Roles for ORC in M Phase and S Phase

See allHide authors and affiliations

Science  13 Mar 1998:
Vol. 279, Issue 5357, pp. 1733-1737
DOI: 10.1126/science.279.5357.1733


The origin recognition complex (ORC), a six-subunit protein, functions as the replication initiator in the yeast Saccharomyces cerevisiae. Initiation depends on the assembly of the prereplication complex in late M phase and activation in S phase. One subunit of ORC, Orc5p, was required at G1/S and in early M phase. Asynchronous cells with a temperature-sensitiveorc5-1 allele arrested in early M phase. In contrast, cells that were first synchronized in M phase, shifted to the restrictive temperature, and then released from the block arrested at the G1/S boundary. The G1/S arrest phenotype could not be suppressed by introducing wild-type Orc5p during G1. Although all orc2 and orc5 mutations were recessive in the conventional sense, this dominant phenotype was shared with other orc5 alleles and an orc2 allele. The dominant inhibition to cell-cycle progression exhibited by theorc mutants was restricted to the nucleus, suggesting that chromosomes with mutant ORC complexes were capable of sending a signal that blocked initiation on chromosomes containing functional origins.

In Saccharomyces cerevisiae, replication initiates from specific DNA sequences called autonomous replication sequences (ARSs), many of which have proven to be chromosomal origins of replication. A six-subunit protein complex, the ORC, binds to ARSs in an ATP-dependent manner and is required for initiation (1-3). Homologs of ORC subunits have been identified in other eukaryotes including humans, suggesting that the mechanism by which ORC initiates replication is highly conserved (4). Additionally, these homologs are essential for in vitro replication of Xenopus egg extracts (5-7) and for amplification of the chorion gene cluster of Drosophila (8).

Eukaryotes can initiate replication at a given origin only once per cell cycle, and there are hundreds of origins whose activation is coordinated. Some origins initiate early and others initiate late in S phase. ORC remains bound to origins throughout the cell cycle (9), so initiation of replication, at least in S. cerevisiae, is not regulated by simply controlling the binding of ORC to origins.

Origin initiation is regulated by a two-step mechanism. The first step, referred to historically as origin licensing, occurs in M phase, and the second step, origin activation, occurs in S phase (10). Once a licensed origin has been activated during S phase, it is incapable of initiating again until it is licensed in the next M phase. The factors that make up a licensed origin are not diffusible, because otherwise these factors could diffuse from a late origin that had not initiated to a recently initiated early origin, allowing its reinitiation.

ORC has properties consistent with its being the target of factors that control initiation. From S phase to late M phase, the in vivo footprint at an origin is similar to the footprint created in vitro by purified ORC, suggesting that ORC is the only factor bound during this time. During late M phase, when origins are licensed, the footprint is extended, reflecting the assembly of a prereplication complex (pre-RC) (9). Genetic and molecular data suggest that the pre-RC contains ORC, Cdc6p, and the MCM family of proteins (11-13). Mutation of the MCM genes causes defects in minichromosome propagation (14), and MCM homologs in Xenopus are components of licensing factor (6, 7, 15). The Xenopus MCM proteins bind chromatin in an ORC-dependent manner, suggesting that ORC interactions with MCM proteins make an origin competent to initiate replication (6).

This study revealed that at least one subunit of the ORC complex, Orc5p, was required for at least two steps in the cell cycle, G1/S and early M phase. The execution point for the G1/S function occurred before Start, probably in M phase. Moreover, the level of ORC5 function required for entry into M phase was higher than the level of function required for entry into S phase.

Strains with either orc2-1 or orc5-1 recessive mutations are compromised for replication initiation, but differ in their arrest point in the cell cycle upon shift to the nonpermissive temperature. Haploid cells with a defective Orc2 protein (Orc2p) arrest with 1C DNA content (16). In contrast, orc5-1mutant cells arrest with an apparent 2C DNA content, suggesting that these cells are in G2 or M phase of the cell cycle (12) (Fig. 1) or arrested in late S phase, with most of the genome replicated. A late S phase arrest would suggest that the orc5-1 allele was defective at a small subset of origins of replication, perhaps the late-initiating ones. Arrest after S phase would argue that the function of Orc5p was not confined to S phase. To distinguish between these possibilities, we examined the arrest phenotype of orc5-1 mutant cells.

Figure 1

Arrest of orc5 ts strains with a 2C DNA content. Cells grown at 37°C for 0, 4, or 6 hours were harvested and subjected to FACScan analysis to assess cellular DNA content. All 25 alleles were tested; 7 representative alleles are shown.

