Reaction of the Potent Bacterial Mutagen 3-Chloro-4-(dichloromethyl

The potent bacterial mutagen and drinking water disinfection byproduct 3-chloro-4-. (dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) was reacted with ...
0 downloads 0 Views 106KB Size
Download PDF
226

Chem. Res. Toxicol. 1998, 11, 226-233

Reaction of the Potent Bacterial Mutagen 3-Chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) with 2′-Deoxyadenosine and Calf Thymus DNA: Identification of Fluorescent Propenoformyl Derivatives Tony Munter, Frank Le Curieux, Rainer Sjo¨holm, and Leif Kronberg* Department of Organic Chemistry, A° bo Akademi University, Akademigatan 1, FIN-20500 Turku/A° bo, Finland Received October 29, 1997

The potent bacterial mutagen and drinking water disinfection byproduct 3-chloro-4(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) was reacted with 2′-deoxyadenosine and calf thymus DNA in aqueous solutions at neutral conditions. HPLC analyses of the 2′-deoxyadenosine reaction mixtures showed that two previously unidentified products were formed. The products were isolated by preparative C18 chromatography, and their structures were characterized by UV absorbance, fluorescence emission, 1H and 13C NMR spectroscopy, and mass spectrometry. It was concluded that in both products a propeno bridge had been incorporated between N-1 and N6 of the adenine unit. In one of the products, the propeno bridge carried a formyl group [3-(2′-deoxy-β-D-ribofuranosyl)-7H-8-formyl[2,1-i]pyrimidopurine (pfA-dR)], and in the other the substituents consisted of a formyl group and a chlorine atom [3-(2′-deoxy-β-D-ribofuranosyl)-7H-8-formyl-9-chloro[2,1-i]pyrimidopurine (Cl-pfA-dR)]. These novel adducts exhibited fluorescence in the visible region with emission maxima around 460 nm. The yields of the products in reactions performed at pH 7.4 and 37 °C were about 0.03 mol %. In reaction of MX with calf thymus DNA, the adduct pfA-dR was formed and its yield was about 0.6 adduct/105 nucleotides.

Introduction In the late 1970s, it was reported that drinking water disinfected with chlorine generated mutagenicity in the Ames assay (1). Search for the compounds accounting for the mutagenicity resulted in the detection of the potent bacterial mutagen 3-chloro-4-(dichloromethyl)-5hydroxy-2(5H)-furanone (MX1) in extracts of chlorinated drinking water (Scheme 1) (2, 3). A few years earlier, MX had been isolated from a strongly mutagenic extract of chlorine-bleached pulp mill effluents, and its structure had been tentatively determined (4). Confirmation of the structure was provided by the synthesis of MX (5). In recent studies carried out in our laboratory, it has been shown that several compounds structurally related to MX are present in chlorine-bleached pulp mill effluents and in chlorine-disinfected drinking waters (6, 7). In pulp mills, MX and the other related compounds are formed when chlorine reacts with lignin, and in drinking water treatment plants, the compounds are formed when chlorine reacts with humic substances not removed during the purification of the raw water (6, 8, 9). In * Author for correspondence. E-mail: [email protected]. Phone: +358-2-2654138. Fax: +358-2-2654866. 1 Abbreviations: MX, 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)furanone; CMCF, 3-chloro-4-(chloromethyl)-5-hydroxy-2(5H)-furanone; MCF, 3-chloro-4-methyl-5-hydroxy-2(5H)-furanone; pfA-dR, 3-(2′-deoxyβ-D-ribofuranosyl)-7H-8-formyl[2,1-i]pyrimidopurine; Cl-pfA-dR, 3-(2′deoxy-β-D-ribofuranosyl)-7H-8-formyl-9-chloro[2,1-i]pyrimidopurine; NOE, nuclear Overhauser effect; NOESY, nuclear Overhauser enhancement spectroscopy; CHSHF, C-H shift correlation NMR spectroscopy; COLOC, C-H shift correlation NMR spectroscopy via longrange coupling; PEG, poly(ethylene glycol).

Scheme 1

lignin and humic substances, phenolic subunits are present, and it has been suggested that these subunits act as precursors for MX and the related compounds (10, 11). MX is one of the most potent, direct-acting bacterial mutagens known (12-14). The compound has been found in drinking water at concentrations ranging from a few nanograms per liter to about 60 ng/L, and it has been estimated that MX accounts for 30-50% of the Ames mutagenicity of drinking water (3, 7, 15-19). Also, the structurally related compounds generate mutagenicity in bacteria, but they are weaker mutagens than MX and account for at most about 5% of the drinking water mutagenicity (7). MX has been found to be mutagenic in mammalian cells in vitro (13, 20-22) and in vivo (23, 24). In a very recent work of Komulainen et al., it was found that the compound was a potent carcinogen in rodents (25). This finding, in connection with the findings of several epidemiological studies that show an association between cancer in humans and the consumption of chlorinated drinking water, suggests that MX should be considered as a risk factor for human health (26-28). It is generally considered that chemical reactions of genotoxic compounds with the base units in DNA may

S0893-228x(97)00195-1 CCC: $15.00 © 1998 American Chemical Society Published on Web 02/25/1998

MX Forms Propenoformyl Adducts with 2′-dAdo

cause gene mutations and contribute to cancer initiation (29). In previous contributions from this laboratory, it has been shown that the MX-related compounds 3,4dichloro-5-hydroxy-2(5H)-furanone (MCA) and 3-chloro4-methyl-5-hydroxy-2(5H)-furanone (MCF) form adducts upon reactions with nucleosides (30-34). Very recently, this laboratory reported that MCA, MCF, and MX form some previously identified adenosine adducts when reacted with calf thymus DNA (35). The current report is concerned with the structural determination of two novel fluorescent adducts produced in reaction of MX with 2′deoxyadenosine. Further it is shown that one of the adducts is also formed when MX is incubated with calf thymus DNA.

