Reactions of a modified lysine with aldehydic and diketonic dicarbonyl compounds: An electrospray mass spectrometry structure/activity study

JOURNAL OF MASS SPECTROMETRY

J. Mass Spectrom. 2006; 41: 216–228

Published online 13 January 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.980

Reactions of a modiﬁed lysine with aldehydic and

diketonic dicarbonyl compounds: an electrospray mass

spectrometry structure/activity study

1,2∗

Marco A. Saraiva,^1,2Carlos M. Borges^1,2and M. Helena Florencio

ˆ

1

Department of Chemistry and Biochemistry, Faculty of Sciences of the University of Lisbon, Campo Grande, C8, 1749-016 Lisbon, Portugal

Center for Chemistry and Biochemistry, Faculty of Sciences of the University of Lisbon, Campo Grande, C8, 1749-016 Lisbon, Portugal

2

Received 7 July 2005; Accepted 1 November 2005

The phenomenon known as non-enzymatic glycation is described as the reaction of reducing sugars with

basic amino groups of proteins and nucleic acids, as well as with simple amines, without enzyme mediation.

Non-enzymatic model glycation reactions that make use of low-molecular-weight compounds make an

important contribution in the elucidation of glicated processes in vitro and in vivo. Four a-dicarbonyl

compounds, aldehydic (glyoxal, methylglyoxal and phenylglyoxal) and ketonic (diacetyl), were reacted

with the modiﬁed amino acid N_a-acetyl-L-lysine (AcLys) in an attempt to establish structure/activity

relationships for the reactivity of a-dicarbonyls with the amine compound. Electrospray ionization

mass spectrometry (ESI-MS) combined with tandem mass spectrometry (MS/MS) and collision-induced

dissociation (CID) was used to identify and characterize reagents, intermediates and reaction products. The

formation of dicarbonyl-derived lysine dimers was observed exclusively. Especially, attention is drawn to

alkyl- (asymmetrical dicarbonyl systems) and carboxyl- (glyoxal system) substituted imidazolium ions, at

ring position 2.

The main differences observed in the reactions studied were related to the reactivity with the diimine

intermediate. This intermediate can react either with a non-hydrated dicarbonyl molecule at the aldehydic

carbonyl, or with a mono-hydrated one at the ketonic carbonyl, particularly for asymmetrical dicarbonyls.

For 2-carboxyl-substituted imidazolium ion (glyoxal reaction), besides the usual keto–enol rearrangement

from the diol group, an alternative reaction pathway (proton abstraction) appears to contribute also for

the imidazolium ring-closure process. Moreover, the formation of imidazolium ring structures can depend

on several factors, namely, the presence (or absence) of electron donor substituents at the formed diol,

the degree of stability of the new electrophile generated and/or the equilibrium concentration of the non-

and mono-hydrated dicarbonyl forms in solution, the last being particularly important for asymmetrical

dicarbonyls.

The results reported reveal the complexity of reactivity as well as the diversity of imidazolium

molecular structures. Copyright  2006 John Wiley & Sons, Ltd.

KEYWORDS: non-enzymatic model glycation reactions; ESI-MS/tandem mass spectrometry; aldehydic and ketonic

˛-dicarbonyls; acetyl-lysine; imidazolium ions

and a protein amino group in equilibrium with the cyclic

glycosamine form. This step is followed by a rearrangement,

i.e. the Schiff base gives rise to an enaminal form and subse-

quently to a relatively stable ketoamine compound (Amadori

rearrangement). The Amadori compound is stabilized by its

cyclization to a furanose or pyranose ring.²In the inter-

mediate stage, the stable Amadori compound is degraded

by a series of reactions to yield small molecules containing

dicarbonyl groups, which are more reactive when compared

with the starting sugar molecules. In the ﬁnal stage, chem-

ically irreversible as the preceding one, a wide variety of

structurally diverse adducts, known as advanced glycation

end products (AGEs) are formed. The AGE adducts have,

frequently, chromophores, ﬂuoropores and protein cross-

links. AGEs can be distinguished in a number of aspects

INTRODUCTION

The Maillard reaction,¹or non-enzymatic glycation, has a

key role in the understanding of some pathogenesis result-

ing from diabetic complications and the ageing process. A

number of important reports have been published on this

subject in the past decade.^2,3In food chemistry also the

Maillard reaction has a recognized importance.⁴In a simpli-

ﬁed way, the Maillard reaction can chemically be analysed in

three stages. In the initial stage, there is a reversible formation

of a Schiff base (unstable form) between a reducing sugar

^ŁCorrespondence to: M. Helena Floreˆncio, Department of

Chemistry and Biochemistry, Faculty of Sciences of the University

of Lisbon, Campo Grande, C8, 1749-016 Lisbon, Portugal.

E-mail: mhﬂorencio@fc.ul.pt

Copyright  2006 John Wiley & Sons, Ltd.

Reaction of a modiﬁed lysine with ˛-dicarbonyls

217

O

Glyoxal

Methylglyoxal

Phenylglyoxal

Diacetyl

HN

OH

NH₂

O

N_α-acetyl-L-lysine (AcLys)

Lys

N

Lys

N

Lys

N,N(-di(N_ε−lysino))imidazolium

(glyoxal-lysine dimer, GOLD)

N,N(-di(N −lysino))-4-methyl-imidazolium

ε

(methylglyoxal-lysine dimer, MOLD)

Figure 1. Chemical structures of ˛-dicarbonyls, acetyl-lysine (AcLys) and imidazolium ions GOLD and MOLD.

