Stepwise assembly of initiation complexes at budding yeast replication ...

67

Journal of Cell Science, Supplement 19, 67-72 (1995) Printed in Great Britain © The Company of Biologists Limited 1995

Stepwise assembly of initiation complexes at budding yeast replication origins during the cell cycle John F. X. Diffley, Julie H. Cocker, Simon J. Dowell*, Janet Harwood and Adele Rowley* Imperial Cancer Research Fund, Clare Hall Laboratories, South Mimms, Hertfordshire, EN6 3LD, UK 'Present address: Glaxo-Wellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK

SUMMARY DNA replication is a pivotal event in the cell cycle and, as a consequence, is tightly controlled in eukaryotic cells. The initiation of DNA replication is dependent upon the comple­ tion of mitosis and upon the commitment to complete the cell cycle made during Gi. Characterisation of the protein factors required for initiating DNA replication is essential to understand how the cell cycle is regulated. Recent results indicate that initiation complexes assemble in multiple stages during the cell cycle. First, origins are bound by the multi­ subunit origin recognition complex (ORC) which is essential for DNA replication in vivo. ORC, present at little more than one complete complex per replication origin, binds to origins immediately after initiation in the previous cell cycle. ORC binding occurs by the recognition of a bipartite sequence that includes the essential ARS consensus sequence (ACS) and the functionally important B1 element adjacent to the

ACS. A novel pre-replicative complex (pre-RC) assembles at origins at the end of mitosis in actively cycling cells and remains at origins until DNA replication initiates. Finally, Dbf4, which is periodically synthesised at the end of Gi, interacts with replication origins. Dbf4-origin interaction requires an intact ACS strongly suggesting that interaction occurs through ORC. Dbf4 interacts with and is required for the activation of the Cdc7 protein kinase and together, Dbf4 and Cdc7 are required for the Gi-S transition. Separate regions of Dbf4 are required for Cdc7- and origin-interaction suggesting that Dbf4 may act to recruit Cdc7 to repli­ cation origins where phosphorylation of some key component may cause origin firing.

INTRODUCTION

yeast origins. First, all ARSs contain an 11 bp A+T rich sequence, the ARS consensus sequence (ACS), that is essential for origin function. Second, this ACS is not sufficient for origin function; in all cases tested a flanking sequence 3' to the T-rich strand of the ACS is essential for origin function. Analysis of several origins has provided strong evidence that these flanking sequences are composed of multiple subdomains which are each important, but not essential, for origin function. The arrangement of sequence elements within one of the best studied yeast replication origins, ARS1, is shown in Fig. 1 (Marahrens and Stillman, 1992).

Eukaryotic genomes are often much larger than their prokary­ otic counterparts. Moreover, genomic DNA is usually divided among multiple chromosomes. To deal with this, DNA repli­ cation in eukaryotic cells initiates from large numbers of repli­ cation origins during each S phase. Replication from these origins does not initiate synchronously at the beginning of S phase, but, instead, can initiate throughout S phase. Conse­ quently, the co-ordination of replication origin function in eukaryotes poses a considerable problem; replication origins must be used efficiently in each S phase to ensure rapid and complete DNA replication, yet origins m ust never be reused in any S phase since this would lead to unbalanced DNA synthesis, abnormal DNA structures and alterations in gene dosage. Understanding this ‘once and only once’ mechanism for initiation on a molecular level represents an important goal in cell cycle study (reviewed by Rowley et al., 1994). In Saccharomyces cerevisiae, specific, short DNA sequences known as autonomously replicating sequences (ARSs) serve as replication origins both genetically and bio­ chemically. That is, ARSs are required for the initiation of DNA synthesis in vivo and DNA synthesis begins within these sequences. Analysis of the sequences required for origin function has revealed a common architecture to all budding

Key words: initiation complex, assembly, Saccharomyces cerevisiae, replication origin, cell cycle

RESULTS We have taken several approaches to characterise protein complexes at budding yeast replication origins in vivo.