We performed pulsed-field gel electrophoresis (PFGE) on the chromosomal DNA of arrested orc5-1 cells (17). Chromosomal DNA of wild-type (WT) cells arrested in S phase by treatment with hydroxyurea does not enter the gel matrix (18). In contrast, the fully replicated chromosomal DNA of cells arrested either in G1 by α-factor or in M phase by nocodazole entered the gel and migrated in a characteristic manner. By this assay,orc5-1 cells arrested with a 2C DNA content and contained fully replicated chromosomal DNA similar to that of orc5-1cells grown at 23°C and to that of WT cells grown at either 23° or 37°C (Fig. 2).

Figure 2

PFGE analysis of orc5-1 cells. The chromosomes of cells in S phase do not enter the gel matrix during PFGE (hydroxyurea-treated cells, lane 2). The chromosomes of cells arrested in G1 (α-factor–treated cells, lane 1) or M phase (nocodazole-treated cells, lane 3) enter the gel matrix. The chromosomes of orc5-1 cells (JRY4253) grown at 37°C for up to 10 hours migrated identically to those of WT cells (JRY3009) (35).

The RAD9 checkpoint pathway is activated in response to small amounts of unreplicated and damaged DNA. For example, both DNA ligase (CDC9) and DNA polymerase α (CDC2) and γ (CDC17) mutant cells arrest in G2, and this arrest depends on RAD9 (19). Double mutants containing the rad9Δ and cdc9, cdc2, or cdc17 mutations lose viability rapidly at the nonpermissive temperature and form microcolonies of dead cells. To test if the orc5-1 arrest required the RAD9checkpoint, we analyzed the viability of an orc5-1 rad9Δ double mutant (20). There was no enhanced lethality in theorc5-1 rad9Δ double mutant (21). Moreover, after 4 days at 37°C, the double mutants remained as single, large budded cells, indicating that the arrest of orc5-1 cells was independent of RAD9 (22). We also tested whether other checkpoint genes involved in monitoring DNA metabolism were required for the orc5-1 arrest. Combining themec1-1, mec2-1, or rad17Δ mutations with orc5-1 also did not alter the viability oforc5-1 strains (21). Thus, orc5-1cells appeared to arrest after S phase with fully replicated chromosomes.

Activity of the Cdc28 protein peaks at Start and rapidly declines upon entry into S phase, then starts to appear again during S phase, finally peaking shortly before entry into anaphase. Cells arrested before M phase have low Cdc28p kinase activity, whereas cells arrested in M phase before anaphase have high activity (23). After 6 hours, orc5-1 cells had Cdc28p activity equivalent to that of cells arrested in early M phase with nocodazole at the permissive temperature (Fig. 3) (24). These mutant cells arrested with a single nucleus positioned at the bud neck and a short spindle that spanned the nucleus (22). Because orc5-1 cells arrest at the restrictive temperature with fully replicated chromosomes, a high Cdc28p kinase activity, and short nuclear spindles less than half the length of anaphase spindles (22), they appear to be arrested in early M phase.

Figure 3

Cdc28p kinase activity inorc5-1 cells. (A) orc5-1 cells (JRY4253) were shifted to the nonpermissive temperature and samples, harvested at the indicated times, were prepared for assay of Cdc28p kinase (24). Control cells (JRY4253) were arrested in G1 (α-factor), S (HU), or M phase (Nocodazole) at the permissive temperature. (B) Quantitation of the gel in (A). Relative Cdc28 kinase activity was determined by Imagequant Software. (Inset) The control lanes from (A) are represented as a bar graph.

The M phase arrest phenotype of orc5-1 was unexpected given the central role of ORC in the initiation of replication. We therefore tested whether this arrest phenotype was unique to theorc5-1 allele by analyzing the arrest of 20 unique temperature-sensitive (ts) alleles (25). At the nonpermissive temperature, all alleles caused the cells to arrest with a single, large bud after 4 hours at 37°C (22) with a 2C DNA content, with some alleles showing a stronger arrest than others (Fig. 1) (26). Additionally, these alleles were distinct from the original orc5-1 allele because they showed no silencing defect (22). Thus, the M phase arrest phenotype oforc5-1 represented a loss of ORC5 function in this region of the cell cycle and was not allele specific.

We examined whether Orc5p might have a second essential role in the cell cycle. If Orc5p function were required only in early M phase, then orc5-1 mutant cells would arrest only in M phase regardless of how ORC5 function was inactivated. We synchronized orc5-1 cells in M phase with nocodazole at the permissive temperature, and then shifted the cells to the restrictive temperature while maintaining the M phase block. The cells were then released from the block and monitored for cell cycle progression (27). Single unbudded cells containing 1C DNA appeared 40 min after release from the nocodazole block, indicating that Orc5p function was not required for exit from mitosis. Cells were able to pass Start, as indicated by the emergence of buds. However, the cells arrested as large budded cells with 1C DNA content (Fig.4). Thus, Orc5p function was required to begin S phase.

