The Regulation of Cdc20 Proteolysis Reveals a Role for the APC Components Cdc23 and Cdc27 During S Phase and Early Mitosis Susanne Prinz, Ellen S

750 Research Paper

The regulation of Cdc20 proteolysis reveals a role for the APC components Cdc23 and Cdc27 during S phase and early mitosis Susanne Prinz, Ellen S. Hwang, Rosella Visintin and Angelika Amon

Background: In eukaryotic cells, a specialized proteolysis machinery that Address: Whitehead Institute for Biomedical targets proteins containing destruction-box sequences for degradation and that Research, 9 Cambridge Center, Cambridge, uses a ubiquitin ligase known as the anaphase-promoting complex/cyclosome Massachusetts 02142, USA. (APC) plays a key role in the regulation of mitosis. APC-dependent proteolysis Correspondence: Angelika Amon triggers the separation of sister chromatids at the metaphase–anaphase E-mail: [email protected] transition and the destruction of mitotic cyclins at the end of mitosis. Recently, two highly conserved WD40-repeat proteins, Cdc20 and Cdh1/Hct1, have Received: 16 February 1998 Revised: 20 April 1998 been identified as substrate-specific regulators of APC-dependent proteolysis Accepted: 7 May 1998 in the budding yeast Saccharomyces cerevisiae. Here, we have investigated the cell cycle regulation of Cdc20 and Cdh1/Hct1. Published: 4 June 1998

Current Biology 1998, 8:750–760 Results: Whereas the levels of CDH1/HCT1 RNA and Cdh1/Hct1 protein are http://biomednet.com/elecref/0960982200800750 constant throughout the cell cycle, CDC20 RNA and Cdc20 protein are present only during late S phase and mitosis and Cdc20 protein is unstable © Current Biology Ltd ISSN 0960-9822 throughout the entire cell cycle. The instability of Cdc20 depends on CDC23 and CDC27, which encode components of the APC. During the G1 phase, a destruction box within Cdc20 mediates its instability, but during S phase and mitosis, although Cdc20 destruction is still dependent on CDC23 and CDC27, it does not depend on the Cdc20 destruction box.

Conclusions: There are remarkable differences in the regulation of Cdc20 and Cdh1/Hct1. Furthermore, the APC activator Cdc20 is itself a substrate of the APC-dependent proteolysis machinery, and the APC subunits Cdc23 and Cdc27 have a role in the degradation of Cdc20 during S phase and early mitosis that is not mediated by its destruction box.

Background comprised of the ubiquitin-conjugating enzymes Ubc4 Progression through mitosis requires the precisely timed and UbcX [10–13] and a ubiquitin ligase known as the ubiquitin-dependent degradation of specific substrates. anaphase promoting complex/cyclosome (APC) [11,14]. Degradation of an as yet unidentified protein in egg The APC is composed of multiple subunits that include extracts from Xenopus [1], of Pds1 in the budding yeast the four previously identified proteins Cdc16, Cdc23, Saccharomyces cerevisiae [2] and Cut2 in Schizosaccharo- Cdc27 and Apc1/BimE [11,15–19] and the cullin homolog myces pombe [3] is required for the metaphase–anaphase Apc2 [20,21]. Following ubiquitination, substrates are transition. Proteolysis of mitotic cyclins (cyclins A and B in degraded by the 26S proteasome (reviewed in [22]). higher eukaryotes, Cdc13 in S. pombe, and Clb cyclins in S. cerevisiae; reviewed in [4]) is important for the inactiva- The stability of substrates of the APC-dependent proteol- tion of mitotic kinase during exit from mitosis. Destruc- ysis machinery varies greatly during the cell cycle. Sub- tion of Ase1 — a protein associated with the mitotic strates are stable during S phase, G2 and early mitosis spindle — promotes disassembly of the mitotic spindle (during which, for example, Clb2 has a half-life of more during exit from mitosis [5]. The efficient degradation of than 2 hours), but are unstable during exit from mitosis these proteins requires that they have a specific amino and G1 (Clb2 has a half-life of less than 1 minute during acid sequence termed the destruction box [6], which is G1) [2,5,23–25]. Work using extracts from Xenopus and usually located near the amino terminus of a protein. clam has shown that APC activity is a target for cell cycle Deletion of this 9 amino acid motif from substrates pre- regulation [11,14]. vents them from becoming multi-ubiquitinated and thus rapidly degraded (reviewed in [4,7–9]). Regulation of the APC-dependent proteolysis machinery is complex. In S. cerevisiae, inactivation of APC-dependent The machinery responsible for ubiquitinating mitotic proteolysis at the G1–S phase transition requires kinases cyclins and other proteins has been identified. It is activated by the G1 phase Cln cyclins [24]. Repression Research Paper The regulation of Cdc20 proteolysis Prinz et al. 751

