Free Radical–Initiated Peptide Sequencing Mass Spectrometry for Phosphopeptide Post-translational Modification Analysis
Free radical–initiated peptide sequencing mass spectrometry (FRIPS MS) was employed to analyze a number of representative singly or doubly protonated phosphopeptides (phosphoserine and phosphotyrosine peptides) in positive ion mode. In contrast to collision-activated dissociation (CAD) results, a loss of a phosphate group occurred to a limited degree for both phosphoserine and phosphotyrosine peptides, and thus, localization of a phosphorylated site was readily achieved. Considering that FRIPS MS supplies a substantial amount of collisional energy to peptides, this result was quite unexpected because a labile phosphate group was conserved. Analysis of the resulting peptide fragments revealed the extensive production of a-, c-, x-, and z-type fragments (with some minor b- and y-type fragments), suggesting that radical-driven peptide fragmentation was the primary mechanism involved in the FRIPS MS of phosphopeptides. Results of this study clearly indicate that FRIPS MS is a promising tool for the characterization of post-translational modifications such as phosphorylation.
KeywordsFree radical–initiated peptide sequencing (FRIPS) Radical-driven tandem mass spectrometry Phosphopeptides Phosphorylation Post-translational modifications (PTMs)
Tandem mass spectrometry is an essential analytical method in the identification and characterization of peptides and proteins, making it as a valuable tool in proteomics research. Collision-based tandem mass spectrometry, such as collision-induced dissociation (CID) (or collision-activated dissociation (CAD)), is the most widely used technique [1, 2, 3]. However, in recent years, electron-based tandem mass spectrometry methods (electron-activated dissociations (ExDs)), such as electron capture (transfer) dissociation, are also used as complementary tools for collision-based dissociation methods [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]. It is well known that ExDs are exceptionally effective at characterizing post-translational modifications (PTMs), including phosphorylations and glycosylations [5, 6, 12] because labile PTM groups typically remain intact, whereas the peptide backbone undergoes dissociation. Through extensive theoretical and experimental mechanistic studies, it was revealed that an H-radical species is essential to the unique ExD process [4, 5, 8, 9, 15, 18].
The discovery and confirmation of radical species’ important role in peptide-backbone dissociation mass spectrometry has led to the development of a variety of techniques for introducing a radical site in peptides and proteins. Collisional activation (CA) of the transition metal–ligand ternary complexes, [Mn+(L)m−(P)](n − m)+, was developed by Siu and colleagues to introduce a radical site in peptides, where M is a transition metal, L is a ligand, and P is a peptide [25, 26, 27, 28, 29, 30, 31]. Alternatively, UV photolysis of the iodine-containing compound was mainly explored by the Julian group [32, 33]. Photodetachment of anionic peptides offers another effective method of introducing a radical in peptides [34, 35]. As another innovative method, Beachamp and Porter pioneered a method wherein a radical site is generated from a precursor that is conjugated with peptides [36, 37, 38, 39]. Once a radical site is generated on a peptide, the unique radical-initiated peptide-backbone dissociation process ensues usually via thermal heating by CA. So far, extensive experimental and theoretical studies have enhanced the understanding of the radical’s stability as well as that of radical transfer and radical-initiated dissociation mechanisms [28, 29, 31, 32, 33, 38, 39, 40, 41].
Our group has focused on prolonged efforts to improve the use of TEMPO-assisted FRIPS MS as a practical tool for the proteomics research. To achieve this, a few drawbacks associated with this method had to be overcome. First, it was found that the lysine-containing peptides (particularly formed by tryptic digestion) underwent double conjugations with o-TEMPO–Bz–C(O)–NHS at the N-terminal amine and the side chain of the lysine residue. To circumvent the double conjugation issue, a guanidination reaction strategy was introduced, whereby the primary amine side chain of lysine was blocked from further conjugation with o-TEMPO–Bz–C(O)–NHS by being guanidinated to homoarginine . Second, it would be desirable if the TEMPO-assisted FRIPS MS could proceed by a single application of CA instead of two ensuing applications of CAs (one for radical generation and the other for radical-driven peptide dissociation). Recently, it was demonstrated that the TEMPO-assisted FRIPS MS could proceed in a single step in the Orbitrap mass spectrometer and the hybrid quadrupole time-of-flight (Q-TOF) mass spectrometer even in positive ion mode, whereas a two-step application of CA was required in the (linear) ion-trap mass spectrometer in positive ion mode . Higher collision energies and multiple collisions of the produced benzyl radical–containing peptide species in the abovementioned two instruments (Orbitrap and Q-TOF MS) presumably enabled the single-step FRIPS MS (note that the single-step FRIPS MS was also possible even in the ion-trap instrument in negative ion mode) .
