Analyst View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
PAPER
Cite this: DOI: 10.1039/c4an01265h
View Journal
Quantitative prediction of enantioseparation using b-cyclodextrin derivatives as chiral selectors in capillary electrophoresis† Xin Guo,a Zhiqiang Wang,a Lihua Zuo,b Zhixu Zhou,a Xingjie Guo*b and Tiemin Sun*a b-Cyclodextrin derivatives as chiral selectors are becoming increasingly important for enantioseparations in capillary electrophoresis (CE). Nevertheless, there are some enormous challenges in choosing effective selectors from a variety of compounds, and up to now no systematic quantitative studies for predicting the possibility of enantiomeric separation before CE experiments have been reported. In this paper, in order to resolve previous confusions, we investigated the enantioseparations of ten chiral drugs using a method of combining experiments with theoretical calculations. MMFF, PM3, DFT and ONIOM2 methods were simultaneously utilized during the course of our computer simulations. The results indicated that a
Received 14th July 2014 Accepted 15th October 2014
specific value of greater than or approximately equal to 6 kJ mol1 for the interaction energy difference
DOI: 10.1039/c4an01265h
separation. This discovery offers a meaningful reference to predict enantiomeric separations, so as to
www.rsc.org/analyst
design and synthesize some more efficient chiral selectors.
(DDE) between a pair of enantiomers with a selector is required in order to achieve enantiomeric
1. Introduction Capillary electrophoresis1–3 has been established as an important technique for the separation of chiral drugs because of its short analysis time, high efficiency, and the diversity of chiral selectors. Recently, various theoretical studies have been performed to investigate the enantioseparation mechanism of bcyclodextrin derivatives4–6 as chiral selectors in CE. It has been conrmed that interactions between enantiomers and chiral selectors may affect the mobilities of enantiomers, and that the retention times of the free and complexed forms are decided by the interaction energies, because parameters that inuence the separation, such as type and concentration of the selectors, composition and pH of the background electrolyte (BGE), temperature and voltage, are similar during the electrophoretic process.7–11 Some irregular energy differences in the interactions between enantiomers and selectors were found by using various molecular modeling methods.12–17 Nevertheless, because of the computational limitations of low levels of theory, or the non-quantitative nature of theoretical studies with respect to qualitative analyses, the results of previous studies
could only be used to interpret the mechanism of enantiomeric separation, but could not be used to predict the possibility of separation before the CE experiment. In this paper, we investigated ten chiral drugs (Fig. 1) that were separated by CE, from the perspective of the relationship between the possibility of chiral separation and the value of the interaction energy difference, by using a method of combining experiments with computer simulations, in order to obtain the critical value which satised enantioseparation. Some of the selected compounds (anticholinergic drugs including Homatropine hydrobromide, Anisodamine, Tropicamide, Atropine, Homatropine methylbromide and Atropine methobromide; Clorprenaline and Tulobuterol belong to antiasthmatic drugs)
a
Key Laboratory of Structure-Based Drug Design and Discovery, Shenyang Pharmaceutical University, Ministry of Education, Wenhua Road 103, Shenyang 110016, P. R. China. E-mail: [email protected]; Fax: +86-24-23986398
b
School of Pharmacy, Shenyang Pharmaceutical University, Wenhua Road 103, Shenyang 110016, P. R. China. E-mail: [email protected]
† Electronic supplementary information (ESI) available: The relationship between the interaction energy difference (DDE) and the resolution (Rs) for several chiral drugs separated by capillary electrophoresis using b-cyclodextrin derivatives as selectors. See DOI: 10.1039/c4an01265h
This journal is © The Royal Society of Chemistry 2014
Fig. 1 The chemical structures of the ten chiral compounds under investigation.
Analyst
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Analyst
could be efficiently separated with favorable resolutions (Rs), while the separations of other compounds (antibacterial drugs including Gatioxacin and Ooxacin) could not be achieved. We assumed that a specic value of interaction energy difference between of each pair of enantiomers and a chiral selector to achieve separation could be obtained by theoretical simulations. To clarify the computational procedure, Tulobuterol (TUL) and Clorprenaline (CLO) separated by sulfated b-cyclodextrin (S-b-CD) were chosen from the ten compounds under investigation as the examples. Until now, various computational approaches have been applied to investigate systems of inclusion complexes such as molecular dynamics (MD),18 molecular mechanisms (MM),19 quantum mechanism (QM),20 and hybrid techniques like quantum mechanism-molecular mechanism (QM-MM).21 MD and MM are oen used to simulate the systems at the molecular level. Quantum mechanism including semi-empirical methods22 and density functional theory (DFT)23 calculations, especially using the popular B3LYP functional combined with different standard basis sets, has been used reliably to describe host and guest interactions. In addition, QM-MM is an important method which has attracted more and more attention due to its excellent properties of accurate and fast calculations for supramolecular systems. As the majority of previous studies were inclined to apply one of the MD, MM, or QM methods for qualitative investigations, it is not necessary to use extremely precise methods. However, we investigated the mechanism from the perspective of quantitative analysis as some more accurate data than that previously attained should be obtained to support the purpose of our studies. Therefore, the MM (MMFF), QM (PM3, DFT) and QM-MM (ONIOM2) methods were simultaneously performed in this paper, and all computer simulations were implemented using Spartan 10 (ref. 24–26) and Gaussian 09 soware packages.27
2. 2.1
Methods Experimental section
Chemicals and reagents. Racemates of Tulobuterol, Clorprenaline, Tropicamide, Homatropine hydrobromide, Homatropine methylbromide, Atropine, Atropine methobromide, Anisodamine, Gatioxacin and Ooxacin were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Enantiomers of the ten compounds under investigation were kindly provided by the Pharmaceutical Chemistry Laboratory of Shenyang Pharmaceutical University. Sulfated b-cyclodextrin and carboxymethyl b-cyclodextrin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sulfamic acid b-cyclodextrin was provided by the Pharmaceutical Chemistry Laboratory of Shenyang Pharmaceutical University. Sodium hydroxide (NaOH), chromatographic grade phosphoric acid, tris(hydroxymethyl) aminomethane (Tris), triethanolamine (TEA) and sodium dihydrogen phosphate (NaH2PO4) were supplied by the Tianjin Bodi Chemical Holding Co. Ltd. (Tianjin, China). Ultra-pure water was used throughout and all solutions were passed through 0.22 mm pore size lters. Analyst
Paper
Apparatus. All experiments were carried out on a Beckman P/ ACETM MDQ capillary electrophoresis system (Beckman Fullerton, CA, USA), equipped with a diode array detector for absorbance measurements at the wavelength of 210 nm with a data sampling rate of 4 Hz, a bandwidth of 10 nm, and normal ltering. 32 Karat 8.0 Soware (Beckman, Fullerton, CA, USA) was used for the instrumental control, data acquisition and data analysis. The temperature control of the sample carousel was kept constant using a capillary cartridge coolant (Beckman Coulter Instruments, CA, USA). An uncoated fused-silica capillary column (Beckman Coulter Instruments, CA, USA) with a total length of 50.2 cm (effective length 40 cm) 50 mm i.d. was used for separation throughout the experiments. Capillary conditions. The new capillary was activated by ushing at 20 psi with methanol for 10 min, 1.0 mol L1 HCl for 10 min, water for 10 min, 1.0 mol L1 NaOH for 40 min and water for 10 min. At the beginning of each working day, the capillary was rinsed successively at 20 psi with 0.1 mol L1 NaOH for 10 min, water for 10 min, and conditioned with background electrolyte for 20 min. Between runs, the capillary was ushed with background electrolyte for 2 min. At the end of each working day, the capillary was rinsed for 10 min with water at 20 psi and then dry stored. Preparation of background electrolyte and sample preparation. The background electrolyte for the electrophoretic experiments was prepared by dissolving the appropriate phosphate in ultra-pure water, and the BGE pH was adjusted with phosphoric acid (1 M) or sodium hydroxide (1 M). The chiral selectors were added to the BGE at the desired concentration. The running buffer was passed through a 0.22 mm lter before use. The injection was performed by pressure at 0.5 psi for 5 s. To achieve reproducible separations, all experiments were carried out in triplicate. The BGE solution was refreshed aer 10 runs. The stock solutions of Tulobuterol, Clorprenaline, Tropicamide, Homatropine hydrobromide, Homatropine methyl bromide, Atropine, Atropine methobromide, Anisodamine, Gatioxacin and Ooxacin at concentrations of 1.0 mg L1 were prepared in ultra-pure water. Standard solutions of different concentrations were prepared by dilution of the stock solution with ultra-pure water. All running buffers and analyte solutions were passed through a 0.22 mm syringe lter before injection. 2.2
Computational section
The initial structures of (R/S)-Tulobuterol and (R/S)-Clorprenaline obtained by CS ChemDraw Ultra 12.0 were imported into the Spartan 10 soware program for conformational searching using the MMFF method.24–26 The molecular mechanics calculations supplied minimum energies for all conformers. The optimal energy conformers were selected within a region of 5 kcal mol1. As a more progressive level, the precision of the density functional method is higher than that of molecular mechanics. Therefore, the above conformers were further calculated using the Gaussian 09 soware program at the density functional theory level with a B3LYP functional and a 631G(d,p) basis set. Then, energy values were redistributed according to the Boltzmann distribution, discarding the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Paper
conformers which possess inappreciable contribution to conformational equilibrium, permitting decline to the most stable conformers of (R/S)-TUL and (R/S)-CLO for further simulations. Furthermore, the b-cyclodextrin structure was obtained from the crystallographic parameters provided by the Cambridge Crystallographic Data Centre, and based on the initial structure, sulfated b-cyclodextrin was constructed and then fully optimized at the PM3 level of theory without imposing any symmetry.28 Subsequently, in order to get the most stable structures of the inclusion complexes, the separately optimized host and guest molecules were used to construct the complexes using a coordinate system (Fig. 2) of the potential energy scanning method. The glycosidic oxygen atoms of S-b-CD were positioned on the XY at surface and their centre was established as the origin of coordinates for the complete system. The reference ˚ on the Z axis, and then the atom was originally situated at 10 A ˚ whole guest molecule entered into the S-b-CD cavity up to 10 A ˚ The structures generated at along the Z axis with a step of 1 A. each step were then optimized, allowing them to change from the initial conformations while keeping the movement of the reference atoms and the S-b-CD structure totally restricted. As during the electrophoretic process chiral compounds t into the CD cavity either completely or with their hydrophobic part only, entering through one of the two openings, two different inclusion orientations were considered. In the rst orientation (Model I), guest molecules passed through the cavity of S-b-CD from its wide side, with the benzene ring leading. In the second orientation (Model II), the benzene ring was docked into the narrow side of the cavity. The interaction energies (DE) between the host and the guest molecules are calculated according to eqn (1),where Ecomplex, Eguest and Ehost stand for the total energy
Analyst
of the inclusion complex and the free energy of the guest and host molecules, respectively. Furthermore, the interaction energy difference (DDE) is calculated on the basis of eqn (2),where DER and DES represent the energies of the R-enantiomer and the S-enantiomer interacting with the host molecule, respectively.
