolysis, abolishing biological activity (MA et al.. 1994 ; SINGH and MAURIZI 1994 ; YAHRAUS et al.. 1996 ). Interestingly, in some proteins, a conservative substitution of arginine for this lysine may not significantly affect nucleotide binding or nucleotide-dependent protein association, although it abrogates ATPase activity (SUNG et al.. 1988 ; REHRAUER and KOWALCZYKOWSKI 1993 ; SUNG and STRATTON 1996 ). Previous data from our laboratory showed that when the invariant lysine present in the A motif of Mcm2p is substituted with arginine, this protein is able to complement both mcm2-ts and mcm2 strains, but a K to A change gives a nonfunctional protein, suggesting that it is unlikely that this Mcm2p domain functions as a bona fide ATPase (FORSBURG et al.. 1997 ).

To determine the importance of the nucleotide-binding and hydrolysis motifs for the in vivo function of the MCM complex, we introduced mutations into the conserved Walker A and B domains of the core MCM proteins. These sequences are shown in Fig 1. In particular, we made conservative K R and nonconservative K A mutations in the Walker A motif and D A mutations in the Walker B motif. The mutant genes were expressed on plasmids and tested for complementation of the cognate mcm-ts and mcm mutations. Where possible, these results were correlated with complex assembly, nuclear localization, and chromatin binding by expressing epitope-tagged forms of the proteins in wild-type cells. However, in several cases the epitope tag caused additional phenotypes, preventing this analysis.

View larger version (62K): In this window In a new window Download PPT slide   Figure 1. Protein sequence alignment of the S. pombe MCM Walker A and B motifs. The MCM homology domains of Mcm2p through Mcm7p were aligned with the ClustalW 1.8 program and formatted with MacBoxShade v2.01 and Microsoft Word 98. The Walker A and B motifs alignment is shown. White text in black boxes indicates identical amino acids present in the six MCMs. Black text in gray boxes indicates identical amino acids present in four or more MCMs. Asterisks indicate mutated amino acids. Walker A and B motifs consensus sequences are shown. Numbers to the left of each sequence indicate position of first amino acid in the Walker motifs.

For clarity, we use the MCM designation for each of these genes. However, several are also known by other gene names in the literature: mcm2⁺ corresponds to cdc19⁺ and nda1⁺ (MIYAKE et al.. 1993 ; FORSBURG and NURSE 1994 ); mcm4⁺ is cdc21⁺ (COXON et al.. 1992 ); mcm5⁺ is nda4⁺ (MIYAKE et al.. 1993 ); and mcm6⁺ is mis5⁺ (TAKAHASHI et al.. 1994 ).

Mcm4p nucleotide-binding and hydrolysis motifs are essential for its function in vivo:
The wild-type mcm4⁺ and mutant derivatives were cloned under the control of the nmt promoter with or without a 3' triple-HA epitope tag and under their own promoter with a 3' triple-HA epitope tag (Table 2 and MATERIALS AND METHODS). We verified that the wild-type Mcm4p clones were functional by ensuring that they could rescue both the temperature-sensitive and null alleles of mcm4⁺ (Table 1, Fig 2A, and data not shown). Next we tested if the mutant proteins Mcm4D A, K A, and K R, with and without a C-terminal HA tag, were functional. None of the Mcm4 mutant proteins could rescue the mcm4-M68 temperature-sensitive or null phenotype (Fig 2A and Fig B). When expressing the untagged Mcm4K R protein, microcolonies with very elongated and sick cells were sometimes observed but could not be propagated in liquid media (data not shown). Taken together, these results indicate that conservative and nonconservative mutations in the Walker A motif, and the nonconservative change in the B motifs, give rise to nonfunctional Mcm4p proteins.

View larger version (46K): In this window In a new window Download PPT slide   Figure 2. Mcm4p, Mcm-6p, and Mcm7p Walker A and B motifs mutational analysis. mcm4+, mcm6+, and mcm7+ and the corresponding mutant derivatives were cloned in different expression vectors. (A) The complementation results. A schematic depiction of each construct is shown. Shaded arrows represent the nmt promoter. Open arrows symbolize endogenous promoters. Open rectangles stand for mcm4 (Mcm4p), mcm6 (Mcm6p), and mcm7 (Mcm7p) genes as indicated. The vertical line at the end of the mcm6 contruct symbolizes the 3' truncated mcm6 gene. Shaded boxes represent the MCM central homology domains, where the relative positions of the Walker A and B motifs are shown. The triple-HA epitope tags are represented by solid boxes. + and - signs indicate complementation and no complementation of the temperature-sensitive and null alleles, respectively. ND, not determined. +/- indicates complementation of the temperature-sensitive strains and no complementation or partial complementation of the null alleles. DE AA and KS AA double mutants were tested for Mcm6p only. (B) Complementation results boxed in A. Plasmids with the untagged mcm4+, mcm6+, and respective mutant genes, under the control of the nmt promoter, were transformed into the mcm4-ts strain FY786 or mcm6-ts strain FY961, respectively. Isolated transformants were streaked on EMM + adenine + leucine + thiamine. Plasmids with the untagged mcm7+ and mutant genes under the mcm7 promoter were transformed into mcm7-ts strain FY1199, and the isolated transformants were streaked on EMM + adenine + leucine. Cells were grown at 36° for 2–3 days.

