international journal of mass spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 x....

13
International Journal of Mass Spectrometry 308 (2011) 142–154 Contents lists available at ScienceDirect International Journal of Mass Spectrometry jou rn al h om epa ge: www.elsevier.com/locate/ijms Investigation of VUV photodissociation propensities using peptide libraries Xiaohui Liu a , Yong Fuga Li b , Brian C. Bohrer a , Randy J. Arnold a , Predrag Radivojac b , Haixu Tang b , James P. Reilly a,a Department of Chemistry, Indiana University, Bloomington, IN 47405, United States b School of Informatics and Computing, Indiana University, Bloomington, IN 47408, United States a r t i c l e i n f o Article history: Received 17 February 2011 Received in revised form 25 April 2011 Accepted 27 April 2011 Available online 5 May 2011 Keywords: De novo sequencing Photodissociation Mass spectrometry MALDI-TOF/TOF a b s t r a c t PSD does not usually generate a complete series of y-type ions, particularly at high mass, and this is a limitation for de novo sequencing algorithms. It is demonstrated that b 2 and b 3 ions can be used to help assign high mass x N2 and x N3 fragments that are found in vacuum ultraviolet (VUV) photofragmentation experiments. In addition, v N -type ion fragments with side chain loss from the N-terminal residue often enable confirmation of N-terminal amino acids. Libraries containing several thousand peptides were examined using photodissociation in a MALDI-TOF/TOF instrument. 1345 photodissociation spectra with a high S/N ratio were interpreted. © 2011 Elsevier B.V. All rights reserved. 1. Introduction An important component of characterizing biological systems is protein identification [1–3]. Peptides are typically generated by enzymatically digesting proteins and analyzed using a variety of chromatographic and mass spectrometric (MS) methods [4–7]. Most often, the products of peptide ion fragmentation are identi- fied by comparison with predicted fragments derived from protein sequence databases. In this approach, precursor masses are used to pre-filter peptide candidates. There are several database search engines available such as Mascot [8], Sequest [9], X!Tandem [10], Spectrum Mill [11], Omssa [12] and others [13–20]. One limi- tation to this method is that it is only applicable to organisms with sequenced genomes. Incomplete or incorrect databases can preclude identification or lead to errors [21]. In addition, pep- tides with unexpected post-translational modifications (PTM) are a challenge since their precursor masses are shifted. The study of post-translational modifications is important for investigating the functions of proteins. It is generally recognized that de novo sequencing can alleviate some of these problems. In this approach, peptide sequences are directly derived from mass spectromet- ric data without the use of a protein database. Mass differences between appropriately chosen peaks in spectra are measured and Corresponding author at: Department of Chemistry, College of Arts and Sciences, Indiana University, Chemistry Building, 800 E. Kirkwood Avenue, Bloomington, IN 47405-7102, United States. Tel.: +1 812 855 1980; fax: +1 812 855 8300. E-mail address: [email protected] (J.P. Reilly). interpreted. Several de novo sequencing algorithms such as PEAKS, PepNovo, and MSNovo have been developed to interpret low- energy CID spectra [22–28]. Since they rely on mass differences, a complete series of fragment ions is essential. Unfortunately, most CID spectra do not provide complete sequence information due to preferential bond cleavages and spectra are often dominated by a limited number of peaks [29]. This makes de novo sequencing difficult. Likewise, different types of ion fragments must be dis- tinguished for this approach to be successful. Incorrect fragment assignments can lead to peptide sequence errors. Finally, isomeric leucine and isoleucine, which together typically represent more than 16% of the amino acids in a database, are not distinguished by low-energy fragmentation [30]. Various fragmentation techniques have been developed to gen- erate more complete sequencing information [31]. For example, electron capture dissociation (ECD) and electron transfer dissocia- tion (ETD) are now widely used [32,33]. These two radical-induced processes generate a series of c- and z ions [34,35]. The frag- ments from ECD and ETD can be employed for de novo sequencing since they extend through entire sequences except N-terminal to proline [36]. C-type ions are 17 Da heavier than corresponding b- type ions and y-type ions are 16 Da heavier than z -type ions. Therefore, the combination of CID with ECD or ETD data yields b/c and y/z ion pairs that can confirm peak identifications and improve the confidence of sequencing results [37–39]. However, ECD and ETD require multiply charged ions formed by ESI and not by MALDI. High-energy CID fragmentation is an alternative that improves fragmentation yields and produces more fragment ions than low-energy CID. High-energy CID also produces side chain loss 1387-3806/$ see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ijms.2011.04.008

Upload: others

Post on 13-Oct-2019

9 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

I

XPa

b

a

ARRAA

KDPMM

1

iboMfisteStwptaotsprb

I4

1d

International Journal of Mass Spectrometry 308 (2011) 142– 154

Contents lists available at ScienceDirect

International Journal of Mass Spectrometry

jou rn al h om epa ge: www.elsev ier .com/ locate / i jms

nvestigation of VUV photodissociation propensities using peptide libraries

iaohui Liua, Yong Fuga Lib, Brian C. Bohrera, Randy J. Arnolda,redrag Radivojacb, Haixu Tangb, James P. Reillya,∗

Department of Chemistry, Indiana University, Bloomington, IN 47405, United StatesSchool of Informatics and Computing, Indiana University, Bloomington, IN 47408, United States

r t i c l e i n f o

rticle history:eceived 17 February 2011eceived in revised form 25 April 2011ccepted 27 April 2011

a b s t r a c t

PSD does not usually generate a complete series of y-type ions, particularly at high mass, and this is alimitation for de novo sequencing algorithms. It is demonstrated that b2 and b3 ions can be used to helpassign high mass xN−2 and xN−3 fragments that are found in vacuum ultraviolet (VUV) photofragmentationexperiments. In addition, vN-type ion fragments with side chain loss from the N-terminal residue often

vailable online 5 May 2011

eywords:e novo sequencinghotodissociationass spectrometry

enable confirmation of N-terminal amino acids. Libraries containing several thousand peptides wereexamined using photodissociation in a MALDI-TOF/TOF instrument. 1345 photodissociation spectra witha high S/N ratio were interpreted.

© 2011 Elsevier B.V. All rights reserved.

ALDI-TOF/TOF

. Introduction

An important component of characterizing biological systemss protein identification [1–3]. Peptides are typically generatedy enzymatically digesting proteins and analyzed using a varietyf chromatographic and mass spectrometric (MS) methods [4–7].ost often, the products of peptide ion fragmentation are identi-

ed by comparison with predicted fragments derived from proteinequence databases. In this approach, precursor masses are usedo pre-filter peptide candidates. There are several database searchngines available such as Mascot [8], Sequest [9], X!Tandem [10],pectrum Mill [11], Omssa [12] and others [13–20]. One limi-ation to this method is that it is only applicable to organismsith sequenced genomes. Incomplete or incorrect databases canreclude identification or lead to errors [21]. In addition, pep-ides with unexpected post-translational modifications (PTM) are

challenge since their precursor masses are shifted. The studyf post-translational modifications is important for investigatinghe functions of proteins. It is generally recognized that de novoequencing can alleviate some of these problems. In this approach,

eptide sequences are directly derived from mass spectromet-ic data without the use of a protein database. Mass differencesetween appropriately chosen peaks in spectra are measured and

∗ Corresponding author at: Department of Chemistry, College of Arts and Sciences,ndiana University, Chemistry Building, 800 E. Kirkwood Avenue, Bloomington, IN7405-7102, United States. Tel.: +1 812 855 1980; fax: +1 812 855 8300.

E-mail address: [email protected] (J.P. Reilly).

387-3806/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.ijms.2011.04.008

interpreted. Several de novo sequencing algorithms such as PEAKS,PepNovo, and MSNovo have been developed to interpret low-energy CID spectra [22–28]. Since they rely on mass differences,a complete series of fragment ions is essential. Unfortunately, mostCID spectra do not provide complete sequence information due topreferential bond cleavages and spectra are often dominated bya limited number of peaks [29]. This makes de novo sequencingdifficult. Likewise, different types of ion fragments must be dis-tinguished for this approach to be successful. Incorrect fragmentassignments can lead to peptide sequence errors. Finally, isomericleucine and isoleucine, which together typically represent morethan 16% of the amino acids in a database, are not distinguished bylow-energy fragmentation [30].

