Advertisement

Characterization of Disulfide-Linked Peptides Using Tandem Mass Spectrometry Coupled with Automated Data Analysis Software

  • Zhidan Liang
  • Kenneth N. McGuinness
  • Alejandro Crespo
  • Wendy ZhongEmail author
Focus: 29th Sanibel Conference, Peptidomics: Bridging the Gap Between Proteomics and Metabolomics by MS: Research Article

Abstract

Disulfide bond formation is critical for maintaining structure stability and function of many peptides and proteins. Mass spectrometry has become an important tool for the elucidation of molecular connectivity. However, the interpretation of the tandem mass spectral data of disulfide-linked peptides has been a major challenge due to the lack of appropriate tools. Developing proper data analysis software is essential to quickly characterize disulfide-linked peptides. A thorough and in-depth understanding of how disulfide-linked peptides fragment in mass spectrometer is a key in developing software to interpret the tandem mass spectra of these peptides. Two model peptides with inter- and intra-chain disulfide linkages were used to study fragmentation behavior in both collisional-activated dissociation (CAD) and electron-based dissociation (ExD) experiments. Fragments generated from CAD and ExD can be categorized into three major types, which result from different S–S and C–S bond cleavage patterns. DiSulFinder is a computer algorithm that was newly developed based on the fragmentation observed in these peptides. The software is vendor neutral and capable of quickly and accurately identifying a variety of fragments generated from disulfide-linked peptides. DiSulFinder identifies peptide backbone fragments with S–S and C–S bond cleavages and, more importantly, can also identify fragments with the S–S bond still intact to aid disulfide linkage determination. With the assistance of this software, more comprehensive disulfide connectivity characterization can be achieved.

Graphical Abstract

Keywords

DisulFinder software tool Disulfide containing peptide EID Mass spectrometry 

Introduction

Disulfide linkage between two cysteine residues is a widely occurring protein post-translational modification (PTM) critical for maintaining structural stability and protein function. Understanding disulfide linkages between cysteine residues in proteins can facilitate structure characterization and also aid in verifying disulfide linkage designs. Over the past two decades, analytical approaches using mass spectrometry (MS) have become central to the characterization of proteins and their PTMs, including disulfide bond assignment [1]. The deduction of protein sequence information is achieved by tandem mass spectrometry (MS/MS) [2], where a peptide ion is fragmented using different dissociation techniques followed by subsequent m/z measurement of fragment ions. The most widely used activation method is collision-activated dissociation (CAD), also known as collision-induced dissociation (CID). During the CAD process, peptide ions of interest are isolated in the gas phase and subjected to collisions with noble gas atoms such as argon or nitrogen to induce breakage of amide bonds in the peptide backbone [2]. However, it has been well documented that PTMs tend to be unstable and are often lost during the CAD process [3]. Electron-based dissociation (ExD) [4], where precursor ions react with electrons carrying different electron energies to induce fragmentation, has been developed as an alternative activation technique. Across the range of electron energy, ExD methods include electron capture dissociation (ECD) [5, 6] (energy electrons ~1 eV), hot ECD (HECD) [7, 8] (electron energies to 8–14 eV), and electron-induced dissociation (EID) ( electron energies >16 eV) [9, 10]. ECD has been shown to induce random peptide backbone cleavages while preserving labile modifications [11]. The application of ECD and HECD has been limited to multiply charged ions, but EID has been successfully applied to study singly charged drug metabolites [9] and peptides [10] since the first study of excitation of singly charged organic cations at high electron energies by Cody and Freiser [12], which was also called EIEIO (electron impact excitation of ions from organics). Alternatively, Liu et al. [13] have reported a disulfide mapping workflow using EThcD (electron transfer higher energy dissociation), where electron transfer dissociation cleaves disulfide bond first followed by subsequent higher energy dissociation of disulfide cleaved peptide precursors. The resulting MS/MS data was then searched against linear peptides database.

Sample preparation approaches also play an important role in analytical strategy establishment. Bottom-up approaches are often used for protein structure elucidation and PTM analysis by reducing the disulfide bond and performing enzymatic digestion prior to MS analysis [14]. Higher fragmentation efficiency can often be achieved with the disulfide bond reduced; however, information about disulfide linkage is obviously lost. Additionally, during the enzymatic digestion process, disulfide bond scrambling can also occur, thus giving misleading results [15]. In contrast, top-down MS has unique strength in the comprehensive analysis of proteins [16] and PTMs by preserving labile modifications [17]. Top-down MS approaches have been successfully applied to the identification of disulfide linkages in peptides, thereby avoiding possible disulfide rearrangement while providing more extensive molecular connectivity information [18].

Despite the advances in fragmentation techniques, the lack of appropriate software tools for disulfide-linked peptide MS data analysis constitutes an analytical bottleneck. There has been great effort in developing proper computer algorithm to aid MS/MS data analysis for disulfide-linked peptide identification, such as MassMatrix search engine [19] and DBond algorithm [20]. However, they are still bottom-up approach- and database search-based, and cannot be readily used for top-down work flow. Up to this point, most of the top-down MS/MS data generated for disulfide-linked peptides have been manually interpreted to avoid missing identification information. Tandem mass spectra of disulfide-linked peptides are typically very complicated. Fragments are generated from multiple peptide bond cleavages [21], including cysteine thioaldehyde, cysteine persulfide, or dehydroalanine, attributable to the breakage of S–S or C–S bonds, respectively, in addition to peptide backbone cleavage. Manually examining MS/MS spectra for all possible peptide fragmentation combinations can be very time-consuming and unreliable, especially if the target protein is cysteine-rich and contains unknown disulfide bonds. In addition, currently available software programs are generally incapable of providing detailed interpretation of MS/MS data generated for disulfide-linked peptides. The typical approach for these software programs is to assume that disulfide bonds are reduced or the additional chain(s) are considered as modifications of the main chain. Consequently, peptide backbone fragments with disulfide bonds still intact are often missed using this approach. As a result, determination of disulfide linkage is often very challenging.

