Filter-Aided Sample Preparation Procedure for Mass Spectrometric Analysis of Plant Histones

Characterization of histone post-translational modifications (PTMs) is still challenging, and robust histone sample preparation is essential for convincing evaluation of PTMs by mass spectrometry. An effective protocol for extracting plant histone proteins must also avoid excessive co-extraction of the numerous potential interfering compounds, including those related to secondary metabolism. Currently, the co-existence of histone marks is addressed mostly by shotgun proteomic analysis following chemical derivatization of histone lysine residues. Here, we report a straightforward approach for plant histone sample preparation for mass spectrometry, based on filter-aided sample preparation coupled with histone propionylation. The approach offers savings in sample handling and preparation time, enables removal of interfering compounds from the sample, and does not require either precipitation or dialysis of histone extract. We show the comparison of two protocol variants for derivatization of histone proteins, in-solution propionylation in the vial and propionylation on the filter unit. For both protocols, we obtained identical abundances of post-translationally modified histone peptides. Although shorter time is required for histone protein labeling on the filter unit, in-solution derivatization slightly outweighed filter-based variant by lower data variability. Nevertheless, both protocol variants appear to be efficient and convenient approach for preparation of plant histones for mass spectrometric analysis.

Supplementary Data 1.1 Chemical derivatization technique for histone proteins adapted to on-membrane performance using mammalian histone extracts

In-solution protein propionylation of mammalian histones (PROP-in-SOL_mh) and proteolytic digestion followed by in-solution propionylation of peptides
Histones extracted from MEC-1 cells were subjected to a double round of propionic anhydride derivatization (at both protein and peptide levels). Briefly, a 20 μg portion of histone sample was diluted with acetonitrile (MeCN) and deionized water to a final volume of 20 μL and final MeCN concentration of 50% (v/v). NH4OH (1 μL) was added, then propionylation reagent was prepared by mixing propionic anhydride with MeCN in a 1:3 ratio and a portion equal to 25% of the sample volume was immediately added. The pH was adjusted to 8-9 by NH4OH, the sample was incubated in thermomixer at 25 °C and 750 rpm for 20 min, then the sample volume was reduced in a Savant SPD121P concentrator (SpeedVac; Thermo Scientific) to 5 μL. The second round of propionylation was carried out with the same protocol. Propionylated histone proteins were reconstituted in 50 μL of 50 mM triethylammonium bicarbonate (TEAB) and trypsin was added in a 1:40 enzyme:protein ratio. After overnight digestion at 37 °C, the sample was dried in the SpeedVac. The generated peptides were subjected to a double round of propionylation at N-termini using the protocol described above. Each sample was dried in a SpeedVac overnight and dissolved in 50 μL of 50% MeCN. The resulting solution was concentrated in the SpeedVac to 20 μL and its peptide concentration was determined using a Micro BCA™ Protein Assay Kit. Prior to LC-MS/MS analysis, the sample was acidified by adding formic acid (FA) to a final concentration of 1%.

On-membrane protein propionylation of mammalian histones (PROP-on-FILTER_mh) and proteolytic digestion followed by in-solution propionylation of peptides
A 20 μg portion of histone extract dissolved in water was placed in a YM-10 Microcon filter unit (Millipore) with 300 μL of 50 mM TEAB (pH 8.5), centrifuged (14 000 g, 30 min, 25 °C) and washed three times with 200 μL of 50 mM TEAB. The sample was diluted with 50 mM TEAB to a volume of 30 μL then 2 μL of NH4OH was added. A 10 μL portion of propionylation reagent, freshly prepared for each batch of three samples by mixing propionic anhydride and isopropanol in a 1:3 ratio, was immediately added to the sample. The pH was adjusted to 8-9 by NH4OH, then the sample was incubated in a thermomixer (50 °C, 700 rpm, 40 min) and centrifuged (14 000 g, 10 min, 25 °C). The second round of propionylation was carried out with the same protocol. After derivatization, the sample was washed three times with 100 μL of 50 mM TEAB, and trypsin diluted in 50 μL of 50 mM TEAB was added in a 1:40 enzyme:protein ratio. Following overnight digestion at 37 °C, the digest was collected by centrifugation (14 000 g, 10 min, 25 °C), subjected to two additional washes with 50 μL of 50 mM TEAB, then concentrated in the SpeedVac to 5-10 μL and diluted with 50 mM TEAB to a volume of 20 μL. Peptides were propionylated at N-termini as follows. One μL of NH4OH was added to the sample, then 5 μL of the propionylation reagent. The pH was adjusted to 8-9 by NH4OH. The sample was incubated at 37 °C for 40 min, concentrated in the SpeedVac to 5 μL, reconstituted in 20 μL of 50 mM TEAB and propionylated at the peptide level again. The sample was then dried in the SpeedVac overnight and dissolved in 50 μL of 50% MeCN. The solution was concentrated in the SpeedVac to 20 μL and the peptide concentration was determined using a Micro BCA™ Protein Assay Kit. Finally, before LC-MS/MS analysis, the sample was acidified by adding FA to a final acid concentration of 1%.

