Tobacco Documents Online

9 . ir 19 x r ~ H H 12 !A0 Scheme= I. Note that rotation of one ring with respect', to the other effectively interchanges the spatial orientation of H2 and H4 with ,respect to the pyrrolidine ring. the resonances of H2 or H4 when the N'-methyl group resonance was irradiated in the mixture of 11-12, though significant enhancements were observed for the resonances of H2' and Hs'P. The lack of enhancements at the resoi5antes of' H2 and H4 implies that the N'-methyl group is trans to the pyridine`ririg. The observed enhancements at the resonances of H2' and HS', could only be obtained if the N'-methyt group were cis to H2, and HS,S, consistent with the methyl group being trans (11) to the pyridine ring.25

10 It is well known that protonation of amines like nicotine with strong acids is diffusion controlled and thus faster than nitrogen inversion. Accord- ingly, the observation that 11 is the major diprotonated isomer implies that 8 is the major free base invertomer. In fact, barring the well-documentated experimental difficulties associated with kinetic quenching of amines, the ratio 11:12 is equal to the ratio 8:7. 36-37 The former ratio could be obtained by assigning specific resonances for both 11 and 12; e.g., for the res- onances of H2, and N'-CH3. We obtained additional evidence by preparing nicotine-3',3',4',4',5',5'-ds and subjecting it to the kinetic quenching proce- I dures discussed above. These results are illustrated in Figures 2A and 2B and Table II. On this basis, we concluded that the major isomer, 8, was present to an extent >90% in nicotine.25 B. . Pyrrolidine Ring Conformation ' While pyrrolidine ring geometries are available for nicotine dihydro- iodide28 and nicotine:salicylic acid complex29 from their respective x-ray analyses, this 'information may not be transferable to solution phase conforma- tions and may be somewhat unreliable due to the errors associated with locating hydrogen atoms by x-ray methods, especially for large compounds containing halogen atoms. During the last twenty years, conformational information has become available by evaluation of NMR proton-proton coupling data using Karplus-type relationships. Pitner, et al. reported in 1978 the analysis of the seven spin (seven attached protons) pyrrolidine ring system of nicotine.38'39 By obtaining NMR spectra for a number of specifically deuteratod -nicotines, they were able to assign the chemical shifts and rel- evant codpling constants for nicotine itself. From these evaluations, they conciuded'. that nicotine's pyrrolidine ring exists in an envelope conformation with H2, and Hs,o protons in a pseudo axial orientation and the 3' and 4' .

11 protons eclipsed, as indicated in Fig. 3.38 There are at least three reserva- tions that tag along with these conclusions: (a) the Karplus equations are empirical and depend on, the choice of coefficients used; (b) the treatment assumes that there are not a number of equally stable minimum energy conformations on a rather flat ground state potential energy surface; and (c) there are no significant bond angle distortions in the pyrrolidine ring which affect the validity of the Karplus equations. . Whidby, Edwards, and Pitner subsequently reported similar analyses of two nicotine analogues, 2-isonicotine (15) and 4-isonicotine (16) using deute- J rated derivatives to obtain chemical shifts and coupling constants.39 Interest- ingly, although 1, 15 and 16 have different physical and chemical properties, their vicinal and long-range coupling constants indicated that they have 39 . virtually identical conformations in CDCI3 solution. 15 4VI 16 01P Recently, we have obtained 1H NMR spectra for a series of additional nicotine analogues at high field (300 and 360 MHz) which resulted in separate resolved resonances for each of the pyrrolidine ring protons. Analyses of these spectra have led to information regarding the effect of substitutents on pyrrolidine.ring conformations, given the same three reservations mentioned above.44• In addition, for nicotine itself, this work has resulted in chemical shifts and coupling constants identical to those reported by Pitner et al.;38 thereby mutually confirming each other's experimental techniques.

12 There have been additional investigations directed at discerning informa- tion regarding nicotine's pyrrolidine ring conformation. Ohashi, Morishima and Yonezawa have reported a lanthanide induced shift (LIS) study of nic- otine,41 and Castagnoli and Cushman have disclosed analogous LIS exper- iments on trans-3'-methylnicotine (17).42 In both cases, significantly greater shifts were observed for the pyridyl ring protons rather than for the pyrrol- idine ring protons, even though in both 1 and 17, the pyrrolidine nitrogen is significantly more basic than the pyridine nitrogen. Since the magnitudes of the LIS are dependent on the distance between the lanthanide atom and I the protons, the lanthanide atom in both cases must be complexing with the respective pyridine nitrogen. Steric inhibition to complexation at the more basic site thus foils the use of this technique to the application under discus- sion in . this section. Interestingly, we examined the ~ LIS of the related N-methylpyridinium iodide (18) and found that complexation occurs at the quaternary N-methylpyridinium iodide site.43 This curious result established V that the countierions.,of quaternary ammonium salts are powerful *LIS donors, stronger than for,; amines like N,N-dimethyldodecylamine and N-methyldodecyl- amine. N UJ 'N l _ +p 1 CH3 CHA I I ~H 3 18 Ov. C. - Pyridine-Pyrrolidine Ring-Ring Orientations As part of their careful analysis of the high resolution specfrum of nicotine, Pitner, et al. observed a small but finite (<0.05 Hz): long-range

