Characterization of Multiple Products of Cytochrome P450 2A6

The primary metabolite of nicotine in smokers is cotinine. Cotinine is further ... In this work, we report that P450 2A6 metabolizes cotinine not only...
0 downloads 0 Views 111KB Size
Download PDF
Chem. Res. Toxicol. 1999, 12, 639-645

639

Characterization of Multiple Products of Cytochrome P450 2A6-Catalyzed Cotinine Metabolism Sharon E. Murphy,* Lisa M. Johnson, and Dominic A. Pullo Department of Biochemistry, Molecular Biology and Biophysics, University of Minnesota Cancer Center, Minneapolis, Minnesota 55455 Received January 28, 1999

The primary metabolite of nicotine in smokers is cotinine. Cotinine is further metabolized to trans-3′-hydroxycotinine, the major urinary metabolite of nicotine in tobacco users. It was recently reported that cytochrome P450 2A6 catalyzes the conversion of cotinine to trans-3′hydroxycotinine. In this work, we report that P450 2A6 metabolizes cotinine not only to trans3′-hydroxycotinine but also to 5′-hydroxycotinine, norcotinine, and a fourth as yet unidentified metabolite. The products of baculovirus-expressed P450 2A6 [methyl-3H]cotinine metabolism were analyzed by radioflow HPLC. Three 3H-labeled metabolites were detected and were present in approximately equal amounts. The identities of two of the metabolites were confirmed to be 5′-hydroxycotinine and trans-3′-hydroxycotinine by LC/MS/MS and LC/MS analysis and comparison to standards. The third product was not identified. A fourth product of P450 2A6catalyzed cotinine metabolism was detected by LC/MS. It was identified by cochromatography with a standard and MS and MS/MS data to be norcotinine. An attempt was made to further characterize the unidentified 3H-labeled metabolite by comparison to the cotinine metabolites generated by hamster liver microsomes. Hamster liver microsomes contain a P450, 2A8, which is closely related to P450 2A6, and have previously been shown to metabolize cotinine to three hydroxylated products, trans-3′-hydroxycotinine, 5′-hydroxycotinine, and N-(hydroxymethyl)norcotinine. We were unable to confirm that N-(hydroxymethyl)norcotinine was the unidentified cotinine metabolite generated by P450 2A6.

Introduction

Scheme 1. Metabolism of Cotinine

Cotinine is the primary metabolite of nicotine in smokers (1). It has been estimated that 80% of the nicotine absorbed by a smoker is metabolized to cotinine (2). This occurs in two steps. The first is cytochrome P450-catalyzed oxidation of nicotine to nicotine ∆-1′,5′imminium ion (3). The imminium ion is then converted to cotinine by cytosolic aldehyde oxidase (4). Several studies support the role of P450 2A6 in the oxidation of nicotine to the imminium ion (5-8). Recently, it has been reported that P450 2A6 also catalyzes the conversion of cotinine to trans-3′-hydroxycotinine, albeit with a relatively high Km (265 µM) and a low Vmax (37 pmol min-1 mg-1) (9). The conversion of cotinine to trans-3′-hydroxycotinine is a major route of nicotine metabolism in people (1). In smokers, urinary trans-3′-hydroxycotinine and its glucuronide conjugate account for more than 50% of the nicotine dose received by a smoker (2). The cis isomer of 3′-hydroxycotinine is present at less than 0.3% of the level of the trans isomer (10). Some of the pathways of cotinine metabolism are illustrated in Scheme 1. In vivo, cotinine may be metabolized to 5′-hydroxycotinine, norcotinine, and cotinine N-oxide as well as trans-3′-hydroxycotinine (11). More than 30 years ago, 3′-hydroxycotinine was isolated from smoker’s urine as well as from the urine of dogs, rats, and humans treated with cotinine (reviewed in ref 1). This 3′-hydroxycotinine metabolite was determined to be the trans isomer by Dagne and Castagnoli in 1972 (12). But not until 1987 was it established that trans-3′hydroxycotinine was a major nicotine metabolite in smokers (13). 5′-Hydroxycotinine which is in equilibrium

with 4-(3-pyridyl)-4-oxo-N-methylbutyramide (keto amide, Scheme 1) has been isolated from the urine of rats and monkeys treated with cotinine (14, 15). Norcotinine and cotinine N-oxide are urinary metabolites of cotinine in rats and monkeys (14-17). In smokers’ urine, norcotinine, cotinine N-oxide, and 5′-hydroxycotinine are each present at about 2% of the nicotine dose (1, 2). In

10.1021/tx990015e CCC: $18.00 © 1999 American Chemical Society Published on Web 06/11/1999

