FBXO32 Targets c-Myc for Proteasomal Degradation and Inhibits c-Myc Activity*

Background: FBXO32 is an E3 ubiquitin ligase that plays important roles in tumorigenesis and muscle atrophy. Results: c-Myc was found to be a target of FBXO32 for proteasomal degradation. Conclusion: FBXO32 targets Lys-326 of c-Myc to form polyubiquitin chains, resulting in inhibition of cell proliferation. Significance: FBXO32 may mediate c-Myc proteasomal degradation. FBXO32 (MAFbx/Atrogin-1) is an E3 ubiquitin ligase that is markedly up-regulated in muscle atrophy. Although some data indicate that FBXO32 may play an important role in tumorigenesis, the molecular mechanism of FBXO32 in tumorigenesis has been poorly understood. Here, we present evidence that FBXO32 targets the oncogenic protein c-Myc for ubiquitination and degradation through the proteasome pathway. Phosphorylation of c-Myc at Thr-58 and Ser-62 is dispensable for FBXO32 to induce c-Myc degradation. Mutation of the lysine 326 in c-Myc reduces c-Myc ubiquitination and prevents the c-Myc degradation induced by FBXO32. Furthermore, overexpression of FBXO32 suppresses c-Myc activity and inhibits cell growth, but knockdown of FBXO32 enhances c-Myc activity and promotes cell growth. Finally, we show that FBXO32 is a direct downstream target of c-Myc, highlighting a negative feedback regulation loop between c-Myc and FBXO32. Thus, FBXO32 may function by targeting c-Myc. This work explains the function of FBXO32 and highlights its mechanisms in tumorigenesis.

FBXO32 (MAFbx/Atrogin-1) is an E3 ubiquitin ligase that is markedly up-regulated in muscle atrophy. Although some data indicate that FBXO32 may play an important role in tumorigenesis, the molecular mechanism of FBXO32 in tumorigenesis has been poorly understood. Here, we present evidence that FBXO32 targets the oncogenic protein c-Myc for ubiquitination and degradation through the proteasome pathway. Phosphorylation of c-Myc at Thr-58 and Ser-62 is dispensable for FBXO32 to induce c-Myc degradation. Mutation of the lysine 326 in c-Myc reduces c-Myc ubiquitination and prevents the c-Myc degradation induced by FBXO32. Furthermore, overexpression of FBXO32 suppresses c-Myc activity and inhibits cell growth, but knockdown of FBXO32 enhances c-Myc activity and promotes cell growth. Finally, we show that FBXO32 is a direct downstream target of c-Myc, highlighting a negative feedback regulation loop between c-Myc and FBXO32. Thus, FBXO32 may function by targeting c-Myc. This work explains the function of FBXO32 and highlights its mechanisms in tumorigenesis.
FBXO32 (also known as MAFbx or Atrogin-1) was originally identified as a muscle-specific gene required for muscle atrophy (1,2). FBXO32 was designated as a muscle-specific E3 ubiquitin ligase because it contains the F-box domain. This is a characteristic of E3 ligases that function as one component of a SCF (Skp1, Cullin, F-box protein) ubiquitin ligase complex (1,2). FBXO32 lacks leucine-rich repeats or WD40 repeats but it does contain a class II PDZ domain (2), which interacts with specific sequences at the carboxyl terminus of the target proteins (3,4). In addition, FBXO32 also contains two nuclear localization signals, which suggests that it may target transcription factors or other nuclear proteins for ubiquitination (5).
As an ubiquitin E3 ligase, FBXO32 has been shown to target several proteins for proteasomal degradation (6 -8). The two most widely known targets of FBXO32 in skeletal muscle are the initiation factor, eIF3-f, and the myogenic regulatory factor, MyoD. In addition to targeting the proteins mediating FBXO32 function in muscle atrophy, FBXO32 also targets MKPK phosphatase-1 for proteasomal degradation and is probably involved in ischemia/reperfusion-induced cardiomyocyte apoptosis (8).
Some evidence suggests that FBXO32 might also play an important role in tumorigenesis. The expression level of FBXO32 is closely correlated with the methylation status of the FBXO32 promoter in ovarian cancer cell lines (9). Restoration of FBXO32 in ovarian cancer cells inhibits colony formation in vitro and xenograft tumor growth in athymic nude mice. Moreover, patients with higher FBXO32 promoter methylation tend to have shorter progression-free survival. This suggests a tumor-suppressive role of FBXO32 (9). In addition, the FBXO32 expression is also decreased in esophageal squamous cell carcinoma (10). EZH2 supports the survival of alveolar rhabdomyosarcoma by repressing FBXO32 (11), but, up-regulation of FBXO32 is a hallmark of cancer cachexia caused by muscle wasting (12,13). Therefore, the role of FBXO32 in tumorigenesis is still unclear, and the underlying mechanism is poorly understood.
