Proteolysis and Tyrosine Phosphorylation of p34 cdc2 /Cyclin B

Previously, it has been shown thatAspergillus cells lacking the function of nimQand the anaphase-promoting complex (APC) componentbimE APC1 enter mitosis without replicating DNA. Here nimQ is shown to encode an MCM2 homologue. Although mutation of nimQ MCM2 inhibits initiation of DNA replication, a few cells do enter mitosis. Cells arrested at G1/S by lack ofnimQ MCM2 contain p34 cdc2 /cyclin B, but p34 cdc2 remains tyrosine dephosphorylated, even after DNA damage. However, arrest of DNA replication using hydroxyurea followed by inactivation of nimQ MCM2 andbimE APC1 does not abrogate the S phase arrest checkpoint over mitosis. nimQ MCM2, likely via initiation of DNA replication, is therefore required to trigger tyrosine phosphorylation of p34 cdc2 during the G1to S transition, which may occur by inactivation ofnimT cdc25. Cells lacking bothnimQ MCM2 and bimE APC1are deficient in the S phase arrest checkpoint over mitosis because they lack both tyrosine phosphorylation of p34 cdc2 and the function of bimE APC1. Initiation of DNA replication, which requires nimQ MCM2, is apparently critical to switch mitotic regulation from the APC to include tyrosine phosphorylation of p34 cdc2 at G1/S. We also show that cells arrested at G1/S due to lack of nimQ MCM2 continue to replicate spindle pole bodies in the absence of DNA replication and can undergo anaphase in the absence of APC function.

During progression into G 1 and S phase, cell cycle-specific proteolysis and tyrosine phosphorylation of p34 cdc2 need to be coordinated in some way to ensure that mitosis does not occur during G 1 or before DNA replication has been completed. For instance, if the APC failed to function during late mitosis or G 1 , then accumulation of cyclin B could potentially form a complex with p34 cdc2 and, if the balance between Wee1 and Cdc25 still favored dephosphorylation of p34 cdc2 , MPF could be generated and so promote cells back into a mitotic state. Therefore, during the transition from metaphase into interphase, a coordinated inactivation of APC activity (to allow accumulation of cyclin B) and re-activation of Tyr-15 phosphorylation of p34 cdc2 (to prevent premature accumulation of MPF activity) must occur.
In the current study, we have asked how tyrosine phosphorylation of p34 cdc2 and proteolysis mediated by the APC are coordinated to allow normal progression from G 2 into M and then from M into G 1 and S phase. Our studies indicate that during G 1 the function of the APC helps prevent progression into mitosis. The function of nimQ MCM2 , a homologue of MCM2 (37) that plays a key role in licensing DNA for a single round of replication per cell cycle (38 -41), is shown to allow transfer of checkpoint control over mitotic initiation from the APC to include tyrosine phosphorylation of p34 cdc2 upon initiation of DNA replication. Our data further show that cells lacking the nimQ MCM2 function undergo multiple rounds of spindle pole body (SPB) duplication in the absence of the chromosomal cycle. The molecular mechanisms proposed help explain the logic of the cell cycle in organisms regulating mitotic initiation by tyrosine dephosphorylation of p34 cdc2 and mitotic exit via the APC.

Reciprocal Shifts and [ 3 H]Adenine
Labeling Studies-To determine the precise cell cycle arrest point of the ts Ϫ lethal nimQ20 mutation, we performed reciprocal shift experiments (42); however, instead of monitoring nuclear division, we measured DNA replication using [ 3 H]adenine to label cells of strain SWJ186 (nimQ20, choA1) adapted from Ref. 43. The experiment was reproduced two times with the same results.
