|Home | About | Journals | Submit | Contact Us | Français|
Skp2, a F-box protein that determines the substrate specificity for SCF ubiquitin ligase, has recently been demonstrated to be degraded by Cdh1/APC in response to TGFβ signaling. The TGFβ-induced Skp2 proteolysis results in the stabilization of p27 that is necessary to facilitate TGFβ cytostatic effect. Previous observation from immunocytochemistry indicates that Cdh1 principally localizes in the nucleus while Skp2 mainly localizes in the cytosol, which leaves us a puzzle on how Skp2 is recognized and then ubiquitylated by Cdh1/APC in response to TGFβ stimulation. Here, we report that Skp2 is rapidly translocated from the cytosol to the nucleus upon the cellular stimulation with TGFβ. Using a combinatorial approach of immunocytochemistry, biochemical-fraction-coupled immunoprecipitation, mutagenesis as well as protein degradation assay, we have demonstrated that the TGFβ-induced Skp2 nucleus translocation is critical for TGFβ cytostatic effect that allows physical interaction between Cdh1 and Skp2 and in turn facilitates the Skp2 ubquitylation by Cdh1/APC. Disruption of nuclear localization motifs on Skp2 stabilizes Skp2 in the presence of TGFβ signaling, which attenuates TGFβ-induced p27 accumulation and antagonizes TGFβ-induced growth inhibition. Our finding reveals a cellular mechanism that facilitates Skp2 ubiquitylation by Cdh1/APC in response to TGFβ.
Transforming Growth Factor β (TGFβ) signaling regulates a variety of cellular processes, including cell proliferation, differentiation, apoptosis and fate specification during embryogenesis. Transduction of the complex signaling starts on the cell surface where TGFβ binding induces the formation of type I and II receptor complex. Type II receptor phosphorylates the GS domain of Type I receptor, resulting in kinase activation. Type I receptor then propagates the signal through phosphorylation of Smad2 or Smad3 (receptor-regulated Smad, R-Smad) on the carboxy-terminal SXS motif. Upon phosphorylation, Smad2 or Smad3 releases from SARA (Smad anchor for receptor activation), forms oligomeric complex with Smad4 (co-mediator Smad, Co-Smad) and translocates to the nucleus where they regulate target gene transcription in collaboration with DNA-binding co-factors such as forkhead family member FOXH1, co-activators such as p300 or co-repressors such as Ski.1,2
TGFβ signaling plays a critical role in tumorigenesis of several types of epithelia, paradoxically switching from a role as a tumor suppressor to a promoter of invasiveness and metastasis during tumor progression.3 The cytostatic effect of TGFβ in early stages of tumorigenesis primarily involves activated transcription of the cyclin-dependent kinase inhibitors (CDKI) p15,4 p21,5 and p57,6 and repressed transcription of the growth promoting transcription factors c-Myc and Id. Repression of c-Myc and Id facilitates the induction of p15 and p21, which is a critical prerequisite for the execution of the full cytostatic response of epithelial cells to TGFβ.7–11 Increased expression of p15 facilitates its binding to CDK4 and CDK6 and displaces p27 from these kinases to cyclin E-CDK2, causing CDK2 inhibition. Concomitantly, p21 is induced and binds to CDK2 as well, ensuring a maximal suppression of CDK2 activity and cell cycle arrest at G1/S transition.12 In addition to transcriptional control, the proteolytic regulation of CDKIs is also involved in cell cycle arrest, as mediated by TGFβ. Our recent studies demonstrate that p27, a short-lived regulatory protein, is stabilized in response to TGFβ stimulation through Skp2SCF degradation targeted by Cdh1/APC.13 Since p27 is constitutively degraded by Skp2SCF, the stabilization of p27 by TGFβ facilitates its complete and successful redistribution from CDK4 and CDK6 to cyclin E-CDK to achieve arrest of cell growth.
