Citrullination of DNMT3A by PADI4 regulates its stability and controls DNA methylation.
Bottom Line: We found that DNMT3A and PADI4 interact, from overexpressed as well as untransfected cells, and associate with each other's enzymatic activity.Finally, we showed that PADI4 overexpression increases DNA methyltransferase activity in a catalytic-dependent manner and use bisulfite pyrosequencing to demonstrate that PADI4 knockdown causes significant reduction of CpG methylation at the p21 promoter, a known target of DNMT3A and PADI4.Our results shed new light on how post-translational modifications might contribute to shaping the genomic CpG methylation landscape.
Affiliation: Laboratory of Cancer Epigenetics, Faculty of Medicine, Université Libre de Bruxelles, 808 route de Lennik, 1070 Brussels, Belgium email@example.com.Show MeSH
Mentions: To learn more about how PTMs affect DNMT function, we focused on the de novo DNA methyltransferase DNMT3A, about which very little is known. We first performed pull-down assays to see if this enzyme might associate with the peptidylargininedeiminase PADI4, which catalyzes conversion of protein arginine residues to citrulline. As shown in Figure 1A, GST-fused full-length PADI4 was able to bind in vitro translated (IVT) radio-labeled full-length DNMT3A (Figure 1A, lane 3). GST failed to do so (lane 2). The gel was Coomassie blue stained to check for equal loading (Figure 1A, lower part). To further substantiate the interaction between DNMT3A and PADI4, we used co-immunoprecipitation assays. (It is worth noting that we used 293T and U2OS cells for most subsequent cell studies, consistently with several previous works on PADI4 (e.g. 23,35,36). In particular, 293T cells have often been employed for PADI4 overexpression studies, while U2OS cells have also been used for PADI4 overexpression and in endogenous/RNAi experiments.) As shown in Figure 1B, when 293T cells were transfected with GAL4-tagged DNMT3A and HA-tagged PADI4, we found DNMT3A to interact with PADI4 (Figure 1B, lane 3). No signal was detected in the precipitate after transfection of cells with either the GAL4-DNMT3A-encoding or HA-PADI4-encoding plasmid alone (Figure 1B, lanes 1 and 2, respectively). It is noteworthy that the level of DNMT3A protein in the total lysate was higher when both PADI4 and DNMT3A were ectopically expressed, suggesting a potential role for PADI4 in regulating DNMT3A protein stability (Figure 1B, Input controls/GAL4-DNMT3A, lane 3) (see below). To demonstrate an interaction between the endogenous PADI4 and DNMT3A proteins, we immunoprecipitated DNMT3A from untransfected U2OS cells and found the immunoprecipitate to contain PADI4 (Figure 1C, lane 3). Control (IgG) antibodies gave only a background signal (Figure 1C, lane 2). Since DNMT3A and PADI4 are both known to interact with chromatin (19,37), we also treated the lysates with DNAse to rule out the possibility that these proteins might co-immunoprecipitate because of indirect interactions mediated solely by chromatin (Figure 1C, lane 4 and Supplementary Figure S1). We also performed endogenous CoIP between PADI4 and DNMT3A in U2OS cells as above (cf. Figure 1C), this time using synchronized cells. After treatment of cells with nocodazole or thymidine, we observed association of DNMT3A with PADI4 in different phases of the cell cycle (G1, G1/S, S and M) (Supplementary Figure S2). These observations suggest no preferential interaction of PADI4 with DNMT3A during a specific phase of the cell cycle.
Affiliation: Laboratory of Cancer Epigenetics, Faculty of Medicine, Université Libre de Bruxelles, 808 route de Lennik, 1070 Brussels, Belgium firstname.lastname@example.org.