Single point mutations in the ATP-binding and hydrolysis motifs of Mcm6p do not affect the function of the protein:
Equivalent mutations were constructed for the mcm6⁺ gene. In our lab it was previously shown that the protein expressed by pSLF225 (nmt-HAmcm6⁺; FORSBURG et al.. 1997 ) could rescue both the mcm6 temperature-sensitive (mcm6-268; TAKAHASHI et al.. 1994 ) and null alleles (SHERMAN et al.. 1998 ). HA-mcm6 mutant plasmids, derived from pSLF225, were constructed (Table 2).

Surprisingly, all of these changes gave rise to functional Mcm6 proteins; the three mutants could complement the mcm6-ts mutation and the mcm6 null strain (FY857; Table 2).

Previous published results using baculovirus recombinant mouse MCM proteins showed that the double mutant Mcm6 K401A S402A or Mcm6 D459A E460A had significantly reduced or completely abolished helicase activity in vitro (YOU et al.. 1999 ). Thus, we were surprised to find that single substitutions in the Walker motifs of Mcm6p gave functional in vivo proteins in Schizosaccharomyces pombe. To compare our in vivo results with the in vitro mutational analysis previously reported (YOU et al.. 1999 ), we constructed the double mutants Mcm6p KS477/478AA and Mcm6p DE527/528AA (see Table 2) and tested their ability to complement mcm6-ts and null mcm6 strain. Interestingly, the Walker B motif Mcm6DE AA mutant was functional while the Walker A domain double mutant Mcm6KS AA was not.

During the construction of the mcm6 mutants we discovered that the plasmids used to make our mutant vectors (pTZmcm6 and pSLF225), which were constructed using the originally published sequence (TAKAHASHI et al.. 1994 ), were missing 73 nucleotides on the basis of the sequence in the S. pombe genome database (accession no. AL139314). As a result, all our constructs also had a premature stop codon, resulting in a shorter but functional Mcm6p. The correct 3' end was introduced by PCR (Table 2). These revised constructs behaved the same as their counterparts lacking the last 73 nucleotides, suggesting that the last 29 amino acids are not required for Mcm6p activity.

Similar results were observed using plasmids without a tag under the nmt promoter (Fig 2A and Fig B). Complementation at 36° of the temperature-sensitive strain occurred with all three thiamine concentrations, indicating that overproduction of Mcm6p and the mutant derivatives is not toxic for the mcm6-ts. There was no deleterious phenotype associated with overproduction in wild-type cells (data not shown).

Unexpectedly, when the expression of the N-terminally tagged HA-Mcm6p was analyzed by Western blot we observed that an extremely high percentage of the Mcm6p-tagged protein lost its tag (data not shown). We constructed a C-terminally HA-tagged Mcm6p but this protein was unable to complement the null mcm6 strain and could rescue the mcm6-ts defect only when overproduced (Fig 2A, Table 2). Without a tag to distinguish wild-type and mutant proteins further analysis of the Mcm6p mutants was not possible.

Nonconservative mutations in the ATP-binding and hydrolysis motifs of Mcm7p abolish its activity in vivo:
Similar mutations in the Walker A and B motif sequences were introduced in the mcm7⁺ gene. Wild-type and mutant genes mcm7D A, mcm7K A, and mcm7K R were cloned into the pUR19-HindIII vector (Table 2). These constructs included the complete mcm7⁺ open reading frame plus 2549 nucleotides upstream and 540 nucleotides downstream, corresponding to the mcm7⁺ promoter and its endogenous stop sequence, respectively. As shown in Fig 2A and Fig B, the wild-type Mcm7p as well as the Mcm7K R mutant could rescue the mcm7-98 temperature defect, while D A and K A could not. Equivalent results were obtained when testing the ability of these proteins to complement the null allele (Fig 2A). These results show that nonconservative changes in crucial residues in the A and B motifs abolish the function of Mcm7p.

To study in more detail which aspects of the Mcm7p function were affected by the mutations, we needed tagged versions of these proteins and an inducible system to control their expression. mcm7⁺ under the control of the nmt promoter with a 3' triple-HA epitope tag could not complement the mcm7-98 defect or the null strain (Fig 2A and LIANG and FORSBURG 2001 ). Curiously, as shown in LIANG and FORSBURG 2001 , mcm7HA in the chromosome is functional, suggesting that the cells are very sensitive to precise levels of the Mcm7p protein or mcm7 mRNA.

We constructed N-terminally tagged Mcm7p and found that the protein complemented the mcm7-ts strain and the null allele. However, in contrast to the untagged protein, HA-Mcm7K R could not rescue the mcm7-98 defect at 36° (Table 2), suggesting that the HA epitope negatively affects the activity of the Mcm7p K409R mutant. These results indicate that the HA tag is not neutral in Mcm7p.

