Protein Carbonylation and Metal-Catalyzed Protein Oxidation in A Cellular Perspective

J O U R NA L OF PR O TE O MI CS 74 ( 20 1 1 ) 2 2 2 82 2 4 2
available at www.sciencedirect.com
www.elsevier.com/locate/jprot
Review
Protein carbonylation and metal-catalyzed protein oxidation in

a cellular perspective
Ian M. Mller a,, Adelina Rogowska-Wrzesinskab , R.S.P. Raoc
a
Department of Genetics and Biotechnology, Aarhus University, Forsgsvej 1, DK-4200 Slagelse, Denmark
Department of Biochemistry and Molecular Biology, University of Southern Danmark, Campusvej 55, DK-5230 Odense M, Denmark
c
CH20, 3rd cross, 7th main, Saraswathipuram, Mysore 570009, India
b
AR TIC LE I N FO
ABS TR ACT
Article history:
Proteins can become oxidatively modified in many different ways, either by direct oxidation
Received 16 February 2011
of amino acid side chains and protein backbone or indirectly by conjugation with oxidation
Accepted 3 May 2011
products of polyunsaturated fatty acids and carbohydrates. While reversible oxidative
Available online 11 May 2011
modifications are thought to be relevant in physiological processes, irreversible oxidative

modifications are known to contribute to cellular damage and disease. The most well-
Keywords:
Carbonylation
Frataxin
Metal binding
Metal-catalyzed oxidation
Mitochondria
Reactive oxygen species
studied irreversible protein oxidation is carbonylation. In this work we first examine how
protein carbonylation occurs via metal-catalyzed oxidation (MCO) in vivo and in vitro with
an emphasis on cellular metal ion homeostasis and metal binding. We then review
proteomic methods currently used for identifying carbonylated proteins and their sites of
modification. Finally, we discuss the identified carbonylated proteins and the pattern of
carbonylation sites in relation to cellular metabolism using the mitochondrion as a case
story.
2011 Elsevier B.V. All rights reserved.
Contents
1.
2.
Introduction . . . . . . . . . . . . . . . . . . . . . . . .
Protein carbonylation and metal-catalyzed oxidation in
2.1. What causes protein carbonylation? . . . . . . .
2.2. Metal ions in vivo . . . . . . . . . . . . . . . . .
2.3. Metal binding in vivo . . . . . . . . . . . . . . .
2.4. MCO in vitro . . . . . . . . . . . . . . . . . . . .
2.5. MCO in vivo . . . . . . . . . . . . . . . . . . . .
. . .
vivo
. . .
. . .
. . .
. . .
. . .
. . . .
and in
. . . .
. . . .
. . . .
. . . .
. . . .
. . .
vitro
. . .
. . .
. . .
. . .
. . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2229
2229
2229
2230
2230
2231
2231
Abbreviations: 4-HNE, 4-hydroxynonenal; AGE, advanced glycation end products; ALE, advanced lipoxidation end products; CML, N(6)carboxymethyllysine; DNP, 2,4-dinitrophenyl; DNPH, 2,4-dinitrophenylhydrazine; HICAT, hydrazide-functionalized isotope-coded affinity
tag; MCO, metal-catalyzed oxidation; MRM, multiple reaction monitoring; O-ECAT, oxidation-dependent carbonyl-specific element-coded
affinity mass tag; PIC, phenyl isocyanate; PUFA, polyunsaturated fatty acid; SCX, strong cation exchange; SPH, solid phase hydrazide; SRM,
single reaction monitoring.
Corresponding author. Tel.: +45 8999 3633, +45 2087 2100 (mobile).
E-mail address: ian.max.moller@agrsci.dk (I.M. Mller).
1874-3919/$ see front matter 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.jprot.2011.05.004
2229
J O U R NA L OF PR O TE O MI CS 7 4 ( 2 01 1 ) 2 2 2 82 2 4 2
3.
Proteomic methods for the detection and quantification of protein carbonylation . . .

3.1. 2D gel based proteomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2. Identification of carbonylation sites in carbonylated proteins . . . . . . . . . .
3.3. Affinity enrichment based high throughput mass spectrometry . . . . . . . . .
3.4. Mass spectrometry-based quantitation of carbonylated residues. . . . . . . . .
4.
Carbonylated sites identified MCO specificity. . . . . . . . . . . . . . . . . . . . . .
5.
Biological implications of protein carbonylation the mitochondrion as a case story.
5.1. Carbonylated mitochondrial proteins . . . . . . . . . . . . . . . . . . . . . . . .
5.2. Carbonylation via conjugation with oxidized carbohydrates and fatty acids . .
5.3. Protein turnover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.4. Intracellular signalling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.
Introduction
The production of Reactive Oxygen Species (ROS) is an

unavoidable consequence of aerobic metabolism. In mammalian cells, the majority of ROS production is localised to
mitochondria. Also in non-photosynthesizing plant cells
mitochondria are the main source of ROS. In contrast, in
photosynthesizing plant cells, it is the chloroplasts and the
peroxisomes that are by far the dominant ROS sources [1]. ROS
can oxidize DNA, carbohydrates, unsaturated fatty acids and
proteins. Proteins can be oxidized in many different ways.
While some of these oxidations are reversible, e.g., disulphide
formation, which plays a role in metabolic regulation, others
are not and can lead to inactivation of the modified protein.
The cell therefore has developed ways to minimize ROS
production, remove ROS once formed and repair at least
some of the damage [2,3]. At the same time ROS can act as
messenger. Under stress such as disease, pathogen attack,
nutrient deficiency, and toxicity, ROS production and concomitant damage generally increases. When the cellular defence
mechanisms are overwhelmed, the amount of damage including that of protein oxidation accumulates and this can lead to
cell death. Understanding protein oxidation is therefore an
important part of understanding cellular stress response.
Protein carbonylation is the most common and best
studied irreversible protein oxidation and we will here first
describe what causes protein carbonylation. We will then
briefly describe proteomic methods for identifying carbonylated proteins and their sites of modification. In the final
sections, we will review the consequences of protein carbonylation for the function of the cells and organisms. Throughout this review, case stories will be selected from amongst
mitochondrial proteins.
2.
Protein carbonylation and metal-catalyzed
oxidation in vivo and in vitro
2.1.
What causes protein carbonylation?
Carbonyl groups (reactive aldehydes and ketones) can be

introduced into proteins by oxidation of amino acid side
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2231
2232
2232
2235
2236
2236
2237
2237
2238
2238
2239
2239
2239
2239
chains or they can be generated through oxidative cleavage of

proteins by -amidation pathway or by oxidation of glutamyl
side chains, leading to formation of peptides with an -keto
derivative at the N-terminal. Protein carbonylation can also be
generated by conjugation with aldehydes produced during
peroxidation of polyunsaturated fatty acids (PUFA) (so-called
Advanced Peroxidation End Products, ALE) and with reactive
carbonyl derivatives generated by reaction with reducing
carbohydrates (so-called Advanced Glycation End Products,
AGE) (e.g., [4]).
Reactive oxygen species (ROS) that lead to protein oxidation can be generated via a number of physiological and nonphysiological processes, primarily as by-products of normal
mitochondrial metabolism. For experimental purposes different methods can be employed in the generation of ROS, for
example X-ray or UV irradiation or chemicals. Fortunately,
living organisms are rarely exposed to that type of damage.
The most common mechanism of protein carbonylation in
living cells appears to be metal-catalyzed oxidation (MCO).
MCO typically occurs when reduced metal ions like Fe2+ or Cu+
interact with H2O2 in the so-called Fenton reaction and
produces the extremely reactive hydroxyl radicals [5]:
Fe2 H2 O2 Fe3 HO HO
1 Fenton reaction
The hydroxyl radical oxidizes amino acid side chains or

causes protein backbone cleavage both resulting in the
formation of carbonyl groups. It has been argued that in
bacteria MCO is the only source of protein carbonylation [6].
It is becoming evident that there exists a close interplay
between the different types of protein carbonylation and MCO,
but the physiological mechanisms controlling these processes
are not yet completely understood. It has been shown that
MCO and free radicals play a major role in the formation of
AGEs and AGE-induced protein cross-linking [7,8]. And MCO of
PUFA in the presence of proteins can also lead to formation of
N(6)-carboxymethyllysine (CML) by reaction of the PUFA
breakdown product, glyoxal, with the lysine side-chain. This
suggests that lipid oxidation plays a role in AGE formation as
well [9]. Dihydroxyacetone phosphate, a glycolytic intermediate, spontaneously decomposes to methylglyoxal, which can
also react with protein amino acid to give AGE products [10].
2230
This is an example of protein carbonylation formed in the

absence of MCO.
2.2.
Metal ions in vivo
To get an idea of the importance and mechanism of MCO, let

us first briefly consider the amounts of metal ions present in
living organisms. We will here concentrate on two of the metal
ions catalysing the Fenton reaction, Fe and Cu (Table 1).
The average total concentration of Fe and Cu ions in
healthy living cells is generally in the micromolar range
although specialized cells like the erythrocytes with their
massive haemoglobin content may contain as much as 20 mM
Fe. In contrast, the concentration of free ions is orders of
magnitude lower. In fact, with an intracellular concentration
of 10 18 M, a yeast cell contains not a single free Cu ion! [11].
The concentration of total and free or labile Fe ions appears
to be highly variable between different cell types (Table 1). At
least for the concentration of labile Fe ions, this is a
methodically defined quantity (see [12] for a discussion), but
the concentration appears to increase in diseased humans
[13].
In mitochondria about 75% of the Cu and Fe ions are
associated with the membrane fraction (containing mainly
the inner membrane), which is consistent with the fact that
the electron transport chain contains a number of abundant
metal-binding proteins [14].
2.3.
Metal binding in vivo
The use of ferric and ferrous ions in living cells developed in

the absence of molecular oxygen in the atmosphere. With the
advent of oxygen-evolving photosynthesis the concentration
of oxygen in the atmosphere rose and it became necessary to
prevent the damaging Fenton reaction (e.g., [15]). Eukaryotic
cells contain hundreds of metalloproteins that require iron
and/or copper ions to carry out their function many of which
are found inside organelles such as the mitochondria [16].
These metalloproteins include metal ion transporters, proteins stabilized by metal binding and enzymes that use the
metal ions in the catalytic site. From the moment the ion is
recognized outside the cell it is bound by dedicated metalbinding proteins and handed down a bucket-brigade of other
such proteins and transporters until it reaches its final