related to their biological and physiological functions: pro-

tein cross-links; recognition factors for speciﬁc AGE-binding

cell-surface receptors; markers and risk predictors of dis-

ease processes.⁵In the past few years, more than one dozen

AGEs have been discovered. Some are related to diabetes

and other pathologies, and their accumulation in tissue pro-

teins, particularly in long-lived structural proteins such as

collagen, ﬁbronectin, tubulin, cristallin lens and myelin, has

been observed. As a result of AGE formation, protein mod-

iﬁcation has been observed due to structure and functional

changes. Furthermore, ˛-dicarbonyl compounds have been

recognized as the major intermediates and/or precursors

in AGE formation, in vivo. Recently, it has been proposed

that ˛-dicarbonyl stress is among the major factors in the

pathogenesis of diabetic complications.^6–8Glyoxal, methyl-

glyoxal (MGO), 3-deoxyglucosone, and glucosone are the

four reactive dicarbonyls that form the group of heteroge-

neous AGEs. The amine groups of the proteins, involved

in the reaction towards sugars and/or dicarbonyls com-

pounds, correspond to (1) N-terminal amino groups of the

polypeptide chains, (2) the ε-amino groups of lysine residues

and (3) guanidino groups in the arginine residue. The extent

of glycation in physiological systems is very low, typically

0.01–1% of the lysine and arginine residues.⁹Its recognized

importance justiﬁes, therefore, in vitro studies in which the

extent of glycation can be improved.⁹The glyoxal dicar-

bonyl is a highly reactive dicarbonyl and it is generated

during the autoxidation of glucose and glycolaldehyde, as

the major product observed. The oxidation of polyunsatu-

rated fatty acids (PUFAs) involving formation of hydrogen

peroxide and ˇ-fragmentation also generate glyoxal.¹⁰DNA

oxidation seems also to produce this dicarbonyl.¹¹Glyoxal

reacts with lysine, cysteine and arginine residues on pro-

teins to form irreversible AGEs. N_˛-(carboxymethyl)lysine

(CML),^12–14glyoxal–lysine dimer (GOLD)^15–17(Fig. 1) and

N_ε-(carboxymethyl)arginine (CMA)^18,19are good examples

of glyoxal-derived AGEs. MGO, similar to glyoxal, is a

by-product of glycolysis.²⁰Minor routes such as PUFA²¹

degradation and acetone metabolism and catabolism of thre-

onine also generate MGO.^22,23Other minor biochemical path-

ways for MGO formation are non-enzymatic ˇ-elimination

of phosphate from triose 3-phosphate intermediates, glycer-

aldehyde 3-phosphate and dihydroxyacetone phosphate.^5,24

The main pathway for MGO production in vivo is unknown,

since it is formed during both non-oxidative and oxidative

chemistry.²Furthermore, MGO is a physiological substrate

of the gluthatione-dependent glyoxalase pathway, which

catalyses the detoxiﬁcation of MGO to ^D-lactate.^25,26This

asymmetrical dicarbonyl reacts fast with amino, guani-

dino and thiol functional groups in proteins²⁷and even

with nucleic acids.²⁸MGO is considerably more reactive

than glucose (20 000 times) and has the capability to dis-

rupt metabolic functions including mitochondrial respiration

and glycolysis.^29,30A number of important MGO-derived

AGEs are reported in the literature, e.g. the hydroimida-

zolones (and isomers),⁹argpyrimidine,³¹tetrahydropyrimi-

dine (THP)⁷and MGO–lysine dimer (MOLD) (Fig. 1).¹⁶

In an effort to elucidate the complex Maillard reaction and

to access AGEs, reaction models with low-molecular-weight

compounds have been studied. These models are based on

the reaction of simple amine compounds, modiﬁed amino

acids (and similar compounds) and peptides with dicarbonyl

compounds and even with sugars (and derivatives). In the

past two decades, this approach has been systematically

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

218

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

undertaken, especially in food chemistry.³²Since dicarbonyl

compounds are among the most important intermediates

in the Maillard reaction, the reactivity towards the amino

compound N_˛-acetyl-L-Lysine (AcLys) of four dicarbonyl

molecules, aldehydic (glyoxal, MGO and phenylglyoxal),

diketonic (diacetyl) (Fig. 1), and their different forms in

solution, hydrated and non-hydrated, were investigated in

our study. The proposed study attempts also to establish a

relationship between the structure of the dicarbonyl and its

ability to react with the amino compound AcLys. This is an

aspect that has not been systematically explored from the

chemical point of view. Moreover, some considerations on

the electrophilic character of the dicarbonyl molecules are

also reported. Results from another study,³³involving the

same dicarbonyl molecules but with modiﬁed arginine, have

been taken into account in the current investigation.

Recently, mass spectrometry has been used for protein

glycation and cross-linking research owing to its high

sensitivity and speciﬁcity.^34–36Electrospray ionization (ESI)

and ion trap mass spectrometry techniques were used as

the methods for our investigation on the non-enzymatic

reactions of AcLys with the four dicarbonyls. To identify as

well as to elucidate the structural aspects of the unknown

compound ions formed in those reactions, ESI was combined

with tandem mass spectrometry (MS/MS) and collision-

induced dissociation (CID).

chromatography/mass spectrometry (GC-MS). Moreover, in

other experiments, organic layers of the hydrolysed mixture

were analysed by GC-MS and revealed that pyruvate is

a possible contaminant, although to a lesser extent, in

comparison with MGO.³⁷

Preparation and incubation of the reaction

mixtures

The experiments were initially performed at 30 C. The

°

continuous detection of many intermediates and products

in the course of several weeks stimulated us to increase the

°

temperature up to 70 C. Experimentally, 1 ml of the buffer

solution (200 m^M) at pH 7.5, previously prepared, was added

to 1 ml of the aqueous solution of the modiﬁed amino acid

AcArg (100 m^M). To this mixture was also added, separately,

1 ml of the aqueous solution of the dicarbonyls glyoxal,

MGO, phenylglyoxal and diacetyl (100 m^M) up to a total

volume of 3 ml. After adding the dicarbonyl compounds,

°

the solutions were maintained incubated at 70 C in a stove

(Memmert, model 1500) with air circulation for 28 days.

Further details concerning the incubation procedure at the

latter temperature are presented in the Appendix. For MGO

and diacetyl reaction systems only, the reaction was initiated

°

at room temperature (r.t.) and kept at 70 C immediately

afterwards. At previously selected ﬁxed time intervals,

°

aliquots of 100 µl were taken and kept at ꢀ80 C immediately.

In order to exclude ions derived from the reaction and/or

interaction with the buffer compound, HEPES, incubations

with the amine compound and the four dicarbonyls using this

buffer were conducted individually at the same temperature

and pH. The buffer HEPES has been used throughout

our studies. Nevertheless, for conﬁrmation purposes, the

buffer system MOPS–KOH has also been used, where

necessary. Both buffer systems HEPES and MOPS–KOH are

convenient because they are temperature resistant, which is

important for the reaction conditions used. They have also

minimum salt effects and form no complexes with metal

ions.^39,40In addition, when the buffer HEPES was used, the

ionization/desolvation mass spectrometry process for the

reaction systems studied was found to improve.

EXPERIMENTAL

Materials

The dicarbonyls, glyoxal, phenylglyoxal and diacetyl, the

modiﬁed amino acids AcLys and AcArg, the sulfonic buffers,

HEPES (sodium salt), MOPS and the MGO precursor,

methylglyoxal 1,1⁰-dimethylacetal, were purchased from

Sigma. In the preparation of the reaction mixtures, ultrapure

water (18.4 Mꢀ.ms) was used, produced in a Wasserlab G. R.

apparatus. Saturated solutions of sodium- and potassium

hydroxide, NaOH and KOH, respectively, (solid pellets

both from Merck) and/or a concentrated solution of

hydrochloride acid, HCl (Riedel-de Hae¨n), were used to

adjust the pH of the reaction solutions and to prepare the

sulfonic buffer solutions.

Methods

Methylglyoxal identiﬁcation and characterization was per-

formed on a gas chromatograph coupled to a mass spec-

trometer, GC-MS system, Fisions GC 8000 Series, cou-

pled to a mass detector Trio 1000; column Rtx-4Sil MS

30 m ð 0.25 mm; i.d. 0.25 µm; carrier gas – helium at 100 psi.

Mass spectrometry experiments were also performed on

a LCQduo (Finnigan, San Jose, CA, USA) ion trap mass

spectrometer equipped with an electrospray ion source,

maintained at 4.5 kV. The coaxial sheath gas (nitrogen) ﬂow

rate was ca 20 psi, with an auxiliary gas (nitrogen) ﬂow rate

of 20 psi. The capillary was kept at ca 10 V and it was also

MGO synthesis via methylglyoxal

1,1^ꢀ-dimethylacetal acid hydrolysis

The commercially available MGO solution (40% in water)

was not adequate, since it was contaminated mainly with

formaldehyde (30% in the solution), as well as with other

impurities such as pyruvate, lactate and formate.³⁷For this

reason, MGO was synthesized following a method described

by Kellum et al.³⁸This method follows its preparation

from methylglyoxal 1,1⁰-dimethylacetal precursor by acid

hydrolysis. Experimentally, a mixture containing 2500 µl

of sulfuric acid, H₂SO₄, 1/10 (v/v) in millipore water,

2300 µl of millipore water and 200 µl of methylglyoxal 1,1⁰-

dimethylacetal was heated in a water bath and kept boiling

for 25 min. The resulting pale yellow mixture was cooled and

kept in ice water until use. For conﬁrmation purposes, the

hydrolysed mixture was distilled, and the MGO yield was ca

50%. This compound was identiﬁed and characterized by gas

°

maintained heated at 220 C. The pressure, measured with

the convectron gauge during electrospray experiments, was

normally 0.92 Torr. The base pressure in the ion trap with

helium was ca 1.12 ð 10^ꢀ5Torr. To obtain ESI mass spectra,

the samples were diluted 200 times in Millipore water and

infused directly into the ESI source at a ﬂow rate of 5 µl/ min.

