Application of mixed peptide arrays to study combinatorial readout of chromatin modifications

The N-terminal tails of histone proteins are massively decorated with post-translational modifications (PTMs), which play important roles in the regulation of gene expression. Several highly conserved chromatin interacting proteins can bind to histone modifications in a sequence and modification specific manner employing specific reading domains. These proteins often contain several reading domains, which can cooperate in the readout of different PTMs. To gain a better insight into the combinatorial readout of PTMs, we developed a method to study the binding of double reading domains to mixed peptide arrays containing two different peptides in each spot. For that, differently modified and unmodified peptides were prepared by SPOT synthesis and solubilized. Then, two peptides were mixed and spotted onto a glass slide creating peptide spots presenting two modifications on two different peptides. Different combinations of mixed spots containing modified and unmodified peptides were generated and incubated with recombinant double reading domains to study their synergistic binding. For validation of the method, we used the well-studied BPTF subunit of the NURF chromatin-remodeling complex. BPTF contains a plant homeodomain finger (PHD) and a Bromodomain recognizing H3K4me3 and H4K16ac, respectively. We first confirmed with peptide arrays and Fluorescence Anisotropy (FA) measurements that the BPTF PHD-Bromo (PB) domain interacts specifically with the expected modifications. Using our novel tool, we observed a strong and synergistic binding only to peptide spots containing both modifications, which was lost if one of the domains was inactivated by a mutation. These data indicate that BPTF-PB simultaneously interacts with both target modifications using its PHD and Bromodomain. In agreement with the synergistic peptide interaction on mixed peptide arrays, we also show that chromatin pulldown by BPTF-PB depends on the activity of both reading domains. We conclude that mixed peptide spot arrays are a powerful, cheap and novel method for screening the combinatorial interaction space of multidomain reading proteins. Using this approach hundreds of mixed peptide spots can be prepared and tested for binding in principle allowing for an unbiased medium throughput investigation.