In the present study, the fragmentation behavior of two model peptides, one with an inter-chain disulfide linkage and the other with an intra-chain disulfide linkage, were thoroughly studied using both CAD and ExD activation techniques, which provided more comprehensive peptide fragmentation. A software tool, DiSulFinder, was then designed and implemented to assist MS/MS data analysis of disulfide-linked peptides. In conjunction with currently available data processing platforms, DiSulFinder is capable of identifying peptide backbone fragments with and without S–S or C–S bond cleavages. This tool quickly identifies fragments retaining S–S bond connectivity as well as those from S–S or C–S bond cleavages, providing critical information for disulfide linkage determination, which cannot be achieved via existing commercial software packages. In addition, the program is a vendor-neutral platform that can handle CAD or ExD data from any MS system. DiSulFinder can potentially be used to identify unknown disulfide linkages in cysteine-rich proteins, which could be useful for better determination of their folding patterns.

Experimental

Materials

Inter-chain disulfide-linked peptide, insulin_frag (A chain: QLENYVCN-NH2; B chain: LVCGER-NH2; Cys connected by S–S) was synthesized by AnaSpec (AnaSpec, Fremont, CA, USA). Intra-chain disulfide-linked peptide, crustacean cardioactive peptide (CCAP, PFCNAFTGC-NH2, disulfide bond linked between Cys 3 and 9) was purchased from Bachem (Bachem Americas, Inc., Torrance, CA, USA). All solvents were Optima LCMS grade from Fisher Chemical.

MS/MS Analysis

Each sample was directly infused into the mass spectrometer by a TriVersa NanoMate robot (Advion, Inc., Ithaca, NY, USA) at a concentration of 5 pmol/μL in a spray solution of 50:50 acetonitrile:water with 0.1% formic acid.

CAD and ExD experiments were performed on a 9.4T Solarix qQq-Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer (Bruker Daltonics, Billerica, MA, USA). Mass spectra were collected in the positive ion mode, with 2 M data points. Collision energy used in the CAD experiment was 25 V with argon as collision gas and MS/MS spectra were summed over 50 scans. ECD MS/MS spectra were summed over 200 scans with electron energies that ranged from 1.7 to 21 eV. The transient length was 1.12 s, and the estimated resolving power was ~200,000 at m/z 400. The signal-to-noise ratio (S/N) threshold was set to 3; signals below that threshold were ignored. Data were analyzed using DataAnalysis 4.4 and BioTools 3.2 SR5 (Bruker Daltonics, Billerica, MA, USA) with a mass accuracy of <5 parts per million (ppm).

DiSulFinder Algorithm and Data Analysis Workflow

Mass spectral data were processed using DataAnalysis 4.4 software and peaks with signal-to-noise ratio >3 were exported as a mass list and used as input for fragment matching using our own program, DiSulFinder. Our code was written in Python 2.7.12 using the periodic table 1.5.0 package [22] and a modified molmass.py [23] program. Peptide sequences (Figure 1a, b) were treated as linear peptides for theoretical fragmentation and the resulting fragments were labeled Ai, Bi, Ci, corresponding to a, b, c peptide fragment ions, respectively; Xi, Yi, Zi, representing x, y, z peptide fragment ions; where i is the position along the N or C- terminus (Figure 1c, d). Each sequence was split by indices into five sections: section 1 goes from N-terminus to the amino acid before the S–S bond; section 2 continues from the first cysteine involved in the S–S bond to either the residue in the middle of the connected cysteine residues (if disulfide is intra-molecular) or to the C-terminus of the first chain; section 3 goes from the residue directly after section 2, to the residue before the second cysteine involved in the S–S bond; section 4 goes from the second cysteine in the S–S bond to the C-terminus of the linear sequence; and section 5 are the residues in-between each cysteine. Figure 1e, f show the graphical representation of the five sections for the two selected peptides in this work. The five sections of peptide sequences were then combined or split further (i.e., section 5) to form all possible fragments that retained and/or broke the disulfide bond. Nomenclature for multiple backbone cleavages was assigned corresponding labels and indices. Fragments involving the breaking of C–S bonds include a ‘–S’ or ‘+S’ to denote the removal or addition of a sulfur atom, respectively. Precursor ions were denoted M and in the case of peptides with multiple chains, each chain was numbered accordingly (i.e., 1, 2). Finally, a mass error of 5 ppm was used for approximating match identity.
Figure 1

Disulfide-linked peptides and the example of four sections used by DiSulFinder. (a), (b) Peptide sequences, where Infc1 and Infc2 are chain 1 and 2, respectively. (c), (d) Linear arrangement and fragmentation nomenclature of sequences. (e), (f) Five sequence domains used in fragment generation

Results and Discussion

Two model peptides, one insulin peptide (insulin_frag) with an inter-chain disulfide linkage and one with intra-chain disulfide-linked peptide (CCAP) were used for fragmentation studies. CAD and ExD were both performed to obtain more comprehensive peptide fragmentation to facilitate the characterization of disulfide-linked peptides. BioTools was employed for initial CAD and ExD MS/MS data processing. BioTools can only identify fragments assuming the S–S bond was already reduced, or alternatively the program could treat one chain as a fixed modification to another. Manual data analysis was also performed to identify additional fragments, which were not identified via BioTools. The results were compared with the new software we developed to evaluate and validate the utility of the new software tool.

