N-Terminal amino acid side-chain cleavage of chemically modified peptides in the gas phase: A mass spectrometry technique for N-terminus identification

Article abstract of DOI:10.1016/j.bmc.2006.05.060

Although genome databases have become the key for proteomic analyses, de novo sequencing remains essential for the study of organisms whose genomes have not been completed. In addition, post-translational modifications present a challenge in database sear

Full text of DOI:10.1016/j.bmc.2006.05.060

Bioorganic & Medicinal Chemistry 14 (2006) 6213–6222
N-Terminal amino acid side-chain cleavage of chemically modified
peptides in the gas phase: A mass spectrometry technique for
N-terminus identification
Almary Chacon,a Douglas S. Masterson,b Huiyong Yin,a
Daniel C. Lieblerc and Ned A. Portera,*
aDepartment of Chemistry, Vanderbilt University, Nashville, TN 37235, USA
bDepartment of Chemistry and Biochemistry, University of Southern Mississippi, Hattiesburg, MS 39406, USA
cDepartment of Biochemistry and Mass Spectrometry Research Center, Vanderbilt University, USA
Received 7 February 2006; accepted 30 May 2006
Available online 19 June 2006
Dedicated to a leader in the field of free radical chemistry, Bernd Giese, the 2005 Tetrahedron Prize winner for Creativity in Organic Chemistry.
Abstract—Although genome databases have become the key for proteomic analyses, de novo sequencing remains essential for the study
of organisms whose genomes have not been completed. In addition, post-translational modifications present a challenge in database
searching. Recognition of the b or y-ion series in a peptide MS/MS spectrum as well as identification of the b1- and ynꢀ1-ions can facil-
itate de novo analyses. Therefore, it is valuable to identify either amino-acid terminus. In previous work, we have demonstrated that
peptides modified at the e-amino group of lysine as a t-butyl peroxycarbamate derivative undergo free radical promoted peptide back-
bone fragmentation under low-energy collision-induced dissociation (CID) conditions. Herewe explore the chemistry of the N-terminal
amino group modified as a t-butyl peroxycarbamate. The conversion of N-terminal amines to peroxycarbamates of simple amino acids
and peptides was studied with aryl t-butyl peroxycarbonates. ESI-MS/MS analysis of the peroxycarbamate adducts gave evidence of a
product ion corresponding to the neutral loss of the N-terminal side chain (R), thus identifying this residue. Further fragmentation
(MS3) of product ions formed by N-terminal residue side-chain loss (-R) exhibited an m/z shift of the b-ions equal to the neutral loss
of R, therefore labeling the b-ion series. The study was extended to the analysis of a protein tryptic digest where the SALSA algorithm
was used to identify spectra containing these neutral losses. The method for N-terminus identification presented here has the potential
for improvement of de novo analyses as well as in constraining peptide mass mapping database searches.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
ed from the MS/MS fragmentation data with the help of
sequencing algorithms.3–5 Due to the complexity of pep-
tide MS/MS spectra, a number of methods have been
developed to facilitate interpretation. N-terminal and
C-terminal charged derivatives were developed in the late
1980s to be used with fast atom bombardment (FAB)
analyses in order to simplify fragmentation spectra pro-
duced from high-energy collision-induced dissociation.
In addition, these charged derivatives were useful in
increasing signal intensity of peptide fragment ions in
such experiments.6,7 Keough and co-workers derivatized
the N-termini of peptides with a sulfonic acid group.8,9
Fragmentation of these peptides by collision-induced
dissociation (CID) and postsource decay (PSD) matrix-
assisted laser desorption ionization (MALDI) yielded
selectively the y-ion series. Work by Geskell et al. inves-
tigated N-terminus derivatization using phenylthiocar-
bamoyl derivatives similar to those employed in
The rapid development and improvement of mass spec-
trometry, along with the availability of large genome dat-
abases,
have
made
possible
high-throughput
characterization of complex protein mixtures.1 Never-
theless, the presence of unexpected post-translational
modifications often leads to unsuccessful or incorrect
identifications in database searches. In addition, the
study of proteins derived from organisms whose gen-
omes are not complete is limited.2 In such cases, more la-
bor-intensive data analysis such as de novo sequencing
must be used, where the peptide sequence is reconstruct-
Keywords: Proteomics; Free radical; Peroxycarbonate; Backbone
fragmentation; Gas phase digestion.
*
Corresponding author. Tel.: +1 615 343 2693; fax: +1 615 343
1234; e-mail: n.porter@vanderbilt.edu
0968-0896/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2006.05.060

6214
A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
Edman degradation.10 CID experiments in this case pro-
duced predominantly b1- and ynꢀ1-ions. James and co-
workers carried out the modification of N-termini of
peptides with nicotinyl-N-hydroxysuccinimide which
resulted in suppression of the y-ion series in MALDI tan-
dem time-of-flight (MALDI-TOF-TOF) analyses.11
More recently, Reilly and co-workers investigated the ef-
fect of acetaminidation of N-termini on fragmentation
patterns in CID spectra.12 In most cases studied, there
was an enhancement of ynꢀ1-ions, while diminishing
other fragmentations, and in some cases contiguous ser-
ies of b-NH3 ions were observed. Another approach to
improving de novo methods involves stable isotopic
labeling. Esterification of the C-termini with deuterium
labeled methanol results in a mass shift of the y-ions with
respect to the unlabeled peptide ester thus identifying the
y-ion series.13 In a similar manner, O18 labeling of the C-
termini of peptides provides information about the y-ion
series.14 Cannon and co-workers presented a computa-
tional method to obtain information from the natural
isotopic distributions of peptide fragments in MS/MS
spectra, in order to facilitate identification of b- and y-
ions without prior derivatization.15
C(O)NHCH(R)CO2 R0 [C4H9OOC(O)M] displayed
fragmentation involving the loss of m/z = 116, C4H8OO-
C(O), to give the underivatized amino acid ester
NH2CH(R)CO2R0, M([C4H9OOC(O)M—116]). In
addition, the neutral loss of the t-butyl peroxycarbonate
moiety, C4H9OOC(O), plus the amino acid side chain,
R, yielded the imine NH@CHCOOR0 ([C4H9OO
C(O)M—117 –R]). A sample CID spectrum of modified
valine methyl ester is shown in Figure 1. Lithium was
used as the counter ion for ESI of the modified amino
acids due to the unavailability of the N-terminus for
protonation. Cleavage of the amino acid side chain
was not observed in the CID spectra of the parent amino
acid esters not bearing the peroxycarbamate moiety. In
most cases [C4H9OOC(O)M—116] was the predominant
fragment; however, the imine fragment was present in
most spectra.
