Report

Mass Spectrometric Analysis of the Anaphase-Promoting Complex from Yeast: Identification of a Subunit Related to Cullins

See allHide authors and affiliations

Science  20 Feb 1998:
Vol. 279, Issue 5354, pp. 1216-1219
DOI: 10.1126/science.279.5354.1216

Abstract

Entry into anaphase and exit from mitosis depend on a ubiquitin–protein ligase complex called the anaphase-promoting complex (APC) or cyclosome. At least 12 different subunits were detected in the purified particle from budding yeast, including the previously identified proteins Apc1p, Cdc16p, Cdc23p, Cdc26p, and Cdc27p. Five additional subunits purified in low nanogram amounts were identified by tandem mass spectrometric sequencing. Apc2p, Apc5p, and the RING-finger protein Apc11p are conserved from yeast to humans. Apc2p is similar to the cullin Cdc53p, which is a subunit of the ubiquitin–protein ligase complex SCFCdc4 required for the initiation of DNA replication.

The APC mediates cell cycle–regulated ubiquitination, and thereby degradation, of proteins containing sequences called destruction boxes (1-4). Entry into anaphase depends on the degradation of proteins such as Pds1p and Cut2p, which inhibit sister chromatid separation (5,6). Degradation of mitotic cyclins inactivates cyclin-dependent kinases (CDKs), which is important for exit from mitosis and is a prerequisite for DNA-replication in the subsequent cell cycle (7).

Five subunits of the yeast APC (Apc1p, Cdc16p, Cdc23p, Cdc26p, and Cdc27p) have been identified through genetic analysis (3,8). However, additional subunits were detected in APC particles purified from yeast and Xenopus oocytes (8, 9). Advances in the analysis of proteins by mass spectrometry and sequencing of the entire yeast genome provide a strategy to identify the components of multiprotein complexes that can be biochemically purified (10). We used this approach to identify five additional subunits of the APC.

To analyze the composition of the APC, we labeled cells expressing Cdc16p with Myc epitopes (Cdc16-Myc6p) with 35S, and the complex was immunoprecipitated with an antibody to Myc (11). Proteins of ∼90 (Apc2p) and 70 kD were detected in addition to Apc1p, Cdc16-Myc6p, Cdc27p, and Cdc23p (Fig.1A). Mass spectrometric analysis revealed that the 70-kD band consists of two proteins, p70 (Apc5p) and p68 (Apc4p). Proteins migrating at ∼40 (Apc9p), 32 (Apc10p), 23 (Apc11p), 20 [Cdc26p (8)], and 19 kD (Apc13p) were also detected (Fig. 1B). All of these proteins were detected in precipitates from strains expressing different epitope-tagged APC subunits but not from control strains, indicating that the yeast complex contains at least 12 different subunits.

Figure 1

Subunit composition of the APC. (A) Detection and purification of APC subunits. Proteins immunoprecipitated with an antibody to Myc from extracts from CDC16 (wild type or PDS1-myc18) and CDC16-myc6 cells were separated in SDS-polyacrylamide gels. Proteins from35S-labeled cells (5 × 107) were detected by fluorography (11) (left). For mass spectrometry, immunoprecipitates from 1010 cells were detected by silver staining (12, 13) (right). (◃) A protein coimmunoprecipitating with Pds1-Myc18p. Pds1p is stained only weakly. (∗) Proteins whose precipitation is not Myc-dependent. (B) Small APC subunits. Immunoprecipitates from 35S-labeled cells expressing Myc-tagged APC subunits were separated in a 4 to 20% gradient gel. Molecular sizes are indicated on the left (in kilodaltons).

To identify these proteins, we purified the APC fromCDC16-myc6 or CDC23-myc9 strains (12). One-step immunoprecipitations from unfractionated cell extracts yielded enough material to detect individual subunits on silver-stained gels (Fig. 1A). Proteins were identified by nanoelectrospray tandem mass spectrometric sequencing (13). The identification of two proteins in the 70-kD band is shown in Fig.2. Because of the small amount of protein, a normal nanoelectrospray mass spectrum failed to provide clear peptide candidates. However, peptide ions could be distinguished from chemical noise by parent ion scans (13) (Fig. 2A). Fragmentation of all peptide ions led to the identification of four peptides from two different yeast proteins (Fig. 2, B and C). One of the peptides was identified in a mixture with a peptide from trypsin (Fig. 2C). The mass spectrometric analysis identified Apc2p, Apc5p, Apc4p, Apc9p, and Apc11p as the gene products of the open reading frames [ORFs (14)] YLR127c [853 amino acids (aa)], YOR249c (685 aa), YDR118w (652 aa), YLR102c (265 aa), and YDL008w (165 aa), respectively [Fig. 2 and (15)].

