Pyrrolization and antioxidant function of proteins following oxidative stress

Article abstract of DOI:10.1021/tx000215m

The consequences of oxidative stress on microsomal proteins were analyzed by studying their pyrrolization and the antioxidative activity of the modified proteins produced. The microsomal system consisted of freshly prepared trout muscle microsomes, which were oxidized in the presence of 5 μM Cu²⁺, 1 mM Fe³⁺/5 mM ascorbate, or 1 mM Cu²⁺/10 mM H₂O₂. Pyrroles on proteins were detected by forming Ehrlich adducts with p-(dimethylamino)benzaldehyde and by determination of ε-N-pyrrolylnorleucine (Pnl) by capillary electrophoresis. Their antioxidative activity was studied by testing two model pyrrolized proteins (dimeric and monomeric modified bovine serum albumin: DBSA and MBSA, respectively), which were produced in the reaction of BSA and 4,5(E)-epoxy-2(E)-heptenal. These proteins were assayed at a concentration of 10-40 μg/mL, which was selected because at this concentration both DBSA and MBSA had a concentration of Pnl similar to the Pnl concentration produced in oxidized microsomes. Both DBSA and MBSA significantly (p < 0.05) protected against lipid peroxidation, assessed by the formation of thiobarbituric acid reactive substances (TBARS), and protein damage, evaluated by amino acid analysis, for the three systems assayed, and this protection was always higher than that exhibited by BSA, which was used as control. The order of effectiveness was DBSA > MBSA > BSA and was parallel to the Pnl content in the assayed proteins. These results suggest that antioxidative activity of BSA may also be related to its ability to react with lipid oxidation products and to produce modified BSA with antioxidative activity. This mechanism may also be contributing to the antioxidative activity exhibited by many proteins.

Full text of DOI:10.1021/tx000215m

582
Chem. Res. Toxicol. 2001, 14, 582-588
P yr r oliza tion a n d An tioxid a n t F u n ction of P r otein s
F ollow in g Oxid a tive Str ess
Francisco J . Hidalgo, Manuel Alaiz, and Rosario Zamora*
Instituto de la Grasa, Consejo Superior de Investigaciones Cient´ıficas,
Avenida Padre Garcı´a Tejero 4, 41012-Sevilla, Spain
Received October 10, 2000
The consequences of oxidative stress on microsomal proteins were analyzed by studying their
pyrrolization and the antioxidative activity of the modified proteins produced. The microsomal
system consisted of freshly prepared trout muscle microsomes, which were oxidized in the
presence of 5 µM Cu2+, 1 mM Fe3+/5 mM ascorbate, or 1 mM Cu2+/10 mM H2O2. Pyrroles on
proteins were detected by forming Ehrlich adducts with p-(dimethylamino)benzaldehyde and
by determination of ꢀ-N-pyrrolylnorleucine (Pnl) by capillary electrophoresis. Their antioxidative
activity was studied by testing two model pyrrolized proteins (dimeric and monomeric modified
bovine serum albumin: DBSA and MBSA, respectively), which were produced in the reaction
of BSA and 4,5(E)-epoxy-2(E)-heptenal. These proteins were assayed at a concentration of 10-
40 µg/mL, which was selected because at this concentration both DBSA and MBSA had a
concentration of Pnl similar to the Pnl concentration produced in oxidized microsomes. Both
DBSA and MBSA significantly (p < 0.05) protected against lipid peroxidation, assessed by the
formation of thiobarbituric acid reactive substances (TBARS), and protein damage, evaluated
by amino acid analysis, for the three systems assayed, and this protection was always higher
than that exhibited by BSA, which was used as control. The order of effectiveness was DBSA
> MBSA > BSA and was parallel to the Pnl content in the assayed proteins. These results
suggest that antioxidative activity of BSA may also be related to its ability to react with lipid
oxidation products and to produce modified BSA with antioxidative activity. This mechanism
may also be contributing to the antioxidative activity exhibited by many proteins.
can produce many of the above cited major interrelated
derangements of cell metabolism.
In tr od u ction
Reactive oxygen species (ROS)1 and reactive nitrogen
species (RNS) are unavoidably and continuously pro-
duced under normal metabolic conditions because they
are critically involved in normal metabolism, especially
in inter- and intracellular signaling (1). However, ROS/
RNS can directly modify DNA, lipids, and proteins, which
results in mutations, strand breaks, aldehyde formation,
LDL oxidation, and alterations in enzyme activities and
signal transduction pathways (2-4). To minimize these
deleterious effects, aerobic organisms have evolved elabo-
rate antioxidant systems, which consist of low molecular
weight antioxidants and antioxidant enzymes (5, 6).
Under physiological conditions, the range of antioxidant
defenses available within the cells is adequate to protect
them against oxidative damage. However, the protective
balance can be lost because of overproduction of free
radicals that overwhelm the antioxidant defenses or by
inadequate intake of nutrients that contribute to the
defense system. This imbalance, namely oxidative stress,
As a part of the cell defense system, previous studies
from this laboratory have pointed out that oxidative
stress is also, in some extent, delayed by the oxidized
lipid/amino acid reaction products (OLAARPs) that are
produced as an ultimate step in the lipid peroxidation
process, suggesting that this might be a protective
mechanism by which highly toxic aldehydes are blocked
and endogenous antioxidants are generated (7). As a
continuation of those studies, the present investigation
was undertaken to analyze the potential physiological
role of OLAARP-containing proteins by studying their in
vitro formation as a consequence of oxidative stress and
their antioxidant activity. Although their physiological
functions are unknown at present, OLAARP-containing
proteins have been detected in the sera of healthy
individuals (8), and they are likely to be the most
important source of OLAARPs in living beings.
Among these OLAARPs, pyrroles have been shown to
be produced in the reaction of amino groups with many
lipid oxidation products, including 4,5-epoxy-2-alkenals
(9), 4-hydroxy-2-alkenals (10, 11), unsaturated epoxyoxo
fatty acids (12), and lipid hydroperoxides (13). Therefore,
protein pyrrolization appears to be an important step in
the lipid peroxidation process initiated by the oxidative
stress. In addition, pyrrolized proteins have been found
in vivo in human plasma and vasculature as levuglandin
E2-protein adducts (14).
* To whom the correspondence should be addressed. Phone: +34
95 461 1550. Fax: +34 95 461 6790. E-mail: rzamora@cica.es.