Genomic footprinting: the origin recognition complex binds replication origins in vivo To begin to understand DNA-protein interactions at replication origins, we have used the technique of genomic footprinting to achieve a nucleotide level resolution map of the regions of replication origins bound by proteins in chromatin. In this technique, cells are permeabilised and immediately treated

68

J. F. X. Diffley and others

TRP1

GAL3 ACS

B3 — B2

Bl

Chromosome IV

A

A RSI

Fig. 1. Schematic view o f ARS1. ARS 1 is found on chromosome IV between the TRP1 and GAL3 genes. Celniker et al. showed it to be composed of three domains designated A, B and C. Analysis from Marahrens and Stillman (1992) showed domain B to be composed of three subdomains designated B l, B2 and B3. The essential ARS consensus sequence lies within domain A. The other elements are non-essential but make significant contribution to origin function both on plasmids and in their normal chromosomal location (Marahrens and Stillman, 1994).

w ith nuclease. D N A is purified and the positions o f cleavage sites is determ ined by prim er extension into the region of interest with an end-labelled oligonucleotide. The pattern of cleavage sites in chrom atin can then be com pared to the pattern o f nuclease cleavage sites in purified DNA. Initial experim ents on unsynchronised cells revealed a num ber o f interesting features (D iffley and Cocker, 1992). M ost im portantly, it show ed that the essential A CS was protected from nuclease cleavage in chrom atin and that sequences adjacent to the ACS, especially w ithin the B l elem ent, w ere hypersensitive to cleavage by nuclease, indicating that it w as bound by a novel cellular factor. C oncom itantly, Bell and Stillm an (1992) iden­ tified a m ulti-subunit protein factor that interacted specifically with the ACS in vitro. A cross the ACS (in dom ain A) and the B 1 elem ent, our in vivo cleavage patterns were rem arkably

B3

sim ilar to their in vitro patterns suggesting that this origin recognition com plex (ORC) binds to replication origins in vivo (Fig. 2A). T ogether, our data argued that a factor critical for initiation o f replication had been identified. Since our genom ic footprints were perform ed on unsyn­ chronised cells, the saturated footprint at the ACS suggested that ORC binding was not transient during the cell cycle. This is discussed in further detail below. O ur experim ents also provided evidence that the above-m entioned ABF1 protein was bound at the B3 elem ent o f ARS1 in vivo. Furtherm ore, ARS1 is bounded by a tightly positioned nucleosom e on the dom ain B-distal side o f dom ain A. T hese experim ents gave a view o f initiation com plexes at a eukaryotic replication origin at nucleotide resolution (sum m arised in Fig. 2B).

Identification and characterisation of the RRR1/ORC2 gene: genetic evidence that ORC plays a role in DNA replication and transcriptional silencing in vivo A lthough the dem onstration that ORC required the essential ACS for ARS binding in vitro (Bell and Stillm an, 1992) and appeared to be bound at replication origins in vivo (Diffley and Cocker, 1992) strongly suggested a role for ORC in replica­ tion, genetic evidence w as lacking. O RC is com posed o f subunits o f 120, 72, 62, 56, 53 and 50 kDa. In collaboration with Gos M icklem and K im N asm yth, w e isolated the gene, designated RRR1, w hich encodes the 72 kD a subunit o f O RC (O R C2) (M icklem et al., 1993). RRR1 is essential for prolifer­ ation and r r r l m utants are unable to m aintain an endogenous plasm id, the 2 m icron circle and have aberrant cell cycle kinetics (M icklem et al., 1993). The sam e gene, designated

