Journal of Chromatography A, 1363 (2014) 119–127
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Evaluation of perphenylcarbamated cyclodextrin clicked chiral stationary phase for enantioseparations in reversed phase high performance liquid chromatography Limin Pang a , Jie Zhou a , Jian Tang a,∗∗ , Siu-Choon Ng b , Weihua Tang a,∗ a Key Laboratory of Soft Chemistry and Functional Materials, Ministry of Education, Nanjing University of Science and Technology, Nanjing 210094, People’s Republic of China b Division of Chemical and Biomolecular Engineering, College of Engineering, Nanyang Technological University, 16 Nanyang Drive, Singapore 637722, Singapore
a r t i c l e
i n f o
Article history: Received 30 April 2014 Received in revised form 26 June 2014 Accepted 11 August 2014 Available online 17 August 2014 Keywords: Chiral separation Cyclodextrin Chiral stationary phase Reversed phase-HPLC
a b s t r a c t In this study, perphenylcarbamated cyclodextrin clicked chiral stationary phase (CSP) was developed with high column efficiency. The characteristics of the column were evaluated in terms of linearity, limit of detection and limit of quantification. The enantioselectivity of the as-prepared clicked CSP was explored with 26 recemates including aryl alcohols, flavanoids and adrenergic drugs in reversed phase highperformance liquid chromatography. The effect of separation parameters including flow rate, column temperature, organic modifier and buffer on the enantioselectivity of the clicked CSP was investigated in detail. The correlation study of the analytes structure and their chiral resolution revealed the great influence of analytes’ structure on the enantioseparations with cyclodextrin CSP. Methanol with 1% of triethylammonium acetate buffer (pH 4) was proved to be the best choice for the chiral separation of basic enantiomers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Chiral compounds may significantly differ from each other in their biological, pharmacological and toxicological effects. A notorious example would be thalidomide which was debuted as a racemic drug [1]. Chiral separation has thus gained wide attention from both academia and industry. To date, chromatographic techniques have evolved as routine approaches for chiral separation, which feature high efficiency, quick analysis and precise resolutions. The most widely used techniques include high performance liquid chromatography (HPLC), gas chromatography, capillary electrophoresis, supercritical fluid chromatography and stimulated moving bed. By affording direct enantioseparation, HPLC coupled with chiral stationary phases (CSPs) has developed as one of the most important techniques for both detection of enantiomeric purity and quick preparation of pure enantiomers [2,3]. The CSPs include Pirkle-type,
∗ Corresponding author. Tel.: +86 25 8431 7311; fax: +86 25 8431 7311. ∗∗ Corresponding author. Tel.: +86 25 8431 7311. E-mail addresses: [email protected] (J. Tang), [email protected] (W. Tang). http://dx.doi.org/10.1016/j.chroma.2014.08.040 0021-9673/© 2014 Elsevier B.V. All rights reserved.
macrocyclic antibiotic, crown ethers, imprinted polymers and chiral ligand exchange CSPs as well as polysaccharide-, protein- and cyclodextrin (CD)-based CSPs [4–13]. The CD based CSPs have thus been extensively explored for the enantioseparation by functionalizing the CD rims to construct additional interactions such as – stacking, dipole–dipole, ion-paring, electrostatic and steric repulsive effects between analytes and the resulted CDs. Moreover, CD based CSPs are especially attractive for their versatility and durability under various conditions [8–13]. The CD-CSPs can be used in three modes including reversed phase (RP), normal phase (NP) and polar organic modes [13]. RP is the most popular mode in which the solute molecules are distributed between the relatively polar mobile phase and the non-polar stationary phase to afford separation. Under RP conditions, the enantioselectivity of CDs is dependent upon the analyte structure, CD’s type and the functionalities on the CD rims. The extent of host–guest inclusion generally depends on CD’s cavity dimension and the structure of the enantiomers. According to the concept of size-fit for inclusion complexation, better enantioselectivity for the CSP-enantiomers pairs generally occur when the size of hydrophobic portions of analytes matches with the CD cavity [14]. “Click chemistry” was first proposed by Sharpless and the [3 + 2] dipolar cycloaddition between azides/alkynes catalyzed by copper
120
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
OH
OH
F
Br
Cl
OH
NH2
OH
OH
F Aryl-OH-5
OH
O
O
OH
Aryl-OH-11
Aryl-OH-10
Aryl-OH-8
Aryl-OH-7
Aryl-OH-6
OH Aryl-OH-9
OH
NH2
Cl Aryl-OH-4
Aryl-OH-3
Aryl-OH-2
Aryl-OH-1
OH
OH
4-Chlorobenzoin
Benzoin
Aryl-OH-12
OH
Cl
OH O
O
O
O
O
O
O 7-Hydroxylflavanone
O
HO
O O
O
O
Flavanone
6-Methoxyflavanone
7-Methoxyflavanone
OH
OCH3 HO
O
OH
HO
O O
Acebutolol
OH
N H
H2N O
O OH
OH O Naringenin
Hesperetin
N H
O
O
OH O
O
4-Hydroxylflavanone
OH H N
Atenolol
Propranolol
OH H N
Cl H2N
OH H N
HO
Clenbuterol
O
O O Si O
N H
N N N
OR 6
silica gel
HO
Cl
N H
R=
Isoprenaline
H N
(OR)14 O
CD clicked CSP
Fig. 1. Structures of racemate studied.
can be regarded as the primary one [15]. By affording high selectivity towards reagents and excellent flexibility in various solvents, click reaction has dramatically burgeoning in the application of immobilization of ligands to solid and polymer supports [16,17]. As a continuation of our long-term research project, our recent studies have demonstrated that the CD based CSPs prepared via click chemistry can afford good enantioselectivities and durability in broad separation conditions in HPLC [5,7,9,13,18]. In order to explore the potential of triazolyl-linked CD CSPs for enantioseparation in HPLC, the perphenylcarbamated CD clicked CSP (structure in Fig. 1) was prepared to pack chiral column with greatly improved column efficiency and CD surface loading. The column efficiency greatly enhanced CSP afforded significant improvement in enantioselectivities to larger library of racemates in comparison to our first report [18]. We herein report the detailed characteristics of column. The enantioseparation capability of the clicked CD CSP was further evaluated with 26 racemates including aryl alcohols, aryl amines, flavanoids, and adrenergic drugs. The effect of the separation conditions including flow rate, column temperature, mobile phase composition, and organic modifiers on the enantioseparation of the model analytes were thoroughly investigated in RPLC. The correlation of analytes’ structures with their chiral resolutions by the clicked CD CSP was also described.
2. Experiments 2.1. Chemicals and materials All reagents such as -CD and phenyl isocyanate were purchased from Tokyo Chemical Industry (TCI, Japan). HPLC-grade acetonitrile (ACN) and methanol (MeOH) were purchased from Tedia (USA). HPLC-grade acetic acid, phosphoric acid and triethylamine (TEA) were obtained from J&K (Shanghai, China). Deionized water for the experiments was purified by Milli-Q system (Millipore, Bedford, MA, USA). All racemic analytes were purchased from the Meryer (Shanghai, China) (structures in Fig. 1). Kromasil ˚ was purchased from Eka Chemispherical silica gel (5 m, 100 A) cals (Bohus, Sweden). All other chemicals used were of analytical reagent grade without purification prior to use. 2.2. Apparatus All enantioseparations with CD clicked CSP were performed on an Agilent HPLC system, which was comprised of an Agilent 1260 system consisting of with G1315D diode array detection (DAD) system, G1329B quaternary pump, a G1331C automatic injector, a G1316A temperature controller and Agilent Chem Station data manager software (Agilent Technologies, Palo Alto, CA, USA).
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
121
Fig. 2. The standard curves of the first enantiomer (a) and the second enantiomer (b) of flavanone (c) the chiral resolutions of flavanone at different sample concentrations. Conditions: methanol as mobile phase, 250 nm detection wavelength.
