A method for the chemical generation of N-terminal peptide sequence tags for rapid protein identification

Article abstract of DOI:10.1021/ac9911847

We describe a method for generating multiple small sequences from the N terminal of peptides in unseparated protein digests by stepwise thioacetylation and acid cleavage. The mass differences between a series of N-terminally degraded peptides give short sequences of defined length. Such short sequence tags together with the mass of the parent peptide can be used to identify the protein in a database. The sequence ladders are generated without the use of chain terminators or sample aliquoting and the degradation reagents are water soluble so that the chemistry can be carried out on peptides immobilized on C-18 reversed-phase supports without any peptide loss due to washing with organic solvents as occurs in Edman type sequencing. The entire procedure can be automated, and we describe a prototype device for the parallel analysis of multiple samples. We demonstrate the effectiveness of this chemical tagging method in a comparison with Edman sequencing, peptide mass fingerprinting, and MS/MS analysis of crude protein fractions obtained from an HPLC separation of the Escherichia coli ribosome complex which consists of 57 proteins. We show that chemical tagging is a viable first-pass high-throughput identification method to be used prior to an in depth MS/MS analysis.

Full text of DOI:10.1021/ac9911847

Anal. Chem. 2000, 72, 1006-1014
A Me t h o d fo r t h e Ch e m ic a l Ge n e ra t io n o f
N-Te rm in a l P e p t id e S e q u e n c e Ta g s fo r Ra p id
P ro t e in Id e n t ific a t io n
Sjouke Hoving, Ma rtin M u1 nc hba c h, Holge r Sc hm id, Luc a Signor, Anton Le hm a nn,
We rne r Sta ude nm a nn, Ma nfre do Qua droni, a nd Pe te r J a m e s *
Protein Chemistry Laboratory, Universit a¨ tsstrasse 16, Swiss Federal Institute of Technology, 8092 Z u¨ rich, Switzerland
microchips and arrays,3 differential display PCR,4 and serial
analysis of gene expression.5
We describe a method for generating multiple small
sequences from the N terminal of peptides in unseparated
protein digests by stepwise thioacetylation and acid cleav-
age. The mass differences between a series of N-terminally
degraded peptides give short sequences of defined length.
Such short “sequence tags” together with the mass of the
parent peptide can be used to identify the protein in a
database. The sequence ladders are generated without the
use of chain terminators or sample aliquoting and the
degradation reagents are water soluble so that the chem-
istry can be carried out on peptides immobilized on C-1 8
reversed-phase supports without any peptide loss due to
washing with organic solvents as occurs in Edman type
sequencing. The entire procedure can be automated, and
we describe a prototype device for the parallel analysis of
multiple samples. We demonstrate the effectiveness of
this chemical tagging method in a comparison with Edman
sequencing, peptide mass fingerprinting, and MS/ MS
analysis of crude protein fractions obtained from an HP LC
separation of the Esch er ich ia coli ribosome complex
which consists of 5 7 proteins. We show that chemical
tagging is a viable first-pass high-throughput identification
method to be used prior to an in depth MS/ MS analysis.
The level of protein expression and activity is modulated to a
great extent by a wide variety of posttranslational modifications,
such as phosphorylation, and by the relative rate of synthesis and
degradation and is, to a large extent, independent of the level of
6
mRNA expression. In order for proteome-level analysis to be
viable, rapid and sensitive protein identification methodologies
must be developed. The earliest methods, such as sequence
determination by Edman degradation or compositional analysis
such as amino acid analysis, are still useful but are increasingly
being replaced by mass spectrometry based methods. The first
advance was the development of database-searching algorithms
to identify proteins on the basis of the masses of the peptides
produced by sequence-specific chemical or proteolytic diges-
tion.7,8,9,10,11 This was subsequently extended to the use of
orthogonal digest data,12 partially interpreted MS/ MS,13 and raw
MS/ MS data for DNA database searching.14 Peptide mass
fingerprinting is very rapid (1 sample per 30 s in fully automated
mode) but the data is only useful for searching protein databases.
MS/ MS searching is much more effective for DNA database
searching however data accumulation is relatively slow at ∼10
min per sample. We therefore decided to develop an intermediate
solution, in which data accumulation is very rapid but in which
some sequence data is also generated to allow DNA database
searching and which gives a much higher confidence level in the
search result.
The development of high-throughput DNA sequencing and
computer algorithms for the rapid assembly of the random
sequence fragments into large contiguous sequences has resulted
in an exponential growth in database size. Currently, 17 bacterial
and 2 eukaryotic genomes have been completed, with another 62
bacterial and 9 eukaroytic genomes due before the turn of the
(3) Schena, M.; Shalon, D.; Davis, R. W.; Brown, P. O. Science (Washington,
D.C.) 1 9 9 5 , 270, 467-70.
(
(
4) Liang, P.; Pardee, A. B. Science (Washington, D.C.) 1 9 9 2 , 257, 967-71.
5) Velculescu, V. E.; Zhang, L.; Vogelstein, B.; Kinzler, K. W. Science
(Washington, D.C.) 1 9 9 5 , 270, 484-7.
century. These, together with extensive Expressed Sequence Tag
1
(6) Gygi, S. P.; Rochon, Y.; Franza, B. R.; Aebersold, R. Mol. Cell. Biol. 1 9 9 9 ,
(
ESTs, partial mRNA sequences ) libraries, allow the entire
1
9, 1720-30.
potential protein complement of organisms to be defined (the
(
7) Henzel, W. J.; Billeci, T. M.; Stults, J. T.; Wong, S. C.; Grimley, C.; Watanabe,
Proteome). Genome-wide studies of gene expression are now
possible at the mRNA level,2 since the development of DNA
C. Proc. Natl. Acad. Sci. U.S.A. 1 9 9 3 , 90, 5011-5.
(8) James, P.; Quadroni, M.; Carafoli, E.; Gonnet, G. Biochem. Biophys. Res.
Commun. 1 9 9 3 , 195, 58-64.
(
9) Mann, M.; Hojrup, P.; Roepstorff, P. Biol. Mass. Spectrom. 1 9 9 3 , 22, 338-
*
Corresponding author: (phone) 0041 1632 2919; (fax) 0041 1632 1591; (e-
45.
mail) peter.james@bc.biol.ethz.ch.
(10) Pappin, D. J. C.; Hojrup, P.; Bleasby, A. J. Curr. Biol. 1 9 9 3 , 3, 327-32.