1
Abbreviations: BHT, butylated hydroxytoluene; BSA, bovine
serum albumin; DBSA, dimeric modified BSA; DMAB, p-(dimethy-
lamino)benzaldehyde; EH, 4,5(E)-epoxy-2(E)-heptenal; HPCE, high-
performance capillary electrophoresis; HPLC, high-performance liquid
chromatography; KRPB, Krebs-Ringer phosphate buffer; MBSA, mon-
omeric modified BSA; OLAARPs, oxidized lipid/amino acid reaction
products; PI, protection index; Pnl, ꢀ-N-pyrrolylnorleucine; RNS,
reactive nitrogen species; ROS, reactive oxygen species; TBARS,
thiobarbituric acid reactive substances.
10.1021/tx000215m CCC: $20.00 © 2001 American Chemical Society
Published on Web 05/21/2001

Antioxidant Function of Proteins
Chem. Res. Toxicol., Vol. 14, No. 5, 2001 583
acid conditions as described previously (25). Briefly, incubated
microsomes (2 mL) were treated with 320 µL of a 2% solution
of DMAB in 3.5 N HCl:ethanol (4:1), and the resulting solution
was incubated at 45 °C for 30 min. The protein was precipitated
with 1.2 mL of 30% trichloroacetic acid at 0 °C for 1 h and, then,
centrifuged at 2250g for 15 min. The pellet was washed with
ethanol:ethyl acetate (1:1), and, finally, treated with 2 mL of 6
M guanidine HCl containing 20 mM potassium phosphate:
trifluoroacetic acid, pH 2.3, for 30 min at 37 °C with vortexing.
This treatment partially solubilized the pyrrolized proteins and
their absorption spectra could be obtained.
Pnl was determined as described previously (21). Briefly,
incubated samples (15 mL) were centrifuged at 100000g for 1 h
and the resulting pellets, which contained the oxidized mi-
crosomes, were suspended in 1 mL of deionized water and
submitted to basic hydrolysis with 2 N NaOH at 120 °C for 18
h. Pnl was determined by capillary electrophoresis after deriva-
tization with diethyl ethoxymethylenemalonate and using ho-
moarginine as standard.
To test antioxidative activity of pyrrolized proteins, two
model purified proteins were prepared from a bovine
serum albumin/4,5(E)-epoxy-2(E)-heptenal (BSA/EH) mix-
ture, which was selected for a variety of reasons. BSA
has been suggested to act as antioxidant (15, 16) and to
protect apolipoprotein B from damage by sequestering
lipid peroxidation products (17). In addition, it represents
a major class of animal proteins; it is free of prosthetic
groups and other complicating factors; and its primary,
secondary, and tertiary structure has been well charac-
terized in the literature (18). Among the many different
aldehydes which can be formed during lipid peroxidation
and that pyrrolize proteins, EH was selected because it
is highly reactive with the lysine amino groups (19), and
because the mechanisms involved in these reactions are
now well-known (9, 20).
Ma ter ia ls a n d Meth od s
Model pyrrolized microsomes were obtained by incubating
microsomes (50 µL of the stock solution suspended in 5 mL of
KRPB) at 37 °C for 3 h in the presence of 1 mM EH.
Effect of OLAARP -Con ta in in g P r otein s on Oxid a tive
Da m a ge In d u ced by ROS in Micr osom es. Microsomes (50
µL of the stock solution) were suspended in 5 mL of KRPB and
incubated at 37 °C in the presence of the ROS generating system
and the protein tested as antioxidant. At different incubation
times, lipid peroxidation was assessed by the formation of
thiobarbituric acid reactive substances (TBARS) using a pro-
cedure of Kosugi et al. (26), which was modified as described
previously (7). At the end of the incubation period, incubated
samples (1.8 mL) were hydrolyzed with 6 N HCl for 20 h (110
°C) and, then, submitted to amino acid analysis by high-
performance liquid chromatography (HPLC) using a previously
described procedure (27).
Ma ter ia ls. EH was prepared from 2(E),4(E)-heptadienal
analogously to 4,5(E)-epoxy-2(E)-decenal (13). 2(E),4(E)-Hepta-
dienal and 3-chloroperoxybenzoic acid were purchased from
Aldrich Chemical Co. (Milwaukee, WI). Essentially fatty acid
free BSA, Sephacryl S-200-HR, PD-10 columns packed with
Sephadex G-25 medium, and butylated hydroxytoluene (BHT)
were purchased from Sigma Chemical Co. (St. Louis, MO). Other
reagents and solvents used were analytical grade and were
purchased from reliable commercial sources.
P r ep a r a tion of OLAARP -Con ta in in g P r otein s. BSA (30
mg) was dissolved in 3 mL of 50 mM sodium phosphate buffer,
pH 7.4, and treated with 10 mM EH. The mixture was allowed
to react in a closed vessel for 19 h at 37 °C, and, then, the brown
solution was fractionated chromatographically on a Sephacryl
S-200-HR column. One milliliter of the protein solution was
injected in the column (1 × 56 cm) and eluted with 50 mM
sodium phosphate buffer, pH 7.4, at a flow rate of 11 mL/h. Five-
hundred microliter fractions were collected and tested for
protein by absorption at 280 nm. This sequence was used
repeatedly until the whole sample was fractionated. Fractions
corresponding to monomeric and dimeric modified BSA (MBSA
and DBSA, respectively) were pooled, desalted using a PD-10
column, and freeze-dried. This procedure yielded 16.2 mg of
MBSA and 4.7 mg of DBSA. OLAARP presence in MBSA and
DBSA was analyzed by determining OLAARP product ꢀ-N-
pyrrolylnorleucine (Pnl) by high-performance capillary electro-
phoresis (HPCE) after basic hydrolysis as described previously
(21).
P r epar ation an d Oxidation of Tr ou t Mu scle Micr osom es.
Muscle microsomes from freshly killed rainbow trout (Salmo
gairdnerii) were prepared by differential centrifugation accord-
ing to a procedure of Parkin and Hultin (22) as described
previously (21). Washed microsomes were dispersed in Krebs-
Ringer phosphate buffer (KRPB) (23), and protein concentration,
measured according to Bradford’s method using BSA as stan-
dard (24), was adjusted to 3.0 mg/mL for experiments.