2.3. HPLC procedures The mobile phases were prepared by mixing appropriate amounts of acetonitrile or methanol into water to separate neutral compounds, while for basic analytes, the mobile phases were prepared by adding 1% (v/v) triethylammonium acetate buffer (1% TEAA, pH adjusted with acetic acid) or 1% (v/v) triethylammonium phosphate buffer (1% TEAP, pH adjusted with phosphoric acid) in water. All mobile phases were filtered with 0.22 m Millipore membrane and ultrasonicated 30 min for degassing before use. The detection wavelength range was 200–300 nm. The racemic samples were dissolved in methanol to prepare stock solutions of 1–5 mg/mL. If needed, stock solutions were further diluted with methanol to prepare sample solutions of 100–500 g/mL. All sample solutions were stored at 4 ◦ C prior to use. All chromatograms were obtained from 10 L injection of sample solution with 1.0 mL/min flow rate of mobile phase under 25 ◦ C, unless otherwise specified. The following resolution parameters were determined: k1 and k2 were calculated using (t1 − t0 )/t0 and (t2 − t0 )/t0 , respectively, where t0 is the retention time at which the first baseline disturbance by the solvent peak occurred, t1 and t2 are the retention time of the first and the second enantiomer. The separation factor (˛) was calculated using k2 /k1 , while the chiral resolution (Rs ) was evaluated using the equation Rs = 1.18(t2 − t1 )/(w1 + w2 ), where w1 and w2 were the half-peak width of each enantiomer (based on USP standards). 2.4. Preparation of CD-CSP The perphenylcarbamated CD clicked CSP was prepared according to our recently reported procedure [18]. Starting from mono-6A -azido--CD, mono-6A -azido-perphenylcarbamated CD was obtained via a nucleophilic addition with phenyl isocyanate and a followed immobilization onto alkynyl functionalized silica gel via “click” chemistry. The elemental analysis results for the as-prepared CSP were 13.69% C, 0.546% N and 2.1% H. The immobilization efficiency was much improved than that in our first report (10.2% C, 1.90% N and 1.71% H) [18]. The surface loading of the CSP was calculated to be 0.021 mol/m2 with [C%/(12 × Nc × Ssilica )] × 1000, where C% is derived from elemental analysis of carbon, Nc is the carbon atom number per CD molecule (182) and Ssilica is the surface area of silica gel (300 m2 /g). The CD clicked CSP was then packed into columns by conventional slurry method, where the CSP suspension in methanol was packed into stainless steel columns (4.6 × 250 mm) under constant pressure (9000 psi) using packing system (Lab Alliance Scientific). The packing pressure was maintained for at least 0.5 h. Detected with toluene as marker sample and methanol as mobile phase under 25 ◦ C, the as-packed column afforded high column efficiency
of 12,000, which was 50% higher than our previous report [18]. The minimum plate height was determined as 20.9 m, corresponding to a theoretical plate number of 48,000 plates/meter. 3. Results and discussions 3.1. Characteristics of the column The linearity, limit of detection (LODs), limit of quantification (LOQs), relative standard deviation (RSD) and sample capacity of this column were evaluated with flavanone. The sample concentration increased from 1 g/mL to 800 g/mL at the fixed flow rate of 1.0 mL/min for methanol mobile phase. The standard curve by plotting sample peak area against sample concentration was obtained based upon the least square linear regression fit in the concentration range of 100–600 g/mL. As shown in Fig. 2a and b, impressive results were obtained with linearity coefficients R2 > 0.999 for both enantiomers of flavanone. Meanwhile, increased concentration could lead to a reduced chiral resolution probably due to the overloading of the CSP (Fig. 2c). LOD and LOQ for both enantiomers of flavanone were determined at an S/N (signal/noise ration) of 3 and 10, respectively. LOD was thus determined to be 4.12 × 10−12 mol and 4.46 × 10−12 mol, while LOQ to be 1.35 × 10−11 mol and 1.48 × 10−11 mol for the first and second enantiomer, respectively. Besides, the sample capacity (the sample amount corresponding to 10% decreased column efficiency) was 5.36 × 10−8 mol. Precision of the method was carried out with repetitive runs in RPLC via 10 injections of solutions (100 and 200 g/mL). Linearity parameters are summarized in Table 1, RSD values (%) obtained for both enantiomers were lower than 0.2% for the migration times and lower than 0.3% for the peak areas of flavanone enantiomers, revealing the excellent repeatability and reproducibility for the chiral column. 3.2. Effect of flow rate on enantioseparations In HPLC enantiomeric separations, flow rate is of great importance for both retention time and chiral resolutions. The flow rate was optimized by gradually increasing from 0.2 mL/min to 1.0 mL/min for methanol mobile phase (Fig. 3). The retention time and chiral resolutions of flavanone were found to decrease dramatically with the increment of methanol flow rate while the selectivity maintained unchanged. As revealed by Berthod et al. [19] that the formation of inclusion complexes was slower than the complex formed through the external interaction, the constant selectivity factor at varied flow rates indicated that inclusion complex did not play the decisive role for the chiral separation at this conditions. Taking selectivity, chiral resolution and retention time into
122
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Table 1 Characteristics of the column on the test of flavanone. Items
First enantiomer
Concentration (g/mL) Peak area, RSD% (n = 10) Migration time, RSD% (n = 10) Linear range (g/mL, injection 20 L) Linear equation Standard error Linear coefficient (R2 ) LOD (mol) LOQ (mol) Sample capacity (mol)
100 0.16 0.09 100–600 y = 0.01845x − 0.06053 Sa = 0.06583, Sb = 0.00017 0.99983 4.12 × 10−12 1.35 × 10−11 5.36 × 10−8
Second enantiomer 200 0.16 0.12
ln ˛ = −
100 0.21 0.14 100–600 y = 0.01838x − 0.06740 Sa = 0.08376, Sb = 0.00022 0.99958 4.46 × 10−12 1.48 × 10−11
H ◦ S ◦ + R·T R
200 0.19 0.14
(2)
where R is the gas constant, k is the retention factor, ϕ is the column phase ratio, H◦ and S◦ represent the respective changes of enthalpy and entropy when one enantiomer transfers from mobile phase to the CSP. Since the value of ϕ is often not known, the S◦ * values [[S◦ * = S◦ + Rln ϕ]] calculated from the intercepts of the plots via equation (1) are generally used. H◦ and S◦ represent the differences of S◦ and S◦ * for both enantiomers, respectively. The plot of ln ˛ against 1/T is thus a straight line if enantioselective interactions do not change over the temperature range studied [23]. According to Fig. 5a and b, a good linear relationship between ln ˛ and 1/T was observed with correlation coefficients over 0.99, which shows both the conformation of CD CSP and the enantioselective interactions attributed for the enantioseparation were unaffected at varied column temperature. As the values of H◦ and S◦ were always negative, the chiral separation was indicated as enthalpy driven process [23–25]. 3.4. Evaluation of enantioseparations ability of CSP
Fig. 3. The effect of flow rate on the enantioseparation of flavanone in methanol mobile phase detected at 250 nm.
