Enantiodifferentiation of chiral baclofen by β-cyclodextrin using capillary electrophoresis: A molecular modeling approach

Journal of Molecular Structure 1019 (2012) 43–49

Contents lists available at SciVerse ScienceDirect

Journal of Molecular Structure journal homepage: www.elsevier.com/locate/molstruc

Enantiodifferentiation of chiral baclofen by b-cyclodextrin using capillary electrophoresis: A molecular modeling approach FakhrEldin O. Suliman a,⇑, Abdalla A. Elbashir b a b

Department of Chemistry, College of Science, Sultan Qaboos University, Box 36, Al-Khod 123, Oman Department of Chemistry, Faculty of Science, University of Khartoum, Box 321, Khartoum, Sudan

a r t i c l e

i n f o

Article history: Received 16 February 2012 Received in revised form 26 March 2012 Accepted 26 March 2012 Available online 3 April 2012 Keywords: Chiral separation Capillary electrophoresis Baclofen Molecular modeling PM6 b-Cyclodextrin

a b s t r a c t Using capillary electrophoresis baclofen (BF) enantiomers were separated only in the presence of b-cyclodextrin (bCD) as a chiral selector when added to the background electrolyte. Proton nuclear magnetic resonance and electrospray ionization mass spectrometry (ESI–MS) techniques were used to determine the structure of the BF–bCD inclusion complexes. From the MS data BF was found to form a 1:1 complex with a- and bCD, while the NMR data suggest location of the aromatic ring of BF into the cyclodextrin cavity. A molecular modeling study, using the semiempirical PM6 calculations was used to investigate the mechanism of enantiodifferentiation of BF with cyclodextrins. Optimization of the structures of the complexes by PM6 method indicated that separation is obtained in the presence of b-CD due to a large binding energy difference (DDE) of 46.8 kJ mol1 between S-BF–bCD and R-BF–bCD complexes. In the case of aCD complexes DDE was 1.3 kJ mol1 indicating poor resolution between the two enantiomers. Furthermore, molecular dynamic simulations show that the formation of more stable S-BF–bCD complex compared to R-BF–b-CD complex is primarily due to differences in intermolecular hydrogen bonding. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Chiral recognition is attracting increasing attention owing to the enormous importance in many fields such as drug discovery, life sciences, food science, and agrochemicals and also in environmental studies. Of a particular importance is the significant impact of chirality in pharmaceutical research. More than half of the currently developed drugs possess one or more chiral centers. Moreover, the majority of these drugs are prescribed and used as racemates [1]. The general perception is that one enantiomer may possess the desired pharmacological effects while the other is of a lower potency or even inactive. However, for many drug candidates, the effect of individual enantiomers on the pharmacodynamics or the pharmacokinetics remains unraveled. Since the 1980s capillary electrophoresis (CE) techniques have emerged as powerful tools in enantiomeric separation due to their simplicity, high efficiency and versatility [2]. In CE chiral selectors of various types are added to the background electrolyte (BGE). Electrophoretic methods usually separate charged species in an electric field due to differences in sizes and charges which will lead to different mobilities [2,3]. Enantiomers of a chemical species possess similar charges and the same size, therefore are expected not to be separated by electrophoretic techniques. ⇑ Corresponding author. Tel.: +968 24141480; fax: +968 24141469. E-mail address: [email protected] (F.O. Suliman). 0022-2860/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.molstruc.2012.03.055

To achieve separation of enantiomers in CE chiral selectors are added to the BGE and chiral recognition is obtained by formation of transient reversible complexes between both enantiomer species and the chiral selector [2,4–6]. Enantioseparation can then be achieved when the affinities of individual enantiomers towards the chiral selector are different or when the diastereomers’ mobilities are different. These two mechanisms, the binding selectivity and the complex mobility selectivity, may counteract each other or work cooperatively. However, the complex binding selectivity is the most predominant one. Cyclodextrins (CDs) are optically active molecules that can form diastereomeric pairs by complexation with racemic mixtures through host–guest mechanism. An arsenal of cyclodextrin hosts has found extensive use in the chiral separations and in many chromatographic systems as chiral stationary phases or chiral mobile phase additives [7,8]. In CE cyclodextrins are considered an important class of chiral selectors used as additives to BGE. Baclofen (BF), 4-amino-3-(4-chlorophenyl) butyric acid (Fig. 1) is a c-aminobutyric acid analog and is extensively used as a stereoselective agonist for GABAB receptor [9,10]. BF has also been used as a muscle relaxant, and to treat spasticity due to multiple sclerosis, cerebral and spinal cord injury, cerebral palsy, and complex region pain syndrome [9,11,12]. BF is usually prescribed and used as a racemic mixture. However, it is claimed that only the R-enantiomer is stereoselectively active on GABAB-receptors and is more active than the S-enantiomer [13,14].