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

Reaction of a modiﬁed lysine with ˛-dicarbonyls

219

0.00001

0.0001

0.001

0.01

0.1

0.0001

0.001

0.01

0.1

mz/

m/z

1

10

100

1000

10000

1000

1.0E-02 1.0E-03 1.0E-04 1.0E-05 1.0E-06 1.0E-07 1.0E-08

[AcLys] / M

(a)

[AcArg] / M

(b)

Figure 2. Relative abundance of the modiﬁed amino acid protonated molecules, AcLys and AcArg, observed in the ESI mass

spectra, normalized with respect to m/z 261 ions (most prominent buffer ion in the ESI mass spectra) as a function of solution

^{concentration: (a)}♦ m/z 189 ion (AcLys protonated molecule) (buffer solution concentration 10^ꢀ3M); ꢀ m/z 189 ion (buffer solution

concentration 10^ꢀ5M); ◊ predicted variation for m/z 189 ions, accounting for the different buffer solution concentrations used;

^(b)♦ m/z 217 ion (AcArg protonated molecule) (buffer solution concentration 10^ꢀ3M); ꢀ m/z 217 ion (buffer solution concentration

10^ꢀ5M); ◊ predicted variation for m/z 217 ions, accounting for the different buffer solution concentrations used. ESI-MS operating

conditions are as stated in the Experimental section.

Experiments were carried out in the positive ion mode and

full scans were recorded in the mass range m/z 50–1000.

The mass spectrometer was operated using three microscans

with a maximum ion injection time of 50 ms (default values).

In the ESI-MS/MS and CID experiments, the ions were

isolated in the ion trap by the development of appropriate

voltages in the different lens segments of the instrument.

The ions, especially M^C, were forced to collide with helium

gas retained in the ion trap segment of the apparatus. For

the excitation of the ionic species, 20–50% of the maximum

available collision energy was applied. This energy value

was selected in order to ensure that the precursor ions could

be observed in the ESI-MS/MS spectra, i.e. a remaining

relative peak intensity of about of 1–5% for precursor ions

was achieved. The total ion current was above 5 ð 10³in all

cases.

systems studied also did not change signiﬁcantly with

reaction time. For this reason, m/z 261 ions were used as the

standard. Measurements were based on the ratio between

the relative ion abundance of m/z 261 ions and the total

ion current (TIC) of the corresponding mass spectra (results

not shown). Several aspects regarding the use of the buffer

HEPES in the present study should also be pointed out:

(1) The buffer ions do not seem to suppress the protonated

molecules of AcLys and AcArg in experiments conducted

individually. The relative abundances of protonated amino

acid molecules plotted as a function of m/z 261 buffer ions

at two different buffer concentrations (10^ꢀ5and 10^ꢀ3M)

are shown in Fig. 2. (2) A linear relationship between the

amino acid concentration in solution and the corresponding

normalized relative abundances of protonated molecules in

the ESI mass spectra (range: 5 ð 10^ꢀ6–5 ð 10^ꢀ3M for the

protonated AcLys molecule and 5 ð 10^ꢀ8–10^ꢀ4M for the

protonated AcArg molecule) is observed (Fig. 2). Moreover,

the solution pH also did not change signiﬁcantly; the

buffer may possess electrolyte behaviour, levelling the total

ionic current. (3) Stabilization of the reaction pH occurs,

an important factor for the maintenance of the ion signal

stability considering that some of the molecules can easily be

protonated in solution.

Reactions systems and ESI-MS analysis were performed

in duplicate. ESI-MS/MS spectra were repeated two to three

times.

RESULTS AND DISCUSSION

Buffer behaviour (HEPES)

The ESI mass spectra (not shown) of the reactions of the

amino compound AcLys and of the single component

solutions with the dicarbonyls revealed that no signiﬁcant

interaction with the buffer occurred. Nevertheless, to study

the effect of the analyte ion signal suppression in non-volatile

buffered solutions and for the purpose of comparison with

AcLys, experiments using AcArg (an amino acid having

also a basic side chain and a similar magnitude of the

physical properties such as polarity, hydrophobicity and

media bulkiness) have been performed. The abundance of

the m/z 261 buffer ion in the ESI spectra of all reaction

Behaviour of the dicarbonyl reaction systems

under ESI mass spectrometry

Four reaction systems were studied in this investigation,

i.e. the reactions of the modiﬁed amino acid AcLys with

the dicarbonyls glyoxal, MGO, diacetyl and phenylglyoxal.

At selected ﬁxed reaction times, aliquots were recovered,

and the ESI mass spectra obtained (not shown) exhibit

several M^Cions and one [M C H]^Cion, with very little/no

fragmentation. As an example, the ESI mass spectra of

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

220

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

aliquots of MGO reaction system solutions are presented in

Fig. 3. The M^Cions correspond to: simple imidazolium ions

at m/z 411 (glyoxal), 425 (MGO), 487 (phenylglyoxal) and

453 (diacetyl), ion 2; 2-alkyl-substituted imidazolium ions at

m/z 439 (MGO) and 563 (phenylglyoxal), ions 3; 2-carboxyl-

susbtituted imidazolium ion at m/z 455 (glyoxal), ions 4.

The [M C H]^Cion corresponds to a substituted dihydro-

1,4-diazepine ion at m/z 497 (MGO), ion 5. All M^Cand

[M C H]^Cions are referred to in Table 1, and their relative

abundances were determined at two different reaction times.

Together with these ions, dicarbonyl-derived lysine dimer

adduct ions with sodium and potassium atoms are also

included in Table 1. In most cases, the relative abundance

of the identiﬁed ions (reaction products) was less than 10%.

This observation can be attributed to a suppression effect by

the buffer ions. It must however be emphasized that the TIC

obtained in the ESI spectra was, in general, ca 10⁷. Moreover,

the amine compound AcLys was found to react at different

rates with the four dicarbonyls. In general, phenylglyoxal

and glyoxal appear to react more slowly, whereas MGO and

diacetyl react much faster, accounting for the lower relative

abundance of m/z 189 ion, as can be observed in Table 1

for AcLys protonated molecule, when comparing relative

intensities and corresponding reaction times. Furthermore,

and despite the rapid reaction of the AcLys compound with

MGO and diacetyl dicarbonyls, in less than 4 h of incubation,

some of the ions of interest were found to be diminished in

their relative abundances, e.g. m/z 439 ions in the MGO

reaction (Fig. 3 and Table 1).

x8

x30

497*

469 483

x8

100

80

60

40

20

0

*

439

*

425

519

541

400

420

440

460

480

500

520

x8

540

(a)

m/z

x8

x30

100

80

60

40

20

0

497 *

469

541

483

*

425

*

439

400

420

440

460

480

m/z

500

520

540

(b)

Figure 3. ESI mass spectra of aliquots of MGO reaction

system solutions: (a) 30 min and (b) 240 min. ^ŁDenotes the

ions of interest. All other ions assigned in the spectra refer to

imidazolium adduct ions with sodium. The m/z 261 ion (buffer

HEPES) is the base peak of the spectra.

structure, also consistent with the literature reports for

imidazolium compounds GOLD and MOLD.^15,16

Identiﬁcation and characterization of reaction

products

The ions at m/z 411, 425, 487 and 453 (ions 2), for

glyoxal, MGO, phenylglyoxal and diacetyl reaction systems,

respectively, (Tables 3 and 4) were found to possess the

same fragmentation pattern, namely, with losses from the

carboxyl group, amide group and part of the side chain

of the modiﬁed amino acid, AcLys (Tables 3 and 4). These

losses can be attributed to water, to the protective acetyl

On the basis of ESI-MS/MS and CID analysis (not shown),

ion structures for the ions 2–5 (Table 1) are proposed and

depicted in Table 2. These ion structures are consistent with

the fragment ion composition, depicted in Tables 3 and 4.