Various fragmentation techniques have been developed to gen-erate more complete sequencing information [31]. For example,electron capture dissociation (ECD) and electron transfer dissocia-tion (ETD) are now widely used [32,33]. These two radical-inducedprocesses generate a series of c- and z•ions [34,35]. The frag-ments from ECD and ETD can be employed for de novo sequencingsince they extend through entire sequences except N-terminal toproline [36]. C-type ions are 17 Da heavier than corresponding b-type ions and y-type ions are 16 Da heavier than z•-type ions.Therefore, the combination of CID with ECD or ETD data yieldsb/c and y/z•ion pairs that can confirm peak identifications andimprove the confidence of sequencing results [37–39]. However,

ECD and ETD require multiply charged ions formed by ESI and notby MALDI. High-energy CID fragmentation is an alternative thatimproves fragmentation yields and produces more fragment ionsthan low-energy CID. High-energy CID also produces side chain loss
Page 2: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154 143

Table 1Composition of synthetic peptide libraries.

1 2 3 4 5 6 7 8 9 10 11 12 # of peptides

BB10AN T M I L S I C T R

5184E L A D H E P V L KG Y G Q

BB11AY Q A V G P N P G I R

4608A T E H L S T L S E KI F

G N D I G H P E M Q A RT

fiarMpfPvf

d[tttdmgdltritost[tuoittsp1taMotnotptpopam

BB12A F L A S V S

ragments including w-type ions that distinguish leucine andsoleucine [40,41]. However, high-energy CID also generates anbundance of internal fragments particularly in the low-massegion and sequence information is still typically incomplete. InALDI TOF/TOF experiments, the most common fragmentation

rocess is post source decay (PSD) that involves precursor ionragmentation during its flight in the mass spectrometer [42,43].SD typically generates a-, b- and y-type ions. Because it is notery efficient, PSD rarely provides a complete series of peptideragments.

We have shown that 157 nm photodissociation generates abun-ant high-energy peptide ion fragments along the entire backbone44–49]. It produces a series of x-type ions for peptides with C-erminal arginine or C-terminal lysine, if the lysine is convertedo homoarginine by guanidination [44,50,52]. Abundant v- and w-ype ions are also observed enabling leucine and isoleucine to beistinguished. Photodissociation has been implemented in a com-ercial MALDI-TOF/TOF instrument and high quality spectra are

enerated with this sensitive mass spectrometer [50]. Postsourceecay (PSD) spectra can be recorded with the photodissociation

ight turned off. A de novo sequencing algorithm that combines pho-odissociation and PSD data has been developed [51]. It exploitselationships such as the 26 Da spacing between x- and y-typeons to identify x/y ion pairs. Mass differences between adjacent x-ype ions lead to peptide sequence information. The identificationf corresponding v- and w-type ions increases the confidence ofequence assignments. The algorithm has been successfully appliedo the identification of Deinococcus radiodurans ribosomal proteins52]. Five or more residues were sequenced in 298 ribosomal pep-ides using this sequencing algorithm. One shortcoming that wasncovered is that the N-terminal regions of tryptic peptides areften incompletely sequenced. This is a consequence of the lim-ted number of high-mass y-type ions that are detected leadingo a limited number of high-mass x/y pairs. In order to improvehe de novo sequencing algorithm and further investigate photodis-ociation propensities, three libraries containing several thousandeptides were synthesized. The libraries contain peptides with0–12 residues all with C-terminal arginine or lysine. Each of thehree peptide library mixtures was chromatographically separatednd analyzed in both photodissociation and PSD modes in ourALDI TOF/TOF instrument. Since the MALDI ionization efficiency

f lysine-terminated peptides is not as high as it is for arginine-erminated peptides, very few of the former were detected, andone of that data is included in this study. Although guanidinationf lysine eliminates these problems [52], for the present mechanis-ic study, it was simpler to just analyze the arginine-terminatedeptides. Because the library peptide sequences are known andhere are no post-translational modifications, the number of falseositive identifications should be limited and the effect that vari-

us residues have on photodissociation propensities can be directlyrobed. The observations of low-mass b-type ions and high-mass v-nd w-type ions that are discussed below should lead to improve-ents in peptide de novo sequencing.

4096V L T S K

2. Experiment

2.1. Materials

Acetonitrile (ACN) and trifluoacetic acid (TFA) wereobtained from EMD chemicals (Gibbstown, NJ). �-Cyano-4-hydroxycinnamic acid (CHCA) was purchased from Sigma (St.Louis, MO) and Protea Biosciences (Morgantown, WV). Phenylac-etamidomethyl (PAM) resin beads were bought from MidwestBiotech (Indianapolis, IN). 3-(Diethoxy-phosphoryloxy)-3H-benzo[d][1,2,3]triazin-4-one (DEPBT) and N-tert-butoxycarbonyl (Boc)-protected amino acids were obtained from National Biochemicals(Twinsburg, OH).

2.2. Peptide synthesis

Three peptide libraries ranging from 10 to 12 residues in lengthwere synthesized [53]. The sequences of components are listedin Table 1. A total of 13,888 sequences were represented in thethree libraries. The hydrophobicity distribution of each library wasdesigned to imitate the peptides from human protein tryptic digests[53]. The peptides were synthesized following common solid-phasesynthesis and split-and-mix techniques [54]. The synthesis startedwith PAM resin beads with arginine preloaded. 10-fold molarexcess of Boc-protected amino acids and DEPBT was added to thevessel containing PAM resin beads to assure that all the resin wascoupled with the appropriate amino acid. Then the beads weredivided into equal portions and transferred to separate vessels inorder to incorporate multiple residues at one position. The couplingof the next amino acid took place in individual vessels with thepresence of Boc-protected residues and DEPBT at a 10-fold molarexcess. When the synthesis was complete, peptides were cleavedfrom the resin and lyophilized.

2.3. LC separation and MALDI spotting

The three synthetic libraries were chromatographically sepa-rated and analyzed respectively. Synthetic peptides were separatedby reverse phase chromatography (Eksigent nanoLC-2D System,Dublin, CA) and spotted onto MALDI plates using a robotic spot-ter (Eksigent, Dublin, CA). Mobile phase A was 100% water with0.1% TFA and mobile phase B was 100% acetonitrile with 0.1%TFA. The spotter combined matrix solution containing 10 g/L CHCAin 70% acetonitrile, 30% water and 0.1% TFA with the reversephase effluent. The flowrates of the nanoLC and matrix solutionwere 600 nL/min and 300 nL/min respectively. 9 �L of 0.2 mg/mLsample was loaded into the column (∼0.3 pmol of each pep-

tide). 0.6 �L spots were deposited every 40 s. Most peptideswere deposited onto one or two spots, so each spot contain0.1–0.3 pmol of each peptide. The reverse phase gradient lasted for120 min.
Page 3: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

1 f Mass

2

mFscFwpsiPoMfp

2

d[paMllpdliitamtTTcppdttti

2

pa0wMc

2

pbalb

44 X. Liu et al. / International Journal o

.4. MALDI TOF/TOF mass spectrometry

Samples deposited on MALDI plates were analyzed by ourodified 4700 MALDI TOF/TOF instrument (Applied Biosystems,

ramingham, MA) that has previously been described [50]. Masspectra were first collected using the reflectron-TOF. Photodisso-iation was performed by vacuum ultraviolet light (VUV) from an2 laser (Coherent Lambda Physik, Germany). Peptide PSD spectraere separately collected with the VUV light turned off. In eitherhotodissociation or PSD experiments, the 10 most intense precur-or ions with S/N above 80 from each MALDI spot were selected andsolated by a timed ion gate. Precursor ions, photofragments andSD fragments all contributed to photodissociation spectra. A totalf 2000 shots were averaged for each photodissociation and PSDS/MS spectrum. For de novo sequencing, peak lists were extracted

rom both sets of data. The threshold for the software’s automatedeak detection was set at a signal to noise ratio of 10.