destination. In this way the cell can regulate the metal ion
supply while at the same time keeping the concentration of
free metal ions extremely low. This will minimize metal ion
binding to undesirable proteins and limit the concomitant risk
of oxidative damage due to the interaction with O2 and H2O2
[17].
Cu and Fe ions are bound to proteins by the side chains of
46 amino acid residues, mainly Cys, Met, His, Glu, Asp or Tyr
[18]. The total relative abundance of these amino acids in
proteins from eukaryotes is 21% (Uniprot database), which
means that there is a very large number of potential metal
binding sites in a cell, most of which are rarely or never
occupied. The ligands are often found on neighbouring
helices, which means that it is virtually impossible to predict
metal binding to proteins especially where the secondary and
tertiary structure is unknown [16].
To obtain the right picture we need to compare the
total concentration of Fe and Cu ions, which is normally in
the micromolar range (Table 1), to the total protein concentration in a selected cellular compartment. For instance, in
the mitochondrial matrix the protein concentration is
>500 mg/ml [19]. Assuming that the proteins are 50 kDa on
the average, this is equivalent to >10 mM of proteins. Each of
these protein molecules may contain one to several potential
metal binding sites, most of which are quite unspecific. Thus,
the number of potential metal binding sites on proteins
exceeds the total number of metal ions by several orders of
magnitude.
Not all bound Fe and Cu ions can catalyze the Fenton
reaction. If all the coordination sites of the metal ion are
occupied by ligands then binding of oxygen or ROS is not
possible. This is what happens when EDTA binds one of these
ions, which is used to terminate in vitro MCO (see below). In
fact, the inclusion of EDTA in the homogenization medium
when making extracts of biological tissues is recommended to
prevent Fenton reactions to take place during sample handling. On the other hand, oxygen or ROS binding to a proteinbound metal ion is not a guarantee of the Fenton reaction. If
the metal ion is strongly bound with slow on-off rates the
oxygen intermediate will be stable and will not give rise to the
hydroxyl radical. This type of mechanism is observed in
enzymes like catalase, superoxide dismutase (SOD) and
Table 1 Concentration of metal ions in living cells.

Cell/tissue
Total iron
Free iron
Mammalian cells
Rat plasma
Rat hepatocytes
Human cell lines
Human plasma
0.00110 M
0.253 M
1000 M
Yeast cells
Yeast cells
Arabidopsis mitochondria c
4060 M
3200 M
9.8 M
0.21.5 M
Undetectable a
0.55.8 M b
a
b
c
Total copper
Free copper
68 M
0.262.9 M
[17]
[112]
[113]
[114]
[13]
170 M
5 M
750 M
<10 18 M
[11]
P. Pedas pers. comm.
[14]
Healthy people.
Patients with renal disease.
Concentrations calculated assuming 1 l matrix volume per mg mitochondrial protein.
Reference
cytochrome c oxidase, which all interact with either H2O2 or O2

(e.g., [20]).
Fe ions are sequestered in several cellular compartments
by ferritins, a class of proteins that store the metal ions in a
safe and easily accessible form [21,22]. In mammalian cells
ferritins are found in the cytosol, nucleus and mitochondria,
while in plant cells they are also found in plastids. Knock-out
mutants of ferritins typically give rise to cells that are
oxidatively stressed although the presence of several ferritin
isoforms can mask that effect [23].
Frataxin is another protein contributing to Fe homeostasis
specifically in the mitochondria where it is involved in the
biosynthesis of FeS clusters and cytochromes [2427]. Frataxin
deficiency causes severe illness (Friedreich ataxia) or embryo
lethality in both humans and plants [25,28] and, in knock-out
mutants, the absence of frataxin induces severe oxidative
stress in mitochondria [29,30]. In contrast, overexpression of
frataxin resulted in yeast cells that were more resistant to
oxidizing agents and contained fewer carbonylated proteins
[24].
2.4.
MCO in vitro
MCO has often been used in model studies as a convenient

way to generate protein carbonylation. The pure protein or
biological protein mixture is incubated at 2030 C for few
minutes and up to several hours with Fe3+ and ascorbate or
Cu2+ and ascorbate and the reaction is stopped by chelating
the metal ions with e.g. EDTA followed by dialysis. Typically,
1001000 M metal salts have been used, but their low
solubility at neutral pH would have kept the concentration
of free metal ions at a much lower level but probably orders of
magnitude above that observed in biological systems. This
would ensure binding to sites, which might not normally bind
metal ions. In some cases the metal ion concentration used
exceeded the protein concentration many-fold, so each
protein molecule would have bound several or many metal
ions mostly at lower-affinity sites. The protein modifications
(carbonylation and other oxidative modifications) subsequently identified have no doubt been very useful for
developing methods for studying protein oxidation, but the
sites and the type of modification found may not reflect the in
vivo response [6,3135].
We are only aware of one study in which in vitro MCO was
conducted at a series of metal ion concentrations ranging in
stoichiometry from 1 Fe per 1500 protein molecule (BSA) to 1 Fe
per 1.5 protein molecule [6]. This type of study potentially
yields much more useful information about the MCO mechanism (see Section 4). However, none of the above studies
have linked carbonylation with function, so we still know
relatively little about the effect of the specific modifications
observed on the structure and function of the affected
proteins.
2.5.
MCO in vivo
As we have seen, the concentration of free metal ions in the

living cell is extremely low and the proteins involved in the
uptake, transport and processing of metal ions bind them in
such a way that they are unable to catalyze the Fenton
2231
reaction. This is also true for most of the proteins using metal
ions for structure and/or function. A special case is provided
by the enzymes involved in ROS or oxygen metabolism. These
enzymes must be able to bind molecules like H2O2 without
catalyzing the Fenton reaction.
In some cases, the Fenton reaction is thought to be part of
the desired mechanism. The transcription factor PerR in
Bacillus subtilis contains two His residues coordinating bound
Fe2+. Upon exposure to low levels (<10 M) of H2O2 one or both
of the His are oxidized presumably by HO generated by a
Fenton reaction involving the bound Fe 2+. The oxidized
transcription factor derepresses the PerP regulon, encoding
enzymes acting to detoxify peroxides. The His oxidation leads
to degradation of the protein. Thus, the transcription factor
acts as an H2O2 sensor and it is sacrificed to help minimize
damage from Fenton reactions elsewhere in the cell [36].
Free metal ions are occasionally released by mistake
when metal-containing proteins are damaged or oxidized. A
case in point is the mitochondrial aconitase containing an FeS
centre, which is very sensitive towards superoxide, a ROS
produced at several sites in the mitochondrial electron
transport chain as well as in the mitochondrial matrix [3,37].
When the FeS centre is oxidized by superoxide, Fe 3+ is
released. The free iron levels increases dramatically in E. coli
cells overproducing an FeS-center-containing protein when
they are under oxidative stress [38]. The released iron
presumably binds to the first metal-binding site it comes
across, which may not be protected. The Fenton reaction can
therefore proceed causing damage very close (within a few
nm) to the Fe3+ binding site, because the hydroxyl radical
reacts with virtually the first amino acid residue it encounters.
Thus, protein carbonylation sites are likely to be bioindicators
of adjacent metal binding sites. Contrarywise, if we can
predict metal binding sites then we may be able to predict
sites of oxidative modifications.
As mentioned earlier, the absence of frataxin induces
severe oxidative stress in the mitochondria, which has been
used as a model system for studying oxidative stress in yeast
[30] and in higher plants [29].
3.
Proteomic methods for the detection and
quantification of protein carbonylation
It is relatively straightforward to detect and quantify global
protein carbonylation. It can be done for example by derivatisation of carbonyl groups with 2,4-dinitrophenylhydrazine
(DNPH) followed by spectrophotometric measurements or
immunodetection with DNPH-specific antibodies either in
gels or in ELISA assay [39]. While this methodology proved
robust and is widely used, there exists a possibility to generate
artifactual results if care is not taken [40]. Recently it has been
reported that DNPH does not react selectively only with
carbonyl groups, but it can also form hydrazone derivative
with thioaldehyde derived from sulfenic acid [41]. It is widely
accepted that MCO introduces reactive carbonyl groups in the
side chains of Arg, Lys, Pro and Thr, which can react with
DNPH forming stable 2,4-dinitrophenyl (DNP) hydrazone
products. However, other products of amino acid side chain
oxidation like for example aspartate semialdehyde (oxidation
2232
product of Met) or N-formylkynurenine and kynurenine

(oxidation products of Trp) also contain aldehyde and ketone
groups, respectively. In spite of this, these products are never
mentioned in studies involving DNP-conjugation. This is an
important point that should be resolved.
The ultimate goal, to identify carbonylation type and
localise the affected residue, is technically challenging and
only recently appropriate methods have been developed and
tested in vitro and in vivo [42]. In this section we will
summarize proteomic approaches used for identification
and quantitation of carbonylated proteins and give an
overview of the emerging mass spectrometry- (MS-) based
advances that allow direct identification of carbonylation sites
in the protein. Challenges that are associated with this type of
analysis will be discussed.
3.1.
2D gel based proteomics
To date the majority of proteomic studies of protein carbonylation have used two-dimensional gel electrophoresis (2-DE)
coupled with MS as a means of protein separation, quantification and identification. Supplementary Table S1 lists over 50
proteomic studies identifying carbonylated proteins using 2DE and carbonyl-specific detection methods. This approach
has been successfully applied for studying various diseases in
human subjects and in animal models e.g. Alzheimer disease,
systemic lupus erythematosus, chronic obstructive pulmonary disease, glaucoma and diabetes. In plants it has been
used to follow protein oxidation during the plant's life cycle,
germination and O3 and CO2 stress. Using cultured cell lines,
bacteria and yeasts this method was applied to determine the
impact of various stress factors on protein carbonylation e.g.
H2O2, MCO, arsenite, X-ray irradiation and drug-induced
toxicity as well as displacement of iron and copper homeostasis. It has also been used extensively to study protein
carbonylation during ageing. Recently, two new areas of
application have emerged, following changes in protein
oxidation in sentinel aquatic species and assessing the quality
of frozen stored fish meat (see supplementary Table S1 for
references).
By combining 2-DE with carbonyl-specific probes it is
possible not only to isolate and identify carbonylated proteins,
but also to quantify the degree of carbonylation of each
protein in relation to its overall quantity. Several different
chemical probes for detection of protein carbonyls have been
developed including DNPH, tritiated sodium borohydride,
biotin hydrazine-containing probes and fluorescent probes.
Properties of these probes have been reviewed recently [43].
The most widely used system for detecting carbonylated
proteins on 2D gels is based on DNPH derivatisation and
immunodetection with anti-DNPH antibody. It was first
published in 1994 by Shacter and colleagues [44] and is
commercially available under trade name OxyBlotTM. Three
variants of the protocol are used depending on when in the
process the DNPH derivatisation step is carried out. It can be
carried out before isoelectrofocusing [45]; right after isoelectrofocusing [46,47] or post electrophoretically [48]. Pre-electrophoretic DNPH derivatisation of proteins requires low pH
and the excess reagent has to be removed, which can lead to
uncontrolled loss of proteins. DNPH derivatisation changes
protein mobility and therefore it is not possible to compare the

