The Carcinogenic Significance of Reactive Intermediates Derived from

Ester Hydrolysis and Nitrosative Deamination of Novocaine in Aqueous Solutions. Emilia Iglesias-Martinez, Isabel Brandariz, and Francisco Penedo. Chem...
0 downloads 0 Views 236KB Size
Download PDF
Chem. Res. Toxicol. 2005, 18, 1955-1966

1955

The Carcinogenic Significance of Reactive Intermediates Derived from 3-Acetoxy- and 5-Acetoxy-2-hydroxy-N-nitrosomorpholine Richard N. Loeppky,* Sunil Sukhtankar, Feng Gu, and Misun Park Department of Chemistry, University of Missouri, Columbia, Missouri 65211 Received July 23, 2005

N-Nitroso-2-hydroxymorpholine (NHMOR), a relatively reactive metabolite of two potent carcinogens, N-nitrosodiethanolamine (NDELA) and N-nitrosomorpholine (NMOR), has been reported to not be carcinogenic. Two isomeric acetate esters of the R-hydroxynitrosamines expected to be produced from the cytochrome P450-mediated metabolism of NHMOR have been synthesized, and their hydrolytic decomposition products, hydrolysis rates, and deoxyguanosine (dG) reaction adducts have been determined. N-Nitroso-3-acetoxy-2-hydroxymorpholine was prepared in high yield from the reaction of N-nitroso-2,3-dehydromorpholine with dry peracetic acid in glacial acetic acid or by the reaction of its dimethyldioxirane-produced epoxide with glacial acetic acid. The hydrolysis of this R-acetoxynitrosamine gave acetaldehyde (10%), ethylene glycol (55%), glyoxal (95%), and acetic acid. The pH rate profile for the hydrolysis of this nitrosamine was abnormal in that it exhibited pronounced base-catalyzed hydrolysis beginning at pH 5. The mechanism of hydrolytic decomposition is proposed to involve neighboring group participation with the formation of a reactive epoxide intermediate. N-Nitroso-3-acetoxy-2-hydroxymorpholine reacted with dG to give these guanine adducts after acidic deglycosylation: 1,N2-glyoxal (65%), 7-(2-hydroxyethyl)guanine (9%), and O6-hydroxyethylguanine (3%). N-Nitroso-5-acetoxy-2-hydroxymorpholine was synthesized from 2-hydroxyethylvinylnitrosamine by its oxidative conversion to the corresponding aldehyde followed by reaction with dry peracetic acid in glacial acetic. The hydrolytic decomposition products of this nitrosamine were 2-acetoxyacetaldehyde (65%), a rearrangement product, glycol aldehyde (15%), a trace of glyoxal, and acetic acid. The pH rate profile for the hydrolysis of this acetate is similar to other R-acetoxynitrosamines in that it exhibits a pH-independent region which gives way to base-catalyzed ester hydrolysis beginning at pH 7. The lower pH (≈ 7 < 9) onset of base catalysis is proposed to involve base-catalyzed opening of the hemiacetal and intramolecular acyl transfer to give an unstable R-hydroxynitrosamine. N-Nitroso-5-acetoxy2-hydroxymorpholine was less reactive toward dG and gave the 1,N2-etheno-dG adduct (44%). The products from both of the isomeric R-acetoxy nitrosamines were judged to arise from diazonium ions produced from unstable R-hydroxynitrosamine intermediates. The high yield of the rearrangement product 2-acetoxyacetaldehyde could explain the low carcinogenic potential of NHMOR if it is mainly R-hydroxylated at the 5 carbon. Hydroxylation of NHMOR at carbon 3 is expected to yield a carcinogenic outcome.

Introduction 1)1

N-Nitroso-2-hydroxymorpholine (NHMOR, is an intermediary in vitro metabolite of N-nitrosomorpholine (NMOR, 3) (1-3) and N-nitrosodiethanolamine (NDELA, 4) (3-6), both of which are potent animal carcinogens. Yet, Lijinsky and Hecht failed to observe carcinogenicity for 1 when it was administered to either the A/J mouse or the Sprague Dawley rat (7). As shown in Scheme 1, N-nitroso-2-hydroxyethylglycine (NHEG, 5) is a prominent noncarcinogenic, in vivo metabolite of both NDELA and NMOR (1, 4). Since the biological oxidation of NHMOR to 5 is very reasonable, it has been speculated * To whom inquiries should be addressed. E-mail: LoeppkyR@ missouri.edu. Phone: 573 882-4885. 1 Abbreviations: dG, deoxyguanosine; gdG, 1,N2-glyoxal-deoxyguanosine adduct; gG, 1,N2-glyoxal-guanine adduct; NDELA, Nnitrosodiethanolamine; NHEG, N-nitroso-N-2-hydroxyethylglycine; NHMOR, N-nitroso-2-hydroxymorpholine; NMOR, N-nitrosomorpholine.

that the conversion of NDELA and NMOR to NHMOR is part of a detoxification pathway for these nitrosamino carcinogens. In contrast, NHMOR is the hemiacetal of the R-nitrosamino aldehyde 2, which is reactive and both deaminates bases in DNA by nitroso transfer and produces glyoxal deoxyguanosine, gdG, adducts in DNA in vitro without any enzymatic transformations (3, 6, 8-19). We have also demonstrated that microsomal oxidation of NHMOR produces glyoxal (6) and that the administration of NDELA, NMOR, or NHMOR to male Wistar rats results in the formation of gdG DNA adducts in the livers of these animals (18, 19). The conversion of NDELA to NHMOR is catalyzed by both alcohol dehydrogenase and cytochrome P450 (probably 2E1) (6, 20). Glyoxal can be metabolically produced from these compounds in several ways. There is good evidence that the carcinogenic activation of nitrosamines involves R-hydroxylation by means of the cytochrome P450-