during S phase and early mitosis requires the continuous gene (CDC20–HA) encoding Cdc20 tagged with a triple activity of Cln-associated and Clb-associated kinases [26]. hemagglutinin epitope and under the control of the wild- Activation of APC-dependent proteolysis during mitosis type CDC20 promoter — a gene which complements the also depends on cyclin-dependent kinase activity. Recon- growth defect of a temperature sensitive cdc20-1 mutant. stitution of cyclin B proteolysis in vitro using partially puri- Cells carrying CDC20–HA were synchronized in G1 by fied components suggests that cyclin B–Cdc2 kinase addition of the mating pheromone α-factor. Cells were activates APC-dependent proteolysis [14,27–29]. then released from the block and Cdc20 RNA and protein Recently, several regulators of the APC-dependent prote- levels were analyzed. CDC20 RNA levels were high during olysis machinery have been identified. Protein kinase A the α-factor-induced G1 arrest, declined during S phase has been shown to inhibit APC activity [17,30] whereas and accumulated again during G2 and mitosis the protein kinase Cdc5/Polo/Polo-like kinase activates (Figures 1a,2a). The increased RNA levels in the the APC [30–32]. Two highly conserved WD40-repeat pheromone-induced G1 arrest were not observed as cells proteins, Cdc20 and Cdh1/Hct1, have also been identified entered the subsequent G1 phase (Figure 1a,2a), suggest- as activators of APC-dependent proteolysis [31,33,34]. In ing that they were specific to the pheromone-induced G1 cdc20-1 mutants, degradation of Pds1 is defective but Clb2 arrest. Despite high levels of CDC20 RNA in pheromone- and Ase1 are degraded properly. Deletion of CDH1/HCT1 arrested cells, Cdc20 protein was absent during G1 results in stabilization of Clb2 and Ase1 but not of Pds1. (Figure 1a) suggesting that post-transcriptional mecha- Overexpression of CDC20 or CDH1/HCT1 is sufficient to nisms downregulate Cdc20 protein during G1. Through- induce the APC-dependent proteolysis of the appropriate out the rest of the cell cycle, Cdc20 protein levels had a target in stages of the cell cycle when substrates are nor- cell cycle profile similar to that of CDC20 RNA and that of mally stable. These findings define Cdc20 and Cdh1/Hct1 the mitotic cyclin protein Clb2 (Figure 1a). Cdc20 protein as substrate-specific rate limiting activators of APC- was absent during S phase, peaked when cells were in dependent proteolysis. mitosis and declined as cells entered G1 (compare Figure 1a and 1c). We conclude that Cdc20 RNA and As Cdc20 and Cdh1/Hct1 are critical regulators of APC- protein levels fluctuate during the cell cycle, being dependent proteolysis, it is important to understand how maximal during mitosis. these proteins are themselves regulated. Here, we report that the levels of Cdh1/Hct1 RNA and protein do not fluc- To determine the degree to which transcriptional control tuate during the cell cycle. In contrast, the levels of Cdc20 was responsible for the fluctuation in Cdc20 protein levels RNA and protein are cell cycle regulated. CDC20 RNA is during the cell cycle, we expressed CDC20 from the galac- detectable only during pheromone arrest, late S phase and tose-inducible GAL1-10 promoter. To allow moderate rates mitosis. The transcriptional controls operative on CDC20 of transcription, low levels of galactose (0.1%) were added. might be similar to those operative on the gene encoding Expression of CDC20 from the GAL1-10 promoter after the mitotic cyclin Clb2, as transcription of both CDC20 and release from a pheromone-induced G1 arrest gave levels of CLB2 requires active mitotic kinase. Furthermore, Cdc20 CDC20 RNA that were mostly constant throughout the cell protein is unstable throughout the cell cycle. This instabil- cycle (Figure 1b). Cdc20 protein levels were mostly con- ity is in part dependent on two APC subunits, Cdc23 and stant throughout the cell cycle under these conditions, Cdc27, suggesting that APC-dependent proteolysis is except during G1, when Cdc20 protein levels were signifi- partly responsible for the instability of Cdc20. Unlike other cantly lower than in other stages of the cell cycle APC substrates, proteolysis of Cdc20 depends on CDC23 (Figure 1b, time point 0). Our results suggest that transcrip- and CDC27 throughout the cell cycle. During G1, the insta- tional regulation of CDC20 plays an important role in bility of Cdc20 is mediated by a destruction box located in restricting Cdc20 protein to late S phase and mitosis. the amino-terminal part of the protein. In contrast, the During G1, additional post-transcriptional mechanisms help instability of Cdc20 during S phase and mitosis, though downregulate the Cdc20 protein, because neither ectopic dependent on CDC23 and CDC27, does not depend on the transcription from the GAL1-10 promoter nor α-factor- destruction box within Cdc20. Our results suggest a role for induced transcription led to accumulation of Cdc20 protein. the APC in protein degradation during S phase and early mitosis. Remarkably, for the degradation of Cdc20, this Similar transcriptional controls may be operative on activity does not depend on the destruction box in Cdc20 CDC20 and CLB2 necessary for its degradation during G1. The similar cell cycle oscillation of CDC20 and CLB2 RNA prompted us to investigate whether CDC20 and CLB2 tran- Results scription are under similar controls. Transcription of CLB2 Cdc20 protein levels are cell cycle regulated depends on an active mitotic kinase; when grown at the As a first step towards understanding how Cdc20 activity restrictive temperature, strains temperature-sensitive for the is regulated, we analyzed the levels of Cdc20 RNA and mitotic kinase (clb-ts strains) fail to accumulate CLB2 tran- protein during the cell cycle. To this end, we engineered a scripts [35]. Similarly, in the clb-ts strain, accumulation of 752 Current Biology, Vol 8 No 13

Figure 1

(a) (b) (c) (d) 1C 2C 1C 2C 0 15 30 45 60 75 90 105 120 135 150 165 180 195 Time (min) 0 15 30 45 60 75 90 105 120 135 150 165 180 195 Time (min) - Cdc20 Time Time - Cdc20 (min) (min) 195 195 - Clb2 - Clb2 180 180 165 165 Protein

Protein 150 150 - Cdc28 - Cdc28 135 135 120 120 105 105 Time (min) 90 90 0 15 30 45 60 75 90 105 120 135 150 165 180 195 Time (min)

0 15 30 45 60 75 90 105 120 135 150 165 180 195 75 75 60 60 - CDC20 45 45 - CDC20 30 30 15 15 RNA 0 0 - CDC28 RNA - CDC28 DNA content DNA content

Current Biology

Cdc20 protein levels fluctuate during the cell cycle. Cells carrying galactose. Samples were withdrawn at the indicated times to analyze either (a) CDC20–HA (A1182; MATa YCp-TRP1-CDC20–HA) or the levels of Cdc20 and Clb2 protein and CDC20 RNA. Cdc28 (b) GAL–CDC20–HA, a gene encoding Cdc20–HA under the control protein and CDC28 RNA were used as internal loading controls in of the GAL1-10 promoter, (A1052; MATa YCp-TRP1- immunoblot and northern blot analysis, respectively. Parallel samples GAL–CDC20–HA) were arrested in G1 with α-factor pheromone were taken from cells expressing (c) CDC20–HA and (2.5 µg/ml) in YEP medium containing raffinose. When arrest was (d) GAL–CDC20–HA to determine the cell cycle position by DNA complete, 0.1% galactose was added for 30 min followed by removal content measurements; the DNA content of cells in G1 (1C) and of α-factor and release into YEP–raffinose medium containing 0.1% G2/M (2C) is indicated.

CDC20 RNA after release from a pheromone-induced G1 at the level of transcription. During G1, however, neither arrest was severely delayed compared to the accumulation ectopic transcription nor transcription induced by α-factor in wild-type cells (Figure 2a,b). Control of CDC20 transcrip- led to accumulation of Cdc20 protein, suggesting that post- tion is not identical to that of CLB2, however, as it was transcriptional mechanisms must contribute to the lower induced by mating pheromone (Figures 1a,2a). We con- protein levels observed during this phase of the cell cycle. clude that transcription of CDC20, like that of CLB2, To identify these post-transcriptional controls, we deter- depends on an active mitotic kinase, but that additional con- mined whether protein turnover was responsible for the trols induce transcription in response to mating pheromone. downregulation of Cdc20 during G1. Cells were arrested at various stages of the cell cycle and the half-life of Cdc20 Cdc20 protein is unstable throughout the cell cycle was measured after transient expression from the GAL1-10 From our data, it seems that the fluctuation in the amount promoter. Expression from the GAL1-10 promoter was of Cdc20 protein during the cell cycle is determined largely brief (30 minutes) to avoid inappropriate activation of

Figure 2

(a) (c) 1C 2C (d)1C 2C (e)

0 15 30 45 60 75 90 105 120 135 150 Time (min) Time Time 100 -CDC20 (min) (min) 150 80 150 135 135 120 60 -CDC28 120 105 105 40 90

90 cells (%) (b) 75 20 Wild type 75 clb-ts 60 60 -CDC20 45 45 0 30 30 Proportion of budded 0 30 60 90 120 150 15 15 0 0 Time (min) -CDC28 Current Biology