44

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

Fig. 1. Chemical structure of (a) Baclofen (BF) and (b) cyclodextrin, n = 6aCD, n = 7bCD.

A number of chiral separation protocols were developed for the analysis of BF. In these methods various chiral selectors were used, including neutral native and modified CDs, charged CDs, chiral crown ethers, and a mixture of crown ethers and CDs [15–20]. In all these studies the mechanism underlying the chiral separation of BF enantiomers was not discussed. Recently, there has been an increase in interest in molecular modeling studies on the formation and stability of inclusion complexes of cyclodextrins with a variety of molecules and other aspects of supramolecular chemistry. Various theoretical approaches have been applied in these studies such as molecular mechanics, molecular dynamics, semiempirical methods, as well as hybrid techniques such as quantum-mechanics–molecular mechanics (QM–MM) techniques [21–28]. Hartree–Fock (HF) and density functional theory (DFT) calculations especially using the popular B3LYP functional combined with different standard basis set have been used reliably to describe host guest interactions of CDs with several molecules [29–32]. Despite this rapid development and use of ab initio and DFT semiempirical methods still attract great deal of attention owing to their less computational demands. The recently introduced semiempirical PM6 method has been found to give more accurate estimates of molecular properties comparable to HF and DFT methods at even lower computational cost, making it an attractive method for the description of inclusion complexes [33]. Of special interest to molecular modeling studies are those associated with the enantioseparations of chiral molecules using the supramolecular chemistry. These studies aimed to rationalize and predict the experimental results obtained using different chromatographic separation techniques such as capillary electrophoresis, gas chromatography and liquid chromatography. In the present study, we investigate the inclusion complexation of baclofen with aCD and bCD and their chiral recognition properties using NMR and ESI–MS techniques. Furthermore, to investigate and predict the enantiodifferentiation of baclofen by the two cyclodextrins we employed molecular simulations using molecular mechanics methods with Autodock to determine the mode of inclusion of BF within the CD. The obtained structures were further optimized by the semiempirical method PM6 to obtain the binding energies of the studied inclusion complexes. We also employed molecular dynamic (MD) simulation for ten nanosecond for R- and S-BF–bCD complexes in aqueous media to further investigate the hydrogen bonding patterns of these complexes in aqueous media to understand the mechanism of separation. 2. Experimental 2.1. Materials and reagents Phosphoric acid (85% w/w), sodium dihydrogen phosphate and the racemic mixture and the optically active forms of the baclofen were purchased from Sigma–Aldrich (St. Louis, MO, USA). b-Cyclodextrin (bCD) and a-cyclodextrin (aCD) were obtained from Fluka

(Buchs, Switzerland). The phosphate buffer solutions were prepared by using a 50 mM sodium dihydrogen phosphate solution and adjusted to desired pH with phosphoric acid. All samples were prepared in aqueous solution at a concentration of 0.5 mg mL1. The concentrations of the chiral selectors bCD, and aCD were prepared by dissolving the pure materials in phosphate buffer. 2.2. CE instrument and conditions Analytical separations were carried out on a Waters Capillary Ion Analyzer (Milford, MA, USA) which was interfaced to a Waters PC 800 Workstation using an uncoated fused-silica capillary (total length, 35 cm and internal diameter, 75 lm; effective length, 27.5 cm). The separations were conducted at 25 °C by applying a voltage of 15 kV. Samples were injected hydrodynamically for 10 s by applying a pressure of 50 mbar. Detection was performed at 254 nm. New uncoated fused-silica capillary was conditioned by flushing with 1 M NaOH for 30 min, then 0.1 M NaOH for 10 min and finally water and buffer each for 15 min. The running buffer solution was passed through 0.2 lm cellulose nitrate membrane filters (Whatman, UK) and degassed by sonication prior to use. Prior to each analysis, the capillary column was rinsed with 0.1 M NaOH for 2 min, and then Milli-Q water, followed by the carrier electrolyte, each for 3 min between the runs. 2.3. NMR studies All 1H NMR spectra are recorded on a Bruker-400 instrument in D2O at 23 °C. Chemical shifts (d) are reported in ppm with the residual HOD signal being used as the internal reference. Sample solutions were prepared by dissolving BF and CDs in D2O to obtain the required stock solutions. In a typical experiment, 1.0 mL of CD was mixed with increasing amounts of BF to cover the range of BF concentration 1–20 mM while keeping that of CD constant at 10 mM. 2.4. ESI–MS A Quattro Ultima LC–MS system (Waters Corporation, Milford, USA) equipped with electrospray ionization (ESI) interface was used. The MS conditions were as follows: source temperature 100 °C, desolvation temperature 200 °C, cone voltage 35 V, and capillary voltage setting 3.0 kV. The low and high mass resolution were set at 13.0 (arbitrary units), and the multiplier was set at 650 V. In each run 1.0 mL of CD was mixed with different amounts of BF in the range 1–5 mM while keeping that of CD constant at 5 mM, appropriate dilutions of the solutions mixtures were prepared using acetate buffers. 2.5. Molecular modeling The starting structure of BF was constructed by Chem3D Ultra (Cambridgesoft.com) and was fully optimized using the semiem-