Materials and Methods Caution: MX has been found to be a rodent carcinogen and is one of the strongest known direct-acting mutagens tested in the Ames mutagenicity assay with Salmonella typhimurium (TA100). Therefore, caution should be exercised in the handling and disposal of the compound. Chemicals. Calf thymus DNA, DNase from bovine pancreas, nuclease P1 from Penicillium citrinum, alkaline phosphatase from bovine intestinal mucosa, acid phosphatase from white potato, and 2′-deoxyadenosine were obtained from Sigma Chemical Co. (St. Louis, MO). 3-Chloro-4-(dichloromethyl)-5-hydroxy2(5H)-furanone (MX) and 13C-3-labeled MX were synthesized according to the method of Franze´n and Kronberg (36). The purity of MX was at least 98%, as estimated by 1H NMR and GC. Chromatographic Methods. HPLC analyses were performed on a Kontron Instruments liquid chromatographic system consisting of a model 322 pump, a 440 diode-array detector (UV), a Jasco FP-920 fluorescence detector, and a KromaSystem 2000 data handling program (Kontron Instruments S.P.A, Milan, Italy). The reaction mixtures were chromatographed on a 5-µm, 4- × 125-mm reversed-phase C18 analytical column (Spherisorb ODS2, Hewlett-Packard, Espoo/ Esbo, Finland). In the spiking experiments two additional columns were used: a C8 column (5 µm, 4 × 125 mm, Lichrospher 100, RP-8, Hewlett-Packard) and a 250-mm long C18 column (5 µm, 4 × 250 mm, Spherisorb ODS2). The columns were eluted isocratically for 5 min with 5% acetonitrile in 0.01 M phosphate buffer, pH 7.1, and then with a gradient from 5% to 30% acetonitrile in 25 min at a flow rate of 1 mL/ min. Preparative isolation of the products was performed by column chromatography on a 2.5- × 10-cm column of preparative C18 bonded silica grade (40 µm, Bondesil, Analytichem International, Harbor City, CA). The products were further purified on a semipreparative 8-µm, 10- × 250-mm (Hyperprep ODS, Hypersil, Krotek, Tampere/Tammerfors, Finland) reversedphase C18 column. The column was eluted isocratically with 7% acetonitrile in 0.01 M phosphate buffer solution (pH 7.1) for 2 min and then with a gradient from 7% to 30% acetonitrile in 28 min at a flow rate of 4 mL/min. The column was coupled to a Shimadzu HPLC system, which consisted of two Shimadzu LC-9A pumps and a variable wavelength Shimadzu SPD-6A UV spectrophotometric detector (Shimadzu Europe, Germany). Spectroscopic and Spectrometric Methods. The 1H and 13C NMR spectra were recorded at 30 °C on a JEOL JNM-A500 Fourier transform NMR spectrometer at 500 and 125 MHz, respectively (JEOL, Japan). The samples were dissolved in Me2SO-d6, and TMS was used as an internal standard. The chemical shift assignments were based on H-H and C-H shift correlation spectra and on H-H NOE interactions. In addition, data obtained from proton-coupled 13C NMR were used. For the 2-D NMR spectra standard pulse sequences of the instrument (CHSHF, COLOC, COSY, and NOESY) were used. The determination of the shifts and the coupling constants of the multiplets of the proton signals in the deoxyribose unit was

Chem. Res. Toxicol., Vol. 11, No. 3, 1998 227 based on a first-order approach. The NMR experiments were performed on 1.2 mg of each compound. The NMR tube size was 5-mm i.d. The longest accumulation time was about 33 h, and it was required in order to obtain the nondecoupled 13C NMR spectra. The electrospray ionization mass spectra were recorded on a Fisons ZABSpec-oaTOF instrument (Manchester, U.K.). Ionization was carried out using nitrogen as both nebulizing and bath gas. A potential of 8.0 kV was applied to the ESI needle. The temperature of the pepperpot counter electrode was 90 °C. The samples were introduced by loop injection at a flow rate of 20 µL/min (H2O/CH3CN/acetic acid: 80/20/1). PEG 200 was used as standard for exact mass determinations. The mass spectrometer was working at a resolution of 7000. The UV spectra of the isolated compounds were recorded with the diode-array detector as the peaks eluted from the HPLC column. A Shimadzu UV-160 spectrophotometer (Shimadzu Europe) was used for the determination of the molar extinction coefficients (). The fluorescence spectra were recorded with a Hitachi F-2000 fluorescence spectrophotometer (Hitachi Ltd., Japan). Preparation of 3-(2′-Deoxy-β-D-ribofuranosyl)-7H-8formyl[2,1-i]pyrimidopurine (pfA-dR) and 3-(2′-Deoxy-βD-ribofuranosyl)-7H-8-formyl-9-chloro[2,1-i]pyrimidopurine (Cl-pfA-dR). MX (5 g, 23.1 mmol) was reacted with 2′deoxyadenosine (2.91 g, 11.6 mmol) in 900 mL of 0.5 M phosphate buffer solution (pH 7.4). The reaction was performed at 37 °C for 8 days and was followed by HPLC analyses on the C18 analytical column. The reaction mixture was filtered and then passed through the preparative C18 column. The column was first eluted with 250 mL of H2O and then with 100 mL of 5%, 6%, 10%, 13%, 17%, and 20% acetonitrile solutions in water. Fractions of 30 mL were collected. The compound pfA-dR eluted from the column with the 6% and 10% acetonitrile washes, and Cl-pfA-dR eluted with the 10% and 13% washes. The fractions containing the products were combined (the combined fraction contained both compounds) and concentrated by rotary evaporation to about 20 mL. Separation and purification of the compounds were carried out using the semipreparative column. The collected fractions were then desalted by use of the preparative C18 column. The desalted solutions were rotary evaporated to dryness, and the residues (green powder) were subjected to spectroscopic and spectrometric studies. The isolated amounts of the compounds were 1.2 mg. The isolated compounds had the following spectral characteristics: UV spectra pfA-dR UVmax (HPLC eluent, 10% acetonitrile in 0.01 M phosphate buffer, pH 7.1) 408 ( 23 300 M-1 cm-1), 212, 240 nm, UVmin 284 nm; Cl-pfA-dR UVmax (HPLC eluent, 14% acetonitrile in 0.01 M phosphate buffer, pH 7.1) 414 ( 25 000 M-1 cm-1), 218, 248 nm, UVmin 338, 240 nm; fluorescence spectra (H2O) pfA-dR λex 379 nm, λem,max 462 nm; Cl-pfA-dR λex 381 nm, λem,max 466 nm. In the positive ion electrospray mass spectra the following ions were observed (m/z, relative abundance, formation): pfAdR, 318 (100, MH+), high-resolution mass spectrometry gave the protonated molecular formula as C14H16N5O4 (MH+ 318.1204, calcd 318.1202); Cl-pfA-dR, 354/352 (33/100, MH+), high-resolution mass spectrometry gave the protonated molecular formula as C14H15N5O4Cl (MH+ 352.0804, calcd 352.0812). The 1H and 13C NMR spectroscopic data of pfA-dR and ClpfA-dR are presented in Tables 1 and 2, respectively. Small-Scale Reactions of MX with 2′-Deoxyadenosine. MX (17 mg, 0.08 mmol) was reacted with 2′-deoxyadenosine (10 mg, 0.04 mmol) in 3 mL of 0.5 M phosphate buffer solutions at pH 7.4, 6.0, and 4.6. The reactions were performed at 37 °C. The progress of the reaction was followed by HPLC analyses of aliquots of the reaction mixtures using the C18 analytical column. Reaction of 13C-3-Labeled MX with 2′-Deoxyadenosine. MX (173 mg, 0.80 mmol) mixed with 13C-3-labeled MX (56 mg, 0.26 mmol) was reacted with 2′-deoxyadenosine (133 mg, 0.53 mmol) in 90 mL of 0.5 M phosphate buffer solution at pH 7.4.