3.1
The process of complexation for R-Tulobuterol with S-b-CD: A (Model I) and B (Model II) represent the guest molecules passing though S-b-CD from its wide and narrow sides, respectively. Reference atoms are labelled with red stars.
This journal is © The Royal Society of Chemistry 2014
(1)
DDE ¼ DER DES
(2)
To further identify the possibility and ease of formation of inclusion complexes, we ran frequency calculations for the above stable complexes at 1 atm and 298 K in water by PM3 methods. It is well known that the enthalpy (H), entropy (S) and Gibbs free energy (G) are the most important parameters in a thermodynamic system, as changes in them can be used to conrm the exothermic or endothermic nature of a reaction, clarify the chemical stabilities, and predict the equilibrium and spontaneity, respectively.29 Therefore, we then calculated the enthalpy change (DH), the entropy change (DS) and the Gibbs free energy change (DG) by using statistical thermodynamics methods. Then, with the purpose of improving the accuracy of the computational modelling and reducing the computational requirement, a rotation was performed to establish the preferred angular orientation of the guest molecules during the inclusion process based on the structural characteristics of the cavity; an ONIOM2 [B3LYP/631G(d,p):PM3] method as a superior level of calculation was used in order to approach the ideal geometry for the above complexes which were obtained by a PM3 method. In our ONIOM2 calculations, TUL and CLO were dened as the layer of high-level, while S-b-CD was dened as the layer of low-level, which were calculated by a B3LYP/631G(d,p) level of theory and a PM3 method, respectively. Eventually, the natural bond orbital (NBO) and frontier molecular orbital (FMO) calculations were carried out for the most stable complexes obtained by the ONIOM2 method. NBO can be used to elucidate the intermolecular interaction between host and guest molecules via the determination of the stabilization energy E(2), which was correlated with the delocalization tendency of electrons on the basis of the perturbation method.30 Furthermore, FMOs were used to investigate the chemical stabilities of the above inclusion complexes.31,32
3.
Fig. 2
DE ¼ Ecomplex (Eguest + Ehost)
Results and discussion Experimental analysis
During the process of our experiments, we investigated the parameters that have an effect on the enantioseparation, including the background electrolyte pH, the concentration of the b-CD derivatives, the buffer concentration, temperature and voltage. As seen from Table 1, the enantiomers of Homatropine hydrobromide, Tulobuterol, Tropicamide, Atropine, Homatropine methylbromide, Atropine methobromide, Clorprenaline and Anisodamine could be effectively separated by simple,
Analyst
View Article Online
Analyst
Paper
Conditions used to achieve baseline resolution for ten chiral drugs with b-CD derivatives as chiral selectors separated by capillary electrophoresis
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Table 1
Chiral drug
b-CD derivative
Buffer concentration (mmol L1)
pH
Voltage (kV)
Temperature ( C)
Resolution
Tulobuterol Clorprenaline Tropicamide Homatropine hydrobromide Homatropine methylbromide Atropine Atropine methobromide Anisodamine Gatioxacin Ooxacin
6-Sulfated-b-CD 6-Sulfated-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 2,6-Sulfamic acid-b-CD 2,6-Sulfamic acid-b-CD
30 30 30 30 30 30 30 30 40 40
3.0 3.0 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5
20 20 25 25 25 25 25 25 30 30
25 25 20 20 20 20 20 20 25 25
6.17 1.50 5.13 6.29 4.67 4.85 4.59 1.13/6.92/2.58 — —
fast and precise methods, but the enantiomeric separations of Gatioxacin and Ooxacin could not be achieved by optimizing the experimental conditions. Among them, the enantioseparations of Tulobuterol and Clorprenaline with S-b-CD as chiral selectors were obtained within short analysis times, and no interferences were observed at the migration times of the peaks corresponding to the two analytes, which demonstrated the selectivity of this method (Fig. 3). The migration orders of the Rand S-enantiomers were identied by adding the S-enantiomer into the racemic mixture. Notably, not only Tulobuterol, but also Clorprenaline followed the trend that R-enantiomers are separated prior to S-enantiomers. Therefore, the S-enantiomers preferably tted into the S-b-CD cavity through its wide or narrow sides, resulting in different interaction energies compared with R-enantiomers. This led to different mobilities for each pair of enantiomers, and thus the enantiomeric separations could be achieved.
Electropherograms of Tulobuterol (A) and Clorprenaline (B) separated by capillary electrophoresis.
Fig. 3
Analyst
3.2
Structures of isolated host and guest molecules
Fig. 4 summarizes the most stable structures of the isolated host and guest molecules optimized by PM3 and DFT methods, respectively. As the difference in geometrical parameters is a primary factor that inuences the chiral separation, which makes the inclusion complex formed by one enantiomer more stable than that formed for the other one, it is necessary to nd the low-lying energy conformers in order to improve the precision for further theoretical calculations. 3.3 Determination of the initial structures of the inclusion complexes The results of the potential energy scanning method showed that the DE varied at different coordinates (Fig. 5). The more negative the interaction energy, the more stable the complex formed by the host–guest molecules and eight DE minima were found for Model I and Model II. As seen from Table 2, the complexes of (R/S)-TUL and (R/S)-CLO possess energies that are lower than the sum of the energy of the isolated guest and host molecules. This indicated that it is favorable to form inclusion complexes for all enantiomers and in both orientations.
Fig. 4 Geometrical structures of R-TUL (A), S-TUL (B), R-CLO (C), SCLO (D), and the S-b-CD (E) optimized by using the DFT and PM3 methods, respectively.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper
Analyst
Therefore, inclusion complexes A, B, C and D were the initial structures obtained using this calculational method. Meanwhile, the above results claried that the interaction energy difference between a pair of enantiomers with a selector was a signicant factor contributing to the enantiomeric separation.
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
3.4
Thermodynamic analysis
The thermodynamic parameters of the complexation process, including enthalpy change (DH), entropy change (DS) and Gibbs free energy change (DG) are presented in Table 3. Clearly, the interactions between the enantiomers and S-b-CD were exothermic as evidenced by the large enthalpy decreases. Furthermore, the DS and DG were also negative for the inclusion processes of forming the most stable complexes for TUL/ CLO with S-b-CD, indicating that the formations of the complexes were enthalpy-driven and spontaneous procedures. The statistical thermodynamic parameters for the above obtained stable inclusion complexes revealed a propensity to form inclusion complexes between host and guest molecules. Thus, the main factors which contributed to the thermodynamics are associated with the insertion of the guest molecules into the S-b-CD cavity. 3.5 Interaction energies of R-TUL (A), S-TUL (B), R-CLO (C), and SCLO (D) entered into S-b-CD from its wide (Model I) and narrow (Model II) sides.
