fragmentation behavior of amadori-peptides obtained by non-enzymatic glycosylation of lysine...

66

4

Research ArticleReceived: 23 February 2010 Accepted: 22 April 2010 Published online in Wiley Interscience: 17 May 2010

(www.interscience.com) DOI 10.1002/jms.1758

Fragmentation behavior of Amadori-peptidesobtained by non-enzymatic glycosylationof lysine residues with ADP-ribose in tandemmass spectrometryMaria Fedorova, Andrej Frolov† and Ralf Hoffmann∗

Mono- and poly-adenosine diphosphate (ADP)-ribosylation are common post-translational modifications incorporated bysequence-specific enzymes at, predominantly, arginine, asparagine, glutamic acid or aspartic acid residues, whereasnon-enzymatic ADP-ribosylation (glycation) modifies lysine and cysteine residues. These glycated proteins and peptides(Amadori-compounds) are commonly found in organisms, but have so far not been investigated to any great degree.In this study, we have analyzed their fragmentation characteristics using different mass spectrometry (MS) techniques.In matrix-assisted laser desorption/ionization (MALDI)-MS, the ADP-ribosyl group was cleaved, almost completely, at thepyrophosphate bond by in-source decay. In contrast, this cleavage was very weak in electrospray ionization (ESI)-MS. Thesame fragmentation site also dominated the MALDI-PSD (post-source decay) and ESI-CID (collision-induced dissociation) massspectra. The remaining phospho-ribosyl group (formed by the loss of adenosine monophosphate) was stable, providing a directand reliable identification of the modification site via the b- and y-ion series. Cleavage of the ADP-ribose pyrophosphate bondunder CID conditions gives access to both neutral loss (347.10 u) and precursor-ion scans (m/z 348.08), and thereby permitsthe identification of ADP-ribosylated peptides in complex mixtures with high sensitivity and specificity. With electron transferdissociation (ETD), the ADP-ribosyl group was stable, providing ADP-ribosylated c- and z-ions, and thus allowing reliablesequence analyses. Copyright c© 2010 John Wiley & Sons, Ltd.

Keywords: ADP-ribosylation; collision-induced dissociation; electron transfer dissociation; electrospray ionization; glycation; matrix-assisted laser desorption/ionization

Introduction

Adenosine diphosphate (ADP)-ribosylation is a covalentmodification of proteins and nucleic acids, which has been rec-ognized since the early 1960s.[1] It plays important roles in inter-and intracellular signaling, transcription, DNA reparation, cell cycleregulation, apoptosis and necrosis.[2] ADP-ribosyltransferases (AD-PRTs) transfer the ADP-ribose group from β-NAD+ to arginine (lessoften to asparagine, glutamic acid or aspartic acid) residues in spe-cific protein motifs,[3 – 6] yielding mono- or poly-ADP-ribosylatedproteins. Initially, mono-ADP-ribosyltransferases (MARTs) were de-scribed as members of the ADP-ribosylating exotoxin family inbacteria[7] that targeted mostly nucleotide-binding eukaryoticproteins (e.g. EF-2, heteromeric GTP-binding proteins and actin) inorder to disturb the physiological cellular processes.[7,8] Recently,however, orthologous genes encoding ectogenic or endogenicMARTs were also characterized in mammals.[9]

Besides this enzymatic reaction, ADP-ribose can also react di-rectly with the ε-amino group of lysine[10] and the thiol-groupof cysteine residues,[11] though not with arginine residues.[12]

This lysine-directed, non-enzymatic mono-ADP-ribosylation be-longs to protein glycation, which describes the reaction ofcarbonyl functions in aldoses and ketoses with free aminogroups yielding Amadori-[13] or Heyns-compounds,[14] respec-tively. Intracellular ADP-ribose is produced by the degradationof poly-ADP-ribose in the nucleus[15] and in other cell compart-ments by the turnover of NAD+ and cyclic-ADP-ribose[16] and

mono-ADP-ribosylhydrolase activity.[17] Thus, enzymatic and non-enzymatic ADP-ribosylation can occur at different positionsin parallel,[12] especially in long-lived basic proteins like hi-stones. Although ADP-ribosylation has been characterized forseveral proteins, the underlying mechanisms, especially of non-enzymatic ADP-ribosylation, have not been well investigated.Moreover, sensitive analytical methods to access non-enzymaticADP-ribosylation, for example using tandem mass spectrometry(MS/MS), have not been developed, although glycation of serumproteins and hemoglobin with glucose is well-studied.[18]