All intermediate/product ions identiﬁed were found to

possess the dicarbonyl-derived lysine dimer-type molecular

Table 1. Identiﬁed ions based on ESI-MS analysis and their normalized relative abundances at different reaction times

Compounds of

interest

Reactions systems (acetyl-lysine (AcLys) with glyoxal, MGO, phenylglyoxal

and diacetyl) m/z (M^Cor [M C H]^C) (Rel. abundance (%))

Glyoxal

Methylglyoxal

Phenylglyoxal

Diacetyl

AcLys (1)

189 (46.7 ^a; 27.5 ^b) 189 (29.9 ^a; 31.5 ^c) 189 (61.2 ^a; 26.0 ^d) 189 (42.4 ^a; 40.9 ^e)

[M]^C(2) (Simple imidazolium ions)

411 (3.1; 16.1)

425 (6.6; 4.7)

439 (10.5; 2.9)

f

487 (17.1; 198.4)

563 ( – ; 6.3)

453 (0.6; 1.1)

[M]^C(3) (2-Alkyl-substituted imidazolium ions)

[M]^C(4) (2-Carboxyl-substituted imidazolium ion)

[M C H]^C(5) (Substituted dihydro-1,4-diazepine)

[M ꢀ H C Na]^C(2a)/[M ꢀ 2H C 2Na]^C(2b)

[M ꢀ H C Na]^C(3a)/M ꢀ 2H C 2Na]^C(3b)

[M C Na]^C(5a)/[M ꢀ H C 2Na]^C(5b)

f

455 (29.5 ^g,h; – )

f

497 (8.8; 0.9)

– /469

– /483

519/541

f

433/487 ^h

509/531

585/ –

f

Reaction times: ^a30 min; ^b7 days; ^c240 min; ^d14 days; ^e120 min; ^fNot applicable; ^g60 min.

Relative abundances normalized to the m/z 261 HEPES buffer ion.

^hIon obtained from different reaction conditions: MOPS-KOH buffer system. The last three lines of the Table refer to adduct ions with

sodium (or potassium) atoms identiﬁed in the ESI mass spectra.

–

Not detected.

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

Reaction of a modiﬁed lysine with ˛-dicarbonyls

221

Table 2. Proposed ion structures for the ions 2–5, mentioned in Table 1

Ions

of

interest

Reaction systems (AcLys with glyoxal, MGO, phenylglyoxal and diacetyl)

Possible ion structures

Glyoxal

Methylglyoxal

Phenylglyoxal

Diacetyl

H₃C

CH₃

H

CH₃

H

R

N

N R

2

R

N

N R

R

N

N R

N

R

N

CH₃

H

(m/z 425)

(m/z 411)

(m/z 487)

(m/z 453)

CH₃

H

3

4

—

R

N

N R

N

R

N

CH₃

(m/z 439)

H

(m/z 563)

—

R

N

N R

O

OH

(m/z 455)

+

H

CH₃

R

N

N R

5

—

CH₃

H₃C

H

OH

(m/z 497)

R D C(COOH)ꢁNHCOCH₃ꢂꢁCH₂ꢂ₄. – Not detected.

group (42 Da loss), to CO₂(44 Da loss) and, to a greater

extent, to the complete side chain of the AcLys compound.

The common loss of 171 Da (Tables 3 and 4) is an example

of the latter. Such a 171-Da loss, also observed in the

imidazolium ions GOLD and MOLD MS/MS spectra, is

consistent with literature reports^41,42and accounts for the

presence of the acetal protective group in the product

ions formed in the present case. MS/MS data of triosidine

compounds which are formed in the reaction of AcLys

with glyceraldehyde revealed equivalent losses from the

modiﬁed amino acid residue. These losses refer to H₂O,

acetyl (42Da), acetyl C CO C H₂O (88 Da), acetyl C NH₃C

CO C H₂O (105 Da) and acetyl-norleucine (171 Da).⁴³The

simple imidazolium ions, ions 2, at m/z 411, 425, 487 and

453 for glyoxal, MGO, phenylglyoxal and diacetyl reaction

systems, respectively, exhibit a similar fragmentation trend

with respect to the main fragmentation pathways (leading

to abundant fragment ions) as the one observed for the

AcLys protonated molecule (Table 3). A second loss of

171 Da and the subsequent release of the imidazolium ring

are encountered only for the phenylglyoxal reaction system

(Table 4). Furthermore, the relative abundance of the m/z

487 ion was signiﬁcantly higher, when compared to m/z

425 ions abundance, corresponding to simple imidazolium

ions, ions 2, of MGO and phenylglyoxal reaction systems,

respectively (Table 1). Note that the m/z 487 ion was found

to be the base peak in the ESI spectrum after the ﬁrst 7 h

of reaction. Interestingly, 2-alkyl-substituted imidazolium

ions, ions 3, at m/z 439 (MGO) and 563 (phenylglyoxal)

(Table 1), with a fragmentation pattern (Table 4) similar

to the one of ions 2 for the corresponding MGO and

phenylglyoxal reaction systems (Table 3), were also observed

for the asymmetrical dicarbonyls reaction systems. The m/z

439 (MGO) and m/z 563 (phenylglyoxal) ions, however,

may possess, in comparison with ions 2, a methyl and

a phenyl substituent group, respectively. Although not

reported in the literature, for the MGO reaction system,

the m/z 439 ion was found to have a higher relative

abundance, when compared to the one of the m/z 425

ion of the acetylated MOLD compound, an important AGE.

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

222

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

Table 3. Relative abundances (%) of fragment ions, derived from the precursor ions identiﬁed in the ESI mass spectra of

glyoxal and diacetyl reaction systems (ions 1, 2 and 4 mentioned in Table 1)

Relative ion

abundances (%)

Fragment ion

composition

m/z (Precursor ion)

Glyoxal; diacetyl

455; ^c433;^d475

Acetyl-lysine (AcLys)

189^a

211^b

411;453

–

487;^c,e

–

[M ꢀ NH₃]^C

[M ꢀ H₂O]^{C f}

9.7

100

98.2

–

6.1

–

1.7

h

–

1.1

100

–

h

–

61.7; 36.7

11.6; 4.1

9.1; 15.9

100; 100

60.7; 82.2

20.2; 31.3

68.0^g;

–

54.9; 30.1

35.3; 30.8

30.3; 21.9

16.9; 12.5

1.6; 5.9

2.0; 1.5

12.1; 7.1

100; 100

1.4; 4.5

2.0; 8.0

1.8; 6.4

–

1.1; –

[M ꢀ H₂O ꢀ CH₂CO]^C

[M ꢀ H₂O ꢀ CO₂]^C

[M ꢀ CH₂CO]^{C f}

–

1.9; –

[M ꢀ CO₂]^C

100^g;

–

h

[M ꢀ H₂O ꢀ CO]^C

[M ꢀ CH₂CO ꢀ CO ꢀ H₂O]^{C f}

[M ꢀ NaC₂H₅NO]^C

[M ꢀ NaC₂H₅NO ꢀ CO]^C

[M ꢀ 104]^Cž

–

h

–

h

–

5.8

–

h

–

[M ꢀ CH₂CO ꢀ NH₃ꢀ CO ꢀ H₂O]^{C f}

[M ꢀ 106]^Cž

–

1.7; 1.2

37.1; 7.4

–

4.5; 1.8

–

[M ꢀ 171]^{C f}

30.0; 6.4

2.6;

–

0.8;

100; –

[M ꢀ 171 ꢀ H₂O]^C

–

[M ꢀ 171 ꢀ CO₂]^C

h

[M ꢀ 171 ꢀ H₂O ꢀ CO₂]^C

–

^a,bAcLys protonated molecule and AcLys sodiated molecule, respectively.

^cIon obtained using the buffer system MOPS-KOH.

^dImidazolium adduct ion with one sodium atom (2a) (Table 1).

^eImidazolium adduct ion with two potassium atoms (2b) (Table 1).

^fLosses also observed from acetyl-trihydroxy-triosidine and acetyl-triosidine-carbaldehyde protonated molecules, having

one acetyl-lysine residue in the ion structure (Ref. 43).

^gRelative abundances of fragment ions, derived from the imidazolium ring, instead of the amino acid residue, as usual.

^hNot applicable.

–

Not detected.

104, 106 and 171 Da losses refer to C₃H₆O₃N, C₃H₈O₃N and C₈H₁₃O₃N (acetyl-norleucine) residues, respectively.

