Son and Its Alternatively Spliced Isoforms Control MLL Complex-Mediated H3K4me3 and Transcription of Leukemia-Associated Genes

SON is a poorly characterized nuclear protein that is particularly abundant in hematopoietic cells/organs and embryonic stem cells. This protein was recently identified as an SR-like splicing co-factor required for proper cell cycle progression (Mol. Cell. 2011, 42:185) and maintenance of stem cell pluripotency (Nat. Cell. Biol. 2013, 15:1141). Although SON’s function in RNA splicing of several genes, such as TUBG1, AKT1, OCT4 and HDAC6, was recently highlighted, SON was originally identified as a DNA-binding protein, suggesting a potential function in transcription.  Our recent study found that SON represses the promoter of the miR-23a~27a~24-2 cluster in hematopoietic cells (J. Biol. Chem. 2013, 288:5381). However, whether SON is directly associated with specific chromosomal loci in mammalian genome and how SON represses gene transcription are completely unknown.

In order to investigate SON’s function in genome-wide DNA-binding and gene regulation, we performed chromatin immunoprecipitation-sequencing (ChIP-seq) in K562 cells using SON antibodies. Analysis of the genomic distribution of SON ChIP-seq peaks demonstrated that most of SON-binding sites are located near the promoter regions of target genes which include transcription factors, signaling mediators and cell cycle regulators. Interestingly, SON-binding regions near transcription start sites precisely overlap with the locations of tri-methylation of histone H3 lysine 4 (H3K4me3) which is associated with the open chromatin status and gene activation. Our ChIP-qPCR analyses revealed that knockdown of SON causes enhanced recruitment of the mixed lineage leukemia (MLL) complex, a writer of H3K4me3, to the SON target chromatin, significantly increasing H3K4me3 levels. SON depletion also leads to enhanced protein-protein interactions between the MLL complex components. These results indicate a repressive function of SON in H3K4me3 and MLL complex assembly.  Surprisingly, our immunoprecipitation (IP) with SON antibody revealed that SON interacts with menin, an MLL complex component critical for MLL function in oncogenesis. Furthermore, SON overexpression increased the menin-SON complex and simultaneously decreased menin interaction with MLL. These findings demonstrate an inhibitory effect of menin-SON interaction on menin-MLL interaction.  

In addition to full-length SON (SON F), two splice variants of SON (SON B and E), which are C-terminus-truncated forms, have been predicted in genome databases. To address the functional significance of SON splice variants, we examined whether SON splice variants are differentially expressed in the condition of acute myeloid leukemia (AML). Interestingly, the expression levels of alternatively spliced SON isoforms, but not full-length SON, were significantly increased in human AML patient bone marrow and peripheral blood cells as well as in mouse AML1-ETO9a leukemic blasts. Our data also demonstrated that the short isoforms of SON retain its DNA-binding ability, thereby competing with the full-length SON for target chromatin interaction. However, unlike the full-length SON, the short isoforms lack the menin-binding ability and could not inhibit MLL complex assembly. We further demonstrated that short isoforms of SON can abrogate the full-length SON function in lowering the H3K4me3 level, and overexpression of the short isoform enhanced MLL complex assembly.

Taken together, our study reveals that the MLL complex activity is competitively regulated by full-length SON and its alternatively spliced isoforms, and that target genes of MLL-menin are aberrantly controlled by overexpressed “short SON” in AML patients. Furthermore, our findings strongly suggest the significant roles of SON in transcriptional regulation of multiple genes associated with leukemia, stem cell maintenance and survival signaling through histone modification near transcription start sites.