c-Myc is a short-lived protein and a classic oncogene regulated at multiple steps. One of the most prominent mechanisms for c-Myc degradation in cells is through the ubiquitin-proteasome pathway (14). As a component of the RING finger domain, ubiquitin ligase complex (15,16), Fbw7 is the best studied SCF-type E3 ubiquitin ligase for mediating c-Myc degradation. The FBW7 recognizes phosphorylated c-Myc at Thr-58, which is mediated by glycogen synthase kinase 3 (Gsk3) (15,16). Another RING finger E3 ligase, Skp2, recognizes a conserved sequence element in the amino terminus of c-Myc (MBII) as well as the HLH-LZ motifs (amino acids 367-439). It promotes its polyubiquitination and degradation, resulting in the enhancement of c-Myc transcriptional activity (17,18). The third RING finger E3 ligase, ␤-TrCP, binds to the amino terminus of c-Myc and uses the UbcH5 ubiquitin-conjugating enzyme (E2) to form heterotypic polyubiquitin chains on c-Myc. This enhances c-Myc stability (19). Importantly, the expression pattern of FBXO32 is inversely correlated with c-Myc expression during starvation treatment (2). Serum stimulation induces expression of c-Myc immediately (20,21), but removal of growth factors at any point in the cell cycle results in down-regulation of the c-Myc (22,23). In contrast, IGF-1 inhibits the transcription of FBXO32 through the PI3K/Akt/FOXO pathway and blocks dexamethasone-induced atrophy (24); food deprivation increases FBXO32 expression and results in rapid muscle wasting (2). Moreover, activation of FOXO3a leads to a considerable reduction in c-Myc (25,26).
Given that FBXO32 is a direct target of FOXO3a (27), we sought to determine whether FBXO32 participates in the degradation of c-Myc. In this study, we identify c-Myc as a substrate of FBXO32 E3 ubiquitin ligase. FBXO32 targets c-Myc for ubiquitination and degradation through the proteasome pathway. Mutation of the lysine 326 in c-Myc reduces c-Myc ubiquitination and prevents c-Myc degradation induced by FBXO32. Moreover, overexpression of FBXO32 suppresses c-Myc activity and inhibits cell growth. In addition, we reveal that FBXO32 is a direct downstream target of c-Myc.
Plasmid Constructs-The original wild-type c-Myc and its domain constructs were kindly provided by Stephen Hann. HA-tagged c-Myc(P57S) and c-Myc(T58A) were kindly provided by Scott Lowe. The original Gadd45␣ promoter luciferase reporter was provided by Linda Penn. The pGL3-E2F2 promoter luciferase reporter was constructed by PCR amplification (28). pGL3-E2F2-mut promoter luciferase reporter was constructed by mutating three E-boxes in the promoter (the sequence information will be provided upon requested). The FBXO32 coding sequence was amplified from cDNA derived from 293T cell total RNA by PCR with the primers 5Ј-CCCG-AATTCGGATGCCATTCCTCGGGCAGGAC-3Ј (forward) and 5Ј-CCCGGTACCTCAGAACTTGAACAAGTTGATA-AAG-3Ј (reverse) and then subcloned into pCMV-HA vector (Clontech). Phage-CMV-MCS-IZsGreen was used as a retroviral expression vector. FBXO32-WT, FBXO32-⌬F-box, c-Myc-WT, and c-Myc-K326R were subcloned into phage-CMV-MCS-IZsGreen and confirmed by sequencing.
Luciferase Reporter Assays-HEK293T or HCT116 cells were seeded in 24-well plates and transfected with the indicated plasmids by VigoFect (Vigorous Biotech, Beijing). The pRL-SV40 luciferase reporter (Promega) was included in all transfections for normalization. Luciferase activities were measured 24 h after transfection using the Dual-Luciferase reporter assay system (Promega). Data were normalized to Renilla luciferase. Data are reported as the means Ϯ S.E. of three independent experiments performed in triplicate. The statistical analysis was performed using GraphPad Prism version 5 (unpaired Student's t test) (GraphPad Software Inc.).
Colony Formation Assays-The stable transfected SKOV3 cells via lentivirus infection were seeded in 6-well plates at 4 ϫ 10 3 cells/well and cultured in DMEM containing 10% fetal bovine serum for 14 days. The colonies were fixed and stained with 0.4% crystal violet in 50% methanol for 30 min. Dishes were rinsed with water and left for drying at room temperature. Colonies were photographed with a stereomicroscope and counted by ImageJ software. Only the colonies containing more than 50 cells were set for counting. Data are reported as the means Ϯ S.E. of three independent experiments performed in triplicate. The statistical analysis was performed using GraphPad Prism version 5 (unpaired Student's t test) (GraphPad Software Inc.).
