Chem. Res. Toxicol. 1996,8, 875-883
875
Prostaglandin H Synthase-Catalyzed Ring Oxygenation of 2-Naphthylamine: Evidence for Two Distinct Oxidation Pathways John F. Curtis, Kenneth Tomer, Steve McGown, and Thomas E. Eling* National Institute of Environmental Health Sciences, P.O. Box 12233, Research Triangle Park, North Carolina 27709 Received November 21, 1994@ Previous studies showed that prostaglandin H synthase (PHS) cooxidizes 2-naphthylamine (2-NA) to ring-oxygenated products. These metabolites are atypical for a peroxidase-mediated reaction and are completely different from the polymeric nonoxygenated metabolites of 2-NA t h a t are generated with the model peroxidase horseradish peroxidase (HRP). In this study, we investigated possible explanations for the PHS-catalyzed formation of ring-oxygenated 2-NA metabolites. We found that introduction of a peroxyl radical-generated cosubstrate into the HRP/2-NA system resulted in the formation of the same ring-oxygenated products observed in the PHS/2-NA system. l8Oz incorporation studies were utilized to further characterize the source of oxygen in the ring-oxygenated 2-NA metabolites. The data show that, in the case of PHS, ring oxygenation can occur both by peroxyl radical-mediated attack on 2-NA and by direct transfer of peroxide oxygen from PHS to 2-NA.
Introduction 2-Naphthylamine (2-NA)lis one of a series of aromatic amines which is a documented human bladder carcinogen (1,2). The mechanism of 2-NA-induced bladder carcinogenesis is as yet not clearly understood, although a number of proposals exist. A potential explanation for 2-NA-induced carcinogenesis that has been studied in this and other laboratories is direct activation of 2-NA in the bladder (3-5). An enzyme potentially responsible for this metabolic activation is prostaglandin H synthase (PHS), an enzyme which contains both cyclooxygenase and peroxidase activity (6). During the reduction of a peroxide, the peroxidase requires two electrons to return to the ground state. This return to the ground state typically proceeds by two one-electron oxidation steps. A wide range of compounds, including some aromatic amines, are suitable electron donors and have been demonstrated to serve as reducing cosubstrates for PHS peroxidase (7). Some interesting observations have been made regarding the PHS/2-NA system. Previous work from this laboratory demonstrated that 2-NA is metabolized by PHS to products that are mutagenic (8)and are reactive with DNA (3). The 2-NA metabolite profile seen with PHS is unusual for a peroxidase-mediated reaction. Oxidation of 2-NA by PHS yields low levels of ring oxygenation products, instead of the predicted free radical addition products (3). A subsequent report by Bull demonstrated that 2-NA is either a very poor cosubstrate
* Corresponding author. Phone # (919) 541-3911; FAX # (919) 5411460. @Abstractpublished in Advance ACS Abstracts, July 1, 1995. Abbreviations: 2-NA, 2-naphthylamine; PHS, prostaglandin H synthase; ROO', peroxyl radical; PAH, polycyclic aromatic hydrocarbons; DTPA, diethylenetriaminepentaaceticacid; BSTFA, bis(trimethylsily1)trifluoroacetamide; BHA, butylated hydroxyanisole; RSVM, ram seminal vesicle microsomes; HRP, horseradish peroxidase; 13-HPODE, 13-hydroperoxyoctadeaclienoicacid; AQI, 2-amino-1,4-naphthoquinone N4-2-naphthylimine;ADN, 2-aminodinaphthylamine; DBP, dibenzo[a,hlphenazine; PGGz, prostaglandin Gz; INQ, S-imino-l-naphthoquinone; 'OOSO3, sulfite-derived peroxyl radical; BP-7,8-diol, benzo[alpyrene 7,8-dihydrodiol.
