Mutanon Research, 230 (1990) 263-272

263

Elsevier MUT 04872

Effects of monosaccharides and disaccharides on the formation of food mutagens in model systems Kerstin Skog and Margaretha J~gerstad Food Chemtstry, Chemtcal Center, Umverstty of Luna~ S-221 O0 Lund (Sweden)

(Received 13 November 1989) (Revision received2 February 1990) (Accepted 5 February 1990)

Keywords" Food mutagens; Ames test; Model systems; MelQx, DtMeIQx; Creatm(m)e; Glycme; Hexoses

Summary The formation of the mutagenic imidazoquinoxalines (MeIQx, DiMeIQx) was studied using a modification of a previous model system. Creatine or creatinine (0.9 mmole) was heated together with glycine (0.9 mmole) and various sugars (0.45 mmole) dissolved in diethylene glycol and water (3 nil, 5 : 1) for up to 15 rain at 180°C. This system produced the same amount of mutagenicity after 10 min at 1 8 0 ° C as a previous one during 2 h of reflux boiling at 128°C. MeIQx (4 n m o l e / m m o l e creatin(in)e) was the major mutagen produced together with minor amounts of DiMeIQx, both 4,8- and 7,8-DiMeIQx according to HPLC-MS. A few other mutagenic peaks were also separated on HPLC, but they were not identified. Varying the concentration (0-2.4 mmole) and type of monosaccharides and disaccharides greatly affected the yields of all the mutagenic compounds. Sugar in molar amounts lower than the creatin(in)e concentration increased the yield until an optimum was reached. In higher concentrations the formation of all the mutagens was markedly reduced. The same was found for glucose, fructose, sucrose, and lactose, though the monosaccharides showed the most pronounced inhibitory effects. The inhibition of the formation of the mutagenic compounds by an excess of sugars is proposed to be an effect of Malllard reaction products, which may block the formation of imidazoquinoxalines by attacking creatine. Support for this mechanism is given by data showing a lower recovery of unreacted creatine with increasing concentration of glucose and also by an inhibitory effect on the formation of these mutagens after adding a typical Maillard reaction product, 5-hydroxymethyl-2-furfural.

A bbremattons and CA S registry numbers: IQ, 2-amino-3-methyl-

lmidazo[4,5-f]quinolme (76180-96-6); MeIQ, 2-amlno-3,4-dlmethylimidazo[4,5-f]qumoline(77094-11-2); MeIQx, 2-amino3,8-dimethylirmdazo[4,5-f]quinoxaline (77500-04-0); 4,8-DiMeIQx, 2-amino-3,4,8-trimethylimidazo[4,5-f]quinoxaline (95896-78-9); 7,8-DiMeIQx, 2-amino-3,7,8-trimethylinudazo-

[4,5-f]qtunoxahne; TMIP, 2.anuno-n,n,n-tnmethyhrmdazo[4,5-b]pyrichne; PhIP, 2-anuno-l-methyl-6-phenyhnudazo[4,5b hayndine (105650-23-5). Correspondence: KersUn Skog, Food Chemistry, Chemical Center, P.O. Box 124, S-221 00 Lund (Sweden).

0027-5107/90/$03.50 © 1990 Elsevier Science Pubhshers B.V. (Biomedical Divlslon)

264

The major mutagens isolated to date from heat-treated meat and fish products belong chemically to one class of heterocyclic amines, which all contain an imidazo group with an amino moiety in the 2-position. The imldazo group is linked to either quinoline (IQ, MelQ), quinoxallne (MelQx, DiMelQx), or pyridine (TMIP, PhlP) (Sugimura and Sato, 1983; Felton and Hatch, 1986). Based on the molecular structure of these mutagenic compounds, we postulated a route for the formation of aminoimidazoquinolines and armnoimidazoquinoxalines some years ago (J~igerstad et al., 1983). The compounds were assumed to be formed from 3 precursors naturally occurring in foods of animal origin: creatine, free amino acids, and monosaccharides. Creatine was assumed to be transformed into creatinine by heating, and forms the imidazole moiety of the molecule. The qulnoline or quinoxaline part was suggested to arise from Maillard reacuon products, especially pyndines or pyrazines and Strecker degradation products, like aldehydes. By heating creatine or creatlnine, hexoses, and certain free amino aods in a model system, several of the mutagens such as MelQx, DiMelQx, MelQ, and IQ were isolated and identified (for references, see Jagerstad et al., 1989). Also in meat-cooking experiments, we have shown a close relationship between the concentrations of the 3 precursors, especially creatine, and the amount of mutagenic activity produced (Laser Reutersw~ird et al., 1987a,b). The significance of creatine for the formation of meat crust mutagens has also been shown by other groups (Miller, 1986; Nes, 1986; Becher et al., 1988; Knize et al., 1988). However, the mechanisms by which these mutagenic compounds are formed need further studies. According to our hypothesis the Maillard reaction ts involved because glucose and amino acids are necessary as precursors. On the other hand, Yoshida et al. (1984) have reported the formation of IQ after dry-heating creatine and proline without glucose or any other sugar at 180°C for 1 h. The formation of IQ and PhlP after dry-heating phenylalanine and creatine, also without glucose, at 200 ° C for 1-2 h, was reported by Taylor et al. (1987). The possibility of alternative routes therefore cannot be ruled out.

