A Method to Separate and Characterize Modified Forms of Histones Using HILIC and Electron Transfer Dissociation-Mass Spectrometry

Histone proteins are the highly basic core proteins of nucleosomes and play an important role in DNA regulation through post-translational modifications (PTMs), primarily at their N-terminal tails.1 There are many potential sites of modification on histones, and the exhaustive combination of these sites could theoretically lead to millions of distinctly modified histone forms. Previous work has shown, however, that the number of modified forms is significantly reduced by biological specificity, and often only some or specific modifications are observed at each site. Site-specific antibodies have traditionally been used to characterize and investigate sites of histone modification. Nevertheless, this approach, as well as bottom-up mass spectrometry approaches, are not capable of identifying the biological combinations of modifications on a given protein, which give rise to unique multiply modified forms that drive chromatin function.1

The determination of modified histone forms presents significant analytical challenges due to the basic nature of histone proteins and the similar characteristics between the many modified forms. Physical separation by chromatographic methods has proved especially difficult since numerous different modified forms coelute under most conditions. For example, almost no separation is achieved with any type of reversed-phase liquid chromatography. Because of the difficulty of separation and the need to decipher information about the coexistence of PTMs across large sections of the protein, expensive high-resolution Fourier transform mass spectrometers have primarily been used to study the individual modified forms of histones.2

This article presents a method adapted from the work of Garcia et al.3 for determining the modification states of the histone H3 N-terminal tail using cation exchange-hydrophilic interaction liquid chromatography (HILIC) and a benchtop linear quadrupole  ion trap mass spectrometer equipped with electron transfer dissociation (ETD).4,5 The 1–50 amino acid (aa) peptide GluC digestion product from the N-terminus of the H3 histone protein was studied to decrease analytical complexity while preserving information about the coexistence of modifications in the region of the protein (N-terminus) where almost all known modifications have been previously observed. Using this method, the sample requirements were reduced and the specificity of the analysis improved by nanoflow liquid chromatographic separation of differentially modified forms before mass spectrometric analysis and efficient fragmentation by ETD. Although the instrumentation used is not currently widely distributed, it is commercially available and is intended for widespread distribution within the budgets of most bioanalytical laboratories.

Histone preparation

HeLa S3 cells were grown in suspension and histone acid extracted as previously described.6 The histone acid extract was separated into the constituent histone proteins (H2A, H2B, H3, and H4); the H3.2 was then digested with the GluC protease. The 1–50 aa peptide of the H3.2 histone protein was then further purified as described previously.2

Cation exchange-HILIC

Approximately 20 μg of purified histone H3.2 1–50 aa peptide was loaded on a 2.1 × 250 mm Poly CAT A column (PolyLC, Columbia, MD) and the various modified forms thereof were separated using a multistep gradient from 20% solvent A to 70% solvent B at a flow rate of 0.2 mL/min. Solvent A consisted of 75% acetonitrile, 0.02 M TEA (triethylamine)/H3PO4, pH 4.0; solvent B contained 60% acetonitrile, 0.02 M TEA/H3PO4, pH 4.0, plus 0.5 M NaClO4. The gradient used was from 0 to 20%B in 5 min, from 20 to 70%B in 160 min, and then to 100%B over 10 min. One-minute fractions were collected, concentrated by vacuum evaporation, and precipitated in 20% trichloroacetic acid (TCA) overnight. The resulting pellets were washed three times with 20% TCA and then once each with acetone/0.1% HCl and acetone to prepare the sample for MS analysis.

