Genotoxicity of neurotoxic triaryl phosphates - American Chemical

Nov 4, 1992 - Chem. Res. Toxicol. 1993, 6, 294-301. Genotoxicity of Neurotoxic Triaryl Phosphates: Identification of DNA Adducts of the Ultimate Metab...
0 downloads 0 Views 1MB Size
Chem. Res. Toxicol. 1993,6, 294-301

294

Genotoxicity of Neurotoxic Triaryl Phosphates: Identification of DNA Adducts of the Ultimate Metabolites, Saligenin Phosphates Anke Mentzschel, Gabi Schmuck, Wolfgang Dekant,* a n d Dietrich Henschler Institut fur Toxikologie, Universitat Wurzburg, Versbacherstrasse 9,0-8700 Wurzburg, Federal Republic of Germany Received November 4, 1992

2-Phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide is an electrophilic and neurotoxic metabolite of o-tolyl phosphates. We have investigated the genotoxicity of this saligenin phosphate and the structure of adducts formed by incubation of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide with nucleosides and DNA. o-Tolyl phosphate was mutagenic in the Ames test (695 revertants/pmol, Salmonella typhimurium T A 100) only with metabolic activation. 2-Phenoxy4H-1,3,2-benzodioxaphosphorin 2-oxide, which is a cyclization product similar to those expected from o-tolyl phosphate, was a potent mutagen in bacteria (1452 revertants/bmol, S. typhimurium TA 100) which did not require metabolic activation. Incubation of 2-phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide with guanosine, deoxycytidine, and deoxyadenosine resulted in formation of guanosine, deoxyuridine, and adenine adducts. These were identified as N-2(o-hydroxybenzyl)guanosine,N-3-(o-hydroxybenzyl)deoxyuridine,N-l-(o-hydroxybenzy1)adenine, and N-3-(o-hydroxybenzyl)adenineby IH-NMR spectroscopy, thermospray mass spectrometry, and pH-dependent electronic spectrometry. The deoxyuridine adduct is formed by an alkylation a t N-3 gf deoxycytidine followed by conversion of the adjacent exocyclic imino group to carbonyl (hydrolytic deamination). The formation of N-2-(o-hydroxybenzyl)deoxyguanosine, N-3- (o-hydroxybenzyl)deoxyuridine, and N- 140-hydroxybenzy1)deoxyadenosine was also demonstrated when 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide was incubated with calf thymus DNA. Adducts formed with nucleosides in calf thymus DNA reacted with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide in vitro were detected by the 32P-postlabeling technique and identified by comparison with synthetic references. DNA adducts are formed by an o-hydroxybenzylation from cyclic phosphoranes derived from o-alkyl-substituted triaryl phosphates. The results indicate that some o-tolyl phosphates, besides their neurotoxic effects, may have a genotoxic and carcinogenic potential. Introduction Triaryl phosphates are industrial chemicals and widely used as plasticizers in lacquers and varnishes, in highspeed brake fluids, as hydraulic fluids, as flame retardants, and as heat exchanger fluids (1). They have a mild cholinergic effect and induce a delayed neurotoxic effect in several species, including humans (2-4). o-Alkylsubstituted triaryl phosphates such as o-tolyl phosphate (Figure 1) are metabolized to several products in vivo and in vitro. One of these metabolites, 2-(o-tolyloxy)-4H-l,3,2benzodioxaphosphorin 2-oxide (lb, Figure l),has been shown to be 5 times as neurotoxic as the parent compound (5). The initial step in 2-(o-tolyloxy)-4H-l,3,2-benzodioxaphosphorin 2-oxide formation is a cytochrome P-450 dependent oxidation of a methyl group to organophosphate IIa followed by a serum protein catalyzed cyclization to give 2-(o-tolyloxy)-4H-l,3,2-benzodioxaphosphorin 2-oxide (6). Low, but toxicologically significant, amounts of this metabolite were found in liver, kidney, and testes of rats after administration of high doses of o-tolyl phosphate (7). A derivative, 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide IIIa (Figure l),is an electrophilic organophosphateand can be regarded as a reactive benzyl ester.

* Address correspondence to this author at the Institut fur Toxikologie, Versbacherstr. 9,D-8700Wiirzburg, FR Germany. Phone: 49-931-2013449;FAX: 49-931-201-3446.

R

R'

@

H3

la, Ib

R

R'

@ 118,

H2-oH

Ilb

Illa, lllb

Figure 1. Mechanism of formation of dioxaphosphorins from o-tolyl phosphates. o-Tolyldiphenyl phosphate (R = R' = H, Ia) is oxidized to IIa, catalyzed by a microsomal mixed function oxidase. IIa is converted to 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide (IIIa), catalyzed by plasma enzymes. An identical cyclization reaction occurs with-tri-o-tolyl phosphate (R = R = CH3, Ib, IIb, IIIb).

Hydrolytic ring cleavage of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide results in formation of o-hydroxybenzyl phosphate ester which benzylates nucleophiles. Thus, certain oximes and sulfur-containing

