Chem. Res. Toxicol. 1994, 7, 313-318
313
Activation of the Maillard Reaction Product 5-(Hydroxymethy1)furfural to Strong Mutagens via Allylic Sulfonation and Chlorination Young-Joon Surht and Steven R. Tannenbaum* Division of Toxicology and Department of Chemistry, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139 Received December 6, 1993”
5-(Hydroxymethyl)furfural (HMF), one of the major intermediate products in the Maillard reaction, is present in a wide variety of foods. This aldehyde is formed as a decomposition product of glucose and fructose in foodstuffs subject to cooking or heat sterilization. It has been found to possess mutagenic and DNA strand-breaking activity. However, the mechanisms by which HMF exerts its genotoxicity remain unclear. The present study was undertaken to determine if HMF could be metabolically activated via esterification of the allylic hydroxyl group. In support of this concept, the chemically synthesized sulfuric acid ester, 5- [(sulfooxy)methyllfurfural (SMF), exhibited direct mutagenicity at both thymidine kinase and hypoxanthine-guanine phosphoribosyltransferase loci in human lymphoblasts. This reactive ester also induced 8-azaguanine-resistant mutants in Salmonella typhimurium TM677 in a dosedependent manner. The intrinsic mutagenicity of SMF was enhanced by addition of extra chloride ion to the assay medium. The model allylic derivative, 5-(chloromethyl)furfural,was also mutagenic and cytotoxic in bacteria, but much more active than the sulfuric acid ester in this regard. In contrast to (su1fooxy)methyland chloromethyl derivatives of HMF, 2-[(sulfooxy)methyl]- and 2-(chloromethy1)furans which lack the aldehyde functionality did not exhibit significant mutagenicity. Rodent hepatic cytosols contained sulfotransferase activity responsible for the formation of the reactive allylic sulfuric acid ester metabolite from HMF. Chart 1
Introduction Heat treatment of foods containing reducing sugars and amino acids during cooking or sterilization triggers a sequence of nonenzymatic browning reactions,collectively known as the “Maillard reaction” (1-3). 5-(Hydroxymethy1)furfural (HMF;1 structure shown in Chart 1) is one of the most common intermediate products in the Maillard reaction. It is present in a wide variety of foods, including milk, fruit juices, spirits, honey, etc. (4-8). In addition to foods, cigarette smoke contains small amounts of HMF (9). This aldehyde is also found in parenteral solutions (reviewed in ref 10 and references therein). Recently, HMF and related furans have been isolated from the roots of Cirsium chlorolepis (Compositae), which are used in Chinese folk medicine to treat many diseases (11). Toxicological effects of HMF have been investigated in various experimental animals and in cultured mammalian cells (for review, see ref 10). In chronic feeding study in rats, no adverse effects were observed at a level of 250 mg/kg body weight of HMF, and its LDw values ranged from 0.5 to 5 g/kg body weight depending on the animal species and routes of administration (10). Cellular dam-
ages and growth inhibition were observed in cultured chicken embryo fibroblasts treated with HMF. Attempts have been made by several investigators to determine the short-term genotoxicity of Maillard reaction products including HMF (12-18). Omura and his co-workers demonstrated the formation of mutagenic substances in the model browning systems with glucose and a variety of amino acids (18). They isolated a product formed by the Maillard reaction between glucose and lysine and proved it to be HMF by comparison with an authentic specimen. HMF thus formed had mutagenic and DNA strandbreaking activity. HMF and some of its derivatives contained in caramelization products have been found to possess a clastogenic activity in Chinese hamster ovary cells exposed to these substances (19,20) and also to induce mitotic gene conversion in the yeast (21). It has very recently been reported that HMF may act as both an initiator and a promoter in the induction of colonic aberrant cryptic foci in rats (221,suggesting its implications for colon carcinogenesisby thermolyzed sucrose. However, the mechanisms by which HMF exerts its genotoxic and tumorigenic activity remain unclear. Esterification of the allylic hydroxyl group of HMF appears to be one of the possible activation pathways for this hydrocarbon. If the resulting allylic ester bears a good leaving group such as sulfate, it would produce a highly electrophilic allyl
*Author to whom correspondence should be addressed at the Department of Chemistry and Division of Toxicology, Massachusetts Institute of Technology, 77 Massachusetts Ave., Cambridge, MA 02139. Phone: 617-253-3729; Fax: 617-258-8676. t Present address: Division of Environmental Health Sciences, Department of Epidemiology and Public Health, Yale University School of Medicine, 60 College St., New Haven, CT 06510. e Abstract published in Aduance ACS Abstracts, March 15, 1994. 1 Abbreviations: HMF, 5-(hydroxymethyl)fal;SMF,5-[(sulfmxy)methyllfurfural; CMF, 5-(chloromethyl)furfural;PAPS, 3’-phosphoadenosine 5’-phosphosulfate; DCCI, dicyclohexylcarbodiimide; DMF, dimethylformamide; 8-AGR,8-maguanine resistance; MF, mutant fraction; tk, thymidine kinase; hprt, hypoxanthine-guanine phosphoribosyltransferase.
