Nonenzymatic posttranslational protein modifications in ageing

Available online at www.sciencedirect.com

Experimental Gerontology 43 (2008) 247–257 www.elsevier.com/locate/expgero

Mini Review

Nonenzymatic posttranslational protein modifications in ageing Vukic´ Sˇosˇkic´, Karlfried Groebe, Andre´ Schrattenholz

*

ProteoSys AG, Carl Zeiss Strasse 51, 55129 Mainz, Germany Received 13 August 2007; received in revised form 7 November 2007; accepted 4 December 2007 Available online 14 December 2007

Abstract One of the most fundamental molecular aspects of aging is accumulating oxidative damage caused by reactive oxygen species (ROS) as proposed by the free radical theory of aging. These unwanted chemical side products of normal metabolism lead to the formation of altered, less active and potentially toxic species of DNA, RNA, proteins, lipids, and small molecules. Due to gradually accumulating small contributions of irreversible reactions during ageing, uncatalyzed chemical side reactions occur with increasing frequencies and repair functions decline. Eventually key biochemical pathways are impaired by increasingly less efficient cellular stress management. In this review, we describe the chemical nature of nonenzymatic age-related modifications of proteins and provide an overview of related analytical challenges and approaches, with a focus on mass spectrometry. We include the description of a strategy to rapidly exploit the wealth of mass spectrometric information from standard MALDI-TOF peptide fingerprints for the characterisation of age-related oxidative amino acid modifications.  2007 Elsevier Inc. All rights reserved. Keywords: Ageing; Oxidation; Reactive oxygen species; Mass spectroscopy; Posttranslational modification; Carbonylation; Nitrosylation; Proteomics; Advanced glycation; AGE

1. Introduction Age-related chemical side reactions that can occur on proteins include: racemisation (McCudden and Kraus, 2006), deamidation (Robinson and Robinson, 2001), oxiAbbreviations: 2D-PAGE, two-dimensional polyacrylamide gel electrophoresis; AGE, advanced glycation end-product; ALA, advanced lipoxidation end-product; Asn, asparagine; Asp, aspartic acid; Trp, tryptophan; Tyr, tyrosine; CEL, Ne-(carboxyethyl)lysine; CML, Ne-(carboxymethyl)lysine; DNPH, 2,4 dinitrophenylhydrazine; DNP-hydrazone, 2,4dinitrophenylhydrazone; DOGDIC, 3-deoxyglucosone-derived imidazoline cross-link; GODIC, glyoxal-derived imidazoline cross-link; GOLD, glyoxal-lysine dimer; LC-MS/MS, liquid chromatography tandem mass spectrometry; MALDI-TOF, matrix-assisted laser desorption/ionization; MetSOx, methionine sulfoxide and methionine sulfone; MOLD, methylglyoxal-lysine dimer; MODIC, methylglyoxal-derived imidazoline crosslink; MS/MS, tandem mass spectrometry; PMF, peptide mass fingerprinting; Q-TOF, quadrupole time of flight mass spectrometry; ROS, reactive oxygen species; RNS, reactive nitrogen species. * Corresponding author. Tel.: +49 6131 5019215; fax: +49 6131 5019211. E-mail address: [email protected] (A. Schrattenholz). 0531-5565/$ - see front matter  2007 Elsevier Inc. All rights reserved. doi:10.1016/j.exger.2007.12.001

dation of amino acids (Stadtman, 2004, 2006; Stadtman et al., 2005), formation of adducts involving reactive nitrogen and chlorine species (van der Vliet et al., 1995), chemical modification of proteins by products of lipid peroxidation reactions (lipoxidation) and Maillard reaction products (Baynes, 2000, 2001, 2002); Table 1 shows a summary. 2. Oxidative modifications It is now beyond doubt that reactive oxygen species (ROS) and/or reactive nitrogen species (RNS) generated in vivo, play a role in aging, as already proposed in 1956 (Harman, 1956; Beckman and Ames, 1998). Since reactive by-products of normal metabolism also lead to damage (Hayflick, 2007), this theory has recently been extended to the oxidative ‘‘garbage catastrophe theory’’ where ROS or reactive oxygen intermediates are responsible for the accumulation of age-related cellular damage of biomolecules (Stadtman, 2004, 2006; Stadtman et al., 2005). Studies on oxidatively modified proteins have revealed an

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

248

Table 1 Age-related amino acid residue modifications Amino acid

Reaction type/product formed Oxidation

Maillard reaction

Isomerisation

Deamidation

Arg

Glutamic semialdehyde

na

na

Asn Asp

na na

Advanced glycation end-product (AGE) and advanced lipoxidation end-product (ALE) Na na

na

Asp na

Cys Gln Glu

Sulfinic acids and Cysteic acid na 4-Hydroxy-glutamate

na na na

His Leu Lys

2-Oxo-histidine 5- Hydroxyl-leucine 2-Aminoadipic-semialdehyde, 3-, 4- and 5-hydroxy lysine MetSOx 2-, 3-phenylanine and Tyr glutamic semialdehyde, pyroglutamic acid, 2pyrrolidone, 4-hydroxy-proline 2-Amino-3-keto-butyric acid N-formyl-kynurenine, kynurenine, 2-,4-,5-, 6- and 7hydroxy-tryptophan 3,4-Dihydroxy phenylanine Tyr–Tyr cross-linked proteins, 3-nitro-tyrosine, 3,5-dichloro-tyrosine 3- and 4-hydroxy valine

na na Advanced glycation end-product (AGE) and advanced lipoxidation end-product (ALE) na na na

Met Phe Pro Thr Trp Tyr Val

L-Isoaspartate D-Aspartate na na L-Isoglutamate D-Glutamate na na na

na Glu na na na na

na na

na na cis/trans isomerisation na na

na na

na

na

na

na

na

na

age-related increase in the level of protein carbonylation (Levine, 2002), oxidized methionine (Wells-Knecht et al., 1997), cross-linked (Squier and Bigelow, 2000) and glycated proteins (Baynes, 2001), as well as the accumulation of catalytically less active enzymes (Rothstein, 1985) that are more susceptible to heat inactivation and to proteolytic degradation (Stadtman, 2001). One of the best-known markers of age-related protein oxidation is the carbonyl group. The carbonyl content of proteins has been observed to increase with age (Levine, 2002). The main carbonyl products of metal-catalyzed oxidation of proteins in vitro have been shown to be glutamic and aminoadipic semialdehydes (Requen2001; Pamplona et al., 2005). Other markers may be derived from ROS-induced protein oxidation, but may be susceptible to further reactions. Tyrosine residues may be oxidized by hypochlorite, peroxynitrite or by radicals formed in transition metal ion-catalyzed Fenton and Haber-Weiss reactions (e.g. hydrogen peroxide/Fe2+). The ensuing tyrosyl radicals may subsequently form intra- or intermolecular Tyr–Tyr bonds (Balasubramanian and Kanwar, 2002; van der Vliet et al., 1995). Several other oxidized residues, like hydroperoxides of amino acid side chains are highly unstable. N-formylkynurenine (an oxidation product of Trp) can be generated enzymatically and non-enzymatically (Korlimbinis and Truscott, 2006). MetSOx and disulfides may be enzymatically reduced (Chao et al., 1997). Intermolecular cross-links are the most important age-related chemical alterations in collagen and elastin. These cross-links are initially formed (through lysyl oxidase) to provide optimal function during