In Vivo Ubiquitination Assays-HEK293T cells were cotransfected with the indicated plasmids using Vigofect. After transfection for 24 h, the cells were collected, lysed, and subjected to immunoprecipitation by Ni 2ϩ -nitrilotriacetic acid beads (Novagen) and then examined by Western blotting using anti-HA antibody. The ubiquitin mutants have been described previously (30).
Lentivirus Package-Lentiviruses for gene overexpression were generated by transfecting HEK293T cells with combinations of transducing vector and two packaging vectors, PSPAX2 and pMD2.G. For lentiviral shRNA production, transducing vector was co-transfected with three packaging vectors, PMDLg/ pRRE, VSVG, and RSV-Rev, in HEK293T cells. After transfection for 6 h, the medium was replaced with fresh DMEM with 10% FBS. 48 h later, the medium containing the lentivirus particles was harvested, filtered, and transduced into target cells or frozen at Ϫ70°C for subsequent using. Polybrene (8 g/ml) was added to the medium for improving infection efficiency.
Co-immunoprecipitation and Western Blot Analysis-For Western blot analysis and co-immunoprecipitation, the experimental procedures have been described previously (31). Anti-HA antibody and anti-FLAG antibody-conjugated agarose beads were purchased from Sigma-Aldrich. Protein A/G-Sepharose beads were purchased from GE Healthcare. Glutathione S-transferase (GST)-Bind resin was purchased from Novogen. For endogenous immunoprecipitation, the mouse leg muscle was used for extracting protein. The Fuji Film LAS4000 miniluminescent image analyzer system was used to photograph the blots. Multi Gauge version 3.0 was used for quantifying the protein levels based on the band density obtained in Western blot analysis.
Statistical Analysis-Data are presented as mean Ϯ S.E. of three independent experiments performed in triplicate; the statistical analysis was performed using GraphPad Prism version 5.0 (unpaired Student's t tests).  Fig. 1B). To verify the specificity of FBXO32-shRNA-2-mediated FBXO32 knockdown, we performed a rescue experiment. As shown in Fig. 1C, overexpression of FBXO32-shRNA-2-resistant FBXO32 in HCT116 cells neutralized the effect of FBXO32-shRNA-2-mediated c-Myc up-regulation. To determine whether FBXO32-mediated c-Myc reduction was due to FBXO32-induced c-Myc instability, we added the new protein synthesis inhibitor, cycloheximide (CHX) 2 and examined the protein levels at different time points. Without ectopically expressed FBXO32, the overexpressed c-Myc gradually degraded (left panel in Fig. 1D). However, co-expression of wild-type FBXO32 caused c-Myc to degrade more rapidly (middle panel in Fig. 1D). In contrast, co-expression of the F-box domain-deleted mutant of FBXO32 obviously prolonged the half-life of c-Myc (right panel in Fig. 1D). It appears that the mutant, FBXO32-⌬F-box, exhibits the dominant negative form activity in mediating c-Myc degradation. The protein levels of c-Myc were further quantified (Fig. 1D). To further evaluate the data, we knocked down endogenous FBXO32 by transiently transfected pSuper-FBXO32-shRNA-1 and pSuper-FBXO32-shRNA-1 in HCT116 cells and examined the protein level at different time points in the presence of CHX (50 g/ml). Versus the control, which was transiently transfected with pSuper-GFP-shRNA (left panel in Fig. 1E), knockdown of FBXO32 prolonged the half-life of endogenous c-Myc (middle and right panels in Fig. 1E). This is in contrast to what was seen in overexpression of FBXO32 (Fig. 1E). The c-Myc levels were then further quantified (Fig. 1E).

FBXO32 Promotes c-Myc
It is of note that anti-c-Myc antibodies from different companies could detect different expression patterns of endogenous c-Myc. The anti-c-Myc antibody from ABclonal could only detect one obvious band of endogenous c-Myc, but the antibody from Cell Signaling could detect two bands of endogenous c-Myc. Regardless of the number of bands, the tendency of c-Myc expression to be affected by knockdown of FBXO32 remained consistent ( Fig. 1, B, C, and E). Taken together, these observations suggest that FBXO32 induces c-Myc degradation and that the F-box domain is required for FBXO32 to mediate c-Myc degradation.