Figure 2

Protein identification by nanoelectrospray tandem mass spectrometry. (A) Parent ion scan spectrum (13) of the tryptic digest of the 70-kD band. Tandem mass spectra were obtained from all labeled peaks that were identified as peptides from trypsin (∗), keratins (k), Apc5p (T1 +2, CVILLLK; T2 +2, ALEEDDFLK), and Apc4p (t1 +2, LAVIPIR; t2 +2, IIIYVEK) (29). (B) Identification of Apc4p from the tandem mass spectrum of the doubly charged ion t2 +2 [unfilled arrow in (A), mass-to-charge ratio (m/z) = 439.2]. The partial tandem mass spectrum above the parent ion was acquired separately (31 scans accumulated) and then combined with the full-scan spectrum (8 scans accumulated). This eliminated the chemical noise in the most informative part of the spectrum. Databases were searched with peptide sequence tags derived from the mass differences between adjacent COOH-terminal ions (Y" ions) (30). After retrieval of the matching sequence, detection of all predicted ions confirmed the identification. (C) Identification of Apc5p in the 70-kD band. The tandem mass spectrum of the peak atm/z = 539.8 [filled arrow in (A)] yields an ion series (filled arrowheads) from an Apc5p peptide (upper sequence) and another series (unfilled arrowheads) from a trypsin peptide (lower sequence; C, cysteine S-acetamide).

We confirmed the identity of these proteins as APC subunits by modifying the endogenous genes to encode epitope-tagged variants (16). Immunoprecipitates from APC2-myc9,APC4-myc9, APC5-myc9, and APC11-myc9cells all contained the same set of proteins, which included all known constituents of the yeast APC (Fig. 3). Coimmunoprecipitation of Apc9p with Cdc16-Myc6p or Apc5-Myc9p was dependent on the APC9 gene (Fig. 3A). Recently, theDOC1 gene, which encodes a 33-kD protein required for cyclin proteolysis, was identified through genetic analysis (17). Immunoprecipitations from 35S-labeled DOC1-myc9and CDC16-myc6 DOC1-HA3 cells revealed that DOC1encodes the APC subunit Apc10p (18).

Figure 3

Characterization of the identified proteins as APC subunits. Wild-type cells and cells expressing Myc-tagged proteins were labeled with 35S for 150 min at 30°C (A), for 180 min at 23°C (B), or for 90 min at 25°C and then for 60 min at 37°C (C) and processed for immunoprecipitations (11,16). Cse1p is not an APC component (8). (∗) Proteins whose isolation was not Myc-dependent. (A) APC subunits immunoprecipitated from APC5-myc9 and Δapc9 cells. (B) Identification of Apc11p as an APC subunit. (C) APC subunits immunoprecipitated from APC4-myc9 and Δcdc26 cells.

One copy of each ORF in diploid strains was replaced by aHIS marker (16). Tetrad analysis revealed thatAPC2, APC4, APC5, and APC11are all essential genes. His+ spores arrested as large, budded cells within one to three cell divisions after germination.

Haploid cells containing a deletion of APC9 were viable at 25° and 37°C. However, Cdc27p was largely absent in precipitates from Δapc9 strains (Fig. 3A). Apc9p might stabilize the interaction of Cdc27p with the APC. After release of small G1 cells at 37°C, spindle elongation and sister chromatid separation were delayed by 15 min in Δapc9 cells (18). Although not an essential gene, APC9 is required for efficient entry into anaphase. CDC26 encodes another nonessential APC subunit whose function is only required for growth at increased temperatures (8). In immunoprecipitates from Δcdc26 strains, the amounts of Cdc16p, Cdc27p, and Apc9p were reduced, whereas the other subunits were still associated with each other [Fig. 3C and (18)]. Cdc26p might be required for the incorporation of a set of subunits into the APC, especially at increased temperature.

Database searches (19) identified Apc11p as a member of a conserved family of proteins containing a RING finger (Fig.4A). The RING domain contains two zinc ions, is found in many eukaryotic proteins with diverse functions, and is thought to mediate protein-protein interactions (20). Incubation at 37°C of an APC11-myc9 strain led to the accumulation of cells with short and long spindles, indicating a defect in the onset of anaphase and exit from mitosis. (18). Extracts from G1-arrested APC11-myc9 cells shifted to 37°C were defective in the ubiquitination of mitotic cyclins (Fig. 4B).

Figure 4

Characterization of Apc11p. (A) Similarity of Apc11p to a family of RING-finger proteins. The 112 NH2-terminal residues of Apc11p (Sc) were aligned with ORFs encoded by ESTs from humans (Hs, gb AA541685), rat (Rn, gb H32307), Drosophila melanogaster (Dm, gb AA202488), and S. pombe (Sp, dbj AB001022; an intron was removed) (19). (∗) Zinc-binding residues conserved in RING-finger proteins (20). Residues identical in at least three sequences are shaded (29). (B) Cyclin ubiquitination in extracts from wild-type (WT) and APC11-myc9 cells. Strains (MATaΔpep4 Δbar1) were arrested in G1 with α-factor at 23°C and shifted to 37°C for 30 min. Extracts were incubated with adenosine 5′-triphosphate and HA3-tagged cyclins (4). Cyclin-ubiquitin conjugates were detected by immunoblotting with an antibody to HA. Clb2ΔDBp lacks the destruction box.