Aspergillus Transformations and Libraries and cDNA Isolation-For cloning and isolation of nimQ, transformation of Aspergillus was performed as described (44). Analysis of transformants, media for propagation and genetic analysis of Aspergillus nidulans were as described previously (35,45,46). cDNA clones were isolated from a library constructed in gt10 (46). Standard procedures for recombinant DNA manipulations were used (47). DNA sequencing was performed on singlestranded templates using a Sequenase version 2.0 DNA sequencing kit (U.S. Biochemical Corp.). nimQ was cloned by complementation of the nimQ20 ts Ϫ lethal mutation, taking advantage of its previously mapped position on linkage group VII (48). The wild-type nimQ gene was isolated from a cosmid library of 468 independent LGVII-specific clones (49) obtained from the Fungal Genetics Stock Center. Cosmid DNA from five pools of approximately 95 individual clones each (provided by J. H. Doonan) were cotransformed into a nimQ20, argB2 strain (SWJ028: nimQ20, argB2, pabaA1, wA2) using the A. nidulans autonomously replicating helper plasmid pDHG25 (50) to eventually identify a single clone, W18C11, capable of fully complementing nimQ20. The complementing DNA was localized (51) to a 3.6-kb HindIII-KpnI subclone. Using this subclone as a probe, eight cDNA clones were isolated and purified. Six of the cDNAs appeared to be related, as they shared at least two out of three EcoRI restriction fragments, and the two longest cDNAs (each ϳ3.1 kb) were able to fully complement the ts Ϫ lethality of nimQ20 as phage clones or after subcloning into pBluescript KS ϩ and KS Ϫ . Site-specific integration of a cDNA into its own locus can often complement the mutant allele, whereas heterologous integration of a promoterless cDNA does not. The ability of the cDNAs to fully complement nimQ20 at high frequency (n ϭ 228 total nimQ ϩ transformants from four experiments, using approximately 5 g of DNA in each transformation) confirms that we had cloned the nimQ gene. The two remaining cDNAs appeared unrelated to the others. Restriction endonuclease mapping with nine different enzymes showed that the 3.1-kb complementing cDNA clones closely matched the 3.6-kb HindIII-KpnI genomic clone in a 1.7-kb region where they overlapped. Based upon comparison of the genomic restriction map with the cDNA sequence, the genomic clone harbors part of the gene, and is predicted to lack the C-terminal 334 amino acid residues. A stable transformant that fully complemented nimQ20 and argB2 was made by co-transforming this partial genomic clone and pDHG25 helper plasmid into a nimQ20, argB2 strain (SWJ028). Transformants were then outcrossed to a wildtype strain, and analyzed for recovery of nimQ20 ts Ϫ lethal progeny. Various other genetic markers for spore color and nutritional auxotrophies showed the expected 1:1 segregation in these crosses. However, no ts Ϫ lethal segregants were recovered from among 157 progeny, indicating tight linkage between the nimQ chromosomal locus and the integrating genomic DNA, again confirming that we had cloned nimQ.

RESULTS
nimQ Is Required for DNA Replication and Encodes an MCM2 Homologue, a Component of the DNA Replication Licensing System-Previous flow cytometric studies of nimQ20arrested conidia (uninucleate dormant spores) germinated in minimal media revealed a general defect in DNA synthesis (35). To characterize the arrest point of nimQ20 cells grown in complex media, we labeled cells in vivo with [2,8-3 H]adenine (43) and measured the synthesis of labeled DNA at 43°C. Conidial spores containing nimQ20 germinated at permissive temperature underwent three rounds of DNA synthesis and began a fourth during the 8-h experiment (Fig. 1A). By contrast, essentially no DNA synthesis occurred at the restrictive temperature (Fig. 1A). The very slight and gradual increase in DNA labeling observed is likely to be mitochondrial DNA synthesis that can occur without nuclear DNA replication during germination (53).
We also analyzed whether nimQ function was required for continued DNA synthesis after arrest of DNA replication using hydroxyurea (HU) in a reciprocal shift analysis (42). Conidia containing nimQ20 grown at 30°C (permissive temperature) in the presence of 20 mM HU were unable to synthesize DNA and were also incapable of DNA synthesis after removal of the HU and transfer to 43°C (restrictive temperature) (data not shown). The analysis indicates that nimQ is required for the initiation of DNA replication and for continued DNA replication after the HU arrest point.
We cloned nimQ by complementation and isolated a cDNA that was then sequenced (data not shown, GenBank™ accession no. AF014813). The cDNA encodes a putative open reading frame of 889 amino acids corresponding to a protein of 99.95 kDa, assuming that the first in-frame methionine is used for translational initiation. An in-frame stop codon occurs upstream of the putative initiation codon (at base pair 25), indicating that the coding region does not extend beyond the 5Ј end of the clone. A data base search revealed strong homology between the presumed nimQ polypeptide and budding yeast Mcm2 (37) and Nda1 of fission yeast (58). NIMQ showed 58.2% identity and 73.5% similarity to Mcm2 (Fig. 1B), and similar resemblance to Nda1 and also to related proteins in Drosophila (59) and human (60) (data not shown). MCM2 and its relatives encode the universal six-member MCM (minichromosome maintenance) family of proteins that form a complex near origins of replication and are required for DNA synthesis. The MCM proteins are required for initiation of DNA synthesis and have been implicated as components of the licensing factor system that limits DNA replication to once per cell cycle (39 -41, 61).