0893-228x19412707-0313$04.5010
0
1994 American Chemical Society
314
Chem. Res. Toxicol., Vol. 7, No. 3, 1994
carbocation which could be stabilized by distribution of charges on the furan ring. Subsequent interaction of such a n electrophilic intermediate with critical cellular nucleophiles (e.g., DNA, RNA, a n d proteins) may result in structural damages in these informational macromolecules, thereby causing toxicity a n d mutagenicity. As one step toward testing this possibility, we synthesized a presumed electrophilic ester, 5- [(sulfooxy)methyl]furfural (SMF) and determined its mutagenicity in both bacterial and mammalian cells. Metabolic sulfonation of HMF to SMF by hepatic sulfotransferase activity was also investigated. In addition t o SMF, 5-(chloromethyl)furfural (CMF) was included in this study as a model reactive ester of HMF.
Materials and Methods Caution. Nfl-Dicyclohexylcarbodiimide(DCCI) is a contact allergen. Thionyl chloride is strongly irritating and corrosive. Therefore, these chemicals should be handled with adequate skin and eye protection, and all synthetic manipulations should be conducted in an efficient fume hood. Chemicals. HMF (99% pure), furfuryl alcohol, DCCI, (2,4dinitrophenyl)hydrazine,and thionyl chloridewere obtained from Aldrich Chemical Co. (Milwaukee, WI). GSH, bovine serum albumin, and 3'-phosphoadenosine 5'-phosphosulfate (PAPS) were products of Sigma Chemical Co. (St. Louis, MO). [36S]PAPS was purchased from New England Nuclear Product Co. (Boston MA). All other reagents and solvents used were commercial products of analytical grade. Synthesis of Allylic Derivatives of HMF and Furfuryl Alcohol. SMF was prepared by reacting HMF with sulfuric acid using DCCI as a condensing agent. Briefly, to a solution of 5 mmol of DCCI in 3 mL of dry dimethylformamide (DMF) kept at 0 "C was added 1 mmol of HMF in 0.5 mL of DMF with vigorous stirring followed by dropwise addition of sulfuric acid (2 mmol) diluted in 1 mL of ice-chilled DMF. After 30-min reaction at 0 "C, the reaction mixture was centrifuged. The supernatant was neutralized with 1 N methanolic NaOH and centrifuged at 2000 rpm for 15 min in a Fisher Centrific Model 225 centrifugewith a fixed-anglerotor. The precipitate was rinsed with a small volume of DMF and centrifuged at 2000 rpm. The combined organic solvent was concentrated in vacuo. The resulting residue was redissolved in 5 mL of ethanol and centrifugedas above. SMF was precipitated from the supernatant as a sodium salt after slow addition of 10 volumes of anhydrous ethyl ether. The precipitate collected by centrifugationwas rinsed three times with ethyl ether and dried in vacuo (yield 48%). When the product obtained from above reaction was applied on a silica gel thin-layer chromatogram and developed with ethyl ether-hexane (6:l v/v), it remained on the origin and developed an orange color after spraying with (2,4-dinitrophenyl)hydrazine (0.5% solution in 2 N HC1). Under these conditions, HMF migrated to the upper half region. It also produced a single (2,4dinitropheny1)hydrazine positive spot with an Rf value of 0.5 on a cellulose thin-layer chromatogram (Sigma, Type 100) in n-butanol-acetic acid-water (2:2:1 v/v). HMF under these conditions migrated near to the solvent front. Hydrolysis of the above sulfonation product formed a single ethyl ether-extractable material which showed the same R, value as that of HMF. Negative ion fast atom bombardment mass spectral analysis resulted in a strong peak at m/z 205 ([M - Nal-) which was consistent with a structure characteristic of the sulfuric acid ester of HMF (Na salt). Likewise, 5- [(sulfooxy)methyllfuran was prepared as a sodium salt from furfuryl alcohol. CMF was synthesizedby reacting HMF (2.5mmol) with 3-foldmolar excess of thionyl chloride in 10 mL of dry dioxane in the presence of zinc chloride (0.25 mmol) as a catalyst (23). The initial yellow color of the reaction mixture changed to violet as the reaction progressed. After 20-min reaction at 0 "C, the reaction mixture was quenched with 1mL of distilled water, taken up in methylene chloride (40 mL), and briefly washed with saturated NaCl. The
