RAPID COMMUNICATIONS IN MASS SPECTROMETRY
Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rcm.1090

Electrochemical oxidation and cleavage of peptides
analyzed with on-line mass spectrometric detection

Hjalmar P. Permentier, Ulrik Jurva

{

, Begon˜a Barroso and Andries P. Bruins*

University of Groningen, Centre for Pharmacy, A. Deusinglaan 1, 9713 AV Groningen, The Netherlands

Received 11 March 2003; Revised 8 May 2003; Accepted 8 May 2003

An on-line electrochemistry/electrospray mass spectrometry system (EC/MS) is described that
allows fast analysis of the oxidation products of peptides. A range of peptides was oxidized in
an electrochemical cell by application of a potential ramp from 0 to 1.5 V during passage of the sam-
ple. Electrochemical oxidation of peptides was found to occur readily when tyrosine was present.
Tyrosine was found to be oxidized between 0.5 and 1.0 V to various oxidation products, including
peptide fragments formed by hydrolysis at the C-terminal side of tyrosine. The results confirm ear-
lier knowledge on the mechanisms and reaction products of chemical and electrochemical peptide
oxidation. Methionine residues are also readily oxidized, but do not induce peptide cleavage. At
potentials higher than about 1.1 V, additional oxidation products were observed in some peptides,
including loss of 28 Da from the C-terminus and dimerization. The tyrosine-specific cleavage
reaction suggests a possible use of the EC/MS system as an on-line protein digestion and peptide
mapping system. In addition, the system can be used to distinguish phosphorylated from unpho-
sphorylated tyrosine residues. Four forms of the ZAP-70 peptide ALGADDSYYTAR with both,
either or neither tyrosine phosphorylated were subjected to a 0–1.5 V potential ramp. Oxidation
of, and cleavage adjacent to, tyrosine was observed exclusively at unphosphorylated tyrosine
residues. Copyright # 2003 John Wiley & Sons, Ltd.

Protein digestion and analysis of a peptide mixture by mass
spectrometry, or peptide mapping, has become a powerful
method for protein identification and characterization in
the past decade, mainly due to advances in mass spectro-
meter design and the application of new ionization techni-
ques. The digestion method of choice nowadays is usually
enzymatic, and particularly tryptic, but in specific cases che-
mical digestion is useful when enzymes do not have the right
specificity or are not desirable for other reasons.

Oxidative chemical cleavage of proteins has been

employed for a long time already: oxidative cleavage of
peptide bonds was first observed in the 1950s.

1

It was found

that the peptide bond C-terminal to tyrosine or its deami-
nated derivative phloretic acid was the target for hydrolysis
by bromine or N-bromosuccinimide (NBS); the main tyrosine
oxidation product was identified as a brominated spirodie-
none-lactone.

2

Brominating and iodinating agents are still

used for selective oxidative cleavage of tryptophyl- (by o-
iodosobenzoic acid or BNPS-skatole) and tyrosyl-(by NBS)
peptide bonds to yield protein digests.

3– 5

Other agents used

for selective protein cleavage are formic acid (cleaves
between aspartic acid and proline), hydroxylamine (C-
terminal to asparagine), CNBr (C-terminal to methionine),
and NTCB (N-terminal to cysteine).

6

Oxidative modification of amino acids, peptides and

proteins is widely studied, because of the role of oxidative
damage in disease. Peroxyacids, halogenating agents and
various (photochemical) radical species are among the
implicated oxidizing agents.

7

Some may be produced

enzymatically in vivo, e.g. by peroxidases. Currently, a lot
of interest is focused on the oxidative nitration of tyrosine and
tryptophan in proteins by peroxynitrite, which is possibly
produced in vivo from nitric oxide.

8

Nitration is considered to

damage the protein, but it may also have a cellular signalling
function.

9

The specificity and reaction products of different oxidizing

agents vary, but in general the amino acids that most readily
undergo oxidation are tyrosine, tryptophan, cysteine,
methionine, and histidine.

10– 13

In aqueous environments

tyrosine and tryptophan are easily hydroxylated, or, depend-
ing on the oxidizing agent, halogenated or nitrated. Cysteine
and methionine are converted to their sulfonic acid and
sulfone derivatives, respectively. Cystine disulfide bonds
may be cleaved, with formation of two separate cysteic
acids.

14

Electrochemistry of proteins is widely employed as a

sensitive detection technique and for the study of their redox
properties, but electrochemical modification has not received
much attention. As an alternative to chemical oxidation,
electrochemical oxidation and cleavage of peptide bonds was
reported 30 years ago.

