Here, we present unbiased quantification of site-specific protein acetylation and/or succinylation occupancy (stoichiometry) of an entire proteome through a ratiometric analysis of endogenous modifications to modifications introduced after quantitative chemical acylation using stable isotope-labeled anhydrides. In combination with sensitive data-independent acquisition mass spectrometry, accurate site occupancy measurements are obtained.
Post-translational modification (PTM) of protein lysine residues by NƐ-acylation induces structural changes that can dynamically regulate protein functions, for example, by changing enzymatic activity or by mediating interactions. Precise quantification of site-specific protein acylation occupancy, or stoichiometry, is essential for understanding the functional consequences of both global low-level stoichiometry and individual high-level acylation stoichiometry of specific lysine residues. Other groups have reported measurement of lysine acetylation stoichiometry by comparing the ratio of peptide precursor isotopes from endogenous, natural abundance acylation and exogenous, heavy isotope-labeled acylation introduced after quantitative chemical acetylation of proteins using stable isotope-labeled acetic anhydride. This protocol describes an optimized approach featuring several improvements, including: (1) increased chemical acylation efficiency, (2) the ability to measure protein succinylation in addition to acetylation, and (3) improved quantitative accuracy due to reduced interferences using fragment ion quantification from data-independent acquisitions (DIA) instead of precursor ion signal from data-dependent acquisition (DDA). The use of extracted peak areas from fragment ions for quantification also uniquely enables differentiation of site-level acylation stoichiometry from proteolytic peptides containing more than one lysine residue, which is not possible using precursor ion signals for quantification. Data visualization in Skyline, an open source quantitative proteomics environment, allows for convenient data inspection and review. Together, this workflow offers unbiased, precise, and accurate quantification of site-specific lysine acetylation and succinylation occupancy of an entire proteome, which may reveal and prioritize biologically relevant acylation sites.
NƐ-acylation of protein lysine residues is an important regulator of protein function. Lysine acetylation and other acylations, such as succinylation, malonylation, and glutarylation, are thought to regulate metabolism, cell signaling, and other cellular processes1,2,3,4, and have implications in metabolic disorders5. Numerous studies have found that lysine acyl modifications undergo large fold-changes under different conditions in mammalian tissues and cell lines5,6,7,8 as well as bacteria9,10,11, however, these relative fold changes do not provide insights into the proportion of the total protein modified. Studies reporting measurements of acylation site occupancy are scarce12,13,14, despite the relevance and need for such studies as they provide more detail than relative fold change. For example, a 10-fold change could represent a site occupancy increase from 0.01 to 0.1%, 1 to 10%, or even 10 to 100%. Accurate stoichiometry measurements are required to interpret biological significance of acylation and to predict impact regarding the magnitude of protein structural, and possibly, functional changes.
One previous method to quantify site occupancy utilizes heavy stable-isotope chemical labeling of unmodified lysine residues followed by mass spectrometry to measure the ratio between endogenous "light" acetylation in comparison to the exogenous, chemically-labeled "heavy" acetyl-lysine using precursor ion intensities12. Another recent study by Zhou et al.13 described a similar approach to assess lysine acetylation stoichiometry that also employed a complete chemical acetylation of all unmodified lysine residues in proteins with stable isotope labeling, but used fragment ion intensities as measured by DDA. Nakayasu et al.14 used a similar DDA approach, but instead used the ratio of light and heavy acetyl-lysine immonium ions for quantification. Quantification based on fragment ions, immonium or sequence ions (MS2), in most cases results in less signal interferences compared to processing intact peptide precursor ions (MS1). However, quantification of fragment ions from DDA-generated MS/MS spectra can suffer from stochastic sampling deficiencies, where the high-abundance precursor ions are more likely to be selected for MS/MS and thus, lead to a biased and incomplete sampling of precursor ions.
This novel workflow4 (Figure 1) uses conceptually a stable isotope chemical labeling approach originally developed by Baeza et al.12, however the workflow is subsequently coupled with DIA to collect both precursor and multiple fragment ion abundances over the detectable mass range15,16 providing accurate stoichiometry calculations.
Peak areas from fragment ions that contain the acylation site of interest, or 'differentiating fragment ions', are used to quantify acylation site occupancy (Figure 2). Fragment ions that do not contain the modification have identical light and heavy m/z values, and are used for peptide identification but not for quantification. Skyline software17 is used to extract precursor and fragment peak areas. The presence of multiple fragment ions containing a given acylation site provides flexibility if interferences are detected in some of the fragment ions. Data visualization in Skyline allows for critical inspection of light-to-heavy fragment ion ratios. In addition to manual data analysis, an open-source R package written in-house, StiochiolyzeR (https://github.com/GibsonLab/StoichiolyzeR), was developed, which uses the precursor ion and fragment ion data collected through DIA and quantifies site-specific acylation stoichiometry from peptides containing multiple lysine residues, a feature not possible with precursor ion-only intensity measurements4.