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

pirical method PM6 using the mopac2009 package (www.mopac.net). Structures of a- and b-cyclodextrin were extracted from the crystallographic parameters provided by the Structural Data Base System of the Cambridge crystallographic data center. Both CDs were fully optimized with the semiempirical method PM6. The chemical structures and the atom labeling of BF and cyclodextrins is given in Fig. 1. Thereafter, molecular docking studies were undertaken using Autodock program (version 4.2) [34]. We used a Lamarckian genetic algorithm (LGA) for the docking study to generate the inclusion complexes. Autodock defines the conformational space implementing grids over all the possible search space. Grid maps of 40  40  40 Å grid box with 0.375 Å spacing were obtained using Autogrid 4 program. The center of mass of the CD was set as the center of the box. The initial torsions and positions of BF were generated randomly. With the help of Autodock tools the partial charges were calculated using the Gasteiger–Marsili method [35,36]. For the search we used a population of 250 LGA runs with a maximum number of energy evaluations of 2.5  107 and a maximum number of generations of 27,000. An elitism value of 1 was used and a probability of mutation and cross-over of 0.02 and 0.08 were used respectively. At the end of each run the solutions are separated into clusters based on their lowest root mean square deviation (RMSD) and the best score based on a free energy function. Cluster solutions with average scores that are over 1.0 kcal mol1 with respect to the best energy obtained in the respective run were selected. Based on the most predominant conformations the selected final complexes were further optimized using the semiempirical PM6 method. Amber 11 software package was used for the molecular dynamic (MD) simulations using the general force field parameter set [37]. The Sander module of Amber 11 package was used to perform MD simulations on structure of the complexes obtained by the aforementioned docking procedures. Each system was solvated in a truncated octahedral box of TIP3P water molecules [38] with a minimum solute–water distance of 10 Å. The solvated complexes were equilibrated and energy minimized prior to MD simulations. Long-range electrostatic interactions were dealt with by particle mesh Ewald (PME) method [39]. Non-bonded cutoff of 10 Å was used to treat noncovalent interactions. Solvated complexes were heated to 300 K in 50 ps followed by an equilibration step of 600 ps at 300 K and 1 atm. The production step was run for an additional ten nanoseconds. Periodic boundary conditions were adopted in NPT ensabmle coupled to Brendsen thermostat at 300 K and 1 atm with isotropic molecule based scaling [40]. The time step for MD simulation was 1 fs. The SHAKE algorithm [41] was used to constrain all covalent bonds involving hydrogen atmos. Analysis of MD trajectories generated was performed by Ptraj module in Amber 11 [42]. For hydrogen bond analysis, a hydrogen bond cut distance 63.0 Å and angle P120° were used. The computations were conducted at Bahla high performance computing facility at Sultan Qaboos University.

3. Results and discussion 3.1. Electrophoretic separation of baclofen enantiomers Baclofen (BF) is an amphoteric molecule and has pKa values of 3.7 and 9.5 [43–45]. The chiral separation was investigated at pH 7.0. Two chiral selectors, aCD and bCD, were used as buffer additives for the chiral separation of BF enantiomers. When aCD was used as running buffer additive no separation was obtained for the BF at all tested concentrations (5–20 mM). However, a chiral separation of BF was achieved when bCD was used as a running buffer additive as illustrated in Fig. 2.