In the present study, our previous model system has been modified to an open-air system working at 180°C, to simulate normal cooking conditions. The mutagens produced with the modified system have been compared both quantitatively and qualitatively with those produced in the previous system. In addition, the effects of the Maillard reactions on the yield and speoes of mutagens produced have been investigated by adding monosaccharides and disaccharides in different concentrations to the model system. Also, the creatln(in)e recovery was studied to learn to what extent creatme or creatinine reacted m the model system and how the sugars, the key Maillard reactants, affected its participation. Because creatme or creatimne is assumed to form the 2-aminoinudazo part of the molecule, one way to block the formation of these mutagens might be to decrease the availability of creatine or creatmine for these reactions. Materials and methods

Chemtcals All commercially available chermcals were of analytical grade. Diethylene glycol was purchased from Kebo (Stockholm). Other solvents e.g., acetone and HPLC buffers of Llchosorb grade, were purchased from Merck. The material used for chromatographic purification was XAD-2 from Bio-Rad (Richmond). Blue cotton was obtained from Funakoshi Pharmaceutical Co. (Tokyo). Synthetic mutagens IQ, MelQx and DiMelQx were kindly supplied by Professor Kjell Olsson and Dr. Spiros Grivas, Department of Chemistry and Molecular Biology, Uppsala (Sweden). Synthetic PhlP was a kind gift of Dr. Mark G. Knize, Lawrence Livermore National Laboratory, University of California (U.S.A.). Model system The precursors - creatlne, glucose and glyclne - were dissolved in 3 ml diethylene glycol containing 0.5 ml distilled water. Several test tubes were heated simultaneously in a thermostated metal block. The water was allowed to evaporate during heating. The temperature in the reaction mixture was monitored by a thermocouple inserted in the sample and connected to a recorder.

265 In the first set of experiments the reactants creatine or creatinine (0.9 mmole), glucose or fructose (0.45 mmole), and glycine (0.9 mmole) were heated for 5, 10, or 15 min at 140, 160 or 180 ° C. In the other experiments the reactant mixtures were heated to 180°C for 10 min. The amount of glucose ranged from 0 to 2.4 mmole, while the amounts of creatine and amino acid were kept at 0.9 mmole, as above. In some experiments, glucose was exchanged for fructose, lactose, or sucrose (range 0-2.4 mmole).

nm. The flow rate was 3 ml/min and fractions were collected every 30 s and lyophilized before the assay of mutagenicity. In some experiments, where only 20-/~1 samples were injected, an analytical Nucleosil C18 column (250 mm × 4.6 mm, 5 /~m particle size) and a slightly different gradient was used. The flow rate was 1 ml/min, and fractions were collected every minute.

Mutagenictty assay

The mutagens produced in the model system were fractionated on HPLC as described above. The identities of the mutagens were established by comparing the retention times of the mutagenic peaks with the retention times of synthetic compounds, viz., IQ, MelQ, MelQx, DiMelQx, and PhlP. In addition, some samples were spiked with synthetic references before injection. Mutagenic HPLC fractions were lyopbilized and redissolved in 50 #1 methanol. Aliquots were subjected to LC-MS analysis to confirm the identities of the mutagenic compounds. The plasmaspray mass spectra were recorded on a VG Trio-3 instrument using authentic samples as references. The same analytical column and mobile phase as above were used. The column was eluted isocratically (40% B) at a flow rate of 0.9 ml/min. L C / M S measurements were made by single-ion recording (SIR) of the protonated molecular ion (MH ÷) of each mutagen. The mutagens were quantified by comparing the area of the sample with the area of a known amount of standard.

