F-box protein Fbxl18 mediates polyubiquitylation and proteasomal degradation of the pro-apoptotic SCF subunit Fbxl7

Fbxl7, a subunit of the SCF (Skp-Cul1-F-box protein) complex induces mitotic arrest in cells; however, molecular factors that control its cellular abundance remain largely unknown. Here, we identified that an orphan F-box protein, Fbxl18, targets Fbxl7 for its polyubiquitylation and proteasomal degradation. Lys 109 within Fbxl7 is an essential acceptor site for ubiquitin conjugation by Fbxl18. An FQ motif within Fbxl7 serves as a molecular recognition site for Fbxl18 interaction. Ectopically expressed Fbxl7 induces apoptosis in Hela cells, an effect profoundly accentuated after cellular depletion of Fbxl18 protein or expression of Fbxl7 plasmids encoding mutations at either Lys 109 or within the FQ motif. Ectopic expression of Fbxl18 plasmid-limited apoptosis caused by overexpressed Fbxl7 plasmid. Thus, Fbxl18 regulates apoptosis by mediating ubiquitin-dependent proteasomal degradation of the pro-apoptotic protein Fbxl7 that may impact cellular processes involved in cell cycle progression.

Y Liu 1 , T Lear 1 , Y Zhao 1 , J Zhao 1 , C Zou 1 , BB Chen 1 and RK Mallampalli* , 1,2 Fbxl7, a subunit of the SCF (Skp-Cul1-F-box protein) complex induces mitotic arrest in cells; however, molecular factors that control its cellular abundance remain largely unknown. Here, we identified that an orphan F-box protein, Fbxl18, targets Fbxl7 for its polyubiquitylation and proteasomal degradation. Lys 109 within Fbxl7 is an essential acceptor site for ubiquitin conjugation by Fbxl18. An FQ motif within Fbxl7 serves as a molecular recognition site for Fbxl18 interaction. Ectopically expressed Fbxl7 induces apoptosis in Hela cells, an effect profoundly accentuated after cellular depletion of Fbxl18 protein or expression of Fbxl7 plasmids encoding mutations at either Lys 109 or within the FQ motif. Ectopic expression of Fbxl18 plasmid-limited apoptosis caused by overexpressed Fbxl7 plasmid. Thus, Fbxl18 regulates apoptosis by mediating ubiquitindependent proteasomal degradation of the pro-apoptotic protein Fbxl7 that may impact cellular processes involved in cell cycle progression. It is estimated 480% proteins undergo ubiquitin-dependent degradation either by the proteasome or lysosome. 1 The ubiquitylation of a target protein occurs in a carefully orchestrated manner by an enzymatic cascade involving an E1 ubiquitin-activating enzyme, an E2-conjugating enzyme and an E3-ubiquitin ligase. 2,3 Among hundreds of E3 ligases, the SCF superfamily has emerged as a class of E3 proteins that modulates diverse biological processes. 4 The SCF complex contains a catalytic core including Skp1, Cul1 and an F-box protein. [5][6][7] The carboxyl terminus of Cul1 recruits the small RING protein Rbx1, which bridges the E2 enzyme to the E3 ligase, and the amino terminus of Cul1 binds to Skp1 and a variable F-box protein as a substrate receptor. 8 Based on the pivotal roles of SCF E3 ligases in cellular function through the elimination of substrates, SCF components have recently attracted significant attention on cancer therapeutic drug development and potential to modulate inflammation and innate immunity. 9,12 F-box proteins contain two domains: an NH 2 -terminal F-box motif and a carboxyl terminal leucine-rich repeat (LRR) motif or tryptophan-aspartic acid (WD) repeat motif, which are identified within two subclasses: Fbxl or Fbxw proteins, respectively 13 and. 14 The third subclass of F-box proteins are recognized as Fbxo (F-box only) proteins contain no identifiable LRR or WD motifs. The F-box motif binds to Skp1, whereas the LRR/WD motif is a receptor-like motif that recognizes diverse substrates for ubiquitin ligation and subsequent degradation. 