B2

ill i4 i: i; n W W AATTTCGTCAAAAATGCTAAGAAATAGGTTATTACTGAGTAGTATTTATTTAAGTATTGTTTGTGCACTTGCCTGCAGGC ABF1

ORC Genomic

TTAAAGCAGTTTTTACGATTCTTTATCCAATAATGACTCATCATAAATAAATTCATAACAAACACGTGAACGGACGTCCG

.. ABF1

ft

t

t

USD

t

t tt t

Genomic ORC

Fig. 2. Summary of in vitro and

in vivo footprinting results. (A) Regions of protection from DNase 1 at ARS 1 are shown as B1 boxes while hypersensitive sites are shown as arrows. Data from genomic footprints is ORC ■U' taken from Diffley and Cocker Genomic eaj • ____________________________ ________ CTTTTGAAAAGCAAGCATAAAAGATCTAAACATAAAATCTGTAAAATAACAAGATGTAAAGATAATGCTAAATCATTTGG (1992) (grey rectangles), data 910 GAAAACTTTTCGTTCGTATTTTCTAGATTTGTATTTTAGACATTTTATTGTTCTACATTTCTATTACGATTTAGTAAACC for in vitro ORC footprints is Genomic 0RC taken from Bell and Stillman (1992) (black rectangles) and data for in vitro ABF1 footprints is taken from Diffley and Stillm an (1988) (white rectangles). D N asel cleavage sites within the positioned nucleosome adjacent to domain A are indicated as black filled circles. The positions o f the individual ARS 1 sequence elem ents are indicated above the sequence. (B) Results o f footprinting analysis suggest that ORC binds across the A and B 1 elem ents with the DNA in the B 1 elem ent lying on the ORC protein surface while ABF1 binds to the B3 element. Weak protection at B2 suggests that it may also be bound by a protein in vivo. This figure is derived from Fig. 4 (Diffley and Cocker, 1992).

t^E

B

™ tt tt

Stepwise assembly of initiation complexes ORC2, was isolated in Jasper Rine’s laboratory (Bell et al., 1993; Foss et al., 1993). Together with Bruce Stillman and col­ leagues, they showed that orc2 mutants were defective in maintaining plasmids and were unable to enter and complete S phase at the non-permissive temperature (Bell et al., 1993; Foss et al., 1993). Together our results indicated a crucial role for ORC2 in initiating DNA synthesis. In addition to being defective in DNA replication, rrrl/orc2 mutants are also defective in transcriptional silencing of the HMR mating type locus (Foss et al., 1993; Micklem et al., 1993). In budding yeast two mating type loci, HML and HMR, are main­ tained in a transcriptionally silent state by flanking sequences known as ‘silencers’. These silencers each contain an ACS and can act as replication origins on plasmids. Thus, the finding that rrrl/orc2 mutants are defective in silencing further supports an interesting link between origin function and silencing.

Biochemical characterisation of ORC: a protein present at low levels in vivo with complex sequence requirements for origin binding both in vitro and in vivo The inability to detect ORC DNA binding activity in crude yeast extracts (Bell and Stillman, 1992, and data not shown) together with the generally low yields from ORC protein purifi­ cations led us to examine the amount of ORC per cell. The availability of the cloned RRR1/ORC2 gene allowed us to express the ORC2 subunit in Escherichia coli and generate polyclonal antisera to the purified subunit. From quantitative immunoblots on whole cell extracts we estimate that logarith­ mically growing haploid yeast cells contain only approxi­ mately 600 molecules of the ORC2 subunit (Rowley et al., 1995), which appears to be directly involved in origin recog­ nition (Bell and Stillman, 1992) and ORC integrity (Bell et al., 1993). Previously it has been estimated that there are approx­ imately 350-470 origins per haploid genome. Assuming haploid cells spend approximately half their time with a 2C DNA content (i.e. G 2 and M phases), the average number of origins per cell in an asynchronous population is approxi­ mately 525-705. Thus, ORC is present at levels of approxi­ mately one complex per replication origin. The low abundance of ORC in vivo together with the fact that ORC binds origins almost immediately after initiation and remains bound throughout the cell cycle (see below) suggested that the mechanism by which ORC locates and binds to repli­ cation origins would be interesting. We have purified ORC to apparent homogeneity and begun to characterise its require­ ments for DNA binding. It had previously been shown that ORC requires ATP and an intact ACS for efficient ORC binding (Bell and Stillman, 1992) and our results were consis­ tent with this. Importantly, we also found that mutations in the B l element o f ARS1 reduced ORC binding efficiency 5- to 10fold in vitro. This B 1 effect was dependent upon the presence of a non-specific competitor DNA in the binding reaction. Fur­ thermore, genomic footprinting experiments demonstrated that ORC requires both the ACS and the B 1 element for efficient binding in vivo indicating that the B l element appears to act in vivo to increase the affinity of ORC for replication origins.