The mutagenicity of the samples was tested as described by Ames et al. (1975) using the Salmonella strain TA98 with the addition of 2 ml $9 mix containing 5% Aroclor 1254-induced rat liver per plate. The optimal concentration of $9 had previously been examined. Sample, bacteria, and $9 mix were incubated at 37°C for 20 min before adding top agar. For each determination of the heated samples, 3 different doses, run in duplicate, were used to produce dose-response curves. The number of revertants per #mole creatine was calculated from the linear part of the curve as described by Bjeldanes et al. (1982). The lyophilized HPLC fractions were tested at only one dose. Synthetic MelQx was used as a positive control.

Purification and fractionation of the mutagentcity Heated samples were diluted with 20 volumes of distilled water and subjected to a 0.9 cm × 15 cm column of XAD-2 resin. The mutagenic compounds were eluted with acetone according to Bjeldanes et al. (1982). The acetone fraction was concentrated by evaporation and further purified using blue cotton as described by Hayatsu et al. (1983). After evaporation to dryness, the residue was dissolved in 0.5 ml methanol. Aliquots (150 #1) were injected into a Varian 5000 Liquid Chromatograph equipped with a Polygosil ClS column (300 mm × 10 mm, 5 #m particle size) and eluted with a mobile phase of acetonitrile:0.5% acetic acid (60:40) (B) and 0.5% acetic acid (A). A gradient of 0-50% B over 30 rain was followed by 50% B for another 10 rain and back to 0% B in the last 10 min. The effluent was monitored at 263

Identtfication and analysis of the mutagemctty using HPLC and LC-MS

Creatm(in)e analysis Creatin(in)e means the sum of creatine and creatinine. These 2 compounds were measured separately in the samples by an enzymatic method of Wahlefeld et al. (1974). All analyses were performed in duplicate using an analytical kit purchased from Boehringer Mannheim GmbH. R ~

The model system A temperature profile of a sample heated at 180 °C for 10 rain in the modified model system,

266

203

t~°c

TABLE 1 EFFECTS OF REACTANTS, ON THE MUTAGENICITY CREATIN(IN)E, GLYCINE, IN D I E T H Y L E N E G L Y C O L

:i i

8 heatzr~ tzme (mznutes)

Fzg 1 Temperature m a nuxture of creatme, glucose, and glycme heated to 180 ° C m dlethylene glycol and water (5" 1)

now in use, is shown in Fig. 1. The temperature rose rapidly and reached 1 8 0 ° C within 5 min. Samples (creatine : glucose : glycine) heated at different temperatures for 15 rmn showed the following mutagenicity calculated as revertants per /xmole original creatine. At 1 4 0 ° C no significant mutagenicity was demonstrated; at 1 6 0 ° C the mutagenicity was only half that found at 180 ° C: 210 and 460 revertants//~mole creatine, respectively. The influence of heating time was also studied; the reactants were heated at 180 ° C for 5, 10, or 15 min. As shown in Table 1, the yield of mutagenicity after 5 min was only half that after 10 or 15 min: 220 revertants//~mole creatine compared with 440 and 460 revertants//~mole creatine, respectively. The experiment with 3 different heating times was performed with other combinations of reactants: (1) glycine, creatine, fructose or (2) glycine, creatinine, glucose or (3) glycine, creatinine, fructose. All combinations showed similar results; already after 10 min of heating, the formation of mutagenicity reached a plateau (Table 1). Creatinine produced more mutagenic activity than creatine and so did fructose as compared with glucose.

In some experiments where one of the reactants creatlne, glucose or glycine - was omitted from the reaction mixture, no significant mutagenic ac-

Ternperature

Mutagemcaty (revertants/~t mole creatme)

(°C)

5 man

Creatme/glucose/glycme

140 160 180

222

440

NS 210 459

Creattmne/glucose/glycme Creatme/fructose/glycme Creatmme/fructose/glycme

180 180 180

324 546 575

651 645 1061

702 752 977

Reactants

0

TIME AND TEMPERATURE OF HEATED MIXTURES OF AND MONOSACCHARIDES AND WATER

10 man

15 nun

NS, not significant - , not tested.

tivity was found. And heating pure diethylene glycol and water did not produce any mutagenic actiwty. So, in accordance with our previous model

FI,I

I i /ji I,

~1:%

'-o,y'; ,,j

\.

i

21XIC

S

/ 5OO

5

lO

15

2C

25

3O

reten~lon ~ime (mzn)

Fig 2 UV a b s o r p u o n at 263 n m during H P L C fracuonatlon of a heated rmxture of creatine, glucose, and glycine m diethylene glycol and water ( 1 8 0 ° C / 1 0 rmn). The lower curve is the mutagenic acuwty of the fractions. Retention times for MeIQx and 4,8-DIMeIQx are indicated by arrows

267

system (J~igerstad et al., 1983, 1984), the modified system also required all 3 precursors (glycine, creatin(in)e, and a sugar) to be present at the same time in order to produce any significant mutagenicity.