15 With the increasing characterization of F-box proteins and its cognate substrates, this has led to an improved understanding of how SCF complexes have multiple roles in cell cycle progression, gene transcription, circadian oscillation and inflammation. 16 However, with the exception of a few SCF components, the substrates for the majority of~70 F-box proteins remain largely unknown. In addition the mechanisms that regulate the abundance of the majority of F-box proteins remain unclear and this represents an area of active investigation. It is known that some components within the SCF apparatus are targeted by the ubiquitin ligase anaphase-promoting complex or by autoubiquitylation within their own SCF complex. 17,18 Recent studies also show that one F-box protein may target another F-box protein for ubiquitylation to mediate protein degradation thereby impairing the SCF activity. 12 We have previously characterized the role of a relatively new F-box protein, Fbxl7, on cell cycle progression. Fbxl7 within the SCF complex targets Aurora A for polyubiquitylation and proteasomal degradation leading to mitotic cell cycle arrest. 19 Additionally, Fbxl7 was reported to be associated with an ovarian cancer gene and its expression is linked to steroid responsiveness in human subjects with asthma. 20,21 Although these data suggest a role for Fbxl7 in controlling cell proliferative activity and viability, further studies investigating both its biological role and molecular expression are needed. Here we identified that Fbxl7 is itself ubiquitylated and degraded via the proteasome. Interestingly, we identified that an orphan F-box protein, Fbxl18, targets Fbxl7 for its ubiquitylation using distinct molecular signatures that leads to its disposal in cells. Ectopically expressed Fbxl7 displayed pro-apoptotic activity in Hela cells, an effect abrogated by co-expression of Fbxl18. These observations suggest a molecular interplay between two SCF components that govern availability of a biologically important pro-apoptotic protein.

Results
Fbxl7 undergoes proteasomal degradation. High levels of Fbxl7 in cells appear to critically regulate mitotic arrest; however, the lifespan of this F-box protein in cells is unknown. We first measured the protein stability of endogenous Fbxl7 in MLE12 cells. As shown in Figures 1a and b, cells treated with the protein synthesis inhibitor cycloheximide (CHX) indicated that the half-life of Fbxl7 is~1 h. To identify the degradation pathway involved in Fbxl7 decay, MLE12 cells were also treated with the proteasome inhibitor (MG132) or a lysosome inhibitor (leupeptin) prior to CHX treatment. MG132 treatment, instead of leupeptin, resulted in accumulation of immunoreactive Fbxl7 in cells, which suggests that the proteasome, rather than the lysosome is involved in degradation of Fbxl7.
Fbxl7 is polyubiquitylated. To test if Fbxl7 is modified by ubiquitylation, we examined steady-state levels of Fbxl7 protein after overexpression of cells with a plasmid encoding ubiquitin. The levels of endogenous Fbxl7 decreased after expression of increasing amounts of ectopically expressed HA-tagged ubiquitin (Figure 1c). In co-immunoprecipitation (co-i.p.) studies, we also detected an interaction between endogenous Fbxl7 and ubiquitin (Figure 1d). We further confirmed the binding of Fbxl7 to ubiquitin after expression in cells with Fbxl7-V5 and HA-ubiquitin (HA-Ub) plasmids under MG132 treatment (Figure 1e). Collectively, these results indicate that Fbxl7 degradation is mediated by the ubiquitinproteasome system.
