2744 Yiying Wang Jie Zhou Yun Liu Jian Tang Weihua Tang Key Laboratory of Soft Chemistry and Functional Materials (Ministry of Education of China), Nanjing University of Science and Technology, Nanjing, P. R. China
Received April 14, 2014 Revised June 22, 2014 Accepted June 22, 2014
Electrophoresis 2014, 35, 2744–2751
Research Article
Evaluation of the chiral separation ability of single-isomer cationic -cyclodextrins in capillary electrophoresis Eight single-isomer ammonium--cyclodextrin derivatives with different side chains were successfully developed as chiral selectors for the chiral separation of selected racemates in capillary electrophoresis. The number of substituted groups at N-atom as well as the alkyl chain length greatly influenced the chiral separation. With the numbers of hydroxylalkyl groups at N-atom growing, the aqueous solubility of resolving agents were distinctly decreased and chiral separation ability was also significantly reduced. The apparent complex stability constants between CDs and analytes were further determined. The best enantioseparations of hydroxyl acids was achieved with the use of mono-6A -(3hydroxypropyl)-1-ammonium--cyclodextrin chloride and mono-6A -(3-methoxypropyl)-1ammonium--cyclodextrin chloride. The nuclear magnetic resonance experiments were carried out using them with mandelic acid as guest molecules, revealing the inclusion pattern as well as electrostatic interactions and hydrogen bonding interactions as additional chiral driving force. The contribution of potential interaction sites in the sidearm could enhance the enantioseparations. Keywords: Chiral separation / Hydrogen bonding / Single-isomer cationic cyclodextrin DOI 10.1002/elps.201400198
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Chiral separation has gained extensive attention in the fields of pharmaceutical [1–4], clinical biological [5–7], environmental, and food analysis [8–12]. CE has proven to be a powerful analytical technique affording low consumption of samples
Correspondence: Dr. Weihua Tang, Key Laboratory of Soft Chemistry and Functional Materials (Ministry of Education of China), Nanjing University of Science and Technology, Nanjing 210094, P. R. China E-mail: [email protected] Fax: +86-25-84317311
Abbreviations: BHEtAMCD, mono-6A -bis(2-hydroxyethyl)-1ammonium--cyclodextrin chloride; BMEtAMCD, mono-6A bis(2-methoxyethyl)-1-ammonium--cyclodextrin chloride; HBuAMCD, mono-6A -(4-hydroxybutyl)-1-ammonium--cyclo dextrin chloride; HEtAMCD, mono-6A -(2-hydroxyethyl)-1ammonium--cyclodextrin chloride; HPrAMC, mono-6A -(3hydroxypropyl)-1-ammonium--cyclodextrin chloride; MEt AMCD, mono-6A -(2-methoxyethyl)-1-ammonium--cyclode xtrin chloride; MPrAMCD, mono-6A -(3-methoxypropyl)-1ammonium--cyclodextrin chloride; THEtAMCD, mono-6A tri(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
and solvents, high efficiency, short analysis time [13–18]. To date, CDs and their derivatives proved to be the most effective chiral selectors in CE separations attributable to their favorable enantioseparation properties and relatively low costs [19–21]. CD can be modified via mono- or multisubstitution of hydroxyl groups on the rims. However, the development of randomly multisubstituted CD derivatives is limited due to poor reproducibility in synthetic process and relative uncertainty in practical applications. Hence, the demand for effective chiral selectors has led to an outburst of new strategies toward single-isomer CD derivatives [22]. Among of these certainly functionalized CDs, ionizable CDs are more versatile than neutral ones since they are able to separate both charged and neutral enantiomers [23–25]. Systematic study of anionic sulfated single-isomer CDs was undertaken by Vigh’s group [26–31]. Negatively charged CD derivatives can interact strongly with basic compounds and high migration differences of enantiomers are also observed [32–34]. In terms of cationic CD derivatives, they have not been used as much as anionic ones probably due to their adsorption character to the column wall and relatively complicated
Colour Online: See the article online to view Figs. 1–7 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2744–2751
synthesis procedure. Despite that, the use of cationic CDs in CE can lower the migration time of analytes compared with anionic ones. The majority of positively charged CD derivatives are substituted in the C6 position by quaternary ammonium groups or weak electrolytes [35–40]. The chiral separation in CE using CDs as chiral selectors is fundamentally based on the formation of complexation (inclusion or external) and various additional interactions between analytes and functional moieties in substituted groups [19, 41]. Accordingly, the position and structure of only one substituent can greatly influence the enantioseparation ability of a CD derivative. In terms of single-isomer cationic ammonium--cyclodextrin derivatives, the length and number of alkyl chain would play a principle role in the separation process [42]. On the basis of our previous work [37, 38, 43], a series of versatile CD derivatives with (methoxy)alkylamino or (hydroxy)alkylamino substituents have been synthesized in this study. Mono-6A bis(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride, mono-6A -tri(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride, and mono-6A -bis(2-methoxyethyl)-1-ammonium-cyclodextrin chloride were first reported as chiral selectors in CE. The eight CD derivatives were used for the separation of six selected analytes at different concentrations. Based on the separation results, the influence of substituted groups on the enantioseparation ability of these single-isomer CD derivatives was demonstrated. The apparent complex’s stability constants of CD and analytes were also investigated. In particular, the existence of hydrogen bonding during chiral separation was revealed by NMR measurements. This work presents a good correlation study between the alkylammonium substituted CDs and their enantioseparation ability, especially the alkyl chain length and number of substituents on the cationic center of the CD rim. The theoretical determination of binding constants and NMR study prove to be useful approaches in explanation of macroscopic chiral separation in CE experiments.
2 Materials and methods 2.1 Materials and chemicals -Cyclodextrin was obtained from Tokyo Chemical Industry (TCI, Japan). All the racemates were purchased from Sigma (St. Louis, MO, USA). The CD derivatives were synthesized in the lab according to our previous approach [36]. 6-Monotosyl-CD and appropriate amount of alkylamine were added into dry DMF and reacted for 5–12 h at 90°C. The acetone was added into the solution to precipitate the white solid. The crude product was obtained via further chloride ion exchange. All the CDs were further purified via recrystallization in DIwater before characterization and chiral separation. The structures of analytes and CD derivatives are depicted in Fig. 1 (synthesis of new CDs is provided in Supporting Information). C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2745
2.2 Apparatus and CE procedure NMR measurements were performed on a Bruker AVANCE 500 (500 MHz, Bruker Daltonics, Bremen, Germany). The 1D TOCSY NMR studies were conducted in D2 O for both CD/R-mandelic acid and CD/S-mandelic acid complexes. All CE separations were accomplished on P/ACE MDQ CE unit (Fullerton, CA, USA), equipped with an uncoated fused–silica capillary (50 m id, 363 m od, 49 cm effective length). All samples were injected at 3.45 kPa for 5 s and detected at 214 nm assisted with an online photodiode array (PDA) detector. During the separation process, the initial BGE was composed of 50 mM NaH2 PO4 solution with appropriate pH. The CD solutions were obtained via dissolving appropriate amount of CD derivatives into the BGEs. Stock analytes were set at 50 mg/mL with a mixture solvent of methanol/water (50/50, v/v) and stored in the refrigerator at +2°C. All buffers and sample solutions were undertaken pretreatment of filtration with a 0.45 m syringe-type Millipore membrane and ultrasonication prior to use.