na na na

development and maturation, but can subsequently overstiffen and compromise the structure and function of the fibers throughout the body when present in excess (Bailey, 2001). 3. Spontaneous deamidation, isomerization, and racemization of aspartyl and asparaginyl residues Asparagine and aspartyl residues represent hot spots for spontaneous protein degradation under physiological conditions (Clarke, 2003). For both types of residues, the nucleophilic attack of the peptide-bond nitrogen atom of the following residue on the side chain carbonyl group results in the formation of a five-membered succinimide ring intermediate as shown in Fig. 1 (Dehart and Anderson, 2007). The succinimidyl residue is hydrolyzing with half-times of hours under cellular conditions to give a mixture of aspartyl and isoaspartyl forms. The latter residues induce kinks in polypeptide chains. The succinimide is also racemization-prone (Radkiewicz et al., 2001) and generates the D-succinimidyl, D-aspartyl and D-isoaspartyl forms. Thus, from the original L-aspartyl and L-asparaginyl residues encoded by protein biosynthesis reactions, spontaneous aging results in the formation of at least five altered forms, i.e. D-aspartyl-, D- and L-isoaspartyl-, and D- and L-succinimidyl isoforms. Of these, the L-isoaspartyl form is the most frequently found. Spontaneous direct hydrolysis of asparagine residues by water attack on the side chain amide group can also result in aspartyl residue formation (Robinson and Robinson, 2001; Robinson, 2002). However, at neutral pH, the rate

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257 H N

O H N COO-

Protein

L-isoaspartyl

O N

L-Asp or L-Asn N H

Protein methyltransferase

O O

H N

H N CO2Me L-isoaspartyl methyl ester

Fig. 1. Mechanism of Asp and Asn deamidation, isomerization, and racemisation.

of this reaction appears to be much slower than that of the succinimide pathway. Glutamine and glutamic acid residues are also capable of undergoing similar degradation reactions, but the rates of these reactions are much slower than at those of asparagine and aspartic acid residues (Won et al., 2004). The effect of neighbouring amino acid side chains in the context of protein structures upon succinimide formation is now fairly well understood from studies of synthetic peptides where there is much conformational flexibility. It is clear that the first step is the deprotonation of the attacking peptide-bond nitrogen to form a more nucleophilic anion. The acidity of the nitrogen atom depends for most residues on the electron-withdrawing power of the side-chain of the following residue (Radkiewicz et al., 2001). In general, the half-times of aspartyl and asparaginyl peptide degradation under physiological conditions (pH 7.4, 37 C) vary between about 1 and 1000 days (Brennan and Clarke, 1995). Asparagine residues form succinimides about ten times more rapidly than comparable aspartyl residues (Stephenson and Clarke, 1989). In many proteins, structural constraints on succinimide formation are imposed, and it is very difficult to predict whether a given site on a protein will be particularly susceptible or resistant to succinimide formation without knowledge of the three-dimensional structure (Clarke, 1987). It appears that the crucial factor is the flexibility of the polypeptide chain in the region of the aspartyl or asparaginyl residue so that the peptide-bond nitrogen can rotate around to attack the side chain carbonyl when it is transiently exposed. High-resolution and high mass accuracy Fourier transform mass spectrometry and electron capture dissociation have been used recently to map quantitatively deamidation (Cournoyer et al., 2007).

known as Maillard reaction (Baynes, 2001). The Maillard reaction involves reaction of amino groups on proteins with aldehydes and ketones to produce advanced glycation end-products, or AGEs (Baynes, 2001; Baynes et al., 1989). For example, the first intermediar of protein reaction with glucose is an Amadori rearrangement product known as fructoselysine. Glycation is a reversible reaction and the Amadori compound is not an age-dependent chemical modification of protein. In contrast, most AGE’s irreversibly accumulate with age (Table 2), particularly in longlived proteins (Biemel et al., 2002; Sell et al., 2005). The term’advanced’ refers to the fact that AGE’s arise through a series of reactive intermediates formed by rearrangement, dehydration, oxidation and fragmentation reactions of karbonyl contained compounds or its adducts to proteins (Table 3). Many different AGE’s have been detected in tissue proteins (Table 2). They are formed from a wide range of carbohydrates, including glucose, ascorbate, triose-phosphates or methylglyoxal. Because of the mutual intermediates generated from different carbonyl containing compounds none of the AGE’s provides unambiguous evidence of its origin. Most, but not all AGE’s that accumulate in proteins with age are, in fact, glycoxidation products, however, some can be formed from glucose without oxidation, e.g. CEL, the precursor methylglyoxal may be formed by oxidation of a glucose adduct to protein or by non-oxidative decomposition of glyceraldehyde-3-phosphate, an intermediate of glycolysis (Degenhardt et al., 1998).

Table 2 Major nonenzymatic protein modification that are known to accumulate with age Amino acid

Adducts

Crosslink products

Lys

Ne-(carboxymethyl) lysine (CML) Ne-(carboxyethyl) lysine (CEL)

Glucosepane, DOGDIC, MODIC, GODIC, GOLD and MOLD, pentosidine, vesperlysines, crosslines

Arg

Not known

Glucosepane, DOGDIC, MODIC, GODIC, argpyrimidine, pyrraline, and imidazolone derivatives of Arg

Table 3 Precursors and mechanisms of AGE formation Agent

Intermediates

Mechanism

Aldoses

a-Hydroxy carbonyl compounds a,b-Unsaturated carbonyl compounds Hydroxyalkenals

Dehydratation

Ketoses Ascorbate

3.1. AGE formation A process of cross-linking reaction of proteins with glucose or its metabolites that occurs with age is known as

249

1- and 3deoxyglucosamine Miscellaneous metabolic intermediates

Enediols Dicarbonyl compounds

Rearrangement Canizzaro reaction Michael addition Oxidative decomposition

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

250

Sugars, ascorbate or polyunsaturated fatty acids

Sugars

Proteins

ROS Glycolysis ROS

Schiff‘ base

Glyoxal

Serine

Proteins ROS

Amadori or Heyns adduct

ROS CML-proteins

Glycolaldehyde

Fig. 2. Pathways to formation of Ne-(carboxymethyl)lysine-CMLproteins.

