Epigenetic programming of Dnmt3a mediated by AP2α is required for granting preadipocyte the ability to differentiate

Abstract

Adipogenesis has an important role in regulating energy homeostasis in mammals. 3T3-L1 preadipocytes have been widely used as an in vitro model for analyzing the molecular mechanism of adipogenesis. Previous reports indicated that the stage of contact inhibition (CI), through which the proliferating cells exit from the cell cycle, was required for granting preadipocyte the ability to differentiate. While this kind of the granting mechanism remains elusive. In the present study, we showed that DNA (cytosine-5) methyltransferase 3a (Dnmt3a) was upregulated at both the mRNA and protein level during the CI stage, and resulted in increasing promoter methylation of adipogenic genes. We further identified that the expression of Activator protein 2α (AP2α), a member of the transcription factor activator protein 2 (AP2) family, was highly correlated with the expression of Dnmt3a during the CI stage. In addition, we showed that AP2α transcriptionally upregulated Dnmt3a by directly binding to its proximal promoter region. Importantly, treatment of 3T3-L1 preadipocytes with AP2α-specific siRNAs inhibited the preadipocyte differentiation in a stage-dependent manner, supporting the conclusion that AP2α has an important role during the CI stage. Furthermore, overexpression of Dnmt3a partially rescued the impairment of adipogenesis induced by AP2α knockdown. Collectively, our findings reveal that AP2α is an essential regulator for granting preadipocyte the ability to differentiate through the upregulation of Dnmt3a expression during the CI stage.

Figure 1

Dnmt3a is upregulated by AP2α during the CI stage. (a) Illustration of the murine Dnmt3a promoter (−1.7 to +0.1 kb) is shown and two reported AP2α binding sites were identified in the proximal region. (b) Correlation coefficient analysis of AP2α and Dnmt3a mRNA expression profiles during the CI stage. The mRNA expressions of AP2α and Dnmt3a at different time points during the CI stage were determined by real-time PCR and normalized to glyceraldehyde-3-phosphate dehydrogenase. (c) Real-time PCR analysis of the expression of Dnmt3a in AP2α knockdown cells (si-1 and si-2) at ci48 h. The results were analyzed as means±S.D. (n=3). Statistical significance is indicated: ***P<0.001. (d) Western blotting analysis of the expression of Dnmt3a in AP2α knockdown cells at ci48 h. Cyclin-dependent kinase 4 served as the loading control. Quantification analysis of the relative protein level of Dnmt3a in AP2α knockdown cells was shown on the right panel. The results were analyzed as means±S.D. (n=3). Statistical significance is indicated: **P<0.01. (e) A luciferase reporter plasmid of the murine Dnmt3a promoter (−1.7 to +0.1 kb) and an HA-AP2α cDNA plasmid were constructed for dual-luciferase assay. Results were expressed as firefly luciferase activity normalized to renilla luciferase activity. Error bars indicate S.D. (n=3). Statistical significance is indicated: *P<0.05, **P<0.01. Gradient expression of HA-AP2α was detected by western blotting (bottom panel)

Figure 2

Dnmt3a is upregulated with increasing promoter methylation of adipogenic transcription factors during the stage of CI. (a) The mRNA expression of Dnmt3a during the process of 3T3-L1 preadipocyte differentiation. The mRNA level was determined by real-time PCR and normalized to glyceraldehyde-3-phosphate dehydrogenase mRNA. 'cyc' indicates the stage of cycling cells, 'ci' indicates the CI stage, 'mdi' indicates the stage of differentiation induction by hormone cocktail. Results were expressed as means±S.D. (n=3). Statistical significance is indicated: **P<0.01, ***P<0.001, NS=non-significant. (b) The protein expression of Dnmt3a was detected at indicated time points as described in (a) (left panel). Cyclin-dependent kinase 4 was used as the loading control. Relative quantification analysis of band densities was shown on the right panel. Data were presented as means±S.D. (n=3). Statistical significance is indicated: **P<0.01, ***P<0.001, NS=non-significant. (c) For promoter methylation analysis of adipogenic transcription factors, CpG island analysis of the 5' flanking and 5' untranslated sequences of murine C/EBPβ, C/EBPα, PPARγ and Egr2 are shown. The red vertical line stands for a CpG site. The region indicated by arrows stands for the sequence for bisulfite sequencing. (d) Bisulfite sequencing results of murine C/EBPβ, C/EBPα, PPARγ and Egr2 from 3T3-L1 preadipocytes under the indicated conditions. Each column stands for a CpG site. The methylation percentage of each CpG site was shown by blue in each column. And the non-methylation percentage of each CpG site was shown by yellow in each column. Data were presented as mean from three independent samples

