Báo cáo khoa học: Glycation and glycoxidation of low-density lipoproteins by glucose and low-molecular mass aldehydes Formation of modified and oxidized particles pot

Glycation and glycoxidation of low-density lipoproteins by glucose
and low-molecular mass aldehydes
Formation of modified and oxidized particles
Heather M. Knott*, Bronwyn E. Brown, Michael J. Davies and Roger T. Dean†
The Heart Research Institute, Camperdown, Australia
Patients with diabetes mellitus suffer from an increased
incidence of complications including cardiovascular disease
and cataracts; the mechanisms responsible for this are not
fully understood. One characteristic of such complications
is an accumulation of advanced glycation end-products
formed by the adduction of glucose or species derived from
glucose, such as low-molecular mass aldehydes, to proteins.
These reactions can be nonoxidative (glycation) or oxidative
(glycoxidation) and result in the conversion of low-density
lipoproteins (LDL) to a form that is recognized by the
scavenger receptors of macrophages. This results in the
accumulation of cholesterol and cholesteryl esters within
macrophages and the formation of foam cells, a hallmark of
atherosclerosis. The nature of the LDL modifications
required for cellular recognition and unregulated uptake are
poorly understood. We have therefore examined the nature,
time course, and extent of LDL modifications induced by
glucose and two aldehydes, methylglyoxal and glycolalde-
hyde. It has been shown that these agents modify Arg, Lys
and Trp residues of the apoB protein of LDL, with the extent
of modification induced by the two aldehydes being more
rapid than with glucose. These processes are rapid and
unaffected by low concentrations of copper ions. In contrast,
lipid and protein oxidation are slow processes and occur to a
limited extent in the absence of added copper ions. No evi-
dence was obtained for the stimulation of lipid or protein

oxidation by glucose or methylglyoxal in the presence of
copper ions, whereas glycolaldehyde stimulated such reac-
tions to a modest extent. These results suggest that the earliest
significant events in this system are metal ion-independent
glycation (modification) of the protein component of LDL,
whilst oxidative events (glycoxidation or direct oxidation of
lipid or proteins) only occur to any significant extent at later
time points. This Ôcarbonyl-stressÕ may facilitate the forma-
tion of foam cells and the vascular complications of diabetes.
Keywords: AGE; atherosclerosis; glycolaldehyde; methyl-
glyoxal; protein modification.
The correlation between diabetes and cardiovascular disease
(CVD) has been well established [1], although the precise
mechanisms that facilitate the many complications associ-
ated with diabetes, including CVD and cataracts, are poorly
understood. Uncontrolled plasma glucose concentrations
and the ability of glucose to either oxidatively or nonoxid-
atively (covalently) modify proteins have been proposed to
be instrumental in the development of CVD in both the
insulin-deficient and insulin-resistant forms of diabetes
[2–4]. Both free and protein-bound glucose are known to
undergo nonenzymatic and enzymatic modifications which
can result in the formation of low molecular mass aldehydes
such as methylglyoxal (MG), glyoxal, and glycolaldehyde
(GA). These aldehydes form adducts with Lys and Arg
residues resulting in Schiff base formation, Amadori
rearrangements, and the formation of advanced glycation
end products (AGEs) [5–9]. Thus, reaction of GA with Lys
results in the formation of the well characterized AGE
carboxymethyllysine whilst reaction of MG with Lys gives

carboxyethyllysine [10,11]. The levels of these small reactive
a-dicarbonyls are known to be elevated in diabetics [12–14]
and the accumulation of AGEs has been implicated in the
pathogenesis of diabetes and ageing [4,7].
Modification of low-density lipoprotein (LDL) can lead
to alteration of the apoB protein to the extent that it is no
longer recognized by the regulated cholesterol-feedback
receptors [15]. Instead, this modified LDL is taken up via
scavenger receptors leading to cholesterol and cholesteryl
ester loading of macrophages [16]; this is believed to be the
primary step in foam cell formation and the development of
Correspondence to Heather M. Knott, Centenary Institute of Cancer
Medicine and Cell Biology, Locked Bag Number 6, Newtown,
Sydney, NSW 2042, Australia.
Fax: +61 2 9565 6101, Tel.: +61 2 9565 6156,
E-mail:
Abbreviations: CVD, cardiovascular disease; AGE, advanced gly-
cation end-products; GA, glycolaldehyde; MG, methylglyoxal; LDL,
low-density lipoprotein; DOPA, 3,4-dihydroxyphenylalanine;
m-Tyr, m-tyrosine (3-hydroxyphenylalanine); o-Tyr, o-tyrosine
(2-hydroxyphenylalanine); PTQ, phenanthrenequinone; TOH,
a-tocopherol; FC, free cholesterol; CA, cholesteryl arachidonate;
CL, cholesteryl linoleate; 7K, 7-ketocholesterol; apoB, apolipoprotein
B100; CEO(O)H, cholesteryl ester hydro(pero)xides; REM,
relative electrophoretic mobility.
*Present address: Centenary Institute of Cancer Medicine and Cell
Biology, Newtown, Sydney, Australia.
Present address: University of Canberra ACT 2601 Australia.
(Received 17 February 2003, revised 19 June 2003,
accepted8July2003)

Eur. J. Biochem. 270, 3572–3582 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03742.x
atherosclerosis [17]. Furthermore, the two- to threefold
increase in nonenzymatic glycosylation of serum albumin in
hyperglycaemia has been suggested to alter the antioxidant
(radical scavenging) role of this protein which may, in
concert with increased levels of redox-active copper and iron
levels [18], contribute to the complications of diabetes [19].
In this study the nature, time course, and extent of the
covalent and oxidative changes that occur on LDL particles
exposed to glucose, GA, and MG have been quantified in
order to determine the significance of glycation and
glycoxidation to the pathogenesis of atherogenic complica-
tions related to diabetes.
Materials and methods
Materials
All solutions were prepared with nanopure water (Milli Q
system, Millipore-Waters) and treated with washed Chelex-
100 (Bio-Rad) to remove transition metals prior to use [20].
PD-10 columns (Sephadex G-25
M
) were from Amersham
Biosciences. Pre-cast 1% agarose gels were from Helena
Laboratories (Mt. Waverly, VIC, Australia). Fatty acid-
free BSA, fluorescamine, MG, GA, glucose, arginine, and
lysine were all from Sigma-Aldrich. All other chemicals
were of analytical grade and all solvents were of HPLC
grade.
Isolation of LDL
LDL (density 1.019–1.063 gÆmL
)1

) was prepared from
plasma of fasted, healthy volunteers by density gradient
ultracentrifugation as described previously [21]. After iso-
lation, LDL was dialysed against four to five 1 L changes of
degassed NaCl/P
i
containing 0.1 mgÆmL
)1
chlorampheni-
col, filter sterilized, and used immediately in most cases.
When necessary, dialysis buffers were supplemented with
1mgÆmL
)1
EDTA and the LDL stored until required. In
the latter case, EDTA was removed immediately prior to
use by either dialysis (as above) or by passage of LDL
(<2 mL) through two PD-10 columns as per the manu-
facturer’s instructions; NaCl/P
i
was used as both the
equilibration and elution buffer. Removal of the EDTA
was confirmed by carrying out a small-scale oxidation of
0.4 mgÆmL
)1
LDL with 10 l
M
CuSO
4
for 2–3 h at 37 °C.
After this incubation, a wavelength difference spectrum was

acquired using the control (no copper ion) sample as a
blank. Gross changes in the spectrum, particularly the loss
of the carotenoid peak at 450 nm and changes in the
absorbance at 234 nm, were accepted as indicative of
susceptibility to oxidation [22]. If such changes were not
observed, the LDL was subject to further dialysis or
rechromatographed. Oxidation assays were only performed
using LDL tested in this manner.
Data analysis
For the lipid and protein oxidation data, the variation in
absolute data obtained from LDL extracted from different
donors, containing different levels of antioxidants and pre-
existing lipid peroxides and hence different Ôlag phasesÕ [22],
made collation of all data impractical. This variation results
in differing absolute kinetics, though the trends with each
donor were the same. Therefore, although all data were
analysed, we present here that from a single experiment
representative of all with each data point representing the
average of two samples and the error bars representing
standard deviation (half-range). The protein modification
data are presented as mean ± SEM of data from four
samples and normalized to the control (no added modifier)
at each time point after initial assessment indicated minimal
variationinthezerotimerawdata.Statisticswere
performed by one-way analysis of variance (
ANOVA
) with
Tukey’s multiple comparison post-test analysis using
GRAPHPAD PRISM
(version 3.0a for MacIntosh, GraphPad