Fig. 5
Furthermore, comparing which side, wide or narrow, the guest molecules entered the cavity from, it is clear that the interaction energies of Model I complexes are lower than those for Model II complexes for all enantiomers, demonstrating that complexes formed by guest molecules which passed through from the wide side are more stable and available, and thus we selected Model I for further calculations. On the other hand, compared with the DE obtained from the wide orientation, the DDE of the interactions between R-enantiomers and S-enantiomers with S-b-CD for TUL and CLO were 10.58 kJ mol1 and 9.68 kJ mol1, respectively. This revealed that the complexes which were formed by S-enantiomers and S-b-CD were more stable than those formed by R-enantiomers, that is, the S-enantiomers demand slightly more energy than R-enantiomers for the purpose of accommodating to their structures within the S-bCD cavity, and thus the R-enantiomer can be separated rst by CE, and the S-enantiomer can be separated aerwards.
Table 2
Accurate optimization of the inclusion complexes
Although the PM3 method was able to explain the mechanism of enantiomeric separation, it seems that the result was not accurate. If we used a higher level of theory, some more accurate results might be obtained. However, for such a large system, higher theoretical methods would suffer from a high computational requirement. Therefore, for the sake of overcoming the above deciencies, a multi-layered ONIOM method was used during the process of our theoretical studies. This QM-MM approach not only retains the reliability of quantum mechanics, but also possesses the property of fast calculations for
Table 3 Statistical thermodynamic parameters of Model I calculated by PM3 methods
S-b-CD complex
DHa (kJ mol1)
DSb (J mol1 K1)
DGc (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
28.53 38.77 34.28 45.16
58.82 75.19 62.07 81.94
10.96 16.31 15.49 20.73
a DH ¼ Hcomplex Hhost Hguest. b D ¼ Scomplex Shost Sguest. c DG ¼ DH TDS.
The DE and DDE of the inclusion complexes formed by Tulobuterol and Clorprenaline with S-b-CD
S-b-CD complex
DEModel Ia (kJ mol1)
DEModel IIa (kJ mol1)
DDEModel Ib (kJ mol1)
DDEModel IIb (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
30.25 40.83 37.60 47.28
26.97 35.18 34.11 42.25
10.58
8.21
9.68
8.14
a
DE ¼ Ecomplex (Eguest + Ehost).
b
DDE ¼ DER DES.
This journal is © The Royal Society of Chemistry 2014
Analyst
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Analyst
Paper
macromolecular systems. Table 4 summarizes the results of the ONIOM2 calculations, indicating that the total energies of the inclusion complexes calculated by the ONIOM2 method (EONIOM2) are more negative than those obtained from the PM3 method (EPM3), demonstrating that the higher levels of theory lead to more rational and stable structures for the complexes. Furthermore, the DDE of the interactions between R-enantiomers and S-enantiomers with S-b-CD were 15.41 kJ mol1 and 10.92 kJ mol1 for TUL and CLO respectively, and thus the inclusion complexes formed by S-enantiomers were more stable than those formed by R-enantiomers. The more stable the inclusion complex, the longer the migration time and therefore the R-enantiomer is prone to separate rst, which is in agreement with the enantiomer migration orders observed in the CE experiment. The results of the above theoretical calculations show that the energy difference in the interactions between a pair of enantiomers and a host makes for different migration times, and thus enantioseparation can be achieved. In addition, we found that the benzene rings of the guest enantiomers come very close to the centre of the S-b-CD cavity as a result of hydrophobic interactions between the hydrophobic cavity and the benzene ring, making the inclusion complexes more stable than free host and guest molecules (Fig. 6).
3.6
Hydrogen bonding interactions
The detailed results of the E(2) associated with hyperconjugative interactions are summarized in Table 5, which can be utilized to quantify the range of hydrogen bonding interactions. It suggested that, in general, the E(2) value of strong hydrogen bonding is larger than 2.0 kcal mol1, while the E(2) value of a weak hydrogen bonding interaction is determined to be from 0.5 kcal mol1 to 2.0 kcal mol1. The stabilization energies E(2) of the R-TUL and R-CLO inclusion complexes were all larger than 2.0 kcal mol1, indicating the presence of four strong hydrogen bonding interactions. In addition, the E(2) for the intermolecular C–H/O interaction [LP(O125):BD*(O190– H205)] in the S-TUL complex was also larger than 2.0 kcal mol1, which demonstrated a strong hydrogen bond. Furthermore, both the E(2) of the C–H/O interaction in the S-CLO complex and the E(2) for the intermolecular C–H/O interaction [LP(O190):BD*(O7–H8)] in the S-TUL complex were found to be smaller than 0.5 kcal mol1, revealing two very weak hydrogen bonding interactions. This donor–acceptor electron transfer model of NBO calculations reveals that the stabilizations of the above four stable inclusion complexes are entirely inuenced by
Structures of inclusion complexes obtained from ONIOM2 calculations for R-TUL (A), S-TUL (B), R-CLO (C), and S-CLO (D) with Sb-CD.
Fig. 6
intermolecular hydrogen bonding interactions which arise from orbital overlap. Therefore, we conrmed that the formation of intermolecular hydrogen bonding, combined with hydrophobic interactions, was the mechanism for the stabilization of the inclusion complexes. The detailed intermolecular hydrogen bonding interactions are shown in Fig. 7, as indicated by dashed lines.
3.7
Frontier molecular orbitals
The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) are named the frontier molecular orbitals, and they can be used to gain knowledge of the nature of reactivities, and of some of the structural and physical properties of molecules. The energy difference between the HOMO and the LUMO, called the energy gap, is a critical parameter for determining chemical reactivities, kinetic stabilities and the hardness of molecules. Therefore, in order to investigate the chemical stabilities of the above inclusion complexes, the energies of the HOMOs and LUMOs and their orbital energy gaps were calculated using the above ONIOM2 method, and the results are listed in Table 6. A pictorial illustration of the FMOs for the S-Tulobuterol/S-b-CD complex is shown in Fig. 8; their respective positive and negative regions are represented by red and green colors. The values of the energy separations between the HOMOs and LUMOs were found to be 7.41, 8.73, 5.82 and 6.79 eV for R-
Table 4 The EPM3, EONIOM2, DEONIOM2 and DDEONIOM2 of the inclusion complexes formed by Tulobuterol and Clorprenaline with S-b-CD
S-b-CD complex
EPM3 (kcal mol1)
EONIOM2 (kcal mol1)
DEONIOM2a (kJ mol1)
DDEONIOM2b (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
3063.74 3066.27 3059.66 3061.80
667 042.68 667 050.64 642 371.40 642 478.99
33.28 48.69 35.26 46.18
15.41
a
DE ¼ Ecomplex (Eguest + Ehost).
Analyst
b
10.92
DDE ¼ DER DES.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper Table 5
Analyst The results of the NBO calculations
Table 6
d (2)a
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Donor i
Acceptor j
E (kcal mol1)
O/H
˚) (A
Angleb ( )
R-Tulobuterol/S-b-CD complex LP(1)O3 BD*(1)O189–H205 LP(2)O3 BD*(1)O189–H205 LP(1)O189 BD*(1)C9–H10 LP(2)O189 BD*(1)C9–H10
0.31 11.00 0.47 19.31
1.96
125.51
2.03
123.62
S-Tulobuterol/S-b-CD complex LP(1)O125 BD*(1)O190–H205 LP(2)O125 BD*(1)O190–H205 LP(2)O190 BD*(1)O7–H8
0.11 15.05 0.42
1.92
135.26
2.39
109.23
R-Clorprenaline/S-b-CD complex LP(1)O3 BD*(1)O189–H205 LP(2)O3 BD*(1)O189–H205 LP(1)O189 BD*(1)C9–H10 LP(2)O189 BD*(1)C9–H10
0.25 14.50 0.53 20.37
1.89
120.73
1.95
121.12
S-Clorprenaline/S-b-CD complex LP(1)O3 BD*(1)O190–H205 LP(2)O3 BD*(1)O190–H205 LP(2)O190 BD*(1)O29–H30 LP(2)O190 BD*(1)O29–H30
0.20 0.93 0.39 0.05
2.48
116.20
2.16
91.51
a
The energies of ELUMO, EHOMO, and DE
S-b-CD complex
ELUMO (eV)
EHOMO (eV)
DEa (eV)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
1.21 1.24 1.43 1.40
8.62 9.97 7.25 8.19
7.41 8.73 5.82 6.79
a
DE ¼ EHOMO ELUMO.
E(2) is the energy of hyperconjugative interaction. b The angle of X–H/Y.
The highest occupied and lowest unoccupied molecular orbitals of the S-Tulobuterol/S-b-CD complex.