Cell cycle regulation of protein-origin complexes: identification of a novel G 1 origin binding factor As described above, the fact that the ORC footprint was

69

detectable at replication origins from asynchronous popula­ tions of cells suggested that its binding was not transient during the cell cycle. Therefore, interactions of other proteins with origins might play important roles in the cell cycle regulation of replication. To address this, we undertook an extensive analysis of complexes at replication origins through the cell cycle by genomic footprinting. Consistent with our experiments on unsynchronised cells, we found that the ACS remains protected from nuclease digestion throughout the cell cycle, suggesting that ORC is bound during the entire cell cycle. Post-replicative origin complexes, which appear shortly after DNA replication initiates in S phase and are present through S, G 2 and M phases, are nearly indistinguishable from complexes generated in vitro with purified ORC and ABF1. These results indicate that ORC and ABF1 rebind to replication origins very quickly after repli­ cation initiates. Furthermore, these results argue strongly that the binding of ORC and ABF1 to origins is not sufficient to drive initiation. In rapidly cycling cells, origin chromatin undergoes a change at the end of mitosis consistent with the binding of an additional factor. This ‘pre-replicative complex’ (pre-RC) remains at origins until DNA replication initiates, after which origins return to the post-replicative state described above. When cells enter stationary phase from Gi, they lose the preRC. Stationary phase origin complexes are indistinguishable from post-replicative origins at the level of genomic footprint­ ing. Thus, ORC and ABF1 remain stably bound to origins during long periods of quiescence while the pre-RC is quickly lost.

The G 1 -S transition: Dbf4, the regulatory subunit of the Cdc7 protein kinase, interacts with replication origins through ORC In an attempt to identify other gene products that interact with yeast replication origins, we developed a ‘one hybrid’ genetic screen (Fig. 3). In this screen, an intact and functional replica­ tion origin was placed upstream of the lacZ reporter gene. By itself, this sequence does not activate transcription in budding yeast (Buchman and Kornberg, 1990). W e reasoned that proteins interacting with replication origins either directly, such as ORC and ABF1, or indirectly might activate tran­ scription from this reporter gene when fused to the strong tran­ scriptional activation domain from the GAL4 protein (GAD). Using this approach to screen libraries of budding yeast cDNA-GAD fusions, we isolated five clones that induced sig­ nificant levels of (3-galactosidase activity from reporters con­ taining ARS1 upstream of the lacZ gene but little or no (3galactosidase activity from reporters without a replication origin upstream of lacZ. All of the positive clones contained overlapping regions of the previously identified DBF4 gene (Dowell et al., 1994). DBF4, originally isolated in Lee Johnston’s lab, is specifically required at the end of Gi for the initiation of DNA synthesis. The product of another gene, CDC7, which encodes a protein kinase, is also required at this point. The Cdc7 protein is present throughout the cell cycle, but the Cdc7-associated kinase activity peaks at the Gi-S tran­ sition. Dbf4 appears to play a critical role in the activation of the Cdc7 kinase: Cdc7 kinase activity is reduced in dbf4 mutants (Jackson et al., 1993), Cdc7 and Dbf4 physically interact in vivo (Dowell et al., 1994; Jackson et al., 1993;

70

J. F. X. D iffley and others

lacZ ~¥i— 1

CSt7C''

r*

-W M --------------

Fig. 4. Effect of Dbf4 deletions on ARS1 and Cdc7 interactions. Full length (FL) Dbf4 and a series o f D bf4 deletions were fused to the GAL4 transcriptional activation domain and tested for ARS 1 interaction in the one-hybrid assay described above. Black bars indicate the region of Dbf4 present in the fusions. The same deletions were fused to the Lex A D NA binding domain and tested for their interaction with full length Cdc7 fused to the GAL4 transcriptional activation domain in a two-hybrid assay. All of the Dbf4 deletion constructs were also tested for their ability to complement the dbf4-2 mutation. Adapted from Fig. 4 of Dowell et al. (1994).

Fig. 3. A one-hybrid approach to identify origin-interacting gene products. Gene products that interact either directly (shown in white) or indirectly (shown in grey) with either the A or B elements of ARS1 are expected to activate transcription from the adjacent lacZ gene when they are fused to the transcriptional activation domain o f GAL4 (GAD). Colonies o f cells harbouring such fusions are expected to turn blue in qualitative (3-galactosidase assays using Xgal. Adapted from Fig. 1 of Dowell et al. (1994).

