PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Theor Chem Acc. Author manuscript; available in PMC 2010 October 1.
Published in final edited form as:
Theor Chem Acc. 2009 October 1; 124(3-4): 269–278.
doi:  10.1007/s00214-009-0607-z
PMCID: PMC2796847
NIHMSID: NIHMS151639

First-Principles Determination of Molecular Conformations of Indolizidine (−)-235B′ in Solution

Abstract

Indolizidine (−)-235B′ is a particularly interesting natural product, as it is the currently known, most potent and subtype-selective open-channel blocker of the α4β2 nicotinic acetylcholine receptor (nAChR). In the current study, extensive first-principles electronic structure calculations have been carried out in order to determine the stable molecular conformations and their relative free energies of the protonated and deprotonated states of (−)-235B′ in the gas phase, in chloroform, and in aqueous solution. The 1H and 13C NMR chemical shifts calculated using the computationally determined dominant molecular conformation of the deprotonated state are all consistent with available experimental NMR spectra of (−)-235B′ in chloroform, which suggests that the computationally determined molecular conformations are reasonable. Our computational results reveal for the first time that two geminal H atoms on carbon-3 (C3) of (−)-235B′ have remarkably different chemical shifts (i.e. 3.24 and 2.03 ppm). The computational results help one to better understand and analyze the experimental 1H NMR spectra of (−)-235B′. The finding of remarkably different chemical shifts of two C3 geminal H atoms in a certain molecular conformation of (−)-235B′ may also be valuable in analysis of NMR spectra of other related ring-containing compounds. In addition, the pKa of (−)-235B′ in aqueous solution is predicted to be ~9.7. All of the computational results provide a solid basis for future studies of the microscopic and phenomenological binding of various receptor proteins with the protonated and deprotonated structures of this unique open-channel blocker of α4β2 nAChRs. This computational study also demonstrates how one can appropriately use computational modeling and spectroscopic analysis to address the structural and spectroscopic problems that cannot be addressed by experiments alone.

Introduction

Natural products exhibit a remarkable diversity of structural motifs that can be systematically exploited for drug discovery.1 Thus, natural products provide one of the richest sources of scaffold structures for the identification of novel leads for rational drug design. Alkaloids detected in amphibian skin contain over 20 structural classes, in which various biologically active, lipid-soluble alkaloids have been discovered.2 Besides the epibatidines that are leads in the drug design of new analgesics,3,4 the largest subclass of these alkaloids, i.e. the 5,8-disubstituted indolizidines, has attracted much interest.5

So far, about 80 entities of 5,8-disubstituted indolizidines have been found in skin extracts from frogs and toads.2 None of these alkaloids have been identified from any other natural source. Many research groups across the world have expended great effort into the synthesis5,6,7,8 and characterization of such compounds.9,10 The 5,8-disubstituted indolizidine alkaloid, (−)-235B′ (Figure 1, the C numbering system is taken from Edwards and Daly11) isolated from Dendrobates pumilio (the poison dart frog), is one of the well-characterized ligands of nicotinic acetylcholine receptors (nAChRs). Unlike epibatidine, which is a potent agonist of α4β2 nAChRs,3 (−)-235B′ acts as a noncompetitive and subtype-selective antagonist of α4β2 nAChRs.10 Neuronal nAChRs are members of a superfamily of ligand-gated ion channels, which modulate the function of many major neurotransmitter systems and influence a broad range of brain functions, including cognition, learning, and memory.12,13,14,15,16 Selective α4β2 nAChR agonists and antagonists have therapeutic potential in the treatment of Alzheimer’s disease,17 Parkinson’s disease,18 pain,19 and tobacco-use addiction etc.20 Currently, (−)-235B′ is the most potent and subtype-selective open-channel blocker (IC50 = 74 nM) of α4β2 nAChRs known. This noncompetitive antagonist of α4β2 nAChRs is being considered as a promising lead for new drug discovery to treat cholinergic disorders such as autosomal dominant nocturnal frontal lobe epilepsy.9,10

Figure 1
Molecular structure of (−)-235B′.

