Smad6 Interacts with Runx2 and Mediates Smad Ubiquitin Regulatory Factor 1-induced Runx2 Degradation*

Runx2 is a bone-specific transcription factor that plays a critical role in bone development, postnatal bone formation, and chondrocyte maturation. The protein levels of Runx2 are regulated by the ubiquitin-proteasome pathway. In previous studies we discovered that E3 ubiquitin ligase Smad ubiquitin regulatory factor 1 (Smurf1) induces Runx2 degradation in a ubiquitin-proteasome-dependent manner, and Smurf1 plays an important role in osteoblast function and bone formation. In the present studies we investigated the molecular mechanism of Smurf1-induced Runx2 degradation. Smurf1 interacts with the PY motif of substrate proteins, and a PY motif has been identified in the C terminus of the Runx2 protein. To determine whether Smurf1 induces Runx2 degradation through the interaction with the PY motif of Runx2, we created a mutant Runx2 with a PY motif deletion and found that Smurf1 retained some of its ability to induce the degradation of the mutant Runx2, suggesting that Smurf1 could induce Runx2 degradation through an indirect mechanism. Smurf1 has been shown to interact with Smads 1, 5, 6, and 7, and Smads 1 and 5 also interact with Runx2. In the present studies we found that Smads 1 and 5 had no effect on Smurf1-induced Runx2 degradation. Although Smads 6 and 7 bind Smurf1, it is not known if Smads 6 or 7 interacts with Runx2 and mediate Runx2 degradation. We performed immunoprecipitation assays and found that Smad6 but not Smad7 interacts with Runx2. Smad6 enhances Smurf1-induced Runx2 degradation in an ubiquitin-proteasome-dependent manner. These results demonstrate that in addition to its interaction with the PY motif of Runx2, Smurf1 induces Runx2 degradation in a Smad6-dependent manner. Smurf1-induced Runx2 degradation serves as a negative regulatory mechanism for the BMP-Smad-Runx2 signaling pathway.

Runx2 (Runt-related gene 2) is a bone-specific transcription factor that belongs to the runt-domain gene family. DNA-binding sites for Runx2 have been identified in the promoter regions of many osteoblastspecific genes (1)(2)(3)(4)(5)(6), and Runx2 binds responsive elements in these promoters and regulates the transcription of these genes. Targeted disruption of Runx2 in mice reveals that Runx2 expression is absolutely required for bone development in vivo. Homozygous Runx2-deficient mice die soon after birth due to an inability to breathe. The most pronounced effect is a complete lack of both endochondral and intramembranous ossification (7,8), with an absence of mature osteoblasts throughout the body. Heterozygous mutant mice have skeletal abnor-malities similar to those seen in a human mutation called cleidocranial dysplasia syndrome (9,10) and delayed development of intramembranous bones (7,8). In transgenic mice overexpressing a dominant-negative Runx2 DNA binding domain (mRunx2) driven by the osteocalcin promoter, skeletons are normal at birth, but the mice suffer from osteopenia due to a decrease in bone formation rate 3 weeks after birth (11). These results indicate that Runx2 plays a crucial role not only in bone development but also in postnatal bone formation.
Runx2 is also an important regulator for chondrocyte differentiation and maturation. Supporting evidence includes histomorphologic findings showing an alteration of chondrocyte maturation in long bones of Runx2 null mutant mice as well as cell culture studies indicating that Runx2 is a positive regulator for chondrocyte differentiation (12)(13)(14). Mis-expression of Runx2 in proliferating chondrocytes induces chondrocyte hypertrophy and partially rescues the chondrocyte phenotype of Runx2 null mutant mice (15,16). Selective inactivation of Runx2 in chondrocytes results in severe shortening of the limbs due to a disturbance in chondrocyte differentiation (17). The type X collagen (colX) gene is induced by Runx2 in chondrocytes, and Runx2 binding elements have been identified and characterized in the promoter of the colX gene (18,19).