This protocol also demonstrates the use of endoproteinase Glu-C, which is specific for C-terminal cleavage at glutamic and aspartic acid residues, instead of using trypsin or Arg-C protease, as the latter often generate very large and potentially multiply acylated proteolytic peptides resulting from blocked trypsin cleavage at acylated lysine residues. The chemical labeling procedure was also extended to include use of succinic anhydride (Figure 1), thereby enabling the quantification of succinylation stoichiometry succinylation. In addition to improvements to the sample preparation, the implementation of peptide fractionation by offline, basic pH, reversed-phase (bRP) separation of peptides further decreased quantification interferences, allowing better de-convolution of acylation stoichiometry in whole proteome samples. Together, this method features and highlights several advantages: (1) increased efficiency of chemical acylation; (2) measurement of protein succinylation stoichiometry in addition to acetylation; and (3) improved quantification accuracy. The improved quantification is due to decreased interferences in fragment ion signals from DIA compared to DDA precursor signal, as well as the implementation of off-line peptide pre-fractionation by bRP.
1. Quantitatively Acetylate and/or Succinylate Proteins Using Isotope-labeled Acetic or Succinic Anhydride
2. Digest Reacted Protein Samples Using Glu-C Endoproteinase
3. Fractionate Peptides Using Offline Basic-pH Reversed Phase HPLC
4. DIA Analysis of Fractionated Peptide Samples
5. Example Data Analysis Tutorial
After data acquisition, differentiating MS2 fragment ions were determined from the proteolytic acylated peptides, subsequently extracted ion chromatograms (XIC) were processed in Skyline, and corresponding light and heavy peak areas were exported which were finally used to calculate site occupancy. Figure 2A shows a conceptual illustration of how the precursor-ion XIC may appear featuring both light and heavy peptide signals, and Figure 2B presents an example of the corresponding fragment ion XICs with color coding (red = light; blue = heavy) for differentiating fragment ions containing the lysine acylation modifications.
Figure 3 shows data resulting from an occupancy experiment, such as described above analyzing pre-defined ratios of light/heavy succinylated BSA generated in-house at 1, 10, 50, or 100%, and the determination of succinylation occupancy thereof. Importing DIA occupancy data into Skyline enables easy visualization of the target tree, displaying peptide sequences, and fragment ions both for the light and heavy precursor ions (Figure 3A). Data from DIA-MS of each defined succinylated BSA ratio revealed immanently the relative differences in ratio of light-to-heavy y7 ion, the highest-ranked differentiating fragment ion from the identified peptide-spectra match (indicated in red, Figure 3B). Figure 4 shows an overview of the processed results from the MS2 occupancy calculation that confirmed the succinylation percentages occupancy of input protein, here consisting of the pre-succinylated BSA. Measured lysine succinylation occupancy for 20 proteolytic succinylated peptides obtained from commercial pre-succinylated BSA, succinylated at defined percentages (e.g., at 0, 1, 10, 50, and 100%) are shown in Table 2. For each of the 20 proteolytic BSA peptides, the highest ranked, differentiating MS2 fragment ion was used to calculate lysine succinylation (Ksucc) occupancy, as L/(L+H) in %. Overall, the MS2-based quantification showed very good accuracy in determining acylation occupancy even at low stoichiometries.
Figure 1: Stoichiometry workflow. (A) Proteins are incubated three times with acetic anhydride-d6 to acetylate unmodified lysine residues. Next, acetylated proteins are digested with endoproteinase Glu-C, followed by HPLC fractionation of the proteolytic peptides using basic-pH reversed-phase chromatography. Finally, peptides are analyzed by LC-MS using DIA with variable precursor window widths. (B) The same workflow is used to determine the succinylation stoichiometry as described in (A), except the heavy acylation reagent is changed to succinic anhydride-d4. Figure adapted from Meyer et al.4 (J. Am. Soc. Mass Spectrom., Open Choice). Please click here to view a larger version of this figure.