CDC20 RNA accumulation is delayed in cells lacking mitotic kinases. were released into the cell cycle at 37°C. Samples were withdrawn at (a,c) Wild-type cells (K699) and (b,d) cells temperature-sensitive for the indicated times to determine (a,b) the total amount of CDC20 the mitotic kinase (clb-ts strain A544; MATa clb1∆ clb2-VI clb3::TRP1 RNA, with CDC28 used as a control, (c,d) DNA content clb4::HIS3) were arrested in G1 using α-factor pheromone measurements as in Figure 1, and (e) the proportion of budded cells. (2.5 µg/ml) at 23°C. When arrest was complete (after 165 min) cells Research Paper The regulation of Cdc20 proteolysis Prinz et al. 753

Figure 3 Figure 4

1C 2C (a) α 05 10 30 Time (min) Time -factor Hydroxyurea Nocodazole (a) (min) 015405 015405 015405 Time (min) - Cdc20 Cycling 30 - Cdc20

- Cdc28 0 (b) α-factor Hydroxyurea Nocodazole (b) α-factor - Cdc20 30 (Early G1 phase) - Cdc28 0 (c) 1C 2C 1C 2C 1C 2C cdc4-1 - Cdc20 30 Current Biology (Late G phase) - Cdc28 (d) 0 The endogenous Cdc20 protein is unstable during G1, S phase and early mitosis. Wild-type cells carrying CDC20–HA (A1182) were Hydroxyurea - Cdc20 30 arrested using 15 µg/ml nocodazole, 10 mg/ml hydroxyurea or 5 µg/ml (S phase) α-factor. The half-life of Cdc20–HA was measured as described in - Cdc28 Materials and methods, taking samples to determine (a) the level of 0 (e) Cdc20–HA protein at the indicated times after a pulse-chase with [35S]methionine and (b) the DNA content of arrested cells. cdc13-1 - Cdc20 30 (G2/M phase) - Cdc28 0 Cdc20 seemed to be even less stable (Figure 3b). The (f) increased instability of Cdc20 during G1 might contribute Nocodazole - Cdc20 30 to the low levels of protein observed even when CDC20 is (G2/M phase) - Cdc28 expressed from the GAL1-10 promoter. Interestingly, in 0 (g) cells arrested due to inactivation of the APC (Figure 3g), cdc23-1 - Cdc20 30 the half-life of Cdc20 was prolonged. (G2/M phase) - Cdc28 The endogenous Cdc20 protein was as unstable as Cdc20 0 (h) transiently expressed from the GAL1-10 promoter. - Cdc20 cdc5-1 30 Endogenous Cdc20 had a half-life of less than 5 minutes (Anaphase) in cells arrested using pheromone and those arrested in - Cdc28 S phase and early mitosis (Figure 4). Our results show that 0 (i) Cdc20 is a short lived protein throughout the cell cycle but cdc15-2 - Cdc20 30 that it is partially stabilized in cells with a defective APC. (Anaphase) - Cdc28 0 The instability of Cdc20 throughout the cell cycle depends Current Biology on an active APC The finding that Cdc20 was partially stabilized in an arrest Cdc20 is unstable throughout the cell cycle. Wild-type cells carrying due to inactivation of the APC raised the possibility that GAL–CDC20–HA were either (a) left to cycle, or arrested at 36°C in YEP–raffinose medium at the indicated cell cycle stages using either Cdc20 instability was at least partly mediated by the APC. (b) 5 µg/ml α-factor, (d) 10 mg/ml hydroxyurea or (f) 15 µg/ml Cdc20 might be degraded by the APC only during exit nocodazole. (c,e,g–i) Various temperature-sensitive cell division cycle from mitosis and during G1 — as are all other known APC (cdc) mutants carrying GAL–CDC20–HA were arrested at 36°C at substrates. The finding, however, that Cdc20, unlike the indicated cell cycle stages. After 3 h, galactose was added for 30 min to induce the production of Cdc20. Then glucose (2%) and other APC substrates, was unstable throughout the cell cycloheximide (1 mg/ml) were added to repress transcription and cycle raised the possibility that Cdc20 was degraded by translation, respectively. Samples were taken at the indicated times to the APC throughout the cell cycle. To distinguish determine the amount of Cdc20 protein and to determine cell cycle between these possibilities, we analyzed the conse- position by DNA content measurements. quences of inactivating APC on the turnover of Cdc20 in cells arrested at various stages of the cell cycle. We APC-dependent proteolysis. Half-life measurements arrested wild-type cells and cells defective for the APC revealed that Cdc20 was unstable throughout the cell (cells carrying a temperature sensitive cdc23-1 mutation) cycle. The protein had a half-life of less than 3 minutes in using either α-factor pheromone (Figure 5b), hydroxyurea exponentially growing cells (Figure 3a), during S phase (Figure 5c), or nocodazole (Figure 5d) at the permissive (Figure 3d), during G2 and early mitosis (Figure 3e,f), and temperature (23°C). When arrest was complete, cells were in cells arrested in anaphase (Figure 3h,i). During early G1, shifted to 36°C to inactivate the cdc23-1 gene product and 754 Current Biology, Vol 8 No 13

Figure 5

(a) Wild type cdc23-1 cdc27-A (c) Wild type cdc23-1 cdc27-A 010305 010305 010305 Time (min) 010305 010305 010305 Time (min) - Cdc20 - Cdc20

- Cdc28 - Cdc28

1C 2C 1C 2C 1C 2C 1C 2C 1C 2C 1C 2C Time Time Time Time Time Time (min) (min) (min) (min) (min) (min) 30 30 30 30 30 30

0 0 0 0 0 0 Wild typecdc23-1 cdc27-A Wild type cdc23-1 cdc27-A

(b) Wild type cdc23-1 cdc27-A (d) Wild type cdc23-1 cdc27-A 010305 010305 010305 Time (min) 010305 010305 010305 Time (min) - Cdc20 - Cdc20

- Cdc28 - Cdc28

1C 2C 1C 2C 1C 2C 1C 2C 1C 2C 1C 2C Time Time Time Time Time Time (min) (min) (min) (min) (min) (min) 30 30 30 30 30 30