In the case of modified proline, the parent adduct ion
was observed but the CID spectrum did not contain
an imine fragment. In CID of the peroxycarbamate-
modified glycine we did not observe an imine fragment
which has a mass difference of 2 with respect to the
unmodified amino acid, instead, we observed a fragment
which matches the loss of only 1 hydrogen with respect
to the unmodified amino acid [C4H9OOC(O)M—117]
accompanied by the usual [C4H9OOC(O)M—116] frag-
ment. Modified asparagines also did not lead to an
imine fragment; however, an imine fragment was ob-
served in the CID spectrum of modified peptides bearing
asparagine at the N-terminus. A list of modified amino
acids and corresponding CID fragmentations are pre-
sented in Table 1.
We report here derivatization of the N-terminal a-amino
group of amino acids and peptides to give a reactive ad-
duct which, under low-energy CID conditions, promotes
cleavage of the N-terminal amino acid side chain
(Scheme 1). This fragmentation provides information
about the nature of the N-terminal residue and may also
serve as a labeling technique for elucidating the b-ion
series in de novo analyses. Thus, single amino acids
and peptides were modified at the N-terminus with aryl
peroxycarbonates, 1a–1c, as shown in Scheme 1. Frag-
mentation of the resulting peroxycarbamates in an ESI
triple quadrupole instrument gave evidence of a product
ion corresponding to the neutral loss of the N-terminal
side chain (R), thus identifying this residue.
2.2. Peptide modifications
Peptide YRVRFLAKENVTQDAEDNC (entry 21 in
Table 2) was modified under different pH conditions,
ranging from 5.2 to 8.6, in order to determine the pH
optimum for N-terminal modification over e-N modifi-
cation of Lys. Previous studies have shown that selectiv-
ity for reactions at the N-terminal a-amino group of
peptides can be achieved by pH control, due to the lower
pKa of the N-terminus compared to the e-amino group
of lysine.18–21 In addition to pH control, a less reactive
reagent than 1a (1b) offers higher selectivity for reaction
2. Results
2.1. Single amino acids and peptides
CID spectra of the lithium adducts of modified amino
acid ester derivatives in the form C4H9OO-
O
O
O
Z
OO
X
O
O
H2N CHC
R
OO
N CHC
+
H
R
Y
1a X = H, Y = H, Z = NO2
1b X = H, Y = NO2, Z = H
1c X = Cl, Y =H, Z = NO2
Fragmentation
by ESI-MS/MS
O
O
HN CHC
O
+
R
OO
N CHC
H
R
Scheme 1. Modification and fragmentation of derivatized N-termini.

A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
6215
[C4H9OOC(O)M – 116]
Li+
O
[C4H9OOC(O)M – 117 – R]
H2N CH
C
O
Li+
O
HN CH
C
O
[C4H9OOC(O)M]
O
Li+
O
O
C
OO
N
H
CH
Figure 1. CID spectrum of N-terminal-modified valine methyl ester.
Table 1. Amino acid adducts and their CID fragmentation
Amino acida
[M+Li+] m/z
[tBuOOC(O)M+Li+] m/z
Neutral lossb
R
H
R m/z
Gly
Ala
Val
Leu
Ile
96.1
110.1
138.0
152.1
152.2
210.0
168.0
194.4
176.1
170.1
186.0
126.0
202.1
225.0
140.2
208.3
181.1
137.0
—
212.0
226.1
254.2
268.2
268.2
326.0
283.9
311.1
291.9
285.9
302.0
242.0
318.0
341.0
256.1
325.1
297.2
251.9
—
117.0
132.0
160.1
174.1
174.2
204.1
190.0
—
1
15
CH3
CH(CH3)2
43
57
CH2CH(CH3)2
CH(CH3)CH2CH3
(CH2)2(CO)OC2H5
CH2(CO)OCH3
CH2(CO)NH2
CH2C3H3N2
(CH2)2SCH3
CH2C6H5
57
Glu
Asp
Asn
His
Met
Phe
Ser
101
73
58
81
198.0
192.1
208.1
148.1
224.1
247.0
162.1
189.0
189.0
—
75
91
CH2OH
CH2C6H5O
CH2C8H6N
31
Tyr
Trp
Thr
Gln
Lys
Proc
Cysd
Arg
107
130
46
CH(OH)CH3
(CH2)2(CO)NH2
(CH2)4NH2
72
72
C3H7
CH2SH
(CH2)3NHC(NH2)2
43
47
—
217
189.0(H+)
305.0
100
a Amino acids are methyl ester derivatives (M = NH2CH(R)CO2CH3) with the exception of Asp and Glu which are the dimethyl ester and diethyl
ester derivatives, respectively, Lys is the ethyl ester and Gln and Asn are the t-butyl ester derivatives.
b Neutral loss corresponds to 117 +R where 117 is the m/z of the C4H9OOC(O) group +1 hydrogen and R is the amino acid side chain.
c Cleavage of the side chain was not observed for Pro, only loss of the C4H9OOC(O) group was seen.