consideration, the optimal flow rate of 1.0 mL/min was selected for the following study. 3.3. Effect of column temperature on enantioseparations The column temperature is also important consideration for successful enantioseparations. Three aspects are thought to be influenced by column temperature: (i) mass transfer in mobile phase, (ii) the interactions between CSP and the enantiomer, and (iii) the transfer of an analyte between CSP and mobile phase [20,21]. The influence of temperature on the enantiosepration of CD clicked CSP was studied with flavanone and 7-methoxyflavanone as model analytes in column temperature range of 10–50 ◦ C with methanol as mobile phase. As the enantioseparation data shown in Table 2, k1 , k2 , ˛ and Rs all decreased when the column temperature increased (Fig. 4). This can be attributed to not only the increasing mass transfer of solutes in the mobile phase, but also the interactions between the enantiomers and the mobile phases, which were affected by the thermodynamic properties of the analytes on CSPs [22]. To investigate the thermodynamic functions of enantioselective adsorption, Van’t Hoff plots were interpreted in terms of the mechanistic aspects of chiral recognition: ln k = −
H ◦ H ◦ S ◦ S ◦ ∗ + + ln ϕ = − + RT R RT R
(1)
3.4.1. Enantioseparations of racemic aryl alcohols and flavanoids Several racemic aryl alcohols and flavanoids have been tried on the CD clicked CSP in our previous report [18]. In order to explore the potential of enantioselectivities of the perphenylcarbamated CD clicked CSP, the resolutions of 19 racemates including 10 aryl alcohols, 2 aryl amines and 7 flavanoids were optimized at different separation conditions. The optimal enantioseparations data are summarized in Table 3 at corresponding detection wavelength. From the results, it was found the structure of analytes could have great influence on retention factor, selectivity and resolution. With an S/N of 1 ∼ 4, the selectivity and resolution of chiral aryl alcohols demonstrate obvious improvement when the para-substituent on the benzene ring were changed from H to fluoro/chloro/bromo, exhibiting the preferential inclusion sequence of organohalides was bromide > chloride > fluoride, which agrees with the previous findings [26]. Meanwhile, Aryl-OH-5 exhibited higher selectivity and resolution than Aryl-OH-3 and Aryl-OH-4 due to the additional double bond at chiral carbon center. With the para-substituent was changed from electron withdrawing –Cl (Aryl-OH-5) to electron donating –CH3 (Aryl-OH-6), both ˛ and Rs decreased. These behaviors demonstrated that the chiral discrimination of this CD clicked CSP was largely dependent on the inclusion complexation, – interaction, dipole–dipole interaction and steric hindrance. When the para-substituent was electron withdrawing, the aromatic ring of analyte exhibited a -acidic property, interacting strongly with naturally -basic phenylcarbamate group in CSP. Consequently, it contributed much to the chiral discrimination effect. For ArylOH-2, Aryl-OH-3 and Aryl-OH-4, their enantioseparations with our as-prepared CSP were much better than those with our earlier reported CSP with 50% lower column efficiency at the similar separation conditions [18].
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
123
Table 2 Effect of column temperature on enantioseparation of flavanone and 7-methoxyflavanone in methanol mobile phase. Temp. (◦ C)
10 20 30 40 50
flavanone
7-methoxyflavanone
k1
k2
˛
Rs
k1
k2
˛
Rs
0.41 0.31 0.24 0.20 0.17
0.96 0.67 0.48 0.37 0.29
2.35 2.17 1.99 1.83 1.67
6.4 5.75 4.74 3.71 2.78
0.91 0.65 0.49 0.38 0.30
2.51 1.64 1.11 0.79 0.57
2.76 2.50 2.28 2.07 1.88
7.51 7.53 6.98 5.95 4.69
Aryl-OH-8 could be baseline resolved while Aryl-OH-9 not, mainly due to the different location of chiral center. The results suggested when ␣-substituted naphthalene ring would form stronger steric repulsion when enter into CD cavity. Interestingly, the chiral separation for Aryl-OH-11 was significantly better than that for Aryl-OH-2, probably due to the more flexible and dense -electron of double bond than triple bond, which was more conducive to – conjugation. In terms of flavanoids, baseline separations were obtained for the non-substituted flavanoids. Besides, the obtained resolution values of all flavanoids except hesperetin and naringenin are superior to those in literature [26,27]. 7-Methoxyflavanone exhibited higher ˛ and Rs than 7-hydroxylflavanone with the substituent changed from methoxyl to hydroxyl. Meanwhile, the position of methoxy group would lead to the difference on enantioseparation when comparing 7-methoxyflavanone and 6-methoxyflavanone.
Interestingly, longer retention time could not guarantee better chiral separation. For instance, the naringenin and hesperetin could achieve the highest retention factors but their baseline separations were inaccessible in the studied separation condition. 3.4.2. Enantioseparation of adrenergic drugs and benzoin derivatives Adrenergic drugs can be classified into adrenergic agonists and antagonists (-blockers). The agonist drugs are mainly developed for the treatment of chronic obstructive pulmonary diseases and asthma. The ability of these compounds to interact with dual hydrogen bonding functionalities i.e., the hydroxyl group at the chiral center and a secondary amine in -position with the chiral selector has a large impact on their enantioseparation [28]. For adrenergic agonists such as isoprenaline and clenbuterol, both of them were well resolved (Table 3). Especially, the retention of isoprenaline
Fig. 4. Van’t Hoff equation for (a) flavanone and (b) 7-methxylflavanone, (c) the effect of temperature on column efficiency.
Fig. 5. Influence of different organic modifiers on enantioseparation of (a) Aryl-OH-5 (b) clenbuterol and (c) flavanone with 1% TEAA (pH 4) buffer.
124
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Table 3 Optimal enantioseparations results for 26 neutral and basic racemates at the maximum adsorption wavelength () of analytes, with data of three aryl alcohols compared with ref. [15]. Analytes
k1
k2
˛
Rs
(nm)
Condition
Aryl alcohols Aryl-OH-1 Aryl-OH-2 Aryl-OH-2 [15] Aryl-OH-3 Aryl-OH-3 [15] Aryl-OH-4 Aryl-OH-4 [15] Aryl-OH-5 Aryl-OH-6 Aryl-OH-7 Aryl-OH-8 Aryl-OH-9 Aryl-OH-10 Aryl-OH-11 Aryl-OH-12
1.37 1.11 0.86 1.38 1.06 1.46 1.40 2.47 2.68 2.09 18.70 – 3.95 5.23 4.28
1.46 1.28 0.95 1.85 1.28 2.24 1.77 4.21 3.91 3.46 20.36 – 4.46 11.34 4.71
1.06 1.15 1.11 1.34 1.20 1.54 1.26 1.70 1.46 1.66 1.09 – 1.13 2.17 1.10
0.61 1.65 0.91 3.05 1.41 4.38 1.64 6.21 2.84 3.57 1.02 – 1.12 4.81 0.95
205 225 225 225 225 225 225 225 250 225 250 220 250 250 250
II I III I III I III I I I II II IV I I
Flavanoids Flavanone 7-Methoxyflavanone 6-Methoxyflavanone 7-Hydroxylflavanone 4-Hydroxyflavanone Hesperetin Naringenin
1.10 1.96 3.15 0.96 0.82 9.14 7.23
1.92 4.10 5.54 1.10 1.25 9.42 7.65
1.74 2.09 1.76 1.15 1.52 1.03 1.06
5.58 6.82 6.79 1.77 2.97 0.36 0.46
250 270 225 215 225 225 225
V V V V V I I
Adrenergic drugs Isoprenaline Clenbuterol Propranolol Atenolol Acebutolol Benzoin 4-Chlorobenzoin
0.19 3.00 8.39 0.81 2.80 1.89 –
0.34 4.31 9.99 0.85 3.14 1.97 –
1.82 1.44 1.20 1.05 1.12 1.04 –
1.40 3.04 2.33 0.37 0.98 0.52 –
280 245 230 230 230 250 250
VI VI VI VI VI I I
Conditions: (I) MeOH/H2 O = 50/50 (v/v), 1.0 mL/min flow rate; (II) MeOH/H2 O = 20/80 (v/v), 1.0 mL/min flow rate; (III) MeOH/H2 O = 50/50 (v/v), 0.7 mL/min flow rate; (IV) MeOH/1%TEAA (pH 4) = 10/90 (v/v), 1.0 mL/min flow rate; (V) MeOH/H2 O = 80/20 (v/v), 1.0 mL/min flow rate; (VI) ACN/1% TEAA (pH 4) = 10/90, 1.0 mL/min flow rate.
