Initiation of Genome Replication:
An Overview
THE EUKARYOTIC CELL CYCLE
The eukaryotic genome is organized into multiple chromosomes because of its size and complexity. To produce two daughter cells each receiving a copy of the entire genome, the DNA in each chromosome must first be duplicated accurately, before the two copies are distributed to the two daughter cells. These processes define the two major phases of the cell cycle. DNA replication occurs during S-phase (S for synthesis), which requires about half of the cell cycle time in a typical mammalian cell (~9 h). After S-phase, chromosome segregation and cell division occur in mitosis (or “M-phase”), which requires much less time (~1 h). At the same time, dividing cells must coordinate their growth (i.e. their increase in cell mass) to maintain their size. Time for cell growth is provided through gap-phases between M- and S-phase (G1) and between S-phase and mitosis (G2). Thus, the eukaryotic cell cycle is divided into four sequential phases: G1, S, G2 and mitosis. G1, S and G2 together are called interphase (see FIG 3-1). The two gap phases are not essential and can be skipped in rare cases when rapid cell division is essential, e.g. during embryogenesis. On the other hand, most cells in the body persist in G1 during most of their life time, or – once they exited the cell cycle upon differentiation, in a closely related phase called G0. The progression from one cell cycle phase to the subsequent one is tightly regulated, and cells sense the completion of critical processes (such as DNA replication, or repair after DNA damage) before the next phase is initiated. Information about the completion of cell cycle events, as well as signals from the environment, can cause a control system to arrest the cell cycle at specific “checkpoints”. The G1-phase is especially important in this respect. Its length can vary greatly depending on external conditions and extracellular signals from other cells. If signals to grow and divide are present, cells in early G1-phase progress through the commitment point near the end of G1, known as “Start” or restriction point, and enter a new cell cycle.
Eukaryotic cells have developed a complex network of regulatory proteins known as the cell cycle control system, that governs progression through the cell cycle. Key components of this system are protein kinases grouped together in a family known as cyclin-dependent kinases (CDKs). The activity of individual kinases of this family oscillates as the cell progresses through the cell cycle. These oscillations lead to a cyclical change in the phosphorylation state of target proteins that in turn trigger different events of the cycle to drive DNA replication, mitosis or cytokinesis, and links the cell cycle to gene expression (reviewed in Morgan (1995)) (see FIG 3-1).
An active CDK is composed of a catalytic kinase subunit (cdk) and the regulatory cyclin subunit that activates the kinase and confers substrate specificity. The number of genes encoding for the cdk- and cyclin- subunits varies among eukaryotes. While in yeast only a single catalytic kinase subunit seems to be involved in the regulation of the cell cycle, higher eukaryotes possess at least seven different cdks (cdk1-7). As to the regulatory subunits, a large number of different cyclins could be identified both in yeast (CLN1-3 and CLB1-6) and vertebrates (cyclin A-K). Each cdk interacts with a specific subset of cyclins and, in addition, a single mammalian cyclin can sometimes interact with multiple cdks (Sherr, 1993). This array of combinations explains the broad variety of substrates.
CDK activity is tightly controlled by several complex mechanisms. While cdk levels tend to remain constant throughout the cell cycle, cyclins are synthesized and degraded in a cell cycle-dependent manner involving the ubiquitin-dependent proteolytic machinery (Glotzer et al., 1991). In addition to cyclin binding, complete CDK activation requires phosphorylation of a conserved threonine in the catalytic subunit (Thr161) by the CDK-activating kinase (CAK), while the active cdk/cyclin complex can be inhibited by phosphorylation of a conserved threonine/tyrosine pair (Thr14 and Tyr15) by Wee1 kinase (reviewed in Morgan (1995)).
Another major mechanism for CDK regulation involves a diverse family of proteins, termed the CKIs (cyclin-dependent kinase inhibitors), which bind and inactivate cdk/cyclin complexes (Elledge and Harper, 1994). p21CIP1, the first CKI identified in mammalian cells, has a preference for cdk2- and cdk4-cyclin complexes and is regulated mainly on the level of transcription. p21CIP1 transcription is induced by p53, a transcriptional regulator that mediates cell cycle arrest following DNA damage (Dulic et al., 1994; el-Deiry et al., 1994).