(11) Yates, J. R. D.; Speicher, S.; Griffin, P. R.; Hunkapiller, T. Anal. Biochem.
1 9 9 3 , 214, 397-408.
(12) James, P.; Quadroni, M.; Carafoli, E.; Gonnet, G. Protein Sci. 1994, 3, 1347-
50.
(13) Mann, M.; Wilm, M. Anal. Chem. 1 9 9 4 , 66, 4390-9.
(14) Eng, J. K.; McCormack, A. L.; Yates, J. R. J. Am. Soc. Mass Spectrom. 1 9 9 4 ,
5, 976-989.
(
(
1) Adams, M. D.; Kelley, J. M.; Gocayne, J. D.; Dubnick, M. Polymeropoulos,
M. H.; Xiao, H.; Merril, C. R.; Wu, A.; Olde, B.; Moreno, R. F.; Kerlavage,
A. R.; McCombie, W. R.; Venter, J. C. Science (Washington, D.C.) 1 9 9 1 ,
2
52, 1651-6.
2) Lashkari, D. A.; DeRisi, J. L.; McCusker, J. H.; Namath, A. F.; Gentile, C.;
Hwang, S. Y.; Brown, P. O.; Davis, R. W. Proc. Natl. Acad. Sci. U.S.A. 1 9 9 7 ,
9
4, 13057-62.
1006 Analytical Chemistry, Vol. 72, No. 5, March 1, 2000
10.1021/ac9911847 CCC: $19.00 © 2000 American Chemical Society
Published on Web 02/04/2000

Ladder sequencing, in which a sequence is read by the mass
differences between sequential degradation products, was devel-
oped first as an enzymatic technique15 and then subsequently as
a modified Edman type chemical degradation.16,17,18 The use of
exopeptidases to generate ladders is limited by the extreme
variability of the activity of the protease toward the substrate and
is only useful for individual isolated peptides.19,20 Chemical
sequencing using phenylisothiocyanate with a small percentage
of phenylisocyanate as a chain-terminating reagent has been used
by Chait et al.16 to sequence long isolated peptides in a modified
commercial Edman sequencer. The main disadvantages of the
method are the loss of peptide during the washing steps, which
limits the sensitivity, as well as the terminating reagent, which
removes the alpha N terminus as a charge carrier, thereby
diminishing the relative effectiveness of ionization. The volatile
trifluoroethylisothiocyanate analogue developed by Pappin et al.
removes the need for washes with organic solvent but requires
that the parent peptide be added back in aliquots in order to
generate the ladder which again causes losses through sample
handling. We therefore chose to develop a method that could be
carried out in aqueous solution with an unseparated mixture of
peptides immobilized on a reversed-phase material and that leaves
the N terminal unmodified so that it may act as a proton acceptor.
was bubbled for 1 h through the lower phase until a white solid
appeared (carboxymethyl-thioimidate). CAUTION, HCl gas is
extremely agressive and all handling should be carried out in an
appropriate fume cupboard. The solvent was then evaporated
under vacuum with a rotary evaporator and 250 mL of dry pyridine
2
were added. H S was bubbled through for 4 h until the reaction
was completed. 100 mL ice water was added and stirred to dissolve
the ammonium chloride. The liquid was then poured over a
mixture of 300 mL concentrated hydrochloric acid, 100 mL water
and 300 g ice. The product was extracted with 300 mL diethyl
ether. The aqueous phase was extracted twice more with 150 mL
diethyl ether. The combined diethyl ether layers were then
washed with 3 M hydrochloric acid, following which the diethyl
ether was dried with Na SO . The diethyl ether was removed by
2
4
rotary evaporation, yielding about 20 g of oil. The TATG was
crystallized from warm carbon tetrachloride, the yield was 5 g of
1
yellow solid. H NMR (CDCl , 300 MHz): δ ) 2.89 (s, CH ), 4.13
3
3
13
(s, CH ) ppm. C NMR (CDCl , 75 MHz, proton decoupled): δ
2
3
) 38.57 (CH ), 38.86 (CH ), 173.9 (COOH), 230.8 (CdS) ppm.
3
2
ESI-MS: 45 (6.88), 59 (80.9), 76 (4.86), 91 (7.51), 117 (4.83), 150
(100) M+.
Ribosome Isolation. Escherichia coli MC4100 F- araD139
∆(argF-lac) U169 rspL150 relA1 deoCl ptsF25 rpsR flbB5301 was
2
2
obtained from the Cold Spring Harbor Laboratory collection.
EXPERIMENTAL SECTION
Bacteria were cultivated in a sulfur-free, synthetic glucose-salts
medium, with the addition of 500 µM inorganic sulfate as
described before.23 The cultures were grown aerobically on a
rotary shaker (180 rpm) at 37 °C, and growth was monitored
spectrophotometrically at 600 nm. Cells were harvested in the mid-
exponential phase (A600 ) 1) by centrifugation at 9000 rpm in a
Sorval GSA rotor for 10 min and washed twice with 50 mM Tris-
HCl, pH 7.0. The cells (10 g of wet cells) were resuspended in 70
4
Materials. Acetonitrile, mercaptoacetic acid, N-ethylmorpho-
line, pyridine, thioacetylthioethane, and trifluoroacetic acid (for
protein sequence analysis) were purchased from Fluka AG
(
Buchs, Switzerland). R-Cyano-4-hydroxycinnamic acid, N-methyl-
piperidine, and S-(thiobenzoyl)thioglycolic acid were purchased
from Aldrich GmbH (Buchs, Switzerland). Trizma base was
purchased from Sigma Chemical Co. (Buchs, Switzerland). Am-
monium hydrogen carbonate, â-mercaptoethanol, sodium sulfide,
and sodium sulfate were purchased from Merck AG (Darmstadt,
Germany). Acetic acid, carbon tetrachloride, diethyl ether, hy-
drochloric acid (min. 37%), and HPLC grade acetonitrile were
purchased from Riedel-de Ha e¨ n AG (Seelze, Germany). Sequenc-
ing-grade modified trypsin was purchased from Promega (Z u¨ rich,
Switzerland). DNase was purchased from Boehringer (Mannheim,
Germany). PTFE membranes with C-18 embedded material
mL of extraction buffer (20 mM Tris-HCl, pH 7.4, 40 mM NH Cl,
10 mM MgCl , and 7 mM â-mercaptoethanol). One tablet of a
2
cocktail of protease inhibitors (Boehringer, Mannheim, Germany)
was added, and the cells were ruptured by two passes through a
chilled French pressure cell at 20 000 psi. Ten microliters of
DNase (10 µg/ µL) was added to the ruptured cells and incubated
at 37 °C for 40 min. Cell debris was removed by centrifugation at
16 000 rpm for 30 min at 4 °C in a Sorvall SS-34 rotor. Ribosomes
were isolated as previously described.24 The supernatant was
centrifuged at 45 000 rpm for 2 h at 4 °C in a Beckman Ti45 rotor.
