(The FASEB Journal. 1999;13:1157-1168.)
© 1999 FASEB
New developments in the isoprostane pathway: identification of novel highly reactive 
-ketoaldehydes (isolevuglandins) and characterization of their protein adducts
L. JACKSON ROBERTS, II*1,
ROBERT G. SALOMON
,
JASON D. MORROW* and
CYNTHIA J. BRAME*
* Departments of Pharmacology and Medicine, Vanderbilt University, Nashville, Tennessee, 37232-6602, USA; and
Department of Chemistry, Case Western Reserve University, Cleveland, Ohio 44106, USA
1Correspondence: Department of Pharmacology, Vanderbilt University, Nashville, TN 37232-6602, USA. E-mail: jack.roberts{at}mcmail.vanderbilt.edu
 |
ABSTRACT
|
|---|
The bicyclic endoperoxide prostaglandin (PG) H2
undergoes nonenzymatic rearrangement not only to PGE2 and
PGD2, but also to levuglandins (LG) E2 and
D2, which are highly reactive 
-ketoaldehydes.
Isoprostanes (IsoPs) are PG-like compounds that are produced by
nonenzymatic peroxidation of arachidonic acid. PGH2-like
endoperoxides are intermediates in this pathway. Therefore, we explored
whether the IsoP endoperoxides also undergo rearrangement to form
IsoLGs. Oxidation of arachidonic acid in vitro resulted
in the formation of abundant quantities of compounds that were
established to be IsoLGs by using mass spectrometric analyses. However,
the formation of IsoLGs could not be detected in biological systems
subjected to an oxidant stress. We hypothesized that this was due to
extremely rapid adduction of IsoLGs to proteins. This notion was
supported by the finding that LGE2 adducted to albumin at a
rate that exceeded that of 4-hydroxynonenal by several orders of
magnitude: >50% of LGE2 had adducted within 20 s. We
therefore undertook to characterize the nature of LG adducts. Using
liquid chromatography electrospray tandem mass spectrometry, we
established that LGs form oxidized pyrrole adducts (lactams and
hydroxylactams) with the 
-amino group of lysine. Oxidation of low
density lipoprotein resulted in readily detectable IsoLG adducts on
apolipoprotein B after enzymatic digestion of the protein to individual
amino acids. These studies identify a novel class of ketoaldehydes
produced by the IsoP pathway that form covalent protein adducts at a
rate that greatly exceeds that of other known aldehyde products of
lipid peroxidation. Elucidation of the nature of the adducts formed by
IsoLGs provides the basis to explore the formation of IsoLGs in
vivo and investigate the potential biological ramifications of
their formation in settings of oxidant injury.—Roberts, L. J.,
II, Salomon, R. G., Morrow, J. D., Brame, C. J. New
developments in the isoprostane pathway: identification of novel highly
reactive -ketoaldehydes (isolevuglandins) and characterization of
their protein adducts.
Key Words: free radical • oxidant injury • lipid peroxidation • aldehyde
|
BACKGROUND
|
|---|
IN 1990, WE REPORTED the discovery of prostaglandin
(PG)2
(1)
F2-like compounds, now termed
F2-isoprostanes (F2-IsoPs),
which are produced in vivo by nonenzymatic free
radical-induced peroxidation of arachidonic acid (1)
. We
subsequently demonstrated that F2-IsoPs are
initially formed in situ esterified to phospholipids and
subsequently released in free form by a phospholipase(s)
(2)
. F2-IsoPs are present in readily
detectable levels in all normal animal and human biological fluids and
tissues. This indicates a level of ongoing lipid peroxidation in the
normal state that is incompletely suppressed by the elaborate system of
antioxidant defenses that have evolved to prevent oxidative damage.
Although the archetypical compounds discovered to be formed by this
mechanism contained an F-type prostane ring, subsequent studies have
greatly expanded the number of different types of compounds that can be
formed as products of the IsoP pathway. Central to the formation of
IsoPs are intermediate PGH2-like bicyclic
endoperoxides. F2-IsoPs are formed by reduction
of the endoperoxides. Recently we reported that glutathione is an
important effector of this reduction (3)
. However,
endoperoxides are unstable in aqueous solution, and if not efficiently
reduced will undergo rearrangement. In this regard, we have reported
that PGE2-like and
PGD2-like compounds
(E2/D2-IsoPs) and
thromboxane-like compounds (isothromboxanes) are formed in
vivo as rearrangement products of the IsoP endoperoxides
(Fig. 1
) (4
, 5)
.
View larger version (13K):
[in this window]
[in a new window]
|
Figure 1. IsoP pathway leading to the formation of F-ring IsoPs by reduction of
IsoP endoperoxides and formation of E-ring, D-ring, and
thromboxane-ring compounds by rearrangement of IsoP endoperoxides.
|
|
The importance of the discovery of IsoPs encompasses two areas. One is
the use of measurements of F2-IsoPs to assess
oxidative stress status in vivo. The second relates to the
biological actions exerted by IsoPs that may be relevant in mediating
some of the effects of oxidant injury. Measuring
F2-IsoPs has emerged as one of the most valuable
and reliable approaches to assess lipid peroxidation in
vivo. Using measurements of F2-IsoPs has
implicated a role for free radicals in the pathogenesis of a wide
variety of disease processes. As this report will focus on recent
developments related to the IsoPs, readers are referred to recent
reviews that discuss in some detail various diseases in which a role
for oxidant injury has been suggested by a finding of overproduction of
F2-IsoPs and the currently known biological
actions of IsoPs (6
7
8
9)
.