228 Chem. Res. Toxicol., Vol. 11, No. 3, 1998 Table 1.

1H

and

13C

Chemical Shifts (δ)a and Spin-Spin Coupling Constants, JH,H and JC,H (Hz), of Protons and Carbons in pfA-dR δ

proton

Munter et al.

multiplicity

H-2 (1H) H-5 (1H)

8.49 8.31

m m

H-9 (1H)

7.50

t

H-7a, H-7b (2H) CHO (1H) H-1′ (1H) H-2′ (1H) H-2′′ (1H) H-3′ (1H) H-4′ (1H) H-5′ (1H) H-5′′ (1H)

5.00, 5.01 9.28 6.32 2.64 2.33 4.41 3.88 3.60 3.52

mc t dd ddd ddd ddd dt dd dd

JH,H

0.8 14.4 0.6 7.0, 6.3 13.2, 7.0, 5.9 13.2, 6.3, 3.7 5.9, 3.7, 3.1 4.6, 3.1 11.8, 4.6 11.8, 4.6

1J C,H

>1J C,H

215.7 211.8

4.1 1.8 12.9, 5.4, 2.3 10.9, 5.7 11.6, 1.3 1.3, 2.6 24.8, 6.7, 5.4 4.1, 3.6 5.7, 1.4

carbonb

δ

multiplicity

C-2 C-5 C-3a C-10a C-10b C-9 C-8 C-7 CHO C-1′ C-2′

141.4 147.9 146.3 148.2 124.0 155.4 114.1 44.7 187.5 83.6 39.7

dd dt ddd ddm dd ddt ddt tdt ddt d t

152.6 172.0 168.6 133.6

d d td

150.5 149.5 139.1

C-3′ C-4′ C-5′

70.5 88.0 61.4

176.1

3.6

a Relative to TMS. b The signals of the sugar carbons were further split into complex multiplets. c An AB system further split into multiplets.

Table 2.

1H

and

proton H-2 (1H) H-5 (1H)

H-7a, H-7b (2H) CHO (1H) H-1′ (1H) H-2′ (1H) H-2′′ (1H) H-3′ (1H) H-4′ (1H) H-5′ (1H) H-5′′ (1H) OH OH

13C

Chemical Shifts (δ)a and Spin-Spin Coupling Constants, JH,H and JC,H (Hz), of Protons and Carbons in Cl-pfA-dR δ 8.61 8.52

5.06, 5.07 9.70 6.35 2.65 2.35 4.41 3.89 3.61 3.54 5.37 4.96

multiplicity

JH,H

m s

mc t t dt ddd m dt dd dd br s br s

14.2 0.5 6.6 13.3, 6.6 13.3, 6.6, 3.6 4.5, 3.1 11.7, 4.5 11.7, 4.5

carbonb

δ

multiplicity

1J C,H

C-2 C-5 C-3a C-10a C-10b C-9 C-8 C-7 CHO C-1′ C-2′

142.6 147.3 147.5 147.6 123.7 155.0 104.8 46.4 185.1 83.8 39.5

dd d ddd d dd t dt tdd dt d t

216.2 214.2

149.0 176.9 172.8 134.0

d d td

148.5 149.0 140.3

C-3′ C-4′ C-5′

70.4 88.0 61.3

>1J C,H

4.2 13.4, 5.4, 2.8 4.9 11.4, 1.5 3.4 23.8, 5.3 3.9, 2.8 1.4

3.2

a Relative to TMS. b The signals of the sugar carbons were further split into complex multiplets. c An AB system further split into multiplets.