In recent years, MS/MS has been applied in the analysis of afew enzymatically ADP-ribosylated peptides and proteins. Kemp-tide (cAMP-dependent protein kinase substrate), ADP-ribosylatedat Arg2, for example produced a favorable fragmentation pat-tern, which provided sufficient sequence information, follow-

∗ Correspondence to: Ralf Hoffmann, Institute of Bioanalytical Chemistry,Center for Biotechnology and Biomedicine, Faculty of Chemistry andMineralogy, Universitat Leipzig, Deutscher Platz 5, 04103 Leipzig, Germany.E-mail: [email protected]

† Current address: Department of Secondary Metabolism, Leibniz-Institute ofPlant Biochemistry, Weinberg 3, 06120 Halle (Saale), Germany.

Institute of Bioanalytical Chemistry, Center for Biotechnology and Biomedicine,Faculty of Chemistry and Mineralogy, Universitat Leipzig, Deutscher Platz 5,04103 Leipzig, Germany

J. Mass. Spectrom. 2010, 45, 664–669 Copyright c© 2010 John Wiley & Sons, Ltd.

66

5

Fragmentation behavior of ADP-riboslyated peptides

ing electron-capture dissociation (ECD), while predominantlyADP-ribose-derived fragments were detected with collision-induced dissociation (CID).[19] These authors also suggested anew nomenclature of p- and m-ions, resulting from fragmentationof the ADP-ribose site chain, which we have used in this study.Although fragmentation of the ADP-ribose moiety in CID com-plicates the spectral interpretation, it permits easy identificationof accordingly modified peptides in a precursor-ion scan usingthe adenosine monophosphate (AMP) (m/z 348.08), or the ADP-ribose carbodiimide (m/z 584.44).[20] Despite these studies, andreports on glucose-derived Amadori-products,[21] a comprehen-sive mass spectrometric study on the fragmentation pattern ofnon-enzymatic ADP-ribosylation at lysine residues is still missing.

Here, the MS/MS-fragmentation patterns of peptide glycatedwith ADP-ribose at lysine was studied, to the best of our knowledge,for the first time. Tandem mass spectra were acquired on amatrix-assisted laser desorption/ionization-time-of-flight/time-of-flight-mass spectrometry (MALDI-TOF/TOF-MS) in post-sourcedecay (PSD)- and CID-mode, and on an electrospray ionization(ESI)-linear ion trap (LTQ)-Orbitrap-MS in CID- and electron transferdissociation (ETD)-mode.

Experimental

Peptide synthesis and ADP-ribosylation of the model peptide

Peptide Ac-PAAPAAPAPAEKTPV-OH (human histone H1.4, po-sitions 6–20) was synthesized by Fmoc/tBu-strategy on solidphase.[22] The peptide was purified by RP-HPLC and dissolved inphosphate buffer (8 mM Na2HPO4, 1 mM KH2PO4, pH 7.4) con-taining ADP-ribose in a 76 molar access over the peptide (0.017mM). The sample was split into three aliquots of 25 µl each. Onealiquot was incubated immediately, the second was adjusted topH 8.2 using a sodium hydroxide solution (1 M), and sodiumcyanoborohydride (120 mM) was added to the third, in order toreduce the formed Amadori-compound. All samples were then in-cubated at 37 ◦C for 7 days. The reaction progress was monitoredby MALDI-TOF-MS every day.