Experiments performed with the buffer system MOPS–KOH

conﬁrmed the above-mentioned observation. An ion at

m/z 455 was observed in the ESI mass spectra of the

glyoxal reaction system (Table 3). This ion remained present

independently of the buffer system used. Its fragmentation

pattern showed prominent losses of H₂O and CO₂. Such

a fragmentation pattern differs from the one observed for

simple- and 2-alkyl-substituted imidazolium, parent ions 2

and 3, respectively (Tables 3 and 4). This fragmentation can

be attributed to the presence of a carboxyl group in the

imidazolium ring. Note that when the buffer used is HEPES,

the ion at m/z 455 has a main contribution from the doubly

sodiated moiety. Furthermore, the m/z 455 ion relative

abundance of 29.5% (Table 1) in the ﬁrst hour of reaction

decreased during the ﬁrst day of incubation (MOPS–KOH

buffer system). In the ESI mass spectra (not shown) of the

MGO reaction system, a substituted dihydro-1,4-diazepine

ion at m/z 497, ion 5, was observed (Table 1). This ion appears

to be one of the most prominent ions of interest detected in

the ESI spectra of the MGO reaction system, especially in

the ﬁrst 2 h of incubation (Table 1). The fragment ions from

the substituted dihydro-1,4-diazepine parent ion of MGO

reaction system are shown in Table 4. This ion results from

a fragmentation similar to the one occurring for precursor

ions 2 and 3 (Tables 3 and 4). A possible ion structure for

m/z 497 ions may therefore be a lysine dimer with a 1,4-

diazepine ring (seven-membered ring) (Table 2), instead

of the ﬁve-membered imidazolium ring ions structures

described for ions 2, 3 and 4 (Table 2). Indeed, the formation

of a 1,4-diazepine ring involves a more stable dicarbonyl-

derived reaction intermediate which is known to react

with arginine residues, particularly in the formation of the

argpyrimidine compound.^31,44Moreover, the fragment ions

resulting from m/z 497 (Table 4) are consistent with stable

pyrimidine ring structures (argpyrimidine). In addition,

this proposal (substituted dihydro-1,4-diazepine ion) is

consistent with the fragment ion composition that we

observed for argpyrimidine molecule, an analogue of the

m/z 497 ion, namely, with losses from the amino acid residue

of the modiﬁed amino acid (unpublished results).³³

Besides all the identiﬁed ions mentioned above, adducts

ions with sodium and potassium have also been observed

in the reactions with both buffer systems HEPES–HCl and

MOPS–KOH, with the exception of 2-carboxyl-substituted

imidazolium ion, ion 4. In Table 4, an example of m/z 487

imidazolium adduct ions with one and two sodium atoms,

i.e. m/z 509 and 531 ions, respectively, is presented. Adduct

ions with one and two sodium (or potassium) atoms were

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

Reaction of a modiﬁed lysine with ˛-dicarbonyls

223

Table 4. Relative abundances (%) of fragment ions, derived from precursor ions identiﬁed in the ESI mass

spectra, of MGO and phenylglyoxal reaction systems (ions 2, 3 and 5, mentioned in Table 1)

Relative ion

abundances (%)

Fragment ion

composition

m/z (precursor ion MGO; phenylglyoxal)

425; 487

439; 563

497; –

52.6;

– ; 509^a

– ; 531^b

[M ꢀ H₂O]^Cc

46.6; 48.3

3.4; 8.2

3.8; 11.1

100; 100

58.8; 71.6

7.5; 22.4

37.5; 35.3

2.0; 2.1

13.5; 6.8

100; 99.6

76.1; 100

18.5; 22.3

–

– 37.3

– ; 29.3

– ; 26.9

– ; 18.3

– ; 5.5

– ; 5.1

– ; 16.2

– ; 100

– ; 6.8

– ; 10.0

– ; 7.0

– ; 0.5

– ; 29.2

– ; 5.0

–

– ; 12.0

–

– ; 7.5

– ; 7.8

–

[M ꢀ H₂O ꢀ CH₂CO]^C

[M ꢀ H₂O ꢀ CO₂]^C

[M ꢀ CH₂CO]^Cc

7.5;

–

12.3;

100;

–

[M ꢀ CO₂]^C

75.4;

28.9;

–

[M ꢀ H₂O ꢀ CO]^C

[M ꢀ CH₂CO ꢀ CO ꢀ H₂O]^Cc

[M ꢀ NaC₂H₅NO]^C

[M ꢀ NaC₂H₅NO ꢀ CO]^C

[M ꢀ 104]^Cž

19.7; 36.3

27.7; 32.1

38.1;

d

0.5;

–

[M ꢀ CH₂CO ꢀ NH₃ꢀ CO ꢀ H₂O]^{C c}

[M ꢀ 106]^Cž

–

0.5;

–

0.8;

33.8;

9.2;

1.8;

0.7;

–

[M ꢀ 171]^Cc

8.9; 43.0

9.8; 57.1

–

1.0; 4.7

–

– ; 100

[M ꢀ 171 ꢀ H₂O]^C

[M ꢀ 171 ꢀ CO₂]^C

–

– ; 4.6

–

[M ꢀ 171 ꢀ H₂O ꢀ CO₂]^C

[M ꢀ 2^Ł171]^C

– ; 1.2

–

– ; 0.6

^a,bImidazolium adduct ions with one (2a) (Table 1) and two (2b) (Table 1) sodium atoms, respectively.

^cLosses also observed from acetyl-trihydroxy-triosidine and acetyl-triosidine-carbaldehyde protonated molecules,

having one acetyl-lysine residue in the ion structure (Ref. 43).

^dNot applicable.

–

Not detected.

104, 106 and 171 Da losses refer to C₃H₆O₃N, C₃H₈O₃N and C₈H₁₃O₃N (acetyl-norleucine) residues, respectively.

observed for the most abundant ions at m/z 411 (glyoxal), 497

(MGO) and 487 (phenylglyoxal). These ions, besides having

a fragmentation similar to imidazolium ions, exhibit the loss

of NaC₂H₃O₂(82 Da), assigned to adduct ions with one

sodium atom, and conﬁrmed by a similar proeminent loss

of KC₂H₃O₂(98 Da) from adduct ions with one potassium

atom. Further conﬁrmation is given by the adduct ions in

the MS/MS spectra with two sodium (or potassium) atoms,

which also shows an abundant loss of 171 Da, this being

characteristic of precursor ions that do not interact with

sodium and potassium. Since metal cations (Na^Cor K^C)

can interact individually with the two carboxylic groups of

lysine residues in the charge/proton transfer processes, such

a phenomenon may account for this observation.

imidazolium ring structures, would react more easily with

the ketonic carbonyl in a MGO mono-hydrated dicarbonyl

form than in a phenylglyoxal mono-hydrated one. This

situation is indeed observed, since formation of 2-alkyl-

substituted imidazolium ions, ions 3, occurs exclusively for

asymmetrical dicarbonyl reaction systems. The mechanism

of formation proposed for these ions 3 (Scheme 1) is similar to

the mechanism reported for simple imidazolium ions MOLD

and GOLD.^15,16The fact that the ion at m/z 439 (MGO), ion

3 (Table 1), possesses a higher relative abundance, at least

in the ﬁrst 4 h of reaction, when compared with the relative

abundance of the ion at m/z 425 (MGO), ion 2, may justify

the more favoured reaction of the diimine intermediate

with the mono-hydrated MGO dicarbonyl form. It can

also be observed that the abundance of phenylglyoxal

simple imidazolium ions, ions 2, (Table 1) rapidly increases,

whereas phenylglyoxal ions 3 are observed after a period

of several weeks. We believe that these aspects contribute

to conﬁrm the importance of the mono-hydrated dicarbonyl

MGO and phenylglyoxal forms in solution and also their

dynamic equilibrium concentrations in the formation of

the imidazolium ions. The analysis of the mechanism of

formation of ions 3 (Scheme 1), together with literature

reports for non-acetylated ions 2 formation, suggests that

two important factors are involved in the formation of

the imidazolium ring structures. One is the presence (or

absence) of electron substituents in the diol group, which

Role of the dicarbonyl forms and their hydration

equilibria in the reaction with AcLys – mechanistic

considerations

MGO exists in aqueous solution in three major forms:

non-, mono- and di-hytrated. These forms are in dynamic

equilibrium in the approximate ratio of 1 : 71 : 28. The MGO

non-hydrated form is more reactive than the mono-hydrated

one, although the latter is present at a much higher

concentration in equilibrium.^37,45Phenylglyoxal behaves

similarly, although its extent of reactivity, with respect

to the mono-hydrated form, is less pronounced than in

mono-hydrated MGO forms. It should therefore be expected

that the diimine intermediate, involved in the formation of

Copyright  2006 John Wiley & Sons, Ltd.