K itada et al., 1992, and S. D ow ell and J. D iffley, unpublished data); and D B F4 is specifically transcribed in late Gi at approx­ im ately the tim e that Cdc7 kinase becom es activated (Chapm an and Johnston, 1989). The finding that D bf4 interacts w ith D N A replication initiation com plexes suggested that D bf4 m ight act to recruit the Cdc7 protein kinase to initiation com plexes. A lternatively, D bf4 m ight interact w ith initiation com plexes through Cdc7. T o distinguish betw een these possibilities, we assayed a series o f D bf4 deletions for their ability to interact w ith ARS1 in a one-hybrid assay and with Cdc7 in a tw ohybrid assay (Fig. 4). This experim ent indicated that the origininteraction dom ain o f D bf4 m aps to the am ino term inus w hile the C dc7-interaction dom ain m apped to the m iddle o f the protein. T hat is, the origin- and C dc7-interaction dom ains are separable and, therefore, D bf4 does not interact w ith initiation com plexes through Cdc7.

DISCUSSION T he results sum m arised in this paper have led to the m odel show n in Fig. 5 in w hich initiation com plexes assem ble at yeast replication origins in several discrete stages. O RC rebinds replication origins shortly after initiation and rem ains bound at replication origins throughout the cell cycle and during long periods o f quiescence. T hat O RC is bound post-replicatively indicates that O RC binding is not sufficient to drive the

initiation o f replication. H ow ever, the fact that the ACS is essential for D N A replication together w ith genetic results w ith the RRR 1/O RC 2 gene described above indicate that O RC is necessary for initiation. It is upon a ‘scaffold’ o f O R C that additional com ponents o f initiation com plexes assem ble during the cell cycle. A t the end o f m itosis, origin chrom atin is converted to the pre-replicative state or com plex (pre-RC). R esults discussed here and data not show n suggest that, in the pre-R C , an additional protein com plex binds across the ACS and the A CS-flanking sequence w hich includes dom ains B l and B 2 at ARS1 consistent w ith the hypothesis that the pre-RC contains O RC and a novel G ispecific protein. C andidates for com ponents o f this G i-specific protein include the M em proteins and the C dc6 protein. The M em proteins have been genetically im plicated in the initiation o f D N A synthesis (Tye, 1994). Furtherm ore, three fam ily m em bers (Cdc46/M cm 5, M cm 2 and M cm 3) enter the nucleus at the end o f m itosis - approxim ately the sam e tim e that the pre-R C form s - and disappear from the nucleus upon entry into S phase- the tim e o f disappearence o f the pre-RC. The Cdc6 protein has also been im plicated in origin function in budding yeast (review ed by Bell, 1995) and Cdc6 m utants blocked at the non-perm issive tem perature do not contain pre-RCs (D iffley et al., 1994). The role o f these proteins in the pre-R C is currently under investigation. T he pre-R C appears at the end o f m itosis at the tim e that the m itotic form o f the Cdc28 protein kinase is destroyed (Surana et al., 1993). Several lines o f evidence from the fission yeast Schizosaccharom yces po m b e indicate that it is inactivation of the m itotic form o f the p34cdc2 kinase and not m itosis itself that is critical for resetting nuclei for another round o f replication. First, transient inactivation o f cdc2+ kinase prom otes diploidisation (B roek et al., 1991). Second, overexpression o f a cdk inhibitor, r u m l+, drives m ultiple rounds o f re-replication (M oreno and N urse, 1994). A nd finally, deletion o f c d c l3 +, the single B-type cyclin required for m itosis (H ayles et al., 1994),

• C L N S ynthesis —^

STA R T • CDC28 K inase A ctivation

DBF4 S ynthesis

CLB5, 6 S ynthesis

Fig. 5. Summary of interactions at yeast replication origins in vivo. Details of the model are described in the text.