Figure 2: Example of possible data obtained and stoichiometry calculations. (A) Light (L) and heavy (H) MS1 precursor ions containing one acetylated lysine differ by 3 mass units. XICs are generated for both light and heavy isotopic envelopes indicated in red and blue, respectively. (B) Quantification from the MS2 fragment ion XICs from DIA acquisitions can be performed using 'differentiating' light and heavy fragment ions that contain the acetylation site indicated in red and blue. Common fragment ions are displayed in dotted black and do not contain the site of modification. Figure adapted from Meyer et al.4 (J. Am. Soc. Mass Spectrom., Open Choice). Please click here to view a larger version of this figure.
Figure 3: Skyline visualization of succinylated proteolytic BSA peptide at different occupancy levels. (A) Skyline Target Tree showing the proteolytic peptide LCKsuccVASLRE and its resulting fragment ions. (B) Skyline extracted fragment ion chromatograms at varying levels of succinylation occupancy. The highest ranked differentiating ion, y7, increases in peak area 1, 10, 50, and 100% correlating to the input percentage of pre-succinylated BSA (generated in-house). Please click here to view a larger version of this figure.
Figure 4: Assessing BSA acylation stoichiometry of pre-succinylated BSA. Five different BSA samples at defined percentages of heavy modification (e.g., at 0, 1, 10, 50, and 100%) were subjected to MS2 occupancy determination. Site occupancy for 20 succinylated BSA peptides were determined from the highest ranked differentiating fragment ions for five different BSA samples at defined percentages of light modification (e.g., at 0, 1, 10, 50, and 100%). The Box Whisker plot displays the distribution of the 20 succinyl peptides, values from 50% of peptides are in the 'box' (25% percentile to 75% percentile), the upper whisker indicates the values from 75% percentile to maximum (100%), and the lower whisker indicates the values from 25% percentile to minimum (0% percentile). (J. Am. Soc. Mass Spectrom., Open Choice). Quantification of measurements can be found in Table 2. Please click here to view a larger version of this figure.
Time (minute) | % A | % B |
0.00 | 100 | 0 |
7.27 | 92 | 8 |
45.27 | 73 | 27 |
49.27 | 69 | 31 |
65.27 | 61 | 39 |
72.27 | 40 | 60 |
80.00 | 10 | 90 |
85.00 | 10 | 90 |
86.00 | 100 | 0 |
120.00 | 100 | 0 |
Flow rate: 0.7 mL/min | ||
Buffer A: 10 mM ammonium formate in water, pH 10 | ||
Buffer B: 10 mM ammonium formate in 90% ACN and 10% water, pH 10 | ||
Note: The pH of both mobile phases adjusted to 10 with neat ammonia |
Table 1: Gradient for offline basic-pH reversed-phase HPLC fractionation. Gradient length: 120 min. Buffer A: 10 mM ammonium formate in water, pH 10. Buffer B: 10 mM ammonium formate in 90% ACN and 10% water, pH 10.
L/L+H in % | L/L+H in % | L/L+H in % | L/L+H in % | L/L+H in % |
Ksucc 0% | Ksucc 1% | Ksucc 10% | Ksucc 50% | Ksucc 100% |
0.2 | 1.7 | 11.1 | 50.9 | 99.2 |
1.2 | 2.3 | 12.3 | 49.4 | 97.6 |
2.9 | 3.8 | 14 | 48.6 | 99.5 |
0.5 | 1.8 | 11.7 | 50.8 | 99.5 |
0.2 | 1.7 | 11.1 | 48.3 | 98.2 |
0.2 | 1.2 | 11 | 47.5 | 96.1 |
0.3 | 1.6 | 12.8 | 51 | 99.2 |
3.7 | 5.2 | 14.7 | 51.9 | 89.5 |
1.5 | 1.8 | 12.5 | 47.2 | 91.1 |
0.8 | 1.5 | 11.3 | 48.4 | 96.8 |
0.2 | 1.6 | 13.9 | 49.7 | 98.9 |
0.1 | 1.1 | 10.7 | 48.2 | 98.6 |
0.2 | 0.9 | 10.3 | 49.4 | 99.4 |
0.5 | 2.5 | 17.2 | 52.3 | 97.5 |
0.1 | 3.5 | 20.8 | 51 | 99 |
0.3 | 1.9 | 10.7 | 49.6 | 98.2 |
2 | 1.7 | 10.9 | 47.7 | 96.3 |
0.3 | 1.2 | 10 | 53 | 98 |
1.1 | 1.8 | 12.9 | 57.9 | 94.1 |
0.2 | 1.4 | 11.6 | 48.6 | 98.8 |
Table 2: Quantification of measured BSA lysine succinylation occupancy for 20 proteolytic succinylated peptides. Succinylated peptides obtained from commercial pre-succinylated BSA, succinylated at defined percentages (e.g., at 0, 1, 10, 50, and 100%). For each of the 20 proteolytic BSA peptides, the highest ranked, differentiating MS2 fragment ion was used to calculate lysine succinylation (Ksucc) occupancy, as L/(L+H) in %.