FIG 3-1 Overview of the cell cycle control in mammals. The core of the cell cycle control system consists of a series of cdk/cyclin complexes. Entry into the cell cycle is induced by the expression of D-type cyclins via signal transduction induced by growth factors. Cdk4/cyclin D activity initiates phosphorylation of the tumor suppressor protein pRb (Rb: retinoblastoma), which is complexed to the transcription factor E2F. This inactivates pRb, freeing E2F to activate the transcription of S-phase genes, including the genes for a G1/S-cyclin (cyclin E) and S-cyclin (cyclin A). The resulting appearance of cdk2/cyclin E induces transition into S-phase. At the beginning of S-phase cyclin E is degraded and replaced by cyclin A, important for progression through S-phase. Entry into mitosis is regulated by the maturation promoting factor (MPF) consisting of cdk1 and cyclin B. At the end of mitosis MPF is inactivated by the degradation of cyclin B by the anaphase-promoting complex (APC), which leads to cytokinesis and transition to G1-phase. (adapted from E.-M. Ladenburger (2002))
ORIGINS OF DNA REPLICATION
Origins of replication are DNA sequences that direct the assembly of multiprotein machines that eventually form two replication forks, which then move in opposite directions until they meet the boundaries of the adjacent replication units, called replicons. Depending on the size of the genome and the developmental stage, replication start sites occur every thousands to ten thousands of base pairs. Therefore, the precise duplication of the entire genome requires the coordination of multiple initiation events.
Yeast Origins
The initiation of DNA replication in eukaryotic chromosomes is best understood in the budding yeast Saccharomyces cerevisiae. In this organism, DNA replication initiates at DNA regions of about 150 bp called “autonomously replicating sequences” (ARS) according to their ability to act as origins for replication of free episomal elements or minichromosomes in transfected yeast cells (Stinchcomb et al., 1979).
Replicator activity of these sequences depends on an essential 11 bp AT-rich consensus A-element (A/TTTTAC/TA/GTTTA/T), termed the ARS consensus sequence (ACS), and two or three auxilliary B-elements (B1-B3) (Marahrens and Stillman, 1992). Although less conserved than the A-element, the B-elements are essential for origin activity and are located in a defined distance and orientation to the A-element. Both the A-element together with the B1-element act as the recognition sequence for the “origin recognition complex” (ORC). ORC, a six-polypeptide complex, is essential for the initiation of DNA replication and appears to be the yeast replication initiation factor analog to the dnaA protein of Escherichia coli and T-antigen of Simian Virus 40 (SV40). Like these proteins, ORC recognizes and binds the DNA sequence of the replication origin, and then recruits other replication proteins to this site (Bell and Stillman, 1992). The B2-element, located between the ORC binding site and the B3-element, is part of the DNA unwinding-element (DUE), where parental DNA begins to be unwound to establish new replication forks. Some origins contain the B3-element, a binding site for the transcription factor Abf1, which has a stimulatory effect on the initiation of replication (Diffley and Stillman, 1988) (see FIG 3-2).
FIG 3-2 Origin structure in the yeast S. cerevisiae. Schematic of ARS1, a typical DNA region where DNA replication starts in S. cerevisiae, with the essential consensus A-element (ACS) and the three B-elements (OBR: origin of bidirectional replication). The A-element together with the B1-element acts as the core of the recognition sequence for ORC. The B2-element is part of the DNA unwinding-element (DUE), while the B3-element acts as a binding site for the transcription factor Abf1, which has a stimulatory effect on the initiation of replication. OBR (adapted from M. L. DePamphilis (1999) with modifications)
Many ARS sequences in the yeast genome display origin activity in the context of a plasmid, but only few of those sequences are used as origins in living cells. In addition, the frequency and time of activation differs among individual ARS (reviewed in DePamphilis (1999)). This strongly suggests that additional factors contribute to the origin usage, and it has been proposed that chromatin structure plays an important role in the selection of origins.
