BIOLOGICAL OVERVIEW

Drosophila Cdc2 (cdc stands for cell division control) is the catalytic subunit (the cyclin dependent kinase or cdk) of the cyclin/cdk heterodimer responsible for the execution of the mitotic (M) phase of the cell cycle. While cyclin (the regulatory subunit in all cyclin/cdk heterodimers) alternates between synthesis and degradation during the cell cycle, steady levels of the cdk are maintained; cdk activity is regulated through phosphorylation. The cyclin dependent kinase Cdc2, also known as Cdk1, is maintained in an inactive hyperphosphorylated state during G1, S and G2 stages. The inhibition of Cdc2 activation during these stages involves phosphorylation at conserved residues threonine 14 and tyrosine 15, which overlap the ATP binding site of Cdc2.

The dephosphorylation of Cdc2 is highly regulated by what is termed checkpoint control. Checkpoint control ensures that M phase entry depends on the successful completion of S phase; entry into M phase is allowed only if DNA is fully replicated and is not damaged. Cdc2 is suddenly dephosphorylated by the phosphatase known as String at the boundary between G2 and M, thus driving the cell into mitosis. At this point Cdc2 fulfills a second biological role; the promotion of M phase. Cdc2 targets many cellular proteins involved in mitosis, activating them by catalytically adding phosphate residues to threonine and serine residues. In this second biological role, the Cdc2/Cyclin heterodimer is termed the mitosis promoting factor, or MPF. For more information about the G2 checkpoint, see Cyclin B.

Although the activity of Cdc2 is central to checkpoint control, inputs from developmentally regulated transcription factors, such as Escargot, are required. Escargot is involved in maintaining diploidy in imaginal cells, ensuring that these cells retain the ability to proliferate and differentiation upon metamorphosis. This requirement exemplifies the complex regulation that occurs at the G2/M boundary (Hayashi, 1996).

The mutant phenotypes of cdc2 are similar to those of escargot: many diploid cells in imaginal discs, salivary glands and the central nervous system enter an endocycle, characterized by DNA replication without a subsequent mitotic phase. Such endocycling cells are often polytene, possessing thick chromosomes with DNA replicated many times over. When escargot function is eliminated, diploid imaginal cells that were arrested in G2 lose Cyclin A, a regulatory subunit of G2/M cdk, and entered endocycle. escargot genetically interacts with cdc2, suggesting an intimate biological interaction. Since mitotically quiescent abdominal histoblasts still require cdc2 to remain diploid, the inhibitory activity of Cdc2 on DNA replication appears to be separable from its activity as the mitosis promoting factor. These results suggest that in G2, escargot is required to maintain a high level of G2/M cdk, which actively inhibits the entry into S phase. Expression of Cyclin A is lost in escargot mutants, suggesting that Cdc2 activity (dependent on its regulatory subunit Cyclin A), indirectly depends on escargot (Hayashi, 1996).

The notion of a correlation between cell fate and time of cell division remains a cornerstone of theories that explain the origin of patterns during development. For example, in the pupal wing disc, at pupariation, cells arrest in G2 before progressing through two programmed cell divisions, separated by an intervening S-phase. Cells in vein regions divide in a reproducible pattern before the cells in the adjacent intervein regions. In the early developing embryo, cells with a common developmental fate can be recognized as mitotic domains after cellularization because they divide almost synchronously at defined developmental stages. In the growing amphibian limb, treatment with colchicine inhibits cell divisions, reduces the size of the limb and also reduces the number of digets formed (references in Weigmenn, 1997).

During larval development, Drosophila imaginal discs increase in size about 1000-fold; cells are instructed to acquire distinct fates as a function of their position. The secreted signaling molecules Wingless and Decapentaplegic have been implicated as sources of positional information that globally control growth and patterning. Evidence has also been presented that local cell interactions play an important role in controlling cell proliferation in imaginal discs. A test was made of the role of cell division in pupal wing patterning in Drosophila by inactivating the mitotic kinase Cdc2 in developing imaginal discs. Clones of temperature sensitive cdc2 mutant cells were generated by mitotic recombination and the resulting patterns of tissue development were observed (Weigmenn, 1997).

