Cyclin D1 is destabilized during S phase through the ubiquitin-proteasome pathway in cancer cells. (A, B) Expression profile of cyclin D1 during cell cycle progression in cells released from quiescence. Normal cells (A) and cancer cells (B) were tested. (A) NIH 3T3 mouse fibroblasts and CCD841 CoN normal colon epithelium. (B) HCT 116 and SW480 cells. CCD841 CoN cells were synchronized at G0/G1 phase by treatment with ACL-4 media without EGF for 24 hrs, and then stimulated with complete ACL-4 media containing EGF. Other cells were released from quiescence by serum stimulation. Samples were collected at indicated time points. Cell cycle distributions were determined by flow-cytometry and percentages of each phase are indicated. Western-blots were performed with cyclin D1 and β-actin antibodies, respectively. (C, D) Pulse-chase analysis of cyclin D1 in NIH 3T3 cells (C) and HCT 116 cells (D). Cells were released from quiescence and labeled with ³⁵S-methionine for 1 hour when most of the populations were in G1 phase (9 hrs for NIH3T3 and 6 hrs for HCT 116) or in S phase (21 hrs for NIH3T3 and 15 hrs for HCT 116). The cells were chased with cold methionine for the indicated times and then lysed. Cyclin D1 was immunoprecipitated and analyzed with SDS-PAGE. Autoradiography was performed. Levels of metabolically labeled-cyclin D1 were estimated by quantitative scanning using Quantity One software (Bio-Rad) and blotted on the graph to determine the half-life of cyclin D1. Cell cycle distribution is indicated below the figures. (E) Turnover of cyclin D1 is mediated by the ubiquitin-proteasome pathway. NIH3T3 and HCT 116 cells were released from quiescence by serum stimulation and treated in the presence (+) or absence (−) of MG132 for 2 hrs prior to harvesting at each time point. Western blot was performed with cyclin D1 and β-actin antibodies. (F–H) Polyubiquitination of cyclin D1. (F) HCT 116 colon cancer cells were transfected with HA-tagged ubiquitin cDNA (lanes 1–3) or without it (lane 4) and then synchronized to S phase through the sequential manipulation of serum starvation and stimulation. Cells were treated with 25 µM MG132 (lanes 3 and 4) or without it (lanes 1 and 2) for one hour. Lysates were immunoprecipitated with antibodies to cyclin D1 antibody (lanes 2–4) or control IgG (lane 1) and immunoblotted with a HA antibody (upper panel) or a cyclin D1 antibody (lower panel). Asterisks indicate background non-specific bands (F–G). (G) Nuclear (N) and cytoplasmic (C) proteins were fractionated (Fr.) from cell lysates collected in Panel F, lane 3. Nuclear and cytoplasmic extracts were immunoprecipitated with antibodies to cyclin D1 (lanes 1 and 2) or IgG (lane 3) and immunoblotted with a HA antibody (upper panel) or a cyclin D1 antibody (lower panel). (H) Cell cycle distributions.

FBXW8 ubiquitinates cyclin D1 in a Thr286 phosphorylation-dependent manner. (A) IP-IB analysis (left). Protein from exponentially growing HCT 116 cells was precipitated with antibodies to cyclin D1 or IgG. Immunoprecipitates were subjected to SDS-PAGE and sequentially blotted with cyclin D1, CDK4, SKP1, CUL1 and CUL7 antibodies. IB analysis with 5% of total cell lysates was provided as a control (right). (B) IB analysis following depletion of SKP1 expression for 48 hrs after treatment with SKP1 siRNA double-strand oligonucleotides in HCT 116 cells. Non-targeting siRNA (Control) and mock transfection (−) served as controls. (C) IP-IB analysis. Twenty-six F-box full-length encoding cDNAs were cloned into V5 or Flag epitope tag expression vectors. These V5 or Flag-tagged F-box protein DNA plasmids were transfected together with HA-tagged cyclin D1 (HA-Cyc D1) and CDK4 expression vectors into T98G cells. Cells were collected 24 hrs later. Samples were precipitated with a HA epitope tag antibody. Immunoprecipitates were subjected to SDS-PAGE and subsequently stained with V5 or Flag (F-box proteins), HA (Cyc D1) antibodies. IB analysis with 10% of total cell lysates was provided (bottom). (D) IP-IB analysis (top). V5-tagged F-box protein DNA plasmids were transiently transfected together with either HA-tagged cyclin D1 (Cyc D1) wild type (WT) or T286A mutant, and CDK4 expression vectors in T98G cells respectively. Samples were precipitated with a HA epitope-tag antibody. Immunoprecipitates were subjected to SDS-PAGE and subsequently blotted with V5 (F-box proteins), HA (Cyc D1), and Thr286 phosphorylated cyclin D1 (pCyc D1 Thr286) antibodies. IB analysis with 10% of total cell lysates was provided for comparison (bottom). (E) In vitro ubiquitination assay. In vitro translated F-box proteins with recombinant GST-full-length cyclin D1 (Cyc D1) wild type, HeLa cell extracts Fraction II with ATP, Ubiquitin and ERK2, and in vitro-translated either SKP1, RBX1 and CUL1, or SKP1, RBX1 and CUL7 proteins were incubated at 30°C for 2 hrs. Samples were separated by SDS-PAGE and immunoblotted with a cyclin D1 antibody. (F) In vitro polyubiquitination of cyclin D1 through the SCF-like (SCFL) complex FBXW8 (SKP1-CUL7-FBXW8-RBX1/SCFL^FBXW8). WT or T286A GST- Cyc D1 was incubated in the presence of purified ERK2 (lanes 1, and 3) or its absence (lane 2) at 30°C for 2 hrs. Samples were separated by SDS-PAGE and immunoblotted with a cyclin D1 antibody. Asterisk indicates non-specific bands. (G) Reconstitution of polyubiquitination of cyclin D1 through SCFL^FBXW8 in vitro using purified E1 and E2. GST-WT Cyc D1 was incubated with recombinant SCFL^FBXW8 in the presence or absence of E1 and E2/UbcH5C. Samples were separated by SDS-PAGE and immunoblotted with a cyclin D1 antibody.