In other organisms, the cis-acting sequences required to direct the initiation of replication are more complex. ARS assays used in experiments with the fission yeast Schizosaccharomyces pombe revealed that sequences spread over 500-1000 bp direct initiation (Clyne and Kelly, 1995; Dubey et al., 1994). Detailed analysis of these sequences identified multiple short AT-rich elements of 20-50 bp of length that contribute to origin function, but they do not exhibit the sequence similarities observed in the ACS of S. cerevisiae.
Metazoan Origins
The identification of metazoan origins has been shown to be much more demanding because replication of extra-chromosomal DNA in cultured mammalian cells is inefficient. In turn, metazoan systems that allow efficient replication – such as Xenopus egg extract or Drosophila tissue culture – showed that almost all genomic fragments exhibit ARS activity. For a while, these results favored a sequence-nonspecific mechanism underlying replication initiation in higher eukaryotes (Coverley and Laskey, 1994; Gilbert, 1998; Smith and Calos, 1995). However, upon introduction of several methods to map origins and their sites of initiation of DNA synthesis, a different picture emerged.
The technique of two-dimensional agarose gel electrophoresis allows the distinction of structures emerging from replication initiation at an origin (replication bubbles) and elongating replication forks, and revealed that metazoan origins were much larger than in yeast (Brewer and Fangman, 1987; Brewer and Fangman, 1991). These “initiation zones” were heterogeneous in size and ranged from 2-55 kb (reviewed in DePamphilis (1999) and Hamlin and Dijkwel (1995)). Using the more sensitive approach of determining the abundance of nascent strands allowed origins to be mapped to as small a region as 0.5-2 kb indicating that in metazoan cells origin activity is also defined to certain sites comparable to the situation in yeast (Giacca et al., 1994; Kobayashi et al., 1998) (reviewed in Bielinsky and Gerbi (2001)).
Although a handful of replication start sites have been identified (reviewed in DePamphilis (1999)), metazoan origins are still less well defined than their yeast counterparts. Apparently, origins can extend over thousands of base pairs of DNA and share only few common sequence features. These include tracts of AT-base pairs, which probably serve as start sites for local unwinding of the DNA. These AT-elements are often surrounded by CpG-rich elements, which are often methylated (Araujo et al., 1998; Rein et al., 1997a; Rein et al., 1997b). Sequence analysis of cloned nascent DNA
also revealed an enrichment of CpG-islands. Interestingly, most identified origins are located in CpG-islands in upstream regions of mammalian housekeeping genes (Delgado et al., 1998). Therefore it has been frequently proposed that the local organization of chromatin, like the more open chromatin structure in the promoter regions of actively transcribed genes, determines whether given sites in the metazoan genome are used as origins of replication (DePamphilis, 1996).
Origins in early embryos of Drosophila and Xenopus appear to be located closer to each other and require little or no sequence specificity, presumably to allow an extremely rapid S-phase (reviewed in Blow (2001) and DePamphilis (1999)). Thus, it remains elusive why some sites consistently act as origins of replication, while others are used only in certain developmental stages. Many more experiments will be necessary before we understand the mechanisms that select these sites as origins, and the sequences that determine their location in the three-dimensional context of the cell nucleus.
ESTABLISHMENT OF THE PRE-REPLICATION COMPLEX
The large and complex genomes of metazoan organisms need regulatory mechanisms to ensure that each section of the genome is replicated precisely once each cell cycle.
Mechanism to prevent additional rounds of replication within one cell cycle could use either positive or negative signals on the chromatin template. This concept was originally deduced from cell fusion studies (Rao and Johnson, 1970) and was further elaborated in the “licensing factor” model (Blow and Laskey, 1988). According to this model, the breakdown of the nuclear membrane in mitosis allows transient access for a
“licensing factor” to chromatin. This licensing factor supports a single initiation event where it is bound to DNA, and is inactivated by either initiation or elongation of the replication fork. However, identification of various factors involved in this process over the past years gave a closer insight into the molecular mechanism of the licensing reaction, which appears to be even more complex.
