light from maillard reaction: photon counting, emission spectrum, photography and visual perception

J BlOLUMlN CHEMILUMIN 1995; 10: 277-84

Light from Mai l lard Reaction: Photon Counting, Emission Spectrum, Photography and Visual Perception

Georg Wondrak,' Thomas Pier2 and Roland Tressl' lnstitut fur Biotechnologie' and lnstitut fur Festkorperphysik,* Technische Universitat, Seestrasse 13, 13353 Berlin, Germany

Several authors have reported on high-sensitivity measurement of oxygen-dependent low-level chemiluminescence (CL) from Maillard reactions (MR), i.e. nonenzymatic amino-carbonyl reactions between reducing sugars and amino acids (also referred t o as nonenzymatic browning). Here we report for the first time, that light from Maillard reactions can be seen by the human eye and also can be photographed. In parallel w i th visual perception and photography CL was monitored by means of a CL-detection programme of a liquid scintillation counter (LSC, single photon rate counting). CL emission spectrum was recorded by a monochromator-microchannel plate photomultiplier arrangement. CL intensity from reaction of 6-aminocaproic acid wi th D-

ribose (200 mg each) in 5 mL H20 a t pH 11 at 95°C was high enough for visual perception after adaptation t o absolute darkness. Reaction in dimethylsulphoxide (DMSO) exhibited strongly enhanced CL (10mg each in 5mL were sufficient for visual detection) and could be photographed (15 minutes' exposure, ASA 6400); all characteristics of Maillard specific CL (02-dependence, no CL from nonreducing sugars, inhibition bysul- phur compounds) remained. Visual detection of CLand measurement by LSC were in ful l concordance. The CL emission spectrum showed t w o broad peaks a t around 500 nm and 695nm. Fluorescence emission of the brown reaction mixture matched the blue- green part of the CL emission spectrum. Emission of visible light during Maillard reactions may partly originate from oxygen-dependent generation of excited states and energy transfer t o simultaneously formed fluorescent products of the browning react ion.

Keywords: Maillard reaction; chemiluminescence; visual perception; photography

I NTRO D U CTlO N

Recently, several authors have reported on high- sensitivity measurements of very weak chemiluminescence (CL) during early stages of Maillard reactions (MR) (1,2). Maillard reactions, i.e. nonenzymatic amino-carbonyl reactions between amino acids (free or peptide bound) and reducing sugars with subsequent browning of the reaction mixture, have been under investigation by food

chemists for decades; most roast aromas are formed by Maillard reactions (e.g. N-containing heterocyclic compounds such as pyrazines, pyrroles, pyridines, thiazoles and oxazoles) (3). Var- ious Maillard reaction pathways have been elucidated using stable isotope labelling techniques (4): generally, after initial Schiff base formation a complex set of reactions such as dehydration, re- arrangement, aldol cleavage and condensation leads to the formation of various heterocyclic

CCC 0884-3996/95/050277-08 0 1995 by John Wiley & Sons, Ltd.

Received 19 November 1994 Revised 24 April 1995

278 G . WONDRAK H A L .

compounds; browning of the reaction mixture occurs by polymerization. Additionally, early stages of Maillard reactions proceed via stable radical intermediates (5).

Biochemical interest in physiologically relevant Maillard reactions arose when it was shown, that nonenzymatically glycosylated hemoglobin (HbA1,) is elevated in human diabetes. Even in healthy individuals HbA1, ranges between 3.0% and 8.0% of total hemoglobin, pointing to the general in vivo relevance of Maillard reactions (6). Since then, a whole range of physiologically long- lived proteins such as extracellular matrix col- lagens has been identified as being nonenzymatically modified by Maillard reactions in vivo (7).

Kurosaki et al. (1,2) described an extra-weak chemiluminescence arising from amino-carbonyl reactions between reducing sugars and amino acids and presented ESR evidence for the implica- tion of a carbon radical in the CL phenomenon. Namiki et al. (8) did experiments proposing a pyrazinium radical cation as the key intermediate for production of CL during early stages of Mail- lard reactions.

The purpose of our experiments was to further characterize the nature of the light-producing reaction. Here we report, for the first time, that light from MR can be seen by the human eye and- with dimethylsulphoxide (DMSO) as solvent- also can be photographed.

MATERIALS AND METHODS

Reagents

All chemical reagents were purchased from Fluka AG. Switzerland.

