Alkyl Analogues of Methylphenidate (Ritalin) Salts: Avoidance of

Oct 5, 2011 - ABSTRACT: Alkyl analogues of methylphenidate (Ritalin) salts are slow onset, long duration dopamine reuptake inhibitors with a potential...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/joc

Solution- and Solid-State Conformations of C(α)-Alkyl Analogues of Methylphenidate (Ritalin) Salts: Avoidance of gauche +gauche − Interactions Avital Steinberg,† Mark Froimowitz,‡,∥ Damon A. Parrish,§ Jeffrey R. Deschamps,§ and Robert Glaser*,† †

Department of Chemistry, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel Massachusetts College of Pharmacy and Health Sciences, Boston, Massachusetts 02115, United States § Laboratory for the Structure of Matter, Naval Research Laboratory, Washington, D.C. 20375, United States ‡

S Supporting Information *

ABSTRACT: Alkyl analogues of methylphenidate (Ritalin) salts are slow onset, long duration dopamine reuptake inhibitors with a potential use as a cocaine abuse pharmacotherapy. X-ray crystallographic studies and nuclear magnetic resonance (NMR) investigations strongly suggest that avoidance of sterically unfavorable gauche −gauche + orientations effectively influences both the C(α)-alkyl side chain conformation and the formation of a predominant rotamer about the CHCH bond ligating piperidine and C(Ar)R moieties. The favored CHCH rotamer in D2O and in CD2Cl2 of the pharmacologically interesting i-Bu and CH2-cyc-Pnt (RS,RS)-salts has the same antiperiplanar arrangement that was found in the crystal structures, although there clearly is a fast equilibrium involving smaller amounts of synclinal partners. While the rotamer in the (RS,SR)-i-Bu HCl crystal structure exhibits a synclinal orientation for the vicinal pair of adjacent methine protons, the weighted time-averaged arrangement for these protons becomes almost completely antiperiplanar when the crystals are dissolved in D 2O. Increased steric congestion around the CHCH bond in the analogous N-methyl tertiary ammonium salts seems to augment the quantity of the preferred rotamer within the mixture. The stereochemistry of the species observed via NMR seems to arise from specific combinations of N-methyl orientation and avoidance of sterically unfavorable gauche −gauche + arrangements.

■

INTRODUCTION

Methylphenidate (α-phenyl-2-piperidineacetic acid methyl ester HCl, Ritalin, 1)1 is a psychostimulant approved for the treatment of a number of disorders, the most well-known of which is attention-deficit hyperactivity disorder (ADHD). The threo isomer is the pharmacologically active diastereomer responsible for its therapeutic effect of blocking the reuptake of dopamine (DA) and norephinephrine (NE) in their respective central nervous system transporters.2−4 The active (+)-enantiomer was shown to have the (R,R)-absolute configuration as depicted in iconic drawing 1.5,6 On the basis of conformational analyses of a series of potent DA reuptake blockers, including (−)-(2R,3S)-cocaine (2) (the natural enantiomer) and (−)-(2R,3S)-2-β-carbomethoxy-3-β-(4fluorophenyl)tropane (β-CFT, 3),7 a pharmacophore model was proposed for this activity in which the orientation of the ammonium N−H proton played an important role: equatorial disposition for methylphenidate and axial for the tropane analogues.6 This pharmacophore model correctly predicted the reduced activity of N-methyl analogues of methylphenidate.8 The desired axial N−H orientation for efficient DA reuptake inhibition was demonstrated by preparation of the more active bridging tropane analogue 4 versus less active 5.9,10 © 2011 American Chemical Society

Utilizing the pharmacophore model described above6 and the demonstrated activity of compounds in which the methoxycarbonyl function in cocaine (or in CFT) was replaced with simple alkyl groups,9−13 Froimowitz et al.14 synthesized an extensive series of C(α)-alkyl analogues of methylphenidate (e.g., 6−10). Compared to methylphenidate (1), many of these alkyl analogues proved to have high potency and selectivity for blocking the DA transporter relative to the NE transporter. 14 For example, the binding affinity of the 2-methylpropyl compound (i-Bu-7) at the DA transporter was 53 times higher Received: July 8, 2011 Published: October 5, 2011 9239

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

■

than that at the NE transporter, and its ability to block reuptake at the DA transporter was 14 times that measured at the NE transporter. In comparison, the binding affinity of methylphenidate (1) at the DA transporter was only 6 times stronger than that at the NE transporter, and its capability to block reuptake at the DA transporter was only 0.77 times that determined at the NE transporter. Similar to the N-methyl methylphenidates with their carbomethoxy groups,8 the Nmethyl-i-Bu alkyl analogue 8 was found to have reduced activity in this series.14 The CH2-cyc-Pnt and i-Bu groups are isosteric replacements for the methoxycarbonyl group of methylphenidate. As Froimowitz et al. have illustrated, 14 when the carboxylic CO and methoxy functions are interchanged by 180° rotation about the C(α)H−C(O)OMe bond in the crystallographically determined structure of (−)-cocaine 2, then a peripheral fragment is now common to that of corresponding atoms in i-Bu-7 (see the green cleft in Figure 1).