Mass spectrometric analysis

A combination nanoflow column and nanospray emitter was prepared by pulling a 50-μm-i.d. fused-silica capillary to a tip with a P2000 laser tip puller (Sutter Instruments, Novato, CA). This was then packed with 5-μm-diameter, 100-Å pore size Magic C18 AQ (Michrom BioResources, Auburn, CA) to approximately 10 cm. Fractions collected from the HILIC separation were dissolved in 1.4% acetonitrile, 0.1 M acetic acid, and loaded onto the nanoflow column/nanospray emitter. Using 0.1 M acetic acid as buffer A and 70% acetonitrile, 0.1 M acetic acid as buffer B, histone H3.2 was eluted by a gradient from 2 to 100%B in 30 min. The column eluent was ionized by nanospray ionization and introduced into an LTQETD mass spectrometer (Thermo Fisher Scientific, Waltham, MA). Every cycle, a full MS was acquired from 300 to 2000 m/z, followed by another full MS of 525–575 m/z to observe the 10th charge state of H3.2 1–50 aa peptide more accurately. Five data-dependent MS2 ETD zoom scans with five microscans were acquired each cycle based on the 525–575 m/z scan. Multiple ETD scans were added across the chromatographic peak to further improve signal. High-resolution mass spectrometry was performed in a similar manner on an LTQ-Orbitrap XL (Thermo Fisher Scientific) with full MS 300–2000 m/z at 30,000 target resolution.

Results and discussion

Figure 1 - Cation exchange-hydrophilic interaction chromatogram of the various modified forms of the histone H3.2 (1–50 aa peptide) with mass spectra of the 10th charge state at selected time points. The more highly modified forms elute early in the chromatogram, with the unmodified form eluting last at 74.5 min. From 65 to 75 min, primarily unacetylated forms elute; from 61 to 65 min, mainly monoacetylated forms elute; and mostly diacetylated forms elute from 57 to 61 min. Individual peaks generally correlate to different degrees of methylation, also progressing from least to the heavier modified with decreasing retention time. The mass spectra of selected fractions are inserted, showing different degrees of methylation.

For each HILIC fraction, a unique set of masses from 2 to 5 histone H3 major modification states were found. The masses observed corresponded to theoretical masses expected for the histone H3.2 1–50 aa peptide modified with methylations and/or acetylations. For example, the most highly modified forms eluted earlier in the HILIC gradient, with progressively less modified forms eluting in later fractions and the unmodified form eluting last at 74.5 min, as shown in Figure 1. On a unit resolution ion trap instrument, it is not possible to distinguish between an acetylation and a trimethylation (Δm = 0.036 Da) on a given peptide from the molecular masses alone; thus the modification states were labeled as methyl equivalents until unambiguous assignment was made by ETD tandem mass spectrometry.

The acetylated forms were chromatographically resolved from the isobaric methylated masses, eluting earlier in the HILIC separation. The chromatographic peaks from 65 to 75 min were primarily unacetylated forms, and those from 61 to 65 min were mainly monoacetylated, with diacetylated forms eluting from 57 to 61 min. This can be seen in Figure 1, where the fraction collected at 61 min differs by only three methyl equivalents (or one acetylation) in mass from the 66-min fraction, but eluted about 3 min earlier than expected based on the progression of increased methyl equivalents with decreased retention time from the unmodified peak at 74 min.

Figure 2 - Comparison of the low-resolution LTQ mass spectrum (a) and high-resolution LTQ-Orbitrap XL mass spectrum (b) for the fraction at 66 min, demonstrating accurate assignment of peaks based on unit resolution mass and retention time. The error from the expected mass for the five methyl form of the histone H3.2 1–50 aa peptide (b) is 1.4 ppm.

Where previous methods have relied on accurate mass rather than chromatographic separation for specificity, the method presented here enables characterization of the many forms on a unit resolution mass spectrometer with a combination of chromatographic separation and emerging tandem mass spectrometry technology. Figure 2 demonstrates that the chromatographically resolved methylation states, assigned using unit mass resolution linear ion trap data (Figure 2a), were properly assigned according to accurate mass data acquired on an Orbitrap mass spectrometer (Figure 2b). The monoisotopic peak of the five methyl form (541.8211 m/z) agrees with the theoretical exact mass (541.82185 m/z) with 1.4-ppm error within the expected accuracy of the instrument. The assignment of charge state, made based on the theoretical mass and the distribution of isotopes, as seen in Figure 3a, was validated by the 0.1-m/z separation of isotopes. Thus, the chromatographically separated unit resolution data suffice for proper assignment of the degree of modification of parent masses.