Some compounds that are commonly described as AGE’s may not, in fact, be AGE’s (Baynes and Thorpe, 2000). Thus, CML (Table 2, Fig. 2) and CEL (Table 2) may be formed from both carbohydrates and polyunsaturated fatty acids (Pamplona et al., 2005; Pennathur et al., 2005). Proteins cross links such as GOLD and MOLD (Table 2) are also likely to be derived from lipids. When derived from lipids, these compounds should be termed advanced lipoxidation end-products (ALE) (Januszewski et al., 2003). Besides, Sell and Monnier (2004) have postulated conversion of arginine into ornithine by advanced glycation in senescent human collagen and lens crystallins. CML may occur from: (a) carbohydrates by autoxidation of the Amadori or Heyns adduct or via glyoxal formed by autoxidation of sugars or sugar derivatives, (b) glyoxal formed by autoxidation ascorbate (c) glyoxal formed by autoxidation of fatty acids or from (d) glycolaldehyde formed by autoxidation of serine or in glycolysis. The interplay between glycative and oxidative modification of proteins during aging is complex. Oxidative stress is involved in AGE formation, and AGE’s can induce oxidative stress.

4. Detection of age-related modifications 4.1. Mass spectrometric modification analysis Mass spectrometry is the method of choice for clarifying molecular details of age-related posttranslational modifications in nearly any type of biomolecules. Here we will focus on proteins. The direct detection of such modifications at distinct sites in individual proteins by mass spectrometry is not straightforward. The typical situation is characterized by complex mixtures of multiple redundant isoforms of proteins, which in the first place require efficient strategies of resolution to quantify differential biochemical species. There is no golden bullet approach; every sample requires an individual optimization of analytical methods

(Schrattenholz and Groebe, 2007). However in most cases, 2D-PAGE and subsequent MALDI-TOF and MS/MS mass spectrometry offer the best compromise (Hunzinger et al., 2006). The method is fast, well established, but most importantly, bioinformatic and statistical tools and methods have been developed to a fairly advanced stage. Advanced methods like high-resolution and high mass accuracy Fourier transform mass spectrometry and electron capture dissociation have the potential to focus on the quantification of discrete oxidative events at certain residues like Asp or Cys (Cournoyer et al., 2007; Zhao et al., 2006)). Ideally, it would be required to sequence candidate proteins in question, e.g., by quadrupole time of flight mass spectrometry (Q-TOF) and obtain direct proof of agerelated modifications of amino acid side chains. Indeed this approach has recently been applied in a variety of related proteomic investigations, with a focus on carbonylation of residues, like N-formyl-kynurenine formation, which is relatively stable and easily detectable by mass spectrometry (Poon et al., 2006a,b;Vaishnav et al., 2007b;Sultana et al., 2006; Hunzinger et al., 2006). However, in view of the complexity of protein mixtures in cell homogenates, as well as the large number of potential sites for modifications and a huge dynamic range, this undertaking requires more elaborate approaches. In the following, a procedure is described which may be applied after proteins of interest have been isolated (e.g., by 2D-PAGE), enzymatically fragmented, and identified by MALDI-TOF mass spectrometry and consecutive peptide mass fingerprinting (PMF). In typical age-related experiments, differential protein abundances of two or more conditions (e.g., juvenile and senescent) have been quantified by visualizing spots or peaks appropriately, and PMF identifications have been performed for all differential proteins. Frequently, there are multiple redundant PMF identifications, i.e., the very same protein is found at different locations on a gel or in different peaks of a chromatogram. (For simplicity, gel locations and chromatogram peaks will both be denoted ‘‘spots’’ in the following.) To clarify the question, whether a modification might quantitatively be differential across conditions, one evidently wishes to include all spots with the same PMF identification in a joint modification analysis. The exact assignment of age-related oxidative modifications by mass spectrometry is complicated by considerable ambiguity of standard data analysis tools like Mascot. I.e., spots of the same protein may exhibit different (sets of) PMF identifications and, vice versa, identifications possessing almost identical protein descriptions may turn out to show only few or no homologies.

4.1.1. Step 1. Selecting all spots with matching PMF identifications In the first step, the basic task consists in identifying all spots which are attributable to the same protein and which

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

may, e.g., be true if the proteins represent distinct subunits of a larger functional protein complex. For protein pairs related in this way, it would make no sense at all to be included in a joint modification analysis.

need to be included in the modification analysis of this protein. This task is complicated by a number of factors: • From one and the same mass list, current PMF search engines typically retrieve several different identifications with very comparable scores but different accession numbers. • Re-processing of the same protein spot frequently yields slightly different peptide mass lists which may also result in different sets of identifications. • If the same protein is located in different spots, some deviations in the amino acid backbone and/or in residues bound to some of the amino acids are highly likely. This leads to different peptide compositions of the pertinent enzymatic digests, to different peptide mass lists, and again to different sets of identifications.

All in all, it is not a trivial task to arrange groups of different but related accession numbers and associated sample spots which should be entered in a joint modification analysis. Manual selection of all homologous identifications along with corresponding spots in a set of experiments may be quite tedious if the number of redundant isoforms increases. Automatisation is a prerequisite for systematic large scale screening of masses associated with oxidative modifications from complex protein mixtures. A simple example of this situation is shown in Fig. 3, visualizing relations between accession numbers of proteins (top) and spot positions in which these accession numbers were identified (bottom) by straight lines. The following procedure is employed to define the ‘‘domains’’ of modification analyses. Each domain is composed of a set of accession numbers plus the protein spot positions in which at least one of the accession numbers was discovered. Starting with the first accession number, one searches for all spot locations in which this accession number was detected. Next, one finds all accession numbers which were also identified in one of the latter spot locations. For these newly added accession numbers one searches for pertinent spot locations again, and so on and so forth. After a finite number of cycles, no more additional accession numbers nor spot locations will be found. In the next cycle, one identifies an accession number that has not been assigned a domain yet and starts the above steps over again. The procedure is finished after all

