Juan Méndez

Cyclin-dependent kinases (CDKs) have been likened to the “traffic lights” of the cell cycle (Nasmyth, 2001). They promote and coordinate DNA replication during S phase and chromosome segregation during mitosis. They also ensure that nuclear DNA is replicated once and only once in each cell cycle. In mammalian cells, the D type cyclins (D1, D2, D3) associate with CDK4 and CDK6 kinases, and their primary function is to phosphorylate the retinoblastoma (Rb) family of transcriptional repressors during G1. The E type (E1, E2) and A type (A1, A2) cyclins associate with CDK2 and regulate initiation of DNA replication and progression through S phase. Finally, the B type cyclins (B1, B2, B3) associate with CDK1 to control entry into and exit from mitosis.

It turns out that mouse cells may survive with a couple of “broken lights.” Geng et al. (this issue of Cell) have generated strains of mice in which cyclins E1 or E2 have been deleted. Both E1−/− and E2−/− mice develop normally and are viable, except that half the E2−/− males are sterile due to incomplete testis development. Crossing of the two transgenic strains did not yield viable double knockout mice. However, E1−/−E2−/− embryos survived until the 10th day of gestation. Prenatal death was caused by the mutant placenta, and some E1−/−E2−/− embryos successfully completed gestation when they developed in a wild-type placenta. In culture, fibroblasts derived from the E1−/−E2−/− embryos were able to undergo several rounds of division before becoming senescent. These studies led to the rather provocative conclusion that cyclin E is dispensable for cell proliferation and development in the mouse.

In an independent study, Ortega et al. (2003) discovered that CDK2, the kinase activated by E type and A type cyclins, is also dispensable for mouse development. CDK2−/− animals are viable and do not display anatomical or behavioral abnormalities, except for severe gonad atrophy. In addition, CDK2−/− embryo fibroblasts proliferate normally in culture, indicating that CDK2 activity is not necessary for mitotic cell division. In contrast, CDK2 appears to be essential for meiosis, for CDK2−/− mice are sterile due to a block in spermatogenesis and oogenesis during prophase I.

The fact that E type cyclins are not essential for fibroblast proliferation is not without precedent. In yeast, the cyclins that promote entry into S phase can be deleted without major consequences because the mitotic cyclins can take over their functions (reviewed by Kelly and Brown, 2000). There are also cases of cyclin complementation in mammalian cells: a previous study from P. Sicinski's laboratory revealed that cyclin E introduced at the endogenous cyclin D1 locus rescued the phenotypic defects of cyclin D1−/− mice (Geng et al., 1999). Therefore, it is a reasonable assumption that cyclin A can complement the lack of cyclin E, at least during mitotic cycles.

In the case of CDK2−/− cells, viability is likely due to surrogate kinases associating with cyclin A and cyclin E. It is known that cyclin A can associate with CDK1, but it is not clear whether cyclin E can associate with a kinase other than CDK2. In fact, an immunoprecipitate of cyclin A from CDK2−/− cell extracts contained kinase activity, whereas an immunoprecipitate of cyclin E did not (Ortega et al., 2003). However, the fact that CDK2−/− mice are viable but cyclin E1−/−E2−/− mice die in utero suggests that cyclin E has an essential role that can be executed independently of CDK2. Further studies will be necessary to address the intriguing possibility that this putative CDK2 independent function of cyclin E does not require an associated kinase activity.

Detailed analyses of the cyclin E1−/−E2−/− mice revealed that cyclin E is essential in particular cell cycle situations (Figure 1) . First, the defective placenta in the E1−/−E2−/− mice indicated a role for cyclin E during endoreplicative cell cycles (Figure 1B). Like every rule, the canon “once and only once in each cell cycle” has exceptions: certain cell types replicate their DNA multiple times without intervening cell divisions, because genomic amplification is required for their physiological function. The DNA content of the giant trophoblasts found in a normal placenta can be as high as 1000C. In contrast, the majority of trophoblasts of the E1−/−E2−/− placenta were severely underdeveloped. The requirement for cyclin E in endoreplication was further confirmed by analysis of the platelet precursor megakaryocytes from E1−/−E2−/− mice, which accumulated much less DNA than similar cells from wild-type mice. Interestingly, cyclin E is a central regulator of all Drosophila endocycles (reviewed by Edgar and Orr-Weaver, 2001).