MALDI-MS

Samples were dissolved in 60% (v/v) aqueous acetonitrilecontaining 0.5% (v/v) formic acid. An aliquot of this solution(0.5 µl) was mixed with an equal volume of α-cyano-4-hydroxy-cinnamic acid solution (4 mg/ml in 50% aqueous acetonitrile;Bruker Daltonics GmbH, Bremen, Germany) on the MALDI targetand air dried. The mass spectra were recorded on a MALDI-TOF/TOF mass spectrometer (4700 proteomic analyzer, AppliedBiosystems GmbH, Darmstadt, Germany) operated in positive ion-reflector TOF (reTOF)-mode (acceleration voltage of 20 kV, 70%grid voltage, 1.277 ns delay, detector voltage of 2 kV) in the m/zrange from 700 to 4000 with a focus mass of 1700. Typically,40 subspectra with 50 laser shots each were accumulated at afixed laser intensity of 5000. The instrument was calibrated onthe same plate using six calibration spots containing the 4700Analyzer Calibration mixture (Applied Biosystems). Product ionspectra of selected precursor ions were acquired in reflectorTOF/TOF mode (acceleration voltage of 8 kV, 70% grid voltage,1.277 ns delay, detector voltage of 2.1 kV, collision energy of 1 kV)by accumulation of 6000 laser shots at a fixed laser intensity of5500. This acquisition mode is called MALDI-CID here. Product ion

spectra recorded in TOF/TOF mode without the use of collisiongas in the collision cell are referred to as MALDI-PSD. Mass spectrawere analyzed with the Data Explorer software package (Version4.6, Applied Biosystems).

Nano-ESI-MS

Samples were dissolved in 60% (v/v) aqueous acetonitrilecontaining 0.5% (v/v) formic acid and analyzed on an LTQOrbitrap XL ETD mass spectrometer equipped with a nano-ESIsource (Thermo Fisher Scientific GmbH, Bremen, Germany). Thetemperature of the transfer capillary was set to 200 ◦C and the tubelens voltage to 120 V. The ion spray voltage (1.3 kV) was appliedto a Nanospray TYP II nano-ESI emitter (BioMedical Instruments,Zollnitz, Germany). The spectrum was acquired with an Orbitrapresolution of 60 000 over an m/z range of 400–2000. Tandem massspectra were acquired using CID (isolation width 2, normalizedcollision energy 35%, activation Q 0.25, activation time 30 ms) orETD (isolation width 2, activation time 75 ms for triply chargedprecursor ion, supplementary activation 5%) activation in the LTQ.Data were acquired and analyzed with Xcalibur software (version2.0.7).

Nano-UPLC–Nano-ESI-MS

Samples were separated on a nano-Acquity UPLC (Waters GmbH,Eschborn, Germany) coupled on-line to the above described LTQOrbitrap XL ETD mass spectrometer, equipped with the same nano-ESI source and using the same transfer capillary temperature andthe tube lens voltage. The ion spray voltage (1.5 kV) was appliedto a PicoTip on-line nano-ESI emitter (standard coating) fornano-UPLC (outer diameter 360/20 µm and tip internal diameter10 µm; New Objective, Berlin, Germany). Samples were dissolvedin eluent A (0.1% formic acid in 3% aqueous acetonitrile, 20 µl) andinjected via the autosampler at a flow rate of 10 µl/min onto a trapcolumn (nanoAcquity UPLC Symmetry C18, 180 µm × 20 mm,particle diameter 5 µm). The samples were then separated ona nanoAcquity UPLC BEH130-column (C18, 100 µm × 100 mm,particle diameter 1.7 µm) using a linear gradient from 3% to 90%aqueous acetonitrile with 0.1% formic acid over 20 min (flow rate400 nl/min). The precursor-ion survey scan and the tandem massspectra were acquired as described above by CID. The ion trap scanwas repeated for the six most intense peaks with charge statesof two or higher. Data were acquired and analyzed with Xcalibursoftware (version 2.0.7).

Results

For glycation of the ε-amino group with ADP-ribose in the histonepeptide, we first used the standard procedure established forD-glucose in neutral buffers.[23] The yields monitored with MALDI-MS and ESI-MS were very low, even after 1 week of incubationat 37 ◦C, with the unmodified peptide always dominating thespectra (Fig. 1A and B). Interestingly, the expected product was notdetected by MALDI-MS in positive-ion mode (Fig. 1A) or negative-ion mode, even at low laser energies (data not shown). Onlya glycated by-product carrying phospho-ribose was detected atm/z 1623.79 (theoretical increment mass of 194.10). The expectedproduct (monoisotopic mass 1970.84) was only detected in the ESImass spectra as a doubly protonated signal at m/z 985.93 (Fig. 1B).The ESI mass spectra were also dominated by the unmodified

