Substrate-specific regulation of ubiquitination by the anaphase-promoting complex

Abstract

By orchestrating the sequential degradation of a large number of cell cycle regulators, the ubiquitin ligase anaphase-promoting complex (APC/C) is essential for proliferation in all eukaryotes. The correct timing of APC/C-dependent substrate degradation, a critical feature of progression through mitosis, was long known to be controlled by mechanisms targeting the core APC/C-machinery. Recent experiments, however, have revealed an important contribution of substrate-specific regulation of the APC/C to achieve accurate cell division. In this perspective, we describe different mechanisms of substrate-specific APC/C-regulation and discuss their importance for cell division.

Introduction

By determining the stability of cell cycle regulators, ubiquitination plays a crucial role in controlling cell division.1–3 The aberrant activation of ubiquitination enzymes can lead to unwanted cell division or survival, which often results in cancer. Dissecting regulatory mechanisms that control ubiquitination enzymes is, therefore of paramount importance to understanding tumorigenesis and to providing novel therapeutic strategies against this devastating disease.

The modification of proteins with ubiquitin requires a cascade of three distinct enzymes, referred to as E1 (ubiquitin-activating enzyme),4 E2 (ubiquitin-conjugating enzyme)5 and E3 (ubiquitin ligase).6 Based on bioinformatic studies, human cells have between 600 and 1,000 different E3s.6 Many of these E3s contain a characteristic RING-domain to recruit an E2 charged with ubiquitin, in addition to a wide variety of domains used to interact with specific substrates. In the cullin-family of RING-E3s (CRLs), a large cullin subunit provides a scaffold to bind both a RING-domain subunit and a substrate-recruitment factor. Several of these CRLs, including the Skp1-Cullin-F-box complex (SCF) and the anaphase-promoting complex (APC/C), have long been recognized as essential E3s required for cell cycle control.1,3

Although the SCF and APC/C both target cell cycle regulators for degradation, they differ in important aspects. For example, by recruiting the E2 Ube2R (yeast Cdc34), the SCF assembles K48-linked ubiquitin chains, the canonical proteasomal targeting signal. By contrast, the human APC/C utilizes an initiating E2, Ube2C (UbcH10) and an elongating E2, Ube2S, to catalyze the formation of K11-linked ubiquitin chains.7–10 The K11-linked ubiquitin chains also trigger the degradation of their substrates by the 26S proteasome.7,11 The reasons for why the APC/C assembles K11-linked ubiquitin chains to promote protein degradation are not known.

In addition to these differences in catalysis, the SCF and APC/C are controlled by distinct regulatory mechanisms. The SCF is active throughout all stages of the cell cycle, but usually recognizes its substrates only after these have been tagged for degradation through phosphorylation or related posttranslational modifications.12 Thus, most of SCF-regulation is thought to occur on the level of substrates. By contrast, the activity of the APC/C is restricted to mitosis and G₁ by phosphorylation of APC/C-subunits,13 binding of several distinct APC/C-inhibitors14 or degradation of APC/C-specific E2s.8,15 The full activation of human APC/C during mitosis leads to a massive upregulation of K11-linked chain formation, resulting in the sequential degradation of its many substrates.11,16 Much of this substrate-ordering can be recapitulated in purified systems, indicating that substrates do not have to be modified for APC/C-recognition.17 These observations have led to the suggestion that most of APC/C-regulation occurs on the level of the E3, and not the substrate.

Recent studies, however have revealed that in addition to the intrinsic regulation of the core APC/C machinery, substrate-specific control mechanisms are also critical for accurate cell cycle progression. Here, we describe how substrate-specific APC/C-regulation can be accomplished and discuss potential advantages of this type of APC/C-regulation for cell cycle control.

Substrate-Specific APC/C-Regulation through Post-Translational Modifications

The APC/C recruits its substrates through the adaptor proteins Cdc20 and Cdh1, which directly bind D-box and/or KEN-box sequences (Table 1).14,18–20 Following substrate binding, the E2 Ube2C or a core subunit of the APC/C recognizes another substrate motif, referred to as TEK-box, to rapidly promote ubiquitin chain initiation (Table 1).7 Additional APC/C-interaction motifs with less understood binding partners or functions have also been described (Table 1).21–23 The modification of residues within or in proximity to such substrate motifs has been found to allow for substrate-specific APC/C-regulation.

