Smurf1 inhibits mesenchymal stem cell proliferation and differentiation into osteoblasts through JunB degradation

Abstract

Ubiquitin ligase Smurf1-deficient mice develop an increased-bone-mass phenotype in an age-dependent manner. It was reported that such a bone-mass increase is related to enhanced activities of differentiated osteoblasts. Although osteoblasts are of mesenchymal stem cell (MSC) origin and MSC proliferation and differentiation can have significant impacts on bone formation, it remains largely unknown whether regulation of MSCs plays a role in the bone-mass increase of Smurf1-deficient mice. In this study we found that bone marrow mesenchymal progenitor cells from Smurf1(-/-) mice form significantly increased alkaline phosphatase-positive colonies, indicating roles of MSC proliferation and differentiation in bone-mass accrual of Smurf1(-/-) mice. Interestingly, Smurf1(-/-) cells have an elevated protein level of AP-1 transcription factor JunB. Biochemical experiments demonstrate that Smurf1 interacts with JunB through the PY motif and targets JunB protein for ubiquitination and proteasomal degradation. Indeed, Smurf1-deficient MSCs have higher proliferation rates, consistent with the facts that cyclin D1 mRNA and protein both are increased in Smurf1(-/-) cells and JunB can induce cyclinD1 promoter. Moreover, JunB overexpression induces osteoblast differentiation, shown by higher expression of osteoblast markers, and JunB knock-down not only decreases osteoblast differentiation but also restores the osteogenic potential to wild-type level in Smurf1(-/-) cells. In conclusion, our results suggest that Smurf1 negatively regulates MSC proliferation and differentiation by controlling JunB turnover through an ubiquitin-proteasome pathway.

Fig. 1

Smurf1^−/− mice have increased bone marrow–derived osteogenic colony formation. (A) Representative dishes of ALP⁺ or total colonies formed by wild-type and Smurf1^−/− bone marrow stromal cells. Colony-forming unit (CFU) assays were performed as described in “Materials and Methods.” CFU-ALP, the CFU assays stained with BCIP/NBT, showed ALP⁺ colonies; CFU-total = the CFU assays counterstained with eosin. (B) Representative microscopic (×4) images of wild-type or Smurf1^−/− colonies stained with BCIP/NBT and eosin. (C) Statistical analyses of CFU assays. The number of colonies (each containing a minimum of 20 cells) was calculated after the total number of Smurf1^−/− colonies was set to 100%. Bars represent mean ± SD of three dishes per genotype. *p < .05 versus wild-type cells, n = 3. (D) Bone marrow stromal cells from wild-type and Smurf1^−/− mice were cultured in osteoblast differentiation medium, and ALP mRNA levels were examined at different time points using quantitative RT-PCR. *p < .05 versus wild-type cells, n = 3. (E) C2C12 cells were treated with BMP2 (100 ng/mL) with indicated time. Osteocalcin, JunB, and Runx2 mRNA levels were examined using real-time qPCR. *p < .05 versus time 0, n = 3.

Fig. 2

Smurf1 induces JunB degradation in vivo and in vitro. (A, B) JunB protein level was increased in the primary calvarial (A) and bone marrow stromal cells (B) of Smurf1^−/− mice compared with that of wild-type mice (n = 4). c-Jun protein level also was determined by Western blot analysis. (C) Real-time qPCR data showed that JunB mRNA level underwent no significant changes in the bone marrow stromal cells of Smurf1^−/− and wild-type mice (n = 3). (D) Smurf1 expression decreases JunB protein levels in a dose-dependent manner (n = 5). HEK293T cells were transfected with Flag-JunB and Myc-Smurf1 plasmids. Myc-Smurf1 was added in different amounts. Myc-Smurf1 (C710A) is a mutated Smurf1 in which an active-site cysteine in its catalytic domain is converted into alanine to abolish its activity as an ubiquitin ligase. (E, F) HEK293T cells transfected with Myc-Smurf1 and Flag-JunB were treated with protein translation inhibitor cycloheximide (80 µg/mL) for indicated time and subjected to Western blot analysis of JunB protein levels to determine the half-life of JunB proteins in the presence or absence of Smurf1 overexpression in HEK293T cells (n = 3). Both Western blot (E) and its quantification (F) are shown.

