Developing tools for, and characterization of a family of novel human protein methyltransferases

Abstract

Post-translational modifications (PTM) of proteins are nature’s way of expanding its inventory while keeping organism’s genomes compact. Many diverse PTM’s are well known, such as phosphorylation, acetylation, glycocylation and methylation. Methylation is the process of transferring a methyl group from a methyl donor such as SAM, to various types of biomolecules, including proteins. The proteins that perform this task, denoted methyltransferases (MTases), are well studied in the case of histone proteins and their role in gene regulation. However non-histone MTases have also been characterized recently, such as the protein MTase METTL21D which is a Class I SAM-dependent methyltransferase. This led to the discovery of 9 other proteins with high structural and sequence similarity and together these 10 MTases make up the human members of the Family 16 Methyltransferases.

In this study, a member of this family, METTL21A was investigated as well as the sub-cellular localization of all 10 human family members. Antibodies for both METTL21A and METTL21D were generated, which were sufficient to recognize the recombinant protein, however endogenous protein was not detected, indicating that the MTases are in low abundance within the cell.

In attempt to identify key residues involved in the METTL21A methylation process, several well-conserved amino acids were mutated. When performing in vitro MTase assays on these mutants, three of these residues appeared to be crucial to the MTase activity. Also, the variable N-terminus of the protein was deleted to test the hypotheses that it is important in substrate recognition. This abolished all activity of the enzyme, suggesting that the hypothesis is correct.

METTL21A’s substrates are members of the human HSP70 chaperone family, where the identified substrate, HSPA1, is a stress-inducible chaperone and the HSPA8 is constitutively expressed. When the wt and mutant enzymes were assayed with either substrate, HSPA8 appeared to be the better of the two when the enzyme was mutated.