Inter-Chain Disulfide-Linked Peptide

Figure 2 shows the CAD MS/MS spectrum of the insulin_frag peptide. Analysis of the spectrum was quite complicated, partially due to the cleavage of C–S or S–S bonds, in addition to peptide backbone cleavages. The MS/MS data were first processed via BioTools followed by manual data interpretation. The resulting fragments were categorized into three main types as summarized in Table 1, including peptide backbone fragments with Cys already cleaved, Cys-fragments bearing additional sulfur or losing sulfur due to C–S bond cleavage, and those with the S–S bond still intact. As shown in Table 1, BioTools was able to identify peptide backbone fragments with Cys already cleaved, such as y2 and y3 of chain 1, b3-5 of chain 2, as well as some peptide backbone fragments with the S–S bond intact, such as y4 of chain 1 connected to chain 2 (where chain 2 is considered as a modification of chain 1). However, BioTools lacked the ability to identify any C–S bond cleavage fragments expected to result from the CAD fragmentation process. Thus, we manually examined the MS/MS spectrum, specifically looking for this type of fragments. The results confirmed that the cleavage of C–S bond in the CAD experiment was present and we identified a number of fragments resulting from C–S bond cleavage in addition to peptide backbone fragmentation, such as x6–S and y4+S of chain 1 as shown in Table 1. During this manual data analysis process, we also identified 15 additional fragments not found by BioTools where the S–S bond remained intact. The majority of those S–S connected fragments were generated via peptide backbone bond cleavages on both chains, which could provide peptide sequence information. Such fragments could potentially be valuable for future characterization of unknown disulfide linkages. As summarized in the Table 1, S–S intact fragments such b4 ions of chain 1 may connect to y4 ions of chain 2 via an S–S bond, which may also provide information not only on individual peptide chain amino acid sequence but also on the determination of S–S linkage sites. The identification of S–S intact fragments is critical for disulfide-linkage site determination. Therefore, the development of proper software tools to automatically identify those fragments is needed.
Figure 2

CAD MS/MS spectrum of insulin_frag peptide acquired at collision energy of 25 V and summed over 50 scans; possible cleavage sites of disulfide-linked Cys residues in MS/MS, between S–S bond or C–S bond are shown in the inset

Table 1

Summary of Different Fragment Types of insulin_frag Peptide in CAD Experiment Obtained by Combining the Results from Both BioTools and Manual Data Interpretation

Fragment types

Ions

Sequences

m/z

Peptide backbone cleavage fragments

y2 (chain 1)

ER-NH2

303.17709

y3 (chain 1)

GER-NH2

360.19917

b3 (chain 2)

QLE

371.19171

b4 (chain 2)

QLEN

485.23632

b5 (chain 2)

QLENY

648.29984

C–S cleavage fragments

M+S (chain 1)

LVC(+S)GER-NH2

707.33413

y4+S (chain 1)

C(+S)GER-NH2

495.18131

M (chain 1)-S

LVC(-S)GER-NH2

641.37423

M (chain 2)+S*

QLENYC(+S)N-NH2

914.35127

x6-S (chain 2)

LENYC(-S)N-NH2

748.36098

S–S bond intact fragments

Chain 2 [z2]- Chain 1 [y4]

CN-NH2/CGER-NH2

678.24622

Chain 2[y2]- Chain 1 [y5]

CN-NH2/VCGER-NH2

794.34140

Chain 2 [y3]- Chain 1 [y4]

YCN-NH2/CGER-NH2

858.33570

Chain 2 [y3]- Chain 1 [z4]

YCN-NH2/CGER-NH2

841.31080

Chain 2 [y4]-Chain 1 [b4]

NYCN-NH2/LVCG

880.33502

Chain 2 [z2]- Chain 1

CN-NH2/LVCGER-NH2

890.39920

Chain 2 [z4]- Chain 1 [y4]

NYCN-NH2/CGER-NH2

955.35361

Chain 2 [y4]- Chain 1 [y4]

NYCN-NH2/CGER-NH2

972.37955

Chain 2 [z4]-Chain 1 [b5]

NYCN-NH2/LVCGE

994.37695

Chain 2 [z3]- Chain 1

YCN-NH2/LVCGER-NH2

1053.46297

Chain 2 [y5]-Chain 1 [z4]

ENYCN-NH2/CGER-NH2

1084.39575

Chain 2 [z4]- Chain 1

NYCN-NH2/LVCGER-NH2

1167.50779

Chain 2 [z5]- Chain 1

ENYCN-NH2/LVCGER-NH2

1296.54947

Chain 2 [y6]- Chain 1 [y5]

LENYCN-NH2/VCGER-NH2

1313.57698

Chain 2- Chain 1 [y4]

QLENYCN-NH2/CGER-NH2

1325.53789

a+S, gaining an extra sulfur during C–S fragmentation of the corresponding peptide chain.

b–S, losing sulfur during C–S fragmentation.

c* Represents radical cation.

d‘/’ Between residues represents the two Cys residues on both chains are linked via S–S bond.