Surh and Tannenbaum organic layer was dried over anhydrous sodium sulfate and evaporated to a small volume under reduced pressure. The crude product was chromatographed on a short column of silica gel with methylene chloride as an eluting solvent to give an oily material which solidified during storage at -80 OC (yield 53%). Ita purity was monitored by thin-layer chromatography (silica gel with fluorescence indicator, ethyl ether-hexane, 6:1), which showed a single spot with aRfvalueof 0.8. Under this condition, HMF migrated to the middle (Rf 0.46). Electron impact mass spectroscopy for the purified product gave satisfactory M+ (m/z 144) and a characteristic M+ 2 (m/z 146)molecular ion peaks. The isotope ratio based on their relative intensities (3:l) was consistent with that observed for a monochlorinated compound. The chloromethyl derivative of furfuryl alcohol was prepared in a manner analogous to that described for CMF. Preparation of Rat Liver Cytosolic Fraction. Hepatic postmitochondrial supernatant (S-9) fractions from male Sprague-Dawley rats or CD-1 mice were supplied from Molecular Toxicology,Inc. (Annapolis,MD) and centrifuged at 105000gfor 1h a t 4 "C. The supernatant obtained was stored at -80 "C until used. The protein content of cytosol was determined using the Coomassie Blue G-250 reagent (Pierce Chemical Co., Rockford, IL) with bovine serum albumin as a standard. Bacterial Mutagenicity Assays. The forward mutation assay to 8-azaguanine resistance @-A@) in Salmonella typhimurium strain TM677 originally developed by Skopek et al. was used to assess mutagenic activity. The experimental details have been well documented (24). Briefly, exponentially growing bacteria were suspended in medium and treated with various concentrations of test compounds. After 2-h incubation at 37 "C, cells were resuspended in phosphate-buffered saline and aliquots were plated under selective (50 pg/mL 8-azaguanine) and nonselective conditions. Colonies were counted 48 h after incubation at 37 "C. The 8-AGR mutant fraction (MF) was determined as the number of colonies observed under selective conditions divided by the number of colonies formed under nonselective (permissive) conditions, which was multiplied by the appropriate dilution factor as summarized in the equation: MF = (D)(M/V)where M is the number of colonies (mutants) on selectiveplates, Vis the number of colonies(viables)on toxicity plates, and D is the dilution factor between the number of bacteria plated on the two types of plates. The 95% confidence limit on the mean mutant fractions in untreated control cultures was the criterion used to determine an observed mutation. Mammalian Mutagenicity Assays. HMF and SMF were tested for mutagenic activity at the thymidine kinase (tk) and hypoxanthine-guanine phosphoribosyltransferase (hprt) loci in TK6 human lymphoblast. In the standard assay, 4 X l o 7 cells per replicate culture (100 mL) were exposed to the test substance in 500-mL Erlenmeyer flasks. The cultures were then incubated for 20 h. After incubation, 50-80 mL of each culture (depending on the expected toxicity) was centrifuged at lOOOg and resuspended in 100 mL of fresh media. Cultures were passaged daily after treatment to 4 X l o 7 cells in 100 mL. The positive control was 4-nitroquinoline N-oxide (70 ng/mL). After a 6-day phenotypic expression period, cultures were plated in the presence of trifluorothymidine (three 96-wellplates at 2 cells/well). After incubation for 12 days, plates were scored for the presence of a colony in each well. The observed mutant fractions for the negative and positive controls were within the acceptable range at both loci. The exposure concentrations were chosen on the basis of a preliminary toxicity experiment. In the first experiment cultures were exposedto six concentrations of each test chemical. For HMF, cultures were exposed to 5,10,20,30,50, and 75 pg/ mL. The exposure concentrations for SMF were 5,10,20,30,40, and 60 pg/mL. None of the exposure concentrations of each test substance were excessivelytoxic, and the cultures exposed to the four highest test concentrations were plated to determine the mutant fraction. In the second experiment cultures were exposed to only those four highest test concentrations. Sulfotransferase Assay for HMF. Metabolic sulfonation of HMF was measured by a modification of the radiometric
+
Chem. Res. Toxicol., Vol. 7, No. 3, 1994 316
Allylic Actiuation of 5-(Hydroxymethyl)furfural ~~
Table 1. Bacterial Toxicity and Mutagenicity of SMF* concn (pM) survival (% control) mutant fraction (XlV) 0 100 8.2 25 85 16.5 100 42 77.9 250 13 170
aVarious amounts of SMF dissolved in 10 pL of MezSO were incubated with S.typhimurium TM677 at 37 "Cfor 2 h as described in Materiala and Methods.
1
1001
I
120
I
I
I
-
h
v)
g 100Y
0
Y
=
E
00-
e
60-
5
40-
s
0.5
1.o
1.5
2.0
CMF ( P W
Figure 2. Effect of glutathione on bacterial mutagenicity of CMF. CMF (0.25-2.0 pM)was preincubated with S. typhimurium TM677 with (closedcircles)and without (open circles) GSH (0.8 mM).