Article

RESULTS AND DISCUSSION Solid-State Stereochemistry of α-Alkyl Analogues of Ritalin HCl Salts. Figures 2−5 illustrate ORTEP numbering

Figure 2. ORTEP plot for the (2R,7R)-enantiomer of rac-7. Diastereotopic 8a and 8b protons are colored pink and white, respectively. Displacement ellipsoids are drawn at the 50% probability level. Characteristic torsion angles for this conformation are as follows: −52° for N(1)−C(2)−C(7)−C(8), −176° for C(2)−C(7)−C(8)− C(9), and 60° for C(7)−C(8)−C(9)−H(9).

Figure 1. Superimposition of the corresponding skeletal atoms in the crystallographic structures of both methylphenidate (R,R)-i-Bu analogue 7 (magenta atoms and black bonds; within the green cleft) and (−)-cocaine 2 (dark gray atoms and light gray bonds). The X denotes the location of a hypothetical hydrogen-bonding acceptor site in the pharmacophore model.

Figure 3. ORTEP plot for the (2R,7R)-enantiomer of axial N-methyl rac-8. Diastereotopic 8a and 8b protons are colored pink and white, respectively. Displacement ellipsoids are drawn at the 50% probability level. Characteristic torsion angles for this conformation are as follows: −57° for C(18)−N(1)−C(2)−C(7), −60° for N(1)−C(2)−C(7)− C(8), 180° for C(2)−C(7)−C(8)−C(9), and 60° for C(7)−C(8)− C(9)−H(9).

Hydrolytically labile methylphenidate (as opposed to i-Bu-7) suffers from the liability of the ester function’s rapid metabolism to the inactive acid.15 Locomotor assays of 7 showed it to exhibit the auspicious attributes of slow onset and long duration.14 There is evidence that cocaine abuse is linked to its fast onset and short duration of action. Thus, a compound with the opposite pharmacokinetic profile would be expected to have reduced abuse potential, which might allow it to be a useful substitution pharmacotherapy for the treatment of cocaine abuse. Preliminary animal studies suggest that compound 7 has indeed a lower abuse potential. This paper reports solid- and solution-state conformational studies of active (2RS,7RS)-6−8 and less active (2RS,7SR)-9 and -10 diastereomers utilizing X-ray crystallographic and NMR spectroscopic methods.

diagrams for crystal structures containing the active diastereomer (RS,RS)-i-Bu-7, its (RS,RS)-N-methyl-i-Bu-8 analogue, the inactive (RS,SR)-i-Bu-9 diastereomer, and its (RS,SR)-Nmethyl-i-Bu-10 monohydrate salt. For the sake of simplicity, the atom numbering throughout this paper is based on that arbitrarily chosen for the crystallographic numbering diagrams. The (RS,RS)-7 and -8 crystal structures discussed in this paper are representative of a larger series as are also the (RS,SR)-9 and -10 crystal structures.14 All the crystal structures in both series of Ritalin HCl α-alkyl analogues exhibited chair piperidine ring conformations with equatorially disposed CH(Ar)R1 groups. An antiperiplanar disposition of the H(2) and H(7) methine protons about the C(2)−C(7) bond is a common feature in all 12 (RS,RS)-diastereomer crystal structures [average H(2)− 9240

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

Article

C(7)−H(7) torsion angle of 71(8)° for the (2R,7S)enantiomers as illustrated in Figures 4 and 5. The alkyl substituents in this second group include Me, n-Bu, i-Bu-9, cyclopentylmethyl, benzyl, and N-Me-i-Bu-10. For the sake of simplicity, in a consistent (but arbitrary) manner from now on, we will utilize the antiperiplanar/synclinal descriptors of the H(2)−C(2)−C(7)−H(7) torsion angle to assign the C(2)− C(7) bond conformation without further specification that the methine H(2) and H(7) protons are involved in the definition. There are two independent molecules (A and B) in the asymmetric unit of the N-Me-i-Bu-10 monohydrate crystal. Both molecules show the same C(2)−C(7) bond synclinal sense conformation because superimposition of molecule A and B piperidine non-hydrogen atoms and C(7), C(8), C(12) (ipso), and C(18) (methyl) gives a root-mean-square difference of only 0.109 Å, which is reduced to 0.090 Å if C(methyl) is removed from the comparison. All the alkyl group conformations in the (RS,RS)/(RS,SR)diastereomer crystal structures of this study appear to be governed by avoidance of oppositely signed gauche +gauche − interactions16 involving non-hydrogen atoms in pairs of adjacent C(2)−C(7)/C(7)−C(8) and also C(7)−C(8)/ C(8)−C(9) bonds. These unfavorable steric arrangements are analogous to adverse cis-1,3-diaxial interactions in cis-1,3disubstituted cyclohexanes.16 Both crystal structures of threo-N-methyl methylphenidate HCl (ax-N-Me-1) (CCDC ref code NODBIR) and threo-p,Ndimethyl methylphenidate HCl (p-Me-ax-N-Me-1) (CCDC ref code NODBEN) contained the ax-N-Me-ax-CH(Ph)CO2Me conformer, while the former also contained the ring-inverted Nepimeric minor contributor present in a 0.71:0.29 disorder involving the N-methyl and piperidinyl ring C(4)−C(6) atoms, respectively (see Figure 6).17 The disorder involving the

Figure 4. ORTEP plot for the (2R,7S)-enantiomer of rac-9. Diastereotopic 8a and 8b protons are colored pink and white, respectively. Displacement ellipsoids are drawn at the 50% probability level. Characteristic torsion angles for this conformation are as follows: 83° for N(1)−C(2)−C(7)−C(8), 171° for C(2)−C(7)−C(8)−C(9), and −56° for C(7)−C(8)−C(9)−H(9).