Microsomes (50 µL of the stock solution) were suspended in
5 mL of KRPB and incubated at 37 °C in the presence of a ROS
generating system. Three different procedures were used to
generate ROS. Procedure 1 consisted in the treatment with 5
µM CuCl2. Procedure 2 consisted in the treatment with 1 mM
FeCl3 and 5 mM ascorbic acid. Procedure 3 consisted in the
treatment with 1 mM CuCl2 and 10 mM H2O2. These systems
were very efficient to produce lipid and protein oxidation in
incubated microsomes, and they were used previously to test
antioxidant activities for unbound OLAARPs (7).
For comparison purposes, when differences were significant,
a protection index (PI) was calculated according to the following
equation:
PI ) 100 × [1 - ((x - y)/(a - b))]
where a is the microsomes with ROS, b is the microsomes
without ROS, x is the microsomes plus modified proteins treated
with ROS, and y is the untreated microsomes plus the modified
proteins.
Sta tistica l An a lysis. All results are expressed as mean
values of three experiments unless otherwise indicated. Statisti-
cal comparisons between two groups were made using Student’s
t-test. With several groups, ANOVA was used. When significant
F values were obtained, group differences were evaluated by
the Student-Newman-Keuls test (28). All statistical procedures
were carried out using Primer of Biostatistics: The Program
(McGraw-Hill, Inc., New York). Significance level is p < 0.05
unless otherwise indicated.
Resu lts
Ch a r a cter iza tion of Non en zym a tica lly Br ow n ed
P r otein s. The incubation of BSA with EH rapidly
produced browning and fluorescence in the protein (9).
Figure 1 shows the gel filtration chromatogram on
Sephacryl S-200-HR obtained for a BSA/EH incubation
mixture. Two main fractions, corresponding to dimeric
and monomeric modified BSA (DBSA and MBSA, respec-
tively) were observed. Both fractions, which exhibited
fluorescence at 440 nm when excited at 350 nm (data not
shown), were isolated and studied for amino acid com-
position.
P r otein P yr r oliza tion in In cu ba ted Micr osom es. The
production of OLAARP-containing proteins in incubated mi-
crosomes was analyzed by both: detection of pyrrolized proteins
with the Ehrlich reagent and determination of OLAARP product
Pnl after basic hydrolysis. Pyrrole determination was carried
out by using p-(dimethylamino)benzaldehyde (DMAB) under
Table 1 collects the amino acid composition of BSA,
MBSA, and DBSA. In accordance with previous studies
(9, 29), the reaction of EH with BSA mainly modified the

584 Chem. Res. Toxicol., Vol. 14, No. 5, 2001
Hidalgo et al.
F igu r e 1. Gel filtration chromatogram of a EH/BSA reaction
mixture on Sephacryl S-200-HR. BSA was incubated at 37 °C
in the presence of 10 mM EH for 19 h and injected in the column.
Fractions were collected and tested for protein by absorption
at 280 nm. Two fractions were isolated, which mainly cor-
responded to dimeric and monomeric modified BSA (DBSA and
MBSA, respectively).
Ta ble 1. Am in o Acid Com p osition of Non en zym a tica lly
Br ow n ed P r otein sa
F igu r e 2. Reaction scheme for protein pyrrolization produced
by EH. P represents the protein. Polymerization is produced
both intra- and intermolecularly.
amino
acid
BSA
MBSA
DBSA
Ala
0.68 ( 0.02 (9.1)
0.29 ( 0.03 (3.9)
0.74 ( 0.04 (10.0) 0.74 ( 0.02 (9.7)
Ta ble 2. OLAARP F or m a tion in Non en zym a tica lly
Br ow n ed P r otein sa
Arg
0.29 ( 0.02 (3.9) 0.28 ( 0.02 (3.7)
Asxb
0.83 ( 0.03 (11.1) 0.79 ( 0.04 (10.7) 0.78 ( 0.02 (10.2)
0.14 ( 0.01 (1.9) 0.16 ( 0.02 (2.1)
amino acid
BSA
MBSA
DBSA
cystine 0.14 ( 0.01 (1.9)
Glxb
Gly
His
Ile
1.16 ( 0.03 (15.4) 1.17 ( 0.02 (15.8) 1.21 ( 0.01 (15.9)
Arg
Lys
Pnl
0.35 ( 0.02
0.75 ( 0.01b
0.32 ( 0.02
0.46 ( 0.01c
0.09 ( 0.01b
0.33 ( 0.02
0.46 ( 0.02c
0.12 ( 0.01c
0.27 ( 0.02c (3.6)
0.16 ( 0.02 (2.1)
0.18 ( 0.01c (2.4)
0.35 ( 0.01d (4.7) 0.36 ( 0.01d (4.7)
0.15 ( 0.01 (2.0) 0.18 ( 0.01 (2.4)
0.22 ( 0.01d (3.0) 0.21 ( 0.02d (2.8)
a
Leu
Lys
Met
Phe
Ser
Thr
Tyr
Val
0.89 ( 0.03 (11.9) 0.90 ( 0.02 (12.1) 0.93 ( 0.01 (12.2)
Values are mean ( SD for three experiments and are given
in micromoles per milligram protein. b,c Means in the same row
with different superscripts are significantly different (p < 0.05).
0.82 ( 0.03c (10.9) 0.45 ( 0.02d (6.1) 0.43 ( 0.01d (5.6)
0.06 ( 0.01c (0.8)
0.38 ( 0.01c (5.1)
0.40 ( 0.01c (5.3)
0.47 ( 0.01c (6.3)
0.26 ( 0.01 (3.5)
0.52 ( 0.01 (6.9)
0.07 ( 0.01c (0.9) 0.09 ( 0.01d (1.2)
0.37 ( 0.01c (5.0) 0.41 ( 0.02d (5.4)
0.47 ( 0.01d (6.3) 0.47 ( 0.02d (6.2)
0.52 ( 0.02d (7.0) 0.56 ( 0.02e (7.3)
is stable enough to be determined by HPCE after basic
hydrolysis (21). Table 2 shows arginine, lysine, and Pnl
composition of BSA, MBSA, and DBSA determined after
basic hydrolysis. As expected (9), the decreases observed
in the amino acid lysine of MBSA and DBSA were
partially compensated by the formation of the new amino
acid Pnl. According to HPCE, about 40% of lysine
residues were lost in BSA upon EH treatment and the
13% of these residues appeared as Pnl. The other 27% of
lysine residues should be converted, in a first step, into
hydroxyalkylpyrroles and, later, into polymers. Although
both MBSA and DBSA exhibited the same losses of lysine
after acid and basic hydrolysis, Pnl was not produced in
the same proportion in both modified proteins and DBSA
exhibited a higher proportion of Pnl residues than MBSA
(p ) 0.021).