Figure 3 - Unambiguous assignment of the sites of modification on the five methyl peak. a) Full mass spectrum allows assignment of charge state. b) Mass spectrum focusing on the 10th charge state of the histone H3.2 1–50 aa peptide. c) ETD mass spectrum of the five methyl form of the peptide observed in (b). d) ETD ion fragment map where the sites of modification are assigned based on the ETD data in (c).

The sites of modification were identified by ETD, a recent development in tandem mass spectrometry technology that is now becoming available on benchtop instruments.5 The analysis time required to obtain sufficient sequence information to identify each modified form is much shorter than in previous methods that used electron capture dissociation (ECD) (minutes versus hours per fraction).3 With many modified forms in each sample and several samples needed for any comparative biological analysis, this is a distinct advantage. The amount of sample required is much less due to on-column concentration and the relative efficiency of the ETD analysis. Only seconds of accumulated ETD data as the peptide elutes are required as opposed to hours of direct infusion averaged data for ECD analysis. This reduced sample requirement makes the scale of cell cultures necessary to perform each successful experiment practical.

Figure 3a shows the full MS from 300 to 2000 m/z for the fraction at 67 min. Figure 3b is a separate full MS from 525 to 575 m/z. Figure 3c is a sum of several data-dependent ETD scans of the mass at 543.33 m/z as the histone H3.2 1–50 aa peptide elutes. The ETD data contain full sequence information for the first 15 amino acids from both the N- and C-termini of the histone H3.2 1–50 aa peptide. Of particular importance are the c9+1me2 peak at 1087 m/z and the z15+1me2 peak at 1822 m/z, which identify the major component of the five methyl form to be dimethylated at Lys9 and dimethylated at Lys36. The remaining methylation can be assigned to Lys27 because the other potential sites of methylation (Lys18 and Lys23) within the unfragmented region have not been found to be methylated in extensive previous studies.2


The method presented allows for effective characterization of the modified forms of histone N-terminal tails with significantly reduced analysis time and sample requirements. This methodology will enable further studies on the biological relevance of multiply modified forms of the histone proteins, and how the effects of the combination of specific sites of modification function in concert to regulate gene expression.


  1. Bonisch, C.; Nieratschker, S.M.; Orfanos, N.K.; Hake, S.B. Chromatin proteomics and epigenetic regulatory circuits. Expert Rev. Proteomics 2008, 5(1), 105–19.
  2. Garcia, B.A.; Shabanowitz, J.; Hunt, D.F. Characterization of histones and their posttranslational modifications by mass spectrometry. Curr. Opin. Chem. Biol. 2007, 11(1), 66–73.
  3. Garcia, B.A.; Pesavento, J.J.; Mizzen, C.A.; Kelleher, N.L. Pervasive combinatorial modification of histone H3 in human cells. Nat. Methods 2007, 4(6), 487–9.
  4. Boersema, P.J.; Mohammed, S.; Heck, A.J. Hydrophilic interaction liquid chromatography (HILIC) in proteomics. Anal. Bioanal. Chem. 2008, 391(1), 151–9.
  5. Syka, J.E.; Coon, J.J.; Schroeder, M.J.; Shabanowitz, J.; Hunt, D.F. Peptide and protein sequence analysis by electron transfer dissociation mass spectrometry. Proc. Natl. Acad. Sci. USA  2004, 101(26), 9528–33.
  6. Thomas, C.E.; Kelleher, N.L.; Mizzen, C.A. Mass spectrometric characterization of human histone H3: a bird’s eye view. J. Proteome Res. 2006, 5(2), 240–7.

Dr. Young is Postdoctoral Research Associate, and Prof. Garcia is Assistant Professor, Department of Molecular Biology, Princeton University, 415 Schultz Laboratory, Princeton University, Princeton, NJ 08544, U.S.A.; tel.: 609-258-8854; fax: 609-258-1035; e-mail: bagarcia@princeton.edu. Ms. Plazas-Mayorca, doctoral candidate, and Prof. Garcia are with the Department of Chemistry, Princeton University.