Although our findings indicate that TGFβ-induced Skp2 degradation is governed by Cdh1/APC, the mechanism by which APC is regulated in response to TGFβ signaling and how Skp2 is ubiquitylated for destruction by Cdh1/APC still remain unclear. Result from our recent immunocytochemistry provided us an unexpected indication that Skp2 resides mainly in the cytosol of Mv1Lu mink epithelial cells in the absence of TGFβ. Given the notion that Cdh1 is principally located in the nucleus, we face a puzzle on how Cdh1/APC and Skp2 communicate with each other given that they are in different cellular compartments not to permit recognition and ubiquitylation of Skp2 by its E3 ligase Cdh1/APC. Using a combinatorial approach encompassing cell biology and biochemistry, we have revealed that Skp2 cellular localization is quite dynamic in response to TGFβ signaling, while Cdh1 localization is relatively stable. Our present results show, upon stimulation with TGFβ, Skp2 is rapidly translocated from the cytosol to the nucleus. The measurement of TGFβ-induced nuclear localization of Skp2 by immunocytochemistry and biochemical cellular fractionation further suggests that Skp2 translocation is a critical condition precedent for the ubiquitin protein ligase Cdh1/APC to physically interact with its substrate Skp2. Characterization of Skp2 translocation by altering its NLS (nuclear localization signal) results in the stabilization of Skp2 in the presence of TGFβ signaling, which consequentially results in a failure to accumulate p27 that is necessary for TGFβ-induced growth inhibition. This present work fills a knowledge gap on how Skp2 is regulated by Cdh1/APC in response to TGFβ for growth inhibition, which further advances our understanding of the molecular basis of TGFβ signaling pathway.
Recent studies have demonstrated that TGFβ signaling pathway is tightly regulated by the UPS (ubiquitin-proteasome system),14 where UPS targets various components of the TGFβ pathway including cytoplasmic second messengers, transmembrane bound receptors and accumulated nuclear proteins.15–17 Our endeavor to search for TGFβ-induced fast turnover proteins led us to identify Skp2 as a rapidly degraded protein in response to TGFβ signaling. As shown in Figure 1A–C, Skp2 is rapidly degraded in response to TGFβ stimulation, which in turn results in accumulation of p27. The half-life of Skp2 in response to TGFβ signaling is approximately 60 minutes (Fig. 1B). Furthermore, Skp2 degradation is blocked by incubating cells with 50 µM of a proteasome inhibitor, MG-132 (Fig. 1C). To exclude the possibility that the observed change in Skp2 protein levels is due to altered transcriptional regulation, we measured Skp2 mRNA levels after stimulation with TGFβ by RT-PCR. As shown in Figure 1A, Skp2 mRNA remains constant, suggesting that the drop in Skp2 protein levels is due to protein degradation.
Previous study indicated that Skp2 protein abundance is regulated by Cdh1/APC during cell cycle.18,19 We previously demonstrated that Cdh1/APC activity is enhanced in response to TGFβ stimulation.20 Thus, we tested the possibility that Cdh1/APC is a putative E3 ligase that governs the TGFβ-induced Skp2 degradation. As shown in Figure 1D and E, depletion of Cdh1 by RNA interference in Mv1Lu results in significant attenuation of Skp2 degradation in response to TGFβ, while Skp2 is drastically destroyed in response to TGFβ in the wild-type cells.
To evaluate the biological significance of Skp2 degradation upon TGFβ stimulation, we constructed a stable Skp2 with deletion of its destruction box (ubiquitin-targeting-degron) and further engineered a Mv1Lu cell that stably expresses stabilized Skp2 (D box deletion) (Fig. 1F). As shown in Figure 1F and G, failure of Skp2 proteolysis in response to TGFβ stimulation significantly antagonizes TGFβ-induced growth inhibition. Taken all together, our results suggest an important role for Skp2 proteolysis by Cdh1/APC in TGFβ-mediated growth inhibition.
Previous work suggested that the cellular presence of Cdh1 is in the nucleus.21,22 Localization of Skp2 has been variable depending on cell type.23–27 To assess the localization of Cdh1 and Skp2 in Mv1Lu cell, we performed immunocytochemistry in the presence and absence of TGFβ. Localization of Cdh1 was measured by using monoclonal antibody against Cdh1 and FITC conjugated anti-mouse secondary antibody. As show in Figure 2A, expression of Cdh1 is localized in nucleus, which is consistent with previous observation. To test the localization of Skp2, we used rabbit antibody against Skp2 coupled with Texas Red conjugated secondary antibody. As shown in Figure 2A, the majority of Skp2 in the absence of TGFβ is localized in the cytosol, which is consistent with known previous Skp2 localization in prostate, melanoma, colon, lymphoma and breast cancer tissues.23–27 To examine the potential effect of TGFβ stimulation on cellular translocation of Cdh1 and Skp2, a similar study was performed as described in the above. As shown in Figure 2A, TGFβ stimulation results in significant translocation of Skp2 from the cytosol to the nucleus, while no obvious redistribution of Cdh1 was observed.28 The finding of TGFβ-induced Skp2 nuclear translocation provides us an important clue to explain the possible mechanism of how Skp2 is recognized and catalyzed by Cdh1/APC for ubiquitylation and degradation in response to TGFβ signaling.