0 0 0 0 0 0 Wild typecdc23-1 cdc27-A Wild type cdc23-1 cdc27-A

Current Biology

The instability of Cdc20 during the cell cycle depends on CDC23 and 30 min followed by a shift to 36°C for another 30 min. Transcription CDC27. Wild-type cells (A1052), cdc23-1 mutants (A1178) and and translation were repressed by the addition of glucose (2%) and cdc27-A mutants (A1391) carrying GAL–CDC20–HA were either cycloheximide (1 mg/ml), respectively (at time zero). Samples were (a) grown to exponential phase, or arrested with either (b) 5 µg/ml taken at the indicated times to determine the amount of Cdc20 protein α-factor, (c) 10 mg/ml hydroxyurea, or (d) 15 µg/ml nocodazole in and the cell cycle position by DNA content measurements. YEP–raffinose medium at 23°C. After 3 h, galactose was added for

the half-life of Cdc20 was measured after transient expres- proteolysis [33]. Thus, expression of CDC20 from the sion from the GAL1-10 promoter (see Materials and GAL1-10 promoter, though brief, might inappropriately methods). The cdc23-1 mutation caused a significant activate APC-dependent proteolysis in stages of the cell increase in the half-life of Cdc20 in all the cell cycle stages cycle when it is normally inactive. This would result in analyzed: in exponentially growing cells (Figure 5a), and Cdc20 appearing to be unstable throughout the cell cycle. cells arrested in G1 (Figure 5b), S phase (Figure 5c), and To test this possibility, we constructed a strain that mitosis (Figure 5d). Defects in the APC subunit Cdc27 expressed the genes for both the APC substrates Pds1 and [11,19] also caused stabilization of Cdc20. In temperature- Cdc20 under the control of the GAL1-10 promoter. After sensitive cdc27-A mutants, Cdc20 was stabilized in all the brief transient expression from the GAL1-10 promoter, we cell cycle stages analyzed (Figure 5a–d). The finding that measured the half-life of Pds1 and Cdc20. Whereas Cdc20 throughout the cell cycle proteolysis of Cdc20 depended was unstable during G1, S phase and mitosis, Pds1 stabil- on the APC subunits Cdc23 and Cdc27 indicates that ity was characteristic of that of other APC substrates — it APC-mediated proteolysis is responsible for the instability was unstable during G1, but stable during S phase and of Cdc20 throughout the cell cycle. early mitosis (Figure 6). This result shows that the instability of Cdc20 throughout the cell cycle that is depen- We have previously shown that prolonged overexpression dent on CDC23 and CDC27 was not due to inappropriate of CDC20 induces ectopic activation of APC-dependent activation of APC-dependent proteolysis brought about Research Paper The regulation of Cdc20 proteolysis Prinz et al. 755

Figure 6 Figure 7

Time1C 2C (a) Clb2 25 RLALNNVTN 010 30 60 Time (min)(min) Pds1 85 RLPLAAKDN - Cdc20 Cut2 33 RAPLGSTKQ Pds1 30 52 RTVLGGKST Cycling Cdc20 Box1 17 RSVLSIASP Box2 60 RPSLQASAN - Cdc28 0 Wild type ∆ Box1 ∆ Box2 ∆ Box1+2 - Cdc20 α -factor Pds1 30 0 5 10300 5 1030 0 5 10300 5 1030 Time (min) (Early G1phase) (b) - Cdc20 Cycling - Cdc28 0 - Cdc28

- Cdc20 (c) Hydroxyurea Pds1 30 - Cdc20 α-factor (S phase) - Cdc28 - Cdc28 0 (d) - Cdc20 - Cdc20 30 Hydroxyurea cdc13-1 Pds1 - Cdc28 (G2/M phase) - Cdc28 0 (e) - Cdc20 - Cdc20 Nocodazole - Cdc28 Pds1 30 Nocodazole Current Biology (G2/M phase) - Cdc28 0 The destruction box of Cdc20 contributes to the instability of Cdc20 - Cdc20 during G1. (a) Alignment of the destruction boxes of Clb2, Pds1 and cdc23-1 Pds1 30 Cut2 with two putative destruction boxes of Cdc20, Box1 and Box2; (G2/M phase) the residue number of the first amino acid in each sequence is - Cdc28 (b–e) 0 indicated and identical residues are shown in bold. Cells (A1242) carrying GAL–CDC20–HA or the same construct lacking ∆ Current Biology either the region encoding Box1 ( Box1), the region encoding Box2 (∆Box2), or both these regions (∆Box1+2) were either (b) grown to exponential phase, or arrested using either (c) α-factor, Cdc20, but not Pds1, is unstable throughout the cell cycle. Various cdc (d) hydroxyurea, or (e) nocodazole. The half-life of Cdc20 and Cdc20 mutants and wild-type cells carrying both GAL–CDC20–HA and a lacking one or both destruction boxes was determined as described in GAL–PDS1–HA fusion were arrested as described in the legend to the legend to Figure 3. Figure 3. The half-lives of Cdc20 and Pds1 were measured after transient expression from the GAL1-10 promoter. Samples were taken at the indicated times to determine the amount of Cdc20 and Pds1 protein the APC (Figure 5b). In contrast, deletion of Box1 had and to determine cell cycle position by DNA content measurements. no stabilizing effects during S phase or mitosis (Figure 7d,e). Deleting both Box1 and Box2 did not sta- bilize Cdc20 any more than did deleting Box1 alone, by transient ectopic expression of CDC20 from the confirming that Box2 did not contribute to the instability GAL1-10 promoter. Our results suggest that, in contrast to of Cdc20 (Figure 7b–e). other APC substrates, APC-dependent proteolysis of Cdc20 is active throughout the cell cycle. Our data suggest that the APC subunits Cdc23 and Cdc27 are required for degradation of Cdc20 throughout the cell A Cdc20 destruction box contributes to its instability cycle. APC-dependent proteolysis acts on Cdc20 in two during G1 but not during other stages of the cell cycle ways, however. APC-dependent proteolysis that is medi- Cdc20 contains two destruction-box-like motifs in its ated by the Cdc20 destruction box is required for degrad- amino terminus (Box1 and Box2; Figure 7a). To deter- ing Cdc20 during G1, whereas a previously unidentified mine whether the instability of Cdc20 that is dependent function of the APC subunits Cdc23 and Cdc27, which is on CDC23 and CDC27 is mediated by these destruction not mediated by the destruction box of Cdc20, is neces- boxes, we constructed Cdc20 proteins lacking either one sary for degrading Cdc20 during S phase and early mitosis. or both of the boxes. Deletion of Box2 did not cause a change in the half-life of Cdc20 in cells arrested in G1, S Cdh1/Hct1 RNA and protein levels are constant throughout phase or mitosis (Figure 7b–e). In contrast, deletion of the cell cycle Box1 stabilized Cdc20 but, surprisingly, not throughout The role of the Cdc20 homolog Cdh1/Hct1 in the degra- the cell cycle. Deletion of Box1 caused stabilization in dation of the mitotic cyclin Clb2 and the mitotic-spindle- G1 (Figure 7c) to the same extent as did inactivation of associated protein Ase1 is thought to be analogous to that 756 Current Biology, Vol 8 No 13

Figure 8

Cdh1/Hct1 RNA and protein levels are (a) (b) (d) 1C 2C constant during the cell cycle. (a) Extracts

0 15 30 45 60 75 90 105 120 135 150 165 Time (min) from wild-type cells (no tag, K699) and cells - Cdh1/Hct1 Time carrying CDH1–MYC (CDH1–MYC, A1342) - Cdc28 (min) were probed with an anti-Myc antibody to

Protein 180 165 visualize Cdh1–Myc. (b–d) Cells carrying 150 (c) 135 CDH1–MYC (A1342) were arrested in G1

