The Transcription Factor Atf1 Binds and Activates the APC/C Ubiquitin Ligase in Fission Yeast*

Fission yeast Atf1 is a member of the ATF/CREB basic leucine zipper (bZIP) family of transcription factors with strong homology to mammalian ATF2. Atf1 regulates transcription in response to stress stimuli and also plays a role in controlling heterochromatin formation and recombination. However, its DNA binding independent role is poorly studied. Here, we report that Atf1 has a distinct role in regulating the anaphase-promoting complex/cyclosome (APC/C) ubiquitin ligase. We have identified atf1+ as a dose-dependent suppressor of apc5-1, a mutation causing mitotic arrest. Remarkably, the suppression is not dependent upon the bZIP domain and is therefore independent of the ability of Atf1 to bind DNA. Interestingly, Atf1 physically binds the APC/C in vivo. Furthermore, we show that addition of purified Atf1 proteins into a cell-free system stimulates ubiquitylation of cyclin B and securin by the APC/C. These results reveal a novel role for Atf1 in cell cycle control through protein-protein interaction.

Ubiquitylation is a post-translational modification that occurs through the action of an enzymatic cascade consisting of three enzymes E1, E2, and E3, and typically controls the proteolysis, localization, or activity of a protein (1). It is a tightly regulated, highly specific and temporally controlled process, and thus plays an important role in many processes such as the cell cycle, signal transduction, transcription, DNA repair, development, and regulation of the immune response.
The anaphase-promoting complex/cyclosome (APC/C) 2 is a large (1.5 MDa) multisubunit E3 ubiquitin ligase, which has essential functions in key events during the cell cycle, such as chromosome segregation and mitotic exit by degrading securin/Cut2/Pds1 and cyclin B/Cdc13/Clb2, respectively (2)(3)(4). The APC/C belongs to the RING finger family of E3 ubiquitin ligases, which include Ubr1, c-Cbl, and Skp1-Cullin 1-F-box (SCF) complex (5), but unlike other members of this family, it is exceptionally large and complex, consisting of at least 11 conserved subunits. The APC/C appears to consist of two separable subcomplexes which associate independently with the scaffold subunit, Apc1/Cut4 (6). The first subcomplex includes a RING finger subunit Apc11 and a cullin domain subunit Apc2. This subcomplex is capable of ubiquitylating proteins to some extent, but lacks substrate specificity (7)(8)(9). The specificity seems to rely on the second subcomplex, consisting of three subunits Apc3/Nuc2/Cdc27, Apc6/Cut9/Cdc16, and Apc8/Cut23/Cdc23, which contain a tetratricopeptide repeat (TPR) domain. A member of the Fizzy family of APC/C activators, Cdh1, has been shown to bind Apc3/Cdc27 via its C-terminal IR motifs and recruit a substrate to the APC/C (10 -12). However, the exact roles of the remaining subunits and other TPR subunits, such as Apc5 and Apc7, are unknown.
The activity of the APC/C is regulated in part by the association of the Fizzy family of WD40-containing activator proteins (2)(3)(4). There are two major activators during the cell cycle, Fizzy/Slp1/Cdc20 and Fizzy-related/Ste9/Cdh1, which are required for the APC/C activity in anaphase and in G1 phase, respectively. In addition, APC/C activity is controlled by the positive effect of mitotic phosphorylation of the APC/C core subunits and the negative effect of Fizzy-related/Ste9/Cdh1 phosphorylation. Hence, the sequential and exclusive binding of the APC/C with the specific activator is achieved and thus the correct substrates are destroyed at the right time. Furthermore, apart from the well established role of the C-terminal WD40 domain in substrate recognition, we have recently shown that the interaction between the C-box of the Fizzy family of activators and the APC/C stimulates the activity, thereby providing an additional role in triggering ubiquitylation (13). Interestingly, the Apc10/Doc1 subunit that promotes processivity of the ubiquitylation reaction, also contains a similar C-box-like (CL) motif on the ligand-binding interface. Previously, we have shown that the transcriptional co-activator, CBP/p300 directly binds Apc5 and Apc7 subunits and potentiates the APC/C ubiquitylating activity (14). Therefore, the specificity and processivity of the APC/C might be regulated via multiple protein interactions.