J. Mass. Spectrom. 2010, 45, 664–669 Copyright c© 2010 John Wiley & Sons, Ltd. www.interscience.wiley.com/journal/jms

66

6

M. Fedorova, A. Frolov and R. Hoffmann

Figure 1. MALDI-reTOF (left) and ESI-Orbitrap mass spectra (right) recorded in positive-ion mode for peptide Ac-PAAPAAPAPAEKTPV-OH incubated withADP-ribose for 7 days (37 ◦C) at pH 7.4 (top) and pH 8.2 (bottom). The samples were either directly mixed with matrix solution on the MALDI target (left)or dried and dissolved in 60% (v/v) aqueous acetonitrile containing 0.5% (v/v) formic acid before analysis on the LTQ Orbitrap XL ETD-MS (right).

peptide, i.e. it’s singly and doubly charged ions. As MALDI-MS incontrast to ESI-MS did not display the ADP-ribosylated peptide,it appears likely that the ADP-ribosylated peptide fragmentedcompletely within the ionization source forming the phospho-riboslyated peptide. This in-source decay (ISD) is further supportedby the high resolution of the obtained signal and its correct m/z-value. The glycation reaction was accelerated by using slightlyalkaline conditions (pH 8.2),[24] which supposedly increases boththe concentrations of the free ε-amino group of the lysine residueand the open aldose form of ribose relative to the cyclic hemiacetal(Fig. 1, bottom). In addition, the alkaline conditions favor therearrangement of the initially formed aldemine to the more stableAmadori-product. More basic conditions were not tested in thisstudy, as it was previously reported that these compounds degradeat pH 9 with a half live of 15 min.[25] Thus, the modificationwas already detected within 12 h, and its content increasedcontinuously over the following days to reach the maximumconcentration within around 1 week. The higher contents of theADP-ribosylated peptide under these conditions permitted itsdetection in MALDI-MS at an intensity of approximately 5%, relativeto the unmodified peptide. The phospho-ribosylated peptide,presumably mostly formed by ISD, was present at a level of 20%.This dominant fragmentation was also observed in negative-ion

mode (data not shown). Both MALDI- and ESI-MS were moresensitive in positive- than in negative-ion mode for detection ofthe modified peptides, despite the negatively charged phosphategroups. Thus, it is likely that protonation of the glycated ε-aminogroup and the adenine ring compensated for the acidity of theincorporated pyrophosphate ester.

Glycation of the right position in the peptide was confirmed byMS/MS using MALDI-PSD and ESI-CID (Fig. 2). The MALDI-PSDspectrum of m/z 1972.84 was dominated by the phospho-ribosylated peptide with a mass loss of 347.10 u, correspondingto AMP, i.e. a cleavage of the pyrophosphate bond yielding ap5-fragment ion (Fig. 2A, Scheme 1).[19] This dominant loss alsoconfirmed the above interpretation that this fragmentation hadalready occurred at a high degree in the MALDI source. Thefirst 11 residues of the sequence were confirmed by the b- andy-ion series, although the intensities of these signals wereless than 20% relative to the base peak. All glycated b- andy-ions only carried the phospho-ribose in the side chain, whereasno ADP-ribosylated fragments were detected. The structure ofthe phospho-ribosylated peptide ion (m/z 1623.79, p5-ion) wassimilarly confirmed (Fig. 2B and Scheme 1). Interestingly, the y-ionswere all phospho-ribosylated and did not show a significant loss ofthe phosphate group. The ESI-CID spectra of m/z 1972.84 displayed

www.interscience.wiley.com/journal/jms Copyright c© 2010 John Wiley & Sons, Ltd. J. Mass. Spectrom. 2010, 45, 664–669

66

7


Figure 2. MALDI-PSD (A and B), ESI-CID (C) and ESI-ETD (D) mass spectra recorded for peptide Ac-PAAPAAPAPAEKTPV-OH, either ADP-ribosylated (A, Cand D) or phospho-ribosylated (B) at the ε-amino group of lysine. The small inset shows the fragmentation pattern of AMP (m/z 348.08) recorded in anMS3 experiment.