or not a cosubstrate for PHS peroxidase (9),an observation which is consistent with the absence of free radicalmediated 2-NA metabolites. Studies by Yamazoe et al. showed the in vivo formation of ring-oxygenated 2-NADNA adducts in canine urothelium (4). These workers also found that PHS generated these same DNA adducts in vitro. These observations led us to conclude not only that the 2-NAIPHS system poses an interesting mechanistic question, but that the unusual characteristics of this reaction may also be of importance in understanding the mechanism of 2-NA-induced bladder cancer. With this perspective in mind, we decided to investigate the mechanism of PHS-catalyzed 2-NA oxidation. As expected, we found that the formation of ringoxygenated 2-NA metabolites is not the result of a traditional peroxidase-mediated reaction. Instead, PHScatalyzed 2-NA ring oxygenation apparently occurs by two distinct mechanisms. Our results show that peroxyl radicals (ROO') generated during peroxidase-mediated reactions are capable of catalyzing 2-NA ring oxygenation. This type of ROO'-mediated reaction has been previously demonstrated to be responsible for the PHScatalyzed oxidation of polycyclic aromatic hydrocarbons (PAH) (10, 11). Our data also show that direct oxygen transfer from the peroxidase to 2-NA also occurs. PHSmediated direct transfer of the peroxide oxygen to aromatic amines has not been previously demonstrated.
Materials and Methods Caution: 2-Naphthylamine is classified as a carcinogen and should be handled with extreme caution. Work with solid 2 - N A (i.e., preparation of stock solutions) should be conducted i n a fume hood at the very least. Materials. [ r i r ~ g - ~ H ] - 2 - N (196 A mCUmmo1) was purchased from Midwest Research Institute (Kansas City, MO). 2-NA and [3H]-2-NAwere purified as needed by solid phase extraction or HPLC to ensure purity of 299%. Horseradish peroxidase (type VI-A), soybean lipoxygenase (type I), hematin, diethylenetriaminepentaacetic acid (DTPA), 2-NA, and phenylbutazone were from Sigma Chemical Co. (St. Louis, MO). l80z (98% purity)
This article not subject to U S . Copyright. Published 1995 by the American Chemical Society
876 Chem. Res. Toxicol., Vol. 8, No. 6, 1995 was purchased from MSD Isotopes (Montreal, Canada). Anhydrous sodium sulfite and HPLC grade methanol were from Fisher Scientific (Pittsburgh, PA). Linoleic acid and arachidonic acid were purchased from NuChek Prep (Elysian, MN). Bis(trimethylsily1)trifluoroacetamide(BSTFA) was purchased from Supelco (Bellefonte, PA). Butylated hydroxyanisole (BHA) was from Fluka (Ronkonkoma, NY).Ram seminal vesicles were purchased from Oxford Biomedical (Oxford, MI). Metabolism of 2-NA. Ram seminal vesicle microsomes (RSVM)were used as a source of PHS and were prepared as previously described (12). Metabolism of 2-NA was measured as follows: RSVM or horseradish peroxidase (HRP) was preincubated with 50 ,uM [3H]-2-NA(0.02 pCi/mmol) for 2 min a t 37 "C. If present, reducing cosubstrates (phenylbutazone, sodium sulfite, or phenol) were also included in this preincubation. Reactions were initiated by the addition of arachidonic acid, 13hydroperoxyoctadecadienoic acid (13-HPODE), or hydrogen peroxide. The reactions were conducted in 100 mM potassium phosphate (pH 7.6) for studies with RSVM. Incubations with HRP were conducted in 100 mM sodium acetate (pH 6.0). Specific conditions for incubations are listed in the figure legends. 2-NA solutions were prepared in methanol. Arachidonic acid, 13-HPODE, and phenylbutazone were in ethanol. Phenol was prepared in water. Sodium sulfite was prepared in 100 mM potassium phosphate containing 1 mM DTPA. Incubations were conducted a t 37 "C, and the final reaction volume was 2 mL. The reactions were terminated after 5 min by pouring the incubation mixture onto a n activated C-18 PrepSep solid phase extraction column (Fisher Scientific) and eluting the sample with a partial vacuum. Columns were rinsed with 5 mL of water, and the products were subsequently eluted with 5 mL of methanol. The solvent was removed using a Speedvac Model SS-3 (Farmingdale, NY).Samples were reconstituted in 400pL of HPLC grade methanol. HPLC analysis was conducted using a previously described method (13). Sodium citrate (10 mM, pH 6.0) was used as solvent A and methanol as solvent B. Initial solvent composition was 95% A which was reduced to 10%A using a 20 min linear gradient. Solvent composition was held at 10% A for the remainder of analysis. The total flow rate was 2.0 m u m i n during the course of the run. The following Waters (Milford, MA) HPLC system was used: two Model 6000A pumps, a Model 710B WISP, a Model 