A RUNX2/PEBP2alpha A/CBFA1 mutation displaying impaired transactivation and Smad interaction in cleidocranial dysplasia

Abstract

Cleidocranial dysplasia (CCD), an autosomal-dominant human bone disease, is thought to be caused by heterozygous mutations in runt-related gene 2 (RUNX2)/polyomavirus enhancer binding protein 2alphaA (PEBP2alphaA)/core-binding factor A1 (CBFA1). To understand the mechanism underlying the pathogenesis of CCD, we studied a novel mutant of RUNX2, CCDalphaA376, originally identified in a CCD patient. The nonsense mutation, which resulted in a truncated RUNX2 protein, severely impaired RUNX2 transactivation activity. We show that signal transducers of transforming growth factor beta superfamily receptors, Smads, interact with RUNX2 in vivo and in vitro and enhance the transactivation ability of this factor. The truncated RUNX2 protein failed to interact with and respond to Smads and was unable to induce the osteoblast-like phenotype in C2C12 myoblasts on stimulation by bone morphogenetic protein. Therefore, the pathogenesis of CCD may be related to the impaired Smad signaling of transforming growth factor beta/bone morphogenetic protein pathways that target the activity of RUNX2 during bone formation.

Figure 1

The CCDαA376 mutant lacks transactivating ability. (a) Schematic illustration of RUNX2 (shown as αA), its truncated constructs, and the CCDαA376 mutant. The amino acid sequence comparison between ADs of PEBP2αA/RUNX2 and PEBP2αB/RUNX1 and the position of the premature stop codon in CCDαA376 (marked by an asterisk) are also shown. NLS, nuclear localization signal. (b) The protein expression pattern of RUNX2 and its deletion constructs in COS-7 cells detected by anti-αA1N and anti-αA1C antibodies. (c) Transactivation activities of RUNX2 and its deletion mutants. NIH 3T3 fibroblasts were transfected with the wild-type 1050.rOC-luc (0.5 μg) and the indicated polypeptide chain elongation factor 1α promoter (EF-BOS)-based expression plasmids (0.5 μg) with or without pEF-β2 (0.2 μg). (d) Comparison of the transactivation activities of RUNX2 and CCDαA376. Increasing amounts (0.1, 0.2, 0.4, and 0.8 μg) of pEF-αA or pEF-CCDαA376 in the absence or presence of pEF-β2 (0.2 μg) were cotransfected into NIH 3T3 cells with the wild-type 1050.rOC-luc or the mutant 1050.rOC-luc, in which all three PEBP2 binding sites were mutated.

Figure 2

The region αA(341-424) confers transactivation activity to GAL4-DBD. (a) Schematic illustration of GAL4-DBD-αA fusion constructs. The structures of Til-1 and PEBP2αA also are shown. (b) Expression of GAL4-DBD-αA fusion constructs in COS-7 cells detected by anti-GAL4-DBD antibody. (c) NIH 3T3 cells were transfected with tk-GALpx3-luc (0.4 μg) and the indicated effectors (0.2 μg).

Figure 3

Failure of CCDαA376 to interact with Smads. (a) The physical interaction between RUNX2 and Smad3 as examined by GST–pull-down assay (43). In vitro-translated [³⁵S]methionine-labeled RUNX2 and its deletion mutants were incubated with GST or GST-Smad3 conjugated to glutathione Sepharose beads. The coprecipitated samples were separated by SDS/12% PAGE. INPUT (1/10), 10% of in vitro-translated products used for binding assay was loaded. (b) In vivo interaction. COS-7 cells were cotransfected with expression plasmids coding for Flag-tagged Smad3 (for lanes 1–6) or Flag-tagged Smad1 (for lanes 7 and 8), a constitutively active form of type I TGF-β (TβR-I; for lanes 1–6) or BMP receptor (BMPR-IA; for lanes 7 and 8), and RUNX2 mutants, the structures of which are shown in Fig. 2a. RUNX2 proteins were immunoprecipitated from cell lysates with anti-Flag antibody followed by immunoblotting (WB) using anti-αA8G5 antibody. (Bottom) The coimmunoprecipitation (IP) of RUNX2 and αA(1-424) with Smads. (Top and Middle) Expression levels of individual proteins are shown.

Figure 4

Ligand-dependent interaction between endogenous RUNX2 and Smad1. Two 15-cm plates each of P19 or C2C12 cells (1 × 10⁸) were prepared. One plate each was treated with BMP-7 for 1 h and the other was mock-treated. Immunoprecipitation from 10% of the cell extracts each was performed by using either polyclonal anti-αA or polyclonal anti-Smad1. Western blotting of the immunopresipitates was performed by using monoclonal antibodies specific to RUNX2 (Upper) or Smad1 (Lower). The positions of RUNX2 (shown as αA) and Smad 1 are indicated. Nonspecific bands representing IgG also are indicated.

Figure 5

Impaired responsiveness of CCDαA376 to TGF-β signaling. P19 cells were cotransfected with the reporter plasmid of Cα (WT), Cα (TβRE-mS), Cα (TβRE-mP), or Cα (mSP) (0.2 μg) and the expression plasmids for RUNX2 (shown as αA) or CCDαA376 (0.1 μg) and Smad3/4 (0.1 μg each) with or without TβR-I (0.1 μg). Relative luciferase activities are shown as -fold induction.

Figure 6

Western blot showing expression of RUNX2 and CCDαA376 in C2C12 cells stably expressing the genes. Lane 1, mock; lane 2, RUNX2; lane 3, CCDαA376.

Figure 7

Alkaline phosphatase activities (42) induced by cooperation between Runx2 and Smads. (a, d, g, and j) Control C2C12 cells. (b, e, h, and k) C2C12 cells stably expressing Runx2. (c, f, i, and l) C2C12 cells stably expressing CCDAα376. (a–f) Infected with recombinant adenovirus expressing β-galactosidase as a control of unrelated protein. (d–f) C2C12 cells treated with BMP-7 (150 ng/ml) for 3 days. (g–i) C2C12 cells infected with recombinant adenoviruses expressing Smads 1 and 4 for 24 h and incubated for 3 more days. (j–l) C2C12 cells infected with recombinant adenoviruses expressing Smads 1 and 4 for 24 h and then treated with BMP-7 (150 ng/ml) for 3 days.