The reaction was allowed to proceed at 37 °C for 8 days. The reaction products (pfA-dR and Cl-pfA-dR) were isolated from the reaction mixture and purified in the same way as described for the compounds in the preparative scale reaction mixture. Determination of Product Yields. Quantitative 1H NMR analysis, using 1,1,1-trichloroethane as an internal standard, was performed on aliquots of the adducts. Standard solutions were made for HPLC analyses by taking an exact volume of the NMR sample and diluting it with an appropiate volume of water. The quantitative determination of the adducts in the reaction mixtures was made by comparing the peak area of the adducts in the standard solutions with the area of the adduct peaks in the reaction mixtures. The adducts were quantified using UV detection at 400 nm. The molar yields were calculated from the original amount of 2′-deoxyadenosine in the reaction mixture. Reactions of MX with Calf Thymus DNA. MX (18.25 mg) was reacted with double-stranded calf thymus DNA (3.75 mg) in 1.5 mL of 0.1 M phosphate buffer at pH 7.4. The mixture was stirred and incubated at 37 °C for 4 days. The pH of the incubation mixture was monitored daily and readjusted when necessary. The modified DNA was recovered by precipitation with cold ethanol. To the incubation mixture were added 0.2 mL of 5 M NaCl and 3 mL of cold 96% ethanol. This mixture was centrifuged (10 min, 3000 rpm), and the supernatant was removed. The precipitated DNA was washed with 1 mL of 70% ethanol and then redissolved in 1.5 mL of water. This precipi-

tation/washing procedure was performed (at least twice) until there was no more unreacted MX left in the supernatant (controlled by HPLC analyses). The enzymatic hydrolysis of the DNA was carried out following essentially the procedure described by Martin et al. (37). Briefly, the modified DNA was dissolved in 2.5 mL of 0.1 M phosphate buffer, pH 7.4, containing 5 mM of MgCl2. DNase I (dissolved at 10 mg of DNase/mL in 0.9% NaCl) was added to obtain 0.1 mg of DNase/mL. The mixture was incubated and stirred for 3 h at 37 °C. Nuclease P1 (dissolved at 0.5 mg of nuclease P1/mL in 1 mM ZnCl2) was added to obtain 20 µg of nuclease/mL as the final concentration. Finally, alkaline phosphatase (87 U/mL in water) and acid phosphatase (20 U/mL in water) were added to give final concentrations of 0.5 and 0.3 U/mL, respectively. The mixture was then incubated and stirred at 37 °C for 18 h. The mixture of the hydrolyzed DNA was rotary evaporated to near dryness. The residue was washed four times with 2.5 mL of ethanol/ methanol (1/1). The washes were combined and insoluble particles were removed by centrifugation (20 min, 3000 rpm). Finally, the solution was evaporated to near dryness, 0.1 mL of water was added, and 20 µL of the solution was injected on the analytical C18 HPLC columns. To ensure the quantitative extraction of the adducts by ethanol/methanol, the residual insoluble particles were dissolved in water and the solutions were subjected to HPLC analyses. Blanks. A blank sample was prepared by allowing calf thymus DNA to stand for 2 days at 37 °C and then performing

MX Forms Propenoformyl Adducts with 2′-dAdo

Chem. Res. Toxicol., Vol. 11, No. 3, 1998 229

Figure 1. C18 column HPLC separation of the reaction mixture of MX and 2′-deoxyadenosine held at 37 °C and pH 7.4 for 8 days. The retention time of 2′-deoxyadenosine is 5.0 min. For analysis conditions, see Materials and Methods.

Chart 1

Figure 2. (A) UV absorbance and (B) fluorescence emission spectra of pfA-dR (λexc 379 nm) and Cl-pfA-dR (λexc 381 nm). The UV spectra were recorded with the diode-array detector as the compounds eluted from the HPLC column.

the precipitation, enzymatic hydrolysis, and HPLC sample preparation exactly as described above.

Results and Discussion HPLC analyses of the small-scale reactions of MX with 2′-deoxyadenosine showed the formation of two previously unidentified product peaks with longer retention times than that of 2′-deoxyadenosine (Figure 1). The compound marked pfA-dR eluted from the reversed-phase C18 column at 10.5 min and compound Cl-pfA-dR at 14.5 min. In the reaction carried out for 8 days at pH 7.4 and 37 °C, the yields of both compounds were about 0.03 mol %. At pH 6.0 only trace amounts of the compounds were detected, and they were not observed in the reaction mixture held at pH 4.6. For the purpose of determining the structures of the two compounds, a large-scale reaction was performed and the compounds were isolated from the reaction mixture by preparative C18 column chromatography and HPLC. On the basis of compiled data from UV, fluorescence, and NMR spectroscopy and mass spectrometry, it was concluded that in pfA-dR a formylpropeno bridge and in ClpfA-dR a chloroformylpropeno bridge had been incorporated between N-1 and N6 of 2′-deoxyadenosine (Chart 1). The UV spectra of the two compounds were very similar (Figure 2A). A bathochromic shift was observed for Cl-pfA-dR (UVmax 414 nm) relative to pfA-dR (UVmax 408 nm). This slight red shift may be explained by resonance interactions of the unshared chlorine electrons with the conjugated aldehyde function (38). The fluorescence emission spectra of the products were almost identical, exhibiting emission maxima in the

visible region (Figure 2B). The compound pfA-dR showed an emission maximum at 462 nm when it was excited with light at 379 nm. A slight red shift was observed for the emission maximum of Cl-pfA-dR relative to pfAdR. The chloro compound showed an emission maximum at 466 nm when excited at 381 nm. In the positive ion electrospray mass spectrum of pfAdR and Cl-pfA-dR, the protonated molecular ion peaks were observed at m/z 318 and 352, respectively. The presence of one chlorine atom in Cl-pfA-dR was indicated by the presence of an ion two mass units higher than MH+ in an abundance of 33% of the MH+ ion. The 1H NMR spectrum of pfA-dR displayed, besides the signals of the protons of the deoxyribose moiety, oneproton signals at δ ) 9.28, 8.49, 8.31, and 7.50 ppm (Table 1). Moreover, signals displaying an AB pattern were centered at δ ) 5.005 ppm. The signal at δ ) 9.28 ppm was assigned to the formyl proton on the basis of the downfield chemical shift and the one-bond shift correlation with the carbon signal at δ ) 187.5 ppm. The signals of the AB protons at δ ) 5.01 and 5.00 ppm both correlated with a carbon signal at δ ) 44.7 ppm and were assigned to the two geminal protons of a methylene group (H-7a and H-7b). The signals of the AB pattern were further split by several weak, nonresolved couplings. The signal of the formyl proton and the signals at δ ) 8.31 and 7.50 ppm displayed splittings that disappeared on irradiation at δ ) 5.005 ppm. Consequently, splittings were due to weak couplings to the methylene protons. The signals at δ ) 8.31 and 7.50 ppm were thus assigned to H-5 and H-9. The resonance at δ ) 8.31 ppm could be attributed H-5 by an observed NOE interaction between H-7 and H-5 (2-D NOESY experiment). Consequently, the signal at δ ) 7.50 ppm was assigned to

230 Chem. Res. Toxicol., Vol. 11, No. 3, 1998

Munter et al.