Mutagens produced m the model system An HPLC chromatogram of a mixture of creatine, glucose and glycine, heated at 180 °C for 10 rain, is shown in Fig. 2. This figure also shows the mutagenicity of the fractions tested with the Ames test. Three major mutagenic peaks were separated. Comparing their retention times with authentic samples, as indicated by the arrows, two of the peaks seemed to correspond to MelQx and 4,8-DiMelQx, respectively. Further studies comparing 100!

2:

90. 80. 20. 60. 50. 40. 30. 20. 10_ 15o

100!

eeo

250

300

zso

~00

214

90. ,30. 20. $0_ 50_ 40_ 30_

20_ 1O_ 0_ ~-,-,---.-~,.~, -r----150 200

Fig. 3. Plasma-spray mass spectra. Single-ion recording of the protonated molecular ion ( M H + ). Top: Mutagenic fraction from a heated (180 ° C / 1 0 rmn) mixture of creatme, glucose and glycme m diethylene glycol and water (5:1). Bottom: synthetic M e l Q x ( M H + ).

authentic samples and mutagenic fractions of the model system using LC-MS confirmed the presence of MelQx and DiMelQx. Fig. 3 shows the plasma-spray mass spectrum for the first major mutagenic fraction corresponding in retention time and mass spectrum to authentic MelQx. The amount of MelQx in this fraction was estimated to be about 600 ng. This fraction also contained minor amounts of 7,8-DiMelQx according to retention time and plasma-spray mass spectrum that coincided with authentic 7,8-DiMelQx. The second mutagenic fraction corresponding in retention time to authentic 4,8-DiMelQx was shown to contain 4,8-DiMelQx using LC-MS. The amount was, however, low, only about 30 ng, corresponding to 10% of the total mutagenicity of this fraction. Thus, 90% of the mutagenicity of this fraction might be another mutagenic compound(s). Judged from the retention time IQ and MelQ had retention times shorter than this peak. Authentic PhlP, on the other hand, was eluted later and might correspond to the last mutagenic fraction. This third mutagenic fraction was, however, not identified because of lack of material. HPLC fractionation and the Ames test of samples heated with fructose, sucrose or lactose gave similar results: 3 major mutagenic peaks that might contain MelQx and DiMelQx.

Effects of added sugar on the mutagemcity Usually the glucose concentration in our model system was half the molar amount of creatine and glycine. To check the effects on the yield of mutagenicity, the glucose concentration was changed. Without any glucose, no mutagenicity was observed. By increasing the amount of glucose the mutagenicity first increased, reaching a maximum when the glucose concentration, on a molar basis, was about half that of the other 2 precursors. Increasing the glucose concentration further resuited in markedly reduced mutagenicity. The inhibitory effect of high concentraUons of glucose was confirmed with HPLC fractionation and the Ames test. Fig. 4 clearly shows that high glucose concentrations yielded lower mutagenic activity in the HPLC fractions. Exchanging glucose for either fructose, sucrose, or lactose (also in increasing amounts) in the model system gave similar results; the mutagenic-

268 300 3fI[~2

~Creat

inlne

~Creat

ine

2OO 2OOO

I c

1oo u

--

I

0

GI u c o s e 150 pmo I/ml

2013

2OOO

i ||m 400

glucose concentration

600

800

10013

(pmol/ml)

Fig 5 Effects of vanous glucose concentrations on the creatm(m)e concentratxons m samples of creatme, glucose and glycme heated for 10 nun at 180°C m ðylene glycol and water (5- 1)

1000

.....

400 $mta1/mI

/

0

20

25 R e t e n t lon

t~me

....

800 pmal/ml

3~3 (mln)

Ftg 4. The mutagemc actlvaty m HPLC fracttons from creatlne, glucose and glycme heated for 10 nun at 180°C in diethylene glycol and water (5 : 1). For each HPLC run 30% of the samples was used. (a) Glucose concentration 150 pmole/ml; (b) 400/~mole/ml; (c) 800/~mole/ml.