Fig. 8
energy gaps than the R-enantiomer complexes, which indicates that the S-enantiomer inclusion complexes are more stable and that their electron density changes more easily than that of the R-enantiomer complexes. 3.8
Detailed intermolecular hydrogen bonding interactions, obtained from NBO calculations for R-TUL (A), S-TUL (B), R-CLO (C), and S-CLO (D) with S-b-CD.
Fig. 7
Tulobuterol, S-Tulobuterol, R-Clorprenaline, and S-Clorprenaline complexes, respectively. The large HOMO–LUMO gaps specify high excitation energies of the excited states, good stabilities and a high chemical hardness. We can conclude that the S-enantiomer complexes possess lower HOMO–LUMO
This journal is © The Royal Society of Chemistry 2014
Quantitative studies of ten chiral drugs separated by CE
In the present work, besides the two computational models, a variety of chiral drugs (Table SI†), separated by CE experiments with b-cyclodextrin derivatives as selectors, were investigated by using a similar theoretical method, and an essential condition for achieving enantiomeric separation was discovered. It is shown in Table 7 that the DDE of Homatropine hydrobromide, Tulobuterol, Tropicamide, Atropine, Homatropine methylbromide, Atropine methobromide, Clorprenaline, and Anisodamine enantiomers interacting with b-cyclodextrin derivatives are 17.38, 15.41, 12.60, 11.82, 10.27, 9.53, 10.92 and 6.05/18.97/ 9.22 kJ mol1, respectively. The DDE are all larger than 6 kJ mol1, and the chiral selectors could separate these chiral compounds with relatively ideal resolutions (6.29, 6.17, 5.13, 4.85, 4.67, 4.59, 1.50 and 1.13/6.92/2.58, respectively). The energy differences of the interactions between the enantiomers
Analyst
View Article Online
Analyst
Paper
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Table 7 The DDE and Rs for ten chiral drugs separated by CE
Chiral drugs
Substituent groupsa
DDE (kJ mol1)
Rs
Tulobuterol Clorprenaline Tropicamide Homatropine hydrobromide Homatropine methyl bromide Atropine Atropine methobromide Anisodamine Gatioxacin Ooxacin
6-Sulfated 6-Sulfated 6-Carboxymethyl 6-Carboxymethyl
15.41 10.92 12.60 17.38
6.17 1.50 5.13 6.29
6-Carboxymethyl
10.27
4.67
6-Carboxymethyl 6-Carboxymethyl 6-Carboxymethyl 2,6-Sulfamic acid 2,6-Sulfamic acid
11.82 9.53 6.05/18.97/9.22 5.83 5.96
4.85 4.59 1.13/6.92/2.58 — —
a
The substituent groups of b-cyclodextrin.
and the chiral selectors represent the energetic contribution to enantiomeric separation. Gatioxacin and Ooxacin exhibited single electrophoretic peaks, indicative of no separation, and the DDE were 5.83 and 5.96 kJ mol1, respectively. Furthermore, compared with other chiral drugs (Table SI†) separated by bcyclodextrin derivatives, we discovered that the energy difference of the interactions between a pair of enantiomers with a chiral additive must be greater than or approximately equal to 6 kJ mol1 in order to achieve enantiomeric separation. This can be used to predict the possibility of chiral separation before performing experiments.
4. Conclusions In this paper we put forward a novel explanation for CE, that the different interaction energies play a signicant role in the enantiomeric separation. The more stable the inclusion complex, the longer the migration time and therefore we can predict the enantiomer migration orders through theoretical calculations. Simultaneously, we found that the energy difference of the interactions between each pair of enantiomers with a chiral additive must be greater than or approximately equal to 6 kJ mol1 in order to achieve chiral separation. Thus, we can predict the possibility of enantiomeric separation, and design selectors to separate specic chiral drugs by molecular simulations before performing experiments. Meanwhile, we will also be able to synthesize new selectors for separating chiral drugs, which could be effectively used to reduce experimental requirements and increase separation efficiencies.
Acknowledgements This work was supported by the National Natural Science Foundation of China (no. 81273476). The authors of this manuscript acknowledge the Innovative Research Team of the Ministry of Education and Program for Liaoning Innovative Research Team in University for providing nancial assistance. The theoretical calculations were conducted on the ScGrid and Deepcomp 7000 the Supercomputing Center, Computer Network Information Center of Chinese Academy of Sciences.
Analyst
Notes and references 1 V. Subramaniam, L. Griffith and A. J. Haes, Analyst, 2011, 136, 3469–3477. 2 R. J. Flaherty, B. J. Huge, S. M. Bruce, O. O. Dada and N. J. Dovichi, Analyst, 2013, 138, 3621–3625. 3 L. R. Vemireddy, S. Archak and J. Nagaraju, J. Agric. Food Chem., 2007, 55, 8112–8117. 4 E. Junquera and E. Aicart, J. Pharm. Sci., 1999, 88, 626–631. 5 S. Kodama, A. Yamamoto, T. Ohura, A. Matsunaga and T. Kanbe, J. Agric. Food Chem., 2003, 51, 6128–6131. 6 S. R. Gratz and A. M. Stalcup, Anal. Chem., 1998, 70, 5166– 5171. 7 A. M. Rizzi and L. Kremser, Electrophoresis, 1999, 20, 2715– 2722. 8 A. Rizzi, Electrophoresis, 2001, 22, 3079–3106. 9 B. Chankvetadze, W. Linder and G. K. E. Scriba, Anal. Chem., 2004, 76, 4256–4260. 10 B. Chankvetadze, J. Chromatogr. A, 2007, 1168, 45–70. 11 G. K. E. Scriba, J. Sep. Sci., 2008, 31, 1991–2011. 12 A. A. Elbashir and F. O. Sulimana, J. Chromatogr A, 2011, 1218, 5344–5351. 13 S. L. Ma, S. Shen, N. Haddad, W. J. Tang, J. Wang, H. Lee, N. Yee, C. Senanayake and N. Grinberg, J. Chromatogr A, 2009, 1216, 1232–1240. 14 S. Stachowicz, A. Styrcz, J. Korchowiec, A. Modaressi and M. Rogalski, Theor. Chem. Acc., 2011, 130, 939–953. 15 J. H. Shi, Z. J. Ding and Y. Hu, J. Mol. Model., 2012, 18, 803– 813. 16 N. Leila, H. Sakina, A. Bouhadiba and F. Madi, J. Mol. Liq., 2011, 160, 8–13. 17 G. L. Alexandrino, A. Calderini, N. H. Morgon and F. B. T. Pessine, J. Inclusion Phenom. Macrocyclic Chem., 2013, 75, 93–99. 18 J. R. Garbow, J. J. Likos and S. A. Schroeder, J. Agric. Food Chem., 2001, 49, 2053–2060. 19 T. N. Doman, C. R. Landis and B. Bosnich, J. Am. Chem. Soc., 1992, 114, 7264–7272. 20 W. S. Cai, T. T. Sun, P. Liu, C. Chipot and X. G. Shao, J. Phys. Chem. B, 2009, 113, 7836–7843.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Paper
21 P. Walla, V. B. Arion and U. H. Brinker, J. Org. Chem., 2006, 71, 3274–3277. 22 T. J. Cleij, S. K. Y. Tsang and L. W. Jenneskens, Macromolecules, 1999, 32, 3286–3294. 23 A. W. Lange, M. A. Rohrdanz and J. M. Herbert, J. Phys. Chem. B, 2008, 112, 7345. 24 W. J. Hehre, L. Radom, P. V. R. Schleyer and J. A. Pople, Ab Initio Molecular Orbital Theory, Wiley., New York, 1986. 25 W. J. Hehre, A. J. Shusterman and W. W. Huang, A Laboratory Book of Computational Organic Chemistry, Wavefunction Inc., California, 1996. 26 W. J. Hehre, A Guide to Molecular Mechanics and Quantum Chemical Calculations, Wavefunction Inc., California, 2003. 27 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd,
This journal is © The Royal Society of Chemistry 2014
Analyst
28 29 30 31 32
E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, GAUSSIAN 09 (Revision A.1), Gaussian, Inc., Wallingford, CT, 2009. T. Aree and N. Chaichit, Carbohydr. Res., 2002, 337, 2487– 2494. H. Hayashi and K. Kojima, Ind. Eng. Chem. Res., 1994, 33, 1631–1635. Y. Tsujii, A. Tsuchida, M. Yamamoto and Y. Nishijima, Macromolecules, 1988, 21, 665–670. P. Bultinck, D. V. Neck, G. Acke and P. W. Ayers, Phys. Chem. Chem. Phys., 2012, 14, 2408–2416. M. D. King, C. E. C. Mas and R. P. Wayne, Phys. Chem. Chem. Phys., 1999, 1, 2231–2238.