640 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

summary, all these alternate pathways of cotinine metabolism appear to be minor in smokers. The in vitro metabolism of cotinine has not been well studied. Li and Gorrod reported that hamster liver microsomes metabolize cotinine to cotinine N-oxide, trans-3′-hydroxycotinine, 5′-hydroxycotinine, and N-(hydroxymethyl)norcotinine (Scheme 1) (18). While this was not a quantitative study, the relative amounts of these four metabolites appeared to be similar. The authors suggested that N-(hydroxymethyl)norcotinine might be the metabolic precursor to norcotinine. In their study of the P450 2A6 metabolism of cotinine, Nakajami and co-workers did not look for any cotinine metabolites other than 3′-hydroxycotinine (9). They refer to the 3′-hydroxycotinine product that they detected as the trans isomer, although their assay would probably not have distinguished between the cis and trans isomers. We previously reported that the P450 2A6-catalyzed metabolism of the nicotine-derived nitrosamine N′-nitrosonornicotine (NNN)1 is specific (19). That is, P450 2A6 catalyzes only the 5′-hydroxylation of NNN. In the study reported here, we determined whether P450 2A6 catalyzes the metabolism of cotinine specifically to trans3′-hydroxycotinine or whether other metabolites are also formed.

Experimental Procedures Chemicals. NADP, glucose 6-phosphate, and glucose-6phosphate dehydrogenase were purchased from Sigma Chemical Co. Cotinine was purchased from Aldrich Chemical Co. (St. Louis, MO) and its purity confirmed by GC (20) and HPLC (system II, below). [3H-me]Cotinine was obtained from Amersham in 1988 and stored at -80 °C (specific activity of 80.7 Ci/ mmol). The purity of this preparation was 60% when analyzed by HPLC (system II). Therefore, it was purified by HPLC (system II) and stored in methanol at -80 °C until it was used. Solutions with specific activities from 1.0 to 0.02 Ci/mmol were prepared as needed. The purity of cotinine was confirmed immediately prior to use, and if necessary, it was again purified by HPLC. This was often found to be necessary since a breakdown product of [3H-me]cotinine coeluted with a cotinine metabolite. trans-3′-Hydroxycotinine was provided by S. Amin (American Health Foundation, Valhalla, NY). 4-(3-Pyridyl)-4-oxo-N-methylbutyramide (keto amide) was provided by S. Hecht (University of Minnesota Cancer Center, Minneapolis, MN). Its purity was confirmed by GC/MS and identity confirmed by NMR in D2O: 1H NMR (D O) δ 8.63 (1H, s, pyr-2H), 8.50 (1H, d, pyr-6H), 7.85 2 (1H, dd, pyr-4H), 7.47 (1H, dd, pyr-5H), 2.6 (2H, m, 3′- or 4′-H), 2.50 (3H, s, CH3), 2.3 (2H, m, 3′- or 4′-H). NMR spectra confirmed that in H2O the dominant tautomer was the cyclic form 5′-hydroxycotinine. Cotinine N-oxide was synthesized as previously described (16): 1H NMR (CD3OD) δ 8.3 (2H, m, pyr2, 6H), 7.6-7.5 (2H, m, pyr-4, 5H), 4.72 (1H, m, 5′-H), 2.7-2.4 (3H, m, 3′,4′-CH2CH), 2.53 (3H, s, CH3), 1.9 (1H, m, 3′- or 4′-H). Cotinine Metabolism by BV-Expressed P450 2A6. Microsomes were prepared from a P450 2A6 BV expression system as described previously (21). The coumarin 7-hydroxylase activity was 0.4 nmol min-1 µL of microsomes-1. BV-expressed human NADPH-P450 reductase (53 units/µL, 1 unit reduces 1 µmol of cytochrome c/min) was purchased from Panvera Corp. (Madison, WI). P450 2A6 (3-30 µL) was incubated for 5 min at 30 °C with reductase in 100 mM sodium phosphate buffer (pH 7.4). Then an NADPH-generating system (0.4 mM NADP+, 100 mM glucose 6-phosphate, and 0.4 unit/mL glucose-6-phosphate