J. Mass Spectrom. 2006; 41: 216–228

224

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

R

N

R

R₁

O

R₁

H

R₁

O

NH₂

+

- 2 H₂O

N

R

HO

H

HO

R

Glyoxal

2 AcLys

Diimine

Absence of an electron

donor substituent in the

diol group

R

N

R₁

N

OH

H

O

N

R

N

R

H

OH

H

Intermediate A

Less efficient

nucleophile

Stabilized

new generated

electrophile

R

R₁

H

R₁

N

OH

R₁

OH

R₁

- HCOOH

N

R

N

R

H

More efficient

nucleophile

R

R₁

H

N

R₁= CH₃(Methylglyoxal); Ph (Phenylglyoxal)

R = C(COOH)(NHCOCH₃)(CH₂)₄

R₁

N

R

2-Alkyl-substituted imidazolium ions - ions 3

[m/z 439 (Methylglyoxal); m/z 563 (Phenylglyoxal)]

Scheme 1. Proposed reaction mechanism for the 2-alkyl-substituted imidazolium ions formation, ions 3, in both MGO and

phenylglyoxal reaction systems.

can favour (or not) carboxylic acid release. The other is

the degree of stability of the newly generated electrophile.

For the latter, less stabilization favours the nucleophilic

attack by the reinforced nucleophile (N atom available

in the diimine intermediate) (intermediate A – Scheme 1).

It should, however, be emphasized that the keto–enol

rearrangement in the diol group controls the process of

the new electrophile generation. The latter is in turn

responsible for the imidazolium ring closure and the

extent of its formation depends on the strength of the

nucleophile. For the simple imidazolium ions, ions 2, and

for asymmetrical dicarbonyls, both mentioned processes,

i.e. the presence of electron donor substituents to promote

carboxylic acid release and the stabilization of the newly

generated electrophile, seem to favour the formation of the

imidazolium ring. For the 2-alkyl-substituted imidazolium

ions, ions 3, (Table 1) none of these processes, absence

of electron donor substituents in the diol group and

stabilization of the newly generated electrophile, seem to

signiﬁcantly contribute to the formation of the imidazolium

ring. In this case, the magnitude of these two processes for

imidazolium ring formation seem to compensate each other,

and hardly any modiﬁcations are observed in the reaction

mechanism. Another aspect that should be considered in the

formation of ions 3, which seems to be an important feature,

is the equilibrium concentration of the mono-hydrated

dicarbonyl forms involved.

With respect to the glyoxal reaction system, it should

be taken into consideration that glyoxal is highly hydrated

in aqueous solution and exists as ca 0.005% non-hydrated

form, ca 0.5% mono-hydrated form, ca 1–2% as dimers and

the remainder as di-hydrated.⁴⁵The rapid formation of the

m/z 411 ion (ion 2 – glyoxal; Table 1) can be explained by

the consumption of the non-hydrated glyoxal form, i.e. its

J. Mass Spectrom. 2006; 41: 216–228

Reaction of a modiﬁed lysine with ˛-dicarbonyls

225

most reactive form. The hydration equilibrium is, therefore,

displaced from the di-hydrated form to the non- and mono-

hydrated glyoxal forms. The formation of the carboxyl-

substituted imidazolium ion (ion 4 – m/z 455), along with the

formation of the simple imidazolium ion, ion 2, is attributed

to the fact that, although depending on the equilibrium

concentration of non- and di-hydrated glyoxal forms, mono-

hydrated glyoxal forms should also be available in solution.

The fact that 2-carboxyl-substituted imidazolium ions, ion 4,

appear in the beginning of the reaction and disappear rapidly

in the ﬁrst day of the reaction (MOPS–KOH buffer system)

(results not shown) seems to reinforce this assumption. The

appearance of the mono-hydrated form in the beginning of

the incubation and its reaction with the diimine intermediate

seem also to be consistent with this assumption, since

once formed, it can rapidly be consumed by displacement

of the hydration equilibrium towards the non-hydrated

form (short life-time form). Furthermore, a comparison of

the formation of simple imidazolium ions and 2-carboxyl-

substituted imidazolium ion reaction mechanisms, ions 2

and 4, respectively (Scheme 2), reveals that no carboxylic

acid release is observed in the latter. Furthermore, the

diol group and C˛ hydrogen atoms of intermediate A

(Scheme 2) are almost equally acidic. Therefore, in the

process of imidazolium ring closure, proton abstraction

on the C˛ hydrogen appears to be an alternative pathway

for the generation of the new electrophile, instead of the

usual keto–enol rearrangement. Both the processes may

R

O

H

N

H

O

NH₂

+

- 2 H₂O

N

R

HO

H

HO

R

Diimine

2 AcLys

Glyoxal

Proton abstraction;

generation of a

Destabilized

new generated

electrophile

new electrophile

R

N

R

H

OH

H

N

OH

N

R

N

R

H

OH

H

Absence of an electron

donor substituent in the

diol group

Intermediate A

R

N

R

N

H

OH

H

OH

N

R

N

H

OH

H

R

Oxidation

(diol)

R

N

H

O

R = C(COOH)(NHCOCH₃)(CH₂)₄

N

R

OH

2-Carboxyl-substituted imidazolium ion - ion 4

[m/z 455 (Glyoxal)]

Scheme 2. Proposed reaction mechanism for the 2-carboxyl-substituted imidazolium ion formation, ion 4, in the glyoxal reaction

system.

J. Mass Spectrom. 2006; 41: 216–228

226

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

however coexist, and simple imidazolium ions and formation

of 2-carboxyl-substituted imidazolium ion, ions 2 and 4,

respectively, can result from competitive reactions. Another

possible explanation for the rapid decrease in abundance

of ion 4 with time in the ESI mass spectra (results not

shown) could be related to the rapid consumption of the

OH^ꢀin solution available for proton abstraction, since

the reaction pH is neutral (ca 7.5). Furthermore, if we

consider the two factors mentioned above for ion 4, i.e.

the absence of electron donor substituents in the diol

group to promote carboxylic acid release and the stabilized

new electrophile generated, important for imidazolium ring

closure, one might suggest that in the formation of 2-

carboxyl-substituted imidazolium ion, no electron donor

substituent is present in the diol group and the new

electrophile generated is destabilized. This could possibly

induce the development of another competitive reaction

pathway that might more easily favour the generation of

the new electrophile to promote imidazolium ring closure

(Scheme 2).

Acetyl derivatives of CML and N_ε-(carboxyethyl)lysine

(CEL) have not been detected in our study. This was

somewhat surprising, since these two compounds seem

to be quite relevant according to the literature.^12–14,27An

explanation could be the fact that these compounds are acid

labile, and can also be developed in acidic medium workup

procedures, which may render the conﬁrmation of their

existence difﬁcult.

We are convinced that to confer equal stability and sym-

metry to both MGO molecule elctrophiles, which could

be achieved, for example, by reaction of a simple sub-

stituted amine with the MGO molecule, would minimize

imidazolium ion formation. Moreover, a minimization of

imidazolium ion formation has been observed when diacetyl

reacts with AcLys (Table 1). This latter dicarbonyl has been

used for a number of years as a standard arginine modiﬁer,

and this modiﬁcation is reasonably considered a read-

ily reversible process.^46–48In addition, a modiﬁed MGO

molecule, with more stability (and symmetry) in both devel-

oped electrophiles, as is the case of the diacetyl molecule,

might result in less different products with less stability. This

stabilization and symmetry of both dicarbonyl electrophiles

seem, therefore, to have an important role, considering that

imidazolium compounds and their derivatives are relatively

stable in aqueous solution.

behaves as a nucleophile when reacting with aldehydic

and ketonic carbonyl group of the dicarbonyl molecules.

