Abstract
MLL2 and MLL3 histone lysine methyltransferases are conserved components of COMPASS-like co-activator complexes. In vertebrates, the paralogous MLL2 and MLL3 contain multiple domains required for epigenetic reading and writing of the histone code involved in hormone-stimulated gene programming, including receptor-binding motifs, SET methyltransferase, HMG and PHD domains. The genes encoding MLL2 and MLL3 arose from a common ancestor. Phylogenetic analyses reveal that the ancestral gene underwent a fission event in some Brachycera dipterans, including Drosophila species, creating two independent genes corresponding to the N- and C-terminal portions. In Drosophila, the C-terminal SET domain is encoded by trithorax-related (trr), which is required for hormone-dependent gene activation. We identified the cara mitad (cmi) gene, which encodes the previously undiscovered N-terminal region consisting of PHD and HMG domains and receptor-binding motifs. The cmi gene is essential and its functions are dosage sensitive. CMI associates with TRR, as well as the EcR-USP receptor, and is required for hormone-dependent transcription. Unexpectedly, although the CMI and MLL2 PHDf3 domains could bind histone H3, neither showed preference for trimethylated lysine 4. Genetic tests reveal that cmi is required for proper global trimethylation of H3K4 and that hormone-stimulated transcription requires chromatin binding by CMI, methylation of H3K4 by TRR and demethylation of H3K27 by the demethylase UTX. The evolutionary split of MLL2 into two distinct genes in Drosophila provides important insight into distinct epigenetic functions of conserved readers and writers of the histone code.