Sequence identity among NIMQ, Mcm2, and Nda1 spans most of the coding region except the N and C termini. Although the frequency of identical amino acids in the N-terminal ϳ175 amino acids is relatively low, the distribution of charged residues is well conserved. Each protein begins with a short basic region, followed in order by a region rich in proline and serine, a highly acidic domain ϳ25 residues long, and then a region of ϳ100 amino acids with a high proportion of acidic and basic residues. The novel CX 2 CX 19 CX 2 C postulated zinc finger common to Mcm2 and relatives in other species is also preserved in NIMQ (Fig. 1B). G 1 /S nimQ20-arrested Cells Are Deficient in Tyr-15 Phosphorylation of p34 cdc2 -We have shown previously that inhibition of DNA replication by HU increases the level of Tyr-15phosphorylated p34 cdc2 (13). This increase in Tyr-15phosphorylated p34 cdc2 occurs both at 32°C and 42°C after addition of HU to wild type cells ( Fig. 2A). It also occurs if DNA replication is inhibited by temperature-sensitive inactivation of nimP (Fig. 2C, nimP19 and nimP22), which encodes DNA polymerase ⑀. 2 However, this response is lacking after inhibition of DNA replication by inactivation of nimQ MCM2 or when HU is added after nimQ20 arrest ( Fig. 2A, nimQ20).
Lack of p34 cdc2 Tyr-15 phosphorylation is not due to an inability of p34 cdc2 to bind to NIME cyclin B because a com-parable amount of p34 cdc2 is co-immunoprecipitated with NIME cyclin B in cells with or without nimQ MCM2 function (Fig.  2B). In addition, this defect is not simply caused by lower levels of NIME cyclin B in cells lacking nimQ MCM2 , as high levels of NIME cyclin B accumulate in the nimQ20 ϩ bimE7 APC1 double mutant at 42°C, but p34 cdc2 still remains Tyr-15 dephosphorylated (Fig. 2E) and consequently helps promote premature mitosis from G 1 /S without DNA replication (Figs. 3 and 4C).
Tyr-15 phosphorylation of p34 cdc2 occurs after DNA damage caused by MMS (56). However, this response is defective in cells first arrested without functional nimQ (Fig. 2D). nimQ is not an essential component of the system that Tyr-15 phosphorylates p34 cdc2 (an essential activator of Wee1 for example) because cells arrested in mitosis due to lack of BIME APC1 increase the level of Tyr-15-phosphorylated p34 cdc2 in response to DNA damage equally well with, or without, the function of NIMQ MCM2 (Fig. 2E).
Mitosis Is Partially Uncoupled from DNA Replication by the nimQ20 MCM2 Mutation-Previous studies failed to reveal a mitotic phenotype associated with the nimQ20 mutation (35,48). However, germination in complex media, paying particular attention to early time points, revealed a slightly elevated mitotic index caused by the single nimQ20 MCM2 mutation germinated at 42°C and, as reported previously (35), a dramatic premature entry into mitosis in the nimQ20 MCM2 ϩ bimE7 APC1 double mutant strain (Fig. 3A).
As the slightly elevated mitotic index seen for nimQ20 MCM2 was transient (Fig. 3A), it is possible that in the double nimQ20 MCM2 ϩ bimE7 APC1 mutant lack of APC function caused by the bimE7 APC1 mutation (25,29) is trapping cells at a mitotic stage after they have prematurely progressed through the cell cycle. nimQ20 MCM2 cells were therefore germinated at 42°C in the presence of nocodazole to trap cells at the first mitosis. This treatment partially elevated the mitotic index observed for the nimQ20 MCM2 strain (Fig. 3B) but still the mitotic index fell well short of that observed in the nimQ20 MCM2 ϩ bimE7 APC1 double mutant strain (Fig. 3A). Thus, inhibition of APC function appears to markedly enhance the premature mitosis phenotype observed when nimQ MCM2 function is impaired and not just trap cells in a mitotic state. In further support of this contention, it should be noted that lack of both nimQ MCM2 and bimE APC1 promotes mitosis significantly faster than lack of either individually (Fig. 3A).
Importantly, addition of 100 mM HU was able to prevent premature mitosis in the double nimQ20 MCM2 ϩ bimE7 APC1 mutant strain if it was added at the permissive temperature to arrest cells at S phase prior to shift to the restrictive temperature. After S phase arrest caused by HU, inactivation of BIME APC1 , or a combination defect in BIME APC1 and NIMQ MCM2 , had a limited capacity to promote mitosis (Fig.  3D). This is in marked contrast to the premature mitosis induced in these strains when HU was present at the same time that BIME APC1 and NIMQ MCM2 were inactivated, during which cells will arrest initially at the nimQ20 arrest point (Fig.  3C). This demonstrates that there is a difference in mitotic regulation when DNA replication is arrested by HU (checkpoint over mitotic initiation fully engaged) compared with arrest by absence of nimQ MCM2 (checkpoint not fully engaged). The major difference we have observed is tyrosine phosphorylation of p34 cdc2 at the HU arrest point but not at the nimQ20 MCM2 arrest point (Fig. 2, A and C).