Visual detection of chemiluminescence

Maillard reaction in water. The standard reaction was done by dissolving 6-aminocaproic acid (6- AC) and D-ribose (200 mg each) in 5 mL of water at pH 1 1. After dark adaptation for at least 15 minutes, the reaction was started by heating the test tube on a water bath at 95°C with gentle shaking. After 5 minutes a very weak glow of the solution could be seen.

Maillard reaction in DMSO. Reactions were done as for in water but DMSO (pure DMSO without

addition of acid or base) was used as solvent and only 100mg of each reactant was employed. For better detectability, reactions in water or DMSO could be enhanced by bubbling oxygen through the reaction mixture (1000 mL/min). The visual detection experiments were reproduced by five independent observers.

Monitoring chemiluminescence by liquid scintillation counting (LSC)

Chemiluminescence from Maillard reactions in DMSO was monitored by means of a liquid scintillation counter following a standard procedure (9). The reaction mixture (100mg of each reactant in 5mL DMSO) in an LSC plastic vial was heated at 95°C for three minutes under air, cooled to 60°C and introduced into a Beckman LS1701 in the single photon rate counting mode. Counting period was 2 minutes.

Browning of Maillard reaction mixtures in water and DMSO

Time-dependent browning of the reaction mixtures was assayed by absorbance measurements with a Hewlett-Packard diode array spectrophotometer HP 8451A. 40mL of distilled water or DMSO (60 mmol/L aminocaproic acid, 60 mmol/L D- ribose) were heated at 100°C for 180min and the absorbances of the solutions at 420 nm (A420) against solvent blanks were determined at different times. Samples were diluted if A4*,, exceeded 1.

Photography of c hemi luminescence

D-Ribose and 6-aminocaproic acid (0.5 g each) were dissolved in 50mL of DMSO in an 100-mL Erlenmeyer flask and heated at 120°C on a mag- netic stirrer with electric heater. Oxygen was sup- plied continuously to the reaction mixture through a Pasteur pipette (1000 mL/min). After the onset of visible CL (after about 3 min) a photograph was taken on Konika colour SR-G 3200 with 15 min of exposure. The camera was a Canon Al. The film was developed for high-sensitivity (ASA 6400). Peak intensity of this light was determined after introduction of 0.2mL aliquots of the reaction mixture into a Dynatech Microlite2 plate mea- suring in an Anthos lucyl luminometer with l s

LIGHT FROM MAILLARD REACTION 279

integration time on the photomultiplier tube. Light intensity, in watts was calculated according to the supplier’s recommendation. (60 counts/second equal lOOaW). For calibration of the photon flux a standard plate from Biolink, Austria was used.

Fluorescence spectra

Fluorescence excitation and emission spectra of chemiluminescent Maillard reaction mixtures were measured at room temperature using a Shimadzu spectrofluorophotometer Model RF- 5000. Reaction conditions were the same as for photography.

Chemiluminescence emission spectrum

The CL emission spectrum in the visible range from Maillard reactions was recorded by means of a monochromator (Jobin Yvon, HR 320)/microchannel plate photomultiplier (Hamamatsu, C3360, 3.1 kV) arrangement; spectral resolution was 25nm. Reaction conditions were the same as for photography.

RESULTS

Visual detection and measurement of CL

Visual adaptation to complete darkness for at least 15 minutes was essential for visual detection of CL from MR. Visual detection of CL from Maillard reaction between D-ribose and 6-aminocaproic acid in water at pH 11 was possible and even surface phenomena could be observed: unless the reaction vessel was vigorously shaken or oxygen bubbled through the reaction mixture, only the surface of the reaction liquid glowed weakly under air.

After changing the reaction solvent from water to pure DMSO CL was dramatically enhanced as judged by visual detection. Addition of base to the reaction mixture in DMSO gave no further enhancement of CL. For visual detection of CL from the model reaction between 6-AC and D- ribose lOmg each in 5mL DMSO were sufficient. Even upon stopping the oxygen and heat supply after induction of CL in DMSO the visible light persisted for several minutes. Browning of the reaction mixtures in DMSO was very much enhanced (Fig. 1). Especially noteworthy is that all reaction

mixtures producing CL developed a strong browning during the reaction. Reaction in DMSO did not change requirements for light production from the Maillard reaction: only Maillard reactions with reducing sugars produced detectable light, whereas reaction under nitrogen did not and sul- phur compounds (e.g. cysteine or acetylcysteine) immediately blocked CL after their addition to the reaction mixture. Visual detection and measurement of CL by LSC were in full concordance (Table 1). Remarkably, amino-carbonyl reactions of acetaldehyde, butyraldehyde, and 2-deoxy-~- ribose in neutral DMSO turned out to be chemiluminescent as proved by visual detection and photon counting (Table 1); intense browning occurred in parallel with CL.