The particular characteristics of (−)-235B′ indicate that understanding how this alkaloid interacts with neuronal nAChRs may aid in the rational drug design of more potent antagonists that alter nicotine cholinergic functions in a subtype-selective manner. In addition to quantitative structure-activity relationship (QSAR) studies,21,22,23 molecular modeling and simulation of nAChR binding with agonists and antagonists have provided a better understanding of the molecular mechanisms involved in nAChR function, in addition to the results obtained using [3H]nicotine binding assays, electrophysiology, and molecular biology.24 As demonstrated in our previous computational studies on other nAChR ligands,25,26,27 in order to study how a ligand binds with an nAChR, we first need to know various possible molecular structures and conformations of the ligand itself. This is because an alkaloid, such as (−)-235B′, can exist in multiple possible molecular species (protonated and deprotonated states) in aqueous solution. Furthermore, for a given protonation state, the ligand could also exist in various possible molecular conformations. All of the possible protonation states/conformations could bind to a receptor protein with different binding affinities. It has been demonstrated in our previous computational studies25 on nicotine and deschloroepibatidine that the most stable molecular species and conformer in solution is not necessarily the most important molecular species and conformer that interacts with the binding site of an nAChR. Therefore, all molecular species of (−)-235B′, whether protonated or deprotonated, and all of the possible conformations are potentially important for binding with the protein when the differences between the relative concentrations of different molecular species and conformations of the free ligand in solution are not very large. The relative Gibbs free energies of different molecular species and conformations are important information required to determine their concentration distribution in solution and their relative contributions to the receptor-ligand binding. In order to determine the phenomenological binding free energy, one must know both the relative free energies of various molecular species/conformations of the free ligand in solution and the microscopic binding free energies associated with all molecular species/conformations. Knowing the concentration distribution of all structural forms of (−)-235B′ in solution, one can further study the theoretical binding of each structural form with the nAChR and evaluate the microscopic binding free energy. The appropriately combined use of the calculated microscopic binding free energies and the concentration distribution of different structural forms of (−)-235B′ can then lead to a theoretical prediction of the overall, phenomenological binding free energy for (−)-235B′ binding with the nAChR, as demonstrated in our recent computational studies on other nAChR ligands.25,26,28

With a goal to eventually understand the microscopic and phenomenological binding of frog toxin (−)-235B′ with α4β2 and many other nAChRs, as well as the open-channel blocking effects of this ligand on the acetylcholine-activated α4β2 nAChRs in electrophysiological measurements, in the present study we carefully examined various possible molecular conformations of both the protonated and deprotonated states of (−)-235B′ and predicted their 1H and 13C nuclear magnetic resonance (NMR) chemical shifts and relative Gibbs free energies in solution by carrying out extensive first-principles electronic structure calculations that accurately account for solvent effects. The calculated results and analysis have led us to better understand the detailed molecular structures and NMR spectra of (−)-235B′ in solution.

Results and Discussion

Optimized geometries and calculated relative free energies

We first carried out a detailed conformational search for various possible conformations of the neutral (deprotonated) state of (−)-235B′. The conformational search was focused on various possible conformations of the rings, with the understanding that the six-carbon chain external to the bicyclic portion should exist in the all-anti conformation. Depicted in Figure 2 are four stable conformations that were identified: two chair conformations, denoted by (−)-235B′-A and (−)-235B′-B, and two boat conformations, denoted by (−)-235B′-C and (−)-235B′-D. The geometry optimization starting from any one of the initial structures considered in this study eventually went to one of these four stable geometries, or to an unstable geometry with at least one imaginary vibrational frequency. The relative Gibbs free energies calculated for these deprotonated conformers in the gas phase and in solution (i.e. chloroform and aqueous solution) are summarized in Table 1.

Figure 2
Stable conformations of the deprotonated state of (−)-235B′ optimized at the B3LYP/6-31+G(d) level.
Table 1
Calculated relative Gibbs free energies (ΔG in kcal/mol) of the conformations of the deprotonated state of (−)-235B′ in the gas phase (298.15 K and 1 atm) and in solution.

As one can see from Table 1, the energetic results calculated in the gas phase and in solution using two different solvation models and various basis sets are all qualitatively consistent with each other. Particularly, the calculated relative free energies are not sensitive to the basis set used in the calculations. When the same solvation method was used, the differences between the B3LYP/6-31+G(d) results and the corresponding 6-31++G(d,p) results are only about 0.1 or 0.2 kcal/mol. Further increasing the basis set from 6-31++G(d,p) to 6-311++G(d,p) and to 6-311++G(2d,2p), the changes of the calculated relative free energies are all within 0.2 kcal/mol. The consistent energetic results suggest that even 6-31+G(d), the smallest one of the four basis sets examined in this study, is adequate for studying the relative stability of the various conformations involved in this work. For convenience, unless otherwise indicated, the discussion below will only refer to the energetic results associated with the SVPE solvation model and the largest one of the four basis sets, i.e. 6-311++G(2d,2p).

The energetic results in Table 1 reveal that the most stable conformation of the deprotonated state is always (−)-235B′-A, no matter whether it is in the gas phase, in chloroform, or in aqueous solution. The free energies of the other conformations are higher than that of (−)-235B′-A by ~6.3 to 11.9 kcal/mol in the gas phase, by ~6.9 to 12.0 kcal/mol in chloroform, and by ~6.4 to 11.4 kcal/mol in aqueous solution. Based on the calculated relative free energies, (−)-235B′-A should always be the dominant conformation of the deprotonated state both in the gas phase and in solution. For example, the calculated free energy difference of ~6.9 kcal/mol between (−)-235B′-A and (−)-235B′-C in chloroform suggests that the ratio of the concentration of (−)-235B′-A to that of (−)-235B′-C is ~1.1 × 105 at room temperature (298.15 K).