Cyclin D1 degradation in the cytoplasm is essential for cell proliferation. (A) Viable HCT 116 cells after siRNA-mediated knockdown of FBXW8, CUL1, or CUL7 expression. siRNAs were transfected on days 0, 1, 2, and 4. Cells were collected as indicated, stained with trypan blue, and counted with a haemocytometer. (B) Western blot analysis with total and fractionated nuclear proteins. Protein blotting was performed with total cyclin D1 (Cyc D1), pCyc D1 Thr286, CDK1, Histone H1, and MEK1 antibodies. HCT 116 cells were treated with control siRNA (Cont) or FBXW8 siRNA (W8) for 48 hrs. Samples were then fractionated into nuclear or cytoplasmic proteins, or prepared as total cell lysates. The bottom panel shows a CDK1-associated Histone H1 in vitro kinase assay using nuclear protein. (C) Generation of cyclin D1 ecdysone-inducible (IND) system in HCT 116 cells. Ectopic expression of HA-tagged T286A was induced in by 10 µM Ponasterone A (Pon A). (D) Viable cells after T286A cyclin D1 induction. T286A IND HCT 116 cells were cultured in the presence of Pon A (T286A) or its absence, and control (Cont) or FBXW8 siRNA. (E) Colony formation assay. One hundred single T286A IND HCT 116 cells were cultured in the presence of Pon A (+) or its absence (−), and control siRNA (Cont) or FBXW8 siRNA. Cells were cultured for 2 weeks, and stained with 0.5% crystal violet containing 20% ethanol. (F) A model of cyclin D1 ubiquitination mediated by the complex containing FBXW8. FBXW8 recognizes cyclin D1 through a WD40 repeat motif in an ERK/MAPK-mediated Thr286 phosphorylation-dependent manner. SKP1 interacts with FBXW8 together with CUL1 or CUL7 via a domain in the N-terminus of FBXW8, called F-box. CUL1 or CUL7 recruits RBX1, which in turn conscripts E2, a ubiquitin-conjugating enzyme, to add a multiubiquitin chain to cyclin D1.

