APC/C-mediated multiple monoubiquitylation provides an alternative degradation signal for cyclin B1

Abstract

The anaphase-promoting complex or cyclosome (APC/C) initiates mitotic exit by ubiquitylating cell-cycle regulators such as cyclin B1 and securin. Lys 48-linked ubiquitin chains represent the canonical signal targeting proteins for degradation by the proteasome, but they are not required for the degradation of cyclin B1. Lys 11-linked ubiquitin chains have been implicated in degradation of APC/C substrates, but the Lys 11-chain-forming E2 UBE2S is not essential for mitotic exit, raising questions about the nature of the ubiquitin signal that targets APC/C substrates for degradation. Here we demonstrate that multiple monoubiquitylation of cyclin B1, catalysed by UBCH10 or UBC4/5, is sufficient to target cyclin B1 for destruction by the proteasome. When the number of ubiquitylatable lysines in cyclin B1 is restricted, Lys 11-linked ubiquitin polymers elaborated by UBE2S become increasingly important. We therefore explain how a substrate that contains multiple ubiquitin acceptor sites confers flexibility in the requirement for particular E2 enzymes in modulating the rate of ubiquitin-dependent proteolysis.

Figure 1

Ubiquitin vinyl sulfone (UbVS) inhibits cyclin B1 degradation by depleting available ubiquitin. (a) ³⁵S-labeled cycB1-NT and 20 or 44 μM of wild-type (WT) Ub were introduced into mitotically-arrested Xenopus extract that had been pre-treated with UbVS (20 μM) or buffer (untreated) for 30 min. Proteolysis was measured by release of trichloroacetic acid (TCA) soluble counts, and plotted as percent of input radiolabeled cycB1-NT. Trends are representative of three or more independent experiments. (b) Wild-type ubiquitin or forms of ubiquitin (44 μM) bearing single-point mutations in distinct interaction surfaces on ubiquitin, along with radiolabeled substrate, were added to UbVS-treated extract. For panels (c–e), mitotically-arrested Xenopus extract was pre-treated with UbVS (20 μM) or buffer (untreated) for 30 min. Wild-type ubiquitin (44 μM) was added to extract, as indicated. Aliquots were withdrawn at indicated times and analyzed by SDS-PAGE and western blot. (c) Ubiquitin status in Xenopus extract was examined by anti-ubiquitin western blot. (d) Levels of Ub-charged endogenous UBCH10 were examined by anti-UBCH10 western blot. Aliquots were removed at the indicated times and quenched with either nonreducing sample buffer to examine levels of UBCH10~Ub or reducing sample buffer to examine total levels of UBCH10. (e) Same as in d, but levels of endogenous ubiquitin-charged UBCH5 were examined by western blot. Uncropped images of immunoblots are presented in Supplementary Fig. S9

Figure 2

Ubiquitin chain formation is not essential for cyclin B1 degradation in UbVS-treated Xenopus extract. ³⁵S-labeled cycB1-NT and different forms of Ub, where indicated, were introduced concomitantly into mitotically-arrested Xenopus extract that had been pre-treated with UbVS (20 μM) or buffer (untreated) for 30 min. Proteolysis was measured by release of trichloroacetic acid (TCA) soluble counts, and plotted as percent of input radiolabeled cycB1-NT. Trends are representative of three or more independent experiments. (a) Ubiquitin types (44 μM) with single lysine-to-arginine mutations at indicated positions or at all three positions Lys11, 48 and 63 (Ub^triR) were added to UbVS-treated extract. Ub^WT refers to wild-type ubiquitin and Ub^me refers to methylated ubiquitin. (b) Ub^11R and substrate were introduced into UbVS-treated extract, and supplemented with Ub^11R or buffer control 15 min after initiation of degradation. (c) Ubiquitin types (44 μM) used, as indicated. Ubiquitin^Konly refers to ubiquitin that has all of its lysines, except for those specified, mutated to arginines. (d) Degradation was measured in the presence of different ubiquitin types (44 μM) containing arginine substitutions at two of the three principle sites of ubiquitin-ubiquitin conjugation by the APC/C.

Figure 3

Cyclin B1 proteolysis depends on Lys11-linked Ub chain formation only when the number of available lysine residues is restricted. (a) Sequence comparison of the cyclin B1 N termini from multiple species. Lysine residues are colored in blue; the destruction box (D-box) is colored in orange. (b) Schematic representation of the N terminal region (residues 1–88) of human cyclin B1 with lysine residues denoted K in blue and with the D-box motif denoted with an orange rectangle. Cyclin B1 mutants (cyc^K64only and cyc^{K59, 63, 64, 67only}) were generated by substituting lysine with arginine at all but the specified lysine residues within the first 115 amino acids (residues 89–115 not shown). (c) Purified full-length wild-type (cyc^WT) or single-lysine (cyc^K64only) cyclin B1, in complex with CDK1, and forms of Ub (20 μM) as indicated were added to mitotic Xenopus egg extract that had been pre-treated with UbVS (20 μM) for 30 min. Stability of the exogenous substrate over time was assessed by SDS-PAGE and cyclin B1 western analysis. (d) Performed as in c except the behavior of full-length cyc^{K59, 63, 64, 67only}, in complex with CDK1, was analyzed. Uncropped images of immunoblots are shown in Supplementary Fig. S9.