patterns of carbonylated and non-carbonylated proteins
directly. Preparation of control samples by treating protein
extracts in the same way as for DNPH labelling, but without
DNPH, is mandatory. Post-electrophoretic or isoelectrophoretic staining overcomes those problems and allows direct
comparison between labelled and non-labelled patterns,
which facilitates the quantitation process and MS identification [4648].
Protein carbonyls can also be detected by labelling with
fluorescent carbonyl-reactive probes, for example with fluorescent hydroxylamine [49] or fluorescein-5-thiosemicarbazide [50]. A similar approach based on biotinhydrazide
derivatisation followed by visualisation with avidin fluorescein probes has also been used [51]. Immunoprecipitation of
carbonylated proteins prior to electrophoretic separation has
also been successfully applied for identification of oxidised
proteins in the matrix of rice leaf mitochondria [52] and
susceptibility of endoplasmic reticulum proteins in HL-60 cells
[53].
The advantage of 2-DE based approach is that it is relatively
easily implemented in a molecular biology lab and MS analysis
of proteins can be carried out in a MS dedicated facility. The
separation power of 2-DE simplifies subsequent MS analysis
and the digestion of each spot on a 2D gel gives rise to peptides
from typically one or two proteins. Therefore subsequent
interpretation of the obtained data is relatively simple and one
would expect that it should promote identification of modified/carbonylated peptides. That is unfortunately not the case
with the exception of oxidised Trp residues, which have been
reported in a number of studies [5460]. However, oxidized Trp
has recently been shown to be introduced during gel
electrophoresis [61]. Until now only a single study has
identified a peptide carrying a carbonylated Arg residue in
protein extracted from a 2D gel spot [62].
3.2.
Identification of carbonylation sites in carbonylated
proteins
MS is a central technology in the discovery of post-translational modifications of proteins, enabling mapping of modification sites and subsequent quantification of the abundance
of the modified peptides. It also allows detection of new types
of structures. Modifications are detected from tandem mass
spectra by observing shifts in the expected positions of
fragment peaks compared to the native peptides. In recent
years approaches dedicated to identification of several
different types of modifications (e.g. phosphorylation, glycosylaton, acetylation) have been developed [63,64].
Investigation of any post-translational modification of
proteins presents immense analytical challenges (for reviews
please see [65,66]) mainly due to the fact that they exist in cells
at substoichiometric levels. This means that a modification of
a given site is often present in only a small fraction of the
protein molecules of a given type. This phenomenon is also
true for carbonylated proteins.
In positive-ion operating conditions, ionisation of peptides
strongly depends on the presence of basic sites. These sites
include the N-terminal amine and the side group of Lys, Arg
and His residues. Generation of carbonyl products of Arg and
2233
Table 2a Single amino acids modifications induced by oxidative damage. a Atomic composition and monoisotopic mass of
the difference between the native amino acid and the oxidized product are given.
Amino
acid
Product
Modification of amino acid side chain

Ala
Serine
Arg
Glutamic semialdehyde
Arg
+14 Da b
Arg
Hydroxyarginine
Asn
3-hydroxyasparagine
Asp
Decarboxylation
Asp
3-hydroxyaspartic acid
Cys
Dehydroalanine
Cys
Serine
Cys
Sulfenic acid
Cys
Sulfinic acid
Cys
Sulfonic acid
Gln
+14 Da b
Gln
Hydroxyglutamine
Glu
Decarboxylation
Glu
+14 Da b
Glu
Hydroxyglutamic acid
His
Aspargine
His
Aspartic acid
His
Aspartylurea
His
Formylaspargine
His
2-oxohistidine
Ileu
+14 Da b
Ileu
4-hydroxyisoleucine
Leu
+14 Da b
Leu
3-hydroxyleucine
Lys
Aminoadipic semialdehyde
Lys
+14 Da b
Lys
Hydroxylysine
Lys
Lysine hydroperoxide
Met
Aspartate semialdehyde
Met
Methionine sulfoxide
Met
Methionine sulfone
Met
Homocysteic acid
Phe
Hydroxyphenylalanine, tyrosine
Phe
Dihydroxyphenylalanine (DOPA)
Phe
Trihydroxyphenylalanine (TOPA)
Pro
Pyrrolidinone
Pro
Pyroglutamic acid
Pro
Glutamic semialdehyde
Pro
Hydroxyproline
Pro
Glutamic acid
Ser
Hydroxyserine
Thr
2-amino-3-ketobutyric acid
Thr
Hydroxythreonine
Trp
Kynurenine
Trp
Oxolactone
Trp
2,4,5,6 and 7-hydroxytryptophan
Trp
Hydroxykynurenine
Trp
b-unsaturated-2,4-bis-tryptophandione
Trp
Tryptophandione, dihydrodioxoindole
Trp
N-formylkynurenine, dioxindolylalanine,
dihydroxytryptophan
Trp
Hydroxy-bis-tryptophandione
Trp
Hydroxy-N-formylkynurenine
Trp
Dihydroxy-N-formylkynurenine
Tyr
Dihydroxyphenylalanine (DOPA)
Tyr
Trihydroxyphenylalanine (TOPA)
Val
+14 Da b
Val
Hydroxyvaline
Composition
Monoisotopic
mass change
+1O
5H 1C 3N +1O
2H +1O
+1O
+1O
2H 1C 1O
+1O
2H 1S
1S +1O
+1O
+2O
+3O
2H +1O
+1O
2H 1C 1O
2H +1O
+1O
1H 2C 1N +1O
2H 2C 2N +2O
2H 1C 2N +2O
1H 1C 1N +2O
+1O
2H +1O
+1O
2H +1O
+1O
3H 1N +1O
2H +1O
+1O
+2O
4H 1C 1S +1O
+1O
+2O
2H 1C +3O
+1O
+2O
+3O
2H 1C 1O
2H +1O
+1O
+1O
+2O
+1O
2H
+1O
1C + 1O
2H +1O
+1O
1C + 2O
4H +2O
2H +2O
+2O
+ 15.99492
43.05343
+ 13.97927
+ 15.99492
+ 15.99492
30.01056
+ 15.99492
33.98772
15.97716
+ 15.99492
+ 31.98983
+ 47.98474
+ 13.97927
+ 15.99492
30.01056
+ 13.97927
+ 15.99492
23.01598
22.03197
10.03200
+ 4.97900
+ 15.99492
+ 13.97927
+ 15.99492
+ 13.97927
+ 15.99492
1.03163
+ 13.97927
+ 15.99492
+ 31.98983
32.00846
+ 15.99492
+ 31.98983
+ 33.96910
+ 15.99492
+ 31.98983
+ 47.98475
30.01057
+ 13.97927
+ 15.99492
+ 15.99492
+ 31.98983
+ 15.99492
2.01560
+ 15.99492
+ 3.99490
+ 13.97927
+ 15.99492
+ 19.98983
+ 27.95853
+ 29.97418
+ 31.98983
4H +3O
+3O
+4O
+1O
+2O
2H +1O
+1O
+ 43.95345
+ 47.98474
+ 63.97966
+ 15.99492
+ 31.98983
+ 13.97927
+ 15.99492
Reactive aldehyde
or ketone group
Known to be
induced by MCO
c
c
(continued on next page)
2234
Table 2a (continued)
Amino
acid
Product
Composition
Nitrosylation and chlorination

Phe
2-nitrophenylalanine
Trp
6-Nitrotryptophan
Tyr
3-cholorotyrosine
Tyr
3-nitrotyrosine
a
b
c
d
1H
1H
1H
1H
1N + 2O
1N + 2O
+1Cl
1N + 2O
Monoisotopic
mass change
Reactive aldehyde
or ketone group
Known to be
induced by MCO
+44.98508
+44.98508
+33.69103
+44.98508
Compiled from [60,69,101,115122];

Exact structure of these products is unknown;
Residues potentially containing reactive aldehydes and ketone groups;
Based on [115,122].
Lys (to give glutamic and 2-amino-adipic semialdehydes) is

associated with loss of guanidine and ammonia groups, which
carry the positive charge. This leads to a decreased ionisation
efficiency and reduces the chance of detecting those peptides
using positive-ion mode MS.
What makes analysis of oxidised proteins exceptionally
challenging is the fact that there are many types of modifications of proteins that result in creation of carbonyl residues
and they come in many different sizes (Tables 2a and 2b).
Apart from those modifications there are also many other
oxidative modifications that can be introduced in different
amino acids which can co-exist in oxidised proteins together
with carbonylated residues. Table 2a compiles known products of single amino acid modifications induced by oxidative
damage excluding cross-linked products. Atomic composition
and monoisotopic mass of the difference between the native
amino acid and the oxidized product is given for an easier
inclusion of the modifications into MS data searches. It should
be noted that many of the single amino acid modifications
included in the list have not yet been observed in biological
samples. However, as proved by various experiments using
isolated amino acids and model peptides, these modifications
potentially can be induced by hydroxyl radicals and they
should therefore be included as possible products of MCO.

Additionally a number of products (Table 2a) potentially
containing carbonyl residues, but not commonly regarded as
markers of oxidative stress, is listed. These products, although
probably not abundant, may also contribute to the overall pool
of protein carbonyls in biological systems and they merit
further study.
Yet another angle of complexity is the fact that two
different oxidation products can have the same mass difference. For example, Pro hydroxylation to hydroxyproline
results in addition of +16 Da. The same mass difference is
observed during the conversion of Pro to glutamic semialdehyde. Thus these two very distinct modifications of Pro cannot
be distinguished solely by MS.
Together with the problem of the dynamic range comes also
the complexity challenge, which makes analysis of MS data
significantly more difficult. Identification of low-occupancy
sites among an excess of unmodified peptides is addressed by
techniques that rely on affinity enrichment methods described
below. However, to date no bioinformatics approaches have
been developed to solve the complexity problem.
Typical MS data searching algorithms (e.g. Mascot) can
search with a limited number of set modifications and the
Table 2b Typical advanced glycation and lipoxidation products. a Atomic composition and monoisotopic mass of the
difference between the native amino acid and the oxidized product are given.
Amino acid
Product
Composition
Monoisotopic
mass change
Reactive aldehyde
or ketone group
Advanced glycation products (AGE)