Software, San Diego, California, USA). For each compar-
ison, statistical significance was set at P < 0.05 unless
stated otherwise. For each of the protein modification
assays, control experiments were performed to exclude the
possibility that direct interference between the fluorescamine
or phenanthrequinone and either aldehyde contributed to
the changes in fluorescence observed.
Protein assays
LDL protein concentration was measured by the bicin-
choninic acid (BCA; Pierce) method using 0.4 mgÆmL
)1
BSA as a standard, with incubations performed at 60 °Cfor
45–60 min.
Glycoxidation of LDL
Glucose, MG, and GA stock solutions were prepared
immediately prior to use and filter-sterilized. EDTA-free
LDL was diluted to 350–450 lgÆmL
)1
in filter-sterilized
NaCl/P
i
containing the required concentration of aldehyde
and/or copper ions and incubated at 37 °C under 5% (v/v)
CO
2
. The screw caps on the incubation vials were loosely
fastened to enable oxygen ingress without compromising
the sterility of the sample.
Lipid extraction and analysis
At the required times, 0.1 mL samples were collected

aseptically into tubes containing 10 lL20m
M
EDTA,
10 lL0.2m
M
butylated hydroxytoluene (BHT), and
400 lL nanopure water (or alternatively 0.2 mL sample/
20 lLEDTA/20lLBHT/300lL water) and the samples
mixed immediately. One mL methanol and 5 mL hexane
were added and the samples were mixed thoroughly.
Extracts were stored at )20 °C until use. Samples were
centrifuged at 2060 g for 5–10 min, 4 mL of the hexane
layer was removed and evaporated to dryness and the
residue was resuspended in 200 lL of isopropanol. Lipid
content was analysed by reversed-phase HPLC on a
Shimadzu system with a Supelco ODS LC 18 column
(25 · 0.46 cm, 5 lm particle size) and a 2 cm Pelliguard
guard column. The mobile phase, ethanol/methanol/iso-
propanol at 19.7 : 6 : 1 (v/v/v), was degassed immediately
prior to use. Lipophilic components were separated iso-
cratically at 1 mLÆmin
)1
over 30 min with the eluent
monitored using a diode array UV detector (PDA-
M10AVP) set at 205 and 234 nm and a RF-10AXL
fluorescence detector (for TOH; k
ex
290 nm and k
em
Ó FEBS 2003 Glycation and glycoxidation of LDL (Eur. J. Biochem. 270) 3573

330 nm). Metabolites were quantified by comparison with
standards and expressed as nmolÆmg
)1
apoB protein.
Protein isolation and analysis
Samples (0.5 mL) of the LDL incubations were collected
aseptically into 50 lL20m
M
EDTA and 50 lL0.2m
M
BHT and mixed thoroughly. Samples were subsequently
stored at )20 °C before further processing. Protein samples
were reduced by addition of 10 lL10 mgÆmL
)1
NaBH
4
and
incubation at room temperature for >30 min. The samples
were delipidated by the addition of 100 lL0.3%(w/v)
sodium deoxycholate, precipitated using 50 lL 50% (w/v)
trichloroacetic acid, and centrifuged at 4000 g for 2 min.
After removal of the supernatant, the protein pellet was
washed twice with ice-cold acetone and once with diethyl
ether, with the samples centrifuged (6610 g, 1 min) and the
supernatant discarded after each addition. The samples
were then evaporated to dryness and subjected to gas-phase
hydrolysis in Pico-Tag vessels (Waters) containing 1 mL
6
M
HCl and 50 lL 2-mercaptoacetic acid. The vessels were

sealed under vacuum and incubated at 110 °Covernight.
After hydrolysis, the samples were dried, resuspended in
200 lL nanopure water, filtered through 0.22 or 0.45 lm
membranes, and analysed by HPLC.
Analysis of oxidized amino acids was performed by
reversed-phase HPLC using UV and fluorescence detec-
tion as described previously [23]. Quantification was
performed by integration of peak areas and comparison
with standards. Oxidized amino acid levels are expressed
as lmol per mol parent amino acid to account for any
losses during sample preparation. o-Tyr data was con-
verted from lmolÆmol
)1
Tyr to lmolÆmol
)1
Phe using the
abundance of these amino acids in the apoB100 molecule
(152 Tyr/223 Phe). Preliminary experiments that quanti-
fied the consumption of Tyr under the incubation
conditions used indicated that the (low) loss of Tyr had
negligible impact on this calculation (data not shown).
Previous studies using this methodology have demonstrated
minimal artefactual oxidation during such sample hand-
ling and analysis [23,24].
Relative electrophoretic mobility gels
Samples (10–15 lL) were removed from the LDL incuba-
tions and loaded onto precast 1% (w/v) agarose gels. Native
and acetylated LDL (Ac-LDL, prepared as described
previously [25]) were loaded as negative and positive
controls, respectively, and the samples were run at 90 V

for 45 min. Gels were fixed in 100% methanol (1 min),
stained for 5–10 min with Fat Red 7B (Sigma-Aldrich),
destained for 5–10 min with 70% methanol, and dried at
60 °C. The relative electrophoretic mobility (REM) is
defined as the ratio of the distances travelled by modified
LDL and native LDL.
Tryptophan consumption
Samples (0.2 ml) were removed from the LDL incubations,
0.1 vol. each of 20 m
M
EDTA and 0.2 m
M
BHT were
added, and the volume adjusted to 1 mL with NaCl/Pi. Trp
fluorescence was measured on a Perkin Elmer Lumines-
cence LS50B Spectrometer with k
ex
280 nm and k
em
335 nm.
Lysine consumption
Samples (80 lL) were removed from the LDL incubations
and 0.1 vol each of 20 m
M
EDTA and 0.2 m
M
BHT and
670 lL borate buffer (pH > 8) were added. During vortex
mixing, 250 lL0.15mgÆmL
)1

fluorescamine dissolved in
acetone was added. The fluorescence of these samples,
representative of free amine groups, was measured using k
ex
390 nm and k
em
475 nm and expressed as a percentage of
the zero time value for the control (no added aldehyde/
glucose) samples [26].
Arginine consumption
One-tenth vol. each of 20 m
M
EDTA and 0.2 m
M
BHT
were added to 0.25–0.50 mL of the LDL incubation and the
volume adjusted to 1 mL with NaCl/P
i
. Three millilitres
120 l
M
9,10-phenanthrenequinone (PTQ; in absolute eth-
anol) was added and the reaction initiated by addition of
0.5 mL 2
M
NaOH with the samples then incubated at
60 °C for 3 h. At both t ¼ 0andt¼ 3 h, 0.5 mL samples
were removed and the reaction stopped by addition of an
equal volume of 1.2
M

HCl. The fluorescence was recorded
using k
ex
312 nm and k
em
392 nm with the Arg-dependent
fluorescence calculated as the difference between the t ¼ 0
and t ¼ 3 h samples [27].
Results
Glucose-mediated oxidation of LDL
Lipid oxidation. Six different parameters of lipid peroxi-
dation were measured (Fig. 1A–F). Three concentrations
of glucose representing normal (5 m
M
), pathological
(25 m
M
), and suprapathological (100 m
M
) levels with
1 l
M
Cu
2+
were examined, as well as samples containing
1 l
M
Cu
2+
alone, 100 m