As mentioned above, central in the pathway of formation of IsoPs
are the IsoP endoperoxides, which are reduced to form
F2-IsoPs but also undergo rearrangement in
vivo. It had been shown in the 1970s that
PGH2 undergoes rearrangement in aqueous buffers
to form PGE2 and PGD2 with
a t1/2 of ~5 min (10)
. In 1984, Salomon
also discovered that PGH2 rearranges to
form acyclic -ketoaldehydes at about a 20% yield (11)
.
These compounds have been termed levuglandin (LG)
E2 and D2 because of their
structural similarity to levulinaldehyde. Interest in LGs stems from
the fact that they have been found to be remarkably reactive molecules
that rapidly adduct to proteins and also undergo further reaction to
form extensive protein/protein and protein/DNA cross-links
(12
13
14
15)
. In light of the fact that IsoP endoperoxides
undergo rearrangement in vivo, we explored whether LG-like
compounds are formed as products of the IsoP pathway, which we propose
to term IsoLGs. In the IsoP pathway, four
endoperoxide regioisomers are formed, each comprised of eight racemic
diastereomers for a total of 64 compounds (Fig. 1)
. We hypothesized
that the endoperoxides could undergo rearrangement to form IsoLGs as
outlined in Fig. 2
. As noted, four IsoLGD2 and four
IsoLGE2 regioisomers are formed, each of which is
theoretically comprised of four racemic diastereomers. According to the
nomenclature system for IsoPs approved by the Eicosanoid Nomenclature
Committee, sanctioned by JCBN of IUPAC, the four
regioisomers shown are designated by the carbon number on
which the side chain hydroxyl is located (16)
.
View larger version (18K):
[in this window]
[in a new window]
|
Figure 2. Predicted pathway of formation of IsoLGD2 and
IsoLGE2 compounds. Four IsoLGE2 and
IsoLGD2 regioisomers are formed, each theoretically
comprised of four racemic diastereomers.
|
|
|
RECENT RESULTS
|
|---|
Initially we explored whether we could detect the formation of
IsoLGs during oxidation of arachidonic acid in vitro.
Arachidonic acid (5 mg) was oxidized using a mixture of
FeCl3 (1 mM)/ADP (200 mM)/ascorbate (100 mM) in
50 mM phosphate buffer (pH 7.4) at 37°C for 6 h. Compounds were
then converted to O-methyloxime derivatives by addition of
3% methoxamine and incubated for 45 min at room temperature. The
sample was acidified to pH 3, loaded onto a C-18 SepPak cartridge, and
washed sequentially with 10 ml pH 3 water, 10 ml heptane/ethyl acetate
(99:1,v/v); compounds were eluted with 10 ml of heptane/ethyl acetate
(1:1,v/v). The compounds were then converted to a pentafluorobenzyl
(PFB) ester derivative by treatment with 40 µl of 10%
pentafluorobenzyl bromide in acetonitrile and 20 µl of 10%
diisopropylethylamine in acetonitrile for 20 min at 37°C. Synthetic
LGE2 was converted to a bis
[2H3]
O-methyloxime, PFB ester derivative and 1–4 ng was added to
the sample as an internal standard. Compounds were then subjected to
thin-layer chromatography using a solvent system of heptane/ethyl
acetate (60:40, v/v). Compounds migrating 2 cm above and 0.5 cm below
the bis O-methyloxime, PFB ester derivative of
LGE2 were then converted to a trimethylsilyl
(TMS) ether derivative by treatment with 10 µl dimethylformamide
and 10 µl of N,O-bis(trimethylsilyl)trifluoroacetamide and
analyzed by gas chromatography (GC) negative ion chemical ionization
(NICI) mass spectrometry (MS). The
M-·CH2C6F5
(M-181) ions m/z 481 and 487 were monitored for IsoLGs and
the [2H6]
LGE2 internal standard, respectively. The results
obtained are shown in Fig. 3
. In the lower m/z 487 ion current chromatogram are seen the
incompletely resolved 4 syn and ante
O-methyloxime isomers of the internal standard. In the top
m/z 481 ion current chromatogram are seen a series of
unresolved peaks with a similar retention time as the internal
standard, consistent with the presence of IsoLGs.
View larger version (27K):
[in this window]
[in a new window]
|
Figure 3. Selected ion current chromatograms obtained from GC/NICI/MS analysis
for IsoLGs formed during oxidation of arachidonic acid in
vitro. The compounds were analyzed as a PFB ester,
O-methyloxime, TMS ether derivative. The
[M-·CH2C6F5]- ion
m/z 481was monitored for IsoLGs and the corresponding
ion at m/z 487 was monitored for the bis
[2H3] O-methyloxime derivative
of synthetic LGE2, which was added as an internal
standard.
|
|
Additional experiments were then carried out to further confirm that
the m/z 481 peaks detected during oxidation of arachidonic
acid represented IsoLGs. First, all of the m/z 481 peaks
disappeared and shifted upward 6 Da when analyzed as a
[2H3]
O-methyloxmine derivative, indicating the presence of two
carbonyl groups (not shown). Further, all of the m/z 481
peaks disappeared and shifted upward 9 Da when analyzed as a
[2H9] TMS ether
derivative, indicating the presence of a single hydroxyl group (not
shown). Finally, the compounds were then analyzed by GC/electron impact
ionization (EI)/MS after partial purification by high-performance
liquid chromatography (HPLC). The mixture of compounds was converted to
O-methyloxime derivatives and then subjected to HPLC using a
solvent system of 45% acetonitrile in water with 0.1% acetic acid run
at 1 ml/min, 1 ml fractions. Aliquots of fractions collected were
analyzed by GC/NICI/MS to detect where the putative IsoLGs eluted.
IsoLGs were detected eluting over ~30 fractions. Eight fractions that
eluted with a retention volume of ~30 ml contained a high
concentration of IsoLGs and were pooled and converted to a PFB ester,
TMS ether derivative for analysis by EI/MS.