Table 3. 1H-13C Connectivities for pfA-dR Established by 2-D NMR Heterocorrelation Spectra (CHSHF and COLOC) C-2

C-3a

C-5

C-7 C-8 C-9 C-10a C-10b

CHO

H-2 CH CNCH CNCH H-5 CNCH CH H-7a CNCH CH CCH H-7b CNCH CH CCH H-9 CCH CH CNCH CCCH CHO CCH CH

H-9. This assignment was confirmed by the observed NOE interaction of H-9 with the proton in the aldehyde group. The signal at δ ) 8.49 ppm could then be assigned to H-2, and the assignment was confirmed by the observed H-H correlation with the H-1′ signal in the deoxyribose moiety (δ ) 6.32 ppm). H-2 and H-1′ also displayed a NOE interaction. The other proton signals of the deoxyribose moiety (H-2′-H-5′) were assigned using H-H correlation data. The shifts and couplings are based on a first-order interpretation. As the proton signals could be unambiguously assigned, the 13C NMR signals of the formyl carbon, the olefinic methine carbons (C-2, C-5, and C-9), the methylene carbon (C-7), and the carbons of the deoxyribose moiety (C-1′-C-5′) were assigned using data from the C-H correlation (CHSHF and COLOC) spectra and the fully proton-coupled carbon spectrum (Table 1). One- and multiple-bond connectivities are shown in Table 3. The assignment of the signal at δ ) 114.1 ppm to C-8 was confirmed by its strong two-bond C-H coupling (2J ) 24.8 Hz) to the formyl proton (39). The further splitting of the signal to a doublet of triplets is then due to the couplings to H-9 and H-7, respectively. The signal of C-8 also showed C-H connectivities with the signals of H-7, H-9, and the formyl proton. The signal at δ ) 146.3 ppm was assigned to C-3a based on its C-H correlation with the signals of H-2 and H-5. Strong three-bond couplings can be expected as H-2 and H-5 both are trans to C-3a. Due to the π-bond between N-4 and C-5, the larger coupling is probably to H-5 (39). The signal at δ ) 124.0 ppm was assigned to C-10b based on the chemical shift and the correlation with the signal of H-2. The signal showed a three-bond coupling of 3J ) 11.6 Hz. A coupling of this magnitude can be expected between C-10b and H-2 due to their trans relationship. The signal at δ ) 148.2 ppm, assigned to C-10a, showed correlation with the signal of H-9 and was split into a doublet of doublets due to two three-bond C-H couplings of 10.9 and 5.7 Hz. Three-bond C-H couplings can be expected from C-10a to H-5 and H-9 due to trans relation of the carbon to these hydrogens. A stronger coupling is expected to H-9 than to H-5 due to the presence of the π-bond between C-10a and N-10. The suggested structure was supported by the observation of small triplet splittings in the signals of C-5, C-8, C-9, and the formyl carbon. These splittings must be derived from couplings to the protons (H-7) of the methylene group. The 1H NMR spectrum of Cl-pfA-dR showed essentially the same features as the spectrum of pfA-dR (Table 2). The low-field part of the spectrum, however, showed only three one-proton signals. The signals were found at δ ) 9.70, 8.61, and 8.52 ppm and were assigned to the formyl proton, H-2, and H-5, respectively. An AB system (JAB ) 14.2 Hz), centered at δ ) 5.065 ppm and further split by weak couplings, showed the presence of a methylene group (H-7a and H-7b). The signals of the formyl proton

and H-5 showed small splittings, which disappeared on irradiation at δ ) 5.065 ppm. The assignment of the signal at δ ) 8.61 ppm to H-2 was confirmed by its H-H connectivity to the H-1′ signal at δ ) 6.35 ppm. The formyl proton signal as well as the signal of H-5 showed H-H correlation with the signals of H-7. The position of the methylene group was confirmed by a NOE interaction observed between H-5 and H-7. NOE correlation was also observed between H-2 and H-1′. The assignment of the carbon resonances of Cl-pfAdR (Table 2) was based on a discussion analogous to the discussion for pfA-dR. The carbon resonances of all protonated carbons were assigned using the one-bond C-H correlation data. The signal at δ ) 147.3 ppm, assigned to C-5, showed a C-H correlation with the signals of H-7. The signal at δ ) 104.8 ppm was assigned to C-8 based on the observed strong coupling to the formyl proton (2J ) 23.8 ppm). The signal of C-8 also correlated with the signals of H-7. The upfield shift of C-8 relative to that of C-8 in pfA-dR can be ascribed to the +R effect of the chlorine atom at C-9. This extended conjugation was also reflected in the UV absorption. The signal at δ ) 147.5 ppm was assigned to C-3a, and the assignment was based mainly on the long-range coupling pattern which was almost the same as that of the C-3a in pfAdR. On the basis of the chemical shift and the correlation with the signal of H-2, the signal at δ ) 123.7 ppm was assigned to C-10b. This signal also showed a three-bond coupling of 3J ) 11.4 ppm. The signal at δ ) 147.6 ppm, assigned to C-10a, displayed only one coupling of 3J ) 4.9 Hz as expected (cf. the corresponding signal of pfAdR). The signal at δ ) 155.0 ppm was assigned to C-9 based on comparison with the spectrum of pfA-dR. The signal of C-9 and the signals of C-5, C-8, and the formyl carbon showed small triplet splittings due to coupling with the methylene protons at C-7. The NMR and UV spectroscopic and the mass spectrometric data are consistent with the structures of pfA-dR and Cl-pfA-dR presented in Chart 1. A mechanistic explanation for the formation of pfAdR and Cl-pfA-dR was obtained by reacting 13C-3-labeled MX with 2′-deoxyadenosine. The 13C NMR spectra of the labeled adducts showed a significantly more intensive signal for the methylene carbon (C-7) than for any other carbon present in the spectra. This finding provided a plausible mechanism for the formation of the products. Initially, a proton transfer takes place from the dichloromethyl carbon in the open form of MX to the labeled carbon atom (Scheme 2). Next, the exocyclic amino group of adenosine attacks the former dichloromethyl carbon via a conjugate addition reaction, and the enol A is obtained. Subsequent displacement of HCl and decarboxylation will yield the intermediate B. The terminal chlorine in B is displaced by a conjugate attack via the endocyclic nitrogen (N-1) of adenine, and Cl-pfA-dR is obtained. The mechanism by which the chlorine at C-9 is replaced by a hydrogen as in pfA-dR is unknown. An alternative mechanism for the formation of the compound cannot be excluded. In a preliminary work, we have found that the compound 3-chloro-4-(chloromethyl)-5hydroxy-2(5H)-furanone (CMCF) forms pfA-dR when reacted with 2′-deoxyadenosine. In CMCF a chloromethyl group is located at C-4 in the furanone, and the mechanism for the formation of pfA-dR from CMCF is analogous to the formation of Cl-pfA-dR from MX. The yield of pfA-dR from CMCF (about 1 mol %) is much