Photography

Photography was done with the reaction in neutral DMSO. A highly sensitive colour film was chosen according to the broad emission spectrum of the reaction and developed with high sensitivity (ASA 6400). Fig. 2 shows a photography of CL from the reaction of D-ribose and 6-aminocaproic acid (0.5 g each) in DMSO (40 mL), at 120°C with continuous oxygen supply (1000 mL per minute). Peak light intensity was 54fW corresponding to a photon flux of 4,800,000 photons/second/sterra- dian. Visible CL could be detected during the first 15 minutes of this reaction; after that time CL dis- appeared. Exposure time as long as this CL-phase of the reaction was chosen.

*I-

t [min]

Figure 1. Browning of a Maillard reaction mixture (D-ribose and 6-AC, 60mmol/L each) in neutral water or DMSO at 100°C. measured at 420 nm (A420)

280 G. WONDRAK ETAL.

Table 1. Visual perception and measurement of chemiluminescence from different amino-carbonyl reaction mixtures in DMSO

Reactants Visible CL' Measured CL (cpm)t

- - 33 D-G lucose - 34 - 6-AC - 1341 D-G lucose 6-AC + 51,902 D-Mannitol 6-AC - 802 D - Maltose 6-AC + 28,773 D-Trehalose 6-AC - 2721 D-Ribose 6-AC ++% 44,383% D- Ribose 6-AC + L-cysteineS -% 382% 2- Deoxy-D-ribose 6-AC + 16,550 Acetaldehyde 6-AC ++ 937,109

70 Acetaldehyde -

Butyraldehyde 6-AC + 54,132

" l 0 0 m g each in 5mL DMSO at 95°C for at least 3 minutes and subsequent O2 supply for 5 seconds. t 100 mg each in 5 mL DMSO a t 95°C for 3 minutes under air; Beckman LS 1701, single photon rate counting (after cooling to 60°C). SlOmg each. § 20 mg.

-

-

-

Figure 2. Photograph of CL from Maillard reaction between D-ribose and 6-AC in DMSO with continuous oxygen supply


rel. intensity

400 450 500 550 600 650 700 750 780

l. “4

Figure 3. (A) CL emission spectrum after 3 minutes of reaction between D-ribose and 6-AC in DMSO (conditions as in Fig. 2). (6) Fluorescence spectra after 3 minutes of reaction between D-ribose and 6-AC in DMSO, recorded at room temperature (conditions as for photography); for the emission spectrum on the right, excitation was 422 nm, for the excitation spectrum on the left, emission was 495 nm

At the beginning the reaction mixture was quite clear and then turned dark brown during the CL- phase.

CL emission spectrum

The corresponding CL emission spectrum is presented in Figure 3(A). Two broad emission max- ima at about 500 nm (blue-green) and 695 nm (red) with different relative intensities coexist.

Fluorescence spectra

The fluorescence excitation/emission spectra of the standard reaction mixture (D-ribose and 6-aminocaproic acid) in DMSO are shown in Fig. 3(B). Apparently, the fluorescence emission spectrum (excitation at 422 nm, broad emission maximum at 495 nm) matched the blue-green part of the CL emission spectrum.

DISCUSSION

Visual perception of light from amino-carbonyl reactions between reducing sugars and amino

acids is possible either in basic water or neutral DMSO. Full adaptation to complete darkness for at least 15 minutes was absolutely crucial for visual perception. All attempts to photograph the visible CL from MR in water failed, but were successful in DMSO. Reaction in DMSO seems to follow the principal pathways of the Maillard reaction in aqueous solution as suggested by the GC-MS identity of characteristic pentan/ether extractable products in both reaction solvents (data not shown). Solvent effects on quantum eff- ciencies have been described for many CL reactions. In the case of Maillard reactions, DMSO- dependent amplification may be due to:

(a) acceleration of dehydration and ketoenoltau- tomerization during early Maillard reactions (e.g. the initial Schiff base formation is based on a reversible dehydration step) rapidly speed- ing up the browning of the reaction mixture compared to reaction in water (Fig. 1);

(b) hydrophobic caging of the emitting species pre- venting quenching of the excited state;

(c) allowance of higher reaction temperature.