45

The pH of the running buffer is an important parameter and need to be carefully adjusted as it affects the ionization of the silanol group of the capillary wall, which in turn affects the magnitude of the electroosmotic flow (EOF). The pH of the media also influences the ionization of BF and in many cases a compromise need to be made. In this study, the effect of pH on the resolution over the range 3–9 was investigated. At pH 3.0 no separation was observed and it was noticed that the two enantiomers start to separate at pH 4.0 and the resolution increases with the increase in pH up to 7.0 and then it starts to decrease. The best separation was obtained at pH 7.0 where a resolution of 1.2 was obtained. Similar results were obtained by Ali and Aboul-Enein [19]. 3.2. 1H NMR studies The formation of inclusion complexes usually associate with a significant change of the chemical environment of guest molecules with a subsequent effect on the chemical shift values of both host and guest protons. Therefore, changes in chemical shift (Dd) of proton NMR resonance provides a direct evidence for the formation of inclusion complexes [46,47]. Proton NMR studies were performed in order to determine the mode of inclusion of BF in bCD. The Dd values of the protons in bCD and BF (between the free and complexed forms in D2O) are shown in Table 1. In the presence of BF the signals of protons located inside the cavity of bCD (H3 and H5) are consistently shifted upfield which demonstrate a clear participation of these hydrogen atoms in the guest–host interactions. This upfield shift is due to anisotropic effects caused by the presence of a rich p-electrons group of the guest inside the CD cavity. It is clear from Table 1 that these signals are significantly influenced by addition of BF compared to the protons outside the cyclodextrin cavity. These observations suggest penetration of BF molecule into the cyclodextrin and insertion of the aromatic ring of BF into the cavity. Furthermore, signals originating from aromatic protons of BF (Ha and Hb) have shown large Dd values exhibiting downfield shifts. This downfield shift is probably associated with the changes in the local polarity around the guest molecule or to the deshielding effects exerted by van der Waal’s interactions between aromatic group of BF molecule and the macrocycle cavity inner walls. Other protons of BF exhibited no significant shifts upon complexation with bCD. These findings suggest that the aromatic ring is located inside bCD cavity leaving the amine and the carboxyl groups just exposed to the solvent via the wider side of the host molecule. 3.3. ESI–MS The ESI mass spectra for a mixture of BF with a- and bCD are shown in Fig. 3. The peaks at 995 correspond to [aCD + Na]+. BF exhibits a clear peak at m/z 214 attributed to [BF + H]+. Additionally, a peak at 427 is observed in both spectra of mixtures of aand bCD with BF indicating the formation of the dimeric ion [2BF + H]+ favored by high concentration of the solute. In Fig. 3a the peaks at m/z 1186 and m/z 1208 represents [aCD–BF + H]+ and [aCD–BF + Na]+ suggesting the formation of a 1:1 inclusion complex between BF and aCD. Similarly, in Fig. 3b the ions at m/ z 1371 is attributed to [bCD–BF + Na]+. It is clear that BF also forms a 1:1 host:guest complexes with bCD. 3.4. Molecular modeling To have a further insight into the enantiodifferentiation of BF by bCD and to rationalize our experimental results, we performed molecular docking of the inclusion of BF into aCD and bD cavities using Autodock 4.2 [34]. We performed cluster analysis for all complexes using a cutoff of 1.0 Å root mean square deviation (RMSD).

46

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

Fig. 2. Electropherogram of separation of baclofen with bCD. Conditions: 15 mM bCD, 50 mM sodium dihydrogen phosphate pH 7.0, injection time 10 s; applied voltage 15 kV; and capillary temperature 25 °C.

Table 1 Chemical shift changes of signals of some protons of BF and bCD upon formation of inclusion complex for different ratios of concentrations of BF to bCD. [BF]/[bCD]

0.16 0.64 0.96 1.6

Dd H2

H3

H4

H5

H6

Ha (BF)

Hb (BF)

0.001 0.002 0.004 0.006

0.008 0.032 0.058 0.061

0.001 0.002 0.008 0.011

0.008 0.016 0.055 0.053

0.000 0.006 0.004 0.003

0.083 0.140 0.177 0.192

0.034 0.069 0.091 0.097

Fig. 3. Electrospray ionization mass spectra of baclofen in the presence of (a) aCD and (b) bCD at pH 7.