A mass spectrum obtained from this analysis is shown in Fig. 4
B. Shown in Fig. 4A
is a mass spectrum of
synthetic LGE2. Notable are the striking
similarities in the high mass ions present and their relative abundance
in the two mass spectra. Intense high mass ions are present at
m/z 662 (M+); m/z 631
(M-31), loss of ·OCH3 from a methoxamine group;
m/z 591 (M-71), loss of
·CH2(CH2)3CH3
from the lower side chain; m/z 559 (M-71–32), loss of 71 +
HOCH3; m/z 541 (M-90–31), loss of
Me3SiOH + 31; m/z 501 (M-90–71);
m/z 489 (M-173), loss of
·CH2(OSiMe3)(CH2)4CH3
from the lower side chain. The origin of the ion at m/z 418
is unclear but is present in both mass spectra. Key fragment ions are
present in the mass spectrum of synthetic LGE2 at
m/z 392 and 270, arising from fragmentation between C8 and
C12, representing the upper and lower side chains, respectively. The
material obtained for this mass spectral analysis was only partially
purified. Less abundant ions are present at both m/z 270 and
m/z 392 in the IsoLG mass spectrum, which would be
consistent with the presence of a 15 series
IsoLGE2 compound as a minor component. The base
ion in the IsoLG mass spectrum is m/z 284. This ion would be
consistent with the lower portion of an IsoLGD2
compound. An ion representing the upper portion of the molecule, as is
seen in the mass spectrum of synthetic LGE2, is
not present in this mass spectrum. Possible reasons for the absence of
this ion include different relative abundances of these two ions in
different IsoLG isomers or in IsoLGD2 compared
with IsoLGE2 compounds. The ions resulting from
fragmentation adjacent to the lower chain TMS ether carbon indicate
that the IsoLGD2 compound is a 15-series
regioisomer.
View larger version (32K):
[in this window]
[in a new window]
|
Figure 4. El mass spectra obtained from the analysis of synthetic
LGE2 (A) and partially purified IsoLGs
(B) as a PFB, O-methyloxime, TMS ether
derivative. See text for details of the interpretation of specific
ions.
|
|
To explore the quantitative potential importance of the formation of
IsoLGs, we compared the amount of IsoLGs formed during oxidation of
arachidonic acid with that of F2-IsoPs and
D2/E2-IsoPs (Fig. 5
). The IsoPs were measured by GC/NICI/MS, as described, except that
integration of all peaks was used for quantification (4
, 17)
. Of interest was that the amount of IsoLGs formed was only
slightly less than the amount of
D2/E2-IsoPs formed and
slightly exceeded the amount of F2-IsoPs formed.
This suggests that IsoLGs can be formed in quantities that may have
biological relevance.
View larger version (30K):
[in this window]
[in a new window]
|
Figure 5. Comparison of the relative amounts of IsoLGs,
D2/E2-IsoPs, and F2-IsoPs formed
after oxidation of 5 mg of arachidonic acid in vitro for
4 h with iron/ADP/ascorbate.
|
|
We then sought to assess the formation of IsoLGs in biological systems
in vitro and in vivo in plasma and urine, during
oxidation of liver microsomes and low density lipoproteins (LD)L, and
in liver after administration of CCl4 to rats.
However, in none of these were we able to detect the formation of
IsoLGs, even though there was a dramatic increase in the formation of
IsoPs during oxidation of microsomes and low density lipoproteins (LDL)
and in CCl4-treated rats, as previously reported
(1
, 18
, 19)
. Given the abundant amounts of IsoLGs formed
during oxidation of arachidonic acid in vitro, this was
unexpected. However, there was one fundamental difference between the
conditions present in these experiments and those in which arachidonic
acid was oxidized in vitro, namely, the presence of
proteins. Therefore, we hypothesized that the failure to detect free
IsoLGs in the biological systems may be explained if, once formed, they
adduct to proteins with extraordinary rapidity.
To gain support for this hypothesis, we assessed the time course of
adduction of LGE2 to bovine serum albumin (BSA)
as a model protein and compared this with the rate of adduction of
4-hydroxynonenal (4-HNE). We thought that comparing the rates of
adduction of LGE2 with 4-HNE would be informative
since 4-HNE is considered one of the most reactive products of lipid
peroxidation that has been identified, yet free 4-HNE can be detected
in biological fluids and tissues containing protein (20)
.
In these experiments, 0.1 mM LGE2 and 0.1 mM
4-HNE were incubated in 5 ml of HBSS containing 20 mg/ml of BSA.
Adduction of LGE2 and 4-HNE was assessed by
monitoring the fall in free levels over time in aliquots removed from
the incubation. Free LGE2 was measured by
GC/NICI/MS as described previously and free 4-HNE was measured by
colorimetric assay (Oxis International, Portland, Oreg.). The
results obtained were revealing (Fig. 6
). Levels of free LGE2 fell precipitously during
the initial 60 s; more than 50% had adducted within the first
20 s. In striking contrast, ~50% of 4-HNE still remained
unadducted after 1 h. The rate of adduction of 4-HNE observed in
these experiments agrees closely with that reported previously
(21)
. This indicated that the rate of adduction of
LGE2 to proteins exceeds that of 4-HNE by several
orders of magnitude. It should be pointed out that a small amount of
free LGE2 remained detectable even at the later
time points. This can likely be explained by the presence of a small
amount of LGE2 in which the lower side chain
double bond had migrated from the 
13 position
to the 12 position. This double bond migration
decreases the reactivity of the compound and can be catalyzed by ions
in the incubation buffer (22)
, but analysis by GC/MS
cannot distinguish between these two species.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 6. Comparison of the relative rates of covalent adduction of
LGE2 and 4-HNE to BSA. The formation of adducts was
assessed by monitoring the disappearance of free compounds (closed
circles for 4-HNE and open circles for LGE2) over time and
expressed as the amount of free compound present prior to the addition
of BSA. The proposed explanation for why the levels of LGE2
do not fall to undetectable levels but plateau between 5 and 10% of
the amount of LGE2 present at time zero is discussed in the
text.