One of the mechanisms by which transcription factors are regulated is by modulation of degradation. Runx2 undergoes ubiquitin-mediated proteasomal degradation (20). Protein ubiquitination involves a sequential cascade of enzymatic reactions catalyzed by the E1 ubiquitin-activating enzyme, the E2 ubiquitin-conjugating enzymes, and the E3 ubiquitin ligases (21). Among these enzymes, E3 ubiquitin ligases play a crucial role in defining substrate specificity and subsequent protein degradation by 26 S proteasomes. Smurf1 2 is a member of the Hect domain family of E3 ubiquitin ligases and interacts with Smads 1 and 5 and BMP receptors (22,23), thereby triggering the ubiquitination and degradation of these proteins. We have recently found that the E3 ubiquitin ligase Smurf1 interacts with Runx2 and induces Runx2 degradation in osteoblasts in a ubiquitin-proteasome-dependent manner (24). Overexpression of Smurf1 in osteoblasts inhibits osteoblast differentiation and postnatal bone formation (25). In the present studies we further investigated the molecular mechanism of Smurf1 on Runx2 degradation. In a series of experiments using COS cells as a cell culture model, which permit relatively high levels of protein expression (22, 26 -28) and have lower Smad6 expression, we investigated the molecular mechanism of Smurf1-induced Runx2 degradation and found that Smad6 interacts with Runx2 and mediates Smurf1-induced Runx2 degradation in a PY motif-independent manner. * This work was supported by National Institutes of Health Grants AR051189 and
Luciferase Assay-6xOSE2-OC-Luc reporter construct was cotransfected with Runx2 and different amounts of Smurf1 or mutant Smurf1 (C710A) expression plasmids into COS cells. Cell lysates were extracted 48 h after transfection, and luciferase activity was measured using the Promega Dual Luciferase Reporter Assay kit (Promega, Madison, WI).
Pulse-Chase Assay-COS or C2C12 cells were seeded in 6-cm culture dishes at ϳ50% confluence. Cells were transfected with an equal amount of FLAG-Runx2. 24 h after transfection cells were starved with Met/Cys-free DMEM (Invitrogen Cell Culture) with 5% dialyzed FCS (Invitrogen) for 1 h. Cells were then pulsed with 100 Ci/ml [ 35 S]methionine for 40 min followed by a chase in regular DMEM medium with 2 mM methionine and 2 mM cysteine for 0, 15, 30, 60, and 120 min. The sample at each time point was collected for immunoprecipitation using 5 g of anti-FLAG antibody. Samples were separated by SDS-PAGE gel, then dried and followed by autoradiographic exposure. The signals were quantified using Lab Works 4.0 Image analysis software (UVP, Inc. Upland, CA).
Immunoprecipitation-COS or 293 cells were transiently transfected with expression plasmids using Lipofectamine 2000 reagents (Invitrogen) and incubated for 24 h before analysis. After transfection cells were washed once with phosphate-buffered saline, lysed for 30 min in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% Nonidet P-40) containing protease inhibitors (10 g/ml leupeptin, 10 g/ml pepstatin A, and 10 g/ml aprotinin) and phosphatase inhibitors (1 mM NaF and 1 mM Na 3 VO 4 ), and clarified by centrifugation at 4°C for 15 min. Supernatants were pre-cleared with EZ View Red protein G-Sepharose (Sigma) for 1 h at 4°C. Then 5 g of antibodies specific for each target protein were added in each sample. Immune complexes were precipitated by EZ View red protein G-Sepharose overnight at 4°C, washed 5 times with the lysis buffer. The immune complexes were boiled for 10 min in SDS sample buffer (100 mM Tris-HCl, pH 8.8, 0.01% bromphenol blue, 20% glycerol, 4% SDS) containing 10 mM dithiothreitol and analyzed by 10% SDS-PAGE. Western blot analysis was performed after immunoprecipitation (IP).