P yr r oliza tion of Micr osom a l P r otein s P r od u ced
by ROS. A pyrrolization of microsomal proteins analo-
gous to that observed for the reaction with EH was also
observed following their exposition to ROS. Figure 3
shows the UV spectra obtained for the Ehrlich adducts
produced with three pyrrolized proteins: (A) BSA incu-
bated in the presence of EH; (B) microsomes treated with
0.24 ( 0.01 (3.2)
0.54 ( 0.02 (7.3)
0.24 ( 0.01 (3.1)
0.57 ( 0.03 (7.5)
a
Values are mean ( SD for three experiments and are given
in micromoles per milligram protein (amino acid percentage is
b
shown in parentheses). Asx, aspartic acid + asparagine; Glx,
glutamic acid + glutamine. c,d,e Means in the same row with
different superscripts are significantly different (p < 0.05).
amino acid lysine, which decreased about 45% after acid
hydrolysis. A similar decrease in this amino acid was
observed for both MBSA and DBSA. These losses are a
consequence of the reactivity of EH for the amino groups.
Thus, EH modified some of the ꢀ-amino groups of lysine
residues into pyrrole amino acids (19). The reaction
scheme has been summarized in Figure 2. When EH
reacts with the ꢀ-amino group of lysine, two different
pyrrole derivatives are produced: 1-(5′-amino-5′-carboxy-
pentyl)-2-(1′′-hydroxypropyl)pyrrole (I) and ꢀ-N-pyrrolyl-
norleucine (Pnl) (II). The hydroxyalkylpyrrole derivative
I is unstable and polymerizes spontaneously (both intra-
and intermolecularly) producing lipofuscin-like polymers,
which are the responsible for the brown color and
fluorescence of MBSA and DBSA (9, 20). However, Pnl

Antioxidant Function of Proteins
Chem. Res. Toxicol., Vol. 14, No. 5, 2001 585
F igu r e 3. UV spectra of Ehrlich adducts obtained from
pyrrolized proteins. Ehrlich adducts were produced in the
reaction of DMAB with (A) BSA incubated overnight with 1 mM
EH; (B) microsomes incubated for 3 h in the presence of 1 mM
EH; (C) microsomes oxidized for 3 h in the presence of the Fe3+
/
ascorbate system; (D) control microsomes. Spectra B, C, and D
were taken using the same conditions and, therefore, are
comparable. Spectrum A was reduced to be incorporated into
the figure.
EH; and (C) microsomes oxidized with Fe3+/ascorbate.
These three spectra exhibited similar absorption maxima,
which was characteristic of the chromophore pyrrole/
DMAB (25) and that was not shown by control mi-
crosomes (unoxidized microsomes treated with DMAB,
Figure 3D), suggesting that oxidative stress produced
protein pyrrolization analogously to the treatment with
EH.
A further confirmation of this modification of microso-
mal proteins was obtained by determination of Pnl by
HPCE after basic hydrolysis. Figure 4 shows the elec-
tropherogram obtained from hydrolyzed unoxidized mi-
crosomes, and microsomes oxidized for 3 h in the presence
of Fe3+/ascorbate. Pnl concentration in control mi-
crosomes was 0.5 ( 0.4 µM, and this concentration
increased significantly (p < 0.001) when microsomes were
incubated in the presence of the Fe3+/ascorbate system.
Pnl concentration in these oxidized microsomes was 2.3
( 0.6 µM.
F igu r e 4. Electropherograms obtained for (A) control mi-
crosomes; (B) microsomes oxidized for 3 h in the presence of
the Fe3+/ascorbate system. Peaks corresponding to amino acids
Pnl, arginine, and lysine are indicated in the figure.
and Cu2+/H2O2 systems were 24, 3, and 1 h, respectively
(7). Figure 5 shows TBARS values obtained in mi-
crosomes oxidized in the presence of 20 µg/mL of BSA,
MBSA, and DBSA as a function of the incubation time.
When the microsomes were incubated in the presence of
5 µM Cu2+, a significant decrease in TBARS production
was observed for DBSA after 6 h and for BSA and MBSA
after 24 h (Figure 5A). At the end of the incubation period
(24 h) the PI observed were 15, 19, and 32 for BSA,
MBSA, and DBSA, respectively.
Analogous results were obtained for the Fe3+/ascorbate
and Cu2+/H2O2 systems (Figure 5, panels B and C,
respectively). In the treatment with Fe3+/ascorbate, both
MBSA and DBSA decreased significantly TBARS pro-
duction after 1 min, and BSA after 15 min. At the end of
the incubation period (3 h) the PI observed were 6, 16,
and 27 for BSA, MBSA, and DBSA, respectively. In the
treatment with the Cu2+/H2O2 system, DBSA exhibited
a significant decrease in TBARS production after 15 s
and for MBSA and BSA after 1 min. At the end of the
incubation period (1 h) the PI observed were 6, 12, and
23 for BSA, MBSA, and DBSA, respectively.
Effect of Non en zym a tica lly Br ow n ed P r otein s on
Oxid a tive Da m a ge In d u ced by ROS in Micr osom es.
The nonenzymatically browned proteins MBSA and
DBSA protected microsomes against both lipid peroxi-
dation and protein damage. The concentrations of the
proteins used to test this protection were selected to
obtain a Pnl concentration in the added proteins that
were similar to Pnl concentration produced after Fe3+
/
ascorbate treatment. Thus, a concentration of 40 µg/mL
of MBSA or DBSA corresponded to about 4 µM Pnl, and
therefore, this was the maximum concentration used in
these studies.
This protection of lipid peroxidation depended on the
concentration of the tested OLAARP-containing protein
and increased when its concentration also increased.
Figure 6A shows the effect of protein concentration on
TBARS production in microsomes treated with Cu2+. In
the treatment with Cu2+, the PI obtained at the three
tested concentrations (10, 20, and 40 µg/mL) were,
respectively, 3, 15, and 21 for BSA, 7, 19, and 40 for
MBSA, and 12, 32, and 54 for DBSA. Analogous results
Microsomal protection of lipid peroxidation by MBSA
and DBSA was evaluated by studying TBARS production
in microsomes incubated in the presence of Cu2+, Fe3+
/
ascorbate, and Cu2+/H2O2. Previous studies showed that
the best incubation periods for the Cu2+, Fe3+/ascorbate,

586 Chem. Res. Toxicol., Vol. 14, No. 5, 2001
Hidalgo et al.
