ChIP-less analysis of chromatin states

Abstract

Background: Histone post-translational modifications (PTMs) are key epigenetic regulators in chromatin-based processes. Increasing evidence suggests that vast combinations of PTMs exist within chromatin histones. These complex patterns, rather than individual PTMs, are thought to define functional chromatin states. However, the ability to interrogate combinatorial histone PTM patterns at the nucleosome level has been limited by the lack of direct molecular tools.

Results: Here we demonstrate an efficient, quantitative, antibody-free, chromatin immunoprecipitation-less (ChIP-less) method for interrogating diverse epigenetic states. At the heart of the workflow are recombinant chromatin reader domains, which target distinct chromatin states with combinatorial PTM patterns. Utilizing a newly designed combinatorial histone peptide microarray, we showed that three reader domains (ATRX-ADD, ING2-PHD and AIRE-PHD) displayed greater specificity towards combinatorial PTM patterns than corresponding commercial histone antibodies. Such specific recognitions were employed to develop a chromatin reader-based affinity enrichment platform (matrix-assisted reader chromatin capture, or MARCC). We successfully applied the reader-based platform to capture unique chromatin states, which were quantitatively profiled by mass spectrometry to reveal interconnections between nucleosomal histone PTMs. Specifically, a highly enriched signature that harbored H3K4me0, H3K9me2/3, H3K79me0 and H4K20me2/3 within the same nucleosome was identified from chromatin enriched by ATRX-ADD. This newly reported PTM combination was enriched in heterochromatin, as revealed by the associated DNA.

Conclusions: Our results suggest the broad utility of recombinant reader domains as an enrichment tool specific to combinatorial PTM patterns, which are difficult to probe directly by antibody-based approaches. The reader affinity platform is compatible with several downstream analyses to investigate the physical coexistence of nucleosomal PTM states associated with specific genomic loci. Collectively, the reader-based workflow will greatly facilitate our understanding of how distinct chromatin states and reader domains function in gene regulatory mechanisms.

Keywords: Affinity enrichment; Antibody-free; Chromatin; Histone; PTM; Peptide microarray; Reader domain.

Figure 1

Design of the combinatorial PTM histone peptide microarray. (A) Construction of the combinatorial PTM peptide library. Combinatorial PTM histone peptide microarray featuring a tiled design and a high content of linked PTMs. Peptide sequence (underlined) covers core histones and histone variants; colored amino acids were modified according to the key. Colored arcs represent new PTM combinations and black arcs display PTM combinations used in previous arrays. (B) Histone peptide microarray binding assay. Peptides from Figure 1A were synthesized as cellulose conjugates and spatially spotted onto microarray, on which reader binding was detected and quantified by dual-channel fluorescence. Red star represents the binding signal at 635 nm, which was used for reference binding intensities. (C) Layout of the combinatorial PTM histone peptide microarray. On the left is a representative image of an entire microarray before protein binding (detected at 532 nm, green channel). On each microarray, two full libraries were included and each peptide was printed as triplicate spots. On the right is a zoom-in of a subarray from the big library detected at dual-channel (532 nm/635 nm, green/red channel) after protein binding. On each subarray, biotin peptides were included as positive controls (boxed in red dashed line). Misprinting (boxed in white dashed line) can be easily identified from the green channel. Protein bindings (boxed in yellow dashed line) were quantified from the signal intensities at 635 nm.

Figure 2

Histone peptide microarray analysis reveals high selectivity for amino-acid sequence and combinatorial PTM states by recombinant reader domains. (A-C) Heat map representation of binding specificity of reader fusions and histone antibodies determined from histone peptide microarrays. Reader fusions were probed on the slide at concentrations of 10 nM to 500 nM and antibodies were used as 1:10,000 to 1:2,000 dilution. (A) Effect of amino-acid sequence on binding. Averaged fluorescence signal intensities were normalized to the range between 0 and 1 by the lowest and the highest values of selected peptides for individual array. (B) Effect of combinatorial PTMs on binding. Averaged signal intensities for peptides with combinatorial PTMs were normalized to those with single modifications of interests as log₂ ratio. (C) Correct N-terminal context is required for H3K9me3 binding by ATRX-ADD. The signal intensities of K9me3-containing peptides that covered H3 5 to 17 amino acids were quantified and normalized to the highest signals from individual arrays. ATRX-ADD domain was able to bind H3K9me3 peptides covering H3 1 to 13 aa (A) but not the peptides of H3 5 to 17 aa, whereas the antibody binds to both sequences.

Figure 3

Specific binding with native mononucleosome library by reader modules in matrix-assisted reader chromatin capture (MARCC). (A) Scheme of MARCC. Reader domain characterized by peptide microarray is immobilized on resin and incubated with MNase-digested native mononucleosomes. Enriched nucleosomes are released by TEV protease cleavage and analyzed for coexisting histone PTMs and associated DNA. See text for details. (B,C) AIRE-PHD, ING2-PHD and ATRX-ADD were immobilized on resin and incubated with native mononucleosomes. Bound nucleosomes were boiled on beads and separated on 12% SDS-PAGE and probed with corresponding histone antibodies. HaloTag protein was included as a negative control. Bound nucleosomes were quantified as percentage input. (D) ING2-PHD MARCC enriched for active chromatin. qPCR for active chromatin marker GAPDH and heterochromatin marker SAT-2 was performed for DNA extracted from MARCC by ING2-PHD and ChIP by H3K4me3 antibody (ab8580). IgG and HaloTag were used as negative controls for MARCC and ChIP, respectively. The DNA was quantified by input DNA standard curve and plotted as percentage input.

Figure 4

A unique nucleosomal PTM combination captured by ATRX-ADD MARCC. Relative amounts of different PTM states for selected histone peptide fragments from MARCC-enriched chromatin determined by qMS. Peptide sequence and modified lysine is labeled in the chart. Different PTM states for the same peptide sequence are represented by different colors. Relative quantity of the dominant state is labeled in the pie chart with unmodified state colored in gray and dimethylated or trimethylated state colored in red (red encircling lines mark the pooled dimethylation and trimethylation states). A high level of near stoichiometric enrichment of 4 PTM states (H3K4unmod, H3K9me2/3, H3K79unmod and H4K20me2/3) was identified from MARCC with ATRX-ADD. For AIRE-PHD MARCC, H3K4unmod but not the other three PTM states was enriched to 90%.

Figure 5

Quantitative profiling of chromatin PTM patterns identifies heterochromatic signature captured by MARCC with ATRX-ADD. (A,B) Relative PTM enrichment with ATRX-ADD MARCC and AIRE-PHD MARCC. Relative enrichment was represented by log₂ of the ratio of MARCC over input. PTM states above the x-axis were enriched by MARCC, whereas PTM states in the opposite direction were depleted by MARCC. Error bars were calculated from two separate MARCC-qMS experiments. For the complete qMS data set, refer to Additional file 11: Table S1. (C) Relative enrichment of heterochromatic DNA (SAT-2) by qPCR obtained from MARCC with ATRX-ADD. ChIP with H3K9me3 antibody (ab8898) or IgG and HaloTag were included as controls.