0893-228~/93/2706-0294$04.~0/00 1993 American Chemical Society

Genotoxicity and DNA Binding of Organophosphates

Chem. Res. Toxicol., Vol. 6, No. 3,1993

295

Electronic spectra of aqueous solutions a t pH 2 (1N HCl),pH 6 (H20), and pH 12 (1 N NaOH) were recorded on a Kontron spectrophotometer (Rotkreuz, Switzerland) and with a diodearray detector coupled to an HPLC system. Mutagenicity Assay. The Salmonella typhimurium tester strain TA 100, provided by Dr. B. N. Ames (Berkeley, CA), was used for all experiments; UV and crystal violet sensitivity, ampicillinresistance, and mutability were checked regularly. The mutagenicity assay employed was a plate test as described in the literature (IO). Cultures were grown in Oxoid Nutrient Broth No. 2 for 10 h. Spontaneous revertant frequencies were comparable with values established in our laboratory (11). The plate test was performed with 500 pL of S-9 mix or 500 pL of 0.1 M phosphate buffer (pH 7.2), 100 pL of the bacterial culture, and 100 pL of a solution of the test compound in methanol. After shaking, 2 mL of top agar (45 "C) containing histidine (0.05 mM) and biotin (0.05 mM) were added and the mixture was plated on Vogel and medium. After 2 days of incubation, colonies of revertants were counted with an automated colony counter, and the counts were correlated for overlapping colonies with a computer program. All determinations were made in duplicate, and all experiments were performed twice. Experimental Procedures Synthesis of 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin Caution: Saligenin phosphates are potent mutagens and 2-Oxide. The synthetic procedure was modified from ref 12. highly neurotoxic. They are readily absorbed through the skin. 2-Hydroxybenzyl alcohol (80 mmol) was dissolved in 40 mL of All manipulations should be performed under an efficient hood dry pyridine. Phenyl dichlorophosphate (81 mmol) was added using proper safety equipment. dropwise under stirring and cooling in an ice bath. During Materials. 2-Hydroxybenzylalcohol and phenyl chlorophosaddition, the temperature in the reaction mixture was kept phate were obtained from Aldrich Chemical (Steinheim, FRG) between 0 and -5 OC. The mixture was then kept a t 0 "C for 1 and were used without further purification. Micrococcalnuclease h and at 22 "C for 3 h. Then, 70 mL of cold 10% hydrochloric (EC 3.1.31.1, grade IV), spleen exonuclease(EC 3.1.161), adenine, acid was added and the mixture was extracted three times with nucleosides, deoxynucleotides, and deoxynucleosides were pur50 mL of chloroform each. The organic phases were separated, chased from Sigma Chemical Co. (Deisenhofen, FRG). [ Y - ~ ~ P I - pooled, dried over sodium sulfate, and concentrated under ATP was purchased from Amersham (Braunschweig,FRG). T4 reduced pressure. The residue was kept at -20 "C for crystalpolynucleotide kinase (EC 2.71.78) and calf thymus DNA were 2-oxide was lization. 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin obtained from Boehringer (Mannheim, FRG). HPLC-grade recrystallized from diethyl ether. The yield was 2.1 g, 10% of methanol and NMR supplies were obtained from Merck (Dumthe theoretical yield; mp was 78 "C (lit. mp 76-79 "C). The stadt, FRG). All other chemicals were of the highest purity purity, as determined by 'H-NMR and GC-MS, was 99 7%. available and were purchased from several suppliers. Synthesis of N-2-(o-Hydroxybenzyl)guanosineand N-3Instrumental Analyses. Samples were analyzed with two (o-Hydroxybenzy1)deoxyuridine.One hundred milligrams of HPLC systems: (A) The first was a Hewlett Packard Series 1055 guanosineor deoxycytidinewas reacted with 120mg of 2-phenoxyHPLC system coupled to a Hewlett Packard 1040 diode-array 4H-1,3,2-benzodioxaphosphorin 2-oxide in 40mL of 0.1 M sodium detector. Separations were performed using steel columns filled phosphate buffer (pH 7.4) a t 23 "C for 7 days. The reaction with Partisil ODS I11 (250 X 4 mm) and Supelcosil LC-18s (250 mixtures were then frozen and lyophilized. The residues were X 4 mm). Electronic spectra were recorded in 1.28-9 intervals resuspended in 10 mL of water/methanol (1:lv/v) and filtered. with a threshold setting of 3 mAU. (B) The second was a Waters Two milliliters of each solution was separated by HPLC using HPLC system consisting of two M6000A pumps, a Model 660 system b. The adducts were isolated by HPLC using a 250- x gradient controller, and a U6K injector coupled to a Waters UV 32-mm column. N-3-(o-Hydroxybenzyl)deoxyuridinewas eluted detector (ModelLambda Max 481). Separations were performed with a linear gradient from 0% to 57% methanol in water (pH using steel columns filled with Partisil ODS I11 (10 pm, 250 X 6.3) in 45 min; flow rate was 9.9 mL/min, with a retention time 10 mm and 125 X 10 mm). The absorption of the eluate was was eluted with a of 38 min. N-2-(o-Hydroxybenzyl)guanosine monitored at 254 nm. retention time of 36 min by a linear gradient from 0% to 65% methanol in water pH 6.3 in 40 min; flow rate was 9.9 mL/min. Thermospray mass spectrometry was performed with a FinniThe eluates were collected and lyophilized. gan thermospray interface coupled to a Finnigan 4510 mass spectrometer. Ammonium acetate (0.05 M) in methanol/water (A) N-2-(o-Hydroxybenzyl)guanosine: Electronic spec(1:4 v/v) at a flow rate of 1.3 mL/min was introduced into the MS tra: pH 2, A,, = 262 nm, 278 nm; pH 6, A,, = 255 nm, 275 nm; system with a Waters (Milford, MA) M6000A pump. The pH 13, A,, = 258 nm, 278 nm; lH-NMR (Me2SO) 6 (ppm) = 7.92 vaporizer temperature was 160 "C, and the jet temperature was ( ~ H, , l8-H), 7.2 (dd, J13,14 = 7.4 Hz, 513,15 = 1.5 Hz, 1 H, 13-H), 220 "C. The samples were introduced into the interface by direct 7.1 (dt, 5 1 4 ~ 6= 8.5 Hz, J 1 6 , 1 7 = 8.0 Hz, 1H, 15-H), 6.9 (dd, 516,15 loop injection. Mass spectrometry data were processed by a = 8.0 Hz, 5 1 6 ~ 4= 1.1Hz, 1H, 16-H), 6.7 (dt, J i 4 , 1 3 = 7.5,515,16= Finnigan-INCOS data system. 8.5 HZ, 1 H, i 4 - ~ )5.74 , (d, J ~ , , ~=, 5.37 HZ, 1 H, it-^), 4-48 (t, 52r,1~ = 5.47 Hz, 1H, 2'-H), 4.45-4.40 (m,2 H, 10-H),4.11 (dd, 53,,2, Proton NMR spectra were recorded in 5-mm tubes on a Bruker = 5.08 Hz, 53,,4' = 3.75 Hz, 1H, 3'-H), 3.88 (q,54',3, = 5 4 , , 5 , = 4.0 WM spectrometer (Rheinstetten, FRG) a t 400.136 MHz. 'HHz, 1 H, 4'-H). The 5'-H2 resonance is hidden by the broad NMR spectra were recorded with a spectral width of 6024.09 Hz, signal of water (6 = 3.3-3.5 ppm). an acquisition time of 2.720 s, 32K data points, and a pulse width of 3.0 ps. Chemical shifts are reported in parts per million with (B) N-3-(o-Hydroxybenzyl)deoxyuridine: Electronic spectra: pH 2, A,, = 265.5 nm; pH 6, A,, = 266.5 nm; lH-NMR dimethyl sulfoxide (6 = 2.5 ppm) as internal standard. '3C-NMR spectra were recorded with the same instrument operating at (Me2SO) 6 (ppm) = 8.03 (d, J6,5= 8.1 Hz, 1H, 6-H), 7.1-6.8, (m, 100.627 MHz. Broad-band gated proton decoupling was used. 4 H, 10-H, 11-H, 12-H, 13-H), 5.85 (d, J 5 , 6 = 8.2 Hz, 1 H, 5-H), Spectra were recorded with a spectral width of 25000.00 Hz, an = 6.75, 1 H, l'-H), 4.91 (8, 2 H, 7-H), 4.25 (m, 1 H, 6.18 (t,J1t,2, acquisition time of 0.328 s, 16K data points, and a pulse width , Hz, J4,,5, = 6.9 Hz, 3.6 Hz, 1 H, 4'-H), 3'-H), 3.80 (q,J ~ J ,=~ 6.9 of 2.5 ps. 3.60 (m, 2 H, 5'-H), 2.20 (m, 2 H, 2a-H, 2P-H);W-NMR(Me2SO)