0 15 30 45 60 75 90 105 120 135 150 165 Time (min) 120 using α-factor pheromone (2.5 µg/ml). After kDa 105 97- - No tag - Cdh1–Myc 90 2 h, cells were released from the block and - Cdh1 - CDH1/HCT1 75 66- 60 samples were withdrawn at the indicated

RNA 45 times to analyze the amount of (b) Cdh1/Hct1 - CDC28 30 0 protein and (c) CDH1/HCT1 RNA. (d) In Current Biology parallel, samples were taken to determine cell cycle position by DNA content measurements.

of Cdc20 in the degradation of Pds1 [33,34]. To determine Transcription of CDC20, like that of CLB2 and SWI5, whether the cell cycle control of Cdh1/Hct1 and Cdc20 depends on an active mitotic kinase. Thus, the mitotic are similar, we analyzed the levels of Cdh1/Hct1 RNA and kinases activate APC-dependent proteolysis by inducing protein during the cell cycle. Cells carrying a functional transcription of the critical activator CDC20. MYC-tagged CDH1/HCT1 gene (CDH1–MYC; see Materi- als and methods) were synchronized in G1 by addition of Transcription of CDC20 is also distinct from that of CLB2 the mating pheromone α-factor. After release from the and SWI5. CDC20 transcription is induced by mating block, Cdh1/Hct1 RNA and protein levels were analyzed. pheromone which may be important to ensure continuous As shown in Figure 8, the levels of both Cdh1/Hct1 RNA APC-dependent proteolysis activity during pheromone and protein were constant throughout the cell cycle. It is arrest [36]. We have previously shown that CDC20 is worth noting that the migration of Cdh1/Hct1 protein in required for Pds1 degradation during a pheromone- polyacrylamide gels containing sodium dodecyl sulfate induced G1 arrest [33]. During G1, however, Cdc20 is (SDS) varies during the cell cycle. Cdh1/Hct1 protein highly unstable and does not accumulate to levels that are extracted from G1-phase cells migrates faster than detectable by immunoblot analysis (Figure 1) [31]. The Cdh1/Hct1 obtained from mitotic cells (compare the 0 and finding that CDC20 transcription is highly induced during 75 minute time points in Figure 8b). We do not know the the pheromone arrest might help to explain this apparent reason for this differential migration, but phosphorylation discrepancy. Although Cdc20 is highly unstable during has been shown to alter migration of proteins in G1, the protein is nevertheless synthesized at high levels SDS–polyacrylamide gels. Our result suggests that during the arrest (Figure 4). Thus, Cdc20 might be able to although Cdc20 might perform a similar role in the degra- affect APC-dependent protein degradation. Before being dation of Pds1 to that of Cdh1/Hct1 in the degradation of degraded by the APC, Cdc20 might modify the APC. Clb2 and Ase1, there are remarkable differences in their Alternatively, Cdc20 might target substrates such as Pds1 regulation during the cell cycle. for degradation and by doing so becomes degraded itself by the APC. Discussion Transcriptional controls help to restrict Cdc20 activity to Cdc20 is degraded by a pathway dependent on CDC23 and late stages of the cell cycle CDC27 and by a pathway independent of CDC23 and CDC27 We have characterized the multiple controls operative on The half-life of Cdc20 is less than 3 minutes in exponen- Cdc20 that restrict its activity to late stages of the cell tially growing cells. In cdc23-1 or cdc27-A mutants, the half- cycle. In contrast to the levels of Cdh1/Hct1, which are life of Cdc20, although prolonged compared to wild-type constant throughout the cell cycle, Cdc20 levels fluctuate cells, is still short. One possible explanation for this during the cell cycle. The amount of Cdc20 in a cell is finding is that the defective APC components expressed regulated at the level of transcription, which appears to be by cdc23-1 and cdc27-A mutants are not completely inac- responsible mainly for restricting Cdc20 protein to late tive. We do not think that this is likely as other APC sub- stages of the cell cycle. When CDC20 is expressed at low strates are dramatically stabilized in cdc23-1 and cdc27-A levels from the GAL1-10 promoter, Cdc20 protein is mutants under similar experimental conditions (S.P. and present at constant levels throughout the cell cycle as A.A., unpublished observations) [15]. We therefore soon as cells enter S phase. The transcriptional controls believe that APC-dependent proteolysis is not the only operative on the CDC20 promoter might be similar to mechanism responsible for degrading Cdc20: there must those operative on the CLB2 and SWI5 promoters [35]. also be a mechanism that is independent of CDC23 and Research Paper The regulation of Cdc20 proteolysis Prinz et al. 757