2.3 Calculation of complex stability constants Direct chiral separation in CE is based on an electrophoretic technique as well as a chiral environment [43, 44]. During the separation, the chiral selectors could stereoselectively interact with enantiomers of analytes to form temporary diastereomeric association. Thus, the slight difference in the formation of reversible association will generate a mobility difference, resulting in a successful separation. For CE separation with charged CD derivatives, the mobility greatly depends on difference in binding constants for enantiomers with selectors. Assuming a 1:1 interaction between the enantiomer and CD, the correlation of the effective mobility with CD concentration can be described as the following equation [45]: eff =
f + c K [C] 1 + K [C]
(1)
Where eff is the effective mobility of enantiomers, f is the mobility of analytes in free form, c is the mobility of complex, [C] is the concentration of CD selector and K is the binding constants for enantiomers with CD. Equation (1) can be transformed into the same mathematical form as shown below [46]: eff − f = −K (eff − f ) + K (c − f ) [C]
(2)
According to Eq. (2), the apparent complex stability constants can be determined via the x-reciprocal method by (graphical) solution of plotting form of (eff − f )/[C] versus (eff − f ), which affords −K as slope. www.electrophoresis-journal.com
2746
Y. Wang et al.
Electrophoresis 2014, 35, 2744–2751 OH
OH C * COOH
OH CH2C * COOH
C * COOH
H
Cl
Mandelic acid (MA)
H
H
2-Chloromandelic acid (2-Cl MA)
3-Phenyllactic acid (3-PLA) O
H
H
C * COOH
S
O C * COOH
CH2CH3
O
H3C
CH3
H NH C* COOH CH2CH3
N H3C
2-Phenylbutyric acid (2-PBA)
OH
6
ClNH2
OH
(OH)14
2-Phenoxypropionic (2-POPA)
OH
6
Cl NH2
OH
OH
Cl 6
NH2
OH
Cl
OH
6
OH N H
OH
(OH)14
(OH)14
(OH)14 HPrAMCD
HEtAMCD
Dansyl-DL- -aminonbutyric acid (Dns-Aba)
BHEtAMCD
HBuAMCD
OH OH
Cl 6
(OH)14 THEtAMCD
N
OH
OH
Cl 6
NH2
OCH3
OH
Cl 6
NH2
OCH3
OH
Cl 6
OH
(OH)14 MEtAMCD
(OH)14
In this study, five single-isomer 6-mono-(hydroxyl)alkylammonium--CD (mono-6A -(2-hydroxyethyl)-1-ammonium-cyclodextrin chloride (HEtAMCD); mono-6A -(3-hydroxypropyl)-1-ammonium--cyclodextrin chloride (HPrAMCD); mono-6A -(4-hydroxybutyl)-1-ammonium--cyclodextrin chloride (HBuAMCD); BHEtAMCD and THEtAMCD) and three mono-6A -methoxyalkylammonium--CD, mono-6A -(2-methoxyethyl)-1-ammonium--cyclodextrin chloride (MEtAMCD), mono-6A -(3-methoxypropyl)-1 - ammonium -  - cyclodextrin chloride (MPrAMCD) and BMEtAMCD (see Fig. 1) were applied to study the influence of substituents on the enantioseparations. These amino--CD derivatives were obtained with good reproducibility in synthesis and CE separation. All CDs exhibited good aqueous solubility except THEtAMCD, which was only soluble up to 5 mM in the running buffer. Six enantiomeric pairs (2-POPA, 2-PBA, 3-PLA, MA, 2-Cl MA, Dns-Aba) were investigated and their physical and chemical properties are summarized in Table 1.
3.1 Enantioseparations with different CD at varied concentrations If the CD concentration is zero or very large, the apparent mobility difference will be zero, indicating that an appropriate CD concentration would afford an optimal C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
OCH3 OCH3
(OH)14 BMEtAMCD
MPrAMCD
3 Results and discussion
N H
Figure 1. Chemical structures of racemates and CD derivatives studied.
Table 1. Physical and chemical properties of the six acids studied
Racemates
pKaa)
Molar volume/cm3b)
Mandelic acid 2-Chloromandelic acid 3-Phenyllactic acid 2-Phenylbutyric acid 2-Phenoxypropionic acid Dansyl-DL-␣-aminonbutyric acid
3.41 ± 0.2 3.30 ± 0.1 3.72 ± 0.1 4.34 ± 0.1 3.22 ± 0.1 3.30 ± 0.1
115.1 ± 3.0 127.0 ± 3.0 131.2 ± 3.0 150.6 ± 3.0 141.4 ± 3.0 256.6 ± 3.0
a) All pKa values were determined by titrimetric experiments. b) All values were calculated by ACD/ChemSketch v6.0.
separation of enantiomers. In order to study the effect of concentration on the chiral separation, carboxylic acids (2-POPA, 2-PBA), hydroxy acids (3-PLA, MA, 2-Cl MA), and dansyl amino acid (Dns-Aba) were enantioseparated by gradually increasing CDs concentration from 1 to 20 mM with pH 6.0 BGEs. Most cases can be baseline separated except for BHEtAMCD, THEtAMCD, and BMEtAMCD (see Fig. 2). The selectivity of 2-POPA was significantly increased with higher CD concentration. HPrAMCD and BMEtAMCD could afford 1.5 of selectivity for 2-POPA at 20 mM. In the case of 2-POPA, the superior order for chiral resolution of CD with hydroxyalkyl groups was HPrAMCD ⬎ HBuAMCD ⬎ HEtAMCD. HPrAMCD and MPrAMCD with three-carbon alkyl chain exhibited the best chiral ability, achieving chiral resolution over 5 when the concentrations were higher than 5 mM. And the maximum chiral resolutions were obtained at www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2744–2751
CE and CEC
2747
Figure 2. Chiral resolution (Rs ) of analytes obtained with single-isomer cationic CDs at varied concentrations.
approximate 20 mM for HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD. Interestingly, the chiral resolution even increased linearly with increased concentration for HEtAMCD from 5 to 20 mM. In terms of 2-PBA, a maximal resolution was obtained at 5 mM CD (except THEtAMCD and BMEtAMCD), with chiral resolution over 1.05. It was obvious that the number of substituted groups greatly influenced the chiral resolution. BHEtAMCD and BMEtAMCD with two alkyl chains could not enantioseparate 2-PBA at low concentration, meanwhile, THEtAMCD possessing three alkyl chains on N-atom could not even resolve 2-PBA at all studied concentrations. This indicated that steric hindrance may exist in the formation of temporary diastereomeric association. Taking the alkyl chain length into account, HEtAMCD was the best choice compared with HPrAMCD and HBuAMCD for 2-PBA. Meanwhile, MEtAMCD could afford better selectivity than MPrAMCD at most concentrations. Hydroxyl acids showed increased chiral resolution at higher concentration of chiral selectors for most CDs studied. Among of them, HPrAMCD could resolve them with highest selectivity. When the concentration of HPrAMCD reached
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
20 mM, all the selectivities were over 1.5. The representative electropherograms of 2-POPA and MA using different concentrations of HPrAMCD were depicted in Fig. 3. It revealed that both selectivity and chiral resolutions increased with higher CD concentrations, though the shortest elution time was obtained at 3 mM. Additionally, analytes have more affinity to different CD selectors. For example, 3-PLA and MA were better resolved by HBuAMCD than MPrAMCD, while MPrAMCD could afford better enantioselectivity toward 2-Cl MA. HPrAMCD could resolve MA and 2-Cl MA with a resolution as high as over 18 at 20 mM, meanwhile, MPrAMCD can afford them a resolution over 10. It was also found that CD BMEtAMCD exhibited more excellent chiral ability than BHEtAMCD toward three hydroxyl acids, indicating that the introduction of methoxy group on alkyl chain sometimes could enhance the chiral separation compared to hydroxyl group. In our study, one dansyl amino acid (DnsAba) was also applied in CE separation. All CDs could afford higher selectivity with increased selectivity during concentration range applied, especially that of HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD. HPrAMCD and HBuAMCD exhibited similar selectivity, reaching over 2.0 at 20 mM.
www.electrophoresis-journal.com
2748
Y. Wang et al.
Figure 3. Representative chromatograms of (A) 2-POPA and (B) MA using HPrAMCD at different concentrations.