The Atf1 transcription factor is essential for the response to stress and in conjunction with the MAP kinase Sty1/Spc1, for directing many of the transcriptional changes that occur in response to a number of stressful stimuli (15)(16)(17)(18)(19). Atf1 binds DNA as a heterodimer, preferentially with another bZIP protein (Pcr1) (20,21) and plays a role in directing heterochromatin formation and modulating meiotic recombination (21)(22)(23). All of these functions rely on the ability of Atf1 to bind to DNA through its basic leucine zipper (bZIP) domain.
Here, we investigated a TPR-subunit Apc5 using fission yeast Schizosaccharomyces pombe genetics and identified new regu-lators of the APC/C. Our results indicate that the transcription factor Atf1 functions not only through its ability to bind DNA, but also has a role in activating the APC/C ubiquitin ligase.

EXPERIMENTAL PROCEDURES
Yeast Strains and General Methods-Methods for handling S. pombe were as described (24). Strains used in this study are listed in Table 1. A PCR-based gene targeting method (25) was used for constructing gene deletion or C-terminal-tagged or deletion strains under the native promoter. Rich YEϩ5S or minimal EMM medium was used. Thiamine (4 M) was added to the medium to repress the nmt1 promoter. For spot tests for the temperature sensitivity, 8 l of 5-fold serial dilutions were spotted out on the medium and incubated at the indicated temperatures for 3ϳ5 days. To determine mating efficiency, homothallic h 90 strains were grown in EMM to a density of 5 ϫ 10 6 cells/ml, then shifted to EMM without a nitrogen source (EMM-N) and incubated at 30°C for 24 h. The efficiency of conjugation was calculated as the following ratio: (2ϫ number of asci and cells with conjugation formed)/(total number of non-conjugated cells ϩ 2ϫ number of asci and cells with conjugation formed).
Construction of Temperature-sensitive (ts) Mutant Strains-Genomic DNA was prepared from a strain in which the 3HA tag linked with the G418-resistance marker kanMX6 was inserted into the C terminus of apc5 ϩ under its native promoter. This apc5 ϩ -3HA-kanMX cassette was amplified with mutagenic PCR due to an unbalanced ratio of dNTPs (dGTP:dAT-P:dTTP:dCTP, 10:1:1:1) followed by integration into a wild-type genome. G418-resistant colonies were selected at 26°C and ts mutant alleles then screened by replica plating to 36°C.
Screen to Identify Multicopy Suppressors of the apc5-1 ts Mutation-HYY653 (apc5-1 ts) mutant strain was transformed with an S. pombe cDNA library, pTN-RC5 (a gift from C. Shimoda), containing cDNA fragments constructed in the expression vector pREP42. Colonies that could grow at the restrictive temperature (36°C) were collected and inserts of the plasmids were examined by Southern blotting using apc5 ϩ DNA as a probe. Plasmid DNAs were recovered from the Ura ϩ ts ϩ colonies and those not bearing apc5 ϩ were used to re-transform the apc5-1 strain. Plasmid inserts were sequenced upon confirmation of suppression of the ts phenotype.
Plasmid Construction-The coding regions of apc5 ϩ , ste11 ϩ , and atf1 ϩ were amplified from an S. pombe cDNA library, and subcloned into the pREP41 or pREP42 vectors (26). For domain analysis of atf1 ϩ , the C-terminal deletion (bZIP⌬ or C343⌬) and the N-terminal deletion (N120⌬ or N224⌬) constructs of atf1 were amplified by PCR and cloned into the pREP41 vector.
To set up a cell-free ubiquitylation assay, the coding regions of ptr3/uba1 ϩ , ubc1 ϩ , ubc4 ϩ , and ubc11/ubcp4 ϩ were isolated from a cDNA library and subcloned into pMal or pET16b vectors. All constructs were confirmed by DNA sequencing (Cogenics).
Northern Blot Analysis-Total RNA was prepared from fission yeast cells essentially as described previously (27). RNAs were analyzed by electrophoresis on a 1.5% agarose gel containing 0.66 M formaldehyde. Probes for blotting were prepared from DNA fragments containing hsp9 ϩ , ctt1 ϩ , pyp2 ϩ , or cdc2 ϩ by random oligonucleotide priming with [ 32 P]dCTP using the Megaprime DNA Labeling system (GE Healthcare).
Protein Expression in Sf9 Cells and Purification-For the production of transcriptionally inactive S. pombe Atf1, recombinant baculoviruses encoding Atf1-bZIP⌬ proteins tagged at the N terminus with either His 6 or GST were constructed, using the BaculoGold system (BD Pharmingen). Sf9 cells were infected with the appropriate viruses, and the proteins were purified using Ni-NTA beads (Qiagen) or GSH Sepharose (GE Healthcare) as described by the supplier.