Scheme 1. Structure and obtained fragment ions of a peptide glycated with ADP-ribose at the site chain of a lysine residue. The fragment ions are namedaccording to the nomenclature of Hengel et al.[19] .

very similar fragmentation patterns for both compounds, butprovided a better sequence coverage at the C-terminal residues(Fig. 2C). Both the b- and y-series also included the modified lysine,confirming the right modification side. The pyrophosphate groupwas relatively labile once more, yielding the phospho-ribosylatedpeptide ion as the base peak and mostly phospho-ribosylatedfragment ions, although ions b12 to b14 were ADP-ribosylated. Inaddition, AMP was detected at m/z 348.08 (m6-ion) and confirmedby CID-fragmentation (small inset in Fig. 2C), verifying this neutralloss. In ETD, the ADP-ribosyl group was very stable (Fig. 2D).Cleavage of the pyrophosphate yielded only a minor signal at m/z1625.42 ([M+3H]+), accompanied by AMP at m/z 348.08 (m6-ion).

This fragmentation increased significantly, however, when thesupplementary activation was applied. Thus, the intensity of thesecharacteristic reporter ions can be easily adjusted. The sequencewas completely retrieved from the c- and z-ion series, includingthe modified lysine residue. The longer z- and c-fragment ionsfrom z5 to z14 and c12 and c13, containing the modification sitewere ADP-ribosylated. The corresponding phospho-ribosylatedfragments were not detected.

As ADP-ribosylation was not very stable under either MALDI-PSD or ESI-CID conditions, the keto-group of the Amadori-product was reduced with cyanoborohydride. The monoisotopicmass increased quantitatively within 1 h by 2 u. The recorded


66

8

M. Fedorova, A. Frolov and R. Hoffmann

Figure 3. MALDI-PSD (A and B), ESI-CID (C) and ESI-ETD (D) mass spectra recorded for peptide Ac-PAAPAAPAPAEKTPV-OH carrying a reduced ADP-ribosyl-(A, C and D) or phospho-ribosyl group (B) at the ε-amino group of lysine.

MALDI-PSD, ESI-CID- and ESI-ETD tandem mass spectra displayedvery similar fragment ions series as obtained for the non-reduced peptide (Fig. 3). Even the relative signal intensitieswere almost identical, indicating that the reduction did notinfluence protonation of the ε-amino group of lysine, nor didit influence either the fragmentation of the peptide backboneor the pyrophosphate group. The signal intensities, however,increased, providing significantly better signal-to-noise ratios(Fig. 3), which might be advantageous for real samples with,presumably, lower modification degrees. Moreover, the mass shiftof 2 U, indicative of the glycation site, could provide an easier andmore reliable identification of the modification site, if spectra fromboth forms (native and reduced) are recorded. As reduction didnot influence the fragmentation, sodium cyanoborohydride wasadded directly to the reaction mixture to obtain the reduced ADP-ribosylated peptide. The idea being to capture the initial reversiblyformed aldemine by reducing it immediately and irreversibly.Using sodium cyanoborohydride, glycation occurred much faster,yielding the reduced product at a relatively high content, asindicated by its signal intensity relative to the unmodified peptidein MALDI-MS.

Discussion

The optimized reaction conditions allowed synthesis of the ADP-ribosylated peptide in reasonable yields. This ADP-ribosylatedpeptide was then easily separated from the unmodified peptideby reversed-phase chromatography. Reduction of the keto-group

improved the yields further, which could prove favorable forsynthesizing reduced ADP-ribosylated peptides in larger scales, inorder to study structural and functional aspects in vitro. Althoughit was possible to analyze the modified peptide by both ESI- andMALDI-MS, the dominant loss of AMP in MALDI, most likely viaISD, limits its applicability. It should be noted, however, that thisfragmentation might be attributed to the TOF/TOF-instrument,and the need to use α-cyano-4-hydroxy-cinnamic acid as thematrix. Other matrices in combination with other lasers mightreduce the fragmentation, and thus provide a better directaccess to ADP-ribosylated peptides. It would be interesting tostudy such effects with ‘cooler’ matrices on other instrumenttypes: unfortunately, these were not available for this study.It was, however, still possible to identify the modification sitevia the stable phospho-ribosyl group, which provided a goodsequence coverage in both MALDI-PSD and ESI-CID at the lowpicomole range. ESI-ETD had the advantage of preventing site-chain fragmentations, due to its well-known favorable cleavageof the peptide backbone. All three MS/MS techniques allowedidentification of the ADP-ribosylation site, and thus enableddistinctions to be made between the non-enzymatic glycationat lysine residues, from arginine, aparagine, glutamic acid, andaspartic acid residues modified enzymatically by MARTs.