However, the chiral resolution obtained by HEtAMCD and HPrAMCD was much higher than that by HBuAMCD. Meanwhile, BHEtAMCD and BMEtAMCD only partially resolve it and THEtAMCD could not achieve baseline separation. The electropherograms at lower CD concentrations (ࣘ3 mM) exhibit certain drift in baseline, which may be due to the absorption of cationic CDs onto capillary inner wall to cause the disturbance to the system. Based on the enantioseparations data, the number and length of alkyl chain at ammonium center was of great importance for chiral resolution. The comparable resolutions were shown in Fig. 4. In this case of separation with 10 mM of HEtAMCD, HPrAMCD, and HBuAMCD, the superior order of single alkyl chain with hydroxyl group was HPrAMCD ⬎ HBuAMCD ⬎ HEtAMCD (see Fig. 4A). 2-Cl MA approximately achieved two folds chiral resolution with HPrAMCD over HEtAMCD, indicating a proper alkyl chain length would influence the enantioseparations. Compared the chiral resolutions obtained by HEtAMCD, BHEtAMCD, and THEtAMCD, HEtAMCD with only single hydroxyamino group proved to be the superior choice for all analytes (see Fig. 4B). HEtAMCD could afford a resolution of 5.0 toward MA, while BHEtAMCD and THEtAMCD could not resolve it. With the number of substituted groups at N-atom increased, the water solubility and separation ability dramatically declined. THEtAMCD with three alkyl chains on N-atom was the worst C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2744–2751
Figure 4. Comparison of chiral resolutions for six acids with (A) HEtAMCD, HPrAMCD, and HBuAMCD at 10 mM concentration; and (B) HEtAMCD, BHEtAMCD, and THEtAMCD at 5 mM concentration.
chiral selector in this comparison study. This might be explained by the impeded complex stability constants through steric hindrance.
3.2 Calculation of binding constants, apparent complexation selectivity, theoretical optimum selector concentration The apparent complex stability constants (K) were calculated from the measured mobility values in the applied concentration range according to the x-reciprocal method [47] (see Table 2). Since cationic CDs used in CE can cause change in solution viscosity and ionic strength in the BGE, all mobilities for the modeling here were corrected by using the migration times of analytes referenced to zero CD concentration using neutral marker (MeOH). For the plots of (eff − f )/[C] against (eff − f ), the correlation coefficient (R) of the fitting curves could reach at least 0.974, 0.950, 0.969, 0.995, 0.998, and 0.991 for 2-POPA, 2-PBA, 3-PLA, MA, 2-Cl MA, and Dns-Aba with HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD, respectively. Representative example of the fitting curve for Dns-Aba and 2-POPA using MPrAMCD was shown in Fig. 5. www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2744–2751
2749
Table 2. Apparent complex stability constants (K1 and K2 ), apparent complexation selectivity (␣cplx ), maximum resolution (Rs max ), and calculated optimal concentration (copt ) of CD studied
Analytes
3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA ␣cplx 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD
HPrAMCD
HBuAMCD
MPrAMCD
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
87 36 31 210 58 119
121 47 44 229 71 132
124 35 73 186 67 121
131 40 84 191 87 129
87 22 52 178 32 160
113 34 56 181 45 171
89 45 53 139 78 134
108 53 63 145 93 140
HEtAMCD 1.39 1.31 1.42 1.09 1.22 1.11
HPrAMCD
HBuAMCD
1.06 1.14 1.15 1.03 1.30 1.07
1.30 1.55 1.08 1.02 1.41 1.07
MPrAMCD 1.21 1.18 1.19 1.04 1.19 1.04
Rs max 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD 6.05 (10) 8.16 (20) 10.28 (20) 2.96 (20) 6.63 (20) 1.67 (5)
HPrAMCD 8.45 (10) 18.57 (20) 19.88 (20) 3.90 (20) 9.21 (20) 1.56 (5)
HBuAMCD 6.22 (20) 10.05 (20) 10.05 (20) 3.16 (15) 7.45 (20) 1.63 (5)
MPrAMCD 9.95 (15) 11.19 (20) 14.36 (20) 3.83 (20) 8.70 (20) 1.48 (5)
Copt 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD 10 24 27 5 16 8
HPrAMCD 8 27 13 5 13 8
HBuAMCD 10 37 19 6 26 6
MPrAMCD 10 20 17 7 13 7
Complex stability constants (K1 and K2 ) calculated from 4–7 different concentrations between 0–20 mM by linear regression. Each data point was the mean of at least three injections. Rs max represented the maximum chiral resolution obtained in the experiments. The apparent complexation selectivity (␣cplx ) was calculated according to equation: K1 /K2 ; Theoretical optimum selector concentration (copt ) was obtained from equation: (K1 K2 )−1/2 .
was needed if the analytes possesses lower binding constants for the purpose of optimal selectivity [42]. The K values lower than 50 M−1 for ionic MA with all CDs were observed, consistent with previous report compared with other analytes. The dansyl amino acids possessed the highest apparent complex stability constants with these CDs and the migration time was thus shortest compared with other analytes racemates acid, which was attributed to its matched size of naphthalene ring with CD cavity. Accordingly, the theoretical optimal CD concentration for achieving maximum selectivity of Dns-Aba was lower than others. And the highest selectivity for 3-PLA was obtained with HEtAMCD compared with other CDs. Figure 5. X-reciprocal plot for MPrAMCD:Dns-Aba (lines 1 and 2) and MPrAMCD: 2-POPA (line 3 and 4).
Furthermore, the apparent complexation selectivity (␣cplx ) and theoretical optimum CD concentration were obtained (see Table 2). In general, higher CD concentration C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.3
NMR evaluation of chiral recognition
Chiral resolutions were variant that may occur due to different substituted group on CD rims. The manner of the intermolecular complexes responsible for a significant change in www.electrophoresis-journal.com
2750
Y. Wang et al.
Electrophoresis 2014, 35, 2744–2751
Figure 6. 1 H NMR spectra of MA and HPrAMCD in D2 O as well as the inclusion complexes of HPrAMCD/R-MA and HPrAMCD/S-MA (1:1, 10 mM, D2 O).
Figure 7. Portion of the 2D ROESY of mixture HPrAMCD/R-MA in DMSO-d6 .
the chiral recognition ability might be reflected by the NMR spectroscopy. Herein, each enantiomer of MA was selected as guest molecules and HPrAMCD played as host molecules. The 1 H NMR spectra for HPrAMCD was depicted in Fig. 6. All the aromatic protons of each MA enantiomers shifted to high field for all cases, indicating that inclusion complex occurred in D2 O solution. Moreover, Ho protons exhibited the largest chemical shift increment among of the three aromatic protons due to it might interact with H5 and H3 located inside the cavity of CD close to the primary rims. Meanwhile, low-field shifted chemical shifts were also observed for –CH ␣ of MA, mainly resulting from the electrostatic in C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
teractions formed between ammonium center of CD and carboxylic group of MA. Thus, it could be presumed that aromatic ring of MA molecule has inserted into cavity of CD form the narrow ring. The low-frequency shifted absorptions were also obtained for all protons (a–c) on alkylammonium moiety of CD. Moreover, the chemical shifts for CD/R-MA were larger than that of CD/S-MA might due to stronger interactions for R-enantiomer complex, which was verified by the later elution of R-MA in the CE experiment. 2D ROESY experiments were also carried out with the DMSO-d6 . The portion of ROESY obtained for HPrAMCD/RMA complex was shown in Fig. 7. It was obviously observed that the cross-peaks occurred between aromatic protons (Ho , 7.32–7.32 ppm; Hm , 7.24–71 ppm) and the H5 and H3 protons (3.60–3.40 ppm), confirming the existence of the inclusion complex. Additionally, the proton on chiral carbon afforded cross-peaks with the proton adjacent to ammonium center. Since the protons with cross-peaks observed were close in space, the results were consistent with the insert of phenyl ring into cavity in an axial orientation close to the narrow rim of CD (see Fig. 7).