721 solvent programmer, and a Model 990 photodiode detector. The HPLC system was also equipped with a Radiomatic Model CR detector (Tampa, FL) for monitoring radioactivity. Ecolume (ICN, Costa Mesa, CA) was used as the scintillant. Analysis was conducted using a Waters pBondapak C-18 column (3.9 mm x 30 cm). Aliquots (150 uL) were analyzed by HPLC. Aliquots (150 ,uL) were also counted for radioactivity using a Packard (Downer's Grove, IL) Model 2000 CA scintillation counter to calculate recoveries. Reaction of ROO with 2-NA. Sodium sulfite (5 mM) was added to a n oxygraph cell containing 1.5 mL of sodium citrate (pH 6.0) a t 37 "C. L3H1-2-NA(50,uM, 1.2 uCi), in acetone, was added either before or after addition of the sulfite. Oxygen uptake was monitored using a Gilson (Middleton, WI) Model 5i6 oxygraph equipped with a Clark type electrode. After the reaction was complete, a 1.0 mL aliquot was directly injected onto the HPLC. This necessitated collecting fractions (at 30 s intervals) instead of using our flow detector, due t o interference from the sample with the scintillation cocktail. This was a transient artifact that disappeared by allowing the vials containing Ecolume and sample to sit before scintillation counting. UVNis Spectrophotometric Measurement of 2-NA Metabolism. All spectrophotometric studies were conducted using a Hewlett-Packard (Palo Alto, CA) Model 8450A spectrophotometer equipped with a heated stirring module. Incubations were conducted as described in the "Metabolism of 2-NA" section above. Solubilized RSVM, prepared in 100 mM potassium phosphate (pH 7.81, containing 1% Tween 20 were used as a source of PHS for the spectrophotometric assays. Preparation of [leO1-13-HPODE. [la0]-13-HPODE was biosynthesized adapting the method described by Funk et al. for the synthesis of 15-hydroperoxyeicosatetraenoicacid (14).
Curtis et al. One hundred milliliters of 100 mM Tris (pH 9.0) was added to the l 8 0 2 apparatus. The reaction flask was immersed in ice and purged with argon for 15 min. The apparatus was then evacuated and filled to atmospheric pressure with l a 0 2 . The buffer was stirred vigorously to enhance the rate of oxygen solvation. Soybean lipoxygenase ( 5 mg) was added to the buffer through a septum. The reaction was initiated by addition of 50 mg of linoleic acid. The reaction mixture was stirred vigorously for 5 min, at which time the reaction was terminated by acidification to pH 3 with 6 N HC1. Purification then proceeded as described (14). Analysis of products by HPLC yielded one peak which coeluted with authentic 13-HPODE.GC/ MS analysis determined the isotopic purity of the biosynthesized [1a0]-13-HPODEto be greater than 95%. Measurement of lSO Incorporation into 2-NA. Incorporation of l a 0 2 into 2-NA was conducted in the same apparatus used to synthesize [1a01-13-HPODE. In order to ensure experimental consistency, the same procedure was used when either l 6 0 2 or laO2 was used as a n atmosphere. One hundred milliliters of 100 mM potassium phosphate (pH 7.6) was added to the reaction apparatus, which was maintained a t 37 "C. l 6 0 2 or l a 0 2 was added to the vessel as described above. RSVM (0.5 mgimL) were then added to flask through a septum. 2-NA (50 ,uM)was then added and allowed to preincubate for 2 min. The reaction was initiated by the addition of arachidonic acid (100 pM), [l80]-l3-HPODE(50 ,uM),or hydrogen peroxide (100 uM). After 5 min, the reaction was terminated by the addition of 500 ,uM BHA through the septum. The reaction mixture was then briefly flushed with argon and then processed by distributing among 5 (2-18 Prep-Sep columns. Each column was rinsed with 10 mL of water, followed by 10 mL of 50/50methanovwater to remove 2-NA. The metabolites were then eluted with 5 mL of methanol. The methanol fractions were pooled, and the solvent was removed using rotary evaporation. The metabolites were reconstituted in 1 mL of diethyl ether and resolved on a semipreparative TLC plate (20 x 5 cm, 1000 pm thickness). Benzeneiacetone (8:l) was used as a developing solvent. The reddish band corresponding to 2-amino-1,4-naphthoquinone N42-naphthylimine (AQI) was isolated, and the metabolite was extracted with 2 x 2 mL of diethyl ether. The solvent was pooled and removed under argon. The AQI was then derivatized for G C N S by dissolving the sample in 25 pL of BSTFA and heating for 1 h at 80 "C. The sample was stored in BSTFA a t -80 "C until G C M S analysis. GCMS Analysis of AQI. G C M S analysis was conducted using a Hewlett-Packard Model 5890 gas chromatograph coupled to a Kratos (Ramsey, NJ) ISQ mass spectrometer. Analysis conditions are given in the figure legends.
