Article pubs.acs.org/ac
Compositional Analysis of Asymmetric and Symmetric Dimethylated H3R2 Using Liquid Chromatography−Tandem Mass SpectrometryBased Targeted Proteomics Qingqing Xu, Feifei Xu, Liang Liu, and Yun Chen* School of Pharmacy, Nanjing Medical University, 818 Tian Yuan East Road, Nanjing, 211166, China S Supporting Information *
ABSTRACT: Protein arginine methylation is one of the common post-translational modifications in cellular processes. To date, two isomeric forms of dimethylated arginine have been identified: asymmetric NG,NG-dimethylarginine (aDMA), and symmetric NG,N′G-dimethylarginine (sDMA). Evidence indicated that these isomers can coexist and have different or even opposite functions, with aDMA and sDMA forms of arginine 2 on histone H3 (i.e., H3R2me2a and H3R2me2s) being an example. Thus, specific detection and quantification of each isomeric form is important. Current methods are capable of predicting and detecting thousands of methylarginine sites in proteins, whereas differentiation and stoichiometric measurement of dimethylated protein isomers are still challenging. Liquid chromatography coupled with tandem mass spectrometry (LC−MS/MS)-based targeted proteomics has emerged as a promising technique for site-specific quantification of protein methylation using enzymatic peptides as surrogates of target proteins. However, it should be pointed out that a routine targeted proteomics strategy cannot easily distinguish sDMA- and aDMA-containing surrogate peptides due to their common nature. The estimated amount should be considered as the sum of both arginine dimethylated isomers. In this study, compositional analysis based on a linear algebra algorithm as an add-on to targeted proteomics was employed to quantify H3R2me2a and H3R2me2s (i.e., surrogate peptides of AR(me2a)TK(me1/2)QT and AR(me2s)TK(me1/2)QT). To achieve this simultaneous quantification, a targeted proteomics assay was developed and validated for each isomer first. With the slope and intercept of their calibration curves for each multiple reaction monitoring (MRM) transition, linear algebraic equations were derived. Using a series of mock mixtures consisting of isomers in varying concentrations, the reliability of the method was confirmed. Finally, the H3R2 dimethylation status was analyzed in normal MCF-10A cells, parental drug-sensitive MCF-7/WT cancer cells, and drug-resistant MCF-7/ADR cancer cells. Dimethylated H3R2 was also monitored in MCF-7/WT cells with the treatment of doxorubicin (DOX) for confirmation.
P
observed to coexist, and this type of bifunctional dimethylated site in proteins is often thought of as an on−off switch.4,6 However, there is increasing evidence indicating their simultaneous occurrence and that the site is more like a rotary switch.7 For example, the tails of histones, as a primary target of PRMTs, could potentially be either symmetrically or asymmetrically dimethylated at one time.8 Thus, the distinction between these isomers could give important added value in the study of biological systems. Histones are the major structural proteins of chromosomes, and their N-terminal tail has several known methylarginine sites.9,10 Among these sites, arginine 2 on histone H3 (H3R2) is among the most studied. To date, its methylation has been described in a number of reports.4,7 However, the specific function of its dimethylated forms remains unclear, especially
ost-translational modifications (PTMs) control a variety of cellular processes, and thus, their analysis is a crucial step in the investigation of biological systems. Among various modification patterns, protein methylation at arginine residues is a prevalent PTM.1 To date, two isomeric forms of dimethylated arginine have been identified: asymmetric NG, NG-dimethylarginine (aDMA) and symmetric NG, N′Gdimethylarginine (sDMA) (Figure 1).2 To form these dimethylated isomers, two principle categories of protein arginine N-methyltransferases (PRMTs) are involved.3 Thus, it is reasonable to assume that symmetric and asymmetric arginine dimethylated proteins may have different or even opposite functions.4 Consistent with this speculation, aDMA and sDMA amino acid residues originating from protein arginine dimethylation after protein hydrolysis have shown different clinical diagnostic values.5 However, different from the released amino acid residues, it is traditionally considered unnecessary or difficult to distinguish between potential aDMA and sDMA-containing proteins because they have been rarely © XXXX American Chemical Society
Received: January 7, 2016 Accepted: July 25, 2016
A
DOI: 10.1021/acs.analchem.6b00076 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
Figure 1. Chemical structures of nonmethylated, monomethylated (MMA), and dimethylated arginines as well as the product ion spectra of the doubly charged H3R2me2aK4me2 and H3R2me2sK4me2 at m/z 380.8 using the Agilent 6410 Triple Quadrupole LC/MS and Masshunter Workstation Software. The y-ions and b-ions observed were marked on the peptide sequence. An exhaustive list of theoretical product ions was also provided. The experiments were repeated three times, and representative mass spectra are shown.