F igu r e 6. TBARS production in microsomes incubated with
ROS and nonenzymatically browned proteins at different con-
centrations. Microsomes were suspended in KRPB at 37 °C in
the absence (stripped bars) or in the presence of BSA (open
bars), MBSA (crosshatched bars), or DBSA (horizontally striped
bars), and oxidized with (A) 5 µM Cu2+; (B) 1 mM Fe3+/5 mM
ascorbate; (C) 1 mM Cu2+/10 mM H2O2. The incubation time
depended on the ROS generating system used and was 24 h for
Cu2+, 3 h for Fe3+/ascorbate, and 1 h for Cu2+/H2O2. Bars with
different letters are significantly different (p < 0.05).
F igu r e 5. TBARS production in microsomes incubated with
ROS and nonenzymatically browned proteins. Microsomes were
suspended in KRPB and incubated at 37 °C in the absence (O)
or in the presence of 20 µg/mL of BSA (4), MBSA (3), or DBSA
()), and oxidized with (A) 5 µM Cu2+; (B) 1 mM Fe3+/5 mM
ascorbate; (C) 1 mM Cu2+/10 mM H2O2. Control microsomes with
or without proteins added (0) are included in graphs for
comparison purposes.
were obtained for the Fe3+/ascorbate and the Cu2+/H2O2
systems when the proteins were tested at 20 or 40 µg/
mL, respectively (Figure 6, panels B and C).
ment and the amino acid studied, DBSA was always more
protective than MBSA, and both of them were much more
protective than BSA. Thus, for example, lysine damage
was protected by 8, 38, and 50% in the treatment with
Fe3+/ascorbate, and -17, 53, and 62%, in the treatment
with Cu2+/H2O2. These two systems were slightly differ-
ent from Cu2+ because some amino acids were not
protected at the tested concentrations of MBSA and
Addition of modified proteins to microsomes also
induced a protection against protein damage as could be
determined by amino acid analysis after acid hydrolysis.
Table 3 shows the amino acid analysis obtained for
microsomes treated with Cu2+ when 40 µg/mL of BSA,
MBSA, or DBSA was added. This table only includes the
amino acids that were destroyed in microsomes following
Cu2+ oxidation (7). Addition of BSA induced a slight
protection in histidine and lysine residues but not in
arginine, cyst(e)ine, or tyrosine. BSA had negative PI for
these last three amino acids. These results suggest that
BSA did not protect arginine, cyst(e)ine and tyrosine
residues in microsomes, and a part of these residues in
the added BSA were also destroyed by the ROS generat-
ing system. Addition of 40 µg/mL of MBSA was signifi-
cantly more protective of all the amino acids, except
cyst(e)ine. Finally, addition of 40 µg/mL of DBSA exhib-
ited an almost complete protection, and all the amino
acids were recovered almost quantitatively after the 24
h incubation period.
DBSA. These were cyst(e)ine and methionine for the Fe3+
/
ascorbate system, and cyst(e)ine, methionine, and ty-
rosine for the Cu2+/H2O2 system. Nevertheless, the PI
also increased in these cases in the order BSA < MBSA
< DBSA.
Discu ssion
The results obtained in this study suggest that one
consequence of oxidative stress is the pyrrolization of
proteins. This effect, which was previously observed in
other systems (see, for example, ref 25), seems to be
general with independence of the lipids implicated.
Protein pyrrolization has been determined by using two
different procedures. The first of them consisted in the
treatment of oxidized microsomes with the Ehrlich
reagent. The Ehrlich adducts obtained for oxidized mi-
crosomes were analogous to those obtained for mi-
crosomes incubated in the presence of EH and for BSA
Analogous results were obtained for the Fe3+/ascorbate
and the Cu2+/H2O2 systems (Tables 4 and 5, respectively).
However, the protection exhibited depended on the amino
acid studied and the ROS generating system employed.
As a general rule, and with independence of the treat-

Antioxidant Function of Proteins
Chem. Res. Toxicol., Vol. 14, No. 5, 2001 587
a
Ta ble 3. Effect of Non en zym a tica lly Br ow n ed P r otein s on Am in o Acid Com p osition of Micr osom es In cu ba ted w ith Cu 2+
none
BSA
treated
1.16 ( 0.06b 1.11 ( 0.03c 1.51 ( 0.03b 1.45 ( 0.02c
MBSA
treated
-20 1.29 ( 0.01b 1.28 ( 0.01b 100 1.27 ( 0.01b 1.27 ( 0.01b 100
DBSA
amino
acid
control
treated
control
PI
control
PI
control
treated
PI
Arg
cystine 0.15 ( 0.04b 0.09 ( 0.01c 0.39 ( 0.01b 0.25 ( 0.02c -133 0.30 ( 0.01b 0.24 ( 0.01c
0
0.28 ( 0.01b 0.27 ( 0.01b 100
83 0.66 ( 0.02b 0.65 ( 0.01b 100
63 2.15 ( 0.01b 2.09 ( 0.01c
83
56 0.82 ( 0.01b 0.81 ( 0.01b 100
His
Lys
Tyr
0.53 ( 0.07b 0.35 ( 0.06c 0.80 ( 0.05b 0.67 ( 0.06c
2.00 ( 0.10b 1.65 ( 0.06c 2.97 ( 0.07b 2.68 ( 0.04c
0.68 ( 0.03b 0.59 ( 0.02c 1.00 ( 0.02b 0.85 ( 0.03c
28 0.68 ( 0.01b 0.65 ( 0.01c
17 2.20 ( 0.02b 2.07 ( 0.01c
-67 0.84 ( 0.01b 0.80 ( 0.01c
a
Values (in µmol/mg of microsomal protein) are mean ( SD for three independent experiments and are given for the amino acids
which decreased significantly in microsomes following Cu2+ treatment (7). When control and treated samples were significantly different
(p < 0.05), percentages of inhibition (PI) were calculated as described in the Materials and Methods. When control and treated samples
were not different, PI was equal to 100. b,c Means for control and treated samples with different superscripts are significantly different
(p < 0.05).