tty first reached a p l a t e a u a n d was then reduced. But the yield of total m u t a g e n i o t y was slightly different w h e n c o m p a r e d with glucose. T a b l e 2 shows the effect of different sugars in increasing

c o n c e n t r a t i o n s o n the m u t a g e m c i t y in the model system. F r u c t o s e or sucrose p r o d u c e d more m u t a genicity t h a n glucose o n a n e q u i m o l a r basis, b u t with all the sugars there was a t e n d e n c y to i n h i b i t the f o r m a t i o n of m u t a g e n i c i t y at high c o n c e n t r a tions. T h e m l u b i t o r y effect of high sugar conc e n t r a t i o n s was a g a i n c o n f i r m e d with H P L C fract i o n a t i o n a n d the A m e s test. Sucrose consists of glucose a n d fructose m e q u i m o l a r a m o u n t s . This disaccharide i n d u c e d m u c h higher m u t a g e n i c i t y over a wide range of c o n c e n t r a t i o n s t h a n glucose, fructose or lactose. T h e i n l u b i t o r y effect of excess sucrose was therefore less p r o n o u n c e d t h a n that of the other sugars.

TABLE 2

Effects of added sugar on the creatmOn)e recove~

EFFECTS OF VARYING SUGARS ON THE MUTAGENICITY OF MIXTURES OF SUGAR, GLYCINE, AND CREATINE HEATED FOR 10 MIN AT 180 °C

C r e a t i n e is converted to c r e a t i n i n e when heated, especially at low p H (Lempert, 1959). The effect of d i f f e r e n t glucose c o n c e n t r a t i o n s o n the creatin(in)e recovery is shown in Fig. 5. After h e a t i n g creatine a n d glycine w i t h o u t glucose for 10 m i n at 1 8 0 ° C a b o u t 90% of the original creatine r e m a i n e d u n r e a c t e d . W h e n glucose was a d d e d to the m o d e l system at half the m o l a r a m o u n t of creatine a n d glycine, still a b o u t 90% of the original a d d e d creatine was recovered, m a i n l y as creatinine. W h e n glucose in m o l a r excess was a d d e d to the system, the recovery of creatin(in)e decreased markedly. F o r instance, a 3 - 5 - f o l d increase of a d d e d glucose reduced the final creatln(in)e con-

Sugar

Mutagemcity (rev//~mole creatme)

(/1mole/ml)

Glucose

Fructose

Sucrose Lactose

25 50 100 150 200 400 600 800

233 481 568 238 113 69 NS

241 530 547 399 172 171 -

172 588 847 732 1176 944 692 539

NS, not s~gruficant -, not tested

179 274 280 207 135 -

269 TABLE 3 EFFECT OF ADDED HMF TO MIXTURES OF GLYCINE (300/~mole/ml), CREATININE (300/~mole/ml) AND GLUCOSE (150 /~rnole/ml) ON THE MUTAGENICITY AND CREATIN(IN)E RECOVERY AFTER HEATING FOR 10 MIN AT 180 o C IN DIETHYLENE GLYCOL AND WATER (5 : 1) Added HMF /xmole/ml

~ a

Mutagemoty (rev/#mole creatme)

Analyzed creatin(in)e (/~mole/ml)

0 3 30 75 150

0 1 10 25 50

438 361 279 337 104

246 249 260 233 174

a Percent of added creatlne

centration by corresponding figures. An excess of fructose, lactose or sucrose in the model system also showed reducing effects on the creatin(in)e recovery. These results suggest that glucose itself or, more likely, some Maillard reaction products may combine directly with creatine or creatinine, and that such a reaction may be competitive with the reaction forming the mutagenic compounds. One Maillard reaction product, 5-hydroxymethyl-2-furfural (HMF), also showed some inhibitory effect when added to the model system at concentrations corresponding to 1, 10, and 25% of the added creatine. But when HMF at half the molar amount of creatine was added to the model system, the original mutagenicity was reduced by 75% (Table 3). (Adding corresponding amount of HMF showed no effect in the Ames test.) Also the creatin(in)e recovery was markedly reduced using 50% HMF, while HMF at the other concentrations showed no such effect. Discussion

The modified model system used in the present study is very versatile. All the physical parameters such as time, temperature, water activity, and pH - can easily be changed to simulate various cooking conditions. The modified system allows the water to evaporate, which also occurs in the crust during the frying of foods. Moreover, inter-