Starting from the optimized geometries of the deprotonated state, geometries of the corresponding conformations in the protonated state were optimized. Depicted in Figure 3 are the optimized geometries: (−)-235B′H+-A, (−)-235B′H+-B, (−)-235B′H+-C, and (−)-235B′H+-D, which are the protonated structures of (−)-235B′-A, (−)-235B′-B, (−)-235B′-C, and (−)-235B′-D, respectively. The calculated relative free energies are summarized in Table 2. As shown in Table 2, the calculated relative free energies are also insensitive to the basis set used in the calculations. The differences between the B3LYP/6-31+G(d) results and the 6-31++G(d,p) results are about 0.1 or 0.2 kcal/mol. The differences between the B3LYP/6-31++G(d,p) results and the 6-311++G(d,p) results are between 0.0 and 0.2 kcal/mol. Further increasing the basis set from 6-311++G(d,p) and to 6-311++G(2d,2p), the changes of the calculated relative free energies are all within 0.1 kcal/mol. The discussion below will only refer to the energetic results calculated at the B3LYP/6-311++G(2d,2p) level.

Figure 3
Stable conformations of the protonated state of (−)-235B′, denoted by (−)-235B′H+, optimized at the B3LYP/6-31+G(d) level.
Table 2
Calculated relative Gibbs free energies (ΔG in kcal/mol) of the conformations of the protonated state of (−)-235B′ in the gas phase (298.15 K and 1 atm) and in solution.

As seen in Table 2, the relative free energies of the four conformations of the protonated state of (−)-235B′ are qualitatively the same as those of the corresponding four conformations of the deprotonated state. The most stable conformation of the protonated state is (−)-235B′H+-A in both the gas phase and in aqueous solution. We did not calculate the protonated state in chloroform solution because the experimental NMR spectra were available only for the free base (i.e. the deprotonated state) of (−)-235B′ in chloroform solution. The free base of (−)-235B′ is not expected to be protonated in chloroform solution. The free energies of the other conformations are higher than that of (−)-235B′H+-A by ~6.4 to 10.3 kcal/mol in the gas phase and ~6.9 to 9.7 kcal/mol in aqueous solution. Protonation slightly decreases the differences in free energy between the most stable conformation and the others. The calculated free energy difference of ~6.9 kcal/mol between (−)-235B′H+-A and (−)-235B′H+-C in aqueous solution predicts that the ratio of the concentration of (−)-235B′H+-A to that of (−)-235B′H+-C is ~1.1 × 105 at 298.15 K.

NMR chemical shifts

Witala, Cramer, and Hoye recently used computed 1H and 13C NMR chemical shifts to distinguish stereoisomers of diastereomeric penam beta-lactams.29 In order to compare the optimized geometries with available experimental structural information (i.e. 1H and 13C NMR spectra in chloroform)30,31 on this nAChR antagonist, we calculated the 1H and 13C NMR chemical shifts for each of the optimized geometries of the deprotonated state in chloroform. The predicted 1H and 13C NMR chemical shifts are summarized in Tables 3 and and44 in comparison with the available experimental data.30,31 It should be pointed out that an alkaloid such as (−)-235B′ in chloroform solution can only exist in the deprotonated state, although the protonated state could be dominant in aqueous solution. Thus, the experimental NMR chemical shifts30 should be compared to the results calculated for the deprotonated state. The assignment of the experimental NMR chemical shifts is based on the chemical shifts calculated for (−)-235B′-A, since this should be the dominant conformation, as discussed above.

Table 3
1H NMR chemical shifts (δH in ppm) of (−)-235B′ in chloroform calculated by using the GIAO method at the IEFPCM-B3LYP/6-311++G(d,p) level in comparison with available experimental data.
Table 4
13C NMR chemical shifts (δC in ppm) of (−)-235B′ in chloroform calculated calculated by using the GIAO method at the IEFPCM-B3LYP/6-311++G(d,p) level in comparison with available experimental data.

The computational results summarized in Table 3 show a remarkable difference in the pattern of the distribution of the calculated 1H NMR chemical shifts between conformations (−)-235B′-A and (−)-235B′-C, and conformations (−)-235B′-B and (−)-235B′-D. In particular, in (−)-235B′-B and (−)-235B′-D, there are four H atoms with 1H NMR chemical shifts between 2.7 and 3.3 ppm, whereas in (−)-235B′-A and (−)-235B′-C, there is only one H atom with a chemical shift between ~2.3 and ~4.8 ppm. Only the patterns corresponding to (−)-235B′-A and (−)-235B′-C are consistent with the available 1H NMR spectra30 (see Table 3 for the experimental 1H NMR chemical shifts). As seen in Table 3, only one H atom has a chemical shift (3.25 ppm) between ~2.3 and ~4.8 ppm. Thus, based on the patterns for the distribution of the calculated 1H NMR chemical shifts, conformations (−)-235B′-B and (−)-235B′-D can be excluded as being present in chloroform solution. However, both (−)-235B′-A (chair) and (−)-235B′-C (boat) conformations are possible.