Figure 5. ORTEP plot for the (2R,7S)-enantiomer of equatorial Nmethyl rac-10 monohydrate (molecule B in the Z′ = 2 asymmetric unit). Diastereotopic 8a and 8b protons are colored pink and white, respectively. Displacement ellipsoids are drawn at the 50% probability level. Characteristic torsion angles for this conformation are as follows: 54° for C(18)−N(1)−C(2)−C(7), 70° for N(1)−C(2)−C(7)−C(8), −160° for C(2)−C(7)−C(8)−C(9), and −52° for C(7)−C(8)− C(9)−H(9). Figure 6. PLUTO plot of the threo-N-methyl methylphenidate ammonium ion in the ax-N-Me-1 crystal structure showing the major ax-N-Me-ax-CH(Ph)CO2Me contributor (magenta) and its axN-Me-eq-CH(Ph)CO2Me minor partner (gray) to the 0.71:0.29 disorder.

C(2)−C(7)−H(7) torsion angle of 175(4)°]. This arrangement places the aryl and alkyl groups anti and syn to nitrogen, respectively (see Figures 2 and 3). However, in the crystal structures of a series of 13 (RS,SR)-analogues, there are two solid-state rotamer arrangements about the C(2)−C(7) bond. Five of the crystal structures showed the same antiperiplanar disposition of the H(2) and H(7) methine protons in which the relative syn/anti locations of the aryl and alkyl groups vis-à-vis nitrogen were now reversed [average H(2)−C(2)−C(7)−H(7) torsion angle of 175(3)°]. The other eight (RS,SR)-crystal structures exhibited an average (+)-synclinal H(2)−C(2)−

piperidinyl and N-methyl carbons bears a general resemblance to a bicyclo[3.3.1]nonane geometry. gauche +gauche − steric avoidance is also readily observed in the crystal structures of Nmethyl i-Bu-diastereomers 8 and 10. Figure 3 illustrates (1R,2R,7R)-i-Bu salt 8 with an axial N-methyl group resulting in favorable gauche −gauche − torsion angles for the two adjacent 9241

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

Article

N(1)−C(2)/C(2)−C(7) pairs of bonds: −56.8° for C(18)− N(1)−C(2)−C(7) and −59.6° for N(1)−C(2)−C(7)−C(8). Figure 5 depicts an equatorial N-methyl group in the (1S,2R,7S)-i-Bu-11 salt with gauche +gauche + torsion angles for the two adjacent N(1)−C(2)/C(2)−C(7) pairs of bonds: 54.0° for C(18)−N(1)−C(2)−C(7) and 69.9° for N(1)− C(2)−C(7)−C(8). Solution-State Weighted Time-Averaged Stereochemistry of α-Alkyl Analogues of Ritalin HCl Salts. Aliphatic 1H and 13C NMR spectral parameters for salts 6−10, measured in solvents with different dielectric constants (D 2O and CD2Cl2), are listed in Tables S1−S4 of the Supporting Information. Density functional theory B3LYP/6-311+G(2d,p) basis set geometry-optimized models18 were calculated using the crystal geometries of (2R,7R)-enantiomers for rac-7 and rac-8 (axial N-methyl) and the (2R,7S)-enantiomers for rac-9 and rac-10 (equatorial N-methyl) (Figures 2−5) as input structures. Density functional theory B3LYP/6-311+G(2d,p) basis set total nuclear spin−spin coupling Jcalcd parameters and also δ calcd values relative to those from a Td symmetry TMS model (NMR = SpinSpin keyword)18 were based on the geometry-optimized models as input structures and are also listed in Tables S2 and S4 of the Supporting Information. In major species 6−10 in both solvents, the 11−12 Hz J(2− 3ax) and 2−3 Hz J(2−3eq) vicinal coupling constants are typical for antiperiplanar and synclinal arrangements of the respective vicinal protons and are in accord with equatorially disposed CH(Ar)CH2-cyc-Pnt and CH(Ar)-i-Bu substituents ligated to the piperidinyl ring C(2) position. The disparate magnitudes of these values do not appear to result from significant weighted time averaging arising from conformational changes involving the C(2)−C(3) bond (e.g., ring inversion). An axial orientation for these substituents would have afforded synclinal magnitude values for both J(2−3ax) and J(2−3eq). The methoxycarbonyl groups in D2O solutions of threodiastereomers of 4-chloro methylphenidate (p-Cl-1) and of axial N-methyl methylphenidate (ax-N-Me-1) also have equatorially orientated CO2Me groups as shown by the respective J(2−3ax) values of 12.5(5) and 11.4(1) Hz and J(2−3eq) values of 2.6(1) and 2.5(1).17 Only the NMR spectra of (RS,RS)-i-Bu-7 in both D2O and CD2Cl2 showed the presence of a sole solvated species. One predominant form in the presence of minor species was observed at slow exchange in the spectra of other compounds in this study. With the exception of N-methylated 8 and 10, this paper will only discuss the predominant species found in each solvent. A solution-state antiperiplanar [H(2)−C(2)−C(7)− H(7)] conformational preference for the weighted time average of C(2)−C(7) bond rotamers at fast exchange was found for many of the compounds in this study. The evidence for this interpretation consists of measurements of 9−11 Hz for J(2− 7)exptl coupling constants, e.g., (RS,RS)-CH2-cyc-Pnt-6 [8.8(1) Hz in CD2Cl2], (RS,RS)-i-Bu-7 [8.6(2) Hz in CD2Cl2], (RS,RS)-ax-N-Me-i-Bu-8 [11.3(2) Hz in D2O and 9.5(2) Hz in CD2Cl2], p-Cl-1 [9.2(1) Hz in D2O and 10.3(1) Hz in CD2Cl2], and ax-N-Me-1 [11.0(1) Hz in D2O and 9.9(1) Hz in CD2Cl217]. These weighted time-averaged coupling constants are all consistent with a high proportion of the same antiperiplanar H···H geometry about the C(2)−C(7) bond that is found in the (RS,RS)-i-Bu-7 crystal structure (Figure 2). The average J(2−7)exptl values for both (RS,SR)-i-Bu-9 [9.8(8) Hz in D2O] and (RS,SR)-ax-N-Me-i-Bu-10 [9.2(3) Hz in D2O] are not commensurate with (+)-synclinal H(2) and H(7)