actions by components of enhancing or inhibiting capacity can easily be evaluated in this new system. The use of diethylene glycol mixed with water makes it possible to study the reaction in temperatures between 100 and 245°C (boiling point of diethylene glycol). All the reactants are easily dissolved in this mixture, and we have not been able to show that diethylene glycol contributes to any formation of mutagenicity. Diethylene glycol of the highest purity was used and heated mixtures of diethylene glycol and water alone or together with binary combinations of the reactants have not been shown to produce any significant mutagenicity. Because cooking times are sometimes very short, especially during frying, it is an advantage that the model system generates mutagenicity within 5 rain. With increasing time the mutagenicity increased but was about the same after 10 min and 15 rain. Such plateaus have been shown previously in cooking experiments on hamburgers (Spingarn and Weisburger, 1979; Pariza et al., 1979; Felton and Hatch, 1986). There are several explanations for this phenomenon: volatilization or decomposition of the mutagenic compounds, insufficient water supply, or other blockings of the reactions that produce the mutagenic compounds. The amount and type of mutagens produced in the modified system are similar to those obtained with our previous model system. By heating creatine, glucose, and glycine for 2 h at 128 °C in the previous system, 4.2 nmole MelQx per mmole creatine and 0.4 nmole 7,8-DiMelQx per mmole creatinine were produced. The modified model system produced a similar amount of MelQx (4 nmole) per mmole creatine, and also minor amounts of DiMelQx, about 10% of 4,8-DiMelQx and 5% of 7,8-DiMelQx according to HPLC-MS. When Grivas et al. (1985) heated creatine, glycine, and fructose at 128°C for 2 h, 6.6 nmole MelQx/mmole creatine was isolated together with 1.0 nmole IQ/mmole creatine. In this experiment, fructose was used instead of glucose, which might have enhanced the yield (see also Table 1) and caused the formation of IQ instead of DiMelQx. However, there are also mutagenic compounds in our modified model system that we have not identified yet. The HPLC fractionation indicated 3 major mutagenic peaks of which the first was

270

clearly shown to contain MelQx and traces of 7,8-DiMelQx. The next mutagemc peak contained 4,8-D1MelQx, accounting for only 10% of the mutageniclty of this fraction. The other part of this peak, as well as the third mutagenic peak, has not been identified yet. The present study implies that the Matllard reaction can both enhance and inhibit the formation of mutagenic compounds. According to our originally postulated mechanism for the formation of imidazoquinohnes and lmidazoquinoxalines, the Maillard reaction forms products such as pyrazines or pyridines and Strecker aldehydes that condense to vlnylpyridines or vinylpyrazlnes (J~igerstad et al., 1983). Aldehydes and vlnylpyridines or vinylpyrazines might then condense with creatinine and subsequently undergo ring closure, water elimination, and aromatization yielding lmidazoquinolines or imidazoquinoxalines. However, we and others have failed to produce IQ compounds by reacting vinylpyridines or vmylpyrazmes together with creatinine. Another possible route suggested by Nyhammar (1986) is that creatlmne first undergoes an aldol condensation w~th Strecker aldehydes before subsequent conjugated addition of pyridine or pyrazme. A similar route has also been suggested by Jones and Welsburger (1988), who reported formation of a mutagenic compound, 2-amino-5-ethylidene-l-methylimidazol-4-one, by reflux boiling creatinine with either acetaldehyde or threonine in a water-diethylene glycol mixture. Threonine was assumed to decompose to acetaldehyde or lactic aldehyde durmg the reaction (150 ° C / 2 h). The condensation between aldehyde and creatmine was shown to be Inhibited by adding indohc compounds, especially tryptophan, which was proposed to react with the aldehyde in competition with creatme. We report in the present study that mcreaslng concentrations of both monosaccharides and &saccharides showed a curve that peaked regarding the amount of mutagenic compounds produced. The monosaccharides in excess seem to be more powerful inhibitors than the disaccharides. The mechamsms behind the inhibitory effects are not clear, but studies on creatinine recovery indicated a decrease with increasing amounts of added glucose. This effect required the presence of glycine, which implies that Maillard reactions are in-