MX Forms Propenoformyl Adducts with 2′-dAdo

Chem. Res. Toxicol., Vol. 11, No. 3, 1998 231

Scheme 2

higher than from MX, but since we could not find even traces of CMCF (detection limit about 0.2% by GC/MS) in the pure MX used in the current study, we exclude the possibility of CMCF accounting for the formation of pfA-dR. In a previous contribution from this laboratory, it was shown that when 3-chloro-4-methyl-5-hydroxy-2(5H)furanone (MCF, a furanone with a methyl group at C-4 in the furanone ring) was reacted with adenosine, the compound 4-(N6-adenosinyl)-3-formyl-3-butenoic acid was formed (33). The initial steps in the mechanism for the formation of this product are the same as for formation of pfA-dR and Cl-pfA-dR from MX. However, decarboxylation and ring formation do not take place in the reaction of MCF since the only chlorine in the compound is lost by displacement of HCl and not by an attack through the N-1 nitrogen of the adenine unit. Previously, Roques et al. and Petra et al. prepared fluorescent [2,1-i]pyrimidopurine derivatives (40, 41). In the paper of Petra et al. (41) it was reported that the compound, where a carbonyl group was located at C-7 and a carboxymethyl group was bound to C-9, exhibited strong fluorescence in the UV/vis region: λexc 370 nm and λem 420 nm. Reaction of MX with Calf Thymus DNA. HPLC analyses of the hydrolysate of the calf thymus DNA reacted with MX showed the presence of a peak at exactly the same retention time, 10.5 min, as that of pfA-dR. Following spiking the DNA hydrolysate with pure pfAdR, it was found that the compounds coeluted when chromatographed on a short (125-mm) and a longer (250mm) C18 column and also on a C8 column (Figure 3). Since also the UV spectra of the adduct in the hydrolysate and of pure pfA-dR were in all essential features identical, we conclude that pfA-dR was formed in the reaction of MX with calf thymus DNA. The yield of formation of pfA-dR in DNA was about 0.6 adduct/105 nucleotides. When HPLC analyses were performed using the fluorescence detector, it was found that the sensitivity for detection of pfA-dR was more than 100 times higher than when using the UV detector. A blank sample of calf thymus DNA was treated in exactly the same way as the DNA incubated with MX. The chromatograms obtained by HPLC analysis of the blank sample showed no peaks that could represent the adducts formed by MX.

Figure 3. C18 analytical column (column length 125 mm) HPLC separation of the enzymatically hydrolyzed DNA after incubation of the DNA with MX: (A) detection with the UV detector and (B) detection with the fluorescence detector, λexc 379 nm. The chromatograms of the DNA hydrolysate spiked with pfA-dR are superimposed on the original chromatograms. For analysis conditions, see Materials and Methods.

Although it was not possible to detect the adduct ClpfA-dR in the DNA hydrolysate, we cannot exclude the possibility that trace amounts of this adduct are formed in DNA. Both pfA-dR and Cl-pfA-dR are novel adducts that could represent DNA lesions leading to mutations. In these adducts, sites N-1 and N6 of adenine are blocked and thus base pairing with thymine is hindered. MX is an extremely potent, direct-acting mutagen in the S. typhimurium strain TA100. The mutational events in this strain take place mainly at the hisG46 allele, where the target sequence is CCC (42). It has been shown that MX induces primarily GC f TA transversions in TA100 (43). These transversions cannot be due to the formation

232 Chem. Res. Toxicol., Vol. 11, No. 3, 1998

of adducts at the adenine moiety in DNA, and therefore, it seems unlikely that pfA-dR and Cl-pfA-dR would be important premutagenic lesions in S. typhimurium strain TA100. However, Knasmu¨ller et al. (43) have also found AT f TA transversions in the mutational spectrum of MX in S. typhimurium strain TP2428. It is possible that the propenoformyl derivatives identified in this study act in that strain as mediators of the genotoxic effects of MX. A most interesting question is the role these adducts may play in the generation of the various forms of cancer observed in rodents (25).

