The mitotic checkpoint is a cellular safeguard that prevents chromosome missegregation in eukaryotic cells [1, 2]. Suboptimal functioning may foster chromosome missegregation in cancer cells [3]. Checkpoint signaling produces the "mitotic checkpoint complex" (MCC), which prevents anaphase by targeting Cdc20, the activator of the anaphase-promoting complex/cyclosome (APC/C). Recent biochemical and structural studies revealed that the human MCC binds two Cdc20 molecules, one (Cdc20M) through well-characterized, cooperative binding to Mad2 and Mad3/BubR1 (forming the "core MCC") and the other one (Cdc20A) through additional binding sequences in Mad3/BubR1 [4-6]. Here, we dissect the different functionality of these sites in vivo. We show in fission yeast that, at low Cdc20 concentrations, Cdc20M binding is sufficient for checkpoint activity and Cdc20A binding becomes dispensable. Cdc20A binding is mediated by the conserved Mad3 ABBA-KEN2-ABBA motif [7, 8], which we find additionally required for binding of the MCC to the APC/C and for MCC disassembly. Strikingly, deletion of the APC/C subunit Apc15 mimics mutations in this motif, revealing a shared function. This function of Apc15 may be masked in human cells by independent mediators of MCC-APC/C binding. Our data provide important in vivo support for the recent structure-based models and functionally dissect three elements of Cdc20 inhibition: (1) sequestration of Cdc20 in the core MCC, sufficient at low Cdc20 concentrations; (2) inhibition of a second Cdc20 through the Mad3 C terminus, independent of Mad2 binding to this Cdc20 molecule; and (3) occupancy of the APC/C with full MCC, where Mad3 and Apc15 are involved.