Eukaryotic origins of replication direct the formation of a number of protein complexes leading to the assembly of two bidirectional DNA replication forks (Bell and Dutta, 2002). These events are initiated during the G1-phase, when “pre-replication complexes” (pre-RCs) form at origins of replication by the ordered assembly of a number of replication factors including ORC, Cdc6p, Cdt1p, and Mcm2-7p. These protein factors render chromatin competent for replication, a state referred to as being
“licensed”. The regulation of pre-RC formation is a key element of the mechanism coordinating DNA replication with the cell cycle. Once formed, this complex awaits activation by at least two types of kinases (CDK and DDK) that trigger the transition to S-phase. As with the formation of the pre-RC, the transition to active replication involves the ordered assembly of additional replication factors, which facilitate unwinding of the DNA at the origin and culminate in the association of the eukaryotic DNA-polymerases with the unwound DNA.
Even though the organization of the origin structure in complex eukaryotes remains unclear, the initiation factors, in particular ORC, Cdc6p and Mcm2-7p seem to be evolutionary conserved from yeast to humans. This strongly suggests that the mechanism of the initiation of replication is conserved in all eukaryotes.
FIG 3-3 A model for pre-replicative complex formation in yeast. In the yeast S. cerevisae ORC is bound to specific origin sites in the genome as a hexameric complex. In early G1-phase ORC recruits the loading factors Cdc6p and Cdt1p, which in turn are responsible for the assembly of the six subunit MCM complex to form a pre-RC rendering chromatin competent for replication, a state refered to as being
“licensed”. Upon transition to S-phase the pre-RC is activated by at least two types of kinases: cyclin-dependent kinases (CDKs) and the Dbf4-cyclin-dependent kinase cdc7 (DDK). Function of both kinases leads to the dissociation of Cdc6p. For transition to origin unwinding and replication the association of Cdc45p to the Mcm2-7 complex is required. It is, however, not clear whether, as indicated in the figure, CDK and DDK leave the initiation complex after they have accomplished the loading of Cdc45p. Cdc45p is required for the loading of the DNA-polymerase α/primase on chromatin. All these events finally lead to the establishment of replication forks. (adapted from E.-M. Ladenburger (2002) with minor modifi-cations)
The Origin Recognition Complex
The origin recognition complex (ORC) was originally identified in the yeast S. cerevisiae as a factor that binds specifically to ARS elements in an ATP-dependent manner (Bell and Stillman, 1992). ScORC consists of six subunits encoded by the essential yeast genes ORC1-ORC6. The subunits ScOrc1p, ScOrc4p and ScOrc5p contain putative nucleotide binding sites (the Walker A- and Walker B-motifs), which are often found in DNA binding proteins that possess helicase activity (Koonin, 1993).
DNA binding is mediated by the largest subunit ScOrc1p that has been shown to hydrolyze ATP in vitro (Klemm et al., 1997). However, upon binding to stranded DNA, the ATPase reaction is inhibited and ATP is stabilized. Unlike double-stranded origin DNA, single-double-stranded DNA stimulates ATP hydrolysis. In addition, electron microscopy shows that ORC alternates between two conformations: an extended conformation, when bound to double-stranded DNA, and a bent conformation, when bound to single-stranded DNA (Lee et al., 2000). These observations are consistent with a model where ORC is associated with DNA in an ATP-bound state, followed by conformational changes in a downstream event that produces single-stranded DNA (e.g. initiation) promoting hydrolysis of the bound ATP. On the basis of published results, ORC itself does not seem to be involved in the unwinding of DNA, and in this respect differs from simpler initiator protein like Escherichia coli dnaA or Simian Virus 40 (SV40) T-antigen (Kornberg and Baker, 1992). More recent results demonstrate, that ORC might modulate the local chromatin structure, for example by affecting the positioning of nucleosomes to facilitate initiation (Lipford and Bell, 2001).
