Discovery of NSD2 non-histone substrates and design of a super-substrate.

The human protein lysine methyltransferase NSD2 catalyzes dimethylation at H3K36. It has very important roles in development and disease but many mechanistic features and its full spectrum of substrate proteins are unclear. Using peptide SPOT array methylation assays, we investigate the substrate sequence specificity of NSD2 and discover strong readout of residues between G33 (-3) and P38 (+2) on H3K36. Unexpectedly, we observe that amino acid residues different from natural ones in H3K36 are preferred at some positions. Combining four preferred residues led to the development of a super-substrate which is methylated much faster by NSD2 at peptide and protein level. Molecular dynamics simulations demonstrate that this activity increase is caused by distinct hyperactive conformations of the enzyme-peptide complex. To investigate the substrate spectrum of NSD2, we conducted a proteome wide search for nuclear proteins matching the specificity profile and discovered 22 peptide substrates of NSD2. In protein methylation studies, we identify K1033 of ATRX and K819 of FANCM as NSD2 methylation sites and also demonstrate their methylation in human cells. Both these proteins have important roles in DNA repair strengthening the connection of NSD2 and H3K36 methylation to DNA repair.

Distinct specificities of the HEMK2 protein methyltransferase in methylation of glutamine and lysine residues.

The HEMK2 protein methyltransferase has been described as glutamine methyltransferase catalyzing ERF1-Q185me1 and lysine methyltransferase catalyzing H4K12me1. Methylation of two distinct target residues is unique for this class of enzymes. To understand the specific catalytic adaptations of HEMK2 allowing it to master this chemically challenging task, we conducted a detailed investigation of the substrate sequence specificities of HEMK2 for Q- and K-methylation. Our data show that HEMK2 prefers methylation of Q over K at peptide and protein level. Moreover, the ERF1 sequence is strongly preferred as substrate over the H4K12 sequence. With peptide SPOT array methylation experiments, we show that Q-methylation preferentially occurs in a G-Q-X3 -R context, while K-methylation prefers S/T at the first position of the motif. Based on this, we identified novel HEMK2 K-methylation peptide substrates with sequences taken from human proteins which are methylated with high activity. Since H4K12 methylation by HEMK2 was very low, other protein lysine methyltransferases were examined for their ability to methylate the H4K12 site. We show that SETD6 has a high H4K12me1 methylation activity (about 1000-times stronger than HEMK2) and this enzyme is mainly responsible for H4K12me1 in DU145 prostate cancer cells.