Cdc20 is required for the post-anaphase, KEN-dependent degradation of centromere protein F

Abstract

Progression through mitosis and cytokinesis requires the sequential proteolysis of several cell-cycle regulators. This proteolysis is mediated by the ubiquitin-proteasome system, with the E3 ligase being the anaphase-promoting complex, also known as the cyclosome (APC/C). The APC/C is regulated by two activators, namely Cdc20 and Cdh1. The current view is that prior to anaphase, the APC/C is activated by Cdc20, but that following anaphase, APC/C switches to Cdh1-dependent activation. However, here we present an analysis of the kinetochore protein Cenp-F that is inconsistent with this notion. Although it has long been appreciated that Cenp-F is degraded sometime during or after mitosis, exactly when and how has not been clear. Here we show that degradation of Cenp-F initiates about six minutes after anaphase, and that this is dependent on a C-terminal KEN-box. Although these two observations are consistent with Cenp-F being a substrate of Cdh1-activated APC/C, Cenp-F is degraded normally in Cdh1-null cells. By contrast, RNAi-mediated repression of APC/C subunits or Cdc20 does inhibit Cenp-F degradation. These findings therefore suggest that the APC/C does not simply 'switch' upon anaphase onset; rather, our observations indicate that Cdc20 also contributes to post-anaphase activation of the APC/C. We also show that the post-anaphase, KEN-box-dependent degradation of Cenp-F requires it to be farnesylated, a post-translational modification usually linked to membrane association. Because so many of the behaviours of Cenp-F are farnesylation-dependent, we suggest that this modification plays a more global role in Cenp-F function.

Fig. 1.

GFP-tagged SecurinΔD and Aurora A are degraded after anaphase. (A) Timeline showing experimental strategy. (B) Line graphs showing the mean degradation over time of securin-GFP (WT, light grey circles, n=15 cells from four experiments), securinΔD-GFP (ΔD, dark grey circles, n=14 from four experiments) and Aurora-A—GFP (dark grey circles, n=7 from one experiment). Immunoblots show the expression of the GFP-fusion proteins. BubR1 was used as a loading control. Note that the lower mobility forms are more apparent in the securinΔD-expressing line. This is because securin is also degraded in interphase by a phosphorylation-mediated SCF-dependent mechanism (Gil-Bernabe, 2006). The ΔD mutant seems refractory to SCF-dependent degradation, resulting in the accumulation of the hyper-phosphorylated forms.

Fig. 2.

GFP-tagged Cenp-F is degraded after anaphase in a KEN-dependent manner. (A) Schematic of Cenp-F showing the domain encoding C630. (B) Immunoblots showing the tetracycline-induced expression of GFP-tagged C630 (WT), C630 ΔKEN6 and C630 ΔKEN7 mutants. Bub3 was used as a loading control. (C) Immunoblot showing the tetracycline-induced expression of GFP-tagged Cenp-F (WT) and Cenp-F ΔKEN7. BubR1 was used as a loading control. (D) Immunofluorescence of mitotic cells expressing GFP-tagged C630 and C630 ΔKEN7, stained as indicated. Enlargements of highlighted (boxed) areas show kinetochore localisation as judged by Bub1. Scale bars: 5 μm. (E) Immunofluorescence of mitotic cells expressing GFP-tagged Cenp-F and Cenp-F ΔKEN7, stained as indicated. Enlargements of highlighted (boxed) areas show kinetochore localisation as judged by Bub1. Scale bars: 5 μm. (F) Line graphs showing the mean degradation over time of GFP-C630 (WT, green circles, n=27 from five experiments), GFP-C630 ΔKEN6 (yellow circles, n=12 from four experiments), GFP-C630 ΔKEN7 (red circles, n=14 from four experiments), GFP—Cenp-F (WT, green circles, n=8 from two experiments) and GFP—Cenp-F ΔKEN7 (red circles, n=14 from four experiments).

Fig. 3.

