ETD Outperforms CID and HCD in the Analysis of the Ubiquitylated Proteome
- First Online:
- Cite this article as:
- Porras-Yakushi, T.R., Sweredoski, M.J. & Hess, S. J. Am. Soc. Mass Spectrom. (2015) 26: 1580. doi:10.1007/s13361-015-1168-0
Comprehensive analysis of the ubiquitylome is a prerequisite to fully understand the regulatory role of ubiquitylation. However, the impact of key mass spectrometry parameters on ubiquitylome analyses has not been fully explored. In this study, we show that using electron transfer dissociation (ETD) fragmentation, either exclusively or as part of a decision tree method, leads to ca. 2-fold increase in ubiquitylation site identifications in K-ε-GG peptide-enriched samples over traditional collisional-induced dissociation (CID) or higher-energy collision dissociation (HCD) methods. Precursor ions were predominantly observed as 3+ charged species or higher and in a mass range 300–1200 m/z. N-ethylmaleimide was used as an alkylating agent to reduce false positive identifications resulting from overalkylation with halo-acetamides. These results demonstrate that the application of ETD fragmentation, in addition to narrowing the mass range and using N-ethylmaleimide yields more high-confidence ubiquitylation site identification than conventional CID and HCD analysis.
KeywordsUbiquitin Ubiquitylation Mass spectrometry Proteomics K-ε-GG antibody Electron transfer dissociation Collisional-induced dissociation
Proteomic studies have sought to understand the role of ubiquitin conjugation by identifying proteome-wide ubiquitin modified proteins. Recent advances in the use of mass spectrometry have greatly enhanced our current understanding of the repertoire of proteins that are ubiquitylated [1, 2, 3, 4, 5, 6]. The combined use of the di-glycine (GG) remnant antibody [7, 8] and advances in orthogonal peptide prefractionation, have led to the identification of more than 10,000 putative ubiquitylation sites in human cell lines [6, 9, 10, 11, 12, 13]. Advances in mass spectrometers have been instrumental in the characterization of the ubiquitylome; however, optimization of specific parameters needs further exploration. Currently, the highest number of ubiquitylation sites reported in Saccharomyces cerevisiae, Baker’s yeast, is 2299, but this is believed to be a fraction of the total sites that can occur in the whole proteome, especially when one considers that most proteins are eventually targeted for degradation by the ubiquitin proteasome system (UPS) [12, 14].
Ubiquitin is a regulatory protein that is conjugated to target proteins via a cascade of E1, E2, and E3 enzymes that determine the site and complexity of ubiquitylation. Conjugation of ubiquitin to a specific lysine residue in the target protein results in the formation of an isopeptide bond between the ε-amino group of lysine and the C-terminus of ubiquitin. Fortunately, the C-terminus of ubiquitin ends in RGG, allowing ubiquitylated proteins to yield tryptic peptides that have a GG remnant covalently attached to the modified lysine, giving rise to K-ε-GG peptides. Therefore, the GG moiety can be used to identify specific lysine residues in the target protein where ubiquitin was linked, based on a mass modification of +114.0429.
Promising studies using electron capture dissociation (ECD) on a Fourier transform ion cyclotron mass spectrometer (FT-ICR) was able to use ECD fragmentation to fully characterize the ubiquitylation sites on a tryptically digested purified protein and additionally characterize the autoubiquitylation sites on ubiquitin demonstrating that a softer fragmentation method could better preserve the GG moiety at the site of modification . Later fundamental studies, using electron transfer dissociation (ETD), suggested that a higher yield of ubiquitylation identifications could be achieved using ETD instead of CID and HCD because it is more amenable to fragmenting higher charged species and preserving PTMs than CID [19, 20, 21]. The use of ETD techniques for the preferential fragmentation of GG modified peptides has been tested on single ubiquitylated proteins and has been shown to result in greater site coverage than CID [20, 22]. However, the use of ETD on a complex mixture of digested proteins enriched in ubiquitylated peptides has so far not been tested. Although it has previously been shown that individual standard GG peptides have benefited from ETD analysis, the proteomics field has been hesitant to adopt ETD fragmentation on global ubiquitylation site analyses, possibly because the increased length of the duty cycles is expected to negatively affect the outcome. Additionally, ETD requires specialized equipment and optimization, which may have discouraged adoption. In this study, we demonstrate that ETD outperforms both CID and HCD in a global ubiquitylation analysis and explore methodological and mass spectrometric factors that will encourage the use of ETD for global ubiquitylation studies.