Fig. 3

Smurf1 interacts with JunB through a PY motif and induces its ubiquitination and degradation. (A) GST-Smurf1 directly interacts with JunB protein. In vitro–translated JunB protein was copurified with GST-Smurf1 in a GST pull-down assay (n = 3). (B) JunB-Smurf1 interaction in mesenchymal C2C12 cells. Smurf1 was identified in the immunoprecipitates by anti-JunB antibodies (n = 3). (C) A diagram showing a YF mutation in the PY motif of JunB that was generated to examine if the PY motif is required by Smurf1-induced JunB degradation. (D) Wild-type JunB protein interacts with Smurf1, whereas JunB (YF) mutant cannot (n = 3). HEK293T cells were transfected with Myc-Smurf1-C710A and Flag-JunB or Flag-JunB-YF. IP assays were carried out using anti-Flag antibodies, and immunoprecipitates were subjected to Western blot using anti-Myc antibodies; whole-cell lysates also were examined to ensure that both JunB and Smurf1 were expressed. (E) Smurf1 overexpression could not decrease the protein level of JunB-YF (n = 4). HEK293T cells were transfected with plasmids as indicated. (F) Proteasome inhibitor MG132 rescued JunB protein from Smurf1-induced degradation (n = 3). HEK293T cells were transfected with plasmids as shown. Cells transfected with Myc-Smurf1 and Flag-JunB were treated with MG132 (10 µM) for 4 hours before being harvested and subjected to Western blot analysis. (G) JunB ubiquitination assay (n = 3). 293T cells transfected with HA-ubiquitin, Myc-Smurf1, and Flag-JunB or Flag-JunB-YF were treated with MG132 for 4 hours before harvesting, and polyubiquitinated JunB was immunoprecipitated by anti-Flag and immunoblotted with anti-HA.

Fig. 4

Smurf1 negatively regulates MSC proliferation. (A) Western blot results showed that the cyclin D1 protein level was increased in the primary calvarial cells or in the marrow stromal cells of Smurf1^−/− mice compared with wild-type mice (n = 3). (B) Real-time qPCR data showed that the cyclin D1 mRNA level was increased in marrow stromal cells of Smurf1^−/− mice. *p < .05 versus wild-type cells, n = 3. (C) C3H10T1/2 mesenchymal stem cells were transfected with cyclin D1 reporter and different doses of JunB overexpression plasmids. Luciferase activities showed that cyclin D1 promoter was induced by JunB expression in a dose-dependent manner. *p < .05 versus pGL3-Luc, n = 3. (D) C2C12 myoblast/osteoblast progenitor cells were transfected with cyclin D1 reporter and different doses of JunB overexpression plasmids. Luciferase activities showed that cyclin D1 promoter was induced by JunB expression in a dose-dependent manner. *p < .05 versus pGL3-Luc, n = 3. (E) The numbers of Smurf1^−/− bone marrow stromal cells were higher than those of wild-type cells after 7 days of culture in 12-well plates. *p < .05 versus wild-type cells, n = 4. (F) Cell proliferation rates were increased in Smurf1^−/− BMSCs. Thus 3 × 10⁵ or 9 × 10⁵ BMSCs per well were seeded into 96-well plate, and cell proliferation was examined using a colorimetric assay. *p < .05 versus wild-type cells, n = 4. (G) Smurf1^−/− CD45⁻ MSCs have a higher proliferation rate than wild-type cells. Bone marrow cells were stained with APC-CD45 and FITC–anti-BrdU after in vivo labeling of BrdU by i.p. injection. CD45⁻ and BrdU⁺ populations of wild-type and Smurf1^−/− cells were analyzed (n = 3).

Fig. 5

Smurf1-induced JunB degradation has a negative impact on osteoblast differentiation of MSCs. (A) Western blot showed that the JunB protein level was increased in C2C12 cells infected with JunB-overexpressing retrovirus (n = 3). (B) Enhanced ALP activity by JunB overexpression in C2C12 myoblast/osteoblast progenitor cells shown by ALP staining of control or JunB-overexpressing cells on BMP2 treatment at the doses indicated (n = 3). (C, D) JunB knockdown by siRNA was confirmed by Western blotting (C) or RT qPCR (D). *p < .05 versus wild-type cells + control siRNA, n = 3. Calvarial cells of wild-type and Smurf1^−/− mice were isolated and transfected with JunB siRNA or control siRNA. Cells were cultured in osteoblast differentiation medium with or without BMP2. (E) ALP activity of the calvarial cells transfected with siRNAs was measured to evaluate whether JunB knockdown restored increased osteogenic differentiation in Smurf1^−/− cells. *p < .05 versus wild-type cells + control siRNA, n = 3. (F–H) CD45⁻ MSCs were isolated from in vitro–cultured bone marrow stromal cells of wild-type and Smurf1^−/− mice, transfected with JunB siRNA or control siRNA and then cultured in osteoblast differentiation medium with or without BMP2. JunB (F), ALP (G) and osteocalcin (H) mRNA levels were measured by real-time qPCR. *p < .05 versus wild-type cells + control siRNA, n = 3.