As the APC/C needs to recognize D-boxes for substrate-binding, phosphorylation of residues near these motifs provides an intuitive means of stabilization. Indeed, such a mechanism was reported to prevent the degradation of the APC/C-substrate Cdc6 in cells that have entered the cell cycle after quiescence (Fig. 1).24 Cdc6 is required for the assembly of pre-replication complexes during G₁, although the APC/C is still active at this time of the cell cycle.24–27 To counteract the APC/C, the cyclin E-Cdk2 kinase phosphorylates serine residues in the vicinity of the D-box of Cdc6, which interferes with binding of Cdc6 to the substrate recruitment factor of the APC/C, Cdh1, and disrupts the APC/C-dependent degradation of Cdc6.28 A similar mechanism was shown to ensure the proper timing of the APC/C^Cdc20-dependent degradation of the yeast anaphase inhibitor securin.29 Moreover, recognition of the ribonucleotide reductase subunit RRM2 by APC/C^Cdh1 is inhibited through phosphorylation near a KEN-box motif, showing that such regulation is not restricted to D-boxes (Michele Pagano, pers. communication). The phosphorylation of residues in proximity to APC/C-recognition motifs can, therefore, selectively inhibit the degradation of specific APC/C-substrates by blocking their recognition through Cdc20 or Cdh1.

A slightly different mechanism regulates the levels of Aurora A kinase, a critical regulator of mitosis. The APC/C-dependent degradation of Aurora A was found to be inhibited by phosphorylation of a specific serine residue within an N-terminal sequence motif referred to as the “A-box”.21 This serine residue is phosphorylated during early stages of mitosis, most likely by cyclin B1/Cdk1, in both Xenopus laevis and human cells.30 Consistent with these observations, Aurora A needs to be dephosphorylated by the protein phosphatase PP2A in order to be efficiently degraded during mitotic exit.31 The function of the A-box in APC/C-dependent reactions is not fully understood, and it has remained unclear how exactly serine phosphorylation leads to stabilization of Aurora A. As epitope-tagging of Aurora A at its N-terminus interferes with its proper degradation,30 the A-box might serve to correctly orient Aurora A on the APC/C, and phosphorylation might interfere with this event. Nevertheless, these results suggest that phosphorylation of substrate residues might also interfere with APC/C-dependent ubiquitination at steps other than substrate binding.

Although most is known about phosphorylation-dependent substrate stabilization, it is possible that other posttranslational modifications serve similar roles in regulating the proteolysis of specific APC/C-substrates. Indeed, it has been suggested that acetylation of cyclin A or BubR1 is able to modulate their stability during mitosis, potentially by interfering with APC/C-dependent ubiquitination.32,33 In addition, farnesylation has recently been reported to regulate the APC/C-dependent degradation of CENP-F and CENP-E.34 Farnesylation also promotes the targeting of these proteins to the kinetochore, suggesting that their APC/C-dependent ubiquitination and intracellular localization are tightly connected to each other.35 Together, these observations indicate that posttranslational modifications provide powerful and versatile means of regulating the ubiquitination of specific substrates by the APC/C.

Substrate-Specific APC/C-Inhibition through Reversible Protein Interactions

Similar to phosphorylation, reversible protein interactions are able to inhibit the APC/C-dependent degradation of specific substrates, as recently observed for the turnover of the spindle assembly factors (SAFs) HURP and NuSAP.36 HURP and NuSAP promote spindle formation by bundling microtubules, thereby stabilizing kinetochore fibers that connect the mitotic chromosomes to the spindle.37,38 Both SAFs are inhibited by binding to nuclear transport receptors of the importin-family, with the precise mechanism of this inhibition still being poorly understood. It is known, however, that the interaction between SAFs and importins can be disrupted by the small protein Ran, when it is bound to GTP (Ran^GTP).39,40 As the guanine-nucleotide exchange factor of Ran, Rcc1, associates with mitotic chromosomes,41,42 Ran^GTP is enriched in the proximity of chromatin, leading to the localized activation of HURP and NuSAP at these sites. The reversible binding of importins to HURP and NuSAP can, therefore, regulate their activity in chromatin-driven spindle formation.

In a beautiful example of co-regulation, Ran and importins also regulate the APC/C-dependent degradation of HURP and NuSAP.36 The ubiquitination of these SAFs by the APC/C depends on specific D-, KEN- and TEK-boxes. By binding to overlapping motifs, importin-β blocks the association of HURP and NuSAP with the substrate-adaptors of the APC/C, Cdc20 and Cdh1 (Fig. 2).36 Ran^GTP in turn disrupts the interaction between HURP and NuSAP with importin-β, thereby exposing the SAFs to recognition by Cdc20/Cdh1 and ubiquitination by the APC/C. The regulation of APC/C-dependent degradation by Ran and importin-β is thus conceptually similar to phosphorylation, in that importin-β reversibly interferes with substrate recognition by Cdc20 or Cdh1.