10.1021/tx0502037 CCC: $30.25 © 2005 American Chemical Society Published on Web 11/03/2005

1956

Chem. Res. Toxicol., Vol. 18, No. 12, 2005 Scheme 1

Scheme 2

mediated introduction of a hydroxyl group at the carbon atom adjacent to the nitrosamino nitrogen atom. Extensive chemical studies have shown that almost all R-hydroxynitrosamines are unstable in aqueous media and decompose to aldehydes and reactive diazonium ions (21-27). This process is depicted for both NDELA and NMOR in Scheme 2. We have shown that the hydrolysis of R-acetoxy-N-nitrosodiethanolamine, which generates 6, gives glycolaldehyde 7 and products derived from the decomposition of the 2-hydroxyethyldiazonium ion 8 (28). Similarly, Fishbein has recently prepared and examined the properties of 3-hydroxy-N-nitrosomorpholine 9 and demonstrated that its decomposition gives 11 and other products through the diazonium ion 10 (26). The acid 12, an oxidation product of 11, is an in vivo metabolite of NMOR (1). While neither of these processes gives glyoxal, we have shown that it is produced from the in vitro microsomal oxidation of glycolaldehyde 7 and ethylene glycol which are generated from the hydrolysis of 8 (6). The autoxidation of 7 also readily generates glyoxal.

Loeppky et al. Scheme 3

While the formation of gdG adducts from NDELA does not absolutely require the intermediacy of NHMOR, its R-hydroxylation at either the 3 or 5 position would generate glyoxal either directly through the decomposition of 13 or through the oxidation of glycoladehyde 7 produced from the decomposition of 14, as shown in Scheme 3. The pathway involving the 3-R-hydroxylation of NHMOR has the advantage of producing both glyoxal and 8-13. We have observed the in vivo production of both gdG and O6-hydroxyethyldeoxyguanosine (OHEdG) DNA adducts from rats fed NMOR (18). The 2-hydroxyethyldiazonium ion 8 is the logical precursor of OHEdG adducts. These experiments support the involvement of NHMOR in the carcinogenic activation of both NDELA and NMOR. The failure to observe carcinogenic action from the direct administration of NHMOR could be due to several factors. Low doses were administered to the animals (7). The A/J mouse does not give a significant response with NDELA (7), suggesting that it may not have significant levels of the metabolically requisite enzymes. It is possible that the oral route of administration of NHMOR does not give either the same pharmacologic disposition or metabolism as results from the consumption of NDELA or NMOR. It is also possible that the decomposition of the diazonium ions produced from the R-hydroxynitrosamines results in chemistry which has a significant detoxifying component. An example is provided by the 2-hydroxyethyldiazonium ion. About 35% of the aqueous decomposition of 8 involves a hydride migration from the HO-bearing carbon to the incipient carbocation as the nitrogen molecule departs (28). This results in a much more stable carbocation, an oxonium ion, the oxygenprotonated conjugate acid of acetaldehyde. This type of chemistry is a common feature of diazonium ions with β-oxygen functionality and could play a role in properties of the diazonium ions derived from the R-hydroxylation of NHMOR. To test this hypothesis, we have prepared and examined the chemistry of the R-acetates of 13 and 14 which we report here.

Experimental Procedures Caution: Most nitrosamines are potent carcinogens. Considerable care should be taken in their use so as to avert exposure to humans and to avoid environmental contamination. We routinely perform all operations with these substances, except for dilute solutions thereof, in well-ventilated fume hoods. We rinse all nitrosamine-contaminated glassware with a solution of concentrated HBr in glacial acetic acid which is effective in cleaving the NO group form the amine nitrogen atom. The action of this agent in aprotic solutions is also effective in nitrosamine