d The MS spectrum of modified Cys was complex and did not contain a peak corresponding to [C4H9OOC(O)M], suggesting decomposition of the
amino acid may occur during the modification reaction. Modified Asn was also an exception where no side-chain loss was observed.
at the N-terminus. LC/MS analysis of the modified pep-
tide revealed that under basic conditions (pH 7.5–8.5)
and with reagent 1a, modification occurs at both amino
group sites, but with greater efficiency on the lysine side
chain. Slightly acidic conditions (pH 5.5) and reagent 1b
gave better selectivity for the N-terminus. Figure 2
shows the LC/MS trace of the modified peptide at pH
8.6 and 5.5. Peak 1 in the chromatogram corresponds
to m/z = 758.2 ([M+3H]3+) where M = YRVRFLA-
C(O) and the N- terminal tyrosine side chain (–CH2–
C6H4OH, m/z = 107). This fragment was absent in the
CID spectrum of peak 3. At lower pH and with the
use of reagent 1b, peak 3 is diminished, which implies
a preference for N-terminal modification under these
conditions.
Several peptides containing different N-terminal amino
acids were modified using reagent 1b at pH 5.5 and
the N-termini were identified in CID experiments (Table
2). Thus, CID spectra of peptide peroxycarbamates
bearing alanines at the N-terminus give product ions
that indicate the loss of CH3 ([C4H9OOC(O)M—117–
15]), whereas terminal valines show a similar loss of
CH3CHCH3. In cases where cysteine and proline were
the terminal residues, the CID spectra did not exhibit
side-chain cleavage product ions. Instead, as seen for
KENVTQDAEDNC. Peaks
2 and 3 indicate a
single modification with m/z = 796.8 ([C4H9OO-
C(O)M+3H]3+). Peak 2 corresponds to modification at
the N-terminal group, whereas peak 3 corresponds to
a-N modification of lysine. This was confirmed by the
CID fragmentation of peak 2, also shown in Fig. 2,
where the ion with m/z = 722.2 ([C4H9OOC(O)M—
117–107+3H]3+) indicates the neutral loss of C4H9OO-

6216
A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
Table 2. Peptides analyzed by CID for N-terminal side-chain loss
Entry
Sequence
Ma m/z
[C4H9 OOC(O)M]b m/z
Imine fragmentc m/z
R
R m/z
1
2
VYIHPF
HPKRPWIL
775.4(1H+)
523.4(2H+)
962.0(1H+)
335.2(3H+)
516.4(2H+)
597.4(2H+)
768.4(1Na+)
761.4(1H+)
882.5(1H+)
518.0(3H+)
468.5(2H+)
630.9(2H+)
795.2(1Na+)
868.6(2H+)
251.1(2H+)
561.9(2H+)
904.3(1H+)
900.3(1H+)
660.1(3H+)
758.2(3H+)
523.5(2H+)
891.5(1H+)
581.8(2H+)
1078.0(1H+)
373.8(3H+)
574.4(2H+)
655.4(2H+)
885.2(1Na+)
877.5(1H+)
998.5(1H+)
556.4(3H+)
526.5(2H+)
689.0(2H+)
911.5(1Na+)
926.8(2H+)
308.9(2H+)
619.8(2H+)
1020.4(1H+)
1016. 5(1H+)
—
731.1(1H+)
483.1(2H+)
887.2(1H+)
324.5(3H+)
486.9(2H+)
551.9(2H+)
679.7(1Na+)
760.3(1H+)
752.1(1H+)
592.6(3H+)
877.7(1H+)
602.0(2H+)
722.0(1Na+)
815.11(2H+)
205.8(2H+)
553.1(2H+)
—
CH(CH3)2
43
81
75
31
58
91
72
1
CH2C3H3N2
(CH2)2SCH3
CH2OH
3
MEHFRW
SRVSRRSR
4
5
NRVYVHPF
FHPKRPWIL
KNPYIL
CH2(CO)NH2
CH2C6H5
6
7
(CH2)4NH2
H
9
GFLRRI
WHWLQL
10
11
12
13
14
15
16
17
18
19
20
21
22
CH2C8H6N
(CH2)2(CO)NH2
CH2CH(CH3)2
CH(CH3)CH2CH3
(CH2)2(CO)OH
CH2C6H4OH
CH(OH)CH3
CH3
130
72
57
57
73
107
46
15
43
100
47
107
59
QRPRLSHKGPMPF
LSRLFDNAl
ISRPPGFSPFR
EDKDLD
YGSLPQKSQRSQDEN
TPRK
APRLRFYSL
PPGFSPFR
RPKPQQF
C3H7
(CH2)3 NHC(NH2)2
CH2SH
801.2(1H+)
—
CNLAVAAASHIYQNQFQ
YRVRFLAKENVTQDEDC
DRVYIHPF
796.8(3H+)
582.1(2H+)
722.2(3H+)
493.8(2H+)
CH2C6H4OH
CH2(CO)OH
a M = m/z of unmodified peptide.
b m/z of peptide modified with reagent 1b at pH 5.5.
c [C4H9OOC(O)M—117 –R] fragment or imine fragment in CID spectrum resulting from the neutral loss of theC4H9OOC(O) group +1 hydrogen
(117) and the N-terminal amino acid side chain (R).