was very short attributed to the presence of hydroxyl groups on phenyl ring. The introduction of hydroxyl groups may lead to weak inclusion complexation. The similar phenomenon was observed for 7-methoxyflavanone and 7-hydroxylflavanone. In terms of adrenergic antagonists, propranolol was the best resolved than atenolol and acebutolol probably due to the size of naphthalene ring could embed tightly with CD cavity than benzene ring. Benzoin could be better separated than 4-chlorobenzoin, which might be ascribed to different interactions caused by different substituent on both compounds with the selectors. 3.4.3. Effect of the content of organic modifiers Selected flavanoids were used to investigate the influence of methanol content on chiral separations. The methanol content in aqueous mobile phase was reduced from 90% to 60%. The Rs values of flavanone, 4-hydroxyflavanone and 7-hydroxyflavanone increased from 4.8 to 6.1, 2.8 to 3.6 and 1.75 to 2.2, respectively. However, the Rs of 6-methoxyflavanone (∼6.79) and 7-methoxyflavanone (6.7–6.8) remained almost unchanged at varied methanol contents. The former may be due to a lower content of methanol in the mobile phase could lead to an increase in solvent polarity, with less solvent molecules entering into the hydrophobic cavity of CD. This would lead to the enhancement of the hydrophobic interactions between the analytes and the CSPs. This latter phenomenon can be explained by an excessive increase in peak widths noticed at long retention times. The retention times of all flavanoids were found to increase with the decreased methanol content. Additionally, the effect of the content of acetonitrile was also evaluated using some aryl alcohols as samples. ACN concentration was varied from 20% to 50%. Different from methanol, both
resolution and retention factor increased with the decreased ACN content. The highest Rs of 7.2, 6.3, 5.1, 4.5, 4.2, 2.5 and 0.8 could be achieved for Aryl-OH-5, Aryl-OH-4, Aryl-OH-7, Aryl-OH-3, ArylOH-6, Aryl-OH-2 and Aryl-OH-1, respectively, with 20% ACN. This was probably attributable to the aprotic nature of ACN, leading to the enhanced affinity with the preferred sites in the hydrophobic CD cavity via the competition with the enantiomers [29]. 3.4.4. Effect of organic modifier type MeOH and ACN were used for the optimization of chiral separation in RPLC. As shown in Fig. 5a and b, ACN could lead to faster elution, while MeOH afforded higher chiral resolutions of them. For example, both baseline separation of Aryl-OH-5 and clenbuterol were achieved with MeOH/water and MeOH/1% TEAA (pH 4) as mobile phase, respectively, while relatively lower resolutions were obtained for the same composition of ACN/water and ACN/1% TEAA (pH 4). As showed in Fig. 5c, both faster elution and higher resolution were obtained for flavanone with MeOH (100%) than ACN/water (50/50). Similar retention behavior was observed due to the weaker displacing effect of MeOH than ACN, allowing for a stronger inclusion complexation between CD cavity and analytes [30]. Moreover, the polar interactions (hydrogen bonding, ionic interactions, etc.) formed with MeOH made a contribution to the chiral recognition [31]. 3.5. Effect of buffer on enatioseparation for basic compounds 3.5.1. Effect of buffer type In our study, basic compounds showed no sign of baseline separation with only water and organic solvent as mobile phase. Hence, a buffer solution was added into the mobile phase to enable or
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
125
Table 4 Effect of buffer type on the enantioseparation of basic compounds. ˛
Rs
Condition
Aryl-OH-10
3.95 2.24 2.33 1.37
4.46 2.59 2.45 1.46
1.13 1.16 1.05 1.07
1.12 1.33 0.38 0.61
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
Isoprenaline
0.45 0.18 0.19 0.01
1.34 0.78 0.34 0.14
2.97 4.43 1.82 1.05
1.87 2.10 1.40 1.85
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
13.13 4.68 8.36 4.34
18.98 6.47 9.99 5.62
1.45 1.38 1.20 1.05
1.73 1.24 2.33 1.87
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
Acebutolol
2.34 1.55 – –
2.60 1.73 – –
1.11 1.12 – –
0.92 0.84 – –
MeOH/TEAA = 30/70 MeOH/TEAP = 30/70 ACN/TEAA = 30/70 ACN/TEAP = 30/70
Clenbuterol
2.02 1.24 0.94 0.42
3.78 2.27 1.03 0.47
1.87 1.83 1.09 1.11
3.63 3.47 0.87 0.62
MeOH/TEAA = 30/70 MeOH/TEAP = 30/70 ACN/TEAA = 30/70 ACN/TEAP = 30/70
Atenolol
1.30 1.00 0.81 0.55
1.46 1.17 0.85 0.61
1.12 1.17 1.05 1.11
0.67 0.62 0.37 0.56
MeOH/TEAA = 5/95 MeOH/TEAP= 5/95 ACN/TEAA= 5/95 ACN/TEAP= 5/95
Analytes
Propranolol
k1
k2
All TEAA and TEAP buffers were 1% in volume concentration with pH 4.
enhance the enantioseparations [32]. In our cases, 1% TEAA (pH 4) and 1% TEAP (pH 4) were employed to investigated the influence of buffer type on the enantioseparation of basic analytes. The separation data summarized in Table 4 reveal the influence of MeOH or ACN as organic modifiers on the chiral separation of neutral and basic analytes. Firstly, MeOH/buffer afforded better selectivity and chiral resolution than ACN/buffer for selected analytes exception propranolol. For example, the k1 , ˛ and Rs of clenbuterol were 2.20, 1.87 and 3.63 with MeOH/buffer, while 0.94, 1.09 and 0.87 with ACN/buffer, respectively. The retention factors of all analytes with the use of 1% TEAA were larger than those with 1% TEAP. Meanwhile, TEAA buffer also delivered higher ˛ and Rs than 1% TEAP except Aryl-OH-10 and isoprenaline. It was widely acknowledged that TEAA can be accommodated into the CD cavity and compete
with analytes at the hydrogen bonding sites on CD [33,34], while the phosphate acids was almost inert with CDs [35]. Hence, the former may be more favor of chiral separation. Accordingly, TEAA buffer was chosen for our following study.
Fig. 6. The effect of TEAA (pH 4) content on the chiral resolution of propranolol.
Fig. 7. The effect of mobile phase pH on the enantioseparation of propranolol.
3.5.2. Effect of concentration of buffer The influence of TEAA (pH 4) concentration on the resolution of analytes was demonstrated with propranolol as model racemate by gradually increasing TEAA concentration from 0.1 to 4%
126
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Fig. 8. Typical chromatograms of selected racemates.