nucleophiles, such as cysteine and glutathione, react with 2-phenoxy-4H-l,3,2-benz~dioxaphosphorin 2-oxide to form the corresponding o-hydroxybenzyl ethers (8,9). These experiments demonstrated that saligenin phosphate IIIa is an electrophile. Since most DNA-reactive carcinogens are electrophiles or are transformed to electrophilic metabolites, 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide was expected to react with DNA and could thus cause mutations. Therefore, we have investigated the genotoxicityof o-tolyl phosphate and 2-phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide and the structure of DNA adducts formed from 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide. The formation of these adducts may be one of the mechanisms responsible for genotoxic effects of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide. The results obtained demonstrate the direct mutagenicity of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide and its ability to react with DNA constituents and DNA in vitro.

296 Chem. Res. Toxicol., Vol. 6, No. 3, 1993 6 (ppm) = 160 (C-4), 153 (C-2),148 (C-9),137 (C-6), 125,123,120,

118, 113 (benzene-C), 99 (C-5), 85 ((2-49, 83 (C-l’), 67 (C-3’),59 (C-5’), 39 (C-2’). The resonance of C-7 coincides with the resonance of the solvent MezSO-d6. Reaction of Deoxycytidine and Deoxyuridine with 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin2-Oxide. Deoxycytidine or deoxyuridine (0.1 mg) was reacted with 0.5 mg of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide in 400 pL of 0.1 M sodium phosphate buffer (pH 7.3) at 37 “C. After 4 days the mixtures were analyzed for the formation of N-~-(ohydroxybenzy1)deoxyuridine by HPLC system a using a 5-pm C18 (125- X 4-mm) column by a gradient elution [solvent A: 50 mM sodium acetate buffer (pH 6.0); solvent B: methanol; 0-3575 B in 10 min, then to 65% B in 25 min; flow rate 1.0 mL/min]. Comparison of both HPLC profiles showed that the reaction of deoxyuridine with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide resulted in trace amounts of N-3-(o-hydroxybenzyl)deoxyuridine eluting at a retention time of 21.2 min. Synthesis of N-1-(eHydroxybenzy1)adenineand N - 3 - ( e Hydroxybenzy1)adenine. Twenty milligrams of adenine was reacted with 45 mg of 2-phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide in 2.5 mL of 40% ethanol in water for 1 2 h a t 50 “C. This solution was separated by HPLC system a using a 5-pm C18 (125- X 4-mm) column by gradient elution [solvent A: 25 mM sodium phosphate buffer (pH 6.0); solvent B: methanol; 0-35% B in 10 min, then to 46% B in 30 min; flow rate 1.2 mL/min. The main product, N-1-(0-hydroxybenzy1)adenineeluted at a retention time of 21 min. The minor product N-3-(o-hydroxybenzyl)adenine eluted at a retention time of 23.8 min. The fractions were collected and lyophilized. (A)N-1-(eHydroxybenzy1)adenine:Electronic spectra: pH 2, A,, = 267 nm; pH 6, A,, = 270 nm; pH 13, A,, = 272 nm; lH-NMR (DMF-&)l 6 (ppm) = 8.80 (s, 1 H, H-2), 8.30 (s, 1 H, H-8), 7.43 (dd, 513.14 = 7.6 Hz, 513,15 = 1.6 Hz, 1 H, H-13), 7.23 (dt, 5 1 5 16 = 515,14 = 7.3 Hz, 515.13 = 1.6 Hz, 1 H, H-15), 6.99 (dd, 515,16 = 7.25 Hz, 516.14 = 1Hz, 1 H, H-16), 6.83 (dt, J 1 4 , 1 3 = 7.5 Hz, 5 1 4 , l s = 7.3 Hz, 514.16 = 1.1Hz, 1 H, H-14), 5.5 (s,2 H, H-lo), 8.01, 2.91, 2.74 (dimethylformamide-d7). (B) N-1-(eHydroxybenzy1)deoxyadenosine: Electronic = 260 nm; pH 13, A,, = 264 nm. spectra: pH 2, A,, (C) N-3-(eHydroxybenzyl)adenine:Electronic spectra: pH 2, A,, = 275 nm; pH 6, A, = 273 nm; pH 13, A, = 273 nm; lH-NMR(DMF-d7)6(ppm) = 8.7 (s,lH,H-2),7.9(s,lH,H-8), 7.45 (dd, J i n . 1 5 = 7.5 Hz, 513,16 = 2 Hz, 1H, H-13), 7.21 (dt, 515,14 = 7.5 Hz, 5 1 5 , i ~= 8.0 Hz, J i 5 , 1 3 = 2 Hz, 1H, H-15), 6.94 (dd, 5 1 6 , 1 6 = 8.0 Hz, 516,13 = 1.1Hz, 1H, H-16),6.84 (dt, 5 1 4 13 = J i 4 , i s = 7.5 Hz, 514,16 = 1.1Hz, 1 H, H-14), 5.3 (s, 2 H, H-10). Modified Deoxynucleoside 3’-Phosphates. These were prepared by the reaction with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide in aqueous buffers. After 10 h, the mixtures were analyzed for the formation of 2-(o-hydroxybenzyl) adducts by HPLC. The structure of the adducts was confirmed by their electronic spectra recorded by the diode-array detector. ( A ) N - 2 - ( 0 - H y d r o x y b e n z y 1 ) - 2 ’- d e o x y g u a nos i n e 3’-monophosphate: For HPLC separation, a 4 X 250-mm steel column filled with Partisil ODS-I11 (3 pm) was used. A linear gradient from 0% to 35 % methanol in 10 mM sodium phosphate buffer (pH 6.0) in 35 min followed by a linear gradient to 65% methanol in 40 min was used for separation. Flow rate was 1 mL/min. N-2-(o-Hydroxybenzyl)-2’-deoxyguanosine3’-monophosphate eluted with a retention time of 22.9 min (A,, = 255 nm, 275 nm). ( B ) N - 3 - ( 0 - H y d r o x y b e n z y 1 ) - 2’-d e o x y u r i d i n e 3’-monophosphate: For HPLC separation, a 4- X 250-mm steel column filled with Supelcosil LC-18 (3 pm) was used. A linear gradient from 0% to 25% methanol in 10 mM sodium phosphate buffer (pH 6.0) in 35 min followed by a linear gradient to 33% methanol in 20 min was used for separation at a flow rate of 1 I Abbreviations: DMF, dimethylformamide; PEI, poly(ethy1ene imine); rev, revertants.