Phosphorylation of c-Myc at Thr-58 and Ser-62 Is Dispensable for FBXO32-induced c-Myc Degradation-It has been reported that FBW7, an F-box E3 ligase, destabilizes c-Myc, but the FBW7 mutant lacking the F-box domain delayed it (15,16), which is similar to what was observed for FBXO32. Of note, the turnover of c-Myc by FBW7 is largely dependent on the phosphorylation of Thr-58 and Ser-62 in MBI (15,16). In addition, a common c-Myc mutant, P57S, in Burkitt lymphoma is able to abrogate Thr-58 phosphorylation (32). To determine whether FBXO32 also has functions similar to that of FBW7 on these 2 The abbreviations used are: CHX, cycloheximide; K8R,

FBXO32 Targets c-Myc for Degradation
mutants, we examined the effect of FBXO32 on the c-Myc mutants T58A, S62A, T58A/S62A, and P57S as well as the constitutive phosphorylation mutant S62E. As shown in Fig. 2, A and B, co-expression of HA-FBXO32 promotes degradation in all of these mutants. In addition to these three sites, four other sites (Ser-71, Tyr-74, Thr-358, and Ser-373) in c-Myc have also been reported to be phosphorylation sites (33)(34)(35)(36). We found that HA-FBXO32 also promotes degradation in all four mutants (S71A, Y74A, T358A, and S373A) (Fig. 2C). These observations suggest that the turnover of c-Myc by FBXO32 is neither dependent on phosphorylation of Thr-58 and Ser-62 in MB1 nor dependent on phosphorylation of other sites, which is in contrast to FBW7 (15,16). FBXO32 Interacts with c-Myc-To determine the mechanisms of FBXO32-mediated c-Myc degradation, we examined whether FBXO32 could interact with c-Myc. As shown in Fig.  3C, ectopically expressed HA-c-Myc could pull down ectopically expressed FLAG-FBXO32 in HEK293T cells by immunoprecipitation assays using anti-HA-conjugated agarose beads (Fig. 3A). This interaction was endogenous and direct because a polyclonal anti-c-Myc antibody (A0309, ABclonal) pulled down endogenous FBXO32 in mouse muscle lysates (Fig. 3B), and GST-FBXO32 expressed in Escherichia coli pulled down His-c-Myc in E. coli (Fig. 3C). In addition, the c-Myc mutants T58A, S62A, and T58A/S62A could interact with FBXO32 when they were overexpressed in HEK293T cells (Fig. 3D).
We performed domain mapping to further study the mechanism of FBXO32-mediated c-Myc degradation. Whereas FLAG-tagged FBXO32 could be effectively pulled down by the truncated c-Myc mutants, 1-144 and 1-367 (Fig. 3, E and F), the other two truncated c-Myc mutants (168 -367 and 168 -439) only weakly pulled down FLAG-tagged FBXO32 (Fig. 3, E  and F). The c-Myc truncated mutant (mutant 1-144) contains two conserved domains, MBI and MBII (Fig. 3E) (37). As reported, MBI is required for FBW7 interactions (15), whereas MBII is required for Skp2 interactions (18). To determine whether MBI or MBII is required for FBXO32 interaction, we performed fine domain mapping. Deletion of MBI or MBII did not impact the interaction between FBXO32 and the c-Myc mutant (mutant 1-144; Fig. 3G). We also examined whether the interaction between FBXO32 and the c-Myc mutant (mutant 168 -367) requires the domain MBIII or MBIV or the We next sought to determine the FBXO32 domains required for c-Myc interaction. It appeared that all of the domains in the FBXO32 could pull down c-Myc, whereas the amino terminus (residues 1-223) or the carboxyl terminus alone interacted with c-Myc much more strongly. Of note, the truncated mutant of FBXO32 lacking the F-box domain (1-355(⌬F-box)) still interacted with c-Myc. To further determine whether the F-box domain is required for c-Myc interactions, we fused GFP to 218 -355 and 271-355, respectively, and performed immunoprecipitation assays. As shown in Fig. 3J, c-Myc could interact with the domain 272-355 of FBXO32 as well as the domain 218 -355 of FBXO32. This suggests that the F-box is definitely not required for FBXO32 to interact with c-Myc. In sum, the data suggest that FBXO32 interacts with c-Myc endogenously and directly.
FBXO32 Catalyzes c-Myc to Form Lys-48-linked Polyubiquitin Chain and Induces c-Myc for Proteasomal Degradation-To determine whether FBXO32 acts as an E3 ubiquitin ligase to induce c-Myc proteasomal degradation, we examined whether the potent proteasomal inhibitor, MG132, could block FBXO32-induced c-Myc degradation. We transfected HA-c-Myc into HEK293T cells as well as an HA empty vector control or HA-FBXO32. We then treated the cells with MG132 or DMSO (control) for 6 h before the harvest. The addition of MG132 effectively blocked FBXO32-induced c-Myc degradation (Fig. 4A), which indicates that FBXO32 mediates c-Myc proteasomal degradation.