Ta ble 4. Effect of Non en zym a tica lly Br ow n ed P r otein s on Am in o Acid Com p osition of Micr osom es In cu ba ted w ith
F e3+/Ascor ba tea
none
BSA
treated
MBSA
treated
DBSA
treated
amino
acid
control
treated
control
PI
control
PI
control
PI
Ala
2.54 ( 0.13b 2.43 ( 0.18b 3.23 ( 0.07b 3.37 ( 0.06b
1.16 ( 0.06b 0.97 ( 0.11c 1.51 ( 0.03b 1.37 ( 0.08c
3.84 ( 0.20b 3.63 ( 0.19c 4.80 ( 0.10b 4.93 ( 0.07b
100 2.89 ( 0.01b 2.91 ( 0.01b
26 1.29 ( 0.01b 1.17 ( 0.04b
100 4.33 ( 0.02b 4.34 ( 0.06b
100 2.87 ( 0.03b 2.97 ( 0.06b
37 1.27 ( 0.01b 1.18 ( 0.03c
100 4.28 ( 0.10b 4.38 ( 0.09b
100
53
100
-67
100
100
62
60
100
50
Arg
Asxd
cystine 0.15 ( 0.04b 0.12 ( 0.02c 0.39 ( 0.01b 0.29 ( 0.01c -233 0.30 ( 0.01b 0.21 ( 0.01c -200 0.28 ( 0.01b 0.23 ( 0.01c
Glxd
Gly
His
Ile
Leu
Lys
Met
Phe
Ser
Thr
Tyr
Val
3.00 ( 0.16b 2.79 ( 0.14c 4.42 ( 0.08b 4.26 ( 0.06b
2.42 ( 0.17b 2.19 ( 0.18c 2.42 ( 0.18b 2.72 ( 0.06b
0.53 ( 0.07b 0.27 ( 0.07c 0.80 ( 0.07b 0.56 ( 0.01c
1.74 ( 0.10b 1.54 ( 0.14c 1.77 ( 0.04b 1.61 ( 0.04c
2.50 ( 0.13b 2.27 ( 0.19c 3.49 ( 0.06b 3.42 ( 0.03b
2.00 ( 0.10b 1.52 ( 0.10c 2.97 ( 0.07b 2.53 ( 0.03c
0.90 ( 0.19b 0.74 ( 0.15b 0.98 ( 0.05b 0.76 ( 0.01c
1.19 ( 0.06b 1.02 ( 0.10c 1.62 ( 0.02b 1.51 ( 0.02c
1.69 ( 0.08b 1.53 ( 0.13c 2.12 ( 0.05b 2.01 ( 0.02c
1.80 ( 0.15b 1.54 ( 0.14c 2.26 ( 0.20b 2.41 ( 0.04b
0.68 ( 0.03b 0.54 ( 0.05c 1.00 ( 0.02b 0.83 ( 0.04c
2.50 ( 0.13b 2.27 ( 0.18c 2.90 ( 0.03b 3.00 ( 0.06b
100 3.71 ( 0.02b 3.72 ( 0.09b
100 2.43 ( 0.02b 2.35 ( 0.05b
100 3.65 ( 0.04b 3.76 ( 0.06b
100 2.43 ( 0.06b 2.42 ( 0.08b
-4 0.66 ( 0.02b 0.56 ( 0.03c
35 1.66 ( 0.01b 1.58 ( 0.04c
100 2.95 ( 0.02b 3.01 ( 0.06b
38 2.15 ( 0.01b 1.91 ( 0.03c
8
0.68 ( 0.01b 0.41 ( 0.02c
20 1.68 ( 0.01b 1.55 ( 0.03c
100 2.99 ( 0.01b 3.02 ( 0.03b
8
2.20 ( 0.02b 1.90 ( 0.03c
-38 0.95 ( 0.01b 0.41 ( 0.04c -238 0.88 ( 0.01b 0.50 ( 0.05c -138
35 1.40 ( 0.01b 1.37 ( 0.01c
31 1.95 ( 0.01b 1.85 ( 0.02c
100 2.16 ( 0.01b 2.20 ( 0.03b
-21 0.84 ( 0.01b 0.76 ( 0.03c
100 2.61 ( 0.01b 2.55 ( 0.04b
82 1.38 ( 0.01b 1.36 ( 0.02b
38 1.91 ( 0.01b 1.83 ( 0.03c
100 2.12 ( 0.02b 2.22 ( 0.06b
43 0.82 ( 0.01b 0.75 ( 0.04c
100 2.58 ( 0.02b 2.59 ( 0.07b
100
50
100
50
100
a
Values (in µmol/mg of microsomal protein) are mean ( SD for three independent experiments. When control and treated samples
were significantly different (p < 0.05), percentages of inhibition (PI) were calculated as described in the Materials and Methods. When
control and treated samples were not different, PI was equal to 100. b,c Means for control and treated samples with different superscripts
d
are significantly different (p < 0.05). Asx, aspartic acid + asparagine; Glx, glutamic acid + glutamine.