Mentzschel et al. mL/min. N-3-(o-Hydroxybenzyl)-2‘-deoxyuridine3‘-monophosphate eluted with a retention time of 13.5 min (A,, = 266 nm). ( C ) N - 1- ( 0 - Hydroxy be n z y 1 ) - 2’-d eo x y a d e n os i n e 3’-monophosphate: For HPLC separation, a 4- X 250-mm steel column filled with Supelcosil LC-18 (3 pm) was used. A linear gradient from 0% to 35 % methanol in 10 mM sodium phosphate buffer (pH 6.0) over 10 min followed by a linear gradient to 65% methanol for 40 min was used for separation. Flow rate was 1 mL/min. N-l-(o-Hydroxybenzyl)-2’-deoxyadenosine3’-monophosphate eluted with a retention time of 14.4 min (pH 2, A,, = 260.5 nm; pH 13, A, = 264 nm). Reaction of 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin 2-Oxide with DNA. Calf thymus DNA (0.5 mg) was incubated 2-oxide (1mg) with 2-phenoxy-4H-l,3,2-benzodioxaphosphorin for 3 days at 37 “C in 500 pL of 0.1 M sodium phosphate buffer (pH 7.4). After the addition of 100 pL of 5 M sodium chloride, the reacted DNA was precipitated with 2 mL of cold ethanol. The pellets were washed with cold ethanol, dried, and dissolved in 700 pL of 0.01 X SSC (lox: 87.6 g of NaCl and 44.1 g of sodium citrateil, pH 7.0) buffer. 32P-Postlabeling.One microgram of DNA was digested to 2’-deoxynucleotide 3’-monophosphates in 5 pL of buffer (0.2 M sodium succinate, 0.1 M CaC12, pH 6.0) by incubation with spleen exonuclease (1p g ) and micrococcal nuclease (0.12 unit) at 37 “C and 3 h. An equivalent of 8 pg of DNA after digestion was used for the postlabeling piocedure. The DNA adducts were postlabeled with [32PlATPaccording to the intensification method of Guptaet al. (21). The phosphorylation reaction was performed in a total volume of 16 pL by mixing 8 pL of DNA digest and 8 pL of buffer containing 800 mM Bicine-NaOH, pH 9.7,lOO mM MgC12,40 mM spermidine, 400 mM dithiothreitol, T4 polynucleotide kinase (3’-phosphatase free, 10 units/pL), and [-p32P]ATP (3000 pCi/nM). The mixture was incubated at 37 “C for 45 min. Apyrase (10 milliunits) was added and incubated for another 30 min. Adducts and modified nucleosides were separated on poly(ethy1eneimine)-cellulose plates using the following solvent systems: D 1 , l M sodium phosphate (pH 6.8); D3,3.8 M lithium formate, 6.8 M urea (pH 3.4); D4, 0.61 M sodium phosphate,0.61MTrizma,6.53Murea(pH8.2);D5,1.7Msodium phosphate (pH 6.0). After chromatography, the plates were washed with water and air-dried. The radioactive spots on the PEI sheets were located by autoradiography using Kodak X AR-5 film and an intensifying screen.

