Coumarin tags for analysis of peptides by MALDI-TOF MS and MS/MS. 2. Alexa fluor 350 tag for increased peptide and protein identification by LC-MALDI-TOF/TOF MS

Article abstract of DOI:10.1021/ac048375g

The goal of this study was the development of N-terminal tags to improve peptide identification using high-throughput MALDI-TOF/TOF MS. Part 1 of the study was focused on the influence of derivatization on the intensities of MALDI-TOF MS signals of peptides. In part 2, various derivatization approaches for the improvement of peptide fragmentation efficiency in MALDI-TOF/TOF MS are explored. We demonstrate that permanent cation tags, while significantly improving signal intensity in the MS mode, lead to severe suppression of MS/MS fragmentation, making these tags unsuitable for high-throughput MALDI-TOF/TOF MS analysis. In the present work, it was found that labeling with Alexa Fluor 350, a coumarin tag containing a sulfo group, along with guanidation of ε-amino groups of Lys, could enhance unimolecular fragmentation of peptides with the formation of a high-intensity y-ion series, while the peptide intensities in the MS mode were not severely affected. LC-MALDI-TOF/TOF MS analysis of tryptic peptides from the SCX fractions of an E. coli lysate revealed improved peptide scores, a doubling of the total number of peptides, and a 30% increase in the number of proteins identified, as a result of labeling. Furthermore, by combining the data from native and labeled samples, confidence in correct identification was increased, as many proteins were identified by different peptides in the native and labeled data sets. Additionally, derivatization was found not to impair chromatographic behavior of peptides. All these factors suggest that labeling with Alexa Fluor 350 is a promising approach to the high-throughput LC-MALDI-TOF/TOF MS analysis of proteomic samples.

Full text of DOI:10.1021/ac048375g

Anal. Chem. 2005, 77, 2085-2096
Coumarin Tags for Analysis of Peptides by
MALDI-TOF MS and MS/MS. 2. Alexa Fluor 350 Tag
for Increased Peptide and Protein Identification by
LC-MALDI-TOF/TOF MS
Anna Pashkova, Hsuan-Shen Chen, Tomas Rejtar, Xin Zang, Roger Giese, Victor Andreev,
Eugene Moskovets, and Barry L. Karger*
Barnett Institute and Department of Chemistry, Northeastern University, 341 Mugar, 360 Huntington Avenue,
Boston, Massachusetts 02115
in the MS mode of MALDI-TOF MS has been examined.4
However, with constantly growing protein databases,5,6 and with
recent improvements in MS/MS instrumentation7-9 and database
searching algorithms,10-12 confident peptide identification is now
based on MS/MS fragmentation, in addition to exact molecular
masses of peptides measured in the MS mode.13-15 For this reason,
new peptide derivatization reagents are designed, among other
purposes, for improved MALDI MS/MS fragmentation.16-20 The
goal of the present study was to develop a derivatization approach
that would enhance peptide identification in high-throughput LC-
MALDI-TOF/TOF MS, while enhancing or maintaining signal
intensity in the MS mode. In the future, isotopic labels could be
incorporated in the developed tags, to allow differential quantita-
tion of the proteins expressed in various cell states.
The goal of this study was the development of N-terminal
tags to improve peptide identification using high-through-
put MALDI-TOF/TOF MS. Part 1 of the study was focused
on the influence of derivatization on the intensities of
MALDI-TOF MS signals of peptides. In part 2, various
derivatization approaches for the improvement of peptide
fragmentation efficiency in MALDI-TOF/TOF MS are
explored. We demonstrate that permanent cation tags,
while significantly improving signal intensity in the MS
mode, lead to severe suppression of MS/MS fragmenta-
tion, making these tags unsuitable for high-throughput
MALDI-TOF/TOF MS analysis. In the present work, it was
found that labeling with Alexa Fluor 350, a coumarin tag
containing a sulfo group, along with guanidation of E-ami-
no groups of Lys, could enhance unimolecular fragmenta-
tion of peptides with the formation of a high-intensity y-ion
series, while the peptide intensities in the MS mode were
not severely affected. LC-MALDI-TOF/TOF MS analysis
of tryptic peptides from the SCX fractions of an E. coli
lysate revealed improved peptide scores, a doubling of the
total number of peptides, and a 30% increase in the
number of proteins identified, as a result of labeling.
Furthermore, by combining the data from native and
labeled samples, confidence in correct identification was
increased, as many proteins were identified by different
peptides in the native and labeled data sets. Additionally,
derivatization was found not to impair chromatographic
behavior of peptides. All these factors suggest that labeling
with Alexa Fluor 350 is a promising approach to the high-
throughput LC-MALDI-TOF/TOF MS analysis of pro-
teomic samples.
(3) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R.
Nat. Biotechnol. 1999, 17, 994-9.
(4) Pashkova, A.; Moskovets, E.; Karger, B. L. Anal. Chem. 2004, 76, 4550-
4557.
(5) SwissProt/TrEMBL, http://us.expasy.org/sprot/.
(6) NCBInr, http://www.ncbi.nih.gov/Entrez/.
(7) Medzihradszky, K. F.; Campbell, J. M.; Baldwin, M. A.; Falick, A. M.; Juhasz,
P.; Vestal, M. L.; Burlingame, A. L. Anal. Chem. 2000, 72, 552-558.
(8) Schwartz, J. C.; Senko, M. W.; Syka, J. E. P. J. Am. Soc. Mass Spectrom.
2002, 13, 659-669.
(9) Hager, J. W. Rapid Commun. Mass Spectrom. 2002, 16, 512-526.
(10) Perkins, D. N.; Pappin, D. J. C.; Creasy, D. M.; Cottrell, J. S. Electrophoresis
1999, 20, 3551-3567.
(11) Yates, J. R.; Eng, J. K.; McCormack, A. L.; Schieltz, D. Anal. Chem. 1995,
67, 1426-1436.
(12) Zhang, W.; Chait, B. T. Anal. Chem. 2000, 72, 2482-2489.
(13) Smith, R. D.; Anderson, G. A.; Lipton, M. S.; Pasa-Tolic, L.; Shen, Y.; Conrads,
T. P.; Veenstra, T. D.; Udseth, H. R. Proteomics 2002, 2, 513-523.
(14) Clauser, K. R.; Baker, P.; Burlingame, A. L. Anal. Chem. 1999, 71, 2871-
2882.
(15) Zubarev, R. A.; Hakansson, P.; Sundqvist, B. Anal. Chem. 1996, 68, 4060-
4063.
(16) Roth, K. D. W.; Huang, Z.-H.; Sadagopan, N.; Watson, J. T. Mass Spectrom.
Various derivatization approaches to increase the information
content from a sample, as well as for differential quantitation, are
widely used in proteome studies.1-3 In part 1 of the present study,
the importance of derivatization agents to improve peptide signals
Rev. 1998, 17, 255-274.
(17) Keough, T.; Lacey, M. P.; Youngquist, R. S. Rapid Commun. Mass Spectrom.
2002, 16, 1003-1015.
(18) Peters, E. C.; Horn, D. M.; Tully, D. C.; Brock, A. Rapid Commun. Mass
Spectrom. 2001, 15, 2387-92.
(19) Pappin, D. J.; Bartlet-Jones, M. Proceedings of the 50th ASMS Conference on
Mass Spectrometry and Allied Topics, Orlando, FL, June 2-6, 2002.
(20) Ross, P.; Huang, Y.; Marchese, J.; Khainovski, N.; Williamson, B.; Hattan,
S.; Juhasz, P.; Daniels, S.; Pillai, S.; Purkayastha, S.; Martin, S.; Pappin, D.
Proceedings of the 52th ASMS conference on Mass Spectrometry and Allied
Topics, Nashville, TN, May 23-27, 2004.
* Corresponding author. Phone: (617) 373-2867. Fax: (617) 373-2855.
E-mail: b.karger@neu.edu.
(1) Tao, W. A.; Aebersold, R. Curr. Opin. Biotechnol. 2003, 14, 110-118.
(2) Keough, T.; Youngquist, R. S.; Lacey, M. P. Anal. Chem. 2003, 75, 156A-
165A.
10.1021/ac048375g CCC: $30.25 © 2005 American Chemical Society
Published on Web 02/25/2005
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2085