What is the relevance of this mechanism for cell cycle control? Importin-β-binding deficient mutants of HURP are prematurely degraded during spindle assembly, at the time when the SAF is normally active.36 Conversely, aberrant spindles form with high frequency in cells that express higher levels of HURP or NuSAP,36,43,44 suggesting that the levels of active SAFs have to be tightly controlled to achieve proper spindle assembly or spindle dynamics. By using importin-β as a common regulator of activity and degradation, the cell ensures that APC/C specifically targets the active pool of SAFs. This might provide spatial (i.e., degrade SAFs that were released from importin-β at an inappropriate location in the mitotic cell) or temporal (i.e., degrade SAFs that had been active for a sufficient amount of time) control of spindle formation. These two mechanisms of regulation are not mutually exclusive, as motor-proteins could remove active SAFs from spindle microtubules, thereby leading to SAF-degradation. Consistent with this hypothesis, the yeast APC/C-substrate Cik1 appears to require an interaction with the kinesin Kar3 for efficient degradation.45

Substrate-Specific APC/C-Activation through Protein Interactions

While the masking of APC/C-recognition motifs through modifications or binding partners provides means of APC/C-inhibition, offering additional routes of APC/C-binding should accelerate the degradation of specific substrates. An example for this type of regulation is centered around the homologous proteins Cks1 and Cks2, which bind to complexes between cyclin-dependent kinase Cdk1 and cyclin A or cyclin B, two long known substrates of the APC/C.46 Cks1/2 facilitate the degradation of the cyclins,47,48 but not other APC/C-substrates, demonstrating that they act as substrate-specific activators of the APC/C (Fig. 3).

How do Cks1/2 promote cyclin degradation on a mechanistic level? The Cks proteins were known to contain a phospho-peptide binding domain, the function of which is important for mitotic progression.49 The APC/C is heavily phosphorylated during mitosis,13,50 and consistent with this, it was shown that Cks1/2 are able to directly bind phosphorylated APC3 during prometaphase.48,51,52 The interaction between Cks1/2 and the APC/C occurs in parallel to the interaction between the substrate, cyclin, and the APC/C-substrate recruitment factors, Cdc20 or Cdh1.48,53 Such multivalent binding is predicted to increase the affinity of the substrates for the APC/C, thereby positively affecting the processivity of ubiquitin chain formation. It had been shown before that the processivity of ubiquitin chain formation is a major determinant of the timing of APC/C-substrate degradation.17 Thus, binding partners that provide additional ways to APC/C-recruitment are an effective means of increasing the processivity of chain formation, thereby accelerating substrate degradation during the cell cycle.

Regulated Ubiquitination of the APC/C-Specific E2s Ube2C and Ube2S

Several activators of the APC/C, including Cdc20, Cdh1 or the E2s Ube2C and Ube2S, are also substrates of the APC/C.8,14,15 To ensure the correctly timed degradation of canonical APC/C-substrates, it is critical to prevent the premature turnover of APC/C-activators. Especially for the E2s Ube2C and Ube2S, the mechanisms achieving this regulation are different from those restricting the binding of APC/C-substrates, thereby representing another layer of substrate-specific APC/C-control.

Ube2C acts as the physiological initiating E2 for the APC/C,7,11,54 which adds the first ubiquitin to most, if not all APC/C-substrates. In addition to modifying substrates, Ube2C is also subject to APC/C-dependent autoubiquitination and degradation.8,15 The latter reaction is inhibited, when substrates are bound to the APC/C, as Ube2C then ubiquitinates the substrate, but not itself (Fig. 4). Given the recent finding that APC/C-activation leads to a dramatic upregulation of bulk K11-linked chain formation,11 the APC/C is likely saturated with substrates throughout mitosis, which efficiently delays Ube2C-degradation and APC/C-inactivation until G₁. The correct timing of Ube2C-degradation was not observed with amino-terminally tagged Ube2C,55 most likely because the epitope tag interferes with known functions of the N-terminus of Ube2C in controlling charging by E1,56 binding to the APC/C,57 and catalyzing ubiquitin transfer to itself.15 These observations, however, implied that the peculiar timing of Ube2C-degradation is linked to sequences in its N-terminus, which requires further investigation. Results from these studies should also provide insight into mechanisms underlying the competition between canonical APC/C-substrates during mitosis.17