To develop a more comprehensive understanding of disulfide-linked peptide fragmentation, the ECD experiment was also performed on the insulin_frag peptide. At 1.7 eV of electron energy, the fragmentation efficiency was rather low. HECD experiment of the same peptide with electron energy of 16 eV generated much more fragmentation, especially in the lower mass range as shown in Figure 3. The fragmentation patterns were similar to those obtained from CAD, and the resulting fragments could be grouped into the same three categorizes. In addition to b/y type ions, S–S and C–S bond cleavage together with extensive peptide backbone fragmentation also resulted in the generation of various c/z ions, giving greater peptide sequence coverage than was obtained in CAD. A large number of fragments with the S–S bond intact were also identified, including those with one or more peptide backbone cleavages. HECD generated many more fragments than were observed with ECD, mainly attributable to excess energy deposited into molecules when higher energy electrons collided with the molecules. Extensive fragmentation obtained from the HECD experiment requires more efficient data processing as it is impractical to manually identify all the possible fragments.
Figure 3

HECD MS/MS spectrum of insulin_frag peptide acquired at electron energy of 16 eV, and irradiation time of 500 ms and summed over 200 scans

The in-depth knowledge and understanding of disulfide fragmentation under CAD and ECD conditions garnered from these processes was then incorporated into the development of software that will facilitate the identification of these different types of fragments. A software tool, DiSulFinder, was specifically developed to interpret CAD and ECD MS/MS data for the identification of S–S and/or C–S bond cleavage fragments and those in which the S–S bond remained intact. The identification of CAD and HECD fragments using DiSulFinder is summarized in Table 2 (only fragments with intensity higher than 1+E06 and up to 2 possible structures of the same peak are reported in Table 2; complete list summarized in Supplementary Tables S1 and S2, respectively. As seen in Table 2, the software outputs the ion type using the nomenclature defined in this software development as well as corresponding sequence information. Figure 4 depicts the nomenclature that DiSulFinder adapted to illustrate the four types of fragments identified by the software. Figure 4a represents a single bond cleavage fragment, Z2, from peptide backbone cleavage of chain 2 with two C-terminal residues remaining. Figure 4b shows fragment B5Z4 formed by cleavage of one peptide bond on each of the two chains (one C-terminal residue of chain 1 was cleaved and three N-terminal residues were cleaved from chain 2). Figure 4c shows a fragment ion, X11B3C11, formed from three bond cleavages, at different locations: two peptide backbone cleavages occurred along chain 1 (amino acid residues L of N-terminus and GER of C-terminus were lost); the remaining fragment was connected to chain 2 via a disulfide linkage (C-terminus N residue was cleaved off of chain 2). Figure 4d illustrates a four bond cleavage fragment, A5X10Y6A11. Two peptide backbone cleavages occurred on each chain: L, and V, from the N-terminus and R of C-terminus of chain 1 were cleaved; Q from the N-terminus and N from C-terminus of chain 2 were also cleaved.
Table 2

Summary of insulin_frag CAD Fragments Identified Using DiSulFinder

Ion

Sequence

m/z

Charge

ppm

Intensity

[1]/Y2

LVCGER-NH2/CN-NH2

907.4258

1

2.39

2.56E+08

[1]/Z2

LVCGER-NH2/CN-NH2

890.3992

1

2.34

1.38E+08

[1]+S

LVCGER-NH2

707.3332

1

0.6

1.25E+08

Y9

GER-NH2

360.1992

1

0.49

1.06E+08

[1]/Y4

LVCGER-NH2/NYCN-NH2

1184.533

1

2.44

8.34E+07

Y4C11

LVCGER-NH2/NYC

1070.491

1

3.39

7.26E+07

[1]/Y3

LVCGER-NH2/YCN-NH2

1070.491

1

3.39

7.26E+07

[1]/Y5

LVCGER-NH2/ENYCN-NH2

1313.577

1

3.32

7.11E+07

Y11Y6

VCGER-NH2/LENYCN-NH2

1313.577

1

3.32

7.11E+07

Z10/[2]