15–17

These studies have shown that the

electrochemical oxidation products of peptides are formed in
a similar way to their chemical oxidation counterparts.

Copyright # 2003 John Wiley & Sons, Ltd.

*Correspondence to: A. P. Bruins,

University of Groningen,

Centre for Pharmacy, A. Deusinglaan 1, 9713 AV Groningen,
The Netherlands.
E-mail: [e-mail]

{

Current address: AstraZeneca, R&D Mo¨lndal, S-431 83

Mo¨lndal, Sweden.

Although the technique was recognized to have great
promise, further developments have been limited. A few
papers have been published in recent years reporting the
electrochemical cleavage of peptides: a study of adsorption of
peptides to metal electrodes,

18

and a study of electrochemical

nitration of proteins.

19

Walton and coworkers recognized the

possibilities for electrochemical digestion of proteins, but
have mainly focused on developing a specific and efficient
method for electrochemical nitration of proteins.

20–22

Electrochemical oxidation or reduction has several advan-

tages over chemical methods: it generally is faster, there is no
need to remove reagents, and the specificity and selectivity
can be controlled dynamically, to some extent, by adjusting
the cell potential. It becomes an even more useful technique
when it is used in conjunction with mass spectrometry. An
electrochemical cell can be coupled on-line with electrospray
ionization mass spectrometry (ESI-MS) and this has proven
to be a powerful instrument for studying oxidation reactions
of a variety of compounds,

23,24

including the mimicking of

in vivo oxidative drug metabolism.

25,26

The installation of an

HPLC column between the electrochemical cell and the mass
spectrometer allows for separation and detailed identifica-
tion of the oxidation products.

27

ESI-MS is obviously well

suited for identification of protein oxidation products, but
has so far only been applied in an off-line manner.

10,13,28

Reductive electrochemistry of metalloproteins on-line with
mass spectrometric detection has been reported by Johnson
et al.

29

The target for reduction was the iron-sulfur center,

not the protein backbone.

The electrospray ionization technique is also inherently

capable of oxidizing the sample at the electrospray needle
tip.

30

This phenomenon is usually limited in scale and not

efficient for complete oxidative modification of a sample
component, due to the limited current that can be achieved.

In the present study the electrochemical oxidation of

several peptides is analyzed by coupling the electrochemical
cell directly on-line with electrospray mass spectrometry
(ESI-MS). Oxidation products, including cleavage products,
are directly detected and further analyzed by collision-
induced dissociation (CID) MS/MS. Prospects for future use
as a fast on-line protein digestion and peptide mapping
system are assessed and the potential for studying protein
oxidation and identifying tyrosine phosphorylation is
demonstrated.

EXPERIMENTAL

Chemicals

The peptides angiotensin I (human), angiotensin II (human),
renin substrate tetradecapeptide (porcine), ACTH fragment
1–10 (human), and ACTH fragment 18–39 (human) were
obtained from Sigma (Zwijndrecht, The Netherlands).
b

-Endorphin (human) was a kind gift from Organon (Oss,

the Netherlands). The four phosphorylation forms of the
ZAP-70 peptide (tryptic fragment 485–496 of the human
tyrosine kinase ZAP-70) were a kind gift from AstraZeneca
(Lund, Sweden).

31

Acetonitrile (gradient grade), formic acid

(analytical grade) and nitric acid (65%, analytical grade) were
obtained from Merck. Water was purified by a Maxima Ultra-
pure water system (ELGA, High Wycombe, Bucks, UK).

Instrumentation

A Brownlee microgradient system (Brownlee Labs, Santa
Clara, CA, USA) delivered an isocratic flow of 50 mL/min to
a Coulochem 5020 electrochemical guard cell or a Coulochem
5011 analytical cell (ESA Inc., Bedford, MA, USA). The ESA
working electrode is porous graphite and all reported cell
potentials are versus a palladium reference electrode. The
solvent consisted of 50% acetonitrile, 50% water and 0.1% for-
mic acid (pH 2.8), primarily to ensure good mass spectro-
metric detection, while the formic acid is expected to
function as the main electrolyte in the electrochemical cell.
Samples were dissolved in the same solvent to a concentra-
tion of 5 mM and were injected into a 500-mL loop on a Rheo-
dyne injector (Rheodyne, Rohnert Park, CA, USA) directly in
front of the cell. Between injections the cell was flushed with
50–100% acetonitrile, or, in the case of strong adsorption
(observed in mass spectra by tailing of the peptide signal),
cleaned by flushing off-line with 3 M HNO

3

, followed by

flushing with water and re-equilibration on-line with mobile
phase.

