Regulation of AML1/RUNX1 Function by Protein Arginine Methyltransferase 4 (PRMT4) in Myeloid Differentiation

Abstract 549

RUNX1 (also known as AML1) is the DNA binding component of the Core Binding Factor (CBF)-transcriptional regulatory complex, which plays an important role in hematopoiesis. Upon binding to the common binding sequence -PyGpyGGTPy (Py = pyrimidine) in the regulatory regions of promoters and enhancers of its target genes, RUNX1 acts either as an activator or a repressor, depending on promoter context and its interacting partners. Thus, modulation of the network of RUNX1 interactions can influence hematopoiesis. However, how RUNX1 selects one set of partners over another to assemble a functional complex is largely unknown. Posttranslational modifications, including ubiquitination, phosphorylation, acetylation and methylation, present a viable mean to fine-tune its functions.

Here we shown that RUNX1 is arginine methylated at a specific residue, R223, by PRMT4, a type I arginine methyltransferase generally thought of as a co-activator molecule. We hypothesized that arginine methylation of RUNX1 by PRMT4 affects its protein-protein interactions, therefore, to identify proteins that specifically interact with unmethylated and/or methylated-R223 RUNX1, in an unbiased manner, we performed a peptide pull-down experiment, using a methyl-R223 RUNX1 peptide and an unmodified RUNX1 peptide as bait, following by mass spectrometry analysis. We identified several proteins that preferentially interacted with the R223 methyl peptide, but focused on a novel interacting protein, DPF2 (double PhD Finger 2), which is a widely expressed member of the d4 protein family, characterized by the presence of a tandem plant-homodomain (PHD domain). We confirmed the specific interaction between methylated-RUNX1 with DPF2 in vivo by immunoprecipitation. We generated an antibody specific for the R223 methylated-RUNX1 protein, and found that RUNX1 methylation decreases during the myeloid differentiation of human CD34+ haematopoietic stem/progenitor cells (HSPCs), without a change in the total level of RUNX1 protein, and this occurred co-incident with a downregulation of PRMT4 protein expression. Having determined that PRMT4 expression declines during myeloid differentiation, we examined the role of PRMT4 in this process, using short hairpin RNAs to knockdown PRMT4 expression in CD34+ cells. Knockdown of PRMT4 accelerates the myeloid differentiation of the cells, whereas overexpression of PRMT4 in human CD34+ cells blocked their myeloid differentiation. When analyzing the expression of several “master” regulators of myeloid differentiation, we identified microRNA-223, a myeloid specific microRNA, as a common target gene of PRMT4 and RUNX1. Furthermore, we have found that by promoting the assembly of a functional complex containing R223 methylRUNX1 and DPF2 at the transcriptional regulatory region of the microRNA-223 promoter, PRMT4 can control miR-223 expression and myeloid differentiation. We have verified the role of DPF2 in this process, as DPF2 represses miR-223 expression and loss of DPF2 promotes myeloid differentiation. Thus, DPF2 acts in a common pathway with PRMT4 to regulate myeloid differentiation.

In conclusion, our study elucidates a novel mechanism, where the arginine methylation of RUNX1 regulates its recruitment of interacting partner(s). In addition to demonstrating that PRMT4 can trigger repression of gene expression, we have identified a novel role for PRMT4 (aka CARM1) in myeloid differentiation.