Cell growth was seen to continue after inactivation of Cdc2, with little effect on overall patterning. The final size and shape of the pupal wing is not affected by inhibiting pupal cell division. Blocking cell division does not affect the spatial and temporal patterning of DNA replication in pupal wings. DNA replication is seen to continue at its normal pace in the absence of cell division, and larger than normal cells are generated, these cells attaining the architectural arrangement that smaller, normal cells achieve in the wild-type situation. Thus cell growth continues in the absence of cell division during larval stages and inhibition of cell division in mid-larval stages fails to block the patterning process. It is concluded that the mechanisms that regulate the size of the disc do not function by regulating cell division, but appear to act primarily by regulating size in terms of physical distance or tissue volume. In other words, overall size and shape of discs with mutant clones is unaffected and therefore disc size and structure is independent of cell number and pattern of cell division (Weigmann, 1997).

There are clear limitations to the universality of the above observations. If larvae are shifted to restrictive temperatures at early stages of development, wings fail to grow to normal size. In this case, where the cell number and overall dimensions of the disc are greatly reduced, pattern elements of gene expression do not resolve properly. This is consistent with the results of experiments with growing amphibian limbs (Weigmann, 1997).

There is already some precedence for these findings in the literature of Drosophila development. Simpson and Morata (1981) produced small clones of genetically marked, but otherwise wild-type cells in discs carrying large clones of Minute+ cells (which had been induced earlier). Mutant Minute cells divide more slowly than wild-type Mutant+ cells. Clones of wild-type cells located near the edge of the Minute+ clone grow larger than genetically equivalent clones far from the clone border. These observations suggest that cells within the Mutant+ territory but near the Minute mutant territory are stimulated to proliferate more strongly than cells surrounded by wild-type cells. It is concluded that the ability of cells to respond appropriately to long range patterning cues is relatively independent of cell number and that cell size can continue to increase if cell division is blocked, thus compensating for the presence of fewer cells (Weigmann, 1997). It might be that within a certain range, normal patterning mechanisms and compensatory mechanisms take presence over a deficiency in cell number to attain a genetically preprogrammed pattern.

Studies in unicellular systems have established that DNA damage by irradiation invokes a checkpoint that acts to stall cell division. During metazoan development, the modulation of cell division by checkpoints must occur in the context of gastrulation, differential gene expression and changes in cell cycle regulation. To understand the effects of checkpoint activation in a developmental context, a study was performed of the effect of X-rays on post-blastoderm Drosophila embryos. In Drosophila, DNA damage delays anaphase chromosome separation during cleavage cycles that lack a G2 phase. In post-blastoderm cycles that include a G2 phase, irradiation delays the entry into mitosis. Gastrulation and the developmental program of string (Cdc25) gene expression, which normally regulates the timing of mitosis, occurs normally after irradiation. The radiation-induced delay of mitosis accompanies the exclusion of mitotic cyclins from the nucleus. Furthermore, a mutant form of the mitotic kinase Cdk1 that cannot be inhibited by phosphorylation drives a mitotic cyclin into the nucleus and overcomes the delay of mitosis induced by irradiation. It is concluded that developmental changes in the cell cycle, for example, the introduction of a G2 phase, dictate the response to checkpoint activation, for example, delaying mitosis instead of or in addition to delaying anaphase. This unprecedented finding suggests that different mechanisms are used at different points during metazoan development to stall cell division in response to checkpoint activation. The delay of mitosis in post-blastoderm embryos is due primarily to inhibitory phosphorylation of Cdk1, whereas nuclear exclusion of a cyclin-Cdk1 complex might play a secondary role. Delaying cell division has little effect on gastrulation and developmentally regulated string gene expression, supporting the view that development generally dictates cell proliferation and not vice versa (Su, 2000).