Control of the cell potential was done by a potentiostat that

was in turn controlled by a MacLab system with Chart 3.5.7
software (AD Instruments, Castle Hill, NSW, Australia). This
allowed the applied potential to be accurately programmed
in time. The potential and current through the cell were
continuously recorded by the MacLab system.

A sample volume of 500 mL was introduced into the loop in

front the electrochemical cell. At a flow rate of 50 mL/min this
corresponds to a 10 min long sample plug. A potential ramp
was programmed starting at 1 min after injection and rising in
5 min to 1.5 V with 2 s steps of 0.01 V. Then the potential was
decreased, first to 1.4 V in 10 s, and then to 0 V with 10 s steps
of 0.2 V.

The cell outlet was connected to the TurboIonSpray source

of an API 3000 mass spectrometer (MDS-Sciex, Concord,
Ontario, Canada). The turbo heat gun was not used in the
experiments. Full-scan mass spectra of the cell effluent
were continuously recorded, with steps of 0.3 or 1 u. The
orifice voltage was set at 20–30 V to prevent up-front frag-
mentation of the peptides. Product ion spectra were obtained
by CID MS/MS, with nitrogen as collision gas and with 1 u
steps.

RESULTS

Angiotensin I was the initial test peptide (the amino acid
sequences of the peptides are listed in Table 1) used for elec-
trochemical oxidation with on-line mass spectrometric detec-
tion of its oxidation products. Oxidation of tyrosine and the
two histidines was expected to occur a priori. Figure 1 shows
the mass spectrum accumulated in the time range 3.7–
5.7 min after injection, corresponding to a cell potential of
0.8–1.4 V (see below). The spectrum was corrected for the
background signal recorded in the first 2 min after injection.
The peaks corresponding to intact angiotensin I had largely
disappeared at potentials higher than 0.8 V. The observed
oxidation products and their masses can be explained by
the pathway shown in Fig. 2. Two electrons and a proton
are initially abstracted from the phenol group of tyrosine,
converting it into a phenoxonium group.

14

This intermediate

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

1586

H. P. Permentier et al.

can then undergo further reactions to form the products num-
bered (1) to (6) in Fig. 2.

15,17

Products (2) and (3) cannot be dis-

tinguished by MS or CID MS/MS, but the quinol product (3)
is expected to be easily oxidized further to the quinone pro-

duct (4). The phenoxonium intermediate and the protonated
cyclohexadienone product (1) are also not distinguishable by
mass, but the phenoxonium-containing ion is unlikely to be
sufficiently stable to reach the mass spectrometer.

Table 1. Electrochemically oxidized peptides and their main oxidation products (the numbers correspond to those in Fig. 2).
Peptides (M) with a single tyrosine residue are represented as XYX

00

, and those with two tyrosines as XY

1

X

0

Y

2

X

00

. The molecular

weights of the peptides and products are given, calculated from the experimental m/z values

MW (Da)

Non-cleavage oxidation

products of tyrosine (1)–(4)

N-terminal cleavage

product (5)

C-terminal

cleavage

product (6)

angiotensin I,

1295.7

1293 M
2

549

762

DRVYIHPFHL

1309 Mþ14
1311 Mþ16

angiotensin II,

1045.5

1043 M
2

549

512

DRVYIHPF

1059 Mþ14
1061 Mþ16

renin substrate,

1757.9

1754 M
2
2

549 X(Y

1

2)

1240 X

0

(Y

2

þ16)X

00

a

DRVYIHPFHLLVYS

1772 M
2þ16

a

1667 X(Y

1

2)X

0

(Y

2

2)

1135 X

0

(Y

2

2)

c

1790 Mþ16þ16

a

1685 X(Y

1

þ16)X

0

(Y

2

2)

a

ACTH 18–39,

2464.2

2462 M
2

758

1722

RPVKVYPNGAEDESAE

2480 Mþ16

a

APFLEF
b

-endorphin,

1857.9

1856 M
2

—

1695 X

00

YGGFMTSEKSQTPLVTL

1873 Mþ16

b

1711 X

00

þ16

b

1889 Mþ32

b

ZAP-70 peptide,

1301.6

1299 M
2

808 X(Y

1

2)

346 X

00

ALGADDSYYTAR

1297 M
2
2

970 XY

1

(Y

2

2)

507 Y

2

X

00

1313 M
2þ14

986 X(Y

1

þ16)(Y

2

2)

509 (Y

2

2)X

00

1315 M
2þ16
1329 Mþ14þ14
1333 Mþ16þ16

bradykinin,
RPPGFSPFR

1059.6

At > 1.2 V, loss of 28 Da and
loss of C-terminal arginine

ACTH 1–10,
SYSMEHFRWG

1298.6

Adsorption to electrode

a

Mþ14 and Mþ16 were hard to distinguish because of high charge states.

b

Includes methionine oxidation (þ16, þ32 Da).

c

Internal fragment, produced by cleavage adjacent to both tyrosines.