In a final attempt to have tagged Mcm7 proteins that would behave as the untagged counterparts, we cloned the mcm7 derivatives with a 3' HA epitope tag under the control of the mcm7⁺ promoter (Table 2). The wild-type Mcm7p-HA protein was able to rescue mcm7-98 and the null allele, in contrast to the episomal nmt-mcm7⁺-HA. Since the only difference is the promoter, we suggest that the mcm7⁺ gene may require some form of transcriptional regulation. However, the mutant Mcm7-K R-HA did not complement as well as the untagged version (Fig 2A). We conclude that Mcm7K409R is a functional protein that becomes defective when a triple-HA epitope is added at either end: a synthetic phenotype.

Interestingly, overexpression of the functional wild-type HA-Mcm7p from an episome or an integrated allele was toxic for normal cell growth and viability.

Analysis of MCM complex formation by Mcm4p mutants:
To investigate the ability of mutant MCMs to form a complex or localize properly in vivo, the proteins must be expressed in a wild-type cell (to maintain viability), with an epitope tag (to distinguish it from the wild-type protein). Only in the case of Mcm4 was there no significant difference in the behavior of tagged and untagged proteins, so our protein analysis focused on this MCM subunit.

To analyze complex assembly, we used a co-immunoprecipitation competition strategy in which mutant proteins are expressed at low levels and immunoprecipitated; other members in the MCM complex are detected by Western blot (SHERMAN et al.. 1998 ). Plasmids coding for wild-type and mutant derivatives were transformed into a wild-type strain. To compare relative complex affinity, the expression levels of the tagged proteins have to be similar to one another and to the endogenous wild-type Mcm4p levels. When cells were grown with the same concentration of thiamine, we noted that the expression levels of the different tagged proteins varied reproducibly (data not shown). We accommodated this by culturing the cells in different thiamine concentrations as indicated in Fig 3A.

View larger version (61K): In this window In a new window Download PPT slide   Figure 3. Nonconservative mutations in Mcm4p Walker motifs reduce MCM complex formation. (A) Plasmids pMGC32, pEBG4, pEBG5, and pEBG6, encoding for Mcm4p-HA (WT), Mcm4D A-HA (DA), Mcm4K A-HA (KA), and Mcm4K R-HA (KR), respectively, were transformed into the wild-type strain FY254. Cells were grown at 32° in minimal media plus the indicated concentrations of thiamine. Protein lysates were separated by SDS-PAGE and transferred to Immobilon. Endogenous Mcm4p and the HA-tagged proteins were detected with anti-Mcm4p antibodies (anti-mcm4), and HA antibodies (anti-HA) revealed the HA-tagged proteins. (B) The HA-tagged proteins were selectively immunoprecipitated with antibodies to HA. The immunoprecipitated proteins were separated by SDS-PAGE, transferred to Immobilon, and blotted with the indicated MCM antibodies. Note that Mcm4D A-HA has a faster mobility.

The HA-tagged proteins were immunoprecipitated and Western blot analysis was used to detect coprecipitating MCM proteins (Fig 3B). The amount of immunoprecipitated HA-tagged proteins was determined by blotting with an anti-Mcm4 antibody (Fig 3B). The endogenous wild-type Mcm4p was not detected in the anti-HA precipitates, suggesting that there is only one molecule of Mcm4p protein per complex. Similar results were reported for Mcm2p (SHERMAN et al.. 1998 ).

Compared to the amount of MCMs bound to Mcm4p-HA, normal levels of MCMs coprecipitated with Mcm4-K R-HA, while somewhat reduced levels were associated with Mcm4D A-HA and Mcm4K A-HA (Fig 3B). These results show that all the mutants can assemble into a complex, although the nonconservative mutants have reduced affinity. Interestingly, all the coprecipitated MCMs were affected to the same extent. These data suggest that the Walker domains of Mcm4p contribute to the formation of a stable MCM2-7 complex.

In situ chromatin localization:
Next, we investigated the ability of the mutant HA-tagged Mcm4p to localize to the nucleus and to bind to chromatin. In fission yeast, MCMs are nuclear throughout the cell cycle (MAIORANO et al.. 1996 ; OKISHIO et al.. 1996 ; SHERMAN and FORSBURG 1998 ; PASION and FORSBURG 1999 ; LIANG and FORSBURG 2001 ) but periodic association with chromatin has been reported (OGAWA et al.. 1999 ; KEARSEY et al.. 2000 ). Recently a new technique was published to analyze the chromatin association of fission yeast proteins on the basis of detection of Mcm4p-green fluorescent protein (GFP) fluorescence in permeabilized cells after extraction with a nonionic detergent (KEARSEY et al.. 2000 ). In permeabilized cells that were not detergent washed, Mcm4p-GFP was nuclear throughout the cell cycle, but following detergent extraction, Mcm4p-GFP nuclear localization was visualized only in the nucleus of binucleate (M/G1/S phase) cells, where it is bound to chromatin.