The pellet contains the crude ribosomes and was resuspended in
20 mL of high-salt buffer (20 mM Tris-HCl, pH 7.4, 400 mM NH4-
(
(
Empore, 3M) were from Varian (Zug, Switzerland). Peptide C21W
WFRGLNRIQTQIRVVNAFRSS) was synthesized manually using
Fmoc chemistry.
Synthesis of Thioacetylthioglycollic Acid (TATG). The
synthesis was based on the description of Mross and Doolittle.21
Mercaptoacetic acid (46 g, 0.50 mole) and acetonitrile (22.5 g,
Cl, 10 mM MgCl , and 7 mM â-mercaptoethanol). Then, 5 mL of
2
the crude ribosomal solution was layered onto a 7-mL 17.5%
0
.55 mole) were mixed in a reaction flask and cooled on ice. The
sucrose cushion (in high-salt buffer) and centrifuged at 45 000
mixture was overlaid with 2-3 cm petroleum ether and HCl gas
rpm in a Beckman Ti45 rotor for 3 h at 4 °C. The pellet (containing
7
0S ribosomes) was resuspended in 10 mL of low-Mg2 buffer
+
(
(
(
15) Aimoto, S.; Takao, T.; Shimonishi, Y.; Hara, S.; Takeda, T.; Takeda, Y.;
Miwatani, T. Eur. J. Biochem. 1 9 8 2 , 129, 257-63.
16) Chait, B. T.; Wang, R.; Beavis, R. C.; Kent, S. B. Science (Washington, D.C.)
(
4 2
20 mM Tris-HCl, pH 7.4, 40 mM NH Cl, 0.3 mM MgCl , and 7
mM â-mercaptoethanol) layered onto four continuous 10-30%
sucrose gradients, and centrifuged in a Beckman Ti28 swing-out
rotor at 18 000 rpm for 14 h at 4 °C. The fractions containing
1
9 9 3 , 262, 89-92.
17) Bartlet-Jones, M.; Jeffery, W. A.; Hansen, H. F.; Pappin, D. J. Rapid Commun.
Mass Spectrom. 1 9 9 4 , 8, 737-42.
(
(
18) Gu, Q. M.; Prestwich, G. D. J. Pept. Res. 1 9 9 7 , 49, 484-91.
19) Korostensky, C.; Staudenmann, W.; Dainese, P.; Hoving, S.; Gonnet, G.;
James, P. Electrophoresis 1 9 9 8 , 19, 1933-40.
(22) Silhavy, T. J.; Berman, M. L.; Enquist, L. W. Experiments with Gene Fusions;
Cold Spring Harbor Laboratory: New York, 1984.
(23) Kertesz, M. A.; Leisinger, T.; Cook, A. M. J. Bacteriol. 1 9 9 3 , 175, 1187-
90.
(24) Traub, P.; Mizushima, S.; Lowry, C. V.; Nomura, M. Methods Enzymol. 1979,
49, 391-407.
(
(
20) Patterson, D. H.; Tarr, G. E.; Regnier, F. E.; Martin, S. A. Anal. Chem. 1 9 9 5 ,
6
7, 3971-8.
21) Mross, G. A.; Doolittle, R. F. In Advanced methods in protein sequence
determination; Needleman, S. B., Ed.; Springer-Verlag: Berlin, 1977.
Analytical Chemistry, Vol. 72, No. 5, March 1, 2000 1007

protein were pooled (60 mL total) and dialyzed overnight against
5
or 2 µL of 100 mM TATG in 2% NEM (cycle 3 and 4) and incubate
5 min, (d) purge with N for 5 min at 1 bar, (e) TFA gas delivery
for 3 min at 0.5 bar, (f) purge with N for 10 min at 0.5 bar, and
(g) wash with 50 µL of water. Then, the cycle was usually repeated
(steps a-g) three times. Finally, the peptides were slowly eluted
with 50 µL of 60% acetonitrile/ 0.1% TFA.
L of low-Mg2 buffer. In the presence of a low Mg2+ concentra-
+
2
tion, the two subunits dissociate. The dialyzed solution was
centrifuged at 45 000 rpm for 4 h at 4 °C and the ribosomes were
resuspended in 7.5 mL of low-Mg2 buffer and stored at -20 °C.
Protein was determined, according to a modified method of
Lowry,25 as 7.5 mg/ mL.
2
+
MALDI Mass Spectrometry. The masses of the intact
proteins from the HPLC separation as well as those of the tryptic
peptides and the degraded peptides were determined by MALDI-
MS. Approximately 1 pmol in 0.5 µL was added to the same
amount of a saturated matrix solution (15 mg of R-cyano-4-
hydroxy-cinnamic acid in 50% acetonitrile and 1.25% TFA in water)
and allowed to dry at ambient temperature. Spectra were recorded
using a Voyager Elite MALDI-TOF mass spectrometer (Perseptive
Biosystems, Framingham, MA). Samples were analyzed in delayed
extraction reflector mode using an accelerating voltage of 20 kV,
a pulse delay time of 150 ns, a grid voltage of 60%, and a guide
wire voltage of 0.05%. Spectra were accumulated for 32 or 64 laser
shots.
MS/ MS Sequencing of Ribosomal P roteins. Two microli-
ters of each tryptic digest (∼5 pmol) from the ribosomal proteins
was desalted on the C18 membranes and redissolved in 50%
methanol and 1% acetic acid before being introduced into the mass
spectrometer at a flow rate of 0.2 µL/ min with a syringe pump.
MS/ MS sequencing was performed on a Finnigan MAT LCQ ion
trap mass spectrometer (San Jose, CA). The peaks of interest were
selected with a mass window of (3 amu, and fragmentation of
each peptide was established using a relative collision energy of
35-60 for MH+ ions and 20-30 for MH2+ ions. On average, 20
MS/ MS spectra were measured within 15 min for each protein
digest.