In all cases the lowest energy structures of BF–CD correspond to the cluster with the maximum number of BF conformation indicating convergence of the docking procedure. The binding free energies of the R- and S-BF complexes with aCD obtained from Autodock were 21.2 and 21.3 kJ mol1 respectively whereas those obtained for R- and S-BF complexes with bCD were 19.5 and 18.9 kcal mol1 respectively. The predominant conformations of the inclusion complexes obtained for both enantiomers with a- and bCD are those in which the aromatic ring penetrates into the cavity through the secondary hydroxyl group side. Further

we evaluated the docking results using a previously reported probability score [22], which confirmed the aforementioned results (Table S4). The differences in binding energies obtained by Autodock are insignificant compared to the standard errors reported for the docking simulations; therefore further calculations are required. We then performed quantum mechanical calculations on the inclusion complexes obtained by the docking procedures. Here we used the semiempirical method PM6 to understand the molecular recognition between BF and cyclodextrins, and to estimate more accurately the energies of the inclusion complexes. The

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

47

calculated binding energies (DE) associated with the formation of the 1:1 BF–CD complexes are obtained using following equation:

DE ¼ Ecomp  ðEBF þ ECD Þ

ð1Þ

where Ecomp, EBF, and ECD are the energies of BF–CD inclusion complex, free guest and the free host molecules respectively. The results of the interaction energies of BF with the aCD and bCD obtained by the PM6 semiempirical method are shown in Table 2. It is apparent from these results that all complexes of BF enantiomers with aCD and bCD possess energies that are always lower than the sum of the isolated guest and host molecules. This is indicative of formation of stable complexes between R- and S-BF with these cyclodextrins. The optimized structures of these inclusion complexes of BF with aCD and bCD are presented in Figs. 4 and 5 respectively. Inspection of these figures reveals that BF molecule penetrates into CD nanocavity in both cyclodextrins. The hydrophobic phenyl moiety locates itself into the middle of the cavity while BF polar amino and the

Table 2 Interaction energies and thermodynamic properties of BF–CD inclusion complexes. Parameter

R-BF/aCD

S-BF/aCD

R-BF/bCD

S-BF/bCD

E (kJ mol1) DE (kJ mol1) DDE (kJ mol1) DH (kJ mol1) DS (J mol1 K1) DG (kJ mol1)

5503.5 128.3 1.3 132.3 310.4 39.7

5500.0 127.1

6451.4 131.8 46.8 131.2 243.2 58.6

6496.1 178.5

129.3 285.2 44.3

181.8 295.5 93.8

DDE = DER  DES, negative DDE implies that R isomer is eluted first. DA = Acom ACD  Aguest, A = H, S; DG = DH  TDS.

Fig. 5. Geometries of the most favorable inclusion complexes of (a) R-BF–bCD and (b) S-BF–bCD.

Fig. 4. Geometries of the most favorable inclusion complexes of (a) R-BF–aCD and (b) S-BF–aCD.

carboxylic groups adopt a position onto secondary hydroxyl rim of the CD-cavity to maximize favorable interactions. The quantum mechanical calculations show that R-BF–aCD and S-BF–aCD possess almost similar binding energies of 128.3 and 127.1 kJ mol1 respectively. This result suggests that aCD as chiral selector is incapable of the enantiodifferentiation of BF racemates. This motif characterized by a marginal difference in energy (DDE = 1.3 kJ mol1) between the inclusion complexes of R-BF and S-BF with aCD is in line with the results obtained from the CE experimental results. The optimized structures of BF enantiomers with bCD are illustrated in Fig. 5. It is evident from this figure that the guest molecules adopt similar geometries inside bCD cavity as in aCD cavity. It is clear from Table 2 that inclusion complex of S-BF with bCD are more stable than the R-BF–bCD inclusion complex by a formidable energy difference of 46.8 kJ mol1. This result also indicates that bCD is an excellent chiral selector capable of separation of BF enantiomers. This result is in agreement with the experimental results obtained by CE. We also performed statistical thermodynamic calculations at 1 atm and 298.15 K using PM6 methods. The enthalpy change (DH), the entropy change (DS) and the Gibbs free energy change (DG) are calculated and listed in Table 2. It is clear from this table that the complexation process between BF and the two cyclodextrins is an exothermic process as suggested by the large negative DH values. On the other hand the enthalpy change for R-BF–aCD complex is only 4 kJ mol1 higher than that of S-BF–aCD inclusion complex. This indicates that the stabilities of the two diasteromers are comparable suggesting that there is little or no chance for their separation using aCD as chiral selector. For the BF–bCD system the thermodynamic results suggest that S-BF–bCD inclusion complex with DH of 181.8 kJ mol1 is more stable that R-BF–bCD complex (DH = 131.2 kJ mol1). Also from the PM 6 calculations we obtained high negative entropy changes in the range of 240– 310 J K1 mol1 for the BF complexes with aCD and bCD. This is due to the decrease in the translational and rotational degrees of