As tyrosine phosphorylation of p34 cdc2 is regulated by the opposing activities of the Wee1 tyrosine kinase and Cdc25 tyrosine phosphatase (62), lack of tyrosine-phosphorylated p34 cdc2 in nimQ20-arrested cells could be caused by elevated Cdc25 phosphatase activity or lack of Wee1 kinase activity. If the Wee1 kinase is inactive at the nimQ20 arrest point, leading to lack of tyrosine-phosphorylated p34 cdc2 , then inactivation of the Cdc25 phosphatase should not prevent entry into mitosis. However, if the Wee1 kinase is active at the nimQ20 arrest point, and p34 cdc2 is being dephosphorylated by Cdc25 to produce non-tyrosine-phosphorylated p34 cdc2 , then inactivation of Cdc25 may prevent entry into mitosis. In the latter scenario, inactivation of Cdc25 would allow Wee1 tyrosine-phosphorylated p34 cdc2 to accumulate, which would consequently prevent any premature mitosis in nimQ20-arrested cells. When a nimQ20 ϩ nimT23 cdc25 double mutant strain was germinated at 42°C, the number of cells that entered premature mitosis was markedly reduced as compared with those in the single nimQ20 mutant strain (Fig. 3E). This indicates that Wee1 is active at the nimQ20 arrest point but that nimT cdc25 has greater activity to promote dephosphorylation of p34 cdc2 and allow some entry into mitosis. These data suggest that lack of tyrosine-phosphorylated p34 cdc2 in nimQ20-arrested cells is largely due to the balance between the Cdc25 phosphatase and the Wee1 kinase favoring dephosphorylation, perhaps due to FIG. 2. nimQ20 G 1 /S-arrested cells are deficient in Tyr-15 phosphorylation of p34 cdc2 . A, Ex., exponentially growing cells at 32°C; 42°C, exponentially growing cells shifted to 42°C for 3 h; 42°C ϩ HU, exponentially growing cells shifted to 42°C for 3 h and then treated with 100 mM HU for 90 min; HU, exponentially growing cells (32°C) treated with 100 mM HU for 90 min. A wild type strain and nimQ20 strain as indicated were treated in an identical manner. B, Ex., exponentially growing nimQ20 cells at 32°C; 42°C, exponentially growing nimQ20 cells shifted to 42°C for 3 h. NIME cyclin B was immunoprecipitated using NIME cyclin B -specific antibodies. The immune complex of NIME cyclin B was then detected for NIME cyclin B and co-immunoprecipitation of p34 cdc2 by Western blotting or assayed for the associated-H1 kinase activity. C, Ex., exponentially growing cells at 32°C; 42°C, exponentially growing cells shifted to 42°C for 3 h. nimP mutant alleles were used as indicated. D, Ex. and 42°C and HU sample treatments are described in A; MMS, exponentially growing cells (32°C) treated with 0.04% MMS for 90 min; 42°C ϩ MMS, exponentially growing cells shifted to 42°C for 3 h and then treated with 0.04% MMS for 90 min. E, Ex. and 42°C sample treatments are described in A, and 42°C ϩ MMS as in C. To determine the level of tyrosine-phosphorylated p34 cdc2 , p34 cdc2 was first immunoprecipitated from total protein extracts derived from cells described above using affinity-purified antisera raised against a C-terminal peptide of A. nidulans NIMX cdc2 (53). The level of Tyr-15 phosphorylation of p34 cdc2 was then detected by Western blotting using a monoclonal anti-phosphotyrosine antibody. After phosphotyrosine (P-Tyr) detection, the blot was stripped and then detected for NIMX cdc2 using the NIMX cdc2 -specific antiserum (E-77).
high Cdc25 phosphatase activity still being present, as it is known to be activated at the G 2 /M transition.
Initiation of mitosis in A. nidulans requires activation of both p34 cdc2 /cyclin B H1 kinase and the NIMA kinase. To determine if NIMA kinase is also required for the premature mitosis in the nimQ20-arrested cells, we generated a nimQ20 ϩ nimA5 double mutant strain and examined the ability of the double mutant cells to initiate premature mitosis when germinated at 42°C. As shown in Fig. 3E, lack of NIMA function prevented premature mitosis in the nimQ20-arrested cells, demonstrating that NIMA kinase is required for the low level of premature mitosis at G 1 /S seen in the nimQ20-arrested cells.