Improvement in the fragmentation efficiency of peptides is of
particular importance for MALDI-generated ions, because the
predominant singly charged ions in MALDI generally fragment
worse than doubly charged ions.21 Compared to the MS/MS
spectra of doubly charged ESI-generated ions, MS/MS spectra
of singly charged MALDI ions contain ions from various fragment
ion series,22 resulting in the dissipation of MS/MS ion current
and low S/N of individual fragment ion peaks.23 This low S/N
will cause poor MS/MS spectra, leading to low levels of identifica-
tion and false positives.
Several derivatization approaches for improved fragmentation
of singly charged ions have been presented.2,16 The most com-
monly used derivatization agents have been cationic or strongly
basic moieties, as they were reported to improve peptide detection
sensitivity,16,24 and to simplify high-energy collision-induced dis-
sociation (CID) MS/MS and postsource decay (PSD)25,26 spectra
of peptides, leading to preferential formation of a- and d-ion
series.27,28
Recently introduced MALDI-TOF/TOF mass spectrometers,7,29
which employ both unimolecular (PSD) and high-energy CID
fragmentation,30 are well suited for high-throughput analysis of
complex peptide mixtures. It has been demonstrated that MALDI-
TOF/TOF CID spectra are similar to MALDI-CID and liquid
secondary ion-CID spectra obtained earlier with tandem double
focusing EBE-TOF mass spectrometers.31 As a consequence, the
cation tag labeling approach for the improvement of fragmentation
efficiency of peptides was expected to be applicable for MALDI-
TOF/TOF MS as well. On the other hand, there have been
suggestions that the use of cationic tags for MALDI-TOF MS/
MS may not be effective, but no detailed investigation has been
published.2,32
Figure 1. Structures of N-terminal peptide tags used in the study.
led to an increase in MASCOT scores and improved de novo
sequencing analysis.35,36 Recently our laboratory reported on a
similar fragmentation pattern found for N-terminal coumarin tags.37
Another advantage of the coumarin tags is that they enhance
intensities of peptide signals in the MS mode via improved
incorporation in hydrophobic MALDI matrixes during cocrystal-
lization.4
Finally, N-terminal sulfonic acid derivatives were advanced
recently for peptide sequencing by MALDI-TOF MS.2,38-40 Despite
a decrease in MALDI MS signal intensity in the positive ion mode,
sulfonation was shown to be beneficial for MS/MS unimolecular
fragmentation (PSD). This is due to the fact that the sulfo group
facilitates MS/MS fragmentation of singly charged peptide ions
by providing an additional “mobile” proton, the acidic proton from
the sulfo group, which lowers amide bond strength, allowing facile
unimolecular decay.2,21 Up to now, N-terminal sulfonation has been
primarily used in the analysis of in-gel digest samples and de novo
sequencing. Only one attempt to use a sulfo tag, 4-sulfophenyl
isothiocyanate (SPITC), with LC separation has been published,
with MS/MS analysis conducted by ESI-ion trap MS.41
In this paper, part 2, following the previously published part
1,4 we have utilized a combination of the coumarin structure with
a sulfo group, commercially available Alexa Fluor 350 dye, as a
new N-terminal derivatizing agent for LC-MALDI-TOF/TOF MS
analysis (see Figure 1). Additionally, we guanidated ꢀ-amino
groups of lysine, to create a high proton affinity site at C-terminus,
as well as protect lysines from being labeled by a second sulfo
tag.40 To demonstrate the advantages of the use of Alexa Fluor
350 labeling for the analysis of complex proteomic samples, a
strong cation exchange (SCX) LC fraction of a tryptic digest of
Escherichia coli lysate was divided into two parts; one was
Another derivatization approach that offers advantages for
peptide analysis in both the MS and MS/MS modes is N-teminal
labeling with weakly basic tags, which have high gas-phase proton
affinity, e.g., pyridine, morpholine, and piperazine derivatives.19,33,34
Labeling of peptides with N-succimidyl-3-morpholine acetate was
shown to improve coverage of sequence-informative ions in
MALDI-TOF/TOF MS spectra, in particular, b-ion series, which
(21) Dongre, A. R.; Jones, J. L.; Somogyi, A.; Wysocki, V. H. J. Am. Chem. Soc.
1996, 118, 8365-8374.
(22) Rouse, J. C.; Wen, Y.; Martin, S. A. J. Am. Soc. Mass Spectrom. 1995, 6,
822-835.
(23) Wattenberg, A.; Organ, A. J.; Schneider, K.; Tyldesley, R.; Bordoli, R.;
Bateman, R. H. J. Am. Soc. Mass Spectrom. 2002, 13, 772-783.
(24) Liao, P. C.; Allison, J. J. Mass Spectrom. 1995, 511-512.
(25) Spengler, B.; Luetzenkirchen, F.; Metzger, S.; Chaurand, P.; Kaufmann, R.;
Jeffery, W.; Bartlet-Jones, M.; Pappin, D. J. C. Int. J. Mass Spectrom. Ion
Processes 1997, 169/170, 127-140.
(26) Huang, Z. H.; Shen, T.; Wu, J.; Gage, D. A.; Watson, J. T. Anal. Biochem.
1999, 268, 305-317.
(35) Ross, P. L.; Huang, Y.; Pillai, S.; Purkayastha, B.; Smirnov, I.; Pappin, D.
Proceedings of the 51th ASMS Conference on Mass Spectrometry and Allied
Topics, Montreal, Canada, June 8-12, 2003.
(36) Summerfield, S. G.; Bolgar, M. S.; Gaskell, S. J. J. Mass Spectrom. 1997,
32, 225-231.
(37) Pashkova, A.; Zhang, X.; Giese, R.; Karger, B. L. Proceedings of the 51th
ASMS conference on Mass Spectrometry and Allied Topics, Montreal, Canada,
June 8-12, 2003.
(38) Burlet, O.; Yang, C.-Y.; Gaskell, S. J. J. Am. Soc. Mass Spectrom. 1992, 3,
337-344.
(27) Zaia, J.; Biemann, K. Anal. Chem. 1995, 6, 428-436.
(28) Huang, Z.-H.; Wu, J.; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T.
Anal. Chem. 1997, 69, 137-144.
(29) Yergey, A. L.; Coorssen, J. R.; Backlund, P. S.; Blank, P. S.; Humphrey, G.
A.; Zimmerberg, J.; Campbell, J. M.; Vestal, M. L. J. Am. Soc. Mass Spectrom.
2002, 13, 784-791.
(30) Campbell, J. M. Proceedings of the 51th ASMS Conference on Mass
Spectrometry and Allied Topics, Montreal, Canada, June 8-12, 2003.
(31) Medzihradszky, K. F.; Adams, G. W.; Burlingame, A. L.; Bateman, R. H.;
Green, M. R. J. Am. Soc. Mass Spectrom. 1996, 7, 1-10.
(32) Darryl Pappin, personal communication.
(39) Keough, T.; Youngquist, R. S.; Lacey, M. P. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 7131-6.
(33) Cardenas, M. S.; Heeft, E. v. d.; Jong, A. P. J. M. d. Rapid Commun. Mass
Spectrom. 1997, 11, 1271-1278.
(40) Keough, T.; Lacey, M. P.; Youngquist, R. S. Rapid Commun. Mass Spectrom.
2000, 14, 2348-2356.
(34) Munchbach, M.; Quadroni, M.; Miotto, G.; James, P. Anal. Chem. 2000,
72, 4047-4057.
(41) Lee, Y. H.; Kim, M.-S.; Choie, W.-S.; Min, H.-K.; Lee, S.-W. Proteomics 2004,
4, 1684-1694.
2086 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