CGER-NH2/QLENYCN-NH2

1325.538

1

1.36

4.42E+07

Y10+S

CGER-NH2

495.1813

1

2.14

3.94E+07

[1]/Z3

LVCGER-NH2/YCN-NH2

1053.463

1

2.37

3.31E+07

Y4B11

LVCGER-NH2/NYC

1053.463

1

2.37

3.31E+07

[1]-S*

LVCGER-NH2

641.3742

1

2.02

2.64E+07

[1]/Z4

LVCGER-NH2/NYCN-NH2

1167.507

1

3.33

2.61E+07

C11Z2

LVCGER-NH2/C

776.3565

1

2.91

1.81E+07

B11Y2

LVCGER-NH2/C

776.3565

1

2.91

1.81E+07

[1]/Z5

LVCGER-NH2/ENYCN-NH2

1296.549

1

2.51

1.41E+07

Z11Y6

VCGER-NH2/LENYCN-NH2

1296.549

1

2.51

1.41E+07

Z10Y2

CGER-NH2/CN-NH2

678.2462

1

2.33

1.40E+07

Y10Z2

CGER-NH2/CN-NH2

678.2462

1

2.33

1.40E+07

Z11Y2

VCGER-NH2/CN-NH2

777.3155

1

3.08

1.36E+07

Y11Z2

VCGER-NH2/CN-NH2

777.3155

1

3.08

1.36E+07

Z4B11

LVCGER-NH2/NYC

1036.437

1

3.03

7.99E+06

B8

QLE

371.1917

1

2.16

7.53E+06

Z3C11

LVCGER-NH2/YC

939.4207

1

3.39

7.50E+06

Y3B11

LVCGER-NH2/YC

939.4207

1

3.39

7.50E+06

Y10Y4

CGER-NH2/NYCN-NH2

972.3796

1

2.16

7.30E+06

C9Z5

EN

244.0924

1

1.54

7.23E+06

B9Y5

EN

244.0924

1

1.54

7.23E+06

Z10Y4C11

CGER-NH2/NYC

841.3108

1

3.37

7.15E+06

Z10Y3

CGER-NH2/YCN-NH2

841.3108

1

3.37

7.15E+06

Y10Y4C11

CGER-NH2/NYC

858.3357

1

1.38

7.14E+06

Y10Y3

CGER-NH2/YCN-NH2

858.3357

1

1.38

7.14E+06

Z10Y4

CGER-NH2/NYCN-NH2

955.3534

1

2.58

6.61E+06

Y10Z4

CGER-NH2/NYCN-NH2

955.3534

1

2.58

6.61E+06

M

LVCGER-NH2/QLENYCN-NH2

1554.72

1

3.29

6.54E+06

Y10Y5

CGER-NH2/ENYCN-NH2

1101.423

1

2.37

4.43E+06

Y10/[2]

CGER-NH2/QLENYCN-NH2

1342.567

1

2.98

4.05E+06

A4X10A11

CG/QLENYC

879.3145

1

2.39

3.82E+06

X11X6

VCGER-NH2/LENYCN-NH2

683.2684

2

1.13

3.40E+06

Y11Y2

VCGER-NH2/CN-NH2

794.3414

1

2.26

3.39E+06

B9

QLEN

485.2363

1

1.82

3.22E+06

Z11/2

VCGER-NH2/QLENYCN-NH2

1424.608

1

2.46

2.90E+06

Z10Z4

CGER-NH2/NYCN-NH2

938.3268

1

2.64

2.88E+06

Y9

ER-NH2

303.1771

1

1.45

2.85E+06

B7

QL

242.1495

1

1.64

2.47E+06

Z10+S

CGER-NH2

478.1546

1

1.78

2.43E+06

Y11C3X2

VC/CN-NH2

478.1546

1

1.78

2.43E+06

A10Z5

ENY

362.1353

1

1.65

2.41E+06

A11Z2

LVCGER-NH2/C

731.3353

1

3.51

2.34E+06

Z10Z2

CGER-NH2/CN-NH2

661.2196

1

2.33

2.31E+06

[1]/Y6

LVCGER-NH2/LENYCN-NH2

1426.66

1

2.25

2.25E+06

Y6B11

LVCGER-NH2/LENYC

648.2998

2

4.09

2.05E+06

B10

QLENY

648.2998

1

1.66

2.05E+06

Z4A11

LVCGER-NH2/NYC

1008.441

1

2.06

1.97E+06

C10Z5

ENY

407.1569

1

1.88

1.97E+06

B10Y5

ENY

407.1569

1

1.88

1.97E+06

Z11Z6

VCGER-NH2/LENYCN-NH2

1279.522

1

2.21

1.84E+06

Z10Z3

CGER-NH2/YCN-NH2

824.2837

1

2.74

1.78E+06

Z10Y4B11

CGER-NH2/NYC

824.2837

1

2.74

1.78E+06

Z3B11

LVCGER-NH2/YC

922.3939

1

3.2

1.68E+06

Z10Y5

CGER-NH2/ENYCN-NH2

1084.396

1

2.09

1.60E+06

Y10Z5

CGER-NH2/ENYCN-NH2

1084.396

1

2.09

1.60E+06

Z11C4Y6B11

VCG/LENYC

880.335

1

2.54

1.55E+06

B5Z3

LVCGE/YCN-NH2

880.335

1

2.54

1.55E+06

Z10B11

CGER-NH2/QLENYC

1194.467

1

0.64

1.41E+06

B3B11*

LVC/QLENYC

1063.449

1

2.47

1.34E+06

Z10C11

CGER-NH2/QLENYC

1211.496

1

2.67

1.33E+06

Y10B11

CGER-NH2/QLENYC

1211.496

1

2.67

1.33E+06

[1]/B11

LVCGER-NH2/QLENYC

1423.65

1

3.16

1.31E+06

A5Y10X6A11

CGE/LENYC

880.2974

1

1.1

1.26E+06

A5X10Y6A11

CGE/LENYC

880.2974

1

1.1

1.26E+06

A11Y2

LVCGER-NH2/C

748.361

1

2.28

1.20E+06

A10Y6

LENY

492.2463

1

2.08

1.19E+06

Y11-S*

VCGER-NH2

528.2897

1

1.53

1.17E+06

Z11Y4C11

VCGER-NH2/NYC

940.3788

1

2.53

1.14E+06

Z11Y3

VCGER-NH2/YCN-NH2

940.3788

1

2.53

1.14E+06

[2]+S

QLENYCN-NH2

914.3513

1

1.94

1.13E+06

X11C11

VCGER-NH2/QLENYC

677.286

2

1.97

1.10E+06

B3Z2

LVC/CN-NH2

531.2067

1

2.39

1.01E+06

a/’ Between residues represents the two Cys residues on both chains are linked via S–S bond.

b* Represents radical cation.

c[1] and [2] in ‘Ion’ column represents intact chain 1 and chain 2, respectively.