 
Figure 1. Cellular Processes that Strictly Require Cyclin E Activity

(A) Exit from quiescence.

(B) Endoreplication cycles.
 

View larger version: [In this window] [In new window]
 

In addition, serum starvation experiments with the cyclin E1−/−E2−/− fibroblasts uncovered a specific role for cyclin E during the exit from the quiescent G0 state (Figure 1A). Upon mitogenic stimulation of wild-type cells, the Rb protein is phosphorylated and releases the E2F1-3 transcription factors, which drive the expression of the proteins required to license the origins of DNA replication, such as the MCM (minichromosome maintenance) proteins. In the E1−/−E2−/− fibroblasts, the Rb protein was phosphorylated, presumably by cyclin A-CDK2, but one of the MCM proteins failed to associate with the chromatin, a key step for origin activation and hence the initiation of DNA replication. This interesting observation is in agreement with a previous report demonstrating that, in quiescent 3T3 nuclei induced to replicate their DNA in a cell-free system, cyclin E opens the “window of opportunity” to activate origins of replication (Coverley et al., 2002). In this context, cyclin E activity could be required to phosphorylate MCM proteins and promote their loading onto chromatin, or to inactivate an inhibitor of MCM loading, such as geminin or the Ran-GTP/Crm-1 complex (McGarry and Kirschner, 1998; Yamaguchi and Newport, 2003).

Finally, Geng et al. (2003) show that cyclin E1−/−E2−/− fibroblasts are resistant to oncogenic transformation. Conversely, cyclin E is upregulated in a variety of human tumors (reviewed by Malumbres and Barbacid, 2002). Perhaps cyclin E is required to overcome the senescence crisis preceding cell transformation, as it is required to overcome the quiescent state caused by serum starvation.

With these recent reports, many of the key regulators of the cell cycle have been deleted individually in animal models. Further crosses between those mutant mice that are fertile will test the redundancy limits of the CDK network. How many traffic lights can be broken before we get into a jam?

Coverley, D., Laman, H., and Laskey, R.A. (2002). Nat. Cell Biol. 4, 523-528. [Medline]

 Edgar, B.A. and Orr-Weaver, T.L. (2001). Cell 105, 297-306. [Medline] [Full Text]

 Geng, Y., Whoriskey, W., Park, M.Y., Bronson, R.T., Medema, R.H., Li, T., Weinberg, R.A., and Sicinski, P. (1999). Cell 97, 767-777. [Medline] [Summary] [Full Text]

 Geng, Y., Yu, Q., Sicinska, E., Das, M., Schneider, J.E., Bhattacharya, S., Rideout, W., Bronson, R.T., Gardner, H., and Sicinski, P. (2003). Cell 114, 431-443. [Medline] [Summary] [Full Text]

 Kelly, T.J. and Brown, G.W. (2000). Annu. Rev. Biochem. 69, 829-880. [Medline]

 Malumbres, M. and Barbacid, M. (2002). Nat. Rev. Cancer 1, 222-231. [Medline]

 McGarry, T.J. and Kirschner, M.W. (1998). Cell 93, 1043-1053. [Medline] [Summary] [Full Text]

 Nasmyth, K. (2001). Cell 107, 689-701. [Medline] [Full Text]

 Ortega, S., Prieto, I., Odajima, J., Martin, A., Dubus, P., Sotillo, R., Barbero, J.L., Malumbres, M., and Barbacid, M. (2003). Nat. Genet., (10.1038/ng1232) ; published online 17 August, 2003.

 Yamaguchi, R. and Newport, J. (2003). Cell 113, 115-125. [Medline] [Summary] [Full Text]