In all of the above cases, the accession numbers returned by PMF are regularly not the same, typically with some overlaps, and their ranking is altered (in the case of identical accession numbers). Mostly, these identifications are closely related to each other and exhibit the same functionality. This observation is owing to the fact that the proteins exhibit a high degree of homology and that, hence, a number of enzymatic fragment peptides are identical. On this background, diverging identifications are likely to be artefacts brought about by some blur and/or small dissimilarities in the respective mass spectra – which is destined to create difficulties regarding potentially modified aberrant peaks. • On the other hand, homologies also may be entirely absent even though the respective descriptions are suggestive of a close similarity between the proteins. This

gi|85074641

ps|27916

251

gi|88176383

ps|24700

ps|20945

B7 C11 E4

L2

L3

L4

L5

N2

N3

N4

N5

O2

P2

P3

P4

P5

D1

E6

I19 I21 L16 L21

Fig. 3. Computational strategy for rapidly finding peptides with posttranslational modifications from MALDI-TOF peptide mass fingerprints of highly homologous redundant protein spots. Graph of PMF identifications for ATP synthase b-subunit (top) and corresponding protein spot positions (bottom) in Podospora anserina. Identifications are connected to spots in which they were detected by straight lines. gij. . . and psj. . . are protein accession numbers all of which are described as ATP synthase b-subunit. (psj. . . are internal notations for proteins that have not been incorporated in the NCBI database yet.) L2, L3 . . . denote the sample protein spot positions from which PMF identifications were obtained. A joint modification analysis is performed for all connected subgraphs of maximum order (i.e. for the three subgraphs (i) proteins gij85074641, gij88176383, psj27916 and spots L2, L3, L4, L5, N2, N3, N4, N5, O2, P2, P3, P4, P5; (ii) protein psj24700 and spots B7, C11, E4; (iii) protein psj20945 and spots D1, E6, I19, I21, L16, L21. For more details see text.).

252

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

accession numbers and all spot locations have been assigned a domain for modification analysis. In this way, the totality of accession numbers and spot locations gets partitioned in disjoint subsets. Using the terminology of graph theory, the presentation of accession numbers and corresponding protein spot locations according to Fig. 3 may be interpreted as a graph, and each domain of a modification analysis is a connected subgraph of maximum order. The rationale for this approach is as follows: For two identifications to be detected by PMF with similarly high scores from the same mass list, the peptide masses resulting from digests of the two proteins need be in good agreement, i.e., many of the peptide masses from the two digests are identical and their differences are not larger than some ppm. Given the fact that both proteins serve the same function in the same or closely related species, there is a good chance that the proteins are modifications of one another and that therefore many of the underlying peptides are also identical. Even though this approach often yields the desired results very reliably, there is no direct proof of it. Therefore, after automatic grouping a reasonable degree of homology between all identifications assigned to one and the same group should be verified manually – which is an easy and rapid task using a suitable software package like jemboss (download from http://emboss.sourceforge.net/). In practice, one feeds a list of all accession numbers for protein identifications suspicious of homology into a program which sequentially checks through the identifications obtained from all spots and controls whether they include one of the identifications in the list. In this way, for each accession number one collects all spots containing identical identifications (corresponding to the connecting lines in Fig. 3). From this information, the disjoint groups of accession numbers and spot IDs are computed which in turn are suspicious of posttranslational modifications. Here the resolution of the separation method is crucial, insufficiently resolved mixtures of proteins cause complications (Schrattenholz and Groebe, 2007). In this case, one needs to perform separate grouping runs none of which must contain more than one of the unrelated identifications for this spot. For each group of identifications and corresponding spot locations, and for each protein in the group, the masses of all potential modified fragment peptides need to be matched against the actually measured mass lists. This takes us to: 4.1.2. Step 2. Generating lists of masses of all potentially modified peptides In this step, one computes all peptide masses which may occur if the modifications under study were present at one or more sites in the amino acid sequence of any one of the identified proteins. To that end, one performs a theoretical digest of the protein with the enzyme actually used in the wet experiment, resulting in a set of theoretical peptide

amino acid sequences which are assembled in a list of unmodified peptides. In this course of action, a realistic number of missed cleavages need to be accounted for. Next, modifications are applied to all of the peptides. During this in silico process, each peptide is added to the list of unprocessed peptides again after having been modified in order to potentially undergo further modifications. Thus, in an iterative procedure one generates all possible combinations of all modifications under study for each peptide. After the iteration has completed, all peptides containing one of the modifications of interest are selected, and their molecular masses are stored for the actual comparison with measured masses (step 3 of the analysis). If PMF analysis resulted in two or more identifications from the same mass list with comparable scores and similar functionality, a sound decision as to which protein really underlies the spectrum is sometimes not possible. Since the current approach to modification analysis is based on the prior knowledge of the underlying protein it consequently needs to be applied to each single one of the competing identifications – multiplying computational and manpower requirements. As an alternative, one may perform the described theoretical digests and modifications for all PMF identifications as described above and then proceed with the list of all modified peptides arising from anyone of these digests. In cases of highly homologous PMF identifications this will only moderately increase the total number of modified peptides over and above the number obtained from only one protein but drastically reduce the required effort. Usually, only those modifications are of interest that exhibit differential abundances in the framework of the actual experimental setup (e.g., modifications that occur more frequently in older versus younger individuals). To discern these modifications from others, one also needs to detect the presence of the corresponding unmodified peptide masses and to compare the quantitative relations of modified and unmodified peptides. Consequently, when generating the sets of potential modifications one needs to also store the associated unmodified peptide masses along with each modified peptide.1 4.1.3. Step 3. Detecting candidate masses for modifications in measured mass lists The third step is the identification of potentially modified peptide masses in the actually measured monoisotopic mass lists (i.e non-canonical masses). For each of the above domains of individual modification analyses and for each protein spot location contained in the 1

For a peptide in which neither missed cleavages nor multiple modifications emerge, the definition of the base peptide is obvious. In a miscleaved and/or multiply modified peptide, however, this concept may lead to a vast host of different base peptides. In our routine, it has proven practical and adequate to restrict the sets of base peptides to the ones containing no modification at all plus the ones immediately ‘‘preceding’’ the modification which is currently being analyzed.

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

domain, one searches for modified peptides by scanning the measured mass list for peptide masses which are attributable to one of the modifications under study to one of the identified proteins. Each experimental mass meeting certain inclusion criteria (e.g., 650 ppm deviation from a theoretically modified peptide mass) is accepted as a candidate. Due to the limited accuracy of measurement, one frequently finds a multitude of measured masses that are consistent with some modified peptide computed from one of the identified protein sequences. This is particularly true if the latter are very large, the modified amino acids are prevalent, and missed cleavages as well as cleavage preventing or cleavage enabling modifications are considered. In order to reduce the number of peptide masses to be entered in the concluding MS/MS analysis, one requires, as an additional exclusion criterion, that also the corresponding unmodified masses be present in at least one of the protein spots. Typical modification analyses include more than one sample with same identifications, originating, e.g., from different spots on a 2D-PAGE gel or from matching spots in two or more gels representing different experimental conditions (e.g., juvenile and senescent individuals). Only differential abundances of modifications, based on biological dynamics are of interest. This is another reason why it is required that not only the mass of the modified peptide is contained in the experimental measurements but that at least one of the corresponding unmodified masses can also be retrieved from the same set of experiments. 4.1.4. Step 4. Confirming modifications in candidate masses by MS/MS Definite proof of a modification is ultimately achieved by sequencing its mass peak in a Q-TOF mass spectrometer or by lifting the respective peak in a MALDI-TOF machine. While all of the preceding steps can be fully automated, it is advisable to inspect the relevant mass ranges of spectra and to perform the final selection of the candidates for sequencing manually. In this way, typical peak patterns that are highly indicative of a particular modification (like in the case of the N-formyl-kynurenine oxidation of tryptophan; see below) may reliably be assigned by additional biophysical parameters. Software tools have been developed to automatically load relevant spectra in a viewer program and tag the mass peaks of interest for further bioinformatic and MS processing. The selected masses are subsequently submitted to automatic MS/MS sequencing. 4.1.5. Step 5. Quantitating modifications in biological conditions under study Last but not least, quantitative relations between modified and unmodified sites are crucial for biological interpretation. The question of biological significance can be decided by comparing normalized intensities of mass spec-