Results Mutagenicity of 2-Phenoxy-4H-l,3,2-benzodioxaphosphorin 2-Oxide. To investigate the genotoxic potential of o-tolyl phosphate and 2-phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide, the mutagenicity of both compounds was assessed in the Ames test. o-Tolyl phosphate undergoes cytochrome P-450 mediated cyclization to a dioxaphosphorin 2-oxide structurally anal2-oxogous to 2-phenoxy-4H-1,3,2-benzodioxaphosphorin ide. Figure 2a shows that 0-tolyl phosphate was mutagenic in the Ames test only after metabolic activation by liver subcellular fractions (695 rev/pmol). 2-Phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxidewas also a potent mutagen in bacteria (1452rev/pmol;S. typhimurium TA loo),which did not require metabolic activation (Figure 2b). No increase of revertant frequencieswas observed after adding 53-9 mix (981 rev/pmol; S. typhimurium TA 100). 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide is an electrophilic compound; its ability to benzylate nucleophiles has been demonstrated (8). Its reactivity with DNA constituents was investigated by incubation with nucleosides. Reaction of 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin 2-Oxide with Guanosine. The reaction of guanosine with 2-phenoxy-4H-l,3,2-benzodioxaphospho-

Genotoxicity and DNA Binding of Organophosphates

Chem. Res. Toxicol., Vol. 6, No. 3, 1993 297 391

1

a

600

L

m n fn "

400

m

E

280

s! 200

320

360

1400

M/Z

1

8

b

0 0

4000

0

-n

0,Ol

0 , 0 3 0,05

0,l

0,2

0,3

3 pM

b

1

3000

L

n 5

v)

2000

m m

2

I

I

I

8

8

7

I

I

I

6

5

4

6 (PPM) 1000

0

0

0,l 0 , 3

1

2

3

5 ILM

Figure 3. (a) Thermospray mass spectrum of the product of the reaction of guanosine with 2-phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide (recorded in methanol/water). (b) 400-MHz 'HNMR spectrum of N-2-(o-hydroxybenzyl)guanosine in dimethyldg sulfoxide after purification by preparative HPLC (for details of spectra collection, see Experimental Procedures).

Figure 2. Mutagenic activity of tri-o-tolyl phosphate (a) and 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide (b) in S. typhimurium TA 100 without (m) or with (hatched boxes) the addition of rat liver microsomes (0.1 mg of protein/plate). Each value represents the mean of four determinations from two independent experiments.

rin 2-oxide was performed in aqueous buffer at pH 7.2-7.5 and 37 "C for 4 days. The reaction resulted in the formation a new UV-absorbing product as indicated by HPLC analysis. The product was isolated by reversedphase chromatography and showed a thermospray mass spectrum with a major fragment mlz = 391, likely to be the molecular ion (M + 2H)+ (Figure 3a). This finding suggests the addition of guanosine (mlz = 282) to an o-hydroxybenzyl group (mlz = 107). For further elucidation of the structure, 'H-NMR spectra (Figure 3b) of the isolated compound were recorded. The resonance at 6 = 7.9 ppm is assigned to the proton at C-8 in the purine ring of guanosine. Signals at 6 = 6.7,6.9,7.1, and 7.2 ppm, integrated as one proton each, correspond to the four protons of the aromatic ring. The splitting pattern and J Hz, J5',6' = 8.5 Hz, and the coupling constants J ~ I =, ~7.3 J6',7' = 8.0 Hz are similar to those reported for 1,2disubstituted benzenes. The broad signal at 4.4 ppm corresponds to the benzylicprotons bound to the exocyclic nitrogen of guanosine. Therefore, the structure was assigned as N-2-(0-hydroxybenzy1)guanosine. The electronic spectra at pH 2.0, 6.2, and 13 (Figure 4) were identical with to those reported for N-2-Q-methoxybenzyllguanosine (14). Reaction of 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin 2-Oxide with Deoxycytidine. 2-Phenoxy-4H1,3,2-benzodioxaphosphorin2-oxide and deoxycytidine were incubated in aqueous buffer at pH 7.2-7.5 and 37 "C for 4 days. Formation of a new UV-absorbing compound

0

C

4 38

0.150

245

280

315

nm

Wavelength

Figure 4. pH-dependent electronic spectra of N-2-(o-hydroxybenzy1)guanosineat pH 2 (HCI),pH6 (HzO),and pH 10 (NaOII). was indicated by HPLC. Electronic spectra taken at pH 2.0 and 6.2 showed maxima at 220 nm, 265 nm and 225 nm, 266.5 nm (data not shown). These findings suggest the formation of an N-3-alkylated uridine and not of an deoxycytidine derivative. To verify this proposed reaction, NMR and thermospray mass spectra were recorded. The thermospray mass spectra of the isolated product (Figure 5a), recorded in methanollwater, showed an intensive fragment with mlz = 335 and a minor fragment with mlz = 219. The molecular mass of N-3-(o-hydroxybenzyl)deoxycytidine is 333 amu; that of N-3-(o-hydroxybenzyl)uridine is 334 amu. The loss of the deoxy sugar moiety from (M + 2H)+ is represented at mlz = 219. Further support for the proposed transformation of the deoxycytidine moiety to deoxyuridine can be derived from the IH-NMR spectra. Figure 5b shows the lH-NMR of the adduct. Signals at 6 = 7.1 and 6.9 ppm, integrated as one proton singulet each, and a multiplet at 6 = 6.8 ppm integrated as two protons, are assigned as protons on the

298 Chem. Res. Toxicol., Vol. 6, No. 3, 1993 loo

1

335

250

300

Mentzschel et al.