4 Concluding remarks Direct chiral separation of six acidic enantiomers was performed in CE with a series of single-isomer ammonium-CD derivatives with different number or length of alkyl chain. In terms of binding constants, selectivity, and chiral resolution, these cationic CD derivatives demonstrated different recognition ability toward selected analytes. The more number of alkyl chains at N-atom afforded worse enantioseparation due to the impeded complex and apparent intrinsic chiral www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2744–2751
CE and CEC
2751
ability by steric hindrance. In most cases, greater difference in complex stability constants between enantiomers would lead to higher selectivity. Furthermore, 1D and 2D NMR revealed that the additional electrostatic interaction and hydrogen bonding would, to a certain extent, enhance the chiral recognition.
[19] Fanali, S., Electrophoresis 2009, 30, S203–S210.
We gratefully acknowledged the financial support from the National Natural Science Foundation of China (Grant No. 21305066), Program for New Century Excellent Talents in University (NCET-12-0633), Doctoral Fund of Ministry of Education of China (No. 20103219120008), the Jiangsu Province Natural Science Fund for Distinguished Young Scholars (BK20130032), and the Fundamental Research Funds for the Central Universities (30920130111006).
[23] Cucinotta, V., Contino, A., Giuffrida, A., Maccarrone, G., Messina, M., J. Chromatogr. A 2010, 1217, 953–967.
The authors have declared no conflict of interest.
5 References [1] Deeb, S. E., Watzig, H., EI-Hady, D. A., Albishri, H. M., de Griend, C. S., Scriba, G. K., Electrophoresis 2014, 35, 170–189. [2] Stalcup, A. M., Ann. Rev. Anal. Chem. 2010, 3, 341–363. [3] Ward, T. J., Ward, K. D., Anal. Chem. 2010, 82, 4712–4722. [4] Solinova, V., Kasicka, V., Sazelova, P., Holy, A., Electrophoresis 2009, 20, 2245–2254. [5] Castro-Puyana, M., Crego, A. L., Marina, M. L., Electrophoresis 2008, 29, 274–293.
´ [20] Juvancz, Z., Kendrovics, R. B., Ivanyi, R., Szente, L., Electrophoresis 2008, 29, 1701–1712. [21] Tsioupi, D. A., Stefan-Vanstaden, R. I., KapnissiChristodoulou, C. P., Electrophoresis 2013, 34, 178–204. [22] Tang, W., Ng, S. C., Sun, D., Modified Cyclodextrins for Chiral Separation, Springer, Berlin 2013.
[24] Dai, Y., Wang, S., Zhou, J., Tang, J., Tang, W., Electrophoresis 2013, 34, 833–840. [25] de Boer, T., Zeeuw, R. A., Jong, G. J. de., Ensing, K., Electrophoresis 2000, 21, 3220–3239. [26] Tutu, E., Vigh, G., Electrophoresis 2011, 32, 2655–2662. [27] Busby, M. B., Vigh, G., Electrophoresis 2005, 26, 3849–3860. [28] Busby, M. B., Vigh, G., Electrophoresis 2005, 26, 1978–1987. [29] Estrada, R., Vigh, G., J. Chromatogr. A 2012, 1226, 24–30. [30] Tacker, M., Glukhovskiy, P., Cai, H., Vigh, G., Electrophoresis 1999, 20, 2794–2798. [31] Maynard, D. K., Vigh, G., Electrophoresis 2001, 22, 3152–3162. [32] Boonleang, J., Stobaugh, J. F., Electrophoresis 2013, 34, 1232–1240. [33] Koval, D., Severa, L., Adriaenssens, L., Vavra, J., Teply, F., Kasicka, V., Electrophoresis 2011, 32, 2683–2692. [34] Chankvetadze, B., Electrophoresis 2009, 30, S211–S221.
[6] Mikus, P., Marakova, K., Electrophoresis 2009, 30, 2773–2802.
[35] Tang, W., Ng, S. C., Nat. Protoc. 2007, 2, 3195–3200.
[7] Gharbi, A., Humblot, V., Turpin, F., Pradier, C. M., Imbert, C., Berjeaud, J. M., FEMS Immunol. Med. Mic. 2012, 65, 257–269.
[37] Zhou, J., Ai, F., Zhou, B., Tang, J., Ng, S.-C., Tang, W., Anal. Chim. Acta. 2013, 800, 95–102.
[8] Garcia-Canas, V., Simo, C., Castro-Puyana, M., Cifuentes, A., Electrophoresis 2014, 35, 147–169. [9] Rocco, A., Aturki, Z., Fanali, S., Trends Anal. Chem. 2013, 52, 206–225. ´ Marina, M. L., J. Chro´ ´ [10] Perez-Fern andez, V., Garc´ıa, M. A., matogr. A 2011, 1218, 6561–6582.
[36] Tang, W., Ng, S. C., Nat. Protoc. 2008, 3, 691–697.
[38] Wang, S., Dai, Y., Wu, J., Zhou, J., Tang, J., Tang, W., J. Chromatogr. A 2013, 1277, 84–92. [39] Tang, W., Ong, T.-T., Muderawan, I. W., Ng, S.-C., Anal. Chim. Acta. 2007, 585, 227–233. ´ C., Garcia-Canas, V., Cucinotta, V., [40] Giuffrida, A., Leon, Cifuentes, A., Electrophoresis 2009, 30, 1734–1742.
[11] Herrero, M., Simo, C., Garcia-Canas, V., Fanali, S., Cifuentes, A., Electrophoresis 2010, 31, 2106–2114.
[41] Servais, A.-C., Rousseau, A., Dive, G., Frederich, M., Crommen, J., Fillet, M., J. Chromatogr. A 2012, 1232, 59–64.
[12] Astray, G., Gonzalez-Barreiro, C., Mejuto, J., Rial-Otero, ´ R., Simal-Gandara, J., Food Hydrocolloid. 2009, 23, 1631–1640.
[42] Ivanyi, R., Jicsinszky, L., Juvancz, Z., Roos, N., Otta, K., Szejtli, J., Electrophoresis 2004, 25, 2675–2686.
[13] Jac, P., Scriba, G. K., J. Sep. Sci. 2013, 36, 52–74.
[43] Dai, Y., Wang, S., Zhou, J., Liu, Y., Sun, D., Tang, J., Tang, W., J. Chromatogr. A 2012, 1246, 98–102.
[14] Opekar, F., Coufal, P., Stulik, K., Chem. Rev. 2009, 109, 4487–4499.
[44] Rizzi, A., Electrophoresis 2001, 22, 3079–3106.