253

trometric peaks of modified and unmodified peptides obtained from the respective spots/experimental conditions. For normalization, one may, e.g., employ the average peak intensities of all masses that (i) belong to the particular protein and that (ii) can be detected and quantified in the spectrum of every spot/condition to be compared. Taken together the combination of mass spectrometric methods which are fast, highly automated and well supported by broadly available database and software tools (like Mascot) with appropriate differential strategies (on biological and analytical levels) should be used to reliably identify and quantify potential age-related modifications from complex protein mixtures. Once non-canonical mass peaks suspicious for age-related chemical alterations have been assigned, the molecular detail will be clarified by MS-based sequencing. 4.2. Carbonylation, detection of N-formyl kynurenine in mass spectrometry Among the various types of carbonylation (see also Table 1) which are not too common and ambiguous for a straightforward downstream analysis, the dioxygenation of tryptophan residues by ROS is a prime candidate. Tryptophan is a direct target of ROS and thus an early and direct event in the chain of oxidative damage, unlike downstream events like, e.g., glycation. The resulting posttranslational modification, a carbonyl named N-formyl kynurenine can be easily detected in mass spectra and even quantified by relating the very characteristic mass increments of 4, 16 and 32 of modified peptides to the corresponding peaks of unmodified peptides. In Fig. 4 an age-related example shows the corresponding spectra of aconitase-2 (gi27806769) (Hunzinger et al., 2006). 4.3. Age-related protein nitration Elevated levels of protein tyrosine nitration have been found in various age-related pathologies and methods have been developed for selective enrichment of nitrotyrosinecontaining peptides from complex proteome samples and subsequent analysis with LC-MS/MS: 3-nitrotyrosine is used as a stable marker of protein oxidative damage. One procedure uses selective conversion of nitrotyrosine to a derivative with a free sulfhydryl group followed by high efficiency enrichment of sulfhydryl-containing peptides with thiopropyl sepharose beads (Zhang et al., 2007). Other approaches use nano HPLC tandem mass spectrometry (Hong et al., 2007; Wang et al., 2007) or selective isolation of nitrated proteins using immunoprecipitation, followed by SDS–PAGE and nano-electrospray-MS/MS analysis (Gokulrangan et al., 2007) or just MALDI-TOF mass spectrometry (Fugere et al., 2006; Ahmed et al., 2005). Interpretation of such results in terms of functional consequences of protein nitration, requires the same basic considerations regarding solution and differential quanti-

254

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

Fig. 4. Example of oxidatively modified peptides found following the strategy described above (Fig. 3); here different ratios of oxidation can be quantified according to peak intensities. Representative MALDI spectrum showing the unmodified and N-formyl kynurenine modified tryptic peptides 371–378 and 657–671 of aconitase-2. The characteristic patterns of masses (+4, +16, and +32; solid arrows) and the corresponding signal intensities of the unmodified peptide 371–378 (dashed arrow) and the modified one, suggest that tryptophan 373 is predominantly oxidized in one condition. The opposite is true for another spot separated at a different position in a 2D gel of the identical sample, yielding a peptide 657–671 of aconitase-2; leading to the conclusion that tryptophan 657 is predominantly non-modified in that particular protein isoform.

fication discussed above (Schrattenholz and Groebe, 2007) and moreover would probably be most fruitful when related to concomitant oxidative molecular events or even the inclusion of ROS-inducible biological parameters like nitric oxide synthase isoforms, age-receptor and superoxide dismutase expressions (Freixes et al., 2006).

ski et al., 2005), which subsequently can be submitted to independent MS-based confirmation. As in the case of carbonylation, similar Western blots technologies can also be used to detect age-dependent nitrosylation/nitration of amino acid residues. 5. Conclusion/discussion

4.4. Immunological detection of carbonylation: Oxyblots An alternative to mass spectrometry, which is frequently used, is quantification of protein oxidation by immunoblotting, i.e. the Oxyblot technique, like described e.g. in: (Bulteau et al., 2001, 2005; Korolainen et al., 2005, 2007; Aguilaniu et al., 2003). There are commercially available kits on the market, essentially carbonyls are derivatized by reaction with 2,4 dinitrophenylhydrazine (DNPH) to 2,4-dinitrophenylhydrazone (DNP-hydrazone). After separation by gel electrophoresis and subsequent Western blotting, carbonylated proteins are detected by antibodies specific to the DNP moiety of proteins (Barreiro et al., 2007; Kriebardis et al., 2006, 2007; Scheckhuber et al., 2007). Two-dimensional multiplexed oxyblotting has been suggested and applied for the study of both the concentrations and carbonylation of disease-related protein oxidation (Korolainen et al., 2007), or differential redoxproteomics of age-related samples (Vaishnav et al., 2007a). In a number of cases the combination of immunoblotting and mass spectrometry has been particularly useful, with the immunological method serving as a first filter for tracing carbonylated posttranslational modifications (Kan-

Taken together, a set of useful and highly complementary techniques have been developed to unambiguously identify age-related posttranslational modifications in considerable molecular detail. Some, like N-formyl-kyurenine or 3-nitrotyrosine, are stable enough for a variety of mass spectrometry-based approaches, and can alternatively be detected by immunological methods for oxidative protein carbonylation or nitration. Mass spectrometry and immunological approaches should be supplemented with appropriate biological levels of analysis, like enzymatic or receptor activities or quantification of kinetic changes of functional parameters like calcium or metabolite concentrations. In any case the merely phenomenological description of certain age-related molecular changes in some appropriate biological sample will most probably remain anecdotic, unless a reliable differential quantification can be performed, preferentially correlated to the kinetics of biochemical signatures of ageing. More than any other biological phenomenon, ageing occurs in a truly systems biology sense (Kriete et al., 2006; Raghothama et al., 2005; Franceschi et al., 2007). Oxidative damage certainly