volved. When glucose was present m half the molar amount of creatlne or creatinine, almost 90% of creatinine was still recovered after being heated together with glycine. This means that only a very small part of creatin(in)e takes part in the formation of the IQ compounds. Also in meat crust, it is possible to measure creatin(in)e concentrations after frying, in&cating that only a small part is revolved in the formation of mutagens (Laser Reutersw~ird et al., 1987b). By increasing the concentration of monosaccharides and disaccharldes as compared with the amino acids, other Maillard reaction products might have been produced. One common product is 5-hydroxymethyl-2-furfural (HMF). Adding this compound to the model system decreased both the creatinine recovery and the mutageniclty produced. Creatmine is known to react readily with aromatic aldehydes (Lempert, 1959). H M F is both a common Maillard reaction product and an aromatic aldehyde. It was, however, required m almost half the molar amount of creatine to block mutagen formation. If the blocking effect of the Maillard reaction is directed towards creatme, all the mutagenic peaks should contain an lmidazo group, because all the mutagenic peaks separated on HPLC decreased in the presence of increasing amounts of monosaccharides and disaccharides. Although monosaccharides and disacchandes m excess inhibited the formation of IQ compounds, they seem to be required in certain amounts to make mutagen formation possible in our model system and also in pan-fried beef (J~igerstad et al., 1983). The role of reducing sugars in the formation of imidazoquinohnes or imldazoquinoxalines is probably catalytic, which means that it might be possible to produce IQ compounds in the absence of reducing sugar if the right cooking time and temperature are used. Those groups that have reported formation of mutagemc irmdazoazaarenes without sugar have dry-heated the reactants for 1-2 h between 180 and 2 0 0 ° C (Yoshida et al., 1984; Taylor et al., 1987, 1988). In our more aqueous and sugar-containing system, the formation of mutagens occurred rapidly and already peaked within 10 min. (Cooking of hamburgers is generally performed by pan-frying for 6-18 min at 180-250 ° C.) Taylor et

271

al. (1987) dry-heated creatine and phenylalanine with and without glucose at 200 °C and mutagenicity appeared within 10 min but the yield first peaked 1-2 h later. Interestingly, meat muscle may contain half to equal molar amounts of monosaccharides, free or phosphate-bound, compared with creatine (Fabiansson and Laser Reutersw~ird, 1985; Laser Reutersw~d et al., 1987), which means that the conditions in meat with regard to the precursors favor mutagen formation. Liver, however, contains glucose in marked excess and extremely low levels of creatine; and, as we have already reported, frying of liver produces almost negligible amounts of mutagenicity. The fact that sugars in excess are able to limit the formation of mutagens could be exploited in certain meat products where carbohydrates are added to the recipe. Meatballs, a popular meat dish in Sweden, are prepared by mixing minced beef with potato starch, milk powder (lactose), or golden bread crumbs. Work is now in progress to study the formation of mutagens in meat products containing ingredients rich in starch and carbohydrates.

Acknowledgements We thank Ms Marie Birger, Department of Food Chemistry, for excellent technical assistance. The gifts of synthetic compounds from Professor Kjell Olsson, and Dr Spiros Grivas (Uppsala, Sweden), and Dr Mark Knize (Livermore, CA, U.S.A.) are gratefully acknowledged. Salmonella typhimurium, strain TA98, was kindly provided by Professor Bruce Ames, University of California (Berkeley, CA, U.S.A.). This study was supported by the Swedish Cancer Foundation (1824-B8806XB) and the Swedish Council for Forestry and Agricultural Research (855/88L-96:2).

References Ames, B.N, J. McCann and E. Yamasakl (1975) Methods for detecting carcinogens and mutagens with the Salmonella/ mammahan-microsome mutagenicity test, Mutation Res., 31, 347-364 Becher, G., M.G. Kmze, I.F. Nes and J.S. Felton (1988) IsolaUon and tdenUfication of mutagens from a fried Norwe~an meat product, Carcinogenesis, 9, 247-253

