BIOLOGICAL OVERVIEW

The Drosophila homolog of the DNA replication initiation inhibitor Geminin has all of the properties of Xenopus and human Geminin. During Drosophila development, Geminin is present in cycling cells; protein accumulates during S phase and is degraded at the metaphase to anaphase transition. Overexpression of Geminin in embryos inhibits DNA replication, but cells enter mitosis arresting in metaphase, as in mutants of double parked (the Drosophila homolog of mammalian Cdt1), and undergo apoptosis. Overexpression of Geminin also induces ectopic neural differentiation. geminin mutant embryos exhibit anaphase defects at cycle 16 and increased numbers of S phase cells later in embryogenesis. In a partially female-sterile geminin mutant, excessive DNA amplification in the ovarian follicle cells is observed. These data suggest roles for Drosophila Geminin in limiting DNA replication, in anaphase and in neural differentiation (Quinn, 2001).

Strict regulation of DNA replication is essential for accurate propagation of the genetic material, since aberrant chromosome replication results in ploidy or mitotic defects, which can lead to tumorogenesis or cell death. Therefore, DNA replication must be restricted to once per cell cycle. To explain how DNA replication is limited to once per cell cycle, an agent termed replication licensing factor (RLF) has been proposed. RLF allows DNA to become replication-competent after mitosis, but once initiation occurs, RLF is inactivated until the next cell cycle. Biochemical purification of RLF has revealed that it is composed of the Mcm complex (RLF-M) and RLF-B (Cdt1) (see Tada, 2001). Cdt1 (Drosophila homolog: Double parked) together with Cdc6 promotes loading of the Mcm (see Disc proliferation abnormal) complex onto the Orc complex (see Origin recognition complex 2) at replication origins in M/G1 phase to form the pre-replication complex (pre-RC) and is inactivated after replication origin firing in S phase (Quinn, 2001 and references therein).

DNA replication initiation is best described in Saccharomyces cerevisiae, which has well-defined DNA replication origins. Entry into S phase is driven by the regulated activity of G1 cyclin/Cdk protein kinases and Cdc7/Dbf4 protein kinases by triggering recruitment of Cdc45 to the pre-RC, which allows the binding of DNA polymerase alpha-primase. Conversely, phosphorylation of Cdc6 and Mcms by Cdks or Ddk results in their release from the origins and clearance from the nucleus. Re-replication is prevented in S phase by disassembly of the pre-RC and in G2/M phase by mitotic Cyclin B/Cdk activity, maintaining Mcms in a hyper-phosphorylated state and inhibiting new synthesis of Cdc6. At the completion of mitosis, inactivation of mitotic cyclin-associated Cdk activity marks the end of inhibitory signals that block new pre-RC formation and sets the cell into a pre-replicative state (Quinn, 2001 and references therein).

In more complex eukaryotes, replication initiation is similar to yeast, however, there are a number of differences in the regulatory mechanisms as well as additional tiers of regulation. In contrast to yeast pre-RC, bound Cdc6 and Mcms are not degraded nor exported from the nucleus after the initiation of DNA replication. An important regulation to prevent re-replication in both yeast and metazoans is high Cdk activity, although low Cdk activity by itself is not sufficient to allow the re-initiation of DNA replication in G2 in the Xenopus cell free system. This additional regulation to prevent re-replication is provided by Geminin (McGarry, 1998; Wohlschegel, 2000; Tada, 2001). Geminin was originally identified in Xenopus as a protein that undergoes ubiquitin-dependent degradation in mitosis and as an inhibitor of DNA replication initiation in vitro that functions by preventing loading of Mcms onto chromatin (McGarry, 1998). Recently two studies have revealed that Geminin prevents Mcm loading by binding to and inhibiting Cdt1 activity (Wohlschegel, 2000; Tada, 2001). Consistent with Geminin functioning to inhibit Cdt1 and prevent re-replication, Geminin is high in G2 cells but is degraded in mitosis thereby allowing Cdt1 function and re-assembly of the pre-RC in G1. Geminin is present in mammals but has not been identified in budding or fission yeast or C. elegans. Therefore, Geminin appears to be restricted to more complex eukaryotes where it acts as an additional layer of regulation to prevent re-replication. Geminin also appears to have a role in inducing neural differentiation (Kroll, 1998). Many questions, however, remain unanswered concerning the in vivo role of Geminin (Quinn, 2001 and references therein).