Separation of Ribosomal Subunits. Before separating the
ribosomal proteins by reversed-phase HPLC, the rRNA was
2
extracted. MgCl and acetic acid were added to the ribosomes to
a final concentration of 67 mM and 67%, respectively, and then
the entire solution was incubated on ice for 1 h. The mixture was
centrifuged at 10 000 rpm for 10 min at 4 °C in a Sorvall SS-34
rotor. The procedure was repeated once more with the pellet.26
The supernatants were combined and concentrated in the speed-
vac. The proteins were redissolved in 3% acetic acid, centrifuged
at 10 000 rpm for 20 min at room temperature in an Eppendorf
centrifuge to remove any undissolved particles, and injected onto
a preparative reversed-phase HPLC system (L-6220 Intelligent
Pump, L-4250 UV-vis Detector, Merck-Hitachi AG, Darmstadt,
Germany). A gradient of 10-25% B in 30 min, 25-35%B in 40
min, 35-36% B in 30 min, 36-40% B in 40 min, 40-55%B in 60
min, and 55-90% B in 40 min was run at 2 mL/ min (A ) 0.1%
TFA, B ) 80% acetonitrile/ 0.08% TFA), using a C18 preparative
column (250 × 21 mm, Nucleosil 100-12 µm, Macherey-Nagel
AG, Oensingen, Switzerland). The absorbance was measured at
2
20 nm. Fractions of 2 mL were collected.27,28
P rotein Digestion. Fractions containing protein from the
preparative reversed-phase HPLC were dried in a speed-vac and
redissolved in 100 µL of water. For tryptic digestion, 20 µL of
protein was taken (approximately 120 µg of protein), and porcine
trypsin (sequencing grade) was added at 2% (w/ w) with respect
N-terminal Sequencing of Ribosomal P roteins. Intact
ribosomal proteins (0.5-1 µL) from the preparative HPLC separa-
tion were spotted onto methanol-wetted PVDF membrane. N-
terminal sequence analysis was directly performed on a Hewlett-
Packard G1000A protein analyzer, equipped with four cartridges.
Released PTH amino acids were analyzed on a Hewlett-Packard
HPLC series 1100 (Palo Alto, CA). Six sequence cycles were
performed according to the standard protocols provided by the
manufacturer.
to the ribosomal protein in a buffer of 100 mM NH
4 3
HCO , pH
8
.0. Digestion was performed for 48 h at 37 °C in an Eppendorf
shaker and stopped by addition of 5 µL of formic acid. The
solutions were dried in the speed-vac to remove the volatile buffer,
resuspended in 100 µL of water, and dried again, and finally, the
peptides were redissolved in 50 µL of water and stored at -20
°
C.
P eptide Degradation. The peptide degradation was carried
out with a prototype device which was designed to deliver reagents
in the gas phase (N-methylpiperidine (NMP) or N-ethylmorpholine
Database Searching. N-terminal sequences were searched
against the Swissprot and trEMBL databases using the FASTA
program. Peptide mass fingerprinting searches were carried out
using either the MassSearch or PeptideSearch algorithms.8,9
Database searching using the chemical tag data was carried out
with MassDynSearch. The program searches the SwissProt and
trEMBL by peptide mass after N- or C-terminal degradation
(enzymatic or chemical). The algorithm was developed in col-
laboration with the group of Professor Gaston Gonnet (Computa-
tion Biology Research Group, ETH Z u¨ rich) and a detailed
description of the algorithm and its application to exopeptidase-
based ladder searching has been published.19 Information for
obtaining the code for MassDynSearch is available at the web
site http:/ / cbrg.inf.ethz.ch. MS/ MS searches were carried out
using the Sequest program and the nonredundant database.14
(
NEM) and trifluoroacetic acid (TFA)) at low pressure to 10
samples, in parallel, in a thermostatically controlled block.
Standard peptides or tryptic digests (usually 1-20 pmol in 0.1%
TFA) were loaded slowly (2-5 µL/ min) onto PTFE membranes
with embedded C18 reversed-phase material in a multiple sample
loader and washed with 50 µL of 0.1% TFA. Nitrogen gas at 0.5
bar pressure was used to deliver either gas or liquid reagents.
The ladder generating steps were performed at 40 °C. The
degradation steps were carried out as follows: (a) purge with N
for 10 min at 1.0 bar, (b) NMP base (gas) delivery for 3 min at
.5 bar, (c) load 2 µL of 50 mM TATG in 1% NEM (cycle 1 and 2)
2
0
(
25) Markwell, M. A.; Haas, S. M.; Bieber, L. L.; Tolbert, N. E. Anal. Biochem.
9 7 8 , 87, 206-10.
26) Hardy, S. J.; Kurland, C. G.; Voynow, P.; Mora, G. Biochemistry 1 9 6 9 , 8,
897-905.
1
(
2
RESULTS AND DISCUSSION
(
(
27) Kamp, R. M.; Wittmann-Liebold, B. FEBS Lett. 1 9 8 4 , 167, 59-63.
28) Kamp, R. M.; Bosserhoff, A.; Kamp, D.; Wittmann-Liebold, B. J. Chromatogr.
Ladder Sequencing Chemistry. Peptide sequencing can be
1
9 8 4 , 317, 181-92.
carried out by generating a series of degradation fragments so
1008 Analytical Chemistry, Vol. 72, No. 5, March 1, 2000

S c h e m e 1 . Ge n e ra l Ou t lin e o f t h e De g ra d a t io n
Ch e m is t ry a n d Re a g e n t s Us e d
that the sequence can be read out by sequentially measuring the
mass differences between the products starting from the parent
mass. The ladder can be generated by either chemical or
enzymatic means and was first shown in the early 1980s15 using
carboxypeptidase digestion of a small enterotoxin. The develop-
ment of MALDI mass spectrometers made this method very
attractive, and the method was rediscovered. Single purified
peptides were digested with either amino- or carboxypeptidases
to generate N- or C-terminal sequence ladders which were then
analyzed by MALDI-MS20 or by a modified Edman degradation
to generate ladders.16,17,18 Unfortunately, the exopeptidase ap-
Fig u re 1 . Comparison of the degradation reagents. (A) MALDI-
TOF spectrum of the peptide C21W after two rounds of degradation
with thiobenzoylthioglycollic acid. * indicates the peak is a byproduct.