Arg
Glyoxal derived imidazolone
Arg
Glyoxal derived hydroimidazolone
Lys
Glyoxal derived carboxymethyllysine (CML)
Lys
Methylglyoxal derived carboxyethyllysine (CEL)
Arg
Methylglyoxal derived argpyrimidine
Arg, Lys
Methylglyoxal derived tetrahydropyrimidine
Arg
3-deoxyglucosone derived imidazolone A
+2C + 1O
+2H + 2C + 1O
+2H + 2C + 2O
+3H + 3C + 2O
+4H + 6C + 1O
+6H + 7C + 4O
+8H + 6C + 4O
+39.99491
+42.01056
+58.00548
+71.01331
+92.02622
+154.02661
+144.04226
Advanced lipoxidation
Arg
Lys
Lys
Lys
Cys, His, Lys
Cys, His, Lys
+4H + 3C + 1O
+4H + 3C + 1O
+6H + 6C + 1O
+10H + 8C + 1O
+11H + 9C + 2O
+16H + 9C + 2O
+56.02622
+56.02622
+94.04187
+122.07317
+154.09990
+156.22480
products (ALE)
Malondialdehyde derived N-propenalarginine
Malondialdehyde derived N-propenallysine
Acrolein derived FDP-lysine
Croton aldehyde derived dimethyl-FDP-lysine
4-oxo-2-nonenal (Michael adduct)
4-hydroxy-2-nonenal (Michael adduct)
Compiled from [123125].
searching time dramatically increases with an increased

number of modifications. Therefore it is necessary to search
the same data separately several times [67] and all possible
combinations of modifications cannot be included. To be able
to find a modified residue using computer-aided search
algorithms (e.g. Mascot) it is required to define the mass
difference between the modified and non-modified residues.
Therefore only known modifications can be found with this
approach. New types of modifications can only be found if
manual data analysis is carried out, but this approach is labour
intense and time consuming, and cannot be used effectively
with complex protein mixtures.
Identification of oxidatively modified peptides with a search
algorithm poses several other technical problems. MS-based
analysis of proteins inevitably involves the enzymatic degradation of proteins to peptides. The enzyme of choice is trypsin.
This protease has high cleavage specificity by cleaving Cterminally to Arg or Lys residues. Oxidised Lys and Arg residues
become inaccessible to trypsin proteolysis, leading to a higher
number of missed cleavages than normal [68]. As a consequence
the resulting peptides might be outside the analytical range of
mass spectrometers and the possibility of a higher numbers of
missed cleavages has to be included in search algorithms
increasing the search space and search time exponentially and
causing high false discovery rates. Oxidative attack on proteins
results in protein backbone cleavage [69]. Peptides produced by
oxidative fragmentation lack the C-terminal enzyme specificity
(Lys or Arg residues for trypsin) that is required for enzymespecific identification. Thus search algorithms used must
include an option of searching for semitryptic or non-tryptic
peptides, which again increases search space and search time
and can cause high false discovery rates [68]. Oxidation can also
induce protein and peptide cross-linking. Finding cross-linked
peptides in a mixture of proteins is very difficult and requires
special search algorithms.
In spite of the difficulties in identifying carbonylation sites,
the combination of the methods described above with the
enrichment methods described in the following Section 3.3
has lead to the identification of several hundred carbonylation
sites in specific proteins. This will be treated in Section 4.
3.3.
Affinity enrichment based high throughput mass
spectrometry
Proteomic methods that rely exclusively on MS as a means of
protein identification, characterisation and quantitation require a pre-fractionation step that allows for the enrichment
of a certain group or type of proteins in this case
carbonylated proteins. Some of the specific chemical probes
that have been developed for the detection of carbonyl
residues, but also new probes that have been developed, can
serve to affinity purify carbonylated proteins. The use of
different chemical probes for specific enrichment of carbonylated proteins has been reviewed recently [42].
Biotin-hydrazide is an aldehyde and ketone reactive probe;
it reacts with carbonyl groups to form a Schiff base, which is
then reduced with sodium cyanoborohydride to prevent
hydrolysis. The resulting biotin hydrazone can act as a
handle by which carbonylated proteins can be selected
with immobilized avidin or streptavidin resins. This approach
2235
has been used in a number of proteomic studies to enrich for

carbonylated peptides and identify carbonylation sites in
model proteins [31] and in complex protein mixtures e.g.
yeast lysate [34,68,70], rat plasma [71] and human plasma [67].
An important fact is that the interaction between biotin and
avidin is very strong and therefore in most studies monomeric
avidin resins are used. Even though the interaction is so
powerful that efficient elution of peptides cannot be achieved,
this approach is preferred for protein enrichment [42].
2-Iminobiotin, the cyclic guanidino analogue of biotin, first
described by Hofmann and Axelrod in 1950 [72] is an attractive
alternative to biotin. This tag has a pH-dependent interaction
with avidin. At high pH the free base of 2-iminobiotin forms a
stable and tight complex with avidin (KD = 3.5 10 11 M). At low
pH, the interaction becomes much weaker (KD < 10 3 M) [73].
Consequently, the elution of iminobiotinylated peptides can
be achieved by a simple pH shift. The method can be
efficiently used for enrichment of carbonylated peptides and,
combined with de novo sequencing, allows for detection of
carbonylated amino acids [74].
Girard's P reagent provides an alternative approach for
selecting carbonylated peptides. It contains a carbonylreactive group and a quaternary amine group that can be
used for selection with strong cation exchange (SCX) at pH 6.0.
Proteins derivatised by Girard P reagent are digested into
peptides. Peptides containing one or more amino acid residues
with covalently bound Girard P reagent can then be selected by
SCX fractionation and analysed by MS [33]. The advantages of
this approach are that the quaternary amine group present
within a peptide enhances its ionization. However, SCX
selection has a low specificity towards Girard P-derivatised
peptides and many unmodified peptides may contaminate
the sample. This difficulty can be overcome by adding a
quantitative angle to the analysis, where only pairs of peptides
containing heavy and light Girard-P reagent are included in
the analysis [32].
O-ECAT (Oxidation-dependent carbonyl-specific ElementCoded Affinity Mass Tag) is used to tag protein residues
oxidized to aldehydes or ketones. O-ECAT can be synthesized
with a variety of metals, which allows generating peptide pairs
with a defined mass shift and is suited for multiplex analysis.
The O-ECAT moiety also serves as means for affinity
purification, which is achieved using an immobilized antibody
[35]. Although this approach proved applicable for identification of carbonylation sites in oxidized Lys, Arg, Pro, Thr, Glu
and Asp, O-ECAT reagent and antibodies for its affinity
purification are not commercially available limiting its
broader use.
Solid Phase Hydrazide (SPH) chemistry provides another
simple mode to enrich carbonylated peptides in particular 4HNE modified peptides [75]. 4-HNE modified peptides are
captured on hydrazide-coated glass beads by formation of a
covalent, but reversible, hydrazone bond. Peptides are released under acidic conditions and can be analyzed by MS.
This approach has been successfully applied to yeast lysate
treated with 4-HNE [75], 4-HNE modified apomyoglobin [76]
and rat skeletal muscle homogenate treated with 4-HNE [77]. It
has not been shown to identify carbonylated amino acid sites
in complex mixtures of endogenously modified proteins,
highlighting the need for further development of the method.
2236
3.4.
Mass spectrometry-based quantitation of
carbonylated residues
Table 3 MCO specificity as indicated by the number of

times different amino acids have been carbonylated.
Amino acid
In order to achieve a comprehensive understanding of the

mechanisms governing protein oxidation in complex biological systems we need quantitative tools to follow the oxidative
damage to proteins in time and within different cellular
compartments. While MS has been mainly used to map
carbonylation sites within proteins, spectrophotometry and
immunochemistry combined with 2-DE have been the
methods of choice to do the quantitative measurements.
The ultimate goal of redox proteomic studies is to detect and
quantify the oxidized amino acids within proteins. Relative
quantification studies using stable isotope coding allows
comparing the degree of oxidation of a particular site between
two or more samples. Isotopomers of DNPH [78], Girard-P
reagent [32], O-ECAT [35], HICAT [79], iTRAQ [80,81], PIC
reagent [82] and most recently targeted 18O-Labeling [77]
have been used in relative quantification studies of carbonylated proteins. Even though relative quantitation provides
invaluable insights into the mechanisms of protein carbonylation very little is known about the fraction of any particular
protein or protein site being oxidized in selected condition.
Absolute quantitation by MS can be achieved by adding an
internal standard. Because quantitation in MS is based on
integrated ion intensities it is crucial to take into account that
MS signal intensities depend not only on the amount of
sample, but most importantly on the chemical properties of
the peptide. Therefore to measure the absolute quantity of a
peptide a synthetic heavy isotope labelled peptide of the same
sequence and modification has to be used. The method
utilizing internal standard is often referred to as single
reaction monitoring (SRM) or multiple reaction monitoring
(MRM) if several peptides are being followed in a single
experiment. In recent years this technique has been rapidly
developing and is considered one of the most effective tools
for quantitative proteomics, especially clinical proteomics
[83]. In proteomic studies it has not yet been widely used for
monitoring protein carbonylation, but successful quantification of 4-HNE and glutathione adducts to carnosine and
anserine dipeptides in rat skeletal muscles [84] provides a
valuable proof of concept.
4.
Carbonylated sites
identified MCO specificity
Using the methods described in Section 3, it is possible to
identify carbonylated proteins and the site of carbonylation.
We have found 12 proteomic studies in which specific
carbonylated Arg, Lys, Pro and Thr residues have been
identified in a variety of organisms although no plant study
has been found. A total of 456 non-redundant sites in 208
proteins have been identified (Supplementary Table S2). Most
of these studies were in vitro MCO where oxidation was
effected at one fixed (high) metal ion concentration or they
were in vivo MCO where samples were taken under control
conditions (e.g. healthy humans) and/or under stressed
conditions (patients or oxidatively stressed cells). The most
commonly carbonylated amino acid was Lys followed by Pro
Arginine
Lysine
Proline
Threonine
Total
Carbonylated frequency a
90
147
129
90
456
From 208 non-redundant carbonylated proteins [6,3135,62,67,68,