M
glucose only, and no additions.
The zero time levels of all lipids and products correlate
well with published plasma data (e.g. [28]) and are not
statistically different between experiments. The oxidation
of the lipid moieties of LDL in the presence of glucose
and 1 l
M
Cu
2+
proceeded slowly with complete destruc-
tion of a-tocopherol (TOH; Fig. 1A) occurring between
1 and 2 w. These results correlate with loss of free
cholesterol (Fig. 1B, FC), cholesteryl arachidonate
(Fig. 1C, CA), and cholesteryl linoleate (Fig. 1D, CL)
by 2 weeks as well as the accumulation of the cholesterol
oxidation product, 7-ketocholesterol (Fig. 1E, 7K), to a
maximum of 200 nmolÆmg
)1
apoB at 4 w. Cholesteryl
ester hydroxide and hydroperoxide accumulation (Fig. 1F,
CEO(O)H) was not detected for the Cu
2+
-containing
samples; this presumably reflects the production and
subsequent rapid degradation of these unstable products.
With 100 m
M
glucose alone and with the incubated
control (no addition) samples, oxidation proceeded more

slowly than in any of the samples containing 1 l
M
Cu
2+
.
A small acceleration of the rate of oxidation was detected
in the 100 m
M
glucose samples compared with the
controls at 2 and 4 weeks in all the parameters measured,
3574 H. M. Knott et al. (Eur. J. Biochem. 270) Ó FEBS 2003
with the exception of FC and 7K. In these samples, no
conversion of FC to 7K was detected, whilst TOH was
not fully depleted until 7 weeks in the presence of 100 m
M
glucose, with a corresponding 84% loss in the control.
Due to the lower rate of oxidation of these samples,
accumulation of CEO(O)H was detected and reached a
maximum (170 nmolÆmg
)1
apoB) by 4 weeks. The small
increases in the rate of oxidation observed in the presence
of 100 m
M
glucose compared with the controls is
consistent with a very low rate of radical formation in
the absence of added metal ions such as Cu
2+
.No
additive effect of the various glucose concentrations over

the rate of oxidation induced by Cu
2+
alone was
discernible.
Protein oxidation and modification. The formation of the
Tyr and Phe oxidation products, DOPA and o-Tyr, were
examined in identical incubations to those described above.
The levels of these products are shown in Fig. 1G and H.
Zero time levels of DOPA (900 lmolÆmol
)1
Tyr) and o-Tyr
(550 lmolÆmol
)1
Phe) are in accord with literature data for
plasma [23] and not statistically different between prepara-
tions. By week 2, accumulation of both DOPA (Fig. 1G,
four- to sixfold increase) and o-Tyr (Fig. 1H, seven- to
eightfold increase) was evident in all samples containing
added Cu
2+
, independent of the glucose concentration with
no statistically significant change in the two controls. The
levels of DOPA detected tended towards a decrease at
longer incubation times, consistent with the further oxida-
tion of this material to undetected indolic materials [29].
To complement the above markers of oxidative damage
to apoB, analysis of the potential loss of other amino acid
side chains (Trp, Lys and Arg), that would be expected to be
targets of glycation reactions (i.e. covalent modification),
was quantified. The loss of each of these side chains was

examined using fluorescence spectroscopy and in each case
changes were measured relative to control (no addition)
samples. Care was taken to eliminate any possible inter-
ference from the added reagents. As previous data has
suggested that glycation reactions are rapid (e.g. [6,30]),
these analyses were only carried out over the time frame
prior to the formation of significant levels of lipid and
protein oxidation products (i.e. up to 2 weeks, see above).
No evidence was obtained for significant consumption of
Trp or Lys residues under any of the incubation conditions
studied (data not shown) indicating that neither glucose nor
low levels of Cu
2+
induce significant Trp or Lys modifica-
tion. Furthermore the absence of any peak shifts in the Trp
fluorescence spectra imply only minor, if any, structural
changes in the vicinity of these residues, as the fluorescence
of this residue is environment dependent. Changes in the
level of Arg were detected with high, but not low,
concentrations of glucose with this being independent of
the presence of added Cu
2+
. Thus a loss of 67% of Arg
residues was detected after 14 days’ incubation with
100 m
M
glucose when compared with incubated controls;
no significant loss of Trp or Lys residues was detected under
these conditions (data not shown).
The relative electrophoretic mobility of the modified

LDL particles was also examined using agarose gels, as this
technique yields information on the overall net positive
charge on the particle (i.e. the total contribution of Arg, Lys,
and protonated His residues together with the N terminus,
relative to Glu, Asp, and the C terminus). Minor increases
in electrophoretic mobility were evident for the samples
incubated with low concentrations of glucose (data not
shown), though these were only 1.2–1.3-fold greater than
native (nonincubated) LDL, and similar changes were
detected with the incubated controls. These minor changes
were Cu
2+
-independent and consistent with minor extents
of modification of Lys and Arg residues. With samples
incubated with 100 m
M
glucose for 14 days a small increase
Fig. 1. Glucose-mediated oxidation of the lipid and protein moieties of
LDL. LDL ( 0.4 mg proteinÆmL
)1
) was incubated with varying
concentrations of glucose and/or copper ions. At the indicated times,
0.2 mL samples were removed and extracted with 1 mL methanol and
5 mL hexane. Four millilitres of the hexane fraction was then dried to
completion, resuspended in 200 lL isopropanol, and the levels
(expressed as nmolÆmg LDL protein
)1
) of TOH (A),FC(B),CA(C),
CL (D), 7-KC (E), and CEO(O)H (F) were determined by HPLC (see
Materials and methods). Concurrently, 0.5 mL samples were removed,

delipidated, hydrolysed, and the resulting oxidized and parent amino
acids quantified by reversed-phase HPLC. The levels of the oxidized
amino acids DOPA (G) and o-Tyr (H) are expressed relative to their
parent amino acids, Tyr and Phe, respectively. Each data point rep-
resents the mean (± SD) of two replicates from a single experiment
representative of several. Black bars, 100 m
M
glucose + 1 l
M
Cu
2+
;
horizontal striped bars, 25 m
M
glucose + 1 l
M
Cu
2+
;darkstippled
bars, 5 m
M
glucose + 1 l
M
Cu
2+
; light stippled bars, 100 m
M
glucose
only; white bars, 1 l
M

Cu
2+
only;diagonalstripedbars,control(no
addition) incubations. Asterisks indicate the first time point at which
the data becomes statistically different when compared with the t ¼ 0
value.
Ó FEBS 2003 Glycation and glycoxidation of LDL (Eur. J. Biochem. 270) 3575
in REM was detected but this was not significantly different
from the incubated controls (data not shown). Acetylated
LDL samples run under identical conditions as positive
controls gave REM values of 3–4.
MG-mediated oxidation of LDL
Lipid oxidation. LDL was incubated with a wide range of
MG concentrations (10 l
M
)100 m
M
) in both the absence
and presence of low concentrations of added Cu
2+
(1 l
M
).
No evidence for lipid oxidation was obtained up to the
longest time period studied (17 days, data not shown). That
is, there was no change in the level of TOH, no loss of
cholesterol or the cholesteryl esters examined (CA, CL), and
no accumulation of CEO(O)H or 7K (data not shown).
Protein oxidation and modification. As the above lipid
data was negative, the generation of DOPA and o-Tyr,

which are radical-mediated products [31,32], was not
examined. Protein modification by MG was examined by
quantifying the loss of Trp, Lys, and Arg residues and the
changes in REM using the same incubation conditions as
described above. Fig. 2 shows the data obtained, expressed
as a percentage of control incubations (no added MG). At
the initial time point there was a significant decrease in the
measurable level of Trp in the presence of 100 m
M
MG
(P < 0.001) while there was no statistical difference
between any of the other conditions. No statistically
significant changes in Trp levels were detected with low
concentrations of MG (10 and 100 l
M
). With higher
concentrations (10 and 100 m
M
MG) a significant decrease
in the level of this residue was observed by day 5 (Fig. 2A);
this decrease occurred in the absence of added Cu
2+
but
was dependent on the MG concentration (by 1 day the
difference between 100 m
M
and 10 m
M
is significant to
P < 0.01). The concentration of Lys residues was not

significantly different at zero time. Significant Lys loss
(Fig. 2B) was detected by 1 day in the samples containing
either 100 or 10 m
M
MG (P < 0.001 relative to all other
samples) and at this time has occurred to a greater extent in
the 100 m
M
than in the 10 m
M
MG samples (P < 0.001),
indicating that that the loss of this residue is dependent on
the concentration of MG. In the experiments using 100 l
M
MG, loss of lysine was only seen in the samples containing
1 l
M
Cu
2+
(P < 0.05 relative to other samples) and not
until day 14 (P < 0.01 relative to previous time points).
Whilst this is suggestive of a copper-ion dependence it is not
supported by any of the other data.
Modification of Arg residues by MG has also been
quantified and the data obtained is presented in Fig. 2C
(expressed relative to control samples at each time point,
with the latter set to 100%). With low concentrations of
MG (10 l
M
or 100 l