We performed ubiquitination assays to confirm that FBXO32 can indeed catalyze c-Myc to form polyubiquitin chains (Fig. 4B). However, polyubiquitination of c-Myc reduced dramatically when the wild-type FBXO32 was replaced by the FBXO32 mutant that lacked the F-box domain. This suggested that FBXO32 requires the F-box domain to perform its E3 ubiquitin ligase role.

FBXO32 Targets c-Myc for Degradation
Lys-only mutants (Fig. 4D). These phenomena indicate that FBXO32 could indeed catalyze c-Myc to form Lys-48-linked polyubiquitin chains, which is consistent with its role in mediating c-Myc for proteasomal degradation.
FBXO32 Targets c-Myc Ubiquitination at Lys-326 -To identify the lysine residue(s) in c-Myc catalyzed by FBXO32, we systematically mutated them to arginine. The protein degradation efficiency by FBXO32 was used to monitor the potential ubiquitination site(s) in c-Myc. Fig. 5A illustrates that overexpression of FBXO32 did not promote degradation of one mutant (K326R). Although the mutant (K326R) showed the most dramatic increase of resistance to FBXO32-mediated degradation, it looks like additional sites (K51R/K52R, K126R, K157R, K289R, K371R, K389R, and K412R) also exhibited some resistance. To further confirm that Lys-326 is the key site catalyzed by FBXO32, we made additional mutants K51R, K52R, K51R/K52R/K126R/K157R/K289R/K371R/K389R/K412R (K8R), and K326RϩK8R and repeated the experiments. The mutants K51R, K52R, and K8R could still be degraded by FBXO32, although K8R exhibited some resistance (Fig. 5, B and C). When Lys-326 was mutated in K8R (K326RϩK8R), the mutant showed complete resistance (Fig. 5C). The quantitative data are summed in Fig. 5D. Versus the wild-type c-Myc, polyubiquitination of the c-Myc(K326R) mutant by the wild-type FBXO32 was clearly decreased, but it did not completely disappear (Fig. 5E). It appears that this reduction was not a result of the reduction of interaction between FBXO32 and c-Myc(K326R) because c-Myc(K326R) could still interact with c-Myc as well as the wild-type c-Myc (Fig. 5F). Interestingly, versus the wild-type c-Myc, polyubiquitination of the c-Myc K8R mutant by the FBXO32 was also decreased, and polyubiquitination of the c-Myc K326RϩK8R mutant induced by FBXO32 completely disappeared.
FBXO32 Suppresses c-Myc Activity-To evaluate the biological consequences of FBXO32-mediated c-Myc degradation, we examined the effect of FBXO32 on the transactivation of two well defined c-Myc target genes, Gadd45␣ (42)and E2F2 (43). As expected, ectopic expression of c-Myc inhibited the Gadd45␣ promoter activity ϳ0.23-fold (p ϭ 0.0050) (Fig. 6A) in HEK293T cells. However, co-expression of FBXO32 together with c-Myc restored the suppressive role of c-Myc in Gadd45␣ promoter activity (p ϭ 0.0044) (Fig. 6A). The c-Myc activated the E2F2 promoter activity ϳ2.55-fold. Co-expression of FBXO32 together with c-Myc abrogated the E2F2 promoter activity induced by c-Myc (p ϭ 0.0190). In contrast, c-Myc had no activating role in the mutated E2F2 promoter activity (p ϭ 0.1107) (Fig. 6C). Co-expression of FBXO32 also could not exhibit its suppressive role in the mutated E2F2 promoter activity (Fig. 6C). The expressions of ectopically expressed HA-c-Myc and HA-FBXO32 in HEK293T cells were confirmed by Western blot analysis (Fig. 6D).
To further evaluate the role of FBXO32 in c-Myc activity, we established stable A673 cell lines (a rhabomyosarcoma cell line) with overexpression of FBXO32 or knockdown of FBXO32 by transduced A673 cells using lentiviral infection. The endogenous FBXO32 protein levels in mouse muscle, A673 cells, and HCT116 cells were examined by Western blot assay (data not shown). Overexpression of FBXO32 reduced the E2F2 mRNA level but caused an increase in the Gadd45␣ mRNA level, as revealed by semiquantitative RT-PCR analysis (Fig. 6, E and F). In contrast, knockdown of FBXO32 by either FBXO32-shRNA-1 or FBXO32-shRNA-2 caused an increase in the E2F2 mRNA level, but this reduced the Gadd45␣ mRNA level, as revealed by semiquantitative RT-PCR analysis (Fig. 6, G and H).
These observations are consistent with the data obtained from the promoter assays.
