HIGHLIGHTS

TRANSCRIPTION cells. The researchers found that, in Gata-3 that induces IL-4 in Mbd2-

Mbd2-deficient cells, the normal pat- mutant TH1 cells. In line with this tern of T-cell differentiation was dys- observation, the authors proposed a regulated. For example, Mbd2-defi- quantitative competition model for Competing forces cient cells produced ectopic activators (such as Gata-3) and

interleukin (IL)-4 under TH1-polariz- silencers (such as Mbd2) whereby Chromatin structure has long been ing conditions, and the normally each regulator has an independent suspected to have a role in regulating essential transcriptional activator quantitative effect. cell fate but there has been no con- Gata-3 was dispensable for IL-4 To test this hypothesis, cells with vincing evidence for this. However, induction. Moreover, expression of three different Mbd2 gene doses data published in Molecular Cell now the signature genes encoding IL-4 and (Mbd2+/+, Mbd2+/−, Mbd2−/−) were γ support a model whereby transcrip- interferon- was increased in TH2 and cultured under TH1 conditions (that

tion factors and epigenetic silencing TH1 cells, respectively, suggesting that is, where Gata-3 is not produced) and are integrated quantitatively to con- Mbd2 might be crucial in silencing transduced with a bicistronic retro-

trol cell differentiation. these cytokine genes during TH-cell virus encoding Gata-3 and green fluo- Hutchins et al. used helper-T-cell differentiation. rescent protein (GFP). Cells with dif-

(TH cell) differentiation as a model To investigate the relationships ferent levels of Gata-3 were evaluated

system, and analysed TH cells from between both Gata-3 and Mbd2, and based on the level of green fluores- mice deficient in methyl-CpG-bind- the IL-4 gene, the authors determined cence intensity. Increasing levels of ing-domain protein 2 (Mbd2), one of relative Gata-3 messenger-RNA levels Gata-3 rendered cells competent for the molecules that is thought to link using the reverse-transcription poly- IL-4 production in a manner that is DNA methylation and chromatin- merase chain reaction. Gata-3 mRNA inversely correlated with Mbd2 gene based silencing. levels were repressed as normal in dosage.

In this study, cytokine signalling cells cultured under TH1 conditions To investigate the basis for the was manipulated to generate polarized regardless of the Mbd2 genotype, competition between Gata-3 and

TH1 and TH2 cells from uncommitted indicating that it is not an excess of Mbd2 further, the authors performed

CELL CYCLE anaphase-promoting complex (APC). This mitotic exit can progress without either Clb5 is activated by a molecule called Cdc20, degradation or Sic1. which then recruits . CDKs Next, Wäsch and Cross analysed the mitotic phosphorylate and thereby activate Clb2 as a potential Cdc20 target, and Cyclin’ around Cdc20–APC, which triggers cyclin showed that CLB2∆db yeast cells were unable destruction and mitotic exit. As a result, to exit . Clb2 might therefore be the the CDK activity decreases, which causes cyclin that needs to be destroyed for mitotic Cyclin destruction and the associated drop APC inactivation and re-accumulation of exit to occur. in cyclin-dependent kinase (CDK) activity cyclin. Previous studies had shown that cells that are necessary for mitotic exit. In yeast, The cell cycle in somatic cells and yeast, lacked Cdh1 and Sic1 are not viable, degradation of the S-phase cyclin Clb5 was however, is more complex because it suggesting that both proteins are essential thought to be essential, but a study includes a that is important for for mitotic exit, and that the Cdh1–Sic1 reported in Nature now raises doubts cell growth and differentiation. G1 oscillator might be the more important about the importance of Clb5 in regulating regulation might be achieved by a second regulator. By contrast, Wäsch and Cross mitotic exit. oscillator that differs from the first one in found that cells lacking these proteins do In frog embryos, a single oscillator can using the APC-regulatory subunit Cdh1 not have a marked defect in mitotic exit, manage the cell cycle. This negative- instead of Cdc20. CDKs inhibit Cdh1–APC, and indeed survive, although poorly. feedback oscillator alternates between S which means that, in late mitosis, CDK However, the cells showed some and M phases, and involves the regulation proteolysis causes activation of Cdh1–APC, abnormalities associated with an unstable of a ubiquitin–protein ligase called the which then maintains cyclin destruction G1 phase. This finding also highlights the throughout G1 (see diagram). important role of the Cdc20 oscillator in Mitotic Mitotic entry exit To address the question of whether the destroying Clb2. Clb5 cyclin must be destroyed, Wäsch and So, the destruction of mitotic cyclins such CDK Cross constructed a yeast strain (CLB5∆db) as Clb2 is crucial to mitotic exit and the in which the CLB5 gene lacked the Cdc20 and Cdh1–Sic1 oscillators seem to

Cdc20– Cdh1– destruction-box sequence that targets Clb5 work in a complementary manner, with the Cyclin APC APC to Cdc20–APC. However, these cells did not latter being important for a stable G1 phase. seem to have any defect in mitotic exit, Arianne Heinrichs implying that proteolysis of Clb5 is not References and links required. Exit also occurred in CLB5∆db ORIGINAL RESEARCH PAPER Wäsch, R. & Cross, F. R. cells that lacked Sic1, a CDK inhibitor APC-dependent proteolysis of the mitotic cyclin Clb2 is essential for mitotic exit. Nature 418, 556–562 (2002) S/G2MMG1 S/G2 G1 responsible for further reducing CDK FURTHER READING Morgan, D. O. & Roberts J. M. © (2002) Macmillan Magazines Ltd. activity in late mitosis. This suggests that Oscillation sensation. Nature 418, 495–496 (2002)

632 | SEPTEMBER 2002 | VOLUME 3 www.nature.com/reviews/molcellbio

© 2002 Nature Publishing Group