M
) no significant changes were
observed in the presence or absence of added copper ions,
except with 100 l
M
MG in the absence of Cu
2+
from day 5
onwards. With higher (millimolar) concentrations, rapid
concentration-dependent loss of this side-chain was detected
even in the absence of Cu
2+
(P < 0.05 by 1 day for both 10
and 100 m
M
MG); this is in accord with a previous report
[33]. Loss of Arg was particularly rapid, with significant
losses observed immediately after mixing (i.e. in the t ¼ 0
samples) with the 100 m
M
MG samples (P < 0.01 relative
to other samples at time zero). No additional loss of Arg
was observed at later time points. For the samples
containing 10 m
M
MG, a significant (P < 0.01) loss of
arginine was seen at 1 day with, again, no further losses
during the remainder of the time course.
As expected on the basis of the above data, changes in the
REM of LDL incubated with MG (conditions as above)

were detected (Fig. 2D). For the sake of clarity, data for
native LDL (REM set to 1) and AcLDL (REM > 3) are
not shown. With 10 l
M
MG, minor changes occurred after
5 days of incubation and increased at subsequent time
points (approximately twofold relative to native LDL by
14 days). More pronounced and more rapid changes were
detected with higher concentrations of MG in the absence of
added Cu
2+
.
GA-mediated oxidation of LDL
Lipid oxidation. LDL was incubated with 0, 0.1, or 1 m
M
GA with and without added 1 l
M
copper ions. Fig. 3 shows
typical data; similar trends were observed with other
Fig. 2. MG-mediated modification of apoB. LDL ( 0.4 mg
proteinÆmL
)1
) was incubated with 10 or 100 l
M
MG with and without
added 1 l
M
Cu
2+
or LDL ( 1mgproteinÆmL

)1
) was incubated with
10 or 100 m
M
MG in the absence of Cu
2+
. Trp residues (A) were
quantified by fluorescence (k
ex
280 nm, k
em
335 nm) after dilution in
NaCl/P
i
. Lys residues (B) were quantified by fluorescence (k
ex
390 nm,
k
em
475 nm) after dilution in borate buffer (pH > 8) and derivatiza-
tion with fluorescamine. Arg residues (C) were quantified by fluores-
cence (k
ex
312 nm, k
em
392 nm) after derivitization with
9,10-phenanethrequinone. For further details see Materials and
methods. Data are means (n ¼ 4) ± SEM. For REM analysis (D),
10–15 lL samples were loaded onto precast 1% (w/v) agarose gels,
subjected to electrophoresis, and the distance migrated relative to

native LDL (REM set to 1) calculated. Data are expressed as a per-
centage of control incubations with no added MG. Black bars,
100 m
M
MG;horizontalstripedbars,10m
M
MG; dark stippled bars,
100 l
M
MG + 1 l
M
Cu
2+
; light stippled bars, 100 l
M
MG; white
bars, 10 l
M
MG + 1 l
M
Cu
2+
;diagonalstripedbars,10l
M
MG.
Asterisks indicate the first time point at which the data becomes
statistically different when compared with the t ¼ 0 value.
3576 H. M. Knott et al. (Eur. J. Biochem. 270) Ó FEBS 2003
concentrations of GA (data not shown). No significant
differences were seen for any of the lipid peroxidation

parameters between treated and nontreated samples at the
zero time point. By 15 days, the most rapid lipid peroxi-
dation occurred in those samples containing 1 m
M
GA
supplemented with 1 l
M
copper ions. In this case, TOH
(Fig. 1A) and CA (Fig. 1C) were completely consumed and
CL (Fig. 1D) partly consumed ( 70% and  20%,
respectively) with a concomitant accumulation of
CEO(O)H (Fig. 1E) (loss of TOH significant at 7 days,
loss of CA and accumulation of CEO(O)H significant at
12 days, and loss of CL significant at 15 days. The two
other Cu
2+
-containing conditions showed significant
changes in these parameters at day 15, except in the case
of consumption of CL which did not decline at any time
point. In no case was there any statistically significant
change in the level of FC (Fig. 1B) nor any accumulation of
7K (data not shown). No statistically significant changes
were observed with the Cu
2+
-free conditions, suggesting
that Cu
2+
acts as a catalyst for GA-mediated LDL
oxidation. Experiments were performed to exclude the
possibility of GA and Cu

2+
competing for the same sites on
LDL and therefore interfering with the ability of each agent
to initiate oxidation. In these experiments, incubations were
carried out as normal but either the GA or Cu
2+
were left
out initially and then added after 1 h (an arbitrary time
frame anticipated to enable binding of either oxidant to the
surface of the LDL). In these experiments, the order of
preincubation had no impact on the rate or extent of
oxidation (data not shown) suggesting that competition for
particular sites on the LDL particle was not affecting the
extent of lipid oxidation.
Protein oxidation and modification. Astheaboveexperi-
ments demonstrated little lipid oxidation at short incuba-
tion times, it was expected that protein oxidation might also
be modest. This was confirmed by the examination of
DOPA and o-Tyr formation as a result of the incubation of
LDL with either 0.1 or 1 m
M
GA with and without added
Cu
2+
over 4 days; no significant generation of either
material was detected under these conditions (data not
shown).
Protein modification was examined in LDL samples
incubated with 1, 10, and 100 m
M

GA alone (with
1mgÆmL
)1
LDL) and 1 m
M
GA with added Cu
2+
(1 l
M
). The data obtained are presented in Fig. 4 and
expressed as percentage of control samples (no added GA).
A rapid and extensive loss of Trp fluorescence was detected
with high concentrations of GA that was time- and
concentration-dependent (Fig. 4A) and these changes
occurred in the absence of Cu
2+
; for 10 and 100 m
M
GA
loss of Trp was significant at 1 day, with higher loss of Trp
at this time point occurring in the presence of 100 m
M
GA.
For the lower concentrations of GA, losses did not become
significant until 7 days when compared with time zero.
Fig. 3. GA-mediated oxidation of LDL lipids and alpha-tocopherol.
LDL ( 0.4 mgÆmL protein
)1
) was incubated with 0, 0.1 or 1 m
M

GA
with and without 1 l
M
Cu
2+
; control samples were incubated with
1 l
M
Cu
2+
alone. At the indicated time points, 0.2 mL samples were
removed and the levels of TOH (A),FC(B),CA(C),CL(D),and
CEO(O)H (E) quantified as indicated in the legend to Fig. 1. Each data
point represents the mean (± SD) of two replicates from a single
experiment representative of several. Black bars, 1 m
M
GA + 1 l
M
Cu
2+
;horizontalstripedbars,1m
M
GA;darkstippledbars,0.1m
M
GA + 1 l
M
Cu
2+
; light stippled bars, 0.1 m
M

GA; white bars,
1 l
M
Cu
2+
. Asterisks indicate the first time point at which the data
becomes statistically different when compared with the t ¼ 0value.
01714
Time (days)
0
20
40
60
80
100
120
Arg (% of control)
01714
Time (days)
0
20
40
60
80
100
120
Lys (% of control)
01714
Time (days)
0

20
40
60
80
100
120
Trp (% of control)
140
01714
Time (days)
0
1
2
3
4
REM
5
6
140
A
*
*
*
*
*
C
*
B
D
*