200
25
50
75
100
SMF (PM)
Figure I. Effect of extra chloride ion on the cytotoxicity and mutagenicity of SMF. SMF (25-100 pM) was incubated with S. typhimurium TM677 with (closed circles) and without (open circles) addition of extra chloride ion (154 mM) to the media. After incubation at 37 OC for 2 h, cells were plated and further incubated for 48 h to allow the growth of AGR colonies. Table 2. Intrinsic Toxicity and Mutagenicity of CMF in S. tyddmurium TM677 concn (pM) survival (% control) mutant fraction ( X 1P) 0 100 6.2 71 24.3 0.5 1.0 25 65.5 2.5 13 64.3 7 40.0 5.0 method using [SSIPAPS as a sulfate donor (25). The incubation mixture in a total volume of 100 p L consisted of 3 mM MgC12, 5 mM 2-mercaptoethanol, bovine serum albumin (0.0625%), 20 p L of liver cytosol (0.3-0.4 mg protein), 50 pM [%S]PAPS(0.025 pCi), 50 mM phosphate buffer (pH 7.0), and various concentrations of HMF. Incubations were performed for 20 min at 37 OC and terminated with 66 p L of an equal volume mixture of 0.1 M barium hydroxide and 0.1 M barium acetate, followed by immediate addition of 33 p L of 0.1 M zinc acetate. The mixtures were vortex-mixed and centrifuged. To the supernatant collected were added 33 p L of barium hydroxide and 33 pL of zinc sulfate in succession. After centrifugation, aliquota of supernatant were counted using Liquiscint as a cocktail.
Results Comparative Bacterial Mutagenicities of Allylic Derivatives of HMF. SMF, the electrophilic allylic
sulfuric acid ester of HMF, showed direct mutagenicity toward S. typhimurium TM677 when incubated with these bacteria without a metabolic activation system. Thus, this reactive ester induced 8-AGR mutants in a dosedependent manner a t concentrations from 25 to 250 pM (Table 1). Toxicity of SMF was also enhanced with increased mutagenicity. The acetic acid ester of HMF was neither mutagenic nor cytotoxic under the same experimental conditions. The intrinsic mutagenicity of SMF was significantly enhanced by addition of extra chloride ion (154mM) to the incubation medium (Figure 1, bottom). In parallel with increased mutagenicity, cytotoxicity of SMF was also enhanced in the presence of chloride ion (Figure 1,top). This chloride ion-dependent increase in mutagenicity and cytotoxicity of SMF might have been mediated through the formation of a less polar chloromethyl derivative by interaction of SMF with chloride ion. To test this possibility, bacterial mutagenicities of the chemically synthesized chloromethyl and (su1fooxy)methyl derivatives of HMF were compared. While SMF showed a dose-dependent mutagenicity over the concentration ranges from 25 to 100pM, CMF was too toxic to assess the correct mutagenicity, but considerable toxicity was still observed above the concentration of 2.5 pM (Table 2). GSH nonenzymatically protected against the strong mutagenicity and toxicity of CMF (Figure 2). Mutagenic activitiesof 2-[(sulfooxy)methyll- and 2-(chloromethy1)furan without the aldehyde functionality were also determined. In contrast to SMF and CMF containing an aldehyde group in the furan ring, no significant mutagenicity was observed with the (su1fooxy)methyland chloromethyl derivatives of furfuryl alcohol (Table 3). MammalianMutagenicity of HMF and SMF. There was no indication of a mutagenic activity of HMF in TK6 human lymphoblast cells. No exposure concentration
316 Chem. Res. Toxicol., Vol. 7, No. 3, 1994
Surh and Tannenbaum
Table 3. Comparative Bacterial Mutagenicity of SMF, CMF, 24(Sulfooxy)methyl]furan, and 2-(ChIoromethyl)furan
ba
R1430 -CH2R2
compd MezSO SMF
f-[(sulfooxy)methyllfuran
mutant concn survival fraction R1 R2 (pM) (7% control) (x106) Experiment 1 100 6.3 CHO OSOsNa 25 58 18.2 50 51 41.9 100 36 88.1 200 16 241 H OSOsNa 25 86 6.0 50 100 200 Experiment 2
MezSO CMF
CHO C1
2-(chloromethyl)- H furan
C1
1 2.5 5 10 100 lo00
t
79 73 59
6.5 7.0 7.7
100 21 18 6 85
4.7 74.2 83.7 200 7.6
82 25
7.9 16.4
Table 4. Comparison of Mutagenic Activity of HMF and SMF in TK6 Human Lymphoblasts mutagenicity at concn relative test tk hPd locus locus (pg/mL) survival compd 0 1.00 2.84 f 1.34" 2.11 i 0.74 1.11 1.94 f 0.32 2.12 f 0.34 HMF 20 1.01 2.58 f 0.90 1.68 i 0.52 30 1.12 2.75 f 0.67 1.52 f 0.25 50 75 1.08 3.12 f 1.34 1.02 f 0.73 SMF 20 0.98 4.20 f 1.99 2.09 f 0.57 30 0.69 5.78 i 1.28 3.61 f 1.62 0.63 6.75 i 1.86b 3.55 i 1.33 40 60 0.21 12.9 f 2.75b 8.17 f 2.38b 4-nitroquinoline 0.07 1.02 57.0 f 12.2b 31.1f 6.32b N-oxide a Mean f standard deviation. Significantly different from the concurrent control (p < 0.05).