in buffer/ACN (80/20) mobile phase. The results are shown in Fig. 6. The Rs first increased from 1.57 at 0.1% to 1.74 at 1%, then decreased to 0.98 at 4%. In the case of separating basic analytes on crown ether-based CSP, Machida et al. [36] proposed two kinds of mechanisms to explain the effect of salt additives: one is the competition between the analytes and buffer cations, another is ion-pair formation between analyte cations and counter anions in the mobile phase. When the former interaction is dominant, retention factor decreases with the increase of salt’s concentration. Otherwise, it would increase. Similar separation mechanisms were proposed on another cavity-containing calixarene-based CSP [37]. In addition, Dai et al. [38,39] clarified the role of ion pairing in the retention of cation separations in RPLC on the basis of ionpair formation constants measured with capillary electrophoresis. Similar to CHIROBIOTIC V [40], -CD possesses both the cavity and the characteristics in reversed phase. The separation mechanisms mentioned above could thus be applied to explain the effect of salt concentration. In the range of 0.1 ∼ 1.0% TEAA, the ion-pair
formation is dominant, while the competition becomes a major factor for the retention at TEAA concentration higher than 1%. TEAA concentration of 1% was chosen for further study. 3.5.3. Effect of pH of buffer For basic analytes, buffer pH can affect their interactions with CSP due to the presence of amine groups. The effect of buffer pH on the chiral resolution was examined by changing 1% TEAA buffer in 4.0 ∼ 7.0 pH rage for water/ACN (80/20) mobile phase. For comparison, the separation in water/ACN (80/20) mobile phase without TEAA was also provided. As the pH-dependent resolutions of propranolol shown in Fig. 7, no enantioseparation was observed without the addition of TEAA, and the chiral resolution of propranolol increased with the decrease of pH of TEAA. In addition, longer retention time of basic analytes was observed when increasing TEAA pH, which may be explained with the interactions between analytes with the residual silanol groups of CSP. For acidic analytes such as carprofen and ibuprofen, however, good enantioseparation
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
can still be achieved at higher pH (
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Evaluation of perphenylcarbamated cyclodextrin clicked chiral stationary phase for enantioseparations in reversed phase high performance liquid chromatography Limin Pang a , Jie Zhou a , Jian Tang a,∗∗ , Siu-Choon Ng b , Weihua Tang a,∗ a Key Laboratory of Soft Chemistry and Functional Materials, Ministry of Education, Nanjing University of Science and Technology, Nanjing 210094, People’s Republic of China b Division of Chemical and Biomolecular Engineering, College of Engineering, Nanyang Technological University, 16 Nanyang Drive, Singapore 637722, Singapore
a r t i c l e
i n f o
Article history: Received 30 April 2014 Received in revised form 26 June 2014 Accepted 11 August 2014 Available online 17 August 2014 Keywords: Chiral separation Cyclodextrin Chiral stationary phase Reversed phase-HPLC
a b s t r a c t In this study, perphenylcarbamated cyclodextrin clicked chiral stationary phase (CSP) was developed with high column efficiency. The characteristics of the column were evaluated in terms of linearity, limit of detection and limit of quantification. The enantioselectivity of the as-prepared clicked CSP was explored with 26 recemates including aryl alcohols, flavanoids and adrenergic drugs in reversed phase highperformance liquid chromatography. The effect of separation parameters including flow rate, column temperature, organic modifier and buffer on the enantioselectivity of the clicked CSP was investigated in detail. The correlation study of the analytes structure and their chiral resolution revealed the great influence of analytes’ structure on the enantioseparations with cyclodextrin CSP. Methanol with 1% of triethylammonium acetate buffer (pH 4) was proved to be the best choice for the chiral separation of basic enantiomers. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Chiral compounds may significantly differ from each other in their biological, pharmacological and toxicological effects. A notorious example would be thalidomide which was debuted as a racemic drug [1]. Chiral separation has thus gained wide attention from both academia and industry. To date, chromatographic techniques have evolved as routine approaches for chiral separation, which feature high efficiency, quick analysis and precise resolutions. The most widely used techniques include high performance liquid chromatography (HPLC), gas chromatography, capillary electrophoresis, supercritical fluid chromatography and stimulated moving bed. By affording direct enantioseparation, HPLC coupled with chiral stationary phases (CSPs) has developed as one of the most important techniques for both detection of enantiomeric purity and quick preparation of pure enantiomers [2,3]. The CSPs include Pirkle-type,
∗ Corresponding author. Tel.: +86 25 8431 7311; fax: +86 25 8431 7311. ∗∗ Corresponding author. Tel.: +86 25 8431 7311. E-mail addresses: [email protected] (J. Tang), [email protected] (W. Tang). http://dx.doi.org/10.1016/j.chroma.2014.08.040 0021-9673/© 2014 Elsevier B.V. All rights reserved.
macrocyclic antibiotic, crown ethers, imprinted polymers and chiral ligand exchange CSPs as well as polysaccharide-, protein- and cyclodextrin (CD)-based CSPs [4–13]. The CD based CSPs have thus been extensively explored for the enantioseparation by functionalizing the CD rims to construct additional interactions such as – stacking, dipole–dipole, ion-paring, electrostatic and steric repulsive effects between analytes and the resulted CDs. Moreover, CD based CSPs are especially attractive for their versatility and durability under various conditions [8–13]. The CD-CSPs can be used in three modes including reversed phase (RP), normal phase (NP) and polar organic modes [13]. RP is the most popular mode in which the solute molecules are distributed between the relatively polar mobile phase and the non-polar stationary phase to afford separation. Under RP conditions, the enantioselectivity of CDs is dependent upon the analyte structure, CD’s type and the functionalities on the CD rims. The extent of host–guest inclusion generally depends on CD’s cavity dimension and the structure of the enantiomers. According to the concept of size-fit for inclusion complexation, better enantioselectivity for the CSP-enantiomers pairs generally occur when the size of hydrophobic portions of analytes matches with the CD cavity [14]. “Click chemistry” was first proposed by Sharpless and the [3 + 2] dipolar cycloaddition between azides/alkynes catalyzed by copper
120
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
OH
OH
F
Br
Cl
OH
NH2
OH
OH
F Aryl-OH-5
OH
O
O
OH
Aryl-OH-11
Aryl-OH-10
Aryl-OH-8
Aryl-OH-7
Aryl-OH-6
OH Aryl-OH-9
OH
NH2
Cl Aryl-OH-4
Aryl-OH-3
Aryl-OH-2
Aryl-OH-1
OH
OH
4-Chlorobenzoin
Benzoin
Aryl-OH-12
OH
Cl
OH O
O
O
O
O
O
O 7-Hydroxylflavanone
O
HO
O O
O
O
Flavanone
6-Methoxyflavanone
7-Methoxyflavanone
OH
OCH3 HO
O
OH
HO
O O
Acebutolol
OH
N H
H2N O
O OH
OH O Naringenin
Hesperetin
N H
O
O
OH O
O
4-Hydroxylflavanone
OH H N
Atenolol
Propranolol
OH H N
Cl H2N
OH H N
HO
Clenbuterol
O
O O Si O
N H
N N N
OR 6
silica gel
HO
Cl
N H
R=
Isoprenaline
H N
(OR)14 O
CD clicked CSP
Fig. 1. Structures of racemate studied.
can be regarded as the primary one [15]. By affording high selectivity towards reagents and excellent flexibility in various solvents, click reaction has dramatically burgeoning in the application of immobilization of ligands to solid and polymer supports [16,17]. As a continuation of our long-term research project, our recent studies have demonstrated that the CD based CSPs prepared via click chemistry can afford good enantioselectivities and durability in broad separation conditions in HPLC [5,7,9,13,18]. In order to explore the potential of triazolyl-linked CD CSPs for enantioseparation in HPLC, the perphenylcarbamated CD clicked CSP (structure in Fig. 1) was prepared to pack chiral column with greatly improved column efficiency and CD surface loading. The column efficiency greatly enhanced CSP afforded significant improvement in enantioselectivities to larger library of racemates in comparison to our first report [18]. We herein report the detailed characteristics of column. The enantioseparation capability of the clicked CD CSP was further evaluated with 26 racemates including aryl alcohols, aryl amines, flavanoids, and adrenergic drugs. The effect of the separation conditions including flow rate, column temperature, mobile phase composition, and organic modifiers on the enantioseparation of the model analytes were thoroughly investigated in RPLC. The correlation of analytes’ structures with their chiral resolutions by the clicked CD CSP was also described.
2. Experiments 2.1. Chemicals and materials All reagents such as -CD and phenyl isocyanate were purchased from Tokyo Chemical Industry (TCI, Japan). HPLC-grade acetonitrile (ACN) and methanol (MeOH) were purchased from Tedia (USA). HPLC-grade acetic acid, phosphoric acid and triethylamine (TEA) were obtained from J&K (Shanghai, China). Deionized water for the experiments was purified by Milli-Q system (Millipore, Bedford, MA, USA). All racemic analytes were purchased from the Meryer (Shanghai, China) (structures in Fig. 1). Kromasil ˚ was purchased from Eka Chemispherical silica gel (5 m, 100 A) cals (Bohus, Sweden). All other chemicals used were of analytical reagent grade without purification prior to use. 2.2. Apparatus All enantioseparations with CD clicked CSP were performed on an Agilent HPLC system, which was comprised of an Agilent 1260 system consisting of with G1315D diode array detection (DAD) system, G1329B quaternary pump, a G1331C automatic injector, a G1316A temperature controller and Agilent Chem Station data manager software (Agilent Technologies, Palo Alto, CA, USA).
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
121
Fig. 2. The standard curves of the first enantiomer (a) and the second enantiomer (b) of flavanone (c) the chiral resolutions of flavanone at different sample concentrations. Conditions: methanol as mobile phase, 250 nm detection wavelength.