proach is limited to single peptides and the outcome depends
strongly on the sequence specificity of the exopeptidase being
used. The length of the ladder and especially its starting point is
very variable and hence is not practical for use with peptide
mixtures. Chemical methods are to be preferred since the number
of amino acids removed can be easily controlled, greatly simplify-
ing spectral interpretation. However the Edman type isothiocy-
anate degradation is not easy to carry out with low picomole
amounts of material since the peptides must either be immobilized
on a support, which must be washed after the coupling step to
remove excess reagent,16 or the peptides must be aliquoted and
added back after each cycle when using the volatile reagents.17
We therefore explored the use of alkyl alkoxydithioformates
for peptide degradation as had been first demonstrated by Kenner
and Khorana in 1952.29 We tested three reagents for their
suitability as ladder-generating reagents (see Scheme 1); reagents
A and B are commercially available, and C was synthesized as
described by Mross and Doolittle.21 The reagents were coupled
in the presence of either N-ethylmorpholine (NEM) or N-
methylpiperidine (NMP). Other bases, such as N,N-dimethylami-
no(pyridine), trimethylamine, and pyridine, were found to be
effective but were not used due to their unpleasant smell. The
(B) As above but using the reagent thioacetylthioethane. (C) As above
but using the reagent thioacetylthioglycollic acid.
coupling of all reagents was found to occur rapidly and completely
at temperatures between 40 and 80 °C and reaction times of 5-30
min. Higher temperatures led to a number of side reactions. One
reaction which occurs if air is not rigorously excluded is the S f
O substitution in the thioacetyl-peptide intermediate after the
coupling reaction which leads to blocking of the degradation since
the acetyl-peptide cannot be cleaved. The thioacetyl-peptide
derivative is fairly stable; however, if water is present during the
cleavage step with the TFA, the back reaction, cleavage to
thioacetate, competes with the release of the thioazolinone amino
acid. This is the basis of the generation of the ladder sequence.
Figure 1 shows the mass spectra obtained from the peptide C21W
immobilized on reversed-phase material after two cycles of
degradation with each of the three reagents.
Reagent A, thiobenzoylthioglycollic acid (TBTG), reacted well
and gave the two expected degradation products (-W and -WF);
however, there were also a number of side products whose
formation could not be avoided and that we could not remove
using different conditions (Figure 1A). The main problem with
(
29) Kenner, G. W.; Khorana, H. G. J. Chem. Soc. 1 9 5 2 , 2076-81.
Analytical Chemistry, Vol. 72, No. 5, March 1, 2000 1009

reagent A is that it causes considerable loss of more hydrophilic
peptides from the reversed-phase support, probably by a bulk
elution effect. Reagent B, thioacetylthioethane (TATE), allows only
the removal of the N-terminal amino acid, after which no further
cleavage is seen to occur (Figure 1B). This was found to be due
to the buildup of hydrophobic salt (base-TFA) which buffers
subsequent base additions, lowering the coupling efficiency and
peptide loss. The third reagent C, thioacetylthioglycollic acid
(Figure 3). Forty-seven fractions were obtained. Each fraction was
analyzed on an automated Edman sequencer to obtain the
N-terminal sequence(s) of the protein(s) present, and a MALDI
spectrum of the intact proteins was measured. An aliquot of each
HPLC fraction was then digested with trypsin and split into two
for either MS/ MS analysis on an electrospray ion trap mass
spectrometer or for chemical tagging in the prototype instrument
described above. Figure 4 shows the results obtained with fraction
36. Since several proteins were observed to be present in the
fraction by MALDI-MS (Figure 4A), the mixture of sequences
obtained from N-terminal sequencing was impossible to interpret.
There were not many peptides present, and the digest was
obtained from a mixture of proteins, so no clear identification could
be obtained by protein mass fingerprinting (Figure 4B). The partial
degradation approach gave three clear tag sequences (Figure 4C)
which matched the E. coli 50S subunit ribosomal proteins L1 and
L11 as well as the 30S subunit protein S7. This agreed well with
the masses obtained by MALDI analysis of the intact proteins as
well as the results obtained by MS/ MS analysis of the digest
mixture (Figure 4D) and database searching using Sequest.
The results are summarized in Table 1. The forty-seven HPLC
fractions each contained between 1 and 7 proteins, giving a total
of 106 proteins (from 62 different genes), since many proteins
occurred in more than one fraction. All of the proteins S1-21 in
the 30S subunit were found to be present as well as all of the 50S
proteins, L1-36, except for L8, L9, and L21. However, L8 does
not really existsit is a complex of L10, L7/ 12, all of which were
identified. Five ribosome-associated proteins were identified and
confirmed by manual interpretation of the MS/ MS data (Fraction
15, P52098 yaeO, a hypothetical open reading frame product;
Fraction 38, P21507, RNA helicase and P07012, peptide chain
release factor 2; Fraction 39, P07019, GTP binding export factor;
and Fraction 43, P52084, another hypothetical open reading frame
product). The ‘Gold Standard’ for the comparison of methods were
the twenty or so MS/ MS spectra obtained per fraction which were
used to search the databases automatically using the Sequest
program. Proteins identified by less than 5 independent MS/ MS
spectra searches were confirmed by manual sequence interpreta-
tion of the data.
(
TATG), is the most hydrophilic of the series and gives an
excellent ladder with very few side reactions (Figure 1C). A
considerable amount of Na+ adduct was found as well as a
significant amount of uncleaved product, especially at large
sterically hindered residues. This was subsequently corrected by
adding a longer cleavage step after the final coupling reaction and
including a final water wash of the peptides on the membrane. A
series of experiments using standard peptides were carried out
to determine the sensitivity level of the sequencing method. The
chemistry described in the Experimental Section worked well
without modification from the mid femtomole (100 femtomole of
C21W gave a clear signal after three cycles) to the low nanomole
range. The limiting factors are sample handling and the final
elution volume from the support.
Construction of a Sequencer P rototype. A prototype se-
quencer was constructed to allow multiple samples to be analyzed
and to ensure reproducibility of reagent delivery (Figure 2). Five
reagent/ solvent bottles are connected to two 6-way rotary valves
the sixth position is used to flush the lines between deliveries.
