Biochemical characterization and identification of novel substrates of protein lysine methyltransferases

Abstract

The methylation of lysine side chains is a prevalent post-translational modification (PTM) of proteins, which is introduced by protein lysine methyltransferases (PKMTs). Histone methylation can have different effects on chromatin structure, lysine methylation of non-histone proteins can regulate protein/protein interactions and protein stability. For most PKMTs currently not all methylation sites are known which limits our understanding of the regulatory role of these enzymes in cells. Therefore, it is an important research aim to gain more information about the substrate spectrum of PKMTs. The identification of the substrate specificity of a PKMT is a very important step on the way to identify new PKMT methylation sites. The focus of this study was the analysis of the substrate specificity of different PKMTs by SPOT peptide arrays and based on this on the identification and validation of possible new methylation substrates. The analysis of the substrate specificity of human SUV39H2 revealed significant differences to its human homolog SUV39H1, although both enzymes methylate the same histone substrate (H3K9). SUV39H2 is more stringent than the SUV39H1, which could be demonstrated by the lack of methylation of SUV39H1 non-histone targets by SUV39H2 and by the fact that it was not possible in this study to identify non-histone substrates for SUV39H2. Kinetic studies showed that SUV39H2 prefers the unmethylated H3K9 as substrate. Moreover, it was shown that the N324K mutation of SUV39H2 which leads to a genetic disease in Labrador retrievers causes a change in folding finally leading to the inactivation of the enzyme. It had been reported by another group that the histone variant H2AX is methylated by SUV39H2. However, the sequence of H2AX K134 does not fit to the substrate specificity profile of SUV39H2 determined in the present work. Follow-up in vitro peptide and protein methylation studies indeed showed that H2AX K134 is not methylated by SUV39H2. This indicates that H2AX methylation by SUV39H2 is most probably a wrong assignment of a substrate to a PKMT. Based on already available specificity data for the SUV39H1 PKMT, the SET8 protein was validated as novel substrate in cellular studies. SET8 is a PKMT itself and it could be shown in this thesis that methylation of SET8 at residue K210 by SUV39H1 stimulated the SET8 activity. In humans, there exist different PKMTs, which methylate H3K36. For example, NSD1, NSD2 and SETD2 which were investigated in this thesis. In literature, it was shown that the oncohistone mutation K36M inactivates NSD2 and SETD2. Steady-state methylation kinetics using a peptide substrate and a K36M peptide as inhibitor revealed that NSD1 is inhibited by this histone oncomutation as well. The steady-state inhibition parameters for all enzymes showed a better binding of the PKMTs to the inhibitor peptide than to the substrate, suggesting some mechanistic similarities in target peptide interaction. The SETD2 is a methyltransferase, which is able to introduce trimethylation of H3K36. During this thesis two substrate specificity motifs of SETD2 were determined using peptide array methylation experiments. Additionally, based on the substrate specificity investigations a super-substrate at peptide and protein level was determined. Furthermore, one novel substrate (FBN1) for SETD2 was discovered and validated. The Legionella pneumophila RomA PKMT was shown previously by our collaborators to methylate H3 at K14. Based on the specificity profile of RomA determined in this study it could be shown that this enzyme methylates seven additional human non-histone proteins. Collaborators tested the methylation of one of the non-histone targets (AROS) and could demonstrate its methylation during the infection of human cells with L. pneumophila. The role of these methylation events in the infection process must be studied in future experiments.