.5. De novo sequencing

A de novo sequencing program was previously developed toerive peptide sequences from photodissociation and PSD data51,52]. The de novo sequencing algorithm combines peak lists fromhotodissociation and PSD data to identify x/y ion pairs. As notedbove, for the lysine terminated peptides, the ionization yield inALDI source is low and thus the precursors of these peptides are

ow in abundance. Therefore, no attempt was made to sequenceysine-terminated library peptides. In a previous investigation wehotodissociated ribosomal peptides from D. radiodurans and theseata provided access to some sequences not present in our peptide

ibraries [52]. These peptides were guanidinated so they yieldedntense MALDI signals. Two examples of these spectra are includedn the present study. For the peptides with C-terminal arginine,he spectra contain numerous x- and y-type ions. Although y ionsppear in both PD and PSD spectra, x ions only occur in the for-er. The program looks for x-type ions from photodissociation data

hat are 25.98 Da heavier than y-type ions found in PSD spectra.he spacing between adjacent x-type ions identifies an amino acid.he presence of corresponding v-, w- and y-type ions enables theonfidence of sequence assignments to be estimated. If an x/y ionair is missing, a gap is left in the sequence. A homology searchrogram was developed to identify peptides by matching partiale novo sequences against a database that included all peptides inhe library. Precursor masses with a tolerance of 0.3 Da are usedo confirm homology matches. 80–90% of the peptide residues areypically derived with the de novo sequencing algorithm. Fragmenton mass accuracy was assumed to be 0.3 Da.

.6. Mascot search

The database search engine Mascot was also used to matchhotodissociation data with the peptide library sequence databasessuming tryptic digestion with no missed cleavages and the same.3 Da mass error. Fragments including a-, b-, x-, y-, y-NH3-, y-H2O,- and v-type ions were all considered by Mascot. Peptides withascot scores above 40 were considered to be confident identifi-

ations.

.7. Candidate b–x ion pairs

There are two steps to identify candidate b2/xN−2 and b3/xN−3airs. First, the 20 amino acids lead to 400 or 8000 different possible

2 or b3 sequences. However, leucine and isoleucine are isomersnd some b2 or b3 ion masses with different sequences differ byess than 0.0001 Da and for all practical purposes their masses cane considered identical. The error tolerance to assign b2 or b3 ions

Spectrometry 308 (2011) 142– 154

was set as 0.1 Da since low-mass b ions are usually measurable toabout 0.05 Da. Only 177 and 1484 b2 and b3 ion masses that differby more than 0.0001 Da exist. The observed peaks in a spectrumare compared with these 177 or 1484 precomputed theoreticalmono-isotopic m/z values. When a peak mass matches one of thesenumbers to within 0.1 Da, it is considered as a candidate b2 or b3ion. The sum of the mass of a bn ion and its complementary xN−n

ion is 26.99 Da larger than the precursor mass. This relation is usedto search for a tentative xN−n ion fragment. An attempt was madeto estimate how often false b–x ion pairs were detected using thisapproach. Spectra were divided into two groups based on Mascotpeptide identifications. In the first group, Mascot identified b2/xN−2or b3/xN−3 pairs and in the second group of spectra, none werefound. Results of comparing the two groups to estimate the numberof random matches are discussed below.

3. Results and discussion

3.1. De novo sequencing

Each synthetic library was chromatographically separated andMALDI ionized respectively. Ion fragmentation occurred from post-source decay and was induced by VUV photodissociation. Fig. 1shows MS/MS spectra for a typical library peptide, EGADLSIVTRgenerated by these two processes. The PSD spectrum contains low-energy fragments including a-, b- and y-type ions. In addition tothese, photodissociation produces x-, v- and w-type ions. (Notethat with our apparatus, PSD products necessarily appear in pho-todissociation spectra, sometimes with higher and sometimes withlower intensities than in the PSD spectrum.) Isoleucine and thre-onine each have two different �-substituents on their side chainsand as a result can produce two types of ions – wa and wb. How-ever, the latter is normally much lower in intensity than the former.This means that the ethyl group loss of isoleucine and the hydroxylgroup loss of threonine are usually favored in photodissociation,and this is the case in Fig. 1. PSD yields y1 to y6 as well as y9ions. However, y7 and y8 are missing. Consequently our de novoalgorithm recognizes only six consecutive x/y ion pairs along withthe x9, y9 pair leading to a partial sequence of [372.230] LSIVTR.In this notation, [372.230] represents the mass gap between thesequenced portion of the peptide and the precursor mass. Obser-vation of v2, v3 and v5 ions confirms threonine, valine and serineassignments. The w4 and w6 fragments also distinguish leucine andisoleucine in the peptide. Nevertheless, the program fails to identifythe four N-terminal residues. Thus, although photofragmentationprovides a complete series of x-type ions from x1 to x9, the absenceof high-mass y7 and y8 PSD ions that would confirm the identitiesof x7 and x8 leads to a sequencing gap. This is consistent with ourprevious observation that high-mass y-type ions are often missing[52]. Other data interpretation strategies are therefore needed toconfirm the identities of large x-type ions.

3.2. Synthetic peptide libraries

The three synthetic libraries of peptides were studied to extractphotofragmentation rules that might improve de novo sequencing.The synthetic libraries contain equal numbers of peptides with C-terminal arginine and lysine. However, as noted above, data foronly arginine-terminated peptides were interpreted in this study.Although it would have been helpful to have spectra with intensefragments, each synthetic library contained thousands of peptides

with similar sequences and the same length (Table 1), so the pep-tides could not be completely separated. As a result, many peptideswith similar masses were deposited on the same MALDI spot andionization competition led to some low precursor ion intensities.
Page 4: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154 145

PD

PSD

00

MH-H2O

y9

y6

y5b6y4 b6#b5a5

y3b4

y3*b4#y2y2*

y1*

y1

200 400 600 800 1000

w2

b3

MH-H2Ov10+1y9 x9x8

w4x5 x7x6

y5b6

y6

w6

b6#

y4 v5b5

x4a5v4y3

b4

x3

b4#

w3x2 v3y2

y2*

v2

x1

b2y1*

y1

tsourc

PtwcMeoasiRcp4ats(ltNypc

3

tbticpbsflc

ETMD, ATAH, AQIH and AQEH yielded almost no b3 ions. How-ever, fragmentation propensities can still be derived from othersequences. Fig. 3a displays 32 different b2 ions, their sequences

PD

PSD

y9y8y7

b3b2a2

y1

1000800600400200

w9y7 y9y8

x7 x8x9

y1

a2

b2

b3

200 400 6

Fig. 1. Photodissociation (PD) and pos

hotodissociation spectra generated from these precursors con-ained correspondingly weak peaks. Because each MS/MS spectrumas recorded by averaging 2000 laser shots, repeating this pro-

ess for the ten most intense precursor ions largely consumed aALDI spot. Therefore, the last few photodissociation spectra from

ach spot tended to be mediocre. Moreover, the automated processf depositing MALDI spots could split some peptide samples ontodjacent spots, also weakening precursor ion signals. Thus, whileynthetic peptide libraries are in principle attractive for study-ng peptide fragmentation, the experiment was still challenging.eflecting this, when Mascot was applied to interpret photodisso-iation MS/MS spectra, only 1345 of the 6944 arginine-terminatedeptides in the three libraries were identified with scores above0. The 1345 spectra were used to derive some new insightsbout photofragmentation. These include a new appreciation forhe ubiquity of low-mass b-type ions in VUV photodissociationpectra, and the surprising observation of vN- and wN-type ionswhere N is the number of residues in the peptide). Similarly, theibrary data hint at an unexpected enhanced fragmentation of pep-ides next to threonine but only when it is close to the peptide-terminus. Finally, Y-type ions (that are 2 Da lighter than their-type counterparts) were found to be much more abundant inhotofragmentation data than had been previously reported. Aomplete discussion of these observations follows.

.3. Assignment of high-mass x-type ions using b2 and b3 ions

b2 ions are commonly formed by low-energy activation of pep-ide ions [55,56]. Fig. 2 shows an example of PSD-generated b2 and3 ions. In this case, their intensities increase with photofragmen-ation. In the 1345 photodissociation spectra of synthetic peptidesdentified by Mascot, numerous small b ions appeared. Table 2 indi-ates that 1337 and 1297 of these spectra contain b2 and b3 ioneaks, respectively. Over 96% of photodissociation spectra containoth. The b2 and b3 ion intensities depend somewhat on peptide

equence. In our synthetic libraries, 14 and 32 different sequencesor b2 and b3 ions, respectively, were represented. In addition, theibraries contain a variety of different third and fourth residues thatould affect b2 and b3 ion fragmentation. Intensities of b2 and b3

800 1000

e decay (PSD) spectra of EGADLSIVTR.

ions were extracted from our 1345 peptide spectra and normalizedwith respect to the sum of the intensities of all assigned peaks.Results are displayed in Fig. 3. The numbers of identified b2 orb3 ions for each sequence are also noted in the figure. The pep-tide detection exhibits some sequence preference probably due toionization competition. Therefore, some N-terminal sequences arerarely detected. For example, very few peptides with N-terminalsequences NTM and GND were found resulting in very few b2 ions.Likewise peptides with N-terminal sequences NTMI, NTMD, GNDS,

1000800600400200

Fig. 2. b2, b3 ions and complementary x- and y-type ions in the photodissociation(PD) and postsource decay (PSD) spectrum of NLADLGIQLR.