This protocol provides a novel and accurate method to quantify site-specific lysine acetylation and succinylation occupancy that can be applied to an entire proteome. This method relies on measurement of endogenous light peptides and exogenous heavy peptides, the latter of which are generated in vitro using quantitative chemical acylation of proteins with deuterated acetic anhydride-d6 or succinic anhydride-d4 (Figure 1). Similar methods have used stable isotopic labeling of native unmodified lysine residues and performed site occupancy quantification based on either precursor ion-only12 or fragment ions from DDA acquisitions13,14. This protocol applies several steps during sample preparation to improve acylation efficiency and extends the chemical labeling to succinylation. Data collection using DIA acquisitions obtains both precursor ion and MS2 fragment ion intensities over the entire detectable mass range, thus reducing the fragment ion interferences. Analysis and quantification via Skyline and custom scripts developed in-house allow the site occupancy calculation for peptides containing more than one lysine residue and more than one acylation at one site.
There are several critical steps during the sample preparation stage of this protocol that should be followed closely. Since the entire protocol relies on efficient chemical modification of all lysine residues, this step is of utmost importance. Anhydrides react with free amines, so amine-containing buffer or contaminant molecules in the protein lysate must be avoided. Also during the chemical per-acylation step, ensure that the pH of the reaction mixture is adjusted back to pH 8 after each of the three incubations with anhydride reagent as this reaction acidifies the mixture, and O-acylation side-reactions may form. Furthermore, dilution of the 8 M urea-containing reaction mixture to around 0.8 M urea prior to digestion and checking that the pH is within the optimal range is important for the optimum activity of the endoproteinase Glu-C. Another key component of our protocol is the data collection step and the introduction of a DIA workflow, which overcomes any DDA data sampling inconsistencies and which allows for quantification at the MS2 fragment ion level. One main advantage of the DIA methodology is the decrease of interferences which are typically much more prone and problematic when quantifying the MS1 precursor ion signal (as some of the previous publications have suggested).
Several modifications to the workflow can be made when preparing highly complex samples. The number of fractions pooled after the offline basic-pH reversed-phase HPLC fractionation can be decreased to lower the complexity of the pooled samples that will be acquired by LC/MS. Additionally, longer chromatographic gradients can also be used during acquisition to allow for better peptide separation. Alternatively, multiple proteases may be used to digest samples to produce a larger variety of cleaved peptides and increase coverage of protein lysine residues. Trial runs and optimization can be conducted using commercial BSA containing specific percentages of acetylation or succinylation.
One limitation to this protocol is potentially slight site occupancy overestimation due to unaccounted other lysine modifications, such as methylation or ubiquitination, etc., at the same site4. Calculated from the peak areas of light and heavy acyl modifications, L/(L+H), the site occupancy is based on the assumption that the total level of modification at a lysine site consists of only endogenous 'light' acylation (L) and chemically labeled 'heavy' acylation (H). However, the actual total acylation occupancy at a site in vivo may include other modifications beyond acetylation and/or succinylation. Additionally, the reported isotopic purity of the purchased reagents acetic anhydride-d6 (99%) and succinic anhydride-d4 (>98%) may contribute to a small overestimation of up to 2% (succinyl) or 1% (acetyl) of the endogenous acylation level4. Together, these slight overestimations add to the difficulty in quantifying sites with less than 1% occupancy. Large abundance differences between the complete chemically acylated peptides and the native acylated peptides also contribute to potential site occupancy quantification errors, especially for low endogenous acylated peptides27. A recent study by Weinert et al.27 found that decreased abundance differences between complete per-acetylated peptides can reduce the quantification error27.
Stoichiometry quantifications gathered using this protocol can pinpoint important acylation hotspots in a particular protein and form hypotheses for biological follow-up experiments, such as site-directed mutagenesis of certain sites of interest. This protocol could also reveal combined effects of low acylation stoichiometry sites that could exert indirect or subtle influences on protein structural stability and localized cellular environments. Additionally, implementation of this protocol to measure acetylation and succinylation and other possible lysine acylation modifications beyond acetylation would uniquely offer insights into the dynamic acylation effects on proteins under different biological conditions.
The authors have nothing to disclose.