[15] Preinerstorfer, B., Lammerhofer, M., Lindner, W., Electrophoresis 2009, 30, 100–132. [16] Scriba, G. K. E., Electrophoresis 2009, 30, S222–S228. [17] Chankvetadze, B., J. Chromatogr. A. 2007, 1168, 45–70. [18] Chankvetadze, B., Chem. Soc. Rev. 2004, 33, 337–347.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[45] Ann Van Eeckhaut, A. V., Michotte, Y., Electrophoresis 2006, 27, 2880–2895. [46] Wren, S. A. C., Rowe, R. C., J. Chromatogr. A 1992, 603, 235–241. [47] Rundlett, K. L., Armstrong, D. W., J. Chromatogr. A 1996, 721, 173–186.
www.electrophoresis-journal.com
Received April 14, 2014 Revised June 22, 2014 Accepted June 22, 2014
Electrophoresis 2014, 35, 2744–2751
Research Article
Evaluation of the chiral separation ability of single-isomer cationic -cyclodextrins in capillary electrophoresis Eight single-isomer ammonium--cyclodextrin derivatives with different side chains were successfully developed as chiral selectors for the chiral separation of selected racemates in capillary electrophoresis. The number of substituted groups at N-atom as well as the alkyl chain length greatly influenced the chiral separation. With the numbers of hydroxylalkyl groups at N-atom growing, the aqueous solubility of resolving agents were distinctly decreased and chiral separation ability was also significantly reduced. The apparent complex stability constants between CDs and analytes were further determined. The best enantioseparations of hydroxyl acids was achieved with the use of mono-6A -(3hydroxypropyl)-1-ammonium--cyclodextrin chloride and mono-6A -(3-methoxypropyl)-1ammonium--cyclodextrin chloride. The nuclear magnetic resonance experiments were carried out using them with mandelic acid as guest molecules, revealing the inclusion pattern as well as electrostatic interactions and hydrogen bonding interactions as additional chiral driving force. The contribution of potential interaction sites in the sidearm could enhance the enantioseparations. Keywords: Chiral separation / Hydrogen bonding / Single-isomer cationic cyclodextrin DOI 10.1002/elps.201400198
Additional supporting information may be found in the online version of this article at the publisher’s web-site
1 Introduction Chiral separation has gained extensive attention in the fields of pharmaceutical [1–4], clinical biological [5–7], environmental, and food analysis [8–12]. CE has proven to be a powerful analytical technique affording low consumption of samples
Correspondence: Dr. Weihua Tang, Key Laboratory of Soft Chemistry and Functional Materials (Ministry of Education of China), Nanjing University of Science and Technology, Nanjing 210094, P. R. China E-mail: [email protected] Fax: +86-25-84317311
Abbreviations: BHEtAMCD, mono-6A -bis(2-hydroxyethyl)-1ammonium--cyclodextrin chloride; BMEtAMCD, mono-6A bis(2-methoxyethyl)-1-ammonium--cyclodextrin chloride; HBuAMCD, mono-6A -(4-hydroxybutyl)-1-ammonium--cyclo dextrin chloride; HEtAMCD, mono-6A -(2-hydroxyethyl)-1ammonium--cyclodextrin chloride; HPrAMC, mono-6A -(3hydroxypropyl)-1-ammonium--cyclodextrin chloride; MEt AMCD, mono-6A -(2-methoxyethyl)-1-ammonium--cyclode xtrin chloride; MPrAMCD, mono-6A -(3-methoxypropyl)-1ammonium--cyclodextrin chloride; THEtAMCD, mono-6A tri(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
and solvents, high efficiency, short analysis time [13–18]. To date, CDs and their derivatives proved to be the most effective chiral selectors in CE separations attributable to their favorable enantioseparation properties and relatively low costs [19–21]. CD can be modified via mono- or multisubstitution of hydroxyl groups on the rims. However, the development of randomly multisubstituted CD derivatives is limited due to poor reproducibility in synthetic process and relative uncertainty in practical applications. Hence, the demand for effective chiral selectors has led to an outburst of new strategies toward single-isomer CD derivatives [22]. Among of these certainly functionalized CDs, ionizable CDs are more versatile than neutral ones since they are able to separate both charged and neutral enantiomers [23–25]. Systematic study of anionic sulfated single-isomer CDs was undertaken by Vigh’s group [26–31]. Negatively charged CD derivatives can interact strongly with basic compounds and high migration differences of enantiomers are also observed [32–34]. In terms of cationic CD derivatives, they have not been used as much as anionic ones probably due to their adsorption character to the column wall and relatively complicated
Colour Online: See the article online to view Figs. 1–7 in colour.
www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2744–2751
synthesis procedure. Despite that, the use of cationic CDs in CE can lower the migration time of analytes compared with anionic ones. The majority of positively charged CD derivatives are substituted in the C6 position by quaternary ammonium groups or weak electrolytes [35–40]. The chiral separation in CE using CDs as chiral selectors is fundamentally based on the formation of complexation (inclusion or external) and various additional interactions between analytes and functional moieties in substituted groups [19, 41]. Accordingly, the position and structure of only one substituent can greatly influence the enantioseparation ability of a CD derivative. In terms of single-isomer cationic ammonium--cyclodextrin derivatives, the length and number of alkyl chain would play a principle role in the separation process [42]. On the basis of our previous work [37, 38, 43], a series of versatile CD derivatives with (methoxy)alkylamino or (hydroxy)alkylamino substituents have been synthesized in this study. Mono-6A bis(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride, mono-6A -tri(2-hydroxyethyl)-1-ammonium--cyclodextrin chloride, and mono-6A -bis(2-methoxyethyl)-1-ammonium-cyclodextrin chloride were first reported as chiral selectors in CE. The eight CD derivatives were used for the separation of six selected analytes at different concentrations. Based on the separation results, the influence of substituted groups on the enantioseparation ability of these single-isomer CD derivatives was demonstrated. The apparent complex’s stability constants of CD and analytes were also investigated. In particular, the existence of hydrogen bonding during chiral separation was revealed by NMR measurements. This work presents a good correlation study between the alkylammonium substituted CDs and their enantioseparation ability, especially the alkyl chain length and number of substituents on the cationic center of the CD rim. The theoretical determination of binding constants and NMR study prove to be useful approaches in explanation of macroscopic chiral separation in CE experiments.
2 Materials and methods 2.1 Materials and chemicals -Cyclodextrin was obtained from Tokyo Chemical Industry (TCI, Japan). All the racemates were purchased from Sigma (St. Louis, MO, USA). The CD derivatives were synthesized in the lab according to our previous approach [36]. 6-Monotosyl-CD and appropriate amount of alkylamine were added into dry DMF and reacted for 5–12 h at 90°C. The acetone was added into the solution to precipitate the white solid. The crude product was obtained via further chloride ion exchange. All the CDs were further purified via recrystallization in DIwater before characterization and chiral separation. The structures of analytes and CD derivatives are depicted in Fig. 1 (synthesis of new CDs is provided in Supporting Information). C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2745
2.2 Apparatus and CE procedure NMR measurements were performed on a Bruker AVANCE 500 (500 MHz, Bruker Daltonics, Bremen, Germany). The 1D TOCSY NMR studies were conducted in D2 O for both CD/R-mandelic acid and CD/S-mandelic acid complexes. All CE separations were accomplished on P/ACE MDQ CE unit (Fullerton, CA, USA), equipped with an uncoated fused–silica capillary (50 m id, 363 m od, 49 cm effective length). All samples were injected at 3.45 kPa for 5 s and detected at 214 nm assisted with an online photodiode array (PDA) detector. During the separation process, the initial BGE was composed of 50 mM NaH2 PO4 solution with appropriate pH. The CD solutions were obtained via dissolving appropriate amount of CD derivatives into the BGEs. Stock analytes were set at 50 mg/mL with a mixture solvent of methanol/water (50/50, v/v) and stored in the refrigerator at +2°C. All buffers and sample solutions were undertaken pretreatment of filtration with a 0.45 m syringe-type Millipore membrane and ultrasonication prior to use.