Reactive Intermediates from 2-Hydroxy-N-nitrosomorpholine destruction. Aqueous solutions are treated with either Ni(R) or Al in concentrated sodium hydroxide. This process may produce hydrazines. Equipment, Methods, and Materials. Flash column chromatography was performed on 230-400 mesh silica gel (MerckKieselgel 60) with ACS grade solvents without further purification. Analytical TLC was performed on ALUGRAM separations silica gel plates (Silica gel 60, 20 cm × 20 cm, 0.25 mm thickness). Preparative TLC was performed on Analtech Uniplates (F 254, 20 cm × 20 cm, 1.0 mm thickness). All the compounds on TLC were visualized under UV lamp or by iodinesilica gel. 1H and 13C NMR spectra were recorded on either a Bruker AMX 500 MHz or Bruker ARX 250 MHz spectrometer. Gas chromatography was performed on a HP 5890 gas chromatograph fitted with a 30 m × 0.25 Supelco SPB-20 capillary column with a Flame Ionization detector. UV-vis and kinetics experiment were performed on a Hewlett-Packard 8453 UVVis spectrophotometer with UV-Visible Chemstation software, magnetic stirring equipment, and an HP 89090A constanttemperature bath. High-pressure liquid chromatography (HPLC) was performed with a Waters chromatography system containing a Waters Maxima 820 system controller, Waters model 490 programmable multiwavelength detector, Waters model 712 WISP auto sampler, and two waters model 510 pumps. A programmable FC 203 Gilson fraction collector was utilized for collecting eluted samples from HPLC system. Unless otherwise stated, HPLC column used was Phenomenex C18 column (250 mm × 4.6 mm). Liquid chromatography mass spectrometry (LC/ MS) was performed on a Finnigan TSQ 7000 triple quadrapole mass spectrometer. Reagents and solvents were used as obtained from Aldrich, Fischer, Sigma, or other chemical distributors. HPLC solvents were degassed and filtered prior to use. Water was glassdistilled. THF and diethyl ether were purified and dried by distilling over sodium and followed by distillation from LAH. Methylene chloride was refluxed with calcium hydride and distilled. All other chemicals were purified by conventional methods such as distillation and recrystallization whenever necessary. 4-Nitroso-1,7-dioxo-4-azabicyclo[4.1.0]heptane (N-Nitroso-2,3-epoxymorpholine) 20. A flame-dried 250 mL flask, equipped with a rubber septum and gas inlet and outlet needles, was charged with 50 mg of N-nitroso-2,3-dehydromorpholine (29) 19 in 5 mL of CH2Cl2. Thoroughly dry dimethyldioxirane in acetone (0.03 M, total 70 mL) was added slowly via a sharpened transfer needle to the stirring solution through a syringe at room temperature with protection from light and moisture. A stream of dry nitrogen gas was passed through the flask to remove the solvents. NMR data showed that the resulting oily residue was the epoxide 20 (90+ % yield), but it proved to be too unstable for elemental composition analysis. 1H NMR (CDCl ): δ 5.69 (d, 1H, J ) 2.3 Hz), 5.35 (d, 1H, J ) 3 2.3 Hz), 4.02 (m, 2H), 3.87 (m, 1H), 2.94 (m, 1H). 13C NMR (CDCl3): δ 76.94, 62.01, 58.32, 31.32. GC-MS (50 °C for 4 min, then to 250 °C at 15 °C/min, injector temperature 100 °C): 9.1 min. 130 (M+, 1), 101 (2), 86 (1), 73 (100), 56 (4). N-Nitroso-2,3epoxymorpholine 20 is stable in dry CH2Cl2 at 0 °C for at least 2 days; it decomposes slowly in methanol. It decomposes very rapidly in water. It was used without purification in the next synthetic step. N-Nitroso-3-acetoxy-2-hydroxymorpholine 17. 1. Method 1. N-Nitroso-2,3-epoxymorpholine 20 (30 mg, 0.23 mmol), prepared as described above, was dissolved in 15 mL of dry CH2Cl2 in a flame-dried 50 mL round-bottom flask. To the solution was added 13.8 mg of glacial acetic acid (1.0 equiv), and the mixture was stirred at room temperature for 30 min. After the solvent was removed in vacuo, the residue was purified by flash chromatography, eluting with methylene chloride/ethyl acetate (1:1 v/v). UV (pH 3.0):  (235) ) 17 320,  (350) ) 320. IR (neat) 3414, 3000, 2947, 2900, 2367, 1762, 1486, 1446, 1374, 1307, 1229, 1143, 1045, 975, 947, 908, 702 cm-1. 1H NMR (CDCl3): δ E isomer (major) 6.97 (s, 1H), 5.38 (s, 1H), 4.64 (dd, 1H, J )