methine proton arrangements found in crystal structures of these same compounds (see Figures 4 and 5). The 8.9−9.6 Hz DFT J(2−7)calcd italicized values listed in Tables S2 and S4 of the Supporting Information based on antiperiplanar H(2) and H(7) models are reduced in magnitude relative to more usual calculated values of 11−12 Hz (and also experimentally measured) for ca. 180° H···H arrangements in other bonds. This is likely to be a result of the electron withdrawing groups ligated to both termini of the C(2)−C(7) bond. The smaller J(2−7)exptl coupling constants of the major species of (RS,RS)CH2-cyc-Pnt-6 [7.2(1) Hz in D2O], (RS,RS)-i-Bu-7 [7.2(1) Hz in D2O], and (RS,SR)-i-Bu-9 [6.5(1) Hz in CD2Cl2] may be rationalized in terms of a larger contribution of synclinal rotamers to their solution-state weighted time-average geometries. Epimerization at the Methylated Nitrogen. Similar to solutions of threo- or erythro-N-methyl methylphenidate,17 dissolution in D2O or CD2Cl2 of either (RS,RS)-ax-N-Me-iBu-8 or (RS,SR)-eq-N-Me-i-Bu-10 crystalline salts allowed the labile stereogenic nitrogen atoms19,20 to undergo a prototropic shift and nitrogen inversion. The outcome of this facile process is a mixture of axial/equatorial N-methyl epimeric diastereomers. The N-methyl diastereomeric ratio is dependent on both the solvent and the configuration at C(7). The (RS,RS)-NMe-i-Bu-8 ax-N:eq-N ratio is 1:1 in D2O and 4:1 in CD2Cl2, while this ratio is now reversed for the (RS,SR)-N-Me-i-Bu-10 diastereomer (9:11 in D2O and 1:3 in CD2Cl2). The same configurational and conformational geometries as in crystal structures 8 and 10 were found for the major solution-state species of 8 [ax-N-methyl, antiperiplanar H(2) and H(7) disposition in CD2Cl2] and for the major species of 10 [eq-Nmethyl, synclinal H(2) and H(7) arrangement in both solvents]. The quantity of the mixture’s major species was found to be enhanced in solutions of the lower-dielectric constant CD2Cl2 (compared to D2O). An N-methyl epimerization of ax-N-Me-8 producing an equatorial N-methyl diastereomer in which its original antiperiplanar H(2) and H(7) crystalline-state arrangement remains invariant would lead to a sterically unfavorable oppositely signed gauche −gauche + interaction as shown by the blue double-headed arrow in 11. An antiperiplanar to synclinal conformational change of the C(2)−C(7) bond concomitant with a prototropic shift or nitrogen inversion removes this problematic steric outcome. The observation of synclinal 4.3(3) and 4.3(2) Hz J(2−7)exptl weighted time-averaged coupling constants in D2O and CD2Cl2 together with characteristic equatorial Nmethylpiperidine chemical shifts of 43.70 and 42.22 ppm (in D2O and CD2Cl2,17 respectively) strongly suggests that a configurational change via a prototropic shift−nitrogen inversion mechanism together with a change in the predominant C(2)−C(7) bond conformation is indeed a valid rationalization of the observations. The Jcalcd coupling constants italicized in Table S2 of the Supporting Information for an eq-N-Me-(+)-synclinal H(2)···H(7) DFT p-Cl-8 geometry-optimized model (Figure 7) are completely consistent with this interpretation. The proposed (+)-synclinal H(2) and H(7) (2R,7R)-candidate has the lowest DFT B3LYP/6-311+G(2d,p) relative calculated energy of the three C(2)−C(7) bond rotamer models: 0.00 kcal/mol for (+)-synclinal, 3.03 kcal/mol for (−)-synclinal, and 4.30 for kcal/mol antiperiplanar. The Boltzmann distribution for these models is 99.3:0.6:0.1 at 25 °C under vacuum. Assuming that a Boltzmann distribution of the actual solvated rotamers is similar 9242

