Role of Hcn1 and Its Phosphorylation in Fission Yeast Anaphase-promoting Complex/Cyclosome Function*

The anaphase-promoting complex/cyclosome (APC/C) is a conserved multisubunit ubiquitin ligase required for the degradation of key cell cycle regulators. The APC/C becomes active at the metaphase/anaphase transition and remains active during G1 phase. One mechanism linked to activation of the APC/C is phosphorylation. Although many sites of mitotic phosphorylation have been identified in core components of the APC/C, the consequence of any individual phosphorylation event has not been elucidated in vivo. In this study, we show that Hcn1 is an essential core component of the fission yeast APC/C and is critical for maintaining complex integrity. Moreover, Hcn1 is a phosphoprotein in vivo. Phosphorylation of Hcn1 occurs at a single Cdk1 site in vitro and in vivo. Mutation of this site to alanine, but not aspartic acid, compromises APC/C function and leads to a specific defect in the completion of cell division.

The anaphase-promoting complex/cyclosome (APC/C) 4 is a multiprotein ubiquitin-protein isopeptide ligase that was first identified based on its role in facilitating the ubiquitination of A-and B-type cyclins, thereby targeting them for proteasomemediated destruction during mitosis (1)(2)(3). Although mitotic phase cyclins were the first targets known, many other APC targets have been identified subsequently (4), including securin, the destruction of which is required for chromosome segregation (5)(6)(7). Substrates are recognized by the APC/C based on the presence of short consensus motifs within them, such as the D-box, the KEN motif, and the A-box (8).
The majority of APC/C subunits (at least 13 in yeast) are stably associated throughout the cell cycle in an inactive core complex. The addition of transiently expressed CDC20 protein family members and phosphorylation events contribute to APC/C activation during mitosis and G 1 phase. Although active APC/C preparations treated with phosphatase lose activity (9 -11), it remains unclear which phosphorylation event(s) on which subunit(s) are responsible for this observation. It is also possible that no individual phosphorylation event alone has much effect, but rather it is the combination of many phosphorylation events that serves to alter complex activity. In this regard, it is clear that both CDC20 and core APC/C components are phosphorylated during mitosis. Recently, roles of CDC20 phosphorylation in both positive and negative APC/C regulation have emerged (12)(13)(14)(15)(16), suggesting that CDC20 phosphorylation might be a common target of cell signaling pathways that affect mitotic progression by altering the timing of APC/C activity (17).
Several components of the core APC/C are phosphoproteins (18 -22), and in vitro, the purified mitotic protein kinases Cdk1 and Plk (Polo in Drosophila, Cdc5 in Saccharomyces cerevisiae, and Plo1 in Schizosaccharomyces pombe) phosphorylate multiple subunits (1). To begin tackling the contribution of Cdk1 phosphorylation to APC/C function, all consensus Cdk1 phosphorylation sites in Cdc27 (APC3), Cdc16 (APC6), and Cdc23 (APC8) were altered to alanines (21). There were some defects in mitotic exit and Cdc20 binding and increased sensitivity to the spindle checkpoint in cells expressing the APC/C complex missing these serine and threonine residues (21). However, the lack of significant defects in the context of previous data indicating an essential role of APC/C phosphorylation argues that critical phosphorylation events were not identified. Indeed, phosphorylation of other subunits in vitro and residual phosphorylation of the subunits targeted for mutagenesis in vivo were detected. A recent investigation of human APC/C phosphorylation by mass spectrometry was more comprehensive in terms of site identification and reported that there are Ͼ43 sites of phosphorylation within the human APC/C, 34 of which are specific to mitosis (23). This study also demonstrated that Cdk1 is able to phosphorylate many of the identified sites on more than three subunits; Plk is able to phosphorylate others, and still other sites are not generated by either of these protein kinases in vitro, suggesting the involvement of additional protein kinases in APC/C regulation. Despite the wealth of data from this study on APC/C phosphorylation sites, the picture is still incomplete, as three APC/C subunits were not identified in the mass spectrometric analysis (23). In terms of the role of these phosphorylation events in APC/C regulation, Cdk1 and Plk phosphorylation increases APC/C activity in vitro in a CDC20dependent manner (21,(23)(24)(25). However, the roles of particu-lar phosphorylation events in particular core subunits have not been determined. Interestingly, the timing of phosphorylation events in different subunits was found to vary during mitosis, raising the possibility that distinct phosphorylation events might have different functional consequences (23).