In both (−)-235B′-A and (−)-235B′-C, a unique H atom with a chemical shift between ~2.3 and ~4.8 ppm, i.e. one of the two H atoms on C3 (see Figure 1), is observed. The chemical shift of Ha-3 is predicted to be 3.24 ppm in (−)-235B′-A and 3.23 ppm in (−)-235B′-C, and all are very close to the experimental chemical shift of 3.25 ppm30 or 3.23 ppm.31 The chemical shift of the other H atom (Hb-3) on C3 is predicted to be 2.03 ppm in (−)-235B′-A and 1.89 ppm in (−)-235B′-C, and all are close to the observed chemical shift of 1.80–1.97 ppm.30 The computational results clearly predict that the chemical shifts of the two geminal H atoms on C3 are considerably different, with a remarkable difference of 1.21 ppm in the (−)-235B′-A (chair) conformer and 1.34 ppm in the (−)-235B′-C (boat) conformer. This implies that the two geminal H atoms on C3 in these two conformations experience remarkably different chemical environments/magnetic shielding such that they have remarkably different chemical shifts. With this point in mind, it is not surprising to note that the difference in the chemical shift between the two geminal H atoms on C3 is dependent on the conformation of the molecule. As seen in Table 3, in the (−)-235B′-B (chair) or (−)-235B′-D (boat) conformer, the difference between the calculated chemical shifts of the two geminal H atoms on C3 is negligible (0.02 or 0.05 ppm).

The calculated 1H NMR chemical shifts can help us to better understand and correctly analyze the experimental 1H NMR spectra. The experimental 1H NMR chemical shifts were assigned by Collins et al.31 It is notable that Collins et al.31 assigned the chemical shift at 3.23 ppm to the H atom on C8a. This assignment is not supported by our computational data, since the calculated chemical shift of H-8a is only ~1.55 ppm. It is more reasonable to assign the chemical shift at 3.23 ppm31 (or 3.25 ppm in another experimental measurement30) to one of the H atoms on C3, based on the aforementioned discussion. Except for the chemical shifts of the H atoms on C3 and C8a, the assignments by Collins et al. for all the other 1H chemical shifts (i.e. at 5.77, 4.95, 2.89, 2.00, 0.86–1.95, and 0.83 ppm)31 are supported by our calculated results. It should be pointed out that the relative magnitudes of our calculated chemical shifts (1.10, 0.84, and 0.59 ppm) for the three H atoms on C9 are significantly different from those of the chemical shifts observed for these three H atoms (all at 0.83 ppm). This is because our NMR calculations were carried out on a single conformation of the structure (local minimum). Practically, any molecule will undergo thermal motions. The methyl group (consisting of C9 and its attached three H atoms) of (−)-235B′ can easily rotate during the experimental process, so that the three H atoms are statistically equivalent, unlike the two non-rotatable geminal H atoms on C3. Thus, the experimentally observed 1H NMR chemical shifts of the three H atoms of the methyl group should reflect the statistical average of the chemical shifts of the three rotatable H atoms. Consequently, it is better to compare the average value of the calculated three 1H chemical shifts with the experimental chemical shift for the three H atoms of the methyl group. The average value of the calculated chemical shifts of the three H atoms is (1.10 + 0.84 + 0.59)/3 = 0.84 ppm, which is very close to the experimental values of 0.83 ppm31 or 0.85 ppm.30

Further, we examined the linear correlation between the calculated and experimental 1H NMR chemical shifts. The obtained values of the correlation constant (r2) and root-mean-square devision (RMSD) are also summarized in the last two rows of Table 3. As seen in Table 3, the linear correlation between the calculated and experimental 1H NMR chemical shifts for (−)-235B′-A (chair) and (−)-235B′-C (boat) is significantly better than that for (−)-235B′-B (chair) and (−)-235B′-D (boat). The best linear correlation is associated with (−)-235B′-A (chair).

The predicted 13C NMR chemical shifts are summarized in Table 4. The patterns for the distribution of the 13C NMR chemical shifts calculated for all the conformations of the deprotonated state are reasonably close to that of the experimental 13C NMR chemical shifts.30 A comparison of the 13C NMR chemical shifts still does not allow one to distinguish whether the experimentally observed conformation should be (−)-235B′-A (chair) or (−)-235B′-C (boat). We also examined the linear correlation between the calculated and experimental 13C NMR chemical shifts. The obtained r2 and RMSD values are summarized in the last two rows of Table 4. Data summarized in Table 4 show that the linear correlation between the calculated and experimental 13C NMR chemical shifts for (−)-235B′-A (chair) and (−)-235B′-C (boat) is significantly better than that for (−)-235B′-B (chair) and (−)-235B′-D (boat). The best linear correlation is also associated with (−)-235B′-A (chair).