Overalkylation is an added problem that must be considered while performing ubiquitylome analysis [23, 24]. Nielsen et al. demonstrated that overalkylation of proteins with iodoacetamide (IAA) can result in the addition of two carbamidomethyl groups at lysine residues, equal in molecular composition (C8H12N4O4) and mass (+114.04292 Da) to a GG remnant . The authors suggested instead using chloroacetamide (ClAA) because it is less reactive . However, the underlying problem can remain since overalkylation with IAA and ClAA yield the same product leading to potential false PTM assignments. N-ethylmaleimide (NEM) is an alternate alkylating agent that can be used to prevent disulfide bridges from reforming, but most importantly the use of NEM adds a mass tag of +125.0477 Da. By using NEM to alkylate the side chains of cysteine residues, we circumvent the potential problem since the mass modification by NEM would be different than that of the GG remnant.
In this study, we explored GG peptide identification by CID, HCD, ETD, and data-dependent decision tree (DT) fragmentation. We demonstrate the vast improvement made by the use of ETD fragmentation for the proteome-wide identification of ubiquitylation supported by the prevalence of these peptides in higher charge states. In our study, ETD fragmentation resulted in the identification of a higher number of unique GG peptides in a complex proteomic sample than traditional CID or HCD analysis.
Synthetic Peptide Sample Processing
Custom synthetic peptides purchased from JPT Innovative Peptide Solutions (Berlin, Germany) were synthesized with the following amino acid sequence AMLK(GG)SEQNR, where (GG) represents a di-glycine unit connected to the peptide via the ε-amino group of lysine. Dried peptides were resuspended in LC-MS water to a concentration of 1 nmol/μL and stored at –20°C. Immediately before analysis, peptide solutions were diluted to a concentration of 500 fmol/μL and a total of 1 pmol was analyzed by nanoLC-MS/MS.
Yeast strains and Growth Conditions
Wild type S. cerevisiae strain, RJD360 (W303 background) was used in this study. Yeast cells were grown at 30°C in YPD media using standard methods and conditions. A culture of 800 mL was inoculated at a starting OD600nm of 0.1 and grown to an optical density of 0.6–1.0. Cells were then harvested by centrifugation at 5000 × g for 10 min at 4°C, washed once with 20 mL of ice-cold water, and harvested at 5000 × g for 5 min at 4°C. The cell pellet was then frozen in liquid nitrogen and stored at –80°C until lysis.
Digestion and Desalting
Cell lysis, digestion and peptide desalting procedures were adapted from the PTMScan Ubiquitin Remnant Motif (K-ε-GG) Kit #5562 Cell Signaling Technology product manual. Briefly, yeast cells were lysed in 5 mL of lysis buffer (20 mM HEPES (pH 8.0), 9 M urea, 1× protease inhibitor cocktail (Promega, Madison (WI), USA), 1 mM PMSF) and 4 mL of glass beads by vortexing 1 min followed by a 1 min incubation on ice, seven times. The lysate was collected and centrifuged at 16,000 × g for 15 min. Protein concentration of the lysate was then determined by Bradford. Cleared lysate containing 10 mg of protein was reduced for 45 min by adding 1/278th (v/v) of 1.25 M DTT. Alkylation of cysteines was performed by treating the lysate with 250 mM NEM dissolved in H2O (25× stock) to achieve a final concentration of 10 mM NEM, for 30 min at room temperature in the dark. For trypsin digestion, lysate was diluted to 2 M urea by adding 100 mM Tris (pH 8.0). Proteins were digested by trypsin using a ratio of 1:100. Digestion was carried out overnight (≥15 h) at room temperature in the dark. The following morning the reaction was quenched by the addition of formic acid to a final concentration of 0.2%.