produced a statistically significant response at either the tk or hprt locus. HMF was also not toxic to TK6 cells at the concentration tested. The relative survival of the cultures exposed to the highest test concentration, 75 pgl mL, was as high as that observed in solvent-treatedcontrols (Table 4). Under these assay conditions, however, SMF was clearly mutagenic at both the tk and hprt loci (Table 4). Thus, the responses a t the tk locus of exposure to 40 and 60 rg/mL were statistically significant in comparison to the concurrent negative control. Exposure to 60 pgl mL also resulted in mutation at the hprt locus. SMF was toxic to the cells at the concentrations tested whereas the parent hydroxymethyl hydrocarbon was not (Table 4). Thus, the relative survival of the cultured cells exposed to the highest concentration (60 pg/mL) of SMF was 0.21 (Table 4). Metabolic Sulfonation of HMF. For determining the metabolic formation of SMF, the sulfotransferase assay for HMF was conducted by incubating the various concentrations of HMF with liver cytosol and [35SlPAPS. The apparent enzyme activity responsible for sulfonation of HMF was determined by measuring the radioactivity (3%) incorporated into the substrate during 20 min. The
400
I I
l
0
//
300
,0
0
2.5
5
7.5
4
10
HMF (PM)
Figure 3. Metabolic sulfonation of HMF by rat liver cytosolic sulfotransferase activity. HMF was incubated with liver cytosols from male Sprague-Dawley rats in the presence of [SSIPAPS. The incorporation of radioactivity during the initial 20-min incubation period was determined as described in the text. The enzyme activity is expressed as pmol of HMF sulfonated/(mg of cytosolic protein.20 min).
cytosolic sulfotransferase activity for HMF appears to be dose-dependent up to the 10mM substrate concentration (Figure 3). When HMF (10 mM) was incubatedat various time intervals with rat liver cytosol and [36SlPAPS, radioactivity incorporated in the product increased during the first 15 min, but therefore diminished gradually to the background level by 60 min (data not shown). This timedependent decline in the amounts of the reaction product appears to be associated with hydrolysis of the presumed reactive sulfuric acid ester metabolite of HMF in an aqueous environment, thereby releasing 35S04previously incorporated in the product. Structural identification of the reaction product formed from HMF in the presence of sulfotransferase activity needs further investigation.
Discussion Considering the wide distribution of HMF in heatprocessed foods and parenteral solutions, the correct evaluation of its genotoxicity is very important. Although HMF has been found to be mutagenic in the Ames Salmonella assay and to possess DNA strand-breaking activity in vitro, its mechanism of action is largely unknown. Our data presented herein suggest the metabolic activation of HMF via the formation of a reactive allylic ester bearing a good leaving group such as sulfate. This supposition is based on the high mutagenicity of SMF and presence of cytosolic sulfotransferase activity responsible for the formation of this reactive ester from HMF. Sulfu conjugation has been shown to play a role in metabolic activation, mutagenicity, and carcinogenicity of a variety of environmental toxicants that include aromatic amines (26, 27), alkenylbenzenes (28, 291, and polynuclear aromatic hydrocarbons (30-37). The K mvalue for enzymatic sulfonation of HMF appears to be relatively high (>5 mM) since no saturation of enzyme activity was observed at a substrate concentration as high as 10 mM. A structurally related compound, furfuryl alcohol, has been shown to be sulfonated by arylsulfotransferase IV, and the K m value of this reaction was about 1mM (38). It seems likely that the same enzyme also catalyzesthe sulfuric acid esterificationof HMF, which is structurally analogous to furfuryl alcohol. Since HMF has been shown to be tumorigenic in rat colon, determi-
Allylic Activation of 5-(Hydroxymethyl)furfural
nation of sulfotransferase activity in colon cytosol using HMF as a substrate merits further investigation. The present study indicates that CMF is much more mutagenic in S. typhimurium TM677 than SMF as determined by the increase in the number of 8-AGR colonies in these bacteria. Since CMF is more lipophilic than SMF, it may penetrate the bacterial cell membrane to a greater extent than the relatively polar sulfuric acid ester, thereby producing more damage to cellular DNA. The increased bacterial mutagenicity of SMF in the presence of extra chloride ion in the assay medium further supports the above concept. Similar findings have been recently demonstrated with sulfuric acid esters of certain hydroxymethyl polycyclic aromatic hydrocarbons (33,36, 37, 39). Mutagens lipophilic per se enough to readily penetrate the bacterial cell membrane or those not subject to possible chloride ion displacement did not exhibit a profound increase in mutagenic activities in the presence of chloride anion (37). Thus, it is likely that the different lipophilicity of SMF and CMF accounts for their remarkable differences in mutagenic potency. The intrinsic mutagenicity of relatively polar benzylic sulfuric acid esters of some polynuclear aromatic hydrocarbons has been attributed to the formation of less polar (more lipophilic) chloromethyl derivatives by interaction of these reactive sulfuric acid esters with chloride ion (33, 35, 37, 39). Likewise, SMF in the presence of chloride ion may form a highly mutagenic chloromethyl derivative. At the present time, however, there is no clear-cut evidence for the formation of the chloromethyl derivative directly from HMF in biological systems. A potential pathway is chlorination of HMF in gastric juice where the concentration of chloride ion is relatively high. This possibility will be further discussed elsewhere (Y.