This observation is particularly important for asymmetrical

dicarbonyls, and also contributes to the high equilibrium

concentration of the respective mono-hydrated dicarbonyl

molecules, as in the case of 2-alkyl-substituted imidazolium

ions, ions 3. Furthermore, and despite the important

concentration in solution of mono-hydrated phenylglyoxal

forms, its reactivity with the ketonic carbonyl group, such as

in ions 3, is not favoured, and may account for the higher

electrophile stability of this latter group. Nevertheless, when

the difference in the stability of both developed dicarbonyl

electrophiles is not so prominent, which is the case for the

MGO molecule, the attack towards the ketonic group is

favoured. Carboxylic acid release was absent in 2-carboxyl-

substituted imidazolium ion, ion 4 – glyoxal reaction system.

In the formation of ions 4, when compared with that of

ions 2, a competitive reaction pathway becomes important

for imidazolium ring closure. Furthermore, aspects related

with the type of electron donor substituent character (or

absence) in the diol group formed and the stability of the

new electrophile generated seem to have an important role in

the formation of imidazolium ions. The instability observed

for some of the ions identiﬁed, namely, the ions at m/z 455

and 497, for glyoxal and MGO reaction systems, respectively,

should also be mentioned. These compounds, if able to

survive in vivo, might be used as novel standards to monitor

and/or approach new therapies in the disease processes

they would be related with, especially in an early stage.

Furthermore, alkyl- and carboxyl-substituted imidazolium

ions and a substituted dihydro-1,4-diazepine were obtained

in these reactions.

We believe our observations combining both solution

and gas-phase experiments may contribute to a better

understanding of the solution chemistry of imidazolium

salts. They may also be used as possible alternative pathways

for the syntheses of substituted imidazolium salts and

substituted 1,4-diazepine compounds.

Acknowledgements

The authors gratefully acknowledge Dr Carlos Cordeiro and Prof.

Ana Ponces Freire of the enzymology group (CQB-FCUL, Portugal)

for their inestimable contribution to this work, especially in the

non-enzymatic glycation area. Discussions with Prof. Susana Santos

are also acknowledged. One of the authors, M. Saraiva, thanks

Fundac¸a˜o para a Cieˆncia e a Tecnologia (FCT) for a Ph.D. scholarship

(SFRH/BD/3161/2000).

CONCLUSIONS

APPENDIX

In the four reaction systems studied, several dicarbonyl-

derived lysine dimer ions were identiﬁed and characterized.

Besides the identiﬁcation of the acetylated GOLD and MOLD

ions, all other ions, especially the ones corresponding to the

most relevant dicarbonyls, glyoxal and MGO, have not, to our

knowledge, been reported before. Simple imidazolium ions,

ions 2, in particular, possess the higher relative abundances

in the ESI mass spectra of all reaction systems, except

for the MGO system. In terms of reaction mechanics, the

imidazolium ions appear to have versatile structures upon

closure of the imidazolium ring. The diimine intermediate

Experimental details: preparation and incubation

of the reaction mixtures

The solutions of the dicarbonyls were pre-incubated at

°

70 C except for MGO and diacetyl dicarbonyl solutions,

because at this temperature reversible acetal formation

(the MGO precursor form) occurred together with some

hydration of the dicarbonyl molecules. The glyoxal and

phenylglyoxal solutions did not show signiﬁcant alterations

in their reactivity due to the pre-incubation procedure, and

were, therefore, the only two dicarbonyl solutions submitted

°

to this procedure (70 C).

J. Mass Spectrom. 2006; 41: 216–228

Reaction of a modiﬁed lysine with ˛-dicarbonyls

227

to complication in insulin-dependent diabetes mellitus. J. Clin.

Invest. 1993; 91: 2463.

15. Kevin J, Wells-Knecht KJ, Brinkmann-Frye E, Baynes JW.

Characterization of an imidazolium salt formed from glioxal

and N_˛-hippuryllysine: a model for Maillard reaction crosslink

in proteins. J. Org. Chem. 1995; 60: 6246.

16. Brinkmann-Frye E, Wells-Knecht KJ, Thorpe SR, Baynes JW.

Characterization of an imidazolium compound formed by

reaction of methylglyoxal and N^˛-hippuryllysine. J. Chem. Soc.

Perkin Trans. 1 1995; 22: 2817.

17. Thornalley PJ, Langborg A, Minhas HS. Formation of glyoxal,

methylglyoxal and 3-deoxyglucosone in the glycation of proteins

by glucose. Biochem. J. 1999; 344: 109.

18. Iijima K, Murata M, Takahara H, Irie S, Fujimoto D. Identiﬁ-

cation of N^ω-carboxymethylarginine as a novel acid-labile

advanced glycation end product in collagen. Biochem. J. 2000;

347: 23.

19. Glomb MA, Lang G. Isolation and characterization of glyoxal-

arginine modiﬁcations. J. Agric. Food Chem. 2001; 49: 1493, DOI:

10.1021/jf001082d.

Phenylglyoxal solubility in water is limited; therefore

it was necessary to place the solution in an ultrasonic

°

water bath (Sonorex TK 52H) at 40 C to ensure its complete

solubility.

Methylglyoxal was the only reaction system for which pH

of the resulting solution needed to be adjusted, already at the

beginning of its incubation, using a pH meter, E516 titriskop,

connected to a glass membrane electrode. In general, the pH

of the reaction mixtures, starting at 7.5, gradually diminished

with time. In order to keep an acceptable range of one pH

unit difference, 7.5 ꢁ pH ꢁ 6.5,⁴all reaction solutions

were adjusted in the ﬁrst 4 h of incubation.

°

At ꢀ20 C, sample colour and/or peak relative intensity

in the ESI mass spectra changed signiﬁcantly in a few

hours. Since sample solidiﬁcation did not produce relevant

°

alterations, the temperature of ꢀ80 C was therefore used.

20. Hayasi T, Mase S, Namiki N. Formation of three-carbon sugar

fragments at an early stage of the browning reaction of sugar

with amine or amino acids. Agric. Biol. Chem. 1986; 50: 1959, DOI:

10.1021/jf960067z.

21. Portero-Otin M, Bellmunt MJ, Requena JR, Pamplona R. Protein

modiﬁcation by advanced Maillard adducts can be modulated

by dietary polyunsaturated fatty acids. Biochem. Soc. Trans. 2003;

31: 1403.

22. Ray M, Ray S. Aminoacetone oxidase from goat liver. Formation

of methylglyoxal from aminoacetone. J. Biol. Chem. 1987; 262:

5974.

23. Reichard GA, Skutches CL, Hoeldtke RD, Owen OE. Acetone

metabolism in humans during diabetic ketoacidosis. Diabetes

1986; 35: 668.

24. Abordo EA, Minhas HA, Thornalley PJ. Accumulation of

˛-oxoaldehydes during oxidative stress. A role in citotoxicity.

Biochem. Pharmacol. 1999; 58: 641.

25. Thornalley PJ. The glyoxalase system: new developments

towards functional characterization of a metabolic pathway

fundamental to biological life. Biochem. J. 1990; 269: 1.

26. Philips SA, Thornalley PJ. The formation of methylglyoxal

from triosephosphates. Investigation using a speciﬁc assay for

methylglyoxal. Eur. J. Biochem. 1993; 212: 101.

27. Ahmed MU, Brinkmann-Frye E, Degenhardt TP, Thorpe SR,

Baynes JW. N_ε-(carboxyethyl)lysine, a product of the chemical

modiﬁcation of proteins by methylglyoxal increases with age in

human lens proteins. Biochem. J. 1997; 324: 565.

REFERENCES

1. Maillard L-C. Action des acides amines sur les sucres: formation

des melanoidines por voie methodique. C. R. Hebd. Seances Acad.

Sci. Ser. 1912; 154: 66.

2. Lacovella GM, Alderson NL, Thorpe SR, Baynes JW. The role of

reactive ˛-dicarbonyls in AGE/ALE formation and implication

in diabetic complications. South Carolina Journal of Molecular

Medicine 2004; 5: 73.