Digested peptides were centrifuged at 16,000 × g for 15 min to remove insoluble material. Cleared peptides were desalted by SepPak using a 500 mg capacity column. Briefly, resin was hydrated using 7 column volumes of acetonitrile (21 mL), followed by equilibration with 7 column volumes of Buffer A (0.2% TFA in H2O) (21 mL). Peptides were loaded onto the resin by gravity flow. After binding, the resin was washed with 7 column volumes of Buffer A and 3 column volumes of wash buffer (0.2% TFA, 5% acetonitrile in H2O). Desalted peptides were recovered using 2 column volumes of elution buffer (0.2% TFA, 40% acetonitrile in H2O) and lyophilized to dryness.
K-ε-GG Antibody Cross-Linking and Immunoprecipitation
In short, K-ε-GG peptide-specific antibody (PTMScan Ubiquitin Remnant Motif (K-ε-GG) Kit #5562, Limited Use License, Cell Signaling Technology) was washed with 3 × 1 mL aliquots of 100 mM sodium borate (pH 9.0). Antibody bound beads were pelleted after each wash by centrifugation at 2000 × g for 30 s and kept on ice whenever possible. After washing, the beads were incubated for 30 min in 1 mL of DMP cross-linking solution (100 mM sodium borate, pH 8.0, 20 mM dimethyl pimelimidate, DMP) for 30 min at room temperature with gentle rotation. The cross-linking reaction was quenched by first washing the beads with 3 × 1 mL aliquots of 200 mM ethanolamine blocking buffer (pH 8.0) then incubating with 1 mL of ethanolamine blocking buffer for 2 h at 4°C. After blocking the antibody-bound beads were washed with 3 × 1 mL aliquots of 1X IAP buffer (50 mM MOPS, pH 7.2, 10 mM sodium phosphate, and 50 mM NaCl), then incubated with the desalted peptide sample for 1 h at 4°C. Before incubating with cross-linked antibody, the desalted peptide sample was first resuspended in 1.0 mL of 1X IAP buffer, the pH was measured (should be pH ≅ 7), and cleared by spinning at maximum speed for 5 min. After incubating the beads with the peptide sample, the beads were pelleted by centrifugation at 2000 × g for 1 min, resuspended in 500 μL of 1X IAP, and transferred to a 0.67-mL tube and washed three times with 500 μL of 1X IAP buffer. Following the IAP washes, the beads were washed twice with 1X PBS and once with mass spectrometry grade water (Fluka, Seelze, Germany). Finally, the bound K-ε-GG peptides were eluted with 2 × 150 μL aliquots of 0.15% TFA, each time incubating the beads with elution buffer for 10 min at room temperature with constant mixing. The eluents were combined, dried, desalted by HPLC using a Michrom Bioresources, (Auburn (CA), USA) C18 macrotrap, (Buffer A: 0.2% formic acid in H2O; Buffer B: 0.2% formic acid in acetonitrile) and concentrated in vacuo.