Chem. Res. Toxicol., Vol. 18, No. 12, 2005 1957 11.8, 3.6 Hz), 4.06 (ddd, 1H, J ) 4, 11.8, 12 Hz), 3.74 (dd, 1H, J ) 12, 3.9 Hz), 3.27 (br s, 1H), 3.02 (ddd, 1H, J ) 12, 3.9, 12 Hz), 2.15 (s, 3H). 13C NMR (CDCl3): δ E isomer, 169.26, 90.23, 80.57, 56.71, 36.09, 20.77. HRCIMS found 191.066 9, C6H11N2O5 (M + H) requires 191.066 8. This compound is stable in CH2Cl2 while it decomposes at aqueous solution. 2. Method 2, Preferred. N-Nitroso-2,3-dehydromorpholine 19 was converted to 17 in a single step using dry peracetic acid. Commercially available peracetic acid (10 mL) was kept overnight over anhydrous sodium sulfate (3 g) and then treated with acetic anhydride (2.5 mL) to remove water. The substrate (50 mg) was dissolved in 2 mL of dry CH2Cl2 in a flame-dried 10 mL flask. Dry peracetic acid (2 mL) was added through a syringe, and the mixture was stirred at room temperature for 2 h. After the solvent was removed in vacuo, the residue was purified by flash chromatography with methylene chloride/ethyl acetate (1:1 v/v). Yield: 52%. 2-Hydroxyethylvinylnitrosamine 21. A dried 100 mL flask equipped with a septum and a nitrogen balloon was charged with KOtBu (2.50 g, 22.32 mmol, 1.04 equiv) and 18-crown-6ether (5.90 g, 22.32 mmol, 1.04 equiv). The mixture was dissolved by adding diethyl ether (50 mL) through a syringe. N-Nitrosomorpholine (2.50 g, 21.55 mmol, 1. 0 equiv) dissolved in diethyl ether (10 mL) was added dropwise while stirring at room temperature and continued for 40 min. The reaction was monitored by TLC using ethyl acetate/methylene chloride (1:3) as solvent. The reaction was quenched with H2O (1.2 mL, 3 equiv) and stirred for 10 min. Magnesium sulfate (10 g) was added and stirred for 30 min. The solid was removed by filtration and washed with methylene chloride. The combined filtrate was concentrated on a rotary evaporator. The residue was chromatographed using ethyl acetate/methylene chloride (1:3) as eluent to give the product (2.12 g, 85% yield). 1H NMR (500 MHz, CDCl3): δ E isomer (83.8%) 7.73 (dd, 1H, J ) 9.37, 16.18 Hz), 5.02 (dd, 1H, J ) 1.99, 16.08 Hz), 4.77 (dd, 1H, J ) 2.03, 9.19 Hz), 3.91 (t, 2H, J ) 5.89 Hz), 3.61 (t, 2H, J ) 5.97 Hz), 2.70 (s, 1H, br); δ Z isomer (16.2%) 7.49 (dd, 1H, J ) 9.58, 16.27 Hz), 4.97 (dd, 1H, J ) 1.09, 16.59 Hz), 4.77 (dd, 1H, J ) 2.03, 9.19 Hz), 4.40 (t, 2H, J ) 5.43 Hz), 3.97 (t, 2H, J ) 5.55 Hz), 2.70 (s, 1H, br). 13C NMR (125.8 MHz, CDCI3): δ E isomer 135.98, 97.82, 58.29, 42.71; δ Z isomer 124.44, 99.71, 59.21, 51.44. IR (neat) 3392, 1841, 1472, 1310 cm-1. Anal. found (calcd) C 41.13 (41.38), H 7.04 (6.94), N 24.39 (24.12). 2-Oxoethylvinylnitrosamine 22. A dry 50 mL flask equipped with a magnetic stir bar, a septum and a nitrogen balloon was charged with 2-hydroxyethylvinylnitrosamine 21 (0.259 g, 2.24 mmol, 1 equiv) and dry CH2Cl2 (10 mL) to give a 0.22 M solution. Dess-Martin periodinane (1.05 g, 2.47 mmol, 1.1 equiv) dissolved in dry CH2Cl2 (10 mL) was added dropwise via a syringe. The reaction mixture was allowed to stir at room temperature for 1 h, while it was monitored by TLC. A white precipitate formed was removed by filtration. The solvent was evaporated, and the desired product was purified by flash chromatography (ethyl acetate/CH2Cl2 (1:5 v/v) to give 2-oxoethyllvinylnitrosamine 22, 72%. 1H NMR (CDCl3): E-isomer (major isomer 94%) δ 9.366 (s, 1H), 7.95 (dd, 1H, J ) 9.305, 16.05 Hz), 4.83 (dd, 1H, J ) 2.54, 9.305 Hz), 4.76 (dd, 1H, J ) 2.56, 16.04 Hz), 4.56 (s, 2H); 13C NMR (CDCl3): δ 191.78, 135.33, 97.32, 49.34. The compound polymerizes readily and was used without further purification in the next step. N-Nitroso-2-hydroxy-5-acetoxymorpholine 18. 2-Oxoethyllvinylnitrosamine 22 (50 mg) was dissolved in 2 mL of dry CH2Cl2 in a flame-dried 10 mL flask. Dry peracetic acid (2 mL) was added through a syringe, and the mixture was stirred at room temperature for 1 h. The solvent was removed in vacuo, and the residue was purified by flash chromatography with methylene chloride/ethyl acetate (1:1 v/v) to give 18, 62%. HRCIMS found 191.0663, C6H11N2O5 (M + H) requires 191.0668. 1H NMR (CDCl ): E-isomer (major) 7.18 (d, 1H, J ) 1.89 Hz), 3 5.42 (d, 1H, J ) 3.085 Hz), 4.68 (d, 1H, J ) 14.03 Hz), 4.56 (dd, 1H, J ) 2.18, 12.82 Hz), 4.05 (d, 1H, J ) 12.785 Hz), 3.03 (dd, 1H, J ) 3.16, 14 Hz), 2.14 (s, 3H); 13C NMR: δ 169.1, 89.1, 78.9,

1958

Chem. Res. Toxicol., Vol. 18, No. 12, 2005

Loeppky et al.

Table 1. Kinetic Data for the Hydrolysis of r-Acetoxy NHMOR Isomers 17 and 18 18 substrate buffer (0.1 M) pH

Phosphate kobs × 103

2 3 4 5 6 7 8 9 10

0.130 ( 0.004 0.120 ( 0.002 0.110 ( 0.001 0.090 ( 0.002 0.120 ( 0.001 0.370 ( 0.004 1.88 ( 0.03 14.78 ( 0.12 74.89 ( 0.46

17

Acetate kobs × 103

Phosphate (set 1) kobs × 103 37 °C

Phosphate (set 1) induction time (s)