-J. Surh, A. Liem, J. A. Miller,and S. R. Tannenbaum, manuscript in preparation). It has recently been reported that myeloperoxidase catalyzesthe chlorination of certain xenobiotics as well as endogenous substrates (40). It may be of interest to determine if CMF can be formed by this enzyme activity. The aldehyde functional group of SMF and CMF has an electron-withdrawing effect and is considered to deactivate the resonance stabilization of a presumed allyl carbocation produced from these reactive derivatives. If this does occur, non-aldehyde [(sulfooxy)methyll- and (chloromethy1)furans should be more susceptible than SMF and CMF to substitution ( s N 1 ) reactions. On the contrary, however, the non-aldehyde analogs, 2-[(sulfooxy)methyll- and 2-(chloromethyl)furans, were not mutagenic while their corresponding aldehyde derivatives showed profound intrinsic mutagenicity. Hence, the aldehyde functionality contained in SMF and CMF seems to be critical in determining the mutagenicity of these reactive allylic derivatives of HMF. The a,@-unsaturated aldehyde group of SMF and CMF which is the Michael acceptor may also interact with cellular nucleophiles, thereby making these molecules act as bifunctional alkylating agents. Cross-linking of the DNA molecule has been well illustrated with certain bifunctional alkylating agents such as mitomycin C ( 4 1 ) . Therefore, it seems likely that the aldehyde group of SMF and CMF might be involved in producing DNA cross-linkage, leading to increased mutagenicity. Both SMF and CMF form adducts with calf thymus DNA.2 However, it is unclear R. Kerdar and S. R. Tannenbaum, unpublished data.
Chem. Res. Toxicol., Vol. 7, No. 3, 1994 317
whether these DNA adducts were produced as a result of cross-linking or covalent binding through the presumed allyl carbocation. Certain dialdehydes have been reported to possess mutagenic and alkylating activity (42). When chemically synthesized 2,5-furandialdehyde was tested in S. typhimurium TM677, it was found to be extremely mutagenic and cytotoxic,3which might result from crosslinking of bacterial DNA by this reactive dialdehyde. It would be worthwhile to determine if this dialdehyde derivative could be formed metabolically from HMF by such enzyme activity as alcohol dehydrogenase. So far as carcinogenicity of HMF is concerned, only limited information is available. Multiple doses (10 X 50 pmol) of HMF were topically applied to the shaved backs of mice followed by promotion with 12-0-tetradecanoylphorbol 13-acetate, but the incidence and multiplicity of papillomas produced were not statistically significant (43). Zhang et al. have very recently reported the oncogenic potential of HMF in the rat colon (22). Results from our preliminary study indicate that both (su1fooxy)methyl and chloromethyl derivatives of HMF potentially possess initiating activity in mouse skin while the activity of the parent hydroxymethyl hydrocarbon is not very significant (44).CMF was also hepatocarcinogenic in infant male B6C3F1 mice (44). The marked tumorigenicity of SMF and CMF, together with their strong mutagenicity and reactivity to form DNA adducts, suggesta that these allylic derivatives, if formed in vivo, might play a role as ultimate electrophilic and carcinogenic metabolites of HMF. Further research on this issue is in progress.
Acknowledgment. The authors gratefully acknowledge the technical assistance of Walter Bishop in conducting bacterial mutagenicity assays. The mammalian mutagenicity testa were performed under contract by Gentest Corp., Woburn, MA. Special appreciation is expressed to Dr.Charles L. Crespi for analysis of data and valuable discussions. This study was financially supported by Grants ES 05622 and ES 02109 from the National Institutes of Health.
References Mauron, J. (1981) The Maillard reaction in food; a critical review from the nutritional standpoint. B o g . Food Nutr. Sci. 5, 5-35. Adrian, J. (1974)Nutritionaland physiological consequences of the Maillard reaction. World Rev. Nutr. Diet. 19,71-112. Ashoor, S.H., and Zent, J. B. (1984) Maillard browning of common amino acids and sugars. J. Food Sci. 49,1206-1207. Morales,F. J., Romero,C.,andJimenez-PBrez,S.(1992)An enhanced liquid chromatographic method for 5-hydroxymethylfurfural determination in UHT milk. Chromatographia 33, 45-48. Blanco Gomis, D.,Gutibrrez Alvarez, M. D., Sopefia Naredo, L. S., and Mangas Alonso, 3. 3. (1991) High-performance liquid chromatographic determination of furfural and hydroxymethylfurfural in apple juices and concentrates. Chromatographia 32,45-48. Meydav, S., and Berk, Z. (1978) Colorimetric determination of browningprecursorsin orangejuice products. J.Agric. Food Chem. 26,282-285. Jeuring, H. J., and Kuppers, F. J. E. M. (1980) High performance liquid chromatography of furfural and hydroxymethylfurfural in spirits and honey. J. Assoc. Off. Anal. Chem. 63,1215-1218. Tharasyroulou,A. T. (1986)The use of HMF and diastase as criteria of quality of greek honey. J. Agric. Res. 25, 186-195. Black, D. K. (1966) Isolation of 5-hydroxymethylfurfural from cigarette smoke condensate. Chem. Znd. 32, 1380. Ulbricht, R. J., Northup, S.J., and Thomas, J. A. (1984) A review of 5hydroxymethylfurfural(HMF) inparenteralsolutions. Fundam. Appl. T O X ~ C4, O843-853. ~. Y.-J. Surh and S.R. Tannenbaum, unpublished observation.