Figure 4

Arrest of orc5-1cells with 1C DNA content. orc5-1 cells (JRY4253) were arrested in M phase at 23°C, then shifted to 37°C for 2 hours while maintaining the M phase block. Cells were released from the block into prewarmed 37°C medium, and samples were harvested at the time points indicated. Cytology of the arrested cells is shown to the left of the corresponding FACScan profile.

To test whether these cells arrested completely before S phase or whether they entered S phase, we analyzed the chromosomal DNA of cells from the 200-min time point of Fig. 4 by PFGE. The chromosomal DNA from these cells was indistinguishable from that of WT cells arrested in G1 (22). Therefore, cells released from M phase with inactivated Orc5p traversed Start but arrested before the onset of S phase, consistent with a second requirement for Orc5p in the cell cycle, at the G1/S boundary.

Because orc5-1 is recessive, one would expect that addition of WT Orc5p to orc5-1 cells arrested at the G1/S boundary would restore replication if the pre-RC can assemble at any point before the beginning of S phase. However, if there is a critical point for forming the pre-RC before S phase, then injection of WT Orc5p might not rescue the replication block. We tested these models by using mating itself, which occurs at Start in G1, to introduce Orc5p protein genetically into anorc5-1 mutant (Fig. 5A). Theorc5-1 cells were arrested in M phase with nocodazole and shifted to the restrictive temperature. The cells were then released from the M phase block and mixed with asynchronous WT cells of the opposite mating type at the restrictive temperature. After 4 hours the mixture was plated onto medium that selected for diploids at the restrictive temperature. A parallel experiment was also performed in which all incubations were at 23°C, including the M phase arrest (28). Both the MAT α orc5-1 andORC5 strains could form diploids at 23°C with theMAT a WT strains, indicating that nocodazole arrest did not affect mating. In contrast, at 37°C, the orc5-1 strain failed to form diploid colonies when mated with the WT strain, whereas the ORC5 strain did (Fig. 5B). We isolated 11 zygotes during the mating reaction by micromanipulation and followed their progression by microscopy at 37°C. All 11 zygotes derived from theorc5-1 cross formed a single, large bud that did not divide, even after 4 days at 37°C on rich medium, which could support the growth of either haploid parent strain (22). This dominant arrest phenotype was not allele-specific because eitherorc5-2 or orc5-3 mutant cells behaved identically to orc5-1 cells (22). Thus, cells lacking Orc5p from M to G1, when fused with a WT cell, inhibited cell cycle progression of the resulting heterozygous diploid cell in a dominant manner.

Figure 5

Failure of addition of WT ORC by mating to rescue the G1 arrest of orc cells. (A) Diagram of the experimental procedure and possible results. orc5-1 cells were arrested in M phase at 23°C, shifted to 37°C for 2 hours, and then released from the M phase block at 37°C with a mixture of WT cells of the opposite mating type. Control cultures were incubated at 23°C throughout all stages. Solid nucleus represents an orc5-1 nucleus that was inactivated at the restrictive temperature in M phase; clear nucleus represents a WT ORC5nucleus; striped nucleus represents a nucleus containing chromosomes from an orc5-1 and an ORC5 nucleus. The growth of diploid colonies would indicate rescue of the replication defect by WT Orc5p added in G1 (Outcome 1). The growth of haploid colonies containing only the chromosomes contributed by theORC5 nucleus would suggest no rescue by Orc5p and cis-dominant inhibition of the chromosomes contributed by theorc5-1 parent (Outcome 2). Absence of colonies would suggest trans-dominant inhibition by the chromosomes from theorc5-1 strain on the replication or inheritance of the chromosomes from the ORC5 parent (Outcome 3). (B) Nocodazole-arrested orc5-1 (JRY5493) ororc2-1 (JRY4503) cells incubated at 23°C, but not at 37°C, were able to form diploid colonies when mated to a WT strain (JRY2334 and JRY4012, respectively). Additionally, WT cells (JRY3009xJRY4012) formed diploids at 23°C at a much higher frequency than at 37°C, due to the temperature sensitivity of karyogamy (30).