also drives m ultiple rounds o f re-replication. It is tem pting to speculate that inactivation o f the m itotic cdc2+/C dc28 kinase directly prom otes pre-R C form ation. In this regard, it is inter­ esting that a B -type cyclin is also required for entry into S phase (Schw ob et al., 1994) suggesting that the very conditions required to activate D N A replication m ay also block the form ation o f new pre-R C s and thus prevent re-replication in a single cell cycle. T he ability to directly exam ine the pre-RC at origins together w ith the ability to m anipulate levels o f B cyclin-C dc28 kinase activity in vivo should allow exam ination o f this hypothesis. T he product o f the CD C 14 gene may also play a key role in assem bly o f the pre-R C . c d c l4 m utants block at the end o f m itosis before the pre-RC form s (Diffley et al., 1994). Both c d c l4 and cdc6 m utants exhibit an elevated rate o f plasm id loss in vivo. This loss is prim arily 1:0 loss suggesting a defect in D NA replication rather than chrom osom e segregation. F ur­ therm ore, this high plasm id loss rate can be suppressed by inclusion o f m ultiple replication origins on the plasm id sug­ gesting that these m utants may have a defect specifically in the initiation o f replication. W hy c d c l4 m utants block in m itosis and not S phase w hile cdc6 m utants arrest prior to entering S phase (Bueno and Russell, 1992) is unclear and may suggest a second essential function for C d c l4 in m itosis. T he pre-R C is lost rapidly w hen cells enter quiescence (D iffley et al., 1994). W hether this loss o f the pre-RC is sim ply a consequence o f degradation o f inherently unstable protein com ponents, decay o f an unstable post-translational m odifica­

tion state or is due to a m ore active m echanism is unknow n. C ells re-entering the cell cycle from quiescence form the preRC before D NA synthesis (data not show n) dem onstrating that the resetting of replication origins can occur at tim es other than the end o f mitosis. T he products o f the C D C 7 and D B F 4 genes are req u ired at the en d o f Gi fo r entry into S phase. C ells blocked at the end o f G i by raising a cd c7 ts m utant to the restrictiv e tem perature can com plete S phase in the p resence o f cyclo h ex im id e indi­ cating that new protein synthesis is not required after this p o in t for D N A replication. Thus, the finding that D bf4, w hich interacts w ith and positively regulates Cdc7 protein kinase activity, also interacts directly w ith initiation com plexes suggests strongly that Cdc7 also acts directly at initiation com plexes to trigger initiation. T he identification o f the direct targ et o f D bf4 interaction as w ell as the target o f C dc7 ph o s­ p h o rylation rep resen t im portant challenges for the near future. H ow ever, the activation o f D NA synthesis is not likely to be controlled entirely by Dbf4 synthesis and activation o f the C dc7 kinase for two reasons. First, Sclafani and co-w orkers have isolated a m utant ib o b l) that bypasses the requirem ent for both Dbf4 and Cdc7 (Jackson et al., 1993). W hile the nature o f the b o b l m utation is still unclear, its existence suggests that there m ight be a second pathw ay also required for initiating D N A synthesis. And second, as described above, at least one B cyclin is required for S phase entry from Gi (Schw ob et al., 1994). T hus there are at least tw o S T A R T-dependent kinases

72

J. F. X. Diffley and others

required to initiate DNA synthesis. Whether these kinases operate in the same or parallel pathways is currently unknown. Together, the results described in this paper and the model presented suggest that DNA replication is regulated at least in part by co-ordinating the assembly of initiation complexes with major cell cycle events such as the completion of mitosis and START.

REFERENCES Bell, S. P. and Stillm an, B. (1992). Nucleotide dependent recognition of chrom osom al origins of DNA replication by a m ulti-protein complex. Nature 357, 128-134. Bell, S. P., Kobayashi, R. and Stillm an, B . (1993). Yeast origin recognition com plex functions in transcription silencing and DNA replication. Science 262, 1844-1849. Bell, S. P. (1995). Eukaryotic replicators and associated protein complexes. Curr. Opin. Genet. Dev. 5, 162-167. Broek, D., Bartlett, R., Crawford, K. and Nurse, P. (1991). Involvem ent of p34cdc2 in establishing the dependency of S phase on mitosis. N ature 349, 388-393. Buchm an, A. R. and K ornberg, R. D. (1990). A yeast ARS-binding protein activates transcription synergistically in combination with other weak activating factors. Mol. Cell. Biol. 10, 887-897. Bueno, A. and Russell, P . (1992). Dual functions of Cdc6 - a yeast protein required for DNA replication also inhibits nuclear division. EM BO J. 11, 2167-2176. Celniker, S. E., Sweder, K., Srienc, F., Bailey, J. E. and Cam pbell, J. L. (1984). Deletion mutations affecting autonom ously replicating sequence AES1 o f Saccharomyces cerevisiae. Mol. Cell. Biol. 4 , 2455-2466. Chapman, J. W . and Johnston, L. H. (1989). The yeast gene, DBF4, essential for entry into S phase is cell cycle regulated. Exp. Cell. Res. 1 8 0 ,419-428. Diffley, J. F. X. and Stillm an, B. (1988). Purification o f a yeast protein that binds to origins of DNA replication and a transcriptional silencer. Proc. Nat. Acad. Sci. USA 85, 2120-2124. D iffley, J. F. X. and Cocker, J. H. (1992). Protein-DNA interactions at a yeast replication origin. Nature 357, 169-172. Diffley, J. F. X., Cocker, J. H ., Dowell, S. J. and Rowley, A. (1994). Two