We modified this technique to analyze fission yeast HA-tagged proteins during the cell cycle (see MATERIALS AND METHODS). Our revised method differs primarily in the use of 0.025% (low) Triton X-100 washes to enhance nuclear signal without delocalizing the HA-tagged proteins. To test this method we analyzed the localization and chromatin binding of Mcm2p-HA (FY798), Mcm4p-HA (FY1718), and Orp1p-HA (FY793) integrated at the endogenous locus (Fig 4A). The proteins were localized to the nucleus in all cells when treated with low Triton X-100. Orp1p-HA remained nuclear throughout the cell cycle when cells were washed with 1% (high) Triton X-100, while Mcm2p-HA and Mcm4p-HA were visible in binucleate cells only. These results agree with the previous Mcm4p-GFP published data indicating that Mcm2p and Mcm4p bind to chromatin only at the end of mitosis and during G1/S phase (binucleate cells; KEARSEY et al.. 2000 ), while Orp1p associates with chromatin throughout the cell cycle (LYGEROU and NURSE 1999 ).

View larger version (87K): In this window In a new window Download PPT slide   Figure 4. The Mcm4p Walker A and B domain mutants can localize to the nucleus but only Mcm4p K to R can bind to chromatin. Cells were permeabilized and washed with 0.025% Triton X-100 or 1% Triton X-100 as indicated. DNA was visualized by DAPI staining (DNA). HA-tagged proteins were visualized using the HA antibody and a secondary antibody conjugated to Cy3 (HA). Bar, 10 µm. (A) Chromatin-binding analysis of S. pombe HA-tagged Mcm2p, Mcm4p, and Orp1p. The Mcm2p-HA (FY798), Mcm4p-HA (FY1718), and Orp1p-HA (FY793) proteins were expressed from the native promoters as the only copies in the cells. (B) mcm4+ and the mutant versions, HA tagged at the 3' end and under the control of the nmt promoter, were integrated at the leu1 locus of strain FY254. Cells were grown at 32° in the presence of thiamine. Cells were photographed using the same exposure times. 4 WT, Mcm4p-HA; 4 K R, Mcm4K R-HA; 4 K A, Mcm4K A-HA; 4 D A, Mcm4D A-HA. (C) Mcm4p K to R mutants also bind to chromatin in hydroxyurea S-phase-arrested cells. Strains with the integrated HA-tagged mcm4 genes under the control of the nmt promoter were incubated for 3.5 hr with 15 µM hydroxyurea in the presence of thiamine. Cells were photographed using the same exposure times.

Mcm4-HA clones under the control of the nmt promoter were integrated into the leu1⁺ locus of strain FY254 to eliminate cell-to-cell variation observed in expression from episomes (e.g., PASION and FORSBURG 1999 ). As for the association assays, the HA-tagged proteins have to compete with the endogenous Mcm4p for association with the other MCMs, and the mutant MCM complexes have to compete to enter the nucleus and to bind to chromatin. Cells were grown asynchronously in media containing thiamine, permeabilized, treated with low or high Triton X-100, and fixed. Western blot analysis confirmed the expression of similar levels of HA-tagged proteins (data not shown). In cells treated with low Triton X-100, Mcm4p-HA and the HA-tagged mutant proteins were nuclear throughout the cell cycle (Fig 4B), indicating that all the analyzed proteins can localize to the nucleus. After detergent extraction (+ 1% Triton X-100), nuclear localization was lost in mononucleate G2 cells. Binucleate cells retained nuclear localization only when Mcm4p-HA and Mcm4-K R-HA proteins were expressed (Fig 4B). The intensity of the signal was weaker in the K to R mutant when compared with Mcm4p-HA. Thus chromatin binding appears particularly sensitive to mutations of the MCM4 Walker motifs.

To further analyze the ability of these mutants to bind to chromatin, we arrested the cells as mononucleates in early S phase with hydroxyurea (Fig 4C). After 3.5 hr in 15 mM hydroxyurea, a high proportion of the cells had a 1C DNA content (data not shown). Cells washed with low Triton X-100 had the HA-tagged proteins in the nucleus (Fig 4C). When treated with high Triton X-100, only Mcm4p-HA and Mcm4K R-HA proteins remained chromatin bound (Fig 4C).

From these results we conclude that the three nonfunctional Mcm4p mutant proteins can localize to the nucleus, but only the Mcm4K R-HA mutant protein can bind to chromatin. However, the mutant Mcm4-K R-MCM complexes appear less efficient at binding to chromatin than wild-type MCM complexes.

Walker A domain mutant proteins have a dominant negative phenotype when overproduced:
As previously reported (MAIORANO et al.. 1996 ; FORSBURG et al.. 1997 ), when mcm4⁺ is expressed by full-strength nmt promoter on an episome at 32° it is dominant negative in wild-type cells. When nmt-mcm4HA was present as a single copy in the cell (strain FY1602), overproduction of Mcm4p-HA protein was not toxic (Fig 5A). To test the effect of overproducing the Mcm4 mutant proteins, strains FY1603, FY1604, and FY1605 were streaked on plates without thiamine and incubated at 32°. After 3 days, overproduction of Mcm4D A-HA had no phenotype: Colony formation was normal and cell and colony morphology was similar to cells overproducing Mcm4p-HA (Fig 5A). In contrast, overproduction of Mcm4-K A-HA and to a lesser extent Mcm4K R-HA was toxic for the cells: Colonies were much smaller and cells were elongated (Fig 5A). To analyze this dominant negative overproduction phenotype, cells were grown in liquid media lacking thiamine for a maximun of 23 hr (Fig 5B and Fig C). The DNA fluorescence profiles of cells overproducing Mcm4D A-HA and Mcm4K R-HA were indistinguishable from one another and slightly wider than when Mcm4p-HA was overproduced (Fig 5B). In contrast, the peak of DNA content was much broader when Mcm4K A-HA was overproduced (Fig 5B). Cells were DAPI stained and visualized under the microscope (Fig 5C). Cells overproducing Mcm4p-HA and Mcm4D A-HA were slightly elongated and one DAPI-stained body was visualized per cell. When Mcm4K R-HA was overproduced, cells were elongated and some nuclear fragmentation was detected; this was dramatically exacerbated when Mcm4-K A-HA was overproduced (Fig 5C). Interestingly, only the Walker A motif mutants had a dominant negative phenotype, suggesting that nucleotide-binding and hydrolysis domains might be important for different aspects of Mcm4p function. However, these phenotypes did not correlate with chromatin binding or complex assembly. Table 3 summarizes these results.