developed to quantify product ions from these peptides.22 However, it should be pointed out that routine targeted proteomics strategies cannot easily distinguish sDMA- and aDMA-containing surrogate peptides due to their common nature.6 Hence, the estimated amount should be considered as the sum of both isomers.12 In some situations, the characteristic fragmentation properties of dimethylated peptides, such as neutral losses or specific reporter masses, can be employed.15 For example, the dimethylammonium ion (NH2(CH3)2+, m/z 46.065 Da) was considered as a specific reporter of aDMA, while the monomethylammonium ion (CH3NH3+, m/z 32.049 Da) was indicative of the presence of sDMA.6 Nevertheless, their extensive application was subject to a number of issues, which will be discussed in the following section. In this study, compositional analysis based on a linear algebra algorithm as an add-on to targeted proteomics was employed to simultaneously quantify H3R2me2a and H3R2me2s (i.e., surrogate peptides of AR(me2a)TK(me1/2)QT and AR(me2s)TK(me1/2)QT). A LC−MS/MS-based targeted proteomics assay was developed and validated for each arginine dimethylated isomer first. With the slope and intercept of their calibration curves for each MRM transition, linear algebraic equations were developed. Solutions of the linear algebraic equations were an assignment of the amount values to the isomers. To evaluate the reliability of this approach, a series of mock mixtures consisting of isomers in varying concentrations were prepared and subjected to analysis. Finally, this approach was applied to determine the H3R2 dimethylation status in normal MCF-10A cells, parental drug-sensitive MCF-7/WT cancer cells, and drug-resistant MCF-7/ADR cancer cells. The result was compared to that obtained using Western blotting.
for the aDMA and sDMA forms of H3R2 (i.e., H3R2me2a and H3R2me2s). The only available information is that dimethylated H3R2 is closely related to the adjacent residue lysine 4 (H3K4), which is enriched at the 5′ end of active genes and whose trimethylated product recruits chromatin remodeling activities during transcriptional activation.11,12 During this process, H3R2me2a acts as a repressive marker that can antagonize H3K4 trimethylation,7 whereas H3R2me2s has an effect opposite to that of H3R2me2a.13 Therefore, the status of H3R2 dimethylation could be a critical modulator of gene activation and repression. To provide more insight into this modulation, differentiation of these arginine dimethylated isomers and specific quantification of each isomeric form become important. Currently, the widely used methods for studying protein methylation include amino acid analysis, the incorporation of radioactive methyl-3H groups, Edman sequencing, and peptide mass mapping by matrix-assisted laser desortion ionization (MALDI) mass spectrometry.2 While these methods provide much valuable information, they may only yield qualitative information about the presence of methylation in a given protein,14 and most of the assays could easily fail to identify the type of methylation.15 Liquid chromatography−tandem mass spectrometry (LC−MS/MS)-based targeted proteomics is a technology for detecting proteins of interest and site-specific PTMs with high sensitivity, quantitative accuracy, and reproducibility.16 Our previous work has demonstrated the capability of targeted proteomics.17−21 The key concept of the targeted analysis is to specifically explore the proteins at the peptide level. Surrogate peptides are produced via enzymatic digestion of the target protein, and assays are subsequently B
DOI: 10.1021/acs.analchem.6b00076 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
VA) were cultured in DMEM media supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin at 37 °C under a 5% CO2 atmosphere. MCF-7/ADR (Keygen Biotech, Nanjing, China) cells were maintained routinely in RPMI 1640 media (with L-glutamine and sodium pyruvate) supplemented with 10% fetal bovine serum. The cells were split every 5−7 days using 0.25% trypsin. To maintain a highly drug-resistant cell population, MCF-7/ADR cells were periodically reselected by growing them in the presence of 1000 ng/mL DOX.25 Experiments were performed using the cells cultured without DOX for 48 h. Cell viability was assessed by trypan blue (0.4%) exclusion and the viable cell number was counted with a hemocytometer (Qiujing, Shanghai, China). Most of the available histone extraction methods rely on the solubility of histones in acidic conditions. Under such conditions, most other nuclear proteins and nucleic acids precipitate.26 In this study, cells were first pelleted at 1480g for 10 min and resuspended in 2 mL of ice-cold lysis buffer (10 mM Tris-HCl, 50 mM sodium bisulfite, 1% Triton X-100, 10 mM MgCl2, and 8.6% sucrose, pH = 6.5) containing a protease inhibitor cocktail (Sigma-Aldrich, MO). After dounce homogenization, the samples were spun at 1000g for 7 min. The pellet was