LC-MS/MS analysis of derivatized histone peptides
The acquired peptide mixtures were analyzed by LC-MS/MS using a RSLCnano system (Thermo Fisher Scientific) connected on-line to an Impact II Ultra-High Resolution Qq-Time-Of-Flight mass spectrometer equipped with a CaptiveSpray nanoBooster ion source (Bruker). Portions of the solutions containing 40 ng of the derivatized human peptides (prepared as described above) were injected into the LC system, respectively. Prior to LC separation, peptides were concentrated online on a trapping column (100 μm × 30 mm), filled with 3.5-μm X-Bridge BEH 130 C18 sorbent (Waters), that had been equilibrated (together with the analytical column) with the initial mobile phase before injecting the sample into the sample loop. The peptides were separated using an Acclaim Pepmap100 C18 column; 3 µm particles, 75 μm × 500 mm; Thermo Fisher Scientific analytical column. The gradient elution was as follows: 0-90 min, 1-70% B; 90-100 min, 70-98% B; 100-120 min, 98% B (where the mobile phases A and B consisted of 0.1% FA in water and in 80% MeCN, respectively). After each injection, the flow rate was set at 500 nL/min for 13 minutes to load peptides on the column, then linearly decreased to 300 nL/min over 2 minutes. The analytical column's outlet was directly connected to the NanoBooster ion source, which was filled with MeCN. MS and MS/MS spectra were acquired in a data-dependent strategy with 3 s cycle time. The mass range was set to 150-2200 m/z and precursors were selected from 300 to 2000 m/z. The acquisition speed of MS and MS/MS scans (the latter varied according to precursor intensity) was 2 and 4-16 Hz, respectively. DataAnalysis software (version 4.2 SR1; Bruker) was used for pre-processing the mass spectrometric data (including recalibration, compound detection and charge deconvolution).

Database searches and quantification
In-house Mascot search engine (version 2.4.1; Matrixscience) was used to search for matches to exported MS/MS spectra obtained from analyses of human samples in the UniProtKB Human (version 2017_02; 21031 protein sequences), in-house Histone Human (version 2017_02; 114 protein sequences in total) and cRAP contaminants databases. Settings for all searches included semispecific Arg-C enzyme specificity and up to two missed cleavages. In addition, the following variable modifications were set for searches against the Histone human database: methyl (R, K), dimethyl (K), trimethyl (K), propionyl (K, N-term, S, T, Y), acetyl (K, protein N-term) and deamidation (N, Q). Propionyl (K, N-term, S, T, Y), acetyl (K, protein N-term) and deamidation (N, Q) were set as variable modifications in UniProtKB Human database searches. Mass tolerances of peptides and MS/MS fragments for MS/MS ion searches were 20 ppm and 0.05 Da, respectively. Manual peak labelling and peptide precursor area calculation were done via Skyline 3.6 software. A spectral library was created using the Proteome discoverer platform (version 1.4; Thermo Fisher Scientific).
Only peptides with statistically significant peptide scores (p < 0.01) were included. portion of each histone sample was used for quality control -SA (red), MQ (blue), or SDS (pink). Considerable amount of protein or non-protein contaminants was observed in both SA and MQ chromatograms while histone protein peaks were not detected. Due to presence of SDS, poor quality of SDS sample separation was observed. The histone proteins are likely to be hidden in the broad peaks. Human recombinant histone standards and human histone extract after TCA precipitation redissolved in MilliQ water are also presented (black and grey, respectively). Figure 3. Chemical derivatization technique for histone proteins adapted to onmembrane performance using mammalian histone extracts. (A) An illustrative scheme of the PROPon-FILTER_mh workflow including comparison with a commonly used in-solution derivatization (PROP-in-SOL_mh). (B) PROP-on-FILTER_mh performance compared to PROP-in-SOL_mh approach. Grayscale pie charts showing proportions of identified histone H3 and H4 peptides in four categoriesdesired (peptides cleaved and propionylated as expected), underpropionylated (peptides with at least one unmodified amino group on lysine residue or N-terminus), overpropionylated (peptides with at least one propionylated hydroxyl group on S, T or Y residue), non-specifically cleaved (peptides with cleavage at lysine C terminus or missed cleavage at arginine C terminus). Color pie charts showing proportions of assignable peptides, i.e. peptides enabling correct quantification. (C) Box-plots and scatter-plots of the means and standard deviations of abundances of histone H3 and H4 peptide forms detected in the samples. The boxplots show extremes, interquartile ranges and medians (N=39). Means and standard deviations were compared by Mann-Whitney tests (p-values) and Spearman's correlation coefficients (SCC values). (D) Radar charts showing relative abundances of individual peptide forms of histone H3 and H4 (means; N = 2), determined from the ratio of the XIC peak areas of particular assignable products to the summed XIC peak areas of the total pool of all quantified H3 or H4 peptides, respectively. The Y axes have a binary logarithm scale, with zero located in the center. Figure 4. Comparison of PROP-in-SOL and PROP-on-FILTER performance in term of inter-sample variability of selected plant histone mark levels. The distribution of log2transformed peptide precursor XIC peak areas is represented by a pair of boxplots. The boxplots show extremes, interquartile ranges and medians obtained from analyses of five samples. The numbers on the x-axis correspond to the peptide order in the Supplementary Table 1. Figure 5. PROP-in-SOL performance in term of inter-sample variability of selected plant histone mark levels in seven weeks old leaves and seven days old seedlings. The distribution of log2-transformed peptide precursor XIC peak areas is represented by a pair of boxplots. The boxplots show extremes, interquartile ranges and medians obtained from analyses of five samples. The numbers on the x-axis correspond to the peptide order in the Supplementary Table 1 and 2, respectively.

Supplementary Tables
Supplementary Table 1 Comparison of PROP-on-FILTER and PROP-in-SOL performance using plant histone extracts. Log2-transformed peptide precursor areas and relative abundances (RA) of individual peptide forms of histone H3 and H4 are presented for each replicate. RA was determined as the ratio of the XIC peak areas of particular assignable products to the summed XIC peak areas of the total pool of all quantified H3 or H4 peptides, respectively.