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

is a key event (Harman, 1956), but not alone: there are other components like neuronal vulnerability (Robert et al., 2007): crocodiles and turtles regenerate neurons in contrast to mammals and thus potentially can live longer (Lutz et al., 2003; Font et al., 2001;Castanet, 1994), some sponges don’t even have brains at all and for this and other reasons are potentially immortal (Leys et al., 2005). Hormonal regulation of anti-oxidative damage repair systems, by e.g. dehydroepiandosterone in mammals (Arlt, 2004; Anisimov, 2006), as well as genetic and epigenetic factors (Reznick, 2005; Weinreb et al., 2007), e.g. telomerase activity (Sampedro et al., 2007; Blasco, 2007), diet, unsaturated membrane lipids (Hulbert, 2006) and antioxidants like resveratrol (Blagosklonny, 2007; Chen and Guarente, 2007; Holme and Pervaiz, 2007), which are quite differently managed across species all contribute to life spans of organisms and contribute to controversies and complexity. Exact knowledge of the quantitative kinetics of nonenzymatic posttranslational modifications in age-related models and their systematic correlation to above mentioned biological effectors can be expected to reveal novel information about the interplay and mutual regulation of biochemical pathways contributing to life spans in a sense of generally conserved mechanisms of ageing. References Aguilaniu, H., Gustafsson, L., Rigoulet, M., Nystrom, T., 2003. Asymmetric inheritance of oxidatively damaged proteins during cytokinesis. Science 299, 1751–1753. Ahmed, N., Ahmed, U., Thornalley, P.J., Hager, K., Fleischer, G., Munch, G., 2005. Protein glycation, oxidation and nitration adduct residues and free adducts of cerebrospinal fluid in Alzheimer’s disease and link to cognitive impairment. J. Neurochem. 92, 255–263. Anisimov, V.N., 2006. Premature ageing prevention: limitations and perspectives of pharmacological interventions. Curr. Drug Targets 7, 1485–1503. Arlt, W., 2004. Dehydroepiandrosterone and ageing. Best Pract. Res. Clin. Endocrinol. Metab. 18, 363–380. Bailey, A.J., 2001. Molecular mechanisms of ageing in connective tissues. Mech. Ageing Dev. 122, 735–755. Balasubramanian, D., Kanwar, R., 2002. Molecular pathology of dityrosine cross-links in proteins: structural and functional analysis of four proteins. Mol. Cell Biochem. 234-235, 27–38. Barreiro, E., Nowinski, A., Gea, J., Sliwinski, P. 2007. Oxidative Stress: In The External Intercostals of Obstructive Sleep Apnea Patients. Thorax June 15; [Epub ahead of print]. Baynes, J.W., 2000. From life to death – the struggle between chemistry and biology during aging: the Maillard reaction as an amplifier of genomic damage. Biogerontology 1, 235–246. Baynes, J.W., 2001. The role of AGEs in aging: causation or correlation. Exp. Gerontol. 36, 1527–1537. Baynes, J.W., 2002. The Maillard hypothesis on aging: time to focus on DNA. Ann. N.Y. Acad. Sci. 959, 360–367. Baynes, J.W., Thorpe, S.R., 2000. Glycoxidation and lipoxidation in atherogenesis. Free Radic. Biol. Med. 28, 1708–1716. Baynes, J.W., Watkins, N.G., Fisher, C.I., Hull, C.J., Patrick, J.S., Ahmed, M.U., Dunn, J.A., Thorpe, S.R., 1989. The Amadori product on protein: structure and reactions. Prog. Clin. Biol. Res. 304, 43–67. Beckman, K.B., Ames, B.N., 1998. The free radical theory of aging matures. Physiol. Rev. 78, 547–581.

255

Biemel, K.M., Friedl, D.A., Lederer, M.O., 2002. Identification and quantification of major maillard cross-links in human serum albumin and lens protein. Evidence for glucosepane as the dominant compound. J. Biol. Chem. 277 (28), 24907–24915. Blagosklonny, M.V., 2007. An anti-aging drug today: from senescencepromoting genes to anti-aging pill. Drug Discov. Today 12, 218– 224. Blasco, M.A., 2007. The epigenetic regulation of mammalian telomeres. Nat. Rev. Genet. 8, 299–309. Brennan, T.V., Clarke, S., 1995. Effect of adjacent histidine and cysteine residues on the spontaneous degradation of asparaginyl- and aspartylcontaining peptides. Int. J. Pept. Protein Res. 45, 547–553. Bulteau, A.L., Lundberg, K.C., Ikeda-Saito, M., Isaya, G., Szweda, L.I., 2005. Reversible redox-dependent modulation of mitochondrial aconitase and proteolytic activity during in vivo cardiac ischemia/reperfusion. Proc. Natl. Acad. Sci. USA 102, 5987–5991. Bulteau, A.L., Verbeke, P., Petropoulos, I., Chaffotte, A.F., Friguet, B., 2001. Proteasome inhibition in glyoxal-treated fibroblasts and resistance of glycated glucose-6-phosphate dehydrogenase to 20 S proteasome degradation in vitro. J. Biol. Chem. 276, 45662–45668. Castanet, J., 1994. Age estimation and longevity in reptiles. Gerontology 40, 174–192. Chao, C.C., Ma, Y.S., Stadtman, E.R., 1997. Modification of protein surface hydrophobicity and methionine oxidation by oxidative systems. Proc. Natl. Acad. Sci. USA 94, 2969–2974. Chen, D., Guarente, L., 2007. SIR2: a potential target for calorie restriction mimetics. Trends Mol. Med. 13, 64–71. Clarke, S., 1987. Propensity for spontaneous succinimide formation from aspartyl and asparaginyl residues in cellular proteins. Int. J. Pept. Protein Res. 30, 808–821. Clarke, S., 2003. Aging as war between chemical and biochemical processes: protein methylation and the recognition of age-damaged proteins for repair. Ageing Res. Rev. 2, 263–285. Cournoyer, J.J., Lin, C., Bowman, M.J., O’Connor, P.B., 2007. Quantitating the relative abundance of isoaspartyl residues in deamidated proteins by electron capture dissociation. J. Am. Soc. Mass Spectrom. 18 (1), 48–56. Degenhardt, T.P., Thorpe, S.R., Baynes, J.W., 1998. Chemical modification of proteins by methylglyoxal. Cell Mol. Biol. 44, 1139–1145. Dehart, M.P., Anderson, B.D., 2007. The role of the cyclic imide in alternate degradation pathways for asparagine-containing peptides and proteins. J. Pharm. Sci. 96 (10), 2667–2685. Font, E., Desfilis, E., Perez-Canellas, M.M., Garcia-Verdugo, J.M., 2001. Neurogenesis and neuronal regeneration in the adult reptilian brain. Brain Behav. Evol. 58, 276–295. Franceschi, C., Capri, M., Monti, D., Giunta, S., Olivieri, F., Sevini, F., Panourgia, M.P., Invidia, L., Celani, L., Scurti, M., Cevenini, E., Castellani, G.C., Salvioli, S., 2007. Inflammaging and anti-inflammaging: a systemic perspective on aging and longevity emerged from studies in humans. Mech. Ageing Dev. 128, 92–105. Freixes, M., Rodriguez, A., Dalfo, E., Ferrer, I., 2006. Oxidation, glycoxidation, lipoxidation, nitration, and responses to oxidative stress in the cerebral cortex in Creutzfeldt-Jakob disease. Neurobiol. Aging 27, 1807–1815. Fugere, N.A., Ferrington, D.A., Thompson, L.V., 2006. Protein nitration with aging in the rat semimembranosus and soleus muscles. J. Gerontol. A. Biol. Sci. Med. Sci. 61, 806–812. Gokulrangan, G., Zaidi, A., Michaelis, M.L., Schoneich, C., 2007. Proteomic analysis of protein nitration in rat cerebellum: effect of biological aging. J. Neurochem. 100, 1494–1504. Harman, D., 1956. Aging: a theory based on free radical and radiation chemistry. J. Gerontol. 11, 298–300. Hayflick, L., 2007. Biological aging is no longer an unsolved problem. Ann. N.Y. Acad. Sci. 1100, 1–13. Holme, A.L., Pervaiz, S., 2007. Resveratrol in cell fate decisions. J. Bioenerg. Biomembr. 39, 59–63. Hong, S.J., Gokulrangan, G., Schoneich, C., 2007. Proteomic analysis of age dependent nitration of rat cardiac proteins by solution isoelectric