Bjeldanes, L.F., K G. Rose, P.H. Davas, D.H. Stuemer, S.K. Healy and J.S. Felton (1982) An XAD-2 resin method for effioent extraction of mutagens from fried ground beef, Mutation Res., 105, 43-49. Fabiansson, S., and A Laser Reutersward (1985) Low voltage electrical stimulation and post mortem energy metabohsm in beef, Meat Scl, 4, 205-223. Felton, J , and F Hatch (1986) Mutagens m cooked foods, Energy Tech Rev, Jan., 1-15 Felton, J.S., M.G Kmze, N.H. Shen, R Wu and G Becher (1988) Mutagemc heterocychc lnudazoamlnes in cooked foods, m: C M lOng, M S. Romano and D. Schuetzle (Eds.), Carcmogemc and Mutagemc Responses to Aromatic Attunes and Nltroarenes, Elsevier, Amsterdam, pp 73-85. Hayatsu, H , Y Matsul, Y Ohara, T. Oka and T. Hayatsu (1983) Characterization of mutagemc fractions m beef extract and m cooked ground beef. Use of blue-cotton for efficient extraction, Gann, 74, 472-482 J~igerstad, M., A Laser Reutersward, R Oste, A. Dahlqvlst, S Gnvas, K. Olsson and T Nyhammar (1983) Creatlmne and Maallard reaction products as precursors of mutagemc compounds formed in fried beef, in G.R Waller and M S. Feather (Eds), The Madlard Reaction in Foods and Nutrition, Symposium Series 215, Waslungton DC, pp. 507-519. Jagerstad, M , K. Olsson, S. Gnvas, C Neglshi, K Wakabayashi, M Tsuda, S Sato and T Suglmura (1984) Formation of 2-anuno-3,8-dimethyhnudazo[4,5-f]qumoxahne in a model system by heating creaumne, glyclne and glucose, Mutation Res., 126, 239-244 Jagerstad, M., K Skog, S. Gnvas and K Olsson (1989) Mutagens from model systems, in M W. Panza (Ed), Mutagens and Carcinogens in the Diet, Proceedings of a Satelhte Symposium of the Fifth International Conference on Enwronmental Mutagens, in press. Jones, R.C., and J.H Weisburger (1988) Intubttion of amanotmadazoqtunoxahne-type and amlnomudazol-4-onetype mutagen formation m liquid reflux models by L-tryptophan and other selected mdoles, Gann, 79, 222-230 Kmze, M G , N.H. Shen, S.K Healy, F.T. Hatch and J.S. Felton (1987) The use of bactenal mutation tests to morntot chermcal isolation of the mutagens in cooked foods, Dev Ind Mlcrobiol, 28, 171-180. K_mze, M.G., N H. Shen and J.S. Fetton (1988) A comparison of mutagen production m fried-ground chicken and beef effect of supplemental creatme, Mutagenesis, 3, 503-508 Laser Reutersward, A., K Skog and M. J~igerstad (1987a) Effects of creatine and creatimne content on the mutagemc acuvlty of meat extracts, bouillons and gravaes from different sources, Food Chem Toxacol, 25, 747-754 Laser Reutersw~ird, A., K Skog and M. J~gerstad (1987b) Mutagemc~ty of pan-fried bowne tissues in relation to their content of creatme, creatmme, monosacchandes and free anuno acids, Food Chem. Toxacol, 25, 755-762 Lempert, C (1959) Chermstry of the glycocyanudmes, Chem. Rev., 59, 667-736. Miller, A.J (1986) Precursor effects on lmldazoquinohne-type mutagen formation in heated model and muscle systems,

272 m. I B. Knudsen (Ed.), Genetic Toxacology of the Diet, Llss, New York, p 320 (Abstract P15) Muramatsu, M., and T Matsusluma (1985) Formation of MelQx and 4,8-D1MelQx by heating mixtures of creatmlne, amino acids and monosacchandes Selected abstracts of papers presented at the 13th annual meeting of the Environmental Mutagen Society of Japan, 12-13 October, 1984, Tokyo, Mutation Res, 147, 266 Nes, I.F (1896) Mutagen formation in fried meat emulsion contammg various amounts of creatme, Mutation Res, 175, 145-148 Nyhammar, T (1986) Studies on the Mafllard Reaction and its Role in the Formation of Food Mutagens, Doctoral Thesis, Swedish University of Agrtcultural Sciences, ISBN 91-5762658-8, pp. 23-29 Panza, M.W, S.H. Ashoor, F S Chu and D B. Lund (1979) Effects of temperature and time on mutagen formation in pan-fried hamburger, Cancer Lett, 7, 63-69 Splngarn, N E., and J.H. Weisburger (1979) Formation of mutagens in cooked food I Beef, Cancer Lett, 7, 259-264

Sug~mura, T., and S Sato (1983) Mutagen-carclnogens m food, Cancer Res (Suppl), 43, 2415s-2421s Taylor, R T., E Fultz and M Knlze (1985) Mutagen formation m a model beef boihng system III Purification and ~dent~ficat~on of three heterocychc amine mutagens-carclnogens, J Enxaron. Scl Health, A20, 135-148 Taylor, R T., E. Fultz, M G Kmze and J.S Felton (1987) Formation of the fried ground beef mutagens 2-amlno-3methyhmldazo(4,5-f)qulnohne (IQ) and 2-armno-l-methyl6-phenyhmldazo(4,5-b)pyrldme (PhlP) from 1-phenylalamne (Phe)+creatlmne (cre) (or creatme), Environ Mutagen., 9, 276 (Abstract) Wahlefeld, A.W., G Holz and H.U. Bergmeyer (1974) Methods of Enzymatic Analysis, Vol 1, 1st edn, Verlag Cherme, Welnhelm and Acadermc Press, New York, London, pp. 1786-1790. Yosluda, D , Y Smto and S. Mlzusakl (1984) Isolation of 2-amino-3-methyl-lmldazo-(4,5-f)qumohne as mutagen from the heated product of a rmxture of creatme and prohne, Agnc Biol. Chem, 48, 241-243