maintained native and the other modified by guanidation of lysines,
followed by N-terminal labeling with Alexa Fluor 350. Both
fractions were then analyzed in parallel by high-performance
reversed-phase nanoLC-MALDI-TOF/TOF MS using a home-built
multiplexed LC-MALDI deposition device.42,43 All studies were
performed using an AB 4700 MALDI TOF/TOF mass spectrom-
eter in both PSD (unimolecular fragmentation) and CID modes.
Significant improvements in the number and confidence of peptide
and protein identifications as a result of labeling were found.
followed by digestion with trypsin at 37 °C overnight. The resulting
digest was concentrated and cleaned with 1-mL Bond Elut C18
cartridges (Varian, Palo-Alto, CA).
Strong Cation Exchange Fractionation. SCX was conducted
on a Series II 1090 liquid chromatograph (Agilent Technologies,
Palo Alto, CA). E. coli lysate digest sample (500 µL, ∼1 mg of
total protein) was injected into a polysulfoethyl A column (2.1 mm
× 15 cm, PolyLC, Columbia, MD) and separated with a 1-h linear
gradient at a flow rate of 200 µL/min. The composition of solvent
A was 25% (v/v) ACN, 10 mM phosphate buffer at pH 3.0, and
solvent B was 25% (v/v) ACN, 10 mM phosphate buffer at pH 3.0
with 350 mM KCl. The separation was monitored using a Waters
486 UV detector (Waters Corp., Milford, MA). Fractions of ion
exchange eluent, collected in 6-min time intervals (1200 µL each),
were concentrated by a SpeedVac (Thermo Savant, Holbrook, NY)
to a final volume of 100 µL and frozen at -80 °C. Several SCX
fractions were selected for labeling and analysis by reversed-phase
LC-MALDI-TOF and TOF/TOF MS.
Guanidation of E-Amino Groups of Lysine Residues. The
guanidation reaction was performed with a slight modification of
a protocol described elsewhere.40 Fresh solution of O-methyl-
isourea hemisulfate, 13.2 mg in 50 µL of water, was prepared
immediately before the reaction. A 10-µL aliquot of this solution
was added to 10-20 µL of an aqueous tryptic digest of a standard
protein (10-4-10-6 M) or a fraction of E. coli cell lysate digest
after SCX. Roughly 5 µL of pure triethylamine (TEA) was added
to raise the pH to 10.5-11.0 (as measured by indicator paper
strips). After mixing, the samples were placed in a heater (Heat
Block, VWR Scientific, San Dimas, CA) at 65 °C for 10 min.
Typically, 100% labeling efficiency for amino groups of lysines was
achieved under these conditions, provided that the reaction was
maintained at pH 10.5-11.0. In some cases, especially for peptides
with N-terminal glycine, N-termini partially reacted (producing an
MS signal up to 30% of the intensity of the main product peak).
This side reaction did not cause problems in analysis of the other
portion of the peptide that was not N-terminally blocked, as it could
subsequently be labeled with Alexa Fluor 350.
N-Terminal Labeling of Peptides. Labeling of peptides was
performed with a minor modification of the manufacturer’s
protocol.44 A 2-µL aliquot of buffer concentrate (1 M triethyl-
ammonium bicarbonate, pH 8.3) was added to 10 µL of the
guanidated tryptic digest of a standard protein (10-4-10-6 M) or
cell lysate digest. The tagging solution was prepared immediately
before labeling: 20 µg of the tag (N-hydroxysuccinimide ester)
was dissolved in 100 µL of 50% (v/v) solution of ethanol in water,
and 5 µL of the given solution was added to the digest. The
reaction was carried out at room temperature for 30 min, and then
a 5-µL aliquot of a fresh tag solution was added every 30 min, for
a total of 4-6 times within 2-3 h. The reaction products were
purified and analyzed by MALDI-TOF MS. The labeling was found
to be complete for the majority of the peptides for the digests of
standard proteins.
EXPERIMENTAL SECTION
Chemicals. All organic solvents, reagents, salts, standard
peptides, proteins, and indicator paper strips were purchased from
Sigma (St. Louis, MO). R-Cyano-4-hydroxycinnamic acid (CHCA)
was obtained from Bruker Daltonics (Billerica, MA) and used
without further purification. Coumarin dyes (N-hydroxysuccin-
imide esters), including Alexa Fluor 350, were purchased from
Molecular Probes (Eugene, OR), and trypsin was from Promega
(Madison, WI). An E. coli cell suspension was kindly provided
by Prof. Kim Lewis (Department of Biology, Northeastern
University). Novagen brand reagents for protein extraction were
purchased from VWR Scientific (San Dimas, CA).
Synthesis of Tris(2,4,6-trimethoxyphenyl)phosphonium-
acetic acid, N-Hydroxysuccinimide Ester (TMPP-Ac-OSu).
A 25-mL dry round-bottom flask, equipped with an N2 inlet, was
filled with bromoacetic acid N-hydroxysuccinimde (100 mg), 4
mL of anhydrous toluene, and a 4-mL toluene solution of tris-
(2,4,6-trimethoxyphenyl)phosphine (216.8 mg), and the resulting
reaction mixture was stirred for 1/2 h. The solid product was
collected via filtration, dried under vacuum and used without
further purification.
Preparation of a Cell Lysate Digest. A protein extract of an
E. coli cell lysate was prepared according to a protocol described
by Novagen (EMD Biosciences, Madison, WI). Briefly, a wet cell
pellet of E. coli (300 µg) was resuspended in BugBuster protein
extraction reagent (5 mL of the reagent/g of wet cell paste) at
room temperature. Additionally, for the same amount of wet cell
paste, two enzymes were added: 1 µL (25 units) of Benzonase
Nuclease (Novagen, EMD Biosciences) for DNA and RNA
cleavage and 1000 units of rLysozyme (Novagen, EMD Bio-
sciences) for efficient lysis of E. coli cell walls. The suspension
was incubated on a shaking platform for 10-20 min at room
temperature. Insoluble cell debris was removed by centrifugation
at 16000g for 20 min at 4 °C. The protein concentration was
estimated by the Bradford assay. Next, the protein extract was
cleaned from the lysis detergents by acetone precipitation,
specifically, treatment with 90% acetone at -80 °C for 15 min,
followed by centrifugation at 14000g for 15 min. The precipitated
protein was collected, reconstituted in 0.1 M triethylammonium
bicarbonate (TEA-HCO3) buffer, pH 8.5, denatured with 6 M
guanidine hydrochloride, and reduced with 5 mM dithiothreitol
(DTT). The protein extract was then diluted with TEA-HCO3
buffer to a final guanidine concentration of less than 0.5 M,
Sample Cleanup. Sample cleanup was required, both after
the guanidation reaction for efficient N-terminal labeling and after
N-terminal labeling, before nanoLC-MALDI TOF MS. In each case,
the cleanup was performed with Zip-Tip C18 (Millipore, Bedford,
(42) Rejtar, T.; Chen, H.-S.; Moskovets, E.; Li, L.; Andreev, V.; Karger, B. L.
Proceedings of the 51th ASMS Conference on Mass Spectrometry and Allied
Topics, Montreal, Canada, June 8-12, 2003.
(43) Chen, H.-S.; Moskovets, E.; Rejtar, T.; Karger, B. L. Proceedings of the 52th
ASMS Conference on Mass Spectrometry and Allied Topics, Nashville, TN,
2004.
(44) Handbook of Fluorescent Probes and Research Products, www.probes.com/
handbook.
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2087