1

a

400

350

ME

N-H

6 I1

I

I

I

I

8

7

6

5

I 4

I 9

6 (PPM)

I 8

I

7

I 6

I

I

5

4

I 3

I

1

2

1

6 (PPM) 0.600

T

160

I

150

I

140

I 130

I

120

I 110

I 100

I

90

I 80

I

I

70

60

6 (PPM)

Figure 5. (a) Thermospray mass spectrum of the product of the reaction of deoxycytidine with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide (recorded in methanoliwater). (b)400-MHz 'H-NMR spectrum of 3-(o-hydroxybenzyl)deoxyuridinein dimethyl-& sulfoxide. (c) 100-MHz '3C of 3-(o-hydroxybenzyl)deoxyuridine in dimethyl-& sulfoxide (off resonance decoupled). Signals not labeled are identified under Experimental Procedures.

aromatic ring. The singulet at 6 = 4.9 ppm corresponds to the benzylic proton. The 13C-NMR spectrum (Figure 6) demonstrated the presence of two carbonyl resonances at 6 = 152 and 161 ppm assigned as C-2 and (3-4, respectively. These are very similar to the signals of the carbonyl groups observed for uridine in Me&3O-ds. The electronic spectra and the 'H-NMR spectra of the isolated compound were very similar with those of N-3-benzyluridine (15). These considerations identify the adduct formed by incubation of deoxycytidine as N-34o-hydroxybenzy1)deoxyuridine. Under conditions of shorter incubation, HPLC analysis indicated the formation of a nucleotide adduct with the characteristic electronic spectrum of N-3-benzyldeoxycytidinederivate. Unfortunately, the adduct could not be isolated due ita instability in the aqueous media required for purification. Reaction of 2-Phenoxy-4H-1,3,2-benzodioxaphosphorin 2-Oxide with Adenine. 2-Phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide and adenine were reacted in 40 5% ethanollwater at 50 "C for 24 h. Formation of two new UV-absorbing compounds were indicated by HPLC. The compounds accounted for 41 % and 21 9% of the total areas of the UV absorbance at 254 nm and were isolated

0.200

220

240

260

280

300

Wavelength (nm)

Figure 6. (a) 400-MHz 'H-NMR spectrum of N-1-(0-hydroxybenzy1)adenine recorded in dimethylformamide-d,. (b) UV absorption spectra of N-1-(0-hydroxybenzy1)adenineat pH 2,6, and 12. by reversed-phase chromatography. The thermospray mass spectra of the isolated compounds were identical (data not shown) and showed minor ions at mlz = 242, likely the molecular ion (M + HI+, and major ions at m/z = 136, the molecular ion (M + H)+ of adenine, which is consistent with the loss of the o-hydroxybenzyl group (mlz = 107). The 'H-NMR spectra of the major product were recorded and confirmed the presence of a 1,Zdisubstituted benzene ring (Figure 6a). The splitting pattern and the coupling constants are similar to those observed for N-2-(o-hydroxybenzyl)deoxyguanosine and N-3- (0-hydroxybenzy1)deoxyuridine. The NMR spectra of the adenine derivative showed a singulet, integrated as a signal representing two protons at 6 = 5.5 ppm. Signals at 6 = 8.6 ppm and 6 = 8.3 ppm were assigned as protons on the purine ring at C-2 and C-8, respectively (Figure 6a). The electronic spectrum of the major product is shown in Figure 6b. The electronic

Chem. Res. Toxicol., Vol. 6, No. 3,1993 299

Genotoxicity and DNA Binding of Organophosphates

spectra of the isolated minor product recorded at pH 2, 6, and 13 showed maxima at 275 and 273 nm (data not shown). On the basis of these data we assigned the structure of the major compound as N-1-(0-hydroxybenzy1)adenine and of the minor product as N-3-(o-hydroxybenzy1)adenine. In incubations with deoxyadenosine, only one product identified as N-1-(0-hydroxylbenzy1)deoxyadenosine was identified by thermospray mass spectrometry and pH-dependent UV spectroscopy (data not shown). The electronic spectra recorded at pH = 6 showed a maximum at X = 260 nm. Synthetic N-6benzyldeoxyadenosineshowed a maximum at X = 269 nm, supporting the structural assignment of benzylation a t the N-1 position of deoxyadenosine. Identification of Modified Nucleotides in DNA Reacted with 2-Phenoxy-4H-l,3,2-benzodioxaphosphorin 2-Oxide. Using the 32P-postlabelingprocedure to detect DNA adducts, at least seven products of the reaction of DNA and 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide were found by incubation of calf thymus DNA with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide. To determine which of these compounds was identical to the characterized adducts, chromatography standards were prepared by reacting 2-phenoxy4H-1,3,2-benzodioxaphosphorin 2-oxide with 2’-deoxyguanosine 3’-monophosphate and 2’-deoxycytidine 3’-monophosphate under aqueous conditions. N-2-(o-Hydroxybenzyl)-2’-deoxyguanosine 3’-monophosphate and N-3(o-hydroxybenzyl)-2’deoxyuridine3’-monophosphate were characterized by their electronic spectra and isolated by HPLC. By the reaction of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide with 2’-deoxyadenosine 3‘monophosphate in 40 % ethanol/water, only N-1-(0hydroxybenzyl)-2’-dexyadenosine3’-monophosphate was formed and isolated by reversed-phase chromatography. 32P-ContainingN-2-(0-hydroxybenzyl)-2’-dexyguanosine 3’,5’-bisphosphate and N-3-(o-hydroxybenzyl)-2’-deoxyuridine 3’,5’-bisphosphatehave the same mobility on PEI plates as adducts 1and 2 (Figure 7, top left and top right). Adducts formed by the treatment of calf thymus DNA with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin2-oxide and the synthetic standards comigrated after 32p-postlabeling and confirmed the structures of adducts 1and 2 (Figure 7, top left and top right) as the deoxyguanosine and deoxyuridine adducts. When the products of the reaction of 2’-deoxyadenosine 3’-monophosphate and 2-phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide were treated with [32P]ATP,radioactive spots were indicated after 32P-postlabelingand TCC separation (Figure 7). One of these spots could be assigned the structure of N-140hydroxybenzyl)-2’-dexyadenosine3’-monophosphate;the other could not be identified. It was, however, not identical to N-3-(o-hydroxybenzyl)-2’-deoxyadenosine3’-monophosphate. Comparison of the migration characteristics of N-l-(o-hydroxybenzyl)-2’-deoxyadenosine3’-monophosphate and of the adducts formed by treating DNA with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin2-oxide indicated that adduct 3 (Figure 7, lower left) represents N-1-(o-hydroxybenzyl)-2’-deoxyadenosine3’,5’-bisphosphate. These results show that DNA adducts are also formed by treatment of intact DNA with 2-phenoxy-4H1,3,2-benzodioxaphosphorin2-oxide in vitro.