Figure 2. LC/MS of modified peptide YRVRFLAKENVTQDAEDNC. (A) Selected Ion Monitoring (SIM)m/z = 758.2 ([M+3H]3+) where M is the
unmodified peptide YRVRFLAKENVTQDAEDNC. (B) SIM m/z = 796.8 ([C4H9OOC(O)M+3H]3+) where M was modified with reagent 1b at pH
5.5. Peaks 2 and 3 correspond to N-terminus and lysine modifications, respectively. (C) SIM m/z = 796.8 ([C4H9OOC(O)M+3H]3+) where M was
modified with reagent 1a at pH 8.6. (D) CID spectrum of peak 2, parent m/z = 796.8 ([C4H9OOC(O)M+3H]3+). The product ion m/z = 722.2
corresponds to the loss of C4H9OOC(O) and the N-terminal tyrosine side chain, CH2C6H4OH ([C4H9OOC(O)M—117–107+3H]3+).
the single amino acids, modified cysteine gave a complex
MS spectrum and the CID spectrum of modified proline
only showed the loss of the C4H9OOC(O). Nevertheless,
the presence of these amino acids anywhere else in the
sequence did not affect the fragmentation chemistry at
the N-terminus of modified peptides. CID of a modified
peptide bearing glycine at the N-terminus yielded similar
fragmentation to the CID of the modified single amino
acid, where fragment [C4H9OOC(O)M—117] was ob-
served instead of the expected fragment [C4H9OO-
C(O)M—117–1]. Results from the CID of a modified
peptide bearing Asn at the N-terminus (entry 5 in Table
2) indicated loss of the Asn side chain, which is in con-
trast to results obtained with the modified Asn amino
acid derivative.
When a mixture of the two isobaric peptides HPKRP-
WIL and DRVYIHPF (entries 2 and 22 in Table 2)
was modified at pH 5.5 using reagent 1b, the two pep-
tides were distinguishable in LC–MS selected reaction

A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
6217
2.3. Analysis of modified protein tryptic digest
A myoglobin tryptic digest standard was modified using
reagent 1c at pH 5.5. LC–MS/MS analysis was done on
the unmodified as well as the modified digest. Visual
inspection of the MS data identified 12 most abundant
peptides, 7 of which were modified (Table 3). Identifica-
tion of the peptides from MS/MS spectra was confirmed
by database search algorithm SEQUEST22 and the P-
Mod algorithm.23 N-terminal side-chain loss was ob-
served in data-dependent MS2 experiments for entries
2, 4, 5, 7, and 8. MS3 gave a product ion spectrum of
the [C4H9OOC(O)-M—117 –R] fragment for entry 2.
This MS3 spectrum exhibited a labeled b-ion series indi-
cating the loss of the N-terminal valine side chain + 1
hydrogen with respect to the umodified peptide (mass
shift = ꢀ44), while the y-ion series remained unchanged
(Table 4).
The SALSA algorithm17 was used to search through
tandem MS data for spectra indicating the N-terminal
side-chain losses. The SALSA algorithm was designed
to find characteristics in the MS/MS spectra such as
charged and neutral losses, a particular product ion,
or a sequence motif. It can also search for a combination
of two or more of these characteristics and it does not
require a peptide sequence input. SALSA scores do
not indicate the probability of false positives, but in-
stead rank spectra based on intensity and number of
ions found that match a search criteria. We searched
our modified myoglobin digest data for the neutral loss
of 116, which leads to the fragment [C4H9OOC(O)M—
116] or simply [M]. This generated a list of precursor
ions that match the m/z of modified peptides in the di-
gest. To find peptides bearing a certain amino acid at
the N-terminus, we added a secondary search character-
istic corresponding to the neutral loss of 117 + R, which
leads to the side-chain loss fragment [C4H9OOC(O)—
117 –R]. Thus, ions corresponding to (neutral loss of
117 + R) were scored only in CID spectra also contain-
ing the primary loss of 116. This scoring scheme reduced
the number of false positive matches. For example, a
SALSA search for MS/MS spectra of peptides bearing
histidine at the N-terminus scored a primary neutral loss
Figure 3. LC-SRM trace of peptide mixture HPKRPWIL and
DRVYIHPF modified with reagent 1b at pH 5.5. (A) SIMm/z = 582
([C4H9OOC(O)M+2H]2+) where M is either peptide. (B) SRM of m/
z = 582 ([C4H9OOC(O)M+2H]2+)!m/z = 483 ([ C4H9OOC(O) M—
117–81+2H]2+) where M is HPKRPWIL, and the histidine side chain
(CH2-C3H3N2) has m/z = 81.(C) SRM of m/z = 581 ([C4H9OOC
(O)M—117–81+2H]2+) ! m/z = 494
([C4H9OOC(O)M—117–
59+2H]2+) where M is DRVYIHPF and the aspartic acid side chain
(–CH2COOH) has m/z = 59.
monitoring (SRM) and neutral loss experiments by N-
terminal side-chain cleavage (Fig. 3). Thus, the peptide
bearing an N-terminal histidine can be distinguished
from the peptide bearing aspartic acid without further
sequencing. Figure 3(B) shows the SRM of m/z = 582
([C4H9OOC(O)M+2H]2+) ! m/z = 483
([C4H9OO-
C(O)M—117–81+2H]2+) where M is HPKRPWIL and
the side chain of histidine (CH2–C3H3N2) has m/
z = 81. Figure 3(C) shows the SRM of m/z = 582
([C4H9OOC(O)M+2H]2+) ! m/z = 494
([C4H9OO-
C(O)-M—117–59+2H]2+) where M is DRVYIHPF
and the side chain of aspartic acid (–CH2COOH) has
m/z = 59.
Table 3. Peptides detected by LC–MS from a modified myoglobin digest
a Entries in blue are peptides for which modification at the N-terminus is observed.
b m/z of modified peptides ([C4H9OOC(O)M]).
c Mass shift is adjusted for singly charged species.

6218
A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
Table 4. b- and y-ions from MS2 of unmodified peptide VEAKIAGHGQEVLIR compared to b- and y-ions from MS3 of the modified
peptide([C4H9OOC(O)M])
a Parent ion m/z = 804.1(M+2H)2+ where M is the peptide VEAKIAGHGQEVLIR.
b Parent ion m/z = 782.1([C4H9OOC(O)M—117–43+2H]2+), where valine (–CH(CH3)2) has m/z = 43.
c The difference between the b-ions corresponds to the valine side chain +1 hydrogen.