2.3. HPLC procedures The mobile phases were prepared by mixing appropriate amounts of acetonitrile or methanol into water to separate neutral compounds, while for basic analytes, the mobile phases were prepared by adding 1% (v/v) triethylammonium acetate buffer (1% TEAA, pH adjusted with acetic acid) or 1% (v/v) triethylammonium phosphate buffer (1% TEAP, pH adjusted with phosphoric acid) in water. All mobile phases were filtered with 0.22 m Millipore membrane and ultrasonicated 30 min for degassing before use. The detection wavelength range was 200–300 nm. The racemic samples were dissolved in methanol to prepare stock solutions of 1–5 mg/mL. If needed, stock solutions were further diluted with methanol to prepare sample solutions of 100–500 g/mL. All sample solutions were stored at 4 ◦ C prior to use. All chromatograms were obtained from 10 L injection of sample solution with 1.0 mL/min flow rate of mobile phase under 25 ◦ C, unless otherwise specified. The following resolution parameters were determined: k1 and k2 were calculated using (t1 − t0 )/t0 and (t2 − t0 )/t0 , respectively, where t0 is the retention time at which the first baseline disturbance by the solvent peak occurred, t1 and t2 are the retention time of the first and the second enantiomer. The separation factor (˛) was calculated using k2 /k1 , while the chiral resolution (Rs ) was evaluated using the equation Rs = 1.18(t2 − t1 )/(w1 + w2 ), where w1 and w2 were the half-peak width of each enantiomer (based on USP standards). 2.4. Preparation of CD-CSP The perphenylcarbamated CD clicked CSP was prepared according to our recently reported procedure [18]. Starting from mono-6A -azido--CD, mono-6A -azido-perphenylcarbamated CD was obtained via a nucleophilic addition with phenyl isocyanate and a followed immobilization onto alkynyl functionalized silica gel via “click” chemistry. The elemental analysis results for the as-prepared CSP were 13.69% C, 0.546% N and 2.1% H. The immobilization efficiency was much improved than that in our first report (10.2% C, 1.90% N and 1.71% H) [18]. The surface loading of the CSP was calculated to be 0.021 mol/m2 with [C%/(12 × Nc × Ssilica )] × 1000, where C% is derived from elemental analysis of carbon, Nc is the carbon atom number per CD molecule (182) and Ssilica is the surface area of silica gel (300 m2 /g). The CD clicked CSP was then packed into columns by conventional slurry method, where the CSP suspension in methanol was packed into stainless steel columns (4.6 × 250 mm) under constant pressure (9000 psi) using packing system (Lab Alliance Scientific). The packing pressure was maintained for at least 0.5 h. Detected with toluene as marker sample and methanol as mobile phase under 25 ◦ C, the as-packed column afforded high column efficiency
of 12,000, which was 50% higher than our previous report [18]. The minimum plate height was determined as 20.9 m, corresponding to a theoretical plate number of 48,000 plates/meter. 3. Results and discussions 3.1. Characteristics of the column The linearity, limit of detection (LODs), limit of quantification (LOQs), relative standard deviation (RSD) and sample capacity of this column were evaluated with flavanone. The sample concentration increased from 1 g/mL to 800 g/mL at the fixed flow rate of 1.0 mL/min for methanol mobile phase. The standard curve by plotting sample peak area against sample concentration was obtained based upon the least square linear regression fit in the concentration range of 100–600 g/mL. As shown in Fig. 2a and b, impressive results were obtained with linearity coefficients R2 > 0.999 for both enantiomers of flavanone. Meanwhile, increased concentration could lead to a reduced chiral resolution probably due to the overloading of the CSP (Fig. 2c). LOD and LOQ for both enantiomers of flavanone were determined at an S/N (signal/noise ration) of 3 and 10, respectively. LOD was thus determined to be 4.12 × 10−12 mol and 4.46 × 10−12 mol, while LOQ to be 1.35 × 10−11 mol and 1.48 × 10−11 mol for the first and second enantiomer, respectively. Besides, the sample capacity (the sample amount corresponding to 10% decreased column efficiency) was 5.36 × 10−8 mol. Precision of the method was carried out with repetitive runs in RPLC via 10 injections of solutions (100 and 200 g/mL). Linearity parameters are summarized in Table 1, RSD values (%) obtained for both enantiomers were lower than 0.2% for the migration times and lower than 0.3% for the peak areas of flavanone enantiomers, revealing the excellent repeatability and reproducibility for the chiral column. 3.2. Effect of flow rate on enantioseparations In HPLC enantiomeric separations, flow rate is of great importance for both retention time and chiral resolutions. The flow rate was optimized by gradually increasing from 0.2 mL/min to 1.0 mL/min for methanol mobile phase (Fig. 3). The retention time and chiral resolutions of flavanone were found to decrease dramatically with the increment of methanol flow rate while the selectivity maintained unchanged. As revealed by Berthod et al. [19] that the formation of inclusion complexes was slower than the complex formed through the external interaction, the constant selectivity factor at varied flow rates indicated that inclusion complex did not play the decisive role for the chiral separation at this conditions. Taking selectivity, chiral resolution and retention time into
122
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Table 1 Characteristics of the column on the test of flavanone. Items
First enantiomer
Concentration (g/mL) Peak area, RSD% (n = 10) Migration time, RSD% (n = 10) Linear range (g/mL, injection 20 L) Linear equation Standard error Linear coefficient (R2 ) LOD (mol) LOQ (mol) Sample capacity (mol)
100 0.16 0.09 100–600 y = 0.01845x − 0.06053 Sa = 0.06583, Sb = 0.00017 0.99983 4.12 × 10−12 1.35 × 10−11 5.36 × 10−8
Second enantiomer 200 0.16 0.12
ln ˛ = −
100 0.21 0.14 100–600 y = 0.01838x − 0.06740 Sa = 0.08376, Sb = 0.00022 0.99958 4.46 × 10−12 1.48 × 10−11
H ◦ S ◦ + R·T R
200 0.19 0.14
(2)
where R is the gas constant, k is the retention factor, ϕ is the column phase ratio, H◦ and S◦ represent the respective changes of enthalpy and entropy when one enantiomer transfers from mobile phase to the CSP. Since the value of ϕ is often not known, the S◦ * values [[S◦ * = S◦ + Rln ϕ]] calculated from the intercepts of the plots via equation (1) are generally used. H◦ and S◦ represent the differences of S◦ and S◦ * for both enantiomers, respectively. The plot of ln ˛ against 1/T is thus a straight line if enantioselective interactions do not change over the temperature range studied [23]. According to Fig. 5a and b, a good linear relationship between ln ˛ and 1/T was observed with correlation coefficients over 0.99, which shows both the conformation of CD CSP and the enantioselective interactions attributed for the enantioseparation were unaffected at varied column temperature. As the values of H◦ and S◦ were always negative, the chiral separation was indicated as enthalpy driven process [23–25]. 3.4. Evaluation of enantioseparations ability of CSP
Fig. 3. The effect of flow rate on the enantioseparation of flavanone in methanol mobile phase detected at 250 nm.