ScORC requires the five largest subunits (ScOrc1p-ScOrc5p) to recognize DNA, four of which are in close contact with the origin DNA as determined by cross-linking studies (Lee and Bell, 1997). ScOrc1p together with ScOrc2p and Orc4p are in close contact with one DNA strand in the major groove of the ARS element A, while subunit ScOrc5p interacts with the B1-element. Contacts to the subunits ScOrc3p and ScOrc6p are probably mediated by protein-protein interactions.
Although first identified in S. cerevisiae, subsequent studies have found the multi-subunit ORC complex in all eukaryotes studied (reviewed in Quintana and Dutta (1999)). However, the properties of the single ORC subunits in these organisms are not as well characterized as in yeast and are still under investigation.
The Loading Factors Cdc6p and Cdt1p
The origin recognition complex (ORC) recruits Cdc6p (Cdc: cell division cycle), which in turn is required for the loading of MCMs onto chromatin (Liang et al., 1995;
Mizushima et al., 2000). In fact, Cdc6p is related to Orc1p and, to a lesser extent, to Orc4p on the amino acid level and it also contains a nucleotide binding domain composed of Walker-motifs (Bell et al., 1995; Quintana et al., 1997). Analysis of mutations in the Walker A- and Walker B-motifs revealed that the Walker A-motif is essential for an efficient interaction with ORC, while the Walker B-motif plays an
important role in the recruitment of the MCM proteins to the pre-RC (Tye, 1999).
Interestingly, Cdc6p also displays sequence similarities to a superfamily of factors that load ring-shaped “processivity factors” of DNA-polymerases onto DNA (reviewed in Tye (1999)). These loading factors include subunits of the RF-C (replication factor C), which load the eukaryotic sliding-clamp protein PCNA (proliferating-cell nuclear antigen) onto DNA (Jonsson and Hubscher, 1997). PCNA is a trimeric ring-shaped complex with pseudo-sixfold symmetry and acts as a processivity factor of the DNA-polymerase δ holoenzyme. Due to the similar structure of the hexameric MCM complex, which has been shown to resemble a ring with a central opening by electron microscopy studies (Adachi et al., 1997), it is tempting to speculate whether the loading of the MCM-complex by Cdc6p is analogous to the loading of PCNA by RF-C.
Apart from Cdc6p, a second factor called Cdt1p (Cdt: cdc10-dependent transcript) is important for the recruitment of the MCM proteins to chromatin. Chromatin binding of both Cdc6p and Cdt1p depends on the presence of ORC on origin DNA, but occurs independently of each other (Maiorano et al., 2000; Nishitani et al., 2000), although they appear to function synergistically to load the MCM complex. Interestingly, once the MCM proteins have been loaded on chromatin, ORC and Cdc6p (and probably Cdt1p) can be removed from chromatin without affecting subsequent DNA replication, suggesting that the primary role of the pre-RC is the loading of MCMs (Hua and Newport, 1998; Rowles et al., 1999).
The MCM Proteins
All eukaryotes express six MCM proteins (MCM: minichromosome maintenance), Mcm2p-Mcm7p, a protein family with considerable sequence conservation between different species. Homology between the MCM subunits is concentrated in a central region of about 200 amino acids, which contains an element similar to the Walker-A-motif described above (Koonin, 1993).
Although immunoprecipitation of one of the MCM proteins often leads to co-precipitation of all six members of the MCMs (Adachi et al., 1997; Brown and Kelly, 1998; Kubota et al., 1997; Richter and Knippers, 1997), a variety of subcomplexes have also been purified. In particular, complexes containing Mcm2/4/6/7p, Mcm4/6/7p and Mcm3/5p are commonly detected (Burkhart et al., 1995; Fujita et al., 1997; Kimura et
Although immunoprecipitation of one of the MCM proteins often leads to co-precipitation of all six members of the MCMs (Adachi et al., 1997; Brown and Kelly, 1998; Kubota et al., 1997; Richter and Knippers, 1997), a variety of subcomplexes have also been purified. In particular, complexes containing Mcm2/4/6/7p, Mcm4/6/7p and Mcm3/5p are commonly detected (Burkhart et al., 1995; Fujita et al., 1997; Kimura et