Anaphase Can Occur without APC Function in the Absence of nimQ MCM2 -DAPI staining of mitotic cells seen when nimQ20 MCM2 conidia were germinated at 42°C revealed that many (55%, Fig. 4B) had progressed into anaphase and contained segregated DNA masses, some of which were not equal in intensity (data not shown, but see Fig. 4A, c and e). Addition of HU did not affect anaphase progression in this strain, but microtubule depolymerization utilizing nocodazole completely suppressed DNA segregation, indicating it to be mitotically driven (Fig. 4B). Progression into anaphase was similarly observed (60%) in the double nimQ20 MCM2 ϩ bimE7 APC1 mutant strain with or without HU addition (Fig. 4A, c and e; and B) but was totally absent in the single bimE7 APC1 strain (Fig. 4A, a  and b; and B). The actual percentage of mitotic cells progressing into anaphase measured in these experiments is conservative as visualization of DNA segregation is dependent upon the orientation of DNA within individual cells. Inactivation of FIG. 3. Cells lacking NIMQ MCM2 have a limited ability to initiate mitosis but cells lacking both NIMQ MCM2 and BIME APC1 initiate precocious mitosis in the absence of DNA replication. A, chromosome mitotic index (CMI%) of nimQ20, bimE7 and nimQ20 ϩ bimE7 mutant cells germinated at 32°C (the permissive temperature) and at 42°C (the restrictive temperature). Samples were taken at the time points indicated, fixed in glutaraldehyde, and stained with DAPI. CMI represents percentage of cells containing condensed mitotic chromatin. As the bimE7 single and the nimQ20 ϩ bimE7 double mutant cells had very similar chromosome mitotic indexes when germinated at 32°C, the chromosome mitotic index for the bimE7 cells is shown only. B, chromosome mitotic index of a wild type and the nimQ20 mutant strains germinated at 42°C in the presence and absence of 5 g/ml nocodazole. Nocodazole was added to trap cells in mitosis. C, a time course of chromosome mitotic index (CMI%) of nimQ20, bimE7, and nimQ20 ϩ bimE7 mutant cells germinated at 42°C in the presence of 100 mM HU. D, chromosome mitotic index (CMI%) of bimE7 and nimQ20 ϩ bimE7 mutant cells, germinated at 32°C for 7 h and then treated with or without HU at 32°C for 2 h before temperature shift to 42°C to inactivate NIMQ MCM2 and BIME APC1 . Note the synergy between lack of NIMQ MCM2 and lack of BIME APC1 in promoting mitosis when HU was added to germinating cells at 42°C (C, q, nimQ20 ϩ bimE7 ϩ HU) but the lack of synergy when HU was added first to cells at 32°C to arrest S phase before shifting to 42°C to inactivate NIMQ MCM2 and BIME APC1 (D, ‚, nimQ20 ϩbimE7). E, chromosome mitotic index (CMI%) of nimQ20, nimQ20 ϩ nimT23 and nimQ20 ϩ nimA5 double mutant cells germinated at 42°C. nimQ MCM2 therefore allows anaphase progression in the apparent absence of APC function.
Absence of NIMQ MCM2 function could promote anaphase in the double nimQ20 MCM2 ϩ bimE7 APC1 mutant strain by restoring APC function. Alternatively, absence of NIMQ MCM2 could bypass the requirement for the APC to promote anaphase. We therefore assayed p34 cdc2 H1 kinase activity and NIME cyclin B levels in bimE7 APC1 , nimQ20 MCM2 , and double mutant strains germinated for 6 h at 42°C (Fig. 4C). Levels of NIME cyclin B and H1 kinase activity were equally elevated in both the bimE7 APC1 and nimQ20 MCM2 ϩ bimE7 APC1 strains. This suggests that lack of nimQ MCM2 function when bimE7 APC1 is inactivated does not allow activation of the APC, because NIME cyclin B levels are elevated and cells are arrested in mitosis. Instead, the data indicate that lack of nimQ MCM2 bypasses the requirement for the APC to promote anaphase.
Anaphase progression without APC function, as seen in the double nimQ20 MCM2 ϩ bimE7 APC1 strain, is most likely due to the absence of sister chromatids as without sister chromatid cohesion a reductional anaphase has previously been observed in S. cerevisiae cdc6 mutants that fail to enter S phase but go into mitosis (63) and also occurs in a cdc6 mutant lacking APC function (64).