Results and Discussion 2-NA Metabolism by PHS and HRP. PHS and HRP metabolize 2-NA to different products (31, as shown in Figure 1. The HRP-catalyzed metabolism 2-NA free radical undergoes additional reactions, to ultimately form the stable end products which were isolated (Figure 1A). As was previously reported (31, there were two major metabolites, 2-aminodinaphthylamine (ADN; t~ = 19.1 min) and dibenzo[a,h]phenazine (DBP; t~ = 24 min), which were identified by their W/vis spectra and reported HPLC retention times. Two other previously observed minor metabolites ( t =~ 21 and 22 min) were not identified; based on their W/vis spectra (data not shown), they are presumably of polymeric origin. The PHS-catalyzed oxidation of many aromatic amines proceeds by one-electron oxidation (7).Aromatic amines that are metabolized by this mechanism are, in general, good cosubstrates for PHS in peroxidase assays. However, there is another class of aromatic amines that is not oxidized by PHS in this manner. These compounds are poor reducing cosubstrates in standard peroxidase
Chem. Res. Toxicol., Vol. 8, No. 6, 1995 877
Ring Oxidation of 2-Naphthylamine 20,000
20,000
A 2-NA
\
I
2-NA
DBP
E
/
AQl
E
Q 0
P 0
INQ
a 2,000
5
0
B
0
1
rr n 2-NA
\
E
Q 0
min Figure 1. HPLC profiles of 2-NA metabolites catalyzed by HRP and PHS. The chromatogram in panel A shows the metabolites from the HRP-catalyzed oxidation of 2-NA. Incubation conditions were as follows: 0.5 pg/mL HRP, 50 pM L3H1-2-NA(0.02 pcilnmol), and 100 pM HzOz in 100 mM sodium acetate (pH 6.0). Panel B shows the metabolites formed from the PHScatalyzed oxidation of 2-NA. Incubation conditions were as follows: 0.5 mg/mL RSVM, 50 pM L3H1-2-NA(0.02 pcilnmol), and 100 pM HzOz in 100 mM KPOI (pH 7.6).
assays (91, have very limited turnover, and do not yield the predicted free radical reaction products. 2-NA is a n aromatic amine that fits this profile. As can be seen in Figure lB, PHS oxidizes 2-NA to two major products. The major metabolite was identified as AQI ( t R = 21 min) on the basis of cochromatography, UV/vis spectroscopy, and GC/MS. Although AQI had the same retention time as one of the minor peaks in Figure lA, the HRP-derived metabolite was not AQI. A second peak was observed (unknown 1, t R = 18.5 min) that had a distinctive UV spectrum nearly identical to that of AQI. Attempts to identify this metabolite were unsuccessful. None of the characteristic 2-NA one-electron oxidation products were found, which is in complete contrast with what was found with the HRPI2-NA system. The Effect of Peroxyl Radicals on HRP-Catalyzed 2-NAOxidation. One of the basic differences between HRP and PHS is that ROO' are generated as intermediates during PHS-catalyzed conversion of arachidonic acid to prostaglandin Gz (PGG2) (7). In contrast, there are no corresponding ROO' generated in the HRP-catalyzed reduction of H2Oz. Because ROO' are oxidizing agents
Figure 2. The effect of peroxyl radicals on the HRP-catalyzed oxidation of 2-NA. The chromatogram shows the effect of peroxyl radicals on HRP-catayzed 2-NA oxidation. Incubation conditions were as follows: 0.5 pglmL HRP, 50 pM L3H1-2-NA (0.02 pcilnmol), 100 pM sodium sulfite, and 100 pM HzOz in 100 mM sodium acetate (pH 6.0).