To extend the applicability of the TEMPO-assisted FRIPS MS approach, the possibility of characterizing phosphopeptides was investigated. Although the characterization of phosphopeptides can be readily achieved by ExD methods [5, 7, 11], it is not so obvious whether TEMPO-assisted FRIPS MS can characterize phosphopeptides because TEMPO-assisted FRIPS MS uses CA. The phospho–ether bond is so fragile that the phospho [HPO3 (80 Da) or H3PO4 (98 Da)] group easily detaches, hindering the localization of the phosphorylation sites [56, 57, 58, 59, 60]. If the CA energy requirement for the homolytic cleavage of the C–O bond between the TEMPO and benzene to generate a radical site is higher than the bond dissociation energy of the phospho–ether bond, the characterization of the phosphopeptides using the TEMPO-assisted FRIPS is not possible. In the other case, the phosphopeptides can be readily analyzed by our FRIPS approach. The detailed description of the FRIPS MS results for a number of model phosphopeptides, including a few phosphoserine and phosphotyrosine peptides is presented below.
All phosphopeptides were custom-synthesized by Anygen (Gwangju, Korea) and used as provided from the supplier. o-TEMPO–Bz–C(O)–NHS was obtained from FutureChem (Seoul, Korea), and water and methanol of HPLC grade were purchased from Burdick & Jackson (Ulsan, Korea). Anhydrous dimethyl sulfoxide (DMSO), acetic acid, and triethylammonium bicarbonate (TEAB) were obtained from Sigma (St. Louis, MO, USA).
Peptides were conjugated with o-TEMPO–Bz–C(O)–NHS as described in previous studies [42, 43, 44, 45, 46, 47, 48, 49, 50]. A dried phosphopeptide was dissolved in anhydrous DMSO at a concentration of 100 μM. Separately, 1 mg of o-TEMPO–Bz–C(O)–NHS was dissolved in 100 μL anhydrous DMSO solvent. Twenty microliters of the peptide, 12 μL o-TEMPO–Bz–C(O)–NHS, and 100 mM TEAB buffer were all added to 100 μL of DMSO solution. This solution was vortexed for 1 min and reacted for 6 h. The final o-TEMPO–Bz–C(O)–peptide was diluted ten times with a solution of acetic acid:methanol:water (2:49:49, v/v/v).
FRIPS MS and CAD experiments were performed on an electrospray ionization (ESI) linear ion-trap (LTQ XL) and LTQ-Orbitrap mass spectrometer (Thermo Fischer Scientific, San Jose, CA, USA) in its positive ion mode. The o-TEMPO–Bz–C(O)–peptide sample solution was directly infused at a flow rate of 3–5 μL/min using a built-in syringe pump. The LTQ XL instrumental parameters were set as follows: sheath gas (N2) flow rate, 15 (arbitrary units); spray voltage, + 4–5 kV; capillary temperature, 275°C; capillary voltage, + 35 V; tube lens, + 70 V; isolation window, ± 3 m/z width; AGC target, full MS (30,000), MSn (10,000); and maximum ion injection time, 100 ms. Normalized collision energies (NCEs) at MS/MS and MS3 were optimized in the range of 20–30% to achieve high-yield fragmentations. For higher-energy collision dissociation (HCD) experiments using an LTQ-Orbitrap instrument, the following experimental conditions were used: a heated ESI (HESI) source; source temperature, 40°C; flow rate, 5 μL/min; resolution, 60,000; sheath gas (N2) flow rate, 15; spray voltage, + 3.5 kV; capillary temperature, 300°C; capillary voltage, + 35 V; tube lens offset voltage, + 60 V; isolation width, ± 3 m/z width; HCD NCE, 30%; AGC target, full MS (500,000), MS/MS (200,000); and maximum ion injection time, 500 ms. Each spectrum was averaged over a minimum of 50 scans and recorded.