The CL of the reaction of 2-desoxyribose, acetaldehyde and butyraldehyde strongly suggests that the presence of an a-hydroxycarbonyl group is not essential for CL from amino-carbonyl reactions and therefore raises the question whether the phenomenon of CL from Maillard reactions is only dependent on a Schiff base structure that interacts with oxygen; in this sense, sugars contribute to the visible CL phenomenon owing to their carbonyl group but not owing to their specific a- hydroxycarbonyl structure.

As to the nature of the light-emitting species gen- erated during Maillard reactions at least three reaction mechanisms may be considered (Scheme 1):

(1) Schif base CL after oxidative formation of an intermediate dioxetan. McCapra et al. (1 0,ll) showed that the oxidation of certain aromatic Schiff bases in strongly basic DMSO (e.g. from 9-aminoanthracene and isobutyralde- hyde) can be a very efficient chemiluminescent reaction with the resulting aromatic fluorescent formamido-anion being in the light-emitting excited state.

In the case of the Maillard reaction with aliphatic amines, CL could occur after energy transfer from the excited aliphatic formamido-anion to an adjacent fluorescent molecule (e.g. an aromatic heterocyclic browning


4 0 OH

H,N-R'

T H20 OH

Schiff base

11 OH

I excited state formation I chemiluminescence

energy- transfer

H

+y /~ ' - OH H

Scheme 1. Proposed oxidative pathways during Maillard reactions leading to excited state formation and chemiluminescence: (1 ) Schiff base CL, (2) CL from Schiff base induced peroxidative decay of the sugar, (3) formation of stable pyrazinium radical cation systems (R , = sugar residue, R2 = residue of amino acid)

product). The fact that even simple aliphatic aldehydes and 2-desoxy-~-ribose reacting with amines give rise to visible CL (Table 1) and browning strongly supports the relevance of this mechanism, but may not exclude a sugar (i.e. a-hydroxycarbonyl) specific contribution to the generation of CL during Maillard reactions. Furthermore, the finding that CL is not enhanced after addition of base to the neutral reaction mixture with reducing sugars in DMSO demonstrates that classical Schiff base CL does not account for all characteristics of the Maillard reaction related CL phenomenon.

(2) CL from Schif base promoted peroxidative decay of the sugar by radical mechanism. The pathway may be regarded as homologous to the radical autoxidation of unsaturated fatty acids. Addition of oxygen to a Schiff base sta- bilized radical during Maillard reactions in neutral DMSO may produce sugar peroxides with formation of excited carbonyl products upon decay. Kurosaki et al. (1) suggested formation of a carbon radical followed by formation of the corresponding peroxyl radical to be

the key steps for generation of excited states during Maillard reactions. Additionally, CL from electrochemical oxidation of polyols in alkaline solution without involvement of amines has been described by Karatani (12). Consistently, we found that CL from sugar alone is easily visible, when the sugar is heated in very basic DMSO (data not shown). How- ever, if the amino-compound is absent, neither sugar nor polyalcohol alone give rise to CL during heating in neutral DMSO under oxygen (Table 1).

( 3 ) Radical ion CL after generation of stablepyrazi- nium radical cation systems. ESR studies by Namiki and Hayashi (5 ) revealed the formation of a pyrazinium radical cation during the early stage of MR and this structure has been proposed to be the key intermediate for generation of oxygen-dependent CL. The hypotheti- cal pathway to a lucigenin-like oxygen- dependent CL starts from dimerization of the initially formed enaminol tautomer of the Schiff base; the 8-T-electron containing dihy- dropyrazine stabilizes by a single electron


\’ H,C-d .N-W,

I 1 H,C CH,

CH,

A 0 C

Scheme 2. Structural analogy between radical cation systems of oxygen-dependent chemiluminescent molecules: (A) Iuci- genin, (B) tetrakis-(dimethylamino)-ethylene, (C) Maillard reaction derived N,N’-disubstituted pyrazinium system. ( R , = sugar residue, R2 =residue of amino acid)

oxidation step leading to the stable pyrazinium radical cation. This may be regarded as a simple single ring analogue of the lucigenin radical cation (a N,N’-dimethylbisacridinium structure-Scheme 2), reacting with the super- oxide radical anion to form a dioxetan (13). Upon decay this could then yield to formation of excited species and-after energy transfer- to CL. Interestingly, another well-known autoxidation-dependent chemiluminescent molecule with close resemblance to the Mail- lard reaction derived pyrazinium radical cation is tetrakis-(dimethylamino)-ethylene (14). As for lucigenin the crucial structure is the formation of a radical cation system (Scheme 2).