Analyst
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
PAPER
Cite this: DOI: 10.1039/c4an01265h
View Journal
Quantitative prediction of enantioseparation using b-cyclodextrin derivatives as chiral selectors in capillary electrophoresis† Xin Guo,a Zhiqiang Wang,a Lihua Zuo,b Zhixu Zhou,a Xingjie Guo*b and Tiemin Sun*a b-Cyclodextrin derivatives as chiral selectors are becoming increasingly important for enantioseparations in capillary electrophoresis (CE). Nevertheless, there are some enormous challenges in choosing effective selectors from a variety of compounds, and up to now no systematic quantitative studies for predicting the possibility of enantiomeric separation before CE experiments have been reported. In this paper, in order to resolve previous confusions, we investigated the enantioseparations of ten chiral drugs using a method of combining experiments with theoretical calculations. MMFF, PM3, DFT and ONIOM2 methods were simultaneously utilized during the course of our computer simulations. The results indicated that a
Received 14th July 2014 Accepted 15th October 2014
specific value of greater than or approximately equal to 6 kJ mol1 for the interaction energy difference
DOI: 10.1039/c4an01265h
separation. This discovery offers a meaningful reference to predict enantiomeric separations, so as to
www.rsc.org/analyst
design and synthesize some more efficient chiral selectors.
(DDE) between a pair of enantiomers with a selector is required in order to achieve enantiomeric
1. Introduction Capillary electrophoresis1–3 has been established as an important technique for the separation of chiral drugs because of its short analysis time, high efficiency, and the diversity of chiral selectors. Recently, various theoretical studies have been performed to investigate the enantioseparation mechanism of bcyclodextrin derivatives4–6 as chiral selectors in CE. It has been conrmed that interactions between enantiomers and chiral selectors may affect the mobilities of enantiomers, and that the retention times of the free and complexed forms are decided by the interaction energies, because parameters that inuence the separation, such as type and concentration of the selectors, composition and pH of the background electrolyte (BGE), temperature and voltage, are similar during the electrophoretic process.7–11 Some irregular energy differences in the interactions between enantiomers and selectors were found by using various molecular modeling methods.12–17 Nevertheless, because of the computational limitations of low levels of theory, or the non-quantitative nature of theoretical studies with respect to qualitative analyses, the results of previous studies
could only be used to interpret the mechanism of enantiomeric separation, but could not be used to predict the possibility of separation before the CE experiment. In this paper, we investigated ten chiral drugs (Fig. 1) that were separated by CE, from the perspective of the relationship between the possibility of chiral separation and the value of the interaction energy difference, by using a method of combining experiments with computer simulations, in order to obtain the critical value which satised enantioseparation. Some of the selected compounds (anticholinergic drugs including Homatropine hydrobromide, Anisodamine, Tropicamide, Atropine, Homatropine methylbromide and Atropine methobromide; Clorprenaline and Tulobuterol belong to antiasthmatic drugs)
a
Key Laboratory of Structure-Based Drug Design and Discovery, Shenyang Pharmaceutical University, Ministry of Education, Wenhua Road 103, Shenyang 110016, P. R. China. E-mail: [email protected]; Fax: +86-24-23986398
b
School of Pharmacy, Shenyang Pharmaceutical University, Wenhua Road 103, Shenyang 110016, P. R. China. E-mail: [email protected]
† Electronic supplementary information (ESI) available: The relationship between the interaction energy difference (DDE) and the resolution (Rs) for several chiral drugs separated by capillary electrophoresis using b-cyclodextrin derivatives as selectors. See DOI: 10.1039/c4an01265h
This journal is © The Royal Society of Chemistry 2014
Fig. 1 The chemical structures of the ten chiral compounds under investigation.
Analyst
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Analyst
could be efficiently separated with favorable resolutions (Rs), while the separations of other compounds (antibacterial drugs including Gatioxacin and Ooxacin) could not be achieved. We assumed that a specic value of interaction energy difference between of each pair of enantiomers and a chiral selector to achieve separation could be obtained by theoretical simulations. To clarify the computational procedure, Tulobuterol (TUL) and Clorprenaline (CLO) separated by sulfated b-cyclodextrin (S-b-CD) were chosen from the ten compounds under investigation as the examples. Until now, various computational approaches have been applied to investigate systems of inclusion complexes such as molecular dynamics (MD),18 molecular mechanisms (MM),19 quantum mechanism (QM),20 and hybrid techniques like quantum mechanism-molecular mechanism (QM-MM).21 MD and MM are oen used to simulate the systems at the molecular level. Quantum mechanism including semi-empirical methods22 and density functional theory (DFT)23 calculations, especially using the popular B3LYP functional combined with different standard basis sets, has been used reliably to describe host and guest interactions. In addition, QM-MM is an important method which has attracted more and more attention due to its excellent properties of accurate and fast calculations for supramolecular systems. As the majority of previous studies were inclined to apply one of the MD, MM, or QM methods for qualitative investigations, it is not necessary to use extremely precise methods. However, we investigated the mechanism from the perspective of quantitative analysis as some more accurate data than that previously attained should be obtained to support the purpose of our studies. Therefore, the MM (MMFF), QM (PM3, DFT) and QM-MM (ONIOM2) methods were simultaneously performed in this paper, and all computer simulations were implemented using Spartan 10 (ref. 24–26) and Gaussian 09 soware packages.27
2. 2.1
Methods Experimental section
Chemicals and reagents. Racemates of Tulobuterol, Clorprenaline, Tropicamide, Homatropine hydrobromide, Homatropine methylbromide, Atropine, Atropine methobromide, Anisodamine, Gatioxacin and Ooxacin were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Enantiomers of the ten compounds under investigation were kindly provided by the Pharmaceutical Chemistry Laboratory of Shenyang Pharmaceutical University. Sulfated b-cyclodextrin and carboxymethyl b-cyclodextrin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sulfamic acid b-cyclodextrin was provided by the Pharmaceutical Chemistry Laboratory of Shenyang Pharmaceutical University. Sodium hydroxide (NaOH), chromatographic grade phosphoric acid, tris(hydroxymethyl) aminomethane (Tris), triethanolamine (TEA) and sodium dihydrogen phosphate (NaH2PO4) were supplied by the Tianjin Bodi Chemical Holding Co. Ltd. (Tianjin, China). Ultra-pure water was used throughout and all solutions were passed through 0.22 mm pore size lters. Analyst
Paper
Apparatus. All experiments were carried out on a Beckman P/ ACETM MDQ capillary electrophoresis system (Beckman Fullerton, CA, USA), equipped with a diode array detector for absorbance measurements at the wavelength of 210 nm with a data sampling rate of 4 Hz, a bandwidth of 10 nm, and normal ltering. 32 Karat 8.0 Soware (Beckman, Fullerton, CA, USA) was used for the instrumental control, data acquisition and data analysis. The temperature control of the sample carousel was kept constant using a capillary cartridge coolant (Beckman Coulter Instruments, CA, USA). An uncoated fused-silica capillary column (Beckman Coulter Instruments, CA, USA) with a total length of 50.2 cm (effective length 40 cm) 50 mm i.d. was used for separation throughout the experiments. Capillary conditions. The new capillary was activated by ushing at 20 psi with methanol for 10 min, 1.0 mol L1 HCl for 10 min, water for 10 min, 1.0 mol L1 NaOH for 40 min and water for 10 min. At the beginning of each working day, the capillary was rinsed successively at 20 psi with 0.1 mol L1 NaOH for 10 min, water for 10 min, and conditioned with background electrolyte for 20 min. Between runs, the capillary was ushed with background electrolyte for 2 min. At the end of each working day, the capillary was rinsed for 10 min with water at 20 psi and then dry stored. Preparation of background electrolyte and sample preparation. The background electrolyte for the electrophoretic experiments was prepared by dissolving the appropriate phosphate in ultra-pure water, and the BGE pH was adjusted with phosphoric acid (1 M) or sodium hydroxide (1 M). The chiral selectors were added to the BGE at the desired concentration. The running buffer was passed through a 0.22 mm lter before use. The injection was performed by pressure at 0.5 psi for 5 s. To achieve reproducible separations, all experiments were carried out in triplicate. The BGE solution was refreshed aer 10 runs. The stock solutions of Tulobuterol, Clorprenaline, Tropicamide, Homatropine hydrobromide, Homatropine methyl bromide, Atropine, Atropine methobromide, Anisodamine, Gatioxacin and Ooxacin at concentrations of 1.0 mg L1 were prepared in ultra-pure water. Standard solutions of different concentrations were prepared by dilution of the stock solution with ultra-pure water. All running buffers and analyte solutions were passed through a 0.22 mm syringe lter before injection. 2.2
Computational section
The initial structures of (R/S)-Tulobuterol and (R/S)-Clorprenaline obtained by CS ChemDraw Ultra 12.0 were imported into the Spartan 10 soware program for conformational searching using the MMFF method.24–26 The molecular mechanics calculations supplied minimum energies for all conformers. The optimal energy conformers were selected within a region of 5 kcal mol1. As a more progressive level, the precision of the density functional method is higher than that of molecular mechanics. Therefore, the above conformers were further calculated using the Gaussian 09 soware program at the density functional theory level with a B3LYP functional and a 631G(d,p) basis set. Then, energy values were redistributed according to the Boltzmann distribution, discarding the
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Paper
conformers which possess inappreciable contribution to conformational equilibrium, permitting decline to the most stable conformers of (R/S)-TUL and (R/S)-CLO for further simulations. Furthermore, the b-cyclodextrin structure was obtained from the crystallographic parameters provided by the Cambridge Crystallographic Data Centre, and based on the initial structure, sulfated b-cyclodextrin was constructed and then fully optimized at the PM3 level of theory without imposing any symmetry.28 Subsequently, in order to get the most stable structures of the inclusion complexes, the separately optimized host and guest molecules were used to construct the complexes using a coordinate system (Fig. 2) of the potential energy scanning method. The glycosidic oxygen atoms of S-b-CD were positioned on the XY at surface and their centre was established as the origin of coordinates for the complete system. The reference ˚ on the Z axis, and then the atom was originally situated at 10 A ˚ whole guest molecule entered into the S-b-CD cavity up to 10 A ˚ The structures generated at along the Z axis with a step of 1 A. each step were then optimized, allowing them to change from the initial conformations while keeping the movement of the reference atoms and the S-b-CD structure totally restricted. As during the electrophoretic process chiral compounds t into the CD cavity either completely or with their hydrophobic part only, entering through one of the two openings, two different inclusion orientations were considered. In the rst orientation (Model I), guest molecules passed through the cavity of S-b-CD from its wide side, with the benzene ring leading. In the second orientation (Model II), the benzene ring was docked into the narrow side of the cavity. The interaction energies (DE) between the host and the guest molecules are calculated according to eqn (1),where Ecomplex, Eguest and Ehost stand for the total energy
Analyst
of the inclusion complex and the free energy of the guest and host molecules, respectively. Furthermore, the interaction energy difference (DDE) is calculated on the basis of eqn (2),where DER and DES represent the energies of the R-enantiomer and the S-enantiomer interacting with the host molecule, respectively.