In summary, the comparison in both 1H and 13C NMR spectra between the computational results and experimental data cannot conclusively determine whether the experimentally observed conformation should be (−)-235B′-A (chair) or (−)-235B′-C (boat). These two conformations must be distinguished by use of the predicted relative free energies (discussed above). The calculated relative free energies show that the experimentally observed conformation should be (−)-235B′-A (chair), rather than (−)-235B′-C (boat). It is especially interesting to note that the computationally predicted 1H NMR chemical shifts for the two C3 geminal hydrogen atoms of (−)-235B′ in the most stable conformation are remarkably different, which helps to better understand the experimental 1H NMR spectra reported in the literature. As shown in Tables 3 and and4,4, the 1H and 13C NMR chemical shifts calculated for (−)-235B′-A are all reasonably close to the corresponding experimental data,30 according to the detailed assignments summarized in the tables.

Energetics and pKa

In addition to the determination of dominant molecular conformations, it is also interesting to predict the pKa of (−)-235B′ in aqueous solution to eventually understand the microscopic and phenomenological binding of the receptor protein with the protonated and deprotonated structures of this important open-channel blocker of α4β2 nAChRs. Knowing the pKa value will allow evaluation of the relative concentrations of the protonated and deprotonated states at any pH. The pKa of (−)-235B′ is determined by the Gibbs free energy change in aqueous solution (ΔGa) of the following protonation process:

()235BH+()235B+H+
(1)

The pKa can be evaluated via

pKa=ΔGa/(2.303RT).
(2)

Prediction of the free energy change of a deprotonation process requires knowing the absolute free energy of the proton (H+) in aqueous solution, ΔGhyd298(H+), in addition to the free energies calculated for all of the molecular species (i.e. protonated and deprotonated states) discussed above. Due to the inherent difficulty of measuring absolute solvation free energy of an ion, the reported “experimental” ΔGhyd298(H+) values have a wide range from −252.6 to −264.1 kcal/mol.32 We predicted ΔGhyd298(H+) (by using a high-level, ab initio method of incorporating a hybrid supermolecule-continuum approach based on the same SVPE procedure used in the present study) to be -262.4 kcal/mol.40b The same computational protocol has been used to calculate the pKa for cocaine, nicotine, and other organic amines, and the calculated pKa values are all in good agreement with the corresponding experimental data. Thus, the pKa of (−)-235B′ in aqueous solution was calculated to be ~9.7 in the present study. The calculated pKa value of ~9.7 predicts that the concentration of the protonated state of (−)-235B′ should be higher than that of the deprotonated state when the pH < ~9.7, but the concentration of the deprotonated state of (−)-235B′ should be higher than that of the protonated state when the pH > 9.7. Particularly, in aqueous solution at a physiological pH (7.4), ~99.5% of (−)-235B′ should exist in the (−)-235B′H+-A conformer, while only ~0.5% of (−)-235B′ should exist in the (−)-235B′-A conformer. The other conformations (B to D) are negligible for both the protonated and deprotonated states in aqueous solutions. Further computational studies on (−)-235B′ binding with nAChRs may be focused on the (−)-235B′H+-A conformer of the protonated species and the (−)-235B′-A conformer of the deprotonated species.

Computational Methods

In order to search for stable molecular geometries, we first needed to build various possible initial structures of (−)-235B′. The initial structures of the deprotonated state used in the geometry optimizations were built in consideration of various possible chair and boat conformations. Geometries of all structures considered in this study were optimized by using gradient-corrected density functional theory (DFT) with the B3LYP functional33 and 6-31+G(d) basis set.34 The optimized geometries were used to carry out the harmonic frequency analysis at the same B3LYP/6-31+G(d) level, to make sure that each of the optimized geometries is associated with a local minimum on the potential energy surface, and to evaluate the zero-point and thermodynamic corrections to the Gibbs free energy. The geometries optimized at the B3LYP/6-31+G(d) level were also used to carry out single-point energy calculations using larger basis sets at the B3LYP/6-31++G(d,p), B3LYP/6-311++G(d,p), and B3LYP/6-311++G(2d,2p), levels, in order to make sure that the basis set used in this study is adequate.

