Enzymatic Oxidation of Nicotine to Nicotine A1’c5’) Iminium INTERMEDIATE IN THE METABOLISM OF NICOTINE

SUMMARY Oxidation of nicotine in the presence of rabbit liver 10,000 x g supernatant, NADPH, and cyanide results in the production of 5’-cyanonicotine. NaBDl reduction of the reaction mixture from a similar incubation of nicotine, NADPH, and supernatant results in the production of deuteronicotine having a single deuterium in the pyrrolidine ring. Oxidation of nicotine to cotinine under 1802 did not lead to incorporation of I80 in the cotinine produced. These results are explained by the formation of nicotine AlfC6’) iminium ion as an intermediate in the degradation of nicotine. The formation of the iminium ion is catalyzed by an enzyme system having the properties of a mixed function oxidase. The reaction re-quires NADPH and O2 and is inhibited by carbon monoxide. It is proposed the nicotine A1’(6’) iminium ion is formed by loss of water from 5’-hydroxy nicotine. The iminium ion formed from nicotine represents a nicotine analogue with a markedly altered chemical reactivity and must therefore be considered in studies on the physiological action of nicotine.

x g supernatant, NADPH, and cyanide results in the production of 5'-cyanonicotine.
NaBDl reduction of the reaction mixture from a similar incubation of nicotine, NADPH, and supernatant results in the production of deuteronicotine having a single deuterium in the pyrrolidine ring. Oxidation of nicotine to cotinine under 1802 did not lead to incorporation of I80 in the cotinine produced.
These results are explained by the formation of nicotine AlfC6') iminium ion as an intermediate in the degradation of nicotine. The formation of the iminium ion is catalyzed by an enzyme system having the properties of a mixed function oxidase.
The reaction requires NADPH and O2 and is inhibited by carbon monoxide. It is proposed the nicotine A1'(6') iminium ion is formed by loss of water from 5'-hydroxy nicotine.
The iminium ion formed from nicotine represents a nicotine analogue with a markedly altered chemical reactivity and must therefore be considered in studies on the physiological action of nicotine.
The metabolism of nicotine in mammals has been studied by many authors both in vitro and in viva (1). In vitro, liver enzymes catalyze the oxidation of nicotine to cotinine,l desmethylnicotine, nicotine N-oxide, hydroxycotinine and y-(3-pyridyl) y-oxo-N-methyl butyramide (2). Cotinine is the major product of nicotine metabolism in man as well as in a number of mammalian speties (1). Hucker et al. presented the first study on the enzymatic nature of the conversion of nicotine to cotinine (3). These authors proposed that nicotine is initially oxidized to 5'-hydroxy nicotine by an NADPH-dependent oxidase present in rabbit microsomes.
The production of cotinine from the aminocarbinol was presumed to be catalyzed by a soluble enzyme similar, if not identical, to aldehyde oxidase.
In support of these contentions they found that cyanide inhibited the production of cotinine but not the disappearance of nicotine.
Since aldehyde oxidase is sensitive to cyanide, while mixed function oxidases are not, the reaction was thought to be stopped at the aminocarbinol stage. Indeed, when the reaction was performed on a large scale in the presence of cyanide the authors isolated a compound with properties similar to those of the proposed aminocarbinol.
In order to more clearly define the enzymatic steps involved in the production of cotinine from nicotine we have re-examined the oxidation of nicotine by rabbit liver preparations in vitro. When nicotine was incubated with reconstituted rabbit liver 10,000 x g supernatant in the presence of an NADPH generating system and 0.01 M KCN, the major product isolated was identified as 5'-cyanonicotine. It is postulated that this compound is produced by the reaction of the nicotine Al'(5') iminium ion with cyanide.2 MATERIALS All cofactors were purchased from Boehringer Mannheim. Nicotine was obtained from Eastman Kodak Co., Rochester, New York.
1802 was purchased from Miles Laboratories, Inc., Elkhart, Indiana. NaBD4 was a product of Alfa Inorganics, Beverly, Massachusetts.