submerged in 400 μL of 0.4 N H2SO4 and placed on a shaker at 4 °C to extract the histone proteins overnight.27 After further centrifugation at 12 000g for 10 min, the supernatant was mixed with 1 mL of acetone. Air-dried histone proteins were resuspended in 200 μL of water and stored at −20 °C. The protein concentration was determined using a BCA protein assay kit (Beyotime, Jiang Su, China). In-Solution Enzymatic Digestion. A quantity of 100 μL of each sample was mixed with 50 μL of 50 mM NH4HCO3. Afterward, the proteins were reduced by adding 50 mM DTT to obtain a final concentration of 10 mM. The sample was subsequently incubated at 60 °C for 20 min. The sample was alkylated by the addition of 400 mM IAA until a final concentration of 50 mM was achieved and incubated at room temperature for 1 h in the dark. Finally, thermolysin was added, and the sample was incubated at 80 °C for 6 h in the dark. The digestion was stopped by adding 10% FA to a final concentration of 0.5%. Then, 100 μL of the internal standard solution was added to the peptide mixture. Finally, the sample was evaporated to dryness and resuspended in 100 μL of ACN/ water (50:50, v/v) containing 0.1% FA. LC−MS/MS Assay Development and Validation. Quantification was performed on an Agilent series 1200 HPLC system (Agilent Technologies, Waldbronn, Germany) coupled with a 6410 Triple Quad LC/MS mass spectrometer (Agilent Technologies, Santa Clara, CA). The liquid chromatography separation was performed on a hydrophilic interaction liquid chromatography (HILIC) column (5 μm, 100 mm × 2.1 mm; Agela Technologies, China) at room temperature. The mobile phase consisted of solvent A (10 mM CH3COONH4, pH = 4.2) and solvent B (ACN). A linear gradient at a flow rate of 0.2 mL/min was applied with the following proportions of solvent B: B 80% (0 min) → 80% (1 min) → 40% (3 min) → 40% (9 min) → 80% (10 min). The injection volume was 5 μL. The mass spectrometer was equipped with an electrospray ion source. The peptides were detected in the positive MRM mode. The parameter settings were Q1 and Q3 at unit resolution, capillary voltage at 4000 V, nebulizer pressure at 45 psi, drying gas temperature at 350 °C, drying gas flow at 10 L/
Dimethylated H3R2 was also monitored in MCF-7/WT cells with the treatment of doxorubicin (DOX) for confirmation.
■
MATERIALS AND METHODS Chemicals and Reagents. Nonmethylated (i.e., ARTKQT) and methylated peptides (i.e., AR(me2a)TK(me1)QT, AR(me2s)TK(me1)QT, AR(me2a)TK(me2)QT, and AR(me2s)TK(me2)QT) as well as an internal standard containing stable-isotope coded amino acids were synthesized by the China Peptides Co., Ltd. (Shanghai, China). The purities of the peptides were provided by the manufacturer. The stable isotope-labeled amino acid was purchased from Cambridge Isotope Laboratories, Inc. (Andover, MA). Thermolysin was obtained from Promega (Madison, WI). Ammonium bicarbonate (NH4HCO3) was obtained from the Qiangshun Chemical Reagent Co., Ltd. (Shanghai, China). DLdithiothreitol (DTT) and iodoacetamide (IAA) were supplied by Sigma-Aldrich (St. Louis, MO). Ammonium acetate (CH3COONH4) was supplied by Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Phosphate buffered saline (PBS) was obtained from the Beyotime Institute of Biotechnology (Jiangsu, China). Acetonitrile (ACN) was purchased from the Tedia Company, Inc. (Fairfield, OH). Formic acid (FA), sulfuric acid (H2SO4) and acetone were provided by the Lingfeng Chemical Reagent Co., Ltd. (Shanghai, China). Dulbecco’s Modified Eagle Media (DMEM) and penicillin-streptomycin solution were obtained from the Thermo Scientific Gibco (Logan, UT). Fetal bovine serum was obtained from the Thermo Scientific HyClone & Gibco (Logan, UT). Trypan blue was purchased from the Generay Biotech Co., Ltd. (Shanghai, China). Sodium dodecyl sulfate (SDS) was obtained from the Biosharp Co., Ltd. (Hefei, China). Water was purified and deionized with a Milli-Q system manufactured by Millipore (Bedford, MA). Preparation of Stock Solutions, Calibration Standards, and Quality Controls (QCs). A 1 mg/mL stock solution was first prepared by accurately weighing the methylated peptides and dissolving them in deionized water. The solution was stored at −20 °C in a brown glass tube. An isotope-labeled synthetic peptide (nonmethylated form) was used as a single internal standard. The selection of a suitable internal standard will be described below. In a similar manner, an internal standard stock solution at 5 μg/mL was prepared. Then, a 100 ng/mL internal standard solution was prepared in an ACN/water mixture (50:50, v/v) containing 0.1% FA. Because matrix complexity is one of the significant issues in the quantification of an endogenous analyte for which a true blank is not available, H3-depleted cell extract was used as a matrix in the present study.23 The experimental details about immuno-depletion of cellular extract are provided in the Supporting Information. The concentrations of the calibration standards were 50, 100, 200, 400, 600, 800, and 1000 ng/mL. The QC standards (i.e., lower limit of quantification (LLOQ), low QC (LQC), mid QC (MQC), and high QC (HQC)) were prepared at 50, 150, 400, and 800 ng/mL, respectively. These calibration and QC standards were in the same matrix and frozen prior to use. A series of mock mixtures were made to a total concentration of 1000 ng/mL and were prepared in different concentrations of arginine dimethylated isomers according to the simplex lattice design using the JMP Software Package (SAS Institute Inc., Cary, NC).24 Cell Culture and Histone Isolation. MCF-7/WT (ATCC, Manassas, VA) cells and MCF-10A cells (ATCC, Manassas, C