This work was supported by the NIH National Institute of Diabetes and Digestive and Kidney Diseases NIH-NIDDK grant R24 DK085610 (PI: E. Verdin). JGM was supported by a NIH T32 fellowship (T32 AG000266, PI: J. Campisi). The authors acknowledge support from the NIH shared instrumentation grant for the TripleTOF 6600 system at the Buck Institute (1S10 OD016281, PI: B.W. Gibson).
aqueous acetic anhydride-d6 | Sigma Aldrich, St. Louis, MO, USA | 175641-1G | 1% isotopic purity |
succinic anhydride-d4 | Sigma Aldrich, St. Louis, MO, USA | 293741 | 2% isotopic purity |
sequencing grade endoproteinase Glu-C | Roche, Indianapolis, IN, USA (through Sigma) | 10791156001 | 1:50 protease to substrate protein ratio (wt:wt) |
Oasis HLB Solid-Phase Extraction (SPE) cartridges | Waters Corp., Milford, MA, USA | WAT 094225 | |
HPLC acetonitrile | Burdick and Jackson, Muskegon, MI, USA | BJAH015-4 | |
HPLC grade water | Burdick and Jackson, Muskegon, MI, USA | BJAH365-4 | |
iodoacetamide | Sigma Aldrich, St. Louis, MO, USA | I1149-25G | |
dithiothreitol (DTT) | Sigma Aldrich, St. Louis, MO, USA | D9779-5G | |
FLUKA formic acid (FA) for mass spec | Thomas Scientific, Swedesboro, NJ, USA | C988C27 | ~98% |
urea | Thermo Fisher Scientific, Waltham, MA, USA | PI29700 | |
bovine serum albumin (BSA) | Pierce, Rockford, IL, USA | 88341 | |
triethylammonium bicarbonate (TEAB) | Sigma Aldrich, St. Louis, MO, USA | T7408-100ML | |
anhydrous dimethyl sulfoxide (DMSO) | Alfa Aesar, Tewksbury, MA, USA | 43998 | |
sodium hydroxide (NaOH) | EMDMillipore, Burlington, MA, USA | SX0590 | |
50% hydroxylamine solution in water | Sigma Aldrich, St. Louis, MO, USA | 438227-50ML | |
LC-MS grade water | Burdick and Jackson, Muskegon, MI, USA | BJLC365-4 | |
LC-MS acetonitrile | Burdick and Jackson, Muskegon, MI, USA | BJLC015-4 | |
Eppendorf Themomixer Compact | Eppendorf AG, Hamburg, Germany | T1317-1EA | |
Thermo Scientific Savant SPD131DDA SpeedVac Concentrator | Thermo Fisher Scientific, Waltham, MA, USA | SPD131DDA-115 | |
Waters 1525 binary HPLC pump system | Waters Corp., Milford, MA, USA | WAT022939 | |
Waters 2487 Dual Wavelength UV detector | Waters Corp., Milford, MA, USA | WAT081110 | |
Waters 717plus Autosampler | Waters Corp., Milford, MA, USA | WAT022939 | |
Waters Fraction Collector III | Waters Corp., Milford, MA, USA | 186001878 | |
Agilent Zorbax 300Extend C18 column | Agilent Technologies, Inc., Santa Clara, CA, USA | 770995-902 | |
Labconco Freeze Dry System Freezone 4.5 Lyophilizer | Labconco, Kansas City MO, USA | 7751000 | |
indexed retention time (iRT) normalization peptide standard | Biognosys AG, Schlieren, Zurich, Switzerland | Ki-3002-2 | |
Ultra Plus nano-LC 2D HPLC system | Sciex LLC, Eksigent Division, Framingham, MA, USA | model# 845 | |
orthogonal quadrupole time-of-flight (QqTOF) TripleTOF 5600 mass spectrometer | Sciex LLC, Framingham, MA, USA | per quote | |
orthogonal quadrupole time-of-flight (QqTOF) TripleTOF 6600 mass spectrometer | Sciex LLC, Framingham, MA, USA | per quote | |
C18 pre-column chip (200 μm × 6 mm ChromXP C18-CL chip, 3 μm, 300 Å) | Sciex LLC, Framingham, MA, USA | 804-00006 | |
SWATH 2.0 plugin into PeakView 2.2 | Sciex LLC, Framingham, MA, USA | software download Sciex | |
Spectronaut | Biognosys AG, Schlieren, Zurich, Switzerland | Sw-3001 | |
C18-CL chip (75 µm x 15 cm ChromXP, 3 µm, 300 Å) | Sciex LLC, Framingham, MA, USA | 804-00001 |