Preparation of Virus-Platinum-E cells in the 6-well plate were cultured in DMEM medium (2 ml) containing 10% FCS, 1 g/ml puromycin, 10 g/ml blasticidin, and antibiotics (penicillin and streptomycin). When cells reached 30% confluence, retrovirus vector (pRetro-H1G), which expresses Smurf1-or Smad6-small hairpin RNA (shRNA) or control shRNA (Cellogenetics Inc., Baltimore, MD), was transfected into the cells using FuGENE 6 (Roche Applied Science). 1 g of vector was incubated with 3 l of FuGENE 6 in DMEM (0.1 ml) for 15 min at room temperature, and then the mixture was added to the cells. After 2 days supernatants were harvested and filtered with a 0.45-m membrane filter.
Infection of Smurf1 or/and Smad6-shRNA-Cell density of 2T3 cells was adjusted to 2 ϫ 10 5 cells/ml in serum-free ␣MEM. An equal volume of virus supernatant and Polybrene (32 g/ml) was incubated for 5 min at room temperature. The mixture was added to the cell suspension, and then 2 ml of cell mixture was plated into a 6-cm dish (final: Polybrene, 8 g/ml; 2T3 cells, 2 ϫ 10 5 cells/dish). After cells were cultured in a CO 2 incubator for 4 h, 2 ml of ␣MEM containing 10% FCS was added to cells to dilute the Polybrene concentration (total medium volume 4 ml). When cells reached 90% confluence, they were washed with serum-free ␣MEM (3 ml) 3 times and cultured for 48 h in ␣MEM containing 10% FCS. Cells were lysed with M-PER (50 l, Pierce) containing Complete Mini (Roche Applied Science) and 1 mM phenylmethylsulfonyl fluoride and subjected to Western blot analysis. The experiment was performed twice, Western blot data were quantified using Lab Works 4.0 Image Analysis Software (UVP, Inc. Upland, CA), and the average fold change was provided in Fig. 5, c and d.

Smurf1 Mediates Runx2 Degradation in COS Cells-
In previous studies we found that E3 ligase Smurf1 induces Runx2 degradation in myoblast precursor C2C12 cells (24). In the present studies we have examined the effect of Smurf1 on Runx2 degradation using COS cells. COS and 293 cells have been extensively used for studies of protein degradation of Smad2 and transforming growth factor ␤ and BMP receptors (26,27,32). Expressed Runx2 has been demonstrated to have a similar phosphorylation pattern in COS and 293 cells compared with that of endogenous Runx2 in SAOS-2 cells (33). When Runx2 expression plasmid was co-transfected with different amounts of Smurf1 or mSmurf1 (C710A, lacking catalytic activity) expression plasmid into COS cells, Smurf1 induced Runx2 degradation in a dose-dependent manner, and mSmurf1 had no significant effect on Runx2 degradation (Fig. 1a). When these expression plasmids were co-transfected with Runx2 reporter construct, 6xOSE2-OC-Luc, Smurf1 inhibited Runx2induced luciferase activity of the reporter in a dose-dependent manner, whereas mSmurf1 had no significant effect on Runx2-activated 6xOSE2-OC-Luc reporter (Fig. 1b).
Smurf1 Induces Runx2 Degradation through a PY Motif-independent Mechanism-It has been reported that Smurf1 interacts with the PY motif of its substrate Smads 1 and 5 (22). In the Runx family members a conserved PY motif has been identified by amino acid sequence analysis, and the PY motif of Runx2 is located in the C terminus of the protein Smad6 Mediates Smurf1-induced Runx2 Degradation (Fig. 1c). It has been recently reported that Smurf1 mediates Runx3 degradation and that this activity is partially lost in 293 cells transfected with a mutant Runx3 having a PY motif deletion (34). To determine whether Smurf1-induced Runx2 degradation requires the PY motif of Runx2, we generated a mRunx2 expression plasmid with a PY motif deletion (Fig. 1c). Wild-type and mRunx2 expression plasmids were co-transfected with Smurf1 expression plasmid into COS cells. Although Smurf1 induced degradation of both wild-type and mRunx2, the level of Runx2 degradation was reduced in COS cells expressing mRunx2 compared with the wild-type Runx2 (Fig. 1d). These results suggest that the PY