Table 5. First 6 results from SALSA search for peptides with histidine at the N-terminus in a myoglobin digest
a Doubly charged precursor ions which lead to product ions in the MS/MS spectra matching a neutral loss of 116 (loss of C4H9OOC(O)) alone or
accompanied by the neutral loss of 198 (loss of C4H9OOC(O) +1 hydrogen –R of histidine).
b Doubly charged product ions corresponding to the neutral loss of 116 with respect to the precursor ion.
c Doubly charged product ions corresponding to the neutral loss of 198 with respect to the precursor ion.
of 116 linked to a secondary neutral loss of 198. To im-
prove matching accuracy, we also specified doubly
charged ions, since these dominate the spectra.
3. Discussion
The chemistry of peroxycarbamates is based on previ-
ous work from our laboratory on the derivatization of
lysine residues.16,24 In this study, a-N-acetyl lysine
methyl ester was converted at the e-amino group to
its peroxycarbamate derivative by reaction with 4-
nitrophenyl-t-butyl peroxycarbonate (1a). CID experi-
ments on the lysine derivative as the lithium adduct
yielded a unique fragmentation, which included amide
bond cleavage at the C-terminus of lysine. Similar re-
sults were obtained when peptides containing one or
more lysine residues were modified under the same
conditions. The mechanism for this fragmentation is
thought to be initiated by the dissociation of the t-bu-
tyl peroxy group to form an aminyl radical intermedi-
ate, followed by remote hydrogen abstraction and b-
fragmentation of the resulting carbon radicals, see
Scheme 2. Introduction of the labile peroxide bond
provides a low-energy fragmentation pathway to free
radicals that ultimately affect the peptide fragmenta-
tion patterns. Recently, Beauchamp and co-workers25
made use of reagents that modified peptide lysines
with an azo functional group. Decomposition of the
labile azo linkage led to a protocol they dubbed ‘Free
Radical Initiated Peptide Sequencing (FRIPS)’. Again,
The first 6 of 24 results in the SALSA output for this
search are shown in Table 5. Given the scoring scheme
described above, the highest scoring spectra met both
the primary and secondary criteria. The top score
matched
the
m/z
of
modified
peptide
C4H9OOC(O)HGTVVLTALGGILK (entry 2 in Table
5). This result was validated by visual inspection of the
MS/MS data. Other peptides bearing histidine at the
N-terminus (entries 4 and 6) exhibited low scores due
to lower intensity of product ions corresponding to the
scored losses. The peptide in entry 3 does not contain
histidine at the N-terminus; however, this precursor
ion leads to an abundant [C4H9OOC(O)M—116] prod-
uct ion (neutral loss of 116), therefore it scores higher
than entries 4 and 6. It addition, this precursor ion also
leads to a neutral loss of 198, which is unrelated to the
N-terminal side-chain loss; however, this ion is present
in low abundance. Entry 5 has a higher score than entry
6 even though the peptide does not bear histidine at the
N-terminus, because this precursor ion also gives an
abundant [C4H9OOC(O)M—116] product ion; however,
it does not match the neutral loss of 198.

A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
6219
R
HN
R
HN
O
O
O
O
NH
R'
NH
R'
O
O
N
O
OO
N
O
H
H
-CO2
R'
O
R
R
HN
R
HN
HN
HN
O
C
O
O
H2N
NH
R'
NH
R'
NH
R'
NH
R'
H2N
O
HN
O
H2N
O
H2N
O
Scheme 2. Free radical fragmentation of lysine e-amino peroxycarbamates.
the azo linkage provides a low-energy pathway to free
radicals that affect peptide fragmentation patterns.
flate.26 The resulting hydroperoxide was treated with
3-nitrophenyl chloroformate providing 2. The chloro-
formate was prepared from 3-nitrophenol as described
in Section 5.
3.1. N-Terminal amino acid side-chain loss by CID
ESI-MS spectra of N-terminal peroxycarbamate modi-
fied single amino acids and peptides showed a molecular
ion, [C4H9OOC(O)M], where M is the amino acid or
peptide and exhibited a fragment whose m/z is equal
to the m/z of the unmodified species [C4H9OO-
C(O)M—116]. This suggests that gas phase decomposi-
tion of the t-butyl peroxycarbamate group is an
important process. In most cases [C4H9OOC(O)M—
116] is the dominant fragment and it is accompanied
by another fragment corresponding to the neutral loss
of the N-terminal amino acid side chain [C4H9OO-
C(O)M—117 –R]. MS/MS of the [C4H9OOC(O)M—
116] fragment obtained in Q1 (by in-source CID) does
not lead to side-chain fragmentation (data not shown).
Instead, this ion generates fragment ions normally ob-
served for the unmodified amino acid or peptide. This
suggests that the [C4H9OOC(O)M—116] fragment can-
not be an intermediate in the side-chain loss, [C4H9OO-
C(O)M—117 –R]. Two mechanisms must be in play,
with one leading to formation of M from C4H9OO-
C(O)M and the other giving side-chain loss.
Compound 2 was then used to modify the N-terminus of
the peptide WHWLQL (entry 10 in Table 2) using con-
ditions previously described for modifications with the
reagent 1b. The deuterium labeled peroxycarbamate
(C4H9-d9OO(CO)M, 3 in Scheme 4) was subjected to
CID and the resulting product ions included the N-ter-
minal side-chain loss fragment, and a fragment corre-
sponding to M+1, suggesting that deuterium transfer
to the peptide attends peroxycarbamate dissociation.
This indicates that the decomposition of the peroxycar-
bamate 3 follows a pathway in which a hydrogen (deu-
terium) atom is transferred from the t-butyl group to the
a-amino of the N-terminus resulting in a CID fragment
identical to the unmodified peptide (Scheme 4). See sup-
porting information for details in the preparation of 2
and CID spectra of the modified peptide. The mechanis-
tic details of this hydrogen (deuterium) atom transfer re-
main unclear, although an intermediate carbamic acid
could be involved, as could either radical or ion pair
species.