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

Article

Figure 8. PLUTO plot for a DFT geometry-optimized model of (2S,7R)-ax-N-Me-p-Cl-10. Diastereotopic 8a and 8b protons are colored pink and white, respectively. Characteristic torsion angles for this conformation are as follows: 62° for C(18)−N(1)−C(2)−C(7), 168° for N(1)−C(2)−C(7)−C(8), 176° for C(2)−C(7)−C(8)− C(9), and 54° for C(7)−C(8)−C(9)−H(9).

antiperiplanar H(2)···H(7) candidate has the lowest DFT B3LYP/6-311+G(2d,p) relative calculated energy of the three C(2)−C(7) bond rotamer models: 0.00 kcal/mol for antiperiplanar, 6.01 kcal/mol for (−)-synclinal, and 6.00 for kcal/mol (+)-synclinal. The Boltzmann distribution concentration for the antiperiplanar model is essentially 100% at 25 °C under vacuum. A comparison of the 13C δ parameters for the N-methyl diastereomeric mixtures of 8 and 10 listed in Tables S1 and S3 of the Supporting Information shows the expected closer to TMS γ-gauche effect21,22 values for piperidine ring NCH3, C(3), and C(5) signals in the axial N-methyl epimer. The average γgauche effects for ax-N-Me-8 and -10 measured in both solvents are 9(1) (NCH3), 2.1(9) [C(3)], and 4.4(9) ppm [C(5)] closer to TMS versus average δ values for corresponding atoms in the spectra of the eq-N-Me-8 and -10 diastereomers. Solution-State Conformation of the C(α)-Alkyl Side Chains. The conformations of the CH2-cyc-Pnt side chain in 6 and of the i-Bu side chains in 7−10 are all very similar irrespective of the relative C(2)−C(7) bond configuration, the dielectric constant of the solvent, or the packing arrangement of the crystal structure. This sterically auspicious spatial disposition encompasses antiperiplanar [C(2)−C(7)−C(8)− C(9)] and synclinal [C(7)−C(8)−C(9)−H(9)] torsion angles so that H(9) points inward toward the flat face of the neighboring aryl ring. In all the side chains of the crystal structures, H(8a) (pink in Figures 2−7) is antiperiplanar to H(7) while its diastereotopic H(8b) partner (white in Figures 2−7) is antiperiplanar to H(9). Tables S2 and S4 of the Supporting Information show that these crystal structure arrangements are clearly maintained in both solutions as evidenced by 11.7(9) Hz average J(7−8a)exptl and 10.1(5) Hz average J(8a−9)exptl values whose large magnitudes do not appear to result from weighted time averaging. The largest difference in chemical shifts for diastereotopic H(8a) and H(8b) was measured for eq-N-Me-10 in D2O (1.98 and 1.19 ppm, respectively) and is in accord with the respective DFT δ calcd values of 2.10 and 0.42. Further evidence of the preferred side chain conformation mentioned above is seen in the fact that the diastereotopic anisochronous 13C methyl chemical shifts in the six i-Bu entries in Tables S1 and S3 of the Supporting Information appear to exhibit a γ-gauche effect for one of the methyl carbons in accord with the proposed structures. Using DFT geometry-optimized

Figure 7. PLUTO plot for a DFT geometry-optimized model of (2R,7R)-eq-N-Me-p-Cl-8. Diastereotopic 8a and 8b protons are colored pink and white, respectively. Characteristic torsion angles for this conformation are as follows: 59° for C(18)−N(1)−C(2)−C(7), −173° for N(1)−C(2)−C(7)−C(8), 165° for C(2)−C(7)−C(8)− C(9), and 53° for C(7)−C(8)−C(9)−H(9).

to that noted above, the few very low-intensity multiplets from minor forms in the spectrum were neglected. The 4.3(3) Hz J(2−7)exptl value measured in both solvents is in nice agreement with the 4.1 Hz J(2−7)calcd parameter calculated from the lower-energy 54° H(2)···H(7) (+)-synclinal rotamer model. The J(2−7)exptl value mentioned above is clearly in disagreement with the very small J(2−7)calcd value of 0.9 Hz based on the higher-calculated energy −84° H(2)···H(7) (−)-synclinal partner.

The blue double-headed arrow in 12 illustrates that an analogous N-methyl epimerization of eq-N-Me-10 while its original (+)-synclinal H(2) and H(7) crystalline-state conformation was kept invariant would lead to a new axial Nmethyl analogue exhibiting a sterically unfavorable oppositely signed gauche −gauche + interaction. Once again, a C(2)−C(7) bond antiperiplanar to synclinal conformational change concomitant with a prototropic shift and nitrogen inversion removes this problematic steric outcome. The measurement of an antiperiplanar 9.2(3) Hz J(2−7)exptl (in D2O) weighted time-averaged coupling constant together with characteristic γgauche effect20,21 axial N-methylpiperidine chemical shifts of 34.61 and 34.02 ppm (in D2O and CD2Cl2, respectively) supports the validity of this rationalization. The Jcalcd coupling constants in Table S4 of the Supporting Information for an axN-Me antiperiplanar H(2)···H(7) DFT p-Cl-10 geometryoptimized model (Figure 8) are completely consistent with the interpretation described above. Furthermore, the proposed 9243