*
*
*
*
Fig. 4. GA-mediated modification of apoB. LDL ( 0.4 mg proteinÆ
mL
)1
) was incubated with 1 m
M
GA ± 1 l
M
copper ions or LDL
( 1mgproteinÆmL
)1
) was incubated with 1, 10 and 100 m
M
GA.
Quantification of Trp (A), Lys (B) and Arg (C) residues was carried
out as described in the legend to Fig. 2. The data are mean
(n ¼ 4) ± SEM. REM analysis (D) was carried out as described in the
legend to Fig. 2. Black bars, 1 mg LDL proteinÆmL
)1
with 100 m
M
GA;horizontalstripedbars,1mgLDLproteinÆmL
)1
with 10 m
M
GA; dark stippled bars, 1 mg LDL proteinÆmL
)1

with 1 m
M
GA; light
stippled bars, 0.4 mg LDL proteinÆmL
)1
with 1 m
M
GA + 1 l
M
Cu
2+
; white bars, 0.4 mg LDL proteinÆmL
)1
with 1 m
M
GA only.
Asterisks indicate the first time point at which the data becomes sta-
tistically different when compared with the t ¼ 0value.
Ó FEBS 2003 Glycation and glycoxidation of LDL (Eur. J. Biochem. 270) 3577
Similar behaviour was observed for Lys (Fig. 4B). In this
case, an effect of Cu
2+
could be seen ) the high concentra-
tion GA samples, and those containing Cu
2+
, each showed
statistically significant (relative to zero time) loss of Lys
residues by day 1, while those samples containing 1 m
M
GA

with no added Cu
2+
did not show significant loss of Lys
until day 7. The effect of GA on the Arg residues of apoB
is shown in Fig. 4C; these data are expressed relative to the
controls at each time point and zero time values of all
samples (data not shown) were compared to exclude the
possibility of interference by GA with the assay. In the
presence of 100 m
M
GA,  50% of the Arg residues were
observed to be lost in the samples examined immediately
after mixing (t ¼ 0 samples) and no further loss was
observed at longer time points. In none of the other samples
was any statistically significant loss of Arg residues detected.
Fig. 4D shows the changes in REM for analogous LDL
incubations. Native LDL and AcLDL were also examined
but these data are not shown for reasons of clarity; the
values obtained for these materials were within the expected
range (see above). Incubations with 10 or 100 m
M
GA
showed maximal changes in REM as early as 24 h (up to
4.8-fold increase), with little increase after this time, while
the lower concentrations of GA facilitated small increases
by 24 h ( 1.8-fold) increasing to  2.7-fold by 14 days.
These changes in LDL mobility also occurred in the absence
of added copper ions.
Discussion
Although there is a direct parallel between increased

blood sugar and the clinical state of diabetes, and
extensive support for the theory that increased oxidative
stress is involved in this pathology [4,7], detailed studies
on the processes of glycation and glycoxidation have not
provided a conclusive causative mechanism (or mech-
anisms) for the damage induced by high glucose concen-
trations. It has been clearly established that products of
metal ion catalysed oxidation of glucose and protein-
bound glucose (i.e. glycoxidation) can accumulate at
elevated levels on proteins from diabetic patients, as do
products of direct covalent modification (glycation)
[3,4,7]. Whether one or both of these processes is
causative in the development of the complications of
diabetes, such as atherosclerosis, is less well established
[4]. It is therefore pertinent to establish the relative roles
of both oxidative processes and direct glycation (covalent,
nonoxidative) reactions in the development of atheroscle-
rosis, in particular the modification of LDL which might
promote the formation of lipid-laden (foam) cells in the
artery wall; a hallmark of early atherogenesis. Several
studies have reported elevated levels of oxidized- and/or
AGE-modified LDL in diabetic subjects [34–36] and in
human atherosclerotic lesions [37]. It has also been
reported that glycoxidized and peroxidized LDL colocal-
ize with the macrophage scavenger receptor [38] indicat-
ing a plausible involvement of the accumulation of AGEs
and increased oxidative stress in this pathology. Further-
more, protein-bound sugars have been demonstrated to
generate free radicals (particularly in the presence of
metal ions [39]), which could potentiate further damage,

including lipid peroxidation [40–42], while the suscepti-
bility of LDL to Cu
2+
-induced oxidation has been shown
to increase in the presence of glucose [43,44].
In addition to the hyperglycaemia seen in poorly
controlled diabetic patients, and its potential involvement
in the accumulation of AGEs, a number of studies have
examined the role of low molecular mass aldehydes such as
glyoxal, MG, and GA. These materials are formed both as a
consequence of oxidative processes and AGE modifications
of proteins as well as by a variety of nonrelated metabolic
processes [45,46]. Evidence has been presented for an
elevated level of these materials (or products arising from
them) in diabetics [47,48] and the presence of antibodies to
MG-derivatized proteins in corneal collagen and plasma
proteins [49]. We have therefore carried out a comprehen-
sive analysis of the relative efficacies of glucose and two
aldehydes (MG and GA) in inducing lipid and protein
oxidation and antioxidant depletion of LDL particles as
well as glycation of the apoB protein by measuring specific
parameters of these processes. As previous workers
[44,50,51] have presented data demonstrating that lipid
peroxidation and protein modification of LDL by glucose
can be dependent on the presence of transition metal ions,
studies were also carried out in the presence of low levels of
Cu
2+
. This transition metal ion dependence may indicate a
cooperative effect of glucose and Cu

2+
as glucose has been
reportedtoincreaseCu
2+
-induced LDL oxidation without
affecting oxidation by aqueous peroxyl radicals [52]. Glu-
cose- and transition metal ion-dependent protein oxidation
hasalsobeendetectedinstudiesonrattailcollagenas
measured by accumulation of the specific protein side chain
oxidation products DOPA, m-Tyr, di-Tyr, and Leu and Val
alcohols [53].
In the current study it has been shown that the time
course of oxidation of antioxidants, lipids, and protein side
chains in LDL is slow in the presence of glucose alone,
though marginally faster than in control samples with no
additions. The time course of oxidation was much more
rapid in the presence of low concentrations of Cu
2+
,
compared with its absence, but there were no significant
differences between the samples treated with Cu
2+
alone
compared with those containing Cu
2+
plus any of the
glucose concentrations. This suggests that the observed
reactions are primarily due to oxidation catalysed by Cu
2+
alone and that glucose does not play a major role in these

reactions over the concentration range studied, the highest
ofwhichiswellinexcessofthatobservedeveninvery
poorly controlled diabetics. It has been shown that the
oxidation of LDL by Cu
2+
is saturable, due to the limited
number of high-affinity Cu
2+
-binding sites on the LDL
particle [54], but the experiments performed here were
carried out with Cu
2+
:LDLratios( 1.2 : 1) that are well
below the lower threshold of 5–6 postulated by the these
workers, and hence the absence of any stimulatory effect of
glucose cannot be ascribed to a saturation effect. This
marginal effect of glucose is in accord with some previous
reports which have shown that elevated levels of glucose
failed to potentiate the accumulation of the protein
oxidation product o-Tyr in skin collagen [55] and urine
[56] from diabetics compared with nondiabetics, suggesting
that the accumulation of this oxidation product is unaffect-
ed by the level of hyperglycaemia. The absence of any effect
of glucose on Cu
2+
-stimulated oxidation is, however, in
3578 H. M. Knott et al. (Eur. J. Biochem. 270) Ó FEBS 2003
contrast to another recent study which has reported a
stimulatory effect of glucose on Cu
2+

-mediated LDL
oxidation [44]. The latter study used Cu
2+
:LDL ratios
which were higher that those used in the current study (3 : 1
and 31 : 1) which may account for the observed differences
in behaviour; the lower ratios used in the current study are
likely to be the more physiologically relevant.
A similar absence of any stimulatory effect on the
oxidation of the lipid, protein, and antioxidants of LDL,
above that seen with Cu
2+
alone, was observed with MG.
Within this system, even the presence of Cu
2+
alone
induced only very limited oxidation. In contrast GA, which
is more chemically reactive than either glucose or MG, did
induce the oxidation of lipids and consumption of
a-tocopherol in LDL with this process being GA concen-
tration-dependent although GA in the absence of Cu
2+
had
little effect. No protein oxidation was detected in the Cu
2+
plus GA system, suggesting that such oxidation occurs after
the induction of lipid oxidation and potentially as a result of
damage transfer from the oxidized lipids to the protein
component; this possibility was not investigated further.
Overall, the oxidation of lipids and protein side chains in