The chain elongating E2 of the APC/C, Ube2S, is also subject to APC/C-dependent degradation during G₁, which requires a functional active site in Ube2S itself.8 In addition to providing E2 activity towards substrates, Ube2C and Ube2S collaborate with p31^comet to silence the spindle checkpoint, thus driving full APC/C-activation at metaphase.8,9,57–59

The role of the APC/C in silencing the spindle checkpoint was recently also observed upon chemical inhibition of the APC/C.60 Accordingly, the tightly regulated degradation of Ube2C and Ube2S facilitates APC/C-inactivation during G₁, but it may also function to protect cells from premature checkpoint silencing and inaccurate sister chromatid separation. In agreement with this hypothesis, the overexpression of Ube2C and Ube2S has been linked to the generation of aneuploidy and tumorigenesis,61–63 underscoring the importance of substrate-specific APC/C-regulation.

Conclusions

As an E3 required for the accurate execution of the mitotic program, the APC/C has to be tightly controlled. This is in part achieved by mechanisms that modulate the activity of the core APC/C-machinery, including the spindle checkpoint or inhibitors, such as Emi1. The APC/C is also able to discriminate between its many substrates by catalyzing their ubiquitination with distinct degrees of processivity, thereby providing a blueprint for the timing of substrate degradation. As discussed here, these intrinsic features of APC/C-regulation are overlain by substrate-specific control mechanisms, including phosphorylation or sequestration in specific protein complexes. In this manner, the APC/C can fine-tune the degradation of especially important mitotic regulators, such as cyclins or spindle assembly factors, to faithfully orchestrate cell cycle progression.

Figures and Tables

Figure 1 Inhibition of APC/C-dependent ubiquitination by posttranslational modifications. The ubiquitination of Cdc6 by the APC/C can be inhibited by phosphorylation of serine residues in proximity to an important D-box in the substrate. Cdc6 is phosphorylated by the cyclin E/Cdk2 kinase during late stages of G₁. It can be assumed that a phosphatase is required for Cdc6 degradation during the next G₁ phase.

Figure 2 Inhibition of APC/C-dependent ubiquitination by reversible protein interactions. The ubiquitination of the spindle assembly factor HURP by the APC/C can be inhibited by importin-β, which binds to sites in HURP that overlap with the D-, KEN- and TEK-boxes required for HURP-degradation. GTP-charged Ran can dissociate importin-β from HURP, thereby exposing it to recognition by Cdc20/Cdh1 and ubiquitination by the APC/C.

Figure 3 Stimulation of APC/C-dependent ubiquitination by activator-independent APC/C-binding motifs. The ubiquitination of cyclin B1 is stimulated by Cks1, a component of the cyclin B1/Cdk1 complex. Cks1 contains a phospho-peptide binding motif, which interacts with phosphorylated APC3 during mitosis. The multivalency resulting from the simultaneous recognition of the D-box in cyclin B1 and Cks1 by phosphorylated APC/C accelerates cyclin B1 ubiquitination and degradation during mitosis.

Figure 4 Competitive inhibition of APC/C-dependent ubiquitination. Ube2C is the physiological chain-initiating E2 for the APC/C. When substrates are bound to the APC/C, Ube2C transfers ubiquitin to substrates lysine residues. By contrast, after substrates have been degradation, Ube2C modifies lysine residues within itself, leading to Ube2C-degradation.

Table 1 Substrate motifs involved in APC/C-dependent ubiquitination

APC/C recognition motif	Consensus/Representative sequence (key residues are indicated in bold)	Original paper
D-box	Rxx Lxx xxN	19
KEN-box	KEN xxx N	20
A-box	QRI LGP S*NV PQR V	21
GxEN-box	GxE N	22
O-box	PAS PLT EKN AK	23
TEK-box	KKM TEK TVK AK	7

* The phosphorylation of the serine residue (indicated in bold) blocks the ubiquitination of Aurora A by APC/C.

Acknowledgements

We thank Julia Schaletzky for discussions and for critically reading the manuscript. We are also grateful to all members of our lab for many discussions and ideas. Work in the lab is funded by March of Dimes, the NIGMS, an NIH Director's New Innovator Award, and the Pew Foundation.