br APC C Uses a
APC/C Uses a Dynamic Cullin–RING Mechanism to Elongate Polyubiquitin Chains Human APC/C generates Lys11-linked poly-Ub chains through an entirely different mechanism, via the distinctive E2 enzyme, UBE2S 30, 31, 32. Although APC2 and APC11 are necessary and sufficient to activate UBE2S, the mechanism was not overtly obvious because UBE2S lacks the hallmark E2 residues known to engage RING domains . Indeed, mutation of the APC11 residues binding UBE2C do not impair UBE2S-mediated polyubiquitylation 74, 75. Furthermore, unlike the reaction with UBE2C that is stimulated by coactivator, the intrinsic Ub chain building activity of UBE2S (i.e., transfer of a donor Ub onto a free acceptor Ub) is activated by recombinant APC/C that can be prepared independently of coactivator 21, 74. Instead, a primary function of APC/C is recruiting and positioning the acceptor Ub for its Lys11 to accept another Ub from UBE2S 74, 75. A structural model for the unprecedented interactions between UBE2S, its target (i.e., a Ub that has been linked already to an APC/C substrate), and APC/C was generated by hybrid structural studies merging information from NMR, crosslinking, and mutational data with a low-resolution cryo-EM map of APC/CCDH1 complexed with a proxy for the Ub chain NVP 231 intermediate (Figure 4C) . Although the details of these interactions await high-resolution studies, it is clear that APC/C engages UBE2S in a bipartite manner, but this differs completely from interactions with UBE2C. UBE2S is anchored to APC/C by a flexibly tethered extension C terminus of the catalytic domain of UBE2S 30, 31, 32. Here, the extreme C-terminal residues of UBE2S pack into a pocket between the APC2 N-terminal domain and APC4 β-propeller 47, 50. Additionally, the UBE2S catalytic domain interacts with an APC2 surface that differs completely from previously described E3–E2 interactions 50, 74, 75. The RING is also crucial, as it possesses a distinctive binding site recruiting Ub for modification by UBE2S . Apparently, the catalytic geometry whereby the RING domain of APC11 can present Ub for modification by UBE2S is attainable even with APC2-APC11 in the “down position”, which accounts for APC/C activating UBE2S-dependent generation of unanchored poly-Ub chains even in the absence of a coactivator and a substrate 21, 52, 74, 75. Overall, the data suggest that UBE2S can be activated by various positions of the APC2–APC11 cullin–RING catalytic core, although at this point, it remains unknown whether the position of the APC11 RING domain required to engage UBE2C allows simultaneous positioning of Ub for modification by UBE2S, or how UBE2C and UBE2S might cofunction with APC/C to generate branched Ub chains.
Inhibiting APC/C during Interphase and prior to Anaphase Because ubiquitylation by APC/C triggers cell division, it is essential that APC/C is restrained until cells are prepared for its substrates to be degraded. In addition to regulation by phosphorylation, an additional layer of control comes from cellular inhibitors also restricting ubiquitylation until needed. This is best understood for APC/CCDH1 inhibition by early mitotic inhibitor (EMI)1 in interphase , and APC/CCDC20 inhibition by the MCC prior to anaphase 76, 77, 78, 79, 80. EMI1 and MCC may also play roles in localizing APC/C to specific subcellular structures or substrates , although the structural basis for this regulation remains unknown. EMI1 and MCC hijack the mobile substrate recognition and catalytic modules by binding multiple sites on APC/CCDH1 and APC/CCDC20, respectively (Box 1). However, they achieve these functions through completely different routes. In addition, MCC does not block UBE2S-dependent polyubiquitylation and even binds APC/C in distinct configurations that differentially modulate ubiquitylation, while the available structure of a single EMI1 molecule bound to APC/C shows inhibition of all ubiquitylation. EMI1 regulates the coupling of mitosis and DNA replication . Briefly, after cells divide, APC/CCDH1 activity must be restrained. Although this is ultimately achieved by CDH1 phosphorylation, the required CDK activity is too low in G1 due to APC/CCDH1-dependent degradation of cyclins. Thus, accumulation of cyclins during G1 depends on EMI1 inhibiting APC/CCDH1. The best-characterized portion of the 50-kDa EMI1 protein is its 16-kDa C-terminal domain, which consists of four inhibitory elements: a D box, linker, zinc binding region (ZBR), and C-terminal tail 47, 58, 59, 60. Each individual EMI1 element only weakly interacts with APC/CCDH1, but together they synergistically bind numerous APC/CCDH1 domains to avidly inhibit (Figure 4D and Box 1) 47, 58, 59, 60. The EMI1 D box binds simultaneously to CDH1 and APC10 to block substrate access 47, 58, 59, 60. The linker and ZBR together act like a wedge to simultaneously capture the UBE2C-binding surface of the APC11 RING domain and elements of APC2 and APC1 47, 59. This has several effects including seizing the RING domain away from a catalytic conformation, while also walling off a portion of the central cavity. Finally, the C-terminal tail of EMI1 shares sequence homology with the C-terminal tail of UBE2S, docks in the same groove between APC4 and APC2, and prevents UBE2S binding 47, 59, 60. Although EMI1 locks the APC/C structure in a rigid conformation, the intrinsic flexibility of the APC11 RING domain enables its initial capture by EMI1.