Nitrogen is delivered from a pressure-reducing valve via the first
rotary valve to a selected bottle to pressurize it before the second
rotary valve is opened to allow the chemical to be delivered to
the sample block. It is important that all the components are either
PEEK or made of glass since TFA is very aggressive and will
corrode most materials. The entire bottle assembly is contained
with an explosion/ TFA proof chamber for safety reasons. One-
way valves are installed on all the lines to prevent TFA seeping
back to the rotary and pressure-reducing valves. The reagent flow
is directed from the bottles to either waste or to a stainless steel
block. The block consists of three metal plates, each with a heating
element inside, which sandwich a Teflon membrane containing
C-18 reversed-phase particles (Empore). In the version shown in
Figure 2, 10 positions have been drilled in the bottom two plates
which serve as sample-loading and reaction chambers. The protein
digests are acidified and then loaded into the funnel-shaped
vessels. The top plate is then attached, and the digests are slowly
pushed by gas pressure through the Empore membrane. The
membrane is then washed with 0.1% TFA to remove any buffers
prior to starting the degradation steps. The block temperature is
maintained at a constant value that is preset 15 min before starting
the degradation.
The relative effectiveness versus the time consumption of each
method is highlighted below. The most effective method is clearly
the use of MS/ MS data in conjunction with a database searching
program such as Sequest. One hundred and six proteins could
be identified in a total sample preparation/ data accumulation/
database searching time of 30 h. Edman degradation was the least
effective with only 30 proteins identified in 450 h (cost-effective-
ness not even considered). The most interesting comparison is
between peptide mass fingerprinting and chemical tagging. The
partial degradation method using unseparated protein digests
could identify 50 proteins in a first-pass search (using only 1-3
very clear sequence tags and the peptide masses) and a further
30 in a second-pass search using the masses left over from the
first search (after removing those which matched the highest
scoring protein) with a total data accumulation and search time
of 5 h. The sample loading and degradation is very rapid and can
be carried out on a MALDI target-sized block with 50 positions
in less than 2 h. The final wash on the support desalts the samples,
and they are simultaneously eluted, using matrix in 60% acetoni-
P rotein Identification by Chemical Tag Searching. We
chose the ribosome complex from Escherichia coli as a test system
to investigate the effectiveness of the degradation procedure on
unseparated protein digests to generate sequence tags for protein
identification by database searching. The E. coli ribosome consists
of 57 proteins arranged in two subunits, the 30 S made up of
proteins S1-21 and the 50S proteins L1-36. The complex was
isolated, and the rRNA was removed before a rough separation
of the subunits was carried out by reversed-phase chromatography
1010 Analytical Chemistry, Vol. 72, No. 5, March 1, 2000

Fig u re 2 . Prototype reagent delivery device and reaction chamber. (A) The general layout of the tubing and valves is indicated. (B) The
sample holder and heating elements are shown (indicated in A). (C) Photograph of the actual prototype.
trile, 0.1% TFA, directly onto the MALDI target in a minute.
Peptide mass fingerprinting (mass accuracy first pass, 0.1 mass
units; second pass, 0.04 units) identified only 30 proteins in 3 h.
Chemical tagging is thus a viable alternative to simple peptide
mass fingerprinting. It is formally analogous to the peptide
sequence tag approach advocated by Matthias Mann in which the
peptide sequence tag is derived from manual inspection of an MS/
MS spectrum.13 The main advantages of the “Chem-Tag” approach
is that it can be used with nonsequence specific digestions such
as elastase or partial acid hydrolysis since the sequence tag is
always N-C-terminal and the length is predefined. The database
search data can include the sequence specificity of the digestor
and hence the database search has all the advantages of a peptide
fingerprint search with the added advantage of partial sequence
information. The confidence level of such a search containing
partial sequence information is on average about 1000-fold higher
than that based on mass alone (given the same mass accuracy).
A related approach using conventional Edman chemistry-derived
N-terminal sequence information from intact proteins and mass-
spectral data has been suggested,30 but this is much slower, relies
Analytical Chemistry, Vol. 72, No. 5, March 1, 2000 1011

Fig u re 3 . HPLC chromatogram of E. coli ribosomal subunits. The isolated ribosomes were injected onto a preparative reversed-phase HPLC
and 2-mL fractions were collected each minute.
Fig u re 4 . Mass Spectrometric analysis of fraction 36. (A) MALDI-TOF spectrum of the fraction before digestion. (B) MALDI-TOF spectrum of
the tryptic digest. (C) MALDI-TOF spectrum of the tryptic digest after ladder sequencing. (D) ESI-MS/MS spectrum of a peptide at m/z 380.5
(2+ ion).
on only a single sequence rather than multiple short sequences,
and is much less stringent.
fingerprinting alone. Data extraction using chemical tag is much
simpler than that from MS/ MS data and is much less time
consuming in terms of data accumulation and database searching
CONCLUSIONS
(
due to the predefined directionality and starting point). The main
We have demonstrated that the use of multiple short chemi-
cally generated sequence tags is much more effective in terms of
protein identification as well as confidence level than peptide mass
advantage of Chem-Tag searching is that it can be directly applied
to a DNA sequence (especially EST libraries) with much more
confidence than simple mass fingerprinting and much less effort
than MS/ MS peptide-tag searching. The ribosomal analysis
emphasizes this advantage in the case of small proteins that
(
30) Wilkins, M. R.; Ou, K.; Appel, R. D.; Sanchez, J. C.; Yan, J. X.; Golaz, O.;
Farnsworth, V.; Cartier, P.; Hochstrasser, D. F.; Williams, K. L.; Gooley, A.
A. Biochem. Biophys. Res. Commun. 1 9 9 6 , 221, 609-13.
1012 Analytical Chemistry, Vol. 72, No. 5, March 1, 2000

Ta b le 1 . Id e n t ific a t io n o f P ro t e in s in t h e HP LC Fra c t io n s
N-terminal
peptide tag
protein
protein
protein
fraction
no.a
protein
mass calc.