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

Article

signals. HMQC two-dimensional NMR spectroscopy was used to correlate the 13C and 1H chemical shifts. COSY, NOESY, and NOEDIFF experiments were performed to aid in signal assignment. General Methods for X-ray Crystallography..23 Crystallizations were performed by vapor diffusion of hexane into ethyl acetate solutions of salts 7−10. Crystals were mounted on a glass fiber and then fixed to the goniometer head of the X-ray diffractometer. Crystallographic diffraction measurements were taken on a Bruker SMART 1000 CCD diffractometer for 7−9 (or SMART Apex II CCD diffractometer for 10) with sealed tube source graphite monochromated Mo Kα (λ = 0.71073 Å) radiation for 7 and 10 in the ω scan mode and for 8 and 9 in the ϕ−ω scan mode. Data were reduced with SAINT.24 Analysis of the data showed negligible decay during data collection; the intensities were corrected for Lorentz and polarization factors, and an empirical absorption coefficient was applied via SADABS.25 A “multiscan” absorption correction was used. The structures were determined by application of direct methods and refined by full matrix least-squares on F 2 using SHELXTL.26 Hydrogens were introduced at calculated positions as riding atoms and were refined isotropically. All non-hydrogen atoms were refined anisotropically. X-ray Crystallography of (RS,RS)-2-[1-(4-Chlorophenyl)-3methylbutyl]piperidinium Chloride (rac-7) (CCDC deposit number 255963). A clear thin plate of C 16H25Cl2N having approximate dimensions of 0.57 mm × 0.18 mm × 0.04 mm was chosen. Data were collected to a maximum θ value of 28.23° (96.2% completeness to θ). Cell constants correspond to a monoclinic system C2/c cell with the following dimensions at 93(2) K: a = 28.907(10) Å, b = 7.097(2) Å, c = 19.340(6) Å, β = 120.174(6)°, V = 3430.3(19) Å3, and Z = 8. When Z = 8 and fw = 302.27, the calculated density is 1.171 g/cm3. An empirical absorption coefficient (μ = 0.367 mm−1) was applied, and the transmission coefficients were as follows: Tmin = 0.811, and Tmax = 0.985; 12088 reflections were measured, 4082 of them being unique [R(int) = 0.0416]. At convergence, the final discrepancy indices on F were as follows: R(F) = 0.0428, Rw(F 2) = 0.1077, and GOF S = 1.078 for the 2914 reflections with Inet > 2σ(Inet) and 172 parameters refined with zero constraints. The residual negative and positive electron densities in the final map were −0.412 and 0.566 e/Å3, respectively. X-ray Crystallography of (RS,RS)-2-[1-(3,4-Dichlorophenyl)3-methylbutyl]-1-methylpiperidinium Chloride (rac-8) (CCDC deposit number 827232). A clear prism of C17H26Cl3N having approximate dimensions of 0.34 mm × 0.28 mm × 0.16 mm was chosen. Data were collected to a maximum θ value of 28.29° (97.5% completeness to θ). Cell constants correspond to a monoclinic system P21/n cell with the following dimensions at 103(2) K: a = 12.720(5) Å, b = 11.077(4) Å, c = 13.974(6) Å, β = 112.51(1)°, V = 1818.9(13) Å3, and Z = 4. When Z = 4 and fw = 350.74, the calculated density is 1.281 g/cm3. An empirical absorption coefficient (μ = 0.498 mm−1) was applied, and the transmission coefficients were as follows: Tmin = 0.844, and Tmax = 0.923; 14733 reflections were measured, 4402 of them being unique [R(int) = 0.0405]. At convergence, the final discrepancy indices on F were as follows: R(F) = 0.0610, Rw(F 2) = 0.1601, and GOF S = 1.037 for the 3117 reflections with Inet > 2σ(Inet) and 185 parameters refined with zero constraints. The residual negative and positive electron densities in the final map were −0.690 and 1.962 e/Å3, respectively. X-ray Crystallography of (RS,SR)-2-[1-(3-Chlorophenyl)-3methylbutyl]piperidinium Chloride (rac-9) (CCDC deposit number 827231). A clear prism of C16H25Cl2N having approximate dimensions of 0.34 mm × 0.18 mm × 0.16 mm was chosen. Data were collected to a maximum θ value of 28.31° (95.6% completeness to θ). Cell constants correspond to a monoclinic system P21/n cell with the following dimensions at 93(1) K: a = 7.8496(15) Å, b = 26.389(5) Å, c = 8.7973(15) Å, β = 115.512(4)°, V = 1644.6(5) Å3, and Z = 4. When Z = 4 and fw = 302.27, the calculated density is 1.221 g/cm 3. An empirical absorption coefficient (μ = 0.383 mm−1) was applied, and the transmission coefficients were as follows: Tmin = 0.723, and Tmax = 0.941; 13271 reflections were measured, 3911 of them being unique [R(int) = 0.0485]. At convergence, the final discrepancy indices on F

structures, the DFT average δ(CH3)calcd values are 24.0(2) and 19.6(2) for the antiperiplanar and synclinal methyl carbons, respectively, in the six models [numbered C(10) and C(11), respectively, in Figures 2−7]. The average δ(CH3)exptl values are 25.6(3) and 22.6(3) for the two diastereotopic methyl carbons (six examples measured in D2O), versus average values of 23.7(4) and 20.8(2), respectively, for the six examples measured in CD2Cl2. On this basis, one can propose that the δ(CH3)exptl values closer to TMS be assigned to a synclinal [C(7)−C(8)−C(9)−C(11)] methyl carbon (CH3′), while those farther from TMS be assigned to the antiperiplanar [C(7)−C(8)−C(9)−C(10)] methyl carbon partner (CH3).