Identification of Multicopy Suppressors of an apc5
Mutant-To identify novel regulators of the APC/C, we investigated one of the least studied APC/C subunits, Apc5. We first created conditionally lethal ts mutants of apc5. We made two ts alleles of fission yeast apc5 (apc5-1 and apc5-96) by random mutagenesis followed by chromosomal integration (see "Experimental Procedures"). Both alleles showed typical mitotic defect phenotypes such as "cut" or "nuc" (30) at the restrictive temperature (36°C) as expected from the mitotic role of the APC/C (supplemental Fig. S1). The mutation sites are shown in Table 2. We reasoned that increased concentrations of APC/C regulators might be able to restore the compromised activity of the APC/C in the apc5-1 mutant, thereby allowing the apc5 ts strain to grow at the restrictive temperature (36°C). Thus, we screened for multicopy suppressors of the temperature sensitivity of the apc5-1 mutant. Accordingly, strain HYY653 (apc5-1 ts) was transformed with an S. pombe cDNA library and plasmids were recovered from viable colonies at 36°C. Sequence analysis identified suppressors and these are listed in Table 3, including apc4/cut20/lid1 ϩ (Apc4 subunit), rad24 ϩ (14-3-3 protein), a translation initiation factor eIF3b and a transcription factor ste11 ϩ (Fig. 1A).
Ste11 is a transcription factor involved in sexual differentiation (31) and is regulated, in part, by the Sty1/Spc1 stress-activated MAPK pathway (Fig. 1B). This finding prompted us to examine whether activation of the MAPK pathway by osmotic stress could rescue the apc5-1 mutant. In fact, 0.5 M KCl (Fig. 1, C and D) as well as 2 M sorbitol (data not shown) could clearly suppress the ts phenotype of apc5-1. As expected, this suppression was dependent upon sty1/spc1 ϩ and wis1 ϩ (data not shown). Intriguingly, the suppression was also dependent upon atf1 ϩ , but not ste11 ϩ , suggesting that Atf1 is more critical than Ste11 in the osmotic stress-dependent apc5-1 suppression (Fig.  1D). Therefore, we tested whether atf1 ϩ is a dosage-dependent suppressor of the apc5-1 mutant. As shown in Fig. 1E, overproduction of atf1 ϩ rescued the temperature sensitivity of apc5-1 regardless of the presence of ste11 ϩ ( Table 3).
Overexpression of atf1 ϩ Specifically Rescues the apc5-1 Mutant-The application of osmotic stress resulted in suppression of temperature-sensitive growth of mutations in nearly all of the APC/C subunits (supplemental Fig. S2). We next wished to examine whether overexpression of atf1 ϩ likewise globally rescued compromised APC/C function or whether suppression was specific to the apc5-1 mutant allele. To ask the latter question, another apc5 ts allele, apc5-96, that has mutations in not only the C-terminal domain but also in the N-terminal domain (see Table 2) was used. Overexpression of atf1 ϩ rescued the apc5-1 mutant, but neither the apc5-96 mutant nor any of the other APC/C subunit mutants tested (supplemental Fig. S3). On the other hand, overexpression of another suppressor we identified, apc4/cut20/lid1 ϩ , rescued both apc5-1 and apc5-96 mutants (data not shown). These results demonstrated three major points. First, Atf1 displays genetic interactions with Apc5. Second, Apc4 works together or in close relationship with Apc5 to control APC/C activity, in agreement with previous reports (6,10,32). Indeed, we also observed that overexpression of apc5 ϩ rescued the apc4/cut20-100 ts mutant strain. Third, Atf1 seems to regulate APC/C activity. This last point is further supported by the finding that atf1⌬ had synthetic effects on the mitotic defects of apc5-1; mitotic defects, such as the "cut" or "nuc" phenotypes in apc5-1 at the restrictive temperature, were increased from 23.9 to 47.6% in apc5-1 atf1⌬ double mutant background. In view of this, we asked whether a heterodimeric partner of Atf1, Pcr1 has a similar role in regulating the APC/C. However, overexpression of pcr1 ϩ did not suppress the temperature sensitivity of the apc5-1 mutant and deletion of pcr1 ϩ did not result in any synthetic effects with apc5-1 (data not shown), suggesting that atf1 ϩ specifically rescues the apc5-1 mutant allele in a dose-dependent manner.   The bZIP Domain of Atf1 Is Not Necessary to Suppress the apc5-1 Mutant-To investigate which domain of Atf1 is necessary to rescue the ts phenotype of apc5-1, we made several constructs of Atf1 including a deletion of the C-terminal basic leucine zipper domain (bZIP⌬) and a deletion of the N-terminal 120 or 224 residues (N120⌬ or N224⌬) (Fig. 2A). The basic leucine zipper (bZIP), present in many transcription factors, determines dimerization specificity with heterodimeric partners as well as mediating binding to DNA (33)(34)(35)(36). Notably, overexpression of atf1-bZIP⌬ still suppressed the ts phenotype of apc5-1, although the suppression was slightly weaker than the full-length atf1 ϩ (Fig. 2B). Further deletion (C343⌬) caused loss of the rescue activity. Conversely, Atf1 lacking the N-terminal 120 residues (N120⌬) partially rescued apc5-1, whereas further deletion (N224⌬) resulted in loss of suppression.