Page 5: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

146 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154

Table 2Number of spectra for which various fragment ions were observed.

b2 xN−2 yN−2 b3 xN−3 yN−3 b2/xN−2 yN−2/xN−2 b3/xN−3 yN−3/xN−3

# of peptides 1337 1076 318 1297 1125 485 1053 247 1056 368

Fig. 3. Normalized intensities of (a) b2 ions; (b) b3 ions with various sequences and different residues to provide amine groups for cleavage bonds. In square brackets, thenumbers specify the number of [identified b2 ions/identified peptides/peptides in libraries] respectively.

Page 6: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

f Mass Spectrometry 308 (2011) 142– 154 147

airinosatffao

td1wl1cicwifA1ogmecssobwpicmtisputacm

0

50

100

150

200

250

300

350

400

>=9876543210

# of

spe

ctra

# of false candidate b 2:x N-2 pairs

0

50

100

150

200

250

300

350

>=131211109876543210#

of s

pect

ra# of false candidate b 3:xN-3 pairs

a

b

X. Liu et al. / International Journal o

nd the number of different peptides from which they derived. Its obvious that peptides with glycine or threonine on the secondesidue from the N-terminus residue show relatively low b2 ionntensities compared to glutamine, especially when isoleucine isext to glutamine. However, b3 ion intensity seems to depend moren the fourth residue in the peptide (Fig. 3b). No peptide with GLDSequence at its N-terminus was found with a Mascot search scorebove 40 even though 128 peptides in the libraries contain this N-erminal sequence. 75 identified b3 ions are listed in Fig. 3b. Barsor which standard deviations are not noted correspond to casesor which only a single peptide was detected. Peptides containingspartate or histidine at the fourth position produce relatively weakr no b3 ions.

It has been noted above that missing high-mass x/y pairs leado de novo sequencing failures even when photodissociation pro-uces complete x-type ion series. Table 2 indicates that 1076 out of345 confidently detected peptides yielded xN−2 photofragments,hile only 318 yielded corresponding yN−2 ions (where N is the

ength of the peptide). Analogously, xN−3 ions were observed for125 out of 1345 peptides but yN−3 ions were detected in only 485ases. Because of their high abundances, b2 and b3 ions can helpdentify x-type ions. It is well known that the sum of a bn and aomplementary yN−n ion is the precursor mass plus 1.01 Da. Like-ise, the sum of a bn-type ion and its complementary xN−n-type ion

s 26.99 Da larger than the precursor mass. This relation can there-ore be applied to identify corresponding high-mass x-type ions.s indicated in Table 2, in these 1345 photodissociation spectra,053 b2/xN−2 and 1056 b3/xN−3 ion pairs can be recognized, whilenly 247 xN−2/yN−2 and 368 xN−3/yN−3 ion pairs are found. This sug-ests that the identification of b2/xN−2 and b3/xN−3 ion pairs can beuch more effective for high-mass x-type ion confirmation. How-

ver, to exploit these ion pairs for sequencing, they must first beorrectly identified. Random combinations of peaks can possiblyatisfy the above mass relations. Internal fragment ions of a givenequence have the same mass as b2 and b3 ions of that sequence. Inrder to determine how often random ion pairs satisfy the relationn + xN−n = M + 26.99 (where M is the precursor mass), a programas developed to search for the b2/xN−2 and b3/xN−3 ion pairs inhotodissociation data. The 1345 spectra with confident Mascot

dentifications were examined. All low-mass fragment ions wereompared to a list of the 177 unique b2 and 1484 b3 ion massesentioned above. All matches were included in a list of poten-

ial b2/b3 candidates. The program then tested each one of theseons by looking for complementary high-mass peaks that could beummed with the candidate b2/ b3 ions to yield the precursor masslus 26.99 Da. The numbers of false b2/xN−2 and b3/xN−3 ion pairsncovered by this procedure are plotted in Fig. 4a and b respec-

ively. In each case, the abscissa is the number of false matchesnd the ordinate shows the number of peptides (i.e., spectra) thatontained this number of false matches. It can be concluded thatore than 90% of the 1345 spectra have four or fewer false b2/xN−2

Fig. 5. Comparison of the accuracy of identifying b2 and b3 fragm

Fig. 4. The number of spectra containing various numbers of false b/x ion pairs. (a)b2/xN−2 ion pairs; (b) b3/xN−3 ion pairs.

matches and over 700 spectra have zero or one false identifica-tion. Likewise, about 90% of the spectra yield five or fewer falseb3/xN−3 matches and about half have two false matches or fewer.One ameliorating factor is that in this peptide library study, morethan one precursor may be contributing to some spectra due to thewide ion isolation window. In that case, some spectra may havemore than one legitimate b2/xN−2 or b3/xN−3 ion pair. Although thisprocedure has found multiple false b/x ion pairs, one other checkcan be applied to the data: the xN−1/xN−2, xN−2/xN−3 and xN−3/xN−4mass spacings should match the mass of some amino acid. Appli-cation of these constraints will increase the number of valid x ionidentifications and reduce the number of false positives.

It has been reported that accurate masses of dipeptides canbe used to extract sequence information for peptide identification[57]. In the present case, the residues in potential b2 ion candidatescan be identified if the masses of these fragment ions are mea-sured accurately enough. All the peptides in our synthetic librariesare arginine-terminated and all of the spectra contain a 175.117 Da

y1 ion. Therefore, this mass can be used as an internal calibrant.The Data Explorer software from ABSciex was capable of recali-brating the spectra using a peak assigned with an accurate mass.

ents before and after recalibration of spectra using y1 ions.

Page 7: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

148 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154

Table 3Percentage of wN , vN , vN+1 and xN−1ions for various N-terminal amino acids.

N-terminal residue wN vN vN+1 vN+2 xN−1

Asna 69% 46% 15% 93%Glua 42% 90% 89%Alaa 28% 20% 86%Tyra 51% 85% 44%Phea 72% 76%Glya – – – – 51%

Aspb x xLeub xIleb xValb xHisb x xTrpb x xGlnb x xSerb x x xThrb x x xMetb x

a Peptides derived from synthetic libraries.

f

SigbdBmbati

3

tfvoisSNora

vcewanfsfgGitatIi

996 99 8 100 0

v10

*

v10+1

1042 104 4 104 6 1048

*

*

w10

v10

1080 108 2 108 4 108 6

*v11

v11+1

1032 103 4 1036

*

*

v10

w10

810 81 2 814 81 6

*

v8+2

v8+1v8

1312 13 14 13 16 1318

*

*v12

v12+1

v12+2

a b

d

fe

c

Fig. 6. Assignments of (a) vN+1 radical ion for EGAILSIVLR; (b) vN , wN ions ofDAIQQAFGAR; (c) v radical ion of QLVGQVAANVR; (d) v , w ions of NGAYL-

b Peptides derived from D. radiodurans ribosomal proteins. x indicates that theragment is present. – indicates ions that cannot be formed for N-terminal glycine.

ince the calibrant mass is 175.117 Da, the recalibration particularlymproves the mass accuracy of the low-mass region. We investi-ated the accuracy with which b2 and b3 ions could be measuredefore and after spectral recalibration using y1 as internal stan-ard. The results for 100 typical spectra are displayed in Fig. 5.efore recalibration, only about half of the b2 and b3 measuredasses were within 0.02 Da of their correct values. After y1 recali-

ration, about 95% of b2 and b3 ions were within 0.02 Da and almostll of these ions were within 0.03 Da mass error. This recalibra-ion increases the mass accuracy of MS/MS ion masses and therebymproves the possibility of predicting peptide N-terminal residues.