Mass Spectrometry Settings for ETD and DT Analysis
ETD analysis values
DT analysis values
MS mass range
Minimum signal required for MS/MS
Number of microscans
Number of data-dependent MS/MS
FT master scan preview
Automatic gain control (AGC) Target Value for MS
1 × 106
1 × 106
AGC Target value for MS/MS
5 × 103
5 × 103
Maximum ion injection time for MS/MS
Dynamic exclusion duration
ETD reaction time
Use decision tree or other parameters
Default charge state parameter
CID = 2+ (300–1200 m/z)
ETD ≥ 3+ (300–1200 m/z)
For the ubiquitylome analysis, a GG peptide-enriched sample immunoprecipitated from yeast whole cell lysate was analyzed by five technical replicates of CID, HCD, ETD, and DT. A total of 20 raw files were processed together using MaxQuant (ver. 18.104.22.168) [27, 28, 29]. Spectra were searched against the yeast proteome from SGD (5898 entries, download 1/15/2010) and a contaminant database (245 entries) as well as a decoy database of equal size. Protein, peptide, and site false discovery rates were less than 1% across the entire dataset and were estimated using a target-decoy approach . Trypsin was specified as the digestion enzyme with up to two allowed missed cleavages. Precursor mass tolerance was 6 ppm and fragment ion tolerance was 0.5 Da for CID and ETD analyses, whereas a fragment mass tolerance of 20 ppm was used for HCD. At the spectrum level, a posterior error probability of less than 0.05 gave a spectrum FDR of less than 0.001 across the entire dataset. N-ethylmaleimide modification of cysteine (+125.0477) was specified as a fixed modification. Variable modifications included protein N-terminal acetylation (+42.0106), methionine oxidation (+15.9949), and the GG remnant (+114.0429). Losses of 57.0215 and 114.0429 Da were used to account for fragmentation in the GG remnant. Peptides modified at the C-terminus by a GG adduct were excluded from the MaxQuant analysis and consequently from the final results. Please note, by “unique site” we are referring to a unique position in a protein that is modified by ubiquitin, whereas “unique peptide” refers to a specific combination of residues and modifications in a peptide. Figures 2 and 4 were plotted to show the differences between the mean performances for the various fragmentation methods as recommended by Kryzywinski and Altman .
Results and Discussion
Comparing Fragmentation Strategies
Our initial study began by trying to understand the nature of GG peptides and, consequently, use this information to select the optimal fragmentation strategy for their analysis. A synthetic peptide containing a GG remnant covalently attached to the ε-amino group of the single internal lysine residue in the peptide was analyzed by conventional CID and HCD fragmentation and compared with ETD fragmentation. The sequence of the synthetic peptide is depicted in Figure 1a and illustrates the prevalence of GG peptides in higher charge states attributable to the presence of an additional N-terminus. The abundance of the triply charged species was compared with that of the doubly charged species in Skyline  and found to be considerably higher (Figure 1b). Additional synthetic GG containing peptides were evaluated and found to display a similar trend (data not shown). MS/MS spectrum produced during ETD fragmentation of the synthetic peptide, AMLK(GG)SEQNR, is illustrated in Figure 1c. Complete sequence coverage was observed when the AMLK(GG)SEQNR peptide was fragmented by ETD. We were able to reliably observe fragment ions c2-c8 along with z2-z8 in the nine-residue peptide. CID fragmentation of the synthetic peptide was also performed and we observed similar fragmentation coverage of the synthetic peptide sample (Supplementary Figure 1).
Another consideration in the identification of ubiquitin modification sites in global analyses is the possibility of observing false positive identifications resulting from overalkylation. During bottom-up ubiquitylome analyses, before digestion with trypsin, proteins are reduced, then alkylated to prevent the reforming of disulfide bridges between cysteine residues. Unfortunately, as with all proteomic analyses, the more treatments performed, the greater the likelihood of observing a larger mixture of contaminant peptides. In the case of using the alkylation reagent iodoacetamide (IAA), the possibility of observing false positive identification of GG-containing peptides increases, resulting from the equivalent mass and atomic composition of a carbamidomethyl group and a glycine [22, 23]. If two carbamidomethyl groups become attached to the ε-amino group of lysine, it is practically indistinguishable from the addition of a GG remnant and may lead to a false identification. This observation has been reported before and the authors concluded that the use of chloroacetamide (ClAA) is better suited for ubiquitylome analyses because it is less reactive than IAA and should prevent the addition of carbamidomethyl groups at lysine residues . However, the problem remains that the mass tag of the reagent is still equivalent to the mass tag of a glycine residue and, therefore, indistinguishable from a true GG addition, if overalkylation occurs. We chose to preclude any false positive identification of ubiquitylation sites due to overalkylation by using NEM to alkylate reduced cysteine residues in our peptide samples. The conditions for treating a complex mixture with NEM were determined (data not shown) and found to be optimal at a concentration of 10 mM for 30 min at room temperature in the dark. At this concentration, we achieved adequate modification of cysteine residues, while minimizing NEM adduct addition to non-cysteine residues. For the remainder of the ubiquitylation analyses described, 10 mM NEM was used for the alkylation of cysteine residues.