The geometries optimized at the B3LYP/6-31+G(d) level were also used to perform self-consistent reaction field (SCRF) energy calculations in aqueous solution, in order to determine the solvent shifts. Florian and Warshel35 performed a manual geometry search in solution along the corresponding gas phase intrinsic reaction coordinate for the hydrolysis of mono-methyl phosphate, and demonstrated that the contributions of the solvent-induced structural changes to the overall energetics are small and can be safely neglected. The calculated free energies in aqueous solution were taken as the energies calculated in the gas phase with the B3LYP/6-31+G(d) zero-point and thermal corrections (at 298 K and 1 atm) plus the corresponding solvent shifts determined by the SCRF calculations. The main SCRF method used in this study for the energy calculations is known as the surface and volume polarizations for electrostatics (SVPE).36,37,38,39 This SVPE method is also known as the fully polarizable continuum model (FPCM),40,41 because it fully accounts for both surface and volume polarization effects on the solute-solvent electrostatic interaction. The SVPE method, which was initially developed in the GAMESS program42 by one of us with Chipman and Bentley,36 has been implemented recently in our local version43 of the Gaussian03 program.44 Some advantages of the SVPE method over other SCRF methods can be found in our previous reports concerning reaction field calculations using various SCRF methods on the energy barriers of a series of carboxylic acid esters39 and phosphate esters, including a simplified cAMP model.45 The SVPE calculations in solvent water (with dielectric constant ε = 78.304) or chloroform (ε = 4.9) were carried out at the HF/6-31+G* level, as it has been shown46 that electron correlation does not significantly affect the calculated solvent shifts.

In addition, we also employed the geometries optimized at the B3LYP/6-31+G(d) level to carry out nuclear magnetic shielding calculations at the B3LYP/6-311++G(d,p) level using the Gauge-Independent Atomic Orbital (GIAO) method47,48,49,50 in combination with the integral equation formalism of the polarizable continuum model (IEFPCM)51 implemented in the Gaussian03 program.44 The IEFPCM method used in the GIAO calculations accounted for the effects of solvent chloroform (ε = 4.9) on the nuclear magnetic shielding tensors, as experimental NMR spectra available for (−)-235B′ were measured in chloroform. In order to determine the isotropic 1H and 13C NMR chemical shifts, we also carried out the GIAO calculations on the reference compound, i.e. tetramethylsilane (TMS), at the same IEFPCM-B3LYP/6-311++G(d,p) level.

All of the electronic structure calculations were carried out using the Gaussian03 program on an IBM X-series Cluster (with 340 nodes or 1,360 processors) at the University of Kentucky Center for Computational Sciences, and on a 34-processors IBM Linux Cluster in our own laboratory.

Conclusion

Extensive first-principles electronic structure calculations have led to a determination of the stable molecular conformations and the corresponding relative free energies of the protonated and deprotonated states of indolizidine (−)-235B′ in the gas phase, in chloroform, and in aqueous solution. The 1H and 13C NMR chemical shifts calculated using the computationally determined dominant molecular conformation of the deprotonated state are all consistent with available experimental NMR spectra of (−)-235B′ in chloroform, which suggests that the computationally determined molecular conformations are reasonable.

The calculated NMR chemical shifts in comparison with available experimental NMR data also help one to better understand and analyze the experimental NMR spectra of (−)-235B′. In particular, the available experimental 1H NMR spectra showed only one H atom with a chemical shift (3.23 or 3.25 ppm) between ~2.1 and ~4.8 ppm. This unique 1H chemical shift was assigned to the H atom on C8a (see Figure 1 for the structure numbering), in previously reported experimental studies. However, our calculated chemical shift of H-8a is only ~1.55 ppm. Our computational results reveal that the two geminal H atoms on C3 experience distinctly different magnetic shielding and, thus, have remarkably different chemical shifts (i.e. 3.24 and 2.03 ppm). Hence, the chemical shift at 3.23 ppm (or 3.25 ppm in another experimental measurement) should be assigned to one of the geminal H atoms on C3, rather than the H atom on C8a. It has also been demonstrated that the difference in the chemical shift between the two geminal H atoms on C3 is critically dependent on the conformation of (−)-235B′, as the difference becomes negligible in some other high-energy conformations based on our calculated results. The finding of the remarkably different chemical shifts of these geminal H atoms in a certain molecular conformation may also be valuable in understanding and analyzing the NMR spectra of other structurally related ring-containing compounds.

In addition, the pKa of (−)-235B′ in aqueous solution has been predicted to be ~9.7. All of the computational results obtained in this study provide a solid basis for future comprehensive studies on the microscopic and phenomenological binding of various receptor proteins with the protonated and deprotonated species of this potent and selective open-channel blocker of α4β2 nAChRs.

This study not only helps to understand the detailed structures and conformations of (−)-235B′ in solution, but also demonstrates how one can appropriately use computational modeling and spectroscopic analysis to address the structural and spectroscopic problems that cannot be addressed by experiments alone.

Acknowledgments

The research was supported in part by the Kentucky Science & Engineering Foundation (grant KSEF-925-RDE-008) and the National Institutes of Health (grant U19DA017548). The authors also acknowledge the Center for Computational Sciences (CCS) at University of Kentucky for supercomputing time on the IBM X-series Cluster (with 340 nodes or 1,360 processors).