View larger version (57K): In this window In a new window Download PPT slide   Figure 5. Walker A motif mutant proteins have a dominant negative phenotype when overproduced. Strains carrying the nmt-mcm4 HA-tagged wild-type and mutant genes integrated at the leu1 locus of strain FY254 were analyzed. WT, D A, K A, and K R correspond to cells expressing Mcm4p-HA, Mcm4-D A-HA, Mcm4K A-HA, and Mcm4K R-HA, respectively. (A) Cells were streaked on + thiamine and - thiamine plates as indicated and grown at 32° for 3 days. (B) Cells were grown at 32° in liquid medium lacking thiamine for a maximum of 23 hr. Cells were ethanol fixed and analyzed by flow cytometry; shown are histograms of DNA content. (C) DAPI staining of cells analyzed in B. Bar, 10 µm.

Mcm4p and Mcm7p with mutations in the consensus CDK phosphorylation sites are functional proteins:
We investigated the role of CDK phosphorylation of Mcm4p and Mcm7p by mutating the two CDK consensus sites at positions 15 and 112 of Mcm4p to alanine (mimics an unphosphorylated protein), serine (alternate phosphate acceptor), or glutamate (mimics a constitutively phosphorylated protein). We found that Mcm4p with all of these mutations is fully functional. These mutant proteins complement the temperature sensitivity of mcm4-M68 (Fig 6A) and are wild type for growth rate and cell cycle progression when integrated into the genome as the only copy of mcm4 (data not shown). Further, these Mcm4p mutants did not affect the rate of rereplication in fission yeast cells overexpressing Cdc18p or Rum1p (data not shown). To determine if Mcm4p lacking the CDK consensus sites is no longer a substrate for CDK phosphorylation, we expressed the first 384 amino acids of wild-type or T15A T112A Mcm4p in bacteria. When purified protein was added to Cdc2p/Cdc13-HAp immunoprecipitated from fission yeast extracts, we found that both Mcm4p fragments were phosphorylated equivalently (Fig 6B). In addition, when we constructed a version of Mcm7p with a mutation of the single CDK consensus site at position 546 to alanine, we found that this protein was able to complement the growth defect of mcm7-98 at the restrictive temperature (Fig 6A) as well as complement mcm7 (data not shown). Thus, the CDK consensus sequences are not required for protein activity.

View larger version (35K): In this window In a new window Download PPT slide   Figure 6. Mcm4p and Mcm7p with mutations in the consensus CDK phosphorylation sites are functional proteins. (A) mcm4-M68 (FY784) transformed with empty vector (pIRT2) or wild-type (pAC1), 2T 2A (T15A T112A, pMGC39), 2T 2S (T15S T112S, pMGC40), or 2T 2E (T15E T112E, pMGC41) mcm4-containing plasmid (top) and mcm7-98 (FY1199) transformed with empty vector (pUR19-H), wild-type (pEBG30), or T A (T546A, pEBG43) mcm7-containing plasmid (bottom) were grown at 36° for 3 days. (B) Cdc2/Cdc13 was immunoprecipitated from fission yeast extracts and incubated with [-32P]ATP and histone H1 (positive control for kinase activity) or the first 384 amino acids of wild-type or 2T 2A (T15A T112A) Mcm4p produced in bacterial cells. 32P incorporation (left) and protein loading (right) are shown.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The MCM proteins are members of the large AAA+ ATPase family and are required not only for replication initiation but also for elongation. Mammalian (ISHIMI 1997 ; YOU et al.. 1999 ) as well as fission yeast (LEE and HURWITZ 2000 ) "core" MCMs (MCM4, -6, and -7) have been shown to have helicase activity in vitro, leading to the model that the MCM complex is a replicative helicase at the moving fork. However, the complete MCM hexamer (MCM2–7), which is found in vivo, does not have helicase activity in vitro (ADACHI et al.. 1997 ; ISHIMI et al.. 1998 ; LEE and HURWITZ 2000 ; SATO et al.. 2000 ).