Figure 3

AP2α knockdown impairs both genome-wide DNA methylation and the promoter methylation of adipogenic TFs during the CI stage. (a) 5-methylcytosine immunofluorescence of 3T3-L1 preadipocytes at ci48 h. The cells were transiently transfected with either scramble siRNA (ctrl) or AP2α siRNAs (si-1 and si-2) at ci0 h and fixed at ci48 h for 5-methylcytosine immunofluorescence. Scale bars: 10 μm. (b) Quantification analyses of total nuclear 5-methylcytosine densities in AP2α knockdown cells at ci48 h. Data were collected at the same voltage and were presented as mean±S.D. (n=3). Statistical significance is indicated: ***P<0.001. (c) DNA methylation status of C/EBPβ, C/EBPα, PPARγ and Egr2 promoters. 3T3-L1 preadipocytes were transiently transfected with either scramble siRNA (ctrl) or AP2α siRNAs (si-1 and si-2) at ci0 h and collected at ci48 h. Genomic DNA samples were extracted and then subjected to bisulfite sequencing. Each column stands for a CpG site. The methylation percentage of each CpG site was shown by blue in each column. And the non-methylation percentage of each CpG site was shown by yellow in each column

Figure 4

AP2α transactivates the Dnmt3a promoter by directly binding to a small proximal promoter region. (a) Schematic diagram of promoter segments of the murine Dnmt3a gene (−1.7 to +0.1 kb) inserted in the pGL3 basic luciferase vector. (b) A series promoter segments of the Dnmt3a gene were co-transfected with pRL-TK (Renilla) in 3T3-L1 preadipocytes at ci0 h. Results were expressed as firefly luciferase activity normalized to renilla luciferase activity. Error bars indicate S.D. (n=3). Statistical significance is indicated: ***P<0.001. (c) ChIP analysis of AP2α binding to the target region on the Dnmt3a promoter (−0.6 to −0.4 kb). Chromatin samples of 3T3-L1 preadipocytes at ci48 h were subjected to ChIP assays by using an AP2α antibody and normal rabbit IgG as a control. An upstream region (−1.1 to −0.9 kb) and a downstream region (−0.1 to +0.1 kb) in the Dnmt3a promoter were used as negative controls. (d) EMSA assay on the left panel (lane 1–12) showed the binding of endogenously expressed AP2α to the target region on the Dnmt3a promoter under the condition of biotin-labeled DNA probes and cold competitors. The right panel (lane 13–16) showed the supershift (SS) results of endogenously expressed AP2α in 3T3-L1 preadipocytes at ci48 h by anti-AP2α antibody. Addition of anti-AP2α antibody to the reaction resulted in the formation of the SS complex in lane 15. (e) Inactivation of the Dnmt3a promoter by mutating AP2α binding site in the Dnmt3a promoter. 3T3-L1 preadipocytes at ci0 h were co-transfected with a luciferase reporter plasmid of the Dnmt3a promoter (wt or mut) and pRL-TK. Results were expressed as firefly luciferase activity normalized to renilla luciferase activity. Error bars indicate S.D. (n=3). Statistical significance is indicated: ***P<0.001

Figure 5

Overexpression of Dnmt3a rescues the impairment of adipogenesis induced by AP2α knockdown. (a) 3T3-L1 preadipocytes at ci0 h were transiently transfected with either scramble siRNA (ctrl) or AP2α-specific siRNAs (si-1 and si-2), and then induced with hormone cocktail at ci48 h. On day 8, the cells were stained with Oil-Red-O (left panel), and then extracted with isopropanol and measured at OD_510 nm. The data were presented as means±S.D. (n=3) and shown on the right panel. Statistical significance is indicated: *P<0.01. (b) Western blotting analysis of adipogenic factors on the cells induced by MDI for 8 days. (c) For rescue study, 3T3-L1 cells at ci0 h were transiently co-transfected with three pairs of plasmids (scramble siRNA (ctrl) and GFP plasmid, AP2α-specific siRNA (si-1) and GFP plasmid, AP2α-specific siRNA (si-1) and Dnmt3a plasmid), respectively. The cells were then induced with hormone cocktail after 48 h transfection. On day 8 of MDI induction, the cells were stained with Oil-Red-O (top panel) and extracted with isopropanol and measured at OD_510 nm (bottom panel). The data were presented as means±S.D. (n=3). (d) Western blotting analysis of adipogenic factors on the cells induced by MDI for 8 days. (e) The expressions of Dnmt3a and AP2α were analyzed with western blotting. Cyclin-dependent kinase 4 served as the loading control