Fbxl18 targets Fbxl7 for its ubiquitylation and degradation. Recent studies indicate that some F-box proteins target others for their ubiquitylation and degradation. 12 Thus, we investigated the interaction between Fbxl7 and randomly chosen F-box proteins by co-immunoprecipitation. Fbxl7 specifically bound to Fbxl18 (Supplementary Figure 1A). We further screened ability of ectopically expressed plasmids encoding~50 F-box proteins from L, W and O families that might mediate Fbxl7 degradation. As shown in Figure 2a, randomly chosen V5-tagged F-box proteins in L family were ectopically expressed in MLE12 cells. Interestingly, only Fbxl18 overexpression decreased endogenous Fbxl7 protein levels. Additionally, we observed that overexpressed F-box proteins from W and O families displayed no effect on endogenous Fbxl7 protein abundance, as shown in Supplementary Figures 1B and C, respectively. We further observed that overexpression of Fbxl18 plasmid decreased endogenous Fbxl7 protein levels in a dose-dependent manner ( Figure 2b). However, transfection of cells with increasing amounts of Fbxl16 plasmid had no effect on endogenous Fbxl7 levels. Particularly, MG132 treatment abolished the Fbxl7 decrease induced by overexpression of Fbxl18 ( Figure 2c). In co-i.p. studies we detected the association of endogenous Fbxl18 with Fbxl7 ( Figure 2d) in Hela cells. We further confirmed that Fbxl18 mediated the polyubiquitylation of Fbxl7 using in vitro ubiquitylation assays, and we also excluded the possibility that Fbxl7 might display autoubiquitylation activity ( Figure 2e). Additionally, we investigated Fbxl18-mediated polyubiquitylation using a neddylation inhibitor MLN4924 (Figure 2f). Neddylation-deficient Cul1-Rbx1 largely decreased the polyubiquitylation of Fbxl7, in support that Fbxl7 polyubiquitylation is dependent on a   (Figure 2g). In separate studies knockdown of Fbxl18 with shRNA-stabilized Fbxl7 protein levels in a time-dependent manner (Figures 2h and i). These results indicate that Fbxl18 mediates Fbxl7 polyubiquitylation and proteasomal degradation.
Lys 109 is an Fbxl7 ubiquitin acceptor site. To localize the ubiquitin receptor site within Fbxl7, we ectopically coexpressed individual Fbxl7 plasmids encoding lysine point mutations with an Fbxl18 plasmid in MLE12 cells. Mutants whose protein levels decreased with overexpressed Fbxl18 were excluded from further analysis (data not shown). We then assessed protein stability of the remaining mutants by exposure to CHX treatment after Fbxl7 cellular expression.
We identified that only transfection of cells with a Lys 109E mutant stabilized the decay of Fbxl7, compared to the wildtype or other lysine mutants as shown in Figure 3a. Further, in MLE12 cells co-transfected with Fbxl18 and Fbxl7 plasmids, overexpressed Fbxl18 plasmid dose-dependently decreased wild-type Fbxl7 levels, but had no effect on Fbxl7 K109E protein levels ( Figure 3b). To exclude an amino acid charge effect on protein stability, we further compared the protein turnover rate after expression in cells with plasmids encoding additional Lys 109 substitutions. As indicated in Figures 3c and d, expression in cells of plasmids encoding three distinct Lys 109 point mutants resulted in stabilization of Fbxl7 proteins levels compared with cellular expression of wild-type Fbxl7 plasmid. Here, the addition of a tag to the protein extended the lifespan of wild-type Fbxl7 versus endogenous Fbxl7 (Figure 1). In the in vitro ubiquitylation assays, the expressed Lys 109 point mutants displayed resistance to SCF Fbxl18 -induced polyubiquitylation versus wild-type Fbxl7 (Figure 3e). We also compared the polyubiquitylation levels between wild-type Fbxl7 and the K109R mutant through ubiquitin pull downs in MLE12 cells. The lysine 109 mutation within Fbxl7 dramatically decreased Fbxl7 polyubiquitylation (Figure 3f). These results demonstrate that the ubiquitin acceptor site Lys 109 is essential for Fbxl7 decay through ubiquitin-proteasomal processing.  