The SPB Duplication Cycle Is Uncoupled from the DNA Replication Cycle in nimQ20-arrested Cells-S phase arrest caused by the nimQ20 MCM2 mutation was also remarkable because this arrest allowed uncontrolled duplication of spindle pole bodies. Arrest in S phase or G 2 normally completely prevents spindle formation and SPB re-duplication (Fig. 5). For example, conidia of nimP22, nimA5, and nimT23 containing strains germinated at 42°C totally prevents spindle formation as judged by IIF using ␣-tubulin staining and limits SPB duplication to one round of replication (Fig. 5, shown for nimP22 POL⑀ in Fig.  6). Under the same conditions, 10% of the nimQ20 MCM2 cells displayed mitotic spindles (Fig. 5A), many of which contained multiple spindle pole bodies giving rise to triangular spindles, splayed spindles, and other spindle morphological novelties (Fig. 6, nimQ20 and data not shown). An additional 20% of these cells had other less clearly defined abnormalities typified by thick microtubule bundles (data not shown).
Using indirect immunofluorescence to visualize ␥-tubulin, the number of SPBs (57,65) in the nimQ20 MCM2 strain was significantly greater after germination at 42°C than for the S and G 2 arresting nim mutants (Fig. 5B). In addition, the number of SPBs in the nimQ20 MCM2 strain was higher than in a wild type strain. For the non-mutant strain, the average number of SPBs per cell was 1.81, but for the nimQ20 MCM2 strain, each cell had on average 2.78 SPBs after 6 h germination at 42°C (Fig. 5B). This suggests not only that the SPB cycle is uncoupled from the DNA replication cycle in the nimQ20 MCM2arrested cells but that the SPB replication cycle proceeds significantly faster than normal. DISCUSSION The nimQ20 MCM2 mutation joins a group of mutations that interfere with temporal control of S phase and mitosis, which HU or nocodazole on DNA segregation, the nimQ20 and the nimQ20 ϩ bimE7 mutant cells were also germinated in the presence of 100 mM HU and 5 g/ml nocodazole, respectively. C, p34 cdc2 H1 kinase and levels of NIME cyclin B in bimE7, nimQ20, and nimQ20 ϩ bimE7 mutant cells 6 h after germination at 32°C or at 42°C. After immunoprecipitation with NIMX cdc2 -specific antiserum, the p34 cdc2 H1 kinase activity was assayed using histone H1 as substrate in the presence of [␥-32 P]ATP and was detected by autoradiography. NIME cyclin B was detected by ECL following Western blotting of 100 g total protein extract using anti-NIME cyclin B -specific antibodies (E-8). After NIME cyclin B detection, the blot was stripped and then detected for NIMX cdc2 by ECL with NIMX cdc2 (E-77)-specific antibodies. Conidiospores derived from nimQ20, bimE7, and nimQ20 ϩ bimE7 mutant strains were germinated at 42°C for 6 h, and then fixed and stained with DAPI. Percentages of mitotic cells with segregated DNA were then determined by fluorescence microscopy. To see the effect of also play a role in initiation of DNA replication (63, 66 -71). These observations have led to the proposal that the act of initiating DNA replication activates a negative control system preventing initiation of mitosis during S phase (69,72).
The nature of this negative control system has not been demonstrated previously. Here we demonstrate that upon execution of nimQ20 MCM2 function Tyr-15 phosphorylation of p34 cdc2 becomes activated to help prevent mitotic initiation during S phase. Thus, lack of nimQ MCM2 not only prevents DNA replication, it also prevents Tyr-15 phosphorylation of p34 cdc2 . By contrast, S phase arrest caused by mutation of the nimP pol⑀ gene or by HU (at which initiation complexes would be activated but immediately stalled due to lack of deoxyribonucleotides) allows Tyr-15 phosphorylation of p34 cdc2 , and after the HU arrest point, there is no longer a requirement for nimQ MCM2 in the Tyr-15 phosphorylation of p34 cdc2 . The function of nimQ MCM2 therefore acts like a trigger to allow Tyr-15 phosphorylation of p34 cdc2 upon initiation of DNA replication, after which it is not required for Tyr-15 phosphorylation of p34 cdc2 .