Figure 1. Mass spectrum of oxidized angiotensin I obtained by accumulation of the mass spectra
recorded between 0.8 and 1.4 V cell potential. The resulting spectrum was then corrected for
background signal. The numbers correspond to the oxidation products shown in Fig. 2. The labelled,
experimental m/z values are 1: 324.7 (4

þ), 432.4 (3þ), 648.1 (2þ); 2 & 3: 438.4 (3þ), 657.1 (2þ); 4:

437.8 (3

þ), 656.2 (2þ); 5: 275.8 (2þ), 550.3 (1þ); 6: 255.4 (3þ), 382.3 (2þ), 763.6 (1þ). Asterisks

indicate oxidized background ions, not attributable to products from angiotensin I.

On-line MS of electrochemically oxidized peptides

1587

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

In order to determine at which potential the various

products are generated, the extracted ion chromatogram of a
particular m/z value was converted into an extracted ion
voltammogram for the linear part of the ramp, from 0 to 1.5 V,
taking into account the delay of about 1 min due to the
internal volume of the cell and the connecting tubing to the
mass spectrometer.

25

Figure 3 shows the extracted ion

voltammograms of angiotensin I and its oxidation products.
The signal intensity of unoxidized angiotensin I (solid line)
starts decreasing at 0.5 V and reaches its minimum at about
0.9 V. The five different oxidation products start to appear
between 0.4 and 0.5 V and reach their maximum intensity
between 0.7 and 0.9 V. This concurrent appearance of all five
products is consistent with their formation from a common
phenoxonium intermediate. The intensity ratios of products
1, 2 þ 3, 4 and 5 þ 6 as plotted in Fig. 3 are 1.0:0.8:0.4:0.4,
respectively. However, the actual signal intensities cannot
simply be used to calculate the relative yields of the oxidation
products: in particular the cleavage products, which have
different numbers of basic amino acids, will have different
ionization efficiencies. Assuming that these efficiencies do

not differ substantially, we determined the yield of cleavage
products to be approximately 15% of the total amount of
oxidation products.

The intensity of the unoxidized angiotensin I signal

decreased to less than 10% of the original intensity, so
oxidation is nearly complete under the present, essentially
not optimized, conditions. The intensity of the third isotope
peak of unoxidized angiotensin I (m/z 433.5 (3þ)) is plotted in
Fig. 3; this was chosen because the first isotope peak at m/z
432.9 (3þ) overlaps with the third isotope peak of the M-2
oxidation product.

The oxidation products of angiotensin I were fragmented

by CID to determine which amino acid residue(s) were
oxidized. Figure 4 shows the product ion spectra of the 549 Da
(Fig. 4(a)) and the doubly charged 1311 Da (Fig. 4(b))
products. The 549 Da product has the amino acid sequence
DRV(Y-2), the (Y-2) residue having a mass of 161 Da, 2 Da
lower than that of unoxidized tyrosine, and consistent with
the spirodienone-lactone product (5) shown in Fig. 2. Apart
from the regular a-, b- and y-ions, a fragment ion with a mass
44 Da lower than the precursor was observed. This is

Figure 2. Electrochemical oxidation scheme of a tyrosine-containing peptide, RYR

0

. R and

R

0

are the parts of the peptide N-terminal and C-terminal to the tyrosine residue,

respectively. The numbered oxidation products have the following masses: (1) RYR

0

2 Da,

(2) and (3) RYR

0

þ16 Da, (4) RYR

0

þ14 Da, (5) RY
2 Da, and (6) R

0

.

1588

H. P. Permentier et al.

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

presumably due to loss of CO

2

from the C-terminal lactone;

the resulting fragment is then expected to have the structure
of an a

4

-ion of DRVY.

The product ion spectrum of the 1311 Da oxidation product

(Fig. 4(b)) shows that it has the sequence of angiotensin I with

a 16 Da mass increase of the tyrosine residue; the residue
masses of the other amino acids have remained unchanged.
Likewise, the product ion spectra of the 1309 and 1293 Da
oxidation products only showed a mass change of the
tyrosine residue, of þ14 Da and
2 Da, respectively. These
mass changes imply the presence of tyrosine oxidation
products (4) and (1), respectively, in Fig. 2. The fragment at
m/z 760 was confirmed to have the sequence IHPFHL,
representing the peptide fragment C-terminal to tyrosine
(product (6) in Fig. 2).