for an additional 2 h before boiling in PBS containing 2% SDS. The lysates were next diluted in TBS to a final 0.2% SDS concentration before pull down using a rabbit ubiquitin antibody. The co-i.p. products were processed for immunoblotting with mouse V5 antibody and mouse ubiquitin antibodies, respectively Fbxl18 regulates Fbxl7-induced apoptosis Y Liu et al Fbxl7, we constructed a series of NH 2 -terminal and C-terminal truncated Fbxl7 expression plasmids, as schemed in Figure 4a. In Hela cells, endogenous Fbxl18 interacted with all the Fbxl7 C-terminal deletion mutants. However, loss of the NH 2 -terminal region containing the F-box domain largely abolished the interaction between Fbxl7 and Fbxl18 (Figure 4b). We then focused on the Fbxl7 NH 2 -terminal region and further localized the Fbxl18 interacting region within Fbxl7 via co-immunoprecipitation (Figures 4c and e). Figures 4d and e demonstrate that an amino-terminal region spanning residues 63-69 harbors a putative Fbxl18 binding site within Fbxl7. Indeed, by generating and expressing plasmids encoding specific Fbxl7 point mutants in cells followed by co.ip., we observed that an FQ motif within Fbxl7 was essential for Fbxl18 interaction with Fbxl7 (Figure 4e). To further confirm that the FQ motif is the interaction site between Fbxl18 and Fbxl7, we employed a biotin-labeled peptide binding assay and detected in vitro synthesized V5-tagged Fbxl18 bound to wild-type Fbxl7 peptides.
However, a Fbxl7 FQ-AA mutant largely exhibited a decrease in its association with Fbxl18 ( Figure 4f). We also compared Fbxl7 protein stability after expression in cells of a plasmid encoding either wild-type or an FQ mutant. As shown in Figure 4g, cellular expression of the FQ mutant plasmid resulted in an Fbxl7 variant that exhibited extended lifespan in cells versus the wild-type Fbxl7 protein. Moreover, similiar to wild-type Fbxl7, this FQ mutant displayed ability to retain association with other components within the SCF complex ( Figure 4h) and to mediate degradation of a known substrate, Aurora A, for proteasomal degradation (Figure 4i). These results strongly suggest that an Fbxl7 FQ motif functions as the molecular recognition site for Fbxl18 binding. The binding capacity of Fbxl7 through this motif does not impact its effect on substrate degradation.   19 Here, we observed that overexpressed Fbxl7 induces apoptosis. The expression of Fbxl7 plasmid in cells led to appearance of a larger percentage of apoptotic cells in both a dose and timedependent manner, as determined by propidium iodine staining compared with the cells transfected with a control vector (Figures 5a and b). These data indicate that Fbxl7 is pro-apoptotic.
Fbxl18 counteracts Fbxl7-induced apoptosis. As shown above, overexpression of Fbxl7 plasmid causes apoptosis. We next investigated if Fbxl18 differentially modulates Fbxl7induced apoptosis. Cellular depletion of Fbxl18 with three different Fbxl18 shRNAs caused increased apoptosis, compared with effects of scrambled shRNA-treated cells (Figure 6a). However, the apoptosis induced by knockdown Fbxl18 was largely rescued by the additional knockdown of Fbxl7 ( Figure 6b). Further, as shown in Figure 6c, double transfection of the Fbxl7 FQ mutant and Fbxl18 plasmids showed a impaired cellular proliferation, compared with the cells transfected with Fbxl7 wild-type and Fbxl18 plasmids or a control vector. These data suggest that the Fbxl7 docking site mutant protein that lacks ability to interact with Fbxl18 harbors an extended lifespan in cells to induce apoptosis. Further, Fbxl18 counteracts the pro-apoptotic activity of Fbxl7 in a biological model of cell injury.