Murphy et al. dehydrogenase) and 1-200 µM [3H-me]cotinine with varying specific activities were added, and the samples were incubated for an additional 30 min at 37 °C. The final volume was 0.21.0 mL. Reactions were terminated on ice with the addition of a 10% volume of 25% zinc sulfate and a 10% volume of saturated barium hydroxide. The sample was centrifuged, and the supernatant was analyzed by HPLC with a β-Ram radioflow detector (IN/US Systems Inc., Tampa, FL). Samples for LC/MS analysis were prepared in duplicate using 200 µM cotinine. One sample contained [3H-me]cotinine, and the other did not. The former was used to determine the amount of each metabolite present in the sample. The sample containing unlabeled cotinine was either analyzed directly by LC/MS or injected onto HPLC system II and the eluant collected at the correct retention time for each of the cotinine metabolites. The collected fractions were concentrated under a stream of nitrogen prior to LC/MS analysis. Cotinine Metabolism by Hamster Liver Microsomes. Hamster liver microsomes, prepared from male Syrian golden hamsters, were purchased from Xenotech (Kansas City, KS). Hamster liver microsomes were incubated in 100 mM sodium phosphate buffer (pH 7.0) containing an NADPH-generating system and 500 µM cotinine for 45 min at 37 °C. The reaction was terminated by the addition of 1 g of sodium chloride, and then cotinine and its metabolites were extracted with acetonitrile as described previously (18). Note that only 25% of the cotinine was recovered by this procedure. Metabolites were resuspended in methanol and analyzed by cation exchange HPLC (system III, below) with detection by UV (λ260). HPLC Analysis of Cotinine Metabolites. Three HPLC systems were used. System I was similar to that of Nakajima et al. (9). A Phenomenex (Torrance, CA) Bondex 10 µm, C18 column (2.9 mm × 250 mm) was used. The mobile phase consisted of A (0.01% acetic acid, 1.5% acetonitrile, and 1 mM heptane sulfonate) and B (0.1% acetic acid, 40% acetonitrile, and 1 mM heptane sulfonate). The elution conditions were isocratic from 1 to 10 min at 75% A/25% B, followed by a linear gradient to 50% A/50% B from 10 to 12 min, and then isocratic at 50% A/50% B from 12 to 22 min. The flow rate was 1 mL/ min. No difference in the retention times of trans-3′-hydroxycotinine or cotinine was observed when heptane sulfonate was excluded from the mobile phase. System II was analysis by reverse phase HPLC on a C18 Waters µBondapak column (3.9 mm × 300 mm, Waters Corp., Milford, MA). Cotinine and its metabolites were eluted with a linear gradient from 100% A′ [10 mM ammonium acetate buffer (pH 7)] to 70% A′ and 30% methanol over the course of 60 min. The flow rate was 1 mL/ min. System III used a Phenomenex partisil 10 µm, SCX column (4.6 mm × 250 mm) eluted isocratically for 35 min with A′′ (10 mM ammonium acetate and 0.1% TFA), followed by a linear gradient to 90% A′′/10% methanol over the course of 5 min, and then continued isocratic elution at 90% B/10% methanol for 30 min. The flow rate was 1 mL/min. LC/MS and LC/MS/MS Analysis. MS analysis was carried out on a Finnegan 7000 TSQ instrument with an APCI source interfaced with an Alliance HPLC system (Waters Corp.). The APCI settings were as follows: capillary temperature of 250 °C, vaporizer temperature of 350° C, and current of 5 µA. The instrument was tuned to polypropylene glycol, and then the capillary and lens voltages were optimized for detection of 5′hydroxycotinine via several 5 µL injections directly into the APCI source. A collision energy of either 20 or 27 eV with a collision gas pressure of 1 mTorr was used for MS/MS. LC conditions were as described for system II except a 2 mm × 300 mm Waters µBondapak C18 column was used with a flow rate of 0.2 mL/min.

Results 3

1 Abbreviations: NNN, N′-nitrosonornicotine; [3H-me]cotinine, [methyl-3H]cotinine; BV, baculovirus; CID, collision-induced dissociation; APCI, atmospheric pressure chemical ionization.

Metabolism of [ H-me]Cotinine by P450 2A6. The products of [3H-me]cotinine metabolism by BV-expressed P450 2A6 were analyzed by HPLC with radioflow detec-

Cotinine Metabolism by P450 2A6

Figure 1. Radioflow HPLC analysis (system I) of 10 µM cotinine metabolism by BV-expressed P450 2A6. HPLC and enzyme incubation conditions are described in Experimental Procedures.

tion. The first HPLC system (I) used in this study was similar to that of Nakajima and co-workers. These investigators recently reported that P450 2A6 catalyzed the 3′-hydroxylation of cotinine (9). A sample chromatogram of cotinine metabolite analysis with this system is illustrated in Figure 1. Two radioactive peaks were detected; neither was present in control incubations. The first elutes prior to the trans-3′-hydroxycotinine standard, and the second peak, 4.5 min, coelutes with the standard. To further characterize these products, a second HPLC system (II) was developed. Using this system, many known cotinine metabolites were well-separated (Figure 2A). These metabolites include cotinine N-oxide (20.4 min), cis- and trans-3′-hydroxycotinine (29.1 and 30.1 min, respectively), 5′-hydroxycotinine (35.1 min), and norcotinine (36.5 min). When the products of P450 2A6catalyzed cotinine metabolism were analyzed with this system, three radioactive peaks, approximately equal in size, were detected (Figure 2B). One of these, peak 2, coeluted with trans-3′-hydroxycotinine, while another, peak 1, eluted about 0.5 min past cotinine N-oxide. The third peak, 3, coelutes with 5′-hydroxycotinine; however, there appear to be two radioactive peaks eluting in this region. 5′-Hydroxycotinine exists in equilibrium with 4-(3pyridyl)-4-oxo-N-methylbutyramide (keto amide, Scheme 1). The equilibrium favors the cyclic tautomer which exists in a 4-fold excess (22); for simplicity, we will refer to this compound as 5′-hydroxycotinine. To obtain additional evidence that peak 3 is 5′-hydroxycotinine, it was reduced with sodium borohydride, and then reanalyzed by radioflow HPLC. The entire radioactive peak shifted with the 5′-hydroxycotinine standard to an earlier retention time, that of the reduction product, 4-(3-pyridyl)-4hydroxy-N-methylbutyramide. Therefore, peak 3 was tentatively identified as 5′-hydroxycotinine, and we assume the two closely eluting peaks are the two tautomers. It is important to note that if the [3H-me]cotinine was