70,126,127].
and Arg/Thr (Table 3). In most cases there is no information

about metal ion binding and how that might contribute to the
oxidation pattern observed.
In one study, in vitro MCO was conducted at a series of
metal ion concentrations ranging in stoichiometry from 1 Fe
per 1500 protein molecule (BSA) to 1 Fe per 1.5 protein
molecule [6]. The ascorbate concentration was also varied so
that there were always 250 ascorbate molecules for every Fe3+
ion thus ensuring the production of an average of 250 HO at
each metal-binding site. This kind of study yields more
information about the MCO mechanism.
Based on the observation of in vitro carbonylation in BSA,
Maisonneuve et al. [6] suggested that some positions in the
amino acid chain are more prone to carbonylation, so-called
hotspots. A hot spot was defined as a four-residue window
containing three Arg, Lys, Pro or Thr (RKPT) residues out of
which at least one is a Pro. Further, there should be an ironbinding residue and a hydrophobic residue in the proximity.
Although the majority of carbonylated sites occurred in RKPT
rich regions in BSA, the hot spot rule was not very effective in
predicting carbonylations in other carbonylated proteins [6].
Maisonneuve et al. [6] also identified carbonylated sites in
BSA at different MCO levels. It was observed that carbonylated
sites at higher MCO levels contained all the sites that were
carbonylated at lower levels. It was therefore proposed
that there is an order in the propagation of carbonylation in
proteins.
Finally, Maisonneuve et al. [6] made the interesting
suggestion that RKPT sites may be more susceptible to
carbonylation near (in the protein primary structure) a
carbonylated site. For instance, many of the carbonylated
sites observed at higher MCO levels were very close to the sites
that were carbonylated at lower levels, indicating some sort of
clustering.
To test whether carbonylated sites have a tendency to
cluster in the protein primary structure as suggested by
Maisonneuve et al. [6], we performed a Monte Carlo simulation
(~ random sampling) of the distance between carbonylation
sites. The pattern of simulated distance between all RKPT
sites, i.e. all potential carbonylation sites, is nearly identical to
the actual observed pattern of carbonylation (Fig. 1A). In
contrast, the simulated pattern of proportion/distance
obtained for carbonylated sites is very different from the
actual pattern (Fig. 1B). Due to the large number of available
RKPT sites, carbonylation has a wide scope to vary. Any
deviation from this (random) expected pattern is therefore
seen as a clear indication of some selective force at work.
2237
outer membrane through porin into the cytosol and when a

free metal ion is available, the Fenton reaction can take
place. Although the total metal ion concentration in mitochondria is very high, especially in the membranes [14], the
free metal ion concentration is strictly regulated by metalbinding proteins such as ferritin and frataxin (see Section 2.3).
That this is important is illustrated by the observation that
frataxin deficiency is the cause of Friedreich's ataxia, a human
cardio- and neurodegenerative disease [25]. At the cellular
level, frataxin deficiency in yeast leads to the accumulation of
a number of carbonylated proteins many of which are
mitochondrial [24,86]. Frataxin overexpression, on the other
hand, leads to a marked decrease in the amount of carbonylated protein especially in the mitochondria and an increased
stress tolerance [24].
When carbonylations are assumed to fall in surface-exposed

region, the proportion at close distance is slightly higher.
When carbonylations are assumed to be in RKPT rich regions
the simulated pattern is very similar to the actual pattern
(Fig. 1B). These are just two extreme assumptions; the actual
pattern may be the cumulative result of several factors
including the presence of a metal binding site. However, the
simulation confirms the conclusion of Maisonneuve et al. [6]
that protein carbonylations have a strong tendency to cluster
probably near metal binding sites. More studies of this type
should be performed to gain a better understanding of the
MCO mechanism.
5.
Biological implications of protein
carbonylation the mitochondrion as a case story
5.1.
Carbonylated proteins have been identified in many different
studies in various biological systems (see Supplementary
Table S2). It would not be meaningful to discuss all of these
studies here. Instead we have decided to select the mitochondrion as a case story.
The mitochondrion is the main site of ROS production in
mammalian cells [37] and in non-photosynthesizing plant
cells [1,3]. Superoxide is produced at seven sites in mitochondria, the two major being complexes I and III in the respiratory
chain. Most of the superoxide is released into the matrix space
where it can be converted into H2O2 by MnSOD. There are
several enzyme systems present in the mitochondrial matrix
that can remove the H2O2 [3]. H2O2 can also diffuse across the
inner mitochondrial membrane into the intermembrane
space presumably through aquaporins [85] and across the
In the respiratory chain of rat skeletal muscle, most carbonylated subunits were found in complexes I and III [81], the sites
of the major part of mitochondrial ROS production [37].
However, carbonylation has also been found in subunits of
the other respiratory complexes [52,81,8789], which produce
only small amounts of ROS (complexes II and IV) or none at all
(complex V) [37]. Likewise many matrix enzymes are carbonylated although they are not themselves ROS producers
[52,81,86,88]. As discussed in Section 2, it requires an
unprotected metal ion as well as H2O2 to give the Fenton
reaction and, since superoxide and H2O2 diffuse throughout
the inner membrane and matrix, the carbonylated proteins
may indicate which proteins are neighbours when metal ions
are released? A major site of iron ion release appears to
25
Actual
Simulated
Simulated surface exposed
Simulated RKPT rich
Actual
Simulated
20
% of sites
Carbonylated mitochondrial proteins
15
10
0
1
10
Distance (amino acid)
100
10
100
Distance (amino acid)
Fig. 1 Monte Carlo simulation of distance between carbonylation sites. A. Over all, simulated profile of distance between
potential carbonylation RKPT (Arg, Lys, Pro, Thr) sites is nearly identical to the actual pattern observed. B. The proportion of
carbonylated sites at lower distance (below 10 amino acids) is very low as per random mechanism. The proportion is slightly
higher when carbonylated sites are considered to be in highly surface exposed regions. However, the simulation pattern is
quite similar to the actual one when carbonylated sites are assumed to be in RKPT rich regions. Monte Carlo simulations were
performed programmatically in Python scripting language (ver. 2.6). Pseudo-random numbers were generated to match the
actual number of RKPT sites or carbonylated sites in a given protein. For simulating RKPT sites, numbers were drawn from
within the range of protein length whereas for simulating carbonylation sites they were drawn from the list of actual positions
of RKPT sites in proteins. Sites (centered in 21-residue window) with hydropathy score (using Kyte-Doolittle scale) below 0.5
were assumed to be in highly surface exposed regions. RKPT rich sites were considered as sites with at least four RKPT residues
within a moving window of seven residues around the RKPT site. Simulations were run for 100,000 iterations.
2238
be aconitase, which is very sensitive to superoxide (see

Section 2.5). If the Krebs cycle is organized in a metabolon in
the matrix [90], then one would expect many Krebs cycle
enzymes to be carbonylated under oxidative stress because of
the proximity of aconitase. Indeed many Krebs cycle enzymes
are known to be carbonylated [52,81,86,88]; however, these
enzymes are quite abundant and would also be expected to be
oxidized frequently if a random diffusion-collision model was
applied. Although we know that oxidative stress generally
leads to loss of enzymatic function e.g., [86,91], in most cases
we do not know how carbonylation at specific sites affects the
function of the modified protein.
It has been suggested that Fe ions replace Mg2+ at binding
sites on the carbonylated proteins or on interacting nucleotides and that the MCO damage occurs near these sites [86,92].
An Fe ion binding instead of Mg2+ could well be unprotected
and able to catalyze the Fenton reaction, since the coordination requirements of the two metals differ markedly.
In two studies, N-formylkynurenine doubly oxidized Trp
residues sites were identified in mitochondrial proteins and
taken as an indicator of in vivo protein oxidation [57,58].
Although it has recently been shown that Trp oxidation can be
a preparation artifact occurring during gel electrophoresis [61],
that may not be true for all oxidized Trp. We also need to keep
in mind that DNPH may react with both singly and doubly
oxidized Trp residues. In fact, nine of the ten rice mitochondrial proteins containing N-formylkynurenine had previously
been identified as carbonylated proteins although the site of
carbonylation was not identified [52,58]. Overall, the pattern of
proteins containing oxidized Trp was similar to that of
carbonylation with modification of many subunits of electron
transport complexes and Krebs cycle enzymes.
5.2.
Carbonylation via conjugation with oxidized
carbohydrates and fatty acids
Another mode of protein carbonylation is by formation of AGE
adducts. In the presence of high levels of reducing carbohydrates, peptides and proteins can become glycated by forming
a Schiff base, which further rearranges to stable Amadori
products that contain carbonyl residues [9]. Oxidative degradation of these products can lead to extensive formation of
carbonyls, which induce protein cross-linking and aggregation
in the cell. Formation and accumulation of AGE adducts is a
major route of protein carbonylation in untreated diabetes
[93], but it is also known to be associated with ageing and
chronic diseases such as Alzheimer's disease [94] and vascular
dementia [95]. Table 2b lists the main AGE products.
Protein carbonylation, primarily on Lys residues, can also
be generated by secondary reaction with aldehydes produced
during PUFA peroxidation. Some of these ALE products are
listed in Table 2b. Cells contain a number of carbonyl
reductases, distributed in all compartments including the
mitochondria, which help to remove low-molecular-weight
compounds containing reactive carbonyl groups before they
can form conjugates with proteins and other macromolecules
[96]. A mitochondrial aldehyde dehydrogenase restores
pollen fertility in cytoplasmic male sterile maize (CMS-T)
possibly because it removes the aldehydes formed by PUFA
oxidation before they can form conjugates with mitochon-
drial proteins and in that way prevent oxidative damage

[97,98]. HNE-conjugation to a number of mitochondrial proteins including subunits of mitochondrial electron transport
complexes has been reported [87,99]. When tested on
mitochondrial enzymes, HNE-conjugation had little or no
effect on the activity of the affected enzyme, but the
association between electron transport complexes appeared
to be affected [99].
It is becoming evident that there exists a close interplay
between the different types of protein carbonylation and MCO,
but the physiological mechanisms controlling these processes
are not yet completely understood. It has been shown that
MCO and free radicals play a major role in the formation of
AGEs and AGE-induced protein cross-linking [7,8]. And MCO of
PUFA in the presence of proteins can also lead to N(6)carboxymethyllysine (CML) formation suggesting that MCO
plays a role in ALE formation as well [9]. In senescent human
fibroblast a number of oxidatively modified proteins
carbonylated, HNE-conjugated, AGE-modified were detected;
half of them were of mitochondrial origin [100]. Several
structural proteins (e.g. vimentin) were found to contain
both AGE and HNE modifications.
5.3.
Protein turnover
Whenever quantifying the amount of a component in a

biological system and its changes with time, we need to
keep in mind that we are observing steady-state levels, which
are the net result of synthesis and degradation (e.g., [101])
Mitochondrial protein carbonylation is a case in point, since
we know that damaged mitochondrial proteins are degraded
by specific proteases [102]. It is therefore entirely possible that
proteins with a low steady-state level of carbonylation in fact
have a very high turnover rate. We do not have any
information about that at the moment.
The quantification of protein carbonylation using the
methods described above typically give values of 14 nmol/
mg of protein rising to 8 nmol/mg of protein under oxidative
stress e.g. caused by disease or age (e.g., [69,103] and
references therein). This is equivalent to 0.050.4 carbonyl
per 50 kDa protein or, expressed differently, as many as 40% of
all protein molecules have one carbonyl group. This is a high
value and the cost of replacing the proteins must be heavy
burden for cells. In mitochondria, which are one of the centres
of oxidative stress, protein turnover may cost as much as 2
20% of the total energy output [103].
We know very little about the extent to which MCO
contributes to protein turnover. The turnover times of three
mitochondrial proteins, uncoupling protein, Mn-SOD and
peroxiredoxin IIf were reported to be 6 h, 72 h and 72 h,
respectively [104]. They differ markedly in their metal binding
and ROS interaction. The uncoupling protein does not contain
bound metal ion and it is activated by superoxide via an
indirect process [105,106]. Mn-SOD containing an Mn ion
converts superoxide to H2O2, while peroxiredoxin IIf containing no bound metal ion reduces a wide range of peroxides
using a pair of Cys residues [107]. Mn-SOD and peroxiredoxin
were both reported to be carbonylated in potato tuber
mitochondria [52], which could be due to the conjugation
with HNE originating from superoxide-mediated PUFA
oxidation [106]. In apple fruit mitochondria only the MnSOD