3. Baynes JW, Thorpe SR. Role of oxidative stress in diabetic

complications: a new perspective on an old paradigm. Diabetes

1999; 48: 1.

4. Martins SIFS. Unravelling the Maillard reaction network by

multiresponse kinetic modelling. Ph.D. Thesis, Wageningen

University: The Netherlands, 2003.

5. Ahmed N, Thornalley PJ. Chromatographic assay of glycation

adducts in human serum albumin glycated in vitro by

derivatization with 6-aminoquinolyl-N-hydroxysuccinimidyl-

carbamate and intrinsic ﬂuorescence. Biochem. J. 2002; 364: 15.

6. Chellan P, Nagaraj RH. Protein crosslinking by the Maillard

reaction: dicarbonyl-derived imidazolium crosslinks in aging

and diabetes. Arch. Biochem. Biophys. 1999; 368: 98.

7. Oya T, Hattori N, Mizuno Y, Miyata S, Maeda S, Osawa T,

Uchida K. Methylglyoxal modiﬁcation of protein. Chemical and

immunochemical characterization of methylglyoxal-arginine

adducts. J. Biol. Chem. 1999; 274: 18492.

8. Singh R, Barden A, Mori T, Beilin L. Advanced glycation end-

products: a review. Diabetologia 2001; 44: 129.

9. Ahmed N, Argirov OK, Minhas HS, Cordeiro CAA, Thornal-

ley PJ. Assay of advanced glycation endproducts (AGEs): sur-

veying AGEs by chromatographic assay with derivatization by 6-

aminoquinolyl-N-hydroxysuccinimidyl-carbamate and applica-

tion to N_ε-carboxymethyl-lysine- and N_ε-(1-carboxyethyl)lysine-

modiﬁed albumin. Biochem. J. 2002; 264: 1.

10. Mlakar A, Spiteller G. Reinvestigation of lipid peroxidation of

linolenic acid. Biochim. Biophys. Acta 1994; 1214: 209, DOI:

10.1016/0005-2760(94)90046-9.

11. Murata-Kamiya N, Kamiya H, Kaji H, Kasai H. Glyoxal, a major

product of DNA oxidation, induces mutation at G : C sites on

a shuttle vector plasmid replicated in mammalian cells. Nucleic

Acids Res. 1997; 25: 1897.

12. Dunn JA, Patrick JS, Thorpe SR, Baynes JW. Oxidation of

glycated proteins: age-dependent accumulation of N_ε-

(carboxymethyl)lysine in lens proteins. Biochem. J. 1989; 28: 9464.

13. Dyer DG, Blackledge JA, Thorpe SR, Baynes JW. Formation of

pentosidine during nonenzymatic browning of proteins by

glucose. Identiﬁcation of glucose and other carbohydrates as

possible precursors of pentosidine in vivo. J. Biol. Chem. 1991;

266: 11654.

28. Murata-Kamiya M, Kamiya H. Methylglyoxal, and endogenous

aldehyde, crosslinks DNA polymerase and the substrate DNA.

Nuc. Acids Res. 2001; 29: 3433.

29. Bowes JH, Cater CW. The interaction of aldehydes with collagen.

Biochem. Biophys. Acta 1968; 168: 341.

30. Halder J, Ray M, Ray S. Inhibition of glycolysis and

mitochondrial respiration of Ehrlich ascites cells by

methylglyxal. Int. J. Cancer 1993; 54: 443.

31. Al-Abed Y, Misuhashi T, Ulrich P, Bucala R. Novel modiﬁca-

tions of N^˛-BOC-arginine and N^˛-CBZ-lysine by methylglyoxal.

Bioorg. Med. Chem. Lett. 1996; 6: 1577.

32. Lee K-G, Shibamoto T. Toxicology and antioxidant activities of

non-enzymatic browning reaction products: review. Food Rev.

Int. 2002; 18: 151.

33. Saraiva MA, Borges CM, Floreˆncio MH. (in preparation).

34. Lapolla A, Fedele D, Plebani M, Aronica R, Garbeglio M,

Seraglia R, D‘Alpaos M, Traldi P. Evaluation of glycated

globins by matrix-assisted laser desorption/ionization mass

spectrometry. Clin. Chem. 1999; 45: 288.

35. Lapolla A, Fedele D, Traldi P. The role of mass spectrometry in

the study of non-enzymatic protein glycation in diabetes. Mass

Spectrom. Rev. 2000; 19: 279.

14. Dyer DG, Dunn JA, Thorpe SR, Bailie KE, Lyons TJ, McCance

DR, Baynes JW. Maillard reaction products and theor relation

J. Mass Spectrom. 2006; 41: 216–228

228

M. A. Saraiva, C. M. Borges and M. H. Floreˆncio

36. Yeboah FK, Yaylayan VA. Analysis of glycated proteins by mass

spectrometry techniques: qualitative and quantitative aspects.

Nahrung 2001; 45: 164.

37. Kalapos MP. Methylglyoxal in living organisms chemistry,

biochemistry, toxicology and biological implications. Toxicol.

Lett. 1999; 110: 145.

increased in serum proteins of uremic patients: evidence for

increased oxidative stress in uremia. FEBS Lett. 1998; 427: 381.

43. Tessier FJ, Monnier VM, Sayre LM, Kornﬁeld JA. Triosidines:

novel Maillard reaction products and cross-links from the

reaction of triose sugars with lysine and arginine residues.

Biochem. J. 2003; 369: 705.

38. Kellum MW, Oray B, Norton SJ. A convenient quantitative

synthesis of methylglyoxal for glyoxalase I assays. Anal. Biochem.

1978; 85: 586.

39. Good NE, Winglet GD, Winter W, Conolly TN, Izawa S,

Singh RMM. Hydrogen ion buffers for biological research.

Biochemistry 1966; 5: 467.

40. Good NE, Izawa S. Hydrogen ion buffers. Methods Enzymol. 1972;

24: 53, DOI: 10.1016/0076-6879(72)24054-x.

41. Nagaraj RH, Shipanova IN, Faust FM. Protein cross-linking

by the Maillard reaction. Isolation, characterization, and

in vivo detection of a lysine-lysine cross-link derived from

methylglyoxal. J. Biol. Chem. 1996; 271: 19338.

44. Shipanova IN, Glomb MA, Nagaraj RH. Protein modiﬁcation

by methylglyoxal: chemical nature and synthetic mechanism

of a major ﬂuorescent adduct. Arch. Biochem. Biophys. 1997;

344: 29.

45. Thornalley PJ, Yurek-George A, Argirov OK. Kinetics and

mechanism of the reaction of aminoguanidine with alpha-

oxoaldehydes glyoxal, methylglyoxal, and 3-deoxyglucosone

under physiological conditions. Biochem. Pharmacol. 2000; 60: 55,

DOI: 10.1016/S0006-2952(00)002887-2.

46. Grossberg AL, Pressman D. Modiﬁcation of arginine in the

active sites of antibodies. Biochemistry 1968; 7: 272.

47. Yankeelov JA Jr. Modiﬁcation of arginine in proteins by

oligomers of 2,3-butanedione. Biochemistry 1970; 9: 2433.

48. Yankeelov JA Jr. Modiﬁcation of arginines by diketones.

J. Methods Enzymol. 1972; 15: 566.

42. Odani H, Shinzato T, Usami J, Matsumoto Y, Brinkmann-

Freye E, Baynes JW, Maeda K. Imidazolium crosslinks derived

from reaction of lysine with glyoxal and methylglyoxal are

J. Mass Spectrom. 2006; 41: 216–228

Article Doi

Reactions of a modified lysine with aldehydic and diketonic dicarbonyl compounds: An electrospray mass spectrometry structure/activity study

DOI: 10.1002/jms.980

Source and publish data:

Authors:

Article abstract of DOI:10.1002/jms.980

Full text of DOI:10.1002/jms.980

Products guided by the article

R&D Labs maybe for 78-98-8

Relevant to this article

Hot Product