256

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257

focusing coupled to nanoHPLC tandem mass spectrometry. Exp. Gerontol. 42, 639–651. Hulbert, A.J., 2006. The links between membrane composition, metabolic rate and lifespan. Comp. Biochem. Physiol. A Mol. Integr. Physiol. Hunzinger, C., Wozny, W., Schwall, G., Poznanovic, S., Stegmann, W., Zengerling, H., Scho¨pf, R., Groebe, K., Cahill, M.A., Osiewacz, H.D., Ja¨gemann, N., Bloch, M., Dencher, N.A., Krause, F., Schrattenholz, A., 2006. Comparative profiling of the mammalian mitochondrial proteome: multiple aconitase-2 isoforms including N-formylkynurenine modifications as part of a protein biomarker signature for reactive oxidative species. J. Proteome Res. 5, 625–633. Januszewski, A.S., Alderson, N.L., Metz, T.O., Thorpe, S.R., Baynes, J.W., 2003. Role of lipids in chemical modification of proteins and development of complications in diabetes. Biochem. Soc. Trans. 31, 1413–1416. Kanski, J., Behring, A., Pelling, J., Schoneich, C., 2005. Proteomic identification of 3-nitrotyrosine-containing rat cardiac proteins: effects of biological aging. Am. J. Physiol. Heart Circ. Physiol. 288, H371– H381. Korlimbinis, A., Truscott, R.J., 2006. Identification of 3-hydroxykynurenine bound to proteins in the human lens. A possible role in agerelated nuclear cataract. Biochemistry 45 (6), 1950–1960. Korolainen, M.A., Auriola, S., Nyman, T.A., Alafuzoff, I., Pirttila, T., 2005. Proteomic analysis of glial fibrillary acidic protein in Alzheimer’s disease and aging brain. Neurobiol. Dis. 20, 858–870. Korolainen, M.A., Nyman, T.A., Nyyssonen, P., Hartikainen, E.S., Pirttila, T., 2007. Multiplexed proteomic analysis of oxidation and concentrations of cerebrospinal fluid proteins in Alzheimer disease. Clin. Chem. 53, 657–665. Kriebardis, A.G., Antonelou, M.H., Stamoulis, K.E., Economou-Petersen, E., Margaritis, L.H., Papassideri, I.S., 2006. Membrane protein carbonylation in non-leukodepleted CPDA-preserved red blood cells. Blood Cells Mol. Dis. 36, 279–282. Kriebardis, A.G., Antonelou, M.H., Stamoulis, K.E., Economou-Petersen, E., Margaritis, L.H., Papassideri, I.S., 2007. Progressive oxidation of cytoskeletal proteins and accumulation of denatured hemoglobin in stored red cells. J. Cell Mol. Med. 11, 148–155. Kriete, A., Sokhansanj, B.A., Coppock, D.L., West, G.B., 2006. Systems approaches to the networks of aging. Ageing Res. Rev. 5, 434–448. Levine, R.L., 2002. Carbonyl modified proteins in cellular regulation, aging, and disease. Free Radic. Biol. Med. 32, 790–796. Leys, S.P., Rohksar, D.S., Degnan, B.M., 2005. Sponges. Curr. Biol. 15, R114–R115. Lutz, P.L., Prentice, H.M., Milton, S.L., 2003. Is turtle longevity linked to enhanced mechanisms for surviving brain anoxia and reoxygenation? Exp. Gerontol. 38, 797–800. McCudden, C.R., Kraus, V.B., 2006. Biochemistry of amino acid racemization and clinical application to musculoskeletal disease. Clin. Biochem. 39, 1112–1130. Pamplona, R., Dalfo, E., Ayala, V., Bellmunt, M.J., Prat, J., Ferrer, I., Portero-Otin, M., 2005. Proteins in human brain cortex are modified by oxidation, glycoxidation, and lipoxidation. Effects of Alzheimer disease and identification of lipoxidation targets. J. Biol. Chem. 280 (22), 21522–32150. Pennathur, S., Ido, Y., Heller, J.I., Byun, J., Danda, R., Pergola, P., Williamson, J.R., Heinecke, J.W., 2005. Reactive carbonyls and polyunsaturated fatty acids produce a hydroxyl radical-like species: a potential pathway for oxidative damage of retinal proteins in diabetes. J. Biol. Chem. 280 (24), 22706–22714. Poon, H.F., Calabrese, V., Calvani, M., Butterfield, D.A., 2006a. Proteomics analyses of specific protein oxidation and protein expression in aged rat brain and its modulation by L-acetylcarnitine: insights into the mechanisms of action of this proposed therapeutic agent for CNS disorders associated with oxidative stress. Antioxid Redox. Signal. 8, 381–394. Poon, H.F., Vaishnav, R.A., Getchell, T.V., Getchell, M.L., Butterfield, D.A., 2006b. Quantitative proteomics analysis of differential protein