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 641

Figure 2. HPLC analysis (system II) of cotinine metabolites. (A) UV detection of standards. (B) Radioflow detection of 3Hlabeled metabolites formed when 10 µM [3H-me]cotinine was incubated for 30 min with BV-expressed P450 2A6. The [3Hme]cotinine peak detected is relatively small since it was collected for reuse. The details of the HPLC and incubation conditions are described in Experimental Procedures. Abbreviations: 3′-OH cotinine, 3′-hydroxycotinine; 5′-OH cotinine, 5′hydroxycotinine.

not purified by HPLC immediately prior to use a radioactive peak coeluting with 5′-hydroxycotinine was present in control incubations. However, this peak was not reduced by sodium borohydride. The metabolites in peak 1 were partially sensitive to reduction by sodium borohydride. That is, 10-20% of the peak was reduced to a radioactive product that coeluted with cotinine. Another 20-30% of the radioactivity present in the peak was not recovered after reduction, suggesting the loss of the [3H]methyl group under the conditions of the reduction. It was unclear whether these results suggest the presence of multiple metabolites in this peak or whether one metabolite is being converted to multiple products under the conditions of the reduction. We did not further pursue this approach to characterizing the metabolite(s) present in peak 1. LC/MS and LS/MS/MS Analysis of Cotinine Metabolites. To identify the products of P450 2A6-catalyzed cotinine metabolism, LC/MS and LC/MS/MS analysis was used. Standards were available for trans-3′-hydroxycotinine and 5′-hydroxycotinine, and we prepared cotinine N-oxide. Mass spectral conditions were established to optimize the detection of these three compounds by MS/ MS analysis. The characteristic ions present in product ion spectra of the (M + H)+ ion of these compounds as well as norcotinine are presented in Table 1. Using a collision energy of 27 eV, the major fragment from both 3′- and 5′-hydroxycotinine is the protonated pyridine fragment, m/z 80. A second common fragment is m/z 134. Two additional fragments are unique to 5′-hydroxycotinine (m/z 162, loss of CH3NH2; and m/z 106, PyCO+). The former was the most abundant ion when a collision energy of 20 eV was used.

642 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

Murphy et al.

Table 1. Major Fragments Present in Product Ion Spectra of the (M + H)+ Ion of Cotinine Metabolitesa metabolite cotinine N-oxide trans-3′-hydroxycotinine

m/z (%) fragment lost 96 (100) C5H8NO 98 (38) C5H5NO 134 (55) NCH3, CdO 80 (100) -c

ion compositionb (C5H5NO)H+ C5H8NO+ PyC3H4O+ (C5H5N)H+

5′-hydroxycotinine 173 (10) (-27 eV)d 162 (15) 134 (58) 106 (38) 80 (100)

H 2O NHCH3 NCH3, CdO -c -c

(PyC5H6NO)H+ (PyC4H4O2)H+ PyC3H4O+ PyCdO+ or (PyC2H3)H+ (C5H5N)H+

5′-hydroxycotinine 175 (15) (-20 eV) 162 (100) 134 (68) 106 (12) 80 (45)

H 2O NHCH3 NCH3, CdO -c -c

(PyC5H6NO)H+ (PyC4H4O2)H+ PyC3H4O+ PyCdO+ or (PyC2H3)H+ (C5H5N)H+

norcotinine

CdO -c C4H5NO C 5H 5N

PyC3H4O+ PyC3H4+ (C5H5N)H+ (C4H6NO)+

134 (18) 118 (18) 80 (100) 84 (15)

a Fragments with a relative abundance of g15% are included. These are proposed compositions; Py is pyridyl. c Multiple fragments lost. d The collision energy was 27 eV for all MS/MS analyses of cotinine metabolites; daughter spectra for 5′-hydrocotinine were also obtained at 20 eV.

b

Cotinine P450 2A6 metabolites were generated from a reaction of 200 µM cotinine. The three peaks were collected via HPLC (system II) and analyzed by LC/MS and LC/MS/MS. The major ion present in peak 2 and in peak 3 when analyzed by LC/MS was m/z 193. This is the protonated molecular ion of any hydroxylated cotinine metabolite. Mass spectra from peaks 2 and 3 were characteristic of 3′-hydroxy- and 5′-hydroxycotinine, respectively. Daughter spectra obtained from CID of m/z 193 from these two peaks are presented in Figure 3. These spectra are identical to the standards (Table 1). We were unable to obtain any mass spectral data for collected peak 1. This could be due to the instability of this metabolite, which was suggested by our sodium borohydride experiments. Therefore, the cotinine metabolites of P450 2A6 were analyzed directly by LC/MS. Full spectral data were obtained for both peaks 2 and 3. The spectra were identical to those of the trans-3′hydroxycotinine and 5′-hydroxycotinine standards (data not shown). Again, we were unable to obtain any spectral data for peak 1. However, if the sample was analyzed by LC/MS, with select ion monitoring (SIM), immediately following termination of the reaction a major ion of m/z 193 was detected. This is illustrated by the chromatogram in Figure 4A. Three peaks were detected with SIM of m/z 193. The retention times of these peaks are consistent with them being peaks 1-3 in Figure 2B, respectively. The first peak again eluted very close to cotinine N-oxide, and the latter two coeluted with 3′- and 5′-hydroxycotinine, respectively. To confirm that the m/z 193 ion detected at 26.4 min (Figure 4A) is not due to the presence of cotinine N-oxide, LS/MS/MS analysis with CID of m/z 193 was carried out on the same sample. Cotinine N-oxide has a very characteristic pair of daughter fragments, m/z 98 and 96 (Table 1). Our limit of detection with LC/MS/MS was such that we would obtain good product ion spectra of cotinine N-oxide if one tenth of the m/z 193 ion detected by SIM was the molecular ion of the N-oxide metabolite. However, no product ion spectra were detected for the 26.4