.4. Identification of xN−1, wN, vN and vN+1 ions

Photodissociation generates a series of x-type ions in the spec-ra. The mass spacing between xN−1 and the precursor is uniqueor every N-terminal residue. Therefore, the presence of xN−1 pro-ides an opportunity to identify the peptide N-terminus. Over 75%f the 1345 library spectra contained xN−1 ions, although as shownn Table 3, the probability was quite residue-dependent. The masspacings between each xN−1 and the precursor are listed in Table 4.ome masses spacings in this table are not unique for different-terminal residues. However, the peaks in the high-mass regionf photodissociation spectra are as redundant as in the low-massegion. Therefore, these mass spacings can be used to tentativelyssign xN−1 ions and the N-terminal residues.

Photodissociation of various peptides in our library yielded wN,N or vN+1 ions (where N is the number of residues) that are asso-iated with side chain loss from the N-terminal residue. Somexamples are shown in Fig. 6. Observation of these ions was some-hat surprising since it has been proposed that v- and w-type ions

re formed from x ion precursors (Scheme 1a and b) [45] and it isot possible to form xN ions. Since vN and wN ions are generated

rom unique side chain losses that are residue-dependent, the masspacings between these vN or wN type ions and the precursor shouldacilitate confirmation of peptide N-terminal residues. To investi-ate this, our library peptides with six different N-termini, Asn,lu, Ala, Tyr, Phe and Gly were examined. The top part of Table 3

ndicates the percentage of the 1345 identified spectra that con-ain five different ion fragments. N-terminal glycine can produce

n xN−1, but not a vN or wN ion. It is apparent that different N-erminal residues produce different types of vN- or wN-type ions.n our spectra, N-terminal phenylalanine did not lead to a vN+1 rad-cal, and N-terminal asparagine was the only residue that yielded a

N+1 N N

GIQTR; (e) vN , wN and vN+1 radical ions of TVQALGLR; (f) vN , wN and vN+1 radical ionsof SLDLDSIIAEVKg. *Shows the isotopic peaks and g means that K is guanidinated.

wN ion. 285 of our confidently identified peptides have N-terminalglutamic acid. 90% of these generate vN+1 radicals after photodis-sociation and about 42% yield vN ions. Although only about 40% ofthese peptides produce both vN+1 and vN ions, the vN ion inten-sity is typically three times weaker suggesting that the former isthe dominant product. Fig. 6a presents the high-mass region of aPD spectrum displaying an intense vN+1 radical ion and relativelyweak vN ion. The peaks marked with asterisks predominantly resultfrom 13C isotope contributions and they are not treated as differ-ent fragments by the instrument software. 69% of peptides withN-terminal asparagines produce wN ions and 46% of them generate

vN ions but only 33% of them yield both wN and vN fragments. Fig. 6ddisplays one example with N-terminal asparagine. Table 3 indicatesthat 15% of these peptides yield vN+1 radical ions. However, the13C isotope peak associated with the adjacent vN ion occurs at this
Page 8: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154 149

Table 4Mass differences between precursor and xN−1, vN and wN ions for different N-terminal amino acids.

Precursor-xN−1 Precursor-vN Precursor-(vN+1) Precursor-wN/(vN+2)

Ala 45.058 Da 16.031 Da 15.024 DaAsn 88.064 Da 59.037 Da 58.030 Da 60.0324 Daa

Asp 89.048 Da 60.021 Da 59.014 Da 61.0164 Daa

Gln 102.080 Da 73.053 Da 72.046 DaGlu 103.063 Da 74.037 Da 73.030 DaGly 31.042 Da – –His 111.080 Da 82.053 Da 81.046 DaIle 87.105 Da 58.078 Da 57.071 DaLeu 87.105 Da 58.078 Da 57.071 DaMet 105.061 Da 76.035 Da 75.028 DaPhe 121.089 Da 92.063 Da 91.056 DaSer 61.053 Da 32.026 Da 31.019 Da 30.012 Dab

Thr 75.068 Da 46.042 Da 45.035 Da 44.028 Dab

Trp 160.100 Da 131.074 Da 130.067 DaTyr 137.084 Da 108.058 Da 107.051 DaVal 73.089 Da 44.063 Da 43.056 Da

Arg 130.122 Da 101.095 Da 100.088 DaCys 77.030 Da 48.003 Da 46.996 DaLys 102.116 Da 73.089 Da 72.082 DaPro 71.074 Da – –

snpNNc

ra

a Precursor-wN.b Precursor-(vN+2).

ame mass, so asparagine’s vN+1 ions may well be artifacts that willot be helpful for peptide sequencing. Interestingly, tyrosine andhenylalanine have similar aromatic side chains but peptides with-terminal phenylalanine only generate vN ions and peptides with-terminal tyrosine produce both vN and vN+1 ions by photodisso-iation.

Whether or not a vN+1 ion can be used to identify an N-terminalesidue depends on the uniqueness of its mass (as listed in Table 4)nd whether it can be distinguished from side chain loss m-type

Scheme 1. Pathways to form various fragment ion

ions that would have the same mass. The latter can result fromside chain loss from any residue in the peptide [44]. For example,peptides with N-terminal glutamic acid produce strong vN+1 rad-ical ions. The m ion formed from non-terminal glutamic acid sidechain loss would have the same mass, and this would confuse vN+1radical ion identification. In our synthetic library study, only the

peptides containing glutamic acid and aspartic acid are found tohave m ions from photodissociation. We compared the intensitiesof m ions from interior glutamic acid side chain loss with those of

s: (a) v-ions; (b) w ions; (c) vN+1 and vN ions.

Page 9: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

150 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154

Sa

vopcgfbws

efP

cheme 2. Pathways to form wN and vN from (a)–(b) aspartic acid and (c)–(d)sparagine.

N+1 ions from peptides with N-terminal glutamic acid. Over halff the peptides with non-terminal glutamic acid yielded m ions inhotodissociation but their intensities were 10–20% of those asso-iated with vN+1 radical ions. Because m ions from non-terminallutamic acid side chain loss are much weaker than the vN+1 ionsrom the N-terminal glutamic acid side chain loss, the latter shoulde recognizable by sequencing programs. N-terminal aspartic acidas not included in our libraries, and this residue requires further

tudy.Because our synthetic peptide libraries contain only six differ-

nt N-terminal residues, photodissociation data of tryptic peptidesrom D. radiodurans ribosomal proteins were also briefly examined.eptides with 16 different N-termini were confidently identified

Scheme 3. Proposed pathways to form vN+2 ions from peptides with N-terminal (a)threonine; (b) serine.

using Mascot. In this sample set, only one to ten peptides for eachN-terminal residue were found, and so accurate appearance per-centages for various fragment ions were not derived. The lowersection of Table 3 simply indicates whether or not vN, wN or vN+1ions were observed in the examples that were studied. No peptideswith N-terminal arginine, cysteine, lysine and proline were in thissample set. vN ions were generally observed except for peptideswith N-terminal methionine. It can be seen that some residues yieldvN+1 radical ions but others do not. Fig. 6b and c display examplesof PD spectra of peptides with N-terminal aspartic acid and glu-tamine. Glutamic acid and glutamine terminated peptides yield thesame weak vN ions and intense vN+1 ions. By comparison, asparticacid and asparagine terminated peptides produce strong wN and vN

ions (Fig. 6b and d). The side chains of glutamic acid and glutamineare only one CH2 longer than those of aspartic acid and asparaginesand the wN ions at the latter two are apparently stabilized by for-mation of a stable CO2 side product. It is less obvious what affectsthe vN+1/vN peak ratios but the possibility of conjugation in the glu-tamic acid side product may be involved. Scheme 1c displays howvN or vN+1 fragment ions form.