2.3 Calculation of complex stability constants Direct chiral separation in CE is based on an electrophoretic technique as well as a chiral environment [43, 44]. During the separation, the chiral selectors could stereoselectively interact with enantiomers of analytes to form temporary diastereomeric association. Thus, the slight difference in the formation of reversible association will generate a mobility difference, resulting in a successful separation. For CE separation with charged CD derivatives, the mobility greatly depends on difference in binding constants for enantiomers with selectors. Assuming a 1:1 interaction between the enantiomer and CD, the correlation of the effective mobility with CD concentration can be described as the following equation [45]: eff =
f + c K [C] 1 + K [C]
(1)
Where eff is the effective mobility of enantiomers, f is the mobility of analytes in free form, c is the mobility of complex, [C] is the concentration of CD selector and K is the binding constants for enantiomers with CD. Equation (1) can be transformed into the same mathematical form as shown below [46]: eff − f = −K (eff − f ) + K (c − f ) [C]
(2)
According to Eq. (2), the apparent complex stability constants can be determined via the x-reciprocal method by (graphical) solution of plotting form of (eff − f )/[C] versus (eff − f ), which affords −K as slope. www.electrophoresis-journal.com
2746
Y. Wang et al.
Electrophoresis 2014, 35, 2744–2751 OH
OH C * COOH
OH CH2C * COOH
C * COOH
H
Cl
Mandelic acid (MA)
H
H
2-Chloromandelic acid (2-Cl MA)
3-Phenyllactic acid (3-PLA) O
H
H
C * COOH
S
O C * COOH
CH2CH3
O
H3C
CH3
H NH C* COOH CH2CH3
N H3C
2-Phenylbutyric acid (2-PBA)
OH
6
ClNH2
OH
(OH)14
2-Phenoxypropionic (2-POPA)
OH
6
Cl NH2
OH
OH
Cl 6
NH2
OH
Cl
OH
6
OH N H
OH
(OH)14
(OH)14
(OH)14 HPrAMCD
HEtAMCD
Dansyl-DL- -aminonbutyric acid (Dns-Aba)
BHEtAMCD
HBuAMCD
OH OH
Cl 6
(OH)14 THEtAMCD
N
OH
OH
Cl 6
NH2
OCH3
OH
Cl 6
NH2
OCH3
OH
Cl 6
OH
(OH)14 MEtAMCD
(OH)14
In this study, five single-isomer 6-mono-(hydroxyl)alkylammonium--CD (mono-6A -(2-hydroxyethyl)-1-ammonium-cyclodextrin chloride (HEtAMCD); mono-6A -(3-hydroxypropyl)-1-ammonium--cyclodextrin chloride (HPrAMCD); mono-6A -(4-hydroxybutyl)-1-ammonium--cyclodextrin chloride (HBuAMCD); BHEtAMCD and THEtAMCD) and three mono-6A -methoxyalkylammonium--CD, mono-6A -(2-methoxyethyl)-1-ammonium--cyclodextrin chloride (MEtAMCD), mono-6A -(3-methoxypropyl)-1 - ammonium -  - cyclodextrin chloride (MPrAMCD) and BMEtAMCD (see Fig. 1) were applied to study the influence of substituents on the enantioseparations. These amino--CD derivatives were obtained with good reproducibility in synthesis and CE separation. All CDs exhibited good aqueous solubility except THEtAMCD, which was only soluble up to 5 mM in the running buffer. Six enantiomeric pairs (2-POPA, 2-PBA, 3-PLA, MA, 2-Cl MA, Dns-Aba) were investigated and their physical and chemical properties are summarized in Table 1.
3.1 Enantioseparations with different CD at varied concentrations If the CD concentration is zero or very large, the apparent mobility difference will be zero, indicating that an appropriate CD concentration would afford an optimal C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
OCH3 OCH3
(OH)14 BMEtAMCD
MPrAMCD
3 Results and discussion
N H
Figure 1. Chemical structures of racemates and CD derivatives studied.
Table 1. Physical and chemical properties of the six acids studied
Racemates
pKaa)
Molar volume/cm3b)
Mandelic acid 2-Chloromandelic acid 3-Phenyllactic acid 2-Phenylbutyric acid 2-Phenoxypropionic acid Dansyl-DL-␣-aminonbutyric acid
3.41 ± 0.2 3.30 ± 0.1 3.72 ± 0.1 4.34 ± 0.1 3.22 ± 0.1 3.30 ± 0.1
115.1 ± 3.0 127.0 ± 3.0 131.2 ± 3.0 150.6 ± 3.0 141.4 ± 3.0 256.6 ± 3.0
a) All pKa values were determined by titrimetric experiments. b) All values were calculated by ACD/ChemSketch v6.0.
separation of enantiomers. In order to study the effect of concentration on the chiral separation, carboxylic acids (2-POPA, 2-PBA), hydroxy acids (3-PLA, MA, 2-Cl MA), and dansyl amino acid (Dns-Aba) were enantioseparated by gradually increasing CDs concentration from 1 to 20 mM with pH 6.0 BGEs. Most cases can be baseline separated except for BHEtAMCD, THEtAMCD, and BMEtAMCD (see Fig. 2). The selectivity of 2-POPA was significantly increased with higher CD concentration. HPrAMCD and BMEtAMCD could afford 1.5 of selectivity for 2-POPA at 20 mM. In the case of 2-POPA, the superior order for chiral resolution of CD with hydroxyalkyl groups was HPrAMCD ⬎ HBuAMCD ⬎ HEtAMCD. HPrAMCD and MPrAMCD with three-carbon alkyl chain exhibited the best chiral ability, achieving chiral resolution over 5 when the concentrations were higher than 5 mM. And the maximum chiral resolutions were obtained at www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2744–2751
CE and CEC
2747
Figure 2. Chiral resolution (Rs ) of analytes obtained with single-isomer cationic CDs at varied concentrations.
approximate 20 mM for HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD. Interestingly, the chiral resolution even increased linearly with increased concentration for HEtAMCD from 5 to 20 mM. In terms of 2-PBA, a maximal resolution was obtained at 5 mM CD (except THEtAMCD and BMEtAMCD), with chiral resolution over 1.05. It was obvious that the number of substituted groups greatly influenced the chiral resolution. BHEtAMCD and BMEtAMCD with two alkyl chains could not enantioseparate 2-PBA at low concentration, meanwhile, THEtAMCD possessing three alkyl chains on N-atom could not even resolve 2-PBA at all studied concentrations. This indicated that steric hindrance may exist in the formation of temporary diastereomeric association. Taking the alkyl chain length into account, HEtAMCD was the best choice compared with HPrAMCD and HBuAMCD for 2-PBA. Meanwhile, MEtAMCD could afford better selectivity than MPrAMCD at most concentrations. Hydroxyl acids showed increased chiral resolution at higher concentration of chiral selectors for most CDs studied. Among of them, HPrAMCD could resolve them with highest selectivity. When the concentration of HPrAMCD reached
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
20 mM, all the selectivities were over 1.5. The representative electropherograms of 2-POPA and MA using different concentrations of HPrAMCD were depicted in Fig. 3. It revealed that both selectivity and chiral resolutions increased with higher CD concentrations, though the shortest elution time was obtained at 3 mM. Additionally, analytes have more affinity to different CD selectors. For example, 3-PLA and MA were better resolved by HBuAMCD than MPrAMCD, while MPrAMCD could afford better enantioselectivity toward 2-Cl MA. HPrAMCD could resolve MA and 2-Cl MA with a resolution as high as over 18 at 20 mM, meanwhile, MPrAMCD can afford them a resolution over 10. It was also found that CD BMEtAMCD exhibited more excellent chiral ability than BHEtAMCD toward three hydroxyl acids, indicating that the introduction of methoxy group on alkyl chain sometimes could enhance the chiral separation compared to hydroxyl group. In our study, one dansyl amino acid (DnsAba) was also applied in CE separation. All CDs could afford higher selectivity with increased selectivity during concentration range applied, especially that of HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD. HPrAMCD and HBuAMCD exhibited similar selectivity, reaching over 2.0 at 20 mM.
www.electrophoresis-journal.com
2748
Y. Wang et al.
Figure 3. Representative chromatograms of (A) 2-POPA and (B) MA using HPrAMCD at different concentrations.