motif of Runx2 is not essential for Smurf1-induced Runx2 degradation. To compare the effects of Smurf1 on the degradation of Runx2 to that of Runx3, Smurf1 was co-transfected with either wild-type or mRunx3(ϪPY) expression plasmids. Similar to Runx2, Smurf1 induced mRunx3 degradation, although the degree of the degradation was reduced compared with wild-type Runx3 (Fig. 1e). To confirm the effects of Smurf1 on mRunx2 and mRunx3 degradation, we transfected Smurf1 with either mRunx2 or mRunx3 into C2C12 cells and found that Smurf1 induced the degradation of mRunx2 and mRunx3 in a manner similar to that observed in COS cells (Fig. 1f). Altogether, the results suggest that in addition to the direct involvement of the PY motif of Runx2 and Runx3 proteins, Smurf1 also induces the FIGURE 1. a and b, Smurf1 induces Runx2 degradation in COS cells. a, FLAG-tagged Runx2 (F-Runx2) expression plasmid was co-transfected with different amounts of Smurf1 (0.1, 0.2, and 0.4 g/dish, 6-cm culture dish) or mSmurf1(CA) (0.4 g/dish) expression plasmid into COS cells. Smurf1 induced Runx2 degradation in a dose-dependent manner, and mSmurf1 had no effect on Runx2 degradation. WB, Western blot. b, Runx2 reporter construct, 6xOSE2-OC-Luc, was co-transfected with different amounts of Smurf1 (0.1, 0.2, and 0.4 g/dish) or mSmurf1 expression plasmid (0.4 g/dish) into COS cells. Smurf1 inhibited Runx2-induced luciferase activity of the Runx2 reporter in a dose-dependent manner, and mSmurf1 had no effect on Runx2-induced reporter activity. *, p Ͻ 0.05, one-way analysis of variance followed by Dunnett's test, compared with Runx2 alone (n ϭ 4). c and d, the PY motif is not essential for Smurf1-induced Runx2 degradation. c, amino acid sequence analysis of Runx family members reveals that there is a conserved PY motif among Runx family members, and a mRunx2 expression plasmid with a PY motif deletion, mRunx2(ϪPY), has been generated. d, transfection of wild-type and mRunx2(ϪPY) expression plasmids with Smurf1 expression plasmid into COS cells resulted in a degradation of wild-type as well as mRunx2 proteins, although the intensity of the degradation is reduced for the mRunx2(ϪPY). e, transfection of wild-type and mRunx3(ϪPY) expression plasmids with Smurf1 expression plasmid into COS cells resulted in a similar result. Smurf1 induced wild-type and mRunx3(ϪPY) degradation with higher efficiency to wild-type Runx3. f, F-mRunx2 and F-mRunx3 plasmids were co-transfected with or without Smurf1 plasmid into C2C12 cells (0.4 g/dish). Smurf1 induced mRunx2 and mRunx3 degradation in C2C12 cells. FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6 degradation of Runx2 and Runx3 by another PY motif-independent mechanism.

Smad6 Mediates Smurf1-induced Runx2 Degradation
Smad6 Interacts with Runx2-Comparing the protein stability of Runx2 in C2C12 and COS cells, we noticed that the levels of expressed Runx2 protein were decreased over a 48 -72-h period after transfection in C2C12 cells (Fig. 2a). In contrast, levels of expressed Runx2 protein are not changed up to 72 h in COS cells (Fig. 2b). The expression levels of Smurf1 are similar in C2C12 and COS cells (data not shown). To compare the degradation rates of Runx2 protein in these two types of cells, we performed pulse-chase experiments and found that the Runx2 degradation rate is much slower in COS cells compared with C2C12 cells (Fig. 2, c and d). It has been reported that Smurf1 has an ability to interact with Smads 1, 5, 6, and 7 (26), and Smads 1 and 5 also interact with Runx2 (35). We hypothesize that the Smad protein(s) may serve as an adaptor to mediate Smurf1-induced Runx2 degradation. We first tested if Smad1 mediates Smurf1-induced Runx2 degradation. Transfection of Smad1 with Runx2 and Smurf1 expression plasmids did not change Smurf1-induced Runx2 degradation in COS cells (Fig. 2e). A similar result was also obtained with Smad5 (data not shown). These results suggest that Smads 1 and 5 are not involved