CDC27, about which we know next to nothing. It is proteins are degraded by the S/early M CDC23/CDC27- unclear whether it is ubiquitin dependent and which part dependent proteolysis machinery also remains to be deter- of Cdc20 mediates this mode of proteolysis, but it appears mined. It is tempting to speculate that proteins important to be active throughout the cell cycle. The half-life of for restricting DNA replication to once per cell cycle Cdc20 in a cdc23-1 or cdc27-A mutant is similar whether [37,38] are degraded by this machinery. cells are arrested in S phase, mitosis or G1. Our data indicate that the S/early M, CDC23/CDC27- Two modes of protein degradation dependent on CDC23 dependent proteolysis machinery is active during S phase and CDC27 and early mitosis but might not be very active during G1. We have shown that the instability of Cdc20 during G1 is Deletion of Box1 of Cdc20 prolongs the half-life of Cdc20 dependent on the APC subunits Cdc23 and Cdc27. This during G1 almost to the extent seen in cdc23-1 mutants, finding, together with the finding that during G1 the suggesting that proteolysis mediated by this destruction instability of Cdc20 is dependent on its destruction box, box is the main mechanism responsible for the instability strongly suggests that APC-dependent proteolysis is of Cdc20 during the G1 phase. Our data also suggest that responsible for the instability of Cdc20 during G1. Similar the S/early M, CDC23/CDC27-dependent proteolysis findings were recently reported by Shirayama et al. [31]. machinery might not be as active as the APC-dependent Here, we have also shown that the instability of Cdc20 protein degradation machinery that requires destruction during S phase and early mitosis is independent of this boxes. Cdc20 appears less stable in G1-arrested cells than destruction box, although the APC components Cdc23 in cells arrested using hydroxyurea or nocodazole and Cdc27 are required for the degradation of Cdc20 (Figures 3,5,6). during these cell cycle stages. It is possible that Cdc23 and Cdc27 are not only APC components but that they are also Is there a role for the APC during S phase and early mitosis? components of an as yet unidentified complex that medi- We found that Cdc20 is unstable throughout the cell cycle ates protein degradation during S phase and early mitosis. and that this instability depends on two APC components, Given that the instability of Cdc20 during G1 shows all Cdc23 and Cdc27. During different stages of the cell the characteristics of APC-dependent proteolysis and that cycle, however, Cdc23 and Cdc27 function with different degradation during S phase and early mitosis requires the modes. Because of the complex pattern of protein degra- APC subunits Cdc23 and Cdc27, however, we favor the dation responsible for keeping Cdc20 unstable throughout idea that the requirement of CDC23 and CDC27 for Cdc20 the cell cycle and because the part of Cdc20 mediating the degradation during S phase and early mitosis reflects a S/early M, CDC23/CDC27-dependent proteolysis has not requirement for APC-dependent proteolysis during these been identified, we have not been able to directly analyze cell cycle stages. We do not know yet whether Cdc23 and the consequences of expressing Cdc20 that is stable Cdc27 directly ubiquitinate Cdc20 during S phase and throughout the cell cycle on cell cycle progression. Dele- early mitosis. Irrespective of whether the effects of inacti- tion of the destruction box stabilizes Cdc20 only during vating CDC23 or CDC27 on Cdc20 stability are direct or G1 which has no consequences on cell cycle progression indirect, however, our data demonstrate a previously when it is expressed under the control of the CDC20 pro- unknown function for Cdc23 and Cdc27, and possibly the moter (data not shown). Regulating the amount of Cdc20 APC, during S phase and early mitosis. protein might, however, be critical for cell cycle progression and cell viability. We have previously shown that We propose that APC has at least two functions. One is ectopic expression of CDC20 from the GAL1-10 promoter mediated by destruction boxes in its substrates, the other is lethal [33]. Cells cease to divide, but do not arrest at a one is not. The proteolysis that is dependent on destruc- specific stage of the cell cycle. Overexpression of Cdc20 is tion boxes, as previously reported for numerous APC sub- also sufficient to cause a bypass of the DNA damage and strates, is cell cycle regulated, being active during G1 but mitotic spindle assembly checkpoint arrest [39]. Thus, inactive during S phase and early mitosis. For Cdc20, keeping Cdc20 protein levels under tight control is not another proteolytic activity is mediated by Cdc23 and only critical for normal cell cycle progression but might Cdc27, but this is not dependent on either of Cdc20’s also be important during a cell cycle arrest induced by destruction boxes. Throughout the remainder of this dis- DNA damage or by mitotic spindle damage. cussion, we will term this activity the ‘S/early M, CDC23/CDC27-dependent proteolysis machinery’. We Cdc20 versus Cdh1/Hct1 regulation know little about the S/early M, CDC23/CDC27-depen- Although Cdc20 and Cdh1/Hct1 are highly related proteins dent proteolysis machinery. The part of Cdc20 that medi- and probably perform similar roles in the degradation of ates this mode of protein degradation has to be identified. Pds1, and Clb2 and Ase1, respectively, their roles and reg- Cryptic motifs similar to destruction boxes or new motifs ulation during the cell cycle differ remarkably. Cdc20 plays could mediate this activity. Furthermore, the degradation a critical role in the activation of Pds1 degradation [31,33]. machinery itself needs to be characterized. Whether other CDC20 is also required to initiate Clb degradation: cdc20-1 758 Current Biology, Vol 8 No 13

mutants arrest in metaphase with high levels of stable Clb2 regulated, being maximal during mitosis. The transcrip- [31] (S.P. and A.A., unpublished observations) which tional controls operative on CDC20 might be similar to reflects the requirement of Pds1 destruction for the initia- those regulating CLB2, as transcription of both CLB2 and tion of the proteolytic program responsible for degrading CDC20 depends on an active mitotic kinase. Further- Clb cyclins. Cdc20 also performs an essential function more, Cdc20 is unstable throughout the cell cycle and during exit from mitosis. Deletion of PDS1 in a cdc20-1 this instability is partly dependent on two APC subunits, mutant causes cdc20-1 cells to escape the metaphase arrest, Cdc23 and Cdc27, suggesting that it depends on the but cells go on to arrest in anaphase instead [31,40]. CDC20 APC-dependent proteolysis machinery. Surprisingly, two may also play a role in mitotic spindle dynamics. At the types of proteolytic mechanisms that are dependent on restrictive temperature, cdc20-1 mutants form abnormal CDC23 and CDC27 act on Cdc20. One mechanism, which mitotic spindles containing excessive numbers of micro- is dependent on the destruction box of Cdc20, is active tubules [41,42]. Thus, Cdc20 plays a central role in the ini- during exit from mitosis and G1, whereas the second tiation of the APC-dependent proteolytic program that mode is responsible for the instability of Cdc20 during triggers the metaphase–anaphase transition and exit from S phase and early mitosis and does not require the Cdc20 mitosis. In contrast, Cdh1/Hct1 is dispensable for the destruction box. Our results suggest that there is a previ- metaphase–anaphase transition but plays an important, ously unidentified activity for the APC-dependent prote- though not essential, role during exit from mitosis and G1. olysis machinery that is independent of destruction boxes Cdh1/Hct1 triggers degradation of mitotic cyclins and Ase1 and that acts during S phase and early mitosis. during these stages of the cell cycle [33,34]. Materials and methods The different roles of Cdc20 and Cdh1/Hct1 in the regula- Plasmid and strains tion of APC-mediated proteolysis and the differences in All strains were derivatives of strain W303 (K699; MATa ade2-1 leu2-3 ura3 trp1-1 his3-11,15 can1-100 GAL psi+). DNA manipula- their regulation might allow us to explain the substrate- tions were performed according to published methods [43]. To create specific differences in APC-mediated protein degradation. CDC20–HA, DNA encoding a triple HA tag was inserted in frame at It could explain why APC-mediated proteolysis, although the BstEII site 40 bp downstream of the initiation codon in the CDC20 activated at the metaphase–anaphase transition, com- gene or at the MluI site 510 bp downstream of it. Functionality of the pletely degrades Pds1 at this transition, but does not com- fusions was determined by their ability to complement the growth defect of a temperature-sensitive cdc20-1 mutant strain. The fusions plete degradation of mitotic cyclins until exit from mitosis were then cloned such that they were under the control of the (at the telophase–G1 transition). The Cln-dependent and GAL1-10 promoter as described [33] and placed into plasmid Clb-dependent kinases are continuously required to YCplac22 [44]. The destruction boxes within Cdc20 (Box1, amino prevent activation of APC-dependent proteolysis that is acids 17 to 25; Box2, amino acids 60 to 68) were deleted using the Chameleon double-stranded, site-directed mutagenesis kit (Strata- mediated by a destruction box during S phase and early gene) according to the manufacturers’ instructions. To create mitosis [24,26]. We speculate that Cdh1/Hct1, whose CDH1–MYC, a triple MYC tag was integrated at the initiation codon of protein levels are constant throughout the cell cycle, is the chromosomal CDH1/HCT1 locus as described [45]. inhibited by Cln-dependent and Clb-dependent kinases. In contrast, Cdc20, which is not present during S phase Growth conditions Conditions for growth and release of synchronous cultures from arrest and early mitosis might be less susceptible to the by α-factor (Figures 1,2,8) were as described [46], except cells were inhibitory activity of kinases dependent on Cln and Clb. released into YEP (yeast extract, peptone) medium containing raffinose Therefore, when synthesized during mitosis, Cdc20 initi- and 0.1% galactose in the experiment described in Figure 1. To deter- ates the proteolytic program responsible for progression mine the half-life of Cdc20 and versions of Cdc20 lacking either one or both destruction boxes in cells arrested at various stages of the cell through mitosis even in the presence of high levels of Cln- cycle (Figures 3,6,7), cells were grown in YEP–raffinose medium at dependent and Clb-dependent kinase activity. In contrast, 23°C. Cells were then shifted to 37°C for 3 h and 5 µg/ml α-factor, Cdh1/Hct1 cannot become active until mitotic kinases are 10 mg/ml hydroxyurea or 15 µg/ml nocodazole was added as appropri- at least partially inactivated. But how is mitotic kinase ate. Thereafter, galactose was added for 30 min. Then 2% glucose and 1 mg/ml cycloheximide were added to repress transcription and trans- inactivation initiated? Cdc20 might initiate inactivation by lation, respectively. The amount of Cdc20 was then analyzed by one or more of several mechanisms, including promoting immunoblot analysis. accumulation of the inhibitor of cyclin-dependent kinases Sic1 or inducing degradation of a fraction of mitotic To determine the half-life of the endogenous Cdc20 protein cyclins, which would then allow activation of Cdh1/Hct1 (Figure 4), cells were arrested with α-factor, hydroxyurea or nocodazole in medium containing 0.1 mM methionine. Cells from a 20 ml and accumulation of Sic1. culture were then harvested by filtration, washed with two volumes of medium lacking methionine and containing glucose (–metD medium) Conclusions and resuspended in 2 ml –metD medium. After a 4 min incubation, Our results suggest that RNA and protein levels of the cells were labeled with 0.5 mCi 35S-Translabel (ICN; 85% [35S]methionine, 15% [35S]cysteine) for 7 min. Cells were then cen- APC activator Cdh1/Hct1 are constant throughout the trifuged and resuspended at time zero in YEPD (yeast extract, cell cycle. In contrast, the APC activator Cdc20 is regu- polypeptone and dextrose) medium containing 2 mM methionine, lated at multiple levels. CDC20 RNA levels are cell cycle 2 mM cysteine and 1 mg/ml cycloheximide. Research Paper The regulation of Cdc20 proteolysis Prinz et al. 759