Given this finding, we wanted to assess whether the bZIP deletion construct, Atf1-bZIP⌬ was indeed transcriptionally inactive. To this end, the atf1 deletion strain (atf1⌬) was transformed with either an empty vector (pREP41) or with the plas-mids pREP41-atf1 ϩ , or pREP41-atf1-bZIP⌬. The resulting strains were assessed for levels of Atf1-dependent transcripts such as hsp9 ϩ or ctt1 ϩ . As shown by Northern blotting in Fig.  2C, overexpression of atf1 ϩ induced hsp9 ϩ and ctt1 ϩ transcription, whereas overexpression of atf1-bZIP⌬ had no effects. Furthermore, to validate this result, we created a strain with atf1-bZIP⌬ under its native promoter in the chromosome and examined whether atf1-bZIP⌬ is responsive to osmotic stress. As seen with atf1⌬, the atf1-bZIP⌬ strain was unable to induce Atf1-dependent transcription upon stress, whereas the atf1 ϩ wild type strain responded fully under the same conditions (Fig.  2D). These results indicate that deletion of the bZIP domain of Atf1 abolishes its ability to direct transcription. Taking the above results together, we conclude that Atf1 regulates the APC/C ubiquitin ligase in a manner independent of its role as a transcription factor and independent of its ability to bind to DNA.
Atf1 Physically Binds the APC/C-We hypothesized that Atf1 might physically interact with Apc5 and stimulate APC/C activity. To investigate whether there is a physical interaction between Atf1 and APC/C, we expressed GST-tagged Atf1 in a wild-type (apc5 ϩ -HA) strain and pulled-down Atf1 using GSH-Sepharose beads followed by immunoblotting with anti-HA to detect the HA-tagged Apc5. Apc5 was co-purified with GST-Atf1, but not with GST alone (Fig. 3A). We also checked whether endogenous Atf1 binds APC/C under normal physiological conditions. As shown in Fig. 3B, anti-Atf1 immunoprecipitates (IP) contained Apc5 whereas the mock IP did not. These results indicate that Atf1 physically binds the APC/C, presumably via Apc5. We therefore speculated that the ts mutation in apc5-1 might weaken the interaction between Atf1 and APC/C, and thus overexpression of atf1 ϩ suppresses the ts phenotype of apc5-1 by promoting the interaction. Consistent with this model, Apc5 ts mutant protein was less efficiently retained by immobilized Atf1 than wild-type Apc5 (Fig. 3C, see lanes 3  and 4). Conversely, when atf1 ϩ was overexpressed in the apc5-1 ts mutant cells, the interaction between Atf1 and APC/C was restored (Fig. 3D).