Figure 3. Product ion spectra of m/z 193 for P450 2A6generated cotinine metabolites, 2 (A), and 3 (B). Metabolites were collected from HPLC at the appropriate retention time, concentrated, and reanalyzed by LC/MS/MS. The daughter spectra obtained were identical to standards of trans-3′-hydroxycotinine (A) and 5′-hydroxycotinine (B). Collision energies for MS/MS were -27 and -20 eV, respectively.

min peak. Product ion spectra of the peaks at 37.8 and 43.4 min were obtained and again were identical to standards analyzed under the same conditions. The data presented so far for peak 1 suggest it is not a stable metabolite, that it is partially reduced to cotinine by sodium borohydride, and that it may have a protonated molecular ion of m/z 193. This molecular ion is consistent with it being a hydroxylated cotinine metabolite. A candidate metabolite is N-(hydroxymethyl)norcotinine (Scheme 1). This metabolite could partially reduce to cotinine since it would be in equilibrium with the corresponding imminium ion and would decompose by deformylation, resulting in the loss of the tritium label and the formation of norcotinine, m/z 162. The retention time of norcotinine is 1 min past that of 5′-hydroxycotinine (Figure 2A). When the P450 2A6 cotinine metabolites were analyzed by LC/MS with SIM of m/z 163, (M + H)+, one major peak was detected at 45 min (Figure 4B). This peak coelutes with norcotinine and was not present in control incubations. The broad peak at 54 min coelutes with cotinine. The same sample was analyzed

Cotinine Metabolism by P450 2A6

Figure 4. LC/MS analysis of the products of P450 2A6catalyzed metabolism of 200 µM cotinine. Detection is by positive ion select ion monitoring for m/z 193 (A) and 163 (B). Details of the reaction and LC/MS conditions are provided in Experimental Procedures.

Figure 5. Product ion spectra of m/z 163 obtained from the peak at 45 min in Figure 4B. Analysis was by LC/MS/MS of the complete incubation mixture, and details are provided in Experimental Procedures.

by LC/MS/MS for product ion spectra of m/z 163. A daughter spectrum identical to that from the norcotinine standard was obtained for the peak at 45 min (Figure 5 and Table 1). Therefore, norcotinine is another product of the P450 2A6-catalyzed metabolism of cotinine, and may be the product of the unstable metabolite, peak 1. Norcotinine would not have been detected in the experi-

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 643

Figure 6. Ion exchange HPLC analysis of the products of (A) the hamster liver microsomal metabolism of cotinine and (B) peak 1 (Figure 2B) from the P450 2A6 metabolism of cotinine. Abbreviations are given in the legend of Figure 2. The labeled peaks in panel A all coelute with the indicated standard. The identities of 3′-OH cotinine, 5′-OH cotinine, and cotinine N-oxide were confirmed by LC/MS/MS. The arrow in panel B denotes the retention time of the cotinine N-oxide standard co-injected with the P450 2A6 3H-labeled metabolite.

ments with [3H-me]cotinine since the tritium is in the methyl group. Comparison of Peak 1 to Hamster Liver Microsome-Generated Cotinine Metabolites. Hamster liver microsomes have been reported to metabolize cotinine to N-(hydroxymethyl)norcotinine (18). Therefore, we attempted to generate N-(hydroxymethyl)norcotinine with hamster liver microsomes. The products of cotinine metabolism by hamster liver microsomes were analyzed by cation exchange HPLC (system III) with detection by UV, A260 (Figure 6A). Five metabolite peaks were detected, eluting at 31, 36, 43, 47, and 50 min. The first three of these peaks were identified as 5′-hydroxycotinine, trans-3′-hydroxycotinine, and cotinine N-oxide, respectively. Each peak coeluted with the standard in both HPLC system II (C18 reverse phase) and system III (cation exchange). LC/MS and LC/MS/MS spectra obtained from each of these metabolites were identical to standards (data not shown). The small peak at 58 min coeluted with norcotinine; this peak was not further characterized, but norcotinine was detected by LC/MS analysis of the total hamster liver microsome metabolites. Four products of cotinine metabolism by HLM, trans-3′hydroxycotinine, 5′-hydroxycotinine, cotinine N-oxide, and N-(hydroxymethyl)nornicotine, were previously identified by Li and Gorrod (18). However, we detected five metabolites and were unable to confirm that any of them were N-(hydroxymethyl)nornicotine. The hamster cotinine metabolite peak eluting at 50 min (unknown Y, Figure 6A) coeluted with 5′-hydroxycotinine when reanalyzed on a C18 column (system II) but not on system III (cation exchange). The peak eluting at 47 min (unknown X, Figure 6A) when collected and reanalyzed by HPLC on a C18 column (system II) eluted

