Histone lysine methyltransferases (HKMTs) catalyse the post-translational methylation of
lysine residues. G9a, a HKMT, was found to be overexpressed in several cancer types, where
elevated H3K9me2 levels suppress the expression of important tumor suppressor genes,
and is associated with metastasis and poor prognosis. G9a specifically dimethylates H3K9
(H3K9me2), as well as non-histone targets, such as p53 and HIF . While several potent
inhibitors targeted at the substrate site of HKMTs have been reported, there is a substantial
shortage in co-factor competitive inhibitors for G9a. Co-factor competitive probes, however
would allow to understand the effect of G9a inhibition on chromatin binding and on H3K9,
as well as non-histone methylation, depending on the drug binding site.
Herein, we used three different approaches to develop a novel co-factor competitive probe
for G9a. An initial structural based approach to obtain a crystal structure of G9a in complex
with the previously reported co-factor competitive inhibitor BIX-01338, was unsuccessful
while co-crystallisation of G9a with the SAM analogue sinefungin gave a crystal structure at
1.8  . As a second method we performed a fragment screening of 2500 fragments, yielding
some potential promising candidates. These hits, however, could not be further validated.
This necessitated a shift in focus of the project to understand the effect of different molecular
scaffolds towards G9a binding and affinity. The need of an adenine or adenosine like core
structure to ensure successful co-factor site binding has been discussed in recent publications
for other HKMTs. Based on two different co-factor competitive inhibitors published for
the HKMTs SMYD2 and MLL1, three different scaffolds were investigated to establish the
determinants of a SAM inspired core to allow binding, as well as the role of attached linkers
and their potential to induce further important protein-ligand interactions. This approach
resulted in three promising candidates for further investigation, with binding affinities in a low
micromolar range for G9a, as well as potential candidates with affinities in mid-micromolar
range for G9a.