In Vitro IncubalionsLivers
were obtained from rabbits that were pretreated with phenobarbital for 1 week prior to sacrifice (0.2 mg per ml in drinking water). Rabbit liver microsomes were prepared by differential centrifugation of 20~~ liver homogenates prepared in 0.25 RI sucrose (4). The 100,000 x g pellet was resuspended in the original volume of 0.05 M phosphate buffer pH 7.4. Reconstituted 10,000 x g supernatant was prepared by the addition of 1 ml of 100,000 x g supernatant fraction to 2 ml of resuspended microsomes.
Incubations were performed at 3i" for a maximum of 30 min using 1 mM nicotine as the substrate.
The reaction rates were linear with respect to cotinine or nicotine Al'(5') iminium ion formation during this time period.
The reaction rates were also linear with respect to enzyme concentration throughout the concentration range used in these studies. Exlraction Procedures-(u) In the initial experiments, extractions were performed essentially as described by Hucker et al. (3). Incubations were transferred to 30 ml of 1.55; isoamyl alcoholheptane.
The samples were mixed for 30 min on a rotary mixer, and then the heptane layer was drawn ofi.
A 15.ml portion of the heptane layer was then mixed with 10 2 A preliminary account of this work was presented in the Proceedings of the Second International Symposium on Microsomes and Drug Oxidation, July 28-30, Stanford, California, 1972. ml of 0.1 N HCl.
The HCl layer was separated and adjusted to pH 9 with Na?COz. The basic solution was extracted with 10 ml of CHC& and this extract was concentrated to dryness. The extracts were redissolved in 100 ~1 of CH&12 for gas chromatographic analysis.
(b) After the structure of cyanonicotine was determined the extraction procedure was simplified.
Incubations (2 to 3 ml) were transferred to 5 ml of CH&12 and mixed with a rotary vibrator.
The CH&l? extract was drawn off and a second 5-ml portion of CH&lZ was added to the aqueous layer.
After mixing, the CH#& layer was drawn off and combined with the first extract.
The combined extracts were dried over MgS04 and then concentrated to dryness and redissolved in 100 ~1 of CHzClz for gas chromatographic analysis. Analyses-Gas chromatography was performed on a Hewlett-Packard 5750B gas chromatograph using a flame ionization detector.
A 4-foot column was packed with 3.0% UC-W98 silicone gum rubber (methyl vinyl) on diatoport S. The flame detector and flash heater were maintained at 230". The helium flow rate was 60 ml per min. The column temperature at the time of injection was 100". After holding the temperature at 100" for 2 min. the temperature was raised to 200" at a rate of 10" per min. The retention times of nicotine, cyanonicotine, and cotinine under these conditions are 3.6, 7.6, and 8.4 min, respectively.
Synthesis 0s 5'-Cyanonicotine-Nicotine was oxidized in the presence of mercuric acetate and EDTA according to the method described by Knabe (5). After removal of excess mercury with hydrogen sulfide the pH was adjusted to 7.0 by the addition of 5 N KOH.
An amount of KCN was added equimolar to the amount of nicotine used in the reaction.
The solution was stirred for 1 hour and then extracted with CH&12.
The extract was washed with 0.1 N HCl and then dried over Na2S04. After concentration the residual oil was distilled in VUCUO. The infrared spectrum had an absorption at 2450 cm+ characteristic of the nitrile moiety.
The mass spectrum was obtained on an LKB 9000 GC-MS combination.
The mass ion was at 187 with prominent peaks at 160, corresponding to the loss of HCN, and at 109, corresponding to the N-methyl cyanopyrrolidine fragment.
The 100 MHz nuclear magnetic resonance spectrum contained 4 multiplets in the region between 3.0 and 4.5 ppm. Since none of the multiplets appear to be coupled to one another, the cyano group is assigned to position 5' rather than position 2'. RESULTS

AND DISCUSSION
Gas chromatographic analysis of the extracts obtained after incubation of nicotine with reconstituted rabbit liver 10,000 X g supernatant fraction in the presence of 0.01 M cyanide indicated the presence of a product having a gas chromatographic retention time of 7.6 min. Mass spectrometric analysis of this product indicated that the molecular weight was 187 (Fig. 1). This corresponds to the displacement of a proton of nicotine by a cyan0 group.