Figure 4

UBCH10 and APC/C catalyze rapid multiple monoubiquitination of cyclin B1 that is sufficient for binding ubiquitin receptors. (a) Western analysis of in vitro ubiquitination reaction containing full-length cyclin B1, APC/C immunopurified from mitotically-arrested Xenopus extract, recombinant UBCH10 (100 nM), and forms of Ub (118 μM), as indicated. Ubiquitin types with lysine-to-arginine mutations at one, two, or at all three positions Lys11, 48 and 63 (Ub^triR), as well as methylated ubiquitin (Ub^me) were used. Control “−APC/C” reactions containing all components except for the E3 ligase were performed in parallel. Reactions were allowed to proceed for 15 or 90 min before analysis by SDS-PAGE/western blotting against cyclin B1. (b) Time-course of the in vitro ubiquitination of full-length wild-type cyclin B1 with Ub^WT or Ub^triR and remaining components as in a. (c–f) Binding of ubiquitinated cyclin B1 to GST-tagged Ub receptors. Cyclin B1-Ub conjugates were incubated with immobilized receptor proteins for 1 h at 4 °C before reaction products were subjected to SDS-PAGE/western blot analysis against cyclin B1. Equivalent amounts of input (I), flow-through (FT) and bound (B) were loaded in adjacent lanes. Binding experiments with wild-type Rpn10 (c) and Rad23 (e). Binding with corresponding versions of the receptors lacking the Ub recognition domains, with engineered block substitution of the UIM domain (LAMAL → NNNNN) of Rpn10 (d) or deletion of the UBA domains of Rad 23 (f).

Figure 5

Multiply monoubiquitinated cyclin B1 is rapidly degraded by purified proteasomes and in Xenopus extract. (a) In vitro degradation assay with cyclin B1-Ub species generated with immunopurified Xenopus APC/C, recombinant UBCH10 (250 nM), and forms of Ub (145 μM), as indicated and USP14-deficient human proteasomes (20 nM). Aliquots were removed at the indicated times and reaction products analyzed by SDS-PAGE and anti-cyclin B1 western. (b) Same as in a, but conjugates were generated with UBC4 as the E2 enzyme. (c) Autoradiograph of in vitro APC/C-UBCH10 catalyzed ubiquitination of ³⁵S-cycB1-NT (1–88) with immunopurified Xenopus APC/C, recombinant UBCH10 (100 nM) and forms of Ub (145 μM) as indicated. Products from a 60-minute ubiquitination assay were separated by SDS-PAGE and analyzed using a phosphorimager. (d) CycB1-NT-Ub species from c were incubated with purified human proteasomes (20 nM) reconstituted with or without 20-fold molar excess of GST-tagged wild-type USP14. At indicated times, reactions were terminated by addition of trichloroacetic acid (TCA). Proteolysis was measured by release of TCA soluble counts, and plotted as percent of input radiolabeled cyclin B1 protein. See Supplementary Fig. S5c for additional controls. (e) CycB1-NT-Ub species from c were added to interphase Xenopus extract that had been pre-treated with UbVS (15 μM) or buffer control for 30 min. Reactions were terminated by addition of TCA at indicated times. Proteolysis was measured by release of TCA soluble counts, and plotted as percent of input radiolabeled cycB1-NT.

Figure 6

UBE2S is required for cyclin B1 proteolysis only when ubiquitination is constrained to a single lysine. (a) Mitotically-arrested Xenopus extract was immunodepleted with UBE2S antibody or control IgG. Samples were further incubated with Ub agarose (10:1 ratio of extract to resin) to enrich for E2 enzymes and bound proteins were analyzed by SDS-PAGE/immunoblotting. Lanes 4–6 represent 25-fold enrichment of E2 enzymes on Ub agarose. Levels of UBE2S, UBCH10 and APC/C were examined by immunoblotting. Asterisks, nonspecific signal. Uncropped images of immunoblots are shown in Supplementary Fig. S9. (b) Rate of degradation of ³⁵S-labeled cycB1-NT in UBE2S- or control-depleted mitotic Xenopus extract from a. Recombinant His-UBE2S (10 nM), where indicated, was added to reactions concomitantly with substrate. Proteolysis was measured by release of TCA soluble counts, and plotted as percent of input radiolabeled cycB1-NT. (c) Time-course of degradation of full-length cyclin B1 (cyc^WT) or single-lysine-containing mutant (cyc^K64only), each in complex with CDK1, in control- or UBE2S-depleted mitotic Xenopus extract. Recombinant His-UBE2S (10 nM), where indicated, was added to reactions concomitantly with substrate. Cyclin B1 proteolysis was analyzed by SDS-PAGE and immunoblotting. Asterisks represent nonspecific signal. Uncropped images of immunoblots are shown in Supplementary Fig. S9. (d) Model of cyclin B1 degradation in Xenopus cell-cycle extract. The APC/C and the E2 UBCH10 collaborate to transfer ubiquitin monomers to multiple lysine residues on cyclin B1, with subsequent elaboration of short ubiquitin chains containing K63, K48 and K11 linkages, with K11 linkages predominating. Upon achievement of a threshold of ubiquitin mass, which appears to be 4–5 ubiquitin monomers, multiply monoubiquitinated substrate can associate with proteasome-associated ubiquitin receptors and be degraded efficiently. However, when the number of lysine residues in cyclin B1 is restricted, ubiquitination catalyzed by UBCH10 is insufficient for rapid proteolysis and the activity of UBE2S in extending K11-linked ubiquitin polymers becomes important for efficient degradation.