3.1
The process of complexation for R-Tulobuterol with S-b-CD: A (Model I) and B (Model II) represent the guest molecules passing though S-b-CD from its wide and narrow sides, respectively. Reference atoms are labelled with red stars.
This journal is © The Royal Society of Chemistry 2014
(1)
DDE ¼ DER DES
(2)
To further identify the possibility and ease of formation of inclusion complexes, we ran frequency calculations for the above stable complexes at 1 atm and 298 K in water by PM3 methods. It is well known that the enthalpy (H), entropy (S) and Gibbs free energy (G) are the most important parameters in a thermodynamic system, as changes in them can be used to conrm the exothermic or endothermic nature of a reaction, clarify the chemical stabilities, and predict the equilibrium and spontaneity, respectively.29 Therefore, we then calculated the enthalpy change (DH), the entropy change (DS) and the Gibbs free energy change (DG) by using statistical thermodynamics methods. Then, with the purpose of improving the accuracy of the computational modelling and reducing the computational requirement, a rotation was performed to establish the preferred angular orientation of the guest molecules during the inclusion process based on the structural characteristics of the cavity; an ONIOM2 [B3LYP/631G(d,p):PM3] method as a superior level of calculation was used in order to approach the ideal geometry for the above complexes which were obtained by a PM3 method. In our ONIOM2 calculations, TUL and CLO were dened as the layer of high-level, while S-b-CD was dened as the layer of low-level, which were calculated by a B3LYP/631G(d,p) level of theory and a PM3 method, respectively. Eventually, the natural bond orbital (NBO) and frontier molecular orbital (FMO) calculations were carried out for the most stable complexes obtained by the ONIOM2 method. NBO can be used to elucidate the intermolecular interaction between host and guest molecules via the determination of the stabilization energy E(2), which was correlated with the delocalization tendency of electrons on the basis of the perturbation method.30 Furthermore, FMOs were used to investigate the chemical stabilities of the above inclusion complexes.31,32
3.
Fig. 2
DE ¼ Ecomplex (Eguest + Ehost)
Results and discussion Experimental analysis
During the process of our experiments, we investigated the parameters that have an effect on the enantioseparation, including the background electrolyte pH, the concentration of the b-CD derivatives, the buffer concentration, temperature and voltage. As seen from Table 1, the enantiomers of Homatropine hydrobromide, Tulobuterol, Tropicamide, Atropine, Homatropine methylbromide, Atropine methobromide, Clorprenaline and Anisodamine could be effectively separated by simple,
Analyst
View Article Online
Analyst
Paper
Conditions used to achieve baseline resolution for ten chiral drugs with b-CD derivatives as chiral selectors separated by capillary electrophoresis
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Table 1
Chiral drug
b-CD derivative
Buffer concentration (mmol L1)
pH
Voltage (kV)
Temperature ( C)
Resolution
Tulobuterol Clorprenaline Tropicamide Homatropine hydrobromide Homatropine methylbromide Atropine Atropine methobromide Anisodamine Gatioxacin Ooxacin
6-Sulfated-b-CD 6-Sulfated-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 6-Carboxymethyl-b-CD 2,6-Sulfamic acid-b-CD 2,6-Sulfamic acid-b-CD
30 30 30 30 30 30 30 30 40 40
3.0 3.0 3.5 3.5 3.5 3.5 3.5 3.5 3.5 3.5
20 20 25 25 25 25 25 25 30 30
25 25 20 20 20 20 20 20 25 25
6.17 1.50 5.13 6.29 4.67 4.85 4.59 1.13/6.92/2.58 — —
fast and precise methods, but the enantiomeric separations of Gatioxacin and Ooxacin could not be achieved by optimizing the experimental conditions. Among them, the enantioseparations of Tulobuterol and Clorprenaline with S-b-CD as chiral selectors were obtained within short analysis times, and no interferences were observed at the migration times of the peaks corresponding to the two analytes, which demonstrated the selectivity of this method (Fig. 3). The migration orders of the Rand S-enantiomers were identied by adding the S-enantiomer into the racemic mixture. Notably, not only Tulobuterol, but also Clorprenaline followed the trend that R-enantiomers are separated prior to S-enantiomers. Therefore, the S-enantiomers preferably tted into the S-b-CD cavity through its wide or narrow sides, resulting in different interaction energies compared with R-enantiomers. This led to different mobilities for each pair of enantiomers, and thus the enantiomeric separations could be achieved.
Electropherograms of Tulobuterol (A) and Clorprenaline (B) separated by capillary electrophoresis.
Fig. 3
Analyst
3.2
Structures of isolated host and guest molecules
Fig. 4 summarizes the most stable structures of the isolated host and guest molecules optimized by PM3 and DFT methods, respectively. As the difference in geometrical parameters is a primary factor that inuences the chiral separation, which makes the inclusion complex formed by one enantiomer more stable than that formed for the other one, it is necessary to nd the low-lying energy conformers in order to improve the precision for further theoretical calculations. 3.3 Determination of the initial structures of the inclusion complexes The results of the potential energy scanning method showed that the DE varied at different coordinates (Fig. 5). The more negative the interaction energy, the more stable the complex formed by the host–guest molecules and eight DE minima were found for Model I and Model II. As seen from Table 2, the complexes of (R/S)-TUL and (R/S)-CLO possess energies that are lower than the sum of the energy of the isolated guest and host molecules. This indicated that it is favorable to form inclusion complexes for all enantiomers and in both orientations.
Fig. 4 Geometrical structures of R-TUL (A), S-TUL (B), R-CLO (C), SCLO (D), and the S-b-CD (E) optimized by using the DFT and PM3 methods, respectively.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper
Analyst
Therefore, inclusion complexes A, B, C and D were the initial structures obtained using this calculational method. Meanwhile, the above results claried that the interaction energy difference between a pair of enantiomers with a selector was a signicant factor contributing to the enantiomeric separation.
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
3.4
Thermodynamic analysis
The thermodynamic parameters of the complexation process, including enthalpy change (DH), entropy change (DS) and Gibbs free energy change (DG) are presented in Table 3. Clearly, the interactions between the enantiomers and S-b-CD were exothermic as evidenced by the large enthalpy decreases. Furthermore, the DS and DG were also negative for the inclusion processes of forming the most stable complexes for TUL/ CLO with S-b-CD, indicating that the formations of the complexes were enthalpy-driven and spontaneous procedures. The statistical thermodynamic parameters for the above obtained stable inclusion complexes revealed a propensity to form inclusion complexes between host and guest molecules. Thus, the main factors which contributed to the thermodynamics are associated with the insertion of the guest molecules into the S-b-CD cavity. 3.5 Interaction energies of R-TUL (A), S-TUL (B), R-CLO (C), and SCLO (D) entered into S-b-CD from its wide (Model I) and narrow (Model II) sides.