The MS/MS results clearly indicated that all oxidation

products were modified at, or adjacent to, the tyrosine
residue. The other amino acids, including the two histidines,
do not give rise to oxidation products.

Angiotensin II, which lacks the C-terminal histidine and

leucine residues of angiotensin I, was used to verify the
specificity of the oxidation reactions of angiotensin I. The
same types of oxidation products were observed in both
peptides, and in roughly the same ratios. This indicates that
the reactions are indeed specific, and that the additional
histidine in angiotensin I does not appreciably influence the
oxidation reaction at the tyrosine residue.

Two additional types of oxidation products were found in

angiotensin II that were not observed for angiotensin I. The
first oxidation product, formed between 0.8 and 1.0 V (Fig. 5,
long-dashed line), has a triply charged ion at m/z 697. This is
most likely a didehydro dimer of angiotensin II, cross-linked
through dityrosine. This is consistent with the calculated
mass of the dimer, which is 2 Da lower than twice the mass of
angiotensin II. Ions corresponding to non-covalently bound
dimers were not observed for unoxidized angiotensin II.

Figure 4. Product ion spectra obtained by CID MS/MS of the angiotensin I oxidation products
observed at m/z 550 (a), and m/z 657 (2

þ) (b). Only the singly charged fragment ions are

labelled. The b* labels denote b-NH

3

ions. Labels with ‘‘

þ16’’ and ‘‘
2’’ indicate that these ions

have a mass of 16 Da higher or 2 Da lower, respectively, than expected for an unoxidized
tyrosine.

Figure 3. Extracted ion voltammograms of angiotensin I
(MW 1295.7 Da). For clarity of presentation the data were
smoothed with an 8-point window. Solid: M (m/z 433.5 (3

þ));

long dash: M-2 (m/z 432.1 (3

þ)); short dash: Mþ16 (m/z

438.3 (3

þ)); dotted: Mþ14 (m/z 437.5 (3þ)); dash-dot:

DRVY-2 (m/z 550.6 (1

þ)); dash-dot-dot: IHPFHL (m/z

763.6 (1

þ)).

On-line MS of electrochemically oxidized peptides

1589

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

The second new product was found only at voltages higher

than 1.1 V, and showed a loss of 28 Da with respect to the
oxidized peptides M
2 (Fig. 5, dotted line) and Mþ16,
presumably due to a net loss of CO from the C-terminal
carboxylic acid.

In order to determine whether the oxidation reactions and

in particular the cleavage reactions at tyrosine are reprodu-
cible in different protein ‘‘environments’’, several other
tyrosine-containing peptides were oxidized using the same
procedure. These additional peptides also contained a
number of amino acids not found in angiotensin I and II.
The amino acid sequences of the peptides are given in Table 1,
which also summarizes the oxidation products that were
observed for each peptide during application of a potential
ramp of 0–1.5 V. All tyrosine-containing peptides, except
ACTH 1–10 which was strongly adsorbed to the electrode,
yielded products with masses consistent with cleavage
C-terminal to tyrosine and with oxidation of tyrosine shown
in Fig. 2. A few expected product ions were not detected, such
as the single amino acid N-terminal cleavage fragment of
b

-endorphin and the C-terminal cleavage fragment of renin

substrate. As expected, the methionine of b-endorphin is
apparently oxidized to a sulfoxide; the Mþ32 product is then
due to a combination of tyrosine and methionine oxidation,
both residues gaining 16 Da. When two tyrosine residues are
present, both are oxidized and give rise to cleavage products,
as in renin substrate and ZAP-70 peptide. The net oxidation
spectra (oxidized minus unoxidized: the accumulated spec-
tra between 0.8 and 1.4 V were background-subtracted with
those accumulated between 0 and 0.5 V) of these two peptides
were rather complex. A large number of peaks were present,
which could be identified as various combinations of tyrosine

oxidation and cleavage products, indicated in Table 1. No
specificity or preference for oxidation of either tyrosine
residue was observed.

Two additional oxidation reactions occurred at high

voltages (above 1.1 V) in angiotensin II: loss of 28 Da, and
dimerization. Loss of 28 Da was only observed in ACTH 18–
39, while dimerization appears to take place both in ZAP-70
peptide (Fig. 6(a)) and in b-endorphin. No clues from the
amino acid sequences were found as to what determines the
occurrence of these two additional reactions in a peptide.