Owing to the intrinsic heterogeneity of the Mail- lard reaction with a whole range of reaction pathways occurring simultaneously leading to extremely complex mixtures of distinct compounds and polymers, an unequivocal identification of the chemiluminescent pathway is complicated. In general terms, oxidative generation of excited states followed by energy transfer to strongly fluorescent heterocyclic compounds formed during the browning of the amino acid-sugar mixture seems to be the most probable reaction mechanism (Scheme 1). Consistently, all studied chemiluminescent reaction mixtures strongly exhibited browning, whereas non-chemiluminescent mixtures did not. Involvement of energy transfer to fluorescent molecules of advanced stages of Maillard reactions is supported by spectroscopic evidence: fluorescence emission spectra and CL emission spectra are very similar on the blue-green part of the spectrum. However, the molecular origin of the less

intense red emission at about 700nm remains to be clarified,

The definite evaluation of the relative contribu- tions of the discussed reaction mechanisms leading to the remarkable phenomenon of visual detection of CL from Maillard reactions remains to be done by future experiments. Organic synthesis of sus- pected chemiluminescent key intermediates of Maillard reactions without disturbance by the het- erogeneous fluorescent matrix is in progress. Other experiments arising from these CL studies will address the role of Maillard reactions between peptide bound amino groups and reducing sugars under physiological conditions as an endogenous source of excited states in the living cell.

Acknowledgements

Parts of this work were presented as a poster at the 8th International Symposium on Bioluminescence and Che- miluminescence, University of Cambridge, England, 5- 8 September 1994.

The authors wish to thank Prof F. McCapra and Dr E. A. Chandross for helpful discussion of the poster.

REFERENCES

Kurosaki Y, Sat0 H, Mizugaki M. Extra-weak chemiluminescence of drugs. VIII. Extra-weak chemiluminescence arising from the amino- carbonyl reaction. J Biolumin Chemilumin

Kurosaki Y, Sat0 H, Ishizawa F, Mizugaki M. Extra-weak chemiluminescence of drugs. XII. Effect of the molar ratio of amino acid to

1989;3: 13-9.


sugar on extra-weak chemiluminescence in the Maillard reaction. J Biolumin Chemilumin

3. Tressl R, Kersten E, Rewicki D. Formation of pyrroles, 2-pyrrolidones, and pyridones by heating of 4-aminobutyric acid and reducing sugars. J Agric Food Chem 1993;41:

4. Tressl R, Wondrak G, Kersten E, Rewicki D. Structure and potential crosslinking reactivity of a new pentose specific Maillard product. J Agri Food Chem 1994;42:2692-7.

5. Namiki M, Hayashi T. Development of novel free radicals during the aminocarbonyl reaction of sugars with amino acids. J Agri Food Chem 1975;23:487-91.

6. Koenig RJ, Peterson CM, Kilo C, Cerami A, Williamson JR. Hemoglobin A,, as an indica- tor of the degree of glucose intolerance in diabetes. Diabetes 1976;25:230-2.

7. Ruderman NB, Williamson JR, Brownlee M. Glucose and diabetic vascular disease.

8. Namiki M, Oka M, Otsuka M, Miyazawa T, Fujimoto K, Namiki K. Weak chemilumines-

1991;6:185-8.

2 125-30.

FASEB J 1992;6:2905-14.

cence at an early stage of the Maillard reaction. J Agric Food Chem 1993;41:1704-9.

9. Harris DA. ‘Luminescent’ rate assays. In: Har- ris DA, Bashford CL, editors. Spectrophoto- metry and spectrofluorimetry: a practical approach. Stadt: IRL Press, 1987:73-7.

10. McCapra F, Burford A. Chemiluminescent Schiff bases. JCS Chem Comm 1976;607-8.

11. McCapra F, Burford A. Mechanism of Schiff base chemiluminescence: evidence for an intermediate dioxetan. J C S Chem Comm

12. Karatani H. Polyol chemiluminescence induced by mediated oxidation. J Biolumin Chemilumin 1994;9:292. [abstract]

13. Janzen EG, Pickett JB, Happ JW, DeAngelis W. Electron spin resonance studies of radical formation in nucleophilic addition reactions. I. Radical formation and chemiluminescence in the hydroxide ion addition to N,N’- dimethyl-9,9’-biacridinium dinitrate (lucigenin). J Org Chem 1970;35:88-95.

14. Urry WH, Sheeto J. The autoxidation of tetrakis-(dimethylamino)-ethylene. Photochem Photobiol 1965;4: 1067-83.

1977;874-6.

light from maillard reaction: photon counting, emission spectrum, photography and visual perception

Documents