was observed to be carbonylated [88], in rat skeletal muscle
mitochondria only peroxiredoxin was observed to be carbonylated [81] while Mn-SOD contained an oxidized Trp residue
in rice leaf mitochondria [58] but not in human or bovine heart
mitochondria [57]. These few and disjointed observations do
not give a clear picture of the contribution of MCO to protein
turnover and more focussed experiments are required.
5.4.
Intracellular signalling
ROS not only cause damage to cellular components, but they

are also involved in intracellular signalling and H2O2 is often
presented as being the most likely messenger (e.g., [108])
partly because it can cross membranes through aquaporins
[85]. However, H2O2 does not have the requisite information
content to allow the interaction partner to identify its point of
origin. It has therefore been proposed that irreversibly
(carbonylated) peptides deriving from the proteolytic degradation of oxidized proteins can be the specific secondary ROS
messengers from organelles such as mitochondria or chloroplasts [109]. In this connection it is interesting that carbonylation and subsequent proteolytic degradation of annexin A1
has been suggested to be involved in endothelin-mediated cell
growth and survival although the mechanism by which the
signal is transmitted is unknown [110,111].
6.
Summary
MCO is a common cause of irreversible protein modification,

which increases in living cells as a result of stress, age and
disease. It occurs when a metal ion, often Fe3+ or Cu2+, is
released from its normal protected environment and binds to
an unprotected site. There it interacts with H2O2 in the Fenton
reaction to produce hydroxyl radicals, which oxidize adjacent
amino acid side chains. Carbonyl groups are formed on several
amino acids and this modification can be tagged, e.g. by
reaction with hydrazide derivatives, which in turn can be
recognized e.g. by use of antibodies. This recognition can be
used for quantification of protein carbonylation and in
proteomic strategies (with or without 2-DE) where the
carbonylated proteins are enriched, separated and identified
by use of LC-MS/MS methods. The large number of possible
oxidative modifications makes it particularly difficult to
identify peptides with specific modifications. In spite of this
about 450 carbonylation sites have been identified and they
have a tendency to cluster in RKPT-rich regions. The
mitochondrion is a major site of ROS production in the cell
and many carbonylated mitochondrial proteins have been
identified. Some of these are close to the sites of ROS
production, but many are not and may instead be close to
sites of metal ion release. The carbonylated mitochondrial
proteins are degraded by dedicated proteases, but the extent
to which this contributes to mitochondrial protein turnover is
not understood. We also know very little about the effect of
carbonylation on the properties of the affected proteins and on
their turnover.
Supplementary materials related to this article can be
found online at doi:10.1016/j.jprot.2011.05.004.
2239
Acknowledgements
We are grateful to Dr. Pai Pedas for making unpublished data
available, to Morten J. Bjerrum and Pai Pedas for useful
suggestions and to Dr. Kristian Kristensen for help with the
statistical analysis. This work was supported by a grant from
the Faculty of Agricultural Sciences, Aarhus University to IMM
and by European 6th Framework Program Grant Proteomage
contract no. LSHMCT-518230 to ARW.
REFERENCES
[1] Foyer CH, Noctor G. Redox sensing and signalling associated
with reactive oxygen in chloroplasts, peroxisomes and
mitochondria. Physiol Plant 2003;119:35564.
[2] Halliwell B, Gutteridge JMC. Free Radicals in Biology and
Medicine. Oxford: Oxford University Press; 2007.
[3] Mller IM. Plant mitochondria and oxidative stress: electron
transport, NADPH turnover, and metabolism of reactive
oxygen species. Annu Rev Plant Physiol Plant Mol Biol
2001;52:56191.
[4] Richter T, Munch G, Luth HJ, Arendt T, Kientsch-Engel R,
Stahl P, et al. Immunochemical crossreactivity of antibodies
specific for advanced glycation endproducts with
advanced lipoxidation endproducts. Neurobiol Aging
2005;26:46574.
[5] Valentine JS. Dioxygen reactivity and toxicity. In: Bertini I,
Gray HB, Stiefel EI, Valentine JS, editors. Biological Inorganic
Chemistry Structure and Reactivity. Sausalito: University
Science Books; 2007. p. 3141.
[6] Maisonneuve E, Ducret A, Khoueiry P, Lignon S, Longhi S,
Talla E, et al. Rules governing selective protein
carbonylation. PLoS One 2009;4:e7269.
[7] Sajithlal GB, Chithra P, Chandrakasan G. The role of
metal-catalyzed oxidation in the formation of advanced
glycation end products: an in vitro study on collagen. Free
Radic Biol Med 1998;25:2659.
[8] Sajithlal GB, Chithra P, Chandrakasan G. An in vitro study on
the role of metal catalyzed oxidation in glycation and
crosslinking of collagen. Mol Cell Biochem 1999;194:25763.
[9] Singh R, Barden A, Mori T, Beilin L. Advanced glycation
end-products: a review. Diabetologia 2001;44:12946.
[10] Li Y, Cohenford MA, Dutta U, Dain JA. In vitro nonenzymatic
glycation of guanosine 5-triphosphate by dihydroxyacetone
phosphate. Anal Bioanal Chem 2008;392:118996.
[11] Rae TD, Schmidt PJ, Pufahl RA, Culotta VC, O'Halloran TV.
Undetectable intracellular free copper: the requirement of a
copper chaperone for superoxide dismutase. Science
1999;284:8058.
[12] Petrat F, de Groot H, Sustmann R, Rauen U. The chelatable
iron pool in living cells: a methodically defined quantity. Biol
Chem 2002;383:489502.
[13] Nilsson UA, Bassen M, Savman K, Kjellmer I. A simple and
rapid method for the determination of free iron in
biological fluids. Free Radic Res 2002;36:67784.
[14] Tan Y-F, O'Toole N, Taylor NL, Millar AH. Divalent metal ions
in plant mitochondria and their role in interactions with
proteins and oxidative stress-induced damage to respiratory
function. Plant Physiol 2010;152:74761.
[15] Crichton RR, Pierre JL. Old iron, young copper: from Mars to
Venus. Biometals 2001;14:99112.
[16] Bertini I, Rosato A. From Genes to Metalloproteins: A
Bioinformatic Approach. Wiley-VCH Verlag; 2007. p. 254655.
[17] Lyons TJ, Edie DJ. Transport and storage of metal ions in
biology. In: Bertini I, Gray HB, Stiefel EI, Valentine JS, editors.
2240
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
Biological Inorganic Chemistry Structure and Reactivity.

Sausalito: University Science Books; 2007. p. 3141.
Bertini I, Turano P. Metal Ions and Proteins: Binding,
Stability, and Folding. In: Bertini I, Gray HB, Stiefel EI,
Valentine JS, editors. Biological Inorganic Chemistry
Structure and Reactivity. Sausalito: University Science
Books; 2007. p. 3141.
Srere PA. The infrastructure of the mitochondrial matrix.
Trends Biochem Sci 1980;5:1201.
Valentine JS. Superoxide dismutases and reductases. In:
Bertini I, Gray HB, Stiefel EI, Valentine JS, editors. Biological
Inorganic Chemistry Structure and Reactivity. Sausalito:
University Science Books; 2007. p. 3141.
Arosio P, Ingrassia R, Cavadini P. Ferritins: a family of
molecules for iron storage, antioxidation and more. Biochim
Biophys Acta 2009;1790:58999.
Arosio P, Levi S. Cytosolic and mitochondrial ferritins in the
regulation of cellular iron homeostasis and oxidative
damage. Biochim Biophys Acta 2010;1800:78392.
Tarantino D, Santo N, Morandini P, Casagrande F, Braun HP,
Heinemeyer J, et al. AtFer4 ferritin is a determinant of iron
homeostasis in Arabidopsis thaliana heterotrophic cells.
J Plant Physiol 2010;167:1598605.
Seguin A, Bayot A, Dancis A, Rogowska-Wrzesinska A,
Auchere F, Camadro JM, et al. Overexpression of the yeast
frataxin homolog (Yfh1): contrasting effects on iron-sulfur
cluster assembly, heme synthesis and resistance to
oxidative stress. Mitochondrion 2009;9:1308.
Richardson DR, Huang ML, Whitnall M, Becker EM, Ponka P,
Rahmanto YS. The ins and outs of mitochondrial
iron-loading: the metabolic defect in Friedreich's ataxia.
J Mol Med 2010;88:3239.
Murgia I, Tarantino D, Soave C. Mitochondrial iron
metabolism in plants: frataxin comes into play. Plant Soil
2009;325:514.
Levi S, Rovida E. The role of iron in mitochondrial function.
Biochim Biophys Acta 2009;1790:62936.
Vazzola V, Losa A, Soave C, Murgia I. Knockout of frataxin
gene causes embryo lethality in Arabidopsis. FEBS Lett
2007;581:66772.
Busi MV, Maliandi MV, Valdez H, Clemente M, Zabaleta EJ,
Araya A, et al. Deficiency of Arabidopsis thaliana frataxin
alters activity of mitochondrial FeS proteins and induces
oxidative stress. Plant J 2006;48:87382.
Kim JH, Sedlak M, Gao Q, Riley CP, Regnier FE, Adamec J.
Oxidative stress studies in yeast with a frataxin mutant: a
proteomics perspective. J Proteome Res 2010;9:7306.
Temple A, Yen TY, Gronert S. Identification of specific
protein carbonylation sites in model oxidations of human
serum albumin. J Am Soc Mass Spectrom 2006;17:117280.
Mirzaei H, Regnier F. Identification and quantification of
protein carbonylation using light and heavy isotope labeled
Girard's P reagent. J Chromatogr A 2006;1134:12233.
Mirzaei H, Regnier F. Enrichment of carbonylated peptides
using Girard P reagent and strong cation exchange
chromatography. Anal Chem 2006;78:7708.
Mirzaei H, Regnier F. Affinity chromatographic selection of
carbonylated proteins followed by identification of oxidation
sites using tandem mass spectrometry. Anal Chem 2005;77:
238692.
Lee S, Young NL, Whetstone PA, Cheal SM, Benner WH,
Lebrilla CB, et al. Method to site-specifically identify and
quantitate carbonyl end products of protein oxidation using
oxidation-dependent element coded affinity tags (O-ECAT)
and nanoliquid chromatography Fourier transform mass
spectrometry. J Proteome Res 2006;5:53947.
Lee JW, Helmann JD. The PerR transcription factor senses
H2O2 by metal-catalysed histidine oxidation. Nature
2006;440:3637.
[37] Brand MD. The sites and topology of mitochondrial