644 Chem. Res. Toxicol., Vol. 12, No. 7, 1999

0.5 min after cotinine N-oxide, with the same retention time as the unknown P450 2A6 metabolite, peak 1. The reverse was also true. That is, when the P450 2A6 metabolite, peak 1, was collected from HPLC system II and reanalyzed on a cation exchange column (HPLC system III), the collected 3H-labeled metabolite eluted past cotinine N-oxide and appeared to coelute with metabolite X from the hamster liver microsomes (Figure 6B). Several attempts to obtain spectral data to further characterize the hamster liver microsome metabolites X and Y were unsuccessful. However, LC/MS and LC/MS/ MS data collected for both metabolites were in part characteristic of N-(hydroxymethyl)nornicotine. When peak X was collected from a cation exchange column and analyzed by LC/MS and LC/MS/MS, no spectral data were obtained at the correct retention time, but norcotinine was detected. As noted above, this would be a predicted breakdown product of N-(hydroxymethyl)nornicotine. When peak Y was analyzed in the same manner, a molecular ion of m/z 193 was detected by LC/MS, but we were unable to obtain any product ion spectra of m/z 193. Yet, we did obtain product ion spectra of m/z 163. This spectrum was characteristic of norcotinine. However, the retention time of this metabolite is not that of norcotinine. In summary, we could not identify either of these metabolites as the N-(hydroxymethyl)nornicotine identified by Li and Gorrod (18). However, one of these hamster liver microsomal metabolites behaves chromatographically like the unidentified product of P450 2A6catalyzed cotinine metabolism.

Discussion We report here the characterization of three P450 2A6generated cotinine metabolites, trans-3′-hydroxycotinine, 5′-hydroxycotinine, and norcotinine. In addition, we detected a fourth as yet unidentified product, which may be N-(hydroxymethyl)norcotinine. Previously, it was reported that P450 2A6 metabolized cotinine to trans3′-hydroxycotinine (9). The authors of that study did not detect any other cotinine metabolites. We did not quantify the amount of norcotinine that was formed. However, the unidentified metabolite and 5′-hydroxycotinine were formed in amounts similar to the amount of trans-3′hydroxycotinine. These metabolites have all been detected previously in the urine of cotinine-treated animals (1, 14-17). None of the animal studies were quantitative, and only trans-3′-hydroxycotinine has been reported to be a major metabolite of nicotine in animals or people (13, 23). Li and Gorrod reported and we confirmed in the study presented here that 5′-hydroxycotinine as well as trans3′-hydroxycotinine is a major metabolite of the hamster liver microsomal metabolism of cotinine. trans-3′-Hydroxycotinine accounted for about 50% of the urinary metabolites of nicotine in the hamster but less than 20% in the rat (23). The hamster, but not the rat, has a hepatic P450, 2A8, that is closely related to P450 2A6 (24, 25). Whether P450 2A8 is the enzyme catalyzing the majority of cotinine metabolism in hamster liver microsomes is unknown, but hamster liver microsomes like P450 2A6 catalyze both the 3′- and 5′-hydroxylation of cotinine (Figures 2B and 6A). The rat liver metabolism of cotinine has not been reported. In a detailed study by Benowitz and co-workers, it was estimated that 70% of the dose of nicotine received by

Murphy et al.