Fig. 5
Furthermore, comparing which side, wide or narrow, the guest molecules entered the cavity from, it is clear that the interaction energies of Model I complexes are lower than those for Model II complexes for all enantiomers, demonstrating that complexes formed by guest molecules which passed through from the wide side are more stable and available, and thus we selected Model I for further calculations. On the other hand, compared with the DE obtained from the wide orientation, the DDE of the interactions between R-enantiomers and S-enantiomers with S-b-CD for TUL and CLO were 10.58 kJ mol1 and 9.68 kJ mol1, respectively. This revealed that the complexes which were formed by S-enantiomers and S-b-CD were more stable than those formed by R-enantiomers, that is, the S-enantiomers demand slightly more energy than R-enantiomers for the purpose of accommodating to their structures within the S-bCD cavity, and thus the R-enantiomer can be separated rst by CE, and the S-enantiomer can be separated aerwards.
Table 2
Accurate optimization of the inclusion complexes
Although the PM3 method was able to explain the mechanism of enantiomeric separation, it seems that the result was not accurate. If we used a higher level of theory, some more accurate results might be obtained. However, for such a large system, higher theoretical methods would suffer from a high computational requirement. Therefore, for the sake of overcoming the above deciencies, a multi-layered ONIOM method was used during the process of our theoretical studies. This QM-MM approach not only retains the reliability of quantum mechanics, but also possesses the property of fast calculations for
Table 3 Statistical thermodynamic parameters of Model I calculated by PM3 methods
S-b-CD complex
DHa (kJ mol1)
DSb (J mol1 K1)
DGc (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
28.53 38.77 34.28 45.16
58.82 75.19 62.07 81.94
10.96 16.31 15.49 20.73
a DH ¼ Hcomplex Hhost Hguest. b D ¼ Scomplex Shost Sguest. c DG ¼ DH TDS.
The DE and DDE of the inclusion complexes formed by Tulobuterol and Clorprenaline with S-b-CD
S-b-CD complex
DEModel Ia (kJ mol1)
DEModel IIa (kJ mol1)
DDEModel Ib (kJ mol1)
DDEModel IIb (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
30.25 40.83 37.60 47.28
26.97 35.18 34.11 42.25
10.58
8.21
9.68
8.14
a
DE ¼ Ecomplex (Eguest + Ehost).
b
DDE ¼ DER DES.
This journal is © The Royal Society of Chemistry 2014
Analyst
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Analyst
Paper
macromolecular systems. Table 4 summarizes the results of the ONIOM2 calculations, indicating that the total energies of the inclusion complexes calculated by the ONIOM2 method (EONIOM2) are more negative than those obtained from the PM3 method (EPM3), demonstrating that the higher levels of theory lead to more rational and stable structures for the complexes. Furthermore, the DDE of the interactions between R-enantiomers and S-enantiomers with S-b-CD were 15.41 kJ mol1 and 10.92 kJ mol1 for TUL and CLO respectively, and thus the inclusion complexes formed by S-enantiomers were more stable than those formed by R-enantiomers. The more stable the inclusion complex, the longer the migration time and therefore the R-enantiomer is prone to separate rst, which is in agreement with the enantiomer migration orders observed in the CE experiment. The results of the above theoretical calculations show that the energy difference in the interactions between a pair of enantiomers and a host makes for different migration times, and thus enantioseparation can be achieved. In addition, we found that the benzene rings of the guest enantiomers come very close to the centre of the S-b-CD cavity as a result of hydrophobic interactions between the hydrophobic cavity and the benzene ring, making the inclusion complexes more stable than free host and guest molecules (Fig. 6).
3.6
Hydrogen bonding interactions
The detailed results of the E(2) associated with hyperconjugative interactions are summarized in Table 5, which can be utilized to quantify the range of hydrogen bonding interactions. It suggested that, in general, the E(2) value of strong hydrogen bonding is larger than 2.0 kcal mol1, while the E(2) value of a weak hydrogen bonding interaction is determined to be from 0.5 kcal mol1 to 2.0 kcal mol1. The stabilization energies E(2) of the R-TUL and R-CLO inclusion complexes were all larger than 2.0 kcal mol1, indicating the presence of four strong hydrogen bonding interactions. In addition, the E(2) for the intermolecular C–H/O interaction [LP(O125):BD*(O190– H205)] in the S-TUL complex was also larger than 2.0 kcal mol1, which demonstrated a strong hydrogen bond. Furthermore, both the E(2) of the C–H/O interaction in the S-CLO complex and the E(2) for the intermolecular C–H/O interaction [LP(O190):BD*(O7–H8)] in the S-TUL complex were found to be smaller than 0.5 kcal mol1, revealing two very weak hydrogen bonding interactions. This donor–acceptor electron transfer model of NBO calculations reveals that the stabilizations of the above four stable inclusion complexes are entirely inuenced by
Structures of inclusion complexes obtained from ONIOM2 calculations for R-TUL (A), S-TUL (B), R-CLO (C), and S-CLO (D) with Sb-CD.
Fig. 6
intermolecular hydrogen bonding interactions which arise from orbital overlap. Therefore, we conrmed that the formation of intermolecular hydrogen bonding, combined with hydrophobic interactions, was the mechanism for the stabilization of the inclusion complexes. The detailed intermolecular hydrogen bonding interactions are shown in Fig. 7, as indicated by dashed lines.
3.7
Frontier molecular orbitals
The highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) are named the frontier molecular orbitals, and they can be used to gain knowledge of the nature of reactivities, and of some of the structural and physical properties of molecules. The energy difference between the HOMO and the LUMO, called the energy gap, is a critical parameter for determining chemical reactivities, kinetic stabilities and the hardness of molecules. Therefore, in order to investigate the chemical stabilities of the above inclusion complexes, the energies of the HOMOs and LUMOs and their orbital energy gaps were calculated using the above ONIOM2 method, and the results are listed in Table 6. A pictorial illustration of the FMOs for the S-Tulobuterol/S-b-CD complex is shown in Fig. 8; their respective positive and negative regions are represented by red and green colors. The values of the energy separations between the HOMOs and LUMOs were found to be 7.41, 8.73, 5.82 and 6.79 eV for R-
Table 4 The EPM3, EONIOM2, DEONIOM2 and DDEONIOM2 of the inclusion complexes formed by Tulobuterol and Clorprenaline with S-b-CD
S-b-CD complex
EPM3 (kcal mol1)
EONIOM2 (kcal mol1)
DEONIOM2a (kJ mol1)
DDEONIOM2b (kJ mol1)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
3063.74 3066.27 3059.66 3061.80
667 042.68 667 050.64 642 371.40 642 478.99
33.28 48.69 35.26 46.18
15.41
a
DE ¼ Ecomplex (Eguest + Ehost).
Analyst
b
10.92
DDE ¼ DER DES.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Paper Table 5
Analyst The results of the NBO calculations
Table 6
d (2)a
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Donor i
Acceptor j
E (kcal mol1)
O/H
˚) (A
Angleb ( )
R-Tulobuterol/S-b-CD complex LP(1)O3 BD*(1)O189–H205 LP(2)O3 BD*(1)O189–H205 LP(1)O189 BD*(1)C9–H10 LP(2)O189 BD*(1)C9–H10
0.31 11.00 0.47 19.31
1.96
125.51
2.03
123.62
S-Tulobuterol/S-b-CD complex LP(1)O125 BD*(1)O190–H205 LP(2)O125 BD*(1)O190–H205 LP(2)O190 BD*(1)O7–H8
0.11 15.05 0.42
1.92
135.26
2.39
109.23
R-Clorprenaline/S-b-CD complex LP(1)O3 BD*(1)O189–H205 LP(2)O3 BD*(1)O189–H205 LP(1)O189 BD*(1)C9–H10 LP(2)O189 BD*(1)C9–H10
0.25 14.50 0.53 20.37
1.89
120.73
1.95
121.12
S-Clorprenaline/S-b-CD complex LP(1)O3 BD*(1)O190–H205 LP(2)O3 BD*(1)O190–H205 LP(2)O190 BD*(1)O29–H30 LP(2)O190 BD*(1)O29–H30
0.20 0.93 0.39 0.05
2.48
116.20
2.16
91.51
a
The energies of ELUMO, EHOMO, and DE
S-b-CD complex
ELUMO (eV)
EHOMO (eV)
DEa (eV)
R-Tulobuterol S-Tulobuterol R-Clorprenaline S-Clorprenaline
1.21 1.24 1.43 1.40
8.62 9.97 7.25 8.19
7.41 8.73 5.82 6.79
a
DE ¼ EHOMO ELUMO.
E(2) is the energy of hyperconjugative interaction. b The angle of X–H/Y.
The highest occupied and lowest unoccupied molecular orbitals of the S-Tulobuterol/S-b-CD complex.