To examine the effect of DNA damage on the progression of the cell cycle during Drosophila embryogenesis, embryos 0-4.5 hours of age were exposed to 570 rads of X-rays. At this dose, 40%-60% of cellular embryos die and fail to hatch into larvae. This dose therefore corresponds to the half-maximal lethal dose (LD50). When syncytial embryos are exposed to X-rays: nuclei enter mitosis normally but chromosome segregation is delayed. The delay is transient such that nuclei enter the next interphase without completely separating sister chromosomes, resulting in polyploid nuclei (Su, 2000).

In cellularized embryos, changes in cell cycle indicators that are consistent with a delay in the entry into mitosis are observed. In untreated embryos at these stages, cells divide in stereotypical clusters termed 'mitotic domains'. Both the location of a mitotic domain within the embryo and the time at which it goes through mitosis are invariant from embryo to embryo. The timing of morphogenetic movements that comprise gastrulation is likewise invariant from embryo to embryo. Thus, the wild-type pattern of mitotic cells at any specific time during this period, as indicated by the extent of gastrulation, is easily recognizable. In irradiated samples, embryos were found in which expected mitotic domains were not in mitosis, as judged by the absence of condensed chromosomes and mitotic figures. Antibody staining for a mitotic-specific phospho-epitope on histone H3 (PH3), and staining with wheatgerm agglutinin (WGA) to detect nuclear envelope breakdown, has confirmed the absence of mitoses in these embryos. It is inferred that irradiation delays the entry into mitosis in cellularized embryos, whereas under identical conditions, chromosome segregation is delayed in syncytial embryos. Treatment of cellularized embryos with a DNA-damaging agent, methyl methane sulfonate, results in a similar delay of mitosis. Therefore, the observed effect of irradiation on mitosis is probably due to the DNA-damaging activity of X-rays (Su, 2000).

It is an unprecedented finding that irradiation leads to two different cell cycle responses in a single organism: either the delay of anaphase chromosome segregation or the delay of mitosis. Mitotic chromosome segregation and the initiation of mitosis are regulated by different mechanisms. The former requires the proteolysis of proteins, such as PDS1 in budding yeast and cyclin A in Drosophila, whereas the latter requires the activation of mitotic cyclin-Cdk complexes. It is suggested that checkpoint activation by the same dose of radiation under identical conditions must have used different downstream mechanisms in order to delay chromosome segregation in the syncytium and mitosis in the cellularized embryos. Although mechanisms that operate in the syncytium remain elusive, the mechanisms used by cellularized embryos were addressed in this study (Su, 2000).

Despite the finding that irradiation does not interfere with String expression, it might have antagonized String activity. Cdc25Stg activates Cdk1 by removing the inhibitory phosphates on Thr14 and Tyr15. A Cdk1 mutant in which these residues have been mutated (Cdk1AF) bypasses the requirement for String. If the mechanism by which radiation delays mitosis solely involves inhibitory phosphorylation of Cdk1, Cdk1AF should bypass the radiation-induced delay. To test this hypothesis, Cdk1 or Cdk1AF, in conjunction with a mitotic cyclin, was expressed from a heat-inducible (hs) promoter during interphase 14. It was then asked whether irradiation could delay the onset of mitosis 14 in embryos expressing these transgenes. It was found that many cells of heat-shocked embryos that carried hs-Cdk1AF and hs-mitotic cyclin transgenes fail to delay mitosis after irradiation. This effect was seen with mitotic cyclins A, B or Bs -- a truncated version of cyclin B that is resistant to proteolysis. In contrast, embryos carrying hs-Cdk1, in combination with the same cyclins, behave like wild-type embryos and delay mitosis. It is concluded that Cdk1AF, and not Cdk1, can overcome the radiation-induced delay in mitosis. It is inferred that inhibitory phosphorylation on Cdk1 is required to delay mitosis in response to DNA damage, in agreement with previous results from fission yeast and vertebrates (Su, 2000).