Cleavage of the pyrophosphate group in ESI-CID also providesdirect access to both neutral loss (347.10 u) and precursor-ion (m/z 348.08) scans, as reported recently for enzymaticallyADP-riboslyated kemptide.[19] In particular, the intense adenosine-monophosphate signal at m/z 348.08 permits an easy, andrelatively specific, identification of ADP-ribosylated peptides in

www.interscience.wiley.com/journal/jms Copyright c© 2010 John Wiley & Sons, Ltd. J. Mass. Spectrom. 2010, 45, 664–669

66

9


complex samples. It can also allow their direct quantification bysingle reaction monitoring (SRM), as successfully applied to otherpost-translational modifications.

Conclusions

Incubation of a peptide with ADP-ribose under slightly alkalineconditions at 37 ◦C yielded the expected Amadori-product insufficient quantities to enable the study of its fragmentationpattern. The corresponding quasi-molecular ion was detectedonly in ESI-MS, whereas the pyrophosphate bond was proneto MALDI-ISD, resulting in only the detection of the phospho-ribosylated peptide. This mass loss of 347.10 u also dominated theESI-CID spectra, even at low collision energies. The b- and y-ionscontaining the modified lysine residue carried only the phospho-ribosyl group in all recorded MALDI-PSD and ESI-CID mass spectra.Thus, the modification site could be reliably identified. Thelabile cleavage site allows a direct identification of this non-enzymatic modification by both neutral loss and precursor-ionscans. In contrast, ADP-ribosylation was mostly stable under ETDconditions, providing a direct access to the accordingly modifiedresidues. Although weaker, the reporter ion at m/z 348.08 andthe corresponding neutral loss were still present, allowing a directidentification of the modification type in the MS from the firstanalysis.

Acknowledgements

A PhD-stipend to M. F. provided by the ‘Gottlieb Daimler- und KarlBenz-Stiftung’ and financial support from the European Fond forRegional Structure Development (EFRE, European Union and FreeState Saxony) are gratefully acknowledged. We thank Dr ChristinaNielsen-Marsh for proofreading.

References

[1] P. Chambon, J. D. Weill, P. Mandel. Nicotinamide mononucleotideactivation of new DNA-dependent polyadenylic acid synthesizingnuclear enzyme. Biochem. Biophys. Res. Commun. 1963, 11, 39.

[2] P. O. Hassa, S. S. Haenni, M. Elser, M. O. Hottiger. Nuclear ADP-ribosylation reactions in mammalian cells: where are we todayand where are we going? Microbiol. Mol. Biol. Rev. 2006, 70, 789.

[3] D. D. i. Corda, M. Girolamo. Mono-ADP-ribosylation: a tool formodulating immune response and cell signaling. Sci. STKE. 2002,2002, e53.

[4] D. D. i. Corda, M. Girolamo. Functional aspects of protein mono-ADP-ribosylation. EMBO J. 2003, 22, 1953.

[5] J. Moss, S. J. Stanley, P. A. Watkins. Isolation and properties ofan NAD- and guanidine-dependent ADP-ribosyltransferase fromturkey erythrocytes. J. Biol. Chem. 1980, 255, 5838.

[6] J. Moss, M. K. Jacobson, S. J. Stanley. Reversibility of arginine-specific mono(ADP-ribosyl)ation: identification in erythrocytes ofan ADP-ribose-L-arginine cleavage enzyme. Proc. Natl. Acad. Sci. USA1985, 82, 5603.

[7] K. M. Krueger, J. T. Barbieri. The family of bacterial ADP-ribosylatingexotoxins. Clin. Microbiol. Rev. 1995, 8, 34.