This result provides a molecular explanation for why mitosis can be initiated before DNA replication in the double nimQ20 ϩ bimE7 mutant strain of A. nidulans (35). We have recently demonstrated that arrest of mitotic initiation after S phase in A. nidulans is mediated by the combined actions of bimE APC1 and Tyr-15 phosphorylation of p34 cdc2 (13). Therefore, any treatment interfering with both APC function and the ability to Tyr-15 phosphorylate p34 cdc2 will promote lethal premature mitosis in A. nidulans. As lack of nimQ MCM2 leads to S phase arrest during which p34 cdc2 cannot be tyrosine-phosphorylated, the double nimQ20 ϩ bimE7 mutations in essence phenocopy the p34 cdc2 AF ϩ bimE7 double mutant when DNA replication is inhibited. By preventing the transition from G 1 into S phase FIG. 5. Indirect immunofluorescent microscopy of microtubules and SPBs. A, spindle mitotic index (SMI%). The spindle mitotic index values represent percentages of cells containing mitotic spindles as visualized by indirect immunofluorescent microscopy after immunostaining with a monoclonal antibody against ␣-tubulin. All cells were at 6 h after germination at 42°C. B, average number of spindle pole bodies per cell at 6 h after germination at 42°C. This experiment was repeated four times with at least 500 cells observed for each determination. The SPBs were visualized by immunostaining of ␥-tubulin with affinitypurified antibodies raised against the bacterially expressed A. nidulans ␥-tubulin as described by Oakley et al. (57). It has previously been shown by electron microscopy that nimA5 G 2 -arrested cells have duplicated but not separated SPBs (88). at a stage that blocks activation of Tyr-15 phosphorylation of p34 cdc2 (inactivate nimQ MCM2 ) and also inactivating APC function (bimE7 APC1 mutation), negative controls over MPF, and other mitotic regulators such as NIMA (13), are removed and mitosis is prematurely initiated from G 1 /S.
Coordination of APC and Tyr-15 Phosphorylation of p34 cdc2 -At some point during progression out of mitosis into interphase, the APC needs to be inactivated (it is activated to permit exit from mitosis) to allow mitotic cyclin accumulation for the next mitosis. Additionally, Tyr-15 phosphorylation of p34 cdc2 needs to be turned back on (it is turned off after activation of Cdc25 at G 2 /M) to prevent premature activation of pre-MPF during interphase. We show that Tyr-15 phosphorylation of p34 cdc2 during G 1 /S progression occurs after the function of nimQ MCM2 , which may involve down-regulation of Cdc25, which is activated during G 2 /M of the previous cycle. Because APC activity is known to remain active during G 1 in both yeast (73) and human cells (74), it is likely that cyclin B cannot normally accumulate and bind p34 cdc2 in G 1 of many cell types and so they have no need to negatively regulate p34 cdc2 by Tyr-15 phosphorylation during this cell cycle stage. However, as soon as APC activity is reduced during the G 1 /S transition, and cyclin B begins to accumulate in preparation for the next mitosis, it becomes critical to prevent production of MPF, which is achieved by allowing Tyr-15 phosphorylation of p34 cdc2 . A transition thus occurs, during which negative control over mitotic p34 cdc2 is transferred from the APC in G 1 to include the Tyr-15 phosphorylation of p34 cdc2 in S phase. Here we show that this transition is dependent upon nimQ MCM2 function.
The requirement for nimQ MCM2 to trigger Tyr-15 phosphorylation of p34 cdc2 at G 1 /S is likely to involve its capacity to promote the initiation of DNA replication as nimQ MCM2 is not required for Tyr-15 phosphorylation of p34 cdc2 after DNA replication has been initiated (for instance, at the HU-arrest point). Additionally, as mentioned before, nimQ20 is not unique as numerous mutations have been identified that interfere with temporal control of S phase and mitosis, which also play a role in the initiation of DNA replication (63, 66 -71).
Tightly coupling Tyr-15 phosphorylation of p34 cdc2 to initiation of DNA replication ensures that mitosis is not promoted prematurely during S phase. We further propose that APC inactivation may occur only after Tyr-15 phosphorylation of p34 cdc2 is activated to ensure that mitosis is not initiated in G 1 by accumulation of cyclin B (Fig. 7).
Work using Schizosaccharomyces pombe has also indicated the existence of another level of control over mitotic p34 cdc2 during G 1 because p34 cdc2 is tyrosine-dephosphorylated during M/G 1 until execution of Start (75). This pre-Start checkpoint was suggested to act in an undefined manner by preventing formation of the p34 cdc2 p56 cdc13 complex. Our studies partially support such a mechanism but, importantly, more directly implicate the APC in the pre-Start, or pre-DNA replication, negative control of p34 cdc2 . It is worth pointing out that at the nimQ20 arrest point some cyclin B protein is present and significant H1 kinase activity exists (Fig. 4C). This likely helps explain why some nimQ20-arrested cells enter a mitotic state from G 1 /S. However, only upon inactivation of the APC by the bimE7 mutation do nimQ20-arrested cells increase cyclin B protein levels and H1 kinase activity and so overwhelmingly and rapidly enter mitosis (Fig. 4C). The possibility exists that BIME APC1 also negatively regulates the NIMA kinase as NIMA is also required to promote premature mitosis in the nimQ20arrested cells and it is unstable during exit from mitosis (76). We have shown previously that tyrosine-dephosphorylated p34 cdc2 in combination with lack of BIME function leads to deregulation and activation of NIMA in S phase, a point in the cell cycle when NIMA is normally inactive (13).