|
|
Given the remarkable rapidity with which LGs adduct to proteins, we
undertook studies to identify the nature of LG adducts by using liquid
chromatography (LC)/electrospray ionization (ESI)/MS. Salomon and
colleagues had shown that LGE2 formed a pyrrole
adduct with the -amine of lysine (23)
. Adduction to the
-amino group rather than the 
-amino group is highly favored
because the 
-amine is much more reactive (24)
and is
the most abundant amine available on proteins. After incubation of
approximately a 1:1 mixture of LGE2 with lysine
for up to 4 h at 37°C in phosphate-buffered saline, adducts were
loaded onto a C-18 SepPak cartridge, washed sequentially with 10 ml of
water, 10 ml of heptane, and then eluted with 10 ml of heptane:ethyl
acetate (1:1). Compounds were analyzed by LC/ESI/MS in the positive ion
mode using a 2.1 x 150 mm C18 column (Waters Associates, Milford,
Mass.) and a solvent system consisting of a gradient of
water:acetonitrile gradient (3%/min; hold 5 min) at 0.2 ml/min. Full
scanning analysis did not reveal the predicted
[MH]+ ion for the lysyl
LGE2 pyrrole (m/z 463), but did reveal
intense ions 16 and 32 Da higher at m/z 479 and
m/z 495. Shown in Fig. 7
are selected ion current chromatograms of m/z 479 and
m/z 495 obtained from an analysis of an incubation of
LGE2 with lysine. Of potential relevance to this
finding is that pyrroles have been shown to undergo autoxidation to
form lactams and hydroxylactams with molecular masses 16 and 32 Da
higher than the corresponding pyrrole (25)
. The mechanism
of autoxidation of pyrroles is depicted in Fig. 8
. More than a single m/z 479 and m/z 495 peak is
seen in Fig. 7
, consistent with the fact that autoxidation of the
pyrrole would not be stereoselective and that the C-15 hydroxyl group
in synthetic LGE2 is racemic. Consistent with a
facile autoxidation of the LGE2 pyrrole is that
Salomon and colleagues found the pyrrole to highly unstable unless it
was derivatized or oxygen was rigorously excluded (23)
. To
support the notion that the m/z 479 and m/z 495
peaks represented lactam and hydroxylactam adducts derived from
oxidation of the pyrrole, we carried out incubations under argon to
exclude air, and upon analysis did see an intense signal at the
predicted [MH]+ for the pyrrole at
m/z 463. Although this likely represented the pyrrole
adduct, collision-induced dissociation (CID) of the m/z 463
ion did not yield informative daughter fragment ions that would allow a
conclusive assignment of the structure of the molecule as the
LGE2 pyrrole.
View larger version (20K):
[in this window]
[in a new window]
|
Figure 7. Selected ion current chromatograms of m/z 479 and
m/z 495 obtained from LC/ESI/MS analysis of adducts
formed during an incubation of LGE2 with lysine.
|
|
View larger version (14K):
[in this window]
[in a new window]
|
Figure 8. Proposed mechanism for the formation of lactams and hydroxylactams
by autoxidation of pyrroles (adapted from ref 25
).
|
|
To further structurally characterize the putative lysyl
LGE2 lactam and hydroxylactam adducts, CID
analyses of the respective [MH]+ ions were
performed. The CID mass spectra for the lactam and hydroxylactam
adducts are shown in Fig. 9
A, B, respectively. Ions were present in the CID spectrum of
the lactam at m/z 461 and the hydroxylactam at
m/z 477, representing the loss of one molecule of
H2O, and at m/z 459 in the
hydroxylactam spectrum, representing the loss of two molecules of
H2O. Other prominent ions present and their
proposed structures are listed in Figs. 10
and
11. Although the precise molecular mechanism involved in the formation of
these ions remains speculative, the structures depicted were supported
by the ion shifts noted on analysis of the corresponding adducts formed
with the lysine analogs
[13C6] lysine and
N--acetyl-lysine methyl ester, as indicated in Figs. 10
and 11
.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 10. Proposed structures for the daughter ions formed by CID of the
[MH]+ ion, m/z 479, of the
LGE2-lysine lactam adduct. The proposed structures of these
ions were supported by the ion shifts noted on analysis of the lactam
adducts formed with the lysine analogs,
[13C6]-lysine and N--acetyl-lysine methyl
ester.
|
|
View larger version (31K):
[in this window]
[in a new window]
|
Figure 11. Proposed structures for the daughter ions formed by CID of the
[MH]+ ion, m/z 495, of the
LGE2 hydroxylactam adduct. The proposed structures of these
ions were supported by the ion shifts noted on analysis of the
hydroxylactam adducts formed with the lysine analogs,
[13C6]-lysine and N--acetyl-lysine methyl
ester.
|
|
During the course of these experiments, an additional observation
highlights further the remarkable reactivity of these compounds. After
incubation of LGE2 with a molar excess of
[3H] lysine and subsequent analysis by HPLC, it
was ascertained that the lactam and hydroxylactam adducts do not even
account for the majority of adducts formed by reaction of
LGE2 with lysine. In this regard, a broad slur of
unresolved radioactivity was seen that eluted from the HPLC over many
more fractions than the lactam and hydroxylactam adducts and was widely
separated from unreacted lysine. Although this material is not amenable
to analysis by the conventional approaches used in these studies, we
speculate that this material represents cross-linked species.