We purified the S. pombe APC/C previously using tandem affinity purification and identified a total of 13 core components, including Hcn1, by mass spectrometry (26). Hcn1 had been identified previously as a high copy suppressor of a cut9-665 mutant (27). It was assumed to be an APC/C component because of its homology to the S. cerevisiae APC/C component (Cdc26) and because Cdc26 can also suppress cut9 (27). However, it had not been shown previously to co-purify with bona fide APC/C components. Here, we have extended our analysis of the role of Hcn1 within the S. pombe APC/C and addressed the role of Hcn1 phosphorylation in cell cycle progression. We found that Hcn1, unlike S. cerevisiae Cdc26, is an essential core APC/C component. It binds Cut9 directly and is required to link different APC/C subcomplexes together. We have mapped the single site of phosphorylation within Hcn1 to Ser 48 , a consensus site for Cdk1. Mutational analysis indicated that Hcn1 phosphorylation is not an essential modification, but mutation of Ser 48 to alanine induced a novel S. pombe APC/C phenotype. hcn1-S48A cells displayed a pronounced delay in the final stages of cell division because of a reduction in the amount of the Ace2 transcription factor.

EXPERIMENTAL PROCEDURES
Strains, Media, and Molecular Biology Methods-The S. pombe strains used in this study (Table 1) were grown in yeast extract medium or Edinburgh minimal medium with the appropriate supplements (28). Expression of constructs under the control of the thiamine-repressible nmt promoter system was performed as described previously (29). Standard genetic and recombinant DNA methods were used except where noted. Gene fragments were obtained by PCR amplification from S. pombe genomic DNA. Yeast transformations were performed using either a lithium acetate method (30) or electroporation (31). Cells were labeled with [ 32 P]orthophosphate as detailed previously (32).
Cell Cycle Synchronization-To obtain synchronous cultures of cells, small cells in early G 2 phase were isolated from 4-liter cultures that had been grown in yeast extract medium at 25°C to mid-log phase by centrifugal elutriation using a Beckman JE-5.0 rotor. The isolated small cells were filtered immediately and resuspended in yeast extract medium at the indicated temperatures and allowed to progress through the cell cycle. Samples were collected every 10 or 15 min. Cell cycle synchrony was then monitored by counting the number of nuclei and the septation index.
In Vivo Tagging and Gene Deletion-Strains expressing epitope-tagged versions of wild-type and mutant Hcn1 were constructed using a PCR-based approach as described previously (33). hcn1 ϩ was tagged at its endogenous locus at its 3Ј-end with a variety of tag-Kan R cassettes, including enhanced green fluorescent protein and Myc 13 . Appropriate tagging was confirmed by PCR and immunoblot inspection as appropriate. All tagged strains were viable at temperatures ranging from 25 to 36°C. The hcn1 ϩ coding sequences were replaced with the ura4 ϩ gene using a one-step PCR-based approach as described previously (33). The appropriate deletion was confirmed by PCR amplification using primers inside the ura4 ϩ gene and primers outside the disruption cassette. Immunoprecipitation and Immunoblotting-Whole cell lysates were prepared in Nonidet P-40 buffer, followed by antihemagglutinin (HA), anti-green fluorescent protein (GFP), or anti-Myc immunoprecipitations as described previously (32,34). Denatured lysates were prepared as described previously (35). Protein samples were resolved on 10% SDS-polyacrylamide gels and subsequently transferred to 0.2-m nitrocellulose membrane (Bio-Rad) or Immobilon (Millipore). Immunoblotting was done with anti-HA (12CA5; 0.3 g/ml), anti-Myc (9E10; 1.1 g/ml), anti-PSTAIRE peptide (1:5000; Sigma), or anti-GFP (0.2 g/ml; Roche Applied Science) monoclonal antibody. The primary antibodies were detected with horseradish peroxidase-conjugated goat anti-IgG secondary antibodies (0.4 mg/ml; Jackson ImmunoResearch Laboratories, Inc.) at a dilution of 1:50,000, followed by ECL visualization. Alternatively, affinity-purified Alexa Fluor 680-or IRDye800-conjugated goat anti-IgG antibodies were used as secondary antibodies, followed by scanning and quantitation with Odyssey (LI-COR Biosciences).
Immunoprecipitation/Phosphatase Assay-Following immunoprecipitation from denatured cell lysates, Sepharose-bound proteins were washed two times with 1 ml of Nonidet P-40 buffer and four times with 1 ml of phosphatase buffer (25 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, and 0.1 mg/ml bovine serum albumin), divided in half, and pulsed down, and the supernatant was aspirated off. 10-l reactions composed of 1ϫ phosphatase buffer, 2 mM MnCl 2 , and 1 l of -protein phosphatase (New England Biolabs) or 1 l of H 2 O were then incubated at 30°C for 45 min with gentle mixing. The beads were washed three times with Nonidet P-40 buffer and resuspended in 25 l of 2ϫ SDS sample buffer.