LDL and the depletion of a-tocopherol by glucose and the
two aldehydes examined appears to proceed slowly, even in
the presence of low concentrations of Cu
2+
.
In contrast with the above processes, covalent modifica-
tion (glycation) of LDL has been shown to occur rapidly
with the aldehydes, occur in the absence of added metal
ions, and depend on both the structure of the compound
and its concentration. Of the two aldehydes examined,
modification of Trp and Lys was more rapid with GA than
with MG and the loss of Lys appears to occur earlier than
that of Trp in the MG system. With GA the loss of these
two amino acids occured too rapidly to determine an order
of reaction. Loss of Arg occurred prior to either of these
other amino acids as evidenced by the significant loss of this
residue in the time zero samples that were analysed
immediately after mixing. A more extensive loss of Arg
was detected with GA compared with MG at this time point
but again the reactions were too rapid to analyse statisti-
cally. Thus with MG the order of depletion of these residues
is Arg > Lys > Trp which is in accord with previous
studies on the reaction of MG with LDL [33], and GA and
MG with Lys and Arg residues on other proteins [6,9,57–
59]. Adducts formed with both residues have been identified
on plasma proteins [10,58].
The mechanisms of modification of the Lys and Arg
residues, and the products formed, have not been examined;
it is likely that these reactions proceed by the pathways
outlined previously (e.g. [6,9]). The loss of Trp fluorescence

was observed to occur in the absence of added Cu
2+
.
Whether this reflects conversion of Trp to products (e.g. via
the kynurenine pathway) or alteration in the local environ-
ment of these (hydrophobic) residues cannot be clearly
differentiated from the current data [60]. The observed
alteration of Lys and Arg residues, which are major
contributors (with protonated His side-chains) to the overall
charge of the LDL particle, are mirrored in the observed
changes in the relative electrophoretic mobility of the
modified particles.
In contrast with the behaviour of GA and MG, no
modification of either Trp or Lys residues was detected even
with the highest levels of glucose used (100 m
M
) in either the
absence or presence of Cu
2+
. Limited modification of Arg
residues was however detected with high concentrations of
glucose with and without added Cu
2+
. These observations
areinaccordwithamorerapidrateofmodificationofArg
residues over Lys residues, particularly given the higher
concentration of Lys residues over Arg in apoB (356 Lys vs.
148 Arg; Swiss-Prot file P04114, residues 28–4563). Fur-
thermore, the absence of significant levels of Lys and Trp
modification in systems where Cu

2+
was added, and where
significant levels of lipid oxidation were detected (cf. the
data obtained after 2 weeks of incubation in Fig. 1),
suggests that modification of these residues by lipid
oxidation products does not occur to any significant extent
under the conditions used. This is in contrast to previous
reports which have suggested that lipid oxidation is a major
route to the modification of protein side chains such as Lys
residues on LDL [50]. In accordance with these measure-
ments, only small changes were detected in the REM of the
glucose-treated LDL particles except with both supra-
pathological levels of glucose and long incubation times.
It has been established that neither glucose nor MG are
effective catalysts of oxidation of the major components of
LDL particles in either the presence or absence of added
Cu
2+
; this inability of MG to initiate LDL oxidation agrees
with a previous report [33]. In contrast, GA can facilitate
lipid peroxidation induced by Cu
2+
but is relatively
ineffective in the absence of such metal ions. These results
indicate that the glycoxidation of LDL, as has been
suggested previously [52], is transition metal ion-dependent.
Interestingly, GA, but not MG, can promote oxidation and
this is likely to be due to the oxidizable b-hydroxyaldehyde
function [– CH(OH)-C(O)-] on this molecule which is not
present in MG which contains the corresponding oxidized

a,b-dicarbonyl function [i.e. (– C(O)-C(O)-)]. Reaction of
the b-hydroxyaldehyde function with Cu
2+
is believed to
result in the formation of the reduced metal ion Cu
+
and a
radical anion species from the GA (e.g. [61]). Further
reactions of one or both of these species may give rise to the
observed induction of lipid peroxidation. Similar reactions
cannot occur with MG but are likely to occur, albeit at a
much lower rate, with the open-chain form of glucose; the
low concentration of this latter species in equilibrium
mixtures of glucose anomers is likely to be at least partially
responsible for the slow reaction kinetics detected when
compared with equimolar amounts of GA [61]. In contrast
with such oxidative reactions, covalent modification (gly-
cation) appears to be a much more rapid and potentially
more significant process, particularly as there is controversy
regarding the presence of significant concentrations of
reactive transition metal ions in the artery wall and in
developing atherosclerotic lesions [62–64].
The loss (derivatization, covalent modification) of Lys and
Arg residues may be highly pertinent to the biological effects
of modified LDL particles as it has been shown that
modification of Lys residues promotes recognition of LDL
by the macrophage scavenger receptor [65,66]. The effect of
specific modification of Arg residues on LDL particle recog-
nition is less well established. Preliminary data (B. E. Brown,
H. M. Knott, R. T. Dean and M. J. Davies, unpublished

results) on the uptake of LDL particles modified by GA,
MG, and glucose prepared under similar conditions to
Ó FEBS 2003 Glycation and glycoxidation of LDL (Eur. J. Biochem. 270) 3579
those used in the studies reported here are consistent with the
recognition of GA- (and possibly MG-) modified LDL
particles, but not glucose-modified species, by receptors
present on mouse macrophage-like cells (cf. data with other
modified proteins [59,67]) and the subsequent accumulation
of lipids (cholesterol and cholesteryl esters, as measured by
HPLC) within these cells. These data support a previous
suggestion [68] that covalent modification (glycation), in the
absence of lipid or protein oxidation or antioxidant
consumption, is sufficient for the formation of foam
cells. Such Ôcarbonyl stressÕ may play a significant role in
the modification of LDL particles both in plasma and in the
intima of the artery wall and may therefore contribute to the
elevated (two- to threefold) levels of atherogenesis observed
in diabetic patients compared with nondiabetic patients.
Acknowledgements
This work was supported by the National Health and Medical
Research Council, the Australian Research Council, the Juvenile
Diabetes Foundation International, Diabetes Australia Research Trust,
and the Wellcome Trust. B.E. Brown gratefully acknowledges receipt
of an Australian Postgraduate Award, administered through The
University of Sydney.
References
1. Kannel, W.B. & McGee, D.L. (1979) Diabetes and cardiovascular
disease: the Framingham Study. Diabetes 241, 2035–2038.
2. Thorpe, S. & Baynes, J.W. (1996) Role of the Maillard
Reaction in Diabetes Mellitus and disease of aging. Drugs Aging 9,

69–77.
3. Baynes, J.W. & Thorpe, S.R. (1999) Role of oxidative stress in
diabetic complications: a new perspective on an old paradigm.
Diabetes 48, 1–9.
4. Baynes, J.W. & Thorpe, S.R. (2000) Glycoxidation and lipoxi-
dation in atherogenesis. Free Radic. Biol. Med. 28, 1708–1716.
5. Wolff, S.P., Jiang, A.Y. & Hunt, J.V. (1991) Protein glycation and
oxidative stress in diabetes mellitus and ageing. Free Radic. Biol.
Med. 10, 339–352.
6. Lo, T.W., Westwood, M.E., McLellan, A.C., Selwood, T. &
Thornalley, P.J. (1994) Binding and modification of proteins by
methylglyoxal under physiological conditions. A kinetic and
mechanistic study with N-alpha-acetylarginine, N-alpha-acetyl-
cysteine, and N-alpha-acetyllysine, and bovine serum albumin.
J. Biol. Chem. 269, 32299–32305.
7. Brownlee, M. (1995) Advanced protein glycosylation in diabetes
and aging. Annu. Rev. Med. 46, 223–234.
8. Fu, M X., Requena, J.R., Jenkins, A.J., Lyons, T.J., Baynes, J.W.
& Thorpe, S.R. (1996) The advanced glycation end product
N
e
-(carboxymethyl) lysine is a product of both lipid peroxidation
and glycoxidation reactions. J. Biol. Chem. 271, 9982–9986.
9. Degenhardt, T.P., Thorpe, S.R. & Baynes, J.W. (1998) Chemical
modification of proteins by methylglyoxal. Cell. Mol. Biol. 44,
1139–1145.
10. Nagaraj, R.H., Shipanova, I.N. & Faust, F.M. (1996) Protein
cross-linking by the Maillard reaction. Isolation, characterization,
and in vivo detection of a lysine-lysine cross-link derived from
methylglyoxal. J. Biol. Chem. 271, 19338–19345.