Phosphate (set 2) kobs × 103 25 °C

0.56 ( 0.02 3.51 ( 0.39 10.10 ( 0.29 25.71 ( 0.29 32.46 ( 0.45

0.27 ( 0.00 1.30 ( 0.05 5.84 ( 0.26 20.37 ( 0.35 54.60 ( 0.64

2221 351 66 4 0

0. 45 ( 0.02 2.2 ( 0.1 7.2 ( 0.1 23.5 ( 0.5 49.29 ( 0.09 49.9 ( 0.5 75.2 ( 0.1

60.5, 40.5, 20.9; Z-isomer (minor) 7.13 (m, 1H), 4.86 (dd, 1H, J ) 3.26, 13.57 Hz), 4.74 (dd, 1H, J ) 3.28, 9.06 Hz), 4.30 (dd, 1H, J ) 1.27, 13.05 Hz), 4.02 (dd, 1H, J ) 2.34, 13.05 Hz), 2.73 (dd, 1H, J ) 9.05, 13.05 Hz), 2.14 (s, 3H); 13C NMR: δ 169.1, 92.08, 78.1, 65.7, 41.7, 20.9. Kinetics Experiments. The kinetics of hydrolysis of both N-nitroso-2-hydroxy-3-acetoxymorpholine 17 and N-nitroso-2hydroxy-5-acetoxymorpholine 18 were determined by following the decrease in UV absorbance at 234 nm using an HP spectrophotometer. A solution of the substrate (100 µL, 0.02 M) in THF was transferred to a cuvette, and the preheated buffer solution (37 °C, 1.4 mL) was then added. The cuvette inverted twice, and the change in absorbance was recorded as a function of time. In the case of 17, both 0.1 M phosphate and acetate buffers were used. For 18, kinetic studies were done using 0.1 M citric acid-NaOH buffer solutions having pH ) 2-6, 0.1 M phosphate buffer having pH ) 7-8, and 0.1 M glycine-NaOH buffer having pH ) 9-10, respectively. For all reactions, the ionic strength was maintained with 1 M NaClO4. All transformations were allowed to proceed for at least three half-lives. Rate constants, the average of two to three determinations, were determined by nonlinear least-squares fitting as described in Results and Discussion. Rate constants are given in Table 1. Initial experiments to determine the effect of pH on the hydrolysis rate of 17 at 25 °C in phosphate buffer, reported in Table 1 under the column headed Phosphate (set 2), differed somewhat. Solutions of known concentration of 17 in CH2Cl2 were added to the cuvette and blown dry with dry N2. A 25 °C solution of 0.1 M phosphate buffer and 1 M NaClO4 was added to bring the volume to 4 mL and stirred, and measurements were started. The data as instrument-generated plots of absorbance versus time were digitized using the program Un-ScanIt and the rate constants obtained by nonlinear regression as described in the text. These rate constants are the average of two runs at each pH. Hydrolysis of N-Nitroso-2-hydroxy-3-acetoxymorpholine 17. The hydrolysis of the substrate was carried out at pH 7 in 0.1 M phosphate buffer. The products of the decomposition of 17 were analyzed and determined by directly injecting an aliquot of the aqueous mixture from the kinetic study into a GC with an FID detector. A SUPELCO SPB-20 column was used for this analysis. The program used was as follows: injector temperature, 220 °C; initial temperature, 90 °C for 2 min, then to 220 °C at 15 °C/min. Under these conditions, the retention times (min) were acetaldehyde, 3.91; acetic acid, 4.06; and ethylene glycol, 4.74. The aldehydes (acetaldehyde and glyoxal) from the mixture were also derivatized as follows: 500 µL of solution was taken from the mixture and was mixed with 500 µL of 2,4-dinitrophenylhydrazine (2 mM in 18% phosphoric acid). After 5 min reaction time, the hydrazones and the remaining hydrazine were extracted into 500 µL of methylene chloride. The organic phase (200 µL) was transferred into a new vial, evaporated with nitrogen gas, and redissolved in 200 µL of acetonitrile. The sample was separated on a Zorbax C18 column (4.6 mm × 25 cm) with the program listed below. Hydrazone standards were prepared by refluxing the respective

aldehydes (1 mmol) with 2,4-dinitrophenylhydrazine (1 mmol) in ethanol, containing 10% phosphoric acid. The precipitated hydrazones were filtered, washed and dried, and dissolved in acetonitrile. An HPLC program utilizing acetonitrile (A) and water (B) as eluent (390 nm UV detection, 1 mL/min) was as follows: 0-34 min, 10% A; 34-50 min, 40% A; 50-65 min, 80% A. Hydrolysis of N-Nitroso-2-hydroxy-5-acetoxymorpholine 18. The substrate (12 mg, 0.0631 mmol) was added to 500 µL of 0.1 M phosphate buffer (pH 7) and shaken at 37 °C until 95% of substrate decomposed. In the same manner as described for 17, products from the hydrolysis of 18 were determined by using GC detection in comparison with authentic standards and HPLC of the 2,4-dinitrophenylhydrazones of the aldehydes. The decomposition products (1 µL of 0.0192 M aqueous mixture, pH 7) were injected onto the same column as above, except that the temperature was increased from 70 °C after 2 min at a rate of 10 °C/min to 220 °C. Under these conditions, three major peaks were detected (min): glycolaldehyde, 3.58 min; acetic acid, 3.80; and acetoxyacetaldehyde, 4.39. The 2,4-dinitrophenylhydrazone of the aldehydes was separated on a Phenomenex C18 column, 5 µm, 250 mm × 4.6 mm, with the same program as described above. Retention times (min) were as follows: glycolaldehyde, 37.6; glyoxal, 43.0; and acetoxyacetaldehyde, 43.8. Reaction of 2′-Deoxyguanosine with N-Nitroso-2-hydroxy-3-acetoxymorpholine 17. A solution of 2′-deoxyguanosine (1.1 mg, 0.0082 M) in 0.5 mL of phosphate buffer (pH ) 7) was added to N-nitroso-2-hydroxy-3-acetoxymorpholine 17 (20 mg, 0.210 M). The mixture was stirred at 37 °C for 15 min, then 1 N HCl was added, and the mixture was further stirred at 70 °C for 1 h. The resulting hydrolysate was neutralized with dilute NaOH. Aliquots were injected into the HPLC, and the products were eluted with 0.1% formic acid (A) and acetonitrile (B) at 1 mL/min using a Microsorb-MV C18 column (250 mm × 4.6 mm) using the following program: 0-5 min, 99% A; 5-17 min, 89% A; 17-22 min, 40% A. Under these conditions, the following products were found and identified both by comparison of retention times and by LC/MS: glyoxal-guanine adduct (30), 9.9 min; guanine, 12 min; 7-(2-hydroxyethyl)guanine (31), 14 min; O6-(2-hydroxyethyl)guanine (32), 18.8 min. Relative percentages are given in Results and Discussion. Reaction of 2′-Deoxyguanosine with N-Nitroso-2-hydroxy-5-acetoxymorpholine 18. A solution of 2′-deoxyguanosine (2.6 mg, 0.0097 M) in 1.0 mL of phosphate buffer (pH 7) was added to N-nitroso-2-hydroxy-5-acetoxymorpholine (20 mg, 0.105 M). The mixture was stirred at 37 °C for 2 h, then 1 N HCl was added, and the mixture was further stirred at 70 °C for 1 h. The resulting hydrolysate was neutralized with dilute NaOH. Aliquots were injected into the HPLC, and the products were eluted with acetonitrile (A) and water (B) at 1 mL/min using a Phenomenex C18, 5 µm, 250 mm × 4.6 mm column using the following program: 0-15 min, 1% A; 15-17 min, 11% A; 17-22 min, 60% A. Under these conditions, the following products were found and identified both by comparison of retention times and by LC/MS: glyoxal-guanine adduct, 6.0 min; guanine, 8.1 min; 1-N2-ethenoguanine (33), 15.1 min; and an

Reactive Intermediates from 2-Hydroxy-N-nitrosomorpholine

Chem. Res. Toxicol., Vol. 18, No. 12, 2005 1959

Scheme 4

Figure 1. Chemical shift assignments are given in the lefthand structure (13C in italics). Proton coupling constants as established by COSY are shown in the right-hand structure.