Ta ble 5. Effect of Non en zym a tica lly Br ow n ed P r otein s on Am in o Acid Com p osition of Micr osom es In cu ba ted w ith
Cu 2+/H2O2
a
none
BSA
treated
MBSA
treated
57 2.89 ( 0.01a 2.95 ( 0.04a
18 1.29 ( 0.01a 1.16 ( 0.01c
57 4.33 ( 0.02a 4.36 ( 0.02a
DBSA
treated
100 2.87 ( 0.03a 2.97 ( 0.06a 100
73 1.27 ( 0.01a 1.15 ( 0.02c
76
100 4.28 ( 0.10a 4.37 ( 0.11a 100
amino
acid
control
treated
control
PI
control
PI
control
PI
Ala
2.54 ( 0.13a 1.80 ( 0.16c 3.23 ( 0.07a 2.91 ( 0.12c
1.16 ( 0.06a 0.67 ( 0.08c 1.51 ( 0.03a 1.11 ( 0.04c
3.84 ( 0.20a 2.82 ( 0.29c 4.80 ( 0.10a 4.36 ( 0.07c
Arg
Asxd
cystine 0.15 ( 0.04a 0.00 ( 0.00c 0.39 ( 0.01a 0.00 ( 0.00c -160 0.30 ( 0.01a 0.00 ( 0.00c -100 0.28 ( 0.01a 0.00 ( 0.00c -87
Glxd
Gly
His
Ile
Leu
Lys
Met
Phe
Ser
Thr
Tyr
Val
3.00 ( 0.16a 2.23 ( 0.25c 4.42 ( 0.08a 3.96 ( 0.06c
2.42 ( 0.17a 1.66 ( 0.28c 2.42 ( 0.18a 2.08 ( 0.06c
0.53 ( 0.07a 0.16 ( 0.04c 0.80 ( 0.07a 0.25 ( 0.02c
1.74 ( 0.10a 1.12 ( 0.15c 1.77 ( 0.04a 1.52 ( 0.04c
2.50 ( 0.13a 1.44 ( 0.21c 3.49 ( 0.06a 2.61 ( 0.05c
2.00 ( 0.10a 0.96 ( 0.21c 2.97 ( 0.07a 1.75 ( 0.04c
0.90 ( 0.19a 0.26 ( 0.22c 0.74 ( 0.22a 0.00 ( 0.00c
1.19 ( 0.06a 0.02 ( 0.02c 1.62 ( 0.02a 0.10 ( 0.04c
1.69 ( 0.08a 1.05 ( 0.16c 2.12 ( 0.05a 1.54 ( 0.06c
1.80 ( 0.15a 1.21 ( 0.20c 2.26 ( 0.20a 1.88 ( 0.05c
0.68 ( 0.03a 0.00 ( 0.00c 1.00 ( 0.02a 0.00 ( 0.00c
2.50 ( 0.13a 1.79 ( 0.22c 2.90 ( 0.03a 2.58 ( 0.07c
40 3.71 ( 0.02a 3.76 ( 0.03a
55 2.43 ( 0.02a 2.21 ( 0.04c
-49 0.68 ( 0.01a 0.27 ( 0.02c
60 1.68 ( 0.01a 1.57 ( 0.03c
17 2.99 ( 0.01a 2.94 ( 0.04a
-17 2.20 ( 0.02a 1.71 ( 0.02c
-16 0.95 ( 0.01a 0.00 ( 0.00c
-30 1.40 ( 0.01a 0.56 ( 0.05c
100 3.65 ( 0.04a 3.78 ( 0.08a 100
71 2.43 ( 0.06a 2.41 ( 0.06c 100
-11 0.66 ( 0.02a 0.46 ( 0.03c
46
82 1.66 ( 0.01a 1.70 ( 0.03a 100
100 2.95 ( 0.02a 3.00 ( 0.06a 100
53 2.15 ( 0.01a 1.75 ( 0.02c
62
-48 0.88 ( 0.01a 0.00 ( 0.00c -38
28 1.38 ( 0.01a 0.56 ( 0.02c
67 1.91 ( 0.01a 1.74 ( 0.02c
30
73
9
1.95 ( 0.01a 1.74 ( 0.01c
36 2.16 ( 0.01a 2.18 ( 0.05a
-47 0.84 ( 0.01a 0.00 ( 0.00c
55 2.61 ( 0.01a 2.59 ( 0.07a
100 2.12 ( 0.02a 2.21 ( 0.06a 100
-24 0.82 ( 0.01a 0.00 ( 0.00c -21
100 2.58 ( 0.02a 2.67 ( 0.06a 100
a
Values (in µmol/mg of microsomal protein) are mean ( SD for three independent experiments. When control and treated samples
were significantly different (p < 0.05), percentages of inhibition (PI) were calculated as described in the Materials and Methods. When
control and treated samples were not different, PI was equal to 100. b ,c Means for control and treated samples with different superscripts
d
are significantly different (p < 0.05). Asx, aspartic acid + asparagine; Glx, glutamic acid + glutamine.
also treated with this aldehyde, suggesting that all of
these processes produced analogous pyrroles. Other stud-
ies from this laboratory have shown that the same
Ehrlich adducts were also obtained in the oxidation of
serum proteins with Fe3+/ascorbate or in the treatment
of serum proteins with lipid hydroperoxides, but not in
the treatment of these proteins with hydroxyalkenals
(25). The second procedure consisted in the determination
of OLAARP product Pnl. Pnl was produced in microsomes
oxidized with Fe3+/ascorbate. After 3 h, the Pnl concen-
tration in these microsomes was 2.3 ( 0.6 µM, a concen-
tration that was high enough to delay the oxidative stress
process (7).
When model OLAARP-containing proteins, with a Pnl
concentration similar to that produced in oxidized mi-
crosomes, were tested for antioxidant activity in the
microsomal system, these proteins efficiently protected
microsomes against lipid peroxidation and protein dam-
age induced by three different ROS generating systems.
This protection, which was parallel to the concentration
of the assayed protein and the Pnl content in the protein,
was much higher than that exhibited by the unmodified
protein. These results suggest that OLAARP-containing
proteins, which are produced as a final step in the lipid
peroxidation process, are delaying the whole oxidative
stress by competing effectively with other nucleophiles

588 Chem. Res. Toxicol., Vol. 14, No. 5, 2001
Hidalgo et al.
(7) Zamora, R., Alaiz, M., and Hidalgo, F. J . (1997) Feed-back
inhibition of oxidative stress by oxidized lipid/amino acid reaction
products. Biochemistry 36, 15765-15771.
(8) Kotani, K., Maekawa, M., Kanno, T., Kondo, A., Toda, N., and
Manabe, M. (1994) Distribution of immunoreactive malondialde-
hyde-modified low-density lipoprotein in human serum. Biochim.
Biophys. Acta 1215, 121-125.
(9) Hidalgo, F. J ., and Zamora, R. (2000) Modification of bovine serum
albumin structure following reaction with 4,5(E)-epoxy-2(E)-
heptenal. Chem. Res. Toxicol. 13, 501-508.
(10) Xu, G., Liu, Y., Kansal, M. M., and Sayre, L. M. (1999) Rapid
cross-linking of proteins by 4-ketoaldehydes and 4-hydroxy-2-
alkenals does not arise from the lysine-derived monoalkylpyrroles.
Chem. Res. Toxicol. 12, 855-861.
(11) Salomon, R. G., Kaur, K., Podrez, E., Hoff, H. F., Krushinsky, A.
V., and Sayre, L. M. (2000) HNE-derived 2-pentylpyrroles are
generated during oxidation of LDL, are more prevalent in blood
plasma from patients with renal disease or atherosclerosis, and
are present in atherosclerotic plaques. Chem. Res. Toxicol. 13,
557-564.
(12) Hidalgo, F. J ., and Zamora, R. (1995) In vitro production of long
chain pyrrole fatty esters from carbonyl-amine reactions. J . Lipid
Res. 36, 725-735.
(13) Zamora, R., and Hidalgo, F. J . (1995) Linoleic acid oxidation in
the presence of amino compounds produces pyrroles by carbonyl
amine reactions. Biochim. Biophys. Acta 1258, 319-327.