in Smurf1-induced Runx2 degradation. Very low expression of Smad7 was detected in both C2C12 and COS cells (data not shown). In contrast, expression levels of Smad6 in C2C12 cells were significantly higher (28-fold) compared with COS cells (Fig. 2f). Although Smads 6 and 7 bind Smurf1 and co-localize with Smurf1 in the nucleus (26,27,23), it is not known if Smads 6 or 7 interact with Runx2 and mediate Runx2 degradation. We performed IP assays in COS cells after transfection of Myc-tagged Runx2 with FLAGtagged Smurf1, Smad6, and Smad7 expression plasmids. Runx2 was immunoprecipitated by an anti-Myc antibody, and Smurf1 and Smad6, which interacted with Runx2, were detected by Western blot using an anti-FLAG antibody (Fig. 3a). In this experiment, we found that Smad6 but not Smad7 interacts with Runx2. Because Smad7 may co-migrate with IgG heavy chain, we then performed IP with an anti-FLAG antibody. Again, Smad6, but not Smad7, was co-precipitated with Runx2 in COS cells (Fig. 3b). To determine whether the PY motif of Runx2 is required for its interaction with Smad6, we performed IP experiments using mRunx2(ϪPY) and found that Smad6 also co-precipitated with mRunx2(ϪPY) (Fig. 3c), indicating that the PY motif deletion in Runx2 protein does not interfere with its interaction with Smad6. To further confirm the interaction of Smad6 with Runx2 in other cells, we transfected Myc-tagged Runx2 with FLAG-tagged Smad6 or Smad7 into 293 cells. The IP was performed using an anti-FLAG antibody followed by Western blot using an anti-Myc antibody. The result demonstrated that only Smad6 interacts with Runx2 in 293 cells (Fig. 3d). These results . WB, Western blot. c and d, FLAG-tagged Runx2 (F-Runx2) expression plasmid was transfected into C2C12 and COS cells, and the pulse-chase assays were performed 24 h after transfection. The autoradiographic results from the pulse-chase experiments indicated that the degradation rates of Runx2 were much lower in COS cells than in C2C12 cells. e, Smad1 had no effect on Smurf1-induced Runx2 degradation. Smad1 expression plasmid was co-transfected with Runx2 and Smurf1 expression plasmids into COS cells. Smad1 had no effect on Runx2 protein levels or Smurf1-induced Runx2 degradation in COS cells. f, Smad6 protein is highly expressed in C2C12 cells. Expression of Smad6 protein was compared in C2C12 and COS cells. Smad6 expression is much higher in C2C12 cells than in COS cells.
To determine the effect of Smad6 on Runx2 activity, Smad6 expression plasmid was co-transfected with Smurf1 and 6xOSE2-OC-Luc reporter. Smad6 enhanced the inhibitory effect of Smurf1 on Runx2-induced luciferase activity of the 6xOSE2-OC-Luc reporter (Fig. 4d). It is interesting to point out that Smad6 alone had only a minor effect on Runx2 degradation, although it significantly enhanced the effect of Smurf1 on FIGURE 3. Interaction of Smad6 with Runx2. a, Myc-tagged Runx2 (M-Runx2) expression plasmid was co-transfected with FLAG-tagged Smad6, Smad7, and mSmurf1(CA) expression plasmids into COS cells. IP assays were performed using an anti-Myc antibody followed by Western blot (WB) analysis using an anti-FLAG antibody. Smad6 and Smurf1 but not Smad7 were co-precipitated with Runx2 in COS cells. b, M-Runx2 was co-transfected with F-Smad6 and F-Smad7 in COS cells. IP assays were performed using an anti-FLAG antibody followed by Western blot analysis using an anti-Myc antibody. Smad6 but not Smad7 was co-precipitated with Runx2 into COS cells. c, Myc-tagged Runx2 and mRunx2(ϪPY) were co-transfected with F-Smad6 and F-Smad7 into COS cells. Smad6 but not Smad7 was detected by Western blot after Myc-Runx2 or Myc-mRunx2(ϪPY) proteins were immunoprecipitated by the anti-Myc antibody. d, Myc-tagged Runx2 was co-transfected with FLAG-tagged Smad6 or Smad7 into 293 cells. IP was performed using the anti-FLAG antibody followed by Western blot using the anti-Myc antibody. Only Smad6 interacts with Runx2 in this IP assay in 293 cells.