A Role for nimQ MCM2 in Spindle Pole Body Duplication?-The DNA licensing hypothesis proposed that DNA licensing factors, being essential for initiation of DNA replication, were irreversibly changed upon initiation of DNA replication and could not be replenished to promote DNA replication again until passage through mitosis (38). Based upon movement of MCM proteins in and out of the yeast nucleus during the cell cycle (77)(78)(79) and MCM protein association with DNA and phosphorylation state during the cell cycle (60, 71, 80 -84), this family of proteins (including Mcm2) have been suggested as candidates for licensing factor. Purification of licensing factor from Xenopus egg extracts identified two complexes having licensing activity (85,86). One of these (RFL-M) contains MCM homologues, further supporting the role of this class of protein in the regulation of DNA replication to one round per cell cycle. Our data additionally suggest a potential role for nimQ MCM2 in the regulation of SPB duplication. Inactivation of nimQ MCM2 prevents DNA replication, which normally prevents SPB reduplication (Fig. 5B). However, SPB duplication is uncoupled from these normal controls in the nimQ20 MCM2 -arrested cells, leading to the production of multiple SPBs in the absence of DNA replication. FIG. 7. Coordination between the APC and tyrosine phosphorylation of p34 cdc2 . A model showing the coordination of Tyr-15 phosphorylation of p34 cdc2 and the activity of the APC through the cell cycle. During S phase and G 2 , Tyr-15 phosphorylation of p34 cdc2 plays a major role preventing mitotic initiation. During the G 2 /M transition, p34 cdc2 is rapidly tyrosine-dephosphorylated to promote mitosis. After cells progress into mitosis, the APC is activated to promote anaphase and mitotic exit by targeting proteins for degradation through polyubiquitination. APC activity persists through G 1 phase (73,74), thus keeping the activities of the mitosis-promoting factors low. During the G 1 /S transition, checkpoint control over mitotic initiation is transferred from the APC to include Tyr-15 phosphorylation of p34 cdc2 , as revealed by a defect in the nimQ MCM2 gene, which is required to execute the initiation of DNA replication. Once Tyr-15 phosphorylation of p34 cdc2 is activated, the activity of the APC can decline to allow accumulation of cyclin B in preparation for mitosis once again.
The elevated numbers of SPBs generated when nimQ MCM2 is inactivated were functional because they were capable of nucleating microtubules leading to the generation of tripolar and other multipolar spindles. Indeed, at the resolution of the light microscope, many of the excess SPBs observed were similar in size to normal SPBs. A similar phenotype has been reported in mouse embryonic fibroblasts lacking the p53 tumor suppressor protein. Cells containing multiple centrosomes forming tripolar and other abnormal spindles were observed in cultured mouse cells lacking p53 (87). The defects observed in mouse cells lacking p53 are remarkably similar, in relationship to microtubule organizing centers, to what we have observed for A. nidulans cells lacking the function of an Mcm2 homologue. The potential relationship between the roles of p53 and Mcm2 in microtubule organizing center duplication will be of interest in higher eukaryotic cells.
It is intriguing that a function known to be involved in the regulation of DNA replication may also play a key role in the regulation of SPB duplication. Perhaps these two levels of regulation are linked in some way via the Mcm2 function to ensure normal coupling of the DNA replication cycle and the SPB duplication cycles. However, in the present study it is not certain how Mcm2, in addition to its role in initiation of DNA replication, regulates SPB duplication. Regardless of what potential roles Mcm2 may play in the regulation of SPB duplication, the uncoupling of the SPB from the chromosomal cycle in cells lacking Mcm2 could help us understand how these two cycles are normally coordinated during the cell cycle.
Conclusions-We present a model whereby control of mitosis during the G 1 to S transition is transferred from the APC in G 1 to include the Tyr-15 phosphorylation of p34 cdc2 after initiation of DNA replication (Fig. 7). This transition from APC-controlled inhibition of mitosis to inhibition by Tyr-15 phosphorylation of p34 cdc2 was revealed through examination of a mutation in the nimQ MCM2 gene, which is required for initiation of DNA synthesis. We also show that the SPB duplication cycle is uncoupled from the chromosomal cycle in cells lacking nimQ MCM2 , suggesting a potential role for Mcm2 in the control of SPB duplication. The data provide a molecular logic for negative control of mitosis during G 1 /S, which is likely to be applicable to other eukaryotes that regulate mitotic exit via the APC and mitotic entry by Tyr-15 dephosphorylation of p34 cdc2 .