To determine whether the half-life of Cdc20 was dependent on a func- 16. Tugendreich S, Tomkiel J, Earnshaw W, Hieter P: CDC27Hs tional APC throughout the cell cycle (Figure 5), cells were arrested at colocalizes with CDC16Hs to the centrosome and mitotic spindle 23°C (the permissive temperature for the cdc23-1 and the cdc27-A and is essential for the metaphase to anaphase transition. Cell mutations) using α-factor (5 µg/ml), hydroxyurea (10 mg/ml), or noco- 1995, 81:261-268. 17. Yamashita YM, Nakaseko Y, Samejima I, Kumada K, Yamada H, dazole (15 µg/ml). After 3 h, galactose was added for 30 min, fol- Michaelson D, Yanagida M: 20S cyclosome complex formation and lowed by a shift to 36°C for another 30 min. Then 2% glucose and proteolytic activity inhibited by the cAMP/PKA pathway. Nature 1 mg/ml cycloheximide were added to repress transcription and trans- 1996, 384:276-279. lation, respectively. The amount of Cdc20 was then analyzed by 18. Peters J-M, King RW, Hoog C, Kirschner MW: Identification of BIME immunoblot analysis. as a subunit of the anaphase promoting complex. Science 1996, 274:1199-1201. 19. Zachariae W, Shin TH, Galova G, Obermaier B, Nasmyth K: Techniques Identification of subunits of the anaphase-promoting complex of Immunoblot analysis of the total amount of Clb2, Cdc28, Cdc20–HA Saccharomyces cerevisiae. Science 1996, 274:1201-1204. and Cdh1–Myc was determined as described [47]. The dilutions used 20. Yu H, Peters J-M, King RW, Page AM, Hieter P, Kirschner MW: were 1:2000 for anti-Clb2, anti-HA and anti-Cdc28 antibodies and Identification of a cullin homolog region in a subunit of the 1:500 for anti-Myc antibody. Equal loading of gel lanes was shown by anaphase-promoting complex. Science 1998, 279:1219-1222. probing blots with an anti-Cdc28 antibody. All other techniques were 21. Zachariae W, Shevchenko A, Andrews PD, Galova M, Stark MJR, performed as described ([35] and references therein). Immunoprecipi- Mann M, Nasmyth K: Mass spectometric analysis of the anaphase promoting complex from budding yeast: identification of a tation of 35S-labeled Cdc20 was performed as described [48] using an subunit related to cullins. Science 1998, 279:1216-1219. HA-affinity matrix distributed by BABCO. 22. Hochstrasser M: Ubiquitin, proteasomes, and the regulation of intracellular protein degradation. Curr Biol 1995, 7:215-223. 23. Evans T, Rosenthal ET, Youngblom J, Distel D, Hunt T: Cyclin: a Acknowledgements protein specified by maternal mRNA in sea urchin eggs that is We thank Steve Bell, Terry Orr-Weaver, Brian Cali and Jacqueline Lopez for destroyed at each cleavage division. Cell 1983, 33:389-396. helpful comments and discussions during the course of this project and for 24. Amon A, Irniger S, Nasmyth K: Closing the cell cycle circle in yeast: their critical reading of the manuscript. This research was supported by an G2 cyclin proteolysis initiated at mitosis persists until the activation Erwin Schroedinger Fellowship to S.P. and a National Institutes of Health of G1 cyclins in the next cell cycle. Cell 1994, 77:1037-1050. grant GM 56800-01 to A.A. 25. Brandeis M, Hunt T: The proteolysis of mitotic cyclins in mammalian cells persists from the end of mitosis until the onset of S phase. EMBO J 1996, 15:5280-5289. References 26. Amon A: Regulation of B-type cyclin proteolysis by Cdc28- Anaphase is initiated 1. Holloway SL, Glotzer M, King RW, Murray AW: associated kinases in budding yeast. EMBO J 1997, by proteolysis rather than by the inactivation of maturation- 16:2693-2702. promoting factor. Cell 73: 1993, 1393-1402. 27. Felix MA, Labbe J C, Doree M, Hunt T, Karsenti E: Triggering of 2. Cohen-Fix O, Peters J-M, Kirschner MW, Koshland D: Anaphase cyclin degradation in interphase extracts of amphibian eggs by initiation in Saccharomyces cerevisiae is controlled by the APC- cdc2 kinase. Nature 1990, 346:379-382. dependent degradation of the anaphase inhibitor Pds1p. Genes 28. Luca FC, Shibuya EK, Dohrmann CD, Ruderman JV: Both cyclin 10: Dev 1996, 3081-3093. AD60 and BD97 are stable and arrest cells in M-phase, but only 3. Funabiki H, Yamano H, Kumada K, Nagao K, Hunt T, Yanagida M: cyclin BD97 turns on cyclin destruction. EMBO J 1991, Cut2 proteolysis required for sister-chromatid separation in 10:4311-4320. fission yeast. Nature 1996, 381:438-441. 29. Lahav-Baratz S, Sudakin V, Ruderman JV, Hershko A: Reversible How proteolysis 4. King RW, Deshaies RJ, Peters J-M, Kirschner MW: phosphorylation controls the activity of cyclosome associated drives the cell cycle. Science 1996, 274:1652-1659. cyclin-ubiquitin ligase. Proc Natl Acad Sci USA 1995, 5. Juang Y-L, Huang J, Peters J-M, McLaughlin ME, Tai C-Y, Pellman D: 92:9303-9307. APC-mediated proteolysis of Ase1 and the regulation of the 30. Kotani S, Tugendreich S, Fujii M, Jorgensen P-M, Watanabe N, Hoog mitotic spindle. Science 1997, 275:1311-1314. C, et al.: PKA and MPF-activated Polo-like kinase regulate Cyclin is degraded by the 6. Glotzer M, Murray AW, Kirschner MW: anaphase-promoting complex activity and mitotic progression. ubiquitin pathway. Nature 1991, 349:132-138. Mol Cell 1998, 1:371-380. 7. Murray AW: Cyclin ubiquitination: the destructive end of mitosis. 31. Shirayama M, Zachariae W, Ciosk R, Nasmyth K: The Polo-like 81: Cell 1995, 149-152. kinase Cdc5p and the WD-repeat protein Cdc20/fizzy are The only way out of mitosis. 5: 8. Glotzer M: Curr Biol 1995, 970-972. regulators and substrates of the anaphase promoting complex in 9. Deshaies RJ: The self-destructive personality of a cell cycle Saccharomyces cerevisiae. EMBO J 1998, 17:1336-1349. transition. Curr Opin Cell Biol 1995, 7:781-789. 32. Descombes P, Nigg EA: The polo-like kinase Plx1 is required for M 10. Hershko A, Ganoth D, Sudakin V, Dahan A, Cohen LH, Luca FC, et phase exit and destruction of mitotic regulators in Xenopus egg al.: Components of a system that ligates cyclin to ubiquitin and extracts. EMBO J 1998, 17:1328-1335. their regulation by the protein kinase Cdc2. J Biol Chem 1994, 33. Visintin R, Prinz S, Amon A: CDC20 and CDH1: a family of 269: 4940-4946. substrate-specific activators of APC-dependent proteolysis. 11. King RW, Peters J-M, Tugendreich S, Rolfe M, Hieter P, Kirschner Science 1997, 278:460-463. MW: A 20S complex containing CDC27 and CDC16 catalyzes the 34. Schwab M, Lutum AS, Seufert W: Yeast Hct1 is a regulator of Clb2 mitosis-specific conjugation of ubiquitin to cyclin B. Cell 1995, cyclin proteolysis. Cell 1997, 90:683-693. 81:279-288. 35. Amon A, Tyers M, Futcher B, Nasmyth K: Mechanisms that help the 12. Aristarkhov A, Eytan E, Moghe A, Admon A, Hershko A, Ruderman JV: yeast cell cycle clock tick: G2 cyclins transcriptionally activate G2 E2-C, a cyclin-selective ubiquitin carrier protein required for the cyclins and repress G1 cyclins. Cell 1993, 74:993-1007. destruction of mitotic cyclins. Proc Natl Acad Sci USA 1996, 36. Irniger S, Nasmyth K: The anaphase-promoting complex is 93: 4294-4299. required in G1-arrested yeast cells to inhibit B-type cyclin 13. Yu H, King RW, Peters J-M, Kirschner MW: Identification of a novel accumulation and to prevent uncontrolled entry into S phase. J ubiquitin-conjugating enzyme involved in the mitotic cyclin Cell Sci 1997, 110:1523-1531. degradation. Curr Biol 1996, 6:455-466. 37. Heichman KA, Roberts JM: The yeast CDC16 and CDC27 genes 14. Sudakin V, Ganoth D, Dahan A, Heller H, Hershko J, Luca FC, et al.: restrict DNA replication to once per cell cycle. Cell 1996, 85:39-48. The cyclosome, a large complex containing cyclin selective 38. Heichman KA, Roberts JM: CDC16 controls initiation at ubiquitin ligase activity targets cyclins for destruction at the end chromosome replication origins. Mol Cell 1998, 1:457-463. of mitosis. Mol Biol Cell 1995, 6:185-198. 39. Hwang LH, Lau LF, Smith DL, Mistrot CA, Hardwick KG, Hwang ES, 15. Irniger S, Piatti S, Michaelis C, Nasmyth K: Genes involved in sister et al.: Budding yeast Cdc20: a target of the spindle checkpoint. chromatid separation are needed for B-type cyclin proteolysis. Science 1998, 279:1041-1044. Cell 1995, 87:269-277. 760 Current Biology, Vol 8 No 13

