Direct interaction between the PRDM3 and PRDM16 tumor suppressors and the NuRD chromatin remodeling complex

Abstract

Aberrant isoform expression of chromatin-associated proteins can induce epigenetic programs related to disease. The MDS1 and EVI1 complex locus (MECOM) encodes PRDM3, a protein with an N-terminal PR-SET domain, as well as a shorter isoform, EVI1, lacking the N-terminus containing the PR-SET domain (ΔPR). Imbalanced expression of MECOM isoforms is observed in multiple malignancies, implicating EVI1 as an oncogene, while PRDM3 has been suggested to function as a tumor suppressor through an unknown mechanism. To elucidate functional characteristics of these N-terminal residues, we compared the protein interactomes of the full-length and ΔPR isoforms of PRDM3 and its closely related paralog, PRDM16. Unlike the ΔPR isoforms, both full-length isoforms exhibited a significantly enriched association with components of the NuRD chromatin remodeling complex, especially RBBP4. Typically, RBBP4 facilitates chromatin association of the NuRD complex by binding to histone H3 tails. We show that RBBP4 binds to the N-terminal amino acid residues of PRDM3 and PRDM16, with a dissociation constant of 3.0 μM, as measured by isothermal titration calorimetry. Furthermore, high-resolution X-ray crystal structures of PRDM3 and PRDM16 N-terminal peptides in complex with RBBP4 revealed binding to RBBP4 within the conserved histone H3-binding groove. These data support a mechanism of isoform-specific interaction of PRDM3 and PRDM16 with the NuRD chromatin remodeling complex.

DNA sequence repeats identify numerous Type I restriction‐modification systems that are potential epigenetic regulators controlling phase‐variable regulons; phasevarions

“…71 Our analysis of only full-length annotated genes likely leads us to miss many hsdM and hsdR genes that contain SSRs that are phase varied OFF, and which lead to genes being annotated as out-of-frame. 9 As such, these sequences need to be confirmed by methods that can accurately discern the sequence of SSR tracts, such as PacBio SMRT long-read sequencing technology. The use of short read next-generation sequencing (NGS) technologies that rely on mapping short (<200 bp) reads to reference genomes, and assembling areas where SSRs are present often leads to an underestimation of the repeat tract length due to collapsing the tract down by the assembly software, or alignment to multiple places in the genome 72 as assembly software cannot distinguish between sequences.…”

“…Each hsdS gene is made up of two target recognition domains (TRDs; TRD 1 in red, and TRD 2 in green), with the SSR tract located between the two TRDs (gray boxes). We therefore only defined a Type I hsd gene as "phase-variable" if the repeat tract in an open reading frame (ORF) was at least nine bases long for mononucleotide repeat tracts (eg, G [9] ), five repeats long for dinucleotide repeats (eg, GA [5] ), and three repeat units long for tetra-, penta-, hexa-, and octanucleotide repeat tracts (eg, AGCC [3] ). These likely dimerise via the C-terminal coiled coil region in each truncated HsdS subunit to form a functional HsdS protein.…”

“…[2][3][4][7][8][9][10]17 However, the sole example of an hsdS gene containing SSRs and showing phase-variable expression, that of the NgoAV system in N gonorrhoeae strain FA1090, 31 demonstrated that rather than ON-OFF switching, variation in the length of the SSR contained in the encoding hsdS gene lead to production of a full-length or a truncated HsdS specificity subunit, which resulted in different methyltransferase specificities. [2][3][4][7][8][9][10]17 However, the sole example of an hsdS gene containing SSRs and showing phase-variable expression, that of the NgoAV system in N gonorrhoeae strain FA1090, 31 demonstrated that rather than ON-OFF switching, variation in the length of the SSR contained in the encoding hsdS gene lead to production of a full-length or a truncated HsdS specificity subunit, which resulted in different methyltransferase specificities.…”

“…The human respiratory pathogen H. influenzae encodes an hsdM gene which contains a CGAGA repeat tract (CGAGA [3][4][5][6][7][8][9][10] ). For example, multiple strains of the bovine pathogen Mannheimia haemolytica encode an hsdM gene which contains an ACAGC repeat tract (ACAGC ).…”

“…1 Many host-adapted bacterial pathogens contain phase-variable genes, and these often encode surface associated virulence factors that are subjected to periodic selection, such as iron acquisition systems, 2,3 pili, 4 adhesins, 5,6 and lipooligosaccharide biosynthetic genes. 9 ON-OFF switching of genes encoding Type III methyltransferase expression leads to differential regulation of multiple genes in systems known as phasevarions (phase-variable regulon; Srikhanta et al 2005). We recently characterized the distribution of phase-variable Type III mod genes in the REBASE database of R-M systems, and demonstrated that 17.4% of all Type III mod genes contain SSRs.…”