Results and Discussion
Comparison Between CAD and TEMPO-Assisted FRIPS MS Results
FRIPS MS was also performed in the linear ion-trap mass spectrometer for the two positively charged, phospho-peptides: o-TEMPO–Bz–C(O)–AVLTpSGIELR and o-TEMPO–Bz–C(O)–GLAIRApYGI. In the first step of FRIPS MS (i.e., MS/MS), a ·Bz–C(O)–peptide radical species was exclusively formed. In the second step of FRIPS MS, i.e., MS3, CA was given to the generated ·Bz–C(O)–peptide. As shown in Figure 1b, d, CA of the ·Bz–C(O)–peptides resulted only in a limited loss of the phosphate group while leading to extensive peptide-backbone dissociations. Considering that CA was provided for the ·Bz–C(O)–peptides to initiate the radical-driven peptide dissociations in FRIPS MS, this result is quite surprising. This result was also confirmed in a very recent work by Borotto et al., in which TEMPO-assisted FRIPS MS for phosphopeptides was performed in negative ion mode . Although phosphoserine-containing peptides readily lose the phosphate group, i.e., –HPO3 (80 Da) or –H3PO4 (98 Da), upon CAD, FRIPS MS (i.e., CA of ·Bz–C(O)–AVLTpSGIELR) mostly resulted in radical-driven peptide dissociation without the loss of the phosphate group, except for [y6 − H3PO4 + 2H]+ at m/z 656 and [z7 − H3PO4]+ at m/z 740.
As indicated in Figure 1b, d, a-, c-, x-, and z-type peptide-backbone fragments were the major products, as previously reported [40, 46]. These fragments arose from radical-driven peptide-backbone dissociation pathways, and their fragmentation mechanisms have been previously suggested : note that a suggested mechanism for TEMPO-assisted FRIPS MS is given in the Supplementary material (Supplementary Scheme S1). A benzyl radical can abstract a hydrogen atom from the α-carbon (Cα) of an amino acid side chain, and thus, a radical site is transferred to Cα. The Cα-centered radical then undergoes β fragmentation to produce a/x or c/z peptide-backbone fragments (see Supplementary Scheme S1).
For the phosphotyrosine-containing peptide GLAIRApYGI, the loss of a phosphate or phosphoric acid was observed only for a few fragments (e.g., [z5 − HPO3 − H]+, [y7 − H3PO4 + H]+·, and [rb7 − H3PO4]+) and most of the observed peptide-backbone fragments retained the phosphoric acid. With the phospho group retained, the localization of the phosphorylated group along the peptide backbone of AVLTSGIELR and GLAIRAYGI could readily be made in the TEMPO-assisted FRIPS MS.
Another important feature found in Figure 1b, d is the peaks arising from the neutral loss of side chains, for example the loss of 15 Da, 18 Da, 29 Da, 56 Da, 59 Da, 72 Da, 74 Da, 86 Da, and 118 Da from the ·Bz–C(O)–peptide in the so-called (M· − X) region of Figure 1b . It was previously noted that the side-chain radical losses and neutral losses provided information on the presence of certain amino acid residues, which could improve peptide identification [49, 63]. The characteristic losses frequently found in TEMPO-assisted FRIPS MS are summarized in Supplementary Table S1 (see Supplementary material). These extensive radical losses and neutral losses are also found in the other TEMPO-assisted FRIPS MS spectra shown below. A mechanism for the formation of neutral losses in TEMPO-assisted FRIPS MS is also given in the Supplementary material (Supplementary Scheme S2).
TEMPO-Assisted FRIPS MS of Phosphoserine Peptides
These results indicate that the radical-forming C–O bond cleavage reaction is more energetically favorable than the reactions leading to phosphoric acid loss. Otherwise, the loss of a phosphate group would be the dominant result. Further, the peptide-backbone fragments observed here presumably arose from [rM + H]+·, which had sufficient internal energy to proceed to the consecutive radical-driven peptide-backbone dissociation. A similar result was also observed for the other phosphopeptides.