Drosophila is an ideal model organism in which to study the regulation of DNA replication because the cell division patterns have been well characterized during development. The first 14 S phases of embryogenesis occur in a syncitium (without cell membranes) and consist of very rapid S-M cycles driven by maternally-supplied cell cycle proteins. At G2 of cycle 14, cellularization occurs and the next three divisions are regulated in G2 by zygotic expression of the mitotic inducer Cdc25 (String). Most cells arrest in G1 of cycle 17 attributable to down-regulation of Cyclin E with the notable exception of nervous system cells that undergo several more divisions and the gut cells that undergo endoreplication. In larval stages, gut tissues, including the salivary glands, continue to undergo endoreplication, whereas imaginal discs undergo G1/G2 regulated cell cycles. During oogenesis, endoreplication of the nurse cells and follicle cells that surround the egg chamber also occur. The role of the Drosophila geminin-related gene in the cell cycle and neural differentiation during development has been studied by analyzing Drosophila geminin mutants, the effects of overexpression, and the cell cycle distribution of the protein. These studies provide evidence that during Drosophila development, Geminin has roles in limiting DNA replication, in anaphase and in neural differentiation (Quinn, 2001).

Drosophila Geminin exhibits all the functional characteristics described for Xl and Hs Geminin: (1) it inhibits DNA replication in a Xenopus in vitro assay by blocking Mcms loading onto chromatin and also when overexpressed in flies; (2) it binds to Double parked (Dup) in vivo and genetically interacts with dup; (3) it accumulates in S/G₂ and is degraded at the metaphase to anaphase transition; and (4) when overexpressed, it induces neural differentiation. In addition, the expression of Geminin during development and the phenotype of geminin mutants are consistent with Geminin being a functional homolog of Xl and Hs Geminin (Quinn, 2001).

Geminin protein and mRNA expression is correlated with mitotically proliferating and endoreplicating cells. Geminin is absent in G₁/early S-phase cells, increases during S/G₂ and is degraded at the metaphase to anaphase transition in mitosis, similar to Cyclin B. The similar cell cycle profile observed for Geminin and Cyclin B suggests that these proteins may be degraded by the same mechanism. The destruction box motifs of Drosophila and Xl Geminin, however, show variations compared with the canonical destruction motifs of Xl or Drosophila Cyclin B. Furthermore, whereas Drosophila Cyclin B is degraded by ubiquitin-mediated proteolysis in Xenopus oocyte mitotic extracts, Geminin is not, suggesting that Geminin and Cyclin B may be degraded by different mechanisms in Drosophila. Also, because Xl Geminin is degraded in Xenopus oocyte mitotic extracts (McGarry, 1998) and Drosophila Geminin is not, there may be species-specific differences in the Geminin degradation pathway (Quinn, 2001).

Ectopic expression of Geminin in the embryo can induce some cells to ectopically differentiate as neural cells. The reason that not all cells expressing Geminin differentiate into neural cells as occurs with Xl Geminin (Kroll, 1998) may be attributable to induction of apoptosis or to the requirement of other neuralizing factors. Ectopic neural differentiation may explain the severity of the eye phenotype observed when Geminin is ectopically overexpressed using the GMR driver, which cannot simply be explained by an S-phase inhibitory effect. GMR-driven Geminin results in only a 40%-50% decrease in S-phase cells in the eye imaginal disc, whereas the eye phenotype is more severe than that observed when post-MF S phases were almost completely ablated by GMR-human p21. GMR-driven expression of the apoptosis inhibitor p35 partially rescues the eye phenotype by increasing the size of the eyes, but they are still disorganized, suggesting that apoptosis contributes to the rough eye phenotype but that other mechanisms, perhaps ectopic neural differentiation, may also be involved. Further analysis is required to determine whether Geminin has a physiological role in neurogenesis during development. If Geminin is involved in neurogenesis, however, it must function at an early stage in neural differentiation since Geminin expression in Xenopus and Drosophila is specific to cells that are actively cycling (Quinn, 2001).

Evidence of over-replication was observed in geminin mutants in (1) endoreplicating tissues in stage 16 embryos; (2) the CNS in stage 16 embryos, and (3) DNA amplification cycles of ovarian follicle cells. These data are consistent with excessive DNA replication occurring in geminin mutants. Together with the inhibition of DNA replication by Geminin overexpression, these data show that Geminin has a role in limiting DNA replication. Because Geminin is absent before S-phase initiation in the mitotic cycles, the presence of Geminin may be sufficient to limit DNA replication. Geminin is present throughout the endoreplication and amplification cycles, suggesting that other factors are required to activate Geminin to limit DNA replication in these cycles (Quinn, 2001).