Sucrose Gradient Analysis-For sucrose gradient analysis, cells were grown to mid-log phase in yeast extract medium at 32°C. Approximately 2 ϫ 10 8 cells were collected by centrifugation, and native lysates in Nonidet P-40 buffer were prepared as described above. Lysates were layered on 10 -30% sucrose gradients prepared in Nonidet P-40 buffer. Gradients were ultracentrifuged at 30,000 rpm for 21 h in a Beckman SW 50.1 rotor. Sedimentation markers (thyroglobulin for 19 S and aldolase for 11.3 S) were fractionated on gradients prepared and spun in parallel. Fractions were collected, run on 10% SDSpolyacrylamide gradient gels, and then immunoblotted as described above.
Purification and Analyses of Lid1-TAP Complexes-8-liter cultures of lid1-TAP and lid1-TAP hcn1::ura4 ϩ leu1-32:: nmt81hcn1cDNA were grown to log phase, and the tagged proteins were isolated as described (36). The TAP complex was analyzed by direct analysis of large protein complexes/tandem mass spectrometry as described previously (37). However, the TAP complexes prepared from lid1-TAP hcn1-S48A and cut9-TAP were analyzed differently in that they were run by threephase multidimensional protein identification technology (MudPIT) on an LTQ instrument (Thermo Electron Corp.). The three-phase MudPIT was set up as described (38) with the following modifications. Briefly, 12 elution steps were included in the three-phase MudPIT experiments. For each salt elution, 2 l of ammonium acetate was injected through an autosampler (FAMOS). The concentration of salt used for each step was 0 mM, 10 mM, 25 mM, 50 mM, 100 mM, 200 mM, 300 mM, 400 mM, 600 mM, 800 mM, 1 M, and 5 M, respectively.
Yeast Two-hybrid Analysis-The yeast two-hybrid system used in this study was described previously (39). cDNAs of APC subunits were cloned into the bait plasmid pGBT9 and/or the prey plasmid pGAD424 and sequenced to ensure the absence of PCR-induced mutations and that the correct reading frame had been retained. To test for protein interactions, both bait and prey plasmids were cotransformed into S. cerevisiae strain PJ69-4A. ␤-Galactosidase reporter enzyme activity in the twohybrid strains was measured using a Galacto-Star TM chemiluminescent reporter assay system (Tropix Inc.) according to the manufacturer's instructions, with the exception that cells were lysed by glass bead disruption. Each sample was measured in triplicate. Reporter assays were recorded on a Mediators PhL luminometer (Aureon Biosystems).
In Vitro Kinase Assay-All bacterially produced recombinant proteins were purified on amylose beads (for maltose-binding protein (MBP)). Approximately 100 ng of recombinant Cdk1 kinase complex, purified from baculovirus-infected insect cells as described (26), was used to phosphorylate ϳ1 g of bacterially produced MBP-Hcn1 in HB15 buffer (25 mM MOPS, pH 7. Site-directed Mutagenesis and Gene Replacement-Serines within Hcn1 were changed to alanines using a QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. To generate a gene replacement strain, the hcn1-S48A mutation in a genomic clone contained in pIRT2 was transformed into the heterozygous hcn1 ϩ / hcn1::ura4 ϩ diploid strain. Leu ϩ diploids were allowed to sporulate, and Leu ϩ Ura ϩ haploid progeny were isolated. These were grown overnight in minimal medium containing uracil and leucine. The following day, 3 ϫ 10 7 cells were plated on minimal medium containing uracil, leucine, and 1 mg/ml 5-fluroorotic acid. Ura Ϫ Leu Ϫ colonies were then selected. To confirm that the correct gene replacement was present within these cells, the relevant genomic DNA of hcn1 was amplified by whole cell PCR using oligonucleotides hcn1-For2 (5Ј-cctctcttgctttttcatccttg-3Ј) and hcn1-Rev (5Ј-cagcagtatggaggcaggtgtg-3Ј). The PCR product was sequenced directly using oligonucleotide hcn1-For1 (5Ј-catgttacgaagaaatcctacg-3Ј).

RESULTS
To confirm that Hcn1 is a constitutive APC component, the locus encoding hcn1 ϩ was modified to produce Hcn1-GFP or

Hcn1 Phosphorylation Affects APC/C Function
Hcn1-Myc 13 , and the Myc-tagged allele was combined with cut9-HA 3 . In an anti-HA immunoprecipitate from hcn1-Myc 13 cut9-HA 3 , both Hcn1-Myc 13 and Cut9-HA 3 were detected, and conversely, in an anti-Myc immunoprecipitate, both Hcn1-Myc 13 and Cut9-HA 3 were detected (Fig. 1A). We next tested whether the association between these proteins is constant during the cell cycle. Small G 2 phase cells of the hcn1-Myc 13 cut9-HA 3 strain were collected by centrifugal elutriation and released into fresh medium to obtain a synchronous cell population. Samples were taken every 15 min and used for anti-Myc immunopurification. The amount of Hcn1-Myc 13 from every time point was constant, and the amount of co-purified Cut9-HA 3 was constant through the entire cell cycle, although its phosphorylation state changed as evidenced by lower mobility forms (Fig. 1B) (20). Also, Hcn1-Myc 13 co-sedimented in a sucrose gradient with Cut9-HA 3 (Fig. 1C). These data indicate that Hcn1 is a core component of the APC/C.