However, the chiral resolution obtained by HEtAMCD and HPrAMCD was much higher than that by HBuAMCD. Meanwhile, BHEtAMCD and BMEtAMCD only partially resolve it and THEtAMCD could not achieve baseline separation. The electropherograms at lower CD concentrations (ࣘ3 mM) exhibit certain drift in baseline, which may be due to the absorption of cationic CDs onto capillary inner wall to cause the disturbance to the system. Based on the enantioseparations data, the number and length of alkyl chain at ammonium center was of great importance for chiral resolution. The comparable resolutions were shown in Fig. 4. In this case of separation with 10 mM of HEtAMCD, HPrAMCD, and HBuAMCD, the superior order of single alkyl chain with hydroxyl group was HPrAMCD ⬎ HBuAMCD ⬎ HEtAMCD (see Fig. 4A). 2-Cl MA approximately achieved two folds chiral resolution with HPrAMCD over HEtAMCD, indicating a proper alkyl chain length would influence the enantioseparations. Compared the chiral resolutions obtained by HEtAMCD, BHEtAMCD, and THEtAMCD, HEtAMCD with only single hydroxyamino group proved to be the superior choice for all analytes (see Fig. 4B). HEtAMCD could afford a resolution of 5.0 toward MA, while BHEtAMCD and THEtAMCD could not resolve it. With the number of substituted groups at N-atom increased, the water solubility and separation ability dramatically declined. THEtAMCD with three alkyl chains on N-atom was the worst C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Electrophoresis 2014, 35, 2744–2751
Figure 4. Comparison of chiral resolutions for six acids with (A) HEtAMCD, HPrAMCD, and HBuAMCD at 10 mM concentration; and (B) HEtAMCD, BHEtAMCD, and THEtAMCD at 5 mM concentration.
chiral selector in this comparison study. This might be explained by the impeded complex stability constants through steric hindrance.
3.2 Calculation of binding constants, apparent complexation selectivity, theoretical optimum selector concentration The apparent complex stability constants (K) were calculated from the measured mobility values in the applied concentration range according to the x-reciprocal method [47] (see Table 2). Since cationic CDs used in CE can cause change in solution viscosity and ionic strength in the BGE, all mobilities for the modeling here were corrected by using the migration times of analytes referenced to zero CD concentration using neutral marker (MeOH). For the plots of (eff − f )/[C] against (eff − f ), the correlation coefficient (R) of the fitting curves could reach at least 0.974, 0.950, 0.969, 0.995, 0.998, and 0.991 for 2-POPA, 2-PBA, 3-PLA, MA, 2-Cl MA, and Dns-Aba with HEtAMCD, HPrAMCD, HBuAMCD, and MPrAMCD, respectively. Representative example of the fitting curve for Dns-Aba and 2-POPA using MPrAMCD was shown in Fig. 5. www.electrophoresis-journal.com
CE and CEC
Electrophoresis 2014, 35, 2744–2751
2749
Table 2. Apparent complex stability constants (K1 and K2 ), apparent complexation selectivity (␣cplx ), maximum resolution (Rs max ), and calculated optimal concentration (copt ) of CD studied
Analytes
3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA ␣cplx 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD
HPrAMCD
HBuAMCD
MPrAMCD
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
K1 (M−1 )
K2 (M−1 )
87 36 31 210 58 119
121 47 44 229 71 132
124 35 73 186 67 121
131 40 84 191 87 129
87 22 52 178 32 160
113 34 56 181 45 171
89 45 53 139 78 134
108 53 63 145 93 140
HEtAMCD 1.39 1.31 1.42 1.09 1.22 1.11
HPrAMCD
HBuAMCD
1.06 1.14 1.15 1.03 1.30 1.07
1.30 1.55 1.08 1.02 1.41 1.07
MPrAMCD 1.21 1.18 1.19 1.04 1.19 1.04
Rs max 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD 6.05 (10) 8.16 (20) 10.28 (20) 2.96 (20) 6.63 (20) 1.67 (5)
HPrAMCD 8.45 (10) 18.57 (20) 19.88 (20) 3.90 (20) 9.21 (20) 1.56 (5)
HBuAMCD 6.22 (20) 10.05 (20) 10.05 (20) 3.16 (15) 7.45 (20) 1.63 (5)
MPrAMCD 9.95 (15) 11.19 (20) 14.36 (20) 3.83 (20) 8.70 (20) 1.48 (5)
Copt 3-PLA MA 2-Cl MA Dns-Aba 2-POPA 2-PBA
HEtAMCD 10 24 27 5 16 8
HPrAMCD 8 27 13 5 13 8
HBuAMCD 10 37 19 6 26 6
MPrAMCD 10 20 17 7 13 7
Complex stability constants (K1 and K2 ) calculated from 4–7 different concentrations between 0–20 mM by linear regression. Each data point was the mean of at least three injections. Rs max represented the maximum chiral resolution obtained in the experiments. The apparent complexation selectivity (␣cplx ) was calculated according to equation: K1 /K2 ; Theoretical optimum selector concentration (copt ) was obtained from equation: (K1 K2 )−1/2 .
was needed if the analytes possesses lower binding constants for the purpose of optimal selectivity [42]. The K values lower than 50 M−1 for ionic MA with all CDs were observed, consistent with previous report compared with other analytes. The dansyl amino acids possessed the highest apparent complex stability constants with these CDs and the migration time was thus shortest compared with other analytes racemates acid, which was attributed to its matched size of naphthalene ring with CD cavity. Accordingly, the theoretical optimal CD concentration for achieving maximum selectivity of Dns-Aba was lower than others. And the highest selectivity for 3-PLA was obtained with HEtAMCD compared with other CDs. Figure 5. X-reciprocal plot for MPrAMCD:Dns-Aba (lines 1 and 2) and MPrAMCD: 2-POPA (line 3 and 4).
Furthermore, the apparent complexation selectivity (␣cplx ) and theoretical optimum CD concentration were obtained (see Table 2). In general, higher CD concentration C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
3.3
NMR evaluation of chiral recognition
Chiral resolutions were variant that may occur due to different substituted group on CD rims. The manner of the intermolecular complexes responsible for a significant change in www.electrophoresis-journal.com
2750
Y. Wang et al.
Electrophoresis 2014, 35, 2744–2751
Figure 6. 1 H NMR spectra of MA and HPrAMCD in D2 O as well as the inclusion complexes of HPrAMCD/R-MA and HPrAMCD/S-MA (1:1, 10 mM, D2 O).
Figure 7. Portion of the 2D ROESY of mixture HPrAMCD/R-MA in DMSO-d6 .
the chiral recognition ability might be reflected by the NMR spectroscopy. Herein, each enantiomer of MA was selected as guest molecules and HPrAMCD played as host molecules. The 1 H NMR spectra for HPrAMCD was depicted in Fig. 6. All the aromatic protons of each MA enantiomers shifted to high field for all cases, indicating that inclusion complex occurred in D2 O solution. Moreover, Ho protons exhibited the largest chemical shift increment among of the three aromatic protons due to it might interact with H5 and H3 located inside the cavity of CD close to the primary rims. Meanwhile, low-field shifted chemical shifts were also observed for –CH ␣ of MA, mainly resulting from the electrostatic in C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
teractions formed between ammonium center of CD and carboxylic group of MA. Thus, it could be presumed that aromatic ring of MA molecule has inserted into cavity of CD form the narrow ring. The low-frequency shifted absorptions were also obtained for all protons (a–c) on alkylammonium moiety of CD. Moreover, the chemical shifts for CD/R-MA were larger than that of CD/S-MA might due to stronger interactions for R-enantiomer complex, which was verified by the later elution of R-MA in the CE experiment. 2D ROESY experiments were also carried out with the DMSO-d6 . The portion of ROESY obtained for HPrAMCD/RMA complex was shown in Fig. 7. It was obviously observed that the cross-peaks occurred between aromatic protons (Ho , 7.32–7.32 ppm; Hm , 7.24–71 ppm) and the H5 and H3 protons (3.60–3.40 ppm), confirming the existence of the inclusion complex. Additionally, the proton on chiral carbon afforded cross-peaks with the proton adjacent to ammonium center. Since the protons with cross-peaks observed were close in space, the results were consistent with the insert of phenyl ring into cavity in an axial orientation close to the narrow rim of CD (see Fig. 7).