plex

48

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

freedom upon complexation of the guest molecules with the host molecule compared to the free molecules. Additionally, the negative Gibbs free energy changes clearly indicate that the interaction of BF with aCD and bCD is spontaneous at room temperature. The high DG values reflect the high affinity of BF to the cyclodextrin cavity. Additionally, the thermodynamic properties of these complexes clearly confirm that an efficient enantioseperation is obtained by bCD as a chiral selector. It is worth mentioning here that the thermodynamics of the inclusion complexation is associated with insertion of guest molecule into the hydrophobic host cavity of cyclodextrin molecules. This process is accompanied by formation of tightly bound guest–host complex where various noncovalent interactions play a pivotal role such as van der Waals interactions, hydrophobic interactions and hydrogen bonding. Careful examination of the optimized structures of the inclusion complexes presented in Figs. 4 and 5 reveals the presence of a number of strong hydrogen bonding (Table S5). Evidently from these results the intermolecular hydrogen interactions in the BF–CD inclusion complexes play an essential role with regard to the stability of these complexes as reflected by the thermodynamics data. Needless to say that despite the fact that this system involves a remarkable number of hydrogen bonding interactions, other weaker guest–host interactions, such as van del Waals interactions, dipole–dipole interaction and other hydrophobic interactions, contribute significantly to the stability and the geometry of these inclusion complexes [48]. The enantiorecognition is ascribed to the subtle differences between strong and weak intermolecular interactions involved in the formation of the different complexes. BF is indeed a relatively small molecule that fits into the cavity of both aCD and bCD. However, it is clear that the intermolecular forces of BF with bCD lead to a profound difference in energy between the optimum geometries of R-BF and S-BF within the cyclodextrin cavity resulting in their enantioseparation. 3.5. Molecular dynamics To shed an additional light on the separation mechanism, we further run MD simulation on BF–bCD complexes. The time

dependence of the root-mean square deviation (RMSD) of the MD simulation of the atomic position of R-BF–bCD and S-BF–bCD inclusion complexes compared to the energy minimized structure is illustrated in Fig. 6. As seen from this figure the trajectories of the MD simulation on both complexes in water are stabilized. For R-BF–bCD the MD trajectory rose sharply to 2.5 Å within 1 ns then averaged at around 2.6 Å after 4 ns. On the other hand, the RMSD for S-BF–bCD reached 2.7 Å in the first 1.7 ns and averaged at around 2.3 Å in the final stages of the simulation. In both cases the RMSD fluctuates during the first four nanoseconds reaching up to 3.6 Å in the case of R-BF–bCD, but more stable trajectories are oscillating around an average value of 2.5 Å. Therefore only conformations generated during the last 4 ns were used to determine the properties of these complexes. We further investigated the evolution of the structure of the inclusion complexes with time through the established intermolecular hydrogen bonding. It is imperative to mention here that the length of intermolecular hydrogen bonds in inclusion complexes and the percentage occupancy during MD simulations are indicators of the stability of these complexes. Table 3 shows a list of the most abundant hydrogen bonds formed between S-BF and bCD averaged over the last four nanoseconds. The major hydrogen bonds formed involve the secondary hydroxyl oxygen atoms of the bCD acting as donor atoms with the hydroxyl and the amine moieties in BF acting as acceptor atoms. These results are consistent with the NMR results obtained indicating that the phenyl ring of BF is deeply inserted into the cavity of the cyclodextrin while the amino and the carboxyl group are located at the entrance of bCD cavity that is guarded by the secondary hydroxyl groups in a Table 3 Hydrogen bond occupancy and distance (and standard deviation) calculated during the last four nanosecond of the MD trajectories for S-BF–bCD. Donor

Acceptor

Occupancy%

Distance (SD)

OH OH OH OH OH

OH (BF) OH (BF) NH2 (BF) NH2 (BF) NH2 (BF)

20.4 18.9 16.2 14.8 14.3

2.785 2.743 2.868 2.866 2.876

(CD) (CD) (CD) (CD) (CD)

(0.11) (0.11) (0.08) (0.08) (0.08)

Fig. 6. The time dependence of the root-mean-square deviations (RMSD) of atomic positions in the MD simulated complexes from those in the corresponding energy minimized structures for (a) R-BF–bCD and (b) S-BF–bCD.