MA) on 1-2 µg of total peptide mass, or with Ultra MicroSpin
C18 (The Nest Group, Southborough, MA) for sample amounts
up to 100 µg, according to the manufacturers’ protocols.
developed algorithms, MEND45 and PRESEL,46 to generate the
list of precursor ions for MS/MS analysis. The MS peak height
threshold was set at 200 counts, resulting, for each sample, in a
list of ∼6000 precursors submitted for MS/MS. The MS/MS data
were submitted to the MASCOT server for database searching.
The searches were performed against a SwissProt database. For
taxonomy, “Escherichia coli” was specified for the â-galactosidase
and E. coli fraction analysis, and “mammals” for the standard
proteins. To analyze Alexa Fluor 350-labeled samples, N-terminal
“Alexa” modification (monoisotopic mass increment 296.0228), was
added to the default list of modifications, while guanidation is a
standard option in MASCOT. To decrease the number of false
positives, the probability of random hits (p) was set <0.01,
meaning 99% confidence in the correct peptide identification. The
peptide mass tolerance was set at (50 ppm, and the fragment
mass tolerance at (0.15 Da. Trypsin was specified as the
proteolytic enzyme, and the maximum of two missed cleavages
was allowed. For native samples, methionine oxidation was set
as a variable modification; for tagged samples, two additional
modifications were added: lysine guanidation as permanent and
N-terminal Alexa as variable. Under these conditions, the MAS-
COT significance threshold was 26 for the E. coli peptides. The
“MUDPIT” scoring option in MASCOT was chosen,47 ion score
cutoff was set at 20, and only peptides with rank 1 were
considered.
Instrumentation and Methods. Mass spectra were acquired
on an Applied Biosystems 4700 TOF/TOF Proteomics analyzer
(Framingham, MA), equipped with delayed extraction and a 200-
Hz repetition rate UV laser (355 nm). The instrument was
controlled by the AB 4700 Data Explorer software, version 2.0.
The mass resolution on the AB 4700, operating in the positive
mode, was at least 10 000 (fwhm) for all peptide peaks. Typically,
600 shots/spectrum were accumulated in the MS mode and 1000
shots/spectrum in the MS/MS mode. When CID mode was used
for peptide fragmentation, air was the collision gas, and the CID
chamber pressure was in the range (1-6) × 10-6 Torr, with
accelerations of 1 and 2 kV. In the PSD mode, ion acceleration of
1 kV was used. Standard 192-well stainless steel MALDI sample
plates were used in this study. An internal standard was used for
mass calibration, and the Plate Model calibration option, allowed
by the instrument software, was used to alleviate spatial variations
of the mass accuracy across the MALDI plate. For the dried
droplet samples, the CHCA matrix was prepared in acetonitrile/
water (60:40, v/v) containing 0.1% trifluoroacetic acid (TFA), at a
concentration of 5 mg/mL. For the continuous deposition, 7 mg/
mL CHCA solution in 50:50 (v/v) ACN/water, 0.1% (v/v) TFA,
was used.
NanoLC-MALDI System. A multiplexed four-channel HPLC
system was coupled off-line to MALDI-MS using a laboratory-built
deposition interface, as described elsewhere.43 Each separation
channel consisted of an individual manual injection valve with a
2-mL sample loop and a 15 cm × 100 µm i.d. analytical column
(packed with Magic C18 phase (Alltech, Nicholasville, KY),
particle size 5 µm, pore size 100 Å). The mobile phases were
delivered by an UltiMate system (Dionex, Sunnyvale, CA) and
then split equally into four separation channels. The samples were
directly injected into the analytical columns and then simulta-
neously separated by a 50-min linear gradient at a flow rate of 1
µL/min in each of the four channels. The composition of solvent
A was 2% (v/v) ACN, 0.1% (v/v) TFA, and that of solvent B was
85% (v/v) ACN, 5% (v/v) 2-propanol, 0.1% (v/v) TFA. The eluents
from individual LC columns were then subsequently mixed in
Micro-Tees (Upchurch, Oak Harbor, WA) with MALDI matrix
solution (7 mg/mL CHCA, 0.1% (v/v) TFA in 50% (v/v) ACN/
water) at a 3:2 flow rate ratio and simultaneously deposited onto
MALDI plates in the form of spots. Each spot consisted of 5 s of
chromatographic time. A standard 2 in. × 2 in. MALDI plate could
accommodate four 7 × 28 arrays of spots, representing 16 min of
separation for each channel. During deposition, the interface was
controlled by a LabVIEW program (National Instruments, Austin,
TX). An internal standard, Glu-fibrinopeptide B, m/z 1570.6774,
was used for mass calibration.
RESULTS AND DISCUSSION
The goal of the present study was to develop a derivatization
approach that would enhance peptide identification in high-
throughput MALDI-TOF and TOF/TOF MS. In part 1,4 we
focused on the improvement in MALDI-TOF MS signal, which
was achieved through the use of commercially available N-terminal
coumarin tags. In this study, several approaches for directed MS/
MS fragmentation in TOF/TOF MS, with both high-energy CID
and unimolecular (PSD) fragmentation, were explored. The
advantage of the Alexa Fluor 350 tag for improved MS and MS/
MS analysis was examined in depth in the analysis of an SCX
fraction of an E. coli cell lysate.
Cationic Tag Derivatization. As noted in the introduction,
N-terminal labeling of peptides with cationic tags leads to directed
fragmentation and formation of a- and d-ion series in tandem
double-sector MS/MS instruments, where CID occurs under high-
energy conditions.27,28 In directed fragmentation, peptides fragment
in a predictable way, forming preferentially one (or more) nearly
complete series of fragment ions. It has been observed that
modern MALDI-TOF/TOF MS instruments are able to generate
high-energy CID fragmentation conditions,7,29 with ion acceleration
up to 3 keV and center-of-mass energy deposition up to 255 eV,30
comparable to tandem double-sector instruments.48 Therefore, the
use of cationic tags for MALDI-TOF/TOF MS analysis in the CID
mode was examined.
Data Acquisition and Processing. MASCOT server version
2.0 (Matrix Science, London, U.K.) was used for peptide and
protein identifications. The data for MS/MS ion fragment and ion
current distribution were extracted from MASCOT data using an
in-house-written PERL program. For an SCX fraction of the E.
coli digest, after completion of the data acquisition in the MS
mode, the spectra were processed for peak picking by internally
(45) Andreev, V. P.; Rejtar, T.; Chen, H.-S.; Moskovets, E. V.; Ivanov, A. R.; Karger,
B. L. Anal. Chem. 2003, 75, 6314-6326.
(46) Andreev, V.; Rejtar, T.; Chen, H.-S.; Moskovets, E.; Zang, L.; Karger, B. L.
Anal. Chem., in press..
(47) Mascot Proceedings of the 51th ASMS Conference on Mass Spectrometry and
Allied Topics, Nashville, TN, 2004.
(48) Poulter, L.; Taylor, L. C. E. Int. J. Mass Spectrom. Ion Processes 1989, 91,
183-197.
2088 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