Inhibition of Smurf1 and Smad6 Enhances Endogenous Runx2 Protein Levels-To further determine the role of Smurf1 and Smad6 in regulation of Runx2 protein stability in osteoblasts, we infected osteoblast precursor 2T3 cells with retrovirus expressing double-stranded shRNA of Smurf1 and Smad6. The shRNA of Smurf1 and Smad6 are highly effective in blocking endogenous Smurf1 (81%) or Smad6 (77%) expression in 2T3 cells (Fig. 5, a and b). In 2T3 cells expressing small hairpin Smurf1, protein levels of endogenous Runx2 were increased 1.9-fold. Alone, Smad6 shRNA had no significant effect on Runx2 protein levels. When both Smurf1 and Smad6 shRNAs were infected in 2T3 cells, Runx2 protein levels were increased 3.7-fold (Fig. 5c). Infection of 2T3 cells with retrovirus expressing Smurf1 induced complete Runx2 degradation, and infection of shRNA of Smad6 reversed the effect of Smurf1 on Runx2 degradation in 2T3 cells (Fig. 5d). These results further demonstrate that Smurf1 and Smad6 cooperate to regulate Runx2 protein levels in osteoblast precursor cells.
Smurf2 and WWP1 Induce Runx2 Degradation with Smad6-To determine whether other Hect-domain E3 ligases also regulate Runx2 degradation, we examined the effects of Smurf2 and WWP1 (31) on Runx2 degradation in COS cells. Although Smurf2 had a weak effect on Runx2 degradation compared with Smurf1 (Fig. 6a), WWP1 effectively induced Runx2 degradation (Fig. 6b). In the presence of Smad6, Smurf2and WWP1-induced Runx2 degradation was enhanced (Fig. 6, a and b). These results show that Smurf2 and WWP1 also regulate the protein stability of Runx2 through a mechanism that involves Smad6.

DISCUSSION
Previously we discovered that Smurf1 interacts with bone-specific transcription factor Runx2 and induces Runx2 degradation in osteoblast precursor cells (24). In the present studies we provide new evidence that the interaction of Smurf1 to the PY motif of Runx2 is not essential for Smurf1-induced Runx2 degradation. Here we demonstrate that Smurf1 can degrade Runx2 in a PY motif-independent manner through forma- . a-c, Smad6 mediates Smurf1-induced Runx2 ubiquitination and proteasomal degradation. a, different amounts of F-Smad6 expression plasmid (0.1, 0.2, and 0.4 g/dish, 6-cm culture dish) was co-transfected with FLAG-tagged Runx2 (F-Runx2) and Smurf1 expression plasmids (0.5 g/dish) into COS cells. Smad6 enhanced Smurf1-induced Runx2 degradation. Smad6 itself was also degraded with Runx2 by Smurf1 (sixth through eighth lanes). WB, Western blot. b, F-Runx2, Smurf1 and HA-ubiquitin expression plasmids were co-transfected into COS cells. IP was performed using an anti-FLAG antibody followed by Western blot analysis using an anti-HA antibody. Smad6 enhanced Smurf1-induced Runx2 ubiquitination. Ub, ubiquitin. c, Runx2 expression plasmid was transfected into COS cells with Smad6 and Smurf1 expression plasmids and then treated with 5 M proteasome inhibitor proteasome inhibitor 1 (PS-1; fifth and sixth lanes) for 4 h. Smurf1 induced Runx2 degradation which was enhanced by Smad6. In the fourth lane, Runx2 protein was totally degraded by Smad6/Smurf1. Proteasome inhibitor PS-1 significantly reversed the effects of Smurf1 alone or Smurf1/Smad6-induced Runx2 degradation (fifth and sixth lanes). d, Smad6 enhances the inhibitory activity of Smurf1 on Runx2-induced reporter activity. Runx2 reporter, the 6xOSE2-OC-Luc, was co-transfected with Runx2 expression plasmid into COS cells. Runx2 significantly stimulated the reporter activity. Co-transfection of Smurf1 significantly inhibited Runx2-induced reporter activity, and co-transfection of Smad6 slightly inhibited Runx2-induced reporter activity. Transfection of Smad6 significantly enhanced the inhibitory effect of Smurf1. *, p Ͻ 0.05, unpaired t test, compared with Runx2 alone (n ϭ 4).