smokers and nicotine patch users was metabolized to cotinine (2). The authors measured the total urinary excretion of cotinine, and cotinine glucuronide, trans-3′hydroxycotinine, and its glucuronide as well as cotinine N-oxide and norcotinine. Norcotinine accounted for about 2% of the nicotine dose, while trans-3′-hydroxycotinine and its glucuronide accounted for more than 45% of the dose. Neither 5′-hydroxycotinine nor any metabolites of it were measured. But even in the unlikely event that all the unaccounted for nicotine was metabolized to 5′hydroxycotinine, the 5′-hydroxylation of cotinine would still only occur at 1/4 the level of 3′-hydroxylation. Neurath reported the level of 5′-hydroxycotinine in smokers’ urine to be less than 4% of that of the trans-3′-hydroxycotinine (1). This is in contrast to the approximately equal rates of 3′- and 5′-cotinine hydroxylation by P450 2A6. This apparent discrepancy between the extent of 5′-hydroxylation of cotinine in vivo and in P450 2A6 metabolism in vitro raises questions about the role of P450 2A6 in cotinine metabolism in smokers. We recently completed a study on the relationship of coumarin 7-hydroxylation, a P450 2A6 specific reaction to cotinine 3′-hydroxylation in 30 nonsmokers, and reported a weak but significant correlation (26). This supports the role of P450 2A6 as an important catalyst of cotinine 3′-hydroxylation. However, as in the study discussed above, we found no evidence for significant excretion of urinary metabolites of cotinine 5-hydroxylation. More than 55% of the dose was excreted as either trans-3′-hydroxycotinine or its glucuronide and 30-40% as either cotinine or cotinine N-glucuronide. These data again appear to be inconsistent with the relative rates of cotinine 3′- and 5′-hydroxylation by P450 2A6. Possible explanations for the apparent inconsistency between cotinine metabolism in vivo and P450 2A6catalyzed cotinine metabolism include the following. (1) P450 2A6 is not the major liver P450 that catalyzes cotinine 3′-hydroxylation. (2) There is extensive extrahepatic 3′-hydroxylation of cotinine. (3) There is interconversion of cotinine metabolites. Presently, there is little data to support any of these explanations. However, Nakajima’s study on human liver microsomal metabolism of cotinine (9) and unpublished data in our laboratory do not support the first explanation. Potentially of relevance to the third explanation is a report by Castagnoli et al. (27), who found that 2-hydroxy-1-methyl-5(3-pyridyl)pyrrole is metabolized to cis-3′-hydroxycotinine in the rabbit. A tautomer of this 2-hydroxy pyrrole is ∆4′(5′)-dehydrocotinine, which is the dehydration product of 5′-hydroxycotinine. Therefore, it is possible that 5′hydroxycotinine could be converted to trans-3-hydroxycotinine in people by a pathway analogous to that proposed by Castagnoli to exist in the rabbit (27). Obviously, more detailed studies are required to confirm that hepatic P450 2A6 is the major enzyme catalyzing cotinine metabolism in smokers. While a larger body of data exists supporting the role of P450 2A6 as the catalyst of nicotine conversion to the imminium ion (5-8), the specificity of this reaction has not been investigated. Confirming the role of P450 2A6 in cotinine and nicotine metabolism in smokers is critical to our understanding of both nicotine consumption and tobacco carcinogenesis. Polymorphisms in the P450 2A6 gene exist; it has been reported that individuals heterozygous for mutant P450 2A6 genotypes are less likely to smoke, and if they do smoke they use fewer cigarettes (28). The

Cotinine Metabolism by P450 2A6

authors propose that this change in smoking behavior is due to differences in nicotine metabolism. However, whether differences in P450 2A6 activity result in a significant change in nicotine metabolism in smokers has yet to be investigated. Two tobacco-specific nitrosamines, NNN and 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), are metabolically activated by P450 2A6 (19, 21, 29). If P450 2A6 is the major catalyst of cotinine and nicotine metabolism, this may well affect the efficiency of NNN and NNK activation in smokers. Therefore, understanding the relative importance of P450 2A6 to nicotine metabolism in smokers will also contribute to our understanding of tobacco carcinogenesis.

Acknowledgment. We thank Stephen S. Hecht for numerous discussions of nicotine chemistry and the N-(hydroxymethyl)norcotinine. We also thank Shantu Amin for providing trans-3′-hydroxycotinine. All MS analyses were carried out by the Analytical Chemistry Core Facility of the University of Minnesota Cancer Center (1 P30 CA77598-01). This research was supported by the Minnesota Medical Foundation and by a Grantin-Aid from the Graduate School of the University of Minnesota.

References (1) Neurath, G. B. (1994) Aspects of the oxidative metabolism of nicotine. Clin. Invest. 72, 190-195. (2) Benowitz, N. L., Jacob, P., III, Fong, I., and Gupta, S. (1994) Nicotine metabolic profile in man: comparison of cigarette smoking and transdermal nicotine. J. Pharmacol. Exp. Ther. 268, 296-303. (3) Murphy, P. J. (1973) Enzymatic oxidation of nicotine to nicotine ∆1′(5′) iminium ion. J. Biol. Chem. 248, 2796-2800. (4) Gorrod, J. W., and Hibberd, A. R. (1982) The metabolism of nicotine-∆1′(5′)-iminium ion, in vivo and in vitro. Eur. J. Drug Metab. Pharmacokinet. 7, 293-298. (5) Flammang, A., Gelboin, H., Aoyama, T., Gonzalez, F., and McCoy, G. D. (1992) Nicotine metabolism by DNA expressed human cytochrome P-450. Biochem. Arch. 8, 1-8. (6) Cashman, J. R., Park, S. B., Yang, Z.-C., Wrighton, S. A., Jacob, P., III, and Benowitz, N. L. (1992) Metabolism of nicotine by human liver microsomes: stereoselective formation of transnicotine N′-oxide. Chem. Res. Toxicol. 5, 639-646. (7) Nakajima, M., Yamamoto, T., Nunoya, K.-I., Yokoi, T., Nagashima, K., Inoue, K., Funae, Y., Shimada, N., Kamataki, T., and Kuroiwa, Y. (1996) Role of human cytochrome P4502A6 in C-oxidation of nicotine. Drug Metab. Dispos. 24, 1212-1217. (8) Messina, E. S., Tyndale, R. F., and Sellers, E. M. (1997) A major role for CYP2A6 in nicotine C-oxidation by human liver microsomes. J. Pharmacol. Exp. Ther. 282, 1608-1614. (9) Nakajima, M., Yamamoto, T., Nunoya, K.-I., Yokoi, T., Nagashima, K., Inoue, K., Funae, Y., Shimada, N., Kamataki, T., and Kuroiwa, Y. (1996) Characterization of CYP 2A6 involved in 3′-hydroxylation of cotinine in human liver microsomes. J. Pharmacol. Exp. Ther. 277, 1010-1015. (10) Voncken, P., Rustemeier, K., and Schepers, G. (1990) Identification of cis-3′-hydroxycotinine as a urinary nicotine metabolite. Xenobiotica 20, 1353-1356. (11) Kyerematen, G. A., and Vesell, E. S. (1991) Metabolism of nicotine. Drug Metab. Rev. 23, 3-41.