Notes and References

1. Schneider G, Baringhaus K-H. Molecular Design: Concepts and Applications. Wiley-Vch Verlag GmbH & Co; KGaA, Weinheim: 2008. p. 16.
2. Daly JW, Spande TF, Garraffo HM. J Nat Prod. 2005;68:1556. [PubMed]
3. Carroll FI. Bioorg Med Chem Lett. 2004;14:1889. [PubMed]
4. Bunnelle WH, Daanen JF, Ryther KB, Schrimpf MR, Dart MJ, Gelain A, Meyer MD, Frost JM, Anderson DJ, Buckley M, Curzon P, Cao YJ, Puttfarcken P, Searle X, Ji J, Putman CB, Surowy C, Toma L, Barlocco D. J Med Chem. 2007;50:3627. [PubMed]
5. Michael JP. Beilstein J Org Chem. 2007;3:27. [PMC free article] [PubMed]
6. Michel P, Rassat A, Daly JW, Spande TF. J Org Chem. 2000;65:8908. [PubMed]
7. Barluenga J, Mateos C, Aznar F, Valdes C. Org Lett. 2002;4:1971. [PubMed]
8. Marsden SP, McElhinney AD. Beilstein J Org Chem. 2008;4:8. [PMC free article] [PubMed]
9. Toyooka N, Tsuneki H, Kobayashi S, Dejun Z, Kawasaki M, Kimura I, Sasaoka T, Nemoto H. Curr Chem Biology. 2007;1:97.
10. Tsuneki H, You Y, Toyooka N, Kagawa S, Kobayashi S, Sasaoka T, Nemoto H, Kimura I, Dani JA. Mol Pharmacol. 2004;66:1061. [PubMed]
11. Edwards MW, Daly JW. J Nat Prod. 1988;51:1188. [PubMed]
12. Wonnacott S, Sidhpura N, Balfour DJK. Curr Opin Pharmacol . 2005;5:53. [PubMed]
13. Karlin A. Nat Rev Neurosci. 2002;3:102. [PubMed]
14. Miyazawa A, Fujiyoshi Y, Unwin N. Nature. 2003;423:949. [PubMed]
15. Karlin A. Neuron. 2004;41:841. [PubMed]
16. Lester HA, Dibas MI, Dahan DS, Leite JF, Dougherty DA. Trends Neurosci. 2004;27:329. [PubMed]
17. Kihara T, Shimohama S. Acta Neurobiol Exp. 2004;64:99. [PubMed]
18. Schmaljohann J, Gündisch D, Minnerop M, Bucerius J, Joe A, Reinhardt M, Guhlke S, Biersack H, Wüllner U. Nuclear Med Biology. 2006;33:305. [PubMed]
19. Arneric SP, Holladay M, Williams M. Biochem Pharmacol. 2007;74:1092. [PubMed]
20. Tutka P. Expert Opin Investig Drugs. 2008;17:1473. [PubMed]
21. Zheng F, Zheng G, Deaciuc AG, Zhan CG, Dwoskin LP, Crooks PA. J Enzy Inhib Med Chem. 2009;24:157. [PubMed]
22. Zheng F, Bayram E, Sumithran SP, Ayers JT, Zhan CG, Schmitt JD, Dwoskin LD, Crooks PA. Bioorg Med Chem. 2006;14:3017. [PubMed]
23. Tonder JE, Olesen PH. Curr Med Chem. 2001;8:651. [PubMed]
24. Jensen AA, Frølund B, Liljefors T, Krogsgaard-Larsen P. J Med Chem. 2005;48:4705. [PubMed]
25. Huang X, Zheng F, Crooks PA, Dwoskin LP, Zhan CG. J Am Chem Soc. 2005;127:14401. [PMC free article] [PubMed]
26. Huang X, Zheng F, Chen X, Crooks PA, Dwoskin LP, Zhan CG. J Med Chem. 2006;49:7661. [PubMed]
27. Huang X, Zheng F, Stokes C, Papke RL, Zhan CG. J Med Chem. 2008;51:6293. [PMC free article] [PubMed]
28. Huang X, Zheng F, Zhan CG. J Am Chem Soc. 2008;130:16691. [PubMed]
29. Witala KW, Cramer CJ, Hoye TR. Magn Reson Chem. 2007;45:819. [PubMed]
30. Toyooka N, Tanaka K, Momose T. Tetrahedron. 1997;53:9553.
31. Collins I, Fox ME, Holmes AB, Williams SF, Baker R, Forbes IJ, Thompson M. J Chem Soc Perkin Trans. 1991;1:175.
32. Meijias JA, Lago S. J Chem Phys. 2000;113:7306.
33. (a) Becke AD. J Chem Phys. 1993;98:5648. (b) Lee C, Yang W, Parr RG. Phys Rev B. 1988;37:785. [PubMed] (c) Stephens PJ, Devlin FJ, Chabalowski CF, Frisch MJ. J Phys Chem. 1994;98:11623.
34. Hehre WJ, Radom L, Schleyer PvR, Pople JA. Ab Initio Molecular Orbital Theory. John Wiley & Sons; New York: 1987.