Database searches (19) revealed similarity of the COOH-terminal region of Apc2p to a putative ORF fromCaenorhabditis elegans (K06H7.5) and to a mouse protein whose COOH-terminal 426 amino acids could be assembled from expressed sequence tags (ESTs) [Fig. 5A and (18)]. The mouse sequence is 96% identical to that of human Apc2p (21). Apc2p contains a region of 180 aa with similarity to a family of proteins called cullins, which include yeast Cdc53p (22) (Fig. 5A). Cdc53p is a subunit of the SCFCdc4 ubiquitin–protein ligase complex, which mediates ubiquitination of the CDK inhibitor Sic1p by the ubiquitin-conjugating enzyme Cdc34p (23). Degradation of Sic1p is essential for entry into S phase (24). Cullins are therefore involved in both chromosome duplication and sister chromatid separation. Apc2p and Cdc53p may interact with components commonly required by various ubiquitin–protein ligases such as ubiquitin-conjugating enzymes. Indeed, Cdc53p binds Cdc34p, and this interaction is abolished by mutations in the region with similarity to Apc2p (25).

Figure 5

Analysis of Apc2p. (A) Alignment of Apc2p (aa 490 to 669) with Apc2p-related sequences fromC. elegans (CeK06H7.5) and mouse (MmApc2) and with cullins (19, 22). Residues identical in at least four sequences are shaded (29). (B) DNA content after release at 37°C of small G1 cells containingapc2-1 (28). Other apc2-1 strains gave similar results. (C) Percentage of cells with buds (□), short spindles (○), long spindles (•), separated sister chromatids (two GFP dots, ◊), and staining from Pds1-Myc18p (▾) after release at 37°C of small G1 cells containingapc2-1 tetO tetR-GFP or apc2-1 PDS1-myc18. Budding and spindle formation were similar in both strains. (D) Release at 37°C of small G1 cells containing apc2-1 Δpds1 tetO tetR-GFP. (E) Cyclin ubiquitination in extracts from wild-type (WT), Δapc2 APC2(APC2), Δapc2 apc2-1, and Δapc2 apc2-2 cells. Strains (MATa Δpep4Δbar1) were arrested in G1, and extracts were analyzed as in Fig. 4B.

To analyze the function of Apc2p, we mutagenized the gene in vitro, and two alleles conferring cell cycle arrest at 37°C were integrated into a haploid strain containing a deletion of the genomic APC2gene (26). We monitored sister chromatid separation in cells expressing a tet repressor– green fluorescent protein fusion (tetR-GFP), which binds to an array of tetoperator sites integrated near the centromere of chromosome V (27). We isolated small G1 cells from wild-type and apc2-1 mutant strains and followed their progression through the cell cycle at 37°C (28). In apc2-1cells, DNA replication and the formation of mitotic spindles occurred at the same time (relative to budding) as in wild-type cells. However, most of the mutant cells failed to separate sister chromatids and to elongate their spindles. Cytokinesis and re-replication were completely blocked [Fig. 5, B and C; the wild-type strain was analyzed in (27)]. Degradation of Pds1p, which starts shortly before anaphase, is required for sister chromatid separation (5,27). Detection of Pds1-Myc18p revealed that arrestedapc2-1 cells contained large amounts of Pds1p (Fig. 5C). Deletion of the PDS1 gene allowed apc2-1 cells to separate sister chromatids (Fig. 5D). However, spindle elongation was slower in apc2-1 Δpds1 cells than in wild-type cells. Thus, the inability of apc2-1 cells to enter anaphase may result primarily from a defect in the degradation of Pds1p.apc2-1 cells were also defective in degrading the mitotic cyclin Clb2p (18). Extracts prepared from G1-arrested apc2-1 and apc2-2 cells were defective in the ubiquitination of mitotic cyclins (Fig. 5E), indicating that the defect in proteolysis results from defective ubiquitination.

Yeast Apc5p shows similarity to human Apc5p (21) and to the putative ORF M163.4 from C. elegans. The yeast Apc4p sequence shows weak similarity to the human Apc4p sequence (21) and to the ORF Z97209 from Schizosaccharomyces pombe, which is more closely related to the human protein. Apc4p might represent an APC component that has diverged more during evolution than the other subunits. No homologs have been identified for Apc9p. Thus, in addition to Apc1p, Cdc16p, Cdc23p, Cdc27p, and Apc10p/Doc1p, at least Apc2p, Apc5p, and Apc11p might be conserved subunits of the APC in all eukaryotes.

REFERENCES AND NOTES

View Abstract

Navigate This Article