To determine whether Hcn1 is essential for viability, hcn1 ϩ was deleted by replacing it with the ura4 ϩ gene in a diploid. Tetrad analysis of the correct heterozygous diploids gave rise to two viable Ura Ϫ colonies and two nonviable colonies (data not shown), indicating that hcn1 ϩ is essential. To examine the consequence of Hcn1 loss to APC/C function, a conditional null allele of hcn1 ϩ was constructed by placing it under the control of the thiamine-repressible nmt81 promoter (42) in the hcn1::ura4 ϩ mutant background. The resultant strain was normal when grown in the absence of thiamine, but depletion of hcn1 ϩ through the addition of thiamine led to cell death. These cells failed to segregate their DNA properly, displaying a "cut" phenotype ( Fig. 1D) typical of APC/C mutants (43). These data indicate that Hcn1 is essential for S. pombe viability most likely because of a lack of APC/C activity.
Because Hcn1 loss resulted in a phenotype similar to those of other mutants that affect assembly of the APC/C (20,26,44), the integrity of the APC/C complex in the absence of Hcn1 was examined by sucrose gradient sedimentation. When hcn1 ϩ was normally expressed, both Lid1-Myc 13 and Cut23-Myc 13 sedimented in a peak at ϳ20 S as expected ( Fig. 2A). In contrast, Lid1-Myc 13 and Cut23-Myc 13 shifted into smaller complexes when expression of hcn1 ϩ was repressed ( Fig. 2A). Because Cut23-Myc 13 sedimented in a smaller complex than the subcomplex containing Lid1-Myc 13 , it seemed possible that the loss of Hcn1 split the APC/C into at least two distinct parts. To determine which APC/C subunits remained in the Lid1 subcomplex, we purified Lid1-TAP from the conditional hcn1 ϩ shutoff strain. Compared with the wild-type complex on a silver-stained gel, significant numbers of proteins were diminished in abundance or were absent (Fig. 2B). Mass spectrometric analysis of this complex revealed that Cut4, Apc2, Lid1, Apc5, Apc11, and perhaps Apc15 were present at at least 25% the recovery in hcn1 ϩ cells, but the recovery of other components was reduced more significantly (Table 2). Although this technique is not quantitative, it does suggest that specific components remain associated with Lid1 in the absence of Hcn1, whereas others do not. We concluded that Hcn1 helps link the Lid1-containing subcomplex to other APC/C subunits.
To determine which component(s) interacted directly with Hcn1 to bridge the Lid1-TAP subcomplex to other subunits, we performed a series of directed two-hybrid analyses between Hcn1 and other APC/C components with only Cut4 and Apc15 not included in the analyses. Hcn1 interacted with only Cut9  OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 JOURNAL OF BIOLOGICAL CHEMISTRY 32287 ( Fig. 2C), offering an explanation for the isolation of hcn1 ϩ as a high copy suppressor of the cut9-665 mutation (20). In addition to testing Hcn1 two-hybrid interactions, we examined pairwise interactions among other APC/C subunits, again with the exception of Cut4 and Apc15. The following interactions were the only ones detected: Apc2 showed a strong interaction with Apc11; Nuc2 interacted with Apc10; Lid1 interacted with Apc5; and Cut23 interacted with Apc13 (Fig. 2C). These interactions are consistent with the mass spectrometric data of the Lid1-TAP complex obtained above, as well as with other data regarding subcomplexes of the APC/C observed in other organisms (45,46).