Scheme 5

unknown adduct at 19.5 min. Relative percentages are given in Results and Discussion.

Results and Discussion There are two well-known approaches to the generation of R-hydroxynitrosamines: their direct synthesis and their generation through the hydrolysis of the corresponding R-acetates. We chose the latter approach because of the functional complexity of 13 and 14 as well as the general high reactivity and low stability of R-hydroxynitrosamines. More than 100 R-acetoxynitrosamines have been prepared and utilized in chemical, biochemical, mutagenicity, and carcinogenicity studies to probe and mimic the initiation steps in nitrosamine carcinogenesis. The relevant chemistry and citations have recently been reviewed by Fishbein whose group has provided most of the recent work on both R-hydroxynitrosamines and their acetates (27, 34-37). The methods of their synthesis vary somewhat. Our goal has been to synthesize and examine the chemistry of 3-acetoxy-2hydroxy-N-nitrosomorpholine 17 and 5-acetoxy-2-hydroxyN-nitrosomorpholine 18. The vicinal oxygen functionality of these compounds permitted the utilization of a different synthetic approach. The Synthesis and Characterization of 3-Acetoxy2-hydroxy-N-nitrosomorpholine. Two synthetic methods for the preparation of this nitrosamine (17) were utilized as are shown in Scheme 4. We have previously shown that NHMOR can be converted to N-nitroso-2,3dehydromorpholine 19 (29). Careful epoxidation with dimethyldioxirane gives 20, a highly reactive compound. While exploring synthetic approaches to 20, we discovered that the reaction of 19 with 3-chloroperbenzoic acid gave the 3-chlorobenzoate corresponding to 17 (28). Reaction of 20 with anhydrous acetic acid gave 17, as we have reported in a preliminary account (28). While we have used this two-step pathway to generate a number of multifunctional epoxy nitrosamines and R-acetates, we sought to exploit the latter transformation by directly reacting 19 with peracetic acid. Our first attempts to directly convert 19 to 17 using peracetic acid met with limited success. However, we found that the reaction of 19 with carefully dried peracetic acid gives 17 in 62% yield. Peracetic acid is available as a 32% solution in acetic acid. The dry reagent was prepared by first drying this solution over anhydrous sodium sulfate and then by adding acetic anhydride. This method represents a convenient advance over our initial one, particularly in cases where we do not need the epoxide. 1 H and 13C NMR spectroscopy, employing HMQC, HMBC, and COSY have led to a complete characteriza-

tion and structural assignment for 17 (see Supporting Information for relevant spectra). We were aided by an X-ray crystal structure for the corresponding 3-chlorobenzoate (28). The NMR data show that 17 exists primarily as a single isomer, the E-form of the nitrosamine. Stereochemical assignment is based on a comparison with the data from the 3-chlorobenzoate. The 1H NMR spectrum shows some very small peaks which could be attributed to the corresponding Z-isomer, but their intensity is too low for definitive assignment. Like NHMOR, there is no evidence for the presence of the open chain aldehyde structural isomer. The chemical shift assignments and coupling constants are given in Figure 1. There is no coupling between either of the sets of vicinal equatorial protons. The Synthesis and Characterization of 5-Acetoxy2-hydroxy-N-nitrosomorpholine 18. An epoxidation procedure was also used for the synthesis of this nitrosamine (Scheme 5). 2-Hydroxyethylvinylnitrosamine 21 was oxidized by the Dess-Martin procedure to give the corresponding aldehyde 22. Reaction of this substrate with the dry peracetic acid mixture gave 23, which spontaneously cyclized to 18. 1H, 13C, and multidimensional NMR methods were also used to characterize 18. This nitrosamine exists in solution as a 38:62 mixture of two isomers. When isomers are observed for nitrosamines, a common finding, it is usually due to Z-E isomerism about the N-NO bond. This phenomenon is accompanied by relatively large chemical shift differences at the carbons and attached protons adjacent to the nitrogen. That is not the case here. The data (see Figure 2) are in best agreement with an E diastereomeric assignment for the nitrosamine and isomerism due to the configuration at the hemiacetal carbon 2. In the major isomer, there is no coupling between either protons a and e or c and d, where both proton pairs are diequatorial. These data are consistent with the conformer structure shown (Figure 2). Were the

1960

Chem. Res. Toxicol., Vol. 18, No. 12, 2005

Loeppky et al. Scheme 7

Scheme 8

Figure 2. NMR 1H and 13C chemical shifts and coupling constants for 18. Geminal coupling constants are not shown, but are as follows: for the major isomer, b-c J ) 12.8 Hz and e-f J ) 14 Hz; and for the minor isomer, b-c J ) 13 Hz and e-f J ) 13.5 Hz.

