Middle-down electron capture dissociation and electron transfer dissociation for histone analysis

Abstract

The post-translational modifications (PTMs) of histones play a major role in activating or silencing gene transcription. To gain better understanding of the interplay between the PTMs that occur on histones, they are extensively studied using mass spectrometry techniques. Due to the abundance of lysines and arginines, the typical trypsin digestion has been found less favorable and GluC-digests have been explored as an alternative to yield larger peptides amenable to middle-down approaches. In addition, the use of weak cation exchange hydrophilic interaction liquid chromatography (WCX-HILIC) and the use of electron-based fragmentation techniques were found to be advantageous for the in-depth characterization of histone variants containing multiple PTMs.

As a test model, we used histones from MEL (murine erythroleukemia) cells treated with butyric acid or DMSO. After acid extraction, histone pellets were dried and fractionated using a reversed-phase C3 column. For middle-down analysis, selected histone fractions were digested using GluC. The digested samples were separated on a WCX-HILIC capillary column packed in-house with PolyCAT A resin, coupled to a linear trap quadrupole Fourier transformation ion cyclotron resonance (LTQFT-ICR) instrument. Raw data was acquired on the LTQFT-ICR using electron capture dissociation (ECD). After deconvolution of the raw data, we generated heatmaps to illustrate differential maps between differentially treated histone samples. We also explored the innovative use of Skyline to quantify histone tails. In addition, we report some preliminary data using a synthetic histone peptide acquired on an Orbitrap Fusion using electron transfer dissociation (ETD). Both, ECD and ETD methods are capable of comprehensively analyzing complex histone variations not accessible with conventional techniques.

Review

The post-translational modifications (PTMs) of histones play a major role in activating or silencing gene transcription. Acetylation and methylation are the two most common PTMs on histones. Acetylation occurs on lysines and is transferred by histone acetylases and removed by histone deacetylases. Histone methyl transferases and demethylases regulate the mono-, di- and tri-methylation of lysines and arginines. Acetylation and methylation have opposite effects: methylation stabilizes the charge of the side chain amine and amidine, respectively, while acetylation neutralizes the amine of the lysine by forming an amide bond (Fig. 1a). A general model assumes that the positively charged lysines interact closely with the negatively charged DNA, leading to a closed chromatin formation that is associated with transcriptional repression. In contrast, acetylated and thus neutral lysines cannot interact closely with DNA, leading to an open chromatin structure, which is commonly associated with transcriptional activation (Fig. 1b) (Garcia et al. 2007).

Fig. 1

a Histone lysine side chains can be acetylated via histone acetylases and deacetylated via histone deacetylases. The acetylation creates a newly formed amide bond that cannot be protonated. b Thus, acetylated histones cannot interact closely with negatively charged DNA, opening up chromatins to their active form

Mass spectrometry (MS) and particularly middle-down approaches are playing an increasingly important role in characterizing PTMs on histones (Cannon et al. 2010; Garcia 2010; Kalli et al. 2013; Wu et al. 2005; Moradian et al. 2014). In contrast to a typical bottom-up approach, where proteins are digested with trypsin to generate peptides generally smaller than 2500 Da, middle-down approaches use enzymes that yield larger peptides (circa 3000–10,000 Da). This is particularly useful for histones, where lysines and arginines are abundant in the N-termini. As shown in Fig. 2, trypsin would typically generate histone peptides too small to be reliably detected by MS. In addition, many histone proteoforms share sequence homology (e.g., peptide PGTVALR is shared by histone H3.1 and H3.2). Thus, even if a peptide is detected and characterized, it would be impossible to infer from which proteoform the PTMs originated from. Both AspN and GluC have been successfully used to generate middle-down peptides from histones (Kalli et al. 2013; Bonenfant et al. 2006; da Cunha et al. 2006; Guedes et al. 2011; Jung et al. 2013; Kawasaki et al. 2003; Lu et al. 2009; Kalli and Hess 2012; Phanstiel et al. 2008). The use of weak cation exchange hydrophilic interaction liquid chromatography (WCX-HILIC) enabled the separation of histones based on their charge (acetylation status) and hydrophilic interactions (methylation status) (Young et al. 2009; Young et al. 2010). Fig. 3a shows the resulting broad peaks presenting 5, 4, 3, etc. acetylations. While the peaks may appear relatively broad in comparison to C18 reversed-phase chromatography, WCX-HILIC affords a secondary dimension of separation based on the methylation status (Fig. 3b), separating multiple methylated isoforms that contain the same number of acetylations. Young et al. introduced a pH gradient using WCX-HILIC column to allow direct online coupling of WCX-HILIC with MS (Young et al. 2009; Young et al. 2010). A recent further improvement using a 2D column setup with a C18 precolumn for sample loading has been introduced by Jung et al. (2013). The use of a C18 precolumn allowed Jung et al. to dissolve samples in aqueous solutions rather than acetonitrile solutions that were necessary for the WCX-HILIC approach. This improved reproducibility, since the acetonitrile solutions evaporated too quickly during analysis.