Hcn1 Phosphorylation Affects APC/C Function
Many APC/C components are phosphoproteins, and phosphorylation of the APC/C during mitosis positively regulates its activity (9,10,23,25). To examine Hcn1 phosphorylation during mitosis, we used two different methods of arresting cells in a mitotic state, one using Mph1 overexpression and another using the mts3-1 mutant. Mph1 is a spindle checkpoint kinase; its overexpression mimics activation of the checkpoint and imposes a metaphase arrest (47). mts3 ϩ encodes one subunit of the 26 S proteasome, and its mutant allele (mts3-1) is defective in the metaphase/anaphase transition at non-permissive temperature (48). Hcn1-Myc 13 from metaphase-arrested cells, but not from asynchronous cells, showed a reduced mobility on gels, which was eliminated by phosphatase treatment (Fig. 3A), indicating that the shift was due to phosphorylation. The shift corresponding to altered phosphorylation was also detected in a synchronous population of hcn1-GFP cells as they progressed through mitosis by the appearance of a single more slowly migrating band ( Fig. 3B; see Fig. 4A). The hyperphosphorylated state of Hcn1-GFP peaked prior to septation (Fig. 3B). Phosphorylation of Hcn1-Myc 13 and Hcn1-GFP was also detected by in vivo labeling with [ 32 P]orthophosphate in cells arrested with the mts3-1 mutation (Fig. 3C). Phosphoamino acid analysis of in vivo labeled Hcn1-Myc 13 (Fig. 3D) and Hcn1-GFP (data not shown) revealed that Hcn1 was phosphorylated exclusively at serine residues. Hcn1 has eight serines, and one (Ser 48 ) conforms to the strict consensus site for Cdk phosphorylation (SPXK). Thus, we tested whether recombinant Cdc2-Cdc13 kinase (the S. pombe Cdk1 complex) could phosphorylate  13 . The peaks of thyroglobulin (19 S) and aldolase (11.3 S) collected from parallel gradients are indicated. B, the Lid1-TAP complex was purified from the wild-type strain (KGY3590) or the hcn1 shutoff strain (KGY4578) grown to log phase in the presence of thiamine. Proteins co-purifying with Lid1-TAP were revealed by silver staining. C, strain PJ206 was cotransformed with the bait plasmid pGBT9 and the prey plasmid pGAD424 in which APC/C cDNAs were cloned as indicated or not and then screened for their ability to grow on His-and Ade-deficient (double selection) plates. Streaks are shown in the right panels, with vector control transformants on the left side of plate and test vector transformants on the right. Based on the results from ␤-galactosidase assay, each interaction is shown as a bar graph. Relative light units of ␤-galactosidase activity are designated in parentheses.

TABLE 2 Effect of Hcn1 on Lid1-TAP composition
The values indicate the number of unique tryptic peptides identified from each protein by mass spectrometry. a These data were acquired on the more sensitive LTQ mass spectrometer.

Hcn1 Phosphorylation Affects APC/C Function
MBP-Hcn1 or two mutants of MBP-Hcn1 in which Ser 48 or Ser 69 was altered to alanine by site-directed mutagenesis. Although MBP-Hcn1 and MBP-Hcn1-S69A were phosphorylated by Cdc2-Cdc13, MBP-Hcn1-S48A was not (Fig. 3E), suggesting that Ser 48 is the sole site phosphorylated by Cdc2-Cdc13 in vitro. Phosphopeptide maps of MBP-Hcn1 labeled in vitro by Cdc2-Cdc13 gave rise to two spots that are likely related and that arose from variable tryptic digestion within the SPEKKK sequence (Fig. 3F, right panel). Two similar major spots were also produced from in vivo labeled Hcn1-GFP (Fig. 3F, left  panel) and Hcn1-Myc 13 (data not shown). These two major spots comigrated when equal counts of the two samples were mixed (Fig. 3F, MIX), suggesting that Hcn1 is phosphorylated in vivo at Ser 48 during mitosis.
Interestingly, there was a minor constellation of spots visualized in Hcn1-Myc 13 samples that was absent in the Hcn1-GFP sample, and two pieces of evidence suggest that these arose from phosphorylation within the Myc 13 epitope itself. First, they comigrated with phosphopeptides generated from the otherwise unrelated Sop2-Myc 13 protein; and second, they persisted in labeled Hcn1-Myc 13 lacking all of its eight serine residues (data not shown).
To verify that Ser 48 is the single site of Hcn1 phosphorylation, we generated a strain in which Ser 48 was replaced with alanine (hcn1-S48A) at the hcn1 ϩ genomic locus (see "Experimental Procedures"). The locus was then tagged with GFP or Myc 13 to allow detection of the mutant protein. Both tagged and untagged hcn1-S48A strains were viable, indicating that phosphorylation of Hcn1 at Ser 48 is not essential. Like Hcn1-Myc 13 , Hcn1-GFP migrated as multiple bands during mitosis, and the upper band could be eliminated by phosphatase treatment (Fig. 4, A and B). Mutation of Ser 48 abolished the upper mobility form of Hcn1-GFP (Fig. 4, A and B), indicating that the gel shift was due to this phosphorylation event. Labeling of Hcn1-S48A-GFP in vivo with [ 32 P]orthophosphate revealed that no phosphate was incorporated into the protein relative to Hcn1-GFP (Fig. 4C). Taken together, our data suggest that Ser 48 is the only significant in vivo phosphorylation site within Hcn1.