We then used this information to assess the formation of IsoLG lysine
adducts. Five milligrams of arachidonic acid was oxidized as described
previously, but in the presence of 5 mg of lysine, and then analyzed by
LC/ESI/MS/MS. The analysis revealed the presence of compounds with the
[MH]+ ions and characteristic LC elution
volumes for lysyl IsoLG lactam and hydroxylactam adducts. When the
[MH]+ ions of the IsoLG lactam and
hydroxylactam adducts were subjected to CID analysis (Fig. 12
), the same daughter ions were detected as seen in CID analysis of the
LGE2 lactam and hydroxylactam adducts. There is
one notable difference from the analysis seen after incubation of
LGE2 with lysine: the presence of additional
peaks. This is consistent with the generation of multiple
IsoLGE2 and IsoLGD2
isomers, which formed multiple lactam and hydroxylactam adducts upon
reaction with lysine.
One of the major interests in reactive aldehyde products of lipid
peroxidation stems from the fact that the formation of aldehydes (e.g.,
4-HNE and malondialdehyde) during oxidation of LDL and their subsequent
adduction to apolipoprotein B (Apo-B) are thought to be responsible for
the conversion of LDL to an atherogenic form that is taken up by
macrophages, resulting in the formation of foam cells (26
, 27)
. As might be expected in light of the results reported here,
Salomon and colleagues have demonstrated in vitro that LG
adducts to LDL and converts it to an atherogenic form with an
efficiency that greatly exceeds that of 4-HNE and malondialdehyde
(28)
. Therefore, we explored whether we could detect
IsoLGE2 lactam and hydroxylactam adducts on Apo-B
after oxidation of LDL in vitro. This was regarded as an
informative experiment because, as mentioned previously, we were unable
to detect the presence of free IsoLGs during oxidation of LDL. LDL was
isolated from 10 ml of plasma from normal volunteers using a low
temperature ethanol precipitation procedure (29)
and
oxidized with the azo initiator 2,2'-azobis(2-aminopropane) HCL (AAPH)
for 4 h. The LDL was reprecipitated and delipidated
(29)
. The recovered apolipoprotein B protein was then
treated with 0.2N NaOH for 2 h at room temperature to hydrolyze
any IsoLG protein adducts that may have formed with IsoLG esterified to
LDL lipids. The protein was then subjected to complete enzymatic
hydrolysis to individual amino acids by sequential treatment with
Pronase and leucine aminopeptidase (24)
. Native LDL was
treated in an identical fashion, but not subjected to oxidation. The
amino acid hydrolysate from oxidized LDL and native LDL was then
analyzed by LC/ESI/MS/MS. Before analysis,
[13C6] lysyl IsoLG lactam
and hydroxylactam adducts were added as internal standards. These were
formed by oxidation of arachidonic acid in the presence of
[3H]-,
[13C6] lysine, purified
by HPLC, and quantitated using the specific activity of the
[3H] lysine. Lysyl IsoLG lactam and
hydroxylactam adducts were not detected in native LDL. However, intense
signals indicative of these adducts on Apo-B were present in the
oxidized LDL preparation (Fig. 13
). Approximately 1 nmol of IsoLG adduct was formed per 500 nmol Apo-B.
View larger version (34K):
[in this window]
[in a new window]
|
Figure 13. LC/ESI/MS/MS analysis of IsoLG lysyl adducts on apolipoprotein B after
oxidation of LDL. After copper-induced oxidation of LDL, the
apolipoprotein B protein was enzymatically digested to individual amino
acids and the hydrolysate was analyzed for lysyl-IsoLG lactam and
hydroxylacatm adducts. Internal standards used were
[13C6]-lactam and -hydroxylactam adducts,
which were formed by oxidation of arachidonic acid in the presence of
[13C6]-lysine. The hydrolysate was analyzed
by selected reaction monitoring of the following transitions:
m/z 495.4 to m/z 84.1(hydroxylactam
adducts from oxidized LDL); m/z 501.4 to
m/z 89.1 (internal standard
[13C6] hydroxylactam adducts);
m/z 479.4 to m/z 84.1(lactam adducts from
oxidized LDL); m/z 485.4 to m/z 84.1
(internal standard [13C6] lactam adducts).
|
|
|
UNANSWERED QUESTIONS/DIRECTIONS FOR FUTURE RESEARCH
|
|---|
These studies were performed to determine whether IsoLGs are
formed as products of the IsoP pathway and to characterize the nature
of the adducts formed with proteins. With this goal in mind, we have
demonstrated that IsoLGs are formed as products of the IsoP pathway
in vitro and elucidated that IsoLGs form oxidized pyrrole
adducts in the form of lactams and hydroxylactams. These findings are
an essential first step that provides the key information necessary to
begin to explore many of the unknowns related to the formation of
IsoLGs in vivo and the biological consequences that ensue.
What is intriguing and fascinating about these molecules is their
remarkable reactivity. As demonstrated, LGs adduct to proteins at a
rate that exceeds that of 4-HNE by several orders of magnitude. This
was a dramatic finding since 4-HNE had been considered to be one of the
most reactive products of lipid peroxidation that has been identified
(20)
. 4-HNE has been thought to be responsible, at least
in part, for many of the deleterious effects of oxidant injury
(20)
. However, the discovery that IsoLGs are formed as
products of lipid peroxidation via the IsoP pathway opens numerous
avenues for new scientific inquiry related to biological import of the
formation of these reactive molecules in settings of oxidant injury.