Scheme 6

vicinal protons anti to each other, then large coupling would be observed. Similarly, in the minor isomer, there is only small coupling between protons d and e, while protons a and f show a relatively large coupling. These data are consistent with the structural assignment. r-Acetoxynitrosamine Decomposition Products. 3-Acetoxy-2-hydroxy-N-nitrosomorpholine 17 underwent rapid decomposition in phosphate buffer (pH 7, 25 °C). As shown in Scheme 6, the products were acetaldehyde 26 (10%), glyoxal 15 (95%), ethylene glycol 27 (55%), and acetic acid (65%). Except for glyoxal, identification and quantification were done by GC and GC-MS. The aldehydes were independently quantified as their 2,4dinitrophenylhydrazones. While the nature of the R-acetate decomposition is discussed below, the reaction products are consistent with the generation of the R-hydroxynitrosamine 13. Its facile decomposition then gives rise to less stable diazonium 24, which could decompose by paths a-c. Hydride migration by path a gives rise to 25 from which acetaldehyde 26 and glyoxal 15 are formed. Direct displacement of N2 from 24 (path b) will give ethylene glycol 27 and glyoxal. In contrast, the assisted displacement of N2 by path c followed by H2O addition to the resulting oxonium ion 28 will give 29 which will then hydrolyze to ethylene glycol and glyoxal.

The fact that only 10% of the transformation gives rise to acetaldehyde suggests that path c is competing. We have shown that the decomposition of the 2-hydroxyethyl diazonium ion gives rise to more acetaldehyde, suggesting that the diazonium ion is being captured by another process different from direct hydrolysis (path b). Precedent for the participation of neighboring carbonyl groups in this type of assisted diazonium decomposition is provided in the publications of the Hecht group (38, 39). This intramolecular trapping of the electrophile could be significant in reducing the efficacy of DNA alkylation by the diazonium ion but would always generate glyoxal equivalents, which also bind DNA. As we describe below, following the kinetic analysis, there are several routes from 17 to the R-hydroxynitrosamine 13, but these do not alter our conclusions regarding the pathways from 13 to the products shown in Scheme 6. The hydrolysis of 5-acetoxy-2-hydroxy-N-nitrosomorpholine 18 gives glycolaldehyde (15%) and 2-acetoxyethanal 32 (62%) as shown in Scheme 8. The kinetic data require us to consider two different mechanisms to these products as is discussed below. Kinetics of Hydrolysis. The kinetics of hydrolysis for both 17 and 18 were measured by following the change in UV absorbance with time. Typical data, characteristic of all runs at various pH values in various buffers (see Experimental Procedures), taken for the hydrolysis of 18 in citrate buffer at pH 5, are presented in Figure 3. The decay kinetics are well-behaved and are easily fit to the equation A ) Af + A0e-kt by nonlinear least-squares regression, where A is the observed absorbance and Af, A0, and k (the rate constant) were varied. The rate constants, an average of 2-3 runs, are given in Table 1. The analysis of the absorbance versus time data obtained for the hydrolysis of 17 was more problematic. Data for reactions in 0.1 M acetate buffer could be fit reasonably well by the methods used for 18. In contrast, the data from the hydrolyses of 17 in 0.1 M phosphate

Reactive Intermediates from 2-Hydroxy-N-nitrosomorpholine

Figure 3. The first-order exponential decay (O) of 18 due to hydrolysis in 0.1 M pH 5 citrate buffer at 37 °C is shown with the results of three parameter (Af, A0, k) NLSQ fit (line) of the data to the equation A ) Af + A0e-kt, where A is the experimentally observed absorbance as a function of t, and k is the rate constant.

Figure 4. The change in absorbance resulting from the hydrolytic decay (O) of 17 in 0.1 M pH 4 phosphate buffer at 37 °C is shown. An apparent “induction time” before the commencement of exponential decay is observed. The data were fit to rate equations in two ways. The dashed line shows the fit of the data following the induction time to the single-exponential function of Figure 3. The solid line results from a five-parameter fit to a double-exponential function for two successive first-order reactions (see text).

buffer exhibited apparent induction times. Various mixing procedures were examined. Induction times were not a function of the mixing procedure when a 100 µL volume of a standard solution of 17 in THF was diluted with 3.9 mL of buffer solution. A typical data set is shown in Figure 4. Induction times were inversely proportional to pH (see Table 1 and Figure 5). We experimented with a number of logical fitting procedures. The results of two methods for fitting the data to rate equations are shown in Figure 4. The dashed curve shows the results of a fit of the data following the induction time to the exponential function described above. The rate constants given in Table 1 under the column label Phosphate (set 1) were obtained in this way. The increase in absorbance as a

Chem. Res. Toxicol., Vol. 18, No. 12, 2005 1961

Figure 5. A plot of the log of the average induction time observed for the hydrolysis of 17 in 0.1 M phosphate buffer at 37 °C is shown to be a linear function of pH.

function of time for the early part of the reaction and mechanistic considerations suggested that the reaction was characterized by the relatively rapid production of at least one intermediate with an extinction coefficient greater than the substrate, which then decayed more slowly. Accordingly, we examined the five-parameter fit of the data to the equation A ) Af + A1e-k1t + A2(k1/(k2 k1))(e-k1t - e-k2t), which was derived for two successive first-order reactions, and where A is the observed absorbance, the parameters Af, A1, and A2 are constants proportional to extinction coefficients, and k1 and k2 are rate constants. The solid curve in Figure 4 is the result of such a fit. It returns k1 ) 1.5 ( 0.2 × 10-3 s-1 and k2 ) 0.300 ( 0.009 × 10-3 s-1, respectively. The value of k2 is very close to that obtained from the single-exponential analysis (k ) 0.290 ( 0.005 × 10-3 s-1). This technique did not serve well at higher pH values, however, because the values of k1 and k2 became very close in magnitude and the resulting fit was to a single exponential. As a result, all of the rate constants for the hydrolysis of 17 given under column Phosphate (set 1) in Table 1 were obtained by fitting the decay after the induction period. As mentioned above, the induction times decreased as the base concentration of the media increased. A plot of log (induction time) versus pH is linear with a slope of -0.87 ( 0.05 (see Figure 5). In additon to the analysis described above, the fact that the induction time decreases exponentially with an increase in pH suggests that rate processes are responsible for the induction time. At the wavelength we are following the reaction (234 nm), the absorbance is principally due to the NNO group of the nitrosamine. The decay in the absorbance is due to the destruction of this chromaphore, presumably by conversion of some species into a short-lived diazonium ion. In addition to the substrate, the NNO group could be present in one or two intermediates which have different but similar extinction coefficients to 17. The lifetimes of these intermediates are likely a function of pH and buffer. These phenomena could explain our observations regarding the induction times. Temporally, our initial explorations of the kinetic behavior of the hydrolysis of 17 at various pH values were done at 25 °C. In these experiments, a known amount of