Discussion We investigated the mutagenicity and DNA-binding of 2-phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide, a toxic

. , 1 3 *

Figure 7. 32Pmaps of N-2-(o-hydroxybenzyl)deyguanosine 3’,5’-bisphosphate (1,top left), N-3-(o-hydroxybenzyl)deoxyuridine 3’,5’-bisphosphate (2, top right), N-1-(0-hydroxybenzyl)deoxyadenosine 3’,5’-biaphosphate (3, lower left), and DNA treated with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide (lower right). Spot 4 is most likely N-3-(o-hydroxybenzyl)deoxycytidine3’,5’-bisphosphate(4,lower right). 50 pCi of labeled DNA digest and each 100 pCi of DNA standards (top left, top right, and lower left) (A-C) were chromatographed on PEIcellulose. Autoradiography was performed at -70 “C for 20 h, expect for the lower left and lower right samples, which were exposed for 2 and 12 h, respectively.

llle

OH

\

V

I

NH

dR

OH

VI

II

VI1

dR

Figure 8. Proposed pathways of adduct formation from 2-phe noxy4-I- 1,3,2-benzodioxaphosphorin2-oxide with deoxycytidine, deoxyguanosine, and deoxyadenosine.

metabolite of 0-tolyl phosphate Ib (Figure 1). 2-Phenoxy4H-1,3,2-benzodioxaphosphorin2-oxide was a direct mutagen in the Ames test. The observed mutagenicity is

300 Chem. Res. Toxicol., Vol. 6, No. 3, 1993

likely to be the consequence of a modification of DNA constituents by products formed from 2-phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide (Figure 1). In vitro incu2-oxbation of 2-phenoxy-4H-1,3,2-benzodioxaphosphorin ide with guanosine, deoxycytidine, and adenine resulted in the formation of modified DNA nucleosides. In vitro, four different adducts were identified. The obtained NMR, UV, and thermospray mass spectra identify the compound formed by the incubation of guanosine and 2-phenoxy-4H-l,3,2-benzodioxaphosphorin 2-oxide as N-2(0-hydroxybenzy1)guanosine. The electronic spectra are not markedly different from the electronic spectra of other N-1-, N-2-, 0-6-, or N-7-(p-methoxybenzyl)guanosine compounds (14). Comparison of the chemical shift of the benzylic protons with those of N-1-, N-2-, 0-6-, or N-7(p-methoxybenzy1)guanosinesupported the assumption that the alkylation occurred at the exocyclic nitrogen. N-3-(o-Hydroxybenzyl)deoxyuridine was isolated as product from the reaction of deoxycytidine and 2-phenoxy4H-1,3,2-benzodioxaphosphorin2-oxide in aqueous buffer and not the expected N-3-(o-hydroxybenzyl)deoxycytidine. Deoxycytidine adducts and also deoxycytidine are susceptible to hydrolytic deamination. N-3-(o-Hydroxybenzy1)deoxyuridinewas formed when deoxyuridine and 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide was reacted in aqueous buffer under the same conditions, indicating hydrolytic deamination of the intermediary deoxycytidine adduct. This deoxycytidine derivative could be detected by HPLC separations of incubation mixtures. The product formed after N-3-alkylation of deoxycytidine is the less stable imino tautomeric form, which was hydrolitically converted to N-3-(o-hydroxybenzy1)uridine. Alkylationof deoxycytidine by benzyl halides in aqueous solution gives products alkylated at N-3 and N-4 (20). N-3-Alkylation predominated with benzyl bromide, a “hard” electrophile, and N-4-alkylation with p-methoxybenzyl bromide, a “soft” electrophile, indicating that regioselectivity depends on the hardness of the electrophile. The intermediate formed from 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide can be considered as a relatively “soft”electrophile, which should thus give N-4-alkylation with deoxycytidine. The observed results therefore suggest that factors other than the hardness of the electrophile govern the regioselectivity. On the basis of our observations and the fact that regioselectivity is also determined by the nature of the leaving group, the product formed by the reaction of deoxycytidine and the “hard” electrophile 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide may explain the N-3-alkylation. Incubation of adenine with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide gave both N-1-(o-hydroxybenzy1)adenine and N-3-(o-hydroxybenzyl)adenine. Incubation of deoxyadenosine, 2’-deoxyadenosine 3’-mOnOphosphate, and DNA with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxidegave only N-1-alkylated adenine derivatives. We suggest that the P-D-ribose can exert a shielding effect on the N-3 position of the pyrimidine ring, which is directly adjacent to the site of glycosylation. Therefore, the possibility of formation of N-3-benzylated deoxyadenosine and an N-3 adduct in DNA is unlikely due to steric hindrance. Toobtain the nucleotide adducts, we reacted 2-phenoxy4H-1,3,2-benzodioxaphosphorin2-oxide with 2’-deoxyguanosine 3’-monophosphate, 2’-deoxycytidine 3’-monophos-

Mentzschel et al.