The next goal is to demonstrate that IsoLGs are formed in
vivo in settings of oxidant injury and determine the amounts
produced. Results from preliminary experiments recently performed have
suggested that large quantities of IsoLG adducts are formed in the
liver of rats treated with CCl4 to induce a
severe hepatocellular injury. There are interesting questions to
consider regarding the formation of IsoLGs related to their remarkable
reactivity that may have relevance to the potential biological
alterations they may cause. As mentioned, IsoPs are initially formed
in situ on phospholipids and then released in free form,
presumably by phospholipases (2)
. As might be expected, we
have found that oxidation of 1-palmitoyl,
2-arachidonoyl-glycero-3-phosphocholine results in the formation of
IsoLGs esterified to the phospholipid. However, nothing is known as to
whether esterified IsoLGs are substrates for mammalian phospholipases.
If they are not released in free form from membrane phospholipids, the
biological consequences of their formation may primarily involve
functional and biophysical alterations of cellular membranes due to
adduction to membrane proteins while esterified to phospholipids.
Modification of key membrane proteins by IsoLG adduction, such as
receptor proteins, ion channel proteins, etc., may lead to deleterious
effects on cellular function. However, because of their remarkable
reactivity, even if hydrolyzed from phospholipids, diffusion from the
bilayer may be limited by rapid adduction to membrane proteins. Thus,
in addition to assessing the formation of IsoLGs in vivo,
the cellular localization of adduct formation may provide a basis for
hypotheses as to what biological consequences may ensue as a result of
their formation.
Salomon and colleagues have shown that LGs can form protein–DNA
cross-links when incubated with cultured cells (15)
.
Therefore, it is also important to assess whether IsoLG DNA adducts are
formed during oxidant injury. If they are, this opens up another
important area for future investigation. At present, virtually nothing
is known regarding the consequences of the formation of LG-DNA adducts.
1) Do they cause mutations, and if so, what types of
mutations do they induce? 2) Are IsoLG DNA adducts repaired,
and if so, how rapidly does this occur?
Another area of inquiry relates to our recent finding that IsoP-like
compounds are formed in vivo from free radical-induced
oxidation of docosohexaenoic acid (DHA) (C22,n6,
3)
(30)
. Our interest in the formation of IsoP-like compounds
from oxidized DHA stems from the fact that it is uniquely enriched in
the brain, comprising 25–35% of total fatty acids in
aminophospholipids (31
, 32)
. Thus, these compounds may be
a unique marker of oxidative neuronal injury. As a result, we have
termed these compounds `neuroprostanes' (NPs). As with IsoPs,
intermediates in the pathway of formation of NPs are bicyclic
endoperoxides. The endoperoxides are reduced to form
F4-NPs, but may also undergo rearrangement to
form IsoLG-like compounds. Thus, if formed, IsoLG-like compounds formed
via the NP pathway may be important mediators of oxidative neuronal
injury. Pertinent to this possibility are data obtained from effects of
the hexane metabolite 2,5-hexanedione on neurons. 2,5-Hexanedione, like
LGs, is a -dicarbonyl that shares similar reaction chemistry with
amines. 2,5-Hexanedione has been shown to cause axonal degeneration
through formation of pyrrole adducts and subsequent cross-linking of
neuronal filaments (33)
. Potentially very relevant in this
regard is our recent discovery that levels of
F4-NPs and F2-IsoPs are
significantly increased in cerebrospinal fluid of patients with
Alzheimer's and Huntington's disease (34
, 35)
; free
radicals are thought to play an important role in the neuronal damage
that characterizes these disorders (36
, 37)
. Of interest
is that Alzheimer's disease is also characterized by neurofibrillary
tangles and amyloid plaques that involve protein cross-links. The cause
of the cross-linking has not been clearly established, but is thought
to involve oxidation (38
39
40)
. It would be of great
interest to explore whether there is evidence that IsoLGs or IsoLG-like
compounds formed from DHA participate in the protein cross-linking in
neurofibrillary tangles and amyloid plaques in Alzheimer's disease.
In summary, we describe the formation of extremely reactive
-ketoaldehydes (IsoLGs) as products of the IsoP pathway and have
characterized the nature of the adducts formed with protein.
Previously, LGs were considered to be formed only as by-products of the
cyclooxygenase pathway, and there was little, if any, evidence that
they were formed in vivo. Our new discovery that LG-like
compounds are also formed as products of the IsoP pathway broadens
considerably the potential importance of these molecules in
pathobiology, specifically oxidant injury. Previous work related to LGs
and the work reported here have focused on understanding the chemistry
of these compounds and the reactions they undergo. Essential
information now obtained is the characterization of the nature of the
IsoLG adducts formed with proteins and the development of specific
analytical tools for their detection and quantification. This will
allow us to begin to explore their formation under various conditions
in vivo and the cellular location of the adducts formed.
This information will then form a basis for hypotheses to set into
motion studies into the relatively uncharted area related to the
biological effects exerted by these interesting molecules.
|
ACKNOWLEDGMENTS
|
|---|
This work was supported in part by National Institutes of Health
grants GM 42056, DK 48831, GM 15431, CA 77839, DK 26657, Ca 68485, and
a PhRMA Foundation Fellowship to C.J.B. The authors wish to thank Dr.