We examined mutations that were shown to affect the in vitro DNA helicase activity of the core MCMs (YOU et al.. 1999 ) to determine their effect upon in vivo function. We constructed point mutations in the Walker A and B motifs characteristic of ATPases. Surprisingly, not all corresponding mutations that abolished the in vitro DNA helicase activity of the mouse hexameric Mcm4/6/7p subcomplex (YOU et al.. 1999 ) affected the in vivo function of fission yeast MCM complex (Table 4). That is, there is not a simple correlation between in vitro and in vivo phenotypes.

Substitutions in the mouse Mcm6p Walker B motif completely abolished the in vitro DNA helicase activity of the mutant subcomplex (YOU et al.. 1999 ), although equivalent mutations in fission yeast Mcm6p resulted in complementation. Thus, the Mcm6p Walker B motif appears to be essential for in vitro DNA helicase activity (YOU et al.. 1999 ) but dispensable for in vivo MCM complex function. None of the single mutations in Mcm6p disrupted complementation, although the double Walker A motif mutant KS AA was not able to complement. Data from other ATPases showed that single mutations of the crucial lysine in the Walker A motif abolish ATPase activity and in vivo function (LAURENT et al.. 1993 ; SINGH and MAURIZI 1994 ; RIKKONEN 1996 ). It is possible that the double mutation introduced in fission yeast Mcm6p Walker A domain could alter the protein structure, rather than simply affect its ATP binding. Unfortunately, we could not examine the behavior of this protein in vivo, as it could not be epitope tagged.

Interestingly, YOU et al. 1999 showed that Mcm6p is responsible for ATP binding by MCM4/6/7 and Mcm6p mutant subcomplexes cannot bind ATP (YOU et al.. 1999 ). However, our in vivo results suggest that Mcm6p is not directly involved in the MCM complex ATP-binding and hydrolysis activities. It is possible that in vitro Mcm6p Walker domains are essential for MCM4/6/7 subcomplex ATP binding and thus helicase activity, but in vivo this function can be carried out by a different MCM protein in the intact hexamer. The in vitro studies were carried out using a subset of the MCM proteins, as helicase activity was not obtained with the complete MCM2–7 complex. In contrast, studies in vivo suggest that all MCMs are present and contribute similarly to complex function (APARICIO et al.. 1997 ; SHERMAN et al.. 1998 ; LABIB et al.. 2000 ; LIANG and FORSBURG 2001 ), so perhaps Mcm2, -3, or -5 is responsible for the activity associated with Mcm6p in the in vitro experiments.

Previous results from our lab showed that a conservative substitution of the Mcm2p Walker A domain lysine was able to rescue mcm2-ts and mcm2 mutations, although a nonconservative substitution gave a nonfunctional protein (FORSBURG et al.. 1997 ). These data resemble the results obtained with Mcm7p where the K409R was functional while K409A was not. In several proteins, mutations of the Walker A motif lysine to arginine abolish ATPase activity but not their ability to bind ATP (REHRAUER and KOWALCZYKOWSKI 1993 ; SUNG and STRATTON 1996 ; SUNG et al.. 1988 ). These results suggest the crucial role of ATP in Mcm2p and Mcm7p may be in binding, rather than hydrolysis.

A recent published article (SCHWACHA and BELL 2001 ) shows some discrepancies with our in vivo mutational analysis. Schwacha and Bell analyze the ability of Saccharomyces cerevisiae MCM Walker A and B mutant proteins to complement null MCM strains using tagged MCMp versions under the control of the MCM5 promoter. In our study we found that the HA epitope added to either end of fission yeast MCM proteins is not neutral, especially if combined with other mutations. We speculate that some of the observed differences could be attributed to the use of a tag. However, the intrinsic differences between fission and budding yeasts should also be considered.

An additional finding came out of this study; mutations of the consensus CDK phosphorylation sites (S/T P XX K/R) in Mcm4p and Mcm7p did not have any significant effect on the function of the proteins. In the case of Mcm4p, phosphorylation by CDKs has been demonstrated in higher eukaryotes and correlates with loss of chromatin association (COUE et al.. 1996 ; HENDRICKSON et al.. 1996 ; FUJITA et al.. 1998 ). Mutation of the consensus sites in S. pombe Mcm4p did not have any effect on protein phosphorylation, suggesting that additional SP or TP residues may be phosphorylated; these are numerous throughout the protein. Similar CDK phosphorylation of sites that do not contain the final basic residue has been observed for Orp2p (VAS et al.. 2001 ). As is seen for S. cerevisiae, it may be necessary to combine CDK mutations in multiple proteins to observe a phenotype (NGUYEN et al. 2001 ).

For all the MCMs analyzed, both K A and D A mutations always had the same phenotype. Some reports show that mutations in Walker B motifs abolish ATPase activity but do not affect ATP binding (PAUSE and SONENBERG 1992 ; BROSH and MATSON 1995 ). However, others reported that mutations in this domain abolished binding as well (KLEMM et al. 1997 ). Our findings suggest that neither mutant binds ATP, although in vitro biochemical studies will be required to confirm this.