Smad6 Mediates Smurf1-induced Runx2 Degradation
tion of a complex that includes Smad6. These findings provide new insights about the mechanism of Smurf1-induced Runx2 degradation.
In the present studies we found that Smad6 interacts with Runx2 in COS and 293 cells, and Smurf1 induces the degradation of PY-mutant Runx2 or Runx3 in COS and C2C12 cells. More importantly, we also demonstrate that endogenous Runx2 protein levels are increased by the shRNA of Smurf1 and Smad6. These results demonstrate that this Smad6-mediated Runx2 degradation is not simply due to the overexpression of Smurf1 or Smad6 in a specific cell line.
Recent studies show that Smad6 has a synergistic effect with Smurf1 on inhibition of BMP signaling. Smad6 binds type I BMP receptors and prevents binding and phosphorylation of Smads 1 and 5 (36). Smad6 also facilitates Smurf1-induced type I BMP receptor degradation (23).
To determine the role of Smad6 and Smurf1 in chondrocyte matura-   FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6 tion, Horiki et al. (37) developed Col11-Smad6 and Col11-Smurf1 transgenic mice and demonstrate that Smurf1 enhances the effect of Smad6 on inhibition of chondrocyte differentiation and maturation (37). Our results provide new evidence that Smad6 mediates Smurf1induced Runx2 degradation. Because Smad6 and Smurf1 are co-localized in the nucleus as is Runx2, it is likely that Smad6-mediated Runx2 degradation occurs in the 26 S proteasome, which is located in the nucleus of cells.

Smad6 Mediates Smurf1-induced Runx2 Degradation
Hect-domain E3 ligases frequently use adaptor proteins to mediate the degradation of substrate proteins. For example, Smurf2, another member of the Smurf family, induces the protein degradation of SnoN, a transcriptional repressor of transforming growth factor ␤ signaling, using Smad2 as an adaptor (30). Similarly, Smurfs 1 and 2 target type I transforming growth factor ␤ and BMP receptors for degradation using Smads 6 and 7 as adaptor proteins (23, 26 -28, 38). The E6-AP is a Hect-domain E3 ubiquitin ligase. It binds E6 protein of human Papillomavirus types 16 and 18 and targets the ubiquitin-proteasome degradation of tumor suppressor p53 (39).
Our previous results demonstrated that transgenic mice overexpressing the Smurf1 transgene (Col1a1-Smurf1) in osteoblasts have impaired bone formation and osteoblast activity (25). More recently, it has been reported that bone mass is increased in adult Smurf1 null mutant mice (40). Surprisingly, Smurf1 null mutant mice had normal levels of Smads 1 and 5 and Runx2 proteins that have all previously been shown to be targeted by Smurf1 (22,24). Because these proteins are also targeted by Smurf2 or other E3 ligases (31), it is possible that Smurf2 and other E3 ligases may have played a redundant role in the degradation of these proteins in Smurf1 knock-out cells. In the present studies we found that the degradation of Runx2 is not only induced by Smurf1 but also induced by Smurf2 and WWP1. These findings demonstrate that Runx2 degradation involves multiple E3 ligases and is mediated through both PY motif-dependent and -independent mechanisms.
In summary, our findings indicate that Smad6 interacts with Runx2 and mediates Smurf1-induced Runx2 degradation. These results show that Smad6 and Smurf1 coordinately down-regulate Runx2 protein, which may serve as a negative regulatory mechanism for the BMP-Smad-Runx2 signaling pathway.