Kamaljit Kaur for her expert assistance in the synthesis of
LGE2.
|
FOOTNOTES
|
|---|
2 Abbreviations: IsoPs, isoprostanes; TMS,
trimethylsilyl; PFB, pentafluorobenzyl; MS, mass spectrometry; EI,
electron impact ionization; NICI, negative ion chemical ionization;
CID, collision-induced dissociation; ESI, electrospray ionization; LC,
liquid chromatography; GC, gas chromatography; BSA, bovine serum
albumin; Apo-B, apolipoprotein B; 4-HNE, 4-hydroxynonenal; DHA,
docosahexaenoic acid; HPLC, high-performance liquid chromatography;
LDL, low density lipoproteins; LG, levuglandins; NPs, neuroprostanes;
PG, prostaglandin. 
|
REFERENCES
|
|---|
-
Morrow, J. D., Hill, K. E., Burk, R. F., Nammour, T. M., Badr, K. F., Roberts, L. J., II (1990) A series of prostaglandin F2-like compounds are produced in vivo by a noncyclooxygenase free radical catalyzed mechanism. Proc. Natl. Acad. Sci. USA 87,9383-9387[Abstract/Free Full Text]
-
Morrow, J. D., Awad, J. A., Boss, H. J., Blair, I. A., Roberts, L. J., II (1992) Non-cyclooxygenase derived prostanoids (F2-isoprostanes) are formed in situ on phospholipids. Proc. Natl. Acad. Sci. USA 89,10721-10725[Abstract/Free Full Text]
-
Morrow, J. D., Roberts, L. J., II, Daniel, V. C., Mirotchnechenko, O., Swift, L., Burk, R. F. (1998) Arch. Biochem. Biophys. 353,160-171
-
Morrow, J. D., Minton, T. A., Mukundan, C. R., Campbell, M. D., Zackert, W. E., Daniel, V. C., Badr, K. F., Blair, I. A., Roberts, L. J., II (1994) Free radical induced generation of isoprostanes in vivo: Evidence for the formation of D-ring and E-ring isoprostanes. J. Biol. Chem. 269,4317-4326[Abstract/Free Full Text]
-
Morrow, J. D., Awad, J. A., Zackert, W. E., Wu, A., Daniel, V. C., Roberts, L. J., II (1996) Free radical induced generation of thromboxane-like compounds (isothromboxanes) in vivo. J. Biol. Chem. 271,23185-23190[Abstract/Free Full Text]
-
Roberts, L. J., II, Morrow, J. D. (1997) The generation and actions of isoprostanes. Biochim. Biophys. Acta 1345,121-135[Medline]
-
Morrow, J. D., Roberts, L. J., II (1997) The isoprostanes: unique bioactive products of lipid peroxidation. Prog. Lipid Res. 36,1-21[Medline]
-
Moore, K, Roberts, L. J., II (1998) Measurement of lipid peroxidation. Free Rad. Res. 28,659-671[Medline]
-
Roberts, L. J., II, Morrow, J. D. (1999) Isoprostanes as markers of lipid peroxidation in atherosclerosis. Sehan, C. N. Ward, P. A. eds. Molecular Biology of Inflammation ,141-163 Humana Press Totowa, N.J..
-
Nutgeren, D. H., Hazelhof, E. (1973) Isolation and properties of intermediates in prostaglandin biosynthesis. Biochim. Biophys. Acta 326,448-461[Medline]
-
Salomon, R. G., Miller, D. B., Zagorski, M. G., Coughlin, D. J. (1984) Solvent induced fragmentation of prostaglandin endoperoxides. New aldehyde products from PGH2 and a novel intramolecular 1,2-hydride shift during endoperoxide fragmentation in aqueous solution. J. Am. Chem. Soc. 106,6049-6060
-
Salomon, R. G., Jirousek, M. B., Ghosh, S., Sharma, R. B. (1987) Prostaglandin endoperoxides 21. Covalent binding of levuglandin E2 with proteins. Prostaglandins 34,643-656[Medline]
-
Iyer, R. S., Ghosh, S., Salomon, R. G. (1989) Levuglandin E2 crosslinks proteins. Prostaglandins 37,471-480[Medline]
-
Jirousek, M. R., Murthi, K. K., Salomon, R. G. (1990) Electrophilic levuglandin E2 adducts bind glycine: a model for protein crosslinking. Prostaglandins 40,187-203[Medline]
-
Murthi, K. K., Friedman, L. R., Oleinick, N. L., Salomon, R. G. (1993) Formation of DNA-protein cross-links in mammalian cells by levuglandin E2. Biochemistry 32,4090-4097[Medline]
-
Taber, D. F., Morrow, J. D., Roberts, L. J., II (1997) A nomenclature system for the isoprostanes. Prostaglandins 53,63-67[Medline]
-
Morrow, J. D., Roberts, L., J, II (1998) Mass spectrometric quantification of F2-isoprostanes in biological fluids and tissues as a measure of oxidant stress. Methods Enzymol 300,3-12
-
Longmire, A. W., Swift, L. L., Roberts, L. J., II, Awad, J. A., Burk, R. F., Morrow, J. D. (1994) Effect of oxygen tension on the generation of F2-isoprostanes and malondialdehyde in peroxidizing rat liver microsomes. Biochem. Pharmacol. 47,1173-1177[Medline]
-
Morrow, J. D., Awad, J. A., Kato, T., Takahashi, K., Badr, K. F., Roberts, L. J., II, Burk, R. F. (1992) Formation of novel non-cyclooxygenase derived prostanoids (F2-isoprostanes) in carbon tetrachloride hepatotoxicity, an animal model of lipid peroxidation. J. Clin. Invest. 90,2502-2507
-
Esterbauer, H., Schaur, R. J., Zollner, H. (1991) Chemistry and biochemistry of 4-hydroxynonenal, malondialdehyde, and related aldehydes. Free Rad. Biol. Med. 11,81-128[Medline]
-
Curzio, M, Esterbauer, H., DiMauro, C., Cecchini, G., Dianzani, M. U. (1986) Chemotactic activity of the lipid peroxidation product 4-hydroxynonenal and homologous hydroxyalkenals. Biol. Chem. Hoppe-Seyler 367,321-329[Medline]
-
Iyer, R. S., Miller, D. B., Salomon, R. G. (1990) Decomposition of levuglandin E2. Dehydration and allylic rearrangement products. J. Org. Chem. 55,3175-3180
-
Iyer, R. S., Kobierski, M. E., Salomon, R. G. (1994) Generation of pyrroles in the reaction of levuglandin E2 with proteins. J. Org. Chem. 59,6038-6043
-
DeCaprio, A. P., Olajos, E. J., Weber, P. (1982) Covalent binding of a neurotoxic n-hexane metabolite: conversion of primary amines to substituted pyrrole adducts by 2,5-hexanedione. Tox. Appl. Pharm. 65,440-450[Medline]
-
Smith, E. B., Jensen, H. B. (1967) Autoxidation of three 1-alkylpyrroles. J. Org. Chem. 32,3330-3333
-
Steinberg, D. (1995) Role of oxidized LDL and antioxidants in atherosclerosis. Adv. Exp. Med. Biol. 369,39-48[Medline]
-
Requena, J. R., Fu, M. X., Ahmed, M. U., Jenkins, A. J., Lyons, T. J., Baynes, J. W., Thorpe, S. R. (1997) Quantification of malondialdehyde and 4-hydroxynonenal adducts to lysine residues in native and oxidized human low density lipoprotein. Biochem. J. 322,317-325
-
Hoppe, G., Subbanagounder, G., O'Neil, J., Salomon, R. G., Hoff, H. F. (1997) Macrophage recognition of LDL modified by levuglandin E2, an oxidation product of arachidonic acid. Biochim. Biophys. Acta 1344,1-5[Medline]
-
Mills, G. L., lane, P. A., Weech, P. K. (1984) A Guidebook to Lipoprotein Technique Elsevier Science New York.