DOI: 10.1021/acs.analchem.6b00076 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
sDMA-containing peptide isomers are isobaric (i.e., same mass) and thus indistinguishable in the full mass scan. Second, these isomers contain the same dimethylation site and share the same sequence-type product ions, which is different from our previously investigated isobaric phosphopeptides that could have specific sequence-type product ions related to the phospho-site of interest.18,37 Using H3R2me2K4me2 as an example, the two isomeric forms have identical doubly charged precursor ions at m/z 380.8 (Figure S1) and similar mass spectrometric fragmentation patterns. Their representative product ion spectra are shown in Figure 1 (three replicates were performed). Because a variety of fragment ions could be produced and their specificity was difficult to validate,2 only the observed b ions (m/z 72.1 (b1+), m/z 256.3 (b2+), and m/z 321.4 (b52+)) and y ions (m/z 120.2 (y1+), m/z 404.5 (y3+), and m/z 505.2 (y4+)) were labeled. Even though these product ions are not isomeric-specific, the presence of product ions b2 and y3 provides evidence of dimethylation at H3R2. The relative abundance of the product ions was consistent throughout all of the replicates (Table S1). As mentioned earlier, neutral losses or specific reporter masses from side-chain fragmentation of peptides with methylated arginine (e.g., dimethylammonium ion (NH2(CH3)2+, m/z 46.065 Da) and monomethylammonium (CH3NH3+, m/z 32.049 Da)) have been used in several previous studies to represent aDMA and sDMA-containing peptides.2,6 Despite the advances of these approaches, they may not be eligible under certain circumstances. One of the potential issues is that low-mass fragment ions may not be trapped due to the cutoff values of most tandem mass spectrometers.38 In the case that these ions can be detected, their specificity may not be adequate due to the potential interfering peptide-derived product ions obtained at those m/z values. For ARTKQT with more than one methylated site, the situation could be more complicated. To date, evidence has been provided for monomethylated peptide, in which a signal at m/z 32 was apparent.8 Consistently, a fragment ion of m/z 32 was also observed in the spectra of H3R2me1K4me3 here (Figure S2), thus excluding its use for the detection of H3R2me2s. Furthermore, interference could also arise from the adjacent methylated lysine. As evidence, losses of the dimethylammonium ion have been previously discerned in methyl-lysine-bearing peptides.39 To confirm this observation, the product ion spectrum of H3K4me2 was processed and the existence of m/z 46 is evident (Figure S3). Therefore, the employment of these low mass product ions in LC−MS/MS should be circumvented. Finally, arginine dimethylated isomers often coelute in LC− MS/MS and thus fragment in the same tandem mass spectrum.40 Separation methods have been proposed prior to mass analysis, but they do not always work. In the present work, H3R2me2aK4me2 and H3R2me2sK4me2 had close retention times in the extracted ion chromatograms (approximately 8.9 min, R = 0.03, Figure 2). It should be pointed out that methylated peptides usually contain internal hydrophilic residues, which makes them difficult to be captured on a hydrophobic stationary phase in regular LC−MS/MS columns.15 Thus, hydrophilic-interaction liquid chromatography (HILIC) by separations based on the analyte hydrophilicity was introduced here for better retention of methylated peptides.41 Overall, the quantification of H3R2me2a and H3R2me2s by determining the peak areas of isomeric-specific product ions or by chromatographic separation was not trivial. However, we
min. Data analysis was performed using Agilent MassHunter Workstation Software (version B.01.03). Method validation involves calculating the linear range, precision, accuracy, and limit of quantification (LOQ). The detailed procedures and acceptance criteria utilized to validate the assay have been provided in many previous studies.10,28,29 Safety Considerations. Biosafety Level I was observed in all procedures involving MCF-10A, MCF-7/WT, and MCF-7/ ADR cells. Biological waste was treated with bleach prior to disposal. Used cell culture supplies were autoclaved prior to disposal.