To determine whether FBXO32-mediated regulation of Gadd45␣ and E2F2 is dependent on c-Myc, we generated three stable HCT116 cell lines with lentiviral infection. The first cell line expressed c-Myc-shRNA, which targets the 5Ј-UTR region of c-Myc. The second and third cell lines were established by reinfecting the first cell line with lentiviruses expressing the wild-type c-Myc and c-Myc(K326R) mutant, respectively. In cells with c-Myc-knocked down, overexpression of FBXO32 had no obvious effect on Gadd45␣ and E2F2 promoter activity (Fig. 6, I and M). In contrast, the effects of FBXO32 on Gadd45␣ promoter activity and the inhibition of FBXO32 on E2F2 promoter activity were recovered with c-Myc restoration (Fig. 6, J  and N). However, in the third cell line with c-Myc(K326R) mutant expression, the activation of FBXO32 on Gadd45␣ promoter activity and the inhibition of FBXO32 on the E2F2 promoter activity were not recovered (Fig. 6, K and O). The expressions of wild-type c-Myc, c-Myc(K326R), and HA-FBXO32 were confirmed by Western blot analysis (Fig. 6L). Whenever c-Myc was knocked down by c-Myc-shRNA or wild-type c-Myc was restored in c-Myc-knocked down cells, overexpression of FBXO32 had no effect on the E-box-mutated E2F2 promoter activity (Fig. 6, P and Q). These observations suggest that FBXO32-mediated regulation of Gadd45␣ and E2F2 is depen-  FIGURE 5. FBXO32 targets c-Myc ubiquitination at K326. A, HEK293T cells were transfected with the indicated c-Myc mutants together with either FBXO32 or empty vector. The expressions of c-Myc were detected by Western blot analysis using anti-HA antibody. B, degradation of the c-Myc mutants, K51R, K52R, and K51R/K52R induced by FBXO32 was further confirmed. C, degradation of the c-Myc multiple mutant, K8R, was further confirmed. D, protein levels were quantified based on band density obtained in Western blot assays; the protein level with HA empty vector transfection was treated as 1; the statistical analysis was performed using GraphPad Prism version 5.0 (unpaired Student's t tests); the protein level with HA empty vector transfection was treated as 1. E, the catalytic capability of FBXO32 on c-Myc(Lys-326) polyubiquitination is reduced significantly (the sixth column from the left to the right) compared with that on wild-type c-Myc (the second column from the left to the right). F, the c-Myc Lys/Arg mutant (HA-c-Myc(K326R)), as well as the wild-type c-Myc (HA-c-Myc(WT)), interacts with FBXO32, as revealed by co-immunoprecipitation assays. For ectopic expression, the amount of HA-c-Myc(WT) was 2 times more than that of HA-c-Myc(K326R) (ratio ϭ 1:3) regarding the degradation of wild-type c-Myc by FBXO32. G, the catalytic capability of FBXO32 on the c-Myc multiple mutant (K326R plus K8R) polyubiquitination is further reduced (the eighth column from the left to the right) compared with that on the c-Myc(K326R) (the fourth column from the left to the right). Error bars, S.E. IP, immunoprecipitation; IB, immunoblot. JUNE 26, 2015 • VOLUME 290 • NUMBER 26 dent on c-Myc. Taken together, these data suggest that FBXO32 suppresses c-Myc activity.

FBXO32 Targets c-Myc for Degradation
FBXO32 Inhibits Ovary Cancer Cell SKOV3 Proliferation-To elucidate the biological function of FBXO32 in mediating c-Myc degradation, we examined its effect on cell proliferation using three stable SKOV3 cell lines generated by lentiviral infection, vector control, FBXO32, and FBXO32-⌬F-box. Versus the control cells, the SKOV3 cells with the FBXO32 overexpression proliferated much more slowly from day 2 (Fig. 7A). In contrast, the proliferation rate of the SKOV3 cells with stable FBXO32-⌬F-box mutant expression is similar to that of the control cells. Conversely, the SKOV3 cells with stable knocked down FBXO32 proliferated faster than the control cells expressing GFP shRNA (Fig. 7B).
Through the colony formation assays, we further found that the colony number formed in the SKOV3 cells with the FBXO32 overexpression was fewer than that of the control SKOV3 cells or that of the SKOV3 cells with the FBXO32-⌬Fbox overexpression (Fig. 7D). Thus, FBXO32 can inhibit cell proliferation, which might be mediated by inducing c-Myc degradation.