4 Concluding remarks Direct chiral separation of six acidic enantiomers was performed in CE with a series of single-isomer ammonium-CD derivatives with different number or length of alkyl chain. In terms of binding constants, selectivity, and chiral resolution, these cationic CD derivatives demonstrated different recognition ability toward selected analytes. The more number of alkyl chains at N-atom afforded worse enantioseparation due to the impeded complex and apparent intrinsic chiral www.electrophoresis-journal.com
Electrophoresis 2014, 35, 2744–2751
CE and CEC
2751
ability by steric hindrance. In most cases, greater difference in complex stability constants between enantiomers would lead to higher selectivity. Furthermore, 1D and 2D NMR revealed that the additional electrostatic interaction and hydrogen bonding would, to a certain extent, enhance the chiral recognition.
[19] Fanali, S., Electrophoresis 2009, 30, S203–S210.
We gratefully acknowledged the financial support from the National Natural Science Foundation of China (Grant No. 21305066), Program for New Century Excellent Talents in University (NCET-12-0633), Doctoral Fund of Ministry of Education of China (No. 20103219120008), the Jiangsu Province Natural Science Fund for Distinguished Young Scholars (BK20130032), and the Fundamental Research Funds for the Central Universities (30920130111006).
[23] Cucinotta, V., Contino, A., Giuffrida, A., Maccarrone, G., Messina, M., J. Chromatogr. A 2010, 1217, 953–967.
The authors have declared no conflict of interest.
5 References [1] Deeb, S. E., Watzig, H., EI-Hady, D. A., Albishri, H. M., de Griend, C. S., Scriba, G. K., Electrophoresis 2014, 35, 170–189. [2] Stalcup, A. M., Ann. Rev. Anal. Chem. 2010, 3, 341–363. [3] Ward, T. J., Ward, K. D., Anal. Chem. 2010, 82, 4712–4722. [4] Solinova, V., Kasicka, V., Sazelova, P., Holy, A., Electrophoresis 2009, 20, 2245–2254. [5] Castro-Puyana, M., Crego, A. L., Marina, M. L., Electrophoresis 2008, 29, 274–293.
´ [20] Juvancz, Z., Kendrovics, R. B., Ivanyi, R., Szente, L., Electrophoresis 2008, 29, 1701–1712. [21] Tsioupi, D. A., Stefan-Vanstaden, R. I., KapnissiChristodoulou, C. P., Electrophoresis 2013, 34, 178–204. [22] Tang, W., Ng, S. C., Sun, D., Modified Cyclodextrins for Chiral Separation, Springer, Berlin 2013.
[24] Dai, Y., Wang, S., Zhou, J., Tang, J., Tang, W., Electrophoresis 2013, 34, 833–840. [25] de Boer, T., Zeeuw, R. A., Jong, G. J. de., Ensing, K., Electrophoresis 2000, 21, 3220–3239. [26] Tutu, E., Vigh, G., Electrophoresis 2011, 32, 2655–2662. [27] Busby, M. B., Vigh, G., Electrophoresis 2005, 26, 3849–3860. [28] Busby, M. B., Vigh, G., Electrophoresis 2005, 26, 1978–1987. [29] Estrada, R., Vigh, G., J. Chromatogr. A 2012, 1226, 24–30. [30] Tacker, M., Glukhovskiy, P., Cai, H., Vigh, G., Electrophoresis 1999, 20, 2794–2798. [31] Maynard, D. K., Vigh, G., Electrophoresis 2001, 22, 3152–3162. [32] Boonleang, J., Stobaugh, J. F., Electrophoresis 2013, 34, 1232–1240. [33] Koval, D., Severa, L., Adriaenssens, L., Vavra, J., Teply, F., Kasicka, V., Electrophoresis 2011, 32, 2683–2692. [34] Chankvetadze, B., Electrophoresis 2009, 30, S211–S221.
[6] Mikus, P., Marakova, K., Electrophoresis 2009, 30, 2773–2802.
[35] Tang, W., Ng, S. C., Nat. Protoc. 2007, 2, 3195–3200.
[7] Gharbi, A., Humblot, V., Turpin, F., Pradier, C. M., Imbert, C., Berjeaud, J. M., FEMS Immunol. Med. Mic. 2012, 65, 257–269.
[37] Zhou, J., Ai, F., Zhou, B., Tang, J., Ng, S.-C., Tang, W., Anal. Chim. Acta. 2013, 800, 95–102.
[8] Garcia-Canas, V., Simo, C., Castro-Puyana, M., Cifuentes, A., Electrophoresis 2014, 35, 147–169. [9] Rocco, A., Aturki, Z., Fanali, S., Trends Anal. Chem. 2013, 52, 206–225. ´ Marina, M. L., J. Chro´ ´ [10] Perez-Fern andez, V., Garc´ıa, M. A., matogr. A 2011, 1218, 6561–6582.
[36] Tang, W., Ng, S. C., Nat. Protoc. 2008, 3, 691–697.
[38] Wang, S., Dai, Y., Wu, J., Zhou, J., Tang, J., Tang, W., J. Chromatogr. A 2013, 1277, 84–92. [39] Tang, W., Ong, T.-T., Muderawan, I. W., Ng, S.-C., Anal. Chim. Acta. 2007, 585, 227–233. ´ C., Garcia-Canas, V., Cucinotta, V., [40] Giuffrida, A., Leon, Cifuentes, A., Electrophoresis 2009, 30, 1734–1742.
[11] Herrero, M., Simo, C., Garcia-Canas, V., Fanali, S., Cifuentes, A., Electrophoresis 2010, 31, 2106–2114.
[41] Servais, A.-C., Rousseau, A., Dive, G., Frederich, M., Crommen, J., Fillet, M., J. Chromatogr. A 2012, 1232, 59–64.
[12] Astray, G., Gonzalez-Barreiro, C., Mejuto, J., Rial-Otero, ´ R., Simal-Gandara, J., Food Hydrocolloid. 2009, 23, 1631–1640.
[42] Ivanyi, R., Jicsinszky, L., Juvancz, Z., Roos, N., Otta, K., Szejtli, J., Electrophoresis 2004, 25, 2675–2686.
[13] Jac, P., Scriba, G. K., J. Sep. Sci. 2013, 36, 52–74.
[43] Dai, Y., Wang, S., Zhou, J., Liu, Y., Sun, D., Tang, J., Tang, W., J. Chromatogr. A 2012, 1246, 98–102.
[14] Opekar, F., Coufal, P., Stulik, K., Chem. Rev. 2009, 109, 4487–4499.
[44] Rizzi, A., Electrophoresis 2001, 22, 3079–3106.
[15] Preinerstorfer, B., Lammerhofer, M., Lindner, W., Electrophoresis 2009, 30, 100–132. [16] Scriba, G. K. E., Electrophoresis 2009, 30, S222–S228. [17] Chankvetadze, B., J. Chromatogr. A. 2007, 1168, 45–70. [18] Chankvetadze, B., Chem. Soc. Rev. 2004, 33, 337–347.
C 2014 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[45] Ann Van Eeckhaut, A. V., Michotte, Y., Electrophoresis 2006, 27, 2880–2895. [46] Wren, S. A. C., Rowe, R. C., J. Chromatogr. A 1992, 603, 235–241. [47] Rundlett, K. L., Armstrong, D. W., J. Chromatogr. A 1996, 721, 173–186.
www.electrophoresis-journal.com