13 coupling between H2, and H6; they also observed an NOE of 9 ± 2 and 5 ± 2% for H2, upon saturation of the resonances of H2 and H4, respectively.38 On the basis of these results, they suggested that the pyridine-pyrrolidine ring-ring orientation was perpendicular, i.e., dihedral angle 8=t (H2'C2'CgC2) is 0° and/or 180° (see Fig. 3).38 Subsequently, Pitner, Whidby and Edwards determined the 13C NMR spin-lattice relaxation times (T1's) of nicotine and analyzed these in terms of anisotropic rotational diffusion constants.44 A number of very interesting conclusions were made, one being that inter-ring rotation is slower than overall molecular tumbling in chloroform solution, I since the experimental relaxation times for C2 and C4 differed by more than could otherwise be accounted for. By assuming a fixed geometry for the pyridine and pyrrolidine rings as determined by previous results, they calculated the isC T1's for'different values of t(H2iC2,C3C2). A comparison of Tl,calcd with T1 obsd for all of these dihedral angles indicated that the best fit obtained is for 13 (8=0°), and this fit was far superior to that .~ obtained for 14 (0 =180°). On this basis, they concluded that "the most probable ring-ring orientation is . . . 9=001 (13, c.f. Figure 3).44 Of course', conformational preferences are not "all or nothing" as described previously in the introduction to Section IV. A number of theoret- ical studies have reported the potential energy of nicotine as a function of t (H2iC2,CgC2). Invariably, minima are found in the region of 0° and 180°, extending, as part of energy wells, some 20° on each side.22,30-34 The most recent calculations, those of Dwyer using the INDO algorithm22 and of Lee and_ Park using both ETH and CNDO,34 indicate that these two minima are of nearty equal energy. Given the reliability of these algorithms, we can only conclude that nicotine is likely to be significantly populated in both forms, 13 and 14, shown in Figure 3. This conclusion is not inconsistent

14 with the Pitner, et al. results44 as they "had no feeling for how sensitive the fit [of Tiobsd vs. Tl, calcd] is to population distribution.~,45 We recently compared the effect of substituents on the periphery of nicotine's ring systems on the energy as a function of dihedral angle.22 Using the INDO algorithm, we observed that a methyl substituent at C5, Cr, or C3i, in 5-methyl, 6-methyl, and trans-3'-methylnicotine (5, 6, and 17) respectively had little effect on the overall shape or energetics of rotation about C3-C2i. However, a methyl group at either C2, C4 or C3,a in 2-methyl, 4-methyl and cis-3'-methylnicotine (4, 3 and 19) led to significant changes in - . "'.. . rotational barriers though the energy minima remained essentially at ca. 0° and 180°. See Table Ill. Examination of molecular models indicates that the cross-ring interaction of a C2-or a C4-methyl group on the pyrrolidine ring in 3 and 4 is very similar to the cross-ring interactiois of the C3ia methyl group on the pyridine ring in 19. The results in Table Ill are probably valid as trends in this closely related series, but one cannot push their quantitative capabilities, given the approximate nature of the INDO method and its difficulty; in treating tertiary amines. I CH= 19 VV We-await the development of force fields which incorporate parameters suitabCe -for- both pyridine nitrogen atoms and pyrrolidine nitrogen atoms so that more reliable calculational results may be available using molecular mechanics procedures. Also unquestionably valuable but undoubtedly "CPU

15 expensive" would be the use of ab initio calculations for these rather large molecules. V. THE CHEMICAL REACTIVITY OF NICOTINE AND NICOTINE ANALOGUES. The Development of a Chemical Reactivity Model for Nicotine Pharmacolog- ical Activity. There is a tremendous literature in the area of quantitative structure- activity relationships (QSAR), where the logical and intuitive skills of sci- entists of vastly different backgrounds have made monumental contribu- 46-49 J tions. We have drawn heavily on the results and suggestions of this vast on-going area of research for our own studies. A popular first level of sophistication of QSAR is the utilization of standard parameters, e.g., Hammett- v and Taft E, ,, combined with molecular properties, e.g., pK, and partition coefficient, to derive relationships between -pharmacological prop- erties of compounds and structural/molecular descriptors. We felt it was of primary importance to characterize the "essence" of the nicotine analogues which we were interested in. To do this, we were aided by the pharmacological data in hand and the simplicity of the nicotine molecule. It was clear that the two nitrogen atoms play a central role in the pharmacological properties of the analogues, presumably by interacting in some fashion with nicotine receptor(s). We decided to develop a chemical reaction which would serve as a model for the interaction of these nicotine analogues with their receptors. This model' had to possess the followingg featu res !- - 1. - The reaction should take place exclusively at the nitrogen atoms by the same well-known mechanism for all'compounds studied.