1962

Chem. Res. Toxicol., Vol. 18, No. 12, 2005

Loeppky et al. Scheme 9

Figure 6. The pH rate profiles for the hydrolyses of 17 and 18 are presented.

the substrate was added to the cuvette in CH2Cl2, followed by evaporation of the solvent in a stream of N2. Buffer (25 °C) was added, and absorbance measurements were started after stirring. Induction times were observed here as well, but since these were the first measurements, rate constants were estimated by using t1/2 determinations. While this method was accurate for the runs at higher pH (above 7), we later subjected all the data to the same procedures described above (see Experimental Procedures). The rate constants are given in Table 1 in the column under Phosphate (set 2). While these data are not as precise, they are self-consistent and instructive. Plots of log k for these runs versus pH reveal a shape for the plot that is unusual for R-acetoxynitrosamine hydrolyses (Figure 6). Base catalysis begins at pH 5 but appears to reach a maximum near pH 8-9. Mechanistic Analysis. The more accurate pH rate profiles for the hydrolyses of 17 and 18 obtained at 37 °C from the data of Table 1 are also given in Figure 6. Inspection of these plots shows that the hydrolysis behavior of these two substrates is very different. Fishbein and co-workers have carefully examined the kinetics and mechanisms of hydrolyses of a number of R-acetoxynitrosamines (27, 34-37, 40). Their pH rate profiles for the hydrolysis of secondary R-acetates, such as 1-acetoxyethylethylnitrosamine, or heterocyclic compounds, such as R-acetoxy-NMOR, for example, have the same shape as the pH rate profile for 18, except that our profile is displaced to the left along the pH axis by 2.5 to 3 pH units (27, 41). In this pH range (2-12), the Fishbein group has shown that their compounds hydrolyze by two mechanisms (27, 34-37, 40), one where the rate is independent of pH and one where the reaction is basecatalyzed. The pH-independent mechanism involves the SN-1 solvolysis of the acetate to give the stabilized carbocation, a nitrosiminium ion, which is then captured by a nucleophile, most commonly water, to give the R-hydroxynitrosamine (34). Fishbein’s group could trap the carbocation with azide ion. The second mechanism involves standard base-catalyzed ester hydrolysis and also gives the R-hydroxynitrosamine. Although we did not attempt to trap the carbocation derived from 18 with azide, the similarity of our pH rate profile for the hydrolysis of this substrate to those presented by the

Fishbein group suggests that this compound is hydrolyzing in the pH-independent region (≈2-6) by the same solvolysis mechanism as elucidated by the Fishbein group, which is depicted for 18 in Scheme 7. The R-hydroxynitrosmine 14 then goes on to give products as illustrated in Scheme 8. For the compounds investigated by the Fishbein group, hydrolysis in the basic region (pH > 9) occurs by normal base-catalyzed ester hydrolysis to give the R-hydroxynitrosamine, which decomposes rapidly (24, 25, 27, 3437, 40-42). In the case of 18, we observe a similar increase in the decomposition rate of the R-acetate, but base catalysis is evident already at pH 7 (see Figure 6). At pH 9, 18 is undergoing hydrolytic decomposition approximately 2000 times more rapidly than is the R-acetate of NMOR and related nitrosamines (27, 41). This suggests the inception of some form of intramolecular catalysis (neighboring group participation) near pH 7. We propose that this involves base-catalyzed opening of the hemiacetal to 35, a well-documented process, followed by intramolecular base-catalyzed acyl transfer to give the R-hydroxynitrosamine 36 as shown in Scheme 9. This transacylation (43) is followed by a set of rapid transformations involving conversion of the R-hydroxynitrosamine into a diazo hydroxide, and then to 37 and the diazonium ion 16, decomposition of which is illustrated in Scheme 9. The hydrolysis pH rate profile of 18 tells us that two processes are operating at pH 7, the acidity at which the product study was done. These hydrolytic pathways (Schemes 8 and 9) generate different R-hydroxynitrosamines 14 and 36 and subsequently lead to diazonium ions 31 and 16, respectively. Except for glycol aldehyde, 7, the stable decomposition products of the two pathways are different. The chemistry shown in Scheme 8 generates 2-acetoxyethanal 32, the major decomposition product (62%). Although we consider it less likely, 32 could also be formed by the combination of diazonium ion 16 with the OH group of ethylene glycol mono acetate 37, both of which are formed together from the decomposition of the R-hydroxynitrosamine 36 (Scheme 9). The pathways could be distinguished by labeling experiments, which we have not attempted. Our product analysis would seem to favor the operation of the chemistry shown in Scheme 8 under the conditions used for product isolation and identification. Unless 37 is trapped by the diazonium ion 16, the chemistry shown in Scheme 9 should produce a significant amount of 37, which we did not detect. Reference to Scheme 8 and the product yields show that the most significant reaction channel involves hy-

Reactive Intermediates from 2-Hydroxy-N-nitrosomorpholine

dride migration as the diazonium ion 31 loses nitrogen (path a). This decreases the diazonium alkylating potential to