Histone peptide microarray screen of chromo and Tudor domains defines new histone lysine methylation interactions

Abstract

Background: Histone posttranslational modifications (PTMs) function to regulate chromatin structure and function in part through the recruitment of effector proteins that harbor specialized "reader" domains. Despite efforts to elucidate reader domain-PTM interactions, the influence of neighboring PTMs and the target specificity of many reader domains is still unclear. The aim of this study was to use a high-throughput histone peptide microarray platform to interrogate 83 known and putative histone reader domains from the chromo and Tudor domain families to identify their interactions and characterize the influence of neighboring PTMs on these interactions.

Results: Nearly a quarter of the chromo and Tudor domains screened showed interactions with histone PTMs by peptide microarray, revealing known and several novel methyllysine interactions. Specifically, we found that the CBX/HP1 chromodomains that recognize H3K9me also recognize H3K23me2/3-a poorly understood histone PTM. We also observed that, in addition to their interaction with H3K4me3, Tudor domains of the Spindlin family also recognized H4K20me3-a previously uncharacterized interaction. Several Tudor domains also showed novel interactions with H3K4me as well.

Conclusions: These results provide an important resource for the epigenetics and chromatin community on the interactions of many human chromo and Tudor domains. They also provide the basis for additional studies into the functional significance of the novel interactions that were discovered.

Keywords: Chromatin; Chromodomain; Histone methylation; Peptide microarray; Tudor domain.

Fig. 1

SPIN1 triple Tudor domain interacts with H4K20me3 as well as H3K4me3. a Heat map showing the relative binding detected for each of the indicated domains on the peptide microarray platform. Data represent the average of two independent arrays relative to the most intense binding signal within the indicated set of peptides. b Scatter plot of the relative binding of SPIN1 triple Tudor domain from two independent peptide arrays. H3K4me3-containing peptides are shown in red and H4K20me3-containing peptides are shown in blue. All other peptides are shown in black. c Representative picture of a section of the peptide microarray for SPIN1 triple Tudor domain. The left panel shows both the green (peptide) and red (protein binding) fluorescent channels, while the right panel depicts only the red fluorescence channel for clarity. Positive antibody controls are outlined in white and the positive interaction with the H4K20me3 peptide is outlined in yellow. Full array images are shown in Additional file 3: Figure S1. d Western blot results of peptide pull-down experiments with purified full-length Spindlin family members. The input is shown in Lane 1 and the corresponding bound fraction is shown in Lanes 2–8. e Western blot results of peptide pull-down experiments with whole cell lysates derived from transiently transfected HEK 293T cells (GFP-SPIN1, 2A, 2B, 3 and 4). f Western blot results of peptide pull-down experiments with purified SPIN1 wild type (SPIN1^WT) or aromatic cage mutant in the second Tudor domain (SPIN1^Y170A) demonstrating a loss of H3K4me3 and H4K20me3 in the mutant

Fig. 2

Chromodomains interact with H3K23me2/3 in addition to H3K9me1/2/3. a Heat map showing the relative binding detected for each of the indicated domains on the peptide microarray platform. Data represent the average of two independent arrays relative to the most intense binding signal within the indicated set of peptides. b Western blot results of peptide pull-downs performed with the indicated GST-tagged domain and histone peptide. The input is shown in Lane 1 and the bound fraction is shown in Lanes 2–13