35. Florian J, Warshel A. J Phys Chem B. 1998;102:719.
36. Zhan CG, Bentley J, Chipman DM. J Chem Phys. 1998;108:177.
37. Zhan CG, Chipman DM. J Chem Phys. 1998;109:10543.
38. Zhan CG, Chipman DM. J Chem Phys. 1999;110:1611.
39. Zhan CG, Landry DW, Ornstein RL. J Phys Chem A. 2000;104:7672.
40. (a) Zhan CG, Niu S, Ornstein RL. J Chem Soc Perkin Trans 2. 2001;1:23. (b) Zhan CG, Dixon DA. J Phys Chem A. 2001;105:11534. (c) Dixon DA, Feller D, Zhan CG, Francisco JS. J Phys Chem A. 2002;106:3191. (d) Zheng F, Zhan CG, Ornstein RL. J Phys Chem B. 2002;106:717. (e) Zhan CG, Dixon DA. J Phys Chem A. 2002;106:9737. (f) Zhan CG, Dixon DA, Sabri MI, Kim MS, Spencer PS. J Am Chem Soc. 2002;124:2744. [PubMed] (g) Zhan CG, Dixon DA. J Phys Chem A. 2003;107:4403. (h) Zhan CG, Dixon DA, Spencer PS. J Phys Chem B. 2003;107:2853. (i) Dixon DA, Feller D, Zhan CG, Francisco SF. Int J Mass Spectrom. 2003;227:421.
41. (a) Lu HT, Chen X, Zhan CG. J Phys Chem B. 2007;111:10599. [PubMed] (b) Xiong Y, Zhan CG. J Phys Chem A. 2006;110:12644. [PubMed] (c) Zhan CG, Deng SX, Skiba JG, Hayes BA, Tschampel SM, Shields GC, Landry DW. J Comput Chem. 2005;26:980. [PubMed] (d) Huang X, Zheng F, Crooks PA, Dwoskin LP, Zhan CG. J Am Chem Soc. 2005;127:14401. [PubMed] (e) Xiong Y, Zhan CG. J Org Chem. 2004;69:8451. [PubMed] (f) Chen X, Zhan CG. J Phys Chem A. 2004;108:3789. (g) Zhan CG, Dixon DA. J Phys Chem A. 2004;108:2020.
42. Schmidt MW, Baldridge KK, Boatz JA, Elbert ST, Gordon MS, Jensen JH, Koseki S, Matsunaga N, Nguyen KA, Su SJ, Windus TL, Dupuis M, Montgomery JA. J Comput Chem. 1993;14:1347.
43. Vilkas MJ, Zhan CG. J Chem Phys. 2008;129:194109. [PubMed]
44. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, Montgomery JA, Jr, Vreven T, Kudin KN, Burant JC, Millam JM, Iyengar SS, Tomasi J, Barone V, Mennucci B, Cossi M, Scalmani G, Rega N, Petersson GA, Nakatsuji H, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Klene M, Li X, Knox JE, Hratchian HP, Cross JB, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Ayala PY, Morokuma K, Voth GA, Salvador P, Dannenberg JJ, Zakrzewski VG, Dapprich S, Daniels AD, Strain MC, Farkas O, Malick DK, Rabuck AD, Raghavachari K, Foresman JB, Ortiz JV, Cui Q, Baboul AG, Clifford S, Cioslowski J, Stefanov BB, Liu G, Liashenko A, Piskorz P, Komaromi I, Martin RL, Fox DJ, Keith T, Al-Laham MA, Peng CY, Nanayakkara A, Challacombe M, Gill PMW, Johnson B, Chen W, Wong MW, Gonzalez C, Pople JA. Gaussian 03, Revision A.1. Gaussian, Inc; Pittsburgh, PA: 2003.
45. Chen X, Zhan CG. J Phys Chem A. 2004;108:6407.
46. Zhan CG, Landry DW, Ornstein RL. J Am Chem Soc. 2000;122:2621.
47. McWeeny R. Phys Rev. 1962;126:1028.
48. Ditchfield R. Mol Phys. 1974;27:789.
49. Dodds JL, McWeeny R, Sadlej AJ. Mol Phys. 1980;41:1419.
50. Wolinski K, Hilton JF, Pulay P. J Am Chem Soc. 1990;112:8251.
51. (a) Cances E, Mennucci B, Tomasi J. J Chem Phys. 1997;107:3032. (b) Mennucci B, Cances E, Tomasi J. J Phys Chem B. 1997;101:10506. (c) Cances E, Mennucci B. J Chem Phys. 1998;109:249. (d) Cances E, Mennucci B. J Math Chem. 1998;23:309. (e) Cances E, Mennucci B, Tomasi J. J Chem Phys. 1998;109:260. (f) Tomasi J, Mennucci B, Cances E. J Mol Struct THEOCHEM. 1999;464:211. (f) Amovilli C, Barone V, Cammi R, Cances E, Mennucci B, Pomelli C, Tomasi J. Adv Quantum Chem. 1999;32:227.