The hcn1-S48A mutant strain described above was used to determine whether Hcn1 phosphorylation affects APC/C activity and/or cell cycle progression. Defects in S. pombe APC/C function have manifested themselves previously with the development of a cut phenotype in which chromosome segregation and spindle elongation fail to occur such that subsequent cytokinesis bisects the nucleus or results in segregation of DNA to only one daughter cell (43). This phenotype was not observed in hcn1-S48A cells (Fig. 5A). Instead, many cells in the culture (Ͼ30%) showed cell separation defects resulting in doublets or chains of cells, particularly at 36°C (Fig. 5A). These doublets, arising from incomplete cell division, were separated partially  Hcn1 is a phosphoprotein. A, Hcn1-Myc 13 was immunoprecipitated from wild-type cells (KGY3399), wild-type cells overexpressing Mph1 (O.P. Mph1), or mts3-1 cells (KGY2306) that had been shifted to 36°C for 4 h and treated with either -protein phosphatase (PP) or buffer as a control. Samples were resolved by SDS-PAGE and probed with anti-Myc monoclonal antibody 9E10. B, hcn1-GFP cdc25-22 cells (KGY6080) were synchronized by shifting to 36°C for 4 h and then released to 25°C, and samples were collected every 15 min. Cell cycle synchrony was measured by septation index (provided above the immunoblot), and protein lysates were immunoblotted with anti-GFP antibody. The upper and lower bands correspond to the phosphorylated and unphosphorylated forms of Hcn1-GFP, respectively. C, mts3-1 (KGY2830), hcn1-Myc 13 mts3-1, and hcn1-GFP mts3-1 (KGY5632) strains were labeled with [ 32 P]orthophosphate, and lysates were immunoprecipitated with anti-Myc antibody 9E10 or anti-GFP antibody. The immunoprecipitates were resolved by SDS-PAGE and visualized by autoradiography. D, 32 P-labeled Hcn1-Myc from C was subjected to partial acid hydrolysis, and the resultant phosphoamino acids were separated by two-dimensional thin-layer electrophoresis and detected by autoradiography. E, MBP or MBP-Hcn1 protein (wild-type (WT), S69A, and S48A) was labeled with [␥-32 P]ATP by Cdc2-Cdc13, resolved by SDS-PAGE, and detected by Coomassie Blue staining (upper panel) and autoradiography (lower panel). F, the in vitro 32 P-labeled MBP-Hcn1 protein obtained in E and in vivo 32 P-labeled Hcn1-GFP were digested with trypsin while bound to polyvinylidene difluoride membrane. Equal cpm of tryptic peptides were spotted separately or mixed together (MIX) and resolved in two dimensions on cellulose thin-layer plates by electrophoresis at pH 1.9 with the anode on the left and by ascending chromatography. OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 by sonication and completely by treatment with cell wall-lysing enzymes (data not shown), indicating that independent daughter cells remained connected by undissolved cell wall material. The mutation was recessive because the phenotype of an hcn1 ϩ /hcn1-S48A diploid was wild type (data not shown). We also generated a strain in which Ser 48 was replaced with aspartic acid (hcn1-S48D) in an effort to mimic constitutive phosphorylation. This strain was morphologically wild type (data not shown) and had a normal septation index (13%) at all temperatures. We concluded that the cell separation defect is probably due to a lack of Hcn1 phosphorylation.

Hcn1 Phosphorylation Affects APC/C Function
To determine whether the delayed cell separation in the hcn1-S48A strain was due to reduced APC/C function, we examined genetic interactions between hcn1-S48A and other APC/C mutations. Although no synthetic interactions were detected with cut4-533, cut9-665, apc14⌬, or apc15⌬ (data not shown), hcn1-S48A reduced the restrictive temperatures of both lid1-6 and nuc2-663 (Fig. 5B), suggesting that the lack of Hcn1 phosphorylation compromised APC/C function.
As mentioned above, because of a block in Cut2 (securin) degradation, other S. pombe APC/C mutants display a cut phenotype. Although this phenotype was not observed in the hcn1-S48A strain, the genetic interactions detailed above prompted us to examine hcn1-S48A cells more carefully for defects typi-cally observed in cells with compromised APC/C function. First, the timing of chromosome segregation as an indication of Cut2 degradation was examined. Wild-type and hcn1-S48A cells were synchronized in G 2 phase by centrifugal elutriation, and their progression through mitosis was monitored by examining the formation of binucleate cells and septa in samples taken every 15 min (Fig. 5, C and D). Chromosome segregation and septation began earlier in the hcn1-S48A culture than in the wild-type culture because hcn1-S48A cells were farther along in G 2 phase when isolated. This was due to continued attachment of daughter cells at the end of the previous cell cycle. Although they began earlier, hcn1-S48A cells remained in anaphase slightly longer relative to wild-type cells (Fig. 5, C and D), suggesting that Cut2 degradation might be modestly slower in cells lacking Hcn1 phosphorylation. Once septation began, however, there was a marked accumulation of hcn1-S48A cells that remained connected even through the subsequent round of chromosome segregation (Fig. 5E, right panel), suggesting again that this step in the cell cycle was far more affected than the events of mitosis.
