Bone morphogenetic protein-2 (BMP-2) transactivates Dlx3 through Smad1 and Smad4: alternative mode for Dlx3 induction in mouse keratinocytes

Abstract

Expression of the Dlx3 homeodomain gene is induced in terminally differentiated epidermal cells. Dlx3 regulates gene expression in skin and plays important roles in patterning of the embryonic ectoderm through differential sensitivity to bone morphogenetic protein (BMP) signaling. We analyzed the expression of BMP family members in murine keratinocytes; BMP-2 is expressed in proliferative basal and differentiated suprabasal keratinocytes. BMP-2 induced transcription of Dlx3 within 12 h of treatment of keratinocytes cultured in vitro. We proceeded to delineate the BMP-2-responsive region to an area between -1917 and -1747 in the Dlx3 promoter. Gel shift assays with recombinant Smad1 and Smad4 demonstrated that this DNA fragment (-1917 to -1747) was competent in the formation of protein-DNA complexes. By deletion and mutational analyses we localized a Smad1/Smad4-binding site containing a GCAT motif, which showed similarity to other TGF-beta family responsive elements. Supershift assays with keratinocyte nuclear extracts and antibodies against members of the Smad family showed that this motif was able to form a complex with Smad1. Mutation of the Smad1/Smad4-binding site inhibited transcriptional activation of the Dlx3 gene by BMP-2. In the hair follicle, where Dlx3 is expressed in the hair matrix cells, BMP-2 also activates Dlx3 transcription. These results provide a possible mechanism of action for the BMP signaling pathway on the regulation of Dlx3.

Figure 1

BMP-2 is expressed in mouse keratinocytes and activates the transcription of Dlx3. (A) RT–PCR analysis was performed with cDNA made from total RNA prepared from mouse keratinocytes cultured in 0.05, 0.12 or 1.4 mM Ca²⁺ using specific primers for BMP-2. (Right) RT–PCR analysis of cDNA from basal and suprabasal cells of mouse neonatal skin separated on a Percoll gradient. For normalization of the amount of RNA used for RT–PCR, GAPDH RT–PCR products are shown on the bottom line. (B) Total RNA was prepared from mouse keratinocytes grown in 0.05 mM Ca²⁺ which were treated with 10, 50 or 300 ng/ml BMP-2 for 24 h and then RT–PCR analysis was performed with specific primers for Dlx3 (left). RNA from keratinocytes was treated with 10 ng/ml BMP-2 for 12, 24, 36 and 48 h (center) and 3, 6, 9 and 12 h (right). (C) Mouse keratinocytes were incubated with 20 µg/ml cycloheximide (CHX) for 30 min and cultured in the absence or presence of 10 ng/ml BMP-2 for 24 h and then RT–PCR analysis was performed with Dlx3-specific primers.

Figure 2

Delineation of the BMP-2-responsive sequence in the Dlx3 promoter. (A) Schematic diagram of each 5′-end deletion construct of the Dlx3 promoter. (B) Primary mouse keratinocytes were transfected with 1 µg each deletion construct and treated with 10 ng/ml BMP-2 for 24 h. Forty-eight hours after transfection, cells were harvested and CAT activity was measured. CAT activity was normalized to the protein concentration measured by the Bradford assay. The bars represent the average normalized CAT activity of duplicate plates from three experiments for each construct.

Figure 3

Smad1 and Smad4 bind to the BMP-2-responsive region in the Dlx3 promoter. (A) Gel shift assays were performed using increasing amounts of GST–Smad1, GST–Smad4 and GST–Smad3 fusion proteins and radiolabeled DNA fragment (–1917 to –1747). The GST fusion proteins used in gel shift assays are shown in the center panel (S1, Smad1; S3, Smad3; S4, Smad4). The right panel shows a gel shift assay to demonstrate the binding capability of Smad3. The arrow indicates the complex formed by GST–Smad1 or GST–Smad4 and the DNA fragment. The putative Smad-binding motifs are indicated by underlined and bold letters. (B) Gel shift assays were performed with four overlapping oligonucleotides covering the Smad-binding region (–1917 to –1756) and GST–Smad1 or GST–Smad4. The right panel shows a gel shift assay using SBE4 and Smad1 to determine the specificity of each complex. Each oligonucleotide sequence is shown with the putative Smad-binding motifs indicated by bold and underlined letters.

Figure 4

Smad1 and Smad4 bind to the SBE4-1 element. (A) Gel shift assays were performed with radiolabeled SBE2 and GST–Smad1 or GST–Smad4. As cold competitors, SBE2-1 and SBE2-2 were used at a 25-fold molar excess. (B) Gel shift assays were performed with radiolabeled SBE4 and recombinant GST–Smad1 or GST–Smad4. SBE4-1 and SBE4-2 were used as cold competitors at a 25-fold molar excess. (C) The SBF4-1 oligonucleotide was used as probe for gel shift assays with nuclear extract from mouse keratinocytes untreated or treated with BMP-2. In the supershift assay, each anti-Smads antibody was added to the reaction before addition of probe. The supershifted complex is indicated with an arrow.

Figure 5

A GCAT sequence is responsible for Smad1 and Smad4 binding and activation of Dlx3 expression by BMP-2. (A) Gel shift analysis was performed with radiolabeled mutated SBE4-1 oligonucleotides and GST–Smad1 or GST–Smad4. Each mutated oligonucleotide sequence is shown on the left. The GCAT sequence is boxed. (B) Schematic diagram of constructs used in the CAT assays. (C) Primary mouse keratinocytes were transfected independently with 1 µg p1917CAT, p1917mCAT, p1797CAT or p1747CAT and treated with 10 ng/ml BMP-2 for 24 h. Forty-eight hours after transfection, cells were harvested and CAT activity was measured. (D) The GCAT mutant construct p1917mCAT and the wild-type, p1917CAT, were co-transfected into primary mouse keratinocytes with Smad1/Smad4 expression vectors. Fold activation was calculated based on the CAT activity of each construct co-transfected with the parental vector.

Figure 6

BMP-2 induced the transcription of Dlx3 in hair follicle cells. (A) RT–PCR analysis was performed with specific primers for Dlx3 using total RNA from mouse hair follicle cells treated with 10, 50, 100 or 200 ng/ml BMP-2 for 24 h. (B) Primary mouse hair follicle cells were transfected independently with 1 µg p2317CAT, p1917CAT, p1917mCAT or p1565CAT and treated with 10 ng/ml BMP-2 for 24 h. Forty-eight hours after transfection, cells were harvested and CAT activity was measured.