Discussion
Fbxl7 is a highly evolutionally conserved protein sharing a 98% sequence identity between mouse and humans, and 495% identity among other species. However, Fbxl7 is a relatively uncharacterized protein with regard to its biological role and molecular regulation in cells. Whether Fbxl7 is an oncoprotein or a component within a complex that triggers mitotic arrest leading to cell death requires further study. Our data demonstrate potent activity of this SCF component as it is sufficient to elicit apoptosis in some cells. This effect is relevant as high-level Fbxl7 protein expression in cells, left unchecked, could promote sustained tissue injury through activation of apoptotic programs. This is important given that cell proliferative activity must overcome the activities of proapoptotic factors to ensure optimal healing during tissue injury.
Here, we demonstrate that Fbxl7 protein levels are controlled by the actions of the SCF Fbxl18 complex, perhaps serving as a feedback control mechanism to limit high-level expression of this pro-apoptotic factor. We have demonstrated that Fbxl7 harbors a molecular signature that recruits Fbxl18, identified the acceptor site for Fbxl18 activity within its substrate, and present data suggesting that levels of Fbxl18 affect the ability of cells to remain viable from Fbxl7-induced apoptosis. The data are also the first to characterize the molecular behavior of Fbxl18 in cells and raise the possibility for molecular targeting of an Fbxl18-|Fbxl7 → apoptosis pathway. Fbxl7 is a relatively short-lived protein (t 1/2~1 h) through SCF Fbxl18 -dependent polyubiquitylation and subsequent proteasomal degradation. However, 1 h after cycloheximide treatment, Fbxl7 protein levels remains stable in cells (Figures 1a and b). One possibility could be that Fbxl7 interacts with ubiquitin-specific peptidases. For example, Fbxl7 directly associates to USP1 and USP45, and indirectly is linked to USP12. 22 These deubiquitylation enzymes could counteract SCF Fbxl18 to remove ubiquitin from Fbxl7 to prevent its degradation. More importantly, the Fbxl7 steady-state levels remain detectable with some degree of stabilization following its initial rapid decay that appears to be caused by the accelerated degradation of Fbxl18 based on our observations (data not shown). Although the half-life of endogenous Fbxl7 is short, the addition of a C-terminal V5 tag extended its half-life to~6 h. This might result from a conformation change induced by the tag or that exogenous Fbxl7 exceeded the endogenous SCF Fbxl18 ubiquitylation machinery capacity to impair its function. The underlying mechanism for this disparity is under further investigation. The molecular factors that regulate Fbxl18 levels and potential sensing of Fbxl7 abundance also remain obscure. Fbxl18 may be prone to autoubiquitylation or other modifications and subsequent degradation. Our data suggest that Fbxl18 levels rapidly fluctuate in cells but when elevated, they are linked to stability of Fbxl7 to regulate the apoptotic program. We speculate that there may exist a negative feedback loop driven by Fbxl18 that could be essential to keep a dynamic equilibrium of Fbxl7, which could be critical for execution of vital roles by multiple substrates downstream Fbxl7. The identification of these Fbxl7 substrates in addition to Aurora A also requires further study.
The interaction of F-box protein and its substrates must be tightly controlled. In response to a stimuli, F-box proteins must recruit substrates rapidly and specifically to the catalytic core for ubiquitylation and degradation. For example, phosphorylation of the substrate resulting in a phosphodegron is one of the most common recognition signals for F-box proteins, which either accelerates or inhibits the interaction of F-box protein and substrates. In response to a specific stimuli, kinases such as glycogen synthase kinase 3β regulate biological function by manipulating protein stability through phosphorylating substrates. 23,24 Recently, acetylation is also considered as an important post-translational modification to regulate protein stability. 25 Additionally, other interaction factors also have a vital role in mediating F-box protein to recognize their substrates. 26,27 Interestingly, another F-box protein, Fbxl2 that impacts lipogenesis, innate immune responses and cell cycle progression targets its substrates through an IQ motif on its substrates. 28 In this study, we identified that a similar FQ motif within Fbxl7 is the major docking site for recognition by Fbxl18. We cannot completely exclude the possibility that another minor domain within Fbxl7 might also serve as a recognition signal to recruit Fbxl18 because a series of N-terminal deletions do not totally eliminate the interaction between Fbxl18 and Fbxl7 (Figures 4b-f), and C-terminal truncated Fbxl7 mutants showed slightly different binding intensities to Fbxl18. Whether this FQ motif-based interaction mechanism is unique or undergoes any modification responding to stimuli, is currently under further investigation. The results suggest that some members of the F-box family harboring LRR domains may be recruited to their substrates through these signatures.