We next examined whether Cdc13/cyclin B destruction occurs with altered dynamics in hcn1-S48A cells, resulting in a mitotic exit delay. Because of the difficulty in matching the cell cycle state of elutriated samples for the reasons described above, an nda3-km311 block-and-release synchronization protocol was used. Cdc13 oscillations showed similar kinetics in both wild-type and mutant cells (data not shown), indicating that altered Cdc13 levels were not responsible for the observed delay in cell separation. Thus, typically scored aspects of APC/C function during mitosis appeared to be normal or only modestly affected in hcn1-S48A cells.
Cell separation in S. pombe requires the function of the Ace2 transcription factor (49). Ace2 regulates several genes involved in cell separation, including two encoding the glucanases Eng1 and Agn1, responsible for degrading the primary septum (49 -51). Because cell separation was delayed in hcn1-S48A cells, the amount of Ace2 in these cells was quantified. Whereas Hcn1-GFP and Cdk1 were present at equivalent levels in wild-type and hcn1-S48A cells, the level of Ace2-Myc 13 in hcn1-S48A cells was half that in wild-type cells (Fig. 6A).
If reduced Ace2 function was responsible for the cell separation delay in hcn1-S48A cells, then extra Ace2 would be expected to rescue this defect. To test this prediction, pREP3Xace2 or a control plasmid was transformed into hcn1-S48A cells. Plasmid pREP3Xace2 expresses Ace2 levels that complement the ace2⌬ phenotype and that are not deleterious to cell growth (52). As expected, there was significant rescue of the cell separation phenotype in cells transformed with the ace2-containing vector (Fig. 6B), but no deleterious effect on cell growth (data not shown). Taken together, these results are consistent with Hcn1 phosphorylation specifically affecting a step(s) in the transcriptional program required for the completion of cell division.

DISCUSSION
The yeast APC/C is composed of at least 13 subunits, but the function of many subunits is unknown. We have focused here on understanding the role of the smallest subunit of the S. A, lysates from the indicated strains were subjected to immunoprecipitation with anti-GFP antibody and treated with -protein phosphatase or buffer alone. Immunoprecipitates were resolved by SDS-PAGE, blotted, and probed with antibody to the GFP epitope. B, the hcn1-GFP and hcn1-S48A-GFP strains were transformed with pREP1 or pREP1mph1, and transformants were grown in the absence of thiamine for 18 h. Protein lysates were prepared and subjected to immunoprecipitation with anti-GFP antibody. The immunoprecipitates were resolved by SDS-PAGE, and GFP-tagged proteins were detected by immunoblotting. C, the indicated proteins in a mts3-1 background (strains KGY1834 and KGY5632) were labeled with [ 32 P]orthophosphate during a 4-h shift to the non-permissive temperature, and protein lysates were immunoprecipitated with anti-GFP antibody. The immunoprecipitates were resolved by SDS-PAGE, and labeled proteins were detected by autoradiography.
pombe APC/C, Hcn1. Our results indicate that Hcn1 is essential for cell viability, APC/C integrity, and proper APC/C regulation. Moreover, our analysis of Hcn1 phosphorylation indicated a specific role for the phosphorylation of this subunit late in the cell cycle.
We found that, in the absence of hcn1, the APC/C is split into at least two smaller subcomplexes. The first contains Cut4 (APC1), Lid1 (APC4), Apc2, Apc5, and Apc11. The second consists of the tetratricopeptide repeat proteins and other components (this study and data not shown). Similar subcomplexes have been purified previously from human and S. cerevisiae cells (45,46,53). One subcomplex consisting of APC1, APC2, APC4, APC5, and APC11 was found to interact with a ubiquitin conjugating enzyme and to assemble polyubiquitin chains, but could not conjugate these chains to a substrate (45). A stable subcomplex of the tetratricopeptide repeat proteins Cdc16, Cdc27, and Apc9, as well as of Swm1/Apc13 and Cdc26, has been detected in S. cerevisiae (46,53). Furthermore, in cdc26⌬ cells, the amounts of Cdc16, Cdc27, and Apc9 were reduced in immunoprecipitates of Apc2 (53), implying that, like Hcn1, Cdc26 is required for incorporation of a set of subunits into the APC/C especially at increased temperature. These subunits are implicated in binding activator and substrate proteins (54,55). Putting our purification and mass spectrometric data together with our two-hybrid analyses, we have arrived at a model of APC/C organization that is very consistent with the detailed architecture proposed recently for the S. cerevisiae APC/C by Thornton et al. (46). However, we found one additional interaction that deviates from this proposal: an interaction between Apc10 and Nuc2. Human APC10 was also shown to bind APC3/ CDC27 (56). It is possible that Apc10 has more than a single interaction surface on the APC/C and is present in multiple copies. In support of this possibility, it has been reported that S. cerevisiae Apc10 is present in multiple copies per APC/C complex (57).