Interestingly, the ability of Cdk1AF and cyclins to overcome the delay of mitosis in Drosophila was seen only in certain cells of the embryos, and these cells represent mitotic domains, for example, domain 4. Cells of mitotic domains are distinguished from their neighbors by their accumulation of String protein. Although further experiments are required to demonstrate the importance of String, the perfect coincidence of clusters of irradiated cells that entered mitosis in the presence of Cdk1AF as well as accumulated String, has led to the following suggestion: although cyclin-Cdk1AF activity is not present in sufficient quantities to promote mitosis by itself under these experimental conditions, this activity can induce endogenous String to activate endogenous Cdk1 and induce mitosis. A similar feedback mechanism has been proposed for human Cdk1 and Cdc25. It follows then that endogenous String and Cdk1 might be inhibited by irradiation, but that this inhibition can be overcome by a small amount of Cdk1AF activity (Su, 2000).

The same amount of Cdk1AF activity overcomes another consequence of irradiation, namely, the nuclear exclusion of a mitotic cyclin. Nuclear cyclin/Cdk1 activity is a prerequisite to mitosis and the exclusion of cyclin B1 from the nucleus appears to contribute to the delay in mitosis after irradiation in human cells. In cellular-stage Drosophila embryos, cyclins A and B remain enriched in the cytoplasm in interphase. Cyclin A accumulates in the nucleus of cells that initiate mitosis, as does cyclin B. In irradiated embryos, both cyclins A and B are excluded from nuclei although their levels remain unchanged. In cells that express Cdk1AF (with a mitotic cyclin) that enter mitosis even after irradiation, nuclear accumulation of cyclin A is evident. Thus, a low level of Cdk1 activity, provided by Cdk1AF in these experiments, leads to both the nuclear accumulation of a cyclin and the entry into mitosis (Su, 2000).

Given these two observations -- that Cdk1AF drives the nuclear accumulation of cyclin A and that nuclear accumulation of mitotic cyclins coincides with the entry into mitosis in unperturbed cell cycles -- it has been proposed that Cdk1 activity normally drives the nuclear accumulation of cyclin-Cdk1 complexes. In support of this idea, Cyclin A remains excluded from nuclei in string mutants. In accordance with this, Cdk1AF, in conjunction with endogenous String, overcomes the radiation-induced delay of mitosis because Cdk1AF can start the feedback loop that activates endogenous Cdk1 by endogenous String and Cdk1 activity can drive the nuclear accumulation of cyclin-Cdk1. These ideas help explain previous observations in human cells. In the latter, although the exclusion of cyclin B1 from nuclei appears to be of some importance to regulating mitotic entry, Cdk1AF can overcome the checkpoint-induced delay of mitosis, regardless of whether cyclin B1 or NLS-cyclin B1, which is constitutively localized to the nucleus, is co-expressed. Thus, Cdk1AF in human cells, as in Drosophila, might also drive the nuclear accumulation of cyclin-Cdk complexes and the entry into mitosis by initiating a positive feedback loop for the activation of endogenous Cdk1. Whether a similar feedback loop of Cdk1, String and cyclin-localization operates to control mitosis in other tissues, such as larval imaginal discs, remains to be seen (Su, 2000).

GENE STRUCTURE

Bases in 5' UTR - 45

Bases in 3' UTR - 99

PROTEIN STRUCTURE

Amino Acids - 297

Structural Domains

The two Drosophila Cdc2 proteins, Cdc2 and Cdc2c exhibit a 56% sequence identity. Cdc2 is more closely related to the identified Cdc2 sequences of other species than is Cdc2c. This suggests that the evolutionary separation of Cdc2 and Cdc2c is ancient. The major differences between the Drosophila homologs are confined to regions that are not conserved among known p34cdc2 homologs (Lehner, 1990). Cdc2c functioning during G1 and S phases, partnering Cyclin E, is required for progression through S phase of the mitotic cycle (Knoblich, 1994).

date revised: 10 June 2000 

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.