Figure 4

Illustration of different DiSulFinder output ions, greyed residues being cleaved off in fragments. (a) Single bond cleavage fragment, Z2 was from one peptide backbone cleavage of chain (b) Two bonds cleavage fragments, such as B5Z4 was formed from one peptide backbone cleavage on each chain, one C-terminal residue of chain 1 was cleaved off and three N-terminal residues were cleaved from chain 2. (c) Three bonds cleavage fragments, for example X11B3C11 formed from two peptide backbone cleavages of chain 1, and one peptide backbone cleavage of chain 2: both L of N terminal and GER of C-terminal were cleaved off of chain 1 and N of C-terminal was cleaved off of chain 2. (d) Four bonds cleavage fragments such as A5X10Y6A11 are formed by two peptide backbone cleavages on each chain: L, V of N-terminal and R of C-terminal of chain 1 were cleaved; Q of N-terminal and N of C-terminal of chain 2 were both cleaved

The identification results generated by DiSulFinder were compared with those from BioTools and manual data interpretation for evaluation purposes. DiSulFinder readily identified fragments found by both BioTools and manual data processing. For instance, the m/z 495.18131ion was identified as y4+S of chain 1 manually and DiSulFinder identified the same fragment, and output as Y10+S. Both were correlated to the same amino acid sequence of CGER-NH2 of chain 1, with Cys- bearing additional sulfur from C–S bond cleaved off of chain 2. For the m/z 994.37695 fragment ion, we manually assigned it as chain 2 [z4]-chain 1 [b5], NYCN-NH2/LVCGE as in Table 1. DiSulFinder also identified the same peak, but assigned two possible structures that had the same chemical formula, where the same amino acid composition was named as B5Z4 and a second possible structure was X11B3C11 for VC/QLENYC. Most importantly, DiSulFinder identified many fragments that resulted from four peptide backbone cleavages. These fragments were very challenging to deduce manually, mainly because there was no prior knowledge about their presence. One such example is fragment A5X10Y6A11 (m/z 880.2974), which correlates to CGE/LENYC with two Cys connected via S–S bond, and both chains were cleaved at both terminus.

In summary, the number of fragments identified using DiSulFinder was substantially greater than those identified by BioTools and manual data processing combined. From CAD fragmentation, DiSulFinder identified more than 50 additional fragments that were not identified by BioTools, with the majority of those fragments containing an intact S–S bond and peptide backbone fragmentation from both chains. DiSulFinder successfully identified fragments found and characterized by BioTools as well as the manually assigned fragments, which gave us more confidence about the identification by DiSulFinder.

Intra-Chain Disulfide-Linked Peptide, Crustacean Cardio-Active Peptide (CCAP)

Crustacean cardio-active peptide (CCAP) was used to investigate the fragmentation behavior of disulfide-linked cyclic peptides under different dissociation techniques. CAD often struggles to generate sufficient fragmentation coverage for cyclic peptides since they usually carry a single charge and dissociate poorly, as is the case for CCAP. Here, from CAD MS/MS data, BioTools was only able to identify a few backbone fragments from amino acid residue outside the two disulfide-linked Cys, such as y8 and y7. From manual data processing, as summarized in Table 3, we identified several Cys-containing fragments arising from cleavage of the disulfide bond and an additional peptide bond, such as y6 and b5 ions, both of which contained one Cys residue. Analogous to an inter-chain linked peptide, CAD fragmentation of this cyclic disulfide-linked peptide also led to cleavage of the C–S bond of the Cys residue. We observed b4, b6, and b8 ions with the sulfur lost because of the cleavage of C–S bond as shown in Table 3. Those fragments, however, could not be used to determine the disulfide linkage site.
Table 3

Summary of CCAP Fragments Identified in CAD Experiment

Fragment types

Ions

Sequences

m/z

S–S cleavage fragments

y6

NAFTGC-NH2

611.26208

b6

PFCNAF

678.20291

C–S cleavage fragments

b4-S*

PFC(-S)N

428.19360

b6-S*

PFC(-S)NAF

646.29990

b8-S*

PFC(-S)NAFTG

804.37001

 S–S bond intact fragments

 

PFCNAc|yC

667.23441

y8

FCNAFTGC-NH2

859.32479

y7

CNAFTGC-NH2

712.25580

M-NH2

PFCNAFTGC

939.35141

S–S bond is linked in fragments contain two Cys residues; ‘|’ between residues represents the peptide backbone cleavage site

* indicates the fragment to be radical cation

In an effort to improve fragmentation efficiency of the cyclic disulfide-linked peptide, and obtain more structurally useful information, electron-induced dissociation (EID) was employed. Here at 21 eV, EID fragmentation of the singly charged precursor of CCAP provided complementary fragments to CAD MS/MS. With EID, we were able to obtain S–S or C–S bond-cleaved peptide fragments similar to those seen in CAD MS/MS in Table 3. Subsequently, we found that EID was able to preserve the S–S bond and break peptide bonds between the two linked Cys residues. Such fragments, which involved two or more peptide bond cleavages, were manually identified and are summarized in Table 4. For instance, Asn was cleaved between adjacent Cys- and -Ala residues, and the resulting fragment was represented as PFCz/cAFTGC, where subscript c and subscript z represent the cleavage site on the peptide backbone. Similarly, three connected amino acid residues, Phe-Thr-Gly between Ala and the C-terminal Cys were also cleaved while the two Cys remained linked via an S–S bond. This type of fragment can provide valuable peptide sequence information, and confirm the location of disulfide linkage simultaneously.
Table 4