The inability to observe excessive DNA replication early in embryogenesis is most likely attributable to the maternal supply of Geminin. In Xenopus, however, in vitro DNA replication assays, antibody ablation of >99% of endogenous Xl Geminin results in no apparent re-replication as measured by BrdU labeling and density gradient analysis (McGarry, 1998). Furthermore, the terminal phenotype of Geminin depletion from Xenopus eggs causes abnormalities at the blastula stage that is different to the Drosophila geminin mutant phenotype. The failure to observe excessive replication in Geminin-ablated Xenopus embryos may indicate that DNA replication licensing is regulated differently depending on developmental times or in different organisms (Quinn, 2001).

Functional analysis of Xl Geminin has separated the neuralizing and DNA replication inhibition domains, suggesting that Geminin is dual functional (Kroll, 1998; McGarry, 1998). Dual functional proteins that can act as negative cell cycle regulators and differentiation inducers are not unique because Cdk inhibitors of the p21 family and the Rb family also function in this manner. Interestingly, the Cdk inhibitors human p21^CIP1 and Xenopus p27^XIC1 also have separable cell cycle inhibition and differentiation functional domains. Whether the neuralization and DNA replication inhibition functions of Drosophila Geminin are also functionally separable is currently being investigated (Quinn, 2001).

The anaphase defects observed in cycle 16 geminin mutant embryos may indicate that Geminin also has a role in anaphase. However, given the evidence for Geminin in the block to re-replication in other organisms and the observation of over-replication defects later in development, the most likely reason for the anaphase defects in the geminin mutant cycle 16 embryos is caused by prior S-phase defects. Although BrdU-labeling studies have not revealed over-replication defects before mitosis of cycle 16, a low amount of re-replication would be difficult to detect. Over-replication defects that fail to elicit the DNA replication or damage checkpoints, allowing entry into mitosis, would result in drastic consequences. If chromosomes were partially over-replicated and the over-replicated portion included the centromere, then chromosome bridges would be expected to occur during anaphase. Therefore, the anaphase defects observed may be a manifestation of over-replication in geminin mutants. If this is the case, then Geminin has an important role in ensuring that DNA re-replication does not occur before chromosome segregation (Quinn, 2001).

The anaphase defects of geminin mutant embryos are similar to those observed in mutants of Drosophila lodestar, topI, and barren, which are involved in decatenization and unwinding of replicated DNA. Whether the defects observed in geminin mutants are related to chromosome decatenization and whether lodestar, topI, and barren also show over-replication defects, requires further investigation (Quinn, 2001).

In summary, this study describes the first analysis of the role of Drosophila geminin during development. geminin mutants show over-replication defects, anaphase defects, and neural differentiation defects. Together with the analysis of the ectopic expression of Geminin, these data suggest that Geminin has functions in preventing re-replication, in anaphase and in neural differentiation (Quinn, 2001).

GENE STRUCTURE

cDNA clone length - 891

Exons - 1

PROTEIN STRUCTURE

Amino Acids - 192

Structural Domains

From analysis of the Berkeley Drosophila Genome Project (BDGP) database, a sequence (CG3183) at 42B3 on chromosome 2R has been identified that shows weak, but significant, similarity to Geminin from Xenopus (Xl Geminin) and human (Hs Geminin). By DNA sequencing of cDNA clones, the largest cDNA was determined to be 891 bp, encoding a 192-amino-acid protein (GenBank accession no. AF407275). Analysis of the gene structure reveals that it is an intronless gene. Three P element alleles l(2)k09107, l(2)k14019, and l(2)k03202 were found to be inserted within or 5' to the transcription unit (Quinn, 2001).

The Drosophila Geminin-related sequence shows only limited identity overall with Xl and Hs Geminin, although when specific functional regions are compared, the level of identity is much greater. Geminin can be divided into two functional domains --the DNA replication inhibition domain containing a coiled-coil motif at the carboxyl terminus and a neuralization domain at the amino terminus (Kroll, 1998; McGarry, 1998). The Drosophila Geminin-related sequence shows weak sequence conservation with Xl and Hs Geminin in the neuralization domain but higher levels of homology are observed with the DNA replication inhibition domain and with the coiled-coil motif. The Drosophila Geminin-related sequence has a potential destruction box for ubiquitin-mediated degradation in mitosis (RxALGVIxN) at its amino terminus, which shares a 5/9 match with the Xl Geminin H sequence shown to mediate degradation of the protein (McGarry, 1998). Therefore, although the overall sequence similarity of the Drosophila Geminin-related protein is considerably lower than that observed between Xl and Hs Geminin, the basic arrangement of the protein is conserved (Quinn, 2001).

date revised: 25 November 2001

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.