The fragmentation pattern indicated that the cyano group was present in the pyrrolidine ring.
Chemical oxidation of nicotine followed by reaction with cyanide yielded 5'-cyanonicotine.
Comparison of the mass spectra of the chemical product and the incubation product indicated that the two compounds were identical.
There are two likely explanations for the production of cyanonicotine during the enzymatic oxidation of nicotine and in the presence of cyanide.
These are shown in Equations 1 and 2.
In order to distinguish between these two mechanisms, an experiment was performed wherein the nicotine-microsomal reaction mixture was treated with NaBD4.
If the reaction proceeds through a stable iminium ion it should be reduced by the NaBD4 and yield deuteronicotine.
The mass spectrum of nicotine isolated from NaBD4 treated reaction mixture is shown in Fig. 2A.
A comparison of this spectrum with the mass spec-60 FIG. 1 (left). Mass spectrum of 5'-cyanonicotine. The spectrum was obtained using the extract from an incubation of 1 rnM nicotine with reconstituted rabbit liver 10,000 X g supernatant in the presence of 0.5 ml% NADPH and 0.01 M KCN.
Analysis was performed with an LKB-9000 GC-MS combination using a 3'$& UC-W-98 column as described under "Methods." FIG. 2 (cenler und right). A, mass spectrum of nicotine obtained after treatment of a nicotine incubation mixture with NaBD4. Nicotine (1 mM) was incubated with rabbit liver microsomes in the presence of 0.5 mM NADPH for 20 min at 37". The reaction mixture was extracted as described in Extraction Procedure b of "Methods" section in order to remove unchanged nicotine. The aqueous layer was then treated with 1 mM NaBD, and incubated at 37" for 30 min. The reaction mixture was re-extracted with CH#&.
The CH&lz extract was concentrated and analyzed on an LKB-9000 CC-MS combination using a 3% UC-W98 column as described under "Methods." B, mass spectrum of nicotine obtained from a control incubation of nicotine, NADPH, and rabbit liver microsomes. The conditions used were identical with those described in the legend of A with the exception that NaBD(  Fig. 2B clearly shows that NaBD4 reduction resulted in the production of deuteronicotine.
The mass ion at 163 shows the presence of 1 deuterium in the nicotine molecule, while the fragment at 85 is attributed to the pyrrolidine ring containing a single deuterium atom.
This result indicates the presence of the nicotine iminium ion as a discrete entity in the enzymatic reaction medium.
In light of the known chemical reactivity of iminium salts (6) to nucleophiles such as cyanide it seems likely that the cyanonicotine also was produced from the nicotine iminium intermediate.
We have, therefore, used the production of cyanonicotine as a measure of the oxidation of nicotine to nicotine A1'c5') iminium ion.
The effects of various cofactors on the production of nicotine iminium ion are shown in Table I. NADPH is the preferred cofactor for the reaction.
NADH is approximately one-third as effective, while NAT>+ or NhDP+ is completely devoid of activity.
The effects of changes in oxygen concentration and carbon monoxide on the oxidation of nicotine to either nicotine A1'(5') iminium ion or to cotinine are shown in Table II. There is an absolute requirement for O2 for either cotinine or iminium ion production.
Carbon monoxide inhibits both reactions but is a more effective inhibitor of the production of cotinine. This may be due to the fact that there are tuTo oxidative steps involved in the production of cotinine and only one for the production of the nicotine iminium ion.
The sensitivity to CO and the requirements for 02 and NADPH all indicate that the production of nicotine iminium ion is catalyzed by a mixed function osidase. The most likely pathway for the production of this intermediate is via the formation of 5'-hydroxynicotine. 5 If the equilibrium rates are rapid relative to the irreversible steps designated kj, kp, and ks in Equation 3, then it should be possible to increase the rate of formation of one product at the expense of the other.
This seems to be the case when cyanide is added to a microsomal incubation containing nicotine and NADPH. Fig. 3 shows the effect of cyanide concentration on cotinine and cyanonicotine format'ion. The decline in cotinine formation parallels the increase in cyanonicotine formation.