F.O. Suliman, A.A. Elbashir / Journal of Molecular Structure 1019 (2012) 43–49

position that favors hydrogen bond formation. Surprisingly, R-BF– bCD exhibited no significant hydrogen bonding during the last stages of the simulation suggesting that a less stable complex is formed for R-BF–bCD compared to S-BF–bCD. This confirms the results obtained by the PM6 modeling of the complexes in the gas phase and also in line with the CE results presented in Fig. 2. These findings indicate that enantiorecognition of BF by bCD is mainly due to the differences of stabilities between the inclusion complexes of the two enantiomers. 4. Conclusion The use of bCD as a chiral selector for BF in capillary electrophoresis resulted in separation of R-BF from S-BF with Rs of 1.2. On the other hand when aCD is used as a chiral selector no separation between the two enantiomers was observed. The inclusion complexation of BF with cyclodextrins was also investigated by 1H NMR and ESI–MS. It was inferred from these techniques that an inclusion complex of 1:1 ratio is obtained between the cyclodextrin and BF with the aromatic ring inserted into the cavity. Dockings based on molecular mechanics calculations using Autodock in conjunction with quantum mechanical calculation using PM6 semiempirical method and molecular dynamics simulations were performed to rationalize the experimental results and to explain the mechanism of separation. The optimum conformations of the BF–CD inclusion complexes generated by the Autodock simulation procedures were further optimized by the semiempirical method PM6. The theoretical calculations of the PM6 method predicted that aCD as a chiral selector cannot separate R-BF from S-BF. The same method, however, predicted that bCD can separate the two enantiomers. Furthermore, MD simulations of the complexes in aqueous media supported the results obtained by PM6 calculations suggesting that the differences in stabilities of diastereomers lead to the observed separation. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molstruc. 2012.03.055.

[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35]

[36] [37] [38] [39] [40] [41] [42]

References [1] C.J. Nunez-Aguero, C.M. Escobar-Llanos, D. Diaz, C. Jaime, R. Garduno-Juarez, Tetrahedron 62 (2006) 4162. [2] A. Rizzi, Electrophoresis 22 (2001) 3079. [3] D.C. Harris, Quantitative Chemical Analysis, eigth ed., W. H. Freeman, New York, 2010. [4] B. Chankvetadze, J. Chromatogr. A 1168 (2007) 45. [5] B. Chankvetadze, W. Linder, G.K.E. Schriba, Anal. Chem. 76 (2004) 4256. [6] G.K.E. Scriba, J. Pharm. Biomed. Anal. 27 (2002) 373.

[43] [44] [45] [46] [47] [48]