318 Chem. Res. Toxicol., Vol. 7, No. 3, 1994 (11) Shen, Y.-M., and Mu, Q.-Z. (1990) New furans from Circium
chlorolepis. Planta Med. 56,472-474. (12) Shibamato, T. (1982)Occurrence of mutagenic produds in browning model systems. Food Technol. 3, 59-65. (13) Nishi, Y., Miyakawa, Y., and Kato, K. (1989) Chromosome aberrations induced by hydrolysates of carbohydrates in Chinesehamster V79 cells. Mutat. Res. 227, 117-123. (14) Vagnarelli,P., De Sario, A., Cuzzoni, M. T., Mazza, P., and De Carli, L. (1991) Cytotoxicity and clastogenic activity of ribose-lysine browning model system. J. Agric. Food Chem. 39,2237-2239. (15) Aeschbacher,H. U., Chappuis, C. H., Manganel, M., and Aeschbach, R. (1981)Investigationof Maillard products in bacterial mutagenicity test systems. Prog. Food Nutr. Sci. 5, 279-294. (16) Gazzani, G.,Vagnarelli, P., Cuzzoni, M. T., and Mazza, P. (1987) Mutagenic activity of the Maillard reaction products of ribose with different amino acids. J. Food Sci. 52, 757-760. (17) Aeschbacher, H. R. (1986) Genetic toxicology of browning and caramelizingproducta. In Genetic Toxicology ofthe Diet (Knudsen, I., Ed.) pp 133-144, Alan R. Liss, New York. (18) Omura, H., Jahan, N., Shinohara, K., and Murakami, H. (1983) Formation of mutagens by the Maillard reaction. In The Maillard Reaction in Foods and Nutrition (Waller, G.R., and Feather, M. S., Eds.) pp 537-563, American Chemical Society, Washington, DC. (19) Stich, H. F., Stich, M. P., Rosin, M. P., and Powrie, W. D. (1981) Clastogenicactivityof caramelandcaramelizedsugars. Mutat. Res. 191,129-136. (20) Stich, H. F., Rosin, M. P., Wu, C. H., and Powrie, W. D. (1981) Clastogenicity of furans found in foods. Cancer Lett. 13, 89-95. (21) Rosin, M. P., Stich, H. F., Powrie, W. D., and Wu, C. H. (1982) Induction of mitotic gene conversion by browning reaction products and ita modulating by naturally occurring agents. Mutat. Res. 101, 189-197. (22) Zhang,X.-M., Chan, C. C., Stamp, D., Minkin, S., Archer, M. C., and Bruce, W. R. (1993) Initiation and promotion of colonic aberrant crypt foci in rata by 5-hydroxymethyl-2-furaldehydein thermolyzed sucrose. Carcinogenesis 14,773-775. (23) Squires, T. G., Schmidt, W. W., and McCandlish, C. S., Jr. (1975) Zinc chloride catalysts in the reaction of thionyl halides with aliphatic alcohols. J. Org. Chem. 40, 134-136. (24) Skopek,T.,Liber,H.,Kaden,D.,andThilly, W. (1978)Quantitative forward mutation assay in Salmonella typhimurium using 8-azaguanine as a genetic marker. Proc. Natl. Acad. Sci. USA. 75,410419. (25) Singer, S. S. (1985)Preparation and characterization of the different kinds of sulfotransferases. In Biochemical Pharmacology and Toxicology(Zakim, D., Ed.) pp 95-159, Academic Press, New York. (26) Lai,C.-C.,Miller, J. A.,Miller,E. C., andliem, A. (1985)N-Sulfooxy2-aminofluoreneis the major ultimate electrophilic and carcinogenic metabolite of N-hydroxy-2-acetylaminofluorene in the livers of infant male C57BL/6J X C3H/HeJF, (B6C3F1) mice. Carcinogenesis 6, 1037-1045. (27) Kato, R., and Yamjoe, Y. (1987) Metabolic activation and covalent binding to nucleic acid of carcinogenic heterocyclic amines from cookedfoodandaminoacidpyrolysates.Jpn.J. CancerRes. (Gann) 78,297-311. (28) Boberg, E. W., Miller, E. C., Miller, J. A.,and Liem, A. (1983)Strong evidence from studies with brachymorphic mice and pentachlorophenol that 1’-sulfooxyeafroleis the major ultimate electrophilic and carcinogenic metabolite of 1’-hydroxysafrole in mouse liver. Cancer Res. 43,5163-7173. (29) Fennell,T. R., Wiseman, R. W., Miller, J. A.,andMiller, E. C. (1985) Major role of hepatic sulfotransferase activity in the metabolic activation, DNA adduct formation, and carcinogenicityof 1’-hydroxy2’,3’-dehydroestragole in infant male C57BL/6J X C3H/HeJFl mice. Cancer Res. 45,5310-5320.