FBXO32 Is a Direct Target Gene of c-Myc-As a transcription factor, c-Myc can form heterodimers with MAX that bind to E-boxes in the regulatory regions of the target genes and regulate a broad spectrum of genes involved in multiple physiological and pathological processes (44). After we performed bioinformatic analysis on the promoter region of FBXO32, we realized a potential E-box localized at the positions Ϫ107 to Ϫ102 (the translation initial site is designated as ϩ1). Therefore, we sought to determine whether FBXO32 is a direct target of c-Myc. We first amplified the FBXO32 promoter spanning Ϫ425 to ϩ24 and Ϫ204 to Ϫ5 by PCR and cloned this into the pGL4.10 vector (Promega). The promoter assays showed that ectopic expression of c-Myc caused a significant increase in FBXO32 promoter activity (Fig. 8A). The expression of HA-c-Myc was confirmed by Western blot analysis (Fig. 8B). However, when the core sequence of the E-box (CACGTG) in the Ϫ204 to Ϫ5 region of the FBXO32 promoter was mutated (CAGCTG), ectopic expression of c-Myc had no obvious effect on the activation of the mutated FBXO32 promoter activity versus that of wild-type FBXO32 promoter activity (Fig. 8, C and  D). Moreover, overexpression of c-Myc in HEK293T cells and HCT116 cells caused an increase in FBXO32 mRNA level (Fig.  8, E and F). Consistently, knockdown of c-Myc in HCT116 cells by c-Myc-shRNA-1 and c-Myc-shRNA-2 reduced the FBXO32 mRNA level. The induction of FBXO32 by c-Myc  's t tests). IB, immunoblot. FIGURE 7. FBXO32 inhibits ovary cancer cell SKOV3 proliferation. A, overexpression of wild-type FBXO32 by lentivirus infection in SKOV3 cells inhibits cell proliferation significantly after day 2 compared with the control, but overexpression of F-box-deleted mutant of FBXO32 (FBXO32-⌬F-box) has no obvious effect. The SKOV3 cells transduced with lentivirus vectors stably expressing GFP and FBXO32 or GFP alone (as a control) were seeded in 6-well plates with 8 ϫ 10 3 cells/well. B, knockdown of FBXO32 by FBXO32-shRNA-1 or FBXO32-shRNA-2 in SKOV3 cells enhances cell proliferation after day 2 compared with the control. The SKOV3 cells transduced with lentivirus vectors stably expressing FBXO32-shRNA-1, FBXO32-shRNA-2, or GFP-shRNA (as a control) were seeded in 6-well plates with 8 ϫ 10 3 cells/well. The cell numbers were counted every 2 days using an automated cell counter (Bio-Rad, TC20 TM ). C, overexpression of wild-type FBXO32 inhibits colony formation, but overexpression of F-box-deleted mutant of FBXO32 (FBXO32-⌬F-box) has no obvious effect. Data are presented as mean Ϯ S.E. (error bars) of three independent experiments performed in triplicate; the statistical analysis was performed using GraphPad Prism version 5.0 (t tests). JUNE 26, 2015 • VOLUME 290 • NUMBER 26 was further confirmed by Western blot analysis in the status with either overexpression of c-Myc or knockdown of c-Myc (Fig. 8, H and I).

FBXO32 Targets c-Myc for Degradation
To determine whether c-Myc can directly bind to the promoter of FBXO32, we performed ChIP analysis. Fig. 8, J and K, illustrates that the anti-c-Myc antibody could pull down a 200-bp fragment containing the E-box that is located in the Ϫ204 to Ϫ5 region of the FBXO32 promoter versus the rabbit IgG control. Semiquantitative RT-PCR further confirmed coprecipitation of the anti-c-Myc with the promoter of FBXO32 (Fig. 8L).
Considered collectively, these data suggest that c-Myc could up-regulate FBXO32 expression by directly binding to the FBXO32 promoter. Thus, FBXO32 might be a direct target of c-Myc.

Discussion
FBXO32 Targets c-Myc for Proteasomal Degradation-FBXO32 was identified as a muscle-specific E3 ubiquitin ligase more than 10 years ago. Increasing evidence indicates that FBXO32 plays important roles in skeletal muscle atrophy as well as in tumorigenesis. In this study, we identified c-Myc as a novel target of FBXO32 for proteasomal degradation. First, we showed that FBXO32 promoted c-Myc degradation efficiently, which was effectively blocked by MG132. This suggested that FBXO32 serves as an E3 ubiquitin ligase to mediate c-Myc proteasomal degradation. Similar to the eIF3-f and MyoD degradation targets, the F-box domain is required for FBXO32 to induce c-Myc ubiquitination and degradation, resembling the F-box-containing E3 ubiquitin ligase (6,7,45).