consideration, the optimal flow rate of 1.0 mL/min was selected for the following study. 3.3. Effect of column temperature on enantioseparations The column temperature is also important consideration for successful enantioseparations. Three aspects are thought to be influenced by column temperature: (i) mass transfer in mobile phase, (ii) the interactions between CSP and the enantiomer, and (iii) the transfer of an analyte between CSP and mobile phase [20,21]. The influence of temperature on the enantiosepration of CD clicked CSP was studied with flavanone and 7-methoxyflavanone as model analytes in column temperature range of 10–50 ◦ C with methanol as mobile phase. As the enantioseparation data shown in Table 2, k1 , k2 , ˛ and Rs all decreased when the column temperature increased (Fig. 4). This can be attributed to not only the increasing mass transfer of solutes in the mobile phase, but also the interactions between the enantiomers and the mobile phases, which were affected by the thermodynamic properties of the analytes on CSPs [22]. To investigate the thermodynamic functions of enantioselective adsorption, Van’t Hoff plots were interpreted in terms of the mechanistic aspects of chiral recognition: ln k = −
H ◦ H ◦ S ◦ S ◦ ∗ + + ln ϕ = − + RT R RT R
(1)
3.4.1. Enantioseparations of racemic aryl alcohols and flavanoids Several racemic aryl alcohols and flavanoids have been tried on the CD clicked CSP in our previous report [18]. In order to explore the potential of enantioselectivities of the perphenylcarbamated CD clicked CSP, the resolutions of 19 racemates including 10 aryl alcohols, 2 aryl amines and 7 flavanoids were optimized at different separation conditions. The optimal enantioseparations data are summarized in Table 3 at corresponding detection wavelength. From the results, it was found the structure of analytes could have great influence on retention factor, selectivity and resolution. With an S/N of 1 ∼ 4, the selectivity and resolution of chiral aryl alcohols demonstrate obvious improvement when the para-substituent on the benzene ring were changed from H to fluoro/chloro/bromo, exhibiting the preferential inclusion sequence of organohalides was bromide > chloride > fluoride, which agrees with the previous findings [26]. Meanwhile, Aryl-OH-5 exhibited higher selectivity and resolution than Aryl-OH-3 and Aryl-OH-4 due to the additional double bond at chiral carbon center. With the para-substituent was changed from electron withdrawing –Cl (Aryl-OH-5) to electron donating –CH3 (Aryl-OH-6), both ˛ and Rs decreased. These behaviors demonstrated that the chiral discrimination of this CD clicked CSP was largely dependent on the inclusion complexation, – interaction, dipole–dipole interaction and steric hindrance. When the para-substituent was electron withdrawing, the aromatic ring of analyte exhibited a -acidic property, interacting strongly with naturally -basic phenylcarbamate group in CSP. Consequently, it contributed much to the chiral discrimination effect. For ArylOH-2, Aryl-OH-3 and Aryl-OH-4, their enantioseparations with our as-prepared CSP were much better than those with our earlier reported CSP with 50% lower column efficiency at the similar separation conditions [18].
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
123
Table 2 Effect of column temperature on enantioseparation of flavanone and 7-methoxyflavanone in methanol mobile phase. Temp. (◦ C)
10 20 30 40 50
flavanone
7-methoxyflavanone
k1
k2
˛
Rs
k1
k2
˛
Rs
0.41 0.31 0.24 0.20 0.17
0.96 0.67 0.48 0.37 0.29
2.35 2.17 1.99 1.83 1.67
6.4 5.75 4.74 3.71 2.78
0.91 0.65 0.49 0.38 0.30
2.51 1.64 1.11 0.79 0.57
2.76 2.50 2.28 2.07 1.88
7.51 7.53 6.98 5.95 4.69
Aryl-OH-8 could be baseline resolved while Aryl-OH-9 not, mainly due to the different location of chiral center. The results suggested when ␣-substituted naphthalene ring would form stronger steric repulsion when enter into CD cavity. Interestingly, the chiral separation for Aryl-OH-11 was significantly better than that for Aryl-OH-2, probably due to the more flexible and dense -electron of double bond than triple bond, which was more conducive to – conjugation. In terms of flavanoids, baseline separations were obtained for the non-substituted flavanoids. Besides, the obtained resolution values of all flavanoids except hesperetin and naringenin are superior to those in literature [26,27]. 7-Methoxyflavanone exhibited higher ˛ and Rs than 7-hydroxylflavanone with the substituent changed from methoxyl to hydroxyl. Meanwhile, the position of methoxy group would lead to the difference on enantioseparation when comparing 7-methoxyflavanone and 6-methoxyflavanone.
Interestingly, longer retention time could not guarantee better chiral separation. For instance, the naringenin and hesperetin could achieve the highest retention factors but their baseline separations were inaccessible in the studied separation condition. 3.4.2. Enantioseparation of adrenergic drugs and benzoin derivatives Adrenergic drugs can be classified into adrenergic agonists and antagonists (-blockers). The agonist drugs are mainly developed for the treatment of chronic obstructive pulmonary diseases and asthma. The ability of these compounds to interact with dual hydrogen bonding functionalities i.e., the hydroxyl group at the chiral center and a secondary amine in -position with the chiral selector has a large impact on their enantioseparation [28]. For adrenergic agonists such as isoprenaline and clenbuterol, both of them were well resolved (Table 3). Especially, the retention of isoprenaline
Fig. 4. Van’t Hoff equation for (a) flavanone and (b) 7-methxylflavanone, (c) the effect of temperature on column efficiency.
Fig. 5. Influence of different organic modifiers on enantioseparation of (a) Aryl-OH-5 (b) clenbuterol and (c) flavanone with 1% TEAA (pH 4) buffer.
124
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Table 3 Optimal enantioseparations results for 26 neutral and basic racemates at the maximum adsorption wavelength () of analytes, with data of three aryl alcohols compared with ref. [15]. Analytes
k1
k2
˛
Rs
(nm)
Condition
Aryl alcohols Aryl-OH-1 Aryl-OH-2 Aryl-OH-2 [15] Aryl-OH-3 Aryl-OH-3 [15] Aryl-OH-4 Aryl-OH-4 [15] Aryl-OH-5 Aryl-OH-6 Aryl-OH-7 Aryl-OH-8 Aryl-OH-9 Aryl-OH-10 Aryl-OH-11 Aryl-OH-12
1.37 1.11 0.86 1.38 1.06 1.46 1.40 2.47 2.68 2.09 18.70 – 3.95 5.23 4.28
1.46 1.28 0.95 1.85 1.28 2.24 1.77 4.21 3.91 3.46 20.36 – 4.46 11.34 4.71
1.06 1.15 1.11 1.34 1.20 1.54 1.26 1.70 1.46 1.66 1.09 – 1.13 2.17 1.10
0.61 1.65 0.91 3.05 1.41 4.38 1.64 6.21 2.84 3.57 1.02 – 1.12 4.81 0.95
205 225 225 225 225 225 225 225 250 225 250 220 250 250 250
II I III I III I III I I I II II IV I I
Flavanoids Flavanone 7-Methoxyflavanone 6-Methoxyflavanone 7-Hydroxylflavanone 4-Hydroxyflavanone Hesperetin Naringenin
1.10 1.96 3.15 0.96 0.82 9.14 7.23
1.92 4.10 5.54 1.10 1.25 9.42 7.65
1.74 2.09 1.76 1.15 1.52 1.03 1.06
5.58 6.82 6.79 1.77 2.97 0.36 0.46
250 270 225 215 225 225 225
V V V V V I I
Adrenergic drugs Isoprenaline Clenbuterol Propranolol Atenolol Acebutolol Benzoin 4-Chlorobenzoin
0.19 3.00 8.39 0.81 2.80 1.89 –
0.34 4.31 9.99 0.85 3.14 1.97 –
1.82 1.44 1.20 1.05 1.12 1.04 –
1.40 3.04 2.33 0.37 0.98 0.52 –
280 245 230 230 230 250 250
VI VI VI VI VI I I
Conditions: (I) MeOH/H2 O = 50/50 (v/v), 1.0 mL/min flow rate; (II) MeOH/H2 O = 20/80 (v/v), 1.0 mL/min flow rate; (III) MeOH/H2 O = 50/50 (v/v), 0.7 mL/min flow rate; (IV) MeOH/1%TEAA (pH 4) = 10/90 (v/v), 1.0 mL/min flow rate; (V) MeOH/H2 O = 80/20 (v/v), 1.0 mL/min flow rate; (VI) ACN/1% TEAA (pH 4) = 10/90, 1.0 mL/min flow rate.