Figure 2. Comparison of N-terminal cationic derivative vs native peptide for a peptide that does not contain Arg: MALDI-TOF/TOF CID mass
spectrum of a peptide EGVYVHPV, 1 pmol. MS/MS conditions are described in the Experimental Section. (A) Native peptide, m/z 899.4627; (B)
N-terminally labeled peptide with a cationic phosphonium tag (TMPP), m/z 1471.6438. b* corresponds to loss of the N-terminal tag.
Standard peptides and protein digests were labeled with various
ammonium and phosphonium cation tags and analyzed by MALDI-
TOF/TOF MS under CID conditions. Air was the collision gas,
and the CID chamber pressure was in the range (1-6) × 10-6
Torr, with accelerations of 1 and 2 kV. Under these conditions,
the center-of-mass energy deposited in peptides was calculated
to be in the range 10-40 eV.
Typical results are demonstrated in Figures 2 and 3, showing
MALDI-TOF/TOF CID mass spectra of native and N-terminally
labeled peptides with the tris(2,4,6-trimethoxyphenyl) phospho-
nium acetyl (TMPP-Ac) tag. The spectra of native peptides,
shown in Figures 2 and 3A, contain incomplete series of various
fragment ions, both C-terminal (e.g., y-ions) and N-terminal (a-
and b-ions), with various intensities. When labeled with a cationic
tag, directed fragmentation leading to a preferential formation of
an a-series was observed, if the peptide did not contain arginine
(Figure 2 B, EGVYVHP). On the other hand, if a C-terminal Arg
residue was present in the sequence (Figure 3B, EGVNDNEE-
GFFSAR), no selectively directed fragmentation was observed,
and the spectra of the labeled peptides contained both N-terminal
a-, b-ions, and C-terminal y-ions. Importantly, as can be seen from
Figure 3B, the MS/MS spectra typically showed considerably
lower overall intensity, as well as a decreased number and intensity
of structurally important fragments, compared to the spectra of
native peptides (Figure 3A). Similar results were observed for
several other studied cationic tags (data not shown). For all
cationic tags, fragments, corresponding to the complete loss of
the tag (or its fragment), were the most intense peaks in the MS/
MS spectra of labeled peptides (b* ions in Figures 2 and 3 B).
For MALDI-TOF/TOF MS in the PSD mode (i.e., without a
collision gas), it is known that [M + H]+ ions carry less internal
energy than in the CID mode,22 and the spectra contain a reduced
number of low-mass, high-energy fragment ions.30 From our
experience, for database searching, the PSD mode typically
provides better MS/MS spectral quality than CID. This result is
due to the fact that the fragment ion current in the PSD mode is
distributed mainly over backbone b- and y-ion series, leading to
a higher S/N of ions in the spectrum (see Figure 4, discussed
later). Fragmentation of the cationic tag labeled peptides under
MALDI-TOF/TOF PSD conditions has also been explored, and
results similar to the CID mode were obtained (data not shown).
From all these results, it can be seen that labeling with cationic
tags causes suppression of MS/MS fragmentation under MALDI-
TOF/TOF MS conditions, both in the CID and PSD modes. A
likely explanation is that the addition of a cationic moiety increases
the gas-phase stability of the peptide ions, thus making such ions
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2089

Figure 3. Comparison of N-terminal derivatives vs native peptide for a peptide that contains Arg: MALDI-TOF/TOF CID mass spectrum of a
tryptic peptide EGVNDNEEGFFSAR (Glu-fibrinopeptide B), 1 pmol. MS/MS conditions are the same as in Figure 2. (A) Native peptide, m/z
1570.6774; (B) the same peptide N-terminally labeled with the cationic tag (TMPP), m/z 2142.8585; (C) N-terminally labeled with OMe-coumarin,
m/z 1772.7037; D. N-terminally labeled with Alexa Fluor 350, m/z 1865.6922. Fragments labeled with an asterisk: a*, b*, and y*, correspond
to loss of the N-terminal tag.
2090 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