Fig. 2

Example of a theoretical tryptic digestion of the N-terminus of H3.1 histone leading to peptides generally smaller than six amino acids that would not be amenable to LC-MS/MS in a typical bottom-up approach. Only two peptides (in blue circles) would be detectable. In addition, peptide PGTVALR is shared between histone H3.1 and histone H3.2, which makes it impossible to infer from which protein it originally came from. Middle-down approaches circumvent these problems by looking at larger peptides (in the case of histone H3.1 with 50 amino acids)

Fig. 3

WCX-HILIC separation of histone tails enables the separation based on acetylation (a) and methylation (b) status

As a mild non-ergodic fragmentation technique that preserves PTMs, electron transfer dissociation (ETD) and electron capture dissociation (ECD) have been proven beneficial for the detection of histone modifications. In addition, fragmentation during electron-based dissociation yields more equally intense fragment ions, when compared to collision-induced dissociation (CID) (Fig. 4). An additional benefit is the fact that ratios of fragment ion intensities in electron-based fragmentation techniques can be used for proteoform quantitation (Pesavento et al. 2006).

Fig. 4

By fragmenting the peptide bonds, collision-induced dissociation (CID) produces b- and y-ions of unequal intensity, while electron capture/transfer dissociation (EC/TD) fragments between the alpha carbon and the amine adjacent to the peptide bonds, thus producing c- and z-ions of approximately the same intensity, aiding both the identification of PTMs and the quantification of the peptides. A comparison for the synthetic H3 histone N-terminal fragment ARTKQTARKSTGGKAPRKQLATKAARKSAPATGGVKKPHRYRPGTVALRE is shown with CID (a) and ETD (b)

In our own lab, we use histones from MEL cells as a test model. When treated with the histone deacetylase inhibitor butyric acid or DMSO for 4 days, MEL cells differentiate and produce embryonic or adult hemoglobin, respectively. After acid extraction (Shechter et al. 2007), histone pellets were dried and fractionated using a reversed-phase C3 column. For middle-down analysis, selected histone fractions were digested using GluC. The digested samples were separated on a WCX-HILIC capillary column packed in-house with PolyCAT A resin, coupled to a linear tap quadrupole Fourier transformation ion cyclotron resonance (LTQFT-ICR). After deconvolution of the raw data using YADA (Carvalho et al. 2009), we generated idealized heatmaps to illustrate differences between butyric acid and DMSO-treated histone samples (Fig. 5). As expected, the deacetylase inhibitor butyric acid inhibited deacetylation of histones. In general, the most intense ions were represented by 2–4 acetylations (Fig. 5a). In contrast, the DMSO-treated histones showed more intense ions for 0 and 1 acetylations as shown in Fig. 5b.

Fig. 5

Heat map of the MS1 scans vs. deconvoluted mass of the GluC-digested N-terminus of histone H3 spectra treated with (a) butyric acid and (b) DMSO after WCX-HILIC separation. The histone deacetylase inhibitor butyric acid-treated H3 histones showed the most intense ions with 2 and 3 acetylations, while the DMSO-treated H3 histones showed the most intense ions with 0 and 1 acetylations

To make efficient use of the precursor ion intensities for quantitative purposes, we also explored the innovative use of Skyline to quantify histone tails. Since the +10 fragment ions were the most intense ions in the MS spectra, we imported all theoretical +10 ions for all possible histone H3.2 tails with 0–5 acetylations and 0–7 methylations in Skyline (Schilling et al. 2012; MacLean et al. 2010). The resulting total ion chromatogram (XIC) is shown in Fig. 6, quantifying all acetylation statuses identified in the WCX-HILIC LTQFT ECD analyses.

Fig. 6

Using Skyline to quantify changes of acetylation status in H3 histones after importing all theoretical +10 ions for all possible histone H3 tails with 0–5 acetylations and 0–7 methylations

In addition to ECD analyses on the LTQFT, we have started to explore suitable instrument parameters for ETD on an Orbitrap Fusion. To do so, preliminary data of the synthetic H3 histone N-terminal fragment ARTKQTARKSTGGKAPRKQLATKAARKSAPATGGVKKPHRYRPGTVALRE were acquired on an Orbitrap Fusion instrument.

Since the 5 kDa peptide is more like a small protein than a peptide, we tested initially whether standard pressure settings or intact protein settings should be used for fragmentation. In all cases, the standard pressure settings showed more cleavages than the protein pressure settings. Next, we needed to convert the raw files into mzxml files to be further converted to MGF (Mascot Generic Format) files for database searching. For the raw file to mzxml file conversion, we tested MSConvert (French et al. 2014), and for the mzxml file to MGF conversion, we used MS-Deconv (Liu et al. 2010). The combination of MSConvert and MS-Deconv performed similarly to ReAdW and MS-Deconv for the multiply charged ions as they are observed in middle-down approaches. As shown in Fig. 7, we found that short reaction times of 10 ms together with 40 % supplemental higher energy collision dissociation (HCD) activation gave better results than longer reaction times and lower or higher activation energies. Similarly, 20-ms ETD reaction times together with 20-ms CID supplemental activation resulted in more detectable cleavages (Fig. 7). When we compared the different charge states (+7, +8, +9, +10), we found that the most intense +10 ion generally gave more cleavage product ions.

Fig. 7

Optimization of fusion ETD acquisition parameters. Number of peptide cleavages observed for synthetic H3 fragment using different ETD reaction times vs. supplemental activation with CID or HCD

Conclusions

Taken together, the use of WCX-HILIC with middle-down proteomics is becoming an increasingly popular method for histone characterization. Our preliminary data furthermore show that both the LTQFT with ECD acquisition and the Orbitrap Fusion with ETD acquisition are capable of identifying and quantifying histone PTMs. We are predicting that the Orbitrap Fusion will play an important role in the comprehensive and reproducible analysis of histones. Fully automated bioinformatics pipelines as they exist for bottom-up proteomics are not yet available for middle-down proteomics and will require substantial development. Once in place, they will likely trigger another significant advancement in histone modification analysis.