-
Roberts, L. J., II, Montine, T. J., Markesbery, W. R., Tapper, A. R., Hardy, P., Chemtob, C., Dettbarn, W. D., Morrow, J. D. (1998) Formation of isoprostane-like compounds (Neuroprostanes) in vivo from docosahexaenoic acid. J. Biol. Chem. 273,13605-13612[Abstract/Free Full Text]
-
Skinner, E. R., Watt, C., Besson, J. A. O., Best, P. V. (1993) Differences in fatty acid composition of the grey and white matter of different regions of the brains of patients with Alzheimer's disease and control subjects. Brain 116,717-725[Abstract/Free Full Text]
-
Salem, N., Jr, Kim, H.-Y., Yergery, J. A. (1986) Docosahexaenoic acid: membrane function and metabolism. Simopoulos, A. P. Kifer, R. R. Martin, R. E. eds. Health Effects of Polyunsaturated Fatty Acids in Seafoods ,263-317 Academic Press Orlando, Florida.
-
Graham, D. G., Amarnath, V., Valentine, W. M., Pyle, S. J., Anthony, D. G. (1995) Pathogenetic studies of hexane and carbon disulfide neurotoxicity. Crit. Rev. Toxicol. 25,91-112[Medline]
-
Montine, T. J., Markesbery, W. R., Morrow, J. D., Roberts, L. J., II (1998) Cerebrospinal fluid F2-isoprostane levels are increased in patients with Alzheimer's disease. Ann. Neurol. 44,410-413[Medline]
-
Montine, T. J., Beal, M. F., Roberts, D., Cudkowicz,
M. E., Brown, R. H., O'Donnell, H., Zackert, W. E.,
Roberts, L. J., II, and Morrow, J. D. (1999) Cerebrospinal
fluid levels of F2-isoprostanes, specific markers
of lipid peroxidation, are elevated in Huntington's disease patients.
Neurology In press
-
Browne, S. E., Bowling, A. C., MacGarvey, U., Baik, M. J., Berger, S. C., Muqit, M. M. K., Bird, E. D., Beal, M. F. (1997) Oxidative damage and metabolic dysfunction in Huntington's disease: selective vulnerability of the basal ganglia. Ann. Neurol. 41,646-653[Medline]
-
Markesbery, W. R. (1997) Oxidative stress hypothesis in Alzheimer's disease. Free Rad. Biol. Med. 23,134-147[Medline]
-
Montine, T. J., Amarnath, V., Martin, M. E., Stritmatter, W. J., Graham, D. G. (1996) E4-Hydroxy-2-nonenal is cytotoxic and cross-links cytoskeletal proteins in P19 neuroglial cultures. Am. J. Pathol. 148,89-93[Abstract]
-
Troncoso, J. C., Costello, A., Watson, A. L., Jr, Johnson, G. V. W. (1993) In vitro polymerization of oxidized tau into filaments. Brain Res 613,313-316[Medline]
-
Dyrks, T., Dyrks, E., Hartmann, T., Masters, C., Beyreuther, K. (1992) Amyloidgenicity of ßA4 and ßA4-bearing amyloid protein precursor fragments by metal-catalyzed oxidation. J. Biol. Chem. 267,18210-18217[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
E. POLIAKOV, M.-L. BRENNAN, J. MACPHERSON, R. ZHANG, W. SHA, L. NARINE, R. G. SALOMON, and S. L. HAZEN
Isolevuglandins, a novel class of isoprostenoid derivatives, function as integrated sensors of oxidant stress and are generated by myeloperoxidase in vivo
FASEB J,
December 1, 2003;
17(15):
2209 - 2220.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
D. H. Wright, D. Abran, M. Bhattacharya, X. Hou, S. G. Bernier, A. Bouayad, J.-C. Fouron, A. Vazquez-Tello, M. H. Beauchamp, R. I. Clyman, et al.
Prostanoid receptors: ontogeny and implications in vascular physiology
Am J Physiol Regulatory Integrative Comp Physiol,
November 1, 2001;
281(5):
R1343 - R1360.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
BACKGROUND