Fig. 8
energy gaps than the R-enantiomer complexes, which indicates that the S-enantiomer inclusion complexes are more stable and that their electron density changes more easily than that of the R-enantiomer complexes. 3.8
Detailed intermolecular hydrogen bonding interactions, obtained from NBO calculations for R-TUL (A), S-TUL (B), R-CLO (C), and S-CLO (D) with S-b-CD.
Fig. 7
Tulobuterol, S-Tulobuterol, R-Clorprenaline, and S-Clorprenaline complexes, respectively. The large HOMO–LUMO gaps specify high excitation energies of the excited states, good stabilities and a high chemical hardness. We can conclude that the S-enantiomer complexes possess lower HOMO–LUMO
This journal is © The Royal Society of Chemistry 2014
Quantitative studies of ten chiral drugs separated by CE
In the present work, besides the two computational models, a variety of chiral drugs (Table SI†), separated by CE experiments with b-cyclodextrin derivatives as selectors, were investigated by using a similar theoretical method, and an essential condition for achieving enantiomeric separation was discovered. It is shown in Table 7 that the DDE of Homatropine hydrobromide, Tulobuterol, Tropicamide, Atropine, Homatropine methylbromide, Atropine methobromide, Clorprenaline, and Anisodamine enantiomers interacting with b-cyclodextrin derivatives are 17.38, 15.41, 12.60, 11.82, 10.27, 9.53, 10.92 and 6.05/18.97/ 9.22 kJ mol1, respectively. The DDE are all larger than 6 kJ mol1, and the chiral selectors could separate these chiral compounds with relatively ideal resolutions (6.29, 6.17, 5.13, 4.85, 4.67, 4.59, 1.50 and 1.13/6.92/2.58, respectively). The energy differences of the interactions between the enantiomers
Analyst
View Article Online
Analyst
Paper
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Table 7 The DDE and Rs for ten chiral drugs separated by CE
Chiral drugs
Substituent groupsa
DDE (kJ mol1)
Rs
Tulobuterol Clorprenaline Tropicamide Homatropine hydrobromide Homatropine methyl bromide Atropine Atropine methobromide Anisodamine Gatioxacin Ooxacin
6-Sulfated 6-Sulfated 6-Carboxymethyl 6-Carboxymethyl
15.41 10.92 12.60 17.38
6.17 1.50 5.13 6.29
6-Carboxymethyl
10.27
4.67
6-Carboxymethyl 6-Carboxymethyl 6-Carboxymethyl 2,6-Sulfamic acid 2,6-Sulfamic acid
11.82 9.53 6.05/18.97/9.22 5.83 5.96
4.85 4.59 1.13/6.92/2.58 — —
a
The substituent groups of b-cyclodextrin.
and the chiral selectors represent the energetic contribution to enantiomeric separation. Gatioxacin and Ooxacin exhibited single electrophoretic peaks, indicative of no separation, and the DDE were 5.83 and 5.96 kJ mol1, respectively. Furthermore, compared with other chiral drugs (Table SI†) separated by bcyclodextrin derivatives, we discovered that the energy difference of the interactions between a pair of enantiomers with a chiral additive must be greater than or approximately equal to 6 kJ mol1 in order to achieve enantiomeric separation. This can be used to predict the possibility of chiral separation before performing experiments.
4. Conclusions In this paper we put forward a novel explanation for CE, that the different interaction energies play a signicant role in the enantiomeric separation. The more stable the inclusion complex, the longer the migration time and therefore we can predict the enantiomer migration orders through theoretical calculations. Simultaneously, we found that the energy difference of the interactions between each pair of enantiomers with a chiral additive must be greater than or approximately equal to 6 kJ mol1 in order to achieve chiral separation. Thus, we can predict the possibility of enantiomeric separation, and design selectors to separate specic chiral drugs by molecular simulations before performing experiments. Meanwhile, we will also be able to synthesize new selectors for separating chiral drugs, which could be effectively used to reduce experimental requirements and increase separation efficiencies.
Acknowledgements This work was supported by the National Natural Science Foundation of China (no. 81273476). The authors of this manuscript acknowledge the Innovative Research Team of the Ministry of Education and Program for Liaoning Innovative Research Team in University for providing nancial assistance. The theoretical calculations were conducted on the ScGrid and Deepcomp 7000 the Supercomputing Center, Computer Network Information Center of Chinese Academy of Sciences.
Analyst
Notes and references 1 V. Subramaniam, L. Griffith and A. J. Haes, Analyst, 2011, 136, 3469–3477. 2 R. J. Flaherty, B. J. Huge, S. M. Bruce, O. O. Dada and N. J. Dovichi, Analyst, 2013, 138, 3621–3625. 3 L. R. Vemireddy, S. Archak and J. Nagaraju, J. Agric. Food Chem., 2007, 55, 8112–8117. 4 E. Junquera and E. Aicart, J. Pharm. Sci., 1999, 88, 626–631. 5 S. Kodama, A. Yamamoto, T. Ohura, A. Matsunaga and T. Kanbe, J. Agric. Food Chem., 2003, 51, 6128–6131. 6 S. R. Gratz and A. M. Stalcup, Anal. Chem., 1998, 70, 5166– 5171. 7 A. M. Rizzi and L. Kremser, Electrophoresis, 1999, 20, 2715– 2722. 8 A. Rizzi, Electrophoresis, 2001, 22, 3079–3106. 9 B. Chankvetadze, W. Linder and G. K. E. Scriba, Anal. Chem., 2004, 76, 4256–4260. 10 B. Chankvetadze, J. Chromatogr. A, 2007, 1168, 45–70. 11 G. K. E. Scriba, J. Sep. Sci., 2008, 31, 1991–2011. 12 A. A. Elbashir and F. O. Sulimana, J. Chromatogr A, 2011, 1218, 5344–5351. 13 S. L. Ma, S. Shen, N. Haddad, W. J. Tang, J. Wang, H. Lee, N. Yee, C. Senanayake and N. Grinberg, J. Chromatogr A, 2009, 1216, 1232–1240. 14 S. Stachowicz, A. Styrcz, J. Korchowiec, A. Modaressi and M. Rogalski, Theor. Chem. Acc., 2011, 130, 939–953. 15 J. H. Shi, Z. J. Ding and Y. Hu, J. Mol. Model., 2012, 18, 803– 813. 16 N. Leila, H. Sakina, A. Bouhadiba and F. Madi, J. Mol. Liq., 2011, 160, 8–13. 17 G. L. Alexandrino, A. Calderini, N. H. Morgon and F. B. T. Pessine, J. Inclusion Phenom. Macrocyclic Chem., 2013, 75, 93–99. 18 J. R. Garbow, J. J. Likos and S. A. Schroeder, J. Agric. Food Chem., 2001, 49, 2053–2060. 19 T. N. Doman, C. R. Landis and B. Bosnich, J. Am. Chem. Soc., 1992, 114, 7264–7272. 20 W. S. Cai, T. T. Sun, P. Liu, C. Chipot and X. G. Shao, J. Phys. Chem. B, 2009, 113, 7836–7843.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Published on 15 October 2014. Downloaded by Universiteit Utrecht on 28/10/2014 02:40:37.
Paper
21 P. Walla, V. B. Arion and U. H. Brinker, J. Org. Chem., 2006, 71, 3274–3277. 22 T. J. Cleij, S. K. Y. Tsang and L. W. Jenneskens, Macromolecules, 1999, 32, 3286–3294. 23 A. W. Lange, M. A. Rohrdanz and J. M. Herbert, J. Phys. Chem. B, 2008, 112, 7345. 24 W. J. Hehre, L. Radom, P. V. R. Schleyer and J. A. Pople, Ab Initio Molecular Orbital Theory, Wiley., New York, 1986. 25 W. J. Hehre, A. J. Shusterman and W. W. Huang, A Laboratory Book of Computational Organic Chemistry, Wavefunction Inc., California, 1996. 26 W. J. Hehre, A Guide to Molecular Mechanics and Quantum Chemical Calculations, Wavefunction Inc., California, 2003. 27 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd,
This journal is © The Royal Society of Chemistry 2014
Analyst
28 29 30 31 32
E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, GAUSSIAN 09 (Revision A.1), Gaussian, Inc., Wallingford, CT, 2009. T. Aree and N. Chaichit, Carbohydr. Res., 2002, 337, 2487– 2494. H. Hayashi and K. Kojima, Ind. Eng. Chem. Res., 1994, 33, 1631–1635. Y. Tsujii, A. Tsuchida, M. Yamamoto and Y. Nishijima, Macromolecules, 1988, 21, 665–670. P. Bultinck, D. V. Neck, G. Acke and P. W. Ayers, Phys. Chem. Chem. Phys., 2012, 14, 2408–2416. M. D. King, C. E. C. Mas and R. P. Wayne, Phys. Chem. Chem. Phys., 1999, 1, 2231–2238.
Analyst