A reasonable mechanism for loss of R, the N-terminal
amino acid side chain, proceeds through the aminyl rad-
ical, which is formed by homolysis of the weak O–O
bond and loss of CO2. Loss of R by b-fragmentation
of the aminyl radical appears to be mechanistically
reasonable.
It is unclear how the peroxycarbamate decomposes by
CID to provide the unmodified peptide or amino acid
and to address this question we prepared compound 2
shown in Scheme 3. We suspect that the t-butyl group
of the peroxycarbamate is a possible source of an H
atom during the CID process leading to the peptide or
amino acid fragment, [C4H9OOC(O)M] ! M. Com-
pound 2 was prepared from the commercially available
2-bromo-2-methylpropane-d9 by reaction with anhy-
drous hydrogen peroxide in the presence of silver tri-
Side-chain loss fragmentation was not observed for
modified Pro and Cys either as single amino acid deriv-
atives or as N-terminal amino acids of peptides. In the
case of Cys the peroxycarbamate may not form or
O2N
CD3
OOH
CD3
O2N
O
O
O
CD3
Br
CD3
OCOCl
H2O2
D3C
D3C
O
AgOTf
D3C
2
CD3
D3C
Scheme 3. Preparation of a deuterated peroxycarbonate reagent.

6220
A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
O
O
O
O
CID
D3C
D3C
HN
HN
O O
NH
Peptide
N
H
Peptide
N
H
Peptide
N
H
3
- R
R
CD3
R
(Side Chain Loss Fragment)
(Aminyl Radical Intermediate)
CID
O
DHN
Peptide
N
H
R
(D-Labeled Native Peptide)
Scheme 4. Proposed mechanism for dissociation of peroxycarbamate.
may not be stable due to the reactive thiol group nearby.
The observation that methionine does not present this
problem supports the hypothesis of the negative effect
of the free thiol group. In most cases in ‘real’ proteomics
analyses, Cys residues will be alkylated (e.g., with iodo-
acetamide) to produce S-carboxamido derivatives.
These, similarly to methionine, will likely be better be-
haved than a free Cys. In the case of modified Pro, the
N-terminus is part of a five-membered ring structure,
which would prevent side-chain loss by the mechanism
shown in Scheme 4. The presence of Cys or Pro any-
where else in the peptide sequence, however, did not af-
fect the formation of the N-terminal peroxycarbamates
or N-terminal side-chain cleavage. Modified Gly did
not appear to yield N-terminal loss of the side chain
hydrogen by CID, but instead, a fragment consistent
with the m/z of the aminyl radical (Scheme 4) was ob-
served. It is unclear why modified Asn does not exhibit
side-chain loss; however it is evident that this is not a
problem when Asn is at the N-terminus of a peptide.
concentration of peptides in the digest modification
reaction (lM) compared to the peptide concentrations
in the modification reactions shown in Table 2 (mM)
would slow the reaction rate so that a more reactive re-
agent is needed.
3.3. Applications of N-terminal amino acid side-chain
fragmentation
Results shown in Fig. 3 illustrate the use of N-terminal
side-chain loss as a quick diagnostic peptide identifica-
tion method without the need for a complete sequencing
experiment. These results lead the way to the use of this
technique in peptide mass mapping studies, where N-ter-
minus identification can be used to eliminate false posi-
tive matches in database searching.28 Because MS/MS
spectra of unmodified peptides lack information about
the N-terminus (b1 and ynꢀ1 ions),29 identification of
N-terminal residues as described here would aid in
obtaining complete sequences in de novo analyses. In
addition, results in Table 3 demonstrate that gas phase
cleavage of the N-terminal side chain by CID labels the
N-terminus so that a subsequent CID process (MS3)
leads to a mass shift in the b-ions when compared to
the b-ions from the intact peptide. This mass shift corre-
sponds to the m/z of the N-terminal side chain, and
serves to elucidate the b-ion series in de novo sequencing.
3.2. Peptide modifications
In order to selectively modify single peptides at the N-
terminus, pH and reagent for the modification reaction
were altered from the conditions used previously for
modification of lysine side chains.16 Reagent 1b is less
reactive towards nucleophilic attack by the amine nitro-
gen than 1a, therefore modification occurs at a slower
rate with this reagent. The reaction can be accelerated
by increasing the temperature to 37 ꢁC, conditions under
which reagent 1a undergoes decomposition. Lowering
the pH below the pKa of the terminal a-amino group
(approximate pKa = 8) increases the selectivity for reac-
tion at the N-terminus with respect to the e-amino group
of lysine (approximate pKa = 10.7).18 Results shown in
Fig. 2 show that at pH 5.5 modification with reagent
1b occurs mostly at the N-terminus. When modifying
the myoglobin tryptic digest, reagent 1c appeared to
provide better conversion to peroxycarbamates than
either 1a or 1b. Reagent 1c is more reactive than 1a
and 1b, which would suggest that this reagent would
not be as selective for the N-terminus as reagent 1b;
however, both 1a and 1b required longer reaction times
under conditions necessary for selective N-terminus
modification (over 4 h). The modification reactions de-
scribed here are pseudo 1st order with rate = kobs[pep-
tide] due to the large excess reagent used.27 The lower
The analysis of a myoglobin tryptic digest demonstrates
that this N-terminus identification method is feasible in
a proteomics setting. The SALSA algorithm finds
charged and neutral losses, a particular product ion,
or a sequence motif and it does not require a peptide se-
quence input. This analysis identifies MS/MS spectra of
possible peptides of interest, which can then be con-
firmed by visual inspection.