Chem. Res. Toxicol., Vol. 12, No. 7, 1999 645 (12) Dagne, E., and Castagnoli, N. (1972) Structure of hydroxycotinine, a nicotine metabolite. J. Med. Chem. 15, 356-360. (13) Neurath, G. B., Dunger, M., Orth, D., and Pein, F. G. (1987) trans3′-Hydroxycotinine as a main metabolite in urine of smokers. Int. Arch. Occup. Environ. Health 59, 199-201. (14) Morselli, P. L., Ong, H. H., Bowman, E. R., and McKennis, H. (1967) Metabolism of (()-cotinine-2-14C in the rat. J. Med. Chem. 10, 1033-1036. (15) Dagne, E., Gruenke, L., and Castagnoli, N. (1974) Deuterium isotope effects in the in vivo metabolism of cotinine. J. Med. Chem. 17, 1330-1333. (16) Dagne, E., and Castagnoli, N. (1972) Cotinine N-oxide, a new metabolite of nicotine. J. Med. Chem. 15, 840-841. (17) McKennis, H., Jr., Turnbull, L. B., Bowman, E. R., and Wada, E. (1959) Demethylation of cotinine in vivo. J. Am. Chem. Soc. 81, 3951-3954. (18) Li, Y., and Gorrod, J. W. (1994) N-Hydroxymethylnorcotinine, a new primary in vitro metabolite of cotinine. Xenobiotica 24, 409415. (19) Patten, C. J., Smith, T. J., Tynes, R., Friesen, M., Lee, J., Yang, C. S., and Murphy, S. E. (1997) Evidence for cytochrome P450 2A6 and 3A4 as major catalysts for N′-nitrosonornicotine R-hydroxylation in human liver microsomes. Carcinogenesis 18, 16231628. (20) Curvall, M., Kazemi-Vala, E., and Enzell, C. (1982) Simultaneous determination of nicotine and cotinine in plasma using capillary column gas chromatography with nitrogen sensitive detection. J. Chromatogr. 232, 283-293. (21) Patten, C. J., Smith, T. J., Murphy, S. E., Wang, M.-H., Lee, J., Tynes, R. E., Koch, P., and Yang, C. S. (1996) Kinetic analysis of activation of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone by heterologously expressed human P450 enzymes, and the effect of P450 specific chemical inhibitors on this activation in human liver microsomes. Arch. Biochem. Biophys. 333, 127-138. (22) Nguyen, T. L., Dagne, E., Gruenke, L., Bhargava, H., and Castagnoli, N. (1981) The tautomeric structures of 5-hydroxycotinine, a secondary mammalian metabolite of nicotine. J. Org. Chem. 46, 758-760. (23) Nwosu, C. G., and Crooks, P. A. (1988) Species variation and stereoselectivity in the metabolism of nicotine enantiomers. Xenobiotica 18, 1361-1372. (24) Pearce, R., Greenway, D., and Parkinson, A. (1992) Species differences and interindividual variation in liver microsomal cytochrome P450 2A enzymes: effects on coumarin, dicumarol, and testosterone oxidation. Arch. Biochem. Biophys. 298, 211225. (25) Kimura, S., Kozak, C. A., and Gonzalez, F. J. (1989) Identification of a novel P450 expressed in rat lung: cDNA cloning and sequence, chromosome mapping and induction by 3-methylcholanthrene. Biochemistry 28, 3798-3803. (26) Murphy, S. E., Johnson, L. M., Losey, L. M., Pullo, D. A., and Hatsukami, D. K. (1998) Role of cytochrome P450 2A6 in cotinine metabolism in vitro and in vivo. Proc. Am. Assoc. Cancer Res. 39, 488. (27) Castagnoli, N., Jr., Liu, X., Shigenaga, M. K., Wardrop, R., and Castagnoli, K. (1998) Studies on the metabolic fate of (S)-nicotine and its pyrrolic analog β-nicotyrine. In Nicotine Safety and Toxicity (Benowitz, N. L., Ed.) pp 57-65, Oxford University Press, New York. (28) Pianezza, M. L., Sellers, E. M., and Tyndale, R. F. (1998) Nicotine metabolism defect reduces smoking. Nature 393, 750. (29) Hecht, S. S. (1998) Biochemistry, biology, and carcinogenicity of tobacco-specific N-nitrosamines. Chem. Res. Toxicol. 11, 559-603.

TX990015E