■

CONCLUSIONS

■

EXPERIMENTAL SECTION

For the compounds studied here, avoidance of sterically unfavorable gauche −gauche + orientations plays a major role in the C(α)-alkyl side chain conformation and in the formation of a predominant C(2)−C(7) bond rotamer. In both solvents, the J(2−7)exptl parameters appear to be weighted time averages of J values arising from a mixture of rotamers in fast exchange about the C(2)−C(7) bond. In this series of salts, the primary contributor to the C(2)−C(7) bond rotameric mixture of the solution-state major species appears to have a preferred antiperiplanar relationship for vicinal methine protons H(2) and H(7). The contribution of this preferred conformation to the rotameric mixture varies from compound to compound, and from solvent to solvent. In the case of the pharmacologically interesting salts (2RS,7RS)-CH 2 -cyc-Pnt-6 and (2RS,7RS)-i-Bu-7, this solution-state favored rotamer has the same geometry that was found in the crystal structures, although it is clearly in fast equilibrium with smaller quantities of synclinal partners. While the crystal structure of (RS,SR)-iBu-9 has a synclinal orientation for vicinal protons H(2) and H(7), when these crystals were dissolved in D2O, the weighted time-averaged arrangement for these protons became almost completely antiperiplanar. Judging by the weighted timeaveraged J(2−7)exptl parameters [average value of 10(1) or 3.8(6) Hz] of the four entries of N-methyl tertiary ammonium salts (8 and 10) in Tables S2 and S4 of the Supporting Information, increased steric congestion around the C(2)− C(7) bond seems to augment the quantity of the preferred contributor to the rotameric mixture. The stereochemistry of the NMR-observed species seems to arise from specific combinations of N-methyl configuration and C(2)−C(7) bond conformation in which sterically unfavorable gauche −gauche + arrangements are avoided. Thus, axial N-methyl diastereomers favored antiperiplanar H(2)···H(7) orientations [(2RS,7RS)-ax-N-Me-i-Bu-8 and (2RS,7SR)-ax-N-Me-i-Bu-10], while equatorial N-methyl diastereomers exhibited synclinal arrangements of H(2) and H(7) [(2RS,7RS)-eq-N-Me-i-Bu-8 and (2RS,7SR)-eq-N-Me-i-Bu-10]. NMR Spectroscopy. Solution-state 1H and 13C NMR spectra were recorded at ambient temperature on Bruker FT-NMR spectrometers at either 500.13 and 125.76 MHz, respectively, or 800.13 and 201.19 MHz, respectively. The deuterium solvent signal was utilized as an internal lock. 13C and 1H chemical shifts were based on either the respective CD2Cl2 (53.8 ppm) and residual CHDCl2 (5.32 ppm) signals as internal references or the internal 4,4-dimethyl4-silapentanesulfonate sodium salt (DSS) when samples were measured in D2O. Homonuclear 1H decoupling was used to assign the spin−spin coupling constants. The DEPT-135 and DEPT-90 pulse sequences were used to ascertain the hydrogen multiplicity of the 13C 9244

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245

The Journal of Organic Chemistry

Article

were as follows: R(F) = 0.0466, Rw(F 2) = 0.1187, and GOF S = 1.037 for the 2973 reflections with Inet > 2σ(Inet) and 174 parameters refined with zero constraints. The residual negative and positive electron densities in the final map were −0.300 and 0.505 e/Å3, respectively. X-ray Crystallography of (RS,RS)-2-[1-(3,4-Dichlorophenyl)3-methylbutyl]-1-methylpiperidinium Chloride Monohydrate (rac-10) (CCDC deposit number 827230). A clear plate of C17H28Cl3NO having approximate dimensions of 0.49 mm × 0.16 mm × 0.04 mm was chosen. Data were collected to a maximum θ value of 25.09° (95.0% completeness to θ). Cell constants correspond to a monoclinic system C2/c cell with the following dimensions at 103(2) K: a = 37.226(19) Å, b = 8.122(4) Å, c = 27.186(13) Å, β = 109.884(12)°, V = 7730(7) Å3, Z = 16, and Z′ = 2. When Z = 16 and fw = 745.49, the calculated density is 1.281 g/cm 3. An empirical absorption coefficient (μ = 0.477 mm−1) was applied, and the transmission coefficients were as follows: Tmin = 0.7605, and Tmax = 0.9767; 17430 reflections were measured, 6544 of them being unique [R(int) = 0.1247]. At convergence, the final discrepancy indices on F were as follows: R(F) = 0.0909, Rw(F 2) = 0.2150, and GOF S = 1.039 for the 3417 reflections with Inet > 2σ(Inet) and 406 parameters refined with zero constraints. The residual negative and positive electron densities in the final map were −0.544 and 0.895 e/Å3, respectively.