While examining the associations of Hcn1, we recognized that Hcn1 is post-translationally modified and confirmed that it is phosphorylated during mitosis at a single site, Ser 48 . Ser 48 conforms to a strict consensus Cdk1 site, and Hcn1 Ser 48 is phosphorylated in vitro by Cdk1. Thus, Hcn1 is most likely phosphorylated by Cdk1, although we cannot exclude the possibility that other kinases target this site. Given the interaction of Hcn1 with Cut9 and the overall organiza-FIGURE 5. hcn1-S48A has a cell separation defect. A, hcn1-S48A mutant cells (KGY1630) were fixed and stained with methyl blue and 4Ј,6-diamidino-2-phenylindole to visualize the cell wall and DNA, respectively, and the percentage of various phenotypes was quantitated. A typical doublet (*) and a multi-septated cell (rare; #) are indicated. WT, wild-type cells. B, the indicated strains (KGY1551, KGY1552, KGY2229, and KGY2231) were spotted at 10-fold dilutions and grown at the indicated temperatures for 5 days on yeast extract plates. C and D, wild-type (KGY246) and hcn1-S48A (KGY1630) cells, respectively, were synchronized by elutriation at 25°C and shifted to 36°C, and samples were collected every 15 min to monitor binucleate cell formation (OE) and septation index (f). E, left panel, the percentage of cells from C and D that displayed the doublet and chain phenotype was quantitated; right panel, shown is a 4Ј,6-diamidino-2-phenylindole-stained image of cells corresponding to the sample indicated by the arrow in the left panel. OCTOBER 27, 2006 • VOLUME 281 • NUMBER 43 JOURNAL OF BIOLOGICAL CHEMISTRY 32291 tional model of the APC/C that places Hcn1 proximal to the tetratricopeptide repeat proteins (46), it is perhaps not surprising that Hcn1 might be targeted by Cdk1. Cdk1 phosphorylation of the APC/C, especially of the tetratricopeptide repeat proteins, has been implicated in stabilizing its association with Cdc20 (21,24,25). Of note, Hcn1 homologs are unlikely to be similarly phosphorylated by Cdk1 because human, mouse, and budding yeast orthologs lack consensus Cdk1 sites. Indeed, the overall level of phosphorylation site conservation in APC/C subunits is quite low (21,23), which has complicated assessing their individual contributions.

Hcn1 Phosphorylation Affects APC/C Function
Somewhat unexpectedly, mutation of Hcn1 Ser 48 to alanine, but not aspartic acid, had a very noticeable effect on cell cycle progression. There was a pronounced delay in completing the final step in cell division, cell separation, whereas progression through anaphase was slowed by 10 min at most. This phenotype would not be expected if Cdc20 binding to the APC/C were reduced as a consequence of decreasing Hcn1 phosphorylation. Consistent with this, the association between Cdc20 and Hcn1 in mitotic cells was not affected by the S48A mutation (data not shown), suggesting that ubiquitination of major mitotic targets occurred normally. Indeed, cell separation defects have not been noted previously in mutations of the S. pombe APC/C (43) possibly because loss of global APC/C function principally affects chromosome segregation. Although we cannot completely exclude the possibility that Hcn1 functions outside the APC/C, the hcn1-S48A phenotype appears to be more consistent with the possibility that a specific substrate(s) involved in cell separation is not targeted efficiently. There is growing evidence that substrates are targeted in a temporally programmed manner (4). Furthermore, phospho-specific antibodies to individual phosphorylation sites on human APC/C components react with their targets at different times during mitosis (23), the first suggestion that APC/C phosphorylation might affect substrate specificity temporally.
Although new for S. pombe APC/C mutants, a cell separation phenotype associated with the APC/C has been reported previously in S. cerevisiae lacking the Swm1/Apc13 subunit (53,58). Similar to hcn1-S48A cells, the Ace2 transcription factor is functionally limiting in swm1⌬ cells (53,58). In the case of swm1⌬ cells, however, it is present at normal levels, but mislocalized to the cytoplasm (53,58). S. pombe Ace2 is regulated differently from its S. cerevisiae counterpart (52). ace2 ϩ is a direct target of the Sep1 transcription factor (59 -61), the activity of which appears to be primarily responsible for dictating Ace2 abundance and activity (52). What controls Sep1 activity and S. pombe Ace2 post-translationally is still unknown, and therefore, it is presently unclear how the APC/C might influence S. pombe Ace2 levels. Moreover, providing additional Ace2 through use of a heterologous promoter did not completely rescue the cell separation defect; and hence, other targets of the APC/C also might be involved in the process of cell separation in S. pombe.