N-terminal protein
identified by
identified byi identified j
b
c
d
e
f
(MH mono)g
+
Mass-DynSearchh Mass-Search by FASTA
ID
(av.) / obs. (av.)
sequence DB / Exp
1
2
3
4
5
50S subunit L34
50S subunit L32
50S subunit L33
50S subunit L36
50S subunit L27
5380/ 5384
MKRTFQ/ MKRTFQ (ragged)
MAVQQN/ MAVQQN
MAKGIR/ AKGIRE
MKVRAS/ MKVRAS
MAHKKA/ AHKKAG
919.6 T
50S L34
50S L32
50S L33
50S L34
50S L32
50S L33
50S L36
50S L27
6446/ 6320 (-M)
1232.6 HH
1568.5 LVS
none
6371/ 6259 (-M + CH
3
)
4364/ 4370
9124/ 8995 (-M)
1404.6 FGGE
50S L27
30S S12
1300.5 V
6
50S subunit L14
13541/ 13681
unreadable
1873.5 IG (chymo)
(
conflict in sequence)
3
0S subunit S12
13606/ 13681 (-M)
7871/ 7872
7
50S subunit L31
MKKDIHPKYE/ MKKDI
MAHKKA/ no sequence
MTDKIR/ no sequence
MKKDIHPKYE/ MKKDI
MSITKD/ no sequence
MPVIKV/ ragged N terminus
MAAKIR/ no sequence
MKRTFQPSVLK/ FQPSVL
MAAKIR/ no sequence
MPVIKV/ no sequence
MPKIKTVRGAA/ PKIKT
MAKQSM/ AKQSMK
MSRVCQ/ SRVCQV
none
none
50S L31
50S L31
5
3
0S subunit L27
0S subunit S17
9124/ 8998 (-M)
9704/ not detected
7871/ 7874
8
9
50S subunit L31
5
30S subunit S21
50S subunit L24
5
50S subunit L24
3
5
0S subunit L7/ 12 12295/ not detected
8500/ 8371 (-M)
11316/ 11184 (-M)
5380/ not detected
11316/ 11184 (-M)
8500/ 8371 (-M)
7289/ 7161 (-M)
11580/ 11449 (-M)
9006/ 8876 (-M)
13844/ 13880 (-M + CH
13844/ 13805 (-M)
10430/ 10446
13844/ 13729 (-M)
9698/ not detected
9684/ 9556 (-M)
5380/ not detected
9684/ 9555 (-M)
8986/ 8899 (-M + Acetyl)
11316/ not detected
10693/ 10694
10268/ 10139 (-M)
29860/ 29712 (-M)
10268/ 10137 (-M)
13133/ 12999 (-M)
29860/ not detected
13541/ 13537
29860/ 29678 (-M)
17603/ not detected
6541/ 6412 (-M)
6541/ 6413 (-M)
29860/ 29707 (-M)
22243/ 22245
1189.5 E
1804.1 EAA
30S S21
50 S L24
1
1
0
50 S L24
50 S L24
0S subunit L34
50S L34
1
none
50 S L24
0S subunit S21
0S subunit L35
50S L35
30S S14
50S L28
30S S11
1
2
30S subunit S14
0S subunit L28
1085.6 G
50S L28
5
1
1
3
4
30S subunit S11
30S subunit S11
3
)
MAKAPI/ ????
unreadable
none
1280.6 KG
3
0S subunit S19
30S S19
30S S11
30S S19
1
1
1
5
6
7
30S subunit S11
ACP52098, yaeO
30S subunit S20
no sequence
2478.0 pEVS
1319.8 AF
MSMNDT/ MNDTYQ
MANIKS/ ANIKSA
MKRTFQPSVLKRNR/ VLKRNR
unreadable
ACP52098
30S S20
50S L34
30S S20
5
0S subunit L34
30S subunit S20
1240.7 YLSL
3
5
0S subunit S18
0S subunit L24
30S S18
50S L25
1
1
2
8
9
0
50S subunit L25
0S subunit S15
50S subunit L2
0S subunit S15
50S subunit L19
unreadable
unreadable
unreadable
1080.7 M
1027.7 LQ
2632.2 DAND (chymo) 30S S15
922.3 Y
1315.9 (QA
50S L25
3
3
50S L19
50S L2
50S L2
50S L2
5
5
0S subunit L2
0S subunit L14
2
1
50S subunit L2
unreadable
1202.8 H+
50S L2
3
5
0S subunit S5
0S subunit L30
(+Na )
50S L30
50S L30
50S L2
50S L3
2
2
2
2
2
3
4
5
50S subunit L30
0S subunit L2
50S subunit L3
0S subunit S17
50S subunit L13
0S subunit L3
50S subunit L13
0S subunit L17
MAKTIK/ AKTIKI
913.6 AT
50S L30
50S L3
50S L30
5
MAVVKC/ no sequence
MIGLVG/ MIGLVG
MTDKIR/ TDKIRT
MKTFTA/ MKTFTA
MIGLVG/ no sequence
MKTFTA/ MKTFTA
MRHRKS/ no sequence
MRHRKS/ MRHRKS
MDKKSA/ MDKKSA
unreadable
2187.7 IFT
2187.5 IFT
50S L3
30S S17
50S L13
3
9704/ 9574 (-M)
16018/ 16017
22243/ 22252
16018/ 16019
14364/ 14365
14364/ 14363
12769/ 12769
23469/ 23240
50S L13
50S L3
50S L13
50S L17
50S L17
50S L18
30S S4
50S L13
5
2233.4 AE
50S L13
50S L17
50S L17
50S L18
30S S4
50S L13
5
2104.2 VY (C-term)
1634.4 AGDN
951.5 SG
2
2
2
6
7
8
50S subunit L17
50S subunit L18
30S subunit S4
50S L17
50S L18
1456.6 GNT
(
conflict in sequence)
5
3
5
0S subunit L22
0S subunit S3
0S subunit L18
12226/ 12226
25983/ 25828 (-M)
12769/ 12767
13099/ 12965 (-M)
11735/ 11733
12226/ 12224
2
9
0
30S subunit S13
3
5
unreadable
unreadable
1213.8 TS (C-term)
1521.7 L (C-term)
0S subunit S10
0S subunit L22
50S L22
50S L22
3
50S subunit L22
12226/ 12223
3
5
5
3
3
3
0S subunit S13
0S subunit L23
0S subunit L16
0S subunit S8
0S subunit S1
0S subunit S21
13099/ 12967 (-M)
11199/ 11197
15281/ 15312 (+CH
14126/ 13990 (-M)
61158/ 61508
8500/ not detected
12226/ not detected
9190/ 9187
3
)
30S S21
50S L22
3
1
2
50S subunit L22
unreadable
unreadable
1521.7 LA (C-term)
1685.0 KVE
50S L22
50S L22
3
5
3
3
0S subunit S16
0S subunit L23
0S subunit S21
0S subunit S8
11199/ 11193
8500/ not detected
14126/ 13988 (-M)
12226/ not detected
7273/ 7276
15281/ 15319 (+CH
13496/ not detected
10430/ not detected
7273/ 7276
14980/ 14974
14980/ 14974
15703/ 15763
14856/ 14974
30S S21
50S L22
3
50S subunit L22
5
5
5
3
0S subunit L29
0S subunit L16
0S subunit L20
0S subunit S19
3
)
30S S19
50S L29
50S L15
50S L15
3
3
3
4
50S subunit L29
0S subunit L15
50S subunit L15
MKAKEL/ MKAKEL
MRLNTL/ no sequence
unreadable
1643.2 SV
50S L19
50S L15
30S S9
50S L19
5
1989.9 LNT
3
3
0S subunit S6
0S subunit S9
30S S9
(
conflict in sequence)
18903/ not detected
17603/ 17507 (-M + Acetyl)
5
3
0S subunit L6
0S subunit S5
Analytical Chemistry, Vol. 72, No. 5, March 1, 2000 1013

Ta b le 1 (Co n t in u e d )
N-terminal
peptide tag
protein
protein
protein
fraction
no.a
protein
mass calc.