Conclusion The results of this work show that MX modifies the base unit of 2′-deoxyadenosine, by forming fluorescent propenoformyl adducts. One of the adducts (pfA-dR) was formed also upon incubation of MX with calf thymus DNA. The identified adducts are novel and could represent premutagenic lesions that act as mediators of MX mutagenicity in S. typhimurium strain TP2428. The fluorescence properties of the propenoformyl adducts allow for sensitive detection of the adduct(s) in DNA hydrolysates and cellular systems and could enable studies of the role the adducts may play in the carcinogenicity of MX in rats.

Acknowledgment. This work was supported by a research grant from the Maj and Tor Nessling Foundation in Finland (Tony Munter) and by the European Commission (Contract Nr ERBFMBICT961394, Dr. Frank Le Curieux).

References (1) Loper, J. C. (1980) Mutagenic effects of organic compounds in drinking water. Mutat. Res. 76, 241-268. (2) Hemming, J., Holmbom, B., Reunanen, M., and Kronberg, L. (1986) Determination of the strong mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone in chlorinated drinking and humic waters. Chemosphere 15, 549-556. (3) Meier, J. R., Knohl, R. B., Coleman, W. E., Ringhand, H. P., Munch, J. W., Kaylor, W. H., Streicher, R. P., and Kopfler, F. C. (1987) Studies on the potent bacterial mutagen 3-chloro-4(dichloromethyl)-5-hydroxy-2(5H)-furanone: Aqueous stability, XAD recovery and analytical determination in drinking water and in chlorinated humic acid solutions. Mutat. Res. 189, 363-373. (4) Holmbom, B., Voss, R. H., Mortimer, R. D., and Wong, A. (1984) Fractionation isolation, and characterization of Ames mutagenic compounds in kraft chlorination effluents. Environ. Sci. Technol. 18, 333-337. (5) Padmapriya, A. A., Just, G., and Lewis, N. (1985) Synthesis of 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone, a potent mutagen. Can. J. Chem. 63, 828-832. (6) Kronberg, L., and Franze´n, R. (1993) Determination of chlorinated furanones, hydroxyfuranones, and butenedioic acids in chlorine treated water and in pulp bleaching liquor. Environ. Sci. Technol. 27, 1811-1818. (7) Smeds, A., Vartiainen, T., Ma¨ki-Paakkanen, J., and Kronberg, L. (1997) Concentrations of Ames mutagenic chlorohydroxyfuranones and related compounds in drinking water. Environ. Sci. Technol. 31, 1033-1039. (8) Kringstad, K. P., Ljungquist, P. O., de Sousa, F., and Stro¨mberg, L. (1983) On the formation of mutagens in the chlorination of humic acid. Environ. Sci. Technol. 17, 553-555. (9) Fielding, M., and Horth, H. (1986) Formation of mutagens and chemicals during water treatment chlorination. Wat. Supply 4, 103-126. (10) Peters, R. J. (1991) Chemical aspects of drinking water chlorination, Ph.D. Thesis, Technical University of Delft, The Netherlands. (11) Långvik, V.-A., Hormi, O., Tikkanen, L., and Holmbom, B. (1991) Formation of the mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy2(5H)-furanone and related compounds by chlorination of phenolic compounds. Chemosphere 22, 547-555.

Munter et al. (12) Tikkanen, L., and Kronberg, L. (1990) Genotoxic effects of various chlorinated butenoic acids identified in chlorinated drinking water. Mutat. Res. 240, 109-116. (13) Meier, J. R., Blazak, W. F., and Knohl, R. B. (1987) Mutagenic and clastogenic properties of 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone: A potent bacterial mutagen in drinking water. Environ. Mol. Mutagen. 10, 411-424. (14) LaLonde, R. T., Cook, G. P., Pera¨kyla¨, H., and Dence, C. W. (1991) Effect on mutagenicity of the stepwise removal of hydroxyl group and chlorine atoms from 3-chloro-4-(dichloromethyl)-5-hydroxy2(5H)-furanone: 13C NMR chemical shifts as determinants of mutagenicity. Chem. Res. Toxicol. 4, 35-40. (15) Kronberg, L., and Vartiainen, T. (1988) Ames mutagenicity and concentration of the strong mutagen 3-chloro-4-(dichloromethyl)5-hydroxy-2(5H)-furanone and of its geometric isomer E-2-chloro3-(dichloromethyl)-4-oxo-butenoic acid in chlorine-treated tap waters. Mutat. Res. 206, 177-182. (16) Horth, H. (1990) Identification of mutagens in water. J. Fr. Hydr. 21, 135-145. (17) Andrews, R. C., Daignault, S. A., Laverdure, C., Williams, D. T., and Huck, P. M. (1990) Occurrence of the mutagenic compound “MX” in drinking water and its removal by activated carbon. Environ. Technol. 11, 685-694. (18) Suzuki, N., and Nakanishi, J. (1990) Determination of strong mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone in drinking water in Japan. Chemosphere 21, 387-392. (19) Zou, H., Xu, X., Zhang, J., and Zhu, Z. (1995) The determination of MX [3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone] in drinking water in China. Chemosphere 30, 2219-2225. (20) Chang, L. W., Daniel, F. B., and DeAngelo, A. B. (1991) DNA strand breaks induced in cultured human and rodent cells by chlorohydroxyfuranones - mutagens isolated from drinking water. Teratogen. Carcinogen. Mutagen. 11, 103-114. (21) Jansson, K., and Hyttinen, J. M. T. (1994) Induction of gene mutation in mammalian cells by 3-chloro-4-(dichloromethyl)-5hydroxy-2(5H)-furanone (MX), a chlorine disinfection byproduct in drinking water. Mutat. Res. 322, 129-132. (22) Harrington-Brock, K., Doerr, C. L., and Moore, M. M. (1995) Mutagenicity and clastogenicity of 3-chloro-4-(dichloromethyl)5-hydroxy-2(5H)-furanone (MX) in L5178Y/TK+/--3.7.2C mouse lymphoma cells. Mutat. Res. 348, 105-110. (23) Furihata, C., Yamashita, M., Kinae, N., and Matsushima, T. (1992) Genotoxicity and cell proliferative activity of 3-chloro-4(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) in rat glandular stomach. Wat. Sci. Tech. 25, 341-345. (24) Jansson, K., Ma¨ki-Paakkanen, J., Vaittinen, S.-L., Vartiainen, T., Komulainen, H., and Tuomisto, J. (1993) Cytogenetic effects of 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone (MX) in rat peripheral lymphocytes in vitro and in vivo. Mutat. Res. 299, 25-28. (25) Komulainen, H., Kosma, V.-M., Vaittinen, S.-L., Vartiainen, T., Kaliste-Korhonen, E., Lo¨tjo¨nen, S., Tuominen, R. K., and Tuomisto, J. (1997) Carcinogenicity of the drinking water mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)-furanone in the rat. J. Natl. Cancer Inst. 89, 848-856. (26) Morris, R. D., Audet, A.-M., Angelillo, I. F., Chalmers, T. C., and Mosteller, F. (1992) Chlorination, chlorination byproducts and cancer: A meta-analysis. Am. J. Public Health 82, 955-963 [published erratum appears in (1993) Am. J. Public Health 83, 1257]. (27) Cantor, K. P., Hoover, R., Hartge, P., Mason, T. J., Silverman, D. T., Altman, R., Austin, D. F., Child, M. A., Key, C. R., Marrett, L. D., Myers, M. H., Narayana, A. S., Levin, L. I., Sullivan, J. W., Swanson, G. M., Thomas, D. B., and West, D. W. (1987) Bladder cancer, drinking water source and tap water consumption: A case-control study. J. Natl. Cancer Inst. 79, 1269-1279. (28) Koivusalo, M., Jaakkola, J. J. K., Vartiainen, T., Hakulinen, T., Karjalainen, S., Pukkala, E., and Tuomisto, J. (1994) Drinking water mutagenicity and cancers of the gastrointestinal and urinary tract: An ecological study in Finland. Am. J. Public Health 84, 1223-1228. (29) Harris, C. C. (1991) Chemical and physical carcinogenesis: Advances and perspectives for the 1990s. Cancer Res. 51, 5023s5044s. (30) Kronberg, L., Sjo¨holm, R., and Karlsson, S. (1992) Formation of 3,N4-ethenocytidine, 1,N6-ethenoadenosine, and 1,N2-ethenoguanosine in reactions of mucochloric acid with nucleosides. Chem. Res. Toxicol. 5, 852-855. (31) Kronberg, L., Karlsson, S., and Sjo¨holm, R. (1993) Formation of ethenocarbaldehyde derivatives of adenosine and cytidine in reactions with mucochloric acid. Chem. Res. Toxicol. 6, 495-499.