Our most detailed analysis was carried out on Mcm4p, since a tagged version of this protein behaved identically to the untagged version, facilitating protein analysis in cells. We analyzed the importance of the ATP-binding and hydrolysis domains in the formation of the MCM complex. Interestingly, although none of these mutants are functional, their ability to assemble into the MCM complex is different (Fig 3B). Several reports show that formation and stability of some protein complexes depend on nucleotide binding (e.g., SINGH and MAURIZI 1994 ; SAWAYA et al.. 1999 ; SINGH et al.. 1999 ; ZHANG et al.. 2000 ; JERUZALMI et al.. 2001 ). In most hexameric helicases, the formation of the ring hexamer structure requires nucleotide binding and Mg++ (reviewed in PATEL and PICHIA 2000 ). For some AAA+ family members, oligomerization is promoted both by nonhydrolyzable ATP analogs or by creating mutations that block ATP hydrolysis (reviewed in VALE 2000 ). The differences observed between conservative and nonconservative substitutions in the Walker A motif suggest that ATP bound to Mcm4p is necessary for its association with the other MCMs. Interestingly, both K A and D A mutants were similarly affected in their ability to form a stable MCM complex, suggesting that in vivo they might have similar defects.

Although ATP binding is involved in oligomerization of T7 gp4 helicase subunits, other hydrophobic and electrostatic interactions also contribute to the association and stability of the hexameric structure (SAWAYA et al.. 1999 ). The fact that neither Walker A nor B domain mutations completely abolished Mcm4p complex formation suggests that other regions of Mcm4p must be also involved in association. Because these mutant proteins were still able to form a MCM complex, we also conclude that none of these mutations seriously disrupts Mcm4p structure.

The ability of the Mcm4p mutants to localize to the nucleus and bind to chromatin was analyzed using an in situ chromatin-binding assay (KEARSEY et al.. 2000 ). The three HA-tagged mutants could localize to the nucleus similarly to wild-type Mcm4p-HA (Fig 4 and Fig 5). Previous reports from our lab showed that complex formation is necessary for proper MCM complex nuclear localization (PASION and FORSBURG 1999 ). All Mcm4p mutants can, to some extent, assemble into a complex and, therefore, be imported into the nucleus. Interestingly, there was no difference in cytoplasmic staining between Mcm4 wild-type and mutant proteins, suggesting that their nuclear localization is as effective as wild type.

In contrast, chromatin binding was defective in all three mutants. Mcm4-HA K A and D A mutants showed no chromatin binding, while for the conservative mutant K R it was reduced. Since this mutant is likely to have residual ATP binding (SUNG et al.. 1988 ; REHRAUER and KOWALCZYKOWSKI 1993 ; SUNG and STRATTON 1996 ), we suggest that ATP bound to Mcm4p is required for the association of the MCM complex with chromatin and/or association with other prereplicative complex proteins.

Either ATP or a nonhydrolyzable analog stabilizes chromatin binding of human MCMs (FUJITA et al.. 1997 ). Similarly, DNA binding of SV40 and polyomavirus T-antigen is enhanced by ATP and also by AMP-PNP (DEB and TEGTMEYER 1987 ; LORIMER et al.. 1991 ). In budding yeast, origin recognition complex (Orc)1p ATP binding, but not hydrolysis, is responsible for the ATP dependence of ORC DNA binding (KLEMM et al.. 1997 ). These observations are consistent with our results with Mcm4p mutants defective in chromatin binding, but this could also indicate that ATP is required for complex assembly, which is in turn required for chromatin association.

The different phenotypes we observe in subunits with similar mutations suggest that each subunit has a different role in the activity of the MCM complex. Biochemical studies using the 4A' product of T7 g4 DNA helicase showed that the six identical subunits are functionally distinct; similar to the F1-ATPase, 4A' T7 product also has noncatalytic and catalytic sites, and it was proposed that both proteins function in a similar way (HINGORANI et al.. 1997 ). In the case of the bacterial clamp loader, the three -subunits have different tertiary structures (reviewed in ELLISON and STILLMAN 2001 ). As suggested by others (TYE and SAWYER 2000 ), perhaps in the MCM hexamer some subunits are catalytic and others are noncatalytic. This would be consistent with the different activities associated with subcomplexes in vitro. Our mutational analysis provides evidence supporting this hypothesis, showing that the nucleotide-binding and hydrolysis domains of MCM4, -6, and -7 proteins are not equivalent. Perhaps the most important conclusion from this study is that hypotheses derived from biochemical studies must be correlated with in vivo analysis.

	ACKNOWLEDGMENTS

We thank Jill Meissenholder and Tony Hunter for the 12CA5 antibody, Magdalena Benzanilla for her advice with purifying recombinant proteins, and Wei Jiang for his help with the kinase assays. We are grateful to Julie Bailis, Will Dolan, and Debbie Liang for critical reading of the manuscript and all Forsburg laboratory members for helpful discussions. This work was supported by National Institutes of Health (NIH) grant T32 CA64041 to M.G.C. and NIH grant GM59321 to S.L.F., who is a scholar of the Leukemia and Lymphoma Society.

Manuscript received October 22, 2001; Accepted for publication January 10, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADACHI, Y., J. USUKURA, and M. YANAGIDA, 1997  A globular complex formation by Nda1 and the other five members of the MCM protein family in fission yeast. Genes Cells 2:467-479[Abstract].