Scheme 2 proposes a mechanism for producing wN and vN ionsfrom N-terminal aspartic acid and asparagine. A hydrogen bondbetween the N-terminal amine and the side chain enables the for-mation of a six-member ring. If the hydrogen atom from the amine

is bridged to the oxygen or nitrogen of the adjacent side chain, avN ion is produced following elimination of CH2CO and either H2Oor NH3 as shown in Scheme 2b, 2d. If the hydrogen from the side
Page 10: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142– 154 151

a ELAYHEIQLR

b YQ EVGPNPSER

600500400300200100

Y

AYHE

YHE

HEIAYHYH

HE

I/LQ

H

600500400300200100

GPNPS

PSEGPNQE

PSPN

GP

GP-28

R

EN/R

Y

Ftc

cftal

FS

ig. 7. Immonium ions and internal fragments from the low-mass region of pho-odissociation spectra of (a) ELAYHEIQLR; (b) YQEVGPNPSER. Highlighted residuesorrespond to immonium ions and underlined sequences label internal fragments.

hain is bridged to the nitrogen on the N-terminus, then wN ionsorm (Scheme 2a and c). Because glutamic acid and glutamine con-

ain one more CH2 and a seven-member ring is less stable, wN ionsre not formed and vN ions are not the primary product. Instead,ight activation breaks the N-terminal side chain C�–C� bond, the

a

b

c

11001000900800

x9y9x8y8

11001000900800

y8x8 x9y9

11001000900800

y8 y9

x8

x9

ig. 8. High-mass regions of photodissociation spectra of (a) ETAILSIVLR; (b) ELAIL-PVLR; (c) NLADHEIVLR.

Fig. 9. Y-type ions in photodissociation spectra of (a) FQAAAGQLAQPHR; (b)YQEHLSNPFER; (c) EGAYLGPVLKg. The g subscript means that K is guanidinated.

side chain group is eliminated as a radical (that probably loses ahydrogen to form a conjugated species) and the peptide is left as avN+1 radical ion. Presumably N-terminal residues other than aspar-tic acid and asparagine lose their side chains to form either vN andvN+1 ions as shown in Scheme 1c. Interestingly, asparagine andaspartic acid are the only two N-terminal residues that yield wN

ions (Table 3). Even though the mass difference between these wN

ions and precursor is not unique (as shown in Table 4), based onother information derived from the de novo program, the presenceof wN ions for N-terminal aspartic acid or asparagine should helpto identify these N-terminal residues.

Peptides with N-terminal threonine and serine produce strongvN+2 ion peaks as seen, for example, in Fig. 6e and f. Proposed mech-anisms of vN+2 ion formation for threonine and serine are shownin Scheme 3a and b. The hydrogen bond between hydroxyl hydro-gen and carbonyl oxygen facilitates a six-member ring formation,which helps hydrogen transfer and side chain elimination. The vN+2ion is produced after the elimination of CH3CHO and CH2O groupsfor threonine and serine respectively and hydrogen is rearrangedto form a carbonyl group. As shown in Table 4, the mass of a vN+2ion for N-terminal threonine happens to match the mass of theprecursor with CO2 loss. However, the unique mass of vN+2 ion forN-terminal serine is indicative of its presence. Threonine and serineare the only two N-terminal residues that yield all three vN, vN+1

and vN+2 ions (Table 3) and this should be useful for confirmingtheir presence. Since the D. radiodurans ribosomal protein sampleprovided only a few examples for each N-terminal residue, furtherstudy of these peptides is warranted. Table 4 lists the mass spac-
Page 11: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

1 f Mass

iaun

3

itetidfttFitcNd

3

gsaameid2ppBctoo

mlttiwnrrtftnatItmgdds

Acknowledgements

The authors thank Youyou Yang and Dr. Liangyi Zhang for pro-

52 X. Liu et al. / International Journal o

ng between the precursor ion and the largest x, v and w ions forll possible N-terminal residues. Some of these mass spacings arenique. In other cases, they can be used as constraints to limit theumber of N-terminal residue candidates.

.5. Immonium ions and internal fragments

Immonium ions and internal fragments are commonly observedn all photodissociation spectra of synthetic peptide libraries, andheir observation has been discussed previously [44,50,61]. Notvery residue yields an immonium ion. Fig. 7A and B displays spec-ra that include immonium ions of histidine, arginine, glutamine,soleucine and tyrosine. Histidine and tyrosine immonium ions areominant peaks. Photodissociation also produces immonium ionsor proline, phenylalanine and tryptophan that are not shown inhese examples. Such immonium ions should help to confirm pep-ide de novo sequences. Numerous internal fragments also appear inig. 7. Typical of our synthetic library spectra, most of these intensenternal fragments are either dipeptides or tripeptides that con-ain proline or histidine. These internal ions provide 50% sequenceoverage of ELAYHEIQLR and 73% sequence coverage of YQEVGP-PSER. Therefore, their recognition should be useful for confirminge novo sequences and identifying peptides [57].

.6. Glutamic acid, asparagine and threonine effects

It is well known that fragmentation is enhanced C-terminal tolutamic acid [58,59]. Fig. 8 displays a typical example of photodis-ociation spectra of peptides with glutamic acid and asparaginet the peptide N-terminus. The peptides with N-terminal glutamiccid provide intense x9 and y9 fragments (Fig. 8a and b), but y9 isissing for the peptide with N-terminal asparagine (Fig. 8c). The

nhanced cleavage next to glutamic acid facilitates recognition ofts x/y ion pair. In our 1345 peptides whose spectra were confi-ently identified by Mascot, there were 956 glutamic acid residues,84 of them at the N-terminus. 97% of the glutamic acid-containingeptides cleave next to this residue to generate y-type ions. 80% ofeptides with N-terminal glutamic acid produce yN−1 fragments.y comparison, of 618 asparagine-containing peptides, about 42%leave next to asparagine to form y ions, but only 18 out of 272 pep-ides with N-terminal asparagine produce yN−1 fragments. Thesebservations directly impact de novo sequencing that requires thebservation of y-type ions.

Surprisingly, threonine was also found to enhance peptide frag-entation when it is close to the peptide N-terminus. In synthetic

ibraries, 330 out of 1345 identified peptides contain threonine nexto the N-terminus and 268 of them produce yN−2 fragments. By con-rast, the other 1015 peptides generate only 50 yN−2 ions. The yN−2ntensity normalized by the total intensity of all assigned peaks

as extracted and compared for different residues. The average oformalized yN−2 intensity for threonine at the second N-terminalesidue is 0.008431 and it is only 0.00038 for all other secondesidues. However, when threonine is the third residue from the C-erminus, the averaged y2 intensity is 0.066731 and it is 0.053687or all other residues on the same position. Therefore, the fragmen-ation enhancement C-terminal to threonine is not obvious if it isear the C-terminus rather than the N-terminus. Fig. 8 displaysn example of the threonine effect. The peptides involved havehreonine or leucine as the second residue from the N-terminus.n these rather typical spectra, cleavage next to threonine appearso be enhanced. Although the mechanism for this is not clear, it

ay involve the side chain OH interacting with the adjacent amide

roup through hydrogen bonding just as the OH of aspartic acidoes when facilitating fragmentation next to that residue [59]. Theissociation of peptide ions formed by electrospray ionization wastudied by Wysocki and coworkers and no fragmentation enhance-

Spectrometry 308 (2011) 142– 154

ment next to threonine was observed [29,60]. Further work onsingly charged ions with threonine and serine near the N-terminusis warranted.

3.7. Identification of Y-type ions

Photodissociation produces intense Y-type ions terminated byproline [51,52]. The observation of Y/y pairs in photodissociationand PSD data can be used to identify proline residues [41,42].However, examination of over one thousand library spectra led tothe conclusion that cleavage next to glycine, glutamine, alanine,proline, histidine and tyrosine can also yield Y-type ions by pho-todissociation. Fig. 9 shows some typical examples. Only y- andY-type ions are marked in these spectra. The Y ions formed by cleav-age next to proline are certainly the most intense in Fig. 9. The Yions from glutamine and alanine terminated fragments are also sig-nificant but less intense in Fig. 9. Y ions terminated with histidineare typically relatively weak (e.g., Y2 in Fig. 9a and Y8 in Fig. 9b).However, Y-type ions terminated by the residues glycine, glutamicacid, alanine, histidine and tyrosine do not appear consistently inphotodissociation spectra. For example, glycine N-terminated frag-ments have both Y8 and y8 ions in Fig. 9a, but only have y5 and y9ions in Fig. 9c. In photodissociation of synthetic peptides, Y-typeions terminated by glutamine are generated for only 65% of 232spectra containing this residues; Likewise, tyrosine-terminated Yions appear in only 47% of 176 peptide spectra that include thisresidue. Therefore, the presence of Y-type ions in photodissocia-tion spectra cannot be relied upon for sequencing efforts. However,since the library data suggest that Y-type ions terminated by justcertain residues tend to form, this can be used as corroboratinginformation.