Atf1 Stimulates the Activity of the APC/C Ubiquitin Ligase-Finally, to assess the impact of Atf1 on the activity of the APC/C, we expressed and purified Atf1 proteins from insect cells and added them to a cell-free APC/C-dependent ubiquitylation system using fission yeast E1, E2, and E3 (APC/C) enzymes to detect any interaction with Atf1. In control experiments, Cdc13/cyclin B was efficiently ubiquitylated within 10 min and more than four ubiquitin chains were assembled on Cdc13 as seen by the molecular weight shifts (Cdc13-Ubn, Fig.  4A). To investigate whether Atf1 is involved in the kinetics of polyubiquitylated Cdc13, we monitored earlier stages of ubiquitylation (1-5 min). When purified His-or GST-tagged Atf1-bZIP⌬ was added, the APC/C-dependent ubiquitylating activity was accelerated and polyubiquitylated Cdc13 bands were clearly detected in 5 min, whereas these bands were barely observed in the absence of Atf1 (Fig. 4B). We also used another APC/C substrate, Cut2/securin to confirm the effects of Atf1 on the APC/C. Consistently, addition of His-or GST-tagged Atf1-bZIP⌬ enhanced ubiquitylation of Cut2 by the APC/C FIGURE 2. Domain analysis of Atf1 to rescue apc5-1. A, schematic diagram of Atf1 and truncated constructs used in this study. A basic region followed by a leucine zipper domain, that is essential for DNA binding as well as dimerization, is shown as bZIP. B, cultures of apc5-1 atf1⌬ strain expressing indicated plasmids were spotted with decreasing cell number onto minimal medium, and incubated at either 26 or 33°C. C, levels of hsp9 ϩ and ctt1 ϩ mRNA were assessed by Northern blot analysis from cultures of atf1⌬ cells expressing atf1 ϩ , vector, or atf1-bZIP⌬. Ribosomal RNAs stained with ethidium bromide are shown as a loading control. D, cultures of either atf1⌬, atf1-bZIP⌬, or wild-type cells were grown in rich medium and then incubated for the indicated time in the same medium containing 1.2 M sorbitol. The levels of hsp9 ϩ (heat shock protein), ctt1 ϩ (catalase), pyp2 ϩ (tyrosine phosphatase), and cdc2 ϩ were assessed by Northern blot analysis. Ribosomal RNAs stained with ethidium bromide are also shown. (Fig. 4C). These data indicate that Atf1 plays a role in activating the APC/C ubiquitin ligase.
We next asked the physiological relevance of the interaction between Atf1 and APC/C. One of the characteristic phenotypes of the atf1⌬ strain is sterility, although this is not absolute (15)(16)(17). Presumably, the main reason for sterility is due to lack of Atf1-dependent transcription of genes, required for sexual development, such as ste11 ϩ . However, our results suggest that Atf1 might contribute to the stimulation of the APC/C. In support of our model, we found that after nitrogen starvation, Cdc13/cyclin B disappearance was slower in atf1⌬ strain than wild-type (wt) or atf1-bZIP⌬ strain (supplemental Fig. S4). Furthermore, when zygotic meiosis is induced in homothallic h 90 atf1⌬ strain, expression of atf1-bZIP⌬ under the native promoter in the chromosome increased the mating efficiency (percentage of conjugation) from 12.6 to 42.5% (Fig. 4D). Therefore, Atf1 seems to play a role in degrading cyclin B and aiding conjugation under nitrogen starvation by controlling the activity of the APC/C.

DISCUSSION
In this study, we provide five lines of evidence indicating that the transcription factor Atf1 has an additional role in activating the APC/C ubiquitin ligase apart from its well known DNA binding-based roles such as activating transcription. First, overexpression of atf1 ϩ can rescue the ts phenotype of the apc5-1 mutant. Second, the mitotic defect phenotype is increased when atf1⌬ is combined with apc5-1 mutation, although atf1⌬ itself has no evident mitotic defects. Third, the bZIP domain is not required for Atf1's ability to complement the apc5-1 mutant. Fourth, Atf1 physically interacts with the APC/C in vivo. Finally, purified Atf1 can promote the APC/C-dependent ubiquitylation of Cdc13/cyclin B as well as Cut2/securin in a cell-free system.
Atf1 is a member of the helix-loop-helix bZIP transcription factor family that plays important roles in the regulation of many cellular processes including stress response, survival, and death. Undoubtedly, the best-characterized role of this family is fulfilled by its transcriptional regulation of a wide set of genes, however, it is plausible that it may, in part, work through protein-protein interactions. In the context of our present study, it is intriguing to note that ATF2, the mammalian homologue of Atf1, controls the stability of TIP60 (also known as HTATIP), a histone acetyltransferase (HAT) as part of the DNA damage response pathway (37). It appears that ATF2 binds TIP60 as well as Cul3 and affects the E3 ubiquitin ligase in a manner similar to our findings with Atf1 and Apc5. Under non-stressed conditions, ATF2 interacts with TIP60 and stimulates ubiquitylation of TIP60 by the Cul3/Roc1 ubiquitin ligase. Upon DNA damage, ATF2 dissociates from TIP60. TIP60 is therefore stabilized and can now regulate the activation of ATM by acetylation.