was very short attributed to the presence of hydroxyl groups on phenyl ring. The introduction of hydroxyl groups may lead to weak inclusion complexation. The similar phenomenon was observed for 7-methoxyflavanone and 7-hydroxylflavanone. In terms of adrenergic antagonists, propranolol was the best resolved than atenolol and acebutolol probably due to the size of naphthalene ring could embed tightly with CD cavity than benzene ring. Benzoin could be better separated than 4-chlorobenzoin, which might be ascribed to different interactions caused by different substituent on both compounds with the selectors. 3.4.3. Effect of the content of organic modifiers Selected flavanoids were used to investigate the influence of methanol content on chiral separations. The methanol content in aqueous mobile phase was reduced from 90% to 60%. The Rs values of flavanone, 4-hydroxyflavanone and 7-hydroxyflavanone increased from 4.8 to 6.1, 2.8 to 3.6 and 1.75 to 2.2, respectively. However, the Rs of 6-methoxyflavanone (∼6.79) and 7-methoxyflavanone (6.7–6.8) remained almost unchanged at varied methanol contents. The former may be due to a lower content of methanol in the mobile phase could lead to an increase in solvent polarity, with less solvent molecules entering into the hydrophobic cavity of CD. This would lead to the enhancement of the hydrophobic interactions between the analytes and the CSPs. This latter phenomenon can be explained by an excessive increase in peak widths noticed at long retention times. The retention times of all flavanoids were found to increase with the decreased methanol content. Additionally, the effect of the content of acetonitrile was also evaluated using some aryl alcohols as samples. ACN concentration was varied from 20% to 50%. Different from methanol, both
resolution and retention factor increased with the decreased ACN content. The highest Rs of 7.2, 6.3, 5.1, 4.5, 4.2, 2.5 and 0.8 could be achieved for Aryl-OH-5, Aryl-OH-4, Aryl-OH-7, Aryl-OH-3, ArylOH-6, Aryl-OH-2 and Aryl-OH-1, respectively, with 20% ACN. This was probably attributable to the aprotic nature of ACN, leading to the enhanced affinity with the preferred sites in the hydrophobic CD cavity via the competition with the enantiomers [29]. 3.4.4. Effect of organic modifier type MeOH and ACN were used for the optimization of chiral separation in RPLC. As shown in Fig. 5a and b, ACN could lead to faster elution, while MeOH afforded higher chiral resolutions of them. For example, both baseline separation of Aryl-OH-5 and clenbuterol were achieved with MeOH/water and MeOH/1% TEAA (pH 4) as mobile phase, respectively, while relatively lower resolutions were obtained for the same composition of ACN/water and ACN/1% TEAA (pH 4). As showed in Fig. 5c, both faster elution and higher resolution were obtained for flavanone with MeOH (100%) than ACN/water (50/50). Similar retention behavior was observed due to the weaker displacing effect of MeOH than ACN, allowing for a stronger inclusion complexation between CD cavity and analytes [30]. Moreover, the polar interactions (hydrogen bonding, ionic interactions, etc.) formed with MeOH made a contribution to the chiral recognition [31]. 3.5. Effect of buffer on enatioseparation for basic compounds 3.5.1. Effect of buffer type In our study, basic compounds showed no sign of baseline separation with only water and organic solvent as mobile phase. Hence, a buffer solution was added into the mobile phase to enable or
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
125
Table 4 Effect of buffer type on the enantioseparation of basic compounds. ˛
Rs
Condition
Aryl-OH-10
3.95 2.24 2.33 1.37
4.46 2.59 2.45 1.46
1.13 1.16 1.05 1.07
1.12 1.33 0.38 0.61
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
Isoprenaline
0.45 0.18 0.19 0.01
1.34 0.78 0.34 0.14
2.97 4.43 1.82 1.05
1.87 2.10 1.40 1.85
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
13.13 4.68 8.36 4.34
18.98 6.47 9.99 5.62
1.45 1.38 1.20 1.05
1.73 1.24 2.33 1.87
MeOH/TEAA = 10/90 MeOH/TEAP = 10/90 ACN/TEAA = 10/90 ACN/TEAP = 10/90
Acebutolol
2.34 1.55 – –
2.60 1.73 – –
1.11 1.12 – –
0.92 0.84 – –
MeOH/TEAA = 30/70 MeOH/TEAP = 30/70 ACN/TEAA = 30/70 ACN/TEAP = 30/70
Clenbuterol
2.02 1.24 0.94 0.42
3.78 2.27 1.03 0.47
1.87 1.83 1.09 1.11
3.63 3.47 0.87 0.62
MeOH/TEAA = 30/70 MeOH/TEAP = 30/70 ACN/TEAA = 30/70 ACN/TEAP = 30/70
Atenolol
1.30 1.00 0.81 0.55
1.46 1.17 0.85 0.61
1.12 1.17 1.05 1.11
0.67 0.62 0.37 0.56
MeOH/TEAA = 5/95 MeOH/TEAP= 5/95 ACN/TEAA= 5/95 ACN/TEAP= 5/95
Analytes
Propranolol
k1
k2
All TEAA and TEAP buffers were 1% in volume concentration with pH 4.
enhance the enantioseparations [32]. In our cases, 1% TEAA (pH 4) and 1% TEAP (pH 4) were employed to investigated the influence of buffer type on the enantioseparation of basic analytes. The separation data summarized in Table 4 reveal the influence of MeOH or ACN as organic modifiers on the chiral separation of neutral and basic analytes. Firstly, MeOH/buffer afforded better selectivity and chiral resolution than ACN/buffer for selected analytes exception propranolol. For example, the k1 , ˛ and Rs of clenbuterol were 2.20, 1.87 and 3.63 with MeOH/buffer, while 0.94, 1.09 and 0.87 with ACN/buffer, respectively. The retention factors of all analytes with the use of 1% TEAA were larger than those with 1% TEAP. Meanwhile, TEAA buffer also delivered higher ˛ and Rs than 1% TEAP except Aryl-OH-10 and isoprenaline. It was widely acknowledged that TEAA can be accommodated into the CD cavity and compete
with analytes at the hydrogen bonding sites on CD [33,34], while the phosphate acids was almost inert with CDs [35]. Hence, the former may be more favor of chiral separation. Accordingly, TEAA buffer was chosen for our following study.
Fig. 6. The effect of TEAA (pH 4) content on the chiral resolution of propranolol.
Fig. 7. The effect of mobile phase pH on the enantioseparation of propranolol.
3.5.2. Effect of concentration of buffer The influence of TEAA (pH 4) concentration on the resolution of analytes was demonstrated with propranolol as model racemate by gradually increasing TEAA concentration from 0.1 to 4%
126
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
Fig. 8. Typical chromatograms of selected racemates.
in buffer/ACN (80/20) mobile phase. The results are shown in Fig. 6. The Rs first increased from 1.57 at 0.1% to 1.74 at 1%, then decreased to 0.98 at 4%. In the case of separating basic analytes on crown ether-based CSP, Machida et al. [36] proposed two kinds of mechanisms to explain the effect of salt additives: one is the competition between the analytes and buffer cations, another is ion-pair formation between analyte cations and counter anions in the mobile phase. When the former interaction is dominant, retention factor decreases with the increase of salt’s concentration. Otherwise, it would increase. Similar separation mechanisms were proposed on another cavity-containing calixarene-based CSP [37]. In addition, Dai et al. [38,39] clarified the role of ion pairing in the retention of cation separations in RPLC on the basis of ionpair formation constants measured with capillary electrophoresis. Similar to CHIROBIOTIC V [40], -CD possesses both the cavity and the characteristics in reversed phase. The separation mechanisms mentioned above could thus be applied to explain the effect of salt concentration. In the range of 0.1 ∼ 1.0% TEAA, the ion-pair
formation is dominant, while the competition becomes a major factor for the retention at TEAA concentration higher than 1%. TEAA concentration of 1% was chosen for further study. 3.5.3. Effect of pH of buffer For basic analytes, buffer pH can affect their interactions with CSP due to the presence of amine groups. The effect of buffer pH on the chiral resolution was examined by changing 1% TEAA buffer in 4.0 ∼ 7.0 pH rage for water/ACN (80/20) mobile phase. For comparison, the separation in water/ACN (80/20) mobile phase without TEAA was also provided. As the pH-dependent resolutions of propranolol shown in Fig. 7, no enantioseparation was observed without the addition of TEAA, and the chiral resolution of propranolol increased with the decrease of pH of TEAA. In addition, longer retention time of basic analytes was observed when increasing TEAA pH, which may be explained with the interactions between analytes with the residual silanol groups of CSP. For acidic analytes such as carprofen and ibuprofen, however, good enantioseparation
L. Pang et al. / J. Chromatogr. A 1363 (2014) 119–127
can still be achieved at higher pH (