We repeated the mating experiment with a strain carrying theorc2-1 allele. The orc2-1 strain also formed diploids when mated with the WT strain at 23°C, but not at 37°C (Fig. 5B). Taken together, these results indicated that inactivation of ORC in M phase could not be rescued by addition of WT ORC in G1. We also repeated the mating experiments with either acdc6-1 or a mcm5 (cdc46-1) mutant strain. Either strain could form viable diploids when mated with WT cells at 23° and at 37°C (22). Therefore, this phenotype was ORC specific, as other cdc mutants involved in replication initiation were rescued in G1 by mating with a WT strain.

We examined whether the dominant inhibitory effect of ORC mutations resulted from a defect restricted to the nucleus, such as inactivated origins of replication, or whether it was a property distributed throughout the cell. The mating experiments were repeated as before, this time with a strain containing the kar1-1mutation, which blocks nuclear fusion (29). A cell that mates with a kar1-1 mutant cell forms a zygote with two unfused nuclei, one from the WT strain and the other from thekar1-1 strain. The first bud from the zygote typically contains only one haploid nucleus, resulting in a haploid colony containing the cytoplasmic material from both strains but the nuclear contribution from only one parent (30). If the dominant effect of orc5-1 were restricted to the nucleus containing the inactivated ORC, then haploid buds with the WT nucleus would be propagated (31).

Indeed, haploid colonies having the ORC5 kar1-1 nucleus were produced from a cross with the orc5-1 strain at the restrictive temperature with approximately equal frequency to that in crosses with the WT strain (Fig. 6) (28). Thus, the dominant effect on cell-cycle progression of an orc5-1 mutant resulted from a property restricted to the nucleus.

Figure 6

Requirement for nuclear fusion fororc5-1 cells to inhibit cell cycle progression. At 37°C,orc5-1 KAR1 (JRY4249) and WT KAR1(JRY3009) cells formed approximately the same number of viable colonies containing only the nucleus of the kar1-1 strain (JRY5449). Growth of cells on glycerol and cyclohexamide-containing medium required the presence of functional mitochondria, from theKAR1 parent, and the recessivecyh r nuclear gene, from thekar1-1 parent (31).

Because ORC is the yeast replicator, it was not unexpected that one termination point of orc5 mutants was at the G1/S transition. Our data suggest that the execution point for the G1/S function is between the nocodazole block in M phase and the Start point in G1. orc5-1 mutants can recover from arrest at the restrictive temperature in mitosis but not in G1 or S phase (12). Moreover, introduction of WT ORC in G1 failed to rescue the cell cycle defect oforc5-1 cells released from M phase at the nonpermissive temperature. This execution point matches closely the time at which the pre-RC is assembled. In contrast, functional Cdc6p and Mcm5p proteins, which are also part of the pre-RC, could be introduced as late as G1 to support the next cell cycle (22). Similarly, Cdc6p can promote entry into S phase when expression is activated after Start but before activation of cyclin-B/Cdc28p (32).

The G1/S arrest of orc5 andorc2 mutants could not be rescued by introducing WT Orc5p or Orc2p in G1 in a mating reaction. Although we have no direct measure of where in the cell cycle these cells arrested, the simplest model is that the arrest occurred at G1/S, as in the orc5-1 parent. However, we cannot rule out the possibility that these cells are arrested in early M phase. It is unlikely that the arrested cell cycle reflects insufficient amounts of ORC to complete a cell cycle, for two reasons. First, overexpression ofORC5 in the WT nucleus during the mating reaction failed to suppress the lethality (22). Second, diploid cells with a single ORC5 gene divide with normal rates, indicating that Orc5p is present in excess. The inactive Orc5-1p, and, by inference, origins with inactivated ORC complexes, may send a signal that blocks initiation at origins with functional ORC complexes. This signal appears to be restricted to the nucleus because the dominant inhibition caused by orc5-1 did not block the ORC5 haploid nucleus from the kar1-1 cell from forming a fully functional daughter cell when nuclear fusion was blocked.

The G1/S phase arrest of orc5-1 cells was detectable only if the cells were first arrested in M phase and then shifted to the nonpermissive temperature for 2 hours before release from the mitotic block. Thus, the G1/S function of Orc5p required a lower level of activity than its M phase function.

There are several possible roles for ORC in M phase of chromosome segregation. Proper segregation of chromosomes requires that sister chromatids remain paired until the end of metaphase, that chromosomes are attached to a functional spindle, and that chromosomes condense. The positioning of ORC along the length of chromosomes allows it to contribute to mitosis in a variety of ways. The ATPase activity of ORC could have a role in chromosome cohesion or condensation. Alternatively, the role of ORC in M phase may reflect an unanticipated early step in the assembly of the pre-RC in combination with a checkpoint that monitors the execution of this early step.

  • * To whom correspondence should be addressed. E-mail: jrine{at}


View Abstract

Stay Connected to Science

Navigate This Article