in their own right, their presence represents a potentially important difference between PHS and HRP. There are precedents for PHS-generated ROO' acting to oxidize carcinogens (16, 17). The formation of ROO' has been invoked as a prospective mechanism for explaining why PHS, but not HRP, bioactivates selected xenobiotics, for example, PAH, in mutagenicity assays (18). In order to test the hypothesis that ROO' were involved in the C-oxygenation of 2-NA, a ROO' generator was introduced into the HRP/B-NA system. Sulfite was chosen a s a ROO' generator because it is a peroxidase cosubstrate and can generate high concentrations of sulfite-derived peroxyl radicals ('OOSO3-) during the same brief time span in which which 2-NA is oxidized. Sulfite is oxidized by HRP to a sulfur-centered anion radical (19). This free radical, in turn, traps molecular oxygen, resulting in the formation of *OOSO3-. As can be seen in Figure 2, the addition of sulfite to incubations containing HRP dramatically alters the metabolism profile of 2-NA. The free radical addition products ADN and DBP are still present, although in diminished quantities. The major change in the chromatogram, however, is the appearance of ring-oxygenated metabolites: AQI and a second product which was tentatively identified as 2-imino-1-naphthoquinone (INQ; t~ = 13 min). This structure was proposed for this peak on the following basis: (1)Kadlubar et al. (13) report INQ as having a retention time of 13.5 min with the same HPLC method. (2) The compound coelutes with the analogous 1,2-naphthoquinone and has a very similar UV/vis spectra. (3) The compound is also moderately unstable, which while precluding positive identification by mass spectrometry, is also consistent with the proposed structure. To further study the effects of ROO' on HRP-dependent metabolism of 2-NA, two other cosubstrates were coincubated with 2-NA. Phenylbutazone was used as a second ROO' generator (201,while phenol was used as a cosubstrate that would not generate ROO'. The results of this study are shown in Table 1. One-electron oxidation of 2-NA, measured as formation of ADN and DBP, occurred in all samples containing HzO2. Ring oxygen-
878 Chem. Res. Toxicol., Vol. 8, No. 6, 1995
Curtis et al.
Table 1. The Effect of Different Reducing Cosubstrates on the HRP-CatalyzedOxidation of 2-NA ~
coupling products incubation conditionsa control control - H202 control + sulfite control sulfite - Hz02 control phenol control phenol - H z 0 ~ control phenylbutazone control phenylbutazone - H 2 0 ~
+ + + ++
ADNb 11.5 & 2.0 4.2
DBP 15.6 5 2.3 4.8 4.2
-
-
-c
* 3.8 10.4 * 1.4
~
~~~~~~
ring-oxidation products
*
unknown 1 4.1 f 2.6
AQI -
-
11.0 -
-
-
-
-
-
7.0 5 1.4
6.7 5 0.3
8.4 f 1.3
-
-
-
25.2 -
10.8
0.2
* 1.1 * 5.3
a Incubation conditions for the control were as follows: 0.5 pglmL HRP, 50 pM r3H1-2-NA(0.02 pCilnmol), and 100 pM HzOz in 100 mM sodium acetate (pH 6.0). The sample volume was 2.5 mL. Where present, sodium sulfite, phenol, and phenylbutazone were all used at a concentration of 100 uM. The amount of each product formed is shown as total recovered nmol. Not detected.
100-
%