MS3 mass spectra for the o-TEMPO–Bz–C(O)–conjugated VLTpSSARQR and DTHKpSEIAHR are shown in Figure 2b, c, respectively. For the VLTpSSARQR peptide, a large number of peptide-backbone fragments were observed, with all fragments keeping the phosphate group intact (see Figure 2b). No peptide-backbone fragment was observed to have lost a phosphate group, except for [z7 − H3PO4]+ at m/z 770. The presence of two basic arginine residues inhibited the phosphate group loss reactions that are known to be promoted by a mobile proton. Retaining the phosphate group in the peptide backbone and the observation of [ra4 + 2H]+, [x5–18]+, and z5+ helped us unambiguously assign the phosphorylation site to the fourth serine residue between the two possible phosphorylation serine sites in the sequence, which would be otherwise very difficult. For the DTHKpSEIAHR peptide, all the peptide-backbone fragments were also observed to arise without phosphate (or phosphoric acid) loss, except for [x6 − H3PO4–44]+ at m/z 676, [x6 − H3PO4]+ at m/z 720, [z8 − H3PO4–18]+ at m/z 924, and [ra9 − H3PO4 + H]+ at m/z 1091 (see Figure 2c). In Figure 2b, c, a majority of the fragments were a, c, x, and z types (with some y-type ions), suggesting that radical-driven peptide-backbone dissociation pathways prevailed for the phosphorylated peptides, which was the case in the other TEMPO-assisted FRIPS MS.
For comparison, CAD MS/MS mass spectra were acquired for the singly protonated, unconjugated DTHKpSEIAHR and VLTpSSARQR peptides. Similar to the results shown in Figure 1a, c, the loss of a phosphate group (− 80 Da) or a phosphoric acid (− 98 Da) was a dominant result in both cases (see Supplementary Figure S1).
TEMPO-Assisted FRIPS MS of Phosphotyrosine Peptides
MS3 spectra for o-TEMPO–Bz–C(O)–conjugated RRLIEDNEpYTARG, RHPEpYAVSVLLR, and YLpYEIAR are shown in Figure 3b–d. In these spectra, the loss of a phosphate group was observed, but only to a very limited extent. In particular, in Figure 3b, a series of peptide-backbone fragments, such as [ra8 + H]+·, [ra9 − H]+·, [rc9 + 3H]+·, [ra10 + H]+·, z4+, [y5 + 2H]+, and [z6 + H]+·, which retained a phosphate group during FRIPS MS, unambiguously indicated that the tyrosine in the ninth residue from the N-terminus was phosphorylated. More interestingly, for RHPEpYAVSVLLR, tyrosine and serine in the fifth and eighth positions from the N-terminus, respectively, can potentially be phosphorylated. Conservation of the phosphate group during the peptide-backbone dissociation in the TEMPO-assisted FRIPS MS made it possible to locate the phosphorylation site at the tyrosine in the fifth residue from the N-terminus (see Figure 3c). For the short peptide YLpYEIAR with two potential tyrosine phosphorylation sites, TEMPO-assisted FRIPS also made it possible to locate its phosphorylated site at the third tyrosine.
Competition with Mobile Proton Conditions
The peptides investigated above are the ones that contain at least one arginine (Arg) residue. Arg (R) is known to have the highest gas-phase proton affinity (PA) among natural amino acids . Due to its highest PA, a proton tends to be immobilized at the Arg residue, and thus, the so-called mobile proton condition is difficult to be established when Arg is present in the sequence [65, 66]. However, as CAD of tryptic peptides often occurs under mobile proton conditions, i.e., lysine (Lys)-ending peptides, it is interesting to see how TEMPO-assisted FRIP MS works for a number of Lys-ending peptides without Arg: PNGATHpSPK, VNQIGpTLYASIK, and IHYIpYGSIK. Here, in order to allow a proton to be mobilized along the peptide backbone, Lys (K) was not guanidinated (guanidination of the lysine side chain turns Lys into homoarginine with a high PA value) . With careful pH control (pH ~ 10), o-TEMPO–Bz–C(O) conjugation was found to be directed only into the Lys site.
HCD of o-TEMPO–Bz–C(O) Phosphopeptides
Two CA steps were generally required for TEMPO-assisted FRIPS to produce extensive peptide fragmentations when low collision energies were used in an ion-trap instrument . In contrast, recent studies have reported that FRIPS could proceed in a single step when higher collision energies were adopted in LTQ-Orbitrap and hybrid quadrupole time-of-flight mass spectrometers . However, higher collision energies may lead to the cleavage of the phospho–ether bond, which hinders the localization of a phosphorylation site. Thus, HCD experiments were performed on an LTQ-Orbitrap mass spectrometer for phosphopeptides.