11. Glomb, M.A. & Monnier, V.M. (1995) Mechanism of protein
modification by glyoxal and glycolaldehyde, reactive intermediates
of the Maillard reaction. J. Biol. Chem. 270, 10017–10026.
12. Atkins, T.W. & Thornally, P.J. (1989) Erythrocyte glyoxalase
activity in genetically obese (ob/ob) and streptozotocin diabetic
mice. Diabetes Res. 11, 125–129.
13. Odani, H., Shinzato, T., Matsumoto, Y., Usami, J. & Maeda, K.
(1999) Increase in three a,b-dicarbonyl compound levels in human
uremic plasma: specific in vivo determination of intermediates in
advanced Maillard reaction. Biochem. Biophys. Res. Commun.
256, 89–93.
14. Reichard,G.A.,Jr,Skutches,C.L.,Hoeldtke,R.D.&Owen,O.E.
(1986) Acetone metabolism in humans during diabetic keto-
acidosis. Diabetes 35, 668–674.
15. Brown, M.S. & Goldstein, J.L. (1986) A receptor-mediated
pathway for cholesterol homeostasis. Science 232, 34–47.
16. Araki, N., Higashi, T., Mori, T., Shibayama, R., Kawabe, Y.,
Kodama, T., Takahashi, K., Shichiri, M. & Horiuchi, S. (1995)
Macrophage scavenger receptor mediates the endocytic uptake
and degradation of advanced glycation end products of the
Maillard reaction. Eur. J. Biochem. 230, 408–415.
17. Suzuki, H., Kurihara, Y., Takeya, M., Kamada, N., Kataoka, M.,
Jishage,K.,Ueda,O.,Sakaguchi,H.,Higashi,T.,Suzuki,T.,
Takashima, Y., Kawabe, Y., Cynshi, O., Wada, Y., Honda, M.,
Kurihara, H., Aburatani, H., Doi, T., Matsumoto, A., Azuma, S.,
Noda, T., Toyoda, Y., Itakura, H., Yazaki, Y. & Kodama, T.
(1997) A role for macrophage scavenger receptors in athero-
sclerosis and susceptibility to infection. Nature 386, 292–296.
18. Dean, R.T. (1991) Protein damage and repair. In Oxidative
Damage and Repair (Davies, K.J.A., ed), pp. 341–347. Pergamon

Press, Oxford.
19. Bourdon, E., Loreau, N. & Blacher, D. (1999) Glucose and free
radicals impair the antioxidant properties of serum albumin.
FASEB J. 13, 233–244.
20. van Reyk, D.M., Brown, A.J., Jessup, W. & Dean, R.T. (1995)
Batch-to-batch variation of Chelex-100 confounds metal-cata-
lysed oxidation. Leaching of inhibitory compounds from a batch
of Chelex-100 and their removal by a pre-washing procedure. Free
Radic. Res. 23, 533–535.
21. Jessup, W., Simpson, J.A. & Dean, R.T. (1993) Does superoxide
radical have a role in macrophage-mediated oxidative modifica-
tion of LDL? Atherosclerosis 99, 107–120.
22. Esterbauer, H., Striegl, G., Puhl, H. & Rotheneder, M. (1989)
Continuous monitoring of in vitro oxidation of human low density
lipoprotein. Free Radic. Res. Commun. 6, 67–75.
23. Fu, S., Davies, M.J., Stocker, R. & Dean, R.T. (1998) Evidence for
rolesofradicalsinproteinoxidationinadvancedhumanathero-
sclerotic plaque. Biochem. J. 333, 519–525.
24. Bruce, D., Fu, S., Armstrong, S. & Dean, R.T. (1999) Human
apo-lipoprotein B from normal plasma contains oxidised peptides.
Int. J. Biochem. Cell. Biol. 31, 1409–1420.
25. Kritharides, L., Upston, J., Jessup, W. & Dean, R.T. (1998)
Accumulation and metabolism of low density lipoprotein-derived
cholesteryl linoleate hydroperoxide and hydroxide by macro-
phages. J. Lipid Res. 39, 2394–2405.
26. Udenfriend, S., Stein, S., Bohlen, P., Dairman, W., Leimgruber,
W. & Weigele, M. (1972) Fluorescamine: a reagent for assay of
amino acids, peptides, proteins, and primary amines in the pico-
mole range. Science 178, 871–872.
27. Smith, R.E. & MacQuarrie, R. (1978) A sensitive fluorometric

method for the determination of arginine using 9,10-phenan-
threnequinone. Anal. Biochem. 90, 246–255.
28. Knott, H.M., Baoutina, A., Davies, M.J. & Dean, R.T. (2002)
Comparative time-courses of copper-ion-mediated protein and
lipid oxidation in low-density lipoprotein. Arch. Biochem. Biophys.
400, 223–232.
29. Armstrong, S.G. & Dean, R.T. (1995) A sensitive fluorometric
assay for protein-bound DOPA and related products of radical-
mediated protein oxidation. Redox Report 1, 291–298.
30. Fu, M.X., Wells-Knecht, K.J., Blackledge, J.A., Lyons, T.J.,
Thorpe, S.R. & Baynes, J.W. (1994) Glycation, glycoxidation, and
cross-linking of collagen by glucose. Kinetics, mechanisms, and
3580 H. M. Knott et al. (Eur. J. Biochem. 270) Ó FEBS 2003
inhibition of late stages of the Maillard reaction. Diabetes 43,
676–683.
31. Davies, M.J., Fu, S., Wang, H. & Dean, R.T. (1999) Stable
markers of oxidant damage to proteins and their application in the
study of human disease. Free Radic. Biol. Med. 27, 1151–1163.
32. Davies, M.J. & Dean, R.T. (1997) Radical-Mediated Protein
Oxidation: from Chemistry to Medicine. Oxford University Press,
Oxford.
33. Schalkwijk, C.G., Vermeer, M.A., Stehouwer, C.D., te Koppele,
J.,Princen,H.M.&vanHinsbergh,V.W.(1998)Effectof
methylglyoxal on the physico-chemical and biological properties
of low-density lipoprotein. Biochim. Biophys. Acta 1394, 187–198.
34. Bucala,R.,Makita,Z.,Koschinsky,T.,Cerami,A.&Vlassara,H.
(1993) Lipid advanced glycosylation: Pathway for lipid oxidation
in vivo. Proc. Natl Acad. Sci. USA 90, 6434–6438.
35. Bucala, R., Makita, Z., Vega, G., Grundy, S., Koschinsky, T.,
Cerami, A. & Vlassara, H. (1994) Modification of low density

lipoprotein by advanced glycation end products contributes to the
dyslipidemia of diabetes and renal insufficiency. Proc.NatlAcad.
Sci. USA 91, 9441–9445.
36. Frye, E.B., Degenhardt, T.P., Thorpe, S.R. & Baynes, J.W. (1998)
Role of the Maillard reaction in aging of tissue proteins. Advanced
glycosylation end product-dependent increase in imidazolium
cross-links in human lens proteins. J. Biol. Chem. 273, 18714–
18719.
37. Sakata, N., Imanaga, Y., Meng, J., Tachikawa, Y., Takebayashi,
S., Nagai, R., Horiuchi, S., Itabe, H. & Takano, T. (1998)
Immunohistochemical localization of different epitopes of
advanced glycation end products in human atherosclerotic lesions.
Atherosclerosis 141, 61–75.
38. Sakata, N., Uesugi, N., Takebayashi, S., Nagai, R., Jono, T.,
Horiuchi, S., Takeya, M., Itabe, H., Takano, T., Myint, T. &
Taniguchi, N. (2001) Glycoxidation and lipid peroxidation of low-
density lipoprotein can synergistically enhance atherogenesis.
Cardiovasc. Res. 49, 466–475.
39. Sajithlal, G.B., Chithra, P. & Chandrakasan, G. (1998) The role of
metal-catalyzed oxidation in the formation of advanced glycation
end products: an in vitro study on collagen. Free Radic. Biol. Med.
25, 265–269.
40. Hunt, J.V., Smith, C.C.T. & Wolff, S.P. (1990) Autoxidative
glycosylation and possible involvement of peroxides and free
radicals in LDL modification by glucose. Diabetes 39, 1420–1424.
41. Sakurai, T., Kimura, S., Nakano, M. & Kimura, H. (1991) Oxi-
dative modification of glycated low density lipoprotein in the
presence of iron. Biochem. Biophys. Res. Commun. 177, 433–439.
42. Mullarkey, C.J., Edelstein, D. & Brownlee, M. (1990) Free radical
generation by early glycation products: a mechanism for