[8] J. Sun, J. T. Barbieri. Pseudomonas aeruginosa ExoT ADP-ribosylatesCT10 regulator of kinase (Crk) proteins. J. Biol. Chem. 2003, 278,32794.

[9] G. Glowacki, R. Braren, K. Firner, M. Nissen, M. Kuhl, P. Reche,F. Bazan, M. Cetkovic-Cvrlje, E. Leiter, F. Haag, F. Koch-Nolte. Thefamily of toxin-related ecto-ADP-ribosyltransferases in humans andthe mouse. Protein Sci. 2002, 11, 1657.

[10] E. L. Jacobson, D. Cervantes-Laurean, M. K. Jacobson. ADP-ribose inglycation and glycoxidation reactions. Adv. Exp. Med. Biol. 1997,419, 371.

[11] I. Just, P. Wollenberg, J. Moss, K. Aktories. Cysteine-specific ADP-ribosylation of actin. Eur. J. Biochem. 1994, 221, 1047.

[12] S. V. Kharadia, D. J. Graves. Relationship of phosphorylation andADP-ribosylation using a synthetic peptide as a model substrate.J. Biol. Chem. 1987, 262, 17379.

[13] J. E. Hodge. The Amadori rearrangement. Adv. Carbohydr. Chem.1955, 10, 169.

[14] K. Heyns, H. Noack. Die Umsetzung von D-Fructose mit L-Lysineund L-Arginin und deren Beziehung zu nichtenzymatischenBraunungsreaktionen. Chem. Ber. 1962, 95, 720.

[15] R. Alvarez-Gonzalez, F. R. Althaus. Poly(ADP-ribose) catabolism inmammalian cells exposed to DNA-damaging agents. Mutat. Res.1989, 218, 67.

[16] A. H. Guse. Regulation of calcium signaling by the secondmessenger cyclic adenosine diphosphoribose (cADPR). Curr. Mol.Med. 2004, 4, 239.

[17] J. Moss, A. Zolkiewska, I. Okazaki. ADP-ribosylarginine hydrolasesand ADP-ribosyltransferases. Partners in ADP-ribosylation cycles.Adv. Exp. Med. Biol. 1997, 419, 25.

[18] C. S. Ho, C. W. K. Lam, M. H. M. Chan, R. C. K. Cheung, L. K. Law,L. C. W. Lit, K. F. Ng, M. W. M. Suen, H. L. Tai. Electrospray ionisationmass spectrometry: principles and clinical applications. Clin.Biochem. Rev. 2003, 24, 3.

[19] S. M. Hengel, S. A. Shaffer, B. L. Nunn, D. R. Goodlett. Tandem massspectrometry investigation of ADP-ribosylated kemptide. J. Am. Soc.Mass Spectrom. 2009, 20, 477.

[20] H. Osago, K. Yamada, T. Shibata, K. Yoshino, N. Hara, M. Tsuchiya.Precursor ion scanning and sequencing of arginine-ADP-ribosylatedpeptide by mass spectrometry. Anal. Biochem. 2009, 393, 248.

[21] A. Frolov, P. Hoffmann, R. Hoffmann. Fragmentation behavior ofglycated peptides derived from D-glucose, D-fructose and D-ribosein tandem mass spectrometry. J. Mass Spectrom. 2006, 41, 1459.

[22] D. H. Rich, J. Singh. The Carbodiimide Method. The Peptides, E. Grossand J. Meinenhofer (Eds). Academic Press: New York, 1979.

[23] A. Frolov, D. Singer, R. Hoffmann. Site-specific synthesis of Amadori-modified peptides on solid phase. J. Pept. Sci. 2006, 12, 389.

[24] D. Cervantes-Laurean, D. E. Minter, E. L. Jacobson, M. K. Jacobson.Protein glycation by ADP-ribose: studies of model conjugates.Biochemistry 1993, 32, 1528.

[25] E. L. Jacobson, D. Cervantes-Laurean, M. K. Jacobson. Glycation ofproteins by ADP-ribose. Mol. Cell Biochem. 1994, 138, 207.


fragmentation behavior of amadori-peptides obtained by non-enzymatic glycosylation of lysine...

Documents