To biochemically confirm the observation from our immunostaining, we measured the interaction between Cdh1 and Skp2 in the absence and presence of TGFβ by immunoprecipitation. As shown in Figure 2B, coimunoprecipitation of Cdh1 and Skp2 were detected in nuclear fraction after stimulation with TGFβ. In contrast, no obvious interaction of Cdh1 and Skp2 was observed by co-IP in the cytosol in both absence and presence of TGFβ.
To determine where Cdh1 interacts with Skp2 and catalyzes the ubiquitylation of Skp2, we performed an ubiquitylation assay by using either cytosolic or nuclear fraction. Skp2 was pulled down with antibody against Skp2. The ubiquitin conjugated Skp2 was measured by immunoblotting of Skp2 IP complex with antibody against ubiquitin.13,29 As shown in Figure 2C, Skp2 ubiquitin conjugates were significantly detected in the nuclear fraction after stimulation with TGFβ, while no obvious Skp2 ubiquitin conjugates were visualized in the cytosol fraction in both absence and presence of TGFβ. Taken together, the above results demonstrated Skp2 localization is tightly regulated in response TGFβ, where TGFβ-induced Skp2 nuclear translocation permits recognition of Skp2 by Cdh1/APC and its ubiquitylation by Cdh1/APC.
To elucidate the mechanism by which Skp2 translocates into the nucleus in response to TGFβ, we analyzed the Skp2 sequence for a nuclear localization signal (NLS). The search revealed a conserved NLS motif (Fig. 3A) (cubic.bioc.columbia.edu/newwebsite/services/predict NLS/), a sequence necessary and sufficient for nuclear import of other host proteins.30,31 To test whether the putative NLS in Skp2 mediates its nuclear transport, we deleted the NLS. HA-tagged wild-type Skp2 or NLS mutant Skp2 was transfected into Mv1Lu cells and extracts were prepared from cells exposed (or not exposed) to TGFβ. The fractionated cytosol and nuclear preparations were further examined by immunoblotting. Consistent with the data from the immuno-staining assay, the majority of Skp2 was observed in the cytosol in the absence of TGFβ but redistributed to the nucleus in response to TGFβ stimulation (Fig. 3B). As predicted, deletion of the NLS on Skp2 completely blocked TGFβ-induced Skp2 nuclear translocation (Fig. 3C and D). To ask whether nuclear translocation is coupled to Skp2 degradation, we measured the change in Skp2 levels in response to TGFβ. As shown in Figure 3E and F, deletion of the NLS significantly attenuated the Skp2 degradation in response to TGFβ. These results suggest that the NLS in Skp2 is required for TGFβ-induced Skp2 translocation and destruction.
Impaired Skp2 translocation could disturb the recognition and ubiquitylation of Skp2 by Cdh1/APC and cause the stabilization of Skp2 and proteolysis of p27. Failure to degrade Skp2 and the resulting stabilization of p27 in response to TGFβ should antagonize the TGFβ response. To test this hypothesis, wild-type Skp2 or NLS-deleted Skp2 was introduced stably into Mv1Lu cells by retroviral infection (Fig. 4A).13,32 Subsequently, the pools of infected cells were measured for their ability to respond to TGFβ-induced growth inhibition.13 As shown in Figure 4A and B, stably expressed Skp2 degraded in response to TGFβ stimulation while NLS-deleted Skp2 was quite stable. Growth inhibition analysis showed that expression of wild-type Skp2 only moderately blocked the ability of cells to undergo TGFβ-induced cell cycle arrest. The reason for this may be because the wild-type Skp2 was unstable in response to TGFβ stimulation and thus failed to accumulate (Fig. 4C). In contrast, NLS-deleted Skp2 markedly attenuated TGFβ-induced growth inhibition (Fig. 4C). Therefore, failure of Skp2 translocation in response to TGFβ impairs the TGFβ-induced growth inhibitory effect. These results support the hypothesis that TGFβ-induced Skp2 translocation facilitates its degradation by Cdh1/APC, which helps to maintain p27 protein levels to inhibit cyclin E/CDK2.