Degradation of Cenp-F is farnesylation-dependent. (A,B) Immunoblots showing the tetracycline-induced expression of GFP-tagged C630 C:S and GFP-tagged Cenp-F C:S. Bub3 and BubR1 were used as loading controls. (C) Anti-Myc immunoprecipitates of GFP-C630 (WT), GFP-C630 treated with FTI, GFP-C630 C:S and GFP-C630 ΔKEN7, from cells treated with AGOH. The precipitates were immunoblotted with anti-Myc and anti-AGOH, as indicated. (D) Immunofluorescence of mitotic cells expressing GFP-C630 C:S and GFP—Cenp-F C:S, stained as indicated. Enlargements of highlighted (boxed) areas show kinetochore localisation as judged by Bub1. Scale bars: 5 μm. (E) Line graphs showing the mean degradation over time of GFP-C630 treated with FTI (yellow circles, n=16 from four experiments), GFP-C630 C:S (red circles, n=30 from three experiments), GFP—Cenp-F treated with FTI (yellow circles, n=9 from five experiments), and GFP—Cenp-F C:S (red circles, n=13 from four experiments). Note that WT C630 and WT Cenp-F from Fig. 2 are shown as green lines merely for comparison.

Fig. 4.

Turnover of the Cenp-E tail is farnesylation-dependent. (A) Schematic of Cenp-E showing the tail domain. (B) Immunoblot showing the tetracycline-induced expression of GFP-tagged Cenp-E tail and tail C:S in DLD1 cells. Coomassie blue staining is shown as a loading control. (C) Anti-Myc immunoprecipitates of GFP—Cenp-E tail, GFP—Cenp-E tail treated with FTI and GFP—Cenp-E tail C:S from cells treated with AGOH. The precipitates were then immunoblotted with anti-Myc and anti-AGOH, as indicated. Arrow indicates Myc-tail protein. Arrowhead indicates an unknown farnesylated endogenous protein. M, Myc IP; C, control IP. (D) DNA content of asynchronous DLD1 cultures showing either whole cell populations (red) or only the GFP-positive cells. (E) Immunofluorescence of asynchronous cells expressing GFP—Cenp-E tail, GFP—Cenp-E tail C:S and GFP—Cenp-E tail treated with FTI, then imaged to detect GFP and cyclin B1 (red). Scale bars: 10 μm.

Fig. 5.

Cdh1 RNAi partially suppresses Cenp-F degradation. (A) Immunoblot showing the repression of Cdh1 24 hours after transfection with siRNA duplexes. BubR1 was used as a loading control. (B) Line graphs showing the mean degradation over time of Aurora-A—GFP treated with control (light grey, n=8 from two experiments) and Cdh1 RNAi (dark grey, n=10 from two experiments). (C) Line graphs showing the mean degradation over time of GFP-C630 treated with control (light grey, n=6 from four experiments) and Cdh1 RNAi (dark grey, n=10 from four experiments), and GFP—Cenp-F treated with control (light grey, n=5 from four experiments) and Cdh1 RNAi (dark grey, n=12 from four experiments).

Fig. 6.

Cenp-F is degraded normally in Cdh1-null iMEFS. (A) Line graphs showing the mean degradation over time of GFP-C630 (WT) in wild-type (green circles, n=19 from three experiments) and CDH1^−/− iMEFs (green circles, n=13 from four experiments); GFP-C630 C:S in wild-type (yellow circles, n=13 from two experiments) and CDH1^−/− iMEFs (yellow circles, n=8 from one experiment); and GFP-C630 ΔKEN7 in wild-type (red circles, n=12 from two experiments) and CDH1^−/− iMEFs (red circles, n=10 from three experiments). (B) Line graphs showing the mean degradation over time of: GFP—Aurora-B in wild-type (grey circles, n=11 from two experiments) and CDH1^−/− iMEFs (black circles, n=14 from two experiments); GFP-Plk1 in wild-type (grey circles, n=9 from two experiments) and CDH1^−/− iMEFs (black circles, n=3 from two experiments).

Fig. 7.

Inhibition of APC^Cdc20 blocks Cenp-F degradation. Immunoblot showing degradation of securin, cyclin B1, Aurora A and Cenp-F following RNAi repression of (A) APC 2 and APC 8, or (B) Cdc20 and Cdh1. The blots are derived from one experiment, representative of three independent experiments revealing highly comparable results.