(Click on figures to enlarge)

Eukaryotic cells regulate their proliferation by regulating the onset of DNA replication (“S-phase”).  Once cells are committed to DNA replication, they cannot stop to rest for long periods of time (“G0-phase”) until after they have completed mitosis (“M-phase”).  Moreover, checkpoint controls ensure that M-phase does not occur until S-phase is completed and any damaged DNA is repaired.  In addition, eukaryotic cells limit initiation of DNA replication to once per replication origin per cell cycle.  This ensures that each progeny cell receives one and only one complete copy of its genome.  When cell proliferation escapes these controls, the result is cancers in adult animals and developmental failures in embryonic animals.

We and others have shown that DNA replication begins at specific sites in mammalian chromosomes, comparable in size to those found in single cell organisms.  Nevertheless, despite the fact that DNA replication is a highly conserved process throughout the eukaryotic kingdom [reviewed in Bogan et al., 2000; Bell and Dutta, 2002], multicellular animals such as frogs, flies and mammals exhibit two unique features that are absent in unicellular animals such as yeast.  In multicellular animals, the number and distribution of initiation sites during development can change dramatically, and the stability of the six subunit protein complex ("origin recognition complex") responsible for determining where replication begins is cell cycle dependent.  Therefore, we have focused our attention on how mammalian cells determine where they will initiate DNA replication, and how this process is regulated through changes in  the ORC/chromatin complex.

First, the six origin recognition proteins (Orc1 to Orc6) that comprise the origin recognition complex (ORC) bind to specific sites distributed throughout the genome.  Next, Cdc6 (Cdc18 in fission yeast) and Cdt1(also called RLF-B) proteins load mini-chromosome maintenance (Mcm) proteins 2 to 7 onto the ORC/chromatin sites to form pre-replication complexes.  Mcm(2-7) hexamers are the helicases that unwind DNA to create replication forks.  Activation of this complex begins with binding of Mcm10.  Cdc6 is released by the cyclin dependent protein kinase Cdk2/Cyclin A and replaced by Cdc45 with the help of Cdk2/cyclin E and the protein kinase Cdc7 and its cofactor Dbf4.  Cdc45 allows DNA polymerase-a:DNA primase to bind to this complex and initiate RNA-primed DNA synthesis, the first step in de novo DNA synthesis (S-phase).

In multicellular eukaryotes, replication origins vary in size and complexity [discussed in DePamphilis, 1999; Gilbert, 2001] initiation sites can change from "random" to site-specific as development progresses from rapidly cleaving embryos to a normal cell division cycle and zygotic gene expression begins. Developmental acquisition of specific initiation sites is only possible, because they are determined by epigenetic as well as genetic parameters (DePamphilis, 2003; Rein et al., 1999).  Epigenetic parameters appear to include transcription factors, nucleotide pool levels, ratio of initiation proteins to DNA, chromatin structure, nuclear organization and DNA methylation.

Each replication origin is activated once and only once per S-phase.  In this way, eukaryotic cells produce one and only one copy of their genome each time they divide.  Re-replication is prevented by multiple coherent pathways.  First, Cdc6 is phosphorylated and excluded from the nucleus.  Second, Cdt1 is inhibited by the naturally occuring protein geminin.  Third, Mcm proteins are phosphorylated and released from chromatin.  Fourth, Cdk2 is inhibited by p21, a protein whose synthesis is regulated by p53.  These events all contribute towards preventing the assembly and activation of new pre-replication complexes before the cell has undergone mitosis.  In addition, the activity of ORC is cell-cycle dependent.

The "ORC Cycle"

Studies in yeast, frogs and mammals reveal the presence of an "ORC cycle" in which one or more ORC subunits is post-translationally modified during cell division, with the result that ORC activity (i.e. the ability to initiate pre-RC assembly) is cell cycle dependent (DePamphilis, 2003).  In mammalian cells, the affinity of Orc1 for chromatin is cell cycle dependent.  In hamster cells, Orc1 is selectively released from chromatin as cells enter S-­phase, converted into a mono-ubiquitinated form, and then deubiquitinated and rebound to chromatin during the M to G1 transition, concomitant with the appearance of functional pre-replication complexes (pre-RCs) at specific genomic sites (Natale et al., 2000, Li et al., 2000, Li and DePamphilis, 2002).  Orc1 is degraded by the 26S proteasome only when released into the cytosol.  In contrast, Orc2 remains tightly bound to chromatin throughout the cell cycle and is not a substrate for ubiquitination.  Since both Orc1 and Orc2 have the same half-life in vivo, ubiquitination of non-chromatin bound Orc1 presumably facilitates inactivation of ORC by sequestering Orc1 during S-phase and thereby preventing reassembly of functional ORC/chromatin sites.  The mechanism that releases Orc1 and prevents it from reassociating with chromatin during the S to M transition is not clear, but recent evidence from our lab reveals that Cdk1/Cyclin A hyperphosphorylates Orc1 during the G2/M phase, and that inhibition of this phosphorylation results in rapid binding of Orc1 to chromatin (Li et al., 2003).  In human cells, Orc1 is polyubiquitinated and degraded during S-phase, and then resynthesized during the M to G1 transition (Méndez et al., 2002; Kreitz et al. , 2001; Tatsumi et al., 2003). These results reveal that, in mammalian cells, regulation of initiation of DNA replication begins at the very first step: selective association and dissociation of Orc1 from chromatin bound ORC.