Previous studies from us and others have implicated a role for Cdh1/APC in TGFβ signaling pathway.20,33 Elevated activity for Cdh1/APC was measured in the presence of TGFβ. The TGFβ-induced activation of APC was thought to remove SnoN, a transcriptional co-suppressor, for the induction of TGFβ-responsive genes and to degrade Skp2 to stabilize p27 for TGFβ cytostatic effect.13,32,34 However, how Cdh1/APC targets Skp2 for degradation remains a puzzle since both components are localized in different cellular compartments in the testing model cell line-Mv1Lu. In the present work, finding that Skp2 translocates to the nucleus in response to TGFβ and the further identification of Skp2's nuclear translocation motifs have unveiled the mystery of how Skp2 is recognized for ubiquitylation by Cdh1/APC in the presence of TGFβ signaling. This work elucidates the mechanism by which Skp2 is regulated by Cdh1/APC in TGFβ signaling.
Skp2 was first identified as a cyclinA-CDK2 S-phase kinase-associated protein35 and subsequently characterized as a rate-limiting component of the SCF machinery that ubiquitylates p27 for degradation, which in turn releases cyclin E-CDK2 and promotes entry into S phase.36,37 p27 functions as a tumor suppressor and its inactivation predisposes mice to tumorigenesis. It has been proposed that Skp2 confers oncogenic function through accelerated degradation of p27 and other tumor suppressors, such as CDKIs p21 and p57; RASSF1 (Ras association domain family 1); and RBL2 (retinoblastoma-like, also known as p130).38 Indeed, increased levels of Skp2 proteins are often observed in many human cancers where Skp2 overexpression is usually inversely correlated with reduced p27 expression, and positively correlated with tumor malignancy and poor prognosis in terms of cancer treatment.39,40 Mechanistic studies show that, in addition to gene amplification and mRNA accumulation, deregulated proteolysis is also involved in elevated Skp2 expression in human cancers. We and other groups have demonstrated that in normal cell cycle, Cdh1/APC targets Skp2 for degradation, thus preventing premature entry into S phase.18,19 Our recent work further reveals that the reduced expression of Cdh1 and inverse elevation of Skp2 are evident in various human cancers. Moreover, the deregulated Cdh1/Skp2/p27 cascade is implicated in the development of breast and colorectal cancer.41–43 The tumor suppressor potential of Cdh1 is further illustrated in knockout mice. Cdh1 heterozygous mice show increased susceptibility to spontaneous tumors.44 Given the pivotal role of TGFβ signaling in tumorigenesis and the notion that Cdh1/APC mediates TGFβ-induced cell growth inhibition as demonstrated here and by other studies,13,34 abrogation of Cdh1/APC function in TGFβ signaling could be one mechanism for the initiation of tumor formation.