■

(6) Froimowitz, M.; Patrick, K. S.; Cody, V. Pharm. Res. 1995, 12, 1430−1434. (7) Froimowitz, M. J. Comput. Chem. 1993, 14, 934−943. (8) Froimowitz, M.; Deutsch, H. M.; Shi, Q.; Wu, K.-M.; Glaser, R.; Adin, I.; George, C.; Schweri, M. M. Bioorg. Med. Chem. Lett. 1997, 7, 1213−1218. (9) Miles, M. P.; Johnson, K. M.; Zhang, M.; Flippen-Anderson, J. L.; Kozikowski, A. P. J. Am. Chem. Soc. 1998, 120, 9072−9073. (10) Smith, M. P.; George, C.; Kozikowski, A. P. Tetrahedron Lett. 1998, 39, 197−200. (11) Kozikowski, A. P.; Roberti, M.; Johnson, K. M.; Bergmann, J. S.; Ball, R. G. Bioorg. Med. Chem. Lett. 1993, 3, 1327−1332. (12) Kozikowski, A. P.; Saiah, M. K. E.; Johnson, K. M.; Bergmann, J. S. J. Med. Chem. 1995, 38, 3086−3093. (13) Kozikowski, A. P.; Araldi, G. L.; Prakash, K. R. C.; Zhang, M.; Johnson, K. M. J. Med. Chem. 1998, 41, 4973−4982. (14) Froimowitz, M.; Gu, Y.; Dakin, L. A.; Nagafuji, P. M.; Kelley, C. J.; Parrish, D.; Deschamps, J. R.; Janowsky, A. J. Med. Chem. 2007, 50, 219−232. (15) Wolraich, M. L.; Doffing, M. A. CNS Drugs 2004, 18, 243−250. (16) Hoffmann, R. W. Angew. Chem., Int. Ed. 1992, 31, 1124−1134. (17) Glaser, R.; Adin, I.; Shiftan, D.; Shi, Q.; Deutsch, H. M.; George, C.; Wu, K.-M.; Froimowitz, M. J. Org. Chem. 1998, 63, 1795− 1794. (18) Frisch, M. J.; et al. GAUSSIAN 03, revision D.01; Gaussian, Inc.: Wallingford, CT, 2004. (19) Closs, G. L. J. Am. Chem. Soc. 1959, 81, 5456. (20) Glaser, R.; Peng, Q.-J.; Perlin, A. S. J. Org. Chem. 1988, 53, 2172. (21) Lambert, J. B.; Vagenas, A. R. Org. Magn. Reson. 1981, 17, 265. (22) Kleinpeter, E.; Seidl, P. R. J. Phys. Org. Chem. 2004, 17, 680. (23) CCDC 255963 (for 7), CCDC 827232 (for 8), CCDC 827231 (for 9), and CCDC 827230 (for 10) contain the supporting crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Structural Database, Cambridge Crystallographic Data Centre, via www.ccdc.cam.ac.uk/data_request/ cif. (24) SAINT, version 6.44a; Bruker AXS Inc.: Madison, WI, 2002. (25) SADABS, version 2.10; Bruker AXS Inc.: Madison, WI, 2002. (26) SHELXTL, version 6.14; Bruker AXS Inc.: Madison, WI, 2000.

ASSOCIATED CONTENT

S Supporting Information *

cif files for 7−10, 1H and 13C NMR chemical shifts for aliphatic nuclei in (RS,RS)-rac-6, -rac-7, and -rac-8 and also in (RS,SR)rac-9 and -rac-10 salts (Tables S1 and S3), experimental homonuclear spin−spin coupling constants and DFT-calculated total nuclear spin−spin J couplings (hertz) between aliphatic nuclei in the salts mentioned above (Tables S2 and S4), crystal data and structure refinement, non-hydrogen atomic coordinates and equivalent isotropic displacement parameters, bond lengths and angles, non-hydrogen atom anisotropic displacement parameters, hydrogen coordinates and isotropic displacement parameters, torsion angles, and hydrogen bonds for 7−10 (Tables S5−S31), final coordinates and energies for DFT B3LYP/6-311+G(2d,p) geometry-optimized models of i-Bu-7 (antiperiplanar rotamer), ax-N-Me-i-Bu-8 (antiperiplanar rotamer), eq-N-Me-i-Bu-8 [(+)-synclinal rotamer], i-Bu-9 (antiperiplanar rotamer), ax-N-Me-i-Bu-10 (antiperiplanar rotamer), and eq-N-Me-i-Bu-10 [(+)-synclinal rotamer] (Tables S32− S37, respectively), and the full citation for ref 18. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected]. Present Address ∥ 90 Eastbourne Rd., Newton, MA 02459.

■

ACKNOWLEDGMENTS We extend our gratitude to Dr. Rüdiger Weisemann (Bruker Biospin GmbH, Rheinstetten, Germany) for 1H and 13C NMR spectra measured on a DMX-800 spectrometer.

■

REFERENCES

(1) The Merck Index, 14th ed.; O’Neil, M. J., Ed.; Merck: Whitehouse Station, NJ, 2006; p 1053 and references therein. (2) Maxwell, R. A.; Chaplin, E.; Eckhardt, S. B.; Soares, J. R.; Hite, G. J. Pharmacol. Exp. Ther. 1970, 173, 158−165. (3) Ferris, R. M.; Tang, F. L. M.; Maxwell, R. A. J. Pharmacol. Exp. Ther. 1972, 181, 407−416. (4) Solanto, M. V. Behav. Brain Res. 1998, 94, 127−152. (5) Shafi’ee, A.; Hite, G. J. Med. Chem. 1969, 12, 266−270. 9245

dx.doi.org/10.1021/jo201415h | J. Org. Chem. 2011, 76, 9239−9245