MAPK phosphorylates cyclin D1 at Thr286, which triggers subsequent ubiquitination. (A) Identification of the D-domain in cyclin D1 (Cyc D1). The illustration of full-length human Cyc D1 shows the region of the D-domain (A.A. 179-193) and the MAPK phosphorylation site Thr (T) 286 followed by proline (P) (solid bar and an arrow, respectively). The amino acid sequence of the D-domain within Cyc D1 is aligned with other known MAPK-docking sites of various ERK substrates. The doublet of basic (+) and nonpolar (ϕ) amino acids are conserved residues in the core D-domain motif L/I/V-X-L/I/V. Amino acid positions of the most 5′ residues of the D-domains are indicated with numbers in the left of each amino acid sequence respectively. (B, C) p42 ERK2 in vitro kinase assays for cyclin D1 (upper panels). (B) Wild type or T286A mutant recombinant GST- full length Cyc D1 protein was mixed with ³²P-ATP in the kinase assay reaction buffer in the presence or absence of purified ERK2. Reactions were performed at 30°C for 30 min and stopped by adding sample loading buffer. Samples were separated with SDS-PAGE and ³²P-uptake was detected by autoradiography. Immunoblotting (IB) analysis using antibodies to cyclin D1 is provided as a reference to show substrate amounts. (C) GST alone (lane 1), GST-C-terminal WT Cyc D1 fusion protein retaining the binding site of MAPK (amino acids 165-295, lane 2), T286A (lane 3) and a complete deletion of the D-domain ΔD, lane 4) were used. IB analysis using antibodies to GSTs provided as a reference to show substrate amounts. (D) Immunoprecipitation (IP) and IB analysis following ectopic expression of Flag-tagged ERK2 together with either HA-tagged WT or ΔD Cyc D1 in HCT 116 colon cancer cells. A Western blot for input controls (10% total lysates) is also shown. (E) Western blot analysis with Thr286 phosphorylated cyclin D1 (pCycD1 Thr286) and total Cyc D1 following transfection of various forms of HA-tagged Cyc D1 expression vectors in HCT 116 colon cancer cells. (F, G) In vitro ubiquitination assays using HeLa cell extracts Fraction II as a source of the enzymes necessary to conjugate ubiquitin to substrates and ATP. (F) GST-full length Cyc D1 WT, T286A, or ΔD were used for a reaction either with recombinant ERK2 (lanes 3–5) or without it (lanes 1–2). Samples were separated by SDS-PAGE and immunoblotted with a cyclin D1 antibody. ATP was added to all lanes, and no ubiquitin was added to lane 1. (G) GST-full length Cyc D1 WT was used with or without ubiquitin. After SDS-PAGE separation, immunobloting was performed with antibodies to cyclin D1 (lanes 1, 2) or ubiquitin (lanes 3, 4). (H, I) Pulse-chase analysis of cyclin D1 in HCT 116 cells after exposure to U0126. Exponentially growing HCT 116 cells were treated with DMSO or 10 µM U0126 for 30 min. (H) Cells were pulse-labeled with ³⁵S-methionine and were chased for the indicated times. Cyclin D1 was immunoprecipitated and then analyzed with SDS-PAGE. Autoradiography was performed. (I) Levels of metabolically labeled-cyclin D1. (J) Western blot analysis with pCycD1 Thr286, total cyclin D1, phosphorylation specific ERK (pERK), and total ERK antibodies. (K) Cell cycle distributions are shown.

The stability of cyclin D1 is regulated by FBXW8 complexes through the ubiquitin-proteasome pathway. (A, C) IB analysis. HCT 116 cells were infected with a retrovirus expressing FBXW8 (A), a ΔF mutant form (ΔF FBXW8, Panel C) or a control retrovirus expressing GFP (A, C). Forty-eight hours later, cells were harvested and Western blot analysis was performed with antibodies to cyclin D1, cyclin E, Flag (FBXW8 and ΔF FBXW8), GFP and β-actin. (B) IP–IB analysis. Empty (mock) or Flag-tagged WT FBXW8, and ΔF FBXW8 DNA plasmids were transiently transfected in T98G cells. Samples were precipitated with a Flag epitope tag antibody. Immunoprecipitates were subjected to SDS-PAGE and subsequently blotted with antibodies to cyclin D1, SKP1, CUL1, CUL7, RBX1 and Flag (F-box proteins). (D) IB analysis following depletion of FBXW8 expression for 48 hrs through siRNA in HCT 116 cells. Non-targeting siRNA (Control) and mock transfection (−) were controls. (E–G) Knockdown of FBXW8 or its partner CUL1 or CUL7 through siRNA stabilizes cyclin D1 expression in HCT 116 cells. (E) IB analysis following depletion of CUL1, CUL7 or FBXW8 expression for 48 hrs through siRNA or mismatch (MM) oligonucleotides in HCT 116 cells. Non-targeting siRNA (Control) and mock transfection (−) were controls. (F, G) Pulse-chase analysis of cyclin D1 following depletion of CUL1, CUL7 or FBXW8 expression for 48 hrs through siRNA in HCT 116 cells. Control cells were treated with non-targeting siRNA. Cells were pulse-labeled with ³⁵S-methionine for an hour, chased with cold methionine as indicated, and then lysed. Cyclin D1 was immunoprecipitated and analyzed via SDS-PAGE. (G) Levels of metabolically labeled-cyclin D1 were estimated by quantitative scanning using Quantity One software and plotted graphically to determine the half-life of cyclin D1. (H, I) Immunofluorescence. Nuclei were visualized with Hoechst dye. Subcellular localization of E3 ligase. HCT 116 cells were transfected with V5 epitope-tagged FBXW8 pcDNA3. Cells were synchronized 24 hrs later by sequential manipulation of serum starvation and stimulation. At 6 hrs (G1 phase) or 15 hrs (S phase), cells were fixed. Immunofluorescence was performed with a V5 epitope tag antibody followed by Alexa Fluor 488-conjugated anti-mouse IgG antibody (green) and a rabbit cyclin D1 polyclonal antibody followed by Alexa Fluor 594-conjugated anti-rabbit IgG antibody (red). (I) Cell cycle distributions.