Besides aberrant Cdh1 activity, Skp2 cytoplasmic localization might also be another mechanism for the abnormal accumulation of Skp2 protein in human cancers, which allows Skp2 to escape from Cdh1-mediated degradation. Cytoplasmic Skp2 has been observed in many clinical tumor samples and correlated with aggressive malignancy. Recent mechanistic studies suggest that Akt-dependent phosphorylation of Skp2 at Ser 72 is responsible for Skp2 cytoplasmic translocation.27,28 In this study, both immunofluorescence staining and compartmental immunoprecipitation assay shows that Skp2 resides mainly in the cytosol of Mv1Lu mink lung epithelial cells, and translocates to nucleus in response to TGFβ signaling. Identification and characterization of translocation motifs on Skp2 advanced the knowledge about the Cdh1-Skp2-p27-cyclin E/CDC2 axis in TGFβ signaling. Nevertheless, the mechanism of how Skp2 translocation is orchestrated in response to TGFβ is still unknown. Given the report of Skp2 regulation by Akt, it would be interesting to address the possible connection between Akt and the TGFβ-induced localization of Skp2.27,28 It would be helpful to explore how whether TGFβ signaling alters the activity of Akt to control Skp2 cytosol-nucleus trafficking. Ubiquitylation assay indicated that in the presence of TGFβ signaling, Skp2 translocates to the nucleus and is ubiquitylated by Cdh1/APC for degradation. Deletion of the nuclear localization signal (NLS) sequence impairs Skp2 nuclear translocation and inhibits its degradation, which in turn antagonizes growth inhibition by TGFβ. This work implicates that subcellular distribution of Skp2 is an important cellular marker that reflects the status of cell proliferation. Regulation of Skp2 translocation could be a valuable therapeutic strategy to control cell growth. Moreover, cytoplasmic Skp2 has been suggested to promote cell migration and thus has a potential function in tumor metastasis.27 Given the critical role of TGFβ in epithelial-mesenchymal transition (EMT), which correlates to increased invasiveness and metastasis, it would be important to investigate the possible link between TGFβ-induced translocation and degradation of Skp2 and cell motility in the context of TGFβ signaling.
Skp2ΔNLS (nuclear localization signal deleted Skp2) was constructed by PCR using following primers and then cloned into pCS2-HA or pREX-IRES-CD2, a mammalian expression vector:
5′-AAA ATC GAT ATG CAC GTA TTT AAA ACT CCC GGG CC-3′
5′-AAA GTC TTT GTC ACT CCC TTT GGG GCT CTC CGG GTG GCC CAG GTT-3′
5′-AAA GGG AGT GAC AAA GAC TTT GTG-3′
5′-TTG GCG CGC CTA GAC AAC TGG GCT TTT GCA GTG TC-3′.
pREX-IRES-CD2 is a gift from Dr. Xuedong Liu (University of Colorado-Boulder). pREX-HA-Skp2-IRES-CD2 and pREX-HA-Skp2ΔNLS-IRES-CD2 were generated by PCR using following primers:
5′-AAA AGT CGA CCC ACA GGA AGC ACC TCC AGG AG-3′
5′-TTG CGG CCG CTC ATA GAC AAC TGG GCT TTT G-3′.
Western blot analysis was performed using the anti-Skp2 (Santa Cruz, sc-7164), anti-p27 (Santa Cruz, sc-776), anti-Cdh1 (Abcam, DCS-266), anti-PCNA (Santa Cruz, sc-56), anti-HA (Santa Cruz, sc-7392), anti-Myc (Santa Cruz, sc-789), anti-Flag (Sigma, F3165), anti-SnoN (Cascade BioScience, ABM-3002), Ubiquitin (BD bioscience, 550944) and HRP-conjugated goat-anti-mouse (Promega, W4021) or anti-rabbit secondary antibody (Promega, W4011) with ECL detection kit (Amersham, RPN2106). Semi-quantification of data was performed using NIH image.
For immunoprecipitation assay, cell lysate was incubated with anti-Cdh1 antibody overnight at 4° on a rotator, following by addition of UltraLink Immobilized Protein A/G (Pierce, 53133). IP-complex was resolved by SDS-PAGE. Western blotting was conducted by using antibody against Skp2.
For growth inhibition assay, 5 × 103 Mv1Lu cells were incubated with various concentrations of TGFβ1 (R&D, 40-B) for 3–4 days. The growth of cells was determined by counting compared with that of unstimulated cells.13
Cells were washed and fixed in 4% paraformaldehyde solution for 10 min at room temperature and permeabilized with 0.1% Triton X-100 in PBS. Cells with and without stimulation by TGFβ1 were incubated with a rabbit anti-Skp2 and a mouse anti-Cdh1 antibodies followed by incubating with FITC conjugated anti-mouse (Jackson Lab, 115-095-146) and Texas Red conjugated anti-rabbit (Jackson Lab, 111-075-045) secondary antibodies.
We thank Drs. X. Liu, W. Malcolm, W. Kaelin, A. Weissman and D. Zhang for cDNA clones. We are grateful to members of our laboratory for critical reading of the manuscript. This work is supported by NIH grants CA115943. Y. Wan is a scholar of American Cancer Society and V Cancer Research Foundation.