It is worth noting that overexpression of ste11 ϩ rescues the apc5-1 mutation better than that of atf1 ϩ and also rescues sev- FIGURE 3. Interaction between Atf1 and APC/C. A, extracts were prepared from cultures of wild-type (apc5 ϩ -HA) cells expressing GST alone or GST-Atf1. Atf1-binding proteins were isolated by GSH-Sepharose beads and analyzed by SDS-PAGE followed by immunoblotting with anti-HA antibody. B, mock or anti-Atf1 immunoprecipitates from wild-type (apc5 ϩ -HA) were analyzed by SDS-PAGE followed by immunoblotting with anti-HA antibody. C, same as B, but wild-type (lanes 1 and 3) and apc5-1-HA ts mutant cells (lanes 2 and 4), which had been grown at 26°C first and then shifted to 34°C for 4 h were used. The input (2%) and anti-Atf1 immunoprecipitates were analyzed by SDS-PAGE followed by immunoblotting with anti-HA and anti-Atf1 antibodies. D, overexpression of Atf1 restored the interaction between Atf1 and APC/C. Cultures of apc5-1-HA ts mutant cells expressing vector (lanes 1 and 3) or atf1 ϩ (lanes 2 and 4) were grown at 26°C for 8 h in minimal medium and shifted to 34°C for 4 h. The input and anti-Atf1 immunoprecipitates were analyzed by indicated antibodies, same as C.  1-3). B, same as A, but His-tagged Atf1-bZIP⌬, GST-tagged Atf1-bZIP⌬, or mock (buffer) were added into the in vitro ubiquitylation assay, and samples were taken at the indicated time points. C, same as B, but Cut2/securin was used as a substrate. D, cultures of h 90 (wt), h 90 atf1-bZIP⌬ or h 90 atf1⌬ strains were grown in EMM medium and then shifted to EMM-N (without nitrogen) for 24 h. The percentage of conjugation was estimated by microscopic examination. Mean values from three independent experiments are shown. Bars represent S.D.
eral APC/C subunit mutations. Presumably, Ste11 acts as a transcription factor and stimulates APC/C activity or helps exiting from mitosis. In fact, several cell cycle regulators including apc2 ϩ , rum1 ϩ , and mik1 ϩ have been reported to be targets of the Ste11 transcription factor (38,39).
How does the physical interaction between Atf1 and APC/C enhance the activity of the APC/C? Although we do not have direct evidence, Atf1 presumably interacts with the Apc5 subunit (see supplemental Fig. S3). The recent cryo-EM study of the APC/C suggests that Apc5 is located in the platform domain, which interacts with Apc4 and Apc1 on one side and with the TPR subcomplex (i.e. Apc3/Nuc2/Cdc27-Apc6/Cut9/ Cdc16-Apc8/Cut23/Cdc23) on the other side (40,41). Therefore, it is tempting to speculate that the interaction between Atf1 and Apc5 may cause a conformational change of the APC/C core complex, affecting the positioning of recruited substrates relative to the catalytic center, thereby promoting their ubiquitylation. Alternatively, because apc4/cut20/lid1 ϩ is also a multicopy suppressor of apc5-1 mutant, it is possible that Atf1 stabilizes the interaction between Apc4 and Apc5, which might be important to provide a solid Apc1/Cut4 scaffold to accommodate two separable subcomplexes in the correct spatial relationship. However, because we failed to observe a difference in the interaction between Apc4 and Apc5 in vivo regardless of the presence or absence of Atf1 (data not shown), Atf1 may perform a rather more subtle or transient role than can be detected by simple binding experiments. The precise roles of Atf1 await clarification in future studies.
At the moment, we do not know whether this mechanism is conserved in higher eukaryotes. However, it is possible that one of the many members of the ATF/CREB family of proteins may perform a similar function to Atf1 and activate the APC/C. In more general terms, we propose that the activity of the APC/C ubiquitin ligase in vivo might be regulated by multiple protein interactions other than the Fizzy/Cdc20 family of proteins or the spindle assembly checkpoint (SAC), so that the correct substrates are ubiquitylated efficiently at the right time and at the correct location (e.g. chromatin, spindle, centrosome/SPB, nucleus). In addition, considering the complexity of the APC/C, it is possible that a transient and temporal interaction with a "modulator" protein controls the APC/C to fine-tune its performance. Fission yeast Atf1 and human CBP/p300 may be members of a new family of regulators which may offer a new layer of regulation of the APC/C, which may well be conserved in mammalian cells. Clearly, more studies will be required to confirm and extend these observations.