MX Forms Propenoformyl Adducts with 2′-dAdo (32) Asplund, D., Kronberg, L., Sjo¨holm, R., and Munter, T. (1995) Reactions of mucochloric acid with adenosine: Formation of 8-(N6adenosinyl)ethenoadenosine derivatives. Chem. Res. Toxicol. 8, 841-846. (33) Munter, T., Kronberg, L., and Sjo¨holm, R. (1996) Identification of the adducts formed in reactions of adenosine with 3-chloro-4methyl-5-hydroxy-2(5H)-furanone (MCF), a bacterial mutagen present in chlorine disinfected drinking water. Chem. Res. Toxicol. 9, 703-708. (34) Kronberg, L., Asplund, D., Ma¨ki, J., and Sjo¨holm, R. (1996) Reactions of mucochloric and mucobromic acid with adenosine and cytidine: Formation of chloro- and bromopropenal derivatives. Chem. Res. Toxicol. 9, 1257-1263. (35) Le Curieux, F., Munter, T., and Kronberg, L. (1997) Identification of adenine adducts formed in reaction of calf thymus DNA with mutagenic chlorohydroxyfuranones found in drinking water. Chem. Res. Toxicol. 10, 1180-1185. (36) Franze´n, R., and Kronberg, L. (1995) Synthesis of chlorinated 4-methyl-5-hydroxy-2(5H)-furanones and mucochloric acid. Tetrahedron Lett. 36, 3905-3908. (37) Martin, C. N., Beland, F. A., Roth, R. W., and Kadlubar, F. F. (1982) Covalent binding of benzidine and N-acetylbenzidine to DNA at the C-8 atom of deoxyguanosine in vivo and in vitro. Cancer Res. 42, 2678-2696.

Chem. Res. Toxicol., Vol. 11, No. 3, 1998 233 (38) Rao, C. N. R. (1961) Ultra-Violet and Visible Spectroscopy, p 16, Butterworth, London. (39) Kalinowski, H.-O., Berger, S., and Braun, S. (1994) Carbon-13 NMR Spectroscopy, Chapter 4, John Wiley & Sons, Chichester. (40) Roques, P., Le Gall, J. Y., Lacombe, L., and Olomucki, M. (1992) Reaction of nucleic acid bases with R-acetylenic esters. 5. Synthesis and properties of adenosine and cytidine derivatives. J. Org. Chem. 57, 1579-1585. (41) Petra, D. G. I., de Boer, R. F., Koomen, G. J., Meeuwenoord, N. J., Kuyl-Yeheskiely, E., van der Marel, G. A., and van Boom, J. H. (1996) Synthesis and properties of a fluorescent nucleotide derivative. Recl. Trav. Chim. Pays-Bas 115, 99-102. (42) Cebula, T. A., and Koch, W. H. (1990) Analysis of spontaneous and psoralen-induced Salmonella typhimurium hisG46 revertants by oligodeoxyribonucleotide colony hybridization: Use of psoralens to cross-link probes to target sequences. Mutat. Res. 229, 79-87. (43) Knasmu¨ller, S., Zo¨hrer, E., Kronberg, L., Kundi, M., Franze´n, R., and Schulte-Hermann, R. (1996) Mutational spectra of Salmonella typhimurium revertants induced by chlorohydroxyfuranones, byproducts of chlorine disinfection of drinking water. Chem. Res. Toxicol. 9, 374-381.

TX970195X