4. Conclusion

Three synthetic peptide libraries with C-terminal arginine wereanalyzed by a modified 4700 MALDI TOF/TOF instrument tostudy 157 nm photodissociation. Some fragmentation rules wereextracted with the eventual goal of improving de novo sequencing.Sequencing peptide N-terminal regions often fails because high-mass y-type ions are missing from spectra. However, this workdemonstrated that b2 and b3 ions appear in most photodissociationspectra. Their ubiquity enables b2/xN−2 and b3/xN−3 ion pairs to beidentified which should help confirm the identities of high-mass x-type ions and low-mass b2 and b3 ions. Accurate masses of b2 ionsshould also facilitate assigning the first two N-terminal residues.The spacings between precursor mass and xN−1 ion fragmentscan often determine the peptide N-terminal amino acid. In addi-tion, vN-, wN- and vN+1-type ions can help to confirm N-terminalresidue assignments. Immonium ions and internal fragments canbe utilized to confirm de novo derived sequences. The photofrag-mentation propensities observed in this work should improve thephotofragmentation-based sequencing of peptides. Based on theseimprovements, we anticipate being able to identify more of thepeptides in these libraries.

viding photodissociation spectra from D. radiodurans ribosomalproteins. This work is supported by grants from the National Insti-tutes of Health (R01 RR024236-01A1) and the National ScienceFoundation (CHE-1012855).

Page 12: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

f Mass

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

X. Liu et al. / International Journal o

eferences

[1] R. Aebersold, D.R. Goodloett, Mass spectrometry in proteomics, Chem. Rev. 101(2) (2001) 269–295.

[2] R. Aebersold, M. Mann, Mass spectrometry-based proteomics, Nature 422(6928) (2003) 198–207.

[3] J.R. Yates, Mass spectrometry and the age of the proteome, J. Mass Spectrom.33 (1) (1998) 1–19.

[4] D.F. Hunt, J.R. Yates, J. Shabanowitz, S. Winston, C.R. Hauer, Protein sequenc-ing by tandem mass spectrometry, Proc. Natl. Acad. Sci. U.S.A. 83 (17) (1986)6233–6237.

[5] K. Biemann, H.A. Scoble, Characterization by tandem mass spectrometry ofstructural modifications in proteins, Science 237 (4818) (1987) 992–998.

[6] W.J. Henzel, T.M. Billeci, J.T. Stults, S.C. Wong, C. Grimley, C. Watanabe, Iden-tifying proteins from two-dimensional gels by molecular mass searching ofpeptide fragments in protein sequence databases, Proc. Natl. Acad. Sci. U.S.A.90 (11) (1993) 5011–5015.

[7] J.R. Yates, S. Speicher, P.R. Griffin, T. Hunkapiller, Peptide mass maps: a highlyinformative approach to protein identification, Anal. Biochem. 214 (2) (1993)397–408.

[8] D.N. Perkins, D.J.C. Pappin, D.M. Creasy, J.S. Cottrell, Probability-based proteinidentification by searching sequence databases using mass spectrometry data,Electrophoresis 20 (18) (1999) 3551–3567.

[9] J.R. Yates, J.K. Eng, K.R. Clauser, A.L. Burlingame, Search of sequence databaseswith uninterpreted high-energy collision-induced dissociation spectra of pep-tides, J. Am. Soc. Mass Spectrom. 7 (11) (1996) 1089–1098.

10] R. Craig, R.C. Beavis, Tandem: matching proteins with tandem mass spectra,Bioinformatics 20 (9) (2004) 1466–1467.

11] D. Yi, P.D. Perkins, Identification of ubiquitin nitration and oxidation using aliquid chromatography/mass selective detector system, J. Biomol. Technol. 16(4) (2005) 364–370.

12] L.Y. Geer, S.P. Markey, J.A. Kowalak, L. Wagner, M. Xu, D.M. Maynard, X.Y. Yang,W.Y. Shi, S.H. Bryant, Open mass spectrometry search algorithm, J. ProteomeRes. 3 (5) (2004) 958–964.

13] C.X. Xu, B. Ma, Software for computational peptide identification from MS–MSdata, Drug Discov. Today 11 (13–14) (2006) 595–600.

14] N. Zhang, X.J. Li, M.L. Ye, S. Pan, B. Schwikowski, R. Aebersold, ProblDtree: anautomated software program capable of identifying multiple peptides froma single collision-induced dissociation spectrum collected by a tandem massspectrometer, Proteomics 5 (16) (2005) 4096–4106.

15] V. Bafna, N. Edwards, SCOPE: a probabilistic model for scoring tandemmass spectra against a peptide database, Bioinformatics 17S (2001) S13–S21.

16] N. Zhang, R. Aebersold, B. Schwilkowski, ProbID: a probabilistic algorithm toidentify peptides through sequence database searching using tandem massspectral data, Proteomics 2 (10) (2002) 1406–1412.

17] J. Colinge, A. Masselot, M. Giron, T. Dessingy, J. Magnin, OLAV: towards high-throughput tandem mass spectrometry data identification, Proteomics 3 (8)(2003) 1454–1463.

18] Y. Fu, Q. Yang, R.X. Sun, D.Q. Li, R. Zeng, C.X. Ling, W. Gao, Exploiting the ker-nel trick to correlate fragment ions for peptide identification via tandem massspectrometry, Bioinformatics 20 (12) (2004) 1948–1954.

19] Y.H. Wan, A. Yang, T. Chen, PepHMM: a hidden Markov model based scoringfunction for mass spectrometry database search, Anal. Chem. 78 (2) (2006)432–437.

20] D.L. Tabb, C. Narasimhan, M.B. Strader, R.L. Hettich, DBDigger: reorganized pro-teomic database identification that improves flexibility and speed, Anal. Chem.77 (8) (2005) 2464–2474.

21] A. Keller, A.I. Nesvizhskii, E. Kolker, R. Aebersold, Empirical statistical model toestimate the accuracy of peptide identifications made by MS/MS and databasesearch, Anal. Chem. 74 (20) (2002) 5383–5392.

22] J.A. Taylor, R.S. Johnson, Implementation and uses of automated de novo pep-tide sequencing by tandem mass spectrometry, Anal. Chem. 73 (11) (2001)2594–2604.

23] Z.Q. Zheng, De novo peptide sequencing based on a divide-and-conquer algo-rithm and peptide tandem spectrum simulation, Anal. Chem. 76 (21) (2004)6374–6383.

24] M. Bern, D. Goldberg, De novo analysis of peptide tandem mass spec-tra by spectral graph partitioning, J. Comput. Biol. 13 (2) (2006) 364–378.

25] A. Frank, P. Pevzner, PepNovo: de novo peptide sequencing via probabilisticnetwork modeling, Anal. Chem. 77 (4) (2005) 964–973.

26] B. Fischer, V. Roth, F. Roos, J. Grossmann, S. Baginsky, P. Widmayer, W. Gruis-sem, J.M. Buhmann, NovoHMM: a hidden Markov model for de novo peptidesequencing, Anal. Chem. 77 (22) (2005) 7265–7273.

27] B. Ma, K.Z. Zhang, C. Hendrie, C.Z. Liang, M. Li, A. Doherty-Kirby, G. Lajoie,PEAKS: powerful software for peptide de novo sequencing by tandemmass spectrometry, Rapid Commun. Mass Spectrom. 17 (20) (2003) 2337–2342.

28] L.J. Mo, D. Dutta, Y.H. Wan, T. Chen, MSNovo: a dynamic programming algo-rithm for de novo peptide sequencing via tandem mass spectrometry, Anal.

Chem. 79 (13) (2007) 4870–4878.

29] Y.Y. Huang, J.M. Triscari, G.C. Tseng, L. Pasa-Tolic, M.S. Lipton, R.D. Smith, V.H.Wysocki, Statistical characterization of the charge state and residue depen-dence of low-energy CID peptide dissociation patterns, Anal. Chem. 77 (18)(2005) 5800–5813.

[

Spectrometry 308 (2011) 142– 154 153

30] F. Kjeldsen, K.F. Haselmann, E.S. Sorensen, R.A. Zubarev, Distinguishing ofIle/Leu amino acid residues in the PP3 protein by (hot) electron capture disso-ciation in Fourier transform ion cyclotron resonance mass spectrometry, Anal.Chem. 75 (6) (2003) 1267–1274.