(14) Salomon, R. G., Subbanagounder, G., O’Neil, J ., Kaur, K., Smith,
M. A., Hoff, H. F., Perry, G., and Monnier, V. M. (1997)
Levuglandin E2-protein adducts in human plasma and vascula-
ture. Chem. Res. Toxicol. 10, 536-545.
(15) Peters, T., J r. (1996) All About Albumin. Biochemistry, Genetics
and Medical Applications, pp 238-239, Academic Press, San
Diego, CA.
(16) Das, D. K., and Maulik, N. (1994) Antioxidant effectiveness in
ischemia-reperfusion tissue injury. Methods Enzymol. 233, 601-
610.
(17) Deigner, H. P., Friedrich, E., Sinn, H., and Dresel, H. A. (1992)
Scavenging of lipid peroxidation products from oxidizing LDL by
albumin alters the plasma half-life of a fraction of oxidized LDL
particles. Free Radical Res. Commun. 16, 239-246.
(18) Brown, J . R. (1977) Structure and evolution of serum albumin.
In Albumin Structure, Biosynthesis, Function (Peters, T., and
Sjoholm, I., Eds.) Vol. 50, pp 1-10, Pergamon Press, New York.
(19) Zamora, R., and Hidalgo, F. J . (1994) Modification of lysine amino
groups by the lipid peroxidation product 4,5(E)-epoxy-2(E)-
heptenal. Lipids 29, 243-249.
(20) Hidalgo, F. J ., and Zamora, R. (1993) Fluorescent pyrrole products
from carbonyl-amine reactions. J . Biol. Chem. 268, 16190-16197.
(21) Zamora, R., Navarro, J . L., and Hidalgo, F. J . (1995) Determi-
nation of lysine modification product ꢀ-N-pyrrolylnorleucine in
hydrolyzed proteins and trout muscle microsomes by micellar
electrokinetic capillary chromatography. Lipids 30, 477-483.
(22) Parkin, K. L., and Hultin, H. O. (1982) Fish muscle microsomes
catalyze the conversion of trimethylamine oxide to dimethylamine
and formaldehyde. FEBS Lett. 139, 61-64.
F igu r e 7. Antioxidant effect of OLAARP-containing proteins
produced as a consequence of oxidative stress. ROS produce both
lipid and protein oxidation, but oxidized lipids react with
proteins to form OLAARP-containing proteins with antioxidant
properties.
in reacting with the ROS produced. In this process, all
different OLAARPs produced (9, 29) may participate. The
whole process is shown in Figure 7.
The above results also give a much more complete
vision of the oxidative stress process. Proteins are modi-
fied following oxidative stress, and this modification is
able to induce antioxidative activity in these proteins,
which delays the whole process. According to these
results, and in addition to the antioxidant properties that
some proteins have by themselves, all proteins may also
exert an antioxidant effect after reaction with lipid
oxidation products, which might either increase their
ability to sequester metals or generate new molecules
that are antioxidants by themselves. This may be a
protective mechanism by which these toxic compounds
are removed and endogenous antioxidants are produced,
constituting a general pathway for an additional anti-
oxidative effect of proteins. This protective mechanism
may also be acting in parallel with the endogenous
antioxidant activities suggested for some amino acid
residues, like methionine (30).
Ack n ow led gm en t. This study was supported in part
by the Comisio´n Interministerial de Ciencia y Tecnolog´ıa
(CICYT) of Spain (Project ALI97-0358) and the J unta de
Andaluc´ıa (Project AGR 0135). We are indebted to Unio´n
Pisc´ıcola Navarra, S. L. (Granada, Spain) for the gift of
the trouts used in this study, and to Mr. J . L. Navarro
for the technical assistance.
(23) De Luca, H. F. (1972) Methods for preparation and study of tissues
and enzymes. In Manometric and Biochemical Techniques (Um-
breit, W. W., Burris, R. H., and Stauffer, J . F., Eds.) pp 133-
147, Burgess Publishing, Minneapolis, MN.
(24) Bradford, M. M. (1976) A rapid and sensitive method for the
quantitation of microgram quantities of protein utilizing the
principle of protein-dye binding. Anal. Biochem. 72, 248-254.
(25) Hidalgo, F. J ., Alaiz, M., and Zamora, R. (1998) A spectrophoto-
metric method for the determination of proteins damaged by
oxidized lipids. Anal. Biochem. 262, 129-136.
Refer en ces
(26) Kosugi, H., Kojima, T., and Kikugawa, K. (1989) Thiobarbituric
acid-reactive substances from peroxidized lipids. Lipids 24, 873-
881.
(1) Halliwell, B., and Gutteridge, J . M. (1999) Free Radicals in
Biology and Medicine, 3rd ed., Oxford, U.K.
(27) Alaiz, M., Navarro, J . L., Giron, J ., and Vioque, E. (1992) Amino
acid analysis by high-performance liquid chromatography after
derivatization with diethyl ethoxymethylene malonate. J . Chro-
matogr. 591, 181-186.
(2) Ogino, T., Packer, L., and Traber, M. C. (1999) Oxidant stress
and host oxidant defense mechanisms. In Nutritional Oncology
(Heber, D., Blackburn, G. L., and Go, V. L. W., Eds.) pp 253-
275, Academic Press, San Diego, CA.
(28) Snedecor, G. W., and Cochran, W. G. (1980) Statistical Methods,
7th ed., Iowa State University Press, Ames, Iowa.
(3) Keyse, S. M., Ed. (2000) Stress Response. Methods and Protocols,
Humana Press, Totowa, NJ .
(29) Zamora, R., Alaiz, M., and Hidalgo, F. J . (1999) Modification of
histidine residues by 4,5-epoxy-2-alkenals. Chem. Res. Toxicol.
12, 654-660.
(4) Rice-Evans, C., and Bruckdorfer, K. R., Eds. (1995) Oxidative
Stress, Lipoproteins and Cardiovascular Dysfunction, Portland
Press, London, U.K.
(30) Levine, R. L., Mosoni, L., Berlett, B. S., and Stadtman, E. R.
(1996) Methionine residues as endogenous antioxidants in pro-
teins. Proc. Natl. Acad. Sci. U.S.A. 93, 15036-15040.
(5) Scandalios, J . G., Ed. (1997) Oxidative Stress and the Molecular
Biology of Antioxidant Defenses, Cold Spring Harbor Lab., Pla-
inview, NY.
(6) O¨ zben, T., Ed. (1998) Free Radicals, Oxidative Stress, and
Antioxidants: Pathological and Physiological Significance, Ple-
num Press, New York.
TX000215M