16 2. The reaction should be uncomplicated by side reactions, and it should be irreversible. 3. The products should be stable and clearly distinguishable from the starting materials. 4. The reaction should be readily quantifiable by standard, reliable kinetics procedures. 5. For meaningful comparispns, the reaction rates for the different analogues should vary over a wide range of reactivity. Based on these considerations, we chose to study the iodomethylation of . these nicotinoids. The Menschutkin reaction is one of the most well studied in all of organic chemistry, and it has provided an experimental basis for the understanding of many controlling factors in chemical reactivity.50,51 We.were not, of course, the first to be interested in the alkylation of nicotine. v. Planta and Kekule12 in 1853 and von Stahlschmidt52 in 1854 reported the alkylation of nicotine with excess iodoethane and iodomethane and obtained nicotine._diethiodide (20) and nicotine dimethiodide (21) respec- tively (eq. 1). /These studies were aimed at determining the number and type of nitrogen atoms in nicotine with structure determination as a goal. N R

17 Many years later, Pictet and Genequand58 in 1897 reported the prepara- tion of the two monomethiodides of nicotine, N-methylnicotinium iodide (18) and N'-methylnicotinium iodide (22). This latter work proved to be of significant interest to many subsequent investigators,* not only because 18 N + " I I- r, .~ r1 i Al-.6 l \_J l N CH ~ ~ CH3 CH3 3 CH3 22 VV 11 and 22 were of particular. value to pharmacologists given that they served as models for the monoR.rotonated forms of nicotine, but also because it proved extremely difficult to •repeat the original experimental procedures.54 We25 repeated the 1897 experimental conditions52 and subjected the crude reaction mixture;; of nicotine and iodomethane (eq 2) in either methanol or acetonitrile to iH NMR analysis. The resultant NMR spectrum from a typical experiment is shown in Figure 4A. Note that in this experiment, less than one equiv- alent (0.75 equiv. ) of iodomethane was used in order to minimize the forma- tion of the dialkylation product, nicotine-N, N'-dimethiodide (21). Clearly, *Based on the high activity of N'-methylnicotinium iodide, it is currently believed that monoprotonated nicotine is the pharmacologically active species. This is not a trivial issue, given that nicotine is 81% monoprotonated at physiological pH of 7.4. Since protonationr-deprotonation is rapid~ at this pH, considerable concentrations (effectively 100%) of both nicotine free base or monoprotonated nicotine are available.

18 too many methyl singlets are observed. We showed that the product consists of ca. 2.5:1 mixture of 22 + 18 and not of only 22 as previously reported. Because of fortuitous solubility properties (nicotine is soluble in water, ether, and chloroform; N-methylnicotinium iodide is soluble in water and chloroform but insoluble in ether; and' N'-methylnicotinium iodide is soluble in water but insoluble in both ether and chloroform), the preparation of purified quantities of both monomethiodides turns out to be a rather simple . experimental task -- once it is recognized that the product is a mixture of the two compounds.25 The formation of N-methylnicotinium iodide (18) in this alkylation is quite remarkable, since the pyridine nitrogen of nicotine is some three orders of magnitude less basic than its pyrrolidine nitrogen (Table IV). That the rate of the Menschutkin reaction is significantly dependent on nitrogen basicity can be shown by the fact that methylation of the nicotine analogue, N,N-dimethyl-3-aminomethylpyridine (23) was found to yield only one methiodide' 24, to_ the exclusion of 25 (eq 3).25 As shown in Table IV, the pKa's of nicotine and 23 are nearly identical, supporting our suggestion that steric factors associated with nicotine's pyrrolidine ring decrase the nucleophilicity of its pyrrolidine nitrogen.25 + -NM C1i e2 2 CFI2 -NMeZ _ (nrCH2 NMe3 UnT 23 24 Me 25 ~ Alsis listed in Table I V are the pKa's for a number of compounds which are found as substructures of nicotine. Note that the pKa of the pyrrolidine nitrogen of nicotine is substantially smaller than that of N-methylpyrrolidine