steps in the assem bly o f com plexes at yeast replication origins in vivo. Cell 78, 303-316. Dowell, S. J., Rom anowski, P. and D iffley, J. F. X . (1994). Interaction of Dbf4, the Cdc7 protein kinase regulatory subunit, w ith yeast replication origins in vivo. Science 265, 1243-1246. Foss, M ., M cN ally, F. J., Laurenson, P. and Rine, J. (1993). Origin recognition com plex (ORC) in transcriptional silencing and DNA replication in S. cerevisiae. Science 262, 1838-1844. H ayles, J., Fisher, D., W oollard, A. and Nurse, P. (1994). Tem poral order of S phase and m itosis in fission yeast is determ ined by the state o f the p34cdc2m itotic B cyclin complex. Cell 78, 813-22. Jackson, A. L., Pahl, P. M ., H arrison, K ., Rosam ond, J. and Sclafani, R. A. (1993). Cell cycle regulation of the yeast CDC7 protein kinase by association w ith the D B F4 protein. Mol. Cell. Biol. 1 3 ,2899-2908. K itada, K., Johnston, L. H., Sugino, T. and Sugino, A. (1992). Tem peraturesensitive cdc7 m utations o f Saccharomyces cerevisiae are suppressed by the DBF4 gene, which is required for the G l/S cell cycle transition. Genetics 131,21-9. M arahrens, Y. and Stillm an, B. (1992). A yeast chrom osom al origin o f DNA replication defined by m ultiple functional elements. Science 255, 817-823. M arahrens, Y . and Stillm an, B. (1994). Replicator dom inance in a eukaryotic chrom osom e. E M B O J. 13, 3395-3400. M icklem , G ., Rowley, A ., H arwood, J., Nasmyth, K. and D iffley, J. F. X. (1993). Y east origin recognition com plex is involved in DNA replication and transcriptional silencing. Nature 366, 87-89. M oreno, S. and Nurse, P. (1994). Regulation o f progression through the G I phase of the cell-cycle by the ru m l(+ ) gene. Nature 367, 236-242. Rowley, A., D ow ell, S. J. and Diffley, J. F. X. (1994). R ecent developm ents in the initiation of chrom osom al DNA replication: a com plex picture emerges. Biochim. Biophys. A cta 1217, 239-256. Rowley, A., Cocker, J. H., Harwood, J. and Diffley, J. F. X. (1995). Initiation com plex assem bly at budding yeast replication origins begins with the recognition o f a bipartite sequence by limiting am ounts o f the initiator, ORC. E M B O J. 14,2631-2641. Schwob, E., Bohm , T., M endenhall, M . D. and Nasm yth, K. (1994). The Btype cyclin kinase inhibitor p40S IC l controls the G i to S transition in S. cerevisiae. Cell 79, 233-44. Surana, U., A m on, A., Dowzer, C., M cG rew, J., Byers, B. and Nasm yth, K. (1993). Destruction of the CDC28/CLB mitotic kinase is not required for the m etaphase to anaphase transition in budding yeast. E M B O J. 12, 1969-1978. T ye, B.-K. (1994). The MCM 2-3-5 proteins: are they replication licensing factors? Trends Cell Biol. 4, 160-166.