In this study, TEMPO-assisted FRIPS MS for a number of singly or doubly protonated phosphopeptides was shown to extensively produce peptide-backbone fragments with a limited loss of a phosphate group in contrast to the CAD of the phosphorylated peptides, where dominant phosphate loss was observed. The peptide fragments produced were primarily a-, c-, x-, and z-type ions (with minor b- and y-type ions), indicating that radical-driven peptide fragmentation mechanism was mainly responsible for peptide fragmentation. The conservation of a phosphate group in TEMPO-assisted FRIPS MS enabled us to unambiguously localize the phosphorylated sites along the peptide sequence. In a very recent paper published during the revision process, the retainment of a phosphate group was also reported in the negative ion TEMPO-assisted FRIP MS for phosphopeptides . Currently, the separate nano-LC-FRIPS MS studies for α-casein and other phosphorylated protein tryptic digests enriched using a TiO2 solid-phase extraction are in progress. This study demonstrated that FRIPS MS shows promise as an analytical method for phosphoproteome and also for other labile PTM proteomes (e.g., glycoproteome), which are known to be related with many important diseases.
This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2015R1D1A1A01056782 and 2018R1A6A1A03024940).
- 7.Håkansson, K., Chalmers, M.J., Quinn, J.P., McFarland, M.A., Hendrickson, C.L., Marshall, A.G.: Combined electron capture and infrared multiphoton dissociation for multistage MS/MS in a Fourier transform ion cyclotron resonance mass spectrometer. Anal. Chem. 75, 3256–3262 (2003)CrossRefPubMedGoogle Scholar
- 14.Lee, S.Y., Chung, G.S., Kim, J.D., Oh, H.B.: Electron caputure dissociation mass spectrometry (ECD MS) of peptide cations containing a lysine homologue: a mobile proton model for explaining the observation of b-type of product ions. Rapid Commun. Mass Spectrom. 20, 3167–3175 (2006)CrossRefPubMedGoogle Scholar
- 20.Oh, H.B., Leach, F., Arungundram, S., Kanar, A.-M., Venot, A., Boons, G.-J., Amster, J.I.: Multivariate analysis of electron detachment dissociation (EDD) and infrared multiphoton dissociation mass spectra of heparin sulfate (HS) tetrasaccharides differing only in hexuronic acid stereochemistry. J. Am. Soc. Mass Spectrom. 22, 582–590 (2011)CrossRefPubMedPubMedCentralGoogle Scholar
- 28.Chu, I.K., Zhao, J., Xu, M., Siu, S.O., Hopkinson, A.C., Siu, K.W.M.: Are the radical centers in peptide radical cations mobile? The generation, tautomerism, and dissociation of isomeric alpha-carbon-centered triglycine radical cations in the gas phase. J. Am. Chem. Soc. 130, 7862–7872 (2008)CrossRefPubMedGoogle Scholar
- 48.Nam, J.J., Kwon, H.S., Jang, I.A., Jeon, A.R., Moon, J.K., Lee, S.Y., Kang, D.J., Han, S.Y., Moon, B.J., Oh, H.B.: Bromine isotopic signature facilitates de novo sequencing of peptides in free radical initiated peptide seqeuncing (FRIPS) mass spectrometry. J. Mass Spectrom. 50, 378–387 (2015)CrossRefPubMedGoogle Scholar
- 49.Lee, C.S., Jang, I.A., Hwangbo, S., Moon, B.J., Oh, H.B.: Side chain cleavage in TEMPO-assisted free radical initiated peptide sequencing (FRIPS): amino acid composition information. Bull. Kor. Chem. Soc. 36, 810–814 (2015)Google Scholar
- 50.Jang, I.A., Lee, S.Y., Hwangbo, S., Lee, H.K., Kim, H.I., Kang, D.J., Moon, B.J., Oh, H.B.: TEMPO-assisted free radical initiated peptide sequencing mass spectrometry (FRIPS MS) in Q-TOF and orbitrap mass spectrometers: single-step peptide backbone dissociations. J. Am. Soc. Mass Spectrom. 28, 154–163 (2017)CrossRefPubMedGoogle Scholar
- 55.Iacobucci, C., Schäfer, M., Sinz, A.: Free radical-initiated peptide sequencing (FRIPS)-based cross-linkers for improved peptide and protein structure analysis. Mass Spectrom. Rev. (2018). https://doi.org/10.1002/mas.21568