superoxide production. Exp Gerontol 2010;45:46672.
[38] Keyer K, Imlay JA. Superoxide accelerates DNA damage by
elevating free-iron levels. Proc Natl Acad Sci U S A 1996;93:
1363540.
[39] Yan LJ. Analysis of oxidative modification of proteins. Curr
Protoc Protein Sci 2009 Chapter 14:Unit14 4.
[40] Luo S, Wehr NB. Protein carbonylation: avoiding pitfalls in
the 2,4-dinitrophenylhydrazine assay. Redox Rep 2009;14:
15966.
[41] Dalle-Donne I, Carini M, Orioli M, Vistoli G, Regazzoni L,
Colombo G, et al. Protein carbonylation:
2,4-dinitrophenylhydrazine reacts with both
aldehydes/ketones and sulfenic acids. Free Radic Biol Med
2009;46:14119.
[42] Madian AG, Regnier FE. Proteomic identification of
carbonylated proteins and their oxidation sites. J Proteome
Res 2010;9:376680.
[43] Yan LJ, Forster MJ. Chemical probes for analysis of
carbonylated proteins: a review. J Chromatogr B Analyt
Technol Biomed Life Sci 2011;879:130815.
[44] Shacter E, Williams JA, Lim M, Levine RL. Differential
susceptibility of plasma proteins to oxidative modification:
examination by western blot immunoassay. Free Radic Biol
Med 1994;17:42937.
[45] Sultana R, Newman SF, Huang Q, Butterfield DA. Detection of
carbonylated proteins in two-dimensional sodium dodecyl
sulfate polyacrylamide gel electrophoresis separations.
Methods Mol Biol 2008;476:15363.
[46] Choi J, Malakowsky CA, Talent JM, Conrad CC, Gracy RW.
Identification of oxidized plasma proteins in Alzheimer's
disease. Biochem Biophys Res Commun 2002;293:156670.
[47] Conrad CC, Choi J, Malakowsky CA, Talent JM, Dai R,
Marshall P, et al. Identification of protein carbonyls after
two-dimensional electrophoresis. Proteomics 2001;1:82934.
[48] Conrad CC, Talent JM, Malakowsky CA, Gracy RW.
Post-electrophoretic identification of oxidized proteins. Biol
Proced Online 2000;2:3945.
[49] Poon HF, Abdullah L, Reed J, Doore SM, Laird C, Mathura V,
et al. Improving image analysis in 2DGE-based redox
proteomics by labeling protein carbonyl with fluorescent
hydroxylamine. Biol Proced Online 2007;9:6572.
[50] Chaudhuri AR, de Waal EM, Pierce A, Van Remmen H, Ward
WF, Richardson A. Detection of protein carbonyls in aging
liver tissue: a fluorescence-based proteomic approach. Mech
Ageing Dev 2006;127:84961.
[51] Yoo BS, Regnier FE. Proteomic analysis of carbonylated
proteins in two-dimensional gel electrophoresis using
avidinfluorescein affinity staining. Electrophoresis 2004;25:
133441.
[52] Kristensen BK, Askerlund P, Bykova NV, Egsgaard H, Mller
IM. Identification of oxidised proteins in the matrix of rice
leaf mitochondria by immunoprecipitation and
two-dimensional liquid chromatography-tandem mass
spectrometry. Phytochemistry 2004;65:183951.
[53] England K, Cotter T. Identification of carbonylated proteins
by MALDI-TOF mass spectroscopy reveals susceptibility of
ER. Biochem Biophys Res Commun 2004;320:12330.
[54] Finley EL, Dillon J, Crouch RK, Schey KL. Identification of
tryptophan oxidation products in bovine alpha-crystallin.
Protein Sci 1998;7:23917.
[55] Bienvenut WV, Deon C, Pasquarello C, Campbell JM, Sanchez JC,
Vestal ML, et al. Matrix-assisted laser
desorption/ionization-tandem mass spectrometry with high
resolution and sensitivity for identification and
characterization of proteins. Proteomics 2002;2:86876.
[56] Vazquez S, Aquilina JA, Jamie JF, Sheil MM, Truscott RJ. Novel
protein modification by kynurenine in human lenses. J Biol
Chem 2002;277:486773.
[57] Taylor SW, Fahy E, Murray J, Capaldi RA, Ghosh SS. Oxidative
post-translational modification of tryptophan residues in
cardiac mitochondrial proteins. J Biol Chem 2003;278:
1958790.
[58] Mller IM, Kristensen BK. Protein oxidation in plant
mitochondria detected as oxidized tryptophan. Free Radic
Biol Med 2006;40:4305.
[59] Chetyrkin SV, Mathis ME, Ham AJ, Hachey DL, Hudson BG,
Voziyan PA. Propagation of protein glycation damage
involves modification of tryptophan residues via reactive
oxygen species: inhibition by pyridoxamine. Free Radic Biol
Med 2008;44:127685.
[60] Fedorova M, Todorovsky T, Kuleva N, Hoffmann R.
Quantitative evaluation of tryptophan oxidation in actin and
troponin I from skeletal muscles using a rat model of acute
oxidative stress. Proteomics 2010;10:2692700.
[61] Perdivara I, Deterding LJ, Przybylski M, Tomer KB. Mass
spectrometric identification of oxidative modifications of
tryptophan residues in proteins: chemical artifact or
post-translational modification? J Am Soc Mass Spectrom
2010;21:11147.
[62] Oikawa S, Yamada T, Minohata T, Kobayashi H, Furukawa A,
Tada-Oikawa S, et al. Proteomic identification of
carbonylated proteins in the monkey hippocampus after
ischemiareperfusion. Free Radic Biol Med 2009;46:14727.
[63] Reinders J, Sickmann A. Modificomics: posttranslational
modifications beyond protein phosphorylation and
glycosylation. Biomol Eng 2007;24:16977.
[64] Ytterberg AJ, Jensen ON. Modification-specific proteomics in
plant biology. J Proteom 2010;73:224966.
[65] Jensen ON. Interpreting the protein language using
proteomics. Nat Rev Mol Cell Biol 2006;7:391403.
[66] Zhao Y, Jensen ON. Modification-specific proteomics:
strategies for characterization of post-translational
modifications using enrichment techniques. Proteomics
2009;9:463241.
[67] Madian AG, Regnier FE. Profiling carbonylated proteins in
human plasma. J Proteome Res 2010;9:133043.
[68] Mirzaei H, Regnier F. Creation of allotypic active sites during
oxidative stress. J Proteome Res 2006;5:215968.
[69] Stadtman ER, Berlett BS. Reactive oxygen-mediated protein
oxidation in aging and disease. Chem Res Toxicol 1997;10:
48594.
[70] Mirzaei H, Regnier F. Identification of yeast oxidized
proteins: chromatographic top-down approach for
identification of carbonylated, fragmented and cross-linked
proteins in yeast. J Chromatogr A 2007;1141:2231.
[71] Mirzaei H, Baena B, Barbas C, Regnier F. Identification of
oxidized proteins in rat plasma using avidin chromatography
and tandem mass spectrometry. Proteomics 2008;8:151627.
[72] Hofmann K, Axelrod AE. Microbiological activity of the
guanido analogues of biotin and oxybiotin. J Biol Chem
1950;187:2933.
[73] Green NM. Thermodynamics of the binding of biotin and
some analogues by avidin. Biochem J 1966;101:77480.
[74] Rocha AP, Jensen KB, Roepstorff P, Rogowska-Wrzesinska A.
Modification specific proteomics: specific enrichment and
identification of carbonylated sites in proteins. Poster at 57th
Conference on Mass Spectrometry of the American Society
of Mass Spectrometry, Philadelphia, USA; 2009.
[75] Roe MR, Xie H, Bandhakavi S, Griffin TJ. Proteomic mapping
of 4-hydroxynonenal protein modification sites by
solid-phase hydrazide chemistry and mass spectrometry.
Anal Chem 2007;79:374756.
[76] Rauniyar N, Prokai-Tatrai K, Prokai L. Identification of
carbonylation sites in apomyoglobin after exposure to
4-hydroxy-2-nonenal by solid-phase enrichment and liquid
chromatography-electrospray ionization tandem mass
spectrometry. J Mass Spectrom 2010;45:398410.
2241
[77] Roe MR, McGowan TF, Thompson LV, Griffin TJ. Targeted
18O-labeling for improved proteomic analysis of
carbonylated peptides by mass spectrometry. J Am Soc Mass
Spectrom 2010;21:1190203.
[78] Kinumi T, Osaka I, Hayashi A, Kawai T, Matsumoto H,
Tsujimoto K. Protein carbonylation detected with light and
heavy isotope-labelled 2,4-dinitrophenylhydrazine by
matrix-assisted lasre desorption/ionisation time-of-flight
mass spectrometry. J Mass Spectrom Soc Jpn 2009;57:3717.
[79] Han B, Stevens JF, Maier CS. Design, synthesis, and
application of a hydrazide-functionalized isotope-coded
affinity tag for the quantification of oxylipid-protein
conjugates. Anal Chem 2007;79:334254.
[80] Kim JH, Sedlak M, Gao Q, Riley CP, Regnier FE, Adamec J.
Dynamics of Protein Damage in Yeast Frataxin Mutant
Exposed to Oxidative Stress. Omics 2010;14:68999.
[81] Meany DL, Xie H, Thompson LV, Arriaga EA, Griffin TJ.
Identification of carbonylated proteins from enriched rat
skeletal muscle mitochondria using affinity
chromatography-stable isotope labeling and tandem mass
spectrometry. Proteomics 2007;7:115063.
[82] Suman SP, Faustman C, Stamer SL, Liebler DC. Proteomics of
lipid oxidation-induced oxidation of porcine and bovine
oxymyoglobins. Proteomics 2007;7:62840.
[83] Kim K, Kim Y. Preparing multiple-reaction monitoring for
quantitative clinical proteomics. Expert Rev Proteomics
2009;6:2259.
[84] Orioli M, Aldini G, Beretta G, Facino RM, Carini M.
LC-ESI-MS/MS determination of 4-hydroxy-trans-2-nonenal
Michael adducts with cysteine and histidine-containing
peptides as early markers of oxidative stress in excitable
tissues. J Chromatogr B Analyt Technol Biomed Life Sci
2005;827:10918.
[85] Bienert GP, Moller AL, Kristiansen KA, Schulz A, Mller IM,
Schjoerring JK, et al. Specific aquaporins facilitate the
diffusion of hydrogen peroxide across membranes. J Biol
Chem 2007;282:118392.
[86] Irazusta V, Moreno-Cermeno A, Cabiscol E, Ros J, Tamarit J.
Major targets of iron-induced protein oxidative damage in
frataxin-deficient yeasts are magnesium-binding proteins.
Free Radic Biol Med 2008;44:171223.
[87] Choksi KB, Boylston WH, Rabek JP, Widger WR,
Papaconstantinou J. Oxidatively damaged proteins of heart
mitochondrial electron transport complexes. Biochim
Biophys Acta 2004;1688:95101.
[88] Qin G, Wang Q, Liu J, Li B, Tian S. Proteomic analysis of
changes in mitochondrial protein expression during fruit
senescence. Proteomics 2009;9:424153.
[89] Wen JJ, Garg N. Oxidative modification of mitochondrial
respiratory complexes in response to the stress of
Trypanosoma cruzi infection. Free Radic Biol Med 2004;37:
207281.
[90] Robinson Jr JB, Srere PA. Organization of Krebs tricarboxylic
acid cycle enzymes in mitochondria. J Biol Chem 1985;260:
108005.
[91] Stadtman ER. Protein oxidation and aging. Free Radic Res
2006;40:12508.
[92] Irazusta V, Obis E, Moreno-Cermeno A, Cabiscol E, Ros J,
Tamarit J. Yeast frataxin mutants display decreased
superoxide dismutase activity crucial to promote
protein oxidative damage. Free Radic Biol Med 2010;48:
41120.
[93] Giacco F, Brownlee M. Oxidative stress and diabetic
complications. Circ Res 2010;107:105870.
[94] Reddy VP, Zhu X, Perry G, Smith MA. Oxidative stress in
diabetes and Alzheimer's disease. J Alzheimers Dis 2009;16:
76374.
[95] Taguchi A. Vascular factors in diabetes and Alzheimer's
disease. J Alzheimers Dis 2009;16:85964.
2242
[96] Oppermann U. Carbonyl reductases: the complex