N-terminal protein
identified by
identified byi identified j
b
c
d
e
f
(MH mono)g
+
Mass-DynSearchh Mass-Search by FASTA
ID
(av.) / obs. (av.)
sequence DB / Exp
3
5
50S subunit L1
24729/ 24575 (-M)
unreadable
unreadable
1704.0 FV
50S L1
50S L1
5
3
0S subunit L11
0S subunit S5
3
14875/ 14855 (-M + CH )
17603/ 17486 (-M + Acetyl)
24729/ 24586 (-M)
3
6
50S subunit L1
1704.0 FV
1512.6 GLP
2259.0 S
1267.6 ALLA
none
50S L1
50S L11
30S S7
50S L5
30S S2
5
3
0S subunit L11
0S subunit S7
14875/ 14867 (-M + CH
20019/ 19880 (-M)
20301/ 20162 (-M)
26743/ not detected
49914/ not detected
14980/ not detected
3
)
3
3
7
8
50S subunit L5
30S subunit S2
MAKLHD/ AKLHDY
MATVSM/ ATVSM
50S L5
30S S2
50S L5
30S S2
P21507, RNA helicase
0S subunit L15
5
P07012, peptide chain release factor 2 41250/ not detected
3
4
4
4
4
9
0
1
2
3
30S subunit S2
P07019, GTP binding export factor
50S subunit L20
26743/ not detected
49787/ not detected
13496/ 13359 (-M)
26743/ 26712
17711/ 17569 (-M)
22086/ 22079
MATVSM/ ATVSM
MARVKR/ ARVKRG
MATVSM/ no sequence
unreadable
none
30S S2
30S S2
1104.5 IL
50S L20
50S L20
3
0S subunit S2
50S subunit L10
0S subunit L4
50S subunit L10
0S subunit L4
1616.9 AA (chymo)
1446.7 LAT
2561.6 DAF
1688.8 AAA
none
50S L10
50S L4
50S L10
50S L4
50S L10
50S L4
50S L10
50S L4
5
17711/ 17571 (-M)
22086/ 22061
unreadable
unreadable
5
50S subunit L10
EC2215, AC unassigned
50S subunit L10
50S subunit L7/ L12
50S subunit L7/ L12
17711/ 17574 (-M)
50S L10
11305/ 11171
44
45
46
17711/ 17574 (-M)
12295/ 12297 (-M)
12295/ 12158 (-M)
unreadable
MSITKD/ SITKD
MSITKD/ SITKD
none
none
2015.2 FG
50S L7/ L12
50S L7/ L12
50S L7/ L12
50S L7/ L12 50S L7/ L12
9
64.4 T
2014.9 FGV
244.5 A
4
7
50S subunit L7/ L12
12295/ 12199 (-M)
MSITKD/ SITKD
50S L7/ L12
1
a
The fraction number corresponding to the peaks in the HPLC chromatogram shown in Figure 3. b Protein ID gives the protein name (from
Escherichia coli) or if nonribosomal, the SwissProt accession number (AC). All proteins were identified by MS/ MS using the Sequest program and
confirmed by manual interpretation of the spectra. c The average intact protein mass was calculated from the database entry. The experimentally
d
+
determined average intact protein mass, given as average MH . The annotation (-M) indicates that the N-terminal initiator, methionine, is cleaved
off. Acetyl and CH3 indicate that a protein is posttranslationally modified by acetylation or methylation and conflict means that there are sequence
differences between various entries for the same protein. e Representation of the N termini of the protein according to the database. The
experimentally determined N-termini of the protein (by Edman degradation). When no sequence was found or the sequence was unreadable, this
f
g
is stated. The experimentally obtained monoisotopic masses of the peptides with the partial N-terminal sequence tag obtained after digestion
with trypsin and TATG degradation. Peptides generated by chymotryptic activity are indicated by (chymo), those from the C-terminal by (C-term),
and pE indicates that the N-terminal amino acid was pyroglutamate. h Protein identification using the N-terminal chemical tag by searching SwissProt
and trEMBL with the program MassDynSearch (Korostensky et al., 1998). Note that this search includes the masses of nontagged peptides as
i
well. Protein identification from the data of the tryptic peptide masses by searching SwissProt and trEMBL using the program MassSearch
j
(
James et al., 1993). Protein identification from the data obtained from Edman N-terminal degradation and searching SwissProt and trEMBL
using the program FASTA/ TFASTA (Pearson and Lipman, 1988).
generate only a few peptide masses; this is of especial concern
when one considers that the average length of a translated EST
sequence is less than 300 amino acids. Although the procedure
is effective at cleaving all amino acids, the spectra of digests often
show only three or four tags. This is due to suppression effects
which have been demonstrated to occur with MALDI-MS.31 The
removal of amino acids changes the relative ionization efficiency,
giving rise to altered relative intensities. This is clearly seen when
one compares the spectra obtained, from identical samples, by
MALDI-MS and by ESI-MS: the main peaks are usually observed
in both techniques but the tag sequences found are often
completely different. We are currently investigating the use of
degradation reagents that carry a fixed positive charge which we
hope should reduce the suppression effects.
ACKNOWLEDGMENT
Funding was provided by the Swiss National Foundation for
Scientific Research.
Received for review October 13, 1999. Accepted
December 13, 1999.
(
31) Kratzer, R.; Eckerskorn, C.; Karas, M.; Lottspeich, F. Electrophoresis 1 9 9 8 ,
9, 1910-9.
1
AC9911847
1014 Analytical Chemistry, Vol. 72, No. 5, March 1, 2000