phate, and 2’-deoxyadenosine 3’-monophosphate. On the basis of their electronicspectra the products were identified as N-2-(o-hydroxybenzyl)-2’-deoxyguanosine3‘-monophosphate, N-3-(o-hydroxybenzyl)-2’-deoxyuridine3’monophosphate, and N-l-(o-hydroxybenzyl)-2‘-deoxyadenosine 3’-monophosphate. Identical adducts are also formed by treatment of isolated DNA with the electrophile 2-phenoxy-4H-1,3,2benzodioxaphosphorin 2-oxide. Comparison of the chromatographic properties of these 32P-labeled nucleotide standards and [32PlATP-labelednucleosidesisolated from DNA reacted with 2-phenoxy-4H-1,3,2-benzodioxaphosphorin 2-oxide allowed the identification of N-240hydroxybenzyl)-2’-deoxyguanosine3’,5‘-bisphosphate. Adducts 2 and 3 corresponded to N-1-(0-hydroxybenzyl)2’-deoxyuridine 3’,5’-bisphosphate and N-1-(0-hydroxybenzyl)-2’-deoxyadenosine3’,5‘-bisphosphate. Adduct 4 is most likely N-3-(o-hydroxybenzyl)-2‘-deoxycytidine 3’,5’-bisphosphate. Under the conditions of this assay, this adduct seems to be hydrolyzed incompletely and is thus detected as a spot on the TCC plates. In summary, our results show that certain neurotoxic organophosphates may have mutagenic and, probably, carcinogenic properties which should be investigated in further experiments. Acknowledgment. We thank Dr. D. Scheutzow and Mrs. E. Ruckdeschl of the Institut fur Organische Chemie, Universitat Wurzburg, for recording the NMR spectra. This work was supported by the Bundesamt fur Zivilschutz, Ausschuss 6. References Windholz, M., Budavari, S., Bluvetti, R. F., and Otterbein, E. E. (1983) Merck index: a n encyclopedia of chemicals, drugs and biologicals, Merck and Co., Inc., Rahway, NJ. Abou-Donia, M. B. (1981) Organophosphorus ester induced delayed neurotoxicity. Annu. Rev. Pharmacol. Toxicol. 21, 511-548. Henschler, D. (1958) Die Tricresylphosphatvergiftung:Experimentelle Klaerung von Problemen der Aetiologie und Phatogenese. Klin. Wochenschr. 36, 663-674. Johnson, M. K. (1975) Organophosphorus esters causing delayed neurotoxic effects. Mechanism of action and structure/activity studies. Toxicol. Appl. Pharmacol. 7, 227-235. Bleiberg, M. J.,and Johnson, H. (1965) Effect ofcertainmetabolically active drugs and oximes on tri-o-cresylphosphatetoxicity. Toxicol. Appl. Pharmacol. 7, 227-235. Eto, M., Casida, J. E., and Eto, T. (1962) Hydroxylation and cyclization reaction involved in the metabolism of tri-o-cresyl phosphate. Biochem. Pharmacol. 11, 337-352. Somkuti, S. G., and Abou-Donia, M. B. (1990) Disposition, elimination and metabolism of tri-0-cresylphosphatefollowing daily oral administration in Fisher 344 male rats. Arch. Toxicol. 64,572-579. Ohakawa, H., and Eto, M. (1969) Alkylation of mercaptans and inhibition of SH enzymes by saligenin cyclic phosphate and phosphorothiolate esters. Agric. Biol. Chem. 33, 443-451. Eto, M. (1974) Chapter 111. Chemical reactions. In Organophosphoruspesticides: organic and biological chemistry (Zweig,G., Ed.) pp 57-121, CRC Press, Cleveland, OH. Maron, D. M., and Ames, B. N. (1983) Revised methods for Salmonella mutagenicity test. Mutat. Res. 113, 173-215. Vamvakas, S., Dekant, W., Berthold, K., Schmidt, S., Wild, D., and Henschler D. (1987) Enzymatic transformation of mercapturic acids derived from halogenated alkenes to reactive and mutagenic intermediates. Biochem. Pharmacol. 36, 2741-2748. Nomeir, A. A., and Abou-Donia, M. B. (1986) Studies on the metabolism of the neurotoxic tri-o-cresylphosphate. Synthesis and identification by infrared, proton nuclear magnetic resonance spectroscopy and mass spectrometry of five of its metabolites. Toxicology 18, 1-13. Beland, F. A., Allaben, W., and Evans, E. F. (1980) Acyltransferasemediated binding of N-hydroxyarylamides to nucleic acids. Cancer Res. 40, 834-840.

Genotoricity and DNA Binding of Organophosphates (14) Moschel, C.R., Hudgins, W. R., and Dipple, A. (1981)Dissociation of O"(p-methoxybenzyl)guanosine in aqueous solutions. J. Am. Chem. SOC.103,5489-5494. (15) Philips, D. K.,and Horowitz, J. P. (1975)Benzylation-debenzylation studies on nucleosides. J. Org. Chem. 40,1856-1858. (16)Djuric, Z.,and Sinsheimer, J. E. (1984)Reactivity of propyleneoxides towards deoxydeoxycytidine and identification of the reaction products. Chem.-Biol. Interact. 50,219-231. (17) Singer,B., and Grunberger,D. (1983)Acid, neutra1,and basicspectra of bases, nucleosides, nucleotides, and 55 modified derivatives. In Molecular biology of mutagens andcarcinogens,pp297-334,Plenum Press, New York and London.

Chem. Res. Toxicol., Vol. 6, No. 3, 1993 301 (18)Eto, M., and Oshima, Y. (1962)Syntheses and degradation of cyclic phosphorus esters derived from saligenin and its analogues. Agric. Bioi. Chem. 26,452-459. (19) Moschel, R. C.,Hudgins, W. R., and Dipple, A. (1979)Selectivity in nucleoside alkylation in relation to chemical carcinogenesis. J. Org. Chem. 44,3324-3328. (20) Shapiro, R., and Shiuey, S. (1976)Reactions of deoxycytidine with 7-bromomethylbenz(a)anthracen,benzyl bromide and p-methoxybenzyl bromide. J. Org. Chem. 41,1597-1600. (21) Gupta, R. C., Reddy, M. V., and Randerath, K. (1982) 32Ppostlabeling analysis of non radioactive aromatic carcinogen-DNAadducts. Cancer Res. 40,887-896.