31] L. Sleno, D.A. Volmer, Ion activation methods for tandem mass spectrometry, J.Mass Spectrom. 39 (10) (2004) 1091–1112.

32] R.A. Zubarev, N.L. Kelleher, F.W. McLafferty, Electron capture dissociation ofmultiply charged protein cations. A nonergodic process, J. Am. Chem. Soc. 120(13) (1998) 3265–3266.

33] J.E.P. Syka, J.J. Coon, M.J. Schroeder, J. Shabanowitz, D.F. Hunt, Peptide and pro-tein sequence analysis by electron transfer dissociation mass spectrometry,Proc. Natl. Acad. Sci. U.S.A. 101 (26) (2004) 9528–9533.

34] A. Stensballe, O.N. Jensen, J.V. Olsen, K.F. Haselmann, R.A. Zubarev, Electroncapture dissociation of singly and multiply phosphorylated peptides, RapidCommun. Mass Spectrom. 14 (19) (2000) 1793–1800.

35] S.D.H. Shi, M.E. Hemling, S.A. Carr, D.M. Horn, I. Lindh, F.W. McLafferty, Phos-phopeptide/phosphoprotein mapping by electron capture dissociation massspectrometry, Anal. Chem. 73 (1) (2001) 19–22.

36] R.A. Zubarev, D.M. Horn, E.K. Fridriksson, N.L. Kelleher, N.A. Kruger, M.A. Lewis,B.K. Carpenter, F.W. McLafferty, Electron capture dissociation for structuralcharacterization of multiply charged protein cations, Anal. Chem. 72 (3) (2000)563–573.

37] M.L. Nielsen, M.M. Savitski, R.A. Zubarev, Improving protein identification usingcomplementary fragmentation techniques in Fourier transform mass spec-trometry, Mol. Cell. Proteomics 4 (6) (2005) 835–845.

38] R.A. Zubarev, A.R. Zubarev, M.M. Savitski, Electron capture/transfer versuscollisionally activated/induced dissociations: solo or duet? J. Am. Soc. MassSpectrom. 19 (6) (2008) 753–761.

39] M.M. Savitski, M.L. Nielsen, F. Kjeldsen, R.A. Zubarev, Proteomics-grade de novosequencing approach, J. Proteome Res. 4 (6) (2005) 2348–2354.

40] R.S. Johnson, S.A. Martin, K. Biemann, J.T. Stults, J.T. Watson, Novel fragmenta-tion process of peptides by collision-induced decomposition in a tandem massspectrometer: differentiation of leucine and isoleucine, Anal. Chem. 59 (21)(1987) 2621–2625.

41] W.M. Hines, A.M. Falick, A.L. Burlingame, B.W. Gibson, Pattern-based algorithmfor peptide sequencing from tandem high-energy collision-induced dissocia-tion mass spectra, J. Am. Soc. Mass Spectrom. 3 (4) (1992) 326–336.

42] R. Kaufmann, D. Kirsch, B. Spengler, Sequencing of peptides in a time-of-flightmass spectrometer – evaluation of post-source decay following matrix-assistedlaser desorption ionisation (MALDI), Int. J. Mass Spectrom. Ion Processes 131(1994) 355–385.

43] R. Kaufmann, B. Spengler, F. Lutzenkirchen, Mass spectrometric sequencing oflinear peptides by product-ion analysis in a reflectron time-of-flight mass spec-trometer using matrix-assisted laser desorption ionization, Rapid Commun.Mass Spectrom. 7 (10) (1993) 902–910.

44] M.S. Thompson, W.D. Cui, J.P. Reilly, Fragmentation of singly charged peptideions by photodissociation at lambda = 157 nm, Angew. Chem. Int. Ed. 43 (36)(2004) 4791–4794.

45] W.D. Cui, M.S. Thompson, J.P. Reilly, Pathways of peptide ion fragmentationinduced by vacuum ultraviolet light, J. Am. Soc. Mass Spectrom. 16 (8) (2005)1384–1398.

46] A. Devakumar, D.K. O’Dell, J.M. Walker, J.P. Reilly, Structural analysis ofleukotriene C-4 isomers using collisional activation and 157 nm photodisso-ciation, J. Am. Soc. Mass Spectrom. 19 (1) (2008) 14–26.

47] A. Devakumar, Y. Mechref, P. Kang, M.V. Novotny, J.P. Reilly, Laser-inducedphotofragmentation of neutral and acidic glycans inside an ion-trap mass spec-trometer, J. Am. Soc. Mass Spectrom. 21 (8) (2007) 1452–1460.

48] L. Zhang, J.P. Reilly, Use of 157-nm photodissociation to probe structures of y-and b-type ions produced in collision-induced dissociation of peptide ions, J.Am. Soc. Mass Spectrom. 19 (5) (2008) 695–702.

49] L. Zhang, J.P. Reilly, Extracting both peptide sequence and glycan structuralinformation by 157 nm photodissociation of N-linked glycopeptides, J. Pro-teome Res. 8 (2) (2008) 734–742.

50] L.Y. Zhang, J.P. Reilly, Peptide photodissociation with 157 nm light in a com-mercial tandem time-of-flight mass spectrometer, Anal. Chem. 81 (18) (2009)7829–7838.

51] L.Y. Zhang, J.P. Reilly, Peptide de novo sequencing using 157 nm photodissoci-ation in a tandem time-of-flight mass spectrometer, Anal. Chem. 82 (3) (2010)898–908.

52] L. Zhang, J.P. Reilly, De novo sequencing of tryptic peptides derived fromdeinococcus radiodurans ribosomal proteins using 157 nm photodissociationMALDI TOF/TOF mass spectrometry, J. Proteome Res. 9 (6) (2010) 3025–3034.

53] B.C. Bohrer, Y.F. Li, J.P. Reilly, D.E. Clemmer, R.D. DiMarchi, P. Radivojac, H. Tang,R.J. Arnold, Combinatorial libraries of synthetic peptides as a model for shotgunproteomics, Anal. Chem. 82 (15) (2010) 6559–6568.

54] M. Lebl, V. Krchnak, Synthetic peptide libraries solid-phase peptide synthesis289 (1997) 336–392.

55] M.M. Savitski, M. Fälth, Y.M.E. Fung, C.M. Adams, R.A. Zubarev, Bifurcatingfragmentation behavior of gas-phase tryptic peptide dications in collisionalactivation, J. Am. Soc. Mass Spectrom. 19 (12) (2008) 1755–1763.

56] B. Godugu, P. Neta, Y. Simón-Manso, S.E. Stein, Effect of N-terminal glutamic

acid and glutamine on fragmentation of peptide ions, J. Am. Soc. Mass Spectrom.21 (7) (2010) 1169–1176.

57] A. Schlosser, W.D. Lehmann, Patchwork peptide sequencing: extraction ofsequence information from accurate mass data of peptide tandem mass spectrarecorded at high resolution, Proteomics 2 (5) (2002) 524–533.

Page 13: International Journal of Mass Spectrometrypredrag/papers/liu_intjmassspectrom_2011.pdf · 144 X. Liu et al. / International Journal of Mass Spectrometry 308 (2011) 142–154 2.4

1 f Mass

[

[

[

54 X. Liu et al. / International Journal o

58] J. Qin, B.T. Chait, Preferential fragmentation of protonated gas-phase peptide

ions adjacent to acidic aminoacid residues, J. Am. Chem. Soc. 117 (19) (1995)5411–5412.

59] G. Tsaprailis, A. Somogyi, E.N. Nikolaev, V.H. Wysocki, Refining the model forselective cleavage at acidic residues in arginine-containing protonated pep-tides, Int. J. Mass Spectrom. 156–157 (2000) 467–479.

[

Spectrometry 308 (2011) 142– 154

60] Y.Y. Huang, J.M. Triscari, L. Pasa-Tolic, G.A. Anderson, M.S. Lipton, R.D. Smith,

V.H. Wysocki, Dissociation behavior of doubly-charged tryptic peptides: corre-lation of gas-phase cleavage abundance with Ramachandran plots, J. Am. Chem.Soc. 126 (10) (2004) 3034–3035.

61] J.P. Reilly, Ultraviolet photofragmentation of biomolecular ions, Mass Spec-trosc. Rev. 28 (3) (2009) 425–447.