49

G.K.E. Scriba, J. Pharm. Biomed. Anal. 55 (2011) 688. G.K.E. Scriba, J. Sep. Sci. 31 (2008) 1991. A. Madan, M.C. Schiess, Neuromodulation 14 (2011) 176. L.S. Liu, M. Shenoy, P.J. Pasricha, Neurogastroenterol. Motil. 23 (2011) 356. C.K. Thomas, C.K. Hager-Ross, C.S. Klein, Brain 133 (2010) 117. A. Saval, A.E. Chiodo, J. Spinal Cord Med. 33 (2010) 16. J.S. Kroin, Clin. Pharmacokinet. 22 (1992) 319. J. Sawynok, C. Dikson, Prog. Neuropsychopharmacol. Biol. Psychol. 8 (1984) 729. L. Suntornsuk, S. Ployngam, J. Pharm. Biomed. Anal. 51 (2010) 541. C. Desiderioa, D.V. Rossetti, F. Perri, B. Giardinaa, I. Messana, M. Castagnolaa, J. Chromatogr. B 875 (2008) 280. R.I. Stefan van Staden, A.A. Rat’ko, Talanta 69 (2006) 1049. M.P. Vaccher, E. Lipka, J.P. Bonte, C. Foulon, J.F. Goossens, C. Vaccher, Electrophoresis 26 (2005) 2974. I. Ali, H.Y. Aboul-Enein, Electrophoresis 24 (2003) 2064. K. Shimada, K. Mitamura, M. Morita, K. Hirakata, J. Liq. Chromatogr. 16 (1993) 3311. S. Chaudhuri, S. Chakraborty, P.K. Sengupta, J. Mol. Struct. 975 (2010) 160. S. Chakraborty, S. Basu, A. Lahiri, S. Basak, J. Mol. Struct. 977 (2010) 180. M.-J. Huang, Z. Quan, Y.-M. Liu, Int. J. Quantum Chem. 109 (2009) 81. A.A. Elbashir, F.E.O. Suliman, B. Saad, H.Y. Aboul-Enein, Talanta 77 (2009) 1388. A.A. Elbashir, F.E.O. Suliman, B. Saad, H. Aboul-Enein, Biomed. Chromatogr. 24 (2009) 393. C. Yan, Z. Xiu, X. Li, C. Hao, J. Mol. Graph. Model. 26 (2007) 420. C. Jullian, V. Brossard, I. Gonzalez, M. Alfaro, C. Olea-Azar, J. Solut. Chem. (2011) 727. A.A. Elbashir, F.O. Suliman, J. Chromatogr. A 1218 (2011) 5344. Z. Li, E.P.A. Couzijn, X. Zhang, J. Phys. Chem. B 116 (2012) 943. W. Li, B. Lu, F. Chen, F. Yang, Z. Wang, J. Mol. Struct. 990 (2011) 244. P. Liu, D. Zhang, J. Zhan, J. Phys. Chem. A 114 (2010) 13122. J.K. Khedkar, V.V. Gobre, R.V. Pinjari, S.P. Gejji, J. Phys. Chem. A 114 (2010) 7725. J.J.P. Stewart, J. Mol. Model. 13 (2007) 1173. G.M. Morris, D.S. Goodsell, R.S. Halliday, R. Huey, W.E. Hart, R.K. Belew, A.J. Olson, Autodock 4.2, The Scripps Research Institute, La Jolla, CA, USA, 2009. M.F. Sanner, R. Huey, S. Dallakyan, W. Lindstorm, G.M. Morris, A. Norledge, A. Omelchenko, D. Stoffler, G. Vareille, Autodock Tools, The Scripps Institute, La Jolla, CA, USA, 2007. J. Gasteiger, M. Marsili, Tetrahedron 36 (1980) 3219. J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, J. Comput. Chem. 25 (2004) 1157. W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, J. Chem. Phys. 79 (1983) 926. U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, J. Chem. Phys. 103 (1995) 8577. H.J.C. Berendsen, J.P.M. Postma, W.F. Van Gunsteren, A. Dinola, J.R. Haak, J. Chem. Phys. 81 (1984) 3684. J.P. Ryckaert, G. Ciccotti, H.J.C. Berendsen, J. Comput. Phys. 23 (1977) 327. T.A.D.D.A. Case, T.E. Cheatham, III, C.L. Simmerling, J. Wang, R.E. Duke, R. Luo, R.C. Walker, W. Zhang, K.M. Merz, B. Roberts, B. Wang, S. Hayik, A. Roitberg, G. Seabra, I. Kolossvai, K.F. Wong, F. Paesani, J. Vanicek, J. Liu, X. Wu, S.R. Brozell, T. Steinbrecher, H. Gohlke, Q. Cai, X. Ye, J. Wang, M.-J. Hsieh, G. Cui, D.R. Roe, D.H. Mathews, M.G. Seetin, C. Sagui, V. Babin, T. Luchko, S. Gusarov, A. Kovalenko, P.A. Kollman, AMBER 11, University of California, San Francisco., 2010. Donald A. Godwin, N.-H. Kim, R. Zuniga, Hosp. Pharm. 36 (2001). M. Sznitowska, S. Janicki, T. Gosb, Int. J. Pharm. 137 (1996) 125. S. Ahuja, Chromatographia 34 (1992) 411. C.A. Ventura, I. Giannone, D. Paolino, V. Pistarà, A. Corsaro, G. Puglisi, Eur. J. Med. Chem. 40 (2005) 624. L. Ribeiro, R.A. Carvalho, D.C. Ferreira, F.J.B. Veiga, Eur. J. Pharm. Sci. 24 (2005) 1. G.R. Desiraju, Acc. Chem. Res. 29 (1996) 441.