<<

NEWS AND VIEWS RASSF1A, the new guardian of

Endre Máthé

The tumor suppressor gene RASSF1 contributes to the spatiotemporal regulation of mitosis through a new mechanism. By interacting with Cdc20, its protein product RASSF1A inhibits the -promoting complex and prevents degradation of A and until the spindle checkpoint becomes fully operational.

Transcriptional silencing of the gene RASSF1 mosome alignment. In the case of unattached RASSF1A-Cdc20 connection (Ras association domain family 1) has been or uneven tension across the The mitotic function of RASSF1A came to observed in many cancers, and its tumor-sup- sister chromatids, the act as sen- the attention of Song and colleagues5 when pression function has been implicated in sors and foster assembly of inhibitory proteins they tried to ectopically overexpress the gene reducing the proliferation of cancer cells1,2. Mad2 and Mad3 onto APC/C or Cdc20, in human lines and observed that these RASSF1A has a Ras association domain like thereby halting the substrate-targeting and cells accumulated in prometaphase with lev- that of Ras effectors; it heterodimerizes with ubiquitinating activity of APC/C-Cdc20. els of and cyclin B remaining high. the Ras-GTP binding protein Nore1; and it is When all chromosomes have their sister kine- This suggested that the increased levels of predicted to exert its function through a Ras tochores attached to microtubules and face RASSF1A arrested mitosis before the signal transduction pathway3. Several inter- opposite spindle poles, the spindle checkpoint -anaphase transition. Indeed, phase-specific functions have been proposed inhibition ceases, permitting APC/C-Cdc20 to depletion of RASSF1A by RNA interference for RASSF1A, although the signaling path- initiate anaphase and sister-chromatid separa- (mimicking the naturally occurring gene ways involved have not yet been identified3,4. tion by ubiquitinating the separase inhibitor, silencing observed in tumors) accelerated http://www.nature.com/naturegenetics In this month’s issue of Nature Cell Biology, securin. Additional regulatory mechanisms mitotic progression. The destruction of Min Sup Song and colleagues5 identified control local populations of APC/C-Cdc20, as cyclin A and cyclin B was initiated earlier in RASSF1A as a mitosis-specific inhibitor of the it does not ubiquitinate all its substrates in a RASSF1A-deficient cells than in control APC/C (anaphase-promoting complex/cyclo- single sweep across the spindle and cytosol. cells, and these cells exited mitosis with some), a large multisubunit complex that col- Securin ubiquitination occurs when the chro- severe segregation defects, sug- laborates with ubiquitin-conjugating and mosomes are at spindle equator, but cyclin B gesting that their spindle checkpoint was not ubiquitin-activating enzymes to catalyze the ubiquitination starts from the spindle poles. operational. How can RASSF1A affect formation of polyubiquitin chains on its pro- tein substrates, targeting them for degrada- tion by the proteosome. abcdS/G2/ Prometaphase Metaphase Anaphase A number of regulatory mechanisms coop- erate to restrict the activity of APC/C to a spe- cific period in the (Fig. 1). From S © 2004 Nature Publishing Group phase to early prophase, the inhibitor Emi1 Cdc20 APC/C (early mitotic arrest 1, also called regulator of

(Rca1)) predominates. APC/C Mad2 Cdc20 binds first with Cdc20 (also called fizzy) and Cdc20 P the complex APC/C-Cdc20 triggers protein Cdc20 APC/C RASSF1A P degradation from prometaphase until Emil . From telophase to late G1, APC/C- Cdh1 is the active complex. In addition to the P global activators Cdc20 and Cdh1 (Fzr), APC/C APC/C which are thought to act as substrate adap- APC/C Cdc20 P tors, local regulatory mechanisms are used by Cdc20 Cdc20 cells to promote or restrain the activity of APC/C to specific subcellular compartments. Cyclin A The power of spindle checkpoint One such mechanism, the spindle checkpoint, Figure 1 Inhibition of APC/C activity during the cell cycle. (a) From to prophase, Emi1 inhibits preserves genome integrity by keeping cells in the APC/C by binding to Cdc20. (b) During prometaphase, RASSF1A binds to Cdc20 and inhibits APC/C. metaphase until they have the correct chro- At the end of prometaphase, this inhibition ceases and cyclin A degradation initiates. (c) During late prometaphase, the spindle checkpoint becomes activated. At kinetochores, phosphorylation and Mad2 binding will prevent Cdc20 from binding APC/C, with the exception of a certain population that supports Endre Máthé is at Cancer Research UK Cell cyclin A degradation. (d) Cyclin B remains stable as long as the spindle checkpoint is active. APC/C will become multiphosphorylated. The chromosomes become aligned and the spindle checkpoint is Cycle Genetics Group, Department of Genetics, inactivated. As a consequence, nonphosphorylated Cdc20 binds to APC/C and anaphase inhibitors, University of Cambridge, Downing Street, including securin and cyclin B, are degraded. Blue and red solid lines indicate the stable stages for Cambridge, CB2 3EH, UK. cyclin A and cyclin B1, respectively; dotted lines denote their degradation. e-mail: em247@ mole.bio.cam.ac.uk