■
RESULTS AND DISCUSSION Enzyme Selection for Histone H3 Digestion. Trypsin is the most commonly used enzyme to generate proteolytic peptides in proteomic applications,30 but it is not recommended in certain circumstances. For example, tightly folded proteins such as histones generally resist trypsin digestion.31 In addition, inadequate distribution of trypsin cleavage sites could generate peptides that are too long/too short for mass spectrometric analysis. 31 On the other hand, protein modification often presents another challenge. There is evidence indicating that di- and trimethylation of arginine and lysine residues likely make proteins resistant to trypsin digestion.24,32 Another study reported that no peptides resulting from cleavages after dimethylated arginines (either aDMA or sDMA) were identified in the tryptic digests of proteins.2 Thus, another enzyme or multiple enzymes that can better produce methylated peptides should be preferred. In the case of H3R2, thermolysin was selected as the enzyme of choice. Distinct from other enzymes, thermolysin can tolerate high temperatures and preferentially cleaved at the N-terminus of hydrophobic residues such as alanine. These features make thermolysin an ideal protease for the digestion of tightly folded histones and cleavage of previously inaccessible sites.33 After digestion, the sequence of the generated peptide ARTKQT including residues R2 and K4 was unique to histone H3.1 (accession no. P68431 (H31_HUMAN)) and could be used for a quantitative analysis. Characteristics of Dimethylated H3R2-Containing Peptide Isomers. Histone methylation is probably the most diverse and complex PTM investigated to date, as it can take place on arginine or lysine residues.34 In addition, two forms of arginine methylation (mono (me1)- and di (me2)-) and three forms of lysine methylation (mono (me1)-, di (me2)- and tri (me3)-) are progressively produced during the methylation process. Among these methylated products, only dimethylated arginine exists in two isomeric forms, aDMA and sDMA. Theoretically, each isomeric form of H3R2 (i.e., H3R2me2a and H3R2me2s) can coexist with three forms of H3K4 methylation states. However, there is much evidence indicating that H3R2me2a is mutually exclusive specifically with H3K4me3.7,13,35 Thus, only two combinations H3R2me2K4me1 (AR(me2)TK(me)QT) and H3R2me2K4me2 (AR(me2)TK(me2)QT) were analyzed in this study. In the following analysis, AR(me2a)TK(me1/2)QT and AR(me2s)TK(me1/2)QT are simply referred to as H3R2me2a and H3R2me2s, respectively. It deserves to be mentioned that H3R2me2K4me2, as the strict substrate for trimethylation, could be more closely associated with the biological function of histone H3.36 In general, aDMA and sDMA are not easily differentiated by LC−MS/MS due to their common nature. First, aDMA and D
DOI: 10.1021/acs.analchem.6b00076 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
LC−MS/MS assay for H3R2me2aK4me2 and H3R2me2sK4me2 was developed. The calibration curves for each isomer in the two transitions were regressed using a 1/x2 weighting factor. The relative peak area ratio of the analyte to the internal standard was plotted against concentration. The determined slopes were significantly different using GraphPad Prism 6 software (GraphPad Software, CA; F = 981.831, P < 0.0001). Representative calibration curves are shown in Figure S4. The LOQs were 50 ng/mL. The MRM chromatograms of the LLOQ are shown in Figure S5. Because the matrix should ideally be analyte-free biological samples, the H3-depleted protein extract was utilized as the matrix in the present study. The result indicated that no significant interfering peak was found at the retention time of the surrogate peptides in the chromatograms of the blank matrix (LLOQ response was >5 times the response of the blank matrix, Figure S5). The accuracy and precision of the method were evaluated by analyzing QC samples at four different concentrations of methylated peptides in three replicates. The intra- and interday precisions were expressed in terms of the percent coefficient of variation (% CV). The accuracy was obtained by comparing the mean measured concentrations to their nominal values (% bias). The results are listed in Table S2. The intra- and interday precisions and accuracy were satisfactory for all the QC samples of H3R2me2aK4me2 and H3R2me2sK4me2 (