expression and oxidative modification of specific proteins in the brains of old mice. Neurobiol. Aging 27, 1010–1019. Radkiewicz, J.L., Zipse, H., Clarke, S., Houk, K.N., 2001. Neighboring side chain effects on asparaginyl and aspartyl degradation: an ab initio study of the relationship between peptide conformation and backbone NH acidity. J. Am. Chem. Soc. 123, 3499–3506. Raghothama, C., Harsha, H.C., Prasad, C.K., Pandey, A., 2005. Bioinformatics and proteomics approaches for aging research. Biogerontology 6, 227–232. Reznick, D.N., 2005. The genetic basis of aging: an evolutionary biologist’s perspective. Sci. Aging Knowledge Environ. 11, pe7. Robert, K.A., Brunet-Rossinni, A., Bronikowski, A.M., 2007. Testing the ‘free radical theory of aging’ hypothesis: physiological differences in long-lived and short-lived colubrid snakes. Aging Cell 6, 395–404. Robinson, N.E., 2002. Protein deamidation. Proc. Natl. Acad. Sci. USA 99, 5283–5288. Robinson, N.E., Robinson, A.B., 2001. Deamidation of human proteins. Proc. Natl. Acad. Sci. USA 98, 12409–12413. Rothstein, M., 1985. The alteration of enzymes in aging animals. Basic Life Sci. 35, 193–204. Sampedro, C.F., Cano, S.G., Sampedro, S.F., 2007. Telomerase and telomere dynamics in ageing and cancer: current status and future directions. Clin. Transl. Oncol. 9, 145–154. Scheckhuber, C.Q., Erjavec, N., Tinazli, A., Hamann, A., Nystrom, T., Osiewacz, H.D., 2007. Reducing mitochondrial fission results in increased life span and fitness of two fungal ageing models. Nat. Cell Biol. 9, 99–105. Schrattenholz, A., Groebe, K., 2007. What does it need to be a biomarker? Relationships between resolution, differential quantification and statistical validation of protein surrogate biomarkers. Electrophoresis. Sell, D.R., Monnier, V.M., 2004. Conversion of arginine into ornithine by advanced glycation in senescent human collagen and lens crystallins. J. Biol. Chem. 279 (52), 54173–54184. Sell, D.R., Biemel, K.M., Reihl, O., Lederer, M.O., Strauch, C.M., Monnier, V.M., 2005. Glucosepane is a major protein cross-link of the senescent human extracellular matrix. Relationship with diabetes. J. Biol. Chem. 280 (13), 12310–12315. Squier, T.C., Bigelow, D.J., 2000. Protein oxidation and age-dependent alterations in calcium homeostasis. Front Biosci. 5, D504–D526. Stadtman, E.R., 2001. Protein oxidation in aging and age-related diseases. Ann. N.Y. Acad. Sci. 928, 22–38. Stadtman, E.R., 2004. Role of oxidant species in aging. Curr. Med. Chem. 11, 1105–1112. Stadtman, E.R., 2006. Protein oxidation and aging. Free Radic. Res. 40, 1250–1258. Stadtman, E.R., Van, R.H., Richardson, A., Wehr, N.B., Levine, R.L., 2005. Methionine oxidation and aging. Biochim. Biophys. Acta 1703, 135–140. Stephenson, R.C., Clarke, S., 1989. Succinimide formation from aspartyl and asparaginyl peptides as a model for the spontaneous degradation of proteins. J. Biol. Chem. 264, 6164–6170. Sultana, R., Boyd-Kimball, D., Poon, H.F., Cai, J., Pierce, W.M., Klein, J.B., Merchant, M., Markesbery, W.R., Butterfield, D.A., 2006. Redox proteomics identification of oxidized proteins in Alzheimer’s disease hippocampus and cerebellum: an approach to understand pathological and biochemical alterations in AD. Neurobiol. Aging 27, 1564–1576. Vaishnav, R.A., Getchell, M.L., Poon, H.F., Barnett, K.R., Hunter, S.A., Pierce, W.M., Klein, J.B., Butterfield, D.A., Getchell, T.V., 2007b. Oxidative stress in the aging murine olfactory bulb: redox proteomics and cellular localization. J. Neurosci. Res. 85, 373–385. Vaishnav, R.A., Getchell, M.L., Poon, H.F., Barnett, K.R., Hunter, S.A., Pierce, W.M., Klein, J.B., Butterfield, D.A., Getchell, T.V., 2007a. Oxidative stress in the aging murine olfactory bulb: redox proteomics and cellular localization. J. Neurosci. Res. 85, 373–385. van der Vliet, Eiserich, J.P., O’Neill, C.A., Halliwell, B., Cross, C.E., 1995. Tyrosine modification by reactive nitrogen species: a closer look. Arch. Biochem. Biophys. 319, 341–349.

V. Sˇosˇkic´ et al. / Experimental Gerontology 43 (2008) 247–257 Wang, Z., Paik, D.C., Dillon, J.P., Gaillard, E.R., 2007. Tyrosine nitration site specificity identified by LC/MS in nitrite-modified collagen type IV. Exp. Mol. Med. 39, 74–83. Weinreb, O., Drigues, N., Sagi, Y., Reznick, A.Z., Amit, T., Youdim, M.B., 2007. The application of proteomics and genomics to the study of age-related neurodegeneration and neuroprotection. Antioxid. Redox Signal.. Wells-Knecht, M.C., Lyons, T.J., McCance, D.R., Thorpe, S.R., Baynes, J.W., 1997. Age-dependent increase in ortho-tyrosine and methionine sulfoxide in human skin collagen is not accelerated in diabetes. Evidence against a generalized increase in oxidative stress in diabetes. J. Clin. Invest. 100, 839–846.

257

Won, J.I., Meagher, R.J., Barron, A.E., 2004. Characterization of glutamine deamidation in a long, repetitive protein polymer via bioconjugate capillary electrophoresis. Biomacromolecules 5, 618–627. Zhang, Q., Qian, W.J., Knyushko, T.V., Clauss, T.R., Purvine, S.O., Moore, R.J., Sacksteder, C.A., Chin, M.H., Smith, D.J., Camp, D.G., Bigelow, D.J., Smith, R.D., 2007. A method for selective enrichment and analysis of nitrotyrosine-containing peptides in complex proteome samples. J. Proteome Res. 6, 2257–2268. Zhao, C., Sethuraman, M., Clavreul, N., Kaur, P., Cohen, R.A., O’Connor, P.B., 2006. Detailed map of oxidative post-translational modifications of human p21ras using Fourier transform mass spectrometry. Anal. Chem. 78 (14), 5134–5142.