Bradykinin (RPPGFSPFR) was used as a control peptide

without tyrosine or any other readily oxidizable amino acids.
It did not show oxidation at voltages lower than 1.2 V, but
between 1.2 and 1.5 V two oxidation products were detected.
The first one shows a loss of 28 Da with respect to intact
bradykinin, probably the same type of product seen at high
voltages with angiotensin II and ACTH 18–39. Fragmenta-
tion of this product showed that it is the C-terminal arginine
residue that loses 28 Da. The other product was detected at
m/z 904, corresponding to bradykinin minus a single arginine
residue. Sequencing of this product by MS/MS revealed that
the C-terminal arginine was lost. Less than 10% of bradykinin
was oxidized in either way, as judged from the decrease in
signal intensity of the unoxidized peptide.

As mentioned earlier, the relative amounts of oxidation

products cannot be easily quantified. An estimate based on
the observed intensities reveals significant differences among
the peptides. The relative amount of cleavage products
compared with the total amount of oxidation products varies
widely, from less than 5% in b-endorphin to more than 50% in
some forms of the ZAP-70 peptide (see below). Various
factors, including the amino acids adjacent to tyrosine, and
the conformation of the peptide under the experimental
conditions, may play a role in determining the preferred
oxidation pathway. No effort was made to optimize the
conditions for each peptide separately. The conditions that
were initially used for angiotensin I were applied to all other
peptides to allow a direct comparison of the results. The
efficiency of oxidation at these standard conditions, calcu-
lated from the signal intensity decrease of the parent peptide,
was between 80 and 95%.

The observed specific electrochemical oxidation of tyrosine

suggested an interesting method for distinguishing peptides
with phosphorylated tyrosines from unphosphorylated ones.
The phosphorylation of the hydroxyl group of tyrosine was
expected to prevent oxidation and formation of the phenox-
onium intermediate. The ZAP-70 peptide (sequence
ALGADDSYYTAR)

31

was used to verify this assumption.

Four forms of this peptide were available, with neither, one or
both tyrosines phosphorylated. Figure 6 shows the accumu-
lated oxidation spectra of the four forms, recorded between
cell potentials of 1.0 and 1.5 V. The observed oxidation
products clearly show that cleavage occurs C-terminal to
unphosphorylated tyrosines but not C-terminal to phos-
phorylated ones (Figs. 6(a)–6(c)). The doubly phosphor-
ylated ZAP-70 peptide did not give rise to peptide oxidation
and cleavage products (Fig. 6(d)). In all four spectra a number
of peaks due to oxidation products from unidentified non-
peptide components were observed (at m/z 105, 282, 338, 387,
397, 564 and 794), indicated with asterisks in Fig. 6. Most of

Figure 5. Extracted ion voltammograms of angiotensin II
(MW 1045.5 Da). The intensity of each extracted ion was
normalized to its maximum abundance, and smoothed with
an 8-point window for clarity of presentation. Solid: M (m/z
524.3: second isotope peak of doubly charged angiotensin I);
short dash: M-2 (m/z 522.8 (2

þ)), enlarged 2.1; long dash:

2M-2 (m/z 669.9 (2

þ)), enlarged 23; dotted: M-30 (m/z

522.8 (2

þ)), enlarged 8.0.

1590

H. P. Permentier et al.

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

these were also found in varying amounts in the other
peptides listed in Table 1 (see also Fig. 1), and in experiments
where just solvent was injected and subjected to the potential
ramp (not shown). They were particularly prevalent in the
doubly phosphorylated ZAP-70 peptide (Fig. 6(d)), probably
because there is no competition for oxidation with the
peptide itself in this case.

DISCUSSION

Electrochemical oxidation of peptides occurs readily when
tyrosine is present. Tyrosine was found to be oxidized to
various products, including peptide fragments formed by
hydrolysis at the C-terminal side of tyrosine. Methionine
residues are also readily oxidized, but do not induce peptide
cleavage. The results confirm earlier knowledge on the
mechanisms and reaction products of chemical and electro-
chemical peptide oxidation, cited in the Introduction. A clea-
vage reaction very similar to the one for tyrosine has been
described for tryptophan;

4

both tryptophan and tyrosine

are readily oxidized electrochemically in aqueous solution.

32

Electrochemical cleavage of tryptophyl-peptide bonds has
only recently been observed.

19

Electrochemical cleavage at

tryptophan, but not at tyrosine, was detected in hen egg-
white lysozyme at an optimal potential of 1.3 V, and an opti-
mum pH value of 3. In addition, selective oxidation of the two
methionines present in this protein was achieved by using
various cell potentials or mobile phase acetonitrile concentra-
tions.