Through mutant screening, we identified that Lys-326 of c-Myc might be the key position needed for FBXO32 to catalyze the formation of Lys-48-linked polyubiquitin chain because FBXO32 could not induce c-Myc(K326R) mutant degradation. Although the capability of the polyubiquitin chain formation catalyzed by wild-type FBXO32 was clearly reduced, the FBXO32 still could stimulate formation of polyubiquitin chains on c-Myc(K326R) mutants. This phenomenon suggests that FBXO32 might also catalyze formation of other kinds of polyubiquitin chains (except Lys-48 polyubiquitin chains) or catalyze other lysine sites (except Lys-326) to form polyubiquitin chains. Interestingly, we noticed that the c-Myc mutant K326RϩK8R completely lost the ability to  type)) localized at the FBXO32 promoter (Ϫ204 to Ϫ5) and the mutated E-box construct. D, overexpression of c-Myc activates the wild-type FBXO32 promoter reporter (Ϫ204 to Ϫ5) activity but not the mutated FBXO32 promoter reporter activity. E, the mRNA level of FBXO32 is elevated when c-Myc is overexpressed in HEK293T cells, as revealed by a semiquantitative RT-PCR assay (p ϭ 0.0061). F, the mRNA level of FBXO32 is elevated when c-Myc is overexpressed in HCT116 cells, as revealed by semiquantitative RT-PCR assays (p ϭ 0.0003). G, the mRNA level of FBXO32 is reduced when c-Myc is knocked down in HCT116 cells by c-Myc-shRNA-1 and c-Myc-shRNA-2, as revealed by semiquantitative RT-PCR assays. H, the protein level of FBXO32 is elevated when c-Myc is overexpressed in HCT116 cells, as revealed by Western blot analysis using anti-c-Myc antibody. I, the protein level of FBXO32 is reduced when c-Myc is knocked down in HCT116 cells by c-Myc-shRNA-1 and c-Myc-shRNA-2 as revealed by Western blot analysis. J, schematic of locations of fragments amplified in FBXO32 promoter or ␤-actin promoter. K, ChIP analysis indicates that c-Myc interacts directly with the FBXO32 promoter region harboring the E-box. L, semiquantitative RT-PCR analysis confirms that c-Myc binds to the Ϫ204 to Ϫ5 region of the FBXO32 promoter. Data are presented as the means Ϯ S.E. (error bars) of three independent experiments performed in triplicate.

FBXO32 Targets c-Myc for Degradation
form a polyubiquitin chain, and FBXO32 could also catalyze c-Myc to form Lys-11-only polyubiquitin chains. This may support the hypothesis.
The effect of FBXO32 on c-Myc is quite similar to that of FBW7, which is another well defined F-box E3 ubiquitin ligase that targets c-Myc for proteasomal degradation and inhibits c-Myc activity. However, phosphorylation of Thr-58 and Ser-62 is required for c-Myc degradation mediated by FBW7. In contrast, phosphorylation of Thr-58 and Ser-62 is not required for c-Myc degradation mediated by FBXO32, which suggests that FBXO32 mediates c-Myc degradation through a mechanism different from that of FBW7.
FBXO32 May Perform Its Tumor-suppressive Function through Targeting c-Myc-To date, the role of FBXO32 in tumorigenesis remains debatable. Whereas FBXO32 is considered a hallmark of cancer cachexia (13,46), FBXO32 is also found to be down-regulated in some types of cancer through different mechanisms (9,47), suggesting its tumor-suppressive function. In this study, we showed that FBXO32 could effectively induce degradation of c-Myc, a classic oncogene overactivated in many types of cancer (48). Moreover, we found that overexpression of FBXO32 suppresses growth of ovary cancer SKOV3 cells but that knockdown of FBXO32 enhances growth of SKOV3 cells. These findings reinforce the tumor-suppressive function of FBXO32. As a muscle-specific gene, further exploring the role of FBXO32 in the initiation and progression of some special types of cancer that developed from muscle tissues (i.e. rhabdomyosarcoma) will shed light on the role of FBXO32 in tumorigenesis.
Negative Feedback Regulation of c-Myc by FBXO32-In this study, we also identified FBXO32 as a direct downstream target of c-Myc, revealing a negative feedback regulation loop between c-Myc and FBXO32. c-Myc transcriptionally regulates FBXO32 expression, but FBXO32 promotes c-Myc degradation. Actually, another E3 ubiquitin ligase of c-Myc (17,18), Skp2, is also up-regulated by c-Myc, although it serves as a co-factor instead of a suppressor for c-Myc-regulated transcription (41). Notably, this type of regulation has been widely recognized, including the p53/MDM2 (38) pathway, which implicates a strict regulation mechanism between the two genes. Further understanding of this regulation loop will open a new window for understanding the mechanisms of these two genes in tumorigenesis.