This supports the hypothesis that they share a common intermediate.
Mixed function oxidases are characterized by the incorpora-   Incubations were performed at 37" for 20 rnin rlsing 2 ml of resuspended rabbit liver microsomes and 1 ml of rabbit liver 100,000 X g supernatant.
Final volume, 3.5 ml. Nicotine concentration was 1 mM. Incubations were carried out in Warburg flasks with the cofactor and substrate in the side arm. After 10 min of equilibration with the gas phase, the reactions Lvere started by the addition of the cofactor and substrate to the center wells. The gas mixture was allowed to flow through the flask throughout the reaction. tion of 1 atom of molecular oxygen into the oxidized product (7). In the case of nicotine the oxygen of 5'.hydrosynicotine would come from OZ. If there is no exchange of this osygen with water it should be present in the product cotinine.
On the other hand if 5'-hydroxynicotine is rapidly equilibrating with either of the two intermediates shown in Equation 3, the osygen in cotinine should be derived mainly from HzO.
The production of cotinine from nicotine was examined in an incubation performed under an atmosphere of 'So?. Cotinine was isolated and the mass spectrum analyzed for l80. The re- Incubations were performed for 20 min at 37" using 2 ml of a 20$& suspension of rabbit liver microsomes and 1 ml of rabbit liver 100,000 X g supernatant fraction.
Nicotine concentration was 1 mM. Extraction and analyses were performed as described under "Methods." sults are shown in Table III.
There was no incorporation of 180 into cotinine produced in an atmosphere of '4%~. Therefore, there must be an exchange of the oxygen of 5-hydroxy nicotine prior to the dehydrogenation step. In summary, evidence has been obtained that nicotine A1'c5') iminium ion is a discrete intermediate in the oxidation of nicotine in vitro by liver microsomal enzymes. This intermediate can be trapped by reaction with cyanide and is stable enough to be reduced to deuteronicotine by NaBD4. The formation of nicotine Al'(5') iminium ion is catalyzed by an enzyme system having the characteristics of a mixed-function oxidase.
The initial product of the oxidase reaction is probably 5'-hydroxynicotine.

2799
Although it is not possible to rule out formation of nicotine Al'@') iminium ion via direct dehydrogenation, the properties of the enzymatic reaction make this pathway less likely.
The fact that the reaction is inhibited by carbon monoxide and relatively insensitive to cyanide indicates that the enzyme is not similar to the fatty acid desaturase found in rat liver microsomes (8) * The formation of a carbinolamine during the oxidation of substituted amines has been postulated for a wide variety of dealkylation reactions (9). Carbinolamines can break down to give an aldehyde or ketone plus an amine or they can lose water to form an imine or iminium ion. While there are many examples of the former pathway, there are only a limited number of examples of the latter due, presumably, to the instability of the intermediate.
An imine has recently been proposed as an intermediate in the oxidation of amphetamine to amphetamine oxime (10, 11). The instability of this imine has prevented its direct isolation.
Breck and Trager (12) have proposed the formation of an imine during the metabolism of lidocaine. 3 The imine, which would normally break down to acetaldehyde and an amine, is thought to react intramolecularly to form a cyclic addition compound.
The stability of imines and iminium ions is increased in heterocyclic systems (13). Five and six member ring systems containing an imine or iminium moiety can be readily prepared chemically.
A stable cyclic imine has been isolated as a metabolite of medazepam4 by Schwartz and Kolis (14). These authors were also able to isolate 2-hydroxy medazepam proving the existence of a carbinolamine intermediate in the metabolism of medazepam.
The formation of an iminium ion during the oxidation of nicotine in vitro raises important questions concerning its role in vivo. The iminium ion has a chemical reactivity towards nucleophiles not present in the parent nicotine.
At the same time it carries a positive charge and thereby could presumably bind at nicotine receptor sites. It has been reported that a metabolite of nicotine can block the lethal effects of nicotine (15). These authors postulated that the aminocarbinol, 5'-hydroxynicotine, might perform this function.
The iminium ion must also be considered as a possible blocker of nicotine receptor sites. Studies on the physiological properties of the nicotine iminium ion are currently in the progress.