19

The range of peptides that were analyzed in our study is

too small to draw definite conclusions, but to our present
knowledge specific peptide cleavage takes place only when
tyrosine is present, and, in each peptide where tyrosine is
present, at least some degree of cleavage is observed. The
optimal potential range for cleavage adjacent to tyrosine in
our experimental setup was between 0.5 and 1.0 V. However,
at relatively high voltages (above 1.2 V), some oxidation
reactions take place even in the absence of tyrosine: loss of
28 Da from the C-terminus and loss of the entire C-terminal
arginine in the case of bradykinin. The mechanism by which
both of these losses may take place is unclear at present.
Neither appears to be associated with specific amino acids:
the loss of 28 Da is seen in two other peptides at high voltages,
namely angiotensin II and ACTH 18–39, both of which have a
C-terminal phenylalanine residue. The internal arginine and
phenylalanine residues of angiotensin I and other peptides,
and even the C-terminal arginine of the ZAP-70 peptide, do
not induce cleavage as in bradykinin. The precise amino acid
sequence of the peptide and the possibility of competing
oxidation reactions may well dictate the efficiency of
oxidation of the C-terminal residue.

Not all peptides were oxidized successfully; ACTH

fragment 1–10, the only peptide in our study containing a
tryptophan residue, was very strongly adsorbed to the
electrodes of the electrochemical cell, even before application
of the oxidizing potential. No meaningful results could
therefore be obtained for this peptide. It is not clear whether
the adsorption is due to the presence of tryptophan, or if it is

Figure 6. Accumulated mass spectra, recorded between cell potentials of 1.0 and 1.5 V, of
differently phosphorylated ZAP-70 peptides: (a) ALGADDSYYTAR, (b) ALGADDSYpYTAR,
(c) ALGADDSpYYTAR, and (d) ALGADDSpYpYTAR, where pY is a phosphorylated tyrosine
residue. M

ox

denotes the oxidation products M

2, Mþ14 and Mþ16. Asterisks indicate

background ions.

On-line MS of electrochemically oxidized peptides

1591

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592

coincidental. More tryptophan-containing peptides have to
be tested before any conclusions can be made regarding the
oxidation of tryptophan in peptides. Except for tryptophan
and cysteine all naturally occurring amino acids are present
in the peptides tested thus far.

Tyrosine-containing peptides and proteins can be cross-

linked in various ways upon oxidation.

33,34

Cross-linked

dimer formation of peptides was detected in small amounts
in angiotensin II, unphosphorylated ZAP-70 peptide and b-
endorphin. Peaks attributable to triply charged didehydro
dimers were detected in their net oxidation spectra. How-
ever, in the last two peptides the distinction between covalent
bond formation and non-covalent complex formation is not
clear. Non-covalently bound complexes were also detected in
the unoxidized samples. Chromatographic separation will
have to be employed to determine if covalently linked dimers
are indeed present.

The number of possible products in peptides with several

oxidizable amino acids (e.g. as observed in ZAP-70 peptide
and b-endorphin) quickly makes interpretation of the mass
spectra very difficult. Multiple charging of the products
having several basic residues, as in the angiotensin analogs,
further complicates the analysis. Therefore, separation of the
oxidation products by on-line HPLC/MS is desirable.

Although the largest peptide used in this study was only

2.5 kDa, the EC/MS system holds interesting prospects for its
use as an on-line peptide mapping system. The extension of
the method to large peptides and proteins is a future
challenge. Apart from the need to include LC separation of
the products for identification, several points have to be
addressed, including the accessibility of all amino acid
residues to the electrode, possible adsorption to the electrode
and possible inter- and intramolecular side reactions, such as
cross-linking.

Electrochemical digestion may run into some of the same

drawbacks that the chemical method has compared with
enzymatic digestion, such as low yield, lack of specificity and
undesirable side reactions.

6

However, the many adjustable

parameters of the electrochemical method allow these
problems to be addressed and presumably overcome.

An interesting consequence of the specific tyrosine oxida-

tion is that it allows a distinction between phosphorylated
and unphosphorylated forms of peptides or proteins by
analysis of the products formed by oxidative cleavage.
Phosphorylation of tyrosine, but not of serine or threonine,
can be detected in this way. Identification and location of
protein phosphorylation is an important goal in proteomics,
because of the central regulatory role played by phosphor-
ylation in the regulation of protein activity. A number of
chemical and mass spectrometric techniques have been
developed to improve the detection of phosphorylation.
Electrochemistry may become a useful alternative, for the
same reasons of speed and connectivity already discussed for
the application to peptide mapping.

In conclusion we have shown that electrochemistry

coupled on-line with electrospray mass spectrometry has
the potential to become a very useful tool for peptide analysis.
The specific oxidation of, and cleavage adjacent to, tyrosine
shows prospects for the use of EC/MS as a fast on-line
method for protein digestion and peptide mapping, and for
the detection of tyrosine phosphorylation.