accelerated atherogenesis in diabetes. Biochem. Biophys. Res.
Commun. 173, 932–939.
43. Kobayashi,K.,Watanabe,J.,Umeda,F.&Nawata,H.(1995)
Glycation accelerates the oxidation of low density lipoprotein by
copper ions. Endocrinol. J. 42, 461–465.
44.Leoni,V.,Albertini,A.,Passi,A.,Abuja,P.M.,Borroni,P.,
D’Eril, G.M. & de Luca, G. (2002) Glucose accelerates copper-
and ceruloplasmin-induced oxidation of low-density lipoprotein
and whole serum. Free Radic. Res. 36, 521–529.
45. Thornalley, P.J. (1996) Pharmacology of methylglyoxal: forma-
tion, modification of proteins and nucleic acids, and enzymatic
detoxification – a role in pathogenesis and antiproliferative
chemotherapy. Gen. Pharmacol. 27, 565–573.
46. Thornalley, P.J., Langborg, A. & Minhas, H.S. (1999) Formation
of glyoxal, methylglyoxal and 3-deoxyglucosone in the glycation
of proteins by glucose. Biochem. J. 344, 109–116.
47. McLellan, A.C., Phillips, S.A. & Thornalley, P.J. (1992) The assay
of methylglyoxal in biological systems by derivatization with
1,2-diamino-4,5-dimethoxybenzene. Anal. Biochem. 206, 17–23.
48. McLellan, A.C., Thornalley, P.J., Benn, J. & Sonksen, P.H. (1994)
Glyoxalase system in clinical diabetes mellitus and correlation with
diabetic complications. Clin. Sci. 87, 21–29.
49. Shamsi, F.A., Partal, A., Sady, C., Glomb, M.A. & Nagaraj, R.H.
(1998) Immunological evidence for methylglyoxal-derived modi-
fications in vivo. Determination of antigenic epitopes. J. Biol.
Chem. 273, 6928–6936.
50. Kawamura, M., Heinecke, J.W. & Chait, A. (1994) Pathophy-
siological concentrations of glucose promote oxidative modifica-
tion of low density lipoprotein by a superoxide-dependent
pathway. J. Clin. Invest. 94, 771–778.

51. Millican, S.A., Schultz, D., Bagga, M., Coussons, P.J., Mu
¨
ller, K.
& Hunt, J.V. (1998) Glucose-modified low density lipoprotein
enhances human monocyte chemotaxis. Free Radic. Res. 28,
533–542.
52. Mowri, H., Frei, B. & Keaney, J.F. (2000) Glucose enhancement
of LDL oxidation is strictly metal ion dependent. Free Radic. Biol.
Med. 29, 814–824.
53. Fu, S L., Fu, M X., Baynes, J.W., Thorpe, S.R. & Dean, R.T.
(1998) Presence of dopa and amino acid hydroperoxides in pro-
teins modified with advanced glycation end products (AGEs):
amino acid oxidation products as a possible source of oxidative
stress induced by AGE proteins. Biochem. J. 330, 233–239.
54. Gieseg, S.P. & Esterbauer, H. (1994) Low density lipoprotein is
saturable by pro-oxidant copper. FEBS Lett. 343, 188–194.
55. Wells-Knecht,M.C.,Lyons,T.J.,McCance,D.R.,Thorpe,S.R.&
Baynes, J.W. (1997) Age-dependent increase in ortho-tyrosine and
methionine sulfoxide in human skin collagen is not accelerated in
diabetes. Evidence against a generalized increase in oxidative stress
in diabetes. J. Clin. Invest. 100, 839–846.
56. Krapfenbauer, K., Birnbacher, R., Vierhapper, H., Herkner, K.,
Kampel, D. & Lubec, G. (1998) Glycoxidation, and protein and
DNA oxidation in patients with diabetes mellitus. Clin. Sci. 95,
331–337.
57. Oya, T., Hattori, N., Mizumo, Y., Miyata, S., Maeda, S., Osawa,
T. & Uchida, K. (1999) Methylglyoxal modification of proteins.
Chemical and immunohistochemical characterization of methyl-
glyoxal-arginine adducts. J. Biol. Chem. 274, 18495–18502.
58. Westwood, M.E. & Thornalley, P.J. (1995) Molecular character-

istics of methylglyoxal-modified bovine and human serum albu-
mins. Comparison with glucose-derived advanced glycation
endproduct-modified serum albumins. J. Prot. Chem. 14, 359–372.
59. Nagai, R., Matsumoto, K., Ling, X., Suzuki, H., Araki, T. &
Horiuchi, S. (2000) Glycolaldehyde, a reactive intermediate for
advanced glycation end products, plays an important role in the
generation of an active ligand for the macrophage scavenger
receptor. Diabetes 49, 1714–1723.
60. Yang, C Y., Gu, Z W., Yang, M., Lin, S N., Siuzdak, G. &
Smith, C.V. (1999) Identification of modified tryptophan residues
in apolipoprotein B-100 derived from copper ion-oxidized low-
density lipoprotein. Biochemistry 38, 15903–15908.
61. Thornalley, P.J. (1985) Monosaccharide autoxidation in health
and disease. Environ. Health Perspect. 64, 297–307.
62. Evans, P.J., Smith, C., Mitchinson, M.J. & Halliwell, B. (1995)
Metal ion release from mechanically-disrupted human arterial
wall. Implications for the development of atherosclerosis. Free
Radic. Res. 23, 465–469.
63. Smith, C., Mitchinson, M.J., Aruoma, O.I. & Halliwell, B. (1992)
Stimulation of lipid peroxidation and hydroxyl-radical generation
by the contents of human atherosclerotic lesions. Biochem. J. 286,
901–905.
64. Swain, J. & Gutteridge, J.M.C. (1995) Prooxidant iron and cop-
per, with ferroxidase and xanthine oxidase activities in human
atherosclerotic material. FEBS Lett. 368, 513–515.
65. Haberland, M.E., Olch, C.L. & Fogelman, A.M. (1984) Role
of lysines in mediating interaction of modified low density
Ó FEBS 2003 Glycation and glycoxidation of LDL (Eur. J. Biochem. 270) 3581
lipoproteinswiththescavengerreceptorofhumanmonocyte
macrophages. J. Biol. Chem. 259, 11305–11311.

66. Jinnouchi, Y., Sano, H., Nagai, R., Hakamata, H., Kodama, T.,
Suzuki, H., Yoshida, M., Ueda, S. & Horiuchi, S. (1998) Glyco-
laldehyde-modified low density lipoprotein leads macrophages to
foam cells via the macrophage scavenger receptor. J. Biochem.
123, 1208–1217.
67. Westwood, M.E., Argirov, O.K., Abordo, E.A. & Thornalley, P.J.
(1997) Methylglyoxal-modified arginine residues – a signal for
receptor-mediated endocytosis and degradation of proteins by
monocytic THP-1 cells. Biochim. Biophys. Acta 1356, 84–94.
68. Colaco, C.A. & Roser, B.J. (1994) Atherosclerosis and glycation.
Bioessays 16, 145–147.
3582 H. M. Knott et al. (Eur. J. Biochem. 270) Ó FEBS 2003