40. Yamamoto A, Guacci V, Koshland D: Pds1, an inhibitor of anaphase in budding yeast, plays a critical role in the APC and checkpoint pathway(s). J Cell Biol 1996, 133:99-110. 41. Sethi N, Monteagudo MC, Koshland D, Hogan E, Burke DJ: The CDC20 gene product of Saccharomyces cerevisiae, a beta transducin homologue, is required for a subset of microtubule- dependent cellular processes. Mol Cell Biol 1991, 11:5592-5602. 42. O′Toole ET, Mastronarde DN, Giddins TH, Winey M, Burke DJ, McIntosh JR: Three-dimensional analysis and ultrastructural design of mitotic spindles from the cdc20 mutant of Saccharomyces cerevisiae. Mol Biol Cell 1997, 8:1-11. 43. Maniatis T, Fritsch EF, Sambrook J: Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory Press; 1982. 44. Gietz RD, Sugino A: New yeast–E. coli shuttle vectors constructed with in vitro mutagenised yeast genes lacking six-base pair restriction sites. Genes 1988, 74:527-535. 45. Schneider BL, Seufert W, Steiner B, Yang QH, Futcher AB: Use of polymerase chain reaction epitope tagging for protein tagging in Saccharomyces cerevisiae. Yeast 1995, 11:1265-1274. 46. Surana U, Amon A, Dowzer C, Mcgrew J, Byers B, Nasmyth K: Destruction of the CDC28/CLB kinase is not required for metaphase/anaphase transition in yeast. EMBO J 1993, 12:1969-1978. 47. Schwob E, Böhm T, Mendenhall MD, Nasmyth K: The B-type cyclin kinase inhibitor p40SIC1 controls the G1 to S phase transition in S. cerevisiae. Cell 1994, 79:233-244. 48. Moll T, Tebb G, Surana U, Robitsch H, Nasmyth K: The role of phosphorylation and the Cdc28 protein kinase in cell cycle- regulated nuclear important of the S. cerevisiae transcription factor SWI5. Cell 1991, 66:743-758.

Because Current Biology operates a ‘Continuous Publication System’ for Research Papers, this paper has been published on the internet before being printed. The paper can be accessed from http://biomednet.com/cbiology/cub — for further information, see the explanation on the contents page.