To test this hypothesis, chromatin lacking functional ORCs was isolated from metaphase hamster cells and incubated in Xenopus egg extracts in order to initiate DNA replication de novo with Xenopus proteins (Sun et al. , 2002).  Xenopus ORC rapidly binds to hamster somatic cell chromatin in a Cdc6-dependent manner, and is then released, concomitant with initiation of DNA replication.  Once pre-RCs are assembled either in vitro or in vivo, further binding of ORC is inhibited. Release of chromatin bound ORC is prevented by inhibition of pre-RC assembly, but not by inhibitors of either cyclin-dependent protein kinase activity or DNA synthesis.  These results demonstrate a programmed release of Xenopus ORC from somatic cell chromatin as it enters S-phase, consistent with the proposed role for ORC in preventing reinitiation of DNA replication during S-phase.

In previous studies, we developed techniques for mapping the genomic location of replication origins, and used them to definitively identify the first origin of bi-directional replication in mammalian cells [ori-b 17 kb downstream of the DHFR gene (Burhans et al., 1990) and demonstrated that zones of apparently 'random' initiation in mammalian chromosomes actually consist of multiple replication origins (Kobayashi et al., 1998) that are determined by specific DNA sequences (Altman & Fanning, 2001; Kobayashi & DePamphilis, unpublished data).

To further understand the mechanism by which ORC selects specific DNA sites to initiate replication, we have turned our attention to the fission yeast, Schizosaccharomyces pombe .  S. pombe replication origins are 5 to 10 times larger than those in the budding yeast, S. cerevisiae , resembling in size and complexity replication origins in mammalian cells. Using purified proteins, we and others have shown that the S. pombe Orc4 subunit is solely responsible for ORC binding to multiple, specific AT-rich sites within S. pombe DNA replication origins (Kong & DePamphilis, 2001 ; Lee et al., 2001), suggesting that fission yeast replication origins differ significantly from those in the budding yeast, S. cerevisiae , which contain a single ORC binding site whose recognition requires at least five ORC subunits.  To further dissect S. pombe replication origins, we have analyzed ARS3001 for SpORC binding sites, pre-replication complex (pre-RC) assembly sites, and leading strand initiation sites (Kong and DePamphilis , 2002).  ARS3001 contains four genetically required DNA sites; D3 and D9 are absolutely required, and D2 and D6 are moderately required (Kim & Huberman, 1998).  Our results reveal that Orc4p binds strongly to D3, weakly to D6, and not at all to other sites in ARS3001. Orc4p preferentially binds the oligo(T) strand.  Moreover, in situ , the footprint over D3 extends into the adjacent D2 site during G1-phase; D2 is where leading strand DNA synthesis begins. Therefore, the ~100 bp of D3-D2 are similar to a single origin analogous to those in budding yeast.  D6 may faciliate pre-RC assembly at D3-D2, as suggested by analysis of other S. pombe origins (Takahashi et al., 2003).  Although D9 does not bind ORC, it does bind an as yet Takahashi unidentified protein throughout the cell cycle, and it is required specifically for origin function.  Therefore, D9 may represent a novel component of complex origins for which S. pombe is an appropriate paradigm.

To test this hypothesis, the abililty of Xenopus laevis ORC to initiate DNA replication in egg extract was challenged with SpOrc4.  The results revealed that XlORC preferentially targets the same AT-rich DNA sites selected by SpORC (Kong et al., 2003).  Notably, human ORC also binds preferentially to asymmetric AT-rich sequences (Vashee et al., 2003).

We are now exploring how transcription factors and Orc proteins facilitate initiation of site-specific DNA replication through their association with other proteins (for example of the technology, see Vassilev et al. 2001).