Summary of CCAP Fragments Between Disulfide-Linked Cys-Residues in EID

Sequence

m/z

z[NAF]b

316.12909

z[C(+S)NA]c

321.06850

z[NAF]c

333.15569

z[C(-S)NAF]b*

385.15064

y[FC(-S)NA]b*

402.17723

z[FC(+S)NA]a

423.11554

z[NAFTG]b

474.19840

y[CNAFTG]c

611.26082

PFCNAb|*ZC-NH2

633.17929

yFCNAc|yTGC-NH2

686.20468

y[FC(+S)NAFT]c

729.24853

PFCz|cAFTGC-NH2

842.33220

S–S bond is linked in fragments contain two Cys residues; ‘|’ between residues represents the peptide backbone cleavage site

* indicates the fragment to be radical cation

Lower case a, b, c, and z, y represent the bond cleavage sites between amino acid residues; cleavage between Cα-C, and amino acid residue with Cα will be represented as [AA]a; cleavage between amide bond, and resulting fragments bearing C=O group will be [AA]b; fragments with NH will be [AA]y; cleavage between N-Cα and resulting fragments has Cα will be [AA]z; fragments with NH will be [AA]c

We were able to identify eight fragments in total from CAD fragmentation, three of which had the two Cys residues linked via the disulfide bond (Table 3). No fragments between the linked Cys residues were manually identified in CAD MS/MS. In contrast, DiSulFinder was able to identify a greater number of fragments from the CAD data as summarized in Supplementary Table S3. Those results also confirmed the presence of fragments with amino acid cleavages between two linked Cys residues in the CAD data. Examples include the cleavage of Thr in B6Z2, PFCNAF/GC-NH2, and cleavage of Phe in B5Z3, PFCNA/TGC-NH2. There is a risk of missing important information during manual data analysis, especially if there is no prior knowledge to prompt an investigator what to look for or if the peak intensity is relatively small. At the same time, this type of fragment with the S–S bond intact can be useful for the deduction of disulfide linkages in cyclic peptides. The prevalence of radical cations complicating the interpretation of ExD MS/MS spectra provides another example of the utility of DiSulFinder. Manual analysis time for one EID MS/MS spectrum can easily consume tens of hours depending on how extensive of an analysis is needed. In contrast, with DiSulFinder, we successfully identified all of the different types of fragments observed, including radical cations, within a few minutes. Fragments found by DiSulFinder from the EID MS/MS data are summarized in Supplementary Table S4. We can more efficiently and accurately study disulfide-linked peptides with DiSulFinder. The identification of S–S linkage retained fragments in intra-chain disulfide-linked peptide could provide critical information for disulfide linkage site determination.

Conclusions

Structural characterization of disulfide-linked peptides has been challenging because of insufficient fragmentation and lack of proper data interpretation tools. In this work, we demonstrated a new workflow that can quickly and accurately characterize these types of peptides via CAD/ExD experiments coupled with DiSulFinder, a newly developed software tool. We used inter- and intra-chain disulfide-linked peptides as model compounds to study the fragmentation behavior of disulfide-linked peptides under different dissociation techniques. Electron-based dissociation approaches were used to provide complementary fragmentation to CAD, both of which generated three main types of fragment ions. In particular, EID, which utilizes high electron energy, generated unique peptide backbone cleavages between two linked Cys while keeping the S–S bond intact. These fragments provided important peptide sequence information as well as the disulfide linkage. DiSulFinder was developed based on the observed fragmentation patterns to quickly and accurately identify disulfide peptide fragments without the need of disulfide bond reduction prior to CAD and ExD MS/MS analysis. The unique advantage of this tool lies in its ability to identify fragments not only from peptide backbone and S–S or C–S bond cleavage, but also S–S intact fragments, thereby providing critical information on disulfide linkage sites. As a result, the program could expedite the identification of complicated tandem mass spectra of disulfide peptides and can also be applied to the analysis of peptides/proteins with inter- and/or intra-chain disulfide linkages. It will be interesting to explore the potential of the software for better characterization of unknown disulfide linkages within cysteine rich protein molecules.

Notes

Acknowledgments

The authors thank Drs. Gary Martin and R. Thomas Williamson for their fruitful discussion and critical reading of the manuscript. The authors also gratefully acknowledge the support from MRL Postdoc Research Fellow Program.

Supplementary material

13361_2017_1855_MOESM1_ESM.docx (79 kb)
ESM 1 (DOCX 78 kb)