more difficult to fragment. According to the “mobile proton” theory
of MS/MS fragmentation,21 a proton, added to a peptide during
ionization, plays a major role in the backbone fragmentation
process. Sites with high gas-phase basicities, such as the guanidine
group of Arg, retain the proton, thus limiting formation of
sequence-specific product ions. In contrast to the usual [M + H]+
peptide ions, the [M]+ peptide ion, which is formed with cationic
tag labeled peptides, does not contain a mobile proton, and
fragmentation then follows an alternative pathway, i.e., a charge-
remote mechanism.48,49 It is likely that MALDI-TOF/TOF MS, in
contrast to the tandem double-sector MS, does not provide
sufficient energy to support this fragmentation pathway. Therefore,
it can be concluded that cationic tags are not suitable as
derivatizing agents for high-throughput MALDI-TOF/TOF MS
analysis.
Coumarin Tags. In part 1 of this study, it was shown that
N-terminal coumarin tags are beneficial for peptide analysis in
the MS mode, as they enhance the intensities of signals of
hydrophilic peptides.4 Since, as noted above, moieties with high
gas-phase basicity do not favor MS/MS fragmentation,21 an
alternative approach for improvement of both MALDI-TOF MS
and MS/MS of peptides was explored, and small hydrophobic
coumarin tags were tested in the MS/MS mode.
The same peptides and protein digests, as used above for
cationic tags, were labeled with the following coumarin tags:
7-methoxycoumarin, 7-hydroxycoumarin, Marina Blue, and Pacific
Blue (see part 1 of this work4 for structures) and analyzed by
MALDI-TOF/TOF MS under both CID and PSD conditions. The
MALDI-TOF/TOF CID spectrum of the same peptide Glu-
fibrinopeptide B (EGVNDNEEGFFSAR), labeled with the OMe-
coumarin tag, is shown in Figure 3C. Compared to the MS/MS
spectrum of the native peptide (Figure 3A), the spectrum of the
tagged peptide reveals increased intensities of b-ions, in particular,
b1-b3. A common feature of coumarin-tag directed fragmentation
is the appearance of a1 and b1 ions. The intensities of these ions
were enhanced in the CID mode, compared to PSD. b1 ions are
not present in the MS/MS spectra of native peptides, but can be
observed in the spectra of the N-terminal derivatives formed via
the amide bond, RsC(dO)s(NH-peptide).19,35,36 The explanation
for this result is that the carbonyl group of the tag participates in
the formation of the five-member oxazolone ring, typical for
bn-ions (n g 2).50 Analysis of MS/MS fragment ion distribution
revealed that, despite some improvement in the intensity of the
b-ion series, labeling did not result in a significant increase in the
number of b-ions, and MASCOT scores of identified peptides were
not improved (data not shown). Therefore, while useful for peptide
mass fingerprinting (PMF), coumarin tags 7-methoxycoumarin,
7-hydroxycoumarin, Marina Blue, and Pacific Blue were not
selected for high-throughput LC-MALDI-TOF/TOF MS analysis
of complex proteomic samples.
N-terminal labeling with a sulfo tag, SPITC, was suggested.51-53
To obtain a ladder of y-ions for tryptic peptides, lysine-terminated
peptides required guanidation, both to prevent double labeling
and to create a high proton affinity site on the C-terminus. In part
1,4 a sulfo group containing coumarin tag, Alexa Fluor 350, was
found to provide better MALDI-TOF MS signals of peptides
compared to the CAF reagent.17 In the present work, a series of
peptides, labeled with Alexa Fluor 350 were examined by MALDI-
TOF/TOF MS. Figure 3D is the MALDI-TOF/TOF mass spec-
trum in the CID mode of the same Glu-fibrinopeptide B, N-ter-
minally labeled with the Alexa Fluor 350 tag. It can be seen that,
in the MS/MS mode, peptides labeled with Alexa Fluor 350
showed extensive y-series fragmentation, typical for peptides that
have been N-terminally labeled with sulfo group-containing
tags.38-40,51-53 Moreover, the intensities of all y-ions were uniformly
high, which is advantageous for the identification of the peptides
over a wide concentration range (see below).
To explore in more detail the advantages of Alexa Fluor 350
labeling for peptide analysis, digests of five proteins, R-casein,
â-casein, myoglobin, bovine serum albumin, and cytochrome c,
were labeled and analyzed by MALDI-TOF/TOF MS in both the
CID and PSD modes. The resulting MS/MS spectra were
submitted to database searching and processed, as described in
the Experimental Section. MALDI-TOF/TOF MS ion fragment
and ion current distribution data were extracted from the
MASCOT data using an in-house-written PERL program, to reveal
changes in fragmentation caused by Alexa Fluor 350 labeling,
compared to the native peptides, and to determine the optimum
conditions for MS/MS analysis of the labeled peptides.
A summary of the distribution of ion current between the main
TOF/TOF MS fragment ion series for native and labeled peptides
is shown in Figure 4. Fragment ion series, typical for both the
TOF/TOF CID and PSD modes and used by the MASCOT
scoring algorithm, were considered, i.e., a-, b-, y-, immonium, and
internal ions. Fragments corresponding to a neutral loss of water
or ammonia were summed together and are represented in the
figure as “% neutral”. Also considered were nonidentified frag-
ments, i.e., unassigned fragments, whose presence lower MAS-
COT scores. The most important result in Figure 4 is the clear
shift toward preferential formation of y-ion series in the case of
Alexa Fluor 350-labeled peptides, at the expense of all other
fragment ion series. As can be seen, the fraction of y-series for
both native and labeled peptides was significantly higher in the
PSD compared to the CID mode. In contrast, CID spectra of both
native and labeled peptides were richer in less informative
immonium and internal ions, representing up to 50% of the total
MS/MS ion current (see the bars for the native peptides). At the
same time, Alexa labeling resulted in a slightly higher fraction of
unidentified fragments, which could possibly be explained by the
lowered fragmentation energy barrier due to the presence of the
sulfo group. In the PSD mode, the fraction of the y-series was
more than 70% of the total ion current, decreasing to 50% in the
CID mode. This latter result confirms the unimolecular mecha-
nism of fragmentation of sulfonated peptide ions2 and indicates
Labeling with the Coumarin Tag, Alexa Fluor 350. The
use of N-terminal tags containing a sulfo group for improved de
novo peptide sequencing by MALDI-TOF PSD MS was pioneered
by Keogh and colleagues and recently commercialized as the
chemically assisted fragmentation (CAF) reagent kit.2,17,39,40 Later,
(51) Marekov, L. N.; Steinert, P. M. J. Mass Spectrom. 2003, 38, 373-377.
(52) Wang, D.; Kalb, S. R.; Cotter, R. J. Rapid Commun. Mass Spectrom. 2004,
18, 96-102.
(49) Johnson, R. S.; Martin, S. A.; Biemann, K. Int. J. Mass Spectrom. Ion Processes
1998, 86, 137-154.
(50) Yalcin, T.; Khouw, C.; Csizmadia, I. G.; Peterson, M. R.; Harrison, A. G. J.
(53) Chen, P.; Nie, S.; Mi, W.; Wang, X.-C.; Liang, S.-P. Rapid Commun. Mass
Spectrom. 2004, 18, 191-198.
Am. Soc. Mass Spectrom. 1995, 6, 1165-1174.
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2091