Surh and Tannenbaum (30) Watabe, T., Ishizuka, T., Isobe, M., and Ozawa, N. (1982) 7-Hydroxymethyl sulfate ester as an active metabolite of 7,12-dimethylbenz[alanthracene. Science 215,403-405. (31) Surh, Y.-J., Lai, C.-C., Miller, J. A., and Miller, E. C. (1987)Hepatic DNA and RNA adduct formation from the carcinogen 7-hydroxymethyl-12-methylbenz[alanthraceneand ita electrophilicsulfuric acid ester metabolite in preweanling rata and mice. Biochem. Biophys. Res. Commun. 144, 576-582. (32) Surh, Y.-J., Liem,A., Miller, E. C.,andMiller, J. A. (1989)Metabolic activation of the carcinogen 6-hydroxymethylbenzo[a]pyrene: formation of an electrophilic sulfuric acid ester and benzylic DNA adducts in rat liver in vivo and in reactions in vitro. Carcinogenesis 10, 1519-1528. (33) Surh, Y.-J., Blomquist, J. C., Liem, A., and Miller, J. A. (1990) Metabolic activation of 9-hydroxymethyl-10-methylanthracene and I-hydroxymethylpyreneto electrophilic,mutagenic,and tumorigenic sulfuric acid esters by rat hepatic sulfotransferase activity. Carcinogenesis 11, 1451-1460. (34) Surh, Y.-J., Kwon, H., and Tannenbaum, S. R. (1993) Sulfotransferase-mediated activation of 4-hydroxy- and 3,4-dihydroxycycle penta[c,dlpyrene, major metabolites of cyclopenta[c,dlpyrene. Cancer Res. 53,1017-1022. (35) Surh, Y.-J., Liem, A., Miller, E. C., and Miller, J. A. (1991) 7-Sulfooxymethyl-12-methylbenz[alanthracene is an electrophilic mutagen, but does not appears to play a role in carcinogenesis by 7,12-dimethylbenz[a]anthracene or 7-hydroxymethyl-12-methylbenz[alanthracene. Carcinogenesis 12,339-347. (36) Miller, J. A., and Surh, Y.-J. (1994) Sulfonation in chemical carcinogenesis. In Conjugation-Deconjugation Reactions in Chemical Carcinogenesis (Kauffman, F. C., Ed.) Springer-Verlag,Heidelberg (in press). (37) Glatt, H. R., Henschler, R., Phillips, D. H., Blake, J. W., Steinberg, P., Seidel, A,, and Oesch, F. (1990) Sulfotransferase-mediated chlorination of 1-hydroxymethylpyrene to a mutagen capable of penetrating indicator cells. Enuiron. Health Perspect. 88,43-48. (38) Binder, T. P., and Duffel, M. W. (1988)Sulfation of benzylicalcohole catalyzed by aryleulfotransferase IV. Mol. Pharmacol. 33,477-479. (39) Enders, N., Seidel, A., Monnerjahn, S., and Glatt, H. R. (1993) Synthesis of 11benzylic sulfate esters, their bacterial mutagenicity and ita modulation by chloride, bromide, and acetate anions. In Polycyclic Aromatic Compounds. Synthesis,Properties, Analytical Measurements, Occurrence, and Biological Effects (Garrigues, P., and Lamotte, M., Eds.) pp 887-894, Gordon and Beach Science Publishers, France. (40) Uetrecht, J. P., and M i d , N. (1991)N-Chlorination and oxidation of procainamide by myeloperoxidase: toxicological implications. Chem. Res. Toxicol. 4, 218-222. (41) Tomasz, M., Lipman, R., Chowdary, D., Pawlak, J., Verdine, G.L., and Nakanish, K. (1987) Isolation and structure of a covalent crowlink adduct between mitomycin C and DNA. Science 235,12041208. (42) Bjeldanes, L. F.,andChew,H. (1979)Mutagenicityof 1,2-dicarbonyl compounds: maltol, kojic acid, diacetyl and related substances. Mutat. Res. 67, 367-371. (43) Miyakawa, Y., Nishi, Y., Kato, K., Sato, H., Takahashi, M., and Hayashi, Y. (1991) Initiating activity of eight pyrolysates of carbohydrates in a two-stage mouse skin tumorigenesis model. Carcinogenesis 12, 1169-1173. (44) Surh, Y.-J., Liem, A., Miller, J. A., and Tannenbaum, S. R. (1993) High mutagenicity and tumorigenicity of sulfooxymethyl and chloromethyl derivatives of 5-hydroxymethylfurfural. Proc. Am. Assoc. Cancer Res. 34, 710.