4. Conclusion
The work presented here offers an approach to im-
prove de novo sequencing that relies on the gas-phase
chemistry of chemically modified peptides. We empha-
size that relatively inexpensive instrumentation such as
the ESI-triple quadrupole mass spectrometer used here
can be sufficient to carry out N-terminus identification
analyses. The ability to distinguish two isobaric pep-
tides in a mixture by the N-termini without a com-

A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
6221
plete sequence has the potential to be applied in pep-
5.2. Modification reagents
tide mass mapping analyses. Studies on the N-terminal
amino acid side-chain fragmentation induced by
MALDI are currently underway and will be presented
in the near future.
Compounds 1a–1c (Scheme 1) were synthesized accord-
ing to previous work in our laboratory16 starting from 4-
nitrophenylchloroformate, 3-nitrophenylchloroformate,
and 2-chloro-4-nitrophenylchloroformate, respectively.
The two chloroformates preceding 1b and 1c were pre-
pared from their respective phenols by treatment with
one equivalent of triethylamine followed by two equiva-
lents of phosgene in methylene chloride at 0 ꢁC under
5. Experimental
5.1. Materials
1
nitrogen. All compounds were characterized by H and
Amino acids were purchased from Aldrich Chemical
Co, Inc. (Milwaukee, MA) as either methyl, ethyl or
t-butyl ester derivatives in their hydrochloride salt
forms. Peptides were purchased from American Pep-
tide Company (Sunnyvale, CA). Solvents were of
HPLC grade, obtained from Fisher Scientific (Pitts-
burgh, PA). All other reagents were purchased from
Aldrich Chemical Co, Inc. (Milwaukee, MA). Acetate
buffers (10 mM) of various pH values were prepared
from ammonium acetate titrated to pH 5.2, 5.5, and
5.8 with acetic acid. Tris buffers were 0.1 M solutions
titrated to pH 7.5 and pH 8, respectively. Ammonium
bicarbonate buffer was a 0.1 M solution at pH 8.6.
Buffer systems used to modify peptides and amino
acids consisted of a 1:1 mixture of the appropriate
buffer and acetonitrile. The myoglobin tryptic digest
standard mixture (500 pmol) was purchased from All-
tech-Life Sciences (Deerfield, IL).
13C NMR (see supporting information).
5.3. Modification reactions
Reactions with all 20 amino acids as methyl, ethyl, or t-
butyl ester derivatives were carried out by combining
5 ll of amino acid solution (0.1 M in 50/50 acetonitrile
and water), 20 ll of ammonium bicarbonate buffer sys-
tem (50/50 acetonitrile/buffer), and 20 ll of 1a (0.1 M
in anhydrous acetonitrile). Final concentrations for ami-
no acids and reagent were 0.01 M, and 0.04 M, respec-
tively. Reactions were incubated at room temperature
overnight. The crude reaction mixtures were diluted to
1 ml with a 0.4 mM solution of LiCl in methanol for
MS analysis.
The peptide YRVRFLAKENVTQDAEDNC (entry 21
in Table 2) was modified at pH range 5.2–8.6 by addi-
tion of 10 ll of buffer system (50/50 acetonitrile/buffer)
to 10 ll of peptide solution (1 mg/ml in water), followed
by 5 ll of 1a or 1b (0.1 M in acetonitrile). Final concen-
trations of peptide and reagent were 0.7 lM and 33 lM,
respectively. The reaction mixture was incubated at
37 ꢁC for 1 h. The crude mixtures were analyzed by re-
verse phase LC–MS to determine selectivity for N-termi-
nus modification. A series of peptides bearing each of
the 20 amino acids at the N-terminus (Table 2) were
modified using the conditions which gave best selectivity
for N-terminus modification (pH 5.5, buffer system, re-
agent 1b). All peptide reaction mixtures were diluted
to 1 ml with a 0.1% acetic acid solution and analyzed
by MS/MS as direct liquid infusions.
Mass spectrometry analyses of modified single amino
acid derivatives and peptides were performed on a
ThermoFinnigan (San Jose, CA) TSQ 7000 triple-
quadrupole instrument, equipped with an Electrospray
Ionization (ESI) source and a Waters (Milford, MA)
Alliance HPLC system. The capillary temperature
was kept at 200 ꢁC with a voltage of 20 V. The elec-
trospray needle voltage was 4.5 kV, and the tube-lens
voltage was maintained between 70 and 100 V. The
sheath and auxiliary gases (N2) were adjusted to max-
imize signal. The offset voltage in CID experiments
varied between 20 and 35 eV. Samples were intro-
duced into the ESI source as either lithium solutions
or acidic solutions at a rate of 10–20 ll/min in direct
liquid infusion experiments. For LC–MS analyses of
peptides, a Supelco (Bellefonte, PA) Discovery C18
column was used (25 cm · 4.6 mm, 5 lm). The solvent
conditions consisted of a 20 min gradient from 5% to
50% solvent B (95% acetonitrile, 5% water, and 0.05%
TFA) into solvent A (95% water, 5% acetonitrile, and
0.05% TFA). LC–MS/MS analyses of peptide deriva-
tives were performed on a ThermoFinnigan (San Jose,
CA) LCQ Deca XP ion trap instrument, equipped
with an ESI source, and a HP Agilent (Palo Alto,
CA) HPLC system. Instrument parameters were tuned
using angiotensin II peptide (5 pmol) to maximize sig-
nal. LC–MS/MS analyses of the myoglobin digests
were performed using a Grace Vydac (Hesperia, CA)
C18 column (10 cm · 1 mm · 5 lm). Solvent condi-
tions consisted of a 30 min gradient from 100% sol-
vent A (5% acetonitrile, 95% water, and 0.5% acetic
acid) to 100% solvent B (5% water, 95% acetonitrile,
and 0.5% acetic acid).