APARICIO, O. M., D. M. WEINSTEIN, and S. P. BELL, 1997  Components and dynamics of DNA replication complexes in S. cerevisiae: redistribution of MCM proteins and Cdc45p during S phase. Cell 91:59-69[Medline].

BAKER, T. A. and S. P. BELL, 1998  Polymerases and the replisome: machines within machines. Cell 92:295-305[Medline].

BASI, G., E. SCHMID, and K. MAUNDRELL, 1993  TATA box mutations in the Schizosaccharomyces pombe nmt1 promoter affect transcription efficiency but not the transcription start point or thiamine repressibility. Gene 123:131-136[Medline].

BROSH, R. M. and S. W. MATSON, 1995  Mutations in motif II of Escherichia coli DNA helicase II render the enzyme nonfunctional in both mismatch repair and excision repair with differential effects on the unwinding reaction. J. Bacteriol. 177:5612-5621[Abstract/Free Full Text].

BROWN, G. W., P. V. JALLEPALLI, B. J. HUNEYCUTT, and T. J. KELLY, 1997  Interaction of the S phase regulator Cdc18 with cyclin-dependent kinase in fission yeast. Proc. Natl. Acad. Sci. USA 94:6142-6147[Abstract/Free Full Text].

CHONG, J. P. J., M. K. HAYASHI, M. N. SIMON, R.-M. XU, and B. STILLMAN, 2000  A double-hexamer archael minichromosome maintenance protein is an ATP-dependent DNA helicase. Proc. Natl. Acad. Sci. USA 97:1530-1535[Abstract/Free Full Text].

COUE, M., S. E. KEARSEY, and M. MECHALI, 1996  Chromatin binding, nuclear localization and phosphorylation of Xenopus cdc21 are cell-cycle dependent and associated with the control of initiation of DNA replication. EMBO J. 15:1085-1097[Medline].

COXON, A., K. MAUNDRELL, and S. E. KEARSEY, 1992  Fission yeast cdc21⁺ belongs to a family of proteins involved in an early step of chromosome replication. Nucleic Acids Res. 20:5571-5577[Abstract/Free Full Text].

DEB, S. P. and P. TEGTMEYER, 1987  ATP enhances the binding of simian virus 40 large T antigen to the origin of replication. J. Virol. 61:3649-3654[Abstract/Free Full Text].

ELLISON, V. and B. STILLMAN, 2001  Opening of the clamp. An intimate view of an ATP-driven biological machine. Cell 106:655-660[Medline].

FORSBURG, S. L. and P. NURSE, 1994  The fission yeast cdc19⁺ gene encodes a member of the MCM family of replication proteins. J. Cell Sci. 107:2779-2788[Abstract].

FORSBURG, S. L., D. A. SHERMAN, S. OTTILIE, J. R. YASUDA, and J. A. HODSON, 1997  Mutational analysis of Cdc19p, a Schizosaccharomyces pombe MCM protein. Genetics 147:1025-1041[Abstract].

FUJITA, M., T. KIYONO, Y. HAYASHI, and M. ISHIBASHI, 1997  In vivo interaction of human MCM heterohexameric complexes with chromatin. Possible involvement of ATP. J. Biol. Chem. 272:10928-10935[Abstract/Free Full Text].

FUJITA, M., C. YAMADA, T. TSURUMI, F. HANAOKA, and K. MATSUZAWA et al., 1998  Cell cycle- and chromatin binding state-dependent phosphorylation of human MCM heterohexameric complexes—a role for cdc2 kinase. J. Biol. Chem. 273:17095-17101[Abstract/Free Full Text].

HENDRICKSON, M., M. MADINE, S. DALTON, and J. GAUTIER, 1996  Phosphorylation of MCM4 by cdc2 protein kinase inhibits the activity of the minichromosome maintenance complex. Proc. Natl. Acad. Sci. USA 93:12223-12228[Abstract/Free Full Text].

HINGORANI, M. M., M. T. WASHINGTON, K. C. MOORE, and S. S. PATEL, 1997  The dTTPase mechanism of T7 DNA helicase resembles the binding change mechanism of the F1-ATPase. Proc. Natl. Acad. Sci. USA 94:5012-5017[Abstract/Free Full Text].

ISHIMI, Y., 1997  A DNA helicase activity is associated with an MCM4, -6, and -7 protein complex. J. Biol. Chem. 272:24508-24513[Abstract/Free Full Text].

ISHIMI, Y., Y. KOMAMURA, Z. YOU, and H. KIMURA, 1998  Biochemical function of mouse minichromosome maintenance 2 protein. J. Biol. Chem. 273:8369-8375[Abstract/Free Full Text].

ISHIMI, Y., Y. KOMAMURA-KOHNO, Z. YOU, A. OMORI, and M. KITAGAWA, 2000  Inhibition of Mcm4,6,7 helicase activity by phosphorylation with cyclin A/Cdk2. J. Biol. Chem. 275:16235-16241[Abstract/Free Full Text].

JAVERZAT, J. P., G. CRANSTON, and R. C. ALLSHIRE, 1996  Fission yeast genes which disrupt mitotic chromosome segregation when overexpressed. Nucleic Acids Res. 24:4676-4683[Abstract/Free Full Text].

JERUZALMI, D.