relationships of mammalian carbonyl- and
quinone-reducing enzymes and their role in physiology.
Annu Rev Pharmacol Toxicol 2007;47:293322.
[97] Liu F, Cui X, Horner HT, Weiner H, Schnable PS.
Mitochondrial aldehyde dehydrogenase activity is required
for male fertility in maize. Plant Cell 2001;13:106378.
[98] Mller IM. A more general mechanism of cytoplasmic male
fertility? Trends Plant Sci 2001;6:560.
[99] Winger AM, Taylor NL, Heazlewood JL, Day DA, Millar AH.
The Cytotoxic lipid peroxidation product
4-hydroxy-2-nonenal covalently modifies a selective range
of proteins linked to respiratory function in plant
mitochondria. J Biol Chem 2007;282:3743647.
[100] Ahmed EK, Rogowska-Wrzesinska A, Roepstorff P, Bulteau
AL, Friguet B. Protein modification and replicative
senescence of WI-38 human embryonic fibroblasts. Aging
Cell 2010;9:25272.
[101] Berlett BS, Stadtman ER. Protein oxidation in aging, disease,
and oxidative stress. J Biol Chem 1997;272:203136.
[102] Bayot A, Gareil M, Rogowska-Wrzesinska A, Roepstorff P,
Friguet B, Bulteau AL. Identification of novel oxidized protein
substrates and physiological partners of the mitochondrial
ATP-dependent Lon-like protease Pim1. J Biol Chem
2010;285:1144557.
[103] Mller IM, Jensen PE, Hansson A. Oxidative modifications to
cellular components in plants. Annu Rev Plant Biol 2007;58:
45981.
[104] Sweetlove LJ, Mller IM. Oxidation of proteins in
plants mechanisms and consequences. Adv Bot Res
2009;52:123.
[105] Smith AM, Ratcliffe RG, Sweetlove LJ. Activation and
function of mitochondrial uncoupling protein in plants.
J Biol Chem 2004;279:5194452.
[106] Azzu V, Brand MD. The onoff switches of the mitochondrial
uncoupling proteins. Trends Biochem Sci 2010;35:298307.
[107] Gama F, Keech O, Eymery F, Finkemeier I, Gelhaye E,
Gardestrm P, et al. The mitochondrial type II peroxiredoxin
from poplar. Physiol Plant 2007;129:196206.
[108] Forman HJ, Maiorino M, Ursini F. Signaling functions of
reactive oxygen species. Biochemistry 2010;49:83542.
[109] Mller IM, Sweetlove LJ. ROS signaling specificity is
required. Trends Plant Sci 2010;15:3704.
[110] Wong CM, Cheema AK, Zhang L, Suzuki YJ. Protein
carbonylation as a novel mechanism in redox signalling. Circ
Res 2008;102:3108.
[111] Wong CM, Marcocci L, Liu L, Suzuki YJ. Cell signaling by
protein carbonylation and decarbonylation. Antioxid Redox
Signal 2010;12:393404.
[112] Koizumi M, Fujii J, Suzuki K, Inoue T, Gutteridge JM,
Taniguchi N. A marked increase in free copper levels in the
[113]
[114]
[115]
[116]
[117]
[118]
[119]
[120]
[121]
[122]
[123]
[124]
[125]
[126]
[127]
plasma and liver of LEC rats: an animal model for Wilson

disease and liver cancer. Free Radic Res 1998;28:44150.
Petrat F, Rauen U, de Groot H. Determination of the
chelatable iron pool of isolated rat hepatocytes by digital
fluorescence microscopy using the fluorescent probe, phen
green SK. Hepatology 1999;29:11719.
Epsztejn S, Kakhlon O, Glickstein H, Breuer W, Cabantchik I.
Fluorescence analysis of the labile iron pool of mammalian
cells. Anal Biochem 1997;248:3140.
Xu G, Chance MR. Hydroxyl radical-mediated modification
of proteins as probes for structural proteomics. Chem Rev
2007;107:351443.
Toda T, Nakamura M, Morisawa H, Hirota M, Nishigaki R,
Yoshimi Y. Proteomic approaches to oxidative protein
modifications implicated in the mechanism of aging. Geriatr
Gerontol Int 2010;10(Suppl 1):S2531.
Shringarpure R, Grune T, Davies KJA. Protein oxidation and
20S proteasome-dependent proteolysis in mammalian cells.
Cell Mol Life Sci 2001;58:144250.
Dean RT, Fu S, Stocker R, Davies MJ. Biochemistry and
pathology of radical-mediated protein oxidation. Biochem J
1997;324(Pt 1):118.
Grosvenor AJ, Morton JD, Dyer JM. Profiling of residue-level
photo-oxidative damage in peptides. Amino Acids 2010;39:
28596.
Guedes S, Vitorino R, Domingues R, Amado F, Domingues P.
Oxidation of Bovine Serum Albumin: Identification of
Oxidation Products and Structural Modifications. John Wiley
& Sons, Ltd.; 2009. p. 230715.
Bern M, Saladino J, Sharp JS. Conversion of Methionine into
Homocysteic Acid in Heavily Oxidized Proteomics Samples.
John Wiley & Sons, Ltd.; 2010. p. 76872.
Stadtman ER. Oxidation of free amino acids and amino acid
residues in proteins by radiolysis and by metal-catalyzed
reactions. Annu Rev Biochem 1993;62:797821.
Uchida K. Role of reactive aldehyde in cardiovascular
diseases. Free Radic Biol Med 2000;28:168596.
Uchida K. Histidine and lysine as targets of oxidative
modification. Amino Acids 2003;25:24957.
Cotham WE, Metz TO, Ferguson PL, Brock JWC, Hinton DJS,
Thorpe SR, et al. Proteomic analysis of arginine adducts on
glyoxal-modified ribonuclease. Mol Cell Proteomics 2004;3:
114553.
Kim J-H, Sedlak M, Gao Q, Riley CP, Regnier FE, Adamec J.
Dynamics of protein damage in yeast frataxin mutant
exposed to oxidative stress. OMICS J Integr Biol 2010;14:
68999.
Mirzaei H, Regnier F. Protein:protein aggregation induced by
protein oxidation. J Chromatogr B Analyt Technol Biomed
Life Sci 2008;873:814.

Protein Carbonylation and Metal-Catalyzed Protein Oxidation in A Cellular Perspective

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Protein Carbonylation and Metal-Catalyzed Protein Oxidation in A Cellular Perspective

Transféré par

Droits d'auteur :

Formats disponibles

J O U R NA L OF PR O TE O MI CS 74 ( 20 1 1 ) 2 2 2 82 2 4 2

Protein carbonylation and metal-catalyzed protein oxidation in

Received 16 February 2011

Accepted 3 May 2011

products of polyunsaturated fatty acids and carbohydrates. While reversible oxidative

Available online 11 May 2011

modifications are thought to be relevant in physiological processes, irreversible oxidative

Proteomic methods for the detection and quantification of protein carbonylation . . .

The production of Reactive Oxygen Species (ROS) is an

What causes protein carbonylation?

Carbonyl groups (reactive aldehydes and ketones) can be

chains or they can be generated through oxidative cleavage of

The hydroxyl radical oxidizes amino acid side chains or

This is an example of protein carbonylation formed in the

Metal ions in vivo

To get an idea of the importance and mechanism of MCO, let

Metal binding in vivo

The use of ferric and ferrous ions in living cells developed in

such proteins and transporters until it reaches its final

Table 1 Concentration of metal ions in living cells.

cytochrome c oxidase, which all interact with either H2O2 or O2

MCO has often been used in model studies as a convenient

As we have seen, the concentration of free metal ions in the

product of Met) or N-formylkynurenine and kynurenine

2D gel based proteomics

protein mobility and therefore it is not possible to compare the

Modification of amino acid side chain

(continued on next page)

Nitrosylation and chlorination

Compiled from [60,69,101,115122];

Lys (to give glutamic and 2-amino-adipic semialdehydes) is

should therefore be included as possible products of MCO.

Advanced glycation products (AGE)

Compiled from [123125].

searching time dramatically increases with an increased

has been used in a number of proteomic studies to enrich for

Table 3 MCO specificity as indicated by the number of

In order to achieve a comprehensive understanding of the

From 208 non-redundant carbonylated proteins [6,3135,62,67,68,

and Arg/Thr (Table 3). In most cases there is no information

outer membrane through porin into the cytosol and when a

When carbonylations are assumed to fall in surface-exposed

Carbonylated mitochondrial proteins

Distance (amino acid)

Distance (amino acid)

be aconitase, which is very sensitive to superoxide (see

drial proteins and in that way prevent oxidative damage

Whenever quantifying the amount of a component in a

oxidation [106]. In apple fruit mitochondria only the MnSOD

ROS not only cause damage to cellular components, but they

MCO is a common cause of irreversible protein modification,

Biological Inorganic Chemistry Structure and Reactivity.

[37] Brand MD. The sites and topology of mitochondrial

[96] Oppermann U. Carbonyl reductases: the complex

plasma and liver of LEC rats: an animal model for Wilson

Vous aimerez peut-être aussi