Figure 4. MALDI-TOF/TOF MS fragment ion current distribution obtained from MS/MS spectra of tryptic peptides of five proteins: R-casein,
â-casein, myoglobin, bovine serum albumin, and cytochrome c. The percentage of the total fragment ion current for a particular fragment ion
series is the ratio of the sum of the intensities of all fragment ions of this series to a sum of the intensities of all ions in the spectrum. The
intensity of each fragment ion was determined from the specific cluster area, which is the sum of the peak areas in the isotopic cluster for the
given ion.
that PSD is the mode of choice for the MS/MS analysis of the
sulfonated peptide samples by TOF/TOF MS/MS.
to the native peak widths, with an increase in the retention times
of the labeled peptides, compared to the unlabeled peptides. It
can be concluded that the addition of the tag to the peptide did
not appear to significantly affect chromatographic performance.
Analysis of a Strong Cation Exchange Fraction of a Tryptic
Digest of E. coli Lysate. After the successful preliminary studies,
the analysis of more complex peptide mixtures was performed.
An SCX LC fraction of a tryptic digest of an E. coli lysate was
divided into two parts, one of which was maintained native and
the other modified by guanidation of lysines, followed by N-
terminal labeling with Alexa Fluor 350. Both fractions were then
analyzed in parallel by four-channel multiplexed nanoLC-MALDI-
TOF/TOF MS, each sample in duplicate, as described in the
Experimental Section, and the data were submitted to MASCOT
for database searching.
The results showed the expected improved fragmentation for
many peptides, which led to the higher peptide scores due to
labeling. As an example, Figure 6 presents the MALDI TOF/TOF
MS/MS spectra of the peptide GALSAVVADSR from 50S riboso-
mal protein, m/z 1044.5560, which showed poor fragmentation in
the native form, yielding a Mascot score of 10, with the significant
threshold being 26. MS/MS of the same peptide in the labeled
form resulted in the spectrum with a complete ladder of y-ions
with a high MASCOT score of 95, leading to an unambiguous
identification.
Backbone ion series, such as b and y, are the most important
ion series for peptide identification via database searching.10
Typically, a combination of complementary b- and y-ions is
observed in MALDI-TOF/TOF mass spectra of native tryptic
peptides, with the fraction of y-ions being twice as large as that
of b-ions (see Figure 4). While b- and y-ions make the largest
contribution to the MASCOT scores of the peptides, the scores
are improved even more if a longer sequence of these ion series
is present in the spectrum. As a consequence, for the highest
possible MASCOT score, it is beneficial that all fragment ion
current be directed into a single fragment ion series. For MALDI-
TOF/TOF MS, Lys guanidation, followed by N-terminal deriva-
tization with agents containing a sulfo group, is an effective
strategy. Moreover, isotopic labels may be introduced with
guanidation,18,54 with the N-terminal tag, or metabolic incorpora-
tion, allowing flexibility in experimental design for differential
quantitation.
LC-MALDI-TOF/TOF Analysis of Complex Peptide Mix-
tures. Analysis of a Protein Digest. As noted earlier, in other
reports, N-terminal labeling with various sulfo tags has been
employed for the analysis of in-gel digests by MALDI-TOF PSD
MS.40,52,53 Since N-terminal labeling with Alexa Fluor 350 has been
found to be beneficial for both MALDI-TOF MS4 and MS/MS
analysis, we next decided to explore the possibility of analyzing
labeled samples by RPLC followed by MALDI-TOF/TOF MS.
Approximately 100 pmol of a tryptic digest of â-galactosidase was
guanidated and then N-terminally labeled with Alexa Fluor 350,
and ∼1/50 of this sample was analyzed by nanoLC-MALDI-TOF/
TOF MS, using a home-built multiplexed LC-MALDI deposition
device.42,43 Figure 5 shows extracted ion chromatograms of
representative native and derivatized peptides. Analysis of the
chromatograms revealed that narrow (30-40 s) chromatographic
peaks were obtained for the labeled peptides, which is comparable
Another important advantage found for the Alexa Fluor 350
labeling is the ability to obtain high-quality MS/MS spectra even
from very low intensity MS precursors. Figure 7 shows a typical
MALDI-TOF mass spectrum from one of the deposited spots from
the LC separation of the labeled sample. Peptide SVDEAANSDI-
VDK, from ATP synthase B chain, m/z 1698.668, produced an
MS signal of low intensity (350 counts, S/N ∼9). However, the
inset to Figure 7 shows that the MS/MS spectrum contained the
complete y-series, yielding a high MASCOT score of 119, well
above the significant threshold of 26.
To demonstrate further the improvement in peptide identifica-
tion due to the labeling for peptides in various MS signal intensity
(54) Brancia, F. L.; Montgomery, H.; Tanaka, K.; Kumashiro, S. Anal. Chem.
2004, 76, 2748-2755.
2092 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

Figure 5. Extracted ion chromatograms for the native and Alexa Fluor-350-labeled â-galactosidase peptides APLDNDIGVSATR (A) and
RDWENPGVTQLNR (B). In both figures, the peak widths are ∼30 s. See Experimental Section for details.
Figure 6. Example of high-confidence peptide identification in the E. coli fraction due to labeling. MS/MS fragmentation of GALSAVVADSR,
a peptide from 50S ribosomal protein L10 (L8). Monoisotopic mass of the native peptide (neutral) m/z 1044.556, MASCOT score 10; monoisotopic
mass of the labeled peptide (neutral) m/z 1339.571; fixed modifications guanidination (K); variable modifications Alexa (N-term); MASCOT
score 95. The significance threshold was 26 at p < 0.01.
ranges, the ratios (in percentages) of identified peptides in native
and tagged samples were calculated at different intensities of the
MS/MS precursor signals. All MS/MS precursors were grouped
by their monoisotopic MS peak height values into groups from
500 to 100 000 counts. For each group, the ratios of the number
of peptides that yielded MASCOT scores above 26 (significance
threshold) to the total number of submitted precursors of the same
peak height range were calculated. As presented in Figure 8 A in
the form of a histogram, the results indicated that the percentage
of identified peptides was higher for the tagged sample, compared
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2093

Figure 7. Example of high-confidence identification a low MS intensity peptide SVDEAANSDIVDK from ATP synthase B chain, due to labeling.
Monoisotopic mass of the labeled peptide (neutral) m/z 1698.668, fixed modifications guanidination (K), variable modifications Alexa (N-term),
and MASCOT score 119.
ranges as in Figure 8A. From Figure 8B, it can be seen that the
total number of peptides in each group was similar for native and
tagged samples, and a distribution of their MS intensities was
nearly identical. This important observation means that overall
no significant decrease in MS ion intensity was caused by the
labeling. It is important that the largest improvement was observed
in the low-intensity range (below 4000 counts), as over 80% of
the MS/MS precursors fell in this range (see Figure 8B).
While usually the probability of random hits (p) of less than
0.05 is typically used in MASCOT searching, we selected here
the more conservative setting of less than 0.01, meaning higher
than 99% confidence in the correct peptide identification, to
decrease the number of false positive identifications that result
from low-score hits. As seen in Figure 9A, a total of 419 peptides
were identified in the native sample, while 557 peptides were
identified in the labeled sample (each number consists of the
number of nonredundant hits obtained using the combined data
from the duplicate runs). Surprisingly, the overlap between
peptides identified in both the native and labeled samples was
small (14.5%, Figure 9A). At the same time, the overlap between
the peptides identified in each of the duplicate samples, either
native or labeled, was about 80-90% (data not shown). Figure 9B
shows the identification summary on the protein level: a total of
232 proteins were identified in the native sample and 252 proteins
Figure 8. (A) Comparison of percentages of peptides identified in
the native and Alexa Fluor 350-labeled E. coli fraction. Note that the
in the labeled sample. The number of overlapped proteins was
higher than was the case for peptides (43.6%, Figure 9B),
indicating that many proteins were identified in both the native
and labeled samples, but by different peptides. As a result,
confidence in the correct identification was increased for proteins
found in both samples. In particular, the fraction of protein
identifications based on two or more peptides increased from 44.8%
in the native sample to 64.6% in the combined native + labeled
sample data sets. On the whole, 50.8% more peptides and 31.2%
more proteins were identified from one SCX fraction of the E.
coli cell lysate due to the Alexa Fluor 350 labeling.
peak height scale on the X-axis is not linear. (B) From the same
experiment: a percentage of submitted MS/MS precursors of a
particular peak height range, out of all submitted precursors, for native
and labeled samples.
to the native. It can be further seen in Figure 8A that the
percentage of identified peptides was significantly higher (on
average, 2-3-fold) for the low-intensity precursors (500-1000
counts), confirming the results of Figure 7. Figure 8B shows a
histogram of the distribution of the MS signal intensities of the
native and labeled peptides, calculated for the same peak height
2094 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005

