The significance, development and progress of high-throughput combinatorial histone code analysis

Abstract

The physiological state of eukaryotic DNA is chromatin. Nucleosomes, which consist of DNA in complex with histones, are the fundamental unit of chromatin. The post-translational modifications (PTMs) of histones play a critical role in the control of gene transcription, epigenetics and other DNA-templated processes. It has been known for several years that these PTMs function in concert to allow for the storage and transduction of highly specific signals through combinations of modifications. This code, the combinatorial histone code, functions much like a bar code or combination lock providing the potential for massive information content. The capacity to directly measure these combinatorial histone codes has mostly been laborious and challenging, thus limiting efforts often to one or two samples. Recently, progress has been made in determining such information quickly, quantitatively and sensitively. Here we review both the historical and recent progress toward routine and rapid combinatorial histone code analysis.

Fig. 1

The post-translational modifications of histone H1 reported in the literature (ac acetylation, me1 monomethylation, P phosphorylation, ribo ADP ribosylation, ub ubiquitination/ubiquitylation). Unlike other histones, the numbering of H1 has generally included the N-terminal methionine; thus, we start sequence numbering at two. The sequence and numbering scheme for human histone H1.2 is shown, and PTMs of other variants are adjusted to their homologous H1.2 site. Sequence-specific PTMs of other variants not consistent with the H1.2 sequence are shown by including the alternate amino acid above the sequence. The variants for which this alternate amino acid occur at a homologous point are shown in the subscript on the alternate amino acid. Not all modifications are well validated, and there are substantial gaps in our knowledge of which PTMs occur on which variants. Some have only been observed on one or two variants, but are assumed to occur on other variants due to homology. The acetylation at S2 is N-terminal. The ribosylation at K213 is C-terminal

Fig. 2

The post-translational modifications of histone H2A reported in the literature (ac acetylation, me1 monomethylation, P phosphorylation, ribo ADP ribosylation, ub ubiquitination/ubiquitylation, bio biotinylation). The core sequence and numbering scheme for human histone H2a.1 is shown, and PTMs of other variants are adjusted to their homologous H2a.1 site. Divergent sequences of other variants at the termini are shown as such. Sequence-specific PTMs of other variants not consistent with the H2a.1 sequence are shown by including the alternate amino acid above the sequence. The variants for which this alternate amino acid occurs at a homologous point are shown in the subscript on the alternate amino acid. Not all modifications are well validated, and there are substantial gaps in our knowledge of which PTMs occur on which variants. Some have only been observed on one or two variants, but are assumed to occur on other variants due to homology

Fig. 3

The post-translational modifications of histone H2B reported in the literature (ac acetylation, me1 monomethylation, me2 dimethylation, mex unspecified methylation degree, P phosphorylation, ribo ADP ribosylation, ub ubiquitination/ubiquitylation). The sequence and numbering scheme for human histone H2b.1 is shown, and PTMs of other variants are adjusted to their homologous H2b.1 site. Sequence-specific PTMs of other variants not consistent with the H2b.1 sequence are shown by including the alternate amino acid above the sequence. The variants for which this alternate amino acid occurs at a homologous point are shown in the subscript on the alternate amino acid. Not all modifications are well validated, and there are substantial gaps in our knowledge of which PTMs occur on which variants. Some have only been observed on one or two variants, but are assumed to occur on other variants due to homology

Fig. 4

The post-translational modifications of histone H3 reported in the literature [ac acetylation, me1 monomethylation, me1–2 mono- and dimethylation, me1–3 mono-, di- and trimethylation, me1,3 mono- and trimethylation (di- likely but unreported), mex unspecified methylation degree, P phosphorylation, bio biotinylation, iso proline isomerization, prop proprionylation, cit citrulination, but butylation). The sequence and numbering scheme for human histone H3.1 is shown, and PTMs of other variants are adjusted to their homologous H3.1 site. The one sequence-specific PTM of H3.3S31ph is shown by including the alternate amino acid above the sequence. Not all modifications are well validated, and there are substantial gaps in our knowledge of which PTMs occur on which variants. Some have only been observed on one or two variants, but are assumed to occur on other variants due to homology

Fig. 5

The post-translational modifications of histone H4 reported in the literature (ac acetylation, me2 dimethylation, me1–2 mono- and dimethylation, me1–3 mono-, mex unspecified methylation degree, P phosphorylation, bio biotinylation, cit citrulination, but butylation, ub ubiquitination/ubiquitylation, su sumoylation). Unlike the other histones, there is only a single variant of histone H4. Not all modifications listed are well validated

Fig. 6

The relationship between combinatorial data sets, such as those generated by HT-CHCA, and traditional non-combinatorial approaches, such as bottom-up mass spectrometry and Western blot analysis. For each combinatorial data set, there is exactly one corresponding non-combinatorial result. The reverse relationship is not one-to-one, and there are infinite and widely varying possible combinatorial explanations for the non-combinatorial data. The example used is of histone H3 K4 and K9 methylation. These are signals known to be of opposite meaning (gene activating and silencing, respectively), and the dimethylated form will thus have a meaning distinct from either of the monomethylated forms. Even if they act in superposition and cancel each other, this would mean that the top left example would have all gene-activating signals obliterated, while the lower left example would have a full 20% of H3 activated

Fig. 7

The various approaches to analyzing histone H3. The semi-bottom-up approach involves an ArgC-like digest using trypsin but blocking digestion at lysines with a proprionylation reaction. The middle-down approach typically uses GluC digestion to obtain a large 50-amino acid N-terminal peptide, which contains the vast majority of known modification sites. The top-down approach analyzes the undigested protein, maintaining all connectivity, but at the cost of sensitivity, time and, most importantly, confident connectivity information. The semi-bottom-up analysis usually uses reversed phase and collision-induced dissociation mass spectrometry (RP-LC–MS). Both middle-down and top-down analyses use electron transfer dissociation (ETD) or electron capture dissociation (ECD). These samples are usually fractionated by off-line HPLC, but recently on-line methods have been used (WCX-HILIC–LC–MS), dramatically improving throughput and sensitivity

Fig. 8

An example of a high throughput combinatorial histone code analysis. On-line pH gradient-based weak cation exchange-hydrophilic interaction liquid chromatography (WCX-HILIC) is used for chromatographic separation, and electron transfer dissociation (ETD) is used to sequence eluting combinatorial histone codes. a An LC-MS heatmap is shown. How PTMs effect separation are indicated. Increasing acetylation significantly decreases retention time, while methylation modulates retention time more subtly. b An example of an eluting combinatorial histone code being identified on the millisecond time scale. c The ion map corresponding to (b). Quantitation may be achieved on all of the combinatorial histone codes identified using both the precursor ion intensity and the fraction of the spectrum that each combinatorial histone code represents determined computationally for mixed spectra. In this manner, the entire complement of combinatorial histone codes may be identified and quantified relative to each other for a sample in a matter of hours