5.4. Analysis of modified myoglobin tryptic digest
Modification of a myoglobin tryptic digest standard was
achieved by combining 50 ll of digest solution
(500 pmol in 200 ll of water), 50 ll of acetate buffer sys-
tem (pH 5.5), and 25 ll of reagent 1c (0.1 M in acetoni-
trile). The reaction mixture was incubated at room
temperature for 2 h and analyzed by reverse phase
LC–MS. MS/MS data were analyzed with the SALSA
algorithm17 to search for modified peptides and neutral
losses of N-terminal amino acid side chains.
Acknowledgments
We thank the National Science Foundation, the Van-
derbilt Center for Research in Toxicology, NIEHS P01
ES013125 and Vanderbilt University for financial sup-
port. We also thank Vanderbilt’s Mass Spectrometry

6222
A. Chacon et al. / Bioorg. Med. Chem. 14 (2006) 6213–6222
12. Beardsley, R. L.; Reilly, J. P. J. Am. Soc. Mass Spectrom.
2004, 14, 158–167.
13. Goodlett, D. R.; Keller, A.; Watts, J. D.; Newitt, R.; Yi,
E. C.; Purvine, S.; Eng, J. K.; Haller, P.; Aebersold, R.;
Kolker, E. Rapid Commun. Mass Spectrom. 2001, 15,
1221–1224.
Research Center, in particular Dr. David Hachey, Lisa
Manier, and Dawn Overstreet, for advice in use of
instrumentation. In addition, we thank the Proteomics
Research Facility, in particular Dr. Amy Ham and Dr.
David Freeman, for their advice with the myoglobin
tryptic digest analysis.
14. Quin, J.; Herring, C. J.; Zhang, X. Rapid Commun. Mass
Spectrom. 1998, 12, 209–216.
15. Cannon, R.; Jarman, K. D. Rapid Commun. Mass
Spectrom. 2003, 17, 1793–1801.
Supplementary data
16. Masterson, D. S.; Yin, H.; Chacon, A.; Hachey, D. L.;
Norris, J. L.; Porter, N. A. J. Am. Chem. Soc. 2003, 126,
720–721.
17. Hansen, B. T.; Jones, J. A.; Mason, D. E.; Liebler, D. C.
Anal. Chem. 2001, 73, 1676–1683.
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2006.05.060.
18. Stark, G. R. Biochemistry 1965, 4, 1030–1036.
19. Desbuquois, B. Eur. J. Biochem. 1975, 60, 335–347.
20. Bradbury, H. J.; Howell, J. R.; Johnson, R. N.; Warren,
B. Eur. J. Biochem. 1978, 84, 503–511.
21. Wetzel, R.; Hlualani, R.; Stults, J. T.; Quan, C. Biocon-
jugate Chem. 1990, 1, 114–122.
22. Yates, J. R.; Eng, J. k.; McCormack, A. L. Anal. Chem.
1995, 67, 3202–3210.
23. Hansen, B. T.; Davey, S. W.; Ham, A. L.; Liebler, D. C. J.
Proteome Res. 2005, 4, 358–368.
24. Porter, N. A.; Masterson, D. S.; Yin, H.; Caprioli, R. M.;
Norris, J. L.; Chacon, A. Peptide and Protein Fragmen-
tation by Lysine Residue Originated Reactions.
WO2005037857, 2005.
25. Hodyss, R.; Cox, H. A.; Beauchamp, J. L. J. Am. Chem.
Soc. 2005, 127, 12436–12437.
References and notes
1. Liebler, Daniel C. Introduction to Proteomics: Tools for the
New Biology; Humana Press: Totowa, NJ, 2002, p. 198.
2. Reinders, J.; Lewandrowski, U.; Moebius, J.; Wagner, Y.;
Sickmann, A. Proteomics 2004, 4, 3686–3703.
3. Taylor, J. A.; Johnson, R. S. Anal. Chem. 2001, 73, 2594–
2604.
4. Ma, B.; Zhang, K.; Lajoie, G.; Doherty-Kirby, A.;
Hendrie, C.; Liang, C.; Li, M. Rapid Commun. Mass
Spectrom. 2003, 17, 2337–2342.
5. Frank, A.; Pevzner, P. Anal. Chem. 2005, 77, 964–973.
6. Roth, K. D. W.; Huang, Z. H.; Sadagopan, N.; Watson, J.
T. Mass Spectrom. Rev. 1998, 17, 255–274.
7. Barofsky, D. F.; Chen, T. F. Peptide Sequencing Promoted
by N-Terminal Derivatization. In Mass Spectrometry and
Hyphenated Techniques in Neuropeptide Research; Silberr-
ing, J., Ekman, R., Eds.; John Wiley, 2002; pp 345–373.
8. Keough, T.; Youngquist, R. S.; Lacey, M. P. Proc. Natl.
Acad. Sci. U.S.A. 1999, 96, 7131–7136.
9. Keough, T.; Youngquist, R. S.; Lacey, M. P. Anal. Chem.
2003, 75, 156A–165A.
10. Summerfield, S. G.; Steen, H.; O’Malley, M.; Gaskell, S. J.
Int. J. Mass Spectrom. 1999, 188, 95–103.
11. Munchbach, M.; Quadroni, M.; Giovanni, M.; James, P.
¨
26. Milas, N. A.; Harris, A. S. J. Am. Chem. Soc. 1938, 60,
2434–2436.
27. Thomas, H. L.; Richardson, K. S. Mechanism and Theory
in Organic Chemistry, 3rd ed.; Harper Collins Publishers:
New York, 1987, p. 1090.
28. Brancia, F. L.; Butt, A.; Beynon, R. J.; Hubbard, S. J.;
Gaskell, S. J.; Oliver, S. G. Electrophoresis 2001, 22, 552–
559.
29. Reid, E. G.; Tichy, S. E.; Perez, J.; O’Hair, R. A. J.;
Simpson, R. J.; Kenttamaa, H. I. J. Am. Chem. Soc. 2001,
123, 1184–1192.
Anal. Chem. 2000, 72, 4047–4057.