similar to the complementary use of the MALDI and ESI-MS
methods.56,57 The four-channel multiplexed continuous LC-MALDI
deposition and parallel analysis technology, using a 2-kHz repeti-
tion rate laser,43,58 should be well suited for the native and labeled
sample analysis approach utilized in this work. In summary, the
results demonstrate that the proposed analysis strategy leads to
a significant improvement in protein identification by LC-MALDI-
TOF/TOF MS.
CONCLUSIONS
This paper has shown that Lys guanidation followed by
N-terminal labeling with the commercially available sulfo group
coumarin tag, Alexa Fluor 350, is an effective approach for peptide
analysis in both the MS and MS/MS modes in LC-MALDI-TOF/
TOF MS. On average, no significant decrease in MS ion intensity
was caused by the labeling. On the other hand, the percentage of
identified peptides was considerably higher (on average, 2-3
times) for the tagged sample, compared to the native, due to the
high-quality MS/MS spectra that could be obtained, even from
very low intensity precursors.
Analysis of the MALDI-TOF/TOF MS fragment ion current
before and after labeling revealed changes in the fragmentation
pattern caused by the tag. In particular, labeling with Alexa Fluor
350 resulted in a dramatic increase of the y-ion fraction, while
the fractions of the rest of the fragment ions, such as a-, b-,
immonium, and internal ions, decreased. These trends were found
to lead to significantly improved MASCOT scores for the labeled
peptides. The N-terminal sulfo tag labeling approach is well suited
for commercially available MALDI-TOF/TOF MS instruments,
taking advantage of the high sensitivity and high mass accuracy
of such instruments in both the MS and MS/MS modes, as well
as the possibility of precise precursor ion selection.7
LC-MALDI-TOF/TOF MS analysis of the peptides labeled with
a sulfo tag was successfully performed. The analysis of the data
for tryptic peptides from an E. coli lysate revealed improved
peptide scores and an increased number of peptides identified
due to labeling. Moreover, derivatization did not impair chro-
matographic behavior of peptides, as narrow chromatographic
peaks were obtained for the labeled peptides. Importantly, overlap
in peptides identified in either the native or labeled samples was
small (14.5%). At the same time, the overlap between the two
replicate runs of the same sample, either native or labeled, was
close to 90%. The number of overlapped proteins was slightly
higher, suggesting that many proteins were identified in either
the native or labeled samples by different peptides. As a result,
confidence in the correct identification was increased for proteins
found in both samples. On the whole, 50.8% more peptides and
31.2% more proteins were identified from one SCX fraction of the
E. coli cell lysate due to the labeling.
Figure 9. Summary of peptides (A) and proteins (B) identified in
the native and labeled samples from the SCX fraction of an E. coli
lysate digest. A total of 852 unique peptides were identified with 99%
confidence using both the native and labeled data sets. These
peptides belong to 337 proteins (MUDPIT scoring option was used;
see Results and Discussion section for details).
Examination of the experimental data revealed that the nature
of peptides identified in both the native and labeled data sets was
different. Typically, longer peptides with C-terminal arginine
prevailed among the native peptides, while shorter peptides with
C-terminal lysine (guanidated) predominated in the labeled
sample. There are several factors contributing to this trend. From
PMF data acquired by MALDI-TOF MS, it is known that guani-
dation improves sequence coverage by increasing the MS intensi-
ties of lysine-terminated peptides.55 The effect is peptide-dependent
and may vary from no change to a 10-fold enhancement.55 On the
other hand, labeling with Alexa Fluor 350 also caused changes in
the MS signal intensities of the labeled peptides, varying from a
∼ 5-fold decrease to an enhancement (1.5-fold), depending on the
hydrophobicity of the peptide.4 The resulting effect on MALDI
MS signals of peptides was that both guanidation and Alexa Fluor
350 on average counterbalanced each other, although the distribu-
tion of the MS intensities for various peptides differed. In
particular, the decrease in signal due to Alexa Fluor 350 labeling
was more pronounced for the Arg-terminated peptides, which were
not enhanced by guanidation. The decrease in signal intensity was
somewhat higher for large Arg-peptides with MW >2500, as the
mass increase due to labeling (∼300 Da) moved such peptides
out of the optimum mass range for MALDI with the CHCA matrix.
On the other hand, the bias for the detection of shorter Lys-
peptides after labeling can be explained by both the enhancement
caused by guanidation and the relative increase in the mass and
hydrophobicity due to labeling, which facilitate the detection of
such peptides by MALDI-TOF MS.4 The results suggest that the
labeled samples should be run along with the native samples,
The four-channel multiplexed continuous LC-MALDI deposi-
tion and parallel analysis technology developed in our labora-
tory43,58 is well suited for the native and labeled sample analysis
(56) Juhasz, P.; Falick, A.; Graber, A.; Hattan, S.; Khainovski, N.; Marchese, J.;
Martin, S.; Patterson, D.; Williamson, B.; Malmstrom, J.; Westergren-
Thorsson, G.; Marko-Varga, G. Proceedings of the 50th ASMS Conference on
Mass Spectrometry and Allied Topics, Orlando, FL, June 2-6, 2002.
(57) Bodnar, W. M.; Blackburn, R. K.; Krise, J. M.; Moseley, M. A. J. Am. Soc.
Mass Spectrom. 2003, 14, 971-979.
(58) Chen, H.-S.; Rejtar, T.; Andreev, V.; Moskovets, E.; Karger, B. L., manuscript
(55) Beardsley, R. L.; Reilly, J. P. Anal. Chem. 2002, 74, 1884-1890.
in preparation.
Analytical Chemistry, Vol. 77, No. 7, April 1, 2005 2095

approach demonstrated in this paper. To increase the speed of
analysis, TOF/TOF instruments equipped with high-repetition rate
lasers (e.g., 2 kHz) can be used for data acquisition.59 It should
be further noted that fluorescent properties of the Alexa Fluor
350 tag may be beneficial for the detection of low sample amounts,
as well as for quantitation.60 Differential labeling and quantitation
may be performed by metabolic labeling or by using the isotopi-
cally labeled guanidation reagents, such as 13C- and 15N2-labeled
O-methylisourea54 or the commercially available Lys Tag 4H
reagent (Agilent Technologies).18 A 13C version of the Alexa Fluor
350 tag may be synthesized, as well. All these factors suggest
that labeling with Alexa Fluor 350 has real potential for use in
high-throughput multiplexed LC-MALDI-TOF MS/MS analysis of
proteomic samples.
ACKNOWLEDGMENT
Financial support from the NIH (GM15847) is gratefully
acknowledged. The authors thank Prof. Kim Lewis (Department
of Biology, Northeastern University) for providing the E. coli
sample. The authors also thank Dr. Jennifer Campbell (Applied
Biosystems) for assistance with high-energy CID experiments
performed with the 4700 TOF/TOF instrument and Drs. Darryl
Pappin (Applied Biosystems) and Peter O’Connor (Mass Spec-
trometry Facility, BU Medical School) for helpful discussions.
Contribution No. 846 from the Barnett Institute.
Received for review November 3, 2004. Accepted January
19, 2005.
(59) Moskovets, E.; Chen, H.-S.; Rejtar, T.; Andreev, V.; Karger, B. L. Proceedings
of the 52th ASMS Conference on Mass Spectrometry and Allied Topics,
Nashville, TN, May 23-27, 2004.
(60) Hochleitner, E. O.; Sondermann, P.; Lottspeich, F. Proteomics 2004, 4, 669-
676.
AC048375G
2096 Analytical Chemistry, Vol. 77, No. 7, April 1, 2005