NATURE GENETICS VOLUME 36 | NUMBER 2 | FEBRUARY 2004 117 NEWS AND VIEWS

mitotic progression? One clue to explain for normal prometaphase progression, but of cyclin A, whose APC/C-Cdc20–mediated these abnormalities comes from Liu et al.6, its function is distinct and independent from destruction initiates after who found that the C-terminal region of that of Emi1. This cannot fully explain how breakdown in prometaphase. Thus it seems RASSF1A containing the Ras association RASSF1A overexpression overcomes the reasonable to speculate that the RASSF1A- domain bound to microtubules, conferring a effects of Emi1 depletion if their periods of associated Cdc20 at the is strong cytoprotective activity against the Cdc20 association are different. It is possible responsible for the inhibition of cyclin A microtubule-destabilizing drug nocodazol. that RASSF1A is also involved in Cdc20 reg- degradation during prometaphase. At present, it is not clear whether altered ulation during the period in which Emi1 microtubule stability is the primary cause of functions, but that its association with The burden of Cdc20 abnormal mitosis or the consequence of an Cdc20 is more transient than it is in The spindle checkpoint is coordinated by the unknown primary defect. prometaphase, and so not detected by imm- many partners of Cdc20, which binds kineto- munoprecipitation from prophase extracts. chores by its N-terminal region and centro- A bridge between Emi1 and spindle If this is the case, more experiments are somes, spindle microtubules and Mad2 with checkpoint? needed to investigate whether RASSF1A can its C-terminal region. It also uses the N-ter- Both Emi1 and the spindle checkpoint protein regulate Cdc20 during prophase and the G2 minal region to bind the TPR-subunits of Mad2 negatively regulate APC/C through phase of the cell cycle. APC/C and is degraded through its cyclin association with Cdc20, and the phenotypes How is the prometaphase function for B–type destruction boxes and APC/C-Cdh1. resulting from RASSF1A overexpression or RASSF1A related to the spindle checkpoint? It is exciting to imagine a RASSF1A-depen- loss resemble those of cells in which Emi1 and Lung cancer and HeLa cells, treated with dent mechanism working like a cell cycle Mad2 were similarly perturbed5. RASSF1A nocodazol to activate their spindle check- checkpoint. Some of the interactions required probably works in the same way, as RASSF1A point, stayed arrested in a prometaphase-like for this have now been demonstrated (for and Cdc20 associate through their N-terminal state with high levels of mitotic . The example, the binding of the N-terminal regions and polyubiquitination activity of loss of RASSF1A protein had no apparent region of Cdc20 to the N-terminal regions of http://www.nature.com/naturegenetics APC/C-Cdc20 is inhibited in vitro in the pres- effect on this arrest, and levels of cyclin B Emi1 and RASSF1A). The localization of ence of full-length RASSF1A protein. remained unaffected. Because the cells did RASSF1A to spindle poles, spindle micro- Successive overexpression of RASSF1A and not exit mitosis in the presence of an active tubules and the cytosol could mean that this Cdc20 in vivo showed that the relative level of spindle checkpoint, the authors concluded new molecular mechanism coordinates the these antagonists is key in regulating APC/C- that the RASSF1A function was not required activity of population(s) of Cdc20 and APC/C Cdc20 activity and the consequent destruction for the normal checkpoint functioning. But with that of the cytoskeleton or other cellular of both mitotic cyclins. A substantial amount when RASSF1A was overexpressed with the compartments. of RASSF1A colocalized with Cdc20 at the spindle checkpoint off, the cells did not Investigation of cells with both an activated poles and microtubules of prometaphase spin- exit mitosis and were still arrested in a spindle checkpoint and overexpressed dles, and coimmunoprecipitation experiments prometaphase-like state, suggesting that RASSF1A, or abrogated checkpoint and detected relatively low levels of association excess RASSF1A could provide arrest in the diminished RASSF1A, might uncover the ear- between RASSF1A and Cdc20 during mitosis. absence of a spindle checkpoint signal. liest steps of spindle checkpoint activation. Thus, the new RASSF1A-Cdc20 regulatory The authors postulate that the spindle Studies based on antibody injections in cul- © 2004 Nature Publishing Group mechanism operates during prometaphase. checkpoint and RASSF1A exert their negative tured cells can explore the mitotic functions Cells depleted of Emi1 do not accumulate control on APC/C-Cdc20 through indepen- of RASSF1A and Cdc20, advancing our mitotic cyclins or enter mitosis because dent mechanisms and at different time understanding of how various substrates APC/C activity is not properly inhibited. points. Indeed, the peak of Cdc20-RASSF1A located at different compartments could be When RASSF1A was overexpressed in HeLa and Cdc20–spindle checkpoint activities are degraded simultaneously. cells depleted of Emi1, the mitotic cyclins different, indicating that they might operate 1. Dammann, R. et al. Nat. Genet. 25, 315–319 (2000). accumulated and cells progressed into at different time points or in different cellular 2. Burbee, D.G. et al. J. Natl. Cancer Inst. 93, 691–699 mitosis, arresting around prometaphase. compartments. Notably, cyclin A and cyclin B (2001). 3. Ortiz-Vega, S. et al. Oncogene 21, 1381–1390 Comparing the prophase peak of Cdc20- are not degraded in prometaphase-arrested (2002). Emi1 association to the prometaphase peak cells generated by elevated levels of RASSF1A 4. Shivakumar, L. et al. Mol. Cell Biol. 22, 4309–4318 of Cdc20-RASSF1A association by immuno- protein. (2002). 5. Song, M.S. et al. Nat. Cell Biol. advance online publi- precipitation from normal HeLa cells, the The activated spindle checkpoint prevents cation, 25 January 2004 (doi: 10.1038/ncb1091). authors propose that RASSF1A is required only the degradation of cyclin B and not that 6. Liu, L. et al. Oncogene 22, 8125–8136 (2003).

118 VOLUME 36 | NUMBER 2 | FEBRUARY 2004 NATURE GENETICS