References

  1. 1.
    Tsai, P.L., Chen, S.F., Huang, S.Y.: Mass spectrometry-based strategies for protein disulfide bond identification. Rev. Anal. Chem. 32(4), 257–268 (2013)CrossRefGoogle Scholar
  2. 2.
    Hunt, D.F., Yates, J.R., Shabanowits, J., Winston, S. R. H. C.: Protein sequencing by tandem mass spectrometry. Proc. Natl. Acad. Sci. USA 83, 6233–6237 (1986) .: Proteomic analysis of post-translational modifications. Nat. Biotechnol. 21(3), 255–261 (2003)Google Scholar
  3. 3.
    Mann, M., Jensen, O.N.: Proteomic analysis of post-translational modifications. Nat. Biotechnol. 21(3), 255–261 (2003)CrossRefGoogle Scholar
  4. 4.
    Qi, Y., Volmer, D.A.: Electron-based fragmentation methods in mass spectrometry: an overview. Mass Spectrom. Rev. 36, 4–5 (2017)CrossRefGoogle Scholar
  5. 5.
    Zubarev, R.A., Kelleher, N.K., McLafferty, F.M.: Electron capture dissociation of multiply charged protein cations. A nonergodic process. J. Am. Chem. Soc. 120, 3265–3266 (1998)Google Scholar
  6. 6.
    Yu, X., Warme, C., Lee, D., Zhong, W.: Characterization of a low-level unknown isomeric degradation product using an integrated online-offline top-down tandem mass spectrometry platform. Anal. Chem. 85, 8964–8967 (2013)CrossRefGoogle Scholar
  7. 7.
    Savitski, M. M., Nielsen, M. L., Zubarev, R. A., Side-chain losses in electron capture dissociation to improve peptide identification. Anal. Chem. 79(6), 2296–2302 (2007)Google Scholar
  8. 8.
    Yu, X., Zhong, W.: Differentiation of norvaline and valine in peptides by hot electron capture dissociation. Anal. Chem. 88, 5914–5919 (2016)CrossRefGoogle Scholar
  9. 9.
    Liang, Z., Zhang, Z., Wolff, J.W., Thompson, C.J., Zhong, W.: Implementation of electron-induced dissociation mass spectrometry technique for differentiation of isomeric metabolites of diclofenac. Rapid Commun. Mass Spectrom. 31, 1471–1475 (2017)CrossRefGoogle Scholar
  10. 10.
    Lioe, H., O'Hair, R.A.J.: Comparison of collision-induced dissociation and electron-induced dissociation of singly protonated aromatic amino acids, cystine, and related simple peptides using a hybrid linear ion trap-FT-ICR mass spectrometer. Anal. Bioanal. Chem. 389(5), 1429–1437 (2007)CrossRefGoogle Scholar
  11. 11.
    Zubarev, R.A.: Electron-capture dissociation tandem mass spectrometry. Curr. Opin. Biotechnol. 15(1), 12–16 (2004)CrossRefGoogle Scholar
  12. 12.
    Cody, R.B., Freiser, B.S.: Electron-impact excitation of ions from organics – alternative to collision-induced dissociation. Anal. Chem. 51(4), 547–551 (1979)Google Scholar
  13. 13.
    Liu, F., van Breukelen, B., Heck, A.J.R.: Facilitating protein disulfide mapping by a combination of pepsin digestion, electron transfer higher energy dissociation (EThcD), and a dedicated search algorithm SlinkS. Mol. Cell. Proteom. 13(10), 2776–2786 (2014)CrossRefGoogle Scholar
  14. 14.
    Gorman, J.J., Wallis, T.P., Pitt, J.J.: Protein disulfide bond determination by mass spectrometry. Mass Spectrom. Rev. 21(3), 183–216 (2002)CrossRefGoogle Scholar
  15. 15.
    Trivedi, M.V., Laurence, J.S., Siahaan, T.J.: The role of thiols and disulfides on protein stability. Curr. Protein Pept. Sci. 10(6), 614–625 (2009)Google Scholar
  16. 16.
    Tran, J.C., Zamdborg, L., Ahlf, D.R., Lee, J.E., Catherman, A.D., Durbin, K.R., Tipton, J.D., Vellaichamy, A., Kellie, J.F., Li, M.X., Wu, C., Sweet, S.M.M., Early, B.P., Siuti, N., LeDuc, R.D., Compton, P.D., Thomas, P.M., Kelleher, N.L.: Mapping intact protein isoforms in discovery mode using top-down proteomics. Nature. 480(7376), 254–U141 (2011)CrossRefGoogle Scholar
  17. 17.
    Zhang, H., Ge, Y.: Comprehensive analysis of protein modifications by top-down mass spectrometry. Circ. Cardiovasc. Gene. 4(6), 711 (2011)CrossRefGoogle Scholar
  18. 18.
    Peng, Y., Chen, X., Sato, T., Rankin, S.A., Tsuji, R.F., Ge, Y.: Purification and high-resolution top-down mass spectrometric characterization of human salivary alpha-amylase. Anal. Chem. 84(7), 3339–3346 (2012)CrossRefGoogle Scholar
  19. 19.
    Xu, H., Zhang, L., Freitas, M.A.: Identification and characterization of disulfide bonds in proteins and peptides from tandem MS data by use of the MassMatrix MS/MS search engine. J. Proteome Res. 7(1), 138–144 (2008)Google Scholar
  20. 20.
    Na, S., Paek, E., Choi, J.S., Kim, D., Lee, S.J., Kwon, J.: Characterization of disulfide bonds by planned digestion and tandem mass spectrometry. Mol. Biosyst. 11(4), 1156–1164 (2015)CrossRefGoogle Scholar
  21. 21.
    Clark, D.F., Go, E.P., Toumi, M.L., Desaire, H.: Collision induced dissociation products of disulfide-bonded peptides: ions result from the cleavage of more than one bond. J. Am. Soc. Mass Spectrom. 22 (3), 492–498 (2011)Google Scholar
  22. 22.
    Kienzle, P.: Periodic Table 1.5.0. http://www.reflectometry.org/danse/elements.html. Accessed 17 July 2017
  23. 23.
    Gohlke, C.: Molmass 2015.01.29. http://www.lfd.uci.edu/~gohlke/. Accessed 17 July 2017

Copyright information

© American Society for Mass Spectrometry 2018

Authors and Affiliations

  1. 1.Analytical Research and DevelopmentMRL, Merck and Co., Inc.RahwayUSA
  2. 2.Chemistry Modeling and InformaticsMRL, Merck and Co., Inc.KenilworthUSA

Personalised recommendations