REFERENCES

1. Corey EJ, Haefele LF. J. Am. Chem. Soc. 1959; 81: 2225.
2. Schmir GL, Cohen LA, Witkop B. J. Am. Chem. Soc. 1959;

81

: 2228.

3. Wilson JG, Cohen LA. J. Am. Chem. Soc. 1963; 85: 564.
4. Burstein Y, Patchornik A. Biochemistry 1972; 11: 4641.
5. Mahoney WC, Hermodson MA. Biochemistry 1979; 18: 3810.
6. Kellner R, Lottspeich F, Meyer HE (eds). Microcharacteriza-

tion of Proteins (2nd edn). Wiley-VCH: Weinheim, 1999.

7. Rosen GM, Pou S, Ramos CL, Cohen MS, Britigan BE.

FASEB J. 1995; 9: 200.

8. Ischiropoulos H, Al-Mehdi AB. FEBS Lett. 1995; 364: 279.
9. Greenacre SAB, Ischiropoulos H. Free Radical Res. 2001; 34:

541.

10. Chowdhury SK, Eshraghi J, Wolfe H, Forde D, Hlavac AG,

Johnston D. Anal. Chem. 1995; 67: 390.

11. Berlett BS, Levine RL, Stadtman ER. J. Biol. Chem. 1996; 271:

4177.

12. Nightingale ZD, Lancha AH, Jr, Handelman SK, Dolni-

kowski GG, Busse SC, Dratz EA, Blumberg JB, Handelman
GJ. Free Radical Biol. Med. 2000; 29: 425.

13. Kotiaho T, Eberlin MN, Vainiotalo P, Kostiainen R. J. Am.

Soc. Mass Spectrom. 2000; 11: 526.

14. Lund H, Baizer MM (eds). Organic Electrochemistry: An

Introduction and Guide (3rd edn). Marcel Dekker: New York,
1991.

15. Iwasaki H, Cohen LA, Witkop B. J. Am. Chem. Soc. 1963; 85:

3701.

16. Iwasaki H, Witkop B. J. Am. Chem. Soc. 1964; 86: 4698.
17. Farber L, Cohen LA. Biochemistry 1966; 5: 1027.
18. MacDonald SM, Roscoe SG. Electrochim. Acta 1997; 42: 1189.
19. Walton DJ, Richards PG, Heptinstall J, Coles B. Electrochim.

Acta 1997; 42: 2285.

20. Richards PG, Coles B, Heptinstall J, Walton DJ. Enzyme

Microb. Technol. 1994; 16: 795.

21. Walton DJ, Heptinstall J. Prep. Biochem. Biotechnol. 2000; 30:

1.

22. Kendall G, Cooper HJ, Heptinstall J, Derrick PJ, Walton DJ,

Peterson IR. Arch. Biochem. Biophys. 2001; 392: 169.

23. Zhou FM, Van Berkel GJ. Anal. Chem. 1995; 67: 3643.
24. Deng HT, Van Berkel GJ. Electroanalysis 1999; 11: 857.
25. Jurva U, Wikstro¨m HV, Bruins AP. Rapid Commun. Mass

Spectrom. 2000; 14: 529.

26. Jurva U, Wikstro¨m HV, Weidolf L, Bruins AP. Rapid Com-

mun. Mass Spectrom. 2003; 17: 800.

27. Jurva U, Wikstro¨m HV, Bruins AP. Rapid Commun. Mass

Spectrom. 2002; 16: 1934.

28. Petersson A-S, Steen H, Kalume DE, Caidahl K, Roepstorff

P. J. Mass Spectrom. 2001; 36: 616.

29. Johnson KA, Shira BA, Anderson JL, Amster IJ. Anal. Chem.

2001; 73: 803.

30. Van Berkel GJ, Zhou FM. Anal. Chem. 1995; 67: 2916.
31. Miliotis T, Ericsson PO, Marko-Varga G, Svensson R,

Nilsson J, Laurell T, Bischoff R. J. Chromatogr. B 2001;
752

: 323.

32. Harriman A. J. Phys. Chem. 1987; 91: 6102.
33. Burzio LA, Waite JH. Protein Sci. 2001; 10: 735.
34. Michon T, Chenu M, Kellershon N, Desmadril M, Gue´guen

J. Biochemistry 1997; 36: 8504.

1592

H. P. Permentier et al.

Copyright # 2003 John Wiley & Sons, Ltd.

Rapid Commun. Mass Spectrom. 2003; 17: 1585–1592