Analysis of Ubiquitylation Identifications
To increase confidence in the ubiquitylation sites presented, we performed a thorough analysis of the sites identified. Potential false positives were excluded from consideration in MaxQuant by restricting GG remnant modifications to internal or peptide N-terminal lysines. This is because trypsin is unlikely to cleave on the carboxyl side of a modified lysine . Additionally, a localization probability of 0.5 or higher was used to exclude false identification of sites. If a site was identified as having a localization probability equal to 0.5 at two different lysine residues in the same peptide, they were combined into a single identification. A complete list of ubiquitylation sites identified in this study is presented in Supplementary Table 1. Sites identified are listed along with their corresponding protein open reading frame, position in the protein, modified peptide sequence, precursor charge states, and whether the site was previously reported in SCUD, SGD, or other recent studies [12, 14]. Additionally, the fragmentation methods in which each site was observed are included in the column labeled “Fragmentation method site observed” and a column indicating if the site was only observed while using ETD or DT fragmentation has also been included. The specific site of modification for each entry listed in the “Modified peptide sequences” column is represented by a K(gl) highlighted in bold. Interestingly, of the 235 unique ubiquitylation sites identified in this study, 106 sites (45%) were only found by ETD and/or DT, and not by CID and HCD; and 38 of these sites had not been previously reported in the literature [12, 14]. Another interesting aspect of this study is that the yeast samples used were not treated with proteasome inhibitors or peptide pre-fractionated prior to performing the GG-peptide immunoprecipitation. Future studies should demonstrate the usefulness of ETD fragmentation to more highly enriched GG-peptide samples.
Peptide spectrum matches of all ubiquitylation sites identified in this study are included in the Supplementary Data (SiteSpectra.pdf). Sequence information on the GG peptides identified is listed in Supplementary Table 1, whereas sequence coverage information on all proteins identified in this analysis is included in Supplementary Table 2. A total of 1141 proteins were identified in this study and 216 unique proteins were found to be ubiquitylated, of which are included protein isoforms and paralogs. Contaminants and decoys were removed from the tables for brevity. For completeness, we also determined the duty cycle values for all fragmentation methods used. The duty cycle time for CID was ~1.4 s, HCD was ~7 s, ETD was ~4.8 s, and DT was ~4.2 s. Although the duty cycle times were considerably longer for the ETD and DT methods compared with CID fragmentation, we were still able to identify a higher number of ubiquitylation sites using ETD fragmentation.
In this study, we explored several parameters for increasing the number of high-confidence ubiquitylation site identifications. We used the novel approach of applying ETD and DT fragmentation to the proteomic analysis of ubiquitylated peptides. Additionally, we precluded false identification of ubiquitylation sites by using the alkylation agent NEM. Finally, we extensively curated our list of identifications by excluding false positive GG peptide identifications resulting from unlikely tryptic cleavages. We conclusively demonstrate that ETD fragmentation in either an all ion or DT method greatly improves the comprehensive analysis of ubiquitylation proteome-wide. In the future, we anticipate ETD being applied to other global ubiquitylation studies and aiding in better cataloguing sites of ubiquitylation in future biological work.
The authors thank Dr. Raymond J. Deshaies for helpful suggestions and critical discussions of the work described. In addition, they thank Dr. Deshaies for kindly providing them with the yeast strain RJD 360 (W303 background). Lastly, they thank members of the Proteome Exploration Laboratory, housed in the Beckman Institute at the California Institute of Technology, for helpful discussions during the course of this work. This work was supported by the Gordon and Betty Moore Foundation, through grant GBMF775, the Beckman Institute and the NIH through grant 1S10RR029594-01A1.