The ability of Fbxl18 to recruit pro-apoptotic Fbxl7 for ubiquitylation and proteasomal degradation may be of fundamental importance in human disease. For example, aside from neoplasia, several disorders including emphysema, traumatic injury and aging are characterized by an imbalance between cell proliferative activities and apoptosis. The ability of Fbxl18 to protect cells from apoptosis induced by Fbxl7 might provide an opportunity to design small molecule Fbxl18 agonists to enhance tissue repair and regeneration, or alternatively design novel therapeutics that impair Fbxl18:Fbxl7 interaction to stimulate apoptosis in tumors. These data are, however, too preliminary to conclude that Fbxl18 is an anti-apoptotic molecule until larger numbers of Fbxl18-substrate pairs are characterized in multiple systems. We are currently investigating the identities of additional SCF Fbxl18 substrates to better interpret its role in apoptosis. Additionally, the Fbxl7 sequence is highly conserved among species. Specifically, Fbxl7 in mammals is homologous to Grr1 in yeast Saccharomyces cerevisiae. Grr1 is among the beststudied F-box proteins in yeast, which has a vital role in cell cycle regulation by recruiting Cln2. 29 Besides Aurora A, we expect to identify additional kinases or cell cycle regulators under regulation of Fbxl7 to further elucidate the pro-apoptotic behavior of Fbxl7. Cloning and mutagenesis. PCR-based approaches were applied to clone different F-box proteins into pcDNA3.1D/v5-his (Invitrogen) for expression in mammalian cells using appropriate primers. All mutants were constructed using PCR-based approaches or site-directed mutagenesis (Agilent).
Transfection. All plasmids were delivered into MLE12 cells or BEAS-2B cells using nucleofection or FuGENE HD (Hela cells), following the manufacturers' protocols. Cellular expression of GFP-tagged plasmids using these methods was achieved at 490% efficiency.
Immunoblotting and Immunoprecipitation. Cell lysates (normalized to total protein concentration) were subjected to SDS-PAGE, electrotransferred to membranes, and immunoblotted. For ubiquitin pull down, the freshly collected cells in PBS containing 2% SDS were boiled at 100°C for 10 min. The lysates were diluted with TBS to a final concentration with 0.2% SDS containing 5 μM ubiquitin aldehyde. 1 mg total cellular proteins were immunoprecipitated with 2 μg ubiquitin rabbit antibodies at 4°C for 3-4 h followed by the addition of 30 μl of protein A/Gagarose for an additional 1 h at 4°C. The precipitated complex was washed three times with 0.5% Triton X-100 in PBS and analyzed by immunoblotting with an enhanced ECL system. For other immunoprecipitations, 1-mg cell lysates in PBS containing 0.5% Triton X-100 were incubated with 2-μg-specific primary antibodies for 3-4 h at 4°C followed by the addition of protein A/G-agarose and three washes with lysis buffer before immunoblotting.
Apoptosis analysis. Transfected cells were incubated with FITC-Annexin V and propidium iodide for 15 min following manufacturer's protocols (Biolegend, San Diego, CA, USA). FACS samples were analyzed with the AccuriC6 system with De Novo Software. For proliferation assays, Hela cells were transfected with indicated plasmids. Cells were cultured in 35 mm dishes for up to 48 h; at indicated time points, cells were collected and stained with trypan blue. Viable cells were then counted and quantified.
Statistical analysis. Statistical comparisons were performed using an ANOVA or an unpaired t-test with Po0.05 indicative of significance.